This page intentionally left blank
LATENT INHIBITION: COGNITION, NEUROSCIENCE AND APPLICATIONS TO SCHIZOPHRENIA
Latent inhibition is a phenomenon by which exposure to an irrelevant stimulus impedes the acquisition or expression of conditioned associations with that stimulus. Latent inhibition, an integral part of the learning process, is observed in many species. This comprehensive collection of studies of latent inhibition, from a variety of disciplines including behavioral/cognitive psychology, neuroscience, and genetics, focuses on abnormal latent inhibition effects in schizophrenic patients and schizotypal normals. Amongst other things, the book addresses questions such as: is latent inhibition an acquisition or performance deficit? What is the relationship of latent inhibition to habituation, extinction, and learned irrelevance? Does reduced latent inhibition predict creativity? What are the neural substrates, pharmacology, and genetics of latent inhibition? What do latent inhibition research and theories tell us about schizophrenia? This book provides a single point of reference for neuroscience researchers, graduate students, and professionals, such as psychologists and psychiatrists. R.E. LU B O W is a Professor in the Department of Psychology at Tel Aviv University, Israel. His research and theoretical interests focus on normal attentional processes in animal and human learning, and on their disruption as a result of psychopathology, particularly in schizophrenia. He is a Fellow of the American Psychological Association and of the Association for Psychological Science. IN A W E I N E R is a Professor in the Department of Psychology at Tel Aviv University, Israel. She is one of the leading researchers in the study of the pharmacological and neurophysiological basis of latent inhibition, and in particular its implications for the understanding and treatment of schizophrenia.
LATENT INHIBITION: COGNITION, NEUROSCIENCE AND APPLICATIONS TO SCHIZOPHRENIA Edited by Professor R. E. LUBOW and Professor INA WEINER
CAMBRIDGE UNIVERSITY PRESS
Cambridge, New York, Melbourne, Madrid, Cape Town, Singapore, São Paulo, Delhi, Dubai, Tokyo Cambridge University Press The Edinburgh Building, Cambridge CB2 8RU, UK Published in the United States of America by Cambridge University Press, New York www.cambridge.org Information on this title: www.cambridge.org/9780521517331 © Cambridge University Press 2010 This publication is in copyright. Subject to statutory exception and to the provision of relevant collective licensing agreements, no reproduction of any part may take place without the written permission of Cambridge University Press. First published in print format 2010 ISBN-13
978-0-511-72940-9
eBook (NetLibrary)
ISBN-13
978-0-521-51733-1
Hardback
Cambridge University Press has no responsibility for the persistence or accuracy of urls for external or third-party internet websites referred to in this publication, and does not guarantee that any content on such websites is, or will remain, accurate or appropriate.
Contents
page viii xi
List of contributors Preface Lubow & Weiner
1
A short history of latent inhibition research
1
Lubow
Current topics in latent inhibition research Behavior and cognition
2
Latent inhibition and extinction: their signature phenomena and the role of prediction error
23
Westbrook & Bouton
3
Inter-stage context and time as determinants of latent inhibition
40
De la Casa & Pinen˜o
4
Latent inhibition: acquisition or performance deficit?
62
Escobar & Miller
5
Latent inhibition and learned irrelevance in human contingency learning
94
Le Pelley & Schmidt-Hansen
6
Associative and nonassociative processes in latent inhibition: an elaboration of the Pearce–Hall model
114
Hall & Rodriguez
7
From latent inhibition to retrospective revaluation: an attentional-associative model
137
Schmajuk
8
Latent inhibition and habituation: evaluation of an associative analysis Honey, Iordanova & Good
v
163
vi
Contents
9
Latent inhibition and creativity
183
Carson Neurobiology
10
The phylogenetic distribution of latent inhibition
201
Lubow
11
The genetics of latent inhibition: studies of inbred and mutant mice
225
Lipina & Roder
12
A comparison of mechanisms underlying the CS–US association and the CS–nothing association
252
Gould
13
The pharmacology of latent inhibition and its relevance to schizophrenia
276
Weiner & Arad
14
Parahippocampal region–dopaminergic neuron relationships in latent inhibition
319
Louilot, Jeanblanc, Peterschmitt & Meyer
15
Latent inhibition and other salience modulation effects: same neural substrates?
342
Cassaday & Moran
16
What the brain teaches us about latent inhibition (LI): the neural substrates of the expression and prevention of LI
372
Weiner Latent inhibition and schizophrenia
17
Latent inhibition in schizophrenia and schizotypy: a review of the empirical literature
419
Kumari & Ettinger
18
A cautionary note about latent inhibition in schizophrenia: are we ignoring relevant information?
448
Swerdlow
19
Latent inhibition as a function of anxiety and stress: implications for schizophrenia
457
Braunstein-Bercovitz
20
Nicotinic modulation of attentional deficits in schizophrenia Schnur & Hoffman
477
Contents
21
Latent inhibition and schizophrenia: the ins and outs of context
vii
500
Lubow
Summary and conclusions 22 Issues in latent inhibition research and theory: an overview
531
Lubow & Weiner
Index
558
Contributors
M. Arad Department of Psychology, Tel Aviv University, Ramat Aviv 69978, Israel M. E. Bouton Department of Psychology, University of Vermont, Burlington, VT 05405, USA Hedva Braunstein-Bercovitz School of Behavioral Sciences, The Academic College of Tel Aviv-Yafo, Rabenu Yeruham St., P.O.B. 8401, Yafo 61083, Israel Shelley Carson Department of Psychology, Harvard University, 33 Kirkland Street, Cambridge, MA 02138, USA Helen Cassaday School of Psychology, University of Nottingham, University Park, Nottingham NG7 2RD, UK L. Gonzalo De la Casa Department of Experimental Psychology, Seville University, C/ Camilo Jose Cela, s/n 41018 Sevilla, Spain Martha Escobar Department of Psychology, Auburn University, Auburn, AL 36849-5214, USA Ulrich Ettinger Departments of Psychiatry and Psychology, Ludwig-Maximilians-University of Munich, Munich, Germany M. Good School of Psychology, Cardiff University, Tower Building, Park Place, Cardiff CF10 3AT, UK
viii
List of contributors
ix
T. Gould Department of Psychology, Neuroscience Program, Temple University, Philadelphia, PA 19122, USA Geoffrey Hall Department of Psychology, University of York, York YO10 5DD, UK Allison Hoffman National Institute on Drug Abuse, National Institutes of Health, Department of Health and Human Services, 6001 Executive Blvd, Bethesda, MD 20892, USA R. C. Honey School of Psychology, Cardiff University, Tower Building, Park Place, Cardiff CF10 3AT, UK Mihaela D. Iordanova School of Psychology, Cardiff University, Tower Building, Park Place, Cardiff CF10 3AT, UK J. Jeanblanc Gallo Research Center, 5858 Horton Street, Emeryville, Suite 200, CA 94608, USA Veena Kumari Department of Psychology, Institute of Psychiatry, King’s College London, London, UK M. E. Le Pelley School of Psychology, Cardiff University, Tower Building, Park Place, Cardiff CF10 3AT, UK Tatiana Lipina Samuel Lunenfeld Research Institute at Mount Sinai Hospital, 600 University Avenue, room 860, Toronto, Ontario, M5G 1X5, Canada A. Louilot INSERM U 666 and Institute of Physiology, Strasbourg University, Faculty of Medicine, 11 rue Humann, 67085 Strasbourg CEDEX, France R. E. Lubow Department of Psychology, Tel Aviv University, Ramat Aviv 69978, Israel F. Meyer INSERM U 666 and Institute of Physiology, Strasbourg University, Faculty of Medicine, 11 rue Humann, 67085 Strasbourg CEDEX, France Ralph Miller Department of Psychology, SUNY Binghamton, Binghamton, NY 13902, USA
x
List of contributors
Paula Moran School of Psychology, University of Nottingham, University Park, Nottingham NG7 2RD, UK Y. Peterschmitt Center for Brain Research, Medical University Vienna, A-1090 Vienna, Austria Oskar Pinen˜o Department of Psychology, Hofstra University, Hempstead, NY 11550, USA John Roder Samuel Lunenfeld Research Institute at Mount Sinai Hospital, 600 University Avenue, room 860, Toronto, Ontario, M5G 1X5, Canada G. Rodriguez Faculty of Psychology, University of the Basque Country, San Sebastian 20009, Spain Nestor Schmajuk Department of Psychology and Neuroscience, Duke University, Durham, NC 27708, USA Mia Schmidt-Hansen School of Psychology, Cardiff University, Cardiff, UK Paul Schnur National Institute on Drug Abuse, National Institutes of Health, Department of Health and Human Services, 6001 Executive Blvd, Bethesda, MD 20892, USA Neal Swerdlow University of California, San Diego, Department of Psychiatry, 9500 Gilman Drive, La Jolla, CA 92093, USA Ina Weiner Department of Psychology, Tel Aviv University, Ramat Aviv 69978, Israel R. F. Westbrook School of Psychology, University New South Wales, Sydney, NSW 2052, Australia
Preface
Latent inhibition (LI), a phenomenon that reflects an outcome from the processing of irrelevant stimuli, has been of interest to the research community for five decades. And, if anything, its appeal and influence is growing. To mark the fiftieth anniversary of the publication of the first LI experiment, we asked a number of leading scientists to contribute chapters to a volume that would cover the broadest possible range of recent developments in LI research and theory. Amongst other things, we were interested in showing how a simple behavioral experiment conducted so very many years ago on sheep and goats has led to a burgeoning research enterprise that has enlisted many neuroscience disciplines, including those in physiology, chemistry, pharmacology, and genetics, and has branched out from academic concerns with learning theory to theoretical interests and applications related to schizophrenia. Unfortunately, many people working in research-specific areas find it difficult to keep abreast of the broad cross-disciplinary advances in LI, often directly relevant to their own interests. As an example, there is considerable research on the pharmacological, molecular, and cellular mechanisms underlying LI, and LI is a popular paradigm for studying the neurobiological basis of schizophrenia. However, many of the neuroscientists in this field are unaware of the cognitive/information processing theories underlying the LI effect. The opposite is also true; behavioral/cognitive theorists are often uninformed about advances in the neurophysiology of LI. The present volume provides these researchers with a comprehensive survey of current LI research and theory, from genetics to behavior, thereby strengthening the particularist approach to research as well as fostering an interdisciplinary methodology. Fifty years ago no one would have envisaged, or even come close to foreseeing, the possible fruits of that first demonstration of a latent inhibition effect. At one time, United States Senator William Proxmire awarded an annual prize, the Golden Fleece Award, for research projects most likely to waste taxpayers’ money. Surely, the effects of preexposing a to-be conditioned stimulus on the classical conditioning of leg flexion in goats and sheep, although fortunately preceding Senator Proxmire’s rancor, might have appeared to be a worthy candidate for the award (all the more so,
xi
xii
Preface
since many of the subjects were indeed fleeced). The present volume, at least in retrospect, hopefully would have discounted such a nomination. The present volume is deeply indebted to the students and colleagues with whom we have worked with over these many years, several of whom have contributed chapters, and to the research support from a number of foundations and government agencies in the United States and Israel. In no particular order, we would like to convey our gratitude and indebtedness to Hedva Braunstein-Bercovitz, Gonzalo De la Casa, Joram Feldon, Oren Kaplan, and Paul Schnur, and to the US National Institutes of Health, Israel Foundation Trustees, Psychobiology Foundation, Israel National Academy of Sciences, and Scottish Rite Schizophrenia Research Program. The Editors
1 A short history of latent inhibition research R. E. Lubow
The first latent inhibition (LI) paper was published 50 years ago (Lubow & Moore, 1959), and the present book marks that anniversary. As such, it offers a convenient time for providing a historical perspective to a phenomenon that was born by accident, barely survived the first several post-parturitional years, and yet developed into a flourishing research enterprise, with activities cutting across such diverse fields as learning theory, schizophrenia, and even creativity. Indeed, in the weekly episodes of “Prison Break” on American TV, the concept of LI has even reached prime-time television. In spite of the relatively widespread use of the LI paradigm in the laboratory, and, in particular, because of its adoption in research areas that are far removed from its origins, the present editors felt that there was a need to acquaint the larger audience with both the history and recent advances in LI research and theory. Before describing the serendipitous discovery of LI, this apparently simple, yet ubiquitous, phenomenon requires a definition and a description of its adaptive function. Specifically, LI is a name for the decrease in the ability to acquire or express a new association to a stimulus that has previously received passive, non-reinforced preexposures, as compared to a stimulus that is either novel (not preexposed) or one that has been reinforced or attended. Importantly, LI is not a process. It is an effect that, as will be seen, is in search of a process that generates it. In short, at least according to one explanation of LI, some types of familiarity may not breed contempt, but rather an adaptive neglect. And, for that, we have much to be thankful! When one considers that at any wakeful moment there are countless packets of stimuli that flood our sensory surfaces, it is astonishing that we are so successful at separating the relevant from the irrelevant, the trivial from the potentially useful. Undoubtedly, without the means to accomplish this partition, our lives would be unbearable, buried in an unsurvivable, undifferentiated Jamesian blooming, buzzing confusion, “. . . the consciousness of every creature would be a gray chaotic indiscriminateness, impossible for us even to conceive” (James, 1890/1950, pp. 402–403). Latent Inhibition: Cognition, Neuroscience and Applications to Schizophrenia, ed. R. E. Lubow and Ina Weiner. Published by Cambridge University Press # Cambridge University Press 2010.
1
2
R. E. Lubow
The manner in which organisms accomplish the survival-essential task of selecting the important from the insignificant has been a central concern of psychology over the entire course of its history. In the early years of modern experimental psychology, Gestalt psychologists appealed to the innate ability of the brain to organize information, and behaviorists to reinforcement-based learning. More recently, cognitive psychologists have placed the burden of stimulus selection on various somewhat homuncular models of selective attention (many of which tell us more about how the mind of the scientist works rather than the minds of the subjects that generated the data). Yet, from all of this activity, there appears to be one basic conclusion. As one might expect for a fundamental condition for survival, selection is supported by a variety of mechanisms, some probably complementary and overlapping, and others operating quite independently. Nevertheless, until fifty years ago, with the exception of studies of response habituation, processing of unattended, irrelevant events and how they fostered adaptive behavior was largely ignored, particularly in regard to their subsequent fate when they became relevant. How then did researchers in psychology come to attend to the unattended?
The early years (1959–1973) The idea for the initial experiments, the unexpected results of which came to be called “latent inhibition”, grew out of my interest in associative learning theory, which in the late 1950s was still identified with the competing models of Hull (1943, 1952) and Tolman (1932). At that time, I was a graduate student at Cornell University, working as an assistant to Howard Liddell at his animal behavior farm, an institution devoted to using classical Pavlovian procedures to induce so-called experimental neuroses in goats and sheep (and earlier, also in dogs and pigs). This was a particularly gloomy period for me, not only because of animal welfare concerns, but also because of my inability to observe the pathological behaviors that were apparently so clearly visible to other members of the laboratory (the sheep always seemed forlorn). In that era, five decades ago, the exemplar of scientific psychology was to be found in the hard-floor mazes of associative learning theory and not in the straw-strewn barns of psychopathology. Indeed, one of the salient Hull–Tolman controversies concerned whether or not it was possible to learn without reinforcement. An extensive series of experiments by Tolman and his followers established that when a rat is allowed to explore a complex maze without receiving any overt reinforcement, and then later made hungry and rewarded with food when it reached the goal box, it will learn the correct path faster than one that did not have the opportunity to explore the maze (e.g., Tolman & Honzik, 1930). The superior learning of the group that was preexposed to the maze, but without reinforcement, as compared to a control group that was not preexposed to the maze, was called “latent learning”, latent because there was no evidence of learning during the exploratory phase.
A short history of latent inhibition research
3
Figure 1.1. Experimental room with goat in harness (circa 1957). The CSs, rotor and light, are to the right of the goat. Shock electrodes and lever for recording the leg flexion are on the left leg. This was the Lubow and Moore (1959) procedure used in Experiment 2; in Experiment 1, the electrodes and recording lever were on the right side.
Today, it is generally accepted that although learning can benefit from reinforcement, reinforcement is not a necessary condition for learning. Even in the late 1950s, with the recognition of the distinction between learning and performance, the controversy was losing its significance. Nevertheless, to keep myself from following the assigned course of our animals into “neurosis” (the major conceptual variable for producing experimental neurosis was monotony), I began a project to try to demonstrate latent learning, not in a maze with rats, as all of the previous studies had done, but with sheep and goats in a purely classical conditioning paradigm. A classical conditioning procedure, with its clearly defined punctate conditioned stimuli, as opposed to the diffuse stimuli that constitute a maze, might, I thought, provide a clearer appreciation of what actually was learned during the non-reinforced stimulus preexposure stage. With the help of Ulrich Moore, I constructed an experimental set-up for sheep and goats (Figure 1.1), and we began to search for a facilitatory effect of simple preexposure of the to-be-conditioned stimulus on subsequent conditioning. In that first experiment (Lubow & Moore, 1959, Experiment 1), the animals were exposed to
4
R. E. Lubow
10 non-reinforced presentations of either a flashing light or a turning rotor. In the immediately following test stage, the subjects received alternating trials with the light and the rotor (CSs), each paired with a mild shock (US) to the right foreleg. With repeated CS–US pairings, such a conditioning procedure produces an anticipatory leg flexion (CR) during the presence of the CS. Thus, in this within-subject design, each subject generated CRs to the preexposed stimulus and to a novel stimulus. To our enormous disappointment, conditioning to the CS that had been preexposed was significantly slower than to the novel CS, exactly the opposite from what was expected. However, I knew the reason for these unforeseen results, or so I thought. Unintentionally, the layout of the experimental context was such that the preexposed stimulus of stage-1 and the CSs of stage-2 were presented to the right of the animal, and the US was delivered to the right foreleg. During preexposure, stimulus presentations elicited a noticeable turning of the animal’s head to the right, what Pavlov (1927) had called an investigatory reflex. Sherrington’s (1906) earlier analyses of reflex patterns had indicated that turning the head to one side is accompanied by an increase in extensor muscle tonus of the limb on that same side. Thus, the preexposed stimulus in our experiment was reliably paired with leg extension, a response that would be in competition with the conditioned flexion of the CR. Wonderful! The same analysis yielded the prediction that, if the preexposed stimulus and the CS in test are presented to one side of the animal, and the US to the opposite side, then the preexposed stimulus will produce a flexion of the to-be-conditioned limb. Since flexion is the to-be-conditioned response, then we certainly should be able to obtain the facilitatory latent learning effect that we originally had set out to demonstrate. We hurried back to the laboratory to test this sure-bet prediction (Lubow & Moore, 1959, Experiment 2). Except for the placement of the CSs and US on opposite sides of the animal, the procedure was exactly the same as in Experiment 1. By now, you can guess the outcome. Our prediction not only failed to be confirmed, but once again we obtained the opposite effect. Conditioning was poorer to the preexposed stimulus than to the novel stimulus. Evidently, we had stumbled on to a new behavioral phenomenon, to which we gave the name “latent inhibition”. Importantly, but often misunderstood, the freshly coined term was not meant to have any theoretical meaning. It simply was adopted as a complement to the latent learning phenomenon. The LI effect was deemed latent because there was no behavioral evidence of its presence during the stimulus preexposure stage. As for inhibition, it was only meant to describe that stimulus preexposure produced relatively poor learning, as opposed to the better learning of the latent learning studies. In retrospect, these first LI experiments had two conditions that made it surprising that we generated any effect at all. Firstly, the number of stimulus preexposures, ten, was relatively small. Secondly, the within-subject design of the test stage, with alternating presentations of the reinforced preexposed and novel stimuli, promotes
A short history of latent inhibition research
5
stimulus generalization. As such, it would be expected to reduce the differences in conditioning to the two stimuli. Indeed, as can be seen in Figure 1.2, the LI effects in Experiments 1 and 2, although reliable, were quite small. Upon finishing my graduate studies, I accepted a position with the General Electric Co. in Ithaca, New York. As a consequence, for five years I was unable to follow up on the initial LI experiments. Not surprisingly, the 1959 article also languished, neither noticed nor cited, well on its way to becoming another fatal drowning in the publication ocean. We (the publication and I) were rescued in 1963, when I was fortunate to receive a five-year NIH Career Development Award. The purpose of the grant was to investigate the role of eye movements in the development of visual form perception. The idea was to use a contact lens to mount a stable retinal image projection system on the eye of a new-born goat. Since goats are fully alert and ambulatory at birth, this technique would allow one to get at some of the basic nature/nurture interactions in perceptual development. North Carolina State University provided the needed animals, and I accepted an appointment in their Department of Psychology. Although the required equipment was developed, for a variety of reasons the research program became stalled. However, the sheep and goats were still on hand, and I returned to the still-puzzling phenomenon of LI. Given that the LI effects in the first experiments were completely unexpected, quite small, and generated from sheep and goats, I had two concerns. Most importantly, I was apprehensive that it might be difficult to replicate the effect. Secondly, I was troubled by the possibility that LI might not be generalizable beyond our idiosyncratic farm subjects, or beyond the classical conditioning defensive leg flexion paradigm. These issues occupied me for the next five years. It was with considerable relief that the next experiment (Lubow, 1965), also with sheep and goats, but with a between-subject design and with larger numbers of stimulus preexposures, produced a robust LI effect. Having replicated the basic LI effect, my students and I proceeded to address the problem of generality. In fairly short order, we demonstrated LI in rats with conditioned tail flexion (Chacto & Lubow, 1967) and conditioned suppression (Lubow & Siebert, 1969), and in rabbits with classical conditioning of the pinna response (Lubow, Markman, & Allen, 1968). During this same period, we made several unsuccessful attempts to produce LI in adult humans with the then-popular eyelid conditioning paradigm. Later, one of the students involved in this project, Paul Schnur, then at Indiana University, recognized the relevance of Kenneth Spence’s (1966) use of a masking task to slow the extinction of a classically conditioned response in humans, and thereby to make the adult extinction curve similar to that obtained with animals. Schnur and Ksir (1969) presented subjects with the masking task, an unsolvable guessing game that was initiated in the preexposure phase and continued through the acquisition phase. While engaged in the masking task, the subjects received the critical to-be-CS, later followed by the CS–US pairings. The stimulus-preexposed group displayed poorer
6
R. E. Lubow (a)
Percent conditioned responses
100
80
60
40
20
Non-preexposed stimulus Preexposed stimulus
0 1
2
3
4
5
6
7
Blocks of trials (b)
Percent conditioned responses
100
80
60
40
20
Non-preexposed stimulus Preexposed stimulus
0 1
2
3
4
5
6
7
Blocks of trials
Figure 1.2. Percent conditioned responses to the preexposed stimulus and to the non-preexposed stimulus as a function of blocks of trials: Panel A, Experiment 1; Panel B, experiment 2 (data from Lubow & Moore, 1959; graphs were not included in the original article; data were presented in tabular form as the number of trials before reaching a learning criterion of ten CRs).
A short history of latent inhibition research
7
conditioning than the groups that were not preexposed or preexposed to a stimulus that was different from the CS in the test stage. Thus, for the first time, stimulusspecific LI in humans was demonstrated, an effect that apparently required a masking task1. In 1971, after moving to Israel, and after a brief stay at Bar-Ilan University, I accepted a position at Tel Aviv University. In the meantime, Bob Rescorla and Allan Wagner published their influential model of associative learning (Rescorla & Wagner, 1972; Wagner & Rescorla, 1972), for which it was important to define the relationship between the LI effect and conditioned inhibition. They demonstrated that the stimulus preexposures prior to standard conditioning trials produced LI but not conditioned inhibition2 (Reiss & Wagner, 1972; Rescorla, 1971). Those two papers served to place LI outside the standard associative learning domain, and they led to accepting the idea that unreinforced preexposure of the to-be-conditioned stimulus results in a loss of stimulus salience. As a consequence of the novelty of the LI phenomenon, the awkwardness with which it fit into extant learning theories, and the interest of two of the major figures in the field, LI was on its way to becoming an active area of research. By the early 1970s, there was enough literature on LI and related effects to warrant a review article (Lubow, 1973). Essentially, the paper provided evidence for the robustness of the LI effect, summarizing research with different species and learning paradigms, as well as the effects of a number of variables. It was shown that LI was stimulusspecific, increased as a function of the number of stimulus preexposures, survived relatively long delays between preexposure and acquisition, and was marked by important differences between human children and adults. In addition, the case was made as to why LI could not be explained by means of habituation of the orienting response (e.g., Sokolov, 1963), selective filters (Sutherland & Mackintosh, 1971), competing and complementary responses (Lubow & Moore, 1959), or, as already noted, by conditioned inhibition. The paper concluded with a call for a model of LI that combined attentional and learning constructs, a theme that has remained current.
The middle years (1974–1991) Although one anonymous referee of the 1973 article harshly proclaimed that the “. . . review of the literature on latent inhibition is, unfortunately, much ado about practically nothing”, it, together with the Rescorla and Wagner papers, provided 1
2
Some of the issues raised by these early studies are still alive. For example, does the apparent requirement for a masking task for humans but not for animals mean that the LI effects in these two groups represent the outcomes of different processes? Does the masking task exert its influence by diverting attention from the preexposed stimulus or by disguising the transition from the preexposure to test stage? (For related discussions of these questions, see Lubow, this volume [Chapter 21]; Lubow & Weiner, this volume; Pelley & Schmidt-Hansen, this volume.) Conditioned inhibition is defined by performance deficits on acquisition (retardation) and summation tests. Latent inhibition also is characterized by attenuated acquisition, but the effects of stimulus preexposure do not summate with other excitatory or inhibitory stimuli.
8
R. E. Lubow
a turning point for LI research, both empirical and theoretical. Following their publication, numerous behavioral experiments examined LI as a function of a divers array of variables, including duration, intensity, and inter-stimulus interval of stimulus preexposures. LI also was assessed as a function of retention interval, various additions of other stimuli during preexposure, and most importantly of context change from preexposure to acquisition/test (for summaries, see Lubow, 1989). Accompanying the experimental research, ways were sought to encompass the new phenomenon within a general associative learning framework (e.g., Mackintosh, 1975; Pearce & Hall, 1980; Wagner, 1976). Although these theories continued with the Rescorla–Wagner position that stimulus preexposure produced a loss of stimulus salience (attention) and a subsequent reduction of stimulus associability (acquisition deficit), they differed from each other in regard to what is learned during preexposure and how that learning is transferred to the acquisition/test phase so as to reduce associability. My colleagues and I also accepted the general attention/associability position (A-theories), but we developed an account of LI that was specific to the results from the unattended stimulus preexposure paradigm, namely Conditioned Attention Theory (CAT; Lubow, Schnur, & Rifkin, 1976; Lubow, Weiner & Schnur, 1981; Lubow, 1989). We posited that non-reinforced stimulus preexposures retard subsequent conditioning to that stimulus because the subject learns not to attend to the irrelevant stimulus. However, unlike other A-theories, attention was treated as a response that was subject to all of the empirical laws of classical conditioning, with the absence of a consequence following the to-be-CS serving as a US for the conditioning of an inattentional response. As noted, like other A-theories, CAT accepted the position that the reduced attention/salience of the to-be-CS decreased the ability to acquire a new association to that stimulus. Ironically, early on, we had published a paper that would later provide a basis for undermining the acquisition-deficit position (Lubow, Alek, & Rifkin, 1976). In two experiments, one with children and one with rats, both using a two-stage preexposure and acquisition/test procedure, a comparison between the PE and NPE groups showed that a change of context from the preexposure stage to the acquisition/test stage destroyed the LI effect. In other words, LI was context-specific. The results were readily interpretable in terms of modulated attention, namely that an old stimulus in an old context attracts less attention than a new stimulus in an old context, the consequence of which is poorer acquisition of any new association with the familiar stimulus. The attenuated-LI that accompanied a change of context that we had demonstrated in a two-stage procedure was soon confirmed in a variety of three-stage experiments (preexposure, acquisition, and test) where the context change occurred from the preexposure to the conditioning stage (e.g., Hall & Channell, 1986; Lovibond, Preston, & Mackintosh, 1984). However, the attentional/perceptual interpretation of context-specificity of LI (Lubow, Alek, & Rifkin, 1976), particularly as
A short history of latent inhibition research
9
expressed in modulations of subsequent associability, was challenged by Ralph Miller and Mark Bouton, both of whom contended that LI did not represent a failure of acquisition, but rather that it was the result of retrieval/competition processes (R-theories; e.g., Miller, Kasprow, & Schachtman, 1986; Bouton, 1993; also see Escobar & Miller, this volume). A similar conclusion was reached by Weiner (1990) from a review of the psychobiological data. Accordingly, unlike for A-theories, the acquisition of a new association to an old preexposed stimulus proceeds normally. However, when the subject again encounters the target stimulus in the test stage, two competing associations are retrieved: the earlier stimulus–no consequence association from the preexposure stage and the stimulus–US association from the acquisition stage. In such a situation, the non-preexposed group would gain an advantage because, at the time of test, it had not been previously confronted with a competing association. Although this is not the place to compare and evaluate theories of LI, it is of interest to note that the controversy is yet another version of a broader division that has a long history, namely learning versus performance. Although there is compelling evidence for both sides of the argument, no experiment has been designed that explicitly formulates and tests the contrasting predictions. This suggests that these apparently conflicting theories may either lack sufficient preciseness, or deal with different data domains, or overlap in meaning, or even suffer from some combination of these possibilities. These issues are still alive (see Westbrook & Bouton, this volume; De la Casa & Pinen˜o, this volume; Hall & Rodriguez, this volume; Escobar & Miller, this volume; Lubow & Weiner, this volume; Schmajuk, this volume). Thus far, the overview of LI activity has focused on behavioral experimentation and theory, which, indeed, dominated the field. However, just as neuroscience was beginning to affect many other aspects of experimental psychology, so too it was on its way to becoming a dominant factor in LI research. The earliest work in this area took several directions. Two of these tracks, namely lesions of the septo-hippocampal system and administration of drugs that modulate dopamine uptake, which evolved directly from interests in attentional processes, were later linked in an effort to understand anomalous LI effects in schizophrenia. In the first of these studies, Ackil, Mellgren, Halgren, and Frommer (1969) reported that aspiration lesions of the hippocampus abolished LI of two-way active avoidance learning. The next several years saw a number of replications of this effect with a variety of different preparations (e.g., Kaye & Pearce, 1987; Solomon & Moore, 1975), as well as with septal lesions (e.g., Burton & Toga, 1982; see Gray & McNaughton, 1983, for an early review of the similar affects of septal and hippocampal lesions on LI). LI also was examined as a function of manipulations of noradrenergic (e.g., Tsaltas, Preston, Rawlins et al., 1984), serotonergic (e.g., Solomon, Kiney, & Scott, 1978), and cholinergic systems (e.g., Moore, Goodell, & Solomon, 1976). The facts that amphetamine, a dopamine agonist, can produce schizophrenia-like symptoms in humans and that schizophrenia is characterized by high distractibility
10
R. E. Lubow
(attending to irrelevant stimuli) led to several critical studies of the effects of that drug on LI. Beginning with Ina Weiner’s master’s thesis (1979) and doctoral dissertation (1983), a series of experiments in our laboratory (Weiner, Lubow, & Feldon, 1981, 1984, 1988) and that of Paul Solomon (Solomon, Crider, Winkelman et al., 1981; Solomon & Staton, 1982) established that d-amphetamine attenuates LI. Importantly, Solomon et al. (1981) also found that the disruptive effects of d-amphetamine were eliminated by concomitant administration of chlorpromazine, a powerful neuroleptic. Shortly afterwards, Weiner and colleagues reported that haloperidol, a newer dopamine antagonist, actually potentiated LI (Weiner & Feldon, 1987; Weiner, Feldon, & Katz, 1987), and with that LI became a focal point for research relating attentional dysfunction to schizophrenia. This thrust was accelerated by two characteristics of the LI paradigm. Firstly, the same procedure for generating LI could be employed with animals and humans, including those with various pathologies. Secondly, attenuated LI is indexed by better learning in the test phase. Consequently, a reduction of LI in patient groups cannot be attributed to the nonspecific decremental effects which have plagued much of the research on schizophrenia. By the close of the 1980s, there was enough published LI research to warrant two books (Lubow, 1989; Hall, 1991), both of which reviewed much of the material described above, but in considerably more detail.
The current era (1991–2008) It is appropriate to begin this latest period with Jeffrey Gray (1934–2004), a visionary neuroscientist whose broad interests and knowledge made him the quintessential interdisciplinary researcher (see Schmajuk, this volume). Although he had already made a significant contribution to LI research in the previous era (e.g., Baruch, Hemsley, & Gray, J. A. 1988a, 1988b), his widely cited Behavior and Brain Sciences article (Gray, J. A. et al., 1991), which drew from the results of experiments on behavior, pharmacology, neurophysiology, and psychopathology, provided a coherent model for LI deficits in schizophrenia patients with acute, positive symptoms (also see Gray, J. A., 1998). Gray’s influential publications, together with our earlier work on the opposing influences of dopamine agonists and antagonists on LI (e.g., Lubow, Feldon & Weiner, 1982; Weiner, Lubow, & Feldon, 1981, 1984; Weiner & Feldon, 1987), and Lubow’s (1989) book summarizing the burgeoning LI literature, set the stage for much of the more recent developments in LI research and theorizing, and for their applications to schizophrenia. Contemporary LI-schizophrenia research, much of which is summarized in the chapters of the present volume (e.g., Kumari & Ettinger; Swerdlow; Weiner), has continued in a number of the older areas, including behavioral studies with patients and healthy subjects who score high on self-report schizotypal questionnaires (for a
A short history of latent inhibition research
11
summary, see Lubow, 2005; also Kumari & Ettinger, this volume). In addition, efforts have been made to examine LI effects associated with such pathologies as ADHD (e.g., Lubow, Braunstein-Bercovitz, Blumenthal et al., 2005), Obsessive Compulsive Disorder (e.g., Kaplan, Dar, Rosenthal et al., 2006; Swerdlow, Hartston, & Hartman, 1999), Parkinson Disease (Lubow, Dressler, & Kaplan, 1999), and Tourette’s Syndrome (Swerdlow, Magulac, Filion, & Zinner, 1996). Surprisingly, studies that have correlated brain activity and LI in human subjects have been notably sparse throughout the 50-year course of LI research. A literature search indicates a single LI-EEG publication (Savostjanov & Ilyuchenok, 1999). LI experiments with evoked response potentials (ERP) also are uncommon (for exceptions, Guterman, Josiassen, Bashore et al., 1996; Kathmann, von Recum, Haag, & Engel, 2000). Even the newer and fashionable neuro-imaging techniques appear in only two LI papers (Filsinger, Zimmermann, Kirsch et al., 2004; Young, Kumari, Mehrotra et al., 2005). Unfortunately, the small number of papers, spread across a variety of procedures, precludes identifying reliable effects. Future advances in these areas may be linked to the ability to develop a reliable and robust within-subject LI procedure, a topic that has recently received some attention (e.g., De la Casa & Lubow, 2001; Evans, Gray, & Snowden, 2007; Lubow & Kaplan, 2005; Swerdlow, Stephany, Wasserman et al., 2003). Studies with animals and humans on the effects of psychoactive drugs on LI have remained popular, but include new directions which have released LI from the dominance of the dopaminergic model, including investigations of glutamatergic transmission (e.g., Gaisler-Salomon & Weiner, 2003; Gaisler-Salomon et al., 2008), as well as GABAergic (e.g., Lewis & Gould, 2007b), serotonergic (e.g., Boulenguez, Peters, Mitchell et al., 1998) and cholinergic (e.g., Barak & Weiner, 2007) involvements (for earlier summaries, see Moser, Hitchcock, Lister, & Moran, 2000; Tzschentke, 2001; Weiner, 2003). Animal lesion, in vivo voltammetry, and microdialysis experiments, although less popular than pharmacological studies, have successfully mapped the brain substrates underlying the regulation of LI acquisition and expression (e.g., Gal, Schiller, & Weiner, 2005; Peterschmitt, Meyer, & Louilot, 2007; Pothuizen, Jongen-Relo, Feldon, & Yee, 2006; also see Cassaday & Moran, this volume; Honey, Iordanova, & Good, this volume; Louilot et al., this volume; Weiner, this volume). Of special importance for the construct validity of the LI model of schizophrenia, the brain regions that are responsible for modulating LI overlap with those implicated in the pathophysiology of schizophrenia (e.g., Peterschmitt, Meyer, & Louilot, 2007; for review, see Weiner, 2003; for a review of behavioral evidence for the construct validity, see Lubow, 2005). Genetic paradigms have also come to play a role in LI research. On the one hand, there have been within-species comparisons of LI effects in different strains of mice or rats that have been inbred over many generations, but not necessarily for a specific trait (e.g., Gould & Wehner, 1999). On the other hand, studies have compared strains bred for a specific trait that is related to LI. As an example, Kline, Decena,
12
R. E. Lubow
Hitzemann, and McCaughran (1998) assessed LI in mice from neuroleptic-responsive (NR) and neuroleptic non-responsive (NNR) lines and compared them to a heterogeneous stock for responsiveness to haloperidol-induced catalepsy. LI was obtained in NR and control groups, but not in NNR groups. Similarly, Conti, Palmer, Vanella, and Printz (2001) measured LI in rat strains that were characterized by different levels of prepulse inhibition. Attempts have also been made to selectively breed mice for LI itself (e.g., Ely, Gross, Lambert, & Kilts, 1997; Kilts, Scibilia, & Dunn, 1990; for related experiments with honeybees, see Chandra, Hosler, & Smith, 2000). With humans, a genetic connection between LI and schizophrenia was reported by Serra, Jones, Toone, and Gray, J. A. (2001). They found the absence of a significant LI effect in both high and low schizotypals who were first-degree relatives of chronic schizophrenic patients who themselves did not show LI. Molecular biological and genetic engineering procedures, primarily with mice (e.g., Gould & Lewis, 2005; Lewis & Gould, 2007a; Lipina et al., 2007; see Gould, this volume) have become increasingly popular. Not only do mutant mice that exhibit schizophrenic-like behaviors also show LI deficits (e.g., Lipina, Weiss, & Roder, 2007; Rimer, Barrett, Maldonado et al., 2005; also see Lipina and Roder, this volume), but LI anomalies are also seen in neurodevelopmentally compromised mice that exhibit schizophrenia-relevant abnormalities (e.g., Gerdjikov, Rudolph, Keist et al., 2008; Meyer, Feldon, Schedlowski, & Yee, 2006; Meyer, Schwendener, Feldon, & Yee, 2006; Weiner, this volume; Smith, Li, Garbett et al., 2007). Similar effects have been reported in genetically modified mice (Kline et al., 1998; Caldarone, Duman, & Picciotto, 2000; Harrell & Allan, 2003; Miyakawa et al., 2003; Wang et al., 2004; Rimer et al., 2005; Yee et al., 2006; Bruno et al., 2007; Clapcote et al., 2007; Lipina et al., 2007; Gerdjikov et al., 2008; see Lipina & Roder, this volume; Schnur & Hoffman, this volume). On the behavioral-theoretical side, several new or modified accounts of LI have been put forward (e.g., Denniston, Savastano, & Miller, 2001; Hall & Rodriguez, this volume; Lubow & Gewirtz, 1995; McLaren & Mackintosh, 2000; Schmajuk, Lam, & Gray, J. A. 1996), as well as specific hypotheses about the relationship between LI and schizophrenia (e.g., Escobar, Oberling, & Miller, 2002; Lubow, 2005; Schmajuk, 2002; Weiner, 2003). Importantly, Weiner’s (2003) “two-headed” model of schizophrenia went beyond Gray’s version and accounted for negative as well as positive symptoms. Indeed, there is some evidence that the direction of LI modulations, potentiated or attenuated, is related to symptom type in patients (e.g., Cohen, Sereni, Kaplan et al., 2004) and schizotypal normals (e.g., Gray, N. S., Fernandez, Williams et al., 2002). Finally, there is a practical side to LI research. There is some evidence that LI stimulus preexposure procedures may prevent the acquisition of fears and phobias (for a summary, see Lubow, 1998). In addition, the LI paradigm has been used to screen for potentially useful psychoactive drugs (e.g., Bider, Gross, Nemeroff, & Kilts, 2002; for a summary see Weiner, Gaisler, Schiller et al., 2000; also see Weiner, this volume; Weiner & Arad, this volume).
A short history of latent inhibition research
13
In summary, the last 50 years has seen LI research progress from studying the legs of goats and sheep in 1959 to molecular biology and genetic engineering in 2009. Although this would appear to be a triumph of reductionism, one should recall that these current investigations were born from purely behavioral experiments that revealed the entirely unexpected LI effect. LI then became significant for general theories of learning, which, in turn, was followed by demonstrations of behavioral connections to schizophrenia. Building on the link with schizophrenia, LI research has cut a wide swath across different levels of description and analyses, a dynamic that belies the polarized experimental/theoretical psychology of an earlier era that pitted a dry molar behaviorism against a wet molecular physiology. Like many dichotomies, they made for sterile debates that were ultimately fruitless. The collection of chapters in the present volume underscores the significance of the interaction between behavioral and neurobiological experimentation and theorizing, in this case for a better understanding of how irrelevant stimuli are processed and how dysfunctions in the systems underlying that processing may relate to schizophrenia.
References Ackil, J. E., Mellgren, R. L., Halgren, C., & Frommer, G. P. (1969). Effects of CS preexposure on avoidance learning in rats with hippocampal lesions. Journal of Comparative and Physiological Psychology, 69, 739–747. Baruch, I., Hemsley, D. R., & Gray, J. A. (1988a). Differential performance of acute and chronic schizophrenics in a latent inhibition task. Journal of Nervous and Mental Disease, 176, 698–600. Baruch, I., Hemsley, D. R., & Gray, J. A. (1988b). Latent inhibition and “psychotic proneness” in normal subjects. Personality and Individual Differences, 9, 777–783. Binder, E. B., Gross, R. E., Nemeroff, C. B., & Kilts, C. D. (2002). Effects of neurotensin receptor antagonism on latent inhibition in Sprague-Dawley rats. Psychopharmacology, 161, 288–295. Boulenguez, P., Peters, S. L., Mitchell, S. N., Chauveau, J., Gray, J. A., & Joseph, M. J. (1998). Dopamine release in the nucleus accumbens and latent inhibition in the rat following microinjections of a 5-HT1B agonist into the dorsal subiculum: implications for schizophrenia. Journal of Psychopharmacology, 12, 258–267. Bouton, M. E. (1993). Context, time, and memory retrieval in the interference paradigms of Pavlovian conditioning. Psychological Bulletin, 114, 80–99. Bruno, K. J., Freet, C. S., Twining, R. C., et al. (2007). Abnormal latent inhibition and impulsivity in coloboma mice, a model of ADHD. Neurobiology of Disease, 25, 206–216. Burton, H. A., & Toga, A. W. (1982). Successive discrimination performance improves with increasing numbers of stimulus preexposures in septal rats. Behavioral and Neural Biology, 34, 141–151. Caldarone, B. J., Duman, C. H., & Picciotto, M. R. (2000). Fear conditioning and latent inhibition in mice lacking the high affinity subclass of nicotinic acetylcholine receptors in the brain. Neuropharmacology, 39, 2779–2784. Chacto, C., & Lubow, R. E. (1967). Classical conditioning and latent inhibition in the white rat. Psychonomic Science, 9, 135–136.
14
R. E. Lubow
Chandra, S. B. C., Hosler, J. S., & Smith, B. H. (2000). Heritable variation for latent inhibition and its correlation with reversal learning in honeybees (Apis mellifera). Journal of Comparative Psychology, 114, 73–85. Clapcote, S. J., Lipina, T. V., Millar, K. J., et al. (2007). Behavioral phenotypes of Disc1 missense mutations in mice. Neuron, 54, 387–402. Cohen, E., Sereni, N., Kaplan, O., et al. (2004). The relation between latent inhibition and symptom-types in young schizophrenics. Behavioural Brain Research, 149, 113–122. Conti, L. H., Palmer, A. A., Vanella, J. J., & Printz, M. P. (2001). Latent inhibition and conditioning in rat strains which show differential prepulse inhibition. Behavior Genetics, 31, 325–333. De la Casa, G., & Lubow, R. E. (2001). Latent inhibition with a response time measure from a within-subject design: effects of number of preexposures, masking task, context change, and delay. Neuropsychology, 15, 244–253. Denniston, J. C., Savastano, H. I., & Miller, R. R. (2001). The extended comparator hypothesis: learning by contiguity, responding by relative strength. In R. R. Mowrer & S. B. Klein (Eds.), Handbook of Contemporary Learning Theories. Mahwah, NJ: Erlbaum, pp. 65–117. Ely, T. D., Gross, R. F., Lambert, P. D., & Kilts, C. D. (1997). The neuroanatomical substrates of latent inhibition: a behavioral genetic quantitative cytochrome oxidase histochemical study. Society for Neuroscience, 23, 1840. Escobar, M., Oberling, P., & Miller, R. R. (2002). Associative deficit accounts of disrupted latent inhibition and blocking in schizophrenia. Neuroscience and Biobehavioral Reviews, 26, 203–216. Evans, L., Gray, N. S., & Snowden, R. J. (2007). A new continuous withinparticipant latent inhibition task: examining associations with schizotypy dimensions, smoking and gender. Biological Psychology, 74, 365–373. Filsinger, I., Zimmermann, M., Kirsch, P., et al. (2004). Latent inhibition in schizophrenic patients and healthy controls: an event-related fMRI study. Journal of Psychophysiology, 18, 241 (abstract). Gaisler-Salomon, I., Diamant, L., Rubin, C., & Weiner, I. (2008). Abnormally persistent latent inhibition induced by MK801 is reversed by risperidone and by positive modulators of NMDA: differential efficacy depending on the stage of the task at which they are administered. Psychopharmacology, 196, 255–267. Gaisler-Salomon, I., & Weiner, I. (2003). Systemic administration of MK-801 produces an abnormally persistent latent inhibition which is reversed by clozapine but not haloperidol. Psychopharmacology, 166, 333–342. Gal, G., Schiller, D., & Weiner, I. (2005). Latent inhibition is disrupted by nucleus accumbens shell lesion but is abnormally persistent following entire nucleus accumbens lesion: the neural site controlling the expression and disruption of the stimulus preexposure effect. Behavioural Brain Research, 162, 246–255. Gerdjikov, T. V., Rudolph, U., Keist, R., et al. (2008). Hippocampal alpha 5 subunit-containing GABAA receptors are involved in the development of the latent inhibition effect. Neurobiology of Learning and Memory, 89, 87–94. Gould, T. J., & Lewis, M. C. (2005). Coantagonism of glutamate receptors and nicotinic acetylcholinergic receptors disrupts fear conditioning and latent inhibition of fear conditioning. Learning & Memory, 12, 389–398. Gould, T. J., & Wehner, J. M. (1999). Genetic influences on latent inhibition. Behavioral Neuroscience, 113, 1291–1296.
A short history of latent inhibition research
15
Guterman, Y., Josiassen, R., Bashore, R. E., Johnson, M., & Lubow, R. E. (1996). Latent inhibition effects reflected in event-related brain potentials in healthy controls and schizophrenics. Schizophrenia Research, 20, 315–326. Gray, J. A. (1998). Integrating schizophrenia. Schizophrenia Bulletin, 24, 249–266. Gray, J. A., Feldon, J., Rawlins, J. N. P., Hemsley, D. R., & Smith, A. D. (1991). The neuropsychology of schizophrenia. Behavioral Brain Science, 14, 1–84. Gray, J. A., & McNaughton, N. (1983). Comparison between the behavioral effects of septal and hippocampal lesions: a review. Neuroscience & Biobehavior Reviews, 7, 119–188. Gray, N. S., Fernandez, M., Williams, J., Ruddle, R. A., & Snowden, R. J. (2002). Which schizotypal dimensions abolish latent inhibition? British Journal of Clinical Psychology, 41, 271–284. Hall, G. (1991). Perceptual and Associative Learning. Oxford: Clarendon Press. Hall, G., & Channell, S. (1986). Context specificity of latent inhibition in taste aversion learning. Quarterly Journal of Experimental Psychology, 38B, 121–139. Harrell, A. V., & Allan, A. M. (2003). Improvements in hippocampal-dependent learning and decremental attention in 5-HT3 receptor overexpressing mice. Learning & Memory, 10, 410–419. Hull, C. L. (1943). Principles of Behavior. NewYork: Appleton-Century-Crofts. Hull, C. L. (1952). A Behavior System. New Haven: Yale University Press. James, W. (1950). The Principles of Psychology, vol. 1. New York: Dover (Original published in 1890). Kaplan, O., Dar, R., Rosenthal, L., et al. (2006). Obsessive-compulsive disorder patients display enhanced latent inhibition on a visual search task. Behavior Research and Therapy, 44, 1137–1145. Kathmann, N., von Recum, S., Haag, C., & Engel, R. R. (2000). Electrophysiological evidence for reduced latent inhibition in schizophrenic patients. Schizophrenia Research, 45, 103–114. Kaye, H., & Pearce, J. M. (1987). Hippocampal lesions attenuate latent inhibition of a CS and of a neutral stimulus. Psychobiology, 15, 293–299. Kilts, C., Scibilia, R., & Dunn, L. (1990). Selective breeding for distinct stimulus filtering/selective attention behavioral phenotypes. Society for Neuroscience Abstracts, 16, 189.11. Kline, L., Decena, E., Hitzemann, R., & McCaughran, J. (1998). Acoustic startle, prepulse inhibition, locomotion, and latent inhibition in the neuroleptic responsive (NR) and neuroleptic nonresponsive (NNR) lines of mice. Psychopharmacology, 139, 322–331. Lewis, M. C., & Gould, T. J. (2007a). Reversible inactivation of the entorhinal cortex disrupts the establishment and expression of latent inhibition of cued fear conditioning in C57BL/6 mice. Hippocampus, 17, 462–470. Lewis, M. C., & Gould, T. J. (2007b). Signal transduction mechanisms within the entorhinal cortex that support latent inhibition of cued fear conditioning. Neurobiology of Learning and Memory, 88, 359–368. Lovibond, P. F., Preston, G. C., & Mackintosh, N. (1984). Context specificity of conditioning, extinction, and latent inhibition. Journal of Experimental Psychology: Animal Behavior Processes, 10, 360–375. Lipina, T., Weiss, K., & Roder, J. (2007). The Ampakine CX546 restores the prepulse inhibition and latent inhibition deficits in mGluR5-deficient mice. Neuropsychopharmacology, 32, 745–756.
16
R. E. Lubow
Lubow, R. E. (1965). Latent inhibition: the effects of frequency non-reinforced preexposure to the CS. Journal of Comparative and Physiological Psychology, 60, 454–457. Lubow, R. E. (1973). Latent inhibition. Psychological Bulletin, 79, 398–407. Lubow, R. E. (1989). Latent Inhibition and Conditioned Attention Theory. Cambridge: Cambridge University Press. Lubow, R. E. (1998). Latent inhibition and behavioral pathology: some effects of stimulus preexposure. In W. O’Donohue (Ed.), Learning and Behavior Therapy. Boston: Allyn and Bacon, pp. 107–121. Lubow, R. E. (2005). The construct validity of the animal-latent inhibition model of selective attention deficits in schizophrenia. Schizophrenia Bulletin, 31, 139–153. Lubow, R. E., Alek, M., & Rifkin, B. (1976). The context effect: the relationship between stimulus preexposure and environmental preexposure determines subsequent learning. Journal of Experimental Psychology: Animal Behavior Processes, 2, 38–47. Lubow, R. E., Braunstein-Bercovitz, H., Blumenthal, O., Kaplan, O., & Toren, P. (2005). Latent inhibition and asymmetrical visual-spatial attention in children with ADHD. Child Neuropsychology, 11, 445–457. Lubow, R. E., Dressler, R., & Kaplan, O. (1999). Visual search and latent inhibition in de novo Parkinson patients. Neuropsychology, 13, 415–423. Lubow, R. E., Feldon, J., & Weiner, I. (1982). An animal model of attention. In A. Levy & M. Y. Spiegelstein (Eds.), Behavioral Models and the Analysis of Drug Action. Amsterdam: Elsevier. Lubow, R. E., & Gewirtz, J. C. (1995). Latent inhibition in humans: data, theory, and implications for schizophrenia, Psychological Bulletin, 117, 87–103. Lubow, R. E., & Kaplan, O. (2005). The visual search analog of latent inhibition: implications for theories of irrelevant stimulus processing in normal and schizophrenic groups. Psychonomic Bulletin & Review, 12, 224–243. Lubow, R. E., Markman, R. E., & Allen, J. (1968). Latent inhibition and classical conditioning of the rabbit pinna response. Journal of Comparative and Physiological Psychology, 66, 688–694. Lubow, R. E., & Moore, A. U. (1959). Latent inhibition: the effect of non-reinforced preexposure to the conditioned stimulus. Journal of Comparative and Physiological Psychology, 52, 415–419. Lubow, R. E., Schnur, P., & Rifkin, B. (1976). Latent inhibition and conditioned attention theory. Journal of Experimental Psychology: Animal Behavior Processes, 2, 163–174. Lubow, R. E., & Siebert, L. (1969). Latent inhibition within the CER paradigm. Journal of Comparative and Physiological Psychology, 69, 136–138. Lubow, R. E., Weiner, I., & Schnur, P. (1981). Conditioned attention theory. In G. H. Bower (Ed.), The Psychology of Learning and Motivation, vol. 15. New York: Academic Press, pp. 1–49. Mackintosh, N. J. (1975). A theory of attention: variations in the associability of stimuli with reinforcement. Psychological Review, 82, 276–298. McLaren, I. P. L., & Mackintosh, N. J. (2000). An elemental model of associative learning: I. Latent inhibition and perceptual learning. Animal Learning & Behavior, 28, 211–246.
A short history of latent inhibition research
17
Meyer, U., Feldon, J., Schedlowski, M., & Yee, B. K. (2006a). Immunological stress at the maternal-foetal interface: a link between neurodevelopment and adult psychopathology. Brain, Behavior, and Immunity, 20, 378–388. Meyer, U., Schwendener, S., Feldon, J., & Yee, B. K. (2006b). Prenatal and postnatal maternal contributions in the infection model of schizophrenia. Experimental Brain Research, 173, 243–57. Miller, R. R., Kasprow, W. J., & Schachtman, T. R. (1986). Retrieval variability: sources and consequences. American Journal of Psychology, 99, 145–218. Miyakawa, T., Leiter, L. M., Gerber, D. J., et al. (2003). Conditional calcineurin knockout mice exhibit multiple abnormal behaviors related to schizophrenia. Proceedings of the National Academy of Sciences of the United States of America, 100, 8987–8992. Moore, J. W., Goodell, N. A., & Solomon, P. R. (1976). Central cholinergic blockade by scopolamine and habituation, classical conditioning, and latent inhibition of rabbit’s nictitating membrane response. Physiological Psychology, 4, 395–399. Moser, P. C., Hitchcock, J. M., Lister, S., & Moran, P. M. (2000). The pharmacology of latent inhibition as an animal model of schizophrenia. Brain Research Reviews, 33, 275–307. Pavlov, I. P. (1927). Conditioned Reflexes. London: Oxford University Press. Pearce, J. M., & Hall, G. (1980). A model of Pavlovian learning: variations in the effectiveness of conditioned but not unconditioned stimuli. Psychological Review, 87, 532–552. Peterschmitt, Y., Meyer, F., & Louilot, A. (2007). Neonatal functional blockade of the entorhinal cortex results in disruption of accumbal dopaminergic responses observed in latent inhibition paradigm in adult rats. European Journal of Neuroscience, 25, 2504–2513. Pothuizen, H. H. J., Jongen-Relo, A. L., Feldon, J., & Yee, B. K. (2006). Latent inhibition of conditioned taste aversion is not disrupted, but can be enhanced, by selective nucleus accumbens shell lesions in rats. Neuroscience, 137, 1119–1130. Reiss, S., & Wagner, A. R. (1972). CS habituation produces “latent inhibition” effect but no active “conditioned inhibition”. Learning and Motivation, 3, 237–245. Rescorla, R. A. (1971). Summation and retardation tests of latent inhibition. Journal of Comparative and Physiological Psychology, 75, 77–81. Rescorla, R. A., & Wagner, A. R. (1972). A theory of Pavlovian conditioning: variations in the effectiveness of reinforcement and nonreinforcement. In A. H. Black & W. F. Prokasy (Eds.), Classical Conditioning II: Current Research and Theory. New York: Appleton-Century-Crofts, pp. 64–99. Rimer, M., Barrett, D. W., Maldonado, M. A., Vock, V. M., & Gonzalez-Lima, F. (2005). Neuregulin-1 immunoglobulin-like domain mutant mice: clozapine sensitivity and impaired latent inhibition. Neuroreport, 16, 271–275. Savostjanov, A. N., & Ilyuchenok, R. Y. (1999). Spectra and correlations of human brain electrical activity (EEG) in the range of the alpha-rhythm frequency during manifestation of latent inhibition. Zhurnal Vysshei Nervnoi Deyatelnosti Imeni I Pavlova, 49, 610–619. Schmajuk, N. A. (2002). Latent Inhibition and its Neural Substrates. Norwell, MA: Kluwer Academic Publishers. Schmajuk, N. A., Lam, Y. -W., & Gray, J. A. (1996). Latent inhibition: a neural network approach. Journal of Experimental Psychology: Animal Brhavior Processes, 22, 321–349.
18
R. E. Lubow
Schnur, P., & Ksir, C. J. (1969). Latent inhibition in human eyelid conditioning. Journal of Experimental Psychology, 80, 388–389. Serra, A. M., Jones, S. H., Toone, B., & Gray, J. A. (2001). Impaired associative learning in chronic schizophrenics and their first-degree relatives: a study of latent inhibition and the Kamin blocking effect. Schizophrenia Research, 48, 273–289. Sherrington, C. S. (1906). The Integrative Action of the Nervous System. London: Constable. Smith, S. E. P., Li, J., Garbett, K., Minics, K., & Patterson, P. H. (2007). Maternal immune activation alters fetal brain development through interleukin-6. Journal of Neuroscience, 27, 10695–10702. Sokolov, E. N. (1963). Perception and the Conditioned Reflex. New York: Pergamon Press. Solomon, P. R., Crider, A., Winkelman, J. W., et al. (1981). Disrupted latent inhibition in the rat with chronic amphetamine or haloperidol-induced supersensitivity: relationship to schizophrenic attention disorder. Biological Psychiatry, 16, 519–538. Solomon, P. R., Kiney, C. A., & Scott, D. S. (1978). Disruption of latent inhibition following systemic administration of parachlorophenylanine (PCPA). Physiology and Behavior, 20, 265–271. Solomon, P. R., & Moore, J. W. (1975). Latent inhibition and stimulus generalization of classically conditioned nictitating membrane response in rabbits (Oryctolagus cuniculus) following dorsal hippocampal ablation. Journal of Comparative and Physiological Psychology, 89, 1192–1203. Solomon, P. R., & Staton, D. M. (1982). Differential effects of microinjections of d-amphetamine into the nucleus accumbens or the caudate putamen on the rat’s ability to ignore an irrelevant stimulus. Biological Psychiatry, 17, 743–756. Spence, K. W. (1966). Cognitive and drive factors in the extinction of the conditioned eye blink in human subjects. Psychological Review, 73, 445–458. Sutherland, N. S., & Mackintosh, N. J. (1971). Mechanisms of Animal Discrimination Learning. New York: Academic Press. Swerdlow, N. R., Hartston, H. J., & Hartman, P. L. (1999). Enhanced visual latent inhibition in obsessive-compulsive disorder. Biological Psychiatry, 45, 482–488. Swerdlow, N. R., Magulac, M., Filion, D., & Zinner, S. (1996). Visuospatial priming and latent inhibition in children and adults with Tourette’s disorder. Neuropsychology, 10, 485–494. Swerdlow, N. R., Stephany, N., Wasserman, L. C., Talledo, J., Sharp, R., Auerbach, P. P. (2003). Dopamine agonists disrupt visual latent inhibition in normal males using a within-subject design. Psychopharmacology, 169, 314–320. Tolman, E. C. (1932). Purposive Behavior in Animals and Men. NewYork: AppletonCentury-Crofts. Tolman, E. C., & Honzik, H. C. (1930). Introduction and removal of reward, and maze performance of rats. University of California Publications in Psychology, 4, 257–275. Tsaltas, E., Preston, G. C., Rawlins, J. N. P., Winocur, G., & Gray, J. A. (1984). Dorsal bundle lesions do not affect latent inhibition of conditioned suppression. Psychopharmacology, 84, 549–555. Tzschentke, T. M. (2001). Pharmacology and behavioral pharmacology of the mesocortical dopamine system. Progress in Neurobiology, 63, 241–320.
A short history of latent inhibition research
19
Wagner, A. R. (1976). Priming in STM: an information processing mechanism for self-generated or retrieval-generated depression in performance. In T. Tighe & R. N. Leaton (Eds.), Habituation: Perspectives from Child Development, Animal Behavior, and Neurophysiology. Hillsdale, NJ: Lawrence Erlbaum, pp. 95–128. Wagner, A. R., & Rescorla, R. A. (1972). Inhibition in Pavlovian conditioning: application of a theory. In R. A. Boakes & M. S. Halliday (Eds.), Inhibition and Learning. New York: Academic Press. Wang, H., Ng, K., Hayes, D., et al. (2004). Decreased amphetamine-induced locomotion and improved latent inhibition in mice mutant for the M5 muscarinic receptor gene found in the human 15q schizophrenia region. Neuropsychopharmacology, 29, 2126–2139. Weiner, I. (1979). The effects of arousal on latent inhibition: Tests of conditioned attention theory. Unpublished Master’s thesis, Tel Aviv University. Weiner, I. (1983). The effects of amphetamine on latent inhibition: Tests of the animal amphetamine model of schizophrenia using selected learning paradigms. Unpublished doctoral dissertation, Tel Aviv University. Weiner, I. (1990). The neural substrates of latent inhibition. Psychological Bulletin, 108, 442–461. Weiner, I. (2003). The “two-headed” latent inhibition model of schizophrenia: modeling positive and negative symptoms and their treatment. Psychopharmacology, 169, 257–297. Weiner, I., & Feldon, J. (1987). Facilitation of latent inhibition by haloperidol. Psychopharmacology, 91, 248–253. Weiner, I., Feldon, J., & Katz, Y. (1987). Facilitation of the expression but not the acquisition of latent inhibition by haloperidol in rats. Pharmacology, Biochemistry and Behavior, 26, 241–246. Weiner, I., Gaisler, I., Schiller, D., et al. (2000). Screening of antipsychotic drugs in animal models. Drug Development Research, 50, 235–249. Weiner, I., Lubow, R. E., & Feldon, J. (1981). Chronic amphetamine and latent inhibition. Behavioral Brain Research, 2, 285–286. Weiner, I., Lubow, R. E., & Feldon, J. (1984). Abolition of the expression but not the acquisition of latent inhibition by chronic amphetamine in rats. Psychopharmacology, 83, 194–199. Weiner, I., Lubow, R. E., & Feldon, J. (1988). Disruption of latent inhibition by acute administration of low doses of amphetamine. Pharmacology, Biochemistry, and Behavior, 30, 871–878. Young, A. M. J., Kumari, V., Mehrotra, R., et al. (2005). Disruption of learned irrelevance in acute schizophrenia in a novel continuous within-subject paradigm suitable for fMRI. Behavioural Brain Research, 156, 277–288.
Current topics in latent inhibition research Behavior and cognition
2 Latent inhibition and extinction: their signature phenomena and the role of prediction error R. Frederick Westbrook and Mark E. Bouton
Introduction The simplest way to study learning is to expose subjects to a stimulus and then assess whether they show some effect which is absent in subjects lacking that experience (see Rescorla, 1998). One stimulus exposure effect is latent inhibition. Subjects in one group but not another are exposed to a stimulus in the absence of any other scheduled event. Then subjects in both groups are exposed to a signaling relation between that stimulus (the conditioned stimulus (CS)) and a motivationally significant event (an unconditioned stimulus (US)). The responding elicited by the CS in subjects just exposed to the signaling relation is depressed in those pre-exposed to the stimulus. Conditioned responding is said to have been latently inhibited by the prior stimulus-alone exposures. This effect has also been observed in a within-subject design where subjects are first exposed to one stimulus but not to another and then to a signaling relation between each of these stimuli and a US. Responding develops more rapidly to the novel CS than to one that had been pre-exposed (e.g., Killcross & Robbins, 1993; Rescorla, 2002a, 2002b). Another effect of stimulus exposures is extinction. Two groups of subjects are exposed to a signaling relation between a CS and US. Then subjects in one group but not the other are exposed to the CS in the absence of any other scheduled event. The responding elicited by the CS in subjects just exposed to the signaling relation is depressed in those that additionally received the CS-alone exposures. The CS-alone exposures are said to have extinguished conditioned responding. This effect has also been observed in a within-subject design where subjects are exposed to two stimuli each of which signals a US. Then all subjects are exposed to one but not the other CS in the absence of any other event. A subsequent test reveals more conditioned responding to the CS just subjected to the signaling relation than to the one that had also been exposed alone (see Rescorla, 2004). Latent inhibition and extinction are well-documented effects. Each has been observed across a range of stimulus modalities in both appetitive and aversive Latent Inhibition: Cognition, Neuroscience and Applications to Schizophrenia, ed. R.E. Lubow and Ina Weiner. Published by Cambridge University Press # Cambridge University Press 2010.
23
24
R. Frederick Westbrook and Mark E. Bouton
conditioning procedures, and in a number of animal species, including people. Each is characterized by a depression of conditioned responding produced by simple stimulus exposures, differing in whether these exposures depress new (latent inhibition) or existing (extinction) conditioned responding. In spite of such commonalities, latent inhibition and extinction usually receive different explanations and, as a consequence, the search for their neural substrates has proceeded independently of each other. For example, latent inhibition has commonly been explained as a decrease in attention to or in the functional salience of the CS (e.g., Mackintosh, 1973; Pearce & Hall, 1980; Wagner, 1981), whereas changes in attention or salience have rarely been invoked to explain extinction (but see Robbins, 1990). On the other hand, both latent inhibition and extinction are often assumed to result from an interference effect at the level of learning. For example, the end product of decreases in attention or salience in latent inhibition is that it allows less associative strength to accrue to the CS during conditioning. It results in an acquisition deficit. Similarly, models which have identified learning with a single construct such as associative strength or connection weight have often assumed that extinction is due to erasure of the original association (Rescorla & Wagner, 1972) or reversal of the connection weights (Rumelhart, Hinton, & Williams, 1986). Essentially, breaking the signaling relation was held to destroy the learning produced by that relation. There is now considerable evidence that this focus on such learning or encoding mechanisms is incomplete, especially for extinction. For example, a wide variety of evidence, some of which will be presented in the next section, suggests that much if not all of the original learning survives CS-alone exposures in spite of the fact that the CS fails to elicit responding. This implies that CS-alone exposures produce new learning that depresses conditioned responding while leaving the original learning more or less intact (Bouton, 1993). We will suggest that similar evidence is available for latent inhibition. Extinction, like latent inhibition, is due to the learning produced by stimulus-alone exposures. But what is the nature of the learning that underlies extinction and how does this differ from that which underlies latent inhibition? We compare and contrast extinction and latent inhibition with respect to two questions. The first question concerns the contents of the learning produced by stimulus exposures and how this learning is expressed in behavior. We focus here on the signature phenomena characteristic of extinction and latent inhibition and on the role accorded “context” in retrieval of the contrasting memories produced by stimulus exposure and conditioning. The second question concerns the conditions that produce the learning that underlies each of these effects. We focus on the role of prediction error, the discrepancy between the actual and predicted outcome of a stimulus presentation, a process that has been invoked to explain a range of acquisition phenomena in Pavlovian conditioning. We argue that the learning which underlies latent inhibition and extinction is similar in terms of content and conditions, that each involves “context” controlled retrieval of the associations formed across stimulus- and
Latent inhibition and extinction
25
CS-alone exposures, and that each of these associations is regulated by prediction error. We conclude with some qualifications and implications of these arguments, including comments on the search for the neural substrates of extinction and latent inhibition.
Signature phenomena of extinction What is the nature of the learning which underlies extinction? A clue comes from the various signature phenomena exhibited by an extinguished CS. One such phenomenon is “renewal”. This refers to the finding that the ability of an extinguished CS to depress conditioned responding depends critically on context. A CS trained in one context (A) and extinguished in a second context (B) elicits conditioned responding when subsequently tested in A but not in B (Bouton & Bolles, 1979a). This effect has also been observed in a so-called ABC procedure, where conditioning, extinction, and testing occur in contexts A, B, and C. For example, in an experiment by Harris et al. (2000), subjects were separately trained with two CSs, CS1 and CS2. Then one of these CSs (e.g., CS1) was extinguished in a second context (e.g., B) while the other CS (CS2) was extinguished in a third context (C). Subsequent testing of CS1 and CS2 showed less responding to each CS when tested in its own extinction context (e.g., CS1 in B) than when tested in the other CS’s extinction context (e.g., CS1 in C). A second phenomenon is “reinstatement”. This refers to the finding that responding to an extinguished CS is restored when subjects are tested in a context where US-alone exposures have occurred (Bouton & Bolles, 1979b; Rescorla & Heth, 1975; Westbrook et al., 2002). A third signature phenomenon is “spontaneous recovery” of extinguished responding. This refers to the finding that the conditioned responding which is depressed by CS-alone exposures recovers with the elapse of time (Pavlov, 1927; Rescorla, 2004). For example, subjects trained with two CSs (A and B), extinguished to A at one point in time and to B at a later point, show more responding to A than B when tested shortly after extinction of B (Leung & Westbrook, 2008; Rescorla, 1997). A fourth signature phenomenon is “rapid reacquisition”. Here, CS–US pairings after extinction can create a rapid return to conditioned responding (Napier, Macrae, & Kehoe, 1992). Interestingly, under some conditions, including the delivery of a large number of extinction trials and/or a low number of conditioning trials, reacquisition can be slower than original learning (Bouton, 1986; Bouton & Swartzentruber, 1989; Ricker & Bouton, 1996; Leung, Bailey, Laurent, & Westbrook, 2007). Slow reacquisition could result from the attentional and salience changes that are commonly invoked in latent inhibition; for example, as the CS comes to predict “no US” accurately, its associability could decline (Pearce & Hall, 1980). However, Bouton and Swartzentruber (1989) showed that a return to the extinction context after reconditioning suppressed performance, suggesting that retrieval of extinction is sufficient to re-suppress reacquired responding. And other work has shown that
26
R. Frederick Westbrook and Mark E. Bouton
rapid reacquisition itself can be slowed by presenting occasional conditioning trials (CS–US pairings) among many extinction trials, in a way that might make them a feature of both extinction and conditioning (Bouton, Woods, & Pineno, 2004). Thus, rapid reacquisition may be another renewal effect in which reintroduced CS–US pairings restore part of the context of conditioning and yield a recovery of responding. Bouton (1993) has used the signature phenomena to develop a memory-based account of conditioning and extinction. This holds that conditioning and extinction are represented in memory as distinct episodes with contrasting content: a CS–US memory that is held to be encoded relatively free of contextual influences and a CS–no-US memory which is tightly encoded by context. Critically, “context” does not just include the physical cues comprising the place where extinction occurred but also temporal as well as internal cues and emotional states. Thus, renewal and spontaneous recovery occur because subjects are tested outside the physical and temporal cues against which the CS–no-US memory was encoded. Reinstatement occurs because the US-alone exposures bias retrieval of the CS–US memory. The bias is a result of the shift from the conditions under which the CS–no-US memory was formed to the conditions (of context conditioning) under which the CS–US memory was formed. This account thus proposes that subjects emerge from extinction with two contrasting memories and, hence, that subsequent responding is determined by which memory is retrieved. Support for this account comes from the demonstrations that the spontaneous recovery and renewal of conditioned responding were attenuated by preceding the test with exposure to a cue that had been present across the course of extinction (Brooks & Bouton, 1993, 1994). Such a cue is thought to return the subjects to the conditions (cue present) under which the CS–no-US memory was formed. This biases retrieval of the CS–no-US memory and thereby counters the retrieval of the CS–US memory normally produced by the change in temporal or physical context. According to this account, subjects represent the non-occurrence of the predicted US across CS-alone exposures in terms of a no-US memory and presumably form excitatory associations between the CS and that memory. Subsequent activation of this association interferes at the level of retrieval with the original association formed between the CS and US. A variant of this idea (e.g., Bouton, 1994) is that the association is inhibitory in nature so that CS-alone exposures come to depress the representation of the US otherwise elicited by the excitatory association formed at conditioning. However, either version seems to require some degree of specificity with respect to the US whose omission is signalled. For instance, a CS trained with one US and extinguished can be retrained with a different US. Nevertheless, such a CS remains just as sensitive to devaluation of the original US as one that had been trained with that US and not extinguished (Rescorla, 1996). This shows that extinction does not erase the original association, but it also implies that extinction had produced a no-US memory or an inhibitory association that was specific to the original US.
Latent inhibition and extinction
27
Signature phenomena of latent inhibition What is the nature of the learning that underlies latent inhibition? Again a clue comes from the various signature phenomena exhibited by a latently inhibited CS. One of the most interesting, especially given the context-dependence of extinction, is that latent inhibition is also highly context-specific. Several investigators have reported that the latent inhibitory effect present when subjects are pre-exposed, conditioned and tested in the same context (AAA) is attenuated when they are pre-exposed in one context but conditioned and tested in a second context (BAA) (e.g., Hall & Honey, 1989). Moreover, just as extinction does not erase the learning produced by conditioning, conditioning does not erase the learning produced by pre-exposure. For instance, the subjects that show responding when pre-exposure is in one context while conditioning and testing is in a second context (BAA) show latent inhibition when returned to B for testing (BAB; Bouton & Swartzentruber, 1989; Westbrook et al., 2002). The latent inhibition produced by the return shift to the pre-exposure, like extinction, is specific to the stimulus pre-exposed in that context. Westbrook et al. (2000) pre-exposed one stimulus (CS1) in context A and a second stimulus (CS2) in context B. Then each stimulus was trained with a US in a third context. Subsequent testing revealed greater responding when each CS was tested in the other CS’s preexposure context (e.g., CS1 in context B) than when tested in its own pre-exposure context (CS1 in context A). A second characteristic of latent inhibition is its transience or vulnerability to the effects of time. First, an interval introduced between the pre-exposure and conditioning phases weakens the latent inhibition effect (e.g., Hall & Schachtman, 1987). Second, and even more interestingly, conditioned responding that is depressed when testing occurs shortly after conditioning a pre-exposed stimulus (Westbrook et al., 2002) or context (Killcross et al., 1998; Leung et al., 2007) spontaneously recovers when testing occurs some time after that conditioning (see also Aguado, Symonds, & Hall, 1994). The recovery of suppressed responding over time implies the role of a short-term deficit in performance in producing latent inhibition, rather than a deficit in acquiring associative strength or learning. (A performance deficit is also implied by the finding that, under somewhat different circumstances, latent inhibition can increase over time after conditioning; see Lubow & De la Casa (2005)). Latent inhibition is lost not only with the elapse of time but also when US-alone exposures are interpolated between conditioning a pre-exposed CS and testing (Kasprow et al., 1984). However, it remains to be determined whether the latent inhibition which is normally lost with the elapse of time is, like extinction (Brooks & Bouton, 1993), recovered when testing is preceded by exposure to a cue present across pre-exposure. Gordon and Weaver (1989) have at least shown that a retrieval cue, introduced at the onset of the conditioning stage, recovers latent inhibition when it is attenuated by a context switch from the preexposure to the conditioning phase (i.e., BAA).
28
R. Frederick Westbrook and Mark E. Bouton
As noted previously, Bouton (1993) has proposed that conditioning and extinction are represented in memory as distinct episodes with contrasting content: a CS–US memory that is relatively independent of context and a CS–no-US or an inhibitory CS–US memory that is closely tied to context. But he has also proposed that exactly the same is the case with respect to stimulus pre-exposure (see also Miller, Kasprow, & Schachtman, 1986; Spear, 1981). Here subjects encode the stimulus as one that signals nothing, and that is again closely tied to the physical and temporal context where this stimulus–nothing association was formed. Such subjects then encode the CS–US memory while retaining the association produced by pre-exposure. Hence, as with extinction, subsequent responding is determined by whether the CS elicits the US memory formed at conditioning or the CS–nothing memory formed during preexposure. According to this proposal, a long retention interval between conditioning and test produces recovery from the depressing effects of pre-exposure because it changes the temporal context in which the CS–nothing memory was formed. Similarly, pre-exposure in one context while conditioning and testing in a second context attenuates latent inhibition because the test context is unable to retrieve the CS–nothing memory formed in pre-exposure. Finally, the return shift from the conditioning context to a test session in the pre-exposure context reinstates latent inhibition because the test context now retrieves the CS–nothing memory.
The role of prediction error in extinction Contemporary models of Pavlovian conditioning explain a range of acquisition phenomena by proposing that associative formation is regulated by prediction error and that all the cues present are used to calculate this error. Such models thus hold that the omission of a predicted US constitutes the error which drives the learning that underlies extinction and that the amount of this learning is proportional to the size of the error. Evidence consistent with the proposal that all the cues present contribute to prediction error was provided by Wagner (1969), who reported greater depression of responding to a weakly trained target CS extinguished in compound with a well-trained CS than with a moderately trained one; and by Rescorla (2000, 2006) who found greater depression of responding to a target CS when extinguished in compound with another excitatory CS than when extinguished alone. In a complementary way, compounding an excitatory CS with an inhibitory CS during extinction can reduce associative loss to the excitor (“protection from extinction”; e.g., Rescorla, 2003). Furthermore, negative prediction error is sufficient to cause extinction, as in the “overexpectation effect” in which two separately conditioned CSs are presented in a compound and together paired with the US – despite the continued CS–US pairings, responding to each CS declines (e.g., Kremer, 1978; Lattal & Nakajima, 1998). More recently, Leung and Westbrook (2008) used a within-subject design to study the depressive effect produced by subjecting an extinguished CS that was or was not
Latent inhibition and extinction
29
undergoing spontaneous recovery to additional extinction. They reasoned that additional extinction of a CS undergoing spontaneous recovery would produce greater prediction error than one not undergoing recovery, and hence produce better learning with the additional extinction. They confirmed just such an effect across a subsequent common test of each CS. Leung and Westbrook (2008, Experiment 5) also showed that this effect is regulated by a common error term. They trained four CSs (A, B, C, and D) and then subjected A to extinction at one point in time and B to extinction at a later point. Shortly after extinction of B, all subjects received extinction of a compound composed of an extinguished CS undergoing recovery, and of a non-extinguished CS (AC), and one composed of an extinguished CS not undergoing recovery, and a non-extinguished CS (BD). Subsequent testing revealed greater depression of responding to the CS (C) extinguished in compound with the CS (A) undergoing spontaneous recovery than to the CS (D) extinguished in compound with a CS not undergoing recovery. These results show that the learning underlying extinction is regulated by prediction error, and that all the cues present contribute to the calculation of this error. However, the manner in which error acts on this learning remains unclear. One explanation is that prediction error acts directly on the learning which underlies extinction. According to this explanation, the results reported by Leung and Westbrook (2008, Experiment 5) are due to better learning of C when it is combined with an extinguished CS undergoing recovery (A) as compared to learning about D when it is combined with an extinguished CS not undergoing recovery. Learning about C would be greater than D because the error produced by extinction of the AC compound was greater than the error produced by extinction of the BD compound (e.g., Wagner, 1981). However, a second explanation is that prediction error acts indirectly on learning by regulating attention across subsequent exposures to each of the compounds (Pearce & Hall, 1980). According to this explanation of the results reported by Leung and Westbrook, prediction error was less across extinction of the BD than the AC compound because the presence of the recently extinguished B was a better signal of the outcome of the compound presentations than was the remotely extinguished A. Hence, attention to D declined more rapidly across the compound extinction trials than did attention to C. The first consequence of this difference in the decline in attention was that there was less new learning about D than C across the compound extinction trials. The second consequence of this difference in new learning was that D was better able to elicit its original association with the US than was C. Hence, responding on test was greater to D than to C.
The role of prediction error in latent inhibition Does prediction error regulate what is learned across stimulus pre-exposures? Interestingly, the classic models of latent inhibition do propose a role for error correction. For example, Mackintosh (1975) assumes that attention to a pre-exposed CS declines
30
R. Frederick Westbrook and Mark E. Bouton
because it is no better than the context at predicting no US. Pearce and Hall (1980) similarly suppose that attention to the pre-exposed stimulus will decline if it and the context predict no US. And Wagner’s short-term memory model (e.g., 1978) assumes that the association between the context and the CS which controls latent inhibition is driven by the discrepancy between the CS and the extent to which the context already predicts it (e.g., Wagner, 1978, p. 206). A variation of this suggestion is that learning about a context is itself regulated by prediction error. For instance, McLaren and Mackintosh (2000; McLaren, Kaye, & Mackintosh, 1989) have suggested that subjects detect the several cues comprising a context (its shape, floor texture, smell, etc.) and form associations among them via the Rescorla–Wagner delta rule (Rescorla & Wagner, 1972). Thus, cues (e.g., A and B) that are salient and detected together acquire strong associations with each other so that detection of one predicts the other. This means that other less salient cues (e.g., C) will be impaired in being associated with B when they are detected in conjunction with A; that is, A effectively “blocks” the learning of an association between C and B. Learning about a context–CS association during CS pre-exposure will also be controlled by the delta rule so that the amount learned about the context–CS association will be regulated by the amount already learned about the context. Pre-exposure to a context will undermine its association with a CS during pre-exposure, and thus reduce latent inhibition. In contrast, exposure to a to-be-conditioned stimulus in a novel context will allow the stimulus to be integrated with the context representation, thereby rendering it part of the “background” and making its subsequent conditioning difficult (Lubow, Alek, & Rifkin, 1976; McLaren et al., 1994). Another approach is possible. The presentation of any novel stimulus could generate an implicit “prediction” that it will have a potential consequence. (That is why it might elicit an orienting response.) Such a prediction is innate or based on a history in which sudden unexpected stimuli are followed by outcomes: a sound is followed by the appearance of something or vice versa. Thus, a subject that is hungry or thirsty processes the new stimulus in terms of its relevance to the current motivational state; a novel stimulus could predict food or fluid. The subject might additionally process the stimulus in terms of its relevance to danger. Hence, the discrepancy between such a prediction and what actually occurs (nothing) may drive the learning which underlies its latent inhibitory effect. This learning may be relatively specific to the subject’s motivational state so that what is learned about a stimulus when the subject is hungry may not transfer to when it is thirsty. But learning that such a stimulus is irrelevant to danger should be general and independent of hunger and thirst. There is in fact evidence for such motivational specificity. For instance, Killcross and Balleine (1996) reported that rat subjects exhibited a depression of conditioned responding when a pre-exposed stimulus signaled an appetitive US but only when pre-exposure and conditioning occurred under the same motivational state. Rats preexposed in one state (e.g., hunger) failed to show a latent inhibitory effect when made thirsty and exposed to a signaling relation between that stimulus and a fluid US.
Latent inhibition and extinction
31
This evidence is consistent with the possibility that the animal learns something like a CS–no-food (or CS–no-fluid) relationship during stimulus pre-exposure. It also suggests that what is learned about a stimulus across exposures can be relatively specific to the subject’s motivational state. However, there is little evidence as to whether this learning is regulated, as extinction learning is, by prediction error. An exception is a recent study by Rodriguez and Hall (2008). These investigators exposed rats to an odor either in isolation or in compound with a taste and then made the rats sick by injection of a drug (lithium chloride) following exposure to that odor. They found a greater latent inhibitory effect when the odor had been preexposed in compound with the taste than when pre-exposed alone. An additional experiment arranged so that rats were exposed to one odor in isolation and to another odor in compound with the taste. Then half of the rats were made sick after exposure to one odor and the remainder were made sick after exposure to the other. Latent inhibition was again greater to the odor pre-exposed in compound with the taste than to the odor pre-exposed alone. These results were seen as consistent with the Pearce–Hall model (1980), in which attention on trial n is determined by the difference on trial n 1 between the summed prediction of all CSs on the trial and the actual trial outcome, which is 0 during the stimulus pre-exposure phase. According to this mechanism, attention decreases to zero over fewer trials in the compound condition because the combined salience of two CSs allows the compound to predict the outcome (nothing or no event) over fewer trials than when one CS is presented alone. The results are also consistent, however, with the view that prediction error on trial n 1 acts directly on the learning produced by that trial rather than regulating attention on the subsequent trial n. As noted previously, Bouton (1993) has proposed that what is learned in pre-exposure consists in a stimulus–nothing or stimulus–noevent association. Thus, error can be identified with the discrepancy between the outcome predicted by all the cues present on a trial (something or some event) and the actual outcome (nothing or no event). Exposure to the taste–odor compound would generate a greater prediction, and thus greater prediction error, than would the odor alone, and so produce a greater odor–nothing or odor–no-event association than when the odor was exposed in isolation. One way to discriminate between these contrasting explanations is to precede the odor–taste compound exposures by taste-alone exposures. According to the direct learning view, since the taste already signals the outcome (nothing), its presence should block the association otherwise formed between the odor and the outcome (nothing) across the compound exposures. Thus, the odor presented in compound with an already exposed taste should fail to acquire the association (with nothing) that impairs subsequent conditioning. Such an odor should condition more rapidly than an odor just exposed in compound with the taste. Alternatively, according to the indirect (attentional change) view, exposure to the taste in advance of the odor–taste compound would allow the compound to predict the outcome (nothing or no event) on the first compound trial, leading to a decline in attention to the added odor on the
32
R. Frederick Westbrook and Mark E. Bouton
subsequent trial. This decline would occur more rapidly than in the case where the odor–taste compound was not preceded by taste-alone exposures. Hence, an odor exposed in compound with a pre-exposed taste would undergo a greater loss in attention and hence be more latently inhibited than an odor exposed in compound with a novel taste. Holtzman, Siette, Holmes, and Westbrook (in press) have confirmed this unique implication of the attentional change view. In a within-subject design, rats were preexposed to one stimulus (A) and 1 week later to a second stimulus (B). Shortly after exposures to B, rats received exposures to compounds composed of each of these stimuli and a novel stimulus (AC and BD). Rats were then conditioned to C and D and tested with each of these stimuli. Responding (freezing) was less to D than to C, that is, there was more latent inhibition to the stimulus (D) exposed in compound with a recently presented stimulus (B) than there was to the stimulus (C) exposed in compound with the remotely presented stimulus (A). Essentially, the recently presented stimulus (B) predicted the outcome of the compound trials better than the remotely presented stimulus (A) and, hence, produced a more rapid decline in attention to its associate (D) than the decline to the associate (C) of the remotely presented stimulus (A). As noted above, the idea that prediction error regulates learning directly predicts the opposite result. In addition to being consistent with an attentional model, these results suggest that the learning underlying latent inhibition, like that underlying extinction, is regulated by prediction error and that all the cues present are used to calculate this error. In each case, the size of the error between the prediction generated by all the cues present (a potential event in the case of latent inhibition and the US in the case of extinction) and the actual outcome (nothing or no US) drives encoding of the stimulus–nothing or CS–no-US associations. The consequence of this encoding, according to the Pearce–Hall model, is a decline in subsequent attention across stimulus-alone and CS-alone exposures. However, there is a difference between extinction and latent inhibition, which can be illustrated with reference to the compound extinction experiment reported by Leung and Westbrook and described above. In that experiment, extinction of a compound composed of an excitatory CS and a recently extinguished CS (BD) was interpreted to have produced a decline in attention to D because the presence of the recently extinguished B across extinction of the BD compound trials predicted the outcome (no US). The additional consequence of this decline in attention to D was the undermining of its association with the no-US, and, hence, a greater preservation of its original association with the US. In the unpublished latent inhibition experiment just described (Holtzman et al., in press), exposures to a compound composed of a novel stimulus (D) and a recently pre-exposed stimulus (B) were also interpreted to have produced a decline in attention to D because the presence of the recently pre-exposed B across exposures to the compound predicted the outcome (nothing or no event). Thus, a stimulus can undergo a decline in attention because it signals nothing or no outcome (as in the
Latent inhibition and extinction
33
case of the pre-exposed B stimulus) or because it accompanies such a stimulus (as in the case of the novel D stimulus). However, other than the subsequent effect on conditioning, there is no other consequence of this decline in attention to D across pre-exposures. A stimulus which has undergone a decline in attention is effectively one that predicts nothing or no event. We note that an implication of the decline in attention held to occur when a stimulus reliably signals a null outcome (e.g., nothing or no event) is that a preexposed stimulus will be impaired in entering new associations, either excitatory or inhibitory. Indeed, a pre-exposed CS is not just retarded in developing conditioned responding when it signals a US: it is also retarded in developing inhibition of conditioned responding when it signals the absence of a predicted US (e.g., Rescorla, 1971). Thus, the slowing of associative change produced by the learned loss in attention contributes to the impairment in the development of conditioned responding when a pre-exposed CS signals a US (e.g., Rescorla, 2002b). Moreover, this loss may be sufficiently great that it overcomes any facilitation in the development of conditioned inhibition produced by the existing stimulus–no-event association (see Hall, 1991, p. 136). However, we also note that the associations formed between a pre-exposed CS and a US do not wipe out those formed in pre-exposure. Therefore, these new CS–US associations will also have to compete for expression in performance with the stimulus–nothing or stimulus–no-event associations formed across exposures in the manner described previously (and consistent with performance deficit results). We also note that shifts of physical or temporal context may restore attention since the stimulus is presented outside the context of its encoding and outside the context that serves to retrieve its associate (nothing or no event) that competes with its US association for expression in performance. In the long run, the overall pattern of findings in latent inhibition may be consistent with the view that prediction error might cause both attentional changes (which could produce a partial learning deficit) and associative stimulus–nothing changes (which would create the opportunity for a performance deficit).
Conclusions and implications The study of any form of learning requires consideration of various questions. These include questions about its contents – what is learned and how it is expressed in behavior – and about the conditions under which the learning occurs (see Dickinson, 1980; Rescorla, 1988). We have considered these questions with respect to the learning that underlies extinction and latent inhibition. We have argued that the contents of this learning are similar. Specifically, exposure to the CS in the absence of the US results in subjects encoding this relation in terms of a CS–no-US association or an inhibitory CS–US association; exposure to a stimulus in the absence of any other event likewise results in subjects encoding this relation in terms of a stimulus– nothing or an inhibitory stimulus–event association. The contents of extinction and
34
R. Frederick Westbrook and Mark E. Bouton
latent inhibition thus differ with respect to the specificity of the associate: in extinction, the CS comes to excite a specific no-US representation or to inhibit a specific US representation; in latent inhibition, the stimulus comes to activate (perhaps in a motivationally specific manner) a no-event representation or to inhibit such an event representation. We have also proposed that these contents are closely tied to the cues (physical, temporal, and internal, including emotional states) present across the CS-alone or stimulus-alone exposures. Thus, subsequent behavior is determined by the presence or absence of these cues: their presence retrieves the memory formed across exposures whereas their absence allows the CS to excite its US associate. Finally, in line with contemporary models of acquisition, prediction error is held to regulate associative formation across CS and stimulus-alone exposures. In extinction, the omission of the predicted US is the error which drives formation of the CS–noUS or inhibitory CS–US association; in latent inhibition, the absence of any event is the error which violates the innate or the previously learned expectation generated by the stimulus. In each case, all the cues present are used to compute this error, but an active research question is whether error acts directly on what was learned about these cues on a given trial or indirectly by regulating attention to the cues on subsequent trials (e.g., Holtzman et al., in press). Finally, as in the case of associative formation between a CS and US, it is possible that prediction error acts on both – directly changing associative strength and, in so doing, reducing attention across subsequent trials. A major implication of the view presented here is that the learning which underlies latent inhibition and extinction is similar, differing in the specificity of the event predicted by the stimulus (some event) and the CS (the associated US). The learning produced by stimulus exposures “extinguishes” its prediction of a potential consequence, and the learning produced by extinction can “latently inhibit” reconditioning (Bouton, 1986). Let us briefly consider some implications of the claimed similarities and differences in this learning. If the contents of what is learned across stimulus- and CS-alone exposures differ merely in the specificity of the associate (a no-US memory or a no-event one), extinction of a pre-exposed CS will differ from extinction of CS just subjected to conditioning. For instance, extinction of a pre-exposed CS will restore the conditions under which the pre-exposure memory was formed – the preexposed and conditioned CS is again presented alone – and thereby facilitate the loss of conditioned responding. Moreover, since the contents of what is learned in preexposure and extinction are each tightly controlled by time and place, the restoration of conditioned responding by shifts in place and time should be greater when a CS has been subjected to both pre-exposure and extinction. Essentially, such a shift will increase the ability of the CS to retrieve the US memory since it suffers less interference from both the pre-exposure and the extinction contexts. Finally, we note that the searches for the neural substrates of extinction and latent inhibition have proceeded largely independently of each other. This was doubtless due in part to the treatment of these phenomena as distinct, mediated by different processes with
Latent inhibition and extinction
35
different content. Thus, the search for the substrates of latent inhibition has focused on the role of ascending dopamine projections from the ventral tegmental area to the nucleus accumbens and prefrontal cortex in regulating attention or in switching between the control exerted by the pre-exposure and conditioning memories; the search for the substrates underlying extinction (especially of fear extinction) has concentrated on the role of glutamate projections from the infra-limbic cortex to GABAergic neurons in the intercalated cell bodies of the amygdala in regulating inhibition. This search has provided valuable information, but the view presented here suggests that the neural substrates of these phenomena might also overlap to some extent, and that the circuits known to be involved in one might also be involved in the other. But there are also differences in the learning underlying latent inhibition and extinction. One critical difference is that stimulus exposures in latent inhibition extinguish a prediction of a “potential consequence” whereas CS-alone exposures in extinction extinguish the prediction that a particular consequence (the US) will occur. One result of this difference is that emotional effects may be more important in the latter. That is, whereas exposure to a CS in the absence of its appetitive US or of an aversive one produces frustration or (perhaps) relief, stimulus-alone exposures are unlikely to elicit any such emotive reactions. Moreover, the emotional reactions (e.g., frustration and, perhaps, relief) elicited by the omission of the predicted US (appetitive and aversive, respectively) may themselves enter into associations with the CS and thereby contribute to extinction, as proposed by some classic (Amsel, 1962) and current (Rescorla, 2001) theories. Such emotional reactions may also become part of the “context” under which extinction occurs and, like other such cues, subsequently retrieve the CS–no-US memory or excite the inhibitory CS–US one (Bouton, 1993). In contrast, subsequent retrieval of the stimulus–nothing memory in latent inhibition may be less dependent on emotional cues. In short, the omission of a predicted US in extinction, but not of a “potential consequence” during latent inhibition, may be represented at the level of affect and encoded with other information learned about the CS.
References Aguado, L., Symonds, M., & Hall, G. (1994). Interval between preexposure and test determines the magnitude of latent inhibition: implications for an interference account. Animal Learning & Behavior, 22, 188–194. Amsel, A. (1962). Frustrative nonreward in partial reinforcement and discrimination learning: some recent history and a theoretical extension. Psychological Review, 69, 306–328. Bouton, M. E. (1986). Slow reacquisition following the extinction of conditioned suppression. Learning and Motivation, 17, 1–15. Bouton, M. E. (1993). Context, time and memory retrieval in the interference paradigms of Pavlovian learning. Psychological Bulletin, 114, 80–99.
36
R. Frederick Westbrook and Mark E. Bouton
Bouton, M. E. (1994). Context, ambiguity, and classical conditioning. Current Directions in Psychology Science, 3, 49–53. Bouton, M. E., & Bolles, R. C. (1979a). Contextual control of the extinction of conditioned fear. Learning and Motivation, 10, 445–466. Bouton, M. E., & Bolles, R. C. (1979b). Role of conditioned contextual stimuli in reinstatement of extinguished fear. Journal of Experimental Psychology: Animal Behavior Processes, 5, 368–378. Bouton, M. E., & Swartzentruber, D. (1989). Slow reacquisition following extinction: context, encoding and retrieval mechanisms. Journal of Experimental Psychology: Animal Behavior Processes, 15, 43–53. Bouton, M. E., Woods, A. M., & Pineno, O. (2004). Occasional reinforced trials during extinction can slow the rate of rapid reacquisition. Learning and Motivation, 35, 371–390. Brooks, D. C., & Bouton, M. E. (1993). A retrieval cue for extinction attenuates spontaneous recovery. Journal of Experimental Psychology, 19, 77–89. Brooks, D. C., & Bouton, M. E. (1994). A retrieval cue for extinction attenuates response recovery (renewal) caused by a return to the conditioning context. Journal of Experimental Psychology: Animal Behavior Processes, 20, 366–379. Dickinson, A. (1980). Contemporary Animal Learning Theory. Cambridge: Cambridge University Press. Gordon, W. C., & Weaver, M. S. (1989). Cue-induced transfer of CS preexposure effects across contexts. Animal Learning & Behavior, 17, 409–417. Hall, G. (1991). Perceptual and Associative Learning. Oxford: Oxford University Press. Hall, G., & Honey, R. C. (1989). Attenuation of latent inhibition after compound pre-exposure: associative and perceptual explanations. The Quarterly Journal of Experimental Psychology, 41B, 355–368. Hall, G., & Schachtman, T. R. (1987). Differential effects of a retention interval on latent inhibition and the habituation of an orienting response. Animal Learning & Behavior 15, 76–82. Harris, J. A., Jones, M. L., Bailey, G. K., & Westbrook, R. F. (2000). Contextual control over conditioned responding in an extinction paradigm. Journal of Experimental Psychology: Animal Behavior Processes, 26, 174–185. Holtzman, O., Siette, J., Holmes, N. M., & Westbrook, R. F. (in press). Reversing the latent inhibitory effects of recent and remote stimulus presentations: the role of error correction mechanisms. Journal of Experimental Psychology: Animal Behavior Processes. Honey, R. C., & Hall, G. (1989). Attenuation of latent inhibition after compound pre-exposure: associative and perceptual explanations. The Quarterly Journal of Experimental Psychology, 41B, 355–368. Kasprow, W. J., Catterson, D., Schachtman, T. R., & Miller, R. R. (1984). Attenuation of latent inhibition by post-acquisition reminder. The Quarterly Journal of Experimental Psychology, 36B, 53–63. Killcross, S., & Balleine, B. (1996). Role of primary motivation in stimulus preexposure effects. Journal of Experimental Psychology: Animal Behavior Processes, 22, 32–42. Killcross, A. S., Kiernan M. J., Dwyer, D., & Westbrook, R. F. (1998). Loss of latent inhibition of contextual conditioning following non-reinforced context exposure in rats. The Quarterly Journal of Experimental Psychology, 51B, 75–90.
Latent inhibition and extinction
37
Killcross, A. S., & Robbins, T. W. (1993). Differential effects of intra-accumbens and systemic amphetamine on latent inhibition using an on-baseline, within-subject conditioned suppression paradigm. Psychopharmacology, 110, 479–489. Kremer, E. F. (1978). The Rescorla-Wagner model: losses in associative strength in compound conditioned stimuli. Journal of Experimental Psychology: Animal Behavior Processes, 4, 22–36. Lattal, K. M., & Nakajima, S. (1998). Overexpectation in appetitive Pavlovian and instrumental conditioning. Animal Learning & Behavior, 26, 351–360. Leung, H. T., Bailey, G. K., Laurent, V., & Westbrook, R. F. (2007). Rapid reacquisition of fear to a completely extinguished context is replaced by transient impairment with additional extinction training. Journal of Experimental Psychology: Animal Behavior Processes, 33, 299–313. Leung, H. T., & Westbrook, R. F. (2008). Spontaneous recovery of extinguished fear responses deepens their extinction: a role for error-correction mechanisms. Journal of Experimental Psychology: Animal Behavior Processes, 34, 461–474. Lubow, R. E., Alek, M., & Rifkin, B. (1976). The context effect: the relationship between stimulus preexposure and environmental preexposure determines subsequent learning. Journal of Experimental Psychology: Animal Behavior Processes, 2, 38–47. Lubow, R. E., & De la Casa, L. G. (2005). There is a time and a place for everything: bidirectional modulations of latent inhibition by time-induced context differentiation. Psychonomic Bulletin & Review, 12, 806–821. Mackintosh, N. J. (1973). Stimulus selection: learning to ignore stimuli that predict no change in reinforcement. In R. A. Hinde & J. S. Stevenson-Hinde (Eds.), Constraints on Learning: Limitations and Predispositions. Cambridge, UK: Academic Press, pp. 75–100. Mackintosh, N. J. (1975). A theory of attention: variations in the associability of stimuli with reinforcement. Psychology Review, 82, 276–298. Mclaren, I. P. L., Bennett, C., Plaisted, K., & Aitken, M. (1994). Latent inhibition, context specificity, and context familiarity. The Quarterly Journal of Experimental Psychology, 47B, 387–400. Mclaren, I. P. L., Kaye, H., & Mackintosh, N. J. (1989). An associative theory of the representation of stimuli: applications to perceptual learning and latent inhibition. In R. G. M. Morris (Ed.), Parallel Distributed Processing: Implications for Psychology and Neurobiology. New York: Clarendon Press/Oxford University Press, pp. 102–130. Mclaren, I. P. L., & Mackintosh, N. J. (2000). An elemental model of associative learning: I. Latent inhibition and perceptual learning. Animal Learning and Behavior, 28, 211–246. Mercier, P., & Baker A. G. (1985). Latent inhibition, habituation, and sensory preconditioning: a test of priming in short-term memory. Journal of Experimental Psychology: Animal Behavior Processes, 11, 485–501. Miller, R. R., Kasprow, W. J., & Schachtman, T. R. (1986). Retrieval variability: Sources and consequences. American Journal of Psychology, 99, 145–218. Napier, R. M., Macrae, M., & Kehoe, E. J. (1992). Rapid reacquisition in conditioning of the rabbit’s nictitating membrane response. Journal of Experimental Psychology: Animal Behavior Processes, 18, 182–192. Pavlov, I. P. (1927). Conditioned Reflexes. London: Oxford University Press.
38
R. Frederick Westbrook and Mark E. Bouton
Pearce, J. M., & Hall, G. (1980). A model for Pavlovian learning: Variations in the effectiveness of conditioned but not unconditioned stimuli. Psychological Review, 87, 532–552. Rescorla, R. A. (1971). Summation and retardation tests of latent inhibition. Journal of Comparative and Physiological Psychology, 75, 77–81. Rescorla, R. A. (1988). Pavlovian conditioning: it’s not what you think it is. American Psychologist, 43, 151–160. Rescorla, R. A. (1996). Preservation of Pavlovian associations through extinction. The Quarterly Journal of Experimental Psychology, 49B, 245–258. Rescorla, R. A. (1997). Spontaneous recovery after Pavlovian conditioning with multiple outcomes. Animal Learning & Behavior, 25, 99–107. Rescorla, R. A. (1998). Behavioral studies of Pavlovian conditioning. Annual Review Neuroscience, 11, 329–352. Rescorla, R. A. (2000). Extinction can be enhanced by a concurrent excitor. Journal of Experimental Psychology: Animal Behavior Processes, 26, 251–260. Rescorla, R. A. (2001). Experimental extinction. In R. R. Mowre & S. Klein (Eds.), Handbook of Contemporary Learning Theories. Hillsdale, NJ: Erlbaum, pp. 119–154. Rescorla, R. A. (2002a). Savings Tests: separating differences in rate of learning from differences in initial levels. Journal of Experimental Psychology: Animal Behavior Processes, 28, 369–377. Rescorla, R. A. (2002b). Comparison of the rates of associative change during acquisition and extinction. Journal of Experimental Psychology: Animal Behavior Processes, 28, 406–415. Rescorla, R. A. (2003). Protection from extinction. Learning & Behavior, 31, 124–132. Rescorla, R. A. (2004). Spontaneous recovery. Learning and Memory, 11, 501–509. Rescorla, R. A. (2006). Deepened extinction from compound stimulus presentation. Journal of Experimental Psychology: Animal Behavior Processes, 32, 135–144. Rescorla, R. A., & Heth, C. D. (1975). Reinstatement of fear to an extinguished conditioned stimulus. Journal of Experimental Psychology, 104(1), 88–96. Rescorla, R. A., & Wagner, A. R. (1972). A theory of Pavlovian conditioning: variation in the effectiveness of reinforcement and nonreinforcement. In A. H. Black & W. F. Prokasy (Eds.), Classical Conditioning II: Current Theory and Research. New York: Appleton-Century-Crofts, pp. 181–215. Ricker, S. T., & Bouton, M. E. (1996). Reacquisition following extinction in appetitive conditioning. Animal Learning & Behavior, 24, 423–436. Robbins, S. J. (1990). Mechanisms underlying spontaneous recovery in autoshaping. Journal of Experimental Psychology: Animal Behavior Processes, 16, 235–249. Rodriguez, G., & Hall, G. (2008). Potentiation of latent inhibition. Journal of Experimental Psychology: Animal Behavior Processes, 34, 352–360. Rumelhart, D. E., Hinton, G. E., & Williams, R. J. (1986). Learning representations by back-propagating errors. Nature, 323, 533–536. Spear, N. E. (1981). Extending the domain of memory retrieval. In N. E. Spear & R. R. Miller (Eds.), Information Processing in Animals: Memory Mechanisms. Hillsdale, NJ: Erlbaum, pp. 341–378. Wagner, A. R. (1969). Stimulus selection and a “modified continuity theory”. In G. H. Bower & J. T. Spence (Eds.), The Psychology of Learning and Motivation, vol. 3. New York: Academic Press, pp. 1–41.
Latent inhibition and extinction
39
Wagner, A. R. (1978). Expectancies and the priming of STM. In S. H. Hulse, H. Fowler, & W. K. Honig (Eds.), Cognitive Processes in Animal Behavior. Hillsdale, NJ: Erlbaum, pp. 177–209. Wagner, A. R. (1981). SOP: a model of automatic memory processing in animal behavior. In N. E. Spear & R. R. Miller (Eds.), Information Processing in Animals: Memory Mechanisms. Hillsdale, NJ: Erlbaum, pp. 5–47. Westbrook, R. F., Iordanova, M., McNally, G., Richardson, R., & Harris, J. A. (2002). Reinstatement of fear to an extinguished conditioned stimulus: two roles for context. Journal of Experimental Psychology: Animal Behavior Processes, 28, 97–110. Westbrook, R. F., Jones, M., Bailey, G., & Harris, J. A. (2000). Contextual control over conditioned responding in a latent inhibition paradigm. Journal of Experimental Psychology: Animal Behavior Processes, 26, 157–173.
3 Inter-stage context and time as determinants of latent inhibition Luis Gonzalo De la Casa and Oskar Pinen˜o
Introduction: effects of time and context on latent inhibition Despite its apparent simplicity, the phenomenon of latent inhibition (LI) represents one of the most sophisticated and flexible mechanisms that organisms with complex nervous systems have developed through evolution to ensure efficient interaction with the environment. Because the environment is constantly changing, mechanisms that determine the processing of a neutral stimulus depend on a large range of different circumstances. In this chapter we will focus on the role played by two factors, namely, time and context, that seemingly affect LI separately, as well as in combination. The impact of these factors (both apart and conjointly) on LI is still one of the greatest challenges to associative theories of learning. In any learning process there is a series of elements that determine the intensity and type of association that is formed. In the case of classical conditioning, some of the parameters on which Pavlov (1927) concentrated in his original studies were related to the temporal contiguity between stimuli (e.g., whether the stimuli involved in the pairings were presented simultaneously or sequentially, or the order in their presentation when presented sequentially), as well as to the excitatory vs. inhibitory nature of the association acquired under different treatments. Some other elements that have subsequently demonstrated their relevance to associative learning were also pointed out by Pavlov, although sometimes in a quite intuitive manner. For example, he mentioned that conditioned reflexes could be affected by the surrounding stimuli during conditioning in the animal’s environment. Thus, for instance: The environment of the animal, even when shut up by itself in a room, is perpetually changing. Footfalls of a passer-by, chance conversations in neighbouring rooms, slamming of a door or vibration from a passing van, street-cries, even shadows cast through the windows into the room, any of these casual uncontrolled stimuli falling upon the receptors of the dog set up a disturbance in the cerebral hemispheres and vitiate the experiments [. . .] (Pavlov, 1927, Lecture II)
Latent Inhibition: Cognition, Neuroscience and Applications to Schizophrenia, ed. R. E. Lubow and Ina Weiner. Published by Cambridge University Press # Cambridge University Press 2010.
40
Inter-stage context and time as determinants of latent inhibition
41
The role of context for conditioning and associative learning has, since then, always been recognized and accompanied by extensive research (see, e.g., Balsam & Tomie, 1985). In particular, the essential role played by context in the modulation of LI has received extensive empirical and theoretical analysis in recent years (e.g., Hall & Channell, 1985, 1986; Lovibond, Preston, & Mackintosh, 1984; Westbrook, Jones, Bailey, & Harris, 2000). More specifically, using two-stage experimental procedures to induce the LI effect (i.e., preexposure stage with the to-be-conditioned stimulus without consequences, followed by a conditioning stage), it has been clearly established that the change of context between the two stages results in the weakening or even the abolition of the LI effect. More recently, these effects have been explored with three-stage procedures (i.e., preexposure, conditioning, and testing), which allow for a greater number of possible context switches (AAA, ABA, and AAB designs, with each letter representing the context at preexposure, conditioning, and testing stages, respectively). Specifically, Westbrook et al. (2000) demonstrated that conducting preexposure in a context different from that of conditioning and testing (i.e., ABB) increased the conditioned response at test (i.e., attenuation of LI), relative to a condition given no context change throughout training (i.e., AAA). Moreover, if the change of context was limited to the conditioning stage (i.e., ABA) the LI effect was enhanced. The second factor on which we shall focus in this chapter, the effect of the passage of time on previously established associations, was a subject also covered by Pavlov (1927) when he analyzed, for instance, spontaneous recovery, the recovery of conditioned responding to an extinguished CS when testing was delayed. However, after Pavlov’s studies, the interest of researchers of associative learning mainly centered on factors determining the acquisition of associations, effectively sidelining the study of the effects of time on the expression of previously acquired associations. This may explain why the study of the effect of passage of time on learning has usually been linked more strongly to the field of memory than to the study of associative learning (with some exceptions, such as the above-mentioned case of spontaneous recovery or the analysis of the fear-incubation effect, e.g., Eysenck, 1968). As for the effect of the passage of time on LI, empirical findings have been contradictory. The first studies evaluating the role of this variable on LI were conducted in the 1970s and revealed that passage of time by itself did not affect LI (for a review see Lubow, 1989, pp. 67–69). In the 1980s, a series of conditioned taste aversion studies, using a threestage procedure for inducing LI, showed that the introduction of time intervals between conditioning and test stages attenuated LI (e.g., Kraemer & Ossenkopp, 1986; Kraemer & Roberts, 1984). Although these studies did not allow definitive conclusions to be drawn because they presented different tastes across experimental stages, they were soon followed by others which did use the same stimuli throughout the experiment. Under such circumstances, a lengthening of the retention interval resulted in a recovery of the conditioned aversion response. For example, Aguado, Symonds, and Hall (1994) found that when testing was performed 14 days after
42
Luis Gonzalo De la Casa and Oskar Pinen˜o
conditioning LI was weaker than when the test was performed 48 hours after conditioning. This result has also been obtained with other experimental procedures, such as conditioned freezing (e.g., Killcross, Kiernan, Dwyer, & Westbrook, 1998). The recovery of the conditioned response (reduction of LI) with an increase in time between conditioning and test stages has been repeatedly cited as evidence that LI is the result of a failure in information retrieval (e.g., Bouton, 1993; Miller & Matzel, 1988), an account that directly challenged the traditional view whereby LI is the result of a failure in the acquisition of the conditioning association (e.g., Lubow, 1989; see Section 2 for a discussion). The first interpretation of the effect of delayed testing on LI was proposed by Kraemer and Roberts (1984), who considered that passage of time has a differential effect on each of the two associations formed during the preexposure and conditioning stages (CS–nothing and CS–US, respectively). The first, relatively less important association, would tend to lose impact on behavior over time, while the second, relatively more important due to its biological significance, would tend to increase its capacity to control behavior as time elapsed. A different interpretation of the attenuation of LI after the introduction of a delay between conditioning and testing stems from the view that the passage of time acts like a context change (e.g., Bouton, 1993). According to this premise, passage of time would lead to an attenuation of LI similar to that observed when the physical context is changed between conditioning and testing stages (e.g., Lovibond et al., 1984; Westbrook et al., 2000). Despite the success of this view in explaining the reported data, other studies using similar procedures soon showed that this manipulation did not always produce the attenuation of LI. On some occasions, the LI effect remained intact despite the introduction of the delay (e.g., Alvarez & Lopez, 1995). Even more critically, the opposite effect was found: super-LI was observed when a long timeinterval (21 days) was introduced between conditioning and testing (e.g., De la Casa & Lubow, 2000, 2002, 2005; Lubow & De la Casa, 2002). For example, De la Casa and Lubow (2000) found a decrease in aversive responding elicited by the conditioned taste (i.e., an increase in LI) when the test was delayed by 21 days compared to a group in which the test was delayed by only one day. Similar results have been obtained in other studies using conditioned taste aversion (e.g., De la Casa & Lubow, 2000, 2002, 2005), as well as with conditioned suppression procedures (e.g., Wheeler, Stout, & Miller, 2004), and with humans using a contingency learning task (Stout, Amundson, & Miller, 2005). In a recent example from our laboratory (De la Casa, unpublished results), we used an eyeblink conditioning procedure with human participants to ascertain the effect of interposing a retention interval (7 days) between conditioning and testing stages in a three-stage LI treatment. All participants were instructed to watch a silent movie, during which time the stimulus preexposed group (PE, n ¼ 20) was presented with an auditory stimulus (i.e., either a tone or a white noise, counterbalanced). The nonpreexposed group (NPE, n ¼ 20) was exposed to the same experimental situation, but
Inter-stage context and time as determinants of latent inhibition
43
100 Mean Percentage of differential CR
Conditioning
Test (extinction) NPE/No-delay NPE/7-days PE/No-delay PE/7-days
80
60
40
20 1 0
–20
1
2
3
4
6 7 5 4-trial Blocks
8
1
2
Figure 3.1. Mean percentage of differential CR (i.e., % CR to the CSþ % CR to the CS) in 4-trial blocks in each of four experimental conditions during Conditioning (left panel) and Test (right panel) stages. PE, preexposed; NPE, non-preexposed.
without receiving any nominal auditory stimuli. Subsequently, one of the stimuli (either tone or white noise, counterbalanced) was paired 32 times with the air-puff US (CSþ condition); the second stimulus was also presented 32 times, but without the US (CS). The test stage was conducted either 3 s (no-delay) or 7 days after conditioning. The test stage consisted of eight presentations each of CSþ and CS. The four experimental groups resulting from combining the Preexposure and Delay factors were: NPE/No-delay, PE/No-delay, NPE/7-days, PE/7-days. Figure 3.1 depicts the mean differential conditioned responses (the percentage of conditioned responses to CSþ minus the percentage of conditioned responses to CS) for conditioning and testing stages as a function of groups. As can be seen, the expected LI effect emerged across conditioning trials, yielding a lower rate of responding for the PE groups than for the NPE groups. As expected (see below), the Preexposure Delay interaction at testing stage was significant, F(1, 36) ¼ 4.17, P <0.05, due to differential CRs in the test stage being significantly lower for the PE/7-days group than for the PE/No-delay group, thus revealing that the retention interval induced an increase of the LI effect. An analysis of the experiments that have studied the effect of passage of time on LI reveals procedural differences related to intensity of the stimuli, temporal parameters, number of preexposures and conditioning trials, and length of the retention interval. However, one factor is consistently associated with the direction of the timeinduced LI effect, attenuated or potentiated: the context in which the retention interval is spent. Specifically, in cases where the attenuation of LI is observed the retention interval is spent in the same context as that of the other experimental stages
44
Luis Gonzalo De la Casa and Oskar Pinen˜o
(typically the home cage). In those experiments reporting enhanced LI, the delay occurred in a context that was different from the contexts of the preexposure, conditioning, and testing stages. Lubow and De la Casa (2005) proposed that the retention interval context was fundamental for determining the effect of the passage of time on LI (see section Explaining time–context interactions).
Theoretical implications Despite the broad variety of phenomena studied in the associative learning literature, it would be fair to say that a small collection of effects have been the primary target of most research, thereby critically guiding theoretical advance. A classic example is the case of the blocking effect (Kamin, 1969), which encouraged the development of the Rescorla and Wagner (1972) model. Similarly, it could be claimed that the need for a comprehensive explanation of LI encouraged the development of a whole family of learning theories, both general and specific to LI (e.g., Bouton, 1997; Lubow, Weiner, & Schnur, 1981; Mackintosh, 1975; McLaren & Mackintosh, 2000; Miller & Matzel, 1988; Pearce & Hall, 1980; Schmajuk, 2002; Wagner, 1981). In the present section we will discuss these models in regard to LI and, specifically, in relation to their account for the impact of contextual and/or temporal manipulations on LI (for detailed discussions of other aspects of some of these models, see Westbrook & Bouton and Escobar & Miller, in the present volume).
Acquisition-deficit models A common distinction of associative models has to do with their explanations of stimulus competition phenomena (e.g., overshadowing and blocking) and interference effects (e.g., extinction and LI) as either acquisition or performance deficits. Briefly, according to acquisition-deficit models, LI consists of the impaired acquisition of the target CS–US association due to the prior exposure of the CS in the absence of the US. By contrast, performance-deficit models propose that LI consists of the impaired conditioned response elicited by the CS due to CS-preexposure. In other words, according to performance-deficit models CS-alone exposures hinder conditioned responding to the CS, but do not interfere with the acquisition of the CS–US association. In the present section, we will discuss the acquisition-deficit account of LI, represented by attentional models of learning (e.g., Lubow et al., 1981; Mackintosh, 1975; Pearce & Hall, 1980) as well as by Wagner’s (1981) SOP model. (Performance-deficit accounts of LI will be discussed in the next section.) Attentional models Attentional models of learning have provided the most popular explanation of LI, perhaps because it best fits with the common view of stimulus competition and
Inter-stage context and time as determinants of latent inhibition
45
interference phenomena as reflecting the organism’s limited information processing resources. Simply put, since it would be practically impossible for an organism to attend to the immense array of stimuli in its environment, much of this information must be ignored. Although some stimuli will “demand” processing merely based on their inherent physical properties (e.g., high salience), other stimuli will “deserve” this processing based on their acquired informative value, a value that is determined by their reliability as signals for the occurrence of relevant events (i.e., stimuli of biological significance, such as food, water, or pain). Within this framework, models agree that, in an LI treatment, exposure to the CS in the absence of the US will result in a progressive loss of attention to the CS. In turn, because attention to the CS determines, at least in part, its ability to acquire associative strength, the effectiveness of subsequent CS–US pairings will be diminished relative to a condition with no CS-preexposure. As we will see, attentional models diverge, however, in the specific mechanisms leading the decline of attention, as well as in their views of the very nature of attention. Soon after the publication of the model by Rescorla and Wagner (1972), its inability to explain LI became apparent. According to their model, the repeated presentation of the CS in the absence of the US during preexposure has no effect upon the CS’s ability to subsequently acquire associative strength. Mackintosh (1975) was quick in realizing that the Rescorla–Wagner model could be revised in order to extend its explanatory power, thereby allowing it to account for several additional phenomena, including LI. Mackintosh added to the Rescorla–Wagner model a rule by which the associability of the CS (i.e., aCS, which corresponds to the CS salience in the Rescorla–Wagner model) could change its value according to experience with the CS. Specifically, according to Mackintosh, on each presentation of the CS the value of aCS is updated based on a comparison between the CS’s associative strength and the associative strength of all other CSs present on that trial. When the associative strength of the target CS is greater than the associative strength of the other CSs, aCS increases. Conversely, when the associative strength of the target CS is smaller than (or equal to) the associative strength of the other CSs, aCS declines. Because the CS’s associability, or aCS, corresponds to the attentional processing received by the CS, this rule in Mackintosh’s model could be restated as follows: attention to the target CS directly correlates with its relative ability to predict the US. As the CS becomes a better predictor of the US, it receives more attention; conversely, as the CS loses predictive value, it also loses attention. Interestingly, in Mackintosh’s (1975) model, attention to the CS (or associability, or aCS) is considered to be an attribute of the CS, whose value is initially (when the CS is novel) determined by its physical properties (e.g., salience) and then (when the CS is familiar) by its prior associative history. Therefore, a reciprocal relation between attention and learning was proposed: not only attention to the CS determined its subsequent ability to acquire associative strength, but also the CS’s associative strength determined its subsequent attentional processing. Mackintosh’s attentional
46
Luis Gonzalo De la Casa and Oskar Pinen˜o
model explained LI by assuming that the associability of the CS (aCS) declines during its preexposure, a decline that is due to the CS having an identical associative strength (i.e., zero strength) to that of the rest of the CSs or contextual cues1. As a consequence, during the CS–US pairings, the CS will acquire a weak (or no) association with the US due to the low value of aCS. In psychological terms, during CS-alone presentations the animal first learns that the CS is not a reliable predictor of the US and, thus, withdraws its attention from the CS. Because attention allocated to the CS during the subsequent CS–US pairings is reduced, the acquisition of the CS–US association is impaired. Another attentional model, this one proposed by Pearce and Hall (1980), shares Mackintosh’s (1975) view that the attentional processing received by the CS is represented by CS associability (aCS). In the Pearce–Hall model, the value of aCS also changes according to the CS’s associative history. However, this model diverges from that of Mackintosh in regard to the rules determining the change of aCS. Whereas, according to Mackintosh, aCS directly correlates with the associative strength of the CS (i.e., the CS receives more attention as it better predicts the US), in the Pearce–Hall model the value of aCS decreases as the CS becomes a more reliable predictor of the outcome, regardless of whether this outcome consists of the occurrence of the US or the absence of the US (i.e., “no US”; Konorski, 1967). Thus, in the Pearce–Hall model, both CS–US pairings and CS-alone presentations will have an identical detrimental impact on the value of aCS: as the animal learns to predict the outcome of the CS, less attentional processing will be devoted to the CS (i.e., the value of aCS will decrease). In this view, once the “meaning” of the CS has been already learned, the CS will not demand further attentional effort and, therefore, these resources can be devoted to those CSs the animal still needs to learn about. In this model, it is precisely when the outcome of the CS is not completely predictable that attention will be paid to the CS (i.e., aCS will maintain a high value). In the Pearce–Hall model, as in Mackintosh’s model, LI is due to the low value of aCS following CS preexposure, which results in the impaired acquisition of associative strength during the subsequent CS–US pairings. However, contrary to Mackintosh’s model, which claims that the low value of aCS following CS-alone trials is due to the CS being a poor predictor of the US, in the Pearce–Hall model the low value of aCS following CS-alone trials is due to the CS being a reliable predictor of the absence of the US (i.e., the CS fully predicts the occurrence of the “no US”). Despite their multiple merits, the attentional models of Mackintosh (1975) and Pearce–Hall (1980) are unable to explain the effects of contextual or temporal manipulations on LI. Because, according to these models, the value of aCS depends on the CS’s reliability as a predictor of the occurrence of the US (as well as of the 1
The decrease in the value of aCS when the CS is a worse predictor of the US than the other stimuli can be perfectly justified. However, the same cannot be said about the decrease in the value of aCS when the CS and the other stimuli are equally reliable predictors of the US. In fact, there is no reason why this latter scenario should not produce an increase, rather than a decrease in the value of aCS. Alternatively, the model could assume no variation in the value of aCS under these conditions.
Inter-stage context and time as determinants of latent inhibition
47
“no US”, in the Pearce–Hall model), experimental manipulations that cause no change in the associative strength of the CS are of no consequence on the CS’s associability (aCS) and, hence, on the acquisition of the CS–US association following CS-preexposure. However, these models could easily account for the impact of certain contextual/temporal manipulations on LI by merely assuming that the reduced CS associability could be directly restored after these manipulations: if, after a context change or a retention interval, aCS was reset to its original value, LI would be attenuated (i.e., because the CS can now readily enter into an association with the US during CS–US pairings). This would allow the models of Mackintosh and Pearce–Hall to account for attenuation of LI when the preexposure and conditioning phases (i.e., in a two-stage design for LI) are separated by a retention interval (e.g., Hall & Minor, 1984; Pinen˜o, De la Casa, Lubow, & Miller, 2006; Rosas & Bouton, 1997; Wagner, 1979; Westbrook, Bond, & Feyer, 1981), a context change (e.g., Hall & Channell, 1985, 1986; Lovibond et al., 1984; Westbrook et al., 2000), or the presentation of novel stimulation (e.g., Escobar, Arcediano, & Miller, 2005; Lantz, 1973; Pinen˜o et al., 2006; Rudy, Rosenberg, & Sandell, 1977). However, even if this assumption (i.e., aCS reset to its original value after a context switch or a time break) was adopted, these models would still be unable to explain the impact of such manipulations when they are performed between conditioning and test stages of a three-phase LI procedure. In this latter case, because the CS associability (aCS) exclusively determines the CS’s ability to enter into an association with the US (i.e., it has no influence on the expression of a previously acquired association), any putative variation in the value of aCS should have a negligible impact on responding to the CS at test. Thus, even if these models adopted an attentional restoration process (i.e., reset in the value of aCS), they would still be unable to explain the results from studies showing that, when a retention interval is interpolated between conditioning and testing, LI can be either attenuated (e.g., Aguado et al., 1994; Bakner, Strohen, Nordeen, & Riccio, 1991; De la Casa & Lubow, 1995; Kraemer & Ossenkopp, 1986; Kraemer, Randall, & Carbary, 1991; Kraemer & Roberts, 1984; Kraemer & Spear, 1992; Westbrook et al., 2000) or enhanced (i.e., the so-called super-LI effect; De la Casa & Lubow, 2000, 2002; Lubow & De la Casa, 2002; Wheeler et al., 2004; for evidence of enhanced LI with a contextual switch between conditioning and test, see Swartzentruber & Bouton, 1992; Westbrook et al., 2000). An alternative attentional model, Lubow et al.’s (1981) Conditioned Attention Theory (CAT), proposed a view of attention as a response, instead of as an attribute of the CS. According to CAT, any novel CS initially elicits an attentional response (e.g., an orienting response), which is necessary in order to learn other responses to the CS. This attentional response decreases as a function of the number of CS presentations, regardless of whether the CS is presented alone or with a US (although the decay in the strength of the attentional response presumably follows different patterns in each case, with this rate being faster during CS-preexposure than during conditioning). That is, CAT, as in the Pearce–Hall model, states that LI is not an
48
Luis Gonzalo De la Casa and Oskar Pinen˜o
exclusive product of CS preexposure, but of any treatment in which the CS becomes a perfect predictor of a given outcome, including the US (e.g., Hall & Pearce, 1979). Of critical importance for CAT, the attentional response behaves according to the laws of any other conditioned response and, thus, its decline during CS-preexposure can be viewed as a parallel of the decline typically observed during the extinction of Pavlovian conditioned responses. This view of CAT also implies that any experimental manipulation that would affect the acquisition and/or expression of a conditioned response (e.g., a contextual switch or a retention interval) should also affect the attentional response and, thus, have an impact on LI. More specifically, CAT predicts a recovery of the attentional response and, hence, attenuation of LI, when the preexposure and conditioning phases are conducted in different contexts (i.e., attentional recovery akin to a renewal of responding; Bouton & Bolles, 1979), or when they are separated by a retention interval (i.e., attentional recovery akin to spontaneous recovery of responding; Pavlov, 1927). In other words, CAT explains those results in the literature that were not accounted for by Mackintosh and Pearce– Hall, unless they adopted the attentional restoration construct. However, CAT cannot explain LI effects resulting from contextual and temporal manipulations that are conducted between conditioning and test stages. The reason is the same as that previously pointed out in regard to the Mackintosh and Pearce–Hall models. Specifically, because in CAT the attentional response elicited by the CS modulates the acquisition, but not the expression, of a CS–US association, any change in attention elicited by the CS at test after the CS–US pairings should not affect responding to the CS.
Memory priming Wagner’s (1981) SOP model also accounts for LI as an acquisition failure, in this case due to the CS being rehearsed or primed in short-term memory by the context during the CS–US pairings. In the framework of the SOP model, the elements of a stimulus representation in memory (i.e., a node) can be in one of three activation states: inactive (I); a primary activation state (i.e., A1), in which the stimulus is in the attentional focus; and a secondary activation state (i.e., A2), in which the stimulus is in the attentional periphery. Briefly, the learning rules of the SOP model state that an excitatory CS–US association is formed when the elements in the nodes of the CS and the US are simultaneously activated in the A1 state. By contrast, when the elements of the CS and US nodes are simultaneously activated into A1 and A2, respectively, an inhibitory CS–US association is formed. According to this model, no association is formed when the CS node is activated in the A2 state, regardless of the activation state of the US node (but see Dickinson and Burke’s [1996] revision of Wagner’s [1981] SOP model). In the SOP model, the elements of a stimulus node can reach the A2 state through either a decay from the A1 state (i.e., according to its rules of temporal dynamics) or the associative activation of the stimulus-node by another
Inter-stage context and time as determinants of latent inhibition
49
stimulus: when stimuli X and Y maintain an excitatory association, the presentation of stimulus X can activate (prime) the node of stimulus Y in short-term memory, specifically into the A2 state. Of importance, when a stimulus is associatively activated into A2, it will be poorly activated into A1, even if it is subsequently presented, because its nodal elements, already activated into A2, cannot be immediately recruited for the A1 activation state. Based on the previous summary of selected features of the SOP model, its explanation of LI becomes straightforward. First, during CS-preexposure, the CS forms a bidirectional association with the context. Based on this association, the context is able to strongly activate (prime) the CS-node into A2, thereby preventing the activation of most of the nodal elements of the CS into A1. As a consequence, during the subsequent conditioning phase, a weak CS–US association is formed. From this explanation, it directly follows that LI should be attenuated when preexposure and conditioning treatments are conducted in different contexts: because the conditioning context maintains no association with the CS, the CS is not activated into the A2 state during the initial CS–US pairings and, thus, it uneventfully enters into an association with the US. However, the SOP model faces difficulties in explaining attenuation of LI when a retention interval is interposed between preexposure and conditioning. SOP can only explain such data when the retention interval is spent in the context in which CS-preexposure took place, thereby allowing the context–CS association to undergo extinction (i.e., because the animal is exposed to the context in the absence of the CS), thereby subsequently reducing the A2 activation of the CS by the context during the conditioning stage. Notably, although some studies have reported attenuation of LI when the retention interval was spent in the context of preexposure (e.g., Escobar, Arcediano, & Miller, 2003; Hall & Minor, 1984; Wagner, 1979; Westbrook et al., 1981), attenuation of LI has also been reported when the retention interval was spent in a different context (e.g., Rosas & Bouton, 1997). Finally, the SOP model predicts no effect of a context change or a retention interval when such manipulations are performed between conditioning and testing in a threestage design of LI. The model predicts that these manipulations would allow the CS to be more strongly activated into A1 at test (i.e., because of the weaker A2 activation caused by the test context), but this treatment should have no detectable impact on responding elicited by the CS because of the previous failure to acquire the CS–US association.
Performance-deficit models In the framework of performance-deficit models, interference effects (e.g., extinction and LI) and cue competition phenomena (e.g., blocking and overshadowing) are not due to a failure in the acquisition of the target CS–US association, but to a failure in the “translation” of an acquired CS–US association into a conditioning response.
50
Luis Gonzalo De la Casa and Oskar Pinen˜o
Specifically, interference and cue competition effects can be viewed as a failure to either retrieve (Bouton, 1993, 1997) or express (Miller & Matzel, 1988) the target CS–US association. In the present section we will discuss these two models in relation to their explanation of LI, with emphasis on their accounts of the effects of contextual changes and passage of time.
Retrieval-deficit: Bouton’s (1993, 1997) retrieval failure model From early on, the Rescorla–Wagner (1972) model has faced problems with its explanation of interference between outcomes. Not only was the model unable, as previously mentioned, to offer an account for LI, but it also provided an explanation of extinction that had already proven problematic. Although the Rescorla–Wagner model could simulate neat acquisition and extinction curves, its explanation of extinction relied on the assumption that the CS–US association, previously acquired during the conditioning phase, was erased during the subsequent CS-alone trials (i.e., catastrophic interference; McCloskey & Cohen, 1989). However, since Pavlov’s (1927) early studies, it was well known that extinction did not consist of the removal of the CS–US memory: the weak conditioned response elicited by the CS following extinction treatment recovered after a lapse of time (i.e., spontaneous recovery). This problem for the Rescorla–Wagner model soon became more evident when Bouton and his collaborators (e.g., Bouton & Bolles, 1979; Bouton & King, 1983; Bouton & Ricker, 1994) found that conditioned responding could also recover by a context change following extinction – an effect that they named renewal of responding. Both spontaneous recovery and renewal effects strongly indicated that extinction could not be accounted for in terms of erasure of the original CS–US memory. Rather, it seemed that extinction consisted of the acquisition of a second memory, which somehow interfered with the memory of the excitatory CS–US association. This idea was elegantly developed by Bouton (1993), who proposed that the extinction memory (i.e., an inhibitory CS–US association) interfered with the retrieval of the conditioning memory (i.e., the excitatory CS–US association). That is, the excitatory CS–US memory remained intact during the subsequent CS-alone trials, but became poorly retrievable due to the interference caused by the more recently acquired and, hence, more strongly retrieved inhibitory CS–US memory. In addition, this model posited an important asymmetry between the excitatory and inhibitory CS–US associations: the inhibitory association, but not the excitatory association, was encoded as context-specific and, thus, depended upon its training context for retrieval. Thus, when the CS was presented outside the extinction context, the inhibitory association could no longer be retrieved and the excitatory association was thereby released from the interference otherwise caused by the inhibitory association. This allowed the excitatory association to be again retrieved from memory and, thus, expressed in conditioned responding. Spontaneous recovery was analogously explained by Bouton’s retrieval-failure model. Because the model views passage of
Inter-stage context and time as determinants of latent inhibition
51
time as an instance of a contextual change (i.e., in this case, it is the temporal context, rather than the physical context, that is changed), spontaneous recovery and renewal were explained by the same principle, as a failure of the memory of extinction to be retrieved at the time of test, indirectly resulting in the effective retrieval of the memory of conditioning. The explanation of LI according to Bouton’s (1993) model was very similar to that of extinction. In the case of LI, however, interference was presumed to occur proactively (i.e., the first-learned memory interferes with the second-learned memory), rather than retroactively (i.e., the second-learned memory interferes with the first-learned memory). Explaining LI required just a small adjustment to the theory. Because during CS-preexposure the animal has not yet experienced the US, the inhibitory CS–US association cannot be acquired. Consequently, Bouton proposed, in line with Konorski (1967; also see Pearce & Hall, 1980), that during CS-preexposure the animal learns a CS–no-US association, better viewed as a CS–nothing association (i.e., the animal learns that the CS is followed by no event). This CS–nothing association then interferes with the retrieval of the CS–US association during the conditioning stage, resulting in the LI effect. Bouton’s (1993) retrieval model provides a straightforward account for attenuation of LI by the interpolation of either a contextual switch or a retention interval between preexposure and conditioning stages. Because the CS–nothing association will be poorly retrieved by the new context in which CS–US pairings are given, the excitatory CS–US association will be strongly retrieved and, thus, the CS will be able to elicit strong responding (i.e., attenuated LI). Interestingly, the same prediction, attenuation of LI, is made by this model when these contextual/temporal manipulations are performed after conditioning and prior to testing (e.g., Aguado et al., 1994; Westbrook et al., 2000). Assuming that the retrieval of the excitatory CS–US association during the conditioning stage was still incomplete, changing the context (physical or temporal) between conditioning and testing could further impair retrieval of the CS–nothing association, thereby allowing for a better recovery of the excitatory CS–US association at test (i.e., attenuated LI). This prediction provides Bouton’s (1993) model with an advantage over the previously described acquisition-focused models, which predict that LI should not be affected by manipulations conducted after the conditioning stage. In spite of its merits, Bouton’s (1993) model fails to explain the enhancement of LI that results from interpolating a retention interval in a different context (i.e., super-LI; e.g., De la Casa & Lubow, 2000), or from switching the context (e.g., Westbrook et al., 2000) between conditioning and test. A revision of this model (Bouton, 1997), however, does account for enhancement of LI after these contextual/ temporal manipulations. According to Bouton (1997), it is not the nature of the associations (i.e., excitatory vs. inhibitory), but the order in which they are learned, that determines which association becomes context-dependent for retrieval. Specifically, it is not the inhibitory CS–US association (or the CS–no-US association), but
52
Luis Gonzalo De la Casa and Oskar Pinen˜o
the association learned in second place, that depends on the context for its retrieval. The second-learned association becomes context-dependent because it is when this association is learned when the CS acquires a second “meaning” and, thus, becomes ambiguous. When the CS is followed by a second outcome, the animal presumably pays attention to contextual cues in an attempt to disambiguate its new meaning. Although this newer version of Bouton’s retrieval failure model does not make any differential prediction regarding extinction (i.e., because in this case the inhibitory association is also the second-learned association and, thus, it is expected to depend on the context for retrieval), its predictions regarding LI are diametrically different from those of his 1993 model: in an LI experiment, the excitatory CS–US association is second-learned and, hence, it is the association that is context-dependent. Therefore, the CS–no-US (or CS–nothing) association should be relatively contextindependent and, thus, highly retrievable in a new context. In other words, stronger LI should be observed when the CS is tested in a context (physical or temporal) different from the conditioning context. Contrary to his original model, Bouton’s (1997) revised model can readily explain enhancement of LI when the conditioning and test phase are separated by a retention interval (e.g., De la Casa & Lubow, 2000), as well as when these phases take place in different contexts (e.g., Westbrook et al., 2000). Interestingly, as opposed to the earlier version, the newer model (Bouton, 1997) cannot explain attenuation of LI when these contextual/temporal manipulations are performed either between the preexposure and conditioning stages, or between the conditioning and test stages. In sum, although Bouton’s retrieval failure model provides a relatively comprehensive explanation of a variety of LI effects, it is still unclear how it can be revised in order to embrace the aforementioned contrary effects of context/time upon LI.
Expression-deficit: Miller and Matzel’s (1988) comparator hypothesis Another performance-deficit account of LI is provided by Miller and Matzel’s (1988) comparator hypothesis (also see Denniston, Savastano, & Miller, 2001). According to this model, cue competition phenomena (e.g., blocking and overshadowing) are due to the poor expression of the target CS–US association. In this model, cues do not compete for the acquisition of associative strength, nor do their associations with the US compete for retrieval (but see Denniston, Savastano, Blaisdell, & Miller, 2003). Instead, cues compete for the behavioral expression of their associative strength. Briefly, the model posits that on pairings of the target CS (X) with the US, CS X acquires an association with the US, as well as with other stimuli (CSs or contextual cues, Y) that are presented in compound with CS X (i.e., the so-called X’s comparator stimuli). When CS X is later presented at test, it consequently retrieves the memory of the US both directly (i.e., based on the X–US association) and indirectly (i.e., based on the conjoint action of the X–Y and Y–US associations). The response elicited by CS X at test is assumed to be the result of a comparison
Inter-stage context and time as determinants of latent inhibition
53
between the direct and the indirect activation of the US-representation. CS X produces a detectable conditioned response only when the direct US-representation is stronger than the indirect US-representation, with the strength of this response being a function of the balance between the direct and the indirect activation of the US-representation. The comparator hypothesis explains LI based on the same processes posited in its explanation of other cue competition phenomena. First, during CS-preexposure, the CS acquires an association with the context. During the subsequent CS–US pairings, both the CS and the context enter into an association with the US. When the CS is then presented at test, its direct activation of the US-representation (i.e., based on the CS–US association) will not be fully expressed into responding due to the strong retrieval of the memory of the context (i.e., based on the strong CS–context association acquired during preexposure), which, in turn, activates the US (i.e., based on the context–US representation), thereby resulting in a strong indirect activation of the US-representation. Therefore, the comparator hypothesis explains LI as an instance of overshadowing, only with the context, instead of a punctate cue or CS, overshadowing the target CS. Notably, this account of LI yields many unique predictions, many of which have received empirical support (for a review, see Escobar & Miller, this volume). The comparator hypothesis, however, has difficulties in explaining the impact of contextual/temporal manipulations on LI. According to the comparator hypothesis, contextual manipulations can only affect responding to the target CS in an LI treatment to the extent that these manipulations affect the strength of the CS–context association, the context–US association, or both (because only these manipulations can modify the indirect activation of the US-representation and, hence, responding to the CS). Therefore, for example, the comparator hypothesis predicts that conducting preexposure and conditioning phases in different contexts will result in attenuated LI because the preexposure context, in this case, a strong comparator stimulus for the CS, has no association with the US, whereas the conditioning context, which has a strong association with the US, is a poor comparator for the CS. (In an LI experiment, the preexposure context is usually a stronger comparator stimulus for the target CS than the conditioning context because of the large number of CS-alone trials, relative to the number of CS–US pairings.) The above explanation does not apply to those cases in which a change of context follows the conditioning phase, prior to testing. In this case, the context switch should not affect responding to the target CS, basically because the status of the CS–context and context–US associations remains intact after such a manipulation. Moreover, according to the comparator hypothesis, the interpolation of a retention interval between CS-preexposure and conditioning, or between conditioning and test, should attenuate LI only to the extent that this retention interval is spent in the training context: in this case, exposure to the context will cause the extinction of either the CS–context association or the context–US association, or both, thereby reducing the indirect activation of
54
Luis Gonzalo De la Casa and Oskar Pinen˜o
the US-representation and, hence, allowing the CS–US association to be strongly expressed in responding. In sum, the comparator hypothesis can only explain attenuation and enhancement of LI when contextual/temporal manipulations are performed that result in the weakening or strengthening, respectively, of the CS–context and/or context–US associations. When contextual/temporal manipulations leave these associations intact, responding elicited by the target CS should not be affected.
Explaining time–context interactions As reviewed in the Introduction, passage of time and contextual manipulations clearly modulate the size of the LI effect at test. Specifically, LI is attenuated in those situations in which a long retention interval is spent in the same context as the other experimental stages (e.g., Aguado et al., 1994). This result is usually obtained in experiments using the conditioned taste aversion procedure, in which the experimental manipulations are typically conducted in the animals’ home cages (i.e., the context remains constant throughout the experiment). In contrast, the super-LI effect appears when the long delay context is different from the context in which the rest of the experimental stages are conducted (e.g., De la Casa & Lubow, 2000). This situation occurs when the preexposure, conditioning and test stages take place in a location that is not the animals’ home cage. None of the theories of LI proposed to date (see section on Theoretical implications) seems capable of integrating the results stemming from the interaction between the delay and the context in which the delay takes place. This led Lubow and De la Casa (2005) to propose the context differentiation hypothesis. The rationale for the hypothesis begins by accepting the commonly held assumption that, during the different stages in an LI treatment, two associations are formed, an association between the preexposed stimulus and the absence of consequences (during preexposure), and an association between the preexposed stimulus and the US (during conditioning). Although both associations are potentially recoverable, the essential factors that determine the context’s ability to retrieve one association or another are the time elapsing following the establishment of the CS–nothing and CS–US associations and the context in which the organism spends this time. When the period following conditioning is short (typically 24/48 hours in conditioned taste aversion experiments), the context does not play an important role. The LI effect is observed, regardless of the context in which the delay has elapsed (see, for example, De la Casa & Lubow, 2000). However, when a longer time interval is interpolated following conditioning, the delay context takes on a pivotal role in the expression of the conditioned response. More specifically, when the delay elapses in the same context as the other experimental stages (preexposure and conditioning), the animal is exposed for a prolonged period to a context in which a series of biologically
Inter-stage context and time as determinants of latent inhibition
55
significant events have occurred, but now in the absence of such events. As a result, we might expect the hypothetical links established between the context and the different associations established in its presence (CS–nothing and CS–US) to be extinguished, with further extinction taking place the longer the period of context exposure. Moreover, the extinction of the links between the context and the CS–nothing and CS–US associations presumably has a differential impact on the subsequent retrieval and/or expression of these two associations. In general, CS–US associations are relatively context-independent compared to the CS–nothing association (Bouton & King, 1983; Bouton & Peck, 1989; Lubow & De la Casa, 2005). Thus, with passage of time, the link between the context and the CS–nothing association would be extinguished, while the link with the CS–US association would remain relatively intact. As a consequence, presentation of the CS in the test stage would result in strong conditioned responding (i.e., attenuation of LI). In contrast, when the time interval is spent in a context different from that of the other experimental stages, all of the previously established associations would remain intact. To account for the super-LI effect that is obtained under these conditions, one might appeal to a time-induced enhancement of the primacy effect, the evidence that first learning is disproportionately strong as compared with subsequent training (Bouton, 1993; Konorski & Szwejkowska, 1952). A number of human memory experiments have shown that primacy effects increase with the length of the retention interval between study and test stages (e.g. Knoedler, Hellwig, & Neath, 1999; Neath, 1993). Thus, for instance, Neath (1993) found that increasing the delay between the study of a list of non-verbal items (i.e., pictures of snowflakes) and a recognition stage resulted in better performance for the first-learned item than for the last-learned item. This interpretation of the role played by the long retention interval context matches the available empirical evidence on the subject, but it does not offer an explanatory mechanism for the origin of the bidirectional modulation of LI (i.e., attenuation of LI vs. super-LI). To address this limitation, Lubow and De la Casa (2005) have proposed the context differentiation hypothesis, which restates the processes outlined in the previous paragraph in terms of perceptual mechanisms. According to the context differentiation hypothesis, the perceived similarity of two contexts does not exclusively depend on their objective physical similarity. Additional factors that determine the perceived similarity between contexts include time between presentations of the two contexts and the degree of objective similarity/ difference between the interval context and the other context of the other experimental stages. The McLaren, Kaye, and Mackintosh (1989; see also McLaren & Mackintosh, 2000) analysis of perceptual learning offers an explanatory framework that is congruent with the context differentiation hypothesis. From their perspective, the perceptual differentiation between two contexts will depend on the proportion of common to unique elements. Discrimination between contexts would be facilitated by exposure to both of them since the salience of any common stimulus elements
56
Luis Gonzalo De la Casa and Oskar Pinen˜o
would be disproportionately reduced compared to unique elements (i.e., there would be a greater LI effect for the more frequently exposed common stimuli than for the unique stimuli). Thus, when preexposure, conditioning, retention interval, and testing occur in the same experimental context (AAAA, respectively), the common contextual elements will receive extensive preexposure which would reduce their associability/salience. However, this process would potentiate the salience of the differential elements (including those that tend to change over time such as physiological state, uncontrolled environmental changes, etc.), and lead to an enhancement of the perceived difference between the preexposure, conditioning, and test contexts, the context-change condition that reliably disrupts LI (e.g. Channell & Hall, 1981; Hall & Channell, 1986; Hall & Minor, 1984; Lovibond et al., 1984). This process of perceptive differentiation, probably together with the hypothetical extinction of the link between the context and the CS–nothing association previously discussed, would result in stronger conditioned responding being observed at test. In contrast, when the delay context is different from the other experimental contexts (AABA), the reduction in saliency of the common elements would be minimal. This, in turn, would promote the perceptual similarity of the experimental contexts, and consequently, at test, favor the retrieval of the CS–nothing association over the recovery of the CS–US association (i.e., due to the former association being more dependent on the context than the latter association), thereby producing enhanced LI.
Applied implications To date, practical, clinically related, applications of the LI effect have been primarily prophylactic, namely using inconsequential stimulus preexposures in order to prevent the acquisition of, for example, food aversions due to radiation/chemotherapyinduced nausea (e.g., Bovbjerg, 2006; Carey & Burish, 1988; Schwartz, Jacobsen, & Bovbjerg, 2006) and phobias (e.g., Davey, 1989; Eiche & Cook, 2001; for a review, see Lubow, 1998). The post-event time/context manipulations, which can either decrease or increase LI, also have implications increasing the efficacy of preventative measures as well as for treatment. For example, the interaction between time and context discussed in the previous section, applied to the chemotherapy treatment, suggests that the effectiveness of food preexposure in reducing the development of food aversions would be enhanced if the chemotherapy sessions were followed by a long a period in a different context from that in which the target food was eaten. As for reducing the likelihood of developing a phobia, the typical procedure already favors the appearance of the super-LI effect. For example, many dentists require children to make several visits to the dental clinic, before any treatment, in order to familiarize them with the setting and the material used in dental operations. Eiche and Cook (2001) suggested that between two and five visits of this type should be made, each lasting about 20 minutes. To maximize prophylactic effectiveness, the first real intervention should be followed
Inter-stage context and time as determinants of latent inhibition
57
by a relatively long period spent (as would usually be the case anyway) in a context different from that of preexposure and conditioning. In this way, the process of perceptive equalization (see section on Explaining time–context interactions) would favor the recovery of the association established during preexposure without consequences, thereby reducing anxiety responses that are characteristic of phobia.
Conclusions The present chapter has discussed evidence and theoretical explanations of attenuation and enhancement of LI due to inter-stage manipulations of context and time. These two key terms provide an appropriate metaphor for the relevance of LI in current research: in today’s new theoretical context, and after a 50-year retention interval since the original report of the LI effect by Lubow and Moore (1959), the impact of LI research in the literature has not attenuated, it has been greatly enhanced. If we can conclude anything from the empirical evidence discussed in this chapter, it is that the phenomenon of LI is determined by a series of variables that do not act in a linear manner. Instead, these variables interact intricately, giving rise to an enormously flexible and adaptive process. In this chapter, we have focused on the impact of two of these variables, namely, context and time, as well as on their interaction. The results from the studies that have examined this interaction do not fit easily into past or current theories of LI. As a consequence, we have described an alternative account, namely the context differentiation hypothesis (Lubow & De la Casa, 2005), which, although reasonable, is largely post-hoc and speculative. The practical difficulties involved in analyzing this interaction in certain experimental situations, particularly in the case of research with humans, present a formidable challenge, but the prospective theoretical and applied results indicate that it would be well worth the effort.
Acknowledgements This research was supported by Spanish MEC SEJ2006–1489 and Junta de Andalucı´ a SEJ-02618 grants to L.G. De la Casa.
References Aguado, L., Symonds, M., & Hall, G. (1994). Interval between preexposure and test determines the magnitude of latent inhibition: implications for an interference account. Animal Learning & Behavior, 22, 188–194. Alvarez, R., & Lopez, M. (1995). Effects of elements or compound preexposure on conditioned taste aversion as a function of retention interval. Animal Learning & Behavior, 23, 391–399.
58
Luis Gonzalo De la Casa and Oskar Pinen˜o
Bakner, L., Strohen, K., Nordeen, M., & Riccio, D. C. (1991). Post-conditioning recovery from the latent inhibition effect in conditioned taste aversion. Physiology and Behavior, 50, 1269–1272. Balsam, P. D., & Tomie, A. (1985). Context and Learning. Hillsdale, NJ: Lawrence Erlbaum. Bouton, M. E. (1993). Context, time, and memory retrieval in the interference paradigms of Pavlovian learning. Psychological Bulletin, 114, 80–99. Bouton, M. E. (1997). Signals for whether versus when an event will occur. In M. E. Bouton & M. S. Fanselow (Eds.), Learning, Motivation, and Cognition. The Functional Behaviorism of Robert C. Bolles. Washington DC: American Psychological Association. Bouton, M. E., & Bolles, R. C. (1979). Contextual control of the extinction of conditioned fear. Learning and Motivation, 10, 445–466. Bouton, M. E., & King, D. A. (1983). Contextual control of the extinction of conditioned fear: tests for the associative value of the context. Journal of Experimental Psychology: Animal Behavior Processes, 9, 248–265. Bouton, M. E., & Peck, C. A. (1989). Context effects on conditioning, extinction, and reinstatement in an appetitive conditioning preparation. Animal Learning & Behavior, 17, 188–198. Bouton, M. E., & Ricker, S. T. (1994). Renewal of extinguished responding in a second context. Animal Learning & Behavior, 22, 317–324. Bovbjerg, D. H. (2006). The continuing problem of post chemotherapy nausea and vomiting: contributions of classical conditioning. Autonomic Neuroscience – Basic & Clinical, 129, 92–98. Carey, M. P., & Burish, T. G. (1988). Etiology and treatment of the psychological side effects associated with cancer chemotherapy: a critical review and discussion. Psychological Bulletin, 106, 307–325. Channell, S., & Hall, G. (1981). Facilitation and retardation of discrimination learning after exposure to the stimuli. Journal of Experimental Psychology: Animal Behavior Processes, 7, 437–446. Davey, G. C. L. (1989). Dental phobias and anxieties: evidence for conditioning processes in the acquisition and modification of a learned fear. Behavior Research and Therapy, 27, 51–58. De la Casa, L. G., & Lubow, R. E. (1995). Latent inhibition in conditioned taste aversion: the roles of stimulus frequency and duration, and amount of fluid ingested during preexposure. Neurobiology of Learning & Memory, 64, 125–132. De la Casa, L. G., & Lubow, R. E. (2000). Super-latent inhibition with delayed conditioned taste aversion testing. Animal Learning & Behavior, 28, 389–399. De la Casa, L. G., & Lubow, R. E. (2002). An empirical analysis of the super-latent inhibition effect. Animal Learning & Behavior, 30, 112–120. De la Casa L. G., & Lubow, R. E. (2005). Delay-induced super-latent inhibition as a function of order of exposure to two flavours prior to compound conditioning. Quarterly Journal of Experimental Psychology: Comparative and Physiological Psychology, 58B, 1–18. Denniston, J. C., Savastano, H. I., Blaisdell, A. P., & Miller, R. R. (2003). Cue competition as a retrieval deficit. Learning and Motivation, 34, 1–31. Denniston, J. C., Savastano, H. I., & Miller, R. R. (2001). The extended comparator hypothesis: learning by contiguity, responding by relative strength. In
Inter-stage context and time as determinants of latent inhibition
59
R. R. Mowrer & S. B. Klein (Eds.), Handbook of Contemporary Learning Theories. Hillsdale, NJ: Lawrence Erlbaum, pp. 65–117. Dickinson, A., & Burke, J. (1996). Within-compound associations mediate the retrospective revaluation of causality judgements. Quarterly Journal of Experimental Psychology, 49B, 60–80. Eiche, K. D., & Cook, J. H. (2001). A learning-based approach to preventing dental phobias. Clinical Update, 23. Bethesda, MD: Naval Postgraduate Dental School, pp. 20–21. Escobar, M., Arcediano, F., & Miller, R. R. (2003). Latent inhibition in human adults without masking. Journal of Experimental Psychology: Learning, Memory, and Cognition, 29, 1028–1040. Escobar, M., Arcediano, F., & Miller, R. R. (2005). Disruption of latent inhibition by interpolation of irrelevant stimulation between preexposure and conditioning. Learning & Behavior, 33, 371–385. Eysenck, H. J. (1968). A theory of the incubation of anxiety/fear responses. Behaviour Research and Therapy, 6, 309–322. Hall, G., & Channell, S. (1985). Differential effects of contextual change on latent inhibition and on the habituation of an orienting response. Journal of Experimental Psychology: Animal Behavior Processes, 11, 470–481. Hall, G., & Channell, S. (1986). Context specifity of latent inhibition in taste aversion learning. Quarterly Journal of Experimental Psychology, 38B, 121–139. Hall, G., & Minor, H. (1984). A search for context-stimulus associations in latent inhibition. Quarterly Journal of Experimental Psychology, 36B, 145–169. Hall, G., & Pearce, J. M. (1979). Latent inhibition of a CS during CS-US pairings. Journal of Experimental Psychology: Animal Behavior Processes, 5, 31–42. Kamin, L. J. (1969). Predictability, surprise, attention, and conditioning. In B. A. Campbell & M. R. Church (Eds.), Punishment and Aversive Behavior. New York: Appleton-Century-Crofts, pp. 279–296. Killcross, A. S., Kiernan, M. J., Dwyer, D., & Westbrook, R. F. (1998). Loss of latent inhibition of contextual conditioning following non-reinforced context exposure in rats. Quarterly Journal of Experimental Psychology, 51B, 75–90. Knoedler, A. J., Hellwig, K. A., & Neath, I. (1999). The shift from recency to primacy with increasing delay. Journal of Experimental Psychology: Learning, Memory, and Cognition, 25, 474–487. Konorski, J. (1967). Integrative Activity of the Brain. Chicago: University of Chicago Press. Konorski, J., & Szwejkowska, G. (1952). Chronic extinction and restoration of conditioned reflexes: IV. The dependence of the course of extinction and restoration of conditioned reflexes on the “history” of the conditioned stimulus (the principle of the primacy of the training). Acta Biologiae Experimentalis, 16, 95–113. Kraemer, P. J., & Ossenkopp, K. P. (1986). The effects of flavor preexposure and test interval on conditioned taste aversions in rats. Bulletin of the Psychonomic Society, 24, 219–221. Kraemer, P. J., Randall, C. K., & Carbary, T. J. (1991). Release from latent inhibition with delayed testing. Animal Learning & Behavior, 19, 139–145. Kraemer, P. J., & Roberts, W. A. (1984). The influence of flavor preexposure and test interval on conditioned taste aversion in the rat. Learning and Motivation, 15, 259–278.
60
Luis Gonzalo De la Casa and Oskar Pinen˜o
Kraemer, P. J., & Spear, N. E. (1992). The effect of nonreinforced stimulus exposure on the strength of a conditioned taste aversion as a function of retention interval: do latent inhibition and extinction involve a shared process? Animal Learning & Behavior, 20, 1–7. Lantz, A. E. (1973). Effect of number of trials, interstimulus interval, and dishabituation during CS habituation on subsequent conditioning in a CER paradigm. Animal Learning & Behavior, 1, 273–277. Lovibond, P., Preston, G. C., & Mackintosh, N. J. (1984). Contextual control of conditioning and latent inhibition. Journal of Experimental Psychology: Animal Behavior Processes, 10, 360–375. Lubow, R. E. (1989). Latent Inhibition and Conditioned Attention Theory. Cambridge: Cambridge University Press. Lubow, R. E. (1998). Latent inhibition and behavioral pathology: some effects of stimulus preexposure. In W. O’Donohue (Ed.), Learning and Behavior Therapy. Boston: Allyn and Bacon, pp. 107–121. Lubow, R. E., & De la Casa, L. G. (2002). Super-latent inhibition and spontaneous recovery: differential effects of pre- and post-conditioning CS-alone presentations after long delays in different contexts. Animal Learning & Behavior, 30, 376–386. Lubow, R. E., & De la Casa, L. G. (2005). Time-induced super-latent inhibition is dependent on the distinctiveness of the retention-interval context from the other experimental contexts. Learning and Motivation, 36, 322–330. Lubow, R. E., & Moore, A. U. (1959). Latent inhibition: the effect of nonreinforced preexposure to the conditioned stimulus. Journal of Comparative and Physiological Psychology, 52, 415–419. Lubow, R. E., Weiner, I., & Schnur, P. (1981). Conditioned attention theory. In G. H. Bower (Ed.), The Psychology of Learning and Motivation, vol. 13. San Diego, CA: Academic Press, pp. 1–55. Mackintosh, N. J. (1975). A theory of attention: variations in the associability of stimuli with reinforcement. Psychological Review, 82, 276–298. McCloskey, M., & Cohen, N. (1989). Catastrophic interference in connectionist networks: the sequential learning problem. In G. H. Bower (Ed.), The Psychology of Learning and Motivation. San Diego, CA: Academic Press, pp. 109–165. McLaren, I. P. L., Kaye, H., & Mackintosh, N. J. (1989). An associative theory of the representation of stimuli: applications to perceptual learning and latent inhibition. In R. G. M. Morris (Ed.), Parallel Distributed Processing: Implications for Psychology and Neurobiology. Oxford: Oxford University Press, pp. 102–130. McLaren, I. P. L., & Mackintosh, N. J. (2000). An elemental model of associative learning: I. Latent inhibition and perceptual learning. Animal Learning & Behavior, 28, 211–246. Miller, R. R., & Matzel, L. D. (1988). The comparator hypothesis: a response rule for the expression of associations. In G. H. Bower (Ed.), The Psychology of Learning and Motivation, vol. 22. San Diego, CA: Academic Press, pp. 51–92. Neath, I. (1993). Distinctiveness and serial position effects in recognition. Memory & Cognition, 21, 689–698. Pavlov, I. P. (1927). Conditioned Reflexes. London: Clarendon Press. Pearce, J. M., & Hall, G. (1980). A model for Pavlovian learning: variations in the effectiveness of conditioned but not of unconditioned stimuli. Psychological Review, 87, 532–552.
Inter-stage context and time as determinants of latent inhibition
61
Pinen˜o, O., De la Casa, L. G., Lubow, R. E., & Miller, R. R. (2006). Some determinants of latent inhibition in humans. Learning and Motivation, 37, 42–65. Rescorla, R. A., & Wagner, A. R. (1972). A theory of Pavlovian conditioning: variations in the effectiveness of reinforcement and nonreinforcement. In A. H. Black & W. F. Prokasy (Eds.), Classical Conditioning II: Current Research and Theory. New York: Appleton-Century-Crofts, pp. 64–99. Rosas, J. M., & Bouton, M. E. (1997). Additivity of the effects of retention interval and context change on latent inhibition: toward resolution of the context forgetting paradox. Journal of Experimental Psychology: Animal Behavior Processes, 23, 283–294. Rudy, J. W., Rosenberg, L., & Sandell, J. H. (1977). Disruption of a taste familiarity effect by novel exteroceptive stimulation. Journal of Experimental Psychology: Animal Behavior Processes, 3, 26–36. Schmajuk, N. A. (2002). Latent Inhibition and its Neural Substrates. Norwell, MA: Kluwer Academic. Schwartz, M. D., Jacobsen, P. B., & Bovbjerg, D. H. (1996). Role of nausea in the development of aversions to a beverage paired with chemotherapy treatment in cancer patients. Physiology & Behavior, 59, 659–663. Stout, S., Amundson, J. C., & Miller, R. R. (2005). Trial order and retention interval in human contingency judgment. Memory & Cognition, 33, 1368–1376. Swartzentruber, D. E., & Bouton, M. E. (1992). Context sensitivity of conditioned suppression following preexposure to the conditioned stimulus. Animal Learning & Behavior, 20, 97–103. Wagner, A. R. (1979). Habituation and memory. In A. Dickinson & R. A. Boakes (Eds.), Mechanisms of Learning and Motivation. Hillsdale, NJ: Erlbaum, pp. 53–82. Wagner, A. R. (1981). SOP: a model of automatic memory processing in animal behavior. In N. E. Spear & R. R. Miller (Eds.), Information Processing in Animals: Memory Mechanisms. Hillsdale, NJ: Erlbaum, pp. 5–47. Westbrook, R. F., Bond, N. W., & Feyer, A. M. (1981). Short-term and long-term decrements in toxicosis-induced odor-aversion learning: the role of duration of exposure to an odor. Journal of Experimental Psychology: Animal Behavior Processes, 7, 362–381. Westbrook, R. F., Jones, M. L., Bailey, G. K., & Harris, J. A. (2000). Contextual control over conditioned responding in a latent inhibition paradigm. Journal of Experimental Psychology: Animal Behavior Processes, 26, 157–173. Wheeler, D. S., Stout, S. C., & Miller, R. R. (2004). Interaction of retention interval with CS-preexposure and extinction treatments: symmetry with respect to primacy. Learning & Behavior, 32, 335–347.
4 Latent inhibition: acquisition or performance deficit? Martha Escobar and Ralph R. Miller
What one knows and what one shows: acquisition vs. performance Since the time of Aristotle, it has been commonly assumed that two events occurring in close proximity become associated to each other. This learning by contiguity is a central determinant of the phenomenon of classical conditioning, in which a neutral target stimulus (X) is repeatedly presented in close proximity to a stimulus that unconditionally produces a response (i.e., an unconditioned stimulus, US), with the consequence that X comes to elicit a response appropriate to the US. That is, X becomes a conditioned stimulus (CS) that produces a conditioned response (CR). However, there have been reports of various conditions under which CS–US pairings fail to result in an acquired response to the CS. For example, even though pairings of X and the US ordinarily result in a robust CR (i.e., acquisition), X will fail to gain response potential if it is trained in the presence of another, more salient CS, A (i.e., overshadowing; Pavlov, 1927). Thus, even though the contiguity between CS X and the US is the same in the X–US acquisition and AX–US overshadowing conditions, the resulting behavioral control by CS X is not. An easy explanation for this difference is that subjects did not acquire an X–US association in the overshadowing condition. However, if the more salient companion CS, A, is made into a weak predictor of the US through extinction (i.e., A-alone presentations after the acquisition phase), behavioral control by CS X is restored (the so-called recovery from overshadowing effect; e.g., Kaufman & Bolles, 1981; Matzel, Schachtman, & Miller, 1985). This observation suggests that subjects in both the acquisition and overshadowing groups have come to associate CS X with the US, but their behavioral expression of this learning was different. Situations like overshadowing and its recovery highlight the difference between acquisition of information and behavioral performance. The former term refers to the actual learning that occurs in any given situation, whereas the latter term refers to the behavioral manifestation of that learning. Performance does not necessarily reflect all associations that have been acquired; acquired associations may be behaviorally silent (e.g., Tolman & Honzik, 1930). Latent Inhibition: Cognition, Neuroscience and Applications to Schizophrenia, ed. R.E. Lubow and Ina Weiner. Published by Cambridge University Press # Cambridge University Press 2010.
62
Latent inhibition: acquisition or performance deficit?
63
Traditionally, deficits in behavioral control by a CS that has been paired with a US have been explained in terms of failures to acquire a CS–US association. For example, when a stimulus (e.g., X) is repeatedly presented alone prior to its being paired with a US (i.e., X–noUS then X–US), the usual observation is that the development of behavioral control by that stimulus is retarded. This so-called latent inhibition effect (Lubow & Moore, 1959) has been explained in many different ways, but most traditional explanations are based on the assumption that X is retarded in acquiring an association to the US due to the preexposure treatment (e.g., Lubow, Weiner, & Schnur, 1981; Mackintosh, 1975; McLaren & Mackintosh, 2000; Pearce & Hall, 1980; Wagner, 1981). A different approach is to view response deficits such as latent inhibition in terms of performance, rather than acquisition deficits. It is this latter perspective that we explore in this chapter. We propose that latent inhibition reflects, at least in part, a failure to express an acquired CS–US association, and support this claim with several experimental reports. We also present a theoretical framework (the comparator hypothesis; Denniston, Savastano, & Miller, 2001; Miller & Matzel, 1988; Stout & Miller, 2007) that accounts for most of the data we review, and discuss the limitations of such an approach.
Latent inhibition as an acquisition deficit Traditional associative models and the construct of attention Probably the best-known associative learning theory is that proposed by Rescorla and Wagner in 1972. The Rescorla–Wagner model proposes a simple linear operator rule to describe how the strength of association between a CS and a US (so-called associative strength) grows with each pairing. In this model, changes in the associative strength of a given CS (DVCS) are a function of salience-determined learning rate parameters for the CS and US (aCS and bUS, respectively), and the extent to which the US is still unpredicted and thus “surprising”. This term is quantified by subtracting the total associative strength of all CSs present on the trial in question (based on the preceding trial) from the learning asymptote (l). These terms are combined in the simple equation, n n1 VCS ¼ aCS bUS l Vtotal ; ð1Þ where n represents the trial being analyzed and n1 the immediately previous trial. In the Rescorla–Wagner model, learning occurs not only when a CS is paired with an unexpected US, but also when a CS is presented in the absence of an expected US. In this case, n n1 VCS ¼ aCS bnoUS 0 Vtotal ; ð2Þ with bnoUS taking a non-zero value lower than bUS because, although the US is not presented, there is an expectation of its occurrence due to the CS presentation.
64
Martha Escobar and Ralph R. Miller
Note, however, that if the CS does not predict the US and the US is not physically present, bnoUS takes a value of zero. That is, the Rescorla–Wagner model assumes that there is no learning without reinforcement or the expectation of reinforcement, an assumption that makes latent inhibition inexplicable in terms of the model (see Miller, Barnet, & Grahame, 1995, for an extensive assessment of the Rescorla– Wagner model). According to the model, learning about the US during the CS–US pairings should proceed at the same rate as a novel cue rather than being retarded. In 1975, Mackintosh proposed a model that emphasized the relevance of attention to the stimulus for the development of conditioning. In his selective attention model, the salience of a stimulus (aCS) is assumed to vary depending on whether or not it predicts changes in the probability of the US occurring. Briefly, a stimulus that predicts a change in the probability of the US increases in salience, whereas a stimulus that is not predictive of change in the probability of the US decreases in salience (see Equations 3b and 3c). The salience of a cue determines the ease with which it can enter into association with a paired US. Mackintosh formalized his theory using a linear operator equation (Equation 3a) that closely resembles that of Rescorla and Wagner (1972), but the value of a for any given CS is determined by comparing the extent to which the CS is a better predictor of the US than are other stimuli. n n1 VCS ; ð3aÞ ¼ aCS bUS l VCS n1 n1 < l Vother ; and ð3bÞ an is positive if l VCS n1 n1 l Vother : an is negative if l VCS
ð3cÞ
When applied to latent inhibition, the model predicts no learning as determined by Equation 3a to occur during the preexposure phase in which the CS is presented without the US because there is no association to the US to modify during preexposure treatment. However, the model does assume that preexposure to the CS results in changes in the salience of (attention to) that CS. According to the model, accounting for the effects of CS preexposure requires the added assumption that the CS enters into association with the preexposure context (which would then play the role of the “other” stimuli). Because neither the CS nor the context predict changes in the probability of the US, an application of Equation 3c would lead to a decrease in the value of aCS (for a similar account through a different mechanism, see Pearce & Hall, 1980). Other attentional models have provided different accounts of latent inhibition, most of them assuming that CS preexposure results in a reduction of CS salience (e.g., Ayres, Albert, & Bombace, 1987; DeVietti et al., 1987; Reiss & Wagner, 1972). However, probably the major development in attentional accounts of latent inhibition was the proposal of Conditioned Attention Theory (CAT) by Lubow and colleagues (e.g., Lubow, 1989; Lubow, Schnur & Rifkin, 1976; Lubow et al., 1981). According to CAT, all stimuli initially command an attentional response, RA.
Latent inhibition: acquisition or performance deficit?
65
Repeated exposures to the stimulus result in conditioned decrements in RA, a decrease that can be attenuated if the preexposed stimulus is paired with another event (e.g., a US). In the case of latent inhibition, repeated presentations of the stimulus followed by the lack of a consequence (i.e., another stimulus) result in conditioned decrements in RA. Although this process is assumed to be akin to classical conditioning, it differs from it in that classical conditioning usually results in the strengthening of a conditioned response, whereas conditioning of inattention is assumed to decrease an already existing response (RA). Because stimuli that do not command attention do not readily enter into associations, preexposed stimuli exhibit subsequent retarded acquisition. Inattention in CAT is assumed to result from a conditioning process; thus, it should follow the same rules as other types of conditioning. For example, pairing the preexposed stimulus with another event (akin to reinforcement) should attenuate latent inhibition because such pairings maintain attention to the preexposed event. Lubow et al. (1976) preexposed rats to a 3-s Stimulus 1 immediately followed by a 1-s Stimulus 2. This treatment resulted in attenuated latent inhibition as compared to control groups in which Stimulus 1 was preexposed alone or Stimulus 1 and Stimulus 2 were explicitly unpaired. Further supporting CAT, Lubow, Wagner, and Weiner (1982; also see Honey & Hall, 1988; Reed, 1995) reported that preexposure to a stimulus compound, in which one of the stimuli was of higher salience than the other, resulted in overshadowing (cf. Pavlov, 1927) of latent inhibition; that is, the lowersalience stimulus displayed less latent inhibition when preexposed in compound with the higher-salience stimulus than when preexposed alone.
Impaired acquisition due to contextual associations CAT was relatively successful in predicting how a number of variables would affect preexposure. However, other predictions of the theory have not been confirmed. For example, as mentioned above, CAT assumes that manipulations that enhance conditioning of CRs in CS–US procedures should also enhance the conditioning of RAs during preexposure procedures. One such manipulation is the so-called superconditioning effect (e.g., Rescorla, 1971; Urushihara, Wheeler, Pinen˜o, & Miller, 2005), in which reinforcing the compound of a novel target CS and a conditioned inhibitor (a stimulus that signals US omission) results in enhancement of conditioned responding to the target CS. McConnell, Wheeler, Urcelay, and Miller (2009) analyzed the effects of compounding a conditioned inhibitor with a target stimulus during preexposure. In their procedure, CS B received conditioned inhibition training (i.e., A–US, AB–noUS, which usually results in conditioned inhibition of B; cf. Pavlov, 1927) prior to preexposing target stimulus X in compound with B (i.e., BX–noUS; Group Inhibition). Control subjects were preexposed to X alone (Group LI), Y alone (Group Acquisition), or the CX compound (Group Compound). Retarded acquisition of behavioral control by X was observed in Group LI as compared to Group
66
Martha Escobar and Ralph R. Miller
Acquisition. However, when preexposure of X occurred in the presence of inhibitory CS B, this retardation was abolished significantly above the (minor) attenuation of latent inhibition produced by compound preexposure. This observation is contrary to CAT’s prediction of enhanced conditioning of inattention (and consequently enhanced retardation of behavioral control) due to the presence of the inhibitor during preexposure. But perhaps more problematic for CAT (and by extension all attentional theories) have been the repeated reports that CS preexposure seems to result in more than just changes in processing of the CS. For example, Channell and Hall (1983) reported that, if stimulus preexposure occurred in Context 1, but conditioning occurred in a different context (Context 2), preexposure was greatly attenuated. That is, latent inhibition exhibits strong context-specificity. These results suggested that mere changes in CS processing could not entirely account for latent inhibition, and a likely alternative mechanism was the development of associations between the CS and the context, which somehow impaired subsequent conditioning. Several theories were developed based on this possibility. Wagner (1981; also see Brandon, Vogel, & Wagner, 2003; Wagner & Brandon, 1989, 2001) proposed a theory which could account for latent inhibition by assuming that strong associations that are formed between the CS and the context during preexposure make the CS unable to enter into associations with the US during the subsequent conditioning phase. According to Wagner’s sometimes opponent processes (SOP) model, all stimulus representations are composed of multiple elements, some of which are sampled with each stimulus presentation or associative activation. Elements are probabilistically moved into an active state (State A1) when the stimulus is physically present. These elements decay into a state of lower activation (State A2) and then into a long-term inactive (I) memory state. Associative activation of a stimulus’s elements into State A2 can occur if an associate of the stimulus is physically present (and thus represented in State A1). Stimulus elements represented into State I can be immediately retrieved for activation into State A1 or State A2; however, stimulus elements represented in State A2 become unavailable for activation into State A1 until they decay back into State I. Excitatory associations are presumed to be formed between elements simultaneously activated into State A1, whereas inhibitory associations are formed between CS elements represented in State A1 and US elements represented in State A2 (see left panel of Figure 4.1). When a CS is preexposed, it is repeatedly presented alone in a context. Thus, elements of both the CS and the context are activated into State A1. With repeated CS presentations, the context comes to associatively activate some elements of the CS into State A2 between CS presentations; this activation becomes more effective as the number of preexposure CS presentations increases. At the time of conditioning, a large proportion of the CS elements would be represented into State A2 and thus unavailable for activation into State A1 during the CS–US pairings. Because associations are formed exclusively between elements simultaneously active in State A1,
67
Latent inhibition: acquisition or performance deficit? SOP
MSOP
A2 pd2
A2 pd2
pd1
pd1
p2
p2 A1
I
A1
I
p1
p1
Activation state
Association change
Activation state
Association change
Simultaneous, A1
Excitation
Simultaneous, A1
Excitation
Simultaneous, A2
No change
Simultaneous, A2
Excitation
CS in A1, US in A2
Inhibition
CS in A1, US in A2
Inhibition
US in A1, CS in A2
No change
US in A1, CS in A2
Inhibition
Figure 4.1. Schematic representation of Wagner’s (1981) SOP (left panel) and Dickinson & Burke’s (1996; also see Aitken & Dickinson, 2005) MSOP (right panel) models. ‘I’, ‘A1’, and ‘A2’ represent the three hypothetical states of activation of stimuli representations (inactive, high activation, and low activation, respectively). ‘p1’ and ‘p2’ represent probabilities of activation (from I into A1 and from I into A2, respectively). ‘pd1’ and ‘pd2’ represent probabilities of decay (from A1 into A2 and from A2 into I, respectively). Dotted lines recurrent in a state represent the model’s assumption that excitatory associations can be formed between activated representations. The table below each graphic representation of a model lists the assumed associations that can be formed. See text for details.
only a few CS elements are available to enter into associations with the US. Note that this approach can easily account for the context-specificity of latent inhibition: a change of context between preexposure and conditioning would result in few or no elements of the CS being activated into State A2 by the conditioning context; thus, more CS elements would be available in State I for activation into State A1. Recently, Dickinson and colleagues (Aitken & Dickinson, 2005; Dickinson & Burke, 1996) suggested that representations simultaneously activated into the same state, either A1 or A2, would develop excitatory associations, whereas representations simultaneously activated into different states would develop inhibitory associations (see right panel of Figure 4.1). This modified SOP (MSOP) was developed to account for retrospective revaluation phenomena, but the modifications do not change the main predictions of the theory in the case of latent inhibition (see below for elaboration). A similar explanation for latent inhibition, albeit through somewhat different mechanisms, was proposed by McLaren and Mackintosh (2000; also see McLaren,
68
Martha Escobar and Ralph R. Miller
Kaye, & Mackintosh, 1989). In their model, all events are viewed as sets of elements or ‘microfeatures’. These elements are probabilistically sampled on each trial, so that a new set of previously sampled and unsampled elements constitutes the perceptual experience on each trial. Each sampled element is represented by a node, which can be activated with strength O (O is assumed to have a value that ranges between þ1 and 1). Activation of a node can occur either through external input (perception of the element without input from another node) or internal input (received from other nodes). Nodes that are simultaneously activated are assumed to become associated with strength w. Thus, if nodes i and j are simultaneously activated, the change of the strength of the i–j association, wij, is given by dwij =dt ¼ Si Oj ;
ð4Þ
where S represents a learning rate parameter, Oj represents activation of node j, and Di represents the difference between external and internal activation of node i (i.e., eiii). The strength of wij is assumed to increase until internal input to the node matches external input (Di ¼ 0). (Note that the use of a delta rule makes Equation 4 functionally equivalent to the Rescorla–Wagner, 1972, error reduction rule.) As learning progresses, elements are also assumed to become less salient, through a process known as salience modulation. Decreases in salience are proportional to the value of D, which can be conceptualized as the discrepancy between the actual (external) and predicted (internal) inputs. In the case of latent inhibition, repeated presentations of the CS during preexposure should result in salience modulation of the CS: the discrepancy between the internal input (associative activation of elements of the CS by the context and other CS elements) and external input (activation of CS elements due to the actual CS presentation), D, becomes smaller with experience. This is equivalent to assuming that strong associations form among the sampled preexposed CS elements, which render these elements ineligible to enter into associations at a later time (note the similarity between this account and SOP’s assumption of associative activation into State A2). Furthermore, because stimuli are preexposed in a context, CS preexposure has the added effect of resulting in the development of associations between the CS elements and the preexposure context. The reliance on changes in both CS processing and CS–context associations as a result of preexposure makes the model powerful and suitable to explain many latent inhibition phenomena. For example, the development of CS–context associations during preexposure provides a mechanism to explain the context-specificity of latent inhibition (a change of context would render the elements associated to the preexposure context available to enter into associations with the US). The model also predicts that extinction of the context following preexposure should attenuate latent inhibition. That is, if preexposure to the CS is followed by long exposure to the context alone, the context–CS associations should become extinguished and the elements previously associated to the preexposure context should then become available for association with the US during the subsequent conditioning phase.
Latent inhibition: acquisition or performance deficit?
69
For example, Westbrook, Bond, and Feyer (1981) preexposed rats to an odor, which was later conditioned as a predictor of illness. For some animals, preexposure to the odor was followed by exposure to the context alone (prior to odor conditioning). This manipulation attenuated the effects of preexposure on subsequent conditioning (also see Baker & Mercier, 1982, Experiments 2, 3, and 4; Wagner, 1979; but see Baker & Mercier, 1982, Experiments 1 and 5; Hall & Minor, 1984).
CS preexposure prevents expression but not acquisition of a CS–US association The role of context on latent inhibition As mentioned above, retardation from CS preexposure can be attenuated if the CS–context association created during preexposure is extinguished prior to the conditioning phase. This is consistent with the view that strong associations between the preexposed CS and the preexposure context prevent the formation of CS–US associations during conditioning (e.g., McLaren & Mackintosh, 2000; Wagner, 1981) because extinguishing the contextual associations prior to conditioning frees the CS to enter into associations with the US. However, long exposure to the preexposure context has an attenuating effect on latent inhibition even if such context extinction takes place after the CS–US conditioning pairings, which contradicts the assumption that CS–context associations developed during preexposure prevent the subsequent acquisition of a CS–US association. For example, Grahame, Barnet, Gunther, and Miller (1994, Experiment 1) preexposed rats to a white noise (Condition LI) or the context alone (Condition noLI) over several days. This preexposure was followed by pairings of the noise and a footshock US. Conditioning in turn was followed by long exposure to the context (Condition Extinction) or equivalent handling without context exposure (Condition noExtinction). Latent inhibition, measured as retarded acquisition in the LI group as compared to the equivalent noLI group, was evident only in the noExtinction condition (i.e., more retardation was evident in Group noLI–noExtinction than Group LI–noExtinction). In the Extinction condition, latent inhibition was abolished (i.e., acquisition did not differ between the LI–Extinction and LI–no Extinction groups). Thus, context extinction attenuated latent inhibition even if it occurred after the CS–US pairings had occurred. In another test of the potential of the context to interfere with the formation of CS–US associations, Escobar, Arcediano, and Miller (2002, Experiment 2) exposed rats to an auditory stimulus prior to pairing the auditory stimulus with a footshock US. In addition, subjects were given short or long exposure to the experimental context before, during, or after CS preexposure. Although latent inhibition was observed in all groups receiving short exposure to the context, it was attenuated when subjects received long context exposure during or after the CS preexposure, but context exposure before CS preexposure did not affect latent inhibition. Notably, McLaren,
70
Martha Escobar and Ralph R. Miller
Bennett, Plaisted, Aitken, and Mackintosh (1994) reported that preexposure to the context in which CS preexposure later occurs not only failed to attenuate latent inhibition, but also reduced the context specificity of latent inhibition. Presumably, this loss of context specificity is observed because, in the McLaren and Mackintosh model terminology, such context preexposure results in latent inhibition of the context elements which then become unavailable to enter into associations with the CS (but see Killcross, Kiernan, Dwyer, & Westbrook, 1998, for a report that exposure to the conditioning context has a similar effect on the context specificity of latent inhibition). Thus, it seems that context exposure prior to CS preexposure does not disrupt latent inhibition, but any exposure to the context after the preexposure trials have begun alleviates the latent inhibition deficit (see Escobar, Arcediano, et al., 2002, for discussion).
Acquisition-based accounts of contextual effects on latent inhibition Recovery from latent inhibition after long exposure to the context is problematic for theories in which CS preexposure is assumed to decrease the subsequent associability of the CS by rendering the CS elements unavailable to enter into associations with the US during conditioning (e.g., Wagner’s SOP, 1981; McLaren & Mackintosh, 2000). As mentioned above, prevention of latent inhibition by extinction of the context either during or after CS preexposure is consistent with these models because such extinction should make the previously unavailable CS elements become available to enter into associations during the subsequent conditioning phase. However, the observation of attenuated latent inhibition when context extinction occurs after conditioning is not consistent with these views. If the CS has not developed associations with the US, it should not produce conditioned responding. In contrast, Dickinson and Burke’s (1996; also see Aitken & Dickinson, 2005) MSOP model can be successfully applied in these situations because it allows for the development or strengthening of representations that are active in memory, even in the absence of external input (see right panel of Figure 4.1). Applied to latent inhibition, MSOP can explain the same range of phenomena initially addressed by SOP (e.g., retarded acquisition after CS preexposure and context specificity of latent inhibition). However, it extends its reach by accounting for the effects of context extinction on latent inhibition, even if that extinction takes place after the CS–US conditioning pairings. During preexposure, associations are assumed to develop between the elements of the CS and context simultaneously activated into State A1. Due to these associations, the context comes to activate CS elements into State A2. If the conditioning phase (i.e., CS–US pairings) occurs in the same context as preexposure, the context will activate a large proportion of the CS elements into State A2. Because the US is physically presented during this phase, its elements will be represented into State A1 and an inhibitory association should form between the US elements activated into State A1 and the CS elements activated
Latent inhibition: acquisition or performance deficit?
71
into State A2. Thus, MSOP provides a mechanism to account for the effects of context extinction on latent inhibition while continuing to assume that latent inhibition reflects a deficit to acquire a CS–US association as a result of CS preexposure.
Attenuation of latent inhibition independent of context extinction From the previous sections, one might conclude that at least some acquisitionfocused approaches can encompass the data concerning latent inhibition. However, several phenomena appear problematic, if not incompatible, with all contemporary acquisition-focused approaches. Reminder treatments Kasprow, Catterson, Schachtman, and Miller (1984) gave rats conventional latent inhibition training with an auditory CS, followed by CS–US pairings. During a subsequent test trial, the stimulus was presented while animals were actively drinking water and conditioned responding was assessed. Prior to testing, some animals received two presentations of the US, which were intended to serve as ‘reminders’ for the CS–US conditioning phase. Animals receiving this treatment exhibited more responding to the preexposed CS (i.e., disruption of latent inhibition) than their matched controls. The increase in responding observed in animals receiving the US reminders did not appear to result from contextual conditioning summating with conditioning of the CS because the reminder US presentations occurred in a context different from that of preexposure and conditioning. Furthermore, the reminder trials did not increase responding to stimuli that had not received conditioning, suggesting a stimulus-specific effect. These results suggest that CS preexposure does not prevent the subsequent acquisition of the CS–US association; this association appears to have been acquired but not expressed in behavior. Attenuation of latent inhibition by a retention interval Aguado, Symonds, and Hall (1994; also see Killcross et al., 1998; Kraemer, Randall, & Carbary, 1991; Kraemer & Roberts, 1984) exposed rats to a saccharin solution (CS) prior to pairing that solution with an injection of lithium chloride (US). Rats in the Short condition were tested for consumption of the saccharine solution 2 days after conditioning (the delay was necessary to restabilize drinking), whereas animals in the Long condition received their test trials 12 days after conditioning (matching control groups were not preexposed to saccharin). With these parameters, the Long condition exhibited significantly less latent inhibition than the Short condition, suggesting that, although conditioning was retarded in the Short condition, the CS–US association had been acquired but not expressed, and it was uncovered by the 12-day retention interval. Such restoration of responding would not have occurred if acquisition of the CS–US association had been prevented by the
72
Martha Escobar and Ralph R. Miller
preexposure treatment. We must note, however, that attenuation of latent inhibition due to retention intervals is somewhat controversial, as several studies have found an enhancement rather than attenuation of latent inhibition with a retention interval (e.g., De la Casa & Lubow, 2000; Wheeler, Stout, & Miller, 2004). De la Casa and Lubow suggested that this so-called super-latent inhibition effect occurs when the retention interval is spent outside the training context rather than in the training context (as is usual in conditioned taste aversion studies as those of Aguado et al.), which may make the effect of the retention interval an extension of the context extinction effects discussed above (see Wheeler et al. for further discussion).
Renewal of latent inhibition Bouton and Swartzentruber (1989) exposed rats to a light CS in Context 1. This CS preexposure was followed by further CS preexposure in either Context 2 or Context 3. This second round of preexposure was followed by CS–shock US conditioning pairings in the original preexposure context, Context 1. Conditioning proceeded until all animals exhibited reliable conditioned suppression. Testing then occurred in Context 2. Animals that received the second round of preexposure in Context 3 exhibited significantly more responding to the CS than animals that received their second round of preexposure in Context 2 (for simplicity, some groups are omitted from this description). That is, latent inhibition was renewed by testing in a context of preexposure. These results are problematic for acquisition-based accounts of latent inhibition such as SOP, MSOP, and McLaren and Mackintosh’s theories because they assume that the associative value of a specific CS (with respect to a US) is updated with training into a unitary value. However, it seems that training of the CS–US association does not change a unitary associative value of the CS; rather, experiencing the nonreinforced exposures created lasting memories that were not erased by the conditioning training. If latent inhibition was a pure acquisition deficit, conditioning should alter the net associative value of the CS, thereby ‘erasing’ all effects of preexposure. In summary, it seems that an account of latent inhibition based exclusively on acquisition processes does not provide a full explanation of the effects of CS preexposure. It is apparent that subjects acquire information about the first phase of the latent inhibition procedure, which can be conceptualized as decreases in salience or associability of the CS (Mackintosh, 1975; Pearce & Hall, 1980), or unavailability of CS elements to later enter into associations with the US (Aitken & Dickinson, 2005; McLaren & Mackintosh, 2000; Wagner, 1981). However, it is also apparent that whatever information is acquired during CS preexposure does not preclude the acquisition of information about the CS–US relationship during subsequent conditioning. Furthermore, the preexposure and conditioning experiences seem to be retained by subjects and used in adaptive ways to determine behavior. These observations have led to the proposal of learning models that emphasize performance
Latent inhibition: acquisition or performance deficit?
73
rather than acquisition. Performance-based models assume that both phases of the latent inhibition procedure (i.e., preexposure and conditioning) provide subjects with information, and that response rules applied at the time of retrieval (i.e., performance) determine which information will be expressed.
Latent inhibition as a performance deficit Latent inhibition as a type of interference In the early 1990s, Bouton (e.g., 1991, 1993, 1994) proposed that several traditional associative learning phenomena be considered to be instances of interference. Interference can be defined as a situation in which conflicting memories compete for retrieval; the memory that is retrieved (usually at the expense of the competing memories) immediately controls behavior. For example, the phenomenon of extinction (cf. Pavlov, 1927), in which CS-alone presentations following CS–US pairings usually result in attenuated responding to the CS, can be viewed as the second phase of training retroactively interfering with memories of the first phase of training. Conversely, latent inhibition (Lubow & Moore, 1959), in which CS-alone presentations followed by CS–US pairings result in attenuated responding to the CS, can be viewed as a form of proactive interference (attenuated retrieval of the second phase of training). Bouton’s (1993) model assumes that when a CS has two or more consequences (e.g., US vs. no US), the CS becomes ambiguous. This ambiguity is hypothetically represented as multiple associations being formed between the CS and its various consequences. One of the most extensively studied interference phenomena addressed by Bouton is extinction. The model assumes that the first phase of the extinction procedure (i.e., acquisition) results in the development of excitatory CS–US associations, whereas the second phase of the procedure results in the development of inhibitory CS–noUS associations (see Figure 4.2a). A different structure would account for the phenomenon of counterconditioning (cf. Sherrington, 1947), in which the CS is paired with two different USs in successive phases of training (i.e., CS–US1 then CS–US2). In this situation, the model assumes that excitatory CS–US1 associations are formed during the first phase of the procedure. During the second phase, excitatory associations are formed between the CS and US2 and, because US1 is absent during the CS–US2 pairings, inhibitory associations are assumed to form between US1 and US2, as well as between the CS and US1 (see Figure 4.2b). In the case of latent inhibition, the initial CS-alone presentations are regarded as resulting in some type of learning that is neither excitatory nor inhibitory. This assumption is based on the observation that CS preexposure not only retards the development of excitation, but also retards the development of inhibition (e.g., Friedman, Blaisdell, Escobar, & Miller, 1998; Reiss & Wagner, 1972; Rescorla, 1971). Bouton (1993) suggested that latent inhibition be explained through a procedure similar to
74
Martha Escobar and Ralph R. Miller
(a)
(b)
(c)
US
Insignif.
US1
US2 CS
CS Context
US CS
Context
Context
Figure 4.2. Schematic representation of Bouton’s (e.g., 1994) interference model. Panel a represents extinction, panel b represents counterconditioning, and panel c represents latent inhibition. In all cases, the context serves to disambiguate the multiple associations acquired to the test CS, excitation and inhibition to the US (extinction), US1 and US2 (counterconditioning), or Insignificance and the US (latent inhibition). Arrowheads denote excitatory association, Ts denote inhibitory associations, small circles denote modulatory gates. See text for details.
counterconditioning. During the CS-alone presentations, subjects presumably form an association between the CS and ‘Insignificance’ and then, during conditioning, an association between the CS and US (see Figure 4.2c). In situations in which the CS is ambiguous due to its being associated to multiple outcomes, which association comes to control behavior is determined by the context of retrieval, which has both physical and temporal characteristics. Bouton (e.g., 1994) has also suggested that associations that result in elicited behavior (excitatory) are less context-specific than associations that result in behavior attenuation (inhibition and insignificance). In consequence, associations that lead to behavior attenuation are assumed to require input from both the stimulus presentation and the context in order to exert behavioral control. In the case of latent inhibition, expression of the CS–Insignificance association acquired during preexposure would require that the association be assessed in the context in which it was acquired. This assumption allows the model to account for the context-specificity of latent inhibition: a change of context between preexposure and conditioning would not result in retarded acquisition because the contextual input would not activate the CS–Insignificance association. Nelson (2002; also see Bouton, 1997; Sissons & Miller, 2009) later additionally suggested that the association learned second, whether excitatory or inhibitory, should be context-specific. This account of context-specificity explains why retarded acquisition after latent inhibition transfers to a novel physical test context (e.g., Escobar, Arcediano et al., 2002). Furthermore, Bouton’s model easily accounts for renewal of latent inhibition: if latent inhibition is assessed in the context of preexposure, the CS–Insignificance association would be activated and little responding should be observed.
Latent inhibition: acquisition or performance deficit?
75
Bouton (1993; also see Rosas & Bouton, 1997) suggested that his interference approach was compatible with some of the conflicting data on spontaneous recovery from latent inhibition. He suggested that the contextual modulation of the CS–Insignificance association might be labile and decrease fast with the passage of time; this decrease should occur at a higher rate than the decay in the representation of the CS–Insignificance association. Thus, with long retention intervals between conditioning and testing, latent inhibition should be enhanced (i.e., super-latent inhibition [cf. De la Casa & Lubow, 2000] should be observed) because the inhibitorylike modulation by the context would decay rapidly and the CS–Insignificance association can then be expressed. The model also assumes that latent inhibition should decrease at longer retention intervals due to decay of the CS–Insignificance association with the passage of time. Although both increased and decreased latent inhibition at different retention intervals have been reported in the literature, the consistent change from attenuated to enhanced latent inhibition as the retention interval increases predicted by Bouton is not apparent. Increased latent inhibition has been observed with retention intervals of 21 days in a conditioned taste aversion preparation (De la Casa & Lubow), and 27 days in a conditioned suppression preparation (Wheeler et al., 2004), whereas attenuated latent inhibition has been observed with retention intervals of 12 days (Aguado et al., 1994) and 21 days (Kraemer et al., 1991; Kraemer & Roberts, 1984) in conditioned taste aversion, and 14 days in conditioned freezing (Killcross et al., 1998). As mentioned above, De la Casa and Lubow reported that the main variable determining whether enhanced latent inhibition was observed with delayed testing was whether subjects spent the retention interval in the experimental apparatus; long periods in the experimental apparatus in which CS preexposure occurred (i.e., context extinction) seemingly reduce latent inhibition. An account of latent inhibition that shares many features with Bouton’s approach is that proposed by Lubow and Gewirtz (1995). According to this account, the context of preexposure serves as an occasion setter for a CS–no-consequence association. An occasion setter is a stimulus that can modulate the expression of an association independently of the occasion setter’s own associative value (e.g., Holland, 1992; Miller & Oberling, 1998). If conditioning occurred in the context of preexposure, the context sets the occasion for the CS–no consequence association and retardation should be observed. Functionally, this occasion setting model has explanatory power similar to that of Bouton’s (1993) retrieval model.
Latent inhibition as a comparison of acquired associations: the comparator hypothesis Miller and Matzel (1988) proposed an account of conditioned behavior based on the assumption that all experiences are acquired and stored in memory. These experiences are compared at the moment of retrieval to determine behavioral control. Their
76
Martha Escobar and Ralph R. Miller
(a)
(b) 1 Test CS
2 Comparator stimulus representation
Directlyactivated US representation
comparator
3
Indirectlyactivated US representation
1 Preexposed CS
CR
2
Preexposure context
Directlyactivated US representation
comparator
3
cr
Indirectlyactivated US representation
Figure 4.3. Panel a presents a graphic description of Miller and Matzel’s (1988) comparator hypothesis. Rectangles are observable stimuli and responses. Ovals are internal representations. The test presentation of the target stimulus is assumed to directly activate a representation of the US (Link 1) and also a representation of its comparator stimulus (Link 2), which in turn activates a representation of the US (Link 3). Activation of the US through the indirect path is supposed to have strength proportional to the product of Links 2 and 3. The directly and indirectly activated US representations are fed into the comparator mechanism. The probability of responding increases with increases in the directly activated US representation and decreases with increases in the indirectly activated US representation. ‘CR’ represents strong conditioned responding. Panel b presents the comparator hypothesis as applied to latent inhibition. In this situation, the preexposure context acts as a comparator stimulus. Heavier arrows represent greater strengths for associations. The product of (robust) Link 2 and (moderate) Link 3 results in more indirect than direct activation of the US representation and, consequently, attenuated conditioned responding (‘cr’).
comparator hypothesis is based on the assumption that US representations can be activated both directly through presentation of stimuli paired with it and indirectly through associates of stimuli presented at test which also have an association to the US (see Figure 4.3a). If a given CS, X, is paired with the US, subsequent conditioned responding to X will be determined by a comparison of the extent to which X activates a representation of the US versus the extent to which stimuli associated to X (so-called comparator stimuli; in the case of a simple CS, the training context) also activate a representation of the US. That is, the model assumes a comparison is made between the strength of the target CS–US association between the strength of (Link 1) and the product of the strength of the target CS–comparator stimulus (Link 2) and comparator stimulus–US association (Link 3; see Figure 4.3a). Conditioned responding is assumed to vary directly with magnitude of the directly activated US representation, and to vary inversely with magnitude of the indirectly activated US representation. Note that according to Miller and Matzel’s comparator hypothesis, the context is the comparator stimulus only in certain situations. The comparator stimulus is assumed to be the one stimulus, other than the US, that has the strongest association with the test CS. Furthermore, the original 1988 version of the comparator hypothesis does not make predictions about absolute levels of responding; rather, it predicts relative levels of responding between groups that are differentially treated.
Latent inhibition: acquisition or performance deficit?
77
In the case of latent inhibition (see Figure 4.3b), the preexposed CS is assumed to develop strong associations to the context during preexposure treatment. Thus, the context of preexposure should be the preexposed CS’s comparator stimulus. Then, during the conditioning phase, both the CS and the context become associated with the US. At test, presentation of the preexposed CS should directly activate a representation of the US through Link 1. Simultaneously, the CS should activate a representation of the context (its comparator stimulus) through Link 2. In turn, the context then activates a representation of the US through Link 3. The US representation directly activated by the CS is then compared to the US representation indirectly activated by the CS through the context. Because of the preexposure manipulation, it is assumed that the CS–context association (Link 2) is very strong (represented by a heavy line in Figure 4.3b). The product of the strong Link 2 and the moderate Link 3 (i.e., the context–US association) should result in greater activation of the US representation through the indirect pathway than if preexposure had not occurred. Because the comparator hypothesis explains latent inhibition in terms of the preexposed CS–context association, it is well suited to account for the effects of context manipulations on latent inhibition. A change of context between preexposure and conditioning would disrupt latent inhibition because the preexposure context (strong Link 2) does not have an association with the US (i.e. Link 3) and, due to the weaker activation of the indirect pathway, Link 1 should be expressed as conditioned responding to the CS. The model can also account for the effects of context extinction prior to or after conditioning by predicting that such treatment should result in a weaker activation of the US through the indirect pathway. In the former case, context extinction should weaken the CS–context association (Link 2), whereas in the latter case it would weaken both the CS–context association (Link 2) and context–US association (Link 3). Furthermore, the model appears compatible with both the observation that context extinction attenuates latent inhibition (Baker & Mercier, 1982, Experiments 2, 3, and 4; Escobar, Arcediano et al., 2002; Grahame et al., 1994; Wagner, 1979; Westbrook et al., 1981) and the observation that context extinction has no effect on latent inhibition (Baker & Mercier, 1982, Experiments 1 and 5; Hall & Minor, 1984). These seemingly conflicting results can be subsumed by the comparator hypothesis if one assumes that the extinction treatment should be extensive enough to effectively attenuate the indirectly activated representation of the US. Insufficient context extinction would result in no apparent effect on latent inhibition (see Grahame et al., 1994, for elaboration).
Testing for the associative structure of latent inhibition Different theories of latent inhibition each propose that the phenomenon has a distinct associative structure. In general, acquisition-based theories assume that the stimulus has a singular associative value that is updated with each trial (e.g., VCS in
78
Martha Escobar and Ralph R. Miller
Rescorla–Wagner, 1972), whereas retrieval-based theories assume the simultaneous existence of multiple associations that are disambiguated by the test context (e.g., Bouton, 1993, 1994), or multiple associations that compete at the time of retrieval (e.g., Miller & Matzel, 1988). Which of these associative structures is most consistent with the empirical data may be determined through the study of interactions between latent inhibition and other behavioral phenomena. Blaisdell, Bristol, Gunther, and Miller (1998) assessed the effects of combining two response-attenuating treatments. The first treatment was latent inhibition, which according to Miller and Matzel’s (1988) comparator hypothesis should result in the associative structure presented in Figure 4.3b as described above. The second treatment was overshadowing (cf. Pavlov, 1927), in which training of moderately salient CS X in conjunction with highly salient CS A (i.e., AX–US) results in less responding to CS X than if CS X had been trained alone (i.e., X–US). The comparator hypothesis explains the overshadowing deficit in a manner similar to the latent inhibition deficit: test presentation of CS X should directly activate a representation of the US, reflecting the strength of the X–US association (Link 1), and indirectly activate a representation of the US, reflecting the product of the strength of the X–A (Link 2) and A–US (Link 3) associations. Due to its high salience, associations to CS A are assumed to be strong; thus, responding to X should be more attenuated in the overshadowing than in the control group. Blaisdell et al. (1998) used a 2 2 factorial design with the variables being latent inhibition treatment (LI) or no latent inhibition treatment (noLI) in Phase 1, and overshadowing treatment (i.e., AX–US; OV) or no overshadowing treatment (i.e., X–US; noOV) in Phase 2. CS X was a weak auditory signal and CS A was a highly salient visual stimulus. Finally, all groups were assessed for responding to CS X. (A fifth group designed to control for the possibility of generalization decrement due to conditioning with a compound and testing with an element in Condition OV was also included, but is not discussed here.) It seems logical to assume that, if preexposure and overshadowing each produce a response deficit, combination of both treatments should result in an even larger response deficit. However, Blaisdell et al. observed the opposite: even though Groups LI–noOV and noLI–OV exhibited little conditioned responding to CS X, Group LI–OV exhibited levels of responding similar to the noLI–noOV control group, suggesting that the two phenomena had counteracted each other (for analogous results in a conditioned taste aversion preparation, see Ishii, 1999; Loy & Hall, 2002; but see Nakajima, Ka, & Imada, 1999; Nakajima & Nagaishi, 2005, for evidence of summation of latent inhibition and overshadowing rather than counteraction in conditioned taste aversion). The latent inhibition–overshadowing (LI–OV) counteraction effect is not consistent with acquisition-based accounts of latent inhibition. If attention to or associability of the CS is attenuated by the preexposure treatment (e.g., Lubow et al., 1981; Mackintosh, 1975; McLaren & Mackintosh, 2000; Pearce & Hall, 1980; Wagner, 1981), overshadowing should occur without impairment or be enhanced. The LI–OV
Latent inhibition: acquisition or performance deficit?
79
counteraction effect could be explained in terms of interference (e.g., Bouton, 1993) if one assumes that the CS X–Insignificance exposures of the preexposure phase proactively interfered with the establishment (or expression) of the X–A association during overshadowing training. However, this account is unable to explain why associative inflation of the context (i.e., unsignaled outcome presentations in the training context) attenuates the LI–OV counteraction effect but inflation of CS A (i.e., A–US pairings) does not (Blaisdell et al., 1998, Experiment 3). The results of Blaisdell et al. are also incompatible with Miller and Matzel’s (1988) comparator hypothesis, which states that only the one stimulus with the strongest association to the target CS (i.e., strongest Link 2) can serve as a comparator stimulus for the target. To account for their observations, Blaisdell et al. (1998) proposed that the comparator hypothesis could be modified to accommodate the LI–OV counteraction effect, by assuming that the two potential comparators (the context and CS A) each serve as a comparator stimulus for the other as well as for CS X. Thus, establishment of the context as a comparator stimulus for CS X during preexposure should block subsequent establishment of CS A as a comparator stimulus for CS X during overshadowing training. Consequently, inflation of the context should reduce responding to CS X after LI–OV treatment because it strengthens the role of the context as a comparator stimulus. These assumptions are also compatible with Miller, Esposito, and Grahame’s (1992) observation of competition between comparator stimuli. These observations led to the development of an extension of the comparator theory, which has proven to be extremely successful in explaining the effects of combining response-attenuation procedures (see Wheeler & Miller, 2008, for a review).
Competition at the time of retrieval: the extended comparator hypothesis As previously mentioned, one of the potential limitations of Miller and Matzel’s (1988) comparator hypothesis is that it predicts that only the stimulus with the strongest association to the test CS can act as a comparator stimulus. Blaisdell et al. (1998) proposed that higher-order comparator effects can also influence responding to CS X. These higher-order processes presumably serve to determine the effectiveness of the associations that directly modulate responding to CS X. Thus, the effectiveness of Link 2 (the CS–comparator association) is determined by its own comparator process, the result of which would determine activation of the comparator stimulus representation. The comparator process regulating Link 2 would then be considered a second-order comparator process. Link 3 would also be subject to its own second-order comparator process which, in conjunction with the modulated Link 2, determines the strength of the indirectly activated US representation. Denniston et al. (2001) stated the principles governing the modulation of Links 2 and 3 through second-order comparator processes. Their so-called extended comparator hypothesis (ECH) posits that all stimuli associated to the target cue have the
80
Martha Escobar and Ralph R. Miller
1
Directly activated US representation
Test CS 2.1 Test CS
Directlyactivated C1 representation 2
comparator
2.2 C2,2 representation 2.3
Indirectly activated C1 representation
Effective C1 representation
3
C1 representation
Conditioned response
Effective indirectly activated US representation
3.1
Directly activated US representation
3.2 C2,3 representation
3.3
Indirectly activated US representation
Figure 4.4. The extended comparator hypothesis (Denniston, Savastano, & Miller, 2001). The effectiveness of Links 2 and 3 is determined by their own (second-order) comparator processes. C1 represents first-order comparator stimuli, and C2,2 and C2,3 represent second-order comparator stimuli for Links 2 and 3, respectively. Black diamonds represent comparator processes. See text for details.
potential to act as first-order comparator stimuli. Although a large number of comparator stimuli could overwhelm the system, not all of these stimuli act as effective comparator stimuli. The effectiveness of a comparator is given by the extent to which its representation is activated by the test stimulus after the higher-order modulation has been taken into consideration. Figure 4.4 presents a graphic representation of ECH. Test presentation of the target CS is assumed to activate a representation of its first-order comparator stimuli (C1; i.e., the stimuli with which it has a strong within-compound association). For any given first-order comparator stimulus, the strength of Link 2 is assumed to depend on the extent to which the test CS activates a representation of a first-order comparator stimulus (C1) and the extent to which other stimuli associated with both the test CS and C1 also activate C1. Note that this higher-order process is meant to mirror Miller and Matzel’s (1988) comparator mechanism. In the case of C1, its effective activation is determined by a comparison between the strength of the target CS–C1 association (Link 2.1 in Figure 4.5) vs. the product of the strength of the association between the target CS and the second-order comparator stimulus (Link 2.2) and the strength of the association between the second-order comparator stimulus and first-order comparator
81
Latent inhibition: acquisition or performance deficit? (a) 1
US(d)
X
2.1 Context(d)
X
2
2.2
Context(i)
A 2.3
Context′
US′(i)
3
3.1 US(d)
Context 3.2
A 3.3
US(i)
(b) 1
US(d)
X
2.1 A(d)
X
2
2.2
A(i)
Context 2.3
US′(i)
A′ 3
3.1 US(d)
A 3.2
US(i)
Context 3.3
Figure 4.5. Application of the extended comparator hypothesis to the LI–OV counteraction effect (Blaisdell et al., 1998). Panels a and b present the hypothetical structures resulting from using the context and overshadowing CS A as first-order comparator stimuli of target CS X, respectively. Context’, A’, and US’ represent the effective activation of the respective stimuli. The subscripts ‘(d)’ and ‘(i)’ represent activation through the direct and indirect pathway,
82
Martha Escobar and Ralph R. Miller
stimulus (Link 2.3). A parallel process is assumed to determine the effective strength of Link 3. Activation of a representation of C1 will in turn activate a representation of the US with a strength that reflects the comparison of the strength of the C1–US association (Link 3.1) vs. the product of the strengths of Link 3.2 and Link 3.3. Figure 4.5 presents a graphic representation of the LI–OV group described above. Note that there are two potential comparator stimuli for the target CS, X: the context of preexposure and the overshadowing stimulus, A. Figure 4.5a presents the application of ECH to the context as C1, whereas Figure 4.5b presents the application of ECH to the overshadowing stimulus, A, as C1. As can be seen in these figures, the X–context association acquired during preexposure attenuates the effectiveness of A as a first-order comparator stimulus by down-modulating the strength of Link 2 (see Figure 4.5b). Thus, the context should become the most effective first-order comparator for CS X. However, indirect activation of the US representation by the context is not robust enough to attenuate responding to CS X because the strong A–US association acquired during the overshadowing training phase downmodulates the strength of Link 3 for the context as a first-order comparator stimulus (see Figure 4.5a). By specifying that the context is the most effective first-order comparator stimulus for CS X, ECH predicts that inflation of the context (i.e., strengthening of Link 3.1 in Figure 4.5a) should attenuate the LI–OV counteraction effect because it presumably strengthens the effective indirectly activated US representation. In contrast, inflation of CS A should not attenuate the LI–OV counteraction effect because A is rendered an ineffective comparator stimulus by the CS-preexposure treatment. Similarly, any procedure that attenuates the latent inhibition effect should also attenuate the LI–OV counteraction effect because both are based on the development of strong associations between the CS and the context. Blaisdell et al. (1998) provided evidence supporting both of these predictions. In their Experiment 3, they observed that associative inflation of the context disrupted the LI–OV counteraction effect (i.e., attenuated responding to CS X). In contrast, inflation of A did not abolish the LI–OV counteraction effect. In their Experiment 4, Blaisdell et al. observed that a change of context between preexposure and overshadowing treatment disrupted the LI–OV counteraction effect. In this situation, conditioning the AX compound in a novel context prevents the preexposure context from down-modulating the effectiveness of A as a comparator stimulus because Link 2.3 (see Figure 4.5b) would be nonexistent. Caption for figure 4.5. (cont.) respectively. Heavier arrows represent greater strengths for associations. The strong CS–context association developed during the preexposure phase (Link 2.2 of panel b) decreases the effectiveness of CS A as a first-order comparator stimulus, whereas the robust A–US association (Link 3.3 of panel a) decreases the effectiveness with which the context (an effective first-order comparator) activates a US representation. Thus, the two potential firstorder comparators down-modulate each other’s effectiveness as comparator stimuli. See text for details.
Latent inhibition: acquisition or performance deficit?
83
Up- and down-modulation of the effectiveness of Links 2 and 3 has direct and predictable effects on behavior. Changes in the strengths of the target CS–C1 or C1–US links (i.e., associative changes to the first-order comparator stimulus) should have an inverse effect on responding to the target CS. That is, making Links 2 or 3 stronger should decrease conditioned responding to the target (CS X) and making Links 2 or 3 weaker should decrease responding to the target. This is consistent with the observation of increased responding due to extinction of the context after latent inhibition training (weakening Link 2 enhances responding to the target CS; e.g., Baker & Mercier, 1982; Escobar, Arcediano, et al., 2002; Grahame et al., 1994; Wagner, 1979; Westbrook et al., 1981) and decreased responding due to inflation of the context after LI–OV training (strengthening Link 3 decreases responding to the target CS; e.g., Blaisdell et al., 1998, described above). The comparator hypothesis explanation of latent inhibition in terms of CS–context associations can account for observations previously explained in attentional terms. For example, Lubow et al. (1976) reported that if CS X was followed by another CS (e.g., A) during preexposure, latent inhibition was attenuated (as compared to a condition in which CS X was preexposed alone). In terms of the ECH, both A and the context should become comparator stimuli for X, and they should simultaneously down-regulate each other’s effectiveness in serving as comparator stimuli for X. A similar situation would ensue in the case of overshadowing of latent inhibition (e.g., Honey & Hall, 1988; Lubow et al., 1982; Reed, 1995), in which preexposure to the AX compound (with A being more salient than X) results in less latent inhibition than when X is preexposed alone. As in the previous case, A should serve to down-modulate the effectiveness of the context as comparator stimulus for X, resulting in disrupted latent inhibition.
A formal implementation of ECH: the sometimes competing retrieval model (SOCR) Although Miller and Matzel’s (1988) comparator hypothesis and Denniston et al.’s (2001) ECH are models with broad explanatory power, they rely on qualitatively weighing associations to determine conditioned responding. To address this limitation and increase the range of phenomena explained by the model, Stout and Miller (2007) proposed a mathematical implementation of ECH that they termed the sometimes competing retrieval model (SOCR). (SOCR is pronounced /sǒk0 ∂r/, just like the wonderful ball game.) For the purposes of this paper, only those aspects of the model relevant to latent inhibition will be described. (For a previous mathematical implementation of the comparator hypothesis, see Baker, Murphy, Valle´e-Tourangeau, & Mehta, 2001.) The comparator hypothesis and ECH have been described as performance rules and assume acquisition of associations using a noncompetitive version of the delta rule. The rule of choice is equivalent to that of Bush and Mosteller (1955), and the change in the association between two paired stimuli, S1 and S2, on any given trial, n, can be described as,
84
Martha Escobar and Ralph R. Miller n n1 VS1;S2 ¼ s1 s2 ð1 VS1;S2 Þ;
ð5Þ
where s1 and s2 represent the salience of Stimuli S1 and S2. Note the superficial similarity between this acquisition rule and that of Rescorla and Wagner (1972; see Equation 1). The critical difference is that, in SOCR, different stimuli do not compete for a limited amount of associative strength (such competition is implied by Rescorla– n1 Wagner in the term Vtotal ; see Equation 1); rather, each stimulus present can acquire full asymptotic associative strength. Furthermore, SOCR’s acquisition rule is not limited to CSs and USs, but rather accounts for the development of associations between all types of stimuli, whatever their role in the association. This characteristic allows the model to account for the development of associations between the CS and context during CS-preexposure treatment, even before the first US presentation. SOCR accounts for extinction of an association between two stimuli using a modification of Equation 5, which represents the change in association for Stimuli 1 and 2 for any trial, n, n n1 VS1;S2 ¼ k1 s1 ð0 VS1;S2 Þ;
ð6Þ
where k1 represents a constant with a value greater than zero and lower than one, and the asymptote of 1 (see Equation 5) has been replaced by a zero value. The response rule in both the comparator hypothesis and ECH assumes a comparison between Link 1 (test stimulus–outcome stimulus association) and the product of Links 2 (test stimulus–comparator stimulus association) and 3 (comparator stimulus– outcome stimulus association). SOCR implements this comparison by subtracting the strength of the indirect pathway of outcome activation (product of Links 2 and 3) from the direct pathway of outcome activation (Link 1). Thus, for any target stimulus x and comparator stimulus j, the comparison of activation of outcome o activation through the direct and indirect pathways is represented as, n n n Rnx ¼ Vx;o k2 ðVx;j Vj;o Þ;
ð7Þ
n represents the strength of Link 1 where Rnx represents responding to X on trial n, Vx;o n (target CS–outcome stimulus association), Vj;o represents the strength of Link 3 n (comparator stimulus–outcome stimulus association), and Vx;j represents the strength of Link 2 (target CS–comparator stimuli association). Note that the latter term allows for multiple stimuli to serve as comparator stimuli. The combined effect of these comparator stimuli can affect responding to the target up to a certain limit, which is provided by parameter k2 in Equation 7.
Limitations of the comparator view of latent inhibition The comparator view of latent inhibition is not without its failures. One of the main problems with the model is that it fails to account for recovery effects that are unrelated to direct manipulation of the stimuli being studied. The model can easily
Latent inhibition: acquisition or performance deficit?
85
account for the reduction in latent inhibition that follows a change of context between preexposure and conditioning (e.g., Channell & Hall, 1983) and extinction of the context of preexposure (e.g., Escobar, Arcediano, et al., 2002; Grahame et al., 1994). However, it fails to directly account for the effects of testing in the preexposure context (i.e., renewal; e.g., Bouton & Swartzentruber, 1989) and the effects of presenting a so-called reminder of the CS–US conditioning phase (i.e., reinstatement; Kasprow et al., 1984). These phenomena are consistent with the spirit of the theory because they reflect the acquisition of information during both the preexposure and conditioning phases (i.e., with latent inhibition being a performance, rather than an acquisition, deficit), but as currently stated the theory has no specific mechanism to account for these effects (see Stout & Miller, 2007, for further discussion). Comparator theory also has difficulties explaining the loss of context-specificity observed when CS preexposure is conducted in multiple contexts (Wheeler, Chang, & Miller, 2003) or when subjects are preexposed to the context of conditioning (Killcross et al., 1998).
Applications of a performance-deficit view of latent inhibition The view that latent inhibition may reflect a decrement in attention (e.g., CAT; Lubow et al., 1981) has led many researchers to regard the latent inhibition phenomenon as a benchmark for the normal functioning of attention. Conversely, disruption of latent inhibition can then be viewed as an example of abnormal functioning of attention. Indeed, acute schizophrenic patients apparently fail to show the latent inhibition deficit (e.g., Baruch, Hemsley, & Gray, 1988; Gray, Pilowski, Gray, & Kerwin, 1995; Vaitl et al., 2002; Yogev, Sirota, Gutman, & Hadar, 2004; but see e.g., Gray et al., 1995; Lubow, Weiner, Schlossberg, & Baruch, 1987; Swerdlow et al., 1996, 2005; Vaitl et al., 2002; Williams et al., 1998). In consequence, latent inhibition has been regarded as a model for the attentional dysfunction that characterizes acute schizophrenia (e.g., Feldon & Weiner, 1991; Lubow, 1997, 2005; Lubow & Gewirtz, 1995). The approach favored in the present paper suggests that latent inhibition is not necessarily (or exclusively) related to the functioning of attention, but that it may reflect the development of associations between the preexposed CS and the context. This approach can be viewed as consistent with the suggestion that disrupted latent inhibition in acute schizophrenics does not reflect primarily an attentional deficit, but a deficit in acquiring associations to contextual stimuli (e.g., Cohen & ServanSchreiber, 1992; Oberling, Gosselin, & Miller, 1999). Escobar, Oberling, and Miller (2002; also see Oberling et al., 1999) proposed an extension of Miller and Matzel’s (1988) comparator hypothesis to account for some of the empirical data linking latent inhibition and acute schizophrenia. They noted that disrupted latent inhibition in acute schizophrenic patients was not “forever”: latent inhibition is restored as psychotic symptoms decrease with the use of medication (Baruch et al., 1988; Leumann, Feldon, Vollenweider, & Ludewig, 2002; Vaitl et al., 2002), and it is observed in patients experiencing the decrease in symptoms that
86
Martha Escobar and Ralph R. Miller
accompanies chronic schizophrenia (Gray et al., 1995; Vaitl et al., 2002). Restoration of latent inhibition due to the use of antipsychotic medication occurred even when a previous test in the non-medicated state showed disrupted latent inhibition (Baruch et al., 1988), suggesting that acute schizophrenic patients actually learned something about the CS during the preexposure phase. This observation may be viewed as renewal of latent inhibition. Furthermore, Escobar, Oberling, et al. noted that acute schizophrenia not only disrupts latent inhibition, but also the blocking effect (Bender, Muller, Oades, & Sartory, 2001; Jones, Gray, & Hemsley, 1992; Jones, Hemsley, Ball, & Serra, 1997; Moran, Al-Uzri, Watson, & Reveley, 2003; Oades, 1997; Oades, Zimmerman, & Eggers, 1996). In a blocking paradigm (Kamin, 1968), target CS X is paired with the US in the presence of another stimulus, A. In the Control group, A is a novel stimulus (i.e., no treatment, then AX–US), whereas in the Blocking group, A had been paired with the US in a previous phase of treatment (i.e., A–US, then AX–US). Escobar, Oberling, et al. suggested that the comparator hypothesis provided a mechanism to encompass both the latent inhibition and the blocking effect. Specifically, they proposed that one of the cognitive markers of acute schizophrenia might be a difficulty in activating the comparator mechanism. They also suggested that any manipulation (e.g., medication or course of illness) that normalizes the function of the comparator mechanism would restore the response deficit (latent inhibition or blocking) which was previously disrupted by acute schizophrenia.
Summary The multiple theoretical explanations of latent inhibition that have been proposed throughout the years can be grouped into two categories. The more traditional view is that retarded responding due to CS preexposure results from altered processing of the CS due to changes in its salience or associability (e.g., Aitken & Dickinson, 2005; Dickinson & Burke, 1996; Mackintosh, 1975; McLaren & Mackintosh, 2000; Pearce & Hall, 1980; Wagner, 1981). This altered processing of the CS prevents it from developing associations to the US during the subsequent conditioning phase. Thus, the traditional view of latent inhibition is that it reflects an acquisition deficit. However, multiple phenomena suggest that CS preexposure does not prevent the development of associations between the CS and US during conditioning. For example, extinction of the context following either preexposure or conditioning (e.g., Grahame et al., 1994) suggests that the CS–US association was acquired, but was not observable in behavioral tests. These observations led to the development of theoretical explanations of latent inhibition that regarded it not as a failure to acquire associations, but as a failure to express associations that had been previously acquired, that is, as a performance deficit. This deficit can be viewed as resulting from interference between acquired associations (e.g., Bouton, 1991, 1993, 1994; Lubow & Gewirtz, 1995) or comparisons of relative strength of associations at the moment of
Latent inhibition: acquisition or performance deficit?
87
producing a response (e.g., Denniston et al., 2001; Miller & Matzel, 1988; Stout & Miller, 2007). Denniston et al.’s (2001) extended comparator hypothesis provides a performancebased account of latent inhibition by proposing that CS-preexposure results in the development of associations between the CS and the US, as well as between the CS and the preexposure context, and between the preexposure context and the US. Thus, presentation of the CS is assumed to activate two representations of the US, one directly (CS–US association) and one indirectly (through the CS-preexposure context and preexposure context–US associations). A comparison of these two US representations determines conditioned responding, which increases with strength of activation of the US representation through the direct path and decreases with strength of activation of the US through the indirect path. Each of the links in the indirect path is in turn subject to a comparator process, which determines the effectiveness of the link. This associative structure predicts interactions between latent inhibition and other phenomena, such as overshadowing (Blaisdell et al., 1998). Specifically, the model predicts that the two effects would counteract, rather than add to, each other (see Wheeler & Miller, 2008, for a review of counteraction effects). The comparator view of latent inhibition has also been applied to cognitive dysfunction in acute schizophrenia, with the implication that disrupted latent inhibition in this population could be regarded as a difficulty in activating the comparator mechanism. This view of disrupted latent inhibition has the advantage that, with a single mechanism, it can account for disruption of other phenomena (e.g., blocking; see e.g., Escobar, Oberling, et al., 2002). Future development in comparator theory needs to address the phenomena of renewal and reinstatement which, although consistent with the basic assumptions of the model, are not fully explained by it. One recent theoretical development has combined the comparator hypothesis with a priming mechanism in a dual-process model that can account for traditional interference effects and traditional associative learning effects (e.g., Escobar, Arcediano, Platt, & Miller, 2004; Miller & Escobar, 2002; but see Wheeler & Miller, 2007). It appears that CS processing, memory interference, and CS–context associations all contribute to the development of latent inhibition. Thus, a full conceptualization of latent inhibition should describe it as a phenomenon determined by both what is acquired and what is expressed in behavior.
Acknowledgements Manuscript preparation was supported by an Auburn University Research Grant and NIH Grant 81269 to M.E. and NIH Grant 33881 to R.R.M. We thank Jeremie Jozefowiez, Mario Laborda, Bridget McConnell, Cody Polack, James Witnauer, and Gonzalo Miguez for comments on an earlier version of this manuscript. Comments concerning this chapter should be addressed to Martha Escobar, 226 Thach Hall, Department of Psychology, Auburn University, AL 36849–5214; e-mail:
[email protected].
88
Martha Escobar and Ralph R. Miller
References Aguado, L., Symonds, M., & Hall, G. (1994). Interval between preexposure and test determines the magnitude of latent inhibition: implications for an interference account. Animal Learning & Behavior, 22, 188–194. Aitken, M. R. F., & Dickinson, A. (2005). Simulations of a modified SOP model applied to retrospective revaluation of human causal judgments. Learning & Behavior, 33, 147–159. Ayres, J. J. B., Albert, M., & Bombace, J. C. (1987). Extending conditioned stimuli before versus after unconditioned stimuli: implications for real-time models of conditioning. Journal of Experimental Psychology: Animal Behavior Processes, 13, 168–181. Baker, A. G., & Mercier, P. (1982). Extinction of the context and latent inhibition. Learning and Motivation, 13, 391–416. Baker, A. G., Murphy, R., Valle´e-Tourangeau, F., & Mehta, R. (2001). Contingency learning and causal reasoning. In R. R. Mowrer & S. B. Klein (Eds.), Handbook of Contemporary Learning Theories. Mahwah, NJ: Erlbaum, pp. 255–306. Baruch, I., Hemsley, D. R., & Gray, J. A. (1988). Differential performance of acute and chronic schizophrenics in a latent inhibition task. The Journal of Nervous and Mental Disease, 176, 598–606. Bender, S., Muller, B., Oades, R. D., & Sartory, G. (2001). Conditioned blocking and schizophrenia: a replication and study of the role of symptoms, age, onset-age of psychosis and illness duration. Schizophrenia Research, 49, 157–170. Blaisdell, A. P., Bristol, A. S., Gunther, L. M., & Miller, R. R. (1998). Overshadowing and latent inhibition counteract each other: support for the comparator hypothesis. Journal of Experimental Psychology: Animal Behavior Processes, 24, 335–351. Bouton, M. E. (1991). Context and retrieval in extinction and in other examples of interference in simple associative learning. In L. W. Dachowski and C. F. Flaherty (Eds.), Current Topics in Animal Learning: Brain, Emotion, and Cognition. Hillsdale, NJ: Erlbaum, pp. 25–53. Bouton, M. E. (1993). Context, time, and memory retrieval in the interference paradigms of Pavlovian learning. Psychological Bulletin, 114, 80–99. Bouton, M. E. (1994). Context, ambiguity, and classical conditioning. Current Directions in Psychological Science, 3, 49–53. Bouton, M. E. (1997). Signals for whether versus when an event will occur. In M. E. Bouton & M. S. Fanselow (Eds.), Learning, Motivation, and Cognition: The Functional Behaviorism of Robert C. Bolles. Washington DC: American Psychological Association, pp. 385–409. Bouton, M. E., & Swartzentruber, D. (1989). Slow reacquisition following extinction: context, encoding, and retrieval mechanisms. Journal of Experimental Psychology: Animal Behavior Processes, 15, 43–53. Brandon, S. E., Vogel, E. H., & Wagner, A. R. (2003). Stimulus representation in SOP: I. Theoretical rationalization and some interpretations. Behavioural Processes, 62, 5–25. Bush, R. R., & Mosteller, F. (1955). Stochastic Models for Learning. New York: Wiley. Channell, S., & Hall, G. (1983). Contextual effects in latent inhibition with an appetitive conditioning procedure. Animal Learning & Behavior, 11, 67–74.
Latent inhibition: acquisition or performance deficit?
89
Cohen, J. D., & Servan-Schreiber, D. (1992). Context, cortex, and dopamine: a connectionist approach to behavior and biology in schizophrenia. Psychological Review, 99, 45–77. De la Casa, L. G., & Lubow, R. E. (2000). Superlatent inhibition with delayed conditioned taste aversion testing. Animal Learning & Behavior, 28, 389–399. Denniston, J. C., Savastano, H. I., & Miller, R. R. (2001). The extended comparator hypothesis: learning by contiguity, responding by relative strength. In R. R. Mowrer & S. B. Klein (Eds.), Handbook of Contemporary Learning Theories. Hillsdale, NJ: Erlbaum, pp. 65–117. DeVietti, T. L., Bauste, R. L., Nutt, G., et al. (1987). Latent inhibition: a trace conditioning phenomenon? Learning and Motivation, 18, 185–201. Dickinson, A., & Burke, J. (1996). Within-compound associations mediate the retrospective revaluation of causality judgments. Quarterly Journal of Experimental Psychology, 49B, 60–80. Escobar, M., Arcediano, F., & Miller, R. R. (2002). Latent inhibition and contextual associations. Journal of Experimental Psychology: Animal Behavior Processes, 28, 123–136. Escobar, M., Arcediano, F., Platt, T., & Miller, R. R. (2004). Interference and time: a brief review and an integration. Reviews in the Neurosciences, 15, 415–438. Escobar, M., Oberling, P., & Miller, R. R. (2002). Associative deficit accounts of disrupted latent inhibition and blocking in schizophrenia. Neuroscience and Biobehavioral Reviews, 26, 203–216. Feldon, J., & Weiner, I. (1991). An animal model of attention deficit. In A. Boulton, G. Baker, and M. Martin-Iverson (Eds.), Neuromethods, vol. 18: Animal Models in Psychiatry I. Clifton, NJ: Humana Press, pp. 313–361. Friedman, B. X., Blaisdell, A. P., Escobar, M., & Miller, R. R. (1998). Comparator mechanisms and conditioned inhibition: CS preexposure disrupts Pavlovian conditioned inhibition but not explicitly unpaired inhibition. Journal of Experimental Psychology: Animal Behavior Processes, 24, 453–466. Grahame, N. J., Barnet, R. C., Gunther, L. M., & Miller, R. R. (1994). Latent inhibition as a performance deficit resulting from CS-context associations. Animal Learning & Behavior, 22, 395–408. Gray, N. S., Pilowski, L. S., Gray, J. A., & Kerwin, R. W. (1995). Latent inhibition in drug naı¨ ve schizophrenics: relationship to duration of illness and dopamine D2 binding using SPECT. Schizophrenia Research, 17, 95–107. Hall, G., & Minor, H. (1984). A search for context-stimulus associations in latent inhibition. Quarterly Journal of Experimental Psychology, 36, 145–169. Holland, P. C. (1992). Occasion setting in Pavlovian conditioning. In D. L. Medin (Ed.), The Psychology of Learning and Motivation, vol. 28. San Diego: Academic Press, pp. 69–125. Honey, R. C., & Hall, G. (1988). Overshadowing and blocking procedures in latent inhibition. Quarterly Journal of Experimental Psychology, 40B, 163–186. Ishii, K. (1999). Attenuation of latent inhibition after compound conditioning. Japanese Psychological Research, 41, 102–111. Jones, S. H., Gray, J. A., & Hemsley, D. R. (1992). Loss of the Kamin blocking effect in acute but not chronic schizophrenics. Biological Psychiatry, 32, 739–755. Jones, S. H., Hemsley, D., Ball, S., & Serra, A. (1997). Disruption of the Kamin blocking effect in schizophrenia and in normal subjects following amphetamine. Behavioural Brain Research, 88, 103–114.
90
Martha Escobar and Ralph R. Miller
Kamin, J. L. (1968). “Attention-like” processes in classical conditioning. In M. R. Jones (Ed.), Miami Symposium on the Prediction Of Behavior: Aversive Stimulation. Miami, FL: University of Miami Press, pp. 9–31. Kasprow, W. J., Catterson, D., Schachtman, T. R., & Miller, R. R. (1984). Attenuation of latent inhibition by post-acquisition reminder. Quarterly Journal of Experimental Psychology, 36B, 53–63. Kaufman, M. A., & Bolles, R. C. (1981). A nonassociative aspect of overshadowing. Bulletin of the Psychonomic Society, 18, 318–320. Killcross, A. S., Kiernan, M. J., Dwyer, D., & Westbrook, R. F. (1998). Loss of latent inhibition of contextual conditioning following non-reinforced context exposure in rats. Quarterly Journal of Experimental Psychology, 51B, 75–90. Kraemer, P. J., Randall, C. K., & Carbary, T. J. (1991). Release from latent inhibition with delayed testing. Animal Learning & Behavior, 19, 139–145. Kraemer, P. J., & Roberts, W. A. (1984). The influence of flavor preexposure and test interval on conditioned taste aversions in the rat. Learning and Motivation, 15, 259–278. Leumann, L., Feldon, J., Vollenweider, F. X., & Ludewig, K. (2002). Effects of typical and atypical antipsychotics on prepulse inhibition and latent inhibition in chronic schizophrenia. Biological Psychiatry, 52, 729–739. Loy, I., & Hall, G. (2002). Taste aversion after ingestion of lithium chloride: an associative analysis. Quarterly Journal of Experimental Psychology, 55B, 365–380. Lubow, R. E. (1989). Latent Inhibition and Conditioned Attention Theory. Cambridge: Cambridge University Press. Lubow, R. E. (1997). Latent inhibition as a measure of learned inattention: some problems and solutions. Behavioural Brain Research, 88, 75–83. Lubow, R. E. (2005). Construct validity of the animal latent inhibition model of selective attention deficits in schizophrenia. Schizophrenia Bulletin, 31, 139–153. Lubow, R. E., & Gewirtz, J. C. (1995). Latent inhibition in humans: data, theory, and implications for schizophrenia. Psychological Bulletin, 117, 87–103. Lubow, R. E., & Moore, A. U. (1959). Latent inhibition: the effect of nonreinforced preexposure to the conditioned stimulus. Journal of Comparative and Physiological Psychology, 52, 415–419. Lubow, R. E., Schnur, P., & Rifkin, B. (1976). Latent inhibition and conditioned attention theory. Journal of Experimental Psychology: Animal Behavior Processes, 2, 163–174. Lubow, R. E., Wagner, M., & Weiner, I. (1982). The effects of compound stimulus preexposure of two elements differing in salience on the acquisition of conditioned suppression. Animal Learning & Behavior, 10, 483–489. Lubow, R. E., Weiner, I., Schlossberg, A., & Baruch, I. (1987). Latent inhibition and schizophrenia. Bulletin of the Psychonomic Society, 25, 464–467. Lubow, R. E., Weiner, I., & Schnur, P. (1981). Conditioned attention theory. In G. H. Bower (Ed.), The Psychology of Learning and Motivation, vol. 15. New York: Academic Press, pp. 1–49. Mackintosh, N. J. (1975). A theory of attention: variations in the associability of stimuli with reinforcement. Psychological Review, 82, 276–298. Matzel, L. D., Schachtman, T. R., & Miller, R. R. (1985). Recovery of an overshadowed association achieved by extinction of the overshadowed stimulus. Learning and Motivation, 16, 398–412.
Latent inhibition: acquisition or performance deficit?
91
McConnell, B. L., Wheeler, D. S., Urcelay, G. P., & Miller, R. R. (2009). Protection from latent inhibition provided by a conditioned inhibitor. Journal of Experimental Psychology: Animal Behavior Processes, 35, 498–508. McLaren, I. P. L., Bennett, C., Plaisted, K., Aitken, M., & Mackintosh, N. J. (1994). Latent inhibition, context specificity, and context familiarity. Quarterly Journal of Experimental Psychology, 47B, 387–400. McLaren, I. P. L., Kaye, H., & Mackintosh, N. J. (1989). An associative theory of the representation of stimuli: applications to perceptual learning and latent inhibition. In R. G. M. Morris (Ed.), Parallel Distributed Processing: Implications for Psychology and Neurobiology. Oxford, UK: Oxford University Press, pp. 102–130. McLaren, I. P. L., & Mackintosh, N. J. (2000). An elemental model of associative learning. I. Latent inhibition and perceptual learning. Animal Learning & Behavior, 28, 211–246. Miller, R. R., Barnet, R. C., & Grahame, N. J. (1995). Assessment of the RescorlaWagner model. Psychological Bulletin, 117, 363–386. Miller, R. R., & Escobar, M. (2002). Associative interference between cues and between outcomes presented together and presented apart: an integration. Behavioural Processes, 57, 163–185. Miller, R. R., Esposito, J. J., & Grahame, N. J. (1992). Overshadowing-like effects between potential comparator stimuli: covariation in comparator roles of context and punctuate excitor used in inhibitory training as a function of excitor salience. Learning and Motivation, 23, 1–26. Miller, R. R., & Matzel, L. D. (1988). The comparator hypothesis: a response rule for the expression of associations. In G. H. Bower (Ed.), The Psychology of Learning and Motivation, vol. 22. San Diego, CA: Academic Press, pp. 51–92. Miller, R. R., & Oberling, P. (1998). Analogies between occasion setting and Pavlovian conditioning. In N. A. Schmajuk and P. C. Holland (Eds.), Occasion Setting: Associative Learning and Cognition in Animals. Washington DC: American Psychological Association, pp. 3–35. Moran, P. M., Al-Uzri, M. M., Watson, J., & Reveley, M. A. (2003). Reduced Kamin blocking in non paranoid schizophrenia: associations with schizotypy. Journal of Psychiatric Research, 37, 155–163. Nakajima, S., Ka, H., & Imada, H. (1999). Summation of overshadowing and latent inhibition in rats’ conditioned taste aversion: scapegoat technique works for familiar meals. Appetite, 33, 299–307. Nakajima, S., & Nagaishi, T. (2005). Summation of latent inhibition and overshadowing in a generalized bait shyness paradigm of rats. Behavioural Processes, 69, 369–377. Nelson, J. B. (2002). Context specificity of excitation and inhibition in ambiguous stimuli. Learning and Motivation, 33, 284–310. Oades, R. D. (1997). Stimulus dimension shifts in patients with schizophrenia, with and without paranoid hallucinatory symptoms, or obsessive compulsive disorder: strategies, blocking, and monoamine status. Behavioural Brain Research, 88, 115–131. Oades, R. D., Zimmermann, B., & Eggers, C. (1996). Conditioned blocking in patients with paranoid, non-paranoid psychosis or obsessive compulsive disorder: associations with symptoms, personality and monoamine metabolism. Journal of Psychiatric Research, 30, 369–390.
92
Martha Escobar and Ralph R. Miller
Oberling, P., Gosselin, O., & Miller, R. R. (1999). Latent inhibition in animals as a model of acute schizophrenia: a re-analysis. In M. Haug and R. E. Whalen (Eds.), Animal Models of Human Emotion and Cognition. Washington DC: American Psychological Association. Pavlov, I. P. (1927). Conditioned Reflexes. London: Oxford University Press. Pearce, J. M., & Hall, G. (1980). A model for Pavlovian learning: variations in the effectiveness of conditioned but not of unconditioned stimuli. Psychological Review, 87, 532–552. Reed, P. (1995). Compound stimulus preexposure effects in an appetitive conditioning procedure. Learning and Motivation, 26, 1–10. Reiss, S., & Wagner, A. R. (1972). CS habituation produces a “latent inhibition effect” but no active “conditioned inhibition”. Learning and Motivation, 3, 237–245. Rescorla, R. A. (1971). Variation in the effectiveness of reinforcement and nonreinforcement following prior inhibitory conditioning. Learning and Motivation, 2, 113–123. Rescorla, R. A., & Wagner, A. R. (1972). A theory of Pavlovian conditioning: variations in the effectiveness of reinforcement and non-reinforcement. In A. H. Black and W. F. Prokasy (Eds.), Classical Conditioning II: Current Theory and Research. New York: Appleton-Century Crofts, pp. 64–99. Rosas, J. M., & Bouton, M. E. (1997). Additivity of the effects of retention interval and context change on latent inhibition: toward resolution of the context forgetting paradox. Journal of Experimental Psychology: Animal Behavior Processes, 23, 283–294. Sherrington, C. S. (1947). The Integrative Action of the Central Nervous System. Cambridge: Cambridge University Press. Sissons, H. S., & Miller, R. R. (2009). Spontaneous recovery of excitation and inhibition. Journal of Experimental Psychology: Animal Behavior Processes, 35, 419–426. Stout, S. C., & Miller, R. R. (2007). Sometimes-competing retrieval (SOCR): a formalization of the comparator hypothesis. Psychological Review, 114, 759–783. Swerdlow, N. R., Braff, D. L., Hartston, H., Perry, W., & Geyer, M. A. (1996). Latent inhibition in schizophrenia. Schizophrenia Research, 20, 91–103. Swerdlow, N. R., Stephany, N., Wasserman, L. C., et al. (2005). Intact visual latent inhibition in schizophrenia patients in a within-subject paradigm. Schizophrenia Research, 72, 169–183. Tolman, E. C., & Honzik, C. H. (1930). Introduction and removal of reward, and maze performance in rats. University of California Publications in Psychology, 4, 257–275. Urushihara, K., Wheeler, D. S., Pinen˜o, O., & Miller, R. R. (2005). An extended comparator hypothesis account of superconditioning. Journal of Experimental Psychology: Animal Behavior Processes, 31, 184–198. Vaitl, D., Lipp, O., Bauer, U., et al. (2002). Latent inhibition and schizophrenia: Pavlovian conditioning of autonomic responses. Schizophrenia Research, 55, 147–158. Wagner, A. R. (1979). Habituation and memory. In A. Dickinson & R. A. Boakes (Eds.), Mechanisms of Learning and Motivation: A Memorial to Jerzy Konorski. Hillsdale, NJ: Erlbaum, pp. 53–82.
Latent inhibition: acquisition or performance deficit?
93
Wagner, A. R. (1981). SOP: a model of automatic memory processing in animal behavior. In N. E. Spear and R. R. Miller (Eds.), Information Processing in Animals: Memory Mechanisms. Hillsdale, NJ: Erlbaum, pp. 5–47. Wagner, A. R., & Brandon, S. E. (1989). Evolution of a structure connectionist model of Pavlovian conditioning (ÆSOP). In S. B. Klein & R. R. Mowrer (Eds.), Contemporary Learning Theories: Pavlovian Conditioning and the Status of Traditional Learning Theory. Hillsdale, NJ: Erlbaum, pp. 149–189. Wagner, A. R., & Brandon, S. E. (2001). A componential theory of Pavlovian conditioning. In R. R. Mowrer & S. B. Klein (Eds.), Handbook of Contemporary Learning Theories. Mahwah, NJ: Erlbaum, pp. 23–64. Westbrook, R. F., Bond, N. W., & Feyer, A. -M. (1981). Short- and long-term decrements in toxicosis-induced odor-aversion learning: the role of duration of exposure to an odor. Journal of Experimental Psychology: Animal Behavior Processes, 7, 362–381. Wheeler, D. S., Chang, R. C., & Miller, R. R. (2003). Massive preexposure and preexposure in multiple contexts attenuate the context specificity of latent inhibition. Learning & Behavior, 31, 378–386. Wheeler, D. S., & Miller, R. R. (2007). Interactions between retroactive interference and context-mediated treatments that impair Pavlovian conditioned responding. Learning & Behavior, 35, 27–35. Wheeler, D. S., & Miller, R. R. (2008). Determinants of cue interaction. Behavioural Processes, 78, 191–203. Wheeler, D. S., Stout, S. C., & Miller, R. R. (2004). Interaction of retention interval with CS-preexposure and extinction treatments: symmetry with respect to primacy. Learning & Behavior, 32, 335–347. Williams, J. H., Wellman, N. A., Geaney, D. P., Cowen, P. J., Feldon, J., & Rawlins, J. N. P. (1998). Reduced latent inhibition in people with schizophrenia: an effect of psychosis or of its treatment. British Journal of Psychiatry, 172, 243–249. Yogev, H., Sirota, P., Gutman, Y., & Hadar, U. (2004). Latent inhibition and overswitching in schizophrenia. Schizophrenia Bulletin, 30, 713–726.
5 Latent inhibition and learned irrelevance in human contingency learning M.E. Le Pelley and Mia Schmidt-Hansen
Introduction To say that latent inhibition (LI) is a well-established phenomenon of animal conditioning is a serious understatement. Since its first demonstration 50 years ago, the finding of a retardation in conditioning following nonreinforced preexposure to a stimulus has been reported countless times across species from sheep to snails, and from goats to goldfish. Given the ubiquity of LI in animal conditioning, and the suggestion that, under some circumstances at least, a common associative mechanism might also underlie human contingency learning (the acquisition of knowledge about the predictive relationship between a cue and an outcome; Dickinson, Shanks, & Evenden, 1984), we might expect a search of the contingency learning literature to reveal many demonstrations of LI in humans. The conclusions to be drawn from such a search depend largely on whether an empirical or theoretical approach to LI is taken.
An empirical approach to LI If we define LI empirically, as the experimental observation of a retardation in the development of conditioned responding to a stimulus resulting from nonreinforced preexposure to that stimulus, then we must conclude that LI has been demonstrated many times in human contingency learning (e.g. De la Casa, Ruiz, & Lubow, 1993b; Ginton, Urca, & Lubow, 1975; Gray, N.S., et al., 2001; Lubow, Ingberg-Sachs, Zalstein-Orda, & Gewirtz, 1992). For example, during Ginton et al.’s preexposure phase, participants were required to report the number of repetitions of a list of nonsense-syllables (a masking task; see below). The to-be-conditioned stimulus (CS), a burst of noise, was interspersed throughout presentation of this list. In a subsequent test phase participants heard the same recording, but were asked to press a button every time they thought a counter was about to be incremented; these increments were contingent on the noise bursts. Participants who had been preexposed to Latent Inhibition: Cognition, Neuroscience and Applications to Schizophrenia, ed. R.E. Lubow and Ina Weiner. Published by Cambridge University Press # Cambridge University Press 2010.
94
Latent inhibition and learned irrelevance in human contingency learning
95
the noise CS were slower to learn to press the button when the CS occurred than were nonpreexposed participants. Hence, following an empirical definition, this experiment, and the others mentioned above, clearly demonstrates LI.
A theoretical approach to LI This empirical definition, however, tells us nothing of the mechanisms underlying LI. While an arch-behaviourist would be happy with the parallel between animal and human LI effects as defined empirically (after all, the outcome – a retardation in subsequent learning – is the same in both cases), the more interesting question is why these effects occur. Extending this idea, we might ask whether the process underlying the empirical LI effect is the same in animals and humans, an assumption that underpins the use of animal LI preparations to provide a model of human mental disorders that involve a disruption of LI, such as schizophrenia (e.g. Lubow, 2005). These issues provide the focus of this chapter. While, as discussed later, there are parallels between the empirical LI effect observed in humans and animals, there is also one important difference. Animal studies reveal that “simple” preexposure to a stimulus is sufficient to generate LI. For example, Kaye and Pearce (1987) gave rats 72 presentations of a light in a conditioning chamber. These rats were slower to learn a conditioned response to the light when it was subsequently paired with food than were rats who had spent an equivalent duration in the conditioning chamber prior to conditioning but without preexposure to the light. In contrast, simple preexposure with adult humans generally does not produce LI (e.g. Hulstijn, 1978; Perlmuter, 1966). Instead, LI is typically observed in humans only if preexposure is accompanied by a masking task, for example the syllable-counting task used by Ginton et al. (1975). The standard analysis (discussed in more detail later) assumes that this syllable-counting task directs participants’ attention away from the noise, which is irrelevant to this task. Hence preexposure to the noise is said to be masked. The requirement for a masking task represents an important distinction between the conditions required to generate LI in humans and animals. In a review, Lubow and Gewirtz (1995) concluded that “in the adult human, with a few exceptions, LI can be demonstrated only if a masking task is used during the stimulus preexposure stage” (p. 100). Graham and McLaren (1998) subsequently reviewed these “few exceptions” and argued that these studies were either methodologically flawed, or likely to be measuring some effect other than LI (such as habituation of an unconditioned response to the CS). Thus the analysis of the literature, up to 1998 at least, is consistent with the suggestion that the use of a masking task is an absolute requirement for LI to be observed in human contingency learning. To the best of our knowledge, three published studies have since purported to demonstrate LI in human contingency learning in the absence of a masking task.
96
M.E. Le Pelley and Mia Schmidt-Hansen
These studies are discussed below, and we shall see that the most likely mechanism underlying the LI effect observed in each (computation of conditional probabilities) is rather similar to certain accounts previously developed to explain LI in animals (Baker & Mercier, 1982; Testa & Ternes, 1977).
Escobar, Arcediano, & Miller (2003) Escobar et al. argued that one reason for previous failures to find simple (i.e. unmasked) LI in humans is that human studies are typically heavily segmented. Participants receive instructions between preexposure and conditioning phases, which might encourage the belief that they are distinct tasks. Likewise the nature of preexposure and conditioning tasks in human LI studies is sometimes rather different from that in animal studies. Given that LI relies on transfer of participants’ experience from preexposure to training, Escobar et al. argued that such influences would decrease the likelihood of transfer and hence the likelihood of observing LI. Escobar et al. therefore used a task in which preexposure and conditioning occurred without interruption. In Experiment 1, participants experienced ten presentations of each of cues X and B (coloured letters). Following this preexposure, “conditioning” comprised one X–outcome pairing, one A–outcome pairing, and one B-alone presentation, where the outcome was a white cross. Finally, participants were asked to rate the extent to which each cue predicted the appearance of the outcome; these ratings are shown in Figure 5.1a. The mean rating for X was significantly lower than for A, despite both being paired with the outcome once, and was similar to the rating for B, which had never been paired with the outcome. Escobar et al. argued that this provided the “first fully controlled demonstration of retarded acquisition of a cue–outcome association following unmasked preexposure in human adults” (p. 1031) in contingency learning. We would disagree, however, because the rating test, located at the end of training, would presumably encourage participants to integrate their experiences with a particular cue when deciding how strongly it predicts the outcome. At the point of test, cue X had been presented 11 times; ten times alone, and once with the outcome. It seems reasonable to suggest that participants might base their rating on all of their experience with X. Summed over this experience, the conditional probability of the outcome given the presence of X is 1/11 ¼ 0.091, indicating that X predicted the outcome only very weakly (and similarly to B, which was never paired with the outcome, giving probability 0/11 ¼ 0). In contrast, A was experienced once, and was paired with the outcome, hence the conditional probability of the outcome given A is 1/1 ¼ 1. It is therefore unsurprising that participants rated X lower than A: the only circumstances under which this would not occur would be if their ratings were based solely on the single most recent experience of each cue. Figure 5.1a shows conditional probabilities of the outcome given each cue (labelled
Latent inhibition and learned irrelevance in human contingency learning (a)
(b) 100
0.5
Scaled probability
60 40 20
Suppression ratio
Empirical data
80 Rating
97
0.4
Group Red (empirical) Group Context (empirical) Group Red (probability) Group Context (probability)
0.3 0.2 0.1 0
0 X
A Test Stimulus
B
1 2 3 4 5 6 7 8 9 10 Trial
Figure 5.1. (a) Causal judgment ratings for Experiment 1 of Escobar, Arcediano and Miller (2003). Black bars show empirical data, redrawn from Escobar et al. Grey bars show conditional probability of the outcome given the presence of each cue, scaled to match the causal judgment ratings received by cues A and B. (b) Suppression ratios to the red sensor across the eight test trials of Nelson and Sanjuan’s (2006) Experiment 1. Filled symbols show empirical data, redrawn from Nelson and Sanjuan. Open symbols show scores derived from the conditional probability of the outcome given the presence of each cue, scaled as described in the main text. Group Red received preexposure to the red sensor; Group Context did not experience the sensor prior to conditioning.
“Scaled probability”), converted into ratings using a linear scaling that matches a probability of 1 to the empirical rating received by A, and a probability of 0 to the rating received by B.1 The match to the mean rating of X, which was not otherwise fit to the empirical data, is near-perfect. On this analysis, Escobar et al.’s findings reveal nothing about the rate of acquisition of the cue–outcome relationship. Participants may have learnt perfectly that, on its final trial considered in isolation, X predicted the outcome. But their overall perception of the predictiveness of X is bound to be influenced by their previous experiences of nonreinforced presentations of this cue. Hence the “LI” observed merely provides another demonstration that humans can integrate information across multiple trials to make reasoned judgments regarding the predictive status of cues (cf. Allan & Jenkins, 1980). In Experiments 3 and 4 Escobar et al. demonstrated that if preexposure and training occurred in different contexts, or if a break filled with redundant instructions interposed between preexposure and training, the rating advantage for A over X (the putative LI effect) was reduced, mirroring the context-sensitivity of LI in animals (discussed later). This follows naturally from Escobar et al.’s segmentation argument: such manipulations 1
The fact that these probabilities do not correspond to the extremes of the rating scale, 0 and 100, presumably reflects participants’ unwillingness to use these extreme values on the basis of limited experience with each of the cues.
98
M.E. Le Pelley and Mia Schmidt-Hansen
will encourage participants to segment the two phases, making them less likely to integrate information from the preexposure phase into their final conditional probability computations, resulting in similar ratings for A and X.
Nelson and Sanjuan (2006) Nelson and Sanjuan used a video game task in which participants earned points for firing torpedoes at an enemy spaceship by clicking a button; the more torpedoes they fired, the more points they earned. However, the enemy could also fire torpedoes which would freeze their controls for a certain period and prevent them from earning points. Participants were told that the higher their rate of firing immediately before the enemy’s attack, the longer their controls would be frozen. They were also told that their spaceship was equipped with “sensors” of various colours, and that “Particular colours or combinations might mean anything. Keep in mind also that they may mean nothing at all!”. Further, they were informed that “learning what sensors mean will help you avoid frozen controls”, implying that certain sensors might signal an imminent attack by the enemy, allowing participants to suppress clicking pre-emptively and thus avoid having their controls frozen. For participants in Group Red of Experiment 1, during a preexposure phase a red sensor (the CS) was activated ten times without consequence: the enemy did not attack during this phase. Given that the CS was task-relevant, this can reasonably be classified as an example of unmasked preexposure. For participants in Group Context, no sensors were activated during this preexposure phase. In the subsequent conditioning phase (which followed without apparent transition), both groups experienced ten trials on which the red sensor was activated and was paired with an attack from the enemy (the outcome). The rate at which participants learnt to suppress clicking during the CS was plotted as a suppression ratio, calculated using (clicks during CS)/(clicks during pre-CS period þ clicks during CS). Results are shown in Figure 5.1b. Participants in Group Context learnt to suppress clicking during the CS significantly faster than those in Group Red, demonstrating LI as a result of CS preexposure in Group Red. As for the study of Escobar et al. (2003), these results are to be expected if participants integrated their previous experience with the CS to generate an overall perception of the predictive status of that CS on the current trial. Figure 5.1b shows predictions from such an account based on the conditional probability of the outcome given the presence of the CS on each trial. For example, following the first conditioning trial in Group Red, this probability would be 1/11 ¼ 0.091; following the second conditioning trial it would be 2/12 ¼ 0.17; and so on. Since these probabilities are inversely related to suppression ratios (a high probability – indicating that the outcome is likely to occur – will produce a low suppression ratio), we then converted each to (1 – probability). Finally, these inverted probabilities were
Latent inhibition and learned irrelevance in human contingency learning
99
scaled using a general linear scaling, such that (i) the mean score for Group Red matched the empirically observed mean suppression ratio for Group Red, and (ii) the mean score over blocks 3–10 for Group Context matched the empirically observed mean suppression ratio over the same blocks for Group Context (these scores provide a “floor” for suppression). Figure 5.1b shows that these scaled probabilities provide an excellent match to Nelson and Sanjuan’s data. This analysis reduces to the idea that, during preexposure, participants in Group Red are explicitly trained that the red sensor does not predict attack and hence they should maintain a high rate of clicking when it appears: if they do not, then they will effectively lose points. It does not seem unreasonable to suggest that this trained tendency to maintain responding during the CS will persist into the training phase: participants are unlikely to immediately and entirely alter their behaviour on the basis of a single CS–outcome pairing at the start of the training phase. In Experiment 2, Nelson and Sanjuan demonstrated a trend towards contextspecificity of their stimulus preexposure effect, which can be explained in similar terms to Escobar et al.’s study: segmentation of preexposure and training decreases the likelihood that information is integrated over these phases. Experiment 3 indicated that nonreinforced preexposure did not endow the preexposed CS with the properties of a conditioned inhibitor (see Rescorla, 1971). This again follows naturally from the probability-integration account. During preexposure participants merely learn that the red sensor does not predict attack, not that it actively prevents attack, or counteracts the predictive validity of other cues. Hence there is no reason to suppose that it would take on inhibitory properties. For example, if P(attack given red sensor) is 0, and P(attack given green sensor) is 1, a conservative participant might well assume that P(attack given red sensor AND green sensor) is 1.
Evans, Gray, and Snowden (2007) Participants in Evans et al.’s study were instructed to press a key when the letter X appeared, or was about to appear, in a sequence of letters presented individually. During a preexposure phase, a target letter appeared 20 times, intermixed with a set of four non-target letters; the X never appeared. In a subsequent test phase (which followed with no apparent transition), every time the target letter appeared it was followed by the X, as was a nonpreexposed control letter. Participants’ anticipatory responses indicated that they were slower to learn to predict the X on the basis of the preexposed target letter than the nonpreexposed control letter during this test phase. The most likely interpretation of this finding is similar to that for Nelson and Sanjuan’s (2006) study. Since participants are informed from the outset that their job is to predict the X before it appears, the preexposure phase amounts to explicit training that none of the preexposed letters (including the target)
100
M.E. Le Pelley and Mia Schmidt-Hansen
predicts the X. It will then take participants time to overcome the effect of this prior training when one of these preexposed letters is paired with the X during the test phase. Calculated over the whole experiment, the conditional probability of the X following the preexposed target letter was 10/30 ¼ 0.33; the corresponding probability for the control letter was 10/10 ¼ 1. The ratio of these probabilities (0.33:1) is very similar to the ratio of correct anticipatory responses made during the preexposed letter as compared to the control letter in the test phase (2:5).
Simple preexposure in humans versus animals: a summary None of the above detracts from the empirical importance of these studies. In terms of the empirical definition of LI, Nelson and Sanjuan’s (2006) and Evans et al.’s (2007) results clearly demonstrate this effect: there is a retardation in the development of a conditioned response to a cue resulting from preexposure to that cue. (Arguably Escobar et al.’s study falls short of this definition, since with only a single conditioning trial there is no way to track the acquisition of a conditioned response.) The analysis presented above suggests that the results of these studies might reflect a simple form of probabilistic reasoning similar to that observed in many previous studies of human causal learning. Beyond establishing the likely mechanisms underlying LI effects in human learning, our focus in this chapter is on whether the same mechanisms underlie LI effects in animals. We have argued that a likely interpretation of the simple LI effect observed in these human studies relies on computation of the conditional probability of the outcome given the CS, integrated over experience with that CS. This idea is rather similar to certain accounts that have previously been advanced to explain LI in animals (e.g. Baker & Mercier, 1982; Testa & Ternes, 1977); for example Baker and Mercier argued that “When animals are conditioned following exposure to the CS, their response is determined not only by the conditioning trials but, also, by some sample of their previous experience with the CS” (p. 413). It is, in fact, difficult to rule out such an account of LI in animals on the basis of existing empirical evidence. We remain wary, however, of attributing relatively complex and abstract cognitive reasoning abilities – namely the ability to integrate appropriate information over several sessions of training and to compute conditional probabilities on the basis of that integrated information – to the “humble rat”. In particular, the procedures of the human studies seem to encourage the use of probability integration in a way that procedures of animal studies do not. Nelson and Sanjuan (2006) and Evans et al. (2007) informed participants at the outset of preexposure about the possible occurrence of the outcome (alien attack or the letter X), which would encourage them to encode the absence of this outcome during preexposure as a meaningful event. While Escobar et al. (2003) did not mention the outcome until test, their experiment was so simple, and so short, that participants would presumably retrospectively recode the outcome as having been absent during preexposure. Moreover, the test question strongly implied that they should assess the cue–outcome
Latent inhibition and learned irrelevance in human contingency learning
101
contingency as considered over the whole experiment. Animals, on the other hand, do not know that they are participating in an experiment, do not know what the unconditioned stimulus (US) is going to be until the conditioning phase begins, may experience preexposure over several sessions spanning a number of days, and are not “requested” to consider all of their experience with a cue when responding to it. We would argue that these factors make the use of probability integration less likely in the animal case than the human case. Nevertheless (and with this reservation in mind) we must acknowledge that, to the extent that this probability-integration view of LI in animals is correct, the studies of simple LI in humans discussed above may be tapping similar processes.
LI following masked preexposure While studies claiming to demonstrate LI in humans using simple preexposure are rare, demonstrations using masked preexposure are legion. Moreover, it is these masked studies that are typically compared to simple LI preparations in animals to derive “animal models” of human cognition (Lubow, 2005). But why does a masking task make it so much easier to observe LI in humans, when it is not required in animals?
The masking task distracts attention from the CS One possibility relates to the demand characteristic of a typical LI paradigm. A human participant is aware of participating in an experiment and is likely to consider all stimuli of importance and therefore maintain attention to them, ensuring that they remain ready to learn about these stimuli in the future, hence minimizing the possibility of LI. Such a demand characteristic would not apply to animals. On this view, LI is presumed to be a product of automatic processing (Lubow, 1989) and the function of the masking task is to divert controlled attention away from the preexposed stimulus, such that it will be processed automatically. Hence the masking task is seen as functionally “reducing” humans to infrahumans. That is, it engages controlled processes (which, according to Lubow, 1989, separate adult humans from animals), leaving automatic processes (which are shared with animals) to generate LI to the preexposed stimulus in the same way as they do for animals. This approach therefore proposes that the psychological mechanisms that give rise to LI following simple preexposure in animals are the same as those giving rise to LI following masked preexposure in adult humans.
The masking task generates learned irrelevance of the CS Alternatively, the masking task might fundamentally alter the nature of the observed preexposure effect, and of the psychological processes underlying it. This view suggests that it is the presence of the masking task itself that produces the decrement
102
M.E. Le Pelley and Mia Schmidt-Hansen
in learning about the preexposed stimulus, because it creates a situation in which learned irrelevance could lead to a reduction in processing of the preexposed cue. Learned irrelevance refers to the finding that experience of a cue as irrelevant to the occurrence of an outcome, or to the solution of a discrimination, produces a subsequent retardation in the rate of new learning about that cue. For example, Mackintosh (1973) demonstrated that rats given uncorrelated exposure to a tone and water were slower to subsequently learn about a contingent tone–water relationship than were rats given no preexposure to either tone or water. Uncorrelated CS/US exposure also produces a retardation in subsequent inhibitory conditioning of the CS (Baker & Mackintosh, 1979; Bennett, Wills, Oakeshott, & Mackintosh, 2000). Thus past experience of the tone’s irrelevance with respect to water seems to result in the tone becoming less ready to enter into association (either excitatory or inhibitory) with the water. That is, there has been a change in the associability of the tone. In that it involves slower learning about a cue following preexposure to that cue, learned irrelevance is similar to LI. However, the procedure used for generating the two effects is different: LI involves preexposure to the cue in the absence of the US, while learned irrelevance involves uncorrelated presentation of the CS and US. And crucially, evidence suggests that this uncorrelated preexposure produces a deficit in learning over and above that which would be expected on the basis of the LI resulting from preexposure to the CS, or habituation resulting from preexposure to the US (Baker & Mackintosh, 1979; Bennett et al., 2000; Matzel, Schachtman & Miller, 1988), although this idea is not universally accepted (see Bonardi & Ong, 2003). Stronger evidence comes from studies of discrimination learning, where it has often been observed that after animals have learnt a particular discrimination (D1), they are faster to learn a novel discrimination (D2) if the stimulus dimension (e.g. colour) that is relevant to the solution of D2 was previously relevant to D1, than if the dimension that is relevant to D2 was previously irrelevant to D1 (e.g. George & Pearce, 1999; Mackintosh & Little, 1969; Oswald et al., 2001; Schwartz, Schwartz, & Teas, 1971). This advantage for intradimensional shifts (IDS) over extradimensional shifts (EDS) again implies that experience of a stimulus (or, in this case, a stimulus dimension) as irrelevant results in a subsequent retardation in new learning about that stimulus. And once again the IDS/EDS effect cannot be explained in terms of LI resulting from mere preexposure to the stimuli, as all are preexposed the same number of times. Instead, it is their non-correlation with the outcome (the solution of the discrimination) that is important. Moreover, learned irrelevance has been demonstrated in human contingency learning. Table 5.1 shows the design used by Le Pelley and McLaren (2003). Participants, playing the role of an allergist, examined the effects of various foods on fictitious patients. During training, they were told the contents of meals, and learnt the type of allergic reaction (e.g. nausea) that a patient suffered after eating them. Letters A–Y refer to different cues and o1–o4 to different outcomes, e.g. “AV–o1” indicates that cues A and V were presented together, and were paired with outcome o1.
Latent inhibition and learned irrelevance in human contingency learning
103
Table 5.1. Design of Le Pelley and McLaren (2003) Stage 1
Stage 2
Test
AV – o1 AW – o1 BV – o2 BW – o2 CX – o2 CY – o2 DX – o1 DY – o1
AX – o3 BY – o4 CV – o3 DW – o4
AC BD VX WY
In Stage 1, participants learnt about the effects of foods on “Mr. X”. During this stage, cues A–D were relevant, as they predicted the outcome on each trial (A and D were consistently paired with o1; B and C with o2). Cues V–Y were irrelevant, being followed equally often by o1 and o2. At the outset of Stage 2, participants were told that they would now learn about a new patient, Mr. Y, who suffered from different reactions to Mr X. During this stage, compounds of a previously relevant and previously irrelevant cue were paired with novel outcomes o3 and o4. If learned irrelevance led to a decrement in the rate of learning about cues experienced as irrelevant during Stage 1 (relative to those experienced as relevant), then participants should learn less about the relationships between cues V–Y and the Stage 2 outcomes (o3 and o4) than about cues A–D and those same outcomes. Consistent with this idea, in a final test, participants rated compounds AC and BD as significantly more predictive of outcomes o3 and o4 respectively than were compounds VX and WY. This difference in learning cannot be a consequence of LI generated by preexposure to the cues, because all cues were presented an equal number of times during Stage 1, equating exposure (and thus LI) to all. Similarly, the conditional probabilities of the Stage 2 outcomes were equivalent for previously relevant and previously irrelevant cues: the probability of outcome o3 given cue A, for example, is identical to the probability of o3 given cue X. Instead, it is the cues’ experienced relationship to the outcome that is critical: cues previously experienced as irrelevant are learnt about more slowly than those experienced as relevant (for related human studies, see, e.g., Bonardi, Graham, Hall, & Mitchell, 2005; Kruschke, 1996). It is not hard to see how these ideas might apply to LI procedures that use masked preexposure. During this preexposure, the CS is irrelevant to the solution of the masking task. Consequently learned irrelevance would lead to a decrease in the attention to, and hence the associability of, the preexposed CS, resulting in slower learning about that CS during the subsequent training phase (see Graham & McLaren, 1998, for a related interpretation of the role of the masking task). Given that we know that learned irrelevance does occur in human contingency learning,
104
M.E. Le Pelley and Mia Schmidt-Hansen
and that inclusion of a masking task creates exactly the circumstances in which learned irrelevance might be expected, it would seem a prime candidate for explaining the retardation in learning observed after masked preexposure. The “demand characteristic” view of the masking task outlined earlier proposes that the masking task plays a relatively passive role, taking controlled processes out of the picture and thus exposing the CS to different, automatic processes that produce LI (and that are shared with animals). In contrast, this latter learned irrelevance view instead suggests that the masking task plays a more active role; it is instrumental in producing the LI effect that is observed, because it is the masking task that causes learned irrelevance to the preexposed CS (because this CS is irrelevant with respect to the responses required by the masking task).
Theories of LI and learned irrelevance LI is a theoretically ambiguous phenomenon, open to many different accounts, for example in terms of attention, biased retrieval, or conditional probability (see Hall, 1991). In contrast, learned irrelevance is arguably less ambiguous. The most widely accepted view of learned irrelevance holds that it reflects a reduction in the rate of learning about a cue as a result of experience of that cue’s irrelevance with respect to an outcome. This reduction in learning rate is often described as reflecting a decrease in attention to the cue; associative models incorporating this idea have been proposed by Mackintosh (1975) and Kruschke (2001), among others, and experimental evidence supports this approach (Livesey, Harris, & Harris, 2009). According to such accounts then, the masking task itself causes attentional suppression of the preexposed CS. That said, alternative, non-attentional accounts of learned irrelevance do exist (e.g., Oswald et al., 2001). But even these accounts do not equate learned irrelevance with LI: different processes are proposed to underlie the two empirically distinct phenomena. Thus it is fair to say that, whatever the processes are that give rise to LI and learned irrelevance, theoretical accounts agree that they are not the same (although they may be similar, as discussed later).
Similarities between simple and masked LI Proponents of a common mechanism underlying simple LI in animals and masked LI in humans point to findings indicating that these effects show similar sensitivities to certain experimental manipulations (Lubow, 2005; Lubow & Gewirtz, 1995). For example, De la Casa, Ruiz, and Lubow (1993a, 1993b) demonstrated that extending the duration of preexposure increases the magnitude of LI in both humans and animals (but see Lipp, 1999). This is not terribly surprising. LI is defined as a retardation in learning in a preexposed condition (i.e. with CS preexposure duration
Latent inhibition and learned irrelevance in human contingency learning
105
greater than zero) as compared to a nonpreexposed condition (with preexposure duration of zero). So we know by definition that the amount of retardation is influenced by the amount of preexposure, and it would seem strange if this pattern did not generalize beyond this limiting case. Consequently this parallel between simple and masked LI is not persuasive evidence for a common mechanism: almost every mechanism that one might imagine that could give rise to either effect (including learned irrelevance) would predict a similar pattern. Another parallel that has been noted between simple LI in animals and masked LI in humans is that both show a degree of stimulus-specificity (Lubow & Gewirtz, 1995). Lubow (1973) noted that simple LI in animals is stimulus-specific, in that no retardation is observed to distortions of the preexposed stimulus. Similarly, in a human study Lubow, Caspy, and Schnur (1982) found that masked preexposure to a set of circular stimuli did not slow learning of a discrimination between square shapes that were very different from the preexposed circles (but see Graham & McLaren, 1998, and Lipp, 1999, for studies in which LI does generalize to slight distortions of the preexposed stimulus). This parallel of stimulus-specificity in animals and humans is again unsurprising. Considering the limiting case, if the test stimulus is so dissimilar to the preexposed stimulus that the two share absolutely no features in common, then of course we should not expect to observe an LI effect. Given this limiting case, it is clear that as the test stimulus becomes less similar to the preexposed stimulus, at some point we will see a reduction in the magnitude of LI: that is, LI will show a generalization gradient (Siegel, 1969). But once again, almost any theory that can explain LI would predict the existence of such a gradient (although the steepness of this gradient may vary between theories). For example, it is clear that learned irrelevance also demonstrates a degree of stimulus-specificity. If this were not the case then we could never observe within-subjects effects of learned irrelevance (Le Pelley & McLaren, 2003), or an IDS/EDS effect. Thus, if the change between the preexposed and test stimulus is sufficiently large, then almost any explanation of LI would predict a reduction in the size of the masked LI effect observed. As a further parallel, it is well-known that LI in animals is context-sensitive (Channell & Hall, 1983; Lovibond, Preston, & Mackintosh, 1984), and it has been argued that masked LI in humans is also context-sensitive (Gray, N.S., et al., 2001; Zalstein-Orda & Lubow, 1995). This sensitivity should not be overstated, however. The nature of the task faced by participants during preexposure and test is typically quite different in studies involving masked preexposure (see the description of Ginton et al.’s (1975) study above, for example): the effect in humans is clearly not that sensitive to context. Similarly, studies of learned irrelevance in humans demonstrate that it too will generalize across small changes in context. For example, in Le Pelley and McLaren’s (2003) study, discussed earlier, learned irrelevance was established with respect to outcomes o1 and o2 (in Mr X), and this effect transferred to influence learning with respect to outcomes o3 and o4 (in Mr Y) in Stage 2.
106
M.E. Le Pelley and Mia Schmidt-Hansen
Those studies of masked LI in humans that have demonstrated context-sensitivity have involved a very obvious and marked change in context. In Zalstein-Orda and Lubow’s (1995) study, the screen background changed between preexposure and test (being covered in either diagonal red lines or vertical green lines). Context-sensitivity of LI was observed only when participants were given extensive “pre-preexposure” to this context in the absence of the CS, which probably served to make the subsequent context change particularly salient.2 N.S. Gray et al.’s (2001) context change involved preexposure and test taking place in different rooms, using different computers. However, as for the preexposure duration and stimulus-specificity effects outlined above, the mere demonstration that simple LI in animals and masked LI in humans can both show context-sensitivity under certain conditions does not require that a common mechanism underlies both effects, since the reason for this contextspecificity may be different in the two cases. As just one possibility, few would deny that human learning and performance can be influenced by “higher-level” beliefs that participants might develop regarding the experiment in which they are participating. For example, unpublished work has demonstrated that explicitly informing human participants that their prior (Stage 1) experience is irrelevant to what they are about to learn in Stage 2 is sufficient to eradicate a learned irrelevance effect that would otherwise occur, presumably because this instruction causes participants to strategically “reset” their attention to all cues (Le Pelley et al., internal report, details available from first author on request; see also Lovibond, 2003). A sufficiently salient change in context could presumably have a similar effect, causing participants to strategically discount their prior experience as being irrelevant to the new context, resulting in a decrease in LI. For example, a change of rooms and computers (as used by N.S. Gray et al., 2001) might lead participants to believe that they are now participating in a new experiment that is distinct from the preexposure phase, and to therefore believe that their prior experience should be disregarded. Each of these three parallels (exposure duration, stimulus-specificity, and contextsensitivity) considered individually does not provide compelling evidence for a common mechanism underlying simple LI in animals and masked LI in humans. We would argue that the generality of each of these parallels – the former two in particular would be predicted by essentially any sensible theory of LI – means that even in combination they do not present a strong case for a common mechanism. If you were to be told that objects A and B are both man-made, weigh less than a ton, and have one straight edge, you would be unlikely to conclude that A and B must be the same object. Finally there are claimed parallels regarding the pharmacological sensitivity of LI. Amphetamine treatment reduces simple LI in rats (Weiner, Lubow, & Feldon, 1988), 2
N.S. Gray et al. (1995) noted that the screen background context used in this study raises the possibility that the context may have affected the perceptual nature of the preexposed stimulus via simultaneous contrast or similitude effects. Hence the “context-sensitivity” observed by Zalstein-Orda and Lubow may instead reflect stimulus-specificity. Also, this study did not use a nonpreexposed control condition and we therefore cannot be certain that LI was significantly affected by the context change.
Latent inhibition and learned irrelevance in human contingency learning
107
and masked LI in humans (Gray, N.S., Pickering, Hemsley, et al., 1992). Haloperidol enhances LI in animals (Weiner & Feldon, 1987), and similar effects have been reported regarding masked LI in humans (Williams et al., 1997). Crucially, however, animal studies have indicated that these drugs have their impact during conditioning, rather than during preexposure (Killcross, Dickinson, & Robbins, 1994; Young, Moran, & Joseph, 2005). For example, Killcross et al.’s results indicate that amphetamine influences LI only indirectly, by changing the functional effectiveness of the US during conditioning. As such its effects would not be specific to LI preparations and hence are uninformative as to the processes underlying LI. As an analogy, imagine a drug which causes participants to forget all of their previous experience with a cue. This drug, administered between preexposure and conditioning, would certainly eliminate LI in both humans and animals, and yet this parallel would reveal nothing about whether the mechanism underlying the standard LI effect was the same in the two cases.
Animal models of human cognition Given that we know that learned irrelevance influences human learning, and that masked preexposure creates exactly the environment in which learned irrelevance can operate, it seems plausible that demonstrations of masked LI reflect learned irrelevance. This casts doubt on the idea that masked LI in humans is equivalent to simple LI in animals, implying instead that different psychological mechanisms might be responsible for the two effects. If correct, this could undermine the use of animal LI preparations as models of human mental disorders that are associated with a reduction in masked LI. Notably, masked preexposure paradigms have revealed reduced LI in individuals with schizophrenia (e.g. Baruch, Hemsley, & Gray, 1988) and in healthy individuals with high levels of schizotypal characteristics (Gray, N.S., Fernandez, Williams, et al., 2002; Lubow et al., 1992). This reduction in LI has featured strongly in theorizing about the aetiology of schizophrenia (e.g. Gray, J.A., 1998; Lubow et al., 1992) and has encouraged the view that simple LI preparations in animals can be directly related to the neuropsychology of schizophrenia. Thus manipulations that reduce LI in animals have often been claimed to provide an “animal model” of schizophrenia (e.g. Lubow, 2005). The validity of such models depends on the assumption that LI following simple preexposure in animals and masked preexposure in humans is governed by the same psychological mechanism. We have argued instead that LI following masked preexposure in humans reflects learned irrelevance made possible by inclusion of the masking task, suggesting that a reduction in LI associated with schizophrenia (or high schizotypy) indicates that learned irrelevance is reduced in these individuals (see also Gray, N.S., & Snowden, 2005). We have recently provided evidence consistent with this idea
108
M.E. Le Pelley and Mia Schmidt-Hansen AC/BD
Difference Score
8
VX/WY
6
4
2
0 Low
High UnEx group
Figure 5.2. Mean difference scores ( SEM) for compounds composed of previously relevant cues (AC/BD) and compounds composed of previously irrelevant cues (VX/WY), from Le Pelley et al. (submitted). Groups were defined by a tertile split on the basis of their scores of the Unusual Experiences (UnEx) subscale of the Oxford–Liverpool Inventory of Feelings and Experiences: participants with UnEx scores higher than the upper tertile were assigned to the High group; participants with scores lower than the lower tertile were assigned to the Low group.
(Le Pelley, Schmidt-Hansen, Harris, et al., in press). This study used Le Pelley and McLaren’s (2003) learned irrelevance paradigm, shown in Table 5.1, to probe the relationship between schizotypy and learned irrelevance and found reduced learned irrelevance in participants displaying high levels of unusual experiences (UnEx; a subscale of the Oxford-Liverpool Inventory of Feelings and Experiences (Mason, Claridge, & Jackson, 1995) measuring traits that relate to the “positive” symptom cluster of schizophrenia). Figure 5.2 shows the result of a between-participant split on the basis of UnEx scores. The dependent variable is a difference score derived from ratings given to the test compounds, with higher scores indicating better learning during Stage 2. These scores were averaged over test compounds comprising previously predictive cues (AC and BD), and those comprising previously nonpredictive cues (VX and WY). For participants with low UnEx scores, a standard learned irrelevance effect was observed, with better learning about previously predictive cues than previously nonpredictive cues. In contrast, participants with high UnEx scores showed no such difference, and the magnitude of the learned irrelevance effect was significantly greater in the low UnEx group than in the high group. These data provide further indirect support for the idea that masked LI in humans reflects learned irrelevance. A parsimonious approach would suggest that the reason that the masked LI shows a similar sensitivity to schizotypy as does learned irrelevance is because the former is generated by learned irrelevance.
Latent inhibition and learned irrelevance in human contingency learning
109
LI and learned irrelevance: different but similar? To say that learned irrelevance and simple LI are different is not to say that they are necessarily unrelated. While other accounts are certainly possible, it has often been argued that simple LI is generated by an attention-like mechanism, resulting from a reduction in the processing of the stimulus during nonreinforced preexposure (Lubow, 1989; Pearce & Hall, 1980; Wagner, 1981). And as noted earlier, learned irrelevance is typically viewed as reflecting a reduction in the processing of the stimulus (in terms of a change in attention, or associability) as a result of irrelevance pretraining. Hence, to the extent that the attentional account of simple LI is correct (and there are many who would argue that it is not, e.g. Bouton, 1993; Miller & Matzel, 1988), the processes that give rise to these two different phenomena might be related. If that is the case, then there is still potential value in animal LI preparations as models for disorders showing an association with masked LI in humans. Conclusion We have argued that there is a difference in the processes underlying the LI effect observed following simple exposure in animals and that following masked preexposure in human contingency learning. Specifically, this latter effect may reflect learned irrelevance made possible by inclusion of the masking task during preexposure. And crucially, empirical studies demonstrate that the retardation in learning observed following learned irrelevance training is not the same as that observed following simple preexposure (i.e. there is more to learned irrelevance than just CS-preexposure), indicating that the psychological processes underlying these two effects are distinct (although they may perhaps be related). The question then becomes: what evidence would we require in order to accept the idea of a common mechanism underlying animal and human LI effects? This is not an easy question to answer because simple LI in animals is theoretically ambiguous; despite its empirical simplicity, there is still no general agreement on an underlying process, and it may well be multiply determined (Hall, 1991). We noted above that, to the extent that a particular theoretical interpretation of LI in animals (probability-integration) is correct, existing studies of simple LI in humans (Evans, Gray, & Snowden, 2007; Nelson & Sanjuan, 2006) may be tapping similar processes. However, until we better understand the mechanism(s) underlying LI in animals, it seems premature to make any strong claim of a parallel between superficially similar effects in animals and humans, an idea that is well illustrated by the case of masked preexposure. Acknowledgements The authors thank the ESRC (UK) and the Danish Agency for Science, Technology and Innovation for funding the research upon which parts of this chapter are based, and Rob Honey for reading a draft of the chapter.
110
M.E. Le Pelley and Mia Schmidt-Hansen
References Allan, L. G., & Jenkins, H. M. (1980). The judgment of contingency and the nature of the response alternatives. Canadian Journal of Psychology, 34, 1–11. Baker, A. G., & Mackintosh, N. J. (1979). Preexposure to the CS alone, or CS and US uncorrelated: latent inhibition, blocking by context or learned irrelevance? Learning and Motivation, 10, 278–294. Baker, A. G., & Mercier, P. (1982). Extinction of the context and latent inhibition. Learning and Motivation, 13, 391–416. Baruch, I., Hemsley, D. R., & Gray, J. A. (1988). Differential performance of acute and chronic schizophrenics in a latent inhibition task. Journal of Nervous and Mental Disease, 176, 598–606. Bennett, C. H., Wills, S. J., Oakeshott, S. M., & Mackintosh, N. J. (2000). Is the context specificity of latent inhibition a sufficient explanation of learned irrelevance? Quarterly Journal of Experimental Psychology, 53B, 239–253. Bonardi, C., Graham, S., Hall, G., & Mitchell, C. (2005). Acquired distinctiveness and equivalence in human discrimination learning: evidence for an attentional process. Psychonomic Bulletin & Review, 12, 88–92. Bonardi, C., & Ong, S. Y. (2003). Learned irrelevance: a contemporary overview. Quarterly Journal of Experimental Psychology, 56B, 80–89. Bouton, M. E. (1993). Context, time, and memory retrieval in the interference paradigms of Pavlovian learning. Psychological Bulletin, 114, 80–99. Channell, S., & Hall, G. (1983). Contextual effects in latent inhibition with an appetitive conditioning procedure. Animal Learning & Behavior, 11, 67–74. De la Casa, L. G., Ruiz, G., & Lubow, R. E. (1993a). Amphetamine-produced attenuation of latent inhibition is modulated by stimulus preexposure duration: implications for schizophrenia. Biological Psychiatry, 33, 707–711. De la Casa, L. G., Ruiz, G., & Lubow, R. E. (1993b). Latent inhibition and recall/ recognition of irrelevant stimuli as a function of preexposure duration in high and low psychotic-prone normal subjects. British Journal of Psychology, 84, 119–132. Dickinson, A., Shanks, D. R., & Evenden, J. L. (1984). Judgement of act-outcome contingency: the role of selective attribution. Quarterly Journal of Experimental Psychology, 36A, 29–50. Escobar, M., Arcediano, F., & Miller, R. R. (2003). Latent inhibition in human adults without masking. Journal of Experimental Psychology: Learning, Memory, and Cognition, 29, 1028–1040. Evans, L. H., Gray, N. S., & Snowden, R. J. (2007). A new continuous withinparticipants latent inhibition task: examining associations with schizotypy dimensions, smoking status and gender. Biological Psychology, 74, 365–373. George, D. N., & Pearce, J. M. (1999). Acquired distinctiveness is controlled by stimulus relevance not correlation with reward. Journal of Experimental Psychology: Animal Behavior Processes, 25, 363–373. Ginton, A., Urca, G., & Lubow, R. E. (1975). The effects of preexposure to a nonattended stimulus on subsequent learning: latent inhibition in adults. Bulletin of the Psychonomic Society, 5, 5–8. Graham, S., & McLaren, I. P. L. (1998). Retardation in human discrimination learning as a consequence of pre-exposure: latent inhibition or negative priming? Quarterly Journal of Experimental Psychology, 51B, 155–172. Gray, J. A. (1998). Integrating schizophrenia. Schizophrenia Bulletin, 24, 249–266.
Latent inhibition and learned irrelevance in human contingency learning
111
Gray, N. S., Fernandez, M., Williams, J., Ruddle, R. A., & Snowden, R. J. (2002). What schizotypal dimensions abolish latent inhibition? British Journal of Clinical Psychology, 41, 271–284. Gray, N. S., Pickering, A. D., Hemsley, D. R., Dawling, S., & Gray, J. A. (1992). Abolition of latent inhibition by a single 5 mg dose of d-amphetamine in man. Psychopharmacology, 107, 425–430. Gray, N. S., & Snowden, R. J. (2005). The relevance of irrelevance to schizophrenia. Neuroscience and Biobehavioral Reviews, 29, 989–999. Gray, N. S., Williams, J., Fernandez, M., et al. (2001). Context dependent latent inhibition in adult humans. Quarterly Journal of Experimental Psychology Section B – Comparative and Physiological Psychology, 54, 233–245. Hall, G. (1991). Perceptual and Associative Learning. Oxford: Oxford University Press. Hulstijn, W. (1978). The orienting reaction during human eyelid conditioning following preconditioning exposures to the CS. Psychological Research, 40, 77–88. Kaye, H., & Pearce, J. M. (1987). Hippocampal lesions attenuate latent inhibition and the decline of the orienting response in rats. Quarterly Journal of Experimental Psychology, 39B, 107–125. Killcross, A. S., Dickinson, A., & Robbins, T. W. (1994). Amphetamine-induced disruptions of latent inhibition are reinforcer mediated: implications for animal models of schizophrenic attentional dysfunction. Psychopharmacology, 115, 185–195. Kruschke, J. K. (1996). Dimensional relevance shifts in category learning. Connection Science, 8, 225–247. Kruschke, J. K. (2001). Towards a unified model of attention in associative learning. Journal of Mathematical Psychology, 45, 812–863. Le Pelley, M. E., & McLaren, I. P. L. (2003). Learned associability and associative change in human causal learning. Quarterly Journal of Experimental Psychology, 56B, 68–79. Le Pelley, M. E., Reimers, S. J., Calvini, G., et al. (submitted). Stereotype formation: biased by association. Journal of Experimental Psychology: General. Le Pelley, M. E., Schmidt-Hansen, M., Harris, N. J., Lunter, C. M., & Morris, C. S. (in press). Disentangling the attentional deficit in schizophrenia: pointers from schizotypy. Psychiatry Research. Lipp, O. V. (1999). The effect of stimulus specificity and number of pre-exposures on latent inhibition in an instrumental trials-to-criterion task. Australian Journal of Psychology, 51, 77–91. Livesey, E. J., Harris, I. M., & Harris, J. A. (2009). Attentional changes during implicit learning: signal validity protects a target stimulus from the attentional blink. Journal of Experimental Psychology: Learning, Memory, and Cognition, 35, 408–422. Lovibond, P. F. (2003). Causal beliefs and conditioned responses: retrospective revaluation induced by experience and by instruction. Journal of Experimental Psychology: Learning, Memory, and Cognition, 29, 97–106. Lovibond, P. F., Preston, G. C., & Mackintosh, N. J. (1984). Context specificity of conditioning and latent inhibition. Journal of Experimental Psychology: Animal Behavior Processes, 10, 360–375. Lubow, R. E. (1973). Latent inhibition. Psychological Bulletin, 79, 398–407.
112
M.E. Le Pelley and Mia Schmidt-Hansen
Lubow, R. E. (1989). Latent Inhibition and Conditioned Attention Theory. Cambridge: Cambridge University Press. Lubow, R. E. (2005). Construct validity of the animal latent inhibition model of selective attention deficits in schizophrenia. Schizophrenia Bulletin, 31, 139–153. Lubow, R. E., Caspy, T., & Schnur, P. (1982). Latent inhibition and learned helplessness in children: similarities and differences. Journal of Experimental Child Psychology, 34, 231–256. Lubow, R. E., & Gewirtz, J. C. (1995). Latent inhibition in humans: data, theory, and implications for schizophrenia. Psychological Bulletin, 117, 87–103. Lubow, R. E., Ingberg-Sachs, Y., Zalstein-Orda, N., & Gewirtz, J. C. (1992). Latent inhibition in low and high psychotic-prone normal subjects. Personality and Individual Differences, 13, 563–572. Mackintosh, N. J. (1973). Stimulus selection: learning to ignore stimuli that predict no change in reinforcement. In R. A. Hinde & J. S. Hinde (Eds.), Constraints on Learning. London: Academic Press, pp. 75–96. Mackintosh, N. J. (1975). A theory of attention: variations in the associability of stimuli with reinforcement. Psychological Review, 82, 276–298. Mackintosh, N. J., & Little, L. (1969). Intradimensional and extradimensional shift learning by pigeons. Psychonomic Science, 14, 5–6. Mason, O., Claridge, G., & Jackson, M. (1995). New scales for the assessment of schizotypy. Personality and Individual Differences, 18, 7–13. Matzel, L. D., Schachtman, T. R., & Miller, R. R. (1988). Learned irrelevance exceeds the sum of CS-preexposure and US-preexposure deficits. Journal of Experimental Psychology: Animal Behavior Processes, 14, 311–319. Miller, R. R., & Matzel, L. D. (1988). The comparator hypothesis: a response rule for the expression of associations. The Psychology of Learning and Motivation, 22, 51–92. Nelson, J. B., & Sanjuan, M. C. (2006). A context-specific latent inhibition effect in a human conditioned suppression task. Quarterly Journal of Experimental Psychology, 59, 1003–1020. Oswald, C. J. P., Yee, B. K., Rawlins, J. N. P., et al. (2001). Involvement of the entorhinal cortex in a process of attentional modulation: evidence from a novel variant of an IDS/EDS procedure. Behavioral Neuroscience, 115, 841–849. Pearce, J. M., & Hall, G. (1980). A model for Pavlovian conditioning: variations in the effectiveness of conditioned but not of unconditioned stimuli. Psychological Review, 87, 532–552. Perlmuter, L. C. (1966). Effect of CS manipulation on the conditioned eyelid response: compounding, generalization, in the inter-CS-interval, and preexposure. Psychonomic Science Monographs, 1, 271–286. Rescorla, R. A. (1971). Summation and retardation tests of latent inhibition. Journal of Comparative and Physiological Psychology, 75. Schwartz, R. M., Schwartz, M., & Teas, R. C. (1971). Optional intradimensional and extradimensional shifts in the rat. Journal of Comparative and Physiological Psychology, 77, 470–475. Siegel, S. (1969). Generalization of latent inhibition. Journal of Comparative and Physiological Psychology, 69, 157–159. Testa, T. J., & Ternes, J. W. (1977). Specificity of conditioning mechanisms in the modification of food preferences. In L. M. Barker, M. R. Best & M. Domjan (Eds.), Learning Mechanisms in Food Selection. Waco, TX: Baylor University Press.
Latent inhibition and learned irrelevance in human contingency learning
113
Wagner, A. R. (1981). SOP: a model of automatic memory processing in animal behaviour. In N. E. Spear & R. R. Miller (Eds.), Information Processing in Animals: Memory Mechanisms. Hillsdale, NJ: Lawrence Erlbaum, pp. 5–47. Weiner, I., & Feldon, J. (1987). Facilitation of latent inhibition by haloperidol. Psychopharmacology, 91, 248–253. Weiner, I., Lubow, R. E., & Feldon, J. (1988). Disruption of latent inhibition by acute administration of low doses of amphetamine. Pharmacology, Biochemistry, and Behavior, 30, 871–878. Williams, J. H., Wellman, N. A., Geaney, D. P., et al. (1997). Haloperidol enhances latent inhibition in visual tasks in healthy people. Psychopharmacology, 133, 262–268. Young, A. M. J., Moran, P. M., & Joseph, M. H. (2005). The role of dopamine in conditioning and latent inhibition: what, when, where and how? Neuroscience and Biobehavioral Reviews, 29, 963–976. Zalstein-Orda, N., & Lubow, R. E. (1995). Context control of negative transfer induced by preexposure to irrelevant stimuli: latent inhibition in humans. Learning and Motivation, 26, 11–28.
6 Associative and nonassociative processes in latent inhibition: an elaboration of the Pearce–Hall model Geoffrey Hall and Gabriel Rodriguez
Background The analysis of latent inhibition to be developed in this chapter had its origins in the attempt to study a (seemingly) quite different phenomenon – the acquired distinctiveness of cues. James (1890) had suggested that similar cues could be rendered more discriminable by training in which each became linked with a very different associate. The idea was taken up by Lawrence (1949) in his classic experiments with rats, and the 1950s and 1960s saw a rush of experiments (reviewed by Hall, 1991) exploring the phenomenon with human subjects. A theoretical account of the effect (based on that of Sutherland & Mackintosh, 1971) was put forward by Mackintosh (1975). He proposed that the representation of any cue has a given level of associability (represented, usually, by the symbol a) but that the value of the parameter a can be changed by experience. When a cue predicts a consequence reliably (or, at least does so more reliably than other cues that might be present) its a value increases; when it predicts a consequence no better than other cues its a value declines. Initial training in which two cues are reliably followed by different consequences will enhance the a value of each of them and subsequent discrimination tasks involving these cues will be facilitated – the acquired distinctiveness effect. Empirical evidence of the acquired distinctiveness effect came largely from studies of discrimination learning, but Mackintosh’s (1975) account implies that the effect should be found in simple conditioning too. Accordingly we (the work was done in collaboration with J.M. Pearce) set about looking for it in the conditioned suppression preparation; we wanted to test the implication that a stimulus trained as a conditioned stimulus (CS), and thus as a reliable predictor of its outcome (shock in this procedure), would be learned about especially readily (its a value should be high) when used as a CS in a subsequent task. The problem, of course, was to devise a design in which transfer could be unambiguously ascribed to change in a, given that the associative strength acquired by the CS would also be likely to affect performance on the test task. In our first attempt (Pearce & Hall, 1979) we trained rats initially on Latent Inhibition: Cognition, Neuroscience and Applications to Schizophrenia, ed. R.E. Lubow and Ina Weiner. Published by Cambridge University Press # Cambridge University Press 2010.
114
Associative and nonassociative processes in latent inhibition
115
Figure 6.1. Design and results of experiment by Pearce and Hall (1979). A, B, X, and Y represent different stimuli; þ indicates the presentation of a footshock. Both groups learned a discrimination in the first phase. The graph shows the acquisition of conditioned suppression (in Phase 2) to the AB compound for Group C, for whom AB was novel, and for Group E, for whom AB had been pretrained in Phase 1.
a discrimination Aþ/AB (where A and B represent different stimuli, and þ and the presence or absence of reinforcement). We reasoned that, after extensive training on this task, the net associative strength of the AB compound would be zero, but that the associability of each of its components would be high (A being a reasonably good predictor of reinforcement, and B a good predictor of its absence); transfer to further conditioning in which the compound was used as a CS should then reveal the acquired distinctiveness, free from the effects of direct associative transfer. Our results (Figure 6.1) indeed showed a transfer effect on this test, but it was the reverse of that anticipated – the AB compound was learned about rather slowly in rats given the Aþ/AB prior training. In a series of follow-up studies (Hall & Pearce, 1979, 1982a, 1982b; Pearce, Kaye, & Hall, 1982) we confirmed the reliability of this negative transfer effect using a different (but related) design. In this we used a single cue, trained initially with a weak shock as the unconditioned stimulus (US). The shock was strong enough to produce some acquisition, but the level of suppression achieved was small, leaving
116
Geoffrey Hall and Gabriel Rodriguez
Figure 6.2. Design and results of an experiment by Pearce et al. (1982). In the first phase subjects were trained with either a light or a tone preceding the presentation of a weak shock (þ). In Phase 2, all received the tone and the shock intensity was increased (+). Acquisition of conditioned suppression in Phase 2 was quicker for subjects for whom the tone was novel at the start of that phase.
room for further learning when, in the second stage, the intensity of the shock was increased. In this case associative transfer seems guaranteed (the strength acquired in the first stage should give the pretrained CS a headstart when it came to the second stage). Even so, the outcome was essentially the same as that just described (see Figure 6.2) – the pretrained stimulus was learned about more slowly in the test stage than a novel stimulus that had not undergone pretraining. Further studies (e.g., Pearce, Kaye, & Hall, 1982; Swan & Pearce, 1988) confirmed the reliability of this negative transfer effect and demonstrated that it could be alleviated if, in the first stage of training, the stimulus was a poor predictor of its consequences (if the outcome varied from one trial to the next). These experiments seemed to force some novel conclusions. They supported the general position that training can change the properties of (the central representation of) a stimulus, modifying its associability; but that change was not the one postulated by Mackintosh (1975) – rather, a stimulus that had been established as a reliable predictor of its consequences appeared to suffer a loss of associability. The first
Associative and nonassociative processes in latent inhibition
117
conclusion, then, was that the acquired distinctiveness effect must have its source in some quite different process (and this has been taken up separately by, e.g., Hall, Graham, Mitchell, & Lavis, 2003; Honey & Hall, 1989; see Hall, 1991, for a review). The second was that our experiments appeared to have revealed a learning process that might be of general importance, worthy of investigation in its own right. In thinking about this matter we were inevitably drawn to the phenomenon of latent inhibition – the observation that repeated nonreinforced exposure to an event reduced its subsequent effectiveness when trained as a CS. Our proposal (developed as the Pearce–Hall model; Pearce & Hall, 1980) was that a stimulus that is a reliable predictor of its consequence will lose associability, and that latent inhibition is a special case of this general truth, in which the consistent consequence happens to be the absence of any event at all. In what follows we will outline in detail the original proposal and then go on to describe ways in which it has been necessary to modify this account in the light of new evidence. We will describe an elaboration of the original theory that results from these considerations, and new experimental work designed to test the elaborated theory that results.
The Pearce–Hall (1980) model The starting point for the model was the assumption (held in common by most attempts to formalize the process of associative learning) that concurrent activation of the representations of the CS and the unconditioned stimulus (US) (achieved by presentation of these events) would establish or strengthen a link between them. The degree of strengthening will be determined by the extent to which the representations are activated (and also, presumably, by the duration of coactivation, although this factor was ignored by the model). Making no further assumptions (surely the correct starting point), this generates the following equation: V ¼ Sl:
ð1Þ
where DV is the change in associative strength (V) produced by the cooccurrence of a CS that produces an activation level of S in its representation and a US that produces an activation level of l. The values of S and l were assumed to be determined by the intensity of the relevant stimuli. Our further, central, assumption was that the associability of a CS will determine the course of associative learning and this was expressed by the introduction of a further parameter (a) associated with the CS. The value of a was specified as varying between 1 and 0; its starting value was assumed to be determined by CS intensity. This resulted in: V ¼ Sal:
ð2Þ
118
Geoffrey Hall and Gabriel Rodriguez
Our results indicated that a will change as a consequence of experience, making it necessary to specify the principles that govern this additional learning process (i.e., additional to that responsible directly for association formation). Our suggestion for these principles was expressed in the following equation: an ¼ jl SVjn1 :
ð3Þ
Here the value of a on trial n is taken to be determined by events occurring on the preceding trial (n1); specifically the value of a is set to equal the absolute value of the discrepancy between the value of l and the summed associative strength (SV) of all CSs that were present on that trial.1 This formulation implies that repeated pairings of a CS and a US will generate increases in associative strength (Equation 2) that will rise to an asymptote as the concurrent decline in a (Equation 3) occurs. At asymptote the value of a will be zero. The introduction of SV into Equation 3 allows for cue competition effects, such as overshadowing. Thus if two cues are conditioned as a compound, both will acquire strength (Equation 2). Although the associative strength acquired by each element of the compound will be less than l, conditioning will stop when their summed strength reaches l (Equation 3), at which point the a value of each will be zero. Finally, it is necessary to say something about the model’s treatment of inhibitory learning (such as will be produced by extinction), as this has relevance to the account of latent inhibition to be developed later. The view adopted was essentially that of Konorski (1967), which equated inhibition with the formation of a new association that acts to oppose the mechanisms activated by the excitatory association. Acquisition of inhibition thus follows the same basic rules: Vi ¼ Sali :
ð4Þ
where Vi is inhibitory associative strength and li the inhibitory reinforcer. The value of the inhibitory reinforcer will depend on the degree of surprise (or frustration or relief, for motivationally significant USs) that results when an anticipated event fails to occur. For simple extinction this will depend on the excitatory strength acquired in acquisition, as follows: li ¼ SV SVi :
ð5Þ
Taking inhibitory learning into account requires us to amend Equation 3 as follows: an ¼ j1 ðSV SVi Þln1 :
ð6Þ
The basic principle, that the value of a declines when a stimulus predicts its consequences, remains unchanged.
1
In fact associability changes are unlikely to be as rapid as Equation 3 implies, and Pearce, Kaye, and Hall (1982) introduced a modification in which a was determined by a weighted average of values from a run of preceding trials. This modification does not change the principle described in the text.
Associative and nonassociative processes in latent inhibition
119
Application to latent inhibition – and some problems The model just outlined readily generates an account of latent inhibition. In the latent inhibition procedure a potential CS is presented on its own, without any subsequent US (and without, on the face of things, any expectation that a US will occur – the conditions necessary for true inhibitory learning). If we assume that the absence of a US equates to the situation in which the value of l is zero, Equation 3 means that the a value of the presented stimulus will fall to zero. Subsequent conditioning will therefore be retarded – on the first trial there will be no increment in the value of V (Equation 2); increments will start to accrue only as experience of the discrepancy between l and V of Equation 3 restores the value of a. Can it really be as simple as this? Sadly, but not surprisingly, the answer turns out to be no. Hall (1991) presented an extensive review of the experimental literature on latent inhibition then available and uncovered a variety of issues that proved to be problematic for so simple a theory. We will consider two of the problems here, one essentially empirical and one theoretical. The empirical issue was the demonstration, quite unexpected on the basis of this theory (but see Wagner, 1976), that latent inhibition shows a marked sensitivity to contextual change. Figure 6.3 shows the results of an experiment by Channell and Hall (1983), the essential features of which have been amply confirmed by subsequent work (e.g., Lovibond, Preston, & Mackintosh, 1984; Rosas & Bouton, 1997; Westbrook, Jones, Bailey, & Harris, 2000). Rats were given preexposure to a stimulus prior to conditioning and the usual latent inhibition effect was observed when this
Figure 6.3. Acquisition of a conditioned response to a light signaling food delivery in rats preexposed to the light either in the test context itself (Exposed same) or elsewhere (Exposed different). Control subjects received prior exposure to the contexts, but not the light. A higher ratio score means a greater tendency to exhibit conditioned responding. (After Channell & Hall, 1983.)
120
Geoffrey Hall and Gabriel Rodriguez
stimulus was then used as a CS signaling food delivery. But when the conditioning phase was conducted in a different context (a different Skinner box) the latent inhibition effect was much attenuated. Why should this be? A possible answer emerges from consideration of another phenomenon that shows context dependence – conditioning itself. Although the effect is far from robust, it seems clear that in some circumstances at least, the vigor of a previously conditioned response will be diminished when the CS is presented in a new context (Hall & Honey, 1989, 1990). This effect does not depend on the existence of a direct association between the context and the CS, such as is presumably formed during initial training; Hall and Mondragon (1998; see also Mondragon, Bonardi, & Hall, 2003) have shown that the context-dependence effect is enhanced when steps are taken to extinguish this direct association. This observation prompted the suggestion that contextual cues exert their effects because they have acquired occasion-setting properties. Specifically, Hall and Mondragon argued that, during initial training, the context acquires control over the CS–US association, control of a form that fosters retrieval of this association when the CS is presented subsequently. In a changed context retrieval will be less effective and the CR less vigorous. Could it be that the learning process responsible for latent inhibition also becomes dependent on the context in this way, so that the effect is attenuated when the preexposed stimulus is transferred to a new context? The parallel seems appealing, but closer inspection soon reveals problems. In particular the account of occasion setting espoused here assumed that the occasion setter acts by facilitating the action of an associative link (between CS and US); but in the account of latent inhibition derived from the Pearce–Hall model, there is no such link – Equation 3 generates a decline in the value of a, independent of any associative learning. The theoretical issue to be considered next has something in common with the point just raised. The central notion of the Pearce–Hall model was that the associability of a stimulus will decline as it comes to predict its consequences; a-change was thus taken to be a consequence of associative learning. The negative transfer effect demonstrated by Hall and Pearce (1979) occurs, according to the theory, because the formation of the CS–US association in the first stage of training leads to a reduction in a. We referred to this effect as indicating “latent inhibition during CS–US pairings”, but the parallel with true latent inhibition, as it was dealt with by the theory, is inexact. As we have seen, the theory can predict that the a value of a stimulus presented alone will decline, but this is not a consequence of a learning process by which the stimulus becomes an accurate predictor of its consequences. To this extent, our treatment of latent inhibition failed to capture the essential feature of the intuition that formed the basis of the Pearce–Hall model. Addressing this matter, Hall (1991) suggested that a true parallel would require that the latent inhibition procedure be conceptualized as involving a learning process in which, over the course of preexposure, the organism learned about the consequences of the stimulus. Hall described this learning as involving the formation of a
Associative and nonassociative processes in latent inhibition
121
stimulus–no-event association. As with orthodox conditioning, the decline in a would then be a result of this learning, and thus of the acquired predictive accuracy of the stimulus. Hall also noted that a hybrid account of latent inhibition would emerge from this conceptualization. One factor in producing the effect would, as before, be the decline in a. But in addition, the stimulus–no-event association formed during preexposure might be expected to interfere proactively with the formation (e.g., Revusky, 1971), or the expression (e.g., Miller & Matzel, 1988) of the CS–US association trained during subsequent conditioning trials. This would constitute a second mechanism of latent inhibition, complementing the effects of the decline in associability. Accepting this associative interference mechanism supplies an explanation for the empirical issue with which this section of the chapter began. If the associative learning that occurs during nonreinforced preexposure is retrieved less successfully when the context is changed then one of the sources of latent inhibition will be ineffective (or less effective) when the preexposed stimulus is conditioned in a new context; that is, the latent inhibition effect will show contextdependence. We next consider a way in which Hall’s (1991) proposal might be formalized.
A formalization Our starting point in the attempt to develop this account of latent inhibition is the assumption that a form of associative learning goes on during nonreinforced preexposure and that this follows the same rules as those that govern association formation during conditioning. We have described this learning, for purposes of exposition, as the formation of a stimulus–no-event association but this is not a characterization that lends itself readily to formalization. Excitatory conditioning is assumed to occur when CS and US nodes are activated together. It is open to us to postulate the existence of a no-event node, but it is necessary to explain why this should be activated along with the stimulus node during preexposure. Rodriguez and Hall (2008) suggested an approach derived from the analysis of inhibitory learning offered by the Pearce–Hall model. This approach begins with the assertion that the events used as the preexposed stimuli in the latent inhibition procedure are unlikely to be truly neutral. Rather, such stimuli are likely to evoke the expectation of some consequence (that is, in associative terms, they will have an excitatory association with some other event or events). It is possible that this may be an intrinsic property of novel stimuli; but it is not necessary to make this assumption, as there is likely to be some degree of generalization from different but similar stimuli that have been experienced previously (perhaps outside the experimental context). Whatever its source, our assertion leads to the conclusion that presentation of a stimulus will activate a preexisting association with some other event, with a certain strength that we can represent as Vevent.
122
Geoffrey Hall and Gabriel Rodriguez
What follows from this analysis is that the repeated presentation of a stimulus, as in the latent inhibition procedure, will generate inhibitory learning. Presentation of the stimulus will activate the representation of some further event that does not then follow. The rules of the Pearce–Hall model predict the formation of a new link strong enough to negate the original excitation. At this point the net associative strength of the cue will be zero, as will its a value, producing a latent inhibition effect when this cue is used as a CS in future conditioning. This account can be formalized in a way that exactly parallels the model’s interpretation of simple extinction of an excitatory CS (see Equations 4, 5, and 6 above). We will refer to the association being formed as a stimulus–no-event association, which acts to oppose the activation or effects of the stimulus–event association. Its growth over successive trials is given by: Vno event ¼ Salno event : As for the inhibitory reinforcer of Equation 5, the value of lno degree to which an event is expected; that is: lno event ¼ SVevent SVno event :
ð7Þ event
depends on the ð8Þ
The value of a will then change in a way similar to that described by Equation 6: an ¼ jlevent ðSVevent SVno event Þjn1 :
ð9Þ
Applying these equations to nonreinforced exposure generates the following. On the first trial a will be high and learning will occur. Since an event is expected (Vevent has a positive value) but as no consequence occurs, the inhibitory reinforcer will be present, its value given by Equation 8, and the CS–no-event association will be strengthened (Equation 7). As this association grows Vevent will be neutralized and learning will stop as lno event (Equation 8) and a (Equation 9) fall to zero. The results of a simple simulation of these changes are presented in Figure 6.4. In this simulation a value of 0.4 was used for S, the starting value of a was 1, and the initial value of Vevent was 0.4. (The predictions made by the model are reasonably independent of variations in these parameters.) The simulation involved six CS-preexposure trials in which stimulus presentation was followed by no event. Over the course of these trials, a decreases (top panel of figure) as a result of the fact that acquisition of inhibitory strength, Vno event, increasingly counteracts the preexisting excitation, Vevent (bottom panel of figure).
An experimental test This interpretation of latent inhibition generates an interesting, and perhaps counterintuitive prediction that has recently been put to experimental test by Rodriguez and Hall (2008). It concerns the effect to be expected if, during latent inhibition training, the target stimulus is preexposed in compound with some other. As we have noted,
Associative and nonassociative processes in latent inhibition
123
1
Alpha
0.8 0.6 0.4 0.2 0 0
1
2
3 Trials
5
6
5
6
4
Associative strength
0.4 Net Vevent Vevent Vno event
0.3
0.2 0.1
0 0
1
2
3
4
Trials
Figure 6.4. A simulation of changes produced by six nonreinforced presentations of a stimulus of salience of 0.4 (S), and initial values of 1 for a and 0.4 for Vevent. The top panel shows the decline in associability (a); the lower panel shows the growth of Vno event, and the consequent change in the net value of Vevent.
the basic Pearce–Hall model is designed to predict overshadowing when a target stimulus is conditioned in compound with another. In essence, the associative strength of the compound reaches asymptote more rapidly than would one element conditioned on its own, with the result that the a value of each of the elements reaches zero more rapidly and conditioning stops, with each element having less associative strength than it would have acquired if conditioned alone. Our first step was to establish that overshadowing could be obtained with the stimuli and training procedures to be used in our subsequent studies of latent inhibition. This preliminary experiment was necessary because, with the stimuli we used (an odor as the target and a taste as the added element), it has sometimes been found that the presence of the taste can potentiate, rather than overshadow, aversion
124
Geoffrey Hall and Gabriel Rodriguez
Figure 6.5. Results of Experiment 1 of Rodriguez and Hall (2008). Consumption of water with an almond odor by rats that had previously received aversion conditioning with the almond alone (the Element group) or with almond mixed with the taste of salt (the Compound group).
conditioning to the odor. Why such an effect should occur is beyond the scope of the present discussion. It is enough to note that, with appropriately chosen stimuli (a moderately strong concentration of the odorant, mixed in with a flavored drink), the standard overshadowing effect is obtained (Bouton, Jones, McPhillips, & Swartzentruber, 1986). Our first experiment confirmed this result. The target stimulus was the odor of almond (a 2% solution of almond essence). One group of subjects (rats) received two conditioning trials in which consumption of this solution was followed by a nausea-inducing injection (of 0.15 M lithium chloride, LiCl, at 10 ml/kg of body weight). A second group received conditioning in which a taste was added to the almond; for them the fluid consumed was a compound of 0.16 M saline and almond. The results of the test trials with the almond solution are presented in Figure 6.5. Both groups showed an aversion (a suppression of consumption) that extinguished over the course of testing, but the aversion was more profound in the groups trained with the element than in the group trained with the compound; that is, overshadowing was obtained. We may now consider the latent inhibition procedure in which, prior to conditioning with almond as the CS, the rats are given nonreinforced preexposure to this odor. According to our account, the subjects learn that no event follows presentation of the stimulus and, given enough preexposure trials, the a value will fall to zero. What will happen if the odor is preexposed in compound with the taste? Figure 6.6 presents the results of a simulation, using Equations 7, 8, and 9, of the changes in associability and associative strength produced by six preexposure trials. The element condition refers to the case in which the target stimulus (the odor in our experiments)
Associative and nonassociative processes in latent inhibition
125
Figure 6.6. A simulation of changes produced by six nonreinforced presentations of a stimulus of salience of 0.4 (S), and initial values of 1 for a and 0.4 for Vevent. The top panel shows the decline in associability (a); the lower panel shows changes in the net value of Vevent. In the Element condition, the stimulus is presented alone. The Compound condition shows the case in which the target stimulus is presented in compound with another assumed to have a salience (S) of 0.6, and initial values of 1 for a and 0.6 for Vevent.
is presented alone. The same parameter values were used as in the simulation of Figure 6.4 (S fixed at 0.4, a starting value of 1 for a, and an initial value of 0.4 for Vevent). This reproduces, for the element condition, the decline in associability seen in the previous figure and below it the decline in the net strength of Vevent, consequent on the growth of the Vno event association. Also presented are the equivalent results for a simulation in which the target stimulus is presented in compound with a more salient nontarget stimulus (such as the taste is assumed to be in our experimental procedure). For this added stimulus the starting value of a was again taken to be 1, but the value of S was set at 0.6 (i.e., at a higher value than that of the target element); and the initial value of Vevent was set at 0.6, reflecting the intuition that the expectation that an event will follow a stimulus is likely to be greater for more salient stimuli.
126
Geoffrey Hall and Gabriel Rodriguez
As the lower panel of Figure 6.6 shows (and similar effects are obtained with a range of parameter values), the value of Vevent falls very steeply for the target stimulus when it is presented in compound with another. The mechanism is essentially that described recently by Rescorla (2006) in a discussion of simple extinction. In the present context it works as follows: presenting two stimuli together will generate a strong expectation that an event will occur; in the absence of an event, the magnitude of the inhibitory reinforcer (Equation 8) will be greater than for the case in which the element is presented alone, Vno event will grow rapidly, and Vevent will reach zero after fewer trials than are needed when the element is preexposed alone. (Although the results are not shown in the figure, the same will apply to the other element of the compound.) The effect of this learning on the value of a is shown in the top part of the figure. The associability of the stimulus preexposed in compound will fall to zero (Equation 9) as the expectation that some event will follow stimulus presentation does so, and this will occur more rapidly with compound than with element preexposure. These considerations lead to the prediction that latent inhibition might be enhanced by preexposure in which the target stimulus is presented in compound with another, more salient, stimulus. Not only will associability be lower after compound than after element training, the strength of the Vevent association will be lower after compound than after element training, something that might also influence the rate of subsequent conditioning when the CS is followed by a US. Rodriguez and Hall (2008, Experiment 2) tested the implications of this analysis in an experiment using the taste and odor stimuli described above. There were three groups of subjects. The element group received six initial nonreinforced presentations of the target stimulus, the almond odor; the compound group received six presentations of almond mixed with saline; the third, the control group, received no preexposure. We then gave aversion conditioning with the odor as the CS. We hoped to show latent inhibition in the element group and potentiation of the latent inhibition effect in the compound group. It should be admitted that this was indeed a hope, rather than a confident prediction, for two reasons. First our prediction depends on the number of exposure trials given – as Figure 6.6 shows, early in exposure the associability of the target stimulus will be higher for the compound group than for the element group; and at asymptote the a value of all the cues will be zero. To obtain the effect we are seeking requires that we guess correctly the appropriate number of preexposure trials. Secondly, there is the problem of generalization decrement. For latent inhibition acquired during preexposure to carry over into the conditioning phase requires that the test stimulus (the odor) be identified by the subject as the element that was presented in compound with the taste in preexposure. If the presence of the taste modifies the way in which the odor is perceived, transfer across the stages of the experiment cannot be expected, and latent inhibition might be attenuated rather than potentiated. A number of previous studies of compound preexposure in latent inhibition have found just such an attenuation (e.g., Honey & Hall, 1988, 1989a) and have attributed the result to the effects of generalization decrement. In choosing our stimuli (an odor and a taste) for
Associative and nonassociative processes in latent inhibition
127
Figure 6.7. Results of Experiment 2 of Rodriguez and Hall (2008). All subjects received aversion conditioning with almond odor; the results are for the second and third conditioning trials (C2 and C3) and for the final nonreinforced test trial (T). Subjects in the Element group had received preexposure to almond; those in the Compound group had received preexposure to almond plus salt. The Control group was not given preexposure.
this experiment we hoped to avoid the generalization decrement effects of previous studies (Honey & Hall, 1989b, used a mixture of two tastes). But only the experimental results can tells us whether we have been right in our choice of stimuli. The results of the test phase of our experiment are shown in Figure 6.7. All subjects drank 10 ml of the almond solution on the first conditioning trial and were then given an injection of LiCl. The effects of this were seen on the next trial (C2 in the figure) in which consumption of almond was again followed by an injection. A further injection was given on the next trial (C3) and this was followed by a nonreinforced test trial (T). It is evident that the aversion was acquired most readily in the control group (consumption of the almond solution was dramatically suppressed). Suppression occurred less readily in the group preexposed to almond alone (the element group), demonstrating a latent inhibition effect. And the group preexposed to the compound learned most slowly of all; that is, latent inhibition was potentiated by compound preexposure. This result appears to be uniquely predicted by the account of latent inhibition that we have developed here. It is necessary, however, to establish its reliability and to consider possible alternative explanations. Rodriguez and Hall (2008) conducted a further study (their Experiment 5) in which all the subjects were preexposed to two odors (almond, as before, and vanilla), one presented on its own and the other in compound with the taste. For the test phase, the subjects were divided into two groups; one received aversion conditioning with odor that was previously experienced alone, the other with the odor preexposed as part of a compound. Our previous result would be confirmed if conditioning proceeded more slowly in the latter group.
128
Geoffrey Hall and Gabriel Rodriguez
Figure 6.8. Results of Experiment 5 of Rodriguez and Hall (2008): group means for consumption of a solution containing an odor over two conditioning trials (C1 and C2) and for a nonreinforced test trial (T). All subjects had received preexposure to one odor presented alone (O1) and another (O2) presented in compound with saline.
This experimental design, in which the two groups are fully matched in their preexposure experience, allows us to rule out some possible alternative explanations for the results obtained in the previous experiment. It might be argued, for example, that the procedure used in the previous experiment means that for the compound group both odor and taste underwent latent inhibition, whereas, for the element group, only the odor did so. If there is some similarity between the taste and the odor, the latent inhibition suffered by the taste might generalize to the odor in the test phase for the compound group. Slow learning by the compound group might thus be the product of direct generalization rather than the potentiation mechanism of interest here. No such argument will apply to the present design, however, in which both groups received experience of the taste during preexposure. Figure 6.8 shows the critical results of this experiment; the amount consumed by each group of the solution containing the odor over two conditioning trials and a final test trial. On the first conditioning trial all subjects were given a fixed amount of the solution (10 ml) prior to the first LiCl injection. When the animals were given free access on the second trial, the element group showed greater suppression of consumption than the compound group; consumption was suppressed as a result of this second conditioning trial, but the difference between them was maintained on the test trial. This experiment thus successfully replicated the effect seen in our previous study and did so using a procedure in which the groups were equated in their experience of the taste and odor in preexposure.
Complications and issues to be resolved The elaborated version of the Pearce–Hall model, described above, deals perfectly well with the basic facts of latent inhibition and, as we just have seen, generates an interesting new prediction that has been confirmed experimentally. But a number of
Associative and nonassociative processes in latent inhibition
129
issues remain, largely to do with details of the associative component of the hybrid theory, that are not fully resolved.
Reinforcer specificity Our central assumption, that, when first presented, a novel stimulus will evoke the expectation that some event will follow (that Vevent will have some initial strength), has something in common with the analysis of latent inhibition offered by Killcross and Balleine (1996). These authors investigated the effects of giving stimulus preexposure under different motivational states; for example, rats were given exposure to one stimulus (S1) when hungry and another (S2) when thirsty. One group of rats then received conditioning in which both S1 and S2 signaled food; a second group received conditioning in which both signaled a fluid reinforcer (all rats were both hungry and thirsty during this test). Killcross and Balleine found that S1 was learned about rather slowly when it was used as a signal for food, but that S2 was learned about slowly when it was a signal for fluid. They concluded that during preexposure the animal learns something about the relationship between the stimulus and other events (a conclusion in accord with theory being developed here), but they went on to add that their rats must be learning, specifically, that the preexposed stimulus is unrelated to events of relevance to their current motivational state. These observations require us to refine some aspects of our proposal that latent inhibition derives from learning evoked by the ability of a novel stimulus to evoke the representation of some event. Two further assumptions become necessary. First, it is necessary to assume that the nature of the event activated by stimulus presentation is in some way “gated” by the motivational state of the animal – for example, that of the variety of nodes that might be activated by the stimulus, those representative of foodstuffs are more likely to be activated when the animal is hungry. The decline in a consequent on the nonappearance of a poststimulus event will then occur just as before. And we must also accommodate the fact that the latent inhibition effect is specific to a given class of reinforcer; that, to continue with this example, it is learning about the CS–food relation that is particularly retarded. Our second assumption, therefore, must be that the a value of the stimulus is specific to a given reinforcer or class of reinforcers – that a stimulus may have an associability for food that is independent of its associability for water, or for shock, or whatever. This second assumption complicates matters substantially; but it is not without precedent – Mackintosh (1975), in his theory of attention, specifically argued that the associability of a stimulus for one class of events might change independently of its associability for another class. Mackintosh had in mind appetitive and aversive events as the two major classes, but how these classes should be defined is essentially an empirical matter.
130
Geoffrey Hall and Gabriel Rodriguez
Context specificity The starting point for the hybrid theory of latent inhibition was the observation that latent inhibition shows context specificity. This observation prompted the suggestion that at least part of the latent inhibition effect stems from associative interference – that an association formed during preexposure will interfere in some way with new learning, but will not be retrieved (and therefore will not interfere) when the context is changed. In the formulation adopted here, the association acquired during preexposure is that between stimulus and no event, an association that acts to negate the preexisting stimulus–event association. If we assume that the change of context effectively eliminates the contribution of the newly acquired (inhibitory) association then what differentiates animals conditioned in a new context from those conditioned in the preexposure context is that the former will have an expectation that some event will follow the stimulus, whereas the latter will not. It is clearly possible to argue that animals in the former condition might acquire a new CR more readily (that is, show an attenuation of the latent inhibition effect). (Such an effect might thus be more properly described as associative facilitation, rather than associative interference.) Clearly, the details of the mechanism involved here need to be specified more fully. But before attempting this, it is worth noting that another feature of the theory proposed here may make this mechanism unnecessary. The central proposal is that a change of context will render ineffective the association formed during the preexposure phase (that, in the limiting case, the value of Vno event will fall to zero). (The notion that inhibitory, or perhaps, second-learned, associations are particularly sensitive to the effects of a change of context is well established from studies of orthodox conditioning, e.g., Bouton, 1993; Nelson, 2002.) This has implications for the change in associability that will occur on the first trial of the conditioning phase. Recall that this is given by: an ¼ jlevent ðSVevent SVno event Þjn1 :
ð9Þ
For all subjects the value of levent of the previous trial was zero, and for those conditioned in the preexposure context the value of (SVevent SVno event) will also be zero; a will thus remain at zero on this trial. But if a change of context means that the Vno event association cannot be retrieved, subjects in this condition will experience a discrepancy between levent and (SVevent SVno event), as the latter will now have a positive value. As a consequence, a will be restored on this trial by the change of context, and the latent inhibition effect will be attenuated.
Inhibition and latent inhibition The mechanism just described depends on the notion that context change limits the retrieval of associative information, but the consequence for latent inhibition is by way of an effect on a rather than by direct associative interference (or facilitation).
Associative and nonassociative processes in latent inhibition
131
Whether such a mechanism can be developed to deal with all the cases that have been taken as evidence of a role for associative interference in latent inhibition may be doubted. (It is difficult to see, for example, how it might accommodate some of the retention interval effects obtained by Aguado, Brugada, & Hall, 1997.) It is worthwhile, therefore, to attempt to expand a little on the implications of the associative component of our hybrid theory. According to our theory, a newly presented stimulus engenders the expectation that some further event will follow; a preexposed stimulus does not. If this difference plays any part in producing the latent inhibition effect, we need to suppose, then, that the expectation provoked by the novel stimulus fosters the development or expression of a CR when that stimulus is used as a CS. In the absence of any evidence that bears directly on this issue, we have adopted, as a working hypothesis, the notion that the CS–event association interacts with the developing CS–US association to promote performance of the CR. Whether or not this particular formulation is accepted, it is necessary to address a problem that confronts all theories of this general type – the issue of latent inhibition in inhibitory conditioning. If subjects that have been preexposed to a stimulus have no expectation that an event will follow (by virtue of having formed a stimulus–no-event association), is it not reasonable to expect that their subsequent learning will be facilitated when that stimulus is trained as an inhibitor (as a signal for the omission of the US)? That the experimental evidence (e.g., Halgren, 1974; Reiss & Wagner, 1972; Rescorla, 1971; see also Friedman, Blaisdell, Escobar, & Miller, 1998) uniformly shows that preexposure retards inhibitory as well as excitatory conditioning is not, in fact, a major problem for a hybrid theory of the type being developed here. Whatever associative interference (or facilitation effects) may be operating in these procedures, it is also the case that stimulus preexposure will have brought about a reduction in the value of a. Retarded inhibitory learning is thus to be expected if the effects of associability loss outweigh any effects produced by associative interactions (see Hall, Kaye, & Pearce, 1985). Instead we need to turn to the other standard test for conditioned inhibition (Rescorla, 1969), the summation test. In this, no new learning is required (and hence there is no contribution from the associability factor); but a signal for the omission of a US will, when presented in compound with an excitatory CS, reduce the magnitude of the CR elicited by that CS. What will be the effects of latent inhibition training with this test procedure? According to our account a novel stimulus evokes the expectation of an event, whereas a preexposed stimulus activates the no-event representation. It might be expected therefore that a latently inhibited stimulus would be effective in reducing the magnitude of the CR evoked by a separately trained excitor when the stimuli are compounded in a summation test. This issue has been addressed in three studies using animal conditioning procedures, two of conditioned suppression in rats (Kremer, 1972; Rescorla, 1971), and one using the rabbit eyeblink procedure (Reiss & Wagner, 1972). They generated all possible results. All of them found that the addition of another stimulus to the
132
Geoffrey Hall and Gabriel Rodriguez
pretrained excitor produced a reduction in the CR, a result readily explained as an instance of generalization decrement (Pavlov’s external inhibition). The issue, therefore, is whether the size of this reduction is larger for a preexposed than for a novel added stimulus, indicating that some additional factor might be operating in determining the outcome. Rescorla’s study found no difference; Kremer found that a preexposed stimulus was more effective, and Wagner that a preexposed stimulus was less effective in this respect. The explanation for this last finding appears to be that the preexposed stimulus, having been subjected to an habituation procedure during preexposure, will have suffered a loss of effective salience, making it less potent as an external inhibitor. This same general line of argument can supply an explanation for the opposite result obtained by Kremer. Habituation training will make the stimulus less effective at eliciting its unconditioned response, which in this case will be the unconditioned suppression of baseline operant responding that a novel stimulus routinely evokes. Summation of unconditioned suppression with the conditioned suppression evoked by the test excitor could outweigh the operation of any other factors in this situation, producing the result observed. The only safe conclusion on the basis of these results, given the complications introduced by the habituation effects just discussed, is that the central issue remains unresolved. It could be that the associative learning produced by preexposure plays a part in determining the effect seen in a summation test, but the results currently available do not allow us to decide one way or the other. Resolution of this matter must await the result of further experiments in which the various effects produced by habituation are controlled for in some way.
Conclusion For a procedure as simple as latent inhibition, the theoretical mechanisms involved in its explanation turn out to be surprisingly complex. According to our account, simple nonreinforced preexposure engages both associative learning processes (those involved in learning the relation between the stimulus and its consequences) and nonassociative processes (those involved in modulating the value of the associability parameter, a). Both forms of learning play a role in the effect ultimately observed – the retarded development of the CR when the stimulus is used as a CS. And this is not all. In our discussion of the interpretation of summation tests, we noted that preexposure to a stimulus will engage the learning mechanism responsible for habituation. We characterized this as involving a reduction in the effective salience of the stimulus, a change identified by Hall (2003) as a reduction in the value of the parameter S, the parameter equated with salience in our learning equations. We noted the implications of a loss in salience for the summation test (that the stimulus would be less effective as an external inhibitor and at evoking its UR), but we must now point out that it has direct implications for the retardation test – that is, for the latent inhibition effect itself. If the value of S declines during exposure, the
Associative and nonassociative processes in latent inhibition
133
acquisition of associative strength during subsequent CS–US pairings will necessarily be retarded (Equation 2). Loss of salience will combine with loss of associability and associative interference effects in producing the overall result obtained. The proposal that a habituation process might play a part in producing latent inhibition will not be seen as controversial. But to acknowledge its possible role is only a first step. We have identified a learning process that modifies that value of a, the associability of the stimulus, and have devoted some effort to the attempt to identify the rules that govern a change and to specify how this form of learning interacts with associative learning. We are now proposing that the salience parameter, S, can also change with experience. It becomes necessary, therefore, to do a similar job for the learning process responsible for salience change. We have made a start on this (e.g., Hall, Blair, & Artigas, 2006; Hall, Prados, & Sansa, 2005), but a full account of latent inhibition will not be available until this work is completed.
References Aguado, L., de Brugada, I., & Hall, G. (1997). Effects of a retention interval on the US-preexposure phenomenon in flavor aversion learning. Learning and Motivation, 28, 311–322. Bouton, M. E. (1993). Context, time, and memory retrieval in the interference paradigms of Pavlovian learning. Psychological Bulletin, 114, 80–99. Bouton, M. E., Jones, D. L., McPhillips, S. A., & Swartzentruber, D. (1986). Potentiation and overshadowing in odor-aversion learning: role of method of odor presentation, the distal-proximal cue distinction, and the conditionability of the odor. Learning and Motivation, 17, 115–138. Channell, S., & Hall, G. (1983). Contextual effects in latent inhibition with an appetitive conditioning procedure. Animal Learning & Behavior, 11, 67–74. Friedman, B. X., Blaisdell, A. P., Escobar, M., & Miller, R. R. (1998). Comparator mechanisms and conditioned inhibition: conditioned stimulus preexposure disrupts Pavlovian conditioned inhibition but no explicitly unpaired inhibition. Journal of Experimental Psychology: Animal Behavior Processes, 24, 453–466. Halgren, C. R. (1974). Latent inhibition in rats: associative or nonassociative? Journal of Comparative and Physiological Psychology, 86, 74–78. Hall, G. (1991). Perceptual and Associative Learning. Oxford: Clarendon Press. Hall, G. (2003). Learned changes in the sensitivity of stimulus representations: associative and nonassociative mechanisms. Quarterly Journal of Experimental Psychology, 56B, 43–55. Hall, G., Blair, C. A. J., & Artigas, A. A. (2006). Associative activation of stimulus representations restores lost salience: implications for perceptual learning. Journal of Experimental Psychology: Animal Behavior Processes, 32, 145–155. Hall, G., Graham, S., Mitchell, C., & Lavis, Y. (2003). Acquired equivalence and distinctiveness in human discrimination learning: evidence for associative mediation. Journal of Experimental Psychology: General, 132, 266–276.
134
Geoffrey Hall and Gabriel Rodriguez
Hall, G., & Honey, R. C. (1989). Contextual effects in conditioning, latent inhibition, and habituation: associative and retrieval functions of contextual cues. Journal of Experimental Psychology: Animal Behavior Processes, 15, 232–241. Hall, G., & Honey, R. C. (1990). Context-specific conditioning in the conditioned emotional response procedure. Journal of Experimental Psychology: Animal Behavior Processes, 16, 271–278. Hall, G., Kaye, H., & Pearce, J. M. (1985). Attention and conditioned inhibition. In R. R. Miller & N. E. Spear (Eds.), Information Processing in Animals: Conditioned Inhibition. Hillsdale, NJ: Erlbaum, pp. 185–208. Hall, G., & Mondragon, E. (1998). Contextual control as occasion setting. In N. Schmajuk & P. Holland (Eds.), Occasion Setting: Associative Learning and Cognition in Animals. Washington DC: American Psychological Association, pp. 199–222. Hall, G., & Pearce, J. M. (1979). Latent inhibition of a CS during CS–US pairings. Journal of Experimental Psychology: Animal Behavior Processes, 3, 31–42. Hall, G., & Pearce, J. M. (1982a). Restoring the associability of a pre-exposed CS by a surprising event. Quarterly Journal of Experimental Psychology, 34B, 127–140. Hall, G., & Pearce, J. M. (1982b). Changes in stimulus associability during conditioning: implications for theories of acquisition. In M. L. Commons, R. J. Herrnstein, & A. R. Wagner (Eds.), Quantitative Analyses of Behavior: Acquisition, vol. 3. Cambridge, MA: Ballinger, pp. 221–239. Hall, G., Prados, J., & Sansa, J. (2005). Modulation of the effective salience of a stimulus by direct and associative activation of its representation. Journal of Experimental Psychology: Animal Behavior Processes, 31, 267–276. Honey, R. C., & Hall, G. (1988). Overshadowing and blocking procedures in latent inhibition. Quarterly Journal of Experimental Psychology, 40B, 163–186. Honey, R. C., & Hall, G. (1989a). Acquired equivalence and distinctiveness of cues. Journal of Experimental Psychology: Animal Behavior Processes, 15, 338–346. Honey, R. C., & Hall, G. (1989b). Attenuation of latent inhibition after compound preexposure: associative and perceptual explanations. Quarterly Journal of Experimental Psychology, 41B, 355–368. James, W. (1890). Principles of Psychology. New York: Holt. Kaye, H., & Pearce, J. M. (1984). The strength of the orienting response during Pavlovian conditioning. Journal of Experimental Psychology: Animal Behavior Processes, 10, 90–109. Killcross, S., & Balleine, B. (1996). Role of primary motivation in stimulus preexposure effects. Journal of Experimental Psychology: Animal Behavior Processes, 22, 43–50. Konorski, J. (1967). Integrative Activity of the Brain. University of Chicago Press, Chicago. Kremer, E. F. (1972). Properties of a preexposed stimulus. Psychonomic Science, 27, 45–47. Lawrence, D. H. (1949). Acquired distinctiveness of cues: I. Transfer between discriminations on the basis of familiarity with the stimulus. Journal of Experimental Psychology, 39, 770–784.
Associative and nonassociative processes in latent inhibition
135
Lovibond, P. F., Preston, G. C., & Mackintosh, N. J. (1984). Context specificity of conditioning, extinction, and latent inhibition. Journal of Experimental Psychology: Animal Behavior Processes, 10, 360–375. Mackintosh, N. J. (1975). A theory of attention: variations in the associability of stimuli with reinforcement. Psychological Review, 82, 276–298. Miller, R. R., & Matzel, L. D. (1988). The comparator hypothesis: a response rule for the expression of associations. The Psychology of Learning and Motivation, 22, 51–92. Mondrago´n, E., Bonardi, C., & Hall, G. (2003). Negative priming and occasion setting in an appetitive Pavlovian procedure. Learning & Behavior, 31, 281–291. Nelson, J. B. (2002). Context specificity of excitation and inhibition in ambiguous stimuli. Learning and Motivation, 33, 284–310. Pearce, J. M., & Hall, G. (1979). Loss of associability by a compound stimulus comprising excitatory and inhibitory elements. Journal of Experimental Psychology: Animal Behavior Processes, 5, 19–30. Pearce, J. M., & Hall, G. (1980). A model for Pavlovian conditioning: variations in the effectiveness of conditioned but not of unconditioned stimuli. Psychological Review, 87, 532–552. Pearce, J. M., Kaye, H., & Hall, G. (1982). Predictive accuracy and stimulus associability: development of a model for Pavlovian learning. In M. L. Commons, R. J. Herrnstein, & A. R. Wagner (Eds.), Quantitative Analyses of Behavior: Acquisition, vol. 3. Cambridge, MA: Ballinger, pp. 241–255. Reiss, S., & Wagner, A. R. (1972). CS habituation produces a “latent inhibition effect” but no active “conditioned inhibition”. Learning and Motivation, 3, 237–245. Rescorla, R. A. (1969). Pavlovian conditioned inhibtion. Psychological Bulletin, 72, 77–94. Rescorla, R. A. (1971). Summation and retardation tests of latent inhibition. Journal of Comparative and Physiological Psychology, 75: 77–81. Rescorla, R. A. (2006). Deepened extinction from compound stimulus presentation. Journal of Experimental Psychology: Animal Behavior Processes, 32, 135–144. Rescorla, R.A., & Wagner, A.R. (1972). A theory of Pavlovian conditioning: variations in the effectiveness of reinforcement and non-reinforcement. In A. H. Black & W. F. Prokasy (Eds.), Classical Conditioning II: Current Research and Theory. New York: Appleton-Century-Crofts, pp. 64–99. Revusky, S. (1971). The role of interference in association over a delay. In W. K. Honig & P. H. R. James (Eds.), Animal Memory. New York: Academic Press, pp. 155–213. Rodriguez, G., & Hall, G. (2008). Potentiation of latent inhibition. Journal of Experimental Psychology: Animal Behavior Processes, 34, 352–360. Rosas, J. M., & Bouton, M. E. (1997). Additivity of the effects of retention interval and context change on latent inhibition: toward a resolution of the context forgetting paradox. Journal of Experimental Psychology: Animal Behavior Processes, 23, 283–294. Swan, J. A., & Pearce, J. M. (1988). The orienting response as an index of stimulus associability in rats. Journal of Experimental Psychology: Animal Behavior Processes, 4, 292–301.
136
Geoffrey Hall and Gabriel Rodriguez
Sutherland, N. S., & Mackintosh, N. J. (1971). Mechanisms of Animal Discrimination Learning. New York: Academic Press. Wagner, A. R. (1976). Priming in STM: an information-processing mechanism for self-generated or retrieval-generated depression in performance. In T. J. Tighe & R. N. Leaton (Eds.), Habituation: Perspectives from Child Development, Animal Behavior, and Neurophysiology. Hillsdale, NJ: Erlbaum, pp. 95–128. Westbrook, R. F., Jones, M. L., Bailey, G. K., & Harris, J. A. (2000). Contextual control over conditioned responding in a latent inhibition paradigm. Journal of Experimental Psychology: Animal Behavior Processes, 26, 157–173.
7 From latent inhibition to retrospective revaluation: an attentional-associative model Nestor Schmajuk
This chapter summarizes my work with Jeffrey Gray (1934–2004) on latent inhibition (LI), i.e., the retardation in the generation of the conditioned response (CR) during the pairing of the conditioned stimulus (CS) and the unconditioned stimulus (US) following CS preexposure. From 1996 to 2004, we published a series of articles on (a) a neural network model that describes LI and other classical conditioning paradigms (Schmajuk, Lam, & Gray, 1996), (b) the mapping of variables in the model onto brain structures and neurotransmitters (Schmajuk, Cox, & Gray, 2001; Schmajuk, Gray, & Larrauri, 2005), (c) the application of the model to the description of the effect of brain lesions and drug administration on LI and the characterization of the neural activity and neurotransmitter release in different brain areas during LI (Buhusi, Gray, & Schmajuk, 1998; Schmajuk, Buhusi, & Gray, 1998), and (d) how the absence of LI, regarded as a positive symptom of schizophrenia, is caused by hippocampal dysfunction and ameliorated by DA blockers (Schmajuk, Christiansen, & Cox, 2000; Schmajuk, 2001; Schmajuk, 2002). After 2004, still guided by Jeffrey’s suggestions, Jose Larrauri and I (Schmajuk & Larrauri, 2006) successfully applied the model to data that seem to challenge the power of existing models of classical conditioning. Namely, recovery from blocking (the reduced responding to one CS of a conditioned compound when the other CS was separately conditioned first), backward blocking (the reduced responding to one CS of a conditioned compound when the other CS was separately conditioned following compound conditioning), the counteraction between LI and overshadowing (the reduced responding to one or both CSs of a conditioned compound compared with their individual conditioning), and super LI (decreased responding to a preexposed CS after a delay following conditioning,). Most recently, we applied the model to describe the multiple properties of extinction (Larrauri & Schmajuk, 2008) and to some aspects of retrospective revaluation (described below). In all cases, we found that the model primarily developed to describe LI was able to correctly address many experimental results that had proven difficult for other approaches. Latent Inhibition: Cognition, Neuroscience and Applications to Schizophrenia, ed. R. E. Lubow and Ina Weiner. Published by Cambridge University Press # Cambridge University Press 2010.
137
138
Nestor Schmajuk
In this chapter, we introduce the Schmajuk–Lam–Gray (SLG) model, explain how it describes the multiple properties of LI, and illustrate how it applies to other important paradigms of classical conditioning.
A neural network model of classical conditioning and LI Schmajuk, Lam, and Gray (1996) introduced a neural network theory of classical conditioning, with special emphasis on LI. The network includes (a) an attentional mechanism, designed to explain LI, (b) an associative network, designed to describe changes in conditioned stimulus–unconditioned stimulus (CS–US) and CS–CS associations compatible with overshadowing, blocking, as well as conditioned inhibition, and (c) a feedback loop, introduced to describe sensory preconditioning and secondorder conditioning. Surprisingly, the “emergent properties” of the resulting model proved capable of describing many other experimental results for which it was not specifically designed. According to the SLG model, a representation of a CS (or the context, CX), called XCS (or XCX), can become associated with the US (or with other CSs or CXs). Representation XCS is characterized by (a) being a short-term memory of the CS, (b) whose magnitude increases if the novelty detected in the environment is large, but decreases if novelty is small. In the model, novelty is represented by variable Novelty0 , which is proportional to the sum of the mismatches between predicted and observed events. If the magnitude of representation XCS is large, then the rate of change of its association with the US is high; if the magnitude is small, the rate is low. In addition, a large XCS implies a strong activation of its association with the US, and the consequent generation of a strong CR. Therefore, the magnitude of XCS controls both storage and retrieval of CS–US associations. A complete set of differential equations describing the model can be found in Schmajuk et al. (1996). Figure 7.1 shows the neural mechanisms that implement the model. (1) Presentation of a CS activates a short-term memory trace, tCS. (2) Attentional memory, zCS, reflects the association between tCS and the Novelty0 detected by the system in the environment. (3) zCS increases when Novelty0 is large and decreases when Novelty0 is small. (4) Novelty0 is proportional to the sum of the absolute values of the differences between predicted ðBÞ and observed ðlÞ average values of the CS, the context (CX), and the US, Novelty0 ¼ fðNoveltyÞ ¼ fðjlCS BCS j þ jlCX BCX j þ jlUS BUS jÞ. (5) The activity level of the internal representation of CS, XCS, is proportional to the product XCS tCS zCS. (6) VCS,US is the associative strength of XCS. (7) Changes in VCS,US are proportional to XCS and the difference between the actual intensity of the US, and the predicted intensity of the US, BUS (US BUS). Therefore, a large XCS implies a fast change in VCS,US. (8) The prediction of the US, BUS, is proportional to the product XCS VCS,US, BUS XCS VCS,US. Therefore, when VCS,US is not zero, a large XCS implies a large BUS. (9) CR amplitude increases with BUS, CR BUS XCS VCS,US. Therefore, when VCS,US is not zero, a large XCS also implies a large CR.
139
From latent inhibition to retrospective revaluation Predicted CS B CS
τ
CS–US Association
Attention to CS CS
US – B US
Novelty′
1 z CS
CR
3
2 X CS
V CS,US
B US
Figure 7.1. The associative–attentional model. A simplified diagram of the model. CS: conditioned stimulus; US: unconditioned stimulus; tCS: trace of the CS; zCS: attention to the CS; XCS: internal representation of the CS; VCS-US: CS–US association; BCS: predicted CS; BUS: predicted US; CR: conditioned response. Triangles: variable connections between nodes. Arrows: fixed connections between nodes.
The SLG model shares properties with other associability models, including the attentional control of the formation of CS–US associations (Pearce & Hall, 1980), the competition among CSs to become associated with the US (Rescorla & Wagner, 1972) or other CSs (Schmajuk & Moore, 1988), and the combination of attention and competition (Wagner, 1979). Other properties of the SLG model include: (a) its equations portray behavior on a moment-to-moment basis (as in the Wagner, 1981, SOP model). (b) It describes not only CS–US associations, as the above-mentioned models, but also CS–CS associations. (c) Attention to the CS is controlled by the CS–US associations (as in the Pearce & Hall, 1980, model), by CX–CS associations (as in the Wagner, 1979, model), and by CS–CS associations (as in the Schmajuk & Moore, 1988, model, a property that explains why LI can become context-independent). (d) Retrieval of CS–US associations is controlled by the magnitude of the attentionally modulated, internal representation of CS, XCS. Importantly, the “retrieval” property is not an additional assumption but a simple and direct consequence (an emergent property) of the way neural networks store and retrieve information: XCS controls both the rate of change in CS–US associations (dV/dt ¼ f(XCS)) and the output of the neuron storing the association, thereby controlling the CR (CR ¼ f(XCS)). Because XCS is controlled by Novelty0 (X ¼ f(Novelty0 )), both storage (dV/dt ¼ f(Novelty0 )) and retrieval (CR ¼ f(Novelty0 )) are functions of Novelty0 . Therefore, according to the SLG model, experimental manipulations of Novelty0 after conditioning can modify the retrieval of already established associations, thereby modifying CR strength even when CS–US associations remain constant. These Novelty0 manipulations can modify attention to a CS that is absent, if its representation is activated by a present CS through CS–CS associations, a mechanism that is used to explain retrospective revaluation (including the recovery from LI, from overshadowing, backward blocking, and post-blocking additive effects).
140
Nestor Schmajuk
NonPreexposed
CS Preexposed
0.8 0.6 0.4 0.2 0 CS
XCS
VCS, US
BUS
CR
CS
XCS
VCS, US
BUS
CR
CS
XCS
VCS, US
BUS
CR
0.8 0.6 0.4 0.2 0
0.8 0.6 CS Preexposed 0.4 CX Extinction 0.2 0
Figure 7.2. Values of the CS, XCS, VCS–US, and the CR for the non-preexposed, CS preexposed, and CS preexposed followed by CX extinction cases. Y axes in arbitrary units.
Figure 7.2 shows the values of the CS, XCS, VCS–US, and the CR for the NonPreexposed (NPE), CS Preexposed (PRE), and CS Preexposed followed by CX extinction (PRE-EXT) cases. XCS is large in the NPE and PRE-EXT groups and small in the PRE group, VCS,US is large in the NPE group and small in the PRE and PRE-EXT groups, and the CR is small in the PRE group and large in the NPE and PRE-EXT groups. Behavioral properties of latent inhibition The SLG model can describe the following aspects of LI (Schmajuk et al., 1996; Schmajuk & Larrauri, 2006).
The effect of different preexposure procedures on the retardation of conditioning Experimental results show that preexposure to a given CS but not to the context or a different CS retards excitatory conditioning to that specific CS (Lubow & Moore, 1959, Experiment 1; Reiss & Wagner, 1972, Experiment 1), CS preexposure retards inhibitory conditioning to that CS (Rescorla, 1971; Reiss & Wagner, 1972, Experiment 2), and that CS–weak US presentations retard CS–strong US conditioning (Hall & Pearce, 1979). Schmajuk et al. (1996) showed that the SLG model is able to describe these results.
141
From latent inhibition to retrospective revaluation Hall and Pearce (1979) 0.6
Suppression Ratio
0.5 0.4 0.3 0.2 0.1 0 1
2
3
4
5
6
Days Simulation 0.6
Suppression Ratio
0.5 0.4 0.3 0.2 0.1 0 1
2
3
4
5
6
7
8
Trials Tone-shock
Light-shock
Tone alone
Figure 7.3. Conditioning following training with a weak US. Upper panel: suppression ratio to the Tone during Tone-strong shock conditioning, following Tone-weak shock, Light-weak shock, or Tone-alone presentations. Data from Hall & Pearce (1979). Lower panel: suppression ratio during eight CS–US conditioning trials with a strong US, after (a) 10 tone-weak US trials (Tone-shock Group), (b) 10 light-weak US trials (Light-shock Group), and (c) 10 tone-alone trials. The intertrial interval (ITI) was 200 time units (t.u.), the Light and Tone CSs lasted 20 t.u. with intensity 1, the weak US was 0.3 and the strong US was 1.8, overlapping the last 5 t.u. of the CSs. Parameter values used in the simulations are listed in Schmajuk & Larrauri (2006).
Figure 7.3 shows that the model replicates well both the basic LI effect as well as the Hall and Pearce (1979) data. According to the SLG model, LI is manifested because: (1) CS preexposure reduces Novelty0 , thereby reducing zCS and XCS (in Figure 7.2 compare XCS for CS Preexposure vs. Non-Preexposure). (2) A small XCS implies a slow rate of acquisition of VCS,US during conditioning, i.e., the storage
142
Nestor Schmajuk
(formation or read-in) of VCS,US is slower after CS preexposure (in Figure 7.2 compare VCS,US for CS Preexposure vs. Non-Preexposure). (3) A small XCS implies a small value of BUS ¼ XCS VCS,US, i.e., retrieval (activation or read-out) of VCS,US decreases after CS preexposure (in Figure 7.2 compare BUS for CS Preexposure vs. Non Preexposure). (4) A small BUS generates a small CR (in Figure 7.2 compare CR for CS Preexposure vs. Non-Preexposure). A similar explanation applies to the case of CS–weak US presentations retarding CS–strong US conditioning.
The effects of different parameters of preexposure on the strength of LI Experimental data show that CS preexposure first facilitates the CR (Fanselow, 1990) and then produces LI (Kiernan & Westbrook, 1993; Prados, 2000). Also, LI increases with increasing number of CS preexposures (Lantz, 1973, Experiment 1). Finally, LI increases with increasing CS duration (Westbrook, Bond, & Feyer, 1981, Experiment 3), increasing total CS-preexposure time (number of CS preexposures multiplied by CS duration; Ayres, Philbin, Cassidy, & Belling, 1992), increasing CS intensity (Crowell & Anderson, 1972, Experiment 1; Schnur & Lubow, 1976, Experiment 2; but see Solomon, Brennan, & Moore, 1974, Experiment 1), and increasing intertrial interval durations (Lantz, 1973, Experiment 2; Schnur & Lubow, 1976, Experiment 1; but see Crowell & Anderson, 1972; DeVietti & Barrett, 1986). Schmajuk et al. (1996) showed that the SLG model is able to describe these results. Figure 7.4 shows that the model replicates well Fanselow’s (1990) data showing that preexposure facilitates, before slowing down, the CR (freezing). According to the SLG model, all these parametric effects can be described by one simple rule: if the manipulation reduces Novelty0 it will increase LI. Specifically, the facilitation of fear acquisition by CS preexposure – an emergent property of the model – is explained because attention to the CS has first to increase before CS–US associations can be formed and freezing observed. Additional CS presentations decrease Novelty0 and attention to the CS, thereby resulting in LI.
The consequences of preexposing to different combinations of CSs Experimental results demonstrate that preexposing CS(A) and CS(B) in compound attenuates LI to CS(A) or CS(B) (overshadowing of LI; Holland & Forbes, 1980; Mackintosh, 1973; Rudy, Krauter, & Gaffuri, 1976), LI is attenuated with simultaneous but not sequential presentations of CS(A) and CS(B) (Honey & Hall, 1988, Experiment 1), preexposing CS(B) followed by simultaneous preexposure of CS(A) and CS(B) in compound tends to preserve LI to CS(A) (blocking of LI; Honey & Hall, 1988, Experiment 3), and that preexposure to CS(A)–CS(B) is less effective than alternated preexposure to separate presentations of CS(A) and CS(B) in producing LI to the CS(A)–CS(B) compound (Holland & Forbes, 1980; but see Baker,
143
From latent inhibition to retrospective revaluation Fanselow (1990) Experiment 3 70
Percent Freezing
60 50 40 30 20 10 0 0.5
9
27
162
Delay (sec) Simulation
Conditional Responses
0.015
0.01
0.005
0 20
40
60
80
100
Time Units PRE
No PRE
Figure 7.4. Conditioning as a function of the time spent in the context preceding the US presentation. Top panel: percent freezing following 0.5, 9, 27, and 162 s in the context preceding presentation of the US with (PRE) or without (No PRE) 120-s preexposure to the context. Data from Fanselow (1990, Experiment 3). Lower panel: simulated peak CR amplitude after 1 CX–US conditioning trial following 20, 40, 60, or 80 time units in the context (with CXg ¼ 0.5, and CXc ¼ 0.5) with (PRE) or without (No PRE) 20 time units of preexposure to the context. The US was 5 t.u. long and of intensity 2. Simulation parameters as above.
Haskins, & Hall, 1990, Experiment 1). In addition, it has been reported that LI is disrupted by the presentation of a surprising event (Best, Gemberling, & Johnson, 1979; Lantz, 1973, Experiment 3; Rudy, Rosenberg, & Sandell, 1977), as well as by the omission of an expected event (Hall & Pearce, 1982), following CS preexposure. Schmajuk et al. (1996) showed that the SLG model is able to describe these results.
144
Nestor Schmajuk
Figure 7.5 shows that the model replicates well the Lantz (1973) data. In this case, only the SLG model is able to describe the result because Novelty0 can be increased, not only by US presentations but also by the presentation of a CS.
The effects of contextual manipulations on the strength of LI Experimental results show that context preexposure prior to CS preexposure facilitates LI (Hall & Channel, 1985a, Experiments 1, 2, and 3), a phase of exposure to the context alone interposed between CS preexposure and conditioning in the same context might slightly decrease LI (e.g., Baker & Mercier, 1982; Wagner, 1979) or slightly facilitate LI (Hall & Minor, 1984, Experiment 5), LI is disrupted by a change in the context from the CS preexposure phase to the conditioning phase (Hall & Channel, 1985a, Experiment 3; Wickens, Tuber, & Wickens, 1983, Experiment 3) and this attenuation of LI occurs even when conditioning occurs in a context already familiar (Hall & Channel, 1985a; Hall & Minor, 1984; Lovibond, Preston, & Mackintosh, 1984). Schmajuk et al. (1996) showed that the SLG model is able to describe these results. Figure 7.6 shows that the model describes the Hall and Channel (1985a, Experiment 3) results showing that LI is attenuated by a change in CX from the preexposure to the conditioning phase.
The effect of preexposure to a pair of CSs Although CS preexposure typically yields LI, preexposure to a pair of CSs might facilitate performance on subsequent discrimination tasks. This phenomenon is known as perceptual learning. Experimental results show that CS preexposure in the rats’ home cage produced perceptual learning, whereas CS preexposure in the training environment produced LI (Channel & Hall, 1981, Experiment 1). However, preexposure to intramaze cues resulted in perceptual learning only when these cues were presented in the same context during preexposure and discrimination training (Trobalon, Chamizo, & Mackintosh, 1992). Schmajuk et al. (1996) showed that the SLG model is able to describe these results.
Orienting response (OR) and LI Kaye and Pearce (1984) showed that whereas CR acquisition is slowed down by CS preexposure, the OR during acquisition is stronger in preexposed than in notpreexposed animals. This result is well explained by the Pearce–Hall (1980) model in terms of the larger difference between the US and the weak preexposed CS–US association, than between the US and the strong non-preexposed CS–US association. Wilson et al. (1992) reported that percent ORs decreased with randomly alternated
From latent inhibition to retrospective revaluation
145
Lantz (1973) Experiment 3 Time to Completion of 10 Licks (sec)
160
120
80
40
0 Habituated Control
HabituatedDishabituated
Nonhabituated Control
Simulation
Conditioned Responses
0.1
0.08
0.06
0.04
0.02
0 Habituated Control
HabituatedDishabituated
Nonhabituated Control
Figure 7.5. Surprising events previous to conditioning. Upper panel: time to completion of 10 licks after 1 tone–US pairing following (a) 60 tone presentations (Habituated Control Group), (b) 60 tone presentations and 1 light presentation (Habituated–Dishabituated Group), or (c) exposure to the context. Data from Lantz (1973, Experiment 3). Lower panel: simulated peak CR amplitude after two tone–US conditioning trials following (a) 20 tone presentations (Habituated Control Group), (b) 20 tone presentations and 1 light CS presentation preceding the tone CS by 20 t.u. during conditioning (Habituated–Dishabituated Group), or (c) 20 exposures to the context and 1 light CS presentation preceding the tone CS during conditioning (Nonhabituated Control Group). The ITI was 500 t.u., the CSs were 20 t.u. long and of intensity 1, and the US was 5 t.u. long and of intensity 2. Simulation parameters as above.
146
Nestor Schmajuk Hall and Channell (1985) Experiment 3 1
Flap Entries
0.8 0.6 0.4 0.2 0 1
2
3
4
5
4
5
Days Simulation
Conditioned Responses
1 0.8 0.6 0.4 0.2 0 1
2
3 Blocks
PRE-SAME
PRE-DIFF
NPE
Figure 7.6. Latent inhibition is context specific. Upper panel: flap entries during the CS presentation as a ratio of prestimulus response rate. Latent inhibition is present when the light was conditioned in the context of preexposure (PRE-SAME) but not when it was conditioned in another context (PRE-DIFF) or when it was not preexposed (NPE). Data from Hall & Channell (1985b, Experiment 3). Lower panel: simulated peak CR amplitude over five blocks (of three conditioning trials) following (a) 15 CS presentations in the context of conditioning (PRE-SAME), (b) 15 CS presentations in a different context (PRE-DIFF), or (c) 15 exposures to the context (NPE). The ITI was 2000 t.u., the CSs were 20 t.u. in duration and of intensity 1, and the US was 5 t.u. in duration and of intensity 2. Simulation parameters as above.
presentation of light–tone–food trials and light–tone alone trials, but it was restored when the nonreinforced trials contained only the light. Figure 7.7 shows that the SLG model, in addition to correctly describing the Kaye and Pearce (1984) results, also reproduces the Wilson et al. (1992) data demonstrating the restoration of the OR.
147
From latent inhibition to retrospective revaluation Wilson et al. (1992) Percentage Light Orientation
1.2 1 0.8 0.6 0.4 0.2 0 2
4
6
8
10
12
14
16
18
Sessions Simulation
Percentage Light Orienting
1.2 1 0.8 0.6 0.4 0.2 0 2
4
6
8
10
12
Trials Control
Experimental
Figure 7.7. Restoration of the light orienting response. Upper panel: light orienting expressed as a percentage of a total number of responses (magazine approaching and light orienting). Percent orienting responses decreased with randomly alternated presentation of light–tone– food trials and light–tone alone trials, and it was restored when the nonreinforced trials contained only the light. Data from Wilson et al. (1992). Lower panel: orienting responses, calculated as a percentage of the strength of the conditioned responses elicited by the CS light and the internal representation of the CS light. Simulations consisted of (a) 15 presentations of CS1 (10 t.u., salience 1) preceding a CS2 (10 t.u., salience 1) which overlapped with a 5 t.u. US (strength 2), alternated with 15 nonreinforced presentations of the same CSs, and (b) 10 CS1– CS2–US presentations as before, alternated with 10 presentations of CS1 alone. The ITI was 500 t.u. and the salience of the CX was.1. Simulation parameters as above.
148
Nestor Schmajuk
Experimental data also show that LI can be impaired by contextual changes that fail to produce dishabituation of the OR, and can remain undisturbed after a period that produces dishabituation of the OR (Hall & Channell, 1985b; Hall & Schachtman, 1987). According to the SLG model, this dissociation between LI and the OR is the consequence of the delay for changes in the OR, directly related to Novelty0 , to be manifested in terms of attention to the CS, zCS.
Motivational effects on LI Experimental data show that preexposure to a CS results in LI only when the reinforcer is relevant to the motivational state in which CS preexposure was conducted. Killcross and Balleine (1996) reported that a CS preexposed when a rat is hungry (or thirsty) results in LI when the rat, when it is both hungry and thirsty, is conditioned with food US (or saline US). Computer simulations with the SLG model show that the model is able to describe this phenomenon when the motivational states of the animal are represented as “internal” contexts that are associated with their corresponding goals. When food (liquid) US is used during conditioning, Novelty0 is smaller for the CS preexposed in the hunger (thirst) state, which is associated with food (liquid), than for the CS preexposed in the thirst (hunger) state, which is associated with liquid (food). Because a smaller Novelty0 is associated with slower conditioning, the model is able to describe the experimental data showing that preexposure to a CS results in LI only when the US is relevant to the motivational state in which CS preexposure was conducted. Figure 7.8 shows that the model reproduces well Killcross and Balleine’s (1996) data.
The effects of postconditioning manipulations Experimental data show that LI is attenuated by extensive exposure to the training context in the absence of the US (Grahame, Barnet, Gunther, & Miller, 1994), by US presentations in another context (Kasprow et al., 1984), or by the post-conditioning presentation of a solution similar but not identical to the preexposed solution (Killcross, 2001). The SLG model is able to describe these results. Figure 7.9 shows that the model correctly replicates the Grahame et al. (1994) result. According to the model, extinction of the context following conditioning manipulations increases Novelty0 as well as the magnitudes of zCS, XCS, and the CR without increasing VCS,US (in Figure 7.2 compare VCS-US for CS Preexposure vs. CS Preexposure–CX Extinction). In other words, even if CS preexposure results in a decreased VCS,US value and CR magnitude during conditioning (LI, in Figure 7.2 compare VCS-US for CS Preexposure vs. Non-Preexposure), the magnitude of the CR can be increased by increasing zCS and XCS after conditioning (attenuation of LI, in Figure 7.2 compare XCS for CS Preexposure vs. CS Preexposure–CX Extinction).
From latent inhibition to retrospective revaluation
149
Killcross and Balleine (1996)
Magazine Entries
16 12 8 4 0 Relevant
Irrelevant
Conditioned Responses
Simulation 0.6
0.4
0.2
0 Relevant
Irrelevant
Figure 7.8. Role of motivation on latent inhibition. Upper panel: mean magazine entries during conditioning trials of a CS preexposed thirsty and reinforced with either saline (Relevant Group) or pellets (Irrelevant Group) during conditioning. Data from Killcross & Balleine (1996, Experiment1). Lower panel: average simulated peak CR amplitude for Relevant and Irrelevant Groups during conditioning. Simulations consisted of (a) 20 presentations of a CX (0.5) representing hunger together with a 20 t.u. CS representing food (intensity 1), alternated with 20 presentations of another CX (0.5) representing thirst together with a 20 t.u. CS representing water, (b) 10 presentations of the CX (0.5) representing hunger with a 20 t.u. CS (intensity 1), alternated with 10 presentations of the CX (0.5) representing thirst with a different 20 t.u. CS, and (c) 10 presentations of both contexts (hunger and thirst) together with the 20 t.u. CS representing saline either (a) with the CS preexposed with thirst (Relevant) or (b) with the CS preexposed with hunger (Irrelevant). The ITI was 200 t.u. Simulation parameters as above.
The effects of the passage of time LI can decrease (De la Casa & Lubow, 2000, 2002; Kraemer et al., 1991) or increase (De la Casa & Lubow, 2000, 2002; Wheeler et al., 2004) with the passage of time. This latter effect, referred to as super-LI, was seen only when (a) the number of preexposure trials was relatively large, (b) the delay period was introduced between conditioning and testing phases instead of between preexposure and conditioning phases,
150
Nestor Schmajuk Grahame et al. (1994) Mean Latency (Log s)
3
2
1
0 CONTROL
EXTINCTION
Conditioned Responses
Simulation 0.3
0.2
0.1
0 CONTROL PRE
EXTINCTION SIT
Figure 7.9. Recovery from latent inhibition. Upper panel: mean latencies to complete the first 5 cumulative seconds of drinking in the test context in Control and Extinction groups. Data from Grahame et al. (1994, Experiment 1, Figure 3). Lower panel: simulated results showing the strength of the conditioned response to the CS during one test trial, following 75 CS trials, 45 CS–US trials, and 120 CX trials. Simulation values were CS: 30 to 40 time units, salience 0.6; US: 35 to 40 time units, strength 1; Context salience 0.1, intertrial interval 600 time units. PRE: preexposed group. NPE: non-preexposed group. Simulation parameters as above.
(c) when all experimental stages were conducted in different contexts, and (d) the US was relatively strong. Figure 7.10 shows that the model correctly replicates the De la Casa and Lubow (2002) results for different US intensities and short and long delays between conditioning and testing. According to the SLG model, super-LI is not due to a further decrease in attention to the flavor in the PRE group, but to an increased attention to the stimulus representing the water and related stimuli, which becomes a conditioned inhibitor. When the US is relatively weak, introducing a delay between conditioning and testing decreases LI because attention to the water barely changes, as a result of a relatively small increment in Novelty0 . In turn, the small increase in Novelty0 is the
From latent inhibition to retrospective revaluation
151
De la Casa & Lubow (2000) Mean Saccharin Consumption (ml)
10
8
6 4
2
0 1 day
21 days
1 day
Strong US
21 days
Weak US
Simulations
Consumption
0.8
0.6
0.4
0.2
0 1 Trial
20 Trials
1 Trial
Strong US
20 Trials
Weak US NPE
PRE
Figure 7.10. Absence of super latent inhibition after a post-conditioning delay with a weak US. Upper panel: mean saccharin consumption in groups that received no preexposure (NPE) or 4 preexposure days (PRE) followed by 1 day (Short Delay) or 21 days (Long Delay) in the home cage after conditioning using a Strong or a Weak US. Data from Lubow & De la Casa (2000, Experiment 2). Lower panel: five-trial average of simulated consumption after no preexposure (NPE) or 30 preexposure trials (Long PRE) and 1 (Short Delay) or 33 Trials (Long Delay) after conditioning using a Strong or a Weak US. Simulation values were CS: 15 to 40 time units, salience 1; US: 35 to 40 time units, Strong US 1, Weak US 0.8; Home and Training context salience 0.1, and Intertrial Interval 500 time units. Simulation parameters as above.
152
Nestor Schmajuk
consequence of having only the unfulfilled expectation of the training context, the flavor and of a weak US, when the water is presented in their absence in the home cage. Because attention to the water does not increase during the delay, the water–US association becomes only barely inhibitory, and super-LI is absent. In addition, the increase in Novelty0 causes attention to the flavor to increase and, therefore, LI decreases. Latent inhibition and overshadowing counteract each other Blaisdell et al. (1998, Experiment 1) showed that combined LI and overshadowing treatments attenuate the decrease in responding produced by each procedure alone. Figure 7.11 shows that the model reproduces well the Blaisdell et al. (1998) results. Preexposure (PREþCON, LI) causes attention to the CS to decrease, and overshadowing (OVER) is the result of the competition between CSs. Responding in the Preexposure and Overshadowing (PREþOVER) group is stronger than in the PREþCON (LI) group because the introduction of CS2 during conditioning increases Novelty0 and attention to CS1, consequently increasing the CS1–US association and the magnitude of the CR. Moreover, CS1 testing in the absence of CS2 and the US increases Novelty0 more in the PREþOVER group than in the PREþCON group, further increasing attention to CS1 and the relative advantage of the CR in the PREþOVER group. Responding in the PREþOVER group is stronger than in the OVER group because, as indicated above (see Figure 7.4), a limited number of preexposure trials results in an increased attention to CS1. Therefore, at the beginning of overshadowing, attention to CS1 is larger in the PREþOVER group than in the OVER group. Because, in our model, overshadowing is the result of the competition between CS1 and CS2 to gain association with the US, an increased attention to CS1 results in an increased CS1–US association, and a larger CR to CS1 in the PREþOVER group than in the OVER group. In contrast with the above data, LI and overshadowing do not counteract, but summate in conditioned taste aversion (Nagaishi & Nakajima, 2008). In order to capture the methodological peculiarities of conditioning taste aversion, our simulations (a) included a CS representing water, (b) presented sequentially the CSs representing the flavors as it was done in the experiment, and (c) separately reinforced each flavor CS in the presence of water. Under these conditions, simulated results show less responding after LI and overshadowing than after LI or overshadowing separately. Role of attentional mechanisms in extinction and retrospective revaluation As mentioned, the SLG network originally included (a) an attentional mechanism, (b) an associative network, and (c) a feedback loop based on CS–CS associations, in order to describe specific classical conditioning paradigms. As shown above, the model proved later capable of correctly describing other experimental results, such as recovery from LI or the counteraction between LI and overshadowing, for which it was not specifically designed. In addition, Jose Larrauri and I discovered that, in the
From latent inhibition to retrospective revaluation
153
Blaisdell et al. (1998)
Mean Time (log s)
2
1.5
1
0.5
0 PRE + OVER
OVER
PRE + CON
CON
Conditioned Response
Simulation
0.3
0.2
0.1
0 PRE + OVER
OVER
PRE + CON
CON
Figure 7.11. Latent inhibition and overshadowing counteract each other. Upper panel: mean latencies to complete the first 5 cumulative seconds of drinking in the test context. Data from Blaisdell et al. (1998, Figure 5). Lower panel: simulated results showing the strength of the conditioned response to CS1 during one test trial, following 50 CS1 trials in phase 1 (PRE) or 50 CS3 trials, and 15 CS1–CS2–US or CS1–US trials in phase 2 (OVER or CON). Simulation values were CS1: 30 to 40 time units, salience 0.6; CS2: 30 to 40 time units, salience 0.75; US: 35 to 40 time units, strength 0.5; Context salience 0.1, intertrial interval 600 time units. Simulation parameters as above.
SLG model, the attentional mechanism designed to explain LI had an important role for understanding extinction (Larrauri & Schmajuk, 2008), and that the combination of the attentional mechanism and the feedback loop allowed the model to address retrospective revaluation, including backward blocking and retrospective revaluation in causal learning (Schmajuk & Larrauri, 2006). Interestingly, the feedback loop in our model provides information back to the input based on within-compound associations, which Dickinson and Burke (1996) suggested play an important role in retrospective revaluation.
154
Nestor Schmajuk
Extinction Larrauri and Schmajuk (2008) showed that many experimental results regarding extinction of classical conditioning can be explained in terms of the attentional mechanisms in the model. For example, according to the SLG model, extinction might result in the formation of inhibitory CX–US associations, which prevent the complete elimination of the excitatory CS–US association. These inhibitory CX–US associations are not detectable through summation tests (Bouton & Swartzentruber, 1989) because attention to the CS and the CX decreases. Spontaneous recovery refers to the fact that responding to the extinguished CS reappears when the CS is tested some time after extinction. Spontaneous recovery is an increasing function of extinction–test time intervals (Pavlov, 1927; Robbins, 1990), a decreasing function of the acquisition–extinction interval for long intervals (Rescorla, 2004), and decreases with repeated extinction sessions (Thomas et al., 2003). According to the SLG model, during spontaneous recovery, increases in Novelty0 (when the CS is reintroduced after a period of time) increase attention to the excitatory CS faster than attention to the inhibitory CX. Therefore, the CS–US association is activated, and a CR generated, until attention to the CX increases too. External disinhibition refers to the fact that responding to the extinguished CS reappears when a novel CS precedes the CS during testing. External disinhibition (Bottjer, 1982; Pavlov, 1927) is the result of the presentation of a novel CS before the presentation of the target CS that increases Novelty0 and attention to the remaining CS–US association, which results in the generation of CR. Renewal refers to the fact that responding to the extinguished CS reappears when the CS is tested in a context different from that of extinction. Renewal (a) is stronger following ABA and ABC procedures than in an AAB procedure (Thomas et al., 2003), and (b) is eliminated by massive extinction in the extinction context (Denniston et al., 2003). In renewal, inhibitory CX–US associations are formed during extinction but responding recovers when the context is changed during testing. Renewal is relatively strong with ABA and ABC procedures because the change from the context of acquisition (CXA) to the context of extinction (CXB) increases Novelty0 , attention to the CS, and responding during testing. Renewal decreases with an increasing number of extinction trials because attention to the CS decreases during the nonreinforced presentation of the CS. Reinstatement is the fact that responding to the extinguished CS reappears when the CS is tested in the context of extinction following US presentations in that context. Reinstatement (a) is attenuated with increasing number of extinction trials (Delamater, 1997) because prolonged extinction decreases attention to the CS and responding during testing. Reacquisition refers to conditioning that follows acquisition and extinction. Reacquisition is slow (a) with relatively many extinction trials (Ricker & Bouton, 1996), and (b) with relatively few conditioning trials (Ricker & Bouton, 1996). According to the model, reacquisition is slow when attention to the CS decreases during a prolonged extinction, or when attention to the CS is relatively small at the end of conditioning and further decreases during extinction.
From latent inhibition to retrospective revaluation
155
Backward blocking and recovery from backward blocking According to the SLG model, backward blocking (Miller & Matute, 1996; Shanks, 1985) is the consequence of the smaller Novelty0 on Aþ trials (experimental group) than on Bþ trials (control group) that follow AXþ presentations. Because the (surrogate) US retrieves the memory of X through US–X associations, attention to X is larger in the control than the experimental group. Figure 7.12 shows that the SLG model correctly describes Pinen˜o et al.’s (2005, Figure 3) data demonstrating backward blocking as well as the recovery from backward blocking after a retention interval. In addition, also in agreement with Pinen˜o et al. (2005), the model describes how a retention interval results in the recovery from forward blocking. In both cases, recovery is due to the increased representation of X when the context retrieves the memory of X and Novelty0 increases due to the absence of the US and A.
Retrospective revaluation: additivity effects on blocking Beckers et al. (2005) showed that blocking is stronger when the outcome is additive, and information regarding maximality and additivity is provided before or after elemental and compound training. In their Experiment 4, “additive” training consisted of Aþ trials (elemental training) and AXþ trials (compound training) followed by Gþ/Hþ/GHþþ alternated trials, where þþ represents a US twice as strong as þ. In “subadditive” training, elemental and compound training was preceded by Gþ/ Hþ/GHþ alternated trials. According to the SLG model, A–US and X–US associations do not change during additive or subadditive training, and the observed changes in blocking are the consequence of attentional changes. In the additive case, the increased outcome on GHþþ trials predicts the absence of X, because the outcome retrieves the memory of X but now it is not presented. Therefore, when Novelty0 increases during additive training, the representation of X is not very active and attention to X does not increase as much as in the subadditive case. In both the control and experimental cases, attention to the stimuli in the compound decreases to similar levels as Novelty0 decreases because, unlike the experimental case, in the overshadowing control case stimuli in the compound perfectly predict each other. Schmajuk and Kutlu (in preparation) showed that the model is also able to explain all the other experiments in the Beckers et al. (2005) study. Latent inhibition and schizophrenia Seidman (1983) suggested that whereas positive symptoms of schizophrenia (delusions, hallucinations, stereotypy) originate from dysfunction in limbic, midbrain, and upper brainstem regions, negative symptoms (apathy, blunted affect, social withdrawal, and avolition) originate from frontal lobe dysfunction. Among those positive symptoms, absence of LI is considered as a failure to filter out irrelevant information. Baruch et al. (1988; see also Lubow, Weiner, Schlossberg, & Baruch, 1987) reported that LI is absent in acute schizophrenics tested within the first week of
156
Nestor Schmajuk
Mean Suppression (log s)
2.5
Pineño et al. (2005)
2 1.5 1 0.5 0 Backward Blocking
Mean Suppression (log s)
1.2
Control
Simulation
1 0.8 0.6 0.4 0.2 0 Backward Blocking No Delay
Control Delay
Figure 7.12. Backward blocking. Upper panel: mean latencies to complete the first 5 cumulative seconds of drinking in the presence of the backward blocked stimulus. Data from Pinen˜o et al. (2005, Figure 3). Lower panel: simulated results showing the strength of the conditioned response to X during one test trial, following 100 X–A–O trials, 20 A–O trials, 150 O–US trials, and 20 delay trials. Simulation values were X: 30 to 40 time units, salience 1; A, B: salience 0.6; O: salience 0.6; US: 35 to 40 time units; strength 1; context saliences 0.1/0.05/ 0.05; intertrial interval 600 time units.
the beginning of a schizophrenic episode but it is present in chronic, medicated schizophrenics (Figure 7.13, top panel). Schmajuk et al. (2000) examined in the rat eyeblink response preparation the effect of haloperidol administration on the impairment of LI produced by aspiration lesions of the hippocampus. Although lesioned rats injected with saline did not show LI, the phenomenon was reinstated in HFL animals receiving haloperidol injections (Figure 7.13, middle panel). Figure 7.13 (bottom panel) shows that simulations with the SLG model match well the experimental results both in schizophrenic patients and hippocampally lesioned rats. In the
From latent inhibition to retrospective revaluation
157
Figure 7.13. Latent inhibition in schizophrenia, hippocampal lesioned animals, and the SLG neural network model. Top panel: experimental results from Baruch et al. (1988). Cumulative percent of subjects responding to the CS on the first 20 trials of the experiment for NORMAL, SCHIZO-ACUTE, and SCHIZO-CHRONIC groups. Middle panel: experimental results from Schmajuk et al. (2000). Percent CRs on the 4th day of conditioning for SL-SAL, HFL-SAL, SL-HAL, and HFL-HAL groups under PE and NPE conditions. Bottom panel: simulated results from Schmajuk et al. (2000). Simulated percent CRs on the 4th day of conditioning for SL-SAL, HFL-SAL, SL-HAL, and HFL-HAL groups under PE and NPE conditions. SCHIZO, schizophrenic patients; NORMAL, normal subjects; SL, sham lesion; HFL, hippocampal formation lesion; SAL, saline; HAL, haloperidol; PE, CS-preexposed groups; NPE, non-preexposed groups. From Schmajuk et al. (2000).
158
Nestor Schmajuk
model, hippocampal lesions were represented by assuming that CX–CS and CS–CX associations cannot be formed, which results in increased Novelty0 , attention to the CS, and impaired LI. The effect of haloperidol administration was represented by decreasing the effect of Novelty0 (assumed to be coded by dopamine) on attention, with the concomitant normalization of LI.
Conclusion Over the last years, we found that an attentional–associative model (Schmajuk et al., 1996), originally designed to account for some basic classical conditioning paradigms, with a concentration on LI, is able to describe many other experimental results, such as the multiple properties of extinction, backward blocking, and retrospective additive effects on blocking. In all these cases, an emergent property of the model – the fact that it can modify attention to absent CSs when they are predicted by other CSs – plays an important role. Furthermore, by introducing changes in the computation of some of its variables, the model is able to capture the impairing effects of hippocampal lesions and schizophrenia on LI, as well as the amelioration of those effects by drug administration.
Acknowledgements The author is grateful to Gonzalo De la Casa and Oskar Pinen˜o for their comments on an early version of the manuscript.
References Ayres, J. J. B., Philbin, D., Cassidy, S., & Belling, L. (1992). Some parameters of latent inhibition. Learning and Motivation, 23, 269–287. Baker, A. G., Haskins, C. E., & Hall, G. (1990). Stimulus generalization decrement in latent inhibition to a compound following exposure to the elements of the compound. Animal Learning & Behavior, 18, 162–170. Baker, A. G., & Mercier, P. (1982). Extinction of the context and latent inhibition. Learning and Motivation, 13, 391–416. Baruch I., Hemsley D., & Gray J. A. (1988). Differential performance of acute and chronic schizophrenics in a latent inhibition task. Journal of Nervous and Mental Diseases, 176, 598–606. Beckers, T., De Houwer, J., Pinen˜o, O., & Miller, R. R. (2005). Outcome additivity and outcome maximality influence cue competition in human causal learning. Journal of Experimental Psychology: Learning, Memory, and Cognition, 31, 238–49. Best, M. R., Gemberling, G. A., & Johnson, P. E. (1979). Disrupting the conditioned stimulus preexposure effect in flavor aversion learning: effects of interoceptive distractor manipulations. Journal of Experimental Psychology: Animal Behavior Processes, 5, 321–334.
From latent inhibition to retrospective revaluation
159
Blaisdell, A., Bristol, A., Gunther, L., & Miller R. (1998). Overshadowing and latent inhibition counteract each other: further support for the comparator hypothesis. Journal of Experimental Psychology: Animal Behavior Processes, 24, 335–251. Bottjer, S. W. (1982). Conditioned approach and withdrawal behavior in pigeons: effects of a novel extraneous stimulus during acquisition and extinction. Learning and Motivation, 13, 44–67. Bouton, M. E., & Swartzentruber, D. (1989). Slow reacquisition following extinction: context, encoding, and retrieval mechanisms. Journal of Experimental Psychology: Animal Behavior Processes, 15, 43–53. Buhusi, C. V., Gray, J. A., & Schmajuk, N. A. (1998). The perplexing effects of hippocampal lesions on latent inhibition: a neural network solution. Behavioral Neuroscience, 112, 316–351. Channell, S., & Hall, G. (1981). Facilitation and retardation of discrimination learning after exposure to the stimuli. Journal of Experimental Psychology: Animal Behavior Processes, 7, 437–46. Crowell, C. R., & Anderson, D. C. (1972). Variations in intensity, interstimulus interval, and interval between preconditioning CS exposures and conditioning with rats. Journal of Comparative and Physiological Psychology, 79, 291–298. De la Casa, L., & Lubow, R. E. (2000). Super-latent inhibition with delayed conditioned taste aversion testing. Animal Learning & Behavior, 28, 389–399. De la Casa, L., & Lubow, R. E. (2002). An empirical analysis of the super-latent inhibition effect. Animal Learning & Behavior, 30, 112–120. Delamater, A. R. (1997). Selective reinstatement of stimulus-outcome associations. Animal Learning & Behavior, 25, 400–412. Delamater, A. R. (2004). Experimental extinction in Pavlovian conditioning: behavioural and neuroscience perspectives. Quarterly Journal of Experimental Psychology, 57B, 97–132 Denniston, J. C., Chang, R. C., & Miller, R. R. (2003). Massive extinction treatment attenuates the renewal effect. Learning and Motivation, 34, 68–86. DeVietti, T. L., & Barrett, O. V. (1986). Latent inhibition: no effect of intertrial interval of the preexposure trials. Bulletin of the Psychonomic Society, 24, 453–455. Dickinson, A., & Burke, J. (1996). Within-compound associations mediate the retrospective revaluation of causality judgements. The Quarterly Journal of Experimental Psychology, 49B, 60–80. Fanselow, M. S. (1990). Factors governing one-trial contextual conditioning. Animal Learning & Behavior, 18, 264–270. Grahame, N. J., Barnet, R. C., Gunther, L. M., & Miller, R. R. (1994). Latent inhibition as a performance deficit resulting from CS-context associations. Animal Learning & Behavior, 22, 395–408. Hall, G., & Channel, S. (1985a). Latent inhibition and conditioning after preexposure to the training context. Learning and Motivation, 16, 381–481. Hall, G., & Channel, S. (1985b). Differential effects of contextual change on latent inhibition and on the habituation of an orienting response. Journal of Experimental Psychology: Animal Behavior Processes, 11, 470–481. Hall, G., & Minor, H. (1984). A search for context-stimulus associations in latent inhibition. The Quarterly Journal of Experimental Psychology, 36B, 145–169. Hall, G., & Pearce, J. M. (1979). Latent inhibition of a CS during CS-US pairings. Journal of Experimental Psychology: Animal Behavior Processes, 5, 31–42.
160
Nestor Schmajuk
Hall, G., & Pearce, J. M. (1982). Restoring the associability of a preexposed CS by a surprising event. The Quarterly Journal of Experimental Psychology, 34B, 127–140. Hall, G., & Schachtman, T. R. (1987). Differential effects of a retention interval on latent inhibition and the habituation of an orienting response. Animal Learning & Behavior, 15, 76–82. Holland, P. C., & Forbes, D. R. (1980). Effects of compound or element preexposure on compound flavor aversion conditioning. Animal Learning & Behavior, 8, 199–203. Honey, R. C., & Hall, G. (1988). Overshadowing and blocking procedures in latent inhibition. The Quarterly Journal of Experimental Psychology, 49B, 163–186. Kasprow, W. J., Catterson, D., Schachtman, T. R., & Miller, R. R. (1984). Attenuation of latent inhibition by post-acquisition reminder. Quarterly Journal of Experimental Psychology, 36B, 53–63. Kaye, H., & Pearce, J. M. (1984). The strength of the orienting response during Pavlovian conditioning. Journal of Experimental Psychology: Animal Behavioral Processes, 10, 90–109. Kiernan, M. J., & Westbrook, R. F. (1993). Effects of exposure to a to-be-shocked environment upon the rat’s freezing response: evidence for facilitation, latent inhibition, and perceptual learning. Quarterly Journal of Experimental Psychology, 46B, 271–288. Killcross, S. (2001). Loss of latent inhibition in conditioned taste aversion following exposure to a novel flavour before test. Quarterly Journal of Experimental Psychology, 54B, 271–288. Killcross, S., & Balleine, B. (1996). Role of primary motivation in stimulus preexposure effects. Journal of Experimental Psychology: Animal Behavior Processes 96, 32–42. Kraemer, P. J., Randall, C. K., & Carbary, T. J. (1991). Release from latent inhibition with delayed testing. Animal Learning & Behavior, 19, 139–145. Lantz, A. E. (1973). Effect of number of trials, interstimulus interval, and dishabituation on subsequent conditioning in a CER paradigm. Animal Learning & Behavior, 1, 273–277. Larrauri, J. A., & Schmajuk, N. A. (2008). Attentional, associative, and configural mechanisms of extinction. Psychological Review, 115, 640–676. Lovibond, P. F., Preston, G. C., & Mackintosh, N. J. (1984). Context specificity of conditioning, extinction, and latent inhibition. Journal of Experimental Pyschology: Animal Behavior Processes, 10, 360–375. Lubow, R. E., & Moore, A. U. (1959). Latent inhibition: the effect of non-reinforced preexposure to the conditional stimulus. Journal of Comparative and Physiological Psychology, 52, 415–419. Lubow, R. E., Weiner, I., Schlossberg, A., & Baruch, I. (1987). Latent inhibition and schizophrenia. Bulletin of the Psychonomic Society, 25, 464–467. Mackintosh, N. J. (1973). Stimulus selection: learning to ignore stimuli that predict no change in reinforcement. In R. A. Hinde & J. S. Hinde (Eds.), Constraints of Learning. London: Academic Press. Miller, R., & Matute, H. (1996). Biological significance in forward and backward blocking: resolution of a discrepancy between animal conditioning and human causal judgement. Journal of Experimental Psychology: General, 125, 370–386.
From latent inhibition to retrospective revaluation
161
Nagaishi, T., & Nakajima, S. (2008). Further evidence for the summation of latent inhibition and overshadowing in rats’ conditioned taste aversion. Learning and Motivation, 39, 221–242. Pavlov, I. P. (1927). Conditioned Reflexes. London: Oxford University Press. Pearce, J., & Hall, G. (1980). A model for Pavlovian conditioning: variations in the effectiveness of conditioned but not unconditioned stimuli. Psychological Review, 87, 332–352. Pinen˜o, O., Usushihara, K., & Miller, R. R. (2005). Spontaneous recovery from forward and backward blocking. Journal of Experimental Psychology: Animal Behaviour Processes, 31, 172–83. Prados J. (2000). Effects of varying the amount of preexposure to spatial cues on a subsequent navigation task. Quarterly Journal of Experimental Psychology, 53B, 139–148. Reiss, S., & Wagner, A. R. (1972). CS habituation produces a “latent inhibition effect” but no active “conditioned inhibition”. Learning and Motivation, 3, 237–245. Rescorla, R. A. (1971). Summation and retardation tests of latent inhibition. Journal of Comparative and Physiological Psychology, 75, 77–81. Rescorla, R. A. (2004). Spontaneous recovery varies inversely with the trainingextinction interval. Learning & Behavior, 32, 401–408. Rescorla, R. A., & Wagner, A. (1972). A theory of Pavlovian conditioning: variations in the effectiveness of reinforcement and non-reinforcement. In A. H. Black & W. F. Prokasy (Eds.), Classical Conditioning II: Current Research and Theory. New York: Appleton-Century-Crofts, pp. 64–99. Ricker, S. T., & Bouton, M. E. (1996). Reacquisition following extinction in appetitive conditioning. Animal Learning & Behavior, 24, 423–436. Robbins, S. J. (1990). Mechanisms underlying spontaneous recovery in autoshaping. Journal of Experimental Psychology: Animal Behavior Processes, 16, 235–249. Rudy, J. W., Krauter, E. E., & Gaffuri, A. (1976). Attenuation of latent inhibition by prior preexposure to another stimuli. Journal of Experimental Psychology: Animal Behavior Processes, 2, 235–247. Rudy, J. W., Rosenberg, L., & Sandell, J. H. (1977). Disruption of taste familiarity effect by novel exteroceptive stimulation. Journal of Experimental Psychology: Animal Behavior Processes, 88, 665–669. Schmajuk, N. A. (2001). Hippocampal dysfunction in schizophrenia. Hippocampus, 11, 599–613. Schmajuk, N. A. (2002). Latent Inhibition and its Neural Substrates. Norwell, MA: Kluwer Academic. Schmajuk, N. A., Buhusi, C. V., & Gray, J. A. (1998). The pharmacology of latent inhibition: a neural network approach. Behavioural Pharmacology, 9, 711–730. Schmajuk, N. A, Christiansen, B. A., & Cox, L. (2000). Haloperidol reinstates latent inhibition impaired by hippocampal lesions: data and theory. Behavioral Neuroscience, 114, 659–670. Schmajuk, N. A., Cox, L., & Gray, J. A. (2001). The neurophysiological substrates of latent inhibition: a neural network approach. Behavioral Brain Research, 118, 123–141. Schmajuk, N. A., Gray, J. A., & Larrauri, J. A. (2005). A pre-clinical study showing how dopaminergic drugs administered during pre-exposure can impair or facilitate latent inhibition. Psychopharmacology, 177, 272–279.
162
Nestor Schmajuk
Schmajuk, N. A., Lam, Y. W., & Gray, J. A. (1996). Latent inhibition: a neural network approach. Journal of Experimental Psychology: Animal Behavior Processes, 22, 321–349. Schmajuk, N. A., & Larrauri, J. A. (2006). Experimental challenges to theories of classical conditioning: application of a computational model of storage and retrieval. Journal of Experimental Psychology: Animal Behavior Processes, 32, 1–20. Schmajuk, N. A., & Moore, J. (1988). The hippocampus and the classically conditioned nictitating membrane response: a real-time attentional-associative model. Psychobiology, 16, 20–35. Schnur, P., & Lubow, R. E. (1976). Latent inhibition: the effects of ITI and CS intensity during preexposure. Learning and Motivation, 7, 540–550. Seidman, L. J. (1983). Schizophrenia and brain dysfunction: an integration of recent neurodiagnostic findings. Psychological Bulletin, 94, 195–238. Shanks, D. R. (1985). Forward and backward blocking in human contingency judgment. Quarterly Journal of Experimental Psychology, 37B, 1–21. Solomon, P. R., Brennan, G., & Moore, J. W. (1974). Latent inhibition of the rabbit’s nictitating membrane response as a function of CS intensity. Bulletin of the Psychonomic Society, 4, 445–448. Thomas, B. L., Larsen, N., & Ayres, J. J. B. (2003). Role of context similarity in ABA, ABC, and AAB renewal paradigms: implications for theories of renewal and for treating human phobias. Learning and Motivation, 34, 410–436. Trobalon, J. B., Chamizo, V. D., & Mackintosh, N. J. (1992). Role of context in perceptual learning in maze discriminations. The Quarterly Journal of Experimental Psychology, 44B, 57–73. Wagner, A. R. (1979). Habituation and memory. In A. Dickinson & R. A. Boakes (Eds.), Mechanisms of Learning and Motivation. Hillsdale, NJ: Lawrence Erlbaum. Wagner, A. R. (1981). SOP: a model of automatic memory processing in animal behavior. In N. E. Spears and R. R. Miller (Eds.), Information Processing in Animals: Memory Mechanisms. Hillsdale, NJ: Erlbaum, pp. 5–47. Westbrook, R. F., Bond, N. W., & Feyer, A -M. (1981). Short-and long-term decrements in toxicosis-induced odor-aversion learning: the role of duration of exposure to an odor. Journal of Experimental Psychology: Animal Behavior Processes, 7, 362–381. Wheeler, D. S., Stout, S. C., & Miller, R. R. (2004). Interaction of retention interval with CS-preexposure and extinction treatments: symmetry with respect to primacy. Learning & Behavior, 32, 335–347. Wickens, C., Tuber, D. S., & Wickens, D. D. (1983). Memory for the conditioned response: the proactive effect of preexposure to potential conditioning stimuli and context change. Journal of Experimental Psychology: General, 112, 41–57. Wilson, P. N., Boumphrey, P., & Pearce, J. M. (1992). Restoration of the orienting response to a light by a change in its predictive accuracy. The Quarterly Journal of Experimental Psychology, 44B, 17–36.
8 Latent inhibition and habituation: evaluation of an associative analysis Robert C. Honey, Mihaela D. Iordanova and Mark Good
Summary Latent inhibition and habituation are behavioural phenomena that result from the same simple procedure: when an animal receives exposure to a stimulus, the unconditioned responses that were once provoked by that stimulus diminish or habituate, and subsequently the rate at which conditioned responding can be established to the stimulus is reduced. One parsimonious interpretation of these phenomena is that they reflect the operation of the same attentional process: a novel stimulus attracts attention, provokes responding and is rapidly learnt about; exposure to a stimulus produces a decline in attention and concomitant reductions in responding and learning. Here, we present an integrative review of recent research that provides support for a specific account of habituation and latent inhibition that is grounded in a formal analysis of the mnemonic processes involved in associative learning. This analysis also provides one interpretation of the role of the hippocampus, a structure traditionally associated with both mnemonic and attentional processes, in habituation and latent inhibition.
Introduction Latent inhibition (LI) is a phenomenon in both senses of the word: it is a very simple and ubiquitous behavioural effect, and it has stimulated intense empirical and theoretical interest, and some controversy over a protracted period. After a faltering start (see Lubow, this volume), the observation first reported by Lubow and Moore (1959) has become a central concern for theoretical analyses of associative learning: once its generality had been established, LI became the territory of models of associative learning wherein variations in attention play a key role (e.g., Lubow, 1989; Mackintosh, 1975; Pearce & Hall, 1980). Indeed, as is often the case in psychology, labels for phenomena (e.g., LI) and the processes assumed to underlie them (in this case attention) have come to be used, if not interchangeably, then with little Latent Inhibition: Cognition, Neuroscience and Applications to Schizophrenia, ed. R. E. Lubow and Ina Weiner. Published by Cambridge University Press # Cambridge University Press 2010.
163
164
Robert C. Honey, Mihaela D. Iordanova and Mark Good
acknowledgement of the fact that there remains no real agreement concerning how LI is to be explained; or indeed whether there is a viable unitary explanation for it. There have been advances in our understanding of LI over the last 50 years (e.g., Hall, 1991; Lubow, 1989; McLaren & Mackintosh, 2000), but one enduring theoretical issue revolves around the relationship between LI and another simple phenomenon, habituation. Habituation and LI result from the same basic procedure (repeated stimulus exposure) and habituation was the focus of intensive analysis long before LI was discovered. It is the origin of, and relationship between, habituation and LI that is the central concern of the present chapter, where we will draw upon evidence from both behavioural research and investigations of the neural mechanisms that underlie habituation and LI.
Habituation When the behaviour of an animal is recorded across a series of repeated exposures to a stimulus, one can often identify a response that will be provoked by the stimulus when it is novel and that becomes less evident or habituates over successive stimulus presentations. For example, when rats first encounter a localized light in an operant chamber they often move towards the light and explore it; that is, presentation of the light results in a behavioural orienting response (OR). Over a number of presentations of the light, both within a single experimental session and across days, this orienting response (OR) habituates, but when the rat is presented with a different light the OR is restored (i.e., dishabituation occurs). Like LI, habituation has been the subject of intense interest from a variety of perspectives: from simple behavioural analysis, through neuroscience, to psychopathology (e.g., Gray, 1982; Hall, 1991; Hawkins & Kandel, 1984; Horn & Hill, 1964; Konorski, 1967; Mackintosh, 1987; Sokolov, 1963). Habituation and LI are the result of the same procedure, and clearly while this need not mean that they reflect the operation of a common mechanism, it would be very surprising if there was no overlap between their origins.
An associative analysis of habituation and latent inhibition Wagner’s standard operating procedure (SOP) model of associative learning (Wagner, 1981) includes within it the first formal attempt to provide an integrated account of habituation-related phenomena and LI; one alternative basis for the SOP acronym, a sometimes opponent process model, reveals some of its roots (e.g., Solomon & Corbit, 1974) and the fact that it provides a coherent analysis for complex, sometimes opposing patterns of results (see Donegan, 1981). Before describing this model and evaluating it in more detail, it is perhaps helpful to summarize what has unfolded over the intervening 20 or so years and to foreshadow aspects of the remainder of this chapter. Thus, close on the heels of one of the
Latent inhibition and habituation: evaluation of an associative analysis
165
successes of the SOP model, its principled account for why LI should be contextspecific (e.g., Lubow, Rifkin & Alek, 1976; Lovibond, Preston & Mackintosh, 1984), came evidence that, if not fatal, then was certainly embarrassing to SOP: habituation survived a change in context that disrupted LI (e.g., Hall & Channell, 1985; Hall & Honey, 1989). This surprising observation has suggested to some that habituation and LI might be tapping fundamentally different processes: with habituation being a rather primitive process, reflecting the decline in the efficacy of the (untrained) links between the processes activated by the presentation of the stimulus and those responsible for generating the response (e.g., Groves & Thompson, 1970; Hawkins & Kandel, 1984; Horn & Hill, 1964); and LI involving a reduction in attention to the stimulus (Lubow, 1989; Mackintosh, 1975; Pearce & Hall, 1980). Of course, while the apparent dissociation of habituation and LI is a distinct problem for Wagner’s model, the fact that LI is context-specific was really only predicted by Wagner’s model: if the change of context does not alter the animal’s perception of the preexposed stimulus or the representation that was activated by the stimulus, then why else should changing the context between the exposure stage and conditioning influence the attention paid to the stimulus? However, as we will also see, more recent evidence has revealed that habituation, like LI, can be context-specific under some conditions (Honey & Good, 2000a; Honey, Good, & Manser, 1998; Honey, Watt & Good, 1998; Jordan, Strasser, & McHale, 2000). These observations provide some hope that a parsimonious explanation for aspects of LI and habituation is viable.
The SOP model Wagner’s (1981) model of associative learning includes an analysis for both shortterm and long-term changes in stimulus processing. The analysis of long-term habituation and LI, the focus of interest here, is based upon the observation that exposure to a stimulus does not happen in a vacuum, but rather it ordinarily occurs in a specific experimental context. For example, a typical study of LI in rats would consist of an exposure stage in which the stimulus (e.g., a light or tone) is presented in an operant chamber, prior to trials on which that stimulus is paired with an event of motivational significance (e.g., food) in the same chamber. With this description borne in mind, a LI procedure provides the opportunity for the formation of an association between the experimental context and the target stimulus. Wagner (1981) assumes that once this association is formed, placement in the experimental context will activate a memory of the stimulus and this fact influences both performance and learning when the stimulus is again presented in the experimental context. More specifically, in the case of habituation, the influence of the context is assumed to be on performance – making it less likely that the presentation of the stimulus will provoke an unconditioned response. However, in the case of LI the influence that the context has is both on learning (e.g., forming an association between the
166
Robert C. Honey, Mihaela D. Iordanova and Mark Good
preexposed stimulus and a US) and on performance. It is assumed that the expression of what has been learnt about the preexposed stimulus during conditioning will be less evident when the stimulus is presented in the preexposed context than when it is presented elsewhere. To appreciate the sometimes complex nature of the model, its predictions, and behavioural results (e.g., Donegan, 1981), one first needs to unpack some of the model’s assumptions.
The nature of mnemonic activity SOP begins by assuming that the presentation of a stimulus activates a set of representational elements that together constitute the memory for that stimulus. When the memory of a given stimulus is inactive, presentation of that stimulus provokes a proportion of its elements to move from an inactive state (I) to a primary state of activation (A1; see arrow 1 in Figure 8.1). An inactive memory is considered to be in “long-term store” and the A1 state can be allied to the focus of “attention” or “rehearsal”. It is assumed that once in the A1 state, elements decay rapidly to a secondary state of activation (A2; see arrow 2d, d for decay, in Figure 8.1), a state that is allied to being in “short-term store”. Finally, these elements decay into their original inactive state (see arrow 3 in Figure 8.1). It is assumed that a given element cannot be in two states at the same time and that elements in the A2 state cannot move directly into the A1 state. "Short-term store"
A2 3
2d
2a
I
A1 1
"Long-term store"
"Rehearsal"
Figure 8.1. The states of activation (I, A1, and A2) in which elements of a memory can reside according to Wagner’s SOP model. The arrows indicate the permissible transitions between the three states: The presentation of a stimulus results in the movement of a proportion of its memorial elements from the Inactive state (in “Long-term store”) to the A1 state (“rehearsal”; see arrow 1), from which they first decay into the A2 state (see arrow 2d), and finally into the Inactive state (see arrow 3). If the elements of one memory become active, then a proportion of the elements of others with which it has an association will enter the A2 state (see arrow 2a; adapted from Wagner, 1981).
Latent inhibition and habituation: evaluation of an associative analysis
167
Association formation SOP supposes that during a pairing of any two stimuli (e.g., an experimental context and a stimulus), if a large proportion of the elements of both stimuli are concurrently in the A1 state, then an excitatory association would develop between the two memories – an association that lies dormant until such time as one of the memories becomes active (see Wagner & Brandon, 2001, for a version of SOP, C-SOP, where the elements or components of a stimulus can acquire independent associative strengths). Once the memory of the experimental context has become active, the context–stimulus association will engender activity in a proportion of the elements of the stimulus – specifically, it is assumed that a proportion of the elements of the stimulus will move from the inactive state to the A2 state of activation (see arrow 2a, a for association, in Figure 8.1).
Modulation of performance: habituation, sensitization and latent inhibition In order to explain habituation and related phenomena, it is clearly necessary for mnemonic activity to affect performance. The preceding paragraph makes it clear that during the presentation of a stimulus the proportions of its memorial elements that are in the two active states (A1 and A2) will depend on whether or not that stimulus has been recently presented or preceded by a context with which it has an association. When a stimulus has not been recently presented, none of its elements will be in the A2 state and it is assumed that its intensity will determine the proportion of its elements (p1) that will be provoked into the A1 state. However, when a stimulus has been recently presented, a proportion of its elements (pA2) will have decayed into the A2 state before that stimulus has been re-presented; since elements in the A2 state cannot move directly into the A1 state, under such conditions the presentation of the stimulus will be capable of provoking A1 activity only in those of its elements that are still inactive. To be specific, the proportion of the elements of a stimulus that will be provoked into A1 will equal p1(1pA2). Similarly, when a stimulus is not presented in a context in which it has been preexposed, none of its elements will be in the A2 state and its intensity will determine the proportion of its elements (p1) that will be provoked into the A1 state. However, when the stimulus is presented in a context with which it has an association (i.e., the context in which it was preexposed), a proportion of its elements (pA2) will have been placed into the A2 state before that stimulus has been presented. SOP assumes that the vigour of responding (unconditioned or conditioned) elicited by the presentation of the stimulus will reflect the combined effect of (1) the proportion of its elements that have been placed in the A1 state, p1(1pA2), and (2) the proportion that the prime has placed in the A2 state, pA2. Within the model, a separate weighting factor determines the influence on performance of A1 and A2 elements (W1 and W2, respectively). Therefore, if A1 activity is more effective at
168
Robert C. Honey, Mihaela D. Iordanova and Mark Good
generating performance than A2 activity (i.e., W1 > W2), and an unannounced presentation of a stimulus produces A1 activity in all of its elements (i.e., p1 ¼ 1.00), then presenting a stimulus in a context with which it has an association will affect a reduction in the behaviour that can be elicited by the presentation of a stimulus. As a result, the amount of unconditioned responding provoked by a stimulus will be reduced (i.e., habituation will be observed), and any conditioned responding that a stimulus could have elicited (as a result of becoming associated with a US) will be less apparent (i.e., LI will be observed) in a context in which it has been preexposed than in a different context. Of course, if the conditions described above are not met, then a predicted stimulus could elicit more responding; that is, sensitization might be observed. In effect, the associatively generated activity in the memory of a stimulus can add to that provoked by the stimulus itself. More formally, if p1 is less than 1, and pA2 is greater than 0, then the number of elements that will be active when a stimulus is presented in the same context as it was preexposed ([p1(1pA2)] þ pA2) will be greater than when it is presented in a different context (p1(1pA2)). We shall have cause to return to this explanation for sensitization when considering the neural mechanisms that underlie LI and habituation.
Modulation of learning: latent inhibition We have already seen that SOP predicts that the formation of a context–stimulus association can affect the ability of a stimulus to provoke unconditioned and conditioned responding; thereby explaining long-term habituation and providing one mechanism for LI. However, the model also supposes that there is a second source of LI. As we have already noted, SOP assumes that the formation of an excitatory association between one stimulus and another depends upon their presentations generating A1 activity in their representative nodes. The preceding paragraphs should make it clear that the development of an association between one stimulus (e.g., a light) and another (e.g., food) will occur less readily when the light is presented in a context in which it has been preexposed, because the memory node for the light is more likely to be in the A2 state when that light is paired with food.
Evaluation of SOP The general utility and dominant position that SOP occupies as a model of associative learning is well documented (e.g., Hall, 1991), and it has been developed in a number of novel and interesting ways since its publication (e.g., Wagner & Brandon, 1989, 2001). However, for the present purposes, our evaluation of SOP will be restricted to its analysis of LI and habituation, and in particular to the influence that context– stimulus associations are presumed to exert on performance and learning. We will then consider converging evidence from studies of the neural bases of LI and habituation.
Latent inhibition and habituation: evaluation of an associative analysis
169
Contextual specificity of LI The fact that LI is context-specific is now very well established in a variety of procedures. For example, in a fairly typical study, rats might be given several days of preexposure where one auditory stimulus (X; e.g., a light) is presented in one context (A; e.g., an operant chamber with a specific odour and ambient noise level), but not in a second context (B; e.g., a second operant chamber with a different odour and noise level), and then receive conditioning trials in which X is paired with food in contexts A and/or B. Under these conditions, conditioned responding to X (in the form of approach behaviour directed towards the food well) is less evident in context A (the same context as preexposure) than in B (a different context; see Figure 8.2, Hall & Channell, 1985; for additional examples, see Hall & Honey, 1989; Lovibond, Preston, & Mackintosh, 1984). SOP provides a principled account for this observation: stimulus X will enter into association with food less readily in context A than in context B (through a difference in learning), and any associative strength that accrues to X will be less evident in A than B (through a difference in performance). The view that the preexposure context does influence performance has received some support from a related procedure: once rats have received preexposure to X in context A, but not in B, conditioned responding can be established in a third context, C, and then conditioned responding assessed in A and B. Using this procedure, Holt and Maren (1999) showed that conditioned freezing was less evident in A than B (see Figure 8.3; see also Bouton & Swartzentruber, 1989). This observation is consistent with the view that when rats are placed in context A, the memory of X is provoked into the A2 state, which results in the presentation of X being unable to generate A1
Appetitive conditioning
1.00 0.80 0.60 0.40 Same 0.20
Different 1
2 3 4 Days conditioning
5
Figure 8.2. Context-specific latent inhibition. Acquisition of conditioned appetitive responding to a localized light (X) in the same context as it was preexposed (A) and a different, but familiar context in which it was not preexposed (B). Scores above 0.50 indicate that approach to the food well was greater during the light than in the period immediately preceding the light (adapted from Hall & Channell, 1985).
170
Robert C. Honey, Mihaela D. Iordanova and Mark Good 100
Freezing (%)
80 60 40 20 0 Control
Different
Same
Figure 8.3. Contextual control of performance to a stimulus by the context in which it was preexposed. Mean levels of fear (as measured by the freezing behaviour of rats) to a novel stimulus (Control) and to a preexposed stimulus as a function of whether it was presented in the same context as it had been preexposed (Same: A) or another familiar context (Different: B). Conditioned freezing was established in context C (adapted from Holt & Maren, 1999).
activity, thereby limiting conditioned performance (i.e., freezing; see also Honey, Hall, & Bonardi, 1993; Kaye & Mackintosh, 1990). The fact that knowledge about where the preexposed stimulus was presented modulates performance to that stimulus (Bouton & Swartzentruber, 1989; Holt & Maren, 1999) leaves us in something of a quandary that is bound to strike some as paradoxical: where then is the evidence that LI, a retardation in the development of conditioned responding to a preexposed stimulus, reflects anything other than a performance effect; or a failure to retrieve what has been successfully acquired (see Kasprow, Catterson, Schachtman, & Miller, 1984). The suggestion that stimulus preexposure affects learning, as well as performance, receives some support from a simple observation involving an effect, overshadowing, first reported by Pavlov (1927). During an overshadowing procedure two stimuli, often with different intensities, are presented together and paired with a US. This procedure usually results in the less intense stimulus of the two being less able to provoke conditioned responding than if it had been conditioned in isolation (i.e., without the more intense stimulus). According to SOP, overshadowing reflects a limit on the acquisition of associative strength by the overshadowed stimulus (see also Mackintosh, 1975; Pearce & Hall, 1980; Rescorla & Wagner, 1972). The issue of interest is whether preexposure to the overshadowing stimulus changes its capacity to overshadow learning about the less intense target stimulus. On the one hand, if LI solely reflects a performance effect, with a preexposed stimulus and a novel stimulus being learnt about with equal readiness, then both a preexposed and a novel stimulus should be equally able to overshadow a further stimulus. On the other hand, if LI reflects, at least in part, an acquisition failure then a preexposed stimulus should be less likely to limit the
Latent inhibition and habituation: evaluation of an associative analysis
171
Conditioned Suppression
0.25
0.20
0.15
0.10
0.05
N+L
N
Control
N+L
N LI
Figure 8.4. Interaction between latent inhibition and overshadowing. Mean conditioned suppression ratios to a noise after trials in which it had either been paired with shock in isolation (N) or with a light (N þ L). For the control groups, the light has not been preexposed, and for the LI group the light had been preexposed prior to conditioning; scores closer to 0 indicate greater levels of fear than those closer to 0.50 (adapted from Carr, 1974).
acquisition of associative strength by another stimulus in an overshadowing procedure. Carr (1974) confirmed that preexposure to a light disrupted its ability to overshadow conditioning to a noise. Figure 8.4 shows conditioned suppression scores to a noise following conditioning trials with either the noise alone (N) or following conditioning with a compound of the noise with a light (N þ L). Inspection of the left-hand pair of bars shows that that in the Control groups there is greater conditioned fear (i.e., lower conditioned suppression scores) after conditioning with the noise alone than after compound conditioning: a novel light produced overshadowing. The right-hand pair of bars shows that this overshadowing of the noise by the light was abolished in rats from the LI groups who were given preexposure to the light (see also Dwyer & Honey, 2007; Wheeler & Miller, 2007). The aforementioned results, taken at face value, are consistent with the suggestion that LI reflects, at least in part, an acquisition deficit (cf. Holt & Maren, 1999).1
Contextual specificity of habituation The issue of whether or not habituation is context-specific is one that has vexed us for many years. Until relatively recently (in the grand scheme of things) many would have accepted the following general conclusion: provided it is the case that a change 1
It should be acknowledged that a more sophisticated analysis of the role of retrieval processes in the expression of what has been learnt during an LI procedure (and overshadowing) can also provide a potential account for the reduction in overshadowing observed by Carr (1974; see Wheeler & Miller, 2007).
172
Robert C. Honey, Mihaela D. Iordanova and Mark Good
in context between habituation training and test does not alter the animal’s perception of the stimulus (cf. Marlin & Miller, 1981) or alter the animal’s state of arousal (cf. Groves & Thompson, 1970), a change in context does not result in dishabituation. That is, habituation is not context-specific. The acceptance of this conclusion was in no small part due to an important series of experiments conducted by Hall and Channell (1985). They gave rats exposure to a localized visual stimulus in context A and then simply placed them in context B where no stimuli were presented. Once the behavioural OR to the light had habituated, the animals received test trials with the light in context B. The rats were no more likely to orient to the light in context B, a different context than the one in which habituation training had taken place, than in context A, the same context as the one in which habituation had occurred. However, as we have already seen, pairings of the light with food resulted in more rapid acquisition of conditioned responding in B than in A. These results seem to be inconsistent with the predictions made by SOP, and the dissociation appears to be correspondingly damaging to any attempt to provide a unified theory of LI and habituation. In fact, this result appears most damaging because, unlike other experiments where habituation is often very rapid (see, for example, Hall & Honey, 1989), habituation of the OR occurs gradually, and further experiments have revealed that the vigour of the OR is sensitive to other manipulations (Hall & Schachtman, 1987). However, the interpretation of the dissociation of habituation and LI observed by Hall and Channell (1985) is more complex than first appears. This is so for two reasons: one is theoretical and the other concerns an unforeseen feature of the experimental design. First, we have already seen that within SOP there are two sources of long-term LI that are context-specific: learning the association between the light and food and expressing that learning. In the case of habituation, however, a change of context can only influence performance. In short, at a theoretical level, the differential contextual dependence of LI and habituation is not as surprising as it might first appear: a change in context influences two processes (performance and learning) that could augment conditioned responding in an LI procedure, but only one process (performance) that could affect dishabituation of the OR. Second, there are now grounds for supposing that the experimental design used by Hall and Channell (1995) might have underestimated the effectiveness of the change in context in generating dishabituation of the OR. The argument is a little involved, but experimental analysis has confirmed its merit. Several studies since Hall and Channell (1985) have demonstrated that when a tone is paired with a localized light, the subsequent presentation of the tone alone provokes (conditioned) orienting (e.g., Wilson & Pearce, 1988). This rather simple observation has important implications when considered alongside the experimental design employed by Hall and Channell (1985). Thus, in the study by Hall and Channell (1985), during habituation training context A was paired with a visual stimulus whereas context B was not. This state of affairs might have made the rats
Latent inhibition and habituation: evaluation of an associative analysis
173
more likely to approach the location in which the light was to be presented in context A than in context B. The rats would then be more likely to be in the vicinity of the light in context A than in context B, a situation that could potentially obscure or otherwise counteract the result that would be expected if habituation was contextspecific (i.e., greater orienting in context B than in context A). Although this analysis may appear overly complex, and could even strike some as a little contrived, there is evidence that provides support for it. Honey, Good and Manser (1998) exposed rats to sequences of stimuli in which discrete auditory stimuli (A and B: a tone and a clicker) served as phasic contexts that immediately preceded two localized target lights (X and Y, respectively): one light was located on the left side of the stimulus panel in a standard operant chamber and the other was located on the right side of the same panel. Following habituation training, rats received test trials with the same sequences (i.e., A!X and B!Y) and trials on which the contexts in which the visual stimuli were presented were exchanged (i.e., A!Y and B!X). As with Hall and Channell (1985), there was no more orienting on the trials on which X and Y were presented in different contexts than when they were presented in the same, training contexts. However, observation of the rats’ behaviour in the auditory contexts revealed that during A they were in the vicinity where X would be presented and during B they were likely to be in the vicinity of Y. As outlined above, the fact that contexts A and B elicited different conditioned orienting responses complicates interpretation of the fact that a change in the context in which X and Y were presented did not affect orienting. Simply put, the conditioned behaviours elicited by A and B meant that the rats were more likely to be in the vicinity of X and Y when these stimuli were presented in the training contexts (on A!X and B!Y trials) than when the contexts were exchanged (on A!Y and B!X trials). This difference might well be expected to contribute to the levels of orienting on the two types of trials – increasing the likelihood of the rats coming into contact with the localized lights on same-context trials (on A!X and B!Y trials) compared with different-context trials (on A!Y and B!X trials). With the issues described above borne in mind, we conducted additional experiments where the role of conditioned orienting to the contexts could be equated on same and different trials. This was achieved by using visual stimuli that were distinguished by their temporal properties (either constantly illuminated throughout their presentations or flashing on and off) as opposed to their spatial properties (left versus right). Under these conditions, the auditory contexts (A and B) should be no more likely to result in the rats being in the vicinity of the lights on ‘same context’ test trials (A!X and B!Y) than the ‘different context’ test trials (A!Y and B!X). Now, on test trials rats were more likely to orient to X and Y when they were presented on different test trials than when they were presented on same test trials (see Figure 8.5; Honey, Good & Manser, 1998). This result is reliable (see Honey & Good, 2000a) and entirely consistent with predictions derived from SOP. Moreover, the contextual specificity of habituation has also been observed using habituation of
174
Robert C. Honey, Mihaela D. Iordanova and Mark Good
Percentage of trials with an OR
60 Same 50 Different 40 30 20 10
1
2 Test block
Figure 8.5. Associative control of the orienting response. Mean percentages of trials with an orienting response (OR) to visual stimuli as a function of whether they were presented following the same auditory context as during preexposure or following a different auditory context (adapted from Honey, Good, & Manser, 1998).
unconditioned suppression to auditory stimuli and more conventional contextual stimuli (see Jordan et al., 2000).
Neural bases of habituation and latent inhibition The behavioural studies outlined in the preceding two sections provide evidence that is consistent with the associative analysis of habituation and LI offered by SOP. Perhaps the measure of any theory, however, is its applicability to fields of enquiry that are beyond those from which it originated. The fact that habituation and LI are relatively simple and robust behavioural phenomena has encouraged their use as assays or behavioural models of changes in stimulus processing in other domains. Before LI had been well established as a behavioural effect, the neural mechanisms involved in habituation had been the focus of analysis for many years. For example, it had been argued that the hippocampus is an important component of a system that detects or acts upon mismatches between (a) the current stream of stimulation, and (b) stored memories of previous sensory streams (Gray, 1982; Sokolov, 1963; Vinogradova, 1975). This informal analysis clearly resonates with that provided by SOP. There is now converging evidence showing that disrupting hippocampal function can have dramatic effects on both habituation and LI. For example, Honey and Good (1993) showed that while LI is not always influenced by hippocampal lesions (cf. Han, Gallagher, & Holland, 1995; Oswald, Yee, Bannerman, et al., 2002) the contextual specificity of LI is abolished by such lesions. Similarly, Maren and Holt (1999) showed that the influence of the LI training context on performance (described in detail earlier in this chapter) is abolished by
Latent inhibition and habituation: evaluation of an associative analysis
175
Percentage of trials with an OR
50 Same 40
Different
30 20 10 0 Control
HPC
Figure 8.6. Associative control of the orienting response in control rats and rats with lesions of the hippocampus. Mean percentages of trials on which there was an orienting response (OR) to the elements of a visual array as a function of whether they were presented after the same auditory context as during preexposure or after a different auditory context (adapted from Honey & Good, 2000a).
inactivating the hippocampus during the retrieval test. The fact that changes in context have less impact on LI in rats with lesions to, or inactivation of, the hippocampus suggests that the mnemonic processes to which SOP appeals are mediated by the hippocampus. Perhaps the most striking evidence concerning this suggestion comes from studies that have examined the contextual dependence of habituation. We used the simple procedures described above, in which auditory stimuli served as the contexts for visual targets, to assess the role of the hippocampus in changes in stimulus processing. In one experiment, the procedure was adapted so that during training auditory context A was followed by, for example, two lights that were continuously operated through their 10-second duration (i.e., A!X,X) and B was followed by the intermittent operation of the same two lights (i.e., B!Y,Y; Honey & Good, 2000a). During the test, rats received presentations of the auditory contexts that were followed by a visual array consisting of the continuous and flashing components (e.g., A!X,Y), with one visual component (in this case X) being presented in the same auditory context as during training and the remaining component (here Y) being presented in a context that differed from training. In a group of control rats, there was more orienting to the visual component that was presented in a different context (i.e., Y) than the component presented in same context (i.e., X; see Figure 8.6). However, in rats with hippocampal lesions (group HPC), the reverse was the case: they were more inclined to orient to the component presented in the same context (X) than that presented in a different context (Y; see also Honey, Watt & Good, 1998; human fMRI data also reveal a role for the hippocampus in directly analogous procedures; see Kumaran & Maguire, 2006). The pattern of results from the control rats is exactly what SOP would predict; but what of the pattern of results
176
Robert C. Honey, Mihaela D. Iordanova and Mark Good
observed in rats with hippocampal lesions? Both groups, Control and HPC, had encoded the relationships between the auditory and visual components of the training trials; if they had not done so, then there would be no basis upon which the rats could orient preferentially to one of the components during the test (see also Honey, Watt & Good, 1998). However, the way in which this associative knowledge influenced the OR clearly differed in the two groups. One of the key features of SOP is that it provides an integrated analysis for habituation and sensitization, and returning to this analysis provides us with a potential insight into the results described above and thereby into hippocampal function. It will be recalled that SOP’s prediction that habituation (and LI) should be context-specific is one that is subject to constraints: if A1 activity is more effective at generating performance than A2 activity (i.e., W1 > W2), and the presentation of a given stimulus ordinarily produces A1 activity in all of its elements (i.e., p1 ¼ 1.00), then presenting that stimulus in a context with which it has an association will produce a reduction in the behaviour that the stimulus elicits. But, when these conditions are not met, SOP predicts that the training context could augment responding to a stimulus with which it has an association; that is, SOP predicts sensitization. The rats with hippocampal lesions in Honey and Good (2000a; see also Honey, Good & Manser, 1998) are then behaving in a way that would be anticipated if the visual target stimuli lacked intensity, and/or the weighting given to A1 and A2 activity states was relatively similar (i.e., instead of W1 > W2, W2 W1). There is one obvious locus in the mnemonic processes within SOP (see Figure 8.1) that might result in such a state of affairs, namely an especially rapid decay from the A1 state to the A2 state. If for rats with lesions to the hippocampus the rate of decay from the A1 to A2 state was especially rapid, then there would be a greater number of the elements in the A2 state when a stimulus is predicted or primed than when it is not. In effect, once the contribution of A1 to the performance generated by visual targets is reduced, the presence of the training context can augment responding during the presentation of its associated visual target by adding to the number of elements that will be in the A2 state; and changing the context in which a visual target is presented will reduce the number of elements that are in the A2 state. This formal analysis of the role of the hippocampus in long-term habituation (see Honey & Good, 2000b) is consistent with the more general suggestion that the hippocampus contributes to the process of maintaining the trace of a stimulus (e.g., Olton, Becker, & Handelmann, 1979; Rawlins, 1985). The analysis described in the previous paragraph illustrates that the associative analysis under consideration, SOP, has broad explanatory power. However, it also makes a straightforward prediction. SOP not only explains long-term changes in stimulus processing, as indexed by long-term habituation and LI, but it also explains short-term changes in stimulus processing. When a stimulus is presented, its elements move to the A1 state and from there decay into the refractory A2 state. Should the stimulus be re-presented during this refractory period then it will be unable to
Latent inhibition and habituation: evaluation of an associative analysis
Percentage of trials with an OR
Control
177
HPC
50 40
Remote
Remote
Recent
Recent
30 20 10 0
1
2
3
4
1
2
3
4
Test trials
Figure 8.7. Nonassociative control of the orienting response in control rats and rats with lesions of the hippocampus. Mean percentages of trials on which there was an orienting response (OR) to the elements of a visual array, X and Y, that had been presented 10 s (X) or 70 s before (Y) (adapted from Marshall et al., 2004).
provoke A1 activity and ordinarily will be less able to provoke responding. The analysis of hippocampal function offered in the previous paragraph makes a clearcut prediction. The influence of hippocampal lesions on long-term habituation observed by Honey and Good (2000a) should be accompanied by a parallel effect on short-term changes in stimulus processing. That is, while in control rats a stimulus should be less likely to elicit responding if it has recently been presented, in rats with lesions of the hippocampus a stimulus should be more likely to elicit responding if it has recently been presented. Marshall et al. (2004) tested this prediction using a short-term habituation procedure that was based upon the long-term habituation procedure employed by Honey and Good (2000a). Rats received sequences of three 10-second visual arrays: one visual array (Y, Y) was followed after an interval of 50 seconds by another array (X, X), and then after an interval of 10 seconds by the test array (X, Y). Control rats were more likely to orient to the stimulus experienced more remotely (Y) than to the stimulus that had been presented recently (X; see Figure 8.7). However, rats with lesions to the hippocampus (group HPC) showed the reverse pattern of orienting behaviour: at least on the initial test trials, they were more likely to orient to the stimulus that had recently been presented (X) than the stimulus that had been presented more remotely (Y; for a more detailed discussion of the pattern of results, see Marshall et al., 2004). Moreover, this modulation of short-term changes in stimulus processing is not restricted to studies of the OR; it is also evident when other forms of exploratory behaviour are examined. For example, during free and controlled exploration of an open field and conventional context, respectively, rats with hippocampal lesions are more likely to revisit or explore places that they have recently passed or explored than are control rats (see Honey, Marshall, McGregor, Futter, & Good, 2007).
178
Robert C. Honey, Mihaela D. Iordanova and Mark Good
Other determinants of habituation and latent inhibition It would be simple, yet inaccurate, to leave the reader with the impression that the associative analysis provided by SOP is the sole mechanism for habituation and LI. There is relatively clear evidence that is inconsistent with this otherwise parsimonious suggestion. For example, while rats with lesions of the hippocampus do not show the normal pattern of short-term habituation (Honey et al., 2007; Marshall et al., 2004) and long-term associative habituation (Honey & Good, 2000a; Honey, Watt, & Good, 1998), they do show dishabituation when the physical properties of the stimuli change (see Honey, Watt, & Good, 1998; Oswald et al., 2002). The clear implication of this pattern of results is that habituation and dishabituation must have multiple determinants, some involve the hippocampus and operate in the way envisaged by SOP, and other processes do not require the hippocampus (e.g., Groves & Thompson, 1970; Hawkins & Kandel, 1984; see also Dias & Honey, 2002). Similarly, the fact that LI can be observed in rats with lesions of the hippocampus, but that this LI effect is context-independent (Honey & Good, 1993), suggests that LI too has multiple determinants (see also McLaren, Bennett, Plaisted, et al., 1994; Weiner, 1990; Weiner & Feldon, 1997). This suggestion was foreshadowed by Lubow (1973). Finally, Pearce and his colleagues have shown that habituation of the OR is more marked, and LI less evident, to a visual target that has previously been followed by inconsistent consequences (food on some trials and no food on others) than when it is consistently followed by food (or no food; e.g., Kaye & Pearce, 1984; Swan & Pearce, 1988). The latter results provide clear support for the Pearce–Hall model of stimulus processing (Pearce & Hall, 1980), according to which animals will attend to and orient to stimuli whose significance is uncertain. However, it is worth noting that there is very little evidence that these effects of inconsistency can be observed in within-subjects designs, and where the potential effects of inconsistency on, for example, general arousal (cf. Groves & Thompson, 1970) can be controlled (Wilson, Boumphrey, & Pearce, 1992, Experiment 3; for a further discussion of this issue, see Honey et al., 1998).
Concluding comments The 50 years since Lubow and Moore (1959) first demonstrated LI have witnessed many empirical and theoretical articles involving this phenomenon. The current chapter has barely scratched the surface of this endeavour, and the selective approach that we have adopted, with respect to both data and theory, will not be to everyone’s taste. However, it should be apparent, not least from the diverse contributions to this volume, that a consensus has yet to be reached concerning the origins of LI, and indeed of habituation. In itself, this is a surprising observation: as a behavioural phenomenon, LI comes a close second to habituation in terms of its surface simplicity. Notwithstanding these caveats, it is our hope that this chapter
Latent inhibition and habituation: evaluation of an associative analysis
179
illustrates the general utility of one associative approach to the phenomenon from which this volume takes its name.
Acknowledgements The authors thank the BBSRC (UK) for funding the research upon which parts of this chapter are based, and Mike Le Pelley for commenting on a draft of the chapter.
References Bouton, M. E., & Swartzentruber, D. (1989). Slow reacquisition following extinction: context, encoding, and retrieval mechanisms. Journal of Experimental Psychology: Animal Behavior Processes, 15, 43–53. Carr, A. F. (1974). Latent inhibition and overshadowing in conditioned emotional response conditioning with rats. Journal of Comparative and Physiological Psychology, 86, 718–723. Dias, R., & Honey, R. C. (2002). Involvement of rat medial prefrontal cortex in novelty detection. Behavioral Neuroscience, 116, 498–503. Donegan, N. H. (1981). Priming-produced facilitation or diminution of responding to a Pavlovian unconditioned stimulus. Journal of Experimental Psychology: Animal Behavior Processes, 7, 295–312. Dwyer, D. M., & Honey, R. C. (2007). The effects of habituation training on compound conditioning are not reversed by an associative activation treatment. Journal of Experimental Psychology: Animal Behavior Processes, 33, 185–190. Gray, J. A. (1982). The Neuropsychology of Anxiety: An Enquiry into Functions of the Septo-hippocampal System. Oxford: Oxford University Press. Groves, P. M., & Thompson, R. F. (1970). Habituation: A dual-process theory. Psychological Review, 77, 419–450. Hall, G. (1991). Perceptual and Associative Learning. Oxford, England: Clarendon. Hall, G., & Channell, S. (1985). Differential effects of contextual change on latent inhibition and on the habituation of an orienting response. Journal of Experimental Psychology: Animal Behavior Processes, 11, 470–481. Hall, G., & Honey, R. C. (1989). Contextual effects in conditioning, latent inhibition and habituation: Associative and retrieval functions of contextual cues. Journal of Experimental Psychology: Animal Behavior Processes, 15, 232–241. Hall, G., & Schachtman, T. R. (1987). Differential effects of a retention interval on latent inhibition and habituation of an orienting response. Animal Learning and Behavior, 15, 76–82. Han, J. -S., Gallagher, M., & Holland, P. C. (1995). Hippocampal lesions disrupt decrements but not increments in conditioned stimulus processing. Journal of Neuroscience, 15, 7323–7329. Hawkins, R. D., & Kandel, E. R. (1984). Is there a cell biological alphabet for simple forms of learning? Psychological Review, 91, 375–391. Holt, W., & Maren, S. (1999). Muscimol inactivation of the dorsal hippocampus impairs contextual retrieval of fear memory. Journal of Neuroscience, 19, 9054–9062.
180
Robert C. Honey, Mihaela D. Iordanova and Mark Good
Honey, R. C., & Good, M. (1993). Selective hippocampal lesions abolish the contextual specificity of latent inhibition and conditioning. Behavioral Neuroscience, 107, 23–33. Honey, R. C., & Good, M. (2000a). Associative modulation of the orienting response: Distinct effects revealed by hippocampal lesions. Journal of Experimental Psychology: Animal Behavior Processes, 26, 3–14. Honey, R. C., & Good, M. (2000b). Associative components of recognition memory. Current Opinion in Neurobiology, 10, 200–204 Honey, R. C., Good, M., & Manser, K. L. (1998). Negative priming in associative learning: Evidence from a serial-habituation procedure. Journal of Experimental Psychology: Animal Behavior Processes, 24, 229–237. Honey, R. C., Hall, G., & Bonardi, C. (1993). Negative priming in associative learning: Evidence from a serial-conditioning procedure. Journal of Experimental Psychology: Animal Behavior Processes, 19, 90–97. Honey, R. C., Marshall, V. J., McGregor, A., Futter, J., & Good, M. (2007). Revisiting places passed: Sensitization of exploratory activity in rats with hippocampal lesions. Quarterly Journal of Experimental Psychology, 60, 625–634. Honey, R. C., Watt, A., & Good, M. (1998). Hippocampal lesions disrupt an associative mismatch process. Journal of Neuroscience, 18, 2226–2230. Horn, G., & Hill, R. M. (1964). Habituation of the response to sensory stimuli of neurones in the brain stem of rabbits. Nature, 202, 296–298. Jordan, W. P., Strasser, H. C., & McHale, L. (2000). Contextual control of long-term habituation in rats. Journal of Experimental Psychology: Animal Behavior Processes, 26, 323–339. Kasprow, W. J., Catterson, D., Schachtman, T. R., & Miller, R. R. (1984). Attenuation of latent inhibition by postacquisition reminder. Quarterly Journal of Experimental Psychology, 36B, 53–63. Kaye, H., & Mackintosh, N. J. (1990). A change of context can enhance performance of an aversive but not of an appetitive conditioned response. Quarterly Journal of Experimental Psychology, 42B, 113–134. Kaye, H., & Pearce, J. M. (1984). The strength of the orienting response during Pavlovian conditioning. Journal of Experimental Psychology: Animal Behavior Processes, 10, 90–109. Konorski, J. (1967). Integrative Activity of the Brain: An Interdisciplinary Approach. Chicago: University of Chicago Press. Kumaran, D., & Maguire, E. A. (2006). An unexpected sequence of events: Mismatch detection in the human hippocampus. PLoS Biology, 4, e424. Lovibond, P. E., Preston, G. C., & Mackintosh, N. J. (1984). Context specificity of conditioning and latent inhibition. Journal of Experimental Psychology: Animal Behavior Processes, 10, 360–375. Lubow, R. E. (1973). Latent inhibition. Psychological Bulletin, 79, 398–407. Lubow, R. E. (1989). Latent Inhibition and Conditioned Attention Theory. New York: Cambridge University Press. Lubow, R. E., & Moore, A. U. (1959). Latent inhibition: The effect of non-reinforced preexposure to the conditioned stimulus. Journal of Comparative and Physiological Psychology, 52, 415–419. Lubow, R. E., Rifkin, B., & Alek, M. (1976). The context effect: The relationship between stimulus preexposure and environmental preexposure determines
Latent inhibition and habituation: evaluation of an associative analysis
181
subsequent learning. Journal of Experimental Psychology: Animal Behavior Processes, 2, 38–47. Mackintosh, N. J. (1975). A theory of attention: Variations in the associability of stimuli with reinforcement. Psychological Review, 82, 276–298. Mackintosh, N. J. (1987). Neurobiology, psychology, and habituation. Behaviour, Research and Therapy, 25, 81–97. Marlin, N. A., & Miller, R. R. (1981). Associations to contextual stimuli as a determinant of long-term habituation. Journal of Experimental Psychology: Animal Behavior Processes, 7, 313–333. Marshall, V. J., McGregor, A., Good, M., & Honey, R. C. (2004). Hippocampal lesions modulate both associative and nonassociative priming. Behavioral Neuroscience, 118, 377–382. McLaren, I. P. L., Bennett, C., Plaisted, K., Aitken, M., & Mackintosh, N. J. (1994). Latent inhibition, context specificity, and context familiarity. Quarterly Journal of Experimental Psychology, 47, 387–400. McLaren, I. P. L., & Mackintosh, N. J. (2000). An elemental model of associative learning: I. Latent inhibition and perceptual learning. Animal Learning and Behavior, 28, 211–246. Olton, D. S., Becker, J. T., & Handelmann, G. E. (1979). Hippocampus, space and memory. Behavioural Brain Sciences, 2, 313–365 Oswald, C. J. P., Yee, B. K., Bannerman, D. B., et al. (2002). The influence of selective lesions to components of the hippocampal system on the orienting response, habituation and latent inhibition. European Journal of Neuroscience, 15, 1983–1990. Pavlov, I. P. (1927). Conditioned Reflexes. London: Oxford University Press. Pearce, J. M., & Hall, G. (1980). A model of Pavlovian learning: Variations in the effectiveness of conditioned but not unconditioned stimuli. Psychological Review, 87, 532–552. Rawlins, J. N. P. (1985). Associations across time: The hippocampus as a temporary memory store. Behavioural Brain Sciences, 8, 479–496. Rescorla, R. A., & Wagner, A. R. (1972). A theory of Pavlovian conditioning: Variations in the effectiveness of reinforcement and nonreinforcement. In A. H. Black & W. F. Prokasy (Eds.), Classical Conditioning II: Current Research and Theory. New York: Appleton-Century-Crofts, pp. 64–99. Sokolov, E. N. (1963). Perception and the Conditioned Reflex. London: Pergamon Press. Solomon, R. L., & Corbit, J. D. (1974). An opponent-process theory of motivation: I. Temporal dynamics of affect. Psychological Review, 81, 119–145. Swan, J. A., & Pearce, J. M. (1988). The orienting response as an index of stimulus associability in rats. Journal of Experimental Psychology: Animal Behavior Processes, 14, 292–301. Vinogradova, O. S. (1975). Registration of information and the limbic system. In G. Horn & R. A. Hinde (Eds.), Short-term Changes in Neural Activity and Behavior. Cambridge: Cambridge University Press, pp. 95–148. Wagner, A. R. (1981). SOP: A model of automatic memory processing in animal behavior. In N. E. Spear & R. R. Miller (Eds.), Information Processing in Animals: Memory Mechanisms. Hillsdale, NJ: Lawrence Erlbaum Associates, pp. 5–48. Wagner, A. R., & Brandon, S. E. (1989). Evolution of a structured connectionist model of Pavlovian conditioning. In S. B. Klein & R. R. Mowrer (Eds.),
182
Robert C. Honey, Mihaela D. Iordanova and Mark Good
Contemporary Learning Theories: Pavlovian Conditioning and the Status of Traditional Learning Theory. Hillsdale, NJ: Lawrence Erlbaum Associates, pp. 149–190. Wagner, A. R., & Brandon, S. E. (2001). A componential theory of Pavlovian conditioning. In R. R. Mowrer & S. Klein (Eds.), Contemporary Learning: Theory and Application. Mahwah, NJ: Erlbaum, pp. 23–64. Weiner, I. (1990). Neural substrates of latent inhibition: the switching model. Psychological Bulletin, 108, 442–461. Weiner, I., & Feldon, J. (1997). The switching model of latent inhibition: an update of neural substrates. Behavioral Brain Research, 88, 11–25. Wheeler, D. S., & Miller, R. R. (2007). Constrasting reduced overshadowing and blocking. Journal of Experimental Psychology: Animal Behavior Processes, 33, 349–359. Wilson, P. N., Boumphrey, P., & Pearce, J. M. (1992). Restoration of the orienting response to a light by a change in its predictive accuracy. Quarterly Journal of Experimental Psychology, 44B, 17–36. Wilson, P. N., & Pearce, J. M. (1988). Transfer of responding during serial conditioning in the rat. Quarterly Journal of Experimental Psychology, 40B, 259–278.
9 Latent inhibition and creativity Shelley Carson
Latent inhibition (LI) is a robust phenomenon in which repeated preexposure to a stimulus that is not reinforced retards future associability to that stimulus (Lubow, 1989). LI has been uniformly accepted as an adaptive mechanism across a variety of species (Lubow & Gewirtz, 1995). In humans, a deficit in LI has been associated with the active phase of schizophrenia and with psychosis-proneness (Baruch, Hemsley, & Gray, 1988a, 1988b; Lubow, Ingberg-Sachs, Zalstein-Orda, & Gewirtz, 1992). However, attenuated LI has also been reported in non-disordered normal subjects (see Braunstein-Bercovitz, Rammsayer, Gibbons, & Lubow, 2002), suggesting that attenuated LI exists on a continuum that extends from hospitalized psychotics to high-functioning normals. Recent research suggests that there may be situations in which attenuated LI actually confers an advantage to individuals. There is a growing body of evidence that indicates attenuated LI may be present in a subset of highfunctioning and creative individuals (e.g. Carson, Peterson, & Higgins, 2003; Peterson & Carson, 2000). Attenuated LI may increase the probability of making novel or original associations among disparate stimuli by increasing the amount of information available to conscious awareness.
The theoretical relationship of creativity and latent inhibition Latent inhibition is widely accepted as an index of the ability to ignore irrelevant stimuli (Lubow & Kaplan, 2005). When latent inhibition is expressed, the ability to form associations to information deemed irrelevant is reduced. Conversely, when latent inhibition is attenuated, the ability to form associations to seemingly irrelevant information is expanded. This is conceptually related to the associationist view of creativity described by Sarnoff Mednick, which posits that creativity is the result of combining loosely associated information to form novel and original ideas. According to Mednick (1962), creative thinking is “the forming of associative elements into new combinations which either meet specified requirements or are in some Latent Inhibition: Cognition, Neuroscience and Applications to Schizophrenia, ed. R.E. Lubow and Ina Weiner. Published by Cambridge University Press # Cambridge University Press 2010.
183
184
Shelley Carson Latent Inhibition deficit
Creative Genius
Psychosis
Additional Cognitive Deficits Shared Biological Vulnerability
Very High IQ
Additional Cognitive Strengths
Figure 9.1. Creativity and psychosis: the shared vulnerability model.
way useful. The more mutually remote the elements of the new combination, the more creative the process or solution” (p. 221). In theory, then, attenuated LI should increase the ability “to form associative elements into new combinations” by increasing the number of stimuli available for recombination in the cognitive workspace. If creative thought is indeed the ability to combine disparate bits of information in a novel and original synthesis, then a cognitive mechanism that limits available information bits (such as LI) should be detrimental to creativity while a deficit in such a mechanism – at least in special circumstances – should prove beneficial. But what special circumstances would allow a deficit in LI to confer this creative advantage? In order to answer this question, it may be beneficial to examine the LI deficit in the context of a “shared vulnerability” model of the creativity/psychopathology relationship (see Figure 9.1). This model suggests that psychotic pathology and creativity share biological components that are expressed as either mental illness or as creative genius based on the presence of other moderating factors. A deficit in LI may represent one aspect of this shared vulnerability. If reduced LI increases the available stimuli in conscious awareness to include an array of irrelevant information, then a cognitive strength (such as high IQ) that allows an individual to process and manipulate that additional information (rather than being overwhelmed by it) may serve as a moderating factor that protects the organism from psychosis and encourages creativity. In this model, psychotic vulnerability (including the LI deficit) is associated with creativity only in the presence of high IQ or other cognitive strengths. Such a model accounts for data suggesting that highly creative individuals are at greater risk for psychopathology than members of the general public (e.g. Jamison, 1989; Ludwig, 1995; Prentky, 1989). It also explains why not all highly creative individuals express
Latent inhibition and creativity
185
psychopathology and, conversely, why not all psychotic or psychosis-prone individuals express unusual creativity. The shared vulnerability model makes several predictions: (1) a deficit in latent inhibition will be associated with psychosis and psychosis-proneness; (2) psychosis-proneness will be elevated in the creative population; (3) a deficit in LI in the presence of high IQ will enhance creativity.
In the remainder of this chapter, I review the literature relevant to the predictions stemming from this model. Research confirming the first prediction is addressed elsewhere in this volume (e.g., Kumari & Ettinger; Lubow). Therefore, after addressing the different methods of measuring creativity and latent inhibition, I examine support for the final two predictions: that schizotypy and psychosis-proneness will be associated with creativity and that attenuated LI in the presence of high IQ will enhance creativity.
The measurement of creativity One problem that has hampered research on the relationship of creativity to latent inhibition (and indeed to other variables of interest) is the measurement of the construct of “creativity”. Most researchers have adopted Barron’s (1969) view that a creative product, whether it is a poem, a medical cure, or a plan for a new weapon of mass destruction, must be both (1) novel or original, and (2) useful or adapted in some way to reality. However, researchers have not agreed on the most accurate way to measure creativity within a person and they have employed a variety of methods, depending upon the purpose of the individual investigation (Carson, Peterson, & Higgins, 2005). Common methods of measuring creativity include divergent thinking tasks, creative personality scales, inventories of creative accomplishments, markers of eminence, and having research participants manufacture a creative product in the lab. The most widely used measures of creativity are divergent thinking tasks (e.g. Torrance, 1968; Wallach & Kogan, 1965), which assess the potential for creative ideation and the cognitive ability to activate broad associational networks (Runco, 1991). Such tasks include listing alternate uses for a common object (such as a brick) and listing the consequences of a hypothetical situation (e.g., if humans had six fingers instead of five on each hand). Several aspects of divergent thinking are assessed, including fluency (number of responses generated) and originality (unusualness of responses). While divergent thinking tests have been criticized as capturing only the most trivial aspects of creativity, they have been reasonably successful at identifying creative individuals (McCrae, 1987). A second method of measuring creativity assesses creative personality. The most widely used instrument in this category is Gough’s (1979) Creative Personality Scale (CPS). This scale consists of 30 adjectives that have been found to be most predictive (either positively or negatively) of the personalities of highly creative individuals.
186
Shelley Carson
Subjects are scored on how many of the positive adjectives they endorse and how many of the negative adjectives they fail to endorse. Many researchers believe that creativity should be measured in terms of actual creative production. Therefore, a number of creative achievement inventories have been developed. These inventories are typically self-report checklists of accomplishments in assorted creative domains. The Creative Achievement Questionnaire (CAQ; Carson et al., 2005), for example, measures creative training and accomplishment in ten different domains of science and the arts. Another type of creativity measurement is the marker of creative eminence. Markers can include universally accepted signs of creative accomplishment, such as having won a Nobel Prize or being listed in Who’s Who. Researchers can then measure variables of interest in a group of subjects who exhibit this criterion against a group of matched controls who do not. One final way of measuring creativity, known as the Consensual Assessment Technique (Amabile, 1982), asks research participants to manufacture a creative product, usually a collage or a poem, under controlled conditions in the lab. These products are rated for creativity by a panel of judges with artistic expertise. In summary, divergent thinking tasks and creative personality scales measure creative potential or trait creativity (Eysenck, 1995b). Markers of creative eminence, creative achievement inventories, and the Consensual Assessment Technique measure actual creative behavior and output as judged by societal recognition. The variety of measurement methods can make it difficult to compare study results in the field of creativity research.
The measurement of latent inhibition Researchers have also used different types of tasks to measure latent inhibition in humans, further complicating the comparison of studies. The typical measure of latent inhibition in humans includes a task consisting of two parts: a preexposure phase and a test phase. In one type of LI measure, the trials-to-criterion task, participants who have been exposed to a target stimulus without a reinforcer in the preexposure phase typically take more trials to learn an association to that stimulus in the test phase than do participants who have not been preexposed to the target stimulus. The number of trials needed to learn the association is an indicator of LI, with a lower number indicating an attenuation of LI. The trials-to-criterion measure can be delivered as either an auditory or a visual task (e.g., Burch, Hemsley, & Joseph, 2004). In a second measure of LI, the reaction-time task, participants are asked in the test phase to make a decision that is contingent upon a target stimulus. Participants typically display longer reaction times to make a decision in trials involving a target stimulus to which they were preexposed than to trials involving targets to which they have not been preexposed. The difference in reaction
Latent inhibition and creativity
187
time between trials of preexposed and nonpreexposed target stimuli is an indicator of LI, with lower numbers indicating an attenuation of LI (De la Casa & Lubow, 2001).
Psychosis-proneness and creativity Previous research suggests that schizotypal or psychosis-prone individuals may suffer from an LI deficit (e.g., Lubow et al., 1992). If attenuated latent inhibition is a vulnerability factor for creativity as well as for psychosis, as predicted by the shared vulnerability model, then we can expect at least a subset of creative individuals to manifest actual or subclinical symptoms of psychosis. A rationale for this expectation was provided by Eysenck (1995a), who suggested that a deficit in cognitive inhibition would promote “overinclusive thinking”, the loose, associative thinking process that appears to characterize both schizophrenic and creative thought. The psychoticism scale of the Eysenck Personality Questionnaire (EPQ; Eysenck, Eysenck, & Barrett, 1985) was originally employed as a measure of psychotic vulnerability in studies of creativity (e.g., Woody & Claridge, 1977). However, more typically, psychosis-proneness, or schizotypy, is measured by scales that assess symptoms of schizotypal personality in subclinical populations. These symptoms fall into several factors that include magical thinking, perceptual aberrations, social anhedonia, impulsive nonconformity, and cognitive disorganization (Mason, Linney, & Claridge, 2005). Both magical thinking and perceptual aberration have been associated with the positive traits of psychotic disorders (hallucinations and delusions) and include experiences such as marginal hallucinatory experiences (e.g. hearing knocking at the door), feeling a sense of presence, confusing dreams with reality, out-of-body experiences, depersonalization or derealization, de´ja` vu, and sensations that one’s body or face is changing shape (Eckblad & Chapman, 1983). Therefore, magical ideation and perceptual aberration indicators are sometimes combined to create a “positive” factor of schizotypy (Claridge et al., 1996), while impulsive nonconformity and social anhedonia constitute a “negative” factor. There is a rich literature associating psychosis and schizotypy with creative individuals. Anecdotal evidence from the biographies of highly creative artists and scientists suggests that creative people often engage in behaviors and/or report experiences that are decidedly schizotypal. The composer Schumann, for example, reported that Beethoven and Mendelssohn dictated musical compositions to him “from their tombs” (Lombroso, 1891/1976, p. 68). Likewise, William Blake claimed that both his poetry and his paintings were presented to him in visions by visiting spirits who sometimes jostled him while competing for his attention (Shaw, 2000). Tesla, the scientist credited with developing alternating electrical current, is reported to have suffered from columbiphilia (pigeon-love) and triphilia (obsession with the number three), as well as from auditory and visual hallucinations (Pickover, 1998). Charles Dickens is reported to have fended off the imaginary urchins of his novels
188
Shelley Carson
with an umbrella as he walked the streets of London. Beethoven had such disregard for his personal cleanliness that friends had to undress him and wash his clothes while he slept (Shaw, 2000). In addition to anecdotal evidence, a body of empirical evidence associates schizotypy and psychosis-proneness with creative ability and achievement. Early research provided evidence that genetic vulnerability to psychosis may enhance creativity, even if the creative individual is not measurably psychotic. Heston (1966), for example, found that the adopted-away children of schizophrenic mothers were more likely to pursue creative professions than were adopted-away children of controls. Karlsson (1970) found a higher incidence of creative achievers (as evidenced by a listing in Iceland’s Who’s Who) in the relatives of psychotic patients than was found in relatives of members of the general population. In recent decades, some researchers (e.g. Andreasen, 1987) have questioned this early work, noting that the data from these studies were collected prior to the DSM III (APA, 1980) diagnostic changes regarding schizophrenia and bipolar disorder. Many individuals who were diagnosed as schizophrenic prior to 1980 may have been diagnosed as bipolar subsequently. However, there is substantial empirical evidence for high levels of eccentricity and schizotypy among creative subjects versus controls that cannot be attributed to bipolar spectrum disorders. For example, studies of creative achievers performed at Berkeley’s Institute for Personality Assessment and Research (IPAR) found that creative writers and architects had elevated scores on the MMPI scales of Psychopathic Deviation, Schizophrenia, and Paranoia. The architects and writers often reported vivid dreams, altered states of consciousness, and unusual cognitive experiences. These are characteristic of positive schizotypal symptoms (Barron, 1969; MacKinnon, 1962). Barron (1969) noted that creative artists often reported eccentricities also reported by schizophrenics, including unusual perceptual occurrences and odd mystical experiences. More recent research has supported a relationship between psychosis-proneness and divergent thinking measures of creativity (Abraham & Windmann, 2008; Batey & Furnam, 2008; Cox & Leon, 1999; Folley & Park, 2005; Green & Williams, 1999; Poreh et al., 1994; Rawlings, 1985; Rawlings & Locarnini, 2008; Schuldberg, French, Stone, & Heberle, 1988). Research has also indicated that artists score higher on scales of schizotypy than controls (e.g., Burch, Pavelis, Hemsley, & Corr, 2006; Nettle, 2006) and that art students have higher schizotypy scores than humanities students (O’Reilly, Dunbar, & Bentall, 2001). However, levels of schizotypy have been found to be lower in scientists and mathematicians than in artistic individuals (e.g., Nettle, 2006; Rawlings & Locarnini, 2008). In general, these studies have agreed on two points: (1) there is an elevated level of psychosis-proneness/schizotypy in creative individuals and divergent thinkers, and (2) divergent thinking and creativity are primarily associated with the positive factor (magical ideation and unusual experiences) rather than the negative factor of schizotypy (but see Burch, Hemsley et al., 2006; Cox & Leon, 1999).
Latent inhibition and creativity
189
Creativity, attenuated latent inhibition, and IQ The shared vulnerability model predicts that attenuated latent inhibition combined with cognitive strengths, such as high IQ, may enhance creativity. In order to test this prediction, researchers at Harvard and the University of Toronto conducted a series of investigations of latent inhibition within high-IQ samples (Carson et al., 2003; Peterson & Carson, 2000; Peterson, Smith, & Carson, 2002). In a preliminary study, Peterson and Carson (2000) examined the relationship of the personality trait Openness to Experience to LI. Openness, one of the overarching factors in the Five-Factor Model of Personality (NEO-FFM; Costa & McCrae, 1992), incorporates attributes such as imagination, fantasy-proneness, creativity, intellectual curiosity, unconventional attitudes, and divergent thinking (McCrae, 1987). Several studies have linked openness to higher levels of psychosis-proneness and schizotypal personality (e.g., Miller & Tal, 2007). In fact, Haigler and Widiger (2001) have suggested that the openness scale, with its emphasis on openness to fantasy, ideas, and actions, may represent the positive and adaptive side of schizotypal personality. Because individuals scoring high on schizotypy scales have demonstrated reduced LI, it seemed reasonable to hypothesize that individuals who score high on the openness scale might also demonstrate reduced LI. Peterson and Carson (2000) tested this hypothesis on a group of 86 high-IQ (mean ¼ 130) undergraduates, using a trials-to-criterion LI task that had been developed for studies with schizotypal subjects (Lubow et al., 1992). When openness scores were divided by median split, the mean LI scores for the low-openness group were almost twice as high as the LI scores for the high-openness group, suggesting a definite attenuation of LI in the high-openness participants. LI was also negatively associated with the measures of extraversion and EPQ psychoticism. The findings relative to openness and extraversion were later replicated using a group of 79 undergraduates from the University of Toronto (Peterson, Smith, & Carson, 2002). This study, which employed a trials-to-criterion LI task, also found that LI scores were attenuated in subjects who scored high on the Gough’s (1979) Creative Personality Scale. Carson, Peterson, and Higgins (2003) published a report describing a series of three studies that broadened the research on creativity and LI. In Experiment 1, they tested 86 Harvard undergraduates on measures of latent inhibition, creativity, and IQ. They found an overall latent inhibition effect between the preexposed and nonpreexposed groups. The authors then divided scores on each of the creativity measures, including the Creative Personality Scale, Creative Achievement Questionnaire, and divergent thinking tasks, by median split. For participants who had been assigned to the preexposed LI condition, LI scores were lower in the high creative groups than in the low creative groups on all measures, suggesting an association between creativity and attenuated LI (see Table 9.1). Specifically, LI scores were significantly lower in the high (versus low) creative groups on the following measures: the Creative
190
Shelley Carson
Table 9.1. LI scores (preexposed condition) of high-vs.-low creative groups (based on median split)
Creativity measure
LI (preexposed) score of high creativity group (SD)
LI (preexposed) score of low creativity group (SD)
t
Df
P value
d
CAQ CPS Diverg Orig Fluency Flex
14.3 11.0 15.6 15.1 16.4 16.7
21.8 (10.4) 20.8 (10.1) 21.4 (10.9) 21.5 (10.3) 21.1 (11.5) 19.6 (11.4)
2.93 3.63 2.09 2.34 1.58 1.01
55 47 52 52 52 52
0.006 0.0004 0.04 0.02 0.12 0.32
0.79 1.06 0.58 0.65 0.43 0.28
(8.8) (7.0) (9.5) (9.7) (9.5) (9.5)
Notes: CAQ ¼ Creative Achievement Questionnaire; CPS ¼ Creative Personality Scale; Diverg ¼ total divergent thinking score; Orig ¼ originality subscale; Flex ¼ flexibility subscale.
Achievement Questionnaire (CAQ), the Creative Personality Scale (CPS), total divergent thinking scores, and the originality scale of the divergent thinking tasks. Of interest in this study was the finding that, when regressed on CAQ scores, attenuated LI and IQ scores accounted for 26% of the variance in creative achievement scores. This suggested that attenuated LI was related not only to creative thinking processes but also to actual creative accomplishment, at least within this high-IQ sample. The authors replicated this finding in a study of 96 Harvard undergraduates (Carson et al., 2003, Experiment 2). In this investigation, all participants were tested in the preexposed LI condition, using the CAQ to assess creative achievement. LI scores in the high creative achievement group (as determined by median split) were significantly lower than in the low creative group, again supporting an association between creativity and reduced LI. When IQ and LI scores were regressed on the combined highest and lowest quartiles of the CAQ, IQ and LI jointly predicted 20% of the variance in creative achievement scores, with LI negatively predicting 13% of the variance. In Experiment 3, Carson et al. (2003) compared 25 eminent creative achievers (selected from the combined group of 182 Harvard subjects) to 23 low creative control subjects. The eminent creative achievers had each made a significant contribution to a creative domain, such as having sold a novel to a publishing house, having a musical composition recorded and sold, having a prototype invention built and patented, having a private showing of original artwork at a recognized gallery, or winning a scholarship or national prize for a scientific discovery. Eminent creative achievers were seven times more likely to have low rather than high LI scores, while the control subjects were more likely to have high LI scores. When the eminent achievers and controls were pooled and divided into groups of high and moderate
191
Latent inhibition and creativity 50 45
Creative Achievement Score
40 35 30 25
28.6
20 15 10 5
8.1
6
9.8
0
High LI
Low LI Moderate IQ
High IQ
Figure 9.2. CAQ scores of LI IQ groups in pooled eminent creative achievers and controls. Error bars represent the standard error (positive value only).
IQ scores crossed with high and low LI, the high-IQ/low-LI group had creative achievement scores almost three times greater than any of the other crossed groups (see Figure 9.2). Two other groups have reported findings relative to creativity and LI (Burch, Hemsley, Pavelis, & Corr, 2006; Wuthrich & Bates, 2001). Both groups hypothesized that, based on Eysenck’s (1995a) theory of creativity and disinhibition, there would be an association between measures of creativity and attenuated LI. Neither group made predictions about the moderating effect of IQ. Wuthrich and Bates (2001) tested 54 psychology undergraduates on measures of latent inhibition (assessed by trials-to-criterion), personality, schizotypy, and divergent thinking. The authors divided schizotypy scores into five equal groups and reported an inverted-U relationship with LI scores in the preexposed condition, such that subjects with both very low and very high schizotypy scores displayed attenuated LI while those with moderate schizotypy displayed normal LI. They also mentioned that the divergent thinking task scores were positively correlated to schizotypy but showed no relation to latent inhibition. However, the authors presented no statistical information concerning creativity and LI. Although the creativity/LI relationship was not the primary focus of this investigation, the authors suggested that while both LI and creativity were related to schizotypy, they were independently related to that factor. Their data also indicated that
192
Shelley Carson
openness was not correlated to divergent thinking scores, schizotypy, or latent inhibition (contradicting most previous research findings). Burch, Hemsley et al. (2006) also investigated the relationship of creativity to latent inhibition. In this study, creativity in 100 subjects was measured by divergent thinking tasks and the Creative Personality Scale; latent inhibition was measured using a reaction time within-subjects paradigm. Contrary to Carson et al.’s (2003) results and to the authors’ original hypotheses, they found a positive (rather than negative) correlation between latent inhibition and measures of creativity, suggesting that the more creative individuals demonstrated higher levels of LI than the less creative individuals. However, the study also failed to find an overall latent inhibition effect across the sample. Burch, Hemsley et al. (2006) suggested several reasons for their failure to replicate the earlier findings of a relationship between creativity and reduced LI. First, they noted that, since they could not demonstrate an overall latent inhibition effect for this sample, the reaction time within-subjects paradigm might be a less valid test for LI in a non-clinical population. Second, they noted differences in the IQ of their sample (mean ¼ 108) and the samples in the Carson et al. studies (means ranged from 125 to 131). Carson et al. (2003) had noted that the relationship between attenuated LI and creativity decreased as a function of decreasing IQ. Since there is evidence that IQ and creativity are correlated below a level of approximately IQ 120 (Sternberg & O’Hara, 1999), attenuated LI may facilitate creativity only in the presence of very high IQ. At lower levels of IQ, attenuated LI may act to reduce creativity. Finally, Burch, Hemsley et al. (2006) suggested that because latent inhibition also has been shown to be attenuated in individuals with high state-anxiety, the relationship with creativity might be mediated by anxiety rather than an overinclusive thinking style. In summary, while the research that directly tests the relationship of LI to measures of creativity has yielded mixed results, investigations conducted on groups with IQ greater than 120 have demonstrated a significant relationship between attenuated LI and measures of creativity. These findings support the shared vulnerability model, which predicts that a deficit in LI when combined with high IQ may enhance creativity.
Directions for future research The preponderance of evidence suggests that, at least when found in combination with high IQ, latent inhibition may be attenuated among highly creative individuals. However, inconsistencies in the research suggest four directions for future research.
Reaction time versus trial-to-criterion LI testing Burch, Hemsley et al. (2006) found a positive rather than a negative correlation of creativity to LI using a within-subject reaction-time task. Two explanations suggest this result should be investigated further. First, recent research indicates that
Latent inhibition and creativity
193
creative individuals have slower reaction times than controls on tasks that involve interference (Dorfman, Martindale, Gassimova, & Vartanian, 2008). Dorfman et al. suggested that creative individuals automatically attend to more irrelevant information on interference tasks and that this widening of attention comes at the cost of slower processing speed. Previous research on dichotic shadowing tasks (e.g., Dykes and McGhie, 1976; Rawlings, 1985) also found that creative individuals attend to more irrelevant information than controls. Because the reaction-time latent inhibition task involves interference, it is possible that creative individuals demonstrate slower reaction times on the trials with the target stimuli because they are attending to the target stimuli rather than failing to form associations to it. It is interesting to note that research using another reaction-time task to measure cognitive inhibition (the negative priming paradigm) has also failed to yield results in creative subjects (e.g., Green & Williams, 1999). Therefore, as suggested by Burch et al. (2006), the reaction-time paradigm may be unsuitable to test LI in the non-disordered creative population. More research on reaction time, attention, and creativity may clarify this issue.
Attenuated LI versus variable LI in creative subjects LI can be modulated by, for example, drugs (e.g., Gray, Pickering, Hemsley, et al., 1992) or stress (e.g., Braunstein-Bercovitz et al., 2002). It is also possible that creative individuals may be able to modulate their level of LI depending upon task demands. Martindale (1999) demonstrated that creative individuals can automatically switch between focused and defocused states of attention more easily than less creative individuals. He suggested that creativity may be, in part, an ability to switch between states of cognitive inhibition and disinhibition (Martindale, 1999). Future research should focus on determining whether creative individuals can, in fact, manipulate their level of latent inhibition to facilitate idea generation (which could be associated with reduced LI) or information gathering and idea evaluation (which could be associated with more intact levels of LI). To date, no research on the ability to manipulate LI has been conducted using a high-versus-low creative sample.
Creativity, LI, and stress Burch, Hemsley, et al. (2006) found that LI was negatively correlated with one of two measures of neuroticism, suggesting that attenuated LI might have characterized high anxiety-prone individuals in their sample. Other research has found that LI is attenuated in stressful situations, as did Braunstein-Bercovitz et al. (2002). These authors posed the question of whether anxiety, rather than psychosis-proneness, modulates latent inhibition. The literature on the relationship of creativity to anxiety is mixed. In general, studies that have found any relationship between creativity and
194
Shelley Carson
trait neuroticism have reported a positive correlation (see Feist, 1999, for a review). However, studies have also generally found that a state of anxiety is detrimental to creativity (see Baas, DeDreu, & Nijstad, 2008, for a review). Carson et al. (2003) found no relationship of trait neuroticism to either reduced LI or to measures of creativity in their high IQ samples. However, future research should examine how stress affects the expression of latent inhibition in creative individuals.
Creativity, LI, and other cognitive strengths Finally, the shared vulnerability model predicts that creativity will be enhanced by reduced LI in the presence of cognitive strengths that facilitate the processing of the additional stimuli allowed into conscious awareness. As already discussed, Carson et al. (2003) investigated this prediction using IQ as a high-order cognitive strength. However, other more focused cognitive strengths may also combine with the LI deficit to enhance creativity. A preliminary finding, using working memory (WM) capacity for abstract shapes as a potential cognitive strength, suggested that a combination of low WM capacity combined with LI predicted limited creative achievement, whereas high WM combined with low LI predicted high achievement (Carson, 2001). However, more research on the combination of LI and working memory, as well as other cognitive strengths, would illuminate the cognitive underpinnings of high creative achievement.
Conclusions In previous research with human subjects, attenuated latent inhibition has been associated with schizophrenia and schizotypy. In this chapter, I have provided evidence supporting an association between attenuated LI and creativity (as measured by divergent thinking tasks, creative personality measures, and an inventory of creative achievement), but only in the presence of high IQ, a cognitive asset that may enable the processing of the additional stimuli allowed into conscious awareness through the LI deficit. The reviewed body of research suggests that highly creative individuals and psychotic individuals may share a predispositional factor in the form of attenuated latent inhibition. The research also suggests that highly creative people may possess cognitive protective factors that may inhibit, reduce, or delay the expression of psychosis while increasing the probability of creative achievement. These factors include, but are not limited to, high IQ and possibly an increased working memory capacity. An important debate in the creativity literature centers on whether the cognitive processes of individuals capable of the highest levels of creative achievement are qualitatively or merely quantitatively different from the processes used by normal thinkers. It is possible that very high IQ combined with attenuated latent inhibition
Latent inhibition and creativity
195
may characterize a qualitatively different thinking process that defines the creative genius. In this rare condition, the attenuation of latent inhibition may confer a benefit rather than a deficit to the individual and, in turn, to society as a whole.
References Abraham, A., & Windmann, S. (2008). Selective information processing advantages in creative cognition as a function of schizotypy. Creativity Research Journal, 20, 1–6. Amabile, T. M. (1982). Social psychology of creativity: a Consensual Assessment Technique. Journal of Personality and Social Psychology, 43, 997–1013. American Psychiatric Association (1980). Diagnostic and Statistical Manual of Mental Disorders (third edition). Washington DC: American Psychiatric Association. Andreasen, N. (1987). Creativity and mental illness: prevalence rates in writers and their first-degree relatives. American Journal of Psychiatry, 144, 1288–1292. Baas, M., DeDreu, C. K., & Nijstad, B. A. (2008). A meta-analysis of 25 years of mood–creativity research: hedonic tone, activation, or regulatory focus? Psychological Bulletin, 134, 779–806. Barron, F. (1969). Creative Person and Creative Process. New York: Holt, Rinehart, and Winston. Baruch, I., Hemsley, D. R., & Gray, J. A. (1988a). Differential performance of acute and chronic schizophrenics in a latent inhibition task. Journal of Nervous and Mental Disease, 176, 598–606. Baruch, I., Hemsley, D. R., & Gray, J. A. (1988b). Latent inhibition and “psychotic proneness” in normal subjects. Personality and Individual Differences, 9, 777–783. Batey, M., & Furnham, A. (2008). The relationship between measures of schizotypy and creativity. Personality and Individual Differences, 45, 816–821. Braunstein-Bercovitz, H., Rammsayer, T., Gibbons, H., & Lubow, R. E. (2002). Latent inhibition deficits in high-schizotypal normals: symptom-specific or anxiety-related? A review. Schizophrenia Research, 53, 109–121. Burch, G. St. J., Hemsley, D. R., & Joseph, M. H. (2004). Trials-to-criterion latent inhibition in humans as a function of stimulus pre-exposure and positive-schizotypy. British Journal of Psychology, 95, 179–196. Burch, G. St. J., Hemsley, D. R., Pavelis, C., & Coor, P. J. (2006). Personality, creativity, and latent inhibition. European Journal of Personality, 20, 107–122. Burch, G. St J., Pavelis, C., Hemsley, D. R., & Corr, P. J. (2006). Schizotypy and creativity in visual artists. British Journal of Psychology, 97, 177–190. Carson, S. H. (2001). Demons and muses: An exploration of cognitive features and vulnerability to psychosis in creative individuals. Retrieved from Dissertations and Theses database. Harvard University (AAT3011334). Carson, S., Peterson, J. B., & Higgins, D. M. (2003). Decreased latent inhibition is associated with increased creative achievement in high-functioning individuals. Journal of Personality and Social Psychology, 85, 499–506. Carson, S., Peterson, J. B., & Higgins, D. (2005). Reliability, validity and factor structure of the Creative Achievement Questionnaire. Creativity Research Journal, 17, 37–50.
196
Shelley Carson
Claridge, G., McCreery, C., Mason, O., et al. (1996). The factor structure of “schizotypal” traits: a large replication study. British Journal of Clinical Psychology, 35, 103–115. Costa, P. T., & McCrae, R. R. (1992). Revised NEO Personality Inventory and NEO Five-Factor Inventory Professional Manual. Odessa, FL: Psychological Assessment Resources. Cox, A. J., & Leon, J. L. (1999). Negative schizotypal traits in the relation of creativity to psychopathology. Creativity Research Journal, 12, 25–36. De la Casa, L. G., & Lubow, R. E. (2001). Latent inhibition with a response time measure from a within-subject design: effects of number of pre-exposures, masking task, context change, and delay. Neuropsychology, 15, 244–253. Dorfman, L., Martindale, C., Gassimova, V., & Vartanian, O. (2008). Creativity and speed of information processing: a double dissociation involving elementary versus inhibitory cognitive tasks. Personality and Individual Differences, 44, 1384–1390. Dykes, M., & McGhie, A. (1976). A comparative study of attentional strategies of schizophrenic and highly creative normal subjects. British Journal of Psychiatry, 128, 50–64. Eckblad, M., & Chapman, L. J. (1983). Magical ideation as an indicator of schyzotypy. Journal of Consulting and Clinical Psychology, 51, 215–225. Eysenck, H. J. (1995a). Creativity as a product of intelligence and personality. In D. H. Saklofske & M. Zeidner (Eds.), International Handbook of Personality and Intelligence. New York: Plenum Press. Eysenck, H. J. (1995b). Genius: The Natural History of Creativity. Cambridge: Cambridge University Press. Eysenck, S. B. G., Eysenck, H. J., & Barrett, P. (1985). A revised version of the psychoticism scale. Personality and Individual Differences, 6, 21–29. Feist, G. F. (1999). The influence of personality on artistic and scientific creativity. In R. J. Sternberg (Ed.), Handbook of Creativity. Cambridge: Cambridge University Press. Folley, B. S., & Park, S. (2005). Verbal creativity and schizotypal personality in relation to prefrontal hemispheric laterality: a behavioral and near-infrared optical imaging study. Schizophrenia Research, 80, 271–282. Gough, H. G. (1979). A creative personality scale for the Adjective Check List. Journal of Personality and Social Psychology, 37, 1398–1405. Gray, N. S., Pickering, A. D., Hemsley, D. R., Dawling, S., & Gray, J. A. (1992). Abolition of latent inhibition by a single 5 mg dose of d-amphetamine in man. Psychopharmacology, 107, 425–430. Green, M. J., & Williams, L. M. (1999). Schizotypy and creativity as effects of reduced cognitive inhibition. Personality and Individual Differences, 27, 263–276. Haigler, E. D., & Widiger, T. A. (2001). Experimental manipulation of neo–pi–r items. Journal of Personality Assessment, 77, 339–358. Heston, L. L. (1966). Psychiatric disorders in foster home reared children of schizophrenic mothers. British Journal of Psychiatry, 112, 819–825. Jamison, K. (1989). Mood disorders and patterns of creativity in British writers and artists. Psychiatry, 52, 125–134. Karlsson, J. L. (1970). Genetic association of giftedness and creativity with schizophrenia. Hereditas, 66, 177–182. Lombroso, C. (1891/1976). The Man of Genius. London: Walter Scott Publishing.
Latent inhibition and creativity
197
Lubow, R. E. (1989). Latent Inhibition and Conditioned Attention Theory. Cambridge: Cambridge University Press. Lubow, R. E., & Gewirtz, J. C. (1995). Latent inhibition in humans: Data, theory, and implications for schizophrenia. Psychological Bulletin, 117, 87–103. Lubow, R. E., Ingberg-Sachs, Y., Zalstein-Orda, N., & Gewirtz, J. C. (1992). Latent inhibition in low and high “psychotic-prone” normal subjects. Personality and Individual Differences, 13, 563–572. Lubow, R. E., & Kaplan, R. (2005). The visual search analogue of latent inhibition: implications for theories of irrelevant stimulus processing in normal and schizophrenic groups. Psychonomic Bulletin & Review, 12, 224–243. Ludwig, A. (1995). The Price of Greatness: Resolving the Creativity and Madness Controversy. New York: Guilford Press. MacKinnon, D. W. (1962). The nature and nurture of creative talent. American Psychologist, 17, 484–495. Martindale, C. (1999). Biological basis of creativity. In R. J. Sternberg (Ed.), Handbook of Creativity. Cambridge: Cambridge University Press. Mason, O., Linney, Y., & Claridge, G. (2005). Short scales for measuring schizotypy. Schizophrenia Research, 78, 293–296. McCrae, R. R. (1987). Creativity, divergent thinking, and openness to experience. Journal of Personality and Social Psychology, 52, 1258–1265. Mednick, S. A. (1962). The associative basis of the creative process. Psychological Review, 69, 220–232. Miller, G. F., & Tal, I. R. (2007). Schizotypy versus openness and intelligence as predictors of creativity. Schizophrenia Research, 93, 317–324. Nettle, D. (2006). Schizotypy and mental health amongst poets, visual artists, and mathematicians. Journal of Research in Personality 40, 876–890. O’Reilly, T., Dunbar, R., & Bentall, R. (2001). Schizotypy and creativity: an evolutionary connection. Personality and Individual Differences, 31, 1067–1078. Peterson, J. B., & Carson, S. (2000). Latent inhibition and openness to experience in a high-achieving student population. Personality and Individual Differences, 28, 323–332. Peterson, J. B., Smith, K., & Carson, S. (2002). Openness and extraversion are associated with reduced latent inhibition: replication and commentary. Personality and Individual Differences, 33, 1137–1147. Pickover, C. A. (1998). Strange Brains and Genius. New York: Plenum Press. Poreh, A. M., Whitman, D. R., & Ross, T. P. (1994). Creative thinking abilities and hemispheric asymmetry in schizotypal college students. Current Psychology: Developmental, Learning, Personality, Social, 12, 344–352. Prentky, R. (1989). Creativity and psychopathology: gamboling at the seat of madness. In J. A. Glover, R. R. Ronning, & C. R. Reynolds (Eds.), Handbook of Creativity. New York: Plenum Press. Rawlings, D. (1985). Psychoticism, creativity, and dichotic shadowing. Personality and Individual Differences, 6, 737–742. Rawlings, D., & Locarnini, A. (2008). Dimensional schizotypy, autism, and unusual word associations in artists and scientists. Journal of Research in Personality, 42, 465–471. Runco, M. (1991). Divergent Thinking. Norwood, NJ: Ablex. Schuldberg, D., French, C., Stone, B. L., & Heberle, J. (1988). Creativity and schizotypal traits: creativity test scores and perceptual aberration, magical
198
Shelley Carson
ideation, and impulsive nonconformity. Journal of Nervous & Mental Disease, 176, 648–657. Shaw, K. (2000). The Mammoth Book of Oddballs and Eccentrics. New York: Carroll & Graf. Peterson, J. B., Smith, K. W., & Carson, S. (2002). Openness and extraversion are associated with reduced latent inhibition: replication and commentary. Personality and Individual Differences, 33, 1137–1147. Sternberg, R. J., & O’Hara, L. A. (1999). Creativity and intelligence. In R. J. Sternberg (Ed.), Handbook of Creativity. Cambridge: Cambridge University Press. Torrance, E. P. (1968). Examples and rationales of test tasks for assessing creative abilities. Journal of Creative Behavior, 2, 165–178. Wallach, M. A., & Kogan, N. (1965). Modes of Thinking in Young Children. New York: Holt, Rinehart and Winston. Woody, E., & Claridge, G. (1977). Psychoticism and thinking. British Journal of Social and Clinical Psychology, 16, 241–248. Wuthrich, V., & Bates, T. C. (2001). Schizotypy and latent inhibition: non-linear linkage between psychometric and cognitive markers. Personality and Individual Differences, 30, 783–798.
Neurobiology
10 The phylogenetic distribution of latent inhibition R. E. Lubow
Introduction Inspired by early Darwinian theory, cross-species comparisons of learning abilities were once a focal point of experimental psychology. Today, many researchers have turned to simple organic systems, not to compare them to other species, but rather to take advantage of the relative lack of complexity of their cellular and molecular architecture for the purpose of modeling basic processes that are assumed to be operative in more intricate organisms. This approach has been used with a number of invertebrates, including the fruit fly Drosophila, the sea slug Aplysyia, and the honey bee Apis mellifera (for reviews, see, e.g., Davis, 2005; Kandel, 2001; Menzel & Muller, 1996, respectively). Many of these studies have examined the neural pathways involved in classical conditioning. More recently, however, attention also has been directed to an even simpler behavioral phenomenon, at least operationally, namely latent inhibition (LI). In the classical conditioning paradigm, the subject encounters paired presentations of the CS and US and the experimenter records changes in responsivity to the CS (i.e., CRs). In LI, the subject is first presented with a series of to-be-CSs, each of which is not followed by an event of consequence (CS0). Typically, responses to the to-be-CS are not documented. The stimulus preexposure stage is followed by one or more CSUS pairings, during which time CRs are recorded (two-stage procedure). Alternatively, the second stage may be followed by an additional stage in which the CSs again are presented without the US, and the CRs are monitored (three-stage procedure). In either case, the LI effect is represented by poorer evidence of conditioning in a group that was preexposed (PE) to the CS as compared to a group that was not preexposed (NPE). In short, when an organism is repeatedly exposed to a stimulus that is not followed by a significant consequence, it subsequently becomes less effective, as compared to a novel stimulus, in the acquisition/performance of a new
Latent Inhibition: Cognition, Neuroscience and Applications to Schizophrenia, ed. R.E. Lubow and Ina Weiner. Published by Cambridge University Press # Cambridge University Press 2010.
201
202
R.E. Lubow
association. As such, LI has been demonstrated in a wide variety of learning paradigms and in many different species. Lubow (1989, p. 107) summarized the literature as follows: There is little or no evidence for a latent inhibition effect in invertebrate species; the absence of such an effect has been shown in a mollusk and in the honey bee. Similarly, there is no convincing evidence for latent inhibition in lower vertebrates, although one can be more confident of this conclusion with goldfish than with pigeons. Various mammalian species, on the other hand, all exhibit powerful latent inhibition effects, including mouse, rabbit, rat, sheep, goat, cat, and among humans, children below the age of 6 or 7; older children and adults also show the effect, but only under preexposure conditions that mask the presentation of the to-be-tested stimulus.
The phylogenetic distribution of LI is of interest for several reasons. On the one hand, the pervasiveness of the effect in mammals suggests an evolved adaptive advantage, as would be evident for an organism that is biased to more fully process new stimuli as opposed to older, unimportant ones, thereby limiting processing overload. Such a stimulus selection process might be more relevant to organisms that have a wide range of learning capabilities than to ones that are more stimulus-bound through various unlearned reflex pathways. Nevertheless, the conclusions of the earlier summary regarding invertebrates, written over 20 years ago, and largely based on a small number of null-effect experiments, requires a re-examination, particularly in light of a number of recent studies. Many of these experiments were designed to provide evidence for the LI effect itself. When the effect is thought to have been established, the researchers then endeavor to uncover the neural pathways and structures that are required for generating it. In addition to the desire to increase the understanding of stimulus selection processes, many of these research programs have also been driven by the linkage between LI aberrations and schizophrenia (for reviews, see, e.g., Gray, J.A., 1998; Kumari & Ettinger, this volume; Lubow, 2005; Weiner, 2003; Weiner & Arad, this volume). The prospect of using the relatively simple nervous systems of invertebrate organisms to isolate the neural pathways and genetic coding of LI in order to provide a deeper understanding of LI as a general stimulus selection process and/or to gain access to underlying links with schizophrenia is certainly enticing. However, the entire enterprise depends on valid demonstrations of LI. What, then, are the criteria for asserting that the LI effect has been obtained? The most typical design uses two groups, PE and NPE, with the NPE group spending the same amount of time in the apparatus during the preexposure stage as the PE group. In a within-subject design, where the PE and NPE conditions are presented to the same subject, the PE and NPE stimuli are counterbalanced across subjects. In either case, LI effects that are attributable to an additional variable (e.g. a drug) require a 2 2 design (e.g., PE/NPE drug/no drug) with a significant interaction
The phylogenetic distribution of latent inhibition
203
that represents a greater effect of the drug variable on the PE group/condition than the NPE group/condition.1 Using the above design, LI has been demonstrated hundreds of times, mostly in rodents. Indeed, the construct validity of the animal LI model of schizophrenia is primarily based on experiments with mice and rats (for review, see Lubow, 2005). Thus, a critical question is whether the purported LI effect for any given species is the same phenomenon as the mammalian one. Unfortunately, cross-species comparisons are fraught with problems, particularly for tasks that assess learning. On the one hand, it is desirable to equate testing conditions for different species. On the other hand, one must take into account species-specific behaviors that might delegitimize a particular test. In practice, each species has been favored with its own laboratory paradigm, even in the study of such an apparently simple phenomenon as LI. Thus, LI research with rats relies primarily on the use of conditioned suppression and conditioned taste aversion (CTA) paradigms; with rabbits, conditioned nictitating membrane response; with pigeons, autoshaping (for specific examples of the correlation between species and LI procedures, see Lubow, 1989, pp. 10–56). As already noted, LI has been observed in all mammalian species that have been tested (for review, see Lubow, 1989, pp. 101–107). However, the production of LI in adult humans may require conditions that are different from those for young children and animals, specifically the use of a masking task to divert attention from the stimulus that later will be the conditioned stimulus (e.g., Lubow, 1989; Lubow & Gewirtz, 1995). This, of course, suggests a difference in underlying processes amongst these groups, which, if true, would severely undermine the animal models of schizophrenia that rely on LI data (e.g., Lubow, 2005; Weiner, 2003). Consequently, the next section will examine adult human LI and compare it to LI in other mammals. Subsequent sections will assess LI effects in other vertebrates, including birds and fish, and in invertebrate phyla, including arthropoda, mollusca, annelida, and nematoda. This, in turn, will be followed by a summary of LI research across the phylogenetic spectrum and a discussion of the significance of the apparent differences between mammalian and sub-mammalian species and human adults.
Mammals Most LI experiments with human subjects have used instrumental learning tasks with number of trials to reach a learning criterion as the dependent variable. Typically, in the stimulus preexposure stage, subjects in the PE and NPE groups 1
In any LI experiment, the nominal PE/NPE variable is confounded with the type of transition from the preexposure to conditioning stage. For the PE group, the first conditioning trial in stage-2 is marked by the appearance of a CS that is familiar in a context that also is familiar. On the other hand, for the NPE group, when the CS is presented for the first time in the conditioning stage, it appears as a novel stimulus in a familiar context. As such, the NPE stimulus may be more salient than the PE stimulus, thus providing it with an advantage in acquiring new associations (e.g., Lubow, Alek & Rifkin, 1976). Thus, the LI effect, traditionally attributed to a process uniquely occurring in the PE group, may have two sources, one from the PE group and one from the NPE groups (Lubow, 1997).
204
R.E. Lubow
engage in a masking task. For the PE group, the masking task serves to divert attention from the to-be-target stimuli, which are presented together with the stimuli of the masking task. The NPE group also engages in the masking task, but the to-be-target stimuli are absent. Subsequently, subjects in both groups must learn an association between the stimulus that was preexposed to the PE group, but not to the NPE group, and a new event. The dependent variable is some measure of learning such as number of correct responses (e.g., Braunstein-Bercovitz & Lubow, 1998) or, less frequently, response time (e.g., De la Casa & Lubow, 2001; Evans, Gray, N.S., & Snowden, 2007).
Humans The vast majority of experiments that have demonstrated LI in adults have preexposed the to-be-target stimulus while the subject was occupied with the masking task (e.g., Gray, N.S., Hemsley, & Gray, J.A., 1992; Gray, N.S., Pickering, Hemsley et al., 1992; Lubow, Ingberg-Sachs, Zalstein-Orda, & Gewirtz, 1992; Pineno, De la Casa, Lubow, & Miller, 2006; for an exception, Escobar, Arcediano, & Miller, 2003). Experiments with non-masked stimulus preexposures typically have failed to produce an LI effect (e.g., Graham & McLaren, 1998; for a review of the earlier literature, see Lubow, 1973). Most importantly, studies that have explicitly compared masked and non-masked conditions have obtained LI with the former but not the latter (Braunstein-Bercovitz & Lubow, 1998; De la Casa & Lubow, 2001; Ginton, Urca, & Lubow, 1975; Graham & McLaren, 1998; Lubow, Caspy, & Schnur, 1982). Notably, in all of these experiments, the masking task response was qualitatively different from the test task response, usually requiring the subject to indicate whether or not the target stimulus would be followed by an outcome. (For a proposal that the masking condition induces the equivalent of a learned irrelevance paradigm, see Le Pelley and Schmidt-Hansen, this volume.) In spite of the above, the conclusion that a masking task is a necessary condition for obtaining LI in adult humans may have to be qualified. LI without a masking task has been reported when the dependent variable in the test-stage was some measure of autonomic nervous system activity, either directly as with electrodermal conditioning (e.g., Lipp & Vaitl, 1992; Lipp, Siddle, & Vaitl, 1992), or indirectly, as with CTA (e.g., Arwas, Rolnick, & Lubow, 1989; Cannon, Best, Batson, & Feldman, 1983). However, these studies suffer from a number of methodological problems (see Escobar et al., 2003; Graham & McLaren, 1998; Lubow & Gewirtz, 1995). Other mammals By far, the greatest number of LI experiments have been conducted with rats and mice. Indeed, LI is easily demonstrated in these rodents in a variety of learning paradigms, the most popular of which include conditioned suppression
The phylogenetic distribution of latent inhibition
205
and CTA, but also many others (for a review, see Lubow, 1989). LI also has been obtained in sheep and goats (Lubow, 1965; Lubow & Moore, 1959), rabbits (e.g., Robinson, Port, & Stillwell, 1993; Romano, 1999; Shohamy, Allen, & Gluck, 2000; Talk, Stoll, & Gabriel, 2005), cats (McDaniel & White, 1966; Wickens, Tuber, & Wickens, 1983), dogs (Herendeen & Shapiro, 1976), and monkeys (Mineka & Cook, 1986). Although Lubow (1989) cited several other references, it is quite apparent that LI, extensively reported in humans and rodents, has received scant attention with other mammalian species.
Non-mammalian vertebrates Birds Until 1989, all but one LI study with pigeons used an autoshaping procedure. Two studies reported significant LI effects (Reilly, 1987; Tranberg & Rilling, 1978), two no effect (Tomie, Murphy, Fath, & Jackson, 1980, Exp. 2; Wasserman & Molina, 1975, Exp. 2), and one, at best, a very weak effect (Mackintosh, 1973). The latter three experiments did find interference with subsequent learning when presentations of the preexposed CS and grain hopper were unpaired (a “learned irrelevance” effect). The single non-autoshaping LI experiment with pigeons used conditioned heart rate and did not obtain LI, although, again, preexposures to unpaired light and shock interfered with subsequent learning (Cohen & Macdonald, 1971). After an extensive discussion of these experiments, with emphasis on the problems specific to the autoshaping procedure, Lubow (1989, pp. 32–37) concluded that the evidence for LI in pigeons was equivocal. Since that time, several additional LI studies with pigeons have been conducted. Linden, Savage, and Overmier (1997, Exp. 1), also using an autoshaping, did not obtain LI. However, again, preexposures to unpaired CSs and USs had a negative affect on subsequent learning. Using a simultaneous visual discrimination test task, Good and Macphail (1994, Exp. 3) did report LI in the pigeon. However, stimulus preexposure consisted of non-differential reinforcement to the two test stimuli, a procedure that is grossly different from that used in standard LI protocols, namely non-reinforced preexposure of a single stimulus that will later be the conditioned stimulus (also see Hall & Channell, 1980). In a CTA experiment, Thiele and Frieman (1994) preexposed one of several groups of pigeons to a red vinegar solution (PE), and another to a green salt solution (NPE). For both groups, the red vinegar solution was paired with LiCl. In three extinction sessions, animals were given 15 min access to a colorless vinegar solution. Although the PE group drank more than the NPE group on Day 1, there was no evidence of a significant LI effect. De la Casa and Ruiz (1990) also used pigeons in an unconventional procedure. Although they reported an LI effect on the acquisition of conditioned suppression,
206
R.E. Lubow
the differences between the PE and NPE groups were significant only on the last four of eight trials. In addition, the preexposure session was conducted on baseline, i.e. while the pigeons were responding for food on a VI schedule.
Fish Fish are the only cold-blooded vertebrates that have been used in LI experiments. In the first of these studies, Braud (1971) reported LI in goldfish. Amongst four groups, one received 60 preexposures to a 25 s light, and another group no preexposures. A third group received pairings of the light with food, and a fourth with shock. After 72 hours, the fish were given 20 trials of shuttle box avoidance training. The NPE group avoided the shock on significantly more trials than the PE group, thereby suggesting an LI effect. However, Shishimi (1985, p. 317) identified several problems with the experiment. For one, Braud did not present the course of learning over trials. Consequently, the overall difference between the PE and NPE groups may have been the result of early-trial differences in the unconditioned response to the CS (alpha responses), i.e., the novel light may have elicited more escape responses than the familiar light). In addition, given only 20 test trials, conditioned avoidance responding was unusually high (almost 60%) for the groups preexposed to CSfood and CSshock pairings, both of which differed significantly from the NPE group. Shishimi (1985), in a series of four experiments with goldfish, did not find any evidence for an LI effect. Experiment 1 was designed to see whether stimulus preexposure produced a subsequent associative learning deficit. The PE group was given 1000 10-s preexposures to a colored light over five daily sessions. The NPE group was treated similarly, but to a different colored light (colors were counterbalanced). In the second stage, all animals were given conditioned inhibition training. Tone-alone trials were paired with shock; tone–light compound trials were not paired with shock. For one group, the colored light was the same as that in preexposure, for the other group it was different. The dependent measure was general activity during the 10-s tone and tone/light presentations. Both groups acquired the discrimination with equal facility, thereby providing evidence for neither conditioned inhibition nor LI. Experiment 2 avoided the inherent problem of within-subject designs, namely stimulus generalization, which if present might obscure real stimulus preexposure effects. In the between-subject design, the NPE group received no stimulus exposures. Furthermore, an appetitive instrumental learning procedure replaced the classical conditioning of shock-induced activity of Experiment 1. Again, the LI and conditioned inhibition effects were not significant. Experiment 3 used classic shuttle-box avoidance conditioning. Subjects were preexposed to a colored light either 20, 40, 80, or 120 times. In the next stage, subjects were conditioned with either the same or a different colored light from that in preexposure. The NPE group was not exposed to either of the two colored lights,
The phylogenetic distribution of latent inhibition
207
but half were tested with one, and half with the other. Although the NPE group performed better than the PE groups, the effect was not specific to the type of preexposed stimulus, i.e., groups preexposed and tested with the same color did not differ from groups preexposed and tested with different colors. Thus, there was a generalized preexposure effect but no LI. The nonspecific preexposure effect was explored further in Experiment 4, which used the basic design of Experiment 1, but with the shuttle avoidance procedure. Once again, there was a preexposure effect that was not stimulus-specific, and there was neither significant conditioned inhibition nor LI. Kitagawa (1976)2 also studied stimulus preexposure effects in goldfish. In Experiment 1, the PE group was preexposed 22 h a day for 10 days to a pair of visual patterns that would later be used in a simultaneous discrimination task. The NPE group received similar preexposures, but to a pair of stimuli that were irrelevant to the subsequent discrimination task. In the test, the PE group had significantly poorer learning scores than the NPE group. In Experiment 2, two groups were preexposed to the to-be-relevant stimulus pair as in Experiment 1, but for either 22 h or 30 min a day over the 10 day period. The third group was not preexposed to the stimuli. Learning scores for the 30 min PE group did not differ significantly from the NPE group. Although the 22 h PE group had significantly more errors than either of the other two groups, the Figure 2 caption suggests that the two pairs of stimuli were not counterbalanced. Ferrari and Chivers (2006) examined the role of LI in the ability of fathead minnows to recognize predators. As they pointed out, many species that serve as prey for predators must learn to recognize the potential source of danger. Ferrari and Chivers demonstrated such learning by pairing a novel predator odor (nominal CS) with a conspecific skin extract that contained alarm stimuli (US). For 5 days, 60 min per day, two groups of minnows were preexposed to the CS. A third group was preexposed to distilled water. On the 6th day, one of the CS PE groups and the waterpreexposed group (NPE) received a CSUS pairing. The remaining CSPE group was given a CSwater pairing (conditioning control). On the following day, the groups were tested by presenting the CS odor and recording shoaling behavior3 and general activity. An increase in the former and a decrease in the latter reflect antipredator responses. Compared to a pre-CS baseline, the NPE group exhibited significantly more shoaling and decreased activity than the PE and conditioning control group, which did not differ from each other. Although the two measures are probably not independent of each other, the data certainly suggest an LI effect. However, with a very different set of conditions, Morin, Dodson and Dore (1990) reported that preexposing an odor to Atlantic salmon facilitated subsequent 2
3
The publication is in Japanese. The abstract, tables, and figure captions are in English. The tables contain experimental designs, data, and results from statistical analyses. Minnows were preexposed, conditioned, and tested in groups of three. The shoaling measure is the average proximity of the three fish to each other.
208
R.E. Lubow
conditioning of heart rate deceleration when the odor was paired with shock. The conditioning/test trials were administered in two 20-trial sessions, separated by 24 h. In the first session, the percent heart-rate reduction was almost nil for all groups. However, from the start of the second session, the PE groups showed strong conditioned deceleration compared to the NPE groups.
Invertebrate phyla Arthropods Phylum arthropoda is by far the most diverse in numbers of species, behaviors, and habitats. Nevertheless, LI research has been confined to two crustaceans (crab and crayfish) and two hexapods (insects; honey bee and assassin bug). As already noted, the ability to generate LI effects may have important consequences for predator–prey relations. On the one hand, unreinforced presentations of a predatorcue may subsequently interfere with the ability of prey to learn of the potential dangers of that predator. On the other hand, unreinforced presentations of cues that would later be associated with prey could interfere with the ability of the predator to recognize its next meal. The first scenario has been the subject of the two LI experiments with crustaceans. Crab In nature, the crab Chasmagnathus granulatus is preyed upon by gulls, whose overhead presence elicits reflexive escape running. In the laboratory, a similar response can be obtained by passing an opaque card over the crab. With repeated presentations of the US, the escape response habituates, and remains so even after a 5-day retention interval (Pedreira, Dimant, Tomsic et al., 1995). The long-term habituation is adversely affected by a change of context from the habituation stage to the test stage (e.g., Pedreira et al., 1995; Tomsic, Pedreira, Romano et al., 1998). Tomsic et al. (1998) investigated the crab’s acquisition of context–US associations and their role in maintaining the long-term habituation to the US (also see Rankin, 2000). Experiment 4 was designed to see whether exposure to the context prior to presenting the US habituation trials would interfere with the acquisition of the context–US association (LI). One pair of groups was preexposed for 12 h to the context that would later be used in the habituation and test stages; another pair of groups was preexposed to a context that was different from the one to be used in the habituation and test stages. After the context preexposure stage, one group from each pair received 15 presentations of the US (habituation trials) and the other received none. The groups that received the 15 US trials showed an orderly decrease in the magnitude of the escape response over trials, with no difference between them. However, when a test trial was administered 24 h after the habituation session, the group that had neither been preexposed to the context nor received US habituation
The phylogenetic distribution of latent inhibition
209
trials exhibited a significantly stronger escape response to the US than the group that was not preexposed to the context but received the 15 habituation trials (the basic habituation effect). On the other hand, the groups that had received preexposure to the context, one followed by the 15 habituation trials and the other not, did not differ from each other. The authors interpret these data as an LI effect. However, this conclusion is questionable since the difference between the two groups that had habituation training, one preexposed to the context (PE) and the other not preexposed to the context (NPE), was not significant. Crayfish Acquistapace, Hazlett and Gherardi (2003) preexposed one group of crayfish to a neutral stimulus (goldfish odor). A second group was not preexposed. On the day after the last preexposure, the goldfish odor was paired with a predator-associated alarm odor. The crayfish were then transferred to individual aquaria where they spent the next 24 h. In the test, several food-related behaviors were recorded during three 2-min periods, one each for control, food odor, goldfish odor. If the goldfish odor was associated with the alarm odor, then the feeding response to the food should be suppressed. LI would be evident if the PE group showed less suppression of feeding-related movements than the NPE group. Indeed, food-related behaviors in the presence of the goldfish odor were significantly less than in the food-odor period for the NPE group, but not for the PE group. However, as in Tomsic et al. (1998), the PE and NPE groups were not compared, and the PE/NPE X test odor interaction was not evaluated. In short, the experiment falls short of providing convincing evidence for an LI effect. Honey bees The honeybee, unlike most other invertebrates, is a forager. As suggested by Bitterman, Menzel, Fietz, and Schafer (1983), LI might be a highly adaptive specialization that enhances foraging efficiency by devaluing stimuli (i.e., particular colors and odors) not associated with reward (i.e., pollen and nectar). In an early experiment, they classically conditioned an appetitive response in the honey bee. The CS was a 6-s scented air stream that was blown across the antenna and proboscis. The US, sugar water, also applied to the antenna and proboscis, was presented immediately after the CS. The unconditioned response was an extension of the proboscis in the presence of the sugar solution. A response was scored as conditioned when the proboscis extension occurred within 3 s of the onset of the scented air stream. Conditioning was very rapid, developing within two or three trials. Although eight unpaired CS and US preexposures interfered with subsequent conditioning, eight preexposures of CS-alone failed to retard acquisition of the CSUS association, i.e., there was no LI effect. In a subsequent experiment, Abramson and Bitterman (1986) used aversive conditioning to train free-flying bees. During feeding in the laboratory, either the
210
R.E. Lubow
CS (substrate vibration or air stream) was paired with the US (brief shock), or the CS and US were presented unpaired. The shock could be avoided by either withdrawing the proboscis from the food solution or flying up from the food station. In Experiment 1, the PE group received 10 preexposures to the CS while feeding. The NPE group was not preexposed to the to-be-conditioned stimulus. In the subsequent test, an apparent LI effect was demonstrated when the NPE group acquired the conditioned avoidance response significantly faster than the PE group. Experiment 2, based on a discrimination learning task, used three groups of honey bees. One group was preexposed to the to-be reinforced stimulus (CSþ); one group was preexposed to the to-be non-reinforced stimulus (CS); and an NPE group was not preexposed to either of those stimuli. As in Experiment 1, the group preexposed to CSþ exhibited retarded avoidance learning to CSþ compared to the NPE group. However, the group preexposed to CS showed stronger avoidance conditioning to the CS than the NPE group, suggesting that stimulus preexposure may have produced conditioned inhibition rather than LI. This possibility was confirmed by a summation test in Experiment 3. Chandra, Hosler, and Smith (2000) also used appetitive conditioning of proboscis extension to investigate LI in restrained honey bees. They began by providing drones and queens with 30 preexposures to the to-be-conditioned odor. This was followed by six CSUS trials (odor–sucrose). A “low inhibitor” group was created by selecting animals that made two or fewer CRs; a “high inhibitor” group was created by selecting bees that made four or more CRs. Virgin queens and drones with high levels of “inhibition” were mated, as were those with low levels of “inhibition”. The progeny of these two groups were then tested in a standard between-subject LI design (PE and NPE groups). LI was greater in the offspring of the “high inhibition” than in the “low inhibition” group. “High inhibition” progeny also exhibited poorer reversal learning than “low inhibition” progeny, suggesting to the authors that LI and reversal learning might have a common underlying mechanism that is at least partly genetically determined. It should be noted that the selection procedure, aimed at selecting for low and high LI, confounds initial learning ability with assignment to preexposure condition. Nevertheless, if the behaviors of the subsequent generations were such that the differences between the PE and NPE groups were larger for the “high inhibition” group than for the “low inhibition” group, and that this effect was primarily due to differences between the two PE groups, then the consequences of the confounding would be minimized. Although Chandra et al.’s (2000) Figure 4 suggests that the desired interaction was obtained, the appropriate statistic was not reported. In a related study, Ferguson, Cobey, and Smith (2001) selected progeny on the basis of performance on a reversal discrimination task. They obtained a significant LI effect. However, in opposition to the findings of Chandra et al. (2000), only fast reversers showed the effect. The difference between the two studies may be due to the fact that Chandra et al. selected progeny on the basis of the performance of drones
The phylogenetic distribution of latent inhibition
211
and queens, while Ferguson et al. (2001) only selected from drones, thus making their genetic contribution haploid. Chandra, Hunt, Cobey, and Smith (2001) continued this line of research. In the earlier experiments, Ferguson et al. (2001) selected lines on the basis of the speed of discrimination reversal learning and then tested for LI, and Chandra et al. (2000) selected on the basis of LI (at least nominally; see above), and then tested the progeny on a reversal learning task. In the present experiment, honey bees were selected and bred on the basis of differences in speed of reversal learning, followed by testing separate groups of offspring for reversal learning and LI. In regard to LI, subjects received 36 preexposures to the to-be-conditioned odor followed by six odor–sucrose conditioning trials. Although the authors refer to an LI effect, the absence of an NPE group makes such an attribution problematic (also see Latshaw & Smith, 2005). Furthermore, in the above-cited experiments, the CS odor elicits proboscis extension (the nominal CR) before being paired with the US. This can be seen in Figure 2 of Chandra, Hosler, and Smith (2000), where subjects emitted “CRs” on trial-1, i.e., before the first CSUS pairing. Such data are potentially available during the stimulus preexposure stage, but are often not collected or reported. As a consequence, in the test stage URs to the CS may be erroneously scored as CRs. Since the number of such “false CRs” would be different for PE and NPE groups, the inclusion of such responses in the data analyses results in an overestimation of the LI effect. Although the problem can be detected in the data from the first trial, its effects may persist over many trials. Drosophila Beck, Schroeder, and Davis (2000) presented one group of fruit flies with a single exposure to a 10-s odor. A second group was not preexposed. Following the preexposure phase, both groups received a classical conditioning trial in which the odor was paired with shock. In the test, groups were place in a start box of a T-maze with the conditioned odor streaming through one arm and a different odor in the second arm. The PE and the NPE group conditioned very rapidly, and there was no LI effect, the absence of which may have been due to the relatively weak LI manipulation. Assassin bug Abramson, Romero, Frasca, Fehr, Lizano, and Aldana (2005) reported an LI effect in the assassin bug (a vector for Chagas disease via a protozoan parasite). In Experiment 2, bugs received 12 preexposures to one of three odors. This was followed by 12 pairings of the odor with exposures to 42 C temperature (US), which elicits proboscis extension. There was no NPE group, and LI was inferred by comparison to Experiment 1, where the same acquisition/test procedure was not preceded by a preexposure stage. The authors reported significantly more conditioned responding
212
R.E. Lubow
in Experiment 1 compared to Experiment 2 for two of the three CSs. However, the authenticity of the LI effect is doubtful. In addition to using cross-experiment comparisons, the P values were inappropriately based on paired t-tests.
Molluscs Several species of gastropods have been popular subjects for experiments on learning and memory. A few of these studies have also searched for LI effects, the rationale for which is often based on the premise that LI is a basic associative learning phenomenon, and the fact that these organisms have a small number of relatively simple and observable clusters of neurons that are involved in learning. Thus, the LI gastropod preparation should provide a convenient means to study the neural substrates of associative processes. Nevertheless, although the aquatic slug Hermissenda crassicornis is capable of associative learning on the basis of simple contiguity (e.g., Crow & Alkon, 1978; Farley & Alkon, 1982) and contingency (Farley, 1987a, 1987b), Farley (1987a) demonstrated that such learning was not subject to LI effects. Fifty preexposures of the CS, as opposed to zero, failed to produce differences in conditioning performance when the test session was conducted either 24, 48, or 72 h after the conditioning session. However, Loy, Fernandez, and Acebes (2006) did report LI effects in the common garden snail Helix aspersa. Experiment 2 preexposed one group to an apple odor, and another to a pear odor (10 min per day for 6 days). On the next two days, half of each group had the preexposed odor paired with 10 min of access to the US (carrot feeding). For the other half, the non-preexposed fruit odor was paired with the US. In the test, conducted on the next day, the conditioned odor was presented for 2 min, and CRs (lowering of a tentacle) were recorded. The preexposure, conditioning, and test series was repeated three times. On each of the three test trials, the group that was preexposed, conditioned and tested with the same odor (standard PE group) exhibited significantly fewer conditioned feeding responses to that odor than the group that was preexposed to one odor and conditioned and tested with a different odor. Experiment 3 used a within-subject PE/NPE manipulation. There were fewer CRs in the PE condition than in the NPE condition on the second and third test trials. Although the two experiments obtained robust stimulus preexposure effects, the absence of data regarding possible PE stimulus-elicited tentacle lowering during the preexposure stage makes an LI interpretation of the test results disputable.
Annelids Sahley, Boulis, and Schurman (1994) conducted a series of experiments with the semi-intact leech Hirudo medicinalis. Using a tactile stimulus as the CS, and a mild shock that elicits a shortening reflex of the anterior sucker as the US, they examined
The phylogenetic distribution of latent inhibition
213
the affects of CSUS pairings, unpredicted USs, and stimulus preexposures on conditioned performance. Groups that received only CSUS pairings exhibited a reliable increase in the magnitude of the conditioned withdrawal response, an effect that was similar to that reported for the intact leech (Sahley & Ready, 1988). Degrading the CSUS contingency reduced conditioned performance (Experiment 2), as did non-reinforced preexposures to the CS (Experiment 3). Although they reported positive evidence of LI in the leech, the data are compromised by the fact that the unconditioned response to the CS was the same as the conditioned response. Indeed, Experiment 1 demonstrated that alpha responses declined with repeated non-reinforced exposures of the tactile stimulus. The possible inclusion of alpha responses in Experiment 3 is obscured because the test stage data were converted to percent of responding relative to a pretest baseline, thereby eliminating differences in initial response levels.
Nematodes Rankin (2000) performed a series of experiments with the roundworm Caenorhabditis elegans to determine whether context–US associations affect response habituation. After ascertaining that a weak solution of sodium acetate did not interfere with the elicitation of withdrawal responses (distance traveled backwards) to a series of body taps, the solution was subsequently used as a context cue. In Experiments 2 and 5, worms were given 30 habituation trials in one of two contexts, either with or without the sodium acetate solution. One hour later, retention of the habituated response was tested in a context that was either the same or different from that of the habituation stage. Overall, the habituated response was adversely affected by a change of context from the habituation stage to the test stage. If, indeed, the nematodes acquired an association between US and context, and that association affects habituation performance, then exposure of context cues prior to habituation training should lead to an LI effect, namely a strengthened unconditioned withdrawal response in the test as compared to a group that was not preexposed to the context. In Experiment 4, one group of subjects was preexposed to the context, habituated to the US, and tested with the US, all in the sodium acetate context. The second group was treated the same way except that it was not preexposed to the sodium acetate cue. In the test, context-preexposed worms exhibited an insignificant decrease in withdrawal response strength (the first two trials in test compared to the first two trials in habituation). Worms that were not preexposed to the context cue showed a significant decrease in responding from habituation stage to test stage, suggesting that the former group was deficient in acquiring the context–US association as compared to the latter group. However, the apparent LI effect was compromised on two counts. (1) The contexts were not counterbalanced. (2) The presence of the LI effect was based on two separate t-tests, one indicating that the context PE group had significantly weaker withdrawal responses in the test stage as compared to the habituation stage;
214
R.E. Lubow
the other indicating that the comparable data for NPE group were in the same direction as for the PE group, but not significant. In short, the test data from PE and NPE groups were not compared, nor was the habituation/test PE/NPE interaction reported.
Summary and speculative conclusions Latent inhibition in invertebrates As described above, LI research with invertebrates is sparse; with arthropods, it has been limited to two crustaceans (crab and crayfish) and three insects (honey bee, fruit fly, and assassin bug), for a total of 10 experiments, six of which were with the honey bee. Even less plentiful are LI studies with molluscs (aquatic slug, common garden snail), annelids (leech), and nematodes (roundworm). Overall, in spite of claims to the contrary, there is little evidence to support the presence of LI amongst invertebrates. As noted, studies that claim such an effect suffer from a variety of design, statistical, and interpretative problems. In particular, since LI experiments with invertebrates rely on classical conditioning preparations, they are exceptionally vulnerable to confounded effects from habituation4 and sensitization, which can either obscure or falsely identify LI, depending on whether the conditioning procedure calls for an increase or decrease of the target response. These problems are particularly acute when the unconditioned response to the preexposed stimulus is the same as the CR in the conditioning or post-conditioning test-stage, a situation that is common in invertebrate conditioning preparations. Habituation of neophobia during the preexposure stage provides a special case for commonly used feeding-related stimuli. As a consequence of such habituation, the PE group may consume more of the familiar conditioning and/or test stimulus than the NPE group, for whom the stimulus is relatively novel. Overcoming these various sources of contamination requires special control procedures, which, unfortunately, are often lacking (for a fuller discussion of these issues, see Farley, Jin, Huang, & Kim, 2004). As a consequence, and in spite of the fact that there is ample evidence from a variety of species for conditioning with invertebrates (e.g., Hawkins, Greene, & Kandel, 1998), the case for LI with invertebrates has yet to be made.
4
Although the LI effect occasionally has been attributed to the operation of a habituation process during the preexposure stage, there are several reasons for dismissing this position, particularly in regard to simple response habituation. (1) In most LI experiments (with the exception of conditioned taste aversion), the dependent variable is some response other than the one elicited by the preexposed stimulus. (2) LI to a preexposed stimulus will be evident after pairing with either an appetitive or an aversive US. (3) LI and habituation are dissociable, as demonstrated by the context-dependency of LI but not habituation (e.g., Hall & Honey, 1989; for a review, see Hall, 1991), and by their differential sensitivity to selective brain lesions (e.g., Oswald, Yee, Rawlins et al., 2002).
The phylogenetic distribution of latent inhibition
215
Latent inhibition in non-mammalian vertebrates The present review suggests that the evidence for LI in pigeons also is inconclusive.5 Similarly, there is little evidence for LI in fish. In the most carefully controlled of these studies, Shishimi (1985) did not demonstrate LI, but did find that the stimulus preexposure effects were not stimulus-specific. Unfortunately, most LI studies do not differentiate between specific and general preexposure effects, a design feature that may be particularly relevant for experiments with invertebrates as well as non-mammalian vertebrates, where distinctions between associative and non-associative effects are often not clear.
Latent inhibition in mammals: the role of the masking task LI effects in mammals, particularly in humans and rodents, are particularly robust. However, a comparison of conditions that appear to be necessary for the production of LI in humans and in other mammals raises a question regarding the similarity of underlying processes. Specifically, generating LI in adult humans requires a masking task in the preexposure stage, a condition that is not needed with other mammals. As such, there is a potential paradox. On the one hand, LI is present only in organisms with relatively well-developed cortices. On the other hand, humans require a masking task to divert attention from the preexposed stimulus, and it is this functional decorticalization that allows for LI. Some aspect of cortical function would seem to be a necessary but insufficient condition for producing LI. It must be present, but occupied with events other than the preexposed stimulus, a condition that is promoted by the masking task. However, this position would appear to be inconsistent with the scores of studies that have obtained LI in rats and mice without a masking task. There are at least two approaches to this problem. A rat placed in a novel test chamber during the preexposure phase of an LI experiment engages in vigorous exploratory behavior. The to-be-CSs are presented while the animal is sniffing and stretching, investigating the walls and corners of the apparatus, a scene that resembles the masked preexposure session in human LI studies. In both cases, the irrelevant to-be-CS occurs while the subject is engaged in an activity that is not related to that stimulus. If exploratory behavior serves as a masking task for the rat, then preexposing the rat to the apparatus until that behavior has habituated, and then presenting the to-be-CSs, should disrupt LI. Although one might argue that prior exposure of the preexposure phase context interferes with the acquisition of an association between the preexposed CS and context, an association that has been posited as the basis of the LI effect (e.g., Lubow & Gewirtz, 1995; Wagner, 1978), this interpretation would not negate the claim for a functionally common masking task in animal and human LI 5
In spite of the lack of convincing evidence for LI in pigeons, the effect has been invoked as an explanation for a number of different avian experimental findings, from perceptual differentiation during categorization (e.g., Aitken, Bennett, McLaren, & Mackintosh, 1996) to observational learning (e.g., Biederman & Vanayan, 1988).
216
R.E. Lubow
experiments. The masking task in human LI studies may have exactly that role, to promote an association between the unattended to-be-CS and the context in which it is presented. If all mammalian LI is controlled by the same basic processes, then these processes must be relatively “primitive” and fundamental, and independent of higher cognitive functions, which when present must be bypassed. This suggests that the processes governing LI are automatically engaged whenever a task-irrelevant stimulus is present. However, adult humans can inhibit activation of these mechanisms by the conscious, controlled allocation of attention to the preexposed stimulus. The purpose of the masking task in the human LI experiment is to subvert such top-down attempts and thereby allow the acquisition of the S–context association, and the subsequent display of the LI effect. . . . the difference between human adults, on one hand, and infrahumans and young children, on the other hand, derives from the propensity in adults for controlled processing to override automatic processing. Accordingly, the function of the masking task would be to engage controlled attention, thereby not allowing the otherwise present demand characteristic of the to-be-relevant preexposed stimulus to elicit controlled processing. (Lubow & Gewirtz, 1995, pp. 93–94)
The context-preexposed stimulus association Since LI serves to increase the selectivity and efficiency of the acquisition and/or expression of associative learning, the conclusion, albeit tentative, that relatively primitive organisms are deficient in such a capability should not come as a surprise. LI would be expected to be most prominent in those species for which learning is an important factor for survival, and not in those for which survival is largely determined by innate, reflexive responses. Indeed, the presumed absence of invertebrate LI, together with the rich evidence for classical conditioning in those organisms, suggests that LI, in spite of its apparent simplicity, may require an even more complex neural infrastructure than that required for classical conditioning, one that can accommodate the critical Scont–S association.6 A standard CSUS association can be inferred when the CS comes to elicit a response that was formerly restricted to the US. A Scont–S association is more difficult to detect, particularly as reflected in the view that repeated coincident presentation of S and Scont incorporates S into the representation of the background 6
There is considerable evidence that LI is critically dependent on the acquisition of the Scont–S association. (1) A change of context from preexposure to test disrupts LI (in humans, e.g., Gray, N.S., et al., 2001; Lubow, Alek & Rifkin, 1976; Zalstein-Orda & Lubow, 1995, Exp. 1; in rats, e.g., Hall & Honey, 1989; Honey & Good, 1993). (2) LI is dependent on stimulus and context preexposure occurring conjointly (in humans, Zalstein-Orda & Lubow, 1995, Exp. 4); in rats (Channell & Hall, 1981; Hall & Channell, 1986). (3) Presenting the context by itself after the acquisition of Scont–S association in the preexposure stage extinguishes that association and thereby attenuates LI (e.g., in rats, Escobar, Arcediano, & Miller, 2002; Grahame, Barnet, Gunther, & Miller, 1994; Talk, Stall, & Gabriel, 2005; but see Hall & Minor, 1984; Baker & Mercier, 1982; Zalstein-Orda & Lubow, 1995, Exp. 3).
The phylogenetic distribution of latent inhibition
217
context (cf., representational compression, Gluck & Myers, 1996; also see Lubow, this volume). Such an account deprives S of an independently perceived identity; it is unattended. Although its identity is temporarily lost in the surrounding context, the existence of the latent S can be recovered. A representation of S can be detected by procedures based on omission or reorganization. For example, just as the noise from a refrigerator’s compressor becomes retrospectively available only when it ceases, removing S from the context may produce a similar effect, as assessed by self-report (with humans) or by autonomic measures associated with the orienting response (OR, e.g., Sokolov, 1963). However, with the latter, it would be difficult to know whether the OR represented the system’s recognition of the loss of a specific S or of a change in the context, a distinction that would be theoretically significant. The LI procedure provides a different approach to the problem, namely preexposing S in a particular context, such that it presumably becomes an integral part of that context, and then making the solution of a subsequent learning task dependent on the ability of the organism to re-acquire the independent identity of S. The difficulty with which this is accomplished, relative to an appropriate control group, represents the degree to which S had become an integral part of the context. As such, the LI procedure produces an effect that is independent of a phenomenal report. More importantly, it does not depend on omission of the S in the test, and therefore it avoids the problem of whether the induced OR is attributable to the absence of S or a change of Scont. The difference between vertebrates (arguably only mammals) and invertebrates in their capacity to generate LI would appear to lie in the latter’s inability to acquire Scont–S associations. This lack of ability may stem from a limited neural infrastructure that either cannot represent context information or cannot associate/compress two biologically irrelevant stimuli (S and Scont). If the first hypothesis is true, then, of course, the second one becomes unviable. Although vertebrates unquestionably can represent context information, the evidence for invertebrates is problematic. On the one hand, there are reports that US habituation in invertebrates is context-specific (e.g., Pedreira et al., 1995; Rankin, 2000; Tomsic et al., 1998), suggesting that they can acquire US–context associations, and therefore that they can represent context. On the other hand, the relative simplicity of the invertebrate nervous system makes that notion somewhat fanciful, perhaps residing more in the brains of the experimenters than in the relatively simple organisms that they study. The sensory world of the invertebrate may be so restricted that there is no meaningful distinction between stimulus and context. When, for example, a worm is tapped (S) while it is bathed in a “contextual” sodium acetate solution (Rankin, 2000), the two events may be represented as a simple union. The worm’s world may be composed of a single miasmic stream, differentiated and in flux only over time, rather than being composed of a mosaic of sensory experiences that would constitute a context. If true, then it is more likely that the degraded habituated response that accompanies a change of context from the training to test stage is the
218
R.E. Lubow
result of a stimulus generalization decrement rather than a disruption of a Scont–S association. As already noted, to the limited extent that invertebrates have a stimulus selection problem, it is largely solved by bottom-up reflexive mechanisms. A generalization decrement effect would be one such example.
A final note The apparent absence of LI in invertebrates and its doubtful presence in nonmammalian vertebrates supports a model of LI that involves a level of information processing that is capable of representing context. Since there is abundant evidence that the hippocampus is critical for processing context information (e.g., O’Keefe & Nadel, 1978), LI should only be present in species with hippocampal structures. Indeed, although homologous proto-hippocampal formations have been described in fish and birds (e.g., Colombo & Broadbent, 2000; Rodriguez et al., 2002), but not in invertebrates, such structures are vastly more complex in mammals, and peak in adult humans. Notably, highly elaborated hippocampal formations are accompanied by an increase in cortical development and hippocampal–cortical projections, a neurological substrate whose dysfunction is linked to schizophrenia. More specifically, LI, and reversal-related behaviors in general, appears to be modulated by the hippocampus via an interaction with the mesolimbic dopaminergic system on the site of the nucleus accumbens (e.g., for review, see Weiner, 2003). The connections between LI and schizophrenia (e.g., see Kumari & Ettinger, this volume; Weiner & Arad, this volume), the importance of context in the acquisition and maintenance of LI (e.g., see De la Casa & Pineno, this volume; Lubow, this volume; also footnote 6), the involvement of the hippocampal structures in LI (e.g., see Honey, Iordanova, & Good, this volume), and the generally reduced ability of schizophrenics to process context information (e.g., Cohen, Barch, Carter, & ServanSchreiber, 1999) point to the broad significance of research on the processing of irrelevant stimuli, an activity that will have an impact on our understanding of information processing in general, and specifically on its disruption in schizophrenia. Well over 100 years ago, Bernard Berenson, the historian and critic of renaissance art, beautifully captured the role of LI in reducing the competition from irrelevant stimuli and thereby sharpening the focus of attention. Nor did he overlook the consequences for a comparative psychology. For nature is chaos, indiscriminately clamoring for attention. Even in its least chaotic state it has much more resemblance to a freakish and whirlingly fantastical “Temptation of St. Anthony” by Bosch, than to compositions by Duccio . . . or to others by Raphael. . . . To save us from this contagious madness of this cosmic tarantella, instinct and intelligence have provided us with stout insensibility and inexorable habits of inattention, thanks to which we stalk through the universe tunneled in and protected on every hand, bigger than the ants and wiser than the bees. (Berenson, 1952/1980, p. 104)
The phylogenetic distribution of latent inhibition
219
References Abramson, C. I., & Bitterman, M. E. (1986). Latent inhibition in honey bees. Animal Learning & Behavior, 14, 184–189. Abramson, C. I., Romero, E. S., Frasca, J., et al. (2005). Psychology of learning: A new approach to study behavior in Rhodinius prolixus stal under laboratory conditions. Psychological Reports, 97, 721–731. Acquistapace, P., Hazlett, B. A., & Gherardi, F. (2003). Unsuccesful predation and learning of predator cues by crayfish. Journal of Crustacean Biology, 23, 364–370. Aitken, M. R. F., Bennett, C. H., McLaren, I. P. L., & Mackintosh, N. J. (1996). Perceptual differentiation during categorization learning by pigeons. Journal of Experimental Psychology: Animal Behavior Processes, 22, 43–50. Arwas, S., Rolnick, A., & Lubow, R. E. (1989). Conditioned taste aversion in humans using motion-induced sickness as the US. Behaviour Research and Therapy, 27, 295–301. Baker, A. G., & Mercier, P. (1982). Extinction of the context and latent inhibition. Learning and Motivation, 13, 391–416. Beck, C. D. O., Schroeder, B., & Davis, R. L. (2000). Learning performance of normal and mutant Drosophila after repeated conditioning trials with discrete stimuli. Journal of Neuroscience, 20, 2944–2953. Berenson, B. (1980). The Italian Painters of the Renaissance. Ithaca, NY: Cornell University Press. (Originally published in 1952, but based on much earlier texts.) Biederman, G. B., & Vanayan, M. (1988). Observational learning in pigeons: the function of quality of observed performance in simultaneous discrimination. Learning & Motivation, 19, 31–43. Bitterman, M. E., Menzel, R., Fietz, A., & Schafer, S. (1983). Classical conditioning of proboscis extension in honeybees. Journal of Comparative Psychology, 97, 107–119. Braud, W. G. (1971). Effectiveness of “neutral”, habituated, shock-related, and food-related stimuli and CSs for avoidance learning in goldfish. Conditioned Reflex, 6, 153–156. Braunstein-Bercovitz, H., & Lubow, R. E. (1998). Are high schizotypal normal participants distractible or limited in attentional resources? A study of latent inhibition as a function of masking task load and schizotypy. Journal of Abnormal Psychology, 107, 659–670. Cannon, D. S., Best, M. R., Batson, J. D., & Feldman, M. (1983). Taste and familiarity and apomorphine-induced taste aversions in humans. Behavior Research and Therapy, 21, 969–973. Chandra, S. B. C., Hosler, J. S., & Smith, B. H. (2000). Heritable variation for latent inhibition and its correlation with reversal learning in honeybees (Apis mellifera). Journal of Comparative Psychology, 114, 73–85. Chandra, S. B. C., Hunt, G. J., Cobey, S., & Smith, B. H. (2001). Quantitative trait loci associated with reversal learning and latent inhibition in honeybees (Apis mellifera). Behavior Genetics, 31, 275–285. Channell, S., & Hall, G. (1981). Facilitation and retardation of discrimination learning after exposure to the stimuli. Journal of Experimental Psychology: Animal Behavior Processes, 7, 437–446.
220
R.E. Lubow
Cohen, D. H., & MacDonald, R. L. (1971). Some variables affecting orienting and conditioned heart-rate responses in the pigeon. Journal of Comparative and Physiological Psychology, 74, 123–133. Cohen, J. D., Barch, D. M., Carter, C., & Servan-Schreiber, D. (1999). Contextprocessing deficits in schizophrenia: converging evidence from three theoretically motivated cognitive tasks. Journal of Abnormal Psychology, 108, 120–133. Colombo, M., & Broadbent, N. (2000). Is the avian hippocampus a functional homologue of the mammalian hippocampus? Neuroscience and Biobehavioral Reviews, 24, 465–484. Crow, T. J., & Alkon, D. L. (1978). Retention of an associative behavioral change in hermissenda. Science, 201, 1239–1241. Davis, R. L. (2005). Olfactory memory formation in drosophila: from molecular to systems neuroscience. Annual Review of Neuroscience, 28, 275–302. De la Casa, L. G., & Lubow, R. E. (2001). Latent inhibition with a response time measure from a within-subject design: effects of number of preexposures, masking task, context change, and delay. Neuropsychology, 15, 244–253. De la Casa, L. G., & Ruiz, G. (1990). Contextual specificity of latent inhibition in pigeons (Columba livia): attenuation of the effect of the delay in the acquisition of a CER in familiar contexts. Psicologica, 11, 177–188. Escobar, M., Arcediano, F., & Miller, R. R. (2002). Latent inhibition and contextual associations. Journal of Experimental Psychology: Animal Behavior Processes, 28, 123–136. Escobar, M., Arcediano, F., & Miller, R. R. (2003). Latent inhibition in human adults without masking. Learning, Memory and Cognition, 29, 1028–1040. Evans, L., Gray, N. S., & Snowden, R. J. (2007). A new continuous within-participant latent inhibition task: examining associations with schizotypy dimensions, smoking and gender. Biological Psychology, 74, 365–373. Farley, J. (1987a). Contingency learning and causal detection in Hermissenda: I. Behavior. Behavioral Neuroscience, 101, 13–27. Farley, J. (1987b). Contingency learning and causal detection in Hermissenda: II. Cellular mechanisms. Behavioral Neuroscience, 101, 28–56. Farley, J., & Alkon, D. L. (1982). Associative neural and behavioral change in Hermissenda – consequences of nervous system orientation for light and pairing specificity. Journal of Neurophysiology, 48, 785–507. Farley, J., Jin, I., Huang, H., & Kim, J.-I. (2004). Chemosensory conditioning in molluscs: II. A critical review. Learning & Behavior, 32, 277–288. Ferguson, H. J., Cobey, S., & Smith, B. H. (2001). Sensitivity to a change in reward is heritable in the honeybee, Apis mellifera. Animal Behaviour, 61, 527–534. Ferrari, M. C. O., & Chivers, D. P. (2006). The role of latent inhibition in acquired predator recognition by fathead minnows. Canadian Journal of Zoology, 84, 505–509. Ginton, A., Urca, G., & Lubow, R. E. (1975). The effects of preexposure to a nonattended stimulus on subsequent learning: latent inhibition in adults. Bulletin of the Psychonomic Society, 5, 5–8. Gluck, M. A., & Myers, C. E. (1996). Integrating behavioral and physiological models of hippocampal function. Hippocampus, 6, 643–653.
The phylogenetic distribution of latent inhibition
221
Good, M., & Macphail, E. M. (1994). Hippocampal lesions in pigeons (Columbia livia) disrupt reinforced preexposure but not overshadowing or blocking. Quarterly Journal of Experimental Psychology, 47, 263–291. Graham, S., & McLaren, I. P. L. (1998). Retardation in human discrimination learning as a consequence of preexposure: Latent inhibition or negative priming. Quarterly Journal of Experimental Psychology, 51B, 155–172. Grahame, N. J., Barnet, R. C., Gunther, L. M., & Miller, R. R. (1994). Latent inhibition as a performance deficit resulting from CScontext associations. Animal Learning and Behavior, 22, 395–408. Gray, J. A. (1998). Integrating schizophrenia. Schizophrenia Bulletin, 24, 249–266. Gray, N. S., Hemsley, D. R., & Gray, J. A. (1992). Abolition of latent inhibition in acute, but not chronic, schizophrenics. Neurology, Psychiatry and Brain Research, 1, 83–89. Gray, N. S., Pickering, A. D., Hemsley, D. R., Dawling, S., & Gray, J. A. (1992). Abolition of latent inhibition by a single 5 mg dose of d-amphetamine in man. Psychopharmacology, 107, 425–430. Gray, N. S., Williams, J., Fernandez, M., et al. (2001). Context dependent latent inhibition in adult humans. Quarterly Journal of Experimental Psychology 54B, 233–245. Hall, G. (1991). Perceptual and Associative Learning. Oxford: Clarendon Press. Hall, G., & Channel, S. (1980). A search for perceptual differentiation produced by nondifferential reinforcement. Quarterly Journal of Experimental Psychology, 35B, 135–147. Hall, G., & Channell, S. (1986). Context specificity of latent inhibition in taste aversion learning. Quarterly Journal of Experimental Psychology, 38B, 121–139. Hall, G., & Honey, R. C. (1989). Contextual effects in conditioning, latent inhibition, and habituation-associative and retrieval functions of contextual cues. Journal of Experimental Psychology: Animal Behavior Processes, 15, 232–241. Hall, G., & Minor, H. (1984). A search for context stimulus associations in latent inhibition. Quarterly Journal of Experimental Psychology, 36B, 145–169. Hawkins, R. D., Greene, W., & Kandel, E. R. (1998). Classical conditioning, differential conditioning, and second-order conditioning of the Aplysia gill-withdrawal reflex in a simplified mantle organ preparation. Behavioral Neuroscience, 112, 636–645. Herendeen, D. L., & Shapiro, M. M. (1976). Specific and generalized adaptation of salivary conditioning in dogs. Bulletin of the Psychonomic Society, 18, 68–71. Honey, R., & Good, M. (1993). Selective hippocampal lesions abolish the contextual specificity of latent inhibition and conditioning. Behavioral Neuroscience, 107, 23–33. Jin, I., Huang, H., Kim, J.-I., & Farley, J. (2004). Chemosensory conditioning in molluscs: I. Failure of contextual conditioning in Hermissenda. Learning & Behavior, 32, 257–276. Kandel, E. R. (2001). The molecular biology of memory storage: a dialogue between genes and synapses. Science, 294, 1030–1038. Kitagawa, T. (1976). Effects of preexposure to discriminative stimuli on discrimination learning in goldfish. Annual of Animal Psychology, 26, 13–22. Latshaw, J. S., & Smith, B. H. (2005). Heritable variation in learning performance affects foraging preferences in the honey bee (Apis mellifera). Behavioral Ecology and Sociobiology, 58, 200–207.
222
R.E. Lubow
Linden, D. R., Savage, L. M., & Overmier, B. J. (1997). General learned irrelevance: a Pavlovian analog to learned helplessness. Learning and Motivation, 28, 230–247. Lipp, O. V., Siddle, D. A. T., & Vaitl, D. (1992). Latent inhibition in humans: single cue conditioning revisited. Journal of Experimental Psychology: Animal Behavior Processes, 18, 115–125. Lipp, O. V., & Vaitl, D. (1992). Latent inhibition in human Pavlovian differential conditioning: effect of additional stimulation after preexposure and relation to schizotypal traits. Personality and Individual Differences, 13, 1003–1012. Loy, I., Fernandez, V., & Acebes, F. (2006). Conditioning of tentacle lowering in the snail (Helix aspersa): acquisition, latent inhibition, overshadowing, second-order conditioning, and sensory preconditioning. Learning & Behavior, 34, 305–314. Lubow, R. E. (1965). Latent inhibition: The effects of frequency non-reinforced preexposure to the CS. Journal of Comparative and Physiological Psychology, 60, 454–457. Lubow, R. E. (1973). Latent inhibition. Psychological Bulletin, 79, 398–407. Lubow, R. E. (1989). Latent Inhibition and Conditioned Attention Theory. New York: Cambridge University Press. Lubow, R. E. (1997). Latent inhibition as a measure of learned inattention: some problems and solutions. Behavioural and Brain Research, 88, 75–83. Lubow, R. E. (2005). The construct validity of the animal-latent inhibition model of selective attention deficits in schizophrenia. Schizophrenia Bulletin, 31, 139–153. Lubow, R. E., Alek, M., & Rifkin, B. (1976). The context effect: the relationship between stimulus preexposure and environmental preexposure determines subsequent learning. Journal of Experimental Psychology: Animal Behavior Processes, 2, 38–47. Lubow, R. E., Caspy, T., & Schnur, P. (1982). Latent inhibition and learned helplessness in children: similarities and differences. Journal of Experimental Child Psychology, 34, 231–256. Lubow, R. E., & Gewirtz, J. C. (1995). Latent inhibition in humans: data, theory, and implications for schizophrenia. Psychological Bulletin, 117, 87–103. Lubow, R. E., Ingberg-Sachs, Y., Zalstein-Orda, N., & Gewirtz, J. C. (1992). Latent inhibition in low and high “psychotic-prone” normal subjects. Personality and Individual Differences, 13, 563–572. Lubow, R. E., & Moore, A. U. (1959). Latent inhibition: the effect of nonreinforced preexposure to the to-be-conditioned stimulus. Journal of Comparative and Physiological Psychology, 52, 416–419. Mackintosh, N. J. (1973). Stimulus selection: Learning to ignore stimuli that predict no change in reinforcement. In R. A. Hinde & J. S. Hinde (Eds.), Constraints on Learning. London: Academic Press, pp. 75–96. McDaniel, J. W., & White, R. K. (1966). A factorial study of the stimulus conditions of habituation. Perceptual and Motor Skills, 23, 259–270. Menzel, R., & Muller, U. (1996). Learning and memory in honeybees: from behavior to neural substrates. Annual Review of Neuroscience, 19, 379–404. Mineka, S., & Cook, M. (1986). Immunization against the observational conditioning of snake fear in rhesus monkeys. Journal of Abnormal Psychology, 95, 307–318.
The phylogenetic distribution of latent inhibition
223
Morin, P.-P., Dodson, J. J., & Dore, F. Y. (1990). Facilitatory effects of preexposure on heart-rate conditioning to an olfactory cue in Atlantic salmon (Salmo sola). Journal of Comparative Psychology, 104, 340–344. O’Keefe, J., & Nadel, L. (1978). The Hippocampus as a Cognitive Map. Oxford: Clarendon Press. Oswald, C. J. P., Yee, B. K., Rawlins, J. N. P., et al. (2002). The influence of selective lesions to components of the hippocampal system on the orienting response, habituation and latent inhibition. European Journal of Neuroscience, 15, 1983–1990. Pedreira, M. E., Dimant, B., Tomsic, D., Queseda-Allue, L. A., & Maldonado, H. (1995). Cycloheximide inhibits context memory and long-term habituation in the crab Chasmagnathus. Pharmacology, Biochemistry & Behavior, 52, 385–395. Pinen˜o, O., De la Casa, G., Lubow, R. E., & Miller, R. R. (2006). Some determinants of latent inhibition in human predictive learning. Learning and Motivation, 37, 42–65. Rankin, C. H. (2000). Context conditioning in habituation in the nematode Caenorhabditis elegans. Behavioral Neuroscience, 114, 496–505. Reilly, S. (1987). Hyperstriatal lesions and attention in the pigeon. Behavioral Neuroscience, 101, 74–86. Robinson, G. B., Port, R. L., & Stillwell, E. J. (1993). Latent inhibition of the classically conditioned nictitating membrane response is unaffected by the NMDA antagonist MK801. Psychobiology 21, 120–124. Rodriguez, F., Lopez, J. C., Vargas, J. P., et al. (2002). Spatial memory and hippocampal pallium through vertebrate evolution: insights from reptiles and teleost fish. Brain Research Bulletin, 57, 499–503. Romano, A. G. (1999). Variations in CS associability and mutiple unit hippocampal activity in the rabbit. Behavioural Brain Research, 103, 39–46. Sahley, C. L., Boulis, N. M., & Schurman, B. (1994). Associative learning modifies the shortening reflex in the semi-intact leech Hirudo medicinalis: effects of pairing, predictability, and CS preexposure. Behavioral Neuroscience, 108, 340–346. Sahley, C. L., & Ready, D. F. (1988). Associative learning modifies two behaviors in the leech, Hirudo medicinalis. Journal of Neuroscience, 3, 4612–4620. Shishimi, A. (1985). Latent inhibition experiments with goldfish. Journal of Comparative Psychology, 99, 316–327. Shohamy, D., Allen, M. T., & Gluck, M. A. (2000). Dissociating entorhinal and hippocampal involvement in latent inhibition. Behavioral Neuroscience, 114, 867–874. Sokolov, E. N. (1963). Perception and the Conditioned Reflex. New York: Pergamon Press. Talk, A., Stoll, E., & Gabriel, M. (2005). Cingulate cortical coding of contextdependent latent inhibition. Behavioral Neuroscience, 119, 1524–1532. Thiele, T. E., & Frieman, J. (1994). Taste-potentiated color aversions in pigeons: examination of CS preexposure and subsequent effects on postconditioning extinction of the taste aversion. Animal Learning & Behavior, 22, 436–441. Tomie, A., Murphy, A. L., Fath, S., & Jackson, R. L. (1980). Retardation of autoshaping following pretraining with unpredictable food: effects of changing the context between pretraining and testing. Learning and Motivation, 11, 117–134.
224
R.E. Lubow
Tomsic, D., Pedreira, M. E., Romano, A., Hermitte, G., & Maldonado, H. (1998). Context-US associations as a determinant of long-term habituation in the crab Chasmagnathus. Animal Learning & Behavior, 26, 196–209. Tranberg, D. K., & Rilling, M. (1978). Latent inhibition in the autoshaping paradigm. Bulletin of the Psychonomic Society, 11, 273–276. Wagner, A. R. (1978). Expectancies in the priming of STM. In S. H. Hulse, H. Fowler, & W. K. Honig (Eds.), Cognitive Processes in Animal Behavior. Hillsdale, NJ: Lawrence Erlbaum, pp. 53–82. Wasserman, E. A., & Molina, E. J. (1975). Explicitly unpaired key light and food presentations: interference with subsequent auto-shaped key pecking in pigeons. Journal of Experimental Psychology: Animal Behavior Processes, 1, 30–38. Weiner, I. (2003). The “two-headed” latent inhibition model of schizophrenia: modeling positive and negative symptoms and their treatment. Psychopharmacology, 169, 257–297. Wickens, C., Tuber, D. S., & Wickens, D. D. (1983). Memory for the conditioned response: the proactive effect of preexposure to potential conditioning stimuli and context change. Journal of Experimental Psychology: General, 112, 47–70. Zalstein-Orda, N., & Lubow, R. E. (1995). Context control of negative transfer induced by preexposure to irrelevant stimuli: latent inhibition in humans. Learning and Motivation, 26, 11–28.
11 The genetics of latent inhibition: studies of inbred and mutant mice Tatiana Lipina and John Roder
Latent inhibition (LI) refers to the phenomenon of reduced conditioning seen after nonreinforced stimulus preexposure. Animal studies of LI typically employ a between-subject procedure to demonstrate LI. In such procedures, one group of subjects is preexposed (PE) to the to-be-conditioned stimulus (CS), whereas another group of subjects is not preexposed (NPE). The CS is subsequently paired with an unconditioned stimulus (US). LI is measured by the difference in some index of learning (e.g., conditioned fear) of the CS–US association between the PE and NPE groups, and is manifested as poorer learning in the PE group. As detailed in several chapters in the present volume by Weiner, Lubow, Kumari and Ettinger, Escobar and Miller and others, a number of theories have been put forward to explain LI as a form of learned inattention in which subjects learn to ignore or reduce attention to irrelevant stimuli (Lubow, 1989). LI has also been characterized as a form of proactive interference in which the inattentional response (acquired as a consequence of a “CS–no-consequence” association) interferes with, or competes for, the expression or retrieval of the conditioned response (resulting from an effective CS–US association) (Miller & Escobar, 2009; Weiner, 1990, 2003, 2009). Deficits in attention and information processing are central features in schizophrenia, and these deficits may lead to stimulus overload, cognitive fragmentation and thought disorder common in this disorder (Freedman et al., 1991; Perry et al., 1999; Strauss et al., 1993). As elaborated by Weiner, Lubow, and Kumari in this volume, abnormal LI is a well-established model of attentional deficits in schizophrenia. In the recent decade, LI has been used increasingly to assess schizophrenia-like attentional deficits in genetically modified mice (Bruno et al., 2007; Clapcote et al., 2007; Gerdjikov et al., 2008; Lipina et al., 2007; Miyakawa et al., 2003; Rimer et al., 2005; Wang et al., 2004; Yee et al., 2006). Genetic studies have indicated a significant genetic component to the disorder, with a complex pattern of inheritance that is either polygenic or oligogenic, with a small influence of a single gene on the susceptibility to develop the disorder. Latent Inhibition: Cognition, Neuroscience and Applications to Schizophrenia, ed. R. E. Lubow and Ina Weiner. Published by Cambridge University Press # Cambridge University Press 2010.
225
226
Tatiana Lipina and John Roder
Although numerous candidate genes have been identified (Allen et al., 2008; Harrison & Weinberger, 2005; Waddington et al., 2007), only several of these genes have been replicated in independent samples, including neuregulin 1 (NRG1) on 8p21-p22, regulator of G-protein signaling 4 (RGS4) on 1q21-q22, dysbindin (DTNBP1) on 6p22.3, catechol-O-methyltransferase (COMT) on 22q11, proline dehydrogenase (PRODH) on 22q11, Disrupted-In-Schizophrenia 1 (DISC1) on 1q42 (Craddock et al., 2005; Harrison & Weinberger, 2005) and phosphodiesterase-4B (PDE4B) on 1p31 (Numata et al., 2008; Pickard et al., 2007; Tomppo et al., 2009). The decoding of the mouse genome, coupled with modern molecular-genetic technologies enabling examination of genetic influences on behavior, provides unique opportunities to advance psychiatric research. The mouse is an ideal model organism for human disease, with approximately 99% of mouse genes having human counterparts (Tecott & Wehner, 2001). Construction of mice with targeted mutations in the genes associated with schizophrenia via gene knockout, gene knock-in/transgenic systems, or induction of point mutants has demonstrated the functional significance of the targeted gene and its encoded protein in mouse behavioral facets related to schizophrenia, including LI (Desbonnet et al., 2009; Tecott & Wehner, 2001). Therefore, studying LI in genetic mouse models will lead to decoding the molecular basis of LI and will allow us to build a molecular-genetic map of schizophrenia-related psychopathologies. There are two major approaches to studying genetic contributions to behavior: quantitative and single-gene. The most commonly practiced form of quantitative genetics is Quantitative Trait Locus (QTL) analysis (Cookson et al., 2009). QTL is a chromosomal region that contributes to the range of variation of a continuous behavioral trait, whereas single-gene mutant analysis creates its own raw material by inducing genetic mutations affecting neuronal functions and behavior. The two approaches have the same ultimate goal – an understanding of the genetic underpinning of behavior. Several QTLs have been identified for fear conditioning in mice (Owen et al., 1997; Ponder et al., 2007, 2008; Radcliffe et al., 2000; Wehner et al., 1997). Although most LI procedures in the mouse are based on fear conditioning (Bruno et al., 2007; Clapcote et al., 2007; Gerdjikov et al., 2008; Gould & Wehner 1999; Lipina et al., 2007; Miyakawa et al., 2003; Rimer et al., 2005), in order to apply the QTL approach to LI a within-subject design is needed in which conditioning to a preexposed and non-preexposed stimulus is compared within each animal, thereby providing one measure of LI per animal. Thus far genetic studies using LI have used between-subjects designs and consequently they have focused exclusively on the single-gene mutation approach. LI analysis of mice with targeted mutation of susceptibility genes provides a powerful means of identifying the functional roles of genes and how they might contribute to schizophrenia. However, genetic manipulations raise problems such as compensatory up- or down-regulation of genes that can mask the effect of the genetic manipulation or unknown interactions between the mutated gene(s) and a particular
The genetics of latent inhibition: studies of inbred and mutant mice
227
genetic background. Mutant mice are frequently generated using inbred strains as backgrounds. Inbred strains are created by consecutive matings between siblings of many generations, resulting in genetically identical mice within a strain. The genetics of inbred strains are different from one another, and inbred mice may vary in many parameters. Consequently, the use of inbred strains provides a useful approach to elucidate genetically dependent variability in behavior.
LI in inbred strains The one study that assessed LI in several inbred mouse strains showed that genetic background, indeed, modulated the expression of this phenomenon (Gould & Wehner, 1999). Recently, we have used inbred strains to investigate genetic influences on LI and its response to pharmacological manipulations. We phenotyped six mouse strains (C57BL/6J, 129/SvEvTac, A/J, CBA/J, C3H/He, and C3.BliA-þPde6b) for LI in our laboratory, using an experimental protocol that yields LI in the “gold standard” C57BL/6J strain (Gould & Wehner, 1999; Lipina et al., 2005), to identify strains differing in LI expression (Figure 11.1). LI, expressed as weaker fear conditioning of the preexposed compared to the non-preexposed mice, was present in the C57BL/6J and 129/SvEvTac inbred strains, but there was no difference in the level of fear conditioning between the preexposed and the non-preexposed groups (no LI) in the other four strains. These results are consistent with strain differences reported by Gould and Wehner (1999), and provide further evidence that genetic background contributes to the development or the expression of this phenomenon. Furthermore, as found by Gould and Wehner (1999), strain-dependent loss of LI exhibits two distinct patterns. In the C3H/He and C3.BliA-þPde6b strains, loss of LI was due to strong conditioning in the preexposed groups, which learned as well as 0.6
PE NPE
Suppression ratio
0.5 0.4 0.3 0.2
***
0.1 ** 0 C57BL/6J 129/SvEvTac
A/J
CBA/J
C3H/He C3.BliA–+Pde6b
Figure 11.1. Mean suppression ratios of the preexposed (PE) and non-preexposed (NPE) C57BL/6J, 129/SvEvTac, A/J, CBA/J, C3H/He or C3.BliA-þPde6b mice conditioned with two conditioning trials following 40 preexposures (n per group ¼ 7– 9). **P < 0.01; *** P < 0.001 in comparison with PE within each inbred strain.
228
Tatiana Lipina and John Roder
non-preexposed groups. This is the “classical” attentional deficit captured by LI loss in which the preexposed animals lose the capacity to ignore the irrelevant stimulus. The latter is typically observed following amphetamine administration (Killcross & Robbins, 1993; Russig et al., 2002; Weiner et al., 1988) and gene targeting (Bruno et al., 2007; Clapcote et al., 2007; Gerdjikov et al., 2008; Lipina et al., 2007; Miyakawa et al., 2003; Rimer et al., 2005). Conversely, in CBA/J and A/J mice, the absence of LI was due to an associative deficit in the non-preexposed group. Numerous drug studies in mice and humans have demonstrated that genetic background produces differences in drug efficacy. Several studies reported associations between gene polymorphisms (e.g., in 5-HT2A, HTT, DRD2, RGS4) and responses to antipsychotic drugs (APDs) in humans (Campbell et al., 2008; Penas-Lledo et al., 2008; Wang et al., 2007; Xing et al., 2007). Genetic background is therefore a crucial variable when evaluating drug action in general and novel drugs in particular. To date, nothing is known about how genetic background contributes to schizophrenia-relevant pharmacology of mouse LI. It is well documented that the capacity to effectively ignore irrelevant stimuli is strengthened by APDs in rodents as well as humans (Lipina et al., 2005; Moran et al., 1996; Moser et al., 2000; Weiner et al., 1997, 2003; Williams et al., 1997), and that this LI potentiation or facilitation effect is an index of antipsychotic activity (see Weiner, this volume). We have recently demonstrated in the C57BL/6J strain that NMDA receptor (NMDAR) function enhancers such as D-serine and ALX5407 also facilitate LI (Lipina et al., 2005). Genetic variability in LI expression may be of particular importance for screening APDs and NMDAR enhancers because inbred mouse strains with weak LI expression may be suitable for detection of LI potentiation by such compounds. In our study, we focused on two strains, C3H/He and CBA/J, that failed to show LI due to attentional or associative deficit, respectively, to test the efficacy of the typical APD haloperidol, the atypical APD clozapine, and the NMDA enhancer D-serine (unpublished data). Facilitation of LI was observed in APD and D-serine-treated mice of both strains, C3H/He and CBA/Lac, but the pattern of drug action differed among the strains depending on the nature of the deficit underlying LI loss. In the C3H/He strain, the drugs led to LI potentiation exclusively by improving animals’ capacity to ignore the irrelevant stimulus, whereas in the CBA/Lac strain LI potentiation was due to enhanced conditioning. To the best of our knowledge, this is the first clear demonstration that APDs, including haloperidol, facilitate conditioning.
LI in mutant mice Genetically modified mice have provided abundant insights into the specific signaling molecules involved in psychiatric disorders. Mutant lines of mice are available from different resources: RIKEN (http://www.brc.riken.jp/lab/gsc/mouse/), Mutant Mouse Regional Resource Centers (http://www.mmrrc.org), European Mouse Mutant Archive (http://www.emma.rm.cnr.it), Medical Research Council (MRC)
The genetics of latent inhibition: studies of inbred and mutant mice
229
Harwell (http://www.har.mrc.ac.uk/), Jackson Laboratories (http://www.jax.org/), and Taconic Farms (http://www.taconic.com). There are several kinds of mutant mice – mice carrying deletions of single genes, deletions of whole chromosomal regions, genetic insertions and spontaneous point mutations. Since LI has only been adapted to the mouse in the past decade, there are still relatively few studies in genetically manipulated mice but there is a trend of increased usage of mouse in LI in the past few years (see Table 11.1). As detailed by Weiner in this volume, drugs, lesions and neurodevelopmental manipulations can produce two poles of abnormality in LI in comparison to the status of LI in controls: disrupted LI, manifested under conditions in which controls express LI, and persistent, or augmented LI, manifested under conditions that prevent the expression of LI in controls. The former reflects the attentional over-switching that accompanies the positive symptoms of schizophrenia, whereas the latter is associated with cognitive inflexibility and the negative symptoms of schizophrenia (Weiner, 2003; Weiner, this volume). Genetic manipulations in mice also produce these two poles of abnormality. To date, disrupted LI has been reported in 11 mutant lines, whereas augmentation of LI in gene-modified mice was found in six mutant lines (Table 11.1).
Disrupted LI Disrupted LI was observed in mice lines selected for non-response to haloperidol (neuroleptic non-responsive mice – NNR) (Kline et al., 1998); conditional calcineurin knockout mice (Miyakawa et al., 2003); heterozygous mice for immunoglobulin domain-specific mutation of NRG-1 (Ig-NRG1) (Rimer et al., 2005); Bax knockout mice (Perez et al., 2007); in three different Disc1 mutants (Clapcote et al., 2007; Shen et al., 2008), mGluR5 knockouts (Lipina et al., 2007); coloboma mice (Bruno et al., 2007) and a5-(H105R) mutant mice (Gerdjikov et al., 2008).
Neuroleptic non-responsive lines (NNR) The neuroleptic responsive (NR) and neuroleptic non-responsive (NNR) lines of mice have been bidirectionally selected for response to haloperidol-induced catalepsy (Hitzemann et al., 1994). The formation of the NR and NNR lines resulted in the fixation of those alleles associated with response and non-response to haloperidolinduced catalepsy. LI was estimated in NR, NNR and control mice, which were from the 17th generation of selective breeding. NNR mice showed a lack of LI in the active avoidance paradigm. Preexposed to the non-reinforced tone, NNR mice did not show decrements of conditioned avoidance response acquisition. However, the authors mentioned that in addition to disrupted LI, NNR mice performed poorly on the active avoidance task, which makes it difficult to interpret the data.
Disc1-L100P; C57BL/6J
Forebrain-specific calcineurin KO; C57BL/6J Disc1-Q31L; C57BL/6J
Bax KO; C57BL/6J
IL-6 KO; C57BL/6J
LI disruption Neuroleptic nonresponsive mice; Balb/c LP Heterozygous NRG-1 KO (Ig-domain); C57BL/6J
Gene targeted mice; genetic background
Associations between PPP3CC (gene encodes calcineurin Ag subunit) and SCZ Familial, linkage and association studies correlate with SCZ
Altered expression of NRG1 isoforms in prefrontal cortex of SCZ; variants of NRG1 and ErbB4 increase risk for SCZ Association of IL-6 polymorphism with SCZ high Bax/Bcl-2 ratio in the temporal cortex of SCZ
Linkage with schizophrenia
CER
Clozapine reversed LI deficit
Clapcote et al., 2007; Millar et al., 2000; Devon et al., 2001a,b; Canon et al., 2005; Henah et al., 2005 Clapcote et al., 2007
Clozapine had no effect
CER
Smith et al., 2007; Sun et al., 2003 Perez et al., 2007; Jarskog et al., 2004
Rimer et al., 2005; Corvin et al., 2004; Stefansson et al., 2002, Hashimoto et al., 2004; Norton et al., 2006
Kline et al., 1998
References
Miyakawa et al., 2003; Gerber et al., 2003
Conditioned behavior is impaired in NPE group Conditioning is impaired in NPE group at 3–4 months old but LI is present in the old mice
Sensitive to clozapine
Comments
Freezing
CER
CER
Activity
CAR
LI procedure
Table 11.1. Latent inhibition abnormalities in gene-modified mice
CER
Freezing
Association of M5 and a7-Nicotin with SCZ
Association with APDs treatment-resistant SCZ
M5 muscarinic receptors mutants; CD1 129 genetic background 5-HT3 receptors-Tg overexpression in the forebrain
CER
No association with SCZ
D1-KO; C57BL/6
CER
Taste aversion
CAR
Effect was observed in both male and female rats Effect was observed only in females
DSP-4 reversed LI deficit
LI absent in male
Clozapine and CX546 reversed LI deficit
CER
No consistent findings
Increased neuronal mGluR5 mRNA in cortical pyramidal cell layers in SCZ; mGluR5 has been mapped to 11q14, closely to a translocation that segregates with SCZ Reduced level of GAD67; GAT-1 is reduced; a2 subunit of GABAa is increased in SCZ Altered SNAP-25 transcript or protein levels in SCZ
Conditioning is impaired in NPE group
Activity
LI persistence D2-KO; C57BL/6
Coloboma mice (SNAP-25 deficit); C3H/HeSnJ
a5(H105R) mutant mice
Disc1Tr-Tg; 50% CBA/CaCrl 50% C57BL/6J mGluR5 KO; C57BL/6J
Harrel and Allan 2003; Ji et al., 2008
Bay-Richter et al., 2008; Wong et al., 2000 Bay-Richter et al., 2008; Wong et al., 2000 Wang et al., 2004; De Luca et al., 2004
Bruno et al., 2007; Carroll et al., 2009; Knable et al., 2004
Gerdjikov et al., 2008; Volk & Lewis, 2002; Woo et al., 1998; Petryshen et al., 2005
Lipina et al., 2007; Ohnuma et al., 1998; Devon et al., 2001
Shen et al., 2008
Linkage with genes that modulate the NMDAR glycine binding site: d-amino acid oxidase and G72 (D-serine catabolism); serine racemase (D-serine synthesis)
Linkage with schizophrenia
CER
Freezing; CAR; CER
LI procedure
Clozapine, ALX-5407 and D-serine restored LI
Augmentation of LI due to improved conditioning
Comments
Labrie et al., 2008; Duffy et al., 2008; Chumakov et al., 2002; Goltsov et al., 2006; Morita et al., 2006; Schumacher et al., 2004
Yee et al., 2006
References
Note: CER, conditioned emotional response (thirst-motivated); CAR, conditioned avoidance response; SCZ, schizophrenia.
GrinD481N mutant mice C57BL/6J
CamKII_Cre: Glyt1tm1.2fl/fl (GlyT1); C57BL/6J
Gene targeted mice; genetic background
Table 11.1. (cont.)
The genetics of latent inhibition: studies of inbred and mutant mice
233
Calcineurin Calcineurin (also termed PP2B) is a calcium-dependent protein phosphatase that plays an important role in CNS functions, including the regulation of neuronal structure, neurotransmission, activity-dependent gene expression, synaptic plasticity, and learning and memory (Groth et al., 2003; Mansuy, 2003). Calcineurin is a heterodimer consisting of a catalytic A subunit and a regulatory B subunit (Klee et al., 1988). In the brain, only one regulatory subunit, CNB1, is expressed while two isoforms of the catalytic subunit (CNAa and CNAb) are expressed with distinct yet overlapping patterns (Kuno et al., 1992; Takaishi et al., 1991). Significant associations between genetic variation in the 8p21.3 gene, ppp3cc, which encodes a calcineurin Ag catalytic subunit, and schizophrenia have been reported US and South African samples (Gerber et al., 2003), supporting the role of calcineurin in this disease. Behavioral analysis of the forebrain-specific CNB knockout mice (Miyakawa et al., 2003; Zheng et al., 2001) showed phenotypes that resembled those of schizophrenic patients: hyperactivity, working memory impairments, prepulse inhibition deficit, social withdrawal, and disrupted LI as measured by conditioned freezing.
NRG-1 NRG-1 is a neurotrophic factor that contains an epidermal growth factor (EGF)-like domain. NRG-1 binds to ErbB2, 3 and 4, the receptor tyrosine kinases that mediate its biological activities (Buonanno & Fischbach, 2001; Falls, 2003). Numerous roles for NRG-1 in central nervous system development and function have been identified, including synapse formation, neuronal migration, synaptic plasticity and the regulation of neurotransmitter expression and function (Falls, 2003). Polymorphisms in the NRG-1 and ErbB4 receptor genes have been associated with schizophrenia in numerous population and family studies (Stefansson et al., 2002; reviewed in Mei and Xiong, 2008), and biochemical measurements from postmortem prefrontal cortex homogenates suggest that NRG/ErbB signaling is altered in schizophrenia (Law et al., 2007; Silberger et al., 2006). Moreover, recent work indicates that NRG/ ErbB signaling has a role in regulating glutamatergic transmission (Hahn et al., 2006), supporting its relevance to schizophrenia given the postulated role of glutamatergic hypofunction in the pathogenesis underlying schizophrenia. Homozygous mice with deletions in either the EGF-like, immunoglobulin-like (Ig-like), transmembrane (TM) domains or cysteine-rich domain (CRD) die because of their inability to form functional Schwann cells (Wolpowitz et al., 2000; Stefansson et al., 2002). Heterozygous animals with these engineered mutations are viable, and display several schizophrenia-like behaviors. For instance, EGF-NRG-1þ/_ mice showed hyperactivity (Gerlai et al., 2000), and TM-NRG-1þ/_ mice displayed hyperactivity, prepulse inhibition deficit, impaired exploration of and habituation to novel environments (Stefansson et al., 2002). Ig-NRG-1 mice were sensitive to clozapine
234
Tatiana Lipina and John Roder
treatment, similar to TM-NRG-1þ/_ mice (Stefansson et al., 2002), and they displayed behaviors that are thought to indicate a schizophrenia-like phenotype, such as clozapine-induced suppression of open-field and running-wheel activity (Rimer et al., 2005). So far, only Ig-NRG-1 heterozygous mice have been tested in LI, and they showed impaired LI (Rimer et al., 2005).
Bax The control of the production and elimination of adult-generated neurons is thought to be critical for the maintenance of a relatively constant number of neurons in the nervous system. Target-dependent programmed cell death of neurons is mediated by Bax protein, a proapoptotic member of the Bcl-2 family (Deckwerth et al., 1996; Sun & Oppenheim, 2003; White et al., 1998). Genetic studies have not reported any associations or linkages between Bax and schizophrenia. However, a high Bax/Bcl-2 ratio was found in the temporal cortex of schizophrenics (Jarskog et al., 2004). Moreover, the proportion of sub-G0 cells (proportion of cells in the sub-G0 cell cycle fraction in which apoptotic bodies accumulate) was significantly larger in the schizophrenia group (Catts et al., 2006), supporting the increased susceptibility to apoptosis in schizophrenic patients. There is an absence of Bax proapoptotic protein disrupted LI in Bax knockout mice at the age of 3–4 months (unpublished data). However, lack of Bax protein resulted in a general associative deficit in Bax mice as measured in fear conditioning (Perez et al., 2007). Older Bax knockout mice (12–18 months) were able to overcome LI deficits and other behavioral impairments observed earlier (Perez et al., 2007). Hence, Bax deficiency elicits maturation-dependent effects: negative influences were found at the earlier stage of development and positive effects were observed in older animals. Excessive numbers of neurons in Bax knockout mice could result in inappropriate integration into existing neuronal networks at the young stage, disrupting normal neuronal functioning which leads to behavioral impairments including the LI deficit. Additional numbers of neurons would be beneficial in the aged animals. Taken together, these findings suggest a nonspecific role of apoptosis in modulation of LI expression.
Disc1 Disrupted-In-Schizophrenia 1 (Disc1) is a coiled-coil protein involved in neuronal proliferation, differentiation, and migration, cyclic adenosine monophosphate (cAMP) signaling, cytoskeletal modulation, and translational regulation (Camargo et al., 2007; Porteous et al., 2006; Ross et al., 2006). Disc1 was originally identified at one breakpoint of a chromosomal t(1;11) (q42.1;q14.3) translocation that cosegregates in a large Scottish family with major mental illnesses, including schizophrenia,
The genetics of latent inhibition: studies of inbred and mutant mice
235
bipolar disorder, and major depression (Millar et al., 2000). Multiple independent linkage and association studies have reported evidence in support of Disc1 as a generalized risk factor in schizophrenia and mood disorders (Hennah et al., 2005; Porteous et al., 2006). Several groups have attempted to mimic Disc1 haploinsufficiency by knocking out the mouse Disc1 gene, or by generating transgenic lines overexpressing human Disc1 transgenes of various designs (Chubb et al., 2007; Shen et al., 2008). Disrupted LI has been reported in ENU-induced Disc1 mutants (Disc1-L100P and Disc1-Q31L lines; Clapcote et al., 2007) and in Disc1-Tg mice with overexpression of truncated Disc1 protein (Shen et al., 2008), but LI has not been tested in transgenic mice with ectopic promoters (Hikida et al., 2007; Pletnikov et al., 2008), nor in mice that expressed the mutated 129 Disc1 allele (Koike et al., 2006). Homozygous Disc1-L100P and Disc1-Q31L mutants showed severe LI impairments due to disrupted ability to ignore the irrelevant stimulus, whereas Disc1-Tg mice failed to condition. Clozapine effectively reversed LI deficit in Disc1-L100P but not in Disc1-Q31L mutants (Clapcote et al., 2007), suggesting that distinct neuropathological mechanisms underlie the LI deficit observed in these Disc1 mutants. Moreover, although both homozygous Disc1-Q31L and Disc1L100P mutants had LI deficit, a different pattern was seen in heterozygous mice. Thus, Disc1-Q31L heterozygous mice were able to develop LI, whereas Disc1L100P heterozygous animals showed disrupted LI due to poorer conditioning of non-preexposed groups (Clapcote et al., 2007). Hence, Q31L mutation elicits a recessive influence on LI expression, whereas in Disc1-L100P heterozygous and homozygous mice different psychopathological mechanisms may underlie LI impairments.
mGluR5 Metabotropic glutamate receptors (mGluR) add to the complexity of glutamatergic neurotransmission in the brain. The mGluR5 subtype of mGluRs is of special interest due to its functional link with the NMDA receptor (Awad et al., 2000; Doherty et al., 1997) which plays a critical role in schizophrenia (Goff & Coyle, 2001). In addition, there is evidence indicating a role for mGluR5 itself in the psychopathology of schizophrenia. Increased neuronal mGluR5 mRNA has been found in cortical pyramidal cell layers of schizophrenics (Ohnuma et al., 1998). Furthermore, linkage analysis revealed that mGluR5 has been mapped to 11q14, close to a translocation that segregates with schizophrenia and related psychosis in a large Scottish family (Devon et al., 2001b). Absence of mGluR5 allele in the mouse abolished LI, and LI has been restored in these mice by the AMPAR modulator CX546 (Lipina et al., 2007). Preexposed heterozygous mGluR5 mice showed intermediate LI phenotype compared to wild-type and homozygous animals (unpublished data), indicating the dominant role of mGluR5 protein in the regulation of LI.
236
Tatiana Lipina and John Roder
SNAP-25 SNAP-25 is a presynaptic protein involved in calcium-triggered neurotransmitter release (Hess et al., 1992). Identification of the SNAP-25 gene defect in coloboma mice, which exhibit a 2 cM hemizygous deletion encompassing Snap25, provided a candidate gene for studies in humans that revealed an association between SNAP-25 and attention deficit hyperactivity disorder (ADHD) (Barr et al., 2000; Brophy et al., 2002; Kustanovich et al., 2003; Mill et al., 2002). SNAP-25 also has been implicated in the pathogenesis of schizophrenia by numerous neuropathological studies (e.g. Knable et al., 2004; Thompson et al., 1998). A recent intriguing hypothesis has suggested that genetic variation at SNAP-25 may be differentially associated with schizophrenia and ADHD (Carroll et al., 2009). Consistent with the ADHD phenotype, coloboma mice show robust hyperactivity, decreased locomotor activity in response to amphetamine (Hess et al., 1996), and increased impulsivity (Bruno et al., 2007). Coloboma mice were also assessed in LI (Bruno et al., 2007) and shown to have disrupted LI, as found in ADHD (Lubow et al., 2005). This deficit was reversed by DSP-4, a noradrenergic agent which effectively decreased norepinephrine concentrations and hyperactivity in coloboma mice (Jones & Hess, 2003). It would be interesting to test whether disrupted LI in coloboma mutants could be reversed by APDs or, conversely, whether DSP-4 would exhibit antipsychotic action in an LI model. If not, this could suggest different mechanisms underlying attentional deficits in ADHD and schizophrenia.
a5-GABAA Cognitive symptoms of schizophrenia are relatively stable over time, and considered to be the core of the disease. Activity of gamma-aminobutyric acid (GABA) neurons in the dorsolateral prefrontal cortex is essential for normal working memory function (Rao et al., 2000), and it has been suggested that altered GABA neurotransmission in this brain area could contribute to the cognitive impairments in schizophrenia. Genetic manipulations of GABAergic control produce numerous behavioral alterations in learning and memory. Comparative analysis of behavioral phenotypes of mice carrying genetic alterations in a1, a2, a3, and a5 GABAA subunits has pointed to a5-GABAA subunit involvement in hippocampally mediated learning and memory (Mo¨hler et al., 2008). The behavioral phenotypes of knock-in mice bearing a point mutation (H105R) in the GABAA receptor a5 subunit gene (a5(H105R) mutant mice) include mild elevation in spontaneous open-field activity, impaired sensorimotor gating (Hauser et al., 2005), improved trace conditioning (Crestani et al., 2002; Yee et al., 2004), and increased resistance to extinction (Yee et al., 2004). a5(H105R) mutation attenuates the expression of LI without affecting conditioning (Gerdjikov et al., 2008), supporting previous reports showing that classical conditioning remains largely unaffected by a5(H105R) mutation (Crestani et al., 2002). Lack of LI was observed in male a5(H105R) but not female mice.
The genetics of latent inhibition: studies of inbred and mutant mice
237
Augmentation of LI Augmentation of LI was observed in D2 knockout mice, D1 knockout female mice (Bay-Richter et al., 2008), M5 muscarinic receptor mutants (Wang et al., 2004), 5-HT3 receptors-Tg mice (Harrel & Allan, 2003), glycine transporter 1 mutants with reduced expression of GlyT1 in the forebrain neurons (Yee et al., 2006), and Grin D481N/K483Q mutant mice (Labrie et al., 2008). D1 and D2 receptors The notion that hyperactivity of the DA system, or increased sensitivity to dopamine, plays a central role in schizophrenia derives from the observations that drugs used to treat schizophrenia act through DA receptor blockade, and that DA agonists, such as amphetamine, provoke psychosis in normal humans and exacerbate psychosis in schizophrenics (Carlsson et al., 2001; Guilin et al., 2007; Kapur and Mamo, 2003). There is extensive support from studies in rats that the DA system plays a key role in the expression of LI (Weiner, 2003), and recent pharmacological studies in mice have supported this (Chang et al., 2007; Meyer et al., 2005). Given the affinity of antipsychotic drugs for both D1 and D2 receptors, the role of these receptor subtypes in LI has been recently assessed using D2 and D1 knockout mice (Bay-Ritcher et al., 2008). D2-deficient male and female mice showed enhancement of LI under conditions that prevented the expression of LI in wild-type littermates, but only females of D1-KO mice showed enhancement of LI. The profile obtained in males reproduces the established profile of APDs in rat and mouse LI (Lipina et al., 2005; Moran et al., 1996; Moser et al., 2000; Warburton et al., 1994; Weiner et al., 2003), namely, potentiation by D2 blockers and lack of effect by D1 blockade, as well as in clinical trials reporting no benefit of D1 antagonists in schizophrenia patients. However, clinical trials as well as animal studies have been primarily carried out with male subjects (Den Boer et al., 1995). Gender differences revealed in LI in D1 and D2-KOs may be relevant to the understanding of the well-described gender differences in response to APDs in schizophrenia (Goldstein et al., 2002; Lewine et al., 1996; Usall et al., 2007). One possible mechanism is interplay of D1 receptors and their molecular messengers with sex hormones. It is well established that the estrous cycle influences variation in basal extracellular concentration of striatal DA (Becker, 1999; Le Saux et al., 2006), which in turn could lead to different LI expression in males and females of D1-KO mice. Taken together, LI data from D2 and D1 knockout mice support the augmentation effects of antipsychotics on LI and indicate that enhancement of LI may involve differential regulation by D1 and D2 receptor subtypes in males and females.
M5 muscarinic receptors The importance of M5 muscarinic receptors for mesolimbic dopamine functions has suggested a role for these receptors in schizophrenia (Yeomans, 1995). Indeed,
238
Tatiana Lipina and John Roder
mesolimbic input to the nucleus accumbens, which plays an important role in LI expression (Weiner, 2003), originates in the dopamine neurons of the ventral tegmental area (VTA), where M5 muscarinic receptors activate dopaminergic functions (Yeomans et al., 2001). M5 mRNA is colocalized with mRNA for D2 dopamine receptors near VTA and substantia nigra dopamine neurons (Weiner et al., 1990), and both M5 and D2 receptors strongly influence dopamine release from striatal terminals (Zhang et al., 2002). Moreover, the haplotype analysis of the M5 muscarinic receptor and alpha7-nicotinic receptor revealed their transmission in schizophrenia (De Luca et al., 2004), supporting the role of M5 receptors in this psychopathology. M5 muscarinic receptor mutant mice show LI under conditions in which wild-type littermates do not express LI. M5 mutants have lowered dopamine release in the nucleus accumbens together with increased D2 mRNA expression (Wang et al., 2004), which could explain the observed LI augmentation effect. Interestingly, cholinergic hyperactivity has been associated with negative symptoms of schizophrenia (Tandon & Greden, 1989), and the same has been suggested for augmented LI (Weiner, this volume).
5-HT3 receptors The 5-HT3 receptor belongs to the superfamily of ligand-gated ion channels (Ortells & Lunt, 1995). Ligand binding results in channel opening and allows for the influx of Naþ and Ca2þ (Nayak et al., 1999). The 5-HT3 receptor is primary localized presynaptically as a heteroreceptor (van Hooft & Vijverberg, 2000), which allows for the modulation of release of neurotransmitters such as dopamine (Campbell et al., 1996), norepinephrine (Matsumoto et al., 1995), acetylcholine (Consolo et al., 1994) and gamma-aminobutyric acid (McMahon & Kauer, 1997). Ji et al. (2008) have recently reported an association between the 5-HT3 receptor gene and APD-resistant schizophrenia in the Japanese population, a finding that underscores the significance of 5-HT3 in the action of antipsychotics (Olijslagers et al., 2006). Overexpression (OE) of 5-HT3 receptors restricted to the forebrain enhanced LI in 5-HT3 receptors-OE mice (Harrell & Allan, 2003). The latter is consistent with findings that pharmacological blockade of 5-HT3 receptors facilitates LI (Moser et al., 2000).
Glycine transporter-1 The glutamatergic hypothesis proposes that schizophrenia is primarily caused by diminished activity of glutamatergic pathways. Glutamate neurotransmission has been implicated in schizophrenia in large part as a result of the psychotomimetic effects of NMDA receptor antagonists (Carlsson et al., 2001; Coyle, 1996; Javitt & Zukin, 1991; Krystal et al., 1994). Blockade of the NMDAR with noncompetitive antagonists such as phencyclidine produces and exacerbates positive, negative, and
The genetics of latent inhibition: studies of inbred and mutant mice
239
cognitive schizophrenic symptoms (Javitt & Zukin, 1991; Krystal et al., 1994), and facilitation of NMDAR function may be therapeutic in schizophrenia (Javitt et al., 1999; Labrie et al., 2008; Lipina et al., 2005). Altered NMDAR expression and decreased hippocampal NMDAR binding have been found in post-mortem studies (Millan, 2005), and genetic linkage studies have identified a number of susceptible genes that modulate NMDAR function (Chumakov et al., 2002; Ross et al., 2006). Local concentrations of glycine in the forebrain are regulated by the high-affinity glycine transporter type 1 (GlyT1) (Atkinson et al., 2001; Berger et al., 1998; Smith et al., 1992). Therefore, selective inhibition of GlyT1 in the forebrain produces a more potent enhancement of NMDAR activity than that obtained with exogenous glycine/D-serine administration. Selective forebrain neuronal GlyT1 disruption in GlyT1 mutant mice potentiated LI (Yee et al., 2006). Augmentation of LI was due to the enhanced associative learning of the non-preexposed groups, not conforming to the pattern of LI augmentation typically seen after treatment with APDs and glycinergic enhancers, which is mediated via reduced associative learning in the preexposed condition (Lipina et al., 2005; Moser et al., 2000). Yee et al. proposed that cognitive enhancement in GlyT1 mutants could be interpreted as an effective intervention against cognitive impairments in schizophrenia patients.
NR1 glycine binding site NR1 subunits are distributed throughout the CNS, with highest concentrations in the hippocampus, thalamus, frontal cortex, and other structures implicated in schizophrenia. The glycine B recognition site is located in the NR1 subunit (Moretti et al., 2004; Yamakura & Shimoh, 1999). Glycine B agonists and drugs blocking GlyT1 inhibitors increase activity in NMDA receptors, providing a basis for potential antipsychotic properties. Robust decreases in levels of mRNA encoding NR1 subunits have been reported in schizophrenia, together with reduced binding of the glycine B radioligand, [3H]MDL105,519 (Ibrahim et al., 2000; Meador-Woodruff et al., 2003). Decreases in NR1 subunits were found in entorhinal, temporal, and frontal cortices although inconsistent results were reported for PFC (Meador-Woodruff & Healy, 2000; Millan, 2002; Van Berckel, 2003). The GrinD481N mutant line is a mouse genetic model with a five-fold decrease in NMDAR glycine affinity due to the point mutation (aspartate to asparagine substitution at amino acid 481) in their NR1 glycine binding site (Kew et al., 2000). The chronic developmental hypofunction of the NMDAR leads to persistent LI in GrinD481N mutant mice (Labrie et al., 2008). Persistent LI was reversed by D-serine, clozapine, and ALX-5407 (Labrie et al., 2008), supporting the essential role of NMDAR hypofunction in excessively strong LI (Weiner, this volume). The augmentation of LI seen in GrinD481N mutants also partially involved an enhancement in associative learning. Reduced NMDAR activity in hippocampal synapses recruits extra NMDARs to the synapse from extrasynaptic sites (Tovar & Westbrook, 2002), and increases in NMDAR subunit
240
Tatiana Lipina and John Roder
expression were observed in Grin1D481N mutants (Kew et al., 2000). Thus, a larger number of NMDARs may explain enhancement in associative learning in Grin1D481N mutants.
Conclusion The studies of gene-modified mutant mice in LI illustrate successful applications to elucidating the genetic underpinning of selective attention and its dysfunction and, hence, of the neurobiological processes relevant to schizophrenia. Indeed, the existence of mice with impaired function of such gene-candidates for schizophrenia as NRG-1, Disc1, mGluR5, and GlyT1, which generate aberrant LI, supports their role in schizophrenia. Detailed investigations of LI expression in heterozygous/homozygous mutant mice of both genders accompanied by parametric manipulations of the LI procedure could lead to fuller comprehension of the functional role of the particular gene. Moreover, the precise restrictions of mutations to particular brain regions and neuronal cell types will more effectively elucidate their roles in the neuronal circuits involved in LI. To this end, the large number of available inbred strains is a unique resource for studying LI. The systematic characterization of the wide variety of inbred strains in LI can be used to select strains with disrupted and augmented LI to discover new psychopathological mechanisms of schizophrenia. The pharmacogenetic approach to mouse LI can be productively used to predict the response to APDs in humans. For instance, the differential efficacy of clozapine to reverse LI deficits in Disc1-L100P and Disc1-Q31L mutant lines (Clapcote et al., 2007) opens the possibility of finding polymorphism of Disc1 associated with APD response in humans. Indeed, preliminary evidence obtained by functional neuroimaging supports the assumption that Disc1 is implicated in response variability to APD treatments (Blasi & Bertolino, 2006). Recently, a new protein, Neto1, which interacts with NMDAR complex, has been discovered (Ng et al., 2009). Absence of Neto1 leads to an LI deficit in mice, which is effectively reversed by clozapine and CX547 (AMPAR modulator; unpublished data). Parallel findings with mGluR5 and Grin1D481N further support the notion of using NMDAR function enhancers as pharmacotherapy for schizophrenia. The study of potentially useful treatments in the LI model using inbred strains or genetic mouse mutants is a fruitful approach to identifying new treatments and new genes predicting responses to such treatments. To date, there is a paucity of studies that apply pharmacological treatments to genetic mouse models, hence further studies using pharmacogenetic tactics to assess LI in mouse models are needed. Taken together, mutant mice can be valuable tools for testing hypotheses concerning the roles of particular proteins involved in LI regulation. As such, they will provide important insights into the genetic and cellular basis of LI, hopefully leading to a better understanding of the neurobiology of LI and, hence, of the pathology and treatment of schizophrenia.
The genetics of latent inhibition: studies of inbred and mutant mice
241
References Allen, N. C., Bagade, S., McQueen, M. B., et al. (2008). Systematic meta-analyses and field synopsis of genetic association studies in schizophrenia: the Sz Gene database. Nature Genetics, 40, 827–834. Atkinson, B. N., Bell, S. C., De Vivo, M., et al. (2001). ALX 5407: a potent, selective inhibitor of the hGlyT1 glycine transporter. Molecular Pharmacology, 60, 1414–1420. Awad, H., Hubert, G. W., Smith, Y., Levey, A. I., & Conn, P. J. (2000). Activation of metabotropic glutamate receptor 5 has direct excitatory effects and potentiates NMDA receptor currents in neurons of the subthalamic nucleus. Journal of Neuroscience, 20, 7871–7879. Barr, C. L., Feng, Y., Wigg, K., et al. (2000). Identification of DNA variants in the SNAP-25 gene and linkage study of these polymorphisms and attention-deficit hyperactivity disorder. Molecular Psychiatry, 5, 405–409. Baruch, I., Hemsley, D., & Gray, J. A. (1988). Differential performance of acute and chronic schizophrenics in a latent inhibition task. Journal of Nervous and Mental Disease, 176, 598–606. Bay-Richter, C., O’Tuathaigh, C. M., O’Sullivan, G., et al. (2008). Enhanced latent inhibition in dopamine receptor-deficient mice is sex-specific for the D1 but not D2 receptor subtype: implications for antipsychotic drug action. International Journal of Neuropsychopharmacology, 17, 1–12. Becker, J. B. (1999). Gender differences in dopaminergic function in striatum and nucleus accumbens. Behavioural Pharmacology, 18, 583–589. Berger, A. J., Dieudonne, S., & Ascher, P. (1998). Glycine uptake governs glycine site occupancy at NMDA receptors of excitory synapses. Journal of Neurophysiology, 80, 3336–3340. Bethus, I., Lemaire, V., Lhomme, M., & Goodall, G. (2005). Does prenatal stress affect latent inhibition? It depends on the gender. Behavioural Brain Research, 158, 331–338. Blasi, G., & Bertolino, A. (2006). Imaging genomics and response to treatment with antipsychotics in schizophrenia. NeuroRx, 3, 117–130. Brophy, K., Hawi, Z., Kirley, A., Fitzgerald, M., & Gill, M. (2002). Synaptosomalassociated protein 25 (SNAP-25) and attention deficit hyperactivity disorder (ADHD): Evidence of linkage and association in the Irish population. Molecular Psychiatry, 7, 913–917. Bruno, K. J., Freet, C. S., Twining, R. C., et al. (2007). Abnormal latent inhibition and impulsivity in coloboma mice, a model of ADHD. Neurobiology of Disease, 25, 206–216. Buonanno, A., & Fischbach, G. D. (2001). Neuregulin and ErbB receptor signaling pathways in the nervous system. Current Opinion in Neurobiology, 11, 287–296. Camargo, L. M., Collura, V., Rain, J. C., et al. (2007). Disrupted in schizophrenia 1 interactome: evidence for the close connectivity of risk genes and a potential synaptic basis for schizophrenia. Molecular Psychiatry, 12, 74–86. Campbell, D. B., Ebert, P. J., Skelly, T., et al. (2008). Ethnic stratification of the association of RGS4 variants with antipsychotic treatment response in schizophrenia. Biological Psychiatry, 63, 32–41. Cannon, T. D., Hennah, W., van Erp, T. G., et al. (2005). Association of DISC1/ TRAX haplotypes with schizophrenia, reduced prefrontal gray matter, and
242
Tatiana Lipina and John Roder
impaired short and long-term memory. Archives of General Psychiatry, 62, 1205–1213. Carlsson, A., Waters, N., Holm-Waters, S., et al. (2001). Interaction between monoamines, glutamate, and GABA in schizophrenia: new evidence. Annual Review of Pharmacology and Toxicology, 41, 237–260. Carroll, L. S., Kendall, K., O’Donovan, M. C., Owen, M. J., & Williams, N. M. (2009). Evidence that putative ADHD low risk alleles at SNAP25 may increase the risk of schizophrenia. American Journal of Medical Genetics B, Neuropsychiatric Genetics, 150, 893–899. Catts, V. S., Catts, S. V., McGrath, J. J., et al. (2006). Apoptosis and schizophrenia: a pilot study based on dermal fibroblast cell lines. Schizophrenia Research, 84, 20–28. Chang, T., Meyer, U., Feldon, J., & Yee, B. K. (2007). Disruption of the US preexposure effect and latent inhibition in two-way active avoidance by systematic amphetamine in C57BL/6 mice. Psychopharmacology, 191, 211–221. Chubb, J. E., Bradshaw, N. J., Soares, D. C., Porteous, D. J., & Millar, J. K. (2008). The DISC locus in psychiatric illness. Molecular Psychiatry, 13, 36–64. Chumakov, I., Blumenfeld, M., Guerassimenko, O., et al. (2002). Genetic and physiological data implicating the new human gene G72 and the gene for D-amino acid oxidase in schizophrenia. Proceedings of the National Academy of Sciences of the United States of America, 99, 13675–13680. Clapcote, S. J., Lipina, T. V., Millar, K. J., et al. (2007). Behavioral phenotypes of Disc1 missense mutations in mice. Neuron, 54, 387–402. Consolo, S., Bertorelli, R., Russi, G., Zambelli, M., & Ladinsky, H. (1994). Serotonergic facilitation of acetylcholine release in vivo from rat dorsal hippocampus via serotonin 5-HT3 receptors. Journal of Neurochemistry, 62, 2254–2261. Cookson, W., Liang, L., Abecasis, G., Moffatt, M., & Lathrop, M. (2009). Mapping complex disease traits with global gene expression. Nature Reviews Genetics, 10, 184–194. Corvin, A. P., Morris, D. W., McGhee, K., et al. (2004). Confirmation and refinement of an ‘at-risk’ haplotype for schizophrenia suggests the EST cluster, Hs.97362, as a potential susceptibility gene at the Neuregulin-1 locus. Molecular Psychiatry, 9, 208–213. Coyle, J. T. (1996). The glutamatergic dysfunction hypothesis for schizophrenia. Harvard Review of Psychiatry, 3, 241–253. Craddock, N., O’Donovan, M. C., & Owen, M. J. (2005). The genetics of schizophrenia and bipolar disorder: dissecting psychosis. Journal of Medical Genetics, 42, 193–204. Crestani, F., Assandri, R., Ta¨uber, M., Martin, J. R., & Rudolph, U. (2002). Contribution of the alpha1-GABA(A) receptor subtype to the pharmacological actions of benzodiazepine site inverse agonists. Neuropharmacology, 43, 679–684. Deckwerth, T. L., Elliott, J. L., Knudson, C. M., et al. (1996). BAX is required for neuronal death after trophic factor deprivation and during development. Neuron, 17, 401– 411. De Luca, V., Wang, H., Squassina, A., et al. (2004). Linkage of M5 muscarinic and alpha7-nicotinic receptor genes on 15q13 to schizophrenia. Neuropsychobiology, 50, 124–127.
The genetics of latent inhibition: studies of inbred and mutant mice
243
Den Boer, J. A., van Megen, H. J., Fleischhacker, W. W., et al. (1995). Differential effects of the D1-DA receptor antagonist SCH39166 on positive and negative symptoms of schizophrenia. Psychopharmacology (Berlin), 121, 317–322. Desbonnet, L., Waddington, J. L., & O’Tuathaigh, C. M. (2009). Mutant models for genes associated with schizophrenia. Biochemical Society Transactions, 37, 308–312. DeVietti, T. L., Bauste, R. L., Nutt, G., & Barrett, O. V. (1987). Latent inhibition: a trace conditioning phenomenon? Learning and Motivation, 8, 185–201. Devon, R. S., Anderson, S., Teague, P. W., et al. (2001a). The genomic organisation of the metabotropic glutamate receptor subtype 5 gene, and its association with schizophrenia. Molecular Psychiatry, 6, 311–314. Devon, R. S., Anderson, S., Teague, P. W., et al. (2001b). Identification of polymorphisms within Disrupted in Schizophrenia and Disrupted in Schizophrenia 2, and an investigation of their association with schizophrenia and bipolar affective disorder. Psychiatry Genetics, 11, 71–78. Doherty, A. J., Palmer, M. J., Henley, J. M., Collingridge, G. L., & Jane, D. E. (1997). (RS)-2-chloro-5-hydroxyphenylglycine (CHPG) activates mGluR5, but no mGluR1, receptors expressed in CHO cells and potentiates NMDA responses in the hippocampus. Neuropharmacology, 36, 265–267. Duffy, S., Labrie, V., & Roder, J. C. (2008). D-serine augments NMDA-NR2B receptor-dependent hippocampal long-term depression and spatial reversal learning. Neuropsychopharmacology, 33, 1004–1018. Falls, D. L. (2003). Neuregulins: functions, forms and signaling strategies. Experimental Cell Research, 284, 14–30. Freedman, R., Waldo, M., Bickford-Wilmer, P., & Nagamoto, H. (1991). Elementary neuronal dysfunctions in schizophrenia. Schizophrenia Research, 4, 233–243. Gerber, D. J., Hall, D., Miyakawa, T., et al. (2003). Evidence for association of schizophrenia with genetic variation in the 8p21.3 gene, PPP3CC, encoding the calcineurin gamma subunit. Proceedings of the National Academy of Sciences of the United States of America, 100, 8993–8998. Gerdjikov, T. V., Rudolph, U., Keist, R., et al. (2008). Hippocampal alpha 5 subunitcontaining GABA A receptors are involved in the development of the latent inhibition effect. Neurobiology of Learning & Memory, 89, 87–94. Gerlai, R., Pisacane, P., & Erickson, S. (2000). Heregulin, but not ErbB2 or ErbB3, heterozygous mutant mice exhibit hyperactivity in multiple behavioral tasks. Behavioural Brain Research, 109, 219–227. Goff, D. C., & Coyle, J. T. (2001). The emerging role of glutamate in the pathophysiology and treatment of schizophrenia. American Journal of Psychiatry, 158, 1367–1377. Goldstein, J. M., Seidman, C. J., O’Brien, L. M., et al. (2002). Impact of normal sexual dimorphisms on sex differences in structural brain abnormalities in schizophrenia assessed by magnetic resonance imaging. Archives of General Psychiatry, 59, 154–164. Goltsov, A. Y., Loseva, J. G., Andreeva, T. V., et al. (2006). Polymorphism in the 5¢-promoter region of serine racemase gene in schizophrenia. Molecular Psychiatry, 11, 325–326. Gould, J., & Wehner, J. M. (1999). Genetic influence on latent inhibition. Behavioral Neuroscience, 113, 1291–1296.
244
Tatiana Lipina and John Roder
Groth, R. D., Dunbar, R. L., & Mermelstein, P. G. (2003). Calcineurin regulation of neuronal plasticity. Biochemical and Biophysical Research Communications, 311, 1159–1171. Guilin, O., Abi-Dargham, A., & Laruelle, M. (2007). Neurobiology of dopamine in schizophrenia. International Review of Neurobiology, 78, 1–39. Hahn, C. G., Wang, H. Y., Cho, D. S., et al. (2006). Altered neuregulin 1-erbB4 signalling contributes to NMDA receptor hypofunction in schizophrenia. Nature Medicine, 12, 824–828. Harrell, A. V., & Allan, A. M. (2003). Improvements in hippocampal-dependent learning and decremental attention in 5-HT3 receptor overexpressing mice. Learning & Memory, 10, 410–419. Harrison, P. J., & Weinberger, D. R. (2005). Schizophrenia genes, gene expression, and neuropathology: on the matter of their convergence. Molecular Psychiatry, 10, 40–68. Hashimoto, R., Straub, R. E., Weickert, C. S., et al. (2004). Expression analysis of neuregulin-1 in the dorsolateral prefrontal cortex in schizophrenia. Molecular Psychiatry, 9, 299–307. Hauser, J., Rudolph, U., Keist, R., et al. (2005). Hippocampal alpha5 subunitcontaining GABAA receptors modulate the expression of prepulse inhibition. Molecular Psychiatry, 10, 201–207. Hennah, W., Tuulio-Henriksson, A., Paunio, T., et al. (2005). A haplotype within the DISC1 gene is associated with visual memory functions in families with high density of schizophrenia. Molecular Psychiatry, 10, 1097–1103. Hess, E. J., Collins, K. A., & Wilson, M. C. (1996). Mouse model of hyperkinesis implicates SNAP-25 in behavioral regulation. Journal of Neuroscience, 16, 3104–3111. Hess, E. J., Jinnah, H. A., Kozak, C. A., & Wilson, M. C. (1992). Spontaneous locomotor hyperactivity in a mouse mutant with a deletion including the Snap gene on chromosome 2. Journal of Neuroscience, 12, 2865–2874. Hikida, T., Jaaro-Peled, H., Seshadri, S., et al. (2007). Dominant-negative DISC1 transgenic mice display schizophrenia-associated phenotypes detected by measures translatable to humans. Proceedings of the National Academy of Sciences of the United States of America, 104, 14501–14506. Hitzemann, B., Dains, K., Kanes, S., & Hitzemann, R. (1994). Further studies on the relationship between dopamine cell density and haloperidol-induced catalepsy. Journal of Pharmacology and Experimental Therapeutics, 271, 969–976. Ibrahim, H. M., Hogg, A. J., Healy, D. J., et al. (2000). Ionotropic glutamate receptor binding and subunit mRNA expression in thalamic nuclei in schizophrenia. American Journal of Psychiatry, 157, 1811–1823. Jarskog, L. F., Selinger, E. S., Lieberman, J. A., & Gilmore, J. H. (2004). Apoptotic proteins in the temporal cortex in schizophrenia: high Bax/Bcl-2 ratio without caspase-3 activation. American Journal of Psychiatry, 161, 109–115. Javitt, D. C., Balla, A., Sershen, H., & Lajtha, A. (1999). Reversal of phencyclidineinduced effects by glycine and glycine transport inhibitors. Biological Psychiatry, 45, 668–679. Javitt, D. C., & Zukin, S. R. (1991). Recent advances in the phencyclidine model of schizophrenia. American Journal of Psychiatry, 148, 1301–1308.
The genetics of latent inhibition: studies of inbred and mutant mice
245
Ji, X., Takahashi, N., Branko, A., et al. (2008). An association between serotonin receptor 3B gene (HTR3B) and treatment-resistant schizophrenia (TRS) in a Japanese population. Nagoya Journal of Medical Science, 70, 11–17 Jones, M. D., & Hess, E. J. (2003). Norepinephrine regulates locomotor hyperactivity in the mouse mutant coloboma. Pharmacology, Biochemistry & Behavior, 75, 209–216. Kapur, S., & Mamo, D. (2003). Half a century of antipsychotics and still a central role for dopamine D-2 receptors. Progress in Neuro-Psychopharmacology & Biological Psychiatry, 27, 1081–1090. Kew, J. N. C., Koester, A., Moreau, J. L., et al. (2000). Functional consequences of reduction in NMDA receptor glycine affinity in mice carrying targeted point mutations in the glycine binding site. Journal of Neuroscience, 20, 4037–4049. Killcross, A. S., & Robbins, T. W. (1993). Differential effects of intra-accumbens and systematic amphetamine on latent inhibition using an on-baseline, withinsubject conditioned suppression paradigm. Psychopharmacology, 110, 479–489. Klee, C. B., Draetta, G. F., & Hubbard, M. J. (1988). Calcineurin. Advances in Enzymology and Related Areas of Molecular Biology, 61, 149–200. Kline, L., Decena, E., Hitzemann, R., & McCaughran, J. Jr. (1998). Acoustic startle, prepulse inhibition, locomotion, and latent inhibition in the neurolepticresponsive (NR) and neuroleptic-nonresponsive (NNR) lines of mice. Psychopharmacology (Berl.), 139, 322–331. Knable, M. B., Barci, B. M., Webster, M. J., Meador-Woodruff, J., & Torrey, E. F. (2004). Molecular abnormalities of the hippocampus in severe psychiatric illness: postmortem findings from the Stanley Neuropathology Consortium. Molecular Psychiatry, 9, 609–620. Koike, H., Arguello, P. A., Kvajo, M., Karayiorgou, M., & Gogos, J. A. (2006). Disc1 is mutated in the 129S6/SvEv strain and modulates working memory in mice. Proceedings of the National Academy of Sciences of the United States of America, 103, 3693–3697. Krystal, J. H., Karper, L. P., Seibyl, J. P., et al. (1994). Subanesthetic effects of the noncompetitive NMDA antagonist, ketamine, in humans. Psychotomimetic, perceptual, cognitive, and neuroendocrine responses. Archives of General Psychiatry, 51, 199–214. Kumari, V. (2009). Latent inhibition and schizophrenia: A review of the empirical literature. In R. E. Lubow and I. Weiner (Eds.), Latent Inhibition: Data, Theories, and Applications to Schizophrenia. Cambridge: Cambridge University Press. Kuno, T., Mukai, H., Ito, A., et al. (1992). Distinct cellular expression of calcineurin A-alpha and A-beta in rat brain. Journal of Neurochemistry, 58, 1643–1651. Kustanovich, V., Merriman, B., McGough, J., et al. (2003). Biased paternal transmission of SNAP-25 risk alleles in attention-deficit hyperactivity disorder. Molecular Psychiatry, 8, 309–315. Labrie, V., Lipina, T., & Roder, J. C. (2008). Mice with reduced NMDA receptor glycine affinity model some of the negative and cognitive symptoms of schizophrenia. Psychopharmacology (Berlin), 200, 217–230. Law, A. J., Kleinman, J. E., Weinberger, D. R., & Weickert, C. S. (2007). Diseaseassociated intronic variants in the ErbB4 gene are related to altered ErbB4 splice variant expression in the brain in schizophrenia. Human Molecular Genetics, 16, 129–141.
246
Tatiana Lipina and John Roder
LeDoux, J. (2003). The emotional brain, fear, and the amygdala. Cellular and Molecular Neurobiology, 23, 727–738. Le Saux, M., Morissette, M., & Di Paolo, T. (2006). Erbeta mediates the estradiol increase of D2 receptors in rat striatum and nucleus accumbens. Neuropharmacology, 50, 451–457. Lewine, R. R., Walkert, E. F., Shurett, R., Caudle, J., & Haden, C. (1996). Sex differences in neuropsychological functioning among schizophrenic patients. American Journal of Psychiatry, 153, 1178–1184. Lipina, T., Labrie, V., Weiner, I., & Roder, J. (2005). Modulators of the glycine site on NMDA receptors, D-serine and ALX 5407, display similar beneficial effects to clozapine in mouse models of schizophrenia. Psychopharmacology, 179, 54–67. Lipina, T., Weiss, K., & Roder, J. (2007). The Ampakine CX546 restores the prepulse inhibition and latent inhibition deficits in mGluR5 deficient mice. Neuropsychopharmacology, 32, 745–756. Lubow, R. E. (1989). Latent Inhibition and Conditioned Attentional Theory. Cambridge: Cambridge University Press. Lubow, R. E. (2009). Phylogenetic distribution of latent inhibition. In R. E. Lubow and I. Weiner (Eds.), Latent Inhibition: Data, Theories, and Applications to Schizophrenia. Cambridge: Cambridge University Press. Lubow, R. E., Braunstein-Bercovitz, H., Blumenthal, O., Kaplan, O., & Toren, P. (2005). Latent inhibition and asymmetrical visual-spatial attention in children with ADHD. Child Neuropsychology, 11, 445–457. Mansuy, I. M. (2003). Calcineurin in memory and bidirectional plasticity. Biochemical and Biophysical Research Communications, 311, 1195–1208. Matsumoto, M., Yoshioka, M., Togashi, H., et al. (1995). Modulation of norepinephrine release by serotonergic receptors in the rat hippocampus as measured by in vivo microdialysis. Journal of Pharmacology and Experimental Therapeutics, 272, 1044–1051. McMahon, L. L., & Kauer, J. A. (1997). Hippocampal interneurons are excited via serotonin-gated ion channels. Journal of Neurophysiology, 78, 2493–2502. Meador-Woodruff, J. H., Clinton, S. M., Beneyto, M., & McCullumsmith, R. E. (2003). Molecular abnormalities of the glutamate synapse in the thalamus in schizophrenia. Annals of the New York Academy of Sciences, 1003, 75–93. Meador-Woodruff, J. H., & Healy, D. J. (2000). Glutamate receptor expression in schizophrenia brain. Brain Research Interactive, 31, 288–294. Mei, L., & Xiong W. C. (2008). Neuregulin 1 in neural development, synaptic plasticity and schizophrenia. Nature Reviews Neuroscience, 9, 437–452. Meyer, U., Feldon, J., Schedlowski, M., & Yee, B. K. (2005). Towards an immunoprecipitated neurodevelopmental animal model of schizophrenia. Neuroscience and Biobehavioral Reviews, 29, 913–947. Mill, J., Curran, S., Kent, L., et al. (2002). Association study of a SNAP-25 microsatellite and attention deficit hyperactivity disorder. American Journal of Medical Genetics, 114, 269–271. Millan, M. J. (2005). N-methyl-D-aspartate receptors as a target for improved antipsychotic agents: novel insights and clinical perspectives. Psychopharmacology, 179, 30–53. Millar, J. K., Wilson-Annan, J. C., Anderson, S., et al. (2000). Disruption of two novel genes by a translocation co-segregating with schizophrenia. Human Molecular Genetics, 9, 1415–1423.
The genetics of latent inhibition: studies of inbred and mutant mice
247
Miller, R. R., & Escobar, M. (2009). Latent inhibition as a performance effect. In R. E. Lubow and I. Weiner (Eds.), Latent Inhibition: Data, Theories, and Applications to Schizophrenia. Cambridge: Cambridge University Press. Miyakawa, T., Leiter, L. M., Gerber, D. J., et al. (2003). Conditional calcineurin knockout mice exhibit multiple abnormal behaviors related to schizophrenia. Proceedings of the National Academy of Sciences of the United States of America, 100, 8987–8992. Mohler, H., Rudoulph, U., Boison, D., et al. (2008). Regulation of cognition and symptoms of psychosis: focus on GABAA receptors and glycine transporter 10 . Pharmacology, Biochemistry and Behavior, 90, 58–64. Moran, P. M., Fischer, T. R., Hitchcock, J. M., & Moser, P. C. (1996). Effects of clozapine on latent inhibition in the rat. Behavioral Pharmacology, 7, 42–48. Moretti, L., Pentika¨inen, O. T., Settimo, L., & Johnson, M. S. (2004). Model structures of the NMDA receptor subunit NR1 explain the molecular recognition of agonist and antagonist ligands. Journal of Structural Biology, 145, 205–215. Morita, Y., Ujike, H., Tanaka, Y., et al. (2006). A genetic variant of the serine racemase gene is associated with schizophrenia. Biological Psychiatry, 61, 1200–1203. Moser, P. C., Hitchcock, J. M., Lister, S., & Moran, P. M. (2000). The pharmacology of latent inhibition as an animal model of schizophrenia. Brain Research Reviews, 33, 275–307. Nayak, S. V., Ronde, P., Spier, A. D., Lummis, S. C. R., & Nichols, R. A. (1999). Calcium changes induced by presynaptic 5-hydroxytryptamine-3 serotonin receptors on isolated terminals from various regions of the rat brain. Neuroscience, 91, 107–117. Ng, D., Pitcher, G. M., Szilard, R. K., et al. (2009). Neto1 is a novel CUB-domain NMDA receptor-interacting protein required for synaptic plasticity and learning. Public Library of Sciences, Biology, 24, 7:e41. Norton, N., Moskvina, V., Morris, D. W., et al. (2006). Evidence that interaction between neuregulin 1 and its receptor erbB4 increases susceptibility to schizophrenia. American Journal of Medical Genetics B: Neuropsychiatry Genetics, 141, 96–101. Numata, S., Ueno, S., Iga, J., et al. (2008). Positive association of the PDE4B (phosphodiesterase 4B) gene with schizophrenia in the Japanese population. Journal of Psychiatric Research, 43, 7–12. Ohnuma, T., Augood, S. J., Arai, H., McKenna, P. J., & Ernson, P. C. (1998). Expression of the human excitatory amino acid transporter 2 and metabotropic glutamate receptos 3 and 5 in the prefrontal cortex from normal individuals and patients with schizophrenia. Molecular Brain Research, 56, 207–217. Olijslagers, J., Werkman, T., McCreary, A., Kruse, C., & Wadman, W. (2006). Modulation of midbrain dopamine neurotransmission by serotonin, a versatile interaction between neurotransmitters and significance for antipsychotic drug action. Current Neuropharmacology, 4, 59–68. Ortells, M. O., & Lunt, G. G. (1995). Evolutionary history of the ligand-gated ion-channel superfamily of receptors. Trends in Neurosciences, 18, 121–127. Owen, E. H., Christensen, S. C., Paylor, R., & Wehner, J. M. (1997). Identification of quantitative trait loci involved in contextual and auditory-cued fear
248
Tatiana Lipina and John Roder
conditioning in BXD recombinant inbred strains. Behavioural Neuroscience, 111, 292–300. Pen˜as-Lledo´, E. M., Dorado, P., Ca´ceres, M. C., de la Rubia, A., & Llerena, A. (2007). Association between T102C and A-1438G polymorphisms in the serotonin receptor 2A (5-HT2A) gene and schizophrenia: relevance for treatment with antipsychotic drugs. Clinical Chemistry and Laboratory Medicine, 45, 835–838. Perez, G. I., Jurisicova, A., Wise, L., et al. (2007). Absence of pro-apoptotic Bax protein extends fertility and alleviates post-menopausal health complications in aging females. Proceedings of the National Academy of Sciences of the United States of America, 104, 5229–5234. Perry, W., Geyer, M. A., & Braff, D. L. (1999). Sensorimotor gaiting and thought disturbance measured in close temporal proximity in schizophrenic patients. Archives of General Psychiatry, 56, 277–281. Petryshen, T. L., Middleton, F. A., Tahl, A. R., et al. (2005). Genetic investigation of chromosome 5q GABAA receptor subunit genes in schizophrenia. Molecular Psychiatry, 10, 1074–1088. Pickard, B. S., Thomson, P. A., Christoforou, A., et al. (2007). The PDE4B gene confers sex-specific protection against schizophrenia. Psychiatric Genetics, 17, 129–133. Pletnikov, M. V., Ayhan, Y., Nikolskaia, O., et al. (2008). Inducible expression of mutant human DISC1 in mice is associated with brain and behavioral abnormalities reminiscent of schizophrenia. Molecular Psychiatry, 13, 173–186. Ponder, C. A., Huded, C. P., Munoz, M. B, et al. (2008). Rapid selection response for contextual fear conditioning in a cross between C57BL/6J and A/J: behavioral, QTL and gene expression analysis. Behavioral Genetics, 38, 277–291. Ponder, C. A., Kliethermes, C. L., Drew, M. R., et al. (2007). Selection for contextual fear conditioning affects anxiety-like behaviors and gene expression. Genes Brain Behavior, 6, 736–749. Porteous, D. J., Thomson, P., Brandon, N. J., & Millar, J. K. (2006). The genetics and biology of Disc1—an emerging role in psychosis and cognition. Biological Psychiatry, 60, 123–131. Radcliffe, R. A., Lowe, M. V., & Wehner, J. M. (2000). Confirmation of contextual fear conditioning QTLs by short-term selection. Behavioral Genetics, 30, 183–191. Rao, S. G., Williams, G. V., & Goldman-Rakic, P. S. (2000). Destruction and creation of spatial tuning by disinhibition: GABA(A) blockade of prefrontal cortical neurons engaged by tuning working memory. Journal of Neuroscience, 20, 485–494. Reif, A., Schmitt, A., Fritzen, S., & Lesch, K. P. (2007). Neurogenesis and schizophrenia: dividing neurons in a divided mind? European Archives of Psychiatry and Clinical Neuroscience, 257, 290–299. Rimer, M., Barrett, D. W., Maldonado, M. A., Vock, V. M., & Gonzalez-Lima, F. (2005). Neuregulin-1 immunoglobulin-like domain mutant mice: clozapine sensitivity and impaired latent inhibition. Neuroreport, 16, 271–275. Ross, C. A., Margolis, R. L., Reading, S. A., Pletnikov, M., & Coyle, J. T. (2006). Neurobiology of schizophrenia. Neuron, 52, 139–153.
The genetics of latent inhibition: studies of inbred and mutant mice
249
Russig, H., Murphy, C. A., & Feldon, J. (2002). Clozapine and haloperidol reinstate latent inhibition following its disruption during amphetamine withdrawal. Neuropsychopharmacology, 26, 765–777. Schumacher, J., Jamra, R. A., Freudenberg, J., et al. (2004). Examination of G72 and D-amino acid oxidase as genetic risk factor for schizophrenia and bipolar affective disorder. Molecular Psychiatry, 9, 203–207. Shen, S., Lang, B., Nakamoto, C., et al. (2008). Schizophrenia-related neural and behavioral phenotypes in transgenic mice expressing truncated Disc1. Journal of Neuroscience, 28, 10893–10904. Silberberg, G., Darvasi, A., Pinkas-Kramarski, R., & Navon, R. (2006). The involvement of ErbB4 with schizophrenia: association and expression studies. American Journal of Medical Genetics B, Neuropsychiatric Genetics, 141, 142–148. Smith, K. E., Borden, L. A., Hartig, P. R., Branchek, T., & Weinshank, R. L. (1992). Cloning and expression of a glycine transporter reveal colocalization with NMDA receptors. Neuron, 8, 927–935. Smith, S. E. P., Li, J., Garbett, K., Mirnics, K., & Patterson, P. H. (2007). Maternal immune activation alters fetal brain development through interleukin-6. Journal of Neuroscience, 27, 10695–10702. Stefansson, H., Sigurdsson, E., Steinhorsdottir, V., et al. (2002). Neuregulin 1 and susceptibility to schizophrenia. American Journal of Human Genetics, 71, 877–892. Strauss, M. E., Buchanan, R. W., & Hale, J. (1993). Relations between attentional deficits and clinical symptoms in schizophrenic outpatients. Psychiatry Research, 47, 205–213. Sun, S., Wang, F., Wei, J., et al. (2008). Association between interleukin-6 receptor polymorphism and patients with schizophrenia. Schizophrenia Research, 102, 346–347. Sun, W., & Oppenheim, R. W. (2003). Response of motoneurons to neonatal sciatic nerve axotomy in Bax-knockout mice. Molecular and Cellular Neurosciences, 24, 875– 886. Takaishi, T., Saito, N., Kuno, T., & Tanaka, C. (1991). Differential distribution of the mRNA encoding two isoforms of the catalytic subunit of calcineurin in the rat brain. Biochemical and Biophysical Research Communications, 174, 393–398. Tandon, R., & Greden, J. F. (1989). Cholinergic hyperactivity and negative schizophrenia symptoms: a model of cholinergic/dopaminergic interactions in schizophrenia. Archives of General Psychiatry, 46, 745–753. Tecott, L. H., & Wehner, J. M. (2001). Mouse molecular genetic technologies: Promise for psychiatric research. Archives of General Psychiatry, 58, 995–1004. Thompson, P. M., Sower, A. C., & Perrone-Bizzozero, N. I. (1998). Altered levels of the synaptosomal associated protein SNAP-25 in schizophrenia. Biological Psychiatry, 43, 239–243. Tomppo, L., Hennah, W., Lahermo, P., et al. (2009). Association between genes of disrupted in Schizophrenia 1 (DISC1) interactors and schizophrenia supports the role of the DISC1 pathway in the etiology of major mental illnesses. Biological Psychiatry, 65, 1055–1062. Tovar, K. R., & Westbrook, G. L. (2002). Mobile NMDA receptors at hippocampal synapses. Neuron, 34, 255–264.
250
Tatiana Lipina and John Roder
Usall, J., Suarez, D., & Haro, J. M. (2007). Gender differences in response to antipsychotic treatment in outpatients with schizophrenia. Psychiatry Research, 153, 225–231. Van Berckel, B. N. M. (2003) Glutamate and schizophrenia. Current Neuropharmacology, 1, 351–370. Van Hooft, J. A., & Vijverberg, H. P. M. (2000). 5-HT3 receptors and neurotransmitter release in the CNS: a nerve ending story? Trends in Neurosciences, 23, 605–610. Volk, D. W., & Lewis, D. A. (2002). Impaired prefrontal inhibition in schizophrenia: relevance for cognitive dysfunction. Physiology & Behavior, 77, 501–505. Waddington, J. L., Corvin, A. P., Donohoe, G., et al. (2007). Functional genomics and schizophrenia: endophenotypes and mutant models. The Psychiatric Clinics of North America, 30, 365–399. Wang, H., Ng, K., Hayes, D., et al. (2004). Decreased amphetamine-induced locomotion and improved latent inhibition in mice mutant for the M5 muscarinic receptor gene found in the human 15q schizophrenia region. Neuropsychopharmacology, 29, 2126–2139. Wang, L., Yu, L., He, G., et al. (2007). Response of risperidone treatment may be associated with polymorphisms of HTT gene in Chinese schizophrenia patients. Neuroscience Letters, 414, 1–4. Warburton, E. C., Joseph, M. H., Feldon, J., Weiner, I., & Gray, J. A. (1994). Antagonism of amphetamine-induced disruption of latent inhibition in rats by haloperidol and ondansetron – implications for a possible antipsychotic action of ondansetron. Psychopharmacology, 114, 657–664. Wehner, J. M., Radcliffe, R. A., Rosmann, S. T., et al. (1997). Quantitative trait locus analysis of contextual fear conditioning in mice. Nature Genetics, 17, 331–334. Weiner, I. (1990). Neural substrate of latent inhibition: the switching model. Psychological Bulletin, 108, 442–461. Weiner, I. (2003). The “two-headed” latent inhibition model of schizophrenia: modeling positive and negative symptoms and their treatment. Psychopharmacology (Berl.), 169, 257–297. Weiner, I. (2009). The latent inhibition model of schizophrenia. In R. E. Lubow and I. Weiner (Eds.), Latent Inhibition: Data, Theories, and Applications to Schizophrenia. Cambridge: Cambridge University Press. Weiner, D. M., Levey, A. I., & Brann, M. R. (1990). Expression of muscarinic acetylcholine and dopamine receptor mRNAs in rat basal ganglia. Proceedings of the National Academy of Sciences of the United States of America, 87, 7050–7054. Weiner, I., Lubow, R. E., & Feldon, J. (1988). Disruption of latent inhibition by acute administration of low doses of amphetamine. Pharmacology, Biochemistry & Behavior, 30, 871–878. Weiner, I., Schiller, D., & Gaisler-Salomon, I. (2003). Disruption and potentiation of latent inhibition by risperidone: the latent inhibition model of atypical antipsychotic action. Neuropsychopharmacology, 28, 499–509. Weiner, I., Shadach, E., Barkai, R., & Feldon, J. (1997). Haloperidol- and clozapineinduced enhancement of latent inhibition with extended conditioning: implications for the mechanism of action of neuroleptic drugs. Neuropsychopharmacology, 16, 42–50.
The genetics of latent inhibition: studies of inbred and mutant mice
251
White, F. A., Keller-Peck, C. R., Knudson, C. M., Korsmeyer, S. J., & Snider, W. D. (1998). Widespread elimination of naturally occurring neuronal death in Bax-deficient mice. Journal of Neuroscience, 18, 1428–1439. Williams, J. H., Wellman, N. A., Geaney, D. P., et al. (1997). Haloperidol enhances latent inhibition in visual tasks in healthy people. Psychopharmacology, 133, 262–268. Wolpowitz, D., Mason, T. B., Dietrich, P., et al. (2000). Cysteine-rich domain isoforms of the neuregulin-1 gene are required for maintenance of peripheral synapses. Neuron, 25, 79–91. Wong, A. H., Buckle, C. E., & Van Tol, H. H. (2000). Polymorphisms in dopamine receptors: what do they tell us? European Journal of Pharmacology, 410, 183–203. Woo, T. U., Whitehead, R. E., Melchitzky, D. S., & Lewis, D. A. (1998). A subclass of prefrontal gamma-aminobutyric acid axon terminals are selectively altered in schizophrenia. Proceedings of the National Academy of Sciences of the United States of America, 95, 5341–5346. Xing, Q., Qian, X., Li, H., et al. (2007). The relationship between the therapeutic response to risperidone and the dopamine D2 receptor polymorphism in Chinese schizophrenia patients. International Journal of Neuropsychopharmacology, 10, 631–637. Yamakura, T., & Shimoh, K. (1999). Subunit- and site-specific pharmacology of the NMDA receptor channel. Progress in Neurobiology, 59, 279–298. Yee, B. K., Balic, E., Singer, P., et al. (2006). Disruption of glycine transporter 1 restricted to forebrain neurons is associated with a procognitive and antipsychotic phenotypic profile. Journal of Neuroscience, 26, 3169–3181. Yee, B. K., Hauser, J., Dolgov, V. V., et al. (2004). GABA receptors containing the alpha5 subunit mediate the trace effect in aversive and appetitive conditioning and extinction of conditioned fear. European Journal of Neuroscience, 20, 1928–1936. Yeomans, J. S. (1995). Role of tegmental cholinergic neurons in dopaminergic activation, antimuscarinic psychosis and schizophrenia. Neuropsychopharmacology, 12, 3–16. Yeomans, J. S., Forster, G., & Blaha, C. (2001). M5 muscarinic receptors are needed for slow activation of dopamine neurons and for rewarding brain stimulation. Life Science, 68, 2449–2456. Zeng, H., Chattarji, S., Barbarosie, M., et al. (2001). Forebrain-specific calcineurin knockout selectively impairs bidirectional synaptic plasticity and working/ episodic-like memory. Cell, 107, 617–629. Zhang, W., Yamada, M., Gomeza, J., Basile, A. S., & Wess, J. (2002). Multiple muscarinic acetylcholine receptor subtypes modulate striatal dopamine release, as studied with M1-M5 muscarinic receptor knockout mice. Journal of Neuroscience, 22, 6347–6352.
12 A comparison of mechanisms underlying the CS–US association and the CS–nothing association Thomas J. Gould
The ability to adapt to environmental change is essential for the survival of any organism. Adaptations to environmental change include the ability to learn associations and the ability to modify those associations. Numerous studies have demonstrated that deficits in the ability to modify or modulate learning are associated with multiple mental illnesses and disorders (Amieva, Phillips, Della, & Henry, 2004; Baron-Cohen & Belmonte, 2005; Baruch, Hemsley, & Gray, 1988; Clark & Goodwin, 2004; Kaplan et al., 2006; Lubow & Gewirtz, 1995; Lubow & Josman, 1993; Vaitl, Lipp, Bauer et al., 1999; Weiner, Schiller, & Gaisler-Salomon, 2003). Because of the strong link between mental illness and deficits in processes that modulate learning, understanding the neural substrates of these processes could facilitate development of treatments for many diseases. Latent inhibition is one process that can modulate learned associations and is also altered in patients with mental illness (see the chapters on schizophrenia in this book for an in-depth discussion). As described in preceding chapters in this book, latent inhibition is the phenomenon in which pre-exposure to a conditioned stimulus (CS) prior to the pairing of this CS with an unconditioned stimulus (US) decreases the subsequent conditioned responses (CR). There are three phases of latent inhibition: (1) pre-exposure to the CS, (2) training (or conditioning), and (3) testing. The presence of latent inhibition is identified by comparing the degree of conditioned responding between the CS pre-exposed group and the non-pre-exposed group. It has been debated whether preexposure to a CS disrupts subsequent learning or inhibits expression of the learned association (see Chapter 1). Work by Miller and others (Ackil, Carman, Bakner, & Riccio, 1992; Bakner, Strohan, Nordeen, & Riccio, 1991; Bouton, 1993, 2004; Holt & Maren, 1999; Kasprow, Catterson, Schachtman, & Miller, 1984; Talk, Stoll, & Gabriel, 2005), however, clearly showed that pre-exposure to the CS did not prevent learning of the CS–US association but did alter expression of the CR during testing. The findings that CS pre-exposure does not prevent learning of the CS–US association indicate that two separate learning processes occur during latent inhibition: one at CS Latent Inhibition: Cognition, Neuroscience and Applications to Schizophrenia, ed. R.E. Lubow and Ina Weiner. Published by Cambridge University Press # Cambridge University Press 2010.
252
A comparison of mechanisms underlying the CS–US association
253
pre-exposure and another at training; and that these associations have opposing effects on CR expression at recall. Thus, it is thought that during latent inhibition, the subject learns a CS–nothing association that forms in the CS pre-exposure phase and a CS–US association that forms in the training phase (De la Casa, Diaz, & Lubow, 2003; Lubow, 1989). At recall, the CS–nothing association inhibits expression of the CS–US association. If the causal mechanism for latent inhibition is the formation of a CS–nothing association, then there must be a neural process that underlies the formation of this association and a neural mechanism through which the CS–nothing association inhibits expression of the CS–US association. In order to better understand the neural substrates of latent inhibition, this chapter will review and compare the neural substrates of the CS–nothing association and the CS–US association. Latent inhibition can be demonstrated in a variety of behavioral paradigms including cued fear conditioning, conditioned taste aversion, avoidance conditioning, appetitive conditioning, and eye blink conditioning; to name but a few (Gould & Wehner, 1999; Han, Gallagher, & Holland, 1995; Lubow, 1973; Mcintosh & Tarpy, 1977; Nicholson & Freeman, 2002). Because these learning paradigms involve different neural substrates and areas, it is unclear whether one neural substrate exists for latent inhibition regardless of the behavioral paradigm, or whether the base learning task used to demonstrate latent inhibition determines the underlying neural substrates. To facilitate both the comparison between neural substrates underlying the CS–nothing association and the CS–US association and the examination of the mechanism by which the CS–nothing association inhibits the expression of the CS–US association on test day, this chapter will focus on latent inhibition of one learning paradigm, cued fear conditioning. Towards the end of the chapter, the discussion will extend to how the neural substrates of latent inhibition of cued fear conditioning may differ or be in common with the substrates of other types of latent inhibition. Before discussing the neural substrates of the CS–nothing association associated with latent inhibition of fear conditioning, the neural areas and underlying cell signaling cascades involved in fear conditioning will be briefly summarized. Fear conditioning Fear conditioning is one of the most extensively studied learning paradigms. In cued fear conditioning, animals are trained to form an association between a discrete CS and a foot shock US. One session of two training trials produces robust learning. The neural substrates of fear conditioning have been identified from the brain regions to the neurotransmitter systems to the cell signaling molecules and genetic factors involved. Fear conditioning: neural areas The amygdala is the brain area most commonly associated with fear conditioning, and rightly so. Multiple studies have demonstrated that the amygdala is critically involved in formation of the CS–US association during fear conditioning (Farb,
254
Thomas J. Gould
Aoki, & Ledoux, 1995; Li, Stutzmann, & Ledoux, 1996; Phillips & Ledoux, 1992). The thalamus is believed to relay the sensory information necessary for the formation of the CS–US association to the amygdala. In support, lesions of the medial geniculate nucleus, an auditory area of the thalamus, disrupted fear conditioning with an auditory CS but not with a visual CS (LeDoux, Iwata, Pearl, & Reis, 1986). The amygdala is a collection of nuclei that includes the lateral nucleus, basolateral nucleus, and the central nucleus; these nuclei mediate different processes associated with fear conditioning. In general, the lateral and basolateral nuclei are thought to receive sensory information that is integrated to form the CS–US association and this information is relayed to the central nucleus, the main efferent projecting area of the amygdala (Rogan & LeDoux, 1996). The central amygdala communicates with areas involved in the behavioral expression of fear such as the lateral hypothalamus and the periaqueductal gray (Ledoux, Iwata, Cicchetti, & Reis, 1988). Regions of the cerebral cortex and the cerebellum are also involved in cued fear conditioning. The cerebellum may play a role in retention of cued fear conditioning as indicated by the occurrence of learning-related plasticity in the cerebellum after fear conditioning (Sacchetti, Scelfo, Tempia, & Strata, 2004) and by the disruption of cued fear conditioning-associated memories by cerebellar lesions (Sacchetti, Baldi, Lorenzini, & Bucherelli, 2002a). Inactivation of the prefrontal cortex, the perirhinal cortex, or the parietal cortex at training produced long-term deficits in CR retention (Sacchetti, Baldi, Lorenzini, & Bucherelli, 2002b); however, another study found no effect of perirhinal lesions on cued fear conditioning (Bucci, Phillips, & Burwell, 2000). Inactivation of the prelimibic region of the medial prefrontal cortex at testing reduced CRs (Corcoran & Quirk, 2007). Finally, presentation of the CS during testing activated the anterior cingulate cortex (Milad et al., 2007). In addition to brain regions such as the amygdala that are critically involved in all types of fear conditioning, there are other brain areas that are selectively involved in specific types of fear conditioning. An example is when the CS is not a discrete stimulus but rather a configural stimulus such as a context. The hippocampus is critically involved in forming a contextual fear-based memory but is not critically involved in forming the CS–US (e.g., tone–shock) association (Kim, Rison, & Fanselow, 1993; Logue, Paylor, & Wehner, 1997; Phillips & Ledoux, 1992). The nucleus accumbens may also be involved in contextual fear associations. Lesioning the nucleus accumbens disrupted contextual but not cued fear conditioning (Levita, Dalley, & Robbins, 2002; Riedel, Harrington, Hall, & Macphail, 1997). The involvement of the nucleus accumbens, however, may depend on genetic background because lesioning the nucleus accumbens had no effect on cued fear conditioning in C57BL/6 mice but did disrupt cued fear conditioning in DBA mice (AmmassariTeule, Passino, Restivo, & De Marsanich, 2000). It should be noted that DBA mice show abnormal fear conditioning (Logue et al., 1997) and have been proposed as a model of schizophrenia (Hashimoto, Iyo, Freedman, & Stevens, 2005). The findings that the nucleus accumbens is not critically involved in cued fear conditioning are
A comparison of mechanisms underlying the CS–US association
255
particularly interesting in the context of this chapter because, as will be discussed later, the nucleus accumbens plays an important role in latent inhibition of cued fear conditioning (Weiner, 2003).
Fear conditioning: cell signaling In addition to a solid understanding of the neural substrates of fear conditioning, the cell signaling systems involved in fear conditioning have also been identified. NMDA receptors in the lateral and basolateral nuclei of the amygdala are critically involved in fear conditioning. Infusion of the NMDA receptor antagonist D,L-2-amino-5phosphovalerate (APV) into the amygdala blocks acquisition of cued fear conditioning (Goosens & Maren, 2003; Lee & Kim, 1998). NMDA receptors gate calcium and an increase in calcium influx may lead to activation of cell signaling molecules involved in learning. One such molecule is calcium calmodulin-dependent protein kinase II (CaMKII). Infusion of a CaMKII antagonist into the lateral nucleus of the amygdala prior to conditioning disrupted both short-term and long-term cued fear conditioning memories (Rodrigues, Farb, Bauer et al., 2004). Thus, CaMKII may mediate processes involved in short-term memory that are also necessary for the formation of long-term memory associated with fear conditioning. Calcium influx during fear conditioning may also lead to activation of cell signaling molecules directly involved in long-term memory formation. Examples of such molecules are cAMP-dependent protein kinase (PKA) and protein kinase C (PKC) (Micheau & Riedel, 1999). Mice deficient in the PKC subunit beta showed disrupted cued fear conditioning 24 h after training (Weeber et al., 2000). In addition, inhibition of PKC disrupted long-term but not short-term memory for cued fear conditioning (Ahi, Radulovic, & Spiess, 2004). Similar results were seen with PKA (Ahi et al., 2004; Schafe & Ledoux, 2000). A downstream target of PKA is extracellular regulated kinase (ERK) (Adams & Sweatt, 2002); ERK and PKC may interact to regulate cell signaling (Chen, Sarnecki, & Blenis, 1992). Infusion of an ERK inhibitor into the amygdala disrupted long-term memory for cued fear conditioning but spared short-term memory (Schafe et al., 2000). PKC, PKA, and ERK could contribute to long-term memory formation during fear conditioning through regulation of gene expression by activation of gene transcription factors. One such transcription factor is cAMP response element binding protein (CREB). Mice with altered alpha and delta isoforms of CREB had long-term but not short-term memory deficits for cued fear conditioning (Bourtchuladze et al., 1994). CREB-mediated gene transcription may lead to the synthesis of new proteins critically involved in long-term memory storage. In support, numerous studies demonstrated that protein synthesis inhibitors disrupt cued fear conditioning (Lewis & Gould, 2004b; Maren, Ferrario, Corcoran et al., 2003; Schafe and LeDoux, 2000; for review see Hernandez & Abel, 2008).
256
Thomas J. Gould
In summary, the lateral and basolateral nuclei of the amygdala are critically involved in the plasticity underlying fear conditioning. Sensory information converging in amygdala during fear conditioning increases glutamate release leading to the activation of NMDA receptors and associated gating of calcium. An increase in intracellular calcium levels leads to activation of cell signaling involved in short-term memory and long-term memory. Long-term memory is supported by the synthesis of new proteins (for further discussion of the neural substrates of fear conditioning see Kim & Jung, 2006; Rodrigues, Schafe, & LeDoux, 2004).
Latent inhibition: neural areas Just as with fear conditioning, studies examining the neural substrates underlying latent inhibition have demonstrated that there are regions that are critically involved in latent inhibition and regions that modulate latent inhibition. The hippocampus, amygdala, and orbitofrontal cortex appear to play modulatory roles in latent inhibition. The nucleus accumbens and entorhinal cortex, on the other hand, may be critical sites for latent inhibition. The following sections will first review data supporting critical roles for the nucleus accumbens and the entorhinal cortex in latent inhibition of cued fear conditioning, followed by a review of the areas that modulate latent inhibition of cued fear conditioning. Given that an extensive review of all of the neural substrates for latent inhibition is not possible, the goal of this chapter is to directly compare the neural substrates of cued fear conditioning and of the CS–nothing association associated with latent inhibition of cued fear conditioning. For further discussion of the neural substrates of latent inhibition, the reader is referred to the other chapters in the Neurobiology section of this book. The nucleus accumbens is critically involved in latent inhibition of cued fear conditioning. Lesioning the nucleus accumbens in mice (Restivo, Passino, Middei, & Ammassari-Teule, 2002) and in rats (Tai, Cassaday, Feldon, & Rawlins, 1995) disrupts latent inhibition of cued fear conditioning. Examining the contribution of accumbal subareas to latent inhibition of cued fear conditioning demonstrated that nucleus accumbens shell lesions disrupted latent inhibition while lesions of the entire nucleus accumbens resulted in stronger latent inhibition (Gal, Schiller, & Weiner, 2005). Relevant to the current discussion, neither partial nor full lesions of the nucleus accumbens disrupted learning the CS–US association. Gal, Schiller, and Weiner (2005) proposed that the shell and the core areas of the nucleus accumbens interact to regulate the expression of latent inhibition via modulation of dopamine levels. In support of accumbal dopamine levels critically regulating expression of the CS–nothing association versus the CS–US association, conditioning-related release of dopamine in the nucleus accumbens shell was reduced at testing in animals preexposed to the CS; no change was seen in the nucleus accumbens core (Murphy, Pezze, Feldon, & Heidbreder, 2000). Thus, the regulation of dopamine levels in
A comparison of mechanisms underlying the CS–US association
257
the nucleus accumbens may determine whether the CS–nothing or the CS–US association is expressed during testing. If the nucleus accumbens regulates switching between expression of the CS–nothing association and the CS–US association during testing, the question remains as to what area is critically involved in forming the CS–nothing association. The answer may be the entorhinal cortex. Irreversible lesions of the entorhinal cortex disrupted latent inhibition of cued fear conditioning in CS pre-exposed subjects without affecting learning of the CS–US association in non-pre-exposed subjects (Coutureau, Galani, Gosselin et al., 1999; Coutureau, Lena, Dauge, & Di, 2002; Yee, Feldon, & Rawlins, 1997). Recently, we found that reversible inactivation of the entorhinal cortex via direct infusion of the GABA agonist muscimol during either the pre-exposure phase or the testing phase disrupted latent inhibition of cued fear conditioning (Lewis & Gould, 2007a). In contrast, reversible inactivation of the entorhinal cortex during the training phase had no effect on latent inhibition in CS pre-exposed mice or learning the CS–US association in non-pre-exposed mice. Because inactivating the entorhinal cortex at either CS pre-exposure or at testing disrupted latent inhibition of cued fear conditioning, the entorhinal cortex may be the site where the CS–nothing association is formed during CS pre-exposure. Activation of the CS–nothing association in the entorhinal cortex at testing may thus inhibit the expression of the CS–US association. In contrast to the entorhinal cortex, the effect of hippocampal lesions on latent inhibition is mixed. For some tasks, hippocampal lesions disrupt latent inhibition (Han et al., 1995; Kaye & Pearce, 1987; Oswald et al., 2002; Schmajuk, Lam, & Christiansen, 1994; Solomon & Moore, 1975); however, hippocampal lesions failed to disrupt latent inhibition of cued fear conditioning (Clark, Feldon, & Rawlins, 1992; Coutureau et al., 1999; Holt & Maren, 1999). Additional research suggests that the hippocampus may play a context-specific role in latent inhibition. In particular, Holt and Maren (1999) found that inactivation of the dorsal hippocampus failed to disrupt latent inhibition of cued fear conditioning but did disrupt the context specificity of latent inhibition. In other words, disruption of latent inhibition caused by a shift in contextual information failed to occur in animals with hippocampal inactivation. A similar effect was seen with latent inhibition of an appetitive task (Honey & Good, 1993). Finally, neurotoxic lesions of the ventral hippocampus failed to disrupt latent inhibition but activation of the ventral hippocampus did disrupt latent inhibition of cued fear conditioning (Pouzet, Zhang, Weiner et al., 2004). Collectively, these data suggest that the hippocampus is not critically involved but can modulate latent inhibition and provide contextual information necessary for latent inhibition to be context-specific. The amygdala and orbitofrontal cortex may also modulate latent inhibition. The role of the amygdala in latent inhibition depends on the parameters used to produce the latent inhibition. Using standard parameters, lesion of the basolateral nucleus of the amygdala had no effect on latent inhibition of cued fear conditioning
258
Thomas J. Gould
(Weiner, Tarrasch, & Feldon, 1996b). However, when latent inhibition is attempted with suboptimal parameters, the involvement of the basolateral nucleus of the amygdala becomes clear. Normally, latent inhibition is not produced when too few CS pre-exposures or too many conditioning trials are given (Crowell & Anderson, 1972; Gaisler-Salomon & Weiner, 2003; Lipina, Labrie, Weiner, & Roder, 2005; Lubow, 1973). However, latent inhibition was produced under these conditions in rats with basolateral amygdala lesions (Schiller & Weiner, 2004, 2005; Schiller, Zuckerman, & Weiner, 2006). Similarly, lesions of orbitofrontal cortex resulted in expression of latent inhibition of cued fear conditioning when suboptimal latent inhibition procedures were used (Schiller & Weiner, 2004; Schiller et al., 2006). These cortical effects appear to be specific to the orbitofrontal region of prefrontal cortex because multiple studies have shown that lesions of the medial prefrontal cortex do not alter latent inhibition (Joel, Weiner, & Feldon, 1997; Lacroix, Broersen, Weiner, & Feldon, 1998; Lacroix, Spinelli, White, & Feldon, 2000). Thus, the basolateral nucleus of the amygdala and orbitofrontal cortex may inhibit the expression of latent inhibition when conditions exist that normally favor the expression of the CS–US association over latent inhibition; but when basolateral amygdala function or orbitofrontal cortical function is disrupted, the system may shift, giving greater weight to the pre-exposure effect and thus producing latent inhibition. It is clear that multiple neural areas are involved in latent inhibition and the role of these areas in latent inhibition may be determined by behavioral parameters. For instance, the results from lesion studies of the hippocampus appear equivocal but the degree of engagement of the hippocampus may depend on task parameters and on the role of contextual information in the latent inhibition procedure. The basolateral nucleus of the amygdala and the orbitofrontal cortex may not be critically involved in latent inhibition but may modulate the expression of latent inhibition. In contrast, the entorhinal cortex appears critically involved in the development of the CS– nothing association and expression of latent inhibition; and the nucleus accumbens is critically involved in regulating the expression of latent inhibition through modulating dopamine levels.
Latent inhibition: cell signaling Identifying the neural mechanisms by which CS pre-exposure changes the brain to produce the behavioral phenomenon of latent inhibition will further understanding of latent inhibition and the diseases that disrupt it, but this necessitates understanding the cellular and molecular substrates. This section will review the known cell signaling molecules involved in latent inhibition of cued fear conditioning in order to compare and contrast the substrates of the CS–nothing association and the CS–US association. The logical starting point is at the neurotransmitter level because of the information available on the cell signaling molecules at the receptor level and because
A comparison of mechanisms underlying the CS–US association
259
knowing which receptors are activated can guide evaluation of downstream cell signaling cascades. However, the neurotransmitters involved in latent inhibition of cued fear conditioning will only be briefly reviewed here as the chapter on “The pharmacology of latent inhibition” in this book covers this topic in detail. The role of dopamine, as well as other monoamine neurotransmitters such as serotonin, in latent inhibition has been well studied. In addition to monoaminergic neurotransmitters, evidence supports a role for the cholinergic and glutamatergic neurotransmitter systems in latent inhibition. Changes in dopamine levels can significantly impact latent inhibition. Direct infusion of amphetamine into the nucleus accumbens disrupted latent inhibition of cued fear conditioning (Solomon & Staton, 1982). Amphetamine-induced disruption of latent inhibition can be reversed by administration of typical or atypical antipsychotics, which work primarily through antagonism of dopamine D2 receptors in the nucleus accumbens and striatum, before pre-exposure and conditioning (Warburton, Joseph, Feldon et al., 1994; Weiner, Shadach, Tarrasch et al., 1996a). Furthermore, administration of both typical and atypical antipsychotics before pre-exposure and conditioning can produce latent inhibition of cued fear conditioning when preexposure or conditioning parameters are used that normally do not produce latent inhibition (Feldon & Weiner, 1991; Ruob, Weiner, & Feldon, 1998; Trimble, Bell, & King, 1997; Weiner et al., 1996a). These results support the hypothesis discussed in Weiner’s 2003 review that dopamine in the nucleus accumbens allows latent inhibition to be flexible. In situations that normally would not favor the expression of latent inhibition, a decrease in dopamine levels would facilitate latent inhibition; whereas an increase in dopamine levels during conditions that would produce latent inhibition results in disrupted latent inhibition. Thus, involvement of the nucleus accumbens and dopamine varies with task parameters. The ability of dopamine to alter latent inhibition has implications for schizophrenia; for review see Weiner, 1990 and 2003. Serotonin is another monoamine neurotransmitter shown to be involved in latent inhibition. Depletion of serotonin disrupts latent inhibition of cued fear conditioning (Lorden, Rickert, & Berry, 1983; Loskutova, Luk’yanenko, & Il’yuchenok, 1990; Loskutova, 2001; Solomon, Kiney, & Scott, 1978). Studies investigating the involvement of serotonin in latent inhibition demonstrate that different subregions of the raphe nucleus, the main source of serotonergic afferents, are involved in latent inhibition; lesions of the medial raphe nucleus, but not the dorsal raphe, disrupt latent inhibition (Solomon, Nichols, Kiernan et al., 1980). Interestingly, the 5-HT2A/2C agonist 1-(2,5-dimethoxy-4-iodophenyl)-2-aminopropane disrupted latent inhibition of cued fear conditioning (Hitchcock, Lister, Fischer, & Wettstein, 1997) while 5-HT1A antagonists (Killcross, Stanhope, Dourish, & Piras, 1997) and 5-HT2A antagonists (McDonald et al., 2003) all enhanced latent inhibition. These results seem to contradict serotonin-depletion studies; however, it should be noted that 1-(2,5-dimethoxy-4-iodophenyl)-2-aminopropane has been shown to be a potent
260
Thomas J. Gould
inhibitor of serotonin neuronal firing (Wright, Garratt, & Marsden, 1990). Furthermore, 5-HT1A presynaptic autoreceptors can exert an inhibitory control over the firing rate of serotonergic neurons (Evrard et al., 1999) and inhibiting them could increase serotonergic signaling. Thus, it is possible that agonism of 5-HT2A/2C or 5-HT1A receptors could produce an effect similar to that seen with serotonin depletion. Although the monoamines dopamine and serotonin are involved in latent inhibition, not all monoamines appear to be critically involved in latent inhibition. Disruption of noradrenergic signaling had no detrimental effect on latent inhibition (Lorden et al., 1983; Tsaltas, Preston, Rawlins et al., 1984). The overall evidence suggests that the cholinergic neurotransmitter system plays a modulatory rather than a critical role in latent inhibition. Lesions of the nucleus basalis magnocellularis, the source of cortical cholinergic afferents, did not disrupt latent inhibition of cued fear conditioning (Schauz & Koch, 1999). In another study, nucleus basalis magnocellularis lesions appeared to alter CS–US learning (Rochford, Sen, Rousse, & Welner, 1996), thereby precluding clear evaluation of latent inhibition. Further support of the contention that the cholinergic system may not be required for latent inhibition comes from studies in knockout mice lacking the b2 nicotinic acetylcholinergic receptor subunit (Caldarone, Duman, & Picciotto, 2000) and in mice treated with the broad-spectrum nicotinic receptor antagonist mecamylamine (Gould, Collins, & Wehner, 2001; Gould & Lewis, 2005). In these mice, latent inhibition of cued fear conditioning was intact. However, ligand-mediated effects at nicotinic acetylcholinergic receptors may modulate latent inhibition. The effects of the nicotinic acetylcholinergic receptor agonist nicotine on latent inhibition have been mixed. One study found that nicotine enhanced latent inhibition of cued fear conditioning in mice (Gould et al., 2001), one study found that nicotine disrupted latent inhibition of cued fear conditioning in rats (Joseph, Peters, & Gray, 1993), and a third study found that nicotine both enhanced and disrupted latent inhibition of cued fear conditioning in rats depending on the number of CS pre-exposures (Rochford, Sen, & Quirion, 1996). Thus, activation of nicotinic acetylcholinergic receptors may be sufficient to modulate latent inhibition depending on the dose and behavioral conditions, but nicotinic acetylcholinergic receptors may not be critically involved in latent inhibition. Studies of synaptic plasticity have suggested that NMDA glutamate receptor-gated calcium influx is critically involved in learning (for review see Lisman, 2003; Riedel, Platt, & Micheau, 2003). For latent inhibition, there are conflicting results using a variety of NMDA antagonists and treatment schedules, but overall evidence suggests a role for the NMDA receptor in the formation of the CS–nothing association during latent inhibition. For instance, phencyclidine (PCP; 1 and 5 mg/kg) had no effect on latent inhibition of cued fear conditioning (Weiner & Feldon, 1992). In contrast, studies using more potent and selective non-competitive NMDA receptor antagonists demonstrate a critical involvement of NMDA receptors in latent inhibition. Schauz and Koch (2000) found that direct infusion of the NMDA receptor
A comparison of mechanisms underlying the CS–US association
261
antagonist APV into the basolateral amygdala of rats at CS pre-exposure blocked latent inhibition of fear-potentiated startle. Similarly, we found that doses of 0.5 and 1.0 mg/kg of the NMDA receptor antagonist MK801 administered to mice at CS pre-exposure disrupted latent inhibition of cued fear conditioning (Davis & Gould, 2005; Lewis & Gould, 2004). In addition, direct infusion of APV into the entorhinal cortex of mice at CS pre-exposure disrupted latent inhibition (Lewis & Gould, 2007b). These results strongly suggest that NMDA receptors are involved in formation of the CS–nothing association. Discrepancies between the recent studies that used the NMDA receptor antagonists APV and MK801 and the study that used PCP could be due to differences in the effective doses and selectivity of the drugs or could be due to compensatory mechanisms that may overcome the effects of low doses of NMDA antagonists. In support, PCP appears to be less potent and less selective than non-competitive antagonists such as MK801 (Ellison, 1995). Furthermore, we reported that nicotinic acetylcholinergic receptors and NMDA receptors may mediate similar cellular processes during formation of the CS–nothing association (Gould & Lewis, 2005). Thus, latent inhibition may not be disrupted when a low or moderate dose of an NMDA receptor antagonist is used because nicotinic acetylcholinergic receptor-mediated processes may compensate for the effect of the NMDA inhibitor. However, when higher doses of NMDA receptor antagonists are used, processes mediated through nicotinic acetylcholinergic receptors may not be sufficient to compensate for a greater level of NMDA receptor inhibition. We have proposed that activation of NMDA receptors during pre-exposure to the CS initiates cell signaling cascades that mediate latent inhibition in a manner similar to the NMDA receptor cell signaling cascades that underlie various other learning processes. As discussed earlier, the activation of NMDA receptors and subsequent gating of calcium can initiate the cAMP/PKA/ERK second-messenger signaling involved in associative learning (Abel & Kandel, 1998; Abel et al., 1997; Berman, Hazvi, Neduva, & Dudai, 2000; Blum, Moore, Adams, & Dash, 1999; Chetkovich, Gray, Johnston, & Sweatt, 1991; Roberson et al., 1999; Schafe and LeDoux, 2000; Schafe et al., 2000; Selcher, Weeber, Varga et al., 2002; Szapiro, Vianna, McGaugh et al., 2003; Vossler et al., 1997; Waltereit & Weller, 2003; Walz, Rockenbach, Amaral et al., 1999). Evidence suggests that this signaling pathway is also involved in the formation of the CS–nothing association. If NMDA receptor activation of calcium signaling is critically involved in the formation of the CS–nothing association, altering calcium signaling and cAMP/PKA signaling during CS pre-exposure should disrupt latent inhibition. Indeed, inhibition of voltage-gated calcium channels prior to CS pre-exposure disrupted latent inhibition of cued fear conditioning (Barad, Blouin, & Cain, 2004). Furthermore, we found that amplification of the cAMP signaling pathway at CS pre-exposure by rolipram, a selective type 4 cAMP phosphodiesterase inhibitor, reversed MK801 induced impairments in latent inhibition of cued fear conditioning (Davis & Gould, 2005).
262
Thomas J. Gould
Because phosphodiesterases inactivate cAMP, phosphodiesterase inhibition should increase cAMP levels, which may counter the effects of NMDA receptor inhibition by MK801. In further support of cAMP/PKA involvement in latent inhibition, direct infusion of the cyclic-AMP inhibitor Rp-cAMPS into the entorhinal cortex at CS pre-exposure disrupted latent inhibition of cued fear conditioning (Lewis & Gould, 2007b). Together, these results support the hypothesis that NMDA receptors and calcium and cAMP/PKA signaling are involved in the formation of the CS–nothing association during latent inhibition of cued fear conditioning. Because ERK is activated by PKA (Roberson et al., 1999; Vossler et al., 1997; Waltereit & Weller, 2003) and involved in multiple types of learning (Berman et al., 2000; Blum et al., 1999; Schafe et al., 2000; Selcher, Atkins, Trzaskos et al., 1999; Walz et al., 1999), ERK may also be critically involved in latent inhibition. Mice treated systemically with an inhibitor of ERK signaling at CS pre-exposure had disrupted latent inhibition of cued fear conditioning (Lewis, Davis, & Gould, 2004). Furthermore, direct infusion of the ERK inhibitor U0126 into the entorhinal cortex at CS pre-exposure blocked latent inhibition of cued fear conditioning (Lewis & Gould, 2007b). In total, these studies suggests that a NMDA receptor!cAMP/ PKA!ERK cell signaling cascade in the entorhinal cortex is critically involved in the CS–nothing association underlying latent inhibition of cued fear conditioning. The mechanisms through which ERK may be exerting its effects on latent inhibition of cued fear conditioning remain unknown but potentially include two downstream targets of ERK: the gene transcription factor CREB (Davis, Vanhoutte, Pages et al., 2000; Perkinton, Sihra, & Williams, 1999) and/or microtubule-associated proteins (i.e., MAP2) (Quinlan & Halpain, 1996). If ERK activation supports latent inhibition through CREB activation of CRE-mediated genes, than latent inhibition would be dependent on protein synthesis. Work with inhibitors of translation, however, suggests that latent inhibition can occur in the absence of protein synthesis. We found that latent inhibition of cued fear conditioning can be established in the presence of the protein synthesis inhibitor anisomycin administered immediately before or after CS pre-exposure (Lewis & Gould, 2004). In contrast, the same treatment during conditioning of non CS pre-exposed mice disrupted learning. These findings suggest that protein synthesis may not be critically involved in the CS–nothing association formed during latent inhibition of cued fear conditioning. An alternative mechanism through which ERK could mediate the CS–nothing association is changes in microtubule-associated proteins. Lynch and Baudry proposed that cytoskeletal changes that occur during learning could increase the number of receptors available at the synapse and result in an increase in post-synaptic responses (Baudry & Lynch, 2001; Lynch & Baudry, 1984). Thus, during formation of the CS–nothing association, ERK could phosphorylate microtubule-associated proteins, producing an increase in activity-dependent cytoskeletal rearrangement that would support latent inhibition. This is just one possibility that needs testing because alternative explanations are also feasible. It is possible anisomycin disrupted
A comparison of mechanisms underlying the CS–US association
263
cued fear conditioning through mechanisms other than inhibition of protein synthesis and those processes are not critically involved in latent inhibition of cued fear conditioning. For instance, in addition to inhibiting protein synthesis, anisomycin activates JNK and P38 (Curtin & Cotter, 2002; Torocsik & Szeberenyi, 2000). These issues need further examination.
Similarities and differences between the neural substrates of latent inhibition and fear conditioning After reviewing the neural substrates of the CS–US association and the CS–nothing association for latent inhibition of cued fear conditioning, it is clear that while there is some overlap in underlying neural processes, differences exist. The most striking differences are in the neural areas that are critically involved; specifically the basolateral amygdala, the entorhinal cortex, and the nucleus accumbens. The basolateral nucleus of the amygdala is thought to be a critical site involved in the formation of the CS–US association (Farb et al., 1995; Li et al., 1996; Phillips & LeDoux, 1992) but lesions of the basolateral nucleus did not disrupt latent inhibition of cued fear conditioning (Weiner et al., 1996b). Thus, the CS–US association forms in an area not critically involved in latent inhibition. The converse is also true. Inhibition of the entorhinal cortex at CS pre-exposure is sufficient to block latent inhibition (Lewis & Gould, 2007a). Because it must be assumed that the CS–nothing association occurs during CS pre-exposure, the entorhinal cortex must be involved in the formation of the CS–nothing association. The entorhinal cortex, however, is not involved in the formation of the CS–US association during fear conditioning as inhibition of the entorhinal cortex during conditioning did not block learning (Coutureau et al., 1999; Lewis & Gould, 2007a). In addition, the nucleus accumbens plays an important role in latent inhibition but lesions of the nucleus accumbens did not block fear conditioning (Gal et al., 2005; Levita et al., 2002; Riedel et al., 1997). It should not be too surprising that the neural areas involved in the formation of the CS–US association and the CS–nothing association differ because behavioral studies show that CS pre-exposure during latent inhibition does not block the learning of the CS–US association but instead expression of the CS–US association (Bakner et al., 1991; Bouton, 1993; Kasprow et al., 1984). Thus, if the CS–nothing and the CS–US association are separate processes that are both acquired during latent inhibition, they cannot be mediated by identical neural substrates. Even if they involve separate neural areas, they may engage similar cellular processes. Both fear conditioning and latent inhibition of cued fear conditioning appear to depend on the activation of NMDA receptors, PKA, and ERK (Abel et al., 1997; Atkins, Selcher, Petraitis et al., 1998; Davis & Gould, 2005; Gould & Lewis, 2005; Kim, Fanselow, DeCola, & Landeira-Fernandez, 1992; Lewis & Gould, 2007b; Lewis et al., 2004). However, the location of the cell signaling events differs, with the amygdala critically
264
Thomas J. Gould
involved in the CS–US association and the entorhinal cortex critically involved in the CS–nothing association. While the CS–US association and the CS–nothing association may involve common cell signaling molecules, there may be differences in how the CS–US association versus the CS–nothing association are stored. Long-term memory for cued fear conditioning has repeatedly been shown to depend on protein synthesis (for review see Hernandez & Abel, 2008). Latent inhibition of cued fear conditioning occurs in the presence of the protein synthesis inhibitor anisomycin. As reviewed, administration of anisomycin before or after CS pre-exposure had no effect on latent inhibition (Lewis & Gould, 2004). Thus, the mechanism of synaptic plasticity that underlies the storage of the CS–nothing association remains to be elucidated but appears to be different than that involved in the CS–US association. A model of how activation of the neural substrates of the CS–nothing association results in latent inhibition of cued fear conditioning can be put forward. The entorhinal cortex is intimately involved in the formation of the CS–nothing association during CS pre-exposure and the expression of CS–nothing association at testing, but not in the formation of the CS–US association (Lewis & Gould, 2007a). The question remains as to how the entorhinal cortex regulates latent inhibition. One potential mechanism for latent inhibition-associated control over the conditioned response expression may be entorhinal cortical projections to the amygdala. The entorhinal cortex projects to the basolateral amygdala, central nucleus of the amygdala, and also to the intercalated cell masses (ITC) that are located between the basolateral and central nuclei (McDonald & Mascagni, 1997). Royer, Martina, & Pare (1999) have shown that ITC cells are GABAergic interneurons that are well suited to control the flow of information between the basolateral nucleus (known to be a critical site of plasticity of cued fear conditioning; reviewed in Maren, 2005) and the central nucleus (known to be involved in the expression of fear-related behaviors, LeDoux et al., 1988). During pre-exposure to the CS, repeated presentation of the CS may produce a NMDA receptor/PKA/ERK-mediated change in plasticity within the entorhinal cortex that results in greater activation of the entorhinal cortex and associated efferents at subsequent presentation of the CS. This could lead to the entorhinal cortex having great control over amygdala output at testing when the CS is activated. Modulation of basolateral amygdala output to the central nucleus of the amygdala via changes in ITC cell activity would modulate the expression of fear while having little effect on neural activity in the basolateral amygdala associated with the acquisition of the CS–US association. Another potential mechanism for the entorhinal cortical control of latent inhibition is through mediating changes in nucleus accumbal dopaminergic transmission. It is well established that increased dopamine efflux within the nucleus accumbens disrupts latent inhibition of a variety of tasks (reviewed in Weiner, 2003). In support of the entorhinal cortex modulating accumbal dopamine levels, reversible inactivation of the left entorhinal cortex disrupted latent inhibition of conditioned odor
A comparison of mechanisms underlying the CS–US association
265
aversion and increased dopamine levels within the core, but not the dorsomedial shell, of the nucleus accumbens (Jeanblanc, Peterschmitt, Hoeltzel, & Louilot, 2004). However, in another study, bilateral lesions of the entorhinal cortex disrupted latent inhibition of cued fear conditioning without producing any changes in basal levels or amphetamine-induced release of dopamine within the nucleus accumbens (Coutureau et al., 2002). Thus, it remains unclear whether the entorhinal cortex regulates dopamine in the nucleus accumbens during latent inhibition. Both the entorhinal cortex and the nucleus accumbens are critically involved in latent inhibition but they may mediate different phases of latent inhibition. The entorhinal cortex may be involved in development of the CS–nothing association that occurs at CS pre-exposure while changes in dopamine levels in the nucleus accumbens may regulate the future expression of the CS–US or CS–nothing association (see Weiner, 2003, for discussion). Further work is needed to understand the relationship between nucleus accumbens-mediated processes and entorhinal corticalmediated processes during latent inhibition.
Remaining questions Identifying the neural substrates of latent inhibition and how changes in those substrates alter behavior will advance understanding of the mind and diseases that alter the mind. This area of research, however, is only in its infancy as it is still necessary to identify neural and molecular events that occur with each phase of latent inhibition. It appears that the entorhinal cortex is critically involved in events that take place at CS pre-exposure such as the CS–nothing association. This plasticity, at least for latent inhibition of cued fear conditioning, involves signaling through NMDA receptors and activation of PKA and ERK. The changes that occur in the entorhinal cortex during CS pre-exposure could allow the entorhinal cortex to exert inhibitory control over the amygdala on testing day and thus block the expression of cued fear conditioning. Whereas this model works for latent inhibition of cued fear conditioning, it remains to be determined whether a similar model would apply to other types of latent inhibition. That is, is there a common neural substrate of the CS–nothing association that underlies all types of latent inhibition? This may be the case. The entorhinal cortex is involved in many different types of latent inhibition as lesions of this area disrupt latent inhibition of cued fear conditioning, of appetitive learning, and of eyeblink conditioning (Coutureau et al., 1999; Coutureau et al., 2002; Lewis & Gould, 2007a; Oswald et al., 2002; Shohamy, Allen, & Gluck, 2000). While it is easy to envision how entorhinal cortical control of amygdala output could mediate latent inhibition of cued fear conditioning, it is more difficult to understand how entorhinal cortical–amygdala interactions could mediate latent inhibition of eyeblink conditioning, a task not typically associated with the amygdala. The cerebellum is the critical site of learning during eyeblink conditioning (McCormick & Thompson, 1984);
266
Thomas J. Gould
therefore, if the entorhinal cortex is a universal site for the CS–nothing association, then the entorhinal cortex must be able to exert inhibitory control over cerebellar output during expression of the CR. The entorhinal cortex could modulate cerebellar output directly or modulate output through an intermediary brain region. Interestingly, while the amygdala is not a critical structure for eyeblink conditioning, it has been shown modulate eyeblink conditioning (Blankenship, Huckfeldt, Steinmetz, & Steinmetz, 2005; Lee & Kim, 2004) and therefore entorhinal cortex–amygdala interactions could be involved in latent inhibition of eyeblink conditioning. Thus, an important direction for future research on the plasticity underlying latent inhibition is whether the same neural areas and cellsignaling events are universally involved in all types of latent inhibition.
References Abel, T., & Kandel, E. (1998). Positive and negative regulatory mechanisms that mediate long-term memory storage. Brain Research. Brain Research Reviews, 26, 360–378. Abel, T., Nguyen, P. V., Barad, M., et al. (1997). Genetic demonstration of a role for PKA in the late phase of LTP and in hippocampus-based long-term memory. Cell, 88, 615–626. Ackil, J. K., Carman, H. M., Bakner, L., & Riccio, D. C. (1992). Reinstatement of latent inhibition following a reminder treatment in a conditioned taste aversion paradigm. Behavioral and Neural Biology, 58, 232–235. Adams, J. P., & Sweatt, J. D. (2002). Molecular psychology: roles for the ERK MAP kinase cascade in memory. Annual Review of Pharmacology and Toxicology, 42, 135–163. Ahi, J., Radulovic, J., & Spiess, J. (2004). The role of hippocampal signaling cascades in consolidation of fear memory. Behavioral Brain Research, 149, 17–31. Amieva, H., Phillips, L. H., Della, S. S., & Henry, J. D. (2004). Inhibitory functioning in Alzheimer’s disease. Brain, 127, 949–964. Ammassari-Teule, M., Passino, E., Restivo, L., & De Marsanich, B. (2000). Fear conditioning in C57/BL/6 and DBA/2 mice: variability in nucleus accumbens function according to the strain predisposition to show contextual- or cue-based responding. European Journal of Neuroscience, 12, 4467–4474. Atkins, C. M., Selcher, J. C., Petraitis, J. J., Trzaskos, J. M., & Sweatt, J. D. (1998). The MAPK cascade is required for mammalian associative learning. Nature Neuroscience, 1, 602–609. Bakner, L., Strohan, K., Nordeen, M., & Riccio, D. C. (1991). Postconditioning recovery from the latent inhibition effect in conditioned taste aversion. Physiology & Behavior, 50, 1269–1272. Barad, M., Blouin, A. M., & Cain, C. K. (2004). Like extinction, latent inhibition of conditioned fear in mice is blocked by systemic inhibition of L-type voltage-gated calcium channels. Learning & Memory, 11, 536–539. Baron-Cohen, S., & Belmonte, M. K. (2005). Autism: a window onto the development of the social and the analytic brain. Annual Review of Neuroscience, 28, 109–126.
A comparison of mechanisms underlying the CS–US association
267
Baruch, I., Hemsley, D. R., & Gray, J. A. (1988). Differential performance of acute and chronic schizophrenics in a latent inhibition task. Journal of Nervous and Mental Disease, 176, 598–606. Baudry, M., & Lynch, G. (2001). Remembrance of arguments past: how well is the glutamate receptor hypothesis of LTP holding up after 20 years? Neurobiology of Learning and Memory, 76, 284–297. Berman, D. E., Hazvi, S., Neduva, V., & Dudai, Y. (2000). The role of identified neurotransmitter systems in the response of insular cortex to unfamiliar taste: activation of ERK1–2 and formation of a memory trace. Journal of Neuroscience, 20, 7017–7023. Blankenship, M. R., Huckfeldt, R., Steinmetz, J. J., & Steinmetz, J. E. (2005). The effects of amygdala lesions on hippocampal activity and classical eyeblink conditioning in rats. Brain Research, 1035, 120–130. Blum, S., Moore, A. N., Adams, F., & Dash, P. K. (1999). A mitogen-activated protein kinase cascade in the CA1/CA2 subfield of the dorsal hippocampus is essential for long-term spatial memory. Journal of Neuroscience, 19, 3535–3544. Bourtchuladze, R., Frenguelli, B., Blendy, J., et al. (1994). Deficient long-term memory in mice with a targeted mutation of the cAMP-responsive element-binding protein. Cell, 79, 59–68. Bouton, M. E. (1993). Context, time, and memory retrieval in the interference paradigms of Pavlovian learning. Psychological Bulletin, 114, 80–99. Bouton, M. E. (2004). Context and behavioral processes in extinction. Learning & Memory, 11, 485–494. Bucci, D. J., Phillips, R. G., & Burwell, R. D. (2000). Contributions of postrhinal and perirhinal cortex to contextual information processing. Behavioral Neuroscience, 114, 882–894. Caldarone, B. J., Duman, C. H., & Picciotto, M. R. (2000). Fear conditioning and latent inhibition in mice lacking the high affinity subclass of nicotinic acetylcholine receptors in the brain. Neuropharmacology, 39, 2779–2784. Chen, R. H., Sarnecki, C., & Blenis, J. (1992). Nuclear localization and regulation of erk- and rsk-encoded protein kinases. Molecular and Cellular Biology, 12, 915–927. Chetkovich, D. M., Gray, R., Johnston, D., & Sweatt, J. D. (1991). N-methyl-Daspartate receptor activation increases cAMP levels and voltage-gated Ca2þ channel activity in area CA1 of hippocampus. Proceedings of the National Academy of Sciences USA, 88, 6467–6471. Clark, A. J., Feldon, J., & Rawlins, J. N. (1992). Aspiration lesions of rat ventral hippocampus disinhibit responding in conditioned suppression or extinction, but spare latent inhibition and the partial reinforcement extinction effect. Neuroscience, 48, 821–829. Clark, L., & Goodwin, G. M. (2004). State- and trait-related deficits in sustained attention in bipolar disorder. European Archives of Psychiatry and Clinical Neuroscience, 254, 61–68. Corcoran, K. A., & Quirk, G. J. (2007). Activity in prelimbic cortex is necessary for the expression of learned, but not innate, fears. Journal of Neuroscience, 27, 840–844. Coutureau, E., Galani, R., Gosselin, O., Majchrzak, M., & Di Scala, G. (1999). Entorhinal but not hippocampal or subicular lesions disrupt latent inhibition in rats. Neurobiology of Learning and Memory, 72, 143–157.
268
Thomas J. Gould
Coutureau, E., Lena, I., Dauge, V., & Di, S. G. (2002). The entorhinal cortex-nucleus accumbens pathway and latent inhibition: a behavioral and neurochemical study in rats. Behavioral Neuroscience, 116, 95–104. Crowell, C. R., & Anderson, D. C. (1972). Variations in intensity, interstimulus interval, and interval between preconditioning CS exposures and conditioning with rats. Journal of Comparative and Physiological Psychology, 79, 291–298. Curtin, J. F., & Cotter, T. G. (2002). Anisomycin activates JNK and sensitises DU 145 prostate carcinoma cells to Fas mediated apoptosis. British Journal of Cancer, 87, 1188–1194. Davis, J. A., & Gould, T. J. (2005). Rolipram attenuates MK-801-induced deficits in latent inhibition. Behavioral Neuroscience, 119, 595–602. Davis, S., Vanhoutte, P., Pages, C., Caboche, J., & Laroche, S. (2000). The MAPK/ ERK cascade targets both Elk-1 and cAMP response element-binding protein to control long-term potentiation-dependent gene expression in the dentate gyrus in vivo. Journal of Neuroscience, 20, 4563–4572. De la Casa, L. G., Diaz, E., & Lubow, R. E. (2003). Effects of post-treatment retention interval and context on neophobia and conditioned taste aversion. Behavioural Processes, 63, 159–170. Ellison, G. (1995). The N-methyl-D-aspartate antagonists phencyclidine, ketamine and dizocilpine as both behavioral and anatomical models of the dementias. Brain Research. Brain Research Reviews, 20, 250–267. Evrard, A., Laporte, A. M., Chastanet, M., et al. (1999). 5-HT1A and 5-HT1B receptors control the firing of serotoninergic neurons in the dorsal raphe nucleus of the mouse: studies in 5-HT1B knock-out mice. European Journal of Neuroscience, 11, 3823–3831. Farb, C. R., Aoki, C., & Ledoux, J. E. (1995). Differential localization of NMDA and AMPA receptor subunits in the lateral and basal nuclei of the amygdala: a light and electron microscopic study. Journal of Comparative Neurology, 362, 86–108. Feldon, J., & Weiner, I. (1991). The latent inhibition model of schizophrenic attention disorder. Haloperidol and sulpiride enhance rats’ ability to ignore irrelevant stimuli. Biological Psychiatry, 29, 635–646. Gaisler-Salomon, I., & Weiner, I. (2003). Systemic administration of MK-801 produces an abnormally persistent latent inhibition which is reversed by clozapine but not haloperidol. Psychopharmacology (Berl.), 166, 333–342. Gal, G., Schiller, D., & Weiner, I. (2005). Latent inhibition is disrupted by nucleus accumbens shell lesion but is abnormally persistent following entire nucleus accumbens lesion: the neural site controlling the expression and disruption of the stimulus preexposure effect. Behavioural Brain Research, 162, 246–255. Goosens, K. A., & Maren, S. (2003). Pretraining NMDA receptor blockade in the basolateral complex, but not the central nucleus, of the amygdala prevents savings of conditional fear. Behavioral Neuroscience, 117, 738–750. Gould, T. J., Collins, A. C., & Wehner, J. M. (2001). Nicotine enhances latent inhibition and ameliorates ethanol-induced deficits in latent inhibition. Nicotine & Tobacco Research, 3, 17–24. Gould, T. J., & Lewis, M. C. (2005). Coantagonism of glutamate receptors and nicotinic acetylcholinergic receptors disrupts fear conditioning and latent inhibition of fear conditioning. Learning & Memory, 12, 389–398.
A comparison of mechanisms underlying the CS–US association
269
Gould, T. J., & Wehner, J. M. (1999). Genetic influences on latent inhibition in mice. Behavioral Neuroscience, 113, 1291–1296. Han, J. S., Gallagher, M., & Holland, P. (1995). Hippocampal lesions disrupt decrements but not increments in conditioned stimulus processing. Journal of Neuroscience, 15, 7323–7329. Hashimoto, K., Iyo, M., Freedman, R., & Stevens, K. E. (2005). Tropisetron improves deficient inhibitory auditory processing in DBA/2 mice: role of alpha 7 nicotinic acetylcholine receptors. Psychopharmacology (Berl.), 183, 13–19. Hernandez, P. J., & Abel, T. (2008). The role of protein synthesis in memory consolidation: progress amid decades of debate. Neurobiology of Learning and Memory, 89, 293–311. Hitchcock, J. M., Lister, S., Fischer, T. R., & Wettstein, J. G. (1997). Disruption of latent inhibition in the rat by the 5-HT2 agonist DOI: effects of MDL 100,907, clozapine, risperidone and haloperidol. Behavioural Brain Research, 88, 43–49. Holt, W., & Maren, S. (1999). Muscimol inactivation of the dorsal hippocampus impairs contextual retrieval of fear memory. Journal of Neuroscience, 19, 9054–9062. Honey, R. C., & Good, M. (1993). Selective hippocampal lesions abolish the contextual specificity of latent inhibition and conditioning. Behavioral Neuroscience, 107, 23–33. Jeanblanc, J., Peterschmitt, Y., Hoeltzel, A., & Louilot, A. (2004). Influence of the entorhinal cortex on accumbal and striatal dopaminergic responses in a latent inhibition paradigm. Neuroscience, 128, 187–200. Joel, D., Weiner, I., & Feldon, J. (1997). Electrolytic lesions of the medial prefrontal cortex in rats disrupt performance on an analog of the Wisconsin Card Sorting Test, but do not disrupt latent inhibition: implications for animal models of schizophrenia. Behavioural Brain Research, 85, 187–201. Joseph, M. H., Peters, S. L., & Gray, J. A. (1993). Nicotine blocks latent inhibition in rats – evidence for a critical role of increased functional activity of dopamine in the mesolimbic system at conditioning rather than preexposure. Psychopharmacology, 110, 187–192. Kaplan, O., Dar, R., Rosenthal, L., et al. (2006). Obsessive-compulsive disorder patients display enhanced latent inhibition on a visual search task. Behavior Research and Therapy, 44, 1137–1145. Kasprow, W. J., Catterson, D., Schachtman, T. R., & Miller, R. R. (1984). Attenuation of latent inhibition by post-acquisition reminder. Quarterly Journal of Experimental Psychology: Comparative & Physiological Psychology, 36b, 53–63. Kaye, H., & Pearce, J. M. (1987). Hippocampal lesions attenuate latent inhibition of CS and a neutral stimulus. Psychobiology, 15, 293–299. Killcross, A. S., Stanhope, K. J., Dourish, C. T., & Piras, G. (1997). WAY100635 and latent inhibition in the rat: selective effects at preexposure. Behavioural Brain Research, 88, 51–57. Kim, J. J., Fanselow, M. S., DeCola, J. P., & Landeira-Fernandez, J. (1992). Selective impairment of long-term but not short-term conditional fear by the N-methylD-aspartate antagonist APV. Behavioral Neuroscience, 106, 591–596. Kim, J. J., & Jung, M. W. (2006). Neural circuits and mechanisms involved in Pavlovian fear conditioning: a critical review. Neuroscience & Biobehavioral Reviews, 30, 188–202.
270
Thomas J. Gould
Kim, J. J., Rison, R. A., & Fanselow, M. S. (1993). Effects of amygdala, hippocampus, and periaqueductal gray lesions on short- and long-term contextual fear. Behavioral Neuroscience, 107, 1093–1098. Lacroix, L., Broersen, L. M., Weiner, I., & Feldon, J. (1998). The effects of excitotoxic lesion of the medial prefrontal cortex on latent inhibition, prepulse inhibition, food hoarding, elevated plus maze, active avoidance and locomotor activity in the rat. Neuroscience, 84, 431–442. Lacroix, L., Spinelli, S., White, W., & Feldon, J. (2000). The effects of ibotenic acid lesions of the medial and lateral prefrontal cortex on latent inhibition, prepulse inhibition and amphetamine-induced hyperlocomotion. Neuroscience, 97, 459–468. LeDoux, J. E., Iwata, J., Cicchetti, P., & Reis, D. J. (1988). Different projections of the central amygdaloid nucleus mediate autonomic and behavioral correlates of conditioned fear. Journal of Neuroscience, 8, 2517–2529. LeDoux, J. E., Iwata, J., Pearl, D., & Reis, D. J. (1986). Disruption of auditory but not visual learning by destruction of intrinsic neurons in the rat medial geniculate body. Brain Research, 371, 395–399. Lee, H., & Kim, J. J. (1998). Amygdalar NMDA receptors are critical for new fear learning in previously fear-conditioned rats. Journal of Neuroscience, 18, 8444–8454. Lee, T., & Kim, J. J. (2004). Differential effects of cerebellar, amygdalar, and hippocampal lesions on classical eyeblink conditioning in rats. Journal of Neuroscience, 24, 3242–3250. Levita, L., Dalley, J. W., & Robbins, T. W. (2002). Disruption of Pavlovian contextual conditioning by excitotoxic lesions of the nucleus accumbens core. Behavioral Neuroscience, 116, 539–552. Lewis, M. C., Davis, J. A., & Gould, T. J. (2004). Inhibition of mitogen-activated protein kinase/extracellular signal-regulated kinase disrupts latent inhibition of cued fear conditioning in C57BL/6 mice. Behavioral Neuroscience, 118, 1444–1449. Lewis, M. C., & Gould, T. J. (2004). Latent inhibition of cued fear conditioning: an NMDA receptor-dependent process that can be established in the presence of anisomycin. European Journal of Neuroscience, 20, 818–826. Lewis, M. C., & Gould, T. J. (2007a). Reversible inactivation of the entorhinal cortex disrupts the establishment and expression of latent inhibition of cued fear conditioning in C57BL/6 mice. Hippocampus, 17, 462–470. Lewis, M. C., & Gould, T. J. (2007b). Signal transduction mechanisms within the entorhinal cortex that support latent inhibition of cued fear conditioning. Neurobiology of Learning and Memory, 88, 359–368. Li, X. F., Stutzmann, G. E., & Ledoux, J. E. (1996). Convergent but temporally separated inputs to lateral amygdala neurons from the auditory thalamus and auditory cortex use different postsynaptic receptors: in vivo intracellular and extracellular recordings in fear conditioning pathways. Learning & Memory, 3, 229–242. Lipina, T., Labrie, V., Weiner, I., & Roder, J. (2005). Modulators of the glycine site on NMDA receptors, D-serine and ALX 5407, display similar beneficial effects to clozapine in mouse models of schizophrenia. Psychopharmacology (Berl.), 179, 54–67. Lisman, J. (2003). Long-term potentiation: outstanding questions and attempted synthesis. Philosophical Transactions of the Royal Society B: Biological Sciences, 358, 829–842.
A comparison of mechanisms underlying the CS–US association
271
Logue, S. F., Paylor, R., & Wehner, J. M. (1997). Hippocampal lesions cause learning deficits in inbred mice in the Morris water maze and conditioned-fear task. Behavioral Neuroscience, 111, 104–113. Lorden, J. F., Rickert, E. J., & Berry, D. W. (1983). Forebrain monoamines and associative learning: I. Latent inhibition and conditioned inhibition. Behavioural Brain Research, 9, 181–199. Loskutova, L. V. (2001). The effects of a serotoninergic substrate of the nucleus accumbens on latent inhibition. Neuroscience and Behavioral Physiology, 31, 15–20. Loskutova, L. V., Luk’yanenko, F. Y., & Il’yuchenok, R. Y. (1990). Interaction of serotonin- and dopaminergic systems of the brain in mechanisms of latent inhibition in rats. Neuroscience and Behavioral Physiology, 20, 500–505. Lubow, R. E. (1973). Latent inhibition. Psychological Bulletin, 79, 398–407. Lubow, R. E. (1989). Latent Inhibition and Conditional Attention Theory. Cambridge: Cambridge University Press. Lubow, R. E., & Gewirtz, J. C. (1995). Latent inhibition in humans: data, theory, and implications for schizophrenia. Psychological Bulletin, 117, 87–103. Lubow, R. E., & Josman, Z. E. (1993). Latent inhibition deficits in hyperactive children. Journal of Child Psychology and Psychiatry, 34, 959–973. Lynch, G., & Baudry, M. (1984). The biochemistry of memory: a new and specific hypothesis. Science, 224, 1057–1063. Maren, S. (2005). Synaptic mechanisms of associative memory in the amygdala. Neuron, 47, 783–786. Maren, S., Ferrario, C. R., Corcoran, K. A., Desmond, T. J., & Frey, K. A. (2003). Protein synthesis in the amygdala, but not the auditory thalamus, is required for consolidation of Pavlovian fear conditioning in rats. European Journal of Neuroscience, 18, 3080–3088. McCormick, D. A., & Thompson, R. F. (1984). Cerebellum: Essential involvement in the classically conditioned eyelid response. Science, 223, 296–299. McDonald, A. J., & Mascagni, F. (1997). Projections of the lateral entorhinal cortex to the amygdala: a Phaseolus vulgaris leucoagglutinin study in the rat. Neuroscience, 77, 445–459. McDonald, L. M., Moran, P. M., Vythelingum, G. N., et al. (2003). Enhancement of latent inhibition by two 5-HT2A receptor antagonists only when given at both pre-exposure and conditioning. Psychopharmacology (Berl.), 169, 321–331. Mcintosh, S. M., & Tarpy, R. M. (1977). Retention of latent inhibition in a taste-aversion paradigm. Bulletin of the Psychonomic Society, 9, 411–412. Micheau, J., & Riedel, G. (1999). Protein kinases: which one is the memory molecule? Cellular and Molecular Life Sciences, 55, 534–548. Milad, M. R., Quirk, G. J., Pitman, R. K., et al. (2007). A role for the human dorsal anterior cingulate cortex in fear expression. Biological Psychiatry, 62, 1191–1194. Murphy, C. A., Pezze, M., Feldon, J., & Heidbreder, C. (2000). Differential involvement of dopamine in the shell and core of the nucleus accumbens in the expression of latent inhibition to an aversively conditioned stimulus. Neuroscience, 97, 469–477. Nicholson, D. A., & Freeman, J. H. Jr. (2002). Medial dorsal thalamic lesions impair blocking and latent inhibition of the conditioned eyeblink response in rats. Behavioral Neuroscience, 116, 276–285.
272
Thomas J. Gould
Oswald, C. J., Yee, B. K., Rawlins, J. N., et al. (2002). The influence of selective lesions to components of the hippocampal system on the orientating response, habituation and latent inhibition. European Journal of Neuroscience, 15, 1983–1990. Perkinton, M. S., Sihra, T. S., & Williams, R. J. (1999). Ca(2þ)-permeable AMPA receptors induce phosphorylation of cAMP response element-binding protein through a phosphatidylinositol 3-kinase-dependent stimulation of the mitogen-activated protein kinase signaling cascade in neurons. Journal of Neuroscience, 19, 5861–5874. Phillips, R. G., & Ledoux, J. E. (1992). Differential contribution of amygdala and hippocampus to cued and contextual fear conditioning. Behavioral Neuroscience, 106, 274–285. Pouzet, B., Zhang, W. N., Weiner, I., Feldon, J., & Yee, B. K. (2004). Latent inhibition is spared by n-methyl–aspartate (NMDA)-induced ventral hippocampal lesions, but is attenuated following local activation of the ventral hippocampus by intracerebral NMDA infusion. Neuroscience, 124, 183–194. Quinlan, E. M., & Halpain, S. (1996). Emergence of activity-dependent, bidirectional control of microtubule-associated protein MAP2 phosphorylation during postnatal development. Journal of Neuroscience, 16, 7627–7637. Restivo, L., Passino, E., Middei, S., & Ammassari-Teule, M. (2002). The strainspecific involvement of nucleus accumbens in latent inhibition might depend on differences in processing configural- and cue-based information between C57BL/6 and DBA mice. Brain Research Bulletin, 57, 35–39. Riedel, G., Harrington, N. R., Hall, G., & Macphail, E. M. (1997). Nucleus accumbens lesions impair context, but not cue, conditioning in rats. Neuroreport, 8, 2477–2481. Riedel, G., Platt, B., & Micheau, J. (2003). Glutamate receptor function in learning and memory. Behavioural Brain Research, 140, 1–47. Roberson, E. D., English, J. D., Adams, J. P., et al. (1999). The mitogen-activated protein kinase cascade couples PKA and PKC to cAMP response element binding protein phosphorylation in area CA1 of hippocampus. Journal of Neuroscience, 19, 4337–4348. Rochford, J., Sen, A. P., & Quirion, R. (1996). Effect of nicotine and nicotinic receptor agonists on latent inhibition in the rat. Journal of Pharmacology and Experimental Therapeutics, 277, 1267–1275. Rochford, J., Sen, A. P., Rousse, I., & Welner, S. A. (1996). The effect of quisqualic acid-induced lesions of the nucleus basalis magnocellularis on latent inhibition. Brain Research Bulletin, 41, 313–317. Rodrigues, S. M., Farb, C. R., Bauer, E. P., LeDoux, J. E., & Schafe, G. E. (2004). Pavlovian fear conditioning regulates Thr286 autophosphorylation of Ca2þ/ calmodulin-dependent protein kinase II at lateral amygdala synapses. Journal of Neuroscience, 24, 3281–3288. Rodrigues, S. M., Schafe, G. E., & LeDoux, J. E. (2004). Molecular mechanisms underlying emotional learning and memory in the lateral amygdala. Neuron, 44, 75–91. Rogan, M. T., & LeDoux, J. E. (1996). Emotion: systems, cells, synaptic plasticity. Cell, 85, 469–475. Royer, S., Martina, M., & Pare, D. (1999). An inhibitory interface gates impulse traffic between the input and output stations of the amygdala. Journal of Neuroscience, 19, 10575–10583.
A comparison of mechanisms underlying the CS–US association
273
Ruob, C., Weiner, I., & Feldon, J. (1998). Haloperidol-induced potentiation of latent inhibition: interaction with parameters of conditioning. Behavioural Pharmacology, 9, 245–253. Sacchetti, B., Baldi, E., Lorenzini, C. A., & Bucherelli, C. (2002a). From the cover: cerebellar role in fear-conditioning consolidation. Proceedings of the National Academy of Sciences USA, 99, 8406–8411. Sacchetti, B., Baldi, E., Lorenzini, C., & Bucherelli, C. (2002b). Differential contribution of some cortical sites to the formation of memory traces supporting fear conditioning. Experimental Brain Research, 146, 223–232. Sacchetti, B., Scelfo, B., Tempia, F., & Strata, P. (2004). Long-term synaptic changes induced in the cerebellar cortex by fear conditioning. Neuron, 42, 973–982. Schafe, G. E., Atkins, C. M., Swank, M. W., et al. (2000). Activation of ERK/MAP kinase in the amygdala is required for memory consolidation of pavlovian fear conditioning. Journal of Neuroscience, 20, 8177–8187. Schafe, G. E., & LeDoux, J. E. (2000). Memory consolidation of auditory pavlovian fear conditioning requires protein synthesis and protein kinase A in the amygdala. Journal of Neuroscience, 20, RC96. Schauz, C., & Koch, M. (1999). Lesions of the nucleus basalis magnocellularis do not impair prepulse inhibition and latent inhibition of fear-potentiated startle in the rat. Brain Research, 815, 98–105. Schauz, C., & Koch, M. (2000). Blockade of NMDA receptors in the amygdala prevents latent inhibition of fear-conditioning. Learning & Memory, 7, 393–399. Schiller, D., & Weiner, I. (2004). Lesions to the basolateral amygdala and the orbitofrontal cortex but not to the medial prefrontal cortex produce an abnormally persistent latent inhibition in rats. Neuroscience, 128, 15–25. Schiller, D., & Weiner, I. (2005). Basolateral amygdala lesions in the rat produce an abnormally persistent latent inhibition with weak preexposure but not with context shift. Behavioural Brain Research, 163, 115–121. Schiller, D., Zuckerman, L., & Weiner, I. (2006). Abnormally persistent latent inhibition induced by lesions to the nucleus accumbens core, basolateral amygdala and orbitofrontal cortex is reversed by clozapine but not by haloperidol. Journal of Psychiatric Research, 40, 167–177. Schmajuk, N. A., Lam, Y. W., & Christiansen, B. A. (1994). Latent inhibition of the rat eyeblink response: effect of hippocampal aspiration lesions. Physiology & Behavior, 55, 597–601. Selcher, J. C., Atkins, C. M., Trzaskos, J. M., Paylor, R., & Sweatt, J. D. (1999). A necessity for MAP kinase activation in mammalian spatial learning. Learning & Memory, 6, 478–490. Selcher, J. C., Weeber, E. J., Varga, A. W., Sweatt, J. D., & Swank, M. (2002). Protein kinase signal transduction cascades in mammalian associative conditioning. Neuroscientist, 8, 122–131. Shohamy, D., Allen, M. T., & Gluck, M. A. (2000). Dissociating entorhinal and hippocampal involvement in latent inhibition. Behavioral Neuroscience, 114, 867–874. Solomon, P. R., Kiney, C. A., & Scott, D. R. (1978). Disruption of latent inhibition following systemic administration of parachlorophenylalanine (PCPA). Physiology & Behavior, 20, 265–271. Solomon, P. R., & Moore, J. W. (1975). Latent inhibition and stimulus generalization of the classically conditioned nictitating membrane response in
274
Thomas J. Gould
rabbits (Oryctolagus cuniculus) following dorsal hippocampal ablation. Journal of Comparative and Physiological Psychology, 89, 1192–1203. Solomon, P. R., Nichols, G. L., Kiernan, J. M., III, Kamer, R. S., & Kaplan, L. J. (1980). Differential effects of lesions in medial and dorsal raphe of the rat: latent inhibition and septohippocampal serotonin levels. Journal of Comparative and Physiological Psychology, 94, 145–154. Solomon, P. R., & Staton, D. M. (1982). Differential effects of microinjections of d-amphetamine into the nucleus accumbens or the caudate putamen on the rat’s ability to ignore an irrelevant stimulus. Biological Psychiatry, 17, 743–756. Szapiro, G., Vianna, M. R., McGaugh, J. L., Medina, J. H., & Izquierdo, I. (2003). The role of NMDA glutamate receptors, PKA, MAPK, and CAMKII in the hippocampus in extinction of conditioned fear. Hippocampus, 13, 53–58. Tai, C. T., Cassaday, H., Feldon, J., & Rawlins, J. N. P. (1995). Both electrolytic and excitotoxic lesions of nucleus accumbens disrupt latent inhibition of learning in rats. Neurobiology of Learning and Memory, 64, 36–48. Talk, A., Stoll, E., & Gabriel, M. (2005). Cingulate cortical coding of contextdependent latent inhibition. Behavioral Neuroscience, 119, 1524–1532. Torocsik, B., & Szeberenyi, J. (2000). Anisomycin uses multiple mechanisms to stimulate mitogen-activated protein kinases and gene expression and to inhibit neuronal differentiation in PC12 phaeochromocytoma cells. European Journal of Neuroscience, 12, 527–532. Trimble, K. M., Bell, R., & King, D. J. (1997). Enhancement of latent inhibition in the rat by the atypical antipsychotic agent remoxipride. Pharmacology, Biochemistry, and Behavior, 56, 809–816. Tsaltas, E., Preston, G. C., Rawlins, J. N., Winocur, G., & Gray, J. A. (1984). Dorsal bundle lesions do not affect latent inhibition of conditioned suppression. Psychopharmacology (Berl.), 84, 549–555. Vaitl, D., Lipp, O. V., Bauer, U., Schuler, G., & Stark, R. (1999). Latent inhibition and schizophrenia. Psychophysiology, 36, S117. Vossler, M. R., Yao, H., York, R. D., et al. (1997). cAMP activates MAP kinase and Elk-1 through a B-Raf- and Rap1-dependent pathway. Cell, 89, 73–82. Waltereit, R., & Weller, M. (2003). Signaling from cAMP/PKA to MAPK and synaptic plasticity. Molecular Neurobiology, 27, 99–106. Walz, R., Rockenbach, I. C., Amaral, O. B., Quevedo, J., & Roesler, R. (1999). MAPK and memory. Trends in Neuroscience, 22, 495. Warburton, E. C., Joseph, M. H., Feldon, J., Weiner, I., & Gray, J. A. (1994). Antagonism of amphetamine-induced disruption of latent inhibition in rats by haloperidol and ondansetron – implications for a possible antipsychotic action of ondansetron. Psychopharmacology, 114, 657–664. Weeber, E. J., Atkins, C. M., Selcher, J. C., et al. (2000). A role for the beta isoform of protein kinase C in fear conditioning. Journal of Neuroscience, 20, 5906–5914. Weiner, I. (1990). Neural substrates of latent inhibition: the switching model. Psychological Bulletin, 108, 442–461. Weiner, I. (2003). The “two-headed” latent inhibition model of schizophrenia: modeling positive and negative symptoms and their treatment. Psychopharmacology (Berl.), 169, 257–297. Weiner, I., & Feldon, J. (1992). Phencyclidine does not disrupt latent inhibition in rats: implications for animal models of schizophrenia. Pharmacology, Biochemistry, and Behavior, 42, 625–631.
A comparison of mechanisms underlying the CS–US association
275
Weiner, I., Schiller, D., & Gaisler-Salomon, I. (2003). Disruption and potentiation of latent inhibition by risperidone: the latent inhibition model of atypical antipsychotic action. Neuropsychopharmacology, 28, 499–509. Weiner, I., Shadach, E., Tarrasch, R., Kidron, R., & Feldon, J. (1996a). The latent inhibition model of schizophrenia: further validation using the atypical neuroleptic, clozapine. Biological Psychiatry, 40, 834–843. Weiner, I., Tarrasch, R., & Feldon, J. (1996b). Basolateral amygdala lesions do not disrupt latent inhibition. Behavioural Brain Research, 72, 73–81. Wright, I. K., Garratt, J. C., & Marsden, C. A. (1990). Effects of a selective 5-HT2 agonist, DOI, on 5-HT neuronal firing in the dorsal raphe nucleus and 5-HT release and metabolism in the frontal cortex. British Journal of Pharmacology, 99, 221–222. Yee, B. K., Feldon, J., & Rawlins, J. N. (1997). Cytotoxic lesions of the retrohippocampal region attenuate latent inhibition but spare the partial reinforcement extinction effect. Experimental Brain Research, 115, 247–256.
13 The pharmacology of latent inhibition and its relevance to schizophrenia Ina Weiner and Michal Arad
Latent inhibition refers to the observation that under specific conditions, nonreinforced preexposure (PE) to a stimulus retards the efficacy with which this stimulus is conditioned when paired with reinforcement, compared to a nonpreexposed (NPE) stimulus. The pharmacology of LI has been almost exclusively associated with the use of LI as an animal model of schizophrenia, and therefore largely overlaps the pharmacology of schizophrenia. As detailed in our chapter on neural substrates of LI, the widely held notion that nonreinforced stimulus preexposure reduces attention to or salience of stimuli has served to link LI to attentional processing in schizophrenia. Specifically, because schizophrenia is characterized by an inability to filter out, or ignore, irrelevant or unimportant stimuli, abnormal LI was proposed as a tool for modeling deficient attention in schizophrenia (Solomon, Crider, Winkelman et al., 1981; Weiner, Lubow, & Feldon, 1981, 1984, 1988). Here we will focus on our use of LI for the development of pharmacological animal models related to schizophrenia and the identification of viable antipsychotic/anti-schizophrenia medications. Based on our initial pharmacological data, we adopted in our pharmacological investigations a view of LI that distinguished between the acquisition of LI (learning to ignore the nonreinforced stimulus in preexposure) and the expression of LI (subsequent expression of this learning in conditioning) (Weiner, Feldon, & Katz, 1987; Weiner et al., 1984, 1988). This view of LI has been elaborated in the switching model of LI (Weiner, 1990, 2003; Weiner & Feldon, 1997), and has guided our use of LI for modeling schizophrenia. The switching model was described in our chapter on the neural substrates of LI. Briefly, according to the switching model, LI involves the acquisition of two independent and conflicting associations, one in preexposure (stimulus–no-event) and one in conditioning (stimulus–reinforcement), which compete for behavioral expression during conditioning. Presence of LI indicates that the organism remains under the control of the stimulus–no-event association in spite of the fact that the Latent Inhibition: Cognition, Neuroscience and Applications to Schizophrenia, ed. R. E. Lubow and Ina Weiner. Published by Cambridge University Press # Cambridge University Press 2010.
276
The pharmacology of latent inhibition and its relevance to schizophrenia
277
Figure 13.1. Disruption and persistence of LI.
stimulus is now followed by a significant consequence; in contrast, absence of LI indicates that the organism switches to respond according to the new stimulus– reinforcement contingency. Here we will emphasize several derivatives of the response competition concept for modeling schizophrenia by pharmacological means. First, pharmacological manipulations can prevent or promote the expression of LI, producing two poles of LI abnormality in comparison to non-manipulated controls: disrupted LI under conditions which yield LI in normal rats, or abnormally persistent LI under conditions which prevent LI in normal rats (Figure 13.1). In terms of the switching model, disrupted and persistent LI reflect excessive and retarded switching between stimulus–no-event and stimulus–reinforcement associations, respectively. In more conventional attentional terms, these two poles of abnormality can be seen as a failure to ignore irrelevant stimuli and a failure to dis-ignore irrelevant stimuli when they become relevant, or attentional overswitching and attentional perseveration. A corollary of the two-pole LI abnormality, essential for evaluating the effects of drug treatments in the LI model, is that alleviation of disrupted LI requires that the treatment produces LI, but alleviation of abnormally persistent LI requires that the treatment disrupts LI. Thus, labels of “normal” and “abnormal” cannot be accurately applied to presence or absence of LI. Rather, normal and abnormal LI is defined in comparison to LI status in controls and the manipulation used to alter this status. Second, both disruption and persistence of LI can stem from drug action in the preexposure stage or in the conditioning stage. Preexposure-based LI disruption or LI persistence would reflect impairment or facilitation, respectively, of the acquisition of stimulus–nothing association/inattentional response. Conversely, conditioningbased disruption or persistence of LI would reflect facilitation or retardation, respectively, of switching to respond according to stimulus–reinforcement association. However, drugs administered in conditioning can affect conditioning per se, and this
278
Ina Weiner and Michal Arad
effect can confound or mask their effects on LI expression. The nonpreexposed condition controls precisely for this: if a drug exerts effects on conditioning per se, this would be evident in the NPE condition. It is important that drug-induced disruption of LI (namely, the disappearance of PE–NPE difference in performance) and persistence of LI (namely, the emergence of PE–NPE differences in performance) can be attributed to drug action occurring in the preexposed condition, where processes that are relevant for modeling schizophrenia take place. Ideally, drug doses that affect the performance of the PE condition without affecting performance of the NPE condition should be used. We will return to this point when discussing specific drugs. Our use of LI for the development of pharmacological models of schizophrenia focuses on the three aspects of drug action detailed above: (1) direction of LI perturbation – disrupted or persistent LI; (2) stage of drug action – preexposure or conditioning; (3) condition of preexposure – the preexposed or the nonpreexposed condition. To answer these questions, two conditions of LI are required in non-drugtreated controls, one leading to LI and one to disrupted LI. In most of the experiments conducted in our laboratory and described here, a conditioned emotional procedure is used, in which rats are preexposed to 40 tones followed by two or five tone–shock pairings. Two pairings produce LI, whereas raising the number of conditioning trials to five disrupts LI. Preexposure and conditioning are given 24 h apart and the behavioral index of conditioning is assessed 48 h later in a no-drug state. In some of our experiments a low number of preexposures is used instead of a higher number of conditioning trials. In the following, we describe five distinct LI models based upon the manipulations used to alter the expression of LI, the direction of LI abnormality, and responsiveness to typical or atypical antipsychotic drugs and other schizophrenia-relevant treatments: dopamine (DA) agonist; DA antagonist; NMDA antagonist; muscarinic antagonist; and developmental perturbations.
Treatments that produce abnormal LI Amphetamine-induced LI disruption The notion of a hyperactive dopamine system in schizophrenia is supported by several lines of evidence, most prominently the capacity of the DA releaser amphetamine to induce and exacerbate schizophrenia symptoms, the antipsychotic efficacy of D2 receptor antagonists, and enhanced amphetamine-induced dopamine release in schizophrenia patients (Guillin, Abi-Dargham, & Laruelle, 2007; Toda & Abi-Dargham, 2007). Because amphetamine produces only positive (psychotic) symptoms, amphetamine-induced behavioral abnormalities in animals are considered to model positive symptoms. Solomon et al. (1981) and Weiner et al. (1981, 1984, 1988) were the first to show that LI was lost in rats undergoing preexposure and conditioning under
The pharmacology of latent inhibition and its relevance to schizophrenia
279
amphetamine. The original demonstration of amphetamine-induced LI disruption has been often replicated in rats (Gosselin, Oberling, & Di Scala, 1996; Joseph, Peters, Moran et al., 2000; Killcross, Dickinson, & Robbins, 1994a; Killcross & Robbins, 1993; Moran, Fischer, Hitchcock, & Moser, 1996; Ruob, Elsner, Weiner, & Feldon, 1997; Russig, Kovacevic, Murphy, & Feldon, 2003; Weiner, Bernasconi, Broersen, & Feldon, 1997a; Weiner, Shadach, Tarrasch et al., 1996; Weiner, Tarrasch, Bernasconi et al., 1997c), and more recently in mice (Chang, Meyer, Feldon, & Yee, 2007). Amphetamine-induced LI disruption provides strong support for the validity of the animal LI-schizophrenia model, because: (1) it reproduces the widely documented deficit of schizophrenia, loss of the capacity to ignore irrelevant stimuli; (2) amphetamine-treated normal humans, like amphetamine-treated rats, fail to ignore the preexposed stimulus (Gray, Pickering, Hemsley et al., 1992; Salgado, Hetem, Vidal et al., 2000; Swerdlow, Stephany, Wasserman et al., 2003; Thornton, Dawe, Lee et al., 1996); (3) normal humans scoring high on questionnaires measuring schizotypy show reduced LI relative to subjects with low schizotypy scores (Baruch, Hemsley, & Gray, 1988b; BraunsteinBercovitz & Lubow, 1998; Della Casa, Hofer et al., 1999; Lipp & Vaitl, 1992; Lubow & De la Casa, 2002; Lubow, Ingbergsachs, Zalsteinorda, & Gewirtz, 1992; Lubow, Kaplan, & De la Casa, 2001); (4) the extension of the LI model to the clinic has shown that LI is disrupted in acutely psychotic schizophrenic patients tested within the first weeks of the current episode of illness or being in an acute phase of an otherwise chronic disorder (Baruch, Hemsley, & Gray, 1988a; Gray, Hemsley, & Gray, 1992; Gray, Pilowsky, Gray, & Kerwin, 1995; Lubow, Kaplan, Abramovich et al., 2000; Rascle, Mazas, Vaiva et al., 2001; but see also Swerdlow, Braff, Hartston et al., 1996; Swerdlow, Stephany, Wasserman et al., 2005; Vaitl, Lipp, Bauer et al., 2002; Vaitl & Lipp, 1997); (5) APDs reverse amphetamineinduced LI disruption in rats (Gosselin et al., 1996; Solomon et al., 1981; Warburton, Joseph, Feldon et al., 1994; Weiner et al., 1996) and restore LI in schizophrenia (Baruch et al., 1988a). Taken together, these results have strengthened the likelihood that the LI effect observed in rodents and humans is functionally and pharmacologically the same phenomenon, and support amphetamine-induced disrupted LI as a model of positive symptoms of schizophrenia with face, construct, and predictive validity (Gray, Feldon, Rawlins et al., 1991; Lubow, 2005; Moser, Hitchcock, Lister, & Moran, 2000; Weiner, 2003). Weiner et al. (1984, 1988) showed that rats preexposed under amphetamine (given acutely or following 14 daily injections), but conditioned without the drug, showed intact LI (but see Bethus, Muscat, & Goodall, 2006). This implied that amphetamine did not affect the acquisition of the stimulus–no-event association/inattentional response to the preexposed stimulus, but instead modulated its expression in conditioning, promoting rapid switch of responding according to the stimulus–reinforcement association. Indeed, amphetamine-induced LI disruption was shown to be mediated by this drug’s action at the nucleus accumbens (NAC), a region known to play a key role in cognitive
280
Ina Weiner and Michal Arad
and behavioral switching (Gray et al., 1991; Pennartz, Groenewegen, & Lopes da Silva, 1994; Swerdlow & Koob, 1987; Weiner & Joel, 2002), during the time of conditioning (Gray, Moran, Grigoryan et al., 1997; Joseph et al., 2000). LI disruption by acute amphetamine is restricted to low doses of this drug whereas acute high doses of this drug spare LI (Weiner, Izraeli-Telerant, & Feldon, 1987). However, LI is disrupted following repeated amphetamine administration as well as during withdrawal from repeated amphetamine administration (Murphy, Di Iorio, & Feldon, 2001; Russig, Murphy, & Feldon, 2002; Solomon et al., 1981; Tenn, Fletcher, & Kapur, 2005a; Tenn, Kapur, & Fletcher, 2005b). To date, results with direct DA agonists are inconsistent: while the mixed D1–D2 agonist, apomorphine, as well as selective D1 agonist SKF-38393 and the selective D2 agonist, quinpirole, did not disrupt LI (Broersen, Feldon, & Weiner, 1999; Feldon, Shofel, & Weiner, 1991; Weiner, Shofel, & Feldon, 1990), Masson et al. (Masson, Avanzi, Troncoso, & Brandao, 2003) reported that apomorphine injected prior to conditioning disrupted LI, and Swerdlow et al. (2003) reported that the direct DA agonist bromocriptine disrupted LI in healthy humans. Chagas-Martinich et al. (Chagas-Martinich, Carey, & Carrera, 2007) reported that a direct D2–D3 mixed agonist 7-OH-DPAT disrupted LI at doses acting primarily via D2 receptors but potentiated LI at doses selective for D3 receptors.
MK801- and ketamine-induced LI persistence The DA hypothesis of schizophrenia has been complemented during the last decade by the hypo-NMDA hypothesis, kindled by the observation that non-competitive NMDA receptor (NMDAR) antagonists such as phencyclidine (PCP) and ketamine provoke schizophrenia-like symptoms in human volunteers and exacerbate symptoms in schizophrenia patients, as well as by findings of abnormalities of glutamate neurotransmission in schizophrenia patients (Carlsson, Waters, & Carlsson, 1999; Javitt & Zukin, 1991; Jentsch & Roth, 1999; Tamminga, 1998). NMDA blockade is particularly relevant to negative and cognitive symptomatology. First, unlike amphetamine, which evokes primarily positive symptoms, NMDA antagonists also induce negative symptoms and cognitive impairments characteristic of endogenous schizophrenia (Krystal, D’Souza, Mathalon et al., 2003; Lahti, Koffel, LaPorte, & Tamminga, 1995; Malhotra, Pinals, Adler et al., 1997). Second, findings from clinical trials suggest that conjunctive therapy with compounds that enhance glycine-modulated transmission at the NMDAR, either directly by acting as agonists at the glycine B (GlyB) modulatory site on the NMDAR (such as glycine, D-serine, D-cycloserine (DCS) and D-alanine), or indirectly by inhibiting the glycine transporter (GlyT1; e.g., sarcosine), may reduce negative symptoms and improve cognition in patients with schizophrenia receiving concurrent typical antipsychotics. Given the above, behavioral effects of NMDA antagonists have become increasingly popular
The pharmacology of latent inhibition and its relevance to schizophrenia
281
as a pharmacological model of schizophrenia, and in particular, of negative and/or cognitive symptoms (Moghaddam & Jackson, 2003). Unlike amphetamine, acute systemic administration of low doses of noncompetitive NMDA receptor antagonists, including PCP, ketamine, and the more potent and selective NMDA antagonist MK801, were reported to spare LI (Aguado, San Antonio, Perez et al. 1994; Robinson, Port, & Stillwell, 1993; Tenn et al., 2005b; Turgeon, Auerbach, Duncan-Smith et al., 2000; Turgeon, Auerbach, & Heller, 1998; Weiner & Feldon, 1992). These results have led to the suggestion that NMDA antagonist-induced effects in LI cannot provide a valid model of the disorder (Escobar, Oberling, & Miller, 2002; Powell & Geyer, 2007). Other studies, however, have shown that NMDA antagonists do affect LI, albeit in an opposite manner to that of amphetamine. Gaisler-Salomon and Weiner (2003) showed that a low dose of MK801 was without an observable effect under conditions which led to LI in controls, but that its action became evident under conditions which prevented LI expression in control rats. Under the latter conditions, MK801-treated preexposed rats displayed LI. This same effect was shown also in mice (Lipina, Labrie, Weiner, & Roder, 2005). In other words, MK801-treated animals perseverated in ignoring the preexposed stimulus under conditions in which normal rats switched to treating it as relevant. Importantly, MK801 led to persistent LI by reducing conditioning in the preexposed group without reducing it in the nonpreexposed groups, indicating that LI persistence was not due to impaired conditioning per se. Indeed, because NMDA antagonists are known to impair associative learning (Aguado et al., 1994; HoehnSaric, McLeod, & Glowa, 1991), low doses of MK801 that do not impair conditioning in the nonpreexposed animals are imperative for the manifestation of persistent LI, since poorer conditioning of the preexposed compared to nonpreexposed rats cannot be manifested if the drug reduces conditioning also in the nonpreexposed group. It should be noted, however, that because of this, MK801-induced LI persistence is obtained at a very narrow dose range in both rats and mice (Gaisler-Salomon & Weiner, 2003; Lipina et al., 2005). Conversely, higher doses of MK801, which impair conditioning also in the nonpreexposed groups, result in LI abolition (Gaisler-Salomon & Weiner, 2003). Administration of MK801 confined to the preexposure or conditioning stage revealed that MK801 was ineffective when given in preexposure but led to the emergence of LI when given in conditioning (Gaisler-Salomon & Weiner, 2003), and a similar pattern of results was reported for an acute low dose of PCP (Palsson, Klamer, Wass et al., 2005). The fact that low doses of NMDA receptor antagonists impair performance selectively in the preexposed group, and that this action is exerted in the conditioning stage, when the previously nonreinforced stimulus is followed by reinforcement, implies that NMDA blockade does not affect the acquisition of stimulus irrelevance in the preexposure stage but rather impairs the capacity of the preexposed rats to switch responding from the stimulus–nothing association
282
Ina Weiner and Michal Arad
acquired in preexposure to the stimulus–reinforcement association introduced in conditioning, under conditions triggering such a shift in controls. This is consistent with numerous demonstrations of inflexible behavior following NMDA blockade in rats and humans (Carlsson et al., 1999; Jentsch & Taylor, 2001; Krystal, Bennett, Abi-Saab et al., 2000; Moghaddam, Adams, Verma, & Daly, 1997; Svensson, 2000), and, in particular, with findings showing that low doses of NMDAR antagonists selectively impair rats’ capacity to alter responding based upon changed relationships between stimuli and outcomes (Jentsch & Taylor, 2001; Moghaddam et al., 1997; van der Meulen, Bilbija, Joosten et al., 2003). Since the NMDA receptor antagonist used in humans to mimic schizophrenia symptoms and assess treatment effects is ketamine (Krystal et al., 2003), we have recently tested the effects of this drug on LI. With parameters yielding LI in controls, ketamine spared LI at doses of 2, 8, and 20 mg/kg and disrupted LI at 60 mg/kg. As expected, under conditions that disrupted LI, ketamine led to persistent LI at doses of 2, 8, and 20 mg/kg (unpublished observations). Thus, ketamine is more efficacious than MK801 in producing persistent LI over a wider dose range. The efficacy of ketamine in inducing robust persistent LI further supports the relevance of this phenomenon to modeling schizophrenia symptoms. Recently, the role of reduced NMDA receptor activation in persistent LI has been supported by findings of persistent LI in mice following blockade of the NMDA glycine site with the antagonist L-701,324, as well as in Grin1D481N mice that have reduced NMDA receptor glycine site activation due to a point mutation in their NR1 glycine binding site (Labrie, Lipina, & Roder, 2008). The mechanisms by which NMDA antagonists produce persistent LI remain to be investigated. Since the information that enables switching to respond according to the stimulus–reinforcement association is provided by the glutamatergic afferents from limbic and frontal regions to the NAC (Weiner, 2003), one possibility is that NMDA antagonists block the transmission of information from limbic regions to NAC. Another possibility is that they act at the level of the VTA where they block the phasic response of DA neurons that normally occurs in response to salient and reinforcement-predicting stimuli, resulting in behavioral and cognitive perseveration (Gopel, Laufer, & Marcus, 2002; Ikemoto & Panksepp, 1999). Finally, MK801induced LI perseveration could be mediated by enhanced glutamate release within the prefrontal cortex (PFC), particularly as this effect has been related to perseverative behavior (Moghaddam et al., 1997). It has been suggested that NMDAR antagonist-induced perseverative behaviors are particularly relevant to negative symptoms of schizophrenia, which are characterized by inflexible and perseverative behaviors (Carlsson & Carlsson, 1990; Krystal et al., 2000; Moghaddam et al., 1997). We have suggested that NMDA antagonistinduced LI persistence may model attentional perseveration, or impaired attentional set shifting associated with cognitive impairments and negative symptomatology (Gaisler-Salomon & Weiner, 2003; Weiner, 2003). We have also pointed out that
The pharmacology of latent inhibition and its relevance to schizophrenia
283
the implication of this phenomenon to clinical research is that schizophrenia patients suffering from negative symptoms should show enhanced rather than disrupted LI. Indeed, three studies have demonstrated excessively strong LI in schizophrenia patients with negative symptoms (Cohen, Sereni, Kaplan et al., 2004; Gal, Barnea, Biran et al., 2009; Rascle et al., 2001).
Scopolamine-induced LI disruption and persistence Although muscarinic antagonists such as scopolamine and atropine induce a schizophrenia-like syndrome (“antimuscarinic syndrome”) in humans, which includes both positive symptoms and cognitive impairments (Clarke, Cassidy, Catalano et al., 2004), as well as psychotic-like effects in animal models of schizophrenia (Jones, Eberle, Shaw et al., 2005; Shannon, Rasmussen, Bymaster et al., 2000; Ukai, Okuda, & Mamiya, 2004), the cholinergic system has received less attention in schizophreniarelated research than the dopaminergic and the glutamatergic systems. Recent efforts to identify treatments targeting cognitive impairments in schizophrenia (Friedman, 2004; Marder & Fenton, 2004) have directed attention to the cholinergic system because of its well-known role in cognition (Everitt & Robbins, 1997; Hasselmo & McGaughy, 2004; Sarter, Nelson, & Bruno, 2005). The investigation of the role of the cholinergic system in LI has focused on the nicotinic receptors (Joseph, Peters, & Gray, 1993; Rochford, Sen, & Quirion, 1996), which do not appear to play a role in schizophrenia. Our interest was in the pro-psychotic and anti-cognitive effects of the muscarinic antagonist scopolamine in LI. Since scopolamine is known to impair associative learning (Tinsley, Quinn, & Fanselow, 2004), our investigations used doses of the drug that do not impair conditioning in the nonpreexposed animals. Low scopolamine-induced LI disruption Consistent with data that scopolamine induces psychosis-like effects in animals (Chang et al., 2007; Chen, Baxter, & Rodefer, 2004; Hyde & Crook, 2001; Joseph et al., 2000) scopolamine at doses of 0.15 and 0.5 mg/kg abolished LI; at a higher dose of 1 mg/kg scopolamine spared LI (Barak & Weiner, 2007). Scopolamineinduced LI disruption stemmed exclusively from improved performance of the scopolamine-treated preexposed groups, which conditioned as effectively as their nonpreexposed counterparts. In other words, scopolamine-treated preexposed rats behaved as if they were not preexposed, just as found in amphetamine-treated rats. While this finding suggested that scopolamine induces an amphetamine-like psychotic state in animals and by extrapolation in humans, stage-specific administration of scopolamine refuted this option. In contrast to amphetamine, scopolamine disrupted LI if given only before preexposure but not if given only before conditioning. Thus, muscarinic blockade induced by low doses of scopolamine has no effects on LI expression, but instead
284
Ina Weiner and Michal Arad
modulates the acquisition of LI. Specifically, low doses of scopolamine apparently act by attenuating the normal loss of attention to the stimulus occurring during nonreinforced preexposure. This action of scopolamine in the LI model is in line with extensive evidence implicating the cholinergic system in attentional processes (Everitt & Robbins, 1997; Sarter et al., 2005). Moreover, it suggests that muscarinic blockade can produce positive symptoms by abnormally enhancing stimulus salience, a process often suggested as giving rise to this class of symptoms (Kapur, 2003).
High scopolamine-induced LI persistence The fact that the high dose of scopolamine spared LI led us to test the possibility that higher doses of this drug can produce persistent LI. We found that 1.5 mg/kg scopolamine, like 1 mg/kg, spared LI under conditions yielding LI in controls. Furthermore, rats treated with this dose persisted in expressing LI under conditions preventing LI in controls (Barak & Weiner, 2009). Similarly to MK801, scopolamine led to persistent LI by reducing conditioning in the preexposed groups without concomitantly reducing conditioning in the nonpreexposed groups, indicating that, also in this case, LI persistence was not due to impaired conditioning per se, but rather to a failure to switch responding according to the stimulus–reinforcement association. Given the opposite behavioral effects of high and low scopolamine on LI, of particular interest was the question of whether the stage of action of scopolamine-induced persistent LI would be the same as that of scopolamineinduced disrupted LI, reflecting the common neurotransmitter dysfunction underlying these two LI aberrations, or that of MK801-induced persistent LI, reflecting common behavioral manifestation/cognitive deficit (persistent LI). We found that high dose of scopolamine differed from low also in the stage at which it acted: it was ineffective in producing persistent LI if given in preexposure but led to the emergence of LI when given in conditioning. Thus, contrary to low doses, scopolamine at the higher dose did not affect the acquisition of LI but modulated its expression in conditioning, so that high-scopolamine-treated rats persisted in responding according to the stimulus–no-event association. The latter is consistent with other demonstrations that scopolamine can produce perseverative behaviors (Chen et al., 2004; Ragozzino, Jih, & Tzavos, 2002; Soffie & Lamberty, 1987). It should be noted that in terms of performance, low and high scopolamine doses acted exclusively on the preexposed groups but produced opposite effects on their performance, the former increasing and the latter decreasing conditioning following nonreinforced preexposure. The neural mechanisms underlying the dose-dependent contrasting effects of scopolamine on LI remain to be determined. They could stem from dose-dependent effects of scopolamine on cholinergic transmission, whereby high doses cause a much larger increase in ACh levels in the medial prefrontal cortex than low doses (Ichikawa, Chung, Li et al., 2002), or the dose-dependent effects of scopolamine on NAC DA release,
The pharmacology of latent inhibition and its relevance to schizophrenia
285
with low but not higher doses increasing NAC DA release (Ichikawa et al., 2002). Recent results from our laboratory indicate that this differential effect may be mediated via muscarinic blockade in different brain regions, specifically entorhinal cortex for disrupted LI and basolateral amygdala for persistent LI (Barak & Weiner, 2010).
Treatments that alleviate abnormal LI Effects of antipsychotic drugs APDs are divided into two groups, typical and atypical. There are several criteria for this distinction (Arnt & Skarsfeldt, 1998; Kinon & Lieberman, 1996; Miyamoto, Duncan, Marx, & Lieberman, 2005). The most accepted criteria for atypicality are superior therapeutic efficacy, reduced capacity to cause extrapyramidal side effects, and catalepsy in rodents, as well as a broader receptor profile compared to typical APDs. The features most often suggested to account for the greater efficacy of atypical APDs are their mixed DA2–5-HT2 receptor antagonism and lower D2 occupancy (Arnt & Skarsfeldt, 1998; Meltzer & Nash, 1991; Schotte, Janssen, Gommeren et al., 1996). Overall, there is general agreement that both classes of APDs are effective against positive symptoms, whereas atypical APDs may be more effective for negative symptoms and cognitive symptoms, but the latter is a subject of continuous debate. Response to APDs is a prerequisite for the attainment of predictive validity for animal models of schizophrenia. Furthermore, the requirement is that animal models distinguish between the action of typical and atypical APDs in parallel to their presumed differential efficacy in the clinic. Animal models typically achieve the latter aim by using pro-psychotic drug administration. As described below, the LI model can achieve such discrimination with and without pro-psychotic drug priming.
Naι¨ve animals Weiner and Feldon (1987) reported that the prototypical APD, haloperidol, facilitated weak LI. Subsequently, this effect (also termed augmentation, potentiation or enhancement) was shown for other typical APDs as well as for atypical APDs, and it is the most widely reported effect of APDs on LI. Although APDs can augment LI under conditions of weak LI in controls, typically LI enhancement is demonstrated under conditions that are explicitly manipulated so as not to produce LI in control animals, namely, low number of preexposures or high number of conditioning trials (Dunn, Atwater, & Kilts, 1993; Feldon & Weiner, 1991; Gosselin et al., 1996; Gracey, Bell, & King, 2002; Killcross, Dickinson, & Robbins, 1994b; Moran et al., 1996; Peters & Joseph, 1993; Shadach, Feldon, & Weiner, 1999; Trimble, Bell, & King, 1997, 1998, 2002; Warburton et al., 1994; Weiner & Feldon, 1987; Weiner, Shadach, Barkai, & Feldon, 1997b; Weiner et al., 1996).
286
Ina Weiner and Michal Arad
APDs potentiate LI by actions in the conditioning stage; administration of APDs confined to preexposure has no LI potentiating effects (Peters & Joseph, 1993; Shadach et al., 1999; Shadach, Gaisler, Schiller, & Weiner, 2000; Weiner et al., 1997b). Thus, APDs do not facilitate the acquisition of the stimulus–no-event association in the preexposure stage, but promote the expression of this contingency in conditioning by retarding switching to respond according to the stimulus–reinforcement association, in line with the known action of dopamine blockade to retard behavioral switching (Gray et al., 1991; Weiner & Joel, 2002). Intra-accumbens injection of haloperidol has shown that NAC is the brain region responsible for persistent LI (Gray et al., 1997; Joseph et al., 2000). APD-induced potentiation of LI is notable in several respects: (1) it predicts antipsychotic activity for both typical and atypical APDs; (2) doses of APDs that enhance LI correlate with their clinical potency (Dunn et al., 1993); (3) it is obtained also in normal humans (McCartan, Bell, Green et al., 2001; Williams, Wellman, Geaney et al., 1996, 1997); (4) it does not require previous administration of DA agonists or other drugs so that the model does not rely on pharmacological means to elicit the behavioral index of antipsychotic activity. While these features give the LI model strong predictive validity for APDs, APD-induced LI persistence cannot dissociate between typical and atypical APDs. Shadach et al. (2000) set out to demonstrate such a dissociation based on the 5-HT2A antagonism of atypical APDs and the finding that LI is disrupted by the 5-HT2A antagonist, ritanserin (Cassaday, Hodges, & Gray, 1993). Their rationale was as follows: if atypical APDs disrupt LI, this action would have gone undetected because experiments testing the effects of APDs on LI use conditions that disrupt LI in controls; instead, parameters yielding LI in controls need to be employed. Furthermore, since atypical APDs given in conditioning potentiate LI, LI disruption due to serotonergic antagonism must occur via the preexposure stage. To test this hypothesis, Shadach et al. (2000) administered haloperidol, clozapine, or the selective 5-HT2A/5-HT2C antagonist ritanserin in the preexposure stage, or the conditioning stage, or in both, under conditions that yield and do not yield LI in controls. As found previously, with parameters which did not lead to LI in controls, both haloperidol and clozapine potentiated LI when administered in conditioning and in both stages, and were without an effect when administered in preexposure. Ritanserin was without an effect in all three administration conditions. The novel finding was that, with parameters which led to LI in controls, haloperidol and clozapine produced different patterns of results: haloperidol was without an effect in all three administration conditions whereas clozapine had no effect when administered in conditioning and in both stages but disrupted LI when administered in preexposure. Ritanserin disrupted LI when administered in preexposure and in both stages, and had no effect when administered in conditioning. These results provided the first demonstration that atypical APDs, in addition to their “typical” action of potentiating LI via conditioning, exert an “atypical” action of disrupting LI via preexposure.
The pharmacology of latent inhibition and its relevance to schizophrenia
287
In addition the results suggested that the preexposure-based LI disrupting effect of clozapine was due to its 5-HT2A antagonism. The fact that clozapine disrupted LI via preexposure but spared LI when administered in both stages indicated that 5-HT2 and DA2 antagonistic actions of clozapine compete in LI as previously suggested (Dunn et al., 1993; Trimble et al., 1998). The latter has implications for the action of all atypical APDs having a mixed 5-HT2A/ DA antagonism on LI. Since the relative potency of the DA and 5-HT2A antagonistic actions are dose-dependent, with 5-HT2 receptor occupancy predominating at lower doses and DA2 receptor occupancy occurring at higher doses (Leysen, Janssen, Schotte et al., 1993; Schotte et al., 1996), depending on the dose, the serotonergic component should be able to override the dopaminergic component, or vice versa, leading to either potentiated LI, intact LI, or disrupted LI. We showed precisely this pattern with the atypical APD risperidone (0.25, 0.5, 1.2, 2.5 mg/kg) which specifically targets D2 and 5-HT receptors, including 5-HT2A, 5-HT2C and 5-HT7, as well as with clozapine (2.5, 5, and 10 mg/kg) (Masson et al., 2003; Shadach et al., 1999, 2000) (unpublished observations). The competition between the D2 and 5-HT2 antagonism of atypical APDs has critical implications for interpreting the effects of these drugs on LI in animals and humans (Barrett, Bell, Watson, & King, 2004; McCartan et al., 2001), as well as for the evaluations of the clinical efficacy of these drugs, an issue that is beyond the scope of the present chapter. In summary, results with naı¨ ve rats demonstrated that the LI model has the capacity to discriminate between typical and atypical APDs, so that (1) both classes of drugs potentiate LI via their action at the conditioning stage under conditions which do not lead to LI in controls, and (2) atypical but not typical APDs disrupt LI via action at the preexposure stage under conditions which lead to LI in controls. Since it is commonly asserted that an animal model which is sensitive to both classes of APDs has predictive validity for positive symptoms whereas a model which is sensitive to atypical but not typical APDs has predictive validity for negative/cognitive symptoms (Arnt & Skarsfeldt, 1998; Kinon & Lieberman, 1996), APD-induced LI potentiation may have predictive validity for the treatment of positive symptoms whereas APD-induced LI disruption may have predictive validity for the treatment of negative symptoms/treatment-resistant schizophrenia.
Amphetamine-induced disrupted LI Acute amphetamine-induced LI disruption is reversed by both typical and atypical APDs (Gosselin et al., 1996; Solomon et al., 1981; Warburton et al., 1994; Weiner et al., 1996). APDs are effective also in restoring LI following chronic amphetamineinduced LI disruption, indicating that APDs can normalize behavioral LI abnormalities associated with a sensitized DA system (Russig et al., 2002). Similarly to their LI-potentiating effect, reversal of amphetamine-induced disruption by APDs occurs at the time of conditioning. Gray et al. (1997) and Joseph et al. (2000) showed that
288
Ina Weiner and Michal Arad
intra-accumbens haloperidol reversed the disruption of LI caused by acute systemic amphetamine administration, and that this effect was obtained with intra-accumbens haloperidol injection confined to the time of conditioning (Russig et al., 2002). Likewise, reversal of disrupted LI caused by chronic amphetamine was obtained with systemic APD injection given in both preexposure and conditioning or confined to conditioning. LI disruption following repeated amphetamine was also prevented by concomitant administration of haloperidol or clozapine (Tenn et al., 2005a). Interestingly, the latter study found that the selective D1 antagonist SCH23390 administered concomitantly with amphetamine also prevented LI disruption.
NMDA antagonist-induced persistent LI NMDA antagonist effects in healthy humans and schizophrenia patients are reduced by atypical but not typical APDs (Krystal et al., 2003; Lahti et al., 1995; Malhotra, Adler, Kennison et al., 1997), in line with the view that the former are more effective than the latter in improving negative symptoms and cognitive impairments in schizophrenia (Harvey, Rabinowitz, Eerdekens, & Davidson, 2005; Meltzer & Sumiyoshi, 2003; Moller, 2003) (but see Lieberman, Tollefson, Charles et al., 2005). Likewise, NMDA antagonist-induced behavioral abnormalities in animals, such as locomotor hyperactivity, stereotypy, impaired PPI, impaired social interaction, and perseveration, are typically reversed by atypical but not typical APDs (Bakshi, Swerdlow, & Geyer, 1994; Millan, Brocco, Gobert et al., 1999; Moghaddam & Jackson, 2003; Sams-Dodd, 1996), and such selective sensitivity to atypical APDs is considered to lend the NMDA antagonist models predictive validity for cognitive/negative symptoms. Consequently, we expected that MK801-induced LI perseveration would be reversed by clozapine and risperidone but not by haloperidol. Moreover, we expected clozapine and risperidone to act via the preexposure stage since this is the stage where these drugs acted to disrupt LI. Finally, because the 5-HT2A/2C antagonist ritanserin was shown by us to disrupt LI (Shadach et al., 2000), we expected that MK801 persistent LI would be reversed by the selective 5-HT2A/2C antagonist M100907 via the preexposure stage. Consistent with the selective sensitivity of NMDAR antagonist-induced behavioral effects to atypical APDs, MK801-induced persistent LI was reversed by clozapine (Gaisler-Salomon & Weiner, 2003; Lipina et al., 2005) and risperidone (Gaisler-Salomon, Diamant, Rubin, & Weiner, 2008) but was unaffected by the typical APD haloperidol (Gaisler-Salomon & Weiner, 2003), lending the model predictive validity for negative/cognitive symptoms. Recently, we found the same selective efficacy for clozapine compared to haloperidol for ketamine-induced persistent LI (unpublished observations). Clozapine was also found to reverse persistent LI in Grin1D481N mutant mice that have reduced NMDA receptor activation (Labrie et al., 2008). Stage-confined administration revealed, as predicted, that clozapine and risperidone were ineffective when administered in conditioning, but that they
The pharmacology of latent inhibition and its relevance to schizophrenia
289
reversed MK801-induced persistent LI when administered in preexposure. Furthermore, MK801-induced persistent LI was reversed by a very low non-dopaminergic dose of risperidone and the selective 5-HT2A/C antagonist M100907 via preexposure. Since risperidone specifically targets D2 and 5-HT receptors, the findings that preexposure-based reversal of MK801-induced LI persistence is produced by serotonergic dose of risperidone and M100907, whereas the D2 antagonist haloperidol is ineffective, imply that reversal of MK801-induced persistent LI by risperidone is likely to be mediated by its 5-HT2A antagonism, and that the same may apply to the effects of clozapine. The notion that 5-HT2A antagonist action mediates the selective reversal of NMDA antagonist-induced behavioral effects by atypical APDs is common, although it is acknowledged that the interactions between atypical antipsychotics and NMDA antagonists cannot be mediated by competition for a common receptor. Such direct competition can be ruled out in LI, because MK801-induced LI persistence and the reversal of this effect by atypical APDs and 5-HT2A antagonists occur at different stages of the LI procedure, occurring 24 h apart. Rather, the reversal of MK801-induced LI persistence following atypical APDs and 5-HT2A antagonists is likely to reflect complex interactions within the brain circuitry that modulates the expression of LI (Weiner, 2003). Importantly, the stage-based differentiation between the action of MK801 and atypical APDs provides a model of cognitive/negative symptoms in which the schizophrenia-mimicking LI abnormality is drug-induced, but the detection of the atypical antipsychotic action is not dependent on the mechanism of action of the pro-psychotic drug. Such a model is ideal for identifying agents acting through novel mechanisms.
Scopolamine-induced disrupted LI Antimuscarinic psychosis can be alleviated by APDs (Gopel et al., 2002; Perry, Wilding, & Juhl, 1978). Likewise, psychotic-like effects induced by systemic administration of nonspecific muscarinic antagonists such as atropine or scopolamine in animal models of schizophrenia, such as locomotor hyperactivity/stereotypy, and PPI, are reversed by APDs (Jones et al., 2005; Shannon & Peters, 1990). Scopolamine-induced LI disruption was reversed by both haloperidol and clozapine, supporting the notion that disruption of LI by scopolamine can provide a model of the positive symptoms that are seen in antimuscarinic psychosis, and, by extrapolation, of the cholinergic aspects of positive symptoms in endogenous schizophrenia (Barak & Weiner, 2007). Interestingly, stage-specific administration revealed that both APDs restored scopolamine-induced disrupted LI if injected in the conditioning stage alone, but not in the preexposure stage alone, whereas a low scopolamine dose was shown to induce disrupted LI. Thus, reversal of scopolamine-induced disrupted LI by APDs cannot be attributed to a direct interaction between the dopaminergic and the muscarinic cholinergic systems as has been suggested for other scopolamine-induced behavioral deficits (Jones et al., 2005; Mathur, Shandarin,
290
Ina Weiner and Michal Arad
LaViolette et al., 1997). Rather, reversal of scopolamine-induced LI disruption by APDs, like reversal of MK801-induced persistent LI by atypical APDs, is likely to reflect complex interactions within the brain circuitry that modulate the expression of LI. Likewise, the stage-based differentiation between the action of scopolamine and APDs suggests that scopolamine-induced disrupted LI provides a model of positive symptoms that allows the detection of antipsychotic action that is independent of the mechanism of action of the pro-psychotic drug and thus can identify treatments acting via novel mechanisms. Scopolamine-induced persistent LI In stark contrast to their effects on low-dose scopolamine-induced disrupted LI, both haloperidol and clozapine failed to reverse scopolamine-induced persistent LI (Barak & Weiner, 2009). While the failure of haloperidol could be expected as this drug also failed to reverse MK801-induced persistent LI, the failure of clozapine was unexpected also based on findings in the MK801 model. Increasing the doses of both APDs failed to yield the expected differentiation between haloperidol and clozapine. Finally, since clozapine acts to reverse persistent LI in the preexposure stage, and its action in conditioning may interfere with its action in preexposure (see above), we confined clozapine administration to preexposure. This regime also failed to reverse scopolamine-induced persistent LI. In contrast to their inefficacy against scopolamine-induced persistence, both haloperidol and clozapine, at both doses used, were effective in potentiating LI in vehicle-treated rats, as expected. Taken together, these findings suggest that scopolamine-induced persistent LI may provide a novel LI model with a pharmacological profile that sets it apart from that of both scopolamine-induced disrupted LI and MK801-induced persistent LI models, at least with respect to its lack of response to APDs. The latter suggests that scopolamine-induced persistent LI is an APD-resistant cognitive impairment, and thus may model APD-resistant cognitive impairments in schizophrenia.
Effects of glycinergic compounds In recent years, therapeutic strategies have increasingly focused on enhancing NMDA receptor function as an alternative treatment of negative and cognitive impairments in schizophrenia. Findings from clinical trials have suggested that compounds that potentiate NMDA transmission via the glycine B modulatory site on the NMDAR, either directly by acting as agonists (such as glycine, D-serine, D-cycloserine (DCS) and D-alanine), or indirectly by inhibiting the glycine transporter (GlyT1; e.g., sarcosine), can improve some aspects of negative symptomatology and cognitive dysfunction, without interfering with the beneficial effects of antipsychotics on positive symptoms (Goff, Tsai, Manoach, & Coyle, 1995; Javitt, 2002, 2008), although contradictory results were also reported (Buchanan, Javitt, Marder et al., 2007; Goff, Herz, Posever et al., 2005).
The pharmacology of latent inhibition and its relevance to schizophrenia
291
Naι¨ve animals Similarly to APDs, glycinergic compounds potentiate LI, but do not disrupt the phenomenon (Black, Varty, Arad et al., 2009; Lipina et al., 2005). NMDA antagonist-induced persistent LI Positive modulators of NMDA receptor function, although less thoroughly characterized than APDs, were shown to reverse NMDA antagonist-induced motor effects (Dall’Olio & Gandolfi, 1993; Depoortere, Dargazanli, EstenneBouhtou et al., 2005; Harsing, Gacsalyi, Szabo et al., 2003; Javitt, Balla, Sershen, & Lajtha, 1999; Millan et al., 1999; Tanii, Nishikawa, Hashimoto, & Takahashi, 1994; Toth & Lajtha, 1986), and to counter the effects of MK801 in PPI (Lipina et al., 2005). Lipina et al. (2005) were the first to show, in mice, that MK801-induced persistent LI was reversed by D-serine and GlyT1 inhibitor ALX5407. Likewise, MK801-induced persistent LI in rats was shown to be reversed by glycine site agonists glycine and DCS, as well as the older GlyT1 inhibitor glycyldodecylamide (GDA) and novel GlyT1 inhibitors SSR103800 and SSR504734 (Black et al., 2009; Gaisler-Salomon et al., 2008). Recently, Labrie et al. (2008) showed that D-serine and ALX-5407 reversed persistent LI in Grin1D481N mutant mice. Similarly to APDs, glycinergic compounds potentiate LI in mice and rats (Black et al., 2009; Lipina et al., 2005). Gaisler-Salomon et al. (2008) characterized the stage-specific effects of glycinergic compounds on MK801-induced persistent LI in rats, with the expectation that these compounds would act to reverse MK801-induced LI persistence at the same stage at which it is induced by MK801, namely, conditioning, and thus be differentiated from atypical APDs (which act in preexposure; see above). As predicted, we found that glycine was effective if given in conditioning but not in preexposure, and we demonstrated the same conditioning-based reversal of MK801-induced persistent LI by DCS and GDA. These results indicate that reversal of MK801-induced LI abnormality by glycinergic compounds occurs via effects at the level of the NMDA receptor. The mechanism of interaction between glycinergic drugs and MK801 remains to be elucidated. Possibly, stimulation of GlyB sites increases the frequency and duration of channel opening, and in this manner promotes the dissociation of MK801, bound to the PCP site located inside the ion channel, from the receptor pore (Bolshakov, Gmiro, Tikhonov, & Magazanik, 2003; Millan, 2002, 2005). Whatever the mechanisms of action of glycinergic compounds, taken together with the effects of haloperidol and atypical APDs described above, the MK801-induced persistent LI model provides a powerful screening test which enables a three-way dissociation between atypical APDs (effective in preexposure but ineffective in conditioning), NMDA-based treatments (ineffective in preexposure but effective in conditioning), and typical APDs (ineffective in either of the stages).
292
Ina Weiner and Michal Arad
Low-scopolamine-induced disrupted LI and high-scopolamine-induced persistent LI Scopolamine-induced disrupted as well as persistent LI were reversed by glycine (Barak & Weiner, 2009) (unpublished observations). While the latter is consistent with other reports that glycinergic agonists reverse scopolamine-induced cognitive impairments (Ohno & Watanabe, 1996; Sirvio, Ekonsalo, Riekkinen et al., 1992; Viu, Zapata, Capdevila et al., 2000), the former, to the best of our knowledge, is the only demonstration that glycine can reverse scopolamine-induced psychotic-like behaviors. Glycine reversed both scopolamine-induced disrupted and persistent LI via the conditioning stage, thus targeting high-scopolamine-induced persistent LI at the stage where the abnormality is generated, but targeting low-scopolamine-induced abnormality via some indirect route. Remarkably, glycine reversed low-scopolamineinduced disrupted LI also via the preexposure stage. The capacity of glycine to act via both conditioning and preexposure suggests that it can target the underlying dysfunction irrespective of its source. Furthermore, while glycine action in conditioning reflects its capacity to increase behavioral/cognitive flexibility, apparently this drug can also directly affect attentional processes related to salience modulation. The mechanism by which glycine reverses the effects of high scopolamine in conditioning and low scopolamine in preexposure could involve a direct interaction with NMDA receptors since these receptors are present on cholinergic neurons (Ransom & Deschenes, 1989). Consequently, glycine may increase ACh release, which would compete with scopolamine at muscarinic receptors binding sites. Reversal of low-scopolamine-induced LI by glycine administration in conditioning is obviously mediated by complex interactions. Amphetamine-induced disrupted LI Glycine was ineffective in this model, but the novel GlyT1 inhibitor SSR103800 reversed amphetamine-induced disrupted LI (Black et al., 2009). Although glycinergic agents have been shown to block locomotion-enhancing effects caused by amphetamine in neurodevelopmental models (Depoortere et al., 2005; Kato, Shishido, Ono et al., 2001), data on their ability to reverse the effects of acute amphetamine administration are equivocal and limited to amphetamine-induced activity (Harsing et al., 2003; Javitt, Sershen, Hashim, & Lajtha, 1997). Our finding that SSR103800 reversed amphetamine-induced LI disruption is thus an important demonstration of this compound’s activity in a well-validated model of positive symptoms. The capacity of SSR504734 to increase dopamine levels in the prefrontal cortex (Boulay, Pichat, Dargazanli et al., 2008; Depoortere et al., 2005) may explain its activity in the hyperdopaminergic models because increased levels of dopamine in the prefrontal cortex would attenuate the amphetamine-mediated increase in subcortical dopamine tone and thus reverse the behavioral effects of amphetamine (Depoortere et al., 2005; Grace, 1991). Another mechanism of SSR504734 activity
The pharmacology of latent inhibition and its relevance to schizophrenia
293
in hyperdopaminergic models may stem from its capacity to enhance striatal NMDA-mediated function (Depoortere et al., 2005), since such potentiation would be expected to lead to inhibition of striatal DA release (via NMDA-stimulated GABA release), an effect that would be expected to counteract amphetamine effects.
Effects of cholinergic cognitive enhancers AChE inhibitor physostigmine Physostigmine as well as other AChE inhibitors which increase extracellular ACh levels are widely used to treat cognitive deficits, particularly in Alzheimer’s disease (Bianchetti, Ranieri, Margiotta, & Trabucchi, 2006; Brousseau, Rourke, & Burke, 2007). There is some evidence that AChE inhibitors also may have beneficial effects in the treatment of cognitive impairments in schizophrenia (Friedman, 2004; Noren, Bjorner, Sonesson, & Eriksson, 2006; Schubert, Young, & Hicks, 2006), although the extent of this benefit is debatable (Kumari, Aasen, Ffytche et al., 2006; Lee, Lee, Lee, & Kim, 2007). Recent small-size clinical studies indicate that adjunctive use of galantamine also may improve positive symptoms (Noren et al., 2006). Finally, AChE inhibitors including physostigmine alleviate antimuscarinic psychosis (Granacher & Baldessarini, 1975; Nogue, Sanz, Munne, & de la Torre, 1991; Perry et al., 1978). Studies with animal models of schizophrenia have shown that AChE inhibitors can reverse the psychotomimetic effects induced by scopolamine (Hohnadel, Bouchard, & Terry, 2007; Jones & Shannon, 2000; Shannon & Peters, 1990), DA agonists amphetamine or apomorphine (Andersen, Werge, & Fink-Jensen, 2007; Hohnadel et al., 2007; Karan, Ravishankar, & Pandhi, 2000), and MK801 (Csernansky, Martin, Shah et al., 2005; Hohnadel et al., 2007). Naι¨ve animals Physostigmine has no effect on LI when given on its own (Barak & Weiner, 2007, 2009). Low-scopolamine-induced disrupted LI and high-scopolamine-induced persistent LI Physostigmine reversed both low- and high-scopolamine-induced LI abnormalities (Barak & Weiner, 2007, 2009), in line with other data on this drug’s effectiveness in reversing psychotomimetic effects and cognitive deficits induced by muscarinic blockade (Carnicella, Pain, & Oberling, 2005; Hohnadel et al., 2007; Jones & Shannon, 2000). In each of the models, physostigmine acted via the stage at which scopolamine induced the LI abnormality. While this stage-specific targeting would be expected from a treatment that directly antagonizes the effects of the abnormalityinducing agent, it should be noted that in terms of the cognitive processes affected, physostigmine is capable of ameliorating disrupted attentional processing and inflexibility, raising the question of whether it can exert these effects also when such abnormalities are induced by non-cholinergic manipulations.
294
Ina Weiner and Michal Arad
Amphetamine-induced disrupted LI Physostigmine failed to restore amphetamine-induced LI disruption (Barak & Weiner, 2007). The latter is inconsistent with reports that AChE inhibitors reversed the psychotomimetic effects of amphetamine or apomorphine (Andersen et al., 2007; Hohnadel et al., 2007; Karan et al., 2000) and, in particular, blocked amphetamineinduced stereotypy and potentiated haloperidol-induced catalepsy (Karan et al., 2000). However, the ineffectiveness of physostigmine seen here is in line with its inefficacy in reversing amphetamine-induced hyperactivity (Stone, Rudd, & Gold, 1990). The potential difference between the efficacy of AChE inhibitors to reverse the effects of high amphetamine doses and direct DA agonists such as apomorphine vs. low doses of amphetamine may have important implications for understanding their influence on positive symptoms (Noren et al., 2006). Interestingly, our results indicate that although physostigmine was apparently able to strengthen acquisition of irrelevance that was disrupted by scopolamine, it was incapable of doing so in amphetamine-treated rats. MK801-induced persistent LI Recent interest in the relevance of cholinergic cognitive enhancement for cognitive impairments of schizophrenia has led to tests of AChE inhibitors on MK801-induced behavioral abnormalities. Two recent studies have reported positive results (Csernansky et al., 2005; Hohnadel et al., 2007). Consistent with these results, we found that physostigmine reversed MK801-induced LI persistence, providing additional evidence for the potential efficacy of this class of drugs in ameliorating schizophrenia-relevant negative/cognitive deficits. As for a possible mechanism underlying this effect, increased ACh levels can activate muscarinic and nicotinic cholinergic receptors, such as a7 nicotinic and M1 muscarinic, on NMDA neurons which can potentiate NMDA activity (Marino, Rouse, Levey et al., 1998).
a7 nicotinic partial agonist SSR180711 Among cholinergic function enhancers, a7 nicotinic ACh receptor (nAChR) agonists have emerged as particularly promising (Martin, Kem, & Freedman, 2004). There is accumulating evidence that a7-nAChR agonists facilitate cognitive function in rodents and humans (Levin, Bettegowda, Blosser, & Gordon, 1999; Olincy & Stevens, 2007). Of particular relevance to attentional and sensory gating deficits in schizophrenia, a7-nAChR agonists alleviate both types of deficits in humans and animals (Hajos, Hurst, Hoffmann et al., 2005; Olincy, Harris, Johnson et al., 2006; Timmermann, Gronlien, Kohlhaas et al., 2007; Wishka, Walker, Yates et al., 2006). While a7-nAChR agonists have been identified as lead candidates for improving cognition in schizophrenia (MATRICS), no activity has been predicted for these agents on positive symptoms of schizophrenia. SSR180711(4-bromophenyl-1,
The pharmacology of latent inhibition and its relevance to schizophrenia
295
4-diazabicyclo[3.2.2]nonane-4-carboxylate-hydrochloride) is a novel a7-nAChR partial agonist that produces several electrophysiological, neurochemical and behavioral effects predictive of activity against cognitive impairments of schizophrenia (Biton, Bergis, Galli et al., 2007; Hashimoto, Koike, Shimizu, & Iyo, 2005; Pichat, Bergis, Terranova et al., 2007). Naι¨ve animals Similarly to glycinergic compounds, SSR180711 potentiated but did not disrupt LI (Barak et al., 2009). MK801-induced persistent LI MK801-induced persistent LI was reversed by SSR180711 (Barak et al., 2009), consistent with previous findings with this and other nicotinic agonists in antagonizing the behavioral effects of NMDA blockade (Hashimoto et al., 2005; Mastropaolo, Rosse, & Deutsch, 2004; Pichat et al., 2007; Rezvani & Levin, 2003). The activity of SSR180711 in the hypoglutamatergic models is most likely a consequence of its capacity to increase, via activation of presynaptic a7-nAChRs present on glutamatergic neurons, glutamate levels in the PFC, the hippocampus, and the amygdala (Biton et al., 2007; Pichat et al., 2007). SSR180711 could also act via enhancement of the extracellular ACh levels in the hippocampus and PFC (Biton et al., 2007) as suggested above for physostigmine. Amphetamine-induced LI disruption Unlike findings in other models of positive symptoms tested to date, but consistent with its capacity to potentiate LI in normal rats, SSR180711 reversed amphetamineinduced disrupted LI (Barak et al., 2009), indicating it may effectively reduce positive symptoms of schizophrenia. While this is the first demonstration of such activity for an a7 agonist, it was expected based on the same mechanisms that lend this agent efficacy against cognitive/negative symptoms. Thus, the capacity of SSR180711 to increase hippocampal glutamate neurotransmission as well as prefrontal dopamine levels (Biton et al., 2007; Pichat et al., 2007) would be expected to reduce mesolimbic DA function and block behavioral effects of amphetamine (Goto & Grace, 2007). Alternatively, restoration of disrupted LI could be mediated by increased frontal ACh levels (Biton et al., 2007), because such an increase is expected to facilitate attentional processing via both nicotinic and muscarinic receptors (Hasselmo & McGaughy, 2004; Sarter, Bruno, & Givens, 2003; Sarter et al., 2005). Interestingly, in the latter case, a7 partial agonism would be expected to target positive symptoms directly via modulation of aberrant stimulus salience and thus act via preexposure. This remains to be tested. Irrespective of the mechanisms underlying the effects of SSR180711 on LI, this is the only model to date in which this agent was shown to exert both pro-cognitive and antipsychotic effects.
296
Ina Weiner and Michal Arad
The DA antagonist LI model: haloperidol-induced LI persistence This model is presented in a different format from the others because it differs from the other models and conventional pharmacological modeling of schizophrenia in not being based on a pro-psychotic drug, and in fact, paradoxically, it is based on an antipsychotic drug. However, while DA blockers are effective in the treatment of positive symptoms, their efficacy in treating negative symptoms has long been recognized as very limited, and these drugs themselves can lead to a syndrome similar to the negative symptomatology of schizophrenia (Breier & Berg, 1999; Kane, 1995; Kinon & Lieberman, 1996; Miyamoto et al., 2005). Conversely, amphetamine and other DA enhancing agents improve negative symptoms (Angrist, Peselow, Rubinstein et al., 1982; Ogura, Kishimoto, & Nakao, 1976; Sanfilipo, Wolkin, Angrist et al., 1996). D2 radioreceptor imaging studies (AbiDargham, Gil, Krystal et al., 1998; Breier, Su, Saunders, Carson, Kolachana, de Bartolomeis et al., 1997; Laruelle, Abi-Dargham, Gil et al., 1999; Laruelle, AbiDargham, van Dyck et al., 1996) have shown that amphetamine challenge and the concomitant increase in striatal DA transmission in untreated and neuroleptic naı¨ ve schizophrenics were correlated with an exacerbation of positive symptoms and an improvement in negative symptoms, indicating that increased striatal DA transmission is related to positive symptoms, whereas reduced DA function is associated with negative symptoms. Likewise, PET studies found increased rates of DA synthesis, consistent with increased presynaptic activity, in both first-admission and more chronic psychotic schizophrenic patients, whereas decreased DA synthesis characterized schizophrenic patients with primarily negative symptomatology (DaoCastellana, Paillere-Martinot, Hantraye et al., 1997; Hietala, Syvalahti, Vilkman et al., 1999). Thus, the association of increased and decreased DA transmission with disrupted LI, which models positive symptoms, and persistent LI, which models negative symptoms, respectively, would be consistent with the clinical picture. We have therefore begun to investigate the pharmacological profile of haloperidolinduced persistent LI. Our first aim was to test whether this abnormality would be reversed by atypical APDs via the preexposure stage. Indeed, we found that haloperidol-induced persistent LI was reversed by clozapine and risperidone given in the preexposure stage. In contrast, neither glycine nor physostigmine reversed haloperidol-induced persistent LI (unpublished observations). The latter sets haloperidol-induced persistent LI apart from both MK801- and scopolamineinduced persistent LI. Finally, we tested whether haloperidol-induced persistent LI would be reversed by amphetamine, as found for negative symptoms in the clinic, and compared this to the action of amphetamine against MK801-induced persistent LI. Haloperidol- but not MK801-induced persistent LI was reversed by amphetamine, supporting the distinct neurochemical basis of the two persistent LIs as well as the notion that the former may model DA-hypofunction-based negative symptoms.
The pharmacology of latent inhibition and its relevance to schizophrenia
297
Our use of haloperidol-induced persistent LI as a model of negative symptoms does not undermine the predictive validity of this phenomenon or its utility as a screening test for typical APD action; rather, it suggests that persistent LI taps the D2 blocking action of these drugs which is beneficial in states of abnormally increased DA function but is not beneficial and indeed is deleterious on its own in a system with normal DA function.
Neurodevelopmental perturbation-induced disrupted and persistent LI In recent decades, the widespread recognition that schizophrenia is a neurodevelopmental disorder has been paralleled by the development of animal neurodevelopmental models that are based on interference with normal development during vulnerable periods of perinatal neurogenesis, in utero or in neonatal animals. Schizophrenia-like abnormalities observed with such models are typically shown to be associated with neuroanatomical and cellular alterations that have been implicated in this disorder and thus are believed to closely approximate the pathophysiology, adding to the etiological validity of the model. Furthermore, the development of behavioral abnormalities in these models often reproduces the developmental delay characteristic of schizophrenia. To date, the effects of schizophrenia-relevant treatments on developmentally induced LI abnormalities have been shown in three neurodevelopmental models. These models are included here because it is important to show that also neurodevelopmental perturbations can produce disrupted and persistent LI. In the first model, rats exposed from birth to weaning to extreme restriction of environmental stimulation display disrupted LI in adulthood, and this is reversed by haloperidol. The second model, maternal immune activation, is based on the welldocumented association between maternal exposure to infection in pregnancy and increased risk of schizophrenia in the offspring. In the model, injection of pregnant dams with the viral-mimic polyinosinic-polycytidilic acid (Poly I:C) leads to disrupted LI in the adult offspring, which is reversed by both typical and atypical APDs (Zuckerman, Rimmerman, & Weiner, 2003; Zuckerman & Weiner, 2005). In the third model, inhibition of nitric oxide (NO) production is produced during the very early neonatal period (postnatal days 4–5) (Black, Selk, Hitchcock et al., 1999; Black, Simmonds, Senyah, & Wettstein, 2002), presumably modeling disrupted NO function in schizophrenia (Bernstein, Bogerts, & Keilhoff, 2005). This developmental interference is the only one to date that leads to abnormally persistent LI in adulthood. This persistent LI displays an identical pharmacological profile to that of MK801-induced persistent LI: it is reversed by clozapine but not by haloperidol, and by glycinergic NMDA function enhancers glycine, DCS, and the GlyT1 inhibitors (Barak, Arad, De Levie et al., 2009; Black et al., 2009; De Levie & Weiner, 2007a, 2007b; unpublished observations).
298
Ina Weiner and Michal Arad
Summary Beginning with Kraeplin’s classical observation that schizophrenia patients are unable to focus on relevant stimuli, and relatedly that they are irresistibly drawn to irrelevant stimuli, disturbances in switching capacity have been repeatedly noted in schizophrenia (Lyon, 1991; Weiner & Joel, 2002). Thus, on the one hand, the schizophrenic deficit has been described as an inability to maintain a major response set or a dominant interpretation of a given situation, excessive yielding to the immediate situational demands, and enhanced switching from one associative content to another (Bleuler, 1911; Broen, 1968; Frith, 1979; Gray et al., 1991; Payne, 1966; Shakow, 1962; Swerdlow & Koob, 1987). On the other hand, schizophrenics are known to exhibit behavioral inflexibility and perseveration (Carpenter, Heinrichs, & Alphs, 1985; Carpenter, Heinrichs, & Wagman, 1988; Crider, 1997; Karnath & Wallesch, 1992; Robbins, 1991; Wolkin, Sanfilipo, Wolf et al., 1992). As we showed above, depending on drug or developmental manipulations used to alter LI, the LI model can mimic the two extremes of deficient cognitive switching seen in schizophrenia: excessive switching between associations, manifested in disrupted LI under conditions that yield LI in normal rats, and retarded switching between associations, manifested in persistent LI under conditions which disrupt the phenomenon in normal rats. The facts that these behavioral manifestations can be mapped onto underlying neural (see Weiner, this volume) and neurochemical systems which are involved in schizophrenia support their relevance to this disorder. Disrupted and persistent LI can be seen as two poles of dysfunctional attentional control, namely, a failure to inhibit attention to irrelevant stimuli and a failure to re-deploy attention when previously irrelevant stimuli become relevant. The former would likely give rise to aberrantly increased salience perception and distractibility that are associated with psychotic symptoms (Gray et al., 1991; Kapur, 2003; Smith, Li, Becker, & Kapur, 2006; Swerdlow & Koob, 1987; Weiner & Joel, 2002), whereas the latter would likely result in cognitive inflexibility and impaired attentional shifting that are associated with negative/cognitive symptoms (Carlsson & Carlsson, 1990; Krystal et al., 2003; Moghaddam et al., 1997; Weiner, 2003). As discussed throughout the text and summarized in Tables 13.1 and 13.2, there are currently two pharmacological models of LI disruption, namely, amphetamine-induced LI disruption and scopolamine-induced LI disruption, and three pharmacological models of LI persistence, namely, MK801-induced LI persistence, scopolamine-induced LI persistence and haloperidol-induced LI persistence. Each of these models exhibits a unique pattern of responsiveness to typical and atypical APDs and other treatments believed relevant to future pharmacotherapy of schizophrenia. These five models allow a level of discriminability that has not been achieved using a single behavioral paradigm.
þ
þ
þ [COND] þ [COND] þ [COND] þ [PREEX] ?
Low scop.
þ [PREEX] þ [COND] þ [COND] þ
þ [PREEX] –
–
–
–
Low MK801
Controla –
Persistent LI
LI
þ [COND] þ [COND] ?
–
–
High scop.
Notes: þ, effective; –, ineffective; ?, unknown; [COND], acts via conditioning stage; [PREEX], acts via preexposure stage; a LI in naı¨ ve animals; b the active compound is GlyT1 inhibitor SSR103800.
a7 nicotinic agonist
Physostigmine
Glycine –
þ [COND] þ [COND] þb
þ [COND] þ [COND] þ [COND] –
Haloperidol
Clozapine
Low amph.
Control
Drug
Disrupted LI
Model
?
–
þ [PREEX] –
Haloperidol
Table 13.1. Summary of representative antipsychotic and other putative treatments tested against models of disrupted and persistent LI
300
Ina Weiner and Michal Arad
Table 13.2. Five pharmacological LI models proposed to model five domains of pathology of schizophrenia Disrupted LI
Persistent LI
Amphetamine
Scopolamine
Scopolamine
MK801
Haloperidol
Symptom domain
Positive
Positive/ cognitive
Cognitive
Negative/ cognitive
Negative
Reversed by
Typical APDs
Typical APDs Atypical APDs Cognitive enhancers
Atypical APDs Cognitive enhancers
Atypical APDs
Atypical APDs
Resistant to
Cognitive enhancers Typical APDs Atypical APDs
Typical APDs
Cognitive enhancers*
Cognitive enhancers
Amphetamine- and low-scopolamine-induced disrupted LI The fact that both amphetamine and low scopolamine produce disrupted LI which is reversed by typical and atypical APDs supports the notion that these two drug models model positive symptoms. However, in spite of their identical behavioral manifestations and APD response, the “antimuscarinic LI model” and the “dopamine agonist LI model” are distinct in several respects. First, amphetamine disrupts LI via effects exerted at the conditioning stage and spares LI if it is given only before preexposure, whereas scopolamine disrupts LI via effects exerted at the preexposure stage and spares LI when given only in conditioning. In addition to indicating different underlying neural substrates, stage-based dissociation implies that LI disruption induced by these two drugs reflects a disturbance to distinct psychological functions. Scopolamine impairs/prevents the acquisition of inattention to the preexposed stimulus, whereas amphetamine impairs/prevents the subsequent expression of normally acquired inattention. The dissociation between scopolamine- and amphetamine-induced disruption of LI is further highlighted by the manner in which the two abnormalities are reversed by APDs. Thus, although APDs reverse both abnormalities, in the case of amphetamine-induced LI disruption, the pro-psychotic and the antipsychotic actions are exerted at the same stage of the procedure (conditioning) and thus likely reflect a direct interaction, whereas in the case of scopolamineinduced LI disruption, the pro-psychotic and antipsychotic actions are generated in different stages of the procedure (preexposure and conditioning, respectively) and thus
The pharmacology of latent inhibition and its relevance to schizophrenia
301
mediated by distinct mechanisms. Third, scopolamine-induced, but not amphetamineinduced, LI disruption is reversed by physostigmine. The latter reverses scopolamineinduced disruption via preexposure, indicating direct interaction. In psychological terms, physostigmine apparently acts by restoring animals’ ability to inattend to inconsequential stimuli. Finally, scopolamine-induced, but not amphetamineinduced, LI disruption is reversed by glycine. However, since amphetamine-induced LI disruption is reversed by GlyT1 inhibitor, the two deficits may merely differ in their sensitivity to glycinergic enhancers. Taken together, these findings suggest that scopolamine- and amphetamineinduced LI disruption model different aspects of schizophrenic psychoses. Specifically, scopolamine-induced LI disruption may model muscarinic-related positive symptoms which are more linked to attentional deficits/cognitive impairments than dopamine-related positive symptoms. The distinct characteristics of the two LI disruptions allows one to refine the search for treatments that selectively target each of these abnormalities. For example, it is of interest to determine whether specific muscarinic receptor agonists, which were shown to exhibit antipsychotic properties in the clinic (Bymaster, Felder, Ahmed, & McKinzie, 2002; Jones et al., 2005; Stanhope, Mirza, Bickerdike et al., 2001), would reverse both scopolamine- and amphetamine-induced LI deficits as was shown in other animal models or would show selectivity as found here for physostigmine. Irrespective of the above, it should be noted that our results underscore the utility of both typical and atypical APDs for reversing disrupted LI, consistent with their known efficacy for positive symptoms. This is further emphasized by the fact that neurodevelopmentally induced disrupted LI is also sensitive to both APD classes. Thus, disrupted LI is responsive to APDs irrespective of their underlying pathological, neurochemical, and cognitive mechanism. Based on what we know from APD action in LI, we would claim that this utility derives from their capacity to reduce behavioral and cognitive overswitching irrespective of its origin.
MK801-, high-scopolamine- and haloperidol-induced persistent LI In the case of persistent LI, all three drugs lead to LI emergence via the conditioning stage without impairing associative learning, implying common cognitive dysfunction, namely, cognitive/behavioral inflexibility. Nevertheless, the three persistent LIs exhibit distinct pharmacological profiles. Persistent LI induced by MK801, a representative of a pro-psychotic class of drugs, most widely used as a pharmacological model of negative/cognitive symptoms, is resistant to haloperidol but is reversed by clozapine and risperidone, glycine, DCS, GlyT1 inhibitors, physostigmine, and a7 agonist. This treatment responsivity has been shown previously with other NMDA antagonist-induced behavioral deficits considered to be relevant to negative/cognitive symptoms of schizophrenia (see reference above). We are not aware of any single NMDA antagonist model that
302
Ina Weiner and Michal Arad
exhibits such a broad pharmacological profile. It should also be noted that the MK801 persistent LI model, while responsive to both atypical APDs and glycinergic enhancers, is the only NMDA antagonist model that can discriminate between them (based on their stage-specific action). The pharmacological profile of MK801induced persistent LI is reproduced with remarkable accuracy by persistent LI induced by neonatal inhibition of NOS, suggesting that the latter mimics a developmentally propelled dysfunction of NMDA-based brain circuitries. Scopolamine-induced persistent LI is reversed by glycine and physostigmine, but is resistant to both typical and the atypical APDs, haloperidol and clozapine. While the inefficacy of haloperidol can be expected based on its ineffectiveness in models of negative/cognitive symptoms including MK801-induced persistent LI, the inefficacy of clozapine is unexpected and sets this abnormality apart from MK801-induced LI persistence. It should be noted that scopolamine-induced persistent LI is the first instance of such an effect that is insensitive to atypical APDs, as reversal of persistent LI by clozapine has been shown for other NMDA antagonists, as well as for lesions and neurodevelopmental manipulations. Finally, haloperidol-induced persistent LI is reversed by atypical APDs but is resistant to glycine and physostigmine. On the basis of these pharmacological profiles, we tentatively suggest that the three persistent LI models relate to three domains of pathology in schizophrenia: NMDA antagonist-induced persistent LI taps (hypoglutamatergia-driven) negative/cognitive symptoms that respond to atypical APDs and cognitive enhancers but not to typical APDs. Scopolamine-induced persistent LI represents a domain of (antimuscarinicdriven) cognitive symptoms that are responsive to cognitive enhancers but are resistant to APDs. Given its selective sensitivity to cognitive enhancers, scopolamineinduced persistent LI may have considerable utility in detecting effective treatments for APD-resistant cognitive impairments in this disorder. However, given its insensitivity to APDs, scopolamine-induced persistent LI is likely to represent a class of behavioral inflexibility which is common to a variety of neuropsychiatric disorders, including schizophrenia, autism, addictive behavior, Parkinson disease, Alzheimer’s disease and obsessive compulsive disorder (OCD) (Chamberlain, Blackwell, Fineberg et al., 2005; Crider, 1997; Rapin & Katzman, 1998); interestingly, both PD and OCD patients display abnormally enhanced LI (Kaplan, Dar, Rosenthal et al., 2006; Lubow, Dressler, & Kaplan 1999; Swerdlow, Hartston, & Hartman, 1999). Finally, haloperidol-induced persistent LI represents a domain of (hypodopaminergia-driven) negative symptoms that are treatable by atypical antipsychotics but are resistant to cognitive enhancers. Haloperidol-induced persistent LI is also the only persistent LI that is responsive to amphetamine, just like negative symptoms in the clinic (Laruelle et al., 1999, 1996). We suggest that haloperidol-induced persistent LI represents a class of cognitive/behavioral inflexibility that is unique to schizophrenia. Although, responsive to atypical APDs, this domain of pathology may be seen as “old news”, it is important in two respects: the haloperidol-induced persistent LI model points out
The pharmacology of latent inhibition and its relevance to schizophrenia
303
limitations of cognitive enhancers for schizophrenia, at least as used to counter some forms of negative symptoms; it can be used to develop better and selective treatments for negative symptoms.
Conclusion Clearly, extensive work using additional compounds from all the relevant classes is necessary to substantiate the LI model as presented here, and such work is ongoing in our laboratory. However, we believe that the extant data are sufficient to provide a blueprint of a model that can answer many of the needs of drug development as they are seen now. According to Carpenter and Koenig (2008), “the needed paradigm identifies domains of pathology within the schizophrenia construct and proposes independent therapeutic development for each domain. . . . If successful, the field would evolve co-medication strategies with antipsychotic drugs for psychosis, antinegative symptom drugs for this pathology, cognition-enhancing drugs for cognition, and so forth”. In a similar vein, Gray and Roth (2007) write, “ the future of pharmacologic treatment of schizophrenia will likely start with the continued use of polypharmacy and augmentation strategies aimed at treating the multiple symptom domains of schizophrenia. This may be followed by the development of selectively nonselective single compounds that can target multiple domains at once”. Although such new directions are typically linked to “new approaches to drug evaluations and proof of concept testing” or “shift in our approach to drug development”, we believe that behavioral pharmacology models should not be abandoned but rather continually refined, and that the findings obtained using the LI model fit very well these directions.
References Abi-Dargham, A., Gil, R., Krystal, J., et al. (1998). Increased striatal dopamine transmission in schizophrenia: confirmation in a second cohort. American Journal of Psychiatry, 155, 761–767. Aguado, L., San Antonio, A., Perez, L., del Valle, R., & Gomez, J. (1994). Effects of the NMDA receptor antagonist ketamine on flavor memory: conditioned aversion, latent inhibition, and habituation of neophobia. Behavioral and Neural Biology, 61, 271–281. Andersen, M. B., Werge, T., & Fink-Jensen, A. (2007). The acetylcholinesterase inhibitor galantamine inhibits d-amphetamine-induced psychotic-like behavior in Cebus monkeys. The Journal of Pharmacology and Experimental Therapeutics, 321, 1179–1182. Angrist, B., Peselow, E., Rubinstein, M., Corwin, J., & Rotrosen, J. (1982). Partial improvement in negative schizophrenic symptoms after amphetamine. Psychopharmacology (Berl.), 78, 128–130. Arnt, J., & Skarsfeldt, T. (1998). Do novel antipsychotics have similar pharmacological characteristics? A review of the evidence. Neuropsychopharmacology, 18, 63–101.
304
Ina Weiner and Michal Arad
Bakshi, V. P., Swerdlow, N. R., & Geyer, M. A. (1994). Clozapine antagonizes phencyclidine-induced deficits in sensorimotor gating of the startle response. The Journal of Pharmacology and Experimental Therapeutics, 271, 787–794. Barak, S., Arad, M., De Levie, A., et al. (2009). Pro-cognitive and antipsychotic efficacy of the alpha7 nicotinic partial agonist SSR180711 in pharmacological and neurodevelopmental latent inhibition models of schizophrenia. Neuropsychopharmacology, 34, 1753–1763. Barak, S., & Weiner, I. (2007). Scopolamine induces disruption of latent inhibition which is prevented by antipsychotic drugs and an acetylcholinesterase inhibitor. Neuropsychopharmacology, 32, 989–999. Barak, S., & Weiner, I. (2009). Towards an animal model of an antipsychotic drug-resistant cognitive impairment in schizophrenia: scopolamine induces abnormally persistent latent inhibition, which can be reversed by cognitive enhancers but not by antipsychotic drugs. The International Journal of Neuropsychopharmacology, 12, 227–241. Barak, S., & Weiner, I. (2010). Differential role of muscarinic transmission within the entorhinal cortex and basolateral amygdala in the processing of irrelevant stimuli. Neuropsychopharmacology, Jan 13. [Epub ahead of print]. Barrett, S. L., Bell, R., Watson, D., & King, D. J. (2004). Effects of amisulpride, risperidone and chlorpromazine on auditory and visual latent inhibition, prepulse inhibition, executive function and eye movements in healthy volunteers. Journal of Psychopharmacology, 18, 156–172. Baruch, I., Hemsley, D. R., & Gray, J. A. (1988a). Differential performance of acute and chronic schizophrenics in a latent inhibition task. The Journal of Nervous and Mental Disease, 176, 598–606. Baruch, I., Hemsley, D. R., & Gray, J. A. (1988b). Latent inhibition and ‘psychotic proneness’ in normal subjects. Personality and Individual Differences, 9, 777–783. Bernstein, H. G., Bogerts, B., & Keilhoff, G. (2005). The many faces of nitric oxide in schizophrenia. A review. Schizophrenia Research, 78, 69–86. Bethus, I., Muscat, R., & Goodall, G. (2006). Dopamine manipulations limited to preexposure are sufficient to modulate latent inhibition. Behavioral Neuroscience, 120, 554–562. Bianchetti, A., Ranieri, P., Margiotta, A., & Trabucchi, M. (2006). Pharmacological treatment of Alzheimer’s disease. Aging Clinical and Experimental Research 18, 158–162. Biton, B., Bergis, O. E., Galli, F., et al. (2007). SSR180711, a novel selective alpha7 nicotinic receptor partial agonist: (1) binding and functional profile. Neuropsychopharmacology, 32, 1–16. Black, M. D., Selk, D. E., Hitchcock, J. M., Wettstein, J. G., & Sorensen, S. M. (1999). On the effect of neonatal nitric oxide synthase inhibition in rats: a potential neurodevelopmental model of schizophrenia. Neuropharmacology, 38, 1299–1306. Black, M. D., Simmonds, J., Senyah, Y., & Wettstein, J. G. (2002). Neonatal nitric oxide synthase inhibition: social interaction deficits in adulthood and reversal by antipsychotic drugs. Neuropharmacology, 42, 414–420. Black, M. D., Varty, G. B., Arad, M., et al. (2009). Procognitive and antipsychotic efficacy of glycine transport 1 inhibitors (GlyT1) in acute and
The pharmacology of latent inhibition and its relevance to schizophrenia
305
neurodevelopmental models of schizophrenia: latent inhibition studies in the rat. Psychopharmacology (Berl.), 202, 385–396. Bleuler, M. (1911). Dementia Praecox or the Group of Schizophrenias. New York: International Universities Press. Bolshakov, K. V., Gmiro, V. E., Tikhonov, D. B., & Magazanik, L. G. (2003). Determinants of trapping block of N-methyl-d-aspartate receptor channels. Journal of Neurochemistry, 87, 56–65. Boulay, D., Pichat, P., Dargazanli, G., et al. (2008). Characterization of SSR103800, a selective inhibitor of the glycine transporter-1 in models predictive of therapeutic activity in schizophrenia. Pharmacology, Biochemistry, and Behavior, 91, 47–58. Braunstein-Bercovitz, H., & Lubow, R. E. (1998). Are high-schizotypal normal participants distractible or limited in attentional resources? A study of latent inhibition as a function of masking task load and schizotypy level. Journal of Abnormal Psychology, 107, 659–670. Breier, A., & Berg, P. H. (1999). The psychosis of schizophrenia: prevalence, response to atypical antipsychotics, and prediction of outcome. Biological Psychiatry, 46, 361–364. Breier, A., Su, T. P., Saunders, R., et al. (1997). Schizophrenia is associated with elevated amphetamine-induced synaptic dopamine concentrations: evidence from a novel positron emission tomography method. Proceedings of the National Academy of Sciences of the United States of America, 94, 2569–2574. Broen, W. E. (1968). Schizophrenia: Research and Theory. New York: Academic Press. Broersen, L. M., Feldon, J., & Weiner, I. (1999). Dissociative effects of apomorphine infusions into the medial prefrontal cortex of rats on latent inhibition, prepulse inhibition and amphetamine-induced locomotion. Neuroscience, 94, 39–46. Brousseau, G., Rourke, B. P., & Burke, B. (2007). Acetylcholinesterase inhibitors, neuropsychiatric symptoms, and Alzheimer’s disease subtypes: an alternate hypothesis to global cognitive enhancement. Experimental and Clinical Psychopharmacology, 15, 546–554. Buchanan, R. W., Javitt, D. C., Marder, S. R., et al. (2007). The Cognitive and Negative Symptoms in Schizophrenia Trial (CONSIST): the efficacy of glutamatergic agents for negative symptoms and cognitive impairments. American Journal of Psychiatry, 164, 1593–1602. Bymaster, F. P., Felder, C., Ahmed, S., & McKinzie, D. (2002). Muscarinic receptors as a target for drugs treating schizophrenia. Current Drug Targets. CNS and Neurological Disorders, 1, 163–181. Carlsson, A., Waters, N., & Carlsson, M. L. (1999). Neurotransmitter interactions in schizophrenia – therapeutic implications. Biological Psychiatry, 46, 1388–1395. Carlsson, M., & Carlsson, A. (1990). Schizophrenia: a subcortical neurotransmitter imbalance syndrome? Schizophrenia Bulletin, 16, 425–432. Carnicella, S., Pain, L., & Oberling, P. (2005). Cholinergic effects on fear conditioning I: the degraded contingency effect is disrupted by atropine but reinstated by physostigmine. Psychopharmacology (Berl.), 178, 524–532. Carpenter, W. T., Jr., Heinrichs, D. W., & Alphs, L. D. (1985). Treatment of negative symptoms. Schizophrenia Bulletin, 11, 440–452. Carpenter, W. T., Jr., Heinrichs, D. W., & Wagman, A. M. (1988). Deficit and nondeficit forms of schizophrenia: the concept. American Journal of Psychiatry, 145, 578–583.
306
Ina Weiner and Michal Arad
Carpenter, W. T., & Koenig, J. I. (2008). The evolution of drug development in schizophrenia: past issues and future opportunities. Neuropsychopharmacology, 33, 2061–2079. Cassaday, H. F., Hodges, H., & Gray, J. A. (1993). The effects of ritanserin, RU 24969 and 8-OH-DPAT on latent inhibition in the rat. Psychopharmacology, 7, 63–71. Chagas-Martinich, L., Carey, R. J., & Carrera, M. P. (2007). 7-OH-DPAT effects on latent inhibition: low dose facilitation but high dose blockade: Implications for dopamine receptor involvement in attentional processes. Pharmacology, Biochemistry, and Behavior, 86, 441–448. Chamberlain, S. R., Blackwell, A. D., Fineberg, N. A., Robbins, T. W., & Sahakian, B. J. (2005). The neuropsychology of obsessive compulsive disorder: the importance of failures in cognitive and behavioural inhibition as candidate endophenotypic markers. Neuroscience and Biobehavioral Reviews, 29, 399–419. Chang, T., Meyer, U., Feldon, J., & Yee, B. K. (2007). Disruption of the US preexposure effect and latent inhibition in two-way active avoidance by systemic amphetamine in C57BL/6 mice. Psychopharmacology (Berl.), 191, 211–221. Chen, K. C., Baxter, M. G., & Rodefer, J. S. (2004). Central blockade of muscarinic cholinergic receptors disrupts affective and attentional set-shifting. The European Journal of Neuroscience, 20, 1081–1088. Clarke, L. A., Cassidy, C. W., Catalano, G., Catalano, M. C., & Carroll, K. M. (2004). Psychosis induced by smoking cessation clinic administered anticholinergic overload. Annals of Clinical Psychiatry 16, 171–175. Cohen, E., Sereni, N., Kaplan, O., et al. (2004). The relation between latent inhibition and symptom-types in young schizophrenics. Behavioural Brain Research, 149, 113–122. Crider, A. (1997). Perseveration in schizophrenia. Schizophrenia Bulletin, 23, 3–74. Csernansky, J. G., Martin, M., Shah, R., et al. (2005). Cholinesterase inhibitors ameliorate behavioral deficits induced by MK-801 in mice. Neuropsychopharmacology, 30, 2135–2143. Dall’Olio, R., & Gandolfi, O. (1993). The NMDA positive modulator D-cycloserine potentiates the neuroleptic activity of D1 and D2 dopamine receptor blockers in the rat. Psychopharmacology (Berl.), 110, 165–168. Dao-Castellana, M. H., Paillere-Martinot, M. L., Hantraye, P., et al. (1997). Presynaptic dopaminergic function in the striatum of schizophrenic patients. Schizophrenia Research, 23, 167–174. De Levie, A., & Weiner, I. (2007a). Behavioral and pharmacological changes after neonatal nitric oxide inhibition: a model of negative symptoms. European Neuropsychopharmacology, 17(supp 4), S262. De Levie, A., & Weiner, I. (2007b). Perseverative behavior after neonatal nitric oxide blockade is alleviated following enhanced NMDA but not DA function. European Neuropsychopharmacology, 17(supp 1), S60. Della Casa, V., Hofer, I., Weiner, I., & Feldon, J. (1999). Effects of smoking status and schizotypy on latent inhibition. Journal of Psychopharmacology, 13, 45–57. Depoortere, R., Dargazanli, G., Estenne-Bouhtou, G., et al. (2005). Neurochemical, electrophysiological and pharmacological profiles of the selective inhibitor of the glycine transporter-1 SSR504734, a potential new type of antipsychotic. Neuropsychopharmacology, 30, 1963–1985.
The pharmacology of latent inhibition and its relevance to schizophrenia
307
Dunn, L. A., Atwater, G. E., & Kilts, C. D. (1993). Effects of antipsychotic drugs on latent inhibition: sensitivity and specificity of an animal behavioral model of clinical drug action. Psychopharmacology (Berl.), 112, 315–323. Escobar, M., Oberling, P., & Miller, R. R. (2002). Associative deficit accounts of disrupted latent inhibition and blocking in schizophrenia. Neuroscience and Biobehavioral Reviews, 26, 203–216. Everitt, B. J., & Robbins, T. W. (1997). Central cholinergic systems and cognition. Annual Review of Psychology, 48, 649–684. Feldon, J., Shofel, A., & Weiner, I. (1991). Latent inhibition is unaffected by direct dopamine agonists. Pharmacology, Biochemistry, and Behavior, 38, 309–314. Feldon, J., & Weiner, I. (1991). The latent inhibition model of schizophrenic attention disorder. Haloperidol and sulpiride enhance rats’ ability to ignore irrelevant stimuli. Biological Psychiatry, 29, 635–646. Friedman, J. I. (2004). Cholinergic targets for cognitive enhancement in schizophrenia: focus on cholinesterase inhibitors and muscarinic agonists. Psychopharmacology (Berl.), 174, 45–53. Frith, C. D. (1979). Consciousness, information processing and schizophrenia. The British Journal of Psychiatry, 134, 225–235. Gaisler-Salomon, I., Diamant, L., Rubin, C., & Weiner, I. (2008). Abnormally persistent latent inhibition induced by MK801 is reversed by risperidone and by positive modulators of NMDA receptor function: differential efficacy depending on the stage of the task at which they are administered. Psychopharmacology (Berl.), 196, 255–267. Gaisler-Salomon, I., & Weiner, I. (2003). Systemic administration of MK-801 produces an abnormally persistent latent inhibition which is reversed by clozapine but not haloperidol. Psychopharmacology (Berl.), 166, 333–342. Gal, G., Barnea, Y., Biran, L., et al. (2009). Enhancement of latent inhibition in patients with chronic schizophrenia. Behavioural Brain Research, 197, 1–8. Goff, D. C., Herz, L., Posever, T., et al. (2005). A six-month, placebo-controlled trial of D-cycloserine co-administered with conventional antipsychotics in schizophrenia patients. Psychopharmacology (Berl.), 179, 144–150. Goff, D. C., Tsai, G., Manoach, D. S., & Coyle, J. T. (1995). Dose-finding trial of D-cycloserine added to neuroleptics for negative symptoms in schizophrenia. American Journal of Psychiatry, 152, 1213–1215. Gopel, C., Laufer, C., & Marcus, A. (2002). Three cases of angel’s trumpet teainduced psychosis in adolescent substance abusers. Nordic Journal of Psychiatry, 56, 49–52. Gosselin, G., Oberling, P., & Di Scala, G. (1996). Antagonism of amphetamineinduced disruption of latent inhibition by the atypical antipsychotic olanzapine in rats. Behavioural Pharmacology, 7, 820–826. Goto, Y., & Grace, A. A. (2007). The dopamine system and the pathophysiology of schizophrenia: a basic science perspective. International Review of Neurobiology. 78C, 41–68. Grace, A. A. (1991). Phasic versus tonic dopamine release and the modulation of dopamine system responsivity: a hypothesis for the etiology of schizophrenia. Neuroscience, 41, 1–24. Gracey, D. J., Bell, R., & King, D. J. (2002). Differential effects of the CCKA receptor ligands PD-140,548 and A-71623 on latent inhibition in the rat. Progress in Neuro-Psychopharmacology and Biological Psychiatry, 26, 497–504.
308
Ina Weiner and Michal Arad
Granacher, R. P., & Baldessarini, R. J. (1975). Physostigmine. Its use in acute anticholinergic syndrome with antidepressant and antiparkinson drugs. Archives of General Psychiatry, 32, 375–380. Gray, J. A., Feldon, J., Rawlins, J. N. P., Hemsley, D. R., & Smith, A. D. (1991). The neuropsychology of schizophrenia. Behavioral and Brain Sciences, 14, 1–20. Gray, J. A., Moran, P. M., Grigoryan, G., et al. (1997). Latent inhibition: the nucleus accumbens connection revisited. Behavioural Brain Research, 88, 27–34. Gray, J. A., & Roth, B. L. (2007). The pipeline and future of drug development in schizophrenia. Molecular Psychiatry, 12, 904–922. Gray, N. S., Hemsley, D. R., & Gray, J. A. (1992). Abolition of latent inhibition in acute, but not chronic schizophrenics. Neurology Psychiatry and Brain Research, 1, 83–89. Gray, N. S., Pickering, A. D., Hemsley, D. R., Dawling, S., & Gray, J. A. (1992). Abolition of latent inhibition by a single 5 mg dose of d-amphetamine in man. Psychopharmacology (Berl.), 107, 425–430. Gray, N. S., Pilowsky, L. S., Gray, J. A., & Kerwin, R. W. (1995). Latent inhibition in drug naive schizophrenics: relationship to duration of illness and dopamine D2 binding using SPET. Schizophrenia Research, 17, 95–107. Guillin, O., Abi-Dargham, A., & Laruelle, M. (2007). Neurobiology of dopamine in schizophrenia. International Review of Neurobiology, 78, 1–39. Hajos, M., Hurst, R. S., Hoffmann, W. E., et al. (2005). The selective alpha7 nicotinic acetylcholine receptor agonist PNU-282987 [N-[(3R)-1-Azabicyclo[2.2.2]oct3-yl]-4-chlorobenzamide hydrochloride] enhances GABAergic synaptic activity in brain slices and restores auditory gating deficits in anesthetized rats. The Journal of Pharmacology and Experimental Therapeutics, 312, 1213–1222. Harsing, L. G., Jr., Gacsalyi, I., Szabo, G., et al. (2003). The glycine transporter-1 inhibitors NFPS and Org 24461: a pharmacological study. Pharmacology, Biochemistry, and Behavior, 74, 811–825. Harvey, P. D., Rabinowitz, J., Eerdekens, M., & Davidson, M. (2005). Treatment of cognitive impairment in early psychosis: a comparison of risperidone and haloperidol in a large long-term trial. The American Journal of Psychiatry, 162, 1888–1895. Hashimoto, K., Koike, K., Shimizu, E., & Iyo, M. (2005). Alpha7 nicotinic receptor agonists as potential therapeutic drugs for schizophrenia. Current Medicinal Chemistry Central Nervous System Agents, 5, 171. Hasselmo, M. E., & McGaughy, J. (2004). High acetylcholine levels set circuit dynamics for attention and encoding and low acetylcholine levels set dynamics for consolidation. Progress in Brain Research, 145, 207–231. Hietala, J., Syvalahti, E., Vilkman, H., et al. (1999). Depressive symptoms and presynaptic dopamine function in neuroleptic-naive schizophrenia. Schizophrenia Research, 35, 41–50. Hoehn-Saric, R., McLeod, D. R., & Glowa, J. R. (1991). The effects of NMDA receptor blockade on the acquisition of a conditioned emotional response. Biological Psychiatry, 30, 170–176. Hohnadel, E., Bouchard, K., & Terry, A. V., Jr. (2007). Galantamine and donepezil attenuate pharmacologically induced deficits in prepulse inhibition in rats. Neuropharmacology, 52, 542–551. Hyde, T. M., & Crook, J. M. (2001). Cholinergic systems and schizophrenia: primary pathology or epiphenomena? Journal of Chemical Neuroanatomy, 22, 53–63.
The pharmacology of latent inhibition and its relevance to schizophrenia
309
Ichikawa, J., Chung, Y. C., Li, Z., Dai, J., & Meltzer, H. Y. (2002). Cholinergic modulation of basal and amphetamine-induced dopamine release in rat medial prefrontal cortex and nucleus accumbens. Brain Research, 958, 176–184. Ikemoto, S., & Panksepp, J. (1999). The role of nucleus accumbens dopamine in motivated behavior: a unifying interpretation with special reference to reward-seeking. Brain Research. Brain Research Reviews, 31, 6–41. Javitt, D. C. (2002). Glycine modulators in schizophrenia. Current Opinion in Investigational Drugs, 3, 1067–1072. Javitt, D. C. (2008). Glycine transport inhibitors and the treatment of schizophrenia. Biological Psychiatry, 63, 6–8. Javitt, D. C., Balla, A., Sershen, H., & Lajtha, A. (1999). Reversal of phencyclidineinduced effects by glycine and glycine transport inhibitors. Biological Psychiatry, 45, 668–679. Javitt, D. C., Sershen, H., Hashim, A., & Lajtha, A. (1997). Reversal of phencyclidine-induced hyperactivity by glycine and the glycine uptake inhibitor glycyldodecylamide. Neuropsychopharmacology, 17, 202–204. Javitt, D. C., & Zukin, S. R. (1991). Recent advances in the phencyclidine model of schizophrenia. The American Journal of Psychiatry, 148, 1301–1308. Jentsch, J. D., & Roth, R. H. (1999). The neuropsychopharmacology of phencyclidine: from NMDA receptor hypofunction to the dopamine hypothesis of schizophrenia. Neuropsychopharmacology, 20, 201–225. Jentsch, J. D., & Taylor, J. R. (2001). Impaired inhibition of conditioned responses produced by subchronic administration of phencyclidine to rats. Neuropsychopharmacology, 24, 66–74. Jones, C. K., Eberle, E. L., Shaw, D. B., McKinzie, D. L., & Shannon, H. E. (2005). Pharmacologic interactions between the muscarinic cholinergic and dopaminergic systems in the modulation of prepulse inhibition in rats. The Journal of Pharmacology and Experimental Therapeutics, 312, 1055–1063. Jones, C. K., & Shannon, H. E. (2000). Muscarinic cholinergic modulation of prepulse inhibition of the acoustic startle reflex. The Journal of Pharmacology and Experimental Therapeutics, 294, 1017–1023. Joseph, M. H., Peters, S. L., & Gray, J. A. (1993). Nicotine blocks latent inhibition in rats: evidence for a critical role of increased functional activity of dopamine in the mesolimbic system at conditioning rather than pre-exposure. Psychopharmacology (Berl.), 110, 187–192. Joseph, M. H., Peters, S. L., Moran, P. M., et al. (2000). Modulation of latent inhibition in the rat by altered dopamine transmission in the nucleus accumbens at the time of conditioning. Neuroscience, 101, 921–930. Kane, J. M. (1995). Current problems with the pharmacotherapy of schizophrenia. Clinical Neuropharmacology. 18, S154–S161. Kaplan, O., Dar, R., Rosenthal, L., et al. (2006). Obsessive-compulsive disorder patients display enhanced latent inhibition on a visual search task. Behaviour Research and Therapy, 44, 1137–1145. Kapur, S. (2003). Psychosis as a state of aberrant salience: a framework linking biology, phenomenology, and pharmacology in schizophrenia. The American Journal of Psychiatry, 160, 13–23. Karan, R. S., Ravishankar, P., & Pandhi, P. (2000). Effect of muscarinic receptor agonists on animal models of psychosis. Methods and Findings in Experimental and Clinical Pharmacology, 22, 169–172.
310
Ina Weiner and Michal Arad
Karnath, H. O., & Wallesch, C. W. (1992). Inflexibility of mental planning: a characteristic disorder with prefrontal lobe lesions? Neuropsychologia, 30, 1011–1016. Kato, K., Shishido, T., Ono, M., et al. (2001). Glycine reduces novelty- and methamphetamine-induced locomotor activity in neonatal ventral hippocampal damaged rats. Neuropsychopharmacology, 24, 330–332. Killcross, A. S., Dickinson, A., & Robbins, T. W. (1994a). Amphetamine-induced disruptions of latent inhibition are reinforcer mediated: implications for animal models of schizophrenic attentional dysfunction. Psychopharmacology (Berl.), 115, 185–195. Killcross, A. S., Dickinson, A., & Robbins, T. W. (1994b). Effects of the neuroleptic alpha-flupenthixol on latent inhibition in aversively- and appetitively-motivated paradigms: evidence for dopamine-reinforcer interactions. Psychopharmacology (Berl.), 115, 196–205. Killcross, A. S., & Robbins, T. W. (1993). Differential effects of intra-accumbens and systemic amphetamine on latent inhibition using an on-baseline, within-subject conditioned suppression paradigm. Psychopharmacology (Berl.), 110, 479–489. Kinon, B. J., & Lieberman, J. A. (1996). Mechanisms of action of atypical antipsychotic drugs: a critical analysis. Psychopharmacology (Berl.), 124, 2–34. Krystal, J. H., Bennett, A., Abi-Saab, D., et al. (2000). Dissociation of ketamine effects on rule acquisition and rule implementation: possible relevance to NMDA receptor contributions to executive cognitive functions. Biological Psychiatry, 47, 137–143. Krystal, J. H., D’Souza, D. C., Mathalon, D., et al. (2003). NMDA receptor antagonist effects, cortical glutamatergic function, and schizophrenia: toward a paradigm shift in medication development. Psychopharmacology (Berl.), 169, 215–233. Kumari, V., Aasen, I., Ffytche, D., Williams, S. C., & Sharma, T. (2006). Neural correlates of adjunctive rivastigmine treatment to antipsychotics in schizophrenia: a randomized, placebo-controlled, double-blind fMRI study. Neuroimage, 29, 545–556. Labrie, V., Lipina, T., & Roder, J. C. (2008). Mice with reduced NMDA receptor glycine affinity model some of the negative and cognitive symptoms of schizophrenia. Psychopharmacology (Berl.), 200, 217–230. Lahti, A. C., Koffel, B., LaPorte, D., & Tamminga, C. A. (1995). Subanesthetic doses of ketamine stimulate psychosis in schizophrenia. Neuropsychopharmacology, 13, 9–19. Laruelle, M., Abi-Dargham, A., Gil, R., Kegeles, L., & Innis, R. (1999). Increased dopamine transmission in schizophrenia: relationship to illness phases. Biological Psychiatry, 46, 56–72. Laruelle, M., Abi-Dargham, A., van Dyck, C. H., et al. (1996). Single photon emission computerized tomography imaging of amphetamine-induced dopamine release in drug-free schizophrenic subjects. Proceedings of the National Academy of Sciences of the United States of America, 93, 9235–9240. Lee, S. W., Lee, J. G., Lee, B. J., & Kim, Y. H. (2007). A 12-week, double-blind, placebo-controlled trial of galantamine adjunctive treatment to conventional antipsychotics for the cognitive impairments in chronic schizophrenia. International Clinical Psychopharmacology, 22, 63–68.
The pharmacology of latent inhibition and its relevance to schizophrenia
311
Levin, E. D., Bettegowda, C., Blosser, J., & Gordon, J. (1999). AR-R17779, and alpha7 nicotinic agonist, improves learning and memory in rats. Behavioural Pharmacology, 10, 675–680. Leysen, J. E., Janssen, P. M., Schotte, A., Luyten, W. H., & Megens, A. A. (1993). Interaction of antipsychotic drugs with neurotransmitter receptor sites in vitro and in vivo in relation to pharmacological and clinical effects: role of 5HT2 receptors. Psychopharmacology (Berl.), 112(1 Suppl), S40–54. Lieberman, J. A., Tollefson, G. D., Charles, C., et al. (2005). Antipsychotic drug effects on brain morphology in first-episode psychosis. Archives of General Psychiatry, 62, 361–370. Lipina, T., Labrie, V., Weiner, I., & Roder, J. (2005). Modulators of the glycine site on NMDA receptors, D-serine and ALX 5407, display similar beneficial effects to clozapine in mouse models of schizophrenia. Psychopharmacology (Berl.), 179, 54–67. Lipp, O. V., & Vaitl, D. (1992). Latent inhibition in human Pavlovian differential conditioning – effect of addiitional stimulation after preexposure and relation to schizotypal traits. Personality and Individual Differences, 13, 1003–1012. Lubow, R. E. (2005). Construct validity of the animal latent inhibition model of selective attention deficits in schizophrenia. Schizophrenia Bulletin, 31, 139–153. Lubow, R. E., & De la Casa, G. (2002). Latent inhibition as a function of schizotypality and gender: Implications for schizophrenia. Biological Psychiatry, 59, 69–86. Lubow, R. E., Dressler, R., & Kaplan, O. (1999). The effects of target and distractor familiarity on visual search in de novo Parkinson’s disease patients: latent inhibition and novel pop-out. Neuropsychology, 13, 415–423. Lubow, R. E., Ingbergsachs, Y., Zalsteinorda, N., & Gewirtz, J. C. (1992). Latent inhibition in low and high “psychotic-prone” normal subjects. Personality and Individual Differences, 13, 563–572. Lubow, R. E., Kaplan, O., Abramovich, P., Rudnick, A., & Laor, N. (2000). Visual search in schizophrenia: latent inhibition and novel pop-out effects. Schizophrenia Research, 45, 145–156. Lubow, R. E., Kaplan, O., & De la Casa, G. (2001). Performance on the visual search analog of latent inhibition is modulated by an interaction between schizotypy and gender. Schizophrenia Research, 52, 275–287. Lyon, M. (1991). Animal models of mania and schizophrenia. In P. Willner (Ed.), Behavioral Models in Psychopharmacology: Theoretical, Industrial and Clinical Perspectives. Cambridge: Cambridge University Press, pp. 253–310. Malhotra, A. K., Adler, C. M., Kennison, S. D., et al. (1997). Clozapine blunts N-methyl-D-aspartate antagonist-induced psychosis: a study with ketamine. Biological Psychiatry, 42, 664–668. Malhotra, A. K., Pinals, D. A., Adler, C. M., et al. (1997). Ketamine-induced exacerbation of psychotic symptoms and cognitive impairment in neurolepticfree schizophrenics. Neuropsychopharmacology, 17, 141–150. Marder, S. R., & Fenton, W. (2004). Measurement and Treatment Research to Improve Cognition in Schizophrenia: NIMH MATRICS initiative to support the development of agents for improving cognition in schizophrenia. Schizophrenia Research, 72, 5–9. Marino, M. J., Rouse, S. T., Levey, A. I., Potter, L. T., & Conn, P. J. (1998). Activation of the genetically defined m1 muscarinic receptor potentiates
312
Ina Weiner and Michal Arad
N-methyl-D-aspartate (NMDA) receptor currents in hippocampal pyramidal cells. Proceedings of the National Academy of Sciences of the United States of America, 95, 11465–11470. Martin, L. F., Kem, W. R., & Freedman, R. (2004). Alpha-7 nicotinic receptor agonists: potential new candidates for the treatment of schizophrenia. Psychopharmacology (Berl.) 174, 54–64. Masson, S., Avanzi, V., Troncoso, A. C., & Brandao, M. L. (2003). Effects of apomorphine and clozapine on conditioned freezing and latent inhibition. Progress in Neuro-Psychopharmacology and Biological Psychiatry, 27, 935–943. Mastropaolo, J., Rosse, R. B., & Deutsch, S. I. (2004). Anabasine, a selective nicotinic acetylcholine receptor agonist, antagonizes MK-801-elicited mouse popping behavior, an animal model of schizophrenia. Behavioural Brain Research, 153, 419–422. Mathur, A., Shandarin, A., LaViolette, S. R., Parker, J., & Yeomans, J. S. (1997). Locomotion and stereotypy induced by scopolamine: contributions of muscarinic receptors near the pedunculopontine tegmental nucleus. Brain Research, 775, 144–155. McCartan, D., Bell, R., Green, J. F., et al. (2001). The differential effects of chlorpromazine and haloperidol on latent inhibition in healthy volunteers. Journal of Psychopharmacology, 15, 96–104. Meltzer, H. Y., & Nash, J. F. (1991). Effects of antipsychotic drugs on serotonin receptors. Pharmacological Reviews, 43, 587–604. Meltzer, H. Y., & Sumiyoshi, T. (2003). Atypical antipsychotic drugs improve cognition in schizophrenia. Biological Psychiatry, 53, 265–267; author reply 267–268. Millan, M. J. (2002). N-methyl-D-aspartate receptor-coupled glycineB receptors in the pathogenesis and treatment of schizophrenia: a critical review. Current Drug Targets. CNS and Neurological Disorders, 1, 191–213. Millan, M. J. (2005). N-Methyl-D-aspartate receptors as a target for improved antipsychotic agents: novel insights and clinical perspectives. Psychopharmacology (Berl.), 179, 30–53. Millan, M. J., Brocco, M., Gobert, A., et al. (1999). Contrasting mechanisms of action and sensitivity to antipsychotics of phencyclidine versus amphetamine: importance of nucleus accumbens 5-HT2A sites for PCP-induced locomotion in the rat. The European Journal of Neuroscience, 11, 4419–4432. Miyamoto, S., Duncan, G. E., Marx, C. E., & Lieberman, J. A. (2005). Treatments for schizophrenia: a critical review of pharmacology and mechanisms of action of antipsychotic drugs. Molecular Psychiatry, 10, 79–104. Moghaddam, B., Adams, B., Verma, A., & Daly, D. (1997). Activation of glutamatergic neurotransmission by ketamine: a novel step in the pathway from NMDA receptor blockade to dopaminergic and cognitive disruptions associated with the prefrontal cortex. The Journal of Neuroscience, 17(8), 2921–2927. Moghaddam, B., & Jackson, M. E. (2003). Glutamatergic animal models of schizophrenia. Annals of the New York Academy of Sciences, 1003, 131–137. Moller, H. J. (2003). Management of the negative symptoms of schizophrenia: New treatment options. CNS Drugs, 17, 793–823. Moran, P. M., Fischer, T. R., Hitchcock, J. M., & Moser, P. C. (1996). Effects of clozapine on latent inhibition in the rat. Behavioural Pharmacology, 7, 42–48.
The pharmacology of latent inhibition and its relevance to schizophrenia
313
Moser, P. C., Hitchcock, J. M., Lister, S., & Moran, P. M. (2000). The pharmacology of latent inhibition as an animal model of schizophrenia. Brain Research. Brain Research Reviews, 33, 275–307. Murphy, C. A., Di Iorio, L., & Feldon, J. (2001). Effects of psychostimulant withdrawal on latent inhibition of conditioned active avoidance and prepulse inhibition of the acoustic startle response. Psychopharmacology (Berl.), 156, 155–164. Nogue, S., Sanz, P., Munne, P., & de la Torre, R. (1991). Acute scopolamine poisoning after sniffing adulterated cocaine. Drug and Alcohol Dependence, 27, 115–116. Noren, U., Bjorner, A., Sonesson, O., & Eriksson, L. (2006). Galantamine added to antipsychotic treatment in chronic schizophrenia: cognitive improvement? Schizophrenia Research, 85, 302–304. Ogura, C., Kishimoto, A., & Nakao, T. (1976). Clinical effect of L-dopa on schizophrenia. Current Therapeutic Research, Clinical and Experimental, 20, 308–318. Ohno, M., & Watanabe, S. (1996). D-cycloserine, a glycine site agonist, reverses working memory failure by hippocampal muscarinic receptor blockade in rats. European Journal of Pharmacology, 318, 267–271. Olincy, A., Harris, J. G., Johnson, L. L., et al. (2006). Proof-of-concept trial of an alpha7 nicotinic agonist in schizophrenia. Archives of General Psychiatry, 63, 630–638. Olincy, A., & Stevens, K. E. (2007). Treating schizophrenia symptoms with an alpha7 nicotinic agonist, from mice to men. Biochemical Pharmacology, 74, 1192–1201. Palsson, E., Klamer, D., Wass, C., et al. (2005). The effects of phencyclidine on latent inhibition in taste aversion conditioning: differential effects of preexposure and conditioning. Behavioural Brain Research, 157, 139–146. Payne, R. W. (1966). The measurement and significance of overinclusive thinking and retardation in schizophrenic patients. In P. Hoch & J. Zubin (Eds.), Psychopathology of Schizophrenia. New York: Grune and Stratton. Pennartz, C. M., Groenewegen, H. J., & Lopes da Silva, F. H. (1994). The nucleus accumbens as a complex of functionally distinct neuronal ensembles: an integration of behavioural, electrophysiological and anatomical data. Progress in Neurobiology, 42, 719–761. Perry, P. J., Wilding, D. C., & Juhl, R. P. (1978). Anticholinergic psychosis. American Journal of Hospital Pharmacy, 35, 725–728. Peters, S. L., & Joseph, M. H. (1993). Haloperidol potentiation of latent inhibition in rats: evidence for a critical role at conditioning rather than pre-exposure. Behavioural Pharmacology, 4, 183–186. Pichat, P., Bergis, O. E., Terranova, J. P., et al. (2007). SSR180711, a novel selective alpha7 nicotinic receptor partial agonist: (II) efficacy in experimental models predictive of activity against cognitive symptoms of schizophrenia. Neuropsychopharmacology, 32, 17–34. Powell, S. B., & Geyer, M. A. (2007). Overview of animal models of schizophrenia. Current Protocols in Neuroscience. Chapter 9, Unit 9, 24. Ragozzino, M. E., Jih, J., & Tzavos, A. (2002). Involvement of the dorsomedial striatum in behavioral flexibility: role of muscarinic cholinergic receptors. Brain Research, 953, 205–214.
314
Ina Weiner and Michal Arad
Ransom, R. W., & Deschenes, N. L. (1989). Glycine modulation of NMDA-evoked release of [3H]acetylcholine and [3H]dopamine from rat striatal slices. Neuroscience Letters, 96, 323–328. Rapin, I., & Katzman, R. (1998). Neurobiology of autism. Annals of Neurology, 43, 7–14. Rascle, C., Mazas, O., Vaiva, G., et al. (2001). Clinical features of latent inhibition in schizophrenia. Schizophrenia Research, 51, 149–161. Rezvani, A. H., & Levin, E. D. (2003). Nicotinic-glutamatergic interactions and attentional performance on an operant visual signal detection task in female rats. European Journal of Pharmacology, 465, 83–90. Robbins, T. W. (1991). Cognitive deficits in schizophrenia and Parkinson’s disease – neural basis and the role of dopamine. In P. J. Willner & J. Scheel-Kruger (Eds.), The Mesolimbic Dopamine System – From Motivation to Action. Chichester: Wiley. Robinson, G. B., Port, R. L., & Stillwell, E. G. (1993). Latent inhibition of the classically conditioned rabbit nictitating membrane response is unaffected by the NMDA antagonist MK-801. Psychobiology, 21, 120–124. Rochford, J., Sen, A. P., & Quirion, R. (1996). Effect of nicotine and nicotinic receptor agonists on latent inhibition in the rat. The Journal of Pharmacology and Experimental Therapeutics, 277, 1267–1275. Ruob, C., Elsner, J., Weiner, I., & Feldon, J. (1997). Amphetamine-induced disruption and haloperidol-induced potentiation of latent inhibition depend on the nature of the stimulus. Behavioural Brain Research, 88, 35–41. Russig, H., Kovacevic, A., Murphy, C. A., & Feldon, J. (2003). Haloperidol and clozapine antagonise amphetamine-induced disruption of latent inhibition of conditioned taste aversion. Psychopharmacology (Berl.), 170, 263–270. Russig, H., Murphy, C. A., & Feldon, J. (2002). Clozapine and haloperidol reinstate latent inhibition following its disruption during amphetamine withdrawal. Neuropsychopharmacology, 26, 765–777. Salgado, J. V., Hetem, L. A., Vidal, M., Graeff, F. G., Danion, J. M., & Sandner, G. (2000). Reduction of latent inhibition by D-amphetamine in a conditioned suppression paradigm in humans. Behavioural Brain Research, 117, 61–67. Sams-Dodd, F. (1996). Phencyclidine-induced stereotyped behaviour and social isolation in rats: a possible animal model of schizophrenia. Behavioural Pharmacology, 7, 3–23. Sanfilipo, M., Wolkin, A., Angrist, B., et al. (1996). Amphetamine and negative symptoms of schizophrenia. Psychopharmacology (Berl.), 123, 211–214. Sarter, M., Bruno, J. P., & Givens, B. (2003). Attentional functions of cortical cholinergic inputs: what does it mean for learning and memory? Neurobiology of Learning and Memory, 80, 245–256. Sarter, M., Nelson, C. L., & Bruno, J. P. (2005). Cortical cholinergic transmission and cortical information processing in schizophrenia. Schizophrenia Bulletin, 31, 117–138. Schotte, A., Janssen, P. F., Gommeren, W., et al. (1996). Risperidone compared with new and reference antipsychotic drugs: in vitro and in vivo receptor binding. Psychopharmacology (Berl.), 124, 57–73. Schubert, M. H., Young, K. A., & Hicks, P. B. (2006). Galantamine improves cognition in schizophrenic patients stabilized on risperidone. Biological Psychiatry, 60, 530–533.
The pharmacology of latent inhibition and its relevance to schizophrenia
315
Shadach, E., Feldon, J., & Weiner, I. (1999). Clozapine-induced potentiation of latent inhibition is due to its action in the conditioning stage: implications for the mechanism of action of antipsychotic drugs. The International Journal of Neuropsychopharmacology, 2, 283–291. Shadach, E., Gaisler, I., Schiller, D., & Weiner, I. (2000). The latent inhibition model dissociates between clozapine, haloperidol, and ritanserin. Neuropsychopharmacology, 23, 151–161. Shakow, D. (1962). Segmental set. Archives of General Psychiatry, 6, 1–17. Shannon, H. E., & Peters, S. C. (1990). A comparison of the effects of cholinergic and dopaminergic agents on scopolamine-induced hyperactivity in mice. The Journal of Pharmacology and Experimental Therapeutics, 255, 549–553. Shannon, H. E., Rasmussen, K., Bymaster, F. P., et al. (2000). Xanomeline, an M(1)/ M(4) preferring muscarinic cholinergic receptor agonist, produces antipsychotic-like activity in rats and mice. Schizophrenia Research, 42, 249–259. Sirvio, J., Ekonsalo, T., Riekkinen, P., Jr., Lahtinen, H., & Riekkinen, P., Sr. (1992). D-cycloserine, a modulator of the N-methyl-D-aspartate receptor, improves spatial learning in rats treated with muscarinic antagonist. Neuroscience Letters, 146, 215–218. Smith, A., Li, M., Becker, S., & Kapur, S. (2006). Dopamine, prediction error and associative learning: a model-based account. Network, 17, 61–84. Soffie, M., & Lamberty, Y. (1987). Scopolamine disrupts visual reversal without affecting the first discrimination. Physiology and Behavior, 40, 263–265. Solomon, P. R., Crider, A., Winkelman, J. W., et al. (1981). Disrupted latent inhibition in the rat with chronic amphetamine or haloperidol-induced supersensitivity: relationship to schizophrenic attention disorder. Biological Psychiatry, 16, 519–537. Stanhope, K. J., Mirza, N. R., Bickerdike, M. J., et al. (2001). The muscarinic receptor agonist xanomeline has an antipsychotic-like profile in the rat. The Journal of Pharmacology and Experimental Therapeutics, 299, 782–792. Stone, W. S., Rudd, R. J., & Gold, P. E. (1990). Glucose and physostigmine effects on morphine- and amphetamine-induced increases in locomotor activity in mice. Behavioral and Neural Biology, 54, 146–155. Svensson, T. H. (2000). Dysfunctional brain dopamine systems induced by psychotomimetic NMDA-receptor antagonists and the effects of antipsychotic drugs. Brain Research. Brain Research Reviews, 31, 320–329. Swerdlow, N. R., Braff, D. L., Hartston, H., Perry, W., & Geyer, M. A. (1996). Latent inhibition in schizophrenia. Schizophrenia Research, 20, 91–103. Swerdlow, N. R., Hartston, H. J., & Hartman, P. L. (1999). Enhanced visual latent inhibition in obsessive-compulsive disorder. Biological Psychiatry, 45, 482–488. Swerdlow, N. R., & Koob, G. F. (1987). Dopamine, schizophrenia, mania and depression: toward a unified hypothesis of cortico-striato-pallido-thalamic function. Behavioral and Brain Sciences, 10, 197–245. Swerdlow, N. R., Stephany, N., Wasserman, L. C., et al. (2003). Dopamine agonists disrupt visual latent inhibition in normal males using a within-subject paradigm. Psychopharmacology (Berl.), 169, 314–320. Swerdlow, N. R., Stephany, N., Wasserman, L. C., et al. (2005). Intact visual latent inhibition in schizophrenia patients in a within-subject paradigm. Schizophrenia Research, 72, 169–183.
316
Ina Weiner and Michal Arad
Tamminga, C. A. (1998). Schizophrenia and glutamatergic transmission. Critical Reviews in Neurobiology, 12, 21–36. Tanii, Y., Nishikawa, T., Hashimoto, A., & Takahashi, K. (1994). Stereoselective antagonism by enantiomers of alanine and serine of phencyclidine-induced hyperactivity, stereotypy and ataxia in the rat. The Journal of Pharmacology and Experimental Therapeutics, 269, 1040–1048. Tenn, C. C., Fletcher, P. J., & Kapur, S. (2005a). A putative animal model of the “prodromal” state of schizophrenia. Biological Psychiatry, 57, 586–593. Tenn, C. C., Kapur, S., & Fletcher, P. J. (2005b). Sensitization to amphetamine, but not phencyclidine, disrupts prepulse inhibition and latent inhibition. Psychopharmacology (Berl.), 180, 366–376. Thornton, J. C., Dawe, S., Lee, C., et al. (1996). Effects of nicotine and amphetamine on latent inhibition in human subjects. Psychopharmacology (Berl.), 127, 164–173. Timmermann, D. B., Gronlien, J. H., Kohlhaas, K. L., et al. (2007). An allosteric modulator of the alpha7 nicotinic acetylcholine receptor possessing cognitionenhancing properties in vivo. The Journal of Pharmacology and Experimental Therapeutics, 323, 294–307. Tinsley, M. R., Quinn, J. J., & Fanselow, M. S. (2004). The role of muscarinic and nicotinic cholinergic neurotransmission in aversive conditioning: comparing pavlovian fear conditioning and inhibitory avoidance. Learning and Memory, 11, 35–42. Toda, M., & Abi-Dargham, A. (2007). Dopamine hypothesis of schizophrenia: Making sense of it all. Current Psychiatry Reports, 9, 329–336. Toth, E., & Lajtha, A. (1986). Antagonism of phencyclidine-induced hyperactivity by glycine in mice. Neurochemical Research, 11, 393–400. Trimble, K. M., Bell, R., & King, D. J. (1997). Enhancement of latent inhibition in the rat by the atypical antipsychotic agent remoxipride. Pharmacology, Biochemistry, and Behavior, 56, 809–816. Trimble, K. M., Bell, R., & King, D. J. (1998). Enhancement of latent inhibition in the rat at a high dose of clozapine. Journal of Psychopharmacology, 12, 215–219. Trimble, K. M., Bell, R., & King, D. J. (2002). Effects of the selective dopamine D(1) antagonists NNC 01–0112 and SCH 39166 on latent inhibition in the rat. Physiology and Behavior, 77, 115–123. Turgeon, S. M., Auerbach, E. A., Duncan-Smith, M. K., George, J. R., & Graves, W. W. (2000). The delayed effects of DTG and MK-801 on latent inhibition in a conditioned taste-aversion paradigm. Pharmacology, Biochemistry, and Behavior, 66, 533–539. Turgeon, S. M., Auerbach, E. A., & Heller, M. A. (1998). The delayed effects of phencyclidine (PCP) disrupt latent inhibition in a conditioned taste aversion paradigm. Pharmacology, Biochemistry, and Behavior, 60, 553–558. Ukai, M., Okuda, A., & Mamiya, T. (2004). Effects of anticholinergic drugs selective for muscarinic receptor subtypes on prepulse inhibition in mice. European Journal of Pharmacology, 492, 183–187. Vaitl, D., Lipp, O., Bauer, U., et al. (2002). Latent inhibition and schizophrenia: Pavlovian conditioning of autonomic responses. Schizophrenia Research, 55, 147–158.
The pharmacology of latent inhibition and its relevance to schizophrenia
317
Vaitl, D., & Lipp, O. V. (1997). Latent inhibition and autonomic responses: a psychophysiological approach. Behavioural Brain Research, 88, 85–93. van der Meulen, J. A., Bilbija, L., Joosten, R. N., de Bruin, J. P., & Feenstra, M. G. (2003). The NMDA-receptor antagonist MK-801 selectively disrupts reversal learning in rats. Neuroreport, 14, 2225–2228. Viu, E., Zapata, A., Capdevila, J., Skolnick, P., & Trullas, R. (2000). Glycine(B) receptor antagonists and partial agonists prevent memory deficits in inhibitory avoidance learning. Neurobiology of Learning and Memory, 74, 146–160. Warburton, E. C., Joseph, M. H., Feldon, J., Weiner, I., & Gray, J. A. (1994). Antagonism of amphetamine-induced disruption of latent inhibition in rats by haloperidol and ondansetron: implications for a possible antipsychotic action of ondansetron. Psychopharmacology (Berl.), 114, 657–664. Weiner, I. (1990). Neural substrates of latent inhibition: the switching model. Psychological Bulletin, 108, 442–461. Weiner, I. (2003). The “two-headed” latent inhibition model of schizophrenia: modeling positive and negative symptoms and their treatment. Psychopharmacology (Berl.), 169, 257–297. Weiner, I., Bernasconi, E., Broersen, L. M., & Feldon, J. (1997a). Amphetamineinduced disruption of latent inhibition depends on the nature of the stimulus. Behavioural Pharmacology, 8, 442–457. Weiner, I., & Feldon, J. (1987). Facilitation of latent inhibition by haloperidol in rats. Psychopharmacology (Berl.), 91, 248–253. Weiner, I., & Feldon, J. (1992). Phencyclidine does not disrupt latent inhibition in rats: Implications for animal models of schizophrenia. Pharmacology, Biochemistry, and Behavior, 42, 625–631. Weiner, I., & Feldon, J. (1997). The switching model of latent inhibition: an update of neural substrates. Behavioural Brain Research, 88, 11–25. Weiner, I., Feldon, J., & Katz, Y. (1987). Facilitation of the expression but not the acquisition of latent inhibition by haloperidol in rats. Pharmacology, Biochemistry, and Behavior, 26, 241–246. Weiner, I., Izraeli-Telerant, A., & Feldon, J. (1987). Latent inhibition is not affected by acute or chronic administration of 6 mg/kg dl-amphetamine. Psychopharmacology (Berl.), 91, 345–351. Weiner, I., & Joel, D. (2002). Dopamine in schizophrenia: dysfunctional information processing in basal ganglia-thalamocortical split circuits. In G. Di Chiara (Ed.), Handbook of Experimental Pharmacology, vol. 54/II, Dopamine in the CNS II. Berlin: Springer, pp. 418–472. Weiner, I., Lubow, R. E., & Feldon, J. (1981). Chronic amphetamine and latent inhibition. Behavioural Brain Research, 2, 285–286. Weiner, I., Lubow, R. E., & Feldon, J. (1984). Abolition of the expression but not the acquisition of latent inhibition by chronic amphetamine in rats. Psychopharmacology (Berl.), 83, 194–199. Weiner, I., Lubow, R. E., & Feldon, J. (1988). Disruption of latent inhibition by acute administration of low doses of amphetamine. Pharmacology, Biochemistry, and Behavior, 30, 871–878. Weiner, I., Shadach, E., Barkai, R., & Feldon, J. (1997b). Haloperidol- and clozapine-induced enhancement of latent inhibition with extended conditioning:
318
Ina Weiner and Michal Arad
Implications for the mechanism of action of neuroleptic drugs. Neuropsychopharmacology, 16, 42–50. Weiner, I., Shadach, E., Tarrasch, R., Kidron, R., & Feldon, J. (1996). The latent inhibition model of schizophrenia: further validation using the atypical neuroleptic, clozapine. Biological Psychiatry, 40, 834–843. Weiner, I., Shofel, A., & Feldon, J. (1990). Disruption of latent inhibition by low dose of amphetamine is antagonized by haloperidol and apomorphine. Journal of Psychopharmacology, 4, 255. Weiner, I., Tarrasch, R., Bernasconi, E., et al. (1997c). Amphetamine-induced disruption of latent inhibition is not reinforcer-mediated. Pharmacology, Biochemistry, and Behavior, 56, 817–826. Williams, J. H., Wellman, N. A., Geaney, D. P., et al. (1996). Antipsychotic drug effects in a model of schizophrenic attentional disorder: a randomized controlled trial of the effects of haloperidol on latent inhibition in healthy people. Biological Psychiatry, 40, 1135–1143. Williams, J. H., Wellman, N. A., Geaney, D. P., et al. (1997). Haloperidol enhances latent inhibition in visual tasks in healthy people. Psychopharmacology (Berl.), 133, 262–268. Wishka, D. G., Walker, D. P., Yates, K. M., et al. (2006). Discovery of N-[(3R)-1-azabicyclo[2.2.2]oct-3-yl]furo[2,3-c]pyridine-5-carboxamide, an agonist of the alpha7 nicotinic acetylcholine receptor, for the potential treatment of cognitive deficits in schizophrenia: synthesis and structure–activity relationship. Journal of Medicinal Chemistry, 49, 4425–4436. Wolkin, A., Sanfilipo, M., Wolf, A. P., Angrist, B., Brodie, J. D., & Rotrosen, J. (1992). Negative symptoms and hypofrontality in chronic schizophrenia. Archives of General Psychiatry, 49, 959–965. Zuckerman, L., Rimmerman, N., & Weiner, I. (2003). Latent inhibition in 35-day-old rats is not an “adult” latent inhibition: implications for neurodevelopmental models of schizophrenia. Psychopharmacology (Berl.), 169, 298–307. Zuckerman, L., & Weiner, I. (2005). Maternal immune activation leads to behavioral and pharmacological changes in the adult offspring. Journal of Psychiatric Research, 39, 311–323.
14 Parahippocampal region–dopaminergic neuron relationships in latent inhibition A. Louilot, J. Jeanblanc, Y. Peterschmitt and F. Meyer
Historically, the suggestion that dopaminergic (DAergic) neurons are involved in the latent inhibition (LI) phenomenon is linked to psychopharmacological studies carried out by two laboratories reporting an attenuation of LI responses in animals (rats) treated with the indirect DAergic agonist d-amphetamine, after chronic (Solomon, Crider, Winkelman et al., 1981; Weiner, Lubow & Feldon, 1981, 1984) or acute administration (Weiner, Lubow & Feldon, 1988). Involvement of DAergic neurons in LI was further supported by data showing that the LI attenuation induced by d-amphetamine was reversed by the concomitant administration of the neuroleptic chlorpromazine (Solomon et al., 1981), and by subsequent studies showing a facilitation of LI expression after administration of haloperidol (Weiner & Feldon, 1987; Weiner, Feldon & Katz, 1987), a well-known typical neuroleptic with a potent blockade action on DA receptors. Since these first studies, the reversal of the d-amphetamine-induced LI reduction by DAergic antagonists has been found in different LI paradigms after administration of several atypical neuroleptics, including olanzapine (Gosselin, Oberling & Di Scala, 1996) and clozapine (Trimble, Bell & King, 1998; Russig, Murphy & Feldon, 2002). In other respects, enhancement of LI expression has also been reported, with the atypical antipsychotics displaying a selective blockade action on D2 receptors such as sulpiride (Feldon & Weiner, 1991) or remoxipride (Trimble, Bell & King, 1997), whereas selective D1 antagonists were found to have no effect on LI phenomenon (Trimble, Bell & King, 2002). Based on the psychopharmacological data, understanding how and to what extent the DAergic neurons are involved in LI has been somewhat complicated by the fact that it was initially reported in rats that LI expression was unaffected by direct DA agonists, namely apomorphine, a mixed D1/D2 agonist, administered during the preexposure and conditioning stages of a three-stage LI paradigm (Feldon, Shofel & Weiner, 1991). However, attenuation of LI expression with apomorphine and the reversal of this LI reduction by the neuroleptic chlorpromazine have been observed more recently in rats in another LI paradigm (Troncoso, Osaki, Mason et al., 2003). Latent Inhibition: Cognition, Neuroscience and Applications to Schizophrenia, ed. R. E. Lubow and Ina Weiner. Published by Cambridge University Press # Cambridge University Press 2010.
319
320
A. Louilot, J. Jeanblanc, Y. Peterschmitt and F. Meyer
Furthermore, bromocriptine, a direct DA agonist with a preferential D2 activity, has been found to disrupt LI in humans in a within-subject paradigm (Swerdlow, Stephany, Wasserman et al., 2003); in this context it is worth remembering that different authors (Gray, Pickering, Hemsley et al., 1992; Swerdlow et al., 2003) have observed the LI-disruptive effects of d-amphetamine in humans. The reasons for the contradictory results for the effects of the direct DA agonists are not clear and may have to do with aspects of the experimental conditions and/or procedure which remain to be clarified. Another difficulty stemmed from neuropsychopharmacological studies aimed at identifying, in terms of anatomy, the nature of the DAergic neuronal subgroups involved in the reduction of LI expression observed in studies dealing with peripheral d-amphetamine administration. Using local microinjections of d-amphetamine in the dorsal striatum or nucleus accumbens and measuring LI in a conditioned active avoidance task, Solomon and Staton (1982) suggested that only DAergic neurons innervating the nucleus accumbens are involved in LI. In contrast, Killcross and Robbins (1993), who performed similar d-amphetamine microinjections in the nucleus accumbens but measured LI in a different, conditioned suppression paradigm, and who used a within-subject procedure, found no evidence that DAergic terminals were involved in the nucleus accumbens in LI, although in their protocol they were able to observe an abolition of LI after d-amphetamine systemic injection. Using a third paradigm to measure LI, namely a conditioned taste aversion, Ellenbroek, Knobbout and Cools (1997) carried out a comparative study similar to that of Solomon and Staton (1982). Results obtained after intracerebral microinjections of d-amphetamine in the dorsal striatum or nucleus accumbens led Ellenbroek et al. (1997) to conclude that the DAergic terminals in the dorsal striatum are involved in LI expression, but not the DAergic terminal fields in the nucleus accumbens. The reasons for the discrepancies between the observations reported by Solomon and Staton (1982) and the other authors are not obvious and may be connected with the different paradigms used and/or the injection sites chosen in the different studies. The data showing that LI is disrupted in patients with acute schizophrenia (Baruch, Hemsley & Gray, 1988; Gray, Hemsley & Gray, 1992; Gray, Pilowsky, Gray & Kerwin, 1995; Lubow, Kaplan, Abramovich et al., 2000), on one hand, and the repeated suggestions that a striatal DAergic dysfunctioning is central to the pathophysiology of schizophrenia (e.g., Carlsson, Waters, Holm-Waters et al., 2001; Harrison, 1999; Swerdlow & Koob, 1987), on the other hand, highlighted a long time ago how interesting and important it is to continue defining the modalities of the involvement of mesencephalic DAergic neurons in the LI phenomenon. It soon became obvious that the neuropsychopharmacological approach would not be sufficient, and that an approach allowing for direct investigation of the involvement of DAergic neurons in LI was needed. This realization prompted studies using in vivo methods. The first study designed for the purpose was by Young, Joseph and Gray (1993).
Parahippocampal region–dopaminergic neuron relationships in latent inhibition
321
Using in vivo microdialysis in freely moving animals and a conditioned suppression paradigm to study LI, these authors reported that the conditioned DA increase observed in the nucleus accumbens at test in non-pre-exposed rats after presentation of the conditioned stimulus was not obtained in pre-exposed animals. By the mid 1990s it was clear, however, that the nucleus accumbens could not be regarded as a uniform anatomical entity but, on the contrary, should be subdivided anatomically and functionally into a shell part around the core part surrounding the anterior commissure (Brog, Salyapongse, Deutch & Zahm, 1993; Heimer, Zahm, Churchill et al., 1991; Zahm & Brog, 1992; Zahm & Heimer, 1993). This anatomo-functional subdivision of the nucleus accumbens was taken up by Murphy, Pezze, Feldon and Heibreder (2000) who, also using in vivo microdialysis, investigated the DAergic variations in the shell and core subregions of the nucleus accumbens during the expression of LI in a conditioned freezing paradigm. Murphy et al. (2000) reported that the conditioned response of DA to an aversively conditioned tone was eliminated in pre-exposed animals in the shell subregion, parallel to the attenuated conditioned freezing response, whereas in the core subregion of the nucleus accumbens no differences were observed in DA changes in pre-exposed and non-pre-exposed animals. However, these results were difficult to reconcile with the aforementioned reports by Killcross and Robbins (1993) and Ellenbroek et al. (1997), which argued in favour of specific involvement by the DAergic dorso-striatal innervation in LI. Using in vivo voltammetry, which allows for better anatomical resolution than in vivo microdialysis, and a specially designed conditioned olfactory aversion paradigm, we had previously observed specific conditioned DAergic responses in the dorsomedial shell and core parts of the nucleus accumbens (Besson & Louilot, 1995), as well as in the dorsal striatum (Besson & Louilot, 1997). We had also reported data suggesting that DA variations in the core part of the nucleus accumbens and behavioural outputs are related (Louilot & Besson, 2000), whereas the results obtained by Murphy et al. (2000) suggested the opposite. We therefore decided to carry out a systematic study of the involvement of DAergic neurons in LI at the level of several subregions of the nucleus accumbens and dorsal striatum. To that end, we first designed an LI paradigm from a conditioned olfactory paradigm used previously (Besson & Louilot, 1995, 1997; Louilot & Besson, 2000). This LI protocol is a three-stage paradigm (Figure 14.1). The conditional stimulus (CS) is banana odour, the unconditioned stimulus (US) lithium chloride (LiCl), a nausea-inducing toxic substance (Jeanblanc, Hoeltzel & Louilot, 2002). The following protocol was used to obtain LI. During the pre-exposure (first) session, the pre-exposed animals were placed for 1 h in the experimental cage without any olfactory stimulus, before being exposed for 2 h to the to-be-conditioned olfactory stimulus (banana odour). Three days after this first session, the second session consisted in aversively conditioning the animals to banana odour. Accordingly, 72 h after the pre-exposure session, the animals were put back in the experimental cage
322
A. Louilot, J. Jeanblanc, Y. Peterschmitt and F. Meyer
Figure 14.1. Schematic representation of the behavioural procedure used to obtain latent inhibition. (a) Three-stage latent inhibition paradigm. In the pre-exposure session pre-exposed animals (male rats) were placed in the experimental cage (Exp. Cage) for 1 h. They then were exposed for 2 h to the olfactory stimulus (Olf. St.) alone. The aversive conditioning session took place 3 days after the pre-exposure session. The retention session, 3 days later, was the test session. (b) Aversive conditioning procedure. Non-pre-exposed animals were subjected only to the aversive conditioning and retention sessions. (Adapted from Jeanblanc, Hoeltzel & Louilot, 2002.)
for 1 h before being exposed for 1 h to banana odour (the CS). At the end of the second hour, they were given an intraperitoneal (i.p.) injection of either saline (NaCl 0.9%) or an isotonic solution of lithium chloride (LiCl 0.15 M). They then stayed in the cage for a further hour with the olfactory stimulus still present. Seventytwo hours later, during the retention session (test stage), they were put back in the experimental cage for 1 h before being exposed for a further hour to the CS. The protocol used to obtain the olfactory aversive conditioning is exactly the same except that non-pre-exposed animals were subjected only to the conditioning session and the retention session (Figure 14.1). To sum up, during the conditioning session pre-exposed and non-pre-exposed control animals were administered NaCl i.p. whereas pre-exposed and non-pre-exposed conditioned animals were administered LiCl i.p. Attraction or aversion towards the olfactory stimulus (banana odour) was evaluated during the different sessions by measuring the time spent near the olfactive source. More specifically, the olfactory stimulus was fed into the cage through a hole in the wall next to the door. From preliminary experiments the floor of the cage was
Parahippocampal region–dopaminergic neuron relationships in latent inhibition
323
Figure 14.2. Time spent near the hole by animals exposed to the conditional olfactory stimulus in the retention session 3 days after the aversive conditioning session. Pre-exposed (PE) animals were exposed for the third time to the banana odour whereas non-pre-exposed (NPE) animals (shaded columns) were exposed to the olfactory stimulus (arrow) for the second time. The insert shows the percentage of cumulated time spent near the hole in the first 30 min after presentation of the olfactory stimulus. The dotted line corresponds to a neutral, empirically determined, distribution of the animals in the cage. n is the number of rats per group (PE and NPE): PE-NaCl ¼ 37, PE-LiCl ¼ 35, NPE-NaCl ¼ 31, NPE-LiCl ¼ 37. Statistical significance: ****P < 0.0001; *****P < 0.00001. (Bonferroni’s t test.)
divided into two virtual parts. One part, containing the hole, was a semi-circle covering a surface area equal to 35% of the entire floor. The remaining area made up the second part. The position of the animals in the experimental cage was monitored using a small infrared camera placed in the roof of the cage. Behavioural analysis was carried out over 10-min periods. We assumed that if animals moved randomly around in the cage they should spend 35% (210 s) of the 10-min period in the part containing the hole. This paradigm clearly allows the observation of LI, as evidenced by the disappearance of the aversively conditioned behavioural response in the pre-exposed conditioned animals (Figure 14.2). For the voltammetric investigations, and with respect to the nucleus accumbens, we took into account the anatomical data supporting the view that the nucleus accumbens could be subdivided not only into core and shell parts, but also subsequent studies suggesting that the shell could be subdivided into different subareas and, in particular, that at least in the dorsoventral axis the shell should be divided
324
A. Louilot, J. Jeanblanc, Y. Peterschmitt and F. Meyer
into a dorsomedial part near the ventricle, generally called the cone, and a more ventral part, known as the ventromedial shell (Diaz, Levesque, Griffon et al. 1994; Diaz, Levesque, Lammers et al., 1995; Jongen-Reˆlo, Voorn & Groenewegen, 1994; Wright & Groenewegen, 1995). Concerning the dorsal striatum, we took into account the possibility that the difference between the results obtained after similar local microinjection of d-amphetamine in the striatal structure by Solomon and Staton (1982), on the one hand, and Ellenbroek et al. (1997), on the other hand, might be connected with a localization of the involvement of dorsal striatal DA terminals in LI rather than with a difference due to the LI paradigms. Accordingly, we divided the dorsal striatum into two parts, one part anterior to the genu of the corpus callosum (the “anterior dorsal striatum”) and a more posterior part (the “posterior dorsal striatum”). In choosing this striatal subdivision we were initially guided by data which showed that, in rats, the anterior dorsal striatum was functionally linked to the anteromedian prefrontal cortex (Divac & Diemer, 1980), a cerebral region usually described as being involved in cognitive processes (see Fuster, 1989). First of all, it must be stressed that when we began our studies we had no a priori preference for one theoretical explanation of LI rather than another. As pointed out by Lubow in his introductory historical chapter, over the years the LI phenomenon has been the subject of several different theoretical proposals. Thus, it was first proposed that LI corresponds to a reduction in conditional stimulus (CS) associability as a consequence of either conditioned inattention (Lubow, 1989) or, relatedly, reduced CS salience (e.g., Mackintosh, 1975, 1983; Wagner, 1976) to the stimulus acquired during pre-exposure. Alternatively, it has been proposed that LI corresponds not to a delay in acquisition of the conditioned response, but to a failure of expression in the test stage (Bouton, 1993; Kraemer, Randall & Carbary, 1991; Miller & Matzel, 1988; Weiner, 1990, 2003). However, as we shall see below, our results are more easily explained by a defect in the expression of conditioning rather than by a defect in the acquisition of conditioning. Data collected in our three-stage LI paradigm (Figure 14.1) showed that the behavioural responses during the test session in the non-pre-exposed control group, the pre-exposed control, and the pre-exposed conditioned group were not statistically different. These animals approached the conditional olfactory stimulus after its application (Figure 14.2), suggesting that pre-exposed conditioned animals have an interest in the CS similar to the pre-exposed and non-pre-exposed groups of animals. These results are difficult to reconcile with the view that the LI phenomenon observed in our paradigm is connected with a learned inattentional response to the stimulus that is acquired during pre-exposure. Concerning DAergic neurons, our data argue in favour of localization of their involvement in the nucleus accumbens, as in the dorsal striatum, with a specificity of this regionalization for the two telencephalic structures (Jeanblanc, Hoeltzel & Louilot, 2002, 2003).
Parahippocampal region–dopaminergic neuron relationships in latent inhibition
325
Figure 14.3. Changes in dopamine (DA) release in the left dorsal shell (a), left ventral shell (b), and left core (c) subregions of the nucleus accumbens in animals exposed to the conditional olfactory stimulus (banana odour) during the retention session. Extracellular levels of DA were assessed, in parallel to behavioural analysis, using differential normal pulse voltammetry and computer-assisted numerical analysis in freely moving rats (Louilot, Serrano & D’Angio, 1987; Gonzalez-Mora, Guadalupe, Fumero & Mas, 1991). Voltammograms were recorded every minute. For each experiment the mean value of the control period was taken as the 100% value. The arrow corresponds to the presentation of the olfactory stimulus. n is the number of rats per group. Results were analysed using ANOVA. The insert shows typical recording sites (arrows) in the left dorsal shell (a), left ventral shell (b), and left core (c) parts of the nucleus accumbens. (Adapted from Jeanblanc, Hoeltzel & Louilot, 2002.)
Concerning the nucleus accumbens (Figure 14.3), our results showed, first, that the lack of DA increase observed in the core part of the nucleus accumbens in the nonpre-exposed conditioned animals was no longer observed in the pre-exposed conditioned animals, which displayed variations in DA signal similar to those obtained in the pre-exposed control animals; second, they showed that the rapid increase in DA levels observed in the dorsomedial shell in the non-pre-exposed conditioned animals
326
A. Louilot, J. Jeanblanc, Y. Peterschmitt and F. Meyer
is not obtained in the pre-exposed conditioned animals, which, to be more precise, showed a transient decrease in DA signal; and, third, the results showed that at the level of the ventromedial shell a similar and small increase in DA levels was obtained in both the non-pre-exposed and the pre-exposed conditioned animals. To summarize, the results obtained at the level of the nucleus accumbens argue in favour of an involvement of DAergic neurons reaching the core and dorsomedial shell subregions of the nucleus in LI but not of those reaching the ventromedial shell part of the nucleus accumbens, whose response appeared, however, to be dependent on conditioning in pre-exposed conditioned and non-pre-exposed conditioned animals. Concerning the dorsal striatum, results obtained in the anterior part are very similar, albeit not identical, to those obtained at the level of the core part of the nucleus accumbens (Figure 14.4). Thus, the weak increase in DA signal observed in the anterior part of the striatum in the non-pre-exposed conditioned animals was not obtained in the pre-exposed conditioned animals. In these animals, DA signal increased after presentation of the olfactory stimulus to an extent similar to the increase observed with the pre-exposed control animals. By contrast, in the posterior part of the dorsal striatum DA variations characterized by a moderate enhancement of DAergic levels were found, irrespective of the conditioning or pre-exposure status and only connected with the presentation of the olfactory stimulus. This is consistent with the involvement of DAergic neurons reaching this part of the dorsal striatum in sensorimotor processes (Ljungberg & Ungerstedt, 1976). Thus, the clear regionalization of the involvement of DAergic neurons innervating the dorsal striatum in LI (anterior part/posterior part) may well explain the contradictory results obtained after local d-amphetamine microinjections in the dorsal striatum (Solomon & Staton, 1982, vs. Ellenbroek et al., 1997). In other respects, the differential involvement of DAergic neurons reaching the different subregions of the nucleus accumbens in the LI phenomenon may also provide an explanation for the contradictory observations reported with local d-amphetamine microinjections in this nucleus (Ellenbroek et al., 1997; Killcross & Robbins; 1993; Solomon & Staton, 1982). The subregional specificity of the LI-related DA responses within the nucleus accumbens, as observed in our LI paradigm, may also explain why our results were at odds with those obtained by Murphy et al. (2000), insofar as the microdialysis approach these authors used does not allow subregional variations to be detected with any great degree of precision. As indicated above, DA variations obtained at test in pre-exposed conditioned animals in the ventromedial shell part of the nucleus accumbens are no different from those obtained in non-pre-exposed conditioned animals, which strongly suggests that aversive conditioning is acquired normally in pre-exposed animals in our experimental conditions (Jeanblanc et al., 2002). Accordingly, results obtained with our three-stage LI paradigm appeared difficult to reconcile with theories suggesting LI corresponds to a diminution of the CS associability during conditioning (Mackintosh, 1975, 1983; Wagner, 1976) and more in line with explanations of LI as a defect in the behavioural expression of
Parahippocampal region–dopaminergic neuron relationships in latent inhibition
327
Figure 14.4. Changes in DA release in the left anterior part (a) and left posterior part (b) of the dorsal striatum in animals exposed to the conditional olfactory stimulus during the retention session. For other comments see Figure 14.3 legend. The insert shows typical recording sites (arrows) in the left anterior part (a) and left posterior part (b) of the dorsal striatum. (Adapted from Jeanblanc, Hoeltzel & Louilot, 2003.)
conditioning (Bouton, 1993; Miller & Matzel, 1988; Weiner, 1990). More specifically, the nub of these theories, each with its own specificity, is that LI results from competition at the test stage between retrieval of the memories of the CS–no-event acquired in the preexposure context and retrieval of the memories of the CS acquired during conditioning. Above all in this context it is important to remember that we previously reported that
328
A. Louilot, J. Jeanblanc, Y. Peterschmitt and F. Meyer
olfactory aversively conditioned behavioural responses and concomitant DA responses recorded in the core part of the nucleus accumbens were found to be dependent on the basolateral nucleus of amygdala (Louilot & Besson, 2000). Accordingly, one interpretation of our data not different from the aforementioned latter theoretical explanations of LI would be that during the test stage the influence of cerebral structures involved in the expression of aversive conditioning is inhibited in some way by structures involved in some kind of recognition memory of the information related to the CS alone or the CS–no-event acquired during pre-exposure (Jeanblanc et al., 2002). In a second part of our investigations, we therefore decided to look at the consequences of the transitory functional inactivation of structures reported as being involved in LI and recognition memory processes on LI-related behavioural and DA responses. The entorhinal cortex (ENT) and the ventral subiculum of the hippocampus (SUB), whose involvement in both LI phenomenon and recognition memory is described in the literature (Gray et al., 1995; Petrulis, Alvarez & Eichenbaum, 2005; Suzuki & Eichenbaum, 2000; Weiner, 2003; Weiner & Feldon, 1997), appeared to be two suitable structures for our studies. We chose to investigate the consequences of a transitory blockade, rather than lesion, of these two structures so as to avoid the compensatory mechanisms that follow the lesioning of most cerebral regions. In a first step, and from a heuristic perspective, we assumed that if a structure is involved in encoding information (at preexposure) it may also be involved in the retrieval of this information (at test stage). Accordingly, we studied the impact of transitory inactivation by tetrodotoxin (TTX) of either the ENT or SUB at pre-exposure on behavioural and DAergic responses in preexposed animals during the test session (Jeanblanc, Peterschmitt, Hoeltzel & Louilot, 2004; Peterschmitt, Hoeltzel & Louilot, 2005; Peterschmitt, Meyer & Louilot, 2008). DAergic variations were monitored in the core and shell parts of the nucleus accumbens and the anterior part of the dorsal striatum using, as in previous studies, in vivo voltammetry in freely moving animals. As far as functional inactivation of the ENT at pre-exposure is concerned behavioural results obtained during the retention session showed that the responses in preexposed conditioned animals microinjected with TTX are significantly different from responses obtained in pre-exposed conditioned animals microinjected with the solvent (phosphate-buffered saline, PBS), corresponding to LI expression, and similar to the behavioural responses reflecting aversion obtained in non-pre-exposed conditioned animals (Jeanblanc et al., 2004). In other words, functional inactivation of the ENT during the pre-exposure session causes a complete reversal and disappearance of the behavioural expression of LI in the test stage. Regarding the DAergic variations, first concerning the core part of the nucleus accumbens (Figure 14.5a), in the pre-exposed control animals and pre-exposed conditioned animals microinjected with the solvent (PBS) in the ENT, marked enhancements were observed in the DA signal throughout the hour following presentation of the olfactory stimulus. In the pre-exposed control animals microinjected with TTX in the ENT a gradual increase in the DA signal was also observed.
Parahippocampal region–dopaminergic neuron relationships in latent inhibition
329
Figure 14.5. Changes in DA release in the left core (a) and left dorsal shell (b) parts of the nucleus accumbens, and in the left anterior part of the dorsal striatum (c) during the test session in PE male rats after functional blockade of the left ENT throughout the pre-exposure session (left columns), and in NPE rats (right column). PE animals were microinjected with PBS or tetrodotoxin (TTX) in the left ENT 3 h before pre-exposure to banana odour. For other comments see Figure 14.3 legend. (Adapted from Jeanblanc, Peterschmitt, Hoeltzel & Louilot, 2004.)
330
A. Louilot, J. Jeanblanc, Y. Peterschmitt and F. Meyer
In contrast, in the pre-exposed conditioned TTX animals a rapid albeit slight decrease was observed in the DA signal after presentation of the olfactory stimulus until the 15th minute, after which the signal increased markedly until the end of the session. In the non-pre-exposed control animals the DA signal increased gradually during the post-stimulus hour, whereas in the non-pre-exposed conditioned animals a rapid decrease was obtained upon presentation of the olfactory stimulus and until the 20th minute, after which the DA signal returned to the basal levels and thereafter continued to increase during the second half-hour post-stimulus. To sum up, DAergic variations in the core subregion in pre-exposed conditioned animals subjected to the TTX functional blockade of the ENT were mixed DA responses, with an initial reduction in DA levels similar to the DA response obtained in non-pre-exposed conditioned animals, followed by an increase in DA levels similar to, or even higher than, the DA increases observed in pre-exposed conditioned animals microinjected with PBS. Concerning the dorsomedial shell part of the nucleus accumbens (Figure 14.5b) and the anterior part of the dorsal striatum (Figure 14.5c), changes in DA levels in pre-exposed conditioned animals microinjected with TTX in the ENT were no different to those found in non-pre-exposed conditioned animals. More precisely, as regards the dorsomedial shell (Figure 14.5b), in the pre-exposed control animals microinjected with PBS in the ENT, the DA signal gradually increased during the first 50 min after presentation of the stimulus to peak in the last 10 min. In the preexposed conditioned animals microinjected with PBS, the DA signal increased moderately up to the end of the hour. Changes in the DA signal in the pre-exposed groups microinjected with TTX in the ENT were close to those obtained in the non-preexposed groups. In the pre-exposed TTX control animals, a slight increase in the DA signal was observed during the first 50 min, followed by a sharp increase during the last 10 min. In contrast, for the pre-exposed TTX conditioned animals a rapid and marked enhancement in the DA signal was obtained from the moment of exposure to the olfactory stimulus up to the end of the hour. In the non-pre-exposed control animals after exposure to the olfactory stimulus a small increase in the DA signal was observed for the whole hour whereas in the non-pre-exposed conditioned animals the DA signal increased markedly above the pre-stimulus baseline. Regarding the anterior dorsal striatum (Figure 14.5c), in the pre-exposed control animals microinjected with PBS, the DA signal showed a rapid increase for the first 15 min after presentation of the stimulus and then increased further up to the end of the hour. In the pre-exposed conditioned PBS animals the signal increased continuously for the hour following presentation of the olfactory stimulus. Variations in the signal for the control and conditioned pre-exposed groups microinjected with TTX in the ENT were different, but the directions of the changes were similar to those observed in the respective non-pre-exposed groups. In the pre-exposed control TTX animals a marked enhancement of the DA signal was observed throughout the post-stimulus hour. By contrast, in the pre-exposed conditioned TTX animals, the DA signal
Parahippocampal region–dopaminergic neuron relationships in latent inhibition
331
remained close to the baseline and even fell slightly below the basal levels at the end of the hour. In the non-pre-exposed control animals, after presentation of the olfactory stimulus there was a gradual increase in the DA signal which peaked by the end of the hour. By contrast, in the non-pre-exposed conditioned animals the DA signal fluctuated close to the baseline for the whole of the post-stimulus hour. To sum up, in the pre-exposed conditioned animals microinjected with TTX in the ENT and the non-pre-exposed conditioned animals, DA levels displayed marked increases in the dorsomedial shell subregion and remained close to baseline in the anterior part of the dorsal striatum. As far as functional inactivation of the SUB by TTX at pre-exposure is concerned, behavioural data obtained during the test session showed that the responses in preexposed conditioned TTX animals are statistically different from LI responses found in pre-exposed conditioned PBS animals and no different from the aversive responses shown by the non-pre-exposed conditioned animals towards the olfactory stimulus (Peterschmitt et al., 2005, 2008). Thus, transitory blockade of the SUB at preexposure led to a behavioural reversal of LI expression similar to, or even more pronounced than, those observed after temporary blockade of the ENT. As regards the DAergic variations in the core part of the nucleus accumbens in the pre-exposed conditioned animals microinjected with TTX in the SUB, they are distinct from the DA changes obtained in the pre-exposed conditioned animals microinjected with PBS but not different from those observed in the non-pre-exposed conditioned animals (Figure 14.6a). More specifically, in the pre-exposed control animals and pre-exposed conditioned animals microinjected with PBS in the SUB, the DA signal rose steadily after presentation of the olfactory stimulus before levelling off for the last 20 min of the post-stimulus period. By contrast, there were differences in the directions of DA changes between the pre-exposed control animals and conditioned animals microinjected with TTX in the SUB, and these differences were similar to those observed in the respective non-pre-exposed control and conditioned groups. A rapid and sustained increase in the DA signal was observed in the pre-exposed control TTX animals following presentation of the olfactory stimulus, whereas in the pre-exposed conditioned TTX animals the signal fluctuated around the baseline. In the non-pre-exposed control animals the DA signal showed a rapid and marked increase above the pre-stimulus basal levels. Conversely, in the non-pre-exposed conditioned animals the DA signal decreased rapidly and transiently after presentation of the olfactory stimulus and then stayed close to the basal values from the 10th minute to the end of the session. As regards the dorsomedial shell subregion (Figure 14.6b) in the pre-exposed control animals microinjected with the PBS in the SUB, the DA signal rose rapidly above the basal level in the first 10 min, remained close to this high level in the first half-hour and then increased further during the second half-hour. In the pre-exposed conditioned PBS animals, the DA signal rapidly decreased after presentation of the olfactory stimulus and then from the 15th minute onwards continued to fluctuate
332
A. Louilot, J. Jeanblanc, Y. Peterschmitt and F. Meyer
Figure 14.6. Changes in DA release in the left core (a) and left dorsal shell (b) parts of the nucleus accumbens and in the left anterior part of the dorsal striatum (c) during the test session in PE male rats after functional blockade of the left SUB throughout the pre-exposure session (left columns), and in NPE rats (right column). PE animals were microinjected with PBS or TTX in the left SUB 3 h before pre-exposure to banana odour. For other comments see Figure 14.3 legend. (Adapted from Peterschmitt, Hoeltzel & Louilot, 2005; Peterschmitt, Meyer & Louilot, 2008.)
Parahippocampal region–dopaminergic neuron relationships in latent inhibition
333
near the baseline for the whole hour. By contrast, for the first 50 min following exposure to the olfactory stimulus, the DA signal remained relatively stable above or near the basal levels in the pre-exposed control and pre-exposed conditioned animals microinjected with TTX respectively. Changes in the signal differed in the two TTX groups only during the last 10 min, with a slight drop below the baseline in the pre-exposed control TTX animals and a sudden rise above basal values in the pre-exposed conditioned TTX animals. In the non-pre-exposed control animals the DA signal stayed close to the baseline for the first half-hour and fell below the basal levels thereafter, fluctuating between this lower level and the baseline for the second half-hour post-stimulus. To sum up, DAergic variations in the pre-exposed conditioned animals microinjected with TTX were found to be between those obtained in the pre-exposed conditioned PBS animals and the non-pre-exposed conditioned animals. As for the changes in DA levels in the anterior part of the dorsal striatum (Figure 14.6c), in the pre-exposed control and pre-exposed conditioned animals microinjected with PBS in the SUB, the DA signal increased rapidly after presentation of the olfactory stimulus and then gradually from the 10th minute until the end of the hour. By contrast, distinct changes were observed in the two pre-exposed groups microinjected with TTX. In the pre-exposed control animals microinjected with TTX, the DA signal increased gradually for the whole hour. In the pre-exposed conditioned TTX animals, the signal initially showed a rapid drop after presentation of the olfactory stimulus, before returning to basal values after 20 min, remaining at this level until the 40th minute and then rising sharply until the end of the session. In the non-pre-exposed control animals, a rapid increase was observed in the DA signal in the first 10 min. The signal then continued to rise slowly throughout the hour whereas in the non-pre-exposed conditioned animals it fluctuated near or slightly above the basal levels. To sum up DA variations in the anterior part of the dorsal striatum, they are significantly different in the pre-exposed conditioned TTX and PBS animals, but, although DAergic changes in the conditioned animals microinjected with TTX were globally similar to those observed in the non-pre-exposed conditioned animals, DAergic changes in the former group were statistically different for the first 20 min after presentation of the olfactory stimulus (more pronounced) than the changes observed in the latter group. It should be noted first of all that the differences in the time-course of the DAergic responses observed in pre-exposed conditioned TTX animals and non-pre-exposed conditioned animals in the core subregion after functional blockade of the ENT, on one hand, and in the dorsomedial shell and anterior dorsal striatum subregions after functional blockade of the SUB, on the other hand, argued against the reason for the results obtained during the test session simply being a primary olfactory perceptual deficit in the pre-exposure session in the TTX microinjected animals. Secondly, the specificity of the DAergic responses observed in the retention session further suggests that the two temporal regions are both involved in the disruption of LI. Our
334
A. Louilot, J. Jeanblanc, Y. Peterschmitt and F. Meyer
behavioural results are therefore in line with those obtained with lesion studies as far as the ENT is concerned but are at odds with them as regards the SUB, since other authors found no impairment of LI with excitotoxic lesions of the SUB (Coutureau, Galani, Gosselin et al., 1999; Oswald, Yee, Rawlins et al., 2002; Pouzet, Zhang, Weiner et al., 2004; Shohamy, Allen & Gluck, 2000). A likely explanation for the latter discrepancy may be the existence of more important compensatory mechanisms following the SUB than the ENT lesions. Functional data showing a higher c-fos labelling in the SUB of pre-exposed conditioned animals in the conditioned emotional response paradigm also argue in favour of involvement of the SUB in LI (Sotty, Sandner & Gosselin, 1996). Behavioural results obtained after functional inactivation of the ENT and SUB suggest that the two temporal structures are involved in encoding the information relating to the CS–no-event or CS alone at pre-exposure. Whereas our data did not allow us to assert that the two structures are also involved at test stage in the retrieval of the CS-related information, the fact that DA variations in the two accumbal subregions (core and dorsomedial shell) and anterior part of the dorsal striatum are differently affected depending on the structure targeted at pre-exposure (ENT or the SUB) lends support to this view. No structures were inactivated at the test stage in our experiments, but the DAergic responses in the three striatal subregions appeared to be specifically dependent on the structure blocked during the first session, rather like the reported involvement of the two temporal regions in recognition memory processes (Petrulis et al., 2005; Suzuki & Eichenbaum, 2000). Another argument supporting our interpretation stems from the behavioural results obtained recently by Lewis and Gould (2007) in another LI paradigm, which suggest that the parahippocampal region comprising the ENT and SUB support both the encoding of the CS-related information during the pre-exposure session and retrieval of this information during the retention session. In other words, loss of encoding in respect of the information relating to the CS in the pre-exposed conditioned TTX animals in our LI paradigm may prevent its retrieval in the test stage by the ENT or SUB, depending on the structure first inactivated in the pre-exposure. Accordingly, the specific DAergic variations observed in the three subregions in the pre-exposed conditioned TTX animals may reflect the lack of retrieval of the CS-related information by the ENT or SUB. Conversely, it is tempting to suggest that in normal conditions, as is the case with animals microinjected with the solvent (PBS), the two temporal regions participate cooperatively with other structures such as the basolateral nucleus of amygdala or the prefrontal cortex in the accumbal and striatal DA LI-related changes (see Jeanblanc et al., 2003; Peterschmitt et al., 2005, 2008, for detailed discussions). These dynamic regulatory influences exerted on DA levels by the ENT and the SUB are consistent with the anatomo-functional relationships repeatedly described between the ENT and the SUB, on the one hand, and DAergic transmission, on the other hand, as regards the nucleus accumbens (Blaha, Yang, Floresco et al., 1997; Brudzynski & Gibson, 1997; Cano-Cebrian, Zornoza-Sabina,
Parahippocampal region–dopaminergic neuron relationships in latent inhibition
335
Guerri et al., 2003; Howland, MacKenzie, Yim et al., 2004; Legault, Rompre & Wise, 2000; Louilot & Choulli, 1997; Louilot & Le Moal, 1994; Mitchell, Yee, Feldon et al., 2000; Peleg-Raibstein & Feldon, 2006; Taepavarapruk, Floresco & Phillips, 2000). As regards the anterior part of the dorsal striatum, similar anatomo-functional relationships have yet to be investigated in depth, but it is interesting to note that electrophysiological data have shown convergent influences of ENT and SUB inputs on dorso-striatal neurons (Finch, 1996). To summarize, the data we obtained in our LI paradigm showed that DAergic neurons displayed LI-related responses at the level of both the dorsal striatum and the nucleus accumbens. They argued, however, in favour of a limitation of this involvement to the anterior part of the dorsal striatum, on the one hand, and to the core and dorsomedial shell parts of the nucleus accumbens, on the other hand, with specific responses for each of the three subregions. This narrow regionalization may account for the discrepancies initially reported with the neuropsychopharmacological studies. Our data also strongly suggest that LI-related DA responses are dependent on the functional integrity of a parahippocampal region, including both the ENT and the SUB. It is therefore important to remember that different kinds of cytoarchitectural, anatomical and functional abnormalities have been described in patients with schizophrenia at the level of the ENT and the SUB (Arnold, 2000; Arnold, Ruscheinsky & Han, 1997; Falkai, Schneider-Axmann & Honer, 2000; Jakob & Beckmann, 1994; Joyal, Laakso, Tiihonen et al., 2002; Law, Weickert, Hyde et al., 2004; Silbersweig, Stern, Frith et al., 1995). This proves the importance in LI of deciphering the relationships between DA neurons and the two temporal regions, particularly from a neurodevelopmental perspective. Our more recent data showing that neonatal inactivation of the ENT produces a disruption of LI-related DA responses in the core and dorsomedial shell subregions in grown-up animals (Peterschmitt, Meyer & Louilot, 2007), distinct from that observed after functional inactivation in adult animals (presented above), therefore constitute a further step towards the modelling of the pathophysiology of schizophrenia in animals. Acknowledgements Support provided by EDF (A.L.), FRM (J.J.), MESR þ National Academy of Medicine (Y.P.), and INSERM & Region Alsace þ Fondation de France (F.M.). References Arnold, S. E. (2000). Cellular and molecular neuropathology of the parahippocampal region in schizophrenia. Annals of the New York Academy of Sciences, 911, 275–292. Arnold, S. E., Ruscheinsky, D. D., & Han, L. Y. (1997). Further evidence of abnormal cytoarchitecture of the entorhinal cortex in schizophrenia using spatial point pattern analyses. Biological Psychiatry, 42, 639–647.
336
A. Louilot, J. Jeanblanc, Y. Peterschmitt and F. Meyer
Baruch, I., Hemsley, D. R., & Gray, J. A. (1988). Differential performance of acute and chronic schizophrenics in a latent inhibition task. Journal of Nervous and Mental Disease, 176, 698–600. Besson, C., & Louilot, A. (1995). Asymmetrical involvement of mesolimbic dopaminergic neurons in affective perception. Neuroscience, 68, 963–968. Besson, C., & Louilot, A. (1997). Striatal dopaminergic changes depend on the attractive or aversive value of stimulus. Neuroreport, 8, 3523–3526. Blaha, C. D., Yang, C. R., Floresco, S. B., Barr, A. M., & Phillips, A. G. (1997). Stimulation of the ventral subiculum of the hippocampus evokes glutamate receptor-mediated changes in dopamine efflux in the rat nucleus accumbens. European Journal of Neuroscience, 9, 902–911. Bouton, M. E. (1993). Context, time, and memory retrieval in the interference paradigms of Pavlovian conditioning. Psychological Bulletin, 114, 80–99. Brog, J. S., Salyapongse, A., Deutch, A. Y., & Zahm, D. S. (1993). The patterns of afferent innervation of the core and shell in the “accumbens” part of the rat ventral striatum: immunohistochemical detection of retrogradely transported fluoro-gold. Journal of Comparative Neurology, 338, 255–278. Brudzynski, S. M., & Gibson, C. J. (1997). Release of dopamine in the nucleus accumbens caused by stimulation of the subiculum in freely moving rats. Brain Research Bulletin, 42, 303–308. Cano-Cebrian, M. J., Zornoza-Sabina, T., Guerri, C., Polache, A., & Granero, L. (2003). Acamprosate blocks the increase in dopamine extracellular levels in nucleus accumbens evoked by chemical stimulation of the ventral hippocampus. Naunyn Schmiedebergs Archives of Pharmacology, 368, 324–327. Carlsson, A., Waters, N., Holm-Waters, S., et al. (2001). Interactions between monoamines, glutamate, and GABA in schizophrenia: new evidence. Annual Review of Pharmacology and Toxicology, 41, 237–260. Coutureau, E., Galani, R., Gosselin, O., Majchrzak, M., & Di Scala, G. (1999). Entorhinal but not hippocampal or subicular lesions disrupt latent inhibition in rats. Neurobiology Learning and Memory, 72, 143–157. Diaz, J., Levesque, D., Griffon, N., et al. (1994). Opposing roles for dopamine D2 and D3 receptors on neurotensin mRNA expression in nucleus accumbens. European Journal of Neuroscience, 6, 1384–1387. Diaz, J., Levesque, D., Lammers, C. H., et al. (1995). Phenotypical characterization of neurons expressing the dopamine D3 receptor in the rat brain. Neuroscience, 65, 731–745. Divac, I., & Diemer, N. H. (1980). Prefrontal system in the rat visualized by means of labeled deoxyglucose – further evidence for functional heterogeneity of the neostriatum. Journal of Comparative Neurology, 190, 1–13. Ellenbroek, B. A., Knobbout, D. A., & Cools, A. R. (1997). The role of mesolimbic and nigrostriatal dopamine in latent inhibition as measured with the conditioned taste aversion paradigm. Psychopharmacology, 129, 112–120. Falkai, P., Schneider-Axmann, T., & Honer, W. G. (2000). Entorhinal cortex prealpha cell clusters in schizophrenia: quantitative evidence of a developmental abnormality. Biological Psychiatry, 47, 937–943. Feldon, J., Shofel, A., & Weiner, I. (1991). Latent inhibition is unaffected by direct dopamine agonists. Pharmacology, Biochemistry and Behavior, 38, 309–314.
Parahippocampal region–dopaminergic neuron relationships in latent inhibition
337
Feldon, J., & Weiner, I. (1991). The latent inhibition model of schizophrenic attention disorder. Haloperidol and sulpiride enhance rats’ ability to ignore irrelevant stimuli. Biological Psychiatry, 29, 635–646. Finch, D. M. (1996). Neurophysiology of converging synaptic inputs from the rat prefrontal cortex, amygdala, midline thalamus, and hippocampal formation onto single neurons of the caudate/putamen and nucleus accumbens. Hippocampus, 6, 495–512. Fuster, J. M. (1989). The Prefrontal Cortex, 2nd ed. New York: Raven Press. Gonzalez-Mora, J. L., Guadalupe, T., Fumero, B., & Mas, M. (1991). Mathematical resolution of mixed in vivo voltammetry signals. Models, equipment, assessment by simultaneous microdialysis sampling. Journal of Neuroscience Methods, 39, 231–244. Gosselin, G., Oberling, P., & Di Scala, G. (1996). Antagonism of amphetamineinduced disruption of latent inhibition by the atypical antipsychotic olanzapine in rats. Behavioural Pharmacology, 7, 820–826. Gray, N. S., Hemsley, D. R., & Gray, J. A. (1992). Abolition of latent inhibition in acute, but not chronic, schizophrenics. Neurology, Psychiatry and Brain Research, 1, 83–89. Gray, N. S., Pickering, A. D., Hemsley, D. R., Dawling, S., & Gray, J. A. (1992). Abolition of latent inhibition by a single 5 mg dose of d-amphetamine in man. Psychopharmacology, 107, 425–430. Gray, N. S., Pilowsky, L. S., Gray, J. A., & Kerwin, R. W. (1995). Latent inhibition in drug naive schizophrenics: relationship to duration of illness and dopamine D2 binding using SPECT. Schizophrenia Research, 17, 95–107. Harrison, P. J. (1999). The neuropathology of schizophrenia. A critical review of the data and their interpretation. Brain, 122, 593–624. Howland, J. G., MacKenzie, E. M., Yim, T. T., Taepavarapruk, P., & Phillips, A. G. (2004). Electrical stimulation of the hippocampus disrupts prepulse inhibition in rats: frequency- and site-dependent effects. Behavioral Brain Research, 152, 187–197. Heimer, L., Zahm, D. S., Churchill, L., Kalivas, P. W., & Wohltmann, C. (1991). Specificity in the projection patterns of accumbal core and shell in the rat. Neuroscience, 41, 89–125. Jakob, H., & Beckmann, H. (1994). Circumscribed malformation and nerve cell alterations in the entorhinal cortex of schizophrenics. Pathogenetic and clinical aspects. Journal of Neural Transmission, 98, 83–106. Jeanblanc, J., Hoeltzel, A., & Louilot, A. (2002). Dissociation in the involvement of dopaminergic neurons innervating the core and shell subregions of the nucleus accumbens in latent inhibition and affective perception. Neuroscience, 111, 315–323. Jeanblanc, J., Hoeltzel, A., & Louilot, A. (2003). Differential involvement of dopamine in the anterior and posterior parts of the dorsal striatum in latent inhibition. Neuroscience, 118, 233–241. Jeanblanc, J., Peterschmitt, Y., Hoeltzel, A., & Louilot, A. (2004). Influence of the entorhinal cortex on accumbal and striatal dopaminergic responses in a latent inhibition paradigm. Neuroscience, 128, 187–200. Jongen-Reˆlo, A. L., Voorn, P., & Groenewegen, H. J. (1994). Immunohistochemical characterization of the shell and core territories of the nucleus accumbens in the rat. European Journal of Neuroscience, 6, 1255–1264.
338
A. Louilot, J. Jeanblanc, Y. Peterschmitt and F. Meyer
Joyal, C. C., Laakso, M. P., Tiihonen, J., et al. (2002). A volumetric MRI study of the entorhinal cortex in first episode neuroleptic-naive schizophrenia. Biological Psychiatry, 51, 1005–1007. Killcross, A. S., & Robbins, T. W. (1993). Differential effects of intra-accumbens and systemic amphetamine on latent inhibition using an on-baseline, within-subject conditioned suppression paradigm. Psychopharmacology, 110, 479–489. Kraemer, P. J., Randall, C. K., & Carbary, T. J. (1991). Release from latent inhibition with delayed testing. Animal Learning & Behavior, 19, 139–145. Law, A. J., Weickert, C. S., Hyde, T. M., Kleinman, J. E., & Harrison, P. J. (2004). Reduced spinophilin but not microtubule-associated protein 2 expression in the hippocampal formation in schizophrenia and mood disorders: molecular evidence for a pathology of dendritic spines. American Journal of Psychiatry, 161, 1848–1855. Legault, M., Rompre, P. P., & Wise, R. A. (2000). Chemical stimulation of the ventral hippocampus elevates nucleus accumbens dopamine by activating dopaminergic neurons of the ventral tegmental area. Journal of Neuroscience, 20, 1635–1642. Lewis, M. C., & Gould, T. J. (2007). Reversible inactivation of the entorhinal cortex disrupts the establishment and expression of latent inhibition of cued fear conditioning in C57BL/6 mice. Hippocampus, 17, 462–470. Ljungberg, T., & Ungerstedt, U. (1976). Sensory inattention produced by 6-hydroxydopamine-induced degeneration of ascending dopamine neurons in the brain. Experimental Neurology, 53, 585–600. Louilot, A., & Besson, C. (2000). Specificity of amygdalostriatal interactions in the involvement of mesencephalic dopaminergic neurons in affective perception. Neuroscience, 96, 73–82. Louilot, A., & Choulli, M. K. (1997). Asymmetrical increases in dopamine turn-over in the nucleus accumbens and lack of changes in locomotor responses following unilateral dopaminergic depletions in the entorhinal cortex. Brain Research, 778, 150–157. Louilot, A., & Le Moal, M. (1994). Lateralized interdependence between limbicotemporal and ventrostriatal dopaminergic transmission. Neuroscience, 59, 495–500. Louilot, A., Serrano, A., & D’Angio, M. (1987). A novel carbon fiber implantation assembly for cerebral voltammetric measurements in freely moving rats. Physiology and Behavior, 41, 227–231. Lubow, R. E. (1989). Latent Inhibition and Conditioned Attention Theory. Cambridge: Cambridge University Press. Lubow, R. E., Kaplan, O., Abramovich, P., Rudnick, A., & Laor, N. (2000). Visual search in schizophrenia: latent inhibition and novel pop-out effects. Schizophrenia Research, 45, 145–156. Mackintosh, N. J. (1975). A theory of attention: Variations in the associability of stimuli with reinforcement. Psychological Review, 82, 276–298. Mackintosh, N. J. (1983) Conditioning and Associative Learning. Oxford: Oxford University Press. Miller, R. R., & Matzel, L. D. (1988). The comparator hypothesis: A response rule for the expression of associations. In G. H. Bower (Ed.), The Psychology of Learning and Motivation, vol. 22, San Diego, CA: Academic Press, pp. 51–92.
Parahippocampal region–dopaminergic neuron relationships in latent inhibition
339
Mitchell, S. N., Yee, B. K., Feldon, J., Gray, J. A., & Rawlins, J. N. (2000). Activation of the retrohippocampal region in the rat causes dopamine release in the nucleus accumbens: disruption by fornix section. European Journal of Pharmacology, 407, 131–138. Murphy, C. A., Pezze, M., Feldon, J., & Heidbreder, C. (2000). Differential involvement of dopamine in the shell and core of the nucleus accumbens in the expression of latent inhibition to an aversively conditioned stimulus. Neuroscience, 97, 469–477. Oswald, C. J., Yee, B. K., Rawlins, J. N., et al. (2002). The influence of selective lesions to components of the hippocampal system on the orienting [correction of orientating] response, habituation and latent inhibition. European Journal of Neuroscience, 15, 1983–1990. Peleg-Raibstein, D., & Feldon, J. (2006). Effects of dorsal and ventral hippocampal NMDA stimulation on nucleus accumbens core and shell dopamine release. Neuropharmacology, 51, 947–957. Peterschmitt, Y., Hoeltzel, A., & Louilot, A. (2005). Striatal dopaminergic responses observed in latent inhibition are dependent on the hippocampal ventral subicular region. European Journal of Neuroscience, 22, 2059–2068. Peterschmitt, Y., Meyer, F., & Louilot, A. (2007). Neonatal functional blockade of the entorhinal cortex results in disruption of accumbal dopaminergic responses observed in latent inhibition paradigm in adult rats. European Journal of Neuroscience, 25, 2504–2513. Peterschmitt, Y., Meyer, F., & Louilot, A. (2008). Differential influence of the ventral subiculum on dopaminergic responses observed in core and dorsomedial shell subregions of the nucleus accumbens in latent inhibition. Neuroscience, 154, 898–910. Petrulis, A., Alvarez, P., & Eichenbaum, H. (2005). Neural correlates of social odor recognition and the representation of individual distinctive social odors within entorhinal cortex and ventral subiculum. Neuroscience, 130, 259–274. Pouzet, B., Zhang, W. N., Weiner, I., Feldon, J., & Yee, B. K. (2004). Latent inhibition is spared by N-methyl-D-aspartate (NMDA)-induced ventral hippocampal lesions, but is attenuated following local activation of the ventral hippocampus by intracerebral NMDA infusion. Neuroscience, 124, 183–194. Russig, H., Murphy, C. A., & Feldon, J. (2002). Clozapine and haloperidol reinstate latent inhibition following its disruption during amphetamine withdrawal. Neuropsychopharmacology, 26, 765–777. Shohamy, D., Allen, M. T., & Gluck, M. A. (2000). Dissociating entorhinal and hippocampal involvement in latent inhibition. Behavioral Neuroscience, 114, 867–874. Silbersweig, D. A., Stern, E., Frith, C., et al. (1995). A functional neuroanatomy of hallucinations in schizophrenia. Nature, 378, 176–179. Solomon, P. R., Crider, A., Winkelman, J. W., et al. (1981). Disrupted latent inhibition in the rat with chronic amphetamine or haloperidol-induced supersensitivity: relationship to schizophrenic attention disorder. Biological Psychiatry, 16, 519–538. Solomon, P. R., & Staton, D. M. (1982). Differential effects of microinjections of d-amphetamine into the nucleus accumbens or the caudate putamen on the rat’s ability to ignore an irrelevant stimulus. Biological Psychiatry, 17, 743–756.
340
A. Louilot, J. Jeanblanc, Y. Peterschmitt and F. Meyer
Sotty, F., Sandner, G., & Gosselin, O. (1996). Latent inhibition in conditioned emotional response: c-fos immunolabelling evidence for brain areas involved in the rat. Brain Research, 737, 243–254. Suzuki, W. A., & Eichenbaum, H. (2000). The neurophysiology of memory. Annals of the New York Academy of Sciences, 911, 175–191. Swerdlow, N. R., & Koob, G. F. (1987). DA, schizophrenia, mania, and depression: toward a unified hypothesis of cortico-striato-pallido-thalamic function. Behavioral Brain Science, 10, 197–245. Swerdlow, N. R., Stephany, N., Wasserman, L. C., et al. (2003). Dopamine agonists disrupt visual latent inhibition in normal males using a within-subject paradigm. Psychopharmacology, 169, 314–320. Taepavarapruk, P., Floresco, S. B., & Phillips, A. G. (2000). Hyperlocomotion and increased dopamine efflux in the rat nucleus accumbens evoked by electrical stimulation of the ventral subiculum: role of ionotropic glutamate and dopamine D1 receptors. Psychopharmacology, 151, 242–251. Trimble, K. M., Bell, R., & King, D. J. (1997). Enhancement of latent inhibition in the rat by the atypical antipsychotic agent remoxipride. Pharmacology Biochemistry and Behavior, 56, 809–816. Trimble, K. M., Bell, R., & King, D. J. (1998). Enhancement of latent inhibition in the rat at a high dose of clozapine. Journal of Psychopharmacology, 12, 215–219. Trimble, K. M., Bell, R., & King, D. J. (2002). Effects of the selective dopamine D(1) antagonists NNC 01–0112 and SCH 39166 on latent inhibition in the rat. Physiology and Behavior, 77, 115–123. Troncoso, A. C., Osaki, M. Y., Mason, S., Borelli, K. G., & Branda˜o, M. L. (2003). Apomorphine enhances conditioned responses induced by aversive stimulation of the inferior colliculus. Neuropsychopharmacology, 28, 284–291. Wagner, A. R. (1976). Priming in STM: An information processing mechanism for self-generated or retrieval-generated depression in performance. In T. Tighe & R. N. Leaton (Eds.), Habituation: Perspectives from Child Development, Animal Behavior, and Neurophysiology. Hillsdale, NJ: Lawrence Erlbaum, pp. 95–128. Weiner, I. (1990). The neural substrates of latent inhibition. Psychological Bulletin, 108, 442–461. Weiner, I. (2003). The “two-headed” latent inhibition model of schizophrenia: modeling positive and negative symptoms and their treatment. Psychopharmacology, 169, 257–297. Weiner, I., & Feldon, J. (1987). Facilitation of latent inhibition by haloperidol. Psychopharmacology, 91, 248–253. Weiner I., & Feldon, J. (1997). The switching model of latent inhibition: an update of neural substrates. Behavioral Brain Research, 88,11–25. Weiner, I., Feldon, J., & Katz, Y. (1987). Facilitation of the expression but not the acquisition of latent inhibition by haloperidol in rats. Pharmacology, Biochemistry and Behavior, 26, 241–246. Weiner, I., Lubow, R. E., & Feldon, J. (1981). Chronic amphetamine and latent inhibition. Behavioral Brain Research, 2, 285–286. Weiner, I., Lubow, R. E., & Feldon, J. (1984). Abolition of the expression but not the acquisition of latent inhibition by chronic amphetamine in rats. Psychopharmacology, 83, 194–199.
Parahippocampal region–dopaminergic neuron relationships in latent inhibition
341
Weiner, I., Lubow, R. E., & Feldon, J. (1988). Disruption of latent inhibition by acute administration of low doses of amphetamine. Pharmacology, Biochemistry and Behavior, 30, 871–878. Wright, C. I., & Groenewegen, H. J. (1995). Patterns of convergence and segregation in the medial nucleus accumbens of the rat: relationships of prefrontal cortical, midline thalamic, and basal amygdaloid afferents. Journal of Comparative Neurology, 361, 383–403. Young, A. M. J., Joseph, M. H., & Gray, J. A. (1993). Latent inhibition of conditioned dopamine release in rat nucleus accumbens. Neuroscience, 54, 5–9. Zahm, D. S., & Brog, J. S. (1992). On the significance of subterritories in the accumbens part of the ventral striatum. Neuroscience, 50, 751–767. Zahm, D. S., & Heimer, L. (1993). Specificity in the efferent projections of the nucleus accumbens in the rat: comparison of the rostral pole projection patterns with those of the core and shell. Journal of Comparative Neurology, 327, 220–232.
15 Latent inhibition and other salience modulation effects: same neural substrates? Helen J. Cassaday and Paula M. Moran
The complexity of the environment is such that even a snapshot in time from a single modality can exceed the processing capacity of the human brain (Tsotsos, 1990). This means that selection of some incoming stimuli for more detailed analysis (while ignoring other stimuli) is essential for efficient cognitive processing. Without such filtering we would be inundated with competing sensory impressions. Thus notions of stimulus salience and relevance have assumed prominence in a number of domains, from visual attention (Fecteau & Munoz, 2006; Li, 2002; Morris, Friston & Dolan, 1997) to animal cognition (Mackintosh, 1975). But what exactly is meant by the term stimulus salience? There is no straightforward answer to this question, as the apprehension of salience is likely to be determined by a variety of factors. First, the inherent features of the stimulus will contribute to its effective salience, for example its physical intensity (what we will term intrinsic salience). More physically intense stimuli will be learned about more easily. A stimulus’s physical intensity cannot be modulated, but the perception of its intensity can be. Second, associative learning theories tell us that past experience with the stimulus can also influence the amount of attention that is paid to it (what we will term acquired salience). This affects what is known as its associability – how easily it may be learned about – prototypically demonstrated in latent inhibition (LI). Thus past experience can also modulate the effective salience of a stimulus. Salience differences, both learned and unlearned, can subsequently facilitate or retard learning. Although learning differences will not be the only outcome when effective salience is manipulated, this variable can be quantified by measuring how easily the stimulus undergoes associative learning. This volume does more than mark the anniversary of an intriguing behavioural effect (Lubow & Moore, 1959). As a reliable cross-species learning phenomenon, LI is now widely accepted in the neuroscientific community as an essential translational tool for probing the neural substrates of schizophrenia. From a clinical approach, independent lines of enquiry have focused on aberrant processing of stimulus salience Latent Inhibition: Cognition, Neuroscience and Applications to Schizophrenia, ed. R. E. Lubow and Ina Weiner. Published by Cambridge University Press # Cambridge University Press 2010.
342
Latent inhibition and other salience modulation effects
343
in producing the cognitive abnormalities of schizophrenia (Bleuler, 1911; Kapur, 2003, 2004). The present chapter will touch on applications to schizophrenia: we rely on the rest of the volume to persuade the reader that LI has proven to be translationally relevant for examining schizophrenic dysfunction (see Kumari & Ettinger, this volume; Lubow, this volume; Swerdlow, this volume; Weiner, this volume; also see Lubow, 2005). We first review the roles of intrinsic and acquired salience in LI. This will be followed by comparison between LI and other selective learning effects that depend to varying extents on salience modulation. We will then review what is known about the underlying psychological and neural substrates of these learning effects and how they compare with those of LI. Such comparisons have the potential to identify the underlying mechanisms necessary for normal salience modulation. We will then consider the neural substrates of these selective learning effects, discuss how the action of dopamine (DA) may differ between them, and consider the implications for understanding diseases associated with abnormal salience processing such as schizophrenia.
Salience processing in LI Since the discovery of LI (Lubow & Moore, 1959), the effect has provoked controversy. One early view was that LI arises because of reduced salience of the stimulus as a result of past experience of the stimulus without consequence in the non-reinforced pre-exposure phase (e.g., Mackintosh, 1975). Specifically, low salience of the pre-exposed stimulus would be learned. In a related vein, LI has been compared with the habituation that is produced by repeated stimulus presentations during the pre-exposure phase. Habituation can be seen as attentional in that it can refer to the decline in the orienting response, both within and between sessions of stimulus presentation. The latter “long-term” effect has been identified with LI, based on arguments that the orienting response measure of habituation is also said to provide an index of stimulus associability – the loss of associability being the key effect of LI training (Swan & Pearce, 1988). Pearce and Hall (1980) extended this argument, suggesting that the associability of a stimulus declines as its associative strength increases. In support of this assertion, Kaye and Pearce (1984) showed that the orienting response to the conditioned stimulus (CS) declines as its associative strength increases over the course of training; then they went on to demonstrate that the same hippocampal lesions that attenuate LI result in a decline in the orienting response consistent with reduced associability (Kaye & Pearce, 1987). However, LI and long-term habituation have been dissociated behaviourally by their sensitivity to contextual changes. Hall and Channell (1985) showed in the same procedure that a change in context that does not affect habituation (as measured by the orienting response) was sufficient to disrupt the retardation of learning produced by the LI training. Habituation and LI are also affected in opposite ways by changes in intrinsic salience. Paradoxically, stimuli that produce larger orienting reactions, which take longer to show habituation with repeated
344
Helen J. Cassaday and Paula M. Moran
presentations, produce more rather than less LI (see reviews by Hall, 1991; Lubow, 1989). Indeed pre-exposure to diffuse configurations of stimuli sometimes results in the opposite effect of improved (“perceptual”) learning. For example, context preexposure can facilitate (rather than delay) the formation of future context–UCS association (Rudy & O’Reilly, 1999). A fuller account of the interrelationship between habituation and LI is given by Honey, Iordanova and Good (this volume). In any event, there is more to LI than a pre-exposure effect because by definition the effect is demonstrated on later associative learning. Knowledge of “irrelevance” is established in stage 1 of the procedure, by repeated presentation of a stimulus (e.g., a noise) without any consequences. In stage 2, the prior stimulus pre-exposure reliably retards any subsequent conditioning to that same noise stimulus, even though it now predicts a motivationally significant outcome such as food or foot-shock delivery. Thus, when LI is abolished by drug or lesion treatments, mechanisms underlying the changes occurring in the initial pre-exposure stage of LI are unlikely to be the whole story. In a two-stage procedure, when LI is abolished, there is relatively little effect in the second stage of the procedure – conditioning – of previous non-reinforcement of the stimulus. Consistent with the importance of mechanisms underlying the conditioning stage of LI, interference has been advanced as an explanation of normal LI in the rat. The idea is that the association formed in pre-exposure “this stimulus signals nothing” normally disrupts conditioning “this stimulus signals something” through associative or behavioural interference with learning or performance (Bouton, 1993; Kraemer & Roberts, 1984; Kraemer & Spear, 1992; Weiner, 1990, 2003). However, it should be noted that the best-developed retrieval account of LI locates the site of the interference produced by pre-exposure to the later expression of previous learning that is measured by test presentations in a third stage of the procedure (Miller & Matzel, 1998; Escobar, Arcediano & Miller, 2002a; Escobar, Arcediano, Platt & Miller, 2004; Escobar, Oberling & Miller, 2002b). This distinction will be returned to in discussion of how impairments in LI should be interpreted, and the implications of drug studies that confine treatments to different stages of the LI procedure. We will first compare alternative tasks involving salience modulation, at the level of both the psychological processes and the underlying neural substrates.
Other salience modulation effects To recap, the effects of salience can be direct and immediate, for example if an intrinsically loud noise is a more effective conditioned stimulus than a dim light. Alternatively, the effects of salience can be acquired, as is the case in LI. There are other procedures in which prior knowledge retards later learning and in this sense (reduced) salience is acquired, for example blocking (Kamin, 1968). Blocking experiments demonstrate that if a stimulus, say a light, is conditioned as a signal for an outcome in the presence of a pre-established signal for that same outcome, say a noise, learning about the light is poor, compared to the case in which the noise and
Latent inhibition and other salience modulation effects
345
Table 15.1. Basic design to show blocking Experimental stage group 1. Conditioning control 2. Overshadowing control 3. Blocking
Pre-training
Conditioning
Test
Aþ
Bþ ABþ ABþ
A,B A,B A,B
Notes: A and B are counterbalanced CSs; þ denotes reinforced with the UCS outcome. Blocking is best demonstrated as reduced conditioning to B in group 3 relative to group 2. Overshadowing is demonstrated as reduced conditioning to B in group 2 relative to group 1.
light are conditioned together but the noise has not been pre-trained. In this case prior experience with the noise impedes learning about the light. In the blocking group, the additional stimulus is effectively redundant in the absence of any change in the quantity or quality of food or foot-shock delivered. Importantly blocking procedures need to control for the loss of conditioning due to the cue competition when two stimuli are presented in compound, so that overshadowing can be distinguished from the reduced salience due to redundancy with respect to predicting a biologically significant outcome (Table 15.1). Overshadowing is the essential control group to show blocking because it has long been established that the associability of an otherwise effective CS is reduced by the presence of a competing stimulus (Mackintosh, 1976). Overshadowing is itself shown as reduced conditioning to B in group 2 relative to group 1 (Table 15.1). In its own right overshadowing is an important effect in studies of stimulus selection for learning. A fundamental consideration arises in that both LI and blocking are twostage procedures in which the effects of prior experience must transfer to a subsequent conditioning stage to influence associability. Overshadowing, in contrast, is a simple one-stage conditioning task to test effects on salience modulation without invoking the need explicitly to compare present contingencies with prior experience as is the case in two-stage learning tasks. Thus overshadowing procedures provide the cleanest example of learning differences based on intrinsic salience. However, the distinction between intrinsic and acquired salience is still not absolute. For example, under some conditions, overshadowing can be reduced by prior experience with the CS (Carr, 1974; Oberling, Bristol, Matute & Miller, 2000). Finally, to the extent that interference theories of LI are correct and LI depends in part on processing competing associations, then other learning phenomena can be used to study the underlying substrates. For example, despite similarity in name, conditioned inhibition involves processes distinct from LI. Conditioned inhibition is demonstrated as a discrimination that is demonstrated when the meaning of the CS is qualified by an additional stimulus (Pearce & Hall, 1980; Wagner, 1981). Specifically, whilst the CS presented alone reliably predicts the UCS outcome, when presented in
346
Helen J. Cassaday and Paula M. Moran
conjunction with the conditioned inhibitor the otherwise expected outcome will not occur. The resulting inhibitor acquires the ability to counteract the effect of other CSs for the same UCS (Rescorla, 1969). Both the acquisition and expression of conditioned inhibition thus depend on the ability to process two associations simultaneously, and in this sense conditioned inhibition procedures provide a parallel with what is said to occur in LI, according to interference accounts.
Different psychological processes? Before consideration of whether the neural substrates of LI and other salience modulation effects are the same, it is important to first consider how far analysis of their underlying psychological processes supports the idea that there is overlap. In general terms, the theories of associative learning have been developed to explain selective learning phenomena such as LI, overshadowing and blocking. The proposed mechanisms whereby such selectivity is achieved vary between theories. For example, a particular success for the Rescorla–Wagner theory (1972) was the prediction of one-trial blocking. The stage-1 pre-training in the blocking procedure means that the UCS is already fully predicted; there is no further associative strength to be allocated and nothing further to be learned. By contrast, Rescorla–Wagner (1972) did not predict one-trial overshadowing because the trial-1 error term (or discrepancy between what is expected and what actually occurs) cannot be affected by the presence of the additional overshadowing stimulus. Thus, consistent with a distinction between acquired and intrinsic salience, behavioural analysis based on Rescorla–Wagner (1972) would suggest that blocking and overshadowing are different phenomena. Similarly, the mechanisms underlying LI and overshadowing were assumed to be different (Rescorla & Wagner, 1972; Wagner, 1981). Of course the field has progressed since 1972 and one-trial overshadowing was subsequently demonstrated (James & Wagner, 1979; Mackintosh & Reese, 1979). Precisely this kind of evidence leads to revision of theory. Indeed, Wagner (1981), moving on to consider factors intrinsic to the stimuli, did predict one-trial overshadowing. This theory assumes limited-capacity processing: for learning to occur, the elements that represent the CS are required to be in a state of full activation; the presence of the overshadowing stimulus will restrict the number of CS elements that can be fully active. Accordingly, less salient CS elements will disappear from the animal’s attentional focus. The other post-1972 attentional theories can explain LI and overshadowing in the same way, consistent with shared psychological mechanisms underlying salience modulation, irrespective of whether the salience at issue is acquired or intrinsic. For example, they explain LI and overshadowing in terms of reduced associability, on account of pre-exposure and a learned loss of associability
Latent inhibition and other salience modulation effects
347
suffered by the overshadowed cue, respectively (Mackintosh, 1975; Pearce & Hall, 1980). However, in more recent theoretical developments, Mackintosh now attributes LI to a salience modulation process that is distinct from that mediating overshadowing. Specifically, LI is now explained in terms of reduced associability as elements of the CS become associated with each other and with contextual elements, rendering them unsurprising and thus ineffective for learning. By contrast, overshadowing relies on no such acquisition and (because of an additional feature of the model in terms of what counts as a learning trial) can be presumed to occur from the very first trial (McLaren & Mackintosh, 2000), consistent with the aforementioned experimental evidence (James & Wagner, 1980; Mackintosh & Reese, 1979). New “hybrid” models of associability, beyond the scope of this chapter, go even further (Le Pelley, 2004) to embrace evidence that under some circumstances associative history can impair (Hall & Pearce, 1979; Pearce & Hall, 1980) as well as enhance associability (Mackintosh, 1975). Le Pelley (2004) sees access to the learning mechanism or “attentional weight” in learning experiments with competition between simultaneously presented cues (Mackintosh, 1975) as essentially orthogonal to salience for learning based on previous training (Pearce & Hall, 1980). Therefore, in general terms, Le Pelley’s hybrid model echoes the distinction between intrinsic and acquired salience proposed here. However, behavioural analysis alone has yet to resolve the issue of how stimulus salience should best be conceptualised. The intimate relationship between theory and evidence introduces an inevitable circularity as the development of the general theories is driven by the behavioural evidence. Patterns of association and dissociation in terms of underlying substrates will provide an additional level of analysis to determine whether different varieties of learning effects (measured for example in blocking, overshadowing, and conditioned inhibition procedures) are functionally equivalent or not. The evidence reviewed above clearly points to the conclusion that these different tasks are unlikely to show complete overlap in terms of underlying substrates. How far the differences are attributable to additional processes, beyond salience modulation and not common to different kinds of selection for learning, remains to be determined. Figure 15.1 shows how we might imagine the differential and overlapping contributions of salience modulation and other underlying mechanisms in LI, blocking and overshadowing. The extent to which patterns of similarity and difference map onto a broad distinction between intrinsic and acquired salience similarly remains to be determined. However, this kind of simple taxonomy would be appealing because two such distinct routes to modulating stimulus salience map onto the distinction drawn more widely in psychology between “bottom-up” and “top down” processing. To ground different aspects of salience modulation in terms of their biological bases will help to distinguish between theories developed at the behavioural level of analysis, as well as potentially advance our understanding of schizophrenia. If the
348
Helen J. Cassaday and Paula M. Moran
Figure 15.1. The interrelationship between salience modulation and some selective learning effects (LI, blocking, overshadowing, and trace conditioning) in terms of shared underlying mechanisms as represented by the degree of overlap in the Venn diagram.
salience modulation that is central to LI is the same as that involved in blocking, overshadowing and other selective learning phenomena then shared neural substrates should be clearly identifiable.
Same neural substrates? The DA system, nucleus accumbens (n.acc) in particular, is best known as a substrate of reward and reinforcement (Ikemoto, 2007; Koob, 1992; Schultz, 1997; Wise, 2004; Wise & Rompre, 1989). However, there is mounting evidence that n.acc is activated by the salience of stimuli rather than just their reinforcing properties. This was definitively shown in a sensory pre-conditioning paradigm where conditioning occurs to stimuli that are not directly reinforcing or paired with a reinforcer (Young, Ahier, Upton, Joseph & Gray, 1998). It is now clear that the n.acc and its dopaminergic innervation form a crucial component of the neural circuitry of LI (Jongen-Relo, Kaufmann & Feldon, 2002; Joseph, Peters, Moran et al., 2000; Restivo, Passino, Middei & Ammassari-Teule, 2002; Tai, Cassaday, Feldon & Rawlins, 1995; Weiner, 2003; Weiner, Bernasconi, Broersen & Feldon, 1997; Weiner, Gal, Rawlins & Feldon, 1996). The simplest variant of hypothesis as to underlying mechanism would state that DA activity makes stimuli seem more salient and thereby enhances learning about them, despite the pre-exposure (Gray, Feldon, Rawlins et al., 1991; Gray, Kumari, Lawrence & Young, 1999; Gray, Moran, Grigoryan et al., 1997; Kumari, Cotter, Mulligan et al., 1999; Solomon, Crider, Winkelman et al., 1981; Weiner, Lubow & Feldon, 1981, 1988). The neurobiology of LI has already been extensively reviewed within the present volume (Gould; Weiner; Louilot et al.; Honey, Iordanova, & Good). Non-dopaminergic systems clearly play an important role in LI and related learning effects. We focus on the role of DA and brain structures connected to n.acc
Latent inhibition and other salience modulation effects
349
because the number of comparison points is overall highest. Table 15.2 provides a summary of some of the reported drug and lesion effects on LI, blocking and overshadowing, behavioural effects that have been most thoroughly investigated as models for schizophrenic attention deficit. It is necessarily incomplete because for the most part these comparisons have not been systematic: Table 15.2 includes only studies where data from two or more studies of LI, blocking and overshadowing were available. As can be seen, the evidence for equivalence in the neural substrates of LI and blocking is mixed. The results obtained with noradrenergic dorsal bundle lesions and serotonergic raphe lesions are contradictory. With respect to the hippocampal connection, the apparent inconsistency in terms of the outcome of lesion studies is reflected in the fact that the direction of effects reported is variable (some dopaminergic lesions also generate persistent LI: Weiner, 2003; Weiner, Gal & Feldon, 1999), perhaps as a result of differences in procedures (see section CS–UCS interactions and salience modulation below). However, the evidence pertaining to the dopaminergic substrates of LI, blocking and overshadowing consistently underscores their importance. In fact, there are a number of diverse lines of evidence which taken together suggest that DA and its neural pathways are as important for blocking as for LI. Similar to LI, patients with schizophrenia have deficits in blocking, as has been shown using a variety of learning paradigms (Bender, Muller, Oades & Sartory, 2001; Jones, Gray & Hemsley, 1992; Jones, Hemsley, Ball & Serra, 1997; Moran, Al-Uzri, Watson & Reveley, 2003; Moran, Owen, Crookes et al., 2008; Oades, Rivet, Taghzouti et al., 1987; Serra, Jones, Toone & Gray, 2001). Importantly, this blocking deficit has been reported over and above any change in overshadowing (Bender et al., 2001; Jones et al., 1997; Moran et al., 2003, 2008). In electrophysiological studies of primates, the development of blocking has been shown to be mirrored by reduced dopaminergic neuronal firing in the ventral tegmental area as compared to control groups (Schultz, 2006; Waelti, Dickinson & Schultz, 2001), demonstrating a dopaminergic neuronal correlate to the phenomenon. Like LI, blocking has been reported to be abolished by treatment with amphetamine in the rat (Crider, Solomon & McMahon, 1982; Ohad, Lubow, Weiner & Feldon, 1987; O’Tuathaigh, Salum, Young et al., 2003). Intriguingly from the perspective of comparison with LI, n.acc has recently been identified as a neural substrate of blocking (Iordanova, Westbrook & Killcross, 2006). In this study, direct injection of the indirect DA agonist amphetamine into n.acc enhanced blocking. Conversely, blocking was reduced by injection of a mixed DA D1/D2 antagonist but not by a more selective D2 antagonist. In contrast with these blocking results, in LI procedures direct injection of amphetamine in n.acc resembles the effect seen with systemic treatment in that it abolishes LI (Solomon & Staton, 1982), and dopaminergic effects in LI have been found to be DA D2 mediated (see section Different psychopharmacology? below). At face value, these differences point to different underlying mechanisms in LI and blocking procedures. However, the procedure used by Iordanova et al. (2006) was unconventional in that
YES Weiner (2003, for review)
YES Solomon & Staton (1982) NO Joseph et al. (2000) YES Kaye & Pearce (1987) Schmajuk et al. (1994)
Amphetamine disruption reversed by antipsychotic drugs?
Amphetamine infusion into nucleus accumbens
Hippocampal lesion
YES Weiner (2003, for review)
Amphetamine (systemic)
Latent inhibition
YES Solomon (1977) Rickert et al. (1978, 1981) Gallo & Candido (1995)
YES Ohad et al. (1987) O’Tuathaigh et al. (2003) Crider et al. (1982) YES Crider et al. (1982) But see Iordanova et al. (2006) YES Iordanova et al. (2006)
Blocking
Table 15.2. Some studies of drug and lesion effects on latent inhibition, blocking and overshadowing
YES Holland & Fox (2003) Schmajuk et al. (1983)
NO O’Tuathaigh & Moran (2004) ?
YES O’Tuathaigh et al. (2002, 2003) Horsley et al. (2008)
Overshadowing
NO Tsaltas et al. (1984) Lorden et al. (1983) YES Lorden et al. (1983) Asin et al. (1980)
NO Gallo & Candido (1995) Coutureau et al. (1999) Honey & Good (1993) YES Nicholson & Freeman (2002)
NO Lorden et al. (1980)
NO Holland & Fox (2003) Allen et al. (2002) Good & MacPhail (1994) YES Nicholson & Freeman (2002) YES Lorden et al. (1980)
Note: YES indicates impairment. NO indicates no impairment but does not preclude enhancement.
Raphe lesion
Dorsal bundle lesion
Medial thalamic lesion
Hippocampal lesion
?
?
?
NO Garrud et al. (1984) Good & Macphail (1994)
352
Helen J. Cassaday and Paula M. Moran
the blocking (and overshadowing control) stimulus was the experimental context. The neural substrates of discrete cue and contextual conditioning are not identical (e.g., Winocur, Rawlins & Gray, 1987), and dissociable drug effects mediated in n.acc were earlier demonstrated in the same kind of fear conditioning procedure (Westbrook, Good & Kiernan, 1997). Furthermore, the micro-injections used by Iordanova and colleagues were centred on core rather than shell n.acc. Shell n.acc is the appropriate target through which to reproduce the effects of systemic amphetamine on LI and related effects (Weiner, 2003; Weiner et al., 1996, 1999). In contrast with the pattern of effects reported by Iordanova et al. (2006) and consistent with DA D2 mediation, an earlier study reported that the amphetamine-induced disruption of blocking was haloperidol-reversible (Crider et al., 1982). More generally, when the required overshadowing control for blocking is in place (Table 15.1; see section Other salience modulation effects above), further changes in learning that are unequivocally mediated through (effects on) blocking can be difficult to demonstrate (Cassaday, 2009). Although relatively little is known about the neural basis of overshadowing (Garrud, Rawlins, Mackintosh et al., 1984; Good & Macphail, 1994), the DA system is clearly involved (Oades et al., 1987; O’Tuathaigh & Moran, 2002, 2004; O’Tuathaigh et al., 2003). Using an appetitive procedure, we recently found that systemic amphetamine abolished overshadowing in the absence of any effect, in the same overshadowing procedure, of conventional lesions to n.acc. On the other hand, LI is known to be disrupted by both lesions to the n. acc shell and systemic amphetamine, leading to the hypothesis that systemic amphetamine effects are mediated via effects on DA in this structure (Weiner, 1990, 2003; Weiner et al., 1996, 1999). The absence of lesion effects in overshadowing, which suggests that n.acc does not provide a common path for salience modulation, was not the predicted outcome, and we discussed a number of possible explanations for this discrepancy (Horsley, Moran & Cassaday, 2008). The possible contribution of task motivation is discussed in more detail below. However, if task motivation were the explanation, n.acc shell lesions should be similarly without effect on appetitive LI, which has yet to be tested. Second, complete n.acc. lesions that take out both shell and core subfields also spare LI (Jongen-Relo et al., 2002; Konstandi & Kafetzopolous, 1993; Tai et al., 1995; Weiner et al., 1996). Therefore, to the extent that the same psychological processes underlie (the abolition of) overshadowing and LI, inadequate selectivity of the lesion method could in principle explain the failure to abolish overshadowing. However, this possibility was dismissed as the electrolytic lesions that we used showed good selectivity for shell and core subfields of n.acc and the shell lesion increased unconditioned responding, both for food and to the overshadowing tone stimulus (Horsley et al., 2008). Furthermore, the same shell and core lesions showed differential effects in other conditioning procedures (Cassaday, Horsley & Norman, 2005). We therefore concluded that structures outside n.acc must be involved in overshadowing (Horsley et al., 2008). This conclusion does not preclude a role in overshadowing for the wider range of brain structures known to be involved in LI
Latent inhibition and other salience modulation effects
353
Neocortex Prelimbic
Entorhinal cortex
Frontal Infra- cortex limbic
Dorsal Subiculum Ventral
Core Shell Nucleus accumbens
Lateral Ventral pallidum Medial
Nucleus reticularis thalami
Thalamocortical sensory relay nuclei Mediodorsal thalamic nucleus
Figure 15.2. Some of the key nucleus accumbens connections that are implicated in salience modulation, based largely on the known effects of lesions in latent inhibition and including output to perceptual systems (developed from Gray et al., 1999).
(Gray et al., 1999). Figure 15.2 shows the multi-synaptic circuitry identified with salience modulation through lesion studies of LI. This links n.acc with cortical areas in a number of parallel loops. Thalamic structures and frontal cortex would seem likely candidate substrates for overshadowing. The thalamus has earlier been identified with salience modulation because of its key role as a sensory relay with cortical loops that link n.acc output to perceptual systems (Gray et al., 1999). Thalamic lesions have been shown to disrupt functions mediated by n.acc, including LI and blocking (Nicholson & Freeman, 2002). Frontal cortex has been identified because – in addition to its direct projections with n.acc – of its definitive connections with the mediodorsal thalamic nucleus (Rose & Woolsey, 1948) that in turn receives projections from n.acc, particularly from the shell (O’Donnell, Lavin, Enquist et al., 1994). The role of frontal cortex in LI is currently under investigation. The documented role of the frontal cortex in holding task-relevant information “on-line in working memory” (Levy & Goldman-Rakic, 2000) is also consistent with a role for frontal cortex in overshadowing. However, conventional lesion studies on their own only tell us whether a structure is necessary to the behaviours under test; they do not tell us about its neuromodulatory actions. In LI, conventional lesions reliably reproduce the effects of behaviours mediated through DA D2-like receptors, because the effects of agonists at these receptor subtypes are inhibitory. By contrast, DA D1-like receptors are positively coupled to adenylate cyclase and thus excitatory. Therefore, if the abolition of
354
Helen J. Cassaday and Paula M. Moran
overshadowing produced by amphetamine is mediated by DA D1-like receptors, conventional lesions are a priori unlikely to reproduce this effect. This possibility is returned to below. Finally, although little is known about the neural substrates of conditioned inhibition, mesolimbic DA appears to be a plausible substrate. This possibility is consistent with the moderating effects on conditioned inhibition of reward sensitivity and schizotypy, demonstrated in human studies (Migo, Corbett, Graham et al., 2006). In the rat, treatment with amphetamine has been reported to enhance the acquisition of conditioned inhibition (Harmer & Phillips, 1999). In summary, lesion and pharmacological studies applying drugs directly into the brain suggest that LI shares common dopaminergic substrates with blocking and overshadowing and the same brain regions may be involved in conditioned inhibition. However, one of the few studies of the neural substrates of blocking suggests the possibility of dissociation in terms of the critical DA receptor subtypes (Iordanova et al., 2006). We next turn to the evidence that the psychopharmacological mechanisms through which DA mediates its effects are dissociable, both within LI, depending on the stage of procedure, and between LI and overshadowing.
Different psychopharmacology? Drug studies have greatly advanced our understanding of the bases of disrupted LI by the use of treatments selectively introduced at the pre-exposure and conditioning stages of the procedure (Feldon, Shofel & Weiner, 1991; Feldon & Weiner, 1989; Hitchcock, Lister, Fischer & Wettstein, 1997; Killcross, Stanhope, Dourish & Piras, 1997; LaCroix, Spinelli, Broersen & Feldon, 2000; Weiner & Feldon 1987; Weiner, Lubow & Feldon, 1984). Thus, within the LI procedure itself, dissociations in terms of where drugs are effective tell us through which mechanisms they are effective. For example, amphetamine-induced abolition of LI is very clearly demonstrated, but only providing that the drug is present in the conditioning stage (Joseph et al., 2000; Weiner, 2003; Weiner et al., 1984, 1988; Young, Moran & Joseph, 2005, for review). This pattern of effects naturally fits with interference-based accounts of LI (e.g., Bouton, 1993; Kraemer & Roberts, 1984; Kraemer & Spear, 1992). It is not so readily explained by interference-based accounts that emphasise the retrieval of conflicting associations when conditioning is later tested (Escobar et al., 2002a, 2002b, 2004; Miller & Matzel, 1998). Similarly, drug effects mediated at conditioning are apparently inconsistent with the idea that the demonstration of LI rests on changes in the effective salience or associability of the to-be-conditioned CS due to its inconsequential presentation during the pre-exposure phase (see section Salience processing in LI above 1). The effects of drugs on LI mediated at the pre-exposure phase have been demonstrated with compounds that affect GABA (Feldon & Weiner, 1989; LaCroix et al., 2000), as well as with serotonergic compounds (Hitchcock et al., 1997;
Latent inhibition and other salience modulation effects
355
Killcross et al., 1997; Weiner, 2003, for review). On the other hand, it is much harder to modulate LI with dopaminergic treatments that are confined to the pre-exposure phase (Weiner, 2003; Weiner et al., 1984, 1987). Under some circumstances, the indirect agonists nicotine and amphetamine, when administered in the pre-exposure stage, have disrupted LI, while haloperidol enhanced LI, but only when pre-exposure was conducted over a number of days (Bethus, Muscat & Goodall, 2006; Schmajuk, Gray & Larrauri, 2005). However, pre-exposure conducted over several days is likely to engage additional processes because of an increased memory component. This would follow because memory resources would be required for spaced pre-exposures to have an equivalent outcome to massed pre-exposure. The precise procedural variant of LI in use may also be a factor. Specifically, task motivation may be an important determinant of the underlying substrates that must be distinguished from those key to salience modulation per se (see section CS–UCS interactions and salience modulation below). Dissociations with respect to distinct aspects of salience modulation measured in other procedures have started to emerge, particularly with reference to the different receptor families in the DA system. Pharmacological studies have shown that the DA D2 receptor subserves the differences in associability that arise from past experience with the CS as measured in LI. Amphetamine, which releases DA, disrupts LI, while DA antagonists which act at the D2 site (haloperidol, raclopride, sulpiride and clozapine) selectively reverse amphetamine-induced disruption of LI and potentiate LI when given alone (Dunn, Atwater & Kilts, 1993; Feldon & Weiner, 1991; Moran, Fischer, Hitchcock & Moser, 1996; Moran & Moser, 1992; Moser, Lister, Hitchcock & Moran, 2000; Russig, Kovacevic, Murphy & Feldon, 2003; Russig, Murphy & Feldon, 2002; Shadach, Gaisler, Schiller & Weiner, 2000; Solomon et al., 1981; Weiner, 2003; Weiner & Feldon, 1987, 1997; Weiner, Feldon & Katz, 1987; Weiner et al., 1981). Conversely, there is evidence that the D1 receptor subserves differences in associability that arise from intrinsic salience (the physical aspects of the CS) as measured in overshadowing. Amphetamine also abolishes overshadowing and this effect is reversed by the selective D1 antagonist SCH 23390, but not by D2 antagonists such as haloperidol, raclopride or sulpiride. In addition, the partial D1 agonist SKF 38393 abolishes overshadowing (O’Tuathaigh & Moran, 2002, 2004; O’Tuathaigh et al., 2003). Furthermore, the effect of haloperidol in enhancing LI is not readily reproduced with DA D1 antagonists (Trimble, Bell & King, 2002). These findings suggest the hypothesis that the D1 and D2 subclasses of receptor are dissociably involved in different selection processes for learning: D1-like receptors in encoding stimulus salience determined by stimulus intensity, as measured in overshadowing procedures; D2-like receptors in associability as determined by past experience with the stimulus, as measured in LI procedures. In general, although the interpretation of dissociations in terms of neuropharmacology rather than mere neuroanatomy has been neglected, there are some relevant precedents in studies of addiction (Volkow, Fowler, Wang & Goldstein, 2002) and pre-pulse inhibition (Peng, Mansbach, Braff & Geyer, 1990).
356
Helen J. Cassaday and Paula M. Moran
Of course, pharmacological dissociations may not be clear-cut. For example, it is already known that amphetamine effects on LI are not, as should be predicted, reproduced by other more selective agonists for DA D2-like receptor families (Feldon et al., 1991). These findings suggest that a balance of receptor activations (or interactive effects) may mediate the actions of amphetamine in LI, as seems to be the case for unconditioned behaviours (Canales & Iversen, 2000) and blocking (Iordanova et al., 2006; but see Crider et al., 1982). Further systematic comparison with additional learning effects such as blocking and conditioned inhibition will be needed to clarify the role of the DA D1 and DA D2 receptor families, as well as other known neuromodulators of LI (Weiner, this volume). However, for some such comparison tasks, there is evidence to suggest that paradigm variations – and in particular how the task is motivated – also should be taken into account.
CS–UCS interactions and salience modulation It now well accepted that stimuli, or rather particular combinations of stimuli, do not start out equally effective in conditioning. In animals, some CS–UCS relationships are learned more easily than others (Garcia & Koelling, 1966; Shapiro, Jacobs & LoLordo, 1980). In humans, CS salience ratings vary by modality depending on the outcome to be predicted (Barklamb & Cassaday, 2007). In other words there is a role for stimulus– reinforcer interactions and, in the absence of prior experience, some would argue that these can be innate as a result of evolutionary pressures (Rozin & Kalat, 1971). A particularly clear demonstration is provided in the case of conditioned taste aversion (CTA): flavours are much more readily associated with subsequent illness than are other stimuli, and on the basis of a single experience (Garcia & Koelling, 1966). As well as providing evidence for the role for pre-determined salience such stimulus–reinforcer interactions are likely to influence the susceptibility of salience modulation to dopaminergic treatments. There is clear evidence on this point in the case of CTA. Although a CTA procedure for LI showed the same sensitivity to systemic amphetamine as the more widely used conditioned emotional response (CER) procedure (Russig et al., 2003), there was a striking discrepancy with respect to the role of n.acc. In contrast to the expected disruption (Jongen-Relo et al., 2002; Joseph et al., 2000; Restivo et al., 2002; Tai et al., 1995; Weiner, 2003), shell lesions enhanced LI in a CTA procedure (Pothuizen, Jongen-Relo, Feldon & Yee, 2006). Hippocampal lesions too have contrary effects on LI when tested with CTA (Purves, Bonardi & Hall, 1995; Reilly, Harley & Revusky, 1993). Only with a CTA procedure can LI be abolished by amphetamine treatment confined to stimulus pre-exposure (Bethus et al., 2006), in contrast to results with CER (Weiner, 2003; Weiner et al., 1988). (An alternative explanation for this exceptional outcome from stage of procedure drug studies of LI was discussed in section Different psychopharmacology above.) Taste aversion is undoubtedly the most extreme case of how the nature of the UCS can interact with salience modulation of the CS.
Latent inhibition and other salience modulation effects
357
Table 15.3. Consistent amphetamine-induced abolition of latent inhibition and overshadowing in aversively and appetitively motivated procedures Task
Latent inhibition
Overshadowing
Aversive
Solomon et al. (1981) Weiner et al. (1981, 1988, etc.) Joseph et al. (2000) Russig et al. (2003) Killcross et al. (1994) Norman & Cassaday (2004) N.S. Gray et al. (1992) Kumari et al. (1999)
Oades et al. (1987) O’Tuathaigh & Moran (2002, 2004) O’Tuathaigh et al. (2003)
Appetitive
Horsley et al. (2008)
Although appetitive LI has been under-investigated, sensitivity to disruption under amphetamine, comparable to that seen in the typical CER procedure, has definitely been shown (Killcross, Dickinson & Robbins, 1994; Norman & Cassaday, 2004). Aversive CER and appetitive overshadowing studies also show comparable sensitivity to disruption under amphetamine (Table 15.3). Given the above discrepancies reported with CTA procedures, the consistency (with respect to sensitivity to the disruptive effects of amphetamine) between the CER and appetitive tests of LI and overshadowing suggests that the nature of the aversive motivation matters (see also Moser et al., 2000).
Dissociable aspects of salience modulation and schizophrenia We have considered salience modulation in a variety of experimental procedures that can be applied to our understanding of human disease states. Of these the LI-CER procedure has been most widely investigated in animals. The LI model translates to schizophrenia in that the effect is absent in acute schizophrenia (Baruch, Hemsley & Gray, 1988). Moreover it can be abolished by amphetamine (Gray, Pickering, Hemsley et al., 1992; Kumari et al., 1999) and (under some circumstances) enhanced by haloperidol in humans as well as in rats (Kumari et al., 1999; Williams, Wellman, Geaney et al., 1996, 1997). Such disruptions of LI, which result in better conditioning to (previously) irrelevant stimuli as compared to healthy subjects, are used to model the attentional deficits seen in schizophrenia (for reviews see Gray et al., 1991, 1997, 1999; Weiner 1990, 2003; Weiner & Feldon, 1997). Coincident with these findings, there has been a recent revival of the view that in schizophrenia a dysregulated hyperdopaminergic state leads to an aberrant assignment of salience to the elements of experience (Bleuler, 1911; Gray et al., 1991, 1997, 1999; Kapur, 2003, 2004). Symptoms such as delusions arise as a patient tries to make sense of these aberrantly salient experiences, whereas hallucinations reflect a direct experience of the aberrant salience of internal representations. Antipsychotic drugs through their actions on DA
358
Helen J. Cassaday and Paula M. Moran
Overshadowing Abnormal processing of intrinsic salience
DA D1 family in nucleus accumbens loops
Genetic predisposition Dysregulated dopamine system Environmental factors
Cognitive abnormalities (e.g. delusions) to explain aberrant salience Abnormal processing of acquired salience
DA D2 family in nucleus accumbens loops
DA D1 agonists to reduce abnormalities resulting from abnormal processing of intrinsic salience
DA D2 antagonists to block abnormalities resulting from abnormal processing of acquired salience
Latent inhibition
Figure 15.3. The hypothesized role of different dopamine receptor sub-families in mediating intrinsic versus acquired salience modulation, based on pharmacological studies of overshadowing and latent inhibition (developed from Kapur, 2004).
D2 receptors have been proposed to “dampen the salience” of these abnormal experiences, thus alleviating symptoms. Salience, as we have discussed, is influenced by a multitude of factors. Thus, a better understanding of the psychological processes underlying LI and other kinds of salience learning will be fundamental to our understanding of schizophrenia. But what are the underlying mechanisms and how do findings from LI best relate to wider hypotheses of DA function? It has to be acknowledged that evidence on the neural substrates of salience modulation is as yet incomplete. The same structures may not be necessary for overshadowing (Horsley et al., 2008). However, in principle, the same multi-synaptic circuitry identified with LI through lesion studies could modulate salience processing as measured in overshadowing procedures (Figure 15.2). In other tests of learning and attention, there is evidence – in terms of sensitivity to lesions – that this circuitry is dissociably involved in top-down and bottom-up processing (Myers, Gluck & Granger, 1995; Rossi, Bichot, Desimone & Ungerleider, 2007; Stuss, 2006). Such dissociations are consistent with the picture emerging here of distinct neuropharmacological bases for dissociable aspects of salience encoding (Figure 15.3). This generates a number of testable predictions. For example DA D2 antagonists and D1 agonists should be differentially effective on distinct aspects of the attentional disturbance of schizophrenia.
Latent inhibition and other salience modulation effects
359
The evidence for dissociable aspects of salience modulation, shown in Figure 15.3, is drawn from conditioning studies in animals that for the most part use CER procedures. The interactions between task motivation and salience modulation (see section CS–UCS interactions and salience modulation above) could be viewed as a limitation on the usefulness of this particular animal model of selective learning. However, CER procedures produce results consistent with those found in appetitive procedures (Table 15.3). Moreover, differences in motivational state might be an important determinant of salience processing in schizophrenia, as is certainly the case in ADHD (Luman, Oosterlaan & Sergeant, 2005; van Meel, Oosterlaan, Heslenfeld & Sergeant, 2005). The clinical impression that patients with schizophrenia are particularly attuned to threat is supported by questionnaire measurement of their sensitivity to punishment and non-reward, conceptualised as an overactive behavioural inhibition system (BIS; Gray, 1982; relatedly, for the role of anxiety in LI, see Braunstein-Bercovitz, this volume). Moreover, this questionnaire study showed that higher BIS sensitivity correlated with duration of illness (Scholten, van Honk, Aleman & Kahn, 2006).
Switching and selective learning A full account of selective learning and its underlying biology must also include evidence for the role of processes beyond salience modulation. Following from the now well-established role of n.acc in switching between competing behaviours (Bakshi & Kelley, 1991a, 1991b; Evenden & Carli, 1985; Floresco, Ghods-Sharifi, Vexelman & Magyar, 2006; Robbins & Koob, 1980; Van den Bos & Cools, 2003), hypotheses of DA function also have encompassed a role in “attentional” switching between competing contingencies. A switching mechanism has already assumed prominence in analysis of the DA substrates of LI (Weiner, 1990, 2003). Selective lesion studies have located this switching mechanism in the n.acc core, which is normally inhibited by shell n.acc (Schiller, Zuckerman & Weiner, 2006; Weiner & Feldon, 1997). The idea is that LI is disrupted by shell lesions because of excessive switching to the new contingency (in our terms the moderating effects of acquired salience are lost). Conversely, persistent LI is produced by core lesions because of retarded switching (Weiner, 1990, 2003). The reduction in later learning produced by stimulus pre-exposure in LI may be attributed to stimulus pre-exposure setting up a contingency that interferes with the learning of a new contingency. This is not the only interpretation of LI but it is one that can explain why the effect is abolished by DA treatments confined to the conditioning stage of the procedure while these same treatments are typically ineffective when their administration is confined to pre-exposure (Joseph et al., 2000; Weiner et al., 1987, 1988; Young et al., 2005). This pattern of effects is compatible with those associative interference accounts of LI which allow that effects may be mediated at acquisition rather than exclusively at retrieval (Bouton, 1993; Kraemer & Roberts, 1984; Kraemer & Spear, 1992).
360
Helen J. Cassaday and Paula M. Moran
Similarly, accounts of schizophrenia have been framed in terms of the inability to use “memories of past regularities” (Gray et al., 1991; Hemsley, 1977, 1993). In rat studies, amphetamine has effects across a number of two-stage paradigms which are consistent with increased switching when normal performance depends on an interaction between the stimulus currently presented and stored memories of past contingencies (see Weiner, 1990, 2003). Increased switching under these circumstances would have the same outcome as a failure to be influenced by past regularities. Such a deficit could also be described in terms of the “output processes” of attention which are concerned with the retrieval of information from memory before responding and thus inherently top-down (Hemsley, 1977, 1993). The switching mechanism could be fundamental to the expression of acquired salience in all tasks that require comparison between competing contingencies. However, the outcome of switching would depend on the nature of the previously established contingency. For example, conditioned inhibitors have a negative or “dampening” effect on subsequent associative learning. This is established by prior training with a CS; the conditioned inhibitor is then introduced to signal that the otherwise expected UCS will not occur. The switching hypothesis of n.acc function could be extended to conditioned inhibition as follows: excessive switching between conflicting contingencies (e.g., after shell lesion) should promote conditioned inhibition because the representation of a non-event must rely on the concurrent presence of the excitatory association; conversely, retarded switching between contingencies (e.g., after core lesion) should impair conditioned inhibition because the comparison process that generates mismatch would be impeded. Therefore, the role of switching will depend critically on the learning task in question; and whilst normal switching is not synonymous with normal salience modulation, normal switching will be essential for the modulation of new learning by acquired salience.
Conclusions and implications LI has been a key phenomenon in understanding how learning is influenced by stimulus salience. In the present chapter, we have considered a number of other selective learning phenomena that involve salience modulation. Furthermore, we have reviewed the evidence that whilst, in part, they may rely upon the same neural substrates, there are pharmacological dissociations, and the range of underlying psychological processes is most certainly different. Behavioural analysis alone has already suggested some likely dissociation in terms of neuropharmacological substrates because selective learning tasks influence the salience of available stimuli in different ways. For selection based on the information content of stimuli, salience must necessarily be determined by factors beyond the intrinsic physical characteristics of the stimuli. Since one of these factors is the influence of past experience, we have proposed a distinction between intrinsic and
Latent inhibition and other salience modulation effects
361
acquired salience, as applied to associative learning phenomena that involve salience modulation, both of which may influence associative learning. Such a distinction accommodates observations that LI depends on both sources of salience. For example, studies restricting drug administration to different stages of the LI procedure can distinguish between effects on different aspects of salience processing. Although salience is generally agreed to be reduced through learning in LI, the effectiveness of the pre-exposure stage should initially be influenced by factors intrinsic to the stimulus (as well as salience learning). Later, in the conditioning stage of the procedure, acquired knowledge about the stimulus normally retards learning. This assumption generates testable predictions in that those treatments which unequivocally abolish LI through an effect mediated at pre-exposure (see section Different psychopharmacology above; Weiner, 2003, for fuller review) should abolish overshadowing. LI effects mediated at pre-exposure are already of particular interest in that they have predictive validity for the treatment of negative symptoms (Weiner, 2003). In terms of neural substrates, LI has been more widely investigated than other selective learning phenomena. Therefore, not all lesions with known effects on LI have been tested in all selective learning tasks. In some cases, because comparisons have largely to be made between rather than within studies, procedural differences such as task motivation confound the comparisons. As a consequence, our concluding comments on the equivalence of the neural substrates on LI and other kinds of salience modulation can only be tentative. That said, on balance, the available data suggest that the same neural substrates – or at least the same circuitry – are likely to be important for different types of salience modulation. However, dissociations have been found in psychopharmacological studies. The novel implication arising from these dissociations is that the DA D1 and DA D2 receptor families may be differently involved in associative learning processes that are differentially dependent on intrinsic and acquired salience. If this hypothesis withstands further testing, it will suggest novel strategies for the pharmacological treatment of schizophrenia (Figure 15.3) and it will find some accommodation within cognitive psychology. A fundamentally similar distinction can be found in other domains where different sets of procedures and measures have been devised to differentially engage top-down processes based on prior knowledge and bottom-up processes that are driven more directly by the immediate sensory input. As might be expected, these are not always clear distinctions. However, a better understanding of the role of salience in learning, one that includes reference to the variety of underlying substrates, will significantly improve our understanding of the theoretical bases of associative learning and attention. LI presents a paradigm case in this respect.
Acknowledgements We thank Charlotte Bonardi, Andrew Nelson and Karen Thur for their very helpful comments.
362
Helen J. Cassaday and Paula M. Moran
References Allen, M. T., Padilla, Y., Myers, C. E., & Gluck, M. A. (2002). Selective hippocampal lesions disrupt a novel cue effect but fail to eliminate blocking in rabbit eyelid conditioning. Cognitive Affective & Behavioral Neuroscience, 2, 318–328. Asin, K. E., Wirtshafter, D., & Kent, E. W. (1980). The effects of electrolytic median raphe lesions on two measures of latent inhibition. Behavioral and Neural Biology, 28, 409–418. Bakshi, V. P., & Kelley, A. E. (1991a). Dopaminergic regulation of feeding behavior: 1. Differential effects of haloperidol microinfusion into 3 striatal subregions. Psychobiology, 19, 223–232. Bakshi, V. P., & Kelley, A. E. (1991b). Dopaminergic regulation of feeding behavior: 2. Differential effects of amphetamine microinfusion into 3 striatal subregions. Psychobiology, 19, 233–242. Barklamb, K., & Cassaday, H. J. (2007). The salience of internal and external cues in cases of self-reported hunger and illness. Appetite, 49, 511–515. Baruch, I., Hemsley, D. R., & Gray, J. A. (1988). Differential performance of acute and chronic schizophrenics in a latent inhibition task. Journal of Nervous and Mental Disorders, 176, 598–606. Bender, S., Muller, B., Oades, R. D., & Sartory, G. (2001). Conditioned blocking and schizophrenia: a replication and study of the role of symptoms, age, onset-age of psychosis and illness-duration. Schizophrenia Research, 49, 57–170. Bethus, I., Muscat, R., & Goodall, G. (2006). Dopamine manipulations limited to preexposure are sufficient to modulate latent inhibition. Behavioral Neuroscience, 120, 554–562. Bleuler, E. (1911; translated 1950). Dementia Praecox or the Group of Schizophrenias. New York: International Universities Press. Bouton, M. E. (1993). Context, time, and memory retrieval in the interference paradigms of Pavlovian learning. Psychological Bulletin, 114, 80–99. Canales, J. J., & Iversen, S. D. (2000). Dynamic dopamine receptor interactions in the core and shell of nucleus accumbens differentially coordinate the expression of unconditioned motor behaviors. Synapse, 36, 297–306. Carr, A. F. (1974). Latent inhibition and overshadowing in conditioned emotional response conditioning in rats. Journal of Comparative and Physiological Psychology, 86, 718–723. Cassaday, H. J. (2009). Blocking, overshadowing and related concepts. In I. Stolerman (Ed.), Encyclopedia of Psychopharmacology. New York: Springer. Cassaday, H. J., Horsley, R. R., & Norman, C. (2005). Electrolytic lesions to nucleus accumbens core and shell have dissociable effects on conditioning to discrete and contextual cues in aversive and appetitive procedures respectively. Behavioural Brain Research, 160, 222–235. Coutureau, E., Galani, R., Gosselin, O., Majchrzak, M., & Di Scala, G. (1999). Entorhinal but not hippocampal or subicular lesions disrupt latent inhibition in rats. Neurobiology of Learning and Memory, 72, 143–157. Crider, A., Solomon, P. R., & McMahon, M. A. (1982). Disruption of selective attention in the rat following chronic D-amphetamine administration: relationship to schizophrenic attention disorder. Biological Psychiatry, 17, 351–361.
Latent inhibition and other salience modulation effects
363
Dunn, L. A., Atwater, G. E., & Kilts, C. D. (1993). Effects of antipsychotic drugs on latent inhibition: sensitivity and specificity of an animal behavioral model of clinical drug action. Psychopharmacology, 112, 315–323. Escobar, M., Arcediano, F., & Miller, R. R. (2002a). Latent inhibition and contextual associations. Journal of Experimental Psychology – Animal Behavior Processes, 28, 123–136. Escobar, M., Arcediano, F., Platt, T. L., & Miller, R. R. (2004). Interference and time: a brief review and an integration. Reviews in the Neurosciences, 15, 415–438. Escobar, M., Oberling, P., & Miller, R. R. (2002b). Associative deficit accounts of disrupted latent inhibition and blocking in schizophrenia. Neuroscience and Biobehavioural Reviews, 26, 203–216. Evenden, J. L., & Carli, M. (1985). The effects of 6-hydroxydopamine lesions of the nucleus accumbens and caudate nucleus of rats on feeding in a novel environment. Behavioural Brain Research, 15, 63–70. Fecteau, J. H., & Munoz, D. P. (2006). Salience, relevance, and firing: a priority map for target selection. Trends in Cognitive Sciences, 10, 382–390. Feldon, J., Shofel, A., & Weiner, I. (1991). Latent inhibition is unaffected by direct dopamine agonists. Pharmacology, Biochemistry and Behavior, 38, 309–314. Feldon, J., & Weiner, I. (1989). Abolition of the acquisition but not the expression of latent inhibition by chlodiazepoxide in rats. Pharmacology, Biochemistry and Behavior, 32, 123–127. Feldon, J., & Weiner, I. (1991). The latent inhibition model of schizophrenic attention disorder: haloperidol and sulpiride enhance rats’ ability to ignore irrelevant stimuli. Biological Psychiatry, 29, 635–646. Floresco, S. B., Ghods-Sharifi, S., Vexelman, C., & Magyar, O. (2006). Dissociable roles for the nucleus accumbens core and shell in regulating set shifting. Journal of Neuroscience, 26, 2449–2457. Gallo, M., & Candido, A. (1995). Dorsal hippocampal lesions impair blocking but not latent inhibition of taste-aversion learning in rats. Behavioral Neuroscience, 109, 413–425. Garcia, J., & Koelling, R. A. (1966). Relation of cue to consequence in avoidance learning. Psychonomic Sciences, 4, 123–124. Garrud, P., Rawlins, J. N. P., Mackintosh, N. J., et al. (1984). Successful overshadowing and blocking in hippocampectomized rats. Behavioural Brain Research, 12, 39–53. Good, M., & Macphail, E. M. (1994). Hippocampal lesions in pigeons (Columba livia) disrupt reinforced preexposure but not overshadowing or blocking. Quarterly Journal of Experimental Psychology: Comparative and Physiological Psychology, 47B, 263–291. Gray, J. A. (1982). The Neuropsychology of Anxiety: An Enquiry into the Functions of the Septo-hippocampal System. Oxford: Oxford University Press. Gray, J. A., Feldon, J., Rawlins, J. N. P., Hemsley, D. R., & Smith, A. D. (1991). The neuropsychology of schizophrenia. Behavioral and Brain Sciences, 14, 1–20. Gray, J. A., Kumari, V., Lawrence, N., & Young, A. M. J. (1999). Functions of the dopaminergic innervation of the nucleus accumbens. Psychobiology, 27, 225–235. Gray, J. A., Moran, P. M., Grigoryan, G., et al. (1997). Latent inhibition: The nucleus accumbens connection revisited. Behavioural Brain Research, 88, 27–35.
364
Helen J. Cassaday and Paula M. Moran
Gray, N. S., Pickering, A. D., Hemsley, D. R., Dawling, S., & Gray, J. A. (1992). Abolition of latent inhibition by a single 5 mg dose of D-amphetamine in man. Psychopharmacology, 107, 425–430. Hall, G. (1991). Perceptual and Associative Learning. Oxford Psychology Series No. 18. Oxford: Clarendon Press, pp. 29–107. Hall, G., & Channell, S. (1985). Differential effects of contextual change on latent inhibition and on the habituation of an orienting response. Journal of Experimental Psychology – Animal Behavior Processes, 11, 470–481. Hall, G., & Pearce, J. M. (1979). Latent inhibition of a CS during CS–US pairings. Journal of Experimental Psychology – Animal Behavior Processes, 5, 31–42. Harmer, C. J., & Phillips, G. D. (1999). Enhanced conditioned inhibition following repeated pre-treatment with d-amphetamine. Psychopharmacology, 142, 120–131. Hemsley, D. R. (1977). What have cognitive deficits to do with schizophrenic symptoms? British Journal of Psychiatry, 130, 167–173. Hemsley, D. R. (1993). A simple (or simplistic) cognitive model for schizophrenia. Behaviour Research and Therapy, 31, 633–645. Hitchcock, J. M., Lister, S., Fischer, T. R., & Wettstein, J. G. (1997). Disruption of latent inhibition in the rat by the 5-HT2 agonist DOI: effects of MDL 100,907, clozapine, risperidone and haloperidol. Behavioural Brain Research, 88, 43–49. Holland, P. C., & Fox, G. D. (2003). Effects of hippocampal lesions in overshadowing and blocking procedures. Behavioral Neuroscience, 3, 650–656. Honey, R. C., & Good, M. (1993). Selective hippocampal lesions abolish the contextual specificity of latent inhibition and conditioning. Behavioral Neuroscience, 107, 23–33. Horsley, R. R., Moran, P. M., & Cassaday, H. J. (2008). Systemic amphetamine and electrolytic lesions of the nucleus accumbens (shell versus core) have dissociable effects on appetitive overshadowing. Journal of Psychopharmacology, 22, 172–181. Ikemoto, I. (2007). Dopamine reward circuitry: two projection systems from the ventral midbrain to the nucleus accumbens-olfactory tubercle complex. Brain Research Reviews, 56, 27–78. Iordanova, M. D., Westbrook, R. F., & Killcross, A. S. (2006). Dopamine activity in the nucleus accumbens modulates blocking in fear conditioning. European Journal of Neuroscience, 24, 3265–3270. James, J. H., & Wagner, A. R. (1980). One-trial overshadowing: evidence of distributive processing. Journal of Experimental Psychology – Animal Behavior Processes, 6, 188–205. Jones, S. H., Gray, J. A., & Hemsley, D. R. (1992). Loss of the Kamin blocking effect in acute but not chronic schizophrenics. Biological Psychiatry, 32, 739–755. Jones, S. H., Hemsley, D., Ball, S., & Serra, A. (1997). Disruption of the Kamin blocking effect in schizophrenia and in normal subjects following amphetamine. Behavioural Brain Research, 88, 103–114. Jongen-Relo, A. L., Kaufmann, S., & Feldon, J. (2002). A differential involvement of the shell and core subterritories of the nucleus accumbens of rats in attentional processes. Neuroscience, 111, 95–109. Joseph, M. H., Peters, S. L., Moran, P. M., et al. (2000). Modulation of latent inhibition in the rat by altered dopamine transmission in the nucleus accumbens at the time of conditioning. Neuroscience, 101, 921–930.
Latent inhibition and other salience modulation effects
365
Kamin, L. J. (1968). Attention-like processes in classical conditioning. In M. R. Jones (Ed.), Miami Symposium on the Prediction of Behavior: Aversive Stimulation. Miami, OH: University of Miami Press. Kapur, S. (2003). Psychosis as a state of aberrant salience: a framework linking biology, phenomenology, and pharmacology in schizophrenia. American Journal of Psychiatry, 160, 13–23. Kapur, S. (2004). How antipsychotics become anti-‘psychotic’ – from dopamine to salience to psychosis. Trends in the Pharmacological Sciences, 25, 402–406. Kaye, H., & Pearce, J. M. (1984). The strength of the orienting response during Pavlovian conditioning. Journal of Experimental Psychology – Animal Behavior Processes, 10, 90–109. Kaye, H., & Pearce, J. M. (1987). Hippocampal lesions attenuate latent inhibition of a CS and of a neutral stimulus. Quarterly Journal of Experimental Psychology: Comparative and Physiological Psychology, 39B, 107–125. Killcross, A. S., Dickinson, A., & Robbins, T. W. (1994). Amphetamine induced disruptions of latent inhibition are reinforcer mediated: implications for animal models of schizophrenic attentional dysfunction. Psychopharmacology, 115, 185–195. Killcross, A. S., Stanhope, K. J., Dourish, C. T., & Piras, G. (1997). WAY100635 and latent inhibition in the rat: selective effects at preexposure. Behavioural Brain Research, 88, 51–57. Konstandi, M., & Kafetzopoulos, E. (1993). Effects of striatal or accumbens lesions on the amphetamine-induced abolition of latent inhibitions. Pharmacology, Biochemistry and Behavior, 44, 751–754. Koob, G. F. (1992). Neuronal mechanisms of drug reinforcement. Annals of the New York Academy of Sciences, 654, 171–191. Kraemer, P. J., & Roberts, W. A. (1984). The influence of flavor preexposure and test interval on conditioned taste aversions in the rat. Learning and Motivation, 15, 259–278. Kraemer, P. J., & Spear, N. E. (1992). The effect of nonreinforced stimulus exposure on the strength of a conditioned taste aversion as a function of retention interval: do latent inhibition and extinction involve a shared process? Animal Learning and Behavior, 20, 1–7. Kumari, V., Cotter, P. A., Mulligan, O. F., et al. (1999). Effects of D-amphetamine and haloperidol on latent inhibition in healthy male volunteers. Journal of Psychopharmacology, 13, 398–405. LaCroix, L., Spinelli, S., Broersen, L. M., & Feldon, J. (2000). Blockade of latent inhibition following pharmacological increase or decrease of GABAA transmission. Pharmacology, Biochemistry and Behavior, 66, 893–901. Le Pelley, M. E. (2004). The role of associative history in models of associative learning: a selective review and a hybrid model. Quarterly Journal of Experimental Psychology: Comparative and Physiological Psychology, 57B, 193–243. Levy, R., & Goldman-Rakic, P. S. (2000). Segregation of working memory functions within the dorsolateral prefrontal cortex. Experimental Brain Research, 133, 23–32. Li, Z. (2002). A saliency map in primary visual cortex. Trends in Cognitive Sciences, 6, 9–16. Lorden, J. F., Rickert, E. J., & Berry, D. W. (1983). Forebrain monoamines and associative learning: I. Latent inhibition and conditioned inhibition. Behavioural Brain Research, 9, 181–199.
366
Helen J. Cassaday and Paula M. Moran
Lorden, J. F., Rickert, E. J., Dawson, R., & Pelleymounter, M. A. (1980). Forebrain norepinephrine and the selective processing of information. Brain Research, 190, 569–573. Lubow, R. E. (1989). Latent Inhibition and Conditioned Attention Theory. Cambridge: Cambridge University Press. Lubow, R. E. (2005). The construct validity of the animal-latent inhibition model of selective attention deficits in schizophrenia. Schizophrenia Bulletin, 31, 139–153. Lubow, R. E., & Moore, A. U. (1959). Latent inhibition: the effect of non-reinforced preexposure to the conditional stimulus. Journal of Comparative and Physiological Psychology, 52, 415–419. Luman M., Oosterlaan, J., & Sergeant, J. A. (2005). The impact of reinforcement contingencies on AD/HD: a review and theoretical appraisal. Clinical Psychology Review, 25, 183–213. Mackintosh, N. J. (1975). A theory of attention: variations in the associability of stimulus with reinforcement. Psychological Review, 82, 276–298. Mackintosh, N. J. (1976). Overshadowing and stimulus intensity. Animal Learning & Behavior, 4, 186–192. Mackintosh, N. J., & Reese, B. (1979). One-trial overshadowing. Quarterly Journal of Experimental Psychology: Comparative and Physiological Psychology, 31B, 519–526. McLaren, I. P. L., & Mackintosh, N. J. (2000). An elemental model of associative learning: I. Latent inhibition and perceptual learning. Animal Learning & Behavior, 28, 211–246. Migo, E. M., Corbett, K., Graham, J., et al. (2006). A novel test of conditioned inhibition correlates with personality measures of schizotypy and reward sensitivity. Behavioural Brain Research, 168, 299–306. Miller, R. R., & Matzel, L. D. (1988). The comparator hypothesis: A response rule for the expression of associations. In G. H. Bower (Ed.) The Psychology of Learning and Motivation, vol. 22. San Diego, CA: Academic Press, pp. 51–92. Moran, P. M., Al-Uzri, M. M., Watson, J., & Reveley, M. A. (2003). Reduced Kamin blocking in non-paranoid schizophrenia: associations with schizotypy. Journal of Psychiatric Research, 37, 155–163. Moran, P. M., Fischer, T., Hitchcock, J. M., & Moser, P. C. (1996). Effects of clozapine on latent inhibition in the rat. Behavioral Pharmacology, 7, 1–7. Moran, P. M., & Moser, P. C. (1992). MDL 73,147EF a 5-HT3 antagonist, facilitates latent inhibition in the rat. Pharmacology, Biochemistry and Behavior, 42, 519–522. Moran, P. M., Owen, L., Crookes, A. E., Al-Uzri, M. M., & Reveley, M. A. (2008). Abnormal prediction error is associated with negative and depressive symptoms in schizophrenia. Progress in Neuropsychopharmacology & Biological Psychiatry, 32, 116–23. Morris, J. S., Friston, K. J., & Dolan, R. J. (1997). Neural responses to salient visual stimuli. Proceedings of the Royal Society London Series B Biological Sciences, 264, 769–775. Moser, P. C., Lister, S., Hitchcock, J., & Moran, P. M. (2000). The pharmacology of latent inhibition in the rat: a review of latent inhibition as an animal model of schizophrenia. Brain Research Reviews, 33, 247–282. Myers, C. E., Gluck, M. A., & Granger, R. (1995). Dissociation of hippocampal and entorhinal function in associative learning – A computational approach. Psychobiology, 23, 116–138.
Latent inhibition and other salience modulation effects
367
Nicholson, D. A., & Freeman, J. H. (2002). Medial dorsal thalamic lesions impair blocking and latent inhibition of the conditioned eyeblink response in rats. Behavioral Neuroscience, 116, 276–285. Norman, C., & Cassaday, H. J. (2004). Disruption of latent inhibition to a contextual stimulus with systemic amphetamine. Neurobiology of Learning and Memory, 82, 61–64. Oades, R. D., Rivet, J. M., Taghzouti, K., et al. (1987). Catecholamines and conditioned blocking: effects of ventral tegmental, septal and frontal 6-hydroxydopamine lesions in rats. Brain Research, 406, 136–146. Oades, R. D., Zimmermann, B., & Eggers, C. (1996). Conditioned blocking in patients with paranoid, non-paranoid psychosis or obsessive compulsive disorder: associations with symptoms, personality and monoamine metabolism. Journal of Psychiatric Research, 30, 369–390. Oberling, P., Bristol, A. S., Matute, H., & Miller, R. R. (2000). Biological significance attenuates overshadowing, relative validity, and degraded contingency effects. Animal Learning & Behavior, 28, 172–186. O’Donnell, P., Lavin, A., Enquist, L. W., Grace, A. A., & Card, J. P. (1994). Interconnected parallel circuits between rat nucleus accumbens and thalamus revealed by retrograde transynapic transport of pseudorabies virus. Journal of Neuroscience, 17, 2143–2167. Ohad, D., Lubow, R. E., Weiner, I., & Feldon, J. (1987). The effects of amphetamine on blocking. Psychobiology, 15, 137–143. O’Tuathaigh, C. M. P., & Moran, P. M. (2002). Evidence for dopamine D1 receptor involvement in the stimulus selection task overshadowing in the rat. Psychopharmacology, 162, 225–231. O’Tuathaigh, C. M. P., & Moran, P. M. (2004). The effect of sulpiride on amphetamine-induced disruption of overshadowing in the rat. Progress in Neuro-psychopharmacology & Biological Psychiatry, 28, 1249–1253. O’Tuathaigh, C. M. P., Salum, C., Young, A. M. J., et al. (2003). The effect of amphetamine disruption on Kamin blocking and overshadowing. Behavioral Pharmacology, 14, 315–322. Pearce, J. M., & Hall, G. (1980). A model for Pavlovian learning: variations in the effectiveness of conditioned but not of unconditioned stimuli. Psychological Review, 87, 532–552. Peng, R. Y., Mansbach, R. S., Braff, D. L., & Geyer, M. A. (1990). A D2 dopamine receptor agonist disrupts sensorimotor gating in rats. Implications for dopaminergic abnormalities in schizophrenia. Neuropsychopharmacology, 3, 211–218. Pothuizen, H. H. J., Jongen-Relo, A. L., Feldon, J., & Yee, B. K. (2006). Latent inhibition of conditioned taste aversion is not disrupted, but can be enhanced, by selective nucleus accumbens shell lesions in rats. Neuroscience, 137, 1119–1130. Purves, D., Bonardi, C., & Hall, G. (1995). Enhancement of latent inhibition in rats with electrolytic lesions of the hippocampus. Behavioral Neuroscience, 109, 366–370. Reilly, S., Harley, C., & Revusky, S. (1993). Ibotenate lesions of the hippocampus enhance latent inhibition in conditioned taste aversion and increase resistance to extinction in conditioned taste preference. Behavioral Neuroscience, 107, 996–1004.
368
Helen J. Cassaday and Paula M. Moran
Rescorla, R. A. (1969). Pavlovian conditioned inhibition. Psychological Bulletin, 72, 77–94. Rescorla, R. A., & Wagner, A. R. (1972). A theory of Pavlovian conditioning: variations in the effectiveness of reinforcenment and nonreinforcement. In A. H. Black & W. F. Prokasy (Eds.), Classical Conditioning II: Current Research and Theory. New York: Appleton-Century-Crofts, pp. 64–99. Restivo, L., Passino, E., Middei, S., & Ammassari-Teule, M. (2002). The strainspecific involvement of nucleus accumbens in latent inhibition might depend on differences in processing configural- and cue-based information between C57BL/6 and DBA mice. Brain Research Bulletin, 57, 35–39. Rickert, E. J., Bennett, T. L., Lane, P., & French, J. (1978). Hippocampectomy and the attenuation of blocking. Behavioral Biology, 22, 147–160. Rickert, E. J., Lorden, J. F., Dawson, R., & Smyly, E. (1981). Limbic lesions and the blocking effect. Physiology & Behavior, 26, 601–606. Robbins, T. W., & Koob, G. F. (1980). Selective disruption of displacement behaviour by lesions of the mesolimbic dopamine system. Nature, 285, 409–412. Rose, J. E., & Woolsey, C. N. (1948). The orbitofrontal cortex and its connections with the mediodorsal nucleus in rabbit, sheep and cat. Research Publications – Association for Research in Nervous and Mental Disease, 27, 210–232. Rossi, A. F., Bichot, N. P., Desimone, R., & Ungerleider, L. G. (2007). Top-down attentional deficits in macaques with lesions of lateral prefrontal cortex. Journal of Neuroscience, 27, 11306–11314. Rozin, P., & Kalat, J. W. (1971). Specific hungers and poisoning as adaptive specializations of learning. Psychological Review, 78, 459–486. Rudy, J. W., & O’Reilly, R. C. (1999). Contextual fear conditioning, conjunctive representations, pattern completion, and the hippocampus. Behavioral Neuroscience, 113, 867–880. Ruob, C., Elsner, J., Weiner, I., & Feldon, J. (1997). Amphetamine-induced disruption and haloperidol-induced potentiation of latent inhibition depend on the nature of the stimulus. Behavioural Brain Research, 88, 35–41. Russig, H., Kovacevic, A., Murphy, C. A., & Feldon, J. (2003). Haloperidol and clozapine antagonise amphetamine-induced disruption of latent inhibition of conditioned taste aversion. Psychopharmacology, 170, 263–270. Russig, H., Murphy, C. A., & Feldon, J. (2002). Clozapine and haloperidol reinstate latent inhibition following its disruption during amphetamine withdrawal. Neuropsychopharmacology, 26, 765–777. Schiller, D., Zuckerman, L., & Weiner, I. (2006). Abnormally persistent latent inhibition induced by lesions to the nucleus accumbens core, basolateral amygdala and orbitofrontal cortex is reversed by clozapine but not by haloperidol. Journal of Psychiatric Research, 40, 167–177. Scholten, M. R. M., van Honk, J., Aleman, A., & Kahn, R. S. (2006). Behavioural inhibition system (BIS), behavioural activation system (BAS) and schizophrenia: relationship with psychopathology and physiology. Journal of Psychiatric Research, 40, 638–645. Schmajuk, N. A., Gray, J. A., & Larrauri, J. A. (2005). A pre-clinical study showing how dopaminergic drugs administered during pre-exposure can impair or facilitate latent inhibition. Psychopharmacology, 177, 272–279.
Latent inhibition and other salience modulation effects
369
Schmajuk, N., Lam, Y. W., & Christiansen, B. A. (1994). Latent inhibition of the rat eyeblink response: effect of hippocampal aspiration lesions. Physiology & Behavior, 55, 597–601. Schmajuk, N. A., Spear, N. E., & Isaacson, R. C. (1983). Absence of overshadowing in rats with hippocampal lesions. Physiological Psychology, 11, 59–62. Schultz, W. (1997). Dopamine neurons and their role in reward mechanisms. Current Opinion in Neurobiology, 7, 191–197. Schultz, W. (2006). Behavioural theories and the neurophysiology of reward. Annual Review of Psychology, 57, 87–115. Serra, A. M., Jones, S. H., Toone, B., & Gray, J. A. (2001). Impaired associative learning in chronic schizophrenics and their first-degree relatives: a study of latent inhibition and the Kamin blocking effect. Schizophrenia Research, 48, 273–289. Shadach, E., Gaisler, I., Schiller, D., & Weiner, I. (2000). The latent inhibition model dissociates between clozapine, haloperidol and ritanserin. Neuropsychopharmacology, 23, 151–161. Shapiro, K. L., Jacobs, W. J., & LoLordo, V. M. (1980). Stimulus-reinforcer interactions in Pavlovian conditioning of pigeons: implications for selective associations. Animal Learning & Behavior, 8, 586–594. Solomon, P. R. (1977). Role of hippocampus in blocking and conditioned inhibition of rabbits’ nictitating-membrane response. Journal of Comparative and Physiological Psychology, 91, 407–417. Solomon, P. R., Crider, A., Winkelman, J. W., et al. (1981). Disrupted latent inhibition in the rat with chronic amphetamine or haloperidol-induced supersensitivity: relationship to schizophrenic attention disorder. Biological Psychiatry, 16, 519–537. Solomon, P. R., & Staton, D. M. (1982). Differential effects of micro-injections of d-amphetamine into the nucleus accumbens or the caudate putamen on the rat’s ability to ignore an irrelevant stimulus. Biological Psychiatry, 17, 743–756. Stuss, D. T. (2006). Frontal lobes and attention: processes and networks, fractionation and integration. Journal of the International Neuropsychological Society, 12, 261–271. Swan, J. A., & Pearce, J. M. (1988). The orienting response as an index of stimulus associability in rats. Journal of Experimental Psychology – Animal Behavior Processes, 14, 292–301. Tai, C.-T., Cassaday, H. J., Feldon, J., & Rawlins, J. N. P. (1995). Both electrolytic and excitotoxic lesions of nucleus accumbens disrupt latent inhibition of learning in rats. Neurobiology of Learning and Memory, 64, 36–48. Trimble, K. M., Bell, R., & King, D. J. (2002). Effects of the selective dopamine D1 antagonists NNC 01–0112 and SCH 39166 on latent inhibition in the rat. Physiology & Behavior, 77, 115–123. Tsaltas, E., Preston, G. C., Rawlins, J. N. P., Winocur, G., & Gray, J. A. (1984). Dorsal bundle lesions do not affect latent inhibition of conditioned suppression. Psychopharmacology, 84, 549–555. Tsotsos, J. K. (1990). Analyzing vision at the complexity level. Behavioral and Brain Sciences, 13, 423–445. Van den Bos, R., & Cools, A. R. (2003). Switching to cue-directed behavior: specific for ventral striatal dopamine but not ventral pallidum/substantia innominata GABA as revealed by a swimming test procedure in rats. Neuroscience, 118, 141–1149.
370
Helen J. Cassaday and Paula M. Moran
Van Meel, C. S., Oosterlaan, J., Heslenfeld, D. J., & Sergeant, J. A. (2005). Telling good from bad news: ADHD differentially affects processing of positive and negative feedback during guessing. Neuropsychologia, 43, 1946–1954. Volkow, N. D., Fowler, J. S., Wang, G. -J., & Goldstein, R. Z. (2002). Role of dopamine, the frontal cortex and memory circuits in drug addiction: insight from imaging studies. Neurobiology of Learning and Memory, 78, 610–624. Waelti, P., Dickinson, A., & Schultz, W. (2001). Dopamine responses comply with basic assumptions of formal learning theory. Nature, 412, 43–48. Wagner, A. R. (1981). SOP: A model of automatic memory processing in animal behaviour. In N. E. Spear & R. R. Miller (Eds.), Information Processing in Animals: Memory Mechanisms. Hillsdale, NJ: Lawrence Erlbaum, pp. 5–47. Weiner, I. (1990). Neural substrates of latent inhibition: the switching model. Psychological Bulletin, 108, 442–461. Weiner, I. (2003). The “two-headed” latent inhibition model of schizophrenia: modeling positive and negative symptoms and their treatment. Psychopharmacology, 169, 257–297. Weiner, I., Bernasconi, E., Broersen, L. M., & Feldon, J. (1997). Amphetamineinduced disruption of latent inhibition depends on the nature of the stimulus. Behavioral Pharmacology, 8, 442–457. Weiner, I., & Feldon, J. (1987). Facilitation of latent inhibition by haloperidol in rats. Psychopharmacology, 91, 248–253. Weiner, I., & Feldon, J. (1997). The switching model of latent inhibition: an update of neural substrates. Behavioural Brain Research, 88, 11–25. Weiner, I., Feldon, J., & Katz, Y. (1987). Facilitation of the expression but not the acquisition of latent inhibition by haloperidol in rats. Pharmacology, Biochemistry and Behavior, 26, 241–246. Weiner, I., Gal, G., & Feldon, J. (1999). Disrupted and undisruptable latent inhibition following shell and core lesions. Annals of the New York Academy of Sciences, 877, 723–727. Weiner, I., Gal, G., Rawlins, J. N. P., & Feldon, J. (1996). Differential involvement of the shell and core subterritories of the nucleus in latent inhibition and amphetamine-induced activity. Behavioural Brain Research, 81, 123–133. Weiner, I., Lubow, R. E., & Feldon, J. (1981). Chronic amphetamine and latent inhibition. Behavioural Brain Research, 2, 285–286. Weiner, I., Lubow, R. E., & Feldon, J. (1984). Abolition of the expression but not the acquisition of latent inhibition by chronic amphetamine in rats. Psychopharmacology, 83, 194–199. Weiner, I., Lubow, R. E., & Feldon, J. (1988). Disruption of latent inhibition by acute administration of low doses of amphetamine. Pharmacology, Biochemistry and Behavior, 30, 871–878. Westbrook, R. F., Good, A. J., & Kiernan, M. J. (1997). Microinjection of morphine into the nucleus accumbens impairs contextual learning in rats. Behavioral Neuroscience, 111, 996–1013. Williams, J. H., Wellman, N. A., Geaney, D. P., et al. (1996). Antipsychotic drug effects in a model of schizophrenic attentional disorder: a randomized controlled trial of the effects of haloperidol on latent inhibition in healthy people. Biological Psychiatry, 40, 1135–1143.
Latent inhibition and other salience modulation effects
371
Williams, J. H., Wellman, N. A., Geaney, D. P., et al. (1997). Haloperidol enhances latent inhibition in visual tasks in healthy people. Psychopharmacology, 133, 262–268. Winocur, G., Rawlins, J. N. P., & Gray, J. A. (1987). The hippocampus and conditioning to contextual cues. Behavioral Neuroscience, 101, 617–625. Wise, R. A., & Rompre, P. -P. (1989). Brain dopamine and reward. Annual Review of Psychology, 40, 191–225. Wise, R. A. (2004). Dopamine, learning and motivation. Nature Reviews Neuroscience, 5, 483–494. Young, A. M. J., Ahier, R. G., Upton, R. L., Joseph, M. H., & Gray, J. A. (1998). Increased extracellular dopamine in the nucleus accumbens of the rat during associative learning of neutral stimuli. Neuroscience, 83, 1175–1183. Young, A. M. J., Moran, P. M., & Joseph, M. H. (2005). The role of dopamine in conditioning and latent inhibition: what, where, when and how? Neuroscience and Biobehavioural Reviews, 29, 963–976.
16 What the brain teaches us about latent inhibition (LI): the neural substrates of the expression and prevention of LI Ina Weiner
Scores of experiments in the field of animal learning have demonstrated that conditioning to a stimulus depends not merely on its current relationship with a reinforcer, but is affected by animal’s past experience with that stimulus. Latent inhibition (LI) is one case of such a biasing effect of past experience: it reflects the proactive interference of nonreinforced stimulus preexposure on the subsequent performance of a learning task involving this stimulus (Lubow, 1973, 1989; Lubow, Weiner, & Schnur, 1981). LI can be demonstrated in a variety of classical and instrumental conditioning procedures, and in many mammalian species, including humans (Lubow, 1973, 1989; Lubow et al., 1981). While a variety of behavioral tasks are used to demonstrate LI in rodents, all of them share a basic procedure. In the first stage, preexposure, animals from each of two groups are placed in an environment that will later serve as the conditioning–test apparatus. Subjects in the “stimulus preexposed” (PE) group are repeatedly exposed to a stimulus (e.g., tone) which is not followed by a significant consequence. Subjects in the “nonpreexposed” (NPE) group spend an equivalent amount of time in the apparatus without receiving the stimulus. When the preexposure stage is completed, either immediately, or a certain time later, all of the subjects enter the conditioning stage, in which the preexposed stimulus is paired with a reinforcer over a number of trials. Performance is assessed by examining some behavioral index of conditioned responding, either during the conditioning stage or in a third, test stage. LI is manifested in poorer performance of the PE compared to the NPE group (see Weiner, 2001). The situation is very different in LI research in humans, which is characterized by a multiplicity of procedures often unrelated conceptually and theoretically. However, the fact that LI is sensitive to the same environmental (e.g., change of context) and pharmacological (e.g., amphetamine; see below) manipulations in humans and animals supports the view that it can be viewed as reflecting the operation of analogous processes across species (Lubow, 2005; Lubow & Gewirtz, 1995). Latent Inhibition: Cognition, Neuroscience and Applications to Schizophrenia, ed. R.E. Lubow and Ina Weiner. Published by Cambridge University Press # Cambridge University Press 2010.
372
What the brain teaches us about latent inhibition (LI)
373
Based on findings that nonreinforced stimulus preexposure decreases the attention to, or the associability of, that stimulus, without affecting its associative strength, the construct of attention has been adopted by most major theories which have attempted to explain the LI effect (e.g. Frey & Sears, 1978; Lubow, 1989; Lubow et al., 1981; Mackintosh, 1975; Moore & Stickney, 1980; Pearce & Hall, 1980; Schmajuk & Moore, 1985, 1988; Wagner & Rescorla, 1972; Footnote 1). In short, the retarded conditioning to the preexposed stimulus is commonly seen as resulting from decreased attention or associability of that stimulus, and as such is considered to index the capacity not to attend to, or to ignore, stimuli that predict no significant consequences.
LI as a model of disorder of attention in schizophrenia While LI has long been of concern to learning theorists, the interest in its neural substrates has been spawned by the suggestion that its disruption may provide an animal model of schizophrenia. It was the construct of attention, described above, that provided the launching platform for the model. Of particular importance was the conceptual linkage between LI and clinical observations that schizophrenia patients are unable to filter out or ignore irrelevant stimuli. Beginning with Kraepelin’s (1919) observation that a “disorder of attention” is “conspicuously developed” in patients with dementia praecox, and Bleuler’s (1911) analogous description of schizophrenia as the loss of “selectivity which normal attention ordinarily exercises among the sensory impressions”, attentional deficit in schizophrenia has retained its centrality in numerous theoretical formulations (e.g. Anscombe, 1987; Cornblatt, Lenzenweger, Dworkin, & Erlenmeyer-Kimling, 1985; Frith, 1979; Gjerde, 1983; Hemsley, 1993, 1994; Maher, Manschreck, & Molino, 1983; McGhie & Chapman, 1961; Nuechterlein & Dawson, 1984; Oades, 1982; Venables, 1984). Most often, this deficit has been described as an inability to filter out, or ignore, irrelevant or unimportant stimuli, and it has been argued that the major abnormalities of schizophrenia can be derived from this single underlying deficit. Studies of high-risk individuals (e.g. children of schizophrenic parents) indicate that attentional deficits may constitute a biological marker for the liability to schizophrenia (Cornblatt & Erlenmeyer-Kimling, 1984; Cornblatt, Winters, & Erlenmeyer-Kimling, 1989), and there is evidence that the amelioration of schizophrenic symptoms with neuroleptic treatment is related to the normalization of attentional processes (Asarnow, Marder, Mintz et al., 1988; Braff & Saccuzzo, 1982; Kornetsky, 1972; Rappaport, Silverman, Hopkins, & Hall, 1971). Based on the above, Solomon et al. (1981) and Weiner et al. (1981, 1984, 1988) proposed that disrupted LI may provide an animal model of the widely described failure of schizophrenic patients to ignore irrelevant stimuli. Since the reigning hypothesis regarding the pathophysiology of schizophrenia at that time stated that
374
Ina Weiner
excessive dopamine (DA) neurotransmission in the forebrain contributes to schizophrenia (Meltzer & Stahl, 1976; Snyder, 1976), Solomon et al. and Weiner et al. naturally focused on the dopaminergic system. Specifically, these authors showed that rats treated with the DA releaser amphetamine, which produces and exacerbates (psychotic) schizophrenic symptoms (Angrist, Rotrosen, & Gershon, 1980; Angrist, Sathananthan, Wilk, & Gershon, 1974; Angrist, Shopsin, & Gershon, 1971; Hall, Popkin, Beresford, & Hall, 1988; Janowsky & Davis, 1976; Lieberman, Kane, & Alvir, 1987; Lieberman, Kane, Gadaleta et al., 1984; Snyder, 1973; van Kammen, Bunney, Docherty et al., 1982), failed to show LI. These same laboratories later showed that an opposite effect, namely, LI potentiation, is obtained following blockade of DA transmission by the D2 receptor antagonist and antipsychotic drug (APD) haloperidol (Christison, Atwater, Dunn, & Kilts, 1988; Weiner & Feldon, 1987; Weiner, Feldon, & Katz, 1987). Although APDs can augment LI under conditions of weak LI in controls, typically LI enhancement is demonstrated under conditions that are explicitly manipulated not to produce LI in control animals, namely, low numbers of preexposures or high numbers of conditioning trials. The original demonstration of amphetamine-induced LI disruption has been often replicated (Bakshi, Geyer, Taaid, & Swerdlow, 1995; Gosselin, Oberling, & Di Scala, 1996; Joseph, Peters, Moran et al., 2000; Killcross, Dickinson, & Robbins, 1994a; Killcross & Robbins, 1993; Moran, Fischer, Hitchcock, & Moser, 1996; Ruob, Elsner, Weiner, & Feldon, 1997; Weiner, Bernasconi, Broersen, & Feldon, 1997; Weiner, Shadach, Tarrasch et al., 1996b; Weiner, Tarrasch, Bernasconi et al., 1997). Likewise, LI potentiation has been demonstrated with a wide range of antipsychotic drugs (Dunn, Atwater, & Kilts, 1993; Feldon & Weiner, 1991; Killcross, Dickinson, & Robbins, 1994b; Peters & Joseph, 1993; Shadach, Feldon, & Weiner, 1999; Trimble, Bell, & King, 1997, 1998, 2002). Finally, consistent with clinical pharmacology, APDs, which are effective in the treatment of amphetamine-induced psychosis and schizophrenia (Arnt & Skarsfeldt, 1998; Kinon & Lieberman, 1996), reverse amphetamine-induced LI disruption (Gosselin et al., 1996; Solomon et al., 1981; Warburton, Joseph, Feldon et al., 1994; Weiner et al., 1996b). These findings established an animal model that combined the most prominent neurochemical dysfunction implied in schizophrenia, i.e., dopaminergic, and a widely described cognitive dysfunction of this disorder, i.e., an inability to ignore irrelevant stimuli. Two further empirical observations have contributed to the establishment and the validity of the LI model. First, the extension of the finding from the rat laboratory to humans has shown that LI is disrupted in acutely psychotic schizophrenic patients tested within the first weeks of the current episode of illness or being in an acute phase of an otherwise chronic disorder (Baruch, Hemsley, & Gray, 1988a; Gray, Hemsley, & Gray, 1992a; Gray, Pilowsky, Gray, & Kerwin, 1995b; Lubow, Kaplan, Abramovich et al., 2000; Rascle, Mazas, Vaiva et al., 2001; but see also Swerdlow, Braff, Hartston et al., 1996; Williams, Wellman, Geaney et al., 1998). The initial study (Baruch et al., 1988a), using repeated testing in the same patients, reported that
What the brain teaches us about latent inhibition (LI)
375
LI was absent in the first 2 weeks of a schizophrenic episode and was restored to more or less normal levels after 7–8 weeks of APD treatment. Second, it has been shown that amphetamine-treated normal humans, like amphetamine-treated rats, fail to ignore the preexposed stimulus (Gray, Pickering, Hemsley et al., 1992b; Salgado, Hetem, Vidal et al., 2000; Swerdlow, Stephany, Wasserman et al., 2003; Thornton, Dawe, Lee et al., 1996). In addition, normal humans scoring high on questionnaires measuring schizotypy display reduced LI relatively to subjects with low schizotypy scores (Baruch, Hemsley, & Gray, 1988b; Braunstein-Bercovitz & Lubow, 1998; De laCasa& Lubow,1994;De laCasa,Ruiz,& Lubow,1993;Della Casa,Hofer, Weiner,& Feldon, 1999; Lipp & Vaitl, 1992; Lubow & De la Casa, 2002; Lubow, Ingbergsachs, Zalsteinorda, & Gewirtz, 1992; Lubow, Kaplan, & De la Casa, 2001; Wuthrich & Bates, 2001). Finally, normal humans treated with APDs, like the haloperidol-treated rats, exhibit enhanced LI (McCartan, Bell, Green et al., 2001; Williams, Wellman, Geaney et al., 1996, 1997; but see also Williams et al., 1998). These results have strengthened the likelihood that the LI effect observed in humans and rats is indeed functionally and pharmacologically the same phenomenon, supporting the status of disrupted LI as a model of positive symptoms of schizophrenia with face, construct and predictive validity (Ellenbroek & Cools, 1990; Gray, Feldon, Rawlins et al., 1991; Moser, Hitchcock, Lister, & Moran, 2000; Weiner & Feldon, 1997; Weiner, Gaisler, Schiller et al., 2000).
Attention is not everything: the switching model of LI After demonstrating that amphetamine disrupts LI and haloperidol potentiates LI, we sought to determine the locus of these effects, namely in which stage of the LI procedure drug action was exerted. Specifically, we wondered whether amphetamine and haloperidol exerted their actions on LI in the preexposure stage, as would be expected if these drugs affected animals’ capacity to learn to ignore the irrelevant stimulus. Our answer was clear: neither of these drugs worked in preexposure. Thus, rats preexposed under amphetamine (given acutely or following 14 daily injections) but conditioned without the drug, showed intact LI (Weiner et al., 1981, 1984, 1988). Similarly, rats preexposed under haloperidol, but conditioned without it, showed a normal, non-potentiated LI effect (Weiner & Feldon, 1987; Weiner et al., 1987). These results indicated that neither drug affects animals’ capacity to learn to ignore a nonreinforced stimulus in preexposure but instead affects the subsequent behavioral control by this stimulus. Thus, haloperidol enhances and amphetamine impairs the animal’s ability to continue to respond to such a stimulus as irrelevant in conditioning, when it is followed by reinforcement. This in turn implied that explanations of LI cannot be limited to processes occurring in the preexposure stage, but must take into account processes occurring during the conditioning stage. Consequently, contrary to the then prevalent view that the retarded conditioning to the preexposed stimulus is a passive result of its reduced associability or salience that occurred in
376
Ina Weiner
preexposure, we adopted a view of LI that distinguished between the acquisition of LI (learning to ignore the nonreinforced stimulus in preexposure) and the expression of LI (subsequent expression of this learning in conditioning) (Weiner et al., 1987, 1984, 1988). This view of LI has been elaborated in the switching model of LI (Weiner, 1990, 2003; Weiner & Feldon, 1997), and has guided our use of LI for modeling schizophrenia. The switching model retained the assumption that acquisition of irrelevance occurs in preexposure, but it distinguished between the acquisition and the expression of stimulus irrelevance: while in preexposure the organism learns that a stimulus is irrelevant, it is in the conditioning stage that the “irrelevance” of the stimulus is manifested by the organism continuing to treat it as irrelevant in spite of the fact that it signals a significant outcome. In this manner, the model shifted the locus of “ability to ignore” from preexposure, where it had been exclusively located, to conditioning. The same applies to the “inability to ignore”: the deficit may lie not in the ability to ignore the inconsequential stimulus in preexposure but in the failure to continue to ignore it under changed reinforcement contingencies in conditioning. According to the switching model, the LI paradigm involves the acquisition of two independent and conflicting associations in preexposure (stimulus–no-event) and conditioning (stimulus–reinforcement), which compete for expression during conditioning, when a mismatch arises between conflicting predictions signaled by the CS (see Figure 16.1). In other words, LI is a selection problem. In order to show LI, the organism must remain under the control of the information acquired in preexposure (stimulus–no-event); in contrast, absence of LI indicates that the organism switches to respond according to the new stimulus–reinforcement association. Which one of the two associations gains behavioral control depends on the specific balance between the behavioral impact of preexposure and conditioning. The switching model emphasized that LI is a “window” phenomenon: it is expressed only with certain combinations of preexposure and conditioning parameters, and changes in these parameters, such as reduction in the number of preexposures, an increase in the number of conditioning trials, or a context change between preexposure and conditioning, cause the organism to switch responding according to the new stimulus–reinforcement contingency and thus not to express LI. Likewise, physiological manipulations can promote or retard switching to respond according to stimulus–reinforcement association, and thus produce two poles of abnormality in LI: the former will prevent the expression of LI under conditions in which normal rats exhibit LI, and the latter will promote the expression of LI under conditions in which normal rats fail to show LI. Because low doses of amphetamine increase, while APDs and other means of DA blockade reduce, animals’ capacity to switch ongoing behavior in response to changed environmental contingencies (Cools, Coolen, Smit, & Ellenbroek, 1984; Gelissen & Cools, 1988; Oades, 1985; Robbins & Everitt, 1982), the switching model attributed LI disruption and potentiation by amphetamine and haloperidol,
What the brain teaches us about latent inhibition (LI)
377
Figure 16.1. Latent inhibition as a response competition and selection phenomenon. In the preexposure stage, stimulus-preexposed (PE) animals acquire a stimulus–no-event association, which results in a conditioned response of inattention to the preexposed stimulus. Following conditioned attention theory (Lubow et al., 1981), inattention is treated as a classically conditioned response, acquired when stimuli are consistently followed by the lack of a consequence, and governed by the same rules that govern association formation during classical conditioning. In the conditioning stage, the stimulus signals conflicting outcomes, no-event vs. reinforcement, that compete for behavioral expression (conditioned inattention response vs. the conditioned response acquired in conditioning). Which of the two associations gains behavioral control depends on factors that determine their relative behavioral impact during conditioning. The three most conspicuous factors are strength of preexposure (usually manipulated by changing number of stimulus preexposures but can involve any manipulation known to affect classical conditioning such as stimulus intensity, ISI, etc.), strength of conditioning (usually manipulated by changing the number of conditioning trials or intensity of reinforcement), and context (manipulated by changing the context between preexposure and conditioning), but there are other factors as well, such as the time interval between preexposure and conditioning or the motivational state of the animal in the two stages. Pharmacological LI experiments typically manipulate number of preexposures and/or conditioning trials.
respectively, to this action of these drugs. Within the framework of the switching model, the effects of amphetamine and haloperidol were interpreted to imply that the two drugs do not affect the acquisition of the stimulus–no-event association, but modulate its expression in conditioning, so that amphetamine promotes a rapid switch of responding according to the stimulus–reinforcement association, whereas haloperidol retards such switching. As already noted, consistent with the position of the switching model, DA manipulations affect LI via conditioning (Gray, Moran, Grigoryan et al., 1997; Joseph et al., 2000; Peters & Joseph, 1993; Shadach et al., 1999; Weiner et al., 1984, 1988; Weiner, Shadach, Barkai, & Feldon, 1997). In the following sections, I will summarize most of what we know from the rodent data on the neural substrates of LI and integrate these data with the switching model of LI
378
Ina Weiner
(Weiner, 1990; Weiner & Feldon, 1997). Pharmacology of LI is summarized in another chapter. The bulk of the investigations of the neural substrates of LI has naturally focused on the dopaminergic system, as well as on brain regions presumed to play a role in the pathophysiology of schizophrenia, namely, the limbic system (the hippocampus, the entorhinal cortex, the amygdala), the frontal cortex, and the nucleus accumbens, the target of the mesolimbic dopaminergic system which receives afferents from all of the above regions. Consistent with the notion of a competition for behavioral expression between stimulus–no-event and stimulus– reinforcement associations, lesions to these different brain regions do not merely disrupt LI; rather, they produce the same two LI abnormalities that are produced by amphetamine and haloperidol, namely, prevention of LI under conditions that lead to LI in controls, and expression of LI under conditions that prevent LI expression in controls. In other words, they either promote the behavioral expression of the stimulus–no-event association or of the stimulus–reinforcement association. I will argue that the competition between these associations takes place at the nucleus accumbens and that it is mediated via distinct brain circuitries associated with the two NAC subregions, the shell and the core.
The nucleus accumbens: functional dissociation between the shell and core subregions Because LI was disrupted by low doses of amphetamine that act via the mesolimbic DA system but not high doses that act via the mesostriatal DA system, we suggested that the nucleus accumbens (NAC) is the brain region at which amphetamine acts to disrupt LI. Moreover, given that NAC plays a key role in behavioral and cognitive switching, with enhancement of dopaminergic activity in the NAC promoting switching, whereas dopaminergic blockade in this structure gives rise to perseveration (Koob, Riley, Smith, & Robbins, 1978; Le Moal & Simon, 1991; Oades, 1985; Pennartz, Groenewegen, & Lopes da Silva, 1994; Robbins & Everitt, 1982; Stahl, 1988; Swerdlow & Koob, 1987; Taghzouti, Louilot, Herman et al., 1985; Taghzouti, Simon, Louilot et al., 1985; van den Bos, Charria Ortiz, Bergmans, & Cools, 1991; van den Bos & Cools, 1989), the switching model attributed to the NAC and its dopaminergic innervation a key role in the expression of LI, i.e., in determining whether in conditioning, the animal responds according to the stimulus–no-event or to the stimulus–reinforcement contingency. Briefly, the original 1990 model posited that the NAC is not involved in the preexposure stage but is activated in the conditioning stage, when the previously nonreinforced stimulus is followed by reinforcement. Since such activation leads to rapid behavioral and cognitive switching, it promotes a switch of responding according to the stimulus–reinforcement contingency. In the intact brain, the stimulus–no-event association acquired in preexposure continues to control behavior in conditioning because the switching mechanism of the NAC is inhibited. In the original formulation
What the brain teaches us about latent inhibition (LI)
379
of the switching model (Weiner, 1990), this function was attributed to the hippocampus, but subsequent studies led to a refinement of this position (see Weiner, 2000, 2003; Weiner & Feldon, 1997). The central prediction of the model was that blockade of NAC DA activity or NAC lesion should prevent LI abolition by manipulations that normally do so, and, on their own, have no effect on LI or produce persistent LI because they block the ability to switch. Conversely, LI disruption would be associated with enhanced activation of the NAC. Studies investigating NAC involvement in LI have been supportive of this proposition.
Intra-accumbens injections Solomon and Staton (1982) showed that LI was disrupted by intra-accumbal but not by intra-caudate amphetamine infusion; others have reported that LI was left intact following intra-accumbens injection of amphetamine (Ellenbroek, Knobbout, & Cools, 1997; Killcross & Robbins, 1993), indicating that the effects of direct injection of amphetamine into NAC do not always mimic those of systemic low-dose injection. Nevertheless, a series of experiments by J.A. Gray and Joseph provided convincing evidence that the effects of amphetamine and haloperidol are mediated via NAC, and, moreover, that these effects are exerted at the time of conditioning (Gray et al., 1997; Joseph et al., 2000). First, the disruption of LI caused by systemic amphetamine administration was blocked by intra-accumbens injection of haloperidol, and this effect was obtained when the intra-accumbens haloperidol injection was confined to the time of conditioning. Second, although a single intra-accumbens injection of amphetamine prior to conditioning spared LI, such an injection was able to disrupt LI if it were preceded by a single systemic injection of amphetamine 24 h earlier. This finding showed that amphetamine disrupts LI by virtue of an action in the NAC at the time of conditioning, but also indicated that this action may require a sensitized impulse-dependent DA release in the NAC, as suggested by Warburton et al. (1996) and Weiner et al. (1988). As noted by J.A. Gray et al. (1995a; 1997), lack of LI disruption by intra-accumbal amphetamine infusion in previous studies may have stemmed from the fact that such administration does not produce sensitization. Finally, the induction of persistent LI by haloperidol is mediated via the NAC: following weak preexposure, NAC vehicle-injected rats did not show LI, whereas an intra-accumbens injection of haloperidol led to the emergence of LI. Importantly, the LI-potentiating effect was obtained with haloperidol injection confined to the time of conditioning, as it also was following destruction of dopaminergic terminals in the NAC by means of 6-OHDA lesions. The above results indicate that NAC DA is involved in normal LI as well as in its disruption and potentiation. Moreover, the critical alterations of DA transmission occur at the time of conditioning. Finally, the fact that depletion or blockade of DA in this structure produces persistent LI and prevents the disruptive effects of systemic amphetamine administration is in line with the prediction of the switching model that
380
Ina Weiner
animals with blocked or depleted NAC DA are incapable of switching to respond according to the stimulus–reinforcement contingency in conditioning. Microdialysis and lesion studies have further supported and refined the role of NAC in LI. In particular, consistent with the massive evidence that the NAC contains two subregions, the core and the shell, that are cytoarchitecturally, physiologically, pharmacologically, and functionally distinct (e.g. Deutch & Cameron, 1992; Groenewegen, Berendse, Meredith et al., 1991; Groenewegen, Vermeulen-Van der Zee, te Kortschot, & Witter, 1987; Groenewegen, Wright, & Beijer, 1996; Groenewegen, Wright, Beijer, & Voorn, 1999; Heimer, Alheid, de Olmos et al., 1997; Maldonado-Irizarry & Kelley, 1994, 1995; Pennartz et al., 1994; Zahm, 1999, 2000; Zahm & Brog, 1992), these studies revealed a shell–core dissociation in LI, albeit one of a unique nature.
Microdialysis: functional shell–core dissociation Measurement of extracellular DA in NAC with microdialysis showed that whereas a novel stimulus paired with shock subsequently potentiated DA release, such potentiation was not seen if the stimulus received nonreinforced preexposure prior to conditioning (Young, Joseph, & Gray, 1993). Consistent with the switching model, the difference between preexposed and nonpreexposed stimuli in their capacity to affect extracellular NAC DA was seen in the conditioning stage whereas no changes in NAC DA release were seen during preexposure. The difference continued to be present in the third, retention test stage, in which the stimulus was presented on its own. Subsequent microdialysis studies investigating DA release in each of the two NAC subregions, the core and the shell, during a retention test stage, confirmed Young et al.’s results and revealed a functional differentiation between the two NAC subregions. Murphy et al. (2000) reported that extracellular DA levels were increased in the shell subregion upon conditioned stimulus presentation to the NPE rats, but such an increase was not seen in the PE rats. These authors found no differences between the PE and NPE groups in the core subregion. The differential involvement of the NAC subregions in LI has been confirmed by Jeanblanc et al. (2002) using voltammetry. As in Murphy et al.’s (2000) study, increase in DA in response to the conditioned stimulus was seen in the shell of the NPE rats, but was prevented in the PE rats (this difference was confined to the dorsomedial shell). In addition, these authors also found a decrease in core DA levels in the NPE rats, which was prevented in the PE rats. Thus, in this study, conditioning altered DA levels in both the shell and the core, but in opposite directions, and both of these effects were prevented by nonreinforced preexposure to the conditioned stimulus. The finding that preexposure results in decreased DA levels in the shell subregion suggests that the shell encodes the motivational relevance (or in the case of LI, motivational irrelevance) of the preexposed stimulus, as has been repeatedly suggested
What the brain teaches us about latent inhibition (LI)
381
(Bassareo & Di Chiara, 1999; Deutch, 1993; Di Chiara, 1995; Hutcheson, Parkinson, Robbins, & Everitt, 2001; Ito, Dalley, Howes et al., 2000; Kelley, Smith-Roe, & Holahan, 1997; Marcus, Nomikos, & Svensson, 2000; Parkinson, Cardinal, & Everitt, 2000; Parkinson, Olmstead, Burns et al., 1999; Smith-Roe & Kelley, 2000; Swanson, Heath, Stratford, & Kelley, 1997). As for the inconsistent results obtained with regard to core DA levels in the two studies, this could reflect the different magnitudes of the behavioral response in these studies. Core DA levels were suggested to index the response to the conditioned stimulus (Bassareo & Di Chiara, 1999; Deutch, 1993; Di Chiara, 1995; Hutcheson et al., 2001; Ito et al., 2000; Kelley et al., 1997; Marcus et al., 2000; Parkinson et al., 2000, 1999; Smith-Roe & Kelley, 2000; Swanson et al., 1997). Since there was a robust difference between the conditioned response magnitude of the preexposed and nonpreexposed rats in Jeanblanc et al.’s but not in Murphy et al.’s study, this could explain the significant difference in core DA levels between PE and NPE rats in the former but not the latter study. In a more recent series of studies, Louilot and colleagues studied DA release in the NAC subregions following LI disruption by entorhinal cortex or subiculum inactivation during pre-exposure (Jeanblanc, Peterschmitt, Hoeltzel, & Louilot, 2004; Peterschmitt, Hoeltzel, & Louilot, 2005; Peterschmitt, Meyer, & Louilot, 2008). These authors showed that disrupted LI was associated with a different pattern of DA release in the two subregions but the functional differentiation between them remained. Briefly, preexposed rats which did not show LI had increased DA release in the shell as did the nonpreexposed rats. In the core, preexposed rats that did not show LI had decreased DA release, similar to the nonpreexposed rats, but this was followed by a massive DA increase (see Louilot et al., this volume).
Lesions: further support for functional shell–core differentiation Studies testing LI effects from NAC lesions have yielded inconsistent results. Konstandi and Kafetzopoulus (1993) reported that excitotoxic NAC lesion spared LI, as would be predicted by the switching model. Likewise, Restivo et al. (2002) found that NAC lesion spared or potentiated LI. However, Tai et al. (1995) reported that LI was disrupted by electrolytic and excitotoxic lesion of the medial NAC, and this was not easily accommodated by the switching model. Given the functional shell–core differentiation, we investigated whether the inconsistent NAC lesion results reflected different roles of the two accumbens subregions in LI (Gal, 2000; Gal, Schiller, & Weiner, 2005; Weiner, Gal, & Feldon, 1999; Weiner, Gal, Rawlins, & Feldon, 1996a). We replicated the finding of Tai et al. that shell lesion disrupted LI. In contrast, core lesion left LI intact (Weiner et al., 1996a). The same differential LI-effect of shell and core lesions was reported by Jongen-Relo et al. (2002). While the conventional interpretation of these results would be that the shell is involved in LI while the core plays no role in this phenomenon, the assumption of the switching model, that the
382
Ina Weiner
switching mechanism subserving LI disruption resides in the NAC, led to a different explanation (Weiner et al., 1996a). Specifically, if disrupted LI stems from rapid switching to respond according to the stimulus–reinforcement contingency, then the fact that LI is disrupted by shell but not core lesion implies that shell lesion leads to rapid switching (disruption of LI) whereas core lesion spares the capacity not to switch (spared LI). From this it follows that: (a) the switching mechanism of the NAC resides in the core subterritory; (b) in the intact brain, the switching mechanism is inhibited by the shell when LI is present. This account leads to two major predictions: (1) while shell lesion will abolish LI, a larger NAC lesion which includes a lesion to the core in addition to the same shell lesion will restore LI; (2) rats with a combined shell–core lesion or only core lesion will persist in expressing LI under conditions in which normal rats switch to respond according to the stimulus–reinforcement contingency. We investigated these predictions using conditions that lead to LI in controls (weak conditioning and same context in preexposure and conditioning) and conditions that disrupt LI in controls (strong conditioning and different context in preexposure and conditioning). Under conditions which led to LI in normal rats, shell lesion disrupted LI, but a combined shell–core lesion restored LI (Weiner et al., 1999). We next tested whether rats with a combined shell–core lesion will exhibit persistent LI under conditions that disrupt LI in normal rats. Sham-lesioned rats showed LI with weak but not with strong conditioning, and when preexposure and conditioning were conducted in the same context, but not when they were conducted in different contexts. In marked contrast, NAC-lesioned rats showed LI under all of these conditions (Gal et al., 2005; Weiner et al., 1999). The same pattern was obtained with lesion confined to the core subregion: core lesion alone spared LI under conditions that yielded LI in controls and led to persistent LI under conditions that disrupted LI in controls (unpublished observations). The fact that LI was reinstated in shell-lesioned rats when they additionally sustained a lesion to the core indicates that shell lesion does not impair the acquisition of the stimulus–no-event association in preexposure or its expression in conditioning; rather the lesion causes animals to switch to respond according to the stimulus– reinforcement association under conditions in which normal rats continue to respond according to the stimulus–no-event association. This suggests that the intact shell contains a mechanism that inhibits switching. The fact that switching was prevented in rats with a combined shell–core lesion as well as with core lesion alone, such that they persisted in responding according to the stimulus–no-event contingency under conditions in which sham rats switch to respond according to the stimulus– reinforcement contingency (context shift and strong conditioning), indicates that the core contains the mechanism that enables switching in shell-lesioned and intact rats. Finally, the finding that shell lesion disrupted LI only when the core was intact (when the core was lesioned, LI was present whether the shell was lesioned or intact) indicates that the shell contains a mechanism that inhibits core-mediated switching. Thus, the specific pattern of shell and core lesion effects supports the proposition of
What the brain teaches us about latent inhibition (LI)
383
the switching model that in the intact brain, the mechanism which enables switching to respond according to the stimulus–reinforcement contingency in conditioning resides in the core, and that core-mediated switching can be inhibited by the shell. It is of interest to relate the operation of the switching mechanism to shell–core DA release patterns found with microdialysis/voltammetry too. Since in Murphy et al.’s (2000) and Jeanblanc et al.’s (2002) studies DA levels were assessed in rats manifesting LI, the switching mechanism was presumably inhibited. Thus, the fact that the presentation of the preexposed stimulus is accompanied by a decrease in shell DA release seen in the nonpreexposed animals might be taken to suggest that inhibition of shell DA release subserves the operation of the switch-inhibiting mechanism of the shell. With regard to core DA, rats expressing the preexposure effect should not show DA increase in this subregion, and this was the case in both studies. In addition, Jeanblanc et al. (2002) showed that in the conditioned PE rats, decreased DA levels in the shell were concomitant with unchanged (i.e. similar to nonconditioned control levels) DA levels in the core. This suggests that DA reduction in the shell (signaling stimulus–nothing) prevents changes in core DA levels normally resulting from conditioning (as seen in NPE rats), consistent with the assumption of the switching model that shell controls core switching by modulating core DA (see Cadoni, Solinas, & Di Chiara, 2000). Regarding disrupted LI, the switching model postulated that in preexposed rats which do not express the preexposure effect, there would be an increased DA release in the shell, because the switch-inhibiting mechanism of the shell is released, as well as an increased DA release in the core, because switching requires phasic DA release (Weiner, 2003). Such a pattern was indeed observed following LI disruption by entorhinal and subicular inactivation (Jeanblanc et al., 2004; Peterschmitt et al., 2005, 2008). Specifically, the presentation of the preexposed stimulus failed to reduce DA release in the shell. Thus, stimulus-elicited increase in shell DA was similar to that seen with a nonpreexposed stimulus. In the core, a decrease of DA release, like that seen with the nonpreexposed stimulus, was initially observed with the preexposed stimulus, and this was followed by a massive DA release (see Louilot et al., this volume). Thus, loss of the preexposure effect is accompanied by a DA release pattern identical to that of nonpreexposed stimulus with the exception of added DA release in the core, which according to the switching model would enable responding in accord with the stimulus–reinforcement association. In summary, our lesion results are consistent with the long-held view that the NAC plays a critical role in cognitive and behavioral switching and further suggest that the two NAC subregions control two distinct aspects of the switching process, switching and the inhibition of switching.
Brain areas interacting with the mesolimbic dopamine system The findings that manipulations within NAC can both disrupt and promote the expression of LI provide evidence that NAC plays a key role in processes responsible for generating conditioned responding appropriate to the stimulus–no-event or the
384
Ina Weiner
stimulus–reinforcement association. While it is widely accepted that NAC is involved in mediating the expression of motivated/conditioned behavior, the associative, motivational and contextual knowledge underlying this operation is believed to derive from regions afferent to the NAC, namely, the hippocampus, the basolateral amygdala (BLA) and the prefrontal cortex (Everitt & Robbins, 1992; Floresco, Blaha, Yang, & Phillips, 2001; Grace, 2000; Groenewegen et al., 1999; Howland, Taepavarapruk, & Phillips, 2002; Mizumori, Yeshenko, Gill, & Davis, 2004; Mulder, Hodenpijl, & Lopes da Silva, 1998; Pennartz et al., 1994; Phillips, Ahn, & Howland, 2003; Robbins & Everitt, 1996; Zahm & Brog, 1992). Thus, in LI, the NAC would be expected to produce responding appropriate to the stimulus–no-event versus the stimulus–reinforcement association based on the integration of the relevant information from its limbic and cortical sources of input. The major prediction from the latter is that shell–core functional dissociation in LI should be paralleled by effects of lesions to major sources of NAC input. As we shall see below, this is indeed the case.
The limbic system: functional dissociation between the entorhinal cortex, the hippocampus and the basolateral amygdala The entorhinal cortex LI is disrupted by excitotoxic lesions of the entorhinal cortex (EC) (Coutureau, Galani, Gosselin et al., 1999; Coutureau, Lena, Dauge, & Di, 2002; Shohamy, Allen, & Gluck, 2000; Yee, Feldon, & Rawlins, 1995, 1997), suggesting that the normal expression of LI requires the integrity of the EC. Moreover, it has been shown that LI is disrupted by temporary inactivation of the EC using TTX (Jeanblanc et al., 2004; Seillier, Dieu, Herbeaux et al., 2007) or muscimol (Lewis & Gould, 2007a) infusion confined to the preexposure stage, whereas LI is spared when the same infusions are confined to the conditioning stage (Jeanblanc et al., 2004; Lewis & Gould, 2007a). These findings imply that the intact EC plays a critical role in the acquisition of inattention to the preexposed stimulus, or in the encoding of the association between the preexposed stimulus and no consequences, but not in expression/retrieval of this association. Surprisingly, Seillier et al. (2007) also found that intra-EC TTX infusion before both the preexposure and the conditioning stages spares LI, in contrast to the effects of excitotoxic lesion of this structure. The NMDA antagonist APV, the cAMP inhibitor Rp-cAMP and MAPK inhibitor U0126, infused into the entorhinal cortex prior to preexposure, produced disruption of LI, indicating that signal transduction mechanisms within these regions are involved in LI (Lewis & Gould, 2007b; see Gould in this volume). It should be noted that entorhinal cortex lesion-induced LI disruption was reversed by systemic treatment with haloperidol (Schmajuk, Lam, & Christiansen, 1994; Yee et al., 1995), suggesting that LI disruption was underlaid by DA hyperactivity, probably in the NAC.
What the brain teaches us about latent inhibition (LI)
385
The hippocampus and basolateral amygdala: dissociation between context and reinforcement information Studies using conventional hippocampal lesions have shown LI disruption (Ackil, Mellgren, Halgren, & Frommer, 1969; see Buhusi, Gray, & Schmajuk, 1998, for a discussion; Kaye & Pearce, 1987a, 1987b; Schmajuk et al., 1994; Solomon & Moore, 1975), suggesting that intact hippocampus is necessary for LI. As noted above, based on these findings and on Robbins and Everitt’s (1982) suggestion that NAC switching is controlled by the hippocampus, the original switching model (Weiner, 1990) proposed that NAC switching in LI also is controlled by the hippocampus (Weiner, 1990). However, more recently, conventional hippocampal lesions have been reported to spare (Clark, Feldon, & Rawlins, 1992; Oswald, Yee, Rawlins et al., 2002) and even potentiate (Purves, Bonardi, & Hall, 1995) LI, and most studies of hippocampal excitotoxic lesions as well as muscimol inactivation of this structure have been shown to leave LI intact (Coutureau et al., 1999; Holt & Maren, 1999; Honey & Good, 1993; Pouzet, Zhang, Weiner et al., 2004; Reilly, Harley, & Revusky, 1993; Shohamy et al., 2000) (but see Han, Gallagher, & Holland, 1995). Consistent with the latter findings, destruction of the fornix-fimbria was also found not to affect LI (Pouzet, Veenman, Yee et al., 1999; Weiner, Feldon, Tarrasch et al., 1998) (but see Cassaday, Mitchell, Williams, & Gray, 1993). Finally, chemical activation of the hippocampus by local NMDA infusion disrupted LI (Pouzet et al., 2004). Taken together these findings indicate that intact hippocampus is not necessary for either the acquisition or the expression of LI under conditions that normally produce this phenomenon. This latter point is supported by findings that in rats with excitotoxic lesion or muscimol inactivation of the hippocampus LI persists under conditions that normally attenuate it, namely, context change between preexposure and conditioning (Holt & Maren, 1999; Honey & Good, 1993). Since muscimol-induced inactivation was confined to a retention-test stage, LI resistance to context change could not be due to stronger learning of the stimulus–no-event contingency in preexposure, but rather to an intact expression of this learning in different context condition. In other words, whereas normal rats conditioned in different context from that of preexposure switch to respond according to the stimulus–reinforcement association, rats with a lesioned hippocampus continue to respond according to the stimulus–noevent association in spite of context change. The fact that in hippocampal rats LI is intact but its contextual modulation is lost indicates that the hippocampus is not necessary for the expression of the stimulus–no-event contingency in the context of preexposure but rather for the prevention of its expression when conditioning takes place in a different context. Studies of BLA involvement in LI have been limited compared to the hippocampus. Two studies assessing BLA lesion under conditions yielding LI in controls yielded contradictory results. Specifically, LI was reported to be spared by electrolytic lesions of BLA (Weiner, Tarrasch, & Feldon, 1995), but Coutureau et al. (2001)
386
Ina Weiner
reported that an excitotoxic BLA lesion disrupted LI. In addition, Schauz and Koch (2000) showed that LI was disrupted by the infusion of the competitive NMDA antagonist AP-5 into BLA before preexposure. The switching model would expect BLA lesion to produce persistent rather than disrupted LI, because there is extensive evidence that BLA lesions impair the attribution of motivational value to the conditioned stimulus (Cador, Robbins, Everitt et al., 1991; Everitt & Robbins, 1992; Fanselow & LeDoux, 1999; Fendt & Fanselow, 1999; Gallagher & Chiba, 1996; LeDoux, 1992; Mizumori et al., 2004; Mulder et al., 1998; Robbins & Everitt, 1996), and this would promote the expression of the stimulus–no-event association. We have tested the effects of BLA on LI under three conditions that prevent LI in controls, weak preexposure, strong conditioning and context shift (Schiller & Weiner, 2004, 2005). We replicated our previous finding that BLA lesions spare LI under conditions that yield LI in controls, with both electrolytic and excitotoxic lesion. Furthermore, we showed, as predicted by the switching model, that rats with BLA lesion persisted in showing LI under two conditions that led to LI disruption in controls, namely, weak preexposure and strong conditioning. These results are consistent with the prediction of the switching models as well as with other findings and views that BLA-lesioned rats are incapable of altering behavior to stimuli when reinforcement contingencies or reward values are changed (Baxter, Parker, Lindner et al., 2000; Cardinal, Parkinson, Hall, & Everitt, 2002; Everitt & Robbins, 1992; Hatfield, Han, Conley et al., 1996; Holland & Gallagher, 1999; Malkova, Gaffan, & Murray, 1997; Rolls, 2000a; Schoenbaum, Setlow, Nugent et al., 2003). Interestingly, Louilot and Besson (2000) reported that functional BLA inactivation blocked the decrease in core DA release associated with a conditioned stimulus, as was found for stimulus preexposure (Jeanblanc et al., 2002), suggesting that BLA inactivation reduces the impact of conditioning like preexposure itself. This may point to one potential mechanism by which BLA dysfunction may lead to persistent LI. No persistent LI was seen in BLA-lesioned rats with context shift: BLA rats, like controls, did not show LI under this condition (Schiller & Weiner, 2005). Although BLA has been shown to play a central role in contextual conditioning (Fendt & Fanselow, 1999; LeDoux, 1992), this applies to situations in which the context serves as the CS. As detailed above, the function of the context in LI is apparently mediated by the hippocampus. Indeed, we have found that the role of the hippocampus is specific to changes in context: while excitotoxic hippocampal lesion produced persistent LI with context shift, it failed to produce persistent LI with strong conditioning (unpublished observations). Taken together, our results revealed a double dissociation between the effects of hippocampal and BLA lesions on LI: while hippocampal lesion produced persistent LI with context shift but not with strong conditioning, BLA lesion produced persistent LI with strong conditioning but not with context shift. In other words, while hippocampal lesion produces contextindependent LI, BLA lesion produces reinforcement impact-independent LI.
What the brain teaches us about latent inhibition (LI)
387
By extension, just as the function of the hippocampus in the intact brain is to prevent the expression of LI in a different context, the function of BLA is to prevent LI when the impact of reinforcement is increased.
Prefrontal cortex: functional dissociation between medial prefrontal cortex and orbitofrontal cortex Experiments assessing the role of mPFC (medial prefrontal cortex) perturbations on LI under conditions yielding LI in controls have consistently revealed no effect. Electrolytic lesions of the mPFC as well as smaller lesions confined to two of its subregions, the dorsal anterior cingulate area and/or the infralimbic cortex, spared LI (Joel, Weiner, & Feldon, 1997). Intact LI was also shown following excitotoxic lesion of the mPFC (Lacroix, Broersen, Weiner, & Feldon, 1998; Lacroix, Spinelli, White, & Feldon, 2000) or direct injections of DA agonists or antagonists into the mPFC (Broersen, Feldon, & Weiner, 1999; Broersen, Heinsbroek, de Bruin, & Olivier, 1996; Ellenbroek, Budde, & Cools, 1996; Lacroix, Broersen, Feldon, & Weiner, 2000). The fact that mPFC lesion spares LI coupled with the known difficulty of mPFC-lesioned animals to alter behavior in response to changed reinforcement contingencies, typically persisting in responding according to previous contingency (e.g. Aggleton, Neave, Nagle, & Sahgal, 1995; de Bruin, Sanchez-Santed, Heinsbroek et al., 1994; Dias & Aggleton, 2000; Dias, Robbins, & Roberts, 1997; Joel et al., 1997; Kolb, 1984; Morgan & LeDoux, 1995; Ragozzino, Detrick, & Kesner, 1999; Ragozzino, Wilcox, Raso, & Kesne, 1999), suggests that lesion to the mPFC should produce persistent LI. We (Schiller & Weiner, 2004) therefore tested the effects of lesion to the mPFC as well as another PFC subregion known to be involved in behavioral flexibility, the orbitofrontal cortex (OFC) (Birrell & Brown, 2000; Dias, Robbins, & Roberts, 1996; Dias et al., 1997; McAlonan & Brown, 2003), on LI with strong conditioning. Rats with mPFC and OFC lesions showed LI under conditions that yielded LI in sham controls. With strong conditioning, LI was disrupted in sham as well as mPFC-lesioned rats, but OFC-lesioned rats persisted in exhibiting LI in spite of strong conditioning. The finding that mPFC lesion did not produce persistent LI may appear perplexing. However, the mPFC and the OFC may subserve different types of behavioral flexibility. For example, OFC but not mPFC lesions impair reversal learning, whereas mPFC but not OFC lesions impair extradimensional shift. Consequently, it has been suggested that the OFC may subserve switching of stimulus–reinforcement associations while mPFC may be involved in switching of general or higher-order rules or strategies (Birrell & Brown, 2000; Brown & Bowman, 2002; Dias et al., 1996, 1997; Kesner, 2000; McAlonan & Brown, 2003). Because prevention of LI expression requires switching between stimulus–outcome associations involving the same stimulus, this process would be expected to be sensitive to OFC but not to mPFC damage. In any event, the findings that mPFC lesions neither disrupt LI nor induce abnormally persistent LI strongly suggest that this region is not involved in LI.
388
Ina Weiner
The fact that OFC lesion, like BLA lesion, produces LI with strong conditioning is consistent with many findings showing similar outcomes following damage to BLA and OFC in tasks in which reinforcement contingencies or reward values are changed (Baxter et al., 2000; Gallagher, McMahan, & Schoenbaum, 1999; Hatfield et al., 1996; Malkova et al., 1997; Rolls, 2000a, 2000b; Schoenbaum et al., 2003). Such findings have been taken to suggest that BLA and OFC form a functional system that enables stimuli to access the current affective value of the associated outcome for flexible adjustment of conditioned responding (Baxter et al., 2000; Cardinal et al., 2002; Everitt & Robbins, 1992; Gallagher et al., 1999; Gallagher & Schoenbaum, 1999; Hatfield et al., 1996; Holland & Gallagher, 1999; Pickens, Saddoris, Setlow et al., 2003; Rolls, 1996, 1999, 2000a, 2000b; Schoenbaum, Nugent, Saddoris, & Setlow, 2002; Schoenbaum & Setlow, 2001; Schoenbaum et al., 2003; Setlow, Gallagher, & Holland, 2002). Our data are consistent with the above findings and ideas in demonstrating that both BLA- and OFC-lesioned rats failed to reverse their CS–no-event association when the predictive value of the CS was altered.
The role of fronto-limbic-ventral striatal circuitry in LI: response selection according to stimulus–no-event vs. stimulus–reinforcement association The findings surveyed above can be summarized as follows. (1) The NAC and its dopaminergic innervation form a crucial component of the neural circuitry of LI. (2) There is a clear functional differentiation between the shell and the core subregions, whereby damage to the shell disrupts LI and damage to the core renders LI abnormally persistent. (3) The effects of shell and core lesions parallel those produced by lesions to the major sources of input to the NAC: entorhinal cortex lesion, like shell lesion, disrupts LI, whereas hippocampal and BLA lesions, like core lesion, spare LI; moreover, hippocampal lesion, like core lesion, produces contextindependent LI, whereas BLA lesion, like core lesion, produces reinforcement impactindependent LI. (4) Lesion to the OFC produces reinforcement impact-independent LI like NAC core and BLA lesions, whereas mPFC lesion is inactive. These data provide evidence for a functional specialization of the NAC shell and core subregions and their associated circuitries in LI, and confirm that the neural mechanisms through which the CS–no-event and the CS–reinforcement association control performance are dissociable. Specifically, since shell and entorhinal cortex lesions disrupt LI, it follows that these regions comprise the circuitry mediating the information and the behavioral impact of the stimulus–no-event association. Conversely, given that the core, hippocampus and the BLA produce persistent LI, it follows that these regions comprise the circuitry mediating the information and the behavioral impact of the stimulus–reinforcement association. Finally, the fact that lesions to the NAC mimic all of the effects produced by their sources of
What the brain teaches us about latent inhibition (LI)
389
input, namely, disrupted LI and persistent LI with both context change and strong conditioning, indicates that the inputs provided by the two circuitries interact at the level of the NAC. We suggest that in the intact brain, the two circuitries operate in tandem to determine which of the two associations, CS–no-event or CS–reinforcement, gains control over behavior. More specifically, the switching mechanism, which resides in the core, is activated at the time of conditioning, when the previously nonreinforced stimulus is followed by reinforcement. Depending on the information reaching the NAC from the EC, HP and BLA, shell can inhibit the switching mechanism of the core, so that the stimulus–no-event contingency gains control over behavior, or shell’s inhibition is removed, allowing the expression of the stimulus–reinforcement contingency. The critical input to the shell, which subserves the inhibition of the switching mechanism, i.e., “the stimulus signals nothing”, arrives from the entorhinal cortex. This information subserves the attribution of low associability to the stimulus under all circumstances. This stimulus-specific information is modulated by different types of information, including information on context and on impact of reinforcement, that are fed, during conditioning, into the switch circuitry. Since hippocampal lesions produce context shift-resistant LI, it can be concluded that the hippocampus provides the information about the context, consistent with the general role of this region in contextual processing, Likewise, since BLA lesion produces persistent LI with strong conditioning, it is likely to provide the information on the impact of reinforcement/affective value of the stimulus, consistent with BLA role in reinforcement and motivational processes (LeDoux, 1992; Robbins, Cador, Taylor, & Everitt, 1989; Robbins & Everitt, 1996). It remains unclear at present whether the OFC provides information to the NAC or to the BLA, as part of a functional OFC–BLA system which enables stimuli to access the current motivational value of stimuli and thus flexible adjustment of conditioned responding (see above for references). The preexposure effect is expressed under conditions of low mismatch between conditions of preexposure and conditioning, which occurs with the following combination of signals during conditioning: from the entorhinal cortex ¼ same familiar stimulus; from the hippocampus ¼ same context; from the amygdala ¼ low impact of reinforcement. In this case, comparison between past and present conditions results in assigning the CS a low associability value, and inhibiting the switching mechanism, enabling the continuation of responding according to past predictions. High mismatch occurs when the hippocampal and/or amygdalar inputs signal a change in context and/or high impact of reinforcement. In this case, comparison between present and past conditions results in assigning the CS a high associability value, and a removal of inhibition from the switching mechanism, leading to responding according to present predictions. Inhibition of switching is subserved by decreased DA release in the shell and no change in the core, whereas switching is associated with increased DA release in both subregions.
390
Ina Weiner
While it is postulated here that the results of the mismatch analysis are enacted in the core, where the switching mechanism resides, it is not yet clear how the different types of information reach the core, or where mismatch is calculated. It is apparent that the input from the entorhinal cortex does not reach the core directly but is channeled via the shell. There are at least two pathways via which information from the shell can reach the core. The first is an “open pathway” originating in the shell and traversing the ventral pallidum, thalamus and cortex to reach the core (Zahm, 1999; Zahm & Brog, 1992). The second is the shell projections to the VTA, and the latter’s DA projections to the core (Berendse, Groenewegen, & Lohman, 1992; Groenewegen et al., 1991, 1996; Otake & Nakamura, 2000; Zahm & Brog, 1992). The second pathway is of particular importance since it is likely to provide the anatomical substrate for the here postulated shell-mediated control of the switching mechanism of the core. Thus, it is suggested that via its inhibitory projections to the VTA, the shell can attenuate DA input to the core, thus preventing the switch to the stimulus– reinforcement contingency and allowing the expression of the stimulus–no-event contingency. It is also not clear where the information from the entorhinal cortex is integrated with information from the hippocampus and amygdala. Since the latter regions project to both the VTA and the core, it is possible that the calculation of mismatch is performed at the level of either the VTA or core, or both. Given that it has been suggested that DA cells compute an “error signal” between predicted and actual events (Schultz, 1998), and that a phasic increase in accumbal DA has been suggested to facilitate behavioral switching (Pennartz et al., 1994; Redgrave, Prescott, & Gurney, 1999; Weiner & Joel, 2002), information favoring a switch (e.g., a contextshift or large associative strength), which is channeled from the hippocampus/BLA to both the core and the VTA, may act simultaneously to direct switching in the core and to facilitate such a switch via the projections to the VTA. The latter can be counteracted by switch-inhibiting information channeled from the entorhinal cortex via the shell to the VTA. Finally, while it is clear that the information of “same stimulus” provided by the entorhinal cortex to the shell is critical for the inhibition of switching, it is not clear whether the hippocampus and BLA provide information that contributes to the inhibition of switching (same context and low impact of reinforcement), or only information which is critical for switching (different context and high impact of reinforcement). In this regard, it should be pointed out that the status of reinforcement and context information in the conditioning stage is different, because the unexpected occurrence of reinforcement is an inherent characteristic of the LI procedure, whereas change in context is not. Therefore, it is reasonable to assume that BLA participates in conditioning from the very first stimulus–reinforcement pairing, whereas the hippocampus might remain “silent” unless there is a change in context. In sum, it is suggested that: (a) the mechanism responsible for switching to respond according to the new stimulus–reinforcement contingency in the conditioning stage
What the brain teaches us about latent inhibition (LI)
391
resides in the NAC core; (b) the presence of LI, i.e., continuing to respond according to the previous stimulus–no-event contingency, or its absence, i.e., switching to respond according to the stimulus–reinforcement contingency, is mediated by signals from the entorhinal cortex/hippocampus/BLA. If the sum of the converging signals signifies low mismatch, the switching mechanism is inhibited, and LI is present; if the sum of signals signifies high mismatch, inhibition is removed, and LI is absent; (c) damage to the switch-inhibiting mechanism (shell or its entorhinal input) disrupts LI; (d) damage to the core switching mechanism produces LI which persists under all conditions that disrupt it in normal rats; damage to the sources of mismatch information (hippocampus, BLA) produces LI which persists under some of the conditions that lead to its disruption in control rats (only context-independent LI, or only reinforcement impact-independent LI); (e) switching is subserved by increased DA release in the core, which can be inhibited by the shell via its control of DA input to the core (see Figure 16.2; for related theoretical treatments of LI, see Buhusi et al., 1998; Schmajuk, Buhusi, & Gray, 1998; Schmajuk, Cox, & Gray, 2001; Schmajuk, Gray, & Lam, 1996).
Effects at preexposure The above scheme focuses on processes occurring in the conditioning stage, which ultimately determine the choice of the behavioral strategy, and is silent with respect to processes occurring in preexposure, except for postulating the involvement of the entorhinal cortex, which is presumed to encode the stimulus–no-event contingency. However, findings suggesting that the hippocampus and BLA are involved also in the preexposure stage (Romano, 1999; Salafia & Allan, 1980, 1982; Schauz & Koch, 2000), and, more critically, that manipulations of these structures in preexposure can disrupt LI (Salafia & Allan, 1982; Schauz & Koch, 2000), are rather disturbing for the model, because a question arises as to how a dysfunction of a structure can lead to disrupted LI when occurring in preexposure yet lead to persistent LI when occurring in both stages. Based on findings that temporary inactivation of the hippocampus during a retrieval test stage produced persistent (context-independent) LI whereas permanent hippocampal lesions that are present throughout preexposure and test stages disrupted LI, it has been argued that the hippocampus is responsible for both the acquisition of the stimulus–no-event association in preexposure as well as for its retrieval in the context of preexposure (Holt & Maren, 1999; Maren & Holt, 2000). However, Honey and Good (1993) obtained context-independent LI with a permanent excitotoxic lesion, making the two-process explanation less likely. Alternatively, it is possible that it is the context rather than the stimulus that is encoded by the hippocampus during preexposure. Indeed, if the hippocampus is responsible for providing the contextual information during conditioning/retrieval, then it is reasonable that
392
Ina Weiner
(a) OFC NAC HIP context
core switching mechanism
BLA
Switch to respond according to “stimulusreinforcement”
impact of reinforcement shell EC “stimulus-no event”
switch inhibiting mechanism
inhibition of switching
VTA
Figure 16.2. (a) A schematic diagram of brain regions and pathways involved in LI and their function in the expression and disruption of the preexposure effect as postulated by the switching model. BLA, basolateral amygdala; EC, entorhinal cortex; NAC, nucleus accumbens; mPFC, medial prefrontal cortex; VTA, ventral tegmental area. (b) Pathways/inputs subserving LI expression, i.e., responding according to the stimulus–no-event contingency. The critical projections include a pathway from the EC to the NAC shell, via which a “stimulus–no-event” signal is transmitted, and the subsequent projection from the shell to the VTA, via which the shell inhibits DA input to the core, and in this way prevents switching to respond according to the stimulus–reinforcement contingency. (c) Pathways/inputs subserving LI disruption, i.e., responding according to the stimulus–reinforcement contingency. The projections include pathways from the HIP and BLA to the NAC core and to VTA (note that the latter signify HIP and BLA modulation of DA cell activity rather than specific anatomical pathways), via which information on context and impact of reinforcement, respectively, is transmitted. Information signifying high mismatch between preexposure and conditioning (different context and/or high impact of conditioning) overrides the switch-inhibiting information channeled from the EC, resulting in switching to respond according to the stimulus–reinforcement contingency. At present, it is not known what structures receive core output and subserve the actual switch in behavior. It is reasonable that the switching signal is channeled from the core via its two major output nuclei, the ventral pallidum and the substantia nigra, to the basal ganglia–thalamocortical circuits of the dorsal striatum, in which the actual selection of behavioral output is executed (Weiner, 1990; Weiner & Joel, 2002).
it monitors and encodes such information also during preexposure. If this is correct, then an inactivation of this structure that is confined to preexposure should have the same effect as its inactivation during retrieval, i.e., produce context-independent LI, because the hippocampus will not detect “different” context in conditioning/ retrieval.
What the brain teaches us about latent inhibition (LI)
393
(b) OFC NAC HIP context BLA impact of reinforcement
core switching mechanism
Switch to respond according to “stimulusreinforcement”
shell
EC “stimulus-no event”
switch inhibiting mechanism
inhibition of switching
VTA
(c) OFC NAC HIP context
core switching mechanism
BLA impact of reinforcement
shell
EC “stimulus-no event”
switch inhibiting mechanism
VTA
Figure 16.2. (cont.)
Switch to respond according to “stimulusreinforcement”
394
Ina Weiner
A parallel postulate cannot be made with regard to BLA, however, because the information presumed to be provided by this region during conditioning, namely, the impact of reinforcement, is absent during preexposure. Therefore, it is likely that BLA is indeed involved in the processing of the stimulus during preexposure (Schauz & Koch, 2000). The seemingly inconsistent findings that interference with BLA processing during preexposure disrupts LI, while interference with such processing during both preexposure and conditioning leads to persistent LI, can be reconciled if it is assumed that BLA manipulations in preexposure do not prevent the acquisition of the stimulus–no-event association but rather decrease its strength. While this would be expected to disrupt LI if BLA lesion is confined to preexposure, because in such a case the impact of conditioning remains unchanged, when BLA lesion is present in both stages decreased functional impact of preexposure is presumed to be followed by decreased impact of conditioning, and under such conditions LI can emerge. Thus, in our experiments, control rats did not show LI with 40 preexposures and five conditioning trials whereas BLA lesioned rats did, and we assumed that the latter reflected BLA lesion-induced reduction in the impact of conditioning so that BLA rats performed like normal rats receiving 40 preexposures and two conditioning trials (which yield LI in normal rats). However, if BLA lesion reduced the impact of both preexposure and conditioning, BLA rats could perform like normal rats receiving 20 preexposures and two conditioning trials (which also yield LI in normal rats), and thus still show persistent LI. The same possibility could apply to the hippocampus if this region does participate in the processing of the stimulus–no-event contingency; thus, although hippocampal damage would decrease the impact of preexposure, the LI effect could still be expressed in conditioning, because the information of “different context” is lacking. The major prediction that follows from the above argument is that the effects of hippocampal and BLA lesions should be modifiable by changes in the parameters of the LI procedure. In support of the modulating role of the hippocampus, Salafia and Allan (1980, 1982) reported that high-level electrical stimulation of the hippocampus during preexposure disrupted LI, whereas a low level of such stimulation enhanced LI (Schmajuk et al., 1996, 1998, 2001).
The role of context in LI Because LI is a context-specific phenomenon whereby the stimulus–no-event association is expressed only in the context of preexposure, it is widely held that the role of the context in LI is to restrict the expression of the preexposure effect to the context of preexposure (Bouton, 1993; Hall & Channell, 1985; Holt & Maren, 1999; Lubow & Gewirtz, 1995; Westbrook, Jones, Bailey, & Harris, 2000). This view, in somewhat differing theoretical versions, is probably the most conspicuous theme in the present book (see chapters by Escobar & Miller; Honey, Iordanova & Good; Westbrook & Bouton; Lubow). I would
What the brain teaches us about latent inhibition (LI)
395
like to forward a different view based on lesion data. The fact that LI is present in hippocampus- and core-lesioned rats in both the same and different context conditions indicates that similarity of context is not a necessary condition for either the expression or the acquisition of LI. On the contrary, in the absence of contextual information, LI becomes a widely generalizable phenomenon whereby the stimulus–no-event association acquired in preexposure continues to control animals’ behavior in all contexts, irrespective of whether they are the same or different from that of the preexposure stage. It follows from this that in the intact brain, contextual control of LI (mediated by the hippocampus and its NAC inputs) is not achieved by restricting the expression of the preexposure effect to the context of preexposure, but rather by preventing its expression (i.e., disrupting LI) when conditioning takes place in a different context. In contrast to others, our view of LI assigns no special status to context; it is merely one of several modulating factors that determine whether LI is expressed or not expressed. Since, according to the switching model, such expression at any given time is a function of the specific balance between the behavioral impact of preexposure and conditioning, and the latter can be affected by changing the procedural parameters (e.g., an increase in number of conditioning trials will counteract the impact of preexposure, and vice versa), one of the predictions of this approach is that the effects of context change also can be counteracted by manipulations of other factors. In support of this expectation, it was shown that massive preexposure abolishes the context-specificity of LI (Wheeler, Chang, & Miller, 2003). Although these authors considered this outcome as consistent with the conventional models of the role of context in LI, this seems unlikely. As stated by Wheeler et al., “For all of these models, latent inhibition is context specific because the CS–context association developed during preexposure will not affect responding if the CS is conditioned in a different context”. However, is not clear why massive preexposure causes the CS–context association developed during preexposure to lose its capacity to mediate responding to the CS if it is conditioned in a different context. To the best of my understanding no conventional context-dependent model incorporates a mechanism allowing increased preexposure to counteract the effect of context change. We interpret the finding that increased impact of preexposure can counteract the effect of change in context as showing that also in normal rats, the acquisition and expression of LI are not dependent on context. Interestingly, context does not play a role in LI early in development; young rats show context-independent LI (Yap & Richardson, 2005; Zuckerman, Rimmerman, & Weiner, 2003). It is only with brain maturation that context acquires the capacity to modulate LI, so that LI is not expressed in a context different from that of preexposure. Resistance to context change in juvenile rats is accompanied by resistance to all the manipulations known to disrupt LI expression in adult rats, including weak preexposure, strong conditioning, and amphetamine administration (Zuckerman et al., 2003). We suggested that the sensitivity to these manipulations develops as
396
Ina Weiner
the regions responsible for flexibility of behavior mature – these are the precise same regions that feed into the NAC and enable switching and inhibition of switching in response to situational demands.
Disrupted and persistent LI: implications for brain–behavior relationship and schizophrenia The most remarkable conclusion from the above lesion studies (as well as drug studies; see Chapter 13 on the pharmacology of LI) is that the function of most regions (and neurotransmitter systems) in the intact brain that play a role in LI is not to enable its expression but rather to enable its disruption. Thus, the expression of LI does not depend on the integrity of several brain areas, namely, the orbitofrontal cortex, the nucleus accumbens, the hippocampus and the BLA, because damage to these areas spares LI under conditions allowing for it in controls. Instead, these regions play a role in the non-expression of LI, limiting the appearance of LI to certain conditions (e.g., BLA might be involved in LI restriction based on the impact of preexposure and conditioning, while the hippocampus is involved in contextual restriction of LI). In the absence of such a widespread inhibitory function, the effects of inconsequential stimulus preexposure would have been extremely robust, generalizing across contexts and counteracting the expression of subsequent conditioning to the stimulus; LI would be rigid and insensitive to the ever-changing situational demands. Importantly, this pattern of results has general implications for the elucidation of structure–function relationships in LI. The fact that distinct regions are responsible for the expression vs. the disruption of LI implies that a complex pattern of effects should be expected following manipulations of such structures. Damage to structures that normally subserve the expression of LI will lead to its disruption, whereas stimulation of such structures will enhance it. Conversely, damage to structures normally subserving the disruption of LI will leave it intact or enhance it, whereas stimulation of such structures will disrupt it. Moreover, manifested behavioral effects of combined damage/stimulation of structures that subserve the expression and disruption of LI may be misleading, because such regions compete and the effects of their damage/stimulation may cancel or mask each other. Lastly, the fact that many brain regions/neurotransmitter systems are normally responsible for the disruption of LI has important implications for the interpretation of LI abnormalities in schizophrenia. The concept of LI deficit in schizophrenia has been almost exclusively focused on disrupted LI. However, although most clinical reports are consistent in indicating that LI disruption is associated with acute schizophrenia (Baruch et al., 1988a; Gray et al., 1992a, 1995b; Rascle et al., 2001; Vaitl, Lipp, Bauer et al., 2002; Vaitl & Lipp, 1997), other studies have failed to find this effect (Swerdlow et al., 1996; Williams et al., 1998; see Swerdlow and Kumari & Ettinger, in this volume). Indeed, it is evident that LI is normal in chronic schizophrenia (e.g., Lubow, Weiner, Schlossberg, & Baruch, 1987), which is at first sight perplexing.
What the brain teaches us about latent inhibition (LI)
397
The rat LI data may shed light on this puzzle because they show that LI abnormality is not exclusively manifested as a loss of this phenomenon but that it can also take the form of abnormal persistence. In fact, dysfunction of most regions considered critically involved in schizophrenia, namely, prefrontal cortex, amygdala, and hippocampus, as well as NAC core, should not be expected to disrupt LI. Importantly, since persistent LI appears as normal LI unless an appropriate procedure is used to reveal its excess, normal LI found in schizophrenia patients could very well obscure persistent LI. Thus, a major implication of animal studies for research in schizophrenia patients is that “normal” LI is to be expected in most patients, and that procedures need to be developed that can detect abnormally persistent LI in such patients. Finally, the two poles of LI abnormality produced by brain lesions have implications for understanding brain substrates of schizophrenia. Beginning from Kraeplin’s (1919) classical observations that schizophrenia patients are unable to focus on relevant stimuli on the one hand, and are irresistibly drawn to irrelevant stimuli on the other hand, disturbances in switching capacity have been repeatedly noted in schizophrenia (e.g., Anscombe, 1987; Bleuler, 1911; Broen, 1968; Frith, 1979; Lyon, 1991; Magaro, 1980; Payne, 1966; Shakow, 1962; Weiner & Joel, 2002). Lesion results described above indicate that deficient LI can mimic two extremes of deficient cognitive switching seen in schizophrenia: excessive switching between associations, manifested in disrupted LI under conditions that yield LI in normal rats, and retarded switching between associations, manifested in persistent LI under conditions which disrupt the phenomenon in normal rats, and that these abnormalities can be mapped onto underlying neural systems which are considered dysfunctional in schizophrenia: namely, the mesolimbic DA system and its major sources of input, including the hippocampus, the entorhinal cortex and the amygdala (Beckmann & Jakob, 1991; Bogerts, 1991, 1993; Bogerts, Meertz, & Schonfeldt-Bausch, 1985; Csernansky, Murphy, & Faustman, 1991; Grace, 1991; Harrison, 1999; Jakob & Beckmann, 1986; Kovelman & Scheibel, 1984; Moore, Fadel, Sarter, & Bruno, 1999; O’Donnell & Grace, 1998; Swerdlow & Koob, 1987). Importantly, the two LI abnormalities result from a dysfunction of distinct (albeit interconnected) brain circuitries within these regions, with disrupted LI linked with entorhinal cortex and shell dysfunction, and persistent LI linked with amygdalar, hippocampal and core dysfunction. As we have argued elsewhere, disrupted and persistent LI can be seen as two poles of dysfunctional attentional control, namely, a failure to inhibit attention to irrelevant stimuli and a failure to re-deploy attention when previously irrelevant stimuli become relevant. The former would likely give rise to the aberrantly increased salience perception and distractibility that are associated with psychotic symptoms (Gray et al., 1991; Ikemoto & Panksepp, 1999; Kapur, 2003; Kapur, Mizrahi, & Li, 2005; Smith, Li, Becker, & Kapur, 2006; Swerdlow & Koob, 1987; Weiner, 1990, 2003), whereas the latter would likely result in the cognitive inflexibility and impaired attentional shifting that are associated with negative/cognitive symptoms (Carlsson & Carlsson, 1990;
398
Ina Weiner
Krystal, D’Souza, Mathalon et al., 2003; Moghaddam, Adams, Verma, & Daly, 1997; Weiner, 2003). Importantly, lesion-disrupted LI is reversed by both typical and atypical APDs (but see Coutureau, Gosselin, & Di Scala, 2000) whereas lesion-induced persistent LI is reversed only by atypical APDs, consistent with their differential efficacy in the clinic and supporting the possibility that lesioninduced disrupted and persistent LI may be relevant to positive and negative/ cognitive symptoms, respectively. Thus, understanding of brain substrates mediating disrupted and persistent LI may have important implications for the analysis of the brain circuits involved in positive and negative/cognitive symptoms of schizophrenia and for the action of APDs.
References Ackil, J., Mellgren, R. L., Halgren, C., & Frommer, S. P. (1969). Effects of CS preexposure on avoidance learning in rats with hippocampal lesions. Journal of Comparative and Physiological Psychology, 69, 739–747. Aggleton, J. P., Neave, N., Nagle, S., & Sahgal, A. (1995). A comparison of the effects of medial prefrontal, cingulate cortex, and cingulum bundle lesions on tests of spatial memory: evidence of a double dissociation between frontal and cingulum bundle contributions. The Journal of Neuroscience, 15, 7270–7281. Angrist, B., Rotrosen, J., & Gershon, S. (1980). Responses to apomorphine, amphetamine, and neuroleptics in schizophrenic subjects. Psychopharmacology (Berl.), 67, 31–38. Angrist, B., Sathananthan, G., Wilk, S., & Gershon, S. (1974). Amphetamine psychosis: behavioral and biochemical aspects. Journal of Psychiatric Research, 11, 13–23. Angrist, B. M., Shopsin, B., & Gershon, S. (1971). Comparative psychotomimetic effects of stereoisomers of amphetamine. Nature, 234, 152–153. Anscombe, R. (1987). The disorder of consciousness in schizophrenia. Schizophrenia Bulletin, 13, 241–260. Arnt, J., & Skarsfeldt, T. (1998). Do novel antipsychotics have similar pharmacological characteristics? A review of the evidence. Neuropsychopharmacology, 18, 63–101. Asarnow, R. F., Marder, S. R., Mintz, J., Van Putten, T., & Zimmerman, K. E. (1988). Differential effect of low and conventional doses of fluphenazine on schizophrenic outpatients with good or poor information-processing abilities. Archives of General Psychiatry, 45, 822–826. Bakshi, V. P., Geyer, M. A., Taaid, N., & Swerdlow, N. R. (1995). A comparison of the effects of amphetamine, strychnine and caffeine on prepulse inhibition and latent inhibition. Behavioural Pharmacology, 6, 801–809. Baruch, I., Hemsley, D. R., & Gray, J. A. (1988a). Differential performance of acute and chronic schizophrenics in a latent inhibition task. The Journal of Nervous and Mental Disease, 176, 598–606. Baruch, I., Hemsley, D. R., & Gray, J. A. (1988b). Latent inhibition and ‘psychotic proneness’ in normal subjects. Personality and Individual Differences, 9, 777–783. Bassareo, V., & Di Chiara, G. (1999). Differential responsiveness of dopamine transmission to food-stimuli in nucleus accumbens shell/core compartments. Neuroscience, 89, 637–641.
What the brain teaches us about latent inhibition (LI)
399
Baxter, M. G., Parker, A., Lindner, C. C., Izquierdo, A. D., & Murray, E. A. (2000). Control of response selection by reinforcer value requires interaction of amygdala and orbital prefrontal cortex. The Journal of Neuroscience, 20, 4311–4319. Beckmann, H., & Jakob, H. (1991). Prenatal disturbances of nerve cell migration in the entorhinal region: a common vulnerability factor in functional psychoses? Journal of Neural Transmission. General Section, 84, 155–164. Berendse, H. W., Groenewegen, H. J., & Lohman, A. H. (1992). Compartmental distribution of ventral striatal neurons projecting to the mesencephalon in the rat. The Journal of Neuroscience, 12, 2079–2103. Birrell, J. M., & Brown, V. J. (2000). Medial frontal cortex mediates perceptual attentional set shifting in the rat. The Journal of Neuroscience, 20, 4320–4324. Bleuler, M. (1911). Dementia Praecox or the Group of Schizophrenias. New York: International Universities Press. Bogerts, B. (1991). The neuropathology of schizophrenia: pathophysiological and neurodevelopmental implications. In S. A. Mednick, T. D. Cannon, C. E. Barr & M. Lyon (Eds.), Fetal Neural Development and Adult Schizophrenia. Cambridge: Cambridge University Press, pp. 153–173. Bogerts, B. (1993). Recent advances in the neuropathology of schizophrenia. Schizophrenia Bulletin, 19, 431–445. Bogerts, B., Meertz, E., & Schonfeldt-Bausch, R. (1985). Basal ganglia and limbic system pathology in schizophrenia. A morphometric study of brain volume and shrinkage. Archives of General Psychiatry, 42, 784–791. Bouton, M. E. (1993). Context, time, and memory retrieval in the interference paradigms of Pavlovian learning. Psychological Bulletin, 114, 80–99. Braff, D. L., & Saccuzzo, D. P. (1982). Effect of antipsychotic medication on speed of information processing in schizophrenic patients. American Journal of Psychiatry, 139, 1127–1130. Braunstein-Bercovitz, H., & Lubow, R. E. (1998). Are high-schizotypal normal participants distractible or limited in attentional resources? A study of latent inhibition as a function of masking task load and schizotypy level. Journal of Abnormal Psychology, 107, 659–670. Broen, W. E. (1968). Schizophrenia: Research and Theory. New York: Academic Press. Broersen, L. M., Feldon, J., & Weiner, I. (1999). Dissociative effects of apomorphine infusions into the medial prefrontal cortex of rats on latent inhibition, prepulse inhibition and amphetamine-induced locomotion. Neuroscience, 94, 39–46. Broersen, L. M., Heinsbroek, R. P., de Bruin, J. P., & Olivier, B. (1996). Effects of local application of dopaminergic drugs into the medial prefrontal cortex of rats on latent inhibition. Biological Psychiatry, 40, 1083–1090. Brown, V. J., & Bowman, E. M. (2002). Rodent models of prefrontal cortical function. Trends in Neurosciences, 25, 340–343. Buhusi, C. V., Gray, J. A., & Schmajuk, N. A. (1998). Perplexing effects of hippocampal lesions on latent inhibition: a neural network solution. Behavioral Neuroscience, 112, 316–351. Cadoni, C., Solinas, M., & Di Chiara, G. (2000). Psychostimulant sensitization: differential changes in accumbal shell and core dopamine. European Journal of Pharmacology, 388, 69–76. Cador, M., Robbins, T. W., Everitt, B. J., et al. (1991). Limbic-striatal interactions in reward-related processes: modulation by the dopaminergic system.
400
Ina Weiner
In P. Willner & J. Scheel-Kruger (Eds.), The Mesolimbic Dopamine System: From Motivation to Action. Chichester: John Wiley, pp. 225–250. Cardinal, R. N., Parkinson, J. A., Hall, J., & Everitt, B. J. (2002). Emotion and motivation: the role of the amygdala, ventral striatum, and prefrontal cortex. Neuroscience & Biobehavioral Reviews, 26, 321–352. Carlsson, M., & Carlsson, A. (1990). Schizophrenia: a subcortical neurotransmitter imbalance syndrome? Schizophrenia Bulletin, 16, 425–432. Cassaday, H. J., Mitchell, S. N., Williams, J. H., & Gray, J. A. (1993). 5,7-Dihydroxytryptamine lesions in the fornix-fimbria attenuate latent inhibition. Behavioral and Neural Biology, 59, 194–207. Christison, G. W., Atwater, G. E., Dunn, L. A., & Kilts, C. D. (1988). Haloperidol enhancement of latent inhibition: relation to therapeutic action? Biological Psychiatry, 23, 746–749. Clark, A. J., Feldon, J., & Rawlins, J. N. (1992). Aspiration lesions of rat ventral hippocampus disinhibit responding in conditioned suppression or extinction, but spare latent inhibition and the partial reinforcement extinction effect. Neuroscience, 48, 821–829. Cools, A. R., Coolen, J. M., Smit, J. C., & Ellenbroek, B. A. (1984). The striatonigro-collicular pathway and explosive running behaviour: functional interaction between neostriatal dopamine and collicular GABA. European Journal of Pharmacology, 100, 71–77. Cornblatt, B., & Erlenmeyer-Kimling, L. (1984). Early attentional predictors of adolescent behavioral disturbances in children at risk for schizophrenia. In N. F. Watt, E. F. Anthony, L. C. Wynne, & J. E. Rolf (Eds.), Children at Risk for Schizophrenia: A Longitudinal Perspective. New York: Cambridge University Press. Cornblatt, B. A., Lenzenweger, M. F., Dworkin, R. H., & Erlenmeyer-Kimling, L. (1985). Positive and negative schizophrenic symptoms, attention, and information processing. Schizophrenia Bulletin, 11, 397–408. Cornblatt, B., Winters, L., & Erlenmeyer-Kimling, L. (1989). Attentional markers of schizophrenia: evidence from the New York high-risk study. In S. C. Schulz & C. A. Tamminga (Eds.), Schizophrenia: Scientific Progress. New York: Oxford University Press, pp. 83–92. Coutureau, E., Blundell, P. J., & Killcross, S. (2001). Basolateral amygdala lesions disrupt latent inhibition in rats. Brain Research Bulletin, 56, 49–53. Coutureau, E., Galani, R., Gosselin, O., Majchrzak, M., & Di Scala, G. (1999). Entorhinal but not hippocampal or subicular lesions disrupt latent inhibition in rats. Neurobiology of Learning and Memory, 72, 143–157. Coutureau, E., Gosselin, O., & Di Scala, G. (2000). Restoration of latent inhibition by olanzapine but not haloperidol in entorhinal cortex-lesioned rats. Psychopharmacology (Berl.), 150, 226–232. Coutureau, E., Lena, I., Dauge, V., & Di, S. G. (2002). The entorhinal cortex-nucleus accumbens pathway and latent inhibition: a behavioral and neurochemical study in rats. Behavioral Neuroscience, 116, 95–104. Csernansky, J. G., Murphy, G. M., & Faustman, W. O. (1991). Limbic/mesolimbic connections and the pathogenesis of schizophrenia. Biological Psychiatry, 30, 383–400. de Bruin, J. P., Sanchez-Santed, F., Heinsbroek, R. P., Donker, A., & Postmes, P. (1994). A behavioural analysis of rats with damage to the medial prefrontal
What the brain teaches us about latent inhibition (LI)
401
cortex using the Morris water maze: evidence for behavioural flexibility, but not for impaired spatial navigation. Brain Research, 652, 323–333. De la Casa, L. G., & Lubow, R. E. (1994). Memory for attended and nominally unattended stimuli in low and high psychotic-prone normal subjects: the effects of test-anticipation. Personality and Individual Differences, 17, 783–789. De la Casa, L. G., Ruiz, G., & Lubow, R. E. (1993). Latent inhibition and recall/ recognition of irrelevant stimuli as a function of pre-exposure duration in high and low psychotic-prone normal subjects. British Journal of Psychology, 84, 119–132. Della Casa, V., Hofer, I., Weiner, I., & Feldon, J. (1999). Effects of smoking status and schizotypy on latent inhibition. Journal of Psychopharmacology, 13, 45–57. Deutch, A. Y. (1993). Prefrontal cortical dopamine systems and the elaboration of functional corticostriatal circuits: implications for schizophrenia and Parkinson’s disease. Journal of Neural Transmission. General Section, 91, 197–221. Deutch, A. Y., & Cameron, D. S. (1992). Pharmacological characterization of dopamine systems in the nucleus accumbens core and shell. Neuroscience, 46, 49–56. Di Chiara, G. (1995). The role of dopamine in drug abuse viewed from the perspective of its role in motivation. Drug and Alcohol Dependence, 38, 95–137. Dias, R., & Aggleton, J. P. (2000). Effects of selective excitotoxic prefrontal lesions on acquisition of nonmatching- and matching-to-place in the T-maze in the rat: differential involvement of the prelimbic-infralimbic and anterior cingulate cortices in providing behavioural flexibility. The European Journal of Neuroscience, 12, 4457–4466. Dias, R., Robbins, T. W., & Roberts, A. C. (1996). Dissociation in prefrontal cortex of affective and attentional shifts. Nature, 380, 69–72. Dias, R., Robbins, T. W., & Roberts, A. C. (1997). Dissociable forms of inhibitory control within prefrontal cortex with an analog of the Wisconsin Card Sort Test: restriction to novel situations and independence from “on-line” processing. The Journal of Neuroscience, 17, 9285–9297. Dunn, L. A., Atwater, G. E., & Kilts, C. D. (1993). Effects of antipsychotic drugs on latent inhibition: sensitivity and specificity of an animal behavioral model of clinical drug action. Psychopharmacology (Berl.), 112, 315–323. Ellenbroek, B. A., Budde, S., & Cools, A. R. (1996). Prepulse inhibition and latent inhibition: the role of dopamine in the medial prefrontal cortex. Neuroscience, 75, 535–542. Ellenbroek, B. A., & Cools, A. R. (1990). Animal models with construct validity for schizophrenia. Behavioural Pharmacology, 1, 469–490. Ellenbroek, B. A., Knobbout, D. A., & Cools, A. R. (1997). The role of mesolimbic and nigrostriatal dopamine in latent inhibition as measured with the conditioned taste aversion paradigm. Psychopharmacology (Berl.), 129, 112–120. Everitt, B., & Robbins, T. W. (1992). Amygdala-ventral striatal interactions and reward-related processes. In J. P. Aggleton (Ed.), The Amygdala. Neurobiological Aspects of Emotion, Memory and Mental Dysfunction. Oxford: John Wiley and Sons, pp. 401–429. Fanselow, M. S., & LeDoux, J. E. (1999). Why we think plasticity underlying Pavlovian fear conditioning occurs in the basolateral amygdala. Neuron, 23, 229–232.
402
Ina Weiner
Feldon, J., & Weiner, I. (1991). Effects of haloperidol on the multitrial partial reinforcement extinction effect (PREE): evidence for neuroleptic drug action on nonreinforcement but not on reinforcement. Psychopharmacology (Berl.), 105, 407–414. Fendt, M., & Fanselow, M. S. (1999). The neuroanatomical and neurochemical basis of conditioned fear. Neuroscience & Biobehavioral Reviews, 23, 743–760. Floresco, S. B., Blaha, C. D., Yang, C. R., & Phillips, A. G. (2001). Modulation of hippocampal and amygdalar-evoked activity of nucleus accumbens neurons by dopamine: cellular mechanisms of input selection. The Journal of Neuroscience, 21, 2851–2860. Frey, P. W., & Sears, R. G. (1978). Model of conditioning incorporating the Rescorla-Wagner associative axiom, a dynamic attention process, and a catastrophe rule. Psychological Review, 85, 321–341. Frith, C. D. (1979). Consciousness, information processing and schizophrenia. The British Journal of Psychiatry, 134, 225–235. Gal, G. (2000). Disrupted and undisruptable latent inhibition following shell and core lesions: the dual role of the nucleus accumbens in latent inhibition. Tel Aviv University, Tel Aviv. Gal, G., Schiller, D., & Weiner, I. (2005). Latent inhibition is disrupted by nucleus accumbens shell lesion but is abnormally persistent following entire nucleus accumbens lesion: the neural site controlling the expression and disruption of the stimulus preexposure effect. Behavioural Brain Research, 162, 246–255. Gallagher, M., & Chiba, A. A. (1996). The amygdala and emotion. Current Opinion in Neurobiology, 6, 221–227. Gallagher, M., McMahan, R. W., & Schoenbaum, G. (1999). Orbitofrontal cortex and representation of incentive value in associative learning. The Journal of Neuroscience, 19, 6610–6614. Gallagher, M., & Schoenbaum, G. (1999). Functions of the amygdala and related forebrain areas in attention and cognition. Annals of the New York Academy of Sciences, 877, 397–411. Gelissen, M., & Cools, A. (1988). Effect of intracaudate haloperidol and apomorphine on switching motor patterns upon current behaviour of cats. Behavioural Brain Research, 29, 17–26. Gjerde, P. F. (1983). Attentional capacity dysfunction and arousal in schizophrenia. Psychological Bulletin, 93, 57–72. Gosselin, G., Oberling, P., & Di Scala, G. (1996). Antagonism of amphetamineinduced disruption of latent inhibition by the atypical antipsychotic olanzapine in rats. Behavioural Pharmacology, 7, 820–826. Grace, A. A. (1991). Phasic versus tonic dopamine release and the modulation of dopamine system responsivity: a hypothesis for the etiology of schizophrenia. Neuroscience, 41, 1–24. Grace, A. A. (2000). Gating of information flow within the limbic system and the pathophysiology of schizophrenia. Brain Research. Brain Research Reviews, 31, 330–341. Gray, J. A., Feldon, J., Rawlins, J. N. P., Hemsley, D. R., & Smith, A. D. (1991). The neuropsychology of schizophrenia. Behavioral and Brain Sciences, 14, 1–20. Gray, J. A., Joseph, M. H., Hemsley, D. R., et al. (1995a). The role of mesolimbic dopaminergic and retrohippocampal afferents to the nucleus accumbens in latent inhibition: implications for schizophrenia. Behavioural Brain Research, 71, 19–31.
What the brain teaches us about latent inhibition (LI)
403
Gray, J. A., Moran, P. M., Grigoryan, G., et al. (1997). Latent inhibition: the nucleus accumbens connection revisited. Behavioural Brain Research, 88, 27–34. Gray, N. S., Hemsley, D. R., & Gray, J. A. (1992a). Abolition of latent inhibition in acute, but not chronic schizophrenics. Neurology, Psychiatry and Brain Research, 1, 83–89. Gray, N. S., Pickering, A. D., Hemsley, D. R., Dawling, S., & Gray, J. A. (1992b). Abolition of latent inhibition by a single 5 mg dose of d-amphetamine in man. Psychopharmacology (Berl.), 107, 425–430. Gray, N. S., Pilowsky, L. S., Gray, J. A., & Kerwin, R. W. (1995b). Latent inhibition in drug naive schizophrenics: relationship to duration of illness and dopamine D2 binding using SPET. Schizophrenia Research, 17, 95–107. Groenewegen, H. J., Berendse, H. W., Meredith, G. E., et al. (1991). Functional anatomy of the ventral, limbic system-innervated striatum. In P. Willner & J. Scheel-Kruger (Eds.), The Mesolimbic Dopamine System: From Motivation to Action. Chichester: John Wiley, pp. 19–60. Groenewegen, H. J., Vermeulen-Van der Zee, E., te Kortschot, A., & Witter, M. P. (1987). Organization of the projections from the subiculum to the ventral striatum in the rat. A study using anterograde transport of Phaseolus vulgaris leucoagglutinin. Neuroscience, 23, 103–120. Groenewegen, H. J., Wright, C. I., & Beijer, A. V. (1996). The nucleus accumbens: gateway for limbic structures to reach the motor system? Progress in Brain Research, 107, 485–511. Groenewegen, H. J., Wright, C. I., Beijer, A. V., & Voorn, P. (1999). Convergence and segregation of ventral striatal inputs and outputs. Annals of the New York Academy of Sciences, 877, 49–63. Hall, G., & Channell, S. (1985). Latent inhibition and conditioning after pre-exposure to the training context. Learning and Motivation, 16, 381–397. Hall, R. C., Popkin, M. K., Beresford, T. P., & Hall, A. K. (1988). Amphetamine psychosis: Clinical presentations and differential diagnosis. Psychiatric Medicine, 6, 73–79. Han, J. S., Gallagher, M., & Holland, P. (1995). Hippocampal lesions disrupt decrements but not increments in conditioned stimulus processing. The Journal of Neuroscience, 15, 7323–7329. Harrison, P. J. (1999). The neuropathology of schizophrenia. A critical review of the data and their interpretation. Brain, 122, 593–624. Hatfield, T., Han, J. S., Conley, M., Gallagher, M., & Holland, P. (1996). Neurotoxic lesions of basolateral, but not central, amygdala interfere with Pavlovian second-order conditioning and reinforcer devaluation effects. The Journal of Neuroscience, 16, 5256–5265. Heimer, L., Alheid, G. F., de Olmos, J. S., et al. (1997). The accumbens: Beyond the core-shell dichotomy. The Journal of Neuropsychiatry and Clinical Neurosciences, 9, 354–381. Hemsley, D. R. (1993). A simple (or simplistic?) cognitive model for schizophrenia. Behaviour Research and Therapy, 31, 633–645. Hemsley, D. R. (1994). Cognitive disturbance as the link between schizophrenic symptoms and their biological bases. Neurology Psychiatry and Brain Research, 2, 163–170. Holland, P. C., & Gallagher, M. (1999). Amygdala circuitry in attentional and representational processes. Trends in Cognitive Sciences, 3, 65–73.
404
Ina Weiner
Holt, W., & Maren, S. (1999). Muscimol inactivation of the dorsal hippocampus impairs contextual retrieval of fear memory. The Journal of Neuroscience, 19, 9054–9062. Honey, R. C., & Good, M. (1993). Selective hippocampal lesions abolish the contextual specificity of latent inhibition and conditioning. Behavioral Neuroscience, 107, 23–33. Howland, J. G., Taepavarapruk, P., & Phillips, A. G. (2002). Glutamate receptordependent modulation of dopamine efflux in the nucleus accumbens by basolateral, but not central, nucleus of the amygdala in rats. The Journal of Neuroscience, 22, 1137–1145. Hutcheson, D. M., Parkinson, J. A., Robbins, T. W., & Everitt, B. J. (2001). The effects of nucleus accumbens core and shell lesions on intravenous heroin self-administration and the acquisition of drug-seeking behaviour under a second-order schedule of heroin reinforcement. Psychopharmacology (Berl.), 153, 464–472. Ikemoto, S., & Panksepp, J. (1999). The role of nucleus accumbens dopamine in motivated behavior: a unifying interpretation with special reference to rewardseeking. Brain Research. Brain Research Reviews, 31, 6–41. Ito, R., Dalley, J. W., Howes, S. R., Robbins, T. W., & Everitt, B. J. (2000). Dissociation in conditioned dopamine release in the nucleus accumbens core and shell in response to cocaine cues and during cocaine-seeking behavior in rats. The Journal of Neuroscience, 20, 7489–7495. Jakob, H., & Beckmann, H. (1986). Prenatal developmental disturbances in the limbic allocortex in schizophrenics. Journal of Neural Transmission, 65, 303–326. Janowsky, D. S., & Davis, J. M. (1976). Methylphenidate, dextroamphetamine, and levamfetamine. Effects on schizophrenic symptoms. Archives of General Psychiatry, 33, 304–308. Jeanblanc, J., Hoeltzel, A., & Louilot, A. (2003). Differential involvement of dopamine in the anterior part and posterior parts of the striatum in latent inhibition. Neuroscience, 118, 233–241. Jeanblanc, J., Hoeltzel, A., & Louilot, A. (2002). Dissociation in the involvement of dopaminergic neurons innervating the core and shell subregions of the nucleus accumbens in latent inhibition and affective perception. Neuroscience, 111, 315–323. Jeanblanc, J., Peterschmitt, Y., Hoeltzel, A., & Louilot, A. (2004). Influence of the entorhinal cortex on accumbal and striatal dopaminergic responses in a latent inhibition paradigm. Neuroscience, 128, 187–200. Joel, D., Weiner, I., & Feldon, J. (1997). Electrolytic lesions of the medial prefrontal cortex in rats disrupt performance on an analog of the Wisconsin Card Sorting Test, but do not disrupt latent inhibition: Implications for animal models of schizophrenia. Behavioural Brain Research, 85, 187–201. Jongen-Relo, A. L., Kaufmann, S., & Feldon, J. (2002). A differential involvement of the shell and core subterritories of the nucleus accumbens of rats in attentional processes. Neuroscience, 111, 95–109. Joseph, M. H., Peters, S. L., Moran, P. M., et al. (2000). Modulation of latent inhibition in the rat by altered dopamine transmission in the nucleus accumbens at the time of conditioning. Neuroscience, 101, 921–930.
What the brain teaches us about latent inhibition (LI)
405
Kapur, S. (2003). Psychosis as a state of aberrant salience: a framework linking biology, phenomenology, and pharmacology in schizophrenia. The American Journal of Psychiatry, 160, 13–23. Kapur, S., Mizrahi, R., & Li, M. (2005). From dopamine to salience to psychosis – linking biology, pharmacology and phenomenology of psychosis. Schizophrenia Research, 79, 59–68. Kaye, H., & Pearce, J. M. (1987a). Hippocampal lesions attenuate latent inhibition and the decline of the orienting response in rats. The Quarterly Journal of Experimental Psychology. B, Comparative and Physiological Psychology, 39, 107–125. Kaye, H., & Pearce, J. M. (1987b). Hippocampal lesions attenuate latent inhibition of a CS and of a neutral stimulus. Psychobiology, 15, 293–299. Kelley, A. E., Smith-Roe, S. L., & Holahan, M. R. (1997). Response-reinforcement learning is dependent on N-methyl-D-aspartate receptor activation in the nucleus accumbens core. Proceedings of the National Academy of Sciences of the United States of America, 94, 12174–12179. Kesner, R. P. (2000). Subregional analysis of mnemonic functions of the prefrontal cortex in the rat. Psychobiology, 28, 219–228. Killcross, A. S., Dickinson, A., & Robbins, T. W. (1994a). Amphetamine-induced disruptions of latent inhibition are reinforcer mediated: implications for animal models of schizophrenic attentional dysfunction. Psychopharmacology (Berl.), 115, 185–195. Killcross, A. S., Dickinson, A., & Robbins, T. W. (1994b). Effects of the neuroleptic alpha-flupenthixol on latent inhibition in aversively- and appetitively-motivated paradigms: evidence for dopamine-reinforcer interactions. Psychopharmacology (Berl.), 115, 196–205. Killcross, A. S., & Robbins, T. W. (1993). Differential effects of intra-accumbens and systemic amphetamine on latent inhibition using an on-baseline, within-subject conditioned suppression paradigm. Psychopharmacology (Berl.), 110, 479–489. Kinon, B. J., & Lieberman, J. A. (1996). Mechanisms of action of atypical antipsychotic drugs: a critical analysis. Psychopharmacology (Berl.), 124, 2–34. Kolb, B. (1984). Functions of the frontal cortex of the rat: a comparative review. Brain Research, 320, 65–98. Konstandi, M., & Kafetzopoulos, E. (1993). Effects of striatal or accumbens lesions on the amphetamine-induced abolition of latent inhibition. Pharmacology, Biochemistry, and Behavior, 44, 751–754. Koob, G. F., Riley, S. J., Smith, S. C., & Robbins, T. W. (1978). Effects of 6-hydroxydopamine lesions of the nucleus accumbens septi and olfactory tubercle on feeding, locomotor activity, and amphetamine anorexia in the rat. Journal of Comparative and Physiological Psychology, 92, 917–927. Kornetsky, C. (1972). The use of a simple test of attention as a measure of drug effects in schizophrenic patients. Psychopharmacologia, 24, 99–106. Kovelman, J. A., & Scheibel, A. B. (1984). A neurohistological correlate of schizophrenia. Biological Psychiatry, 19, 1601–1621. Kraepeline, E. (1919). Dementia Praecox and Paraphrenia. New York: Kreiger. Krystal, J. H., D’Souza, D. C., Mathalon, D., et al. (2003). NMDA receptor antagonist effects, cortical glutamatergic function, and schizophrenia: toward a paradigm shift in medication development. Psychopharmacology (Berl.), 169, 215–233.
406
Ina Weiner
Lacroix, L., Broersen, L. M., Feldon, J., & Weiner, I. (2000). Effects of local infusions of dopaminergic drugs into the medial prefrontal cortex of rats on latent inhibition, prepulse inhibition and amphetamine induced activity. Behavioural Brain Research, 107, 111–121. Lacroix, L., Broersen, L. M., Weiner, I., & Feldon, J. (1998). The effects of excitotoxic lesion of the medial prefrontal cortex on latent inhibition, prepulse inhibition, food hoarding, elevated plus maze, active avoidance and locomotor activity in the rat. Neuroscience, 84, 431–442. Lacroix, L., Spinelli, S., White, W., & Feldon, J. (2000). The effects of ibotenic acid lesions of the medial and lateral prefrontal cortex on latent inhibition, prepulse inhibition and amphetamine-induced hyperlocomotion. Neuroscience, 97, 459–468. Le Moal, M., & Simon, H. (1991). Mesocorticolimbic dopaminergic network: Functional and regulatory roles. Physiological Reviews, 71, 155–234. LeDoux, J. E. (1992). Brain mechanisms of emotion and emotional learning. Current Opinion in Neurobiology, 2, 191–197. Lewis, M. C., & Gould, T. J. (2007a). Reversible inactivation of the entorhinal cortex disrupts the establishment and expression of latent inhibition of cued fear conditioning in C57BL/6 mice. Hippocampus, 17, 462–470. Lewis, M. C., & Gould, T. J. (2007b). Signal transduction mechanisms within the entorhinal cortex that support latent inhibition of cued fear conditioning. Neurobiology of Learning and Memory, 88, 359–368. Lieberman, J. A., Kane, J. M., & Alvir, J. (1987). Provocative tests with psychostimulant drugs in schizophrenia. Psychopharmacology (Berl.), 91, 415–433. Lieberman, J. A., Kane, J. M., Gadaleta, D., et al. (1984). Methylphenidate challenge as a predictor of relapse in schizophrenia. The American Journal of Psychiatry, 141, 633–638. Lipp, O. V., & Vaitl, D. (1992). Latent inhibition in human Pavlovian differential conditioning – effect of additional stimulation after preexposure and relation to schizotypal traits. Personality and Individual Differences, 13, 1003–1012. Louilot, A., & Besson, C. (2000). Specificity of amygdalostriatal interactions in the involvement of mesencephalic dopaminergic neurons in affective perception. Neuroscience, 96, 73–82. Lubow, R. E. (1973). Latent inhibition. Psychological Bulletin, 79, 398–407. Lubow, R. E. (1989). Latent Inhibition and Conditioned Attention Theory. New York: Cambridge University Press. Lubow, R. E. (2005). Construct validity of the animal latent inhibition model of selective attention deficits in schizophrenia. Schizophrenia Bulletin, 31, 139–153. Lubow, R. E., & De la Casa, G. (2002). Latent inhibition as a function of schizotypality and gender: implications for schizophrenia. Biological Psychiatry, 59, 69–86. Lubow, R. E., & Gewirtz, J. C. (1995). Latent inhibition in humans: data, theory, and implications for schizophrenia. Psychological Bulletin. 117(1), 87–103. Lubow, R. E., Ingberg-Sachs, Y., Zalstein-Orda, N., & Gewirtz, J. C. (1992). Latent inhibition in low and high “psychotic-prone” normal subjects. Personality and Individual Differences, 13, 563–572. Lubow, R. E., Kaplan, O., Abramovich, P., Rudnick, A., & Laor, N. (2000). Visual search in schizophrenia: latent inhibition and novel pop-out effects. Schizophrenia Research, 45, 145–156.
What the brain teaches us about latent inhibition (LI)
407
Lubow, R. E., Kaplan, O., & De la Casa, G. (2001). Performance on the visual search analog of latent inhibition is modulated by an interaction between schizotypy and gender. Schizophrenia Research, 52, 275–287. Lubow, R. E., Weiner, I., Schlossberg, A., & Baruch, I. (1987). Latent inhibition and schizophrenia. Bulletin of Psychonomic Society, 25, 464–467. Lubow, R. E., Weiner, I., & Schnur, P. (1981). Conditioned attention theory. In G. H. Bower (Ed.), The Psychology of Learning and Motivation, vol. 15. New York: Academic Press, pp. 1–49. Lyon, M. (1991). Animal models of mania and schizophrenia. In P. Willner (Ed.), Behavioral Models in Psychopharmacology: Theoretical, Industrial and Clinical Perspectives. Cambridge: Cambridge University Press, pp. 253–310. Mackintosh, N. J. (1975). A theory of attention: variations in the associability of stimuli with reinforcement. Psychological Review, 82, 276–298. Magaro, P. A. (1980). Cognition in Schizophrenia and Paranoia: The Integration of Cognitive Processes. Hillsdale, NJ: Erlbaum. Maher, B. A., Manschreck, T. C., & Molino, M. A. (1983). Redundancy, pause distributions and thought disorder in schizophrenia. Language and Speech, 26, 191–199. Maldonado-Irizarry, C. S., & Kelley, A. E. (1994). Differential behavioral effects following microinjection of an NMDA antagonist into nucleus accumbens subregions. Psychopharmacology (Berl.), 116, 65–72. Maldonado-Irizarry, C. S., & Kelley, A. E. (1995). Excitotoxic lesions of the core and shell subregions of the nucleus accumbens differentially disrupt body weight regulation and motor activity in rat. Brain Research Bulletin, 38, 551–559. Malkova, L., Gaffan, D., & Murray, E. A. (1997). Excitotoxic lesions of the amygdala fail to produce impairment in visual learning for auditory secondary reinforcement but interfere with reinforcer devaluation effects in rhesus monkeys. The Journal of Neuroscience, 17, 6011–6020. Marcus, M. M., Nomikos, G. G., & Svensson, T. H. (2000). Effects of atypical antipsychotic drugs on dopamine output in the shell and core of the nucleus accumbens: role of 5-HT(2A) and alpha(1)-adrenoceptor antagonism. European Neuropsychopharmacology, 10, 245–253. Maren, S., & Holt, W. (2000). The hippocampus and contextual memory retrieval in Pavlovian conditioning. Behavioural Brain Research, 110, 97–108. McAlonan, K., & Brown, V. J. (2003). Orbital prefrontal cortex mediates reversal learning and not attentional set shifting in the rat. Behavioural Brain Research, 146, 97–103. McCartan, D., Bell, R., Green, J. F., et al. (2001). The differential effects of chlorpromazine and haloperidol on latent inhibition in healthy volunteers. Journal of Psychopharmacology, 15, 96–104. McGhie, A., & Chapman, J. (1961). Disorders of attention and perception in early schizophrenia. The British Journal of Medical Psychology, 34, 103–116. Meltzer, H. Y., & Stahl, S. M. (1976). The dopamine hypothesis of schizophrenia: a review. Schizophrenia Bulletin, 2, 19–76. Mizumori, S. J., Yeshenko, O., Gill, K. M., & Davis, D. M. (2004). Parallel processing across neural systems: implications for a multiple memory system hypothesis. Neurobiology of Learning and Memory, 82, 278–298. Moghaddam, B., Adams, B., Verma, A., & Daly, D. (1997). Activation of glutamatergic neurotransmission by ketamine: a novel step in the pathway
408
Ina Weiner
from NMDA receptor blockade to dopaminergic and cognitive disruptions associated with the prefrontal cortex. The Journal of Neuroscience, 17, 2921–2927. Moore, H., Fadel, J., Sarter, M., & Bruno, J. P. (1999). Role of accumbens and cortical dopamine receptors in the regulation of cortical acetylcholine release. Neuroscience, 88, 811–822. Moore, J. W., & Stickney, K. J. (1980). Formation of attentional-associative networks in real time: role of the hippocampus and implications for conditioning. Physiological Psychology, 8, 207–217. Moran, P. M., Fischer, T. R., Hitchcock, J. M., & Moser, P. C. (1996). Effects of clozapine on latent inhibition in the rat. Behavioural Pharmacology, 7, 42–48. Morgan, M. A., & LeDoux, J. E. (1995). Differential contribution of dorsal and ventral medial prefrontal cortex to the acquisition and extinction of conditioned fear in rats. Behavioral Neuroscience, 109, 681–688. Moser, P. C., Hitchcock, J. M., Lister, S., & Moran, P. M. (2000). The pharmacology of latent inhibition as an animal model of schizophrenia. Brain Research. Brain Research Reviews, 33, 275–307. Mulder, A. B., Hodenpijl, M. G., & Lopes da Silva, F. H. (1998). Electrophysiology of the hippocampal and amygdaloid projections to the nucleus accumbens of the rat: convergence, segregation, and interaction of inputs. The Journal of Neuroscience, 18, 5095–5102. Murphy, C. A., Pezze, M., Feldon, J., & Heidbreder, C. (2000). Differential involvement of dopamine in the shell and core of the nucleus accumbens in the expression of latent inhibition to an aversively conditioned stimulus. Neuroscience, 97, 469–477. Nuechterlein, K. H., & Dawson, M. E. (1984). Information processing and attentional functioning in the developmental course of schizophrenic disorders. Schizophrenia Bulletin, 10, 160–203. O’Donnell, P., & Grace, A. A. (1998). Dysfunctions in multiple interrelated systems as the neurobiological bases of schizophrenic symptom clusters. Schizophrenia Bulletin, 24, 267–283. Oades, R. D. (1982). Attention and Schizophrenia: Neurobiological Bases. London: Pitman. Oades, R. D. (1985). The role of noradrenaline in tuning and dopamine in switching between signals in the CNS. Neuroscience & Biobehavioral Reviews, 9, 261–282. Oswald, C. J., Yee, B. K., Rawlins, J. N., et al. (2002). The influence of selective lesions to components of the hippocampal system on the orienting [correction of orientating] response, habituation and latent inhibition. European Journal of Neuroscience, 15, 1983–1990. Otake, K., & Nakamura, Y. (2000). Possible pathways through which neurons of the shell of the nucleus accumbens influence the outflow of the core of the nucleus accumbens. Brain and Development, 22, S17–26. Parkinson, J. A., Cardinal, R. N., & Everitt, B. J. (2000). Limbic cortical-ventral striatal systems underlying appetitive conditioning. Progress in Brain Research, 126, 263–285. Parkinson, J. A., Olmstead, M. C., Burns, L. H., Robbins, T. W., & Everitt, B. J. (1999). Dissociation in effects of lesions of the nucleus accumbens core and shell on appetitive pavlovian approach behavior and the potentiation of conditioned reinforcement and locomotor activity by D-amphetamine. The Journal of Neuroscience, 19, 2401–2411.
What the brain teaches us about latent inhibition (LI)
409
Payne, R. W. (1966). The measurement and significance of overinclusive thinking and retardation in schizophrenic patients. In P. Hoch & J. Zubin (Eds.), Psychopathology of Schizophrenia. New York: Grune and Stratton. Pearce, J. M., & Hall, G. (1980). A model for Pavlovian learning: variations in the effectiveness of conditioned but not of unconditioned stimuli. Psychological Review, 87, 532–552. Pennartz, C. M., Groenewegen, H. J., & Lopes da Silva, F. H. (1994). The nucleus accumbens as a complex of functionally distinct neuronal ensembles: an integration of behavioural, electrophysiological and anatomical data. Progress in Neurobiology, 42, 719–761. Peters, S. L., & Joseph, M. H. (1993). Haloperidol potentiation of latent inhibition in rats: evidence for a critical role at conditioning rather than pre-exposure. Behavioural Pharmacology, 4, 183–186. Peterschmitt, Y., Hoeltzel, A., & Louilot, A. (2005). Striatal dopaminergic responses observed in latent inhibition are dependent on the hippocampal ventral subicular region. The European Journal of Neuroscience, 22, 2059–2068. Peterschmitt, Y., Meyer, F., & Louilot, A. (2008). Differential influence of the ventral subiculum on dopaminergic responses observed in core and dorsomedial shell subregions of the nucleus accumbens in latent inhibition. Neuroscience, 154, 898–910. Phillips, A. G., Ahn, S., & Howland, J. G. (2003). Amygdalar control of the mesocorticolimbic dopamine system: parallel pathways to motivated behavior. Neuroscience & Biobehavioral Reviews, 27, 543–554. Pickens, C. L., Saddoris, M. P., Setlow, B., et al. (2003). Different roles for orbitofrontal cortex and basolateral amygdala in a reinforcer devaluation task. The Journal of Neuroscience, 23, 11078–11084. Pouzet, B., Veenman, C. L., Yee, B. K., Feldon, J., & Weiner, I. (1999). The effects of radiofrequency lesion or transection of the fimbria-fornix on latent inhibition in the rat. Neuroscience, 91, 1355–1368. Pouzet, B., Zhang, W. N., Weiner, I., Feldon, J., & Yee, B. K. (2004). Latent inhibition is spared by N-methyl-D-aspartate (NMDA)-induced ventral hippocampal lesions, but is attenuated following local activation of the ventral hippocampus by intracerebral NMDA infusion. Neuroscience, 124, 183–194. Purves, D., Bonardi, C., & Hall, G. (1995). Enhancement of latent inhibition in rats with electrolytic lesions of the hippocampus. Behavioral Neuroscience, 109, 366–370. Ragozzino, M. E., Detrick, S., & Kesner, R. P. (1999). Involvement of the prelimbicinfralimbic areas of the rodent prefrontal cortex in behavioral flexibility for place and response learning. The Journal of Neuroscience, 19, 4585–4594. Ragozzino, M. E., Wilcox, C., Raso, M., & Kesner, R. P. (1999). Involvement of rodent prefrontal cortex subregions in strategy switching. Behavioral Neuroscience, 113, 32–41. Rappaport, M., Silverman, J., Hopkins, H. K., & Hall, K. (1971). Phenothiazine effects on auditory signal detection in paranoid and nonparanoid schizophrenics. Science, 174, 723–725. Rascle, C., Mazas, O., Vaiva, G., et al. (2001). Clinical features of latent inhibition in schizophrenia. Schizophrenia Research, 51, 149–161. Redgrave, P., Prescott, T. J., & Gurney, K. (1999). The basal ganglia: a vertebrate solution to the selection problem? Neuroscience, 89, 1009–1023.
410
Ina Weiner
Reilly, S., Harley, C., & Revusky, S. (1993). Ibotenate lesions of the hippocampus enhance latent inhibition in conditioned taste aversion and increase resistance to extinction in conditioned taste preference. Behavioral Neuroscience, 107, 996–1004. Restivo, L., Passino, E., Middei, S., & Ammassari-Teule, M. (2002). The strainspecific involvement of nucleus accumbens in latent inhibition might depend on differences in processing configural- and cue-based information between C57BL/6 and DBA mice. Brain Research Bulletin, 57, 35–39. Robbins, T. W., Cador, M., Taylor, J. R., & Everitt, B. J. (1989). Limbic-striatal interactions in reward-related processes. Neuroscience & Biobehavioral Reviews, 13, 155–162. Robbins, T. W., & Everitt, B. J. (1982). Functional studies of the central catecholamines. International Review of Neurobiology, 23, 303–365. Robbins, T. W., & Everitt, B. J. (1996). Neurobehavioural mechanisms of reward and motivation. Current Opinion in Neurobiology, 6, 228–236. Rolls, E. T. (1996). The orbitofrontal cortex. Philosophical Transactions of the Royal Society of London. Series B, Biological Sciences, 351, 1433–1443; discussion 1443–1434. Rolls, E. T. (1999). The Brain and Emotion. Oxford: Oxford University Press. Rolls, E. T. (2000a). Neurophysiology and functions of the primate amygdala, and the neural basis of emotion. In J. P. Aggleton (Ed.), The Amygdala: A Functional Analysis. New York: Oxford University Press, pp. 447–478. Rolls, E. T. (2000b). The orbitofrontal cortex and reward. Cerebral Cortex, 10, 284–294. Romano, A. G. (1999). Variations in CS associability and multiple unit hippocampal activity in the rabbit. Behavioural Brain Research, 103, 163–173. Ruob, C., Elsner, J., Weiner, I., & Feldon, J. (1997). Amphetamine-induced disruption and haloperidol-induced potentiation of latent inhibition depend on the nature of the stimulus. Behavioural Brain Research, 88, 35–41. Salafia, W. R., & Allan, A. M. (1980). Attenuation of latent inhibition by electrical stimulation of hippocampus. Physiology and Behavior, 24, 1047–1051. Salafia, W. R., & Allan, A. M. (1982). Augmentation of latent inhibition by electrical stimulation of hippocampus. Physiology and Behavior, 29, 1125–1130. Salgado, J. V., Hetem, L. A., Vidal, M., et al. (2000). Reduction of latent inhibition by D-amphetamine in a conditioned suppression paradigm in humans. Behavioural Brain Research, 117, 61–67. Schauz, C., & Koch, M. (2000). Blockade of NMDA receptors in the amygdala prevents latent inhibition of fear-conditioning. Learning and Memory, 7, 393–399. Schiller, D., & Weiner, I. (2004). Lesions to the basolateral amygdala and the orbitofrontal cortex but not to the medial prefrontal cortex produce an abnormally persistent latent inhibition in rats. Neuroscience, 128, 15–25. Schiller, D., & Weiner, I. (2005). Basolateral amygdala lesions in the rat produce an abnormally persistent latent inhibition with weak preexposure but not with context shift. Behavioural Brain Research, 163, 115–121. Schmajuk, N. A., Buhusi, C. V., & Gray, J. A. (1998). Psychopharmacology of latent inhibition: a neural network approach. Behavioural Pharmacology, 9, 711–730. Schmajuk, N. A., Cox, L., & Gray, J. A. (2001). Nucleus accumbens, entorhinal cortex and latent inhibition: a neural network model. Behavioural Brain Research, 118, 123–141.
What the brain teaches us about latent inhibition (LI)
411
Schmajuk, N. A., Gray, J. A., & Lam, Y. W. (1996). Latent inhibition: a neural network approach. Journal of Experimental Psychology. Animal Behavior Processes, 22, 321–349. Schmajuk, N. A., Lam, Y. W., & Christiansen, B. A. (1994). Latent inhibition of the rat eyeblink response: effect of hippocampal aspiration lesions. Physiology and Behavior, 55, 597–601. Schmajuk, N. A., & Moore, J. W. (1985). Real-time attentional models for classical conditioning and the hippocampus. Physiological Psychology, 13, 278–290. Schmajuk, N. A., & Moore, J. W. (1988). The hippocampus and the classically conditioned nictitating membrane response: a realtime attentional-associative model. Psychobiology, 16, 20–35. Schoenbaum, G., Nugent, S. L., Saddoris, M. P., & Setlow, B. (2002). Orbitofrontal lesions in rats impair reversal but not acquisition of go, no-go odor discriminations. Neuroreport, 13, 885–890. Schoenbaum, G., & Setlow, B. (2001). Integrating orbitofrontal cortex into prefrontal theory: common processing themes across species and subdivisions. Learning and Memory, 8, 134–147. Schoenbaum, G., Setlow, B., Nugent, S. L., Saddoris, M. P., & Gallagher, M. (2003). Lesions of orbitofrontal cortex and basolateral amygdala complex disrupt acquisition of odor-guided discriminations and reversals. Learning and Memory, 10, 129–140. Schultz, W. (1998). Predictive reward signal of dopamine neurons. Journal of Neurophysiology, 80, 1–27. Seillier, A., Dieu, Y., Herbeaux, K., et al. (2007). Evidence for a critical role of entorhinal cortex at pre-exposure for latent inhibition disruption in rats. Hippocampus, 17, 220–226. Setlow, B., Gallagher, M., & Holland, P. C. (2002). The basolateral complex of the amygdala is necessary for acquisition but not expression of CS motivational value in appetitive Pavlovian second-order conditioning. European Journal of Neuroscience, 15, 1841–1853. Shadach, E., Feldon, J., & Weiner, I. (1999). Clozapine-induced potentiation of latent inhibition is due to its action in the conditioning stage: implications for the mechanism of action of antipsychotic drugs. The International Journal of Neuropsychopharmacology, 2, 283–291. Shakow, D. (1962). Segmental set. Archives of General Psychiatry, 6, 1–17. Shohamy, D., Allen, M. T., & Gluck, M. A. (2000). Dissociating entorhinal and hippocampal involvement in latent inhibition. Behavioral Neuroscience, 114, 867–874. Smith, A., Li, M., Becker, S., & Kapur, S. (2006). Dopamine, prediction error and associative learning: a model-based account. Network, 17, 61–84. Smith-Roe, S. L., & Kelley, A. E. (2000). Coincident activation of NMDA and dopamine D1 receptors within the nucleus accumbens core is required for appetitive instrumental learning. The Journal of Neuroscience, 20, 7737–7742. Snyder, S. H. (1973). Amphetamine psychosis: A “model” schizophrenia mediated by catecholamines. The American Journal of Psychiatry, 130, 61–67. Snyder, S. H. (1976). The dopamine hypothesis of schizophrenia: focus on the dopamine receptor. The American Journal of Psychiatry, 133, 197–202. Solomon, P. R., Crider, A., Winkelman, J. W., et al. (1981). Disrupted latent inhibition in the rat with chronic amphetamine or haloperidol-induced
412
Ina Weiner
supersensitivity: relationship to schizophrenic attention disorder. Biological Psychiatry, 16, 519–537. Solomon, P. R., & Moore, J. W. (1975). Latent inhibition and stimulus generalization of the classically conditioned nictitating membrane response in rabbits (Oryctolagus cuniculus) following dorsal hippocampal ablation. Journal of Comparative and Physiological Psychology, 89, 1192–1203. Solomon, P. R., & Staton, D. M. (1982). Differential effects of microinjections of d-amphetamine into the nucleus accumbens or the caudate putamen on the rat’s ability to ignore an irrelevant stimulus. Biological Psychiatry, 17, 743–756. Stahl, S. M. (1988). Basal ganglia neuropharmacology and obsessive-compulsive disorder: the obsessive-compulsive disorder hypothesis of basal ganglia dysfunction. Psychopharmacology Bulletin, 24, 370–374. Swanson, C. J., Heath, S., Stratford, T. R., & Kelley, A. E. (1997). Differential behavioral responses to dopaminergic stimulation of nucleus accumbens subregions in the rat. Pharmacology, Biochemistry, and Behavior, 58, 933–945. Swerdlow, N. R., Braff, D. L., Hartston, H., Perry, W., & Geyer, M. A. (1996). Latent inhibition in schizophrenia. Schizophrenia Research, 20, 91–103. Swerdlow, N. R., & Koob, G. F. (1987). Dopamine, schizophrenia, mania and depression: toward a unified hypothesis of cortico-striato-pallido-thalamic function. Behavioral and Brain Sciences, 10, 197–245. Swerdlow, N. R., Stephany, N., Wasserman, L. C., et al. (2003). Dopamine agonists disrupt visual latent inhibition in normal males using a within-subject paradigm. Psychopharmacology (Berl.), 169, 314–320. Taghzouti, K., Louilot, A., Herman, J. P., Le Moal, M., & Simon, H. (1985). Alternation behavior, spatial discrimination, and reversal disturbances following 6-hydroxydopamine lesions in the nucleus accumbens of the rat. Behavioral and Neural Biology, 44, 354–363. Taghzouti, K., Simon, H., Louilot, A., Herman, J. P., & Le Moal, M. (1985). Behavioral study after local injection of 6-hydroxydopamine into the nucleus accumbens in the rat. Brain Research, 344, 9–20. Tai, C. T., Cassaday, H. J., Feldon, J., & Rawlins, J. N. (1995). Both electrolytic and excitotoxic lesions of nucleus accumbens disrupt latent inhibition of learning in rats. Neurobiology of Learning and Memory, 64, 36–48. Thornton, J. C., Dawe, S., Lee, C., et al. (1996). Effects of nicotine and amphetamine on latent inhibition in human subjects. Psychopharmacology (Berl.), 127, 164–173. Trimble, K. M., Bell, R., & King, D. J. (1997). Enhancement of latent inhibition in the rat by the atypical antipsychotic agent remoxipride. Pharmacology, Biochemistry, and Behavior, 56, 809–816. Trimble, K. M., Bell, R., & King, D. J. (1998). Enhancement of latent inhibition in the rat at a high dose of clozapine. Journal of Psychopharmacology, 12, 215–219. Trimble, K. M., Bell, R., & King, D. J. (2002). Effects of the selective dopamine D(1) antagonists NNC 01–0112 and SCH 39166 on latent inhibition in the rat. Physiology and Behavior, 77, 115–123. Vaitl, D., & Lipp, O. V. (1997). Latent inhibition and autonomic responses: a psychophysiological approach. Behavioural Brain Research, 88, 85–93. Vaitl, D., Lipp, O., Bauer, U., et al. (2002). Latent inhibition and schizophrenia: Pavlovian conditioning of autonomic responses. Schizophrenia Research, 55, 147–158.
What the brain teaches us about latent inhibition (LI)
413
van den Bos, R., Charria Ortiz, G. A., Bergmans, A. C., & Cools, A. R. (1991). Evidence that dopamine in the nucleus accumbens is involved in the ability of rats to switch to cue-directed behaviours. Behavioural Brain Research, 42, 107–114. van den Bos, R., & Cools, A. R. (1989). The involvement of the nucleus accumbens in the ability of rats to switch to cue-directed behaviours. Life Sciences, 44, 1697–1704. van Kammen, D. P., Bunney, W. E., Jr., Docherty, J. P., et al. (1982). d-Amphetamineinduced heterogeneous changes in psychotic behavior in schizophrenia. The American Journal of Psychiatry, 139, 991–997. Venables, P. H. (1984). Cerebral mechanisms, autonomic responsiveness, and attention in schizophrenia. Nebraska Symposium on Motivation, 31, 47–91. Wagner, A. R., & Rescorla, R. A. (1972). Inhibition in Pavlovian conditioning: Application of a theory. In R. A. Boakes & M. A. Halliday (Eds.), Inhibition and Learning. New York: Academic Press. Warburton, E. C., Joseph, M. H., Feldon, J., Weiner, I., & Gray, J. A. (1994). Antagonism of amphetamine-induced disruption of latent inhibition in rats by haloperidol and ondansetron: implications for a possible antipsychotic action of ondansetron. Psychopharmacology (Berl.), 114, 657–664. Warburton, E. C., Mitchell, S. N., & Joseph, M. H. (1996). Calcium dependence of sensitised dopamine release in rat nucleus accumbens following amphetamine challenge: implications for the disruption of latent inhibition. Behavioural Pharmacology, 7, 119–129. Weiner, I. (1990). Neural substrates of latent inhibition: the switching model. Psychological Bulletin, 108, 442–461. Weiner, I. (2000). The latent inhibition model of schizophrenia. In M. S. Myslobodsky & I. Weiner (Eds.), Contemporary Issues in Modeling Psychopathology. Dordrecht: Kluwer Academic, pp. 197–230. Weiner, I. (2001). Latent inhibition. In J. N. Crawley, C. R. Gerfen, M. A. Rogawski, D. R. Sibley, P. Skolnick & S. Wray (Eds.), Current Protocols in Neuroscience, vol. 8.13. New York: John Wiley & Sons. Weiner, I. (2003). The “two-headed” latent inhibition model of schizophrenia: modeling positive and negative symptoms and their treatment. Psychopharmacology (Berl.), 169, 257–297. Weiner, I., Bernasconi, E., Broersen, L. M., & Feldon, J. (1997). Amphetamineinduced disruption of latent inhibition depends on the nature of the stimulus. Behavioural Pharmacology, 8, 442–457. Weiner, I., & Feldon, J. (1987). Facilitation of latent inhibition by haloperidol in rats. Psychopharmacology (Berl.), 91, 248–253. Weiner, I., & Feldon, J. (1997). The switching model of latent inhibition: an update of neural substrates. Behavioural Brain Research, 88, 11–25. Weiner, I., Feldon, J., & Katz, Y. (1987). Facilitation of the expression but not the acquisition of latent inhibition by haloperidol in rats. Pharmacology, Biochemistry, and Behavior, 26, 241–246. Weiner, I., Feldon, J., Tarrasch, R., Hairston, I., & Joel, D. (1998). Fimbria-fornix cut affects spontaneous activity, two-way avoidance and delayed non matching to sample, but not latent inhibition. Behavioural Brain Research, 96, 59–70. Weiner, I., Gaisler, I., Schiller, D., et al. (2000). Screening of antipsychotic drugs in animal models. Drug Development Research, 50, 235–249.
414
Ina Weiner
Weiner, I., Gal, G., & Feldon, J. (1999). Disrupted and undisruptable latent inhibition following shell and core lesions. Annals of the New York Academy of Sciences, 877, 723–727. Weiner, I., Gal, G., Rawlins, J. N., & Feldon, J. (1996a). Differential involvement of the shell and core subterritories of the nucleus accumbens in latent inhibition and amphetamine-induced activity. Behavioural Brain Research, 81, 123–133. Weiner, I., & Joel, D. (2002). Dopamine in schizophrenia: dysfunctional information processing in basal ganglia-thalamocortical split circuits. In G. Di Chiara (Ed.), Handbook of Experimental Pharmacology, vol. 54/II, Dopamine in the CNS II. Berlin: Springer, pp. 418–472. Weiner, I., Lubow, R. E., & Feldon, J. (1981). Chronic amphetamine and latent inhibition. Behavioural Brain Research, 2, 285–286. Weiner, I., Lubow, R. E., & Feldon, J. (1984). Abolition of the expression but not the acquisition of latent inhibition by chronic amphetamine in rats. Psychopharmacology (Berl.), 83, 194–199. Weiner, I., Lubow, R. E., & Feldon, J. (1988). Disruption of latent inhibition by acute administration of low doses of amphetamine. Pharmacology, Biochemistry, and Behavior, 30, 871–878. Weiner, I., Shadach, E., Barkai, R., & Feldon, J. (1997). Haloperidol- and clozapineinduced enhancement of latent inhibition with extended conditioning: implications for the mechanism of action of neuroleptic drugs. Neuropsychopharmacology, 16, 42–50. Weiner, I., Shadach, E., Tarrasch, R., Kidron, R., & Feldon, J. (1996b). The latent inhibition model of schizophrenia: further validation using the atypical neuroleptic, clozapine. Biological Psychiatry, 40, 834–843. Weiner, I., Tarrasch, R., Bernasconi, E., et al. (1997). Amphetamine-induced disruption of latent inhibition is not reinforcer-mediated. Pharmacology, Biochemistry, and Behavior, 56, 817–826. Weiner, I., Tarrasch, R., & Feldon, J. (1995). Basolateral amygdala lesions do not disrupt latent inhibition. Behavioural Brain Research, 72, 73–81. Westbrook, R. F., Jones, M. L., Bailey, G. K., & Harris, J. A. (2000). Contextual control over conditioned responding in a latent inhibition paradigm. Journal of Experimental Psychology. Animal Behavior Processes, 26, 157–173. Wheeler, D. S., Chang, R. C., & Miller, R. R. (2003). Massive preexposure and preexposure in multiple contexts attenuate the context specificity of latent inhibition. Learning and Behavior, 31, 378–386. Williams, J. H., Wellman, N. A., Geaney, D. P., et al. (1996). Antipsychotic drug effects in a model of schizophrenic attentional disorder: a randomized controlled trial of the effects of haloperidol on latent inhibition in healthy people. Biological Psychiatry, 40, 1135–1143. Williams, J. H., Wellman, N. A., Geaney, D. P., et al. (1997). Haloperidol enhances latent inhibition in visual tasks in healthy people. Psychopharmacology (Berl.), 133, 262–268. Williams, J. H., Wellman, N. A., Geaney, D. P., et al. (1998). Reduced latent inhibition in people with schizophrenia: an effect of psychosis or of its treatment. The British Journal of Psychiatry, 172, 243–249. Wuthrich, V., & Bates, T. C. (2001). Schizotypy and latent inhibition: non-linear linkage between psychometric and cognitive markers. Personality and Individual Differences, 30, 783–798.
What the brain teaches us about latent inhibition (LI)
415
Yap, C. S., & Richardson, R. (2005). Latent inhibition in the developing rat: an examination of context-specific effects. Developmental Psychobiology, 47, 55–65. Yee, B. K., Feldon, J., & Rawlins, J. N. (1995). Latent inhibition in rats is abolished by NMDA-induced neuronal loss in the retrohippocampal region, but this lesion effect can be prevented by systemic haloperidol treatment. Behavioral Neuroscience, 109, 227–240. Yee, B. K., Feldon, J., & Rawlins, J. N. (1997). Cytotoxic lesions of the retrohippocampal region attenuate latent inhibition but spare the partial reinforcement extinction effect. Experimental Brain Research, 115, 247–256. Young, A. M., Joseph, M. H., & Gray, J. A. (1993). Latent inhibition of conditioned dopamine release in rat nucleus accumbens. Neuroscience, 54, 5–9. Zahm, D. S. (1999). Functional-anatomical implications of the nucleus accumbens core and shell subterritories. Annals of the New York Academy of Sciences, 877, 113–128. Zahm, D. S. (2000). An integrative neuroanatomical perspective on some subcortical substrates of adaptive responding with emphasis on the nucleus accumbens. Neuroscience & Biobehavioral Reviews, 24, 85–105. Zahm, D. S., & Brog, J. S. (1992). On the significance of subterritories in the “accumbens” part of the rat ventral striatum. Neuroscience, 50, 751–767. Zuckerman, L., Rimmerman, N., & Weiner, I. (2003). Latent inhibition in 35-day-old rats is not an “adult” latent inhibition: implications for neurodevelopmental models of schizophrenia. Psychopharmacology (Berl.), 169, 298–307.
Latent inhibition and schizophrenia
17 Latent inhibition in schizophrenia and schizotypy: a review of the empirical literature Veena Kumari and Ulrich Ettinger
Introduction Schizophrenia is a severe neuropsychiatric disorder of unknown aetiology. The condition is clinically heterogeneous and affects virtually all areas of life, often resulting in disabling cognitive, perceptual, and emotional symptoms. The symptoms of schizophrenia are generally classified as positive (e.g. hallucinations and delusions), negative (e.g. anhedonia, alogia, thought paucity) and cognitive (e.g. thought disorder, bizarre thinking). The disease course is often chronic and the financial cost on the health-care system and society is tremendous, in addition to the personal consequences of the illness for friends and family of sufferers (Mangalore & Knapp, 2007; McEvoy, 2007). Tragically, a substantial number of patients commit suicide. The most successful treatments for schizophrenia that are currently prescribed are pharmacological, and most of these involve blockade of striatal dopamine receptors (Kapur & Remington, 2001). These treatments are successful in reducing the acute symptoms of the condition but provide no cure; accordingly, the need for a better understanding of the aetiology and pathophysiology of schizophrenia and the development of novel treatments is considerable. Drug development draws on many experimental strategies as well as on serendipity (Carpenter & Koenig, 2008; Javitt et al., 2008); translational models such as latent inhibition (LI) play an important role in the strategic effort to develop new treatments (Weiner, 2003). Indeed, a wealth of animal data exist that support the utility of the LI paradigm in studies of antipsychotic treatments (e.g., Weiner, this volume). Importantly, schizophrenia represents a complex and heterogeneous phenotype that is not dichotomously distributed in the general population (i.e. present/absent) but follows a more continuous pattern of distribution, with subclinical signs and symptoms of schizophrenia observed in otherwise healthy individuals (Eysenck, 1967; Strauss, 1969; van Os et al., 2000). Such symptoms are often termed schizotypy (Rado, 1960, cited in Lenzenweger, 1998). Schizotypy has been conceptualised in a number of different ways (for reviews, see Gruzelier, 2002; Lenzenweger, 1998; Meehl, 1989), Latent Inhibition: Cognition, Neuroscience and Applications to Schizophrenia, ed. R. E. Lubow and Ina Weiner. Published by Cambridge University Press # Cambridge University Press 2010.
419
420
Veena Kumari and Ulrich Ettinger
but most accounts stress the proximity to (and similarity with) schizophrenia. While schizotypy (or schizotypality) is the most common term used to describe this set of traits, others have referred to it as psychosis-proneness. That term is used especially in the Eysenckian tradition, with the Psychoticism (P) scale of the Eysenck Personality Questionnaire (EPQ) being frequently applied. Schizotypy is thought to be related to schizophrenia at different levels of analysis, including phenomenology, neurobiology, cognition, and genetics. Therefore, it is of importance to consider the relationship between LI and schizophrenia not in isolation but to relate it to the subclinical expression of schizophrenic signs and symptoms, viz. schizotypy. The aims of this chapter are to review the empirical literature of LI in people carrying a diagnosis of schizophrenia and those with a high level of schizotypy and to identify areas of future application of the LI model with a view to advance understanding of the pathophysiology, treatment and genetics of schizophrenia.
Latent inhibition in schizophrenia: empirical findings Details of all published empirical studies of LI in schizophrenia are listed in Table 17.1 (also see Lubow, 2005). It should be pointed out that the reduction or absence of LI in this patient group, where found, has been described as “better” performance, given that it represents faster learning of the CS–US association than in healthy controls (Hemsley, 1987, 1993; Gray et al., 1991). This point is notable given that schizophrenia patients fail to achieve normal performance level on a number of cognitive, motor, and perceptual tasks (Heinrichs & Zakzanis, 1998; Nuechterlein & Dawson, 1984; Reichenberg & Harvey, 2007). Such pervasive neurocognitive impairment could point to the operation of general motivational or brain functional deficits in addition to specific processing impairments. As is evident from the findings noted in Table 17.1, acutely ill patients with schizophrenia show disrupted LI, indicated by relatively unimpaired learning about the pre-exposed stimulus in the second phase of the experiment. Such disruption of LI appears to be more clearly seen in patients in the first two weeks of a current medicated psychotic episode or for up to one year in unmedicated patients (Baruch et al., 1988a; Gray et al., 1992a, 1995; Kathmann et al., 2000; Rascle et al., 2001; Vaitl et al., 2002; Yogev et al., 2004). However, some contradictory reports exist (Swerdlow et al., 1996; Williams et al., 1998). The finding of intact, rather than disrupted, LI in acutely ill patients in the studies by Swerdlow and colleagues may be explained by continued use of antipsychotic medication in their patient groups. On the other hand, intact LI in acutely ill unmedicated patients, as reported by Williams et al. (1998), is more difficult to understand. It is possible that the majority of studies in which unmedicated acutely ill patients showed disrupted LI had subjects with a history of drug abuse, since (a) substance abuse is reliably associated with the emergence of psychosis, with about half of the schizophrenia population having a lifetime diagnosis of comorbid substance abuse (Buckley et al., 2009), and (b) drug abuse has
Cohen et al., 2004
Cross-sectional: 26 acute patients (18 males; mean age (in years) ¼ 32.8, SD ¼ 11.8)
Baruch et al., 1988a
53 controls (24 males; mean age ¼ 30.3, SD ¼ 10.5) Longitudinal: 11 (of 26) acute and 13 (of 27) chronic patients retested after 6–7 weeks 30 patients (24 males; mean age ¼ 17.1, range ¼ 13–21); 25 paranoid, 2 disorganised, 2 unspecified, 1 residual; all except one on antipsychotics; average duration of current hospitalisation ¼ 4.73 months (SD ¼ 4.64) 30 controls (16 males; mean age ¼ 16, range ¼ 14–18)
27 chronic patients (16 males; mean age ¼ 41, SD ¼ 9.6)
Sample
Authors
Table 17.1. Studies of latent inhibition in schizophrenia
Visual, within-subjects
Auditory, between-subjects
LI method
Patients were split into high/low positive/ negative symptom groups: high negative with low positive symptoms group showed greatest LI; this was the only of the four subgroups differing from controls. Patients with greater LI were more likely to have been hospitalised for longer duration than subjects with relatively lower LI.
LI observed in controls and chronic patients but not in acute patients (interaction P < 0.01). There were no differences between groups in NPE condition, differences come from PE condition. LI observed in patients with low, but not high, BPRS scores. Both acute and chronic patients showed LI on retesting. Both groups showed LI, but there was a higher number of patients than controls in the group showing greatest LI.
Results
Experiment II: 19 patients (10 males; mean age ¼ 41.8, SD ¼ 9.6) 20 controls (5 males; mean age ¼ 41.8, SD ¼ 9.6) Mean duration of illness ¼ 22.5 years (SD ¼ 10.9); all treated with antipsychotics; “predominantly high-level of negative symptoms” Patients off benzodiazepines and anticholinergics on day of testing 16 acutely ill patients (12 males): 9 within first psychotic episode, 7 in acute phase of a chronic disorder, 2 unmedicated and 1 drug-naı¨ ve, all treated patients tested within days 1–14 of start of neuroleptic treatment 16 chronic patients (12 males): continuously ill for at least 6 months, all receiving neuroleptics 20 healthy controls (10 males) Acute group had significantly higher positive symptom scores, chronic group had nonsignificantly higher negative symptom scores 15 drug-naı¨ ve patients (7 males; mean age ¼ 28.54, SD ¼ 1.52); final N ¼ 13 as two male patients were classified as non-learners
Gal et al., 2009
Gray et al., 1995
Gray et al., 1992a
Sample
Authors
Table 17.1. (cont.)
Auditory, within-subjects
Auditory, between-subjects
Visual, within-subjects
LI method
Both groups showed LI, no difference between groups. LI did not correlate with BPRS scores, but correlated with duration of illness:
LI was reduced in acutely ill patients; normal LI occurred in controls and chronic patients. LI did not correlate with BPRS scores.
Controls showed non-significantly more LI than patients on trials 1–5; patients showed more LI on trials 6–10 with controls showing attenuation of LI on these trials.
Results
10 typically treated chronic patients (8 males; mean age ¼ 40, SD ¼ 12.8, range ¼ 23–63) 12 atypically treated chronic patients (11 males; mean age ¼ 38, SD ¼ 11.7, range ¼ 22–56) 18 controls (13 males; mean age ¼ 35, SD ¼ 5.2, range ¼ 26–42) 20 paranoid patients (18 males; mean age ¼ 45.6, SD ¼ 14), all medicated 19 non-paranoid patients (12 males, mean age ¼ 38.8, SD ¼ 17.5), all medicated 48 student controls (20 males; mean age ¼ 23, SD ¼ 3.7)
Leumann et al., 2002
Auditory, between-subjects
Between-subjects: go/no-go task with auditory conditional stimulus predicting visual go-signal; pre-exposure phase was discrimination task Auditory, between-subjects
17 acute patients (12 males; mean age ¼ 30.4, SD ¼ 10.9); all on neuroleptics 16 remitted patients (8 males; age ¼ 33.2, SD ¼ 8.9); all on neuroleptics 20 controls (9 males; mean age ¼ 27.8, SD ¼ 8.5)
Kathmann et al., 2000
Lubow et al., 1987
Auditory, between-subjects
14 chronic patients (11 males, mean age ¼ 32.29, SD ¼ 6.90); all patients stable on antipsychotics 14 controls (6 males; mean age ¼ 27.8)
Guterman et al., 1996
13 healthy volunteers (7 males; mean age ¼ 31, SD ¼ 2.36); range of illness duration 2–36 months.
All three groups showed LI.
Both patient groups and the controls showed LI; no difference observed between typically and atypically treated patients.
reduced LI associated with shorter duration of illness, normal LI associated with longer duration of illness. Patients with illness duration < 12 months showed no LI, those with > 12 months did show LI. No correlation of LI with basal ganglia D2 binding (using SPET). Controls: PE subjects had lower CNV than NPE subjects in first and second blocks of trials – no such effect found in patients. No significant difference in RT data for PE and NPE subjects in either group. All groups showed LI at behavioural level (RT). PE increased N100 amplitude in acute patients, but caused trend towards decreased N100 in controls. The amplitude of CNV was unaffected.
35 acute patients (23 males; mean age ¼ 28.11, SD ¼ 7.57); 6 of them free of antipsychotics 30 chronic patients (21 males; mean age ¼ 30.57, SD ¼ 8.65); 2 of them free of antipsychotics) 40 controls (14 males, mean age ¼ 31.64, SD ¼ 8.84)
Experiment II: 22 chronic patients (15 males; mean age ¼ 32.86, SD ¼ 4.59) 22 schizotypal relatives (9 males, mean age ¼ 51.26, SD ¼ 17.22) 19 non-schizotypal relatives (6 males; mean age ¼ 43.74, SD ¼ 17.68) 27 controls (13 males; mean age ¼ 27.55, SD ¼ 7.81) 73 patients (49 males; mean age ¼ 34.84, range ¼ 18–62); acute: tested within 14 days of admission; chronic: tested after more than 14 days of hospitalisation or outpatients 13 mood disordered patients (8 males; mean age ¼ 34.85, range ¼ 18–61): 4 with major
Rascle et al., 2001
Serra et al., 2001
Swerdlow et al., 1996
Visual search, withinsubjects
32 patients (17 males; mean age ¼ 35.8); all except two treated with antipsychotics 32 controls (15 males; mean age ¼ 34.1)
Lubow et al., 2000
Experiment I: auditory, between-subjects as in Baruch et al., 1988a (73 controls, 45 schizophrenia, 13 mood disorder, 19 obsessive compulsive disorder)
Auditory, between-subjects
Visual contingency detection paradigm, between-subjects
LI method
Sample
Authors
Table 17.1. (cont.)
Experiment I: Schizophrenia patients showed overall slower learning but had similar LI. Similar levels of LI were also observed in the other two patient groups. Experiment II: Overall slower learning observed in schizophrenic PE and NPE groups, but comparable LI.
Condition (PE/NPE) by Gender by Group interaction: female controls and male patients showed trends towards LI, male controls showed weak trend, and female patients showed the reverse pattern. Controls showed LI, acute patients did not show LI, and chronic patients showed enhanced LI. In acute patients, the presence of LI was associated with higher total PANSS general and negative symptom scores. PANSS negative symptom scores were correlated with speed to find association in phase 2 (r ¼ 0.42, P ¼ 0.02). Controls but none of the three spectrum groups showed LI. The absence of LI in the spectrum groups was due not to faster learning in PE condition, but to slower learning in NPE condition.
Results
Williams et al., 1998
Vaitl et al., 2002
Swerdlow et al., 2005
depressive episode, 7 with bipolar disorder, 2 with schizo-affective disorder 19 patients with obsessive-compulsive disorder (8 males; mean age ¼ 32.63, range ¼ 19–55) 107 controls (42 males; mean age ¼ 24.67, range ¼ 18–57) 20 patients (16 males; mean age ¼ 37.1, range ¼ 21–48); all except two on antipsychotics 55 controls (19 males; mean age ¼ 22.6, range ¼ 18–34) 16 unmedicated patients (8 males; mean age ¼ 33.6, range ¼ 19–61); 6 drug-naı¨ ve and 10 untreated for at least 18 days 16 medicated patients (8 males; mean age ¼ 32.1, range ¼ 20–49) 16 controls (8 males; mean age ¼ 27.9, range ¼ 18–54) 57 schizophrenia patients (23 drug-naı¨ ve, 34 receiving antipsychotics) 23 drug-naı¨ ve (14 males; median age ¼ 24, interquartile range ¼ 21–31) 34 treated (28 males; median age ¼ 30.5, interquartile range ¼ 23–37) 73 controls (39 males; median age ¼ 25, interquartile range ¼ 21–30) History of drug or alcohol abuse not allowed in the patient groups Auditory, between-subjects
Visual CS, auditory tone as US in reaction time task Dependent variable was skin conductance response (SCR)
Visual, within-subjects
Experiment II: visual, between-subjects (34 controls, 23 schizophrenia patients)
LI observed in drug-naı¨ ve but not in medicated patients. Controls given saline showed LI, but controls given haloperidol did not. Controls showed more LI than medicated patients, but did not differ from drug-naı¨ ve patients. Effect of schizotypy: controls with STA score > 16 showed reduced LI compared to those with lower STA. Effect of smoking: smokers show lower LI than non-smokers.
Both controls and patients showed LI. No difference in LI between in- and outpatients (all except one in each group treated, clinical descriptives very similar between groups). Larger SCR during NPE than PE stimuli found in controls and medicated but not in unmedicated patients. Similar LI effects were observed at behavioural levels in all three groups.
28 positive symptom patients (15 males; mean age ¼ 34.3, SEM ¼ 2.7; 13 female; mean age ¼ 38.7, SEM ¼ 1.9) 13 negative symptom patients (8 male; mean age ¼ 39.3, SEM ¼ 2.8; 5 female, mean age ¼ 43.6, SEM ¼ 6.3) All patients free from antipsychotic medication for 4 weeks or longer 24 controls (6 male, mean age ¼ 30.8, SEM ¼ 3.5; 18 female, age ¼ 28.9, SEM ¼ 2.4)
Yogev et al., 2004
Auditory, between-subjects
LI method
Reduced LI observed in patients with high positive symptoms. Generally poor learning and intact LI in patients with high negative symptoms. Evidence of over-switching in positive symptom group and of under-switching in negative symptom group on a (different) attention task.
Results
Notes: BPRS, Brief Psychiatric Rating Scale (Overall & Gorham, 1962); SPET, single photon emission tomography; PE, pre-exposed; NPE, non-pre-exposed; RT, reaction time; PANSS, positive and negative syndrome scale (Kay et al., 1987); CS, conditioned stimulus; US, unconditioned stimulus; STA, Scale of the Schizotypy Questionnaire (Claridge & Broks, 1984).
Sample
Authors
Table 17.1. (cont.)
Latent inhibition in schizophrenia and schizotypy
427
recently been linked with reduced LI by Stevens et al. (2007). The conclusions of Stevens et al., however, have to be treated with much caution because (i) their within-subject procedure did not counterbalance PE and NPE stimuli, (ii) there appears to be a ceiling effect, and (iii) the differences between groups came from the NPE condition. Furthermore, drug abuse may not explain the results of Williams et al. (1998). Yogev et al. (2004), like Williams et al. (1998), had also tested acutely ill unmedicated patients with no prior history of drug or alcohol abuse. Yogev et al.’s findings of disrupted LI in patients with positive symptoms and intact LI in those with negative symptoms within the group of unmedicated patients suggest that the LI effect may be a function of a complex interaction between symptom profiles even in unmedicated patients. A number of independent groups have demonstrated normal LI in chronic medicated schizophrenia samples (Cohen et al., 2004; Kathmann et al., 2000; Leumann et al., 2002; Lubow et al., 1987; Swerdlow et al., 2005). Chronic patients, who tend to experience a high level of negative and a low level of positive symptoms, may in fact show enhanced LI (Cohen et al., 2004; Rascle et al., 2001). Earlier reports suggest that normalisation or enhancement of LI may be facilitated by antipsychotics (Baruch et al., 1988a; Gray et al., 1992a), probably due to dopamine antagonism since, as in animals (Moser et al., 2000), LI is disrupted in human subjects by the indirect dopamine agonist, amphetamine (Gray et al., 1992b; Kumari et al., 1999; Swerdlow et al., 2003; Thornton et al., 1996). There have also been reports of disrupted LI in patients receiving high doses of a typical antipsychotic, haloperidol (Williams et al., 1998), and in healthy volunteers under administration of a high dose of this drug (Kumari et al., 1999), suggesting that the effects of antipsychotics on LI in patients may be highly dose-dependent and their LI-correcting action may be seen only when given in therapeutic doses allowing for tolerance in the chronic phase. Such a possibility is also supported by data in rats showing potentiated LI in rats over a range of doses of antipsychotic drugs, but reduced LI at the highest dose of haloperidol tested, 3 mg/kg (Dunn et al., 1993). There is one cross-sectional report of LI normalisation with a longer than 12 months duration of illness without antipsychotic treatment, suggesting the influences of other factors/processes intrinsic to illness progression (Gray et al., 1995). This finding, however, is based on a small number of patients (N ¼ 13; seven with an illness duration of 12 months). The duration of illness in this particular study can also be described as “the duration of untreated psychosis” and it did not specifically examine positive and negative symptom profiles of patients with regard to LI. Reduced LI in acutely ill schizophrenia patient groups is thought to result from an overactive switching mechanism leading to excessive adherence to concurrent situational factors, thereby failing to show the influence of previously learned regularities (Gray et al., 1991; Hemsley 1987, 1993). Enhanced LI in chronic patient samples, following the two-headed model of LI proposed more recently by Weiner (2003),
428
Veena Kumari and Ulrich Ettinger
may indicate underswitching in this patient group. Also consistent with Weiner’s model is a report of persistent LI in chronic patients with a high level of negative symptoms (score 22–35 on the negative symptoms scale of the Positive and Negative Syndrome scale; Kay et al., 1987) under conditions that attenuate LI in normal controls (Gal et al., 2009). The puzzling observation reported by Gal et al. (2009), however, is that the same chronic patient group failed to show significant LI under conditions that produced LI in healthy controls. In a study by Guterman et al. (1996), it is also unclear as to why chronic medicated patients, unlike controls, failed to show LI, as assessed with contingent negative variation to pre-exposed and non-pre-exposed stimuli. Chronic medicated patients, however, do show weaker learning in LI paradigms (e.g. Swerdlow et al., 1996), and normal LI in the chronic group may sometimes be obscured by poor overall associative learning (Serra et al., 2001). Disrupted LI is unusual among behavioural and cognitive abnormalities commonly reported in schizophrenia in that it seems to be present mainly during the positive symptomatic, unmedicated state of acute schizophrenia. An important issue here is that positive and negative symptoms of schizophrenia often co-exist and, complicating this issue further, there is one report of reduced LI in acutely ill patients with elevated negative, rather than positive, symptoms (Rascle et al., 2001). Such discrepancies may reflect procedural differences between studies, especially the task characteristics in terms of the number of pre-exposures and conditioning trials and the differences in response criterion with some but not others providing continuous measure of LI. In both human (Gal et al., 2009) and animal studies (Ruob et al., 1998; Shadach et al., 1999), the LI effect is critically dependent on the number of pre-exposure and conditioning trials; LI is attenuated with reduced number of preexposures or increased number of conditioning trials. To our knowledge, systematic comparisons of procedurally different LI paradigms in healthy or schizophrenia populations and their effects on the direction and magnitude of group differences have not been carried out. In summary, the most consistent behavioural findings from studies published to date reveal disrupted LI in acutely ill unmedicated samples, and normal or possibly enhanced LI in chronic schizophrenia samples. Further experiments are required to clarify the optimal task conditions for eliciting these effects, their psychological underpinnings (e.g. under- versus over-switching; narrowing versus widening of attention), and how these effects may change with pharmacological influences of antipsychotic medications, particular symptoms or some other processes associated with illness progression. An important point to consider is that normalisation of LI may occur much earlier into the antipsychotic treatment, and not necessarily require a minimum of 14 days, in line with their other clinical effects (Agid et al., 2008; Kapur et al., 2005). The traditional view, as reflected in major psychiatry textbooks, of “delayed onset” of antipsychotic response in the range of 2–3 weeks may not be accurate (Kapur et al., 2005) and may add yet another source of variation in LI
Latent inhibition in schizophrenia and schizotypy
429
studies of acute schizophrenia patients. It may also be very helpful to study LI during the prodromal period of schizophrenia, i.e. before the onset of full-blown psychotic symptoms (e.g. Phillips et al., 2002).
Latent inhibition and schizotypy: empirical findings Both Kraepelin (1919/1971) and Bleuler (1950) described a form of personality aberration that is thought to be a quantitatively less severe expression of schizophrenia. They coined the term “latent schizophrenia” to refer to this type of personality. Later, Rado (1960, cited in Lenzenweger, 1998) used the terms “schizotype” to convey the idea of a “schizophrenia phenotype” and “schizotypal behaviour” to denote the behavioural manifestations of this phenotype. The key features of schizotypal personality include feelings, behaviours, and cognitions such as paranoia, poor interpersonal skills, delusional thinking, superstition, disordered thought, odd behaviour and speech, and social anxiety (Raine, 2006; Raine et al., 1994). Schizotypal personality traits may manifest clinically as schizotypal personality disorder (SPD; see Diagnostic and Statistical Manual (DSM) IV) but are typically measured in the general population using psychometric self-report questionnaires. The Psychoticism (P) subscale of the Eysenck Personality Questionnaire (EPQ; Eysenck & Eysenck, 1975), STA scale of the Schizotypy Questionnaire (Claridge & Broks, 1984), the Schizotypal Personality Questionnaire (SPQ; Raine, 1991), the Rust Inventory of Schizotypal Cognitions (RISC; Rust, 1989) and the Oxford and Liverpool Inventory of Feelings and Experiences (O-LIFE; Mason et al., 1995) have been used by several investigators within the context of LI research. These questionnaires differ in their theoretical context and breadth of assessed traits (review, Chapman et al., 1995; Vollema & van den Bosch, 1995). Briefly, the EPQ-P scale assesses normal behaviour patterns relevant to “schizoid” and “psychopathic” disorders while avoiding clinical symptomatology applicable to pathological populations (Eysenck & Eysenck, 1991). The STA and the SPQ are both modelled after the features of SPD in the DSM III. The RISC provides a measure of mild schizotypal cognitions from the positive symptom spectrum and places an emphasis on avoiding the obviously pathological items found in some other questionnaires. The O-LIFE measures four dimensions of schizotypy, namely, unusual experiences, cognitive disorganisation, introvertive anhedonia, and impulsive non-conformity. Its cognitive disorganisation scale aims to assess attentional difficulties and social anxiety, and supposedly corresponds to thought disorder symptoms of schizophrenia. The unusual experiences subscale examines deviant perceptual and cognitive experiences and corresponds to the reality distortion syndrome. Introvertive anhedonia assesses the inability to experience pleasure, thus corresponding most closely to the negative syndrome. Impulsive non-conformity examines reckless and destructive behaviours and resembles antisocial or psychopathic traits.
430
Veena Kumari and Ulrich Ettinger
Details of all published empirical studies of LI in relation to schizotypal traits are presented in Table 17.2. The findings reviewed in Table 17.2 are consistent in revealing an association between high schizotypy and reduced LI when schizotypy is sampled using the EPQ P-scale, the STA and the O-LIFE. Recent data using the O-LIFE suggest that the positive dimension of schizotypy (Evans et al., 2007) may be more relevant to LI-disruption than the negative dimension, and that this effect may be dopaminemediated. There is experimental evidence demonstrating that dopamine antagonismmediated enhancement or normalisation of human LI depends on schizotypy level of the sample (Kumari et al., 1999; Williams et al., 1997). The association between schizotypy and LI, however, is generally present with only a small-to-moderate effect size. Additionally, it should be noted that most of the investigators have not recorded or ruled out, using reliable methods, current or past drug abuse in their samples. It is possible that drug abuse affects or enhances the influence of schizotypy traits in LI. As discussed elsewhere (Kumari et al., 1999), schizotypy scores are generally lower in drug-screened healthy populations than published norms, and high schizotypy is known to be associated with use of psychoactive compounds (Larrison et al., 1999; Williams et al., 1996). The small-to-moderate effect size of the association between reduced LI and high schizotypy may also indicate that schizotypy is perhaps not the only, or the most, meaningful determinant of LI in healthy populations. Interestingly, one study (Braunstein-Bercovitz, 2000) showed that trait anxiety is related to LI, and this effect is stronger than, and independent of, schizotypy (also see Braunstein-Bercovitz, this volume). This finding led the author to propose that “the high schizotypals’ diminished ability to suppress attention to irrelevant stimuli, the consequence of which is the attenuation of LI, is, indeed related to both high anxiety levels and schizophrenia-like symptoms”. The exact mechanism underlying the association between anxiety and human LI remains obscure at present although the suggestion has been made that it could be mediated via stress (Braunstein-Bercovitz, 2000; Braunstein-Bercovitz, this volume). In animals, dopaminergic activity in the mesolimbic system increases in response to stress (review, Salamone et al., 1997) and increased dopaminergic activity in this system has been proposed as the biological basis of disrupted LI in schizophrenia (Gray, 1998; Gray et al., 1991, 1999). Interestingly, there are no data examining whether anxiety or stress levels might have played a role in LI alterations reported in patients with acute schizophrenia, who are likely to be even more susceptible to task demands as well as the experimenter’s instructions and behaviour than high schizotypals. In summary, susceptibility to schizophrenia as indexed by a high level of schizotypal personality traits, especially on the dimensions corresponding to the positive symptoms of schizophrenia, is associated with a reduced LI effect in otherwise healthy humans. This association is generally present with a small-to-moderate effect size, and may be linear or non-linear depending on task characteristics (with both very low and very high levels of schizotypy abolishing LI; see Table 17.2).
Sample 205 subjects (108 males; mean age ¼ 24.1 years, SD ¼ 6.1)
53 subjects (24 males; mean age ¼ 30.3, range ¼ 18–79)
60 volunteers (11 males; mean age ¼ 22.5, range ¼ 18–26)
Authors
Allan et al., 1995
Baruch et al., 1988b
Braunstein-Bercovitz, 2000
Table 17.2. Studies of latent inhibition in schizotypy
SPQ factor scores derived from a sample of 219 subjects – Anxiety-loaded and PerceptualDisorganisation factors STAI
EPQ P scale, STA, LSHS
STA, STB, EPQ P scale (median split)
Schizotypy method
Visual, betweensubjects
As in Baruch et al., 1988a
Auditory, between subjects
LI method High STA-A subjects showed lower LI than high STA-A subjects (due to faster learning in preexposed high STA). High STA-B subjects showed lower LI than high STA-B subjects (due to faster learning in preexposed high STA). No association of LI with EPQ P. Correlation of LI with EPQ P (P < 0.07) and LSHS (P < 0.13). Low but not high EPQ P subjects showed LI. LI observed in low but not high schizotypy. LI observed at trend (P ¼ 0.052) in low anxiety group, but absent in high anxiety group. Both anxiety and schizotypy independently accounted for LI, but effects of anxiety were stronger.
Results
Experiment I: 83 subjects (8 males; mean age ¼ 21.9, range ¼ 19–28); 44 tested with no masking and 39 with a masking LI procedure Experiment II: 78 subjects (22 males; mean age ¼ 22.8, range ¼ 18–45); 41 tested with a low load masking and 37 with a high load masking LI procedure 77 subjects (12 males; mean age ¼ 23.2, SD ¼ 2.1, range ¼ 20–27)
93 subjects (46 males)
Braunstein-Bercovitz & Lubow, 1998
Burch et al., 2004
Braunstein-Bercovitz et al., 2004
Sample
Authors
Table 17.2. (cont.)
O-LIFE
SPQ (median split)
SPQ, STA (median split)
Schizotypy method
Visual and auditory, betweensubjects
Auditory, between subjects High and low load conditions
Visual, betweensubjects
LI method
In low load condition: LI observed in low but not high schizotypals. In high load condition: LI observed at trend in high but not low schizotypals. Latent facilitation (LF) observed in people with high, but not low, Unusual Experiences scores under conditions of small numbers of pre-exposure trials in the visual task.
In low load masking condition: LI observed in low but not high schizotypals. In high load masking condition: LI observed at trend in high but not low schizotypals.
Results
Sample Experiment I: 120 subjects (43 males; mean age ¼ 21.2)
Experiment I: 105 subjects (52 males; mean age ¼ 31.4, range ¼ 24–46): non-deprived smokers, deprived smokers who smoked, deprived smokers who did not smoke, nonsmokers
Authors
De la Casa et al., 1993
Della Casa et al., 1999
Table 17.2. (cont.)
Experiment I: visual, betweensubjects with either 3, 6, or 15 minutes pre-exposure
Visual and auditory, betweensubjects
STA (median split)
LI method
Experiment I: Paranoia, Schizophrenia and K scales of abbreviated MMPI (median split)
Schizotypy method
Low psychosis-prone group: LI observed at 6 and 15 (but not 3) minutes preexposure duration. High psychosis-prone group: LI observed at 15 but not 3 or 6 minutes preexposure duration. Longer pre-exposure durations slowed learning more strongly in PE groups. Experiment I: visual LI observed in low but not high STA subjects; not affected by smoking status. Auditory LI not affected by STA or smoking.
Results
Evans et al., 2007
Authors
Table 17.2. (cont.)
80 subjects (34 males; mean age ¼ 20.39, range ¼ 18–31, SD ¼ 2.34), 16 of them smokers
Experiment II: 102 subjects (50 males; mean age ¼ 31.9, range ¼ 23–41), groups of deprived/nondeprived smokers/nonsmokers as above
Sample
O-LIFE
Schizotypy method
Visual, withinsubjects
LI method
Experiment II: in visual task, no overall LI effect; high STA had LI at trend level but low STA did not; no effect of smoking. LI observed in auditory task; smokers showed more LI than non-smokers; no effect of STA. Both RT and correct responses of PE trials correlated with Unusual Experiences scores. RT and correct responses of PE trials also correlated with smoking status: smokers show reduced LI; recent smokers had smaller LI than non-recent smokers.
Results
Sample 80 subjects (16 males; mean age ¼ 21.83, SD ¼ 7.89, range ¼ 16–51)
100 subjects (53 males; mean age ¼ 32.1, SD ¼ 0.5)
48 subjects (12 males; mean age ¼ 23.3, range ¼ 16–32)
Authors
Gray et al., 2002
Ho¨fer et al., 1999
Lipp & Vaitl, 1992
Table 17.2. (cont.)
EPQ P scale, LSHS, STA
STA (median split)
O-LIFE
Schizotypy method
Visual, betweensubjects
Visual, betweensubjects
LI method
Reduced LI associated with increased Unusual Experiences, Cognitive Disorganisation, and Impulsive Non-conformity scores but not with Introvertive Anhedonia scores. Low STA subjects showed LI, high STA subjects did not. No relationship of LI to STB (borderline personality) from the CSTQ; no relationship of LI to EPQ P scale (also part of CSTQ). Neither EPQ P nor LSHS correlated with LI. STA was associated with LI as measured using electrodermal response.
Results
Sample 76 subjects (21 males; mean age ¼ 17–51) Experiment III: control group of 24 subjects (10 males; age range ¼ 17–35) Experiment IV: control group of 26 subjects (25 males; age range ¼ 17–40)
Experiment II: 12 subjects (all females, mean age ¼ 18.5, range ¼ 17–23) Experiment III: 44 subjects (22 males; mean age ¼ 19.34, range ¼ 17–35)
Authors
Lipp et al., 1994
Lubow & De la Casa, 2002
Table 17.2. (cont.)
STA (median split)
EPQ P scale, STA, STB, LSHS, VSS, SPQ, MagId, PhysAn, PerAbb, SocAn
Schizotypy method
Visual, withinsubjects
Experiment III: visual, betweensubjects; learning is dependent variable Experiment IV: visual, betweensubjects; SCR is dependent variable
LI method
Experiment II: LI observed in low but not high schizotypy group. Experiment III: LI observed in lowSTA female and high-STA male groups.
Experiment III: There was a trend (P ¼ 0.1) for SPQ to be associated with LI; otherwise no significant effects. Experiment IV: LI reduced in high STA, STB, and LSHS.
Results
Sample
Experiment I: 48 subjects (16 males; mean age ¼ 23.1) Experiment II: 20 subjects (7 males; mean age ¼ 22.2, range ¼ 20–27)
Experiment II: 38 males (mean age ¼ 19.3, SD ¼ 2.72) and 142 females (mean age ¼ 18.3, SD ¼ 1.2)
16 high (9 males; mean age ¼ 20.1, range ¼ 19–21) and 16 low (8 males; mean age ¼ 20.2, range ¼ 19–23) schizotypals selected from a pool of 100 subjects
Authors
Lubow et al., 1992
Lubow et al., 2001
Tsakanikos, 2004
Table 17.2. (cont.)
STA scorers in the lowest (16%) and highest (16%) range
STA (median split)
STA, EPQ P scale (median split)
Schizotypy method
Visual, betweensubjects All participants were also tested on
Visual search, withinsubjects
Experiment I: auditory, betweensubjects Experiment II: visual, betweensubjects
LI method Experiment I: LI reduced in high EPQ P subjects and in high STA subjects (the two scales correlated r ¼ 0.31). Experiment II: LI reduced in high EPQ P and in high STA subjects (the two scales correlated r ¼ 0.41). Trend towards gender by condition (PE, NPE) by schizotypy interaction (P ¼ 0.065). Condition by schizotypy interaction observed in females but not in males. LI observed in low but not high schizotypal subjects; no effect of schizotypy in visual pop-out task.
Results
54 subjects (23 males; age range ¼ 17–49)
Wuthrich & Bates, 2001
EPQ P scale, SPQ intercalated with NEO PI-R using NEO response format Subjects split into 5 groups based on SPQ score
Schizotypy method enhanced stimulus salience using a visual pop-out task Auditory, betweensubjects
LI method
Average schizotypes showed normal LI; high or low groups showed reduced LI. EPQ P or NEO PI-R scores did not correlate with LI.
Results
Notes: PE, pre-exposed; NPE, non-pre-exposed; RT, reaction time; SCR, skin conductance response; STA & STB, two scales of the Schizotypy Questionnaire (STA corresponds to schizotypal personality disorder and STB to borderline personality disorder; Claridge & Broks, 1984); EPQ P, Psychoticism scale of the Eysenck Personality Questionnaire (Eysenck & Eysenck, 1975); O-LIFE, Oxford-Liverpool Inventory of Feelings and Experiences (Mason et al., 1995); SPQ, Schizotypal Personality Questionnaire (Raine, 1991); NEO PI-R, Revised NEO Personality Inventory (Costa & McCrae, 1992); CSTQ, Combined Schizotypal Traits Questionnaire (includes items from STA; Bentall et al., 1989); MMPI, Minnesota Multiphasic Personality Inventory; LSHS, Launay-Slade Hallucinations Scale (Launay & Slade, 1981); VSS, Venables’ Schizotypy Scale (Venables et al., 1990); MagId, Magical Ideation scale (see Chapman et al., 1995); PhysAn, Physical Anhedonia scale (see Chapman et al., 1995); PerAbb, Perceptual Aberration scale (see Chapman et al., 1995); SocAn, Social Anhedonia scale (see Chapman et al., 1995).
Sample
Authors
Table 17.2. (cont.)
Latent inhibition in schizophrenia and schizotypy
439
Latent inhibition and the schizophrenia spectrum: outlook We next consider future applications of the LI model to advance the understanding of schizophrenia and its treatment. We also consider some methodological issues applicable generally to LI research, in addition to those discussed in earlier sections. The LI paradigm is potentially a useful biomarker of schizophrenia (e.g., Lipina & Roder, this volume; Weiner, this volume). A biomarker is an objectively measurable characteristic or feature that is studied as an indicator of normal or pathological processes. Most importantly in this context, biomarkers are popular surrogate markers for changes in disease severity in pharmacological challenge or therapeutic intervention studies. The utility of biomarkers lies principally in their less complex and better understood neural basis compared to the disease process under study. A likely application of the LI paradigm in future work is the development of novel antipsychotic treatments, especially given the emphasis of this work on dopaminergic mechanisms (Kapur & Remington, 2001). However, future studies might also wish to consider in more detail other neurotransmitter systems such as the glutamatergic, serotonergic and cholinergic systems increasingly recognised as playing a role in both schizophrenia and LI (Carpenter & Koenig, 2008; Javitt et al., 2008; Weiner, 2003). A second concept that is widely studied in current biological psychiatry research is that of the endophenotype. An endophenotype is a biological or behavioural marker thought to provide a more direct and specific consequence of genetic risk than the illness phenotype itself (Gottesman & Gould, 2003). The phenotypic and genetic complexity of an endophenotype is assumed to be lower than that of the illness phenotype, making an endophenotype ideally suited to studying the genetics of the illness, or at least of neurophysiologically/anatomically specific facets of the illness. Schizophrenia endophenotypes include deficits in neuropsychological function (review, Skelley et al., 2008; Snitz et al., 2006), prepulse inhibition (Kumari et al., 2005), oculomotor control (Calkins et al., 2008; Hutton & Ettinger, 2006), and macroanatomical brain structure and function (Marcelis et al., 2005; Prasad & Keshavan, 2008; Whalley et al., 2004, 2007; van Haren et al., 2008). The endophenotype research program has flourished in the last years due to advances in molecular genetics, making the investigation of direct links between schizophrenia risk polymorphisms and endophenotypes easily accessible. Regarding LI, much work still needs to be done to establish whether the LI deficit represents an endophenotype for schizophrenia. Key criteria for the validation of an endophenotype include the observation of the marker in the patient group, its trait nature, its heritability, its segregation with the illness in affected families, and its observation in clinically unaffected individuals at genetic risk for illness (Gottesman & Gould, 2003). The most widely studied group to address the latter criterion consists of biological relatives of schizophrenia patients. LI has been studied only once in relatives of schizophrenia patients (Serra et al., 2001), and nothing is known
440
Veena Kumari and Ulrich Ettinger
about the heritability of the phenomenon in healthy humans. Given the observation of associations between LI and schizotypy, however, it is likely that the LI deficit may at least in part represent a trait marker of not only the illness but also its subclinical expression (which shows genetic similarity with schizophrenia; Fanous et al., 2007) and, perhaps, its genetic vulnerability. An interesting synthesis of the biomarker and endophenotype programmes is seen in the field of pharmacogenetics and pharmacogenomics (Arranz & de Leon, 2007; Reynolds, 2007; Thaker, 2007). Should research establish the genetic causes of the LI deficit in the schizophrenia spectrum, future pharmacology could then utilise this genetic information to predict inter-individual differences in drug response. Such an approach will eventually aim at individually tailoring treatment on the basis of a patient’s genetic makeup. One question that seems to be important for future research concerns the relation between what is termed “excessive switching” (Weiner, 2003) in schizophrenic patients with attenuated LI and the established deficit known as perseveration which is also seen in schizophrenia. For example, in the Wisconsin Card Sorting Test (WCST) patients with schizophrenia persevere by adopting a previously successful strategy in the face of feedback that the strategy has now become invalid. It would be of interest to explore how these two seemingly orthogonal deficits can be reconciled, perhaps by exploring problems in response selection at automatic vs. controlled levels of processing. A final avenue for future research that seems to be of importance concerns the nosological specificity of LI within psychiatry. Some studies have shown normal (Swerdlow et al., 1996) or even enhanced LI in obsessive-compulsive disorder (OCD) (Kaplan et al., 2006; Swerdlow et al., 1999). Swerdlow et al. (1996) also observed LI in a mixed group of mood-disordered patients, similar to healthy controls. An interesting investigation of specificity comes from a study by Braunstein-Bercovitz (2000), who found that inter-individual differences in levels of trait anxiety are as strong (and in fact stronger) a predictor of LI than schizotypy (also see BraunsteinBercovitz, this volume). The question that therefore arises is whether LI deficits in different psychiatric syndromes or traits arise for the same psychological, pharmacological, and neural mechanisms as those that are observed in schizotypy and schizophrenia. The psychological explanations of disrupted LI may even be different in acute-phase schizophrenia patients and high schizotypals. Genetic, pharmacological, and imaging studies might be able to address this issue. A number of methodological concerns have been raised regarding the LI paradigm over the last decades. The first studies of LI in schizophrenia and schizotypy used between-subject designs, i.e. randomly allocating participants in the first phase of the experiment to the PE or NPE conditions. While replicable findings have emerged from these studies (see Table 17.1), between-subject designs carry with them a number of disadvantages, ranging from issues of group differences in measured or not measured confounders to the truism that larger samples are required. More
Latent inhibition in schizophrenia and schizotypy
441
recent studies (e.g. Evans et al., 2007; Vaitl et al., 2002) have developed within-subject designs that circumvent these problems. Such tasks are likely to be beneficial in future work. Another point that has been made is that LI is a “window phenomenon” (Weiner, 2003), i.e. it is observed only with certain combinations of parameters in the pre-exposure phase and the conditioning phase, for example, the number and duration of pre-exposures in phase 1 or the intensity of the US in phase 2. Finally, of course, these variables interact with trait and state variables of the assessed organism. This situation means that the LI paradigm has to be carefully designed for its appropriate purpose. Finally, investigators should pay attention routinely to sex and hormonal status (Arad & Weiner, 2008), anxiety and stress (BraunsteinBercovitz, 2000; Braunstein-Bercovitz, this volume), drug and alcohol use (Stevens et al., 2007), and smoking history (Allan et al., 1995) in their research participants to add clarity to their findings and allow meaningful comparisons with other relevant studies. Acknowledgements Veena Kumari was supported by a Wellcome Senior Research Fellowship (067427/z/02). Ulrich Ettinger is funded by the German Research Foundation (ET 31/2–1). References Agid, O., Kapur, S., Warrington, L., Loebel, A., & Siu, C. (2008). Early onset of antipsychotic response in the treatment of acutely agitated patients with psychotic disorders. Schizophrenia Research, 102, 241–248. Allan, L. M., William, J. H., Wellman, N. A., et al. (1995). Effects of tobacco smoking, schizotypy, and number of pre-exposures on latent inhibition in healthy subjects. Personality and Individual Differences, 19, 893–902. Arad, M., & Weiner, I. (2008). Fluctuation of latent inhibition along the estrous cycle in the rat: modeling the cyclicity of symptoms in schizophrenic women? Psychoneuroendocrinology, 33, 1401–1410. Arranz, M. J., & de Leon, J. (2007). Pharmacogenetics and pharmacogenomics of schizophrenia: a review of last decade of research. Molecular Psychiatry, 12, 707–747. Baruch, I., Hemsley, D. R., & Gray, J. A. (1988a). Differential performance of acute and chronic schizophrenics in a latent inhibition task. Journal of Nervous & Mental Diseases, 176, 598–606. Baruch, I., Hemsley, D. R., & Gray, J. A. (1988b). Latent inhibition and “psychotic proneness” in normal subjects. Personality and Individual Differences, 9, 777–783. Bentall, R. P., Claridge, G. S., & Slade, P. D. (1989). The multidimensional nature of schizotypal traits: a factor analytic study with normal subjects. British Journal of Clinical Psychology, 28, 363–375. Bleuler, E. (1950). Dementia Praecox or the Group of Schizophrenias (J. Zinkin, Trans.). New York: International Universities Press. Braunstein-Bercovitz, H. (2000). Is the attentional dysfunction in schizotypy related to anxiety? Schizophrenia Research, 46, 255–267.
442
Veena Kumari and Ulrich Ettinger
Braunstein-Bercovitz, H., Hen, I., & Lubow, R. E. (2004). Masking task load modulates latent inhibition: support for a distraction model of irrelevant information processing by high schizotypals. Cognition and Emotion, 18, 1135–1144. Braunstein-Bercovitz, H., & Lubow, R. E. (1998). Are high-schizotypal normal participants distractible or limited in attentional resources? A study of latent inhibition as a function of masking task load and schizotypy level. Journal of Abnormal Psychology, 107, 659–670. Buckley, P. F., Miller, B. J., Lehrer, D. S., & Castle, D. J. (2009). Psychiatric comorbidities and schizophrenia. Schizophrenia Bulletin, 35, 383–402. Burch, G. S., Hemsley, D. R., & Joseph, M. H. (2004). Trials-to-criterion latent inhibition in humans as a function of stimulus pre-exposure and positiveschizotypy. British Journal of Psychology, 95, 179–196. Calkins, M. E., Iacono, W. G., & Ones, D. S. (2008). Eye movement dysfunction in first-degree relatives of patients with schizophrenia: a meta-analytic evaluation of candidate endophenotypes. Brain and Cognition, 68, 436–461. Carpenter, W. T., & Koenig, J. I. (2008). The evolution of drug development in schizophrenia: past issues and future opportunities. Neuropsychopharmacology, 33, 2061–2079. Chapman, L. J., Chapman, J. P., & Kwapil, T. R. (1995). Scales for the measurement of schizotypy. In A. Raine, T. Lencz, & S. A. Mednick (Eds.), Schizotypal Personality. New York: Cambridge University Press. Claridge, G. S., & Broks, P. (1984). Schizotypy and hemisphere function. I. Theoretical considerations and measurement of schizotypy. Personality and Individual Differences, 5, 633–648. Cohen, E., Sereni, N., Kaplan, O., et al. (2004). The relation between latent inhibition and symptom-types in young schizophrenics. Behavioural Brain Research, 149, 113–122. Costa, P. T., & McCrae, R. R. (1992). Revised NEO Personality Inventory: Professional Manual. Florida: Psychological Assessment Resources Inc. De la Casa, L. G., Ruiz, G., & Lubow, R. E. (1993). Latent inhibition and recall/ recognition of irrelevant stimuli as a function of pre-exposure duration in high and low psychotic-prone normal subjects. British Journal of Psychology, 84, 119–132. Della Casa, V., Ho¨fer, I., Weiner, I., & Feldon, J. (1999). Effects of smoking status and schizotypy on latent inhibition. Journal of Psychopharmacology, 13, 45–57. Dunn, L. A., Atwater, G., & Kilts, C. D. (1993). Effects of antipsychotic drugs on latent inhibition: sensitivity and specificity of an animal behavioural model of clinical drug action. Psychopharmacology, 112, 315–323. Evans, L. H., Gray, N. S., & Snowden, R. J. (2007). A new continuous withinparticipants latent inhibition task: examining associations with schizotypy dimensions, smoking status and gender. Biological Psychology, 74, 365–373. Eysenk, H. J. (1967). The Biological Basis of Personality. Springfield, IL: Charles C Thomas. Eysenck, H. J., & Eysenck, S. B. G. (1975). Manual of the Eysenck Personality Questionnaire. London: Hodder and Stoughton. Eysenck, H. J., & Eysenck, S. B. G. (1991). Manual of the Eysenck Personality Scales. London: Hodder and Stoughton.
Latent inhibition in schizophrenia and schizotypy
443
Fanous, A. H., Neale, M. C., Gardner, C. O., et al. (2007). Significant correlation in linkage signals from genome-wide scans of schizophrenia and schizotypy. Molecular Psychiatry, 12, 958–965. Gal, G., Barnea, Y., Biran, L., et al. (2009). Enhancement of latent inhibition in patients with chronic schizophrenia. Behavioural Brain Research, 197, 1–8. Gottesman, I. I., & Gould, T. D. (2003). The endophenotype concept in psychiatry: etymology and strategic intentions. American Journal of Psychiatry, 160, 636–645. Gray, J. A. (1998). Integrating schizophrenia. Schizophrenia Bulletin, 24, 249–266. Gray, J. A., Feldon, J., Rawlins, J. N. P., Hemsley, D. R., & Smith, A. D. (1991). The neuropsychology of schizophrenia. Behavioral and Brain Sciences, 14, 1–84. Gray, J. A., Kumari, V., Lawrence, N., & Young, A. M. J. (1999). Functions of the dopaminergic innervation of nucleus accumbens. Psychobiology, 27, 225–235. Gray, N. S., Fernandez, M., Williams, J., Ruddle, R. A., & Snowden, R. J. (2002). Which schizotypal dimensions abolish latent inhibition? British Journal of Clinical Psychology, 41, 271–284. Gray, N. S., Hemsley, D. R., & Gray, J. A. (1992a). Abolition of latent inhibition in acute, but not chronic, schizophrenics. Neurology Psychiatry and Brain Research, 1, 83–89. Gray, N. S., Pickering, A. D., Hemsley, D. R., Dawling, S., & Gray J. A. (1992b). Abolition of latent inhibition by a single 5 mg dose of d-amphetamine in man. Psychopharmacology, 107, 425–430. Gray, N. S., Pilowsky, L. S., Gray, J. A., & Kerwin, R. W. (1995). Latent inhibition in drug-naive schizophrenics: relationship to duration of illness and dopamine D2 binding using SPET. Schizophrenia Research, 17, 95–107. Guterman, Y., Josiassen, R. C., Bashore, T. E., Johnson, M., & Lubow, R. E. (1996). Latent inhibition effects reflected in event-related brain potentials in healthy controls and schizophrenics. Schizophrenia Research, 20, 315–326. Gruzelier, J. (2002). A Janusian perspective on the nature, development and structure of schizophrenia and schizotypy. Schizophrenia Research, 54, 95–103. Heinrichs, R. W., & Zakzanis, K. K. (1998). Neurocognitive deficit in schizophrenia: a quantitative review of the evidence. Neuropsychology, 12, 426–445. Hemsley, D. R. (1987). An experimental psychological model for schizophrenia. In H. Hafner, W. F. Gattaz & W. Janzavik W. (Eds.), Search for the Causes of Schizophrenia. Berlin: Springer, pp. 179–188. Hemsley, D. R. (1993). A simple (or simplistic?) cognitive model for schizophrenia. Behavior Research and Therapy, 31, 633–645. Ho¨fer, I., Della Casa, V., & Feldon, J. (1999). The interaction between schizotypy and latent inhibition: modulation by experimental parameters. Personality and Individual Differences, 26, 1075–1088. Hutton, S. B., & Ettinger, U. (2006). The antisaccade task as a research tool in psychopathology: a critical review. Psychophysiology, 43, 302–313. Javitt, D. C., Spencer, K. M., Thaker, G. K., Winterer, G., & Hajo´s, M. (2008). Neurophysiological biomarkers for drug development in schizophrenia. Nature Reviews Drug Discovery, 7, 68–83. Kaplan, O., Dar, R., Rosenthal, L., et al. (2006). Obsessive-compulsive disorder patients display enhanced latent inhibition on a visual search task. Behavior Research and Therapy, 44, 1137–1145.
444
Veena Kumari and Ulrich Ettinger
Kathmann, N., von Recum, S., Haag, C., & Engel, R. R. (2000). Electrophysiological evidence for reduced latent inhibition in schizophrenic patients. Schizophrenia Research, 45, 103–114. Kapur, S., Arenovich, T., Agid, O., et al. (2005). Evidence for onset of antipsychotic effects within the first 24 hours of treatment. American Journal of Psychiatry, 162, 939–946. Kapur, S., & Remington, G. (2001). Dopamine D(2) receptors and their role in atypical antipsychotic action: still necessary and may even be sufficient. Biological Psychiatry, 50, 873–83. Kay, S. R., Fiszbein, P. S., & Opler, L. A. (1987). The positive and negative syndrome scale (PANSS) for schizophrenia. Schizophrenia Bulletin, 13, 261–276. Kraepelin, E. (1919/1971). Dementia Praecox and Paraphrenia (R.M. Barclay, Trans.). Huntington, NY: Krieger. Kumari, V., Cotter, P., Mulligan, O., et al. (1999). Effect of d-amphetamine and haloperidol on latent inhibition in healthy male volunteers. Journal of Psychopharmacology, 13, 398–405. Kumari, V., Das, M., Zachariah, E., Ettinger, U., & Sharma, T. (2005). Reduced prepulse inhibition in unaffected siblings of schizophrenia patients. Psychophysiology, 42, 588–594. Larrison, A. L., Briand, K. A., & Sereno, A. B. (1999). Nicotine, caffeine, alcohol, and schizotypy. Personality and Individual Differences, 27, 101–108. Launay, G., & Slade, P. (1981). The measurement of hallucinatory predisposition in male and female prisoners. Personality and Individual Differences, 2, 221–234. Lenzenweger, M. F. (1998). Explorations in schizotypy and the psychometric high-risk paradigm. In L. J. Chapman, J. P. Chapman, & D. C. Fowler (Eds.), Progress in Experimental Personality and Psychopathological Research, vol. 16. New York: Springer. Leumann, L., Feldon, J., Vollenweider, F. X., & Ludewig, K. (2002). Effects of typical and atypical antipsychotics on prepulse inhibition and latent inhibition in chronic schizophrenia. Biological Psychiatry, 52, 729–739. Lipp, O. V., Siddle, D. A. T., & Arnold, S. L. (1994). Psychosis-proneness in a non-clinical sample. II. A multi-experimental study of ‘attentional malfunctioning’. Personality and Individual Differences, 15, 405–424. Lipp, O. V., & Vaitl, D. (1992). Latent inhibition in human Pavlovian differential conditioning: effect of additional stimulation after preexposure and relation to schizotypal traits. Personality and Individual Differences, 13, 1003–1012. Lubow, R. E. (2005). The construct validity of the animal-latent inhibition model of selective attention deficits in schizophrenia. Schizophrenia Bulletin, 31, 139–153. Lubow, R. E., & De la Casa, G. (2002). Latent inhibition as a function of schizotypality and gender: implications for schizophrenia. Biological Psychology, 59, 69–86. Lubow, R. E., Ingberg-Sachs, Y., Zalstein-Orda, N., & Gewirtz, J. C. (1992). Latent inhibition in low and high “psychotic-prone” normal subjects. Personality and Individual Differences, 13, 563–572. Lubow, R. E., Kaplan, O., Abramovich, P., Rudnick, A., & Laor, N. (2000). Visual search in schizophrenia: latent inhibition and novel pop-out effects. Schizophrenia Research, 45, 145–156.
Latent inhibition in schizophrenia and schizotypy
445
Lubow, R. E., Kaplan, O., & De la Casa, G. (2001). Performance on the visual search analog of latent inhibition is modulated by an interaction between schizotypy and gender. Schizophrenia Research, 52, 275–287. Lubow, R. E., Weiner, I., Scholossberg, A., & Baruch, I. (1987). Latent inhibition and schizophrenia. Bulletin of Psychonomic Society, 25, 464–467. Mangalore, R., & Knapp, M. (2007). Cost of schizophrenia in England. The Journal of Mental Health Policy and Economics, 10, 23–41. Marcelis, M., Suckling, J., Woodruff, P., et al. (2003). Searching for a structural endophenotype in psychosis using computational morphometry. Psychiatry Research, 122, 153–167. Mason, O., Claridge, G., & Jackson, M. (1995). New scales for the assessment of schizotypy. Personality and Individual Differences, 18, 7–13. McEvoy, J. P. (2007). The costs of schizophrenia. Journal of Clinical Psychiatry, 68, 4–7. Meehl, P. E. (1989). Schizotaxia revisited. Archives of General Psychiatry, 46, 935–944. Moser, P. C., Hitchcock, J. M., Lister, S., & Moran, P. M. (2000). The pharmacology of latent inhibition as an animal model of schizophrenia. Brain Research Reviews, 33, 275–307. Nuechterlein, K. H., & Dawson, M. E. (1984). Information processing and attentional functioning in the developmental course of schizophrenic disorders. Schizophrenia Bulletin, 10, 160–203. Overall, J. E., & Gorham, D. R. (1962). The Brief Psychiatric Rating Scale. Psychological Reports, 10, 799–812. Phillips, L. J., Yung, A. R., Yuen, H. P., Pantelis, C., & McGorry, P. D. (2002). Prediction and prevention of transition to psychosis in young people at incipient risk for schizophrenia. American Journal of Medical Genetics, 114, 929–937. Prasad, K. M., & Keshavan, M. S. (2008). Structural cerebral variations as useful endophenotypes in schizophrenia: do they help construct “extended endophenotypes”? Schizophrenia Bulletin, 34, 774–790. Raine, A. (1991). The SPQ: a scale for the assessment of schizotypal personality based on DSM-III-R criteria. Schizophrenia Bulletin, 17, 555–564. Raine, A. (2006). Schizotypal personality: neurodevelopmental and psychosocial trajectories. Annual Review of Clinical Psychology, 2, 291–326. Raine, A., Reynolds, C., Lencz, T., et al. (1994). Cognitive-perceptual, interpersonal, and disorganized features of schizotypal personality. Schizophrenia Bulletin, 20, 191–201. Rascle, C., Mazas, O., Vaiva, G., et al. (2001). Clinical features of latent inhibition in schizophrenia. Schizophrenia Research, 51, 149–161. Reichenberg, A., & Harvey, P. D. (2007). Neuropsychological impairments in schizophrenia: integration of performance-based and brain imaging findings. Psychological Bulletin, 133, 833–858. Reynolds, G. P. (2007). The impact of pharmacogenetics on the development and use of antipsychotic drugs. Drug Discovery Today, 12, 953–959. Ruob, C., Weiner, I., & Feldon, J. (1998). Haloperidol-induced potentiation of latent inhibition: interaction with parameters of conditioning. Behavioural Pharmacology, 9, 245–253. Rust, J. (1989). Rust Inventory of Schizotypal Cognitions Manual. Kent: Psychological Corporation Ltd.
446
Veena Kumari and Ulrich Ettinger
Salamone, J. D., Cousins, M. S., & Snyder, B. J. (1997). Behavioral functions of nucleus accumbens dopamine: empirical and conceptual problems with the anhedonia hypothesis. Neuroscience and Biobehavioral Reviews, 21, 341–359. Serra, A. M., Jones, S. H., Toone, B., & Gray, J. A. (2001). Impaired associative learning in chronic schizophrenics and their first-degree relatives: a study of latent inhibition and the Kamin blocking effect. Schizophrenia Research, 48, 273–289. Shadach, E., Feldon, J., & Weiner, I. (1999). Clozapine-induced potentiation of latent inhibition is due to its action in the conditioning stage: Implications for the mechanism of action of antipsychotic drugs. International Journal of Neuropsychopharmacology, 2, 283–291. Skelley, S. L., Goldberg, T. E., Egan, M. F., Weinberger, D. R., & Gold, J. M. (2008). Verbal and visual memory: characterizing the clinical and intermediate phenotype in schizophrenia. Schizophrenia Research, 105, 78–85. Snitz, B. E., Macdonald, A. W. 3rd, & Carter, C. S. (2006). Cognitive deficits in unaffected first-degree relatives of schizophrenia patients: a meta-analytic review of putative endophenotypes. Schizophrenia Bulletin, 32, 179–194. Stevens, A., Peschk, I., & Schwarz, J. (2007). Implicit learning, executive function and hedonic activity in chronic polydrug abusers, currently abstinent polydrug abusers and controls. Addiction, 102(6), 937–946. Strauss, J. S. (1969). Hallucinations and delusions as points on continua function. Archives of General Psychiatry, 21, 581–586. Swerdlow, N. R., Braff, D. L., Hartson, H., Perry, W., & Geyer, M. A. (1996). Latent inhibition in schizophrenia. Schizophrenia Research, 20, 91–103. Swerdlow, N. R., Hartston, H. J., & Hartman, P. L. (1999). Enhanced visual latent inhibition in obsessive-compulsive disorder. Biological Psychiatry, 45, 482–488. Swerdlow, N. R., Stephany, N., Wasserman, L. C., et al. (2003). Dopamine agonists disrupt visual latent inhibition in normal males using a within-subject paradigm. Psychopharmacology, 169, 314–320. Swerdlow, N. R., Stephany, N., Wasserman, L. C., et al. (2005). Intact visual latent inhibition in schizophrenia patients in a within-subject paradigm. Schizophrenia Research, 72, 169–183. Thaker, G. K. (2007). Schizophrenia endophenotypes as treatment targets. Expert Opinion on Therapeutic Targets, 11, 1189–206. Thornton, J. C., Dawe, S., Lee, C., et al. (1996). Effects of nicotine and amphetamine on latent inhibition in human subjects. Psychopharmacology, 127, 164–173. Tsakanikos, E. (2004). Latent inhibition, visual pop-out and schizotypy: is disruption of latent inhibition due to enhanced stimulus salience? Personality and Individual Differences, 37, 1347–1358. van Os, J., Hanssen, M., Bijl, R. V., & Ravelli, A. (2000). Strauss (1969) revisited: a psychosis continuum in the general population? Schizophrenia Research, 45, 11–20. Vaitl, D., Lipp, O., Bauer, U., et al. (2002). Latent inhibition and schizophrenia: Pavlovian conditioning of autonomic responses. Schizophrenia Research, 55, 147–158. van Haren, N. E., Bakker, S. C., & Kahn, R. S. (2008). Genes and structural brain imaging in schizophrenia. Current Opinion in Psychiatry, 21, 161–167.
Latent inhibition in schizophrenia and schizotypy
447
Venables, P. H., Wilkins, S., Mitchell, D. A., Raine, A., & Bailes, K. (1990). A scale for the measurement of schizotypy. Personality and Individual Differences, 11, 481–495. Vollema, M. G., & van den Bosch, R. J. (1995). The multidimensionality of schizotypy. Schizophrenia Bulletin, 21, 19–31. Whalley, H. C., Harris, J. C., & Lawrie, S. M. (2007). The neurobiological underpinnings of risk and conversion in relatives of patients with schizophrenia. International Review of Psychiatry, 19, 383–397. Whalley, H. C., Simonotto, E., Flett, S., et al. (2004). fMRI correlates of state and trait effects in subjects at genetically enhanced risk of schizophrenia. Brain, 127, 478–490. Williams, J. H., Wellman, N. A., Geaney, D. P., et al. (1997). Haloperidol enhances latent inhibition in visual tasks in healthy people. Psychopharmacology, 133, 262–268. Williams, J. H., Wellman, N. A., Geaney, D. P., et al. (1998). Reduced latent inhibition in people with schizophrenia: an effect of psychosis or of its treatment? British Journal of Psychiatry, 172, 243–249. Williams, J. H., Wellman, N. A., & Rawlins, J. N. (1996). Cannabis use correlates with schizotypy in healthy people. Addiction, 91, 869–877. Weiner, I. (2003). The ‘two-headed’ latent inhibition model of schizophrenia: modeling positive and negative symptoms and their treatment. Psychopharmacology, 169, 257–297. Wuthrich, V., & Bates, T. C. (2001). Schizotypy and latent inhibition: non-linear linkage between psychometric and cognitive markers. Personality and Individual Differences, 30(5), 783–798. Yogev, H., Sirota, P., Gutman, Y., & Hadar, U. (2004). Latent inhibition and overswitching in schizophrenia. Schizophrenia Bulletin, 30(4), 713–726.
18 A cautionary note about latent inhibition in schizophrenia: are we ignoring relevant information? Neal R. Swerdlow
Summary It is clear that the concept of latent inhibition (LI) and the notion that it might be abnormal in schizophrenia (SZ) patients have been powerful heuristic tools for crossspecies studies. Less clear has been the evidence that LI is actually abnormal in SZ, and if so, what the nature of such an abnormality might be, and which types of SZ patients might manifest it. We previously reported in two studies our ability to detect normal LI in a total of 88 SZ patients who successfully learned the non-preexposure task. Normal LI in these subjects could not be easily explained by peculiarities of the design of the LI task, or the characteristics of the study sample. Since submission of the last of these reports in 2004, we identified a total of three Medline papers in which LI was tested in SZ patients: one reported reduced LI only in unmedicated patients with predominant positive symptoms, another found elevated LI in only 6 out of 30 predominantly medicated patients who had the combination of low positive symptoms and high negative symptoms, and the third reported that LI was both reduced and elevated at different times within a single test, among medicated patients, unrelated to positive or negative symptoms. The belief that LI is abnormal in SZ persists, despite a paucity of clear, replicated, direct supportive evidence, and despite the presence of substantial relevant information that might lead us to conclude otherwise. Introduction Latent inhibition (LI) is the normal decrement in the rate of association of a CS and a UCS that occurs when the to-be-conditioned stimulus is first preexposed to the subject absent the UCS (Lubow & Moore, 1959; Lubow, 1973). In one conceptualization, LI occurs when a subject learns to ignore a preexposed (PE) stimulus that does not predict an important event (preexposure phase). When conditions change and the stimulus starts to predict an important event (test phase), the learned
Latent Inhibition: Cognition, Neuroscience and Applications to Schizophrenia, ed. R. E. Lubow and Ina Weiner. Published by Cambridge University Press # Cambridge University Press 2010.
448
A cautionary note about latent inhibition in schizophrenia
449
response must be overcome before a new CS–UCS association can be acquired. LI is seen operationally by a reduction or delay in the acquisition of a CS–UCS association in which the CS is the PE stimulus, compared to the acquisition in which the CS is a non-preexposed (NPE) stimulus. It is a powerful phenomenon that can be detected consistently across species, using a variety of different paradigms. In 1988, it was reported that acutely hospitalized SZ patients exhibit diminished LI, as tested with an auditory LI task (Baruch, Hemsley & Gray, 1988). This observation was later replicated by the same group (Gray, Hemsley & Gray, 1992). However, a number of attempts to reproduce these findings yielded mixed results. Using the identical paradigm, as well as a novel visual LI paradigm, our group detected generalized learning deficits but normal LI in a large sample of SZ patients (Swerdlow, Braff, Hartston et al., 1996). In the ensuing years, others reported either abnormal (Rascle, Mazas, Vaiva et al., 2001) or normal (Leumann, Feldon, Vollenweider & Ludewig, 2002) levels of LI in schizophrenia patients, or detected LI deficits in SZ only in one or another clinical subgroup or sex (Lubow, Kaplan, Abramovich et al., 2000; Lubow, Weiner, Schlossberg & Baruch, 1987), or alternatively, reported that LI deficits reflect the effects of antipsychotic medications (Williams, Wellman, Geaney et al., 1998) or learning deficits in NPE conditions (Serra, Jones, Toone & Gray, 2001). In some cases, LI deficits were detected only in unmedicated, drug-naive schizophrenia patients with illness duration of less than 12 months (n ¼ 6; Gray, Pilowsky, Gray & Kerwin, 1995), while in other cases LI deficits were seen only in medicated (but not unmedicated) patients (Williams et al., 1998). Despite the power and experimental elegance of the LI phenomenon, it seemed that, without parsing the study sample in various and often idiosyncratic ways, it was difficult to see LI deficits in a general sample of SZ patients. Furthermore, the parsing strategies required to demonstrate deficits in one subgroup or another differed from study to study. Many studies reporting LI deficits in SZ patients did so only in the context of generalized deficits in learning or attention. In studies using instrumental trialsto-criterion paradigms, evidence for such performance deficits is seen as impaired performance in NPE conditions (e.g. Gray et al., 1992; Swerdlow et al., 1996). In instances where only a post-hoc clinical subgroup of patients was identified to have LI deficits (e.g. in one case (Gray et al., 1995), unmedicated patients of less than 12-month illness duration or in another case (Lubow et al., 2000) female patients), compared to patients with normal LI, the LI-deficient subgroup exhibited impaired NPE performance. Because the amount of LI is calculated based on a difference in learning in NPE vs. PE conditions, generalized learning or attentional deficits under NPE conditions can yield false positive evidence of LI deficits. Another complexity in the LI literature reflects one strength of the LI phenomenon: LI can be (and is) measured many different ways in humans (cf. Lubow, 2005; Table 1). Each of these approaches differs to such a degree that there is no simple way to reconstruct the basis for contradictory findings, or even to determine whether the same construct is being measured across studies. For example, Guterman, Josiassen,
450
Neal R. Swerdlow
Bashore et al. (1996) used the contingent negative variation (CNV) to assess LI in a between-subject paradigm. These investigators reported CNV deficits in NPE SZ patients (d ¼ 1.87) that were substantially greater than the overall “LI” effect in controls (d ¼ 0.90) (Table 2, p. 321), and which may suggest more generalized attentional impairment. Kathmann, von Recum, Haag and Engel (2000) utilized the N100 event-related potential (ERP) to assess LI. The primary deficit in this report reflected N100 amplitudes to NPE stimuli in SZ patients, while N100 amplitudes to PE stimuli in other SZ patients were comparable to those of controls (Fig. 5, p. 111). In this same study, patients and controls demonstrated comparably robust LI in a go/ no-go task. Vaitl, Lipp, Bauer et al. (2002) developed a within-subject measure of Pavlovian conditioning of electrodermal responses, and reported blunted orienting responses to NPE stimuli, as well as absent LI in unmedicated but not medicated SZ patients. Of course, the difficulty in translating information across different LI paradigms in humans becomes magnified many-fold when trying to bridge findings across species: for example, LI measures in rodents commonly utilize aversive stimuli ranging from electric shocks to malaise-inducing drugs, and differ structurally from human paradigms to such a degree as to make cross-species comparisons precarious at best. Validation of LI deficits in SZ patients, and the reconciliation of differences in findings across studies, is also complicated by the diverse characteristics of the patients to whom these deficits are attributed. Using a within-subject auditory paradigm, Gray et al. (1995) reported that LI deficits in SZ subjects were detected only in a post-hoc subgroup of six drug-naive patients of “new onset” (less than 12 months since the onset of initial symptoms); this temporal cut-off seemed to be critically important, because a subject with an onset date of precisely 12 months actually exhibited the second highest amount of LI among the 13 patients in that study. Importantly, the two earlier reports of LI deficits in SZ in between-subject paradigms were not restricted to either drug-naive or “new-onset” patients (Baruch et al., 1988; Gray et al., 1992), nor were later positive reports (Rascle et al., 2001). In addition to the apparent importance of either or both the chronicity and medication state of the patient populations (defined differently in different reports), there may be complex effects of gender, and perhaps menstrual cycle phase, on schizophrenia-linked inhibitory deficits. For example, some forms of LI appear to exhibit sex-specific patterns (Lubow et al., 2000; Lubow & De la Casa, 2002; Lubow, Kaplan & De la Casa, 2001). One study reported that LI deficits in SZ patients (medicated, not “new onset”) in a visual, reaction time-based paradigm are specific to female patients (Lubow et al., 2000). However, in that paradigm, sex differences are primarily explained by impaired NPE performance in female patients, and male normal control subjects also failed to demonstrate LI (see Fig. 3, p. 151). Lastly, though it may seem on the surface a pedestrian exercise, it is worth considering the difficulty faced by investigators in interpreting and publishing a purely “negative” finding related to LI deficits in SZ patients. Given such an
A cautionary note about latent inhibition in schizophrenia
451
outcome, it is the appropriate instinct of self-critical scientists to pursue a rigorous “quality assurance” evaluation, to make certain that potential confounds or experimental error did not obscure the detection of LI deficits. In this regard, trying to systematically assess whether the sample characteristics are ones that should predict LI deficits is nearly an impossible task, based on the various and somewhat arbitrarily defined characteristics that have been linked to LI deficits in many different reports. Inevitably, peer-reviews ask for additional parsing of the clinical sample, and the resulting subgroup sizes shrink beyond meaningful numbers. Two other hurdles impede publication of findings suggesting normal LI in patients: the knee-jerk reluctance of reviewers and editors to publish “negative” findings, and the burden of proof of any report that fails to replicate a “positive” finding, even if the actual published data do not warrant such a clear “positive” characterization. The net outcome of this process is that many “negative” findings, i.e. normal LI in SZ patients, are never pursued or published, and those reports that are published are more likely to include positive findings that are “forced”, via an exhaustive post-hoc search for subgroups of patients that exhibit abnormal performance. As a result, the literature becomes skewed towards increasingly complex and evanescent patterns of abnormal LI in SZ subpopulations.
Normal LI in SZ patients using a within-subject visual LI paradigm To address some of the difficulties in interpreting between-subject LI paradigms, we (Swerdlow, Stephany, Wasserman et al., 2003) developed a within-subject visual LI paradigm that detects LI in normal male subjects, in a way that is sensitive to disruption by acute treatment with dopamine agonists. When we applied this paradigm to normal men and women, LI exhibited no sex differences or menstrual cyclicity (Swerdlow, Stephany, Wasserman et al., 2005). Compared to normals, SZ patients exhibited learning deficits with both PE and NPE stimuli. Despite these generalized deficits, both acutely hospitalized patients and stable outpatients with SZ exhibited robust LI, as evidenced by significantly faster learning with NPE than PE stimuli. The fact that SZ patients exhibited normal LI in this paradigm could not be attributed to a number of factors, including the metric for assessing LI, as the data were analyzed and presented using a number of different strategies. When performance was assessed using a trials-to-criterion metric, the effect size (PE vs. NPE) for normal control subjects at their learning criterion trial was 0.32, consistent with one auditory within-subject task reported by Gray, Snowden, Peoples et al. (2003), while the effect size for the SZ group at their criterion trial was 0.69. This difference did not reflect significantly greater LI in SZ patients, but at least supported the conclusion that SZ patients did not exhibit less LI compared to normal control subjects.
452
Neal R. Swerdlow
Post-hoc analyses suggested that intact LI in SZ patients could not be explained by simple variables such as age or acuity of illness (acute inpatient vs. stable outpatient), or the disproportionate number of females among control vs. SZ groups. No patient in our 2005 study was within the first 12 months of illness, and this sample also did not include a sufficient number of unmedicated patients to assess the possibility that LI had been “normalized” by psychotropic medications: 16 out of 20 SZ patients were taking atypical antipsychotics at the time of testing. To summarize our findings through 2004, in two substantive studies, using a variety of test designs, analytic strategies, and post-hoc parsing of study samples, our data demonstrated that SZ patients exhibit generalized learning deficits, and intact LI. Collectively, we have “failed” to detect LI deficits – or, stated in the affirmative, detected normal levels of LI in SZ patients – using three LI paradigms, in a total of 88 SZ patients who successfully learned the NPE task.
More recent evidence Since January 2004, when we submitted our report on these findings of normal visual LI in SZ patients, three papers have been listed on Medline, under the search term, “latent inhibition schizophrenia”, that report tests of SZ patients in LI paradigms: Cohen, Sereni, Kaplan et al. (2004), Yogev, Sirota, Gutman & Hadar (2004) and Gal, Barnea, Biran et al. (2009). Many reports studied rodent models of LI, and others utilized measures of “learned irrelevance” in humans, but this review of more recent evidence will focus specifically on the total of three Medline-identified reports of LI measures in schizophrenia patients that have been published since January 2004. It is very possible that other relevant reports since that date have appeared outside the Medline database. Cohen et al. (2004) used a within-subject visual search task to measure reaction time LI in 30 healthy controls and 30 young (13–21 yrs) largely medicated (29/30) previously hospitalized SZ patients. Both patients (P < 0.02) and matched controls (P < 0.025) demonstrated statistically significant LI, and there was no significant interaction of group condition, i.e. no statistical evidence that LI was different across the groups. Despite this, a post-hoc procedure was undertaken: individual LI scores were computed for each subject, and scores were ranked, and divided into five groups of 12 subjects. A chi-square was then performed based on the claim that data were not normally distributed, and the outcome was interpreted to mean that there was a difference in the distribution of control vs. SZ scores across the five ranked subgroups. Further post-hoc analyses based on the presence of positive and negative symptoms revealed that neither positive nor negative symptoms were related to this apparent distributional difference of LI scores, but that SZ patients characterized by the combination of high negative symptoms and low positive symptoms (n ¼ 6) were the only clinical subgroup to differ significantly from controls. Importantly, these six
A cautionary note about latent inhibition in schizophrenia
453
subjects exhibited significantly greater LI, when a parametric statistic (t-test) was used to compare them to the 30 control subjects. The subgroup showing the next greatest amount of LI – nearly 3 times the amount shown by controls – was the subgroup characterized by the exact opposite clinical profile: high positive symptoms and low negative symptoms (n ¼ 6; p. 117, Table 1). Most importantly, SZ patients as a group exhibited normal LI. Only after a very forced series of post-hoc comparisons did the authors conclude that LI was deviant among the 20% of their subjects with a biologically undefined combination of clinical characteristics; in so doing, they ignored the fact that the patient subgroup most similar in performance to the “deviant” one (and certainly not statistically separable from it) exhibited the exact opposite set of clinical characteristics. Yogev et al. (2004) used a between-subject, mixed auditory–visual reaction time attentional task to assess LI in 41 adult, non-medicated, previously hospitalized patients in the midst of another hospitalization. Patients were characterized by predominantly positive (n ¼ 28) or negative (n ¼ 13) symptoms, and were assigned to either PE (n ¼ 16 positive, n ¼ 7 negative) or NPE (n ¼ 12 positive, n ¼ 6 negative) conditions. Two measures were used to assess LI, and in both cases LI was statistically significant across all groups (main effect of condition), and no significant interaction of condition diagnostic group was reported. Curiously, the authors claim that LI was impaired among the positive symptoms patients, a claim that is not supported by the necessary group condition interaction. LI was not enhanced in patients with negative symptoms, who instead appear to demonstrate a level of LI that was arithmetically mid-way between positive symptom patients and normal controls (Fig. 2, p. 720). Gal et al. (2009) studied LI using a within-subject visual reaction time-based task in 19 stable medicated chronic SZ outpatients. Across the entire test session, both patients and matched controls exhibited significant LI that did not differ significantly across groups. This was supported by a significant effect of condition, but no significant interaction of group condition. When the test session was divided into two halves, controls demonstrated significant LI on the first half but not the second half, while SZ patients exhibited significant LI in the second half but not the first half. This difference was supported by a three-way interaction of group condition test half, and a series of post-hoc comparisons. The amount of LI differed significantly between groups in the second half of the test (patients had more than controls), but not in the first half of the test. Test results in SZ patients were not related to the levels of positive or negative symptoms. These complex results were interpreted to suggest an abnormal “persistence” of LI in SZ patients. It is difficult to understand how LI “persisted” in SZ patients, since (according to the authors) it was not present in patients during the first half of the test. Perhaps one might view this more precisely as a “late appearing” LI, rather than “persistent” LI. The results present other challenges, beyond this semantic issue. LI in the controls was very weak: in the first half of the test, when LI in patients was not
454
Neal R. Swerdlow
statistically significant, LI in controls did not differ significantly from LI in patients. Inspection of Figure 2 reveals mean control PE and NPE reaction times that are nearly identical in the first vs. second half of the test, with greater variability in the second half accounting for the failure to achieve statistical separation. In fact, according to Figure 3, LI in controls did not differ significantly between the first and second half of the test, i.e., it “persisted” across the entire test. Thus, it is difficult to know exactly what is abnormal about the pattern of LI in SZ patients in this study. If there is something abnormal, at the least it appears to be unrelated to the levels of negative or positive symptoms. Another post-January 2004 report of potential relevance to the LI puzzle comes from Barrett, Bell, Watson & King (2004), who demonstrated the disruption of some (but not all) forms of visual (but not auditory) LI in healthy male volunteers by some (risperidone) but not other (chlorpromazine) antipsychotic medications. Thus, discrepancies across studies with auditory vs. visual LI might reflect a differential impact of antipsychotics, rather than differences in the LI per se or its relationship to SZ.
Discussion Existing data warrant skepticism about the generalized implications of LI abnormalities in SZ. These abnormalities remain elusive, “moving targets” – variably observed or not observed among specific subgroups of patients that are defined quite differently across studies, using different LI paradigms. In many cases, large samples of SZ subjects demonstrate intact LI, but these samples are then divided by various, idiosyncratic criteria, to identify small subgroups in which LI is either deficient or excessive. These subgroups typically represent a small minority of the study sample, yet their response pattern is the one often described in the title of the paper. It would be appropriate for such post-hoc findings to be used as the basis for a targeted replication strategy, to determine whether the same outcome is identified in a fully powered sample limited to the apparently critical clinical characteristics. Lacking such validation, perhaps these preliminary observations should be the “footnote”, and not the title. We are in danger of ignoring relevant information if we “explain away” negative findings based on experimental differences, and embrace positive findings, despite reasons to be skeptical of them. As discussed above, the number of reports that fail to detect LI abnormalities in SZ patients, or that report such deficits only in a post-hoc subgroup, or in the presence of generalized learning deficits, outnumber those that actually detect LI abnormalities in SZ patients. Given the substantial effort spent in the pursuit of reliable, quantitative SZ phenotypes, it will be very important for investigators to maintain a degree of circumspection regarding the robustness and generalizability of LI abnormalities in SZ patients.
A cautionary note about latent inhibition in schizophrenia
455
Acknowledgements Supported by MH59803. N.R.S. has had grant support from Pfizer Pharmaceuticals and Allergan, Inc., and is a paid Consultant to Sanofi/Aventis. Some data described in this review, and issues discussed thereof, have been reported in previously published work from our group, as referenced in the text.
References Barrett, S. L., Bell, R., Watson, D., & King, D. J. (2004). Effects of amisulpride, risperidone and chlorpromazine on auditory and visual latent inhibition, prepulse inhibition, executive function and eye movements in healthy volunteers. Journal of Psychopharmacology, 18, 156–172. Baruch, I., Hemsley, D. R., & Gray, J. A. (1988). Differential performance of acute and chronic schizophrenics in a latent inhibition task. Journal of Nervous and Mental Disorders, 176, 598–606. Cohen, E., Sereni, N., Kaplan, O., et al. (2004). The relation between latent inhibition and symptom-types in young schizophrenics. Behavioural Brain Research, 149, 113–122. Escobar, M., Oberling, P., & Miller, R. R. (2002). Associative deficit accounts of disrupted latent inhibition and blocking in schizophrenia. Neuroscience and Biobehavioral Reviews, 26, 203–216. Gal, G., Barnea, Y., Biran, L., et al. (2009). Enhancement of latent inhibition in patients with chronic schizophrenia. Behavioural Brain Research, 197, 1–8. Gray, N. S., Hemsley, D. R., & Gray, J. A. (1992). Abolition of latent inhibition in acute, but not chronic schizophrenics. Neurology Psychiatry and Brain Research, 1, 1–7. Gray, N. S., Pilowsky, L. S., Gray, J. A., & Kerwin, R. W. (1995). Latent inhibition in drug naive schizophrenics: relationship to duration of illness and dopamine D2 binding using SPET. Schizophrenia Research, 17, 95–107. Gray, N. S., Snowden, R. J., Peoples, M., Hemsley, D. R., & Gray, J. A. (2003). A demonstration of within-subjects latent inhibition in the human: limitations and advantages. Behavioural Brain Research 138, 1–8. Guterman, Y., Josiassen, R., Bashore, R. E., Johnson, M., & Lubow, R. E. (1996). Latent inhibition effects reflected in event-related brain potentials in healthy controls and schizophrenics. Schizophrenia Research, 20, 315–326. Kathmann, N., von Recum, S., Haag, C., & Engel, R. R. (2000). Electrophysiological evidence for reduced latent inhibition in schizophrenic patients. Schizophrenia Research, 45, 103–114. Leumann, L., Feldon, J., Vollenweider, F. X., & Ludewig, K. (2002). Effects of typical and atypical antipsychotics on prepulse inhibition and latent inhibition in chronic schizophrenia. Biological Psychiatry, 52, 729–739. Lubow, R. E. (1973). Latent inhibition. Psychological Bulletin, 79, 398–407. Lubow, R. E. (2005). The construct validity of the animal-latent inhibition model of selective attention deficits in schizophrenia. Schizophrenia Bulletin, 31, 139–153. Lubow, R. E., & De la Casa, G. (2002). Latent inhibition as a function of schizotypality and gender: implications for schizophrenia. Biological Psychology, 59, 69–86. Lubow, R. E., Kaplan, O., Abramovich, P., Rudnick, A., & Laor, N. (2000). Visual search in schizophrenia, latent inhibition and novel pop-out effects. Schizophrenia Research, 45, 145–156.
456
Neal R. Swerdlow
Lubow, R. E., Kaplan, O., & De la Casa, G. (2001). Performance on the visual search analog of latent inhibition is modulated by an interaction between schizotypy and gender. Schizophrenia Research 52, 275–87. Lubow, R. E., & Moore, A. U. (1959). Latent inhibition: the effect of nonreinforced exposure to the conditioned stimulus. Journal of Comparative and Physiological Psychology, 52, 415–419. Lubow, R. E., Weiner, I., Schlossberg, A., & Baruch, I. (1987). Latent inhibition and schizophrenia. Bulletin of the Psychonomic Society, 25, 464–467. Rascle, C., Mazas, O., Vaiva, G., et al. (2001). Clinical features of latent inhibition in schizophrenia. Schizophrenia Research, 51, 149–161. Serra, A. M., Jones, S. H., Toone, B., & Gray, J. A. (2001). Impaired associative learning in chronic schizophrenics and their first-degree relatives: a study of latent inhibition and the Kamin blocking effect. Schizophrenia Research, 48, 273–289. Swerdlow, N. R., Braff, D. L., Hartston, H., Perry, W., & Geyer, M. A. (1996). Latent inhibition in schizophrenia. Schizophrenia Research, 20, 91–103. Swerdlow, N. R., Stephany, N., Wasserman, L. C., et al. (2003). Dopamine agonists disrupt visual latent inhibition in normal males using a within-subject paradigm. Psychopharmacology, 169, 314–320. Swerdlow, N. R., Stephany, N., Wasserman, L. C., et al. (2005). Intact visual latent inhibition in schizophrenia patients in a within-subject paradigm. Schizophrenia Research, 72, 169–183. Vaitl, D., Lipp, O., Bauer, U., et al. (2002). Latent inhibition and schizophrenia: Pavlovian conditioning of autonomic responses. Schizophrenia Research, 55, 147–158. Weiner I. (2003). The “two-headed” latent inhibition model of schizophrenia: modeling positive and negative symptoms and their treatment. Psychopharmacology, 169, 257–297. Williams, J. H., Wellman, N. A., Geaney, D. P., et al. (1998). Reduced latent inhibition in people with schizophrenia: an effect of psychosis or of its treatment. British Journal of Psychiatry, 172, 243–249. Yogev, H., Sirota, P., Gutman, Y., & Hadar, U. (2004). Latent inhibition and overswitching in schizophrenia. Schizophrenia Bulletin, 30, 713–726.
19 Latent inhibition as a function of anxiety and stress: implications for schizophrenia Hedva Braunstein-Bercovitz
Selective attention is a process which requires that an individual devotes attention to a task-relevant stimulus and processes it, while ignoring a distracting-irrelevant stimulus. The ability to selectively attend and respond effectively to relevant stimuli is essential for adaptive behavior and normal cognitive functioning, since, according to capacity approaches (e.g., Kahneman, 1973; Navon & Gopher, 1979), attention exists in limited amounts and might be allocated to irrelevant as well as relevant information (e.g., Gopher, 1992). That normal selective attention is essential for adaptive behavior is supported by the literature which shows a relationship between psychopathology and the impairment of selective attentional processes. The most prominent example of such a relationship is the attentional deficit of schizophrenics and “normal” schizotypals. Studies have shown that, as compared to normal individuals, schizophrenics and schizotypals are distracted by irrelevant stimuli (see below). In addition, there is evidence that individuals characterized by high levels of anxiety also are distracted by irrelevant stimuli, although most studies suggest that anxiety leads to an attentional bias only for irrelevant threat-related stimuli. Surprisingly, despite the similarity of selective attention impairments in schizophrenia/schizotypy and anxiety, there are hardly any studies that have tested whether the anxiety that characterizes schizophrenia/schizotypy accounts for (even only partially) the difficulties schizophrenics/schizotypals have in ignoring irrelevant information. This chapter will present data about selective attentional dysfunction in groups rated low and high on self-report anxiety and schizotypy scales and in situations that elicit anxiety (stress). The following two questions will be addressed: (a) does anxiety impair the ability to filter out irrelevant stimuli in general, rather than only threat-related stimuli? (b) Is the selective attentional impairment of schizophrenics/schizotypals due to the anxiety that accompanies those conditions?
Latent Inhibition: Cognition, Neuroscience and Applications to Schizophrenia, ed. R. E. Lubow and Ina Weiner. Published by Cambridge University Press # Cambridge University Press 2010.
457
458
Hedva Braunstein-Bercovitz
In order to examine these questions, the chapter will focus on latent inhibition (LI; Lubow, 1989) as a tool for measuring selective attention. LI was chosen because its relationship with schizophrenia and schizotypy has been widely investigated, both pharmacologically and neuropsychologically, and it has been integrated into a neuropsychological model for schizophrenia (e.g., Gray, 1998; Gray et al., 1991; Weiner, 2003). However, findings relevant to selective attentional dysfunctions in schizophrenia/schizotypy and anxiety from negative priming (Stroop, 1935; Tipper, 1985) experiments also will be presented.
Latent inhibition as a tool for measuring selective attentional dysfunctions LI is defined as slower learning to a previously irrelevant preexposed stimulus than to a novel one (Lubow, 1989). In a conventional two-phase LI procedure, during the preexposure phase, one group is exposed to repeated non-reinforced presentations of an irrelevant to-be-target stimulus. In the subsequent test phase, learning an association between the preexposed stimulus and another event is slower in the preexposed group (PE), for whom the target is familiar, than in a non-preexposed group (NPE), for whom the target is novel. In humans, in order to obtain LI, the participant must engage in a primary task which demands attentional resources (masking task), and thereby diverts attention from the preexposed stimulus (for review, see Lubow & Gewirtz, 1995). As such, LI appears to promote stimulus selection, providing a learning bias in favor of potentially important stimuli by degrading those stimuli that, in the past, have been registered as irrelevant or inconsequential. As suggested, since LI magnitude is a function of the amount of attention allocated to the irrelevant stimulus (Braunstein-Bercovitz & Lubow, 1998b), LI can be considered to be the product of a selective attention process. On this basis, LI has been widely discussed as a promising tool in areas of experimental personality research and psychopathology that are characterized by dysfunctions in such a process. LI has been studied as a function of personality traits such as creativity (Burch, Hemsley, Pavelis & Corr, 2006; Carson, Peterson & Higgins, 2003; also see Carson, this volume), and field-dependency (Braunstein-Bercovitz, 2003b), and in a number of pathological populations, including those in which the primary symptoms are psychological/behavioral and those in which they are neurological/motor. With regard to psychopathology, there is much evidence for dysfunctional cognitive processing that accompanies the primary symptomology, frequently pointing to attentional deficits. Thus, LI has been investigated in participants with Parkinson’s disease (Lubow, Dressler & Kaplan, 1999), Tourette’s syndrome (Swerdlow, Magulac, Filion & Zinner, 1996), clinically diagnosed anxiety disorders (see below), obsessive-compulsive disorder (Kaplan, Dar, Rosenthal et al., 2006; Swerdlow, Hartston & Hartman, 1999), and schizophrenia (see below), as well as in normal
Latent inhibition as a function of anxiety and stress
459
populations divided on the basis of self-report questionnaire scores for anxiety (see below), type-A personality (De la Casa, 1994), and schizotypality/psychosis proneness (see below). Relatedly, there are LI studies that have manipulated stress (Braunstein-Bercovitz et al., 2001), and even those that have administered psychoactive drugs such as amphetamine (Gray, N.S., et al., 1992; Kumari, Cotter, Mulligen et al., 1999; Thornton et al., 1996) and haloperidol (Kumari et al., 1999; Williams, Wellman, Geaney et al., 1996, 1997) to normal participants. Of the various groups and disorders for which LI has been assessed, only schizotypality, schizophrenia, and anxiety have a sufficient number of studies to warrant extended discussion.
Selective attentional dysfunctions in schizophrenics and “normal” schizotypals In addition to LI, selective attention in schizophrenia has been studied using negative priming and Stroop paradigms. Negative priming (Tipper, 1985) is defined as poorer performance on a previous distractor that had to be ignored than to a novel target, and it is attenuated if too much attention is devoted to the distractors (e.g., Lavie, 1995, 2000). In a Stroop task (Stroop, 1935), individuals who are asked to name the color of a word (the relevant dimension) that spells a conflicting color name (e.g., the word red in green letters) experience interference from the spelled word (the irrelevant dimension) in the speed and often the accuracy of responding. Stroop interference increases with the amount of attention devoted to the spelled word (MacLeod, 1991). These three paradigms assess distraction of attention by irrelevant stimuli or dimensions. Therefore, the relatively high degree of distractibility in schizophrenics should result in attenuated LI and NP, and enhancement of the Stroop effect. Indeed, acute non-medicated schizophrenics show less LI than controls (Baruch et al., 1988a; Gal, Mendlovic, Bloch et al., 2005; Gray, N. S., Hemsley & Gray, 1992; Gray, N. S., Pilowsky, Gray & Kerwin, 1995; Lubow, 2005; Lubow, & Kaplan, 2005; Swerdlow, Stephany, Wasserman et al., 2005). Similarly, negative priming is disrupted in acute schizophrenics as compared to normal controls (e.g., Beech, Powell, McWilliams & Claridge, 1989; MacQueen, Galway, Goldberg & Tipper, 2003; Salo, Robertson & Nordahl, 1996; Williams, 1996; Zabal & Buchner, 2006). In addition, with Stroop, schizophrenics show more interference from the irrelevant dimension than controls (e.g., Abramczyk, Jordan & Hegel, 1983; Buchanan et al., 1994; Cantor-Graae, Warkentin & Nilsson, 1995; Molina, Montes, De Luxa´n et al., 2008; Perlstein, Carter, Barch & Baird, 1998; Verdoux, Magnin & Bourgeosis, 1995). Similar research has been conducted with normal participants who differ on selfreport schizotypal questionnaires. The rationale for such a tactic is based on evidence from family studies which indicate that the genetic vulnerability to schizophrenia may be manifested in non-psychotic individuals as a schizophrenia-like personality.
460
Hedva Braunstein-Bercovitz
Relatedly, scores from these self-report questionnaires have been treated within the framework of a dimensional model of psychotic phenomena (e.g., Chapman, Chapman, Kwapil et al., 1994; Claridge, 1997, 1994, 1985; Claridge & Broks, 1984; Eysenck, 1952; Eysenck & Eysenck, 1976; Vollema & van den Bosch, 1995). Such a model is based on the premise that psychotic tendencies exist on a continuum, with one extreme being the well-adapted normal, and the other a hospitalized patient group such as schizophrenics (e.g., Claridge & Broks, 1984; Eysenck & Eysenck, 1976). If one accepts the assumption of continuity, then one should expect to find that high-schizotypal “normals” will exhibit attenuated LI and negative priming and enhanced Stroop interference as compared to low-schizotypal “normals”. With regard to LI this has been a quite consistent finding (e.g., Allan et al., 1995; Baruch et al., 1988b; Braunstein-Bercovitz & Lubow, 1998a; Della Casa, Hoefer, Weiner & Feldon, 1999; Evans, Gray & Snowden, 2007; Hoefer, Della Casa & Feldon, 1998; Lipp, Arnold & Siddle, 1994; Swerdlow, Braff, Hartston et al., 1996; Tsakanikos & Reed, 2004). The effect has been achieved with several different learning tasks and with a variety of psychosis-proneness and schizotypy questionnaires. Recent studies that have examined task variables that modulate the effects of schizotypality on LI report that LI in low- and high-psychosis-prone “normals” is differentially affected by the load characteristics of the masking task (BraunsteinBercovitz & Lubow, 1998a; Braunstein-Bercovitz, Hen & Lubow 2004; Hoefer, Della Casa & Feldon, 1998). With regard to negative priming, the effect is reduced in high- as compared to lowschizotypals (e.g., Beech, Baylis, Smithson & Claridge, 1989; Beech & Claridge, 1987; Ferraro & Okerlund, 1996; Moritz, Andresen, Probsthein et al., 2000; Moritz, Mass & Junk, 1998; Williams, 1995). Relatedly, Stroop interference is enhanced in high- as compared to low-schizotypals (Cimino & Haywood, 2008; Mohanty et al., 2005). As with schizophrenia, the LI, negative priming, and Stroop effects with highschizotypals have been interpreted in terms of relatively high degrees of distractibility (e.g., Fox, 1995). Of the various psychometric questionnaires, the Schizotypy Personality Scale (STA, Claridge & Broks, 1984) and Schizotypal Personality Questionnaire (SPQ; Raine, 1991) appear to reflect one aspect of schizotypy that modulates responding on all three test paradigms. Furthermore, the parallel deficits from LI, negative priming, and Stroop tests in schizophrenic patients and highschizotypal normals support the “continuity” hypothesis. As such, the understanding of attentional/cognitive dysfunctions in schizophrenia can be advanced by taking advantage of the fact that with high-schizotypal normals, the predisposition to schizophrenia is isolated from possible confounding factors, such as hospitalization and medication (Mednick & McNeil, 1973). In summary, the results from schizophrenic patients and “normal” schizotypals with LI, negative priming, and Stroop that have been interpreted in terms of high distractibility and which reflect a dysfunction in selective attention have been regarded as a specific characteristic of schizophrenia and as a primary deficit that is
Latent inhibition as a function of anxiety and stress
461
related to other major cognitive disturbances in schizophrenics, such as distorted thinking and perception (e.g., Anscombe, 1987; Franke, Maier, Hardt et al., 1994; Frith, 1979; Gray et al., 1991; Knight, 1984; Maher, 1983; Mirsky & Duncan, 1986). However, the notion that distractibility by irrelevant stimuli is exclusively associated with schizophrenia and schizotypy contradicts other theoretical approaches and experimental data, which suggest that such distractibility is also related to high levels of anxiety.
Selective attentional processes and anxiety/stress Eysenck (1992) has argued that anxious individuals are particularly sensitive to signs of threat or potential danger. As a consequence, selective attention in high anxious individuals is biased towards threat-related stimuli, and they devote more attentional resources to threat-related than to neutral stimuli. Indeed, there is evidence to support such a claim. High-anxiety individuals are overly distracted by threat-related irrelevant stimuli on emotional Stroop ( judging the ink color of negative and neutral words) and negative priming tasks. For example, patients diagnosed as having a generalized anxiety disorder (e.g., Eysenck, MacLeod & Mathews, 1987; Mathews & MacLeod, 1985; Mogg & Bradley, 2005; Mogg, Bradely, Millar & White, 1995; Mogg, Mathews & Weinman, 1989; Price & Mohlman, 2007), obsessive-compulsive disorder (e.g., Amir, Cobb & Morrison, 2008; Enright & Beech & Claridge, 1995; Foa, Ilai, McCarthy et al., 1993; Lavy, van Oppen & van den Hout, 1994; MacDonald, Antony, MacLeod & Swinson, 1999; McNally, Wilhelm, Buhlmann & Shin, 2001), and post-traumatic-stress disorder (Cassiday, McNally & Zeitlin, 1992; Ehlers & Clark, 2000; Thrasher, Dalgleish & Yule, 1994; Vythilingam et al., 2007), as well as individuals with high self-report trait-anxiety (e.g. Berner & Maier, 2004; Dalgleish, 1995; Fox, 1993, 1994; Jansson & Lundh, 2006; MaCleod & Hagan, 1992; van den Hout, Tenney, Huygens et al., 1995) are more distracted by irrelevant threat-related stimuli than normal controls (for review see Bar-Haim, Lamy, Pergamin et al., 2007). With stress, an experimental condition designed to elicit anxiety, results are less clear. Some Stroop studies have reported decreased interference from the irrelevant dimension with high as compared to low stress (e.g., Folkard & Greeman, 1974; Kofman, Meiran, Greenberg et al., 2006; O’Malley & Gallas, 1977), others have reported increased interference in the Stroop task and NP effects (e.g., BraunsteinBercovitz, 2003b; Hartly & Shirely, 1976; Keinan et al., 1999; Roelofs, Bakvis, Hermans et al., 2007), while yet others have not found any differences in Stroop effects as a function of stress (Brand, Schneider & Arntz, 1995; Brand, Verspui & Oving, 1997). However, in addition to a bias towards threat-related stimuli, there is evidence that points to a general inability of anxious/stressed individuals to maintain attentional
462
Hedva Braunstein-Bercovitz
focus (Enright & Beech, 1993; Fox, 1993, 1994; Mathews, May, Mogg & Eysenck, 1990; for a review, see Eysenck, Derakshan, Santos & Calvo, 2007). It appears that high anxious individuals may be characterized by a general distractibility by irrelevant information, rather than a specific bias towards threat-related information, although threatening stimuli seem to amplify such a distraction. The LI–anxiety studies are particularly important because LI uses irrelevant, meaningless stimuli, which, of course, are non-threatening, and therefore allows one to examine whether anxiety accounts for a general breakdown of selective attentional processes. In addition, since LI experiments are an integral part of established models for schizophrenia (e.g., Lubow, 2005; Weiner, 2003), it is important to determine whether the disrupted LI effect is specific to that and related disorders, or whether they are nonspecific effects associated with high anxiety. Unfortunately, there are only a few relevant anxiety studies. Braunstein-Bercovitz (2000) examined LI as a function of trait anxiety and schizotypy levels. During the preexposure stage, all participants engaged in a masking task, which consisted of same/different judgments to pairs of letters (TT, TL, LT, LL). For the NPE groups, only the masking task stimuli were presented during preexposure. For the PE groups, participants were presented with the masking task stimuli and the currently irrelevant to-be-target stimuli, which consisted of a pair of identical, irregularly shaped polygons. During the test phase, LI was assessed with a rule-learning procedure, requiring the acquisition of an association between the previously to-be-target stimulus (the irregularly shaped polygons) and a new event (a change in a counter value). Trait anxiety was measured by the State-Trait-Anxiety Inventory (STAI; Spielberger, Gorsuch & Lushene, 1970), and schizotypy by SPQ (Raine, 1991). LI was attenuated in participants with high as compared to low trait-anxiety scores (see below, the results with SPQ). In another set of experiments, Braunstein-Bercovitz et al. (2001) examined whether stress disrupts LI. Although anxiety and stress are not identical constructs, they are related. In brief, stress is produced in situations characterized by high levels of uncertainty or unpredictability that call for a coping response (e.g., Levine & Ursin, 1980). Stress does not necessarily result in experienced anxiety; however, biochemical responses to stress, such as increases in catecholamine excretion, are similar to the ones observed with anxiety states (e.g., Frankenhaeuser, 1978). Braunstein-Bercovitz et al. (2001) used the same procedure that was described above for measuring the affect of trait-anxiety on LI. In the first experiment, stress was induced in the laboratory by threats to participants’ (college students) self-esteem with a difficult number-series completion test that was purportedly related to intelligence (high stress condition). For the low stress group, the number-series task was easy and not related to intelligence (Keinan, Friedland, Kahneman & Roth, 1999). Subjects in the lowbut not high-stress condition exhibited LI. The second experiment obtained similar results when stress was induced in a non-laboratory situation. For participants in the high-stress condition, the LI task was presented as an integral part of an employment
Latent inhibition as a function of anxiety and stress
463
interview procedure. Participants in the low-stress condition were told that the LI task was only for research purposes, and that their performance would not affect job placement decisions. More recently (Braunstein-Bercovitz, 2003a), negative priming was examined as a function of stress manipulation (the same procedure as in Braunstein-Bercovitz et al., 2001, first experiment), and supported the results obtained with LI, i.e., negative priming was also disrupted by stress. Gibbons, Bengs and Rammsayer (1998) also provided support for the modulating effect of anxiety-related personality traits on LI. They found increased LI in participants with a combination of low-Psychoticism and high-Neuroticism scores as compared to the remaining sample. A visual search version of LI also found attenuated LI in high-anxiety children (Lubow, Toren, Laor & Kaplan, 2000). Indirect evidence for anxiety-disrupted LI comes from a study that reported reduced LI in children diagnosed as attention-deficit/hyperactive (Lubow, Braunstein-Bercovitz, Blumenthal et al., 2005), a disorder accompanied by high levels of anxiety. A variety of animal studies provide additional confirmation that anxiety and stress affect selective attention, as reflected in disrupted LI. First, stressed rats show increased dopaminergic activity in the mesolimbic system (Salamone, Cousins & Snyder, 1997), a treatment that abolishes LI (Gray et al., 1995; Weiner, 1990; Weiner & Feldon, 1997). Interestingly, Hellman, Crider and Solomon (1983) have shown that behavioral stress (tail pinch) can be substituted for amphetamine to produce attenuated LI. Second, animals injected with corticosterone, which is secreted in response to stress (Shalev, Feldon & Weiner, 1998b), or are prenatally exposed to dexamethasone (Hauser, Feldon & Pryce, 2006) also display disrupted LI, as do rats who have been subjected to behavioral stress manipulations, such as early maternal separation (e.g., Bethus, Lemaire, Lhomme & Goodall, 2005; Lehmann, Stohr & Feldon, 2000), and early non-handling (e.g., Shalev, Feldon & Weiner, 1998a), an effect that is particularly strong in males (e.g., Weiner, Schnabel, Lubow & Feldon, 1985). Furthermore, several lines of evidence from experiments with animals (e.g., Bartoszyk, 1998; Espejo, 1997; Gendreau, Petitto, Gariepy & Lewis, 1998; Glavin, 1993; Talalaenko et al., 1994; Troncoso, Osaki, Mason et al., 2003; Yoshioka, Matsumoto, Togashi & Saito, 1996) as well as with humans (e.g., McIvor et al., 1996; Nutt, Bell & Malizia, 1998; Peroutka, Price, Wilhoit & Jones, 1998) point to an involvement of dopamine in the regulation of anxiety (although the underlying neurobiological mechanisms are still to be described). Finally, anxiogenic drugs, like drugs that increase dopamine, disrupt LI (e.g., Feldon & Weiner, 1989; Lacroix, Spinelli, Broersen & Feldon, 2000; Mongeau, Marcello, Andersen & Pani, 2007). Nevertheless, it should be noted that neuroleptic drugs, typical and atypical, that increase dopamine binding have been shown to enhance LI. As yet, there is no such evidence for anxiolytic drugs (for review see, Moser et al., 2000). To summarize, both human and the animal studies indicate that high levels of anxiety and stress disrupt LI (for review, see Braunstein-Bercovitz et al., 2002). These studies suggest that anxiety/stress enhances distractibility, so that anxious/stressed
464
Hedva Braunstein-Bercovitz
individuals devote more attentional resources to irrelevant information than normal controls, irrespective of the threat-value capacity of that information. This contention is in accord with attentional control theory (for review, see Eysenck, Derakshan, Santos & Calvo, 2007), which suggests that anxiety impairs the efficient functioning of the goal-directed attentional system by decreasing attentional control, and that the attentional bias to threat-related stimuli is secondary. In addition, the findings presented above suggest that the attentional processes that govern LI are impaired by trait-anxiety, as well as by conditions, such as stress, that evoke situational anxiety. Therefore, rather than viewing attention-impairing anxiety only as a chronic, trait-like vulnerability, one should consider that transient states of anxiety either acting alone or interacting with various pathologies will also produce attentional dysfunctions.
Does anxiety/stress account for the selective attentional dysfunction in schizophrenics and schizotypals? The attenuation of LI in high-schizotypals and schizophrenics is most commonly attributed to the relatively high degree of distractibility in these groups. As such, these individuals continue to allocate attentional resources to the preexposed irrelevant stimulus, thereby precluding a loss of attention or conditioning of inattention and the consequent LI effect. However, as already suggested, diminished inhibition of attention to irrelevant stimuli also may be a consequence of the high levels of anxiety that accompany schizophrenia/schizotypy. Indeed, SPQ schizotypality scores (Raine, 1991) and STAI anxiety scores (Spielberger, Gorsuch & Lushene, 1970) scores are highly correlated (Braunstein-Bercovitz, 2000), as are STA-schizotypal and STAI-Trait scores (Gibbons & Rammsayer, 1998; Stelton & Ferraro, 2008), as well as schizotypy scale scores and Eysenck’s Neuroticism-scale scores (e.g., Bentall Claridge & Slade, 1989; Eysenck 1992; Gibbons & Rammsayer 1998; Kendler & Hewitt, 1992; Montag & Levin, 1992; Muntaner, Garcia-Sevilla, Fernandes & Torrubia, 1988). Furthermore, several studies have found correlations between types of symptoms in schizophrenic patients and levels of anxiety (e.g., Huppert, Weiss, Lim et al., 2001; Lysaker, Davis, Lightfoot et al., 2005; Lysaker & Salyers, 2007; Norman, Malla, Cortese & Diaz, 1998). For example, Lysaker et al. (2005) found that schizophrenic patients with low positive symptoms and low state anxiety coped better with stress than their higher-scoring counterparts. Recently, Lysaker and Salyers (2007) reported that higher levels of anxiety were associated with greater hallucinations, withdrawal, depression, hopelessness, better insight and poorer function. In addition, as already described, LI is mediated by dopaminergic activity (for reviews, see Gray, 1998; Moser, Hitchcock, Lister & Moran, 2000; Weiner & Feldon, 1997), and both schizotypy (Caplan & Guthrie, 1994; Silver, 1994, 1995) and anxiety (e.g., McIvor et al., 1996; Nutt, Bell & Malizia, 1998; Peroutka, Price, Wilhoit & Jones, 1998) also are characterized by increased dopaminergic activity.
Latent inhibition as a function of anxiety and stress
465
The high correlations of schizotypal scores with anxiety and neuroticism scores suggest that schizotypal scales may contain an anxiety factor (or vice versa). This, together with data that indicate dopaminergic involvement in schizotypality, and that anxious as well as schizotypal individuals are distracted by irrelevant stimuli as measured by LI, negative priming, and Stroop, supports the notion that an anxiety component of the disorder may account for selective attentional deficits in highschizotypal “normals” (and by extension in schizophrenia), and that the diminished LI in schizotypals may be a result of the high anxiety which accompanies that personality type. To investigate this hypothesis, Braunstein-Bercovitz (2000) conducted a factor analysis of SPQ scores, correlated the factor scores with trait-anxiety scores (STAI), and examined LI as a function of these factors. The analysis produced two factors. Factor-I (composed of social anxiety, odd or eccentric behavior, no close friends, odd speech, constricted affect, and suspiciousness sub-scales) was correlated with trait-anxiety scores (r1 ¼ 0.53), and was therefore labeled “anxiety-loaded” (AL). Factor-II (composed of ideas of reference, odd beliefs or magical thinking, and unusual perceptual experiences sub-scales) was labeled “perceptual-disorganization” (PD). PD was also correlated with anxiety (r2 ¼ 0.24). However, r1 was significantly larger than r2. Braunstein-Bercovitz (2000) also examined LI as a function of the AL and PD factors, trait-anxiety (STAI), and schizotypy (SPQ). The results confirmed previous studies. LI was attenuated in high- as compared to low-schizotypals. LI also was disrupted in participants with high scores on trait-anxiety (see above), and on the AL factor. LI was not affected by the PD factor scores. In addition, a regression analysis indicated that both trait-anxiety scores (STAI) and schizotypal scores (SPQ) independently accounted for LI disruption in high-schizotypals, but that the contribution of STAI was stronger. It would seem, then, that the anxiety component of schizotypy, more than the perceptual-disorganization (schizophrenia-like) component, accounts for the attentional dysfunction in high-schizotypals, and for the greater than normal distraction by irrelevant stimuli. LI deficits, then, appear not to represent a specific marker for schizophrenia/ schizotypy. Instead, LI deficits may be a contribution of the heightened anxieties that accompany many different pathologies. Such a proposition may be relevant to the nature of the impaired selective attention in schizotypy and schizophrenia, and, especially, it may contribute to those models of schizophrenia that incorporate stress in the description of the development of schizophrenic symptoms, as in diathetic models (e.g., Jones & Fernyhough, 2007; Nuechterlein & Dawson, 1984; Nuechterlein et al., 1992; Walker & Diforio, 1997). The notion that stress, which elicits anxiety, may exacerbate schizophrenic symptoms and episodes and lead to relapse (for reviews, see Norman & Malla, 1993a; 1993b; Walker & Diforio, 1997) receives additional support from the studies of LI, with schizotypy, anxiety and stress.
466
Hedva Braunstein-Bercovitz
Summary and conclusions At the close of the 1980s, for the first time, diminished LI was reported in non-treated schizophrenics, and in normal participants with high-schizotypal scores. Concurrently, neuroscience investigations were undertaken to determine the physiological basis of LI (for early reviews, see Gray et al., 1991; Weiner, 1990; more recently, Gray, 1998; Moser, Hitchcock, Lister & Moran, 2000; Weiner, 2003; Weiner & Feldon, 1997), while more behaviorally oriented research focused primarily on the relation between schizotypy and LI. For the many reasons already discussed, we were led to expect relatively simple correlations between scores on schizotypal scales and LI, ones that might signify that attenuated LI was a marker for psychosis-proneness. However, current research suggests that anxiety, trait and state, which cut across many different psychopathologies, may modulate LI. The diminished LI in high-schizotypal normals and in a variety of pathologies may be more related to the high levels of anxiety that accompany these states than to any of their component symptoms. The fact that stressful factors have been repeatedly implicated in precipitating and/ or exacerbating psychotic episodes (e.g., Breier, Wolkowitz & Pickar, 1991; Jones & Fernyhough, 2007; Meehl, 1962, 1990; Nuechterlein & Dawson, 1984; Nuechterlein, Dawson & Green, 1994; Nuechterlein et al., 1992) makes the putative relationship between anxiety, LI disruption, and high-schizotypality/schizophrenia even stronger. To conclude, this chapter suggests that anxiety impairs the ability of an individual to gate out irrelevant information, resulting in high distractibility and a difficulty in focusing attention on those aspects of a situation that are task-relevant. Thus, different psychological states, pathological such as schizophrenia, and normal, such as high-schizotypy, both of which are characterized by high levels of anxiety, may result in attenuated LI (depending on task demands). Although such a possibility calls for a reevaluation of claims that attentional dysfunctions in schizophrenia and schizotypy, at least as indexed by LI deficits, are specific markers of those states, more research about LI and anxiety is needed. Future LI studies should include personality traits that go beyond psychosis-proneness, with a particular emphasis on anxiety. Moreover, LI data from patients who are clinically diagnosed with anxiety disorders would contribute to assessing the validity of the above hypothesis.
References Abramczyk, R. R., Jordan, D. E., & Hegel, M. (1983). Reverse Stroop effect in the performance of schizophrenics. Perceptual and Motor Skills, 56, 99–106. Allan, L. M., Williams, J. H., Wellman, N. A., et al. (1995). Effects of tobacco smoking, schizotypy and number of preexposures on latent inhibition in healthy subjects. Personality and Individual Differences, 19, 893–902. Amir, N., Cobb, M., & Morrison, A. S. (2008). Threat processing in obsessive-compulsive disorder: evidence from a modified negative priming task. Behaviour Research and Therapy, 46, 728–736.
Latent inhibition as a function of anxiety and stress
467
Anscombe, R. (1987). The disorder of consciousness in schizophrenia. Schizophrenia Bulletin, 13, 241–260. Bar-Haim, Y., Lamy, D., Pergamin, L., et al. (2007). Threat-related attentional bias in anxious and nonanxious individuals: a meta-analytic study. Psychological Bulletin, 133, 1–24. Bartoszyk, G. D. (1998). Anxiolytic effects of dopamine receptor ligands: involvement of dopamine autoreceptors. Life Science, 62, 649–663. Baruch, I., Hemsley, D. R., & Gray, J. A. (1988a). Differential performance of acute and chronic schizophrenics in a latent inhibition task. Journal of Nervous and Mental Disease, 176, 598–606. Baruch, I., Hemsley, D. R., & Gray, J. E. (1988b). Latent inhibition and “psychotic proneness” in normal subjects. Personality and Individual Differences, 9, 777–784. Beech, A., & Claridge, G. (1987). Individual differences in negative priming: relation with schizotypal personality traits. British Journal of Clinical Psychology, 78, 349–356. Beech, A., Baylis, G. C., Smithson, P., & Claridge, G. (1989). Individual differences in schizotypy as reflected in measures of cognitive inhibition. British Journal of Clinical Psychology, 78, 117–129. Beech, A. R., Powell, T. J., McWilliams, J., & Claridge, G. S. (1989). Evidence of reduced cognitive inhibition in schizophrenia. British Journal of Clinical Psychology, 28, 109–116. Bentall, R. P., Claridge, G. S., & Slade, P. D. (1989). The multidimensional nature of schizotypal traits: a factor analytic study with normal subjects. British Journal of Clinical Psychology, 28, 363–375. Berner, M. P., & Maier, M. A. (2004). The direction of affective priming as a function of trait anxiety when naming target words with regular and irregular pronunciation. Experimental Psychology, 51, 180–190. Bethus, I., Lemaire, V., Lhomme, M., & Goodall, G. (2005). Does prenatal stress affect latent inhibition? It depends on the gender. Behavioural Brain Research, 158, 331–338. Brand, N., Schneider, N., & Arntz, P. (1995). Information processing efficiency and noise: interactions with personal rigidity. Personality and Individual Differences, 18, 571–579. Brand, N., Verspui, L., & Oving, A. (1997). Induced mood and selective attention. Perceptual and Motor Skills, 84, 455–463. Braunstein-Bercovitz, H. (2000). Is the attentional dysfunction in schizotypy related to anxiety? Schizophrenia Research, 46, 255–267. Braunstein-Bercovitz, H. (2003a). Does stress enhance or impair selective attention? The effects of stress and perceptual load on negative priming. Anxiety, Stress and Coping, 16, 345–357. Braunstein-Bercovitz, H. (2003b). The modulation of latent inhibition by field-dependency: is it related to the attentional dysfunction in schizotypy? Personality and Individual Differences, 35, 1719–1729. Braunstein-Bercovitz, H., Dimentman-Ashkenazi, I., & Lubow, R. E. (2001). Stress affects the selection of relevant from irrelevant stimuli. Emotion, 1, 172–182. Braunstein-Bercovitz, H., Hen, I. &., & Lubow, R. E. (2004). Masking task load modulates latent inhibition: support for a distraction model of irrelevant information processing by high schizotypal participants. Cognition and Emotion, 18, 1135–1144.
468
Hedva Braunstein-Bercovitz
Braunstein-Bercovitz, H. &., & Lubow, R. E. (1998a). Are high schizotypal normal participants distractible or limited in attentional resources? A study of latent inhibition as a function of masking task load and schizotypy level. Journal of Abnormal Psychology, 107, 659–670. Braunstein-Bercovitz, H. &., & Lubow, R. E. (1998b). Latent inhibition as a function of modulation of attention to the preexposed irrelevant stimulus. Learning and Motivation, 29, 261–279. Braunstein-Bercovitz, H., Rammsayer, T., Gibbons, H., & Lubow, R. E. (2002). Latent inhibition deficits in high-schizotypal normals: symptom-specific or anxiety-related? Schizophrenia Research, 53, 109–121. Breier, A., Wolkowitz, O. M., & Pickar, D. (1991). Stress and schizophrenia. In C. A. Tamminga, & S. C. Schultz (Eds.), Schizophrenia Research. New York: Raven Press, pp. 141–152. Buchanan, R. W., Strauss, M. E., Kirkpatrick, B., et al. (1994). Neuropsychological impairments in deficit versus nondeficit forms of schizophrenia. Archives of General Psychiatry, 51, 804–811. Burch, G. S. J., Hemsley, D. R., Pavelis, C., & Corr, P. J. (2006). Personality, creativity and latent inhibition. European Journal of Personality, 20, 107–122. Cantor-Graae, E., Warkentin, S., & Nilsson, A. (1995). Neuropsychological assessment of schizophrenic patients during a psychotic episode: persistent cognitive deficit? Acta Psychiatrica Scandinavica, 91, 283–288. Caplan, R., & Guthrie, D. (1994). Blink rate in childhood schizophrenia spectrum disorder. Biological Psychiatry, 35, 228–234. Carson, S. H., Peterson, J. B., & Higgins, D. M. (2003). Decreased latent inhibition is associated with increased creative achievement in high-functioning individuals. Journal of Personality and Social Psychology, 85, 499–506. Cassiday, K. L., McNally, R. J., & Zeitlin, S. B. (1992). Cognitive processing of trauma cues in rape victims with post traumatic stress disorder. Cognitive Therapy and Research, 16, 283–295. Chapman, L. J., Chapman, J. P., Kwapil, T. R., Eckblad, M., & Zinser, O. (1994). Putatively psychosis-prone subjects 10 years later. Journal of Abnormal Psychology, 103, 171–183. Cimino, M., & Haywood, M. (2008). Inhibition and facilitation in schizotypy. Journal of Clinical and Experimental Neuropsychology, 30, 187–198. Claridge, G. S. (1985). Origins of Mental Illness. Oxford: Blackwell. Claridge, G. S. (1994). Single indicator of risk for schizophrenia: probable fact or likely myth? Schizophrenia Bulletin, 20, 151–168. Claridge, G. S. (Ed.) (1997). Schizotypy: Implications for Illness and Health. Oxford: Oxford University Press. Claridge, G. S., & Broks, P. (1984). Schizotypy and hemisphere function – I: theoretical considerations and the measurement of schizotypy. Personality and Individual Differences, 5, 633–648. Dalgleish, T. (1995). Performance on the emotional Stroop task in groups of anxious, expert, and control subjects: a comparison of computer and card presentation formats. Cognition and Emotion, 9, 341–362. De la Casa, L. G. (1994). Latent inhibition and recall of irrelevant stimuli in Type-A and Type-B subjects. Revista de Psicologia General y Aplicada, 47, 289–300. Della Casa, V., Hoefer, I., Weiner, I., & Feldon, J. (1999). Effects of smoking status and schizotypy on latent inhibition. Journal of Psychopharmacology, 13, 362–368.
Latent inhibition as a function of anxiety and stress
469
Ehlers, A., & Clark, D. M. (2000). A cognitive model of posttraumatic stress disorder. Behaviour Research and Therapy, 38, 319–345. Enright, S. J., & Beech, A. R. (1993). Reduced cognitive inhibition in obsessivecompulsive disorder. British Journal of Clinical Psychology, 32, 67–74. Enright, S. J., Beech, A. R., & Claridge, G. S. (1995). A further investigation of cognitive inhibition in obsessive-compulsive disorder and other anxiety disorders. Personality and Individual Differences, 19, 535–542. Espejo, E. F. (1997). Selective dopamine depletion with the medial prefrontal cortex induces anxiogenic-like effects in rats placed on the elevated plus maze. Brain Research, 762, 281–284. Evans, L. H., Gray, N. S., & Snowden, R. J. (2007). A new continuous withinparticipants latent inhibition task: examining associations with schizotypy dimensions, smoking status and gender. Biological Psychology, 74, 365–373. Eysenck, H. J. (1952). Schizothymia-cyclothymia as a dimension of personality. II. Experimental. Journal of Personality, 20, 345–384. Eysenck, H. J. (1992). The definition and measurement of psychoticism. Personality and Individual Differences, 13, 757–785. Eysenck, M. W., Derakshan, N., Santos, R., & Calvo, M. G. (2007). Anxiety and cognitive performance: attentional control theory. Emotion, 7, 336–353. Eysenck, H. J., & Eysenck, S. B. G. (1976). Psychoticism as a Dimension of Personality. London: Hodder and Stoughton. Eysenck, M. W., MacLeod, C., & Mathews, A. (1987). Cognitive functioning and anxiety. Psychological Research, 49, 189–195. Feldon, J., & Weiner, I. (1989). Abolition of the acquisition but not the expression of latent inhibition by chlordiazepoxide in rats. Pharmacology, Biochemistry and Behavior, 32, 123–127. Ferraro, F. R., & Okerlund, M. (1996). Failure to inhibit irrelevant information in non-clinical schizotypal individuals. Journal of Clinical Psychology, 52, 389–394. Foa, E. B., Ilai, D., McCarthy, P. R., Shoyer, B., & Murdock, T. (1993). Information processing in obsessive-compulsive disorder. Cognitive Therapy and Research, 17, 173–189. Folkard, S., & Greeman, A. L. (1974). Salience, induced muscle tension, and the ability to ignore irrelevant information. Quarterly Journal of Experimental Psychology, 26, 360–367. Fox, E. (1993). Attentional bias in anxiety: selective or not? Behaviour Research and Therapy, 31, 487–493. Fox, E. (1994). Attentional bias in anxiety: a defective inhibition hypothesis. Cognition and Emotion, 8, 165–195. Fox, E. (1995). Negative priming from ignored distractors in visual selection: a review. Psychonomic Bulletin & Review, 21, 45–173. Franke, P., Maier, W., Hardt, J., Hain, C., & Cornblat, B. (1994). Attentional abilities and measures of schizotypy: the variation and covariation in schizophrenic patients, their siblings, and normal control subjects. Psychiatry Research, 54, 259–272. Frankenhaeuser, M. (1978). Psychoneuroendocrine approaches to the study of emotion as related to stress and coping. Nebraska Symposium on Motivation, 26, 123–16. Frith, C. D. (1979). Consciousness, information processing and schizophrenia. British Journal of Psychiatry, 134, 225–235.
470
Hedva Braunstein-Bercovitz
Gal, G., Mendlovic, S., Bloch, Y., et al. (2005). Learned irrelevance is disrupted in first-episode but not chronic schizophrenia patients. Behavioural Brain Research, 159, 267–275. Gendreau, P. L., Petitto, J. M., Gariepy, J. L., & Lewis, M. H. (1998). D2-like dopamine receptor mediation of social-emotional reactivity in a mouse model of anxiety: strain and experience effects. Neuropsychopharmacology, 18, 210–221. Gibbons, H., Bengs, C., & Rammsayer, T. H. (1998). Prozeaeigenschaften latenter hemmung im Paradigma “Visual Search” [Process characteristics of latent inhibition in a visual search paradigm]. In H. Lachnit, A. Jacobs, & F. Rosler (Eds.), Experimentelle Psychologie. Lengerich: Pabst Science Publishers, p. 93. Gibbons, H., & Rammsayer, T. H. (1998). Persoenlichkeitsmerkmale und latente Hemmung [Personality traits and latent inhibition]. Zeitung fuer Differentielle und Diagnostische Psychologie, 19, 20. Glavin, G. B. (1993). Vulnerability to stress ulcerogenesis in rats differing in anxiety: a dopaminergic correlate. Journal of Physiology, 87, 239–243. Gopher, D. (1992). The skill of attention control: acquisition and execution of attention strategies. In D. E. Meyer & S. Kornblum (Eds.), Attention and Performance, vol. 14. Cambridge, MA: MIT Press, pp. 187–204. Gray, J. A. (1998). Integrating schizophrenia. Schizophrenia Bulletin, 24, 249–266. Gray, J. A., Feldon, J., Rawlins, J. N. P., Hemsley, D. R., & Smith, A. D. (1991). The neuropsychology of schizophrenia. Behavioral and Brain Sciences, 14, 1–84. Gray, J. A., Joseph, M. H., Hemsley, D. R., et al. (1995). The role of mesolimbic dopaminergic and retrohippocampal afferents to the nucleus accumbens in latent inhibition: implications for schizophrenia. Behavioural Brain Research, 71, 19–31. Gray, N. S., Hemsley, D. R., & Gray, J. A. (1992). Abolition of latent inhibition in acute, but not chronic, schizophrenics. Neurology, Psychiatry and Brain Research, 1, 83–89. Gray, N. S., Pilowsky, L. S., Gray, J. A., & Kerwin, R. W. (1995). Latent inhibition in drug-naive schizophrenics: relationship to duration of illness and dopamine D2 binding using SPET. Schizophrenia Research, 17, 95–107. Hartly, L. R., & Shirely, E. (1976). Color-name interference at different times of day. Journal of Applied Psychology, 61, 119–122. Hauser, J., Feldon, J., & Pryce, C. R. (2006). Prenatal dexamethasone exposure, postnatal development, and adulthood prepulse inhibition and latent inhibition in Wistar rats. Behavioural Brain Research, 175, 51–61. Hellman, P. A., Crider, A., & Solomon, P. R. (1983). Interaction of tail-pressure stress and d-amphetamine in disruption of the rat’s ability to ignore an irrelevant stimulus. Behavioral Neuroscience, 97, 1017–1021. Hoefer, I., Della Casa, V., & Feldon, J. (1998). The interaction between schizotypy and latent inhibition: modulation by experimental parameters. Personality and Individual Differences, 26, 1075–1088. Huppert, J. D., Weiss, K. A., Lim, R., Pratt, S., & Smith, T. E. (2001). Quality of life in schizophrenia: contributions of anxiety and depression. Schizophrenia Research, 51, 171–180. Jansson, B., & Lundh, L. G. (2006). The interactive role of worried mood and trait anxiety in the selective processing of subliminally presented threat words. Personality and Individual Differences, 41, 1195–1204.
Latent inhibition as a function of anxiety and stress
471
Jones, S. R., & Fernyhough, C. (2007). A new look at the neural diathesis-stress model of schizophrenia: the primacy of social-evaluative and uncontrollable situations. Schizophrenia Bulletin, 33, 1171–1177. Kahneman, D. (1973). Attention and Effort. Englewood Cliffs, NJ: Prentice Hall. Kaplan, O., Dar, R., Rosenthal, L., et al. (2006). Obsessive-compulsive disorder patients display enhanced latent inhibition on a visual search task. Behaviour Research and Therapy, 44, 1137–1145. Keinan, G., Friedland, N., Kahneman, D., & Roth, D. (1999). The effect of stress on the suppression of erroneous competing responses. Anxiety, Stress and Coping, 12, 455–476. Kendler, K. S., & Hewitt, J. (1992). The structure of self-report schizotypy in twins. Journal of Personality Disorders, 6, 1–17. Knight, R. A. (1984). Converging models of cognitive deficit in schizophrenia. In W. Spaulding & J. K. Cole (Eds.), Theories of Schizophrenia and Psychosis. Lincoln, NE: University of Nebraska Press. Kofman, O., Meiran, N., Greenberg, E., Balas, M., & Cohen, H. (2006). Enhanced performance on executive functions associated with examination stress: evidence from task-switching and Stroop paradigms. Cognition & Emotion, 20, 577–595. Kumari, V., Cotter, P. A., Mulligen, O. F., et al. (1999). Effects of d-amphetamine and haloperidol on latent inhibition in healthy male volunteers. Journal of Psychopharmacology, 13, 398–405. Lacroix, L., Spinelli, S., Broersen, L. M., & Feldon, J. (2000). Blockade of latent inhibition following increase or decrease of GABA(A) transmission. Pharmacology, Biochemistry and Behavior, 66, 893–901. Lavie, N. (1995). Perceptual load as a necessary condition for selective attention. Journal of Experimental Psychology: Human Perception and Performance, 21, 451–468. Lavie, N. (2000). Selective attention and cognitive control: dissociating attentional functions through different types of load. In S. Monsell & J. Driver (Eds.), Attention and Performance XVIII. Cambridge, MA: MIT Press. Lavy, E. H., van Oppen, P., & van den Hout, M. A. (1994). Selective processing of emotional information in obsessive compulsive disorder. Behaviour Research and Therapy, 32, 243–246. Lehmann, J., Stohr, T., & Feldon, Y. (2000). Long-term effects of prenatal stress and postnatal maternal separation on emotionality and attentional processes. Behavioural Brain Research, 107, 133–144. Levine, S., & Ursin, H. (1980). Coping and Health. New York: Plenum Press. Lipp, O. V., Arnold, S. L., & Siddle, D. A. T. (1994). Psychosis proneness in a non-clinical sample I: a psychometric study. Personality and Individual Differences, 17, 395–404. Lubow, R. E. (1989). Latent Inhibition and Conditioned Attention Theory. New York: Cambridge University Press. Lubow, R. E. (2005). Construct validity of the animal latent inhibition model of selective attention deficits in schizophrenia. Schizophrenia Bulletin, 31, 139–153. Lubow, R. E., Braunstein-Bercovitz, H., Blumenthal, O., Kaplan, O., & Toren P. (2005). Latent inhibition and asymmetrical visual-spatial attention in children with ADHD. Child Neuropsychology, 11, 445–457. Lubow, R. E., Dressler, R. &., & Kaplan, O. (1999). The effects of target and distractor familiarity on visual search in de novo Parkinson’s disease patients: latent inhibition and novel pop-out. Neuropsychology, 13, 415–423.
472
Hedva Braunstein-Bercovitz
Lubow, R. E., & Gewirtz, J. C. (1995). Latent inhibition in humans: data, theory, and implications for schizophrenia. Psychological Bulletin, 117, 87–103. Lubow, R. E., & Kaplan, O. (2005). The visual search analogue of latent inhibition: implications for theories of irrelevant stimulus processing in normal and schizophrenic groups. Psychonomic Bulletin & Review, 12, 224–243. Lubow, R. E., Toren, P., Laor, N., & Kaplan, O. (2000). The effects of target and distractor familiarity on visual search in anxious children: latent inhibition and novel pop-out. Journal of Anxiety Disorders, 14, 41–56. Lysaker, P. H., Davis, L. W., Lightfoot, J., Hunter, N., & Stasburger, A. (2005). Association of neurocognition, anxiety, positive and negative symptoms with coping preference in schizophrenia spectrum disorders. Schizophrenia Research, 80, 163–171. Lysaker, P. H., & Salyers, M. P. (2007). Anxiety symptoms in schizophrenia spectrum disorders: associations with social function, positive and negative symptoms, hope and trauma history. Acta Psychiatrica Scandinavica, 116, 290–298. MacDonald, P. A., Antony, M. M., MacLeod, C. M., & Swinson, R. P. (1999). Negative priming for obsessive-compulsive checkers and noncheckers. Journal of Abnormal Psychology, 108, 679–686. MacLeod, C., & Hagan, R. (1992). Individual differences in the selective processing of threatening information, and emotional responses to a stressful life event. Behavior Research and Therapy, 30, 151–161. MacLeod, C. M. (1991). Half a century research on the Stroop effect: an integrative review. Psychological Bulletin, 109, 163–203. MacQueen, G. M., Galway, T., Goldberg, J. O., & Tipper, S. P. (2003). Impaired distractor inhibition in patients with schizophrenia on a negative priming task. Psychological Medicine, 33, 121–129. Maher, B. A. (1983). A tentative theory of schizophrenia utterance. In B. A. Maher & W. B. Maher (Eds.), Progress in Experimental Personality Research, vol. 12. New York: Academic Press. Mathews, A., & MacLeod, C. (1985). Selective processing of threat cues in anxiety states. Behavior, Research and Therapy, 23, 563–569. Mathews, A., May, J., Mogg, K., & Eysenck, M. W. (1990). Attentional bias in anxiety: selective search or defective filtering. Journal of Abnormal Psychology, 99, 166–173. McIvor, R. J., Davies, R. A., Wiech, A., et al. (1996). The growth hormone response to apomorphine at 4 days postpartum in women with history of major depression. Journal of Affective Disorders, 40, 131–136. McNally, R. J., Wilhelm, S. , Buhlmann, U., & Shin, L. M. (2001). Cognitive inhibition in obsessive-compulsive disorder: application of a valence-based negative priming paradigm. Behavioural and Cognitive Psychotherapy, 29, 103–106. Mednick, S. A., & McNeil, T. F. (1973). Current methodology in research on the etiology of schizophrenia: serious difficulties which suggest the use of the high-risk-group method. Psychological Bulletin, 70, 681–693. Meehl, P. E. (1962). Schizotaxia, schizotypy, schizophrenia. American Psychologist, 17, 827–838. Meehl, P. E. (1990). Towards an integrated theory of schizotaxia, schizotypy, and schizophrenia. Journal of Personality Disorders, 4, 1–99.
Latent inhibition as a function of anxiety and stress
473
Mirsky, A. F., & Duncan, C. C. (1986). Etiology and expression of schizophrenia: neurobiological and social factors. Annual Review of Psychology, 37, 291–319. Mogg, K., & Bradley, B. P. (2005). Attentional bias in generalized anxiety disorder versus depressive disorder. Cognitive Therapy and Research, 29, 29–45. Mogg, K., Bradley, B. P., Millar, N., & White, J. (1995). A follow-up study of cognitive bias in generalized anxiety disorder. Behavior Research and Therapy, 33, 927–935. Mogg, K., Mathews, A., & Weinman, J. (1989). Selective processing of threat cues in anxiety states: a replication. Behavior, Research and Therapy, 27, 317–323. Mohanty, A., Herrington, J. D., Koven, N. S., et al. (2005). Neural mechanisms of affective interference in schizotypy. Journal of Abnormal Psychology, 114, 16–27. Molina, V., Tamayo, P., Montes, C., et al. (2008). Clozapine may partially compensate for task-related brain perfusion abnormalities in risperidoneresistant schizophrenia patients. Progress in Neuro-Psychopharmacology & Biological Psychiatry, 32, 948–954. Mongeau, R., Marcello, S., Andersen, J. S., & Pani, L. (2007). Contrasting effects of diazepam and repeated restraint stress on latent inhibition in mice. Behavioural Brain Research, 183, 147–155. Montag, I., & Levin, J. (1992). Personality correlates of schizotypy factors. Personality and Individual Differences, 13, 545–548. Moritz, S., Andresen, B., Probsthein, E., et al. (2000). Stimulus onset asynchronicity as a modulator of negative priming in schizotypy. Personality and Individual Differences, 29, 649–657. Moritz, S. H., Mass, R., & Junk, U. (1998). Further evidence of reduced negative priming in positive schizotypy. Personality and Individual Differences, 24, 521–530. Moser, P. C., Hitchcock, J. M., Lister, S., & Moran, P. M. (2000). The pharmacology of latent inhibition as an animal model of schizophrenia. Brain Research Reviews, 33, 275–307. Muntaner, C., Garcia-Sevilla, L., Fernandez, A., & Torrubia, R. (1988). Personality dimensions, schizotypal and borderline personality traits and psychosis proneness. Personality and Individual Differences, 9, 257–268. Navon, D., & Gopher, D. (1979). On the economy of the human processing system. Psychological Review, 80, 214–255. Norman, R. M., & Malla, A. K. (1993a). Stressful life events and schizophrenia: I. A review of the research. British Journal of Psychiatry, 162, 161–166. Norman, R. M., & Malla, A. K. (1993b). Stressful life events and schizophrenia: II. Conceptual and methodological issues. British Journal of Psychiatry, 162, 166–174. Norman, R. M. G., Malla, A. K., Cortese, L., & Diaz, F. (1998). Aspects of dysphoria and symptoms of schizophrenia. Psychological Medicine, 28, 1433–1441. Nuechterlein, K. H., & Dawson, M. E. (1984). A heuristic vulnerability/stress model of schizophrenic episodes. Schizophrenia Bulletin, 10, 300–309. Nuechterlein, K. H., Dawson, M. E., Gitlin, M., et al. (1992). Developmental processes in schizophrenic disorders: longitudinal studies of vulnerability and stress. Schizophrenia Bulletin, 18, 387–425. Nuechterlein, K. H., Dawson, M. E., & Green, M. F. (1994). Information-processing abnormalities as neuropsychological vulnerability indicators for schizophrenia. Acta Psychiatrica Scandinavia, 90, 71–79.
474
Hedva Braunstein-Bercovitz
Nutt, D. J., Bel, C. J., & Malizia, A. L. (1998). Brain mechanisms of social anxiety disorder. Journal of Clinical Psychiatry, 59, 4–11. O’Malley, J. J., & Gallas, J. (1977). Noise and attention span. Perceptual and Motor Skills, 44, 919–922. Perlstein, W. M., Carter, C. S., Barch, D. M., & Baird, J. W. (1998). The Stroop task and attention deficits in schizophrenia: a critical evaluation of card and single-trial Stroop methodologies. Neuropsychology, 12, 1–12. Peroutka, S. J., Price, S. C., Wilhoit, T. L., & Jones, K. W. (1998). Comorbid migraine with aura, anxiety, and depression is associated with dopamine D2 receptor (DRD2). Molecular Medicine, 4, 14–21. Price, R. B., & Mohlman, J. (2007). Inhibitory control and symptom severity in late life generalized anxiety disorder. Behaviour Research and Therapy, 45, 2628–2639. Raine, A. (1991). The SPQ: a scale for the assessment of schizotypal personality based on DSM-III-R criteria. Schizophrenia Bulletin, 17, 555–564. Roelofs, K., Bakvis, P., Hermans, E. J., van Pelt, J., & van Honk, J. (2007). The effects of social stress and cortisol responses on the preconscious selective attention to social threat. Biological Psychology, 75, 1–7. Salamone, J. D., Cousins, M. S., & Snyder, B. J. (1997). Behavioral functions of nucleus-accumbens dopamine – empirical and conceptual problems with the anhedonia hypothesis. Neuroscience and Biobehavioral Reviews, 21, 341–359. Salo, R., Robertson, L. C., & Nordahl, T. E. (1996). Normal sustained effects of selective attention are absent in schizophrenic patients withdrawn from medication. Psychiatry Research, 62, 121–130. Shalev, U., Feldon, J., & Weiner, I. (1998a). Gender and age-dependent differences in latent inhibition following pre-weaning non-handling implications for a neurodevelopmental animal model of schizophrenia. International Journal of Developmental Neuroscience, 16, 279–288. Shalev, U., Feldon, J., & Weiner, I. (1998b). Latent inhibition is disrupted by acute and repeated administration of corticosterone. International Journal of Neuropsychopharmacology, 1, 103–113. Silver, L. J. (1994). Biologic factors in schizotypal personal disorders. Acta Psychiatria Scandinavia Supplementum, 90, 45–50. Silver, L. J. (1995). Brain structure/function and the dopamine system in schizotypal personality disorder. In A. Raine & T. Lencz (Eds.), Schizotypal Personality. New York: Cambridge University Press, pp. 272–286. Spielberger, C. D., Gorsuch, R. L., & Lushene, R. E. (1970). Manual for the State-Trait Anxiety Inventory (Self-Evaluation Questionnaire). Palo Alto, CA: Consulting Psychologists Press. Stelton, S., & Ferraro, F. R. (2008). The effect of anxiety on the cognitive functioning in non-clinical schizotypal individuals. Current Psychology, 27, 16–28. Stroop, J. R. (1935). Studies of interference in serial verbal reactions. Journal of Experimental Psychology, 18, 643–662. Swerdlow, N. R., Braff, D. L., Hartston, H., Perry, W., & Geyer, M. A. (1996). Latent inhibition in schizophrenia. Schizophrenia Research, 20, 91–103. Swerdlow, N. R., Hartston, H. J., & Hartman, P. L. (1999). Enhanced visual latent inhibition in obsessive-compulsive disorder. Biological Psychiatry, 45, 482–488. Swerdlow, N. R., Magulac, M., Filion, D., & Zinner, S. (1996). Visuospatial priming and latent inhibition in children and adults with Tourette’s disorder. Neuropsychology, 10, 485–494.
Latent inhibition as a function of anxiety and stress
475
Swerdlow, N. R., Stephany, N., Wasserman, L. C., et al. (2005). Intact visual latent inhibition in schizophrenia patients in a within-subject paradigm. Schizophrenia Research, 72, 169–183. Talalaenko, A. N., Abramets, I. J., Stakhovski, Y. V., et al. (1994). The role of dopaminergic mechanisms on the brain in various models of anxious states. Neuroscience, Behavior, and Physiology, 24, 284–288. Thornton, J. C., Dawe, S., Lee, C., et al. (1996). Effects of nicotine and amphetamine on latent inhibition in human subjects. Psychopharmacology, 127, 164–173. Thrasher, S. M., Dalgleish, T., & Yule, W. (1994). Information processing in post traumatic stress disorder. Behaviour Research and Therapy, 32, 247–254. Tipper, S. P. (1985). The negative priming effect: inhibitory priming by ignored objects. Quarterly Journal of Experimental Psychology, 37a, 571–590. Troncoso, A. C., Osaki, M. Y., Mason, S., Borelli, K. G., & Branda˜o, M. L. (2003). Apomorphine enhances conditioned responses induced by aversive stimulation of the inferior colliculus. Neuropsychopharmacology, 28, 284–291. Tsakanikos, E., & Reed, P. (2004). Latent inhibition and context changes in psychometrically defined schizotypy. Personality and Individual Differences, 36, 1827–1839. van den Hout, M., Tenney, N., Huygens, K., Merckelbach, H., & Kindt, M. (1995). Responding to subliminal threat cues is related to trait anxiety and emotional vulnerability: a successful replication of MacLeod and Hagan (1992). Behaviour Research and Therapy, 33, 451–454. Verdoux, H., Magnin, E., & Bourgeois, M. (1995). Neuroleptic effects on neuropsychological test performance in schizophrenia. Schizophrenia Research, 14, 133–139. Vollema, M. G., & van den Bosch, R. J. (1995). The multidimensionality of schizotypy. Schizophrenia Bulletin, 21, 19–31. Vythilingam, M., Blair, S., McCaffrey, D., et al. (2007). Biased emotional attention in post-traumatic stress disorder: a help as well as a hindrance? Psychological Medicine, 37, 1445–1455. Walker, E. F., & Diforio, D. (1997). Schizophrenia: a neural-diathesis-stress model. Psychological Review, 104, 667–685. Weiner, I. (1990). Neural substrates of latent inhibition: switching model. Psychological Bulletin, 108, 442–461. Weiner, I. (2003). The “two-headed” latent inhibition model of schizophrenia: modeling positive and negative symptoms and their treatment. Psychopharmacology, 169, 257–297. Weiner, I. &., & Feldon, J. (1997). Neural substrates of latent inhibition: an update. Behavioural Brain Research, 88, 11–25. Weiner, I., Schnabel, I., Lubow, R. E., & Feldon, J. (1985). The effects of early handling on latent inhibition in male and female rats. Developmental Psychobiology, 18, 291–297. Williams, J. H., Wellman, N. A., Geaney, D. P., et al. (1996). Anti-psychotic drug effects in a model of schizophrenic attentional disorder: a randomized control trial of the effects of haloperidol on latent inhibition in healthy subjects. Biological Psychiatry, 40, 1135–1143. Williams, J. H., Wellman, N. A., Geaney, D. P., et al. (1997). Haloperidol enhances latent inhibition in visual tasks in healthy people. Psychopharmacology, 133, 262–268.
476
Hedva Braunstein-Bercovitz
Williams, L. M. (1995). Further evidence for a multidimensional personality disposition to schizophrenia in terms of cognitive inhibition. British Journal of Clinical Psychology, 34, 193–213. Williams, L. M. (1996). Cognitive inhibition and schizophrenic symptom subgroups. Schizophrenia Bulletin, 22, 139–151. Yoshioka, M., Matsumoto, M., Togashi, H., & Saito, H. (1996). Effects of conditioned fear stress on dopamine release in the rat frontal cortex. Neuroscience Letters, 209, 201–203. Zabal, A., & Buchner, A. (2006). Normal auditory negative priming in schizophrenic patients. The Quarterly Journal of Experimental Psychology, 59, 1224–1236.
20 Nicotinic modulation of attentional deficits in schizophrenia Paul Schnur and Allison Chausmer Hoffman
An introduction: nicotine, schizophrenia, and tobacco use The purpose of this chapter is to review the role of nicotine in the modulation of attention, with particular application to the latent inhibition model of schizophrenia. Nicotine is the primary psychoactive chemical in tobacco. It is addictive and promotes continued tobacco use. According to the Centers for Disease Control and Prevention, the adverse health effects from cigarette smoking account for an estimated 438,000 deaths, or nearly 1 of every 5 deaths, each year in the United States – more than all deaths from human immunodeficiency virus (HIV), illegal drug use, alcohol use, motor vehicle injuries, suicides, and murders combined. The current estimate of cigarette smoking among adults in the United States is 20.8%; however, people with mental illness are much more likely to smoke, and consume a disproportionately large number of cigarettes (Goff, Sullivan, McEvoy et al., 2005; Grant, Hasin, Chou et al., 2004; Lasser, Boyd, Woolhander et al., 2000). Schizophrenia is a cognitive disorder characterized by hallucinations and disturbances in memory, attention, and executive function. It is also characterized by an extremely high prevalence of smoking. Reliable estimates indicate that approximately 85–90% of schizophrenics smoke cigarettes (Hughes, Hatsukami, Mitchell, & Dahlgren, 1986). The co-morbidity between schizophrenia and nicotine addiction is striking and suggests that there may be common mechanisms in the pathways that lead to these diseases. Indeed, the self-medication hypothesis proposes that nicotine serves to alleviate the cognitive impairment of schizophrenia, to reduce the side effects of anti-psychotic medication or to enhance the anti-psychotic efficacy of the medications. Nicotine’s use by schizophrenics might also be maintained by some of the same factors that keep non-schizophrenics smoking, namely nicotine’s reinforcing properties. Since the aforementioned possibilities are not mutually exclusive, smoking among schizophrenics might be maintained by more than one or all of nicotine’s effects. In this chapter, we will focus on nicotine’s modulation of attentional deficits in schizophrenia, as documented in the latent inhibition (LI) and related paradigms. Latent Inhibition: Cognition, Neuroscience and Applications to Schizophrenia, ed. R. E. Lubow and Ina Weiner. Published by Cambridge University Press # Cambridge University Press 2010.
477
478
Paul Schnur and Allison Chausmer Hoffman
A cardinal feature of the cognitive disturbance that characterizes schizophrenia is the disruption of the normal processes of attention. The ability to attend selectively, to screen out irrelevant stimuli in order to focus on a stimulus of particular importance or salience is reduced in schizophrenia. Indeed, deficits in attentional control might underlie other cognitive dysfunctions seen in this disease, such as the hallucinations and disturbances in memory and executive function that are among the positive symptoms observed in schizophrenia (Kumari & Postma, 2005; Patkar, Gopalakrishnan, Lundy et al., 2002). Deficits in attention among schizophrenic patients have been documented in a variety of paradigms (Sacco, Bannon, & George, 2004). Schizophrenics show deficits in prepulse inhibition (PPI), in sensory gating (revealed by reduced P50 amplitudes), in smooth pursuit and anti-saccadic eye movements and in LI and learned irrelevance (LIrr) paradigms (Gray & Snowden, 2005; Kumari & Postma, 2005). As we will discuss below, nicotine and nicotinic agonists typically reverse these deficits while nicotinic antagonists exacerbate them. We begin with a review of nicotine receptor types, their distribution in the CNS and nicotine pharmacology.
Nicotinic acetylcholine receptor subtypes and schizophrenia Acetylcholine was first synthesized in 1867 (Valenstein, 2002). Then, in the early 1900s, Henry Dale determined that while acetylcholine and muscarine had similar effects (decreasing heart rate and lowering blood pressure) at a number of nervous system sites, some sites (e.g., junctions of parasympathetic nerves onto smooth muscle) were only affected by muscarine. At sites where muscarine had no effect (e.g., junctions of nerves and skeletal muscle), low doses of nicotine acted like acetylcholine. This gave rise to the distinction between the two main classes of acetylcholine receptors, muscarinic and nicotinic, that is still in use today (Valenstein, 2002). It is through these nicotinic acetylcholine receptors (nAChRs) that nicotine, the primary psychoactive ingredient in tobacco, exerts its effects. Nicotinic acetylcholine receptors are membrane-bound ligand-gated ion channels that, when bound by nicotine (or nicotinic agonist), open to permit the flow of sodium and potassium cations through the channel pore (Changeux & Edelstein, 2005; Miyazawa, Fujiyoshi, & Unwin, 2003). In the early 1970s, nAChR proteins were isolated, which led to the development of receptor purification techniques. Initially, affinity chromatography was used to purify the solubilized receptors, which were then characterized using ligands, such as a-bungarotoxin (Salvaterra & Mahler, 1976). Technological advancements soon allowed cloning and visualization of nAChRs (Halliwell, 2007). Since the 1980s, molecular genetics techniques have been used to identify and study the genes associated with different nAChR subunits. mRNA expression, for example, has been used to find the expression of individual subunits (Boulter, Connolly, Deneris et al., 1987; Boulter, O’Shea-Greenfield,
Nicotinic modulation of attentional deficits in schizophrenia
479
Duvoisin et al., 1990; Deneris, Connolly, Boulter et al., 1988). These subunits are differentially expressed, with certain subunits (e.g., a4 and non-a) being more abundant than others (e.g., a2 and a3) (Nauright, 1987; Nef, Oneyser, Alliod et al., 1998). In other cases, pharmacology and binding studies using monoclonal antibodies have been used to identify and characterize new subunits (Wada, Ballivet, Boulter et al., 1988). Early experiments used a-bungarotoxin as a selective, irreversible nicotinic receptor antagonist to establish the location of brain nAChRs. One of the first published studies established that nAChRs were located in the hippocampus and thalamus (Hunt & Schmidt, 1978). As visualization and pharmacological tools advanced, more precise localization was possible, with high levels found in the hippocampus, thalamus (including the habenulo-interpeduncular system), accessory olfactory bulb, and hypothalamus (Tribollet, Marguerat, & Raggenbass, 2004). Technological advances have led to the identification of 12 nAChR subtypes, including nine a (a2–a10) and three b (b2–b4) subunits. This leads to a wide variety of possible nAChRs, with heteromeric receptors being “built” using a combination of a and b subunits (e.g., a4b2), and homomeric receptors comprising a single subunit (e.g., a7). Different combinations change the pharmacological properties of the receptor (Luetje & Patrick, 1991; Mihalak, Carroll, & Luetje, 2006). Although the general classification of nAChRs includes both muscle and neuronal receptors, for the purposes of this chapter nAChR will refer to the neuronal nAChR. Among the best-characterized nAChR subtypes are those containing a4b2 and a7 subunits, which have been linked to nicotine reinforcement and dependence. For example, mice lacking b2- or a7-containing nAChR display altered responses to nicotine reward and nicotine withdrawal. b2-containing receptors have been implicated in nicotine reward and the affective signs of nicotine withdrawal, while a7-containing receptors appear to underlie the physical signs of withdrawal (Jackson, Martin, Changeux, & Damaj, 2008; Salas, Main, Gangitano, & De, 2007; Walters, Brown, Changeux et al., 2006). In addition to the well-characterized a4b2 and a7 nAChRs, other nAChRs have been linked to nicotine dependence through genomewide association scans, including a3, a5, and b4 receptor subtypes (Bierut, Stitzel, Wang et al., 2008); however it is too early to know their particular roles in smoking for either normal or schizophrenic populations. Nicotinic acetylcholine receptors appear to be either differentially expressed, or perhaps have modified pharmacological properties in the schizophrenic population. For example, the post-mortem brains of schizophrenic patients were found to have fewer hippocampal nicotinic receptors as compared to nonschizophrenic subjects (Freedman, Hall, Adler, & Leonard, 1995). However, there is not a general decrease in nAChR levels throughout the brains of people with schizophrenia, as others have found increased levels of nAChRs in the cortex (Breese, Lee, Adams et al., 2000). Moreover, changes in the densities of nAChRs in the brains of schizophrenics might be represented differently among the subtypes, with some increasing, others decreasing.
480
Paul Schnur and Allison Chausmer Hoffman
Since the densities of a4b2- and a7-containing nAChR subtypes are different in schizophrenic versus control populations, differences in receptor density may be an important factor in either the initiation of smoking (e.g., greater reward value of nicotine) or for the later cognitive effects (e.g., impaired cholinergic transmission). For example, the density of a7-containing nAChRs is significantly altered in the brains of schizophrenics, with significantly lower levels in the corpus callosum (Severance & Yolken, 2008) and dorsolateral prefrontal cortex (among those with a particular genetic allelic variant for neuregulin; Mathew, Law, Lipska et al., 2007). We know that a7-containing nAChRs are involved in mediating nicotine withdrawal, so altered levels of a7-containing nAChRs may result in altered manifestations of nicotine withdrawal. In another example, schizophrenics appear to have lower levels of a4b2-containing nAChRs (Durany, Zo¨chling, Boissl et al., 2000). Since cigarette smoking has been found to saturate a4b2 nAChRs in the thalamus, brainstem, cerebellum, frontal cortex, and corpus callosum (Brody, Mandelkern, London et al., 2006), altered levels of a4b2-containing nAChRs may lead to heavier smoking in schizophrenics due to a lower reward response. That is, they need to smoke more to get a “normal” level of effect. Additional research will be necessary to address the role of specific nAChRs in smoking and schizophrenia.
Nicotine and cognition Support for the self-medication hypothesis of schizophrenic smoking comes from studies demonstrating beneficial effects of nicotine on cognitive function in experimental models. Nicotine’s effects on cognition have been documented in a variety of species and across numerous learning, memory and attention paradigms. Our concern in this chapter is with nicotine’s effects on attentional modulation; suffice to say, nicotine has beneficial effects on memory and learning in rats, mice, zebrafish, rabbits, monkeys and humans (Levin, McClernon, & Rezvani, 2006). The effects of nicotine agonists and antagonists have been demonstrated in animals and humans, smokers and nonsmokers. In rats, sensory gating, measured by pre-pulse inhibition (PPI) of acoustic startle, is enhanced by acute injections of low doses of nicotine (0.001, 0.01 mg/kg), but not by higher doses (0.1–5.0 mg/kg) (Acri, Morse, Popke, & Grunberg, 1994). Although PPI is a measure of pre-attentive processes, it is used to model attention-related processes since it is deficient in schizophrenia (Braff, Grillon, & Geyer, 1992) and in children with attention deficit disorder (Anthony, 1990). A five-choice serial reaction time task also has been used to investigate nicotinic modulation of attention in rats. The five-choice serial reaction task requires animals to localize a brief visual stimulus that appears in one of five locations and that indicates the presence of food reward. A high level of attention is indicated by accurate stimulus detections and fast response times, whereas premature responding and perseverative responding are
Nicotinic modulation of attentional deficits in schizophrenia
481
errors denoting impulsive and compulsive behavior, respectively (Robbins, 2002). The choice task is analogous to similar tasks used to assess sustained attention in humans. Semenova et al. (2007) used the task to investigate the effects of acute and chronic nicotine, as well as nicotine withdrawal on attention in male rats. Acute (0.14 mg/kg) and chronic (7 days, osmotic minipump) nicotine generally enhanced attention (increased correct responding and decreased response latency), though the effects were somewhat dependent on rat strain and specific conditions of testing. In contrast, nicotine withdrawal induced small, but detectable, attentional deficits (Semenova, Stolerman, & Markou, 2007). Similarly, others have reported that acute nicotine at doses of 0.1, 0.2, and 0.4 mg/kg increased attention (accuracy though not latency) in the choice task among male rats (Day, Pan, Buckley et al., 2007). Rezvani and Levin (2003) tested the ability of nicotine to restore attentional performance that had been impaired by the NMDA antagonist MK801. Female rats were required to respond one way if a signal occurred and another way if no signal occurred. Nicotine reversed the MK801-induced performance deficit, but the effect was not on responding to the signal (“hits”), but on correctly responding to its absence (Rezvani & Levin, 2003). Bushnell, Oshiro, & Padnos (1997) used the same task in male rats and reported that nicotine (0.8 mg/kg) increased attention by increasing “hits”, but only transiently. Finally, Terry et al. (2002) tested the effects of a nicotine agonist, SIB-1553A, in rats and monkeys. In rats, SIB-1553A reversed attentional deficits impaired by MK801 and in monkeys the same agonist improved accuracy by reducing the effects of distraction in the delayed matching to sample task (Terry, Risbrough, Buccafusco, & Menzaghi, 2002). Nicotine, therefore, improves attention in rats and monkeys, but the facilitation is complex and appears to depend upon details of the experimental protocol. In humans, nicotine’s facilitation of attention is relatively straightforward and consistent across studies. Gilbert et al. (2007) measured ERPs in smokers who abstained overnight and who were then given nicotine or placebo in the morning via patch during a vigilance task that presented smoking-related, emotional, and neutral stimuli as distractors. Measures of the P3b brain response indicated that nicotine increased attention to targets and decreased the distracting effects of smoking-related and negative emotional stimuli. Another measure of attention, contingent negative variation, confirmed nicotine’s effect on attention (Gilbert, Sugai, Zuo et al., 2007). Similarly, Baschnagel and Hawk (2008) tested nonsmokers in a prepulse inhibition of acoustic startle task in two laboratory sessions, one with a nicotine patch, one without. In this study, nicotine was found to facilitate attentional filtering, i.e., selective attention. They also reported that individuals with poorer baseline attention benefited most by nicotine administration (Baschnagel & Hawk, 2008), consistent with the self-medication hypothesis. Azizian et al. (2008) tested smokers and nonsmokers using a paper/pencil measure of attention to correlate severity of nicotine dependence with attentional performance. Smokers with a Fagerstrom score of 6 and above (high dependence) had lower attention scores than
482
Paul Schnur and Allison Chausmer Hoffman
nonsmokers and smokers with Fagerstrom scores of 5 or below (Azizian, Monterossa, Brody et al., 2008). Myers et al. (2008) tested smokers on a battery of tests under conditions of nicotine deprivation and non-deprivation. Nicotine nasal spray improved performance on a continuous attention task when smokers were nondeprived and it reversed attentional deficits induced by nicotine deprivation (Myers, Taylor, Moolchan, & Heishman, 2008). The facilitation observed under non-deprivation conditions was especially noteworthy since there has been uncertainty about nicotine’s effects on performance under non-deprivation conditions (Heishman, 1998; Heishman, Snyder, & Henningfield, 1993). Additional support for a facilitatory effect of nicotine on attention can be found in a number of recent papers using a variety of behavioral and physiological measures (Hitsman, Spring, Pingitore et al., 2007; Holmes, Chenery, & Copland, 2008; Mansvelder, van Aerde, Couey, & Brussaard, 2006; Rycroft, Rusted, & Hutton, 2005). In addition to modulating attention in non-psychiatric smokers and nonsmokers, nicotinic effects on attention have been demonstrated in a number of different psychiatric conditions, including schizophrenia (Kumari & Postma, 2005; Leonard, Adler, Benhammou et al., 2001; Sacco et al., 2004). For example, Barr et al. (2008) investigated the effects of nicotine on cognition in nonsmokers with schizophrenia (medicated) compared to nonsmoking healthy controls. Nicotine was administered by patch and subjects were tested once with nicotine patch and once with placebo patch. Two measures of attention were assessed along with measures of memory and psychomotor speed. The results indicated that compared to placebo, nicotine improved scores on the attentional tasks both in schizophrenics and in controls. Moreover, on errors and Stroop performance, the facilitatory effects of nicotine were greater among the patients than the controls (Barr, Culhane, Jubelt et al., 2008). Others have reported modest improvements in attention/vigilance performance among schizophrenics who were chronic smokers after being given nicotine spray following overnight abstinence (Smith, Warner-Cohen, Matute et al., 2006). Sacco et al. (2005) compared schizophrenic and healthy control smokers on a test of sustained attention following overnight abstinence, and then following smoking reinstatement. Hit rates on the sustained attention task were impaired by overnight abstinence and restored by smoking reinstatement. That these effects were the result of interaction with nicotine receptors was confirmed by the fact that the nonselective nicotine receptor antagonist, mecamylamine, dose-dependently reversed the effects of smoking on attention (Sacco, Termine, Seyal et al., 2005). In a recent report that compared to healthy control smokers, schizophrenic smokers were more impaired by abstinence on the Attention Network Test but showed greater improvement after nicotine (patch) administration (AhnAllen, Nestor, Shenton et al., 2008). Postma et al. (2006) used prepulse inhibition (PPI) of tactile startle as the measure of selective attention to compare healthy smokers and nonsmokers with smoking schizophrenics. Subcutaneous injections of nicotine (12 mg/kg, compared to saline) increased PPI in both patients and non-patient groups (Postma, Gray, Sharma et al., 2006). Others
Nicotinic modulation of attentional deficits in schizophrenia
483
have reported a significant negative correlation (r ¼ 0.23) between Fagerstrom scores and a measure of attention among smoking schizophrenic patients (Patkar et al., 2002). Similarly, impairment in PPI of the auditory evoked response (P50) in schizophrenics (Freedman, Coon, Myles-Worsley et al., 1997) was reversed by smoking or nicotine administration (Adler, Hoffer, Griffith et al., 1992; Adler, Hoffer, Wiser, & Freedman, 1993). Additional support for the self-medication hypothesis comes from studies demonstrating that nicotine agonists can attenuate the debilitating physical (Anfang & Pope, 1997; Goff et al., 2005; Yang, Nelson, Kamaraju et al., 2002) and cognitive (Levin, Wilson, Rose, & McEvoy, 1996; Rezvani & Levin, 2004) side effects of pharmacotherapy in schizophrenia.
Smoking, nicotine and dopamine – the reward of smoking In addition to its effects on cognition, nicotine has well-documented reinforcing effects (Corrigall & Coen, 1989; Goldberg, Spealman, & Goldberg, 1981; Harvey, Yasar, Heishman et al., 2004). Thus, schizophrenics might smoke compulsively for the same reason that non-psychiatric individuals do; namely, smoking is both positively and negatively reinforcing. Nicotine is experienced subjectively by humans as pleasurable, stimulating, or mood-elevating (positively reinforcing; Chausmer, Smith, Kelly, & Griffiths, 2003; Harvey et al., 2004); nicotine reverses the negative sequelae associated with abstinence in the habitual user (negatively reinforcing). These effects might be concomitants of nicotine’s effects on cognition and they might be complementary to those effects. They might precede nicotine’s effects on cognition and they might follow the cognitive effects. That is to say, nicotine’s effects are many and varied. Nicotine might also enhance the reinforcing value of nicotine-associated stimuli (Liu, Chang, Liao et al., 2007; Palmatier, Liu, Caggiula et al., 2007). Nicotine, like all other drugs of abuse and many natural rewards, causes a release of dopamine in the mesolimbic pathway, specifically in the nucleus accumbens (Di Chiara, 2000). Di Chiara (2000) reviewed the considerable evidence for dopamine involvement in nicotine’s discriminative stimulus properties, nicotine-induced facilitation of intracranial self-stimulation, nicotine i.v. self-administration, nicotineinduced conditioned place preference and nicotine-induced disruption of latent inhibition. Although dopamine’s involvement in nicotine’s various effects may be necessary, whether it is sufficient remains an open question. Moreover, there is considerable disagreement and lively controversy over dopamine’s precise role in reward (Di Chiara, 2000). Thus, while release of dopamine is certainly an important concomitant of nicotine administration in animals and humans, the mechanisms of nicotine’s effects are not settled. Whether the prevalence of heavy smoking in schizophrenia is related to the dysregulation of dopamine systems in schizophrenia remains largely unanswered. Nevertheless, it is interesting to speculate that the high
484
Paul Schnur and Allison Chausmer Hoffman
prevalence of smoking in schizophrenia might be related to dopamine’s role in signaling unpredictable stimulus events (Berns, McClure, Pagnoni, & Montague, 2001). If the perceptual world of the schizophrenic is indeed a “booming, buzzing confusion” (James, 1890), then nicotine, acting via DA release, might enhance stimulus salience (discriminability of signal from noise) by improving functional integration among brain areas that subserve attention (Dolan, Fletcher, Frith et al., 1995) and thereby reduce a chaotic stimulus overload. If nicotine reward is dependent upon dopamine, it makes sense to inquire about the role of dopamine in reward. If the role of dopamine in reward were known, it might be possible to understand addiction to nicotine in terms of dopamine regulation and perhaps understand also the high prevalence of smoking in schizophrenia, which is characterized by a dysregulated dopamine system. The role of dopamine in reward has been investigated for almost three decades and, while there is continuing debate about the issue, it is clear that dopamine in the mesolimbic dopamine pathway is central to reward processing and that dopamine is certainly not the only transmitter involved in reward processing (Aston-Jones & Kalivas, 2008; Kalivas, 2007; Kalivas, Lalumiere, Knackstedt, & Shen, 2009; Kalivas & O’Brien, 2008; Lalumiere & Kalivas, 2008; Torregrossa & Kalivas, 2008). Berridge recently reviewed the various dopamine hypotheses of reward (Berridge, 2007). Insofar as nicotine’s effects depend upon dopamine-mediated neural events, these same hypotheses and Berridge’s analysis are relevant to nicotine reward. This is not the appropriate forum to review Berridge’s analysis (Berridge, 2007; Berridge & Robinson, 1998, 2003; Robinson & Berridge, 2000), but it is certainly recommended to the interested reader. For the present purposes, it is sufficient and relevant to note that Berridge (2007) concludes that, with a few exceptions (Schultz, Dayan, & Montague, 1997; Volkow, Fowler, & Wang, 2002; Wise, 2004), a consensus is emerging that the principal role of dopamine in reward is best described as the attribution of incentive salience to reward stimuli. Cast in attentional terms, dopamine causes an increase in the attention-grabbing properties of reward and reward-associated stimuli so that the motivational value of these stimuli is increased.
Nicotine, attention modulation in LI, schizophrenia Latent inhibition, the retardation in associative learning that follows nonreinforced preexposure of the conditioned stimulus, is frequently explained in terms of attention modulation or salience loss (Lubow, Schnur, & Rifkin, 1976; Mackintosh, 1975) and has been proposed as a test bed for schizophrenic attentional/cognitive deficits. Acute or unmedicated schizophrenics fail to show latent inhibition, whereas chronic schizophrenics or those on typical as well as atypical antipsychotic medication do exhibit latent inhibition. The failure to show latent inhibition is indicative of a failure to ignore irrelevant stimuli that contributes to the extreme distractibility and
Nicotinic modulation of attentional deficits in schizophrenia
485
cognitive dysfunction that characterizes schizophrenia. It is perhaps the result of excessive dopamine that indiscriminately attributes salience to multiple concurrent stimuli resulting in “failure to ignore” and stimulus overload. The reversal of the LI deficit in acute or unmedicated schizophrenics and LI restoration with antipsychotic medication, dopamine antagonists, is consistent with the dopamine hyperactivity hypothesis for the positive symptoms of schizophrenia. As will be discussed below, DA agonists reduce LI in animals and humans whereas DA antagonists antagonize the disruption. In addition, DA antagonists enhance LI, that is, enhance the ability to learn to ignore. Evidence that nicotine disrupts LI (i.e., maintains stimulus salience, so there is a failure to ignore) in animals and humans would be consistent with the self-medication hypothesis of smoking prevalence among schizophrenics (discussed below). The validity of the LI model of schizophrenia has been reviewed in this volume and elsewhere (Lubow, 2005). Here, we describe only a few representative findings. Feldon and Weiner (1991) used ten CS preexposures in an LI experiment with rats. Ten CS preexposures ordinarily do not produce an LI effect, and they did not do so among saline-treated control animals in this experiment. Animals given the antipsychotics haloperidol and sulpiride during preexposure, however, manifested the usual LI decrement in learning (Feldon & Weiner, 1991). Thus, dopamine antagonists enhanced LI. In an extension of this work, Weiner et al. (1996) tested the effect of the atypical neuroleptic, clozapine, to enhance LI and to reverse amphetamineinduced disruption of LI. In the first two experiments, clozapine (5 and 10 mg/kg) enhanced the expression of LI, where none occurred in vehicle controls. In two other experiments, they demonstrated that amphetamine disrupted LI, but that clozapine antagonized that disruption (Weiner, Shadach, Tarrasch et al., 1996). These findings and others led Schmajuk et al. (1998) to test a neural network model based on the assumption that indirect DA agonists increase the effect of novelty on attention whereas DA antagonists decrease it. The model correctly simulated a variety of previous LI findings (Schmajuk, Buhusi, & Gray, 1998). In humans, the validity of the LI model has been tested among schizophrenics and normals using a variety of tasks that assess attentional processes. Only a few representative findings will be reviewed here. Swerdlow et al. (2003) compared learning to preexposed (PE) and nonpreexposed (NPE) stimuli (within-subjects manipulation) to test the effects of amphetamine and bromocriptine, two DA agonists, on LI in normal humans. Compared to placebo controls who showed LI (fewer correct responses and more trials to criterion for PE than for NPE stimuli), subjects given amphetamine or bromocriptine did not (no difference between PE and NPE stimuli) (Swerdlow, Stephany, Wasserman et al., 2003). In a test with healthy human volunteers, the dopamine antagonist haloperidol increased LI in a visual learning task (Williams, Wellman, Geaney et al., 1997). Vaitl et al. (2002) compared LI in acute non-medicated schizophrenics with chronic medicated patients and healthy controls. LI was observed in both the chronic patient and control groups but not in the acute
486
Paul Schnur and Allison Chausmer Hoffman
group. LI was also attenuated in patients who had suffered their first psychotic episode (Vaitl, Lipp, Bauer et al., 2002). Others have found disrupted LI in acute schizophrenics and enhanced LI in chronic patients (Rascle, Mazas, Vaiva et al., 2001). A variation of the latent inhibition paradigm is the learned irrelevance paradigm (LIrr) in which both the to-be-CS and the to-be-US are preexposed but in an unassociated manner. This preexposure procedure also retards subsequent learning about the CS and is considered to affect attentional processes. A number of investigators have reported that LIrr is disrupted in acute schizophrenia but evident in chronic schizophrenics and healthy controls, suggesting that LI and LIrr share mechanistic properties (Gal, Mendlovic, Bloch et al., 2005; Orosz, Feldon, Gal et al., 2008; Young, Kumari, Mehrotra et al., 2005).
Nicotine and LI: the modulation of attention Evidence reviewed above indicates that nicotine, like other psychostimulants (e.g., amphetamine), enhances attentional processing and facilitates dopamine transmission in reward pathways. If nicotine is being used to normalize attention in schizophrenia, as the self-medication hypothesis proposes, and if schizophrenics fail to show LI as a result of that same attentional dysfunction, then nicotinic agonists should disrupt LI by enhancing attention to the CS (increase salience) and nicotinic antagonists should enhance LI by impairing attention (reduce salience). Similarly, DA agonists should disrupt LI and DA antagonists should enhance LI. There is evidence from animal and human models to support these expectations, but the evidence is not consistent. Indirect dopamine agonists, such as amphetamine or nicotine, have been shown to disrupt LI in animals (Alves, Delucia, & Silva, 2002; Bethus, Muscat, & Goodall, 2006; Chang, Meyer, Feldon, & Yee, 2007; Joseph, Peters, & Gray, 1993; Joseph, Peters, Moran et al., 2000; Ruob, Elsner, Weiner, & Feldon, 1997; Ruob, Weiner, & Feldon, 1998; Solomon, Crider, Winkelman et al., 1981; Weiner, Lubow, & Feldon, 1981, 1984, 1988) and humans (Buhusi, Gray, & Schmajuk, 1998; Dunn, Atwater, & Kilts, 1993; Gray, Pickering, Hemsley et al., 1992; Kumari, Cotter, Mulligan et al., 1999; Thornton, Dawe, Lee et al., 1996). Dopamine antagonists have been reported to enhance LI (Bethus et al., 2006; Buhusi et al., 1998; Dunn et al., 1993; Joseph et al., 1993; Ruob et al., 1997, 1998; Weiner & Feldon, 1987; Williams, Wellman, Geaney et al., 1996; Williams et al., 1997). To illustrate these effects, consider a recent report by Bethus et al. (2006), which tested the effects of amphetamine and haloperidol presented during preexposure on LI of taste aversion learning in mice. Mice were conditioned to associate the taste of sucrose with the effects of lithium chloride. Half the animals were preexposed to the sucrose CS before conditioning and half of those were given an amphetamine injection before preexposure. LI was evident in preexposed mice given saline injections, but not in those given amphetamine during preexposure. They reported also
Nicotinic modulation of attentional deficits in schizophrenia
487
that amphetamine blocked LI whether given before all three preexposure sessions or only before the last session. Using the same procedures, injection of haloperidol before preexposure was found to enhance the LI effect (Bethus et al., 2006). These findings and others (Weiner, 1990; Weiner et al., 1988) emphasize the fact that modulation of dopaminergic tone only during preexposure is sufficient to affect LI and are consistent with an attentional view of the effects of indirect dopamine agonists and antagonists during preexposure. In contrast to the consistent, disruptive effects of amphetamine on LI, the reported effects of nicotine and nicotine antagonists on LI have been somewhat inconsistent and dependent upon details of experimental procedure. Joseph et al. (2003) reported that nicotine blocks LI in rats when given prior to preexposure and conditioning or prior to conditioning only, but not when given only prior to preexposure. This finding, then, is inconsistent with the hypothesis that nicotine enhances attentional processing during preexposure. They concluded that nicotine disrupts LI, not by preventing inattention, but by blocking the utilization of the preexposed stimulus during conditioning (Joseph, Datla, & Young, 2003). Rochford et al. (1996) found that the effect of nicotine occurred during preexposure but that nicotine could disrupt LI or enhance LI, depending upon experimental parameters. When rats were preexposed to 60 presentations of a 60 s tone, nicotine enhanced LI. Under these same conditions, nicotinic agonists cytosine and lobeline also augmented LI and the nicotinic antagonists hexamethonium and mecamylamine reversed the augmentation. Nicotine, however, disrupted LI when rats were preexposed to 40 presentations of a 5 s stimulus (Rochford, Sen, Rousse, & Welner, 1996). In the Rochford et al. (1996) experiments, nicotine’s effects on LI, disruptive or facilitative, occurred when nicotine was administered prior to preexposure. Gould et al. (2001) reported nicotine enhancement of LI in mice and nicotine was presented only prior to preexposure. This effect was blocked by the nicotine antagonist mecamylamine. The findings were interpreted as due to an effect of nicotine on attention, albeit an enhancement of attentional processes (Gould, Collins, & Wehner, 2001). In humans, the effects of nicotine (or smoking status) on LI are inconsistent from one report to another (Allan, Williams, Wellman et al., 1995; Della Casa, Hofer, & Feldon, 1999; Evans, Gray, & Snowden, 2007). Allan et al. (1995) reported that LI was reduced among smokers, whereas Della Casa et al. (1999) reported the opposite effect; viz., enhanced LI for smokers. Evans et al. (2007) reported that individuals who smoke show reduced LI compared to those who don’t smoke and that individuals who smoked recently (< 4 h since last cigarette) showed reduced LI compared to those who hadn’t smoked recently (at least 10 h since their last cigarette). The difference in LI outcomes may be due to heaviness of smoking. In both the Allan et al. (1995) and Evans et al. (2007) studies, which indicated reduced LI in smokers, an average of 6 cigarettes per day were smoked. In the case of Della Casa et al. (1999), who found enhanced LI in smokers, at least 15 cigarettes per day were smoked.
488
Paul Schnur and Allison Chausmer Hoffman
Judging by these findings, the role of nicotinic receptor activation in LI is not straightforward in either rodents or humans. There are many possible explanations for the inconsistencies in the results of these studies, including species, nicotine dose and drug administration differences, conditioning preparation and parameters of preexposure, etc. Yet another explanation concerns the possibility that nicotine’s effect on LI involves nicotine’s stimulation of and/or interaction with one or more receptor systems (e.g., nAChRs, dopamine, glutamate, NMDA). For example, Gould and Lewis (2005) have demonstrated just such effects in a recent study examining the effects of the nicotine receptor antagonist mecamylamine, AMPA receptor antagonist NBQX, and NMDA antagonist MK801, alone and in combination on LI. Although none of the agents affected LI when administered alone, mecamylamine disrupted LI when co-administered with either NBQX or NMDA prior to preexposure (Gould & Lewis, 2005). In this study, then, blocking both nAChRs and glutamate receptors (AMPARs or NMDARs) disrupted LI, whereas blocking one or the other was not sufficient. Why are the effects of nicotine on LI variable and what do these results say about nicotinic modulation of attention? It seems clear from much work reviewed above that nicotine enhances cognitive processes, including attention, and that it does so through its interactions with the DA system and possibly other neurotransmitter systems. As we’ve seen, psychostimulants and DA agonists, therefore, disrupt LI by maintaining attention to the preexposed stimulus or by interfering with learning to ignore that stimulus. Nicotine by contrast seems to sometimes maintain attention to the preexposed stimulus (disrupted LI) and sometimes facilitates the learning of inattention to the stimulus (enhanced LI). Clearly, nicotinic modulation of attention is complex and the use of smoking by schizophrenics as self-medication is further complicated by nicotine’s interactions with antipsychotic or other medications. If, as Berridge (2007) proposes, DA is a neurotransmitter of incentive salience, we might suppose that nicotine is used to titrate the salience of incentive and perhaps nonincentive stimuli as well so that stimulus discriminability (target versus context) is maintained at an acceptable level.
Nicotinic agonist therapy: smoking and schizophrenia If schizophrenics are smoking to self-medicate an attentional/cognitive dysfunction, then nicotinic agonists directed at specific receptor targets might have a role to play in therapy for schizophrenia. Moreover, the availability of LI and related paradigms in animals and humans provides a convenient test bed for screening compounds that might prove useful in schizophrenic pharmacotherapy. The obvious targets for ameliorating the cognitive deficits in schizophrenia are the a4b2 and a7 subunits, both of which have been shown to be deficient in the brains of people with schizophrenia (Durany et al., 2000; Freedman, Adams, & Leonard, 2000).
Nicotinic modulation of attentional deficits in schizophrenia
489
The high-affinity a4b2 subtype is the most prevalent in the brain and research has shown that it is activated, desensitized and up-regulated by nicotine (Ochoa & Lasalde-Dominicci, 2007); the lower-affinity a7 has been shown to mediate the cognitive/attentional dysfunctions in schizophrenia (Levin & Rezvani, 2006, 2007; Olincy, Harris, Johnson et al., 2006). A number of studies have examined the possible therapeutic effects of nicotine agonists in schizophrenia. Olincy et al. (2006) tested the ability of DMXB-A, an a7 agonist, to improve scores on a cognitive test battery among a group of nonsmoking, medicated schizophrenic patients. Results showed that DMXB-A improved cognition and specifically sensory gating, based on the improvement in P50 inhibition (Olincy et al., 2006). Smith et al. (2002) tested the effects of cigarette smoking and nicotine nasal spray on cognitive functioning in schizophrenia. Subjects were medicated schizophrenics who abstained overnight and were tested 10–12 h abstinent. They were tested after smoking a nicotinized or placebo cigarette and after a nicotine or placebo nasal spray. Effects of smoked nicotine were greater than those of nasal spray, but in either case the effects on cognition were modest and transient (Smith, Singh, Infante et al., 2002). Others have found that nicotine via patch improved cognitive performance including processing speed, working memory and attentional function in medicated schizophrenics (Levin et al., 1996). And, as reviewed above, Barr et al. (2008) showed that transdermal nicotine improved attentional performance in nonsmokers with schizophrenia as well as in non-psychiatric controls. Complicating the assessment of the efficacy of nicotinic agonist therapy is the fact that nicotine and antipsychotic drugs may have reciprocal effects on one another through their actions on common neurotransmitter systems. Antipsychotic medications have a broad spectrum of antagonistic actions on dopamine, serotonin, norepinephrine, and histamine systems in the brain. Similarly, nicotine has actions on cholinergic, dopaminergic, serotonergic, GABA-ergic, and glutamatergic systems. The potential for drug–drug interactions is very complex, indeed. Levin and Rezvani have recently found that clozapine can attenuate nicotine’s facilitatory effects on cognitive function, possibly through their actions at a common serotonergic receptor (Levin & Rezvani, 2007). For example, Rezvani et al. (2005) tested the effects of nicotine and ketanserin, a 5-HT2A receptor antagonist, on attentional performance in rats. Antagonism of serotonin 5-HT2A receptors might underlie the effectiveness of atypical antipsychotics in treating attentional dysfunction in schizophrenia. In this study, ketanserin alone had no effect on attentional performance, but it did block nicotine’s facilitatory effect on attention in a signal-detection task. Thus, the development of nicotinic agonists for the treatment of attentional dysfunction in schizophrenia must be mindful of nicotinic interactions with antipsychotic medications that involve serotonergic systems. Similarly, nicotinic interactions with glutamate systems can be complex. Rezvani et al. (2008) administered the NMDA antagonist MK801 to female rats in order to model the cognitive deficits seen in schizophrenia. MK801 impairs performance on the signal-detection attention task but clozapine alone and
490
Paul Schnur and Allison Chausmer Hoffman
nicotine alone reversed the MK801-induced disrupted performance. When clozapine and nicotine were co-administered, however, they were ineffective against MK801 (Rezvani, Tizabi, Getachew et al., 2008). Since clozapine is often used to treat attentional dysfunction in schizophrenia, its effectiveness might be attenuated by smoking. Finally, there is some evidence that nicotine treatment affects schizophrenics differently from normal controls. Jacobsen et al. (2004), for example, used functional magnetic resonance imaging (fMRI) to examine nicotine’s effects on cognitive performance in smokers, either schizophrenics or normal controls. As compared to placebo, treatment with nicotine improved performance of the schizophrenic subjects, but worsened performance by normal controls. Moreover, fMRI revealed that, compared to the normal controls, nicotine enhanced activation of a network of regions in the thalamus and cortex, and increased connectivity in the thalamocortical region in the schizophrenics, suggesting that nicotine may improve schizophrenics’ performance by enhancing activation and functional connectivity (Jacobsen, D’Souza, Mencl et al., 2004).
Conclusion In this chapter we have reviewed the role of nicotine in modulating attentional deficits in schizophrenia. Since the prevalence of smoking is notably high in schizophrenia and since nicotine has well-documented facilitatory effects on a variety of cognitive processes, it is plausible that schizophrenics smoke as a form of self-medication; i.e., because nicotine improves attentional focus and other cognitive processes. As we noted, however, schizophrenics might smoke also because nicotine possesses both positive and negative reinforcing properties. According to the dopamine hypothesis of reward, nicotine reinforces smoking via its actions in the mesolimbic dopamine pathway, specifically in the nucleus accumbens. In addition, nicotine might act as a negative reinforcer via its actions in counteracting the cognitively disruptive effects of psychotropic medications. Nicotine acts through a variety of cholinergic receptor subtypes, the a4b2 and a7 being the most closely linked to addiction and dependence. Recent work has shown that the levels of these might be altered in the brains of schizophrenics, leading to testable hypotheses about possible mechanisms underlying both schizophrenic symptomatology and the high prevalence of smoking in schizophrenia. In addition, genome-wide association scans have pointed to several other subtypes as possibly related to risk for smoking. This research, deriving from a variety of paradigms, suggests a multiplicity of potential nAChR targets for pharmacotherapy directed at ameliorating cognitive dysfunctions seen in schizophrenia. The latent inhibition paradigm, and its variants, recommends itself for investigating nicotinic modulation of attention. The LI effect is well represented across
Nicotinic modulation of attentional deficits in schizophrenia
491
species and the empirical literature documenting its parameters and neurobiological substrates is comprehensive, as evidenced by the literature reviewed in other chapters in this volume. The effects of nicotine on LI vary depending upon details of experimental procedure, sometimes facilitating and sometimes disrupting LI. Additional research will surely explain why the effects of nicotine vary across experimental preparations and provide further understanding of the prevalence of smoking in schizophrenia and whether pharmacotherapy directed at specific nAChRs will have broad applicability to the amelioration of schizophrenic symptomatology.
References Acri, J. B., Morse, D. E., Popke, E. J., & Grunberg, N. E. (1994). Nicotine increases sensory gating measured as inhibition of the acoustic startle reflex in rats. Psychopharmacology (Berl.), 114, 369–374. Adler, L. E., Hoffer, L. D., Wiser, A., & Freedman, R. (1993). Normalization of auditory physiology by cigarette smoking in schizophrenic patients. American Journal of Psychiatry, 150, 1856–1861. Adler, L. E., Hoffer, L. J., Griffith, J., Waldo, M. C., & Freedman, R. (1992). Normalization by nicotine of deficient auditory sensory gating in the relatives of schizophrenics. Biological Psychiatry, 32, 607–616. AhnAllen, C. G., Nestor, P. G., Shenton, M. E., McCarley, R. W., & Niznikiewicz, M. A. (2008). Early nicotine withdrawal and transdermal nicotine effects on neurocognitive performance in schizophrenia. Schizophrenia Research, 100, 261–269. Allan, L. M., Williams, J. H., Wellman, N. A., et al. (1995). Effects of tobacco smoking, schizotypy and number of pre-exposures on latent inhibition in healthy subjects. Personality and Individual Differences, 19, 893–902. Alves, C. R., Delucia, R., & Silva, M. T. (2002). Effects of fencamfamine on latent inhibition. Progress in Neuro-Psychopharmacology & Biological Psychiatry, 26, 1089–1093. Anfang, M. K., & Pope, H. G., Jr. (1997). Treatment of neuroleptic-induced akathisia with nicotine patches. Psychopharmacology (Berl.), 134, 153–156. Anthony, B. (1990). Blink modulation in attention-deficit disorder. Psychophysiology, 27, S6. Aston-Jones, G., & Kalivas, P. W. (2008). Brain norepinephrine rediscovered in addiction research. Biological Psychiatry, 63, 1005–1006. Azizian, A., Monterossa, J. R., Brody, A. L., Simon, S. L., & London, E. D. (2008). Severity of nicotine dependence moderates performance on perceptual motor tests of attention. Nicotine & Tobacco Research, 10, 599–606. Barr, R. S., Culhane, M. A., Jubelt, L. E., et al. (2008). The effects of transdermal nicotine on cognition in nonsmokers with schizophrenia and nonpsychiatric controls. Neuropsychopharmacology, 33, 480–490. Baschnagel, J. S., & Hawk, L. W. (2008). The effects of nicotine on the attentional modification of the acoustic startle response in nonsmokers. Psychopharmacology (Berl.), 198, 93–101.
492
Paul Schnur and Allison Chausmer Hoffman
Berns, G. S., McClure, S. M., Pagnoni, G., & Montague, P. R. (2001). Predictability modulates human brain response to reward. Journal of Neuroscience, 21, 2793–2798. Berridge, K. C. (2007). The debate over dopamine’s role in reward: the case for incentive salience. Psychopharmacology (Berl.), 191, 391–431. Berridge, K. C., & Robinson, T. E. (1998). What is the role of dopamine in reward: hedonic impact, reward learning, or incentive salience? Brain Research Reviews, 28, 309–369. Berridge, K. C., & Robinson, T. E. (2003). Parsing reward. Trends in Neuroscience, 26, 507–513. Bethus, I., Muscat, R., & Goodall, G. (2006). Dopamine manipulations limited to preexposure are sufficient to modulate latent inhibition. Behavioral Neuroscience, 120, 554–562. Bierut, L. J., Stitzel, J. A., Wang, J. C., et al. (2008). Variants in nicotinic receptors and risk for nicotine dependence. American Journal of Psychiatry, 165, 1163–1171. Boulter, J., Connolly, J., Deneris, E., et al. (1987). Functional expression of two neuronal nicotinic acetylcholine receptors from cDNA clones identifies a gene family. Proceedings of the National Academy of Sciences of the United States of America, 84, 7763–7767. Boulter, J., O’Shea-Greenfield, A., Duvoisin, R. M., et al. (1990). Alpha 3, alpha 5, and beta 4: three members of the rat neuronal nicotinic acetylcholine receptor-related gene family form a gene cluster. Journal of Biological Chemistry, 265, 4472–4482. Braff, D. L., Grillon, C., & Geyer, M. A. (1992). Gating and habituation of the startle reflex in schizophrenic patients. Archives of General Psychiatry, 49, 206–215. Breese, C. R., Lee, M. J., Adams, C. E., et al. (2000). Abnormal regulation of high affinity nicotinic receptors in subjects with schizophrenia. Neuropsychopharmacology, 23, 351–364. Brody, A. L., Mandelkern, M. A., London, E. D., et al. (2006). Cigarette smoking saturates brain alpha 4 beta 2 nicotinic acetylcholine receptors. Archives of General Psychiatry, 63, 907–915. Buhusi, C. V., Gray, J. A., & Schmajuk, N. A. (1998). Perplexing effects of hippocampal lesions on latent inhibition: A neural network solution. Behavioral Neuroscience, 112, 316–351. Bushnell, P. J., Oshiro, W. M., & Padnos, B. K. (1997). Detection of visual signals by rats: effects of chlordiazepoxide and cholinergic and adrenergic drugs on sustained attention. Psychopharmacology (Berl.), 134, 230–241. Chang, T., Meyer, U., Feldon, J., & Yee, B. K. (2007). Disruption of the US pre-exposure effect and latent inhibition in two-way active avoidance by systemic amphetamine in C57BL/6 mice. Psychopharmacology (Berl.), 191, 211–221. Changeux, J. P., & Edelstein, S. J. (2005). Allosteric mechanisms of signal transduction. Science, 308, 1424–1428. Chausmer, A. L., Smith, B. J., Kelly, R. Y., & Griffiths, R. R. (2003). Cocaine-like subjective effects of nicotine are not blocked by the D1 selective antagonist ecopipam (SCH 39166). Behavioural Pharmacology, 14, 111–120. Corrigall, W. A., & Coen, K. M. (1989). Fixed-interval schedules for drug self-administration in the rat. Psychopharmacology (Berl.), 99, 136–139.
Nicotinic modulation of attentional deficits in schizophrenia
493
Day, M., Pan, J. B., Buckley, M. J., et al. (2007). Differential effects of ciproxifan and nicotine on impulsivity and attention measures in the 5-choice serial reaction time test. Biochemical Pharmacology, 73, 1123–1134. Della Casa, V., Hofer, I., & Feldon, J. (1999). Latent inhibition in smokers vs. nonsmokers: Interaction with number or intensity of preexposures? Pharmacology, Biochemistry and Behavior, 62, 353–359. Deneris, E. S., Connolly, J., Boulter, J., et al. (1988). Primary structure and expression of beta 2: a novel subunit of neuronal nicotinic acetylcholine receptors. Neuron, 1, 45–54. Di Chiara, G. (2000). Role of dopamine in the behavioural actions of nicotine related to addiction. European Journal of Pharmacology, 393, 295–314. Dolan, R. J., Fletcher, P., Frith, C. D., et al. (1995). Dopaminergic modulation of impaired cognitive activation in the anterior cingulate cortex in schizophrenia. Nature, 378, 180–182. Dunn, L. A., Atwater, G. E., & Kilts, C. D. (1993). Effects of antipsychotic drugs on latent inhibition: sensitivity and specificity of an animal behavioral model of clinical drug action. Psychopharmacology (Berl.), 112, 315–323. Durany, N., Zo¨chling, R., Boissl, K. W., et al. (2000). Human post-mortem striatal a4b2 nicotinic acetylcholine receptor density in schizophrenia and Parkinson’s syndrome. Neuroscience Letters, 287, 109–112. Evans, L. H., Gray, N. S., & Snowden, R. J. (2007). A new continuous withinparticipants latent inhibition task: examining associations with schizotypy dimensions, smoking status and gender. Biological Psychology, 74, 365–373. Feldon, J., & Weiner, I. (1991). Effects of haloperidol on the multitrial partial reinforcement extinction effect (PREE): evidence for neuroleptic drug action on nonreinforcement but not on reinforcement. Psychopharmacology (Berl.), 105, 407–414. Freedman, R., Adams, C. E., & Leonard, S. (2000). The alpha7-nicotinic acetylcholine receptor and the pathology of hippocampal interneurons in schizophrenia. Journal of Chemical Neuroanatomy, 20, 299–306. Freedman, R., Coon, H., Myles-Worsley, M., et al. (1997). Linkage of a neurophysiological deficit in schizophrenia to a chromosome 15 locus. Proceedings of the National Academy of Sciences of the United States of America, 94, 587–592. Freedman, R., Hall, M., Adler, L. E., & Leonard, S. (1995). Evidence in postmortem brain tissue for decreased numbers of hippocampal nicotinic receptors in schizophrenia. Biological Psychiatry, 38, 22–33. Gal, G., Mendlovic, S., Bloch, Y., et al. (2005). Learned irrelevance is disrupted in first-episode but not chronic schizophrenia patients. Behavioural Brain Research, 159, 267–275. Gilbert, D. G., Sugai, C., Zuo, Y., et al. (2007). Brain indices of nicotine’s effects on attentional bias to smoking and emotional pictures and to task-relevant targets. Nicotine & Tobacco Research, 9, 351–363. Goff, D. C., Sullivan, L. M., McEvoy, J. P., et al. (2005). A comparison of ten-year cardiac risk estimates in schizophrenia patients from the CATIE study and matched controls. Schizophrenia Research, 80, 45–53. Goldberg, S. R., Spealman, R. D., & Goldberg, D. M. (1981). Persistent behavior at high rates maintained by intravenous self-administration of nicotine. Science, 214, 573–575.
494
Paul Schnur and Allison Chausmer Hoffman
Gould, T. J., Collins, A. C., & Wehner, J. M. (2001). Nicotine enhances latent inhibition and ameliorates ethanol-induced deficits in latent inhibition. Nicotine & Tobacco Research, 3, 17–24. Gould, T. J., & Lewis, M. C. (2005). Coantagonism of glutamate receptors and nicotinic acetylcholinergic receptors disrupts fear conditioning and latent inhibition of fear conditioning. Learning & Memory, 12, 389–398. Grant, P. F., Hasin, D. S., Chou, S. P., Stinson, F. S., & Dawson, D. A. (2004). Nicotine dependence and psychiatric disorders in the United States. Archives of General Psychiatry, 61, 1107–1115. Gray, N. S., Pickering, A. D., Hemsley, D. R., Dawling, S., & Gray, J. A. (1992). Abolition of latent inhibition by a single 5 mg dose of d-amphetamine in man. Psychopharmacology (Berl.), 107, 425–430. Gray, N. S., & Snowden, R. J. (2005). The relevance of irrelevance to schizophrenia. Neuroscience and Biobehavioral Reviews, 29, 989–999. Halliwell, R. F. (2007). A short history of the rise of the molecular pharmacology of ionotropic drug receptors. Trends in Pharmacological Sciences, 28, 214–218. Harvey, D. M., Yasar, S., Heishman, S. J., et al. (2004). Nicotine serves as an effective reinforcer of intravenous drug-taking behavior in human cigarette smokers. Psychopharmacology (Berl.), 175, 134–142. Heishman, S. J. (1998). What aspects of human performance are truly enhanced by nicotine? Addiction, 93, 317–320. Heishman, S. J., Snyder, F. R., & Henningfield, J. E. (1993). Performance, subjective, and physiological effects of nicotine in non-smokers. Drug and Alcohol Dependence, 34, 11–18. Hitsman, B., Spring, B., Pingitore, R., Munafo, M., & Hedeker, D. (2007). Effect of tryptophan depletion on the attentional salience of smoking cues. Psychopharmacology (Berl.), 192, 317–324. Holmes, A. D., Chenery, H. J., & Copland, D. A. (2008). Transdermal nicotine modulates strategy-based attentional semantic processing in non-smokers. Neuropsychopharmacology, 11, 389–399. Hughes, J. R., Hatsukami, D. K., Mitchell, J. E., & Dahlgren, L. A. (1986). Prevalence of smoking among psychiatric outpatients. American Journal of Psychiatry, 143, 993–997. Hunt, S. P., & Schmidt, J. (1978). The electron microscopic autoradiographic localization of a-bungarotoxin binding sites within the central nervous system. Brain Research, 142, 152–159. Jackson, K. J., Martin, B. R., Changeux, J. P., & Damaj, M. I. (2008). Differential role of nicotinic acetylcholine receptor subunits in physical and affective nicotinic withdrawal signs. Journal of Pharmacology and Experimental Therapeutics, 325, 302–312. Jacobsen, L. K., D’Souza, D. C., Mencl, W. E., et al. (2004). Nicotine effects on brain function and functional connectivity in schizophrenia. Biological Psychiatry, 55, 850–858. James, W. (1890). Principles of Psychology. New York: Henry Holt. Joseph, M. H., Datla, K., & Young, A. M. (2003). The interpretation of the measurement of nucleus accumbens dopamine by in vivo dialysis: the kick, the craving or the cognition? Neuroscience and Biobehavioral Reviews, 27, 527–541. Joseph, M. H., Peters, S. L., & Gray, J. A. (1993). Nicotine blocks latent inhibition in rats: evidence for a critical role of increased functional activity of dopamine in
Nicotinic modulation of attentional deficits in schizophrenia
495
the mesolimbic system at conditioning rather than pre-exposure. Psychopharmacology (Berl.), 110, 187–192. Joseph, M. H., Peters, S. L., Moran, P. M., et al. (2000). Modulation of latent inhibition in the rat by altered dopamine transmission in the nucleus accumbens at the time of conditioning. Neuroscience, 101, 921–930. Kalivas, P. W. (2007). Cocaine and amphetamine-like psychostimulants: neurocircuitry and glutamate neuroplasticity. Dialogues in Clinical Neuroscience, 9, 389–397. Kalivas, P. W., Lalumiere, R. T., Knackstedt, L., & Shen, H. (2009). Glutamate transmission in addiction. Neuropharmacology, 56 (Suppl. 1), 169–173. Kalivas, P. W., & O’Brien, C. (2008). Drug addiction as a pathology of staged neuroplasticity. Neuropsychopharmacology, 33, 166–180. Kumari, V., Cotter, P. A., Mulligan, O. F., et al. (1999). Effects of d-amphetamine and haloperidol on latent inhibition in healthy male volunteers. Journal of Psychopharmacology, 13, 398–405. Kumari, V., & Postma, P. (2005). Nicotine use in schizophrenia: the self medication hypotheses. Neuroscience and Biobehavioral Reviews, 29, 1021–1034. Lalumiere, R. T., & Kalivas, P. W. (2008). Glutamate release in the nucleus accumbens core is necessary for heroin seeking. Journal of Neuroscience, 28, 3170–3177. Lasser, K., Boyd, J. W., Woolhander, S., Himmelstein, D. A., & McCormick, D. M. (2000). Smoking and mental illness: a population-based prevalence study. Journal of the American Medical Association, 284, 2606–2610. Leonard, S., Adler, L. E., Benhammou, K., et al. (2001). Smoking and mental illness. Pharmacology, Biochemistry and Behavior, 70, 561–570. Levin, E. D., McClernon, F. J., & Rezvani, A. H. (2006). Nicotinic effects on cognitive function: behavioral characterization, pharmacological specification, and anatomic localization. Psychopharmacology (Berl.), 184, 523–539. Levin, E. D., & Rezvani, A. H. (2006). Nicotinic-antipsychotic drug interactions and cognitive function. EXS, 98, 185–205. Levin, E. D., & Rezvani, A. H. (2007). Nicotinic interactions with antipsychotic drugs, models of schizophrenia and impacts on cognitive function. Biochemical Pharmacology, 74, 1182–1191. Levin, E. D., Wilson, W., Rose, J. E., & McEvoy, J. (1996). Nicotine–haloperidol interactions and cognitive performance in schizophrenics. Neuropsychopharmacology, 15, 429–436. Liu, C. M., Chang, S. S., Liao, S. C., et al. (2007). Absent response to niacin skin patch is specific to schizophrenia and independent of smoking. Psychiatry Research, 152, 181–187. Lubow, R., Schnur, P., & Rifkin, B. (1976). Latent inhibition and conditioned attention theory. Journal of Experimental Psychology. Animal Behavior Processes, 2, 163–174. Lubow, R. E. (2005). Construct validity of the animal latent inhibition model of selective attention deficits in schizophrenia. Schizophrenia Bulletin, 31, 139–153. Luetje, C. W., & Patrick, J. (1991). Both a and b subunits contribute to the agonist sensitivity of neuronal nicotinic acetylcholine receptors. Journal of Neuroscience, 11, 837–845. Mackintosh, N. J. (1975). A theory of attention: variations in the associability of stimuli with reinforcement. Psychological Review, 82, 276–298.
496
Paul Schnur and Allison Chausmer Hoffman
Mansvelder, H. D., van Aerde, K. I., Couey, J. J., & Brussaard, A. B. (2006). Nicotinic modulation of neuronal networks: from receptors to cognition. Psychopharmacology (Berl.), 184, 292–305. Mathew, S. V., Law, A. J., Lipska, B. K., et al. (2007). a7 nicotinic acetylcholine receptor mRNA expression and binding in postmortem human brain are associated with genetic ariation in neuroregulin 1. Human Molecular Genetics, 16, 2921–2932. Mihalak, K. B., Carroll, F. I., & Luetje, C. W. (2006). Varenicline is a partial agonist at alpha4beta2 and a full agonist at alpha7 neuronal nicotinic receptors. Molecular Pharmacology, 70, 801–805. Miyazawa, A., Fujiyoshi, Y., & Unwin, N. (2003). Structure and gating mechanism of the acetylcholine receptor pore. Nature, 423, 949–955. Myers, C. S., Taylor, R. C., Moolchan, E. T., & Heishman, S. J. (2008). Dose-related enhancement of mood and cognition in smokers administered nicotine nasal spray. Neuropsychopharmacology, 33, 588–598. Nauright, L. P. (1987). Toward a comprehensive personnel system: the reward system–Part V. Nursing Management, 18, 58–4. Nef, P., Oneyser, C., Alliod, C., Couturier, S., & Ballivet, M. (1998). Genes expressed in the brain define three distinct neuronal nicotinic acetylcholine receptors. The EMBO Journal, 7, 595–601. Ochoa, E. L., & Lasalde-Dominicci, J. (2007). Cognitive deficits in schizophrenia: focus on neuronal nicotinic acetylcholine receptors and smoking. Cellular and Molecular Neurobiology, 27, 609–639. Olincy, A., Harris, J. G., Johnson, L. L., et al. (2006). Proof-of-concept trial of an alpha7 nicotinic agonist in schizophrenia. Archives of General Psychiatry, 63, 630–638. Orosz, A. T., Feldon, J., Gal, G., Simon, A. E., & Cattapan-Ludewig, K. (2008). Deficient associative learning in drug-naive first-episode schizophrenia: results obtained using a new visual within-subjects learned irrelevance paradigm. Behavioural Brain Research, 193, 101–107. Palmatier, M. I., Liu, X., Caggiula, A. R., Donny, E. C., & Sved, A. F. (2007). The role of nicotinic acetylcholine receptors in the primary reinforcing and reinforcement-enhancing effects of nicotine. Neuropsychopharmacology, 32, 1098–1108. Patkar, A. A., Gopalakrishnan, R., Lundy, A., et al. (2002). Relationship between tobacco smoking and positive and negative symptoms in schizophrenia. Journal of Nervous and Mental Disease, 190, 604–610. Postma, P., Gray, J. A., Sharma, T., et al. (2006). A behavioural and functional neuroimaging investigation into the effects of nicotine on sensorimotor gating in healthy subjects and persons with schizophrenia. Psychopharmacology (Berl.), 184, 589–599. Rascle, C., Mazas, O., Vaiva, G., et al. (2001). Clinical features of latent inhibition in schizophrenia. Schizophrenia Research, 51, 149–161. Rezvani, A. H., Caldwell, D. P., & Levin, E. D. (2005). Nicotinic-serotonergic drug interactions and attentional performance in rats. Psychopharmacology (Berl.), 179, 521–528. Rezvani, A. H., & Levin, E. D. (2003). Nicotinic-glutamatergic interactions and attentional performance on an operant visual signal detection task in female rats. European Journal of Pharmacology, 465, 83–90.
Nicotinic modulation of attentional deficits in schizophrenia
497
Rezvani, A. H., & Levin, E. D. (2004). Nicotine-antipsychotic drug interactions and attentional performance in female rats. European Journal of Pharmacology, 486, 175–182. Rezvani, A. H., Tizabi, Y., Getachew, B., et al. (2008). Chronic nicotine and dizocilpine effects on nicotinic and NMDA glutamatergic receptor regulation: interactions with clozapine actions and attentional performance in rats. Progress in Neuro-Psychopharmacology & Biological Psychiatry, 32, 1030–1040. Robbins, T. W. (2002). The 5-choice serial reaction time task: behavioural pharmacology and functional neurochemistry. Psychopharmacology (Berl.), 163, 362–380. Robinson, T. E., & Berridge, K. C. (2000). The psychology and neurobiology of addiction: an incentive-sensitization view. Addiction, 95 Suppl 2, S91–117. Rochford, J., Sen, A. P., Rousse, I., & Welner, S. A. (1996). The effect of quisqualic acid-induced lesions of the nucleus basalis magnocellularis on latent inhibition. Brain Research Bulletin, 41, 313–317. Ruob, C., Elsner, J., Weiner, I., & Feldon, J. (1997). Amphetamine-induced disruption and haloperidol-induced potentiation of latent inhibition depend on the nature of the stimulus. Behavioural Brain Research, 88, 35–41. Ruob, C., Weiner, I., & Feldon, J. (1998). Haloperidol-induced potentiation of latent inhibition: interaction with parameters of conditioning. Behavioural Pharmacology, 9, 245–253. Rycroft, N., Rusted, J. M., & Hutton, S. B. (2005). Acute effects of nicotine on visual search tasks in young adult smokers. Psychopharmacology (Berl.), 181, 169. Sacco, K. A., Bannon, K. L., & George, T. P. (2004). Nicotinic receptor mechanisms and cognition in normal states and neuropsychiatric disorders. Journal of Psychopharmacology, 18, 457–474. Sacco, K. A., Termine, A., Seyal, A., et al. (2005). Effects of cigarette smoking on spatial working memory and attentional deficits in schizophrenia: involvement of nicotinic receptor mechanisms. Archives of General Psychiatry, 62, 649–659. Salas, R., Main, A., Gangitano, D., & De, B. M. (2007). Decreased withdrawal symptoms but normal tolerance to nicotine in mice null for the alpha7 nicotinic acetylcholine receptor subunit. Neuropharmacology, 53, 863–869. Salvaterra, P. M., & Mahler, H. R. (1976). Nicotinic acetylcholine receptor from rat brain. Journal of Biological Chemistry, 25, 6327–6334. Schmajuk, N. A., Buhusi, C. V., & Gray, J. A. (1998). Psychopharmacology of latent inhibition: a neural network approach. Behavioural Pharmacology, 9, 711–730. Schultz, W., Dayan, P., & Montague, P. R. (1997). A neural substrate of prediction and reward. Science, 275, 1593–1599. Semenova, S., Stolerman, I. P., & Markou, A. (2007). Chronic nicotine administration improves attention while nicotine withdrawal induces performance deficits in the 5-choice serial reaction time task in rats. Pharmacology, Biochemistry and Behavior, 87, 360–368. Severance, E. G., & Yolken, R. H. (2008). Novel a7 nicotinic receptor isoforms and deficient cholinergic transcription in schizophrenia. Genes, Brain, and Behavior, 7, 37–45. Smith, R. C., Singh, A., Infante, M., Khandat, A., & Kloos, A. (2002). Effects of cigarette smoking and nicotine nasal spray on psychiatric symptoms and cognition in schizophrenia. Neuropsychopharmacology, 27, 479–497.
498
Paul Schnur and Allison Chausmer Hoffman
Smith, R. C., Warner-Cohen, J., Matute, M., et al. (2006). Effects of nicotine nasal spray on cognitive function in schizophrenia. Neuropsychopharmacology, 31, 637–643. Solomon, P. R., Crider, A., Winkelman, J. W., et al. (1981). Disrupted latent inhibition in the rat with chronic amphetamine or haloperidol-induced supersensitivity: relationship to schizophrenic attention disorder. Biological Psychiatry, 16, 519–537. Swerdlow, N. R., Stephany, N., Wasserman, L. C., et al. (2003). Dopamine agonists disrupt visual latent inhibition in normal males using a within-subject paradigm. Psychopharmacology (Berl.), 169, 314–320. Terry, A. V., Risbrough, V. B., Buccafusco, J. J., & Menzaghi, F. (2002). Effects of (þ/-)-4-[[2-(1-methyl-2-pyrrolidinyl)ethyl]thio]phenol hydrochloride (SIB-1553A), a selective ligand for nicotinic acetylcholine receptors, in tests of visual attention and distractibility in rats and monkeys. Journal of Pharmacology and Experimental Therapeutics, 301, 284–292. Thornton, J. C., Dawe, S., Lee, C., et al. (1996). Effects of nicotine and amphetamine on latent inhibition in human subjects. Psychopharmacology (Berl.), 127, 164–173. Torregrossa, M. M., & Kalivas, P. W. (2008). Microdialysis and the neurochemistry of addiction. Pharmacology, Biochemistry and Behavior, 90, 261–272. Tribollet, E., Marguerat, A., & Raggenbass, M. (2004). Comparative distribution of nicotinic receptor subtypes during development, adulthood and aging: an autoradiographic study in the rat brain. Neuroscience, 12, 405–420. Vaitl, D., Lipp, O., Bauer, U., et al. (2002). Latent inhibition and schizophrenia: Pavlovian conditioning of autonomic responses. Schizophrenia Research, 55, 147–158. Valenstein, E. S. (2002). The discovery of chemical transmitters. Brain and Cognition, 49, 73–95. Volkow, N. D., Fowler, J. S., & Wang, G. J. (2002). Role of dopamine in drug reinforcement and addiction in humans: results from imaging studies. Behavioural Pharmacology, 13, 355–366. Wada, K., Ballivet, M., Boulter, J., et al. (1988). Functional expression of a new pharmacological subtype of brain nicotinic acetylcholine receptor. Science, 240, 330–334. Walters, C. L., Brown, S., Changeux, J. P., Martin, B., & Damaj, M. I. (2006). The beta2 but not alpha7 subunit of the nicotinic acetylcholine receptor is required for nicotine-conditioned place preference in mice. Psychopharmacology (Berl.), 184, 339–344. Weiner, I. (1990). Neural substrates of latent inhibition: the switching model. Psychological Bulletin, 108, 442–461. Weiner, I., & Feldon, J. (1987). Facilitation of latent inhibition by haloperidol in rats. Psychopharmacology (Berl.), 91, 248–253. Weiner, I., Lubow, R. E., & Feldon, J. (1981). Chronic amphetamine and latent inhibition. Behavioural Brain Research, 2, 285–286. Weiner, I., Lubow, R. E., & Feldon, J. (1984). Abolition of the expression but not the acquisition of latent inhibition by chronic amphetamine in rats. Psychopharmacology (Berl.), 83, 194–199. Weiner, I., Lubow, R. E., & Feldon, J. (1988). Disruption of latent inhibition by acute administration of low doses of amphetamine. Pharmacology, Biochemistry and Behavior, 30, 871–878.
Nicotinic modulation of attentional deficits in schizophrenia
499
Weiner, I., Shadach, E., Tarrasch, R., Kidron, R., & Feldon, J. (1996). The latent inhibition model of schizophrenia: further validation using the atypical neuroleptic, clozapine. Biological Psychiatry, 40, 834–843. Williams, J. H., Wellman, N. A., Geaney, D. P., et al. (1996). Antipsychotic drug effects in a model of schizophrenic attentional disorder: a randomized controlled trial of the effects of haloperidol on latent inhibition in healthy people. Biological Psychiatry, 40, 1135–1143. Williams, J. H., Wellman, N. A., Geaney, D. P., et al. (1997). Haloperidol enhances latent inhibition in visual tasks in healthy people. Psychopharmacology (Berl.), 133, 262–268. Wise, R. A. (2004). Dopamine, learning and motivation. Nature Reviews. Neuroscience, 5, 483–494. Yang, Y. K., Nelson, L., Kamaraju, L., Wilson, W., & McEvoy, J. P. (2002). Nicotine decreases bradykinesia-rigidity in haloperidol-treated patients with schizophrenia. Neuropsychopharmacology, 27, 684–686. Young, A. M., Kumari, V., Mehrotra, R., et al. (2005). Disruption of learned irrelevance in acute schizophrenia in a novel continuous within-subject paradigm suitable for fMRI. Behavioural Brain Research, 156, 277–288.
21 Latent inhibition and schizophrenia: the ins and outs of context R. E. Lubow
Introduction After 50 years of latent inhibition (LI) research, it may come as a surprise that there is, as yet, no generally accepted theoretical account of this apparently simple, robust effect (i.e., a deficit in conditioned responding as a result of prior exposure to the conditioned stimulus). A variety of explanations have been proposed, and many of them, such as conditioned inhibition and habituation, discarded (for a review, see Lubow, 1989, pp. 141–189; but see Honey, Iordanova & Good, this volume; Schmajuk, this volume). Currently, one can identify three basic types of LI theories. A-theories, which provided the first accounts of LI (e.g., Lubow, Weiner & Schnur, 1981; Mackintosh, 1975; Pearce & Hall, 1980; Wagner, 1976, 1981; more recently, McLaren & Mackintosh, 2000), accept that passive, inconsequential stimulus preexposures reduce the ability of that stimulus to enter into new associations. In addition, most A-theories explicitly acknowledge the role of attention and attribute the reduction of associability to the decline in stimulus-specific attention or salience. In contrast to A-theories, R-theories eschew attentional concepts and maintain that stimulus preexposure has no effect on stimulus associability; the stimulus preexposed (PE) and non-preexposed (NPE) groups enter the acquisition stage with the same capability for forming new associations with the unconditioned stimulus. According to R-theories, in the stage-three test, the associations formed during the stage-one preexposure compete for expression with the association formed in the stage-two acquisition (e.g., Bouton, 1993; Miller, Kasprow & Schachtman, 1986; Weiner, 1990). As described above, A- and R-theories cleave on two issues: (1) whether stimulus preexposures affect subsequent associability or retrievability (see Figure 21.1, routes A and R, respectively); (2) relatedly, whether or not preexposures to an irrelevant stimulus engage an attentional process that reduces stimulus salience. The basic distinctions between A- and R-theories can be recognized as yet another version of the learning versus performance controversy that characterized the earlier Hull–Tolman debates, which, if not solved, were at least neutralized by compromise formulations Latent Inhibition: Cognition, Neuroscience and Applications to Schizophrenia, ed. R.E. Lubow and Ina Weiner. Published by Cambridge University Press # Cambridge University Press 2010.
500
Latent inhibition and schizophrenia: the ins and outs of context Stage 1
Stage 2
Stage 3
Route A
Stimulus Preexposure
Conditioning
Test
Route R
Stimulus Preexposure
Conditioning
Test
Route A/R
Stimulus Preexposure
Conditioning
Test
501
Figure 21.1. Three possible paths for stimulus preexposure effects in three-stage preparations (stimulus preexposure, conditioning, and test). The arrows represent the flow of information from one stage to another stage. Route A (top) describes the attentional/associative deficit model, whereby the test-stage results represent the effects of preexposure on conditioning. Route R (middle) describes the retrieval/competition model, whereby the test-stage results represent independent inputs from the preexposure and conditioning stages, the final output of which is resolved by a competition process. Route A/R (bottom) represents a combination of the A and R paths.
(for a brief review, see Miller & Escobar, 2001), and, which, indeed, has had many other incarnations in behavioral and cognitive theories, such as, for example, the encoding/ retrieval debate in the human memory literature (Tulving & Thomson, 1973), the source of amnesic effects in animals (Miller & Springer, 1973; Gold & King, 1974), and, even on a more grand scale, differentiations between learning and memory, or between early and late processing. In large part, such controversies were the consequences of commitments to specific experimental paradigms. Investigators focused on the manipulation of the experimental variables while largely ignoring the impact that a particular procedure might have on the interpretation of the data that it generated. Bearing witness to this problem, A-theories receive their support primarily from data generated from two-stage procedures, while R-theories rely exclusively on data from three-stage procedures. In the first case, this lends a bias to acquisition (learning) accounts of LI, and in the second case to memory (retrieval) accounts. In spite of the differences between A- and R-theories, the two are neither completely separable nor are they mutually incompatible, and, indeed, the major distinctions between them may not be testable. For example, as noted above, the question as to whether stimulus preexposure affects future associability or retrievability is compromised by the fact that the former is assessed in stage-2 (acquisition)
502
R. E. Lubow
and the latter in the stage-3 test. Thus, retrievability may be affected by prior associability. Conversely, a test of the associability of a preexposed stimulus in the subsequent stage-2 must acknowledge that performance on any given trial reflects not only the associability of that stimulus, but also the degree to which the prior events are retrievable. On the other hand, certain sets of data do appear to be more compatible with one type of theory but not the other. Thus, the effects from preexposure-stage stimulus manipulations, such as number and intensity, would seem to favor A-theories, while post-preexposure manipulations, such as context1 change and extended retention intervals, support R-theories. The considerations described above have led to the development of LI theories that incorporate both attentional/associative deficit and retrieval/competition constructs, many of which are described in this volume (e.g., Escobar & Miller; Hall & Rodriguez; Cassaday & Moran). Of particular interest is the fact that most contemporary theories of LI assign a critical role to CS–context associations. In earlier explanations of LI such associations were the basis of degraded acquisition of the new stage-2 association (e.g., Wagner, 1981). Lubow and Gewirtz (1995) also acknowledged the significance of context, proposing that during the preexposure stage, the subject acquired two simple associations, CS–0 and CS–context, and a higher-order association, whereby the context became an occasion setter for eliciting the CS–0 association. Nevertheless, contemporary context-based integrations of the two general models of LI still cannot account for a number of LI effects. For example, LI can be either attenuated or potentiated depending on the length of the retention interval between conditioning and test and where that interval was spent (for review, Lubow & De la Casa, 2005; De la Casa & Pinen˜o, this volume). Although theories of LI primarily rely on data from animal experiments and stress the role of associative learning, they may differ as to what is learned during the preexposure stage and how and when the relevant associations are transferred to the test stage, as well as to the performance rules for resolving the conflict between the associations learned in the preexposure stage and in the CS–US acquisition stage. Several chapters in this book have already provided excellent reviews of this material (e.g., De la Casa & Pinen˜o; Escobar & Miller; Hall & Rodriguez; Schmajuk), and these summaries will not be repeated here. Instead, a different point of view will be offered, one that is based on data from human LI experiments and emphasizes perceptual rather than associative processes. Although the beginnings of a perceptual theory of LI can be seen in the many theories of LI that have referred to changes in stimulus salience, the present account will give salience a major role (also see Cassaday & Moran, this volume), one which identifies the loss of stimulus salience with an amalgamation of the preexposed stimulus and the context in which it appears. 1
Unless otherwise noted, context refers to all of those environmental stimuli that are relatively constant during the course of an experiment. As such, context is synonymous with apparatus, environment, or background cues. For a variety of other definitions of context, particularly as used in research with schizophrenic patients, see Hemsley (2005).
Latent inhibition and schizophrenia: the ins and outs of context
503
Following presentation of the perceptually based approach to LI, it will be applied to an analysis of attentional dysfunctions in schizophrenia, specifically to the attenuated LI in patients with positive symptoms and to the apparently potentiated LI in patients with negative symptoms.
A general framework for a theory of stimulus preexposure effects A theory of LI must not only incorporate a postulate that allows for different stimulus preexposure conditions (PE and NPE) to produce different effects (A-theories), but also one that allows for the same stimulus preexposure conditions to produce different test-dependent effects (R-theories). The two sets of variables correspond to those that are manipulated during the preexposure stage and modulate the acquisition of LI, and to those that are manipulated after the preexposure stage and modulate performance-related LI effects. The bottom row of Figure 21.1 depicts how both acquisition and performance effects can influence test-stage assessment of LI. Within that framework, several processes will be discussed, including the encoding of stimulus properties and the encoding of stimulus relationships. As a final point, a comprehensive explanation of LI effects also must take into account the fact that preexposing a stimulus without a consequence can result in subsequently improved performance, suggesting that a theory of LI cannot be developed independently of a more general theory of stimulus preexposure effects. The two poles, poorer or better learning following non-reinforced stimulus preexposure, are represented by LI and perceptual learning (PL), respectively. Although most early behavioral theories treated LI and PL as independent phenomena (e.g., Gibson, 1969; Lubow, Weiner & Schnur, 1981; Mackintosh, 1975; Pearce & Hall, 1980; Wagner, 1978; but see Hall, 1991; Lubow, Rifkin & Alek, 1976), more recent accounts have sought a common theoretical framework (e.g., McLaren, Kaye & Mackintosh, 1989; McLaren & Mackintosh, 2000). In general, whether test stage performance is better or worse than an appropriate control group depends on several factors, including: (1) the number of stimulus preexposures; (2) whether the context of the test stage is the same or different from that of the preexposure stage; (3) whether the test involves a discrimination between two or more stimuli; (4) the perceptual load of the masking task (for human subjects). To anticipate the next several sections, it will be proposed that the acquisition of LI and related effects can be described in terms of serially ordered, but overlapping, processes during the preexposure stage that encode primitive stimulus properties (e.g., intensity, color, tastant quality) and stimulus relationships (associations, such as CS–0, CS–context). Although the acquisition of these encodings are largely affected by attentionally related variables, their consequences may be influenced by retrieval and competition factors at the time of test, which also may be influenced by attentional variables. In particular, the case will be made for treating CS–context
504
R. E. Lubow
associations that develop during the preexposure stage in terms of a perceptual integration of the two sets of irrelevant stimuli, an assimilation that is maintained when the test context is the same as the preexposure context, but which is partially deconstructed when the two contexts are dissimilar such that the PE stimulus in the test stage stands out from its context (i.e., becomes salient).
Information processing during the preexposure stage and its utilization in the test stage The role of attention in the preexposure stage The analysis of information acquisition in the preexposure stage begins with the classical distinction between automatic and controlled modes of processing (Schneider & Shiffrin, 1977; Shiffrin & Schneider, 1977). The former is relatively effortless and involuntary, and functions in a rapid, parallel fashion. The latter is relatively effortful and voluntary, and proceeds slowly and serially. In addition, it is assumed that the automatic attentional processes in animals and humans are basically similar, but that controlled processes are to be found predominantly in humans. On that basis, the need for a masking task to generate LI in adult human, and the apparent absence of that requirement for infra-humans and young children (e.g., Lubow, Caspy & Schnur, 1982; also see Lubow, this volume), derives from the predisposition in adults for controlled processing to override automatic processing (Lubow & Gewirtz, 1995). Accordingly, the function of the masking task in generating LI in adults is to engage controlled attention, thereby diverting processing resources from the preexposed stimulus and allowing the to-be-CS to be processed in the default automatic mode. Thus, normal LI in animals and humans is preceded by the automatic processing of PE and context stimuli, leading to the encoding of relatively primitive stimulus properties. However, there are two modes of automaticity that are sequentially engaged during the preexposure stage. On the one hand, the initial preexposurestage trials present the subject with a stimulus that is novel. As a result, attention is involuntarily directed to, or captured by, the stimulus, as reflected in the evoking of an electrophysiological and behavioral orienting response (OR; Sokolov, 1963). This “What is it?” reflex (Pavlov, 1927) or, in more modern terminology, “call for processing resources” (Ohman, 1979), enhances the encoding of qualitative stimulus properties. However, with the repeated presentations of the unreinforced preexposed stimulus, the demand for controlled processing resources is attenuated, and automaticity returns to its primary default function. In other words, an otherwise inconsequential novel stimulus elicits a dynamic interplay between the two processes, such that initially controlled processing is automatically engaged. However, repetition of the same stimulus conditions diminishes, and in due course terminates, any attention allocated to that stimulus. As will be proposed in later sections, the repeated irrelevant stimulus ultimately becomes an integral part of the context in which it was presented.
Latent inhibition and schizophrenia: the ins and outs of context
505
Phase-1: stimulus property encoding Although the preceding paragraph emphasized the encoding of stimulus properties that accompanies the OR, primitive stimulus properties, such as color, can be encoded even under conditions of inattention. For example, Rock and Gutman (1981) preexposed two visual patterns such that one stimulus overlapped the other, creating a figure-ground display. Subjects were asked to rate the figure on a scale of aesthetic preference (the equivalent of a masking task in LI experiments). Subsequent to the preexposure stage, measures of recognition memory were obtained for the attended and unattended figures. The shapes of the attended figures were recognized, but the shapes of the unattended figures were not. However, color was correctly identified in both the attended and unattended conditions. Humans can encode a number of other primitive properties of unattended stimuli, including figure-ground segmentation (Kimchi & Peterson, 2008), grouping (e.g., Russell & Driver, 2005), surface completion (Moore, Grosjean & Lleras, 2003), and direction of background motion (Watanabe, Nanez & Sasaki, 2001). The fact that the perception of non-focal objects is affected by the perceptual/cognitive load of the primary task (e.g., Lavie, 1995, 2005) indicates that the distribution of attentional resources is an important factor in acquiring the distinction between relevant, object-related events and irrelevant, contextual stimuli. Thus, normal adults exhibit LI with low- but not high-masking task load during the stimulus preexposure stage (Braunstein-Bercovitz & Lubow, 1998a; Braunstein-Bercovitz, Hen & Lubow, 2004). As opposed to primitive stimulus qualities, object perception requires attention in order to bind the several stimulus features that together constitute the object (e.g., Treisman, 1993). Since LI procedures are designed so that the preexposed stimulus (also the context) is relatively unattended, it follows that such procedures should induce property encoding only of relatively primitive stimulus features. Indeed, a number of pilot experiments in our laboratory failed to obtain LI with stimuli such as words and recognizable objects. If attention is the glue that binds features into objects, then inattention is the solvent that decomposes objects into stimuli. Importantly, it will be argued here, conditions that promote inattention to the preexposed stimulus are conducive to making the preexposed stimulus become an integral part of apparatus context, which itself is subject to the encoding of primitive stimulus properties. To be sure, the observed stimulus- and context-specificities of LI (for reviews, see Lubow, 1989, pp. 58–59 and pp. 76–79, respectively) require the encoding of stimulus properties during the preexposure stage, as does the acquisition of associations in general. Thus, the encoding of the properties of stimuli must precede stimulus relationship encoding (see below). As a consequence of this priority, very few stimulus preexposures should facilitate the subsequent association of that stimulus with a new stimulus, as compared to an NPE group or to a PE group that has received a relatively large number of stimulus preexposures (Lubow, 1989). Such
506
R. E. Lubow
facilitatory effects have been demonstrated. Kiernan and Westbrook (1993) showed that learning was enhanced by few preexposures of the to-be-CS and depressed by more extensive preexposures. Similarly, Prados (2000) found that a small number of preexposure sessions to navigational landmarks in a swimming pool facilitated learning the position of a hidden underwater platform, while an intermediate level of preexposure abolished the facilitation effect, and a longer preexposure produced an LI effect. A number of conditioned taste aversion experiments also have reported that although LI is obtained after extensive stimulus preexposures, with minimal preexposure the effect is reversed (for review, see Lubow, 2009). Encoding the properties of context stimuli also precedes the acquisition of context–US associations (e.g., Fanselow, 1990). Facilitation of subsequent learning with few stimulus preexposures, and the canceling of this effect with larger numbers, also can be found in the sensory pre-conditioning literature (for review, see Thompson, 1972). Phase-1 stimulus property encoding also is reflected in other types of perceptual learning effects, as illustrated by increased discriminability between two sets of preexposed stimuli or between a preexposed and novel set. This, in itself, has been attributed to an LI-related process. Thus, preexposure of two sets of stimuli provides the organism with more contact with their common elements than with their unique elements. As a result, the salience of the common elements would be reduced (LI) more than that of the unique elements, thereby increasing the discriminability of the two sets (McLaren, Kaye & Mackintosh, 1989; McLaren & Mackintosh, 2000).
Phase-2: stimulus relationship encoding In addition to encoding the primitive properties of the PE stimulus and the context in which it occurs, stimulus relationships also are encoded. The latter involves the possible establishment of associative links amongst the PE stimulus, context stimuli, and other non-programmed stimuli that might be present. In principle, four associations can be acquired by the PE group: (1) a CS–0 association, when the PE stimulus is not followed by any event of significance; (2) a context–0 association, when no significant event occurs in that preexposed context; (3) a context–CS association, when the PE stimulus is preexposed in a particular context; (4) a context–(CS–0) association, when the CS–0 association is acquired in a particular context. Although the four relationships have been described in terms of associations, there are compelling reasons for treating CS–0 and context–0 as experimental operations without assuming that they lead to the acquisition of associations. As will be argued later, the theoretically critical relationship is that of context–CS, which might best be viewed in terms of perceptual integration rather than association. The centrality of context for explaining LI was anticipated by Alan Wagner (1976, 1981). He suggested that, in the test stage, the preexposure context “primes” the PE stimulus into short-term memory (STM) or into a secondary state of activation such that the PE stimulus receives less than normal processing. The present analysis also
Latent inhibition and schizophrenia: the ins and outs of context
507
accepts that stimuli in STM or working memory are subject to controlled attentional processing. However, as will be suggested, the integration of the PE stimulus and the context prevents the stimulus from entering STM during the acquisition/test stage, thus ensuring that it continues to be processed in the automatic mode. Ginton, Urca and Lubow (1975) found some support for this hypothesis when human subjects in the PE group reported being unaware of the target stimulus in the test session. Relatedly, De la Casa et al. (1993) found poorer recall and recognition memory for PE stimuli in groups that exhibited standard LI (low psychotic-prone normals) as compared to groups that exhibited attenuated LI (high psychotic-prone normals). Indeed, as noted by Lubow and Gewirtz (1995, p. 99), “common sense would seem to demand that unimportant stimuli (ones that have been followed by no consequence) be kept away from STM, which by nature of its limited capacity can ill afford to be constantly aggravated by ecologically trivial events. If contexts were continuously priming into conscious awareness the very stimuli that characterize insignificance, every moment of our existence would be characterized by a Jamesian booming, buzzing confusion”. In short, stimuli that become integrated into context, namely previously insignificant stimuli, are prevented from entering into STM, or, more likely, are represented in STM not as information-rich objects but rather as primitive information-impoverished stimulus properties. Such a hypothesis will be used to explain why schizophrenic patients with positive symptoms have attenuated LI while those with negative symptoms seem to have potentiated LI.
A perception-based theory of latent inhibition in humans The descriptions of A-, R- and hybrid theories and of stimulus property and stimulus relationship encoding in the preceding sections provide a counterpoint for the already foreshadowed perceptual theory of LI. As depicted, the previous accounts of LI emphasized classical principles of learning, particularly the acquisition of CS–0 and context–CS associations. In contrast, the present theory offers an account of LI that stresses changes in the perception of the preexposed stimulus that occur with repeated non-reinforced preexposures. Past and present theories, particularly those that have emphasized the reduced associability of the preexposed stimulus, have not entirely ignored the role of perception in LI. The stimulus-specific reduction in associability that is acquired in the preexposure stage has been attributed to a decline in stimulus salience, either from the conditioning of inattention (e.g., Lubow, Weiner & Schnur, 1981) or by some mechanism that reduces the value of alpha in the Rescorla–Wagner equation (e.g., Mackintosh, 1975; Pearce & Hall, 1980; Wagner, 1981). More recent theories also have highlighted the role of stimulus novelty and salience (e.g., Cassaday & Moran, this volume; McLaren & Mackintosh, 2000; Schmajuk, this volume; Stout & Miller, 2007). However, conspicuously absent from these presentations is a direct expression of how the salience of the
508
R. E. Lubow
preexposed stimulus is affected by the context in which it is presented. As will be seen, the present approach is distinguished from past efforts in that it is primarily based on data from human studies, the vast majority of which have used twostage procedures (a preexposure stage and a combined acquisition-test stage) and a formal masking task. The questions that this may raise concerning the similarity between animal and human LI have been dealt with elsewhere (Lubow, 2005; Lubow, this volume). Although the theorizing will focus on the data from human studies, there is no reason, at least as yet, to limit the generality of the conclusions. The basic premise of the theory is that the effects from CS–0 presentations can be best understood in terms of a perceptual integration of CS and context, rather than the acquisition of a CS–0 association. Although the latter may, at times, be indistinguishable from the former, the perceptual account offers a number of advantages for understanding normal human LI and its modulations in schizophrenia. The following sections, drawing in part on the theories and formulations in the previous pages, present the new perception-centered theory, and then assess its ability to account for the basic human LI effects and their modulation in schizophrenic patients and high-scoring schizotypal normals.
The empirical support for a perception-based theory of LI The relationship between LI and perceptual learning Almost 40 years ago I pointed out the need for a model of LI that combines attentional and learning constructs (Lubow, 1973), a requirement that has been repeated by me ever since (e.g., Lubow, 1989; Lubow & De la Casa, 2005; Lubow & Gewirtz, 1995). Following up on this early suggestion, Lubow, Rifkin and Alek (1976) observed that LI and perceptual learning (PL) experiments used nearly identical preexposure and test stages, and yet they produced opposite outcomes: in the LI experiments, the stimulus-preexposed group learned more slowly than the nonpreexposed group; in the PL experiments the stimulus-preexposed group learned more rapidly than the non-preexposed group. To explain this apparent oddity, it was noted that the test stages of LI experiments were conducted in the same context as the preexposure stage, but that the test stages in PL experiments were conducted in a different context from that of preexposure. On this basis, it was hypothesized that the opposing LI and PL effects could be explained in terms of a context-induced modulation of attention allocated to the preexposed stimulus at the time of test. In two experiments, one with rats and one with young children, Lubow et al. (1976) produced either LI or PL depending on the contrast in novelty between the test-stage target stimulus and its context. Thus, when a previously irrelevant preexposed stimulus was presented in the same context as that of preexposure (SoCo; old stimulus in an old context), there was poorer learning (i.e., LI) as compared to a novel stimulus presented in a familiar context (SnCo). On the other hand, when a
Latent inhibition and schizophrenia: the ins and outs of context
509
familiar stimulus was presented in a novel context (SoCn), learning was better (i.e., PL) than when a novel stimulus was presented in a novel context (SnCn). Clearly, the PL effect was dependent on a contrast in novelty between the test stimulus and context, one which was not afforded in the comparison condition. On the other hand, the LI effect was due to the absence of such a contrast, one that was present in the comparison condition. Of course, this is exactly the condition, a change of context from preexposure to test, which disrupts LI.
The visual search analog of LI Although Lubow et al. (1976) used a learning task in the test stage, the explanation of the results made no reference to acquired or retrieved associations. Instead, LI and PL were explained in terms of differences in context-modulated salience of the target stimulus in the PE and NPE groups. In spite of the fact that the data were entirely consistent with such an explanation, a more direct test of the salience contrast hypothesis, one that omitted a learning task in the test stage, was undertaken. In order to have convergent operations for the conditions that produced LI and PL, and to make the results relevant to adult humans, a visual search task was designed that could generate LI- and PL-like effects. It differed from the standard learning-based procedures in that it used a within-subject design, with response time (RT), rather than correct responses, as the dependent variable. The first such study was conducted by Lubow and Kaplan (1997). As with conventional LI and PL experiments, the search task had preexposure and test stages. In both stages, the participant had to detect a target that was a unique complex shape amongst a field of identical distractors. In the preexposure stage, the target was equivalent to the masking task stimuli and the distractors to the irrelevant preexposed stimulus, as in the standard human LI procedure. Amongst the seven test-stage conditions, two were germane to LI (SoCo vs. SnCo) and two to PL (SnCo vs. SnCn). A third pair of conditions (SnCo vs. SnCn) was relevant to the so-called novel pop-out effect (NPO; Johnston & Hawley, 1994). NPO is said to occur when a new target presented in an array of familiar figures (SnCo) is detected faster than a new target in an all-novel array (SnCn). The relationship between NPO and LI is of particular interest because the two paradigms share a common condition (SnCo); the control or baseline condition in LI is the experimental condition in the NPO paradigm. The presence of an NPO effect would suggest that LI, defined as the poorer test performance by the stimulus-preexposed group (SoCo) as compared to the non-preexposed group (SnCo), and typically attributed to a process occurring only in the first group, may, in fact, arise from two independent sources. In addition to stimulus preexposure generating poor performance as a result of decreased attention/salience to that stimulus at the time of test, the non-preexposed condition may induce superior performance, because at the time of test the novel target,
510
R. E. Lubow
presented against a background of familiar distractors, attracts attention, thereby increasing its salience. This latter point was already noted in regard to the Lubow et al. (1976) study. All three effects, LI, PL, and NPO, were obtained in the above study, and they have been replicated in over a dozen similar visual search experiments (for review, Lubow & Kaplan, 2005). In addition, a number of variables that affect LI have been shown to produce similar effects with the visual search procedure. Thus, visual search LI is reduced when the context is changed from the preexposure to the test stage (Kaplan & Lubow, 2001). Furthermore, LI is less in high- as compared to low-schizotypal subjects (Lubow, Kaplan & De la Casa, 2001; Tsakanikos, Sverdrup-Thygenson & Reed, 2003), as well as in a mixed group of adult schizophrenic out-patients (Lubow, Kaplan, Abramovich et al., 2000) as compared to matched controls. Together, these data, derived from test stages that are either learning-based (e.g., Lubow, Rifkin & Alek, 1976) or perception-based (e.g., Lubow & Kaplan, 1997), suggest that LI may be governed by top-down learning principles as well as by bottom-up perceptual principles. Contemporary theories of LI have largely ignored the perceptual component, at least until recently (e.g., McLaren & Mackintosh, 2000; Cassaday & Moran, this volume; Hall & Rodriguez, this volume; Schmajuk, this volume). What the masking task in human LI experiments tells us about stimulus salience and the allocation of attention The masking task The vast majority of experiments that have successfully elicited LI in adults have preexposed the to-be-target stimulus while the subject was occupied with a masking task (e.g., Gray, N. S., Hemsley & Gray, 1992; Gray, N. S., Pickering et al., 1992; Lubow et al., 1992; Pinen˜o, De la Casa, Lubow & Miller, 2006). On the other hand, numerous experiments with non-masked stimulus preexposures have failed to produce an LI effect (e.g., Graham & McLaren, 1998; for a review of the earlier literature, see Lubow, 1973). Most significantly, studies that have explicitly compared masked and non-masked conditions have obtained LI with the former but not with the latter (Braunstein-Bercovitz & Lubow, 1998a; De la Casa & Lubow, 2001; Ginton, Urca & Lubow, 1975; Graham & McLaren, 1998; Lubow, Caspy & Schnur, 1982). Notably, in all of these experiments, the masking task response was qualitatively different from the test task response, thereby precluding effects that might be due to simple response habituation (as with electrodermal conditioning; e.g., Lipp, Siddle & Vaitl, 1992; Lipp & Vaitl, 1992) or response interference (e.g., Escobar, Arcediano & Miller, 2003; Evans, Gray & Snowden, 2007; Schmidt-Hansen, Killcross & Honey, 2009; Tsakanikos, Sverdrup-Thygenson & Reed, 2003), a reservation that also applies to the visual search studies (see above).
Latent inhibition and schizophrenia: the ins and outs of context
511
The effects of masking task load As already suggested, the masking task promotes LI because of its ability to divert attention from the preexposed stimulus, thereby precluding controlled processing, and allowing processing to progress in the default automatic mode. This supposition is congruent with the fact that LI magnitude is a function of masking task load/ difficulty (Braunstein-Bercovitz, Hen & Lubow, 2004; Braunstein-Bercovitz & Lubow, 1998a; also see Hofer, Della Casa & Feldon, 1999; Wuthrich & Bates, 2001). As an example, Braunstein-Bercovitz and Lubow (1998a) used three conditions of masking task load: zero-load (i.e., no masking task); low-load (similarity judgments for upright letter pairs: LL, TT, TL, LT); high-load (similarity judgments, but with the letters appearing in one of four rotated positions). LI was absent in the zero-load condition, present in the low-load condition, and reduced in the high-load condition. We explained the differential effects of masking task load as follows. In the zeroload condition, where there was no competition for processing resources, subjects presumably allocated full attention to the preexposed stimulus. Since the preexposed stimulus was the only stimulus that was displayed, subjects failed to disengage from the controlled mode; attention to the preexposed stimulus was maintained, and LI was not obtained. On the other hand, the low-load masking task allowed the subject to divide resources between the masking task stimuli and the preexposed stimulus, with a greater allocation to the former than the latter. Thus, while the low-load masking task elicited controlled processing, there was still sufficient resource capacity to allow the preexposed stimulus to be processed automatically. As noted earlier, such processing, identified with early selection, may be limited to the encoding of relatively primitive stimulus features. Finally, the high load condition may have demanded all processing resources, leaving none available even for minimal automatic processing. As a result of the absence of any processing, the preexposed stimulus was functionally novel in the test stage, and there was no LI effect. Since masking task load affects LI, the question arises as to the source of the effect. Does load affect the encoding of stimulus properties and/or the encoding of one or more of the associations? Alternatively, as opposed to asking what is encoded, one might ask how the information is encoded, emphasizing conditions that might affect retrieval, such as transfer appropriate processing (e.g., Morris, Bransford & Franks, 1977), or relatedly, encoding specificity (e.g., Tulving & Thomson, 1973). Irrespective of the explanation, between-experiment differences in masking task load may account for some of the inconsistencies in the human LI literature, particularly in regard to schizophrenia. Other load-related variables In addition to load, there are other variables that affect the interaction between the preexposed and masking task stimuli. However, since predicting the relative size of an LI effect requires knowledge of the load-dependent distribution of attention, the
512
R. E. Lubow
only experiments of interest will be those that use masking tasks for which LI data are already available. Using the basic masking task procedure described in the previous paragraphs, Braunstein-Bercovitz and Lubow (1998b) manipulated the spatial relationship between the masking task and the PE stimuli. In most LI studies with visual stimuli, the masking task stimuli (M) are presented at the center of the computer screen, and they are flanked by the currently irrelevant but to-be-target stimuli (T), creating a T MM T configuration (e.g., Lubow et al., 1992; Zalstein-Orda & Lubow, 1995). Braunstein-Bercovitz and Lubow (1998b, Exp. 1a) showed that such a composition was, in fact, necessary to produce LI. When the PE stimuli were located centrally, between the two masking stimuli (M TT M), LI was not obtained. The absence of LI in the MTTM condition was attributed to automatic allocation of attention to the PE stimuli because, during the course of scanning the two masking stimuli, the PE stimuli fell on the fovea, or because with such a configuration “. . . attention forms a unitary zone that may expand to encompass multiple relevant locations, but must also include the area between them” [my italics] (Usai, Umilta & Nicoletti, 1995). In either case, when the preexposed stimuli are central to the masking task stimuli, more attention will be allotted to them than if they were peripheral, thereby interfering with the processes necessary to generate LI. LI is not only decremented when the PE stimuli are placed between the masking task stimuli (M TT M) as opposed to peripheral to them (T MM T), but also when the PE stimulus is paired with an adjacent novel stimulus at the end of the preexposure stage (Braunstein-Bercovitz & Lubow, 1998b). Evidently, such a procedure draws attention away from the masking task and invests it in the PE stimulus, the consequence of which is an attenuation of LI.
Anxiety and stress There is considerable evidence that anxiety interferes with the ability to ignore irrelevant stimuli (for reviews, see Braunstein-Bercovitz, this volume; Eysenck, Derakshan, Santos & Calvo, 2007). In terms of the above analysis of the masked LI paradigm, this would mean that, at least in a low-load condition, high- as compared to low-anxious subjects would allocate more processing resources to the peripherally preexposed irrelevant stimulus. Thus, high-anxious subjects should have attenuated LI. Braunstein-Bercovitz, Dimentman-Ashkenazi and Lubow (2001) tested the anxiety hypothesis by manipulating stress in two rule-learning experiments that used the same low-load masking task conditions as in the earlier BraunsteinBercovitz studies. One experiment was conducted in the laboratory and the other in the field. In both cases, adult subjects in low- but not high-stress conditions exhibited LI. Relatedly, anxiety is accompanied by elevated dopaminergic activity (e.g., Nutt, Bell & Malizia, 1998), and, as has been repeatedly shown, dopamine agonists such as amphetamine reduce LI in rats and in humans, while dopamine antagonists such as haloperidol have the opposite effect, again in rats and humans (for reviews, see
Latent inhibition and schizophrenia: the ins and outs of context
513
Weiner, 2003; Weiner, this volume). The contrasting LI effects from dopamine agonists and antagonists may be related to opposing distributions of attentional/ processing resources. Field dependence The perception-based account of LI, with its emphasis on the roles of stimulus salience and context, suggests that LI should be related to subjects’ ability to separate figure from ground. Using the basic LI procedures of her previous experiments, BraunsteinBercovitz (2003) examined the association between field-dependency, as measured by the Cattell embedded figure test, and LI. LI was greater in field-dependent as compared to field-independent individuals. Peterson and Carson (2000) obtained similar results with an openness-to-experience questionnaire. They found that LI was attenuated in high- as compared with low-scoring subjects (also see Carson, this volume). These data support the idea that LI is inversely related to the salience of the preexposed stimulus, which, in turn, is affected by the context in which it was preexposed.
Implications for schizophrenia Based, in part, on some of the earlier suggestions, the present section explores the possibility that the anomalous LI effects in schizophrenia patients and highschizotypal normals can be traced back to defective integration, either too strong or too weak, of the preexposed stimulus and its context. However, some background information is required in regard to: (1) the relationship between LI effects and schizophrenia, particularly since different symptom clusters of the pathology may have dissimilar neurobiological bases, perhaps even being polar opposites (see Weiner, this volume); (2) the role of context in explanations of schizophrenia. Following brief discussions of these points, the perceptual account of LI will be applied to positive and negative symptom patients, and to corresponding schizotypal groups.
Attenuated and potentiated LI effects and their relationship to positive and negative symptoms in schizophrenia patients Until recently, it has been generally agreed that animal LI represents a stimulusspecific loss of salience/attention that reduces the associability of the preexposed stimulus (e.g., see earlier discussions of A-theories). If LI is modulated by attentional processes, and if schizophrenia, as many have claimed, is characterized by attentional dysfunctions, particularly as related to distractibility (e.g., Anscombe, 1987; Braff & Light, 2004; Heinrichs & Zakzanis, 1998; McGhie & Chapman, 1961; Minas & Park, 2007; Mirsky & Duncan, 1986; Nuechterlein & Dawson, 1984), then LI in these groups should be different from that in healthy subjects. Indeed, LI is reduced or abolished in recently medicated, acute, positive-symptom (e.g., agitation, delusions,
514
R. E. Lubow
hallucinations, cognitive disorganization) schizophrenic patients as compared to longerterm, medicated, chronic schizophrenic patients and to normals (Baruch, Hemsley & Gray, 1988; Gray, N. S., Hemsley & Gray, 1992; Gray, N. S., Pilowsky, Gray & Kerwin, 1995; Rascle, Mazas, Vaiva et al., 2001; Sitskoorn, Salden & Kahn, 2001; Vaitl, Lipp, Bauer et al., 2002; but see Swerdlow et al., 1996; Swerdlow, this volume; Williams et al., 1998; also see Kumari & Ettinger, this volume). As opposed to this, chronic, medicated schizophrenic patients, for whom negative symptoms (e.g., emotional apathy, absence of volition) are usually dominant, exhibit normal LI (Baruch et al., 1988; Leumann, Feldon, Vollenweider & Ludewig, 2002; Lubow, Weiner, Schlossberg & Baruch, 1987; Serra, Jones, Toone & Gray, 2001), or even super-LI, as reported for chronic medicated schizophrenics with high levels of observed negative symptoms (Rascle et al., 2001), and when the negative symptoms are combined with low levels of positive symptoms (Cohen, E., Sereni, Kaplan et al., 2004).2 In general, high-schizotypal normals also exhibit less LI than low-schizotypals (e.g., Braunstein-Bercovitz & Lubow, 1998a; for reviews, see Braunstein-Bercovitz, this volume; Kumari & Ettinger, this volume; Lubow, 2005). In addition, several studies have taken advantage of the fact that self-report schizotypal questionnaires such as SPQ (Raine, 1991) and O-LIFE (Mason, Claridge & Jackson, 1995) are composed of sub-scales that can be used to differentiate between positive and negative symptom types. Parallel to the data from schizophrenic patients, LI attenuation is associated with scale-scores that reflect positive symptoms, but not negative symptoms (e.g., Burch, Hemsley & Joseph, 2004; Evans, Gray & Snowden, 2007; Gray, N.S., Fernandez, Williams et al., 2002; Schmidt-Hansen, Killcross & Honey, 2009). The results with positive-symptom patients and otherwise healthy schizotypals, together with the disruption of LI by dopamine agonists such as amphetamine, and the potentiation of LI by dopamine antagonists such as haloperidol, provided the basis for a dopamine/attentional dysfunction model of attenuated LI in schizophrenia (e.g., Gray, J.A., 1998; Gray, J.A., Feldon, Rawlins et al., 1991; Lubow, 1989; Weiner, 1990). However, this broad conceptualization now appears to be appropriate only for the acute, positive-symptom patients, who are particularly vulnerable to distraction on a variety of tasks.3 Recent developments have encouraged a more 2
3
Although the positive–negative classification of schizophrenic symptoms (e.g., Andreasen & Olsen, 1982) is still widely used, three-factor systems, which have added a component for disorganization, have become increasingly popular (e.g., Mass, Schoemig, Hitschfeld et al., 2000; Moritz, Andresen, Jacobsen et al., 2001). As yet, there have been no reported studies of LI studies that have used the tripartite division of symptoms. Attentional deficits in schizophrenic patients that have been described as effects from distractibility include digit recall (e.g., Lawson, McGhie & Chapman, 1967; Oltmanns, Ohayon & Neale, 1978), letter detection (e.g., Cash, Neale & Cromwell, 1972; Neale, McIntyre, Fox & Cromwell, 1969), word recall (e.g., Frame & Oltmanns, 1982), and letter detection (e.g., Cash, Neale & Cromwell, 1972). With dichotic listening, an irrelevant message increases the number of shadowing errors by schizophrenics compared to normals. These errors are reflected primarily in errors of omission, not reporting words in the relevant message (e.g., Hemsley & Richardson, 1980; Schneider, 1976), but can also include intrusions from the nonattended ear (Dykes & McGhie, 1976). Elevated interference effects in schizophrenics also have been demonstrated with the Stroop task, both with word naming by a competing color (e.g., Wapner & Krus, 1960) and with color naming by a competing word (e.g., Abramczyk, Jordan & Hegel, 1983). A meta-analysis by Heinrichs and Zakzanis (1998) on differences in distraction effects between schizophrenic patients and healthy controls on a variety of tests, including digit span, continuous performance, trail marking (Parts A and B), and Stroop yielded effect sizes ranging from 0.62 (digit span) to 1.22 (Stroop).
Latent inhibition and schizophrenia: the ins and outs of context
515
comprehensive approach. On the one hand, newer theoretical accounts of the LI deficits in schizophrenia have emphasized retrieval/competition rather than attentional processes (Escobar, Oberling & Miller, 2002; Oberling, Gosselin & Miller, 1998; also see Lubow & Gewirtz, 1995; Weiner, 2003, and the earlier discussion of R-theories). On the other hand, congruent with the present perceptual approach, context continues to play a central role in relating LI to schizophrenia. As an example, the Miller model attributes the reduced LI in schizophrenic patients to their inability, during the preexposure stage, to establish associations between the context and the to-be-CS, a prerequisite for obtaining normal LI as proposed by their comparator hypothesis (e.g., Grahame, Barnet, Gunther & Miller, 1994), as well as by other explanations of LI (e.g., Lubow & Gewirtz, 1995; Wagner, 1981). However, these models do not account for recent reports that schizophrenic patients with predominantly negative symptoms display enhanced LI (see above).
Schizophrenia and context As would be expected from the critical role of context in the acquisition and expression of LI and from the abnormal LI effects associated with schizophrenia and schizotypality, there is an extensive literature indicating that the behavior of schizophrenics is affected by disrupted context processing. Shakow (1962) was one of the first to discuss the importance of context in affecting the behavior of these patients. In a subsequent influential review, Chapman and Chapman (1973) described a number of studies that demonstrated the inability of schizophrenics to attend to and use contextual restraints to organize verbal material (e.g., Honigfeld, 1963; Lawson, McGhie & Chapman, 1964; also see Rutter, Wishner & Callaghan, 1975). Since that time, context has become a significant theme in schizophrenia research (e.g., Cohen, J. D., & Servan-Schreiber, 1992, 1993; Hemsley, 2005; Uhlhaas & Silverstein, 2005). As noted by Cohen, J. D., Barch, Carter and Servan-Schreiber (1999, p. 120), “. . . an important subset of cognitive deficits in schizophrenia can be understood in terms of a disturbance in a single underlying mechanism, one that is responsible for the representation and maintenance of context information needed to select task-appropriate action”. It even has been suggested that the generally slow RTs of schizophrenic patients may reflect an inability to develop stable representations of context (e.g., Cohen, J. D., et al., 1999; Cohen, J. D., & Servan-Schreiber, 1992; Silverstein, Kovacs, Corry & Valone, 2000). However, the definition of context in these studies is very different from that used in the animal-LI experiments, where context is synonymous with apparatus or home cage stimuli. In experiments with patients, the definition is broader and includes all task-relevant information, such as instructions, and prior or concurrent stimulus conditions that may, for example, disambiguate the target stimulus. In contrast, human studies of perceptual organization use contexts that approach the definition
516
R. E. Lubow
in animal-LI studies, at least if one accepts that the ability to group disparate stimulus elements provides the basis for establishing a representation of context (Grossberg, Mingolla & Ross, 1997; Phillips & Singer, 1997). In this regard, schizophrenic patients who are rated high on disorganization show impaired perceptual grouping of non-contiguous stimulus elements, suggesting a weakened capacity to form a representation of context (e.g., Silvertein, Kovacs, Corry & Valone, 2000; for review, see Uhlhaas & Silverstein, 2005). Schizophrenic patients may be particularly deficient in their ability to perceptually organize novel irrelevant information (Lubow, Kaplan, Abramovich et al., 2000; Place & Gilmore, 1980; Silverstein, Knight, Schwarzkopf et al., 1996; Wells & Leventhal, 1984).
The proposed role of context in attenuating LI in positive-symptom patients and in potentiating LI in negative-symptom patients As already implied, the attenuated LI in acute, positive-symptom schizophrenics can be attributed to the failure of the preexposed stimulus to become integrated with the context, the result of which is that the original salience of preexposed stimulus is maintained throughout the preexposure stage. As such, at the time of test, the salience of the preexposed stimulus is high, and there is little or no LI. Alternatively, the many experimentally un-programmed background stimuli also may sporadically capture the attention of positive symptom patients. As a result, the representation of context itself would be unstable. Under either of these conditions, STM would be inundated with experimentally familiar but phenomenally novel stimuli, a situation that Frith (1979) associates with the inability of schizophrenics to limit the contents of consciousness. Similarly, Pogue-Geile and Oltmanns (1980) suggested that schizophrenic patients suffer from a reduction in “overall capacity to handle information in STM”, and noted that such subjects may have, for example, difficulty in continuously accessing STM during conversation (e.g., Rochester, 1978). These proposals, the data on which they are based, and many of the positive symptoms exhibited by patients with schizophrenia, are all compatible with the idea that, for this group, STM is “preoccupied” with irrelevant stimuli, the product of an overly labile, distraction-sensitive attention that affects context processing. This context-modulated salience approach is broadly compatible with that of Kapur (2003), who proposed that positive symptoms of schizophrenia, such as hallucinations and delusions, are a result of hyper-dopaminergic activity that affects distractibility via stimulus salience, irrespective of whether those stimuli are internally or externally generated (also see Guterman, 2007). Delusions, then, are the products of cognitively related attempts to provide a coherent account of these external, abnormally salient stimuli, while hallucinations represent a direct experience of super-salient internal stimulus representations. In both cases, stimuli/events
Latent inhibition and schizophrenia: the ins and outs of context
517
that are normally inconspicuous and unattended gain in salience and thereby affect behavior, albeit inappropriately. If patients with positive symptoms suffer from overly labile attentional processes, then those with negative symptoms, who are relatively quick to integrate the preexposed stimulus and its context, would seem to have overly inert attention, a condition that is reminiscent of Kraeplin’s (1919/1971) description of schizophrenic patients who are “. . . often rigidly fixed for a long time, so that the patients stare at the same point, or the same object, continue the same line of thought, or do not let themselves be interrupted in some definite piece of work”. In contrast to the positivesymptom, hyper-dopaminergic schizophrenic patient, the negative-symptom patient, arguably hypo-dopaminergic, would seem to be governed by an opposing process. For these patients, the assimilation of CS and context is strengthened, producing the abnormally persistent or super-LI effects that have been reported in animal experiments that manipulate dopamine-related pathways in the context-sensitive nucleus accumbens (e.g., Gal, Schiller & Weiner, 2005; Schiller, Zuckerman & Weiner, 2006; Weiner, this volume). For the overly inert, negative-symptom patient, attention is allocated to that portion of the perceptual field that contains the most salient stimulus. Thus, during the preexposure stage, attention is assigned disproportionately to the masking task stimuli, with relatively limited resources devoted to the nominal PE stimulus. As a result, the PE stimulus recedes into the context, thereby promoting LI. However, since the masking task stimuli, although no longer relevant, are still present in the test stage, the negative-symptom patient may continue to allocate attention to those stimuli, now at the expense of the new relevant stimulus, again a condition that would elevate LI. Although it might appear that super-LI in negative-symptom patients is the result of a summation of effects from two different sources, (1) an overly rapid/strong integration of the preexposed stimulus and its context, and (2) a deficit in the ability to disengage from a previously relevant stimulus, the fact that the same masking task condition is present for the NPE group as for the NPE suggests that test-phase differences between the two groups cannot be attributed to a failure of disengagement.
Resolving the salience/intensity paradox A number of studies have reported that LI is an increasing function of the intensity or salience of the preexposed stimulus (e.g., Crowell & Anderson, 1972; Gilley & Franchina, 1985; Rodriguez & Alonso, 2002; Schnur & Lubow, 1976). Although these results are compatible with some theories of LI (e.g., conditioned attention theory [Lubow, 1989] explicitly predicts such an effect), they appear to violate the basic premise of the perceptual approach to LI. On the face of it, a high-intensity preexposed stimulus should perceptually protrude from the context more than a low-intensity preexposed stimulus. As a consequence, the former should be at a disadvantage in its ability to merge with context, with the result that it should lead
518
R. E. Lubow
to less, rather than more, LI than a preexposed low-intensity stimulus. Moreover, if negative-symptom patients disproportionately allocate attention to the masking task stimuli, then the salience of the PE stimulus should be less than that for normals, with the consequence that these patients should exhibit less, rather then more, LI than healthy controls. Likewise, positive-symptom patients, who pay more attention to the PE stimulus, should show super- rather than attenuated LI. How, then, can the perceptual account of LI be reconciled with the fact that LI increases with an increase in the intensity of the preexposed stimulus? Although all of the above-cited experiments were conducted with rats, if we assume that similar effects would be found with human subjects, several admittedly ad hoc solutions come to mind. (1) Stimulus property encoding should be completed faster by the high- than by the low-intensity preexposed stimulus group. Since both PE groups receive the same number of stimulus-preexposure trials, the high-intensity group would have functionally more preexposure trials in which the acquired CS–0 association could be paired with context (since the acquisition of CS–0 must precede the encoding of CS properties). (2) Relatedly, the specific level of salience at the beginning of the test stage may be less important than how it was achieved. Thus, learning not to attend reduces salience, presumably by creating the conditions for incorporating the PE stimulus into its context. As an example, if the salience of a novel stimulus was equated with that of a preexposed stimulus (after preexposure), then conditioning performance would be poorer for the latter than the former (indicating that salience and relevance are not independent; also see Cassaday & Moran, this volume). (3) Finally, a preexposed high-salience stimulus may be more easily retrieved, in this case with the information that it had not been followed by a consequence, than a similarly treated stimulus of low salience. As such the former would evince more LI than the latter. Given the above possibilities, why do positive-symptom subjects, who presumably allocate more attention to the irrelevant PE stimulus than the negative-symptom subjects, have attenuated LI? Arguably, for these subjects the maintenance of attention to the irrelevant PE stimulus is accompanied by investing that stimulus with significance, with the higher the intrinsic stimulus intensity (bottom-up influence on salience), the greater the investment of significance (top-down influence on salience). Thus, one can account for the attenuated LI in positive-symptom subjects by assuming that the normally LI-potentiating effect of increased stimulus intensity can be overridden by the assignment of significance to the preexposed stimulus, a process that would prevent the PE stimulus from becoming part of the context, thereby maintaining attention to it and preventing LI. This account is compatible with descriptions of the positive symptoms of schizophrenia (see above). Relatedly, one must ask why negative-symptom subjects, who presumably allocate less attention to the irrelevant PE stimulus than the positive-symptom subjects, have potentiated LI. Perhaps, these subjects not only weakly attend to the irrelevant PE stimulus, but the higher the intrinsic intensity (bottom-up) of the preexposed
Latent inhibition and schizophrenia: the ins and outs of context
519
stimulus, the more they invest in trying to degrade its significance (top-down influence on salience). Thus, to account for the potentiated LI in negative-symptom subjects, one has to propose that the LI-attenuating effect of weak stimulus intensity can be overridden by the withdrawal of significance from the PE stimulus, a process that would promote the integration of the PE stimulus into the context, thereby reducing attention to it and potentiating the LI effect. This side of the perceptual model is consistent with descriptions of the negative symptoms of schizophrenia (see above).
Summary and conclusions In summary, unreinforced preexposures of the to-be-CS and the accompanying context stimuli result in the encoding of their respective qualitative properties. This, in turn, is followed by the acquisition of CS–0 and context–0 associations, forming a link between CS and context, such that the CS becomes an integral part of the context. The context–CS association/integration provides the basis for the primary LI effect. While the acquisition of that relationship, perhaps through the acquisition of the component associations, would be affected by attention-related variables such as the intensity (e.g., Crowell & Anderson, 1972; Rodriguez & Alonso, 2002; Schnur & Lubow, 1976) and duration of the preexposed stimulus (e.g., Ayres, Philbin, Cassidy et al., 1992) for rats, and the perceptual load of the masking task for humans (e.g., Braunstein-Bercovitz, Hen & Lubow, 2004; Braunstein-Bercovitz & Lubow, 1998a), post-preexposure variables, such as time and context (for references, see above) would impact on the integrity of the perceived context–CS union. The perceptual account of LI appears to be particularly compatible with the inverted U-shaped function that relates masking task load to LI (e.g., BraunsteinBercovitz & Lubow, 1998a), where the results can be described as reflecting the load-based distribution of attention between currently relevant (masking task stimuli) and irrelevant stimuli (the preexposed, to-be-relevant stimuli). The LI effects from other manipulations that affect the distribution of attention, such as foveal placement of the preexposed stimulus, re-evoking attention to the preexposed stimulus by pairing it with a novel stimulus, experimentally induced anxiety/stress and related pharmacological treatments, all can be interpreted in terms of the distribution of attention, and the degree to which processing resources are allocated to the irrelevant preexposed stimulus. When such resources are at a minimum, but not completely absent, the salience of the repeatedly preexposed stimulus is reduced and it becomes an integral part of context, which itself is composed of non-salient, unattended stimuli. As a consequence, the preexposed stimulus is at a disadvantage in the test stage, where the subject has to learn either a new association to the previously ignored stimulus (as in the rule-based learning task), or to find it as a target in a visual search task. Until attention is allocated to the previously irrelevant test stimulus, as for example by pairing it with a reinforcer, it remains an undistinguished, integral element of context.
520
R. E. Lubow
Finally, the proposed resolution of the stimulus salience/intensity paradox suggests that LI is a joint function of preexposed stimulus intensity and significance. In general, high stimulus intensity promotes LI and high stimulus significance decreases LI. The attenuated LI in positive-symptom schizophrenic patients can then be ascribed to high levels of top-down generated significance that maintain attention to the PE stimulus and prevent its submersion into the context. On the other hand, the potentiated LI in negative-symptom schizophrenic patients can be attributed to low levels of top-down generated significance that lead to withdrawal of attention from the PE stimulus, which, in turn, encourages the assimilation of the PE stimulus into the context in which it was presented. More generally, it would seem that schizophrenia-spectrum disorders can be differentiated on the basis of two mutually exclusive attentional deficits, a hyperdopaminergic state characterized by overly labile attention leading to high distractibility, reduced LI, and positive symptoms; and a hypo-dopaminergic state distinguished by overly inert attention, fixation on salient stimuli, and negative symptoms. Although speculative, the bipolar hypothesis is easily tested. One requires an experimental procedure that generates scores that can reflect either of the two attentional poles. The LI paradigm meets this requirement, as demonstrated by the fact that identical procedures that produce LI in non-drugged subjects will produce attenuated LI with amphetamine, and super-LI with haloperidol. Furthermore, this is true both for animals and for humans. Likewise, unmedicated acute positive-symptom schizophrenics show attenuated LI, but chronic schizophrenics with a predominance of negative symptoms appear to display a super-LI effect, a relationship that is mirrored in recent studies with schizotypal subjects. Finally, considering the relative complexity of the conditions for acquiring the context–CS integration, it would not be surprising to learn that LI is limited to organisms with relatively well-developed nervous systems, arguably perhaps only mammals (Lubow, this volume). With that said, and irrespective of the fact that the observable symptoms that define schizophrenia are not readily apparent in animal behavior, the confluence of new empirical findings and new theories of LI may provide important insights into the processes that underlie schizophrenia. Hopefully in some time less than another 50 years, a younger colleague will produce a book on LI and schizophrenia that will end with a chapter that is less provisional than the one just presented.
References Abramczyk, R. R., Jordan, D. E., & Hegel, M. (1983). Reverse Stroop effect in the performance of schizophrenics. Perceptual and Motor Skills, 56, 99–106. Andreasen, N. C., & Olsen, S. (1982). Negative vs positive schizophrenia. Archives of General Psychiatry, 39, 789–794. Anscombe, R. (1987). The disorder of consciousness in schizophrenia. Schizophrenia Bulletin, 13, 241–260.
Latent inhibition and schizophrenia: the ins and outs of context
521
Ayres, J. J. B., Philbin, D., Cassidy, S., Belling, L., & Redlinger, E. (1992). Some parameters of latent inhibition. Learning and Motivation, 23, 269–287. Baruch, I., Hemsley, D. R., & Gray, J. A. (1988). Differential performance of acute and chronic schizophrenics in a latent inhibition task. Journal of Nervous and Mental Disease, 176, 598–606. Bouton, M. E. (1993). Context, time, and memory retrieval in the interference paradigms of Pavlovian conditioning. Psychological Bulletin, 114, 80–99. Braff, D. L., & Light, G. A. (2004). Preattentional and attentional cognitive deficits as targets for treating schizophrenia. Psychopharmacology, 174, 75–85. Braunstein-Bercovitz, H. (2003). The modulation of latent inhibition by field-dependency: is it related to the attentional dysfunction in schizotypy? Personality and Individual Differences, 35, 1719–1729. Braunstein-Bercovitz, H., Dimentman-Ashkenazi, I., & Lubow, R. E. (2001). Stress affects the selection of relevant from irrelevant stimuli. Emotion, 1, 182–192. Braunstein-Bercovitz, H., Hen, I., & Lubow, R. E. (2004). Masking task load modulates latent inhibition: support for a distraction model of irrelevant information processing by high schizotypal participants. Cognition and Emotion, 18, 1135–1144. Braunstein-Bercovitz, H., & Lubow, R. E. (1998a). Are high schizotypal normal participants distractible or limited in attentional resources? A study of latent inhibition as a function of masking task load and schizotypy. Journal of Abnormal Psychology, 107, 659–670. Braunstein-Bercovitz, H., & Lubow, R. E. (1998b). Latent inhibition as a function of modulation of attention to the preexposed irrelevant stimulus. Learning and Motivation, 29, 261–279. Burch, G. St. J., Hemsley, D. R., & Joseph, M. H. (2004). Trials-to-criterion latent inhibition in humans as a function of stimulus pre-exposure and positive-schizotypy. British Journal of Psychology, 95, 179–196. Cash, T. F., Neale, J. M., & Cromwell, R. L. (1972). Span of comprehension in acute schizophrenics: full report technique. Journal of Abnormal Psychology, 79, 322–336. Chapman, L. J., & Chapman, J. P. (1973). Disordered Thought in Schizophrenia. New York: Appleton-Century-Crofts. Cohen, E., Sereni, N., Kaplan, O., et al. (2004). The relation between latent inhibition and symptom-types in young schizophrenics. Behavioural Brain Research, 149, 113–122. Cohen, J. D., Barch, D. M., Carter, C., & Servan-Schreiber, D. (1999). Context-processing deficits in schizophrenia: converging evidence from three theoretically motivated cognitive tasks. Journal of Abnormal Psychology, 108, 120–133. Cohen, J. D., & Servan-Schreiber, D. (1992). Context, cortex, and dopamine: a connectionist approach to behavior and biology in schizophrenia. Psychological Review, 99, 45–77. Cohen, J. D., & Servan-Schreiber, D. (1993). A theory of dopamine function and its role in cognitive deficits in schizophrenia. Schizophrenia Bulletin, 19, 85–104. Crowell, C. R., & Anderson, D. C. (1972). Variations in intensity, interstimulus interval and interval between CS exposure and conditioning with rats. Journal of Comparative and Physiological Psychology, 79, 291–298.
522
R. E. Lubow
De la Casa, L. G., & Lubow, R. E. (2001). Latent inhibition with a response time measure from a within-subject design: effects of number of preexposures, masking task, context change, and delay. Neuropsychology, 15, 244–253. De la Casa, L. G., Ruiz, G., & Lubow, R. E. (1993). Latent inhibition and recall/ recognition of irrelevant stimuli as a function of preexposure duration in high and low psychotic-prone normals. British Journal of Psychology, 84, 119–132. Dykes, M., & McGhie, A. A. (1976). A comparative study of attentional strategies in schizophrenics and highly creative normal subjects. British Journal of Psychiatry, 128, 50–56. Escobar, M., Arcediano, F., & Miller, R. R. (2002). Latent inhibition and contextual associations. Journal of Experimental Psychology: Animal Behavior Processes, 28, 123–136. Escobar, M., Arcediano, F., & Miller, R. R. (2003). Latent inhibition in human adults without masking. Learning, Memory and Cognition, 29, 1028–1040. Escobar, M., Oberling, P., & Miller, R. R. (2002). Associative deficit accounts of disrupted latent inhibition and blocking in schizophrenia. Neuroscience and Biobehavioral Reviews, 26, 203–216. Evans, L., Gray, N. S., & Snowden, R. J. (2007). A new continuous within-participant latent inhibition task: examining associations with schizotypy dimensions, smoking and gender. Biological Psychology, 74, 365–373. Eysenck, M. W., Derakshan, N., Santos, R., & Calvo, M. G. (2007). Anxiety and cognitive performance: attentional control theory. Emotion, 7, 336–353. Fanselow, M. S. (1990). Factors governing one-trial contextual conditioning. Animal Learning and Behavior, 18, 264–270. Frame, C. L., & Oltmanns, T. F. (1982). Serial-recall by schizophrenic and affective patients during and after psychotic episodes. Journal of Abnormal Psychology, 91, 311–318. Frith, C. D. (1979). Consciousness, information processing, and schizophrenia. British Journal of Psychiatry, 134, 225–235. Gal, G., Schiller, D., & Weiner, I. (2005). Latent inhibition is disrupted by nucleus accumbens shell lesion but is abnormally persistent following entire nucleus accumbens lesion: the neural site controlling the expression and disruption of the stimulus preexposure effect. Behavioural Brain Research, 162, 246–255. Gibson, E. J. (1969). Principles of Perceptual Learning and Development. New York: Appleton-Century-Crofts. Gilley, D. W., & Franchina, J. (1985). Effects of preexposure flavor concentration on conditioned aversion and neophobia. Behavioral and Neural Biology, 44, 503–508. Ginton, A., Urca, G., & Lubow, R. E. (1975). The effects of preexposure to a nonattended stimulus on subsequent learning: latent inhibition in adults. Bulletin of the Psychonomic Society, 5, 5–8. Gold, P. E., & King, R. A. (1974). Retrograde amnesia: Storage failure versus retrieval failure. Psychological Review, 81, 465–469. Graham, S., & McLaren, I. P. L. (1998). Retardation in human discrimination learning as a consequence of preexposure: latent inhibition or negative priming. Quarterly Journal of Experimental Psychology, 51B, 155–172. Grahame, N. J., Barnet, R. C., Gunther, L. M., & Miller, R. R. (1994). Latent inhibition as a performance deficit resulting from CS-context associations. Animal Learning and Behavior, 22, 395–408.
Latent inhibition and schizophrenia: the ins and outs of context
523
Gray, J. A. (1998). Integrating schizophrenia. Schizophrenia Bulletin, 24, 249–266. Gray, J. A., Feldon, J., Rawlins, J. N. P., Hemsley, D. R., & Smith, A. D. (1991). The neuropsychology of schizophrenia. Behavioral and Brain Sciences, 14, 1–84. Gray, N. S., Fernandez, M., Williams, J., Ruddle, R. A., & Snowden, R. J. (2002). Which schizotypal dimensions abolish latent inhibition? British Journal of Clinical Psychology, 41, 271–284. Gray, N. S., Hemsley, D. R., & Gray, J. A. (1992). Abolition of latent inhibition in acute, but not chronic, schizophrenics. Neurology, Psychiatry and Brain Research, 1, 83–89. Gray, N. S., Pickering, A. D., Hemsley, D. R., Dawling, S., & Gray, J. A. (1992). Abolition of latent inhibition by a single 5 mg dose of d-amphetamine in man. Psychopharmacology, 107, 425–430. Gray, N. S., Pilowsky, L. S., Gray, J. A., & Kerwin, R. W. (1995). Latent inhibition in drug naive schizophrenics: relationship to duration of illness and dopamine D2 binding using SPET. Schizophrenia Research, 17, 95–107. Grossberg, S., Mingolla, E., & Ross, W. D. (1997). Visual brain and visual perception: how does the cortex do perceptual grouping? Trends in Neurosciences, 20, 106–111. Guterman, Y. (2007). A neural plasticity perspective on the schizophrenic condition. Consciousness and Cognition, 16, 400–420. Hall, G. (1991). Perceptual and Associative Learning. Oxford: Clarendon Press. Heinrichs, R. W., & Zakzanis, K. K. (1998). Neurocognitive deficit in schizophrenia: a quantitative review of the evidence. Neuropsychology, 12, 426–445. Hemsley, D. R. (2005). The schizophrenic experience: taken out of context? Schizophrenia Bulletin, 39, 43–53. Hemsley, D. R., & Richardson, P. H. (1980). Shadowing by context in schizophrenia. Journal of Nervous and Mental Diseases, 168, 141–145. Hofer, I., Della Casa, V., & Feldon, J. (1999). The interaction between schizotypy and latent inhibition: modulation by experimental parameters. Personality and Individual Differences, 26, 1075–1088. Honigfeld, G. (1963). The ability of schizophrenics to understand normal, psychotic and pseudo-psychotic speech. Diseases of the Nervous System, 24, 692–694. Johnston, W. A., & Hawley, K. J. (1994). Perceptual inhibition of expected inputs: the key that opens closed minds. Psychonomic Bulletin and Review, 1, 56–72. Kaplan, O., & Lubow, R. E. (2001). Context and reminder effects in a visual search analog of latent inhibition. Learning and Motivation, 32, 137–153. Kapur, S. (2003). Psychosis as a state of aberrant salience: a framework linking biology, phenomenology, and pharmacology in schizophrenia. American Journal of Psychiatry, 160, 13–23. Kiernan, M. J., & Westbrook, R. F. (1993). Effects of exposure to a to-be-shocked environment upon the rat’s freezing response: evidence for facilitation, latent inhibition, and perceptual learning. The Quarterly Journal of Experimental Psychology, 46B, 271–288. Kimchi, R., & Peterson, M. A. (2008). Figure-ground segmentation can occur without attention. Psychological Science, 19, 660–668. Kraeplin, E. (1919/1971). Dementia Praecox and Paraphrenia. Huntington, NY: Krieger. Lavie, N. (1995). Perceptual load as a necessary condition for selective attention. Journal of Experimental Psychology: Human Perception and Performance, 21, 451–468.
524
R. E. Lubow
Lavie, N. (2005). Distracted and confused? Selective attention under load. Trends in Cognitive Sciences, 9, 75–82. Lawson, J. S., McGhie, A., & Chapman, J. (1964). Perception of speech in schizophrenia. British Journal of Psychiatry, 110, 375–380. Lawson, J. S., McGhie, A., & Chapman, J. (1967). Distractibility in schizophrenia organic brain disease. British Journal of Psychology, 113, 527–535. Leumann, L., Feldon, J., Vollenweider, F. X., & Ludewig, K. (2002). Effects of typical and atypical antipsychotics on prepulse inhibition and latent inhibition in chronic schizophrenia. Biological Psychiatry, 52, 729–739. Lipp, O. V., Siddle, D. A. T., & Vaitl, D. (1992). Latent inhibition in humans: single-cue conditioning revisited. Journal of Experimental Psychology: Animal Behavior Processes, 18, 115–125. Lipp, O. V., & Vaitl, D. (1992). Latent inhibition in human Pavlovian conditioning: effect of additional stimulation after preexposure and relation to schizotypal traits. Personality and Individual Differences, 13, 1003–1012. Lubow, R. E. (1973). Latent inhibition. Psychological Bulletin, 79, 398–407. Lubow, R. E. (1989). Latent Inhibition and Conditioned Attention Theory. New York: Cambridge University Press. Lubow, R. E. (2005). The construct validity of the animal-latent inhibition model of selective attention deficits in schizophrenia. Schizophrenia Bulletin, 31, 139–153. Lubow, R. E. (2009). Conditioned taste aversion and latent inhibition: a review. In S. Reilly & T. Schachtman (Eds.), Conditioned Taste Aversion: Behavioral and Neural Processes. Oxford: Oxford University Press. Lubow, R. E., Caspy, T., & Schnur, P. (1982). Latent inhibition and learned helplessness in children: similarities and differences. Journal of Experimental Child Psychology, 34, 231–256. Lubow, R. E., & De la Casa, L. G. (2005). There is a time and a place for everything: bi-directional modulations of latent inhibition by time-induced context differentiation. Psychonomic Bulletin & Review, 12, 806–821. Lubow, R. E., & Gewirtz, J. C. (1995). Latent inhibition in humans: data, theory, and implications for schizophrenia. Psychological Bulletin, 117, 87–103. Lubow, R. E., Ingberg-Sachs, Y., Zalstein-Orda, N., & Gewirtz, J. C. (1992). Latent inhibition in low and high “psychotic-prone” normal subjects. Personality and Individual Differences, 13, 563–572. Lubow, R. E., & Kaplan, O. (1997). Visual search as a function of type of prior experience with target and distractor. Journal of Experimental Psychology: Human Perception and Performance, 23, 14–24. Lubow, R. E., & Kaplan, O. (2005). The visual search analog of latent inhibition: implications for theories of irrelevant stimulus processing in normal and schizophrenic groups. Psychonomic Bulletin & Review, 12, 224–243. Lubow, R. E., Kaplan, O., Abramovich, P., Rudnick, A., & Laor, N. (2000). Visual search in schizophrenia: latent inhibition and novel pop-out effects. Schizophrenia Research, 45, 145–156. Lubow, R. E., Kaplan, O., & De la Casa, G. (2001). Performance on the visual search analog of latent inhibition is modulated by an interaction between schizotypy and gender. Schizophrenia Research, 52, 275–287. Lubow, R. E., Rifkin, B., & Alek, M. (1976). The context effect: the relationship between stimulus preexposure and environmental preexposure determines
Latent inhibition and schizophrenia: the ins and outs of context
525
subsequent learning. Journal of Experimental Psychology: Animal Behavior Processes, 2, 38–47. Lubow, R. E., Weiner, I., Schlossberg, A., & Baruch, I. (1987). Latent inhibition and schizophrenia. Bulletin of the Psychonomic Society, 25, 464–467. Lubow, R. E., Weiner, I., & Schnur, P. (1981). Conditioned attention theory. In G. H. Bower (Ed.), The Psychology of Learning and Motivation, vol. 15. New York: Academic Press, pp. 1–49. Mackintosh, N. J. (1975). A theory of attention: variations in the associability of stimuli with reinforcement. Psychological Review, 82, 276–298. Mason, O., Claridge, G., & Jackson, M. (1995). New scales for the assessment of schizotypy. Personality and Individual Differences, 18, 7–13. Mass, R., Schoemig, T., Hitschfeld, K., Wall, E., & Haasen, C. (2000). Psychopathological syndromes of schizophrenia: evaluation of the dimensional structure of the Positive and Negative Syndrome Scale (PANSS). Schizophrenia Bulletin, 26, 167–177. McGhie, A., & Chapman, J. (1961). Disorders of attention and perception in early schizophrenia. British Journal of Clinical Psychology, 34, 103–116. McLaren, I. P. L., Kaye, H., & Mackintosh, N. J. (1989). An associative theory of the representation of stimuli: application to perceptual learning and latent inhibition. In R. G. M. Morris (Ed.), Parallel Distributed Processing: Implications for Psychology and Neurobiology. Oxford: Clarendon Press, pp. 102–130. McLaren, I. P. L., & Mackintosh, N. J. (2000). An elemental model of associative learning: I. Latent inhibition and perceptual learning. Animal Learning & Behavior, 28, 211–246. Miller, R. R., & Escobar, M. (2001). Contrasting acquisition-focused and performance-focused models of acquired behavior. Current Directions in Psychological Science, 10, 141–145. Miller, R. R., Kasprow, W. J., & Schachtman, T. R. (1986). Retrieval variability: sources and consequences. American Journal of Psychology, 99, 145–218. Miller, R. R., & Springer, A. D. (1973). Implications of recovery from experimental amnesia. Psychological Review, 81, 470–473. Minas, R. K., & Park, S. (2007). Attentional window in schizophrenia and schizotypal personality. Applied & Preventive Psychology, 12, 140–148. Mirsky, A. F., & Duncan, C. C. (1986). Etiology and expression of schizophrenia: neurobiological and social factors. Annual Review of Psychology, 37, 291–319. Moore, C. M., Grosjean, M., & Lleras, A. E. (2003). Using inattentional blindness as an operational definition of unattended: the case of surface completion. Visual Cognition, 10, 299–318. Moritz, S., Andresen, B., Jacobsen, D., et al. (2001). Neuropsychological correlates of schizophrenic syndromes in patients treated with atypical neuroleptics. European Psychiatry, 16, 354–361. Morris, C. D., Bransford, J. D., & Franks, J. J. (1977). Levels of processing versus transfer appropriate processing. Journal of Verbal Learning and Verbal Behavior, 16, 519–533. Neale, J. M., McIntyre, C. W., Fox, R., & Cromwell, R. L. (1969). Span of apprehension in acute schizophrenics. Journal of Abnormal Psychology, 74, 593–596. Nuechterlein, K. H., & Dawson, M. E. (1984). Information processing and attentional functioning in the developmental course of the schizophrenic disorder. Schizophrenia Bulletin, 10, 160–203.
526
R. E. Lubow
Nutt, D. J., Bell, C. J., & Malizia, A. L. (1998). Brain mechanisms of social anxiety disorder. Journal of Clinical Psychiatry, 12, 31–38. Oberling, P., Gosselin, O., & Miller, R. R. (1998), Latent inhibition in animals as a model of acute schizophrenia: a re-analysis. In M. Haug & R. E. Whalen (Eds.), Animal Models of Human Emotion and Cognition. Washington DC: American Psychological Association, pp. 87–102. Ohman, A. (1979). The orienting response, attention, and learning: an informationprocessing perspective. In E. D. Kimmel, E. H. van Olst, & J. F. Orlebeke (Eds.), The Orienting Reflex in Humans. Hillsdale, NJ: Erlbaum, pp. 443–471. Oltmanns, T. F., Ohayon, J., & Neale, J. M. (1978). The effects of antipsychotic medication and diagnostic criteria on distractibility in schizophrenia. Journal of Psychiatric Research, 14, 81–91. Pavlov, I. P. (1927). Conditioned Reflexes (Trans. G.V. Anrep). London: Oxford University Press. Pearce, J. M., & Hall, G. (1980). A model of Pavlovian learning: variations in the effectiveness of conditioned but not unconditioned stimuli. Psychological Review, 87, 532–552. Peterson, J. B., & Carson, S. (2000). Latent inhibition and openness to experience in a high-achieving student population. Personality and Individual Differences, 28, 323–332. Phillips, W. A., & Singer, W. (1997). In search of common foundations for cortical computation. Behavioral and Brain Sciences, 20, 657–683. Pinen˜o, O., De la Casa, L. G., Lubow, R. E., & Miller, R. R. (2006). Some determinants of latent inhibition in human predictive learning. Learning and Motivation, 37, 42–65. Place, E. J., & Gilmore, G. C. (1980). Perceptual organization in schizophrenia, Journal of Abnormal Psychology, 89, 409–418. Pogue-Geile, M. F., & Oltmanns, T. F. (1980). Sentence perception and distractibility in schizophrenic, depressed, and manic patients. Journal of Abnormal Psychology, 89, 115–124. Prados, J. (2000). Effects of varying the amounts of preexposure to spatial cues on a subsequent navigational task. Quarterly Journal of Experimental Psychology, 53B, 139–148. Raine, A. (1991). The SPQ: a scale for the assessment of schizotypal personality based on DSM-III-R criteria. Schizophrenia Bulletin, 17, 555–564. Rascle, C., Mazas, O., Vaiva, G., et al. (2001). Clinical features of latent inhibition in schizophrenia. Schizophrenia Research, 51, 149–161. Rochester, S. R. (1978). Are language disorders in acute schizophrenics actually information processing problems? Journal of Psychiatric Research, 14, 275–283. Rock, I., & Gutman, D. (1981). The effect of inattention on form perception. Journal of Experimental Psychology: Human Perception and Performance, 7, 275–285. Rodriguez, G., & Alonso, G. (2002). Latent inhibition as a function of CS intensity in taste aversion learning. Behavioural Processes, 60, 61–67. Russell, C., & Driver, J. (2005). New indirect measures of “inattentive” visual grouping in a change-detection task. Perception & Psychophysics, 67, 606–623. Rutter, D. R., Wishner, J., & Callaghan, B. A. (1975). The predictions of predictability of speech in schizophrenic subjects. British Journal of Psychiatry, 126, 571–576.
Latent inhibition and schizophrenia: the ins and outs of context
527
Schiller, D., Zuckerman, L., & Weiner, I. (2006). Abnormally persisitent latent inhibition induced by lesions to the nucleus accumbens core, basolateral amygdala and orbitofrontal cortex is reversed by clozapine but not by haloperidol. Journal of Psychiatric Research, 40, 167–177. Schmidt-Hansen, M., Killcross, A. S., & Honey, R. C. (2009). Latent inhibition, learned irrelevance and schizotypy: assessing their relationship. Cognitive Neuropsychiatry, 14, 11–29. Schneider, S. J. (1976). Selective attention in schizophrenia. Journal of Abnormal Psychology, 85, 167–173. Schneider, W., & Shiffrin, R. M. (1977). Controlled and automatic human information processing: detection, search and attention. Psychological Review, 84, 1–66. Schnur, P., & Lubow, R. E. (1976). Latent inhibition: the effects of ITI and CS intensity during preexposure. Learning and Motivation, 7, 540–550. Serra, A. M., Jones, S. H., Toone, B., & Gray, J. A. (2001). Impaired associative learning in chronic schizophrenics and their first-degree relatives: a study of latent inhibition and the Kamin blocking effect. Schizophrenia Research, 48, 273–289. Shakow, D. (1962). Segmental set. Archives of General Psychiatry, 6, 1–17. Shiffrin, R. M., & Schneider, W. (1977). Controlled and automatic attentional processing II: perceptual learning, automatic attending and a general theory. Psychological Review, 84, 127–190. Silverstein, S. M., Knight, R. A., Schwarzkopf, S. B., et al. (1996). Stimulus configuration and context effects in perceptual organization in schizophrenia. Journal of Abnormal Psychology, 105, 410–420. Silverstein, S. M., Kovacs, I., Corry, R., & Valone, C. (2000). Perceptual organization, the disorganization syndrome, and context processing in chronic schizophrenia. Schizophrenia Research, 43, 11–20. Sitskoorn, M. M., Salden, M., & Kahn, R. S. (2001). Latent inhibition in first-episode patients with schizophrenia. Schizophrenia Research, 49, 145–146 (abstract). Sokolov, E. N. (1963). Perception and the Conditioned Reflex. New York: Pergamon Press. Stout, S. C., & Miller, R. R. (2007). Sometimes-competing retrieval (SOCR): a formalization of the comparator hypothesis. Psychological Review, 114, 759–783. Swerdlow, N. R., Braff, D. L., Hartston, H., Perry, W., & Geyer, M. A. (1996). Latent inhibition in schizophrenia. Schizophrenia Research, 20, 91–103. Thompson, R. F. (1972). Sensory preconditioning. In R. F. Thompson & M. R. Voss (Eds.), Topics in Learning and Performance. New York: Academic Press, pp. 239–271. Treisman, A. (1993). The perception of features and objects. In A. Baddeley & L. Weiskrantz (Eds.), Attention: Selection, Awareness, and Control. Oxford: Clarendon Press, pp. 5–35. Tsakanikos, E., Sverdrup-Thygenson, L., & Reed, P. (2003). Latent inhibition and psychosis-proneness: visual search as a function of pre-exposure to the target and schizotypy level. Personality and Individual Differences, 34, 575–589. Tulving, E., & Thomson, D. M. (1973). Encoding specificity and retrieval processes in episodic memory. Psychological Review, 80, 352–373.
528
R. E. Lubow
Uhlhaas, P. J., & Silverstein, S. M. (2005). Perceptual organization in schizophrenia spectrum disorders: empirical research and theoretical implications. Psychological Bulletin, 131, 618–632. Usai, M. C., Umilta, C., & Nicoletti, R. (1995). Limits in controlling the focus of attention. European Journal of Cognitive Psychology, 7, 411–439. Vaitl, D., Lipp, O. V., Bauer, U., et al. (2002). Latent inhibition and schizophrenia: Pavlovian conditioning of autonomic responses. Schizophrenia Research, 55, 147–158. Wagner, A. R. (1976). Priming in STM: an information processing mechanism for self-generated or retrieval-generated depression in performance. In T. Tighe & R. N. Leaton (Eds.), Habituation: Perspectives from Child Development, Animal Behavior, and Neurophysiology. Hillsdale, NJ: Lawrence Erlbaum, pp. 95–128. Wagner, A. R. (1978). Expectancies in the priming of STM. In S. H. Hulse, H. Fowler, & W. K. Honig (Eds.), Cognitive Processes in Animal Behavior. Hillsdale, NJ: Erlbaum, pp. 53–82. Wagner, A. R. (1981). SOP: A model of automatic memory processing in animal behavior. In N. E. Spear & R. R. Miller (Eds.), Information Processing in Animals: Memory Mechanisms. Hillsdale, NJ: Lawrence Erlbaum, pp. 5–47. Wapner, S., & Krus, D. M. (1960). Effects of lysergic acid diethylamide and differences between normals and schizophrenics on the Stroop color-word test. Journal of Neuropsychiatry, 2, 76–81. Watanabe, T., Nanez, J. E., & Sasaki, Y. (2001). Perceptual learning without perception. Nature, 413, 844–848. Weiner, I. (1990). The neural substrates of latent inhibition. Psychological Bulletin, 108, 442–461. Weiner, I. (2003). The “two-headed” latent inhibition model of schizophrenia: modeling positive and negative symptoms and their treatment. Psychopharmacology, 169, 257–297. Wells, D. S., & Leventhal, D. (1984). Perceptual grouping in schizophrenia: a replication of Place and Gilmore. Journal of Abnormal Psychology, 93, 231–234. Williams, J. H., Wellman, N. A., Geaney, D. P., et al. (1998). Reduced latent inhibition in people with schizophrenia: an effect of psychosis or its treatment. British Journal of Psychiatry, 172, 243–249. Wuthrich, V., & Bates, T. (2001). Schizotypy and latent inhibition: non-linear linkage between psychometric and cognitive markers. Personality and Individual Differences, 30, 783–798. Zalstein-Orda, N., & Lubow R. E. (1995). Context control of negative transfer induced by preexposure to irrelevant stimuli: latent inhibition in humans. Learning and Motivation, 26, 11–28.
Summary and conclusions
22 Issues in latent inhibition research and theory: an overview R. E. Lubow and Ina Weiner
In composing this final chapter, we were faced with a number of decisions. On the one hand, we thought it useful to point out the shared elements and agreements amongst the chapters, as well as their differences and novel contributions. On the other hand, in spite of our efforts to ensure that the chapters of the book covered all of the important issues in LI research and theory, a number of topics remained either unresolved or unexplored. As a compromise, we have opted for the broadest possible summary, recognizing that it will be at the expense of depth, a debt that can be redeemed only by writing another book. The present chapter, then, begins by summarizing a few latent inhibition (LI) facts and the major themes that are common to several or more chapters. Following this, methodological issues that have not been resolved, and which, we believe, deserve the future attention of researchers, are addressed. These include the role of masking, the need for new LI procedures, the distinction between two- and three-stage LI procedures, and the dual source of LI effects. The next major section takes up theoretical issues, examining specific behavioral theories of LI, with an attempt to integrate them, while pointing out the relevance for future research. The final section discusses the neuropsychology of LI particularly as it relates to schizophrenia.
A few latent inhibition facts Foremost, the LI effect is exceedingly robust. To confirm this statement, one need only to refer to the chapters in the current volume, which together cite hundreds of studies that have shown that performance on a learning task is poorer with a stimulus that was presented earlier (PE) without a consequence (CS–0) as compared to a novel stimulus (NPE). Although after 50 years of research there is still no generally accepted theory of LI, there are now major areas of agreement as well as a number of divergences, as can be seen by comparing several of the chapters in this volume (e.g., Escobar & Miller; Hall & Rodriguez; Schmajuk; Cassaday & Moran; Lubow). On the other hand, the phenomenon Latent Inhibition: Cognition, Neuroscience and Applications to Schizophrenia, ed. R.E. Lubow and Ina Weiner. Published by Cambridge University Press # Cambridge University Press 2010.
531
532
R. E. Lubow and Ina Weiner
of learning, itself, with a history going back to ancient Greece, and a truly massive accumulation of research data in the twentieth century, still finds itself the subject of vigorous theoretical debate. Considering the number of LI theories, and the absence of specific confrontations amongst them, one might conclude that the theories are less important than the fact that LI can be modulated by any one of a long list of variables (e.g., number of preexposures, intensity of the preexposed stimulus, context, retention interval duration), as well as by a large number of pharmacological and neurophysiological interventions (e.g., Gould; Lipina & Roder; Louilot et al., Schnur & Hoffman; Weiner; Weiner & Arad; all this volume), and by individual difference variables such as schizophrenia and schizotypy (Kumari & Ettinger, this volume) anxiety (BraunsteinBercovitz, this volume) or creativity (Carson, this volume). Just as with learning, the absence of a complete theoretical consensus has not impeded the use of the LI effect for practical applications. Thus, LI has been adopted as a means for reducing anticipatory nausea from radiation and chemotherapy, for preventing the acquisition of dental and school phobias (for a review of these and other LI-related prophylactic treatments, see Lubow, 1998), as well as for screening drugs that are potentially therapeutic for schizophrenia (Weiner, this volume).
Methodological issues The present section focuses on methodology. However, as will be seen, methodological concerns often have implications for theory.
LI procedures Animal LI research is very consistent in terms of the procedures used to measure LI, specifically thirst-motivated conditioned suppression, conditioned taste aversion, conditioned avoidance, and, in recent years, conditioned freezing, particularly in mice. Importantly, unlike other widely used procedures (e.g., water maze), rat LI has been easily adapted to mice. The paucity of procedures using appetitive reinforcement and instrumental conditioning is sometimes seen as a problem, but those studies that have used them yield identical results to the aversively motivated ones, both in regard to the basic LI effect (e.g., Lubow, Rifkin & Alek, 1976, Exp. 1; De la Casa, Marquez & Lubow, 2009) as well as the effects from lesion and drug manipulations (e.g., Coutereau, Blundell & Killcross, 2001; Norman & Cassaday, 2004). In contrast, human LI research has used a variety of tasks, many of which have only been used a single time, and whose capacity to measure LI is questionable. Certainly some of the LI tasks have not provided replicable modulation of LI effects with schizophrenia patients (for review, see Swerdlow, this volume). In addition, a number of studies with human subjects that have reported performance differences in PE and NPE groups/conditions may have misidentified the effects as LI (e.g., Escobar,
Issues in latent inhibition research and theory: an overview
533
Arcediano & Miller, 2003; Evans, Gray & Snowden, 2007; Gal, Barnea, Biran et al., 2009). In these experiments, designed to produce LI without a masking task, the response required from the subject in the preexposure stage was the same as the one in the test phase. However, during preexposure the PE stimulus was initially paired with not responding, thus making it likely that test phase data for the preexposed group/condition represented a negative transfer effect. For example, prior to the preexposure stage, Evans et al. (2007) instructed subjects to press the keyboard space bar if they thought that a particular shape would be followed by the letter X. However, the X never appeared in this stage, and only occurred later in the test stage when preceded by either one of the preexposed shapes or a new shape. Thus, the preexposed shape, but not the new shape, was initially paired with not responding to the space bar. As opposed to this, in the traditional human LI task the response to the masking task is qualitatively different from the test task response, thereby precluding response interference effects (e.g., Braunstein-Bercovitz & Lubow, 1998; Ginton, Urca & Lubow, 1975). Many within-subject LIRR experiments (e.g., Gal, Mendlovic, Bloch et al., 2005; Orosz, Feldon, Gal et al., 2008; Young, Kumari, Mehrotra et al., 2004) are open to the same criticism regarding response interference.
Is a masking task necessary for inducing LI? The methodological and theoretical implications of masking were reviewed by Lubow (this volume), where it was concluded that the masking task is required to generate LI in humans because it diverts attention from the preexposed to-be-CS. Although Le Pelley and Schmidt-Hansen (this volume) have proposed that the masking condition used in human LI experiments creates a learned irrelevance paradigm (LIRR; preexposures of unpaired CSs and USs), one can also argue the converse, namely that the LIRR procedure provides an even stronger functional masking task than a typical LI procedure. Presenting the to-be-CS in the same context as the relatively more salient to-be-US may divert attention from the CS. Indeed, this may account for the findings that LIRR effects often are larger than LI effects, a difference that might be enhanced even further by the fact that the US in the LIRR paradigm is present in preexposure and test, thus making the contexts of the two stages more similar to each other than those within an LI experiment (Killcross & Dickinson, 1996). Nevertheless, a number of authors have suggested several reasons why the LIRR preparation offers significant benefits compared to LI: (1) the former effect is stronger than the latter; (2) the apparent ease of using the LIRR procedure in a within-subject PE/NPE design; (3) animal LI and human LI have different underlying processes, based on the contention that the latter, but not the former, requires a masking task (e.g., Gray, N. S., & Snowden, 2004; Le Pelley & Schmidt-Hansen, this volume). On this basis, the LIRR paradigm has been used in some recent studies with schizophrenia patients (Gal et al., 2005; Orosz et al., 2008; Young et al., 2004).
534
R. E. Lubow and Ina Weiner
In summary, on the one hand, LIRR appears to be a variation of LI, but generated with a more complex preexposure procedure than traditional LI. If, on the other hand, LIRR is not LI, then its shortcomings as a research protocol with schizophrenia patients would be even more pronounced. The animal and human LI protocols are more closely matched, both procedurally and theoretically, than those of LIRR. Unlike LIRR, LI and its obligatory masking task have theoretical roots that directly relate to attentional impairments in schizophrenia (Lubow, 2005; Lubow & Gewirtz, 1995). Specifically, as already noted, the directions in which the masking task is able to modulate attention to the preexposed stimulus, producing either attenuated or potentiated LI, should correspond to the types of aberrations in attention in patients with positive and negative symptoms, respectively. Furthermore, there is a rich psychopharmacological literature on LI in animals and humans that directly relates to dopamine and other modulators of schizophrenia (Schnur & Hoffman, this volume; Weiner & Arad, this volume), as well as a large body of animal pharmacological, neurodevelopmental, and genetic research that converges to make LI widely accepted as relevant to schizophrenia. A comparable literature does not exist for LIRR. In short, the knowledge base for, and the theoretical links to, LI and schizophrenia are considerably more extensive than for LIRR and schizophrenia. However, if the masking task is a necessary condition for producing LI in humans but not, as some have claimed, in non-human animals, then this would suggest some difference in the processes that modulate attention to inconsequential stimuli between these groups. Nonetheless, a common ground for human and animal LI can be recaptured by noting that animal LI preexposure procedures elicit exploratory behaviors that may serve the same function as the masking procedure. Each time that the animal is placed in the apparatus during the preexposure stage, it actively explores the new environment. Consequently, on many trials the to-be-CS is presented while the animal is engaged in these behaviors, thereby functioning as a masking task to divert attention from the to-be-CS. The two most often used LI procedures, off-baseline conditioned emotional response and conditioned taste aversion, provide particularly strong cases for inferring the presence of an attention-diverting stimulus in animal LI preparations, since in the former, the preexposed stimulus appears while the water-deprived animal is denied the usual access to water, and in the latter, the preexposed flavor is consumed while the waterdeprived animal is slaking its thirst (see below, “Neuropsychology of latent inhibition”, for a comment on the masking task and the validity of the LI model of schizophrenia).
New procedures for generating LI The need to develop new procedures for assessing LI has been driven primarily by the growing research with pathological groups, such as schizophrenia patients. Specifically, there are needs for a viable within-subject procedure and for one that is sensitive to persistent or super-LI, as well as well as attenuated LI.
Issues in latent inhibition research and theory: an overview
535
Within-subject LI procedures LI experiments, whether with animal or human subjects, typically use procedures in which PE and NPE conditions are administered to different groups. As opposed to this between-subject design, in within-subject procedures each subject serves as its own control for the PE and NPE conditions. Although such a design promotes stimulus generalization between the PE and NPE stimuli, which would weaken the LI effect, there are several good reasons for developing and using reliable within-subject LI procedures. LI research with pathological groups would benefit from having the PE and NPE conditions embedded in a within-subject design for the following reasons. (1) The difficulty in obtaining subjects for these groups dictates a design that reduces variance between the PE and NPE conditions, allowing effects to be detected with fewer subjects than in a between-subject design. (2) Relatedly, the many problems encountered in matching subjects across PE and NPE conditions argues for a within-subject design, which, as such, produces perfect matching. This would be particularly important for assessing effects in notoriously heterogenetic schizophrenia groups. (3) The study of individual differences would benefit from the availability of an index of subject-specific LI. The dangers of using averaged group data, particularly with pathological populations, have been reviewed elsewhere (e.g., Caramazza, 1986; Shallice, Burgess & Frith, 1991). Until recently, only one within-subject LI experiment with adult human subjects had been published (Gray, N. S., Pilowsky, Gray & Kerwin, 1995). Although that study, adapted from Ginton, Urca and Lubow (1975), reported an LI effect in the control group and attenuated LI in the schizophrenia patient group, the PE and NPE stimuli were not counterbalanced, a problem that continues to be encountered (e.g., Barrett, Bell, Watson & King, 2004; McCartan, Bell, Green et al., 2001; Stevens, Peschk & Schwarz, 2007; Vaitl, Lipp, Bauer et al., 2002). A number of more recent studies also have attempted to develop viable within-subject LI procedures, but with mixed results (De la Casa & Lubow, 2001; Evans, Gray & Snowden, 2007; Gal, Barnea, Biran et al., 2009; Swerdlow, Stephany, Wasserman et al., 2003, 2005). For example, De la Casa and Lubow (2001) produced LI in four PE/NPE within-subject experiments. However, the LI effects were obtained with response time (RT) but not with number of correct responses, as would be expected since responding to the test-stage task was largely determined by detection as opposed to rule-based processes (see next section). That RT-LI was reduced by decreasing the number of stimulus preexposures, by omitting the masking task, by changing the context from the preexposure to the test stage, and by introducing a delay between the two stages suggests that the within-subject RT-based LI reflects the same processes as those that govern between-subject LI with correct responses as the dependent measure. Response time and correct response measures Older LI experiments with human subjects almost exclusively used some index of correct responding as the dependent measure. However, many newer studies have come to rely on RT, particularly in within-subject designs. Although RT may be a
536
R. E. Lubow and Ina Weiner
more sensitive measure of human LI than correct responses, several other factors may explain why some within-subject procedures have been more successful in producing LI with the latter than with the former. In general, LI studies in which RT is the major dependent variable deemphasize, or even eliminate, the need for discovering a correct response. For example, the preexposure and test-stage instructions for visual search LI require subjects to view a series of frames with many figures on them, and to press one key if all of the items are the same, and a different key if there is a unique item (Lubow & Kaplan, 1997). In other words, the subject is not required to discover a rule, but rather to detect a perceptual difference amongst stimuli. By comparison, in the test stage of a typical LI experiment that uses correct responses as the dependent measure (e.g., Braunstein-Bercovitz & Lubow, 1998), subjects are informed that they are starting a new task in which anything on the screen may be relevant, and they are told to find the rule that relates “something on the screen” to the immediately following consequence. There are a number of advantages of RTs over correct responses; RTs are relatively free of restrictions, such as from floor and ceiling effects and dichotomous data that frequently accompany correct response measures in human LI experiments. These effects can interfere with the assessment of the preexposure effect. In general, with a rule-learning task, the LI effect should appear with correct responses, but not RTs. On the other hand, with a detection task, the LI effect should emerge only with RTs. Consequently, within each type of task, number of correct responses and RTs ought not to be correlated. However, administering the different types of tasks to the same subjects should, and does, lead to a significant positive correlation between the two measures (Gibbons, Rammsayer & Lubow, 2001), indicating that both correct response and RT measures, when used with their appropriate tasks, namely rule-learning and detection, assess the same underlying attentional process.
Procedures for detecting attenuated and persistent LI The effects of pharmacological, physiological and genetic manipulations in animals are typically measured in terms of a reduction or an abolition of the target behavior in comparison to the non-treated controls. However, for LI research, the assessments of pharmacological and neurophysiological effects have taken a different path, focusing on both the disruption and the induction of the phenomenon. The latter effect, termed LI potentiation, enhancement, or persistence, is indexed by comparison to the absence of LI in non-treated controls (see this volume, Weiner; Weiner & Arad; Cassaday & Moran; Lipina & Roder). The best-documented instance of persistent LI, demonstrated in both animals and humans, is that produced by dopaminergic blockade. In recent years, many other manipulations, including neurodevelopmental and genetic (see Lipina & Roder, this volume), have been shown to produce this effect in animals.
Issues in latent inhibition research and theory: an overview
537
Indeed, because the notion that abnormal LI equals abolished LI is deeply entrenched, much information obtained using LI in various pharmacological, neurodevelopmental and genetic animal models is missed and/or misinterpreted. Consequently, it is important to stress that the presence of LI following pharmacological, neurodevelopmental or genetic manipulations does not allow a conclusion that LI is unaffected, unless it also can be shown that the observed LI is not a case of persistent LI. Likewise, when schizophrenic patients display normal LI, it may well be undetected persistent LI, which might appear as normal LI unless uncovered by an appropriate procedure. Although several studies have explored this possibility (Cohen et al., 2004; Gal et al., 2009; Rascle et al., 2001), much effort is still needed, particularly in regard to within-subject designs. Moreover, groups of schizophrenic patients, no matter how categorized (e.g. by symptoms, stage of illness, etc), will still be heterogeneous, including individuals who will and will not exhibit LI. Therefore, for this population, procedures need to be developed that can identify both LI attenuation and persistence. Such procedures would have a lengthy test phase in which healthy controls initially show poorer performance to the preexposed compared to the nonpreexposed stimulus (LI) followed by recovery of performance to the preexposed stimulus (no LI). Disrupted LI would be manifested as equally good early-stage performance to the preexposed and nonpreexposed stimuli whereas persistent LI would be manifested as longer-lasting poor performance to the preexposed stimulus. The likelihood that some individuals within groups of schizophrenic patients will have LI, and others will not, further underscores the need for within-subject designs. Clearly, group statistics would mask the differential manifestations of LI and thus preclude appropriate interpretation of the results. On the other hand, within-subject designs provide individual LI scores that can be used either as a continuous measure or for grouping subjects into those showing or not showing LI. These data then can be correlated with patient characteristics such as symptom domains, illness duration, drug response, performance on other tasks, etc., thus delineating distinctive subsets of patients, perhaps characterized by attenuated or persistent LI and related to overly labile and overly inert attentional processes.
Two- and three-stage procedures LI experiments use procedures with either two or three stages. With animals, the two-stage LI procedure is typically conducted with conditioned avoidance and with classical conditioning procedures such as eyelid conditioning and conditioned freezing. However, the more common animal LI experiments use three-stage procedures, most often conditioned suppression or conditioned taste aversion (for a review of the latter, see Lubow, 2009). On the other hand, human LI studies, with the exception of several electrodermal conditioning experiments (e.g., Lipp, Siddle & Vaitl, 1992),
538
R. E. Lubow and Ina Weiner
and some variations on the animal conditioned suppression protocol (e.g., Salgado, Hetem, Vidal et al., 2000), are composed of only two stages (e.g., Braunstein-Bercovitz & Lubow, 1998). In addition to the basic difference between two- and three-stage procedures, namely, that the test stage of the former is conducted under conditions of reinforcement and that of the latter under conditions of extinction, the two procedures differ in the duration of conditioning and the amount of time elapsing between stimulus preexposure and test. The confounding of number of stages and type of the experimental protocol, and the fact the results from three-stage procedures cannot differentiate between effects derived directly from the stage-one manipulation and those that also may be mediated by effects from stage one on stage two, may have contributed to the apparent conflict between A- and R-theories of LI.
The PE group and the NPE group both contribute to the LI effect The PE and NPE groups nominally represent “treated” and “un-treated” subjects, respectively. As a consequence, given a significant difference in some dependent measure in the test stage (i.e., LI), the effect is normally attributed to the “treated”, PE group. Indeed, this may be the correct conclusion. Equally logical, although less compelling, is the inference that the LI effect is attributable to some process associated with the NPE group. Recall that in the stimulus preexposure stage, the PE group receives a number of trials each containing the same to-be-conditioned stimulus presented in the same context. The NPE group is treated in an identical manner, except that the preexposed stimulus is absent. Thus, in the test stage, the PE group encounters a familiar stimulus in a familiar context, and the NPE group is faced with a novel stimulus in a familiar context. Since LI is defined as relatively poor performance of the PE group compared to the NPE group, the LI effect could derive from independent processes, some operating on the PE group and the other on the NPE group; for example, a preexposure-induced decline in stimulus attention/salience such that the stimulus is less available for processing in the test-stage (e.g., Lubow, Weiner & Schnur, 1981; Mackintosh, 1975; Pearce & Hall, 1980), and a test-stage factor, in which for the NPE group the novel stimulus appearing in a familiar context captures attention (novel pop-out; see Johnston & Hawley, 1994), and, as such, facilitates learning. In the first case, relatively poor learning by the PE group is induced by an acquired loss of stimulus salience, whereas in the second case relatively good learning by the NPE group is promoted by enhanced intrinsic stimulus salience. Such a two-factor position has direct empirical support in animal and human studies (e.g., Lubow & Kaplan, 2005; Lubow, Rifkin & Alek, 1976). The above analysis has implications also for aberrant LI effects in schizophrenia patients and high-schizotypal normals. Until very recently, attenuated LI in patients has been expected from, and attributed to, a process operating on the preexposed stimulus, as
Issues in latent inhibition research and theory: an overview
539
for example the maintenance of attention to the irrelevant preexposed stimulus (for reviews, Kumari & Ettinger, this volume; Lubow, 2005). However, since the NPE group also contributes to the LI effect, the locus of the disruption may, at least in part, reside in that group. In fact, some studies have reported that disrupted LI in schizophrenia can stem from associative deficits in the NPE group (e.g., Swerdlow, this volume; Serra et al., 2001). Such an effect would be expected if the patient is not able to profit from the normal increase in salience when a novel stimulus is presented in a familiar context. In short, the heterogeneity of LI effects in schizophrenia patients might be accounted for by impaired processing of either the PE stimulus or the NPE stimulus, or both. Knowledge of pharmacological and physiological manipulations that produce differential PE and NPE effects can suggest mechanisms that might be involved in schizophrenia (for a more detailed discussion, see Weiner & Arad, this volume).
Theoretical issues The hallmark of an LI experiment is a series of stimulus exposures that are not followed by consequences. However, the basic preexposure procedure also is central to the generation of many other phenomena, some of which lead to performance decrements, as for example habituation (e.g., Honey, Iordanova & Good, this volume), and, with the addition of a non-contingent US, learned irrelevance (e.g., LePelley & Schmidt-Hansen, this volume), or when stages 1 and 2 are reversed (extinction; Westbrook & Bouton, this volume), while others produce performance enhancement, such as perceptual learning (e.g., Rodriguez, Blant & Hall, 2008; for review, Hall, 1991), or perceptual pop-out (e.g., Johnston & Hawley, 1994). Given that in the preexposure stage of many of these paradigms the organism does not have information about the forthcoming test stage, any one of these effects can be generated, depending on the conditions of the test stage. In general, the different outcomes from stimulus preexposure can be accounted for in three ways: (1) from separate effects that are subserved by multiple mechanisms which are activated in parallel during the preexposure session, such as stimulus property and stimulus relationship encoding, with the products of these operations differentially retrieved by different test conditions; (2) from a single mechanism operating during the preexposure stage that generates one outcome whose weight is determined by the particular test condition; (3) from some combination of 1 and 2. Importantly, the latter does not presume that the CS representation is constant across preexposure trials. Indeed, since LI, habituation and perceptual learning effects are modulated by the number of stimulus preexposures, some aspects of CS representation, such as property encoding and relationship encoding, also must be changing over the course of preexposures. Indeed, the fact that a small number of CS preexposures may enhance subsequent test performance, while a larger number produces LI (e.g., Kiernan & Westbrook, 1993; Prados, 2000), particularly when
540
R. E. Lubow and Ina Weiner
the test stage involves a discrimination between the PE and NPE stimuli (for review of relevant conditioned taste aversion literature, see Lubow, 2009), suggests that different processes are engaged serially, and that the final outcome in the test stage is, in part, determined by which processes had been activated in preexposure and to what extent they have been completed (Lubow, this volume). Notably, generating each of the phenomena noted above involves multiple stages, and changes of context across these stages have important consequences on the outcomes. Early theories of learning largely ignored the role of context in learning, primarily as a result of accepting models built on animal data generated from discrete stimuli that were manipulated by the experimenter; other stimuli available to the organism were often overlooked, if not by the rat, certainly by the theorist. The Rescorla–Wagner model, with its extensive use of context as an explanatory variable (e.g., accounting for the US preexposure effect by appealing to blocking by the training context), helped to point the way to a new approach, and many of the chapters in the present volume give context a central role in theorizing about LI (e.g., De la Casa & Pineno; Escobar & Miller; Honey, Iordanova & Good; Lubow).
Behavioral theories of latent inhibition As can be seen from reading the present volume, there would seem to be almost as many theories of LI as there are chapters in the book. However, the apparent diversity obscures an even greater number of commonalities. This should come as no surprise since current thinking has developed from an amalgamation of two theoretical stances once thought to be incompatible. One position, acquisition failure (A-theory), usually is associated with stimulus-specific attentional decrements in the preexposure stage (Lubow, Weiner & Schnur, 1981; Mackintosh, 1975; Pearce & Hall, 1980; Wagner, 1981). The A-theories attribute LI to some mechanism operative during the preexposure period that affects the subsequent associability of the stimulus, usually as a result of a loss of salience. The second position, retrieval/expression/competition (R-theory), typically ignores attentional constructs, and maintains that LI is a result of retrieval/competition mechanisms (e.g., Bouton, 1993; Miller, Kasprow & Schachtman, 1986; Weiner, 1990). By this account, after stimulus preexposure, the acquisition of the new association to the old stimulus proceeds normally; however, when the animal encounters the target stimulus in the conditioning or test stage, associations from the preexposure stage and the CS–US association from the acquisition stage compete for behavioral expression or retrieval. In short, A- and R- theories of LI are both based on the premise that something is learned during the preexposure period, but they differ in terms of the manner in which such learning impacts subsequent CR generation. Evidence for the involvement of attentional processes in the preexposure stage, already alluded to in previous sections, includes results from experiments
Issues in latent inhibition research and theory: an overview
541
with masking, masking task load, visual search LI, and, in general, the effects of preexposure variables on associative performance in the subsequent acquisition stage. R-theories of LI are supported by two groups of experiments, those that have varied the time between the conditioning and test stages (retention interval) and those that have varied context conditions across stages. The R-theories are supported by data that LI is disrupted by long intervals between conditioning and test stages and by change of context between preexposure and conditioning. In regard to retention interval, if LI is found after a short but not a long delay between acquisition and test stages, and if this difference is due to better conditioning performance by the preexposed–delay group than the preexposed–no-delay group, then this is taken as evidence that, with the short delay, the CS–US association was present at normal strength (i.e., same as that of the NPE short-delay group) but not manifest, and that something occurred during the longer delay that allowed the normally encoded CS–US association to be retrieved. Indeed, such effects have been demonstrated when the retention interval is spent in the same context as that of the other experimental stages. However, when the interval is spent in a different context, the LI effect is potentiated (De la Casa & Pinen˜o, this volume; for review, see Lubow & De la Casa, 2005). Although the delay-induced attenuation of LI is certainly more compatible with R-theory than A-theory, both theories have yet to contend with the super-LI effect. Second, as has been extensively documented, LI is context-specific, i.e., LI is disrupted or abolished when the context in which stimulus preexposure was conducted is changed in the subsequent acquisition/test stage (for a review of early studies, see Lubow, 1989, pp. 74–81; also Bouton, Nelson & Rosas, 1999; Escobar, Arcediano, Platt & Miller, 2004). The relevance of these findings for R-theory lies in the fact that the preexposure stage contains no information regarding what will happen in the subsequent stages. Therefore, in stage-1, stimulus-preexposed animals in the same-context condition must acquire the same associations as stimulus-preexposed animals in the changed-context condition. As a result, any difference in stage-2 performance can only be attributed to a post-preexposure difference in the ability to utilize what was acquired in the first stage. Thus, when a change in context disrupts LI, the most obvious interpretation is that the test-stage context served as a cue for the retrieval of the association (e.g., CS–0) that was acquired in the first stage. This interpretation receives additional support from studies that show that for context and stimulus preexposure to be effective in producing LI, the two must be preexposed conjointly (e.g., Channell & Hall, 1981, Exp. 3; Lovibond et al., 1984; Zalstein-Orda & Lubow, 1995). The disruption of LI by a change of context from the preexposure to the test stage does not preclude an attentional account of what occurs during the stimulus preexposure stage, nor the possibility that stimulus preexposure reduces the associability of that stimulus. However, accepting the latter position requires acknowledging that associability is, in part, a function of the degree of
542
R. E. Lubow and Ina Weiner
context-integrity across stages. In other words, current associability of a previously inconsequential stimulus is dependent on the relationship of that stimulus to the post-preexposure context, presumably because of effects on stimulus salience. Such a position, anticipated by Lubow, Rifkin and Alek (1976), circumvents the apparent conflict between A- and R-theories by adding a perceptual component to what had been exclusively an associationist framework, a move that can be seen in recent theories of LI (e.g., Cassaday & Moran, this volume; Lubow, this volume; Schmajuk, this volume; see Weiner, this volume, for a different interpretation of context effects). The chapters in this volume center on the basic questions raised by the older A- and R-theories. What are the processes that the target CS undergoes in the preexposure stage and how do they affect subsequent conditioning, or, what associations are formed in preexposure and how do these associations affect/ interact with subsequent CS–US associations? The same questions also apply to context. As will be seen from the brief summary below, various chapters have provided different answers to the questions regarding processing in the preexposure and conditioning stages.
Preexposure stage According to Hall and Rodriguez (this volume), repeated presentation of a stimulus generates inhibitory learning that reduces alpha. More specifically, the preexposed LI stimuli are not initially neutral but evoke the expectation or representation of some consequence (CS–event association). However, since no consequence occurs, a CS– no-event association is formed that acts to oppose the preliminary CS–event association. Alpha decreases because acquisition of inhibitory strength increasingly counteracts the preexisting excitation. Westbrook and Bouton (this volume) also assume that the presentation of a novel stimulus in preexposure generates a prediction that it will have a potential consequence. However, the discrepancy (prediction error) between this prediction and the actual outcome (nothing) drives the formation of a stimulus–nothing association rather than inhibitory conditioning, and this association results in a decline of attention to the to-be-CS. Reduced attention (inattentional response) to the preexposed CS, resulting from a classically conditioned CS–nothing association, is also assumed by Lubow et al. (1981) and Weiner (this volume). Mismatch between predicted and observed events plays a key role also in Schmajuk’s (this volume) account of stimulus preexposure. For him, attention to the CS is proportional to the novelty in the environment (determined by the mismatch between all predicted and actual events). CS preexposure reduces novelty, thereby reducing attention to the CS. Honey, Iordanova and Good (this volume) and Escobar and Miller (this volume) account for LI by assigning a particularly important role to context
Issues in latent inhibition research and theory: an overview
543
(also see Lubow, this volume). Following Wagner’s SOP model (1981), they all assert that, during preexposure, an association is formed between the experimental context and the target stimulus. Conditioning stage For Hall and Rodriguez (this volume), impaired conditioning to the preexposed stimulus is a consequence of reduced alpha. When the conditioning context is different from the preexposure context, retrieval of the inhibitory association formed during the preexposure phase is prevented, thereby restoring alpha. Thus, for Hall and Rodriguez, context change affects LI by way of an effect on alpha rather than by direct associative interference. Westbrook and Bouton (this volume) also argue that conditioned performance is impaired due to decline in attention. However, they add that the decline is rapidly removed since the signaling of the US (large prediction error) restores attention. Once formed, the CS–US association competes for expression in performance with the previously acquired stimulus–no-event association, thereby providing a source for a performance deficit. A change in context affects both attention (conditioning) and competition (performance): it restores attention to the preexposed stimulus and prevents the retrieval of the competing no-event associate, which otherwise would compete with its US association for expression in performance. According to Honey, Iordanova and Good (this volume), presenting the preexposed stimulus in a context with which it was associated in preexposure activates a relatively weak memorial representation of the stimulus. The latter results in the impairment of both learning (forming an association between the preexposed stimulus and a US) and performance (reduces the expression of conditioned responding that the stimulus would have elicited). Although a change of context “will render ineffective the association formed during the preexposure phase”, the effects from context change come by way of a restoration of alpha rather than by associative interference. In Schmajuk’s (this volume) scheme, at the time of conditioning, CS novelty is lower in the stimulus-preexposed group (only the US unexpected) compared to the nonpreexposed group (both CS and US unexpected). The reduced attention to the preexposed CS gives rise to three effects: impaired CS–US association, weak activation of this association (generation of a weak CR) and slow retrieval of the CS–US association. With a change of context, novelty is increased, thereby restoring attention to the CS. Weiner (this volume) also accepts a basic competition model. The inattentional response to the CS, acquired in the preexposure stage, competes for behavioral expression with the conditioned response acquired as a result of the CS–US pairing. If the CS–nothing association is strong and the CS–US association is moderate, CR generation is inhibited. For Escobar and Miller (this volume) performance in the test stage is a result of competition among multiple associations formed during training. During the conditioning stage, both the CS and the context become associated with the US.
544
R. E. Lubow and Ina Weiner
At test, the presentation of the CS activates two representations of the US, one directly (CS–US association) and one indirectly (through the CS–preexposure context and preexposure context–US associations), and a comparison of these two US representations determines conditioned responding (which increases with strength of activation of the US representation through the direct path and decreases with strength of activation of the US through the indirect path). Assuming that the CS–context association is very strong, whereas the context–US association is moderate, CS preexposure would lead to greater activation of the US representation through the indirect path, resulting in reduced performance.
Reconciling and expanding A- and R-theories of LI As indicated in the above brief survey, it is generally accepted that a viable theory of LI has to encompass some aspects of A- and R-theories; such a theory must acknowledge that the LI effect has multiple sources from the processing of information in the preexposure stage as well as in the conditioning and test stages. Such a general framework maintains the distinction between learning and performance, and allows for construing these factors in associative and/or attentional–perceptual terms. Moreover, these considerations need to be applied to the NPE group as well as the PE group (see section “The PE group and the NPE group both contribute to the LI effect”). Importantly, all of the explanations of LI in this volume, with the exception of Escobar and Miller, stress that stimulus preexposure reduces associability, attention, or alpha, although there is disagreement as to the nature of the associations formed in preexposure that lead to such decrements. In fact, in some chapters, attentional processing is emphasized not only during preexposure, but also during conditioning (e.g., Hall & Rodriguez; Schmajuk; Westbrook & Bouton; also see Lubow’s perceptual account of LI). In order to reconcile A and R approaches to LI, several authors have opted for a combination of attentional and performance deficits; preexposure retards the acquisition of CS–US association which, once formed, competes with CS–no-US for expression (e.g., Honey et al., this volume; Schmajuk, this volume; Westbrook & Bouton, this volume). Finally, several accounts of LI assign a significant role to stimulus salience (e.g., Cassaday & Moran, this volume; De la Casa & Pinen˜o, this volume; Lubow, this volume), a theme rooted in the alpha construct of the Rescorla–Wagner model, which also was prominent in the first theories of LI (e.g., Pearce & Hall, 1980; Lubow, Weiner & Schnur, 1981; Mackintosh, 1975). However, more complex views of salience, not necessarily synonymous with alpha, are also presented here. For example, Cassaday and Moran (this volume) distinguish between intrinsic and acquired stimulus salience, the former referring to the direct and immediate effects of physical salience and the latter resulting from learning that a stimulus signals nothing. Lubow (this volume) shifts the focus from associative to
Issues in latent inhibition research and theory: an overview
545
perceptual processing underlying changes in salience, identifying the loss of stimulus salience with a merging of the preexposed stimulus and the context in which it is appears. Overall, as already noted, unlike the earliest LI theories, which emphasized processing in the stimulus preexposure stage, or the later ones that relied on retrieval/competition factors in the test stage, the current chapters adopt a multistage model of LI effects. The distinction between models that advocate reduced to-be-CS salience that affects subsequent associability and those that rely on retrieval and competition has been replaced by theories that incorporate aspects of both models. In summary, the scaffold on which to build a comprehensive theory of LI requires a 2 2 format: stimulus preexposure and nonpreexposure conditions preexposure stage and acquisition/test stage. With that in mind, it then becomes necessary to identify the specific processes that are engaged in each of the four cells. Although not formally presented within this framework, several of the chapters in the present volume, as described above, have taken steps in that direction.
Neuropsychology of latent inhibition: latent inhibition as an animal model of schizophrenia The first investigations of the neural basis of LI were concerned with the effects of hippocampectomy on LI (Ackil, Mellgren, Halgren & Frommer, 1969; Kaye & Pearce, 1987a, 1987b; McFarland, Kostas & Drew, 1978; Solomon & Moore, 1975). These investigations were prompted by the view that the hippocampus is essential for excluding from attention stimuli that have no significant outcomes (e.g., Douglas & Pribram 1966; Gray, Feldon, Rawlins et al., 1978; Kimble, 1969; Moore, 1979; Moore & Stickney, 1980; Schmajuk & Moore, 1985, 1988). Interestingly, several early proponents of the attentional models suggested that the tuning-out of nonreinforced stimuli was accomplished by a joint action of the hippocampus and the brainstem reticular arousal system (Douglas, 1972; Kimble, 1969; Moore, 1979), which was later shown to contain the dopaminergic system. Although the investigation of the neural substrates of LI as an attentional and associative phenomenon has continued in its own right (see Gould, this volume), most of this research has been guided by the proposition that abnormal LI can model deficient attentional control in schizophrenia. As detailed in many chapters in this volume (e.g., Cassaday & Moran; Gould; Louilot, Jeanblanc, Peterschmitt & Meyer; Schnur & Hoffman; Weiner; Weiner & Arad), and as expected of a phenomenon whose aberrations are relevant to schizophrenia, dopamine plays a key role in LI, with drug-induced increased and decreased dopaminergic transmission abolishing and augmenting LI, respectively, in both rodents and humans. Together, pharmacology and lesion studies reveal that the neural site of these effects is the nucleus accumbens (NAC), where selective lesions to its two subregions, the shell and
546
R. E. Lubow and Ina Weiner
the core, abolish and augment LI, respectively. In addition, perturbations of the entorhinal cortex (EC) that are restricted to the preexposure stage abolish LI. Furthermore, as shown by Louilot et al. (this volume), these perturbations modulate NAC DA. There is less consistency regarding the hippocampus and the basolateral amygdala, whose perturbations can both abolish and enhance LI, suggesting a modulatory effect. Since both structures can control NAC DA (Louilot et al., this volume), this is not surprising. The fact that manipulations of NMDA transmission within the EC, the BLA, and the hippocampus (Gould, this volume) affect LI is consistent with the increasing body of pharmacological evidence that glutamatergic transmission at the NMDA receptor plays a central role in LI, and that its alterations can both abolish and enhance this phenomenon (Weiner & Arad, this volume). Finally, orbitofrontal cortex lesions enhance LI, whereas the medial prefrontal cortex is not involved. The neural substrates mediating aberrant LI are consistent with the dysfunctional fronto-limbic and mesolimbic DA circuitries as well as the dopaminergic and glutamatergic dysregulations implicated in the pathophysiology of schizophrenia (Javitt & Zukin, 1991; Toda & Abi-Dargham, 2007; Weiner & Joel, 2002). Based on the above data and the behavioral theories of LI, two neuropsychological theories of LI have been proposed that map psychological constructs believed to underlie LI onto neural circuitry that is thought to regulate this phenomenon, one originating with Gray, J. A., et al. (1991) and elaborated in several subsequent papers with Schmajuk and other colleagues (the Schmajuk–Lam–Gray model, SLG; e.g., Schmajuk, Buhusi & Gray, 1998; Schmajuk, Cox & Gray, 2001), and the other with Weiner (1990), and expanded in subsequent papers (the switching model, SM; Weiner & Feldon, 1997; Weiner, 2003). Since SLG combines aspects of Gray et al.’s (1991) and Weiner and Feldon’s (1997) circuits (Schmajuk et al., 2001), the two models share many features. In both models, the core psychological construct is the degree of perceived mismatch between predicted and actual events (denoted as Novelty in SLG) which is represented by the activity of NAC core, with low mismatch resulting in LI and high mismatch resulting in its disruption. NAC core activity can be modulated by the shell via its control of VTA DA input to the core. The information to the shell is transmitted from the EC. However, the specific mechanisms and psychological variables mapped onto the circuitry are different. In the SLG model, associations formed during preexposure in the hippocampus and EC reduce future conditioning by decreasing activity corresponding to Novelty throughout the components of the circuitry. In the SM, conditioning is determined by the sum of competing inputs to the core from the EC (signaling the CS–nothing information via the shell and VTA), the hippocampus (signaling information on context) and the amygdala (signaling information about reinforcement). Another difference between the two models is that for SLG, NAC output controls the formation of the CS–US association (in the amygdala), whereas in the SM, NAC activity only determines whether the CS–no-event association or the
Issues in latent inhibition research and theory: an overview
547
CS–US association is expressed, while the formation of CS–US associations is controlled by other (unspecified) circuitry. Finally, for the SLG model, DA controls learning and thus participates in the formation of associations in the preexposure and conditioning stages, whereas in SM, DA controls only performance (expression) and therefore participates only in conditioning. In summary, there is a consensus as to the brain regions regulating LI, but disagreement as to the nature of such regulation, with SLG being closer to A-theories (although Novelty also determines performance and retrieval in the model) and SM closer to an R-theory (although it uses the construct of CS associability; see Weiner, this volume). While additional pharmacological and lesion studies using intracerebral injections and reversible lesions are needed to delineate the interface between neural mechanisms and theoretical constructs, and to discriminate between A and R models, it should be emphasized that extant neurophysiological data already have clear implications for the one- vs. two-process (attentional or competition vs. some combination of the two) debate in LI. Thus, one of the major themes emerging from the pharmacological and reversible lesion studies is the importance of distinguishing between the processes occurring in preexposure and conditioning stages. Dopaminergic manipulations act selectively via conditioning; serotonergic manipulations affect LI via preexposure; cholinergic and glutamatergic manipulations can act via both stages. Notably, depending on the stage of action, the behavioral effects of the drugs are different. Indeed, they may exert synergistic or opposing effects in the two stages (Weiner & Arad, this volume). The same applies to reversible lesions although here the information is still limited. The fact that drugs and reversible lesions act selectively in preexposure or conditioning and, indeed, affect LI differentially via the two stages clearly demonstrates that LI involves separate processes in the two stages. Moreover, if a drug or lesion acts via the conditioning stage in the PE group without affecting conditioning in the NPE group, then its action in the PE group cannot be due to effects on CS–US acquisition; rather, there must be an additional process in the conditioning stage that is affected, a process that involves the expression of the CS–US association. In this regard, Louilot et al. (this volume) provide an illuminating demonstration of the role of the EC and its modulation of NAC DA in LI that supports a two-process competition model. Interestingly, investigations of the cellular/molecular underpinnings of LI reveal that CS–US and CS–nothing associations may involve different cell signaling molecules, and that there may be differences in how the two associations are stored (Gould, this volume), again supporting two-factor models of LI. In general, because LI involves several distinct processes (see above), the effects of pharmacological and physiological manipulations can be complex. The effects of the NMDA antagonist MK801 provide an excellent example of the complex action drugs can have on the acquisition and expression of LI. Low doses produce persistent LI if given in conditioning and are ineffective if given in preexposure. Conversely, high doses disrupt LI if given only in conditioning as well as if given only in preexposure,
548
R. E. Lubow and Ina Weiner
the former being due to impaired conditioning in the NPE group (consistent with the well-known propensity of NMDAR antagonists to disrupt learning), and the latter being due to better conditioning in the stimulus-preexposed group (Gould, this volume). These data indicate that NMDA receptor antagonists affect processes mediating the acquisition as well as the expression of LI as a function of dose and stage of administration. While such a level of complexity is detrimental to the use of LI as an animal model where high throughput is often the main requirement, this same complexity allows for an accurate discrimination between effects of different drugs and lesions on attention and learning. Equally important, most lesion, pharmacological, and genetic manipulations of LI produce stronger rather than weaker LI. These findings suggest that the function of most regions and neurotransmitter systems in the intact brain that play a role in LI is to modulate its expression, i.e., to enable either the association formed in preexposure or that formed in conditioning to gain control over behavior. Indeed, there is evidence that the behavioral expression of these two associations is mediated via distinct brain circuitries (Weiner, this volume). The latter again supports a two-factor view of LI. In addition, the above findings suggest that the function of many regions and neurotransmitter systems in the intact brain is not to enable but rather to restrict or prevent the expression of LI. It should be noted, however, that if different regions or subregions of the same structure compete in regulating LI expression, enabling both expression and disruption of LI, then a complex and often inconsistent pattern of effects can be expected following disturbances of such structures. As pointed out by Weiner (2003), in the absence of mechanisms that restrict the expression of LI, the effects of inconsequential stimulus preexposure would be extremely robust and maladaptive, generalizing across contexts and counteracting the expression of subsequent conditioning to the stimulus. Indeed, the organism would suffer from attentional inflexibility akin to that associated with negative and cognitive symptoms of schizophrenia. Remarkably, LI begins developmentally as a very persistent phenomenon: in juvenile rats, LI is resistant to all of the manipulations known to disrupt LI in adult rats (including weak preexposure, strong conditioning, context shift, and amphetamine). It is only as the brain matures that LI becomes flexible and adaptable, i.e., a disruptable phenomenon (Zuckerman, Rimmerman & Weiner, 2003). Thus, LI is unique in that it evolves from a robust but inflexible phenomenon to one that is highly responsive to situational demands, a development that is apparently based on post-pubertal maturation of brain systems that regulate LI expression, most likely including NAC and its limbic sources of input (see above). Relatedly, a maturation-dependent dysfunction in these same systems might be responsible for the post-pubertal emergence of disrupted LI observed following various perinatal manipulations (e.g. Zuckerman, Rehavi, Nachman & Weiner, 2003; Weiner & Arad, this volume), and possibly for disrupted LI observed in schizophrenia. This would be consistent with the neurodevelopmental view of schizophrenia, which posits
Issues in latent inhibition research and theory: an overview
549
a disruption of late-maturing, highly evolved temporal and cortical functions that fully manifest themselves only in adult life (Weinberger, 1987), as well as with the mesolimbic DA and temporal-limbic pathology implicated in schizophrenia (Grace, 1991, 2000; Weiner & Joel, 2002). While the above studies have been at least partly concerned with theoretical aspects of LI, it should be acknowledged that most animal neuropsychological LI studies are atheoretical. In place of theoretical considerations, they focus on evaluating models or model organisms relevant to the etiology, pathophysiology and genetics of schizophrenia, and on the discovery and development of novel therapeutics for schizophrenia. Both applications use the LI phenomenon as an operational measure of attentional control/gating without explicit adherence to a particular theory of LI. In the first case, LI aberration, particularly its disruption, is taken to support the relevance of the etiological/pathophysiological or neurodevelopmental model to schizophrenia. In the second application, the ability of drugs to “normalize” LI after it has been experimentally affected by other drugs or manipulations (e.g., neurodevelopmental) is taken to predict the clinical utility of antipsychotic/ antischizophrenia agents. We will not survey the evidence that provides the LI model with face, construct and predictive validity. As can be seen in almost every chapter of the present volume, such evidence has accumulated for more than two decades. However we will address two of the most often voiced criticisms, partly related, of this model. One is the failure of studies with schizophrenia patients to consistently find disrupted LI. As Swerdlow (this volume) points out, investigations of LI in schizophrenia patients have yielded “mixed results”, concluding: “It seemed that, without parsing the study sample in various and often idiosyncratic ways, it was difficult to see LI deficits in a general sample of SZ patients. Furthermore, the parsing strategies required to demonstrate deficits in one subgroup or another differed from study to study”. Although the overall picture clearly points towards disrupted LI in the acute, or early stages of the schizophrenia (e.g., Baruch et al., 1988; Gray et al., 1992, 1995; Rascle et al., 2001; Vaitl et al., 2002; Kumari & Ettinger, this volume), but presence of LI in chronic schizophrenia (Baruch et al., 1988; Gray et al., 1992, 1995; Lubow, Weiner, Schlossberg & Baruch, 1987; Swerdlow et al., 2005), some studies have reported intact LI in acute schizophrenic patients (Swerdlow et al., 1996; Williams et al., 1998). The inconsistency in the patient studies contrasts with the highly reliable results obtained with high-schizotypal individuals and healthy people given amphetamine (for reviews, Kumari & Ettinger, this volume; Lubow, 2005), indicating that the LI deficit in schizophrenia is more complex and/or more difficult to detect than in these other groups. Although a discussion of these problems is beyond the scope of this chapter, the important point is that the absence of attenuated LI in schizophrenia need not indicate the absence of LI abnormality. As discussed above and detailed elsewhere in this volume (Weiner; Weiner & Arad), deviant LI is expressed not only as a loss but also as abnormal persistence. In fact, animal data suggest that
550
R. E. Lubow and Ina Weiner
dysfunctions of all the neurotransmitter systems (dopaminergic, glutamatergic, cholinergic) and many brain regions (prefrontal cortex, amygdala and hippocampus, as well as NAC core) considered to be critically involved in schizophrenia can produce persistent LI. Thus, disease processes associated with schizophrenia can impact the LI circuitry in such a way as to bias it towards weaker or stronger LI. As already noted, such latent persistent LI coupled with the inadequacy of group statistics and the scarcity of reliable within-subject procedures increases the likelihood that “LI deficits in a general sample of SZ patients” will tend to be inconsistent and mixed. However, LI can be made a more reliable and powerful tool for investigating dysfunctional attentional control in schizophrenia by changing the emphasis from “attenuated LI” to one of “abnormal LI” and by developing within-subject procedures that can detect both disrupted and persistent LI. The second criticism is that the validity of the LI model is undermined because the attentional processes that are involved in human and animal LI are not identical. This position is based on the claim that human but not animal LI requires a masking task (see Le Pelley & Schmidt-Hansen, this volume). We have provided evidence in this chapter, as well as elsewhere, that questions this assumption (e.g., Lubow, 2005; Lubow & Gewirtz, 1995). However, even if animal and human LI are not identical, this would not compromise the validity of the animal model. The popularity of LI as an experimental paradigm for understanding schizophrenia comes from its conceptual linkage to clinical observations that schizophrenia patients are unable to filter or “gate” irrelevant stimuli (e.g., Anscombe, 1987; Bleuler, 1911; Hajos, 2006; Kapur, 2003; Kraepelin, 1919; McGhie & Chapman, 1961; Nuechterlein & Dawson 1984; Venables, 1984). The LI model of schizophrenia is intended to deal with that specific attentional deficit, namely the inability to process adaptively and to ignore irrelevant stimuli. As summarized above, leading theoretical treatments of LI overwhelmingly embrace the attentional construct of animal LI and, more specifically, reduced attention or associability. As such, it is generally accepted that disruptions of LI in animals reflect impairments of normal attentional processes, and therefore that they can be used to model attentional deficits in schizophrenia. To the degree that LI in humans is the same as that in other animals, it provides strong support for the validity of animal models of schizophrenia. However, such models should not necessarily be judged by their direct application to the clinic. Rather, they should be evaluated in terms of their success in providing a coherent picture of cognitive/ behavioral deficits, their underlying brain substrates, and their sensitivity to relevant treatments. In the case of LI, the model demonstrates the credibility of the concept that dysfunction of various components of the limbic and the mesolimbic DA system, including one of a neurodevelopmental and genetic origins, can give rise to two behavioral/cognitive aberrations, attentional overswitching and attentional perseveration, or what Lubow (this volume) refers to as overly labile and overly inert attentional processes, which have been accepted as core characteristics of schizophrenia from the very inception of dementia praecox (Kraepelin, 1919).
Issues in latent inhibition research and theory: an overview
551
There is considerable agreement that animal models will remain an important tool in developing and testing hypotheses for the pathogenesis of schizophrenia (e.g., Lipska & Weinberger, 2000). As such, animal LI provides a reliable and robust measure of attention that easily translates across species, has strong predictive validity in drug discovery and development, and can be mapped onto a relatively well-identified neural circuitry. It will, no doubt, continue to be used to test and validate these hypotheses and to generate new hypotheses regarding neural, genetic and cellular mechanisms as well as novel therapeutic strategies.
Epilogue The initial impetus for the book arose when we realized that the fiftieth anniversary of the first LI experiment was approaching, and how far we have come since that time. LI, a phenomenon that reflects the manner in which an organism processes irrelevant stimuli, has never stopped being of interest to the research community, and, in fact, its appeal is growing. As can be appreciated from the present volume, LI has received sophisticated theoretical and behavioral analyses and it has captured the attention of researchers across a broad spectrum of disciplines, from behavior and cognition to neurophysiology and neurogenetics. Having developed out of the tradition of classical learning theory with rats, LI has intersected with the pathophysiology of schizophrenia, where it is exceptionally well suited for what everyone is now searching – translational research. We hope that the present volume will provide scientists with a comprehensive survey of current LI research and theory, across the entire spectrum of fields, thereby strengthening the particularist approach to research as well as fostering an interdisciplinary methodology. The first 50 years may well serve as a prologue, however lengthy, for a more complete story that has yet to be told.
References Ackil, J., Mellgren, R. L., Halgren, C., & Frommer, S. P. (1969). Effects of CS preexposure on avoidance learning in rats with hippocampal lesions. Journal of Comparative and Physiological Psychology, 69, 739–747. Anscombe, F. (1987). The disorder of consciousness in schizophrenia. Schizophrenia Bulletin, 13, 241–260. Barrett, S. L., Bell, R., Watson, D., & King, D. J. (2004). Effects of amisulpride, risperidone and chlorpromazine on auditiory and visual latent inhibition, prepulse inhibition, executive function and eye movements in healthy volunteers. Journal of Psychopharmacology, 18, 156–172. Baruch, I., Hemsley, D. R., & Gray, J. A. (1988). Differential performance of acute and chronic schizophrenics in a latent inhibition task. Journal of Nervous and Mental Disease, 176, 698–600.
552
R. E. Lubow and Ina Weiner
Bleuler, E. (1911). Dementia Praecox or the Group of Schizophrenias. New York: International Universities Press. Bouton, M. E. (1993). Context, time, and memory retrieval in the interference paradigms of Pavlovian learning. Psychological Bulletin, 114, 80–99. Bouton, M. E., Nelson, J. B., & Rosas, J. M. (1999). Stimulus generalization, context change, and forgetting. Psychological Bulletin, 125, 171–186. Braunstein-Bercovitz, H., & Lubow, R. E. (1998). Are high schizotypal normal participants distractible or limited in attentional resources? A study of latent inhibition as a function of masking task load and schizotypy. Journal of Abnormal Psychology, 107, 659–670. Caramazza, A. (1986). On drawing inferences about the structure of normal cognitive systems from the analysis of patterns of impaired performance: the case for single-patient studies. Brain and Cognition, 5, 41–66. Channell, S., & Hall, G. (1981). Facilitation and retardation of discrimination learning after exposure to the stimuli. Journal of Experimental Psychology: Animal Behavior Processes, 7, 437–446. Cohen, E., Sereni, N., Kaplan, O., et al. (2004). The relation between latent inhibition and symptom-types in young schizophrenics. Behavioural Brain Research, 149, 113–122. Coutereau, E., Blundell, P. J., & Killcross, S. (2001). Basolateral amygdala lesions disrupt latent inhibition in rats. Brain Research Bulletin, 56, 49–53. De la Casa, G., & Lubow, R. E. (2001). Latent inhibition with a response time measure from a within-subject design: effects of number of preexposures, masking task, context change and delay. Neuropsychology, 15, 244–253. De la Casa, G., Marquez, R., & Lubow, R. E. (2009). Super-latent inhibition of conditioned taste preference with a long retention interval duration. Learning and Motivation, 40, 329–342. Douglas, R. J. (1972). Pavlovian conditioning and the brain. In R. A. Boakes & M. S. Halliday (Eds.), Inhibition and Learning. London: Academic Press, pp. 529–553. Douglas, R. J., & Pribram, K. H. (1966). Learning and limbic lesions. Neuropsychologia, 4, 197–219. Escobar, M., Arcediano, F., & Miller, R. R. (2003). Latent inhibition in human adults without masking. Learning, Memory and Cognition, 29, 1028–1040. Escobar, M., Arcediano, F., Platt, T. L., & Miller, R. R. (2004). Interference and time: a brief review and an integration. Reviews in the Neurosciences, 15, 415–438. Evans, L., Gray, N. S., & Snowden, R. J. (2007). A new continuous withinparticipant latent inhibition task: examining associations with schizotypy dimensions, smoking and gender. Biological Psychology, 74, 365–373. Gal, G., Barnea, Y., Biran, L., et al. (2009). Enhancement of latent inhibition in patients with chronic schizophrenia. Behavioural Brain Research, 197, 1–8. Gal, G., Mendlovic, S., Bloch, Y., et al. (2005). Learned irrelevance is disrupted in first-episode but not chronic schizophrenia patients. Behavioural Brain Research, 159, 267–275. Gibbons, H., Rammsayer, T. H., & Lubow, R. E. (2001). Latent inhibition depends on inhibitory attentional learning to the preexposed stimulus: evidence from visual search and rule-learning tasks. Learning and Motivation, 32, 457–476.
Issues in latent inhibition research and theory: an overview
553
Ginton, A., Urca, G., & Lubow, R. E. (1975). The effects of preexposure to a nonattended stimulus on subsequent learning: latent inhibition in adults. Bulletin of the Psychonomic Society, 5, 5–8. Grace, A. A. (1991). Phasic versus tonic dopamine release and the modulation of dopamine system responsivity: a hypothesis for the etiology of schizophrenia. Neuroscience, 41, 1–24. Grace, A. A. (2000). Gating of information flow within the limbic system and the pathophysiology of schizophrenia. Brain Research, 31, 330–341. Gray, J. A., Feldon, J., Rawlins, J. N. P., Hemsley, D. R., & Smith, A. D. (1991). The neuropsychology of schizophrenia. Behavioral Brain Science, 14, 1–84. Gray, J. A., Feldon, J., Rawlins, J. N. P., Owen, S., & McNaughton, N. (1978). The role of the septo-hippocampal system and its noradrenergic afferents in behavioral responses to nonreward. In K. Elliott & J. Whelan (Eds.), Functions of the Septo-hippocampal System. Ciba Foundation Symposium, No. 58. Amsterdam: Elsevier, pp. 275–300. Gray, N. S., Hemsley, D. R., & Gray, J. A. (1992). Abolition of latent inhibition in acute, but not chronic, schizophrenics. Neurology, Psychiatry, and Brain Research, 1, 83–89. Gray, N. S., Pilowsky, L. S., Gray, J. A., & Kerwin, R. W. (1995). Latent inhibition in drug-naive schizophrenics: relationship to duration of illness and dopamine D2 binding using SPET. Schizophrenia Research, 17, 95–107. Gray, N. S., & Snowden, R. J. (2004). The relevance of irrelevance to schizophrenia. Neuroscience and Biobehavioural Reviews, 29, 989–999. Hajos, M. (2006). Targeting information-processing deficit in schizophrenia: a novel approach to psychotherapeutic drug discovery. Trends in Pharmacological Sciences, 27, 391–398. Hall, G. (1991). Perceptual and Associative Learning. Oxford: Clarendon Press. Javitt, D. C., & Zukin, S. R. (1991). Recent advances in the phencyclidine model of schizophrenia. American Journal of Psychiatry, 148, 1301–1308. Johnston, W. A., & Hawley, K. J. (1994). Perceptual inhibition of expected inputs: The key that opens closed minds. Psychonomic Bulletin and Review, 1, 56–72. Kapur, S. (2003). Psychosis as a state of aberrant salience: a framework linking biology, phenomenology, and pharmacology in schizophrenia. American Journal of Psychiatry, 160, 13–23. Kaye, H., & Pearce, J. M. (1987a). Hippocampal lesions attenuate latent inhibition and the decline of the orienting response in rats. Quarterly Journal Experimental Psychology, 39B, 107–125. Kaye, H., & Pearce, J. M. (1987b). Hippocampal lesions attenuate latent inhibition of a CS and of a neutral stimulus. Psychobiology, 15, 293–299. Kiernan, M. J., & Westbrook, R. F. (1993). Effects of exposure to a to-be-shocked environment upon the rat’s freezing response: evidence for facilitation, latent inhibition, and perceptual learning. Quarterly Journal of Experimental Psychology, 46B, 271–288. Killcross, S., & Dickinson, A. (1996). Contextual control of latent inhibition by the reinforcer. Quarterly Journal of Experimental Psychology, 49B, 45–59. Kimble, D. P. (1969). Possible inhibitory functions of the hippocampus. Neuropsychologia, 7, 235–244. Kraepelin, E. (1919). Dementia Praecox and Paraphrenia. New York: Kreiger.
554
R. E. Lubow and Ina Weiner
Lipp, O. V., Siddle, D. A. T., & Vaitl, D. (1992). Latent inhibition in humans: single-cue conditioning revisited. Journal of Experimental Psychology: Animal Behavior Processes, 18, 115–125. Lipska, B. K., & Weinberger, D. R. (2000). To model a psychiatric disorder in animals: schizophrenia as a reality test. Neuropsychopharmacology, 23, 223–239. Lovibond, P. F., Preston, G. C., & Mackintosh, N. J. (1984). Context specificity of conditioning, extinction, and latent inhibition. Journal of Experimental Psychology: Animal Behavior Processes, 10, 360–375. Lubow, R. E. (1989). Latent Inhibition and Conditioned Attention Theory. New York: Cambridge University Press. Lubow, R. E. (1998). Latent inhibition and behavioral pathology: some effects of stimulus preexposure. In W. O’Donohue (Ed.), Learning and Behavior Therapy. Boston: Allyn and Bacon, pp. 107–121. Lubow, R. E. (2005). The construct validity of the animal-latent inhibition model of selective attention deficits in schizophrenia. Schizophrenia Bulletin, 31, 139–153. Lubow, R. E. (2009). Conditioned taste aversion and latent inhibition: a review. In S. Reilly & T. Schachtman (Eds.), Conditioned Taste Aversion: Behavioral and Neural Processes. Oxford: Oxford University Press, pp. 37–57. Lubow, R. E., & De la Casa, L. G. (2005). There is a time and a place for everything: bi-directional modulations of latent inhibition by time-induced context differentiation. Psychonomic Bulletin & Review, 12, 806–821. Lubow, R. E., & Gewirtz, J. C. (1995). Latent inhibition in humans: data, theory, and implications for schizophrenia. Psychological Bulletin, 117, 87–103. Lubow, R. E., & Kaplan, O. (1997). Visual search as a function of the type of prior experience with target and distractors. Journal of Experimental Psychology: Human Perception and Performance, 23, 14–24. Lubow, R. E., & Kaplan, O. (2005). The visual search analog of latent inhibition: Implications for theories of irrelevant stimulus processing in normal and schizophrenic groups. Psychonomic Bulletin & Review, 12, 224–243. Lubow, R. E., Rifkin, B., & Alek, M. (1976). The context effect: The relationship between stimulus preexposure and environmental preexposure determines subsequent learning. Journal of Experimental Psychology: Animal Behavior Processes, 2, 38–47. Lubow, R. E., Weiner, I., Schlossberg, A., & Baruch, I. (1987). Latent inhibition and schizophrenia. Bulletin of the Psychonomic Society, 25, 464–467. Lubow, R. E., Weiner, I., & Schnur, P. (1981). Conditioned attention theory. In G. H. Bower (Ed.), The Psychology of Learning and Motivation, vol. 15. New York: Academic Press, pp. 1–49. Mackintosh, N. J. (1975). A theory of attention: variations in the associability of stimuli with reinforcement. Psychological Review, 82, 276–298. McCartan, D., Bell, R., Green, J. F., et al. (2001). The differential effects of chlorpromazine and haloperidol on latent inhibition in healthy volunteers. Journal of Psychopharmacology, 15, 96–104. McFarland, D. J., Kostas, J., & Drew, W. G. (1978). Dorsal hippocampal lesions: effects of preconditioning CS exposure on flavor aversion. Behavioral Biology, 22, 398–404. McGhie, A., & Chapman, J. (1961). Disorders of attention and perception in early schizophrenia. British Journal of Medical Psychology, 34, 103–116.
Issues in latent inhibition research and theory: an overview
555
Miller, R. R., Kasprow, W. J., & Schachtman, T. R. (1986). Retrieval variability: sources and consequences. American Journal of Psychology, 99, 145–218. Moore, J. W. (1979). Brain processes and conditioning. In A. Dickinson & R. A. Boakes (Eds.), Mechanisms of Learning and Motivation: A Memorial Volume for Jerzy Konorski. Hillsdale, NJ: Erlbaum, pp. 111–142. Moore, J. W., & Stickney, K. J. (1980). Formation of attentional-associative networks in real time: role of the hippocampus and implications for conditioning. Physiological Psychology, 8, 207–217. Norman, C., & Cassaday, H. J. (2004). Disruption of latent inhibition to a contextual stimulus with systemic amphetamine. Neurobiology of Learning and Memory, 82, 61–64. Nuechterlein, K. H., & Dawson, M. E. (1984). Information processing and attentional functioning in the developmental course of the schizophrenic disorder. Schizophrenia Bulletin, 10, 160–203. Orosz, A. T., Feldon, J., Gal, G., Simon, A. E., & Cattapan-Ludewig, K. (2008). Deficient associative learning in drug-naı¨ ve first-episode schizophrenia: results obtained using a new visual within-subject learned irrelevance paradigm. Behavioural Brain Research, 193, 101–107. Prados, J. (2000). Effects of varying the amounts of preexposure to spatial cues on a subsequent navigational task. Quarterly Journal of Experimental Psychology, 53B, 139–148. Pearce, J. M., & Hall, G. (1980). A model for Pavlovian conditioning: variations in the effectiveness of conditioned but not unconditioned stimuli. Psychological Review, 87, 332–352. Rascle, C., Mazas, O., Vaiva, G., et al. (2001). Clinical features of latent inhibition in schizophrenia. Schizophrenia Research, 51, 149–161. Rodriguez, G., Blant, C. A. J., & Hall, G. (2008). The role of comparison in perceptual learning: effects of concurrent exposure to similar stimuli on the perceptual effectiveness of their unique features. Learning & Behavior, 36, 75–81. Salgado, J. V., Hetem, L. A., Vidal, M., et al. (2000). Reduction of latent inhibition by D-amphetamine in a conditioned suppression paradigm in humans. Behavioural Brain Research, 117, 61–67. Schmajuk, N. A., Buhusi, C. V., & Gray, J. A. (1998). Psychopharmacology of latent inhibition: a neural network approach. Behavioral Pharmacology, 9, 711–730. Schmajuk, N. A., Cox, L., & Gray, J. A. (2001). Nucleus accumbens, entorhinal cortex and latent inhibition: a neural network model. Behavioural Brain Research, 118, 123–141. Schmajuk, N. A., & Moore, J. W. (1985). Real-time attentional models for classical conditioning and the hippocampus. Physiological Psychology, 13, 278–290. Schmajuk, N. A., & Moore, J. W. (1988). The hippocampus and the classically conditioned nictitating membrane response: a real-time attentional-associative model. Psychobiology 16, 20–35. Serra, A. M., Jones, S. H., Toone, B., & Gray, J. A. (2001). Impaired associative learning in chronic schizophrenics and their first-degree relatives: a study of latent inhibition and the Kamin blocking effect. Schizophrenia Research, 48, 273–289.
556
R. E. Lubow and Ina Weiner
Shallice, T., Burgess, P. W., & Frith, C. D. (1991). Can the neurophysiological case-study approach be applied to schizophrenia? Psychological Medicine, 21, 661–673. Solomon, P., & Moore, J. W. (1975). Latent inhibition and stimulus generalization of the classically conditioned nictitating membrane response in rabbits (Oryctolagus cuniculus) following dorsal hippocampal ablation. Journal of Comparative and Physiological Psychology, 89, 1192–1203. Stevens, A., Peschk, I., & Schwarz, J. (2007). Implicit learning, executive function and hedonic activity in chronic polydrug abusers, currently abstinent polydrug abusers and controls. Addiction, 102, 937–946. Swerdlow, N. R., Braff, D. L., Hartston, H., Perry, W., & Geyer, M. A. (1996). Latent inhibition in schizophrenia. Schizophrenia Research, 20, 91–103. Swerdlow, N. R., Stephany, N., Wasserman, L. C., et al. (2003). Dopamine agonists disrupt visual latent inhibition in normal males using a within-subject design. Psychopharmacology, 169, 314–320. Swerdlow, N. R., Stephany, N., Wasserman, L. C., et al. (2005). Intact visual latent inhibition in schizophrenia patients in a within-subject paradigm. Schizophrenia Research, 72, 169–183. Toda, M., & Abi-Dargham, A. (2007). Dopamine hypothesis of schizophrenia: making sense of it all. Current Psychiatry Reports, 9, 329–336. Vaitl, D., Lipp, O. V., Bauer, U., et al. (2002). Latent inhibition and schizophrenia: Pavlovian conditioning of autonomic responses. Schizophrenia Research, 55, 147–158. Venables, P. H. (1984). Cerebral mechanisms, autonomic responsiveness and attention in schizophrenia. In W. D. Spaulding & J. K. Cole (Eds.), Theories of Schizophrenia and Psychosis. Lincoln, NE: University of Nebraska Press, pp. 47–91. Wagner, A. R. (1981). SOP: A model of automatic memory processing in animal behavior. In N. E. Spear & R. R. Miller (Eds.), Information Processing in Animals: Memory Mechanisms. Hillsdale, NJ: Lawrence Erlbaum, pp. 5–47. Weinberger, D. R. (1987). Implications of normal brain development for the pathogenesis of schizophrenia. Archives of General Psychiatry, 44, 660–669. Weiner, I. (1990). Neural substrates of latent inhibition: the switching model. Psychological Bulletin, 108, 442–461. Weiner, I. (2003). The “two-headed” latent inhibition model of schizophrenia: modeling positive and negative symptoms and their treatment. Psychopharmacology, 169, 257–297. Weiner, I., & Feldon, J. (1997). The switching model of latent inhibition: an update of neural substrates. Behavioral Brain Research, 88, 11–25. Weiner, I., & Joel, D. (2002). Dopamine in schizophrenia: dysfunctional information processing in basal ganglia-thalamocortical split circuits. In G. Di Chiara (Ed.), Handbook of Experimental Pharmacology: Dopamine in the CNS. Berlin: Springer-Verlag, pp. 417–472. Williams, J. H., Wellman, N. A., Geaney, D. P., et al. (1998). Reduced latent inhibition in people with schizophrenia: an effect of psychosis or of its treatment. British Journal of Psychiatry, 172, 243–249. Young, A. M. J., Kumari, V., Mehrotra, R., et al. (2004). Disruption of learned irrelevance in acute schizophrenia in a novel continuous within-subject paradigm suitable for fMRI. Behavioural Brain Research, 156, 277–288.
Issues in latent inhibition research and theory: an overview
557
Zalstein-Orda, N., & Lubow, R. E. (1995). Context control of negative transfer induced by preexposure to irrelevant stimuli: latent inhibition in humans. Learning and Motivation, 26, 11–28. Zuckerman, L., Rehavi, M., Nachman, R., & Weiner, I. (2003). Immune activation during pregnancy in rats leads to a post-pubertal emergence of disrupted latent inhibition, dopaminergic hyperfunction, and altered limbic morphology in the offspring: a novel neurodevelopmental model of schizophrenia. Neuropsychopharmacology, 28, 1778–1789. Zuckerman, L., Rimmerman, N., & Weiner, I. (2003). Latent inhibition in 35-day-old rats is not an “adult” latent inhibition: implications for neurodevelopmental models of schizophrenia. Psychopharmacology, 169, 298–307.
Index
acetylcholine, 260, 283–5, 294, 478 acquired distinctiveness, 114–17 activity state, A1, A2, 48–9, 66–7, 70–1, 166–8 addiction, 355, 477, 490 ALX-5407, 239, 291 AMPA, 488 amphetamine, 9, 259, 278–80, 287–8, 292, 294, 295, 296, 300, 320, 349–56, 378–9, 486, 549 amygdala, 35, 253–6, 257–8, 261, 263–6, 337, 385–7, 391–4, 546, 550 animal model of schizophrenia, 107, 203, 298–303, 373–5, 549 annelids, 212, 214 antipsychotic drugs (APDs), 156, 228–37, 259, 279, 285–90, 297, 301, 349–56, 373–4, 376, 420–7, 439, 486 anxiety, 192, 193–4, 429, 430, 458, 461–5, 512–13 apomorphine, 280, 293, 319 aquatic slug, 201, 212 arousal, 172, 178 arthropods, 208, 214 assassin bug, 208, 211, 214 associability, 44–8, 102, 116–26, 130, 206, 213, 343–8, 354 associative interference, 121, 130, 359 associative strength, 12, 44–8, 63–5, 117–22 attention, 44–8, 63–5, 138–55, 458–64, 481–3, 486–8, 504, 510–13 attention deficit hyperactivity disorder (ADHD), 11, 236 attentional memory (zCS), 138 auditory evoked response, 483 automatic processing, 101, 216, 504, 511 autoshaping, 203, 205
calcium calmodulin-dependent protein kinase II (CaMKII), 255 cat, 202 causal learning, 100 cell signaling molecules, 258 cerebellum, 254 clinical subgroups, 449, 452 comparator hypothesis, 52–4, 75–9 extended comparator hypothesis (ECH), 79–83 comparator stimuli, 52–4, 75–83 Conditioned Attention Theory (CAT), 8, 47, 64, 517 conditioned inhibition, 7, 345–8, 354, 356 Consensual Assessment Technique, 186 context, 8, 24, 25, 27, 41–4, 49, 51–2, 54–6, 65–71, 74, 119, 121, 130, 144, 150, 154, 164–74, 254, 257, 343, 347, 352, 488, 506–7, 508 context differentiation hypothesis, 54–7 context–time interaction, 54–6 controlled processing, 101, 216, 504, 511 crab, 208 crayfish, 208, 209, 214 Creative Achievement Questionnaire (CAQ), 186, 190 Creative Personality Scale, 185, 189, 190 creativity, 183–6, 188, 189–92, 458 CS elements, 66–8 cue competition, 52, 118 cyclic-AMP inhibitor, 262 D-cycloserine (DCS), 291, 297 delta rule, 68, 83 dishabituation, 164, 178 distractibility, 298 divergent thinking, 185–8 D,L-2-amino-5-phosphovalerate (APV), 255 dog, 205 dopamine (DA), 9, 10, 35, 256, 259, 280, 295, 296, 319–21, 328–35, 348–53, 374, 483–4, 487, 488 d-serine, 228, 291
backward blocking, 155 between-subject designs, 440, 450, 451, 453, 535 biomarker, 439 blocking, 142, 155, 345–56, 488 bromocriptine, 280, 485
558
559
Index electrodermal conditioning, 204 electroencephalogram (EEG), 11 emergent properties, 138 endophenotype, 439–40 entorhinal cortex, 257, 262, 263–6, 328–31, 384 event-related potential (ERP), 11, 450 exploratory behavior, 215 extinction, 5, 23–6, 50–2, 53, 56, 83, 153, 236, 539 extracellular regulated kinase (ERK), 255, 262–3 Eysenck Personality Questionnaire (EPQ), 429 fathead minnow, 207 field dependence/independence, 458, 513 fish, 203, 206–8 fruit fly (Drosophila), 201, 214 GABA (gamma-aminobutyric acid), 35, 354 garden snail (Helix aspersa), 212, 214 generalization decrement, 126 glutamate, 35, 256, 260, 488 glycine, 238–40, 290, 291–2, 296, 297 glycine transporter-1 (GlyT1), 280, 290, 292, 297 glycyldodecylamide (GDA), 291 goat, 2–5, 13 goldfish, 202 Gray, Jeffrey, 10 Grin1D481N, 291 habituation, 7, 132, 164–8, 171–4, 178, 343, 539 hexamethonium, 487 hippocampus, 156, 174–8, 254, 257, 349, 385, 391–4, 479 honey bee, 201, 202, 208, 209, 214 Hull, Clark, 2 inattention, 65, 505 inbred mice, 227 information processing, 225 bottom-up, 347, 358 encoding, 334, 505–7, 511 top-down, 347, 358 inhibitory learning, 118 intelligence quotient (IQ), 184, 189–95 invertebrates, 201–3, 208–14, 216–18 James, William, 1 ketamine, 282, 288 Kraeplin, Emil, 298 latent inhibition (LI) applications, 56–7, 532 as acquisition deficit, 8, 44–9, 63–9, 72, 85, 86, 171 as inhibitory learning, 121–8, 130–2
as performance/expression deficit, 27, 50–4, 69–75, 326 associative structure of, 77 conditional probability account of, 96 context-independent, 174, 178, 386, 394–6 disrupted/attenuated, 41, 54, 189–92, 229–36, 277–8, 297, 298, 354, 355, 373, 375, 381–3, 384, 385, 389, 391, 396–8, 440, 459, 462, 486, 488, 537 history of, 8–13 early history, 2–7 in humans, 7, 94–101, 104–7, 186–92, 202, 204, 420–41, 449–51, 532, 534 memory-based account of, 28 methodological issues, summary, 539 modeling domains of pathology in schizophrenia, 302 neurodevelopmental models, 297 neuropsychology of, summary, 545 parameters of preexposure, 142–4 persistent/enhanced, 44, 51, 75, 229, 277–8, 290, 297, 298, 301, 379, 382, 386–9, 391, 396–8, 427, 440, 485, 486, 487–8, 537 pharmacological models, 298–303 recovery from, 27, 70–7 reminder treatments, 71 renewal, 25, 50, 72, 74, 86, 87, 154 theoretical issues, summary, 539 theories of A-theories, 500–3, 540 hybrid, 121, 129, 131, 347 perceptual, 507–10 R-theories, 9, 500–3, 540 two-headed model of, 277, 388–91 visual search analog of, 509 latent learning, 4 learned irrelevance, 101–4, 107, 533, 539 lobeline, 487 long-term memory, 255 mammals, 202, 203–5 masking task, 95, 101–4, 460, 504, 505, 510–12, 517–19, 534 mecamylamine, 260, 482, 487, 488 medial prefrontal cortex, 254, 258, 387–8 memory-based account of conditioning and extinction, 26 mesencephalic dopaminergic neurons, 320 microdialysis, in vivo, 11, 320–1, 380 mismatch, 138, 174, 376, 389 MK801 (dizocilpine), 261, 283, 291, 294, 295, 301, 302, 481, 488 molluscs, 212 monkey, 205 motivational state, 30–1, 129, 148
560 muscimol, 257 mutant mouse 5-HT3, 238 AMPA, 235 Bax, 234 calcineurin, 233 coloboma, 236 D1 receptor, 237 D2 receptor, 237 Disrupted-In-Schizophrenia 1 (DISC1), 234 glycine transporter-1 (GlyT1), 237, 238 GrinD481N, 239 immunoglobulin-like (Ig-like), 233 M5 muscarinic, 237 metabotropic glutamate receptors (mGluR), 235 neuregulin (NRG-1), 233 NR1, 239 SNAP-25, 236 a5-GABAA, 236 NBQX (2,3-dihydroxy-6-nitro-7-sulfamoyl-benzo[f] quinoxaline-2,3-dione), 488 negative priming (NP), 193, 458, 463 negative transfer, 115, 120 nematodes, 213–14 NEO-FFM Five Factor Model of Personality, 189 neural network model, 138–52 neuroimaging, 11, 240 nicotine, 260, 477–8, 479, 480–4, 486–8 alpha4beta2, 479–80, 488 alpha7 nicotinic receptor, 238, 294–5, 489 NMDA (N-methyl-d-aspartic acid), 255–62, 280–3, 288, 291, 294, 302, 481, 488 novel pop-out (NPO), 509 novelty, 138–42, 152, 155 nucleus accumbens, 35, 254, 256–7, 263–6, 286, 320–8, 348–53, 360, 378–83, 483 obsessive-compulsive disorder (OCD), 11, 440, 458, 461 occasion setting, 75, 120, 502 omission of an expected event, 28 orbitofrontal cortex, 257–8, 387–8 orienting response (OR), 144, 164 overshadowing, 78–9, 118, 122–4, 142, 152, 155, 170, 345–56 Parkinson disease, 11, 458 passage of time, 27, 41–4, 51, 75, 149 Pearce–Hall model, 31, 46, 117–22 perceptual learning, 55, 144, 503–6, 508, 539 phobia, 56–7, 532 physostigmine, 293–4, 296 pigeon, 202, 203, 205, 215 postconditioning manipulations, 148
Index post-traumatic stress disorder (PTSD), 461 prediction error, 34 in extinction, 28–9 in latent inhibition, 28–9 prefrontal cortex, 254 prepulse inhibition, 482 priming, 48–9 Prison Break, 1 proactive interference, 73 protein kinase, 255 psychosis-proneness, 420, 459 Quantitative Trait Locus (QTL), 226 rabbit, 202, 203, 205 reacquisition, 25, 154 reinforcement specificity, 129 reinstatement, 25, 154 Rescorla–Wagner model, 45, 63–5, 540 response time versus correct response, 186–7, 192, 509, 536 retention interval, 8, 41–4, 47, 49, 51–2, 53–6, 71, 75–2, 208 retrieval, 24, 27, 42, 73, 75, 78, 120, 344, 359 retrieval deficit, 50–2 retrospective revaluation, 155 reward, 354, 359, 480 roundworm, 213 salience, 44–8, 64, 342, 343, 344, 348, 354, 486–8, 517–19 aberrant, 357 acquired, 342, 344–8 context-modulated, 509 incentive, 484, 488 intrinsic, 343–8, 355 modulation of, 68, 344–8, 356–7 salmon, 207 schizophrenia, 85–6, 107, 155, 279, 282, 287, 357–9, 374, 420–9, 430, 440, 452, 459–61, 464, 484–6, 513–19 schizotypy, 107–8, 187, 188, 189, 429–41, 457, 459–61, 464–5, 513–15, 549 Schmajuk–Lam–Gray (SLG) model, 138–42 scopolamine, 283–5, 289, 290, 292, 300, 301 selective attention model, Mackintosh, 64 sensory gating, 480 serotonin, 9, 259–60 shared vulnerability model, 184–92 sheep, 2–5 short-term memory (STM), 48, 138, 255, 506 short-term memory trace (tCS), 138 SIB-1553A, 481 smoking, 480–4 sometimes competing retrieval model (SOCR), 83–5
561
Index sometimes opponent processes model (SOP), 44, 48–9, 66, 70, 164–71 modified SOP (MSOP), 67 modulation of learning, 168 modulation of performance, 167 spontaneous recovery, 25, 29, 41, 50–1, 75 STA scale of the Schizotypy Questionnaire, 429 State-Trait-Anxiety Inventory (STAI), 462 stress, 193, 420, 430, 459, 461–4, 512–13 striatal subregions, 321–8 striatum, dorsal, 320–8 Stroop task, 459–60, 482 subiculum, ventral, 328–34 superconditioning, 65 super-latent inhibition, 42, 51, 54, 56, 72, 75, 149, 150, 514 switching, 359–60
switching model, 375–8 synaptic plasticity, 260, 264 tetrodotoxin (TTX), 328–35, 384 thalamus, 479 tobacco, 478 Tolman, Edward, 2 Tourette’s syndrome, 11, 458 trait anxiety, 440, 461, 462 type-A personality, 459 vertebrates, 202, 203, 205–8, 215, 217 voltammetry, in vivo, 11, 321 within-subject designs, 4, 23, 32, 441, 509, 535 working memory, 194, 236, 507
