PROGRESS I N BRAIN RESEARCH
ADVISORY B O A R D W. Bargmann
H. T. Chang E. De Robertis
J. C. Eccles J. D. French H. HydCn
J. Ariens Kappers S. A. Sarkisov J. P. SchadC
F. 0. Schmitt
Kiel Shanghai Buenos Aires Canberra Los Angeles
Goteborg Amsterdam Moscow Amsterdam Brookline (Mass.)
T. Tokizane
Tokyo
J. Z . Young
London
P R O G R E S S IN B R A I N R E S E A R C H VOLUME 27
STRUCTURE AND FUNCTION OF THE LIMBIC SYSTEM EDITED BY
W. ROSS A D E Y Brain Research Institute, University of California, Los Angeles ( U.S.A.) AND
T. TOKIZANE Institute of Brain Research, University of Tokyo, Tokyo (Japan)
ELSEVIER PUBLISHING COMPANY A M S T E R D A M / LONDON / N E W YORK
1967
tttttttt
335 J A N VAN GALENSTRAAT, P.O. BOX 211, A M S T E R D A M
A M E R I C A N ELSEVIER P U B L I S H I N G COMPANY, INC. 52 VANDERBILT AVENUE, NEW YORK, N.Y. 10017
ELSEVIER P U B L I S H I N G C O M P A N Y L I M I T E D R I P P L E S I D E C O M M E R C I A L ESTATE BARKING, ESSEX
This volume contains the Proceedings of a SYMPOSIUM O N THE STRUCTURE A N D FUNCTION O F THE LIMBIC SYSTEM
organized by the Brain Research Institute, University of Tokyo, and held in connection with the XXIZZrd Znternational Congress of Physiological Sciences at Hakone, Japan in September 1965
LIBRARY O F C O N G R E S S CATALOG C A R D N U M B E R 67-12719
W I T H 266 I L L U S T R A T I O N S A N D 30 TABLES
A L L R I G H T S RESERVED T H I S BOOK O R ANY P A R T T H E R E O F MAY N O T BE R E P R O D U C E D I N A N Y FORM, I N C L U D I N G PHOTOSTATIC O R M I C R O F I L M F O R M , W I T H O U T W R I T T E N P E R M I S S I O N FROM T H E P U B L I S H E R S
,
I,,
,
I
P R I N T E D I N TWE N E T H E R L A N D S
List of Contributors
W. R. ADEY,Departments of Anatomy and Physiology, and Brain Research, University of California, Los Angeles, Calif. (U.S.A.). E. SH. AIRAPETYANTS, The Pavlov Institute of Physiology, USSR Academy of Sciences and The University of Leningrad, Leningrad (U.S.S.R.). P. ANDERSEN, Laboratory of Neurophysiology, Institute of Anatomy, University of Oslo, Oslo (Norway). F. BERGMANN, Department of Pharmacology, Hebrew University, Hadassah Medical School, Jerusalem (Israel). V. M. BUCHER, Department of Physiology, University of Zurich, Zurich (Switzerland). M. CHAIMOVITZ, Department of Pharmacology, Hebrew University, Hadassah Medical School, Jerusalem (Israel). C. D. CLEMENTE, Department of Anatomy and Brain Research Institute, Univexsity of California, Los Angeles, Calif. (U.S.A.). A. COSTIN,Department of Pharmacology, Hebrew University, Hadassah Medical School, Jerusalem (Israel). M. R. COVIAN, School of Medicine, Department of Physiology, Ribeir2o Pr&to,S. P. (Brad). J. M. R. DELGADO, Department of Physiology, Yale University School of Medicine, New Haven, Conn. (U.S.A.). Department of Pharmacology, University of Michigan, Ann Arbor, E. F. DOMINO, Mich. (U.S.A.). A. T. DREN,Department of Pharmacology, University of Michigan, Ann Arbor, Mich. (U.S.A.). M. D. EGGER,Departments of Anatomy and Psychiatry, Yale University,'School of Medicine, New Haven, Conn. (U.S.A.). E. ENDROCZI, Department of Physiology, Medical School, PCcs (Hungary). J. P. FLY", Department of Anatomy and Psychiatry, Yale University School of Medicine, New Haven, Conn. (U.S.A.). S. S. Fox, Department of Psychology, and Mental Health Research Institute, University of Michigan, Ann Arbor, Mich. (U.S.A.). J. A. GERGEN,Department of Physiology, Bowman Gray School of Medicine, Winston-Salem, N.C. (U.S.A.). K. HIROSE,Shionogi Research Laboratory, Shionogi & Co., Ltd., Fukushima-ku, Osaka (Japan). R. D. HUGHDINGLE,Department of Psychology, and Mental Health Research Institute, University of Michigan, Ann Arbor, Mich. (U.S.A.).
R. W. HUNSPERGFJR, Department of Physiology, University of Zurich, Zurich (Switzerland). N. ITOIGAWA, Second Department of Physiology, Osaka University, Medical School, Osaka (Japan). M. KAWAKAMI,Second Department of Physiology, Yokohama University School of Medicine, Yokohama (Japan). R. =DO, Shionogi Research Laboratory, Shionogi & Co., Ltd., Fukushima-ku, Osaka (Japan). K. F. KILLAM, Department of Pharmacology, Stanford University School of Medicine, Palo Alto, Calif. (U.S.A.). E. KINGKILLAM,Department of Pharmacology, Stanford University School of Medicine, Palo Alto, Calif. (U.S.A.). F. KLINGBFJRG, Department of Clinical Neurophysiology, Neurological-Psychiatric Clinic, Karl-Marx University, Leipzig (German Democratic Republic). N. KOBAYASHJ, Department of Physiology, Faculty of Medicine, Kanazawa University, Kanazawa (Japan). J. C. LIEBESKINDE, Department of Physiology, and Mental Health Research Institute, University of Michigan, Ann Arbor, Mich. (U.S.A.). K. LISSAK, Department of Physiology, University Medical School, PCcs (Hungary).
T. WMO,Laboratory of Neurophysiology, Institute of Anatomy, University of Oslo,
Oslo (Norway). A. MATSUSHITA, Shionogi Research Laboratory, Shionogi & Co., Ltd., Fukushimaku, Osaka (Japan). K. MIYAMOTO, Second Department of Physiology, Osaka University Medical School, Osaka (Japan). F. NAKA,Department of Physiology, Faculty of Medicine, Kanazawa University, Kanazawa (Japan). H. NAKAO, Department of Neuropsychiatry, Kyushu University School of Medicine, Fukuoka (Japan). H. NIKI, Department of Psychology, College of General Education, University of Tokyo, Tokyo (Japan). J. H. O B m , Department of Physiology, and Mental Health Research Institute, University of Michigan, Ann Arbor, Mich. (U.S.A.). J. OLDS,Department of Psychology, The University of Michigan, Ann Arbor, Mich. (U.S.A.). T. ONO, Department of Physiology, Faculty of Medicine, Kanazawa University, Kanazawa (Japan). P. L. PARMEGGIANI, Istituto di Fisiologia umana dell'Universit8, Bologna (Italy). P. PASSOUANT, Laboratoire de Pathologie ExpCrimentale, Facultd de Mtdecine, Universitt de Montpellier, Montpellier (France). L. PICKENHAIN,Department of Clinical Neurophysiology, Neurological-Psychiatric Clinic, Karl-Marx University, Leipzig (German Democratic Republic). K. H. PRIBRAM, Stanford University, School of Medicine, Palo Alto, Calif. (U.S.A.).
C. PTERNITIS, Laboratoire de Pathologie Exptrimentale, Facultt de Mtdecine, Universit6 de Montpellier, Montpellier (France). K. SETO,Second Department of Physiology, Yokohama University School of Medicine, Yokohama (Japan). T. S. SOTNICHENKO, The Pavlov Institute of Physiology, USSR Academy of Sciences and the University of Leningrad, Leningrad (U.S.S.R.). M. B. STERMAN, The Sepulveda V. A. Hospital, Sepulveda, Calif. (U.S.A.). E. TERASAWA, Second Department of Physiology, Yokohama University School of Medicine, Yokohama (Japan). Y. YAMAGUCHI, Second Department of Physiology, Osaka University Medical School, Osaka (Japan). K-I. YAMAMOTO, Department of Neuropharmacology, Shionogi Research Laboratory, Osaka (Japan). N. YOSHII,Second Department of Physiology, Osaka University Medical School, Osaka (Japan).
Neuronal Mechanism of Feeding Y U T A K A OOMURA, HIROSHI OOYAMA, TETSURO YAMAMOTO, FUMIHIKO N A K A , N O B U Y A S U KOBAYASHI A N D TAKETOSHI O N 0 Department of Physiology, Faculty of Medicine, Kanazawa University, Kanazawa (Japan)
tt has been generally acknowledged that food intake is controlled by hypothalamic activity, through the observations of a decrease (Hetherington and Ranson, 1940) or increase (Hetherington and Ranson, 1942; Brobeck, 1946) in food intake due to a lesion within the hypothalamus. In 1951, by systematic experiments, Anand and Brobeck clarified the hypothalamic mechanism of feeding. Using rats and cats, they observed that the animals with bilateral lesions made by current passage through an extreme lateral part of the lateral hypothalamic area (LH) at the same rostrocaudal level as the ventromedial hypothalamic nucleus (HVM), showed aphagia (and adipsia) which finally ended in death, if gastric feeding was not applied, due to starvation in spite of the availability of food, and that bilateral lesions of the HVM produced hyperphagia and obesity supposed to be a release phenomenon from an inhibitory mechanism of the HVM. From their results, the LH (at the rostrocaudal plane of the HVM) was termed the feeding center, and the HVM the satiety center. These observations were supported later by Delgado and Anand's (1953) experiment on cats which produced an increase in food intake by electrical stimulation through chronically implanted electrodes in LH. It was demonstrated further that the animals actually ate food, even to the point of satiety, upon stimulation of the LH (Smith, 1956; Miller, 1960). Conversely, by electrical stimulation of the HVM, the cats stopped chewing in the middle of feeding activity and even dropped the food from their mouths (Oomura et al., 1967). Owing to extensive connections between the hypothalamus and various limbic structures, the dual hypothalamic function on feeding is significantly modulated and integrated by those structures (Stevenson, 1964). Many experiments' of lesions of the amygdala (AM) produced hyperphagia and obesity in cats (Koikegami et al., 1955; Green et al., 1957; Fuller et al., 1957; Wood, 1958; Morgane and Kosman, 1960). Fonberg and Delgado (1961) observed, using cats with implanted electrodes, an inhibition of food intake by electrical stimulation of the baso-lateral nucleus of the AM. Oomura et al. (1965, unpublished observation) also confirmed the same result by giving repetitive electrical stimulation to the lateral nucleus of the AM as in the instance of the HVM stimulation. Similar behavioral changes were also induced by septa1 stimulation (Fonberg and Delgado, 1961; Oomura, 1965, unpublished observation). Further, Morgane (1961a) showed that damage to the pallidofugal fiber system of the globus pallidus (GP) in the far lateral hypothalamic area caused consistent References p. 31-33
2
OOMURA
et al.
aphagia and adipsia, though animals lesioned in the pallido-hypothalamic tract immediately dorsal to the fornix columns developed a hyperphagia. Emphasis was, therefore, centered on the possible importance on feeding behavior of the two directed fiber trajectories to the LH and HVM respectively. The relationship between the LH and the mesencephalic tegmentum is emphasized neuroanatomically (Nauta, 1958) and functionally (Morgane, 1964). Hyperphagia was produced by lesion of the dorsolateral tegmentum (Sprague et al., 1961) and of periaqueductal gray matter beneath the superior colliculus (Skultety and Gary, 1962), but no effect was obtained in cats by electrical stimulation on the dorsomedial tegmentum (griseum centrale) during food intake (Oomura, 1965, unpublished observation). In this way, our explanations for food intake may become more precise but also more complex. Analytical studies by single unit recording within the feeding and satiety centers are still relatively few (Sawa et al., 1959; Barraclough and Cross, 1963; Tsubokawa and Sutin, 1963; Oomura et al., 1964; Anand et al., 1964). Oomura et al. (1964) observed the following reciprocal relations between the HVM and LH, when the simultaneous recording of the spontaneous unitary discharges (SUDs) were carried out from them. SUDs in one center were decreased in number by repetitive electrical stimulation of the other center where the SUDs were increased. Moreover, using a statistical treatment of the time series of SUDs, we proved a significant negative cross-correlation between both centers. Striking control and modifications on electrical activities of the HVM and LH by the influence of the limbic system such as the AM, septum, GP and mesencephalic tegmentum were also verified. From these experiments, the neuronal control mechanism of feeding will become clearer. METHODS
Forty-five cats, 2.54.0 kg in weight, were used under artificial respiration and ether anesthesia. To decrease bronchial secretions, atropin was sometimes administered intraperitoneally in doses of 25 mg/kg body weight. In the later experiments on the effect of stimulation of the septa1 nucleus and of the dorsomedial mesencephalic tegmentum, animals were immobilized with gallamine triethiodide (flaxedil) and all wound margins and pressure points were infiltrated with 1% procaine (reinfiltration every 3 h). Concentric bipolar electrodes (tip diameters about 0.1 mm) for stimulation were implanted unilaterally through small openings in the cranium into the following regions according to Jasper and Ajmone-Marsan’s stereotaxic atlas (1954) : HVM (A 11.5,L 1.5, H 4 . 5 ) , L H ( A 1l.O,L3,H-3.5)lateralandmedialnucleusoftheAM (A 11.0, L 10.5, H -6.0, and A 11.0, L 8.0, H -6.0, respectively), lateral and medial parts of the GP (A 14.0, L 7.5, H -1.0, andA 14.0,L6.0,H-2.0,respectively),septal nucelus (SEP) (A 16.0, L 1.0, H 3.0), and the dorsomedial part of the mesencephalic tegmentum (A 2.0, L 2.0, H --0.5), and the electrodes were fixed there with dental cement. For extracellular recordings of the action potentials of neurons of HVM and LH, recording microelectrodes of glass pipettes filled with 6 M NaCl (about 0.1 p, tip diameter; DC resistance 40-60 MQ) or tungsten microelectrodes (about 1 p, tip
N E U R O N A L M E C H A N I S M OF F E E D I N G
3
diameter) coated with vinyl lacquer, were inserted simultaneously into the HVM (A 11.5, L 1.3, H -5.5) and LH (A 11.5, L 3.0, H -3.5) using motor-driven micromanipulators. In some experiments, an additional recording electrode was introduced into the lateral nucleus of the AM (A 11.5, L 10.5, H -6.0). The intervals of time between successive unitary discharges in each sample taken on films were measured under an enlarger and punched in paper tapes, which were then fed for statistical calculation into a general purpose electroniccomputer, NEAC 2230. When the microelectrodes had been inserted into the desired positions, and stable neurons firing with SUDs were detected, lack of neuronal responses to visual, auditory, and other skin stimuli was first verified by standard procedure. Secondly, changes in the activity of the HVM or LH on stimulation of other parts of the brain were recorded. For electrical stimulation, square pulse currents of 0.1 msec duration, 1-50 A amplitude lasting 3-5 sec were applied. c/sec, and 10-6 to At the end of each experiment, the brain, with the recording and stimulating electrodes in position, was perfused by neutral formol-saline solution (10 %). Thus the tips of the electrodes were fixed at the site of the last recorded unit. Electrode sites were verified histologically. Each neuron used as data for this paper was located within the HVM, LH or other desired region. RESULTS
( A ) Relationship between the HVM and LH activity
( 1 ) Spontanegus unitary discharges (SUDs) from the HVM and LH and the depth OJ anesthesia Depth of anesthesia. Simultaneous recording of SUDs from the HVM and LH longer than 40 sec was so difficult, though the recording electrodes were placed at the desired postions, that only 13 of the experiments could be treated statistically. Our predictable rough identification of HVM neurons during the experiment was derived from the observation of an increase and decrease in SUDs frequency upon repetitive electrical stimulation of 50 c/sec for a few seconds applied to the AM and LH respectively. SUDs were, in general, much influenced by the depth of ether anesthesia (Oomura et al., 1964). When the cat was in a half arousal stage (we called this lignt anesthesia), showing responses or moving the ears spontaneously to various sensory stimuli, the frequency (number of impulses/sec) of SUDs was 8-20 c/sec in HVM, while it was 2-6 clsec in LH. Whereas, under the deeply anesthetized condition, showing no response to stimuli and with their muscle tonus relaxed, it decreased to 0-3 c/sec in HVM and increased to 7-20 in LH. It has been believed that though the neuronal activity of the hypothalamus is affected inhibitorily by pentobarbital (Brooks, 1959; Stuart et al., 1964) and chloralose (Stuart et al., 1964), it is not, or only slightly, affected by ether except at the neocortex. In these experiments, however, SUD in the hypothalamus was undoubtedly affected by ether. Fig. 1 shows 16 instances of the mean frequencies of the simultaneous SUDs recordings from the HVM (Y, ordinate) and LH (X, abscissa) References p. 31-33
4
OOMURA
-
et al.
15L
> v
fI
-H
510-
C
f
5-
L
Fig. 1. Mean frequency (number of impulses/sec) of the spontaneous unitary discharges (SUDs) recorded simultaneouslyfrom the HVM and LH under deep as well as light anesthesia. Summarized from 16 records. Ordinate, mean frequency in HVM or);abscissa, that in LH 0. The straight line drawn by Y = - 1.1 X 18 with a correlationcoefficient r,, = -0.78. In mean frequency,increase in the HVM and decrease in the LH indicates light anesthesia (Oomura er d.,1967).
+
under various anesthetic stages, where the electrode tips were proved histologically to be in the right places. When the HVM neurons increased in frequency, those in LH decreased, and vice versa, showing the relation expressed by the equation, Y = -1.1 X 18, with a correlation coefficient, rxy = -0.78 (P< 0.01). Throughout the whole time course of change in depth of anesthesia produced by altering the amount of ether inhalation, continuous simultaneous SUDs recording from both centers was extremely difficult ,so that only three experiments were successful. At the intqmediate stage from deep to light anesthesia, the frequency was increased in HVM and decreased in LH, then became nearly the same. In another case, the frequenciesof HVM and LH under light anesthesia, 10.1 & 1.1 (N = 44)and 4.4 f 2.8 (N = 44)respectively, became 10.0 f 2.2 (N = 6 2 ) and 9.5 3.4l (N = 60) respectively during the intermediate stage. Discharge pattern. First, the discharge patterns in HVM and LH were determined from the histograms of frequency and interspike intervals of SUDs recorded simultaneously from both. Sampling periods for which these histograms were based were taken from the sequential ordering of neuronal activity, which was verified statistically as the stationary process by the F-test mentioned in the next section. The close accord of the observed and theoretical distributions was estimated by the Chi-square test (P < 0.05). Under deep anesthesia, the frequency distribution agreed with the results which Oomura et al. (1964) have already given, e.g. as illustrated in Nos. 1, 2, 5 and 12 of Table I, whenever the mean frequency of the HVM was as low as 2-3 c/sec, differences between mean values of frequency and standard deviations were small, and the pattern corresponded with the Poisson distribution, while the mean frequency of the LH was as high as 10-20 c/sec, with large differences between mean values and standard deviations, and it corresponded with the Gaussian distribution. In regard to the pattern of the histogram of interspike intervals, as also shown in Nos. 1, 2, 5 and 12 of Table I and Fig. 2. the distribution of HVM, mean intervals 25&900 msec fit an
+
TABLE I
a ! %
2
E
e :: i bJ
DIS T RIB UT IO N PATTERN, T H E M E A N FREQUENCY ( N U MBER OF IMP U LS ES /S EC ) A N D MEA N I N T E R S P I K E I N T E R V A L O F S P O N T A N E O U S UNITARY D I S C H A R G E S (SUDS) RECORDED SIMULTANEOUSLY FROM H V M A N D L H (Oomura el al., 1967)
From the histogram of number of impulses/sec (with mean & standard deviation), the Poisson or Gaussian distribution is determined. From the histogram of the interspike intervals,the exponential (Exp.) or r (order 2, T2,)distribution is determined. Total number in parentheses. For determination of the pattern, X2-test was performed at P < 0.05, e.g. ~5~ > 4.6,5 degrees of freedom. Two and three units from the HVM were recorded simultaneously by the same electrode in Unit Nos. 1 and 5 respectively. In No. 1, 3.2 f 1.5 shows mean frequency measured without differentiation of the two spikes.
LH
HVM Unit No.
1
2 5
Number of impulseslsec Mean
Pattern
Mean
3.2 f 1.5 (45) 2.0 f 1.4 (71) 1.2 f 0.4 1(66) 2.5 & 1.6
Poisson X52 > 4.6 Poisson X2 > 4.8 Poisson Xz2 > 3.2 Poisson Xa2 > 4.2 Poisson Xs2 > 4.6 Poisson x 4 2 > 5.3
250 f 140 (199) 434 & 383 (118) 912 & 648 (52) 330 & 250 (30) 350 f 250 (137) 282 f 2 3 5 (303) 375 i-300 (166)
;:1
(77) 2.5 f 1.4 3.3 f 1.8
-
7 8
9 12
-
Interspike interval (msec)
15.8 f 2.1 (36) 15.1 f 3.1
(44)
14.7 i 2.2 (50) 2.2 & 1.3 (46)
-
not Poisson Gauss
3.3 Poisson X32 > 5.1 Xl2>
65 5 47.5 (310) 70 rt 15 (630)
Pattern
Exp. X72
> 9.6
Exp. X2 > 4.1 Exp. X32 > 3.8 Exp. x32> 1.1 Exp. X32 > 7.0 Exp. Xe2 > 4.0 Exp. X72 > 12.3
XS2
> 9.2
not Fa
Number of impulseslsec Mean
19.1 f 3.2 (45)
Pattern
Gauss x 1 2
> 0.1
Interspike interval (msec) Mean
47 f 1 (825)
Pattern
z
m
C
w
0
z
k
r2
> 11
x72
r m 3 ct X
k
10.8 & 2.3 (77) 11.5 & 1.3 (92)
Gauss x22
> 1.1
Gauss Xi2 > 2.4
110 63.7 (177)
172 X? > 13
5 m 3
0 .rt
w
B
m
ElZ 0
6.3 & 2.7 (36) 2.1 f 2.8 (44)
Poisson x 4 2> 1 Poisson x 2 2>3
19.8 f 3.0 (46)
Gauss
406 f 220 (40) 343 70 (95)
Exp. x 5 2
> 10.1
Exp.
X32 >
7.1
Xa2 > 3.2 v,
6
OOMURA
et al.
exponential curve, whereas that of LH fits a curve of a drawn from the following equation (1): n = N/ 2 -
"Xi
+1
I' distribution of order 2 ( r ~ )
a2xe-ax dx
Xi
where a = =;x, mean interval; N, total number. X
Under light anesthesia, as already mentioned, the SUD frequency increased in the HVM to 10-15 c/sec and decreased in the LH to 3-7 c/sec. Consequently, completely reversed patterns of SUDs could be obtained. As illustrated in Nos. 7, 8 and 9 of Table I, the HVM frequency was no longer the Poisson distribution but a Gaussian one, the Poisson distribution being in LH. Accordingly, the distribution pattern of the interspike interval in the HVM was a TZwith mean intervals of 60-70 msec, as shown in Fig. 2, whereas in the LH it was exponential with mean intervals of 120-550 msec. From these results, it is proved that when the SUD frequency decreased greatly either in the HVM or LH, the neuronal discharges arose from a perfect Poisson process, whereas after an increase in activity level they appeared more or less in a regular manner. The results also indicate an intimate reciprocal relationship between the HVM and LH activities, with a highly significant inverse relation of SUDs with a coefficient of - 0 . 7 8 . When recorded units in one center discharged according to a pattern of distribution, the other units in the same region or center would be expected to discharge in the same pattern, too. This postulate was verified by simultaneous recording of SUDs of two units in one center. As with the two examples shown in Nos. 1 and 5 of Table I, two series of SUDs having different amplitudes (three units in No. 5) were recorded simultaneously by the same microelectrode in HVM. The distribution patterns of frequency and of interspike interval were Poisson and exponential distributions respectively in both units. Their mean values, however, were different, which means that they were not working in perfect synchronism. Thus it is likely that various units in the same center, at least in the vicinity, act approximately uniformly. (2) Correlationfunction between the activity of the HVM and LH Simultaneous recording of SUDs for more than 30 sec in the HVM and LH, was treated for an estimation of the auto- and cross-correlation function. Before these statistical treatments, the stationary state of the SUDs had to be checked. Neither the mean nor the degree of variability of the activity (discharge frequency) under study was expected to change significantlythroughout the sampling perigd. For the stationary state test, an F-test was employed, the analysis of variance described by Werner and Mountcastle (1963), and the criterion of stationary state was that of no significance level at P < 0.05. (i) Correlation calculated from the frequency at relatively long z. The auto- and cross-correlation functions were evaluated from the record in which the stationary state was already verified by the following equation (2) :
NEURONAL MECHANISM OF FEEDING
7
7
\,
Deep anesthesia
HVM (40 rec) N=58
534 f 5N.O msec
Fig. 2. Histograms of interspike interval of SUDS in the HVM (left) and the L H (right), recorded simultaneously. Under deep anesthesia, mean interval and standard deviation (2 r t b ) in the HVM is 534 f 574.0 msec (total number N = 58). The distribvtion pattern fits the theoretical exponential curve drawn from the equation, n (N
= 441), and the distribution pattern
equation, n
=N
,.Xi+l
J Xi
a2x e-&" dx, (a
=N
4eX
X
(continuous curve). In LH, 86 f 44.7 msec
fits the theoretical r-function of order 2 ( r z ) , drawn from the
=
*
5). Under light anesthesia, the mean interval, 65 & 47.5 X
msec (N = 310) in the HVM, and the distribution pattern fits the r a type, while in the LH it was 406 f 222.0 mSec (N = 42) and fits the exponential curve. Continuous lines are theoretical curves. n
(yi + k -
i) (xi - x)
[XY(tk), cross -correlation coefficient from x to y series at the time lag. of ket. yi (y! + k), number of SUD in ith {(i 4- k)th} time bin ti (ti + k) of y series; Xi (Xi + k), References p. 31-33
00
TABLE I1 C R O S S - C O R R E ~ A T I O N SOF
SUDS BETWEEN HVM
AND
LH (Oomura et al., 1967)
Calculatedfrom equation(2). T = 1 sec except where specified as T = 0.5 sec. Total number in parentheses. Cxlxain No.1 : corss-correlations calculated from two SUDS(XI and xa) recorded simultaneously from the HVM (No. 1 in Table I). X = HVM; Y = LH. Unit no.
1
Anesthesia
Location of electrodes
D=P
Correct
-0.46 (42) t = 0.5
D=P
Correct
-0.44
D=P
Correct
D=P
Correct
D=P
Correct
D=P
Correct
-0.32 (36)
Light
Correct
-0.27 (49)
Light
Correct
0
I
2
3
4
5 see
-0.36
5YX1
(64)
-0.36
CXZY
(62)
- 0.32
5YXZ
(65) 2
CYX
3
5YX
(34) t = 0.5 4
-0.55
5YX
(17)
5
5X1Y 5X2Y 5XaY
6
SXY
No correlation (92) 0.21 (92)
-0.21
0.16 (185)
-0.18
t = 0.5 5YX
7
tXY
8
tXY
(90) (1 82) 0.19 (184)
-0.12 (182)
e0 *
b
P
% rn
$w
‘p
Y
9
CYX
10
CXY
-0.31 (41) t = 0.5
0.42 (37)
-0.32
Light
Correct
Light
Correct
(43)
11
5XY
0.28 (64)
0.26 (63)
Intermediate Correct
0.34 (61)
No correlation (46)
12
Deep
LH, increases in frequency by AM stimulation
Deep
LH, correct; HVM, 2 mm dorsal from HVM
0.25 (63) 0.29 (63)
Deep
Correct
0.23 (89)
Light
Correct
Deep
Correct
D ~ P
Correct
i ,
bJ
13
5XY 5YX
14
CYiYa
15
h y a
0.31 (89)
0.24 (84)
0.24 (84)
0.12 0.18 (260) (259) t = 0.5 0.38 (55)
t = 0.5
0.15 0.12 (254) (253)
10
OOMURA ef
al.
the same as above but for the x series; % (i), mean number of SUDs per unit time z calculated for the whole period; n, total number of ti. The criterion of a statistical significance level of the cross-correlation followed the central limit theorem (Freund, 1960), and the value of the correlation function over 1.96/dn is significant(the whole period used for the computation was nt). Table I1 summarizes the data obtained. Only significant values of the cross-correlation function during spontaneous activity between the HVM and LH are included. Negative cross-correlations at 1 to 3 sec were obtained in Nos. 1 , 2 , 3 , 4 , 5 and 6, except No. 5x1. For example, in No. 2, when the LH neuron discharged at more than average frequency for 1 t (0.5 sec), the HVM neuron discharged less than its average, sixfold z (3 sec) afterwards. In other words, low activity of the HVM neuron followed high activity of the LH neuron with a delay of 3 sec. In Nos. 1 and 5, two (XI and XZ)and three (XI, x2 and x3) unitary discharges with differept amplitudes were recorded from the same electrode in two HVMs respectively. In the former, the cross-correlations between one unit of HVM, XI and that of LH, and between the other unit of HVM, x2 and that of LH were also significantly negative at 2, and 4 and 1 sec respectively. In the latter, a significant correlation did not appear between one unit of HVM and LH, but significant correlations were between two other units of HVM and the same LH unit. In No. 6 (also in No. 5-&), positive cross-correlations preceded negative ones, i.e. during and after 1 sec of LH discharging more than the average, the HVM unit also discharged more than the average and then less than the average two sec later. Significant negative cross-correlations between the spontaneous activities in HVM and LH existed under light anesthesia as shown in Nos. 7, 8, 9 and 10 of Table 11. On the other hand, only positive cross-correlations were obtained in Nos. 11 and 13. As to the correlation in No. 1 1 , the same units as those of No. 10;were under an intermediate stage from a light anesthesia to a deep one due to an increase in ether inhalation. In No. 12, a cross-correlation function was insignificant, though the LH electrode was in the designed place in LH, probably due to an abnormal response of the LH neuron. That is, the SUDs were not decreased in number by repetitive stimulation applied to the AM, but augmented instead. From the evidence that the activity of the LH neurons was lowered by AM stimulation in more than 80 % of the recorded units, a typical example is given in Fig. 7, the LH neuron of:No. 12 may be supposed to be a special one not affected by the HVM. In No. 13, the HVM electrodewaslocated 2 mm dorsal from the upper margin of the HVM, while the LH electrode was within the intended place, and only positive cross-correlationwas obtained. The physiological function of the region of the former site in the hypothalamus is still unknown. As described before, and clear from Nos. 1 and 5 of Table I, in the discharge patterns of two units recorded simultaneously from one HVM, the distributions of the frequency were of the Poisson type, and of the interspike interval were exponential, showing that their activities were all Poisson processes though each value was different. The cross-correlation function between the activity of the two units, XI and XZ,cleaily indicated significant positive values (lower part in Table IT), while significant negative correlations appeared between XI and LH and between X2 and LH. Nos. 14 and 15 of Table I1 show similar positive correlations between two units acting spontaneously in
11
N E U R O N A L M E C H A N I S M OF F E E D I N G
LH. Consequently,one may suppose that in the same center, or at least in neurons in the very vicinity, the patterns of neuronal activitiestend to be the same, though asynchronous, and between the two centers, HVM and LH, they are reciprocal. As to the autocorrelation function calculated at t = 0.3 to 1 sec, no periodicity of this function appeared at any time, as it did on the SUD in the thalamic neurons (Werner and Mountcastle, 1963). (ii) Auto- and cross-correlation at short unit time. Serial dependencies of the discharge occurring over a time range shorter than the unit time lag, z of 0.3 to 1 sec, escaped detection by the above method. Therefore, the correlations at short unit time, z of 10 msec, were calculated. A c t o c o r r e I a t iD n N=90,
Deep a n c s l h e s l a
1=900
L 1 8 h t anesthesia
L H
0.10005 -
L H
iarf(il
Hvu
I l f ( ' )
015.010-
ODS-
-005-010-015-
-015-
d
Fig. 3. Autocorrelation (5 (t)) calculated at short unit time (t = 10 rnsec). Upper: under deep anesthesia (No. 2 in Table I and 11; N = 907, i.e. total period, Nt,taken for the calculation is 9.07 sec), (t) of the HVM unit is almost zero except at t = 0 (discharge pattern of the HVM is a Poisson distribution. In the LH, ((t) is negative until about 50 msec except at t = 0. The values of [(t) closely fit the 1 theoretical curve (dotted line) drawn from the equation [(t) = - kAAt (A =.: ,x-, mean discharge X
interval). Lower: under light anesthesia (No. 8 in Table I and II, N = 900), [(t) of the LH unit is almost zero (dischafge pattern shows a Poisson distribution). [(t) of the HVM unit is negative until about 40 msec (discharge pattern, r2 type), also fits the theoretical curve (Oomura el al., 1967). ReJerences p. 31-33
12
OOMURA
et al.
First, Nos. 2 and 8 of Table I1 as typical examples of deep and light anesthesia respectively, were analyzed. The autocorrelation functions of the HVM unit of No. 2 and LH unit of No. 8 were almost zero up to 100 msec, as shown in Fig. 3. Since their discharge patterns indicated the Poisson process, this result was natural. About the LH unit of No. 2 and HVM unit of No. 8, whose discharge patterns were more or less regular with interval distributions of I'z type, the autocorrelation functions were negative up to about 50 msec, except at t = 0, then approached zero, as shown in Fig. 3. When the distribution pattern of discharge intervals corresponded to the r2 type, the autocorrelation function was obtained theoretically, which is given in the equation (3) : CX (t) = -1e-41t (3) 1 il = -..-;x, mean SUD interval X
The significant level of the autocorrelation function is not yet determined, but as shown in Fig. 3, the results and the theoretical curves fit very well, which implies physiologically that once a neuronal unit discharged, a lowered state of excitability lasts for the following 50 msec due to some mechanism. Since the mean intervals of such SUDs were always more than 40 msec, this lowered excitabilityis not attributable to the post-tetanic depression (Eccles, 1964) or post-tetanic hyperpolarization (Eccles, 1953) but probably due to neuronal refractoriness (Lloyd, 1951), accompanied by positive after-potential or recurrent inhibition. The meanings of equation (3) will be discussed in more detail later. As to the autocorrelation function at the short t,no periodicity was found in the present experiment as it was in the thalamic neuron every 50 to 70 msec (Poggio and Viernstein, 1964). (iii) Correlation calculatedfrom discharge intervals. To calculate the auto- and crosscorrelation functions from the discharge intervals, the method described by Gerstein and Kiang (1960) was employed. When the total period of the record for analysis, T, and the total number of SUDs in HVM and LH, X and Y contained in T, is extremely big, a conclusion could be derived from plxu(t),an expression in their method. But in our experiments, the amount of T, X and Y were usually so small (T, 20 sec; 40 impulses for X or Y at minimum) that tpXy(t)had to be normalized and tested for its significance level. Therefore, ex= (t)was evaluated by the following equation (4):
xi, yi, impulse number in HVM and LH respectively during A i t . At AT X--, Y--, mean impulse number in HVM and LH respectively.
T T The statistically significant level of the cross-correlation was also tested by the central limit theorem (Freund, 1961).
P
TABLE I11 C O M P A R I S O N B E T W E E NT H E CROSS-CORRELATION COEFFICIENT (X, HVM ;Y, LH) C A L C U L A T E D FROM EQUATION EQUATION (4) (DIGITAL) (Oomura et al., 1967) t = 333
msec; N, 57. Case of No. 2 in Table I.
1.66
2
2.33
0.14 -0.07
0.21
0.09
0.07 -0.09
-0.29* 4 . 1 3
0.07 4 . 1 9
0.02 -0.07
0.17 -0.12
-0.29* 4 . 2 0 -0.03
(2)
( A N A L O GUE) A N D
Q
Time lag
0.33
Analogue -0.02 Digital
*
-0.09
Significant
0.66 -0.07
0.02
1
1.33
2.66
3
3.33
3.66
4
0.07 -0.25*
4.33
4.66
5 sec
0.19 -0.25*
0.12
-0.22* -0.03
-0.27*
0.19
m
0
e
m m
L
w
14
OOMURA
et al.
Table 111 shows the cross-correlation of No. 2 in Table calculated by this method at t = 333 msec. For comparison, the cross-correlation calculated by equation (2) is shown. The significant value of the cross-correlation appearing at 3, 4 and 4.66 sec by the latter, are also all significant in the result calculated by the equation (4). ( 3 ) Change in SUD provoked by electrical stimulation After stable spontaneous firing neurons had been detected, responses to repetitive electrical stimulations in HVM or LH were investigated. The results were coincident with those obtained by Oomura et al. \1964). When LH was stimulated, the SUDs in 30 units out of 44 of the HVM decreased or ceased, in 5 units increased and in the remaining 9 units did not change. With increased stimulus intensity, this decrease lasted not only during the stimulation but also for several seconds following it, and then the SUD returned to its original level. One typical example (50 c/sec, 3 V for 3 sec) is shown in Fig. 4. Conversely, HVM repetitive stimulation inhibited SUDs in 15 LH units out of 24 (Fig. 4). The remaining 3 LH units tended to be facilitated slightly by 15t
.
*
H.V.M
10
10
30
SIC
Fig. 4. Left two records:inhibition of SUD in the HVM (left) and LH (right) by repetitive stimulation (50 c/sec, 0.1 msec duration, 3 V, for several seconds) upon the LH and HVM respectively. Top to bottom and left to right. Stimulation artifacts are seen in the middle of each left column. Horizontal bar: 1 sec (one sweep). Right two curves, inhibitoryeffect in the HVM (top) and the LH (bottom) as seen in the records. Ordinates, impulses per sec; abscissae, time (sec). Arrows indicate stimulation periods. Right and left units are different (Oomura et al., 1967).
stronger stimulation, and 6 units did not respond at all to stimulation of any strength. The inhibitory effect did not always become clear with the lowering of stimulus frequency. It became apparent, however, when the cross-correlation function between stimulus and SUDs was evaluated by the method mentioned on p. 12. Stimulation of the LH even at 1 c/sec still affected the HVM activity. For example, as shown in Fig. 5, the significantly negative cross-correlation (t = 25 msec) which lasted more than 75 msec, was followed by significantly positive correlations at 125 and 275 msec. This negative value may indicate a long-lasting inhibition caused by hyperpolarization of
NEURONAL MECHANISM OF FEEDING
15
Cross-correlation 1 = 25nrec
N = 1380
L H IWSSlirn - H V M
0.15
SUO
2 -lH&largeqikc)
Jk ...................
?set
ft
........................
- ao
-0.05
1
Mean Freqirec Ilk0 2 6 t I x 6 5rO 3(S t )
1
Mean Freqirec
18f05(St)~37rOd(St)
Fig. 5. Upper: inhibition of HVM activity by LH stimulation. Inset records show a series of HVM SUDs, and 1 c/sec stimulation (0.1 rnsec, 3 V indicated as an artifact by a dot) was applied to the LH for 34.5 sec. To make clear the inhibitory effect of stimulation on the HVM SUD, cross-correlation C(t) between stimulation pulses and SUDSin the HVM are shown. Method described on p. 12. t = 25 msec, N = 1380. Significant negative cross-correlation (statistically significant level indicated by dotted lines) lasts more than 75 msec, then significant positive one follows at 125 and 275 msec, which may indicate a post-inhibitory excitation. Lower: facilitation of LH activity by the same LH stimulation. Two SUDS in the LH with different amplitudes (large, LH1 and small, LHz) recorded by one electrode, increase in discharge number by 10 c/sec LH stimulation (0.1 msec, 5 V, for 59 sec). In LH1, the SUD of 1.1 0.2 (mean frequency standard error) before stimulation increased to 6.5 f 0.3 during stimulation. In LH2,1.8 f 0.5 increased to 3.7 f 0.4 .T(t) between stimulation pulse and LHI unitary discharges indicate a significant positive value at 10 msec after each stimulation, which may be the latency from stimulation to response. This is followed by a significant negative value at 40 to 50 msec. This negativity may correspond to the post-excitatory inhibition. In LH2, a significant negative correlation lasting for 30 msec appears, then becomes positive at 40 msec and again becomes negative at 40-50 msec (Oomura et al., 1967).
*
the membrane followed by a post-inhibitory excitation or rebound. In contrast to the above, SUDs in LH increased up to about twice their original frequency upon LH repetitive stimulation. This facilitation lasted for a few seconds. If a mechanism of spreading depression had been responsible for the HVM inhibition by the LH stimulation, a decrease or even a cessation of SUDs shoud have appeared at first in the LH References p . 31-33
16
OOMURA
et al.
neuron which was in the vicinity of applied electrical stimulation, before giving any inhibition in the HVM neuron which was in a remote place. If, furthermore, the HVM inhibition had been caused secondarily by the increased activity in LH, it should have appeared with more delay than the latency of facilitated activity in LH which was more than 15 msec (Fig. 9). Thus these inhibitions in HVM, considering the long-lasting inhibition more than 75 msec together, imply some inhibitory synaptic mechanism. By much stronger stimulation of LH up to 10 V or so, an increase in its SUD frequency was followed by a decrease and then a complete disappearance. This could be explained by an electrotonic spread of the stimulation to the HVM, which facilitated the HVM activity,and the latter in its turn inhibited the LH. In 33 units of the SUDSof the LH, discharge frequencies increased notably by the LH stimulation in 19 experiments, decreased in 6, and in the remainder was not altered. Although SUDs in the LH were modulated by LH stimulation, statistical analysis gave more precise knowledge. The cross-correlation functions at 10 msec t between stimulation and SUDSare shown in Fig. 5 (lower). Two spontaneous unitary discharges with large and small amplitudes were recorded in one recording electrode during stimulation at 10 c/sec and 5 V applied through the stimulatingelectrode 0.5 mm apart from the recording electrode. The frequency of SUDs of the large unit, whose original value was 1.1 f 0.2 c/sec, increased up to 6.5 f0.3, while that of the small one increased from 1.8 f 0.5 to 3.7 f0.4, on stimulation. A significant positive cross-correlation appeared on the large unit from 10 to 20 msec after the stimulation, then the correlation became negative at 40 to 50 msec. On the small unit, a significant negative correlation lasting for 30 msec appeared at first, then became positive at 40 msec and again became negative at 50 msec. In other experiments, the latency of driven unitary discharges in the LH by a single volley applied to the same LH was also analyzed statistically: the curve of post-stimulus latency and probability of response showed an extremely large latency of approximately 15 msec, and this latency was not shortened by stimulation of multi volleys. Therefore, the significant positive cross-correlations at 10 to 20 msec in the large units may indicate the latency of the response; then the following negative correlations show the post-excitatory inhibition (or post-spike hyperpolarization). In the small unit, the significant positive correlation at 40 msec followed by the negative one from 50 msec, can be explained in the same way, though an elucidation of the first negative correlation is not easy and needs further study. (B) Relationship among the limbic system, HVM and L H ( I ) Functional interaction between the A M and hypothalamus To investigate the functional relationship between the HVM or LH and the AM, simultaneous recordings of SUDs from the LH and lateral nucleus of the AM (lat. principal nucleus) were first carried out with 8 successful results under flaxedil immobilization. In three instances in which clearly theoretical discharge patterns in both AM and LH were yielded, a significant negative cross-correlation between both SUDs of AM and LH resulted, but in the other 5 in which definite distribution patterns were not determined, significant cross-correlations alone did not result, but
17
N E U R O N A L M E C H A N I S M O F FEEDING
sometimes positive and negative correlations appeared mixed. Moreover, in the three experiments, the SUD frequencies were alike in both AM and LH, 10-15 c/sec. As shown in the middle and lower parts of Fig. 6, when the number of SUDs in the AM was low (about 10 c/sec) and high in the LH (about 20 c/sec) the frequencies and interspike intervals were Poisson and exponential distributions in the AM, and Gaussian and rz distributions in the LH respectively, exactly the same as with the HVM and LH in the deep ether anesthesia. A significant negative cross-correlation (t,500 msec) from the LH to AM at 3 sec was also found (upper, Fig. 6). By stimulation of the lateral nucleus of the AM, in general, an increase in SUD frequency resulted in the HVM and a decrease in the LH. One example of simultaneous recordings of SUDs in Cross-corielat ion AH-
LH
LY
P( I )
Interval histogram
n
-1H
T = 500 msec
LH
80
N=311
I =4 9 f 2 6 9 n s e c
i= got29.lmsec
0
80
40
P
E -
200
0
400 0
200 msec
100
Frequency hist ogram
An N= 37
n
i=9B+4.0
2
10
I
=36
'Ol 12 Impmmb. I sec
= 20.4 f 2 . 5
24
36
Fig. 6. Middle aad lower: discharge patterns of SUDS recorded simultaneously from the lateral nucleus of the AM and LH. Histogram of the interspike interval of the AM fits the theoretical exponential distribution (continuous line) with the mean interval of 90 f 29.1 msec (% f a), while that of LH to rs distribution with 49 f 26.6 msec. Histogram of the frequency of the AM fits a Poisson distribution with mean frequency of 9.8 f 4.0 c/sec, while that of the L H fits a Gaussian one with 20.4 f 2.5. Upper: cross-correlation of SUDs between the AM and LH calculated from the lower record, t = 500 msec; total period, 13 sec. Significant negative correlation at 3 sec, indicating a hardness of unitary discharging in the AM more than the average after 3 sec when LH unit discharged more than the average for 500 msec. References p. 31-33
18
OOMURA
et ul.
the HVM and LH is shown in Fig. 7A. In A, to make the effect of the AM stimulation clearer, the moving average of the SUD series, mean impulse numbers per 5 sec of the HVM or LH SUD were successively taken at every 1 sec interval. Responding to 10 c/sec stimulation at 10 V for 15 sec, the frequency of SUDs in the HVM increased up to more than twice its normal frequency in the fist few seconds and stayed at this level during the stimulation. It decreased along with the cessation of stimulation. In the LH, the frequency decreased to zero and even continued in this state for about 15 sec after the cessation of stimulation. To obtain the precise time course of the facilitation and inhibition on the HVM and LH activities due to the AM stimulation, the cross-correlation functions between stimulation and SUDs were calculated on a simultaneous recording of the HVM and LH units different from those in A. In Fig. 7B, by stimulation with 10 c/sec at 10 V for about 20 sec, significant positive correlations (z, 10 msec) at 40 and 130 msec were evident in the HVM, whereas a significant negative one (z, 25 msec) at 175 msec was followed by significant positive ones at 200 and 300 msec in the LH. In one example, the upper part of Fig. 7B, the latency of units in the HVM responding to more than 3 stimuli of the lateral nucleus of the AM at 100 c/sec was approximately 40 msec; on increasing stimulus number at 100 c/sec, the latency was not shortened but a multiple response appeared.A complete inhibition of the SUDs in the LH by the same stimulation lasted more than 100 msec as shown in the lower inset. Taking account of these values, the first positive cross-correlation at 40 msec in the HVM is attributable to the latency of a response from a pulse of the stimulation, and the second one at 130 msec also accounted for the latency, since the second pulse of the stimulation fell 100 msec later. In regard to the LH, the first negative cross-correlation at 175msec would be derived from the inhibitory mechanism, and the second positive correlation at 200 msee would be due to the post-inhibitory excitation, but hardly explicable on that at 300 msec. Furthermore, close functional interactions, as well as other evidence, were found between either the AM and HVM or the LH. First, the summation of synaptic potentials with about 14 msec latency could be recorded in the HVM by repetitive stimulation of the lateral nucleus of the AM, as
A
E f f e c t 0 1 A L s t i m . o n HVM I L H
set
NEURONAL MECHANISM OF FEEDING
-005
- 0.1 0
19
-
1
Fig. 7. (A) Increase in SUD frequency in the HVM (upper) and decrease in the LH (lower) by repetitive stimulation of the lateral nucleus of the AM (0.1 msec, 1-0 c/sec, 10 V, for 15 sec). Ordinates: the moving average of SUDs, i.e. mean impulse numbers/5 sec were successively taken at every 1-sec interval. (B) Upper records of insets in the right show responding unitary discharges in the HVM by volley stimulations of the lateral nucleus of the AM at 100 c/sec (0.1 msec, 10 V). The latency of units responding by more than 3 stimuli is about 40 msec (by increasing stimulus number, the latency measured from the onset of stimulation is not shortened but a multiple response appears). Upper left curve shows cross-correlation, [(t), between stimulation pulse (10 c/sec, 10 V, for 19.8 sec) upon the AM, Z, and unitary discharges in the HVM, 5 = 10 msec. Note significant positive correlations at 40 and 130 msec. These 40 and 130 msec are attributable to the latency of responses by lO-c/sec stimulation. Lower records of insets in the right show inhibitions of LH SUDS lasting more than 100 msec, by volley stimulation of AM (at 100 c/sec, 0.1 msec, 10 V). Lower left curve shows t(t) between 2 (10 c/sec, 10 V, for 19.5 sec) and SUDs in the LH, T = 25 msec. A significant negative correlation at 175 msec is followed by a positive one at 200 and 300 msec. Simultanems recording from the HVM and LH.
shown in Fig. 8, though the neuronanatomical connection between this nucleus and the HVM have not been confirmed, probably either through an unknown direct synaptic pathway or indirectly through the medial principal nucleus. Second, when conditioning single stimuli of the AM with a fixed intensity were followed by test stimuli of the LH with various intensities, the probability of LH unitary response was calculated. One example of this kind of experiment is shown in Fig. 9. At two times, 10 and 40 msec, the probability of LH response was lowered. PresumabIy the first lowering was due to a direct inhibition from the AM to the LH and the second to an inReferences p . 3 1 3 3
20
OOMURA
et al.
-
Fig. 8. Summating responses in the HVM by repetitive stimulation of the lateral nucleus of the AM, with a latency of about 14 msec. These responses are probably synaptic potentials produced in the HVM neuron. Vertical bar, 5 mV; horizontal one, 100 q e c .
direct pathway from the HVM whose activity was driven with a latency less than 40 msec by the conditioning stimulus. Twenty-four out of 40 HVM units were facilitated by the repetitive AM stimulation, though 5 other units were inhibited slightly; histological identification showed that 3 units were actually located in regions adjacent to the HVM, probably related to regions in the LH. Eleven units did not respond to any stimulation. In regard to LH units, 32 out of 55 units were inhibited by AM
I
t 1Y c o d .
1
I
30
1
I
60nstc
Interval O f t B t l Shock Fig. 9.Effect of conditioning stimulation of the lateral nucleus of the AM on LH-LH response. Right record in the inset shows a preceding AM conditioning stimulation (fixed intensity of 10 V, 0.1 msec) (first pulseat theleft)andaunitaryresponseoftheLH(thirdspikefrom theleft) by single test stimulus (0.1 msec) of LH (middle pulse, stimulus artifact). Left curve: probability of LH response (ordinate) against intervals of the conditioning and test stimulus. The probability is lowered at 10 and 40 msec. See text.
N E U R O N A L M E C H A N I S M OF F E E D I N G
21
stimulation, 11 units were facilitated to some extent and 6 units never responded. The remaining 6 units increased in frequency first, then decreased. (2) The efect of septa1 stimulation on the SUDs in the HVM and LH As already clearly demonstrated by Tsubokawa and Sutin (1963), and in the present experiments too, unitary dischargesin the HVM evoked by single or volley stimulation at 100 clsec of the central part of SEP under flaxedil immobilization appeared with latencies extending from 10 to 40 msec. Though the latencies were not definitely shortened by the increase in number of stimuli at 100 c/sec, the mean impulse numbers for 100 msec from the cessation of volley stimulation were proportionally increased by an increasing stimulus number at 100 c/sec with a positive significant correlation coefficient (in one instance, 0.86). Some neurons were fired by both AM and SEP stimulations. In contrast, with these effects of brief stimulation, on repetitive stimulation, say 50 c/sec for 10 sec, only three out of 11 HVM units increased in frequency and two units decreased after cessation of stimulation, though they were not altered during stimulation. No obvious change was seen in the remaining units. LH units were markedly inhibited by SEP stimulation. As shown in Fig. 10, in the upper records (A, B, C), a considerably long period of cessation of the LH SUDs
A
C
o . n l l Cross-corrrlation
-0.8
_____ ____-__ - -
1 ::lI S t P 4* l!ai US stiu. I I=Ill1
Fig. 10. Effect of SEP stimulation upon L H SUDs. Upper left (A, B, C ) : SUDS i0 the L H disappear for about 200 msec (B) and 350 msec (C), after single (A), 3 (B) and 4 (C) stimuli at 100 c/sec (0.1 msec, 7 V). After the inhibition, increases in number of SUDS result. As shown upper right, an increase in number of stimulations at 100 c/sec (7V, abscissa, X ) leads to a decrease in S U D numbers for 50 msec after stimulation (ordinate, Y,mean impulse number/50 msec), and this relationship is expressed by Y = - 0.242 (X - 2.5) 0.625, with a significant negative correlation coefficient of -0.825. Lower: cross-correlation between stimulus pulse of SEP stimulation (3 stimuli at 100 c/sec, 7 V, for 15 sec) and S U D s in the LH, t = 10 msec. Sigdicant negative correlation lasting more than 120 msec was followed by a significant positive one at 210-240 msec.
+
References p . 31-33
22
OOMURA
et al.
followed the stimulations of SEP, even after only a single stimulation, but was more evident after .volley stimulation, e.g. with three stimuli at 100 c/sec there was an inhibition lasting approximately 200 msec and with 4 stimuli more than 350 msec. In the right inset, the relationship between the inhibition and the volley stimuli are shown ; the mean impulse number of LH SUDs for 50 msec after stimulation was approximately inversely proportional to the stimulus number at 100 c/sec with a significant negative correlation coefficient, -0.83. To obtain a precise picture of the temporal pattern, the cross-correlation function (r, 10 msec) between SEP stimulation (three stimuli at 100 c/sec, 7 V for 15 sec) and LH SUDs was also calculated (lower, in Fig. 10). A significant negative cross-correlation lasting more than 120 msec was followed by a signscant positive one at 210-240 msec. The negative correlation is caused by the inhibition mentioned above. The latter positive one is due to post-inhibitory excitation which is proved in record c in Fig. 10. An increase in discharge frequency preceded by the inhibition is clearly seen at about 250 msec after the volley stimulation. Eleven out of 16 LHunits behaved asmentioned above on repetitive SEP stimulation, one unit increased in frequency and 4 were not altered. (3) The effect of lateral and medial parts of the GP on HVM and LH activities Repetitive stimuli of 10-50 c/sec for several seconds were applied to the lateral part of the GP. These effects on both the HVM and LH were found to be similar to the effects of stimulation of the lateral nucleus of the AM. As shown by the movement of the average curves (average frequency/sec>of the simultaneous recordings of SUDs in both the HVM and LH in Fig. 11, the SUDs of the HVM unit were increased in frequency not only during stimulation (10 c/sec, 4 V, for 7 sec), but also thereafter up to 2-3 times the original rate. This state lasted about 10 sec and then gradually returned to the original level. Six out of the 16 HVM units were affected similarly. The latency of the responding unit to single or multiple volleys at 100 c/sec stimulation was 30 to 60 msec as shown in the right upper records of Fig. 1 1. In regard to the LH units, an inhibition of activity was produced by stimulation, i.e. as shown by the moving average curve, the SUD decreased to almost zero in frequency during the stimulation and gradually went up beyond the original level. This increase was believed to be due to the post-inhibitory excitation. As shown in the upper and middle records in the right lower part of Fig. 11, the noticeable inhibitions lasting 50-100 msec produced by single volleys were followed by an increase in number of unitary discharges. As becomes clear in the records, the larger the stimulus intensity, the longer the inhibition. The lowest record shows a considerably longer period of inhibition, more than 250 msec on injury discharges from a single volley. Ten of the 17 LH units behaved almost identically. Three units increased in frequency, and 4 units were not affected by the stimulation. The effects of the medial part of the GP were in the opposite direction from those of the lateral one. In three HVM units out of 5, SUDS were decreased in frequency and gradually returned to the original level when the repetitive stimulation (50 c/sec for 5 sec, 10 V) of the medial part was switched on and off. Two units increased slight-
23
N E U R O N A L M E C H A N I S M OF F E E D I N G
HVM SUO n
L - 6 P 4 HVM L-BP
Train
s r i m . la% 4 v
1 rtlavll I t I W W
lmnr
Fig. 11. Effect of GP stimulation on SUDs in the HVM and LH. Right upper records show unitary discharges in the HVM responding to 2 stimuli of the lateral part of the G P at 100 c/sec with latencies of 30 to 60 msec. Stimulus intensity is different in all records, but around 7 V. Right, lower records: upper twoshow post-inhibitoryexcitationin theLHneuron by singlestimulations of the lateral part of the GP. The inhibitions last 50-100 msec. The lowest record shows a long lasting inhibition of injury discharges of a LH neuron more than 250 msec from a single stimulation of the lateral part of the GP. Left: upper and middle, HVM and LH SUDs increase and decrease in frequency respectively after repetitive stimulations of the lateral part of the GP (10 c/sec, 0.1 msec, 4 V, for 5 sec). Simultaneous recording from the HVM and LH. Lower, LH SUDSincrease in frequency during repetitive stimulation of the medial part of the GP (100 c/sec, 17 V, for 15 sec).
ly in frequency. On the LH SUDs, three of 11 units increased in frequency during the stimulation as shown in the lowest curve of the moving average in Fig. 11. One unit was inhibited, but 7 units were not affected by the repetitive stimulation. The latency of LH units responding to a single stimulation of the medial part of GP was 10-20 msec. ( 4 ) The effects of the mesencephalic tegmentum on HVM and LH activities Stimulation at 5-15 V, 1&50/sec of the dorsomedial part of the periaqueductal gray substance had different effects upon both HVM and LH units under flaxedil immobilization. In two out of the 10 HVM units, the moving average curve of HVM SUDS as shown in Fig. 12, decreased in frequency not only during stimulation for 3 sec but also for some time after it, then gradually returned to the original level in nearly 10 sec. But the frequency of 3 HVM units was increased and of 5 units was not altered by the stimulation. It is therefore difficult to reach a conclusion from the results. Fifteen out of the 33 LH units were considerably increased in frequency up to about five fold on stimulation of the periaqueductal gray substance for 3 sec. This effect usually lasted more than 10-15 sec after cessation of stimulation (Fig. References p . 31-33
24
OOMURA
et al.
12), except in one instance in which the increase in frequency caused by 10 clsec stimulation went back once to the original level on cessation of stimulation but was seen again 14 sec later and continued for about 5 sec. Seven units were inhibited and the other 11 units remained unaffected by the stimulation. Unitary discharges of the LH units responding to the single volley stimulation of the dorsomedial mesencephalic tegmentum were usually 10-15 msec in latency, but in some units, as shown in the lower records in Fig. 12, repetitive firings appeared which increased in frequency due to increased stimulus intensity though the latency was extremely long.
-L I
%-
'150-
100 50
-
LH
SUO
50
I
1
I
.
I
gT
Ch 1 5 V
I
I
Fig. 12. Effect of stimulation of dorsomedial part of the dorsomedial mesencephalic tegmentum CDMT) on SUDs in the HVM (upper) and LH (middle). SUDS in the HVM decreasein frequency for about 10 sec after the stimulation (50 c/sec, 13 V), and increase in the LH for about 10 sec (50 clsec, 7.5 V). Lower records: repetitive M n g in LH neuron by DMT stimulation. On increasing the intensity of volley stimulation (9 stimuli at 50 c/sec), the number of unitary response increases.
DISCUSSION
( I ) Reciprocity of SUDs in H V M and LH. The present experiments have established the intimate reciprocal relations between the activities of the HVM and LH. Examples are as follows. (a) By applying a repetitive stimulation upon the HVM for 2-3 sec, SUDs of the LH disappeared for a considerably long period (3-15 sec, Fig. 4) and exactly the reverse relation holds true, even for single stimuli (Fig. 5). (b) Under light
NEURONAL MECHANISM OF FEEDING
25
anesthesia in which animals showed responses to various kinds of sensory stimulation, SUDs in the HVM were high in frequency and low in the LH, both having characteristically different discharge patterns, and under deep anesthesia the relation between the frequencies was completely reversed as were the discharge patterns. In an unanesthetized and unrestrained cat, when the EEG of the HVM showed an arousal pattern of fast waves and that of the LH slow waves with high amplitude, a very slight anesthesia also produced a completely reversed pattern in EEG (Oomura et al., 1967). (c) From statistical treatments, between the SUDs in the HVM and LH there appeared a significant negative cross-correlation, i.e. when either the HVM or LH discharged more than the average in frequency for a unit time, say 250 msec, several units of time later, the other discharged less than the average. The influences of changes in blood composition (Anand et al., 1964; Iki, 1964; Oomura et al., 1964) and of visceral impulses (Anand 1961) on the HVM and LH activities were excluded from this discussion; (a> HVM and LH units were markedly affected in reverse direction by stimulating the various limbic structures. Behaviorally, as mentioned in the introduction, food intake was elicited by LH stimulation and stopped by HVM stimulation. Mechanisms and physiological functions of these reciprocities are of interest. The mutual inhibition in the eye of Limulud is well known. The frequency of an optic nerve discharge caused by illumination of one ommatidium of the Limulus eye, was decreased by an additional illumination of the neighboring ommatidium (Hartline and Ratliff, 1957). This is called the lateral inhibition which is mediated through some lateral connections in the optic nerve plexus between sensory units (Hartline et al., 1956). A big difference is that the inhibition in the hypothalamus is between neurons in the HVM and those in the LH, but not between two neighboring neurons in the same centers, whereas the lateral inhibition is between sensory units of the same kind. In our experiments, the same kind of discharge patterns and positive cross-correlations were usually confirmed between two neurons in the same center (Tables I and 11).Functionally, however, it is reasonable since the physiological functions of the HVM and LH were supposed to be of another kind; as though in two antagonistic muscles, a hyperexcitation of one center causing the hypoexcitation of the other antagonistic center, and vice versa, is in no way contradictory. The lateral inhibition works to emphasize the contrast between bright and dark areas in the visual field at their borders. The two physiological functions, the motivations of food intake and satiety, may be also emphasized in contrast, through the reciprocity of activities in the LH and HVM. Here, the contrast is not spatial but temporal. In this respect, a detailed spatial differentiation in the same center would not be required as in the eye, so mutual facilitations instead of inhibition exist between two neurons in the same center, LH or HVM. Even though it be teleological, the functions of the reciprocity do not seem difficult to understand; but to know the exact pathway through which the inhibitions are delivered between the two centers, we have much more to study, and at this stage conjectures may be inevitable. Though the HVM and LH are 2-3 mm apart from each other, SzentAgothai et al. (1962, Fig. 22) clearly demonstrated by the Golgi and Cox method that the dendrites of the HVM and LH neurons orient into mediolateral directions more than 2 mm on the References p. 31-33
26
OOMURA
et ul.
same frontal plane, to establish contacts with neurons in each others' nuclei as well as with those in the same nucleus. They showed further by even very small lesion in the HVM, numerous degenerated fibers of extremely fine caliber in the LH, showing lots of contacts of many collaterals of the HVM with the LH neurons (Szenthgothai et al., 1962, p. 67). These connections may be considered as pathways for the reciprocal inhibition. As the neurons in the same nucleus seem to have a similar kind of connection between each other, and their activities are facilitatory rather than inhibitory, the above conjecture seems at first rather unreasonable, because it suggests that activity of one neuron, say in HVM, exerts a facilitatory effect on another nucleus, without any internuncial neuron. A few of our recorded SUDS,not described in this paper, leave some suspicion of such interneurons, though the examples were too few. As for the difficulty of one neuron exerting opposite effects on different neurons without any internuncial neuron, observations on molluscan neurons may furnish some explanation. Upon ACh application in ganglia of several species of molluscus, a type of neuron called a D cell depolarizes, and another called an H cell hyperpolarizes (Tauc and Gerschenfeld, 1962; Strumwasser, 1962; Kerkut and Cottrell, 1963 ; Oomura et al., 1965). Moreover, it was found that the impulses discharged from a single neuron actually exert direct synaptic excitatory actions on D cells, and inhibitory actions on H cells, probably both mediated by ACh. The differentiation of excitation or inhibition in this event is provided by the specific properties of the ACh receptor sites. Oomura et ul. (1965) demonstrated that these receptor sites or postsynaptic membranes are not different in their reaction to the ACh, but the difference in cell metabolism maintains differentlevels of concentration of internal chloride in D and H cells, which lead to the differentiation of depolarization and hyperpolarization, but induced by chloride permeability increase by ACh. In view of this information, the difficulty mentioned earlier may not be so grave a problem. The difference in the chemical affinities between neurons in the LH and those on the HVM is conceivable, from physiological functions attributed to them, and from the results of behavioral observation upon local applications of ACh, adrenalin, etc. to these centers (Grossman, 1962; Wagner and de Groot, 1963).
The slow time course of reactions to electrical stimuli is another problem. The latency of LH unitary discharge responding to the LH single shock was almost 15 msec (Figs. 5 and 9). As the latencies did not decrease much when the strength of stimulation was increased, supposing the latency to be solely the conduction time from the site of stimulation to the recording electrode, around 0.5 and 1 mm maximum, the conduction velocity comes out to be 3-7 cm/sec. There is still the problem as to whether active conduction takes place in the dendrite (Rall, 1962). Though Chang (1951) gave a value of 1-2 m/sec for the conduction velocity of the dendrite, Cragg and Hamlyn (1955) and Andersen (1960) obtained the value of 30-50 cm/sec and 18-60 cm/sec respectively on the apical dendrite of rabbit hippocampus (CAI), and Hild and Tasaki (1962) about 10 cm/sec (38") on the dendrite of the tissue culture neuron of kittens and rats. Our values on the conduction velocity in the LH are smaller than this last and the feast value. The reactions of neurons in these centers seem to be rather more sluggish than those in the mammalian spinal cord or cortex on which most
N E U R O N A L MECHANISM OF FEEDING
27
quantitative analyses were carried out, so the slow velocity may not be inconceivable, or it may imply that each neuron even in the LH is connected functionally by interneurons as Eccles (1951) proposed for the cortex neurons of mammals. While there was little evidence of the interneuronal activity, almost all inhibitions or excitations of neurons in both centers seem to last fifty milliseconds or more in units of 10. The supposition of a few fast-working interneurons between the HVM and LH alone, unless their reactions are slow, will not be enough to explain such a slow onset of inhibition as seen in the negative cross-correlation which was of the order of seconds (Table 1). As already shown by Sawa et al. (1959), Tsubokawa and Sutin (1963), Oomura et al. (1964), and in the present study, the HVM and LH were in an intimate relationship, i.e. also reciprocal with the limbic system, the AM, lateral and medial parts of the GP, SEP and mesencephalic tegmentum, which may be another factor contributing to the slow time course of reciprocal correlation between HVM and LH. (2) The eflect of anesthesia on the HVM andLH. The anesthesia brought about considerable effects upon the activities of both HVM and LH. Under light anesthesia, the HVM neurons continued their activity well. They could keep discharging at the relatively high frequency of 10-20 c/sec and with a more or less regular pattern. This is common with the repetitive firing in the sensory nerve provoked by a long sustained depolarization elicited at the sensory terminals (Hagiwara, 1954). The high frequency in HVM neurons may also be due to the membrane depolarization. Depolarization at a certain level of the membrane potential greatly helped to produce hyperpolarization following spike (Hagiwara and Tasaki, 1958, Fig. lo). This after-hyperpolarization sometimes lasted more than fifty tens of msec (Kolmodin and Skoglund, 1958). In HVM neurons, with a Gaussian distribution of SUD frequency and TIof the interval pattern, the autocorrelation function was negative up to about 50 msec (Fig. 3), and this negativity may mean a lowered excitability (or relative refractory period) provoked by the after-hyperpolarization (Lloyd, 1951 ; Goldberg ef al., 1964) which could be especially marked when the membrane is depolarized. As another possibility, a recurrent inhibition of the Renshaw type (Eccles et al., 1954) may be considered. But so far we have obtained only few and indefinitive observations suggesting the existence of the recurrent interneuron. In the central nervous system, depolarization is usually maintained by facilitatory synaptic bombardments (Kolmodin and Skoglund, 1958 ; Hunt and Kuno, 1959). Similarly, depolarization in HVM neurons may be maintained first by arrivals of facilitatory synaptic bombardments upon them from the AM (Figs. 7A, B and 8), the lateral part of the GP (Fig. 1 l), the SEP, etc. together with those from the already known positive feedback circuit between the AM and HVM (Oomura et al., 1965, unpublished observation) or with other unknown circuits, or by a decrease in inhibitory synaptic bombardments from the LH (Figs. 1 and 4), or by both. On the contrary, in the LH neurons, the SUD frequency was low and its discharge patterns were attributable to complete Poisson processes. The lowered activity may be due to membrane hyperpolarization. When the membrane potential was high, the after-hyperpolarization following the spike was less marked, or rather after-negativity followed a spike (Hagiwara and Tasaki, 1958, Fig. 10). This corresponds to the lack of References p . 31-33
28
OOMURA
et al.
initial negativity in the temporal pattern of the autocorrelation of the LH SUDs (Fig. 3). Hagiwara (1954) also demonstrated that the serial autocorrelation of neighboring intervals on discharge records from a toad muscle spindle in which the interspike intervals showed an exponential pattern, was almost zero except at t = 0. Membrane hyperpolarizationmay be due to the increased inhibitory input or decreased excitatory input, OT both. We already know that HVM activity which was inhibitory to the LH, was high and regular under light anesthesia (Fig. 2). The limbic structures were considered as inhibitory areas to the LH, too (Figs. 7-1 1). The decrease in excitatory input to the LH from the dorsomedial mesencephalic tegmentum (Fig. 12) should also be considered. The neuronal activity of the latter probably receives inhibitory influences from the cortex as well as the medullar reticular formation; this is known to occur under light anesthesia or in the half arousal state (French et al., 1955; Magnes et al., 1961 ; Dell et al., 1961). In unrestrained conditions, the SUDs in the cat visual cortex (Hubel, 1959) and in the mesencephalic tegmentum (Huttenlocher, 1961) tended to become random, together with lowered frequency in the arousal state, and increased during sleep. Because of the increase in IPSP and decrease in EPSP, the LH neuron might keep the membrane potential hyperpolarized. Under deep anesthesia, not only the distribution patterns of the SUDs (Table I), but also the autocorrelation function (Fig. 3), were completely reversed in the HVM and LH. Further, the low frequency and randomness of SUDs in HVM neurons were mainly due to an increase in inhibitory input from the LH (Figs. 4 and 5). Low voltage fast waves in the cat cortical EEG under ether anesthesia (Adrian and Matthews, 1934) were descrided as activation sleep by Jouvet (1963). In this sleep, in the mesencephalic tegmentum, its EEG was a low voltage fast wave (Jouvet, 1963), and its SUD frequency was high (Huttenlocher, 1961). Consequently, the neuronal activity in this region should be high under the deep ether anesthesia of the present experiments, and the LH neurons might be much enhanced in their activity through the mesencephalic tegmentum (Fig. 12), while the HVM neurons, or at least some parts of them, might be inhibited (Fig. 12, Tsubokawa and Sutin, 1963). The decrease in the activity of the HVM neurons may also tend to decrease the AM activity. (3) Neuron activity in the HVM or LH. Two contiguous units in the HVM had not only the same discharge patterns but also a positive cross-correlation to each other, though the mean discharge frequencies or mean spike intervals were a little different. They had both negativecross-correlations to SUD simultaneously recorded in the LH. The same relations seem to hold in the units in the nuclei of both hemispheres: the same patterns on activities among units in either HVMs or L H s in both hemispheres, but negative correlations between units in an HVM and the contralateral LH. These observations show that the hypothalamic nuclei are considerably primitive, acting in the same way in the same nucleus. Different and complicated situations are known in the higher centers. For example, in the motor cortex neurons, Creutzfeldt and Jung (1961) found that two units recorded simultaneously by the same electrode sometimes showed reciprocal activity. Asanuma and Okada (1962) discovered that a pyramidal neuron in the motor area responded to stimulation applied exactly to the corresponding point in the opposite
N E U R O N A L MECHANISM OF FEEDING
29
hemisphere, but was inhibited when the stimulating electrode was moved 1 mm off the point. Recording activities of 3 to 4 units in the lateral geniculate body, by the insertion of a triple or quadruple electrode, whose tips were aligned along a straight line and separated about 150 p from each other, Verzeano and Negishi (1960) recognized circulating neuronal activities, i.e. bursts of impulses were recorded by a component of the electrodes, then by another, repeatedly. But the fourth component of the electrodes often recorded a different or rather inhibiting activity to the other three components. ( 4 ) Relationship between the limbic structure and the hypothalamus. The present experiments on the effects of stimulation of the limbic structure upon the activities of the HVM and LH agreed fairly well with the behavioral experiments carried out by many investigators. The fact that the SUDs in the HVM were increased and those in the LH were simultaneously considerably decreased by AM or SEP stimulation confirmed the findings related to the feeding mechanism in the AM. Koikegami et al., (1955), Green el al., (1957), and Morgane and Kosman (1960) produced hyperphagia in cats by bilateral amygdalectomy, and Fonberg and Delgado (1961) an inhibition of the food intake of hungry cats following stimulation in the AM or SEP. As regards stimulation of the lateral part of the GP, the present results mainly showed the same effects on the HVM and LH activities as with the AM stimulation. This strongly supports the results obtained by Morgane (1961a) that lesions placed on the pallido-hypothalamic pathways from the lateral part of the globus pallidus to the HVM caused considerable hyperphagia. On the other hand, the stimulation of the medial part of the GP resulted in an opposite effect on the SUDs of the HVM and LH. Morgane (1961b) also indicated that rats lesioned in this region showed an aphagia, and postdated that crossing fiber systems, particularly pallidofugal fibers to the LH were the critical elements in the organization of the feeding centers. The significance of the medial mesencephalic tegmentum upon the neuronal feeding mechanism emphasiized by Nauta (1958) and Morgane (1964) was also confirmed by our experiments. However, our results were carried out solely on the region of the periaqueductal gray substance. Further unit investigations are required in order more fully to discuss such interactions.
SUMMARY
To establish neuronal regulation of feeding, the electrical activities of the feeding and satiety centers in the hypothalamus, as well as those of the limbic system and of the mesencephalic tegmentum, were investigated. The cross functional relations among those activities were clarified. Glass pipette electrodes were used to record spontaneous unitary discharges (SUD) simultaneously from the cat’s hypothalamic lateral area (LH), the feeding center, ventromedial nucleus (HVM), satiety center, and lateral principal nucleus of the amygdala (AM), under ether anesthesia or flaxedil-immobilization (routine procaine infiltration into pressure points were carried out). For stimulating, concentric bipolar electrodes were inserted into the regions mentioned above and into the lateral or References p.!31-33
30
OOMURA
et al.
medial principal nucleus of the AM, septum (SEP), medial or lateral part of the globus pallidus (GP) and dorsomedial part of the mesencephalic tegmentum. (1) In a half arousal state under light ether anesthesia, in simultaneous recording from the HVM and LH, the HVM-SUD frequency was 10 to 20 impulses/sec on the average, yielding a Gaussian distribution, and the average interspike interval was 50 msec, the interval histogram being a T-distribution of order two (rz), showing more or less regularity. At the LH, the SUD frequency was low, a few impulses/sec, showing a Poisson distribution. The interspike interval was 200-300 msec on the average, the distribution pattern being exponential, revealing a complete randomness or a perfect Poisson process. When the anesthesia became deep, the HVM-SUD frequency was gradually decreased, while that in LH increased, and a completely reversed result in SUD distribution was obtained in the deep ether anesthetic state. (2) When two SUDs with different amplitudes were recorded simultaneously in one center, both showed quite the same discharge patterns, although they were not synchronized. This may indicate that the neurons in one center act in somewhat the same way. (3) After confirming the stationary state of the time series of SUD by variance analyses, we calculated the auto- and cross-correlations: (a) At a comparatively long t (300-1000 msec), there was a significant negative cross-correlation, revealing a marked reciprocal relation between the activities of the HVM and LH, i.e. when one center discharged more than the average for 1 t, several t’s afterwards, the other discharged less than the average. (b) At relatively short t (10 msec,) the autocorrelation function of the SUD with a Poisson distribution was zero except at t = 0, whereas that with the Fa distribution was negative for some tens of msec. Both these autocorrelation functions agreed with the mathematical analysis. The negative value of the correlation was presumed to show a lowered excitability of the neuron due to the after-spike hyperpolarization. (c) Significant positive correlations were obtained between the two SUDs in one center. This means, again, the same type of neuronal activity in the same center. (4) Under flaxedil immobilization, in simultaneous recordings from the LH and the lateral nucleus of the AM, the frequency of SUD was in general high in the LH, yielding a Gaussian distribution, and low in the AM, yielding a Poisson distribution. The cross-correlation function of the SUD between the LH and AM was calculated, resulting in significant negative correlations. These suggest a marked reciprocal relation of the neuronal activities between LH and AM. ( 5 ) Not only the SUDs, but also the electrical stimulation experiments, confirmed the reciprocal relationships between the HVM and LH. The number of the SUDs in the LH was markedly decreased for a considerable time by single or repetitive electrical stimulation of the AM but the reverse was true for the HVM. (6) Regarding the mutual facilitatory circuit, there were intimate functional associations between the HVM and AM, and between the LH and the dorso-medial part of the mesencephalic tegmentum (DMT). By stimulations of the PGS, a considerable increase in frequency of the SUD in the LH was also recorded, and, sometimes, a decrease in the activity of the HVM.
N E U R O N A L MECHANISM OF F E E D I N G
31
(7) The lateral part of the G P behaved on the HVM and LH approximately in the same direction as did the AM, whereas the medial part of it behaved as did the DMT. (8) The driven unitary discharges were recorded in the HVM by single or volley stimulation of the SEP. (9) The hypothalamic feeding mechanism, therefore, is well maintained by the reciprocal activity between the HVM and LH as well as by a close linkage of the negative and positive feedback system made by the relationship among the hypothalamus, limbic system and dorsomedial mesencephalon. ACKNOWLEDGMENTS
The author is greatly indebted to Dr. S . Kano, Professor of Mathematics, Faculty of Science, Kagoshima Universtiy, Kagoshima, for the statistical treatments. This work was supported in part by aids from the Ministry of Education, the Rockefeller Foundation (Re GA MNS 6194), the Japan Waksman Foundation (1962), and the U.S. Army Research and Development Group (Far East) (DA-92-557-FEC-37352).
REFERENCES ADRIAN,E. D., AND MAITHEWS,B. H. C., (1934); The interpretation of potential waves in the cortex J. Physiol., 81,44M71. ANAND,B. K., (1961); Nervous regulation of food intake. Physiol. Rev., 41,677-708. ANAND,B. K., AND BROBECK, J. R., (1951); Hypothalamic control of food intake in rats and cats. Yale J . Biol. M a t , 24, 123-140. ANAND,B. K., CHHINA,G. S., SHARMA, K. N., DUA, S., AND SINGH,B., (1964); Activity of single neurons in the hypothalamic feeding centers: effect of glucose. Amer. J. Physiol., 207, 11461154. ANDERSEN, P., (1960); Interhippocampal impulses. 11. Apical dendritic activation of CA1 neurons. Acta physiol. scand., 48, 178-208. ASANUMA, H., AND OKUDA,O., (1962); Effects of transcallosal volleys on pyramidal tract cell activity of cat. J. Neurophysiol., 25, 198-208. BARRACLOUGH, C. A., AND CROSS,B. A., (1963); Unit activity in the hypothalamus of the cyclic female rat: effect of genital stimuli and progesterone. J. Endocrinol., 26, 33%359. BROBECK, J. R., (1946); Mechanism of the development of obesity in animals with hypothalamic lesions. Physiol. Rev., 26, 541-559. BROOKS, D. C., (1959); Electrical activity of the ventromedial nucleus of the hypothalamus. Amer. J. Physiol., 197, 829-834. CHANG,H.T., (1951); Dendritic potential of cortical neurons produced by direct electrical stimulation of the cerebral cortex. J. Neurophysiol., 14, 1-21. CRAW,B. G., AND HAMLYN, L. H., (1955); Action potentials of the pyramidal neurones in the hippocampus of the rabbit. J. Physiol., 129, 608427. CREUTZFELDT, O., AND JUNG,R., (1961); Neuronal discharge in the cat's motor cortex during sleep and arousal. The Nature of Sleep, G. E. W. Wolstenholme, and M. O'Connor, Editors, London, Churchill, pp. 131-170. DELGADO, J. M. R., AND ANAND,B. K., (1953); Increase of food intake induced by electrical stimulation of the lateral hypothalamus. Amer. J. Physiol., 172, 162. DELL,P., BONVALLET, M., AND HUGELIN, A., (1961); Mechanisms of reticular deactivation. The Nature of Sleep, G . E. W. Wolstenholme and M. O'Connor, Editors, London, Churchill, pp. 86107. ECCLES, J. C., (1951); Interpretation of action potentials evoked in the cerebral cortex. Electroenceph. elin. Neurophysiol., 3, 449464.
32
OOMURA
et al.
ECCLES,J. C., (1953); The Neurophysiological Basis of Mind, Oxford, Clarendon Press, p. 122. ECCLES, J. C., (1964); The Physiology of Synapses, Berlin, Springer-Verlag, p. 84. ECCLES,J. C., FA^, P., AND KOKETSU, K., (1954); Cholinergicand inhibitory synapses in a pathway from motor-axon collaterals to motoneurons. J. Physiol., 126, 524-562. FONBERG, E., AND DELGADO, J. M. R., (1961); Avoidance and alimentary reactions during amygdala stimulation. J . Neurophysiol., 24,651-664. FRENCH, J. D., HERNANDEZ-PE~N, R., AND LMNGSTON, R. B., (1955); Projections from cortex to cephalic brain stem (reticular formation) in monkey. J. Neurophysiol., 18, 74-95. FREUND, J. E., (1960); Modern Elementary Statistics, 2ndEd., Englewood Cliffs, N. J., Prentice-Hall. pp. 203-220. J. L., ROSVOLD, H. E., AND PRIBRAM K. H., (1957); The effect on affective and cognitive FULLER, behavior in the dog of the lesions of piriform-amygdala-hippocampalcomplex. J . comp. physiol. Psychol., 50,89. GERSTEIN, G. L., AND KIANG,N. Y-S., (1960); An approach to the quantitative analysis of electrophysiological data from single neurons. Biophys. J., 1, 15-28. GOLDBERG, J. M., ADRIAN,H. O., AND SMITH,F. D., (1964); Response of neurons of the superior olivary complex of the cat to acoustic stimuli of long duration. J. Neurophysiol., 27, 706-749. GREEN,J. D., CLEMENTE, C. D., AND DE GROOT,J., (1957); Rhinencephalic Iesions and behavior in cats. J. comp. N2urol., 108, 505-545. GROSSMAN, S. P., (1962); Direct adrenergic and cholinergic stimulation of hypothalamicmechanisms. Amer. J. Physiol., 2Q2,872-882. HAGIWARA, S., (1954); Analysis ofinterval fluctuation of the sensory nerve impulse. Jap. J. Physiol., 4, 234-240. HAGIWARA, S., AND TASAKI,I., (1958); A study on the mechanism of impulse transmission across the giant synapse of the squid. J. Physiol., 143,114-137. HARTLINE, H. K., AND RATLIFF,F., (1957); Inhibitory interaction of receptor units in the eye of Limulus. J. gen. Physiol., 40, 357-376. HARTLINE, H. K., WAGNER, H. G., AND RATL.IFF, F., (1956); Inhibition in the eye of Limulus. J. gen. Physiol., 39, 651-673. HETHERINGTON, A. W., AND RANSON, S. W., (1940); Hypothalamic lesions and adiposity in the rat. Anat. Rec., 78, 149-172. HETHERINGTON, A.W., AND RANSON, S. W., (1942); The spontaneous activity and food intake of rats with hypothalamic lesions. Amer. J. Physiol., 136, 609-617. HILD,W., AND TASAKI,I., (1962); Morphological and physiological properties of neurons and glial cells in tissure culture. J. Neurophysiol., 25, 277-304. HUBEL,D. H., (1959); Single unit activity in striate cortex of unrestrained cats. J. Physiol., 147, 226-238. HUNT, C. C., AND KUNO,M., (1959); Background discharge and evoked responses of spinal interneurones. J. Physiol., 147, 364-384. H W I T E ~ H EP.RR., , (1961); Evoked and spontaneous activity in single units of medial brain stem during natural sleep and waking. J. Neurophysiol., 24,451468. IKI,M., (1964); Firing responses of neurons in the hypothalamic feeding center to variable electrical and chemical stimulations. Acta med. Univ. Kagoshima, 6, 155-175. JASPER, H. H., AND -ONE MARSAN, C., (1954); A stereotaxic Atlas of Dicncephalon of the Cat. Ottawa, The National Research Council of Canada. JOUVET, M., (1963); The rhombencephalic phase. of sleep. Progress in Brain Research, VoI. I , Brain Mecknisms, C. Moruzzi and A. Fessard, Editors, Amsterdam, Elsevier, pp. 406-424. KmU<W, G . A., AND C O ~ ~ E LG. L ,A., (1963); Acetylcholine and 5-hydroxytryptamine in the snail brain. Comp. Biochem. Physiol., 8, 53-63. KOIKEGAMI, H., FUSE, S., YOKOYAMA, T., WATANABE, T., AND WATANABE, H., (1955); Contributions to the comparative anatomy of the amygdaloid nuclei of mammals with some experiments of their destruction or stimulation. Folia psychiat. neurol. jap., 8, 337-370. KOLMODIN, G. M., AND SKOGLUND, C. R., (1958); Slow membrane potential changes accompanying excitation and inhibition in spinal moto- and interneurons in the cat during natural activation. Aeta physiol. s c a d . , 44, 11-54. LLOYD,D. P. C., (1951); Electrical signs of impulse conduction in spinal motoneurons. J. gen. Physiol., 35, 255-288. MAGNES,J., MORUZZI,G., AND POMPEUNO,O., (1961); Electroencephalogram-synchronizing struc-
N E U R O N A L M E C H A N I S M OF FEEDING
33
tures in the lower brain stem. The Nature of Sleep, G. E. W. Wolstenholme and M. OConnor, Editors, London, Churchill, p. 57. MILLER,N. E., (1960); Motivational effects of brain stimulation and drugs. Fed. Proc., 19, 846. MORGANE, P. J., (1961a); Alternations in feeding and drinking behavior of rats with lesions in globi pallidi. Amer. J. Physiol., 201, 420-428. MORGANE, P. J., (1961b); Electrophysiological studies of feeding and satiety centers in the rat. Amer. J. Physiol., 201, 838-844. MORGANE, P. J., (1964); Limbic-hypothalamic-midbrain interaction in thirst and thirst motivated behavior. Thirst, M. J. Wayner, Editor, Oxford, Pergamon Press., p. 429. MORGANE, P. J., AND KOSMAN, A. J., (1960); Relationship of middle hypothalamus to amygdalar hyperphagia. Amer. J. Physiol., 198, 1315-1318. NAUTA,W. J. H., (1958); Hippocampal projections and related neural pathways to the mid-brain in the cat. Brain. 81,319-340. OOMURA, Y., KIMURA, K., OOYAMA, H., MAENO,T., IKI,M., AND KUNIYOSHI, M., (1964); Reciprocal activities of the ventromedial and lateral hypothalamic areas of cats. Science, 143, 484485. OOMURA, Y., OOYAMA, H., AND SAWADA, M., (1965); Ionic basis of the effect of ACh on Onchidium D- and H-neurons. XXIZI Int. Congr. Physiol. Sci., No. 913. OOMURA, Y., OOYAMA, H., YAMAMOTO, T., AND NAKA,F., (1967); Reciprocal relationship of the lateral and ventromedial hypothalamus in the regulation of food intake. Physiol. Behuv., 2, in press. POGGIO,G. F., AND VIERNSTEIN, L. J., (1964); Time series analysis of impulse sequences of thalamic somatic sensory neurons. J. Neurophysiol., 27, 517-545. RALL,W., (1962); Electrophysiology of a dendritic neuron model. Biophys. J., 2, 145-167. SAWA,M., MARUYAMA, N., HANAI,T., AND KAJI, S., (1959); Regulatory influence of amygdaloid nuclei upon the unitary activity in ventromedial nucleus of hypothalamus. Fol.psychiut. neurol.jap., 13, 235-256. SKULTELY, F. M., AND GARY,T. M., (1962); Experimental hyperphagia in cats following destructive midbrain lesions. Neurology, 12, 394401. SMITH,0. A., (1956); Stimulation of lateral and meidal hypothalamus and food intake in the rat. Anat. Rec., 124, 363-364. SPRAGUE, J. M., CHAMBERS,W. W., AND STELLAR, E., (1961); Attentive, affective, and adaptive behavior in the cat. Science, 133, 165-173. STEVENSON, J. A. F., (1964); The hypothalamus in the regulation of energy and water balance. Physiologist, 7, 305-318. STRUMWASSER, F., (1962); Post-synaptic inhibition and excitation produced by different branches of a single neuron and the common transmitter involved. XXIZ Internat. Congr. Physiol. Sci., Vol. 2 No. 801. STUART, D. G., PORTER, R. W., ADEY,W. R., AND KAMIKAWA, Y.,(1964); Hypothalamic unit activity. I. Visceral and somatic influences. Electromceph. clin. Neurophysiol., 16,237-247. SZENTAGOTHAI, J., FLERKO, B., MESS,B., AND H A L ~ Z B.,, (1962); Hypothalamic Control of the Anterior Pituitary, Budapest, Akadtmiai Kiadb. TAUC, L., AND GEPSCHENFELD, H. M., (1962); A cholinergic mechanism of inhibitorysynaptictransmission in a molluscan nervous system. J. Neurophysiol., 25, 236-262. TSUBOKAWA, T., AND S m ,J., (1963); Mesenczphalic influence upon the hypothalamic ventromedial nucleus. Electroenceph. clin. Neurophysiol., 15, 804-810. VERZEANO, M., AND NEGISHI,K., (1960); Neuronal activity in cortical and thalamic networks. A study with multiple microelectrodes. J. gen. Physiol., 43, (6) Suppl. 177-195. WAGNER, J. W., AND DE GROOT, J., (1963); Changes in feeding behavior after intracerebral injections in the rat. Amer. J. Physiol., 201,483-487. WERNER, G.,AND MOUNTCASTLE, V. B., (1963); The variability of central neural activity in a sensory system, and its implications for the central reflexion of sensory events. J. Neurophysiol., 26, 958-977. WOOD,D. C., (1958); Behavioral changes following discrete lesions of temporal lobe structures. Neurology, 8,215-226.
34
Limbic and other Forebrain Mechanisms in Sleep Induction and Behavioral Inhibition * CARMINE D. CLEMENTE AND MAURICE B. STERMAN Department of Anatomy and Brain Research Institute, University of California, Los Angeles, and the Sepulvedu V.A. Hospital, Sepulveda, Calif. (U.S.A.)
INTRODUCTION
Cortical EEG synchronizationis of interest to the behavioral physiologist because it is characteristic of the spontaneously occurring cortical spindle pattern associated with sleep and other states of central nervous system suppression. Perhaps one of the first contributions which implicated the brain stem’s influence ovei the state of the EEG, was the classical observation of Bremer (1935) who showed that transaction of the neuraxis at the high cervical level resulted in an EEG pattern of wakefulness, while a more rostra1 transection at the midbrain level, resulted in an EEG pattern similar to sleep. The functional importance, however, of the intrinsic brain stem mechanisms involved in the maintenance of an alert, active EEG was not appreciated until Magoun with his associates Moruzzi (1949) and Lindsley and Bowden (1949) reported their well known stimulation and lesioning experiments in the late 1940’s. Evidenced further by the work of French and Magoun in the primate (1952), this research led to the formation of the concept of an ascending reticular activating system in the core of the brain stem which extends rostrally into the thalamus. Other experiments performed earlier by Dempsey and Morison (1942a, b; 1943) indicated that low frequency stimulation of the intralaminar thalamic muclei was capable of driving the electrical activity of the cerebral cortex in the opposite direction, producing instead, the recurrent EEG spindle burst pattern termed ‘recruitment’. These findings stressed, in our view, the potential for a differential influence from various areas in the central nervous system upon ongoing electrocortical activity. Hess (1954, 1957) contributed significantly to this concept by demonstrating the induction of behavioral sleep with low frequency electrical stimulation of nuclei in the medial thalamic mass. In his distinguished studies on the functions of the hypothalamus Hess differentiated two areas subserving, generally speaking, antagonistic functions. The sympathetic nature of the response, and the generalized behavioral activation resulting from stimulation of the caudal hypothalamus, led him to call this area the
*
Aided by a grant from the U. S. Public Health Service (MH-10083).
MECHANISMS I N SLEEP I N D U C T I O N A N D B E H A V I O R A L I N H I B I T I O N
35
‘ergotropic zone’. More anteriorally located hypothalamic zones produced generalized somatomotor suppression and parasympathetic responses to stimulation, and to these areas he ascribed the term ‘trophotropic zone.’ He included the medial thalamic areas, capable of inducing sleep in this latter functional system. In an evaluation of other brain stimulation studies, our attention was drawn to the preoptic-basal forebrain area, in which there appeared to exist potent suppressor mechanisms affecting a variety of peripheral functions. We felt that other mechanisms may exist there which would be reflected in the EEG and perhaps also in the animal’s behavior. If our thinking in these matters was consistent with the many correlating observations of others, the effects of stimulation here should tend to synchronize the electrical activity of the cerebral cortex and perhaps also produce behavioral suppression.
METHODS A N D R E S U L T S
I. Acute stimulation and recording experiments The first series of electrophysiological experiments were performed in the brains of acutely prepared adult cats. General anesthetic drugs often depress cortical electrical activity thereby masking the cortical influences of stimulation in subcortical areas. Because of this a neuromusclular blocking agent, Flaxedil, was used in conjunction with local anesthesia at incision sites in an effort to achieve recordings against the background of an activated and responsive cerebral cortex. Operations were carried out under general ether anesthesia and electroencephalographic recording was achieved by the use of phonograph needle electrodes placed into the calvarium over the cerebral cortex. Upon an EEG background characterized by low voltage fast activity, the effects of stimulating various basal forebrain sites with concentric stereotaxically placed bipolar electrodes were observed. It was found that stimulation in certain preoptic and basal forebrain areas produced an immediate, sustained and diffuse cortical synchronization. This response was observed upon stimulation at low voltages (2-2.5 V at 0.75 msec duration) and was found most effective at lower frequencies (5-7 per sec). Examples of this cortical synchronization can be seen in Fig. 1. The subcortical points, which upon stimulation resulted in an onset of cortical synchronization, were mapped anatomically and the results from 25 exploration experiments are shown collectively in 4 cross-sectional diagrams of the cat brain in Fig. 2. At A15.5 and A16 a distinct concentration of positive stimulation sites was observed. This focus extended caudally into the preoptic region of the hypothalamus, but at more posterior hypothalamic sites, electrical stimulation did not result in the onset of cortical synchronization. Other experiments were carried out to assess the interaction between stimulation of both the brain stem reticular activating system and the basal forebrain-preoptic synchronizing sites. Figs. 3 and 4 show the results of such an interaction experiment. In the upper set of traces in Fig. 3 is illustrated the well known cortical arousal phenomenon resulting from high frequency stimulation in the reticular activating system. The second series of traces in Fig. 3 shows, however, that if in the course of this arousal, References p . 47
R.PAn.OCC. L. PAR. occ.
TR. PAR. TR. occ.
9
z
tl
-
PROLONGED STIMULATION (47 secondd
CAT SC 29 l6/S/18611
R. FR. PAR. c.
m
E
yyy*yw
TR. FRONIAL TR. PAR.
rk n ,,
3v
0.751~~
5.5 Y e
“ouv P
I
Fig. 1. These three series of EEG traces show the diffuse bilateral cortical synchronization obtained from stimulation of the basal forebrain in three different cats. Note the ‘recruiting’ nature of the response in the first and third records. (From Sterman and Clemente, 1962a.)
MECHANISMS I N S L E E P I N D U C T I O N A N D BEHAVIORAL I N H I B I T I O N
37
the basal forebrain zone was simultaneously stimulated at low frequency, a cortical synchronization was induced, superimposed upon the original activated EEG. The reverse interaction effectwas also achieved and can be seen in the first series of traces of Fig. 4. If basal forebrain stimulation preceded stimulation of the reticular formation the induced cortical synchronization was then replaced by arousal. The second series of traces in Fig. 4. shows the sensitivity of both the arousal response and the induced cortical synchronization to the intravenous administration of 15 mg/kg of barbiturate anesthesia, Nembutal. This non-anesthetizing dose of barbiturate anesthesia clearly established the ineffectiveness of basal forebrain stimulation following barbiturate administration, and perhaps indicates why this rather pronounced effect has not been observed before. 10
8
6
4
2
8
A15
A I4 10
0 1 0
8
6
4
2
01 1 0 8
6
4
2
0
5 mm
6
4
2
0
8
6 4
2 0
-2 -4 -6
A16
A17
Fig. 2. This figure shows 4 frontal diagrams of the cat brain taken from the Jasper-Ajmone Marsan atlas. The black circles indicate points where stimulation resulted in cortical synchronization;the negative points are designated by minus signs. (From Sterman and Clemente, 1962a). References p. 47
38
CLEMENTE A N D STERMAN
ZI. Chronic behavioral experiments In conjunction with acute experiments, other adult cats and adult monkeys were prepared for long-termed chronic study. Bipolar strut electrodes were placed into the regions which were found capable of inducing synchronous cortical activity upon electrical stimulation in acute preparations. The animals were permitted to recover and the effects of stimulation in these unrestrained animals were observed. EEG activity was recorded through leads taken from a number of small screws placed into the calvarium at the time of operation. Bilateral stimulation at basal forebrain sites in these chronically prepared animals at levels of 2-3 V, 0.75 msec and at frequencies of 5-7 per sec resulted in the appearance of drowsiness and sleep within 1-2 min from the commencement of stimulation. This behavioral transition was accompanied by a correlated shift in the frequency and amplitude characteristics of the electroencephalogram. The low voltage fast activity of the activated EEG, within a few seconds from the onset of stimulation, gave way to the spindling patterns observable during the initial stages of sleep (Fig. 5). When the animals appeared to be behaviorally asleep, they could be aroused by tapping on the cage and then returned to sleep once again by basal forebrain stimulation. In other animals, electrodes were also placed into the reticular activating system and these cats could successivelybe aroused by stimulating the reticular formation at high frequency and put back to sleep by stimulating the basal forebrain region at low frequencies. The sequence of events initiated by the onset of electrical stimulation in the basal forebrain usually consisted of a cessation of ongoing behavior, a retreat on the part of the cat to some corner of the cage, the assuming of the reclining position, and the closing of the eyelids. These behavioral correlates of sleep were accompanied by the EEG spindling patterns characteristic of sleep. The effective stimulation parameters and the time required for the onset of sleep seem to be dependent on several factors. Generally, stimulation at 1-3 V with 0.5-0.75 msec duration was effective for frequencies varying from 5-250 impulses per sec. In some animals the onset of sleep was seen to occur with latencies as short as 5-10 sec after the start of stimulation, whereas others required as long as 2-3 min between the onset of stimulation and the occurrence of drowsiness or sleep. The average time required was about 30 sec. i t was found that factors such as the time of day, time of feeding and the numbers of times that the animal had been stimulated previously proved to be important determinants with respect to the ease with which these effects could be induced. Fig. 6 shows an animal in which sleep was induced with bilateral basal forebrain stimulation at higher frequencies (I 50 c/s, 1 V, 0.75 msec). The transition in this animal from alert behavioral posture seen on the left to the quiescent sleep posture seen on the right occurred approximately 20 sec after the onset of stimulation and was accompanied by a rapid transition from an EEG of wakefulness to an EEG of sleep. III. Conditioning experiments
in another series of adult male cats electrodes were chronically implanted into the brain and recording stainless steel screw electrodesplaced into the skull bilaterally over
STIMULATION OF R.A.S. A2. L4, H t I 3OO/SEC 1.5 V 0.15 MSEC
n
0
z
P
z
U
I
STIMULATION OF R.A.S. A2, L4. H t I 1
300/SEC
STIMULATION OF BASAL FOREBRAIN A14, L4, H-5 6.5/SEC 4 V 0.75 MSEC
I
1.5 V 0.15 MSEC
J
Fig. 3. The upper series of EEG traces shows activation of the EEG upon stimulation of the reticular activating system. The lower traces show first an 'activation of the EEG upon reticular formation stimulation and then superimposed upon this activation can be observed the cortical synchronization elicited during simultaneous stimulation of the basal forebrain synchronizing zone. (From Sterman and Clemente, 1962a) w
v,
CAT SC N0.4 6-8-60
I BASAL FOREBRAIN ON I R.A.S. ON
7/SEC 3 V 0.5 MSEC
I OFF
300/SEC 1.5 V 0.15 MSEC
H
I OFF
I SEC
Fig. 4. The upper series of EEG traces shows the induced cortical synchronization obtained by stimulation in the basal forebrain and its interruption by simultaneous stimulation of the reticular activating system. The lower series of traces shows that both the reticular activation response and the basalforebrain synchronization response are abolished following an injection of 15 mg/kg of Nembutal. (From Sterman and Clemente, 1962a).
MECHANISMS I N SLEEP INDUCTION AND BEHAVIORAL INHIBITION
41
the frontal, parietal, and occipital lobes of the cerebral cortex. After a postoperative period of two weeks, these animals were allowed to become familiar with an observation cage and the effective basal forebrain stimulation values determined,with respect to the induction of synchronized EEG activity and behavioral sleep. Conditioning experiments were carried out in these animals as follows: A tone of a given frequency was presented for 10 sec in advance of basal forebrain stimulation and was continued throughout an employed 20-sec period of brain stimulation. The tone and stimulation overlapped for a period of 20 sec and terminated simultaneously. An intertrial interval of 30 sec lapsed between the termination of one trial and the subsequent presentation of the next tone. Trials were repeated every minute throughout the conditioning session. Initially, the presentation of a 2000 c/s tone evoked either a desynchronization of the EEG in the non-alerted cat or little change in the flat low
S.C. NO. 15 ( Z V , 0.75MSEC. 6.5/SEC BILATERALLY) L. F! -0. R. I?-0. TR. OCC. TR. PAR. ALERT
V
DROWSY ASLEEP
OFF
ON
ALERT
DROWSY
ASLEEP
AROUSED BUT DROWSY
t
BACK TO SLEEP
DROWSY
ASLEEP
AROUSED
OFF
"
ON
-
t
AROUSED
Fig. 5. Behavioral and EEG changes induced repeatedly by basal forebrain stimulation in the freely moving cat. In this continuous EEG recording it can be seen that the slow wave electrical activity elicited by stimulationwas consistent with the concomitantbehavioral pattern reflectingthe initiation of sleep. (From Sterman and Clemente, 1962b.)
voltage pattern of the alerted animal. After a few pairings of the tone with bilateral stimulation in the basal forebrain, a conditioned EEG and behavioral response became apparent. This is illustrated in Fig. 7, where the presentation of the 2000 c/s tone on trial 38 evoked a shift in the predominantly desynchronized EEG to one of low frequency high amplitude synchronized activity. At the same time, the alerted animal suddenly showed sleep preparatory behavior at the onset of the tone. After the establishment of a conditioned EEG synchronization to the onset of a 2000 c/s tone, the References p. 47
42
CLEMENTE A N D STERMAN
frequency of the tone was changed to 4000 c/s between trials 45 and 46. Upon hearing the 4OOO c/s tone for the first time, the animal's EEG response generalized to the previous tone of 2000 c/s. On trial 46, instead of following the 4000 c/s tone by low frequency stimulation in the basal forebrain, the animal was stimulated by high frequency low voltage pulses in the miedal thalamic nuclei. These stimulation parameters characteristically induced an alert EEG. In our conditioning experiment the initial response to
Fig. 6. Behavioral response to high frequency electrical stimulation of the basal forebrain area in the cat (150 c/sec, 1 Volt, 0.75 msec). In this alert but unaroused animal (left) the induction of sleep was accomplishedin approximately 20 sec and was accompaniedby an equallyrapid EEG shift to a pattern of slow wave activity. (From Sterman and Clemente, 1962b.)
this alerting stimulus was a transitory synchronization which was shortly followed by the expected activation of the EEG. Further pairings of the 2000 c/s tone with basal forebrain stimulation and the 4000 c/s tone with high frequency stimulation in the medial thalamic nuclei resulted in the establishment of classical conditioned responses. These experiments indicated that stimulation in the basal forebrain could be paired with a previously neutral conditioning sensory stimulus in such a manner that the presentation of the tone alone was able to establish the conditioned synchronized EEG response. We should like to point out that the basal forebrain area is the only non-specific central nervous system site which has been found capable of producing a conditioned synchronization in the ongoing EEG activity. This classical Pavlovian conditioning procedure lends further evidence to the strong functional relationship existent between the basal forebrain area and the cerebral cortex. IV. Pathway studies
An additional series of acute cat preparations was studied in an effort to determine electrophysiologically the connections of the basal forebrain area with other parts of
EXPL. DESIGN TRIALS
A
TRIAL 15
B
TRIAL 38
1-45; 2000 c/sec 4 6 - 5 6 : 4000 c/sec
TONE PAIRED WITH B.F. STIMULATION TONE PAIRED WITH THALAMIC STIMULATION
R. P. COR -X TR. PAR CORTEX
.
pog$.iL
PAR
CONDITIONED E E G SYNCHRONIZATION RESPONSE TONE O N
C
2000
L
B
A
S
A
L FOREBRAIN STIM. 5 l r e c . 4 5 V , 0 7 5 m r e c
C/WC
TRIAL 4 6
I
I I
l00)JV GENERALIZATION
MEDIAL THALAMUS STIM ; 1501sec , 3 OV. 0 7 5 mrec
'
2 aec.
'
Fig. 7. Development and generalization of conditioned EEG synchronization in the cat. Early in conditioning (A) the 2000 c/sec tone produces no change in ongoing EEG activity. However, after many pairings with basal forebrain stimulation (B) this tone is seen to evoke a specific shift to slow wave, high voltage activity. When a 4000 c/sec tone is subsequently presented (C) the conditioned response is even more marked. This tone is followed by high frequency stimulation of the medial thalamus, to which the animal appropriately responds by a somewhat delayed EEG desynchronization. (From Clemente et al., 1963).
44
CLEMENTE A N D STERMAN
the central nervous system. Adult cats were surgically prepared for artificial respiration under ether anesthesia and placed in a stereotaxic instrument. Various areas of the cerebral cortex were exposed, and cortical and subcortical electrodes were located for evoked potential studies. Wound margins and pressure points were infiltrated with xylocaine and a neuromuscular blocking agent administered. Following recovery from the ether anesthesia stereotaxicplacement of electrodes in the basal forebrain area was verified by electrical stimulation through the induction of EEG slow wave activity as the criterion for correct localization, Single biphasic shocks were then delivered to this area, while the electrical activity of various cortical and subcortical sites was explored systematically on a multiple beam oscilloscope. Evoked potentials were observed in a discrete region of the ventrolateral frontal cortex, in the temporal lobe, amygdala, hippocampus, septum, midline thalamus and mesencephalic reticular formation (Fig. 8). The general region of the anterior sylvian gyrus and the more rostra1 temporal lobe cortex were the only cortical areas which showed a direct electrical evoked potential to basal forebrain stimulation. The electrical activity recorded in the hippocampus consisted of a short and a long latency bimodal response. Subsequent lesion studies indicated that the short latency component was mediated by a basal forebrain-septalhippocampal pathway, whereas the long latency component required the integrity of temporal-ammonic tracts. Responses were also evoked in the basal forebrain region by stimulation of the orbital gyrus in the cerebral cortex and in the hippocampus by stimulation of this same cortical area. These findings suggest the involvement of the orbital gyrus and lateral orbital regions of the cerebral cortex and a ‘limbic loop’ including septum, amygdala, hippocampus and temporal cortex, in a descending inhibitory pathway through the basal forebrain ultimately expressing its influence upon regulatory nuclei in the thalamus, mesencephalon, and pontine brain stem (Fig. 8). Induced synchronization observed in midline thalamic nuclei (Clementeand Sterman, 1963)and more generally throughout the cerebral cortex upon stimulation of the basal forebrain region as well as the orbital gyrus, may, therefore, reflect an integrated inhibitory influence upon brain stem structures.
DISCUSSION
It is our interpretation that the forebrain areas stimulated in these experiments, resulting in the production of widespread cortical synchronization in restrained animals and arrest reactions and sleep in behaving animals constitutes a functional brain system which expresses its behavioral suppressor influence over a wide variety of somatic and visceral functions. Furthermore, it is felt that the level of cortical activation or synchronization may be regulated by dually active systems : the reticular activating system of Moruzzi and Magoun (1949) and this forebrain corticalsynchronizingsystem. We wish to propose that these two systems act reciprocally, perhaps through relays in the limbic circuit and in the diffuselyprojecting thalamo-cortical system, of which the latter has been shown capable of activating the EEG with high frequency stimulation
zm
P
0
4 -4
I
z z
I
v1
I v1
z
I
v,
r
n
m -0
I
Z
U
C
0
2
0
z > 2
CI W
m X
>
<
I
0
F
z r
I
z
z,
2
i 0 2:
Fig. 8. Samples of evoked potentials obtained during electrophysiological studies of basal forebrain projections in acute preparations are here shown superimposed upon a phantom representation of the cat brain. With single shock stimulation of the basal forebrain area (BF) evoked responses were observed at A and A', lateral orbital cortex (15 msec); B, hippocampus (28 msec); C, septum (44 msec); D, reticular formation (10 msec). Reciprocal responses were also observed from some of these structures which, together with a consideration of latencies and specific lesion effects, suggest a descending inhibitory pathway from through BF to D, with an associated recurrent limbic projection from BF through A, B ,and consecutively back to the BF.
%
46
CLEMENTE A N D STERMAN
(accompanied by behavioral arousal) and synchronizing the EEG with low frequency stimulation (Monnier el al., 1963). In addition to this effective forebrain synchronizing mechanism, lower regions of the neuraxis contain mechanisms which are also capable of exerting a generalized EEG synchronizing effect. Thus, Bonvallet et al. (1954) and Magnus et al. (1960) have been able to produce cortical synchronization by either increasing visceral afferent discharge through carotid sinus distension or by low frequency stimulation of the tractus solitarius and adjacent reticular structures in the bulb. Pompeiano and Swetti(1961) have also been able to induce EEG synchronization by low frequency peripheral nerve stimulation. This evidence suggests that EEG synchronization,behavioral suppression and sleep are each capable of being initiated by stimuli from the extreme ends of the nervous system. Environmental conditions causing a monotonous, low frequency afferent discharge of either exter- or interceptors at one end or the active onset of an intracerebral mechanism at the other, could each exert similar influences which eventually act upon the thalamus and, thus, induce cortical synchronization. The question can be asked as to why such a potent cortically directed influence from this frequently explored basal forebrain region eluded description prior to our studies. We feel there are several reasons for this, one being methodological. Usually the acute animal preparation employed for this type of research is the barbiturate anesthetized rat. Thus, Morison et al, (1943) examined the influence of basal forebrain stimulation on the intermittent 6-10 per sec bursts characteristically seen in the EEG of such a preparation. They noted inhibition of these bursts elicitable from a wide distribution of points, and concluded that they were probably stimulating sensory and extrapyramidal fiber systems through these zones on their way to the thalamus. In this regard it should be pointed out that intravenous doses of 15 mg/kg of Nembutal completely abolished the basal forebrain induced cortical synchronization. Hess (1957) stimulated this area unilaterally in unanesthetized cats and noted the development of a diminished motor behavioral pattern which he termed ‘adynamia’. He did not monitor cortical activity simultaneously, and the recent replication of his work by Akert et al. (1952) utilizing EEG recordings was concerned primarily with the effects of thalamic stimulation. A consideration of the anatomy of this area suggests a second reason for the delay in its description. Many neurophysiologists have a compulsion to relate accentuated neuronal nuclear groups to function and this may have prevented many investigators from seeking some physiological unit in amorphous zones such as the preoptic basal forebrain regions. The structures surrounding the basal forebrain sychronizingarea including the septa1 nuclei, anterior commissure and caudate nucleus dorsally, the amygdala and striatum laterally, the posterior orbital cortex, insula and olfactory system rostrally, and the hypothalamus caudally, may have distracted some attention from it; but, additionally, these structures also serve to accent its potential for integrative function. With respect to the mechanisms operant in the onset of normal physiological sleep, we believe that a ‘passive reticular deactivation’ brought about by the volitional reduction of sensory stimuli to the reticular activating system sets the stage for the overt expression of a potently active forebrain inhibition which, as a conditioned phenome-
M E C H A N I S M S I N SLEEP I N D U C T I O N A N D B E H A V I O R A L I N H I B I T I O N
47
non, leads directly to initial stages of sleep. These observations support the popular interpretation that the processes of suppression and behavioral inhibition, like excitation, are active basic functions of the central nervous system. The more rostra1 basal forebrain area, together with certain limbic, thalamic and perhaps lower brain stem sites may constitute such a system -capable of responding to conditions favoring behavioral inhibition and acting to bring the nervous system into line with the physiologic requirements at any instant.
REFERENCES
W. P., AND HESS,JR., R., (1952); Sleep produced by electrical stimulation of the AKERT,K., KOELLA, thalamus. Amer. J. Physiol., 168, 260-267. M., DELL,P., ET HIEBEL, G., (1954); Tonus sympathique et activite electrique corticale. BONVALLET, Electroenceph. clin. Neurophysiol., 6, 119-144. F., (1935); Cerveau 'isol6' et physiologie du sommeil. C . R. SOC.Biol. (Paris), 118, 1235BREMER, 1241. M. B., (1963); Cortical synchronization and sleep patterns in acute CLEMENTE, C. D., AND STERHAN, restrained and chronic behaving cats induced by basal forebrain stimulation. The Physiological Basis of Mental Activity. R. Hernindez-Pe6n, Editor. Amsterdam, Elsevier. Suppl No. 24, Electroenceph. elin. Neurophysiol., 172-181. CLEMENTE, C. D., STERMAN, M. B., AND WYRWICKA, W., (1963); Forebrain inhibitory mechanisms: Conditioning of basal forebrain induced EEG synchronization and sleep. Exp. Neurol., 7,404-417. R. S., (1942a); The production of rhythmically recurrent cortical DEMPSEY, E. W., AND MORISON, potentials after localized thalamic stimulation. Amer. J. Physiol., 135, 293-300. DEMPSEY, E. W., AND MORISON,R. S., (1942b); The interaction of certain spontaneous and induced cortical potentials. Amer. J. Physiol., 135, 301-308. R. S., (1943); Theelectrical activity of a thalamocortical relay system. DEMPSEY, E. w., AND MORISON, Amer. J. Physiol., 138,283-296. FRENCH, J. D., AND MAGOUN, H. W., (1952); Effects of chronic lesions in central cephalic brain stem of monkeys. Arch. Neurol. Psychiat. (Chic.), 68, 591-604. HESS,W. R., (1954); Diencephalon, Autonomic and Extrapyramidal Functions. New York, Grune and Stratton, 79 pp. HESS,W. R., (1957); The Functional Organization ofthe Diencephalon. New York, Grune and Stratton, 180 pp. H. W., (1949); Effect upon the EEG of acute injury LINDSLEY, D. B., BOWDEN, J. W., AND MAGOUN, to the brain stem activating system. Electroenceph. d i n . Neurophysiol., 1, 475-486. MAGNUS,J., MORUZZI,G., AND POMPEIANO, O., (1960); Electroencephalogram-synchronizing structures in the lower brain stem. The Nature of Sleep. G. E. W. Wolstenholme and M. O'Connor, Editors. Boston, Little, Brown and Co., pp. 57-85. MONNIER, M., HOSLI,L., AND KRUPP,P., (1963); Moderating and activating systems in medio-central thalamus and reticular formation. The Physiological Basis of Mental Activity. R. Hemhdez-Pebn, Editor. Amsterdam, Elsevier. Suppl. No. 24, Electroenceph. clin. Neurophysiol., 97-1 12. MORISON, R. S., FINLEY, K. H., AND L~THROP, G. N., (1943); Influence of basal forebrain areas on the electrocorticogram. Amer. J. Physiol., 139, 410-416. MORUZZI, G., AND MAGOUN, H. W., (1949); Brainstem reticular formation and activation of the EEG. Electroenceph. elin. Neurophysiol., 1,455473, O., AND SWEXT,J. E., (1961); Cutaneous nerve fibers inducing EEG synchronization in POMPEIANO, normal unrestrained cats. Proceedings Fifth International Congress of Electroencephalography and Clinical Neurophysiology, Rome. Excerpta med. (Amst.), 37, 213-214. M. B., AND CLEMENTE, C. D., (1962a); Forebrain inhibitory mechanisms: Cortical synSTERMAN, chronization induced by basal forebrain stimulation. Exp. Neurol., 6, 91-102. C. D., (1962b); Forebrain inhibitory mechanisms: Sleep patterns STERMAN, M. B., AND CLEMENTE, induced by basal forebrain stimulation in the behaving cat. Exp. Neurol., 6, 103-117.
48
Limbic System and Free Behavior Jose M. R. DELGADO Department of Physiology, Yale Univers ity School of Medicine, New Haven, Conn. (U.S.A.)
Our knowledge of the limbic system has expanded considerably in the last decade due in part to bridging efforts initiated by physiologists and psychologists to further the correlation of electrical and chemical data with behavioral performance and psychological manifestations (see bibliography in Koikegami, 1964). Crossing the psychophysiological barrier required a considerable evolution of the scientific attitude, because investigatorswho were accustomed to the precise quantificationof phenomena in milliseconds and microvolts were dealing with a new group of responses which were so complex that they defied instrumental recording and required interpretation by trained observers. In addition, the phenomena were expressed in terms like fear, anxiety, pleasure and motivation which were often obscure, controversial and speculative. In spite of semantic difficulties, experimental evidence showed that in animals, limbic stimulation was able to induce learning, conditioning, escape, aggression, sexual activity, and other phenomena (Sheer, 1961), and in humans, electrical stimulation of discrete areas of the brain produced psychological manifestations such as fear, hostility, pleasure, friendliness, familiarity, and memories (Ramey and O’Doherty, 1960). Progress in electron microscopy, in microelectrode methodology, and in microchemistry has led to important discoveries concerning the submicroscopic properties of membranes, synapses and transmitters which certainly are essential for our understanding of behavioral mechanisms, but we should remember that the neuron is only a fraction of the whole organism, and that it is necessary to investigate - and to correlate - both sides of the scale: the microphysiology and also the physiology of the behaving brain, with its complex electrical, chemical and thermal reactions which are manifested as movements, emotions, worries and pleasures. The study of autonomic effects, conditioning and instrumental responses has been very active in neurophysiology and in physiological psychology, while the investigation of the neural basis of spontaneous activities and social behavior has usually been neglected, probably because of the difficulties inherent in recordingdata and introducing experimental variables without disturbing the phenomena under analysis. Implanted electrodes were essential for the study of the waking brain and have been widely used during the last few years, but the leads connecting animals and instruments handicapped the full expression of behavior and could not be used for chronic excitations or social behavior studies. Fortunately, advances in miniaturization, in transis-
L I M B I C SYSTEM A N D FREE B E H A V I O R
49
tors, and in integrated circuits have permitted the construction of telestimulators and teletransmitters small enough to be carried by animals like cats or monkeys. Several methods for remote stimulation of the brain have been described (see summary in Delgado, 1963b), but there is still a lag between physical and biological applications of modern electronics. It is surprising that in this age of guided missiles and radio contact with remote galaxies, the instruments at the disposal of the behavioral scientist are relatively bulky and primitive, and that very few authors are actually investigating the behavior of free animals under radio control. At present, several laboratories are working with or developing remote control techniques for the following purposes (see discussion and bibliography in Delgado, 1964; Robinson et ul., 1965; and Slater, 1963): (1) electrical stimulation of the brain; (2) injection of chemical substances and/or intracerebral perfusion of fluids; (3) recording of electrical activity - including single units - temperature, pressure, pH and other data; and (4) recording of individual and social behavior by time-lapse photography, and recording of vocalization by magnetic tape. In the near future, we may expect the establishment and expansion of physical and chemical two way communication systems between instrumentation and free experimental subjects, and I hope that the results presented in this article will encourage other authors to enter into this field. The purposes of this paper are: (a) to describe methodology and some recent developments for radio stimulation studies ; (b) to present experimental material obtained in monkey colonies -and also in monkeys under restraint -during electrical stimulation of the hypothalamus, fimbria of the fornix and amygdala; and (c) to discuss the physiological and psychological implications of these investigations.
METHODS A N D MATERIALS
The experiments were performed in a permanently established colony of monkeys (Mucuca mulutta) where brain stimulations, conditioning procedures and behavioral recordings were carried out without touching, restraining or disturbing the animals. Social relations in the group were stable and the behavioral profile of each monkey was detemined by time-lapse photography. The colony was located in a sound proof, air-conditioned room, and was formed by 4 to 6 animals, usually 2 males and 2-4 females. Two or three of them were equipped with intracerebral electrodes and were exchanged for new animals when their study had been completed, a process which usually took several weeks. Experiments were performed in a total of 13 monkeys, 5 males and 8 females of 3.1 to 4.8 kg. The electrodes were stereotaxicallyimplanted in the brain under anesthesia and with aseptic precautions, according to procedures previously described (Delgado, 1961). One or two weeks after surgery, the monkey was placed in a chair to record the spontaneous electrical activity of the brain and to establish the thresholds for radio stimulations which were monitored on a dual beam oscilloscope to determine simultaneously voltage and milliamperage. The monkey was then returned to the living quarters. References p. 66-68
50
DELGADO
Other experiments which required the precise measurement of pupillary size were performed with the animals restrained in Foringer chairs. Radio stimulation. The basic characteristics to be considered in instrumentation for radio stimulation are : reliability, sensitivity and working distance, size and weight. battery life, remote control of parameters of stimulation, remote selection of cerebral points, strength to stand abuse, and the prodecure for attaching the instrument and leads to the animal. Obviously, the most important requirement is reliability. If the output of the stimulator depends on the intensity of the received signals, or if the receiving antenna is affected by directionality, as happened in some of the circuits published by other authors, the performance will not be satisfactory (see discussion and bibliography in Delgado, 1963b). In the present investigations, I have employed a very simple procedure based on the Vanguard radio receiver for model airplanes (F&M Electronics, Attica, Ohio), which can be purchased in many hobby stores. The receiver is inexpensive, rugged and reliable, and its only function is to close a circuit to activate a tiny 2 transistor stimulator mounted on a deck inside the receiver box (see Delgado, 1959a, 1963b). As the instrument is crystal controlled, several channels may be used to stimulate different animals. The advantages of this system are reliability, availability, simplicity and economy; the main limitations are that stimulation parameters and selection of cerebral points are manually set and cannot be remotely changed. The characteristics of the stimulations used in the present study were : monopolar, cathodal, exponential falling pulses of 0.5 msec duration, 100 c/s, and intensity adjustable manually. In general, 0.5-2.9 mA were used. The receiver was attached to a harness on the animal's back and through subcutaneous leads was connected with any desired contact in the terminal sockets anchored to the skull. Following the natural tendency to reduce size and weight while increasing functional versatility, in collaboration with Mr. Per Hals we have recently completed the development of a three channel radio stimulator which permits the remote control of pulse duration, frequency and intensity of stimulation independently in each channel. Current consumption of the instrument is 1.5 mA, its weight is 22 g and its size is 30 x 25 x 15 mm. The principle of its design is to use a high frequency carrier (100 Mc) with three independent subcarriers which control the stimulation parameters regardless of possible changes in the strength of the received signal. The ouput is constant current in order to avoid fluctuations related to changes in tissue impedance. Time-Iapsephotography. Scientific investigation is based on the recording, analysis and quantification of phenomena, and one of the greatest problems in the study of behavior is that most of the spontaneous and evoked reactions defy instrumental analysis. Autonomic responses like blood pressure or heart rate can be recorded by the appropriate sensors and amplifiers, but fear, rage, or even playing and fighting are complex manifestationsdifficult to describe and quantify. For this reason, most of the reports on behavioral effects evoked by brain stimulation have been descriptive or even anecdotal. In the study of behavior, there is still no substitute for the human observer, and the only choices are when and how the animal is going to be observed, and what method the investigator will select to register his own visual perceptions. Classical studies involving lever pressing, cup lifting and other instrumental responses
L I M B I C SYSTEM A N D FREE B E H A V I O R
51
have been very productive, but in their normal lives, animals do not use such instruments, and the study of their spontaneous behavior requires other methodology like direct observation or photography. Cinemanalysis has been used successfully by several authors for the recording of spontaneous and evoked behavioral reactions (see bibliography in Delgado, 1964). The method that we have developed consists of a 16 mm motion picture camera which automatically records the activities of a monkey colony, taking 1 picture every 2 sec for 8 h daily. The films are later analyzed with the aid of a time and motion study projector, an IMB output typewriter, and a bank of 24 electronic counters. Individual and social behavioral categories are quantified, and profiles of activity, including range of variability, are established (Delgado, 1962, 1964). Experimental design. On stimulation days, experiments were recorded by the usual time-lapse photography, and began with a 1-h control period of spontaneous activities in the colony, prior to testing of any monkey. Then, for 30 min a tone sounded every minute for 7 sec. During the next hour, in one monkey a preselected cerebral point was radio stimulated every minute for 5 sec, starting 2 sec after the tone. Thus, if the evoked effect was conditioned, the animal began to respond during the 2 sec of tone before the stimulation started. After the stimulation hour, the tone alone was continued for another half hour for extinction purposes. Then it was turned off, and while the monkey was radio stimulated again for 5 sec every minute, a small piece of banana was offered to the colony and thrown inside the cage. The test was repeated 5 times, followed by 5 stimulations during which the colony was threatened at either side with the leather catching gloves. Finally, the stimulation was applied continuously for 5-1 0 min or longer, to determine fatigability of the evoked responses. The experiment ended with another 1-h control period. This procedure was repeated for each cerebral point at least on two different days. Direct observation of the colony supplemented the photographic recording and was especially important for quantification of evoked vocalizations. After completion of the studies, the tested monkey was sacrificed under anesthesia, the brain was perfused with formalin and cut stereotaxically in 10 mm blocks. After several days of fixation, frozen sections 50 p thick were prepared and stained with the Kliiver method, to be analyzed histologically. In agreement with previous experience, microscopic analysis did not reveal any specific lesions or degeneration at stimulation sites. Explanation of the tables of results. In Tables I, 111, IV, V and VI, 'N' indicates the number of monkeys in which the same cerebral structure was studied. Stimulation effects are listed chronologically in order of appearance, although some overlapping occurred. Reliability indicates the percentage of times that each response occurred in a minimum of 120 stimulations. In Table 111, results obtained in each monkey are shown in separate columns, while in the other tables, the evoked responses were less complex and could be expressed in one single column with the mean value of all animals. After-effects refer to manifestations observed during the first 10 sec following stimulation. Sensory reactivity describes whether the normal reaction of the monkey when offered food or threatened was modified during brain stimulation. Modification References p. 66-68
52
DELGADO
of stimulation effects indicates whether stimulation effects were inhibited or changed by offering food or threatening the colony. Fatigability, conditioning and extinction are self-explanatory. RESULTS
Lateral hypothalamus. Stimulation of the inferior part of the lateral hypothalamus with intensities of 0.3 to 0.9 mA produced constriction of the ipsilateral pupil. Five monkeys were radio stimulated in the colony with intensities adjusted to produce maximum pupillary constriction (1.8 to 2.9 mA), and results are summarized in Table I. In all 5 TABLE I Location: Lateral hypothalamus (inferior part) Stimulation effects 1. Constriction of ipsilateral pupil 2. Convergent strabismus 3. Moving head 4. Blinking eyes
N: 5 Reliability in % 100 100
6 12
After-effects None Sensory reactivity To food: No change To threat: No change
Modification of stimulation effects By food: No change By threat: No change
Social relations: Not modified Fatigability: None for pupillary constriction which persisted as long as stimulation was applied. (Tested for 21 days). Conditioning: None after 240 trials.
animals, the pupillary response was 100% reliable, and was not affected by the spontaneous activities of the animals, or by testing them with food or threat. Brain stimulation did not modify sensory reactivity, social relations or normal behavior like walking, eating and grooming. However, in a small percentage of stimulations, the monkeys moved or shook their heads slightly and blinked their eyes, suggesting a visual difficulty, probably related to the evoked strabismus. The paucity of behavioral effects contrasted with the wealth of manifestations obtained when other hypothalamic points were studied. Pupillary constriction did not fatigue, and persisted for as long as stimulation was applied, which in one experiment in the colony was prolonged for 3 days. A rebound effect appeared after cessation of stimulation in this animal, in which the ipsilateral pupil dilated and remained 10-20% larger than the other for 2-3 days. No behavioral changes could be related to this lasting excitation of the hypothalamus.
53
L I M B I C SYSTEM A N D FREE B E H A V I O R
To quantify the pupillary effects in greater detail, three of the monkeys were temporarily restrained in Foringer chairs and their pupil size was accurately measured in different experimental conditions by means of serial photography. One of the monkeys had another contact in the upper part of the lateral hypothalamus, stimulation of which produced ipsilateral dilatation of the pupil, and it was possible to confirm our previous report (Delgado, 1959b),that simultaneous stimulation of both constriction and dilatation points may cancel each other at different strengths of stimulation, demonstrating an artificially evoked dynamic equilibrium which could be modified in either direction, thus pupillary dilatation, for example, could be augmented by increasing the stimulation intensity of the dilator point, or by decreasing the stimulation intensity of the TABLE I1 EFFECTS O N I P S I L A T E R A L DIAMETER OF T H E P U P I L D E T E R M I N E D B Y E L E C T R I C A L S T I M U L A TION O F T H E L A T E R A L H Y P O T H A L A M U S
Light (candle foot2)
Electrical stimulation (m4
Pupil size (mm)
0.0
0.0
7.35 7.40 5.80 3.80 3.40 2.80 2.42 2.10
0.3
0.5
0.7 0.9 1.1 1.3 1.6
0.7 0.9 1.1 1.3 1.6
6.20 6.40 5.80 3.70 2.84 2.72 2.38 2.10
3.2
0.0 0.3 0.5 0.7 0.9 1.1 1.3 1.6
4.82 5.20 4.46 4.50 2.80 2.45 2.38 2.10
25.0
0.0
4.20 4.00 3.97 3.30 2.72 2.25 2.20 2.10
0.4
0.0
0.3 0.5
References p . 66-68
0.3 0.5 0.7 0.9 1.1 1.3 1.6
54
DELGADO
constrictor point. Functional equilibrium with synergic and antagonistic effects is probably one of the basic mechanisms for the regulation of physiological responses, and this problem was investigated further by measuring the size of the ipsilateral pupil related to two variables : illumination of the eye and intensity of electrical stimulation of the hypothalamus. Results are shown in Table I1 and Fig. I, and clearly demonstrate
b0
!3
!s
!S !5
!6
!?
'1.0
!a
'10
!4 '50
.m*.
Vft'
Fig. 1 . Effect of light, electrical stimulation of the hypothalamusand summation of both on pupillary size. Ordinate:diameter of the pupil in mm. First abscissa: electrical stimulation of the hypothalamus in mA. Second abscissa: optical stimulation of the eye in candle foot2.
that illumination of the eye and electrical stimulation of the hypothalamus are interchangeable with respect to their effect on pupillary size, with the important qualifications that the effect of cerebral stimulation is linear with respect to the electrical intensity, while the effect of light is logarithmic. Both optical and hypothalamic stimulation could be applied indefinitely without fatigue of the evoked effect; both represented a functional bias establishing a level of pupillary size which could be modified by new factors like emotional reactions; and the two effects summated. Pupillary constriction induced by electrical stimulation of the hypothalamus was maintained in one experiment for 21 days, demonstrating the indefatigability of neurons, pathways and effectors responsible for the effect. A more detailed study of these pupillary findings will be published elsewhere (Delgado and Mir, 1966). Anterior hypothalamus: supraoptic region. In two monkeys, a male and a female, stimulation of the right side supraoptic region with intensities of 1.2 to 1.6 mA produced an immediate interruption of ongoing activity, and after 0.5-1 sec, in each case the animal turned its head quickly, looking to the left as if something had caught its atten-
55
L I M B I C SYSTEM AND FREE BEHAVIOR
tion. Then it began walking forward at a moderate speed of about 1 m/sec, climbed one of the cage walls, descended and walked back to the starting place. The entire behavioral sequence took place during the 5-sec stimulation period. In both monkeys, the evoked motor performance was perfectly well coordinated, and could not be distinguished from spontaneous activities. Neither the stimulated animal nor the rest of the colony appeared frightened or aggressive. In addition, 1/2 to 3/4 complete penile erection was observed in the male monkey, which lasted during the 5-sec stimulation and subsided 2-3 sec later. Immediately after stimulation, the animal stood on all fours, flipped his ears back and forth, uttered a short low growl and adopted an aggressive attitude, looking straight out of the cage, and moving his lips, mouth and tongue, very fast for 4-8 sec. The other colony monkeys did not react to this behavior, indicating that they did not feel threatened or annoyed. One or two seconds later, normal activities were resumed. During these experiments, 5 times, stimulation started when the male monkey happened to be mounting a female. For the first 1-2 sec, stimulation was ineffective, after which the male performed the evoked response of walking forward which was followed by a subtle aggressive attitude. The general characteristics of the evoked effects were consistent, as shown in Table 111,but the details varied slightly depending on the spontaneous activities of the colony at the moment of stimulation. For example, the TABLE I11 Location : Hypothalamus (right supraoptic region)
N: 2 Reliability in % 82 71 78 56 86 98 95 65 80 20 96
Stimulation effects 1. Interruption of activities 2. Looking briefly to left 3. Penile erection 4. Walking forward (speed 0.7-1.2 m/sec) 5. Climbingwall 6. Walking back to starting place After-eflects 7. Flipping ears 8. Low tone vocalization 9. Fast mouth movements 10. Looking straight with aggressive face 11. Standing on all fours 12. Sitting down and returning to normal Sensory reactivity To food : Slight change To threat: No change
92 42 98 82 78 100
Modification of stimulation eflects By food : Diminished By threat: Diminished
Social relations: Not modified Fatigability: None after 1 min, partial after 10 min Conditioning: None after 150 trials References p . 66-68
84 54 70 45 41 100
56
DELGADO
monkey’s walking path around the cage floor was usually similar, but varied if his starting point changed or if other animals were in the way. When bananas were offered simultaneously with hypothalamic stimulation, the monkey was properly oriented and competed with other animals to take the food although he did not eat until stimulation ceased. This effect was clearly demonstrated in one experiment in which he was stimulated for 30 sec while offered a banana. He grabbed the fruit and started walking around the cage without dropping it, but without eating. As soon as stimulation ceased, he ate the banana voraciously. Threatening the colony produced an escape reaction which was not modified by radio stimulation, and during this escape, cerebral excitation was not effective. If stimulation was applied continuously for 10 min, the animal continued walking around the cage at approximately the same speed for 1-3 min, and then periods of 10-20 sec of walking and climbing alternated with other periods of 10-40 sec of sitting down. Flipping ears, vocalization and aggressive attitude were not observed during or after the continuous stimulation. Posterior hypothalamic nucleus: right side. This area was investigated in only one female monkey. Stimulation with 0.9 mA interrupted ongoing activities, the animal turned her head to the left and took one turn on 4 or 2 feet in less than 1/4 of the cage floor. Then she touched the mesh of the cage wall with 1 or 2 hands, without grabbing it, and looked through the feeding tube for about 1 sec. As an after-effect, the animal TABLE lV Location: Posterior hypothalamic nucleus (right side) Stimulation effects 1. Interruption of activities 2. Head to left 3. One turn to left 4. Touching right side wall 5. Looking through feeding hole After-effects 6. Fast movements of lips and mouth 7. Sitting down 8. Waking back to starting place 9. Sitting down and returning to normal Sensory reactivity To food: Diminished To threat: Slightly diminished Social relations: Not modified Fatigability: Yes,in 20-30 sec Conditioning: In 20 trials Extinction: In 12-15 trials
N: 1 Reliability in % 100 96 91 86 82
96 96 95
100
Modification of stimulation effects By food: Changed By threat: Inhibited
57
L I M B I C SYSTEM A N D FREE B E H A V I O R
moved her lips and mouth very quickly, sat down for 3-10 sec, took 3-4 steps back to her starting place, sat down and resumed her normal behavior (Table IV). When food was offered simultaneously with brain stimulation, there was an interplay between both. The tendency to turn persisted, but in 7 out of 10 trials, the turn was not completed. Taking of food was delayed and hesitant, without actual ingestion until stimulation was over. Waving the gloves induced escape which inhibited the electrically evoked effects. Escape from the threatening gloves was only slightly modified by some tendency to circle to the left. Changes in social relations were not recorded, apart from their interruption during stimulation time. During continuous excitation of this hypothalamic point, the monkey circled to the left 4-6 times, touching the wall and looking through the feeding tube each time, and then walking slowed down and after 20-30 sec, the animal sat down, resuming spontaneous activities a few seconds later in spite of the continuation of the stimulation. Conditioning responses started after the 10th trial and were established with over 90 % reliability after the 20th trial. Without cerebral stimulation, a 7-sec tone interrupted spontaneous activities, changed the facial expression, increasing alertness and inducing walking a few steps without turns. Extinction was complete in 12-15 trials. Fimbria of the fornix. In three monkeys (1 male and 2 females), radio stimulation of the fimbria of the fornix at intensities of 0.5-0.8 mA induced an immediate and violent response which appeared without exception each time that the stimulation was applied. The animals ran from side to side in the cage at great speed (2-2.5 m/sec; see Fig. 2 and Table V) during all the stimulation time, climbing half way up the lateral cage TABLE V Location : Fimbria fornix
N: 3
Stimulation efeets 1. Immediate response: running from side to side at great speed (2-2.4 m/sec) 2. Running into other animals 3. Growling After-effects 4. Sitting down with reduced spontaneous activity Sensory reactivity To food: Absent To threat: Diminished
Conditioning: Yes, in 3-6 trials Extinction: After 2 0 4 0 trials (Residual effects last for days) References p. 66-68
7;
100 4 12
95
Modification of stimulation effects By food: No change By threat: Slight change
Social reiations: The whole colony retreats to cage walls Fatigability: Not in 10 min
Reliability in
58
DELGADO
sides to spring back with great energy. Coordination of movements was excellent, and orientation was well preserved. They were able to avoid clashes with the other colony animals, and when occasionally one of them was run over, there was no fight or aggression. After several trials, the whole floor of the cage was free for the racing monkey, while the others sat on the swings or hung in the corners. No aggressive acts
Fig. 2. (A) = Control. B, C and D
= running induced by radio stimulation of the fimbria of the fornix (see text and Table V).
L I M B I C SYSTEM A N D F R E E B E H A V I O R
59
were recorded in the group during these experiments. As soon as stimulation was over, the stimulated monkey sat down and its spontaneous activities and social relations were diminished for a few minutes. If food was offered during stimulation, the animal paid no attention to it and continued running around. Even threatening was ineffective. In one experiment, one side of the colony cage was opened, and the investigator waved
C
D References p . 6648
60
DE L G A D O
the catching gloves and made threatening gestures while the fimbria was stimulated. Running around was evoked but the monkey modified its path and tried to race in only the opposite half of the cage; 2 out of 6 times, the animal ran the full length of the cage, coming to about one foot from the investigator without attempting to bite him or escape through the open door. Continuous stimulation for 10 min produced unintermpted alternation of racing around the cage, jumping to the swings, displacing other animals, climbing the walls, walking, hiding behind other monkeys, uttering very short low toned vocalizations, and lying down on the floor for a few seconds. High pitched vocalizations, wincing, baring the teeth, fearful expression and other signs possibly related to pain perception were absent. After this long stimulationperiod there was a considerable decrease in spontaneous activity. Local excitability, however, was not modified and a 5-sec stimulation produced the usual effect of running around at great speed. In all three stimulated animals, conditioning was quickly established in only 3-6 trials, and as soon as the tone started, the monkey began to walk. In the absence of stimulation, the tone induced walking around the cage at a reduced speed of only 1-1.5 m/sec. The conditioned response was extinguished in 20-40 trials, but a residual effect persisted for weeks and the tone alone induced walking at the beginning of each session until it was again extinguished in a few trials. When a different point located in the internal capsule was stimulated 3-7 days later, the initial response in 2 of the 3 monkeys was to run for 1 or 2 sec, and then the electrically evoked effect of contraction of the contralateral forelimb took over, requiring 10-18 sessions to induce the new response without traces of fimbria conditioning. These results showed the persistence of the previous motor performance which appeared during stimulation of a completely different cerebral point evoking an unrelated effect. Basolateral amygdala. In each of two monkeys, when the left basolateral amygdala was radio stimulated with 0.8-1.1 mA, the animal crouched, flexing its limbs and lowering its body to a nearly horizontal position with right arm slightly extended and head turned 30-60 degrees, and remained in this position making chewing movements throughout the stimulation (see Table VI). When excitation ceased, the monkey rose up and sat down normally. After 15-20 stimulations, the animal spent over 30 % of the inter-trial time sitting quietly with eyes closed. It was easily roused by noises or movements of other monkeys in the colony, but would close its eyes again if there was no reason for alarm. This effect differed from normal sleeping posture because instead of bending the head and body forward, the monkey maintained them erect. Along with the animal's increased resting time there was a corresponding reduction in its spontaneous individual and social activities. During stimulations, it looked at but did not take food offered, indicating that its orienting reactions were adequate, while motor responses were inhibited or slower than usual. Threatening the colony induced general escape but the stimulated monkey was the last to run away and its movements were clumsy. Under continuous stimulation to test fatigability of this response, it was shown that chewing diminished in 30-40 sec and nearly disappeared after 10 min, while the mon-
61
LIMBIC S Y S T E M AND FREE BEHAVIOR
TABLE VI Location: Amygdala (left basolateral)
N: 2 Reliability in % 98 100 76 92
Stimulation effects 1. Crouching 2. Arrest reaction 3. Extension right arm 4. Head to right 5. Chewing
94
After-effects 6. Sitting down with reduced spontaneous activity 7. Sleep with eyes closed, head and body erect Sensory reactivity To food: Absent To threat: Diminished
92 64
Modification of stimulation eflects By food: No change By threat: Slight change
Social relations: Diminished Fatigability: Not in 10 min Conditioning: Criterion (90%)was not reached after 120 trials, but near the end of each stimulation period (60 trials), conditioned responses appeared in about half of the trials.
key remained motionless, crouching on the floor. Reliable conditioning could not be established, but after 20-30 pairings of sound and stimulation, the monkey often responded to the sound by interrupting spontaneous activities to initiate the crouching position. In these two stimulated animals, the behavioral effects of localized seizures were also explored. With the monkey under restraint, radio stimulation of the amygdala with 2.7 mA induced a 7-18 sec localized seizure which did not spread to the motor cortex or to the contralateral temporal lobe, and could be repeated with similar characteristics 9 out of 10 times, with radio stimulations 5 min apart. Then the experiment was repeated by applying 10 stimulations during 1 h with the monkey free in the colony. Results showed that these seizure-evoking stimulations increased the inhibitory effects which were prolonged 10-1 5 sec after stimulation ceased. The monkey crouched motionless but reacted adequately to strong sensory stimulation such as a loud noise or threat to the colony which provoked escape, showing that the animal was conscious and able to process and interpret sensory information in spite of its evident motor inhibition. For more than half of the inter-trial periods, the stimulated monkey appeared asleep with eyes closed but with head and body erect. Social relations of the stimulated animal were considerably reduced. References p . 66.58
62
DELGADO DISCUSSION
The experiments described in this paper were based on the following methodological innovations in behavioral studies: (a) The animals were stimulated by radio control and were completely free inside a colony, without trailing chains or cables. (b) The study was conducted in an established monkey colony and therefore transporting cages or special testing rooms were not used. Experiments, including conditioning, proceeded in the animals’ customary surroundings, avoiding visual and auditory premonitory cues. (c) Stimulationswere applied to one monkey while he was forming part of the group, This fact was essential for the study of social relations. Besides, the attitude of other animals gave important clues to understand the meaning of the evoked effects. (d) Each cerebral point was electrically stimulated at least 120times on two or more different days, and this systematic repetition permitted a precise analysis and quantification of the elements of the responses. (e) Time-lapse photography provided objective and permanent records of the experiments which could be analyzed in detail, compressing or expanding the temporal scale with time and motion study techniques. (f) Fatigability of brain responses was analyzed for up to 21 days of continuous stimulation. (g) The above-mentionedspecial methodology did not exclude the use of more conventional techniques, and the monkeys were also studied under restraint to record electrical activity, monitor stimulations, and perform other experiments. With respect to the results, the fact that pupillary effects may be evoked by electrical stimulation of the brain has been described by several investigators (Delgado, 1959b; Hess, 1954; Hodes and Magoun, 1942; Kaada, 1951; Koikegami, 1964; Magnus and Naquet, 1961; and Showers and Crosby, 1958), and pupillary reflexes have been used by others as models for the study of biological servo-mechanisms(Fernandez-Guardiola and Harmony, 1964; Stark, 1964). The new findings provided here are the enduring quality of the effect which did not fatigue after days of continuous stimulation; the correlation and summation of optical and electrical excitation; and the absence of behavioral repercussion. Hypothalamic indefatigability requires a continuous stream of impulses with liberation of transmitters at the successive synaptic junctions. Alternation of active neurons and coupling between liberation, destruction and synthesis of chemical mediators may explain the prolonged actions to which the organism must adapt in some economical way. Evidence of a lasting rebound when long term stimulations were discontinued revealed the persistence of the adjusting mechanisms. The enduring qualities of hypothalamic stimulations indicate that the classical concept of quick fatigability of the brain may be true for the motor cortex and posterior hypothalamus (Table IV) which do not respond after 5-20 sec of electrical stimulation, but does not apply to other structures like the amygdala, thalamus, hypothalamus, putamen and crus of the fornix, which are still active after minutes, hours and even days of continuous stimulation (see also Delgado, 1955, 1959b). The precise correlation and possible summation of optical and electrical excitations illustrated the physiological qualities of the latter, and permitted the establishment of an equivalence in pupillary effects between candle foot2 units of light reaching the optic receptors, and milliamperes of electrical current applied to the hypothalamus,
LIMBIC SYSTEM A N D FREE B E H A V I O R
63
these relations being logarithmic and linear, respectively. It was significant that electrical stimulation may change the local set point upon which other factors originating at the receptors or in other points of the brain will summate. Functional reactivity may thus be modulated at central levels, and this set point may be altered and investigated experimentally. In a different type of study, Hammel et al. (1963) have suggested the existence of a hypothalamic set point for temperature regulation which may be modified by sleep, wakefulness, skin temperature and other factors. A parallel may be drawn between pupillary effects and behavioral responses like rage and fear, because in both cases, (a) the cerebral mechanisms ale present but inactive until specific sensory stimulation is received; (b) the effects may be prolonged without fatigue; (c) the responses may be evoked by electrical stimulation of specific cerebral structures located not in a single nucleus but in several different regions; (d) the responses may be induced or inhibited by excitation of different areas of the brain, and if several of them are simultaneously activated, a state of dynamic equilibrium may be established (Delgado, 1959b; Egger and Flynn, 1963; see also Gellhorn, 1957). The role of the limbic system in emotional reactions has been well substantiated in the literature, but many aspects of this relation still remain controversial (see reviews by Akert, 1961 ;Delgado, 1964; Gloor, 1960; Green, 1960; and Koikegami, 1964), and as mentioned in the introduction, its role in free behavior is little known. Results presented in this paper indicate that the combination of cerebral radio stimulation and time-lapse recording of group activities is a useful approach for the systematic investigation of evoked behavior. If the same point of the brain is stimulated in the same animal many times and results are recorded and quantified, the effects may be studied in detail to analyze and to evaluate the reliability of the different behavioral components, as shown in Tables I to VI. Objectivity is thus considerably increased, the experiments may be checked independently by different investigators, and behavior may be expressed in mathematical terms. Several authors like Lilly (1958) have mentioned that cerebral stimulations in monkeys are rather labile: ‘The animal must be isolated from all distracting stimuli, even the very weak ones. Subsonic vibrations in the laboratory ...have destroyed the reproducibility of responses’. It is true that, as shown for example in Table 111, climbing the wall appeared in only 65 % of the stimulations of the supraoptic region, and vocalization occurred only 42 % of the times, but other effects like walking were 95-98 % reliable, and even the speed varied only within narrow limits. In my experience, most experiments of cerebral stimulation have been highly reliable even on different days and months, and stimulation of the same structure induced a similar behavioral pattern in different monkeys, as indicated in this paper and demonstrated also in previous publications (see, for example, Fig. 5 in Delgado, 1963a). Behavioral freedom during radio stimulation experiments permitted the development of evoked effects without restrictions, and demonstrated that electricity was able to induce movements and reactions with characteristics similar to those of voluntary activity. The evoked effects consisted of a sequence of actions, as shown in Tables 111, IV, V and VI, and had a recognizable and reliable pattern in space and in time. Similar phenomena have been observed previously when thalamic nuclei or red References p. 66-68
64
DELGADO
nucleus were stimulated, and have been described as sequential behavior (Delgado, 1964, 1965). The great complexity of these responses contrasts with the simplicity of electrical stimulation.Application of a train of 100c/s pulses cannot explain the smooth and well coordinated succession of different effects, and must be considered only as a trigger which activates mechanisms preexisting inside of the brain, just as the finger of the officerwho presses the button to launch a man into space is not responsible for or even aware of the complex succession of phenomena which he has triggered. To explain the preexistence and the coordination of the functional mechanisms, I have proposed a theory of fragmental organization of behavior (Delgado, 1964) which postulates that each behavioral category is formed by a series of motor fragments organized in space and time which have anatomical and physiological reality within the brain where they can be the subject of experimental analysis. In agreement with this theory, the data of Table 111 show that walking appeared with a reliability of 95-98 % during stimulation of the supraoptic region. Walking also appeared as an after-effect of stimulation of the posterior hypothalamic nucleus and was recorded other times as voluntary activity. The act of evoked walking appeared with a precise temporal correlation in the sequence of effects:during stimulation time but not as an after-effect, in Table 111, and only as an after-effectin Table IV. It is not conceivable that walking has a different organization related to its different purposes; for example, to take food, to go away from something unpleasant, or just to exercise; and it is more probable that walking has an anatomical and functional set represented in the motor cortex, cerebellum, basal ganglia and other structures which may be triggered for different purposes from several areas of the brain related to food intake, curiosity, nociception, etc. The interpretation of our experiments therefore is not that turning the head to the left, walking forward, flipping the ears, standing on all fours, and other manifestations are localized in the supraoptic region of the hypothalamus. We can say only that these effects were triggered from that area, while the performance depended on the activation of preformed mechanisms located probably in several different areas of the brain. In agreement with the patterned time sequence of amygdala responses described by Naquet (1953) and with electrographicstudies by Gloor (1955) about the need of a progressive neuronal recruitment to induce patterned sequences, we could conceive of the limbic structures as organizing centers which modulate the functions of other cerebral structures. The triggering area should be more directly responsible for the behavioral sequence and perhaps for the purpose of the sequence than for the motor performance. Isolated acts like flipping the ears can be evoked from a variety of cerebral locations, while sequences of behavior seem to be speciJic of determined structures. Mapping of brain functions has attracted considerable interest among investigators, and the most thoroughly studied structure, the motor cortex, is usually presented in textbooks full of labels precisely localizing the cortical representation of body musculature. The motor cortex, however, is more complex, and the same cortical point may give a different response if, for example, the proprioceptive information from the muscles is modified (Gellhorn, 1953). In a similar way, the limbic system should not be
LIMBIC SYSTEM AND F R E E BEHAVIOR
65
mapped with labels of single functions without remembering that many responses including sequential behavior -will only be expressed in free animals. The stimulated region cannot be accepted as a ‘center’ with determined motor outputs, but as a planning station which activates a variety of substations and organizes their functions in space and in time according to determined sequences, processing at the same time sensory information which interplays with the organization of the response. Fortunately, in spite of this tremendous functional complexity, the behavioral fragments have enough stability to be analyzed independently during the development of evoked responses. The attempts to analyze correlations between limbic system and behavior may therefore be oriented towards the investigation first of those cerebral areas from which a specific fragment of behavior may be evoked; and second, the behavioral sequences represented in determined structures. Our data show that stimulation of the lateral hypothalamus induced discrete pupillary responses without behavioral involvement, while complex sequences were evoked from the anterior and posterior nucleus of the hypothalamus. In agreement with MacLean and Ploog (1962), penile erection was elicited in one monkey by stimulation of the supraoptic region, but no sexual relations were involved, and the effect seemed to be an autonomic manifestation without behavioral meaning. To the contrary, the several fragments of threatening behavior evoked by supraoptic excitation were well integrated and were directed with predilection towards one of the colony monkeys which reacted with submissivegestureslike grimacing and looking away. The stimulated area was in the neighborhood of the preoptic region considered by Fernandez de Molina and Hunsperger (1962) and by Hunsperger (1959) as part of the rage-eliciting system in the cat. The difference was that in our experiments, threatening appeared only as an after-effect, suggesting a delayed recruiting with activation of some cerebral structures related to the stimulated point. It should be emphasized that these effects could not be conditioned. Results obtained during stimulation of the fimbria of the fornix were interesting because this structure has not been included by other authors as part of the offensivedefensive system to which it seems to belong. The response had 100% reliability in 3 different animals, and it was very easily conditioned, confirming some of our previous results (Delgado, 1955), and supporting the theory that posterior hippocampus and fimbria of the fornix may constitute part of the highest links in the central integration of nociceptive responses, even if in the present experiments high pitched vocalization was absent. SUMMARY
Our studies were conducted in 13 monkeys (Macaca mulatta) with implanted intracerebral electrodes. The animals were forming part of a permanently established colony inside of which they were stimulated by radio control, tested, and conditioned. Behavioral results were recorded by time-lapse photography and later analyzed and quantified. Each cerebral point was stimulated at least 120 times on two different days, permitting the establishment of percentages of reliability of the behavioral reactions. References p. 6668
66
DELGADO
Additional studies were performed with the monkeys restrained in chairs. Results were as follows: (1) Radio stimulation of the lateral hypothalamus produced ipsilateral pupillary constriction with only minor behavioral manifestations. A more detailed study under restraint proved that (a) pupillary constriction could be maintained for 21 days without fatigue; (b) pupillary effects determined by illumination of the eye and by electrical stimulation of the hypothalamus were similar, but the relation was logarithmic in the first case and linear in the second; (c) optical and electrical excitation could summate; (d) electrical stimulation of the hypothalamus modified the functional set point for pupillary reactions for as long as stimulation was applied. (2) Stimulation of anterior hypothalamus produced a complex and specific sequential response including walking, climbing, penile erection, flipping of ears, vocalization and other effects. This response was not conditioned and fatigued only partially after 10 min of continuous stimulation. (3) Stimulation of the posterior hypothalamus produced a different sequential response including turning, touching the walls, looking through a tube and other effects. These responses fatigued in 30 sec, could be conditioned in 20 trials and extinguished in 15 trials. (4) Stimulation of the fimbria of the fornix produced fast running around the cage which did not fatigue in 10 min, was conditioned in 6 trials and extinguished in 40. (5) Excitation of the basolateral amygdala produced crouching, inhibited spontaneous activities and increased sleeping time with the monkey in an abnormal posture. The effects did not fatigue in 10 min and were conditioned with difficulty. (6) Results support the theory of fragmental representation of behavior and suggest that the stimulated cerebral structures are not responsible for the behavioral performance, but for the organization of the temporospatial sequence of behavior. (7) The existence of planning stations with control over motor substations is postulated. The limbic system would be more involved in planning than in behavioral performance.
ACKNOWLEDGEMENTS
This investigation was supported in part by research grants from the United State s Public Health Service and the Office of Naval Research. The collaboration of Dr. Mir and Mr. Maxim in some of the experiments, and the editorial assistance of Mrs. Caroline S. Delgado are warmly acknowledged. REFERENCES AKERT,K., (1961); Diencephalon. Electrical stimulation of the Brain. D. E. Sheer, Editor. Austin (Texas), Univ. Texas Press, pp. 288-310. DELGADO, J. M. R., (1955); Cerebral structuresinvolved in transmission and elaboration of noxious stimulation. J. Neuronhvsiol.. _ . 18. 261-275. DELGADO,J. M. R., (1959a); A transistor timed stimulator. Electroenceph. elin. Neurophysiol., 11, 591-593. ~
I
L I M B I C SYSTEM A N D F R E E B E H A V I O R
67
DELGADO, J. M. R., (195913); Prolonged stimulation of brain in awake monkeys. J. Neurophysiol., 22, 458-475. DELGADO, J. M. R., (1961); Chronic implantation of intracerebral electrodes in animals. Electrical Stimulation of the Brain. D. E. Sheer, Editor. Austin (Texas), Univ. Texas Press, pp. 25-36. DELGADO,J. M. R., (1962); Pharmacological modifications of social behavior. Pharmacological Analysis of Central Nervous Action. W. D. M. Paton, Editor, Oxford, Pergamon Press, pp. 265-292. DELGADO, J. M. R., (1963a); Effect of brain stimulation on task-free situations. The Physiological Basis of Mental Activity. R.Hernhndez P d n , Editor. Electroenceph. elin. Neurophysiol., Suppl. 24, pp. 260-280. DELGADO, J. M. R., (196313); Telemetry and telestimulation of the brain. Bio-Telemetry. L. Slater, Editor. New York, Pergamon Press, pp. 231-249. DELGADO, J. M. R., (1964); Freebehavior and brainstimulation. International Review of Neurobiology, Vol. VI. C. C. Pfeiffer and J. R. Smythies, Editors. New York, Academic Press, pp. 3 4 9 4 9 . DELGADO, J. M. R., (1965); Sequential behavior repeatedly induced by red nucleus stimulation in free monkeys. Science, 148, 1361-1363. DELGADO, J. M. R., AND MIR, D., (1966); Infatigability of hypothalamic stimulation in monkeys. Neurology, 16,939-950. EGGER,M. D., AND FLY”, J. P., (1963); Effects of electrical stimulation of the amygdala on hypothalamically elicited attack behavior in cats. J. Neurophysiol., 26, 705-720. FERNANDEZ DE MOLINA, A., AND HUNSPERGER, R. W., (1962); Organization of the subcortical system governing defence and flight reactions in the cat. J. Physiol., 160, 200-213. FERNANDEZ-GUARDIOLA, A., AND HARMONY, T., (1964); Modulation of visual input by pupillary mechanisms. First Conference on Neurobiology. Feedback Systems Controlling Nervous Activity. A. Escobar, Editor. Mexico, SOC.mex. Cienc. fisiol., pp. 197-210. GELLHORN, E., (1953); Physiological Foundations of Neurology and Psychiatry. Minneapolis, Univ. Minnesota Press. GELLHORN, E., (1957); Autonomic Imbalance and the Hypothalamus. Minneapolis, Univ. Minnesota Press. GLQOR,P., (1955); Electrophysiological studies on the connections of the amygdaloid nucleus in the cat. Part 11:Theelectrophysiologicalpropertiesoftheamygdaloidprojectionsystem. Electroenceph. elin. Neurophysiol., 7 , 243-264. GLOOR,P., (1960); Amygdala. Handbook of Physiology,Vol. 11. J. Field, H. W. Magoun and V . E. Hall, Editors. Washington, D. C., Amer. Physiol. Soc., pp. 1395-1420. GREEN,J. D., (1960); The hippocampus. Handbook of Physiology, Vol. 11. J. Field, H. W. Magoun and V. E. Hall, Editors. Washington, D.C., h e r . Physiol. Soc., pp. 1373-1389. HAMMEL, H. T., JACKSON, D. C., STOLWUK, J., HARDY,J. D., AND STRQW, S. B., (1963); Temperature regulation by hypothalamic proportional control with an adjustable set point. J. appl. Physiol., 18, 1146-1 154. HESS,W. R., (1954); Diencephalon. Autonomic and Extrapyramidal Functions. New York, Grune and Stratton. HODES,R., AND MAGOUN, H. W., (1942); Autonomic responses to electrical stimulation of the f o r e brain and midbrain with special reference to the pupil. J. comp. Neurol., 76, 169-190. HUNSPERGER, R. W., (1959); Les representations central- des reactions affectives dans le cerveau anterieur et dans le tronc ckrkbral. Neuro-chirurgie, 5, 207-233. KAADA,B. R., (1951); Somato-motor, autonomic and electrocorticographic responses to electrical stimulation of ‘rhinencephalic’and other structures in primates, cat and dog. Actaphysiol. scand., 24, Suppl. 83. KOIKEGAMI, H., (1964); Amygdala and other related limbic structures; experimental studies on the anatomy and function. 11. Functional experiments. Acta med. biol., 12, 73-266. LILLY,J. C., (1958); Correlations between neurophysiological activity in the cortex and short-term behavior in the monkey. Biological and Biochemical Bases of Behavior. H. F. Harlow and C. N. Woolsey, Editors. Madison, Wis., Univ. Wisconsin Press, pp. 83-100. MACLEAN, P. D., AND PLOOG,D. W., (1962); Cerebral representation of penile erection. J. Neurophysiol., 25, 29-55. MAGNUS, O., AND NAQUET,R., (1961); Physiologie normale et pathologique de I’amygdale. Les Grandes Activitis du Rhinenciphale. Vol. II. Physiologie et Pathologie du Rhinenckphalz. Paris, Masson, pp. 191-221. NAQUET, R., (1953); Sur les fonctions du rhinencbphale d’aprh les resultats de sa stimulation chez le chat. Thesis. Marseille.
68
DELGADO
RAMEY,E. R., AND ODOHERTY, D. S., Editors, (1960); Electrical Studies on the Unanz-sthetizedBrain. New York, Hoever. ROBINSON, B. W., WARNER, H., AND ROSVOLD, H. E., (1965); Brain telestimulator with solar cell power supply. Science, 148,111 1-1 1 1 3. SHEER,D. E., Editor, (1961); Electrical Sfitnulationof the Brain. Austin (Texas), Univ. Texas Press. SHOWERS, M.J. C., AND CROSBY, E. C., (1958); Somatic and visceral responses from the cingulate gyrus. Neurology, 8, 561-565. SLATER, L., Editor, (1963); Bio-Telemetry. New York, Pergamon Press. STARK, L., (1964); Stability, oscillations, and noise in the human pupil servomechanism. First Conference on Neurobiology. FeecamCk Systems Controlling Nervous Activity. A. Escobar, Editor. Mexico, SOC.mex. Cienc. fisiol., pp. 65-88.
69
Mechanisms in the Limbic System Controlling Reproductive Functions of the Ovary with Special Reference to the Positive Feedback of Progestin to the Hippocampus M A S A Z U M I K A W A K A M I , K A T S U O SETO, EI TERASAWA, AND K A Z U C H I K A Y O S H I D A Second Department of Physiology, Yokohama University School of Medicine, Yokohama (Japan)
In the past, limbic structures were thought to be associated only with olfactory function, but it is now reliably known that they participate in the regulation of a number of autonomic functions, certain endocrine reactions, sexual behavior, reproduction, feeding, memory processes, and the extrapyramidal motor activities. This brief enumeration of the functions of the limbic system illustrates that it comprises highly complex structures. The mechanism by which limbic structures modulate the function of the hypothalamo-hypophysial system is undoubtedly neural in character. On the other hand, it is inferred that the sex hormone and gonadotrophin act on the excitabilities of the limbic structures. Hohlweg and co-workers (1932, 1937) first postulated the concept that feedback control between the anterior pituitary and the target organ is effected by the action of their hormones not on the adenohypophysis but on the nervous system. Although studies on the gonadotrophic functions have produced evidence that the feedback control exerted by the gonadal hormones affects the hypothalamus (Flerk6, 1954, 1962; Harris et al., 1958; Sawyer, 1959; Kawakami and Sawyer, 1959a, b; Michael, 1961; Gorski and Barraclough, 1962; Lisk, 1962), the mode of feedback action of the hippocampo-ovarian loop and the amygdalo-ovarian loop have as yet been scarcely considered. The present experiments were therefore designed to demonstrate the existence of positive feedback control of the hippocampus on ovarian progesterone formation and its output, as well as the existence of negative feedback control of the amygdala on ovarian progesterone formation and its output. MATERIAL AND METHODS
Sexually mature female New Zealand white rabbits (2.5-3.2 kg) were fed an artificial diet supplemented periodically with fresh greens and carrots, and were kept under natural lighting conditions. Before use the rabbits were isolated in individual cages for at least 3 weeks. References p . 100-102
70
KAWAKAMI
et al.
For stimulating, recording, and lesioning, permanent electrodes were stereotaxically placed in the hypothalamus and several parts of the limbic system as well as the hippocampus with the aid of a rabbit brain atlas (Sawyer et al., 1954). The electrode assembly consisted of a 9-place minute socket with connected bipolar concentric stainless steel insulated electrodes enclosed in acrylic resin. The electrode assembly was fixed directly to the skull with acrylic resin and additionally secured by cementing it to 4 stainless steel screws fixed in the skull. Stimulation consisted of monophasic square wave pulses delivered unilaterally with a Sanei stimulator (ES-103-Y) and an isolation unit for 30 min, 60 sec on and off, 250-280 PA, at 0.1 msec duration, 60 c/sec. Current flow was monitored with a radiopulse transformer and oscilloscope, and the EEG activity of the brain was recorded during both stimulation and interstimulationperiods on a Sanei electroencephalograph (EG-908) to detect whether or not the EEG seizure pattern was induced by the stimulation. EEG recordings were taken with a 9-channel EEG amplifier and inkwriter (TyFe G-900, Sanei Co.); and a 2-channel EEG analyzer (Type EA-201, Sanei Co.) was used to analyze frequency components contained in the EEG, and transcribe them directly on the EEG record. During the EEG recording sessions the rabbit was free to move around on the table in the shielded, sound-proof recording chamber, to explore, eat and copulate with a buck introduced for the occasion. In order to determine the location of possible nervous pathways from the hippocampus or the amygdala to the hypothalamus which might control pituitary gonadotrophic function, bilateral lesions in the dorsal fornix, the septum and the stria terminalis were made by electrocoagulation with radio frequency waves between a lesioning electrode tip as anode and an indifferent plate electrode in contact with the abdomen. The small lesions in the periventricular arcuate nucleus were produced electrolytically with a direct current of 4 mA for 30 sec. A 3-week period permitted recovery from cerebral edema before stimulation experiments were made. The implanted rabbits were primed with estradiol benzoate in oil (0.1 mg s.c.) for 2 days before stimulation for ensuring an estrous state after the 3-week recovery period. Electrical stimulation was delivered without anesthesia. Laparotomy was performed after local anesthesia of the abdominal muscle with 1 % novocaine, the ovarian venous and the auricular arterial blood was collected for progesterone assay, and unilateral ovariectomy was performed in order to measure progesterone, estradiol and estrone biosynthesis in the ovary under pentobarbital anesthesia. Forty-five minutes after stimulation the ovarian venous and auricular arterial blood was collected for 10 min. The ovary was removed from the body 7 min later, in order to determine the incorporative rates of progesterone and estrogen from l-[14Cacetate. Further, re-laparotomy of the remaining ovary was made under pentobarbital anesthesia some 48 h later in order to detect ruptured ovarian follicles. The ovarian tissues were incubated in 50 ml Erlenmeyer flasks; and 100 mg of the ovarian homogenate or slices, 5 ml of Krebs-Ringer phosphate buffer at pH 7.2, and 10 pmoles (0.3 pC) of 1-[14C]sodium acetate (Isotope Specialities, U.S.A.) were
L I M B I C SYSTEM A N D P R O G E S T E R O N E F E E D B A C K
71
added to each flask. The flasks were closed and shaken for 60 min at 37" in a Taiyo incubator. After incubation, the progesterone, estradiol and estrone in the reaction mixture were separated by the method of Set0 et al. (1964), and the total radioactivity in each of these fractions was measured by the method of Roberts et al. (1962). The progesterone in the blood was analyzed quantitatively by the methods of Set0 et al. (1964), that is, for separation and determination of the progesterone, estradiol and estrone, each of the carrier steroids or of the radioactive steroids (for internal standard) was added to the incubation medium of ovarian tissues or blood, and the mixture was extracted with ether. The residue of the ether extract was partitioned between ligroin (boiling range 65-1 lo") and 90 % aqueous methanol. The aqueous methanolic phase containing the polar lipid fraction was evaporated, and the steroids were extracted from the water phase with benzene. The phenolic steroid fraction was separated from the neutral fraction by shaking the benzene solution with 1 N sodium hydroxide. The neutral fraction was further purified on a silicic acid chromatographic column, and the steroids were being eluted, with benzene containing ethyl acetate. The neutral steroids were then separated on silicic acid thin layer chromatography in the ligroin/ propylene glycol system. The zone of the progesterone was eluted from the thin layer, and was acetylated with pyrdine and acetic anhydride. The acetylated progesterone was further purified by gas-liquid chromatography (argon ionizing detector, SE-30, 245"). The phenolic steroid fraction was separated by silicic acid thin layer chromatography in the toluene/propylene glycol system. The zones of estradiol and estrone were eluted from the thin layer, and were acetylated with pyridine and acetic anhydride. The acetylated estradiol and estrone were further purified by gas-liquid chromatography (argon ionizing detector, SE-30, 245"). Total radioactivity in progesterone, estradiol and estrone was measured by using 5 ml of a scintillation system in toluene containing 0.5 "/d PBD (phenylbiphenyloxadiazole) and 0.01 % POPOP (I ,4-bis-2(5-phenyloxazoly)-benzene). Counting was carried out in the liquid scintillation spectrometer. The ovarian progesterone output was estimated by the differences in amounts of progesterone between arterial and venous blood. In autopsy, brains were fixed in 10% formalin, and the precise location of electrode tips was determined histologically.
RESULTS
( A ) Feedback eflects of gonadal steroids upon the limbic system (1) Inverse alterations of EEG activities in the amygdala and the hippocampus in the course of estrus With respect to the amygdalar function on the ovarian endocrine activity, the participation of the amygdala in ovulation is best known from the fact that in the several species such as the rabbit, cat and rat, stimulation of the central or medial portion of the amygdaloid complex will induce ovulation (Bunn and Everett, 1951; Koikegami et al., 1954; Shealy and Peele, 1957) and bilateral interruption of the stria References p. 100402
72
KAWAKAMI
et al.
terminalis in the female rat elicits precocious ovarian development (Elwers and Critchlow, 1961). Little attention, however, has been paid to the influence of the hippocampus on endocrine regulation. The present experiment therefore was designed to investigate the alterations in the hippocampal and amygdalar activity at the stage of estrus (the stage of estrogen dominance over progesterone), ovulation and pregnancy (the stage of progesterone dominance over estrogen), i.e. the relationship between the EEG activity of these regions and sex hormone (estrogen and progesterone) in the blood. The results were as follows. In EEG records from the dorsal and ventral hippocampus the 4-13 c/sec regular rhythmic sinusoidal waves, which are usually increased during the arousal or alert state, decreased by 30-50% during the maximum stage of estrus against those at the pre-estrus stage, in spite of showing highly estrous behavior, and at the same time the 20-30 clsec components with low amplitude superimposedupon the above-mentioned characteristic regular rhythmic sinusoidal waves were decreased, while the 4-8 c/sec regular rhythmic sinusoidal waves increased by 25-35 %. A representative example of changes in the EEG patterns and their analysis of the frequency components in the hippocampus during the anestrous, estrous and postcopulatory stage is shown in Fig. 1A and B. The histograms in Fig. 1C indicate the average integrated values of 4 cases for each frequency band. After copulation-induced ovulation these EEG patterns were characterized by a marked domination of the rhythmic sinusoidal waves particularly during the first 2 or 3 h within 10 hours’ change of EEG waves, and their analyzed EEG showed an increase of 15-20 % in the 4-8 c/sec band value as well as a 10-15 % increase in the 8-1 3 c/sec band value during 2-10 h following ovulation. Almost no considerable alterations weie revealed in the amplitude of the other integrated frequency band values. With respect to EEG changes at the stage of post-coitus, the shift in amplitude of integrated values in the 8-13 c/sec band was in marked contrast to that during the estrus stage : the former increased while the latter decreased. Similar changes in the amplitude of the integrated band values were observed at midpregnancy; all the frequency components of EEG waves in the hippocampus, especially the integrated value in the 2-4 c/sec band, increased by 10-20 % during both the EEG spindle burst and the arousal stages. The hippocampus participates in a specific way in the electrical activity of the brain when the animals are stimulated in any manner to alert or arouse them, and during the arousal the shift in the frequency of regular sinusoidal waves from 4-8 c/sec to the 8-13 c/sec in response to much stronger afferent stimuli is usually observed. This activity presented by the hippocampus is simplest in the rabbit. On the other hand, a depression of rhythmic sinusoidal waves and a marked development of 6-waves with superimposed fast waves and sporadic high amplitude sharp spikes appeared after the moderate anesthetic treatment or at the drowsy condition. After extracting about 100 ml of heart blood, the 4-8 c/sec hippocampal 8-wave was replaced by the 2-4 c/sec slow wave. The fast wave decreased simultaneously with this change. Frequency analysis of the integrated values showed an abrupt rise in the 2-4
73
L I M B I C SYSTEM A N D P R O G E S T E R O N E F E E D B A C K
HIPPOCAMPUS
ESTRUS
ANESTRUS
I
lo
lo
I
AFTER COPULATION
3
AMYGDALA ANESTRUS
A
V
ESTRUS
w
w
AFTER COPULATION =
y
-
l
50UV
1 sec
B I/ 10
c
5
0
I=
2=
3=
4=
2-4cls 4 - 0ClS 13cls 13 - 2 O c l s
0
-
5= m - 3 0 y s
Fig. 1. (A) An example of EEG pattern, (B) frequency-analyzed values integrated in 10-secepoch, and (C) the mean and the standard deviation of average integrated values of 10 epochs (100 sec) of 4 rabbits. While the hippocampal EEG showed a diminution in the 4-8 c/sec component, the amygdalar EEG showed anincreasein2-4c/secand20-30c/seccomponentsat the estrous stage. After copulation the 4-8 c/sec frequency component of the hippocampus increased, and the 2&30 c/sec component of the amygdala decreased. It is interesting that a similar change t o the one in the hippocampal and amygdalar EEG after copulation could also be observed after ovulation induced by hippocampal stimulation.
c/sec component and a decrease in the 4-30 c/sec components. Thereafter the whole EEG pattern gradually decreased its amplitude(Kawakami and Uemura, unpublished). The frequent appearance of trains of high amplitude regular slow waves of 4-8 or 8-1 3 c/sec, therefore, can be considered to indicate the enhancement of hippocampal activity. Hence it is possible to presume that the electrical activity in the hippocampus is lowered at the estrous stage and raised at the post-coital stage. References
p. 100-102
74
KAWAKAMI
et al.
In the amygdala, with the progress of estrus, the integrated value in 2-8 c/sec frequency bands including the high amplitude spindle bursts showed an increase of 2035 % as compared with that during the period of anestrus, and also those of 13-30 c/sec frequency components distinguishably increased as compared with those during the anestrous stage. An example of changes in EEG amplitude and the average of their integrated band values of 4 tests for each frequency component in the amygdala are illustrated in Fig. 1C. One to 4 h after copulation-induced ovulation, the dominant fast waves intermixed with the spindle bursts of 2-8 clsec frequency components at the estrous stage were more markedly diminished than before copulation. In analysis of the frequency components the amplitude of the 20-30 c/sec integrated band value reduced itself by 20-35 % within 1 h after copulation, while the 2-8 c/sec components were much diminished, and 8 or 9 h afterwards rebounded to a higher level than that of pre-copulation. From these observations, it might be assumed that the level of activity in the amygdaloid complex was increased at the estrous stage and decreased within 1 h following copulation. This assumption is based on the observations that the EEG changes in the amygdala after moderate anesthetic treatment, or in the drowsy condition, consist of suppression of the fast waves and of the appearance of high amplitude irregulat slow waves and sporadic sharp spikes instead, and after extraction of 70-100 ml of blood from their heart the amygdalar irregular slow waves below 4 c/sec appeared with an increase in the integrated band value of the 2 4 c/sec components, and with a decrease in the amplitude of these slow waves. In other words, they consist of increased fast waves including the spindle bursts related to the olfactory extrinsic activity. On the basis of these results, it is possible to estimate that the activity level in the amygdala was elevated and the hippocampal activity was depressed during the period of estrus, whereas the former was lowered and the latter elevated after copulation. In other words, both the hippocampus and the amygdala have biphasic responses in their excitability. Furthermore there exists a ‘seesaw relationship’ of excitability in the amygdala and the hippocampus between the estrus and the post-coital stages. (2) Influence of progesterone or luteinizing hormone upon hippocampal and amygdalar EEG activity in non-estrogen primed ovariectomized rabbit The hippocampus presented an EEG pattern of sporadic large spikes and 3-4 cJsec irregular slow waves in the non-treated ovariectomized rabbits. These were intermixed with a train of transient sinusoidal waves which became dominant with the advance of estrus. As to the influence of progesterone, the frequency of the hippocampal 6-7 c/sec 8wave was replaced by that of 8-9 c/sec, and its amplitude increased by 20-30 ,uV 1 h through 7 h after subcutaneous injection of progesterone (5 mg). At the same time in the amygdala the extrinsic burst waves of 2.4 c/sec slow waves and 20-30c/secfast waves decreased in amplitude. An example is shown in Fig. 2A. In contrast with the EEG alteration under 5 mg of progesterone administration, 10 mg administration showed the effects of reducing the hippocampal waves both in
LIMBIC SYSTEM A N D PROGESTERONE FEEDBACK
75
Fig. 2. The EEG pattern before (A) and after (B) subcutaneous injection of 0.5 mg progesterone propionate; before (C) and after (D) intravenous injection of 0.5 unit LH; and before (E) and after (F) intravenous injection of 200 unit FSH in the estrous rabbit. DHPC = dorsal hippocampus; AMYG = amygdala.
amplitude and in frequency from about 9 h after injection, attaining its maximum value after 15 h, although little effects occurred in the amygdala. In regard to the influence of luteinizing hormone (LH Armour, Lot No. R377279), about 20 min after LH injection (0.3-0.5 units, i.v.), the intermittent appearance of trains of sinusoidal waves (6-8 c/sec) became more frequent and tended to be more durable. After this phase which lasted for 30-60 min, the hippocampal EEG activity again declined gradually. While the hippocampal EEG revealed this activated pattern, the medial amygdaloid EEG pattern was characterized by a decrease in moderate amplitude fast waves of 20-30 c/sec as seen in Fig. 2B. This diminution of fast activity was sustained for 60-90 min. These alterations in the height of frequency components and in the amplitude of References p . I00402
76
KAWAKAMI
et al.
EEG waves following administration of progesterone or luteinizing hormone (LH) were somewhat similar to those at the post-coital stage. ( 3 ) Changes o j evoked potentials in limbic-hypothalamic loop through the estrus cycle From a systematic study involving ablation of various portions of the nervous system (Fee and Parkes, 1930; Brooks, 1937, 1938; Goldstein, 1957; Sawyer, 1957; Bard and Macht, 1958), it might be assumed that the apparent ‘reflex’is not a reflex in the classical sense, but rather that the activation of the hypophysis results from the full constellation of sensory influx and emotional state of the animal. In the rabbit much of the brain, which even includes parts of the limbic cortex, fornix, septum, lateral amygdala, dorsal thalamus and the dorsal hypothalamus, could be removed without preventing ovulation. But the present experiments on inducement of ovulation, showed the powerful influence of hippocampal activation upon the periventricular arcuate nucleus (ARC). Furthermore, it was revealed that the changes in EEG activity of the limbic system are associated with alterations in estrogen and progesterone levels of the circulating blood throughout the estrous cycle and post-ovulation. On the other hand, nerve impulses that induce LH release in the ‘reflex ovulators’ seem to be conducted through numerous complicated nervous chains before their convergence to the ARC. It is probably safe to assume that the alterations in the hippocampal EEG activity more or less reflect neuroendocrine correlates of ARC feedback mechanism working with alteration in estrogen and progesterone levels of the circulating blood. It is, therefore, the purpose of this report to investigate some of the properties of responses in the posterior basal hypothalamus to excitation of the hippocampus or the amygdala in the course of estrus and ovulation. Experiments were performed on 35 ovariectomized adult rabbits with electrodes chronically implanted into various parts of the brain such as the hippocampus, amygdala and the basal hypothalamus. Concentric bipolar electrodes were used for stimulation, and monopolar electrodes for recording. The endocrinological criterion for progesterone administration following estrogen priming in the present experiments was as follows. A chronic effect of progesterone is primarily to suppress the pre-ovulatory release of luteinizing hormone. Short-term effects of some amounts of progesterone are, by contrast, excitatory for the central ovulation mechanisms. There are indications that estrogen and progesterone are synergistic in this respect, just as they are synergistic in the induction of estrous behavior (Sawyer and Everett, 1959). An injection of progesterone can greatly lower the neural threshold for triggering the hormone release in the estrous rabbit either by reflex stimulation or by electrical stimulation of the brain (Kawakami and Sawyer, 1959a; Kawakami et al., 1965,1966; Terasawa and Kawakami, 1965). Injection of estrogen alone can induce ovulation, but its full effect is less abrupt than that of progesterone and release of ovulating hormone takes place only after a priming period of 2 to 3 days in the adult rabbit. It was on the above basis that the application of a 2.0 mg progesterone injection was made to intensify the estrous state in the present experiment,
L I M B I C SYSTEM A N D P R O G E S T E R O N E F E E D B A C K
77
The evoked potentials in the ARC of the hypothalamus of the ovariectomized female rabbits were recorded at the anestrous and estrous stages, or after progesterone administration. The electrical stimulation was delivered from each of the 4parts: the dorsal hippocampus, the mediobasal nuclei of the amygdala, the dorsal root of the sacral segment and the contralateral sciatic nerve. The anestrous rabbit was intended to represent those castrated with no treatment. The estrus rabbit was produced by priming with estradiol benzoate (0.1 mg s.c., in oil) daily for 2 days, and injecting on the 3rd day with progesterone propionate (2 mg s.c., in oil). (a) Changes of evoked potentials in the periventricular arcuate nucleus (ARC) during anestrus and estrus. The evoked potential in the periventricular arcuate response elicited by stimulation of the dorsal hippocampus showed a biphasic negative-positive wave, the negative component having a latency of 10-12 msec, and the positive a latency of 48-50 msec at the anestrous stage. The amplitude of the negative wave of the evoked potential was between 320 and 400 pV, and that of the positive wave between 500 and 540pV, their duration being from 40 to 50 msec in the negative component, and from 120 to 140 msec in the positive component. The potential evoked in the ARC by stimulation of the mediobasal nuclei of the amygdaloid complex showed a biphasic wave, the negative component with a latency of 7-8 msec, amplitude of 790-820 pV and duration of 40-45 msec, and the positive component with a latency of 4 6 4 8 msec, amplitude of 1000-1020 pA and duration of 110-130 msec in most of the present experiments at the anestrous stage. In the progress of estrus after the second treatment with estrogen, the evoked potential in the ARC upon stimulation of the hippocampus was slightly inhibited. The animal exhibited strong estrus in 2 to 6 h after injection of 2 mg progesterone. The negative wave at the anestrous stage showed a decrease of 20-25 % in amplitude, and the positive wave a decrease of 60-80% as compared with that at the maximum stage of estrus. The inhibited negative and positive waves of the evoked potential gradually recovered and were even facilitated 14-17 h after progesterone. This rebound phenomenon was more relevant in the negative component. Twenty-four to 26 h after progesterone the evoked potential returned to the control level, as illustrated in Fig. 3. A similar inhibitory effect could be observed in the evoked potential of the ARC by stimulation of the dorsal root of the sacral segment during estrus. On the contrary, the potential evoked in the ARC by stimulation of the amygdaloid nuclei was facilitated in the estrous stage. During estrus the negative response showed an increase of 15-20 %, and the positive response an increase of 6O-100%, as shown in Fig. 3. But the rebound phenomenon, as observed in the ARC potential evoked by the hippocampal stimulation 14-1 7 h after progesterone, hardly occurred on amygdaloid stimulation. The biphasic long latency response in the ARC to sciatic stimulation is also seen to be facilitated during estrus as the response to hippocampal stimulation. In the lower part of Fig. 3 are shown graphically the changes with the progress of estrus in amplitude and latency of the ARC potential evoked by hippocampal or amygdalar stimulation. (b) InJuence of progesterone or luteinizing hormone without estrogen pretreatment References p . 100-102
78
KAWAKAMI
HPC
et al.
AMYG
a B C
D
F
G
H I
%
JlOOllV
20 msec
L I M B I C SYSTEM A N D P R O G E S T E R O N E F E E D B A C K
79
upon evoked responses in the ARC of the ovariectomized rabbit. After (single) injection of 5 mg progesterone without estrogen pre-treatment the evoked potential recorded from the ARC by stimulation of the dorsal hippocampus was strongly facilitated as compared with the castrated, non-treated anestrous rabbits. About 4 h after an injection of progesterone ( 5 mg, s.c.) in the anestrous stage the potential evoked in the ARC revealed a gradual shortening in latency and increasing in amplitude. Seven hours after administration of progesterone the amplitude attained its maximum of about 40 % as compared with the control. Sixteen hours after a single injection of progesterone response was slightly lower than the control level, and after 24 h returned to the preinjection level; an example of this effect is shown in Fig. 4. Facilitation of the evoked potential was similarly seen in the ARC response to the viscero-sensory afferent impulse, for example from the dorsal root of the sacral segment. The evoked potentials recorded from the ARC by stimulation of the amygdaloid nuclei fell markedly: the negative response decreased by some 35 %, the positive response by some 25 %. After a large amount of progesterone, as already reported, the potential evoked in the ARC by the somato-sensory afferent impulse, for example from the sciatic nerve, was inhibited. Furthermore, during estrus or after progesterone the potential evoked in the midbrain reticular formation manifested more or less the same tendency as the ARC. The effects of LH (0.3 units) intravenously injected into the non-treated ovariectomized rabbits, upon the ARC potential produced by a single stimulation of the hippocampus or the amygdala, were similar to the effects of progesterone (5 mg s.c.) when administered to non-treated ovariectomized rabbits, as illustrated in Fig. 4. That is, regarding hippocampal stimulation, facilitation of both negative and positive components lasting for 20 to 40 min after injection, and inhibition of both components after amygdalar stimulation. After administration of oxytocin (0.3 u. i.v.) the evoked potential recorded from the ARC or RF produced by stimulation of the hippocampus or the amygdala were inverted in estrus or by injection of a large amount of progesterone. These phenomena were similar to those of the cat (Kawakami et al., 1966). In short, it was observed that during estrogen dominance over progesterone, as at estrus, the hippocampal afferent to the ARC is inhibited while the amygdaloid afferent is facilitated: the opposite was observed when progesterone was predominant over estrogen, as in pregnancy; i.e. the hippocampal afferent to the ARC was facilitated and the amygdaloid afferent inhibited. The above-mentioned result seems to have some relation to the observation that during estrus the afferent impulse to the ARC nucleus from the sacral root, containing
Fig. 3. The alteration of ARC potentials elicited by a single shock stimulation to the hippocampus (left) and amygdala (right) with the progress of estrus (A-H), and its graphical manifestation in percentage in relation to the amplitude of the potential before estrogen priming (I (A) ) control . phase; (B) 24 h after 1st estrogen administration; (C) 24 h estrogen administration; and @) 2 h, (E) 4.5 h, (F) 10 h, (G) 16 h, and (H) 24 h after injection of progesterone. References p . 100-102
'A
HPC
AMYG
A
A
B
B
C
C
P
D
EP
E
F
F
'dB
F 20 msec
PROGESTERONE
JlOOJJV
LUTElNlZlNG HORMONE
20 msec
Fig. 4. The alteration of ARC potentials produced by a single shock stimulation to the hippocampus and the amygdala after estrogen and progesterone administration in ovariectomized rabbit. The left row shows the evoked potentials (A) in the control phase, and (B) 3 h, (C) 6 h, @) 10 h, (E) 16 h and (F) 24 h after progesterone administration. The right row shows the evoked potentials (A) in the control phase, and (B) 5 min, (C) 10 min, (D) 30 min, (E) 1 h and (F) 4 h after injection of luteinking hormone. Progesterone as well as LH has inverse effects on the ARC potentials elicited by hippocampal and amygdalar stimulation.
LIMBIC SYSTEM A N D P R O G E S T E R O N E F E E D B A C K
81
the visceral afferent, is inhibited, and the impulse to the ARC from the sciatic nerve, a somato-sensory afferent, is facilitated. On the contrary, during pregnancy the viscerosensory afferent impulse to the ARC is facilitated, and the somato-sensory impulse to the ARC is inhibited. (B) Progesterone and estrogenformation in the ovary and ovarian progesterone output by electrical stimulation of several parts of the limbic structures Results of unilateral electrical stimulation applied through chronically implanted concentric bipolar electrodes within the brain of the rabbit are summarized in Tables I to v. Considerable information on the neurohumoral mechanisms involved in ovulation has been derived from observations in both cyclic and reflex ovulators. The concept that the mammalian ovulatory process is initiated by an increase in the discharge of LH superimposed on a background activity of follicle stimulating hormone (FSH) was introduced by Hisaw (1947). This concept has been supported by evidence of many experiments, and the dynamic patterns of pituitary gonadotrophin release in response to the changes in excitability in several parts of the brain have been studied. The experimental results at present available do not allow a definite picture to be presented of the role of the hippocampus in modulation of gonadotrophin secretion. The present experiments were therefore undertaken to differentiate the effects of hippocampal excitability from that of the amygdala upon ovarian progesterone formation and ovarian progesterone ouput of the estrous rabbit. Ovulation or large hemorrhagic follicles as evidence of subovulatory gonadotrophic stimulation, in response to the 260-280 or 30MOO p A electrical stimulation of the hippocampus, were obtained in 20 out of 30 estrous rabbits. But the 130-150 p A electrical stimulation resulted in neither ovulation nor follicles in 4 out of 5 estrous rabbits, though one animal showed a large hemorrhagic follicle. Almost no significant changes in somato-motor responses were observed in 260-280 pA electrical current stimulation of this region except for transient pupil dilatation, occasional searching behavior, and gazing forward. The first or second inter-stimulus EEG patterns, as well as those during the stimulation, revealed localized brief seizure patterns or occasional propagated seizure waves to the amygdala and some other regions, as shown in Fig. 5. Furthermore, in some instances of 39WOO p A electrical current stimulation of several parts of the hippocampus, localized or generalized seizure patterns in EEG extended more markedly to the other areas during and after the first or second stimulation. The 60 c/sec, low current, electrical stimulation usually suppressed the hippocampal EEG activity, while the same frequency stimulation with much higher electrical current such as 260-280 ,uA rather elevated the hippocampal EEG activity. The activation of the hippocampalEEGpattern, and at the same time the suppression of the amygdalar EEG pattern, were observed after hippocampal stimulation which could induce ovulation, as shown in Fig. 6. A similar change in EEG pattern was observed after copulation-induced ovulation in an estrous rabbit as described previously. References p . I0&102
P
L
z
82
m
t't a[.
il i
KAWAKAMI
0
L I M B I C S Y S T E M A N D PROGESTERONE F E E D B A C K
83
Electrical stimulation: hippocampus, 60 c/s, 0.1 msec, 260 mA, 6Osec on and off Fig. 6. Activation of the hippocampal EEG and suppression of the amygdalar EEG after hippocampal stimulationin the same manner as in the induction of ovulation. A similar change in EEG pattern was observed to that seen after copulation-induced ovulation in an estrous rabbit, as shwon in Fig. 1.
With reference to the induction of ovulation in response to electrical stimulation, inappreciable differences of localization were observed between the dorsal hippocampus and the ventral hippocampus, as illustrated in Table I. When ovulation (with ruptured ovarian follicles) was induced by 260-280 p A electrical stimulation of the dorsal hippocampus (DHPC), progesterone labeling from 1-[14C]acetate by the ovary homogenates increased by 60% and the ovarian progesterone output increased by 70 % as compared with non-stimulated cases. Further, slightly increased [14C]estrogen (estradiol estrone) formation by the ovary homogenates was observed in the positive case of ovulation induced by DHPC stimulation. Therefore, the ovulation-positive example induced by DHPC stimulation showed a decrease of 29 % in the ratio of [Wlestrogen to [14C]progesteronefrom l-[14C]acetate (E/P ratio) compared with the control. Furthermore, in the rabbits in which the ovulation was not induced by DHPC stimulation, the progesterone formation from l-[W]acetate in the ovarian homogenates was increased by 31 % compared with that
+
References p . 100-102
00 P
TA BLE I EFFECT OF H I P P O C A M P A L S T I M U L A T I O N O N O V U L A T I O N A N D FORMATION O F PROGESTERONE A N D E S T R O G E N B Y T H E R A B B I T ’ S O V A R Y
+
The progesterone and estrogen (estradiol estrone) formation from 1-[14C]acetate is expressed as disintegration per min (dpm) per h per 100 mg tissue homogenate. HPC = hippocampus; DHPC = dorsal hippocampus; VHPC = ventrohippocampus; E/P = ratio of the estrogen formation to the progesterone formation; means that stimulation caused ovulation; - means that stimulation failed to induce either ovulation or large hemorrhagic follicles.
+
Formation
Site of stimulation
Current (PA)
Ovulation
zii::
[14C]Progesterone ( %)
Control
DHPC HPC VHPC
-
-
190-210
-
260-280
“4CI-
PCI-
20a-OH- 17-OH-
A-pregprogesnene-3-one terone
f%)
PCItedione
( %)
______
( %)
_____
[14C]. Estrogen ( %I
100 162 123
100 182 144
100
100
181 126
-
-
100 157 131 160 131
-
-
390-400
-
13 5
152 125
-
390-400
-
3 1
-
-
124
+ +
+
-
Ovarian progesterone output ( %)
~
5 5 5 4 4
i-
EIP ( %)
-
100 -
100
-
-
-
112 98
71 75
-
-
132 104
88 84
170 119 -
-
-
-
-
-
116
96
-
-
-
LIMBIC SYSTEM A N D PROGESTERONE FEEDBACK STIMULATED
SITES
IN
85
HIPPOCAMPUS OVULATION A PARTIAL
X NONE
Fig. 7.Transverse reconstruction of rabbit hippocampus indicating the extent of the area that resulted in ovulation when stimulated. Note: In many ovulated positive cases the tips of electrodes were situated at the stratum lucidum and radiatum and at the layer of the large pyramids; while in negative cases they were situated at the fascia dentata. Cross signs represent the areas where ovulation failed to occur on stimulation. Closed circles indicate the points where stimulation caused ovulation, and triangles the points where stimulation induced large hemorrhagic follicles. Alv. = alveus of the cornus; C.Am. = cornus ammonis; Fde. = fascia dentata; Sub.C.Am. = subiculum cornu ammonis.
in the control; the progesterone output from the ovary was likewise increased, the increment being 19 % (Fig. 7). Thus, electrical stimulation of the DHPC exerted an active influence on progesterone formation in the ovary whether ovulated or not, whereas there was little difference, if any, in estrogen formation in either condition. These results are summarized in Table I. The observations reported above show that hippocampal stimulation does influence the ovarian progesterone formation and the ovarian progesterone output. However, it is very dubious that these effects should really be ascribed to the function of the hippocampus itself, since it was also seen that hippocampal stimulation easily induces seizures in many brain structures. The possibility should be considered that the amygdaloid activation caused by hippocampal excitation is the actual cause of the ovulatory change seen in this experiment, since some parts of the amygdala are both anatomically and electrophysiologically connected to the ventromedial and the periventricular arcuate nuclei in the hypothalamus (Gloor, 1956; Green and Adey, 1956; Nauta, 1956, 1963; Crosby et al., 1962; Koikegami, 1963, 1964; and Ban, 1964). The following series of experiments employing lesioning procedures was made in order to exclude this possibility. By electrical stimulation of the DHPC after the bilateral lesioning of the stria terminalis, large hemorrhagic follicles were induced. Progesterone labeling from 1- [Wlacetate by the ovarian homogenates and ovarian progesterone output considerably increased, while labeled estrogen showed almost no increase ;in consequence, the E/P ratio markedly declined. This observation might imply that the effects of DHPC stimulation upon progesterone formation in the ovary still remain to a large References p. 100-102
86
KAWAKAMI
et a!.
extent after the lesioning of the stria terminalis; and such effects are produced upon the ovary by a way different from the stria terminalis. Four rabbits with bilateral massive lesions of the dorsal fornix, including adjacent regions, or of the septum failed to ovulate in response to 280 or 400 p A electrical stimulation of appropriate areas in the hippocampus. Fig. 8 shows a composite dia-
ME
Fig. 8. Schematic representation of lesion sites on cornal section of the dorsal fornix, stria terminalis and basal hypothalamus. Site of the fornical lesions in 5 rabbits in which ovulation failed to occur on hippocampal stimulation. The ovarian progesterone formation and its output were depressed after the fornical lesion, and the hippocampal stimulation no longer showed the facilitatory effect. ////////, inclusive area destroyed by lesions; -, area common to all lesions.
gram of the bilateral destruction produced by all dorsal fornix lesions associated with both decreased ovarian progesterone, estradiol, and estrone production, as well as their output. The hatching in the figure indicatesthe inclusive area destroyed by lesions and dark shading shows the area common to all lesions. In the rabbits with a fornix lesion, the amount of progesterone and estrogen incorporated from l-[14C]acetateby the ovarian homogenates was lower, as compared with the control, by electrical stimulation of the DHPC; the output of ovarian progesterone was decreased. This fact might suggest that excitation of the hippocampus directly activates the release of LH-releasing factor from the posterior basal hypothalamus. The possibility was excluded that excitation of the amygdala or some other regions had resulted from the spread of the hippocampal stimulation effect, since inhibition of progesterone biosynthesis was observed whereas there was no sign of ovulation when the fornix was destroyed. These results indicate that the facilitatory effects of DHPC stimulation upon ovarian progesterone formation are blocked by bilateral lesions of the dorsal fornix. This might be considered as proof that such effects are exerted upon the ovary through
LIMBIC SYSTEM A N D P R O G E S T E R O N E F E E D B A C K
87
the fornix. It might also be implied that the hippocampus plays a major role in the production of steroid hormones in the ovary, the mechanism of which needs further investigation. Among 11 rabbits with electrodes in several parts of the amygdala, positive results were obtained in 5 rabbits with electrode tips located in the medial amygdaloid nucleus. When ovulation was produced by 260-280 p A electrical stimulation of the intermediate nucleus of the amygdala, it was found that [14C]progesteroneformation from l-[14C]acetateby the ovary homogenates increased (as with 260-280 pA stimulation of the DHPC) by 42% as compared with the non-stimulated controls. A marked increase of 55 % was observed in the ovarian progesterone output. Hence the E/P ratio decreased by 14%. Further, where stimulation was delivered to the prepyriform area and large heniorrhagic ovarian follicles were confirmed, the [14C]progesterone labeling from 1-[14C]acetate by the ovary homogenates increased in amount far above the control level. Such an increase in the ovarian progesterone output is illustrated in Table 11. The incorporation rate of [14C]estrogenfrom l-[14C]acetateshowed that it was the same as the control. The E/P ratio, therefore, was markedly lowered against the control; it was also slightly larger than that of the positive results with ruptured ovarian follicles gained by DHPC stimulation, but smaller than that observed after stimulation of the intermediate principal nucleus of the amygdala. Positive results were also obtained in estrous rabbits that were electrically stimulated with 280 pA or a higher current (390-400 pA) at the periventricular arcuate nucleus and the posterior median eminence. Progesterone labeling from 1-[14C]acetateby the ovary homogenates was increased 42 % by 280-290 pA electrical stimulation of the arcuate nucleus over that observed without stimulation. The level of the ovarian progesterone output was 60 % higher and the E/P ratio was 24 % lower in the former than in the latter. These results are summarized in Table 111. The electrical stimulation of the dorsomedial and posterior hypothalamus or the pre-mammillary body elicited large hemorrhagic follicles in the ovary, and revealed a slight elevation of ovarian progesterone output and [14C]progesterone formation in the ovary. To make a comparison however, between the ovulation-positive results of 260280 p A DHPC stimulation and those by 260-280 pA stimulation of the intermediate nucleus of the AMYG: (1) the DHPC stimulation revealed an elevation of 13% as compared with the latter in formation of 1%-incorporated progesterone from 1-[14C]acetate by ovarian homogenates; (2) there was an increase of 10% in ovarian progesterone output, and (3) there was a reduction of 17% in the E/P ratio, as seen in Table IV. Further, progesterone labeling from l-[14C]acetate by the ovary homogenate and the ovarian progesterone output of the ovulation by 280 p A DHPC stimulation were higher than after ARC stimulation : 13and 6 % respectively, and the E/P ratio was 7 % lower. References p , 100-102
E F F E C T OF THE AMYGDALOID
(AMYG) STIMULATION
TABLE I1 O N OVULATION A N D FORMATION OF PROGESTERONE A N D ESTROGEN B Y RABBIT’S OVARY
f means that large hemorrhagic follicles were observed
Control Intermediate Principal Nucleus
-
-
5
260-280
f
+
-
+
loo*
4 2 1
loo* 142 131 115
116 116 100
loo*+ 86 90 89
100 155 132
-
Medial AMYG Principal Nucleus
260-280
-
1 1
138 122
103 117
80 94
123
Lateral Principal Nucleus
270-280
f
2
126
108
88
-
Pre-pyriform area
270-280
f
2
129
100
78
128
Abbrevations:AMYG: amygdala. * The percentage of average dpm progesterone or estrogen in stimulated cases to that of non-stimulatedcases. ** The percentage of E/P in stimulated cases to that of non-stimulatedcases.
%
$ f
EFFECT OF ARCUATE NUCLEUS
(ARC)
TABLE
m
STIMULATION O N OVULATION AND F O R M A T I O N O F PROGESTERONE A N D ESTROGEN B Y R A B B I T S '
P
OVARY
Formation Site of stimulation
Control ARC
Current ( P A )
Ovulation
No. of rabbits
[14C]progesterone ( %)
-
-
5
loo*
loo*
280-290
-
3 4 1
142 123
104 96
+ +
400
[14CIEstrogen ( %)
E/P ( %)
Ovarian progesterone output ( %)
loo**
100
76 80
160 113
>
2,
U
EFFECT OF
DHPC,
Site of stimulation
Control DHPC
AMYG D A L A A N D
ARC
Current ( P A )
-
TABLE I V
1
S TIMU LA TION U P O N T H E P R O D U C T I O N OF PROGESTERONE BY THE O V A R Y I N THE O V U L A T E D R A BBITS
No. of rabbits
14C-incorporated to progesterone ( %)
Progesterone output ( %)
5
loo*
loo**
100
260-280
4
160
71
390-400
13
152
88
-
170
ARC
280-290
3
142
76
160
Amygdala
260-280
4
142
86
155
m
B
0
5
h 3
m m
U
TABLE V EFFECT OF LESIONS OF THE STRIA TERMINALIS, P E R I V E N T R I C U L A R A R C U A T E N U C L E U S O F T H E H Y P O T H A L A M U S A N D T H E D O R S A L F O R N I X O N E L E C T R I C A L S T I M U L A T I O N OF T H E D O R S A L H I P P O C A M P U S
Formation Site cf lesion
Stria terminalis
DHPC
-
+
-
+
ARC Dorsal fomix Ventral fornix Septum
-
+
-
+
Current (PA)
280
-
280
-
ovu'ation
No. of rabbits 5
PCIProgesterone ( %)
100*
pc1-
Estrogen
EIP ( %)
( %!
loo*
loo**
Ovarian progesterone output ( %) 100
4
160
112
71
170
3
101
100
101
100
3
125
92
78
121
5
105
96
96
89
-
5
106
100
92
94
4
91
92
103
85
280 -
4
92
96
106
79
1
90
96
113
91
280 -
2
92
96
106
81
1
92
104
119
79
280
2
91
88
96
85
280
L I M B I C SYSTEM A N D P R O G E S T E R O N E F E E D B A C K
91
The 260-280 pA electrical stimulation of the DHPC, therefore, is different in the ovulated ovary from that of the amygdala (N.intermediate) and the ARC, with respect to progesterone labeling from 1-[l4C]acetate by the ovary homogenates, ovarian progesterone output, and the Iatio of E/P; while only a small difference in the same order of magnitude is found to exist between the latter two as seen in Table IV. After the arcuate nucleus was lesioned, electrical stimulation of the DHPC exerted no effective influence upon the progesterone and estrogen formation in the ovary or ovarian progesterone output. This fact might suggest that the effects of DHPC stimulation upon the ovary are produced through the arcuate nucleus. These results are summarized in Table V. Though it is beyond the limits of the present results, it might be postulated that marked differences in the E/P ratio between lesions of the fornix and the stria terminalis might be due to the lowered activity of the hippocampus caused by interruption of the afferent impulses passing through the dorsal fornix. From the above-mentioned results, it might be inferred that in the ovulated cases, the electrical stimulation of the DHPC exerts influences somewhat different from those of AMYG and ARC stimulation upon the biosynthesis of steroid hormones in the ovary, as well as upon ovarian progesterone output. As already described, by stimulating the hippocampus with more intensive current, 390-400 pA, a generalized seizure occurs in a more extensive area than by stimulating with 260-280 PA. It seems as though, after hippocampus stimulation, the activity of all the other areas also changes, i.e. the limbic system plainly reacts as a single unit. In the ovulated cases by DHPC stimulation with a much higher current of 390400 pA, progesterone labeling from l-[14C]acetate by the ovarian homogenates increased by 52 % as compared with the non-stimulated controls. This change was similar to that produced by 260-280 pA DHPC stimulation. However, the formation of 14C-incorporatedestrogen was higher than in DHPC stimulation by 260-280 pA ;the former revealed a 20 % increase in the E/P ratio as compared with the lattei. Therefore, the differences of the E/P ratio between the effect of highcurrent hippocampal stimulation and that of low current hippocampal stimulation might be due to the elevation of excitability in the amygdala, and in some regions closely related to the hippocampus, through the influences of high current 390-400 pA electrical stimulation of the DHPC, or due to the elevation or depression of excitability of the hippocampus itself by hippocampal stimulation with high electrical current. The main results of the experiment are summarized in Table VI. DISCUSSION
With reference to biosynthesis of ovarian steroid hormones, it has been reported that gonadotrophin has a stimulatory effect upon biosynthesis of ovarian sex steroids, and that the amount of ovarian progesterone fomration is closely related to the biological activity of luteinizing hormone (LH) (Hilliard et al., 1963, 1964; Hilliard and Sawyer, 1964; Hayward et al., 1964; Rice et al., 1964). As to the progesterone formationin the rabbit ovary, Verly (1951) found that the urinary excretion of pregnane-3a,20aReferences p . 100-102
TABLE VI
T H E I N C O R P O R A T I O N V A L U E S OF P R O G E S T E R O N E (A) A N D E S T R O G E N (B) F R O M T H E 1-[l4C] A C E T A T E A F T E R T H E H I P P O C A M P A L , A M Y G D A L A R O R THE A R C U A T E S T I M U L A T I O N I N I N T A C T A N I M A L S A N D A F T E R T H E H I P P O C A M P A L S T I M U L A T I O N I N A N I M A L S WITH B l L A T E R A L LESIONS EITHER I N T H E FORNIX
(FX) OR
I N T H E STRIA TERMINALIS
E
(ST)
TABLE VI A
DHPC stimulation Control Amygdala stimulation Control ARC stimulation Control DHPC stimulation Amygdala stimulation DHPC stimulation ARC stimulation Amygdala stimulation ARC stimulation DHPC stimulation after FX lesion Control DHPC stimulation after ST lesion Control DHPC stimulation after FX lesion DHPC stimulation DHPC stimulation after ST lesion DHPC stimulation DHPC stimulation after FX lesion DHPC stimulation after ST lesion
Ovulation
N
Mean
S.D.
+ + + + + + + + + -
4
149 93 132 93 132 93 149 132 149 132 132 132 79 93 115 93 79 149 115 149 79 115
f 10.4 f 2.0 f 1.9 f 2.0 f 1.7 f 2.0 f 10.4 f 1.9 f 10.9 f 1.7 f 1.9 f 1.7 f 6.8 f 2.0 f 3.8 f 2.0 f 6.8 f 10.4 f 3.8 f 10.4 f 6.8 f 3.8
tf:
-_
+ *+
-
f
5
4 5
3 5
4 4 4 3 4 3 4 5
3
5
4 5
3 4 4 3
Reliability of diferences * * * 0.05 0.05 0.05 9
E
0.05
0.05
NS 0.05 0.05 0.05 0.05 0.05 -
-
TABLE VI B Reliability of differences* * *
Ovulation
N
Mean
S.D.
DHPC stimulation Control
+
4 5
28 25
f 1.3 f 0.8
0.05
Amygdala stimulation Control
+ + + + + + + +
4 5
29 25
f 1.9 f 0.8
0.05
3 5
26 25
f 0.8 f 0.8
NS
4 4
28 29
f 1.3 f 1.9
NS
4 3
28 26
f 1.3 f 0.8
NS
4 3
29 26
f 1.9 f 0.8
0.05
DHPC stimulation after FX lesion Control
-
4 5
23 25
& 1.7 f 0.8
0.05
DHPC stimulation after ST lesion Control
zt
3 5
24 25
4 5
23 28
f 0.5 f 0.8 f 1.7 f 1.3
3 4
24 28
f 0.5 f 1.3
0.05
4 3
24 23
f 1.7 k 0.5
NS
ARC stimulation Control DHPC stimulation Amygdala stimulation DHPC stimulation ARC stimulation Amygdala stimulation ARC stimulation
DHPC stimulation after FX lesion DHPC stimulation DHPC stimulation after ST lesion DHPC stimulation DHPC stimulation after FX lesion DHPC stimulation after ST lesion Abbreviation:
-
+ f +
f
** * t-test for each pair indicated in the first column NS: statistically non-significant
0.05
NS
.. CA
1
m !a 0
5 w
9 cl
n
94
KAWAKAMI
et al.
diol increases after mating or stimulation by gonadotrophin, by the method of Hooker-Forbes Clanberg’s bioassay (Verly, 1951). Hilliard et al. (1963, 1964) found that directly after copulation, 10-25 min after injecting a sufficient amount of gonadotrophin to elicit ovulation [Armour LH (I.U. 0.3 mg), NIH-LH, NIH-FSH (1.O mg), PMS or HCG (25-250 I.U.)], the release of progestin into the blood markedly increases 1 to 2 h after the injection of gonadotrophin (LH-HCG) following lateral removal of the ovary; and the progestin in the contralateral ovary increases as compared with the extirpated ovary. It was concluded from these observations that the increased release of the progestin into blood is due to the increase in its production in the ovary. It was also observed in the present experiments, as illustrated in Figs. 9 and 10, that by intravenous administration of LH (Armour, Lot No. R377279) to the rabbit, progesterone labeling from l-[Wlacetate by the ovarian homogenates, the ovarian slices and the ovarian progesterone output increased in proportion to the administered dose of LH. Fig. 9 shows that after intravenous injection of LH the release of progesterone into the ovarian vein is in parallel with the increase in the progesterone formation from the 1-[14C] acetate in the ovarian homogenate. Ovulation induced by LH occurs at the dosage of 0.3 unit/kg; however an increase in progesterone formation, and hence the progesterone release, can be observed below this dose. The progesterone formation from the l-[14C]acetate homogenate in the ovarian homogenate, and the progesterone output both increase in proportion to the LH dosage. Theiefore, it might be considered that the LH increases the production of progesterone in the ovary, whether or not it elicits ovulation, thus increasing the progesterone release. Fig. 10 (left) indicates the result of LH addition in vitro which shows the increase of progesterone formation from the 1-[Wlacetate in the sliced ovary preparation. And the intramuscular injection of LH also causes a marked increase in progesterone formation from the 1-[14C]acetate in the ovarian slices, as shown in Fig. 10 (right). These observations seem to show that the LH secreted from the hypophysis acts directly on the ovary and facilitates progesterone production, consequently increasing the release of progesterone into the blood, and that this action occurs whether or not it elicits ovulation. On the other hand, this LH acts upon the hippocampus to enhance its excitability as already mentioned, thereby facilitating the release of LH from the pituitary. Ovulation inducement by electrical stimulation of the DHPC, AMYG and the ARC and its subsequent increased biosynthesis of progesterone in the ovary, might suggest that electrical excitation of appropriate sites within the brain produce increased LH release from the adenohypophysis. This explanation is in accord with the findings of Hilliard et al. (1964) and Hayward et al. (1964), who investigated the relation between electrical stimulation of amygdala, coital stimuli, and the biological activity of LH. AS described previously, there was a seesaw relationship of activity between the amygdala and the hippocampus throughout the period of the sexual cycle from the
95
LIMBIC SYSTEM A N D PROGESTERONE FEEDBACK 2
9 fa
. I 600
9
f z y
0
z 500
E
p a
L 0
400
f
E
w K
300
J
300
L
I
f
k
Y
g P
400
W
c, 0
U 500
~
/
200
I /
200
I
b
:
ioa
f
100
'x 600
x 1000
SO0 800
I
W
2 W
z k
600 LL
i
W 0
3
400
a
i
400
f
0
a 0 W
500
200
200 100 0
100
200
100 0
IrE
100
200
300
400
500 IrB
DOSAGE
DOSAGE
Fig. 10. Effect of LH upon incorporation of [14C]acetateinto progesterone by ovarian slices. Left figure shows the changes of incorporation of [Wlacetate into progesterone when L H was added into incubation flask (in vitro). Right figure shows changes when LH was added by intramuscular injection (in vivo).
References p . 100-102
96
KAWAKAMI
et al.
viewpoint of the changes in EEG waves and EEG frequency components. In other words, the activity of the hippocampus decreased at the stage of estrogen dominance over progesterone (estrous stage), and increased at the stage of progesterone dominance over estrogen (post-coital stage and pregnancy), while the excitability of the amygdala increased at the estrous stage and decreased at the postcoital and pregnant stages. The same relationship was observed in alterations of the amplitude and latency of evoked potentials recorded from the periventricular arcuate nucleus in the hypothalamus by stimulation of the hippocampus and amygdala. In the course of the estrous cycle, the amplitude of both negative and positive responses of the ARC to hippocampal stimulation decreased with the progress of estrus and increased at the stage of progesterone dominance over estrogen, such as the rebound stage or anestrous stage. The response, either positive or negative, in the ARC by amygdalar stimulation showed inverted relationships to that of the hippocampus. Excitation of the hippocampus facilitates the release of ovulatory hormones from the adenohypophysis via the hypothalamus, and these discharged hormones yield an increase both in production and output of ovarian progesterone. This increased concentration of progesterone in the systemic circulation elevates the excitability of the hippocampus, in which state the afferent impulse easily induces the elevation of hippocampal activity. This result, therefore, may indicate the existence of a circuit having a positive feedback control mechanism of ovarian progesterone production and output between the activity of the hippocampus and the ovary. Conversely, there may be a negative feedback control of the above-mentioned production and output of progesterone between the amygdaloid activity and the ovary. That is, the ovulatory hormones released from the pituitary gland by amygdalar activation (less than by hippocampal activation), elicit a release of progesterone from the ovary and consequently lower the excitability of the amygdala. The relationship of these negative and positive control mechanisms of ovarian progesterone production and output between the activity of the amygdala and the ovary, and between the hippocampal activity and the ovary is schematically illustrated in Fig. 11. Electrical stimulation of the DHPC exerts a facilitatory effect on ovarian progesterone formation greater than does stimulation of the AMYG or ARC, and had a lower E/P ratio than the latter as shown in Table IV. This might suggestthat the hippocampus plays a role in varying the E/P ratio of mammals along with the estrous cycle. With reference to the function of the limbic system as the focus for the 'negative feedback' action on ACTH secretion, Mason (1957, 1959a, b) found an increase in the 17-OH-CS level in the blood after stimulation of the amygdala, and a decrease after stimulation of the hippocampus in the monkey. According to the unpublished data of Kawakami and Terasawa, alterations in the hippocampal EEG activity were induced by the administration of 0.5 U.S.P.U. of ACTH (Armour Lat No. Y76310), together with a decrease in the amplitude and the frequency of the hippocampal &waves; this represented a slight depression in the
97
LIMBIC SYSTEM A N D PROGESTERONE FEEDBACK
A
R STIMULATION
1
STIMULATION
HIPPOCAMPUS
Positive
Negative feedback
feedback
PROQESTERONE
I
P
PROQESTERONE) I
I I
t
I I
I
positive feedback.
I
Neghive feedback
.
I I
I
-FAClLlTATlOH ____- lHMBlTlON Fig. 11. Schematic illustrations of the hypothetical mode of feedback control between the limbic areas and progesterone (LH) or ACTH (ACH). (A) The feedback mechanism controlling LH secretion: the amygdala-pituitary4vary system forms a negative feedback loop, the hippocampuspituitary-vary system forms a positive loop. (B) The feedback mechanism controlling the ACTH secretion. The amygdalo-pituitary adrenal system forms a positive feedback loop, while the hippocampo-pituitary adrenal system forms a negative feedback system acting on ACTH secretion. The controlling mechanism for ACH secretion is not too clear, though there is some evidence which suggests the negative feedback action of amygdalo-pituitary system and the positive feedback action of hippocampo-pituitary system.
hippocampal activity. However, in the amygdala, the basic fast waves were replaced by the high amplitude fast waves with sporadic 40 c/sec spindle-like bursts after the ACTH administration, which was taken to indicate a marked increase in EEG activity. Furthermore, 30 sec after ACTH injection the ARC potential evoked by hippocampal stimulation was inhibited for 2 h, while the potential evoked by amygdaloid stimulation was markedly facilitated 2 min after injection. The change caused by ACH injection was less clear but seemed to have an opposite effect, in which the hippocampal impulse to the ARC was slightly facilitated and the amygdaloid impulse inhibited. It is possible to interpret these facts as follows. The increase in ACTH concentration in the blood lowers the level of activity in the hippocampus, and raises the level of activity in the amygdala. Putting these effects together with the observation of Mason (1957, 1959a) and Slusher and Hyde (1961) that hippocampal excitation inhibits and amygdalar excitation facilitates ACTH secretion, a mechanism as follows would be postulated. The increase in ACTH References p. I#-I02
98
KAWAKAMI
et al.
would inhibit the hippocampal activity, and facilitate the amygdalar activity. These changes would work together to inhibit the activity of the periventricular arcuate nucleus. The suppression of the periventricular arcuate nucleus would facilitate the secretion of ACTH, and hence of ACH. The heightened level of ACH secretion would then act back to the hippocampus and the amygdala to enhance the activity in the former and lower that in the latter. These changes would elevate the activity of the periventricular arcuate nucleus, bringing about the suppression of ACTH and ACH secretion in sequence. Thus the hippocampus and the amygdala might be working together in the negative feedback control of the ACTH and ACH secretion. As previously mentioned, LH as well as progesterone forms a negative feedback loop with the amygdala-hypothalamo-hypophysis-gonad system and a positive feedback loop with the hippocampo-hypothalamo-hypophysis-gonad system. Therefore, the feedback mechanism controlling LH secretion acts in a somewhat different manner from that of the control mechanism of ACTH secretion. The mutual interplay between the trophic hormones of the anterior lobe and the hormones of the respective target glands has led to the general acceptance of the ‘feedback’ hypothesis as the main mechanism operating to maintain adequate levels of target gland hormones. This type of mechanism has been suggested for the control of the adrenal cortex and the thyroid, as well as for the gonads. The existence of the ‘negative feedback’ control of estrogen and progesterone to the hypothalamus has been especially supported by many studies on the dynamic sequence of the hypothalamo-pituitary-gonad axis in both the reflex and cyclic ovulators. However, the present experiments have made it clear that the hippocampus plays a role in positive feedback control acting on ovarian progesterone formation and its output, and the amygdala has its role in the negative feedback control on the ovarian steroid hormones in the hypothalamo-pituitary-gonad axis. CONCLUSION A N D SUMMARY
New Zealand white rabbits, with chronically implanted electrodes, were used in studies of the influences of electrical stimulation of the limbic system and hypothalamus upon progesterone and estrogen formation in the ovary and ovarian progesterone output by observation of inducement of ovulation and incorporation of radioactivity by progesterone and estrogen in ovary homogenates in vitro with 1-[14C]acetate. The implanted rabbits were primed with estradiol benzoate in oil (0.1 mg s.c.) for 2 days prior to stimulation to ensure an estrous state. Electrical stimulation consisting of monophasic square wave pulses was delivered unilaterally for 30 min, 60 sec on and off, 260-280 PA, at 0.1 msec duration, 60 c/sec. The results were as follows. (1) The hippocampal EEG waves presented a suppressed feature of characteristic 4-8 c/sec and 8-13 c/sec sinusoidal waves with the progress of estrus, and all these waves markedly increased after copulation-induced ovulation and during pregnancy. The 13-30 c/sec high frequency component in the amygdalar EEG increased with the advance of estrus, while theseincreased fast waves were replaced by high amplitude,
LIMBIC SYSTEM A N D P R O G E S T E R O N E F E E D B A C K
99
irregular waves intermixed with sporadic sharp-spikes after copulation-induced ovulation and during pregnancy. In other words, regarding the EEG activity between the hippocampus and the amygdala, a seesaw relationship of excitability might exist both at the estrous stage and at the post-coital stage: during estrus the excitability of the former was lowered, while that of the latter was elevated; at the post-coital stage, the former increased in excitability, and the latter decreased. After injection of progesterone (5 mg) in non-estrogen-primed ovariectomized rabbit the hippocampal &wave increased its amplitude and frequency. On the other hand, the EEG pattern of the amygdala after injection of progesterone showed a decrease in its amplitude. The effect of LH (0.3-0.5 U.S.P. Units) on the hippocampal EEG pattern was a shift from irregular slow waves mixed with low amplitude sinusoidal waves of 3-4 c/sec to the dominance of 4-8 and 8-13 clsec sinusoidal waves in non-estrogen-piimed ovariectomized rabbits. At the same time the amygdalar EEG pattern changed from fast wave dominance to slow wave dominance interposed by sporadic spikes. (2) The potentials evoked in the ARC to which single shock stimuli were delivered from the medial amygdaloid complex were facilitated during estrus as compared with those during post-estrus, while those in the same ARC produced by electrical stimuli of the dorsal or ventral hippocampus were inhibited at estrus and facilitated at postestrus. Between the hippocampus and the amygdala, an inverse relationship of the excitability was found to exist at the estrous and post-estrous stages. After injection of progesterone (5 mg) the ARC potential evoked by stimulation of the hippocampus was facilitated, whereas the ARC potential evoked by stimulation of the amygdala was inhibited. The effect of LH injection upon the ARC potential evoked by the hippocampal stimulation was the facilitation of both negative and positive components. The effect upon the ARC potential evoked by the medial amygdaloid stimulation was the inhibition of both components. (3) By 260-280 pA electrical stimulation of the hippocampus in the estrogen-primed rabbits, ovulation was induced in 20 out of the 30 animals, and almost no differences were observed between the effect of electrical stimulation of the dorsal and ventral parts of the hippocampus. Thus, localized electrical excitation of the hippocampus or the medial amygdala had a similar effect upon the inducement of ovulation. Further, the ovulation thereby induced was blocked by bilateral lesions of the dorsal fornix or by massive lesion o f the septum. (4) Electrical stimulation of the hippocampus enhanced progesterone formation in the ovary more than stimulation of the amygdala or the arcuate nucleus, while there was an inappreciable difference in estrogen formation among these three regions. The ratio of 14C-incorporated estrogen to 14C-incorporated progesterone (E/P) showed a lower value in the hippocampus than in the others. With respect to progesterone labeling from 1-[l4C]acetate by the ovary homogenates, the ovarian progesterone output and the E/P ratio, only a small order of References p . 100-102
100
KAWAKAMI
et al.
magnitude was found to exist between electrical stimulation of the amygdala and the periventricular arcuate nucleus of the hypothalamus including the posterior tuberal region. When the stria terminalis was damaged bilaterally, electrical stimulation of the DHPC led to the production of hemorrhagic follicles in the ovaries. The incorporation of [Wlacetate into progesterone and the progesterone output showed an increase in these rabbits. With the arcuate nucleus lesions, there was no effect of DHPC stimulation upon ovulation, formation of progesterone and estrogen, or progesterone output. Bilateral lesions of the dorsal fornix failed to produce ovulation in response to DHPC stimulation; further, the synthesis of progesterone and estrogen and the progesterone output were lower than those of the controls. These findings may suggest that the effect of DHPC stimulation upon the ovary was exerted through the fornix and the arcuate nucleus, and not through the stria terminalis. Therefore, it is reasonable to infer that the enhanced activity of the hippocampus exerts influences somewhat different from those of the amygdala upon the biosynthesis of sex steroids in the ovary as well as upon the ovarian progesterone output through the hypothalamo-pituitary-gonad axis. In conclusion, the above-mentioned results support the presumption that the hippocampus and the amygdala are the critical areas of feedback control of progesterone : the hippocampo-hypothalamo-pituitary-gonad axis is a loop of the positive feedback of progesterone and the medial amygdaloid complex a loop ofthe negative feedback of progesterone. A C K N O W L E D G E MEN TS
The authors wish to express their appreciation to Mrs. H. J. Hagino for her guidance in the expression of English. This experiment was supported by a grant from the National Institute of Health, U.S. Public Health Service (NB-03860-4), and a grant from the Ministry of Education, Japan.
BAN,T., (1964); The hypothalamus, especially on its fiber connections, and the septo-preopticohypothalamic system. Med. J. Osaka Univ., 15, 1-83. BARD,P., AND MACHT,M. B., (1958); The behaviour of chronically decerebrate cats. Ciba Foundafion Symposium on the Neurological Basis of Behaviour. G . E. W. Wolstenholme and C. M. OConnor, Editors. London, Churchill, pp. 55-71. BROOKS, C. McC., (1937); The role of the cerebral cortex and of various sense organs in the excitation and execution of mating activity in the rabbit. Amer. J. Physiol., 120, 544-553. C. McC., (1938); A study of the mechanismwhereby coitus excites the ovulation producing BROOKS, activity of the rabbit's pituitary. Amer. J. Physiol., 121, 157-177. BUNN,J. P., AND EVEREIT,J. W., (1951); Ovulation in persistent-estrousrats after electrical stimulation of the brain. Proc. SOC.exp. Biol. ( N . Y.), 96, 369-371. E. C., HUMPHREY, T., AND LAUER,E. W., (1962); Correlative Anaromy of rhe Nervous System. CROSBY, New York, Mac-Millan. ELWERS, M., AND CRITCHLOW,V., (1961); Precocious ovarian stimulation following interruption of stria terminalis. Amer. J. Physiol., 201,281-284.
L I M B I C SYSTEM A N D P R O G E S T E R O N E F E E D B A C K
101
FEE,A. R., AND PARKES, A. S., (1930); Studies on ovulation. 111.Effect of vaginal anesthesia on 0-avul tion in the rabbit. J. Physiol. (Lond.), 70, 385-388. F L E R KB., ~ , (1954); Zur hypothalamischen Steuerungder gonadotrophen Funktion der Hypophyse. Acta morph. Acad. Sci. hun:., 4,475-492. GLOOR,P., (1956); Telencephalic influences upon the hypothalamus. Hypothalamic-Hypophysial Interrelationship. W. S . Fields, Editor. Springfield,Thomas, pp. 74-1 13. GOLDSTEIN, A. C., (1957); The experimental control of sex behaviour in animals. Hormones, Brain Function and Behaviour. H. Hoagland, Editor. New York, Academic Press, pp. 99-119. GORSKI,R. A., AND BARRACLOUGH, C. A., (1962); Studies on hypothalamic regulation of FSH secretion in the androgen-sterilized female rat. Proe. SOC.exp. Biol. ( N . Y . ) , 110, 298-300. GREEN, J. D., AND ADEY,W. R., (1956); Electrophysiologicalstudies of hippocampal connections and excitability. Electroenceph. din. Neurophysiol., 8, 245-262. R. P., AND SCOTT,P. P., (1958); Neurological site of action of stilboestrol HARRIS,G. W., MICHAEL, in eliciting sexual behaviour. Ciba Foundation Symposium on the Neurological Basis of Behaviour. Boston, Little Brown, pp. 236-251. HAYWARD, J. N., HILLIARD, J., AND SAWYER, C. H., (1964); Time of release of pituitary gonadotropin induced by electrical stimulation of the rabbit brain. Endrocrinology, 74, 108-1 13. HILLIARD, J., ARCHIBALD, D., AND SAWYER, C. H., (1963); Gonadotropic activation of preovulatory synthesis and release of progestin in the rabbit. Endoerinoiogy, 72, 59-66. HILLIARD, J., HAYWARD, J. N., AND SAWYER, c.H., (1964); Postcoital patterns of secretionof pituitary gonadotropin and ovarian progestin in the rabbit. Endocrinology, 75, 957-963. HILLIARD, J., AND SAWYER, C. H., (1964); Synthesis and release of progestin by rabbit ovary in vivo. Proe. First International Congress on Hormone Steroids, 1,263-272. HISAW,F. L., (1947); Development of the Graafian follicle and ovulation. Physiol. Rev., 27,95-119. W., AND CHAMORRO, A., (1937); uber die luteinisierende Wirkung des Follikel-hormones HOHLWEG, durch Beeinflussung der luteogenen Hypophysenvorderlappensekretion.Klin. Wschr., 16, 196-197. HOHLWEG, W., AND JUNKMANN, K., (1932); Die hormonal-nervose Regulierung der Funktion des Hypophysenvorderlappens. Klin. Wschr., 11. 321-323. KAWAKAMI, M., AND SAWYER, C. H., (1959a); Induction of behavioral and electroencephalographic changes in the rabbit by hormone administration or brain stimulation. Endocrinology, 65,631-643. KAWAMAKI,M., AND SAWYER, C. H., (1959b); Neuroendocrinecorrelates of changes in brain activity thresholds by sex steroids and pituitary hormones. Endocrinology, 65, 652-668. KAWAKAMI, M., TERASAWA, E., TSUCHIHASHI, S., AND UEMURA, T., (1965); Sex hormone sensitive componentsin the rabbit brain and their physiologicalsignificance. U.S.-Japan Joint Conferencefor 'Dynamics of Steroid Hormones'. U. S.-Japan Cooperative Science Program, pp. 1 4 KAWAKAMI, M., TERASAWA, E., TSU~HIHASHI, S., AND YAMANAKA, K., (1966); Differential control of sex hormone upon brain activity in rabbit and its physiological significance. In the press. KAWAKAMI, M., AND UEMURA, T., Unpublished data. KOIKEGAMI, H., (1963); Amygdala and other related limbic structures; experimental studies on the anatomy and function. Acta med. biol.. 10, 161-277. KOIKEGAMI, H., (1964); Amygdala and other related limbic structures; experimental studies on the anatomy and function. Acta med. biol., 12, 73-266. KOIKEGAMI, H., YAMADA, T., AND Usur, K., (1954); Stimulation of amygdaloid nuclei and periamygdaloid cortex with specialreferenceto its effects on uterine movementsand ovulation. Foliapsychiat. neurol. Jap., 8, 7-31. LISK,R. D., (1962); Testosterone-sensitivecenters in the hypothalamus of the rat. Acta enuocr. Kbh., 41, 195-204. MASON, J. W., (1957); The central nervous system regulation of ACTH section. Reticular Formation ofthe Brain. H. H. Jasper, L. D. Proctor, R. S. Knighton, W. C. Noshay, and R. T. Costello,Editors. Boston, Little Brown, pp. 645-670. MASON,J. W., (1959a); Plasma 17-hydroxycorticosteroidlevels during electrical stimulation of the amygdala complex in conscious monkey. Amer. J. Physiol., 196, 44. MASON, J. W., (1959b); Visceral functions of the nervous system. Ann. Rev. Physiol., 21, 353. MICHAEL, R. P., (1961); An investigation of the sensitivity of circumscribed neurological areas to hormonal stimulation by means of the application of oestrogens directly to the brain of the cat. Regional Neuroehemistry. S. S . Kety, and J. Elkes, Editors. Oxford, Pergamon Press, pp. 465-480. NAUTA, W. J. H., (1956); An experimental study of the fornix in the rat. J . comp. Neurol., 104,247. NAUTA, W. J. H., (1963); Central nervous organization and the endocrine motor system. Advances in Neuroendocrinology. A. V. Nalbandov, Editor. Urbana, Illinois Univ. Press, pp. 5-21.
102
KAWAKAMI
et al.
Rrce, B. F., HAMMERSTEIN, J., AND SAVARD, K., (1964); Steroid hormone formation in the human ovary 11. Action of gonadotropins in vitro in the corpus luteum. J. clin. Endocr., 24, 606-615. ROBERTS, S., SEW, K., AND HA"^, B. H., (1962); Regulation of cerebral metabolism. J. Neurochem., 9,493-501. SAWYER, C . H., (1957); Triggering of the pituitary by the central nervous system. Physiological Triggers. T. H. Bullock, Editor. Baltimore, Waverly Press, pp. 164-174. SAWYER, C. H., (1959); Nervous control of ovulation. Endocrinology of Reproduction. C . W. Lloyd, Editor. New York, Academic Press, pp. 1-18. SAWYER, C. H., AND EVERETT, J. W., (1959); Stimulatory and inhibitory effects of progesterone on the release of pituitary ovulating hormone in the rabbit. Endocrinology, 66, 644-651. SAWYER, C. H., EVERETT, J. W., AND GREEN, J. D., (1954); The rabbit diencephalon in stereotaxic coordinates. J. comp. Neurol., 101, 801-824. SETO,K., SEKIGUCHI, N., Us~mosm,I., AND Umzu, M., (1964); The effects of the pituitary gland on the metabolism of testicles. Tohoku J. Zootech. Sci., 14, 39 (In Japanese). SHEALY,C. N., AND PEE=, T. L., (1957); Studies on amygdaloid nucleus of cat. J. Neurophysiol., 20, 125-139. SLUSHER, M. A., AND HYDE,J. E., (1961); Effect of limbic stimulation on release of corticosteroids into the adrenal venous effluent of the cat. Enabcrinology, 69, 1080-1084. SZENTAGOTHAI, J., FLERK~, B., MESS, B., AND mkz, B., (1962); Hypothalamic Control of the Anterior Pituitary. Akademiai Riad6, Budapest. TERASAWA,E., AND U w w , M., (1965); Differential control of oxytocin upon the evoked potential in the hypothalamus. Abstracts, 23rd lnt. Congr. physiol. Sci., p. 430. VERLY, W.G.,(1951); Theurinaryexcretionofpregnane3~:20a-diolin the femalerabbit immediately after mating. J. Endocrinol., 7,258-259.
103
Affective Behaviour produced by Electrical Stimulation in the Forebrain and Brain Stem of the Cat ROBERT W. H U N S P E R G E R
AND
V E R E N A M. B U C H E R
Department of Physiology, University of Zurich, Zurich (Switzerland)
The present paper is concerned with the affective patterns produced by electrical stimulation of the forebrain and brain stem in the unanaesthetized, freely-moving cat, and will describe the structures underlying these reactions. We shall on!y consider experiments carried out in a neutral, unconditioned milieu that is, experiments to which behavioural criteria conform with natural conditions can be applied. It is hoped thereby to avoid the tendency to study ‘pieces of behaviour’ rather than a complete ‘natural’ pattern when dealing with cerebral organisation. Particular attention will be paid to differences in grossly similar patterns, a typical feature dependent on the site or the level explored. Recent work (unpublished) on seismographic recording of increased postural tonus and heart activity during emotional excitement will also be reported. As the literature dealing with the subject has been recently reviewed by Hunsperger (1963), a brief comment on recent work will suffice, on which discussion can be based. RESPONSE PATTERNS
The affective responses produced by electrical stimulation of the forebrain and brain stem are threat and flight reactions; mewing; and, effects suggestive of reaction to pain. The experiments were carried out on 134 cats in which 840 spots (‘points’) in the aforesaid structures had been stimulated (Fernandez de Molina and Hunsperger, 1959; Hunsperger, 1956; Hunsperger and Gwbidi, 1964), using the Hess technique (1932, 1957), modified by Wyss (1950, 1965). Unipolar stimulation was carried out by means of delayed condenser discharges up to 3 V at frequency of 8 per sec, and up to 2 V at 17 per sec. Threat andJEightpatterns
The threat pattern, also referred to as ‘affective defence reaction’ (Hess and Briigger, 1943; Hunsperger, 1956), elicited in the cat is characterised by lowering of the head, laying back of the ears, hunching of the back, accompanied by growling and hissing and signs of sympathetic discharge, such as pupillary dilatation and pilo-erection. References p . 125-127
104
11. W. HUNSPERGER A N D V. M. BUCHER
Fig. 1. Threat patterns elicited from (a) the intermediate zone of the hypothalamus, (b) dorsomedial parts of the amygdala, and (c) the central gray matter of the midbrain. Note the purposeful ‘acting out’ character of the pattern elicited from the hypothalamus (a), the more restrained display evoked from the amygdala (b), and the less complete pattern obtained from the midbrain (c).
The pattern obtained from the forebrain and the brain stem varies somewhat according to the level stimulated. At the level of the diencephalon, i.e. the intermediate zone of the hypothalamus, the response strikingly resembles the natural reaction, seen when cat meets dog, or an unfriendly fellow cat (Fig. la). To the affective defence pattern just described is added an element pertaining to attack: the cat raises a forepaw, ready to strike. Vigorous hissing accompanies the display at this level. Sometimes, after fierce threat display, the cat resorts to aght. With low frequency stimulation, the response builds up gradually and is always preceded by a state of increased awareness, attention (‘Weckeffekt’). The latencies for the first hissing outburst vary from 5 to 50 sec, and depend on the site and the intensity of stimulation. Higher frequency and intensity parameters, occasionally applied, may, however, produce hissing within less than 1 sec. At the level of the forebrain, i.e. the amygdala and stria terminalis system, growling
AFFECTIVE BEHAVIOUR
105
accompanies the defensive display. Hissing only occurs after prolonged stimulation (Fig. lb). The response, in contradistinction to the reaction obtained from the hypothalamus, shows less ‘acting out’. The latencies for growling vary from 3 to 60 sec and for hissing from 7 to 60 sec. Stimulation at the level of the midbrain, i.e. the central gray matter of the aqueduct, always produces hissing and pupillary dilatation, but less regularly a full threat pattern (Fig. lc). The latencies for hissing are on an average shorter than those that hold at the level of the hypothalamus. The $ight pattern is regularly obtained at brain stem levels, but is rarely produced from the forebrain. The flight reaction elicited from the hypothalamus suggests premeditation. The cat, soon after onset of stimulation, looks about as if in search of an outlet, gazes fixedly at an escape route, and within 1-2 sec jumps over the wall of the table, landing firmly on its feet (Fig. 2). The latencies for this ‘escape jump’ vary from 4 to 15 sec. Stimulation with higher frequencies and intensities may produce flight within 1-2 sec. When the reaction is produced from the posterior hypothalamus and adjacent subthalamus, the jump may be preceded by restless wandering on the table (‘Bewegungsdrang’, Hess, 1949 ; Hunsperger, 1956) associated with investigative activity. The flight reaction is only produced occasionally from the amygdala and appears to
Fig. 2. Flight pattern elicited from the hypothalamus. (A) situation before stimulation; (ED) cat looks about as if in search of an outlet; (EF) eyes arrested on escape route; (G-I) cat jumps over wall of table. (After Hunsperger, 1956). References p . 125-127
Fig. 3. Postural tonus, cardiac activity and affective excitement produced by middle frequency stimulation in the hypothalamus. Seismographic record of increased tonic muscular activity and cardiac activity associated (a) with hissing, threat response (Exp. 222), (b), see next page, with excitement leading finally to flight (Exp. 226). Tarachogram: upper channel, photoflash signal line; middle channel, seismographic record of mechanical forces; lower channel, monitoring of voltage after rectification. Time, sec; calibration 20 g *. Reading from above downwards the sequence is as follows: 1st row, left: situation before stimulation 4th row, left: cat raises head spontaneously (signal), right: minimal tonic muscle activity, heart beats not clearly right: low tonic activity, with breaking in of spontaneous slow revealed. head movements. 2nd row, left: hissing, 5th row, left: head brought back to initial position, cat relaxed (signal), right: with increasing voltage (see lower channel) smooth inright: beginning of tracing, forces produced by this rather abrupt crease in tonic muscle activity reaching its peak during head movement. End of tracing, minimal tonic activity hissing (seesignal, upper channel) and remaining at this during relaxed state (signal), with multiphasic deflections level during the penod of maintained stimulation that due to cardiac activity. follows (see below). 3rd row, left: slow turning of head (at the end of maintained stimulatiright: on), duringfalling phase of current, decrease in tonic activity, with breaking in of further slight head movements (signal).
See also legend under Fig. 3a, p. 107. Fig. 3b. 1st row, left: situation before stimulation right : moderate tonic activity, with breaking in of occasional biphasic deflections due to cardiac activity 2nd row, left: right: beginning of stimulation 3rd row, left: slight arousal right: with increasing voltage, increase in tonic activity, and increased deflections produced by cardiac activity.
4th row, left: the animal is tense right: strong tonic muscular activity, with marked biphasic deflections produced by cardiac activity. Just after signal, the cat stands up and jumps over the wall, extreme right. Heart rate just before the animal stands up is 300 per min. The forces of the heart beats lie between 15 and 20 g*.
110
R. W. HUNSPERGER A N D V. M. BUCHER
legend -----f
AFFECTIVE BEHAVIOUR
111
be sudden and unpremeditated. Soon after onset of stimulation the eyes of the animal dart to and fro, the head follows the eye movements and then, suddenly, the cat rushes off the table. The flight reaction obtained from the midbrain appears to be less deliberately prepared than is the response evoked from the hypothalamus. Postural tonus and cardiac eflects associated with threat and flight behaviour The excitement manifested during the threat and flight responses is associated with an increase in the tonus of the skeletal musculature and with an increase in rate and strength of heart beats. This increase can be made evident by resorting to seismographic recording of the mechanical forces produced by the tonic innervation of the skeletal muscles (microvibrations) and by the heart beat. A technique permitting recording and measurement of these endogenous repercussions of the body, has been developed by Corti et al. (1955). The device used in the experiments to be described permits measurement of the vertical force component in the unrestrained animal (Corti et al., 1961; Hunsperger and Dietiker, 1962). It has been shown by Wyss (1963) that amplitude modulated middle frequency currents in a peripheral nerve produce simultaneous excitation under both electrodes. It has also been shown that slowly rising middle frequency currents of long duration produce asynchronous activation of central nervous structures (Hunsperger, 1965), but do not elicit the stimulus-bound muscle action produced by the low frequency cathodal stimulation. Fig. 3a and b shows the increase in postural tonus and cardiac activity associated with threat and flight reactions. The figures on the left show the behaviour response; on the right, the corresponding seismographic recording (the so-called tarachogram). The increase in rate and strength of heart beats during hissing in Fig. 3a is somewhat masked by strong tonic muscular activity and by the forces produced by the slow head movements. Subthreshold activation of the threat zone eliciting arousal, but no head turning, also evokes a marked increase in rate and strength of heart beats. Fig. 4 shows this cardiac effect. The increase in rate and strength of heart beats during threat and flight responses produced by stimulation in the hypothalamus (the only region tested) proves that cardiac activation forms an integrative part of the patterns obtained. It is well known from the researches of Cannon (1939, 1953), that sympathico-adrenalino activity occurs during the display of energy which rage or fear or pain may inducee. Abrahams et al. (1960) have shown that active (cholinergic) dilatation of the skeletal muscle vessels also forms part of the defence pattern. The synergism between autonomic effects (cardio-vascular, pupillary, pilo-motor) and somato-motor effects (increase in postural tonus) ensures ergotropic adjustment to the external environment (Hess, 1949, 1957). Fig. 4. Increase in rate and strength of heart beats during subthreshold activation of threat zone (Exp. 222). 1st row: beginning of stimulation, weak cardiac activity (recurrent biphasic deflections), heart rate approximately 100 per min; 3rd row: period of maximal stimulation, strong and frequent heart beats (mono- and biphasic deflections), heart rate 220 per rnin, force of heart beats approximately 15 g* ; 5th row: no stimulation, weak cardiac activity (multi-phasic deflections). References p . 125-127
112
R. W. HUNSPBRGER AND V. M. BUCHER
Influence of environmental conditions on threat and flight behaviour When a stuffed cat is placed on the table during stimulation of the threat zone in the hypothalamus, the defence pattern (threat occasionally followed by flight) is modified (Brown, Hunsperger and Rosvold, in preparation).
Fig. 5. ‘Warding off‘ and attack reactions on presentation of a stuffed cat (right), elicited from the threat zone of the hypothalamus. (a) warding off. The reaction was followed by flight (exp. IVL); (b) cat strikes dummy’s face (Exp. Vn);(c) cat clutches dummy with forepaws, prior to biting nape of stuffed cat’s neck (attack!) (Exp. IIL). In all figures note hissing, laying back of ears, protrusion of claws and pilo-erection. (After Brown, Hunsperger and Rosvold, in preparation.)
AFFECTIVE BEHAVIOUR
113
Sometimes the excited animal, hissing, with claws unsheathed, wards off the foe (Fig. 5a). A moment later, it will jump from the table in a frenzy. At other times, the cat, still hissing, with claws unsheathed, strikes the dummy’s face. An element of attack has come in! (Fig. 5b). Or, the hissing cat may clutch the throat of the dummy with both paws, prior to planting its teeth in the nape of the dummy’s neck (Fig. 5c).
Fig. 6. Mewing elicited during passage of current. (a) uneasy mewing obtained on stimulation in the preoptic area; (b) mewing of a protestingcharacter obtained from dorsomedialparts of the thalamus; (c) plaintive mewing on stimulation in the septum. (After Hunsperger and Gwbidi.), 1964). References p. 125-127
114
R. W. HUNSPERGER A N D V. M. BUCHER
The pattern has passed from defence to offense (attack)! Comparable results have been observed if a stick or the hand of the investigator is brought near the cat during stimulation of the threat zones in the hypothalamus and midbrain. If the animal is stopped during the flight produced from the appropriate field in the hypothalamus, a threat display may replace the flight initiated. These modified, progressively full patterns, are more easily evoked from the hypothalamus than from the midbrain, and have, so far, never been obtained from the amygdala (Fernandez de Molina and Hunsperger, 1959; Hunsperger, 1956). Mewing
Mewing occurs either during stimulation, or shortly after cessation of stimulation, or during and after stimulation. The mewing elicited during the passage of current is always associated with increased attention or awareness. Uneasy, protesting or plaintive in character, this vocal reaction depends on the site of stimulation (Hunsperger, 1963). Fig. 6a shows uneasy mewing. The jaws are half-opened, the corners of the mouth are not fully retracted, and the upper lip is slightly pursed. The cat looks about in an inquisitive manner, pupils moderately dilated, and often briefly licks its upper lip. Fig. 6b shows mewing protesting in character. The chin is pushed forward, the jaws are almost fully opened (a movement that draws the ears downward and sideways), and the corners of the mouth are retracted. Fig. 6c shows mewing plaintive in tone. The chin is pushed forward, the jaws are almost fully opened, but the comers of the mouth are barely retracted, and the upper lip remains pursed. The mewing following cessation of stimulation only occurs if stimulation produces a state of fixed attention (Fig. 7a) or an arrest reaction (Hunter and Jasper, 1949) with pupils widely dilated. The moment the fixed attention ceases, and the pupils return to normal, the cat mews repeatedly, either plaintively or protestingly. Fig. 7b shows this moment; the cat lifts its head, pupils have returned to normal, then come the plaintive, protesting mews. This vocal reaction according to Andy and Akert (1959, Cadhillac (1955), Hunter (1950), Hunter and Jasper (1949), Kaada et al. (1953), LissSlk et al. (1957), MacLean (1955,1957a) marks the moment of cessation of bioelectrical after-discharges. Mewing also occurs after a generalised seizureattackinducedby electrical stimulation in the forebrain or thalamus. After cessation of the tonic-clonic motor manifestations, at the moment that the animal recovers from the state of confusion associated with pupillary dilatation into which it had been thrown, loud mews, at first protesting, then plaintive, are uttered (Fig. 7c). Mewing during and after stimulation: The uneasy mews, accompanied by somewhat stronger pupillary dilatation and movements of eyes and head from side to side, produced during stimulation are replaced by protesting or plaintive mews shortly after the current has been turned off. Stronger stimulation suppresses the uneasy mews, but accentuates the atypical arousal pattern. When stimulation is discontinued, very loud and frequent mews are emitted as an effect of rebound (Hunsperger, 1963).
AFFECTIVE BEHAVIOUR
115
Fig. 7. Mewing obtained after cessation of stimulation. (a) stimulus-bound, frozen attention 10 sec before end of stimulation; (b) mewing of a protestingcharacter, 11 sec following cessation of stimulation; (c) paintive mewing, 2 min following a generalised epileptic seizure produced by stimulation of the anterior nuclei ofthe thalamus.
Pattern suggestive of reaction to pain Such effects are sometimes evoked by strong stimulation at midbrain levels, and somewhat equivocally, from the hypothalamus and the forebrain. From the midbrain, piercing cries, interrupted by hissing, were elicited from 3 out of the 21 points that, with weak stimulation, evoked mewing without defence reaction. Pupils are widely dilated. The animal utters a piercing cry, head upright, ears pointed, tail anxiously pressed against the body (Fig. Sa). The hissing and growls that form References p . 125-127
116
R. W. HUNSPERGER A N D V. M. BUCHER
Fig. 8. Patterns suggesting a reaction to pain. (a) piercing cries, with irruptions of hissing, obtained by strong stimulation in ventral parts of the central gray matter of the midbrain, an area yielding mewing as threshold response. Note maximal pupillary dilatation, head erected, ears pointed and tail drawn in; (b) strident shrieks produced by strong stimulation in dorsomedial parts of the amygdala. The stimuli had at first produced growling. Note flattened ears.
part of the defence pattern evoked from the infundibular region, werein oneexperiment replaced by growls giving way to shrieks, a reaction recalling the piercing cries accompanied by strong pupillary dilatation described by Karplus and Kreidl (1909, 1910, 1928) in their historical papers. From the amygdala, shrieks replaced the growls that characteristically form part of the defence pattern obtained as threshold answer from this structure, but only from 5 out of 63 active points (Fig. 8b). The question remains open whether these shrieks expressed discomfort or fury. T O P O G R A P H Y OF A C T I V E AREAS
Substratum governing threat andflight
The field governing threat and flight behaviour delimited by means of electrical stimul-
AFFECTIVE BEHAVIOUR
117
ation, extends from dorsomedial parts of the amygdala by way of the stria terminalis to the stria bed at the level of the anterior commissure (Fernandez de Molina and Hunsperger, 1959), and from there through the intermediate zone of the preoptic area and the hypothalamus (Hess et ul., 1945/46) into the central gray matter of the midbrain (Hunsperger, 1956) (Fig. 9). In the brain stem, the areas yielding threat and flight occupy delimited zones. Threat is obtained from two ‘inner zones’, situated in the intermediate region of the anterior hypothalamus and in the central gray matter of the midbrain respectively; flight is elicited from an unbroken zone surrounding and interconnecting the two fields mediating threat (Hunsperger, 1956). Within the responsive territory of the amygdala and stria terminalis system, the few points yielding flight lie scattered between points yielding threat with growling or growling followed by hissing (Fernandez de Molina and Hunsperger, 1959) (see Fig. 10).
Fig. 9. Schematic illustration representing the threat and flight zones in the forebrain and brain stem. Sagittal section through the brain stem with amygdala and other more laterally situated structures superimposed. x x x x = responsive field of the amygdala, stria terminalis and stria bed at the level of anterior commissure, continuing into )I-)) = active areas of the hypothalamus and central gray matter of the midbrain. Black = ‘inner zones’ yielding threat; Hatched = ‘outer zone’ yielding flight. (Slightly modified after Fernandez de Molina and Hunsperger, 1959.) References p. 125-127
118
R. W. H U N S P E R G E R A N D V. M. BUCHER
The location of the active area for threat in the hypothalamus found by Hess and Brugger (1943) agrees with our findings. The field lies predominantly in a rostro-dorsolateral position with regard to the ventro-medialnucleus of the hypothalamus, although some overlapping occurs. These findings have been reconfirmed by Brown et al. (in preparation) using bipolar stimulation and coaxial electrodes of small dimensions. In cats shut in a cage, Nakao (1958) obtained aggressive behaviour from medial and inferior portions of the middle hypothalamus, attempts to escape preceded by hissing from an anteriorly adjacent region up to the level of the chiasma. Yasukochi (1960), also working with cats shut in a cage, produced anxiety or fear from anterior parts of the hypothalamus, rage and furious expressions of aggression from middle parts (ventro-medial nucleus), and behaviour suggestive of curiosity from posterior parts. Brown, Hunsperger and Rosvold, however, produced premeditated flight or restless wandering associated with investigative activity followed by flight from the posterior hypothalamus and the adjacent subthalamus. To sum up, in spite of small differences, probably due to differences in experimental conditions, our results agree with those of
1
am ygdala
Fig. 10. Frontal section through the middle portion of the amygdala. Shows stria terminalis. Level of greatest extension of responsive field for threat and flight as delimited by Fernandez de Molina and Hunsperger, 1959. 0 = threat with growling; = threat with growling followed by hissing; 0 = flight of unpremeditated character; = negative points with regard to affective reactions, but yielding the effects indicated by the symbols in their vicinity; 0 = represents the areas yielding; // = contralateral turning of eyes and head; > = repeated jerking of eyes and head to the side opposite stimulation.
AFFECTIVE B E H A V I O U R
119
other workers. The problems raised will be discussed more fully in the paper in preparation. Skultety (1963), working with caged cats, produced rage reactions from the two anterior thirds, and attempts to escape from caudal parts of the central gray matter of the midbrain, but not from the adjacent tegmental fields, a territory from which Hunsperger (1956, Fig. 2) produced flight reactions in the freely-moving animal. The location of the active area for threat within the amygdala established by Fernandez de Molina and Hunsperger (1959) agrees well with the findings of Naquet (1953), MacLean and Delgado (1953), Magnus and Lammers (1956), Shealy and Peele (‘1957) and Wood (1958), but does not tally with the findings of Kaada et al. (1954) and Ursin and Kaada (1960). According to these authors, the zone concerned with fear occupies lateral portions of the amygdala, and the zone concerned with anger occupies medial and ventral portions of this structure. Fernandez de Molina and Hunsperger, however, obtained turning of the eyes and head, or repeated jerking of eyes and head to the side opposite stimulation from the lateral amygdaloid nucleus (Fig. lo), a reaction recalling the ‘attention response’ evoked by Ursin and Kaada (1960) from this nucleus and more medially lying structures by applying weak stimulation. For the anxiety reactions obtained by Sano (1958) and Fangel and Kaada (1960) and others from cortical structures, see Hunsperger (1963). Hilton and Zbroiyna (1963) have recently emphasised that defence reactions according to these authors threat, wildly running movements and jumping up the walls of the cage in which the cats were enclosed - can be traced from the basal nucleus of the amygdala into the preoptic area and rostral hypothalamus by way of the anterior amygdaloid area and a narrow connection dorsal to the optic tract. The band connecting the amygdala and preoptic/hypothalamic region, according to these authors, seems to correspond to the ventral amygdalofugal pathway described by Nauta (1961). Multiple electrolytic lesions placed in this band abolished the responses from the amygdala. It is not clear from their illustration (Fig. 5) whether or not the stria terminalis system was involved. Neither threat nor flight was obtained by Fernandez de Molina and Hunsperger (1959) by stimulation of the area lying beneath the pallidum between the anterior amygdaloid region and the preoptic area and rostral hypothalamus, an area comprising the diffuse fibres described by Johnston (1923), Fox (1943), and Nauta (1961). This area has recently been re-explored by Hunsperger using bipolar stimulation. The effects yielded were compared with former results obtained with monopolar stimulation. All threat and flight responses obtained from the forebrain and hypothalamus with stimuli up to 2.5 V at frequency 8 or up to 1.5 V at frequency 17 (in other words, stimuli not exceeding three times the lowest threshold for these responses) were considered positive, and were plotted in diagrams based on the Hess atlas. The extent of the areas in the forebrain and hypothalamus yielding threat and flight is illustrated in Fig. 11 in 5 frontal sections. Negative points with regard to these reactions are also shown. This figure should be compared with Fig. 1 of Hilton and ZbroyZna (1963) and Figs. 5 and 6 of Ursin and Kaada (1960). The stipplled areas, sections 326,352, indicate the region through which the diffuse fibres (d.f.) run. It will be observed that the responsive points for threat and flight are grouped in the References p.
12.5427
120
R. W. H U N S P E R G E R AND V. M. B U C H E R
Fig. 11. Schematic illustration of areas in the amygdala, stria terminalis bed and hypothalamus yielding threat, flight or mixed reactions. Shows region below the pallidurn yielding negative findings with regard to affective reactions. 5 frontal sections (atlas of Hess, S 326-430). Effects obtained at intensities up to 2.5 V a t 8 per sec and up to 1.5 V a t 17 per sec. 0 = threat with growling and/or hissing; 0 = flight; C ) = hissing followed by flight; 0 = negative points with regard to affective reactions.
AFFECTIVE BEHAVIOUR
121
Fig. 12. Effects obtained by bipolar stimulation of region belov, the pallidum and location of electrode serving as cathode. (a, left) ipsilateral twitching of the eyelid and rno’xients of the tongue as if to eject a foreign body (anode in region rostra1 to cathode); (a, right) ips.kteral twitching of the upper lip and repeated jerking of eyes and head to side opposite stimulation (anode in region caudal to cathode). (b) the electrode serving as cathode lies beneath the pallidum at the level of Fox’s association bundle b (see arrow).
References p. 125-127
122
R. W. H U N S P E R G E R A N D V. M. B U C H E R
amygdala (sections 378,404,430). By way of the stria terminalis, they can be traced to the bed of the stria at the level of the anterior commissure, shown in section 326, and from this level, along the hypothalamic component of the stria (section 352) into the preoptic and hypothalamic fields (sections 378,404,430). The question of direction of conduction in this path remains open (Fernandez de Molina, and Hunsperger, 1962; Hilton and Zbroiyna, 1963). Again it will be observed that negative points with regard to threat and flight lie scattered above the optic tract (sections 430, 404) and in the region beneath the pallidum through which the diffuse fibres pass (sections 378, 352, 326). The reactions obtained from these ‘negative points’ (a total of 18) include pupillary dilatation, sniffing, salivation, tongue movements, facial motor effects, and repeated jerking of eyes and head to the side opposite stimulation, the latter an effect, as already mentioned, recalling the ‘attention response’ of Kaada. Mewing was obtained only once. Fig. 12a shows two examples of responses obtained by bipolar stimulation of the region of the diffuse fibres. The track of the electrode that served as cathode is shown in Fig. 12b and lies beneath the pallidum among diffuse fibres, including Fox’s b bundle. Fig. 12aleft, shows the effects obtained when the electrode that served as anode was placed in the innominate substance of Reichert -twitching of eyelids and movements of the tongue as if to eject a foreign body. Fig. 12a right, shows the effects produced when the electrode that served as anode was placed more caudally in the region between the optic tract and the entopeduncular nucleus - twitching of the upper lip and repeated jerking of eyes and head to the side opposite stimulation. All these observations lend no support to the contention that the responsive fields for threat and fiight (defence reaction according to Hilton and Zbroiyna) in the amygdala and the hypothalamus are connected by way of a direct ventral route that passes in the region between the optic tract and pallidum. The role played by the diffuse system -a subject to which Koikegami and Yoshida (1953), Ursin and Kaada (1960), Hilton and ZbroZyna (1963), Karli and Vergnes (1964), have drawn attention requires further investigation. Substratumfor mewing
The field for mewing obtained during passage of current also extends from the brainstem to forebrain. At the level of the midbrain, it occupies ventral portions of the central gray matter adjacent to the more dorsally lying region for threat pattern. At Fig. 13. Schematic illustrationof points in the forebrain and diencephalonyielding mewing. 8 frontal sections (atlas of Hess S 287-469). Lightly shaded areas: hippocampus/fornix system and region of transitionlateral preoptic area/hypothalamus, dotted area: gray matter of the hypothalamus and preoptic area. 0 = mewing evoked duringpassageofcurrent; 0 = mewing occurring after cessation of stimulation ; //// = responsive field for threat response in the amygdala, stria terminalis system and hypothalamus ; \\\\ = responsive field for flight in the hypothalamus and preoptic area; 0 = negative points with regard to affective reactions.
AFFECTIVE B E H A V I O U R
References p . 125-127
I23
124
R. W. HUNSPERGER AND V. M. BUCHER
diencephalicand forebrain levels (Fig. 13, S a-h) active points in the preoptic region and anterior hypothalamus border on and intermingle with the active area for flight and threat, and are further traceable into the septum (including precommissural fornix), fornix (S a, b), and fimbria/hippocampus (S f-h), the latter findings supporting results obtained by Magnus and Lammers (1956), MacLean (1955, 1957b) and Parmeggiani (1960). Another group of active points is situated in anterior and dorso-medial parts of the thalamus, and appears to follow the track of the stria medullaris to the habenula. Eleven points were stimulated in this latter structure and the tractus Meynert (S g,h ), but mewing was only obtained from one. It has been shown by Wallenberg (1902), Burgi and Bucher (1960, 1963), that the stria medullaris, besides ipsilateral fibres, conveys a bundle that crosses in the habenular commissure, runs forward on the contralateral side and distributes (arrows, section c, Fig. 13) to the region of transition lateral preoptic area/hypothalamus, an area from which marked mewing responses were consistently obtained as the sole effect (S c, d). The question arises whether the effects obtained from the stria medullaris region are due to activation of the focal field on the side opposite stimulation. This focal field (transition lateral preoptic area/hypothalamus -the lightly-shaded area in Fig. 13, b-d) also receives fibres from the hippocampus/fornix system and septum according to Valenstein and Nauta (1959). We suggest that mewing obtained from these latter structures may be secondarily reinforced by activation of this focal field. The active points for mewing obtained after cessation of stimulation are widely dispersed in structures of the diencephalon and forebrain and include the hippocampus /fornix system, anterior nuclei of the thalamus, cingulum, and finally, the anterior two thirds of the amygdala (S d, e). These active points either lie in regions adjacent to those yielding mewing during passage of current, or in structures which are anatomically directly or indirectly connected with the focal field for mewing described previously. Substratum for pattern suggestive of reaction to pain
It is difficultto assign a specific substratum to responses suggesting reaction to pain, as such responses were never obtained as threshold answers from the midbrain, hypothalamus or forebrain. The most convincing reaction expressing pain (illustrated in Fig. Sa) was evoked from ventral parts of the periaqueductal gray matter, a region which according to Walker (1942) and Morin et al. (1951) receives spinoreticular fibres that accompany the spinotectal and spinothalamic tracts. It may be possible that this ventral region constitutes a relay station for afferent nociceptive fibres, or an ontophylogenetically ancient system subserving pain reactions. These aspects, and the somewhat different findings of Spiegel et ~ l(1954) . in the cat, have been discussed in the review of Hunsperger (1963, pp. 52, 53). SUMMARY
(1) The data presented show that the affective patterns elicited by electrical stimulation of the forebrain and brain stem in unanaesthetized freely-moving cats are threat
AFFECTIVE BEHAVIOUR
125
and flight, mewing of various types, and an effect suggestive of reaction to pain. (2) The threat and flight reactions and mewing are obtained ast hreshold answers. All three can be assigned to specific areas; the reaction suggesting pain perception, however, is only obtained with strong stimulation from points lying within the responsive fields for mewing or threat. (3) The affective patterns obtained from the midbrain, hypothalamus and forebrain, although grossly similar, vary according to the level stimulated. (4) Seismographic recording of tonic muscular activity and of heart beats supplies further information on the somato-motor and autonomic changes associated with threat and flight. (5) The threat and flight responses elicited from the hypothalamus show adaptation to changes in environmental conditions. REFERENCES ABRAHAMS, V. C., HILTON,S. M., AND ZBRO~YNA, A., (1960); Active muscle vasodilatation produced by stimulation of the brain stem: Its significance in the defence reaction. J. Physiol. (Lond.),154, 491-5 13. ANDY,0.J., AND AKERT,K., (1955); Seizure patterns induced by electrical stimulation of hippocampal formation in the cat. J. Neuropath. exp. Neurol., 14, 198-213. BURGI,S., UND BUCHER, V. M.. (1960); Markhaltige Faserverbindungen im Hirnstamm der Katze. Monographien aus dem Gesamtgebiet der Neurologie und Psychiatrie. Fasc. 87, Berlin, Springer. BURGI,S., AND BUCHER,V. M., (1963); Stria terminalis and related structures. Progress in Brain Research Vol. 3. The Rhinencephalon and Related Structures, W. Bargmann and J. P. Schade, Editors, Amsterdam, Elsevier pp. 163-169. J., (1955); Hippocampe et Epilepsie: A propos d'une Sirie d'Expiriences sur le Cobayeet CADHILLAC, le Chat et de I'Exploration Electrique de la Corne d'dmmon chez 1'Homme. Montpellier, Paul Dehan. W. B., (1939); The Wisdon of the Body. New York, Norton. CANNON, CANNON, W. B., (1953); Bodily Changes in Pain, Hunger, Fear and Rage. 2nd ed. Boston, Branford. CORTI,U. A., GASSMANN F., UND WEBER, M., (1955); Unruhebestimmung bei Menschen und Tieren. Verh. Schweiz. Naturforsch. Ges., Pruntrut, 164-167. R. W., UND WYSS,0. A. M., (1961); Registrierung der vertikalen KraftCORTI, U. A., HUNSPERGER, komponente der endogenen Korpererschiitterung (Haltetonus). Reaktionstisch fir Tierversuche. Pfliigers. Arch, ges. Physiol., 274, 95. FANGEL, CH., AND KAADA, B. R., (1960); Behavior 'attention' and fear induced by cortical stimulation in the cat. Electroenceph. clin. Neurophysiol., 12, 575-588. FERNANDEZ DE MOLINA, A., AND HUNSPERGER, R. W., (1959); Central representation of affective reactions in forebrain and brainstem : Electrical stimulation of amygdala, stria terminalis, and adjacent structures. J. Physiol. (Lond.), 145, 251-269. FERNANDEZ DE MOLINA, A., AND HUNSPERGER, R. W., (1962); Organization of the subcortical system governing defence and flight reaction in the cat. J. Physiol. (Lond.), 160, 200-213. Fox, C. A., (1943): The stria terminalis, longitudinal association bundle and precommissural fornix fibers in the cat. J. comp. Neurol., 79, 277-295. HESS,W. R., (1932); Beitriige zur Physiologie des Hirnstamms. I . Die Methodik der lokalisierten Reizung und Ausschaltung subkortikaler Hirnabschnitte. Leipzig, Thieme HESS,W. R., (1949); Das Zwischenhirn. Basel, Schwabe. HESS,W. R., (1957); The Functional Organization of the Diencephalon.New York, Grune and Stratton. HESS,W. R., UND BRUGGER,M., (1943); Das subkortikale Zentrum der affektiven Abwehrreaktion. Helv. physiol. pharmacol. Acta, 1, 33-52. V., (1945/46); Zur Physiologievon Hypothalamus, Area HESS,W. R., BRUGGER,M., UND BUCHER, praeoptica und Septum, sowie angrenzender Balken- und Stirnhirnbereiche. Mschr. Psychiat. Neurol., 3, 17-59. A. W., (1963); Amygdaloid region for defence reactions and its HILTON,S. M., AND ZBROZYNA, efferent pathway to the brainstem. J. Physiol. (Lond.), 165, 160-173.
126
R. W. HUNSPERGER A N D V. M. B U C H E R
HUNSPERGER, R. W., (1956); Affektreaktionen auf elektrische Reizung im Hirnstamm der Katze. Helv. physiol. pharmaeol. Acta, 14, 70-92. HUNSPERGER, R. W., (1963); Comportements affectifs provoques par la stimulation electrique du tronc drkbral et du cerveau anterieur. J. Physiol. (Paris), 55, 45-97. HUNSPERGER, R. W., (1965); Mektreaktionen auf Mittelfrequenzreizung (5000 Hz) im Hypothalamus der Katze. Helv. physiol. pharmacol. Acta, 23, C25-C28. HUNSPERGER, R. W., UND DIETIKER,M., (1962); Die Methodik der Registrierung der endogenen Korpererschiitterung (Haltetonus) an der Katze. Isolierung von Storschwingungen. Helv. physiol. pharmacol. Acta, 20, C7-C9. HUNSPERGEX, R. W., UND G w ~ Z D Z ,B., (1964); Lokalisation zentralnervoser Strukturen fur Miauen auf Grund von Hirnreizversuchen. Verh. Schweiz. Naturforsch. Ges. Zurich, 228-238. HUNTER,J., (1950); Further observations on subcortically induced epileptic attacks. Electroenceph. clin. Neurophysiol., 2, 193-201. HUNTER,J., AND JASPER, H. H., (1949); Effects of thalamic stimulation in unanaesthetized animals. The arrest reaction and petit mal-like seizures, activation patterns and generalized convulsions. Electroenceph. elin. Neurophysiol., 1, 305-324. JOHNSTON, J. B., (1923); Further contributions to the study of the evolution of the forebrain. J. comp. Neurol., 35, 337481. KAADA,B. R., JANSEN, J. JR., AND ANDERSEN, P., (1953); Stimulation of the hippocampus and medial cortical areas in unanesthetized cats. Neurology, 3, 844-857. KAADA,B. R., ANDERSEN, P., AND JANSEN,J. JR., (1954); Stimulation of the amygdaloid nuclear complex in unanesthetized cats. Neurology, 4,48-64. KARLI,P., ET VERGNES,M., (1964); Nouvelles donnks sur les bases neurophysiologiques du comportement d’agression intersgcifique rat-souris. J. Physiol. (Paris), 56, 384. KARPLUS, J. P., UND KREIDL,A., (1909); Gehirn und Sympathicus. I. Zwischenhirnbasis und Halssympathicus. Ppigers Arch. ges. Physiol., 129, 138-144. KARPLUS,J. P., UND KRUDL, A., (1910); Gehirn und Sympathicus. 11. Ein Sympathicuszentrum im Zwischenhm. Pfliigers Arch. ges. Physiol., 125,401416. KARPLUS, J. P., UND KREIDL,A., (1928);Gehirn und Sympathicus. VIII. (1) Zur zentralen Regulierung der Irisbewegungen; (2) Bemerkungen zur Schmenemphdlichkeit der vegetativen Hypothalamuszentren. Pj7iigers Arch. ges. Physiol., 219, 613-618. KOIKEGAMI,H., AND YOSHLDA,K., (1953); Pupillary dilatation induced by stimulation of the amygdaloid nuclei. Folia psychiat. neurol. jap., 7, 109-126. LISSAK,K., GRASTYAN, E., CSANAKY, A., UKESI, F., AND VEREBY, GY., (1957); A study of hippocampal function in the waking and sleeping animal with chronically implanted electrodes. Acta physiol. pharmacol. neerl., 6, 415459. MACLEAN, P. D., (1955a); The limbic system (‘visceral brain’) and emotional behavior. Arch. Neurol. Psychiat. (Chic.), 73, 13CL134. MA CLEAN,^. D.,(l957b); Chemical and electrical stimulationof hippocampus in unrestrainedanimals. I. Methods and electroencephalographic findings. Arch. Neurol. Psychiat. (Chic.), 78, 113-127. MACLEAN, P. D., (1957); Chemical and electrical stimulation of hippocampus in unrestrained animals. II. Behavioral findings. Arch. Neurol. Psychiat. (Chic.), 78, 128-142. MACLEAN, P. D., AND DELGADO, J. M. R., (1953); Electrical and chemical stimulation of frontotemporal portions of limbic system in the awaking animal. Electroenceph. clin. Neurophysiol., 5, 91-100. MAGNUS, O., AND LAMMERS, H. J., (1956); The amygdaloid nuclear complex. Part I: Electrical stimulation of the amygdala and periamygdaloid cortex in the waking cat. Fofia psychiat. neerl., 55, 555-581. M o m , F., SCHWARTZ, H. G., AND O’LEARY,J. L., (1951); Experimental study of the spino-thalamic and related tracts. Acta Psychiat. (Kbh.), 26, 371-396. NAKAO, H., (1958); Emotional behavior produced by hypothalamic stimulation. Amer. J . Physiol., 194,411418. NAQUET,R., (1953); Sur les fonctions du rhinenckphale d’aprb les resultats de la stimulation chez le chat. Thesis, Marseille. NAUTA,W. J. H., (1961); Fibre degeneration following lesions of the amygdaloid complex in the monkey. J. Anat., 95, 515-531. PARME~IANI, P. L., (1960); Reizeffekte aus Hippocampus und Corpus mammillare der Katze. Helv. physiol. pharmaeol. Acta, 18, 523-536. SANO,T., (1958); Motor and other responses elicited by electrical stimulation of the cat’s temporal
AFFECTIVE BEHAVIOUR
127
lobe. Folia psychiat. neurol. jap., 12, 152-176. SHEALY, c.W., AND PEELE,T. L., (1957); Studies on amygdaloid nucleus of cat. J. Neurophysiol., 20, 125-139. SKULTETY, F. M., (1963); Sti nulation of periaqueductal gray and hypothalamus. Arch. Neurol. Psych iat. (Chic.). 8, 608-620. SPIEGEL, A. E., KLETZKIN, M., AND SZEKELY, E. G., (1954); Pain reactions upon stimulation of the tectum mesencephali. J. Neuropath. exp. Neurol., 13, 212-220. URSDT, H., AND KAADA, B. R., (1960); Functional localization within the amygdaloid complex in the cat. Electroenceph. clin. Neurophysiol., 12, 1-20. E. S., AND NAUTA,W. J. H., (1959); A comparison of the distribution of the fornix VALENSTEIN, system in the rat, guinea-pig, cat and monkey. J. comp. Neurol., 113, 337-363. A. E., (1942); The somatotopical localization of the spinothalamic and secondary trigeminal WALKER, tracts in the mesencephalon. Arch. Neurol. Psychiaf. (Chic.), 48, 884-889. A,, (1902); Das basale Riechbundel des Kaninchens. Anat. Anz., 20, 175-1 87. WALLENBERG, WOOD,CH. D., (1958); Behavior changes following discrete lesions of temporal lobe structures. Neurology, 8, 215-220. WYSS, 0. A. M., (1950); Beitrage zur elektrophysiologischen Methodik. 11. Ein vereinfachtes Reizgerat fiir unabhangige Veranderung von Frequenz und Dauer der Impulse. Helv. physiol.pharmaco1. Acta, 8, 18-24. WYSS,0.A. M., (1963); Die Reizwirkung mittelfrequenter Wechselstrome. Helv. physiol. pharmacol. Acta, 21, 173-188. WYSS,0.A. M., (1965); Beitrage zur elektrophysiologischen Methodik. V. Ein Reizgerat zur konventionellen Impulsreizung. Helv. physiol. pharmacol. Acra, 23, 26-30. YASUKOCHI, O., (1960); Emotional responses elicited by electrical stimulation of the hypothalamus in the cat. Folia Psychiat. Neurol. h p . , 14, 2-267.
128
Facilitation and Inhibition in Centrally Induced Switch-off Behavior in Cats HIROYUKI NAKAO Department of Neuropsychiatry, Kyushu University School of Medicine, Fukuoka (Japan)
(1) Switch-off behavior ( S O B ) An experimental box designed for escape learning, as shown in Fig. 1, has been used throughout the present studies. This box has two small windows in a wall. In front of
Fig. 1 . Apparatus used for the switch-off behavior. The cat switches off the stimulation in response to central stimulation(0)yieldmg a flight response, or to sensory stimulation(0).A buzzer is interrupted by the animal trained with the buzzer as the warning signal of a grid shock.
one window is a plate, pushing of which breaks the stimulation circuit. The animal placed in the box can be trained to push the plate to turn off the stimulation delivered to himself. This type of response will be designated as the switch-off behavior (SOB) in the present experiment.
( 2 ) Switch-off behavior induced by sensory stimulation The animal placed in the experimental box learns to push the plate to terminate a grid shock to the feet or a warning buzzer of the grid shock (Nakao, 1958). Experiments have been carried out to see whether brain stimulation can be substituted for stimulation of sensory receptors in the SOB. Delgado et al. (1954) have shown that
S WITCH-OFF BEHAVIOR I N CATS
129
pain-like responses accompanied by a strong aversive drive can be elicited by stimulation in the neighborhood of the medial lemniscus or the spinothalamic tract. In recent experiments (Aoki and Nakao, 1965), it was found that medial lemniscal stimulation at pontine levels evoked the SOB (Fig. 2C).
( 3 ) Hypothalamic switch-off behavior Hess (1949) demonstrated in cats that flight reactions develop during hypothalamic stimulation. This result suggests that the SOB may occur when a flight response is produced by hypothalamic stimulation. The animal with a hypothalamic electrode that could give rise to a flight response, was placed in the experimental box, and the hypothalamus was stimulated monopolarly. The animal pushed the plate while trying to get out of an inadequate opening of a window, switching off the hypothalamic stimulation. After such an accidental pushing had been repeated several times, the animal learned to remove the hypothalamic stimulation by pushing the plate. Studies revealed that the response time, i.e. the time from the onset of stimulation to switching off, decreased to a certain level after training trials had been repeated a few hundred times. The response time was constant with constant intensity, and increased as the intensity decreased. The term hypothalamic SOB will be applied to the SOB induced by hypothalamic stimulation. The response was considered to be positive if an animal maintained a highly stable value of the response time during several hundred trials. If a response time increased or remained unstable, the response was discarded in the present study. Mapping studies showed that the sites yielding the SOB fall medially in the hypothalamus except for the ventromedial region. A low threshold area was found in the anterior hypothalamic nucleus (Fig. 2A). ( 4 ) Mesencephalic switch-off behavior Studies carried out by stimulation of the brains of freely moving cats indicated that emotional responses were obtained not only from the hypothalamus but also from the central gray matter of the midbrain (Spiegel et al., 1954; Delgado, 1955; Hunsperger, 1956; Skultety, 1963). In these reports there was a diversity of opinion concerning the nature of the response elicited on electrical stimulation of the central gray matter. The purpose of the present study was to test whether the flight response from the midbrain can be used to motivate the SOB. At first, more than 400 points in the border between the hypothalamus and the midbrain, and in the dorsal half of the midbrain, were stimulated in order to find the points yielding the SOB. It was found from these pilot experiments that a few points within the central gray matter at the middle and caudal portions could elicit the SOB. The next experiments were concentrated on the central gray matter of the midbrain to determine more precisely the anatomical structures involved in the SOB. For this purpose more than 300 points were explored within this area. The points evoking the SOB were found within the points inducing forward locomotion. The localization of the electrodes is presented in Fig. 2B. References p . 143
130
H. N A K A O
A
B
Fig. 2. Representative sections of the cat’s brain to show the points stimulated to induce the switchoff behavior in the hypothalamus (A), the midbrain (B), and the pons (C). Numbers above sections indicate the anterior and posterior stereotaxic plane. Abbreviations: F,fornix; Mb, mammillary body; ML, medial lemniscus; NO, nucleus of oculomotor nerve; NT, nucleus of trochlear nerve; NvT, trochlear nerve; OT, optic tract; Vm, nucleus of ventromedialis.
It has been shown that pain-like responses can be evoked by stimulation of the central gray matter (Spiegel et al., 1954; Delgado, 1955). In the SOB experiments cats with electrodes yielding pain-like responses, as shown by high-pitched screeching,were not used. (5) Effect of stimulation of basal forebrain areas an hypothalamic switch-off behavior
In these experiments, the influence of basal forebrain stimulation on the response time of the hypothalamic SOB was studied. Double stimulation is the overlap of a proper stimulus for the hypothalamic SOB and an additional stimulus to another
131
S W I T C H - O F F BEHAVIOR I N CATS
Fig. 3. Theeffect ofstimulationofasecondareaon thehypothalamic switch-off behavior (s). Symbols: A,complete inhibition, i.e. no performance; A,marked inhibition; A, slight inhibition; x, no effect; 0, indefinite effect; 0 , marked facilitation; 0, slight facilitation. Abbreviations: fx, fornix; vm, nucleus of ventromedialis.
0L-
1 2 3 4 s
---1 2 3 4 s
1 2 3 4 s
1 2 3 4 5
1 2 3 4 5 trial
Fig. 4. The effects of additional stimulation on the response time of the switch-off behavior. The location of the electrodes and the intensities of additional stimulation: A, lateral hypothalamus (15.5), 0.5 V; B, lateral hypothalamus (13.0), 1 V; C, diagonal band (17.0), 1 V; D, anterior hypothalamic area (13.5), 1 V; E, anterior hypothalamic area (13.5), 0.5 V, 60 c/s A. C., each. Proper stimulation of the switch-off behavior was applied in anteromedial portion of the hypothalamus. The intensities of proper stimulation: A, 0.5 V; B, 0.5 V; C, 1 V; D, 1 V; E, 0.8 V, 60 c/s,A. C., each. Crosses show no performance of the response. References p . I43
132
H. N A K A O
point. The additional stimulus was started 1 sec before the onset of the proper stimulus, and was adjusted to subthreshold strength when applied to medial hypothalamic points from which the SOB could be elicited. At other sites the stimulus intensity was adjusted to threshold intensity, when it evoked overt responses. The animal was able to turn off both stimuli by pushing a plate in the experimental box. Response time was measured to the proper stimulus. An experimenterinterrupted the stimulation if the animal failed to switch off the stimulation during 10 sec, and the trial was qualified as no performance. The results are illustrated in Fig. 3. Stimulation of the medial hypothalamic area with the exception of the ventromedial nucleus had a facilitatory effect on the hypothalamic SOB (Fig. 4D, E). Stimulation of the ventromedial nucleus caused a reaction accompanied by snarling and other full rage manifestations and, when applied as an additional stimulus, resulted in a complete inhibition of the hypothalamic SOB. It has been reported that the lateral hypothalamus and the ventromedial hypothalamus mediate opposing effects on feeding and self-stimulation (Anand, 1961; Hoebel and Teitelbaum, 1962; Olds, 1962). Stimulation of the lateral hypothalamus, however, inhibited the hypothalamic SOB (Fig. 4A, B) as ventromedial stimulation did. Thus, both areas seemed to have the same effect in regard to the SOB, but the mechanism of the inhibition was different. The SOB was disturbed by exploratory behavior during stimulation of the lateral hypothalamus, while the SOB was inhibited by cessation of locomotion during ventromedial stimulation. Among lateral hypothalamic points which were expected to have an inhibitory effect, some points induced a facilitatory effect on the hypothalamic SOB. This may be due to initiation of a forward movement elicited by lateral hypothalamic stimulation. A zone situated at the base of the brain, just rostra1 to the optic chiasma, which has been termed the basal forebrain synchronizing area (Sterman and Clemente, 1962), showed an inhibitory effect on the hypothalamic SOB (Fig. 4C). This effect was induced by unilateral stimulation of the area which caused sniffing and exploring. Although it is hard to determine whether the change in response time in double stimulation is attributed to modification of the reinforcing property or of some performance capacity, it seems very likely that changes in the intensity or the characteristic of the reinforcing play a main role. ( 6 ) Hypothalamic switch-oJ behavior and limbic after-discharges
Stimulation of the areas described above showed stimulus-bound effects on the hypothalamic SOB. The situation is different in limbic structures, because their stimulation readily causes after-discharges wherever stimulated, and when the discharges are induced it has a tendency to spread throughout limbic structures and even the whole brain (Goodfellow and Niemer, 1961; Walker and Udvarhelyl, 1965). The only exception is the anterior cingulate gyrus whose after-discharges seldom propagate to other areas (Nakao, 1963). Observation of animals during such after-discharges has shown that there is little or no evidence of convulsive phenomena in the body’s musculature despite the intense
7
1
m
i References p . 143
S W I T C H - O F F B E H A V I O R IN C A T S
133
134
H. N A K A O
discharges. It has, however, been reported that there is an apparent loss of an animal’s ability to perform an avoidance task to respond to sensory stimulation during hippocampal after-discharges (Delgado and Sevillano, 1961). It is therefore of interest to assess the effects of after-discharges from some limbic foci on the centrally-induced SOB. Bipolar stimulation and recording were made in the following studies. (a) Hippocampal after-discharges The first of this series of experiments dealt with the effects of dorsal hippocampal after-discharges on the hypothalamic SOB (Nakao, 1962). Symptoms of hippocampal after-discharges have been described in detail (Andy and Akert, 1955; Delgado and Sevillano, 1961). Slight changes in appearance of the animal were produced until the after-discharges propagated to the ipsilateral amygdala. During the after-discharges, which were localized in the dorsal hippocampus, there was no appreciable alteration of the hypothalamic SOB. When there was propagation of the after-discharges to the amygdala, a partial loss of the performance ability occurred (Fig. 5E, F, G). As the discharges developed in the amygdala, ipsilateral facial twitchings were observed, often with salivation. Sometimes the after-discharges in the amygdala suddenly changed into the typical amygdala 4-6 c/sec spiking called the reactive after-discharge (Delgado and Sevillano, 1961), coinciding with the start of masticatory movements. A complete loss of the performance of the hypothalamic SOB was observed when the reactive after-discharges appeared in the amygdala (Fig. 51, J). Our observations of the appearance of these animals indicated that the responses were impaired to some extent in the twitching phase and completely inhibited in the masticatory phase of hippocampal after-discharges. These results appear to indicate that the active participation of the amygdala inhibits the hypothalamic SOB during hippocampal after-discharges, because facial motor effects have a clear correlation with the activity of the amygdala.
(b) Amygdaloid after-discharges
Amygdaloid stimulation commonly produces immediate results, namely ipsilateral facial movements and often masticatory movements accompanying slow spiking afterdischarges. The animal, contrary to expectation, maintained normal performance during the slow after-discharges in the amygdala which corresponded to the reactive after-discharges of the amygdala observed during hippocampal after-discharges (Fig. 6B, C, D, E). In other words, twitching and mastication, which are correlated with inhibition of the hypothalamic SOB during hippocampal after-discharges, are not accompanied by inhibition of the hypothalamic SOB during amygdaloid afterdischarges. When fast activity was superimposed on the propagated slow after-discharges in the ipsilateral dorsal hippocampus, an increased response time was often observed, and the severity of the deterioration increased as the fast discharges developed in the hippocampus. After the propagated hippocampal after-discharges became fast discharges, the animal lost its performance ability (Nakao and Yoshida, 1963) as illus-
SWITCH-OFF BEHAVIOR I N CATS
References p . 143
135
136 H. N A K A O
SWITCH-OFF BEHAVIOR I N CATS
137
a I-1LGv . -a
1202-6-13-65
Fig. 8. The effects of hypothalamic after-discharges on the hypothalamic switch-off behavior. The abbreviations are the same as in Fig. 5.
trated in Fig. 6F. The present result reveals that the functional deficits of after-discharges do not always depend on the slow discharges of the amygdala.
Cingulate after-discharges Anterior cingulate stimulation induces short after-discharges which are localized in the cingulate gyms and which may propagate slightly to other areas. Posterior cingulate stimulation induces after-discharges which propagate to other structures. Localized after-discharges did not disturb the hypothalamic SOB performance (Fig. 7,2B), but the after-discharges with some propagation in the dorsal hippocampus caused an increase in response time. The response was abolished when marked after-discharges appeared in the dorsal hippocampus accompanied by some propagated after-discharges in the amygdala (Fig. 7, IB). It was sometimes observed that after cessation of cingulate after-discharges, propagated after-discharges in the dorsal hippocampus (c)
References p. 143
lreC
138 H. N A K A O
SWITCH-OFF BEHAVIOR IN CATS
139
continued to develop, or spontaneous after-discharges started anew in the dorsal hippocampus as if the dorsal hippocampus were stimulated, and the after-discharges developed and propagated. The effect of these after-discharges on the hypothalamic SOB was similar to the effect of hippocampal after-discharges induced by electrical stimulation of the dorsal hippocampus (Nakao, 1964). ( d ) Hypothalamic after-discharges The site within hypothalamus which upon stimulation elicits after-discharges is the fornix (Nakao, 1963).As the after-discharges propagate to the hippocampus and then to the amygdala, no localized after-discharges in the hypothalamus are obtained. The behavioral responses during the after-discharges, and the inhibitory effects on the hypothalamic SOB were similar to those observed in hippocampal after-discharges. The results of this experiment are shown in Fig. 8. The animal regularly switched off the hypothalamic stimulation during hypothalamic after-discharges (Fig. 8, 1); it therefore was tempting to suggest that hypothalamic after-discharges do not affect the hypothalamic SOB. Hypothalamic after-discharges, however, disrupted the response when the after-discharges developed fully in the amygdala (Fig. 8, 2G).
( 7 ) Mesencephalic switch-of behavior and limbic after-discharges (a) Hippocampal after-discharges
Fig. 9 shows one instance in which a complete inhibition of the mesencephalic SOB occurred during hippocampal after-discharges (Fig. 9E, F).This impairment is attributed to full propagation of the after-discharges into the amygdala, especially to the appearance of slow discharges in the amygdala. Similar results have been obtained in a study of the hypothalamic SOB, as described above. ( b ) Amygdaloid after-discharges A complete performance of the mesencephalic SOB was observed during propagated after-discharges in the hippocampus which were synchronized with the after-discharges in the amygdala and were evoked by amygdaloid stimulation. (Fig. 10, IB, C). The appearance of fast discharges in the propagated after-discharges in the hippocampus was an indication of complete loss of the performance ability (Fig. 10, lD, 2B, C). The findings observed here were also found in the hypothalamic SOB. ( 8 ) Comparison of inhibitory effects of limbic after-discharges on hypothalamic and
mesencephalic switch-off behavior
This study has been extended to investigate whether or not a distinction can be made between the inhibitory effects of limbic after-discharges upon hypothalamic and mesencephalic SOB. Animals were trained to push a plate either on hypothalamic stimulation or onmesencephalic stimulation. These stimulations were delivered alternately before, during, and after limbic after-discharges. References p. 143
140
References pp. 142-143
H. N A K A O
I
.
S W I T C H - O F F BEHAVIOR I N C A T S
141
Fig. 11. Comparison of the hypothalamic switch-off behavior and the mesencephalicswitch-off behavior during hippocampal after-discharges.The conventions are the same as in Figs. 5 and 9.
The inhibitory effects on each SOB can appear in various combinations. Inhibition intensity on each SOB could not be compared unless one response still remained after abolition of the other. In general, the inhibition increased as time passed during afterdischarges. If the mesencephalic SOB suffers from limbic after-discharges much more than the hypothalamic SOB, the latter will be performed after abolition of the mesencephalic SOB. The converse is true if the mesencephalic SOB remains after the hypothalamic SOB is abolished. Sixteen animals were prepared to compare these two kinds of SOB under limbic after-discharges. Results obtained from 12 of them are available for comparison. With only one exception, the animals performed the hypothalamic SOB after the mesencephalic SOB was abolished during hippocampal (Fig. 11G) or amygdaloid after-discharges (Fig. 12G, I). The results indicate that the hypothalamic SOB resists hippocampal or amygdaloid after-discharges more than the mesencephalic SOB. References p.!143
142
H. N A K A O
Fig. 12. Comparison of the hypothalamic switch-off behavior and the mesencephalic switch-off behavior during amygdaloid after-discharges. The conventions are the same as in Figs. 5 and 9
(9) Summary
This report is concerned with a learned response in which the cat pushes a plate to terminate the stimulation applied. This response is termed the switch-off behavior (SOB). Studies have been carried out on the neural mechanism that affects the SOB. The findings were as follows: (1) The hypothalamically induced SOB (hypothalamic SOB) was facilitated by stimulation of the medial hypothalamus except the ventromedial nucleus, and inhibited by stimulation of the lateral hypothalamus, the ventromedial nucleus, or the anterior portion of the basal forebrain areas. (2) Limbic after-discharges initiated from the hippocampus, the amygdala, the cingulate gyrus, or the hypothalamus inhibited the hypothalamic SOB to various extents.
SWIT CH - O FF BEHAVIOR I N C A T S
143
(3) Similar results were obtained with the mesencephalic SOB during hippocampal or amygdaloid after-discharges. (4) The inhibitory effects of the hippocampal or amygdaloid after-discharges on the mesencephalic SOB were more severe than those found on the hypothalamic SOB. REFERENCES ANAND,B. K., (1961); Nervous regulation of food intake. Physiol. Rev., 41, 677-708. ANDY,0. J., AND AKERT,K., (1955);Seizurepatternsinduced by electrical stimulation of hippocampal formation in the cat. J. Neuropath. exp. Neurol., 14, 198-213. AOn, I., AND NAKAO,H., (1967); To be published. DELGADO, J. M. R., (1955); Cerebral structures involved in transmission and elaboration of noxious stimulation J. Neurophysiol., 18, 261-275. DELGADO, J. M. R., ROBERTS,W. W., AND MILLER,N. E., (1954); Learning motivated by electrical stimulation of the brain. Amer. J. Physiol., 179, 587-593. J. M. R., AND SEWLLANO, M., (1961); Evolution of repeated hippocampal seizures in the DELGADO, cat. Electroenceph. clin. Neurophysiol., 13, 722-733. GOODFELLOW, E. F., AND NIEMER, W. T., (1961); The spreadof after-dischargefromstimulationofthe rhinencephalon in the cat. Electroenceph. clin. Neurophysiol., 13, 710-721. HESS,W. R., (1949); Das Zwischenhirn : Syndrome, Lokalisationen, Funktionen, Basel, Schwabe. HOEBEL, B. G., AND TEITELBAUM, P., (1962); Hypothalamic control of feeding and self-stimulation. Science, 135, 375-377. HUNSPERGER, R. W., (1956); Affektreaktionen auf elektrische Reizung im Hirnstamm der Katze. Helv. physiol. Pharnaacol. Acta, 14,70-92. NAKAO,H., (1958); Emotional behavior produced by hypothalamic stimulation. Amer. J. Physiol., 194,411-418. NAKAO,H., (1958); Study of a learned behavior motivated by hypothalamic stimulation in cats. Sei-shin-kei-shi, 60, 1396-1401 (In Japanese). NAKAO,H., (1962); The spread of hippocampal after-discharges and the performance of switch-off behavior motivated by hypothalamic stimulation in cats. Fol. psychiat. neurol. yap., 16, 168-180. NAKAO,H., (1963); Learned behavior motivated by hypothalamic stimulation and brain stimulation and brain lesion in cats. No-shinkei., 15, 1117-1129 (In Japanese). NAKAO,H., (1964); The spread of cingulate after-discharges and the performance of switch-off behavior motivated by hypothalamic stimulation in cats. Fol. psychiat. neurol. jap., 18, 153-160. NAKAO,H., AND YOSHIDA, M., (1963); Effect of amygdaloid after discharges on switch-off behavior motivated by hypothalamic stimulation in cats. Fol. psychiat. neurol. Jap., 17, 221-229. OLDS,J., (1962); Hypothalamic substrates of reward. Physiol. Rev., 42, 554-604. SKULTETY, F. M., (1963); Stimulation of periaqueductal gray and hypothalamus. Arch. Neurol., 8,608-620. SPIEGEL,E. A., KLETZKIN, M., AND SZEKELY, E. G., (1954); Pain reactions upon stimulation of the tectum mesencephali. J. Neuropath. exp. Neurol., 13, 212-220. C. D., (1962); Forebrain inhibitory mechanisms: sleep patterns inSTERMAN, M. B., AND CLEMENTE, duced by basal forebrain stimulation in the behaving cat. Exp. Neurol., 6, 103-117. WALKER, A. E., AND UDVARHELYL, G. B., (1965); Dissemination of acute focal seizures in the monkey. II. From subcortical foci. Arch. Neurol., 12, 357-380.
144
The Limbic System and Behavioral Reinforcement JAMES OLDS Department of Psychology, The University of Michigan, Ann Arbor, Mich. (U.S.A.)
What neurohs of the brain are involved in voluntary actiohs and how are they involved? Evarts (1964) discovered patterns of activity of pyramidal neurons characteristic of sleep, paradoxical sleep, waking, and activity. Can observations of this type be extended to permit an understanding of differences in pattern in these and other neural systems which occur while a neuron is involved in an ongoing voluntary act? Prelirhinary experiments reported here suggest that we will be able to make the appropriate observations. Because the neurons involved in any particular pattern of voluntary behavior might be hard to find in the course of normal microelectrode explorations, a method was adopted to circumvent the difficulty. To assure that a neuron under study would be involved in the final voluntary pattern, a neural discharge pattern itself was chosen as the ‘behavior’ to be reinforced in a conditioning experiment. Under these circumstances, if neurons at a recording site could participate in a voluntary pattern, then some behavior would be ‘shaped’ by these procedures whose performance involved an increment in the pace of a neuron under study. The questions that could then be asked concerned (1) the differences in firing pattern as between ‘voluntary’ and ‘involuntary’ bursts of discharges of the neuron, (2) concomitant variation in neighboring neurons, (3) changes in degree of voluntary control attainable contingent on variation in peripheral stimulation, (4) the relative difficulty involved in bringing neurons of different anatomical structures under voluntary control, (5) the behavior patterns associated with the voluntary control of neural patterns in particular brain areas. METHOD
In these experiments, patterns of neural activity were recorded from the brain during periods of ‘quiet waking’ by means of implanted microelectrodes. Patterns which occurred rarely at the outset were ‘conditioned’ to occur more frequently by methods of operant conditioning. Afterwards, these conditioned brain responses were brought under control of ‘discriminative’ stimuli so that periods of operant neural behavior could be alternated with control periods at the will of the experimenter. This permitted observation of the neural and behavioral concomitants of these neural responses when they occurred as components of voluntary behaviors. The neural concomitants were recorded simultaneously from other implanted microelectrodes, some nearby in the
T H E L I M B I C SYSTEM A N D B E H A V I O R A L R E I N F O R C E M E N T
145
same structure, other farther away in the brain ; behavioral concomitants were recorded cinematographically. The method also permitted observation of pattern differences between the activity of neurons in the ‘random’ state and activity of the same neurons in the ‘operant’ state. The methods used have been described in more detail in a previous paper (Olds, 1965).
( A ) Preparation (1) Nichrome wires of 67 p diameter with formvar insulation of about 10 p thickness were used as microelectrodes. The wires were insulated at the factory (Driver-Harris Company, Harrison, N. J. ( U S A . ) or Johnson-Matthey Co., Ltd., 73-83 Hatton Garden, London, E C1, England) and cut with scissors. The cut cross section of the tip was the uninsulated portion of the probe. (2) Microelectrodes were implanted in ventral midbrain, aimed at the pons, and in dorsal forebrain, aimed at the hippocampus, and in other parts of the brain for special tests. Stereotactic guidance was used to lower each probe to its target area, and electrophysiological guidance was then used to bring it into a region of recordable neural activity. Recurrent negative spikes of constant amplitude (200 to 500 p V ) with signal to noise ratio of more than 2 to 1, and with duration of 0.2 to 0.7 msec were taken to be single unit responses if they were dependent on movement of the microelectrodes in a 10 to 100 p range. If one or several unit responses in the target area were recordable for a 15-min period, the probe was fixed to the skull by means of acrylic. A screw which had been placed in the skull previously, and a Z bend previously placed in the microelectrode wire helped to hold the acrylic to the skull and the wire to the acrylic. Four to nine microelectrodes were placed in each rat. These plus a larger, uninsulated ground electrode which was also planted in the brain were attached to a 10 contact plug which was also fixed to the skull with acrylic. (3) A test for stable recording was made 1 to 2 weeks after surgery. A 12 inch, 10lead cord of microdot cable was fixed a t its lower end to the 10 contact plug. At the upper end it was fixed to a 10 wire swivel commutator which was mounted on a counterbalanced arm. High-impedance, solid-state preamplifiers were also mounted on the counterbalanced arm. The animal was placed in a cylindrical plastic cage which was 10 to 15 inches in diameter and 11 inches high; the counterbalanced arm was mounted above. The animal was relatively free to move around. All probes were tested to find whether unit responses were still recordable. If unit responses were observed in recordings from a given probe a week after surgery they were usually recordable indefinitely. ( B ) Training This was begun about 2 weeks after surgery. (1) Pretraining: a subject was trained continuously for about 4 days in a movement detecting, feeding cage to remain relatively motionless in order to obtain food, and to go rapidly to eat when a pellet dispenser discharged, making a characteristic noise. (2) Setting of unit discriminators: the animal was then transferred again to the plastic recording cage with cable, commutator, and counterbalanced arm. One of the References p. I64
146
J. O L D S
unit responses or a set of indiscriminable unit responses recorded from a single probe was chosen for conditioning. The original choice was made by visual inspection of many ‘stopped’ traces on the cathode ray tube of a storage oscilloscope. A simple pattern recognition device was then set to respond whenever one of the chosen action potentials occurred. The device could be set to respond to the action potentials of one neuron if the response was large and well differentiated. More often, however, it was set to respond to several of the unit responses recorded from the same probe. These were then grouped together for the purposes of the ensuing experiment. The criteria for recognition were (1) minimum amplitude, (2) maximum amplitude, (3) minimum ‘fall’ time, and (4) maximum fall time. These criteria were specified by setting 4 potentiometers. The settings were adjusted repeatedly until almost all identified events were clearly within the chosen unit response class by visual inspection and almost all rejected events were clearly outside the chosen class. Thereafter the response of the pattern recognition device was taken as the definition of an occurrence of a unit response. When compared with human judgments of photographed wave forms, the pattern recognition device ‘correctly’ identified about 9 of 10 unit responses and misidentified about 1 of 100 interfering patterns, particularly those emanating from the muscles of the head. Because the unit responses were typically firing at rates in the 10/ sec range and head muscle artifacts during jaw movements were at a rate of 1000/sec, the 1 out of each 100 which were counted as units constituted a substantial source of interference. For this reason, the whole experiment was conducted during those periods when there was no detectable movement or muscle activity. Movement of the animal and muscle activity from jaws and head were detected directly by 3 detectors, 2 mechanical and 1 electrical. Electrical pulses from any one of these caused the analysers and the recorders to be blocked during the movement and for a period of 2 sec thereafter. Thus all recordings were made during artifact-free periods. Long-run records were made from the neurons by counting each pulse which met the amplitude and time-constant criteria of the primary pattern recognizer. An electronic counter was used to count all the spikes of one class; it was set to readout, reset and start again each time it reached a predetermined number. Each such readout made a spike on a moving paper record. The predetermined number was chosen on the basis of convenience between 2 and 127 with a view to obtaining a record with the maximum of detail compatible with a clear separation of spikes. The paper moved at 2 cm/min. As mentioned previously, both movement of the recording paper and counting of spikes were stopped automatically by an electronic gate whenever detectable movement of the animal occurred or when interfering muscle activity reached a criteria1 level. (3) Setting of the ‘burst’ discriminator: a secondary pattern recognition system was set to respond electrically whenever, during an artifact free period, a particularly high frequency of the chosen unit responses occurred. The frequency chosen was so high that it occurred by chance only once every 5 to 15 min during the first several hours of recording (see Fig. 1). The electrical response of this secondary pattern recognizer was used to trigger an event marking pen which recorded its occurrence. During con-
Fig. 1. Three series of 4 sec tracings; highest spikes = 500 pV. The series on theleft and that in themiddle were taken at random during an extinction period; they illustrate to some degree the unconditioned rate of the unit responses recorded from an experimental probe (in this case it was probe No. 6 medially placed in the hippocampus of rat 1588). The series on the right was taken by a method which photographed only bursts which metthecriterionofthe secondary pattern discriminator; in this case the criterion was 20 responses in a 4 sec period. The photographed burst was occurring ‘purposefully’during an acquisition period. Similar bursts sometimes occurred by ‘accident’ during extinction periods (see Fig. 10) but they tended to have a somewhat different pattern. The dots above the unit responses represent the output of the primary pattern discriminator; they indicate which spikes were counted as members of this particular unit group.
P 4
148
J. OLDS
ditioning periods, this response also triggered the food magazine or the application of a reinforcement stimulus. ( 4 ) Reinforcement training: during the first day, the burst discriminator was continuously coupled to the pellet dispenser and each of the high frequency patterns which occurred during an artifact-free period caused a 45 mg pellet to be presented to the animal. The high frequency burst discharged a pellet only if it started at least 2 sec after the last movement or artifact and providing there was a 100-msec period of artifact-free recording after the completion of the burst. This latter requirement was enforced by a delay of reinforcement device which permitted cancellation of reinforcement if artifact occurred during the delay interval. (5) ‘Discrimination’ training: during the second and all following days of conditioning, 2-min periods of acquisition were alternated with 8-min periods of extinction. The onset of each acquisition period was signaled by a stimulus change which was ordinarily sustained for the whole 2-min acquisition period. In these cases, the beginning of each extinction period was signaled by a reversal of the stimulus change. In special tests, only the onset of the acquisition period was signaled; the only indication of the onset of extinction in these cases was the failure of reinforcement. During acquisition periods the secondary pattern discriminator was coupled to the feeder or to a brain stimulator so that each high frequency burst was reinforced either by a pellet or by a $ sec train of brain stimulation (30 pA, 60 c/s, sine wave) in the hypothalamus which had been shown in pretests to have positively reinforcing value. Training was considered to be complete when the animal showed clear and regular discrimination of the acquisition signal by a stable augmentation in rate of the chosen unit pattern during each acquisition period. ( C ) Recordings
In addition to recordings from the unit responses chosen for conditioning, one to three control unit responses were simultaneously recorded and discriminated by means of other discriminators utilizing sometimes other brain probes, sometimes the same brain probe. These units were also counted electronically and marked in convenient multiples on separate channels of the moving record. ( D ) Histological analysis After experiments were completed, brains were cut in 50 p sections. Every other section was saved and mounted; from these, two series were prepared, one stained with Weil stain and the other with cresyl-violet. This permitted precise determination of the electrode track and estimation of the locus of the recording tip. An unusually high proportion of probes fell in or near their target. This was probably due to the clearly recognizable response patterns which often characterized different brain regions when the large fine wire microelectrodes were used. The characteristic responses served to assist the experimenter in homing on target structures during the neurophysiologically guided phase of the implantation procedures.
T H E LIMBIC SYSTEM A N D B E H A V I O R A L R E I N F O R C E M E N T
149
RESULTS
( A ) Histological findings The hippocampal probes were found to be in the ridge of pyramidal neurons in the hippocampal gyrus (see Fig. 2). The pontine probes were just posterior and ventral to the interpeduncular nucleus. Some were placed medially in the pontine gray, others were placed 1 to 3 mm away from the midline.
Fig. 2. Photomicrograph of Weil-stained section showing the track of a microelectrodein the hippocampus. The probe penetrated through the alveus and was lodged among the pyramidal cells of the Ammons horn. This was the experimental probe No. 5 which was conditioned in animal 1404. It was implanted at 2.5 mm lateral from the midline.
( B ) The recordings ( I ) The spike patterns: in the recordings there were repeating single and multiplespike patterns (see Fig. 3). Action potentials of constant amplitude and large signal to noise ratio were common. Often several of these were observed in recordings from a single probe. A multi-spike pattern which was particularly frequent when probes were in the hippocampus had the form of a decrescendo of 3 to 7 spikes each one about 213 of its predecessor in amplitude; the interspike interval within such a group was relatively brief and constant, giving the impression that despite the amplitude differences, the series of spikes represented a repetitive discharge from a single element. If single or multi-spike patterns were observed in recordings made a week after implantation they usually recurred indefinitely over the 2-week to 3-month period of an experiment. In response to the question of whether these long duration recordings were made from References p . 164
150
J. O L D S
Fig. 3. Typical neuronpatterns recordedfrom one probe in the hippocampus(probeNo. 6 in rat 1588)
the self-same neurons, the answer is that in some cases it appeared that identical responses from the same elements were recorded for the whole period. In these cases, not only was the amplitude relatively stable, but also the relative sizes of several spikes, or the signal to noise ratio, or the pattern of a decrescendo burst would remain relatively constant over the whole period. Other times gradual changes occurred suggesting a modification of tissues or movement or erosion of the probe; even in these cases it often appeared possible to follow a given unitary response for several weeks through the course of such a change with the conviction that despite the change in amplitude or template this was still the ‘same’ response. (2) The long run variability; The unit discriminator was usually set to respond to any member of a group of units rather than to a single unit. However, there were often two non-overlapping groups of units counted separately from the same probe in order to study correlations among proximal groups of neurons. The paper speed of 2 cm/min (i.e. 3 sec/mm) set limits on the analysis reported here. When the analysis system was set to read out at rates appreciably higher than 1 per sec, the detailed record disappeared. Thus gross but not fine changes in rate were observed. The highest sustained rates of the chosen response or response class determined the actual settings of the counting devices. When several units were counted together, rates were often very high and a correspondingly high counter setting was used; when a single unit was counted rates were often very low and the counter was set to read out at some relatively low number. Whatever the setting, the counter upon reaching this number would
THE LIMBIC SYSTEM A N D BEHAVIORAL REINFORCEMENT
151
make a mark on the moving record, and then reset and start counting again within a 0.5-msec interval. The most striking aspect of the records so formed was that all discriminators at one time or another exhibited rate changes that had the appearance of waves with 10 to 20 min periodicity; for several unit groups in the same animal, these waves were to some degree synchronized. The waves had the form of a gradual crescendo of readout rate up to a maximum followed by a relatively abrupt slowing. Because the 10 to 20 min periodicity was similar to that recently reported for rat sleep-cycles by Roldan ‘et al. (1963) it is appealing to suppose that these were waves of sleep and waking, and that the waking occurred at the point of abrupt slowing. Slow wave records were not made however and it is therefore not possible to offer direct evidence in behalf of this supposition. Besides these 10 to 20 min cycles, there were longer run changes which are so far unexplained. When the average response rates for a given discriminator for different 10-min periods were compared it was not unusual for these to vary by a factor of 2 or more during a period of several hours; sometimes there was even a 5-fold variation (Figs. 4 and 5). Several discriminators responding to different groups of neurons recorded from the same probe or from other probes in the same structure often gave evidence of a positive correlation among these long run changes of rate. Correlations appeared to be higher within structures than between structures and to be positively related to anatomical proximity of the two groups of neurons. Negative correlations of these gross changes were not observed frequently ;this was particularly true prior to conditioning. Because two negatively related neurons could occur during alternating small time intervals it was not impossible for such neurons to appear positively correlated when rates were averaged over periods of several minutes. It is only these longer run averages which are presented in the present report and this may account for a predominance of positive correlations. Later, analyses in more detail may reveal a larger number of negative relations. ( 3 ) Classical Pavlovian conditioning ? During operant conditioning experiments, the long run variability of neuron responses (analysed grossly by averaging over several minute periods) was regularly brought under some control so that response rates for the series of 2 min acquisition periods were often systematically different from response rates during the interspersed series of 8 min extinction periods. During the early phases of discrimination training, these controlled changes of rate followed an unexpected course that seemed better interpreted in terms of Pavlovian conditioning than in terms of operant conditioning. This was particularly true in the case of certain hippocampal unit responses. In these cases, before the ‘conditioned’ unit response pattern showed any evidence of discrimination, one of the other ‘control’ patterns would present clear evidence that a discrimination already existed (see Fig. 6). The animal, however, still made the ‘operant’ unit response on a chance basis and by the same token the animal still earned reinforcements on a chance basis. One might say that the animal already knew what to expect but did not yet know what to do. In some cases, during these early phases of conditioning the animal actually made fewer of the operant responses References p. 164
152
J. O L D S
H
Fig. 4. Long run variability in 4 different hippocampal neuron groups recorded simultaneously. The period of 6 h represents artifact-free time taken from a longer period of real time (about 12 to 20 h in this case). In all figures and in the text the time baseline is got by using a clock which ran continuously during quiet periods and stopped completely during and 2 sec after periods of detectable movement or artifact. Rates were averaged for 10-min periods: the top point in each case represents the rate for the most active period of an hour, the corresponding low point represents the rate for the quietest period of the same hour. It should be noted that the different curves are plotted on different ordinates and the origin is not zero in any of the cases; low points and high points of the 6-h curves are noted by numbers indicating responses per minute simply to aid inspection. All probes were in hippocampal gyrus (HPC) with laterality (L) at 1,2.5 and 3.5 mm from the midline as indicated. The two probes at 3.5 mm were a twisted pair of microelectrodes cut to the same depth.
Continuous Extinction
z
3
4 P
w
I I
1
2
I
,
3 SUCC&IVE
,
,
hURS60F "&On
I
,
TI&
,
,
lo
,
I'
Fig. 5. Rates after conditioning in an experimental unit group in the pons. Acquisition scores are for the high and low 2 min acquisition periods of an hour; extinction scores for the first 3 h are for the high and low 8 min extinction periods of an hour. Extinction scores for the remaining hours (5 through 11) are for high and low 10 min periods because extinction was continuous. Not only was there difference between acquisition and extinction rates during a period of alternation but also extinction rates were augmented during such a period.
T H E L I M B I C SYSTEM A N D B E H A V I O R A L R E I N F O R C E M E N T
153
H
Fig. 6. Behavior of units 6 and 7 from Fig. 4 during early phases of discriminationtraining. Hatched envelopes indicate response rates for the 2 min acquisition periods.
during acquisition than during extinction periods. If this happened repeatedly when the animal was hungry and if the difference was stable and significant, it was assumed that the unit responses had an a priori negative relation to some aspect of the acquisition period, being depressed by the visual signal or by the expectancy or immediacy of food or by similar causes unknown. Even in such cases, there was often a full reversal during the later phases of training so that a stable difference in favor of the acquisition period occurred. Because these ‘classically conditioned’ changes in rate of the experimental neurons or other neurons were observed mainly with the hippocampal neurons and not in experiments on pontine neurons it is possible to suppose that they derived from some unconditioned or previously established relation between hippocampal neuron activity and the hunger drive and food rewards which were used to shape operant behavior in these experiments. ( 4 ) Operant conditioning: when this was successful, the high frequency response of the chosen unit group occurred far more frequently during the acquisition periods than during the corresponding extinction periods. While this could occur without any overall shift in response rate (by alternation of very high frequency intervals with silent intervals) it was ordinarily associated with a clear increment in the rate of the chosen unit responses during the acquisition periods. The criteria1 high frequency response often underwent an almost all or nothing change as between acquisition and extinction. That is, there would be no sustained frequencies high enough to meet criterion during extinction periods, even though these would occur often during acquisition. At other times these high frequency responses would occur in acquisition and extinction periods but would occur with considerably greater frequency during acquisition. In either case it was common for the background unit response rate to undergo a sizeable increment in rate during acquisition usually amounting to at least a doubling (see Figs. 7, 8 and 9 and Table I). Sometimes rates went up even 5 or 10 fold during acquisition periods and very rarely the units would be almost silent during extinction periods. ( 5 ) Covariation: in the cases of successful conditioning, it was the rule for other unit groups to show clear covariation or clear inverse variation whenever the acquisiReferences p . 164
154
J. O L D S 1588 "Wuntory" Ccnlrcl of Voriobilil# Ailer Conditioning
1000
600
L 1000 500
EX1
I
I
I
2
H
DAY
I
2
H
I
Dpll
Fig. 7. Behavior of units 1,6 and 7 from Fig. 4 during final phases of discrimination training. Hatched envelopes in this case indicate response rates for the 8 rnin extinction periods (it should be noted that this is the opposite convention as that used in Fig. 6). Also it should be noted that the units on probe 7 which responded more rapidly during acquisition in early stages of conditioning responded more rapidly during extinction in the h a 1 stages of conditioning.
Unit # 3
300
z
mtD1 CDntrnl
I
Unit
Hour $k I
Pons
DW
#
2
wmnfi Conhol unit
lhll
ons
HTH Hour
3
Fig. 8. Covariation of two control unit groups in the pons with a conditioned unit group in the pons; variability (low to high 2 min or 8 min intervals) is indicated by hatching or clear rectangle at the top of each histogram. There was no similar covariation evidenced in the behavior of a hypothalamic unit group shown on the right.
tion or extinction changes occurred. While the degree and sign of the correlation changed depending on the anatomical proximity of neuron groups and depending on other unknown factors, it was a rare occurrence for a relative absence of correlation to obtain between two groups of neural responses, particularly insofar as the changes caused by conditioning were concerned.
c
TABLE I
e
3
z
H IPPO CAMPA L U N I T RESPONSE R A T E S * AFTER C O N D I T I O N I N G FOR THE L OW A N D H I G H PERIO DS O F EAC H HO UR
0
There were six 2-min acquisition (A) periods and six 8-min extinction (E) periods in each hour. Probes were in dorsal hippocampus behind the anterior hippocampal commissure at the lateral coordinates indicated
m
Probe No.
Lateral coordinates
Hour 1 A
Hour 2 E
A
E
A
3
Hour 4
Hour 3 E
A
2el
m
E
*2 u
m
6A** 6B 7 1 2
1 1 2.5 3.5 3.5
1312-1600
528-680
26 800- 992 160- 240
10- 12 656-768 6 4 92
1504-1568 0 - 8
768-1024 176- 224
472-640 7- 15 688-760 72- 88
1600-1888 1472-2496 00 512- 928
880- 940 1808-1920 22 648- 728
1312-1632 2496-2688 00 800-1024
792- 976 1600-1760 25 6 2 4 680
m
m
P
6 0 ;d
P P
FJ
* **
Scores are rates in responses per rnin averaged for the 2 rnin or 8 rnin intervals, animal No. 1588. 'conditioned' response A = acquisition periods (2 rnin each) E = extinction periods (8 rnin each)
c
5
0
FJ
0 m
z
m 2: 4
156
J. O L D S *Following"unit 01
Same Site
Con1 f 5
bnd'd" HPC Unit EXP f 5
I
ion- following" HPC unit con1 f 4
DACQ
0 EXT
ANIMAL
Fig. 9. Covariation of one control unit group in the hippocampus and independence of a second unit group in the same structure.
In the case of pontine probes, all the other units recorded from the pons showed clearcut positive or negative changes as between acquisition and extinction as soon as one group was conditioned to the point where it exhibited a clear discriminative response. This seemed to indicate that in conditioning one pontine neural group, the procedures actually conditioned all of them. Most pontine units were faster in acquisition periods and slower in extinction, thereby following the changes of the conditioned units. In two cases, however, a medial lateral distinction occurred; a lateral unit group was conditioned and other lateral units followed its accelerations and decelerations; medial neurons in these cases, however, exhibited inverse variation. In another experiment, a control unit group in anterior lateral hypothalamus was uncorrelated with a conditioned group in the pons; in this case there was neither positive nor negative correlation. The picture with hippocampal neurons was somewhat different. In these cases also the conditioned change in response rate of the chosen units had widespread ramifications. More units showed positive correlation with the conditioned neuron group than negative. However, independence of two neuron groups within the hippocampus appeared possible (see Fig. 9), and negative correlation between the experimental units and closely neighboring units occurred even when more distant neurons were positively correlated. One interesting observation in a hippocampal experiment was an apparent change in correlation pattern which accompanied conditioning. The unit groups involved had seemed positively correlated when 10-min intervals were used as a basis for comparison prior to conditioning (see Fig. 4 :6 and 7). Later, one was chosen as the experimental group and its rate was greatly augmented during 2 min acquisition periods after training was complete. At this point, the other group was a control set and its rate was greatly depressed in'acquisition (see Fig. 7 : 6 and 7). (6) Interspike intervals: Whenlthe high criteria1 rates of the experimental unit (which constituted the operant response) occurred during extinction periods after
THE L I M B I C SYSTEM A N D B E H A V I O R A L R E I N F O R C E M E N T
157
conditioning was complete, they seemed to have a different pattern from that exhibited during the more 'purposeful' responding of acquisition. Extinction responses seemed to be constituted by one or several very fast bursts of the kind which Evarts found to characterize pyramidal neurons during sleep. There were brief and regular interspike intervals within these bursts and relatively long silent periods between them (see Fig. 10). Acquisition responses on the other hand seemed to be more often com-
Fig. 10. Photographed extinction bursts. These should be compared with correlated acquisitionbursts (right hand group in Fig. 1).
posed of interspike intervals which were not so regular or so brief; and the longer silent periods which intervened between the bursts in extinction did not appear so often in the course of acquisition responses. In the case of one hippocampal neuron group interspike interval histograms were plotted from photographic records of successful high frequency bursts. The histograms did not show any significant difference between acquisition and extinction bursts even though the photographs had seemed to indicate that such a difference existed (see Fig. 11). When a similar set of histograms was plotted for a conditioned pons unit, the expected difference did appear (see Fig. 12). Extinction bursts had more very short and very long intervals; acquisition bursts had more of in-between duration. ( 7 ) Special tests: Tests for the validity of the operant conditioning were made by (a) changing the discriminative C.S. to find whether these were responses to the acquisition signal, (b) changing the reward to test for an apriori relation of the unit response to the References p . 164
158
J. O L D S
expectancy of food, (c), radically changing the visual surroundings, and (d) radically changing the drive level to test for a dependency of the conditioned response on aspects of the internal environment. (a) Discrimination evidenced in the behavior of the experimental neurons was not 351
INTER-WIT INTERVAL YSEC
Fig. 11. Inter-spike interval histograms for acquisition and extinction bursts for the experimental, hippocampal neuron group of animal No. 1404.
A
5
ACQ
$0 ;o
do
I&A I&
ub 140Kb 1$017blBO n;?bo
INTER - UNIT INTERVAL MSCC
Fig. 12. Inter-spike interval histograms for acquisitionandextinctionbursts for the experimental, pons neuron group of animal No. 1261.
greatly impaired by modification of the acquisition signal. Even when this was changed from a light sustained for the whole acquisition period to a bell at the onset only, operant behavior continued to show a marked change at the beginning and shortly after the end of the acquisition period (see Figs. 13 and 14). This appeared to indicate that the pattern involved was not an unconditioned response to the original signal. Moreover, when the unit behavior was extinguished quickly without any extinction signal, it appeared from the record of responses that a definite operant behavior pattern was tried several times and then stopped altogether. This suggested that the operant response might be quite a definite act from the animal's point of view even though it required no overt behavior. Not too much should be made of the absence of
T H E LIMBIC SYSTEM A N D BEHAVIORAL REINFORCEMENT
159
___. /
EXP=#J L-1 PONS
--
BURSTS
-Acg’
, EXT
1
Ac9
-
EXT
--A52
__ ._C
EXT
--Am-
Fig. 13. Photographed tracings showing rate of control (CONT) units and experimental (EXP) units in the pons. Control units made a mark on the top channel of each tracing when a predetermined number was reached. Experimental units caused a step to occur in the second tracing when a predetermined number was reached, thereby converting the response rate into the slope of a line. Criteria], high frequency patterns (BURSTS) of the secondary pattern discriminator made marks on the lower channel of each tracing. Different control probes in the pons were numbered 1 and 5 ; one control probe in anterior hypothalamus was numbered 7.The experimental probe in the pons was numbered 3. The ‘L‘ number for a given pons probe indicates its distance from the midline. Two minute acquisition periods were alternated with 8 min extinction periods; the recorder and the timer ran only during periods of artifact-free time. Therefore, the actual acquisition and extinction periods were of longer and unknown duration.
an extinction signal, however, as the animal still had the cessation of pellet discharges and pellet-dispenser noises to go by. Even though these occurred after the responses, they still could influence much of the behavior because what was after one response was before the next. (b) When animals were shifted from food reinforcement to brain stimulus reward, this often resulted in a relapse to undifferentiated unit behavior. Then during the course of a single day, conditioning would recur. During such a period of reconditioning with a hippocampal unit probe an interesting series of changes took place which might be considered a foreshortened version of the changes that occurred during the original course of conditioning (see Fig. 15). For the conditioned unit, rates not only rose during successive acquisition periods but also declined during successive extinction periods. For a control unit, rates first stayed at initial levels for both acquisition and extinction periods, then a brief discrimination appeared, the control units following the experimental ones, becoming accelerated during acquisition. Finally, the response rate of the control units fell sharply and there was again no discrimination evidenced in their behavior but now they were responding at a rate considerably below their initial level. Referencrs p . 164
160
J. O L D S
EXP=#5 L 2 5 HPC
/
/ BURSTS
-
~
-m-
ACQ
EXT
/
--
-
EXT
AW
i-
1 1I /
Each Mort Rewsrnh 14 Unit
I 5 L 2.5 HPCJ I._._ ., I
COW * Y4
L 15 HPC
____I.
C5 LZ.5 HPC
.. . ... . ..
. .
,
.
,
.
.
,
.
.
, ,,,,
Eqc? ?r!
..
R p r a.t r Z O U.n .H r ( n p o n w S
SW; L-
Fig. 14. Photographed tracings showing rates of control and experimental neuron groups in the hippocampus. In ‘bell only’ tests there was an auditory rather than a visual discriminative stimulus, and it was not sustained throughout the acquisition period. In ‘brain reward’ tests, an electrical stimulus in anterior lateral hypothalamus was used for reward rather than a food pellet. In thecurare test, the animal was paralysed and respirated. During the acquisition period there were very few hippocampal responses at the beginning. Therefore, the animal was rewarded first for ‘2 bursts’; i.e. two discharges in a & sec period; then for ‘4 bursts’, and so on up to ‘8 bursts’. This produced a considerable augmentation of the hippocampal discharge rate; in the ensuing extinction period it dropped back toward the prereinforcement level. It did not, however, become augmented during the second acquisition period, and shortly thereafter the animal died. Apparently, the anterior hypothalamic stimulation combined with the curarization t o cause death.
(c) Major modification of the visual stimulus environment had a debilitating influence on the unit behavior. In these tests the cage doors were opened and photoflood lights were used to illuminate the whole area for cinematographic recordings. Direct observation of the animal before and during these photo-flood tests made it clear that the animal was ‘trying’ to get pellets in roughly the same way after the major change in visual environment. Nevertheless, these efforts were not nearly so effective. The animal made fewer operant responses during acquisition tests; and the unit behavior showed less evidence of discrimination between acquisition and extinction. These tests lasted only for short periods and therefore it was not clear whether the changes would have been gradually erased after continued training in the brightened environment. In spite of the poor performance there was still some evidence of discrimination in the brightly lighted environment. Moreover, some of the poor performance may
I;
T H E L I M B I C SYSTEM A N D B E H A V I O R A L R E I N F O R C E M E N T 1404 Brain Stimulus Reward
161
t
P
900 5
x
1
f
I
I
I
I
2
3
I
.
’ ’
I
7
I
8
H
Fig. 15. Acquisition and extinction rates for ‘conditioned‘ and ‘control’ units during reconditioning with brain reward. Variability shown is the total range for each 1-h period (i.e. high and low 2 min or 8 min period). Points are average rates for the 6 such periods in each hour.
have been due to a general reduction in hunger produced by the warming effect of the flood lights. (d) A surprising but in retrospect predictable result was the very great influence of the drive state even when the animal was responding for a brain reward whose reinforcing value in other tests appeared to be unaffected by such drive changes. Such tests were made in only 2 cases and in both cases unit response probes were in the hippocampus. Discrimination evidenced in the behavior of the experimental unit broke down and was not restored until the animal was returned to the high drive state of the original training. It seems possible that this was due to an intrinsic relation between the visceral drives and hippocampal responses.
( C ) The correlated behaviors ( 1 ) Cinematographic recording: Although neither the unit analyses nor the operant behavior could take place during periods of detectable movement or detectable muscle artifacts, there were nevertheless subtle and undetectable movements which accompanied or preceded many operant responses. Cinematographic recordings showed a slow head movement to left or right as a more or less constant predecessor or concomitant of the ‘purposeful’ pontine unit behavior which occurred during acquisition periods. Even if the animal appeared to be standing quite still during the 2-sec period immediately preceding a reinforced burst of units, the head movement was nevertheless apparent in the recordings which immediately preceded the quiet period. Other times a movement which was barely discernible to the eye and which escaped completely the mechanical detectors appeared as the overt component of the pontine operant response*. When hippocampal units were conditioned, there were sometimes clearcut sniffing movements of the tip of the nose combined with whisker movements that similarly accompanied the operant responses. However, with hippocampal responses there were
*
Stimulation of a pontine probe in later tests often seemed to provoke almost the same movement.
References p . 164
162
J. O L D S
times when the animal emitted apparently purposeful operant unit bursts without any observable overt responding at all. (2) Curarizedpreparation: One animal with conditioned neurons in the hippocampus and positive reinforcement probe in anterior lateral hypothalamus was tested for operant behavior after paralysis induced by i.p. injection of 50 mg/kg of flaxedil. The animal had previously been tested in a high drive condition with hypothalamic stimulation as reward. Clear discrimination behavior had been maintained over a 6-h period immediately preceding the flaxedil induced paralysis. The rat was respirated through the nasal openings by means of a nose grip respirator of a type designed and made in our laboratory and which has been used successfully in many of our experiments. The animal was hyperventilated utilizing air mixed with 5 % carbon dioxide. The animal with respiratory device attached was replaced in the plastic recording cage after having been out of it for about 10 min during curarization procedures. Recording and conditioning procedures had been underway prior to the 10-min interruption and they were reinstituted as soon as the animal was replaced in the recording cage. The rate of the experimental unit responses had fallen considerably, and even though the visual acquisition signal was turned on the animal showed no signs of producing the 16 resp sec burst required to earn a brain stimulus reinforcement. In order to shape the operant behavior, therefore, the burst requirement was set back first from 16 to 2. Then after the first reinforcement it was set forward from 2 to 4; then 4 to 6 and then 6 to 8. During this period the experimental unit responses increased their rate about 60 fold from 32 per min up to 32 per sec. This rate was maintained for 2 min and then acquisition was terminated. During the ensuing extinction period the rate fell again to its low, post curare level; 8 min later a second acquisition test was imposed. No shaping methods were used this time and the response rate did not rise above chance levels. Five min later the animal died, death apparently being caused by interaction of the parasympathetic influences of anterior hypothalamic stimulation and the effects of curare. Non-self-stimulating rats were tested repeatedly before and after the test rat. These withstood and recovered well from the curarization procedure. Animals are now being prepared for further curare tests. The reinforcing electrodes in these new cases are in posterior hypothalamus, where, it is hoped, reinforcing stimulation will not kill the curarized animal. The curarization test was considered a limited success because (1) the reinforcement procedure caused a large increase in rate of the hippocampal response in the paralysed animal at first and (2) when the increment failed to appear a second time the animal was already in a debilitated condition preceding death. However, it would require more evidence than is yet at hand to permit a conclusion that hippocampal neurons can be voluntarily controlled without recourse to overt behavior.
+
SUMMARY A N D CONCLUSIONS
1. Responses composed of patterns derived from pontine and hippocampal units were conditioned to occur more frequently by rewarding the animal after each occurrence of the response.
T H E L I M B I C SYSTEM A N D B E H A V I O R A L R E I N F O R C E M E N T
163
2. These patterns were readily conditioned even though detectable movement and muscle artifacts were interdicted before, during and after the response. 3. By rewarding the animals only for those responses made in the presence of a particular stimulus, these unit patterns were brought under the control of this discriminative stimulus. 4. When unit groups in the pons became conditioned all other units in the pons gave evidence of the conditioning. Most other units followed the conditioned response being accelerated during acquisition and decelerated in extinction. But there were two cases where lateral units were conditioned and medial units gave evidence of inverse variation. 5. No similar correlation was observed between pontine units and units in anterior lateral hypothalamus. 6. When unit groups in the hippocampus were conditioned many other units in the hippocampus varied directly and a few inversely with the operant-induced changes in the conditioned units. However, there were cases of independent hippocampus units which did not follow or invert a conditioned change in the experimental unit. 7. In one case, a hippocampal unit which seemed to vary directly with the experimental unit when hour to hour changes were considered prior to conditioning later showed clear evidence of inverse variation when the acquisition and extinction periods were compared after conditioning was complete. 8. Sustained t sec high frequency responses constituted the operant response. When they occurred ‘accidentally’ during extinction they appeared to have more stereotyped bursts and longer intervening silent intervals than when they occurred ‘purposefully’ during acquisition. 9. When pontine units were conditioned, the operant response seemed to include a movement of eyes or head which escaped the movement detectors. 10. When hippocampal units were conditioned the operant response sometimes included movement of nose or whiskers but sometimes there was no obvious peripheral component of the behavior. 11. Hippocampal neurons responded in controlled fashion in midconditioning prior to the emergence of the discriminated operant response. This suggested an a priori connection between the hippocam pal responses and the feeding schedule cycles produced by the alternated acquisition and extinction. 12. Suggesting a similar connection of some hippocampal neurons to hunger drive was the fact that a neural response shaped under high drive conditions with food reward could be transferred to brain reward only if the high drive of training were maintained. 13. Preliminary evidence gave grounds for the hope that it would eventually be possible to observe operant responses from hippocampal neurons when the animal was paralysed with a curarizing agent.
References I. 164
164
J. O L D S
ACKNOWLEDGMENTS
This work was supported by grants from The National Science Foundation, The National Aeronautics and Space Administration, and the U.S. Public Health Service to Dr. J. Olds. The author is grateful to Fred F. Coury (BSEE) for designing analytic circuits, to Mr. William E. Wetzel for designing brain probes and cages, to Mr. Giulio Baldrighi for surgery and neurophysiological work at the time of surgery and to Mr. Floyd F. Foess for assistance with all aspects of the experimental work. REFERE N C E S EVARTS,E. V., (1964); Temporal patterns of discharge of pyramidal tract neurons during sleep and waking in the monkey. J. Neurophysiol., 27, 152-171. OLDS,J., (1965); Operant conditioning of single unit responses. Proc. XXIIZ Znt. Congr. physiol. Sci., Tokyo, Sept. ROLDAN, E., WEISS,T., AND FIFKOVA, E., (1963); Excitability changes during the sleep cycle of the rat. Electroenceph. clin. Neurophysiol., 15, 775-785.
165
Further Studies on the Effects of Amygdaloid Stimulation and Ablation on Hypothalamically Elicited Attack Behavior in Cats M. D A V I D EGGER A N D JOHN P. FLY" Departments of Anatomy and Psychiatry, Yale University School of Medicine, New Haven, Conn. (U.S.A.)
Evidence on the role of the amygdala in the integration of aggressive behavior is contradictory and confusing. However, most investigators agree that an important function of the amygdala is a modulatory one, modulating the activity of the hypothalamus in particular. If we assume that the amygdala has anatomically distinct regions that suppress or facilitate the hypothalamus, it might be possible to explain many apparently contradictory data. Earlier we found that the amygdala does indeed contain anatomically distinct regions capable of suppressing or facilitating an attack response elicited by stimulation in the hypothalamus (Fig. 1;Egger and Flynn, 1963). However, because we used bipolar needle electrodes with tip separations of 2 mm or more, the anatomical localizations were not precise. To determine more sharply the anatomical loci of suppression or facilitation of the attack response, in 9 cats we stimulated in the amygdala with monopolar electrodes. Experiment I
Methods The details of our stimulation and recording methods have been described (Egger and Flynn, 1963; Wasman and Flynn, 1962). Because the attack response we studied was directed at a rat, we selected our experimental subjects from among 'non-ratters'. Most tame cats do not attack rats. During an experimental session, a cat and an anesthetized rat were placed together in an observation cage. The electrodes, implanted in the cats at least 7 days before testing, were connected to a relay so that they could be switched selectively from connection with an EEG recorder to connection with a stimulator. EEG recordings were taken from the amygdala before and after each stimulation, to determine if electrical after-discharges had occurred. Data recorded during trials on which after-discharges did occur were not included in analyses of the effects of amygdaloid stimulation. Electrical stimulation consisted of biphasic, rectangular-shaped pulses. The electrodes were stainless steel needles, insulated except for 0.6 mm at the tip. Stimulation was unilateral. References p . 180-182
166
M. D. EGGER AND I. P. F L Y N N
Fig. 1.Sitesof the tips of amygdaloidelectrodesprojected on diagramsof frontal sections (based on the atlas of the cat brain in Bur& et al., 1962). Filled circles depict electrodesthrough which stimulation produced a suppression of hypothalamically elicited attack behavior; open circles depict electrodes through which stimulation produced a facilitation of hypothalamically elicited attack behavior; and crosses depict electrodesthrough which stimulation produced no consistent effects. The lettered lines connecting the poles of bipolar electrodes identify the electrodes of individual cats (after Egger and Flynn, 1963). AAA = area amygdalaris anterior; AB = nucleus basalis amygdalae; AC = nucleus centralis amygdalae; ACO = nucleus corticalis amygdalae; AHL = area hypothalamica lateralis; AL = nucleus lateralis amygdalae; AME = nucleus amygdalae medialis; APL = area praeopticalaterak; CA = commiissura anterior; CH = chiasma opticum; CI = capsula interna; EN = nucleus entopeduncularis; GP = globus pallidus; P = putamen; SO = nucleus supra0pticus;TO = tractus opticus.
In order to stimulate in two regions of the brain simultaneously (actually, in very rapid succession), we interlocked the outputs of two electrically isolated stimulators in such a way that the two regions of the brain were never stimulated at the same instant, so no inadvertent stimulation of regions lying between the electrode sites could occur. stimulation in the amygdala was monopolar; that in the hypothalamus was bipolar.
STIMULATION A N D A B L A T I O N I N THE A M Y G D A L A
167
The stimulation frequency was 62.5 pulses per sec, with the onset of hypothalamic stimulation typically preceding that of the amygdala by 6 msec. We measured the latencies of initial movement and of attack following the onset of electrical stimulation with stop watches. The attack latency was defined as the interval from the onset of brain stimulation until the cat touched the rat with tooth or claw. Stimulation trials were separated by at least 4 min. Two levels of hypothalamic stimulation were used; the higher level was obtained by increasing the duration of the stimulus pulses. Typical values of hypothalamic stimulation were 0.40 mA, biphasic, peakto-peak, with 1.0 msec per pulse at the low level of stimulation and 1.5 msec per pulse at the high level. Amygdaloid stimulation intensities were chosen at levels below the threshold for electrical after-discharges. We reasoned that a more meaningful anatomical localization of effects might be obtained if we could avoid after-discharges. Typically, the amygdala was stimulated at 0.15-0.20 mA, biphasic, peak-to-peak, 3.0 msec per pulse. Rarely did stimulation of the amygdala alone at these levels elicit observable behavioral reactions. Experimental sessions were separated by at least 46 h. The effects of simultaneous stimulation in the amygdala and the hypothalamus were evaluated systematically for a selected amygdaloid electrode in two successive sessions. During each session, ‘single’, i.e. stimulation in the hypothalamus alone, and ‘dual’, i.e. stimulation in the amygdala and hypothalamus together, trials occurred in ABBA order. The effects of dual stimulation at both low and high levels of hypothalamic stimulation were tested. Each test session consisted of 12 pairs of single and dual stimulation trials. In addition, at least twice during each session, the amygdala alone was stimulated. During trials in which no attack occurred, stimulation was continued for 20 sec; otherwise, stimulation was terminated when an attack occurred. Those trials during which no attack occurred were included in the data analyses as if an attack had occurred at 20 sec. All data on attack and initial movement latencies were transformed to their reciprocals. We shall refer to these reciprocals as ‘speeds’. Because the speed distributions were less skewed than the latency distributions, where possible, statistical analyses were performed on the speeds. In the remainder of this paper, we shall refer to the cats from our earlier study, stimulated with bipolar electrodes in the amygdala, as members of the ‘bipolar’ group ; and to the cats from the present study, stimulated with monopolar electrodes in the amygdala, as members of the ‘monopolar’ group. Results Amygdaloid stimulation in 5 of the 9 monopolar cats produced a statistically significant suppression of attack, i.e. an increase in the time from the beginning of stimulation to the occurrence of an attack as a result of adding stimulation in the amygdala to that in the hypothalamus (Table I ; Fig. 2). In two of the 9 cats, facilitation of attack occurred, i.e. a decrease in the time to attack when both the amygdala and the hypothalamus were stimulated. References p . 180-182
168
M. D. EGGER A N D J. P. PLYNN
TABLE I EFFECTS OF ELECTRICAL STIMULATION I N THE AMYGDALA ON LATENCIES OF HYPOTHALAMICALLY ELICITED ATTACK ~
~
~
Low level of hypothalamic stimulation
Cat
No. of trials1
Mean attack latencies (see) Singlez
B C I
F
E D H A G
2
3
4.6
8
9 10 5 11 12 12 8 10
6.9 7.5 7.3 11.3 6.4
9.0 6.8 7.7
High level of hypothalamic stimulation
No. of trials1
Dual3 6.8
9.5 16.5 11.2 17.2 5.9 8.3 4.1 5.6
12 5 11 10 5 11 12 11 12
Mean attack latencies (see) Singlez
Dual3
3 .O 3.8
6.3 5.6 10.7 6.4 6.7 2.9 4.5 2.9 3.1
6.0
3.8 5.8 3.5 4.8 3.6 3.7
Statistical signifcance of suppression (S) or of facilitation (F)
S*** S***
S*'*
S** S* F*** F***
Data from trials on which after-dischargesoccurred are not included in this table. 'Single' refers to stimulation in the hypothalamus alone. 'Dual' refers to simultaneous stimulation in the amygdala and the hypothalamus. * P < 0.05; ** P < 0.01; *** P < 0.001.
SEC
m -
SOME EFFECTS OF DUAL STIMULATION FACILITATION
CAT A
W3 EFFECT
SUPPRESSION
CAT D
CAT F.
Y
0
a
I-
%
r
z s w
z I-
Fig. 2. M a latencies of attack during trials of stimulation in the hypothalamus alone (single) and in the hypothalamus and amygdala together (dual) for 3 cats, 1 which showed facilitation, 1 which showed no consistent effect, and 1 which showed suppression of attack when amygdaloid stimulation was added to hypothalamic Stimulation. The total height of each pair of bars represents mean attack latencies for low and high levels of hypothalamic stimulation. Open bars show mean attack latencies during stimulation in the hypothalamus alone; the shaded bars show mean attack latencies during simultaneous stimulation in the amygdala and hypothalamus. The dark portions at the bottom of each bar show mean latencies of initial movement following the onset of stimulation.
STIMULATION A N D ABLATION I N THE AMYGDALA
I69
The localization of effects in the amygdala for the monopolar cats agreed with those of the bipolar, and, as hoped, enabled us to define their anatomical loci more precisely : in general, suppression was elicited most consistently in the magnocellular portion of the basal nucleus of the amygdala, and in the anterior and medial portions of the lateral nucleus of the amygdala (Fig. 3). Facilitation was elicited in the dorsolateral portion of the posterior part of the lateral nucleus. One exceptional cat, A, showed facilitation of attack even though, from the location of its electrode in the amygdala, we would have expected suppression. The electrode of cat D is not depicted in Fig. 3. Stimulation through this electrode, located at the interface of the hippocampus and the optic tract, produced no consistent effects on hypothalamically elicited attack. Similarly located bipolar electrode placements also failed to produce consistent effects. At the end of the two test sessions, we stimulated the amygdala by itself at increasing current levels until after-discharges occurred following not more than 20 sec of stimulation. In only one cat, C, did stimulation below the threshold for after-discharges elicit a behavioral response. During stimulation of the amygdala, this cat assumed a defense-like crouch and hissed. At current levels that evoked after-discharges, growling, hissing, and a characteristic amygdaloid-seizure facial twitch also occurred. Stimulation through this electrode, when paired with hypothalamic stimulation, produced a marked suppression of hypothalamically elicited attack.
Discussion Electrical stimulation in the amygdala suppressed or facilitated hypothalamically elicited attack. Opposing effects were elicited from different parts of the amygdala. These effects were not due to the elicitation of behavioral responses by amygdaloid stimulation, because in only one cat did amygdaloid stimulation alone elicit a behavioral response at the current levels used in dual stimulation. The suppression effect was not due to arrest of all motor activity, because at the levels of amygdaloid stimulation used, the cats were responsive and displayed normal behavior patterns (Egger and Flynn, 1963 ;Fonberg and Delgado, 196 1). Furthermore, amygdaloid stimulation was considered to have been effective in suppressing hypothalamically elicited attack behavior only if attack latencies were increased, but latencies of initial movement were unaffected. Finally, the dual stimulation effects were not associated with after-discharges in the amygdala, because data from trials during which electrical after-discharges occurred were excluded. We would like to argue that the effects of amygdaloid stimulation we have observed are due to modulation by the amygdala of neural activity in the hypothalamus. Neuroanatomists have described at least 3 pathways from the amygdala to the hypothalamus in the cat (Fox, 1940,1943; Hall, 1963; Szentiigothai et al., 1962; Valverde, 1963). One pulse per sec electrical stimulation through the electrodes in the amygdala in our cats evoked electrical responses at the hypothalamic electrodes used to elicit attack (Egger and Flynn, 1963). Furthermore, many units in the hypothalamus are influenced by electrical stimulation in the amygdala (Sawa et al., 1959; Stuart et al., 1964; Tsubokawa and Sutin, 1963; Wendt and Adey, 1960). The firing rates of some References p. 180482
170
M. D. EGGBR A N D J. P. FLY”
SUPPRESSION
0
FACILITATION
Fig. 3. Triangles depict electrodes through which stimulation produced a suppression of hypothalamically elicited attack behavior and open circles depict electrodesthrough which stimulation produced a facilitation of hypothalamically elicited attack behavior. Letters in squares identify the electrodes of individual cats. Broken limes connect the poles of bipolar electrodesof cats, M, N, P, and V from Fig. 1. This figure depicts the locations of monopolar and bipolar electrodesof all cats meeting the following criteria: the effect of amygdaloid stimulation on attack speed, as well as on the interval between initial movement and attack, had to have a statistical signiticance exceeding the 0.01 probability level in the first 24 pairs of trials or less. AHA = area hypothalamica anterior; A1 = nucleus intercalatus amygdalae; APM = area praeoptica medialis; CD = nucleus caudatus; C PYR = cortex pyriformis;F = fornix;HVM = nucleus ventromedialis hypothalami; NCAST = nucleus commissurae anterioris et striae terminalis; NHD =
STIMULATION A N D ABLATION I N THE AMYGDALA
171
units in the hypothalamus can be slowed by stimulation in one part of the amygdala, and speeded up by stimulation in another part of the amygdala, in analogy with the contrasting behavioral effects of amygdaloid stimulation reported here (Egger, 1965). None of these observations, however, tells us anything related directly to attack responses elicited by hypothalamic stimulation. We reasoned that more direct evidence of neural effects of the amygdala on hypothalamic mechanisms concerned with attack behavior might be provided by examining the effects of small lesions placed at the tips of electrodes in the amygdala for which we had determined the stimulation effects. Experiment II
Methods Fifteen cats, 7 monopolar and 8 bipolar, received bilateral amygdaloid lesions. Cats in the bipolar group received anodal lesions at the tips of both poles of the bipolar electrodes in the amygdala. Six cats, 2 monopolar and 4 bipolar, served as shamlesion controls. Anodal electrolytic lesions were made by passing 3.CL5.0 mA direct current for up to 20 sec. A lead clipped to the nape of the cat’s neck served as cathode. Between making a lesion and testing its effect, 27 to 36 days elapsed for the monopolar, and from 2 to 36 days elapsed for the bipolar cats. We assessed the effects of the lesions by comparing attack speeds on corresponding trials of experimental sessions before and after lesioning. Identical intensities of hypothalamic stimulation were used in pre- and post-lesion test sessions. Results Five cats attacked significantly faster after receiving amygdaloid lesions : monopolar cats A and F, and bipolar cats M, N, and 0 (Table 11). The mean pre- and post-lesion attack latencies of cat A are summarized in Fig. 4.Similar data of Cat 0 are summarized in Fig. 5. All 5 cats that attacked faster following lesioning, and 5 of the 10 cats that did not attack faster following lesioning were first tested at least 21 days after lesioning. None of the 6 sham-lesion cats attacked faster during the ‘post-lesion’ test. Four of the 5 cats that attacked faster following lesioning had shown statistically significant suppression of attack during amygdaloid stimulation. The 5th cat, A had its electrodes in an amygdaloid region associated with suppression of attack, namely, the magnocellular portion of the basal nucleus of the amygdala, but it had shown facilitation of attack during amygdaloid stimulation. All 5 cats that attacked faster following lesioning had bilateral lesions that included a common locus : the border region of the magnocellular portion of the basal nucleus of the amygdala and the dorsomedial portion of the lateral nucleus of the amygdala nucleus hypothalamicus dorsalis; NHP = nucleus hypothalamicus posterior; NT OLF LAT = nucleus tractus olfactoriilateralis;NTS = nucleus triangularis septi; PV = nucleus paraventricularis; RE = nucleus reuniens; SCH = nucleus suprachiasmaticus; TMT = tractus mammillo-thalamicus ; V = ventriculuslateralis ;VM = nucleus ventralis medialis;VPL = nucleus ventralis posterolateralis ; ZI = zona incerta. For key to additional abbreviations, see Fig. 1. References p . 180-182
I 72
M. D. EGGER A N D J. P. F L Y N N
Fig. 4. Mean latencies of attack during hypothalamic stimulation for cat A before and after amygdaloid lesions. The total height of each pair of bars represents mean attack latencies for low and high levels of hypothalamic stimulation. The shaded portions at the bottom of each bar show latencies of initial movement. The first post-lesion test (session 2) was given 34 days following lesioning; the last (session 5), 178 days or approximately 6 months following lesioning.
(Fig. 6). These lesions also encroached upon the ventral and medial borders of the central nucleus. None of the 10 cats that did not attack faster following lesioning had lesions that included this region (Fig. 7). Lesions of 3 cats are not depicted in Fig. 7: that of cat D, that had no bilaterally symmetric component to its lesion; and those of cats I and T, that attacked more slowly following lesioning. Cat I had a unilateral amygdalectomy prior to electrode implantation. Post-mortem examination of its brain revealed signs of hydrocephalus and apparent shifts in electrode postitions. The decrease in attack speed following SEC
la
Y
0
a I-
s 0
IW
z I-
5
-
n
n
Fig. 5. Mean latencies of attack during hypothalamicstimulationfor cat 0before and after amygdaloid lesions. The total height of each pair of bars represents mean attack latencies for low and high levels of hypothalamic stimulation. The shaded portions at the bottom of each bar show latencies of initial movement.Thefirstpost-lesiontest(session 3) was given 36 days, the second (session 4) 38 days following lesioning.
STIMULATION A N D ABLATION IN THE AMYGDALA
173
lesioning in cat T is unexplained. Sham-lesion cats W, X, and Y also attacked more slowly in ‘postlesion’ tests. Both of the monopolar cats that had shown facilitation of attack during dual stimulation were lesioned. One, cat A, attacked faster following lesioning. The other, cat G, tended to attack more slowly following lesioning, but this tendency was not statistically reliable (P < 0.20). Cats with amygdaloid lesions were tested in the following behavioral situations : they were handled, placed near barking dogs, and placed together in their home cages with mice. No effects of our small amygdaloid lesions were observed in any of these situations. The destruction of a circumscribed region of the amygdala, in which stimulation generally produced suppression of attack, was followed by statistically significant increases in attack speed at fixed intensities of hypothalamic stimulation. All 5 cats that had lesions in this region attacked faster; not one of the 10 cats with lesions in other parts of the amygdala, or of the 6 cats without lesions, attacked faster. The probability of the consistency in location of the effective lesions occurring by chance among the 15 lesioned cats is less than 0.001. Discussion Although in his review, Gloor (1960) concluded that stimulation in the amygdala failed ‘to reveal any consistency in localization of the various amygdaloid stimulation responses’, others have subsequently published data supporting localization of function in the amygdala (e.g. Fonberg, 1967; Koikegami, 1963, 1964; Kreindler and Steriade, 1963, 1964; Ursin and Kaada, 1960). In our earlier experiment using bipolar stimulation in the amygdala, and in the present experiment in which we used monopolar stimulation in the amygdala, electrical stimulation in different regions of the amygdala produced opposing effects on hypothalamically elicited attack behavior. Some anatomical evidence suggests that these regions in the amygdala may be connected to the hypothalamus via different pathways (Fox, 1943; Ban and Omukai, 1959; Hall, 1963). Yoshida (1963) has also found that stimulation in the amygdala can affect responses elicited by hypothalamic stimulation in cats. Table I11 lists selected effects from studies of cats in which electrical stimulation elicited different effects from different regions in the amygdala. While differing in details, these studies of effects of stimulation are remarkably consonant with a functional division within the amygdala, with the dorsomedial amygdala in one functional category, and the lateral and ventral amygdala in another. This bipartite functional division of the amygdala should be considered only a first approximation of a more complicated organization. The dorsomedial division appears related to some sympathetic functions, arousal, and defense ;the lateral and ventral division to a heterogeneous grouping which Koikegami (1964) referred to as parasympathetic and extrapyramidal motor functions. Koikegami (1963) contrasted a proposed functional division of the amygdala with the traditional anatomical division of the amygdala into corticomedial and basolateral nuclear groups (Fig. 8). A similar functional division of References p. 180-182
174
M. D. EGGER A N D J. P. PLY"
Fig. 6. Bilateral amygdaloidlesions of monopolar cats A and F, and of bipolar cats M, N, and 0.A l l these lesions were followed by faster speeds of attack for fixed levels of hypothalamic stimulation. All lesions involved the border region between the magnocellular basal nucleusand the lateral nucleus.
the amygdala is consistent with many of the data in Table 111, as well as with those reported in this paper (Fig. 9). Table I11 does not include studies or effects that differ with the rest of the Table (e.g. Kaada, et al. 1954; the localizations of effects of stimulation on gastric motility, Shealy and Peele, 1957) or in which agreement was uncertain (e.g. Koikegami et al.,
S T I M U L A T I O N A N D A B L A T I O N I N THE A M Y G D A L A
175
Fig. 7. Bilateral amygdaloid lesions of monopolarcats G, E, and H, and of bipolar cats, P, Q, R, and S. None of these lesions was followed by a change in speeds of attack for fixed levels of hypothalamic stimulation.
1957; Kreindler and Steriade, 1963; Wood et al., 1958). However, the major portion of published data provides strong support for functional localization within the amygdala of the cat. Our experiments showed that hypothalamically elicited attack responses can be suppressed by stimulation in a region in the amygdala in which others have elicited References p . 1 8 6 1 8 2
176
M. D. EGGER A N D J. P. FLY"
f Fig. 8. Schematic diagrams showing two divisions of amygdaloid nuclei. The traditional anatomica division of the amygdala into corticomedial and basolateral divisions on the left is contrasted with a proposed functional division on the right. From Koikegami, 1963.
defense-like responses. However, we used intensities of amygdaloid stimulation below threshold for elicitation of behavior, hence lower intensities than those used by investigators studying defense reactions elicited by stimulation in the amygdala. In addition, attack directed at a rat is a different response from the defense reactions elicited by others during and following amygdaloid stimulation. Defense reactions are often associated with attack at a threatening object or person. But the attack response we used in our studies was directed preferentially at a rat (Egger, 1962; Levison and Flynn, 1965). Defense reactions are characterized by sympathetic activation. Our attack responses often occurred with minimal signs of sympathetic activation. It should not be surprisingthat regions related to defense might act to suppress attack at a non-threatening object such as a rat. Following bilateral amygdalectomy in cats, some investigators have observed increased docility (Schreiner and Kling, 1953;Brady et al., 1954; Shealy and Peele, 1957; Kling and Hutt, 1958), some have observed an increase in aggressive behavior and rage reactions (Spiegel et al., 1940; Bard and Mountcastle, 1948), and others have failed to detect any changes in affective behavior (Summers and Kaelber, 1962). The data on the effects of partial lesions of the amygdala in cats are still incomplete and less consistent than those on the effects of stimulation (Green et al., 1957; Horvath, 1963; Kling et al., 1960; Macchi et al., 1963; Morgane and Kosman, 1959; Nakao, 1960; Ursin, 1965a, 1965b; Wood, 1958). Small bilateral lesions in the amygdala in our cats produced consistent and reproducible effects on the latencies of hypothalamically elicited attack. Nakao (1960) found no effects of amygdaloid lesions on thresholds for eliciting a variety of behaviors during hypothalamic stimulation, not, however, including attack behavior. In 4 of Nakao's cats, lesioning of the amygdala was followed by faster speeds of performing a learned escape response to turn off hypothalamic stimulation. These 4 cats all had bilateral lesions involving the region that, when injured in our cats, was followed by faster attack speeds (Fig. 10). The cat in Nakao's study in which the lesions produced
177
STIMULATION A N D ABLATION IN THE AMYGDALA
T A B L E I1 E F F E C T S OF L E S I O N S I N THE A M Y G D A L A O N L A T E N C I E S OF H Y P O T H A L A M I C A L L Y ELICITED
ATTACK
Mean attack latencies (sec) Prelesion
cafi stimulation efects2
Bilateral lesions4 including ABM-ALDM
Bilateral lesions4 not including ABM-ALDM
M B OB A M N B F M P B E M Q B
H M D M R B S B
G M T B I M
No lesions
BM C M
VB YB WB XB
S***
S* F*** S***
S**
s*** S* S*
F*** S*
s*** s*** s*** F** F**
Low level of
hypothalamic stimulation3
High level of hypothalamic stimulation3
Statistical significance of post-lesion increase ( I ) or of post-lesion decrease ( D ) in attack speed
B
A
B
A
13.9 10.2 8.3 10.2 3.9
7.0 5.4 5.9 5.3 3.5
3.6 5.2 4.9 5.5 2.8
2.7 3.7 3.6 3.0 2.1
I*** I*** I*** I** I**
13.3 9.3 9.3 5.6 6.7 14.8 11.4 10.4 9.3 7.8
11.5 8.7 7.8 5.7 6.4 16.3 8.2 10.8 19.9
0
8.2 3.6 6.4 2.8 3.2 4.3 7.1 4.6 5.3 6.9
6.8 3.7 8.4 2.8 4.0 5.1 6.7 5.3 12.2
D**
5.9 7.0 8.5 8.6 4.5 15.1
7.0 6.3 6.1 17.1 6.3 20.0
4.1 4.0 4.5 6.6 3.0 4.7
4.0
0
3.4 5.5 10.6 5.5 12.4
D* D*** D***
‘B’refers to bipolar and ‘M’ refers to monopolar pre-lesion electrical stimulation in the amygdala. ‘S’ refers to suppression and ‘F’ refers to facilitation of hypothalamically elicited attack during simultaneous stimulation in the amygdala. 3 ‘B’ refers to attack latencies before lesioning; ‘A’ refers to attack latencies after lesioning. 4 ‘ABM-ALDM’ refers to the border region between the magnocellular portion of the basal nucleus of the amygdala and the dorsomedial portion of the lateral nucleus of the amygdala. 5 No attack responses could be elicited. See text. * P < 0.05;** P < 0.01; *** P < 0.001. 1 2
the smallest effect on escape latencies, cat 21 1, had the least involvement of this region. Fernandez de Molina and Hunsperger (1962) found no effect in 3 cats of bilateral lesions in the amygdala on latencies of hissing and fiight elicited by hypothalamic stimulation. But lesions in our cats followed by faster attack speeds had no effects on latencies of initial movements. In conclusion, much evidence supports localization of function in the amygdala in cats. Although denial on the basis of direct evidence (Kling et al., 1960) cautions us against unqualified acceptance of this attractive hypothesis, apparently contradictory reports on the effects of amygdalectomy may after all be explained on the basis of differences in site and extent of amygdaloid damage. References p. 180-182
178
M. D. EGGER A N D J. P. FLY”
Fig. 9. A functional division of the amygdala of the cat consonant with many published data on the effects of electrical stimulation in the amygdala. SUMMARY
Electrical stimulation in the magnocellular portion of the basal amygdaloid nucleus, and in the anterior and medial portions of the lateral amygdaloid nucleus generally suppressed hypothalamically elicited attack behavior in cats. Stimulation in the dorsolateral portion of the posterior part of the lateral nucleus facilitated the attack behavior.
STIMULATION A N D ABLATION I N THE AMYGDALA
-
179
..
3
2
I
0
Fig. 10. In these 4 cats, the amygdaloid lesions depicted were followed by faster speeds of escape from hypothalamic stimulation. From Nakao, 1960.
Lesions in the same region of the amygdala in which stimulation generally suppressed attack produced a facilitation of the hypothalamically elicited attack response. This facilitation was observed in all 5 cats with bilateral amygdaloid lesions involving the border region between the magnocellular basal nucleus and the lateral nucleus. None of the 6 sham-lesion cats, and not one of the 10 cats whose amygdaloid lesions failed to include this circumscribed region showed the facilitation. ACKNOWLEDGEMENTS
Most of the data reported here were collected while the senior author held a United States Public Health Service postdoctoral traineeship in the Department of Psychiatry, Biological Sciences Training Program, Yale University School of Medicine. This research was supported by grants from the National Science Foundation and the United States Public Health Service. We thank Mildred Groves, Zaven Khachaturian, and Dirk van Loon for technical assistance, and Louis G. Audette for drawing the figures. References p . 180-182
180
M. D. EGGER AND J. P. FLYNN
TABLE III SELECTED STUDIES OF EFFECTS OF ELECTRICAL STIMULATION IN THE AMYGDALA OF THE CAT
Experimenters
Egger and Flyp, 1963
Sites of stimulation tending to be in dorsomedial amygdala
Sites of stimulation tending to be in lateral and ventral amygdala
Suppression of hypothalamically Facilitation of hypothalamically elicited attack behavior elicited attack behavior Growling, hissing Sniffing, retching
Femandez de Molina and Hunsperger, 1959 Decreased blood pressure Gastaut, 1952; Morin et ul., Increased blood pressure 1952 Snitting, searching Hilton and Zbroiyna, 1963; Defense reaction, including prowlintz and extension of claws Zbroiyna, 1963 Inkbitionof knee jerk and of Facilitation of knee jerk and of Kaada, 1951 cortically induced movements cortically induced movements Koikegami and Fuse, 1952 Increased amplitude of respira- Decreased amplitude of respiration tion Inhibition of gastrointestinal Rise in body temperature Koikegami et al., 1952 motility Kreindler and Steriade, 1964 Acceleration-desynchronization Synchronizationof neocortical electrical activity in the form of neocortical electrical activity of spindles and slow waves Respiratory inhibition M a c h and Delgado, 1953 Respiratory acceleration Mastication Magnus and Lamers, 1956 Growling Effects other than arrest of Arrest of eating and mousing, Norris, Jr., 1963 activity or arousal arousal Cowering, snitling, licking Undirected rage Shealy and Peele, 1957 Decreased corticosteroid levels Increased corticosteroid levels Slusher and Hyde, 1961 in adrenal vein in adrenal vein Cowering, flight, searching Growling and hissing Ursin and Kaada, 1960 Respiratory acceleration, Respiratory inhibition, searching Wood, 1958 growling and biting Sneezing, seeking Yoshida, 1963 bge Facilitation of firing in short Inhibition of firing in short Zbroiyna, 1963 ciliary nerve ciliary nerve, with dilation of sympathectomized pupil
REFERENCES BAN,T., AND O m , F., (1959); Experimental studies on the fiber connections of the amygdaloid nuclei in the rabbit. J. comp. Neurol., 113,245-279. BARD, P., AND MOUNTCASTLE, V. B., (1948); Some forebrain mcchanisms involved in expression of rage with special reference to suppression of angry behavior. Res. Publ. Ass. nerv. rnent. Dis., 27, 362404. BRADY, J. V., SCHREINER, L., GELLER, I., AND KLING,A., (1954); Subcorticalmechanisms in emotional behavior: the effect of rhinencephalic injury upon the acquisition and retention of a conditioned avoidance response in cats. J. comp. physiol. Psychol., 47, 179-186. BmS, J., PETR,~~, M., AND ZACHAR, J., (1962); Electrophysiological Methods in Biological Research. New York, Academic Press. EGGER, M. D., (1962); Some effects of amygdaloid stimulation and ablation on hypothalamically elicited attack behavior in cats. Diss., Doctor of Philosophy. Graduate School, Yale University, New Haven, Conn. EGG=, M.D., (1965); Effects of amygdaloid stimulation on hypothalamic units. AbstractsofPapers, XXZZZ Znt. Congr. Physiol. Sci., Tokyo, p. 429.
STIMULATION A N D ABLATION IN THE AMYGDALA
181
EGGER, M. D., AND FLYNN,J. P., (1963); Effects of electrical stimulation of the amygdala on hypothalamically elicited attack behavior in cats. J. Neurophysiol., 26, 705-720. FERNANDEZ DE MOLINA, A., AND HUNSPERGER, R. W., (1959); Central representation of affective reactions in forebrain and brain stem: electrical stimulation of amygdala, stria terminalis, and adjacent structures. J. Physiol. (Lond.), 145, 251-265. FERNANDEZ DE MOLINA, A., AND HUNSPERGER, R. W., (1962); Organization of the subcortical system governing defense and flight reactions in the cat. J. Physiol. (Lond.), 160, 200-213. FONBERG, E., (1967); The role of amygdaloid nucleus in animal behaviour. Progr. Brain Res., Vol. 22 (in pres). FONFIERG, E., AND DELGADO, J. M. R., (1961); Avoidance and alimentary reactions during amygdala stimulation. J. Neuroph.vsiol., 24, 651-664. Fox, C. A., (1940); Certain basal telencephalic centers in the cat. J. comp. Neurol., 72, 1-62. Fox,C. A., (1943); The stria terminalis, longitudinal association bundle and precommissural fomix fibers in the cat. f.comp. Neurol., 79, 277-295. GASTAUT, H., (1952); Corrdlations entre le systeme nerveux vdgdtatifet le systbme de la vie de relation dans le rhinenckphale. J. Physiol. (Paris), 44,431470. GLOOR, P., (1960); Amygdala (Chapt. 48). Hancibook of Physiology, Neurophysiology. J. Field, H. W. Magoun and V. E. Hill, Editors. 11, pp. 1395-1420. GREEN,J. D., CLEMENTE, C. D., AND DEGROOT, J., (1957); Rhinencephalic lesions and behavior in cats. An analysis of the Kliiver-Bucy syndrome with particular reference to normal and abnormal sexual behavior. J. comp. Neurol., 108, 505-545. HALL,E. A., (1963); Efferent connections of the basal and lateral nuclei of the amygdala in the cat. Amer. J. Anat., 113, 139-151. HILTON,S. M., AND ZBROZYNA, A. W., (1963); Amygdaloid region for defense reactions and its efferent pathway to the brain stem. J. Physiol. (Lond.), 165, 16CL173. HORVATH, F. E., (1963); Effects of basolateral arnygdalectomy on three types of avoidance behavior in cats. J. conip. physiol. Psychol., 56, 380-389. KAADA,B. R., (1951); Somato-motor, autonomic and electrocorticographic responses to electrical stimulation of ‘rhinencephalic’ and other structures in primates, cat and dog. Acta physiol. scand., SUPPI.83,24, 1-285. KAADA,B. R., A~DERSEN, P., AND JANSEN, JR., J., (1954); Stimulation of the amygdaloid nuclear complex in unanesthetized cats. Neurology, 4, 48-64. KLING,A., AND HUTT,P. J., (1958); Effect of hypothalamic lesions on the amygdala syndrome in the cat. Arch. Neurol. Psychiat. (Chic.), 79, 511-517. KLING,A., ORBACH, J., SCHWARTZ, N. B., AND TOWNE, J. C., (1960); Injury to the limbic system and associated structures in cats. Arch. gen. Psychiot., 3, 391-420. KOIKEGAMI, H., (1963); Amygdala and other related limbic structures; experimental studies on the anatomy and function. I. Anatomical researches with some neurophysiological observations. Acta med. biol. (Niigata), 10, 161-277. KOIKEGAMI, H., (1964); Amygdala and other related limbic structures; experimextal studies on the anatomy and function. 11. Functional experiments. Acta med. biol. (Niigata), 12, 73-266. KOIKEGAMI, H., DODO,T., MOCHIDA, Y., AND TAKAHASHI, H., (1957); Stimulation experiments on the amygdaloid nuclear complex and related structures. Effects upon the renal volume, urinary secretion, movements of the urinary bladder, blood pressure and respiratory movements. Folia psychiat. neurol. jap., 11, 157-206. KOIKEGAMI, H., AND FUSE, S., (1952); Studies on the functions and fiber connections of the amygdaloid nuclei and periamygdaloid cortex. Experiments on the respiratory movements (2). Folia psjchiat. neurol. jap., 6, 96103. KOIKEGAMI, H., KUSHIRO, H., AND KIMOTO, A., (1952); Studies on the functions and fiber connections of the amygdaloid nuclei and periamygdaloid cortex. Experiments on gastrointestinal motility and body temperature in cat. Folia psyehiut. neurol. jup., 6, 76-93. KREMDLER, A., AND STERIADE, M., (1963); Functional differentiation within the amygdaloid complex inferred from peculiarities of epileptic afterdischarges. Electroenceph. clin. Neurophysiol., 15, 811-826. KREINDLER, A., AND STERIADE, M., (1964); EEG patterns of arousal and sleep induced by stimulating various amygdaloid levels in the cat. Arch. ital. Biol.. 102, 576586. LEMSON, P. K., AND FLY”, J. P., (1965); The objects attacked by cats during stirnulation of the hypothalamus. Anim. Behav., 13,217-220. MACCHI, G., CARRERAS, M., AZZALI,G., DALLA ROSA,V., AND LECHI,A., (1963); Rinencefalo e com-
182
M. D. EGGER A N D J. P. F L Y N N
portamento istintivo: Effetti provocati da lesioni operate sul lob0 piriforme. (Studio sperimentale nel gatto). Boll. Soc. ital. Biol. sper., 39, 37-41. MACLEAN, P. D., AND DELGADO J. M. R., (1953);Electricalandchemical stimulation offrontotemporal portion of limbic system in the waking animal. Electroenceph. clin. Neurophysiol., 5, 91-100. MAGNUS, O., AND LAMME=, H. J., (1956); The amygdaloid-nuclear complex. Folia psychiat. neerl., 59, 555-582. MORGANE,P. J., AND KOSMAN, A. J., (1959); A rhinencephalic feeding center in the cat. Amer. J . Physiol., 197, 158-162. M o m , G.,NAQUET,R., AND BADJER,M., (1952); Stimulation electrique de la region amygdalienneet pression artbrielle chez le Chat. J. Physiol. (Paris), 44, 303-305. NAKAO,H., (1960); Hypothalamic emotional reactivity after amygdaloid lesions in cats. Folia psychiat. neurol.jap., 14,357-366. NORRIS, JR., F. H., (1963); Arrest of activity by temporal lobe stimulation. Neurology, 13, 895-898. SAWA,M., MARUYAMA, N., HANAI, T., AND UI, S., (1959); Regulatory influence of amygdaloid nuclei upon the unitary activity in ventromedial nucleus of hypothalamus. Folia psychiat. neurol. jap., 13, 235-256. SCHREINER, L.,AND KLING,A., (1953); Behavioral changes following rhinencephalic injury in cat. J. Neurophysiol., 16, 643-659. SHEALY,C. N., AND PEELE, T. L., (1957); Studies on amygdaloid nucleus of cat. J . Neurophysiol., 20, 125-1 39. SLUSHER, M. A., AND HYDE,J. E., (1961); Effect of limbic stimulation on release of corticosteroids into the adrenal venous effluent of the cat. Endocrinology, 69, 1080-1084. SPIEGEL, E. A., MILLER, H. R., AND OPPENHEIMER, M. J., (1940); Forebrain and rage reactions. J. Neurophysiol., 3, 538-548. STUART,D. G., PORTER, R. W., AND ADEY,W. R., (1964); Hypothalamic unit activity. 11. Central and peripheral influences. Electroenceph. d i n . Neurophysiol., 16, 248-258. Sumwxs, T. B., AND KAELBER, W. W., (1962); Amygdalectomy: effects in cats and a survey of its present status. Amer. J. Physiol., 203, 1117-1119. SZENTAGOTHAI, J., FLERK~, B., Mess, B., AND HALASZ,B., (1962); Hypothlumic Control of the Anterior Pituitary. Budapest, Akadkmiai Kiad6. TSUBOKAWA, T., AND S m , J., (1963); Mesencephalicinfluence upon the hypothalamic ventromedial nucleus. Electroenceph. clin. Neurophysiol., 15, 804-810. URSIN,H., (1965a); The effect of amygdaloid lesions on flight and defense behavior in cats. Exp. Neurol., 11, 61-79. URSIN,H., (1965b); Effect of amygdaloid lesions on avoidance behavior and visual discrimination in cats. Exp. Neurol., 11,298-317. URSIN,H., AND KAADA, B. R., (1960); Functional localization within the amygdaloid complex in the cat. Electroenceph. din. Neurophysiol., 12, 1-20. VALVERDE, F., (1963); Amygdaloid projection field. The Rhinencephalon and Related Structures, Vol. 3, Progress in Brain Research. W. Bargmann and J. P. Schadk, Editors. Amsterdam, Elsevier, pp. 20-30. WASMAN, M., AM) FLY", J. P., (1962); Directed attack elicited from hypothalamus. Arch. Neurol., 6, 220-227. WENDT,R. H., AND ADEY,W. R., (1960); A study of evoked unit activity in the hypothalamus. Anat. Rec., 136, 301 (abstract). WOOD,C. D., (1958); Behavioral changes following discrete lesions of temporal lobe structures. Neurology, 8, 215220. WOOD,C. D., Scxo'rm~ms,B., FROST,L. L., AND BALDWIN,M., (1958); Localization within the amygdaloid complex of anesthetized animals. Neurology, 8,477480. YOSHIDA,M., (1963); Effects of amygdaloid stimulation on emotional responses produced by hypothalamic stimulation in cats. Psychiat. neurol. jap., 65, 863-879 (English summary, pp. 71-72). ZBROZYNA, A. W., (1963); The anatomical basis of the patterns of autonomic and behavioural response effected via the amygdala. The Rhinencephalon and Related Strucrttres, Vol. 3, Progress in Bruin Research. W. Bargmann and J. P. Schadk, Editors. Amsterdam Elsevier, pp. 5 M 8 .
183
Influence of Labyrinthine Stimulation on Hippocampal Activity A. COSTIN, F. B E R G M A N N
AND
M. CHAIMOVITZ
Department of Pharmacology, Hebrew University, Hadassah Medical School, Jerusalem (Israel)
We have shown previously that labyrinthine stimulation has a direction-dependent influence on optic nystagmus. E.g. when both labyrinthine and optic nystagmus are directed towards the same side, they enhance each other, while antagonistic stimulations produce mutual depression (Bergmann and Costin, 1966). Similarly, sideposition - a vestibular stimulus that does not evoke nystagmus - enhances ipsiversive optic nystagmus and inhibits contraversive eye movements (Costin et d.,1966). These phenomena can be satisfactorily explained by the non-symmetrical distribution of the fibers in the photic and vestibular pathways. Th&observations mentioned suggest the possibility that labyrinthine stimulation may also exert an asymmetric influence on other responses induced by unilateral stimulation of a cerebral structure. The present study shows that under certain conditions labyrinthine nystagmus to left or right has a different influence on the afterdischarge following unilateral electrical stimulation of the dorsal hippocampus. METHODS
Adult rabbits of either sex, weighing between 2 and 3.5 kg, were used. All surgical procedures were carried out under ether. At the end of the operation the wounds were infiltrated with 2 % xylocain, and the animal was allowed to recover from general anesthesia, before the experiment was started. Three pairs of bipolar, concentric electrodeswere placed in the dorsal hippocampus, at positions bC-bD (Monnier and Gangloff, 1961). One pair served for stimulation and the other two for recording of the left and right hippocampal potentials on a Schwarzer electroencephalograph. At the end of the experiment the animal was killed, and the position of the electrodes was checked by histological examination. For angular acceleration and deceleration, the rabbit was placed on a Toennies rotating chair. After a predetermined angular velocity had been attained, the rotation was maintained at constant speed for 60 sec before deceleration began. In order to avoid interference by photic stimuli, both eyes were kept covered with small black hoods throughout these experiments. References p . 188
184
A. COSTIN
et al.
a
b
C
t
I
1
5
10
15
20
25
200pv
30
35 sec
Fig. 1. Influence of perrotatory nystagmus on the discharges of the resting hippocampus. Male rabbit, 2 kg. Counterclockwise acceleration at 10"/sec2(between arrows) evokes 14 eye beats to left (see nystagmogram in c). After 10 sec, when an angular velocity of lOO"/sechad been reached, the rotating chair was maintained at this speed. After-nystagmus 9 beats during 12 sec. Note that concomitantly with the nystagmus, a 0-rhythm appears in both records, (a) from the right, and (b) from the left dorsal hippocampus.
a
b
C
I
t 1
I
I
I
I
I
200pv I
I
5 10 15 20 25 30 35 sec Fig. 2. Influence of sudden arrest of rotation on the potentials of the resting hippocampus. The same rabbit as used for Fig. 1. After the animal had been rotated counterclockwise for 1 min at a constant speed of lOO"/sec,it was suddenly arrested (at arrow) causing 26 eye beats to the right during 11 sec (c). The nystagmus caused fast, synchronized potentials in both hippocampi, outlasting the eye movements for about 20 sec (a and b).
L A B Y R I N T H I N E S T I M U L A T I O N OF H I P P O C A M P U S
185
Fig. 3. Asymmetric effect of angular accelerationon hippocampal after-discharge (HAD). Male rabbit, 2.6 kg. Stimulation of left dorsal hippocampus at bC, for 5 sec (arrows); 40 c/sec, pulse duration 2 msec, current strength 0.15 mA. Recording electrodes in left hippocampus at bD. (a) Control: hippocampal stimulation, without rotation, produces an HAD of 30 sec duration; (b) simultaneous counterclockwiseacceleration at 10"/sec2,fcr 10 sec; reduction of HAD to 3 sec; (c) simultaneous clockwise acceleration at 10"/sec2,for 10 sec; prolongation of HAD to 60 sec. RESULTS
I . Effect of labyrinthine stimulation on the discharges of the resting hippocampus In the awake, non-stimulated animal, angular acceleration evoked a hippocampal 8rhythm, expressing itself in synchronization and in an increase in amplitude. In Fig. 1, counterclockwise acceleration at lO"/secZ produced during 10 sec 14 eye beats to the left; the changes in the hippocampogram appeared simultaneously with the first eye movement. When after 10 sec a velocity of 100"/sec was maintained, perrotatory nystagmus persisted for another 7 sec, but with decreasing strength. At the same time, the hippocampal potentials decayed and returned to their resting pattern. Analogous observations were made during deceleration. The duration of the hippocampal aftereffect clearly depends on the strength of the vestibular stimulus. Thus in Fig. 2, after sudden arrest of the rotating chair the synchronization outlasted the postrotatory nystagmus considerably. The changes in the left and right hippocampus parallel each other, and either side may be used to measure the effect of clockwise or counterclockwise rotation. II. Effect of labyrinthine stimulation on the after-discharge, following hippocampal excitation In Fig. 3a, electrical stimulation of the left hippocampus for 5 sec produced an afterdischarge for 30 sec. Counterclockwise acceleration at 1O"/secz reduced this period to 3 sec (Fig. 3b), while clockwise acceleration prolonged it to 60 sec (Fig. 3c). The directional effect of perrotatory nystagmus could be demonstrated only with difficulty in References p . I88
1616
A. COSTIN
et ai.
experiments in which the hippocampal after-discharge (HAD) was short (see Table IA), but if the latter lasted for more than 30 sec, the inhibitory effect of labyrinthine nistagmus became manifest. On the other hand, the enhancement was sometimesweak or absent, even when the HAD extended over a considerable period. The data in the table establish the rule that vestibular nystagmus to the left inhibits the HAD evoked from the left hippocampus. It was observed previously (Gangloff and Monnier, 1957; Costin et al., 1963) that the HAD can be considerably prolonged by application of chlorpromazine. This drug TABLE I EFFECT OF LABYRINTHINE STIMULATION O N HIPPOCAMPAL AFTER-DISCHARGE
(HAD)
Duration of HAD (see) Side of hippocampal stimulation
( A ) Untreated animals : Left Right Right Left Left Left Left Left Right Right Right
Control
Clockwise acceleration
Counterclockwise acceleration
13 28 25
11 28 18
15 24 23
21 29
30 40 50 115
75 35 70 130 60 13 14 5
7 16 15 20 5 55 50 110
12 80 20 100
6 13 22 140
12 90 22 40
70 117
(B) Before and after treatment with 2 mg/kg chlorpromazine i.v. : Right (before) Right (after) Left (before) Left (after)
TABLE 11 LOSS OF ASYMMETRIC INHIBITORY INFLUENCE O N
HAD
W I T H INCREASING ANGULAR
ACCELERATION
Female rabbit, 3 kg; stimulation of left dorsal hippocampus at bC for 5 sec;40 c/sec, 2 msec, 1.5 mA. Angular accelerations lasted for 10 sec. Direction of rotation
Control Counterclockwise clockwise Counterclockwise Clockwise
Angular acceleration
Duration of HAD (see) 130 20 130 4 7
187
L A B Y R I N T H I N E S T I M U L A T I O N OF H I P P O C A M P U S
TABLE I11 LONG-LASTING INHIBITORY EFFECT OF LABYRINTHINE STIMULATON ON HIPPOCAMPAL AFTER-DISCHARGE
(HAD)
Male rabbit, 2.8 kg; stimulation of left dorsal hippocampus for 5 sec at 40 c/sec, 2 msec, 0.3 mA. Counterclockwiseacceleration at 20"/sec2,for 10 sec. ~
Condition
Duration of HAD (see)
Hippocampal stimulation only (control) Simultaneous counterclockwiseacceleration Hippocampal stimulation only, 7 min later Hippocampal stimulation only, 14 min later Hippocampal stimulation only, 24 rnin later Hippocampal stimulation only, 36 rnin later
65
10
17 20 29
70
permitted the demonstration of the asymmetric effect of vestibular nystagmus on the HAD even in animals that before treatment were not affected by angular acceleration (Table IB). The asymmetric effects of Fig. 3 can be demonstrated only when the intensity of labyrinthine stimulation is not too high. Angular accelerations above 20"/sec2 usually produced inhibition of the HAD, irrespective of the direction of rotation (Table 11). Although after a short while the hippocampal potentials reassume their resting appearance, the after-effect of labyrinthine stimulation has not passed. If hippocampal stimulation is repeated without rotation, the inhibitory effect remains manifest for about 30 min (see Table 111). On the other hand, the enhancement of the HAD, represented in Fig. 3c, vanishes shortly after the end of the rotation. DISCUSSION
At first glance, no relation is apparent between vestibular nystagmus and HAD. However, Spiegel(l932) has shown that labyrinthine stimulation facilitates cortical seizures. The present findings on hippocampal discharges thus enlarge the observations of this author. The experiments described here establish a pronounced asymmetric effect on the HAD, following labyrinthine stimulation under certain conditions. The duration of after-discharges, evoked by stimulation of the dorsal hippocampus, is reduced by ipsiversive nystagmus. On the other hand, the effect of contraversive nystagmus movements is not as clear-cut, since the HAD sometimes remains unchanged while in others it is markedly prolonged (see Fig. 3c and Table I). It is most conspicuous that a more intense vestibular stimulus, such as increased angular acceleration,may produce inhibition even when hippocampal excitation is combined with contraversive nystagmus (see Table 11). These observations, together with the identical effect of clockwise and counterclockwise rotation on the discharges of the resting hippocampus (Figs. 1 and 2), suggest that each vestibular pathway has connections to both hippocampi, forming both References p . 188
188
A. COSTIN
et al.
excitatory and inhibitory synapses. However, the quantitative distribution of the two types of synapses to either side appears to be unequal. In addition, the effect of vestibular impulses, arriving at the hippocampus, also depends on the state of activity of the latter structure. Electrical excitation of the hippocampus produces a well-organized pattern of discharges. The stimulated region probably serves as pacemaker for the hippocampal neurons on both sides (Gutman et al., 1963). This means that the activity spreading from the pacemaker suppresses all independent, local discharges. It appears possible that this mechanism also paves the way for the inhibitory effect of labyrinthine nystagmus on the HAD. REFERENCES BERGMANN, F., AND COSTIN,A., (1965); Studies on the physiological relationship between retina and labyrinth. Israel J. Med. Sci., 1, 13661372. COSTIN, A., BERGMANN, F., AND CHAIMOvITz, M., (1966); Influence of side position of the head on central and flash nystagmus in the rabbit. Acta oto-laryng., 61, 323-331. Cam, A., GUTMAN,J., AND BERGMANN, F., (1963); Relationship between caudate nucleus and dorsal hippocampus in the rabbit. Electroenceph. elin. Neurophysiol., 15, 997-1005. GANGLOFF, H., AND MONNIER, M., (1957); Topic action of reserpine, serotonin and chlorpromazine on the unanesthetized rabbit's brain. Helv. physiol. pharmacol. Ada, 15,83-104. GUTMAN, J., a m ,A., AND BERGMANN, F., (1963); Constancy of hippocampal afterdischargeunder various conditions of stimulation. Electroenceph. clin. Neurophysiol., 15, 989-996. MONNIER, M., AND GANGLOFF,H., (1961); Rabbit Brain Research, Vol. 1. Elsevier, Amsterdam. SPIEGEL, E. A., (1932); Rindenemgung (Auslosung epileptiformer Anfalle) durch Labyrinthreizung. Versuch einer Lokalization der corticalen Labyrinthzentren. Z. ges. Neurol. Psychiat., 138,178-196.
189
Studies on the Neurovegetative and Behavioral Functions of the Brain Septa1 Area MIGUEL R. COVIAN School of Medicine, Department ofphysiology, Aibeirrio Prgto, S.P. (Brazil)
The septal area is a small structure of the limbic system almost completely surrounded by cerebrospinal fluid, rostra1 and anterior to the anterior commissure, ventral to the splenium of the corpus callosum, related to autonomic, reproductive, emotional and other behavioral responses. We have studied in anesthetized cats, rats and rabbits, and in unanesthetized cats, the effects of electrical septal stimulation upon blood pressure and respiration (Part I). In anesthetized rabbits a conditioned blood pressure response was obtained (Part 11). In rats the role of the septal area on drinking behavior for water and NaC1, using the standard ‘two bottle’ self-selection procedure, was investigated (Part 111). Part I was done in collaboration with J. Antunes-Rodrigues, E. M. Krieger, J. J. O’Flaherty and C. Tim0 Iaria; Part I1 with M. C. Lico; and Part I11 with C. G. Gentil and A. Negro Vilar. (I) EFEECTS OF S T I M U L A T I O N OF T H E S E P T A L A R E A U P O N B L O O D P R E S S U R E
A N D RESPIRATION
The experimental data available on the central nervous control of cardiovascular activity, as well other bodily activities, show that this control is exerted at different levels of the neuroaxis. Since the pioneer work of Karplus and Kreidl (1909,1910) the hypothalamus has been considered one of the most important structures controlling neurovegetative functions. Changes in blood pressure and respiration following the electrical stimulation of several areas of the cerebral cortex in animals and man have also been reported; Delgado (1960) has published a good review on this subject. During recent decades, as a result of investigations carried out in several laboratories, increasing emphasis is being placed on the role played by the limbic system in the regulation of visceral functions and behavior. The first part of our presentation will deal with the blood pressure and respiratory changes elicited in anesthetized and unanesthetized animals by electrical septal stimulation. Anesthetized animals. Cats and rats were anesthetized with a-chloralose, 70 mg/kg i.v. Rabbits received urethane, 1.25 mg/kg given in half doses i.p. and i.v. respectively. Conventional stereotactictechniques were used to introduce bipolar electrodesinto the References p . 215-217
190
M. R. C O V I A N
septal area according to the co-ordinates of Jasper and Ajmone-Marsan (1954) for cat, of De Groot (1959) for rat, and of Sawycr et al. (1954) for rabbit. An AEL stimulator with an isolation unit was used to deliver unidirectional rectangular pulses of 50 c/sec, 10 msec of pulse duration and an intensity of 3-5 V. Systemic arterial pressure was measured with a Statham P23-AA transducer connected to a catheter placed in a femoral artery. Respiration was recorded through a pneumograph attached around the chest. Changes in blood pressure and respiration were recorded synchronously on a polygraph. A tracheal cannula was inserted into each animal, and artificial respiration was used when required. Body temperature was maintained by radiant heat and measured with a telethermometer with a probe introduced in the colon. At the end of each experiment the head was perfused with 4% formaldehyde, and the brain submitted to routine histological procedures to determine the position of the stimulating electrodes.
Results The effects produced by septal stimulation on blood pressure and respiration in the cat have been reported elsewhere (Covian et al., 1963);they will be brieffly summarized and some new unpublished results will be described in more detail. In 21 of 25 intact cats a fall in blood pressure was elicited that outlasted the electrical stimulation, applied during 30 sec, by about 3-5 min. When a pure depressor effect was elicited it began gradually during the period of stimulation, and was accentuated after the end of the stimulus; however, when respiratory changes were present the latency of the response was shorter. Occasionally both changes were concomitant but either effect could appear independently of the other. The fall in blood ptessure was sometimes preceded by a slight pressor reaction, and in some cats the depressor effect was only observed after the end of stimulation. Fig. 1 shows the depressor reactions obtained in one cat, without any significant change in either the heart rate or the frequency or depth of respiration. The blood pressure measurements made before, during, and after the stimulation were 162/122, 138/98, and 95/64 mm Hg respectively; the fall in pressure lasted for 5 min after stimulation was stopped : in the last period there was a diminution in pulse pressure due to the greater fall of the systolic pressure. Fig. 2
Fig. 1. Depressor reaction in the chloralosed cat. Blood pressure (lower record) values before, during and after electricalstimulation (solid line) of the septal area: 162/122,138/98 and 95/64 mm Hg. The fall lasted for 5 min after cessation of stimulation. No change in respiration (upper record) or heart rate.
NEUROVEGETATIVE A N D BEHAVIORAL FUNCTIONS OF THE SEPTAL
AREA^^^
Fig. 2. Depressor reaction in the chloralosed cat. Blood pressure (upper record) values before, during and after electrical stimulation (solid line) of septal area: 142/98,117/72and 75/30 mm Hg. The fall outlasted by 4 min the stimulation. Respiration (lower record) showed an expiratory apnea of 12 sec during stimulation. Heart rate dropped 12 beats/min.
Fig. 3. Depressor reaction in the chloralosed rat. Mean blood pressure (lower record) fell 45 mm Hg during and after electricalstimulation of the septal area. The fall lasted for 3 min 50 sec after cessation of stimulation.No signifcant change in respiration (upper record) occurred.
Fig. 4. Diagram showing the electrode location in rat of Fig. 3. References p. 215-21 7
192
M. R. COVIAN
Fig. 5. Respiratory change (upper record) in chloralosed rat without any change in blood pressure (lower record) during electrical stimulation of the septal area.
Fig. 6. Diagram showing the electrode location in rat of Fig. 5.
depicts the result elicited in another cat in which the alterations in blood pressure were accompanied by mod5cations in respiration and heart rate. The arterial pressure values for the three periods mentioned above were 142198,117172 and 75/30mm Hg; the fall in pressure outlasted the stimulus by 4 min. There was an expiratory apnea during the first 12 sec of stimulation; afterwards normal frequency and amplitude of respiration were regained slowly. Bradycardia occurred during the period of stimulation and persisted during the poststimulatory pressure drop. The heart rate was 180, 160 and 168/min for the three periods mentioned above. In rats the stimulation of the septal area also elicited a blood pressure fall as is shown in Fig. 3. Immediately after the onset of the stimulation there was a small rise
N E U R O V E G E T A T I V E A N D B E H A V I O R A L FUNCTIONS OF THE SEPTAL AREA
193
Fig. 7. Blood pressure (lower record) and respiration (upper record) in the chloralosed cat before, during, and after electrical stimulation (solid line) of the septa1 area. A, intact animal; B, after bilateral cervical vagotomy ;C, after atropine injection (1 mg/kg i.v.).
in blood pressure followed by a fall which persisted for 3'50" after the end of stimulation. The mean blood pressure values before and during the stimulation were 115 and 70 mm Hg respectively. The stereotaxic settings were F, 7.5; L, 1.0; H, +3. The electrode position is shown in Fig. 4. In some rats only respiratory changes were observed (Fig. 5). During stimulation an increased amplitude of respiratory movements was elicited. The electrode was placed at F, 7.5; L, 1.0; H, 1.0. In another horizontal
+
References p. 215-217
194
M. R. C O V I A N
f
160r
I
80
A
f
160
80 40
I -
5sec
Fig. 8. A, blood pressure fall (lower record) elicited by septal stimulation in the chloralosed rat. B , after the administration of atropine (1 mg/kg i.v.) the depressor reaction still remained.
plane (+2) the respiratory changes were accompanied by a fall in blood pressure. The electrode position is shown in Fig. 6. Midcervical bilateral vagotomy did not abolish the blood pressure fall in cats and rats. In Fig. 7 this is shown for the cat. The effect of septal stimulation was not blocked by atropine in either animal, as shown in Fig. 8 for the rat. To rule out any possible interference of muscular contraction on the depressor reaction, flaxedil was injected in some cats: after flaxedil the blood piessure fall still remained, as can be observed in Fig. 9. The possibility of explaining the effects observed by a substance released by septal stimulation was tested by the crossed circulation technique, and by plasma injection from the stimulated animal into an intact rat. These procedures failed to prove the possibility suggested (Covian et al., 1964). Bradycardia was a concomitant feature very often observed in cats and rats under
NEUROVEGETATIVE A N D B E H A V I O R A L F U N C T I O N S OF THE S E P T A L AREA
195
Fig. 9. A, blood pressure (lower record) fail following septal stimulation in the chloralosed cat. B,
after flaxedil administration.
chloralose. In Fig. 10 this result is shown for the cat: in the upper record the values of the systolic and diastolic blood pressure are measured every 20 sec, and in the lower record the beats per minute are plotted. It is observed that during stimulation there was a diminution of 15 beats/min. When the stimulus ceased the diminution was accentuated, the rate falling to 40 beats/min. The pre-stimulatory rate was regained 5 min after the withdrawal of the stimulus. Eserine (100 ,ug/kg i.v.) enhanced the fall of the heart beat and its duration. Nevertheless in some cats the bradycardia still remained after vagal sections. This observation suggests the possibility of an inhibition of the cardiocelerator fibers playing a role in the bradycardia due to septal stimulation. Baroreceptor reflex. The interplay between septal stimulation and baroreceptor reflex was studied in the cat. The procedure was as follows: the baroreceptor reflex was obtained by bilateral carotid occlusion, and tested at different intervals during the hypotension elicited by septal stimulation. Fig. 11 shows at A: the control baroreceptor reflex, at B, during the last 5 sec of septal stimulation, at C , 30 sec, at D, 1 min, at E, 1.5 min, at F, 2 min and at G , 3 min after the stimulus withdrawal. In this sequence the diminution of the baroreceptor reflex during hypotension due t J septal stimulation is considerable. References p . 215-21 7
196
M. R. COV IA N
:"I
I40
Fig. 10. Effect of septal stimulation on blood pressure (A) and heart rate (B) in the chloralosed cat.
In (A), upper circles, systolic blood pressure; lower circles, diastolic blood pressure (both measured every 20 sec, as well as heart rate). Solid line, stimulation time 3 min.
Fig. 11. Interaction between septal stimulation and baroreceptor reflex in the chloralosed cat. A, central baroreceptor reflex elicited by bilateral carotid occlusion; B, during the last 5 sec of septal stimulation; C,30 sec after stimulus withdrawal; D, E, F, G, 1 min, 1.5 min, 2 min, and 3 min, respectively after stimulus withdrawal.
Fig. 12. Increased blood pressure (lower record) during stimulation of the septal area without change in respiration (upper record) in the chloraiosed cat. The values were 104p8, 142/90, and 106/60 before, during and after stimulation.
N E U R O V E G E T A T I V E A N D B E H A V I O R A L FUNCTIONS OF THE S E P T A L A R E A
197
Fig. 13. Microphotographsto illustrateelectrode locations in cats that showed depressor responses to septal stimulation. A, anterior septal area; B, anterior septal area, electrode more medially placed; C, median septal area; D, posterior septal area.
Hypertensive efects. In a few cats and rats a pressor reaction was also elicited. Contrary to the depressor effect, rises in blood pressure occurred very quickly after the onset of stimulation, and they never outlasted the period of stimulation. Fig. 12 is the record obtained in one cat in which the systolic blood pressure increased by 38 mm Hg and the diastolic by 32 mm Hg during stimulation. Both in cats and rats the hypertensive effect was inhibited by dibenamine. Respiratory effects. In both species changes in the amplitude and/or frequency of respiratory movements as well as apnea, either in inspiration or expiration, were obtained during stimulation. They were concomitant sometimes with blood pressure changes, but either effect could appear independently of the other. The location oj’sites stimulated. The histological examination of the septal area of all cats and rats showed that the effects were independent of electrode position, for depressor effects were elicited from the anterior, median and posterior zones, as well as from the medial and lateral parts of the septum. Nevertheless, some points were more ‘respiratory’ than ‘depressor’, and in some animals the moving of the electrode 1 mm down or laterally changed the type of response. In one rat and in one cat a fall in blood pressure was changed to a rise when the electrode was more medially placed. Because of these findings a mapping study of the septal area in the cat is under way. Fig. 13 illustrates the electrode placement in the brains of some cats. Rabbits. In rabbits anesthetized with urethane, stimulation of the septal area also resulted in a fall of blood pressure. This depressor reaction was sometimes very persistent as is shown in Fig. 14. Between stimulation numbers 6 and 46 about 6 h elapsed; References p . 215-21 7
N V I A 0 3 2I 'W
86 I
NEUROVEGETATIVE A N D BEHAVIORAL F U N C T I O N S O F THE SEPTAL AREA
180 7
.
*a**..
160
0
. . a
140.
1
*:.a.
*
a
*a
a .
2 '
'
.
4'
6'
"I
24
26
**a****
***++*
**
120
60
28
.
H
b b
a*.
32rnin
30
ti
.* 140-• **
80
**a
.
1
loo -
**a
*.*
I *
I20
I6O
.//*.. .*. . .*.a
'
199
.
L *
*aa
0..
.*a
**
**
...
**a**.
~
a
-
1 -,+,
0.
.
0.
. , , 40 r 2
4
,I
1
,
.
28
30
32
34
,
36
,
38min
Fig. 15. Effect of stimulationtime upon blood pressure fall due to septal stimulation in the chloralosed cat. G , 23 min; H, 28.5 min. Upper circles, systolic blood pressure; lower circles, diastolic blood pressure (both measured every 20 sec). Voltage, frequency and pulse length were constant.
note the constancy of the response obtained with no change in the position of the electrode or the parameters of stimulation. It is relevant that in cats we have shown (Covian and Timo-Iaria, 1966), that the depressor reaction is maintained during a stimulation time prolonged to 28.5 min as illustrated in Fig. 15. In rabbits also, as in cats and rats, a pressor reaction was obtained once in a while, as shown in Fig. 16. When the electrode was moved 1 mm the depressor reaction was changed to a pressor one. Unanesthetized cuts. In cats with bipolar electrodes implanted in the septal area, and a permanent polyethylene cannula in the abdominal aorta, the results of electrical septal stimulation were studied. As happened in anesthetized cats, a drop in blood pressure was observed together with bradycardia (Fig. 17). After withdrawal of the References p. 215-21 7
200
M. R. COVIAN
Fig. 16. Effects of stimulation of the septal area upon blood pressure (lowerrecord) in the rabbit under urethane anesthesia. A, pressor reaction; B, depressor reaction. Both stimulations were made within a short interval. Parameters of stimulation were the same.
Fig. 17. Blood pressure fall in the mesthetized cat due to septal stimulation. Bradycardia was also elicited. Stimulation time, 30 sec.
N E U R O V E G E T A T I V E A N D B E H A V I O R A L F U N C T I O N S OF THE S E P T A L
AREA201
Fig. 18. Depressor reaction in the manesthetized cat. A, stimulation time, 2 sec;B, 4 sec.
stimulus the fall as well as the bradycardia persisted and both regained the prestimulatory values after 3 min had elapsed. Brief stimulation times (2 4,s and 16 sec) also elicited a fall in blood pressure as shown in Figs. 18 and 19. Bradycardia, without significant changes in blood pressure, was elicited with a lower stimulus intensity, and it was abolished by atropine (Fig. 10). Comments
The effects of electrical stimulation of the septal area upon blood pressure and respiration in the curarized cat (Torii, 1961), in the chloralosed cat (Covian et al., 1964; Manning et al., 1963) and in chloralosed and conscious dogs (Gorten et al., 1964) have recently been reported. The septal area is one of the anatomical components of the limbic system, the phylogenetically oldest structure of the cerebral hemispheres and a common denominator in the brain of all mammals, as was pointed out by Broca. MacLean (1949,1958,1959) has made it clear that man shares with other mammals the physiological properties of this system. Changes in blood pressure and respiration evoked by septal stimulation have been reported before by several investigators, but the depressor reactions were not subReferences p . 215-21 7
202
M. R. COVIAN
Fig. 19. Depressor reaction in the unanesthetized cat. C, stimulation time 8 sec; D, stimulation time 16 sec.
Fig. 20. A, bradycardia elicited by septal stimulationin the unanesthetized cat. B, bradycardia abolished by atropine (1 mg/kg i.v.).
N E U R O V E G E T A T I V E A N D B E H A V I O R A L F U N C T I O N S OF T H E S E P T A L A R E A
203
jected to any detailed analysis (Hess, 1958; Kabat et al., 1935; Kabat, 1936; Kurotsu et al., 1958; Ranson et al., 1935). MacLean et al., (1960) elicited penile erection (a localized circulatory effect) by electrical stimulation of the septal area in the unanesthetized squirrel monkey. The fact that in our expsriments the same reaction has been obtained in three different species of animals pleads for its physiological importance in the regulation of the neurovegetative functions studied. The most striking feature observed in the experiments described was the consistent, marked and long-lasting fall in blood pressure [this last peculiarity was absent in the rabbit] elicited during stimulation but always accentuated or clearly seen after the stimulus was withdrawn. This reaction did not parallel the changes in respiration or the bradycardia which also occurred frequently. The depressor effect could be the result of either an inhibition of vasoconstrictor fibers (directly and/or indirectly through the vasomotor center) or of a facilitation or stimulation of vasodilator fibers or both. Vagotomy and atropine did not interfere with the fall in blood pressure. Further, Manning et ~ l(1963) . have shown, in the chloralosed cat, that the dexease in contractile force and heart rate induced by septal stimulation was abolished by ruling out the cardioaccelerator sympathetic fibers (stellate ganglionectomy, blocking agents), but the fall in blood pressure was unaffected. This observation points to the peripheral origin of the depressor reaction, that is, a decrease in sympathetic tone to the blood vessels. The assumption that septal impulses could affect the discharge of the medullary vasomotor center is based on the observation that stimulation of a restricted hypothalamic zone, just behind and below the anterior commissure about 2 mm laterally to the midline, resulted in drastic inhibition of sympathetic vasoconstrictor tone. It was suggested by Folkow et al. (1959) that this area constitutes a hypothalamic relay station for cortical inhibitory pathways to lower structures, mainly affecting the discharge of the medullary vasomotor center. The possibility of septal impulses acting through this area is supported by the known connections of the septum with the hypothalamus either directly by way of the medial forebrain bundle or indirectly through the amygdala and fornix. It is also important to recall the connection between the septum and the mesencephalic reticular formation (Nauta, 1956). The interplay between septal stimulation and the baroreceptor reflex studied in the cat also favors the idea of an action of the septal area on the vasomotor center. The blockage of the baroreceptor reflex by the hypotension due to septal stimulation cannot be ascribed to the blood pressure fall, because such a reflex is extremely resistant even when the hypotension is very marked. The suggestion is that impulses originating in the septal area at least partially inhibit the bulbar vasomotor center, thus blocking the action of the reflex mechanisms of the baroreceptor nerves. This idea was already advanced (Candia et al., 1962) to explain the blood pressure fall observed during sleep. A peripheral blockage by a humoral factor can be discarded because we were unsuccessful in demonstrating the release of a humoral factor after septal stimulation (Covian et al., 1964). One argument in favor of some possible participation of vasodilator fibers is the fact that the sympathetic vasodilator tract in its intracerebral course passes through the septal region (Elliasson et d., 1952; Uvnas, 1960). References p . 215-217
204
M. R. COVIAN
Bradycardia following septal stimulation has been a feature which very often, but not always, accompanied the fall in blood pressure. It was inhibited by atropine but in some cases it persisted after the drug injection or vagotomy thus suggesting an inhibition of efferents which produce cardiac acceleration. The long persistence of the response observed in rabbit, as well as its maintenance during a prolonged stimulation time in the cat, means that the septal area is one of those structures of the brain resistant to fatigability. Delgado (1959) has shown that continuous stimulation reveals fatigability in some areas of the brain while others are very resistant. One can argue that those brain regions resistant to fatigue integrate the physiological mechanisms which are constantly at work, such as those which regulate blood pressure. The septal area could be one of the structures that form part of these mechanisms. The hypertensive reactions observed in some animals before the hypotension (or as a pure effect) suggest the facilitation of some sympathetic efferents. Blockage by dibenamine demonstrated their vasoconstrictional origin. There would be two antagonistic systems intermingled in the septal area and acting on blood pressure: one induces hypertension and the other provokes hypotension. During electrical stimulation both systems would be excited but predominantly the hypotension circuits, shifting the dynamic balance to the side of hypotension. By the end of stimulation in our experiments, the sudden withdrawal of the less effective hypertensor system would unchain the hypotensor one, thus allowing its full effect, and explaining the accentuation of the blood pressure fall after stimulus withdrawal that we have observed. An interesting feature of the hypotension is its prolonged duration, often lasting 3-5 min after the cessation of the stimulus. This seems to be a peculiarity of the effects produced by septal stimulation, for MacLean and Ploog (1960) reported that in unanesthetized squirrel monkeys throbbing erections may wax and wane for periods up to 5 min following septal stimulation. On the other hand Folkow et al. (1959) observed the slow restitution of the blood pressure after the interruption of the stimulation, requiring 3-5 min or more. Kabat and co-workers (1935) observed that the blood pressure fall did not return to its former level until some time after cessation of stimulation in the neighborhood of the anterior commissure. This peculiarity is of uncertain explanation, but we must remember that a fall in blood pressure after septal stimulation in the cat, and simultaneously the appearance of fast waves in the hippocampus, have been reported (Torii and Kawamura, 1960); also hippocampal afterdischarge following electrical stimulation of the septal area has been described (Torii, 1961). It is possible that the depressor reaction reported here is mediated through the hippocampus, and its maintenance after the stimulus withdrawal could accounted for the hippocampal after-discharge due to reverberating circuits. A decrease in the heart rate and irregularity have been reported in kittens during sleep (Jouvet et al., 196l), as well as constant and marked pressure fall in cats during the fast cortical activity phase of sleep (Candia et al., 1962). Increased sleeping time under barbiturate after destruction of the septal area in the rat has also been observed (Harvey et al., 1964; Heller et al., 1960). On the other hand lesions ofthe septal area interfere with the rhombencephalic phase of sleep in cats (Jouvet, 1962). It has been
NEUROVEGETATIVE A N D B E H A V I O R A L F U N C T I O N S OF THE S E P T A L A R E A 2 0 5
reported that destruction of the septal area produces a heightened emotional behavior in the rat (Brady and Nauta, 1953, 1955; Tracy and Harrison, 1956>,sham rage in the cat (Spiegel et al., 1940) and increased sympathetic activity in the cat (Bond et al., 1957). Electrical and chemical stimulation of the septal area in the cat resulted in enhancement of pleasure reactions (Trembly, 1956). According to these reports, and those of self-stimulation in rats (Olds and Milner, 1954) and man (Heath, 1963), the septal region could be regarded as a ‘quieting system’. However, some authors have not found behavioral changes after septal lesions (Brady and Nauta, 1953; Harrison and Lyon, 1957). If we take into account that placidity and sleep are associated with falls in blood pressure and diminished heart rates, it can be suggested that there is a relationship of our findings with these states. (11) C O N D I T I O N I N G OF THE B L O O D P R E S S U R E F A L L D U E TO S E P T A L STIMULAT I O N I N THE A N E S T H E T I Z E D R A B B I T
The regularity and magnitude of the evoked arterial hypotension obtained by septal stimulation led us to test its conditionability. Thus, the unconditioned response elicited by stimulation of the septal area in rabbits under urethane anesthesia was used as a reinforcer of a sensorial (acoustic) stimulus. Each pairing consisted of a click train of lO/sec repetition rate applied during 35 sec, and an electrical stimulation of the septal area applied during the last 15 sec of the conditioned stimulus. The aim of this uncommon approach, such as studying an adaptive reaction during attention blockage, was to elucidate to what extent a high level of vigilance is necessary to obtain a neurovegetative conditioning when a low integrating structure such as the limbic system is activated. The results showed that approximately 10-25 associations were sufficient to obtain the initial conditioned hypotension to the acoustic stimulus. Fig. 21 illustrates the characteristics of a conditioned response ; the acquired response was progressively enhanced in successive reinforcements reaching a magnitude similar to the unconditioned one. The ECG recorded on the acoustic area of the brain shows, in a few experiments, desynchronization due to the acoustic stimulus only during the conditioning. Septa1 stimulation also caused desynchronization as was often observed. It is well to make clear that in some rabbits the unconditioned stimulus provoked a hypersynchronization. Fig. 22 depicts the results obtained in another rabbit in which the ECG did not change, either during conditioning or during septal stimulation. The conditioned hypotension presented all the peculiarities of conditioned responses, namely, external and internal inhibition, generalization, differentiation (Fig. 23) and extinction (Fig. 24). Comments
The right level of anesthesia for obtaining the conditioned response was one of the hard problems we faced. It seems that there is an optimal degree in which the conditioning is elicited, this being an intermediate stage between a deep and a light anesthesia. References p. 215-217
206
M. R. COVIAN
Fig. 21. Conditioning of the blood pressure fall due to septal stimulation in the rabbit with urethane anesthesia. A, before conditioning septal stimulation elicited the blood pressure fall and ECG desynchronization. After some associations (B) the blood pressure fall was conditioned. The conditioned stimulus desynchronized the ECG and determined a fall which paralleled that obtained by the unconditioned stimulus.
V 3 X V 7 V L d 3 S 3H1 do S N O I 1 3 N n J 'IVXOIAVH3CI a N V ElAIlVL3DElAOXfl3N
208
M. R. COVIAN
Fig. 23. Conditioning in the rabbit under urethane anesthesia. Differentiation. A, the conditioning has been established to a click train of lO/sec repetition rate; B, there was no response to a stimulus of 31 sec repetition rate; C, the conditioning is again shown to a click train of lo/=.
NEUROVEGETATIVE A N D BEHAVIORAL F U N C T I O N S OF THE SEPTAL A R E A 2 0 9
Fig. 24. Conditioning in the rabbit with urethane anesthesia. Extinction. D, the conditioning being established, the blood pressure fall was elicited by the conditioned stimulus without reinforcement. E, after some applications of the conditionedstimulus alone, extinction appeared. F, after some associations the conditioned response reappeared. References p. 215-21 7
210
M. R. C O V I A N
In this regard it is well to quote Galambos and Morgan (1960): ‘... relatively small amounts of anesthetics unfailingly reduce animals and men to a state where no learning whatever is possible. One might naively expect that between the stages of complete anesthesia and none at all a level would be reached where the learning process was, say, only half impaired. Such a stage has, however, never been defined’. Emphasis has been placed on the role played by subcortical structures in the process of conditioning, namely, the thalamus, hypothalamus and mesencephalic reticular formation (Cardo, 1961). It seems that the septal area is another subcortical structure of importance in conditioning. Since Pavlov’s work it is known but not always remembered that the conditional reflex extends to the autonomic system, and that this autonomic conditioning is usually out of consciousness. Horsley Gantt (1964) has clearly pointed out: ‘Traditionally both the medical profession and psychology refuse to admit the existence of conditioning below the conscious level, in spite of the fact that Pavlov began his conditional reflex research with the gastrointestinal system three score years ago’. Our finding is another example of conditioning obtained out of consciousness which raises a number of questions such as the activity and role of the reticular formation in this conditioning; the possibility that lower levels of integration, e.g. limbic structures, can accomplish all the steps of the process and also that the rabbit because of neural peculiarities is a specially suitable animal for this unconscious conditioning. These and other questions involving neurophysiological and even psychological implications claim clarification. (111) A L T E R A T I O N S I N S O D I U M C H L O R I D E A N D W A T E R I N T A K E AFTER SEPTAL LESIONS I N T H E R A T
Considering that an organism acts as an integrated unity, and that appetite behavior requires the interaction of different parts of the central nervous system, therefore not depending on a circumscribed region of the brain, we planned a study to evaluate which parts of the CNS are engaged in the control of NaCl and water intake. Accordingly, systematic studies were undertaken in our laboratory to determine the changes on the free ingestion of 1.5 % NaCl and tap water after localized lesions in the brain of the rat. The normal variations in the intake of both fluids, and those introduced by the technical procedures preceding the local destruction of the hypothalamus, as well as the specific alterations in sodium chloride and water intake, have already been reported (Antunes-Rodrigues and Covian, 1963, 1965; Covian and Antunes-Rodrigues, 1963). Owing to the known connections of the septal area with the hypothalamus, and because of the physiological characteristics of this structure, the role it plays in the regulation of the ingestion of both fluids was investigated. This third section of our communication will deal with our preliminary results. Detailed reports of other aspects of these studies are in preparation. In general, the procedures applied have been reported elsewhere (Antunes-Rodrigues and Covian, 1963; Covian and AntunesRodrigues, 1963). The self-selection method was used, rats of both sexes and weighing
N E U R O V E G E T A T I V E A N D B E H A V I O R A L F U N C T I O N S OF T H E S E P T A L
AREA21 1
about 200 g being kept in individual cages. Each had a food cup filled with a dry mixed diet containing 0.1 12 mEquiv/g Na, and two graduated drinking bottles, filled with 1.5 % NaCl and tap water, respectively. Daily readings were made of the intakes as well as of some environment factors which could interfere with appetite behavior, such as temperature, relative humidity and atmospheric pressure of the rat house. After a control period of about 4 weeks bilateral septal lesions were made by electrolysis under ether anesthesia using the co-ordinates of de Groot’s atlas (1959). A stainless steel electrode of 0.36 mm thick and insulated to the tip was mounted in a KriegJohnson stereotaxic instrument, and a current of 1.5 mA was applied for 15 sec. A control group of sham-operated rats was also studied. After septal lesions were made. observations were maintained as previously to study any alterations in the ingestions. Routine histological procedures were used to localize the site of the lesions. In those rats that showed a change in fluid intake this consisted in an increase in NaCl and a decrease in water consumption. In 10 rats with this alteration the average intake of NaCl solution during the control period of 1 month was of 8.10 ml/day; after the septal lesions the value for the same period was of 31.60 ml/day, and for the two months of postoperative observation it was of 23.1 ml/day. For water the values observed were 20.1, 6.65 and 9.96 ml/day respectively. The total ingestion of fluids was 28.16 ml/day before the operation, and 38.2 and 33.05 ml/day for the aforementioned postoperative periods. Food ingestion did not show any significative change. Fig. 25 illustrates the differences observed in one rat: after the bilateral lesion was made at co-ordinates F, 8.0; L, 0.5; H, +0.5 there was a sharp increase in NaCl (solid line) and a drop in water (broken line) intake which remained for the 76 days of postoperative observation. The total amount of fluid consumption showed a value of 25.3 ml/day during the 30 days preceding the operation and of 35.48 ml/day during the 30 days following it; considering the whole postoperative period this last value was 31.09 ml/day. For NaCl and for the same periods the intake was of 4.1, 33 and 26.25
I5
<:
15
0
m
LO
m
80 D A I S
Fig. 25. Daily 1.5 NaCl (solid line) and water (broken line) intake of 1 rat in which a bilateral septal lesion was made (arrow) at co-ordinates F, 8.0; L, 0.5; H, +0.5. Ordinate,fluid intake in ml. Abscissa, time in days. References p . 215-217
212
M. R. C O V I A N
ml/day respectively; for water it was of 21.1, 2.48 and 4.84 mllday. The control group did not show any significant change in fluid or food ingestion. Localization of septal lesions. The histological examination of the brains showed that in those rats that developed the changes reported the lesions were large, involving either the anterior or the posterior half of the septal area, extending from the medial to the lateral part. As some structures such as the diagonal band of Broca and the anterior commissure were destroyed in some of these rats, restricted lesions to them were made in some rats but they were not followed by any permanent change in ingestion. The septal lesion of the rat whose intake is depicted in Fig. 24 is shown in Fig. 26.
Fig. 26. Photomicrographof brain frontal section of rat of Fig. 25.
Unilateral lesions. In 4 rats the septal area was destroyed only on one side. These rats showed a temporary increase in NaCl intake together with a decrease in water ingestion. Fig. 27 is a good example of the result obtained. After the lesion there was a sharp fall in water intake which lasted 13 days; during the same period a large increase in NaCl consumption developed reaching the high value of 91 ml on the 7th day. Then the ingestion of both fluids gradually became normal and reached the preoperative level at about 45 days after the operation. In the preoperative period of a month and the month following the operation the values for NaCl were 5.35 and 34.7 ml/day, respectively. For water ingestion and during the same periods the values were 27.9 and 7.94 ml/day. The total fluid consumption for the mentioned periods were 33.24 and 42.64 ml/day respectively. The histological examination of the brain of this rat (Fig. 28) showed a unilateral lesion over the whole half of the septal area in the anterior-posterior and medial-lateral planes.
NEUROVEGETATIVE A N D BEHAVIORAL F U N C T I O N S O F THE SEPTAL A R E A 2 1 3
Fig. 27. Daily 1.5 % NaCl (solid line) and water intake (broken line) of 1 rat in which a unilateral septal lesion (arrow) was made at co-ordinates F, 8.0; L, 0.5; H, $0.5. Ordinates, fluid intake iniml. Abscissa, time in days.
Fig. 28. Photomicrographof brain frontal section of rat of Fig. 27.
Comments
Evidence in the literature for the participation of the septal area in drinking behavior is scarce. It has been reported that cholinergic stimulation of this limbic area affects water intake in sated (Grossman, 1964) and non-sated rats (Fisher and Coury, 1962, 1964) as well as its bilateral destruction (Harvey and Hunt, 1965). Our results indicate that rats with septa1 lesions under self-selection conditions increase their NaCl and diminish their water intake, increasing the total fluid conReferences p. 215-217
214
M. R. C O V I A N
sumption. These changes were prolonged to the end of the experiment (between 2 and 3 months) in rats with bilateral lesions, while in those with a unilateral lesion the effects were temporary, subsiding after about 45 days. The septal area integrates the limbic forebrain structures which keeps an intimate anatomical relationship with the hypothalamus through the fornix system, amygdalohypothalamic circuits, stria terminalis and especially the medial forebrain bundle. This morphological connection suggests a physiological interaction between the hypothalamus and the ‘limbic forebrain area’. Work going on in our laboratory shows that amygdaloid lesions either increase or decrease NaCl intake according to their placement. It seems that on the basis of drinking behavior, there is an intricate circuit connecting the septal area, amygdala and hypothalamus. Such a circuit, when interrupted at different points, interferes with NaCl and water ingestion but no particular neuron assembly can be considered holding the whole function. In this circuit the hypothalamus appears as a funnel where many descending pathways meet. Although NaCl and water consumption can be clearly modified by hypothalamic lesions (Covian and Antunes-Rodrigues, 1963) it seems that this regulation does not function independently of more cephalic influences. Impulses coming down from the septal area reaching the hypothalamus directly or indirectly through the amygdala could modulate the activity of the hypothalamic zones related to NaCl and water intake. What remains to be investigated is how these nodal points of the circuit which regulate NaCl and water intake interact. SUMMARY A N D CONCLUSIONS
In the first part of this study the effects of electrical stimulation of the septal area upon blood pressure and respiration were investigated in chloralosed cats and rats, in rabbits anesthetized with urethane and in unanesthetized cats. In the three species a fall in blood pressure was obtained, which in cats and rats outlasted the stimulus for 3-5 min. Bradycardia was often a concomitant feature abolished by atropine and cervical vagotomy ; however, bradycardia sometimes remained, thus suggesting a diminution in cardiac sympathetic tone. These procedures did not interfere with the blood pressure fall which also persisted after curarization. The respiratory responses consisted in hyperventilation or apnea either in inspiration or expiration. They were sometimes elicited simultaneouslywith blood pressure changes, but either effect could appear independently of the other. In a few instances an increase in blood pressure was obtained which was blocked by dibenamine. Occasionally two stimulations within a short interval at points 1 mm apart changed a depressor reaction for a pressor or vice versa. This would imply the existence of two antagonistic systems intermingled in the septal area and acting on blood pressure, one inducing hypotension and the other provoking hypertension. During electrical stimulation both systems would be excited but predominantly the hypotensor circuit, shifting the dynamic balance to the side of hypotension. A prolonged stimulation time in the cat with persistence of the fall in blood pressure,
NEUROVEGETATIVE A N D BEHAVIORAL F U N C T I O N S OF THE SEPTAL AREA 215
as well as its permanence for 46 stimulations over a period of 6 h in the rabbit, showed that the septal area is very resistant to fatigue. The baroreceptor reflex was blocked by septal stimulation; a partial inhibition of the bulbar vasomotor center by impulses originating in the septal area might be responsible for this effect. The suggestion is made that the depressor reaction is most probably due to an inhibition of vasoconstrictor fibers, but a role for vasodilator facilitation cannot be ruled out. The second part of the paper outlines the conditioning of the decrease in blood pressure due to brain septal stimulation obtained in rabbits under urethane anesthesia. This conditioning required a critical level between a light and deep anesthesia which was difficult toiobtain. Once the conditioning was established,it showed all the characteristics of the conditioned responses, namely external and internal inhibition, generalization, differentiation and extinction. No clear correlation between conditioning and electrocortical activity could be detected. The conditioning of this neurovegetative function in the absence of vigilance opens a number of questions, such as the role played by subcortical structures in the accomplishment of the whole process. The third part of our communication deals with the alterations in sodium chloride and water intake induced by septal lesions which were studied by employing the standard 'two bottle' self-selection procedure. An increase in NaCl and a decrease in water ingestion resulted after the placement of the lesions. In rats with bilateral destruction of the septal area the changes lasted to the end of the experiments (from 2 to 3 months), while in those with unilateral lesions the changes subsided after about 45 days. Taking into account previous findings which showed the role played by the hypothalamus in the regulation of NaCl and water intake, it is suggested that impulses coming down from the septal area, reaching the hypothalamus either directly or indirectly through the amygdala, could modulate the activity of the hypothalamic zones related to NaCl and water intake. All the experiments reported in this paper strengthen the physiological importance of this small structure of the limbic system in the regulation of neurovegetative and behavioral functions. ACKNOWLEDGMENTS
This research was supported by the U.S. Air Force under Grant AF-AFOSR-311-64, monitored by the Air Force Office of Scientific Research of the Aerospace Research and by 'FundaCHo de Amparo a Pesquisa do Estado de SHo Paulo' under Grant 63/160/C. BioMgicas. REFERENCES ANTUNES-RODRIGUFS, J., AND COVIAN,M. R., (1963); Hypothalamic control of sodium chloride and water intake. Acta physiol. 1at.-amer., 13, 94-100. ANTUNES-RODRIGUES, J., AND COVIAN, M. R., (1965); Specificchanges in water intake and adipsia for water and sodium chloride after hypothalamic lesions. Acta physiol. /at.-amer., 15, 251-259.
216
M. R. C O V I A N
BOND,D. D., RANDT,C. T., BIDDER, G. T., AND ROWLAND, V., (1957); Posterior septal, fornical, and anterior thalamic lesions in the cat. Arch. Neurol. Psychiat. (Chic.), 78, 143-162. BRADY, J. V., AND NAUTA, W. J. H., (1953); Subcortical mechanisms in emotional behavior; affective changes following septal forebrain lesions in the albino rat. J. comp.physio1. Psychol., 46,339-346. BRADY, J. V., AND NAUTA, W. J. H., (1955); Subcorticalmechanisms in emotional behavior; duration of affective changes following septal and habenular lesions in the albino rat. J. comp. physiol. Psychol., 48,412-420. CANDIA, O., FAVALE, E., GIUSSANI, A., AND ROSSI,R. F., (1962); Blood pressure during natural sleep and during sleep induced by electricalstimulation of the brain stem reticular formation. Arch. ital. B i d , 100,216-283. CARDO,B., (1961); Rapports entre le niveau de vigilance et le conditionnement chez l'animal. 6tude pharmawlogique et neurologique. J. Physiol. (Paris), Suppl. 53. COWAN,M. R., AND A"ES-RODIUGUES, J., (1963); Spxific alterations in sodium chloride intake after hypothalamic lesions in the rat. Amer. J. Physiol., 205,922-926. COWANM. R., ANTUNES-RODIUGUES, J., AND O'FLAHERTY, J. J., (1964); Effects of stimulationof the septal area upon blood pressure and respiration in the cat. J. Neurophysiol., 27, 394-407. C o w , M. R., AND TIMO-IARIA, C.,(1966); Decreased blood pressure due to brain septal stimulation: parameters of stimulation, bradycardia, baroreceptor reflex. Physiol. Behav., 1,37-43. C o w , M. R, TIMO-IARIA,C., A"ES-RODIUGUES, J., AND CORRADO, A. P., (1964); Modificaciones vasculares y respiratoriaspor estimulacibn del area septal del gat0 cloralosado. Rem'menes de wmmunicaciones libres. VZ Congreso de la Asociacidn Lotino-americana de Ciencias Fisioldgicas, Viia del Mar, Chile, 30. DE GRmr, J., (1959); The rat forebrain in stereotaxiccoordinates. Trans. roy. Neth. Acad. Sci., 52, No. 4. DELGADO, J. M. R., (1959); Prolonged stimulation of brain in awake monkeys. J. Neurophysiol., 22, 458-475. DELGADO, J. M. R., (1960); Circulatory effects of cortical stimulation. Physiol. Rev., 40,146-171. ELLIASSON, S., LINDGREN, P.,AND UVNAS,B., (1952); Representation in the hypothalamus and the motor cortex in the dog of sympathetic vasodilator outflow to the skeletal muscle. Actaphysiol. scand.,27, 18-37. FISHER,A. E., AND COURY,J. N., (1962); Cholinergic tracing of a central neural circuit underlying the thirst drive. Science, 138,691-693. FISHER, A. E., AND COIJRY,J. N., (1964); Chemical tracing of neural pathways mediating the thirst drive. First International Symposium on Thirst in the Regulation of Body Water. M. J. Wayner, Editor. New York, London, Pergamon Press, pp. 515-531. FOLKOW, B., JOHANSSON, B., AND &ERG, B., (1959); A hypothalamic structure with a marked inhibitory effect on tonic sympathetic activity. Acta physiol. scand., 47,262-270. GALAMBOS, R., AND MORGAN, C. T., (1960); The neural basis of learning. Handbook OfPhysiology, Section I, Vol. III. J. Field, H. W. Magoun and V. E. Hall, Editors. Washington, Amer. Physiol. SOC., pp. 1471-1499. GANW,W. HORSLEY, (1964); Autonomic conditioning. Ann. N. Y. Acad. Sci., 117, 132-141. GORTEN, R. J., SMITH,JR., 0. A., AND RUSHMER, R. F., (1964); Vasodepressor response elicited by diencephalic stimulation in dogs. Amer. J. Physiol., 207, 915-920. GROSSMAN, S., (1964); Some neurological aspects of the central regulation of thirst. First Znternational Symposium on Thirst in the Regulation of Body Water. M. J. Wayner, Editor. New York, London, Pergamon Press, pp. 487-514. HARRISON,J. M., AND LYON, M., (1957);The role of the septal nuclei and components of the fornix in the behavior of the rat. J. comp. Neurol., 108,121-138. HARVEY, J. A., HELLER,A., MOORE, R. Y.,HUNT,H. F., AND R o n , L. J., (1964); Effect of central nervous system lesions on barbiturate sleeping time in the rat. J. Pharmacol. exp. Ther., 144,24-36. HARVEY, J. A., AND HUNT, H. F., (1965); Effect of septal lesions on thirst in the rat as indicated by water consumptionand operant responding for water reward. J. comp. physiol. Psychol., 59,49-56. HEATH, R. G., (1963);Electrical self-stimulation of the brain in man. Amer. J.Psychiat. ,120,571-577. HELLER, A., HARVEY, J. A., HUNT,H. F., AND ROTH,L.J., (1960); Effect of lesions in the septal forebrain of the rat on sleeping time under barbiturate. Science, 131,662-664. HESS,W. R., (1958); The Functional Organization of the Diencephalon. New York,Grune and Stratton. H. H., AND AJMONE-MARSAN, C., (1954); A Stereotaxic Atlas of the Diencephalon ofthe Cat. JASPER, Ottawa, National Research Council of Canada. JOUVET, M., (1962); Recherches sur les structures nerveuses et les mkanismes responsabies des
NEUROVEGETATIVE A N D BEHAVIORAL FUNCTIONS OF THE SEPTAL AREA 217
diffkrentes phases du sommeil physiologique. Arch. ital. Biol., 100 125-206. D., VALATX, J. L., AND JOUVET, M., (1961); Etude polygraphique du sommeil du chaton. C.R. SOC.Biol. (Paris), 155, 1660-1664. KABAT,H., (1936); Electrical stimulation of points in the forebrain and midbrain. The resultant alterations in respiration. J. comp. Neurol., 64, 187-211. KABAT, H., MAGOUN, H. W., AND RANSON, S. W., (1935); Electrical stimulation of points in the forebrain and midbrain. Arch. Neurol. Psychiat. (Chic.), 34,931-955. KARPLUS, J. P., AND KREIDL,A., (1909); Gehirn und Sympathicus. I. Mitteilung. Zwischenhirnbasis und Halssympathicus. Pfiigers Arch. ges. Physiol., 129, 138-144. KARPLUS, J. P., AND KREIDL,A., (1910); Gehirn und Sympathicus. 11. Mitteilung. Ein Sympathicuszentrum in Zwischenhirn. Pfiigers Arch. ges. Physiol., 135,401-416. KUROTSU, T . , SAKAI, A., MEGAWA, A., AND BAN,T., (1958); The changes in blood pressure and gastric motility induced by electrical stimulation in the preoptic and septal areas. Med. J. Osaka Univ., 9, 201-226. MACLEAN,P. D., (1949); Psychosomatic disease and the ‘visceral brain’. Recent developments bearing on the Papez theory of emotion. Psychosom. Med., 11,338-353. MACLEAN, P. D., (1958); Contrasting functions of limbic and neocortical systems of the brain and their relevance to psychological aspects of medicine. Amer. J . Med., 25, 611-626. MA CLEAN,^. D.,(1959); Thelimbicsystem with respect to two basiclife principles. Second Conference on Central Nervous System and Behavior. Transactions. Mary A. B. Brazier, Editor. New York, Macy. MACLEAN, P. D., AND Puwx;, D., (1960); Cerebral loci involved in penile erection. Fed. Proc., 19,288. MACLEAN, P. D., PLOOG,D. W., AND ROBINSON, B. W., (1960); Circulatory effects of limbic stimulation, with special reference to the male genital organ. Physiol. Rev., 40,105-112. MANNING, J. W., CHARBON,G. A., AND COTTEN, M. DE V., (1963); Inhibition of tonic cardiac sympathetic activity by stimulation of brain septal region. Amer. J. Physiol., 205, 1221-1226. NAUTA,W. J. H., (1956); An experimental study of the fornix system in the rat. J. comp. Neurol., 104, 247-271. OLDS,J., AND MILNER,P., (1954); Positive reinforcement produced by electrical stimulation of septal area and other regions of rat brain. J. comp. physiol. Psychol., 47,419427. RANSON, S. W., KABAT, H., AND MAGOUN, H. W., (1935); Autonomic responses t o electrical stimulation of hypothalamus, preoptic region and septum. Arch. Neurol. Psychiat. (Chic.), 33, 467-477. SAWYER, C. H., EVERETT,J. W., AND GREEN,J. D., (1954); The rabbit diencephalon in stereotaxic coordinates. J. comp. Neurol., 101, 801-824. SPIEGEL, E. A., MILLER,H. R., AND OPPENHEIMER, M. J., (1940); Forebrain and rage reactions. J. Neurophysiol., 3, 538-548. Tom, S., (1961); Two types of pattern of hippocampal electrical activity induced by stimulation of hypothalamus and surrounding parts of rabbit’s brain. Jap. J. PhysioL, 11, 147-157. Tom, S., AND KAWAMURA, H., (1960); Effects of amygdaloid stimulation on blood pressure and electrical activity of hippocampus. Jap. J. Physof., 10, 374-384. TRACY, W. H., AND HAWSON, J. M., (1956); Aversive behavior following lesions of the septal region of the forebrain in the rat. Amer. J. Psychiol., 69, 443447. TREMBLY, B., (1956); A study of the functions of the septal nuclei. Thesis, Yale University, School of Medicine. UVN&, B., (1960); Sympathetic vasodilator system and blood flow. Physiol. Rev., 40,69-76. VOTAW,CH.L. (1960); Study of septal stimulation and ablation in the macaque monkey. Neurology, 10,202-209. JOWET
218
Hippocampal Slow Wave Activity as a Correlate of Basic Behavioral Mechanisms in the Rat L. PI CKE NHAI N
AND
F. KLINGBERG
Department of Clinical Neurophysiology, Neurological-PsychiatricClinic, Karl-Marx University, Leipzig (German Democratic Republic)
Since the occurrence of a regular slow wave activity in the hippocampus of the cat was described and interpreted as an accompaniment of arousal by Green and Arduini in 1954 many investigations have been carried out to clear up the correlation between different behavioral states and the appearance of this characteristic hippocampal activity. But whereas many details of the eliciting trigger mechanism of the slow hippocampal rhythm (SHR) have been described in the work of Brucke et al. (1959 a,b) and Petsche et al. (1962), up to now no agreement on the relation of this rhythm to special behavioral elements in the animal has been reached. There is only agreement that such a correlation exists, but how is not clear. Grastyhn et al. (1959) and Lisshk and Grastyhn (1960) assume that the SHR is in correlation with an orienting activity of the animal. They found it regularly in the early phase of conditioning in which orienting reactions play an important role, and they observed that its occurrence decreased and iinally ceased completely if the conditioned reflex was M y established. Adey et al. (1960) deny an immediate connection between the SHR and the orienting reflex because they frequently observed head movements without the concomitant appearance of the SHR. They suppose that it is connected with a goal-directed behavior of the animal. In well-established conditioned motor reflexes in cats they did not observe the disappearance of the SHR as Grastyhn et al. did in their experiments. Finally Voronin and Kotljar (1962, 1963), who elaborated instrumental conditioned alimentary reflexes in rabbits, distinguished between a slow form of the SHR (6-8/sec), which is supposed to be connected with the orienting reflex, and a fast form (&lO/sec) being an expression of a ‘synthesis of different nervous stages elicited by the signal, the movement and the unconditioned stimulus’. AU these interpretations are based on experimental observations and are confirmed by many other authors, but it seems that a theoretical scheme is lacking to combine the different behavioral and electrophysiological observations. During the last three years we carried out combined electrophysiologicaland behavioral studies in more than 70 awake, freely moving rats with chronically implanted electrodes (Klingberg and Pickenhain, 1964; Pickenhain and Klingberg, 1965). The
HIPPOCAMPAL SLOW W A V E ACTIVITY
219
electrodes were located in different cortical areas, in the dorsal hippocampus, in the antero-ventral thalamus and occasionally in other subcortical regions. Concomitantly with the EEGs the respiration (by means of electrodes placed near the olfactory bulb) and the motor activity (with the help of a method described by Szab6 et al., 1965) were recorded. During the whole experiment the behavior of the rat was recorded by the experimenter. Different behavioral states were evoked by means of clicks or light flashes given separately or in series, by different familiar or unfamiliar stimuli and by the elaboration and succeeding extinction of avoidance, aversive or alimentary conditioned reflexes. During these experiments we obtained extensive material on the relationship between the SHR and the behavior of the rat. In rats the frequency range of the SHR is higher than in cats. The lowest frequency with which the neurones in the dorsal hippocampus of the rat can follow the triggering salvos from the medial septa1 cells is 6/sec. Under special conditions this frequency can rise to 12/sec. This frequency range is the same as reported on rabbits (Sailer and Stumpf, 1957). We will try to answer two questions. (1) In what behavior situations can we observe the slow hippocampal rhythm? and (2) What is the dynamics of the frequency pattern of the slow regular rhythm in the dorsal hippocampus compared with the concomitant behavioral acts? Not every arousal is accompanied by the SHR. Fig. 1 shows that an arousal is elicited by a click. The animal displays a strong startle reaction (recorded on line M), and the electrographic pattern in the cortical derivations and in the dorsal hippocampus changes from a highly synchronized sleeping pattern into a desynchronized waking pattern. The startle reflex is not followed by any orienting motor reaction. In the derivation from the dorsal hippocampus (DH) one sees a high-frequency low-voltage activity. We confirm the statement of Grastyhn et al. (1959) that the SHR is not a necessary accompaniment of arousal nor of the startle reflex. In accord with these authors we found that a new stimulus which bears for the animal no resemblance to former stimuli does not elicit the SHR. In Fig. 2 we give another example, in which the click (mark on line S) is followed by an arousal without eliciting a startle reflex. As in Fig. 1, the DH displays a low amplitude irregular activity, but 4 sec later a slow regular rhythm sets in. Simultaneously one observes a stronger desynchronization in the cortical derivations and in the antero-ventral thalamus. Half a second later the animal performs an orienting movement (on line M) lasting about 1.5 sec. After this time the SHR (in DH) and the orienting movements (in M) disappear simultaneously, and 1 sec later in all cortical and subcortical derivations the sleeping pattern of the electrical activity reappears. Note that the SHR (in DH) sets in before the overt orienting motor activity is observed. Fig. 3 represents an example in which the animal performs an orienting reaction elicited by a noise in the experimental room. One can see the occurrence of the slow regular rhythm in the DH correlated with the orienting motor activity (M) and an acceleration of respiration (R).During its best representation in the DH the SHR is transferred to the visual cortex, a phenomenon often observed in the rat. The References p . 227
220
L. P I C K E N H A I N A N D F. KLINGBERG
Fig. 1. Electrographic arousal in rat to a click. VC = visual cortex; DH = dorsal hippocampus; TH = antero-ventral thalamus; MC = sensori-motor cortex; AC = auditory cortex; R = respiration; M = motor activity. In M one can see a strong startle reflex. Calibrations:horizontal, 1 sec; vertical, 200 pV.
frequency and amplitude of the SHR decrease during a transitory motor silence, bur they show a new increase with a new strengthening of the orienting motor activity. In our experiments with rats we regularly observed this occurrence of the SHR during the orienting motor activity. The essential feature of all these situations consisted in the fact, that the eliciting stimulation bore information that was in some respect similar to the formerly received information. Like Grastyh el al. (1964), we saw neither an orienting reaction nor the SHR if the stimulation was completely new for the animal. In this event, after repeated applications, the stimulus acquired the property of eliciting these two reactions. Therefore, we can characterize the behavioral situation in which the SHR appears as one in which the animal compares the actual sensory information coming from the situation with previously stored information. The statement, that stimuli given the first time do not elicit the orienting motor reaction or the SHR, is also valid if strong nociceptive stimuli are applied. An example appears in Fig. 4. The single lines represent the derivation from the DH
HIPPOCAMPAL SLOW W A V E A C T I V I T Y
22 1
Fig. 2. Electrographic and behavioral arousal in rat to a click. S = signal mark; T = time, 1 sec. Calibration, 200 pV. The orienting motor reaction (M) is preceded by the appearance of the slow regular rhythm in the dorsal hippocampus @H) and concomitant stronger desynchronization in the other cortical and subcortical derivations.
during the first, second, etc. . . . application of I0 short, strong eIectrica1 stimuli, marked on the top line. During the first applications of the electrical stimulus series from the grid floor no SHR was observed, although the animal exerted chaotic efforts to escape. With the following applications the occurrence of the SHR increased gradually until it occupied not only the period of stimulation but also a great part of the interstimulus period (stimulations 10 and 16). With the 16th stimulus series the animal displayed the first successful avoidance reaction, jumping from the grid floor to a freely moving rod. What happens during this time in the brain of the animal? To the first electrical stimuli, which were completely unexpected, the animal displayed different inco-ordinated motor reactions such as jerking around in the cage, biting the bars of the grid floor, jumping against the cage walls and so on. During the interstimulus period an arresting behavior predominated. With further applications of the stimulus series an orienting behavior prevailed more and more, and the most appropriate elements of the chaotic motor acts were selected and gradually combined to a new, well-adapted behavior. Finally, immediately after the first electrical stimulus of the series, the rat jumped on to the rod and avoided further electrical shocks. Apparently in this experiment the SHR was correlated not only with the orienting motor reactions, but also with the dynamic process of forming the new, well-adapted motor behavior. Essentially in the same way the SHR develops during the elaboration of a condiReferences ~ - 2 2 7
222
L. P I C K E N H A I N A N D F. K L I N G B E R G
Fig. 3. Orienting reaction elicited by a noise in the experimental room. Concomitantly with the orienting motor reaction one can see the appearanceof the slow hippocampal rhythm in DH and VC, and the acceleration of respiration. Abbreviations and calibrations as in Figs. 1 and 2.
tioned reflex. This process was described in detail by Grastybn et al. (1959) who pointed out that the orienting motor reaction plays an important role during the early stage of conditioning. But let us take an example of a later stage when the animal displays a well-adapted conditioned avoidance behavior. Fig. 5 represents such an example. The rat, having received more than 50 combinations of a series of 10 light flashes followed by an electrical shock from the grid floor, now displays a stable avoidance reflex (jumping on to a rod). With the first light flash one can see the appearance of the SHR without any motor reaction or acceleration of respiration. The amplitude and frequency of the SHR decrease, and, as a sign of the inhibited behavior of the animal, photic after-discharges in the visual cortex appear 700 msec later. This is a regular phenomenon in the first, so-called negative, phase of a delayed conditioned reflex. About 2 sec later the SHR again increases in frequency and amplitude, the respiration accelerates, and after the 5th flash the animal displaysa slight movement directed to the rod. Immediately before the jump (marked by an arrow) the SHR has a frequency maximum; immediately after the animal hangs on the rod, both frequency and amplitude decrease. The question arises whether here too the SHR is an expression of an orienting reaction. We think not. Behaviorally, after the first flash the animal exerts no motor components of any orienting reaction, and the respiration also shows no significant acceleration. The motor reaction after the 5th flash is apparently not an accompaniment of an orienting reaction but of a goal-directed behavior, clearly motivated by the whole situation. We think it would overcharge the term ‘orienting behavior’ to use it
HIPPOCAMPAL SLOW WAVE ACTIVITY
223
Fig. 4. Derivation from dorsal hippocampus in the rat during successive (1st. . 16th) applications of sets of 10 strong electrical shocks from the grid floor. The line at the top marks single electrical shocks. During the 16th shock series the rat jumps (arrow) to a freely hanging rod, thus avoiding further shocks. Calibration: horizontal, 1 sec; vertical, 200 pV.
in this connection. We prefer to describe this situation by saying that the animal compares actual sensory information with formerly stored information in order to perform certain well-adapted actions. This comparison is valid as long as the motor behavior has not yet changed into an automatized behavior. This change of the motivated motor behavior into automatized motor acts depends on the special experimental conditions. Therefore, Holmes and Adey (1 960) and Sadowski and Longo (1962) did not observe the disappearance of the SHR in the period of stabilized conditioned reflexes in cats, whereas GrastyLn et al. (1959) saw it. Voronin and Kotljar (1963) observed the SHR in rabbits even after 800 combinations of conditioned and unconditioned stimuli, and we observed the same in rats after more than 200 combinations. This also applies when we use alimentary conditioning. In Fig. 6 we show a ‘spontaneous’ intersignalreaction of a rat that had a stabilized alimentary motor conditioned reflex. The rat was sitting quietly at its starting place and began suddenly to run to the food tray (arrow upward). This motivated motor act began with an SHR whose frequency and amplitude increase during running. After the animal arrived at the feeding place, it displayed some sniffing motions during which the SHR disappeared, References p . 227
224
I.. P I C K E N H A I N A N D F. K L I N G B E R G
Fig. 5. Electrogram from the visual cortex (VC), sensori-motorcertex (MC) and dorsalhippocampus OH), respiration (R) and motor activity (M)during a conditioned avoidancereflex in rat. On line S: marks, single light flashes; arrow, conditioned jump on to the rod. Calibrations: horizontal, 1 sec; vertid, 200 pV.
Fig. 6. Interstimulus reaction during conditioned motor alimentary behavior in rat. VC1 and VCZ, visual cortex, bipolar and unipolar; DH1 and DHa, dorsal hippocampus, bipolar and unipolar; Th. antero-ventral thalamus; R, respiration;M,motor activity; S, arrow upward, beginning of running; arrow downward, beginning of drinking. Calibrations: horizontal, 1 sec; vertical 200 pV.
H I P P O C A M P A L SLOW WAVE ACTIVITY
225
Fig. 7. Conditioned motor alimentary reaction in rat. On line S: marks of light flashes; the downward mark after the 5th light flash, unlocking of the door; arrow upward, beginning of running; arrow downward, beginning of drinking. Calibrations as for Fig. 6.
and after this it began to drink (arrow downward). During the automatized motor act of licking the glucose solution no SHR was observed. Therefore, in this example, too, we can say that the occurrence of the SHR is correlated with a motivated, nonautomatized motor behavior of the rat. The strong correlation between non-automatized motor behavior and SHR on the one hand, and automatized motor behavior and lack of SHR on the other hand are shown in Fig. 7. During the interstimulus period the animal exhibits an SHR owing to the motivated alimentary situation. But as soon as the rat begins to scratch (recorded on line M) the SHR disappears, and in spite of the beginning of the conditioned stimuli it does not return as long as the scratching is continued. Only after the 4th flash, when scratching has finished, does the SHR appear again in a very marked manner, and the animal turns its head to the door through which it must pass to reach the food. But the SHR ceases again because the door is unlocked only between the 5th and 6th flashes (at the downward mark). After this the animal runs quickly to the food tray (arrow upward) and begins to drink (arrow downward). During drinking no SHR is observed. Reviewing our experimental facts we can answer our first question in the following way. The SHR in the dorsal hippocampus appears in all behavioral situations in which the rat displays a motivated behavior. This motivated behavior may include motivated, non-automatized motor acts, or it may be an inhibited motivated behavior as in the first, inhibited period of the conditioned reaction of Fig. 5. In these latter examples one can assume that the animal directs its attention to the changing situation and prepares motivated motor acts. Inversely, all motor acts of the rat are accompanied by the occurrence of the SHR if they are motivated, and if they are not References p . 227
226
L. P I C K E N H A I N A N D P. KLINGBERG
automatized. In other terms one can say that the SHR appears in all situationsin which a comparison of the actual sensory information with formerly storedinformation takes place. This likewise happens during the so-called orienting or searching behavior as when the animal elaborates a new, well-adapted motor behavior using inborn or formerly acquired elementary behavioral acts. During the performance of a welladapted routine the SHR only appears, if the performance of the motor acts is not automatized and further on requires slightly changing, subtle variations of the adapted motor acts. This interpretation of the SHR as an expression of a dynamic comparator mechanism fits well both with former assumptions on the mechanism of the orienting behavior (Sokolov, 1963) and with the concept that the structure of the hippocampus is appropriate for the comparisonof signalsof differentmodalities (McLardy, 1959). As to the second question we have already mentioned that the frequency of the SHR in the rat changes from 6/sec to 12/sec. These frequency changes also show a clear correlation with the animal's behavior. The frequency range lies between 6 and
Fig. 8. Derivations from antero-ventralthalamus (TH), visual cortex (VC)and dorsal hippocampus (DH), respiration (R) and motor activity (M) of a rat when the experimenter suddenly pulls the rod, which the animal touches with its forepaws, out of the cage (arrow).Calibrationsas in Fig. 6.
8/sec if the animal displays no or only slight motivated movements, and it increases up to 12/sec if it shows strongly excited, motivated or goal-directed motor acts. In Fig. 8 the experimenter pulls the rod, which the rat touches with its forepaws, suddenly out of the cage. Immediately, in the dorsal hippocampus a SHR appears with maximal frequency (12fsec) and maximal amplitude, the respiration accelerates, and the animal displays very strong orienting motor reactions. But the rat quickly calms down displaying only slight searching movements, and the SHR decreases both in frequency (6-7/sec) and amplitude. Therefore, we assume that the frequency range of the SHR is correlated with the amount of central activation during motivated behavior. As Sailer and Stumpf
H I P P O C A M P A L SLOW W A V E A C T I V I T Y
227
( 1957) showed in rabbits, in which they stimulated the lateral hypothalamus, a stronger
stimulation gives a higher frequency of the SHR, ranging up to 12/sec. One can assume that this is also valid in the animal that is awake during the influence of natural stimuli. The amount of impulses coming from the lateral hypothalamus and reaching the medial septa1cells determines the frequency of the SHR during motivated behakior. REFERENCES ADEY,W. R., DUNLOP, C. W.,
AND
HENDRIX,C. E., (1960); Hippocampal slow waves. Arch. Neuroi.
(Chic.), 3, 74-90. BRUCKE,F., PETSCHE,H., PILLAT,B.,
UND DEISENHAMMER, E., (1959a); Die Beeinflussung der ‘Hippocampus-arousal-Reaktion’beim Kaninchen durch elektrische Reizung im Septum. Pfiigers
Arch. ges. Physiol., 269, 319-338.
BRUCKE,F., PETSCHE, H., PILLAT,B., UND DEISENHAMMER, E., (1959b); a e r Veranderungen des Hippocampus-Elektroencephalogrammes beim Kaninchen nach Novocaininjektion in die Septumregion. Naunyn-Schmiedeberg’s Arch. exp. Path. Pharmk., 237, 276-284. GRASTY~N, E., (1959); The hippocampus and higher nervous activity. The Central Nervous System and Behaviour, M. A. B. Brazier, Editor. Transactions of the Second Conference. Macy Found., New York, pp. 119-205. GRASTYAN, E., KARMOS, G., VERECZKEY, L., UND LOSONCZY, H V., (1964); Uber den nervalen Mechanismus des Orientierungsreflexes. Wis.2.Karl-Marx-Univ. Leipzig, Math. Naturw. Reihe, 13, 1169-1 174.
GRASTY~N, E., LISSAK,K., M A D A R ~ I., Z , AND DONHOFFER, H., (1959); HipFocampal electrical activity during the development of conditioned reflexes. Electroenceph. din. Neurophysiol., 11, 409430.
GREEN, J. D.,
17, 533-557.
AND
ARDUINI, A., (1954); Hippocampal electrical activity in arousal. J. Neurophysiol.,
HOLMES, J. E., AND ADEY,W. R., (1960); Electricalactivity oftheentorhinalcortex during conditioned behaviour. Amer. J. Physiol., 199, 741-744. KLINGBERG, F., UND PICKENHAIN, L., (1964); Veranderungen des ECG, der kortikalen hervorgerufenen Potentiale und der Startle-Reaktionen bei Ratten wahrend der Ausarbeitung eines bedingten Fluchtreflexes. Acta biol. med. germ., 12, 552-567. L I S S ~ KK., , AND GRASTY~N, E., (1960); The changes of hippocampal electrical activity during conditioning. Moscow Colloquium on Electroencephalography of Higher Nervous Activity. H. H. Jasper and G, D. Smirnov, Editors. Electroenceph. clin. Neurophysiol., Suppl. 13, 271-279. MCLARDY, T., (1959); Hippocampal formation of brain as detector-coder of temporal patterns of information. Perspect. Biol. Med., 2, 443-452. F’ETSCHE, H., STUMPF, CH., AND GOGOLAK, G., (1962); The significance of the rabbit’s septum as a relay station between the midbrain and the hippocampus. I. The control of hippocampal arousal activity by the septum cells. Electroenceph. clin. Neurophysiol., 14, 202-21 1. PICKENHAIN, L. AND KLINGBERG, F., (1965); Behavioural and electrophysiological changes during avoidance conditioning to light flashes in rat. Electroenceph. elin. Neurophysiol., 18,464476. SADOWSKI, B., AND LONGO,V. G., (1962); Electroencephalographic and behavioural correlates of an instrumental reward conditioned response in rabbits, a physiological and pharmacological study. Electroenceph. clin. Neurophysiol., 14,465-476. SAILER, S., UND STUMPF, CH., (1957) ;Beeinflussbarkeit der rhinenzephalen Tatigkeit des Kaninchens. Naunyn-Schmiedeberg’s Arch. exp. Pathol. Pharmak., 231, 63-77.
SOKOLOV, E. N., (1963); Orienting reflex as cybernetic system. Zh. vyssh. nerv. Deyat., Pavlova, 13,
8 16-830. Szm6, I., KELL~NYI, L., AND KARMOS, G., (1965); A simple device for recording movements of unrestrained animals. Actaphysiol. Acnd. Sci. hung., 26, 343-349. VORONIN, L. G., AND KOTLJAR,B. I., (1962); Bioelectrical activity of some parts of the brain during elaboration and extinction of alimentary conditioned reflex. Zh. vyssh. nerv. Deyat., Pavlova, 12, 547-554. (Russian) L. G., AND KOTUAR,B. I., (1963); Cortical electrical activity during formation and stabiliVORONIN, zation of alimentary and avoidance motor conditioned reflexes. Zh. vyssh. nerv. Deyat. Pavlova, 13, 917-927. (Russian)
228
Hippocampal States and Functional Relations with Corticosubcortical Systems in Attention and Learning W.R. ADEY Departments of Anatomy and Physiology, and Brain Research Institute, University of California, Los Angeles, Cali’ (U.S.A.)
The role of the hippocampus in mechanisms of attention and learning has slowly unfolded in recent years. Evidence has accrued that it is intrinsically concerned in both acquisition and subsequent ability to perform learned habits requiring discriminative functions of a high order, as suggested by Pribram and Mishkin (1955). At the same time, however, it has become apparent that the hippocampus does not function in these mechanismsindependently of subcorticalzones, with which it is profoundly interconnected, nor does it appear to be the repository of memory traces in the sense of a bank or store of such engrams. Through its connections with the diencephalon and more caudal brain stem, it appears to influence, and to be influenced by, activity in sensory systems;and to be responsible for the establishment in these extrahippocampal structures of the physiological ‘set’ that is requisite for storage of information therein. It is a challenging notion that these systems may transact information without the ability to effect its storage in the absence of appropriate interrelations with the hippocampal system. The classical trinity of ablation, stimulation and electrophysiologicalrecording have each played a part in the still incomplete picture of hippocampal function. During behavioral training we have noted increased regularity of pattern in both hqpocampal and mesencephalic EEG records with attainment of high performance levels in visual discriminative tasks (Adey, Walter and Hendrix, 1961; Adey and Walter, 1963), and. have seen a variety of complex patterns of conditional firing appear in rostra1midbrain and thalamic units during classical conditioning and extinction (Kamikawa e t al., 1964). In further studies of the role of the hippocampus in learning, we have examined both intimate aspects of its intrinsic electrical activity (Porter et al., 1964), and simultaneous relations with other brain structures. We have also sought evidence by the use of impedance measurements for changing tissue states in the hippocampus, as well as in the amygdala and midbrain reticular formation, in the course of behavioral training with separate alerting, orienting and discriminative stimuli. We have sought differential regional responses relating to these stimuli at different levels of training and in cue reversals with retraining. The findings have further supported the view that impedance changes so evoked relate to the process of information storage in the tissue, rather than
ROLE O F H I P P O C A M P U S I N ATTENTION A N D LEARNING
229
to nonspecific aspects of total tissue activity, as suggested by Jasper (1965). The relationship of impedance responses to normal neuronal content of the tissue, and differential effects of modifications in ionic environments will also be discussed. ( 1 ) Regional aspects of intrahippocampal electrical activity; eflects of small electrolytic lesions
Electrical activity in hippocampal tissue of such mammals as rat (Bremner, 1964), rabbit (Jung and Kornmuller, 1938) and cat (Adey, Dunlop and Hendrix, 1960; Grastyan, 1959) during attentive behavior has been characterized by regular wave trains at 4 to 7 c/s (&activity). Recent studies have emphasized, however, that even in the cat, sharp regional differences exist within the hippocampus in the distribution of @activity as a concomitant of alerted behavior, and in orienting and discriminative responses (Porter et al., 1964; RadulovaEki and Adey, 1965). Our studies of bipolar records from transverse arrays of chronically implanted electrodes in the cat's dorsal hippocampus clearly indicated that the mere aspect of an alerted state, without either gross orienting behavior, or the imposition of a discriminative performance, was uniformly accompanied by a rich gamut of hippocampal slow waves. These waves were present particularly in leads primarily located in the dendritic zone of the pyramidal cell layer from the first moments of exposure to the test situation. By contrast, deeper regions of the dentate fascia and the subiculum typically exhibited 'fast desynchronized' activity in intertrial epochs at the same time as hippocampal dendritic zones showed a spectrum of high amplitude &waves. Nevertheless, leads exhibiting much fast activity during intertrial epochs showed a typical 6 c/s &burst during discrimination, and also frequently a 4 to 5 c/s train during orientation (Fig. 1). The possibility of relating such hippocampal EEG activity to finer shades of behavioral responsiveness will be discussed further below, but it may be pointed out that much continuous 8-activity occurred in leads in hippocampal dendritic zones from the earliest exposure to the test situation, as a correlate of the alerted state. Our findings do not support the view that the waves in these regions arise secondarily to the development of an orienting reflex, following an initial period in which responses to unfamiliar stimuli are characterized by desynchronization and disappearance of slow waves (Grastyan, 1959). A possible solution to these seemingly incompatible findings may lie in evaluation of regional distribution of 0-wave trains, by taking more critical account than has been customary of exact recording sites within the hippocampus. Since there is considerable evidence that extensive hippocampal lesions interfere with learning and learned performance in animals and man (Adey, 1961'; Barbizet, 1963; Drachman and Ommaya, 1964), we have examined the effects of a lesion in one part of the hippocampus on electrical activity in other parts of the structure, and sought changes in learned performances after small bilateral hippocampal electrolytic lesions (Porter et al., 1964). To this end, an array of 6 parallel electrodes 1 mm apart was implanted in the hippocampal arch on each side in the cat. This permitted destruction of the hippocampal arch by electrolysis through one pair of electrodes, leaving the remainder undisturbed for continuing registration of electrical activity. References p . 243-245
230
W. R. A D E Y
EFFECTS O F C L - 4 0 0 ON HIPPOCAMPAL EEG
L. HlPP
DISCRIMINATIVE RESPONSES A CONTROL RECORDS - 3/30/64
RA3
3-4 4-5
5-6 R.HIPP
12-13 13-14
Fig. 1. Typical EEG tracings in control records (A) and on days after a psychotomimetic cyclohexamine (B) during T-maze discriminative responses (left) and orienting responses (right). Vertical bar line indicates moment of behavioral presentation, includingonset of 500 c/s tone. After drug, regularity of 6 c/s activity was enhanced during discrimination,and in orientingresponses, slower wave trains were also enhanced. Note leads exhibiting much fast activity during intertrial epochs (trial 24, 5-6, 12-13) which changed to 6 c/s ‘@-activityduring discrimination,and a 4 to 5 c/s train during orientation (see text). (From RadulovaEki and Adey, 1965).
Behavioral training was in a modified T-maze with approach to a concealed food reward on the basis of a visual cue (light-dark discrimination). Computed averages of
ROLE O F H I P P O C A M P U S I N A T T E N T I O N A N D L E A R N I N G
23 1
L
*
,m
;
IN
y
T
P N
I n A
?
m
'
4 P
0 1
I n K
Fig. 2. Bipolar EEG records from a series of closely spaced leads in right and left dorsal hippocampi during performance of the discriminative task on 6 different days, showing changes produced by electrolytic lesions at LDH 2 , 3 and 4 and at RDH 5 and 6 . (From Porter er al., 1964).
hippocampal electrical activity during behavioral performance were prepared from 40 daily trials. Lesions varied in size from less than 1.0 mm3 to zones extending 4 to 5 References p. 243-245
232
w. R.
ADEY
mm across the hippocampal arch and a similar distance along the structure, but even the largest lesions were essentially confmed within hippocampal structures. Passage of electrolytic current was accompanied by local seizures and often also in the contralateral hippocampus. Seizure discharge usually continued for a few minutes after cessation of current flow and was followed by reduction of EEG amplitude in all leads. Recovery of amplitude occurred after 24 h in the contralateral hippocampus, and in the ipsilateral undamaged regions (Fig. 2). High frequency components in undamaged regions recovered more slowly from the adjacent injury than the amplitude of the slow waves. Recovery of high frequency components began 3 days after the lesion and was complete after about 1 week. Computed averages of electrical activity accompanying the discriminative performance reflected the effects of the lesion. Within the lesion, there was no synchronous slow wave activity, and only slight increments in amplitude occurred thereafter. In partly damaged regions, recovery of high frequency components was also detectable a few days after the lesion. In 4 of 7 animals, computed averages from surrounding undamaged areas showed increased amplitude, apparently arising in increased synchrony of slow wave activity, and persisting for several days after the lesion (Fig. 3). This increased synchrony in the average did not appear to relate to transient loss of high frequency components noted above. Amplitude, latency and duration of synchronous 6 c/s activity in the averaged records, and the variability of these daily averages during constant behavioral test conditions, were not markedly changed by the electrolytic lesions. Bilateral small lesions did not in any way alter the correct discriminative performance, even when tested only lOmin after the lesion,and while electricalactivitywas generally reduced and some abnormal spikes occurred. If the lesion produced continuing seizure activity, performance was affected during the period of electrical abnormality. Cue reversal and retraining also were tested before and after lesions. In only one of 7 cats was there a significant increase in the number of trials to reach a 90% criterion after bilateral dorsal hippocampal lesions. In this animal, cue reversal was deliberately withheld before lesions were made. Another animal showed increased numbers of trials to ‘unlearn’ the discrimination after the lesions, and a slight increase in the number of trials to reach criterion in retraining. In summary, there was little effect of limited bilateral hippocampal lesions on retention of a learned habit, or ability to acquire a new habit. These findings are in striking contrast to the severe but temporary changes in hippocampal EEG and performance capabilityfollowing subthalamic lesions, but not lesions of similar size in overlying thalamic tissue (Adey, Walter and Lindsley, 1962). Hippocampal 0-trains of 6 c/s were absent or grossly modified during the period of impaired performance. Absence of normal @activityfollowing such diencephalic lesions emphasizes the significance of such hippocampo-diencephalic and -mesencephalic connections, both in relation to performance capability and in the genesis of the hippocampal &activity as a correlate of behavioral performance. Normal patterns of &activity thus appear to depend not only on integrity of the septum (Green and Arduini, 1954) but also on these more caudal zones of the subthalamus.
ROLE O F H I P P O C A M P U S I N A T T E N T I O N A N D L E A R N I N G
233
C O M P U T E D A V E R A G E S OF 4 0 T R I A L S O N C O N S E C U T I V E T R A I N I N G DAYS P E R F O R M A N C E S O N A L L D A Y S B E T W E E N 85 A N D 100% CORRECT 6 P. 15 BEFORE HIPPOCAMPAL LESIONS
‘963 DEC 16
AFTER 1964
rl
JAN. 27
D€C 19’
JAN 26
$
JAN 2 9
i
JAN 30
DEC 2 3 .
JAN 31
1964
FEB I
DEC 20
DEC 21
HIPPOCAMPAL LESIONS
JAN. 2 l
JAN. 3
I
JAN. 4
I
FEE ?
FEB
JAN 6 I
TON I ON
APPROACHES
FOOD
0.5SEC.
1
FEB !
TO1
0
APPROACHES
FOOD
Fig. 3. Comparison of computed averages of 40 daily trials from a bipolar channel adjacent to electrolytic lesion on consecutive days before and after production of small regions of damage in both hippocampi, showing transient increase in synchrony. For both series, behavioral performance remained between 85 and 100% correct. (From Porter efal., 1964). References p . 243-245
234
W. R. A D E Y
LATE EFFECTS OF LSD
HIPPOCAMPAL RECORDS DURING DISCRIMINATIVE PERFORMANCE L E F T H I PPOCAM PUS
R I G H T HIPPOCAMPUS WAD 37
A. CONTROL TEST DAYS AVERAGES O F 4 0 RUNS
AVERAGES O F 4 0 R U N S DATE
1125163-
I
/26/63-
6.ADMINISTRATION OF LSD 70,ug/Kg I.P, 1/29/63 I. CONTROL RECORD
I . CONTROL RECORD -20AVERAGES
- 2 0 AVERAGES
2. AVERAGE OF 20 RUNS 90 M I N AFTER LSD 2. AVERAGE OF 2 0 RUNS 90 MIN. AFTER LSD
4
C. AVERAGES ON TEST DAYS AFTER LSD ON 1/29/63 AV E
AVERAGES O F 4 0 RUNS
--1/30/63
GES OF 4 0 RUNS
-/
-1/31/63
-/ I K 3 -2/2/63-
-2/4/63
-2/5/63-
-
%f
-
P
TONE 01
-
A P P R O A C H E S FOOD 0.5 SEC.
Fig. 4. Computed averages of epochs of left and right hippocampal EEG records during approach performanceon days before LSD (A), the day of LSD dosage (B), and days after LSD (C).There was an increase in the amplitude and regularity of the averages following LSD. This maximized 3 days after the drug (C, 2/1/63), and declinedthereafter. (From Adey, Porter et al., 1965).
ROLE O F H I P P O C A M P U S I N A T T E N T I O N A N D L E A R N I N G
235
( 2 ) Effects of psychotomimetic and hallucinogenic agents on the hippocampal system
Previous studies have indicated the hippocampal system as a prime site of action for psychotomimetic cyclohexamines (Adey and Dunlop, 1960; Monroe and Heath, 1961) and hallucinogenic amides (LSD-25, pilocin and psilocybin), with decrements in performance associated with propagation of drug-induced seizure discharges into subcortical and sensory cortical systems (Adey et al., 1962). More recent studies have shown that effects of LSD (75 pglkg) in single doses, at intervals of not less than 3 weeks, persist for many days after the drug, at times when all aspects of seizure discharge have subsided, and baseline EEG appears identical with predrug records. Computed averages of rhythmic 6 c/s hippocampal and entorhinal wave trains during approach performance showed increased amplitude and regularity, maximal about 4 days after LSD, and decaying to control levels after 5 to 7 days (Fig. 4) (Adey et al., 1965). Midbrain reticular activity showed only minor changes over the same period. It was noted that single doses of these drugs were followed by a substantial disinhibition of inhibited orienting behavior in the test situation. Orienting behavior present during initial exposures in the test situation, and inhibited in the course of training, was dramatically reestablished following single doses of LSD. An attempt was therefore made to study aspects of hippocampal activity that might accompany orienting behavior, and to test effects of LSD and cyclohexamines on EEG activity during orientation. (3) Comparison of EEG patterns in orienting und discriminative behavior
The uniqueness of the orienting reflex rests on certain ‘principles’in the intimate behavior of its component reflexes, including their non-specificity with respect to both quality and intensity of the stimulus, and the selectivity of extinction of various properties of the stimulus with repeated presen ation (Sokolov, 1963; Vinagradova, 1961). A specific relationship between hippocampal -wave trains and orienting behavior has been postulated (Grastyan, 1959; Grastyan et al., 1959). However, the exquisite plasticity of hippocampal &rhythms in changingbehavioral states, including the appearance of bursts of waves in a narrow spectral range during performance of a visual discriminative task, have suggested more subtle and specific relations to discriminative functions and judgment capability (Adey, 1965; Adey, Dunlop and Hendrix, 1960; Adey, Bell and Dennis, 1962; Adey, Walter and Lindsley, 1962). RadulovaEki and Adey (1965) distinguished three basic states in hippocampal EEG activity in the cat; in alert but non-performing animals, in the course of discriminative performance, and during orienting behavior. Alert but non-performing animals exhibited a wide spectrum of 0-waves in the range 3 to 7 c/s on first introduction into the test situation, without overt aspects of orienting behavior. This activity persisted in EEG epochs between discriminative and orienting trials throughout many months of training. During T-box discriminative performance, &waves regularized at 6 c/s, as described above. Computed averages in orienting trials, given in the same number on References p . 243-245
236
W. R. A D E Y
EFFECTS O F L S D ON COMPUTED AVERAGES C A T BPI5 L S D 8 0 p g / K g 1.P A. AVERAGES DURING DISCRIMINATION -40 DAILY T R I A L S
3/3/64 B. AVERAGES DURING ORIENTATION WITHOUT DlSCR I M I N ATlON 40 DAILY TRIALS
TEST SCORE 40/40
-3/4/64-
40/40
-3/5/64-
40/40
-3/6/64-
1
40/40
-3/7/64-
40/40
-3/l0/64-
40/40
-3/ll/64-
40/40
-3/13/64
-
-40/40
-3/l4/64
-
40/40
-3/l6/64-
40/40
-3/l9/64-
40/40
-3/23/64-
39/40
-3/24/64-
40/40
-3/26/64-
NO
APPROACH
-
Fig. 5 . Effects ofa single dose of LSD on hippocampal EEG, with typical augmentationof discriminative averages from the 3rd to 10th day after drug (A). In similar averages of orienting responses (B), a more regular average appeared about the 10th day (3/14/64), and persisted for approximately 10 days, but was consitently slower than averages during discrimination (see text). (From RadulovaEki and Adey, 1965).
ROLE OF H I P P O C A M P U S I N ATTENTION A N D L E A R N I N G
237
each test day and randomly interspersed with the discriminative trials, showed slower and less regular averages at 4 to 5 c/s (Fig. 5). Single doses of LSD-25 were followed by prolonged disinhibition of inhibited orienting behavior, and by the gradual appearance of a regular EEG average during orientations 5 to 10 days after the drug, and declining after 15 to 20 days, concurrently with the decline of orienting behavior. Similar but acceleratedchanges in EEG and behavior were induced by a psychotomimetic cyclohexamine,CL-400. It is concluded that hippocampal EEG patterns sensitively reflect a repertoire of behavioral responses, rather than exhibiting non-specificity as an accompaniment of a wide repertoire of specific reflexes. The sensitivity of neuroelectric processes in the hippocampus to subtle shifts in cerebral states, and indications that hippocampal &activity during discrimination has the characteristics of a ‘pacemaker’, with fragmentary and less regular rhythms in midbrain reticular formation, subthalamus and primary sensory cortical areas, have suggestedthat deposition of a ‘memory trace’ in extrahippocampal systemsmay depend on such wave trains; and subsequent recall, on the stochastic reestablishment of similar wave patterns (Adey and Walter, 1963). In continuing studies (Elazar and Adey), specification of interrelations between the hippocampal EEG and activity in the subthalamus, midbrain reticular formation and visual cortex is being undertaken by comprehensive spectral analysis (Walter and Adey, 1965a and b). High levels of coherence occur between hippocampal and subcortical leads, as well as with cortical leads, during discrimination. These high coherences may occur at frequencies higher than the &range, at frequencies between 12 and 20 CIS.Analyses during orienting behavior have also produced widespread coherent relationships but differing in distribution from those during discrimination. Differences in coherence levels have also been detected between correct and incorrect decisions.
( 4 ) Hippocampal impedance measurements during acquisition of a learned discriminative habit Disclosure of patterns in cerebral slow waves consistently related to the performance of discriminative tasks suggested the possible importance of monitoring concurrent changes in functional state in the cerebral tissue generating these waves. We have used impedance measurements in restricted volumes of tissue in the hippocampus, septum, amygdala and reticular formation to provide a series of correlates with states of tissue excitability, and have sought changes relating to acquisition of learned behavior (Adey, Kado and Didio, 1962; Adey, Kado et al., 1963). The latter may not manifest themselves as clearly in aspects of ongoing electrophysiological activity presumably involved in transactional mechanisms, as in other more subtle measures capable of revealing changes in tissue states relating to the storage of information. In particular, there is the possibility that these storage mechanisms may not lie exclusively within the neuronal compartment, but that enveloping neuroglia may also be involved. With microvolt signals at 1000 c/s applied through chronically implanted coaxial electrodes in volumes of cerebral tissue of about 1.0 mm3, it was found that at chance References p. 243-245
238
W. R. A D E Y
A. BRAIN WAVES DURING APPROACH AVERAGES OF 40 TRIALS
B. IMPEDANCE
C & ~ ~ ~ ~ : ~ C AVERAGES E
LEVEL
RECORDS DURING APPROACH
1
OF 30 TRIALS
50 OHMS
-
80%
-
w
APPRDPlOHES FcOD
C. I M PEDANCE
DURING APPROACH
AVERAGES O F 30 TRIALS
Ton€
0
I
I
I
2
I
I
4
3
I
5
I
6
I
7
i
8
S EC
Fig. 6. Development of regularity in EEG wave trains and appearance of impedance responses in the hippocampus during acquisition of visual discriminative performance in modified T-maze. Computed averages of EEG records (A) a t chance levels of performance were essentially irregular; showed some regularity a t 80% correct; and sustained regularity at 100%. Impedance records at the same levels of training (B) initially showed only irregular perturbations; at 80% a small fall at the start of the performance followed by a rise; and at 100% a profound fall, outlasting the performance. In its full conthe response is biphasic and lasts about 5 sec. Approach to food lasts about 1.5 sec. figuration (0,
levels of performance in a modified T-box with visual cues, separate computed averages of hippocampal impedance during correct and incorrect responses showed only irregular deviations around the baseline. At intermediate performance levels, a deep transient fall of 2.0 to 6.0 % of the baseline value immediately followed presentation of the test situation, and persisted beyond completion of the task. It was followed
R O L E OF H I P P O C A M P U S I N A T T E N T I O N A N D L E A R N I N G
239
by a slow rise, exceeding 8.0 % of baseline in some cases, with slow return to the preapproach level after 6 to 8 sec. This ‘evoked’ response persisted undiminished with considerable overtraining (Fig. 6). Extinction of the learned habit abolished these responses, which reappeared with retraining. No baseline impedance shifts were seen in these hippocampal impedance records during acquisition or extinction of the discrimination. It has been suggested by Jasper (1965), that these impedance changes provide a measure of ‘the simple amount of activity’?and are not critically related to any obscure molecular change in learning. If this were the case, it would be necessary to take account of the relative magnitude of impedance shifts in situations where the relative amounts of neuronal activity can be specified withreasonablereliability. Such situations are few, but it may be noted parenthetically that the magnitude of the impedance change accompanying a learned response (approximately 15 % of baseline peak-topeak) is of the same order as that seen in spreading depression, associated with a massive depolarization of all neuronal elements in the population, and possibly involving neuroglial elements in a similar process (Weiss et al., 1964). Yet the neuronal duty cycle, and thus the proportion of time per neuron spent in a depolarized state, is substantially less in the learned response than in spreading depression. More direct evidence of the qualitative, as well as quantitative? relationship of impedance changes to the conditional response has come from a study of impedance shifts in the amygdala, hippocampus and midbrain reticular formation. Computed averages have been made of impedance records from these three structures for 5 days’ training, each involving 30 trials ina paradigm withconsecutive presentation of an alerting stimulus, then orientation to the test situation, followed by a discriminative performance. These computed averages were made at chance performance level, at criterion, immediately following cue reversal, and in the course of retraining (McIlwain, Kado and Adey, in preparation). Computer analysis covered an 8-sec epoch, with approximately 1.5 sec between presentation of the 3 successive stimuli. In all 3 structures, there was a progressive decrease in variance in baseline impedance in the course of traing to a high performance level. Immediately after cue reversal, variance throughout the epoch again increased sharply, but declined with retraining in the new paradigm. Impedance responses during a learned response were larger and more consistent in the hippocampus than in the midbrain reticular formation (Fig. 7). Orienting behavior was associated with larger impedance responses in midbrain reticular formation than in the hippocampus. No consistent impedance responses were detected in the amygdala during orienting or discriminative behavior. Reversal of behavioral cues at first exaggerated hippocampal impedance responses, which declined on subsequent training days, but reappeared on reattainment of high performance levels. ( 5 ) A consideration of the nature of cerebraI impedance responses
The foregoing account has indicated qualitative and quantitative aspects of impedance changes that are both regionally determined and critically dependent on levels of training. They certainly do not appear to relate simply to the amount of activity in the References p. 243-245
240
W. R. A D E Y
IMPEDANCE IN ALERTING. ORIENTATION AND DISCRIMINATION FOR 5 DAYS
COMPUTED MEANS WITH VARIANCE -AVERAGES DORSAL HIPPOCAMPUS - CAT KAM 2.
C A PA C I TAN CE
R ES I STA NC E
A. 100% PERFORMANCE -LIGHT CUE
0. IMMEDIATELY A F E R
CUE REVERSAL
.
20
.
15 10
10
5
5
0
0
C RETRAINING TO DARK. CUE. - .7 6 % PERFORMANCE 15 .
10
.
.
.
.
.
.
..
..
.
.
5
0 TONE ON
LIGHT
ON
APPROACH TO FOOD
.
2 SEC .
Fig. 7.Impedance measurements at lo00 c/s, made in approximately 1.O mm3 of hippocampal tissue, when the cat was fully trained (A), immediately after cue reversal (B), and in the course of retraining after cue reversal (C).Each graph shows 3 traces, the mean and upper and lower limits of variance. Each average was prepared from 5 days’ training with 30 trials daily. Vertical bars indicate successive presentations of tone, visual cues and opportunity to approach food. There was a marked increase in variance throughout analysis epoch following cue reversal. Variance decreased with retraining (B). ‘Evoked’ impedance response during discrimination, amounting t o about 10%of baseline impedance a t 100% performance, waned following cue reversal, but was reestablished with retraining (C).At 100% performance (A) variance in capacitive lead declined progressively from moment tone was presented to commencement of approach and remained narrow throughout remainder of analysis. No comparable ‘evoked‘ responses during discrimination were seen in similar analyses from midbrain reticular formation or amygdala (see text). (From Mcnwain, Kado and Adey, in preparation).
ROLE OF H I P P O C A M P U S I N ATTENTION A N D L E A R N I N G
241
particular tissue, and even if it be argued that they relate to ionic shifts within cerebral compartments (Jasper, 1965), it is obviously necessary that we seek an understanding of underlying mechanisms, which may involve macromolecular systems. We have previously discussed impedance current flow in a tricompartmental model, with neuronal, neuroglial and extracellular compartments (Adey, Kado et al., 1963). The neuroglial compartment, essentially enclosing the neurons in many areas, may be regarded as intervening between the neuron and the blood vascular system in metabolic exchanges, and as forming a micrometabolic module of neuronal and neuroglial elements (Barrnett, 1963; HydCn and Egyhazi, 1962). Preferred current pathways would lie in low resistance shunt paths in the extracellular space and in neuroglia, rather than through neurons. These impedance responses may occur in non-neuronal compartments, including neuroglia. The evidence suggests that impedance responses reflect changes in intrinsic characteristics of cerebral tissue, rather than relating directly to such factors as cerebral blood flow or blood pressure (Adey, Kado and Walter, 1965). These changes may relate more closely to tissue carbon dioxide levels than to vascular factors. Carbon dioxide metabolically produced may undergo conversion to carbonic acid in the presence of carbonic anhydrase. This reaction may occur within the neuroglial compartment. Obviously, these ionic shifts may be controlled by other more complex factors, and the notion that carbon dioxide may be the ultimate arbiter should be treated cautiously. More explicit evidence on the role of the neuronal content of cerebral tissue in these evoked impedance responses has been gleaned from studies in the cat (MacGillivray et al., 1966). Ethyl alcohol at blood levels up to 240 mg/100 ml produced lowered resistance and increased capacitance in the amygdala, hippocampus and latera KANt
RT. VISUAL CORT. AEL. 4-2-64
#
L
.
60
L
30
4
I
0
.
c
30
.
60
#
.
90 rnln
L
I/--
l
120
Fig. 8. Resistive and capacitivechanges in the degenerated (right) and normal (left) lateral geniculate body in the cat following infusion of ethyl alcohol. Blood alcohol 30 min after infusion was 230 mg/100 ml. Note the absence of response on the ablated side. Solid line, effects of alcohol; dashed line, control injection of Ringer lactate solution used as vehicle for alcohol. Infusion began at vertical line. Abbreviations:A.L.G. = coaxial electrode in anterior region of lateral geniculatenucleus; P.L.G. = coaxial electrode in posterior (cellular) region of lateral geniculate. (From MacGillivray et al., 1966) References p. 243-245
242
W. R. A D E Y
geniculate body of the cat. Following retrograde degeneration of about 80% of the neuronal population in the lateral geniculate body, alcohol no longer produced the characteristic response in the degenerated tissue (Fig. 8). Similar differential responsiveness between normal and degenerated geniculate tissue occurred with a psychotomimetic cyclohexaminedrug. The evidencethus suggeststhat a normal neuronal population is essential to the impedance responses. Since the relatively high impedance of neuronal elements makes it unlikely that they are the site of changed conductance discernible in measurements across the tissue as a whole, the question thus arises as to whether neurons may exercise a controlling influence on conductance characteristics in their environment, involving extracellular fluid and neuroglial envelopes. While it would be easy to dismiss the extracellularcompartment as a mere bucket of saline, if our notions of cerebral organization required it merely to contribute sodium or other cations, attention has recently been focused on its content of mucupolysaccaride material, which may exhibit considerable organization in the arrangement of these large molecules. They are capable of modulating and controlling rates of ionic movements through this compartment. They have been shown to be chemically disarranged in the types and location of their sugar molecules in mental disorders (Barker et al., 1962). Their presence in ground substance of brain tissue has long been recognized, but their functional significance, and that of the extracellular compartment generally, has received little attention.
(6) Hippocampal organization as a model of cerebralprocesses in attention and learning We may conclude that hippocampal participation in learning involves establishment of functional patterns with subcortical structures, and thus, with diencephalocortical paths involved in sensory systems. The degree of these interrelations appears to change subtly and swiftly from moment to moment. Assessed on the basis of simultaneous EEG records from hippocampal and extrahippocampal structures, analyzed with such techniques as are currently available for detection of linearly shared spectral components, we have glimpsed some of the plastic and complex patterns characterizing alerting, orienting and discriminative behavior (Adey and Walter, 1963; RadulovaEki and Adey, 1965; Walter and Adey, 1963; Walter, 1963; Walter and Adey, 1965a and b). Within the hippocampus, at least, patterns of &activity established in the course of training exhibit increasing regularity, and their reestablishment with each behavioral performance has a degree of probabilistic scatter suggesting their organization on a ‘best fit’ basis. Do these wave patterns, then, underlie the initial deposition of information in cerebral tissue, and is its subsequent recall dependent on the reestablishment of a pattern of waves to which the neuron had been previously exposed? The effectiveness of these subsequent wave patterns in eliciting neuronal firing might depend on their multivariate relationship to an ‘optimal’ wave pattern, capable of inducing firing of that neuron at its lowest threshold. Evidence that these wave processes recorded grossly reflect an intracellular wave phenomenon having similar frequency characteristics has been found by Fujita and Sat0 (1964) in hippocampal &trains, and in neocortical neurons by Creutzfeldt et al.,
ROLE O F H I P P O C A M P U S I N A T T E N T I O N A N D L E A R N I N G
243
(1964) and by Elul (1964). Elul has noted that, despite similar spectra in the EEG recorded grossly and the intracellular waves, their mathematical coherence is low. This has suggested that the surface EEG may arise in a population of intraneuronal generators as the normal distribution ensuing from their combined activity in a non-linearly related fashion, in accordance with the central limit theorem of statistics (Cramer, 1955). In such a scheme, it would be anticipated that the frequency characteristics of individual generators would relate strongly to the gross EEG, but that phase relations would be lost in the process of summation. The notion that &wave patterns are indicative of hippocampal inhibition is difficult to support from the concomitant unit firing noted by Fujita and Sat0 (1 964), and the evidence for EPSP activity, rather than IPSP, in the course of 0-waves. At least through the window of a microelectrode within the body of the cortical neuron, firing to produce a propagated spike is not a regular concomitant of the depolarizing phase of the intracellular waves, even where this exceeds the threshold level for firing in some cases (Adey and Elul, 1965). The intracellular wave and initiation of a propagated impulse thus appear to involve processes that may occur in parallel in individual neurons, and may bear non-linear interrelations to one another. We have thus come to recognize the critical difference between the sensing of physiological processes that relate to transmission and transaction of information, as opposed to its storage. These studies have emphasized the peculiar nature of cerebral cellular organization, epitomized in many aspects by the hippocampus; a dendritic tree substantially larger than the volume of the cell body, overlapping and Ferhaps physically contiguous with those of neighboring neurons ; a neuroglial compartment intimately involved in neural metabolic exchanges ; and an extracellular compartment disposed between neuronal and neuroglial elements and characterized by a substantial content of macromolecules. ACKNOWLEDGMENTS
These studies were supported by Grants NB-01883 and MH-03708 from the National Institutes of Health, by Contract AF 49(638-1387) from the U.S. Air Force Office of Scientific Research, by Contract NONR 233(91) from the Office of Naval Research, and by Contract NASA NsG 502 from the National Aeronautics and Space Administration. Mr. W. Simpson gave valuable assistance in impedance analyses. REFERENCES
ADEY,W. R., (1961); Brain mechanisms and the learning process. Fed. Proc., 20,617-627. ADEY,W. R., (1965); Electrophysiological patterns and cerebral impedance characteristics in orienting and discriminative behavior. Symposium on Neural Mechanism of Conditioned Reflex and Behavior. Roc. Int. Union Physiol. Sci., XXIII Int. Congr., Tokyo, 1965, 4, 324-329. ADEY,W. R., BELL,F. R., AND DENNIS,B., (1962); Effects of LSD-25, psilocybin, and psilocin on temporal lobe EEG patterns and learned behavior in the cat. Neurology, 12, 591-602. ADEY,W. R., AND DUNLOP,C. W., (1960); The action of certain cyclohexamines on hippocampal system during approach performance in the cat. J . Pharmacol. exp. Ther., 130,418426. ADEY,W. R.,DuNLoP,C.W.,ANDHENDRIX, C. E., (1960);Hippocampal slow waves; distribution and phase relations in the course of approach learning. A.M.A. Arch. NeuroI., 3, 74-90.
244
W. R. AD EY
ADEY,W. R., AND ELUL,R., (1965);Nonlinear relationship of spike and waves in cortical neurons. Physiologist, 8,98. ADEY, W. R., KADO,R. T., AND DIDIO,J., (1962); Impedance measurements in the brain tissue of chronic animals using microvolt signals. Exp. Neurol., 5,47-66. ADEY,W. R., KADO,R. T., DIDIO,J., AND SCFIINDLER, W.J., (1963);Impedancechangesincerebral tissue accompanying a learned discriminative performance in the cat. Exp. Neurol., 7, 282-293. ADEY,W. R., KADO,R. T., AND WALTER, D. O.,(1965);Impedance characteristics of cortical and subcortical structures; evaluation of regional specificity in hypercapina and hypothermia. Exp. Neurol., 11, 190-216. ADEY,W. R., PORTER, R., WALTER, D. O., AND BROWN, T. S., (1964); Prolonged effects of LSD on EEG records during discriminative performance in cat; evaluation by computer analysis. Electroenceph. clin. Neurophysiol., 18,25-35. ADEY, W. R.,AND WALTER,D. O., (1963);Application of phase detection and averaging techniques in computer analysis of EEG records in the cat. Exp. Neurol., 7,282-293. ADEY,W. R., WALTER, D. O., AND HENDICE, C.E., (1961); Computer techniques in correlation and spectral analyses of cerebral slow waves during discriminativebehavior. Exp. Neurol., 3,501-524. ADEY,W. R.,WALTER, D. O., AND LINDSLEY, D. G., (1962);Effects of subthalamiclesions on learned behavior and correlated hippocampal and subcorticalslow wave activity. A.M.A. Arch. Neurol., 6, 194-207. BARBIZET, J., (1963); Defect of memorizing of hippocampo-mamillary origin: a review. J. Neurol. Neurosurg. Psychiat., 26,127-135. BARKER, S . A., BAYYUK, S. H. I., AND STACEY, M.,(1962);Chemistry of a case of juvenile amaurotic idiocy. Nature, 1%, 6 4 6 5 . BARRNETT, R. J., (1963); Fine structural basis of enzymatic activity in neurons. Trans. Amer. Neurol. ASSOC., 88,123-126. BREMNER, F., (1964); Hippocampal activity during avoidance behavior in the rat. J. comp. physiol. Psychol., 58,1622. CRAMER, H., (1955); The Elements of Probability Theory. New York, Wiley. CREUTZFELDT, 0. D., F ~ RJ. ,M., LUX,H. D., AND NACMENTO,A., (1964); Expenmenteller Nachweis von Beziehungen zwischen EEG-wellen und Activitat corticaler Nervenzellen. Naturwissenschjten, 51, 166-167. DRACHMAN, D. A., AND OMMAYA, A. K., (1964); Memory and the hippocampal complex. A.M.A. Arch. Neurol., 10,411-425. ELUL,R.,(1964);Specific site of generation of brain waves. Physiologist, 7, 125. FUJ~A Y., ,AND SATO,T., (1964);Intracellular records from hippocampal pyramidal cells in rabbit during theta rhythm activity. J. Neurophysiol., 27, 1011-1025. GRASTYAN, E., (1959);The hippocampus and higher nervous activity. The Central NervousSystem and Behavior, Transactions of the Second Conference. M. A. B. Brazier, Editor. New York, Josiah Macy Jr. Foundation, pp. 119-193. GRASTYAN, E., LISSAK, K., MADARASZ,I., AND DONHOFFER, H., (1959); Hippocampal electrical activity during the development of conditioned reflexes. Electroenceph. clin. Neurophysiol., 11,409-430. GREEN, J. D., AND ARDUIM,A. A., (1954); Hippocampalelectricalactivityin arousal. J. Neurophysiol., 17,532-557. HYD~N, H., AND EGYHAZI, E., (1962); Nuclear RNA changes of nerve cells during a learning experiment in rats. Proc. nat. Acad. Sci. (Wash.), 48,1366-1373. JASPER, H.H.,(1965); The Analysis of Central Nervous System and Cardiovascular Data Using Computer Methods, National Aeronautics and Space Administration, Washington. L. D. Proctor and W. R. Adey, Editors, NASA Document SP-72,p. 138. JUNG, R., AND KORNMULLER, A. E., (1938); Eine Methodik der Ableitung lokalisierter Potentialschwankungen aus subcorticalen Hirngebieten. Arch. Psychiat. Nervenkr., 109, 1-30. KAMIKAwA, K.,MdLWm, J. T., AND ADEY,W.R., (1964); Response patterns of thalamic neurons during classical conditioning. Electroenceph. clin. Neurophysiol., 17,4854%. MACGILLIVRAY, B., KADO,R. T., AND ADEY,W.R., (1965);The effects of alcohol on brain tissue impedancein animals and maa. Psychosom. &d., 28,464-474. MONROE, R. R., AND HEATH,R. G.,(1961); Effectsoflysergicacid and various derivatives ondepth and cortical electrograms. J. Neuropsychiat., 3, 75-82. PORTER, R.,ADEY,W. R., AND BROWN,T. S., (1964);Effects of small hippocampal lesions on locally recorded potentials and on behavior performance in the cat. Exp. Neurol., 10,216-235. PRIERAM, K.,AND MISHKIN,M., (1955); Simultaneous and successive visual discrimination by
ROLE OF H I P P O C A M P U S I N A T T E N T I O N A N D L E A R N I N G
245
monkeys with inferotemporal lesions. J. comp. physiol. Psychol., 48, 198-202. RADULOVA~KI, M., AND ADEY,W. R., (1965); The hippocampus and the orientingreflex. Exp. Neurol., 12, 68-83. SOKOLOV, E. N., (1963); Higher nervous functions: the orienting reflex. Ann. Rev. Physiol., 25, 545580. VINAGRADOVA, 0. S., (1 961); Orientirovochnyi Reflex i ego Neyrofiziologicheskic Mechanizmi. Izd-vo APN RSFSR, Moskva. WALTER, D. O., (1963); Spectral analysis for electroencephalograms: mathematical determination of neurophysiological relationships from records of limited duration. Exp. Neurol., 8, 155-181. WALTER, D. O., AND ADEY,W. R., (1963); Spectral analysis of electroencephalograms during learning in the cat, before and after subthalamic lesions. Exp. Neurol., 7,481-503. WALTER, D. O., AND ADEY,W. R., (1965a); Analysis of brain-wave generators as multiple statistical time series. Inst. Electrical and Electronic Engineers, Trans. Biomed. Eng., 12, 8-1 3. WALTER, D. O., AND ADEY,W. R., (1965b); Linear and nonlinear analysis of intracerebral relationships by stochastic modeling. Ann. N . Y. Acad. Sci., In press. WEISS,T., KADO,R. T., AND ADEY,W. R., (1964); The reactive component of impedance during spreading depression and anoxia in relation to ohmic resistance, DC potential, EEG and surface temperature in the cortex. Physiologist, 7,283.
246
Involvement of Limbic Structures in Conditioning Motivation and Recent Memory K. LISSAK A N D E. ENDRdCZI Department of Physiology, UniversityMedical School, Pkcs (Hungary)
From recent studies of higher nervous activity we drew the conclusion that in its development conditioned reflex behavior may be regarded as a ‘two-dimensional’ system: the goal-directed motor pattern acquired in the first trials is equivalent to the somatomotor manifestation of the drive, forming one of the dimensions (‘spatial frame’), and the temporary linkage developed between the drive and the environmental signal corresponds to the second dimension (‘temporal frame’). It is a particular feature of this two-dimensional system that the temporary linkage may be broken or restored respectively by inhibitory and facilitatory influences of the environment, but the somatomotor pattern of goal-directed motor activity, if once engraved on the brain, cannot be removed (Koranyi et al,, 1964; Endroczi and LissPk, 1965; Lisshk, 1962; Endroczi, 1965). In recent decades a great number of attempts has been made to understand how the temporary linkage is formed within the neural network and how the different parts of the central nervous system are involved in the decoding, storage and integration of informations arriving through different sensory channels. The midbrain reticular core, or in a broader sense, the ‘centrencephalic’ system, has been implicated not only in integration of the EEG and behavioral arousal, sleep and wakeful states but also in the elaboration of conditioned reflex activity (Moruzzi and Magoun, 1949; Jasper, 1949; Morrell and Jasper, 1956; Gastaut, 1958; Anokhin, 1960; HernPndez-Pe6n et al., 1958). Lesions destroying the dorsal subthalamic region, the postero-lateral hypothalamus or the medial forebrain trajectories may produce an unconscious state, lack of EEG or behavioral arousal and complete abolition of conditioned reflex activity. The sites of lesions and the methods used in these investigations have led the authors to various conclusions as to the extent of the subcortical areas primarily involved in these fundamental processes (Doty et al., 1959; Hernhndez-Pe6n et al., 1958;Lisshk and Endroczi, 1964). A close functional unity of the mesencephalic reticular core with the rostra1 and basal diencephalic and limbic structures has recently been proposed by several authors in the integration of elementary learning processes as well as in motivated behavior (Nauta, 1960; Endroczi, 1965; Lisshk and Endroczi, 1964; Korhnyi et al., 1964).
LIMBIC S T R U C T U R E S I N C O N D I T I O N I N G MOTIVATION
241
The eflect of electrical stimulation of the limbic structures on the establishment of the avoidance conditioned reflex in rats
The rats, bearing chronically implanted electrodes in different parts of the subcortex, were trained to jump up on to a bench when a conditional stimulus (sound of a bell) was presented, and to avoid an electric shock, as unconditional stimulus, administered through the grid-floor. (A detailed description has been given in our previous communication: Koranyi et al., 1964.) After 20 associations of the unconditional andconditional stimuli, 10 non-reinforced conditional signals were presented during which the limbic structures were stimulated with threshold intensity checked by behavioral response (0.7-2.0 v, 0.3 msec, 60-100 c/s). In the early period of conditioning stimulation of the dorsal hippocampus, the amygdaloid complex of nuclei and the basal septa1 region led to a gradual inhibition of the acquired conditioned response. After a 10-min stimulation period, re-establishment of the conditioned reflex required as many trials with reinforcements as had been given to the naive animal. It seemed that the rats had completely ‘forgotten’ the somatomotor pattern of the conditioned response. Stimulation of the mesencephalic reticular formation resulted in facilitation of the conditioned response, and the animals jumped up to the bench even without presentation of the conditional signal. Stimulation of the ‘specific’ thalamic nuclei did not influence the recently-acquired conditioned reflex activity. Inhibition of the conditioned reflex performance after stimulation of the limbic structures was temporary when it was produced in an advanced stage of conditioning. In this experimental series 80-120 trials with random reinforcing were presented for 4 to 6 consecutive days. When on the test day the non-reinforced trials were given simultaneously with electrical stimulation of the dorsal hippocampus or amygdaloid complex of nuclei, inhibition of the conditioned response lasting for 24 h was observed. However, presentation of a few reinforcements re-established normal conditioned reflex activity. The inhibitory action of the limbic structures on conditioned reflex behavior may be regarded as situation-dependent. This assumption was suggested by the following findings : electrical stimulation of the dorsal hippocampus and the amygdaloid complex of nuclei in another training box which was similar to the original one in size, color and smell except for the rods of the grid-floor, which were thinner in diameter, did not influence the conditioned response, which had been tested 10 min after stimulation in the original training box. This observation and our previous findings showed that the rats with very precise tactile modalities could discriminate between different environments, and even virtually minute environmental changes could definitely modify conditioned reflex activity (KorBnyi et al., 1964; KorBnyi and Endroczi, 1965). Stimulation of the mesencephalic reticular formation in rats produced not only the elicitation of conditioned response without presentation of the conditional signal but also disinhibition of the conditioned response during the inhibitory state elicited by stimulation of the limbic structures. The animals received 40 associations at first, then 10 non-reinforced trials simultaneously with electrical stimulation of the dorsal References p. 252-253
248
K. L I S S ~ KAND E. E N D R O C Z I
hippocampus. Inhibition of the conditioned reflex performance due to stimulation of the hippocampus was abolished by stimulation of the mesencephalic reticular formation: the animals showed a typical conditioned response to stimulation and only a few repetitions of the midbrain stimulation resulted in the recovery of normal conditional reflex behavior. The somatomotor pattern of a conditioned reflex was interpreted by us as the somatic manifestation of the drive in a given situation. The present experiments in rats demonstrated that the somatomotor pattern might be inhibited or elicited by electrical stimulation of differentpoints of the midbrain-limbic circuit. A marked impairment of memorization of a recently-acquired somatomotor pattern, caused by stimulation of the limbic structures, may well be compared with the general belief that the limbic system is invoved in memory function on the one hand, and integration of motivated behavior, on the other. Our interpretation assumes a closer relationship between memory function and memorization of a somatomotor pattern as a somatic. manifestation of the drive. In other words, memorization of a somatomotor pattern for a task, cannot be separated from the drive, and any suppression or facilitation of the drive manifests itself through a decreased or increased repetition of goal-directed motor activity. Starting from this view, the effectof the limbic system on the memorization of goal-directed motor activity was studied in a two-choice conditioned reflex situation, and an attempt was made to achieve a separate facilitation or inhibition of somatomotor patterns by electrical stimulation of different subcortical structures in the same conditioned reflex situation in cats.
The influence of electrical stimulation of the temporal cortex, the amygdaloid complex of nuclei, the hippocampus, the mesencephalic reticular formation on alimentary conditioned reflex activity in a two-choice conditioned reflex situation in cats The alimentary conditioned reflex was established in a chamber having a floor of about 4 square meters. The cats (No. 18) were trained in a two-choicereflexsitutationin the following manner: two feeders, at a distance of 50 cm, with a 25 watt bulb above each of them, were used. A piece of meat served as reinforcement. The duration of the conditional signal was 10 sec, and it was given at 2-min intervals. Usually 20 associations were presented in a daily session. During the intertrial intervals the cats were kept in a starting cage in the center of the test-room, about 1.5 m from the feeders. Each animal was permitted to run out when the plastic door of the cage was opened. Associations were presented alternately at the left and right feeder in equal number in each daily session. All cats used in these experiments bore chronically implanted electrodes in the limbic structures as well as in the mesencephalic reticular formation in bilateral positions (Figs. 1 and 2). Before going into the details of the effect of electrical stimulation of the limbic system, we shall make a short survey of the establishment of the conditioned reflex response in the two-choice situation. In an early stage of conditioningthe cats showed no spatial discriminationon presentation of the conditional signal, and ran indiscriminately to one of the feeders without hesitation or latency. As many as 300 to 450
LIMBIC S T R U C T U R E S I N C O N D I T I O N I N G M O T I V A T I O N
249
Fig. 1. Locations of the electrodes in the temporal cortical region, the amygdaloid complex of nuclei, the dorsal hippocampus and in the mesencephalic reticular formation.
H
Y /
I
/
I
U
I
I
I
I
/
I I
n Fig. 2. Schematic drawing of the somatomotor route pattern during the electrical stimulation of the temporal cortex. Threshold stimulation of the temporal cortex unilaterally produces an inhibition of the approach to the feeder on the homolateral side, and running to the contralateral feeder (left). A few repetitions of the stimulation during the approach to the feeder on one side resulted in running to the contralateral feeder or returning to the starting cage (right). The white and/or dark part on the feeders correspond to the presentation of a light signal (white feeder).
associations were necessary for obtaning a near hundred per cent performance of the conditioned response. When the animals had acquired discrimination and ran to the correct feeder, a delay between presentation of the conditional signal and the running out, was introduced. The conditioned response was tested with delays of 5, 10, 20 and 30 sec alternately for both feeders. Most of the cats gave a correct performance of the conditioned response with a 5-sec delay, but after a delay of 10 sec References p. 252-253
250
K. LISSAK A N D E. E N D R O C Z I
or more, some of the cats became confused and ran to the wrong feeder. Only 4 of 12 cats gave consistently correct responses with a delay of 20 or 30 sec. Electrical stimulation of the dorsal hippocampus and the amygdala resulted in an inhibition of the conditioned reflex performance which lasted only for the duration of stimulation, and a few seconds later the animals gave a correct conditioned response. No difference was observed in post-stimulatory behavior whether the stimulation was made when the animal stayed in the starting cage or during the approach to the feeder. Different behavioral reactions could be observed according to whether the dorsal hippocampus or the amygdala was stimulated: hippocampal stimulation resulted in cessation of locomotor activity and a staring expression, while that of both the basolateral and the anteromedial parts of the amygdala resulted in fear, rage and escape behavior. Emotional responses occurred as stimulus intensity-dependent reactions. Stimulation of both structures at threshold intensity (checked by behavioral response), did not influence the conditioned response : elevation of the stimulus intensity resulted in both the inhibition of conditioned reflex and the behavioral reactions mentioned above. Behavioral reactions quite different from those observed during stimulation of the amygdala and the hippocampus were observed on stimulation of the temporal neocortex, pyriform cortex and the entorhinal cortical area. As long as the cats were in the starting cage, stimulation of the temporal cortex only produced a moderate orienting reaction accompaniedby contralateral turning of the head, at higher stimulus intensity ; pure emotional reactions, however, were absent. Repetition of near-threshold stimulation of the temporal cortex 3 to 4 times for 10 sec during presentation of the conditional signal produced a change in conditioned response performance;that is, on presentation of the signal on the homolateral side the cats ran to the contralateral feeder. Elevation of the stimulus intensity led to complete inhibition of the conditioned reflex activity, i.e. the animals did not run out of the starting cage when its door was opened. Depending on the stimulus intensity and the number of associationswith the conditional signal such an inhibitory state lasted for several minutes or for the whole daily session. Stimulation of the temporal cortex produced partial inhibition of the conditioned reflex activity when stimulation was performed during the approach to the feeder on one side. Thus, pairing of the stimulation of the temporal cortex with performance of the conditioned response to the feeder on one side only in several consecutive trials resulted in inhibition of the approach behavior on the stimulated side. The cats mostly ran to the contralateral feeder, and on several occassions while running out, the animals did not jump up to the bench but after a short hesitation returned to the starting cage. The selective inhibition of approach behavior to the feeder on one side lasted for the whole daily session, although the animals gave a correct response on presentation of the conditional signal in the next daily session. On the other hand, the selective inhibition of the somatomotor pattern on one side did not influence the correct response to the contralateral feeder. Attempts were made to interpret the nature of selective inhibition on the somatomotor pattern elicited by stimulation of the temporal pole in a delayed conditioned
LIMBIC STRUCTURES I N CONDITIONING MOTIVATION
=
reflex situation. Four cats were used giving nearly 100% performance of the conditioned response in two-choice test situation with a delay of 30 sec. Threshold stimulation of the temporal cortex was employed to avoid a longer inhibitory state after cessation of stimulation. The cats remained in the starting cage during the inhibitory period, although the door was open. The inhibition usually lasted for 20 to 30 sec, and the cats performed a correct conditioned response when it had ceased. This observation showed that the inhibition evoked by stimulation of the temporal pole did not impair memorization of the conditional signal; its perception was normal but the elaboration of adequate conditioned somatomotor pattern was under inhibition. This assumption seemed also to be confirmed by the observation that stimulation of the mesencephalic reticular formation abolished the inhibitory state and induced disinhibition of the conditioned reflex activity. In these experiments the cats were in the starting cage and the conditioned response to one or both feeders was inhibited by stimulation of the temporal cortex. The door of the starting cage was opened and stimulation of the midbrain reticular formation elicited an approaching response to one of the feeders. A few repetitions of stimulation of the mesencephalic reticular formation resulted in a complete disinhibition and the animals gave a correct conditioned response to the presentation of the conditional signal. DISCUSSION
A goal-directed somatomotor pattern, which may be considered a somatic manifestation of drive, will appear as the first event in the establishment of conditioned reflex activity. Selective inhibition of the somatomotor pattern cannot be attributed to impaired perception or to a deficit in memory, which had been demonstrated in ‘open door’ delayed tests, but it should be regarded as an impairment of integration of the goal-directed somatomotor pattern. One of the consistent phenomena in the observations presented here was that, during the inhibitory state induced by electrical stimulation of the temporal cortex, the animals showed no other somatomotor activity than running to the wrong feeder or returning to the starting cage. Inhibition of an activity in progress has been ascribed by several authors to stimulation of the orbito-insulo-temporal region in higher mammals (Tower, 1936; Kaada, 1951; Fangel and Kaada, 1960; Chow, 1961). This inhibitory state was not interfered with by sectioning the pyramids at a medullary level (Tower, 1936; Kaada, 1951). There is a large amount of controversial data in the literature on the effect of the limbic structures on the learning capabilities of animals. A marked impairment of the conditioned reflex activity could be observed in Chow’s experiments, but it was suggested that the inhibition has a close relationship with spike and wave afterdischarges caused by stimulation of the limbic structures. In the present experiments the electrophysiological control of stimulation of the limbic structures revealed that the appearance of the after-discharges is not necessarily related with the inhibition of conditioned reflex performance. Thus, epileptic discharges induced by stimulation of the dorsal hippocampus led only to an inhibitory period of several minutes while the stimulation of the temporal cortex with threshold intensity produced inhibition withReferences p! 252-253
252
K. LISSAK A N D E. ENDROCZI
out the appearance of the after-discharges in the subcortical and cortical records (Unpublished observations, 1964). At an advanced stage of conditioning in an avoidance test, the impairment of conditioned reflex activity was more serious in rats than had been observed in a two-choice alimentary situation in cats after stimulation of the hippocampus and the amygdaloid complex of nuclei. Differences between behavioral reactions elicited by stimulation of the limbic system may be attributed to differentiation of these structures in the course of evolution and to the differences existing in alimentary versus avoidance situations. It is difficult to see a marked difference between the changes in learning behavior observed following ablation and those after stimulation of the limbic structures. Both types of intervention seem to interfere with integration of the correct goal-directed motor activity rather than abolition of the decoding of environmentalinformation as well as its memorization. On the basis of this assumption, inhibition of the somatomotor pattern in a conditioned reflex situation elicited by stimulation of the limbic system may be ascribed to an impairment of integration of the acquired goal-directed motor pattern. This conclusion seems to be in accord with observations made by Stamm and Pribram (1961), who found that epileptoid discharges from the inferotemporal cortex retarded the acquisition of conditioned reflex activity in a visual discrimination test but did not impair memory for the task. It is a generally accepted view that the limbic system plays a fundamental role in at least two different categories of brain function: (1) in the organization of motivated behavior, and (2) in memory function. According to our view there is a close relationship between the two categories, which even form an inseparable unit. Motivated behavior always appears in the form of a goal-directed motor activity (somatomotor or visceromotor), and integration of the outgoing activity necessarily involves memorization of the acquired somatomotor pattern in a given conditioned reflex situation. SUMMARY
The results presented in this paper have led us to propose that the limbic system exerts situation-dependentaction on the integration of the conditioned somatomotor activity. Furthermore, integration of a goal-directed motor pattern is based on memory function and such activity appears as somatic manifestation of the drive. Viewing the problem from this aspect of brain physiology, we may assume that the primary role of the limbic system is connected with integration of the goal directed motor behavior, and its role may be regarded as secondary in the so-called motivated behavior or in memory functions. REFERENCES ANOKHIN, P. K., (1960); On the relation of cortex and subcortex in theconditioned reflex. An electroencephalographicstudy. Electroenceph. clin. Neurophysiol., Suppl. No. 13. CHOW,K. L., (1961); Anatomical and electrographicalanalysis of temporal neocortex in relation to visual discrimination learning in monkeys. Brain Mechanisms andlearning. CIOMS Symposium, Montevideo, 1959. Oxford, Blackwell, pp. 507-521.
LIMBIC STRUCTURES IN C O N D I T I O N I N G M O T I V A T I O N
253
DOTY,R. W., BECK,E. C., AND Koor, K. A., (1959); Effect of brain-stem lesions on conditioned responses of cats. Exp. Neurol., 1, 36C384. ENDROCZI, E., (1965); The role of the meso-diencephalic activating system in EEG and behavioral arousal, motivation and conditioned reflex processes. Symposium on the ‘Early Manifestations in the Forming of Temporary Connections, Budapest, 1963. Acta physiol. Acad. Sci. hung., 26,69-80. ENDROCZI, E., AND LISSAK,K., (1966); Behavioral reactions evoked by electrical stimulation in the medial forebrain bundle region. J. Physiol. Behavior, 1: 223-227. FANGEL, C., AND KAADA,B. R., (1960); Behavior ‘attention’ and fear induced by cortical stimulation in the cat. Electroenceph.clin. Neurophysiol., 12, 575-588. GASTAUT,H., (1958); The role of the reticular formation in establishing conditioned reactions. Reticular Formation of the Brain. H. H. Jasper, L. D. Proctor, R. S. Knighton, W. C. Noshay and R. T. Costello, Editors. Boston, Little, Brown, pp. 561-580. HERNANDEZ-PEON, R., BRUST-CARMONA, H., ECKHAUS, E., MPEZ-MENDOZA, E., AND ALCOCERC., (1958); Effects of cortical and subcortical lesions on salivary conditioned response. CUARON, Actu neurol. lat.-amer., 4, 111-120. JASPER, H. H., (1949); Diffuse projections systems: the integrative action of thalamic reticular system. Electroenceph. din. Neurophysiol., 1, 405419. KAADA, B. R., (1951); Somatomotor, autonomic and electrographic responses t o electrical stimulation of ‘rhinencephalic’and other structures in primates, dog and cat. Acta physiof. scund., Suppl. 83, 24, 1-285.
KORANYI,L., ENDROCZI,E., AND LISSAK,K., (1964); Avoiding conditioned reflex in blind rats and rats deprived vibrissae. Acta physiol. Acad. Sci. hung., 24, 193-198. LISSAK,K., (1962); The neural and humoral control of the alimentary conditioned reflex behavior. The Second Conference on Bruin andlkhavior, M. A. B. Brazier, Editor. Americain Institute of Biological Sciences, pp. 219-272. LISSAK,K., AND ENDR~CZI, E., (1964); The role of meso-diencephalic activating system in higher nervous activity, its role in habituation, learning mechanisms and conditioned reflex processes. Znt. Conf.Sechenov’s Centenary,Bruin Refixes, Symp. II. USSRSci. Acad. Press, p. 86 (Abstracts). MORRELL,F., AND JASPER, H. H., (1956); Electrographic studies of the formation of temporary connections of the brain. Electroenceph. din. Neurophysiol., 8, 201-215. MORUZZI, G., AND MAGOUN, H. W., (1949); Brain-stem reticular formation and activation ofthe EEG. Electroenceph. din. Neurophysiol., 1,455471. NAUTA,W. J. H., (1960); Limbic system and hypothalamus; Anatomical aspects. Physiol. Rev., 40, 102-104.
STAMM, J. S.,AND PRIBRAM, K. H., (1961); Effects of epileptogenk lesions of inferotemporal cortex on learning and retention in monkeys. J. comp. physiol. Psychol., 54, 614-618. TOWER, S. S., (1936); Extrapyramidal action from cat’s cerebral cortex: motor and inhibitory. Brain, 59,408-444.
254
Mechanisms for Limbic Modification of Cerebellar and Cortical Afferent Information* STEPHEN S. FOX* *, JOHN C. LIEBESKIND * *, JAMES H. O’BRIEN, HUGH DINGLE
AND
R. D.
Department of Psychology, and Mental Health Research Institute, University of Michigan, A m Arbor, Mich. (U.S.A.)
Previous work has shown inhibitory and facilitatory effects of hippocampal origin on evoked potentials in the cortex (Fox and O’Brien, 1962; Redding and Siegfried, 1962; Cazard and Buser, 1963) and in dorsal tegmentum (Adey, 1958; Adey et al., 1957). This similarity between effectsof hippocampal stimulation and caudate or brain stem stimulation on the cortex raises the question of the generality of limbic and particularly hippocampal influenceon sensory afferents to other structures. Since the hippocampus and limbic system in general, similar to caudate and brain stem, receive afferents from a number of sense modalities (Gerard et al., 1936; Green and Machne, 1955) it may be included among those polysensory structures suggested as serving a role in sensory integration (Fessard, 1961). As an extension of earlier work showing hippocampal modification of visual cortical afferents (Fox and O’Brien, 1962) in these studieswe investigated effects of hippocampal and amygdala stimulation on sensory responses of the cortex and the cerebellum. In particular, responses of the folium-tuber complex of the vermis were chosen for study in view of earlier reports showing that the cerebellar cortex provides an electrode site from which evoked potentials can be led in response to stimulation of a number of different sensory modes (Gerard et al., 1936; Snider and Stowell, 1944; and others). While there is no known direct anatomical connection between the hippocampus and the cerebellum, several functional studies have suggested that such a pathway does exist. Morin and Green (1 953) observed long-lasting cerebellar after-discharges following repetitive stimulation of the fimbria and hippocampal commissure in guinea-pigs and cats. Conversely, Iwata and Snider (1959) were able to show that repetitive stimulation of tuber vermis in curarized cats effected a change in hippocampal activity from low voltage fast waves to slow waves of moderate voltage. Hippocampal low voltage fast activity induced by stimulation of the contralateral hippocampus
*
Supported by NSF Grants G16033, G21446, and GB1711.
** Present address: Department of Psychology, University of Iowa, Iowa City, Iowa (U.S.A.). *** Present address: Department of Psychology, University of California, Los Angeles, Calif.
(U.S.A.). 5 Present address: Department of Zoology,University of Iowa, Iowa City, Iowa (U.S.A.).
H I P P O C A M P A L EFFECTS O N C E R E B E L L A R A N D C O R T I C A L A C T I V I T Y
255
could also be changed to slow activity of moderate voltage by stimulation of the vermis. Further, it was shown that seizure activity in the hippocampus and suprasylvian gyrus resulting from hippocampal stimulation could be promptly halted by vermal stimulation and that instead of the typical post-ictal depression, low voltage fast activity followed termination of the seizure. In confirmation of Morin and Green (1953), these authors also observed the spread of hippocampal after-discharges to the cerebellum but were unable to demonstrate any other hippocampal-cerebellar changes. More recent reports, however, have indicated that stimulation of a more direct hippocampal-cerebellar path modifies both cerebellar visual and auditory afferents (Fox and Liebeskind, 1963). The principal aim of these studies was to attempt to elucidate and describe limbic, especially hippocampal effects on cerebellar and cortical activity. A second goal was to investigate in the cerebellum, properties of complex sensory interaction within and between sensory modes and to compare these results with those obtained on the interaction of hippocampal and unimodal sensory stimulation. A final goal was to attempt to derive some generality of mechanisms of sensory interaction and control. METHOD
Thirty-seven cats were used. The animals were anesthetized with ether for the surgical procedures; an intravenous catheter and a tracheal cannula were inserted and the animals were placed in the stereotaxic instrument. A bone flap approximately 2 cm in diameter was made over the anterior cerebellum, or a hole of 5 mm was drilled and, following microdissection of the dura, an electrode was placed visually on the exposed cortex of the folium-tuber of the vermis and lowered into the cerebellum cortex. Electrodes were stainless steel wire, insulated except for approximately 0.5 mm at the tips. A hemostat on the reflected temporal muscle served as the indifferent electrode. Because of the extreme sensitivity of the cerebellar cortex to anesthesia, upon completion of the operative procedures, ether was withdrawn and the animals were maintained for the remainder of the experiment under local anesthesiaand D-tubocurarine chloride. Respiration and body temperature (37") were artificially maintained. Exposed tissue margins and pressure points were treated periodically with 2 % procaine. The pupils were kept maximally dilated with atropine sulfate applied topically. Recording was bipolar or monopolar and did not begin until 3 to 4 h after the withdrawal of ether. Cerebellar recording was from a depth of 500 to 1500 ,u in the granular layer or the underlying white matter. An agar seal over the area of exposed brain served to reduce respiratory and cardiac artifacts. Simultaneous recording was from either association visual cortex (A3.0-4.0) or superior colliculus. Electrodes to these and to the areas to be stimulated were introduced through 2 mm drill holes in the calvarium. Recorded signals were fed through Tektronix model 122 preamplifiers, displayed on a Tektronix model 502 oscilloscope and monitored by a 'slave' CRO for photography. Three types of stimuli were used. Single (sometimes double or triple) electrical pulses (0.1-0.6 msec duration; 4-15 V) were delivered to the hippocampus from a References p. 280
256
S. S . F O X
et al.
modified Grass S4 stimulator through an isolation transformer. Auditory stimuIi consisted of clicks of 82 db sound pressure in reference to 0.002 dynesJcm2as measured with a General Radio Sound Level Meter (Model 1551-B). The visual stimuli, 10-30 psec flashes from a photostimulator, could be varied in intensity from bright (peak intensity of 10 million lux) to dim (peak intensity of 14 million lux), and while usually delivered in total darkness, were occasionally presented in conditions of variable ambient illumination (produced by rheostatic control of a 100 W bulb, 6 ft in front of the animal). Experiments lasted between 6 and 20 h. At the termination of an experiment electrode sites were verified histologically (frozen sections of 80 p thick stained with chromalum gallocyanine). RESULTS
Characteristics of the cerebellar response to light Negative-positive responses to light stimulation were recorded from the depths of the vermal cortex (see also Fadiga et al., 1957). The initial negative component had an amplitude of 50 and 400 pV, a latency of 15-40 msec to its onset, and a duration of 15-20 msec. The positive component was generally of smaller amplitude (50-200 p V ) and longer duration (30-60 and up to 100 msec). Fig. 1A shows characteristic cerebellar responses to light and with simultaneously recorded responses from the visual cortex and the superior colliculus. Differences in wave-form and latency among cerebellar responses and between these responses and those recorded from the visual
Fig. 1. (A) Responses to light from cerebellar vermis are compared with response to the same flash from the visual cortex and the superior colliculus. Note variability among responses in the same and different structures. (B) Responses of the cerebellar vermis to electrical stimulation of the superior colliculus (0.05 msec, 3 V) (left and right). Note similarity to cerebellar response to light. Temporal summation is seen with two threshold pulses (B, right). Negative is downward in this and all succeeding figures unless otherwise indicated. Photocell artifact indicates light flash.
H I P P O C A M P A L EFFECTS O N C E R E B E L L A R A N D C O R T I C A L A C T I V I T Y
257
cortex and the superior colliculus were also observed. Cerebellar response latencies could be grossly divided into three groups: 28-32 msec (most common), 15-20 msec and 35-40 msec (least common). The similarity between the cerebellar response to light and to stimulation of the superior colliculus (Fig. 1B) would appear to lend support to the suggestion (Snider and Stowell, 1944) that photic impulses might reach the cerebellum via the superior colliculi (although similarity of wave-form has been found for most types of cerebellar evoked potentials; Bremer, 1958; Bremer and Bonnet, 1951). Fig. lB, however, presents, additionally, evidence for a synaptic relationship between the superior colliculus and the cerebellar vermis. It also may be observed that both components of the response are facilitated by the presentation of a pulse pair rather than selectively as observed with other forms of stimulation. This modification of both the early and later components will be seen below to be characteristic for the light response in the cerebellum. The efjlects of ambient illumination and$ash intensity Wave-form and amplitude changes in the cerebellar response to light were induced by varying the degree of ambient illumination in which the light flash was presented. With photic stimulation under conditions of dim ambient illumination there was an augmentation of a secondary wave while during bright ambient conditions, the secondary wave was eliminated and the primary was reduced by about 25 % (Fig. 2A). Such
Fig. 2. (A) Effects of ambient illumination on cerebellar evoked response to light flash. Left, flash delivered in dark; center, flash deliveredin bright ambientillumination;right,indim ambient illumination. (B) Latency and wave-form changes of cerebellar evoked responses to dim flash (left), bright flash (center) and superimposed sequential traces to compare responses to bright and dim flash.
a secondary negative component would appear in dim ambient illumination even when absent in the control response. This response doubling occurred during both dim and bright ambient light; but with bright ambient, such double responses were always of considerably attenuated amplitude. Amplitude, wave-form and latency changes also resulted from varying the intensity of the light flash (Fig. 2B). Cerebellar responses to dim light were of lower voltage and longer latency and were of more complex wave-form. Note the relatively small latency shift of the visual cortex response (5 msec) compared with the much larger shift (up to 20 msec) seen in the cerebellum. If this effect were peripheral in origin, one would expect the visual cortex to reflect the shift as clearly as the cerebellum. Since it does not, References p . 280
258
s. s. F O X et al.
one may conclude that the effect is centrally mediated. Another alternative, however, remains which, if substantiated, would support an interpretation of peripheral mediation of the brightness latency shift. Close inspection of Fig. 2B shows that while the original dim flash cerebellar response shifts in latency, the shift of this original response is quite comparable to the shift in visual cortex. The difference in response latency to the bright flash actually results from the early appearance of a new negative component. It is quite possible, therefore, that an additional population of retinal elements becomes active in the bright flash condition and that this new activity is relayed more directly to the cerebellar verrnis, as indicated by the shorter latency. The cerebellar auditory response The cerebellar response to click stimulation occurred with a shorter latency but relatively similar form as the response to light (Figs. 3A and 3B). The distribution of
Fig. 3. Cerebellar responses to sensory stimuli. (A) Cerebellar vermis shows monophasic response to light flash (left), to click (center), and to both stimuli (responses superimposed for comparison) (right). Note latency difference of responses to click and light. (B) Left, more typical biphasic response to click; center, block of responses at left by prior presentation of light flash.(Tracesaresuperimposed; first click alone is presented, then click and light. No response to click is seen in the trace containing the response to light); right, complete block of click response (shown at left) by prior light flash, although the flash elicits no response. See Fig. 4.
the cerebellar vermal click responses was similar to that which was reported for maximal click responses (Snider and Stowell, 1944) and maximal responses to stimulation of auditory cortex I and I1 (Jansen and Fangel, 1961). The initial, negative component of the click response had a latency of 16-25 msec, an amplitude of 50 to 400 pV and a duration of 10 to 20 rnsec. The later, positive component (Fig. 3B) occurred at 30 msec with average amplitude equal to the early component, but as with the light response, the amplitude of this positive wave was considerably more variable. Intramodality sensory interaction Intramodal interaction was regularly observed with the presentation of paired visual or auditory stimuli. Fig. 4 shows some examples of the effect of paired click stimuli at increasing interstimulus intervals between the stimuli as well as the mean recovery cycle (N = 20). The amplitude of the test response was reduced by a preceding click and recovered progressively with increasing separation of the clicks. At a 30 msec
H I P P O C A M P A L EFFECTS O N CEREBELLAR A N D C O R T I C A L A C T I V I T Y
259
CEREBELLAR- RESPONSE RECOVERY LIGHT-CLICK LIGHT- LIGHT ............ CLICK- CLICK
---
120 I
a
d
-
20
40 60 80 100 120 140 STIMULUS SEPARATION (msec)
I60
Fig. 4. Intra- and crossmodality interactions (recovery cycles). Intramodality interactions of cerebellar response to light (top, right) and to click (top, center) as a function of interstimulus interval. Crossmodality interaction between light and click (top, left). Note the differences in recovery times. Major effect in all cases is selectively on late positive component. Graph represents mean intra- and crossmodal effects. N = 12 for light-light, N = 8 for click-click, N = 9 for light4ick curves. Note facilitation for short intervals for crossmodality but not for intramodality interactions and similarity of plateaus for intramodal recoveries.
condition-test separation the test response was completely blocked and by 100 msec was completely recovered, with a plateau in recovery at about 55 % recovery at 65-80 msec. A very similar pattern of intramodal interaction can be seen for paired light stimuli References p. 280
260
S. S. FOX
et al.
(Fig. 4). The test-flash response was completely blocked as with a click at about 30 msec condition-test interval, but increased in amplitude more slowly with increasing separation of the flashes than did the click response and recovery was not complete until 140 msec separation, and a plateau between 80 and 105 msec at 55 % recovery. It may be noted for both modalities in the examples shown in Fig. 4 that as the separation between stimuli decreased, the later, positive component of the test response was affected first. Only at 80 msec (click) and 60 msec (light) after the late positive deflection was considerably diminished, was the initial, negative component appreciably reduced. fntermodality sensory interaction Interactions between visual and auditory stimuli were also observed (Figs. 3B and 4). Such interactions between sensory modalities are particularly relevant to the extent that they reveal the common elements in the two sensory pathways. Click responses were totally blocked by properly timed prior light flashes regardless of whether or not these had elicited a cerebellar response (Fig. 3B). Increasing intervals between the light flash and the click produced recovery cycles similar to those for double click and double flash (Fig. 4). Average intermodality recovery was complete by 94 msec separation, but at separations between 40 and 50 msec, where intramodal occlusion was still complete, partial recovery could already occur for the two different stimuli; at 28 msec separation, the first portion of the negative component was evident. As with the intramodal separation series, it was the late positive deflection which most suffered attenuation. The mean intermodality recovery cycle was faster than that for paired clicks or paired flashes and no plateau in recovery, similar to that seen with paired clicks or flashes, was seen. Facilitatory interaction (greater than control amplitude of the test response)
Fig. 5. Facilitatory interaction of light- and click-evoked responses of cerebellar vermis. Click response amplitude is enhanced by paired presentation with light. Necessary stimulus interval is 20 msec or less. Note greater arithmeticincrease of late but not early components. In this figure only, negative corresponds to upward deflection. See text and Fig. 4.
H I P P O C A M P A L EFFECTS O N CEREBELLAR A N D C O R T I C A L A C T I V I T Y
261
occurring at briefer separations (10-20 msec) between response to two stimuli of similar or dissimilar modes was also observed quite frequently, and emphasizes the dynamic quality of such an interaction (Fig. 4, graph; Fig. 5) (See also Albe-Fessard and Szabo, 1954). In Fig. 5 (lower left-hand frame) at a separation of 20 msec, the light and click responses have summated. Superimposed (lower right) upon a control light response is a combined light and click response of greater magnitude than that which would result from simple wave-form addition,
3
24 22
i
a
b 20
Y
g
18
$ 16 UJ
w 14
a
0
1
2
3
4
5
AMBIENT WHITE NOISE LEVEL
(60-82db)
Fig. 6. Crossmodality interaction. Progessive blocking of vermal evoked response to flash (cerebellar light response) with increasing ambient white noise (A-D). Left and right columns from two different animals. Upper traces = Cerebellar evoked response. Lower traces = output of microphone at animal's ear. In graph at right, mean evoked response amplitudes are given for each of 8 cats, tested 8-12 times each at every stimulus level. Light flash artifact is sharp vertical photocell deflection, and discontinuity in auditory noise base line. Ambient noise is continuous throughout each trace. Blocking effects of ambient noise on response to light were observed to continue for periods up to 10 min with no recovery aod trials spaced 2 sec apart.
A further illustration of intermodal interaction was the attenuation of the visual response observed during the simultaneous presentation of white noise from an air jet (Fig. 6). The degree of attenuation was proportional to the background white noise level which was controlled over 4-3 intensity steps by opening and closing the air valve. (The sound level ranged from 64 db to 82 db sound pressure as measured by the General Radio Level Meter and was essentially flat throughout the intensity range used, f 2 db, showing peaks at 7000-8000 and 11,000c/s.)Attenuation of the click responses by varying the level of ambient illumination was not observed in these studies. Cerebellar responses to hippocampal stimulation Stimulation of either dorsal or ventral portions of the hippocampus produced negative-positive cerebellar responses similar in amplitude, duration and wave-form to cerebellar visual and auditory responses. The initial negative component occurred with a latency of approximately 10-16 msec. Single subthreshold pulses which proReferences p . 280
262
s. s. FOX et al.
Fig. 7. Cerebellar response evoked by hippocampal stimulation (0.1 msec, 6 V). (A) Response to paired pulses. (B) Sequeotial traces superimposed showing increasing amplitude of early and late components with increasing stimulating voltage. (C)Sequential traces superimposed showing latency shift from alternate stimulation with 2 and 3 pulses. (D and E) Latency and amplitude changes in cerebellarevoked response from stimulation with 1,2 and 3 pulses to the hippocampus. (F) Amplitude and latency changes in cerebellarresponsefrom increasingvoltage of paired pulses to the hippocampus. Note similarityto light and click evoked responses.
duced no evoked potential, could be made to elicit responses of increased amplitude and decreased latency when given in pairs or three. Fig. 7A shows a characteristic cerebellar response evoked by a double pulse to the hippocampus as well as modifications in amplitude, duration and latency produced by 1, 2, and 3 pulses (Fig. 7B-F). Latency differences between responses evoked by two pulses and those evoked by three pulses were often quite striking (Fig. 7C). Clear response amplitude differences to 1, 2, and 3 hippocampal pulses were also consistently produced (Fig. 7D). Amplitude and latency changes resulting from the summation of two or three pulses suggest a synaptic, probably multisynaptic hippocampal-cerebellar path. Supportive evidence for this derives from observations that two widely separated single subthreshold pulses to hippocampus failed to elicit a cerebellar response but as the pulse separation was reduced to a point where summation of the impulses could occur (32 msec average), a small negative response became evident at 15 msec latency (Fig. 8).
H I P P O C A M P A L EFFECTS O N CEREBELLAR A N D C O R T I C A L A C T I V I T Y
HIPPOCAMPAL
60 msec
263
$!
220
32 msec
-
28 msec
Y
24 msec
I80 140
W
n
5
3
a
20 msec
3
100
60
10 msec
20 10 and
20
20 msec
40
60
SEPARATION msec)
Fig. 8. Effects of pulse separation of hippocampal stimulation on amplitude, latency and wave-form of cerebellar evoked responses. Sharp positive deflection and discontinuity indicate subthreshold pulses to the hippocampus. Traces at bottom (10 and 20 msec) have been photographically, not oscillographically superimposed. Note progressive development of cerebellar response components with decreasing pulse separation, temporal summation.
d
HIPPO 7/sec
d
HIPPO 9/sec
>
k P
-10
200 msec
Fig. 9. Augmenting responses of cerebellar vermis in response t o repetitive 7/sec (top trace) and 9/sec hippocampal single pulse stimulation (0.3 msec, 3 V). Sequential oscilloscope traces have been joined artificially by horizontal lines. Note somewhat earlier fatigue with 9/sec hippocampal stimulation. Sequential superimposed traces at the end of each record provide further comparison of the two frequencies.
With successively briefer separations, the latency of this negative response decreased to 10 msec and it increased in amplitude. At 28 msec separation a second, negative component appeared at 30 msec latency. With progressively smaller separations this second component increased in amplitude References p. 280
264
s. s. FOX et al.
and decreased in latency (20 msec latency at 10 msec pulse separation). The relationship between amplitude and interpulse separation is shown graphically, also in Fig. 8. At the point at which the second negative component appeared, a late positive component also started developing progressively in its amplitude, reaching finally about 5 times its original value. Some small shift in latency for this late response also was observed throughout such series (ca. 3-5 msec).
Fig. 10. Hippocampal modification of cerebellar and cortical sensory responses. Modification of cerebellar click response (upper right); A, B, and C show sequential superimposed traces on three different occasions. Hippocampal paired pulse stimulation in all traces with progressively greater separation of click and stimulation. Note selective blocking effect on late click response components only and recovery with greater separation. D shows block of late cerebellar click component and completerecovery at greater separation. Note the similarityand consistency of wave-form and amplitude of the cerebellar response to hippocampal stimulation and to click. E, similar blocking and recovery as in D on another occasion. F, increasing voltage of hippocampal stimulation from subthreshold, blocks hippocampal click response. Note that with increased amplitude of late component of response to hippocampal stimulation, late click response component is reduced and is recovered when late component of direct response is smaller with lower voltage stimulation. Modificationof cerebellar light response (lower right): effects of progressive separation of light and hippocampal stimulation (recovery cycle). Upper traces: visual cortex. Lower traces: cerebellar vermis. Note increase in visual cortical evoked response at time of decrease in cerebellar response. Also note appearance of a second late response to light in the cerebellum corresponding to a slow wave in the visual cortex. Stimulation artifact appears as discontinuity in visual cortex traces; light artifact is vertical deflection with discontinuity on cerebellar traces. Modification of visual cortical light response (upper left): opposing effects of hippocampal stimulation on the visual cortex (upper traces: vis. cort.) and the cerebellum (lower traces: cereb.) Top: control observations with light presented alone. Bottom: light preceded by single pulse to the hippocampus. Note reduction of cerebellar light response with simultaneousincrease in the visual cortical response to the same flash. Graph presents mean recovery cycles for cerebellar response to click (N = 120 (6 cats)) and light (N = 100 (locats)) as afunction of separationfrom hippocampal stimulation. Notefacilitatory effects on cortical light response (N= 120 (6 cats)) prolonged recovery of cerebellar light response and that short separations produced facilitationfor cerebellar light and click responses.
HIPPOCAMPAL EFFECTS ON CEREBELLAR A N D CORTICAL ACTIVITY
265
Augmentation and fatigue of cerebellar responses to repetitive hippocampal stimulation at higher frequency (7 and 9 per sec) was also observed, also supportive of a synaptic relationship between these structures (Fig. 9). Somewhat more pronounced augmentation and more rapid fatigue of this activity was seen at the higher stimulating frequency. Repetitive stimulation with double pulses resulted in the cerebellum in no observable augmentation at either 9 or 12 sec. Fatigue, however, did occur, especially at the higher frequency, but was much less evident in simultaneous recordings from dorsal hippocampus, suggesting perhaps the involvement of fewer synapses between the intrahippocampal placements.
H@pOCQmpQlmodifkation of cerebellar and cortical auditory and visual evoked responses The occlusive effects of hippocampal stimulation on cerebellar responses to clicks were, in many ways, comparable to the intra- and intermodal interactions described above. The extent of occlusion of the click responses occurring at 4 different separations from a conditioning pair of pulses delivered to the hippocampus is shown in Fig. 10 (top right, A, B, C). Mean maximum blocking occurred at about 30 msec and complete recovery occurred at approximately 100 msec separation. Mainly, the second positive portion of the response was blocked. Fig. 10 (D and E) also illustrates this effect, showing the responses at approximately 50 and 80 msec separations and the effect of increased stimulating voltage on the magnitude of occlusion (F). The application of
Fig. 11. Hippocampal modification of cerebellar auditory response. Control response (A). Effects of increasing voltage (B-F). Parameters increasing from 0.3 msec, 3 V. (G) shows click response with and without preceding hippocampal stimulation. Artifacts retouched as in preceding figures. References p . 280
266
s. s.
F O X et
al.
higher voltages yielded larger direct responses to stimulation and resulted in further occlusion of the positive component of the click response. The relationship between voltage of hippocampal (conditioning) stimulation and the extent of occlusion is further detailed in Fig. 11. At a constant separation of 45 msec, a progressive increase in voltage produced a progressive diminution of the click response. As before, the positive wave was most affected. Blocking of cerebellar responses to light by hippocampal stimulation was also observed. Unlike the occlusion seen as a function of direct sensory stimulation or in the hippocampal-cerebellar auditory interactions, blocking in this case was complete at separations up to 80 msec and whether partial or complete, took place at a time when all recorded components of the direct response to hippocampal stimulationhad terminated. Fig. 12A shows control responses to light and hippocampal stimulation. A control light response superimposed upon a trace in which the light response was totally blocked by prior stimulation is also shown in Fig. 12B. Additionally, in Fig. 12B
Fig. 12. Hippocampal blocking of cerebellar light responses. (A) Responses to light (left) and hippocampal stimulation (0.2 msec, 3 V); @) Superimposed: response to light without and with (blocked) hippocampal stimulation (left). At right, same as left but with partial block of light response; (C) left: failure of a single hippocampal pulsb to block light response in the cerebellum (traces superimposed): right: three subthreshold pulses summate to block lighr response (traces superimposed); @) A single or double subthreshold pulse m a y block the cerebellar light response.
superimposed traces of a contiol and a partially blocked light response appear. The question of whether blocking can occur following hippocampal stimulation below threshold for eliciting an evoked response, was partially answered in another series of experiments (Figs. 12C and D). In the first experiment it can be seen that a single pulse delivered to the hippocampus was not sufficient to evoke a response in the cerebellum, and in this case no diminution in the light response was observed. Three pulses, how-
HIPPOCAMPAL EFFECTS O N CEREBELLAR A N D CORTICAL ACTIVITY
267
ever, produced a large evoked response and almost completely blocked the response to light (Fig. 12C). On the other hand, the results of a second experiment (Fig. 12D) show that one pulse, while again failing to evoke a direct response, still was capable of blocking the response to light. The extent of hippocampal blocking of the cerebellar and cortical response to light at various interstimulus intervals was also investigated (Fig. 10). Cerebellar responses to light were maximally blocked at separations of 30 msec as with click but remained blocked with separations as long as 80 msec, with full recovery occurring on the average only at 140 msec condition-test separations. This is in contrast with the inter- and intramodal, as well as the hippocampal-cerebellar auditory interactions in which recovery of the test response was more clearly progressive as the separation widened and was quite substantial by 80-100 msec. These data would suggest that the mechanism or site of interaction between hippocampal stimulation and cerebellar visual responses may differ from those mechanisms or sites underlying the other three types of interactions previously described. Hippocampal stimulation had an opposite, facilitatory effect on light responses of visual cortex. Fig. 10 also shows the average time course of such augmentation of the light response in the visual cortex following hippocampal stimulation at increasing intervals. An increase in the amplitude of the two major positive peaks as well as the amplitude of the negative-going deflection between them can be seen. This facilitation was maximal 40 msec separation, and was evident at all subsequent separations up to 130 msec; such facilitation, however, could be observed for condition-test intervals up to 200 msec. Comparison of the visual cortical and cerebellar light responses under control and stimulation conditions is also provided in Fig. 10. At a constant separation of 100 msec, a marked facilitation of the light response in the visual cortex occurred at the same time that the negative and especially the positive components of the cerebellar light response were substantially (50 %) reduced. At condition-test separations shorter than 20 msec, prior hippocampal stimulation produced enhancement of the amplitude of both the cerebellar click and light test responses. Mean hippocampal-click facilitation was 30 % over control amplitude and hippocampal-lightfacilitation was 20 % greater than control response amplitudes. NO parallel inhibition at such short intervals was seen for the cortical response to light.
Modification of the cerebellar response to hippocampal stimulation by preceding light flash Reversing the usual order of presentation of the stimuli also resulted in occlusion of the second response (Fig. 13). At 45 msec separation, light flash totally occluded a subsequent response to hippocampal stimulation, while at a slightly larger separation the occlusion was partial, showing inhibition of the initial, negative and particularly the second, positive component. Also in Fig. 13 (bottom, left) a control response evoked by hippocampal stimulation is superimposed upon a trace in which a light response occurred and partially blocked the subsequent, electrically-inducedresponse. In the top frame at the right are seen superimposed with 45 and 65 msec interstimulus separations. Total occlusion once again appeared at the briefer separation, but comReferences p. 280
268
S. s. FOX et al.
Fig. 13. Light block of cerebellar evoked response to hippocampal stimulation. Top left: response to hippocampal stimulation (0.08 msec, 4 V). Middle left: response to light flash which blocks response to hippocampal Stimulation at 45 msec separation. Bottom left: partial recovery of response to hippocampal stimulation at greater separation of flash and stimulation (55 msec). Superimposed for comparison is response to hippocampal stimulation alone. Top right: sequential, superimposed traces showing completeblock by light flaph of response to hippocampal stimulation (45 msec) and complete recovery at 65 msec separation. Bottom right: light flash and hippocampal stimulation present. Sequential superimposed traces with and without ambient light in room. When response to light flash is blocked by ambient light, response to hippocampal stimulation returns.
plete recovery of the direct response to hippocampal stimulation occurred when the separation was increased to 65 msec. More complex modification by flash of hippocampal blocking of cerebellar afferents is shown in Fig. 14. Using three sequential stimuli, it was possible to show multiple dependencies of the recovery cycles of the separate pairs. Fig. 14 shows the mean results of experiments in which hippocampal stimulation was used to block cerebellar click responses as in Fig. 10 (above). In these experiments, as before, the click response recovery cycle is shown as a function of separation from preceding hippocampal stimulation. In these experiments, however, a light flash is presented at varying intervals preceding the conditioning hippocampal stimulation. With decreasing flashhippocampal stimulation intervals, the recovery cycle for the click response is faster. It may be said that as the preceding flash increasingly attenuates the response to hippocampal stimulation, the effect of this conditioning hippocampal stimulation on the test response to a click is less and the recovery, therefore, is faster. The data of Fig. 13 (bottom, right) is also relevant to the question raised earlier concerning the necessity of appearance of an evoked response to the conditioning stimulus in order for interaction to occur. Complete occlusion at 45 msec separation is seen compared with the response to the light flash delivered in the presence of bright ambient illumination. Under these conditions, the light response failed to occur and the response to stimulation reappeared. In this experiment, therefore, blocking of the test response depended upon the integrity of the response to the conditioning stimulus.
HIPPOCAMPAL EFFECTS ON CEREBELLAR AND CORTICAL A C T I V I T Y
L, 20
33
40
50
60
HIFUCAMPAL STIMUTKN-CUCK S E R W "
70 80 (msec)(b)
93
269
100
Fig. 14. Complex crossmodalityrecovery cycles. Amplitude recovery of click response as function of hippocampal stimulation-click separation. Flash (for each curve) is presented at constant flashhippocampal stimulationseparation.As flash is presented further from stimulationin the other curves, hippocampal stimulation progressively blocks the click more. See text for details.
This finding is consistent with the results of Fig. 12C in which the necessity of the conditioning response for blocking of the test response was also observed, but not with the data of Fig. 12D.
Amygdala modification of cortical evoked potential and single cell responses to light Stimulation of the amygdala just preceding or simultaneous with presentation of a light flash also produced alteration of afferent responses in visual cortex. Both effects on evoked potentials and single units were observed. Evoked potential interactions with amygdala stimulation took the form of wave-form and amplitude changes while effects on single firing were in terms of latency and number of spikes in the response to the light. Stimulation of the amygdala produced up to a three-fold increase in amplitude of the evoked response to light (Fig. 15A) and resulted in considerable increase in complexity of early and late wave components (Fig. 15B). Amygdala modification of single cell firing took a number of forms. Such stimulation produced decreased latency of the response burst as well as a mean increase in total spikes in a burst; and this was the case whether a unit was spontaneously active (Fig. 16A) or silent before stimulation (Figs. 16B and C) or in response to either stimulus alone (Fig. 16B). Thus, cells which responded to neither stimulus alone responded variably to the two stimuli presented together. References p . 280
270
s. s.
FOX
et al.
Fig. 15. Amygdala modification of visual cortical light responses. (A) Control responses to amygdala stimulation (a) and to light (c); (b) and (d) show enhancement of response from combined amygdala stimulation and light. Light stimulation t r a m with and without amygdala stimulation superimposed in (d); (B) Two examples of complex wave-form enhancement and change with combined limbic (amygdala) stimulation and light. (a), (c), and (e) are control responses; (b) and (d) show effects of stimulus separations up to 40 msec. Note minimal direct evoked response to amygdala stimulation.
Decreased mean latency of single cell responses to light in the visual cortex was also observed with amygdala stimulation (Fig. 16C). Comparison of the frequency distribution of latencies of single cell responses to light, with and without preceding amygdala stimulation showed a mean decrease in response latency with amygdala stimulation of 10.1 msec from a mean control response latency of 38.9 msec without stimulation (N = 35) (Fig. 17). This substantial shift in latency with stimulation was not specificto any particular amygdala nucleus or such sets of nuclei, but was observed with the same frequency regardless of the amygdala site of stimulation. Discussion Wave-form, amplitude, latency, and duration of the various components of the cerebellar vermal response were markedly similar for all forms of stimulation used in this work: flash, click, and hippocampal stimulation. This is in agreement with the reports of previous investigators (Snider and Stowell, 1944; Gerard et al., 1936;
H I P P O C A M P A L EFFECTS O N C E R E B E L L A R A N D C O R T I C A L A C T I V I T Y
271
Fig. 16. Modification of single cell responses to light in the visual cortex by amygdala stimulation. (A) Control responses (a) and (b) are shown along with burst enhancement with amygdala stimulation (c) and (d); (B) A unit which responds to neither light nor amygdala stimulation alone (a) and (b), responds with variable burst to the two stimuli combined (c). (d), and (e); Q A cell which fires with fixed latency to light alone (a) and (c), shows decreased latency in response to amygdala stimulation and light (b) and (d).
Bremer and Bonnet, 1951;Levy et al., 1961;Albe-Fessard and Szabo, 1954; Szabo and Albe-Fessard, 1954)and might lead to either of two conclusions. The evoked potentials may reflect the response of a heterogeneous cell population upon which the termination of the different modalities on the vermian cortex overlap (Snider and Stowell, 1944), or alternatively the response similarity may be accounted for in terms of a homogeneous, polysensory population of cells. The present evidence does not permit determination of which of these two alternatives, of different, neighboring sensory cells or individual cells which respond to more than one sensory input, is correct. It is possible to suggest, however, two experimental approaches which might provide additional information. While Snider and Stowell found that the cerebellar visual and auditory projection areas were greatly overlapping, they were not perfectly contiguous. If it were possible to delineate ‘fringe’ areas of single sensory projection from which reReferences p . 280
272
s. s. FOX et al.
Fig. 17. Modification of spike latency in response to light by amygdala stimulation. Distribution of latencies of single cell responses to light in the visual cortex is shown (dotted bars) N = 35. Control = light alone. Amygdala stimulation produces greater proportion of cells showing shorter latencies (hatched bars) (P < 0.001).
sponses to a single sensory mode were still similar to response from the areas of overlap, the hypothesis of polysensory cells might be unnecessary. A second, more direct approach which we have taken, has been to record with microelectrodes single cell responses to different sensory modalities. The presence of cells in our microelectrode studies which respond to both visual and auditory input supports a concept of a cerebellar polysensory cell. The similarity of the response from hippocampal stimulation to the responses from light and click stimuli is not as easily understood. Although responses with very similar form have been described resulting from cortical stimulation (Szabo and AlbeFessard, 1954; Jansen and Fangel, 1961) no such relationship has been elucidatedpreviously for the hippocampus. While diffuse hippocampal cerebellar relationships have been suggested by the spread of hippocampal seizure after-discharge to cerebellar cortex (Morin and Green, 1953; Iwata and Snider, 1959), no studies previous to the present report have demonstrated suchprojections and certainly, therefore, no topological relationships have been established. Overlapping fields of projection may exist for reticular-cerebellar relationships (Whiteside and Snider, 1953)which might provide an organization for hippocampalgcerebellarafferentsvia a path from the hippocampus
H I P P O C A M P A L EFFECTS O N CEREBELLAR A N D C O R T I C A L A C T I V I T Y
273
to the mammillary bodies via the fornix, the mammillotegmental tract, and the tegmento-olivary system of fibersleading to thecerebellum.The assumption of an additional heterogeneous overlapping population of cells to mediate hippocampal responses is made unnecessary by our microelectrode studies which show responses to hippocampal stimulation, of the same polysensory cells which respond to other sense modalities (click and light), suggesting direct hippocampal termination or termination of the more indirect fibers from the hippocampus on these cells. Despite the microelectrode data showing the existence of cerebellar polysensory cells we cannot eliminate the possibility that the responses in these investigations resulted from different but neighboring cells. The interaction data, however, provide some additional information. First, the fact that interaction has been seen not only between the two physiological modes of stimulation but also between hippocampal stimulation and each of the physiological modes might support the view that termination is on common cells for all three rather than on three separate, overlapping fields. The similarity in recovery cycle data may be interpreted as providing further support for a common element. Recovery cycles for intraand intermodal click and light interactions as well as the recovery cycles for hippocampal interaction with these two sensory modes were remarkably similar. While from three separate, overlapping fields one might expect somewhat greater variability in recovery cycle characteristics between the three modes of stimulation, three cell populations of adequate anatomical complexity might provide the necessary interconnectivity to account for these results. Details of the cerebellar responses to electrical stimulation of the hippocampus along with a consideration of some aspects of the interactions observed, raise some additional questions regarding the possibly different specific anatomic sites of these effects. The series of synaptic interactions resulting from the use of paired pulses which showed a dependence of each component of the total response upon the full development of each previous component, as well as a change of sign, might suggest sequential elements, rather than independent afferent fibers, possibly involving presynaptic mossy fiber activity, postsynaptic granular activity and Purkinje cell efferent activity in the three sequential response components respectively. Jansen and Fangel (1961) have seen similar early cerebellar responses to stimulation of the auditory cortex and have attributed these to early and late Purkinje cell discharges. Further, the occlusive effects from paired stimuli are reflected predominantly by modifications of the late, slow, positive component of the cerebellar response, consistent with evidence showing the particular lability of this late component, its susceptibility to depressing influences,its augmentation by local strychnine, and most important, its summation as a result of appropriately timed paired stimuli (Bonnet and Bremer, 1951; Bremer, 1952; Bremer and Bonnet, 1951; Bremer and Gernandt, 1954. ) An exception to the view that cerebellar interaction is reflected by changes in the late, positive component is the hippocampal modification of light responses in the cerebellum. Hippocampal stimulation preceding a light stimulus did not only modify the late, positive wave but diminished the initial, negative component as well. This is in contrast with the effects of hippocampal stimulation on a click in which the late reReferences p. 280
274
s. s. FOX et al.
sponse is selectively diminished with no change in the initial, negative component and may also suggest two different sites of interaction for the two modalities with hippocampal stimulation. The difference between hippocampal effects on light and on click responses appears to lie primarily in the presence or absence of the effects on the early wave; hippocampal stimulation blocked the early wave of the light response but not the click response suggesting closer alignment of the hippocampal and light pathways than the hippocampal-click pathways. Further, interaction between click and light is reflected only in changes of the late, and not the early response component as might be expected if a common path fully subserved both modalities. The implication of such a suggestion would be that, to some extent, visual and auditory afferents maintain relative independence in cerebellum. Koella (1959) has also suggestedfundamentally different sites and modes of termination of visual and auditory afferent fibers in cerebellar cortex. We have also observed that the cerebellar auditory response is considerably more resistant to barbiturates (Nembutal) than the visual response, an observation which has been used as an indication of anatomically different paths (Koella, 1959). The inverse effects of hippocampal stimulation on visual cortical and cerebellar responses to light observed in these studies may be significant from a functional point of view. Within the range in which hippocampal stimulation inhibited the cerebellar visual response, the visual cortical response to light was simultaneously facilitated. Further, this facilitation was observed at separations at which the cerebellar response to stimulation of the hippocampus was fully recovered. Adey et al. (1957) have reported that hippocampal (gyrus) stimulation inhibits conduction within the brain stem. In other studies hippocampal modification of the visual cortical evoked potentials has been demonstrated by single pulse (Fox and O’Brien, 1962; Fox and Liebeskind, 1963) and tetanizing stimulation (Redding and Siegfried, 1962; Cazard and Buser, 1963). Grastyan et al. (1959) have proposed a role of the hippocampus in the control of the orienting reflex. These authors have also suggested a general, regulatory function of the hippocampus ‘probably at the highest functional level of the activating system’. Votaw (1959) has produced movements by hippocampal stimulation of the head, neck, arm, and upper torso which might be important in the orienting response. Finally, Kaada et al. (1953) have described an ‘arrest reaction’ from hippocampal stimulation in the awake, unrestrained cat; and Kaada (1951) has similarly produced inhibition of cortically-evoked movements in the unanesthetized preparation. To the extent that the orienting reflex represents a behavioral condition involving selection of both sensory and motor responses, these studies of hippocampal modification of both sensory and motor structures may be related. Implicationsfor interaction mechanisms The emphasis in these studies was on the demonstration of sensory interaction at a rather complex level in the cerebellum and the demonstration of limbic modification on cerebellar and cortical afferent signals. Sensory interaction or sensory modification was reflected by amplitude, wave-form and latency changes both in the afferent evoked
H I P P O C A M P A L EFFECTS O N C E R E B E L L A R AND C O R T I C A L A C T I V I T Y
275
potential or single cell response. We have not thus far considered the question of what common mechanism underlies these interactions. The following experiments concern not only further demonstrations of sensory interaction at the level of the single cell in the cerebellum and the effect of limbic system stimulation on these afferents, but may have considerably greater generality in terms of the basic neurophysiological mechanisms underlying processes of sensory interaction. In the first of these studies we examined the latency of single cell spikes in response to hippocampal stimulation. Single cells responding to hippocampal stimulation were easily found in the cerebellar vermis and constant latencies similar to the first inflection time of the evoked potentials described above (20-30 msec) were commonly observed. The data of Fig. 18 were taken from a wide variety of observations both within individual animals and across animals. The data were purposely selected from experiments in which the stimulus intensity was constant rather than from the series of experiments in which single unit latency was observed to change as a function of changes in stimulus intensity. Fig. 18 therefore presents the relationship between the normal variability in firing of a single spike response to click stimulation and the
-
95
-
0
L 05 I)
z P
5 2 75 E -I
c g 65 -
a
Y 0 0 W
55-
z
5z
45
-
W
35
25
c
-
a20030 0.40 0.50 0.60 o.maeo as0 LOO 1.10 1.20 1.30 1.40 1.50 1.60 1.70 1.801.90 200 MEAN RELATIVE FREQUENCY OF SPIKES TO CLICK
Fig. 18. Prediction of spike latency by probability of spike to preceding stimulus. Latency of cell response to hippocampal stimulation increases proportionally with increased probability (mean relative frequency in 20 sweeps) of spike response to click. Note that mean failure occurs at approximately 1.60 (1.60 spikes per sweep in response toclick). See text and Fig. 21. References p. 289
276
s. s. FOX et al.
latency of the single cell spike in response to a succeeding hippocampal stimulus. In these cases, the control latency of a single spike in response to hippocampal stimulation was observed and then a click stimulus similar to those described above was presented before the hippocampal stimulation. Under these conditions presentation of a click stimulus before the hippocampal stimulation blocks the hippocampal single-cell response on the vermis. In recording single cell responses to sensory stimuli, however, only with strong sensory stimuli or under unusual conditions was it possible to produce a response every time followingthe sensory stimulus. Rather, it was the case more often that single cell responses observed to sensory stimuli ranged from those seen quite regularly to those seen intermittently. In this case, groups of 20 sweeps have been taken beginning at an arbitrary point in a continuous record and the mean relative frequency of a single spike in response to the click stimulus has beep plotted. These data actually provide the mean probability of discharge but since at times a click stimulus of adequate intensity presented to the cat elicited from the cerebellum two, rather than one single spike, the graph has been presented as mean relative frequency of spike discharge per group of 20 sweeps following the click stimulus. Thus, for example, a mean relative frequency of 0.4 spikes per sweep indicates that a single spike was seen on the average in 8 of 20 sweeps and a mean relative frequency of 2 spikes per sweep, that the cell was firing twice on the average for every click stimulus delivered. As the probability of spikes in response to the click increased, the mean latency of response to the succeeding hippocampal stimulation increased until at about 1.6 spikes per click response (that is, at a point where one response to a click is occurring 100 % of the time and the second spike probability is at about 0.6) the response to hippocampal stimulation fails altogether. It should be noted here that the latency of the hippocampal-stimulated spike is not a function on any given trial of the number of spike responses to a click but is related only stochastically or probabilistically. These findings suggest that since on any given trial the response to a click may occur or not occur or simply occur with greater stability, that it is some other more basic function which is, in fact, determining this probability of response related to the latency of the spike to hippocampal stimulation. It is quite clear that the relative frequency of the preceding single unit spikes in response to the click alters the latency (and finally, probability) of the cell response to hippocampal stimulation, although these data are derived from mean relative frequencies and not from any absolute numbers of any given particular sweeps. It is, therefore, the likelihood of spike occurrence in response to the preceding click which alters the spike latency to stimulation, since the spike does not have to occur but only has to occur on the average. These results strongly lead to the inference that some second process may be responsible for the latency shift since it is difficult to see what the significance of 3/10 of a spike or 8/10 of a spike can have in terms of its effects on the latency of a response to the hippocampal stimulation that follows. The significance we feel is in the suggestion of a continuous process underlying the single unit data presented here. We do know of continuous and additive processes in the nervous system which would behave in this manner. Such underlying facilitatory or inhibitory influences would be additive and a
H I P P O C A M P A L EFFECTS O N CEREBELLAR A N D C O R T I C A L A C T I V I T Y
A. 3000
25004
277
C. W
n Y
4918 SWEEPS
IA 0
a W rn I
z I
B.
see
D.
Fig. 19. Relation between probability of firing of a single cell and evoked potential wave-form. (A) Frequency distribution of spikes from a single cell in the visual cortex of a cat after stimulation with 4918 flashes; (B) averaged evoked potential (200 oscilloscope sweeps) recorded from the same microelectrode, after cell death (r = 0.60; P < 0.001). Similarly, spike distribution for a single cell is shown in (C) (3150 sweeps) and the corresponding averaged evoked potential in (D) (150 sweeps) (r= 0.51 ; P < 0.001). Ordinate (for unit distributions): number of times the cell fired in response to light flash. Abscissa (for unit distributions): time, in 100-msecdivisions. (Reprinted from Fox and OBrien, 1965).
figure of 3/10 of a spike would suggest a 30-70 proportion of facilitatory to inhibitory influences. The suggestion that the evoked potential reflects a continuous process underlying the frequency or probability of spike occurrence has been made before by us in an earlier paper (Fox and OBrien, 1965), in which we demonstrated by computer compilation of a single cell frequency distribution that the probability of firing of a single cell following a stimulus is described by the evoked potential recorded from the same electrode (Fig. 19). The correlations in that study were observed to go as high as 0.89 between the single cell frequency-time distribution and the wave-form of the evoked potential recorded from the same microelectrode. Fig. 20 gives the sequential cross correlation of the pairs of averaged evoked potentials and spike frequency distributions shown in Fig. 19. The graph of Fig. 20 is the result of a computer program providing a point-by-point cross correlation by means of a moving triangular window technique. It should be noted that the high correlations between evoked potentials and spike distributions are consistent throughout the post-stimulus period and the all-over correlation, therefore, is not attributable to any particular synchromus early or late activities. References p. 280
278
s. s. FOX et al.
.-I 1.0
A
Fig. 20. Sequentialcross-correlation of single cell firing probability and evoked potential waveform. (A) Cross-correlationof A and B, Fig. 19. (B) Cross-correlation of C and D, Fig. 19. First (approximately 35) points give mean cross-correlation in each curve. Triangular window of 60 points moves, dropping last point and adding point at leading edge. Correlation is recomputed after each move. Note consistently high correlation throughout. Use of triangular window with weight equal to number in base gives greater weight to points just added, less to points to be dropped (further from leading edge).
The present data (Fig. 18) are strongly supportive of our earlier results and since the probability or mean relative frequency of the spike in response to the click altered the latency significantly, we expected that some aspect of the evoked potential recorded from the same microelectrode should be related in a similar manner. In another series of experiments, we evaluated the relationship between evoked potential amplitude as an arbitrary measure and again, the latency of single cell responses to hippocampal stimulation. The experimentswere the same as those of Fig. 18. A click stimulus was given during each sweep followed by hippocampal stimulation. The single-cell response was recorded from the microelectrode and the evoked potential response was differentiated through separate amplifiers. Fig. 21 gives the mean relationship between evoked potential amplitude in response to a preceding click stimulus and the latency of the single cell response to the hippocampal stimulation
H I P P O C A M P A L EFFECTS O N CEREBELLAR A N D C O R T I C A L A C T I V I T Y
-
-8
125
279
-
j 105
k 5
5
115
95-
2
w 85-
u %
75-
LL
0
&
E
65
-
55-
3z 45 3 z 35 I
I
I
I
I
I
I
I
I
I
I
I
0.10 0.200.m 0.40 0.50 a60 0.70080 0.90 1.00 1.0
I
I
I
I
I
I
I
I
I
120 1.30 1.40 1.50 1.60 1.70 1.801.90 2.00
RATIO DEVIATION~EANCLICK-EVOKED RESPONSE AMPLITUDE
Fig. 21. Prediction of spike latency by amplitude of evoked potential response to precedingstimulus. Latency of single cell response to hippocampal stimulation increases proportionally with increased amplitude of evoked response (relative to control) to preceding click. Note that mean failure of cell to fire occurs at approximately 1.40 (evoked response = 1.4 times control). See text and Fig. 18.
recorded from the same electrode. Since evoked potentials vary considerably from animal to animal and from time to time within the same animal, the measure used was the mean deviation in amplitude of any given evoked potential from the mean evoked potential for that animal for that day. In this way, groups of evoked potentials were averaged in amplitude, the mean deviation derived, and this amplitude measure related to the mean latency of the spike response to hippocampal stimulation. Fig. 21 shows almost in duplicate, the data of Fig. 18. The amplitude of the evoked response to a stimulus preceding the hippocampal stimulation clearly effects the latency of the spike to hippocampal stimulation. It is interesting to note that the shape of the two curves is almost identical and that the failure point is almost exactly the same as when latency is predicted by the probability of a spike occurring in response to click. It is concluded from these two experiments that in this case evoked potential amplitude in response to the first stimulus predicts the latency or probability of the spike to the next stimulus as well as spike probability, and consistent with our earlier findings, that the probability of spike discharge is closely related to the evoked response recorded from the same microelectrode. Different, however, from our previous work, in these experiments we have used evoked potential amplitude as an arbitrary measure and in the earlier paper, evoked potential wave-form was used. The grossness of these measurements, however, particularly in terms of the arbitrariness of choosing a given particular group for averaging, points up the strength of this relationship. Even with such techniques and no clear way to evaluate which particular sequences of evoked potential responses should be averaged together, the evoked response still strongly predicts the spike probability in the interaction with the stimulus that follows. References p . 280
280
s. s. FOX et al. REFERENCES
ADEY,W.R., (1958); Organization of the rhinencephalon. Reticular Formation ofthe Brain. H. H. Jasper et al., Editors. Little, Brown, Boston, pp. 621-644. ADEY,W. R., SEGUNDO, J. P., AND LIVINGSTON, R. B., (1957); Cortical influences on intrinsic brainstem conduction in cat and monkey. J. Neurophysiol., 20, 1, 1-16. D., AND SZABO, TH., (1954); Observations sur l’interaction des afftkences d’origines ALBE-FESSARD, periphbriqueet corticale destinh B I’borce c6rkbelleuse du chat. J.Physiol., 46,225-229. BONNET, V., AND BREMER,F., (1951); Auditory responses of the cerebellum. J. Physiol., 114, 54-55. BREMER, F., (1958); Cerebral and cerebellar potentials. Physiol. Rev., 38, 357-388. BREMER, F., AND BONNET, V., (1951); Caractbres ghkraux de la r6ponse du cervelet A uae volk d’influx affbrents. J. Physiol. (Paris), 43,662-665. BPEMER, F., AND GERNANDT,B. E., (1954); Acoustic response and the strychnine convulsive patterns of cerebellum. Actaphysiol. seand., 30,120-136. CAZARD,G., ET BUSER,P., (1963); Modification des rkponses sensorielles corticales par stimulation de l’hippocampe dorsal chez le lapin. Electroenceph. elin. Neurophysiol., 15, 413425. FADIGA,E., -ELI, G. C., VON BERGER, H.W., (1957); Cerebellar reactions to the visual system’s activation. Acta physiol. p h a r m o l . n?erl., 6, 284-294. FESSARD, A., (1961); The role of neuronal networks in sensory communications within the brain. In: Sensory communication, W.Rosenblith, Editor. Wiley, New York. Fox, S. S., AND LIEBEKIND,J. C., (1963); Hippocampal modification of cerebellar afferent information. Fed. Proc., 22, 451. Fox, S. S., AND O’BRIEN,J. H., (1962); Inhibition and facilitation of afferent information by the caudate nucleus. Science, 137,423-524. Fox, S. S., AND OBRIEN,J. H.,(1965); Duplication of evoked potential waveform by curve of probability of 6 h g of a single ceU. Science, 147,888-890. Gnwu>, R. W., MARSHALL, W. H., AND SAUL,C. J., (1936); Electrical activity of the cat’s brain. AMA Arch. Neurol. Psychiat., 36,675-738. GRASTYAN, E., L r s s ~ K., , MADARASZ, I., AND DONHOFFW, H., (1959); Hippocampalelectrical activity during the development of conditioned reflexes. Electroenceph.clin. Neurophysiol., 11,409429. GREEN,J. D., AND MACHNE,X., (1955); Unit activity of rabbit hippocampus. Amer. J. Physiol., 181,219-224.
IWATA,K., AND SNIDW, R. S., (1959); Cerebello-hippocampalinfluences on the electroencephalogram. Electroenceph. elin. Neurophysiol., 1 1 , 4 3 9 4 6 . JANSEN, J., AND FANGEL, C., (1961); Observations on cerebrocerebellar evoked potentials in the cat. Exp. Neurol., 3, 160-173. KAADA,B. R., (1951); Somato-motor, autonomic and electrocorticographic responses to electrical stimulation of the rhinencephalicand other structures in primates, cat and dog. Actaphysiol. seand., SUPPI.,83, 1-263. KAADA, B. R., JANSEN, I., AND ANDERSEN, P., (1953); Stimulation of the hippocampus and medial cortical areas in unanesthetized cats. Neurology, 3, 844-857. KOELLA, W. P., (1959); Some functional properties of optically evoked potentials in cerebellar cortex of cat. J. Neurophysiol., 22, 61-77. LEVY,C. K., LQESER, J. D., AND KOELLA, W.P., (1961); The cerebellar acoustic response and its interaction with optic responses. Electroenceph. clin. Neurophysiol., 13,235-242. M o m , F., AND GREEN, J. D., (1953); Diffuse afterchanges following stimulation of the fimbria hippocampi. Amer. J. Physiol., 175, 251-257. REDDING, F., AND SIEGFRIED, J., (1962); The effect of hippocampal stimulation on visual evoked potentials. Electroenceph. elin. Neurophysiol., 14,583-588. SNIDW,R. S., (1950); Recent contributions to the anatomy and physiology of the cerebellum. Arch. Neurol. Psychiat., 64,196-219. SNIDER,R. S., AND STOWELL, A., (1944); Receiving areas of the tactile, auditory and visual systems in the cerebellum. J. Neurophysiol., 7,331-357. SZABO,TH.,AND ALBE-FESSARD, D., (1954); Repartitions et caractbres des aff6renca somesthbiques et d’origine corticale sur le lobe paraddian de cervelet du chat. J. Physiol. (Paris), 46, 528-531. VOTAW,C. L., (1959); Certain functional and anatomical relations of the cornu ammonis of the macaque monkey. I. Functional relations. J. c o w . Neurol., 112, 353-382. WHITESIDE, J. A., AND SNIDER, R. S., (1953); Relation of cerebellum to upper brain stem. J. Neurophysiol., 16, 397413.
28 1
A Study on the Invasive Hippocampal 0-Waves to the Cortex Y. YAMAGUCHI, N. YOSHII, K. MIYAMOTO
AND
N. ITOIGAWA
Second Department of Physiology, Osaka University Medical School, Osaka (Japan)
One of the electrographicarousal states has been characterized by cortical desynchronization and hippocampal synchronization. It has been reported by Yoshii and his associates that the cortical desynchronization sometimes contains a well-developed activity at the frequency-band corresponding to the hippocampal &waves. This has been demonstrated by these workers in studies on varieties of conditioned reflexes and some innate behavior. In acute experiments with the cat we have found that, in response to arousal stimulation, some cortical areas exhibit augmentation on activity at &waves frequencies in association with the synchronization of the hippocampal activity. The experiment to be reported here was initiated to elucidate properties of the invasive hippocampal &waves to the cortex. In the following we shall use the term ‘iso-hippocampal rhythm’ (IHR) to denote the activity of extra-hippocampal structures which continues at least for 2 sec at frequencies of the 0 of 4-7 c/sec and waxes and wanes in synchrony with the 0-waves of the hippocampus. The present experiments were designed to investigate : (1) general features of the cortical IHR, (2) the cortical distribution of the IHR, (3) conditions for the appearance of the cortical IHR, and (4) the cortical excitability during the IHR. METHOD
Thirty-eight cats were used in acute experiments. Under preliminary ether anesthesia, necessary operations were performed. The cats were immobilized with Flaxedil, and no additional anesthetic was given except novocaine locally for the animal’s head mounted in a Horseley-Clark stereotaxic instrument. Standard oscilloscopic and electroencephalographic techniques were employed. Bipolar leads of the cortical EEGs were made with screws fixed to the skull. Screws were separated by 5 to 10 mm. In some experiments monopolar recordings were used, the framework of the stereotaxic apparatus being employed as the reference. The hippocampal EEGs were tapped with concentric electrodes. The cortical and hippocampal EEGs were recorded on an ink-writing EEG machine. Some of them were recorded on 8-channel magnetic tapes to facilitate frequency analysis. The frequency analyzer covered the frequency range from 1 to 30 c/sec. To elicit the hippocampal 0waves, electrical stimulation at 1 0 0 c/sec was applied to the brain stem structures such as the preoptic region, the hypothalamus, the anterior ventral nucleus of the thalamus References p . 292
282
Y. Y A M A G U C H I et
al.
and the midbrain tegmentum, a concentric electrode begin used. The cortical primary evoked potentials were produced by single shock stimulation (at a frequency less than 0.5 c/sec) of the medial and lateral geniculate bodies. These evoked potentials were recorded monopolarly with ball-tipped silver electrodes on the exposed cortex, or with screws into the skull. In some experiments, effects of the IHR were examined upon the direct cortical response which was produced by stimulating the exposed cortical surface with bipolar ball-tipped electrodes (tip distance, 2 mm) and was recorded monopolarly at a point 2 to 3 mm distant from the stimulated site. Changes in the direct cortical evoked potentiah were examined in relation to the cortical IHR. RESULTS
The IHR activity could appear spontaneously, but it was much easier to induce it by the arousal stimulation. The IHRs here presented were mainly obtained by the electrical stimulation of the brain stem structures. (1) Generalfeatures of the cortical IHR Fig. 1A shows an example of the cortical IHRs induced by stimulation to the nucleus ventralis anterior of the thalamus. Fig. 1B shows the results of the frequency analysis of the records of Fig. 1A. The cortical IHRs appear in the anterior and posterior temporal areas and the lateral occipital area. The frequency of the &activity (approximately 100 pV) was about 4.5 c/sec in this experiment. In the frontal and lateral parietal areas the EEGs showed an increase in the &band, but its main component differed in frequency from that of the 6-waves of the hippocampus. The temporal and lateral occipital IHRs started without an apparent delay from the onset of the hippocampal synchronization. There are, however, exceptional instances in which a few seconds’ delay was found between the cortical IHR and the hippocampal 8waves. The waxing and waning of the cortical IHR confirmed by the frequency analysis was similar to that of the hippocampus. The induced cortical IHR never continued indefinitely in the manner of afterdischarges after the hippocampus ceased to show the 6-activity upon cessation of the stimulation. It is suggested that the cortical neuronal mechanism of the IHR does not function independently of the hippocampus. It is well known that the induced hippocampal 6-waves become faster with increasing intensity of stimulation. A similar finding was obtained with the cortical IHR (Fig. 2). When the intensity of the stimulation was increased from 2.5 to 3.0 V, the main component of the hippocampal @-activitywas accelerated from 4.5 to 5.5 clsec. Corresponding to this, it was found in the cortical EEG that the main component shifted from 4.5 to 5.5 c/sec, keeping in harmony with the hippocampus. When the stimulus intensity was increased to 5 V, the main component of the hippocampal activity was initially 6.5 clsec, but it decreased to 5.5 c/sec just before the end of the stimulation. The synchrony of the cortical activity with the hippocampal activity of 6.5 c/sec was not complete, but was established with that of 5.5 c/sec toward the end of the stimulation. This slowing down of the hippocampal activity continued for about 4
a> < m
I/)
4
0
Fig. 1. Simultaneous recording of electrical activity from 16 cortical points, the dorsal and ventral hippocampus. (A) Original EEG records. Cortical electrodes were connected bipolarly. Between two arrows the nucleus ventralis anterior was stimulated for 8 sec with parameters as indicated. IHRs are seen in the lateral occipital (channels 9 and lo), anterior and posterior temporal (11 and 13) leads upon inspection. Calibration: 100 pV for cortical leads and 200 pV for hippocampal leads. ( B ) Analyzed EEGs. Leads marked by asterisks in (A) were subjected to frequency analysis. Values to the left indicate frequencies in cycles per second. In dorsal and ventral hippocampal leads (15 and 16), the main component of 8-waves induced was initially at 4.5 c/sec and shifted to 3.75 c/sec during the stimulation. Toward the end of the stimulation it was at 3.5 c/sec. Lateral occipital (10) and posterior temporal (13) leads behaved in the same way. In the remaining cortical leads there were no corresponding signs.
2
m 0
5 $r
284
Y. Y A M A G U C H I et al.
Fig. 2. Effects of increasing intensitiesof reticular formation stimulation. Simultaneous records of the anteriortemporal cortical and ventral hippocampal activity are shown above and below, respectively. In each frame, the uppermost tracing is the original EEG and the remaining ones are analyzed EEGs whose frequencies are indicated to the left. See further in text.
sec to attain 3.75 c/sec finally. During this slowing down of the hippocampal activity, it was seen that the cortical IHRs successively decreased their frequency from 6.5 to 4.5 c/sec. But, when the hippocampal activity had shifted its main component toward 3.75 c/sec, there were no traces of the corresponding cortical IHRs. This is striking because there is a very large amplitude of the hippocampal activity without the association of the cortical IHR. It may therefore be reasonable to conclude that the cortical IHRs are not due to physical spread of the hippocampal activity, but originated in the cortex itself. (2) Cortical distribution of the IHRs
For examining the cortical distribution of the IHRs, 16 screws were fixed to the skull (Fig. 1A). Based upon 7 experiments, the extent of the cortical areas where the IHRs could be observed, determined by bipolar recording, are schematically shown in Fig. 3. The IHRs appear most frequently and clearly in the lateral occipital and posterior temporal areas. The IHR activity was not observed in the frontal area, though occasionally in the sensori-motor area there was seen a weak but well-defined IHR activity. During monopolar recording, the IHRs were seen widely all over the cortex. ( 3 ) Conditionsfor the appearance of the cortical IHR
The cortical IHR mainly depended upon the regularity of the hippocampal 8-
INVASIVE HIPPOCAMPAL @-WAVES TO THE CORTEX
285
Fig. 3. Cortical map illustrating distributionof IHRs. Hatches show the areas of the IHRs. Crossed hatches indicate the most active areas. This map is based upon observations in 7 cats.
activity. That the hippocampus shows well-synchronized,large amplitude &activity at 4-6 c/sec is one of the most favorable conditions for the cortex to exhibit the IHR. Among the factors of the hippocampal activity, the synchronization seems to be a more critical factor than the amplitude or the frequency. Another factor involved in the appearance of the cortical IHR is the level of the background activity of the cortex. The cortical IHRs were easily induced when the cortical EEGs were desynchronized or augmented in fast-wave components to some extent. When the cortical irregular slow waves or sleep spindles occurred while the hippocampal &waves were maintained, though not so regularly, the IHRs were scarcely observed. And also, the cortical desynchronization, produced by too strong arousal stimuli, did not accompany the IHRs, even though the &waves were welldeveloped in the hippocampus. The cortical distribution and amplitude of the IHR and the threshold for elicitation depended upon the stimulation site. The hypothalamic stimulation was most effective in producing cortical IHRs. This is probably because it has a high effectiveness in producing the hippocampal &activity. However, the effectiveness upon the hippocampus is not a sole factor, as shown in Fig. 4. In this experiment a pair of stimulating electrodes was placed each in the hypothalamus and the midbrain reticular formation. By adjusting the intensity of stimulation, the hippocampal 8-waves elicited from the hypothalamus and the reticular formation were made broadly similar to each other. In the comparison of the effects of two stimulations, it was found that the cortical IHR appeared in more restricted cortical areas and with smaller amplitudes in response to reticular than to hypothalamic stimulation. This might indicate that the reticular stimulation caused too strong neocortical arousal of desynchronization and this latter activity disturbed the appearance of the IHRs. The tonucof the brain stem activating system seems to be maintained at a moderate level for inducing the cortical IHR.
(4) Cortical excitability during the IHR In the resting condition the primary potentials of the visual cortex evoked by stimulation of the lateral geniculate body fluctuated, especially in the negative component, more or less spontaneously. When the IHRs were evoked by hypothalamic stimulation in the visual cortex, this fluctuation of the evoked potentials was reduced during the maintenance of the IHRs (Fig. 5). References p . 292
286
Y. Y A M A G U C H I
et al. 0
P
EEG 2.5- 3 3-35
35- 4 4-5 5 -6 6-7
EEG
2.5- 3 3-35
35- 4 4-5 5 -6
iJ, 6 - 7 100%0.lmrC 5v
t
b
t
P
0
EEG 25- 3 3 - 35
35- 4 4-5 5 -6 6-7
EEG
25- 3 3 - 35 35- 4 4-5 5 -6 6-7
Fig. 4. Cortical distributions of the IHRs by ventromedial hypothalamic(below) and reticular (above) stimulation applied between two arrows for 10 sec. The hippocampal activity to stimulationwas made broadly similar in the two experiments. Upon hypothalamic stimulation, cortical IHRs were seen in the medial (channels 3,4 and 5) and lateral (8 and 9) parieto-occipital area, middle (12) and posts rior (13 and 14) temporal area.On reticular stimulation, cortical IHRs were seen in the lateral (8 and 9) parieto-occipital area and middle (12) and posterior (13 and 14) temporal region. To the nght, original and analyzed EEGs are shown which came from parietal (P), Occipital (0), middle temporal (mT)cortical areas and ventral hippocampus (v. Hippo). Calibration: 100 pV for original cortical EEGs and 200 pV for hippocampal EEGs.
During the hypothalamic stimulation, the main component of the hippocampal 8waves was found at 4.5 c/sec initially and then it shifted to 5.5 clsec. The main component of the cortical activity changed its frequency following the hippocampal one.
I N V A S I V E H I P P O C A M P A L 6 - W A V E S TO THE C O R T E X
287
Fig. 5. Effects of cortical IHR upon visually-evoked potentials. IHR was induced by ventromedial hypothalamic stimulation. Evoked potentials, elicited by stimulationof the lateral geniculatebody at 0.3 c/sec, are shown at the top. The negativity is indicated upward. Tracings labeled with VC and Hippo are original EEGs from the visual cortex and the dorsal hippocampus, respectively. Their analyzed components are shown below each original EEG, whose frequencies are indicated to the left. The records are mounted in Figs. 6,8 and 9 in the same way and with the same convention as in Fig. 5.
During this time, the evoked potentials were regularized. Just before the cessation of the hypothalamic stimulation, the main component of the hippocampal &waves shifted from 5.5 clsec to 4.5 clsec. A further shift in frequency toward 2.75 c/sec was accomplished within 7 sec. Such shift in frequency of the main component was seen in the cortex. But the regularization of the cortical responsiveness could be traced in the first two evoked responses during the IHR of 4 clsec after the cessation of the hypothalamic stimulation. It is noteworthy that the negative phase of evoked potentials is augmented in association with the IHRs, although the average amplitude did not exceed the maximal amplitude seen in the pre-stimulation period. When the hypothalamic stimulation acted upon the cortex to cause the welldesynchronized activity instead of the IHR, the testing evoked potential was found to be augmented markedly. But this facilitation soon subsided (Fig. 6). These features are shown graphically in Fig. 7. The functional state of the cortex in the appearance of the IHR is characterized by regularization as well as augmentation of cortical excitability, though they could appear without the IHRs. The regularization and augmentation of the cortical excitability during the IHRs were also seenin the auditory cortex, the primary evoked response produced by stimulation of the medial geniculate body being taken as index (Fig. 8). The effects of IHRs were witnessed in both surface-positive and -negative components. The direct cortical responses were not always affected by the IHR, but regularization and augmentation was observed when the response had been fluctuating very much during the control state (Fig. 9). References p . 292
Y. YAMAGUCHI et al.
288
Fig. 6. Effects of cortical desynchronization upon visual cortical potentials. Immediately following hypothalamicstimulation,the evoked potential was enhanced, but it soon declined.
P " L " " " " " " " "+ - I " '
f 0
E
a
IHR
---___ 4
I -
G L st.
A L
3.5 BeC
Fig. 7.Time course showingeffects of the IHRs and the strong desynchronizationupon the amplitude of the negative phase of the visual cortical evoked potentials. The IHR was induced by hypothalamic stimulation at 100 c/sec. The evoked potentials were obtained by lateral geniculate stimulation at 0.3 c/sec and their amplitudesat thenegativepeaks wereplottedsuccessivelyon the ordinates. Stimulation is indicated by horizontal lines. During stimulation IHR was induced in the upper record and strong desynchronizationin the lower one. After cessation of stimulation IHR occurred in the upper record as indicated by the broken line.
I N V A S I V E H I P P O C A M P A L 6 - W A V E S TO THE CORTEX
289
Fig. 8. Effects of cortical IHRs upon auditory cortical potentials. Stimulation was applied at 0.3 c/sec to the medial geniculate body. The hippocampal &wave, elicited by hypothalamic stimulation, had its main component at 4 c/sec. Corresponding to this, the cortical IHR was maximal at the same frequency. During these cortical IHRs, evoked potentials were facilitated and regularized. This effect continued for about 5 sec after cessation of the stimulation. During this time, the cortical IHR and hippocampal 0-waves shifted their frequency from 4 to 3.75 c/sec.
Fig. 9. Direct cortical responses during the presence of cortical IHR. The direct cortical response was recorded from the gyrus lateralis posterior (Lp). Veutromedial hypothalamic stimulation was used to induce cortical IHR. Hippocampal &waves had their main components at 3.75 c/sec initially and at 4-4.5 c/sec later. During this time direct cortical responses were increased and regularized. References p . 292
290
Y. Y A M A G U C H I et al. DISCUSSION
( I ) On the appearance of cortical IHR Green and Arduini (1954) reported that the hippocampal &waves could be traced in the fornix, mammillary body, mammillo-thalamic tract and anterior thalamus in the rabbit. Yoshii et ul. (1958) stated that the subcortical IHRs were seenin theamygdala, septa1 region, fornix, reticular formation and intralaminar nuclei of the thalamus during the conditioning of the dog. Van Reeth (1959) and Petsche and Stumpf (1960) found the cortical as well as subcorticalIHR in the rabbit. By the method of frequency analysis Yoshii et al. (1966) found, in the experiments on conditioned behavior and activated sleep in dog, that an increase in the 8-component in the cortical EEG was due to the hippocampal &wave. It may be reasonable to consider that the cortical IHRs described by these workers and in this paper are caused by impulses originating in the hippocampus and passing through some of the above-describedsubcorticalstructures. The cortical distribution of hippocampal after-discharges reported by Elul (1964) is similar to that of our cortical IHR. The two phenomena seem to be mediated by a common pathway. Green and Arduini (1954) showed that the cortical IHR was not seen when the tip distance between bipolar cortical electrodes was less than several millimeters. In contrast with their experiment, by a bipolar recording similar to that of Green and Arduini, we often found that the well-developedcortical IHR is detectable without the help of the frequency analyzer. When the frequency analysis was applied, the cortical IHR could be detected from the diffusely desynchronized EEG records which seemed not to contain an apparent IHR activity. The fact that the amplitude of hippocampal &waves is not a decisive factor in inducing the cortical IHR excludes the possibility that the cortical IHR is due to physical spread from the hippocampus. The most important conditions for appearance of the IHR are regularization of hippocampal 8waves and the functional state of the cortex. The cortical excitability necessary for the IHR is between the resting arousal and the strong desynchronization. (2) Cortical responsiveness during the IHR The primary cortical potentials evoked by stimulation of the medial or lateral geniculate body were augmented to some extent and regularized in the presence of cortical IHR. The facilitation was marked in the surface-negative component which are believed to be due to dendritic discharges. When a strong desynchronization appeared in the cortex, these facilitatory effects were much more marked, and the regularization was rarely observed. In regard to the direct cortical response, the facilitation and the regularization were also noted during ,the IHR, although they were difficult to see in some experiments. Among the effectselicited by arousal stimulation upon the cortical excitability, the regularization is characteristic of the IHR, not of the strong desynchronization,whereas this relation is reversed with the facilitation. Bremer (1959), Dumont and Dell (1960) and Demetrescu et ul. (1965) reported on the reticular facilitation of the primary evoked potentials, but did not pay much attention to the regularization and its relation to the cortical IHR. It is supposed that the arousal
INVASIVE HIPPOCAMPAL
8-WAVES
T O THE CORTEX
291
stimulation in their experiments was frequently made so strong that the cortex was activated to show a strong desynchronization instead of the IHR.
(3) Interpretation of the cortical IHR during the conditioning Yoshii et al. (1965) found that the IHRs appear in some cortical areas just before the performance of the food-rewarded lever-pressing. The IHRs were also noted in the defensive conditioned reflexes as an EEG response preceding the peripheral conditioned response. Recently, Yoshii, Itoigawa and Ueno (unpublished) proved a significant relationship between correct performance in the recent memory test and the cortical 0-wave component. According to their analysis the correct performance is seen when the cortical and hippocampal EEGs show an increase in the &wave components. Moreover, the indifferent impulses experimentally given to the brain stem structure during the test period are excluded from the sensori-motor cortex where the IHR is induced during the period of viewing food. The cortical activity with IHR is supposed to be moderately aroused and excludes the indifferent impulses coming to the cortex for inducing the correct performance in the conditioned and learned behavior. Parasomnic behavior of humans, such as sleeptalking, adult enuresis and some sleep-gnashing were found to appear in the activated sleep stage with IHR in the cortical EEGs, which was shown by the frequency analysis. SUMMARY
(1) Properties of the cortical IHRs (iso-hippocampal rhythm) were investigated on immobilized and locally anesthetized cats. The cortical IHR means the activity which continues for at least 2 sec at frequencies of 8-waves (4-7 c/sec) and waxes and wanes in synchrony with the hippocampal &waves. The cortical IHR was most clearly and easily detected by the method of EEG frequency analysis. This type of cortical activity could appear spontaneously, but studies in the present experiment were limited to those elicited by high frequency arousal stimulation of the brain stem structures. (2) The cortical IHRs appeared dominantly in the occipital and temporal areas. (3) The large amplitude of regular hippocampal synchronization at a frequency of 4-6 c/sec facilitated the appearance of the cortical IHR. A moderate neocortical arousal was favorable to the appearance of the cortical IHR. (4) The primary cortical potentials evoked by electrical stimulation of the lateral and medial geniculate bodies were augmented and regularized during the cortical IHR. (5) The direct cortical response had similar trends in augmentation and regularization during the cortical IHR. (6) The functional implication of the IHR is discussed in relation to the conditioned and innate behavior. ACKNOWLEDGEMENT
This work was aided by a grant from the Foundations’ Fund for Research in Psychiatry (FFRP Grant 61-233). The authors express thanks to Dr. K. Iwama for his kind discussions. References p. 292
292
Y. YAMAGUCHI et.aZ. REFERENCES
BREMER, F., (1959); L'interprbtation des potentiels electriques de l'bcorce drbbrale. Structure and Function ofthe Cerebral Cortex. D. B. Tower and J. P.Schad6, Editors. Amsterdam, Elsevier, pp. 173-191.
DEMETRESCU, M., DEMETRESCU, MARIA,AND IOSIF,G., (1965); The tonic control of cortical responsiveness by inhibitory and facilitatory diffuse influences. Electroenceph.elin. Neurophysiol., 18, 1-24. DUMONT, S., AND DELL,P., (1960); Facilitation r6ticulaire des meCanismes visuels corticaux. Electroenceph. elin. Neurophysiol., 12, 769-796. ELUL,R., (1964); Regional differencesin the hippocampus of the cat. 11. Projectionsof thedorsaland ventral hippocampus. Electroenceph. elin. Neurophysiol., 16,489-502. GREEN, J. D., AND ARDUINI,A. A., (1954); Hippocampalelectrical activityin arousal. J. Neurophysiol., 17, 533-557.
Topographic and toposcopic study of originandspreadofthe regular synchronized arousal pattern in the rabbit. Electroenceph. din. Neurophysiol., 12,589-600. VANREETH,unpublished data, c$ BREMEB, F., (1959); L'interprktation des potentiels klectriques de l'korce drkbrale. Structure and Function of the Cerebral Cortex. Amsterdam, Elrevier, p. 184. YOSHII,N., (1965); Background activities controlling conditioned reflex and behavior Proc. 23rd h t . Congr. physiol. Sci.,Tokyo, Proc. IV, 354-358. YOSM,N., MATSUMOTO, J., M m , S., HASEGAWA, Y., YAMAGUCHI, Y., SHIMOKOCHI, M., HOW, Y., AND YAMAZAKI, H., (1958); Conditioned reflex and electroencephalography. Med. J. Osaka Univ., ~ETSCHE,H., AND STUMPF,CH.,(1960);
9,353-375.
YOSHII,N., MATSUMOTO, J., OGURA,H., SHIMOKOCHI, M., YAMAGUCHI, Y., AND YAMASAKI, H., (1960); Conditioned reflex and electroencephalography. Electroenceph. elin. Neurophysiol., SUPP.13, 199-210. YOSM,N., MIYAMOTO, K., AND SHIMOKOSCHI, M., (1965); Electrophysiological studies on the conditioning of frequency specific waves. Med. J. Osaka Univ., 15,321-344. YOSHII,N., SHNOKOCHI, M., MNAMOTO, K., AND ITo, M., (1966); Studies on the neural basis of behavior by continuous frequency analysis of EEG. Progress in Brain Research, 21A. T. Tokizane and J. R. Schadb, Editors. Elsevier, Amsterdam.
293
The Limbic Cortex, Its Connections and Visceral Analyzers E. SH. AIRAPETYANTS
AND
T. S. SOTNICHENKO
The Pavlov Institute of Physiology, USSR Academy of Sciences and The University of Leningrad, Leningrad (USSR)
Complex and little understood systems of various organisms, i.e. inner and external organizers of behavior, are nowadays a subject of research for both related and unrelated branches of science. Even within physiological science which deals with the study of the brain there have appeared various independent groups of research workers with no less varied views and methodological approaches. It becomes extremely difficult, frankly almost impossible, to keep informed about the vast results of this study. The ways of research have widened immensely. It is practically impossible to define in total this or that trend in the study of the most vital brain mechanisms. The necessity for co-operation in our research work is obviously pressing. It is still more important to understand the functions of the brain’s parts and its microstructures in connection with the laws of activity of the brain as a whole. That is exactly where we face the most difficult task, comparable to climbing along the edge of a dangerous Alpine path; we have to stay within the borders of strictly physiological analysis, i.e. to deal with mechanisms. At the same time, the necessity for applying general biological evaluation of nervous activity to the aims of our work, the study of the human brain, draws us to the tempting mirages of anthropomorphism. We say this because the subject of our research is the brain of inner analysis, and it may be that it is most difficult for us to resist the temptation to make generalizations of this kind, though we are firmly convinced that reflex analysis is the only reliable way of research. In co-operation with Balacshina, and then with a number of colleagues in our laboratory it was established that these interoceptive or, to be more specific, visceral conditioned signals differ in accordance with modality of stimulation, i.e. mechano-, chemo-, thermoanalysis. Following the teaching of Pavlov we quite reasonably qualified this system of interoceptive information as analyzers (Airapetyants, 1940, 1952, 1959, 1963). Let us not discuss the unresolved classification of internal analyzers. It is more convenient to use the term ‘visceral analyzers’ when talking about signals from internal organs. The facts show that visceral conditioned reflexes and higher visceral analysis like exteroceptive conditioned reflexes and external environmental analysis are closed and carried out in the brain cortex of higher vertebrates. Thus, for example, in our laboratory, as a result of experiments on dogs with complete decortication, it was proved that both the formation and analysis of visceral conditioned reflexes are impossible (Lobanova, 1965). The description of the cerebral system and References p . 303-304
294
E. SH. A I R A P E T Y A N T S A N D T. S. SOTNICHENKO
e.specially of the cortical localization of central links of visceral analyzers has attracted our attention during recent years. As a result of many years of experiments, data were accumulated which speak in favor of morpho-physiological apportionment of certain afferent regions of the cortex where nuclei of visceral analyzers are concentrated. This part may be called the ‘visceral cortex’. Newly acquired experimental materials make it possible to develop our thesis on the special significance of the sigmoid gyrus region (of 4th and 6th fields) in the cortical apparatus of viscero-mechanical and viscero-chemical and also of respiratory and skeleton analyzers. Of special interest in this cycle of investigations are two groups of regularities that reveal the specific character of the formation of the visceral cortex. First, it is evident that overlapping of cortical centers (nuclei) of visceral and proprioceptive analyzers takes place, and that the area gigantopyramidalisturns out to be this common neuron structure; of no less importance is the overlappingthat also occurs at the thalamus level. Further it was found that, after dissection of the frontal lobe (8 and 12 fields in one group of dogs, and 3 and 43 fields in another group), visceral (from stomach and intestine) food conditioned reflexes are fully preserved (Zubkova, 1966). Does this mean that the above-mentioned cortical fields cannot be attributed to the apparatus of visceral analyzers? Are mechanisms of temporary connection and higher visceral analysis carried out only in fields 4 and 6? (Fig. 1).
Fig. 1. Preliminary scheme of the visceral cortex. The regions, extirpation of which depresses the interoceptive conditioned reflexes, are dotted; the regions, extirpation of which does not influence interoceptive conditioned reflexes, are shaded.
It may be supposed that under these extraordinary conditions (damage to the central part of the visceral cortex) the system of virtual mechanisms acts as the substitute, and thus the functions of fields 4 and 6 are transferred to the same 8, 12, 3 and 43 fields either separately or jointly. This is one of the possible ways of reintegration of analyzer functions, or in other words, one of the mechanisms of reliability of cortica 1
CO NNE CT IO NS O F LIMBIC CORTEX
295
neuron organization. At the same time the search for other afferent cortical projections is justified, and not only in the fields of the frontal lobe. Developing the subject of our research from this point of view, we began many years ago to study limbic fields and simultaneously some subcortical formations of visceral analyzers. The concept of the ‘limbic system’ has become especially popular recently. There are many reasons for this. Some scientists are impressed by an unusual unification of numerous, mainly subcortical, formations into a certain functional unit. According to a number of authors, this limbic focus mysteriously encloses mechanisms of complex elements of behavior - physiological, biological, and even social. It seems original to postulate the relative independence of the limbic system of the neocortex. Most morphological formations now included in the ‘limbic system’ are known to be part of the rhinencephalon. Papez (1937) and MacLean (1949) may be rightfully considered the original initiators of the theory of the limbic system. They considerably widened what was originally Broca’s scheme. Along with the limbic lobe, the insula, temporal tip and also the amygdala, septum, hypothalamus and the nuclei anterior of the thalamus were introduced. In the works of Smith (Smith, 1944) the descriptions of the diversity of functions of the cortical part of the limbic system led different authors to call it either an additional motor field (Kennard, 1955), an inhibitory field (Smith, 1944; Ward, 1948; Kaada, 1955, 1960), an autonomic region (Kramer, 1947; Ward, 1948) or a cortical center of emotions (MacLean, 1940; Fulton, 1953). The data on the limbic system scheme was enriched by the data concerning its connections with other parts of the brain by Brady (1958) and Nauta (1960). Nauta found that the medial diencephalon brain is connected exclusively with the limbic system. Highly respected Japanese colleagues (Tokizane et al., 1963; Torii and Kawamura, 1960; Kato et al., 1960) established that the limbic system has an activating system of its own - the posterior part of the hypothalamus. This system is independent of the activating system of the neocortex, i.e. the reticular formation. If we digress from the limbic system as a whole and consider the limbic cortex directly, we should point out that rich material about its structure and connections has now been accumulated. Nevertheless the problem of the functional significance of this region remains unsolved. In our opinion the inadequacy of data on afferent functions of the limbic cortex, i.e. about the presence of projection of receptors in it, is the major obstacle in the study of limbic cortical physiology. Recently this area has been enriched by the work of a number of Russian scientists (Delov et al., 1961 ; Gaza, 1962; Gulyaeva, 1962; Musyaschikova, 1964; Sovetov, 1963; Tolmusskaya, 1964; Chernigovsky and Zareiskaya, 1962). Our early physiological research showed that previously established conditioned reflexes from the intestine and kidney are temporarily reduced (for 2-3 weeks) in dogs when the posterior part of the limbic region is damaged unilaterally (Adam, 1957). Next we described a prolonged change (over a period of several months) in the conditioned reflex upon stimulation of the stomach and intestine in adult dogs with References p. 303-304
296
E. SH. A I R A P E T Y A N T S A N D T. S. S O T N I C H E N K O
bilateral extirpation of anterior limbic areas (Airapetyants et al., 1960). In puppies this change took place for different periods of time depending on their age (Moiseeva, 1962). Recently our study of the role of the limbic region has been especially attracted to the activity of the viscero-chemical analyzer. Here are some results of the experiments done by Vasilevskaya (1962, 1963, 1964), Vasilevskaya and Udalova (1965), Gasanov (1964), Nikitina (1964) and Pogrebcova (1965). If, in intact dogs, acid loading intensifies the acid defense reflex, and alkali loading reduces it, then in animals with a bilaterally damaged anterior limbic region this adaptive reaction disappears for a long time. When the same phenomenon was studied in the posterior limbic region, it was found that bilateral extirpation does not exclude the possibility of carrying out the closing function in relation to viscero-chemical stimulations, although the normal conditioned defense reaction is disturbed under acid and alkali loadings. The influence of the extirpation of the anterior limbic region of the cortex on food behavior, specifically on the analysis of sodium chloride stimulators under normal conditions and with dehydration of the organism, was studied in dogs by the method of free motor food conditioned reflexes in the form of free running in the menage. It was shown, that prolonged (for 7.5 months) phase changes of food behavior followed as a consequence of extirpation. Proper correlation of the consumption of a particular substance with the content of it in the internal environment of the organism was distorted. Later there appeared a stable inhibition of ‘running’ to the rack with salty food but the reaction to the insipid food signal remained. During another series of experiments we studied changes in blood sugar level with gastric stimulations and with injection of carbocholine. We observed that the unconditioned glycemic reaction, upon carbocholine injection as well as upon mechanical stimulation of the stomach, was disturbed after damage to the anterior limbic part. Correspondingly pre-established positive exteroceptive hyperglycemic reflexes were lost or depressed for a long period of time (up to 3-3.5 months). The positive interoceptive glycemic conditioned reflex turned out to be depressed for 4-4.5 months. The restoration of interoceptive and unconditioned and conditioned reflexes occurred after the above-mentioned periods of time. According to a number of scientists, stimulation of the limbic cortex brings about changes in the activity of the respiratory system (Smith, 1964; Kramer, 1947; Kaada et al., 1949). At the same time the problem of the significance of the limbic cortex in the analysis of afferent impulses coming from breathing apparatus receptors, and playing an important part in the regulation of respiration, remains open for further research. Both efferent and afferent influences of the anterior linbic cortex upon the respiratory system were introduced in our experiments. Exteroceptive respiration conditioned reflexes were formed on the basis of an unconditioned hypercapnic reflex, i.e. breathing of air containing about 5% of carbon dioxide. In another chamber an increase in carbon dioxide concentration was used as conditioned stimulus of the defense motor reflex (jerking back of the leg in response to electrical shock). Respiratory activity is disturbed after bilateral extirpation of the anterior part of the limbic system. In these circumstances analysis of stimuli coming to analyzers of the
297
C O N N E C T I O N S O F LIMBIC CORTEX
respiration system, i.e. the afferent part of the respiration system, suffers most. The state of the efferent part of respiration reflexes, judging by exteroceptive respiration conditioned reflexes, does not change considerably. At the same time there were found certain disturbances in breathing in the state of rest as well as under emotional loadings and also in the state of hypercapnia. The disturbances were temporary, and restoration of distorted functions was observed in about 2 months (Fig. 2).
$ 40 60
8ot
The negative conditioned reflexes
I
100
Fig. 2. The changes in conditioned reflexes from the breathing organs after damage to the anterior limbic region.
The experiments described above make it obvious that the function of the viscerochemical analyzer is considerably disturbed with damage of the limbic cortical complex and especially with extirpation of anterior lobes of the brain. Elaborated signaling is depressed, adequate chemoreception becomes defective, and functions of cortical and subcortical defense mechanisms are distorted for a long time. We found no specific deviations in the behavior of dogs after extirpation in the motor region in comparison with those after extirpation of the limbic cortex. The dogs were not abnormally aggressive or malicious, and no other form of intensified positive or negative state of emotion was observed. We have been unable to find the objective reason for the disagreement of our results with the work of other authors. Research in electrical activity of the anterior and posterior1limbic region, andmotor, parietal and occipital regions was carried out in rabbits with injection of 10 ml of 5 % solution of sodium chloride into the blood. Syilchronous changes in cortical electrical activity were found clearly expressed in motor and limbic regions. They were absent from the occipital part. There were periodical volleys of waves of high amplitude with frequency of 8-12 per sec or low amplitude waves with frequency of 1-2 per sec (Fig. 3). The visceral EEG took a different form depending on qualitative and quantitative differences in chemical substances (hydrochloric acid, alkali, sodium chloride, calcium chloride). Low concentration solutions mainly provoked high amplitude outbreaks. High concentration solutions for the most part provoked slow waves. Administration of aminazine abolished the above-mentioned effects. The influence of References p. 303-304
298
E. SH. A I R A P E T Y A N T S A N D T. S. S O T N I C H E N K O
Fig. 3. EEG of the rabbit after intravenous administrationof sodium chloride.
other substances such as ephedrine, which intensifies reticular formation tone, did not bring about an activating reaction in the limbic system, Specifically for visceral stimuli, changes in electrical activity stopped in the remaining motor and limbic fields after bilateral damage of the anterior limbic region or motor cortex or after application of GABA (Fig. 4). Damage to the posterior part of the limbic system does not hamper the display of the above-described changes in the EEG. The same results were obtained when connections between the anterior limbic region and motor cortex were cut bilaterally.
Fig. 4. EEG of the rabbit: a, after administration of sodium chloride; b,c, in different periods after application of GABA.
Assuming that the data on connections of the gyrus cinguli cortex with the archicortex (hippocampal formation) are satisfactory, we turned to the study of its interrelation with some regions of the neocortex (Sotnichenko, 1962, 1965). First the question arose about connections of the gyrus cinguli cortex with the motor field or in other words with the sigmoid gyrus cortex where, as mentioned above, there exists afferent representation of visceral analyzers. Employing the method of degeneration of axons with impregnation according to Bilshovsky, Gross and Nauta, we found the numerous degenerations but only in field 4 (posterior sigmoid gyrus) after damage to
299
C O N N E C T I O N S OF L I M B I C CORTEX
the anterior limbic region locatzd directly under field 6 (anterior sigmoid gyrus). Recurrent connection distribution shows a selective attitude of the 4th and anterior limbic fields, i.e. local damage of field 6 is restricted to fiber degeneration only within the limits of the cruciate sulcus while the damage of the posterior sigmoid gyrus (field 4) leads to fiber degeneration along a considerable part of the limbic region, almost along the entire length of the sulcus splenialis (Fig. 5). Moreover, a large
Fig. 5. The distribution of degenerated fibers on the medial surface of the cat’s brain afterthe extirpation of the sigmoid gyrus.
number of degenerated fibers appears in lower layers of the cortex after damage to the anterior region. The damage of the sigmoid gyrus mainly caused degeneration of surface fibers, horizontal fibers of the first layer of the cortex of the limbic region and especially of the cortex located in the cruciate sulcus. As a result it was established that the connections between these two regions are mutual but not equivalent since fibers from the anterior limbic region reach the motor cortex lower layers, but motor region fibers exervate the lamina molecularis of the gyrus cinguli. This inequality demands a physiological explanation. Still more unexpected was the discovery of a large number of degenerated deep (V and VI layers) fibers of the parietal and especially The degener: fibres
plen
s.sup
Fig. 6 . The distribution of degenerated fibers on the medial and lateral surface of the dog’s brain after lesion of the anterior limbic region. Sagittal section of the left hemisphere. References p . 303-304
300
E. SH. AIRAPETYANTS A N D T. S. SOTNICHENKO
of anterior regions of the occipital cortex (Fig. 6). When the number of degenerated fibers in parietal fields was reduced as compared with the motor field, a considerable increase in the number of fragmentary fibers of V-VI layers was observed in the visual cortex. The above-described distribution of degenerated fibers along the cortex of the lateral regions after damage of the anterior limbic region was noted for its great regularity in all experimental animals (cats and dogs). Cats in this respect are a still more obvious example because their deep fiber degeneration in the optic cortex embraces the entire optic cortex down to its very caudal parts. In a preliminary investigation massively degenerated pathways were found in fields 13, 14 (the angle formed by s. presylvii and s. rhinalis). It should be emphasized that stimulation of this region leads to visceral effects (Babkin and Kite, 1950;Kaada el al., 1949). Such degeneration was not discovered in parts close to the s. sylvii. Our comparative study of the structure of visceral analyzers, especially in regard to brain phylogenesis, naturally includes the limbic cortex, on which we experimented, not only in dogs and cats, but also in rats. In studying the brain of rats we meet qualitatively different categories of morphological formations. Feebleness of associative connections of the neocortical fields in rats occur in the structures that interest us. The white matter of the cortex of this rodent is poorly developed, small in size and consists almost entirely of projective and callosal fibers. The limbic cortex of the rat differs greatly from the cortical fields of the lateral surface, but not in the brain of dogs and cats. The cells of layer I1 have thick, bright palings and a thick felt-like plexus of fibers of layer I. That is why the limbic of the rat is nearer to the mesocortical than to the neocortical formations. Hypothetically and in general it may be said that the limbic acquired distinguishing features peculiar to the neocortex in the process of evolutionary development. The damage to motor and limbic regions was performed with an electrode inserted to a certain depth. Let it be emphasized that in rats the degeneration of surface fibers of the medial cortex was not found, which is contrary to the results obtained with cats and dogs. Numerous degenerations of fibers in all layers were observed in the immediate proximity to the place of the damage. Their direction was mostly horizontal. Their mainly occipital direction was clearly seen in sagittal sections. The damage to the anterior limbic field was of several types: (a) the damage embraced the limbic cortex and cingulum, (b) the cingulum was damaged locally while the cortex of the limbic region was preserved (the electrode broke into the limbic region from the side of the lateral surface), (c) only the gray substance of the anterior limbic region was damaged. There are very many fragmentary fibers in the limbic cortex in immediate proximity to the place of the damage and at some distance from it when the cingulum alone is damaged. The fibers leave the cingulum, go radially and terminate in layers V-VI and in layer I of the cortex (Fig. 7). The degeneration of fibers of the first layer of the cortex was observed only with this type of damage. With a different type of motor region or with coagulation of cells of the anterior limbic region, the first layer of the medial cortex was left completely untouched. Consequently these fibers of the medial cortex are projectional. Termination of the projectional fibers in the first layer of the cortex
CONN E C TI 0 NS 0EAI M BI C CORTEX The medial 8. part of cingulum \\,
-- - -
.
..
30 1
.
The damage
Fig. 7. The distribution of degenerated fibers on the rat’s cortex after damage to the medial part of the cingulum.
is quite likely the phenomenon specific to the mesocortical formations. Distribution of the fragmented fibers is wide here. They place themselves along the entire surface of the medial cortex as far as its transition to the lateral cortex, and with the damage of the cingulum by a single electrode near the genu corpus callosi they may be found at a considerable distance from the place of its interruption, i.e. in the posterior limbic fields. We shall not discuss the distribution and the nature of degeneration in the regio entorhinalis and presubiculum, since these phenomena were fully studied and described in the work of White (1957). He showed that the degeneration in the regio entorhinalis is connected with a breach in the medial part of the cingulum and is subcortical and projectional. We fully support this point of view. The data accumulated with an isolated injury to the gray substance of the anterior limbic region show that the cell axons of that part of the cortex distribute over the medial segment of the cingulum and penetrate into the limbic at a full length. They are numerous near the place of injury, and the single ones are met in the posterior limbic. We failed to find an obvious degeneration in the medial part of the cingulum as in the previous instance. Correspondingly there was no degeneration in the entorhinal region. There was a small number of fibers in the state of varicose in
the presubiculum.
Thus the intensive degeneration of the fibers in the hippocampal formation caused by damage to the anterior limbic region in rats is mainly ensured by the interruption of thalamus fibers found in the cingulum. In our most recent experiments we failed to find any degenerative processes in the hippocampal formation (entorhinal cortex, presubiculum hippocampus and fascia dentate) whereas the degeneration of fibers in the fields of the lateral surface and especially that of the optical fields was a regular phenomenon. Consequently according to the well-known scheme the switching of the gyrus cinguli into the ‘Papez circle’, i.e. the limbic system, is carried out through that cell station. References p . 303-304
302
E. SH. A I R A P E T Y A N T S A N D T. S. S O T N I C H E N K O
The above observations show that the limbic cortex and especially its anterior region is in direct efferent connection with many regions of the lateral cortex. These limbic efferents are selectively represented in various cortical fields. Though the connections are mutual the ones that go out of the limbic predominate. Spreading of this connection over into the visual cortex is of interest. According to our observations the limbic region is in a closer and more direct connection with the neocortex than with the hippocampal formation, which does not confirm one of the theses of the ‘limbic system’ about its inalienable and specific connection with the hippocampus. We also studied the role of the hippocampus in the structure and functions of viscera1 analyzers. Thus our results make it possible to confirm an important role of the limbic in the cortical visceral ensemble of afferent projections, in the formation of visceral temporal connections and in the functions of visceral analyzers. It should be emphasized that the generally-accepted conception of the ‘limbic system’ necessarily includes its close connection with the hippocampal formation as well as the deep subconsciousemotional influence exerted upon cortical activity, and also the relative independence of this functional unit of the neocortex. In our opinion the limbic is a part of the visceral cortex; it plays a special part in this structure, and it is naturally of vital importance in the structure of the visceral brain under certain conditions. The term ‘visceral brain’ is generally used as a synonym for the ‘limbic system’. There are no grounds for this morphophysiological merging, this identification of the ‘limbic system’ with the visceral brain. It is possible to justify a difference in theory about this phenomenon. The functional and morphological construction of the visceral brain embraces all elements of the complex of central apparatuses of visceral analyzers, both at the cortical and subcortical levels.
303
C O N N E C T I O N S O F LIMBIC CORTEX
REFERENCES ADAM,D., (1957); Rep. USSR. Acud. Sci., 3,113,709. (In Russian) AIRAPETYANTS, E. SH., (1940); Sci. trans. Univ. Leningrad, 59, 13. (In Russian) AIRAPETYANTS, E. SH., (1952); J. high. nerv. act. USSR. 2 (4), 481. (In Russian) AIRAPETYANTS, E. SH., (1959a); Investigations on cortical localization of internal signals. X X Z Znt. Congr. Physiol. Sci., Abstr., Buenos-Aires, p. 8. AIRAPETYANTS, E. SH., (1959b); Investigation of internal signal localization in the cortex functional structure of visceral analysers. Sci. inform. Pavlov Znst. Physiol. USSR Acad. Sci., 2, p. 9. (In Russian) AIRAPETYANTS, E. SH., (1963a); On the visceral cortex. XXZZ Znf. Congr. Physiol. Sci., Abstr., Leiden, II,45. AIRAPETYANTS, E. SH. (1963b); Foreshadowing of internal analysers by Sechenov and the state of current knowledge concerning them. Sechenov Physiol. J. USSR., 44, 1294. (In Russian) AIRAPETYANTS, E. SH., N. E. VASILEVSKAYA AND T. S. SOTNICHENKO, (1960); Transact. Pavlov Znst. Physiol., USSR. Acud. Sci., 9, 261. (In Russian) BABKIN,B. P., AND W. C. KITE,(1950); Central and reflex regulation of motility of pyloric antrum. J. Neurophysiol., 13, 321. BRADY,J. V., (1958); The Puleocortex and Behavioral Motivution. Biological and Biochemical Bases of Behavior. H. F. Harlow and C. N. Woolsey (Eds.), Univ. Wisconsin Press, p. 193. CHERNIGOVSKII, V. N., AND S. M. ZARAISKALA, (1962); Vicarious work of the vagus nerve in the cortex of brain hemispheres and in the limbic lobe of the brain in cats. Rep. USSR. Acud. Sci., 147, 742. (In Russian) DELOV, V. E., N. A. ADAMOVITCH AND A. N. BORGEST, (1961); Influence of afferent impulses from visceral receptors on electrical activity of the limbic cortex. Sechenov Physiol. J. USSR., 47, 1083. (In Russian) FULTON,J. F., (1953); The limbic system: a study of the visceral brain in primates and man. Yule J. Biol. Med., 26, 107. GAZA,N. K., (1962); In The mechanisms of cortico-visceral relations. Zvunovo USSR, p. 384 (In Russian) GASANOV, G. G., (1964); The role played by the posterior limbic cortex of the encephalon in interoceptive unconditioned reflexes of the stomach. Rep. USSR Acad. Sci., 159, 1427. (In Russian) GULYAEVA, L. N., (1962); In The mechanisms of cortico-visceral relations. Zvunovo USSR, p. 392. (In Russian) KAADA,B. K., (1960); Cingulate, posterior orbital, anterior insular and temporal pole cortex. Handbook of Physiol. S.Z., Neurophysiol., Vn, 1345. KAADA,B. R., K. H. -RAM AND J. A. EPSTEIN,(1949); Respiratory and vascular responses in monkeys from temporal pole, insula, orbital surface and cingulate gyrus. J. Neurophysiol., 12,346. KATO,S . K., KOGIAND H. KAWAMURA, (1960); Effect of inhalation anestheticson theelectrical activity of limbic system. X Z Ann. Meet. Jap. Electroencephalogr. Soc., 23. KENNARD,M., (1955); The cingulate gyrus in relation t o consciousness. J. nerv. and ment. Dis., 121, 34. KREMER, W. F., (1947); Autonomic and somatic reactions induced by stimulation of the cingular gyms in dogs. J. Neurophysiol., 10, 371. LOBANOVA, L. V., (1965); Sci. inform. Pavlov Znst., USSR. Acad. Sci., 3, 103. (In Russian) MACLEAN, P. D., (1949); Psychosomaticdisease and the ‘visceral brain’. J.psychosom. Med., 2,338. j uffhfigifmof Fortigo-visceral relations. Zvanovo USSR 61 1.
MvdiiyfiI M, All
(is@)
la The
I
(In Russian) MUSYASCHIKOVA, S. S., (1964); Rep. USSR Acud. Sci., 157, 1257. (In Russian). NAUTA,W. I., (1960); Some neural pathways related to the limbic system. Electrical studies of the unanesthetized brain. E. R. Ramey and D. S. 0’Doherty (Eds.), 4 , l . PAPEZ,J. W., (1937); A proposed mechanism of emotion. Arch. Neurol. Fsychiaf., 38,725. SMITH,W. K., (1944); The results of ablation of the cingular region of the cerebral cortex. Fed. Proc., 3,42. SOTNITCHENKO, T. S., (1962); Some morphological data on intracortical connexions of anterior and posterior limbic and motor areas in cat brain. Arch. Anut. Histol. Embryol., 43, 3. (In Russian, Summary in English) SOVETOV, A. N., (1963); The posterior limbic cortex and the interoceptive reflexes from the stomach and the intestine. Rep. USSR. Acad. Sci., 152,1012. (In Russian)
304
E. SH. AIRAPETYANTS A N D T. S. SOTNICHENKO
SOTNITSHENKO, T. S., (1965); Characteristics of the distribution of degenerated fibers over the cortex of median and lateral brain surface in dogs with destroyed anterior limbic and motor regions. Bull. Biol. Med. Expt. USSR, 3, 18. (Summary in English) TOKIZANE, T., H. KAWAMURA AND G. Imhim~, (1960); Hypothalamic activation upon electrical activities of paleo- and archicortex. Neurology, 2,63. TORE,S., AND H. KAWAMURA, (1960); Effects of amygdaloid stimulation on blood pressure and electrical activity of hippocampus. Jup. J. Physiol, 10,374. TOLMASSKAYA, E. S., (1964); The nervous mechanisms coordination somatic and visceral functions of the orgunism. Moscow Acad. Sci. USSR. (In Russian) VASILEVSKAYA, N. E., (1962); The mechanisms of cortico-visceral relations. Zvanovo USSR, 377. (In Russian) VASILEVSKAYA, N. E., (1963); Role of the reticular formation and of anterior limbic cortical fields in electrocorticographic manifestations of chemical changes in the internal environment. Sechenov Physiol. J. USSR, 49,293. (In Russian) VASILEVSKAYA,N. E., (1964); In m e nervous system. Leningrad Univ., 5, 73. (In Russian) N. E., AND G. P. UDAIDVA, (1965); The electric activity of brain cortex of rabbits VASILEVSKAYA, in the case of salt overload following destruction of motor and posterior limbic cortex area. Rep. Acad. Sci. USSR, 161,1238. (In Russian) WARD, A., (1948); The cingular gyrus: area 24. J. Neurophysiol, 11, 14. WHITE,L. E., (1959); Ipsilateral afferentsto the hippocampalformation in the albino rat. I. Cingulum projections. J. comp. Neurol., 113, 1. N. A,, (1966); The theses oftheIII. Middle Asiatic Conference of physiology, biochemistry, ZUBKOVA, and pharmacology, Dushambe, p. 154. (In Russian)
305
Effects of Hippocampal Ablation on Learning in the Rat * HIROAKI N I K I Department ofPsychology, College of General Education, University of Tokyo, Tokyo (Japan)
In a previous study (Niki, 1962), it was found that bilateral ablation of the hippocampus in rats resulted in an impairment of maze performance and a certain visual discrimination and delayed response. An increase in activity, and the increased resistance to extinction in a runway were also observed after the hippocampal ablation. The present study was designed to analyze further the nature of the hippocampal involvement in learning. In recent years increased attention has been paid to the neural mechanisms which play an important role in the inhibitory control of behavior (Diamond et al., 1963). In the present series of experiments, involvement of the hippocampus in the inhibitory control of behavior was investigated. GENERAL METHOD
Surgery. The animals were anesthetized with nembutal (40 mg/kg) and held in a stereotaxic instrument during the operation. After suitable openings were made in the skull, the dura was opened and the desired portion of the brain was removed by aspiration. In the experimental group the hippocampus was sucked bilaterally through the neocortex as much as possible, care being taken to spare the underlying thalamus. In the operative control group the neocortex overlying the hippocampus was removed bilaterally. After all bleeding ceased, the scalp was closed. All animals recieved intramuscular injections of penicillin for two days postoperatively. A recovery period of 3 or 4 weeks was allowed before testing. Histology. Following the termination of testing, the brains of the operated animals were fixed in 10% formalin, and then embedded in paraffin. Sections were cut 25 p thick, and every fifth section was mounted and stained with hematoxylin and eosin. Sections were examined for evidence of damage. Statistical test. The one-tailed Mann-Whitney U test (Siegel, 1956) was used throughout the study.
*
More complete reports will be published in Jup. Psychol. Res., 7, 1965, and 8, 1967.
References p. 316-317
306
H. N I K I Restoration of inhibited bar-press
I
90
Post-operative
Re-operative
h 2
3
Ext inctin
1
2
Test
Fig. 1. Median number of responses emitted during extinction and test.
Restoration of inhibited bar-press C
I
.-
b 90
" 80
2
Post-operative
Re-operative
70
2 60 5 % 50
-e
40
8 30
5 20
n
z
10
0
1
2 Extinction
1
Test
2
Fig. 2. Median number of responses to criterion during extinction and test. Restoration of inhibited bar-press
900. 800 -
Pre-operative
-
Post operative
0cc
700C
.P 600
b
E
*
500-
0 400c
f
300 200 loo 0
L
2 Extinction
3
1
Test
Fig. 3. Median time (sec)to criterion during extinction and test.
2
I
HIPPOCAMPAL ABLATION A N D LEARNING
307
E X PE R I MEN T I
A restoration of an extinguished bar pressing response fallowing the hippocampal ablation Sixteen male albino rats about 3 months old were trained to press the bar on the continuous reinforcement schedule until they obtained 90 reinforcements (30 trials per day for three days). After the training, the bar pressing response was extinguished. The extinction consisted of a daily 20-min period on three successive days. All animals were then subjected to operation. After the recovery period of 4 weeks, they were tested on the spontaneous recovery of the preoperatively extinguished bar pressing response for two days. The postoperative test procedure was the same as that used in the preoperative extinction. The results are presented in Figs. 1-3. The hippocampal animals showed a higher rate of bar pressing, required more responses and time in reaching the extinction criterion of 3 min without response than did the control animals in the postoperative test. These differences are all statistically significant (P< 0.01). Thus, hippocampal ablation resulted in ‘disinhibition’ of the previously extinguished response. This result is interpreted to mean that hippocampal ablation produced a loss of inhibition normally associated with nonreinforcement. E X PE R I MEN T I 1
Bar pressing under
SD
and SA
Seven animals with hippocampal lesions, 9 animals with neocortical destruction, and 10 sham-operated animals were used as subjects. The apparatus was the same Skinner box as used in EXPERIMENTI. After recovery from surgery the animals were trained to press the bar until a stable rate of bar pressing was shaped. After the completion of bar pressing training they received discrimination ti-aining for 12 days with a bright light serving as SD, and no light as S A . A continuous reinforcement schedule prevailed during S D , and bar pressing during S A was never reinforced. A daily experimental session consisted of 2-min S D - 4-min S A - 2-min SD. Fig. 4 represents the median number of bar presses performed by the three groups of animals during the daily SD and S A periods. The response rate during the S A period was higher in the hippocampal animals than in the other control animals although the rate of response during the SD period was similar for different groups. The higher response rates of the hippocampal animals during S A seem to reflect their inability to inhibit response to nonreinforced stimuli. These findings agree with those of Jarrard (1965) and Clark and Isaacson (1965), who found that hippocampectomized rats could not adjust their response to fit the schedule of reinforcement, and consistently pressed at higher rates than controls under variable interval and DRL. In order to evaluate the discriminatory behavior of the animals, a discrimination ratio was calculated for each animal. The discrimination ratio was defined as the References p. 316-317
308
H. N I K I
Bar--
under SDand
SA
Fig. 4. Median number of bar pressings duringthe SD and S A periods.
ratio of the response rate in S A over the response rate in SD. The median discrimination ratio of the last three days of discrimination training provided a single index of each animal’s discriminatory behavior. The group medians based on this measure were 1.06,0.25, and 0.20 for the hippocampal, cortical control, and sham-operated control group, respectively. It would appear that the impaired performance of the hippocampal animals on the discrimination was closely associated with the increased rates of S d responding without a concomitant increase in SD response rates. E X P E R I M E N T I11
A decline in runway performance due to the delay of reinforcement The animals were the same as those in EXPERIMENT‘II,now reduced in number to 7 in the cortical control group, and 7 in the sham-operated control group. The apparatus was a standard straight-alley runway. After 50 training trials (10 trials per day for 5 days) test trials were begun. Thirty test trials, 10 per day, were given with an intertrial interval of 10 min. The only change in procedure during the test was that the animals had to wait in the goal box for 30 sec before they were given access to food. Starting times (ST) and running times (RT) were measured on all training and test trials. The results are summarized in Fig. 5 where median starting and running times are plotted for successive days. As indicated in the figure there was no difference between the groups during training trials; during test trials, however, the animals in the hippo-
HIPPOCAMPAL ABLATION A N D LEARNING
.e--.
0-occ x---x
309
H
c
delay
1
2
3 4 5 Training test
1
immediate
2
3Days
delay
Fig. 5. Performance curves. The top graph gives starting time, and the bottom graph shows running time.
campal group had shorter ST and RT than the animals in the other two control groups. It is apparent that the runway performances of the hippocampal animals were less affected by the introduction of the delay of reinforcement. Hippocampal ablation, therefore, seems to interfere with the development of response inhibition associated with the delay of reinforcement. Considering the greater resistance to extinction of the hippocampally ablated animals in a runway situation (Niki, 1962; Jarrard et al., 1964), the present result is not surprising. EXPERIMENT IV
Position reversal learning
The 10 hippocampally ablated animals were compared with the 12 cortically ablated control animals in regard to their ability to reverse their acquired position responses. References p . 316-317
310
H. N I K I
The apparatus used in this experiment was a single unit enclosed T maze. Animals were trained on a position response for 10 triaIs per day until they reached the criterion of 2 consecutive errorless trials on 2 successive days. Upon reaching this criterion reversal learning began. The procedure during reversal learning was exactly the same as during initial learning except that the previously correct side was no longer reinforced but the previously incorrect side was. Reversal leaming was continued until the animals met the same criterion employed in the initial learning.
“LllL Initial position discrimination
I
7 rriais
5
O c c
Errors
5
0
0
Position reversal
40
10
0
Fig. 6. Mean number of trials and errors to criterion during position learning and position reversal learning.
The results are shown in Fig. 6. The hippocampal group required a greater number of trials to reach the reversal criterion (P < 0.005), and made more errors in reaching this criterion (P < 0.001), than did the cortical control group although there was no difference between the groups during initial learning with respect to both trials and errors. Thus, hippocampal ablation resulted in an impairment on position reversal learning but not on initial position learning. It is very likely that the deficits in position reversal learning following the hippocampal ablation are due to the increased resistance to extinction of previously acquired responses. A certain support for this statement comes from an analysis of ‘perseverance’ scores of the present data, where ‘perseverance’is defined as the number of trials to the first correct response in reversal learning. The hippocampal group took a mean of 7.2 trials, while the cortical control group required 3.7 trials. Consequently, hippocampal ablation seems to produce
HIPPOCAMPAL ABLATION A N D LEARNING
31 1
deficits in abandoning the previously acquired response tendency that is no longer adaptive. The present results confirm the recent findings by Mahut and Cordeau (1963) and Teitelbaum (1964) of successive reversal deficits in monkeys and cats with hippocampal damage. EXPERIMENT V
Single alternation learning
The animals were the same as those in EXPERIMENT IV, but reduced in number to 9 in each group. Using a modified Skinner box with two bars, the animals were trained to alternate between two bars. Having pressed the right bar and been reinforced, they could obtain a subsequent reinforcement by pressing the left bar and so on. All animals received 50 trials per day for 10 successive days. A trial was defined as one alternated bar pressing. Errors were counted whenever a response to one bar was followed by subsequent response to that same bar.
0-0
cc
Fig. 7. Performance curves on the two-bar alternation task.
Fig. 7 shows the performance curves for the two groups on the two-bar alternation problem. It is evident that the hippocampal animals performed more poorly than did the cortical control animals (P < 0.001). The median numbers of total errors were 1173 for the hippocampal group, and 556 for the control group. The difference between the groups was estimated to be significant (P < 0.001). These results can be interpreted to mean that hippocampal ablation produced a tendency for the animals to repeat rather than to alternate bar pressing responses. In fact, bursts of response on the same bar were frequently observed in the hippocampal animals. References p . 316-317
312
H. N I K l EXPERIMENT VI
Flexibility of solutions in a Dashiell maze Six animals with hippocampal lesions, 7 animals with neocortical destructions, and 8 sham-operated animals were tested in a Dashiell maze (Dashiell, 1930). Each animal received one trial per day for 18 days. Errors were counted whenever the animal entered a blind alley. As measures of 'flexibility' in maze learning, the number of different paths taken, and the number of shifts on consecutive days were scored. Flexibility
151
Number of shtft
Number of different path
1'
n
Errors
Fig. 8. Performance scores in a Dashiell maze.
The results are presented in Fig. 8. The hippocampal group showed a poorer maze performance than did each of the two control groups. Not only did the hippocampal animals make more errors, but they also showed less flexibility of solutions as revealed in the decreased number of different paths chosen and the decreased number of shifts on consecutive days. Hippocampal ablation, therefore, seems to decrease the plasticity of behavior as manifested in the increased response perseveration or, in other words, in the failure of adequate inhibitory control of behavior by nonreinforced events. Such a defect may be related to certain types of learning deficits found in our previous study. EXPERIMENT VII
Response alternation Twenty-eight male albino rats were randomly assigned to one of the three groups :(I) Hippocampal (N = 10); (11) Cortical control (N = 8); (111) Sham-operated control (N = 10). ( A ) Response alternation after two forced turns in the same maze
H I P P O C A M P A L ABLATION A N D L E A R N I N G
313
The animals were run one trial per day for two blocks of a 5 days’ session with 10 days intervened. On the first 5 days half the animals in each group were run on the .?-shaped maze, and on the last 5 days the L-shaped maze was used. For the other half the reverse order was utilized.
~
Response alternation
B
90
’40
V
30 20 10
Fig. 9. (A) Percentage of turn alternationsat the choice point after two forced turns in the same maze. (B) Percentage of response alternations after two forced trials in the Tmaze.
Percentages of turn alternations after two forced turns in the same maze based on the total 10 trials are shown in Fig. 9A. The animals in the two control groups showed an obvious tendency to avoid the preceding turns, while the hippocampal animals alternated at or even below a ‘chance’ rate. (B) Response alternation after two forced trials in a T maze On days 6-15 the animals were given three consecutive trials daily in the T maze with an intertrial interval of zero second. The first two of these trials were forced choices and the third was a test trial. The results are presented in Fig. 9B. In contrast to the control animals, all of which displayed a greater response alternation after two forced trials, the hippocampal animals were found to show far fewer alternation scores. These results indicate that the hippocampally ablated animals tended to persevere rather than to alternate after the forced turns. These findings are similar to those obtained from the. studies of ‘spontaneous alternation’ (Roberts et aE., 1962; Douglas and Isaacson, 1964). References p . 316-31 7
314
H. NIKI
EXPERIMENT VIII
Habituation of the cardiac response to sound stimulatim Animals were the same albino rats as used in EXPERIMENTS IV AND v, and were approximately 180 days of age at the beginning of this experiment. Nine hippocampal and 11 cortical control animals were utilized. The heart rate was recorded from two silver wire loops (0.2 mm in diameter) permanently inserted in the animal's skin just above the right and left forepaws. An electroencephalograph was used to record heart rate. Being adapted to the experimental situation, the animals received 100 trials of 3-sec sound stimulation with an intertrial interval of 60 to 90 sec. The resting heart rate of animals adapted to the experimental situation varied from 280 to 420 beats per min. No significant differences in the resting heart rate between the control and hippocampal animals were observed. The number of heart beats was counted: in each of three consecutive 3-sec intervals for each trial : 3-sec pre-tone, 3-sec during tone, and 3-sec post-tone. This yielded two measures of change in heart rate, (a) tone minus pre-tone and (b) post-tone minus pre-tone.
Habituation of EKG
5:
c
p 1 . zf
.
*\ \\
\
\ '0 '\
0
1-10
H
0---0
cc
\.
0-0-0
0,
0-0
\
'
\
' 6 6 6 0 1e- e11-20 21-30 31-40 41-50 51-60 61-70 71-60 61-90 91-xx)Trials
Fig. 10. Median number of cardiac responses (more than 8 beats/minchange)throughouthabituation shown in 10-trial blocks.
The results are presented in Fig. 10. The rate of habituation of the heart rate change to sound stimulation was slower in the hippocampal group than in the control group. In this experiment the heart rate change to the tone was found to be a decrease in heart rate. Deficits in habituation following the hippocampal ablation are consistent with the hypothesis that the hippocampus is important for the inhibitory control of behavior.
H I P P O C A M P A L ABLATION A N D L E A R N I N G
315
ANATOMICAL RESULTS
Fig. 11 presents a diagrammatic summary of the typical brain lesions for the hippocampal and the cortical control groups. The extent of hippocampal damage varied Diagram of lesions
Hippocampal
Neocortical
Fig. 11. Summary diagrams of lesions.
from about 25% to 75%. The hippocampus was never completely ablated, but the middle portion of the hippocampus was ablated for all experimental animals. The most rostral and ventral portions were spared. Damage in the cortical contIol group was largely restricted to the dorsal neocortex. The extent of cortical damage in the control group was consistently greater than in the hippocampal group. GENERAL D I S C U S S I O N
The results of this study are as follows. (1) Hippocampal ablation resulted in ‘disinhibition’ of previously-extinguishedbar pressing responses; References p. 316-31 7
316
H. N I K I
(2) Hippocampal ablation increased the rate of responding to nonreinforced stimuli without concomitant increases in the response rate to reinforced stimuli ; (3) The runway performances of the. hippocampally ablated animals were less affected by the delay of reinforcement; (4) Hippocampal ablation resulted in an impairment on position reversal learning but not on initial position learning; ( 5 ) Hippocampal abIation produced a severe deficit in a single alternation learning. (6) The hippocampally ablated animals showed less flexibility of solutions in a Dashiell maze; (7) The hippocampally ablated animals tended to persevere rather than to alternate after the forced turns; (8) The rate of habituation of the cardiac response to sound stimulation was slower in the hippocampally ablated animals than in the operated control animals. All these findings suggest that the hippocampus plays an important role in the inhibitory control of behavior. Similar evidence regarding the inhibitory control has been obtained from the studies of passive avoidance (Kimura, 1958; Isaacson and Wickelgren, 1962; Kimble, 1963). Other neural structures such as the amygdala (Brutkowskiet al., 1960; Schwartzbaumet al., 1964), the septal area (McCleary, 1961; Kaada et al., 1962), and the frontal cortex (Brutkowski, 1959; Butter et al., 1963),have been implicated in the behavioral inhibition. It is possible that these neural structures as well as the hippocampus form a common diffuse system which exerts an inhibitory control over the behavior. It would be interesting, therefore, to differentiate the effect of each lesion in these structures. REFERENCES BRUTKOWSKI, S., (1959); Comparison of classical and instrumental alimentary conditioned reflexes following bilateral prefrontal lobectomies in dogs. Acta Biol. exp., 19,291-299. BRUTKOWSKI, S., FONBERG, E., AND MEMPEL,E., (1960); Alimentary typeII (instrumental)conditioned reflexes in amygdala dogs. Acta Biol. exp., 20,263-271. BUTTER,C. M., MSHKIN, M., AND ROSVOLD, H. E., (1963); Conditioning and extinction of a foodrewarded response after selective ablations of frontal cortex in rhesus monkeys. Exp. Neurol., 7, 65-75
CLARK,C. V. H., AND ISAACSON,R. L., (1965); Effect of bilateral hippocampal ablation on DRL performance. J. comp. physiol. Psychol., 59,137-140. DASHIELL, J. F., (1930); Direction orientation in maze running by the white rat. Comp. PsychoZ. Monogr. 7.
DIAMOND, S., B A L . ~R., S., AND DIAMOND, F. D., (1963); Inhibition and Choice: A Neurobehavioral Approach to Problems of Plasticity in Behavior. New York, Harper and Rowe. DOUGLAS,R. J., AND ISAACSON, R. L., (1964); Hippocampal lesions and activity. Psychon. Sci., 1, 187-188.
ISAACSON,R. L., AM) WICKELGREN, W. O., (1962); Hippocampal ablation and passive avoidance. Science, 138, 1104-1 106. JARRARD, L. O., (1965); Hippocampal ablation and operant behavior in the rat. Psychon. Sci., 2, 115-1 16.
JARRARD,L. O., ISAACSON, R. L., AND WICKELGREN, W. O., (1964); Effects of hippocampal ablation and intertrial interval on runway acquisition and extinction. J. comp. physiol. Psychol., 57,442-444. KAADA, B. R., RASMUSSEN, E. W., AND KVEIM,O., (1962); Impaired acquisition of passive avoidance behavior by subcallosal,septal, hypothalamic, and insular lesions in rats. J . comp.physio1. Psychol., 55,661-670.
HIPPOCAMPAL ABLATION A N D LEARNING
317
KIMBLE, D. P., (1963); The effects of bilateral hippocampal lesions in rats. J. comp. physiol. Psychol., 56, 273-283. KIMURA,D., (1958); Effects of selective hippocampal damage on avoidance behavior of the rat. CU~U J . ~Psycho/., . 12, 213-217. MAHUT,H., AND CORDEAU, J. P., (1963); Spatial reversal deficit in monkeys with amygdalohippocampal ablations. Exp. Neuvol., 7, 426-434. MCCLEARY, R. A. (1961); Response specificity in the behavioral effects of limbic system lesions in the cat. J . comp. physiol. Psychol., 54, 605-613. NIKI,H., (1962); The effects of hippocampal ablation on the behavior in the rat. Jup. Psychol. Res., 4, 139-153. NIKI,H., (1965); The effects of hippocampal ablation on the inhibitory control of operant behavior in the rat. Jup. Psychiol. Res., 7, 126-137. NIKI,H . , (1966); Response perseveration following the hippocampal ablation in the rat. Jup. Psychol. Res., 8 1-9. ROBERTS, W. W., DEMBER, W. N., AND BRODWICK, M., (1962); Alternation and exploration in rats with hippocampal lesions. J. comp. physiol. Psychol., 55, 695-700. SCHWARTZBAUM, J . S., THOMPSON, J. B., AND KELLICUTT, M. H., (1964); Auditory frequency discrimination and generalization following lesions of the amygdaloid area in rats. J . comp. physiol. Psychol., 57, 257-266. SIEGEL, S., (1956); Nonparametvic Statistics for the Behavioral Sciences. New York, McGraw-Hill. TEITELBAUM, H., (1964); A comparison of effects of orbitofrontal and hippocampal lesions upon discrimination learning and reversal in the cat. Exp. Neurol., 9, 452-462.
318
The Limbic Systems, Efferent Control of Neural Inhibition and Behavior K A R L H. P R I B R A M Stanford University, School of Medicine, Palo Alto, Calif. (U.S.A.)
The functions of the limbic forebrain have been characterized either in terms of their influence on response regulation (McCleary, 1961; Miller et al., 1960), on emotion (MacLean, 1950; Pribram and Kruger, 19541, or on memory (Milner, 1958; Penfield and Milner, 1958). In and of themselves these characterizations have so far failed to provide the key to the essential nature of the limbic contribution to behavior and to psychological experience. In part, this failure can be attributed to an absence of precision in the concepts invoked. With this in mind the question has been raised whether a clearer picture might be obtained if a possible relationship between limbic function and information processing were pursued. Perhaps with this relationship worked out, the puzzle of the importance of the limbic systems to response regulation, to emotion and to memory will also come into focus. To this end a series of neurophysiological and neurobehavioral experiments were undertaken. The results of the neurophysiological experiments were such that they suggested a model of limbic system function. I shall first introduce the model, then present the neurophysiological and behavioral data which generated it. Finally, I will attempt to summarize the model in its current form. SOME NEUROPHYSIOLOGICAL D A T A AND A M O D E L
The model regards inhibitory neural processes - inhibition defined as a reduction in the excitation of a neural unit. Two major types of neural inhibition are recognized. The first is inherent in afferent activity : active afferent neurons inhibit their neighbors. This lateral or surround inhibition operates through collateral processes distributed among neurons or via amacrine-like cells and is well-demonstrated in the visual (Hartline et al., 1956), auditory (Von Bekesy, 1957), and somatic (Mountcastle, 1957) systems, both at peripheral and central stations. This type of afferent neural interaction corresponds to Pavlov’s ‘external’ inhibition. The second type of afferent inhibition is recurrent (Asanuma and Brooks, 1965; Brooks and Asanuma, 1965a, b). It takes two forms, presynaptic and postsynaptic. Both result in the regulation of afferent activity via negative feedback. In the case of postsynaptic recurrent inhibition,
EFFERENT CO NT RO L OF N E U R A L I N H I B I T I O N A N D BEHAVIOR
319
interneurons of the Renshaw type are assumed via recurrent inhibitory fibers, to function as dampers which control the excitability of active neurons as a consequence of their own activity. The model focuses on the interaction of these two forms of afferent inhibition. The collateral type acts to accentuate the difference between active and less active sites while the recurrent type tends to equalize such differences. Any patterned change in the system will be enhanced by collateral inhibition ; recurrent inhibition works against change, tending to stabilize the status quo. The collateral type is thus conceived to be a labile mechanism sensitive to input and concurrent activity. The recurrent type, on the other hand, works more slowly, countering the rapid fluctuations in the patterns of neural activity that would otherwise occur and stabilizing the changes once they have occurred. The chief concern of the model is with efferent control exerted over this interaction. This control is primarily cerebrofugal. Mechanisms which enhance and inhibit afferent inhibition are assumed to converge upon the afferent pathways. Because of this site of operation, a 4-fold mechanism of efferent or cerebrofugal control should in theory be distinguishable: (a) enhancement of collateral inhibition ; (b) enhancement of recurrent inhibition; (c) inhibition of collateral inhibition; and (d) inhibition of recurrent inhibition. There is already available evidence for corticofugal control over both the presynaptic and postsynaptic forms of recurrent inhibition. Repetitive stimulation of a variety Visual Cortex Recovery Cycle
-
25 rn sec
Wrn sec 100 m sec
Fig. 1 . A representativerecord of the change produced in visual evoked responses by chronic stimulation oftheinferotemporalcortex. Upper set of recordswas taken before stimulation;lower set, during stimulation.All traces were recorded from the visual cortex; in the first column areresponsesproduced by a pair of flashes separated by 25 msec; flash separation is 50 msec in the second column and 100 msec in the thud. References p. 335-3361
320
K. H. PRIBRAM
of sensory-motor points on the lateral cortex influences presynaptic inhibition at the spinal level (Andersen et a!., 1962; Andersen and Eccles, 1962; Eccles, 1962).And the effect of hippocampal stimulation on visual evoked activity has also been recorded (Fox, 1966). The evidence for efferent control of collateral inhibition has been gathered ia my own laboratory, in collaboration with Dr. D. N. Spinelli (Spinelli and Pribram, 1966). I will present these studies in detail and then continue to develop the model which is so extensively rooted in these data. Experiments were performed on fully awake monkeys implanted with small bipolar electrodes and a device which allows chronic repetitive stimulation of one of the electrode sites. The monkeys were presented with pairs of flashes and the interflash interval was varied from 25 to 200 msec. Electrical responses were recorded from the striate cortex and the amplitude of the responses was measured. A comparison of the amplitude of the second to the first response of each pair was expressed and plotted as a function. The assumption underlying the interpretation of t h i s function is that when the amplitude of the second of the pair of responses approximates that of the first, the responding cells have fully recovered their excitability. In populations of cells such as those from which these records are made, the percent diminution of amplitude of the second response is used as an index of recovery of the total population of cells -thus the smaller the percent, the fewer the number of recovered cells in the system. Chonic stimulation (8-l0/sec) of several cerebral sites alters this recovery function. When the inferotemporal cortex of monkeys is stimulated, recovery is delayed (Figs. 1,
I
110 100 -
90 80 -
+ ~ 7
p
0 -
215050 -
40 30 20 -
''0
M Before
stimulation --O Shortly after sti&ulation 0--0 I wk. after o n s t stimulation 8-43 I mo. #or ongot stirnulorion
7 25 50 75 100
rnsec
I50
200
Fig. 2. A plot of the recovery functions obtainedin one monkey Wore and during chronic stimulation of the infbrotemporal (I."IT) cortex.
E F FE R E NT C O N T R O L O F N E U R A L I N H I B I T I O N A N D B E H A V I O R 120-
-
I
I
25
50
I
1
75
100
321
-
110100 -
90
.-
80
.-
$10-
u
26050 40 30 -
20 --
m~ e c
150
200
Fig. 3. A plot of the recovery functions obtained in 5 monkeys before and during chronic cortical
stimulation.
2s
50
75
IW MSEC
I50
261
rig. 9. I ne eiieci U I enrunic StirnuiaiiuIi ui tne a~iiypuaiuiucuiiipiex UII i e c ~ v t x yiuIictiuIi.
i ~ i uutieu c
line indicates the function before, the solid line after, 1 month of stimulation. Bars perpendicular to the curves show variability among subjects. Each curve is based on the average response of 4 subjects.
2, 3). Stimulation from control sites (precentral and parietal) has no such effect. Nor does the stimulation of inferotemporal cortex alter auditory recovery functions. These, however, can be changed by manipulations of the insular-temporal cortex, as was shown in a parallel experiment performed on cats. Here the crucial cortex was removed and recovery functions obtained on responses recorded from the cochlear Rrfircnccs p . 335-336
322
K. H. P R I B R A M
50
25
100 MSEC
75
150
200
Fig. 5. This figure represents the same data as those represented in Fig. 4.However, here %change in recovery is plotted. Shaded area indicates range of recovery for unstimulated subjects.
L2I5
1
50
1
75
I
100
msec
I
125
1
150
1
175
I
200
Fig. 6 . This figure plots the % change in recovery for all subjects in the various experiments. It is thus a summary statement of the findings. Fig. 7. This figure is made up of photographs of a pulse histogram derived from a readout from a computer for average transients. Each vertical segment represents the number of impulses recorded from a neural unit during a 1.24 msec period. The upper three traces show the effects of concurrent stimulation of the frontal (fr.), the bottom three traces the effects of concurrent stimulation of the temporal (it)., cortex of cats on the unit activity evoked in the striate cortex to repeated fla5hes (f.). The first and last trace in each trial are controls; the middle traces were recorded during concurrent stimulation. Note that the first silent period is lengthened by concurrent temporal, and shortened by concurrent frontal, cortex stimulation.
EFFERENT CONTROL OF NEURAL INHIBITION A N D BEHAVIOR
fr.l!
f.
l! References p . 335-336
'
100 msec
'
1
323
324
K. H. P R I B R A M
nucleus (Dewson 111 et al., 1966). Removal of insubtemporal cortex shortens recovery in the auditory system. A great many neurobehavioral experiments have shown the importance of these isocortical temporal lobe areas (and not others) to visual and to auditory discrimination. These studies are reviewed elsewhere (Pribram, 1954, 1966). What concerns us here is that a corticofugal, efferent mechanism is demonstrated and that this mechanism alters the rapidity with which cells in the visual and auditory afferent systems recover their excitability. Further, since stimulation delays and ablation speeds up recovery, the inference is that the normally afferent inhibitory processes which delay recovery are enhanced by the ordinary operation of these temporal lobe isocortical areas. But the opposite effect narqely inhibition of afferent inhibition - can also be obtained when cerebral tissue is chronically stimulated. In these experiments the cortex of the frontal lobe and the cortical nucleus of the amygdala were chronically stimulated and recovery of cells in the visual system were shown to be speeded. This result has suggested that the frontal and anterior medio-basal portions of the forebrain function as efferent systems which inhibit afferent inhibitory processes (Figs. 4, 5, 6). The antagonistic effect of these two efferent control systems is best illustrated by data obtained at the unit level. These unit recordings were made from the striate cortex of flaxidilized cats to whom flashes of light were presented. Note that the silent period of a cell can be lengthened by concurrent inferotemporal stimulation.Notealso that concurrentfrontal stimulation can shorten this d e n t period. Finally, note the unit whose silent period is lengthened by inferotemporal, and shortened by frontal, stimulation (Figs. 7,8). In summary, the model is based on neurophysiological evidence of two forms of afferent inhibition : collateral and recurrent. The reciprocal interaction of these two forms is spelled out. Data are presented which indicate that afferent inhibition is under efferent corticofugal control. Further, such efferent control is shown to be balanced : both efferent enhancement and efferent inhibition of afferent inhibition were found to converge so as to regulate the activity of a single system and even a single cell. The major assumption of the model is that separate forebrain systems can be found to regulate collateral and recurrent afferent neural inhibition. One of the consequences of this model of efferentcontrol over afferent inhibition is a plausible neural explanation of the orienting reaction and its habituation. A series of studies has shown (1) that orienting can be identilied by a specificpattern of behavioral and physiological indices; and (2)that habituation of this set of indices is not a function of a raised neural threshold to input, but to change in some neural configuration against which input is matched (Sokolov, 1960). The reasonable suggestion can be made that habituation reflects incrementsin recurrent inhibition and that the orienting reaction manifests an override on habituation which takes place whenever collateral inhibition is enhanced. There is at least preliminary evidence at the neurophysiological level which is congruent with this suggestion (Thompson, 1966). The following data at the neurobehavioral level can also be interpreted to be in accord with the model.
-
EFFERENT CONTROL OF NEURAL I N H I B I T I O N AND BEHAVIOR
325
f.
f.
f. msec Fig. 8. A pulse histogram obtained in the same fashion as that reproduced in Fig. 7. Here the influence of concurrent temporal (it.) (2nd trace) and concurrent frontal (fr.) (4th trace) cortical stimulation on the flash (f.) evoked activity of the same single unit is shown. Note that the first silent period is lengthened by concurrent frontal, and shortened by concurrent temporal, cortex stimulation.
THE R E G U L A T I O N OF O R I E N T I N G A N D T H E A M Y G D A L O I D C O M P L E X
Bilateral amygdalectomy (Fig. 9) interferes drastically with the orienting reaction as gauged by the galvanic skin response (G. S. R., Bayshaw et al., 1965; Kimble et al., References p. 335-336
326
K. H. PRIBRAM
x-222
X-2W
EFFERENT CONTROL OF NEURAL I N H I B I T I O N AND BEHAVIOR
327
I............... x H 0 .............. e IT .............. LF
ao 0
..
60 AVG PERCENT
GSR
40
;. .............
0 ’
1
2
3 10 TRIAL
.
BLOCKS
6
0
5
Fig. 10. Curvesof average (AVG) % GSR response to the first 30 presentations of the original sthulus for the normal and three experimental groups.
1965; Koepke and Pribram, 1966. See Fig. 10). However, the behavioral effect of this interference is not simple. In a variety of discrimination learning tasks, some of which amygdalectomized monkeys found more difficult than their controls, a behavioral measure of orienting was taken (Bateson, submitted to Science). This measure consisted of noting the flick of the monkey’s ears during the time the cues were presented. Normal monkeys show this flick of theears while they are learning; once a task has been mastered this ear response no longer occurs. Amygdalectomized monkeys show a longer total time during which such ear flicks occur, especially in those tasks in which they showed impairment. These results led to the idea that orienting was made up of two components - one an alerting reaction indicated by the ear flick, the other a focusing function which allowed registration of the event which produced the alerting. In is this second stage which involves the amygdala and is signalled by the appearance of a GSR. The two phases of orienting fit the model presented. The first phase, alerting, can be explained as a consequence of initial disinhibition of collateral inhibition. In the absence of a secondary controlling mechanism this reaction would overcome the stabilizing mechanism provided by recurrent inhibition. Events would continually be noticed but adjustment of the stabilizing mechanism (habituation) precluded. This is believed to be the case after amygdalectomy. By contrast, in normal subjects, collateral inhibition is in turn inhibited by the operation of the amygdaloid mechanism. This provides the reaction with a stop mechanism which increases the likelihood that its specific Fig. 9. Reconstructions of the bilateral lesions of the amygdaloid complex. Black areas denote the lesion. References p . 335-336
328
K. H. P R I B R A M
Fig. 11. Mean percentage changes in total responses of test sessions which followed prolonged deprivation of food. The values in the legend refer t o the range of total responses for the three preceding control sessions on which the percentage changes are based.
SMALL REWARD
-Normals
128-558 c-----Amygdalas 476-581
m
E
20
c
1
2
3
4
5
Session
6
Fig. 12. Mean percentage changes in total responses in test sessions with the small reward. (The values in the upper right refer to the range of total responses for the three preceding large-reward sessions on which the percentage changes are based; group differences are not statistically significant).
0
P . 6 140-
sQ 0-2 e
-
LARGE /.,
\:
.____.__,-a *_.. _...
0-
tE 100%
.-
80-
REWARD
.._
.'../
..
,.%.
o-oNorrnals
o--oArnygdalas
Fig. 13. Mean percentage changes in total responses within test sessions with the large reward.
EFFERENT C O N T R O L OF N E U R A L I N H I B I T I O N A N D B E H A V I O R
329
configuration will be stabilized, i.e. registered. A difficulty in registering events should show up behaviorally in a variety of ways. One of them certainly would be in direct measures of habituation. Short term measures should show an increased speed of habituation; on the other hand, longer term measures should show that such habituation had failed to incorporate the orienting experience. This is exactly what has been found (Schwartzbaum, 1964). Another consequence of difficulty in registration would be the relative inefficacy of reinforcement. And, indeed, a series of experiments has shown that changing the amount of reward or its size (Schwartzbaum, 1960a, b) or the distribution of its occurrence (Schwartzbaum, 1961), has considerably less effects on amygdalectomized monkeys than on their controls (Figs. 11, 12, 13). THE HIPPOCAMPAL FORMATION A N D HABITUATION
Douglas (Douglas and Pribram, 1966) formulated in precise behavioral terms a theory that I have taken the liberty of incorporating into my model. He suggested that the amygdala system operates as a reinforce-register mechanism and that the hippocampal formation serves to evaluate error. Several ingenious experiments were devised to test hypotheses derived from the theory. I shall present three of these. All were performed in an automated discrimination apparatus which allowed programming of tasks by a special purpose computer which could also be used for data reduction and analysis (Pribram et al., 1962; see Figs.). Douglas modified a standard behavioral testing procedure to his purpose. The procedure is called probability matching and in it subjects are trained to discriminate
Fig. 14. Display panel of the automated discrimination apparatus. Note 16 clear hinged windows through which patterns can be displayed, and central tray attached to feeder mechanism. References p.1335-336
330
K. H. PRIBRAM
between two cues. Ordinarily one cue is rewarded 100% of the time and the other is never rewarded. In a probability matching task, however, one cue is rewarded some percentage less than 100 - say 70% - while the other cue is rewarded on the remaining occasions - in this instance, 30 % of the time. This task is, of course, more difficult than the ordinary discrimination. The probability test is more interesting, however, since different organisms demonstrate different strategies in solving the problem. Douglas trained monkeys (bilaterally amygdalectomized, hippocampectomized and sham operated controls) in such a probability matching situation and then paired a novel cue with either the most- or the least-rewarded of the familiar cues. His results were striking.
Fig. 15. Control console and special purpose computer for the automated discrimination apparatus which allowsprogrammingoftasks as well as data reduction and analysis. Fig. 16. Reconstructions of the bilateral lesions of the hippocampus. Note that in this figure dashed areas on the reconstructions denote the lesion, black areas denote sparing. Dotted areas show the overlying cortex removed in the approach. Heavy lines on the cross-sections show the extent of the lesion on the ventral surface.
EFFERENT CONTROL OF N E U R A L I N H I B I T I O N A N D BEHAVIOR
45
References p. 335-336
69
331
332
K. H. P R I B R A M
i4 LEARNING WITH
70-30 SCHEDULE
5 20 W V rn
I
W n
2
3
4
40-TRIAL 5 6 7 BLOCKS 8 9
1011
I21314
Fig. 17. Percent choice of most rewarded cue in probability task involving learning with a 70-30 schedule.
n
80
R z
a 6 NOVEL VS MOST REWARDED CUE
N=
6
p 4
NOVEL VS LEAST REWARDED CUE
Fig. 18. Percent responseto novel stimulus in groups of novel vs. most rewarded cue, compared with groupspf novel vs. least rewarded cue.
First, monkeys with hippocampal lesions learned the probability task more slowly than did the other groups. This slower learning is interpreted as consonant with an impaired error-evaluate system in the hippocampectomized monkeys (cJ:Figs. 16, 17). Second, monkeys with hippocampectomies,when compared with the other groups, chose the familiar cue more often when this was paired with a novel cue, irrespective of whether that familiar cue had been reinforced on 70 % or 30 % of the trials. The choice of the familiar is also consonant with an intact reinforce-register function and an impaired error-evaluate mechanism (Fig. 18). Finally, the cues used in the probability matching task were again presented, this time without reinforcement. As could be predicted, control subjects quickly shifted their responses away from the previously rewarded cue since these responses were now erroneous. And again, hippocampally ablated monkeys came to the support of the theory by failing to shift their responses on the basis of error (Fig. 19).
EFFERENT C O N T R O L OF N E U R A L I N H I B I T I O N A N D BEHAVIOR
333
As already noted, the behavioral process invoked to explain these results is an errorevaluate mechanism. On the basis of the model and data presented, the hippocampus is suggested to provide this mechanism. By inhibiting recurrent inhibition, the erroneous experience is allowed to register. In the absence of the hippocampus, the stabilizing effect of recurrent inhibition is assumed to be sufficiently strong to overcome the registration of nuances: the system of afferent inhibitory processes tends to revert to the status quo ante. This hyperstability is overcome only if the orienting events are overwhelming or if they recur regularly. Probabilistic occurrences, such as errors, fail to ‘get through‘. According to this view, short term habituation should be slowed by hippocampectomy and registration limited to regularly recurring events. There is evidence in support of both of these statements (Douglas and Isaacson, 1964; Roberts et al., 1962; Kimble, 1963). SUMMARY A N D CONCLUSION
The model is now complete. Collateral and recurrent afferent inhibition are bucked against one another, forming a primary couplet of neural inhibition within afferent channels. Four forebrain mechanisms are assumed to provide efferent control on this primary couplet: Frontotemporal
/
Sensory specific-intrinsic
inhibition Self inhibition \
Polysensory-motor
Two of these, frontotemporal and sensory specific-intrinsic (which includes the inferotemporal cortex), work their influence by regulating collateral inhibition ; two others, hippocampal and polysensory-motor, regulate recurrent inhibition. The sensory specific-intrinsic and polysensory-motor ‘association’ cortical systems exert their control by enhancing, while the frontotemporal and hippocampal systems exert control by inhibiting afferent neural inhibition. According to the model, orienting is a function of changes in the pattern of collateral inhibition ;habituation is due to recurrent inhibition elicited in response to this changed pattern. Registration of experience is a function of habituation. Complex problemsolving is dependent on evaluating erroneous experiences : these, because they are Rejerences p. 335-336
334
K. H. PRIBRAM
PFOS
M
N=4
STIMULUS WOKE, WRING EXTINCFi
-2-1
I
2
3
4
5
6
40-TRIAL
7
8 9 BLOCK
-
10 I I
12 I3
14
Fig. 19. Percent choice of most rewarded cue in 40-trial block, where cues used in the probability matching task were again presented, this time without reinforcement.
nuances, can only be registered if the hyperstabilizing influence of recurrent inhibition is in turn inhibited. Data were presented to show that the amygdaloid complex regulates the orienting reaction. These data were interpreted, according to the model, as indicating that the amygdala ordinarily originates a process of efferent inhibition of collateral inhibition. This process stops the change in pattern of collateral inhibition from proceeding at such a rate, and to such an extent, as to preclude revision of the pattern of recurrent inhibition. Registration occurs only when such revision has taken place. Data were also presented to show that the hippocampal formation is involved in the handling of erroneous experience. A case was made to the effect that ordinarily negative instances of experience must be evaluated. According to the model, the hippocampus, by efferently inhibiting recurrent inhibition, ordinarily provides a mechanism for allowing the registration of nuances. Here again, revision of the pattern of recurrent inhibition is dependent on an efferent inhibitory mechanism. In this instance, however, efferent inhibition overcomes the tendency of recurrent inhibition toward hyperstabilitY. This neurological model of information processing has helped me considerably in understanding the wealth of neurophysiological and neuropsychological data available. Some of this understanding has been brought to bear here on the problem of limbic system function. Aside from further tests of the model, the job ahead is to devise experimentswhich will allow extension of the model to such problems as the regulation of action, memory and emotion. ACKNOWLEDGEMENT
This research was supported by U.S. Public Health (NIMH) Grant MH-03732 and U.S. Army Contract DA-49-193-MD-2328.
EFFERENT C O N T R O L OF N E U R A L I N H I B I T I O N A N D B E H A V I O R
335
REFERENCES ANDERSEN, P., AND ECCLES,Sm J., (1962); Inhibitory phasing of neuronal discharge. Nature, 1%, 645-647. ANDERSEN, P., ECCLES, SIRJ., AND SEARS,T. A., (1962); Presynaptic inhibitory action of cerebral cortex on the spinal cord. Nature, 194,740-741. ASANUMA, H., AND BROOKS, V. B., (1965); Recurrent cortical effects following stimulation of internal capsule. Arch. ital. Biol., 103, 22&246. BAGSHAW, MURIELH., KIMBLE,D. P., AND PRIBRAM,K. H., (1965); The GSR of monkeys during orienting and habituation and after ablation of the amygdala, hippocampus and inferotemporal cortex. Neuropsychologia, 3, 11 1-1 19. BATESON, PATRICK, (1966); submitted to Science. VONBEKESY, G., (1957); Neural volleys and the similarity between some sensations produced by tones and by skin vibrations. J. acoust. Soc. Amer., 29, 1059-1069. BROOKS, V. B., AND ASANUMA, H., (1966);Recurrent corticaleffects following stimulation of medullary pyramid. Arch. ital. Biol,, 103, 247-278. BROOKS, V. B., AND ASANUMA, H., (1965b); Pharmacological studies of recurrent cortical inhibition and facilitation. Amer. J. Physiol., 208, 674681. DEWSON 111, J. H., NOBEL,K., AND PRIBRAM, K. H., (1966); Corticofugal influence at cochlear nucleus of the cat: some effects of ablation of insular-temporal cortex. Brain Research, 2, 151-159. DOUGLAS, R., AND ISAACSON, R., (1964); Hippocampal lesions and activity. Psychon. Sci.,1,187-188. ECCLES,SIRJ., (1962); Inhibitory controls on the flow of sensory information in the nervous system. Information Processing in the Nervous System. Proceedings of the International Union of Physiological Sciences, XXII International Congress, Leiden, pp. 22-45 Fox, S. S., This volume, p. 254. HARTLINE, H. K., WAGNER, H. O., AND RATLIFF, F., (1956); Inhibition in the eye of limulus. J. gen. Physiol., 39, 651-673. KIMBLE, D. P., (1963); The effects of bilateral hippocampal lesions in rats. J. comp.physiol. Psychol., 56,273-283. KIMBLE,D. P., BAGSHAW, MURIELH., AND PRIBRAM, K. P., (1965); The GSR of monkeys during orienting and habituation after selective partial ablations of the cingulate and frontal cortex. Neuropsychologia, 3, 121-1 28. KOEPKE, JEAN, E., AND -RAM, K. H., (1966); Habituation of GSR as a function of stimulus duration and spontaneous activity. J. comp. physiol. Psychol., 3, 4424t8. MACLEAN, P. D., (1950);Psychosomaticdisease and the ‘visceralbrain’; Recent developments bearing on the Papes theory of emotion. Psychosom. Med., 11, 338-353. MCCLEARY, R. A., (1961); Response specificity in the behavioral effects of limbic system lesions in the cat. J . comp. physiol. Psychol., 54, 605-613. MILLER,G . A., GALANTER, E. H., AND PRIBRAM, K. H., (1960); Plans and the Structure OfBehavior. New York, Henry Holt and Co. MILNER,BRENDA,(1958); Psychological defects produced by temporal lobe excision. Res. Publ. Ass. nerv. ment. Dis.,36, 244257. MOUNTCASTLE, V. B., (1957); Modality and topographic properties of single neurons of cat’s somatic sensory cortex. J. Neurophysiol., 20,408434. PENFIELD, W. G., AND MILNER,BRENDA, (1958); Memory deficit produced by bilateral lesions in the hippocampal zone. Arch. Neurol. Psychiat., 79,475497. -RAM, K. H., (1954); Toward a science of neuropsychology: (method and data). Current Trends inPsychology andthe BehavioralSciences, R. A. Patton, Editor, Pittsburgh, University of Pittsburgh Press, pp. 115-142. PRIBRAM, K. H., (1966); Memory and the organization of attention and intention: the case history of a model. Brain Function and Learning, V. P. Hall, Editor, Los Angeles, University of California Press, in press. PRmRAM, K. H., GARDNER,K. W., PRESSMAN, G. L., AND BAGSHAW,MURIELH., (1962); An automated discrimination apparatus for discrete trial analysis (DADTA). Psychol. Rep., 11, 247-250. -RAM, K. H., AND KRUGER,L., (1954); Functions of the ‘olfactory brain’. Ann. N. Y. Acad. Sci., 58!109-138. ROBERTSJW. w., DEMBER, w. N., AND BROADWICK, M., (1962); Alternation and exploration on rats with hippocampal lesions. J. comp. physiol. Psychol., 55, 695-700. SCHWARTZBAUM, J. S., (1960a); Changes in reinforcing properties of stimuli following ablation of the amygdaloid complex in-monkeys. J. comp. physiol. Psychol., 53, 388-395.
336
K. H. P R I B R A M
SCHWARTZBAUM, J. S., (1960b); Response to changes in reinforcing conditions of bar-pressing after ablation of the amygdaloid complex in monkeys. Psychol. Rep., 6, 215-221. SCHWARTZBAUM, J. S., (1961); Some characteristicsof amygdaloid hyperphagia in monkeys. Amer. J . Psychol., 14,252-259. SCHWARTZBAUM, J. S., (1964); Visually reinforced behavior following ablation of the amygdaloid complex in monkeys. J. comp. physiol. Psychol., 57,340-347. SOKOLOV, E. N., (1960); Neuronal models and the orienting reflex. The Central Nervous System and Behuvior, Mary A. B. Brazier, Editor, Josiah Macy, Jr. Foundation, New York, pp. 187-276. SPINELLI, D. N., AND PRJBRAM, K. H., (1966); Changes in visual recovery functions produced by temporal lobe stimulationin monkeys. Electroenceph. clin. Neurophysiol., 20,4449. SPINELLI, D. N., AND PRIBRAM, K. H., (1966); Changes in visual recovery functions and unit activity produced by frontal and temporal cortex stimulation. Electroenceph. elin. Neurophysiol. Accepted for publication. THOMPSON, R., (1966); Proceedingsofthe Fourth ConferenceonBrain FunctionandLearning, sponsored jointly by the Brain Research Institute, University of California, Los Angeles, and the United States Air Force Office of Scientilic Research, in press.
337
Pharmacologic Evidence for Cholinergic Mechanisms in Neocortical and Limbic Activating Systems * E D W A R D F. DOMINO, A N T H O N Y T. DREN**
AND
KEN-ICHI YAMAMOTO***
Department of Pharmacology, University of Michigan, Ann Arbor, Mich. (U.S.A.)
For several years this laboratory has been concerned with the actions of cholinergic agonists and antagonists on the central nervous system as meaured with macroelectrode and behavioral techniques. It is well documented (Feldberg, 1945b, 1950; Paton, 1958;Crossland, 1960)that acetylcholine, cholineacetylase, and cholinesterases are present in various areas of the mammalian brain. On the basis of pharmacologic evidence there is little doubt of the important role of cholinergic mechanisms in cerebral function, although the intimate details are far from clear. The role of acetylcholine as a neurotransmitter at discrete central synapses has not been definitely demonstrated. In order that an agent be proven to be a neurotransmitter many rigid criteria must be met. These can be categorized as follows: ( A ) Anatomical. Identification of the neuronal and synaptic organization of the structures under consideration. ( B ) Biological and chemical. (1) Chemical, histochemical, and/or specific bioassay identification of transmitter in presynaptic nerve endings, and evidence of its release during transmission. (2) Evidence in presynaptic nerve terminals of necessary precursors, as well as enzyme(s) which synthesize transmitter. (3) Presence of enzyme(s)which inactivate transmitter. Probably these will be found postsynaptically, but not necessarily (viz. superior cervical ganglia, Koelle, 1962). (4) Evidence that the concentration of the transmitter varies with functional activity. (C) Physiological. (1) External application of pure substance in reasonable concentration to postsynaptic structures should reproduce the specific events of synaptic transmission such as its effects on the membrane potential, whether excitatory or inhibitory. Synaptic or other diffusional barriers may be a problem in meeting this criteria. (2) Kinetics must be compatible with known facts of transmission. That is,
* Supported in part by Grant NB-01311, USPHS, and Council for Tobacco Research, USA. ** Present address: Department of Pharmacology, Abbott Laboratories, North Chicago, Ill. ***
(U.S.A.). Present address: Department of Neuropharmacology, Shionogi Research Laboratory, Osaka (Japan).
References p . 362-364
338
E. F. D O M I N O el
d.
the time course for release, diffusion, breakdown, synthesis, etc. must be compatible with known synaptic delay times. ( D ) Pathological. (1) Loss of transmitter and its associated substrates, enzymes for synthesis,destruction, etc. with denervation. (2) Denervation sensitivity to transmitter. (E} Pharmacological. (1) Evidence of action of agonists related to the transmitter. (2) Evidence of effects of antagonists, competitive equilibrium, competitive nonequilibrium, non-competitive. (3) Effects of blockers of synthesis, antirelease agents, and substances which inhibit the enzymes for synthesis and destruction. (4) Effects of drugs which inhibit storage or binding of transmitter. ( F ) Phylogenetic. Consistency across various species strengthens the generalization of the concept to the human nervous system. However, as long as the other criteria have been met they can be accepted as proving a transmitter function in a given species. Not all of these criteria have been fulfilled even for the best known central synapse presumed to be mediated by acetylcholine, the Golgi recurrent collateral-Renshaw cell synapse in the spinal cord (Eccles, 1964). Most of our knowledge of this synapse is based on physiologic and pharmacologic evidence which is sufliciently convincing to implicate acetylcholine or a closely related substance. When dealing with higher cerebral systems our knowledge of chemical mediators of synaptic transmission becomes even less certain. With macroelectrode recordings one can only surmise a possible role of neurotransmitters in such complicated structures as those involved in neocortical and limbic system activation. A large body of evidence exists to implicate acetylcholine in the neural mechanisms of EEG activation of the neocortex. Bonnet and Bremer (1937) reported that small doses of acetylcholine injected intracarotidly to high spinal cats caused EEG desynchronization. Bradley (1958) reported a similar phenomenon occurred in cats with prepontine brain stem transections. In contrast, Mantegazzini (1957) was unable to show that intracarotid acetylcholine had any significant EEG effect in cats with a prepontine brain stem transection. His data tended to implicate a peripheral action of acetylcholine in this phenomenon. On the other hand, Rinaldi and Himwich (1955a, b) showed that intracarotid acetylcholine caused EEG activation in both intact curarized and prepontine transected rabbits. The data of Longo and Silvestrini (1957) suggests that acetylcholine produces primarily peripheral effects. Although small doses of acetylcholine caused EEG activation in intact curarized rabbits, even massive doses had no significant EEG effect in prepontine transected preparations. One of the major problems using acetylcholine itself is that it is highly charged and therefore may not penetrate the blood-brain barrier easily. There is much evidence to suggest that cholinergicagents which penetrate the blood-brain barrier producemarked EEG activation. As pharmacologists our concern has been to determine the muscarinic (m)and nicotinic (n) cholinergic elements of the activating system using drugs as tools. An additional problem has been to identify the behavioral correlates, if any, of this phenomenon. Although our early interest was primarily with the brain stem activating system and its influence on the neocortex, recently we have become concerned with the limbic system as well. This manuscript describes some of our results with both acute and chronic cat and acute dog preparations.
C H 0 LINER G I C
EEG
A C T I V A T I N G ME C H A N I S MS
339
METHODS
Cats with chronic indwelling brain electrodes. Experiments were performed in 20 cats utilizing a Latin Square design for drug administration at 2 week intervals. The cats were prepared for placement of indwelling brain electrodes using modifications of conventional techniques. Adult cats of both sexes were used. Surgical preparation of the animals was under pentobarbital sodium anesthesia. Stainless steel wires of 0.22 mm in diameter (insulated except for tips of 0.5 mm) were used as the depth electrodes. Bipolar depth electrodes were inserted into the amygdala and hippocampus with the aid of the stereotaxic atlases of Jasper and Ajmone-Marsan (1954) and Snider and Niemer (1961) using physiological recordings of injury discharges by insertion of electrodes for location of the hippocampus, and olfactory-induced waves for the amygdala. Bipolar silver ball electrodes of 0.5 mm in diameter were applied to the epidural surface of the somatosensory cortex. Additional depth electrodes were placed occasionally in the posterior hypothalamus and mesencephalic reticular formation. Each electro de was soldered to a Cannon plug and fixed on the scalp by means of dental cement.Silastic tubing of 0.7 mm in diameter was inserted into the right jugular vein with the other end fixed to a connector on top of the skull. The animals were allowed to recover for a 2 week period before being used for drug studies. In the meantime, they were given antibiotics prophylactically to reduce infection. At the time of the experiment, the EMG of the posterior neck muscles, EKG and respiratory movements were recorded along with brain waves on a Grass polygraph. The animals were each placed in a sound proof box with a one-way viewing window. Behavioral changes were observed and correlated with EEG activity. In order t o promote naturally occurring sleep, the animals were made as warm and comfortable as possible. With care and patience it was possible to observe all stages of natural sleep. The following drugs were given as an intravenous infusion in physiological saline solution over a 1 min period : acetylcholine chloride, atropine methylnitrate, atropine sulfate, arecoline hydrochloride, mecamylamine hydrochloride, (-) nicotine base, pilocarpine hydrochloride,physostigmine salicylate andtrimethidiniummethosulfate. All drugs were given in doses calculated as base. The actions of these drugs and various combinations on behavior were compared with each drug infused in a constant volume of 1.5 ml over a 1 min period. After completion of a series of experiments the position of each electrode was determined histologically by the iron deposition technique (see Domino, 1955 for details). Acute dogpreparations. Adult animals of both sexes were used. Dogs were prepared under 80 % nitrous oxide-20 % oxygen as well as local lidocaine anesthesia following immobilization with decamethonium, 1 mg/kg intravenously. Artificial respiration was maintained with positive pressure ventilation of 300 ml of air/kg/min. End tidal COZwas monitored with an infrared COZanalyzer. Body temperature was maintained at 37-37.5" by placing a heating pad on the dog. Temperature was controlled automatically with a rectal thermistor connected to an electronic control unit. The femoral artery was cannulated for monitoring blood pressure using a Statham P23A pressure transducer and recorded on an Offner polygraph. The femoral vein was cannulated References p . 362-364
340
E. F. D O M I N O et al.
for intravenous administration of drugs. The animal was positioned in a stereotaxic instrument. Concentric stainless steel needle electrodes with tip exposure of 0.5 mm and an interelectrode distance of 0.5 mm were placed in various subcortical areas according to the stereotaxic atlases of Lim et al. (1960) and Adrianov and Mering (1964). Monopolar and/or bipolar recordings were taken from neocortical areas Prc2, Prcl, and 0 1 , and from the dorsal hippocampus, and basal amygdala. Stimulatingelectrodes were placed in the mesencephalic reticular formation and occasionally in the posterior hypothalamus. The midline nasal bone was used as an indifferent site for monopolar recordings. An intraventricular needle was placed in the left lateral ventricle for drug injections. At least 1-2 h elapsed after nitrous oxide-oxygen anesthesia for elimination of the general anesthetic before drug testing. The following drugs were given either intravenously or intraventricularly: arecoline hydrochloride, atropine methylnitrate, choline chloride, D-tubocurarine chloride, =amphetamine sulfate, epinephrine hydrochloride, gallamine triethiodide, hemicholinium bromide (HC-3), (-) nicotine, pilocarpine nitrate, physostigmine salicylate, sodium chloride and sodium bromide. All drug dosage was calculated as base unless otherwise specified. At the conclusion of each experiment the animal was terminated and the position of each electrode determined histologically by the iron deposition technique. Brain acetylcholine assay. Dogs were sacrificed by use of intravenous compressed air. The brain was removed within 2 min and various areas dissected. Excess blood was removed by blotting on filter paper. The tissue samples were rapidly frozen in isopentane contained in beakers previously cooled in dry ice-ethanol. The entire procedure took no more than 4 min. The frozen samples were weighed and pulverize# in a cold stainless steel mortar. The acetylcholine was extracted by homogenizing the pulverized tissue in 0.2% acetic acid-95% ethanol as described by Stone (1955). The extracts were bioassayed on the neostigmine sensitized frog rectus abdominus preparation (Crossland and Merrick, 1954; Crossland, 1961) maintained at 25" in a 4 or 5 ml bath. Two minute isotonic contractions were recorded on a smoked paper kymograph. A 4 min wash interval was used. As suggested by Feldberg (1945a), acetylcholinestandards were prepared in NaOH inactivated extract in order to allow for the presence of sensitizing factors. RESULTS
( A ) Effects of various cholinergic antagonists on the awake-sleep cycle of the cat
Depending upon dose, atropine either had no effect, facilitated slow wave sleep or prevented natural sleep by causing behavioral hyperexcitability and continuous EEG slow waves. Doses of 0.3 mg/kg of atropine produced definite though minimal m cholinergic blockade both centrally and peripherally. In contrast doses twice as large (0.6 mg/kg) of methyl atropine produced no significant effectson the EEG of neocortica1 and limbic areas, but caused definite peripheral m cholinergic blockade. Doses of 1 mgfkg or more of mecamylamine interfered with the awake-sleep cycle of the cat. Doses of 0.7 mg/kg or less did not significantly alter the awake-sleep cycle of the cat,
C H 0L I N E R G I C
EE G
ACT I V A T I N G MECHANISMS
34 1
but caused minimal n cholinergic blockade both centrally and peripherally. In contrast large doses (I mg/kg) of trimethidinium did not have any significant central effects although marked peripheral n cholinergic blockade was evident. ( B ) Interaction of various cholinergic agonists-antagonists on the awake-sleep cycle of the cat Large doses of acetylcholine given intravenously to cats in slow wave sleep caused prompt EEG activation of the neocortical and limbic structures and behavioral arousal. The EEG manifestations of a large dose of acetylcholine given intravenously are shown in Fig. 1. Either atropine or methyl atropine in a dose of 0.3 mg/kg blocked the EEG and behavioral manifestations of intravenous acetylcholine (see Fig. 2). As might be expected, the effects of intravenous acetylcholine were blocked equally well by m cholinergic antagonists with a predominant peripheral (methyl atropine), or central and peripheral (atropine) effect. In contrast ra ganglionic cholinergic antagonists such as mecamylamine or trimethidinium had no effect on the EEG and behavioral manifestations produced by acetylcholine(see Fig. 3). It would appear that intravenous acetylcholine produces primarily peripheral m cholinergic effects which were blocked by m cholinergic antagonists. The insignificant central effects of intravenous acetylcholine are to be contrasted with those of cholinergic agonists which penetrate the blood-brain barrier. Nicotine (0.02 mg/kg) given intravenously produces prompt activation of neocortical and limbic structures which is accompanied by behavioral arousal. The EEG effects of nicotine are illustrated in Fig. 4. EEG manifestations of activation were noted in neocortical and limbic structures within a few seconds after nicotine injection. Pretreatment with atropine or methyl atropine did not block neocortical activation induced by nicotine. There was a slight reduetion in the duration of EEG activation after atropine. Of particular interest is the dissociation of neocortical and limbic system activation induced by nicotine after atropine pretreatment. Although neocortical desynchronization is still evident, characteristic limbic system activation was not observed (see Fig. 5). In contrast equal doses of methyl atropine did not modify nicotine induced neocortical and limbic activation. The n ganglionic cholinergic antagonists trimethidinium and mecamylamine might be expected to modify the effects of nicotine much more than m cholinergic antagonists. After pretreatment with trimethidinium, nicotine induced activation was slightly reduced, but after mecamylamine it was completely blocked (see Fig. 6). There were no significant nicotine induced behavioral manifestations in mecamylamine pretreated animals, although after trimethidinium pretreatment typical EEG and behavioral wake-up effects to nicotine were obtained. Arecoline is a m cholinergic agonist which should penetrate the blood-brain barrier readily. In doses of 0.04 mg/kg intravenously it produced prompt activation in neocortical and limbic structures as well as a behavioral wake-up effect (see Fig. 7). This action of arecoline lasted approximately I2 min. Following methyl atropine pretreatment, arecoline induced EEG activation was decreased slightly. On the other hand, following atropine pretreatment in the same dose, the effects of arecoline were References p. 362-364
LPOSTSIG
LRb
.,.. ..
S L O W WAVE SLEEP
. ., . . .
,
, .
-,.
-
A W E
..
(,,
...
I , .
..
., . "
,,,.. .
,
.
.
',..
.-*,
..
, ..
.
.-.,.
RESF! ACETYLCHOLINE 0007rnglkg iv
O.Smln
- " " -
-
souv'-Ti5i
Fig. 1. EEG effects of acetylcholine. Acetylcholine was given the arrow during slow wave sleep. Note changes in EEG activity and other physiological signs. The records are continuous. All recordings were bipolar. Symbols: L. POST. SIG. = left posterior sigmoid gyrus; AMY. = amygdala; CM = 'centre mbdian'; HIP. = dorsal hippocampus; EKG = electrocardiogram, lead 11; EMG = electromyograph of neck muscles; RESP. = thoracic respiration. Time base and voltage calibration are as indicated. These symbols when used also apply to Figs. 2-9.
95 MIN AFTER TRlMETHlDlNlW I'maAaiv
E
HEAD SHAKING
RESP
n
<
-
tACETYLCHOLlNE 0007rng/kg
v
v
-
*"
0 Smin
105 MIN AFTER MECAMYLAMlNEO7rna/ka -. = = iv
4
~
-1
>
E RESF!
m v
IV
*A0
IP
-
'
HEAD SHAKING
HEAD MOW *
'
a smin
Fig. 3. Failure of n ganglionic cholinergic antagonists to modify the EEG effects of acetylcholine.
0
b
l
l
l
l
J
.
l
EMG
l
l
~
EYES OPEV
RESP
-
r
v
NICOTINE 002 mghg iv
HEADU?
HEAD M O M
-Lh-?\fi% Fig. 4. EEG effects of nicotine.
HEAD SHAKING 0,
3
0,
t NlCOTlNE 002 mgikg iv
lmin
SOAFzizz
Fig. 5. Modificationof EEG effects of nicotine by m cholinergicantagonists.Note that atropine blocked the effects of nicotine in the amygdala and hippocampus, but not in the neocortex indicating a dissociation. Behavioral arousal to nicotine still occurred. In contrast methyl atropine did not modify the EEG and behavioral effects of nicotine.
P 250 MIN AFTER TRMETHIDNIUM I rng/kg I V SLOW WAVE SLEEP
AWAKE
n
1 -
31
0 8
EYES WEN
.
NICOTINE 0.02 rnglkg iv
tNlCOTlNE 0.02 rng/kg iv
r
HEADUP
-
-
*
SOvrPrrs
Fig. 6. Modification of the EEG effects of nicotine by n ganglionic cholinergic antagonists.Note that trimethidinium did not block EEG activation and behavioral arousal to nicotine. On the other hand mecamylamine blocked these responses.
w
5
Refeerenc
t.
362-364
CHOLINERGIC
EEG ACTIVATING MECHANISMS c
9
a
Y0 s c
0
c
c
d
L
.-d
349
200 MIN AFTER TRlMETHlDlNlUM I mg/kgiv LPOST.SIG.
SLOW WAVE SLEEP
AWAKE
,
f ARECOLINE 0.04rng/kg iv
C
"
.
Y
V
z
0
It3 MIN AFTER MECAMYLAMlNEO.7 mglkgiv
L.POST.SIG.
SLOW WAVE SLEEP
2
AWAKE
R
-tw+vwwwm* EYES ?PEN
RESP
HEAD SHAKING
MA3 UP I,' *c
'
HEAD MOVE
"14
I mln
tAREC0I-M 004 mghg
Iv
Fig. 9. Lack of blockade of the EEG and behavioral effects of arecoline by n ganglionlc cholinergic antagonists.
CHOLINERGIC
EEG
A C T I V A T I N G MECHANISMS
351
completely blocked (see Fig. 8). Atropine pretreated animals remained in slow wave sleep which was not altered by arecoline injection. In contrast to the effectiveness of atropine in blocking EEG activation of arecoline, the n ganglionic cholinergic antagonists, trimethidinium and mecamylamine, had no significant effects (see Fig. 9). Similar studies have been carried out using other cholinergic agonists and antagonists. Table I summarizes the interactions of these agents on neocortical and hippocampal EEG activation in the chronic cat. The number of +'s indicate approximately the duration of EEG activation of neocortical structures (as manifested by low voltage, fast wave activity) and the hippocampus (as manifested by 8-activity). Given alone or following saline pretreatment all cholinergic agonists produced varying degrees of activation of the neocortex and hippocampus. Although all agonists were given in maximal doses their duration of action varied considerably. Physostigmine produced EEG activation for approximately 31 min. Arecolineand pilocarpine produced activation for approximately 12 min. Nicotine and acetylcholine produced EEG activation for approximately 4 min and DMPP for approximately 2 min. On the basis of the interactions of intravenously administered cholinergic agonists it may be concluded that the actions of acetylcholine are primarily peripheral and of the m cholinergic type. Pilocarpine also is a m cholinergic agent with peripheral as well as central actions. Arecoline and physostigmine have prominent central actions. The n ganglionic cholinergic antagonists did not block the EEG effects of m cholinergic agonists. The reverse situation, however, was not true. Administration of a centrally acting m cholinergic antagonist (atropine) reduced the activating effects of n cholinergic agonists such as DMPP and nicotine. Particularly striking was the dissociation in neocortical and hippocampal activation following nicotine administration to atropine pretreated animals. The EEG activating effects of DMPP appear to involve primarily peripheral mechanisms. Nicotine has central and peripheral actions which contribute to its overall EEG activation. Only a n ganglionic cholinergic antagonist with central and peripheral actions, mecamylamine, completely blocked the effects of nicotine.
( C ) Eflects of hemicholinium on the neocortical and limbic activating systems of the dog As might be expected, because of its highly cationic character, intravenous hemicholinium (HC-3) did not produce consistent EEG changes in dogs maintained under adequate artificial ventilation. Enormous doses (of 5-10 mg/kg, i.v.) do, in occasional animals, produce marked EEG slowing and failure of activation to all afferent stimuli. It was assumed that the marked variability was due to individual differences in bloodbrain barrier permeability, brain and/or blood levels of choline, etc. Therefore, the drug was given intraventricularly. Via this route, remarkably small amounts of HC-3 produced consistent EEG changes in neocortical and limbic areas. As illustrated in Fig. 10, HC-3 (50 pg total dose) within one and a half hours initially caused spiking in the amygdala and abolished spontaneous hippocampal &activity. However, neocortica1 activation still persisted. After 3 h both neocortical and limbic system activation was blocked. Intraventricular and intravenous choline, after a delayed onset, partially antagonized the effects of HC-3. In the particular animal whose EEG recordings are References p. 362-364
TABLE I CHOLINERGIC AGONIST-ANTAGONIST I N T E R A C T I O N S O N NEOCORTICAL A N D H I P P O C A M P A L
EEG
ACTIVATION
Agonist Antagonist
Acetylcholine 0.007 mglkg
N
H
Arecoline 0.04 mglkg
N
H
Pilocarpine 0.15 mglkg
-N
H
Physostigmine 0.05 mglkg
N
H
DMPP 0.005 mglkg
N
H
Nicotine 0.02 mglkg N
H
Mecamylamine 0.7 mg/kg
+++ +++ ++++ ++++ ++++ ++++ +++++ +++++ ++ ++ +++ +++ 0 0 -I0 0 0 0 + O O + + 0 0 0 +++ +++ ++ ++ +++ +++ +++ +++ + + +++ +++ ++++ ++++ ++++ ++++ +++++ +++++ + + 0 0
Trimethidinium 1 melke
+++ +++ ++++ ++++ ++++ ++++ +++++ +++++
None Atropine 0.3 mg/kg Methyl atropine 0.3 mg/kg
++ +
+ + = Markedactivation; H = Dorsal hippocampus 0-waves
+ = Minimal activation;
0 = No significantactivation;
0
0
++
++
N = Somatosensory cortex desynchronization;
m ?I
tJ
0
5 z
0
s %
r:* M
c, c3
<
3
P
=I 2,
0
E0 X * 5M
Fig. 10. EEG effects of HC-3 and modification by choline. Note the initial effects of HC-3 are in the amygdala and hippocampus and not in the neocortex. Subsequently all EEG channels show spikes and slow waves. Intraventricular choline 6 h after HC-3 altered slightly the hippocampal EEG, but not the other areas. Atropine methylnitrate (0.125 mg/kg, i.v.) war administered 10 min before intravenous choline. Intravenous choline 7 h after HC-3 had a g r a t a antagonistic effect than intraventricular choline. Recovery was noted 25 h after injection. Symbols: Prcl = monopolar recording from postcruciate gyrus; Prcg = monopolax recording from precruciate gyrus; AMYG. = monopolar recording from medial amygdala; DORSAL HIPP. = bipolar recording from dorsal hippocampus; B.P.= femoral arterial blood pressure in rnm Hg. Reference site for monopolar recordings was nasion. Time base as indicated. The vertical bars for voltage calibration indicate 100 pV. These symbols when used apply to the subsequent tigures.
5
w
VI
w
354
E. F. DOMINO et al.
illustrated in Fig. 10, recovery was obtained 25 h after HC-3. Similar effects were not observed following intraventricular administration of comparable amounts of sodium chloride, sodium bromide, Dtubocurarine, and gallamine indicating the specificity of HC-3. Fig. 11 illustrates HC-3 blockade of neocortical and hippocampal activation by electrical stimulation of the midbrain reticular formation and posterior hypothalamus at a time of maximum effect. Similar blockade was observed on activation produced by high frequency stimulation of the diffuse thalamic projection system and a variety of afferent stimuli. Dissociation of both evoked and spontaneous neocortical and limbic system activation occurred with the 10 pg dose of HC-3. A similar dissociation was seen early after large doses of HC-3 which after 2 h usually caused generalized EEG slow waves. ( D ) Eflects of hemicholinium on dog brain acetylcholine levels
To date only the effects of large intraventricular doses (5000 pg) of HC-3 have been studied on brain acetylcholine levels of the dog. Brains were assayed for acetylcholine 4 h after HC-3 administration when EEG activation was completely blocked. HC-3 reduces the acetylcholine-like activity of various subcortical structures. The medial thalamus, hippocampus, amygdala, and medullary reticular formation all showed approximately 50 % decrease in acetylcholine content, in contrast to the neocortex. A bar graph showing the mean acetylcholine content f S.E. in mpM/g of various brain areas after saline or HC-3 injection is illustrated in Fig. 12. All areas showed a significant decrease in acetylcholine content except the neocortex. ( E ) Efects of various cholinergic agonists in HC-3 induced brain acetylcholine deficient animals
As described above, 5 mg of HC-3 intraventricularly does not deplete subcortical brain acetylcholine, but only reduces total levels. It would be expected that some cholinergic agonists would antagonize the functional deficits produced by the lowered levels of brain acetylcholine. Non-cholinergic centrally acting drugs which act on neuronal chains containing a cholinergiclink should be less effective. The EEG actions of nicotine and arecoline in such brain acetylcholine deficient preparations are illustrated in Fig. 13. It can be observed that arecoline, a m cholinergic agent, is a much better antagonist of HC-3 than nicotine, a n cholinergic drug. Similar studies have been conducted with a variety of drugs which produce neocortical EEG activation. The results to date are summarized in Table 11. It may be observed that primariIy the m cholinergic agonists which penetrate the blood-brain barrier are effective. Of course, dose and time after HC-3 are critical factors, but to date it has been observed that this generalization is valid. Much less has been done in studying the actions of such substances on limbic system activation. Further research is now in progress.
CONTROL
5H
A F T E R H C - 3 (2.5 mg LVENTRIC.) I
L Prc, MESENCEPHALIC RETICULAR FORMATION 300 c/s. I msec
1
L.0,
I
L.DORSAL HIPP. I
8.P.
0.75V
POSTERIOR HYPOTHALAMUS IOOC/S.
I msec
2.5 V
L 01
*
L.DORSAL HIPP 0P
1.75V
Fig. 11. Effects of HC-3 on neocortical and hippocampal activation. EEG activation was elicited by mesencephalic reticular and posterior hypothalamic electrical stimulation. Note that in spite of the marked increase in intensity of electrical stimulation, HC-3 slow waves were only slightly modified.
w
u u
356
E. F. DOMINO
et al.
DISCUSSION
The major problem in studying central cholinergic phenomena is that most drugs affecting this system are highly charged molecules that do not penetrate the bloodbrain barrier easily. As a result the effects of most cholinergic agents on brain function cannot be studied adequately when given via a blood-borne route. Direct electrophoretic, intracerebral or intraventricular methods of drug administration likewise have certain basic shortcomings. As a result, a combination of techniques must be used to ascertain cerebral cholinergic mechanisms. In the introduction certain criteria were listed that must be met to prove a neurotransmitter mechanism. It is pertinent to
16
1
SALINE (0.2ml INTRAVENTRICULAR) H C - 3 ( 5 m g INTRAVENTRICULAR)
PrCe
AYYGDALA
HIPPOCAYPUS
THALAMUS
RETIC. FORMATION
Fig. 12. Effects of HC-3 on total acetylcholinelevels of various brain areas. Note that the subcortical areas show a decrease in total acetylcholine content. These reductions are statistically significant (P< 0.05) using the student ‘t’ test, group comparison.N representsnumber of dogs for each analysis. The height of each bar represents the mean acetylcholinecontent in mpM/g of f r o m tissue.
ascertain how much is known concerning cholinergic mechanisms in the neocortical and limbic activating systems. The neural substrates for neocortical and limbic system activation are still far from clear. Since the original postulates of Moruzzi and Magoun (1949) it has become apparent that the activating system is anatomically quite complex. Neocortical activation appears to involve primarily extrathalamic components. The original suggestion of a diffuse thalamic pathway for neocortical activation is incompatiblewith the experiments of Schlag et al. (1961) and Schlag and Chaillet (1963). High frequency thalamic stimuli which elicit EEG activation apparently are channeled caudally into the reticular formation. Tokizane et al. (1960) have presented evidence for a hypothalamic pathway involving neocortical and hippocampal activation. The hippocampal 6-wave system apparently is regulated by pacemaker cells in the septum (Stumpf et al. 1962; Petsche et al. 1962). Acetylcholine has been shown to be localized in various anatomical areas involved in EEG activation (Paton, 1958). Although specific chemical assays for acetylcholine
CHOLINERGIC
EEG
A C T I V A T I N G MEC H A N I S MS
357
are still unsatisfactory, sufficiently sophisticated bioassay techniques have demonstrated that acetylcholine or an extremely similar agent is involved. For practical purposes it suffices to consider this as acetylcholine until specific chemical assays are available. It is of interest that KrnjeviC and Phillis (1963a, b, c) and KrnjeviC (1964, 1965a, b) have shown that electrophoretic injection of acetylcholine affects cells primarily in the deeper layers of the neocortex. These cells characteristicallydischarge slowly and in a prolonged fashion to acetylcholine. The distribution of acetylcholinesterase is in layer 5 in the vicinity of the pyramidal cells (KrnjeviC and Silver, 1964). This is in agreement with the finding that deep neocortical cells are primarily responsive to acetylcholine. The acetylcholinesterase containing fibers are prominent among the U-shaped neurons connecting adjacent cortical areas. Subcortical cells in the corpus striatum and precommissural septum containing acetylcholinesterase appear to send fibers to the cerebral cortex. KrnjeviC and Silver have shown further that the fibers in the forebrain containing acetylcholinesterase originate in cells in the striatal primordium, and later TABLE I1 EFFECTS OF V A R I O U S
EEG
ACTIVATORS ON
HC-3
INDUCED SLOW WAVES
5 mg of HC-3 were given intraventricularly and the drugs tested at least 4 h later
EEG activators Arecoline 0.04 mg/kg Choline 20 mg/kg Pilocarpine 0.50 mg/kg Physostigmine 0.10 mg/kg Nicotine 0.04 mg/kg D-Amphetamhe 2.0 mg/kg Epinephrine 0.005 mg/kg
Neocortieal &synchronization
Hippocampal 8
+++ ++ +++ +++ +
+++ + +++ +++
0
0
0 0 0
+ + + = Marked activation; + = Minimal activation; 0 = No significant effect. grow into the initially acetylcholinesterase free pallium. KrnjeviC (1965b) suggests that phylogenetically recent non-cholinergic neurons in the cortex may be under control of a cholinergic projection from subcortical basal forebrain structures. There is evidence (Shute and Lewis, 1963) that there is a continuous brain stem system of neurons which contain acetylcholinesterase. It is well known that acetylcholine is released from the neocortex, and that the amount released varies with cerebral activity. MacIntosh and Oborin (1953) have shown that acetylcholine is released from the intact cerebral cortex, but not from the isolated cortical slab. The release of acetylcholine is greater during enhanced activity such as via reticular stimulation and spontaneous neocortical activation (Mitchell, 1963; Kanai and Szerb, 1965; Jasper, 1965). As might be expected the neocortical content of acetylcholine is increased on the ipsilateral side of a midbrain pretrigeminal transection (Pepeu and Mantegazzini, 1964). To our knowledge similar experiments involving the hippocampus or other limbic structures have not been performed. References p . 362-364
CHOLINERGIC
EEG
A C T I V A T I N G MECHANISMS
359
Not only are some neocortical cells responsive to electrophoreticallyinjected acetylcholine, but so are reticular, hippocampal and pyriform cortical neurons. The acetylcholine sensitive cells of paleocortex are distributed throughout all cortical layers in contrast to the neocortex (RandiC and Straughan, 1965). At least two reports have appeared pointing out the sensitivity of hippocampal neurons to acetylcholine (Stefanis et al., 1964; Biscoe and Straughan, 1965). Considerable data exists that some brain stem reticular neurons are sensitive to acetylcholine (Curtis and Koizumi, 1961; Salmoiraghi and Steiner, 1963;Bradley and Mollica, 1958; Bradley and Wolstencroft, 1965). Neurons responding to acetylcholine usually show increased firing rates. Only a small percentage exhibit decreased firing. In the neocortex and hippocampus the cells usually show m cholinergic responses, while in the medulla the cells show both m and n cholinergic responses. One of the most disturbing consequences of the electrophoretic studies is the peculiar time course of neuronal responsiveness to acetylcholine. In general, the responses are slow and prolonged, somewhat incongruous with a discrete neurotransmitter function, but very much in agreement with the local hormone concepts of Desmedt and LaGrutta (1957). Not many pathological studies have been performed. Shute and Lewis (1963) have shown that lesions of neuronal systems projecting to amygdaloid and hippocampal areas produce a decrease of acetylcholinesterase in these structures. Section of the fornix also results in the loss of cholineacetylase activity in the hippocampus (Lewis et al., 1964). Pharmacological data suggesting the importance of cholinergic mechanisms in neocortical and limbic system EEG activation is impressive. The Russian pharmacologists Denisenko (1962), Ilyutechenok (1962), Michelson (1961), Valdman (1961, 1963) and others have obtained considerableevidence for the presence of m and n cholinergic receptors in neocortical activation. Our own data with nicotine (Knapp and Domino, 1962; Yamamoto and Domino, 1965) arecoline (Villareal and Domino, 1964) and the results of the present study lead us to similar conclusions. The gross behavioral consequence of such EEG activation is clearly a wake-up or arousal state. It has previously been reported by Wikler (1952), Bradley and Elkes (1957), Bradley and Nicholson 1962; Yamamoto and Domino, 1965),arecoline (Villarreal and Domino, 1964)and the tion from gross behavior. Similar findings were noted in the present study with large doses of atropine. However, it should be pointed out that effective doses of physostigmine and other cholinergic agonists produce behavioral arousal that is associated with neocortical and limbic activation. The emphasis in the literature on EEG dissociation from gross behavior may have been overstated, particularly in relationship to the awake-sleep cycle of the chronic cat. The findings of Bradley and Elkes and others with cholinergic agonists were made at a time when the stage of fast wave sleep was not generally known. It would appear that in some instances investigators may have been observing fast wave sleep and did not recognize it as such. The findings of the present study strengthen the concept of an intimate relationship between EEG effects and behavior in regard to the awakesleep cycle of the cat. By the use of various m and n cholinergic antagonists with differential abilities to penetrate the blood-brain References p . 362-364
360
E. F. DO MIN O et ul.
barrier, it has been possible to determine if the actions of various cholinergic agonists given intravenously were primarily central or peripheral in origin. It should be noted that the method of pharmacological antagonism, as used in this study, is fraught with possible error. Perhaps the most important criticism is that any investigation of agonist-antagonist interactions should be quantitative involving doseresponse curves. Trendelenburg (1963) has pointed out that false conclusions can be drawn in studies involving single doses of agonists and antagonists. In view of the lack of simple and meaningful methods of quantification of EEG activity no direct dose-effectrelationships were obtained. For this reason the doses of various cholinergic agonists used produced nearly maximal EEG and behavioral manifestations. In contrast, the doses of the cholinergic antagonists were those that did not significantly interfere with the awakesleep cycle. Admittedly, the doses of various antagonists were insufficient to completely block peripheral autonomic effects such as alteration in blood pressure or heart rate. In spite of these limitations several predictions were confirmed experimentally. Acetylcholine is a highly charged molecule that does not easily penetrate the bloodbrain barrier. Its actions in producing EEG activation and behavioral arousal should be primarily peripheral in origin. Therefore these effects should be blocked equally well by atropine and methyl atropine. This, in fact, was obtained. These findings are in agreement with the observations of Nakao et ul. (1956) that the EEG effects of intravenous acetylcholine are markedly reduced by sino-aortic denervation. On the other hand, cholinergicagonists which penetrate the blood-brain barrier easily, should have central as well as peripheral actions. Methyl atropine should reduce slightly the actions of such compounds by a peripheral action. In contrast, atropine, which has both central as well as peripheral actions, should produce effective blockade. It is well recognized that the blood-brain barrier is relative. Although Paul-David et al. (1960) have shown that methyl atropine produces EEG slow waves, the doses used in the present study were clearly below those which produce significant central actions. The effects of various m cholinergic agonists such as acetylcholine,arecolinek,pilocarpine, and physostigmine were blocked by the m cholinergic antagonistsand not significantly reduced by the n cholinergic antagonists. In certain instances mecamylamine actually enhanced EEG desynchronization of some cholinergic agents. In contrast, nicotine induced neocortical activation was not markedly affected by methyl atropine or atropine in the doses used. An interesting dissocation between the neocortical and limbic structures was observed following atropine pretreatment and subsequent nicotine administrationas well as following intraventricular HC-3. Recently, Torii and Wikler (1965) noted that atropine dissociated neocortical and limbic activation to electrical stimulation of the reticular formation or posterior hypothalamus. However, it is well known that large doses of atropine effectively block neocortical activation due to a variety of stimuli (Domino and Hudson, 1958; Paul-David et ul., 1960; White and Boyajy, 1959;White and Daugneault, 1959). This data plus the findings of Loeb et ul. (1960) suggest that the terminal neurons for neocortical activation must involve a m cholinergic neuron, perhaps in the cortex. This is in agreement with recent electrophoreticevidencethat the
CHOLINERGIC
EEG
A C T I V A T I N G MECHANISMS
361
excitatory effects of acetylcholine in the cerebral cortex are m cholinergic with no significant n cholinergic component (KrnjeviC, 1965a, b). HC-3 is known to possess at least three pharmacological actions by which it can affect cholinergic mechanisms. These are: (a) inhibition of acetylcholine synthesis with subsequent reduction of tissue acetylcholine levels (MacIntosh, 1963), (b) curare-like blockade of n neuromuscular synapses (Schueler, 1960) and (c) inhibition of cholinesterase (Long, 1963). The inhibition of acetylcholine synthesis by HC-3 is thought to be due to its ability to affect choline transport (Schueler, 1960). This notion is strengthened by the fact that choline is a specific antagonist. Although the majority of evidence concerning the mechanisms of action of HC-3 has been obtained from peripheral cholinergic synapses, several reports have indicated' that HC-3 can lower cerebral acetylcholine levels (MacIntosh, 1963; Metz, 1962) as well as alter cerebral activity (Frazier and Boyarsky, 1964; Szerb, 1965). It is worth noting that intraventricular HC-3 significantly lowered subcortical acetylcholine levels at a time when neocortical and limbic electrical activity was markedly affected. These results as well as the antagonistic actions of choline suggestthat the effect of HC-3 on cerebral electrical activity may be due to a reduction'in brain acetykholine. Furthermore, the m cholinergic agonists were much more effective than n cholinergic agonists or adrenergic agents in antagonizing the EEG effects of HC-3. In conclusion, it seems that the overall evidence for implicating cholinergic mechanisms in neocortical and limbic system EEG activation is quite impressive. SUMMARY
Four predominantly muscarinic (m) cholinergic agonists (acetylcholine, arecoline, pilocarpine, and physostigmine) and two nicotinic ganglionic (n) cholinergic agonists (DMPP and nicotine) were studied on the awake-sleep cycle of cats. The animals had chronic indwelling brain electrodes in various neocortical and limbic areas. The effects of these compounds were compared before and after the following m and n cholinergic antagonists : atropine, methyl atropine, mecamylamine and trimethidinium. Atropine pretreatment blocked EEG activation induced by acetylcholine, arecoline, pilocarpine and physostigmine, but only reduced that produced by DMPP and nicotine. Atropine also blocked nicotine induced hippocampal &wave activity. Methyl atropine, a m cholinergic antagonist with predominant peripheral effects, markedly antagonized EEG activation by acetylcholine, but did not block EEG activation induced by other m or n cholinergic agonists. The n ganglionic cholinergic antagonists mecamylamine and trimethidinium, had no significant effects on EEG activation induced by m cholinergic agonists. On the other hand, the actions of n cholinergic agonists such as DMPP and nicotine were completely blocked by mecamylamine. Trimethidinium blocked EEG activation of DMPP but reduced only slightly that of nicotine. In general, gross behavior of the cats paralleled the neocortical EEG effects of these drugs when given in low dosage. Another pharmacologicalapproach to studying central cholinergicmechanisms was with the drug hemicholinium (HC-3) which decreases acetylcholine synthesis by interReferences p. 362-364
362
E. F. DOMINO
et al.
fering with choline transport. Acute dog preparations were used to study the effectsof HC-3. The actions of the drug were unpredictable on intravenous administration, but highly reproducible when given intraventricularly in total doses of 10-5000 pg. HC-3 produced initially amygdala spiking and blockade of hippocampal 8-wave activity, but not neocortical activation. This demonstratesa dissociationbetween the neocortical and limbic activating systems. Eventually, neocortical slow waves appeared. The EEG effects of HC-3 may be related to lowered levels of acetylcholine, because subcortical acetjdcholine was reduced approximately 50 % 4 h after drug administration. Exogenous choline produced a delayed and transient reversal of the HC-3 effects. Arecoline, pilocarpine, and physostigmine caused EEG activation following HC-3, whereas nicotine, ephinephrine and D-amphetamine were either much less effective or their EEG actions were completely blocked. REFERENCES ADRIANOV,0. S., AND MERING,T. A., (1964); Atlas of the Canine Brain. E. Ignatieff, Translator and E. F. Domino, Editor Ann Arbor, Mich., University of Michigan, p. 340. BISCOE,T. J., AND ST~AUGHAN, D. W., (1965); The pharmacology of hippocampal neurones. J. Pharm. Pharmacol., 17, 60-61. BONNET,V., AND BREMER, F., (1937); Action du potassium, du calcium, et de l'acktylcholine sur les activitk blectriques, spontanks et provoquks, de l'bcorce ddbrale. C. R. SOC.Biol. (Paris), 126, 1271-1275. BRADLEY, P. B., (1958); The central action of certain drugs in relation to the reticular formation of the brain. Reticular Forma?ion of the Bruin. Henry Ford Hospital International Symposium. H. H. Jasper, L. D. Proctor, R. S. Knighton, W. C. Noshay, R. T. Costello, Editors. Boston, Little, Brown, pp. 123-149. BRADLEY, P. B., AND ELKES, J., (1957); The effects of some drugs on theelectrical action on the brain. Brain, 80, 77-1 17. BRADLEY, P. B., AND MOLLICA,A., (1958); The effects of adrenaline and acetylcholine on single unit activity in the reticular formation of the decerebrate cat. Arch. ital. Biol., %, 168-186. BRADLEY, P. B., AND NICHOLSON, A. N., (1962a); The effect of some drugs on hippocampal arousal. Electroenceph. din. Neurophysiol., 14, 824-834. BRADLEY,P.B., AND NICHOLSON, A. N., (1962b); The effects of drugs on the electrical activity of the hippocampus. Physiologie de I'
[email protected]. Passouant, Editor. Editions du Centre National de la Recherche Scientifique, Paris, pp. 445462. BRADLEY, P.B., AND WOLSTENCROFT, J. H., (1965); Actions of drugs on single neurones in the brainstem. Brit. med. Bull., 21, 15-18. ~RWLAND, J., (1%0); Chemical transmission in the central nervous system. J. Pharm. Pharmacol., 12, 1-36. CROSSLAND, J., (1961); Biologic estimation of acetylcholine. Meth. med. Res., 9, 125-129. C k o s s ~J.,~ AND ~ , ~~ERRIcK, A. J., (1954); The effectof anaesthesia on the acetylcholine content of brain. J. Physiol. (Lond.), 125, 56-66. CVRns, D. R., AND KOIZUMI, K., (1961); Chemical transmitter substancs in brain stem of cat. J. Neurophysiol., 24, 80-90. DENISENKO, P. P., (1962); Influence of pharmacological agents upon cholinoreactive and adrenoreactive systems of the reticular formation and other regions of the brain. Pharmacological Analysis of Central Nervous Action. Proc. First Int. Pharmacol. Meet. W.D. M. Paton, Editor. Vol. 8. New York, Macmillan Co., pp. 199-201. DESMEDT,J. E., AND LAGRUTTA, G., (1957); Theeffect of selective inhibition of pseudocholinesterase on the spontaneous and evoked activity of the cat's cerebral cortex. J. Physiol. ( h n d . ) , 136,2&40. DOMINO, E. F., (1955); A pharmacological analysis of the functional relationship between the brain stem arousal and diffuse thalamic projection systems. J. Pharmacol. exp. Ther., 115, 449-463. ECCLES, J. C., (1964); l7ze Physiology of Synapses. Berlin, Springer-Verlag, p. 316. FELDBERG, W., (1945a); Synthesis of acetylcholine by tissue of the central nervous system. J. Physiol., 103,367-402.
CHOLINERGIC
EEG
A C T I V A T I N G MECHANISMS
363
FELDBERG, W., (1945b); Present views on the mode of action of acetylcholine in the central nervous system. Physiol. Rev., 25, 596-642. FELDBERG, W.,(1950); Theroleofacetyclholinein thecentralnervoussystem. Brit.med. BuIl.,6,312-321. D. T., AND BOYARSKY, L. L., (1964); Evidence for cholinergic transmission at a central FRAZIER, synapse. Proc. Soc. exp. Biol. ( N . Y.), 115,876-879. ILYUTECHENOK, R. I., (1962); The role of cholinergic systems of the brainstem reticular formation in the mechanism of central effects of anticholinesterase and cholinolytic drugs. Pharmacological Analysis of Central Nervous Action. Proc. First Int. Pharmacol. Meet. W. D. M. Paton, Editor. Vol. 8. New York, Macmillan Co.,pp. 211-216. JASPER, H. H., (1965); Mechanism for the selection and preservation of acquired stimulus-response patterns. Proceedings of the Znternational Union of Physiological Sciences. XXZZZ Znternational Congr<ss, IV,641-644. C., (1954); A Stereotaxic Atlas of the Diencephalon of the Cat. J4SPER, H. H., AND AJMONE-MARSAN, Ottawa, The National Research Council of Canada. J. C., (1965); Mesencephalic reticular activating system and cortical acetylKANAI,T., AND SZERB, choline output. Nature (Lond,), 205, 80-82. KOELLE,G.,(1962); A new general concept of the neurohumoral functions of acetylcholine and acetylcholinesterase. J. P h r m . Pharmacol., 14, 65-90. E. F., (1962); Action of nicotine on the ascending reticular activating KNAPP,D. E., AND DOMINO, system. Znt. J. Neuropharmacol., 1, 333-351. KRNJEVIC,K., (1964); Micro-iontophoretic studies on cortical neurons. Int. Rev. Neurobiol., 7,41-98. KRNJEVIC, K., (1965a); Actions of drugs on single neurons in the cerebral cortex. Brit. med. Bull., 21 10-14. KRNJEMC, K., (1965b); Transmitters in the cerebral cortex. Proceedings of the Znternational Union of Physiol3gical Sciences. XXIIZ International Congress, IV, p. 435-443. KRNJEVIC, K., AND PHILLIS,J. W., (1963a); Iontophoretic studies of neurones in the mammalian cerebral cortex. J . Physiol. (Lond.), 165, 274-304. J. W., (1963b); Acetylcholine-sensitive cells in the cerebral cortex. J. KRNJEVI~, K., AND PHILLIS, Physiol. (Lond.), 166, 296-327. KRNJEVIC,K., AND PHILLIS,J. W., (1963~);Pharmacological properties of acetylcholine-sensitivecells in the cerebral cortex. J. Physiol. (Lond.), 166, 328-350. KRNJEVIC,K., AND SILVER,A., (1964); The development of cortical acetylcholinesterase (AChE) staining in the brain of the cat. J. Physiol. (Lond.), 175, 22p23p. LEWIS,P. R., SHUTE,C. C. D., AND SILVER,A., (1964); Confirmation from choline acetylase analyses of a massive cholinergic innervation to the hippocampus. J. Physiol. (Lond.), 172,9p-lop. LIM,R. K. S., LIU,C. N., AND MOFFTIT, R. L., (1960); A Stereotaric Atlas of the Dog's Brain. Springfield, Thomas, pp. 93. LOEB,C., MAGNI,F., AND ROSSI,G. F.,(1960); Electrophysiological analysis of the action of atropine on the central nervous system. Arch. ital. Biol., 98, 293-307. LONG, J. P., (1963); Structure-activity relationships of the reversible anticholinesterase agents. Cholinesterasesand Antichalinesterase Agents. G. B. Koelle, Editor. Handbuch der experimentellen Pharmakologie. Vol. 15. Berlin, Springer-Verlag, pp. 374-427. LONGO,V. G., AND SILVESTRINI, B., (1957); Effects of adrenergic and cholinergic drugs injected by intra-carotid route on electrical activity of brain. Proc. Soc. exp. Biol. ( N . Y . ) , 95, 43-47. F. C., (1963); Synthesis and storage of acetylcholine in nervous tissue. Can. J. Biochem. MACINTOSH, Physiol., 41, 2555-2571. F. C., AND OBOIUN,P. E., (1953); Release of acetylcholine from intact cerebral cortex. MACINTOSH, Abstracts of Communications. 19th Int. Physiol. Congr., Montreal. August 31 to Sept. 4, 1953, pp. 580-581. MANTEGAZZINI, P., (1957); Action de la 5-hydroxytryptamine (entkramine) et de l'adtylcholine sur le track 6lectroenckphalographique du chat. Arch. int. Pharmacodyn., 112, 199-21 1. METZ,B., (1962); Correlation between respiratory reflex and acetylcholine content of pons and medulla. Amer. J. Physiol., 202, 80-82. MICHELSON, M. J., (1961); Pharmacological evidences of the role of acetylcholine in the higher nervous activity of man and animals. Activ. Nerv. Super., 3, 1-147. MITCHELL, J. F., (1963); The spontaneous and evoked release of acetylcholine from the cerebral cortex. J. Physiol. (Lond.), 165,98-116. MORUZZI, G., AND MAGOUN, H. W., (1949); Brainstemreticularformationand activation of the EEG. Electroenceph. clin. Neurophysiol,, 1, 455-473.
364
E. F. DOMINO et al.
NAKAO, H., BALLIM, H. M., AND GELLHORN, E., (1956); The role of the sino-aortic receptors in the action of adrenaline, nor-adrenaline and acetylcholine on the cerebral cortex. Electroenceph. elin. Neurophysiol., 8,413420. PATON, W. D. M., (1958); Central and synaptic transmission in the nervous system (Pharmacological aspects). Ann. Rev. Physiol., 20,431-470. PAUL-DAVID, J., RIEHL,J.-L., AND UNNA,K. R., (1960); Quantification of effects of depressant drugs on EEG activation response.J. Pharmacol. exp. m r . , 129, 69-74. PEPEU, G., AND MANTEGAZZINI, P., (1964); Midbrain hemisection: Effect on cortical acetylcholine in the cat. Science, 145, 1069-1070. PETSCHE, H., STLJMPF, CH., AND GOGOLAIC, G., (1962); The significance of the rabbit’s septum as a relay station between the midbrain and the hippocampus. I. The control of hippocampus arousal activity by the septum cells. Electroenceph. clin. Neurophysiol., 14,202-211. RANDIC, M., AND STRAUGHAN, D. W., (1965); Iontophoretic study of palaeocortical neurones. J. Physiol. (Lond.), 177, 61-68. RINALDI, F., AND HIMWICH, H. E., (1955a); Alerting responses and actions of atropine and cholinergic drugs. AMA Arch. Neurol. Psychiat., 73, 387-395. RINALDI, F., AND HIMW~CH, H. E., (1955b); Cholinergic mechanism involved in function of mesodiencephalic activating system. AMA Arch. Neurol. Psychiat., 73, 396.402. SALMOIRAGHI, G. C., AND STEINER, F. A., (1963); Acetylcholine sensitivity of cat’s medullary neurons. J. Neurophysiol., 26, 581-597. SCHLAG, J. D., AND CHAILLET, F., (1963); Thalamic mechanisms involved in cortical desynchronization and recruiting responses. Electroenceph. elin. Neurophysiol., 15, 39-62. SCHLAG, J. D., -LET, F., AND HERZET, J.-P., (1961); Thalamic reticular system and cortical arousal. Science, 134, 1691-1692. SCHLELER,F. W., (1960); The mechanismofactionofthehemicholiniums. Znt. Rev. Neurobiol.,Z,77-97. SHUTE, C. C. D., AND LEWIS,P. R., (1963); Cholinesterase-containingsystems of the brain of the rat. Nature (Lond.), 199, 1160-1164. SNIDER, R. S., AND NIEMER, W. T., (1961); A Stereotaxic Atlas of the Cat Brain. Chicago, University of Chicago Press. STEFANIS, C., (1964) ; Hippocampal Neurons: Their responsiveness to microelectrophoretically administered endogenous amines. Pharmacologist, 6, 171. STONE,W. E., (1955); Acetylcholine in the brain. I. ‘Free’, ‘bound’ and total acetylcholine. Arch. Biochem., 59, 181-192. STUMPF, CH., PETSCHE, H., AND GOGOLAIC, G., (1962); The significance of the rabbit’s septum as a relay station between the midbrain and the hippocampus. II. The differential influence of drugs upon both the septal cell firing pattern and the hippocampus theta activity. Electroenceph. clin. Neurophysiol., 14, 212-219. SZERB, J. C., (1965); Averaged evoked potentials and cholinergic synapses in the somatosensory cortex of the cat. Electroemph. elin. Neurophysiol., 18, 140-146. TOKIZANE, T., KAWAMTJRA, H., AND IMAMURA, G., (1960); Hypothalamic activation upon electrical activities of paleo- and archicortex. Neurol. med. chir., 2, 63-76. Tom, S., AND WIKLER, A., (1965); Effects of atropine on electrical activity of neocortex and hippocampus in cat. Fed. Proc., 24,516 (Abst. No. 2138). TRENDELENBURG, U., (1963); Supersensitivity and subsensitivity to sympathomimetic amines. Pharmacol, Rev., 15,225-216. VALDMAN, A. V., (1961); The Pharmacology of Reticular Formation and Synaptic Transmission. Leningrad, p. 432. VALDMAN, A. V., (1963); Problems ofPharmology of Reticular Formation and Synaptic Transmission. Leningrad, p. 416. VILLARREAL, J. E., AND DOMINO, E. F., (1964); Evidencefor two types of cholinexgicreceptorsinvolved in EEG desynchronization. Pharmologist, 6, 192. WHITE, R. P., AND BOYMY, L. D., (1959); Comparison of physostigmine and amphetamine in antagonizing the EEG effects of C N S depressants. Proc. Soc. exp. Biol. ( N . Y.), 102,47983. WHITE, R. P., AND DAIGNEAULT, E. A., (1959); The antagonism of atropine to the EEG effects of adrenergic drugs. J. Pharmacol. exp. %r., 125, 339-346. WIKLER,A., (1952); Phannacologic dissociation of behavior and EEG ‘sleep patterns’ in dogs.: Morphine, n-allylnormorphine, and atropine. Proc. Soc. exp. Biol. ( N . Y.),79,261-265. YAMAMOTO, K., AND DOMINO, E. F., (1965); Nicotine induced EEG and behavioral arousal. Znt. J. Neuropharmacol., 4, 359-373.
365
Effects of Some Drugs on Aggressive Behaviour and the Electrical Activity of the Limbic System R Y O N O S U K E KIDO, K A l ’ S U M I HIROSE, K E N - I C H I YAMA MOTO MATSUSHITA
AND
AKIRA
Shionogi Research Laboratory, Shionogi & Co., Ltd., Fukushima-ku, Osaka (Japan)
(+)-3-Methoxy-N-methyl- A6-morphinan hydrobromide (175-S), a drug possessing a taming effect, is a compound obtained from sinomenine (an alkaloid of Sinomenium acutum) by Sawa and Tsuji (unpublished) in our laboratory (Fig. 1). Sinomenine, a histamine-releasing compound, has a weak action on the central nervous system. On the other hand, 1 7 5 s has a potent taming effect. In order to determine thesiteand mode of action of the drug on the central nervous system, electroencephalographic and behavioural analyses were performed in animal experiments. The results are also compared with those of the typical central nervous system depressants. A part of data on the effect of 175-S on the motor system is also described here. EXPERIMENTAL
General pharntacological properties
(I) Acute toxicity
LD50 was calculated by Bliss’s method in D/S mice (Bliss, 1938). The oral LD50 was 190.1 mg/kg, the subcutaneous and intravenous LD50’s were 147.5 mg/kg and 68.8 mg/kg respectively.
( 2 ) Spontaneous motor activity Spontaneous motor activity of mice was measured in a light box (Winter & Flataker, 195I ; Dews, 1953) in which interruptions of light beams by a mouse as it moves about were recorded automatically. The 50 increasing dose of 175-S in spontaneous motor activity was calculated by the up and down method, and was 24.5 mg/kg (s.c.). Duration of action was from 3 to 4 h. The 50% increasing dose of morphine was 74.1 mg/kg (s.c.).
( 3 ) Analgesic eflect Analgesic activity was estimated by Haffner’s and D’Amour-Smith’s methods (Haffner, 1929; D’Amour and Smith, 1941). Mice treated with 175-S responded normally by squeaking and turning. Rrfrrcnccs p.T385-387
366
R. K i D o et al.
Fig. 1. Chemical structures of (+)-3-methoxy-N-methyl-d~-morphinan hydrobromide (1 7 5 4 ) (a) and sinomenine (b).
(4) Anticonwlsant activity ( a ) Anti-electroshock.The maximal electroshock convulsant threshold was calculated on a group of 5 mice at three dose levels by the technique of Swinyard et al. (1952). The dose that prevented hindlimb extension in half the mice was measured as the ED50.The ED50 of 1754 was found to be 7.9 mg/kg (s.c.). (b) Antimetrazol, antinikethamide and antistrychnine activity. 1 7 5 s was tested €or its ability to protect mice against convulsion and/or death after subcutaneous injection of 200 mg/kg (s.c.) of metrazol, 400 mg/kg (s.c.) of nikethamide or 2.5 mg/kg (s.c.) of strychnine. The ED50 of 17543 against nikethamide was 2.8 mg/kg. No anticonvulsant activities against metrazol and strychnine were observed. ( 5 ) Potentiated narcosis The minimum anaesthetic effects of thiopental-Na was potentiated by 175s. The pretreatment with 10 mg/kg (s.c.) of 175-S prolonged the duration of the anaesthetic action about three times. (6) Traction test The mouse was hung on a horizontal metal wire by grasping and holding with its fore paws. The sedative EDm was measured as the dose which caused half the mice to fall within 1 min. The ED50 of 1754 was calculated as 9.5 mg/kg (s.c.).
(7) Conditioned avoidance response As a classical conditioned avoidance response, the shuttle box method was used. The rat was trained to move to another box over the hurdle at the sound of a buzzer (conditioned stimulus) in order to avoid an anticipated electroshock (unconditioned stimulus) which the rat had learned to associate with the buzzer. Conditioned avoid-
DRUGS AND AGGRESSIVE BEHAVIOUR
367
ance response was inhibited by 175-S. The ED50 was calculated to be 4.3 mg/kg (s.c.). ‘Sidman’ avoidance experiments (Sidman, 1953, 1955) were carried out in the Skinner type box containing one response lever mounted in one wall and a grid floor for presenting shock to the feet. Each lever press postponed the shock for 15 sec. Whenever the rat failed to press the lever every 15 sec, a brief shock (0.2 sec) was given. At the doses of 1 mg/kg to 2.5 mg/kg (s.c.) the response rates increased. However, at doses of more than 5 mg/kg (s.c.) the response rate decreased. ( 8 ) Maze test Naive male Wister rats weighing 200 g to 250 g were used. A warden multiple-U maze was employed (Warden, 1929). Under a 23 h feeding rhythm each rat was trained to reach the food box from the start box. The trials per day were performed until the rat rached a criterion of 20 consecutive errorless runs on two successivedays. The total running time from the raising of the door in front of the start box to entering the goal box was measured. At the dose of 5 mg/kg (s.c.) errors were increased, but running time was not prolonged.
(9) Taming efect Two hamsters were placed in the same cage and were exposed at the same time to an electric shock through the bottom of the cage. They attacked each other. This kind of fighting behaviour was inhibited by 175-S. The anti-aggressive effect of 1754 was also compared with the inhibitory effect of the defensive behaviour in the Japanese monkey, using the rating scale of Kawai and Tsumori (unpublished). The rating scale was used in the following 4 situations: (1) Observer approaches the monkey cage. Mild aggressive without defensive behaviour is usually seen; (2) Observer threatens the monkey. In this situation, aggressive behaviour is dominant and defensive behaviour rather weak; (3) Observer threatens the animal with a stick. The rate of aggressive and defensive behaviour is equal; (4) The animal is shown a monkey pelt. In this situation, monkeys show instinctive fear of the monkey’s pelt. Accordingly, defensive behaviour is seen dominantly. To score measures of aggressive and defensive behaviour, an increase in score from 0 to 3 is used. The items of observation for aggressive and defensTABLE I S C O R I N G SHEET FOR A G G R E S S I V E A N D D E F E N S I V E B E H A V I O U R OF T H E M O N K E Y
Aggressive behaviour
1. Pounce upon 2. Scratch (or shake off) 3. Swing pearch or net 4. Swing the shoulder back and forth 5. Bring down the shoulder back and forth 6. Open mouth (baring teeth) 7. Flattened ears and open eyelids 8. Glaring 9. Threating bark References p . 385-387
Defensive behaviour
1. Escape (locomotion) 2. Bend backward 3. Turn face away 4. Turn eyes away 5. Crouching 6. Tearful face 7. Bend backward and open eyes widely 8. Turn eyes away and shake off 9. Shriek
368
R. K I D O et al.
1
10
O.lmg/kg
1
10
0.8 me/@
Fig. 2. Anti-aggressive and antidefensive effects of 17543 when the animal was threatened with a bar. Ordinate indicates the mean value of aggressive or defensive score measured by the rating scale of Kawai and Tsumori in 6 Japanese monkeys, and abscissa indicates the time course.
ive behaviour are as shown in Table I. Fig. 2 shows an example of the score curve in the situation where the monkey is threatened with a bar. The black line indicates the score curve of aggressive behaviour and the dotted line that of defensive behaviour. In each dose, aggressive behaviour was inhibited more markedly in comparison with defensive behaviour. The onset of the action was fast, and the duration of action was about 6 h with a larger dose. Fig. 3 shows another example of the score curve in the situation where the monkey pelt was shown. After a larger dose of 175s aggressive behaviour was more markedly inhibited than defensive behaviour in the 4 situations stated above. These effects on the central nervous system and other pharmacological properties are summarized in comparison with other central nervous system depressants in Table 11.
P
TA B LE I1 PHARMACOLOGICAL PROPERTIES OF
Methods
Animals
1. Acute toxicity 2. Spont. mot. activ. 3. Analgesic effect
4. Anticonvulsant activity (a) Antielectroshock (b) Antimetrazol (c) Antinikethamide (d) Antistrychnine 5. Potent. narc. 6. Traction test 7. Cond. avoid. resp. 8. Maze test 9. (a) Agressive behaviour (b) Aggressive behaviour
-,
without activity;
Mouse Mouse Rat Mouse Mouse Mouse Mouse Mouse Mouse Mouse Rat Rat Hamster Monkey
175-s
I N C O M P A R I S O N W I T H OTHER C E N T R A L N E R V O U S SYSTEM DEPRESSA NTS
175-5' LD5o i.v. 68.8 mg/kg oral 190.1 mg/kg
+ + i+++ +++ ++ + ++-+++
Chiorpromazine
U
++
(4)
+ + + (4) Prolong latency + ($1
+ + (4)
+, +, + +, + + +, slight to strong activities; (J), inhibitory effect.
Methyl-Hexabital
z
+ +++
Increase error
+++ (4) +++ ($1
Morphine 350 mg/kg oral 945 mg/kg
LD50 i.v.
+++ +++
+++ +++ (4) Prolong latency ++ (4) + (4)
* * (4) -- * ($1
Increase error
R. K I D O et
370
al.
-Aggressive behaviour (34) ---- Defensive behaviour
(700) 60 123 180 min
__ -- -.’ 60 120 180min
min
Fig. 3. Anti-aggressive and antidefensive effects of 1 7 5 4 when the animal was threatened with a pelt of a monkey. Ordinate indicates the mean value of aggressive or defensivescore measured by the rating scale of Kawai and Tsumori in 6 Japanese monkeys, and abscissaindicates the time course.
EHect on EEG and behaviour
The cat, dog and monkey were prepared for the placing of indwelling brain electrodes using a modification of conventional techniques (Yamamoto, 1959). Both sexes of cats weighing about 3 kg, mongrel dogs weighing about 10 kg and rhesus monkeys (Macaca mlatta) weighing about 4 kg were used. Experiments were performed on 15 cats, 7 dogs and 6 monkeys utilizing a cross-over design in Latin square for drug administration at 2 week intervals.
371
D R U G S A N D AGGRESSIVE BEHAVIOUR
TABLE 111 EFFECTS OF
175-S
O N BEHAVIOUR IN THE CAT, DOG AND M O N K E Y ~
cut 0.1-0.5-1 .o mg/kg i.v.
General behaviour Narcosis Sleep Sedative Alertness Excitement Rage Emotional Disappearance of friendship Fear Disappearance of aggression Disappearance of attention Somatic Skeletal muscle Disturbance of gait Hypersensibility Convulsion Increase of respiration Autonomic Eyelid Pupil Salivation Duration of action
+ ++ + +
-
2-4
mgfkg S.C.
0.1-0.25 mgfkg i.v.
0.5-4.0
mgikg i.v.
0.5-0.7 mglkg i.v.
f
++ + + +
+ ++ ++
+
+
?
Appearance
Appearance
Appearance Appearance (f +) (+)
+
Monkey
Dog
(+)
++ +
(+I
+
f
+ +
+
&
-
Rigidity
Rigidity
Rigidity
+ +
++ -
+ -
It-+
f Open
(+I
mydriases
(+I
f
200-300 min
?
(+I
(+I
Tremor
Tremor
(+I)
(+)
(+I
+
-
+
+
Open
Open
Mydriasis
Mydriasis
Mydriasis
(+ +)
+
(+)
+ Open
(++I
(++)
(+I
f
> 20h
170min
(+)
-
300min-> 1 7 h 150-180min
( I ) Electroencephalographic and behavioural alterations caused by 175-S in animals with chronic indwelling brain electrodes ( a ) Cat. Behavioural observation was noticed from general, emotional, somatic and autonomic activities. The dominant consciousness level observed in the general behaviour due to 0.1 to 1 mg/kg (i.v.) of 17543 was a sedateness. Functional changes in somatic activity scuh as an appearance of moderate rigidity in the hindlimbs, a disturbance of gait, an increase in skin sensibility to touch (hypersensitiveness)and an increase in respiratory rate were observed. Friendship to an observer or an aggressive References p . 385-387
372
R. KIDO et
al.
Fig. 4. Effect of 175-S (1 mg/kg i.v.) on the EEG in the chronic cat. The upper two traces show the control EEG in the waking and the slow wave deep sleeping state respectively. The lower left (33 min) shows a typical depression of the hippocampal @-wavesdissociated from the neocortical desynchronization caused by the drug. The cat shows a sedate behaviour. The lower right (103 min) shows much increase in the spike discharm in the amygdala and the hippocampus in a later period after the drug. TheeyelidiswidelyopenedeveninthisEEGstate.L.CM, left ‘centremedian’nucleus of the thalamus; L.AMY amygdala; L. HIP., hippocampus; L.ANT.SIG., anterior sigmoid gyrus; L.ECT.SYLV., cctosylvian gyrus; L.LAT., lateral gyrus; RESP., respiratory movement.
hehaviour in untamed cats evidently disappeared. On the contrary, an appearance of fear towards an observer was enhanced before injection of the drug. The eyelid was also widely opened, and marked mydriasis and a slight salivation were noticed. These effects continued for about 5 h. With a larger dose of 1 7 5 4 (2 to 4 mg/kg s.c.), cats initially showed a behavioural excitation such as trying to escape from the cage or a biting of the lead wire for the EEG. Cats then lay on the floor, and behavioural sedateness gradually occurred. These behavioural characteristics caused by the drug were enhanced by an increase in the dose (Table TIT). In a moderate dose of 175-S (0.5 to 1
D R U G S A N D AGGRESSIVE BEHAVIOUR
373
Before
L L ECT. SYL
b
b
L LAT
RESP
Mouse
93 min after 1755 0.5 rng/kg i.v.
L
I
Mouse
-
5 0 y V 1sec
Fig. 5. EtTect of 175-S (0.5 mg/kg i.v.) onthe EEG arousal reaction and behavioural attentionreaction caused by the showing of a mouse to the chronic cat. The upper trace shows the appearance of the reaction in the control. The lower trace shows an abolition of the reaction after the drug.
mg/kg i.v.), the most marked change in EEG was an initial disappearance of the hippocampal &waves dissociated from the neocortical desynchronization. About 90 min after injection, multiple spike dischargesin the amygdala and the hippocampusaccompanied by neocortical slow waves were observed. These sleep patterns were more References p . 385-387
374
R. KIDO et al.
evident than those of the normal slow wave sleeping state, however, the eyelid of cats was usually widely opened even in this electroencephalographic condition (Fig. 4). Furthermore, cats completely lost a behavioural attention reaction and an EEG arousal reaction to the showing of a mouse (Fig. 5). (b) Dog. In a smallerdose of 171s(0.1 to 0.25 mg/kg i-v.), a tendency of behavioura1 responses which were seen in general, emotional, somatic and autonomic activities was almost the same as those of the cat (Table 119. Jn a larger dose of the drug (0.5 to 4.0 mg/kg i.v.), however, sedate behaviour was gradually reduced by an increase of the dosage and excitation in general behaviour occurred. The main consciousness level judged from the behaviour was wakefulness, and no sleeping behaviour was seen
Fig. 6. Effect of 1 7 5 4 (0.7 mg/kg iv.) on the EEG m the chronic monkey. The upper two traces ghow thecontrolEEGinthewakhgand thedrowsy states. Thelower left (15 min) shows typical high voltage fast waves dissociated from the neocorticaldesyncbnization pattern caused by the drug. As shown in the lower right (45 min) the slow wave sleep pattern was not so marked as in a later period after the drug injection; L.ANT.FRONT., left anterior part of the frontal cortex; L-PARET., parietalcortex; L.TEMP., temporal cortex.
D R U G S A N D AGGRESSIVE B E H A V I O U R
375
through the experimental procedure. In a moderate dose of the drug (0.25 mglkg i.v.), the hippocampal &waves changed to high voltage fast waves dissociated from the neocortical desynchronization pattern. The electroencephalographic slow wave deep sleep state was never seen corresponding to its awaking behaviour. (c) Monkey. Intravenously administered 1754 (0.5 to 0.7 mg/kg) caused the same tendency of the behavioural responses with those in the cat and the dog. Particularly, an aggressive behaviour was completely abolished at this dosage. However, a fear-like response to external stimulation was still maintained. The main consciousness level judged from the sedateness and sleeping tendency was less than that seen in the cat. At this dosage, high voltage fast waves in the hippocampal lead, dissociated from the neocortical activites, were observed as in cats and dogs. At the same time, an occurrence of the slow wave sleeping pattern was markedly reduced as compared with the control pattern (Fig. 6). The results are summarized in Table 111.Electroencephalographicand behaviovral responses of monkeys on 175-S seem to be intermediate between those of cats and dogs. ( 2 ) Electroencephalographic analysis on the effect of 175s in acute cat preparation Thirty-eight acute cats were used for analysis of the mode of action of 175-S on the central nervous system. ( a ) Alteration of EEGpattern in acute cats. Immediately after an injection of 1753 (0.5 mg/kg i.v.), evident high voltage slow waves and spindle bursts in the neocortex, and multiple spike discharges in the hippocampus and the amygdala, were observed. This type of pattern was not seen in the control of acute cats. About 100 min after the injection, neocortical activities came to show a desynchronization; multiple spike discharges still continued, however, in the amygdala and the hippocampus. (b) Electroencephalographic analysis of the stimulation experiments. As reported by Tokizane et al. (1960), the anterior and the posterior part of the hypothalamus showed an EEG activation mechanism on the amygdala and the hippocampus respectively (hypothalamus activating system). Stimulation of the preoptic area. High frequency stimulation (100 c/sec, 1 msec) of the preoptic area produced a desynchronizationpattern in the amygdala and suppressed the hippocampal &waves. The mean value of threshold elevation showed no significant change after 1754 (119 %). Stimulation of the posterior part of the hypothalamus. High frequency stimulation of the posterior part of the hypothalamus produced an EEG arousal reaction in the amygdala, the hippocampus and the neocortex. After an administration of 175-S, dissociation of threshold elevation between the hippocampus and the neocortex occurred. Namely, the stimulation threshold both for the hippocampus and the neocortex was almost the same before the drug injection. After 1754injection,the threshold for the hippocampal activation was significantlyelevated (132 %); on the contrary, the threshold for the neocortical activation did not show significant elevation (119%). Stimulation of the reticular activating system. High frequency stimulation of the ‘centre median’ nucleus of the thalamus or the mesencephalic reticular formation produced the EEG arousal reaction both in the neocortex and the hippocampus. These References p . 385-387
376
R. KIDO et al.
Before L.CM
Central gray matter, 1 0 0 C'ESec, Imsec
L.VPL
L.AMY
L .ECT. SYL
1.75 V
50min after 175s 0.5 mg/kg 1.v.
I
I
I
I
2.4 V
5oJJvisei
Fig. 7. Effect of 1754 (0.5 mg/kg i.v.) on the EEG arousalreaction in the acute cat. The threshold for an appearance of the hippocampal 0-waves caused by an electrical stimulation of the central gray matter was markedly elevated by the drug.
thresholds did not show significant changes either in the thalamic (1 18 %) or the reticular stimulation (124%). Stimulation of the central gray matter. High frequency stimulation of the mesencephalic central gray matter produced the EEG arousal reaction both in the hippo-
P
h
c v
TA BLE IV Effects of 1754 and various CNS depressants on EEG arousal reaction in the neo-, paleo- and archicortical system (Tokizane et al., 1960) and seizure discharges in the acute cat. Effects of drugs were indicated by the percentage change of stimulation threshold as shown in the text Parameters
Anterior hypothalamus Posterior hypothalamus Reticular formation Central gray matter Septum Amygdala
-+
+ + + -+
--f
Amygdala Hippocampus Neocortex Hippocampus Seizure discharge Seizure discharge
Morphine Chlorpromaziue Meprobamate 5 mglkg i.v. 3 mglkg i.v. 30 mglkg i.v.
- 125% -+ 139
- 122 -+
135
- 100 - 110
160% 185 -+ 141 -+ 133 98 99 -+
-+
-
- 113% - 112 137 -+ 144 - 100 140 -+
-+
175-S Phenobarbital Chlordiazep30 mglkg i.v. oxide 20 mglkg 0.5 mglkg i.v. per 0s
- 125%
149 186 151 - 114 132
- 98% - 112 - 114 141 - 112 - 100
- 119% -+
132
- 124 -+
135
- 100 I-
-,
90-125%;
+, 126-200%.
0 C
d
w
4 4
378
R. K I D O
et al.
campus and the neocortex. As shown in Fig. 7, 175-S depressed this EEG arousal reaction. The mean values of the threshold elevation were 135% for the neocortex and 134% for the hippocampus. Stimulation of the caudate nucleus. A single shock of 1 msec pulse delivered to the caudate nucleus induced a spindle burst (caudate spindle) in the ipsilateral somatosensory cortex and the other newortical sites (Buchwald et al., 1961). The stimulation threshold for the induction of a spindle burst did not show significant change (1 07 %) after 1754. Stimulation of the centre median nucleus of the thalamus and the caudate nucleus. The recruiting response was induced by low frequency stimulation (8 c/sec, 1 msec) of the ‘centre median’ nucleus of the thalamus and the caudate nucleus. The threshold of the recruiting response was not significantly altered by 175-S either in the ‘centre median’ nucleus (106 %) or the caudate nucleus (104 %). Stimulation of the venfro-postero-lateral nucleus of the thalamus. Low frequency stimulation of the specificnucleus of the thalamus caused an augmented response in the limited neocortex and the other related structures. The threshold for the production of the augmented response slightly decreased after an injection of 17545. Stimulation of the amygdala. Stimulation of the amygdala (40c/sec, 3 msec) caused a seizure discharge in the hypothalamic and limbic areas. The threshold for the production of a seizure discharge was not altered by 1 7 1 s (100%). The results are summarized in Table IV. As seen in the table, 1754 seems to have an inhibitory effect on the activation structures related with an occurrence of the hippocampal 0-waves such as the posterior part of the hypothalamus and the central gray matter. (3) Behavioural analysis on the efect of 175-S in cats with chronic indwelling brain electrodes As reported by many authors, electrical stimulati4n of various sites of the brain produced many characteristicbehavioural changes (Kaada; 1951;Kaada and Bruland, 1960; Kimura, 1960; Tokizane, 1958a; Yamamoto and Kido, 1964; Kido et al., 1966). These behavioural changes (‘induced-behaviour’) were therefore used as an indicator in the analysis of the mode of action of 175-S on the cebtral nervous system, particularly the inhibitory effect of the drug on an aggressive.behaviour. The sites and the parameters of stimulation in the brain were exactly the same as those which produced the EEG arousal reaction, recruiting response and seizure discharge in the previous section. The dose of 175-S was 0.5 mg/kg (i.v.). Stimulation of the preoptic area. High frequency stimulation of the preoptic area caused an induced-behaviour, such as an olfactory and a searching response, during the stimulation. The threshold to induce these responses was not significantly changed by 175s (123%). Stimulation of the posterior part of the hypothalamus. High frequency stimulation of the posterior part of the hypothalamus produced behavioural changes such as mydriasis, piloerection and searching response at a lower voltage, and running and rage at a higher voltage. 175.9did not show an inhibitory effect either on the searching response
w aJ
e
TABLE V
LU
2
&v
EFFECTS OF
175-S AND
THE VARIOUS
CNS
D EP R ES S A N TS O N T H E I N D U C E D BEHA V IOU R S C A U S E D B Y S TIM ULATI ON OF V A R I O U S SITES OF B R A I N I N THE C H R O N I C C A T
Parameters
Morphine Chlorpromazine Meprobamate 5 mglkg i.v. 3 mglkg i.v. 30 mglkg i.v.
Phenobarbital Chlordiazep30 mglkg i.v. oxide 20 mglkg per 0s
17.5-S
0.5 mglkg i.v.
-
Anterior hypothalamus
Mydriasis Sniffing Searching response
Posterior hypothalamus
Searching response Rage Mydriasis, piloerection Somatic movement
Reticular formation Central gray matter
Mydriasis, piloerection Head-turning Hissing
?%
-
99 143*
+l80%
- 106
- 104
+
150%
+ 167%
+
+ 165
146
+ 141 + 160
- 106 + 135
- 122 - 115
-f
175
- 100
- 101 - 114 + 130
+
146%
- 115
+
138
+ 132
- 117
- 101
- 108
+ 183
- 117 + 140
- 100 + 159
+ 137
- 112
163
- 110
?. 205*
+ 108
Septum
Sniffing, salivation Searching response Feeding response
- 100
- 105
- 105
Amygdala
Sniffing, salivation feeding response
- 120
- 107
+ 130
-f
- 123%
- 107
380
R. KIDO
et al.
(122 %) or rage (1 15%). Stimulation of the mesencephalic reticular core. High frequency stimulation of the reticular formation caused behavioural changes such as mydriasis, piloerection and somatic movement. This induced behaviour was not altered by 175-S (101 %). Stimulation of the central gray mutter. High frequency stimulation of the central gray matter produced behavioural changes such as mydriasis, piloerection, headturning and hissing. This induced behaviour, particularly hissing, was significantly inhibited by the drug (130 %). Stimulation of the umygdala. Stimulation of the amygdala caused induced behaviour such as mydriasis, salivation, facial spasm and feeding response. This induced behaviour was not suppressed by 175-S (107 %). The results are summarized in Table V. The most evident and limited effect of 175-S on the induced behaviour in cats was an inhibitory action on the hissing induced by stimulation of the central gray matter.
E$ects on the motor system The usual signs seen after subcutaneous injection of 2 mg/kg of 1754 are a slight increase in sensitivity to a mechanical stimulation applied to the skin, over-extension of fingers ,muscle stiffness with weak extension of the upper limbs and intense flexion of the lower limbs, persistence of the tonic neck reflex and absence of the righting reflex, disturbance of the sense of equilibrium, sometimes yielding tremor, apparent ataxia and akinesia. Firstly, the effect on the spinal reflex discharges was studied on anaesthetized, intercollicularly decerebrated and spinal cats. The reflex responses elicited by the tibial or common peroneal nerve stimulation were recorded from the L7 or S1 ventral root. When muscle nerves were cut or completely ligated, in other words when the activity of the y-motor system was eliminated from the spinal reflex arcs to be studied, an intense decrease in the monosynaptic reflex (MSR) and polysynaptic reflex (PSR) response was usually observed in anaesthetized and spinal cats. The result with decerebrate animals was unstable and revealed a quite different feature as compared with those of two other preparations. The activation of MSR activity was observed after an intravenous injection of 1 mg/kg of 1754 in 3 of 4 decerebrate cats. When the y-loop was intact, the elevation of MSR response became more appreciable, accompanied by a moderate increase in the frequency of muscle spindle discharge. However, the muscle spindle activity seemed to attain maximal effect sooner than MSR activation. In addition, during the period that showed a maximal increase in the frequency of muscle spindle discharge, MSR activity was rather depressed. As stated above, there was quite a difference in the EEG pattern between unrestrained and restrained animals after an injection of 175-S. It will be reasonable to consider that this EEG difference may be reflected in the motor activity, so an experiment on spinal reflex activities of the unrestrained animal became desirable. For this purpose, two collar type electrodes (Pompeiano and Swett, 1962; Kubota et aZ., 1965), composed of two short 300 p silver wires and a rectangular piece of vinyl tape, were attached to the trunk of the sciatic nerve and the tibial or common peroneal nerve. The former was
381
D R U G S A N D AGGRESSIVE BEHAVIOUR
used for stimulation and the latter for recording the MSR. When muscle nerves were tightly ligated at their peripheral end in order to avoid a muscle contraction and to obtain a stable reflex response, an intravenously or subcutaneously injected 175-S caused a significant fall in MSR activity. When muscle nerves were intact, on the other hand, a slight and transient depression of the MSR, which lasted for about 40 min, was followed by a slight increase in the response. Although no ligation was performed on the nerve, its impulse conduction might be more or less interrupted by attaching an electrode to it, because the gait of animals was apparently abnormal and the cat dragged its operated leg when walking. Therefore, the loop on the reflex activity to be studied is not in a complete state here. Two other methods for evaluating the muscle tonus without any injury to the reflex arc are (a) observation of the stretch reflex activity of the extremities by electromyography and (b) the mathematical analysis of discharge patterns of the neuromuscular unit (Tokizane and Shimazu, 1964). 1 7 5 s failed to reduce the stretch reflex response of the gastrocnemius muscle of the decerebrate cat, but rather slightly enhanced this activity. In the intact cat, the neuromuscular unit analysis revealed that this compound apparently elicited muscle rigidity in the skeletal muscle of the hindlimbs, because 175-S produced a significant increase in the stretch reflex of the tibialis anterior and the gastrocnemius muscle, particularly
300 Control
a 175-5 l.Omg/kg i.v.
0
0
20 0
Fig. 8. Effect of 1754 on the pattern of neuromuscular unit dischargesrecorded from the gastrocnemius muscle of the cat. Abscissa and ordinate indicate the mean value of discharge interval and the standard deviation (S.D.) respectively. Open circles are obtained in control and filled circles in the ataxic state induced by 1 mg/kg (i.v.) of 175-S.The ataxia usually attainedits peak within 15 min after injection of this dose.
of the flexor muscle, and many tonic neuromuscular discharges were easily and stably obtained. After drug administration, the mean values of discharge intervals plotted against their standard deviations were clustered below the control distribution, and almost all points were concentrated between 50 and 80 msec in regard to their mean values of discharge intervals (Fig. 8). It was assumed from these results that in spite of a slight depression of the phasic motor mechanism, which was inferred from a deReferences p. 385-387
382
R. KIDO et al.
crease in MSR activity, the tonic motor function was rather increased during the ataxia by 175s. DISCUSSION
Electroencephalographicand behavioural analysis of 17543, which showed a dramatic inhibitory effect on an agressive behaviour in wild monkeys, was carried out to determine the site of action on the central nervous system in experimental animals such as cats, dogs and monkeys. In smaller doses of 1754, which failed to show an evident inhibition on consciousness level of the experimental animals, an abolition of the attention reaction caused by showing a mouse, friendship to an observer and aggressivebehaviour for external stimulation was observed in each animal. These kinds of behavioural changes were also observed after treatment with so-called tranquillizing drugs such as morphine, chlorpromazine or chlordiazepoxide (Yamamoto and Kido, 1964). On the other hand, the most evident alteration seen in EEG due to 175-S was an occurrence of the depression in the hippocampal &waves which were dissociated from the neocortical desynchronization in each animal. This characteristic change in the hippocampal activities should be an 'arousal pattern at lower level' as described by Imamura and Kawamura (1962), and was also observed in the meprobamate and chlordiazepoxide treated animals (Yamamot0 and Kido, 1962a, 1964). This kind of depression in the hippocampal activities was a common change seen in animals which abolished the attention reaction, friendship or aggressive behaviour by treatment of central nervous system depressants including 1754. For example, in chlorpromazine or morphine treated cats, in spite of a higher level of the EEG, the behaviour showed a sedateness, and the threshold of the EEG arousal reaction or behavioural attention reaction caused by external stimulation was markedly elevated. With barbiturate, on the contrary, depression of the attention reaction was finally abolised by a narcotic dose level (Yamamoto and Kido, 1962a). No narcosis was obtained even by a large dose of 1754 in each animal, and this might be one of the characteristics of the psychotropic drugs. The pharmacological meaning for the changes in emotional behaviour or in hippocampal lead due to central nervous system depressants is not yet clarified. However, the following explanations have been described : an excitation of the inhibitory area for the hippocampal structures (Preston, 1956; Yamamoto and Kido, 1964, for chlorpromazine);an inhibition of the hypothalamic activating system (Yamamoto et al., 1961, for morphine; Yamamoto and Kido, 1964, for morphine, chlorpromazine and chlordiazepoxide); an inhibition of the central gray matter (Yamamoto and Kido, 1962b, for chlordiazepoxide) and an inhibition of the limbic system (Randall, 1961, for chlordiazepoxide; Schallek and Kuehn, 1960; Clark, 1961, for chlordiazepoxide and meprobamate). In a later period after 1754, dissociation between EEG and behaviour such as evident slow waves in all EEG leads was noticed in the waking state. The mechanisms of dissociation between EEG and behaviour are still unknown, including those brought about by atropine (Wikler, 1952), ether (Yamamoto and Kido, 1964) and fast wave
D R U G S A N D AGGRESSIVE AEHAVIOUR
383
sleep (Dement, 1958; Jouvet, 1961). In the present work, however, it seems to be important to analyse the relationship between the effect of 1754 and the function of the hippocampus, as a higher integration centre of memory mechanism and attention reaction (Tokizane, 1958a; Niki, 1962). On the other hand, species specificity for 1754 was observed in the response of EEG and behaviour. The most evident sedate and sleep responses due to 1754 were obtained in cats; on the other hand, continuous excitation behaviour was obtained in dogs. The monkey showed an intermediate effect between those with cats and dogs. In comparing of the action of 1753 with the other central nervous system depressants, however, differences in species specificity were noticed. As reported, earlier, species specificity caused by central nervous system depressants was frequently observed (Yamamoto and Kido, 1964). For instance, electroencephalographic and behavioural sleep patterns were obtained in chlorpromazine-treated dogs and monkeys; on the other hand, ‘rage-like behaviour’ was noticed in cats with the same drug (Kido and Yamamoto, 1962). With morphine, cats showed a continuous awaking pattern, but dogs showed narcosis. Thus, the action of 1754 was apparently different from those of chlorpromazine and had completely opposite effects between cats and dogs compared with those of morphine. Differences in the EEG patterns between acute and chronic cats as in 1754 were also observed with morphine, chIorpromazine and chlordiazepoxide. In these drugs, in general, much increase in slow waves in all EEG leads was noticed as compared with those in chronic experiments. This phenomenon seems to be one of the characteristic actions of central nervous system depressants that do not show a narcotic action. Electroencephalographic analysis was carried out in order to determine the site of action of 175-S in acute cats, and the results were compared with those of the other central nervous system depressants such as morphine, chlorpromazine, meprobamate, sodium phenobarbital and chlordiazepoxide. As shown in Table IV, the elevation of the stimulating threshold in the activating system concerned with an appearance of the hippocampal &waves, such as the posterior part of the hypothalamus and the central gray matter (Tokizane et al., 1960; Kawamura et ul., 1961), was noticed with morphine, chlorpromazine, chlordiazepoxide and 1753. On the contrary, the threshold of the neocortical EEG desynchronization due to the reticular and the posterior hypothalamic stimulation was not inhibited by morphine, chlordiazepoxide or 1754. From this point of view, the action of 175-S on the central nervous system seems to be electroencephalographically similar to those of morphine and chlordiazepoxide. On the other hand, 175-S inhibited ‘hissing’ induced by a stimulation of the central gray matter in cats. It was difficult to explain the direct relationship between inhibition by 175-S of aggressive behaviour and of hissing. As shown in Table V, however, elevation of the thresholds of ‘rage’ or ‘hissing’ induced by the posterior hypothalamic or the central gray matter stimulation were at least obtained after administration of morphine, meprobamate and chlordiazepoxide, which evidently inhibited spontaneous aggressive behaviour in animals (Clark, 1961; Yamamoto and Kido, 1964). On the other hand, in spite of the inhibitory action of the aggressive behaviour, fear towards an observer was not abolished by 17% in cats and monkeys. The fear response for References p. 385-387
384
R. KIDO
et al.
external stimulation was also inhibited by chlorpromazme, meprobamate and chlordiazepoxide (Schallek and Kuehn, 1960; Clark, 1961; Yamamoto and Kido, 1962a, 1964). The amygdala has been described as a higher integration centre of fear (Tokizane, 1958b; Gloor, 1960). The stimulation threshold in the preoptic area to the amygdala was elevated by the above drugs in behavioural observation except 1 7 5 s (Table V); on the contrary, it was not elevated in the electroencephalographic study except by chlorpromazine(Table IV). The behaviourally analysed results might be an explanation of a lack of effect of 1753 on this site; in other words, an aggressive behaviour and a fear behaviour seem to be essentially independent in their origins. Dissociation of the threshold elevation between EEG and behaviour, however, has not yet been observed as clearly as those seen between spontaneous EEG and behaviour. Besidespsychotropicactions, all neuroleptics exert some action on the motor system, particularly on the extrapyramidal system. The motor syndromes inducCd in animals by neuroleptics are described in many reports as follows: Parkinson-like behaviour, catalepsy, muscle rigidity, tremor, akinesia, etc. Apart from the question whether showing some effects on the motor activity is an indispensable property for a neuroleptic, there are several reports that have discussed the chlorpromazine or reserpine syndrome in animals from an electrophysiological point of view. Many phenothiazinederivatives, represented by chlorpromazine, produce a cataleptic behaviour in animals (Courvoisier et al., 1957). Henatsch and Ingvar (1956) observed an intense depression of y-motor activity with chlorpromazine in cat experiments where actions of the drug were compared on two types of decerebrate rigidities, a- and y-rigidity. Bando et ul. (1960a, b) supposed that the central origin for the depression of y-motor activity of the cat by chlorpromazine might be firstly the posterior part of the hypothalamus and secondly the mesencephalicreticular formation. However, during a cataleptic behaviour which reached a maximum several hours after an administration of a larger dose of chlorpromazine, an increase in muscle tonus inferred from an elevation of stretch reflex activity of the extremities was observed in the cat and dog (Matsushita et ul., in preparation). On the other hand, Steg (1964) performed:a skillful and precise experiment on tail muscle innervation of the rat and analysed the motor syndrome of reserpine with this preparation. According to Steg’s report, the y-system was strongly depressed by reserpine; however, the a-motor function, especially a system innervating the slow muscle unit, was somewhat enhanced. Although the relationship between behavioural change and electrophysiological results was discussed in this report, the analyses of the long-lasting behavioural change by this drug seems to be insufficient. In contrast with chlorpromazine or reserpine, 175-Sbelongs to another category of neuroleptics. In spite of the monosynaptic reflex depression, 1 7 5 s enhanced tonic stretch reflex activity. This type of action is similar to ether rigidity in some degree (Shimazu, 1962). An increase in y-motor activity is inferred from [these results; however, some supraspinal influences on the a-system at present being investigated cannot be neglected.
D R U G S A N D AGGRESSIVE BEHAVIOUR
385
SUMMARY
(I) Pharmacological properties of 1 7 5 s were investigated. 1 7 5 s has a central nervous system depressant effect. Its taming effect is especially characteristic. (2) Electroencephalographic and behavioural analysis on the taming effect caused by 175s was performed in cats, dogs and monkeys. (3) Species specificity in EEG and behavioural response was observed in 1754; a sedateness and sleep in cats, excitation in dogs and intermediate response between cats and dogs in monkeys. No narcotic action was obtained by a large dose of the drug in each animal. (4) The characteristic change in EEG was the inital inhibition of the hippocampal &waves dissociated from the neocortical desynchronization. Multiple spike discharges in the amygdala and the hippocampus were observed corresponding to the neocortical slow waves in later periods, with 175s. ( 5 ) An appearance of the hippocampal &waves due to stimulation of the posterior hypothalamus and the central gray matter was inhibited by the drug. The hissing induced by a stimulation of the central gray matter was also inhibited by the drug. (6) 175-5 intensely depressed the spinal monosynaptic and polysynaptic reflex responses of anaesthetized and spinal cats. The effect on decerebrate cats, however, was rather complicated and a moderate activation of the reflex response, which was preceded by an incerase in muscle spindle discharges, was sometimes observed. This compound also operated its depressive effect on the monosynaptic reflex activity of unrestrained and unanaesthetized cats. (7) Intercollicular decerebrate rigidity manifested on the hindlimb, especially on the flexor muscle of the foot, turned into a stiffer state after an injection of 175-S. An electromyographicstudy revealed that many motor units of the gastrocnemius and the tibialis anterior muscle showed tonic activities during the ataxia by 175s. REFERENCES BANW, T., SAKAI,Y.,SAGAWA, Y.,AND KUBOTA,K., (1960a); Pharmacological studies on muscle tone. (II) Functional differentiationof the y-system and chlorpromazine. Foliu pharmacol. jap., 56, 0 109 (In Japanese). BANDO,T., SAW, Y., SAGAWA, Y.,M ~ U Y A S UK., , AND S m u , H., (1960b); Pharmacological studies on muscle tone. (I Effects ) of chlorpromazine and nembutal ony-system. Foliapharmucol. jup., 56, 0 34 (In Japanese). BLISS,C. I., (1938); The determination of the dosage-mortality curve from small numbers. Quurt. J. Pharmacol., 11, 192-216. BUCHWALD, N. A., WYEM, E. J., OK-, T., AND HEUSER,G., (1961); The 'caudate-spindle'. I. Electrophysiological properties. Electroenceph. elin. Neurophysiol., 13, 509-518. CLARK,W. G., (1961); Electrophysiological correlates of chlordiazepoxide.Dis. nerv. Syst., 22, 16. COURVOISIER, S., DUCROT, R., AND JULOU,L., (1957); Nouveaux aspects exp6rbentaux de l'activit6 centrale des ddrivtk de la phdnothiazine. Psychotropic Drugs. S . Garattini and V. Ghetti, Editors. Amsterdam, Elsevier, pp. 373-391. D'AMOUR,F. E., AND S m , D. L., (1941); A method for determining loss of pain sensation. J. Pharmacol. exp. Ther., 12,7479. DEMENT,W., (1958); The occurrence of low voltage, fast electroencephalogram patterns during behavioral sleep in the cat. Electroenceph. elin. Neurophysiol., 10,291-296. DEWS,P. B., (1953); The measurement of the influence of drugs on voluntary activity in mice. Brit. J. Pharmacol., 8, 46-43.
386
R. K I D O
et al.
GLOOR,P., (1960); Amygdala. Handbook of Physiology. Neurophysiology, 2. Amer. Physiol. Soc., Washington, D. C. pp. 1395-1420. F., (1929); ExperimentellePriifung schmerzstillender Mittel. Dtsch. med. Wschr., 55, 731HAFFNER, 733.
HENATSCH, H.-D., AND INGVAR, D. H., (1956); Chlorpromazine und Spastizitat. Eine experimentelle elektrophysiologische Untersuchung. Arch. Psychiat. Nervenkr., 195, 77-93. IMAMURA, G., AND K A w m m , H., (1962); Activation pattern in lower level in the neo-, paleo-, and archicortices. Jap. J. Physiol., 12,494-505. JOUVET, M., (1961); Telencephalic and rhombencephalicsleep in the cat. Nature of Sleep, Ciba Found. Symp., pp. 188-206. KAADA, B. R., (1951); Somatomotor autonomic and electrocorticographic responses to electrical stimulation of ‘rhinencephalic’ and other structuresin primates, cats and dogs. Actaphysiol. scand., 24, Suppl. 3. KAADA,B. R., AND BRULAND, H., (1960); Effect of chlorpromazine on the ‘attention’ (orienting), fight and anger response elicited by cerebral stimulation. Acta physiol. scand., 50, Suppl. 175. QWN, M., AND TSUMORI, A., Unpublished data. QWAMURA, H., NAKAMURA, Y.,AND TOKIZANE,T., (1961); Effect of acute brainstem lesions on the electrical activities of the limbic system and neocortex. Jap. J. Physiol., 11, 564-575. Kmo, R., AND YAMAMOTO,K., (1962); An analysis of tranquilizers in chronicallyelectrode implanted cat. Int. J. Neuropharmol., 1, 49-53. Urn, R., Y m m o , K., AND MATSUSHITA, A., (1966); Behavioural and electrophysiological study of drugs affecting brain and motor system in animal experiments. Progress in Brain Research, Vol. 21B, Correlative Neurosciences. Clinical Studies, T. Tokizaneand J. P. Schade,Editors,Amsterdam, Elsevier, pp. 113-149. Kmm, T., (1960); The emotional influence of various drugs upon the reaction induced by electric stimulation of the hypothalamus. Report 2. The influence of psychotropic drugs upon the appearance of sham rage and flight reaction in rabbits, especially on the action of tranquilizer. Folio phurmacol. jap., 56,771-783 (In Japanese). KUBOTA, K^, Iw-, Y.,AND NIIMI,Y., (1965); Monosynaptic reflex and natural sleep in the cat. J. Neurophysiol., 28, 125-138. MA~USWA A.,, TAKESUE, H., AND Kmo, R., A comparativestudy of bulbocapnine, proclorperazine and reserpine catalepsy in animal experiments. In preparation. NIKI,H., (1962); The effects of hippocampal ablation on the behaviour of the rat. Jap. psychol. Res., 4, 139-153.
POMPEIANO, O., AND S m , J. E., (1962); EEG and behavioural manifestations of sleep induced by cutaneous nerve stimulation in normal cats. Arch. ital. Biol., 100, 311-342. PRESTON, J. B., (1956); Effect of chlorpromazine on the central nervous system of the cat; a possible neural basis for action. J. Pharmacol. exp. Ther., 118, 100-115. RANDALL,L. D., (1961); Pharmacology of chlordiazepoxide (Librium). Dis. nerv. Syst., 22, 7-15. SAWA,Y.,AND Tsun, N., Unpublished data. SCHALLEK, W., AND KUEHN,A., (1960); Effects of psychotropic drugs on limbic system of cat. Proc. SOC.exp. Biol. (N.Y.),105, 115-117. S w u , H., (1962); Some problems in physiology of the motor system. Seitai-no-Kagaku, 13, 131144 (In Japanese). SIDMAN, M., (1953); Avoidance conditioning with brief shock and no exteroceptive warning signal. Science, 118, 157-158. SIDMAN, M.,(1955); Technique for assessing the effects of drugs on timing behaviour. Science, 122, 925.
STEG,G., (1964); Efferent muscle innervation and rigidity. Actaphysiol. scand., 61, Suppl. 225. S m n , E. A., BROWN, W. C., AND GOODMAN, L. S., (1952); Comparativeassays of anti-epileptic drugs in mice and rats. J. Pharmacol. exp. %r., 106, 319-330.
TOKIZANE, T., (1958a); Physiology of the limbic system. Saishin Zguku, 13, 1959-1982 (In Japanese). TOKIZANE, T., (1958b); Physiology of corpus amygdaloideum. Recent Adv. Res. nerv. Syst., 2, 493-541 (In Japanese). TOKIZANE, T., KAWAMURA,H., AND Immuru, G., (1960); Hypothalamic activation upon electrical activities of paleo- and archicortex. Neurol. med.-chir., 2,63-76. TOKIZANE, T., AND S w u , H., (1964); Functional Differentiation of Human Skeletal Muscle. Corticalization and Spinalization of Movement. Tokyo Univ. Press, Tokyo.
D R U G S A N D AGGRESSIVE B E H A V I O U R
387
WARDEN, C. J., (1929); A standard unit animal maze for general laboratory use. J. genet. Psychol., 36, 174176. WIKLER, A., (1952); Pharmacologic dissociations of behaviour and EEG. ‘Sleep patterns’ in dogs: morphine, N-allylnormorphine and atropine. Proc. Soc. exp. Biol. (N. Y.), 79, 261-265. WINTER, C. A., AND FLATAKER, L., (1951); The effect of cortisone, desoxy-corticosterone and adrenocorticotrophic hormone upon the responses of animals to analgesic drugs. J. Phurmucol. exp. Ther., 103,93-105. YAMAMOTO, K., (1959); Studies on the normal EEG of the cat. Ann. Rep. Shionogi Res. Lab., 9, 1125-1164 (In Japanese). YAMAMOTO, K., AND KIDO,R., (1962a); Comparative studies on the effects of tranquilizers, barbiturates and morphine with implanted electrodes in cats and dogs. Bruin und Nerve, 14, 591-608 (In Japanese). YAMAMOTO, K., AND Kmo, R., (1962b); An analysis of various central nervous system depressants in YAMAMOTO, K., AND KIDO,R., (1964); An analysis of central acting drugs using animal experiments. Association and dissociation between EEG and behaviour. Bruin and Nerve, 16,4458 (In Japanese). YAMAMOTO, K., YOSHIOKA, M., NAKAMURA, Y., AND KAWAMURA, H., (1961); Electrophysiological study of effects of morphine on the central nervous system. Bruin and Nerve, 13,327-350 (In Japanese).
388
Rhinencephalic Activity during Acquisition and Performance of Conditional Behavior and its Modification by Pharmacological Agents K. F. KILLAM A N D F V A KING KILLAM
Department of Pharmacology, Stanford University School of Medicine, Palo Alto, Calif, (U.S.A.)
The purpose of these studies was to record and measure in brain wave recordings bioelectric signals correlated with the acquisition of a conditional behavior, with the performance of a stabilized conditioned behavior and with alteration of the behavior induced by pharmacological agents. Such studies should allow recognition of patterns of electrical activity which covary in a variety of brain structures. Further, a more dynamic conceptualization of the participation of brain structures in the performance of conditioned behavior and in information processing in general may be realized. Drugs were used to separate possible epiphenomena from crucial electrical events and to gain information concerning possible mechanisms of action of certain agents. MET H 0 D S
Cats with bipolar, stainless steel electrodes implanted in deep brain structures and on the surface of the neocortex were conditioned either to discriminateintensity differences between two identical geometric figures or to discriminate between different geometric forms of the same intensity projected on translucent panels. The animals were required to press the correct panel by pawing (CR) to obtain milk from dishes directly under each of the three translucent panels which served as the manipulanda. The test chamber consisted of a sound-attenuated, shielded box with background lighting, a viewing port and one wall containing the stimulus-manipulandum panels and reinforcement dishes. The conditioning stimuli (CS) were projected upon the translucent panels using a slide projector with a motor driven phantom interrupting the projection path. The light/dark ratio used was 1/4 and the frequency 10 c/s. The use of a frequency-coded CS (TCS) allowed the analysis of spontaneous and evoked activity for alterations of waveforms or rhythms in the brain structure monitored. For a discussion of this technique in detail see John and Killam (1959). The intertrial intervals were manually randomized but the sequence of appearance of the positive CS on one of the three translucent panels was preprogrammed. Once the CS was initiated, the termination of stimuli, the delivery of the reinforcement, the coding and notation of responses were automatically controlled and recorded. Following recovery from surgery, the cats were placed on a’dry diet with ad lib.
RHINENCEPHALIC ACTIVITY
389
fluids. When experimentation began, fluid in the home cage was withdrawn and made available only in the test chamber. The cats were subjected to a period of habituation in which milk reinforcement was not paired with the CS. The animals were then trained to press the panel upon which the geometric form was projected. When the behavior was under stimulus control, in that within 15 sec after the initiation of the CS the cats pressed the correct panel 90% of the time to obtain milk, more complex training began. Two types of training were used: (1) pattern discrimination and (2) total light flux discrimination. In pattern discrimination training, the cats were trained to select one of two geometric forms and finally one of three geometric forms. During each session at least one quarter of the trials used the single presentation. This allowed the comparison of changes in evoked activity related to the complexity of training and uncomplicated by multiple figures in the visual field (Fig. 1). In total light flux discrimination training, the cats were presented with two identical geometric forms, one of which was masked at the projection source with neutral density filters. The animals were trained to select the brighter of the two forms. Three
m u mom ucao
c. s.
Fig. 1. Pattern discrimination training. The upper set represents the habituation phase of training wherein reinforcement was not paired with the CS. The second set represents the first stage of training in which the animal was reinforced upon activation of the panel on which the stimuli appeared. The third and fourth sets represent form discrimination training. In each session at least t of the trials in blocks of 10 were the same as in the first stage of training. The sequence of presentation of the blocks was constant for a given animal but was varied between animals.
1
c.5. MILK
YllU
c. 5. MILU
Fig. 2. Total light flux discrimination training. 1 and 2 represent the early stages of training described in Fig. 1. 3 represents the final stage of training in which the animal was required to select the brighter of the two geometric forms. The actual sequence of presentation in blocks of 10 trials was constant for a given animal but was varied between animals. References p . 399
390
K. F. K I L L A M A N D E. K. K I L L A M
levels of difficulty were employed using Eastman Kodak neutral density filters of 2.0, 1.O and 0.6. This is diagrammaticallyrepresented in Fig. 2. As in pattern discrimination at least one quarter of the trials used only the single form without the neutral filters. Daily sessions were recorded for each cat using a Grass IV electroencephalograph for paper recordings and as preamplification for magnetic tape recordings. The paper records were used to edit the experimental data and to select those data to be processed on the LINC computer. Three basic types of data processing were used on the data from these experiments:(1) sequential averaging, (2) sequential autocorrelations and (3) pattern recognition using cross correlation with zero lag time. The present report will deal with treatments 1 and 2 for reasons that will become apparent below. Sequential averaging and autocorrelations were used in an attempt to recognize both stable time series and discontinuitiesin time series over the CS-CR interval, a time in which instabilities could be recognized by inspection. The minimum epoch employed for both calculations was 800 msec. In sequential averaging the evoked responses to the first 8 flashes of the CS were averaged, then the evoked responses to flashes 2-9, 3-10, etc. were similarly treated until the animals responded (Fig. 3). In sequential
..D
I
.-
,-
a
Fig. 3. Sequential averaging. The upper tracing represents the EEG recorded on the electroencephalograph. The lower tracing represents a plot of data digitized by the LINC computer. The data repre sent the first portion of the upper tracing. The center trace represents the sequential averages derived from the data illustratedin the upper and lower traces. The numbers on the center trace correspond to those on the upper and lower t r a m and thereby indicatethe amount of raw data (8 evoked responses).
autocorrelations, the autocorrelation function of the 800 msec commencing with the initiation of the CS was calculated using a 6 z of one data point (1 -56 msec lag) and a maximum t of 256 points. Then, the data from 400-1200 msec, 800-1600, etc. after the initiation of the CS was used to compute successive autocorrelation functions terminating with the animal's response (Fig. 4). From the above transforms, descrip-
RHINENCEPHALIC ACTIVITY
39 1
Fig. 4. Calculation of sequential autocorrelograms. Traces labeled 2, 3 and 6 refer to brain waves derived from the lateral geniculate nucleus of the thalamus, the nucleus reuniens of the thalamus and the lateral gyrus. The horizontal bars above the upper trace delineate the segments of the brain waves used to compute the autocorrelogram displayed directly under each brain wave tracing. 6 t:1.56 msec or 1 data point. Total sample: 512 points, T maximum: 128 points.
tive patterns of change were recognized for all channels throughout the training periods and during the assessments of drug effects. All electrode placements were verified histologically. RESULTS
( A ) Training The discussion of data from these experiments will be confined to a comparison of the evoked activity from several rhinencephalic structures with that derived from the lateral geniculate body of the thalamus. In each of 12 animals there was at least one electrode in rhinencephalic structures. Since all animals had electrodes in the visual afferent system and comparable changes in evoked activity from these structures between animals were observed, the geniculate changes were used as points of reference for comparison across structures. The evoked activity from the lateral geniculate body in response to a single geometric form became more complex as the animal was required to make more difficult discriminations of differences either in form or total light flux. This is illustrated in Fig. 5. The upper half of the figure represents sets of sequential averages derived from the amygdala (Columns 1-4) and lateral geniculate nucleus (Columns 5-8). Column 5 represents a typical trial from a stage of training at which the animal responds correctReferences p . 399
392
K. F. K I L L A M A N D E. K. K I L L A M
f 0 L
TRAININO
L.
"1
Fig. 5. Changes in evoked brain activity during training. See text for explanation.
ly to a single pattern in 50 % of the trials. Column 6 represents a typical trial when the performance level is 100% to a single pattern. Column 7 depicts averaged potentials in a typical single pattern trial at a stage when training in discrimination of 3 forms has begun. Column 8 is representative of typical sequential average responses to a single pattern when the animal is fully trained to discriminate the correct symbol from the three presented. The sequential averages indicate that as performance is improved to 100% the responses to the CS are enhanced and become somewhat changed in form (Column 6). The change in form is more apparent when the 3 pattern discrimination is introduced (Column 7) and the averaged response in the geniculate takes on a much more sharp and complex wave form as the difficult discrimination is mastered (Column 8). Panels immediately below each column of successive averages display sequential autocorrelograms from the same data. The 10 c/s response (two complete waves in each 200 msec tracing) is apparent much earlier in trials in the second stage (Column 6), becomes stable throughout the trial (Column 7), and eventually becomes sharper, greater in amplitude and admixed with faster activity above 30 c/s at the final training stage (Column 8). It must be emphasized that all the trials from which these data were recorded represent presentation of the same single geometric form so that the visual stimulus field is the same throughout. The upper left of the figure depicts simultaneous sequential averages from the nucleus basalis of the amygdaloid complex during the same trials as those illustrated on
RHINENCEPHALIC ACTIVITY
393
the right side. As can be noted, there is synchronous activity at 40 c/sec earlier in the CS-CR interval (Column 1) but the activity dies away before the CR. In Column 2 the hypersynchronous activity persists throughout the trial. However, note that the phase of the evoked responses gradually shifts. For this reason no pattern recognition transforms were calculated on these data. With the initiation of complex form discrimination (Column 3) the synchronous activity becomes more marked with little further change in the pattern at the behavioral criterion (Column 4). The sequential autocorrelograms from the same raw data displayed below each column reveal added information concerning the stability of the 40 c/s activity. In Column l note that peaks of correlation attributable to a basic 40 c/s persist throughout the trial although the activity is not recognizable in the sequential averaging. In Column 2 there is a variable autocorrelation reflecting the instability of the time series even with the short epochs employed (800 msec). This is in contrast to the sequential averages immediately above. In Column 3 at the start of complex training, there appears in addition to the peaks of correlation equivalent to 40 c/s, an enhanced peak at 100 msec or corresponding to 10 c/s, the basic TCS frequency. This is also apparent at the behavioral criterion (Column 4) and should be noted as being not immediately recognizable from the sequential average plots directly above. The data are consistent with the view that two basic frequencies are derived from the evoked activity. In general, in form discrimination tasks the 40 c/s activity derived from nucleus basalis of the amygdaloid complex becomes more dominant and becomes admixed with the TCS frequency as the daily required discrimination tasks become more difficult. The hypersynchronous activity is discernable in the raw EEG as well as in computer-derived transforms. The increase in 40 c/s activity parallels the change in complexity of the evoked responses from the lateral geniculate body of the thalamus. With total light flux discrimination, a similar parallelism between total difficulty of a daily session and hypersynchronous activity in the amygdaloid complex was not apparent despite the shift of complexity of evoked responses in the lateral geniculate. However, these animals when later trained to form discrimination, exhibited the same amygdala responses as described for the animals trained only to discriminate forms. The spontaneous and evoked activity from the entorhinal area and from the hippocampus did not display the 6 c/s activity reported by Adey et al. (1961). Little evoked activity even at the TCS frequency was observed except from the surface recordings in the hippocampus and in the entorhinal area and this appeared only when bioelectric activity was subjected to sequential autocorrelation analysis. This activity is enhanced following anti-cholinergic drugs. ( B ) Drug studies Anticholinergic drugs. Atropine and scopolamine were used to alter the conditioned behavior of the animals. Atropine methylnitrate was studied as a control to eliminate the possibility that effects were due to pupillary and other peripheral side effects as sequelae of cholinergic blockade. Disruption of behavior under atropine (0.5-1 .O mg/kg i.p.) and scopolamine (0.0550).125 mg/kg i.p.) varied with the dosage employed from errors of performance to failure to attempt to respond. In all cases the animals References p . 399
394
K. F. K I L L A M A N D E. K. K I L L A M
would accept fluids. Atropine methylnitrate at doses up to 2.0 mg/kg i.p. produced maximal peripheral side effects with no change in conditioned behavior or brain wave activity. After atropine and scopolamine the degree of behavioral disruption roughly paralleled neocortical hypersynchrony. Spontaneous activity of hippocampus and entorhina1 area was disorganized but during trials the sequential autocorrelograms indicated an emergence of evoked activity close to the TCS frequency. Fig. 6 compares sequential averaged potentials from 3 rhinencephalic areas with those from lateral geniculate nucleus. Only from the latter can evoked activity at the TCS be recognized in the control and marked enhancementunder atropine be observed. By contrast, autocorrelograms of the same data (Fig. 7) indicate time locked activity clearly in ventral hippocampus as well as in lateral geniculate in the controls. Following the administration of atropine evoked activity in the TCS frequency shows in dorsal and ventral hippocampus and in entorhinal cortex to a lesser degree along with an enhancement of
Fig. 6. Effects of atropine sulfate on evoked activity as measured by sequential averaging. See text for explanation.
395
RHINENCEPHALIC ACTIVITY INTO.
v. nirr.
LOIN.
4
B
L
8
h
Fig. 7. Effects of atropine sulfate on evoked activity as measured by computing sequential autocorrelogram. See text for explanation.
lateral geniculate 10 c/s response. These changes were observed from hippocampal placements which were on the ventral surface of the dorsal hippocampus and the lateral margin of the ventral hippocampus. Nine other hippocampal placements exhibited no change with training or with drug treatment. The 40 c/s activity from the amygdaloid complex was either blocked or markedly reduced by doses of 0.5 to 1.0 mg/kg of atropine when the evoked activity from the lateral geniculate body was increased in amplitude and simplified in waveform. Reserpine. L-desoxyphenylalanine (L-DOPA) and D-amphetamine were used to attempt to reverse behavioral alterations induced by reserpine. In all animals L-DOPA and D-amphetamine reversed both the depression of spontaneous motor behavior and peripheral side effects following doses of 70-120 pg/kg of reserpine. In all animals these doses of reserpine abolished conditioned behavior but in only 1 of 12 animals was this alteration of the conditioned behavior reversed by L-DOPA at 100 mg/kg. In contrast D-amphetamine 0.5 mg/kg reversed the behavioral blockade in 8/10 of the animals. In Fig. 8 sequential averages reveal the effects of reserpine and subsequent reversal by L-DOPA on the 40 c/s activity from the amygdaloid complex during the CS-CR interval. Following reserpine the 40 c/s activity was more marked and lasted throughout the prolonged trials. After the administration of L-DOPA and reversal of the References p . 399
396
K. F. K I L L A M A N D E. K. K I L L A M
Fig. 8. Sequential autocorrelograms following the administrationof reserpine and L-DOPA.
motor behavioral deficit autonomic side effects and the decrement of the conditioned behavior, the evoked activity in the amygdala reverted to more or less normal activity. Three hours later the effects of reserpine returned and the brain wave patterns reverted to the reserpinized state. Marked changes were also noted in the evoked activity from the lateral geniculate. Following reserpine the sequential averages indicated a simplification of the waveforms with a broadening of the potentials which was partly reversed by L-DOPA. The sequential autocorrelograms of the same data indicated marked enhancement of synchronous time locked activity (Fig. 8) under reserpine. In the one cat in which behavior was improved, L-DOPA normalized this pattern. Amphetamine in the 8 cats in which it was effective, reversed the reserpinized brain wave patterns slightly more effectively than L-DOPA, but qualitatively changes were similar. No consistent pattern was observed from other rhinencephalic derivations either with reserpine or the effective antagonist. DISCUSSION
The present studies further emphasize the utility of employing a frequency-coded conditioning stimulus (TCS) for the study of the neurophysiological basis for condi-
RHINENCEPHALIC ACTIVITY
397
tioned behavior. In past studies (John and Killam, 1959) the TCS induced generalized alterations in the total environment by pulsing the background lighting in the test chamber. The present studies were designed to limit the stimulus field and to gain indicators for the handling by the animal of more precise bits of information. Evoked activity from the lateral geniculate body and the amygdaloid complex changed to more complex waveforms even in response to a single stimulus configuration when the animal was trained to discriminate geometric forms. Comparable changes were not seen in recordings from the amygdala when the animals were trained to what the investigator considered was a very difficult discrimination of total light flux in the stimulus field. If one hypothesizes that hypersynchronous activity from the amygdala is consistent with some form of anxiety, then it would follow that the discrimination of forms as designed in these experiments is an exceedingly difficult task for the cat. The present design did not allow the animal to discriminate forms on the basis of intensity changes. It could be argued, however, that the number of bits of difference exceeded the ‘channel capacity’ of the animals accounting for the different amygdaloid activity in the two paradigms. The anticholinergic substances atropine and scopolamine disrupted the conditional behavior. The degree of deficit could roughly be quantified by an evaluation of the individual neocortical hypersynchronous activity. Thus, the old concept of the separation of EEG and behavior by these drugs has to be modified in terms of the yardstick of ‘adequate or normal behavior’ since older studies involved spontaneous rather than conditional behavior. The finding that atropine methylnitrate at doses producing peripheral autonomic changes did not significantly change either the brain wave activity or the conditioned behavior would suggest that the atropine and scopolamine effects were central phenomena. Further, the anticholinergic substances blocked the 40 c/s amygdaloid activity during the CS-CR interval while the waveforms evoked from the lateral geniculate body were distorted but increased in amplitude. At the same time evoked activity from the brain stem reticular formation was little affected. In the past the role of cholinergic mechanisms modulating sensory and motor function via interaction between extralemniscal and lemniscal pathways through the reticular core has been emphasized. The present data would suggest that another major interaction would be rhinencephalo-lemniscal. Further evidence for this suggestion is in the data from the evoked activity from the entorhinal and hippocampal areas after the anticholinergic substances. As the behavioral deficit developed evoked activity became more apparent in these structures. The reserpine experiments are of interest because of the implications for the mechanisms of action of reserpine and for an assessment of the role of biogenic amines in the maintenance of respondant behavior. In the past, evidence for the normalizing effects of biogenic amines has been measured by the ability of these substances to reverse the autonomic symptoms and motor depression following reserpine administration. Such studies have implicated 5-hydroxytryptamine, norepinephrine, epinephrine and dopamine as active neurotransmitters. Further, both from studies of peripheral physiology and neurophysiology, the role of other sympathomimetic amines, e.g. wamphetamine, has been considered to be indirect presumably by releasReferences p. 399
398
K. F. K I L L A M A N F E. K. K I L L A M
ing one of the afore-mentioned amines. Seiden and Carlsson (1964) have reported that L-desoxyphenylalanine (L-DOPA) was able to reverse the deficits induced by the administration of reserpine in cats or rats trained in a conditioned avoidance paradigm. Further, they were able to correlate the antagonisms with changes in levels of dopamine in brain. From the present studies the ability of administered L-DOPAto reverse the autonomic signs and motor depression following reserpine has been confirmed. However, the poor showing of L-DOPAin reversing the conditionedbehavioral deficit (1/12) was surprising. When L-DOPA was effective the brain wave correlates were observed to revert to approximately normal patterns. Amphetamine was observed to reverse effectively(8/10) the conditioned behavioral deficit following reserpine as well as the autonomic signs (10/10) and motor depression (lO/lO). The evoked brain waves also indicated a normalization from the deficit following reserpine. These effects were observed when brain levels of biogenic amines are presumably at a low ebb. These findings would suggest that more selective criteria of behavioral effects coupled with neurophysiological monitors are needed to describe more accurately the role of any substance exogenous or endogenous. Further, the peripheral model employed to describe the pharmacology of a drug, e.g. D-amphetamine, is at best an oversimplified approximation of the effects of such agents in the central nervous system. S U MM A R Y
Cats with chronically implanted electrodes were conditioned either to discriminate light intensity differences between identical geometric forms or to discriminate between different forms of the same intensity projected on translucent panels. The brain wave activity was analyzed using a LINC computer by computing sequential averages and autocorrelograms. The course of change in evoked responses was studied over the conditioning period, during stabilization of the behavioral response and following the administration of drugs. During the CS-CR interval hypersynchronous activity at 40 c/s was observed from nucleus basalis of the amygdala during the acquisition and stabilization of the form discrimination behavior but not during total light flux discrimination. No 6 c/s activity was observed either during the intertrial interval or during the CS-CR interval from the entorhinal or hippocampal derivation. Evoked activity at frequencies slightly higher than that of the CS was observed from the ventral surface of the dorsal hippocampus and lateral margin of the ventral portion of the hippocampus. When the anticholinergic drugs atropine and scopolamine blocked either conditioned behavior, the 40 c/s activity from the amygdaloid complex was reduced or eliminated; evoked activity appeared in the derivation from the entorhinal area; and the evoked hippocampal activity more closely approximated the frequency of the CS. Following reserpine the 40 c/s activityfrom the amygdaloidcomplexbecame more marked. The behavior and altered brain responses were normalized by the effective antagonists L-DOPA or D-amphetamine. Inconsistent patterns were observed following reserpine in the activity from the entorhinal area or hippocampus.
RHINENCEPHALIC ACTIVITY
399
ACKNOWLEDGEMENTS
The data reported are portions of a large study supported in part by Public Health Service Research Grant No. MH-03241 from the National Institute of Mental Health and Grant No. DA18-108-AMC-63-208(A) from the Army Chemical Corps. The experiments were carried out and analyzed by a team consisting of the authors, Miss Ruth Carpenter, Dr. A. J. Hance, Dr. Jane Koenig, Dr. David Lindsley, and Dr. L. S. Seiden. Grateful acknowledgements are made to Miss Igea Arrico for the preparation of histological material and to Mr. Kelvin Seifert for his devoted assistance in the computational aspects of study. Reserpine was supplied by Ciba and D-amphetamine by Smith, Kline and French Laboratories. REFERENCES ADEY,W. R., WALTER, D. O., AND HENDRIX,C. E., (1961); Exp. Neurol., 3, 501-524. JOHN, E. R.,AND KILLAM, K. F., (1959); J. PhurmucoI. exp. Ther., 125, 252-274. SEIDEN, L.s., AKD CARLSON, A., (1964); ~sychophurmco~ogiu, 5, 178-181.
400
Control of Hippocampal Output by Afferent Volley Frequency PER ANDERSEN
AND
TERJE L 0 M O
Laboratory of Neurophysiology, Institute of Anatomy, University of Oslo, Oslo (Norway)
INTRODUCTION
One of the crucial factors determining the ouput from a neuronal system is the frequency of the input stimulation. This affects both the electrical and, most likely, the behavioural responses to stimulation of various brain structures. Not least is this so in the cortex. Indeed, both with monosynaptic and polysynaptic cortical connections, the frequency of afferent stimulation plays a decisive role in determining the output. All investigators studying the responses of hippocampal pyramidal neurones are familiar with the great increase, or decrease, down to the level of extinction, of the responses that may occur in different circumstances. An understanding of the mechanisms involved in both increased and decreased probability of discharge of hippocampal neurones is of paramount importance in order to design efficient stimulation experiments. The aim of the present investigationhas been to study the effect of a relatively shortlasting tetanic stimulation on the efficiency of certain excitatory synapses in the hippocampus. Further, possible correlations between neurophysiological responses and processes of learning have been sought. The results have been encouraging, since a relatively short-lasting tetanic stimulation was capable of increasing the number of discharging pyramidal cells by a factor of several hundred. The effect took place during the tetanic stimulation and is therefore different from post-tetanic potentiation. Furthermore, the effect considerably outlasted the stimulation and is, therefore, of some interest in connection with learning processes. METHODS
Adult cats and rabbits were used. The cats were anaesthetized with a mixture of urethane and chloralose (500 and 40 mg/kg body weight, respectively) or with 30 mg/kg sodium pentobarbital, all drugs being given intravenously. The rabbits received urethane-chloralose intravenously (750 mg and 40 mg/kg, respectively). The hippocampal formation was exposed by removal of the overlying neocortex by suction. The skin flaps were sewn to a metal ring to make a pool which was filled with warm liquid paraffin. Various afferent fibre systems were stimulated through bipolar electrodes made of
FREQUENCY POTENTIATION I N HIPPOCAMPUS
40 1
two insulated silver or copper wires of 0.5-0.8 mm diameter, laquered together and cut off squarely to give bare tips. The polar separation was from 0.5-1 mm. Recording was made by glass micropipettes, filled with a solution of either 4 M NaCI, 3 M KC1 or 2 M potassium citrate. The signals were fed through a differential cathode follower to high stability differential DC or AC coupled amplifiers, displayed on an oscilloscope and photographed. The cathode follower output was also made to operate a meter amplifier to give a continuous visual reading of the membrane potential, and was further led to a neutralizing amplifier to compensate for the capacity between the microelectrode and earth. Through a bridge circuit, similar to that described by Araki and Otani (1955), current pulses could be applied through the intracellularly located microelectrode to alter the membrane potential of the neurone from which the record was obtained. The microelectrode was moved by a micromanipulator, the advancement of which was guided by the monitoring of the microelectrode signals on a display oscilloscope. This instrument had a phosphor electrode with long persistence of the trace, a feature which facilitated the comparison between subsequent traces. A major obstacle for intracellular recordings of long duration from the hippocampal pyramids was the respiratory and circulatory movements of the brain. These effects were reduced by the application of a perspex pressure foot with a central hole for the microelectrode and the surface electrode. Further reduction of the movements of the brain tissue was produced by immobilization of the animal with succinylcholine infusion (0.15 mg/kg/min) and artificial respiration with a relatively small stroke volume (600 ml/min in rabbits, 750 ml/min in cats). Recordings were made from the areas CA3 and CAI in both cats and rabbits. In both animals, the area CA2 is too small to allow its identification without subsequent specialized histological examination (Lorente de N6, 1934). When recording was done from the CA3 pyramids, the afferent systems employed (Fig. 1A) were the commissural fibres (Com) which end on the basal dendrites (Blackstad, 1956), and the mossy fibres (MF), the axons of the dentate granule cells that end on the big spines of the proximal portion of the apical dendrite (Cajal, 1911; Lorente de N6, 1934; Blackstad and Kjrerheim, 1961; Hamlyn, 1962). When records were taken from the CA1 neurones (Fig. lB), the afferent fibres activated were the commissural fibres which end on the basal dendrites as well as on the proximal half of the apical dendritic tree (Blackstad, 1956), and the Schaffer collaterals of the CA3 pyramids which end on a fairly restricted portion of the apical dendrite and its branches, peripherally to the territory of the commissural synapses (Cajal, 1911 ; Lorente de N6, 1934). The commissural stimulus was delivered at the homotopic point, opposite the recording electrode. The mossy fibres and the Schaffer collateral system were brought into operation by stimulation of the entorhinal area. In rabbits, the stimulating electrode was located on, or just beneath, the angular bundle. The synaptic action of the mossy fibres was exerted on CA3 cells, while that of the Schaffer collaterals acted on CA1 pyramids. The perforant path synapses were also activated by a stimulus to the entorhinal area, but the latency of the monosynaptic perforant path activity is shorter References p . 41.2
402
P. A N D E R S E N A N D T. LBMO
than the tri-synaptic Schaffer collateral activity initiated by an entorhinal volley (Aedersen et al., 1966). RESULTS
Pathways studied. The synaptic pathways that we have used in the present investigation are diagrammatically represented in Fig. 1. A represents a CA3, and B, a CAI pyramidal neurone. Their relative positions in the hippocampal formation are indicated in Fig. lC, which is a simplified sagittal section. The arrangement of the recording and stimulating electrodes can be seen from Fig. 1D. Recording from the CA3 neurones, two pathways were employed. First, the commissural afferent volleys were initiated by delivering stimuli to the homotopic point of the other side (Corn). The system ends with boutons en passage on the basal dendrites of the CA3 neurones (Fig. IA). This connection is monosynaptic. The other set of synapses activating the CA3 cells was the mossy fibres, MF in Fig. 1A. These fibres were activated, partly directly through stimulation in the dentate fascia, partly transsynaptically by stimulation in the entorhinal area (Ento in Fig. 1C and D). The mossy fibres end with giant synapses in contact with large dendritic spines of the proximal part of the CA3 apical dendrites (Fig. 1A). In recording from CA1 (Fig. lB), three different afferent systemswere employed, all monosynaptic to the CA1 pyramids. The pathways are the commissural (Corn), the Schaffer collaterals of the CA3 pyramids (Sch) and the perforant path (PP) that can be activated by an entorhinal electrode. The territories of the different synapticcontingencies are shown in Fig. 1B. Thus, it is possible to activate monosynaptically different portions of these two types of pyramidal cells.
Fig. 1. Diagrammatical representation of the hippocampal formation with the cell types studied, A CAI neurone (A) and a CA3 neurone (B) are shown with some afferent pathways. (C) Sagittal section showing localization of neuronesand pathways. @) The hippocampalformation viewed from above, indicating the stimulated areas and the regions from which records were taken. Corn, commissural pathway; Ento, entorhinal area; MF, mossy fibres; PP, perforant path; Sch, Schaffer collaterals.
FREQUENCY POTENTIATION I N HIPPOCAMPUS
403
Excitatory eflect of the diflerent afferent pathways. All pathways, except the direct perforant path, were able to initiate an extracellular population spike, indicating the near synchronous discharge of a great number of pyramidal neurones. Examples of such responses can be seen in Fig. 7. Recording from an appropriate area in response to stimulation of a certain afferent pathway, the extracellular electrode revealed a negative field potential indicating inward current at the level of the activated synapses. In accord with the data from the experimental histologists, the field potentials had their maxima at the level of the termination of afferent fibres (Blackstad, 1956, 1958). Superimposed on the negative field potential, there appeared one or two negative short-lasting (2-4 msec) deflections. Since these spikes showed a definite inversion at the pyramidal layer, and occurred in time simultaneously with intracellularly recorded spikes, they indicate the discharge of pyramidal cells. Iittracellular studies of synaptic activity. On the basis of the extracellular recording showing the ease with which the afferent systems employed could elicit pyramidal cell discharges, the almost total lack of an excitatory postsynaptic potential (EPSP), seen by the intracellular recording electrode, was somewhat of a paradox. Even in a cell that for other reasons showed a great tendency to discharge, as judged by spontaneous firing (Fig. 2), single shock stimulation only created a pure hyperpolarization response, an inhibitory postsynaptic potential or IPSP (Fig. 2A). Very rarely a definite EPSP
E lkac '4
--A/-Fig. 2. Normal appearance of synaptically activated spike, (A) shows the usual appearance of an IPSP in response to a single commissural volley. The neurone had a great tendency to fire as indicated by the spontaneously occurring spike. On increasing the rate of stimulation from 1 t o 10 per sec (B and C), there appeared an antidrornic, and later, a synaptically driven spike. The latter rises from the baseline without any prepotential. (D) and (E) are taken at l/sec stimulation, 2 and 4 sec after the tetanic stimulation, respectively. intra, intracellular; extra, extracellular; surf, surface records. References p . 412
404
P. ANDERSEN A N D T. L B M O
was seen in response to a single afferent volley. This probably occurred only when a large number of synapses, situated close to the cell body, was excited. We have seen it in 4 cells only, out of a total of several hundred penetrations. The usual lack of an EPSP is compatible with a localization of excitatory synapses at some distance from the cell body. Thus, under conditions prevailing in our experiments, only a limited number of neurones was discharged by the first afferent volley. The majority of the pyramidal cells were inhibited, most likely through the recurrent inhibitory pathway described by Kandel et al. (1961), and later studied by Andersen et al. (1964). On the basis of this information, one might ask how the hippocampal pyramidal neurones can be recruited to a massive discharge by an afferent volley. An unanaesthetized preparation might show a greater tendency to discharge of the pyramidal cells. However, another major factor determining the output from hippocampal cells is the rate at which the afferent volleys impinge upon the cell population. Thus, by increasing the rate of stimulation from l/sec to lO/sec (Fig. 2C, D), the cell shows both antidromic and synaptically driven discharges. The greater efficiency of synaptic discharges is shown particularly well in Fig. 2C, where the antidromic spike occurs only once and the rest of the 6 superimposed sweeps shows a synaptically excited spike only. A comparison between Fig. 2B and C shows that the situation develops over a period of several seconds. A single afferent volley, elicited 2 sec after the cessation of the tetanic stimulation (Fig. 2D), was still capable of discharging the cell. When 4 sec had elapsed only an IPSP was elicited (Fig. 2E). The enhanced efficiency of the synaptic transmission evolving during the tetanic stimulation and outlasting it, will be called frequency potentiation. We believe this factor to be of major importance in determiningthe behaviour of pyramidal cells in response to a given afferentstimulation. Mechanism of frequency potentiation. Studying Fig. 2, one might get the impression that the increase of the excitatory synaptic efficiency is due to a decrease in the IPSPs. However, the records of Fig. 2 were taken with AC amplification and cannot be used to demonstrate any possible reduction in the recurrent IPSPs during tetanic stimulation. The apparent reduction is due to a hyperpolarization of the cell to a level close to the EIPSP. In order to study the time course, and some features of the mechanism of frequency potentiation, several types of experiment were performed. In Fig. 3, each assemblage of 3 records in the upper two rows are from above: the intracellular record, the just extracellular record, taken with the same gain and polarity as the intracellular, and the surface record. Following an increase in the rate of stimulation from 1 to lO/sec there occurred over a period of several seconds a decrease in the IPSP, followed by the development of a depolarizing potential. The latter had a rise time, duration, and a relation to spike discharge that allowed its classification as an excitatory postsynaptic potential (EPSP). The growth of this EPSP is quite formidable, reaching amplitudes up to 16 mV. Since the excitatory synaptic activity takes place at a distance from the soma, we must assume that the local EPSPs are even larger than those shown in Fig. 3. In the surface records of Fig. 3, the first potential deflection is caused by the antidromic invasion, and the second wave by the monosynaptic activation of CA1 pyramidal
FR E Q U E N C Y P O T E N T I A T I O N I N H I P P O C A M P U S
I /sec
405
-
Fig. 3. Frequency potentiation of EPSPs. Successive records taken at rates of stimulationof l/sec and lO/sec. In each assemblage the records are from above: the intracellular (intra), the extracellular (extra) and the surface record (surf). The numbers under the records indicate the time since the onset of the tetanic stimulation (first and second horizontal rows), and after the cessation of the tetanus (third and fourth horizontal rows). The arrows point to the onset of the EPSP as measured to the point of departure between the intra- and extracellular records.
cells. The latency of the EPSP is about 3.2 msec (arrows), measured to the point of departure between the extracellular and intracellular records. This latency is 1.5-2 msec shorter than the latency of the IPSP, and corresponds to the second wave of the surface record, showing that at least the first part of the EPSP is monosynaptic. A salient feature of the frequency potentiation response is that the enhancement of the EPSP outlasts the tetanic stimulation period by a considerable time. Time course of frequency potentiation. The time course of the frequency potentiation effect may be seen from the graph in Fig. 4. The potential amplitude is plotted as the distance from the base line before the shock artifact to the peak of the IPSP in response to 1.2/sec stimulation (filled circles), and also at another time incidated by the latency to the peak of the fully developed EPSP (open circles; a and b in theinset). Asexpected, the potential values measured at these two times after the shock artifact, parallel each other closely. Apparently, the IPSP was reduced to zero by a tetanic stimulation of 12/sec for 6 sec, being followed by a period with a large EPSP. On cessation of the tetanic stimulation, the amplitude of the EPSP showed an initial rapid and a later slower decline, until the pre-stimulation level of the IPSP was reached, after about 35 sec. Another period of tetanic stimulation produced similar, but slightly smaller response changes. There was a slight tendency to a slower decline in the potential size after the second period of tetanic stimulation, as compared with the first. Provided References p. 412
406
P. A N D E R S E N A N D T. L 0 M O
Fig. 4. Time course of frequency potentiation of hippocampal EPSPs. The amplitude of the postsynaptic potentials elicited in a CA3 neurone by commissural stimulation is plotted against time. The amplitude has been plotted at two times after the stimulus artifact, corresponding to the peak of the IPSP (filled circles) and to the peak of the EPSP (open circles). The thick horizontal bars indicate the periods of tetanic stimulation at l2/sec. The records in the inset bere taken at the points indicated by a and b, respectively.
Fig. 5. Effect of repeating series of tetanic stimulation. The amplitude of the postsynaptic potential o a CA3 neurone in response to a commissural stimulation is plotted against time. The amplitude is measured in relation to the baseline preceding the artefact, The thick horizontal lines indicate three periods of tetanic stimulation at lO/sec. The recovery to the normal hyperpolarizing response takes a longer time after the third period.
the tetanic stimulation did not last long enough to produce extinction of the synaptic system, this decline in the augmented potentials took longer and longer time, being an example of a primitive synaptic learning. The repeatability of the frequency potentiation response is illustrated by Fig. 5. Here, the size of the synaptic potential measured at a given time after the shock artifact is plotted against time. Three periods of lO/sec stimulation follow each other with intervals of l/sec stimulation. The effect had a similar latency each time, and the decline from the depolarized level was slightly slower for each tetanic period. This is,
FREQUENCY POTENTIATION I N HIPPOCAMPUS
407
therefore, an indication of a simple learning process in a cortical synaptic system. If, however, the rate of stimulation was increased too much, or the duration of the tetanic stimulation was too long, the frequency potentiation vanished in spite of continued stimulation. There is an optimal frequency at which the size of the intracellularly recorded EPSP and the extracellularly recorded population spike acquire maximal sizes for an extended length of time. A further increase in the rate of stimulation will give an even larger EPSP or population spike, but the duration of the increased responses is cut short very dramatically. Following a period of tetanic stimulation producing frequency potentiation, there appeared a period of lowered synaptic excitability. This reduced excitability, in some instances a complete extinction, lasted from a few seconds up to several minutes, depending upon the rate and duration of the tetanic stimulation. The role of frequency potentiation in initiating hippocampal cell discharges. With intracellular recording, cells were occasionally met in which a small all-or-nothing spike was seen in response to a single afferent volley. In Fig. 6, the commissural pathway was stimulated and the records were taken from a CA3 neurone. In Fig. 6A, the all-or-nothing spike has an amplitude of about 3 mV and a duration of 2 msec. Following an increase in the rate of stimulation from l/sec to lO/sec (Fig. 6B, C), there appeared rather suddenly an antidromic invasion and also a synaptic activation
Fig. 6. Effect of frequency potentiation in assisting the soma invasion of a dendritic spike. (A)isthe response of a CA3 neurone to a single commissural volley. A small all-or-nothing spike appears. In (B) and (C) the rate of stimulation was raised from 1 to lO/sec. In (C) there suddenly appeared a large antidromic spike, followed by a synaptically driven spike. In @) to (H) the gain was reduced. The records were taken consecutively. In (D), (F), and (H) the rate of sthulation was lO/sec, in (E) and (G) it was l/sec. The tetanic stimulation produced a depolarization of the cell that assisted the spike generated in the basal dendrites t o invade the soma. The arrows indicate the stimulus artifact. References p. 412
408
P. A N D E R S E N A N D T. L B M O
of the neurone. An important feature of the cell responses in these circumstances was the lack of any prepotential of the spike. In Fig. 6C, the last spike is the synaptically induced one, and takes off straight from the baseline without any sign of a preceding EPSP. In Fig. 6D-H, the gain was reduced. Due to an increased capacity between the microelectrode and the surrounding tissue, the stimulus artefact somewhat distorted the first antidromic spike. However, the following synaptically produced spike appeared undistorted, and showed a typical A-B inflection. When the rate of stimulation was reduced to l/sec again (Fig. 6E), the large spike disappeared, and a small all-ornothing spike of about 8 mV amplitude appeared. A new increase in the rate of stimulation to lO/sec produced a new large spike (Fig. 6F). This time the B component was often blocked, leaving the A spike only. A further reduction to l/sec stimulation (Fig. 6G) again gave the little all-or-nothing spike, whereas the last increase to lO/sec stimulation provoked the full-blown spike again. Thus, by altering the rate of stimulation the appearance of this spike could be drastically changed. The interpretation of these records must be based on the fact that the electrode tip is most likely located in the soma. Since the small spikehad an all-or-nothing character, a positive polarity, was mono-phasic, and of a considerable amplitude, it was most likely due to a spike produced in that same neurone, but at a distance from the cell body. The most likely explanation is that the small spike was generated somewhere in the dendritic tree, and that its invasion of the soma is blocked under conditions of l/sec stimulation. Following this interpretation, we call the small spike a dendritic spike or a D spike. When the rate of stimulation was increased to lO/sec, the additional depolariaztion created, as shown in Figs. 4 and 5, helped the invasion of the D spike into the soma. A successful invasion of the soma by the D spike was signalled by the ordinary A-B spike. Careful measurement shows the times of rise and fall of the D spike to be similar to those of the B spike. Thus, the D spike is most likely due to an ordinary B spike generated somewhere in the dendritic tree, but electrotonically attenuated owing to the cable-like properties of the dendrites. The sequence in Fig. 6 clearly shows that the frequency potentiation may be a major factor determining the discharge of a single cell. The functional role of the neurone depends upon whether or not it discharges a B spike down its axon. Any local spike within the dendritic tree, not producing a soma invasion,will have no effect on the output of the system. The power of repetitive stimulation in assisting the soma invasion of dendritic spikes is better seen by extracellular recording of population spikes. This procedure gives a measure of the total number of cells participating in the discharge. In Fig. 7, the recordipg is from the CA3 area, and the input from the entorhinal area. The upper trace is the extracellular CA3 record and the lower trace the surface record. In the extracellular l/sec record, the inflection is due to the distant granule cell activity. Following this activity, a faint up-going deflection is hardly noticeable (arrow). However, when the rate of stimulation was increased to lO/sec, the record taken 1 sec after the onset of stimulation shows the growth of a negative potential of about 4 msec duration (arrow). When the tetanic stimulation had lasted for 6-8 sec, respective-
FREQUENCY POTENTIATION I N HIPPOCAMPUS
409
Fig. 7. Efficiency of frequency potentiation in recruiting pyramidal cells to discharge. Extracellular records taken from the pyramidal layer of CA3 in response to entorhinal stimulation. Single shock activation (l/sec) hardly produced any potential attributable to mossy fibre synaptic activity (arrow). The preceding potential is due to the dentate granule cell activity. When the rate of stimulation was increased to lO/sec, there appeared first a negative wave (l”, arrow), indicating the inward current at the mossy fibre synapses, and later a negative wave with one or more superimposed spikes ( 6 and 8”). Following reduction of the stimulus rate to l/sec, a large negativity is still produced by the mossy fibre synapses 1 sec after the tetanus. Fifteen seconds later only traces of this potential are left. Lower traces are extracellular records.
ly, the negative wave carried a large population spike on its top. It is difficult to estimate the total number of cells participating in this response. However, by dividing the total amplitude of the population spike, amounting to 8 mV in the 8-sec record, with the smallest spike amplitude recorded, one arrives at a value of several hundred. The effect outlasts the tetanic stimulation. On reduction of the rate of stimulation to l/sec, there still prevailed a large negative potential 5 sec after the cessation of stimulation. Fifteen seconds later, there was still a small negative deflection at the place of the earlier large negative wave. Thus, the extracellular recording reveals that the process of frequency potentiation possesses a great power in increasing the number of hippocampal pyramidal neurones taking part in a discharge. There seems to exist an optimal rate of stimulation which is larger than 6/sec but lower than I5/sec, 10-12/sec being the optimal value under the conditions we have used in our experiments. When lower rates of stimulation are employed the number of discharging cells will be much smaller. On the other hand, if higher rates of stimulation are used, the output, after a short intense period of discharge, will rapidly drop to very low values, even to zero. References p . 412
410
P. A N D E R S E N A N D T. L 0 M O DISCUSSION
Role of frequency potentiation in regulating the behaviour of the pyramidal neurones. From the results, it is clear that the rate of stimulation is a decisive factor determining the output from the hippocampal formation. It appears reasonable to take the information about optimal rates of stimulation into account when experiments are designed to produce a maximal output from the hippocampal formation by electrical stimulation. Further, when short bursts of tetanic stimulation are used, they should last beyond a minimum period of 5-10 sec. This is the time during which the EPSP grows to attain a maximum amplitude and hence produce the large depolarization. The latter assists the dendritic spike invasion of the soma, and will therefore determine the number of discharging neurones in the hippocampal output. Employing rates of stimulation below 6/sec will mean that a large number of hippocampal cells will have repetitive IPSPs produced so rapidly as to keep the membrane hyperpolarized. In such circumstances, it is not unlikely that the total number of discharging cells will decrease, the opposite of what one might expect. On the other hand, the application of tetani of more than 15/sec very rapidly leads to the extinction of the hippocampal synaptic activity. Therefore, when such stimulation is used, one can only expect a short-lasting activity of the hippocampal neurones. In conclusion,therefore, the present experimentsindicate that there exists an optimal rate of stimulation of the hippocampal formation of about 8-l0/sec. However, we would like to stress that comparable studies have to be done in unanaesthetized preparations in order to assess the optimal rate of stimulation under such conditions. Such studies are in progress in our laboratory. Possible relation between frequency potentiation and learning processes. If several neuronal channels are lying side by side, and one of these channels is used by a stream of afferent volleys at an optimal frequency, this line will increase its efficiency of synaptic transmission by a factor of several hundred in relation to the neighbouring channels. Further, since the potentiation takes place during the tetanic stimulation, the effect is different from the post-tetanic potentiation. Furthermore, the increase in EPSPs denoting an enhanced efficiency of synaptic transmission outlasts the stimulation by a certain period, from several seconds up to a few minutes. This duration is of the same order of magnitude as that of the post-tetanic potentiation. It is too short to account for the plastic changes in a neuronal circuit that might take place in learning processes of a higher kind. However, if frequency potentiation takes place in a set of neurones constituting a polysynaptic chain, the individual effects may be greatly enhanced since the amplifying factor of the frequency potentiation is of the order of several hundred in relation to a factor of 2-5 for the post-tetanic potentiation.,It appears, therefore, likely that the frequency potentiation of cortical synaptic activity may be a factor involved in the establishmentof neuronal circuits of facilitated synaptic transfer as one might envisage happening during a learning process. Further studies are necessary, especially directed towards the duration of the enhanced synaptic efficiency after the frequency potentiation period. Also, studies must be undertaken to determine whether lasting changes appea; :in a neuronal circuit which has been subjected to a series of periods with tetanic stimulation.
FREQUENCY POTENTIATION I N H I P P O C A M P U S
41 1
Frequency potentiation antagonizing recurrent inhibition. It is a striking fact that every pyramidal cell in the hippocampus seems to be under the influence of the recurrent inhibition. This inhibition is very pronounced and longlasting (Kandel et al., 1961). Since the durations of the recurrent IPSPs are of the order of 2-600 msec, the maximal rate of cell discharge would be 2-5/sec. However, the hippocampal neurones may discharge at a much higher rate. Further, we know that the cells may follow artificial stimulation at rates up to 80/sec. The solution of this apparent paradox may lie in the phenomenon of frequency potentiation. The process of frequency potentiation may act as an antagonist against the inhibitory effect of the recurrent IPSPs. During tetanic stimutation, therefore, there is an initial period of about 5-6 sec in which the recurrent IPSPs keep the total output of the system at a low level by abolishing the discharge of many pyramidal cells. Following this initial period, the EPSPs will have increased in a large number of cells to such levels that the recurrent inhibitory effect is completely counteracted. The great power of the process of frequency potentiation merits a closer study of this phenomenon, both in the hippocampal formation and also in other locations within higher levels of the central nervous system. SUMMARY
In cats and rabbits anaesthetized with nembutal or urethane and chloralose, the effect of different frequencies of stimulation on synaptic transmission in the hippocampus was studied. When various excitatory inputs were stimulated with single shocks at l/sec, EPSPs were recorded from only 4 out of several hundred cells studied. With higher frequencies of stimulation, there often developed, after an initial phase of hyperpolarization of the membrane due to summation of individual IPSPs, large depolarizing potentials (maximum 16 mV), whose criteria justified their interpretation as EPSPs. The optimal frequency of stimulation for their production was 8-12/sec. Higher frequencies led to rapid extinction of the EPSPs. Occasionally, single shock stimulation produced all-or-nothing spikes of small amplitude (3 mV). These spikes are interpreted as dendritic spikes (D spikes) failing to invade the soma. At higher frequencies, these spikes are replaced by normal A-B spikes or only A spikes, indicating that repetitive stimulation aids the invasion of propagated dendritic spikes into the soma. Repetitive stimulation led to the development of large negative waves with population spikes superimposed. The evoked potentials could increase by a factor of several hundred. The process whereby higher frequencies of stimulation increases the synaptic response and aids the propagation of spikes along the dendrites is termed frequency potentiation. Some findings are presented suggesting that trains of tetanic stimulation may lead to facilitation of the response to similar trains of tetani presented seconds or minutes later. This phenomenon may be interpreted as an example of primitive synaptic learning. References p. 412
412
P. A N D E R S E N A N D T. L B M O
ACKNOWLEDGEMENT
This investigation was supported by Public Health Service Research Grant NR 04764, from the National Institute of Neurological Diseases and Blindness, which is gratefully acknowledged. REFERENCES ANDERSEN, P., ECCLES, J. C., AND MYNING, Y.,(1964); Pathway of postsynaptic inhibition in the hippocampus. J. Neurophysiol., 27, 608-619. ANDERSEN, P.,HOLMQVIST, B., AND VOORHOEVE, P. E., (1965);Excitatory synapses on hippocampal apical dendrites activated by entorhinal stimulation. Acta physiol. scand., In press. ARAKI,T., AND OTAM,T., (1955); Response of single motoneurones to direct stimulation in toad’s spinal cord. J. Neurophysiol., 18,472-485. BLACKSTAD,T. W.,(1956);Commissuralconnectionsofthehippocampal regioninthe rat, with special reference to their mode of termination. J. comp. Neurol., 105,417-536. BLACKSTAD, T. W.,(1958); On the termination of some afferents to the hippocampus and fascia dentata. Acta amt. (Basel), 35,202-214. BLACKSTAD, T.W.,AND KJAERHEIM, A., (1961); Special axo-dendritic synapses in the hippocampal cortex: Electron and light microscopic studies on the layer of mossy fibres. J. comp. Neurol., 117, 133-160. CAJAL,S . RAMON Y, (1911); Histologie du Syst2me Nerveux de I’Homme et des Vert&br&, Vol. 2. Paris, Maloine, pp. 993. HAMLYN, L.H ,(1962);The line structure of the mossy fibre endings in the hippocampusof the rabbit. J. Anat. (Lond.), %, 112-120. KANDEL, E. R., SPENCER, W.A., AND BRINLEY, F. J., (1961); Electrophysiology of hippocampal neurons. I. Sequential invasion and synaptic organization. J. Neurophysiol., 24,225-242. DE NO, R., (1934);Studies on the structure of the cerebral cortex. 11. Continuation of the LORENTE study of the Ammonic system. J. Psychol. Neurol. (Lpz.), 46, 113-177.
On the Functional Significance of the Hippocampal &Rhythm* PIER LUIGI PARMEGGIANI * * Isstituio di Fisiologia umana dell'universita, Bologna (Italy)
In 1938, Jung and Kornmuller reported that, in the rabbit, low-frequency waves appear in the hippocampal recordings upon stimulation of peripheral nerves. These and similar findings of other authors (MacLean et at., 1952) remained practically unnoticed until Green and Arduini (1954) stressed the importance of this pattern of response. Indeed, they pointed out the very peculiar fact that in arousal the bioelectrical patterns of hippocampal activation are the reverse of those exhibited by the neocortex. That these low-frequency waves (&rhythm) can be elicited by a great variety of stimuli or by the electrical stimulation of central and peripheral nervous structures, was shown by several authors (Liberson and Cadilhac, 1954; Liberson and Akert, 1955; Passouant et al., 1955; Rimbaud et at., 1955; Gangloff and Monnier, 1956; Grastyan et al., 1959; Tokizane et al., 1959; Iwata and Snider, 1959; Brucke et al., 1959a, b ; Beteleva and Novikova, 1960; Torii and Kawamura, 1960; Corazza and Parmeggiani, 1960a, b ; Parmeggiani and Zanocco, 1961,1963). The afferent pathways involved have also been studied (Green and Arduini, 1954; Mayer and Stumpf, 1958; Briicke et al., 1959a, b ; Corazza and Parmeggiani, 1960a, b, 1961a, b, 1963; Torii, 1961 ;Kawamura et al., 1961 ;Petsche et al., 1962, 1965; Stumpf, 1965; Brugge, 1965). Furthermore, unitary analysis has shown that not only is each of the &waves generally associated with a burst of unit activity (Arduini and Pompeiano, 1955; Green and Machne, 1955), but also that fixed phase relationships may be maintained between spikes and waves (Green et at., 1960, 1961; Fujita and Sato, 1964). Many other important contributions to the electrophysiology of the hippocampus should be mentioned, but because they cover a larger field than is desirable for this condensed paper, the reader is referred to the recent review of Green (1964) as a further source of information and references. A short account of recent findings is presented here with the view of clarifying certain aspects of the functional significance of the hippocampal &rhythm. Although this rhythm appears to be a typical feature of the hippocampograms of lower mammalsIonly, a better understanding of its significance would improve our knowledge of the physiology of the hippocampus of both lower and higher mammals. In lower mam-
* Dedicated to Prof. W. R. Hess. ** Present address: Istituto di Fisiologia urnana delI'Universita, Piazza S. Donato 2, Bologna (Italy). References p. 4 3 8 4 4 1
414
P. L. P A R M E G G I A N I
mals, the effects of such a stereotyped and synchronized output as that of the hippocampus during the appearance of the &rhythm can be detected more easily than those of outputs related to any other physiological condition of hippocampal activity. In higher mammals, the depression or suppression of the 8-rhythm appears to be the final step of an evolutionary trend already apparent in lower mammals. The differences between the hippocampograms of the rabbit and the cat need not to be stressed here: it can be simply stated that these differences seem to be related to the development of the neocortex. What has been gained and what has been lost by higher mammals can only be clarified if the functional significance of the hippocampal &rhythm in lower mammals becomes clear. The following working hypothesis has been derived from the experimental data already mentioned. It seems reasonable to suppose that the hippocampal output during the &rhythm exerts a synchronizinginfluence on the activity of subcortical and neocortical neurones and therefore counteracts the desynchronizing effects of the reticular activating system. In fact, the bursts of hippocampal unit activity associated with the &waves allow one to predict that the hippocampal output in turn may induce a rhythmic firing of neurones receiving projections directly or indirectly from the hippocampus. Furthermore, because such synchronizing action would occur in arousal, an antagonistic role against the effects of the reticular activating system has been tentatively attributed to the hippocampal output during the &rhythm. The validity of this working hypothesis was supported by the results of behavioral experiments (Parmeggiani, 1958, 1962a). Repetitive electrical stimulation of the midbrain reticular formation in unrestrained cats, besides increasing the arousal level, also elicits an amazing rebound sleep behavior. After the arousal the animal yawns, grooms itself, looks for somewhere to lie down, curls up and finally goes more or less deeply to sleep. Because such patterns of sleep behavior occur in a primary fashion during direct stimulation of the hippocampus (Parmeggiani, 1959,1960), the origin of the rebound sleep could itself be due to the intervention of the hippocampus indirectly activated by the reticular stimulation. As shown by Green and Arduini (1954), reticular stimulation elicits the hippocampal &rhythm, which may last longer than the stimulation period. In the present exposition the following points will be considered : (A) Mechanisms underlying the appearance of the hippocampal &rhythm in arousal; (B) Neocortical influences of the hippocampal output associated with the 8-rhythm ; (C) Thalamic influences exerted by the hippocampal output during low-frequency stimulation of the hippocampus dorsalis; (D) Interaction phenomena appearing at the thalamic level between hippocampal and reticular iniluences; (E) Behavioral effects of the suppression of the hippocampal &rhythm; (F) Behavioral effects of the repetitive stimulation of the hippocampus dorsalis, the fimbria and the fornix.
415
FUNCTIONAL SIGNIFICANCE OF @-RHYTHM
(A) Mechanisms underlying the appearance of the hippocampal @-rhythmin arousal A simple method for the elicitation of the hippocampal 0-rhythm in curarized* cats consists in the repetitive stimulation** of the sciatic nerve (Corazza and Parmeggiani, 1960a, b). This method of activation is preferred because it provides at the same time reliable and physiological input to both the reticular formation and the hippocampus dorsalis. In curarized cats (Corazza and Parmeggiani, 1961a, b), low-frequency sciatic stimulation (4-2O/sec, 1 msec, 0.5-3 V) elicits the hippocampal 0-rhythm, whereas with higher frequencies (20-100/sec, 1 msec) the type of hippocampal response depends upon the voltage of the stimuli (0.5-3 V). In the latter instance, by increasing the voltage from 0.5 to 3 V the 0-waves become smaller and smaller in amplitude up to clear-cut desynchronization of the hippocampogram during stimulation (Fig. 1).
A hd hs I
I
hd
hs
C
I
t
hd
hs
hd hs I
I
,
2sec,
Fig. 1. Effects exerted on the bioelectrical activity of the hippocampus by repetitive electrical stimulation of the sciatic nerve. Unanesthetized, curarized cat. Effects of sciatic stimulation (lOO/sec, 1 msec) of increasing voltage (A, 0.6 V; B, 0.8 V; C, 1 V; D, 1.4 V). Note that the 8-waves are smaller in amplitude during the stimulation period of B and C than during that of A. Note also in D the overt desynchronization during the stimulation period, and that the hippocampal @-rhythmappears in a rebound-like manner. The nerve stimulation is marked by the line at the bottom of the tracings; hd, right hippocampus dorsalis; hs, left hippocampus dorsalis. (From Corazza and Parmeggiani, 1961b.)
However, following the end of the stimulation period the &rhythm appears in a rebound-like manner.
*
Barbiturate or chloralose anesthesia suppresses the &rhythm (cf. also Green and Arduini, 1954).
**
In our experimental conditions full activation of group I11 fibers was avoided.
References p. 438-441
416
P. L. P A R M E G G I A N I
After coagulation of the septum the hippocampogram becomes very similar to the neocorticogram, and sciatic stimulation elicits, at all frequencies and voltages used, only desynchronizationof the bioelectrical activity of the hippocampus (Figs. 2 and 3).
A hd hs
k
B h hS
d
~
b
M
$
l
$
#
:
’
~
0
MwWAdwJM@ , Zsec
,
Fig. 2. Effects exerted on the bioelectricalactivity of the hippocampus by repetitive electrical stimulation of the sciatic nerve before and after septal coagulation. Unanesthetized, curarized cat. Effects of sciatic stimulation (lOO/sec, 1 msec, 1 V) before (A) and after (B) coagulation. Note that after septal coagulation the peripheral stimulation elicits only desynchrouking effects. The nerve stimulation is marked by the line at the bottom of the tracings; hd, right hippocampus dorsalis; hs, left hippocampus dorsalis. (From Corazza and Pameggiani, 1961b.)
These results are consistent with the idea that there are two separate afferent systems to the hippocampus dorsalis, the one synchronizing(&rhythm), the other desynchronizing (Corazza and Parmeggiani, 1961a, b; cf. also Torii, 1961). At the hippocampal level, the two systems act concomitantly; consequently, mixed frequency patterns are seen in the absence of any predominant driving influence. A clear-cut prevalence of the effects of the one or of the other system results in the synchronization or desynchronization of the hippocampogram for a limited time, during and/or after peripheral stimulation. The central course of the pathway eliciting the &rhythm in the cat’s hippocampus dorsalis upon sciaticstimulationwas subjectedto further study(Corazza and Parmeggiani, 1963). The results can be summarized as follows. (i) The hippocampal &rhythm elicited by sciatic stimulation is unaffected after unilateral or bilateral coagulation of any one of the following bulbo-pontine structures: fibrae arcuatae internae, n. cuneatus, n. dorsalis nervi vagi, n. gracilis, nn. pontis, n. raphes, n. reticularis gigantocellularis, n. reticularis paramedianus, n. reticularis parvocellularis, n. reticularis pontis caudalis, n. reticularis pontis oralis, reticularis tegmenti pontis, n. reticularis ventralis, n. tractus solitarius, tractus spino-thalamicus. (ii) At the midbrain level, coagulation of the formatio reticularis depresses the hippocampal &rhythm only when placed bilaterally*. As to the other structures explored, unilateral or bilateral coagulations proved to be ineffective when placed in any one of the following structures : commissura posterior, fasciculus longitudinalis
* The fact is here recalled that in these experiments the coagulation was never so extensive as to involve the whole formatio reticularis.
6-RHYTHM
F U N C T I O N A L S I G N I F I C A N C E OF
417
pd ~ ~ .r ~ .,..,~ .~ ,,.~ ~ ~ ~ ,~ ~ ~ ,.~ ,. od
I I I
0s
~~~~--~-~.~~..L..._~.-.LLLUIL
ps
%&j..P"Ju'*.\1?~ ~ ~ : ,q ~ :
fs
~
B
hs ' fd
,
~
~
-
*
\
-
,
~
d a
~
~
~
~
~
,
~
~ 'Hi;i
. ~
~
~
.
,
$ ~ W W ~ ~ . 4 h w * V k * . * & w .
W++.h4-..*)
Pd ~ -
. . r c - : 7 , ; ~ .
~
;.cr-r,~~-~~~~.,.~:~.;'.~~~.~-
:+a
~
,
~
~
~
:
~
-
~
*
-
~
~
~
~
~
&
v
~
~
,
h
~
$ % :-
A'VnAH.wI
dorsalis (Schutz), lemniscus medialis, n. interpeduncularis, n. limitans, n. paralemniscalis, n. tegmenti, nn. tegmenti Guddenii, praetectum, substantia grisea centralis, substantia nigra, tractus habenulo-peduncularis. (iii) Coagulations (uni- or bilateral) of diencephalic structures showed that the appearance of the 6-rhythm in the hippocampus dorsalis upon sciatic stimulation is only dependent on the integrity of structures of the hypothalamus medialis (Fig. 4): area hypothalamica anterior, area hypothalamica dorsalis, area hypothalamica posterior, area tuberalis, n. ventralis medialis hypothalami, pars anterior n. paraventricularis, pars tuberalis n. periventricularis. On the other hand the 0-rhythm is unaffected by unilateral or bilateral coagulation of the following structures : area hypothalamica lateralis, campi Forelii, fasciculus medialis telencephali (median forebrain bundle), fornix, n. mammillaris medialis (pars centralis), n. mammillaris medialis (pars intermedia), n. paraventricularis (pars dorsalis), n. praemammillaris, n. subthalamicus, n. supramammillaris, n. supraopticus, n. tuberomammillaris, pedunculus mammillaris, tractus mammillo-thalamicus, zona incerta. The 8-rhythm is equally unmodified after coagulation of the thalamic nuclei, either specific or non-specific.
.
~
."Cn..-,w,*-*4v$/%fb+
Fig. 3. Effects exerted on the bioelectrical activity of the hippocampus by repetitive electrical stimulation of the sciatic nerve before and after septa1coagulation. Unanesthetized, curarized cat. Note that the same stimuli (20/sec, 1 msec, 2 V) exerting synchronizing effects on the bioelectrical activity of the hippocampus before coagulation (A), after coagulation (B), desynchronize the hippocampogram in a way comparable to that of the neocortical arousal. The nerve stimulation is marked by the line at the bottom of the tracings; os, left occipital; ps, left parietal; fs, left frontal; hs, left hippocampus dorsalis; fd, right frontal; pd, right parietal; od, right occipital. (From Corazza and Parmeggiani, 1961b.)
References p. 438441
,
~
~
i
S
418
P. L. P A R M E G G I A N I
The results reported here indicate that the neural elements responsible for the appearance of the &rhythm in the hippocampogram undergo recognizable structural grouping only from the mesencephalic-hypothalamiclevel, and run upwards to the septum : the synchronizing afferent system to the hippocampus seems, therefore, to originate from the reticular formation of the midbrain. On the basis of the experiments of sciatic stimulation in the cat it appears, moreover, that the &rhythm results from moderate reticular activity, and desynchronization occurs during strong reticular activation. The functional significance of the hippocampal 8-rhythm can be fully understood only if the desynchronization of the bioelectrical activity of the hippocampus dorsalis is also taken into account. Further research has to be performed to solve the hodological problem underlying this desynchronization. However, the data already available show that hippocampal desynchronization can still be obtained in cats after hypothalamic, thalamic and septa1lesions, and that hippocampal and neocortical desynchronization are probably closely interrelated. In the rabbit, the cat and the monkey, the
Fig. 4. Effects exerted on the bioelectricalactivity of the hippocampus by repetitive electrical stimulation of the sciatic nerve before and after coagulation of hypothalamic structures. Unanesthetized, curarized cats. A X , D-F, G-I refer to three different preparations. In A, D, G and in C, F: I (respectively before and after coagulation) EEG tracings are recorded from the neocortex and hippocampus (RF, right frontal; RH, right hippocampus dorsalis; LH, left hippocampus dorsalis; LF, left frontal). The nerve stimulation (20/sec, 1 msec, 1 V) is marked by the line at the bottom of the tracings. In B, E, H histological sections are reproduced (Nissl staining)to illustrate the location and the extent of the coagulations in the hypothalamus of the three animals. (From Corazza and Parmeggiani, 1963.)
F U N C T I O N A L S I G N I F I C A N C E OF 6 - R H Y T H M
419
ease with which the hippocampal 0-rhythm can be observed is inversely related to the development of the neocortex, so that it can be surmised that the desynchronization of the hippocampogram and the depression of the 0-rhythm depend a t least in part on the increased influence of the neocortex on the bioelectrical activity of the hippocampus. In higher mammals, the reticulo-neocortico-hippocampal system may possibly become so dominant that the activity of the hippocampus is maintained desynchronized also at moderate levels of reticular activation.
(B) Neocortical influences of the hippocampal output associated with the 8-rhythm The influence of the hippocampal output, when the &rhythm is observed in the hippocampogram, was detected in the bioelectrical activity of the neocortex of curarized cats. In fact: (i) during the appearance of the hippocampal &rhythm low-frequency recruiting waves depending upon impulses originating in the hippocampus itself wax and wane in the neocorticogram (Corazza and Parmeggiani, 1960a, b); (ii) the amplitude and course of induced neocortical d.c. potential shifts are modified by the hippocampal output during &rhythm (Parmeggiani and Rabini, 1964); and (iii) the nature of the changes in the evoked responses of the auditory and visual cortex which accompany induced arousal reactions depends upon the presence or absence of the hippocampal output associated with the &rhythm (Parmeggiani and Salvatorelli, 1961; Parmeggiani, 1962b, c). (i) Hippocampal &rhythm and neocortical low-frequency recruiting waves. The effects upon the bioelectrical activity of the hippocampus and neocortex (monopolar recording technique) elicited by repetitive electrical stimulation (4-20-1 OO/sec, 1 msec, 0.5-3 V) of the sciatic nerve were investigated, and the following evidence was assembled: ( a ) With variable latency (0.1 to 15 sec) with respect to the beginning of the hippocampal synchronization (@-rhythm)the neocortical recordings show, besides the fast activity typical for the arousal reaction, a rhythm of low-frequency waves (Figs. 5 , 6 and 7) increasing in voltage (100-200 pV, 2-5/sec), and often subjected to regular variations in amplitude (waxing and waning); (b) It may be inferred that the neocortical low-frequency waves are local in origin and not due to physical spread of hippocampal events, because of the above-mentioned latency (Fig. 8), and because the waves are suppressed by neocortical cooling (Fig. 9) and strongly depressed by topical application of cocaine; (c) Furthermore, coagulation experiments (Fig. 10) of rhinencephalic structures (septum, fornix and mammillo-thalamic tracts) and of non-specific thalamic nuclei show that the neocortical low-frequency waves related to the &rhythm depend upon impulses fired by the hippocampus itself and that thalamic non-specific nuclei are involved in the origin of such waves. The same phenomenon that is observed in curarized cats (i.e. $-waves in the hippocampogram and low-frequency waves in the neocortical recordings) could also be proved to occur in unrestrained cats, in attentive or excited wakefulness and particularly during activated sleep (low-voltage fast EEG phase of sleep). In such conditions, on a background of fast activity, both the neocortical recordings and the leads from subcortical structures show low-frequency (4-6/sec) waves (Figs. 11 and 12). Discrete References p. 438-441
420
P. L. P A R M E G G I A N I
05
PS
fd
j
Fig. 5. Effects exerted on the bioelectrical activity of the hippocampus and neocortex by repetitive electrical stimulation of the sciatic nerve. Unanesthetized,curarized cat. Note that sciatic stimulation (lOO/sec, 1 msec, 2 V) elicits the 0-rhythm in the hippocampogramsand a rhythm of low-frequency waves in the occipital and parietal leads. The nerve stimulation is marked by the line at the bottom of the tracings; os, left occipital; ps, left parietal; fs, left frontal; hs, left hippocampus dorsalis; hd, right hippocampusdorsalis; fd, right frontal; pd, right parietal ;od, right occipital. Same conventionsapply to Figs. 6 to 10. (From Corazza and Parmeggiani, 1960b.)
septa1 lesions, preventing the &rhythm from appearing in the hippocampogram, also suppress the neocortical and subcortical low-frequency waves. It is interesting to note that, if one accepts the oneiric correlate of activated sleep (Aserinsky and Kleitman, 1955; Dement and Kleitman, 1957a, b; Dement, 1958), one might infer that the appearance of the hippocampal &rhythm and of archicortically-induced low-frequency waves in neocortical and subcortical recordings during this phase of sleep reveals the
od
+......=
Fig. 6. Effects exerted on the bioelectrical activity of the hippocampus and neocortex by repetitive electrical stimulation of the sciatic nerve. Unanesthetized,curarized cat. Note that the hippocampal 0-rhythm appears in a rebound-like manner after the end of the sciatic stimulation (lOo/sec, 1 msec, 3 V); a rhythm of low-frequency waves is also evident at the same time in the occipital leads. (From Corazza and Parmeggiani, 1960b.)
F U N C T I O N A L S I G N I F I C A N C E OF 6 - R H Y T H M
42 1
*Sf
Fig. 7. Hippocampal 8-rhythm and appearance of low-frequency waves in neocortical leads. Unanesthetized, curarized cat. A and B are a continuous sequence. Note that a rhythm of low-frequency waves is evident in the neocortical leads during the spontaneous appearance of the hippocampal 8rhythm. (From Corazza and Parmeggiani, 1960b.)
importance of the archicortical component (affective, visceral) in the activity of the dreaming brain. In summary, these results support the hypothesis that the hippocampal &rhythm is related to a synchronizing hippocampal influence which counteracts the desynchronizing effects of the reticular activating system on the activity of neocortical and subcortical structures. (ii) Hippocampal &rhythm and induced neocortical d.c. potential shifts. The neocortical d.c. potential shifts elicited by sciatic stimulation (&2&100/sec, 1 msec, 0.5-3 V) were studied both before and after the appearance of the hippocampal &rhythm was prevented by means of septa1 coagulation. The results can be summarized by References p. 438441
422
P. L. P A R M E G G I A N I
o
~
p -
-
s
\
~
s
*\-
~
5
~
~
~
-
W
~
'T*w*w-*+-p
-
~
w
\
V
w
f h
S
hd -Lp?
~
fd + b --w v % -
~
~
~
~
J
~
P
pd 1
o+hL+%w-
',f I
:h
Fig. 8. Effects exerted on the bioelectrical activity of the hippocampus and neocortex by repetitive electrical stimulation of the sciatic nerve. Unanesthetized,curarized cat. Note that upon stimulation of the sciatic nerve (lOO/sec, 1 msec, 2 V) the rhythm of low-frequency waves appears in the neocortica1 leads after the beginning of the hippocampal 0-rhythm. (From Corazza and Parmeggiani, 1960b.)
0s
hs hd
od
Fig. 9. Effects exerted by topical cooling of the neocortex on the low-frequency waves elicited by repetitive electrical stimulation of the sciaticnerve. Unanesthetized, curarized cat. The temperature of the right occipital neocortex was reduced from 37 to 29". Note the effects of sciatic stimulation (100/sec, 1 msec, 1 V) after cooling: the rhythm of low-frequency waves is absent in the lead from the cooled neocortical area. (From Corazza and Parmeggiani, 1960b.)
stating that: (a) in the normal animal sciatic stimulation elicits surface-negative shifts (0.15-1 mV, 10-100 sec) of the neocortical d.c. potential; and (b) after septal coagulation the shiftselicited by the same stimuli are smaller in amplitude (20-60 %) than before the septal lesion (Fig. 13). This depression of the negative shifts does not depend upon variations of the effects of sciatic stimulation on the systemic arterial pressure, but appears to be related to the suppression of the hippocampal &rhythm as a consequence of septal coagulation. These results further support the hypothesis that during &rhythm the hippocampal output consistently influences the neocortical activity. (iii) Hippocampal 8-rhythm and neocortical evoked responses. The influence on the responsiveness of the primary auditory area exerted by repetitive electrical stimulation (100/sec, 1 msec, 0.5-3 V) of the sciatic nerve has been investigated. During and after the nerve stimulation the first positive component of the auditory responses to clicks is depressed, whereas the second positive (cf. Goldstein et al., 1959; Cenacchi and Par-
F U N C T I O N A L S I G N I F I C A N C E OF 8 - R H Y T H M
423
Fig. 10. Effects exerted on the bioelectrical activity of the hippocampus and neocortex by repetitive electrical stimulation of the sciatic nerve before and after coagulation of rhinencephalic structures. Unanesthetized, curarized cat. In A, sciatic stimulation (20/sec, 1 msec, 2 V) beforecoagulation. InB, histological sections (Nissl staining) showing the lesions produced by coagulation of the basal region of the septum. In C , sciatic stimulation (20/sec, 1 msec, 3 V) after coagulation. Note that after coagulation both the hippocampal 8-rhythm and the neocortical low-frequency waves are absent. (From Corazza and Parmeggiani, 1960b.)
meggiani, 1962, 1963) and the main negative components are enhanced. In such experimental conditions the sciatic stimulation does not depress in toto the auditory response to receptor stimulation (cf. also Bremer et al., 1960; Steriade and Demetrescu, 1962), as has been observed by others (Bremer and Bonnet, 1950; Gauthier et al., References p . 438-441
424
P. L. P A R M E G G I A N I
Fig. 11. Bioelectrical activity of cortical and subcorticalstructuresin the cat during relaxed or excited wakefulness. Unanesthetized, unrestrained cat. In A, the animal is relaxed at the beginning of the recording period, and becomes attentive towards the end. During the relaxed state, note: absence of h a v e s , lower frequency of neocortical activity, and spindle-like envelopes. Durhg the attentive state, note: clear-cut hippocampal &rhythm, low-frequency waves in central gray, subthalamus,parietal and occipital leads, and higher frequency of neocortical activity. In B, the animal is very excited and mewing. Note the hippocampal @-rhythmand low-frequency waves in subcortical and neocortical recordings. F, frontal; P, parietal; 0, occipital; CG, substantia grisea centralis; Ht, hypothalamus; Hp, hippocampus dorsalis; St, subthalamus. (From Parmeggiani and Zanocco, 1963.)
ww&*h*t,w~-
%+ .Zrec
,
Fig. 12. Cortical and subcortical recordings during activated sleep in the cat. Unanesthetized, unrestrained cat. In A, behavior of the animal during the phase of activated sleep. In B, hippocampal 0rhythm and low-frequency waves in neocortical and subcortical leads. P, parietal; 0, occipital; M, substantia grisea centralis; H, hippocampus dorsalis; SN, substantia nigra. (From Parmeggiani and Zanocco, 1961.)
1956; Desmedt and La Grutta, 1957; Bremer and Stoupel, 1958, 1959; Bremer et ul., 1960). The observed changes are not only due to the ascending influences of the reticular system, as pointed out by many authors (Bremer and Stoupel, 1958, 1959; Dumont and Dell, 1958,1960; Bremer et al., 1960; Steriade and Demetrescu, 1962; Demetrescu et ul., 1965), but also to hippocampal influences effective when the 0-rhythm appears
in the hippocampogram. In fact, as soon as the 8-rhythm resulting from sciatic stimulation is prevented from appearing by coagulating the septum, the auditory responses are no longer modified by sciatic stimulation as they had been before the septa1lesion. Furthermore, the second positive and the main negative components of the response
FUNCTIONAL SIGNIFICANCE OF @-RHYTHM
425
Fig. 13. Effects of septal coagulation on neocortical d.c. potential shifts elicited by sciatic stimulation. Unanesthetized, curarized cat. Upper tracings in A, B, D, E show surface-negative shifts (upward deflections) elicited by sciatic stimulation, and lower tracings respectively show the effects of sciatic stimulation on hippocampal activity (A, D) and systemic arterial pressure (B, E). The tracings of A and B refer to the intact animal; those of D and E were recorded after septal coagulation. The nerve stimulation (4/sec, 1 msec, 1 V) is marked by the line at the bottom of the tracings. In C, a histological section is reproduced (Nissl staining) to illustrate the location and the extent of the coagulation in the septum of this animal. Note the depression of the d.c. shifts after coagulation. (From Parmeggiani and Rabini, 1964.)
appear depressed (Figs. 14 and 15). The nature of the changes induced in the auditory responses by the hippocampal output during the &rhythm shows that this output has modified the activity of neocortical inhibitory and excitatory circuits (Cenacchi and Parmeggiani, 1963). Other experiments have shown that the neocortical responses to photic stimuli are enhanced by the hippocampal output during the appearance of the 8-rhythm (Parmeggiani, 1962~). In summary, the accumulated evidence indicates that impulses from the hippocampus intervene in the regulation of sensory mechanisms. It is, therefore, significant that the 0-rhythm appears in the hippocampogram during the orientation reaction (Grastyhn et al., 1959; RadulovaEki and Adey, 1965), conditioning (Grastyan et al., 1959) and approach behavior (Holmes and Adey, 1960; Adey, 1961;Grastykn et al., 1965). Moreover, among other factors, the effects of the hippocampal output during the 0-rhythm should also be taken into account to explain the nature of the changes in the evoked neocortical and thalamic responses occurring in arousal and activated sleep. (cf. Jouvet, 1962; Cordeau et al., 1965; Favale et al., 1965; Rossi et al., 1965). ( C ) Thalamic influences exerted by the hippocampal output during low-frequency stimulation of the hippocampus dorsalis The indirect evidence of the hippocampal influences on the neocortex, as shown previously by the modification of various kinds of neocortical activity during the appearance of the hippocampal &rhythm, made it appear profitable to study the subcortical mechanisms involved. To provide information about this problem, the effects exerted by repetitive electrical stimulation (2-5/sec, 0.5 msec, 2-12 V) of the hippocampus dorsalis on thalamic unit activity were studied on curarized cats (Manzoni and Parmeggiani, 1964a, 1965). The frequency of repetitive stimulation was chosen within the range of the rhythm References p.^438-441
.- ._..-._- .... ..
-. . -
, . .
. . . - . .
V
F U N C T I O N A L S I G N I F I C A N C E OF
6-RHYTHM
427
Fig. 15. Schematic representation of the variations elicited in the components of the primary auditory responses by repetitive electrical stimulation of the sciatic nerve before and after septal coagulation. Unanesthetized, curarized cat. Variations in amplitude of the first positive, second positive and negative components of the response, as calculated before (A, B, and C respectively)and after (A1, B1, and CI respectively) septal coagulation, are compared with the mean amplitude of each component. The mean amplitude is calculated from 12 responses recorded immediately before the sciatic sthulation, and is shown in each graph by the horizontal heavy line, the two broken lines giving the standard deviation. The segments on the abscissa indicate the periods during which three auditory responses were recorded, the time being computed from the beginning of sciatic stimulation. Such a stimulation is marked by the segment with two arrows. For each group of three responses, the mean amplitude of each component is plotted in the proper graph as a filled circle, the standard deviation being given by the vertical bar. (From Parmeggiani, 1962b.)
observed for the 6-waves in order to evaluate the driving power of the hippocampal output at physiological levels of activity. It is worth while recalling that bursts of hippocampal unit activity are associated with such waves, a condition that would be similar to that realized artificially in the present research by repetitive low-frequency stimulation of the hippocampus. References p . 438441
428
P. L. PARMEGGIANI
E
-i/"r" ... f
i/y"-
....
0
........ 50H1
........ IOOHZ
-6#&
Is ........50 ......... ........ ......... 100 Hz 100 Hz 50Hr
Fig. 16. Hippocampal influence on the activity of thalamic neurones. Unanesthetized, curarized cats. In A-E, responses to stimulation of the left hippocampus dorsalis (A, B, 8 V; C, 10 V; D, DI, Dz, 6 V; E, El, 8 V) recorded respectively from the n. lateralis dorsalis (A, B) and the n. anterior ventralis (C, D, D1, D2, E, El) of the left-side thalamus. Note that in B, upper tracing shows the response recorded from the right hippocampus dorsalis and lower tracing the response from the n. lateralis dorsalis; D, D1 and Dz, were recorded respectively at the beginning, during and at the end of hippocampal stirnulation; E and El are a continuous sequence. In F, the left column shows seizure waves of the right hippocampus dorsalis, and the right column the unit activity recorded simultaneously from the n. anterior ventralis of the left-side thalamus. In all tracings, upward deflections indicate negativity of active electrode. For further explanation see text. (From Manzoni and Parmeggiani, 1964a.)
Unit responses to repetitive hippocampal stimulation have been recorded thus far from the following ipsilateral thalamic nuclei : anterior ventralis, anterior medialis, ventralis anterior, medialis dorsalis, paracentralis, centralis lateralis, lateralis dorsalis and habenularis. Two principal types of response were observed. (i) Each shock delivered to the hippocampus elicits a triphasic or diphasic wave on
F U N C T I O N A L S I G N I F I C A N C E OF
0-RHYTHM
429
which a unit response is often superimposed as a burst (20&500/sec) of spikes (0.3-1.2 mV). The unit responses often wax and wane* or completely disappear during the stimulation period (Fig. 16). The latency of the first spike of the burst is variable and ranges from 6 to 72 msec**. (ii) Hippocampal stimulation influences the unit activity in the thalamus in such a way that random spikes (0.3-1.2 mV) become grouped in bursts (200-400/sec), and/or previously silent neurones are recruited (Fig. 16). The latency of the first spike of the bursts described ranges from 5 to 160 msec**. Concerning the functional relationship between the hippocampus and the previously mentioned nuclei of the thalamus, some other data deserve mention. During hippocampal seizures induced by repetitive electrical stimulation at higher frequency (loo/ sec, 0.5 msec, 2-12 V) of the ipsilateral hippocampus, each hippocampal seizure-wave is seen to be related to a burst of thalamic unit activity (Fig. 16). If the seizure spreads to the thalamus itself, then each hippocampal spike and wave complex is associated with a spike and wave complex in the thalamus. In our experimental conditions the hippocampal output is able to induce rhythmic variations in the activity of thalamic neurones. Besides the nuclei of the anterior group (see also Green and Adey, 1956; Adey et al., 1958; Cazard, 1963), many others are affected by hippocampal stimulation (see also Green and Adey, 1956; Adey et al., 1958). Therefore, not only the neurones which underlie the regulation of the activity of the gyrus cinguli, but also those projecting to other neocortical regions are influenced by the hippocampus. It follows, finally, that the cortical fields of projection of such thalamic neurones must in turn be affected by the hippocampal output. The hippocampal control of cortical neurones has already been proved for units of the gyrus cinguli (Manzoni and Parmeggiani, 1964b), whereas for other neocortical areas only indirect evidence suggests that a similar condition exists (Green and Morin, 1953; Elul, 1964). As far as the hippocampal &rhythm is concerned, the nature of the thalamic responses to repetitive hippocampal stimulation studied in the present research allows one to propose that a periodical activation of the non-specific nuclei of the thalamus underlies the rhythms of low-frequency recruiting waves which appear in the neocortical recordings during the hippocampal &rhythm (cf. Bi, p. 419). Since, as previously shown, such neocortical waves depend on impulses fired by the hippocampus, there is good reason to relate these impulses with the rhythmic bursts of hippocampal unit activity associated with the 0-waves (Arduini and Pompeiano, 1955; Green and Machne, 1955; Green et al., 1960, 1961 ; Fujita and Sato, 1964). All this, in electrophysiological terms, means that the hippocampal action during the appearance of the hippocampal &rhythm is a synchronizing one (Corazza and Parmeggiani, 1960a, b). A hippocampal rhythmic activation of the thalamic non-specific nuclei may also explain the fact that the neocortical low-frequency recruiting waves (cf. Bi, p. 419) and the
* Referring to the number of spikes of the unit response. ** The actual latency figures for each nucleus explored are given in Table I of the original paper
(Manzoni and Parmeggiani, 1965). References p. 438441
430
P. L. P A R M E G G I A N I
components of the neocortical evoked responses that are modified during the appearance of the hippocampal 6-rhythm (cJ Biii, p. 422), depend on the activity of the superficial layers of the neocortex (Corazza and Parmeggiani, 1960b; Cenacchi and Parmeggiani, 1963), where the non-specific afferents terminate. Finally, the data concerning the neocortical d.c. shifts (cf. Bii, p. 421) may fit into this line of thinking in so far as dendritic mechanisms may underlie their origin (Brookhart et al., 1958;Caspers, 1959; O'Leary and Goldring, 1964). ( D ) Interaction phenomena appearing at the thalamic level between hippocampal and reticular influences
The study of the interaction between hippocampal and reticular influences (Manzoni and Parmeggiani, 1964b, 1965) was performed because of the working hypothesis that the hippocampus operates in opposition to the reticular activating system during the appearance of the &rhythm. The effects of contemporary hippocampal and sciatic repetitive stimulation were studied in some of those thalamic nuclei where the discharge was previously seen to be modified or elicited by hippocampal stimulation (cf. C, p. 425). The stimulation of the sciatic nerve (20-100/sec, 1 msec, 0.5-3 V) was preferred to the direct stimulation of the midbrain reticular formation with the view of achieving a more physiologicalactivation of the ascending reticular system. The results obtained under such experimental conditions can be grouped as follows. (i) The unit responses elicited by low-frequency stimulation of the hippocampus dorsalis in the nuclei anterior ventralis (Fig. 17), lateralis dorsalis and medialis dorsalis are not consistently modified during contemporary sciatic stimulation. Sometimes a moderate enhancement of the unit responses to hippocampal stimulation can be observed during and for a few seconds after the sciatic stimulation. (ii) Stimulation of the sciatic nerve resulted in the depression or abolition of the unit responses to contemporary low-frequency hippocampal stimulation in the nuclei ventralis anterior (Fig. 18), centralis lateralis and sometimes in the nucleus medialis dorsalis. This study has shown that the thalamic limbic pathway to the gyrus cinguli is relatively independent from any reticular control and therefore is able to convey to this cortex patterns of activity that closely resemble those of the hippocampal output. The conditions do exist for the reactivation of the hippocampus via the gyrus cinguli (cf. Brodal, 1947; Adey, 1959): so a circuit is closed which is capable of carrying and maintaining stable patterns of activity. Also the projections of the nuclei dorsalis lateralis, and in part medialis dorsalis, are not subjected to apparent reticular influence. On the contrary, as far as the nuclei ventralis anterior, centralis lateralis, and in part medialis dorsalis are concerned, the hippocampal effects interact with those of the ascending reticular system. In this event it must be assumed that at the subcortical level integration phenomena might already occur, and that the hippocampal influence on the neocortex could in turn be modulated or suppressed outright by the reticular system.
F U N C T I O N A L S I G N I F I C A N C E OF 0 - R H Y T H M
43 1
Fig. 17. Interaction phenomena between the effects of contemporary hippocampal and sciatic stimulation appearing at the thalamic level. Unanesthetized, curarized cat. In A-Az, responses of the n. anterior ventralis to repetitive stirnulation (5/sec, 0.5 msec, 4 V) of the left hippocampus dorsalis respectively before (A), during (A1) and after (Az) repetitive stimulation (20/sec, 1 msec, 2V) of the sciatic nerve. Dots indicate the beginning and the end of sciatic stimulation. Note that during sciatic stimulation the unit responses are not depressed. (From Manzoni and Parmeggiani, 1965.) References p. 438441
432
P. L. P A R M E G G I A N I
Fig. 18. Interaction phenomena between the effects of contemporary hippocampalandsciaticstimulation appearing a t the thalamic level. Unanesthetized, curarized cat. In A-A4, responses of the n. ventralis anterior to repetitive stirnulation (5/sec, 0.5 msec, 8 V) of the left hippocampus dorsalis respectively before (A), during (A1 and As) and after (A3, 11 sec; A4, 20 sec) repetitive stimulation (2O/sec, 1msec, 2V) of thesciaticnerve. Dotsindicatethebeginningandtheend of sciaticstimulation; A1 and A2 are a continuous sequence. Note that during and after sciatic stimulation the unit responses are markedly depressed. (From Manzoni and Parmeggiani, 1965.)
F U N C T I O N A L S I G N I F I C A N C E OF
8-RHYTHM
433
(E) Behavioral eflects of the suppression of the hippocampal @rhythm The behavioral implications of the working hypothesis were studied by experiments performed on unrestrained cats before and after discreteseptalcoagulation suppressing the hippocampal @rhythm (Parmeggiani and Zanocco, 1963; Lena and Parmeggiani, 1964). After such lesion (cf. also Brady and Nauta, 1953, 1955; Green and Arduini, 1954; Harrison and Lyon, 1957; King, 1958; Brugge, 1965; Nielson et al., 1965) the waking behavior of the animal is scarcely modified at all. However, it is noteworthy that certain environmental situations or stimuli, which do not seem to annoy a normal cat, can irritate one with a septal lesion. In fact, a certain hyper-reactivity is visible in its relations with the experimenter and with other cats, towards which it often adopts a rather aggressive attitude. At times it may be no less docile than any normal cat, provided it is treated gently, since the aggressive behavior is always in reply to stimuli and is short-lived. However, the fact that such stimuli are usually insufficient to trigger off a reaction in the cat before coagulation of the septum, must certainly be given considerable weight. When placed in the experimental cage in the non-soundproof room the animal remains motionless, and gives the impression of being apprehensive and of encountering some difficulty in curling up and in falling asleep. Once asleep, the slow-wave EEG phase is less frequently followed by activated sleep, than it was in the same animal and environmental situation before the septal coagulation. It can be observed that the cat wakes up suddenly during the flattening of the EMG, instead of entering the activated sleep phase. If activated sleep develops, it lasts in general for a shorter period of time than jn the normal animal. During activated sleep there is always desynchronized activity in the hippocampogram and never &rhythm patterns. Low-frequency waves are also no longer apparent either in neocortical or subcortical recordings, which only show low-voltage fast waves. The EMG of the neck muscles appears flat. In factual terms, after the septal lesion, the average number and mean total duration of the activated sleep phases recorded during each standard session fall approximately to 60 % of the average values observed before septal coagulation (Fig. 19). As to the results yielded by the experimental sessions in a soundproof room, their evaluation shows that in such conditions the effects of septal lesions on activated sleep are generally less clear-cut than when the animal is not shielded, or in certain conditions completely fail to appear. As regards the effects on waking behavior of septal lesions preventing the &rhythm from appearing in the hippocampogram, these observations incline one to the view that the hippocampal mechanism related to the &rhythm, and operating as a negative feedback with respect to the reticular activating system, might be important above all in relation to the control of the reactivity to environmental stimuli. It would seem that such a feedback improves the animal's adaptation to its surroundings and favors the appearance of behavioral patterns of trophotropic* character (Parmeggiani, 1960; cf. also MacLean, 1958a, b). Actually, it is well known that sleep has a particularly important role among such patterns. One may note in this connection that the animal
*
This term is used according to the concepts developed by Hess (1948, 1949).
References p . 438441
434
P. L. PARMEGGIANI
Fig. 19. Effects of septal coagulation on activated sleep. Unanesthetized, unrestrained cat. In A and C , bioelectrical patterns of activated sleep before and after coagulation respectively. In B, histological section showing extent and location ofthecoagulation. In D, the first group ofhistograms refers to the average number of activated sleep phases per session (frequency), and the second one to their mean total duration per session (duration). In both groups the first histogram (a) refers to average values (equated to 100) before coagulation, the second one (b) to average values (percentage of a) after coagulation. Vertical bars indicate standard error. Lettering is as follows: P, parietal; 0, occipital; RHp, right hippocampus; LHp, left hippocampus; EMG, electrogram of the neck muscles. Note that after coagulation the hippocampal 0-rhythm is suppressed and that both frequency and duration of activated sleep phases are markedly reduced. (From Lena and Parmeggiani, 1964.)
with a septal lesion is less likely to curl up than the normal one. This observation agrees with the results of recent research, attributing to the hippocampus a pre-eminent function in the preparatory phase of sleep (Parmeggiani, 1959, 1960, 1962a). The fact that septal coagulation, which suppresses the hippocampal &rhythm, is followed by a consistent reduction both in the average number and in the mean total duration of the periods of activated sleep during sessions of standard duration in cats placed in the non-soundproof room, seems to indicate that the hippocampus stabilizes and maintains this phase of sleep when its bioelectrical activity is characterized by the
F U N C T I O N A L S I G N I F I C A N C E OF
0-RHYTHM
43 5
&rhythm. In the normal animal the increased reticular activity during the phase of activated sleep (Parmeggiani and Zanocco, 1963) would be counterbalanced by the opposite action of the hippocampus, while such action would be absent or depressed in the animal with the septal lesion. On the other hand, it is noteworthy that in the soundproof room the reduction in the average number of phases of activated sleep and in their mean total duration per session as a result of the septal lesion may be less pronounced or absent altogether. On the basis of the foregoing results it seems justifiable to conclude that the imbalance of the central regulating mechanism, due to suppression of the hippocampal feedback after septal coagulation, is practically latent both in the waking state and in activated sleep if adequate stimuli are not present to reveal it.
( F ) Behavioral effects of repetitive electrical stimulation of the hippocampus dorsalis, the Jimbria and the fornix The very complex results of this study are only briefly summarized here to show that, in the unrestrained cat, discrete activation of the hippocampus dorsalis can elicit behavioral responses as significant as those observed during hippocampal seizures. The experiments were performed by means of the Hess technique (1932). Electrical stimulation was carried out using delayed condenser discharges with a rising phase of 10 msec duration (Wyss, 1950). The reader is referred to the original papers for more details on technique and results (Parmeggiani, 1959, 1960). The nature of the behavioral effects of repetitive stimulation of the hippocampus dorsalis, the fimbria and the fornix is strictly dependent on the stimulation intensity. (i) With low intensity (0.2-2 V, depending on the frequency u s e d 4 or 8.5/sec) a wide spectrum of reactions is obtained such as: orientation reactions, motor activity to change the position of the body, vegetative responses as salivation, protrusion of the nictitating membranes, mydriasis, increase in the respiratory rate etc., as well as somatomotor patterns of trophotropic character (grooming, scratching etc.) or related to sleep behavior (stretching, yawning, rolling up) and to affective behavior (mewing, purring, friendliness,apprehension, restlessness etc.). The responses that are components of the sleep behavior may appear isolated or in a well defined sequence which gives way to sleep. In general, the stimulation effects previously described are variably combined together in patterns changing from stimulation to stimulation and from animal to animal. The resulting global behavior appears, however, very natural. (ii) Strong stimulation (1-3 V, depending on the frequency used-8.5-17/sec) elicits a stereotyped response which is initially characterized by arousal followed by an arrest reaction and successively by a relatively constant sequence of many of the effects obtainable at random with low-intensity stimulation. There is no reason to go into the details of this response, because it is evidently the result of a hippocampal seizure (cf. also MacLean, 1957). The behavioral effects of low-frequency, low-intensity stimulation of the hippocampus dorsalis, the fimbria and the fornix show that during discrete rhythmic activation of these structures, similar to that occurring with the &rhythm, the hippocampal References p. 438-441
436
P. L. P A R M E G G I A N I
output can also elicit significant responses in the somatic, vegetative and affective spheres. Because these responses are organized into variable patterns of global behavior, a goal-directed and integrative activity of the hippocampus seems to underlie the expressiveness of the behavior. CONCLUSIONS
On the basis of the preceding review of experimental data and related discussion, it is now possible to present a unitary view on some aspects of the functional significance of the hippocampal &rhythm. This effort can be considered valuable only within the limits of a working hypothesis stimulating research to prove or to reject it. The bioelectrical activity of the hippocampus is regulated by at least two afferent non-specific systems, the one synchronizing (0-rhythm), the other desynchronizing. In the rabbit, a tonic prevalence of the first is observed, whereas in the cat the effects of' the two systems appear to be in balance. In higher mammals, the condition of equilibrium is lost in favor of the desynchronizing system, which prevails tonically. This reversal in the influence of the two afferent systems seems to be related to the development of the neocortex. As far as concerns the cat, the animal providing the experimental evidence, some further conclusions can be drawn. The appearance of the 0-rhythm in the hippocampogram depends on a moderate level of reticular activation (Fig. 20). In this condition the hippocampal output elicits a rhythmic firing of subcortical neurones. Via the anterior thalamic nuclei and the
Fig. 20. Schematic drawing illustrating some of the functional relations discussed in the text. CA, commissura anterior; CC, corpus callosum; CM, corpus mammillare; CS, colliculus superior; FR, formatio reticularis; GC, gyrus cinguli; HB, habenula; HP, hippocampus; PO, pons; SGC, substantia grisea centralis; SP, septum; TH, thalamus.
F U N C T I O N A L S I G N I F I C A N C E OF
8-RHYTHM
437
gyrus cinguli a hippocampo-hippocampal circuit is closed, and reactivation of the hippocampus at the same rate is quite possible. This circuit as well as the other thalamo-neocortical circuits of the nuclei lateralis dorsalis and in part medialis dorsalis are not consistently affected by the ascending influences of the activating reticular formation. It follows that the hippocampal influences can reach without interference the neocortical neurones of the corresponding projection areas. On the contrary, at the level of the nuclei of the non-specific thalamo-neocortical system, such as the nuclei ventralis anterior, paracentralis, and centralis lateralis, the effects of the hippocampal output during the &rhythm could be depressed or suppressed outright by the reticular influences. Probably this does not happen in conditions of moderate reticular activation, because the hippocampal effects appear at the neocortical level also in the realm of the projections of the thalamic non-specific system. In fact, under the hippocampal influence, low-frequency recruiting waves appear in the neocorticogram, and the neocortical d.c. potential and the neocortical evoked responses show impressive changes. The nature of the changes in the evoked responses shows that the activity of neocortical inhibitory and excitatory mechanisms is modified : sensorial, perceptual and possibly mnemonic processes may in turn be affected by the hippocampal output. In conditions of moderate arousal the hippocampal output during the appearance of the &rhythm has enough driving power taexert definite control on the bioelectrical activity of subcortical and neocortical structures. In terms of behavior a central tendency is translated into behavioral patterns of hippocampal type which are directed to improve the animal's adaptation to its surroundings and to fulfil trophotropic goals. The &rhythm appears, therefore, as the bioelectrical correlate of a hippocampal negative feedback regulating and counterbalancing the effects of the reticular activating system and related ergotropic mechanisms. On the other hand, in conditions of strong reticular activation, the hippocampal desynchronization betrays an increased dominance of the ergotropic mechanisms. However, as soon as the level of reticular activation drops, the hippocampal &rhythm may appear as a rebound effect, giving way to the bioelectrical changes in the activity of subcortical and neocortical structures and to the behavioral patterns mentioned above. It is evident that the preceding considerations cannot be directly extended to higher mammals. The emphasis for regulation is now shifted from the hippocampus to the neocortex. The depression or suppression of the hippocampal &rhythm does not necessarily mean that the hippocampus has lost or basically changed its functional significance. The hippocampal action may only become more discrete and modulated. In conditions of normal or pathological synchronization of the hippocampal activity, however, the hippocampus may again acquire the relevant driving power lost through the development of the neocortex and may, therefore, elicit bioelectrical and behavioral effects, and possibly related contents of consciousness, revealing similarities with those appearing in lower mammals.
As has happened in the past, we are once again obliged to recognize the scholarly foresight of C. J. Herrick who proposed many of these functional relations hypothetically in 1933. References p . 438-441
438
P. L. P A R M E G G I A N I ACKNOWLEDGEMENTS
This work was supported by grants from the Consiglio Nazionale delle Ricerche (Rome, Italy) and the SchweizerischenNationalfond zur Forderung der wissenschaftlichen Forschung (Bern, Switzerland).
REFERENCES ADEY,W. R., (1959); Recent studies of the rhinencepholon in relation to temporal lobe epilepsy and behaviour disorders. Intern. Rev. Neurobiol., 1, 1 4 6 . ADEY,W. R., (1961); Studies of hippocampal electrical activity during approach learning. Brain Mechanisms and Learning. Oxford, Blackwell, pp. 577-588. ADEY,W. R., DUNLOP, C. W., AND S U N D ~ A MS.,) ,(1958); A survey ofrhinencephalicinterconnections with the brain stem. J. comp. Neurol., 110, 173-203. ARDUINI, A., E POMPEIANO, O., (1955); Attivitrl elettrica di singole unitsl dell'ippocampo e sue modificazioni con stimoli afferenti. Arch. Sci. biol. (Bologna), 39, 397-406. ASERINSKY, E., AND KLEITMAN,N., (1955); Two types of ocular motility occurring in sleep. J. appl. Physiol., 8, 1-10. BETELEVA, T. G., AND NOVIKOVA, L. A., (1960); Electrophysiological study of the hippocampus and its reaction to afferent stimulation. Sechenov physiol. J. U.S.S.R., (English edition), 46,48-59, BRADY, J. V., AND NAUTA,W. J. H., (1953); Subcorticalmechanisms in emotional behavior: affective changes following septal forebrain lesions in the albino rat. J. comp. physiol. Psychol., 46,339-346. BRADY,J. V., AND NAUTA,W. J. H., (1955); Subcortical mechanisms in emotional behavior: the duration of affective changes following septal and habenular lesions in the albino rat. J. comp. physiol. PsychoL, 48,412420. BREMER, F., ET BONNET,V., (1950); Interpretation des reactions rythmiques prolong& des aires sensorieks de l'kcorce ckrebrale. Electroenceph. clin. Neurophysiol., 2, 389-400. BREMER, F., ET S T O ~ E N., L , (1958);De la modifkation des rkponses sensorielles corticalesdans l'eveil rkticulaire. Acta neurol. psychiat. belg., 58, 401403. BREMER, F., ET STOUPEL,N., (1959); Facilitation et inhibition des potentiels 6voqueS corticaux dans l'eveil drbbral. Arch. internat. Physiol., 67, 240-275. BREMER, F., STOUPEL, N., ET VANREETH,P. CH., (1960); Nouvelles recherches sur la facilitation et l'inhibition des potentiels 6voqueS corticaux dans l'heil reticulaire. Arch. ital. Biol., 98, 229-247. BRODAL,A., (1947); The hippocampus and the sense of smell. A review. Brain, 70, 179-222. BROOKHART, J. M., h m ,A., MANCIA, M., AND MORUZZI, G., (1958); Thalamocortical relations as revealed by induced slow potential changes. J. Neurophysiol., 21,499-525. BRUCKE,F., PEmcm, H., PILLAT, B., UND DEISENHAMMER, E., (1959a); Ein Schrittmacher in der medialen Septumregion des Kaninchengehirnes. Pfliger's Arch. ges. Physiol., 269, 135-140. BRUCKE, F., PETSCHE,H., PILLAT, B., UND DEISENHAMMER,E., (1959b); Die Beeintlussung der 'Hippocampus-arousal-Reaktion' beim Kaninchen durch elektrische Reizung im Septum. PJliiger's Arch. ges. Physiol., 269, 319-338. BRUGGE, J. F., (1965); An electrographicstudy of the hippocampusand neocortex in unrestrained rats following septal lesions. Electroenceph. elin. Neurophysiol., 18, 36-44. H., (1959); Uber die Beziehungen zwischen Dendritenpotential und Gleichspannung an der CMPERS, Hirnrinde. Pfliiger's Arch. ges. Physiol., 269, 157-181. CAZARD,P., (1963); Activitb 6voqueeS chez le chat au niveau du group nucleaire antkrieur du thalamus et du corps mamillaire interne par stimulation de l'hippocampe dorsal et par stimulation somatique. J. Physiol. (Paris), 55, 220-221. CENACCHI, V., AND PARMEGGIANI, P. L., (1962); Analysis of the primary auditory response of the cat. Proc. XXII int. Congr. Physiol. (Leiden), No. 1011. CENACCHI, V., E PARMEGGIANI, P. L., (1963); Componenti della risposta acustica primaria. Arch. Sci. biol. (Bologna), 47,291-315. CORAZZA, R., m PARMEGGIANI, P. L., (1960a); Auftreten von niederfrequentenWellen im EEG der Katze bei Synchronisierung der bioelektrischen Tatigkeit des Hippocampus. Helv. physiol. pharmacol. Aeta., 18, C61464.
F U N C T I O N A L S I G N I F I C A N C E OF 0 - R H Y T H M
439
COUZZA,R., E PARMEGGIAM, P. L., (1960b); Sincronizzazionedei ritmi pulsatori dell'ippocmpo e comparsa di onde di bassa frequenza negli elettrogrammi registrati dal neocortex. Arch. Sci. biol. (Bologna), 44,435-463. CORAZZA, R., E PARMEGGLANI, P. L., (1961a); Sull'attivazione dell'ippocampo nel gatto per effetto della stimolazione periferica. Boll. SOC.ital. Biol. sper., 37, 377-378. CORAZZA, R.,E PARMEGGLANI, P. L., (1961b); Desincronizzazione dei ritmi bioelettrici dell'ippocampo. Arch. Sci. biol. (Bologna), 45,401-408. CORAZZA, R., AND PARMEGGIANI, P. L., (1963);Central course of the afferent system elicitingthe theta rhythm in the cat's hippocampus upon ischiatic stimulation. Helv. physiol. pharmacol. Acta, 21, ClW12. CORDEAU, P. J., WALSH,J., ET MAHUT,H., (1965); Variations dans la transmission des messages sensoriels en fonction des differents Ctats d'eveil et de sommeil. Aspects Anatomo-Fonctionnels de la Physiologie du Sommeil. Editions du Centre National de la Recherche Scientitique, Paris, No. 127, pp. 477-507. DEMENT, W., (1958); The occurrence of low voltage, fast electroencephalogram patterns during behavioral sleep in the cat. Electroenceph. clin. Neurophysiol., 10, 291-296. DEMENT, W., AND KLEITMAN, N.,(1957a);Cyclic variations of EEG during sleep and their relation to eye movements, body motility and dreaming. Electroenceph. clin. Neurophysiol., 9, 673-690. DEMENT, W., AND KLEITMAN, N.,(1957b); The relation of eye movements during sleep to dream activity: an objective method for the study of dreaming. J. exp. Psychol., 53, 339-346. DEMETRESCU, M., DEMETRESCU,M., AND IOSIF, G., (1965);The tonic control of cortical responsiveness by inhibitory and facilitatory diffuse influences. Electroenceph. clin. Neurophysiol., 18, 1-24. DESMEDT, J.E., AND LAGRUTTA, G., (1957); The effect of selective inhibition of pseudochokesterase on the spontaneous and evoked activity of the cat's cerebral wrtex. J. Physiol. (Lond.),136,2040. DUMONT, S., ET DELL,P., (1958); Facilitations sp6cifiques et non-spkcifiques des reponses visuelles corticales. J. Physiol. (Paris), 50, 261-264. DUMONT, S.,ET DELL,P., (1960);Facilitation reticulaire des mecanismes visuels corticaux. Electroenceph. clin. Neurophysiol., 12, 769-796. ELUL,R., (1964);Regional differences in the hippocampus of the cat. II.Projections of the dorsal and ventral hippocampus. Electroenceph. clin. Neurophysiol., 16,489-502. FAVALE, E., LOEB,C., MANFREDI, M., AND SACCO,G., (1965);Somaticafferent transmission and cortical responsiveness during natural sleep and arousal in the cat. Electroenceph. elin. Neurophysiol., 18, 354368. FUJITA,Y., AND SATO,T., (1964);Intracellular records from hippocampal pyramidal cells in rabbit during theta rhythm activity. J . Neurophysiol., 27, 1011-1025. GANGLOFF, H., AND MONNIER, M., (1956);Electrographicaspects of an 'arousal' or attention reaction induced in the manesthetized rabbit by the presence of a human being. Electroenceph. clin. Neurophysiol., 8,623-629. GAUTHIER, C., PARMA, M., AND ZANCHETII,A., (1956);Effect of electrowrtical arousal upon development and configurationof specific evoked potentials. Electroenceph. clin. Neurophysiol., 8,237-243. GOLDSTEIN, M. H., JR.,KIA",N. Y.-S.,ANDBROWN,R. M.,(1959);Responsesoftheauditorycortex to repetitive acoustic stimuli, J. acoust. SOC.Amer., 31, 356-364. GRASTYAN, E., KARMOS, G., VERECZEKEY, L., MARTIN,J., AND KELLENYI, L., (1965); Hypothalamic motivational processes as reflected by their hippocampal electrical correlates. Science, 149,91-93. GRASTYAN, E., LISSAK,K., MADARAsz, I., AND DONHOFFER,H., (1959); Hippocampal electrical activity during the development of conditioned reflexes. Electroenceph. clin. Neurophysid., 11, 409430. GREEN, J. D., (1964);The hippocampus. Physiol. Rev., 44,561-608. GREEN, J. D.,AND ADEY,W. R., (1956);Electrophysiological studies of hippocampal connectionsand excitability. Electroenceph. clin. Neurophysiol., 8, 245-262. GREEN, J. D., AND ARDUINI, A. A., (1954);Hippocampalelectricalactivity in arousal. J. Neurophysiol., 17.533-557. G&N, J. D., AND MACHNE, X., (1955);Unit activity of rabbit hippocampus. Amer. J. Physwl., 181, 219-224. GREEN,J. D., MAXWELL, D. S., AND PETSCHE,H., (1961); Hippocampal electrical activity. III. Unitary events and genesis o f slow waves. Electroenceph. clin. Neurophysiol., 13, 854.867. GREEN,J. D., MAXWELL, D. S., SCHINDLER, W. J., AND STUMPF,C., (1960); Rabbit EEG 'theta' rhythm: its anatomical source and relation to activity in single neurons. J. Neurophysiol., 23, 403420.
440
P. L. PARMEGGIANI
GREEN,J. D., AND MORIN,F., (1953); Hypothalamic electrical activity and hypothalamo-cortical relationships. Amer. J. Pkysiol., 172, 175-186. HARRISON, J. M., AND LYON,M., (1957); The role of the septal nuclei and components of the fornix in the behavior of the rat. J. comp. Neurol., 108, 121-138. HERRICK, C. J., (1933); The functions of the olfactory parts of the cerebral cortex. Proc. nut. Acad. Sci. (Wash.), 19, 7-14. HESS,W. R., (1932); Beitriige zur Physiologie des Hirnstammes. I. Die Methodik der lokalisierten Reizung und Ausschaltung subkortikaler Hirnabschitte. Leipzig, Thieme. HESS,W. R, (1948); Die funktionelle Organisation &s vegetativen Nervensystems. Basel, Schwabe. HESS,W. R., (1949); Das Zwischenhirn. Syndrome, Lokalisationen, Funktionen. Basel, Schwabe. HOLMES,J. E., AND ADEY,W. R.,(1960); Electrical activity of theentorhinal cortex during conditioned behavior. Amer. J. Physiol., 199,741-744. IWATA, K., AND Smm, R. S., (1959); Gxebello-hippocampal influences on the electroencephalogram. Electroenceph. elin. Neurophysiol,, 11, 439446. JOUVET,M.,(1962); Recherche sur les structures nerveuses et les mkcanismes responsables des d86rente.s phases du sommeil physiologique. Arch. ital. Biol., 100, 125-206. JUNG,R., UND KORNM~LLER, A,, (1938); Eine Methodik der Ableitung lokalisierter Potentialschwankungen aus subcorticalen Himgebieten. Arch. Psychiat. Nervenkr., 109, 1-30. KAWAMURA, H., NAKAMURA, Y.,AND TOKEANE, T., (1961); Effect of acute brain stem lesions on the electrical activities of the limbic system and neocortex. Jup. J. Physiol., 11, 564-575. KING,F. A., (1958); Effects of septal and amygdaloid lesions on emotional behavior and conditioned avoidance responses in the rat. J. nerv. ment. Dis., 126, 57-63. LENA,C., AND PARMEGGIANI, P. L., (1964); Hippocampal theta rhythm and activated sleep. Helv. . physiol. pharmacol. Acta, 22, 120-135. W.T., AND AKERT,K., (1955); Hippocampal seizure states in guinea-pig. Electroenceph. LIB-N, clin. Neurophysiol., 7, 211-222. LIBERSON, W.T., AND CADILHAC, J. G., (1954); Hippocampalresponses to sensory stimulation in the guinea pig. Electroenceph. clin. Neurophysiol., 6, 710-71 1. P. D., (1957); Chemical and electrical stimulation of the hippocampus in unrestrained MACLEAN, animals. 11. Behavioral fiqdings. A.M.A. Arch. Neurol. Psychiat., 78,128-142. MAC^, P. D., (1958a); Contrasting functions of limbic and neocortical systems of the brain and their relevance to psychophysiological aspects of medicine. A m r . J. Med., 25,611-626. MACLEAN, P. D., (1958b); The limbic system with respect to self-preservation and the preservation of the species. J. nerv. ment Dis., 127, 1-11. MACLEAN, P. D., HORWITZ, N. H., AND ROBINSON, F., (1952); Olfactory-like responses in pyriform area to non-olfactory stimulation. Yale J. Biol. Med., 25, 159-172. M ~ Z O MT.,, AND PMGGW, P. L., (1964a); Hippocampal control of the activity of thalamic neurones. Helv. physiol. pharmacol. Acta, 22, C28431. MANZOM, T., E PARMEGGIANI, P. L., (1964b); Ippocampo e attivita di neuroni talamici. Boll. SOC.ital. Biol.sper., 40,409-412. MANZOM, T.,AND PARMEGGIANI, P. L., (1965); The hippocampus as a pacemaker of the activity of thalamic neurones. Helv. physiol. phrmacol. Acta, 23, 180-198. MAYER,CH.,UND STUMPF, CH.,(1958); Die Physostigminwirkung auf die Hippocampus-Tatigkeit nach Septumlibionen. Nawtyn-Schmiedeberg's Arch. exp. Puth. Pharmuk., 234,490-500. NIELSON,H. C., MCIVER,A. H., AND BOSWELL,R. S., (1965); Effect of septal lesions on learning, emotionality, activity, and exploratory behavior in rats. Exp. Neurol., 11, 147-1 57. S., (1964); D-C potentials of the brain. Physiol. Rev., 44, 91-125. O'LEARY,J. L., AND GOLDRING, PARMEGGIANI, P. L., (1958); Himreizversuche mit Schlafeffekt aus subkortikalen Strukturen. Helv. physiol. phurmacol. Acta, 16, C73-06. P m c m u ~P., L., (1959); Schlafverhalten bei elektrischer Reizung von Hippocampus und Corpus mammillare der nichtnarkotisiertenfreibeweglichen Katze. Helv. physiol. p h a r m o l . Acta, 17, C34. P ~ G G W P., L., (1960); Reizeffekte aus Hippocampusund Corpus manmillare der Katze. Helv. physiol. pharmacol. Acta, 18, 523-536. P. L., (1962a); Sleep behaviour elicited by electrical stimulation of cortical and subP-MM, cortical structures in the cat. Helv. physiol. pharmacol. Acta, 20, 347-367. PARMEGGIANI,P. L., (1962b); Sincronkazione dell'attivit8 bioelettrica dell'ippocampo e risposta dell'area cerebrale primaria di proiezione acustica. Arch. Sci. biol. (Bologna), 46, 121-141. PARMEOOIANI, P. L., (19624; Hippocampal theta rhythm and neocortical responses to photic stimuli. Helv. physiol. phmacol. Acta, 20, C71-C73. I
FUNCTIONAL SIGNIFICANCE OF 6-RHYTHM
44 I
PARMEGGIANI, P. L., AND RABINI, C., (1964);Hippocampal theta rhythm and neocortical d.c. potential shifts. Helv. physiol. pharmacol. Acta, 22, C31X34. PARMEGGIANI, P. L.,UND SALVATORELLI, G., (1961); Archicorticale Regulierung der bioelektrischen Tatigkeit der primaren Horrinde. Helv. physiol. pharmacol. Acta, 19, C94-C96. PARMEGGIANI, P. L.,AND ZANOCCO, G., (1961);Cortical and subcorticalrecordings during low voltage fast EEG phase of sleep in the cat. Helv. physiol. pharmacol. Acta, 19, C97-4299. PARMEGGIANI, P. L.,AND ZANOCCO, G., (1963);A study on the bioelectrical rhythms of cortical and subcortical structures during activated sleep. Arch. ital. Biol., 101, 385412. PASSOUANT, P., PASSOUANT-FONTAINE, TH., ET CADILHAC, J., (1955); Hippocampe et rhction d'bveil. C.A. SOC.Biol. (Paris), 149, 164-166. PETSCHE, H., GOGOLAK, G., AND VANZWIETEN,P. A., (1965);Rhythmicity of septal cell discharges at various levels of reticular excitation. Electroenceph. din. Neurophysiol., 19, 25-33. PETSCHE,H., STUMPF, CH., AND GOGOLAK, G., (1962);The significance of the rabbit's septum as a relay station between the midbrain and the hippocampus. I. The control of hippocampus arousal activity by the septum cells. Electroenceph. clin. Neurophysiol., 14,202-211. RADULOVA~KI, M., AND ADEY,W. R., (1965); The hippocampus and the orienting reflex. Exp. Neurol., 12, 68-83. RIMBAUD, L., PASSOUANT, P., ET CADILHAC,J., (1955); Participation de l'hippocampe B la regulation des etats de veille et de sommeil. Rev. neurol., 93, 303-308. ROW, G. F., PALFSTINI,M., PISANO, M., AND ROSADINI, G., (1965);An experimental study of the cortical reactiyity during sleep and wakefulness. Aspects Anatomo-Fonctionnels de la Physiologie du Sommeil. Editions du Centre National de la Recherche Scientifque, Paris, No. 127,pp. 509 526 STERIADE,M., AND DEMETRESCU, M., (1962);Reticular facilitation of responses to acoustic stimuk Electroenceph. d in . Neurophysiol., 14,21-36. STUMPF,CH.,(1965);The fast component in the electrical activity of rabbit's hippocampus. Electroenceph. clin. Neurophysiol., 18,477-486. TOKIZANE,T., KAWAKAMI,M., AND GELLHORN, E., (1959); Hippocampal and neocortical activity in different experimental conditions. Electroenceph. elin. Neurophysiol., 11,431-437. Tom, S., (1961); Two types of pattern of hippocampal electrical activity induced by stimulation of hypothalamus and surrounding parts of rabbit's brain. Jap. J. Physiol., 11, 147-157. TORII,S.,AND KAWAMURA, H., (1960); Effects of amygdaloid stimulation on blood pressure and electrical activity of hippocampus. Jap. J. Physiol., 10, 374-384. WYSS,0.A. M., (1950); Beitrage zur elektrophysiologischen Methodik. 11. Ein vereinfachtes Reizgerat fur unabhangige Verhderung von Frequenz und Dauer der Impulse. Helv.physio1. pharmacol. Ada, 8, 18-24.
442
Functional Properties of the Hippocampus in the Sub-human Primate JOHN A. GERGEN Department of Physiology, Bowman Gray School of Medicine, Winston-Salem, N. C. ( U S A . )
At the present time, the functional role played by the hippocampal formation* in operational activities of the central nervous system remains uncertain. This situation is in contrast to the extensive and growing body of information on the intrinsic anatomical, physiological, and gross behavioral properties of this brain area (Green, 1964). In the following presentation, we wilI review a series of investigations dealing with electrophysiological and behavioral aspects of the hippocampus of a sub-human primate, the South American squirrel monkey (Saimiri sciurars). On the basis of these data it is possible to extend or revise previous conceptual models which attempt to relate anatomical and physiological properties and behavioral functions of the hippocampus. The present results and their interpretations thus provide a basis for further inquiry into structure-function relationships of the ‘limbic system’. SPONTANEOUS ACTIVITY A N D HIPPOCAMPAL AFTER-DISCHARGE
Spontaneous hippocampal electroencephalographic activity and electrically induced recurrent hippocampal seizures have been examined in 8 squirrel monkeys with chronically implanted electrodes over periods of 1 to 18 months (Gergen and MacLean, 1961). Monopolar and concentric bipolar electrodes with 1 mm exposed tips have been placed in 2 or more sites of the ventral hippocampus as well as in other selected subcortical areas for purposes of stimulation and recording. In these studies, no prominent &rhythms have been recorded in the hippocampus associated with alerting or orienting behavior (Fig. 1). Instead, desynchronization patterns paralleling those of the neocortex are seen. Stimulation at various hypothalamic and upper brain stem sites also does not evoke 8-activities concurrently with cortical desynchronization. Irregular 8-activities are evident during drowsiness, however, in association with cortical synchronization.
* In these studies, the term hippocampal formation is used to refer collectively to the dentate gyrus, hippocampus and hippocampal gyrus. The names and abbreviations for subdivisions and special cortical areas of these structura are taken from Lorente de N6 (1933, 1934).
443
F U N C T I O N A L P R O PE R T IES O F H I P P O C A M P U S
I
L H l P W AP 4 I.niwo A P ~
CLAP
I. H I P W AP4 L fflPP0 A P 3
R fNTO AP4 R MTO A P 1
I
157.6
MOWLY -ACHED
1 sec
Fig. 1. Representativeelectroencephalographic activity recorded from the neocortex and hippocampus of the squirrel monkey, demonstrating concurrent presence of low voltage fast activity in the neocortex and ventral hippocampus following hand clap and approach of investigator. Findings here are characteristic of those for multiple electrode placements in various areas throughout ventral hippocampus. Epidural monitor recorded between epidurd electrodes over right and left parietal cortex. L Hippo AP 4 and AP 3, electrodes in left hippocampus; AP 4, located in stratum oriens of are1 CA 3 while AP 3 is approximately 1 mm posterior on mesial surface of the dentate gyrus. R Ento AP 4 and AP 3, electrodes located on mesial surface of right entorhinal cortex with interelectrode distance of approximately 2 mm. AP coordinates of this and subsequent figures based on stereotaxic atlas of Gergen and MacLean (1962). Vertical calibration, 200 ,uV; time marker, 1 sec.
Hippocampal stimulation in these monkeys has typically not elicited reproducible gross behavioral or motor responses. More common findings with stimulation have been cortical desynchronization and intermittent head turning to the contralateral side. A total of 1050 hippocampal after-discharges have been induced in these animals by short trains (2-5 sec) of square wave pulses of 0.1-0.5 msec duration at frequencies of 30-100 per sec. Peak currents have been monitored during stimulating and stimuli are set at 0.1-0.2 mA above threshold for eliciting after-discharges with typical pulse currents of 0.6-0.7 mA. During induced after-discharges, the most obvious behavioral change in the animal is a tendency for intermittent head turning to the contralateral side associated with the appearance of ‘visual searching’ which does not appear to fixate on any real object. Occasional swallowing may occur late in the seizure or in the few seconds following its termination. Heart rate changes are not apparent during the average seizure but bradycardia may appear when seizure durations become prolonged. No generalized motor seizure has ever appeared in any of our animals. Hippocampal seizures themselves remain lateralized with only propagated activity being evident in the contralateral hippocampus. Seizures in our animals do not disrupt on-going behavior such as eating, but there appears to be blunting of ‘emotional responses’ to threatening environmental stimuli. Relatively reproducible seizure durations of 18-30 sec have been obtained in 5 animals by allowing a 20-min recovery period between stimulations. More frequent stimulation in these animals increases the variability and References p. 460461
444
J. A. GERGEN
duration of after-discharge with the appearance of both abortive seizures and, more frequently, prolongation of after-discharge. Three animals have demonstrated greater irregularity in seizures with frequent lengthening of the seizure duration following an interseizure interval of 20 min. These latter animals have all demonstrated on histological examination hemorrhages or excessivegliosis surrounding at least one electrode track in- ttte hippocampus. All animals have demonstrated spontaneous electroencephalographic spiking in the hippocampal formation after repeated seizures. This spiking is augmented by drowsiness and suppressed during periods of hippocampal desynchronization induced by orienting responses or brain stem stimulation. During the initial post-seizure recovery period following the average seizure, low frequency stimulation demonstrates the lability of the hippocampus to develop further seizure activity by recruiting bursts of ‘clonic’ discharge over an additional period of 30-80
Fig. 2. Prolongation of clonic hippocampal after-discharge by low frequency stimulation. Recording is between Hippo AP 3 electrode located in mesial dentate gyrus and scalp reference electrode. Seizure induced by stimulation at 100 per sec for 2.5 sec; vertical arrows designate time for approximately 1 per sec stimuli delivered during and after the period of after-discharge. Note late onset of low voltage fast activity during recovery period from seizure along with slow recovery of evoked response. Lowest trace demonstrates evoked hippocampal response during a control period to approximately 1 per sec stimuli. Vertical calibration, ImV; time marker, 1 sec.
sec. Optimal frequency for such stimulation is 0.5-2 per sec. As seen in Fig. 2, there is a distinct tendency for such low frequency stimulation to both initiate and terminate ‘clonic’ after-bursts. Introduction of repetitive stimuli at intervals longer than 2 sec following termination of seizure activity is usually ineffective in triggering such ‘clonic’ discharges. Some attention has been paid to studyingother manifestationsof the recovery cycle following hippocampal after-discharges. Spontaneous electroencephalographic activity in the hippocampus begins to be evident at intervals of 1-30 sec
F U N C T I O N A L PROPERTIES OF H I P P O C A M P U S
445
following termination of the seizure with return to essentially normal activities requiring several minutes or more. Recovery of evoked responses in the hippocampus to stimulation of the fornix or hypothalamus tends to return concurrently with recovery of spontaneous activity. Subcortical evoked responses in the lateral geniculate nucleus to flash and in the inferior colliculus to click are not significantly modified during or after hippocampal after-discharges. Responses to click, flash, and trigeminal nerve stimulation recorded in the upper brain stem in one animal, however, have been attenuated by after-discharges but recover within 1-2 sec following termination of the seizure. Suggestive evidence of more prolonged recovery cycles has been found when measuring seizure duration as a function of inter-seizure interval. In addition to observations that a 20-min inter-seizure interval appeared optimal for regularity of seizure duration, it has been found that for inter-seizure intervals of 2 h there is a distinct tendency for spontaneous lengthening of after-discharge when compared to the average seizure. In one animal a persistence of this tendency has been evident at 15 h following a series of seizures induced at 20-min intervals. Stimulation of the hippocampus in other mammals has led to a variety of responses including body and facial movements and alterations in visceral, autonomic and endocrine functions (see Green, 1964). It is difficult, however, to find descriptions of consistent responses that might be elicited for more than one site within the hippocampal formation in which the possibility of induction of hippocampal after-discharges has been adequately excluded. In the squirrel monkey, recruitment of slow evoked potentials and hippocampal after-discharges frequently appears during cortical and subcortical stimulation which elicits penile erection ; stimulation of selected portions of the posterior ventral hippocampus may also induce penile erection (MacLean et al., 1963; MacLean and Ploog, 1962). Hippocampal &rhythms have been extensively studied during ‘arousal’ and paradoxical sleep in the rabbit, cat, and other mammals (see Green, 1964). Such &rhythms have been more difficult to isolate in the rhesus monkey (Green and Arduini, 1954), chimpanzee (Rhodes et al., 1963) and man (Kaitor et al., 1957). Two reasonable possibilities which might explain possible species differences are that (1) rhythmic &activities have become restricted to a smaller group of behavioral responses and might thus be less frequently encountered in a controlled laboratory environment and (2) anatomical and physiological refinements have occurred within the hippocampal formation and septa1 area which suppress the amplitude or alter the frequency of these rhythms. Some modification of the first hypothesis has generally been favored by other investigators concerned with this discrepancy (Grastyhn et al., 1959). Data to be presented later in the present discussion, however, suggest that the second hypothesis might also receive further investigation. Hippocampal after-discharges in the squirrel monkey appear to parallel in most respects findings on other mammals. Use of low stimulus intensities and moderate inter-seizure intervals in the squirrel monkey apparently suppresses the extension of hippocampal seizures into other cortical areas and may limit progressive evidences of behavioral alteration and prolongation of seizure duration with repeated seizures reported in the cat (Delgado and Sevillano, 1961). Data in the present study suggest References p.y460-461
446
J. A. G E R G E N
that hippocampal seizuresmay remain 'sub-clinical' in the primate even after induction of repeated seizures. The presence of concurrent hippocampal injury, however, appears to prolong seizure duration, which, in turn, might well potentiate spread to other brain areas. Our data also suggest that prolonged recovery times may exist within the hippocampus following a seizure; this disturbance might, in turn, alter participation of the hippocampal formation in modulating ongoing complex behaviors well beyond the period of gross seizure activity. H I P P O C A M P A L EVOKED S L O W P O T E N T I A L RESPONSES A N D S I N G L E U N I T ACTIVITIES
Evoked slow potential responses and single unit activities have been recorded in the ventral hippocampus of 39 squirrel monkeys following electrical stimulation of the septum, olfactory tract, posterior cingulate gyrus, and other subcortical areas, and following photic stimulation (Gergen and MacLean, 1964). All animals in these studies have been anesthetized with intraperitoneal a-chloralose in doses up to 100 mg/kg. In 16 animals mapping of the distribution of slow evoked responses has been carried out with fine electrode; having an exposed recording surface at their tip of 0.006 in.2. Seven animals have been studied utilizing fixed recording electrodes in the hippocampal formation with roving concentric bipolar electrodesbeing used for stimulation. Recordings of single and group unit activities in the hippocampus and adjacent entorhinal cortex have been obtained in 16 animals utilizing etched steel microelectrodes. Electrical stimuli in all cases are square wave pulses delivered through a stimulus isolation unit. Pulse current and stimulus duration have been monitored with stimuli not exceeding 1 mA for 1 msec. A photic stimulator has been used to deliver brief (0.010 msec) flashes of light. Prominent slow evoked responses can be obtained in the hippocampus following electrical stimuli delivered to the olfactory tract, lateral olfactory stria, diagonal band of Broca, septum, anterior and posterior cingulate gyrus, anterior thalamus, hypothalamus and fornix. No consistent responses can be obtained on the other hand following stimulation of the caudate nucleus, internal capsule, globus pallidus, putamen or claustrum. Unstable long latency responses occur following stimulation of the supra-callosal cingulate gyrus, anterior portions of the gyrus rectus, and posterior thalamus including the centromedian nucleus. Responses to photic stimulation have been found intermittently in 19 of 36 animals using as a criterion the presence of a consistent form and latency (& 20 msec) following more than 50% of photic stimuli delivered successively at rates not exceeding 1 per sec. In 7 other animals, suggestive responses to photic stimulation in terms of form and latency but not regularity have been observed. Slow evoked responses in the ipsilateral hippocampus appearing at a latency of 31-35 msec have been consistently elicited by low frequency (up to 3 per sec) and low intensity stimulation of the olfactory tract and lateral olfactory stria. Evoked activities in the ipsilateral entorhinal cortex include a short latency potential (3 msec), best seen in the anterior part of the ventral hippocampal formation, and a longer latency
F U N C T I O N A L PROPERTIES O F H I P P O C A M P U S
447
response (21-31 msec) which appears to correlate with the evoked response of the hippocampus. Waxing and waning of response amplitude may be present and appear to be largely attributable to interaction with spontaneous chloralose potentials. Evoked activities to olfactory tract stimulation typically demonstrate a positivity at the ventricular surface of the hippocampus with polarity reversal appearing in the stratum radiatum and stratum lacunosum-moleculare of areas CA 2-3. Projection of olfactory responses into the dentate gyrus are variable with polarity reversal of evoked activity becoming evident in the stratum moleculare in some instances; on other occasions, polarity reversal does not occur until after the electrode has penetrated through the stratum moleculare of the entorhinal cortex. Photic responses when present have latencies of 45-1 60 msec within the hippocampus and subiculum. Shorter latency responses (18 msec) of lower amplitude have been encountered in the posterior entorhinal cortex near the cingulum bundle. Intraperitoneal injection of pentylenetetrazol (10-30 mg/kg) has been noted to lead to an increase in amplitude and stability of photic responses at latencies of 45-60 msec within the hippocampus. The distribution and polarity of the typical photic evoked response is similar to that seen following stimulation of the olfactory tract. Photic responses are not blocked by coagulations placed high in the pontine reticular formation, anterior hypothalamus or septum. The form and latency of responses to septa1 and fornix stimulation have been found to be dependent on the locus of the stimulating electrode and the intensity of stimulus
-
S
F
. . . . . ...
... .
?r'
v--
'-"'A
/1;G /GL IL
'k
c1
J L
b/
k
n,
L
Fig. 3. Evoked responses to stimulation of the septum (S) and anterior fornix (F)recorded during exploration of the ventral hippocampal formation with a microelectrode. Electrode track passes sequentially through the alveus, stratum oriens, stratum pyramidale, stratum radiatum, and stratum lacunosum-moleculare and then into the dentate gyrus where it encounters the stratum moleculare and stratum granulosum of dentate granule cells before and after entering the stratum pyramidale of CA 4. In this and subsequent figures, the convention of 'positive-up' has been adopted. Time marks at top of each column are 10 msec; vertical calibration 2.5 mV. References p. 460-461
448
J. A. GERGEN
current among other factors. Responses to stimulation of both the septum and fornix show low amplitude activities at latencies of 3.5-5.0 msec in the fimbria and alveus. This would be consistent with activities conducted over the fornix at velocities of 4-8 m/sec. Lateral septal stimulation typically elicits a more prominent evoked negativity in the stratum oriens of hippocampal areas CA 2-3 at latencies of 11-14 msec (Fig, 3). Similar evoked potentials appear at latencies of 6-8 msec following stimulation of the medial septum and anterior fornix. This negativity typically appears to undergo a polarity reversal as the electrode traverses the stratum pyramidale and enters the stratum radiatum An initial sharp negativity may persist into upper portions of the stratum radiatum, particularly following fornix or higher intensity septal stimulation. Longer latency responses (20-50 msec) demonstrating an evoked positivity in the stratum oriens and stratum pryamidale of areas CA 2-3 are also seen following septal and fornix stimulation. These latter responses characteristically show polarity reversal at a deeper point in the stratum radiatum than that seen for the shorter latency negativity already described. Evoked responses recorded in the dentate gyrus to stimulation of the septum and fornix are more complex in character and have less clearly defined evidences of polarity reversals. No short latency potential demonstrating polarity reversal has been evident near the dentate granule cells following septal Stimulation. Fornix stimulation as seen in Fig. 3 does, however, demonstrate short latency responses having complex configurations of changing amplitude and distribution as the electrode passes through the layer of dentate granule cells into the stratum moleculare. Intermediate and longer latency responses showing distinct polarity reversals in these areas are encountered following both septal and fornix stimulation. .. . . . . . . .
.................
Fig. 4. Unit (column 1) and evoked slow responses (column 2) following septal stimulation record microelectrode during exploration of ventral hippocampal formation. Low frequency activities are filtered in column 1 to permit better visualization of unit discharges. Vertical calibration 500 pV; time markers are 10 msec. Ca = caudate nucleus; Cin = cingulum; GL = lateral geniculatenucleus.
F U N C T I O N A L PROPERTIES O F HIPPOCAMPUS
449
A short latency evoked positivity (5-7 msec) is evident in the stratum pyramidale and stratum oriens of area CA la, prosubiculum, subiculum, and adjacent entorhinal cortex following stimulation of both the lateral septum and anterior fornix (e.g. Fig. 4). Changes in amplitude and form of this response are present as the electrode traverses the stratum pyramidale into the stratum oriens of these latter areas, but there is again no evidence of a distinct polarity reversal. Short latency evoked responses to septal and fornix stimulation are stable in terms of amplitude and configuration on repetitive stimulation at frequencies up to 10 per sec and appear to show little interaction with spontaneous chloralose potentials. Longer latency responses on the other hand are more labile, with failure to follow stimulus frequencies greater than 3-4 per sec and with significantly greater interaction with spontaneous chloralose potentials. In microelectrode studies, 180 single and 31 group unit discharges have been recorded in the hippocampal formation. Unit discharges include negative, positivenegative and positive potentials. Spontaneous activity has been evident in 44 of these units, all of which have had positive or positive-negative spike patterns. These latter units typically demonstrate rapid firing when first encountered but later stabilize to S
F
21-16
Fig. 5. Inhibition of spontaneous unit activities following septal (S) and fornix (F) stimulation. Upper row (21-16) is recorded from unit in area CA 3; spontaneous burst-inactivation discharge following recovery from septal stimulation has been encountered only in cells considered seriously injured by the microelectrode. Lower row (2144) is recorded from unit adjacent to or within dentate granule cell layer; this unit demonstrates multiple firing following septal and fornix stimulation. Note recovery of spontaneous firing in both units associated with disappearance of prolonged negative evoked activities, Vertical calibration 1 mV; time marks are 100 msec.
a lower but persistent rate of spontaneous discharge. Burst inactivation patterns of firing (see Fig. 5) are apparent only in more seriously injured cells and do not appear to be necessarily correlated with evoked responses or spontaneous chloralose ‘spike’ potentials. Single unit discharges are related to spontaneous and evoked slow potentials in a complex fashion. Little apparent correlation is present between unit firing References p. 460461
450
J. A. GERGEN
and the typical spontaneous high amplitude chloralose potential which demonstrates a negativity in the stratum lacunosum-moleculare and polarity reversal in the stratum radiatum. Less frequent, more complex spontaneous potentials may be followed by a pause in firing of spontaneously active units. Of 143 units in the hippocampal formation activated by septal and fornix stimulation, all discharge in the interval 3.5-70.0msec following the stimulus. Only 15of these units have been recruited within the period of 3.5-6.0 msec, a latency considered consistent with possible antidromic activation. It thus appears that the majority of units being recorded are driven through one or more synaptic linkages. Recruitment of single unit following fornix or septal stimulation appears to most closely parallel the early negativity evoked in the stratum oriens and stratum pyramidale of areas CA 2-3 (Fig. 5). Unit firing is only weakly correlated with the longer latency negativity evoked in the stratum radiatum and stratum lacunosum-moleculare. A similar weak positive correlation with variable latency is seen for unit firing following electrical stimulation of the olfactory tract and posterior cingulate gyrus and photic stimulation. Recruitment of units in areas CA 1-2, prosubiculum and subiculum by septal and fornix stimulation correlates with the evoked slow potential positivity recorded in these areas at short latencies. The typical unit in hippocampal pyramidal cell layers discharges only once following stimulation of the fornix or septum. Units recruited at latencies up to 30 msec are frequently able to follow stimulus frequencies up to 10 per sec, while recruitment of additional units at these frequencies is also unusual. Repetitively firing neurons to a single afferent volley from the fornix or septum have been encountered in the present study in the dentate gyrus adjacent to or overlapping the layer of dentate granule cells and in the subiculum and adjacent entorhinal cortex. These latter units are more responsive to varying the frequency of stimulation and may be recruited at stimulus frequencies HIPPOCAMPAL PYRAMIDAL C E L L (CA 3) Auditory System
\I f
Olfactory System
J
Entorhinol cortex Visual System
f
/
/
k-
I
+
I
/
Recurrent --( collateral system
+
Entorhinol cortex
Perforant Path
Septum and Hypotholamus
-Dentate
>--- -. //---
granule cells Commissural afferents
~.
Tholamus
Fig. 6. Provisional model for projection pathways to the ventral hippocampus of squirrel monkey
(see text). Excitatory synaptic activities denoted by plus and inhibitory by minus signs. Entorhinal
projections to dentate granule cells appear to be excitatory, but eventual relays to pyramidal cell are less certain.
F U N C T I O N A L PROPERTIES O F HIPPOCAMPUS
45 1
of 5-10 per sec. Of the spontaneously active neurons, 43 units recorded in all areas of the hippocampus and subiculum demonstrate a suppression of spontaneous firing for periods of 200-1 300 msec following stimulation of the septum or fornix (see Fig. 5). Recovery of spontaneous firing in these units appears to closely correlate with a prolonged slow negativity which is typically of maximal amplitude in the stratum pyramidale and which can be recorded at lower amplitudes in the stratum oriens and stratum radiatum adjacent to the stratum pyramidale. Repetitive stimulation at frequencies up to 20 per sec may lead to a slight prolongation of the period of suppression following the last stimulus. Dying units may also be transiently revived during the later portions of the period of typical unit suppression. Fig. 6 depicts a provisional model for projection pathways to area CA 3 of the ventral hippocampus of the squirrel monkey. At least two afferent projections to the hippocampus from the septal area are suggested. One of these, the ‘direct’ pathway, appears to project to basal dendrites of areas CA 2-3. Excitation of this pathway in the present experiments Is apparently strongly correlated with unit firing. Latency differences between lateral septal and fornix stimulation suggest that one or more rzlays are present within the septum itself. The tentative outflow path following septal stimulation appears to pass through the medial septal nucleus and fornix to the hippocampus. The conduction velocity in this afferent system appears to be slightly less than that for antidromically activated fibers from the pyramidal cells. The possibility of additional synaptic relay in the hippocampal formation (prosubiculum, CA 1a) with a subsequent projection over the alveus to the stratum oriens cannot be excluded, however. Three animals subjected to septal lesions and subsequently stained with a Nauta technique (1954) have demonstrated the presence of degenerating fibers skimming into the stratum oriens from the fimbria and alveus as well as projecting into the prosubiculum and entorhinal cortex. These preparations have not demonstrated projections to dentate granule cells although degenerating fibers can be found in area CA 4. A second or ‘indirect’ projection system appears to relay in the entorhinal cortex with projection over the perforant path to distal portions of apical dendrites. A similar projection path is apparently utilized by olfactory and visual inputs. In each instance this latter pathway is characterized by a weak excitatory linkage in terms of single unit recruitment and by a greater tendency for interaction with other afferent sources or with previous stimulation. Other projection systems having intermediate and longer latencies following septal or fornix stimulation are less clearly resolved in the present experiments. Schaffer collaterals to apical dendrites in areas CA 3-CA l c may well be excitatory to these areas and account for some of the late negative evoked responses recorded in apical dendrites. It is also possible that relays from the contralateral hippocampus contribute to the evoked responses recorded at intermediate and longer latencies, particularly in posterior portions of the hippocampal formation. Concurrent recording from the contralateral hippocampus following septal and fornix stimulation has demonstrated low amplitude evoked potentials at latencies comparable to those recorded ipsilaterally. Relays through dentate granule cells in the present experiments appear to be labile but may well provide additional excitatory projections to pyramidal cells through the supraReferences p. 460-461
452
J. A. G E R G E N
and infra-pyramidal mossy fiber systems. Our data provide good evidence for the presence of long duration inhibition following activation of pyramidal cells in squirrel monkeys. These findings appear consistent with those for the inhibitory recurrent collateral system described in the cat by Spencer and Kandel (1961). The specific neural linkages of recurrent collateral inhibition remain somewhat controversial at the present time (Gloor, 1963). Andersen et al. (1964) have suggested that short latency inhibition in the cat hippocampus is mediated through basket cells with inhibitory relays in the stratum pyramidale. This short latency inhibition and its concurrent positive evoked response in the stratum pyramidale may well be obscured by the short and intermediate latency excitatory responses recorded here in the chloralosed monkey. The later period of inhibition distinguished in the present experiments is associated with a negative evoked response in the stratum pyramidale, suggesting that this inhibition is, in part, mediated through other circuits or has an anomalous pattern of current flow (Gloor et al., 1963). The present data do indicatethat more than one ‘null’ or polarity reversed point should be considered in deciphering correlations of evoked slow potential responses or electroencephalographic wave forms with single unit activities in the hippocampus (Green et al., 1960). S U B CO RT ICAL EFFECTS OF H I PPOC A MPA L ST IMU L A T ION
Effects of hippocampal stimulation on single unit activities of 377 neuronsin theseptum, hypothalamus, thalamus, and basal ganglia have been studied in the ‘cerveau isolt’ preparation. Prior to these studies, squirrel monkeys have been decerebrated by coagulations placed along an approximate stereotaxic plane which passes through the commissure of the inferior colliculus to just posterior to the interpeduncular nucleus. Electrical stimuli have been delivered through stereotaxically implanted bipolar electrodes placed in the posterior ventral hippocampal formation. Stimulus current amplitude and pulse shape have been monitored with typical stimulus parameters, being a square wave pulse of 1mA for 0.5 sec. An additional monitorial electrode for hippocampal slow potential responses and seizure activities has been stereotaxically placed in the anterior ventral hippocampus. On completion of experimental procedures, the animal is perfused with a formalin-agar solution. Frozen brain sectionshave been cut at 50 mp and electrode tracks are plotted from stained sections. Subcorticalunit activities have been followed during stimulation of the hippocampus at frequencies between 0.5-10.0 per sec. Effects of trains of stimuli on single unit activity have been determined by plotting a post-stimulus time histogram (Gerstein and Gang, 1960). Initial post-stimulus time histograms have been plotted from unit activities recorded on film from an oscilloscope display. More recently, a modified Mnemotron 400B computer has been used to facilitate calculations. Following determination of effects of low frequency stimulation, a hippocampal after-discharge has been induced by stimulation at frequencies of 10-60 per sec. Unit activities here are followed by summing total discharges over 1-sec time intervals before, during and after the induced hippocampal seizure. For purposes of the present analysis, no unit has been considered unaffected by low frequency stimulation until tested for at least
F U N C T I O N A L PROPERTIES OF H I P P O C A M P U S
453
two different frequencies. Typically, all units have been tested for at least three frequencies with 100 to 200 stimuli at each frequency. Units have been considered affected by stimulation when the post-stimulus time histogram demonstrates a peak or valley in the first 500 msec following each stimulation which exceeds for two or more successive 5-msec measurement intervals either the range of values obtained in an equivalent post-stimulus time histogram with subthreshold stimulation, or the range of values present for the time interval 500-1000 msec following stimulation. Of the 377 subcortical units characterized at the present time, 169 or 45% appear affected by the low frequency stimulation. In terms of anatomical distribution, brain areas in which 10 or more units have been adequately characterized, and more than 50 % of these units appear affected, include the anterior thalamus, mammillary body, paracentral nucleus and septa1 area. Brain areas in which 25-50% of units are considered affected include the medial dorsal nucleus, particularly the medial and very lateral portions, the dorsal hypothalamus and anterior ventral nucleus. Brain areas in which only 5-10 units have been studied and which show more than 50 % of these units being affected include the central lateral nucleus and adjacent nucleus lateralis posterior, the bed nucleus of the stria terminalis, and the preoptic and posterior hypothalamus. Brain areas with 5 or more units studied and showing less than 25 % of affected units include the medial and lateral ventral posterior nuclei, the H -13
MIDLINE ANTERIOR HYPOTHALAMUS I/<*.-
t
s
100
200
5/sec
100 msec
300 msec
500
400
A(500-1000)
IO/sec
200
1
I00
rec
Fig. 7. Post-stimulustime histogram of single unit encounteredin the midline anterior hypothalamus. Stimulus of posterior hippocampus at frequencies of 1, 5, and 10 per sec. Vertical bar at each time interval represents summed discharges over an interval of 10 mSec for 50 stimuli. S in this and subsequent figures designated time of stimulus. Last bar in upper histogram, A (50&1000), is average firing rate of neuron in time interval 500-1000 msec (9 per sec) and approximate average firing of unit when unaffected by stimulation. References p . 460461
454
J. A.' GERGEN
Fig. 8. Post-stimulus time histogram of single unit encountered in the posterior hypothalamus approximately 1 mm dorsal to mammillary bodies. Stimulation of the posterior hippocampus at frequencies of 1,2,3,5, and 10 per sec. Each vertical bar represents summed discharges over 80 msec for 90 stimuli. Average firing rate of neuron when unaffected by stimulation is 1.4 per sec.
Fig. 9. Post-stimulus time histogram of single unit encountered in the anterior ventral thalamic nucleus. Stimulation of posterior hippocampus at 1, 2.5, and 5 per sec. Each vertical bar represents summed discharges over 25 msec for 50 stimuli. On right, results of 2.5 per sec (dark bars) and 5 per sec (stippled bars) superimposed for sake of better comparison. Average firing rate of neuron when unaffected by stimulation is 16.7 per sec.
F U N C T I O N A L PROPERTIES OF H I P P O C A M P U S
IL
455
K-28
4 R E A PREOPTICA
(midline)
2/5ec
t
S
4i)O
SO0
? S
mrec
400
IO/rec
5/5ac
l I
tS
160
t
s
A
(3.2/rrc)
Fig. 10. Post-stimulus time histogram of single unit encountered in the midline preoptic hypothalamus. Stimulation of posterior hippocampus at frequencies of 1, 2, 5, and 10 per sec. Each vertical bar represents summed discharges over 80 msec; results at faster frequencies of stimulation corrected to represent 60 stimuli for comparison with 1 per sec stimulation. Average firing rate of neuron when unaffected by stimulation is 3.2 per sec.
RESPONSE TO LOW FREOUENCY STIMULATION POST-STIMULUS TIME NISTOGRAWS
Sosic Patiem
_____ _____
Fig. 11. Diagrammatic representation of more characteristic post-stimulus time histograms following 1 per sec stimulation. Dotted line in each pattern represents average firing rate of neuron when unaffected by stimulation. Peak on solid curve represents period of increased activity while valley indicates period of relative suppression of spontaneous firing. See text for additional details. References p. 460461
456
J. A. GERGEN
reticular thalamic nuclei, the centromedian nucleus, and the paraventricular nucleus. Characteristic post-stimulus time histograms for several units are demonstrated in Figs. 7-10. Several of the more characteristicresponse patterns for the post-stimulus time histogram of affectedunits to low frequency stimulation are diagrammatically illustrated in Fig. 11. Units encountered in the septum, mammillary body, anterior thalamic nucleus, and lateral paracentral nucleus, among others, characteristically demonstrate patterns A or B. In pattern A, there is an initial period of increased activity at latencies of 10-30 msec followed by a longer period of suppression which may last from 50-300 msec. This period of suppressionmay be followed by a rebound increase in firing which tends then to recede into background f b g rates at latencies of approximately 500 msec. Pattern B fails to demonstrate short latency recruitment but does demonstrate an initial period of moderate or pronounced suppression now typically followed by a more prominent post-suppression rebound. Units with this initial pattern to low frequency stimulation frequently demonstrate some evidence of short latency recruitment when stimulus frequencies are increased to 5-10 per sec. Patterns C and D are less commonly encountered. Pattern C demonstrates recruitment at latencies of 15-50 msec without subsequent suppression. This type of response has been found in the preoptic and lateral hypothalamus and in the ventral anterior nucleus. Pattern D demonstrates a prominent late excitement phase which roughly corresponds to the rebound periods in patterns A and B. This type of pattern has most commonly been seen in the medial dorsal nucleus but may be occasionally encountered elsewhere. Of 235 neurons examined for effects of hippocampal after-discharge, 169 or 72% appear affected by hippocampal seizures. Brain areas with 5 or more units examined which demonstrate less than 50 % of these units affected by after-discharges include the dorsal hypothalamus, the ventral posterior medial and ventral posterior lateral nucleus, the centromedian nucleus, and the parafascicular nucleus. More characteristic patterns of response for units affected by low frequency stimulation include increased firing during the after-discharge with rebound depression (units with pattern A) and decreased firing during the after-discharge with rebound excitation (units with pattern B). Prolonged periods of excitation or inhibition during and following the after-discharge are found in units not directly affected by low frequency stimulation. Occasional units may show periods of fluctuating excitation and suppression during the after-discharge with variable rebound at the termination of the seizure, while other units have shown only ‘regularization’ of firing. Effects of after-discharge on unit firing can be detected for periods up to 40 sec following termination of the seizure. Results obtained in these current experiments begin to demonstrate the extent of hippocampal control of subcortical neurons and also indicate the likelihood for both excitatory and suppressive influences. It is evident in our data that few neurons, with the possible exception of those in the septa1 area and mammillary body, are directly driven by hippocampal stimulation such that unit discharges occur at a relatively fixed latency following each stimulus. Instead, a more characteristic response pattern is to see only intermittent recruitment at low frequency stimulation (0.5-1 per sec) with strengthening of this early recruitment by stimulation at faster frequencies
F U N C T I O N A L PROPERTIES OF H I P P O C A M P U S
457
(2-10 per sec). This recruitment appears to be analogous to the potentiation described by Andersen (1967) following stimulation of various afferent systems to hippocampal pyramidal cells at frequencies up to 10 per sec. The presence of numerous units showing early suppression (pattern B) in close proximity to units showing the early excitatoiy period followed by suppression (pattern A) suggests that the suppressive period may well be mediated through inhibitory collateral circuits. The extent and duration of this suppressive period in units showing either both pattern A or pattern B is typically sufficientto depress unit activity below spontaneous firing rates. Delayed post-suppressionrebound excitation encountered in patterns A, B, and D does not appear to be necessarily proportional to the extent of suppression experienced by the neuron under study. This suggests that increased activity during the post-suppression period is some combination of rebound hyperexcitability within the neuron itself and increased interneuronal excitation from other rebounding units. Previous investigators have generally stressed either the excitatory or inhibitory aspects of hippocampal stimulation on subcortical evoked slow potential responses and single unit activity. Manzoni and Parmeggiani (1964) and Parmeggiani (1967) have described excitatory effects of dorsal hippocampal stimulation on single unit activity in the anterior ventral, anterior medial, dorsal medial, lateral dorsal, ventral anterior, central lateral, paracentral, and habenular nuclei of the curarized cat. In these studies low frequency stimulation (2-5 per sec) elicited waxing and waning responses of single units including eliciting burst-fire patterns and recruitment of previously silent neurons. Stimulation of the sciatic nerve in these studies appears to depress or abolish unit responses to contemporary hippocampal stimulation in the anterior ventral, central lateral and occasionally in the medial dorsal nuclei. Adey (1958) has previously described prolonged subcortical inhibition of spontaneous unit activity in the central gray substance of the marsupial phalanger (Trichosuuris vulpeculu). Single stimuli to the entorhinal cortex in these animals elicits an inhibition of spontaneous firing which lasts for 300400 msec or longer. Hippocampal stimulation elicits inhibition at a somewhat longer latency, a finding consistent with the suggestion that direct entorhinal connections to the central gray substance mediated the inhibition. Adey et ul. (1957) have also demonstrated that stimulation of the entorhinal cortex in the rhesus monkey suppresses evoked slow potential responses in the mesencephalic reticular formation to stimulation of the caudal reticular formation for periods up to 2000 msec. More recently Feldman (1962) has described inhibition of evoked slow potential activity recorded in the hypothalamus to sciatic nerve stimulation by concurrent stimulation of the hippocampus in the cat. L E A R N I N G SET P E R F O R M A N C E A N D D I S C R I M I N A T I O N REVERSAL
Effects of hippocampal lesions on learning set performance and discrimination reversal are currently being studied in 12 squirrel monkeys. These animals are trained on visual discrimination problems in a modified Wisconsin General Test Apparatus. Criterional training procedures are used in which each animal is brought to an intraproblem criterion of 80 % correct responses over 2 successive sets of 10 trials References p. 460-461
458
J. A. GERGEN
before moving to the next problem (Rumbaugh and McQueeney, 1963). Non-correction procedures are used. Test objects are randomly alternated in left-right positions according to a Gellerman series. For each pair of visual discrimination patterns, alternate subjects are rewarded for opposite patterns to control for pattern preference. Rewards consist of small pieces of grapes. Animals are food deprived for 20 h prior to being introduced to the apparatus. All animals are trained on 20 learning set problems prior to operation. On completion of this initial block of learning set problems, animals are subsequently variously subjected to bilateral hippocampal lesions, lesions of the inferior temporal lobes, unilateral hippocampal lesions and sham operative procedures in which electrodes are introduced but no attempt is made to produce large areas of tissue destruction. Lesions have been produced by electrolytic currents from electrodes stereotaxically implanted at 4-10 sites within the temporal lobes. A 10-day recovery period has been allowed for each animal followingthese operations. Animals are then tested on an additional 20 learning set problems. One week subsequent to completion of all learning set problems, animals have been introduced to discrimination reversal training. In reversal training, 6 animals have been trained to meet a criterion of 80 % correct performance over 5 successive trials before and after reversal has occurred. An additional 5 animals have been trained to a criterion of 80 % over 10 successive trials preceding and subsequent to reversal before moving to the next problem. Histological examination of brain lesions has now been completed on 7 of 12 animals. Two animals of this group have demonstrated evidence of bilateral damage to the hippocampus; one animal has demonstrated extensive unilateral damage to the hippocampus; two additional animals have evidence of bilateral damage to the entorhinal cortex but not the hippocampus; one animal has had damage predominantly confined to the visual system and one animal has had only a sham operative procedure but some inadvertent bilateral damage to visual pathways. All animals, with the exception of the one demonstrating extensive visual system damage, demonstrate postoperative retention of learning set performance. Two animals with bilateral hippocampal damage, however, demonstrate a suggestive deficit in discrimination reversal which has not been evident in animals with lesions placed elsewhere. Fig. 12 illustrates findings in two animals, one with a sham operative procedure and the other with moderate bilateral damage to the hippocampus. These data presently suggest that a major deficit in our animals is an inability to easily give up previously learned or preferred patterns of response rather than any disturbance in recall or retention of recent memory. Findings are thus in agreement with other investigators on the relative prominence of perseverative behavior and resistance to extinction in the hippocampectomized animal (Isaacson et al., 1961 ; Jarrard et al., 1964; Jarrard, 1965; Kimble, 1963; Kimura, 1958; Niki, 1962, 1965). SUMMARY
Stimulation and ablation experiments in the squirrel monkey and other animals appear to implicate the hippocampus as participating in transactional or associative activities of the central nervous system away from processes related to recognition of
459
F U N C T I O N A L PROPERTIES OF H I P P O C A M P U S
Pee Wee (Sham Operate Control)
Otis IHipposompol Lesion) Learning Set
Learning Set
;
5 'L
;
{
*
-
s
A
50
0
c ." !?
0
Y
e 50
1
2
3
4
5
6
Blocks of 5 problems
7
8
..c t o L
0 c
Discrimination Reversal
f 1
2
3
Blocks of 5 problems
4
:H' , - ; - : ; I 1
2 3 4 5 6 7 Blocks of 5 problems
%
Discrimination Reversal C-d Al
*-*
tn h r u l
.-------
1 2 3 Blocks of 5 problems
4
Fig. 12. Comparison of effects of sham operation and placement of bilateral hippocampal lesions on learning set performance and discrimination reversal in the squirrel monkey. Upper and lower graphs on left are from sham operant control; this animal has moderate bilateral damage in visual pathways but hippocampal formation appears intact on histological examination. Post-operative fluctuations in learning set performance appear to be characteristic for animals sustaining visual system lesions. Upper and lower graphs on right present learning set and discrimination reversal performance of a squirrel monkey with moderate bilateral damage (50 %) in the ventral hippocampus. Lesions produce no significant effect on learning set performance but do disturb significantly discrimination reversal. All points on graphs represent results averaged over block of 5 problems each. See text for additional details.
primary sensory information or execution of somatic motor response. Our data suggest that hippocampal neurons receive sensory information from exteroceptive receptors which is relayed over the cingulum bundle to the entorhinal cortex and thence over the perforant path to apical dendrites of the stratum radiatum and stratum lacunosum-moleculare. Hippocampal neurons, however, appear to be more directly affected in terms of neural firing patterns by afferent systems originating in subcortical centers. It has previously been suggested (Gergen and MacLean, 1964) that if this model of hippocampal input circuitry were viewed analogously in terms of classical behavioral conditioning, interoceptive impulses conducted by the septal-fornix system would be comparable to unconditional stimuli since they are capable by themselves of discharging units. In contradistinction, those of exteroceptiveand other origins relayed by pathways from the entorhinal area would be analogous to conditional stimuli, lacking the capacity when at first acting alone to bring about a consummated response. Hippocampal output, in turn, appears to influence firing patterns of a substantial number of subcortical neurons with both excitatory and inhibitory affects. The direction and magnitude of hippocampal bias of subcortical neural activity appears to be related in part to the extent and frequency of hippocampal pyramidal cell output. Assuming complex behaviors to arise from probabilistic (vs. deterministic) interactions of neurons in the central nervous system, the hippocampus thus appears to act as one source of selective bias on behavioral patterns References p . 460461
460
J. A. G E R G E N
elaborated by thalamic and hypothalamic neurons. It is suggested that within this context, hippocampal function may well be related to directional and anticipatory components of ‘motivation’. ACKNOWLEDGEMENTS
Portions of the work reported here were carried out in the section on Limbic Integration and Behavior, Laboratory of Neurophysiology, National Institute of Mental Health, Bethesda, Md., under the guidance of Dr. Paul D. MacLean. Other portions of these studies have been subsequently supported in part by the U.S. Public Health Service Research Grants NB-3992, MH-10061, and FR-0080. The assistance of Mrs. Fabian Jackson with execution and anaIyses of experiments on learning set performance and discriminate reversal is gratefully acknowledged. REFERENCES
ADEY,W. R., (1958); Organization of the rhinencephalon. Reticular Formation of the Brain. Henry Ford Symposium. H. H. Jasper et ul., Editors. Boston, Little, Brown, pp. 621-645. ADEY, W. R., S ~ U N D O J. ,P., AND LIVINGSTON, R. B., (1957); Corticofugal influence on intrinsic brainstem conduction in cat and monkey. J. Neurophysiol., 20, 1-16. ANDERSEN, P., AND -0, T., (1967); Control of hippocampal output by afferent volley frequency. Structure andFunction of the Limbic System. Progress in Brain Research, vol. 27. W. Ross Adey and T. Tokizane, Editors. Amsterdam, Elsevier, pp. 400412. ANDERSEN,P., ECCLES, J. C., AND LBYMNG, Y.,(1964); Location of postsynapticinhibitory synapses on hippocampal pyramids. J. Neurophysiol,, 27, 592-607. DELGADO, J. M. R., AND SEVILLANO, M., (1961); Evolution of repeated hippocampal seizures in the cat. Electroenceph. clin. Neurophysiol., 13,122-733. FELDMAN, S., (1962) ;Neurophysiological mechanisms modifying afferent hypothalamo-hippocampal conduction. Exp. Neurol., 5, 269-291. GERGEN, J. A., AND MACLEAN, P. D., (1961); Hippocampal seizures in squirrel monkey. Electro-
enceph. clin. Neurophysiol., 13, 316-317. GERGEN, J. A., AND MACLEAN, P. D., (1962); A Stereotauic Atlas of the Squirrel Monkey’s Brain. Public Health Service Publication 933. Washington, U.S. Government Printing Office,91 pp. GERGEN, J. A., AND MACLEAN, P. D., (1964); The limbic system: photic activation of limbic cortical areas in the squirrel monkey. Ann. N. Y. Acad. Sci., 117,69-87. GERSTEIN, G. L., AND KIA”, N. Y.S., (1960); An approach to the quantitative analysis of electrophysiological data from single neurons. Biophys. J., 1,15-28. GLOOR, P.,(1963); Identification of inhibitory neurons in the hippocampus. Nature, 199, 699-700. GLOOR,P., VERA,C. L.,AND SPERTI,L., (1963); Electrophysiological studies of the hippocampal neurons. I. Configuratjon and laminar analysis of the ‘resting’ potential gradient, of the maintransient response to perforant path, fimbrial and mossy fiber volleys and of ‘spontaneous’ activity. Electroenceph. clin. Neurophysiol., 15, 353-378. GRASTYAN, E., LISSAK,K., MADARASZ, I., AND DONHOFFER, H., (1959); Hippocampal electrical activity during the development of conditioned reflexes. Electroenceph. clin. Neurophysiol., 11, 409430. GREEN, J. D., (1964); The hippocampus. Physiol. Rev., 44, 561-608. GREEN,J. D., AND ARDUW,A. A., (1954); Hippocampal electrical activity in arousal. Electroenceph. elin. Neurophysiol., 17, 534-557. GREEN, J. D., MAXWELL, D. S., SCHINDLER, W. J., AND STUMPP, C., (1960); Rabbit EEG ‘theta’ rhythm: its anatomical source in relation to activity in single neurons. J. Neurophysiol., 23, 403-420.
ISAACSON, R. L., DOUGLAS, R. J., AND MOORE,R. Y.,(1961); The effect of radical hippocampal ablation on acquisition of avoidance responses. J. comp. physiol. Psychol., 54, 625-628.
F U N C T I O N A L PROPERTIES O F H I P P O C A M P U S
46 1
JARRARD, L. E., (1965); Hippocampal ablation and operant behavior in the rat. Psychon. Sci., 2, 115-1 16. R.L., AND WICKELGREN, W. O., (1964); Effects of hippocampal ablation JARRARD, L. E., ISAACSON, and intertrial interval on runway acquisition and extinction. J. comp. physiol. Psychol., 57,442-444. KAJTOR,F., HULLAY, J., FARAGO, L., AND HABERLAND, K., (1957); Effect of barbiturate sleep on electrical activity of the hippocampus of patients with temporal lobe epilepsy. Electroenceph. clin. Neurophysiol., 9,441451. KIMBLE,D. P., (1963); The effects of bilateral hippocampal lesions in rats. J. comp. physiol. Psychol., 56, 273-283. KIMURA, D., (1958); Effects of selective hippocampal destruction on avoidance behavior in the rat. Cunad. J. Psychol., 12,213-218. LORENTE DE N6, R.,(1933); Studies on the structure of the cerebral cortex. I. The area entorhinalis. J. Psychol. Neurol. (Lpz.), 45, 381438. LORENTE DE N6, R.,(1934); Studies on the structure of the cerebral cortex. 11. Continuation of the study of the ammonic system. J. Psychol. Neurol. (Lpz.), 46, 113-117. MACLEAN, P. D., DENNISTON, R.H., DUA,S., AND P m , D. W.,(1962); Hippocampalchanges with brain stimulation eliciting penile erection. Physiologie de I'Hippocumpe. Colloques Internationaux du Centre National de la Recherche Scientifique,No. 107, Paris, pp. 492-510. MACLEAN, P. D., AND PLOOG, D. W., (1962); Cerebral representation of penile erection. J. Neurophysiol., 25, 29-55. MANZONI, T., AND PARMEGGIANI, P. L., (1964); Hippocampal control of the activity of thalamic neurones. Helv. physiol. pharmocol. Actu, 22, C28-C31. NAUTA, W. 3. H., AND GYGAX, P. A., (1954); Silver impregnation of degeneratingaxons in the central nervous system: a modified technique. Stain Technol., 29, 91-93. NIKI, H., (1962); The effects of hippocampal ablation on the behavior of the rat. Jup. Psychol. Aes., 4, 139-153. NIKI,H., (1967); Effects of hippocampal ablation on the learning behavior of the rat. Structure and Function of the Limbic System. Progress in Bruin Research, Vol. 27, W. Ross Adey and T. Tokizane, Editors. Amsterdam, Elsevier, pp. 305-317. PARMEGGIANI, P. L., (1967); On the functional significanceof the hippocampal 0-rhythm. Structure and Function of the Limbic System. Progress in Bruin Research, Vol. 27, W. Ross Adey and T. Tokizane, Editors. Amsterdam, Elsevier, pp. 41 3-441. J. M., ADEY,W. R., KADO,R.T.,AND REITE,M. L., (1963); Differing EEG sleep stages and RHODES, their response to tone stimulation in the chimp. Physiologist, 6,264. D. M., AND MCQUEENEY, J. A., (1963); Learning-set formation and discrimination RUMBAUGH, reversal: learning problems to criterion in the squirrel monkey. J. comp. physiol. Psychol., 56, 435-439. SPENCER, W. A., AND KANDEL, E. R., (1961); Hippocampal neuron responses to selective activation of recurrent collaterals of hippocampofugal axons. Exp. NeuroI. 4, 149-161.
462
RCactions de 1’ Hippocampe des Hypoxies Oxyprives Repetees chez le Chat Adulte et le Chaton P. P A S S O U A N T
ET
C. PTERNITIS
Laboratoire de Pathologie Expe‘rimentale,Faculte‘ de Me‘decine, Universite‘ de Montpellier, MontpeNier (France)
Les ltsions de la corne d‘Ammon provoqutes par l’anoxie sont bien connues, depuis les ttudes anatomiques de Spielmeyer (1962) et de Scholz (1952). La frtquence de ces ltsions est en faveur d’une atteinte stlective, encore ma1 expliquke et gtntralement rapportte, aprts Scholz (1953, B l’association d’un facteur circulatoire et d’un facteur mktabolique. Du point de vue circulatoire, la vascularisation atypique de la corne d’Ammon avec arttres terminales ‘en rateau’, sans suppltance par des collattrales (Nilges, 1944), favorise le retentissement d’une insuffisance circulatoire. Le secteur de Sommer irrigut par une seule arttre est particulitrement sensible. Les reactions hkmodynamiques, dtclenchtes par la carence en 0 2 , facilitent les lesions cellulaires par la stase capillaro-veineuse qu’elles entrainent. Les territoires ‘limites’ et en ‘bout d’arttre’, telle la formation ammonnienne, sont ainsi particulitrement touchts. Du point de vue mttabolique les donntes ne sont que partielles pour justifier cette atteinte stlective. Certaines substances pharmacodynamiques, telle la 3 acttyl-pyridine entrainent des lesions limitees aux secteurs CA3, CA4 (Coggeshall et MacLean, 1958,) La richesse en dthydrogtnase succinique des secteurs CA3, CA4, contribuerait A la plus grande rtsistance de cette rtgion A l’anoxie comparativement au secteur CA1 (Lammers et Gastaut, 1962). La prtsence de zinc au niveau des secteurs CA3, CA4 (MacLardy, 1960; Von Euler, 1962), la modification de la teneur en zinc aprts des hypoxies rtptttes, accompagntes de convulsions (MacLardy, 1962)’ sont une indication complkmentairesur la vulnkrabilitk ‘intrinshque’de la corne d’Ammon, en rapport avec une activitk mttabolique propre. Les donntes physiologiques concernant les rtactions de l’hippocampe B l’hypoxie manquaient. Les ttudes neuro-physiologiques ont par contre prtcist la sensibilitk des systtmes non sptcifiques du tronc ctrtbral. L‘activation de la formation rtticulaire, due aux dkcharges des chkmo-rtcepteurs carotidiens et aortiques survient au dkbut de l’hypoxie et provoque la ptriode d’activation corticale. Elle persiste durant la pkriode de dtpression qui dkbute par les ondes lentes corticales et au cours de laquelle se produisent les convulsions hypertoniques (Hugelin et al., 1959; Dell et al., 1961; Baumgartner et al., 1961). Afin de rechercher les rtactions de l’hippocampe B l’hypoxie et leurs variations selon
HIPPOCAMPE ET HYPOXIES OXYPRIVES
RBPBTBES
CHEZ LE CHAT
463
Mge, nous avons utilist des chats adultes et des chatons, libres de leurs mouvements et les hypoxies ont Ctt rCpCtCes plusieurs fois chez le mCme animal. Certains rCsultats ont dCjA fait l’objet de publications (Passouant et Pternitis, 1965; Pternitis, 1964; Pternitis et Passouant, 1961; Pternitis et al., 1963).
MATBRIEL
ET M ~ T H O D E S
L’hypoxie, provoquie par un appauvrissement en 0 2 , a CtC faite chez le chat, porteur d’klectrodes implantCes a demeure, selon une technique mise au point par l’un de nous. L‘animal est introduit dans une cage transparente en matitre plastique, facilitant son observation, d’une capacitt de 125 1. Une vanne ii pointeau permet l’introduction de l’azote. L‘tvacuation du gaz est assurCe par plusieurs vannes selon le dCbit choisi. La pression du gaz dans la cage est contr81Ce par un manomttre A eau. Un ventilateur crCe une circulation permanente de l’air de la cage travers les deux compartiments d’un bac perfork contenant le premier un gel de silice pour la fixation de la vapeur d’eau et le second de la chaux sodte pour la fixation du C02. La composition gazeuse de l’atmosphtre de la cage a Ctt CvaluCe par la micromtthode de Schollander ou A l’aide de l’appareil de Bechmann. Un connecteur de Cannon, Ctanche, scellC dans la paroi de la cage relie les electrodes de l’animal A l’appareil enregistreur. Les hypoxies ont CtC de deux types : les hypoxies modCrCes ( 0 2 : 8 %) et les hypoxies profondes ( 0 2 : 4 %). Dans tous les cas elles ont CtC rtpCtCes. Les hypoxiesprofondes: ( 0 2 : 4%) ont CtC soit d’installation rapide (3 A 4 min) soit d’installation progressive (6 B 15 min). Elles ont Ctt conduites jusqu’au coma et au silence Clectrique. La rtanimation obtenue, a p r b ouverture de la cage et parfois massage thoracique, a Ctt suivie jusqu’ii la rtcupCration compltte de l’animal. Ces experiences ont port6 sur 30 chats adultes et 10 chatons. Le nombre d’hypoxies a variC pour chaque animal de 2 B 80 et au total 300 hypoxies ont ttC pratiquCes. Les hypoxies mod6rLes: ( 0 2 : 8%) ont durt de 1 h 30 9 h. Dix chats adultes et 3 chatons ont CtC soumis a ce type d’hypoxie. Le nombre des hypoxies a variC de 2 A 12 et au total 54 hypoxies ont CtC effectukes. En complCment de l’hippocampe les structures CtudiCes ont CtC: la formation rCticulaire du tronc cCrCbral, le cortex, l’tcorce cCrCbelleuse. Les variations Clectriques, les modifications du comportement ont CtC suivies durant l’hypoxie, la rtanimation et parfois plusieurs mois aprts la dernikre hypoxie. Une etude histologique (Nissl et Luxol) des cerveaux a ttC faite pour la plupart des animau x.
HYPOXIES OXYPRIVES
RBPBTBES
CHEZ LE CHAT ADULTE
Hypoxie profande
L‘hypoxie profonde ( 0 2 : 4%) pousste jusqu’au stade de coma et de tract plat a Ctt References p. 474475
464
P. P A S S O U A N T , C. P T E R N I T I S
repetbe en moyenne de 10 a 20 fois chez le mtme animal et exceptionnellementdans un cas jusqu'8 80 fois. Certains animaux, gardCs en observation durant une pkriode allant de quelques jours a plusieurs mois aprts la dernitre hypoxie ont permis l'Ctude de sdquelles intkressant le systtme limbique. Les rkactions durant I'hypoxie profonde Les diffkrents types de reactions l'hypoxie profonde oxyprive dCcrits sur la preparation endphale isold (Creutzfeldt et al., 1957; Hugelin et al., 1959) sont retrouves chez l'animal libre de ses mouvements. La reaction &activation prbcCd6e d'une courte periode (20 21 30 sec) sans modification Clectrique ni de comportement, a une durde de 40 B 60 sec. L'animal est agitC et cherche a sortir de sa cage. L'activitC Clectrique corticale est rapide, l'activite Clectrique de la formation reticulaire est acckldrke (20 a 30 c/sec). La reaction d'inhibition commence par une pCriode de depression corticale et aboutit 8 la disparition de toute activitC Clectrique. Sa durke varie entre 2 et 4 min. Pendant la phase de ddpression corticale, avec ondes lentes corticales l'animal est inquiet et miaule plaintivement. I1 ne peut maintenir la position debout ou assise et se couche. L'activitk electrique de la formation rt%culaire est rapide, en opposition a I'activitC lente corticale. C'est la fin de cette periode que surviennent les convulsions anoxiques de courte
Fig. 1. En haut, pointes hippocampiques au dkbut de la m o d e de dt5pression corticale. En bas. fuseau hippocampique en p6riode de depression corticale prononde (hypoxie profonde Oa: 4%), RM = rkticulk m&enc&phalique; H = hippocampe; CES = cortex ecto-sylvien; CSS = cortex supra sylvien).
H I P P O C A M P E ET H Y P O X I E S O X Y P R I V E S R E P E T E E S C H E Z LE C H A T
465
durte, traduites par des mouvements dtsordonnts et par un spasme hypertonique type rigiditt de dtctrtbration. Ces paroxysmes convulsifs ont 6tt trouvts dans 85 % des cas. La ptriode terminale ou ptriode de coma est associte B un tract plat, coup6 de courtes boufftes d’ondes ralenties ‘en dent de peigne’ qui inttressent diverses structures (cortex, cervelet, formation rtticulaire). L’activite‘dectrique de l’hippocampe: a des caractbres particuliers au cours des difftrentes ptriodes: (1) Lors de la ptriode d’activation, l’activitt tlectrique de l’hippocampe correspond B des rythmes 0 et cette synchronisation rappelle la synchronisation hippocampique de la rtaction d’alerte. (2) Lors de la ptriode de depression corticale l’activitt tlectrique hippocampique d’expression polymorphe avec ondes rapides et ralenties est surchargte de pointes isoltes et de courts fuseaux rapides (Fig. 1). Les pointes hippocampiques, amples (200 B 300 pV) apparaissent en m2me temps que les ondes lentes corticales et persistent jusqu’au dtbut de la ptriode terminale. Ces pointes sont constantes, surviennent dts la premikre hypoxie et sont retrouvtes au cours des hypoxies successives. Les fuseaux rapides (30 A 40 c/sec) sont de courte durte ( 5 B 10 sec) et se produisent & la fin de la ptriode de dtpression corticale. Ces fuseaux localists B l’hippocampe n’ont pas de propagation A d’autres structures, en particulier B la formation rtticulaire, et peuvent apparaitre aprbs la disparition des rythmes rapides rtticulaires. 11s ne surviennent qu’aprbs plusieurs hypoxies et ne sont pas rtgulitrement retrouvts au cours des hypoxies suivantes. 11s ne sont accompagnts d’aucune modification particulitre de comportement. Des fuseaux rapides analogues, de courte durte et localists, ont t t t enregistrts, mais avec une frtquence moindre, au niveau du cortex et du cervelet. Au cours de 300 hypoxies, les fuseaux hippocampiques se sont produits dans 17% des cas, les fuseaux cirtbelleux dans 6 % et les fuseaux corticaux dans 2 %.
Fig. 2. Persistance d’une activitb blectrique dans l’hippocampe (H) lors de la p6riode terminale de l’hypoxie profonde. RM = r&icul& mksenckphalique; CM = centre rn6dian; CS = cortex suprasylvien; CES = cortexeto-sylvien;V = vermis; Cr = crus). References p. 474475
466
P. P A S S O U A N T , C. P T E R N I T I S
(3) Lors de la ptriode terminale, une activitt tlectrique hippocampique peut persister alors que l'activitt tlectrique du cortex et du cervelet ont disparu. Cette constatation contraire i.l'tpuisement rapide de l'activitt hippocampique en cours d'anoxie ischtmique (Sugar et Gerard, 1938) est en faveur d'une resistance particulitre de l'archtocortex 21 l'hypoxie oxyprive (Fig. 2). Les rkactions durant la rkanimation Elles correspondent aux variations klectriques successives dtcrites par Bremer et Thomas (1936) et sont prkcistes par les modificationsde comportement. Dts le dtbut de la reanimation le track tlectrique est plat, surchargt de courtes ptriodes 'd'ondes en dent de peigne'. Puis survient, pendant 20 a '10 sec, une pkriode d'agitation avec convulsions atypiques et spasme tonique type rigiditt de dtckrtbration. Par la suite, l'activitt electriquecorticale se prtcise avec phase d'ondes lentes type sommeil B laquelle succtde une activitt de veille normale. Durant cette dernitre ptriode, l'animal se redresse, les fonctions vtgttatives se normalisent et le comportement habitue1 est retrouvk. L'activite' dectrique de l'hippocampe : est ordinairement la premitre a se rtorganiser sous la forme de rythmes rapides et peu amples. La facilitk convulsivante de la ptriode de rkanimation (Gellhorn et Heymans, 1948) peut se traduire au niveau de l'hippocampe par des pointes, des fuseaux rapides et des decharges organistes. Les pointes hippocampiques peuvent apparaitre dts le dtbut de la rtanimation, elles sont plus frtquentes lors de la pkriode des ondes lentes corticales. Les fuseaux rapides sont comparables a ceux qui surviennent en cours d'hypoxie.
Fig. 3. Dkharge hippocampo-rkticulaire spontanke survenue en cours de rkanimation chez un chat adulte aprks 12 hypoxies profondes.
HIPPOCAMPE ET HYPOXIES O X Y P R I V E S REPETEES CHEZ LE CHAT
467
11s sont localists, non propagts, de courte durte, ils apparaissent au dtbut de la rtanimation et aprbs plusieurs hypoxies. Des fuseaux analogues ont t t t enregistrks, dans les m&mesconditions, au niveau du cortex et du cervelet. Les dtcharges tonico-cloniques organistes, d’une durte de 2 B 3 min, inttressant l’hippocampe, la formation rtticulaire et parfois le cortex surviennent aprbs plusieurs hypoxies, soit dans les premibres minutes de la rkanimation (Fig. 3), soit tardivement (aprbs 1 h). Ces dtcharges sont rares et sont accompagntes d’une rtaction de malaise avec miaulements ou de reactions orales (mastication, salivation). Les dtcharges qui intkressent le systtme limbique sont plus frtquentes que les dtcharges gtntralistes, soit myocloniques, soit crises genkralistes, ces dernitres ne survenant qu’a une pkriode tardive de la rtanimation. Les skquelles post-anoxiques Les stquelles post-anoxiques dtpendant d‘une rtaction de l’hippocampe et du systtme limbique ii des hypoxies oxyprives rtptttes peuvent avoir diverses expressions: (1) soit dtcharges hippocampo-rtticulaire facilities par le sommeil ou par des stimulations sensorielles (auditives, stimulation lumineuse intermittente); (2) soit crises amygdaliennes avec mastication et salivation; (3) soit troubles du comportement avec hyperphagie, exagtration des rtactions de bien-Ctre (grattage et ltchage), dtviations sexuelles dtfaut d’adaptation dans un groupe d‘animaux de mCme espbce, ensemble de manifestations proches du syndrome de Kliiver-Bucy. Au cours d‘un Ctat de mal, survenu aprbs 24 hypoxies le r81e jout par l’hippocampe a pu Ctre pr6cist. Cet etat de ma1 d’une durte de 45 h a evolut en 4 pCriodes: la premibre (4 h) s’est traduite par des crises gtntralistes et des crises partielles hippocamportticulaires; la seconde (24 h) a correspondu & un ttat de stupeur avec activitt Clectrique globalement dtprimte en particulier dans l’hippocampe ; la troisibme (14 h) traduite par une agitation d’intensitt variable de l’animal, ttait accompagnte de dtcharges hippocampiques, soit grandes pointes survoltees et isoltes soit courtes dtcharges hippocampo-rkticulaires, assocites A une recrudescence de l’agitation ; la
Fig. 4. Dkcharges spontanks hippocampo-rkticulairesurvenues en cours d’hypoxie modkr6e (8 %) et reproduites A la 4Rme min de skjour en hypoxie. References p . 474475
468
P. P A S S O U A N T , C. P T E R N I T I S
quatritme (3 h) terminale s'est caractkriste B son debut par des crises gtn6ralisees, puis par des crises partielles hippocampiques et corticales uniquement tlectriques et sans manifestations motrices.
Les Ibsions anatomiques de I'hippocampe Les lbions anatomiques de l'hippocampe sont constantes chez les animaux qui ont subi plusieurs hypoxies, mais elles sont d'importance variable.Les necroses sectorielles sont rares. Elles ont port6 sur le secteur CA1 ou sur les secteurs CA3, CA4 en s'ktendant au gyrus dentatus. Ces necroses sectorielles n'ont pas de relation avec le nombre des hypoxies: obtenues aprks 5 B 10 hypoxies elles n'ont pas 6tt5 trouvtes chez un animal qui a subi 70 hypoxies. Les lesions cellulaires sont constantes et correspondent 21 un oedhme pkricellulaire, B une hypertrophie neuronale avec vacuoles cyto-plasmiques, B une dtgdntrescence ischemique neuronique du type Spielmeyer. Hypoxie modbrbe L'abaissement $ 0 2 A 8 % a t t t progressif et obtenu en 15 min. La dude de l'hypoxie a varit entre 1 h 30 et 9 h. Le nombre des hypoxies rkpet6es chez le mtme animal a CtC compris entre 2 et 12. Les reactions B ce type d'hypoxie varient au cours de 3 pdriodes successives. La premitre correspondant A l'abaissement 802 de 21 % h 12% n'entraine aucune modification tlectrique ou de comportement. La seconde ou periode &activation dkbute pour une teneur en 0 2 de 12% : l'animal cherche A sortir de sa cage, miaule, le coeur et la respiration sont accC16res, les rythmes corticaux sont rapides, les rythmes hippocampiques synchronids. La troisitme correspond A une phase de depression corticale modCree avec periode de sommeil coupe d'tveil et traduit un &at d'accoutumance B l'anoxie (Fig. 4). Les modifications hippocampiques correspondent h un abaissement du seuil convulsivant et A l'apparition de paroxysmes spontanees. (I) L'abaissement du seuil convulsivantde l'hippocampe A la stimulation Clectrique est associd A la prolongation de la post-decharge. Aprts 10 min de skjour en atmosphbre appauvrie en 0 2 la post-dbcharge a une durCe de 50 sec. Aprbs 120 min, la postdCcharge persiste pendant 90 sec et peut se reproduire spontankment. L'abaissement du seuil convulsivant et la prolongation de la post-decharge persistent lorsque l'animal est remis ii l'air libre et leur importance est fonction de la durhe de l'hypoxie. Une post-decharge de 50 sec est obtenue aprts une hypoxie mod6rCe d'une dude de 2 h, une post-dtcharge de 200 sec aprhs une hypoxie de 6 h. (2) Les paroxysmes spontants ne surviennent qu'aprts plusieurs hypoxies et correspondent soit 21 des fuseaux, soit h des dtcharges tonico-cloniques. (a) Les fuseaux hippocampiques de 10 h 15 c/sec ou de 20 h 25 c/sec sont de courte dude et sont propagts h la formation rdticulaire. (b) Les dkcharges tonico-cloniques peuvent ttre prdcdddes de pointes rythmiques et de fuseaux. Elles inttressent la formation reticulaire mtsenctphalique. Elles peuvent se
H I P P O C A M P E ET H Y P O X I E S O X Y P R I V E S REPETEES C H E Z LE C H A T
469
.
Fig. 5. Reactions a I’hypoxie profonde d’un chaton de 6 jours I, Activation hippocampique (H). 11, Fuseau dans la formation reticulaire (RM) et le vermis (V).111, Pkriode de track plat.
propager B l’amygdale et au cortex et se traduisent alors par des clonies faciales, une salivation et des mouvements de mastication (Fig. 5). Ces dCcharges hippocampiques, propagkes aux formations limbiques ont CtC les seules enregistries au cours de ce type d’hypoxie. I1 n’a jamais CtC observC de dCcharges gCnkraliskes du type myoclonique ou de crise globale. La seule reaction corticale a correspondu B l’inscription de pointes rythmiques, rares et sans modification de comportement. Les lesions histologiques sont minimes et se limitent A des atteintes cellulaires rares et dissCminCes. HYPOXIES OXYPRIVES
RBPBTBES
C H E Z LE C H A TO N
Des hypoxies profondes ( 0 2 : 4 %) et des hypoxies modtries ( 0 2 : 8 %) ont CtC rCpetCes au nombre de 2 B 8 pour le m&meanimal, chez des chatons d’hge compris entre 2 et 30 jours. Les animaux ont CtC sacrifiks aprts la dernibre hypoxie et il n’a pas CtC fait d’ktude des stquelles post-hypoxiques. Les rkactions observkes ont CtC diffkrentes au cours de la pCriode nko-natale et aprbs le 15tme jour. La piriode nio-natale Le chaton nouveau-nC est particulitrement rksistant B l’hypoxie. Des hypoxies rCpetkes au nombre de 5 entre le 2bme et le 3bme jour aprks la naissance ont Ctk trbs bien supportCes. Ces rksultats sont comparables B ceux obtenus chez le cobaye B la naissance (Windle et al., 1944) et chez le raton (Stafford et al., 1960). Les rCactions B l’hypoxie profonde sont trts diffkrentesde celles de l’animal adulte et les diffkrentes pCriodes sont ma1 isolCes. Toutefois une activation de la formation rkticulaire, traduite par des rythmes rapides de 10 B 15 c/sec est indentifiCe dis le deuxibme jour aprbs la naissance. Des pointes isoltes, parfois rCpktCes s’inscrivent au niveau du cortex. Des rythmes rapides organisis apparaissent dans le vermis du cervelet. La ptriode terminale avec activitC Clectrique nulle est particulikrementprononcte et References p. 474-475
470
P. P A S S O U A N T , C. P T E R N I T I S
Fig. 6. Activation hippocampique en cours d'hypoxie profonde et de rkanimationchez un chaton de 6 jours.
RM.
~
-
c _ N _
-H CES-A-
css
w
.
>
peut persister durant 1 B 2 min au cours de la reanimation (Fig. 6). Elle est parfois surchargee de courts fuseaux interessant simultantment l'hippocampe, la formation reticulaire et le vermis du cervelet. L'activation hippocampique a des expressions differentes durant l'hypoxie et la reanimation.
H I P P O C A M P E ET HYPOXIES O X Y P R I V E S REPETEES CHEZ LE C H A T
471
En cours d‘hypoxie profonde des pointes isoltes ou rythmiques, des fuseaux avec ondes rapides (12 ii 15 c/sec) ou ralenties (6 a 7 c/sec) sont identifites dbs les premiers jours aprbs la naissance. Les rythmes hippocampiques peuvent Ctre les derniers A disparaitre avant la ptriode de tract nul. En cours d’hypoxie modtrte de longues ptriodes d’activitt synchroniste de 6 A 7 c/sec surviennent la fin de la premike semaine. En cours de reanimation l’activation hippocampique est plus precise. Elle se traduit soit par des pointes rythmiques, rtptttes durant plusieurs minutes, soit par des fuseaux rapides (25 30 c/sec) limit& a la come d‘Ammon. Durant ces modifications tlectriques le chaton presente une agitation dtsordonnte associCe a des miaulements (Fig. 7). Ces rtactions de l’archto-cortex i% l’hypoxie s’opposent par leur prkcision, B celles du nto-cortex traduites uniquement par des pointes isoltes. Cette difftrence, en faveur d’un niveau de maturation hippocampique plus avanct, est appuyte par ailleurs par l’existence dbs la naissance d’une activitt rtgulikre hippocampique de 8 a 12 c/sec et rtactive A certaines stimulations sensorielles (olfacto-trigtminte) (Cadilhac et Passouant-Fontaine, 1962). La pbriode du 15dme au 3Odme jour La pCriode comprise entre le 15bme et 30kmejour est caracttriste par l’organisation progressive des ptriodes d’activation et de depression telles qu’elles existent chez l’animal adulte et par l’apparition de dtcharges tonico-cloniques en cours de rtanimation. L’activation corticale, traduite par des rythmes rapides est indiqute dbs le 15bme jour, la synchronisation hippocampique est precise au 30bme jour. Lors de la ptriode d’inhibition, des ondes ralenties corticales surviennent dbs le 15bmejour et sont comparables B celles de l’adulte, a la fin du premier mois (Fig. 8). En ptriode de rtanimation des dtcharges organistes tonico-clonique apparaissent A partir du 15bmejour. Ces dtcharges de longue durte (3 4 min) inttressent la formation rtticulaire et peuvent se propager au cortex (Fig. 9). Elles sont accompagntes sur le plan comportemental, d’une agitation de l’animal, associte a des miaulements plaintifs. Ces dtcharges sont frtquentes et peuvent se rtptter durant la reanimation. Elles surviennent, aprbs la premibre hypoxie et non, comme chez l’animal adulte, aprks plusieurs hypoxies. Elles se reproduisent, dans une proportion de plus de 50%’ au cours des hypoxies successives. Des ltsions histologiques inttressent les neurones des divers champs de la come d’Ammon, mais il n’a pas Ctt observt de nCcrose sectorielle. DISCUSSION ET CONCLUSIONS
La rtpttition d‘hypoxies oxyprives, chez le chat libre de ses mouvements, prtcise les reactions de l’archto-cortex zi ce type d’hypoxie, indiqutes par ailleurs par les ttudes anatomiques classiques. Les rtactions tlectriques de la come d’Ammon varient avec le degrt de l’hypoxie, sa rtpttition et en fonction de l’lge de l’animal. References p. 474475
472
P. PASSOUANT, C. P T E R N I T I S
Fig. 8. Reactions il l’hypoxie profonde d‘un chaton de 15 jours.I, Pkriode d‘activation; 11, Pkriode de dkpression corticale. III, Pkriode de track plat.
Fig. 9. Decharge spontank hippocampo-dticulaireet corticale aprh 3 min 40 sec de rhnimation, chez un chaton de 15 jo-.
Chez I’animal adulte Trois types de risultats miritent &&re isolis : les rtactions ilectriques constantes, les dtcharges Cpisodiqueset la rtsistance de l’activit6 Clectrique de la corne d’Ammon a l’hypoxie. (a) Les riactions constantes correspondent A la synchronisation hippocampique de la piriode d’activation de l’hypoxie et aux pointes hippocampiques qui surviennent lors de la pCriode de dtpression corticale. Ces 2 types de modifications de l’activitt Blectrique de l’hippocampe, dicrites, l’une en piriode d‘alerte, l’autre au cours du
HIPPOCAMPE ET HYPOXIES OXYPRIVES REPETEES CHEZ LE C H A T
473
sommeil lent (Rimbaud et al., 1955), objectivent au niveau de l’archto-cortex les 2 ptriodes successivesd’activation et de dtpression provoqutes par I’hypoxie (Dell et al., 1961). La frtquence et la prtcision des pointes hippocampiques peuvent Ctre facilities par l’activation rtticulaire qui persiste durant la ptriode de dtpression corticale. (b) Les dtcharges tpisodiques ne surviennent qu’aprbs plusieurs hypoxies et correspondent aux fuseaux hippocampiques et aux dtcharges hippocampo-rtticulaires. (1) Les fuseaux, de la pCriode terminale de l’hypoxie profonde et du dCbut de la rtanimation, sont localisis l’hippocampe et sans propagation. Ce type d’activation parait exprimer une modification des conditions locales de l’hippocampe en liaison avec l’appauvrissement extrgme et rtpttt du milieu en 0 2 . I1 ne parait pas dtpendre de I’activation rtticulaire qui, dans nos conditions d’exptriences, peut avoir disparu lors de la production de ces fuseaux. Des dtcharges analogues enregistrtes au niveau du cortex ecto-sylvien et du cervelet, la ptriode terminale de l’hypoxie profonde et au dtbut de la rtanimation paraissent traduire aussi, un processus d’activation IocalisCe. La plus grande frtquence des fuseaux hippocampiques est en faveur d‘une atteinte sClective de cette structure. Ce type d’activation localiste parait Ctre la traduction tlectrique d’une souffrance precise des neurones dont la constquence pourrait Ctre un abaissement du potentiel de membrane. Un tel mtcanisme a CtC retenu pour l’activation des neurones spinaux au cours d’hypoxie(Kolmodin et Skoglund, 1959), et pour l’activation qui accompagne la ptriode de rtanimation (Baumgartner et al., 1961). (2) Les dtcharges hippocampo-rkticulaires surviennent en milieu plus riche en 0 2 (hypoxie ,modtrte) lors de la rtanimation et en tant que stquelles post-hypoxiques. Ces dtcharges ont probablement un mtcanisme different selon les conditions de leur production. L‘absence de ltsions hippocampiques, chez les animaux soumis B des hypoxies modtrtes est en faveur d‘une perturbation mttabolique rtversible qui prtctde les ICsions anatomiques irrtversibles observtes aprts des hypoxies profondes rtptttes. La frtquence des dtcharges organistes hippocampo-rtticulaires, comparativement aux dtcharges corticales et gtntralistes, souligne B nouveau la reactivitt particulitre de l’hippocampe a I’hypoxie oxyprive. Une telle sensibilitt va de pair avec la rtactivite exquise de l’hippocampe A la stimulation tlectrique, mtcanique ou chimique et qui est facilitte par l’organisation architectonique de cette structure (Passouant et Cadilhac, 1962). (3) L’activitC Clectrique de l’archto-cortex est plus rtsistante a I’hypoxie que celle du nCo-cortex. Cette constatation en faveur d‘un moindre besoin en 0 2 des structures phylogCnCtiquement anciennes, est appuyCe par certains rCsultats sur la consommation en 0 2 de diverses structures ctribrales: cette consommation &ant plus ClevCe pour le cortex que pour la corne d’Ammon (Quastel et Quastel, 1961). Chez le chaton Deux rtsultats mtritent d’&treisolts : (1) la sensibilitt exquise de l’hippocampe a I’hypoxie, (2) I’expression difftrente de la rtaction hippocampique au cours du premier mois. (1) La sensibilitC de l’hippocampe aux hypoxies oxyprives rtpCttes est plus grande References p . 474475
414
P . P A S S O U A N T , C. P T E R N I T I S
que chez l'animal adulte. Les rCactions hippocampiques (pointes rythmiques, fuseaux) sont plus frtquentes. Les dkcharges tonico-cloniques, apparues d5s le 15bme jour, surviennent des la premiere hypoxie et sont retrouvkes dans une proportion de plus de 50 % au cours des hypoxies suivantes. Ces rksultats en prtcisant la rtactivitt de l'hippocampe ii l'hypoxie dks la naissance, confirment la facilitt convulsivante du jeune animal. (2) L'organisation progressive des rkactions hippocampiques B l'hypoxie au cours du premier mois, confirme les indications dbjh obtenues sur l'tvolution de la maturation de l'hippocampe (Cadilhac et Passouant-Fontaine, 1962; Passouant et al., 1965). C'est ainsi que les dtcharges hippocampiques, provoqutes par l'hypoxie ii partir du 1%me jour, prkisent l'organisation h ce moment des relations hippocampo-rtticulaires. Ces ddcharges hippocampiques sont les premitres ii apparaitre, les dkcharges gtnCralisCes ne survenant qu'apres le premier mois. D'aprbs ces premiers rbultats, indiquant les variations de l'activation de l'hippocampe par l'hypoxie selon l'fige, l'ttude de l'hypoxie pourrait Ctre retenue chez le jeune animal comme un test valable dans l'ttude de la maturation ctrtbrale. BIBLIOGRAPHIE
BAUMGARTNER, G., CREUTZFELD, O., AND JUNG, R., (1961); Microphysiology of cortical neuron- in acute anoxia and in retinal ischemia. Cerebral Anoxia and the Electroencephalogram. Springfield, Thomas, pp. 5-34. BREMER, F., ET THOMAS, J., (1936); Action de l'anoxie et de l'hypercapnie sur l'activite klectrique du cortex drkbral. C.R. SOC.Biol. (Paris), 123, 1256-1260. CADILHAC, J., ET PASSOUANT-FONTAINE, TH., (1962); Dkcharges kpileptiques et activitk klectrique de veille et de sommeil dans l'hippocampe au cours de l'onto&&e. Physiologie de I'Hippocampe. Paris, Centre National de la Recherche Scientifique, pp. 429442. C~GGESHALL, R. E. AND MACLEAN,P. D., (1958); Hippocampal lesions followhg administration of 3-acetylpyridine. Proc. SOC.Exp. Biol. N.Y., 98, 687-689. CREUTZFELD, O., KASAMABU,A., UND VAZ-FERREIRA, A., (1957); Aktivitatsanderungen einzelner corticaler Neurone im akuten Sauerstoffmangel und ihre Fkziehungen zum EEG tn?i Katzen. Pfliigers Arch. ges. Physiol., 263,647-667. DELL, P., HUGEF, A., ET BONVALLET, M., (1961); Effects of hypoxia on the reticular and cortical diffusesystems. Cerebral Anoxia and the Electroencephalogram. Springfield, Thomas, pp. 46-58. EULER,C. VON, (1962); On the signifcance of the high zinc content in the hippocampal formation. Physiologie de I'Hippocampe, Paris, Centre National de la Recherche Scientifque, pp. 135-145. GELLHORN, E., AND HEYMANS, C., (1948); Differentialaction of anoxia, asphyxia and carbon dioxide on normal and convulsive potentials. J. Neurophysiol., 11, 261-274. HUGELIN, A., BONVALLET, M., ET DELL,P., (1959); Activation reticulaire et corticale d'origine chemoceptive au c o w de l'hypoxie. Electroenceph. clin. Neurophysiol., 11, 326340. KOLMODIN, G. M., AND SKOGLUND,C. R., (1959); Influence of asphyxia on membrane potential level and action potentials of spinal mot0 and intemeurones. Actaphysiol. scand., 45, 1-18. LAMMERS, H. J., ET GASTAUT, H., (1962); Relations cyto-architectoniqueset enzymo-architectoniques dam l'hippocampe. Physiologie de I'H*pocampe. Paris, Centre National de la Recherche Scientifique, pp. 1-21. MACLARDY, T., (1960); Neuro syncitial aspects of the hippocampal mossy fibre system. Confin. neurol. (Basel), 20, 1-17. MACLARDY,T., (1962); Pathological zinc rich synapse. Nature (Land.), 195, 1315-1316. NILGES, R. E., (1944); Arteries of the mammalian comu Ammonis. J. comp. Neurol., 80, 117-190. PASSOUANT, P., ET CADILHAC, J., (1962); Place de l'hippocampe dans l'organisation fonctiomelle du cerveau. J. Psychol. (Paris), No. 4.
H I P P O C A M P E ET H Y P O X I E S O X Y P R I V E S REPETEES C H E Z LE C H A T
475
PASSOUANT, P., CADILHAC, J., ET PASSOUANT-FONTAINE, TH., (1965); Hippocampe et maturation &rebrale. Actualitis Neurophysiologiques, 6 eme drie. Paris, Masson, (sous presse). PASSOUANT, P., ET PTERNITIS, C., (1965); Activation de l'hippocampe par des hypoxies oxyprives repetks. Acta. physiol. Acad. sci. hung., 26, pp. 123-130. PTERNITIS, C., (1964); Reactions de I'hippocampe a des hypoxies oxyprives repktks. Etude dectrophysiologique et anatomique. Th2se Mid., Montpellier, 106 pp. PTERNITIS, C., ET PASSOUANT, P., (1961); Modifications EEG et de comportement au cours de l'hypoxie repet& et de la reanimation chez le chat libre de ses mouvements. C.R. Soc. Biol. (Paris), 155, 2380-2397.
PTERNITIS, C., PASSOUANT, P., ET CADILHAC, J., (1963); Etat de ma1 post-anoxique, place de l'hippocampe dans l'entretien de l'etat de mal. C.R. SOC.Biol. (Paris), 157, 347-349. QUASTEL, J. H., AND QUASTEL, D. M., (1961); The Chemistry of Brain Metabolism in Healthand Diseases. Springfield, Thomas, 170 pp. RIMBAUD, L., PASSOUANT, P., ET CADILHAC, J., (1955); Participation de l'hippocampe a la regulation des Ctats de veille et de sommeil. Rev. Neurol., 93, 303-308. SCHOLZ, W., (1952); Les nkcroses parenchymateuses electives par hypoxkmie et oligkmie et leur expression topistique. Atti I" Congr. Intern. Istopathol. Sist. Nerv., 1, 321-346. SPIELMEYER, W., (1962); Histopathologie des Nervensystems, Berlin, Springer. SUGAR,O., AND GERARD, R. W., (1938); Anoxia and brain potentials. J. Neurophysiol., 1, 558-572. STAFFORD, A., ASD WEATHERALL, J. A. C., (1960); The survival of newborn rat in nitrogen. J. Physiol. (Lond.), 153,457412. WINDLE,W. F., BECKER, R. F., AND WEILL,(1944); A., Alteration in brain structure after asphyxiation at birth. An experimental study in the guinea-pig. J. Neuropath. exp. Neurol., 3, 224-238.
476
Author Index * Abrahams, V. C., 111 Adam, D., 295 Adamovitch, N. A., 295 Adey, W. R., 3, 85, 171, 218, 228-245, 254, 274, 393,425,429,430,455,457 Adrian, H. O., 27,28 Adrianov, 0.S., 340 Auapetyants, E. Sh., 293-304 Ajmone Marsan, C., 2, 190, 339 Akert, K.,6 6 3 , 1 1 4 , 134,413 Albe-Fessard, D., 271,272 Alcocer-Cuaron, C., 246 Anand, B. K., 1,2,25, 132 Andersen, P., 114, 119, 174, 320,400412, 452,457 Andy, 0. J., 114, 134 Antunes Rodrigues, J., 190,201, 203,210,214 Aoki, I., 129 Archibald, D., 91, 94 Arduini, A., 218,232,290,413-415,429, 430,433,445 Asanuma, H., 28 Asanuma, H., 318 Aserinsky, E., 420 Azzali, G., 176 Babkin, B. P., 300 Badier, M., 180 Bagshaw, M. H., 325,329 Baldwin, M., 175 Ballim, H. M., 360 Balvin, R. S., 305 Ban, T., 85, 173, 203 Bando, T., 384 Barbizet, J., 229 Bard, P., 176 Barker, S. A., 242 Barraclough, C.A., 2, 69 Barrnett, R. J., 241 Baumgartner, G., 462,464,473 Bayyuk, S. H. J., 242 Beck, E. C.,246 Becker, R. F., 469 Bell, F. R., 235 Bergmann, F., 183-188 Beteleva, T. G., 413 Biscoe, T. J., 359 Blackstad, T. W., 401,403
Bliss, C. I., 365 Bond, D. D., 205 Bonnet, V., 257, 271, 338, 423 Bonvallet, M., 23, 46, 462, 464, 473 Borgest, A. N., 295 Boswell, R. S., 433 Bowden, 3. W., 34 Boyajy, L. D., 360 Boyarsky, L. L., 361 Bradley, P. B., 338, 359 Brady, J. V., 176,205,295,433 Bremer, F., 34, 257,271, 290, 338,423,424, 466 Bremner, F., 229 Brinley, F. J., 404,411 Broadwick, M., 333 Brobeck, 5. R., 1 Brodal, A., 430 Brodwick, M., 313 Brookhart, 5. M., 430 Brooks, C. McC., 76 Brooks, D. C., 3 Brooks, V. B., 318 Brown, R. M., 422 Brown, T. S., 228,229,231, 233, 234 Brown, W. C., 360 Briicke, F., 413 Brugge, J. F., 413, 433 Briigger, M., 103, 117, 118 Bruland, H., 378 Brust-Carmona, H., 246 Brutkowski, S., 316 Bucher, V. M., 103-127 Buchwald, N. A., 378 Bunn, J. P., 71 BureS, J., 166 Biirgi, S., 124 Buser, P., 254, 274 Butter, C. M., 316 Candia, O., 203, 204 Cadhillac, J., 114,413,462, 471,473, 474 Cajal, S. Ram6n y, 401 Cannon, W.B., 111 Cardo, B., 210 Carlsson, A., 398 Carreras, M., 176 Caspers, H., 430
* Italics indicate the pages on which the paper of the author in these proceedings is printed.
AUTHOR I N D E X
Cazard, G., 254, 274, 429 Cenacchi, V., 422,430 Chaillet, F., 356 Chairnovitz, M., 183-188 Chambers, W. W., 2 Chamorro, A., 69 Chang, H. T., 26 Charbon, G. A., 203 Chernigovskii, V. N., 295 Chhina, G. S., 2, 25 Chow, K. L., 251 Clark, C. V. H., 307 Clark, W. G., 382-384 Clemente, C. D., 1, 29, 34-47, 176 Coggeshall, R. E., 462 Corazza, R., 413423,429, 430 Cordeau, J. P., 311,425 Corti, U. A,, 111 Costin, A., 183-188 Cotten, M. de V., 203 Cottrell, G. A., 26 Courvoisier, S., 384 Coury, 5. N., 213 Covian, M. R., 189-217 Cragg, B. G., 26 Cramer, M., 243 Creuzfeldt, O., 28, 242, 462,464, 473 Critchlow, V., 72 Crosby, E. C., 62, 85 Cross, B. A., 2 Crossland, J., 340 Csanaky, A , 114 Curtis, D. R., 359 Daigneault, E. A., 360 Dalla Rosa, V., 176 D’Amour, F. E., 365 Dashiell, J. F., 312 De Groot, J., 26, 176, 190,211 Deisenhammer, E., 413 Delgado, J. M. R., 1, 29,4848, 119, 129, 130, 134, 170, 180, 189,204,445 Dernpsey, E. W., 34 Dell, P., 28, 46, 291, 424, 462, 464, 473 Delov, V. E., 295 Dernber, W. N., 313, 316,333 Dement, W., 383,420 Demetrescu, Maria, 290, 424 Demetrescu, M., 290, 423, 424 Denisenko, P. P., 359 Dennis, B., 235 Denniston, R. H., 445 Desmedt, J. E., 359,424 Dews, P. B., 365 Dewson, 111, J. H., 322 Diamond, F. D., 305 Diamond, S., 305 Didio, J., 237 Dietiker, M., 111
477
Dodo, T., 174 Domino, E. F., 337-364 Donhoffer, H., 218, 219, 222, 223,235, 274, 413,425,445 Doty, R. W., 246 Douglas, R. J., 313, 329, 333, 458 D r a c h m a , D. A., 229 Dren, A. T., 337-364 Dua, R. H..445 Dua, S., 2, 25 Ducrot, R., 384 Durnont, S., 290, 424 Dunlop, C. W., 218, 229,235,429 Eccles, 5. C., 12, 27, 338, 404, 452 Eccles, Sir J., 320 Eckhaus, E., 246 Egger, M. D., 63, 165-182 Egyhazi, E., 241 Elkes, J., 359 Elliasson, S., 203 Elul, R., 243, 290, 429 Elwers, M., 72 Endrkzi, E., 246-253 Epstein, 5. A., 296, 300 Evarts, E. V., 144 Everett, J. W., 70, 71, 76, 190 Fangel, C. H., 119, 251,258,273 Farago, L., 445 Fatt, P., 27 Favale, E., 203, 204, 425 Fee, A. R., 76 Feldberg, W., 337, 340 Feldman, S., 457 Fernandez de Molina, A., 65,103, 114, 117-119, 122, 177, 180 Fernandez-Guardiola, A., 62 Fessard, A,, 254 Fifkovk, E., 151 Fisher, A. E., 213 Flataker, L., 365 Flerkb, B., 25,26, 69, 170 Flym, 5. P., 63,165-182 Folkow, B., 203,204 Fonberg, E., 1, 29, 170, 173, 316 Fox, C. A., 119, 170, 173 FOX, S. S., 254-280, 320 Frazier, D. T., 361 French, 5. D., 28, 34 Frost, L. L., 175 Fujita, M., 242, 243, 413, 429 Fuller, J. L., 1 Fulton, J. F., 295 Fuse, S., 1, 29, 180 Fuster, J. M., 242 Galambos, R., 210 Gangloff, H., 183, 186, 413
478
AUTHOR INDEX
Gantt, W., Horsley, 210 Gardner, K. W., 329 Gary, T. M.,2 Gasanov, G. G., 296 Gassmann, F., 111 Gastaut, H., 180, 246,462 Gauthier, C., 423 Gaza, N. K., 295 Geller, I., 176 Gellhorn, E., 63,64,360,413,466 Gerard, R. W., 254, 270,466 Gergen, J. A., 442-461 Gerschenfield, H. M.,26 Gerstein, G. L., 12,452 Giussiani, A., 203, 204 Gloor, P., 63, 64, 85, 173, 384, 452 Gogolslk, G., 218, 356 Gogolof, G., 413 Goldberg, J. M.,27 Goldring, S., 429 Goldstein, A. C., 76 Goldstein, M. H., Jr., 422 Goodfellow, E. F., 132 Goodman, L. S., 366 Gorski, R. A., 69 Gorten, R. A., 201 Grastyh, E., 114,218-223,229,235,274, 413,425,445 Green, J. D., 1,29, 63, 70, 85, 176, 190, 218,232,254,255,272,290,413415, 429,433,442,445 Groot, J., 1, 29 Grossman, S., 213 Grossman, S. P., 26 Gulyaeva, L. N., 295 Gutman, J., 186, 188 Gwbtdt, B., 103, 113 Gygax, P. A., 451 Haberland, K., 445 Haffner, F., 365 Hagiwara, S., 27,28 Halirsz, B., 25,26, 170 Hall, E. A., 170 Hamlyn, M. H., 26,401 Hammel, H. T., 63 Hammerstein, J., 91 Hanai, T., 2,27, 171 Hanking, B. H., 71 Hardy, J. D., 63 Harmony, T., 62 Harris, G. W., 69 Harrison, J. M.,205, 433 Hartline, H. K., 25, 318 Harvey, J. A.. 204,213 Hasegawa, I., 290 Hayward, J. N., 91, 94 Heath, R. G., 205,235 Heller, A., 204
Henatsch, H.- D.. 384 Hendrix, C. E., 218,228,229, 393 Hemhdez-Pebn, R., 28,246 Herzet, J. P., 395 Hess, J., Jr., 46 Hess, W. R., 34,46, 62, 103, 105, 117, 118, 129,203,435 Hetherington, A. W., 1 Heuser, G., 378 Heymans, C., 466 Hiebel, W., 26 Had, G., 46 Hilliard, J., 91,94 Hilton, S. M.,111, 119, 122, 180 Himwich, H. E., 338 Hirose, K., 365-387 Hisaw, F. L., 81 Hodes, R., 62 Hoebel, B. G., 132 Hohlweg, W., 69 Holines, J. E., 223, 425 Holmqvist, B., 402 Hori, Y., 290 Horvath, F. E., 176 Hosli, L., 46 Hugelin, A., 28,462,464,473 Hugh Dingle, R D., 254-280 Humphrey, T., 85 Hunter, J., 114 Hullay, J., 445 Hunsperger, R. W., 65, 103-127, 129,177, 180 Hunt, C. C., 27 Hunt, H. F., 204,213 Hutt, P.J., 176 Huttenlocher, P. R., 28 Hyde, J. E., 96, 180 Hydh, H., 241 Iki, M.,2-4,12,25,27 Ilyutechenok, R. I., 359 Imamura, G., 295, 356 Ingvar, D. H., 384 Iosif, G., 290,424 Isaacson, R. L., 309,313, 316, 333,458 Ito, M., 290 Itoigawa, N., 281-292 Iwamura, Y., 377,380,382, 383 Iwata, K., 254, 272, 413 Jackson, D. C., 63 Sansen, J., 258,273 Jansen, J., Jr., 114, 119, 174 Jarrard, L. O., 307, 309,458 Jasper, H. H., 2, 114, 190,229,239,241, 338, 357 Johansson, B., 203,204 John, E. R., 388,397 Johnston, J. B., 119 Jouvet, D., 204
AUTHOR I N D E X
Jouvet, M., 28, 204, 383, 425 Julou, L., 384 Jung, R., 28,229,413,462,464,472 Junkmann, K., 69 Kaada, B. R., 62, 114, 119, 122, 173, 174, 180,251, 295, 296, 300, 316, 378 Kabat, H., 203, 204 Kado, R. T., 235, 237, 239,241,445 Kaelber, W. W., 176 Kaitor, F., 445 Kaji, S., 2, 27, 171 Kamikawa, K., 228 Kamikawa, Y., 3 Kanai, T., 357 Kandel, E. R., 404,411,452 Karli, P., 122 Karmos, G., 219, 220,425 Karplus, 5. P., 116, 189 Kasamatsu, A., 464 Kato, S., 295 Kasamatsu, A., 464 Kato, S., 295 Kawakami, M., 69-102,413 Kawamura, H., 204,295, 356, 377, 382, 383, 413 Kellhyi, L., 219, 425 Keilicutt, M. H., 316 Kennard, M., 295 Kerkut, G. A., 26 Kiang, N. Y.-S., 422 Kiang, N. Y.-S., 12, 452 Kido, R., 365-387 Killam, K. F., 388-399 Kimble, D. P., 316, 325, 333,458 Kimura, D., 316,458 Kimura, K., 2 4 , 12, 25,27 King, F. A., 433 King Killam, E., 388-399 Kite, W. C., 300 Kjaerheim, A., 401 Kleitman, N., 420 Kletzkin, M., 124, 129, 130 Kling, A., 176, 177 Klingberg, F., 218-227 Knapp, D. A., 359 Kobayashi, N., 1-33 Koella, W. P., 271, 274 Koelle, G., 337 Koepke, J. E., 327 Kogi, K., 295 Koikegami, H., 1,29,48, 62, 63, 71, 85, 122, 173, 174, 180 Koizumi, K., 359 Koketsu, K., 27 Kolmodin, G. M., 27 Kolmodin, G. M., 473 Kooi, K. A., 246 Korhyi, L., 245, 247
Kornmuller, A. E., 229,413 Kosman, A. J., 1,29, 176 Kotljar, B. I., 218, 223 Kreidl, A., 116, 189 Kreindler, A., 173, 175, 180 Kremer, W. F., 295 Krnjevie, K., 357, 361 Kruger, L., 318 Krupp, P., 46 Kubota, K., 380, 384 Kuehn, A., 382,384 Kuniyashi, M., 2-4, 12, 25, 27 Kuno, M., 27 Kurotsu, T., 203 Kushiro, H., 180 Kveim, O., 316 LaGrutta, G., 359, 424 Lammers, H. J., 119, 124, 180,462 Lauer, E. W., 85 Lechi, A., 176 Lena, C., 433,434 Levison, P. K., 176 Levy, C. K., 271 Lewis, P. R., 357, 359 Liberson, W. T.,413 Libeskind, J. C., 254-280 Lim, R. K. S., 340 Lilly, J. C., 63 Lindgren, P., 203 Lindsley, D. B., 34, 322, 235 Lisk, R D., 69 Lisshk, K., 114, 218, 219, 222, 223, 235, 246253,413,425 Liu, C. N., 340 Livingston, R.B., 28,245, 274,451 Lloyd, D. P. C., 12, 27 Lobanova, L. V., 293 Loeb, C., 368,425 Loeser, J. D., 271 L0m0, T., 400-412 Long, J. P., 361 Longo, V. G., 223,338 Lbpez-Mendoza, E., 246 Lorente, de N6, R., 401, 442 Losonczy, H. V.,220 Layning, Y.,404,452 Lux, H. D., 242 Lyon, M., 205,433 Macchi, G., 176 MacGillivray, B., 241 Machne, X.,254,413,429 Macht, M. B., 76 MacIntosh, F. C., 357, 361 MacLardy, T., 462 MacLean, R. D., 65, 114, 119, 124, 180, 201-204,295, 318,433,435,422-446, 459,462
479
480
AUT H O R I N D E X
Madarasz, T., 218,219, 222, 223,235,274 413,425 Maeno, S., 290 Maeno, T., 2 4 , 12, 25,27 Magni, F.,360 Magnus, J., 26, 28, 170 Magnus, O., 119, 124, 180 Magoun, H. W., 34,44,62,203,204,246,356 Mahut, A., 311 Mahut, M., 425 Mancia, M., 430 Manfredi, M., 425 Manning, J. W., 203 Mantegazzhi, P., 338, 357 Manzoni, T., 425,428432,457 Marshall, W. H., 254, 270 Martin, J., 425 Maruyama, N., 2,27, 171 Mason, J. W., 96, 97 Matsushita, A., 365-387 Matsumoto, J., 290 Matthews, B. H. C., 28 Maxwell, D. S., 413,429, 452 Mayer, Ch., 413 McCleary, R. A., 316, 318 McIlwain, J. T., 228 McIver, A. H., 433 McLardy, T., 226 McQueeney, J. A., 458 Megawa, A., 203 Mernpel, E., 316 Mering, T. A., 340 Merrick, A. J., 340 Mess, B., 25,26, 170 Metz, B., 361 Michael, R. P., 69 Michelson, M. J., 359 Miller, G. A., 318 Miller, H. R., 176, 205 Miller, N. E., 1, 128 Milner, B., 318 Milner, P., 205 Mir, D., 54 Mishkin, M., 228, 316 Mitchell, J. F.,357 Mitsuyasu, K., 384 Miyamoto, K., 281-292 Mochida, Y.,174 Moffitt, R. I., 340 Moiseeva, N. A., 296 Mollica, A., 359 Monnier, M.,46, 183, 186, 413 Monroe, R. R., 235 Moore, R. Y.,204,458 Morgan, C. T., 209 Morgane, P. J., 1, 2, 29, 176 Morin, F.,124, 254, 255, 429 Morin, G., 180 Morison, R. S., 34
Morrell, F., 246 Moruzzi, G., 28, 34,44,46,246, 356,430 Mountcastle, V. B., 6, 11, 176, 318 Musyaschikova, S. S., 295 Nacimknto, A., 242 Naka, F., 1-33 Nakamura, Y.,382,413 Nakao, H., 118,128-143, 176, 177, 360 Naquet, R., 62, 64,119, 180 Nauta, W. J. H., 2, 29, 85, 119, 203,205, 246, 295,433,451 Negishi, K., 29 Nicholson, A. N., 359 Nielson, H. M., 433 Niimi, Y.,380 Nilges, R. E., 462 Niemer, W. T., 132,339 Niki, H., 305-317, 383, 458 Nobel, K., 322 Norris, Jr., F. H., 180 Novikova, L. A., 413
bberg, B., 293,204 OBrien, J. H.,254-280 ODoherty, D. S., 48
O'Flaherty, J. J., 194, 201, 203 Okuda, O., 28 Okuma, T., 378 Olds, J., 132,144-164, 205 O'Leary, J. L., 124,430 Ommaya, A. K., 229 Ornukai, F., 173 Ono, T., I-33 Oomura, Y.,1-33 Ooyama, H., 1-33 Oppenheimer, M. J., 176, 205 Orbach, J., 176, 177
Palesthi, M., 425 Papez, J. W., 295 Parkes, A. S.,76 Parma, M., 423 Parmeggiani, P. L., 124, 4 1 3 4 1 Passouant, P., 413, 462475 Passouant-Fontaine, Th., 413, 471, 474 Paton, D. W. M., 356 Paul-David, J., 360 Peele, T. L., 71, 119, 174, 176, 180 Penfield, W. G., 318 Pepeu, G., 357 Petrh, M., 166 Petsche, H., 218,290, 356,413,429 Phillis, J. W., 357 Pickenhain, L., 218-227 Pillat, B., 413 Pisano, M., 425 Ploog, D. W., 65,203,204,445 Poggio, G. F.,12
AUTHOR INDEX
Pompeiano, O., 28,46, 380,413, 429 Porter, R. W., 3, 171, 228, 229, 231, 233, 234 Pressman, G. L., 329 Preston, J. B., 382 Pribram, K., 1,228, 252, 296, 300, 318-336 Pternitis, C., 462475 Quastel, D. M., 473 Quastel, J. H., 473 Rabini, C., 419, 425 RadulovaEki, M., 229,230,235,236,242,425 Rall, W., 26 Ramey, E. R., 48 Randall, L. D., 382 RandiC, M., 359 Ranson, S. W., 1, 203, 204 Rasmussen, E. W., 316 Ratliff, F., 25 Redding, F., 254, 274 Reite, M. L., 445 Rhodes, I. M., 445 Rice, B. F., 91 Riehl, R.-L., 360 Rimbaud, L., 413, 473 Rinaldi, F., 338 Roberts, S., 71 Roberts, W. W., 128, 313, 333 Robinson, B. W., 49,203 Robinson, F., 413 Roldan, E., 151 Rosadini, G., 425 Rossi, G. F., 360,425 Rossi, R. F., 203, 204 Rosvold, H. E., 1, 49, 316 Roth, L. J., 204 Rumbaugh, D. M., 458 Rushmer, R. F., 201 Sacco, G., 425 Sadowski, B., 223 Sagawa, Y., 384 Sailer, S., 219, 226 Sakai, A., 203 Sakai, Y.,384 Salmoiraghi, G. C.. 359 Salvatorelli, G., 419, 426 Sano, T., 119 Sato, T., 242, 243, 413, 429 Sevillano, M., 134 Saul, C. J., 254, 270 Savard, K., 91 Sawa, M., 2, 27, 171 Sawada, M., 1, 26,27 Sawyer, C . H., 69, 70, 76, 91, 94, 190 Schallek, W., 382, 384 Schindler, W. J., 237, 241, 413, 429, 452 Schlag, J. D., 356 Scholz, W., 462
48 1
Schottelius, B., 175 Scholz, W., 462 Schottelius, B., 175 Schreiner, L., 176 Schueler, F. W., 361 Schwartz, N. B., 176, 177 Schwarz, H. G., 124 Schwarzbaum, J. S., 316, 329 Scott, P. P., 69 Sears, T. A., 320 Segundo, J. P., 254, 274,457 Seiden, L. S., 398 Sekiguchi, N., 71 Seto, K., 69-102,71 Sharma, K. W., 2,25 She&', C. N., 71, 119, 174, 176, 180 Sheer, D. E., 48 Shimazu, H., 380 Shirnokochi, M., 290, 291 Showers, M. J. C., 62 Shute, C. C. D., 357, 359 Sidman, M., 367 Siegel, S., 305 Siegfried, J., 254, 274 Silver, A,, 357, 359 Silvestrini, B., 338 Singh, B., 2,25 Skoglund, C. R., 27 Skoglund, C. R., 473 Skultely, F. M., 2 Skultety, F. M., 119 Slater, L., 49 Slusher, M. A., 96, 180 Smith, D. L., 365 Smith, F. D., 27 Smith, 0.A., 1 Smith, 0.A., Jr., 201 Smith, W. K., 295,296 Snider, R. S., 254,257,258,270-272, 339,413 Sokolov, E. N., 226, 235, 324 Sotnichenko, T. S., 293-304 Sovetov, A. N., 295 Spencer, W. A., 404,411,425 Sperti, L., 452 Spiegel, A. E., 12!, 129, 130 Spiegel, E. A., 176, 187, 205 Spielmeyer, W., 462 Spinelli, D. N., 320 Sprague, J. M., 2 Stacey, M., 242 Stafford, A., 469 Stamm, J. S., 252 Stark, L., 62 Stefanis, C., 359 Steg, G., 384 Steiner, F. A., 359 Stellar, E., 2 Steriade, M., 173, 175, 180, 423, 424 Sterman, M. B., 3 4 4 7
482
AUTHOR INDEX
Stevenson, J. A. F., 1 Stolwijk, J., 63 Stoupel, N., 423,424 Stowell, A., 254,257,258,270,271 Straughan, D. W., 359 Stremme, S. B., 63 Strumwasser, F.,26 Stuart, D. G., 3, 171 Stumpf, Ch., 218, 219, 226,290, 356,413 429,452 Sugar, O., 466 Summers, T. B., 176 Sunderland, S., 429 Sutin, J., 2, 21, 27, 28, 171 Swett, J. E., 46 Swett, J. E., 380 Swinyard, E. A., 366 Szabo, I., 219 Szabo, Th., 271,272 Szekely, E. G., 124, 129, 130 Szentitgothai, J., 25, 26, 170 Szerb, J. C., 357, 361 Takahashi, H., 174 Tasaki, I., 26, 27 Tauc, L., 26 Teitelbaum, H., 311 Teitelbaum, P., 132 Terasawa, E.,69-102 Thomas, J., 466 Thompson, J. B.,316, 324 Timo-Iaria, C., 199 Tokizane, T., 295, 356, 377, 378, 380, 383, 384,413 Tolmasskaya, E. S., 295 Torii, S., 201, 204, 295, 360,413, 416 Tower, S. S.,251 T o m e , J. C., 176, 177 Tracy, W. H., 205 Trembly, B., 205 Trendelenburg, U., 360 Tsubokawa, T., 2,21,27,28, 171 Tsuchinashi, S., 76-79 Udalova, G. P., 296 Uemura, T., 76 Udvarhelyl, G. B., 132 Umezu, M., 71 Unna, K. R., 360 Ursin, H., 119, 122, 173, 176, 180 Ushikoshi, I., 71 Usui, K., 71 Uvniis, B., 203 Valatx, J. L., 204 Valdman, A. V., 359 Valenstein, E. S., 124 Valverde, F., 170 Van Reeth, P. Ch., 423,424
Van Zwieten, P. A., 413 Vasilevskaya, N. E.,296 Vaz-Ferreira, A., 464 Vera, C. L., 452 Vereby, Gy., 114 Vereczekey, L., 220,425 Vergnes, M., 122 Verly, W. G., 91 Veneano, M., 29 Viernstein, L. J., 12 Villarreal, J. E., 359 Vinagradova, 0. S., 235 Von Bekesy, G., 318 Voorhoeve, P. E., 402 Voronin, L. G., 218,223 Votaw, C. L., 274 Wagner, H. G., 25 Wagner, H. O., 318 Wagner, J. W., 26 Walker, A. E., 124, 132 Wallenberg, A., 124 Walsh, J., 425 Walter, D. O., 228, 232, 234, 235, 237, 241, 242, 393 Ward, A., 295 Warden, C. J., 367 Warner, H., 49 Wasman, M., 165 Watanabe, H., 1,29 Watanabe, T., 1,29 Weatherall, J. A. C., 469 Weber, M., 111 Wendt, R. H., 171 Weill, A., 169 Weiss, T.,151, 239 Werner, H. G., 25 White, L. E., 301, 360 Whiteside, J. A., 272 Wickelgren, W. O., 309, 316,458 Wikler, A., 360, 382 Windle, W. F., 469 Winter, C. A., 365 Wolstencroft, J. H., 359 Wood, D. C.,1 Wood, D. C., 119, 175, 176, 180 Wyers, E. J., 378 Wyrwicka, W., 43 Wyss, 0. A. M., 103, 111,435 Yamada, T., 71 Yamaguchi, Y., 281-292 Yamamoto, K.-I.,337-36#, 365-387 Yamamoto, T., 1-33 Yamanaka, K., 76,79 Yasukochi, O., 118 Yokoyama, T., 1,29 Yoshida, K., 69-102, 122 Yoshida, M., 134, 173, 180
AUTHOR INDEX
Yoshioka, M., 382 Yoshii, H., 290 Zachar, J., 166 Zanchetti, A., 423
1
Zanocco, G., 413,424,433,435 Zaraiskaia, S. M., Zbroiyna, A., 111, 119, 122, 180 Zubkova, N. A., 294
483
484
Subject Index Acetylcholinelevel, and cholinergic agents, 354-357 and hemicholinium, 354-356 Affective behavior, and brain stem, electrical stimulation, 103-125 and forebrain, electrical stimulation, 103-125 mewing, and brain level, 114, 115, 122-124 zones, 122-124 pain reactions, 115, 116, 124 Alimentary conditioned reflex, and amygdala, 248-251 and hippocampus, 248-251 and reticular formation, 248-251 Amygdala, ablation, and aggressive behavior, 165-180 affective behavior, zones, 119-122 and cortex, evoked responses, 269-272 EEG activity, and ovulation, 81-83, 98, 99 electrical stimulation, and aggression, latency, 168-172 and conditioned reflex, 248-251 estrogen administration, and ARC potential, 79, 80 function, and ovulation, 71-74 and hypothalamus, functional interaction, 16-21, 29 lesions, and attack behavior, 171-175, 177-179 and hyperphagia, 1 , 2 and orienting, regulation, 325-329 progesterone administration, and ARC potential, 79, 80 Amygdala, radio stimulation, responses, 60, 61, 64 stimulation, and affective behavior, 119, 120 and aggressive behavior, 165-180 and attack behavior, 171-177 and estrogen formation, 88-93, 99 and ovulation, 87-91 and pain pattern, 116 and progesterone formation, 88-93,99
Anesthesia, level, and blood pressure, conditioned response, 205-210 Anticholinergic drugs, and behavior, 393-395. Aphagia, and hypothalamus, lesions, 1, 2 Arousal, and hippocampus, theta-rhythm, 415419 Attack behavior, suppression, and amygdala, lesions, 171-175, 177-179 Attention, cerebral processes, and hippocampus, 242,243 and hippocampus, 228-243 Atropine, and behavior, 393-395 Auditory responses, and hippocampus, theta-rhythm, 422427 Avoidance, conditionedreflex, and limbic system, 247,248 Awake-sleep cycle, and cholinergic agonists-antagonists,341-35 1 and cholinergic antagonists, 340, 341 Baroceptor reflex, and septa1 area, stimulation, 195, 196, 203 Behavior, affective -, and brain stem, stimulation, 103-125 and forebrain, stimulation, 103-125 aggressive -, and amygdala, ablation, 165-180 and amygdala, stimulation, 165-180 and drugs, 365-385 scoring sheet, 367 and sinomenhe, 365-385 conditioned reflex, and limbic system, 247,248 and SHR, 221-225 discriminative -, and EEG, patterns, 235,237 and hippocampus, EEG activity, 235-239 and impedance responses, 238-242 efferent control, and limbic system, 318-335 free -, and limbic system, 48-66 and hippocampus,
SUBJECT I N D E X
EEG activity, 229-231 electrical stimulation, 435-437 frequency potentiation, 400, 401, 410 hypoxia, 466, 467 theta-rhythm, suppression, 433-435 inhibition, and sleep induction, 34-47 mechanisms, coordination, 64 and hippocampus, slow-wave activity, 218227 and limbic system, 295 microphysiology, 48-50 orientating reaction, and SHR, 219-222 orienting -, and EEG patterns, 235, 237 and hallucinogenic agents, 235 and psychomimetic drugs, 234-236 reinforcement, and limbic system, 144-163 classical conditioning, 151-153 correlated behavior, 161-163 covariation, 153, 154, 156 histological findings, 149 interspike intervals, 156-158 operant conditioning, 153-155 special tests, 157-161 spike patterns, 149, 150 variability, 15&152 responses, and brain structure, 4 0 W 1 2 and hippocampus, theta-rhythm, 445 sequential -, and limbic system, stimulation, 65 triggering area, specificity, 64, 65 sensory stimulation, 128, 129 and SHR, 218-220 switch-off -, and amygdala, after-discharges, 134, 135, 139, 140, 142 and cingulate, after-discharges, 137, 139 facilitation, 128-142 and hippocampus, after-discharge, 133, 134 139, 141 and hypothalamus, after-discharge, 138, 139 and limbic after-discharge, 132-139 stimulation, 129-132 inhibition, 128-142 mesencephalon, and limbic after-discharge, 139-141 stimulation, 129-132 trigger mechanisms, and hypothalamus, 64 Blood pressure, conditioned response, level of anesthesia, 205-210 and subcortical structures, 210
485
fall, conditioning, and septal area, stimulation, 205-210 and septal area, stimulation, 189-205 Blood sugar, and limbic system, 296 Brain, septal area, and behavioral functions, 189-215 lesion, and sodium chloride intake, 21&215 and water intake, 210-215 and neurovegetative function, 189-215 stimulation, and baroceptor reflex, 195, 196,203 and blood pressure, 189-205 fall, conditioning, 205-210 bradycardia, 203,204 hypertensive effects, 196, 197, 204 localization, 197-201 prolonged duration, 204 and respiration, 189-205 visceral cortex, interoceptive conditioned reflexes, 293, 294 Brain stem, electrical stimulation, and affective behavior, 103-125 and flight pattern, 105 flight zones, 116122 threat zones, 116-122 Bardycardia, and septal area, stimulation, 203, 204 Cardiac response, and sound stimulation, 314, 315 Cerebellum, activity, and hippocampus, 254-279 and limbic system, 254-279 click stimulation, response, 258,275, evoked responses, latency, 257, 258, 263 and hippocampus, 265-269 and light flash, 267-269,273 and spike latency, 276-279 and hippocampus, stimulation, 261-265, 272, 273 latency, 274-279 light stimulation, responses, 256-274 sensory interaction, mechanisms, 274-279 stimulation, intramodal interaction, 258-261, 273 Cholinergic agonists-antagonists, and awake-sleep cycle, 340-351 Cholinergic mechanisms, and limbic system, 337-362 and neocortex, 337-362 Conditioned reflex, blood pressure, and subcortical structures, 210
486
SUBJECT I N D E X
Cortex, evoked responses, and amygdala, 269-272 and hippocampus, 265-269 Discrimination, rate, and hippocampus, 307,308 reversal, and hippocampus, lesions, 357459 Feeding, and hypothalamus, activity, 1-33 dual function, 1 neuronal mechanisms, 1-33 Feeding center, and hypothalamus, lateral area, 1,2 Flight behavior, active areas, 116-122 adaptation, and environmental conditions, 112-114 and cardiac activity, 106-1 11 and postural tonus, 106-1 11 Flight pattern, and brain level, 105, 111 Flight responses, and amygdala, zones, 119-122 Food intake, and hypothalamus, lesions, 1, 2 Forebrain, electrical stimulation, and affective behavior, 103-125 and mewing, 114,122,123 and threat pattern, 103-105 flight zones, 116-122 threat zones, 116-122 Fornix, radio stimulation, evoked effects, 57 responses, 57-60 stimulation, responses, central integration, 65 Frequency, potentiation, and hippocampus, 400-41 1 cell discharges, 407-409 mechanism, 404,405 recurrent inhibition, 411 time course, 4 0 5 4 7 Habituation, and hippocampus, 329-333 Hemicholinium, and acetylcholine levels, 354 and limbic system, 351-354 and neocortex, 351-354 Hippocampus, ablation, and alternation learning, 311
and cardiac response, 314, 315 and discrimination rate, 307, 308 and disinhibition, 307 and learning, 305-3 16 flexibility, 312 and performance, 308, 309 and response alternation, 312, 313 activity, and labyrinth, stimulation, 193-188 after-discharge, and chlorpromazine, 186 and labyrinth, stimulation, 185-187 and spontaneous acitivity, 442-446 and attention, 228-243 cerebral processes, 242,243 cell discharges, and frequency potentiation, 407409 and cerebellum activity, 254-279 evoked responses, 265-269,273 and cortical activity, 254-279 discharges, and labyrinth, stimulation, 185 EEG activity, and behavior, 229-231 and discriminative behavior, 235,237 and estrus cycle, 76-81 and hallucinogenic agents, 235 impedance responses, 238-242 and orientating behavior, 235, 237 and ovulation, 81-83,98, 99 and performance capability, 232 and psychomimetic agents, 234-236 and seizure discharge, 232 and sex hormones, 71-76 slow-wave -, 218-227 and subthalamus, lesions, 232-234 electrical activity, regional aspects, 229-234 electrical stimulation, behavioral effects, 435-437 and conditioned reflex, 248-251 and estrus cycle, ARC potentials, 78-80 frequency potentiation, 400411 and behavior, 400,401,410 and learning, 400,410 mechanism, 404,405 pathways, 402,403 recordings, 401,402 and recurrent inhibition, 411 synaptic activity, 403,404 time course, 405407 function, and endocrine regulation, 72-14 properties, 442460 and habituation, 329-333 hypoxia, and behavior, 466,467 and lesions, histology, 468
SUBJECT INDEX
neonatal period, 469471 impedance measurement, and learning, 237-239 isohippocampal rhythm, see IHR and learning, 228-243 cerebral processes, 242, 243 lesion, and discrimination, reversal, 457-459 and hypoxia, 462-474 and learning set, performance, 457459 potential responses, and single unit activities, 4 4 W 5 2 and reticular formation, interaction, 43-32, 436
slow rhythm, and behavior, 218-220 and orienting reaction, 219-222 slow-wave, activity, and behavior, 218-227 stimulation, and amygdala, activity, 81, 85, 99 and estrogen formation, 84-87, 92, 93 and ovulation, 84-86, 99 photic responses, 447 and progesterone formation, 84-87, 92, 93, 99
projection pathways, 450,451 subcortical effects, 452457 theta-rhythm, and arousal, 415419 and auditory responses, 422427 behavioral responses, 445 functional significance, 413-437 and hypoxia, 465, 466 and neocortex, 419-425 and reticular formation, 41W19, 436 suppression, and activated sleep, 433-435 behavioral effects, 433435 and thalamus, 425430 theta-waves, and cortex, 281-291 and IHR, 282-290 HVM activity, and anesthesia, 3-6, 27-29 anesthesia, and discharge pattern, 4-6,27-29 electrical stimulation and SUDs, 14-25 latency, 18-24 and feeding center, 1, 2 and LH activity, 3-16 crosscorrelation, 6-14$ 28, 29 SUDs (spontaneous unitary discharges), and anesthesia, depth, 3-6 Hyperphagia, and amygdala, lesions, 1, 2 and hypothalamus, lesions, 1, 2
487
Hypotension, conditioned -, and septa1 area, stimulation, 205-210 Hypothalamus, and amygdala, functional interaction, 16-21, 29 electrical stimulation, and conditioning responses, 56, 57, 62 evoked effects, 55, 56 and pupillary response, 52-57 EEG activity, and estrus cycle, 76-81 lesions, and aphagia, 1, 2 and hyperphagia, 1,2 radio stimulation, and pupillary response, 52-57 and sleep induction, 35, 37 stimulation, and attack behavior, latency, 172 Hypoxia, and hippocampus, reactions, 462474 theta-rhythm, 465466 IHR and brain stem, stimulation, 282 cortical -, EEG, frequency analysis, 282, 283 general features, 282-284 and hippocampus, synchrony, 282-284 induction, 284, 285, 290 cortical distribution, 284, 285 and cortical excitability, 295-297 and hippocampus, theta-waves, 282-290 Inhibition, collatera type, and limbic system, 319-323 loss, and hippocampus, ablation, 301 neural -, efferent control, and limbic system, 318-335 recurrent type, and limbic system, 319-324 Labyrinth, stimulation, and hippocampus, activity, 183-188 after-discharge, 185-1 87 discharges, 185 and optic nystagmus, 183, 184 Learning, cerebral processes, and hippocampus, 242,243 and hippocampus, 228-243 frequency potentiation, 400,410 impedance measurements, 237-239 studies, 305-316
488
SUBJECT INDEX
Learning set, performance, and hippocampus, lesion, 457-459 LH, activity, and anesthesia, 3-6,27-29 anesthesia, and discharge pattern, 4-6,27-29 electrical stimulation, and SUDS, 14-25 latency, 18-24 and feeding center, 1 , 2 and HVM, activity, 3-16 crosscorrelation, 6-14, 28, 29 SUDS, and anesthesia, depth, 3-6 Limbic system, activation, and neurovegetative conditioning, 205 and avoidance, conditioned reflex, 247, 248 and behavior, conditioned reflex, 247, 248 efferent control, 318-335 free -, 48-66 mechanisms, 295 recordings, 50-52 and reticular formation, 295 and behavioral functions, 189-215 and behavioral reinforcement, 144-163 classical conditioning, 151-153 correlated behaviors, 161-163 covariation, 153-156 histological findings, 149 interspike intervals, 156458 operant conditioning, 153-1 55 special tests, 157-161 spike patterns, 149, 150 studies, 144-148 variability, 150-152 and blood sugar, studies, 296 and cerebellar activity, 254-279 cholinergic mechanism, 337-362 and conditioning motivation, 246-252 cortex, conditions, 293-302 functional significance, 295 and respiratory system, 296,297 visceral analyzers, 293-302 and cortical activity, 254-279 EEG, and chemical concentration, 297,298 EEG activity, and estrus cycle, 76-81 and sex hormones, 71-76 electrical activity, and drugs, 365-385 and sinomenine, 365-385 electrical stimulation, experiments, 48,49
and estrus cycle,sARC potential, 76-81 and fiber degeneration, 298-301 function, experiments, 69-71, 98-100 mechanisms, 69-100 neurophysiology, 318-325 gonadal steroids, feedback effect, 71-81 and hemicholinium, 351-354 inhibition, collateral type, 319-325 neural -, efferent control, 318-335 recurrent type, 319-325 and memory, recent -, 246-252 and neurovegetative functions, 189-21 5 and progesterone feedback, 69-100 radio stimulation, and behavior, 5&52,63 stimulation, and estrogen formation, 81-91 and progesterone formation, 81-91 and sequential behavior, 65 Memory, recent -, and limbic system, 246-252 Midbrain, electrical stimulation, and pain reactions, 115, 116 Motivation, conditioning -, and limbic system, 246-252 Neocortex, cholinergic mechanisms, 337-362 and hemicholinium, 351-354 and hippocampus, theta-rhythm, 419-425 Orienting, regulation, and amygdala, 325-329 Ovulation, and amygdala, electrical stimulation, 87-91 and ARC, electrical stimulation, 89-91, 99, 100 and hippocampus, electrical stimulation, 81-91, 99 Pain perception, and brain level, 116, 124 Pain reaction, patterns, 115, 116 threshold answers, and brain level, 124 Performance, curves, and hippocampus, 308, 309 Progesterone, feedback control,
SUBJECT I N D E X
and amygdala, 69-100 and hippocampus, 69-100 and limbic system, 69-100 Psychomimetic drugs, and orienting behavior, 234, 235 Pupillary response, and hypothalamus, stimulation, 52-57, 62 Reserpine, and behavior, 395, 396 Respiration, and limbic system, cortex, 296, 297 and septa1 area, stimulation, 189-205 Reticular activating system, and sleep induction, experiments, 3440, 46 Reticular formation, electrical stimulation, and conditioned reflex, 248-251 and hippocampus, interaction, 430-433,436 theta-rhythm, 4 1 M 1 9 , 436 Rhinencephalon, activity, and drugs, 388-399 patterns, 389-393 sequential averaging, 390-394 studies, 388-399 and behavior, conditional -, 388-399 Satiety center, and hypothalamus, ventromedial nucleus (HVM), 1, 2 SHR, see Hippocampus, slow rhythm and conditioned reflex, 221-225 and orientating reaction, 219, 220, 222 and sensory information, 226 Sinomenine, effects, and CNS, 365-385 Sleep induction,
and behavior, chronic experiments, 38, 41,42 and conditioning experiments, 38, 41, 43 electrophysiological experiments, 34-37 and hypothalamus, 35, 37 mechanisms, and behavioral inhibition, 34-47 ergotropic zone, 34, 35 trophotropic zone, 34, 35 pathways, 4 4 4 7 and reticular activating system, experiments, 34-40,46 Spontaneous unitary discharge, HVM, and anesthesia, 3-6 latency, and HVM, electrical stimulation, 18-24 and LH, electrical stimulation, 18-24 LH, and anesthesia, depth, 3-6 SUD, see spontaneous unitary discharge Thalamus, and hippocampus, theta rhythm, 425-430 Threat behavior, active areas, topography, 116122 adaptation, and environmental conditions, 112-1 14 and cardiac activity, 106111 and postural tonus, 106111 Threat responses, and amygdala, zones, 119-122 Threat pattern, and brain level, 103-105 Visceral analyzers, cortex, localization, 293, 294 limbic system, studies, 295-302 subcortical formations, 295-302
489