PROGRESS I N B R A I N RESEARCH VOLUME 1 BRAIN MECHANISMS
This Page Intentionally Left Blank
PROGRESS I N B R A I N RESEARCH VOLUME 1
BRAIN MECHANISMS INTERNATIONAL COLLOQUIUM SPONSORED B Y T H E INTERNATIONAL BRAIN RESEARCH ORGANISATION ( I B R O )
O N SPECIFIC A N D UNSPECIFIC M E C H A N I S M S OF SENSORY MOTOR INTEGRATION; PISA, 1961
EDITED BY
GIUSEPPE MORUZZI Istituto di Fisiologiu, Universitri di Pisa, Pisu, Iiuliu A L F R E D FESSARD Laborutoire de Neurophysiologie Gknndrale, Collige de France, Puris, Frunce HERBERT H. JASPER Montreal Neurological Institute, Montreal, Canudu
ELSEVIER PUBLISHING COMPANY AMSTERDAM - LONDON - NEW YORK
1963
SOLE D I S T R I B U T O R S F O R T H E U N I T E D S T A T E S A N D C A N A D A A M E R I C A N ELSEVIER P U B L I S H I N G C O M P A N Y , I N C .
52 V A N D E R B I L T A V E N U E ,
NEW YORK
17, N.Y.
SOLE DISTRIBUTORS FOR GREAT BRITAIN 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 1 2 B , R I P P L E S I D E C O M M E R C I A L ESTATE R I P P L E ROAD, B A R K I N G , ESSEX
L I B R A R Y OF C O N G R E S S C A T A L O G C A R D N U M B E R
WITH
246
62-19695
ILLUSTRATIONS A N D 1 TABLE
ALL R I G H T S R E S E R V E D T H I S BOOK OR 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, INCLUDING PHOTOSTATIC OR MICROFILM FORM, W I T H O U T WRITTEN PERMISSION FROM T H E PUBLISHERS
This Page Intentionally Left Blank
The participants in the Pisa Colloquium. Front row: OLeary, Eccles, Moruui, Ah-Fessard, Bremer, Anokhin, Fessard, Cranit, tissik, Adey. Second row: Jung, Grey Walrer, Jasper, Brookhart, Magoun, Rossi, Carreras, Pompeiano. Third row: Jouvet, Hagbarth, Dumont, Hugelin, Buser, Dell, Arduini, Zanchetti, Naquet, Ricci.
Participants to Pisa Colloquium
W. R. ADEY,Department of Anatomy and Brain Research Institute, University of California Medical Center, Los Angeles, Calif., U.S.A.
C . AJMONE MARSAN, Branch of EEG and Clin. Neurophysiology, National Institutes of Health, Bethesda, Md., U.S.A. D. ALBE-FESSARD, Centre d’Etudes de Physiologie Nerveuse et d’Electrophysiologie du C.N.R.S., Paris, France P. ANOKHIN, Academy of Medical Sciences, Moscow, U.S.S.R. A. ARDUINI, Istituto di Fisiologia dell’universita di Pisa, Italy
I. BERITASHVILI, Institute of Physiology, Academy of Science of the Georgian SSR, Tbilisi, U.S.S.R. M. A. B. BRAZIER, Brain Research Institute, University of California Medical Center, Los Angeles, California, U.S.A. F. BREMER,Laboratoire de Pathologie GCnCrale, UniversitC Libre de Bruxelles, Belgium
J. M. BROOKHART, Dzpartment of Physiology, University of Oregon Medical School, Portland, Oregon, U.S.A.
P. BUSER,Laboratoire de Neurophysiologie ComparCe, Faculte des Sciences, Universit6 de Paris, France M. CARRERAS, Clinica delle Malattie Nervose dell’universith di Parma, Italy
P. DELL,Laboratoire de Neurophysiologie, HBpital Henri Rousselle, Paris, France J. C. ECCLES,Department of Physiology, The Australian National University, Canb-rra, Australia
E. FADIGA,Istituto di Fisiologia Umana, Universith di Bologna, Italy A. FESSARD, Laboratoire de Neurophysiologie GCnerale, Collbge de France, Paris, France
H. GASTAUT, Unite de Recherches Neurobiologiques de I’lnstitut National d’Hygikne, Marseille, France
VIII
PARTICIPANTS TO PISA COLLOQUIUM
R. GRANIT,The Nobel Institute for Neurophysiology, Karolinska Institutet, Stockholm, Sweden K.-E. HAGBARTH, Department of Clinical Neurophysiology, Akademiska Sjukhuset, Uppsala, Sweden A. HUGELIN,Laboratoire de Neurophysiologie, H6pital Henri Rousselle, Paris, France
H. H. JASPER, Montreal Neurological Institute, McGill University, Montreal, Canada M. JOUVET, Laboratoire de Physiologie, Facultt de Mtdicine, Universitt de Lyon, France R. JUNG,Abteilung fur klinische Neurophysiologie der Universitat, Freiburg i/Br., Germany K. LISSAK,Physiological Institute of the University of PCcs, Hungary H. W. MAGOUN, Graduate Division, University of California, Los Angeles, 24, Calif., U.S.A. G. MORUZZI, Istituto di Fisiologia dell’Universit8 di Pisa, Italy R. NAQUET, Laboratoire EEG, H6pital la Timone, Marseille, France S. NARIKASHVILI, Institute of Physiology, Academy of Sciences of the Georgian SSR, Tbilisi, U.S.S.R. Departments of Neurology and Neurosurgery, Washington University J. L.O’LEARY, School of Medicine, Saint Louis, Mo., U.S.A. 0. POMPEIANO, Istituto di Fisiologia dell’Universit8 di Pisa, Italy G. F. RICCI,Istituto di Farmacologia dell’universita di Pisa, Italy G. F. ROW, Clinica Neurochirurgica dell’Universit8 di Genova, Italy S. TYC-DUMONT, Laboratoire de Neurophysiologie, Hapita1 Henri Rousselle, Paris, France W. GREYWALTER, Burden Neurological Institute, Bristol, Great Britain A. ZANCHETTI, Istituto di Patologia Speciale Medica dell’Universit8 di Siena, Italy
Introductory Remarks on Behalf of IBRO and UNESCO HERBERT H. JASPER
It is with great pleasure that I have accepted the invitation of Professor Moruzzi to speak on behalf of the International Brain Research Organization at the inaugural ceremonies of this Colloquium. In so doing, I am merely the spokesman for the many scientists from varied disciplines and special training in many countries who have formed a world community of colleagues with common interests in the brain sciences. We have been working together and communicating more frequently with each other during recent years. The formation of the International Brain Research Organization in October 1960, less than one year ago, is only the formalization of a growing body of scientific workers determined to promote and improve the quality of basic research on the brain independent of political barriers which separate us, often by chance, into different countries. We are also determined to do what we can to improve the working relations between our various countries by our demonstration of cordial and effective collaboration in spite of political situations which would seem to place us in conflict one with another. But in keeping with the traditions of the first international scientific colloquium held in Pisa in October of 1839, 122 years ago, we will abide by the wishes of the Grand Duke of Tuscany and try to refrain from political discussions during our formal meetings - though we cannot make any promises for many informal sessions which are often the best part of such colloquia. I take pleasure also in bringing greetings and salutations from UNESCO to this first colloquium sponsored by their very young offspring, IBRO, which is actually less than a year old. I must say that Unesco is somewhat astounded by the vigor of their young child, and perhaps fearful at times that we are trying to run before we have learned to walk, but the splendid manner in which Prof. Moruzzi and his co-workers have organized this colloquium with the generous assistance of the Valentino Baldacci Foundation - and the eminent scientists gathered here from near and far - should reassure them that we can not only walk, but we can run and even fly. We are particularly grateful to Prof. Ugo Baldacci who has made this conference possible - even before IBRO was formally organized, and certainly before we have become sufficiently well established financially to undertake such a meeting. We are pleased to be able to pay tribute in this manner to his distinguished father, Dr. Valentino Baldacci of Pisa. This is one of a series of colloquia started by the Laurentian symposium on Bruin
X
INTRODUCTORY REMARKS ON BEHALF OF IBRO AND UNESCO
mechanisms and consciousness held in 1953 (Blackwell, Oxford, 1954). This was followed by the Detroit symposium Reticular formation of the brain (Little, Brown and Co., Boston, 1958), the CIBA conference on Neurological basis of beliaviour (Churchill, London, 1958), and the Moscow colloquium on Electroencephalography of higher nervous activity (Electroenceph. clin. Neurophysiol., 1960, Suppl. 13). Several additional symposia were held along the same lines, as for example the Montevideo symposium on Brain mechanisms of learning (Blackwell, Oxford, 1961). 1 would like to take this occasion to pay tribute to one of our members who has provided much of the initiative and inspiration for many of these symposia, including the first held in 1953 in Canada, that is Professor Henri Gastaut. He has worked in the background for many of these important meetings, and he deserves more credit than is usually given him in the publication of their proceedings. The proceedings of this colloquium are to be published in extenso for the benefit of our many colleagues who are unable to be with us. The Baldacci Foundation is to publish the French edition, while IBRO will publish the English edition, in keeping with our policy to publish in the two working languages of UNESCO.
Introductory Remarks by the Honorary President PROFESSOR FREDERIC BREMER
M y dear Colleagues, I owe to the date of my birthday the privilege and pleasure of expressing on your behalf our gratitude to all who have made this Colloquium such a pleasant reality. We are all especially grateful to the University of Pisa for the hospitality it has offered us at the Istituto di Fisiologia, and for the interest expressed in our work by the presence at this inaugural session of Professor Faedo, Rector of the University, and Professor Puccinelli, Dean of its Medical Faculty. You will all wish me, I am sure, to ask them to accept our sincere thanks for their kind attendance. Our thanks are due also to the International Brain Research Organization, who are the sponsors of the Colloquium, and to our colleague Herbert Jasper, the energetic Executive Secretary of this Organization, who has played a majorrole in the preliminary stages of the excellent arrangements made for this meeting. Equally grateful are we to our dear colleague Giuseppe Moruzzi and to Mrs. Moruzzi, who have devoted so much time and ingenuity to ensure that the Colloquium will be the success that it already promises to be. Further, I should like to thank on your behalf, the Fondazione Valentino Baldacci, whose generous financial aid has been so valuable. The director of this Foundation, our colleague Professor Ugo Baldacci, and its secretary, Doctor F. Suma, have once more shown the meticulous solicitude and cordiality which have been, in the past, characteristic of the contributions made by the Fondazione to the success of scientific enterprises. The Pisa meeting follows, after three years, the memorable Moscow Colloquium. The support given to us by the International Brain Research Organization is an indication that this Organization approves of the idea that the understanding of cerebral integration requires the spatial and spiritual cooperation of those who are studying this supreme problem in neurophysiology. In this era of political unrest in which we are condemned to live, our meeting here will be a vivid symbol of what can be done by friendly cooperation in a domain of science which is so intimately associated with the problems of human destiny and progress. We shall work in the shadow of the great monuments of a glorious city. They are the comforting testimony that great things can be accomplished in the midst of struggle and warfare, though the warfare to which I allude was certainly, I must
XI1
INTRODUCTORY REMARKS BY THE HONORARY PRESIDENT
admit, a warfare performed with the “conventional” weapons of the 12th century! To me it is particularly moving that our session will be held in the Institute of my old friend Giuseppe Moruzzi, an Institute which has made, as you know, outstanding contributions to the themes that we shall discuss. Our only regret is that Professors Beritashvili, Narikashvili and Smirnov, and Doctor Terzian are unable to share in our meetings. As a compensation for this, however, we shall enjoy the full attendance of Professor Richard Jung, who has happily recovered so quickly from the accident which recently, as a result of his passionate interest in Romanesque architecture, he had the misfortune to suffer.
This Page Intentionally Left Blank
Frideric Bremer
Dedication to Professor Bremer This volume is dedicated to a friend whose work has greatly advanced our understanding of the nervous system, whose presence can enliven the dullest meeting and whose 70th birthday gives us the excuse for expressing our feelings. Fortunately FrCdCric Bremer’s age has no relation to his present activity. Forty years in the laboratory have not lessened his keen interest in new investigations and his critical appreciation of the problems they bring. Though he can look back to the string galvanometer, he has retained his mastery of experimental technique. The use of some new technique has often led to a rapid advance in the physiology of the nervous system. Few nowadays can appreciate the difficulties of experiments on the brain before the barbiturates were available: indeed the introduction of Dial as a suitable anaesthetic in 1930 can count as a major advance in technique and the short note on it by Fulton, Liddell and g o c h deserves to rank as a turning point in the history of cerebral physiology. But the brain depressed by drugs is not the normal brain and another turning point came when Bremer introduced the “Cerveau isole” and “EncCphale isole” preparations in 1935-36. His preparation solved the problem of anaesthesia by providing for the division of the pathways for pain below the cerebral level leaving intact much of the regulating mechanism for cortical activity. He found, after mid-brain transection, that the pattern characteristic of sleep would be shown both in the EEG record and in oculomotor behaviour: the pattern changed from time to time to that characteristic of arousal and he found that the change could be brought by appropriate sensory stimuli. Thus his “Cerveau isole” and “EncCphale isole” preparations made it possible to start a new chapter in the analysis of the brain stem regulating centre. They have given a fresh impetus to research on the problems of sleep and attention and have opened up fields which are still the centre of interest. This work on the brain stem in relation to cerebral activity has been a major contribution to the physiology of the central nervous system but it is far from being the only important contribution he has made. His early study of the cerebellum in Sherrington’s laboratory and his more recent work on the auditory and visual pathways and cortical responses have established valuable results, and Bremer has always been attracted by the general problems which have been left unsolved because there is so much detail to be filled in. He has been concerned with the waves as well as the spikes, with the factors which can lead to synchronised rhythms in the cord and in the brain and with the general problem of auto-rhythmicity. It is indeed his concern for the whole advance and his knowledge of the way it has gone which gives him a special claim to our good wishes. Research on the central
XVI
DEDICATION TO PROFESSOR BREMER
nervous system offers an immense variety of topics, from the anatomy of the cell to the psychology of the individual: very few of us can hope to follow all the developments recorded in so many fresh publications, but Bremer has never lost touch with the major issues. In any discussion he can refer to the details and make us see them as part of the whole picture: and we can enjoy his own papers for their light on our immediate problems as well as for the lucid writing which commands our interest. Throughout his career Bremer has influenced the development of research on the central nervous system. We have profited by the methods he has introduced and the ideas he has given us. It is a pleasure to express our thanks to such a colleague and to send our congratulations on his seventieth birthday.
ADRIAN
Contents
Participants in the Pisa Colloquium
. . . . . . . . . . . . . . . . . . .
Introductory remarks on behalf of IBRO and UNESCO byH.H.JAsPER. . . . . . . . . . . . . . . . . . Introductory remarks by the Honorary President byF.BREmR.. . . . . . . . . . . . . . Dedication to Professor Bremer by Lord ADRIAN .. Contents.
vii
.........
ix
.............
xi
.....................
xv
.............................
xvii
Postsynaptic and presynaptic inhibitory actions in the spinal cord by J. C. ECCLES(Canberra, Australia) . . . . . . . . . . . Recurrent inhibition as a mechanism of control by R. GRANIT(Stockholm) . . . . . . . . .
. . . . . .
.............
Studies of the integrative function of the motor neurone by J. M. BROOKHART and K. KUBOTA (Portland, Oreg.)
. . . . . . . . . .
1
23
38
The plasticity of human withdrawal reflexes to noxious skin stimuli in lower limbs by K.-E. HAGBARTH and B. L. FINER (Uppsala, Sweden) . . . . . . . . . 65 Reticular homeostasis and critical reactivity by P. DELL(Paris) . . . . . . . . . . .
...............
Thalamic integrations and their consequences at the telencephalic level by D. ALBE-FESSARD and A. FESSARD (Paris) . . . . . . . . . . . Influence of unspecific impulses on the responses of sensory cortex by S . P. NARIKASHVILI (Tbilisi, U.S.S.R.) . . . . . . . . . . The tonic discharge of the retina and its central effects by A. ARDUINI (Pisa, Italy) . . . . . . . . . . . .
....
82
115
. . . . . . 155
..........
184
XVlIl
CONTENTS
Multisensory convergence on cortical neurons: Neuronal effects of visual, acoustic and vestibular stimuli in the superior convolutions of the cat’s cortex by R. JUNC, H. H . KORNHUBER andJ.S.D~FoNSEcA(Freiburg/Br.,Germany). 207 The direct cortical response (DCR). Associated events n pyramid and muscle during development of movement and after-discharge (St. and J. L. Q’LEARY by S. MINGRINO, W. S. COXE,R. KATZ, S. GOLDRIN( Louis, Missouri) . . . . . . . . . . . . . . . . . . . . . . . . . . 241 Brief survey of direct current potentials of the cortex by J. LO’LEARY (st. Louis, Missouri) . . . . . . . . . . . . . . . . . . 258 Studies of non-specific effects upon electrical responses i n sensory systems by H. H. JASPER (Montreal, Canada) . . . . . . . . . . . . . . . . . . 272 Aspects of sensorimotor reverberation to acoustic and visual stimuli. The role of primary specific cortical areas by P. BUSER,P. ASCHER, J. BRUNER, D. JASSIK-GERSCHENFELD and R. SINDBERG (Paris) . . . . . . . . . . . . . . . . . . . . . . . . . . . 294 New data on the specific character of ascending activations by P. K. ANOKHIN (Moscow) . . . . . . . . . . . . . . . . . . . . .
325
The characteristics and origin of voluntary movements in higher vertebrates by 1. s. BERlTASHVlLl (Moscow). . . . . . . . . . . . . . . . . . . . .
340
Responses in non-specific systems as studied by averaging techniques by MARYA. B. BRAZIER(Cambridge, Mass.) . . . . . . . . .
. . . . . 349
A transcranial chronographic and topographic study of cerebral potentials evoked by photic stinlulation in man by H. GASTAUT, E. BEEK,J. FAIDHERBE, G. FRANCK, J. FRESSY, A. REMOND, C. SMITHand P. WERRE(Marseille, France) . . . . . . . . . . . . . . . 374 Specific and non-specific responses and autonomic mechanisms in human subjects during conditioning by W. GREYWALTER(Bristol, Great Britain). . . . . . . . . . . . . . . 395 Sleep mechanisms. Chairman: J. M. BROOKHART. . . . . . . . . . . . . . . . . . . . . 404
CONTENTS
General discussion on inhibition . Chairman : F. BREMER. . . .
XIX
. . . . . . . . . . . . . . . . . . . .
444
. . . . . . . . . . . . . . . . . . . . . . . . . . .
453
Author index
. . . . . . . . . . . . . . . . . . . . . . . . . . . . .
475
Subject index
...............................
481
Final discussion
This Page Intentionally Left Blank
Postsynaptic and Presynaptic Inhibitory Actions in the Spinal Cord J. C . ECCLES
Department of Physiology, Australian National University, Canberra (Australia)
Sherrington was the first neurophysiologist to appreciate fully the key role of inhibition in the integrative action of the nervous system, particularly in the spinal cord. For him synaptic inhibitory actions ranked equally with synaptic excitatory actions. Despite his leadership both in experiment and in conceptual development, there is still today a tendency to consider the nervous system as operating very largely along excitatory pathways. The present paper is, in part, an attempt to redress this unbalance ; but a more compelling motive derives from the necessity for reconsidering the whole central inhibitory story now that it has been established that there are two quite distinct synaptic inhibitory mechanisms. The recent recognition of presynaptic inhibition(Frankand Fuortes 1957;Eccles 1961a,b; Eccles, Ecclesand Magni 1961;Eccles, Magni and Willis 1962; Eccles, Schmidt and Willis 1962; Eccles, Kostyuk and Schmidt 1962a,b) has led to a reinterpretation of many experimental investigations that have been made over the last few decades. A brief historical rksumt will serve to remind us once again that inhibitory phenomena provide the most fascinating problems in basic neurophysiology. Gasser and Graham (1933) found that dorsal root volleys produced slow positive potentials (P waves) of the cord dorsum, and that the time courses of these waves corresponded approximately to that of the inhibition of flexor reflexes when one dorsal root volley was employed to condition the flexor reflex evoked by another volley. Consequently, they asked “whether the positive potential may not be connected with the process responsible for inhibition?’ In a further communication Hughes and Gasser (1934) provided additional evidence supporting this correlation, and later Gasser (1937) attributed the inhibition to a depression of interneurones in a common central pathway that was produced by the positive after-potential that followed their activation by the conditioning volley. Barron and Matthews (1938) found that dorsal root volleys also gave rise to a depolarization that spread electrotonically along the same or adjacent dorsal roots and postulated that this dorsal root potential was produced by the same potential generator that gave the P wave, and that this generator was also responsible for inhibition; but it was suggested that this inhibition was brought about by electric currents which caused blockage of conduction in the collateral branches of interneurones. References P. 16-18
2
J. C. ECCLES
In recent years the evidence relating both the dorsal root potential and the P wave of the cord dorsum to a central inhibitory action has been lost sight of because the interneuronal theory of inhibition (Gasser 1937; Bonnet and Bremer 1939; Bremer and Bonnet 1942) could not explain inhibition of monosynaptic reflexes (Lloyd 1941, 1946; Renshaw 1941, 1942), and also because there was such convincing evidence that inhibition wasdueto thepostsynapticaction of special inhibitory synapses (Eccles 1953, 1957,1961a; Fatt 1954; Fessard 1959). Byintracellular recording it was shown that inhibitory and excitatory synapses had opposed actions on the postsynaptic membrane ; and inhibitory action seemed to be fully accounted for by the observed interaction of the postsynaptic currents and potentials produced by excitatory and inhibitory synapses, much as was originally postulated by Sherrington (1925) in his concept of algebraic summation. This position is no longer tenable because it has been shown that a large proportion of the inhibitions exhibited in the spinal cord is due to a quite different mechanism, which has been called presynaptic inhibition. Depolarization of excitatory presynaptic fibres causes a diminution in their synaptic action (Hagiwara and Tasaki 1958; Takeuchi and Takeuchi 1962; Eccles, Kostyuk and Schmidt 1962). It has now been shown that virtually all medullated primary afferent fibres in the spinal cord are depolarized by suitable conditioning volleys, and that this depolarization reduces their excitatory effectiveness and so results in the inhibition which has been designated presynaptic inhibition. This presynaptic depolarization that is responsible for presynaptic inhibition is also manifested both in the P wave of the cord dorsum and in the dorsal root potential, which thus re-establishes the earlier hypothesis of Gasser, Matthews, Bremer and their colleagues. However, the detailed mode of operation of presynaptic inhibition is very different from the earlier suggestions. It is now postulated that special interneurones form depolarizing synapses close to the synaptic terminals of the primary afferent fibres; the presynaptic depolarization so produced results in a diminution in the quantity of transmitter which is liberated by the impulses. Thus the postsynaptic excitatory action of the impulses is diminished not by any interaction at the level of the postsynaptic membrane (the postsynaptic inhibitory mechanism), but as a consequence of a diminution of the transmitter liberation, there being thus a smaller excitatory response from an otherwise unaltered postsynaptic membrane. POSTSYN APTlC INHIBITION
Integrative ,functions of postsynaptic inhibition: the Renshaw cell system When motoneurones discharge impulses to muscles, they also activate Renshaw cells via motor axon collaterals (Renshaw 1946). These Renshaw cells in turn inhibit the motoneurones and so tend to suppress the motoneuronal discharge (Eccles, Fatt and Koketsu 1954). The more intense the motoneuronal discharge, the more intense is the activation of Renshaw cells and the consequent inhibition of motoneurones. Thus the Renshaw cell system operates as a negative feed-back to motoneurones. When studied in detail, it is found that the negative feed-back from any particular impulse is not at all selective to the motoneurone that discharges that impulse, nor
3
INHIBITORY ACTIONS IN THE SPINAL CORD
to the whole population of motoneurones that supplies the same muscle, nor even to the whole assemblage of motoneurones supplying muscles of comparable function, e.g., flexion or extension. All that can be stated is that the Renshaw inhibition from the motoneurones of any muscle is more powerful on the motoneurones in close proximity, and in particular on the motoneurones of slow tonic muscles (Fig. 1 ; Granit et al. 1957; Eccles, Eccles, Iggo and Ito 1961). Possibly this selective distribution serves to stabilize the frequency of motoneuronal discharge to tonic muscles
---C
SG
SM
--
Q
MG
PI
LG
Po P
Y -
PB
Fig. 1 A : diagram showing synapses from two motor axons onto three Renshaw cells whose axons in turn
make inhibitory synaptic connections onto the two motoneurones. Note microelectrode in position for recording extracellularly from one Renshaw cell as in B. B : responses of a Renshaw cell evoked by single maximal antidromic volleys in the motor nerves to various muscles. PI - plantaris; Sol - soleus; MG - medial gastrocnemius; LG - lateral gastrocnemius. All other antidromic volleys were ineffective. C : intracellular recording from an anterior biceps motoneurone, showing the responses evoked by single maximal antidromic volleys in the motor nerves to various muscles as indicated by symbols. SG - superior gluteal; Q - quadriceps; MG - medial gastrocnemius; PI - plantaris; SM semimembranosus; PB - posterior biceps; LG - lateral gastrocnemius; Pop - popliteus; AB anterior biceps; ST - semitendinosus; Sol - soleus; FDL - flexor digitorum longus; IG - inferior gluteal; G R - gracilis; PT - posterior tibial. Note that the anterior biceps volley evokes an antidromic spike potential of the motoneurone with a subsequent after-hyperpolarization that is superimposed on the Renshaw IPSP. The arrow marks the approximate size of the IPSP alone
during maintenance of postures (Granit et al. 1957). An alternative suggestion is that it serves to suppress all discharges from tonic motoneurones during the rapid movements of running or jumping. This suppression is functionally desirable, else the slowly contracting and relaxing tonic muscles would impede the rapid movements (Denny Brown 1928; Eccles, Eccles, lggo and Ito 1961). Apart from this special action on tonic motoneurones, the general action of the Renshaw cell system is to produce an unspecific limitation on the activity of motoneurones regardless of their function. References P. 16-16:
4
J. C. ECCLES
The postsynaptic inhibitory action of impulses from annulospiral endings of muscle spindles According to a general rule first enunciated by Lloyd (1946) the large afferent fibres (group Ia of muscle spindles) not only excite the motoneurones of that muscle and muscles of like function, but they also inhibit the motoneurones of antagonistic muscles at the same joint, so forming a “myotatic unit” (Fig. 2, A-C). More extensive investigation has revealed that there are many exceptions to this generalization, particularly for the motoneurones supplying muscles acting at the hip and knee joints (Eccles and Lundberg 1958; Lundberg 1959); nevertheless, the myotatic unit is a useful concept in the attempt to understand the factors involved in the maintenance of joint position. For example, if flexor motoneurones are activated so intensely that they discharge impulses, the resultant flexor contraction may flex the joint and so stretch the extensor muscles. Thus the discharge along the extensor Ia afferent fibres will be increased with the consequent inhibition of the flexor motoneurones and monosynaptic excitation of the extensor motoneurones, as diagrammed in Fig. 2 0 . If extension of the joint supervenes, there is a complementary pattern with activation of the flexor
Q AS
’A Fig. 2
A : diagram showing at the Lo and
L7 segmental levels of the spinal cord the central pathway for
a group Ia afferent fibre (QIa) that comes from an annulospiral ending (AS) in the quadriceps muscle (Q). One branch monosynaptically excites a quadriceps motoneurone, the other branch descends to L7 and excites an interneurone which in turn inhibits a motoneurone of the antagonist muscle, biceps-semitendinosus (BST). Excitatory and inhibitory synapses are labelled E and I respectively. B and C are respectively the intracellular responses (upper traces) evoked by a maximal group Ia quadriceps afferent volley (see lower traces) in a quadriceps motoneurone (B) and in a bicepssemitendinosus motoneurone (C) in the same experiment. D: diagram showing the effect of a reflex discharge of flexor motoneurones (FM) in causing knee flexion (see arrow below knee joint) with a consequent excitation of the annulospiral endings (AS) of the extensor muscle. This discharge up the Ia afferent fibre gives monosynaptic excitation of the extensw motoneurones (EM) and disynaptic inhibition of the flexor motoneurones (FM), an inhibitory interneursne (I cell) being interpolated on this inhibitory pathway.
INHIBITORY ACTIONS IN THE SPINAL CORD
5
motoneurones and inhibition of the extensor. Thus, at our present level of understanding, the myotatic unit provides a reciprocal control tending to stabilize the joint at a position determined by the integrated excitatory and inhibitory influences playing upon the extensor and flexor motoneurones from pathways other than those of the myotatic unit; such as the descending tracts and the various other afferent inputs into the spinal cord. Group Ia impulses from muscle also exert a postsynaptic inhibitory action on the cells of origin of the dorsal and ventral spinocerebellar tracts (Eccles, Oscarsson and Willis 1961; Eccles, Hubbard and Oscarsson 1961). However, detailed investigation of a far greater number of cells is necessary before attempting to develop postulates relating to the possible integrational significance of these inhibitory actions. The postsynaptic inhibitory action of impulses from the Golgi tendon organs of muscle The Golgi tendon organs differ from the muscle spindle receptors because they are in series with the contracting muscle fibres and have a much higher threshold. Thus it requires a powerful muscle contraction in order to excite these receptors and cause an effective discharge of impulses up the group Ib fibres (Lundberg and Winsbury 1960). These Ib impulses cause inhibition of extensor motoneurones, not only of that joint, but of all joints of that limb(Fig. 3; Granit 1955;Laporte and Lloyd 1952;Eccles, Eccles and Lundberg 1957). Thus under conditions where there is intense activation
msec
llllllnTnlllllllll
Fig. 3 A, B: intracellular records from a gastrocnemius motoneurone o IPSP's evoked by a plantaris an1 a flexor digitorum group I volley respectively. There is a record of the dorsal root volley below each intracellular record. C: as in A, B, but for a quadriceps afferent volley in another experiment, the dorsal root record being above the intracellular record. D: diagram showing how an intense discharge of extensor motoneurones (EM) when causing powerful muscle contraction against resistance excites Golgi tendon organs to discharge impulses along Ib fibres, which in turn excite inhibitory cells (I cell) that inhibit the extensor motoneurones. References p . 16-18
6
J. C. ECCLES
of extensor motoneurones, such as in leaping or in decelerating a fall, the tension on the muscle is prevented from reaching a damaging intensity by the Ib inhibitory action on the extensor motoneurones. Ib impulses also have postsynaptic inhibitory actions on the cells of origin of the dorsal and ventral spinocerebellar tracts; but, as with the la inhibition, much more extensive investigation is required.
The postsynaptic inhibitory action of impulses from the flexor reflex afferents ( F R A ) These afferents include both the alpha and delta groups of cutaneous fibres (Mark and Steiner 1958; Hunt and McIntyre 1960), the group 11 and 111 afferent fibres from muscle (Lloyd 1943) and the afferent fibres from joints (Eccles and Lundberg 1959). Their postsynaptic inhibitory action has been demonstrated on motoneurones and the cells of origin of the dorsal and ventral spinocerebellar tracts. The inhibitory postsynaptic potentials may be 50 msec or more in duration (Eccles 1953; Kostyuk 1960). Presumably the preponderant inhibition of extensor motoneurones is part of a protective general flexor reflex in which the limb is withdrawn from some injurious contact (Creed et al. 1932). Much more investigation is required before precise statements are possible, but the detailed studies by Hagbarth on cutaneous afferents (1952 and this symposium) are important developments in this direction. PRESYNAPTIC INHIBITION
Inhibition of central effects of Group l a afferent Jibres By several different experimental procedures it has been shown that volleys in group Ia and group 1b afferent fibres of flexor muscles depolarize the group La fibres of both extensor and flexor muscles. This depolarization is directly observed by intracellular recording from the afferent fibres in the dorsal part of the spinal cord (Fig. 4, A - C ) ; but it is also demonstrated by the increased excitability of the group Ia afferent fibres, which is observed along their whole intramedullary course, being largest close to their terminals in the ventral horn (Fig. 4 D; Eccles, Magni and Willis 1962). Other evidence of the group Ia depolarization is given by the potential field that it produces in the cord, maximum negativity in the ventral horn and positivity in the dorsal columns (Fig. 5 A ; Eccles, Magni and Willis 1962), and by the generation of impulse discharge in the group Ia fibres, which gives rise to the dorsal root reflex (Fig. 5B; Eccles, Kozak and Magni 1961). These various procedures show that the presynaptic depolarization has a characteristic slow time course: latency about 4 msec, summit at about 20 msec and a total duration of at least 300 msec. The briefer duration of the dorsal root reflex is attributed to accommodation. Several afferent volleys in quick succession considerably increase the depolarization, and potentials in excess of I mV are regularly observed in the fibres in the cord dorsum (Fig. 4). Presumably the depolarizations are several times larger in the fibre terminals in the ventral horn. In every respect this presynaptic depolarization accounts for the depression of the monosynaptic EPSP which is observed after conditioning by group 1 volleys from flexor muscles (Fig. 6 A , B ; Frank and Fuortes 1957; Eccles, Eccles and Magni 1961),
INHIBITORY ACTIONS IN THE SPINAL CORD
b
I0
20
30
40
50
60
7
70mre:0
n
Fig. 4 A shows the experimental arrangement for obtaining the responses B-C. The microelectrode is inserted into a group la afferent fibre (GS) from the gastrocnemius-soleus muscle at a depth of 0.6 mm from the cord dorsum, and the potential changes of the upper records of B are produced by 1 , 2 and 4 group I volleys in the combined posterior biceps-semitendinosus (PBST) and deep peroneal nerves (PDP). The lower records of B show the potentials similarly produced, but with the microelectrode withdrawn to a just extracellular position. The actual potential changes across the membrane of that group Ia fibre are given by the differencesbetween the corresponding intracellular and extracellular potentials, as plotted in Cfor 1 , 2 and 4 volleys, upward deflections signalling depolarization. Presumably the potential changes are much larger close to the presumed site of depolarization near the synaptic terminals of these group Ia fibres on motoneurones in the ventral horn, as shown in A. D shows the time course of the excitability changes which a single PBST volley produces in group la afferent fibres of gastrocnemius nerve at the region of their synaptic terminals in the ventral horn (cf. Eccles, Magni and Willis 1962). Ordinates show excitability values as percentages of the control, while abscissae give test intervals.
and also for the depression of a monosynaptic testing reflex (Fig. 6C; Eccles, Schmidt and Willis 1962). The responses of both flexor and extensor motoneurones are equally depressed, which corresponds to the similar depolarizations of the group Ia afferent fibres. It now remains to discuss the possible physiological meaning of this presynaptic inhibitory action. The action of strychnine clearly distinguishes between presynaptic and postsynaptic inhibition, the latter being heavily depressed by intravenous doses of 0.1 mg/kg, whereas the former is unaffected. There are many instances in the literature of inhibitions that were unaffected by strychnine, and at least some of them are typical examples of presynaptic inhibition due to depolarization of group l a afferent fibres. References P . 16-18
8
f . C . ECCLES
1*
4
4
*
%
$
w
2 04&2 2 4 e 2 44
3 04
/
I
f l w
Fig. 5 A : field potentials generated by five group 1 volleys in PBST nerve at the indicated depths in mm along a microelectrode track from the dorsum of the cord in a ventro-lateral direction. Upward deflections signal negativity relative to the indifferent earth lead, there being reversal from the slow positive potential in the dorsal regions to a slow negative potential at depths in excess of 2 mm. B: recording of dorsal root reflexes by a microelectrode inserted into a group Ia afferent fibre of gastrocnemius nerve. The single or double reflex discharges seen in the lower tracings were produced by a combined PBST and PDP volley at a low body temperature (31°C). The upper traces show the cord dorsum potentials. (Eccles, Kozak and Magni 1961.)
For example, Liddell and Sherrington (1925) found that the stretch reflex of quadriceps muscle was inhibited by pulling on the knee flexor muscles, and this inhibition was not affected by injection of 0.1 mg/kg of strychnine (Fig. 7), or even of larger doses. This observation is important for it shows that in large part the inhibition was not due to the postsynaptic inhibitory action of group Ia afferent fibres (cf. Fig. 2); but it is precisely the presynaptic inhibition that would be expected to be produced by the group I afferent impulses from the knee flexors (cf. also, Cooper and Creed 1927). Evidently, presynaptic inhibition is very effectively produced by the asynchronous repetitive discharges of muscle stretch receptors, which is not surprising in view of the finding that presynaptic inhibition is built up by repetitive volleys even at relatively low frequencies, and has a very long duration (Eccles, Eccles and Magni 1961). When employing the usual procedures of testing by a monosynaptic reflex, presynaptic inhibition appears to be more potent than postsynaptic inhibition, even large monosynaptic reflexes being virtually suppressed (cf. Fig. 6C) ; hence it would be expected that presynaptic inhibition would play an important role in the control of the reflex movements. For example, under certain circumstances presynaptic inhibition provides the negative sign in a homeostatic system. Thus in Fig. 8 stretch of flexor muscles by a powerful extensor contraction would depress the activation of
9
INHIBITORY ACTIONS IN THE SPINAL CORD
ioa %
75 50
25
0 0
300
200
100
400
msec 500
Fig. 6 A : time course of depression of the EPSP produced in a plantaris motoneurone by monosynaptic action of a group I afferent volley in plantaris nerve. The conditioning was produced by four PBST volleys, maximal for group I, and the abscissae show intervals between the first of the conditioning volleys and the testing plantaris volley. Specimen records composed of superimposed tracings are shown in B, the ordinates of the plotted curve being the sizes of the testing EPSP's expressed as percentages of the control (CON). C: time course of depression of the monosynaptic reflex discharge evoked by a maximum group I gastrocnemius volley. The conditioning was produced as in A by four PBST volleys, and the abscissae show the intervals between the first of the conditioning volleys and the testing gastrocnemius volley. Specimen records are shown in the inset with testing intervals given in msec, CON being the control reflex spike.
L""" 2
3
4
set
1
2
3
4
Fig. I A : mechanical responses of quadriceps muscle to a stretch applied as shown in T before (M and M') and after (P) severance of its nerve. The additional tensions in M and M1 over P are due to stretch reflex contraction. During each response a stretch of biceps muscle was applied commencing approximately at the points marked by the arrows. M gives the response before and M 1after injection of strychnine in a just sub-convulsive dose. B: as in A, but at the arrow the knee flexor muscles were stretched merely by a passive extension of the knee that commenced approximately at the arrows (Liddell and Sherrington 1925). References P . 16-18
10
J. C. ECCLES
Fig. 8 Diagram showing how a contracting extensor muscle extends joint and strongly stretches a flexor muscle, particularly if it is contracting. The resulting discharge up the la and lb fibres from the flexor excites an interneuronal pathway that leads to presynaptic inhibition of the Ia synapses on the extensor motoneurone, so introducing the negative sign in the circuit that began with powerful extensor activation.
extensor motoneurones by means of the presynaptic inhibitory action on the group l a afferent fibres of the extensors, i.e., on the gamma-loop system of the extensors. The weaker flexor muscles have access to a system which reduces the excitation of extensor motoneurones ; and the flexor muscles would exercise a particularly effective influence of this kind when they were actively contracting against the more powerful extensors. Possibly this action is of importance in terminating the extensor phase of the step. The duration of the presynaptic inhibition corresponds well to the duration of the extensor quiescence in a step. However, this presynaptic inhibitory action would also be equally exerted on the monosynaptic excitation of flexor motoneurones. Evidently the presynaptic inhibition that is initiated by group 1 impulses from flexor muscles has a more complex functional meaning that still eludes our understanding. The complexity of the system will be appreciated when it is realized that group 1 afferent volleys from a flexor muscle at knee or ankle exert presynaptic inhibition on group l a fibres of the flexors and extensors of all joints of a limb (Eccles, Eccles and Magni 1961; Eccles, Schmidt and Willis 1962). A further complication is that the presynaptic depolarization of group la fibres is not entirely produced by the group I fibres of flexor muscles: the knee extensor muscle, quadriceps, often has a small but significant action. Finally, it should be mentioned that group Ib afferent fibres of
INHIBITORY ACTIONS IN THE SPINAL CORD
11
muscle also are depolarized by group 1 afferent volleys; but, in contrast to group Ia fibres, this depolarization is exerted by extensor afferent volleys with an effectiveness not much less than that from flexor volleys. Afferent volleys from cutaneous nerves have virtually no depolarizing action on either group la or Ib fibres of muscles. Inhibition of the central efects of the aferent jibres responsible f o r evoking t h e y e x o r reflex Investigations on the actions of afferent volleys both indirectly on motoneurones via interneuronal pathways and on the cells of origin of various types of ascending tracts have led to the conclusion that cutaneous and joint afferent fibres as well as group 11 and I11 afferent fibres from muscle have in common a diversity of central actions, which includes the setting up of flexor reflexes; hence they have come to be known collectively as flexor reflex afferents (FRA) (Eccles and Lundberg 1959; Holmqvist et al. 1960; Holmqvist and Lundberg 1961). Single volleys in cutaneous nerves produce a large and prolonged depolarization of cutaneous afferent fibres within the spinal cord. This depolarization can be accurately measured by recording with an intracellular electrode (Koketsu 1956; Eccles and KrnjeviC 1959), but it is more readily observed as the dorsal root potential (Fig. 9) or by the centrifugal discharge of impulses, the dorsal root reflex (Tonnies 1938). There is now evidence that all FRA afferent fibres both give and receive depolarization in the spinal cord. For example, group I1 and 111 volleys from muscle produce a dorsal root potential (Fig. 9B) by depolarizing cutaneous fibres, and, complementarily, cutaneous volleys
Fig. 9 A : diagram showing manner of obtaining the responses illustrated in B and C. Recording of the dorsal root potential (DRP) is effected by leads applied to an isolated dorsal rootlet (from L7 root), one lead being about 1 mm from the origin of the rootlet from the spinal cord, the other on its cut distal end. The P potential from the cord dorsum is recorded by one electrode on the cord dorsum close to the dorsal root entry, the other being indifferently placed on the lumbar muscles. The responses of B were elicited by single stimuli to the gastrocnemius-soleus (GS) nerve at the indicated strengths relative to threshold, the upper traces being DRP’s, the other responses being led from the cord dorsum as shown in A . Strengths of stimuli are shown relative to threshold, 2.1 being maximum for group I , 8.4 for group 11 and 42 for group 111. The responses of C were elicited by stimulation of the purely cutaneous superficial peroneal nerve (SP) at 1.5, 4 and 40 times threshold. which are respectively about half maximum for the alpha fibres, maximal for the alpha fibres and maximal for the alpha and delta fibres. (Eccles, Kostyuk and Schmidt 1962a.) R e f e w n u s P. 16-18
12
J. C . ECCLES
depolarize group I1 and 111 muscle afferents (Eccles, Kostyuk and Schmidt 1962a,b). Volleys in cutaneous and high threshold muscle afferents excite the same interneurones in the spinal cord, and some a t least of these interneurones have properties suggesting that they lie on the central pathway concerned in depolarization of the FRA fibres. Fig. 10 gives diagrammatically this postulated interneuronal pathway in its simplest form. Impulses in cutaneous and group I1 and 111 muscle afferent fibres synaptically excite different interneurones, but these in turn converge onto and excite second order interneurones (called D-type), which make synaptic connections with the central terminals of primary afferent fibres, so effecting the presynaptic depolarization.
2mm
Fig. 10 Schematic diagram illustrating the suggested pathway for presynaptic inhibitory action on a cutaneous primary afferent fibre. Three cutaneous afferent fibres (C) and group I1 and 111 muscle afferent fibres (I1 and 111) are shown ending monosynaptically on their respective interneurones. These interneurones in turn have excitatory synaptic connections on an interneurone (D) whose axon makes synaptic connections on the synaptic terminals of the C fibres, which i s the postulated pathway for the presynaptic inhibition.
By four independent methods of investigation it has been shown that a conditioning FRA volley produces a depression of the response to a testing FRA volley, which is attributable to the presynaptic depolarization produced by the conditioning volley, and thus is a further example of presynaptic inhibition. Firstly, there is inhibition of the flexor reflex, which is illustrated in Fig. 11 for two cutaneous volleys that produced almost the same flexor reflexes. There is seen to be approximately the same prolonged time course of inhibition when a volley in either one conditioned the flexor reflex response produced by the other. At testing intervals up to 50 msec there was almost complete inhibition. Secondly, there is inhibition of both the monosynaptic and
INHIBITORY ACTIONS IN THE SPINAL CORD su
13
-
Pl+SP
Fig. 11 Flexor reflex inhibition by a single cutaneous volley. The flexor reflex was evoked by a single volley in another cutaneous pathway and was recorded monophasically from the nerve to PBST muscle. The specimen records in A and B show in the upper line the reflex discharge which is integrated in the lower line. The numbers indicate the intervals in msec between the conditioning and testing volleys, which are specified above the respective series. PT and SP are combined to give about the same flexor reflex response (control (CON) in A ) as SU alone (CON in B). In C the amplitude of the integrated spike is plotted in percent of control values for various intervals between the two volleys. Open circles correspond with A , filled circles with B, the curve being drawn for the open circles. Stimulus strength for all nerves was 4 times threshold. Stimuli were applied every 4.5 sec. In A and B the voltage scale is for the monosynaptic reflexes. Same time scale in msec for all records. (Eccles, Kostyuk and Schmidt 1962b.)
delayed discharges which cutaneous volleys produce in the ascending tract fibres of the dorsolateral funiculus of the same side (Fig. 12). The monosynaptic discharge is always less depressed than the delayed discharge, which is to be expected because it is produced by such a powerful synaptic excitation. Thirdly, there is depression of the monosynaptic EPSP which a cutaneous volley produces in neurones in the dorsal horn, much as in Fig. 6 A , these neurones being either interneurones or relay cells for the monosynaptic tract discharge (Eccles, Eccles and Lundberg 1960). Finally, as shown in Fig. 13, there is depression of the dorsal root potential (the presynaptic depolarization) produced by the testing volley (cfi Barron and Matthews 1938; Bonnet and Bremer 1938).Thus we have the basic performance of a negative feed-back system: an afferent volley produces presynaptic depolarization, which results in presynaptic inhibition, which in turn diminishes the effectiveness of a later afferent volley in producing presynaptic depolarization and presynaptic inhibition. Three of these experimental procedures have shown that presynaptic inhibitory action also occurs between cutaneous fibres and group I1 and I11 afferent fibres from muscle. Intracellular recording of the monosynaptic EPSP’s as in Fig. 6 A has not yet been investigated with this type of interaction. When attempting to understand the physiological significance of the negative feedback provided by presynaptic inhibition, it is helpful to consider the operational References P. 16-18
14
J . C . ECCLES
0
so
100
150
2OD
250
miec
300
Fig. 12 Inhibition of discharges recorded from the ipsilateral cutaneous tract that was isolate1 in the upper lumbar segments and mounted on electrodes for monophasic recording. The action potentials were evoked by a single SU stimulus (4 times threshold) and were conditioned by another single stimulus (4 times threshold) in SP. The specimen records in A show in the upper line the early tract discharges, in the lower one the cord dorsum potentials. Each record consists of three superimposed sweeps. The numbers indicate the interval in msec between the two volleys, and CON is the control record. B shows a similar series recorded with a slower sweep speed, and with integrated traces so that the amplitude of the late discharges can be measured. Potential scales are for the monophasic records only. In C (open circles) the depression of the monosynaptic discharge (i.e., the amplitude of the first spike in A ) is plotted in per cent of the mean control value against the interval between the two volleys in msec. The filled circles show the depression of the late discharges as given by the integrated records of B. These were measured after the time of the monophasic spike in order to reject its contribution to the integrated response, i.e., measurements were from 1.3 msec to 19 msec after the onset of the monophasic discharge. (Eccles, Kostyuk and Schmidt 1962b.)
conditions that normally obtain with all the diversity of afferent input into the spinal cord. Hitherto it has been assumed that all of this sensory information is processed in the spinal cord during transmission through interneuronal pathways which offer opportunities for inhibitory action at each synaptic relay. This inhibition would be exerted by a specific postsynaptic process generating currents that antagonised the postsynaptic depolarizing action of the excitatory synapses. Thus rejection of sensory input could occur only after it had excited interneurones. Presynaptic inhibition provides a mechanism for suppressing the sensory input before it has exerted a n y synaptic action. In this way a powerful afferent input through FRA channels can suppress all trivial inputs before they have an effective action on the central nervous system, which, as a consequence, is “cleared” for the “urgent” reflex actions set in train by the powerful input. It will be objected that the powerful input will also itself be subjected to the diffuse depressant action of the presynaptic inhibition which it generates, but, if it is sufficiently powerful, the depression will be negligible, as is illustrated for example with the monosynaptic tract discharges in Fig. 12. In general terms it can be stated that presynaptic inhibition provides the first stage in what we may term “perceptual attention”, whereby powerful sensory inputs with an implica-
1NHlRlTORY ACTIONS IN THE SPINAL CORD
‘C x 80.
.
60’
0
0
100 100
100
300 300
su SP
400
500
5P
su
msec
Fig. 13 Interaction of cutaneous volleys as revealed by dorsal root potentials (DRP) recorded as in Fig. 9 A . The rootlet was taken from the caudal part of LS dorsal root. A shows the control dorsal root potential (CON) generated by a SP volley and its inhibition by a preceding volley in SU at intervals given by the time scale. In B the dorsal root potential evoked by the SU volley was conditioned by a SP volley. In C are the plotted points of the series partly illustrated in A (open circles) and in B (filled circles). The ordinates are the sizes of the DRP expressed as percentages of the mean control values and measured as the addition to the conditioning DRP. The stimulus strength for the nerves was adjusted to give about the same DRP size for SU and SP in the rootlet under observation. (Eccles, Kostyuk and Schmidt 1962b.)
tion of urgency can suppress all concurrent trivial inputs into the central nervous system. Presynaptic inhibition probably also is responsible for contrast phenomena. DISCUSSION
The general conclusion from the investigations on presynaptic inhibition is that there are in the spinal cord two neuronal systems that operate in producing presynaptic inhibition, with possibly a subsidiary system that acts on group Ib afferent fibres from muscle. One of these systems is virtually restricted to group I muscle afferents, being produced very largely by the group I afferent fibres of flexor muscles and received by the group Ia fibres of both flexor and extensor muscles. The other principal system probably includes all the FRA group on the receiving side, and is activated even more widely because the group I muscle afferents contribute in addition to the FRA system, though apparently by a quite independent interneuronal pathway. The operational relationships of the two systems can be represented as in Fig. 14 in which arrows show the directions of action of the depolarizing pathways. The subsidiary system giving depolarization of group Ib afferent fibres is also indicated, together with the action of group I volleys from extensors on the FRA system of fibres. The two principal systems correspond approximately to the two modes of sensory References p . 16-18
16
J. C. ECCLES
Group la fibres-Group
Group Ib fibres -Group
I volleys of -FRA flexor muscles
I volleys o f
fibres-
FRA volleys
/
extensor muscles
Fig. 14 Block diagram showing the operational relationships between the systems of presynaptic inhibitory action. As described in the text there are three different sets of primary afferent fibres, group la, group Ib and FRA fibres. The arrows show the directions of action of the pathways that depolarize these three sets of primary afferent fibres.
input into the spinal cord. There is firstly the input from limbs functioning as passive and static receptors, e.g., hair movement, touch and pressure on skin, pressure on muscle (group 111; Paintal 1960, 1961; Bessou and Laporte 1961) and static information of muscle tension (group 11 of muscle; Harvey and Matthews 1961). Secondly, there is the group I input from muscle, which is principally dynamic (CJ: Harvey and Matthews 1961 !for group la) and which is largely concerned with the control of movement. It is activated particularly in response to the alpha and gamma motoneuronal discharges to muscle in combination with limb movements and external forces producing muscle tension. As a final comment it can be suggested that presynaptic inhibition may occur at sites remote from the zone of entry of afferent fibres into the spinal cord. For example inhibition with the prolonged time course of presynaptic inhibition occurs with the synaptic relay of group l a fibres on the DSCT cells in Clarke’s column (Eccles, Oscarsson and Willis 1961). Furthermore, Wall (1958) showed that, after a conditioning volley in the S1 dorsal root, the presynaptic terminals of primary afferent fibres in nucleus gracilis in the upper cervical region displayed an increased excitability that had a time course comparable to that observed in the lumbar enlargement; hence it can be presumed that the conditioning volley had effected a presynaptic inhibitory action in the nucleus gracilis by depolarizing the presynaptic terminals therein. As yet there has been no evidence suggesting that presynaptic inhibition occurs at synapses other than those made by primary afferent fibres. Evidently this possibility should be rigorously explored at sites that lend themselves to testing for presynaptic depolarization.
REFERENCES BARRON, D. H. and MATTHEWS, B. H. C. The interpretation of potential changes in the spinal cord. J. Physiol. (Lotid.), 1938, 92: 216-321. BESSOU, P. et LAPORTE, Y. Etude des rkcepteurs rnusculaires innerves par les fibres afferentes du groupe 111 (fibres myelinisees fines) chez le chat. Arch. ital. Biol., 1961, 99: 293-321. BONNET,V. et BREMER, F. Etude des potentiels electriques de la moelle epiniere faisant suite chez la grenouille spinale a une ou deux volees d’influx centripetes. C. R. SOC.Biol. (Paris), 1938, 127: 806-812. BONNET, V. et BREMER, F. D u mecanisme de I’inhibition centrale. C.R. SOC.Biol. (Paris), 1939, 130: 760-161.
INHIBITORY ACTIONS IN THE SPINAL CORD
17
BREMER, F. et BONNET, V. Contributions A l’etude de la physiologie gkndrale des centres nerveux. 11. L‘inhibition reflexe. Arch. int. Physiol., 1942, 52: 153-194. S. and CREED, R. S. More reflex effects of active muscular contraction. J. Physiol. (Lond.), COOPER, 1927, 64: 199-214. CREED, R. S . , DENNY-BROWN, D., ECCLES, J. C., LIDDELL, E. G. T. and SHERRINGTON, C. S. Reflex activity of’the spinal cord. Oxford University Press. London, 1932, 183 pp. DENNY-BROWN, D. On the essential mechanism of mammalian posture. D. Phil. Thesis, Oxford, 1928. ECCLES,J. C. The neurophysiofogical basis of mind: The principles of neurophysiofogy. Clarendon Press, Oxford, 1953, 314 pp. ECCLES, J. C. The physiology of nerve cells. The Johns Hopkins Press, Baltimore, 1957, 270 p. J. C. The nature of central inhibition. Proc. roy. SOC.B, 1961a, 153: 445476. ECCLES, ECCLES, J. C. The mechanism of synaptic transmission. Ergebn. Physiol., 1961b, 51 : 299430. ECCLES,J. C., ECCLES,R. M., ICCO,A. and ITO, M. Distribution of recurrent inhibition among motoneurones. J. Physiol. (Lond.), 1961,159: 479499. ECCLES,J. C., ECCLES,R. M. and LUNDBERG, A. Synaptic actions on motoneurones caused by impulses in Golgi tendon organ afferents. J. Physiol. (Lond.), 1957, 138: 227-252. ECCLES,J. C., ECCLES,R. M. and LUNDBERG, A. Types of neurone in and around the intermediate nucleus of the lumbosacral cord. J. Physiol. (Lond), 1960, 154: 89-114. ECCLES, J. C., ECCLES, R. M. and MACNI,F. Central inhibitory action attributable to presynaptic depolarization produced by muscle afferent volleys. J.Physiol. (Lond.), 1961,159: 147-166. ECCLES,J. C., FATT,P. and KOKETSU, K. Cholinergic and inhibitory synapses in a pathway from motor-axon collaterals to motoneurones. J. Physiol. (Lond.), 1954, 126: 524-562. ECCLES, J. C., HUBBARD, J. I. and OSCARSSON, 0. Tntracellular recording from cells of the ventral spino-cerebellar tract. J. Physiol. (Lond.), 1961, 158: 486-516. ECCLES, J. C., KOSTYUK, P. G. and SCHMIDT, R. F. Central pathways responsible for depolarization of the primary afferent fibres. J. Physiol. (Lond.), 1962a, 161: 237-257. ECCLES, J. C., KOSTYUK, P. G. and SCHMIDT, R. F. Presynaptic inhibition of the central actions of flexor reflex afferents. J. Physiol. (Lond.), 1962b, 161: 258-281. ECCLES, J. C., KOSTYUK, P. G. and SCHMIDT, R. F. The effect of electric polarization of the spinal cord on primary afferent fibres and their monosynaptic excitatory action. J. Physiol. (Lond.),1962, 162: 138-150. ECCLES, J. C., KOZAK,W. and MAGNI,F. Dorsal root reflexes of muscle group I afferent fibres. J. Physiol. (Lond.), 1961,159: 128-146. ECCLES, J. C. and KRNJEVIC, K. Potential changes recorded inside primary afferent fibres within the spinal cord. J. Physiol. (Lond.), 1959, 149: 250-273. ECCLES, J. C., MAGNI,F. and WILLIS,W. D. Depolarization of central terminals of group I afferent fibres from muscle. J. Physiol. (Lond.), 1962, 160: 62-93. ECCLES, J. C., OSCARSSON, 0. and WILLIS,W. D. Synaptic action of group I and 11 afferent fibres of muscle on the cells of the dorsal spino-cerebellar tract. J. Physiol. (Lond.),1961,158: 517-543. ECCLES, J. C., SCHMIDT, R. F. and WILLIS,W. D. Presynaptic inhibition of the spinal monosynaptic reflex pathway. J. Physiol. (Lond.), 1962, 161: 282-297. ECCLES,R. M. and LUNDBERG, A. Integrative pattern of la synaptic actions on motoneurones of hip and knee muscles. J. Physiol. (Lond.), 1958, 144: 271-298. ECCLES, R. M. and LUNDBERG, A. Synaptic actions in motoneurones by afferents which may evoke the flexion reflex. Arch. ital. Biol., 1959, 97: 199-221. FATT,P. Biophysics of junctional transmission. Physiol. Rev., 1954,34: 674-710. FESSARD, A. Les processus de base de l’inhibition centrale. XXI. International Congress of Physiological Sciences. Symposia and special lectures. Buenos Aires, 1959: 4046. FRANK, K. and FUORTES, M. G. F. Presynaptic and postsynaptic inhibition of monosynaptic reflexes. Fed. Proc., 1957, 16: 3940. H. S. The control of excitation in the nervous system. Harvey Lect., 1937, 32: 169-193. GASSER, GASSER, H. S. and GRAHAM, H. T. Potentials produced in the spinal cord by stimulation of the dorsal roots. Amer. J. Physiol., 1933, 103: 303-320. GRANIT, R. Receptors and sensory perception. Yale University Press, New Haven, 1955, 369 p. GRANIT,R., PASCOE, J. E. and STEG,G. The behaviour of tonic and y-motoneurones during stimulation of recurrent collaterals. J. Physiol. (Lond.), 1957, 138: 381400. HAGBARTH, K. E. Excitatory and inhibitory skin areas for flexor and extensor motoneurones. Acta physiol. scand., 1952, 26, Suppl. 94.
18
J . C . ECCLES
HAGIWARA, S. and TASAKI,I. A study of the mechanism of impulse transmission across the giant synapse of the squid. J. Physiol. (Lond.), 1958, 143: 114-137. HARVEY, R. J. and MATTHEWS, P. B. C. The response of de-efferented muscle spindle endings in the cat’s soleus to slow extension of the muscle. J . Physiol. (Lond.), 1961, 157: 370-392. HOLMQVIST, B. and LUNDEIERG, A. Differential supraspinal control of synaptic actions evoked by volleys in the flexion reflex afferents in alpha motoneurones. Acta physiol. scand., 1961, $4, Suppl. 186, 51 pp. HOLMQVIST, B., LUNDBERG, A. and OSCARSSON, 0. Supraspinal inhibitory control of transmission to three ascending spinal pathways influenced by the flexion reflex afferents. Arch. ital. Biol., 1960, 98: 60-80. HUGHES,J. and GASSER,H. S. The response of the spinal cord to two afferent volleys. Amev. J . Physiof., 1934, 108: 307-321. HUNT,C. C . and MCINTYRE,A. K. An analysis of fibre diameter and receptor characteristics of myelinated cutaneous afferent fibres in cat. J . Physiol. (Lond.), 1960, 153: 99-112. KOKETSU, K. Intracellular potential changes of primary afferent nerve fibres in spinal cords of cats J . Neurophysiol., 1956, 19: 375-392. KOSTYUK,P. G. Features of polysynaptic excitation and inhibition in individual motoneurones. Sechenov Physiol. J. USSR, 1960, 46: 471-482 (English translation). LAPORTE,Y . and LLOYD,D. P. C. Nature and significance of the reflex connections established by large afferent fibers of muscular origin. Amer. J. Physiol., 1952, 169: 609-621. LIDDELL, E. G. T. and SHERRINGTON, C. Further observations on myotatic reflexes. Proc. roy. SOC.B, 1925, 97: 267-283. LLOYD,D. P. C. A direct central inhibitory action of dromically conducted impulses. J. Neurophysiol., 1941, 4 : 184-190. LLOYD,D. P. C. Neuron patterns controlling transmission of ipsilateral hind limb reflexes in cat. J. Neurophysiol., 1943,6: 293-315. LLOYD,D. P. C. Facilitation and inhibition of spinal motoneurones. J. Neurophysiol., 1946, Y: 42 1 4 38.
LUNDBERG, A. Integrative significance of patterns of connections made by muscle afferents in the spinal cord. XXI. International Congress o j Physiological Sciences. Buenos Aires, 1959 : 100-105. LUNDBERG, A. and WINSBURY, G. J. Selective adequate activation of large afferents from muscle spindles and Golgi tendon organs. Acta physiol. scand., 1960, 49: 155-164. MARK,R. F. and STEINER, J. Cortical projection of impulses in myelinated cutaneous afferent nerve fibres of the cat. J. Physiol. (Lond.), 1958, 142: 544-562. PAINTAL, A. S. Functional analysis of group 111 afferent fibres of mammalian muscles. J. Physiol. (Lond.), 1960, 152: 250-270. PAINTAL, A. S. Participation by pressure-pain receptors of mammalian muscles in the flexion reflex J . Phypiol. (Lond.), 1961, 156: 498-514. RENSHAW, B. Influence of discharge of motoneurons upon excitation of neighbouring motoneurons. J . Neurophysiol., 1941, 4: 167-183. RENSHAW, B. Reflex discharges in branches of the crural nerve. J . Neurophysiol., 1942, 5: 487-498. RENSHAW, B. Central effects of centripetal impulses In axons of spinal ventral roots. J. Neurophysiol., 1946, 9 : 191-204. SHERRINGTON, C. S. Remarks on some aspects of reflex inhibition. Proc. roy. SOC.B, 1925, 97: 519-545.
TAKEUCHI, A. and TAKEUCHI, N. Electrical changes in pre- and postsynaptic axons of the giant synapse of Loligo. J . gen. Physiol., 1962, 45: 1181-1193. TONNIES, J. F. Reflex discharge from the spinal cord over the dorsal roots. J . Neurophysiol., 1938, I 378-390. WALL,P. D. Excitability changes in afferent fibre terminations and their relation to slow potentials. J. Physiol. (Lond.), 1958, 142: 1-21.
INHIRITORY ACIIONS IN THE SPINAL CORD
19
DISCUSSION
F. BREMER:
I must thank our colleague Eccles for having mentioned the former investigations carried out by us with Valentine and Bonnet, in the interpretation of central inhibition in spinal batracians preparations as blocking of afferents, attributing this block to interneuronal activity. Experiments on barbiturized frogs, which was starting-point of our analysis, consisted of suppression of the reflexogenic effect of a pair of afferent volleys (effect based on temporal summation) by means of a conditioning afferent volley through either the same nerve fibres or those of a contralateral nerve, separated from the reflexogenic pair by an interval sufficient to avoid addition to the temporal summation effect. Of course, the conditioning afferent volley had no visible reflexogenic effect; this eliminated any explanation on the basis of motor neuron refractoriness. Inhibition of the reflex thus produced was not abolished by strychnine, which could even increase it. The curve of evolution of the inhibitory process revealed a progressive development, followed by slow, regular dissipation. Generally, the depression of the reflex was still manifest 200 msec after inhibitory conditioning stimulation. In our experiments the slow dorsal root potential produced by the inhibitory and reflexogenic stimuli, was supposed to give information on the spinal interneuronal activity. Eccles - taking up the suggestion of Barron and Matthews -suggests that the dorsal root potential is the expression, by electrotonic diffusion, of a depolarization of intraspinal terminations of afferent fibres, actively produced by the inhibitory impulses; this means that this depolarization, which blocks synaptic transmission of afferent impulses of the conditioningvolley, is the instrument of the inhibitory effect. But when we consider the dorsal root potential corresponding with the test stimulus, it seems to express an inhibitory effect produced by the conditioning stimulus. There is a difficulty here. However this may be, the investigations of our colleague Eccles et a/. will doubtless elucidate the significance - so far somewhat enigmatic - of the dorsal root potentials! BREMER,F. et BONNET,V. Contribution a I’etude de la physiologie gkntrale des centres nerveux. L‘Inhibition rtflexe. Arch. Znt. Physiol., 1942, 52: 153-194. B. H. The interpretation of potential changes in the spinal cord. BARRON,D. H. and MATTHEWS, J . Physiol. (London), 1938, 92: 276-321.
H. H. JASPER:
To clarify my own thoughts on the rapid advances made by Prof. Eccles in his conception of mechanisms of inhibition, I would like to ask for more details concerning the present conception of the structural, anatomical, chemical or neurohumoral character of presynaptic synapses. Also would Prof. Eccles please elaborate on the mechanism whereby a depolarisation of presynaptic fibres - which should increase excitability - results, instead, in a decreased liberation of the excitatory transmitter substances.
J. L. O’LEARY:
I understood from your presentation, Prof. Eccles, that presynaptic inhibition involves depolarization. I also heard you say that you hypothesized a retardation in the passage of chemical transmitter. In the latter instance is it possible that the same axon might, under different circumstances, provoke an excitatory or an inhibitory effect, in one situation retarding and in the other facilitating passage of transmitter across the synaptic cleft?
20
J. C. ECCLES
J. M. BROOKHART:
I should like to make sure that I understand correctly the basic evidence on which you base your inference that axo-axonic synapsesareresponsible for the primary afferent terminal depolarisation. Is it correct that this evidence consists principally in the differences which can be detected between the potential change recorded from the inside and outside of the afferent fibre?
K.-E. HAGBARTH: It is still true, I suppose, that a conditioning group I afferent volley from a flexor muscle causes direct facilitation of the monosynaptic reflex elicited from a synergetic flexor. But you are now saying that in this case there is also a pre-synaptic inhibitory interaction involved. I want to ask you about the time relationships between the pre- and post-synaptic effects in this case. If the effects coincide, the presynaptic inhibition apparently tends to conceal the postsynaptic facilitation. But if instead the test reflex is elicited from an extensor muscle, the presynaptic inhibition will add to the postsynaptic inhibitory effect in this case. Is that correct?
W. R. ADEY: In connection with Professor Eccles’ very interesting hypothesis of a presynaptic inhibitory mechanism, my colleagues and 1 were interested in a similar possibility several years ago. We investigated the effects of certain convulsant hydrazides, particularly thiosemicarbazide, in inducing seizures in cerebellar cortex. It was our conclusion, reported elsewhere (Adey et al. 1960) that these effects might be explained by interference with normal inhibitory activity in presynaptic terminal mechanisms. Has Professor Eccles investigated the effects of thiosemicarbazide or other convulsant agents on such presynaptic inhibitory mechanisms in the spinal cord?
ADEY,W. R.,DUNLOP, C. W., KILLAM, K. F. and BRAZIER,M. A. B. In: Inhibition in the nervous system and gamma-aminobutyric acid. Pergamon, London, 1960: 3 17-323.
R. JUNG: The most interesting features of presynaptic inhibition at sensory terminals, besides the mechanism of its postulated axon-axonal synapses seem to be its long duration and its large dipole potentiul Jield with electrotonic spread along the posterior roots. Thus Sir John Eccles presents us with a very simple explanation of the slow spinal and dorsal root potentials and possibly also of other slow waves at higher cerebral levels. For a physiologist who believes in the neurone theory it is a great comfort that Sir John found interneurones involved in this inhibition. This again shows similarities with inhibitory mechanisms and slow waves in the cerebral cortex. But we might speak about possible applications of Sir John’s discoveries to higher levels in the discussion of inhibition this afternoon. For the spinal cord it now appears well established that presynaptic inhibition is not a special and exceptional mechanism arising from flexor afferents, but is rather a general regulation of afferent systems. Functionally it might not only be important for negative feedback and homeostasis or for reciprocal innervation, but also for sensory information: It may depress the uninteresting background of afferent impulse flow and enhance the stronger and more important messages and thus work similarly to contrast mechanisms as Sir John has suggested. It seems possible that these spinal regulations of afferent influx might also be under the control of upper levels by cerebral efferents running to interneurones in the cord which subserve presynaptic inhibition. This might be as worthwhile investigating as it was for the Renshaw inhibition and for the gamma system. A dependence on higher centres might also elucidate the still obscure mechanisms of some clinical neurological syndromes. If presynaptic inhibition is unaffected by strychnine which kills postsynaptic inhibition, this would mean survival or even enhancement of presynaptic inhibition in the CNS, poisoned by strychnine. Under such conditions this surviving inhibition would then stand out from a background of released excitation. The large strychnine waves, found in the spinal cord and in the cortex, might then be also
INHIBITORY ACTIONS IN THE SPINAL CORD
21
an electrical manifestation of presynaptic inhibition. The slow rhythmical components which these strychnine waves contain (besides rapid neuronal discharges, probably caused by lack of postsynaptic inhibition) might also be the result of this presynaptic inhibition. This fits in with the long pauses of neuronal discharges, which we have found following “strychnine waves” and the associated high frequency neuronal bursts in the cortex. Not only normal spinal cord potentials or Bremer’s spinal strychnine rhythms but possibly also normal and abnormal brain waves may appear in a new light as a result of Sir John’s findings. I should like to know whether Sir John would find it permissible to apply these rather loose hypotheses and wide applications of his findings to other levels of the CNS. J. C . ECCLES’s replies
To F. Bremer I would agree with Professor Bremer’s implication that the inhibition investigated by Dr. Bonnet and himself is an example of presynaptic inhibition; they were correct in associating this inhibition with the dorsal root potential. It may help to clarify the situation if it is realized that the dorsal root potential generated by an afferent volley is too late to exert any depressant action on the central excitatory action of that afferent volley. To H . H . Jasper I would distinguish between the transmitters responsible for presynaptic and postsynaptic inhibition on two grounds. Firstly, the two types of synapse have different pharmacological properties: strychnine is a very effective depressant of postsynaptic inhibition, acting probably competitively on the receptor sites, just as curare does at the neuro-muscular junction, whereas it has no action on presynaptic inhibition; by contrast presynaptic inhibition is not affected by strychnine, but is depressed by picrotoxin, which has no action on postsynaptic inhibition. The other ground for discrimination is that the synapses responsible for postsynaptic inhibition act by causing a membrane hyperpolarization, whereas presynaptic inhibition is produced by depolarizing synapses. This depolarizing action can also be distinguished from that of the excitatory synapses, because the latter synapses are not affected by picrotoxin, and also because the equilibrium potential is probably quite different, being probably no more than 20 mV depolarization for the presynaptic inhibitory synapses and virtually at zero potential for excitatory synapses, i.e. at about 70 mV depolarization. To J. L . O’Leary The question really relates to the equilibrium potential for the presynaptic action on the primary afferent fibres. Dr. O’Leary is correct in suggesting that it should be possible to depolarize these fibres beyond the equilibrium potential so that the synapses on these fibres have a hyperpolarizing action, with a consequent potentiation of their synaptic excitatory action on motoneurones. We have attempted to do this, but have never succeeded in producing a reversal. We presume that we have not yet been able to produce a sufficient current density to make the required change in the membrane potential of the primary afferent fibres to beyond the equilibrium potential.
To J. M . Brookhart Yes, the most direct evidence is provided by the difference between the potentials recorded inside and just ouside a primary afferent fibre. This recording procedure gives just as reliable a measure of the presynaptic depolarization as does the intracellular recording from motoneurones for postsynaptic depolarization. However, in addition, the presynaptic depolarization is very effectively demonstrated by the increased excitability that is observed with brief testing electrical pulses. This increased excitability occurs along the whole intramedullary course of the primary afferent fibres, which establishes that there must be active depolarizing foci on these primary afferent fibres; for, if the depolarization were passively produced in the fibre by current flow, there must be regions of hyperpolarization with associated depressed excitability; and this is never observed. Finally, the dorsal root potential and the dorsal root reflex discharge provide further evidence of a large and prolonged presynaptic depolarization. All these types of investigation show that there must be a very large transport of electric charge across the surface membranes of the primary afferent fibres. In the vertebrate, chemically transmitting synapses are the only devices that are known to be capable of thiselectrical transport, which occurs by the increased ionic permeability that the chemical transmitter produces in the subsynaptic membrane.
22
J . C. ECCLES
To K.-E. Hngbnrtli It can be observed that these opposed actions do occur, but of course they have quite different time courses. The facilitation of group Ia monosynaptic action is maximum with simultaneous volleys in the nerves to two flexor muscles of the same joint; and has disappeared at an interval of 20 msec. On the other hand the presynaptic inhibitory action of a group I volley from a flexor muscle has a much slower time course, the latency of action being at least 4 msec with a maximum a t about 20 msec and a total duration in excess of 200 msec. Besides this difference in time course, there is of course a difference in distribution. The group Ia facilitation occurs only for the flexor muscles acting a t the same joint, whereas there is no such restriction in the field of action of presynaptic inhibition. There would also be a temporal discrepancy of about the same order between the myotatic and presynaptic inhibitions of extensor niotoneurones. To W. R.Adey Unfortunately picrotoxin and strychnine are the only convulsant drugs that we have so far tested on presynaptic inhibition, though there have of course been several more extensive investigations on postsynaptic inhibition. Our first postulate is that convulsant drugs are of two types: strychnine-like, depressing the postsynaptic inhibitions; and picrotoxin-like, depressing the presynaptic inhibitions. However, there may well be a third classof convulsants that act not on the inhibitions, but for example by intensifying synaptic excitation. Addendum: I have been unable to find any reference to presynaptic terminals and inhibition in the paper cited by Dr. Adey.
To R. Jung 1 am in complete agreement with Professor Jung’s speculations on the possibility that presynaptic inhibition may occur also a t the higher levels of the nervous system, and hence provides a likely cause for the slow potential waves that characterize so much central activity. Experimental investigation should be designed in order to test these suggestions, as also the suggestion that cortical influence on the spinal cord may in part be due to presynaptic inhibition. The long time course of the presynaptic depolarization can, I think, be most probably explained by postulating that the presynaptic terminals are encapsulated so that the diffusion of the transmitter substance is greatly slowed. We are already familiar with a duration of transmitter (acetylcholine) action on Renshaw cells that is as long as 50 msec in the presence of cholinesterase and up to 2 sec when it is inactivated. One suspects that this encapsulation is a function of glial cells and that it is important in prolonging the action of the transmitter to a duration appropriate for its function, and also in giving a more efficient usage of the transmitter substance. The postulate of a diffusion barrier is supported by the observation that nenibutal greatly prolongs the duration of the presynaptic depolarization, for it seems likely that anaesthetics would be adsorbed on surfaces and so further clog the diffusional channels in the barrier. Alternatively, the prolongation by nembutal could be ascribed to an interference with an enzyme system that destroys the transmitter.
Recurrent Inhibition as a Mechanism of Control RAGNAR GRANIT Nobel Institute for Neurophysiology, Karolinska Institutet, Stockholm (Sweden)
It will be known in this circle that Renshaw, in making the experiments on recurrent inhibition (1941, 1946), which have served as a starting point for much later work, also saw recurrent excitation. This has since been studied by Wilson and his co-workers (Brooks and Wilson 1959; Wilson 1959) and everything points to its being the kind of disinhibition (Wilson et al. 1960) that Hartline and Ratliff (1956) first described in the Limulus eye. If I restrict myself to recurrent inhibition, this is because it is far morepotent on tonic extensors. In decerebrate preparations, in fact, we hardly ever see recurrent excitation which influences the motoneurones of the ankle extensors, unless it be identified with the rebound which is a common occurrence. Our work has been done with tonically responding extensor neurones and we were led into this field of study by our original interest in the gamma control of tone, which naturally leads to a study of the various aspects of the physiological mechanisms which center round the maintenance of longlasting tonic discharges and contractions. It was in this way that we encountered the fact that it was practically always possible to inhibit our tonic motoneurones, isolated in root filaments, by antidromic stimulation of the rest of the ventral root, while the phasic ones quite often proved highly resistant to this influence (Granit et al. 1957). We could never influence gamma motoneurones by antidromic stimulation which induced recurrent inhibition. These results were soon confirmed by others (Kuno 1959; Eccles et al. 1960 a ; Eccles et al. 1960 b). Because tonic motoneurones can be set to discharge at semistationary frequencies over a considerable time, this preparation seemed to offer excellent opportunities for trying to build up something quantitative out of a reflex act, including the principles of regulation shown by recurrent inhibition. It has never been shown that a maintained reflex discharge faithfully reflects the level of depolarization of the active region of the membrane of the motoneurone. In order to examine such a postulate, it was necessary to begin by assuming that it is a valid assumption and then to proceed to prove or disprove it. The idea is, of course, merely the well-known concept of the generator potential. Katz’s (1950) experiments with muscle spindles showed that the discharge rate is proportional to the amplitude of the slow potential change at the terminals. This has since been observed in experiments with other sense organs (MacNicholl956; Fuortes 1958,1959; Loewenstein 1960; Wohlbarsht 1960). For motoneurones we have the evidence obtained by Frank and Fuortes (1960 and personal References P . 34-JS
24
R. GRANIT
communication) that, when a stimulus is applied through an intracellular electrode, the discharge rate is proportional to the depolarizing current over a considerable range. Eccles et al. (1954) have shown that recurrent inhibition is of the polarizing type. In the experiments the reflex discharge is obtained by tetanizing muscular afferents, some of which may be also inhibitory and may contribute a certain amount of polarizing current Ppol.But the large spindle afferents would cause depolarization P d e p and this would lead to a tonic discharge in proportion to the net depolarizing current (or depolarizing pressure). Hence we have the fundamental equation for discharge frequency F = k(Pciep f'po~) (1)
+
in which k is the proportionality constant, and the expression within brackets represents the net depolarizing current. It is possible to vary F by increasing the rate or strength of the afferent stimulation, so that we obtain, for any one motoneurone, a number of normal values Fn to put into the equation. In order to test our assumption that inhibition and excitation sum algebraically, the next step is to test, as is shown in the diagram of Fig. 1, by a constant antidromic tetanus of the rest of the ventral root. The stimulus frequency must be well above the natural firing rate of motoneurones. This means the introduction into the equation of a constant quantity F p 0 l . Hence we obtain the inhibited frequency Ft, which is
Fi
= k(Pdep
+
Ppol
+
P'pol)
(2)
DR
Fig. 1 Experimental arrangement. Stimulating electrodes on severed gastrocnemius nerves and on the ventral root (VR), from which a single fibre (S) has been isolated. DR, dorsal root; the recurrent circuit is diagrammatically shown within the motoneurone pool ( M N ) . Below, record of spike discharging in response to repetitive stimulation of the gastrocnemius nerves; antidromic stimulation at 48/sec inserted at moment marked by dot; time marker, 100 c/sec (Granit and Renkin 1961).
RECURRENT INHIBITION AS A MECHANISM OF CONTROL
25
from which it follows that Fn
-
Fi
=
-kP'p,1
(3)
Thus, if reflex firing can really be treated quantitatively on the assumptions made, whatever the rate of Fn (above a certain minimum), the difference between it and the inhibited rate Ft should be a constant. The appended diagram (Fig. 2) perhaps helps us to understand what this means.
I
Rate of discharge ~
Fig. 2 Diagrammatic. Depolarizing pressure (=Paep Pp0l)plotted against the rate of discharge of a single motor cell responding to it. The curve marked F, shows their relationship without the addition of experimental recurrent inhibition. When RI, which is a constant amount of recurrent inhibition (= P'poi),is subtracted, the curve shifts to the dashed line marked Ft.
+
The firing rate is plotted on the abscissa against the net depolarization on the ordinate (upper curve). If we then subtract a constant amount of recurrent inhibition (RI in the graph), this amounts to removing a constant amount of depolarizing current and the inhibitory curve will be shifted downwards by a constant amount. Actually the experimental results have been plotted as in Fig. 3. Fa will be found on the abscissa and Ft on the ordinate. Now, if there were no recurrent inhibition, Fa would always be equal to Ft and the curve would be a 45" line with its proportionality constant 1.0. This curve is the upper one in the graphs of Fig. 3. The lower curve, obtained by the method of least squares, is drawn through the experimental values, and it is clear that our assumptions mean that it can only be moved downwards by a fixed amount, It should, therefore, also possess a regression coefficient of 1.0. This is seen to be so. The regression coefficient is given in the left upper corner; below it is the standard error, which is defined as the standard deviation divided by the square root of the number of observations ( N ) . With eighteen cells and a total of 470 References D, 34-35
26
I<. GRANIT
30 -
25-
20 -
u15-
10-
Fig. 3 Decerebrated cat. Single tonic motor cell. Two experiments plotted to show F , as a function of F, and the regression line drawn through the data. The upper line represents F, = Fi or a 45” line with its proportionality constant = 1.00. The theoretical value of the regression coefficient should also be 1.00. The actual value in both these experiments is 1.01 (upper, left corner). The figure below shows the standard error (Granit and Renkin 1961).
observations, sixteen cells gave a regression coefficient which fell inside the limits 0.9-1 . l . If we regard all the experiments with the eighteen niotoneurones as being part of one experiment devoted to testing the validity of our equation, the regression coefficients should be weighted with respect to the number of observations ( N ) in order to obtain their true average value. This proved to be 0.996, which may well be called an unexpectedly good approximation to the theoretical 1.0. Apparently, in order to have cancelled out so well, systematic errors must have been small. The average inhibitory effect in the whole material was Fn - Fi
5.5 impulses/sec
z-
Thus it seemed possible, on the assumptions used, to turn reflexology into a simple piece of algebra. As some of you may remember, Sherrington used to think of inhibition and excitation as truly opposite processes which sum algebraically upon the motoneurone, but the time was not yet ripe for putting these general notions to an experimental test. The next question is: how does the frequency of the antidromic stimulation influence a constant firing rate of a motoneurone. This question is relevant only within the range of firing frequencies in reflex action. Below 5 impulses/sec no tonic motoneurones fire steadily and a large number does not seem to be able to maintain a discharge below lO/sec. At these low rates of stimulation each antidromic shock leads to a pause which is generally succeeded by a brief rebound increase of frequency, which compensates for the loss of spikes during the pause. Thus the net effect in terms of frequency may
21
RECURRENT INHIBITION AS A MECHANISM OF CONTROL
be zero. It is well known that the discharge of the Renshaw cells in response to isolated shocks is a high-frequency burst which gradually disappears within some 30-50 msec, but, when the rate of repetition of the antidromic stimuli approaches some 8-10 impulses/sec, the whole character of the discharge changes. The tail drops out and the shock produces only a couple of impulses. From then onwards, up to some 30-40 shocks/sec, the effect of the antidromic stimulation measured as Fn - Fi tends to be, as Fig. 4 shows, proportional to the shock frequency. This is the best we have been able to do at the moment.
6-
.-
5-
I
4-
u.
kc
321-
4
I
10
I I 20 30 Antidromic shocks/sec
I 40
1 50
Fig. 4 Experiment as in Fig. 3 but with F,, kept constant and the frequency of antidromic stimulation varied (nbscissa). On the ordinate the effect on the recurrent inhibition in terms of Fn - Fi (Granit and Renkin 1961).
It is a striking fact that both the results presented above form a perfect analogy to the findings of Hartline and his co-workers (Hartline 1949; Hartline and Ratliff 1956) with lateral inhibition in the eye of the horseshoe crab Limulus. The amount of inhibition is proportional to the firing rate of the inhibiting ommatidium, but is independent of the frequency of discharge of the inhibited ommatidium. It seems to be of considerable interest that it is the frequency of discharge which emerges as the decisive factor and lends itself to treatment in terms of simple linear equations. I n LimuZus, Hartline and his group have explored these possibilities in great detail. Because lateral inhibition is mutual and proportional to the firing rate of the two cells concerned, the cell which discharges more slowly will inhibit its fellow less than the latter inhibits in return. The sum total of the two effects can be obtained not only by experimentation, but also by solving two simultaneous equations after their constants have been obtained by experimentation. Mutual inhibition implies, in effect, a mechanism of sensory contrast. In a similar way recurrent inhibition will, as is easily understood from a consideration of Fig. 5, act to produce motor contrast. In Fig. 5 the abscissa is the frequency of stimulation of the motor nerve to the soleus and the ordinate is the isometric tension of the muscle. For this nerve-muscle preparation - values from Matthews (1959) - the curve drawn between the experimental References p . 34-35
28
R . GRANIT
points shows that there is a critical region, which begins around 5-10 impulses/sec, from which tension, as a function of stimulus frequency rapidly rises asymptotically to reach a maximum around 30 shocks/sec. This is the firing range of the soleus motoneurones (Denny-Brown 1929 ; Granit 1958). 2MOy
Stimuli 1 3 s
Fig. 5 Curve with observations in filled circles (from Matthews 1959) which illustrates the isometric tension (ordinate) plotted against the stimulus frequency to the muscle nerve (abscissa) of the soleus at an initial length determined by a tension value below 50 g. It is assumed [see Text)thatthe abscissa also represents the natural firing frequencies across the soleus motoneurone pool, the range being from 25-5 impulses/sec from the centre to the edge of the pool. An approximate idea of what the experimental curve would have looked like in the absence of recurrent inhibition is obtained by subtracting for all points in the curve 6 impulses/sec. This is the upper curve. The lower one is the difference between the two others (Granit and Renkin 1961).
Let us now imagine that the scale of the abscissa also represents a variation in the firing rate from the edge to the centre of the soleus motoneurone pool. The most active focal neurones which fire at high rates are in the centre. These will dominate and so let us, for a first approximation, subtract their constant recurrent inhibition of 6 impulses/sec. This is the upper curve. The lower one is the difference between the two and it shows that the recurrent inhibition has been most potent at the lower frequencies of discharge coming from cells which contribute but little to the tension. If we add the fact that the recurrent inhibitions are mutual, we realize that the motoneurones which fire at low frequencies have been able to suppress those in the centre of the pool but little, while they themselves have been struck by the full inhibitory force of the neurones which fire at high rates. I n this sense there is motor contrast. The just liminal fringe of motoneurones will be suppressed and the active focus will be emphasized. In motor contrast there is another mechanism of considerable importance which tends, like the one just explained, to suppress activity within the fringe. I shall take but one experiment to illustrate it. In Fig. 6 the abscissa is the running time and the ordinate is the impulse frequency of a single tonic cell responding to stretching of the ankle extensors. It is seen that the discharge soon settles down to a constant value for the period between the two horizontal lines. This discharge has been recorded as open circles. The small oblongs on the abscissa show the times when the recurrent
RECURRENT INHIBITION AS A MECHANISM OF OCNTROL
29
Fig. 6 Motoneurone responding to a steady pull of 15 mm on the ankle extensors. Tetanic antidromic inhibition at 114/sec inserted for 0.7 sec at regular intervals, as marked by rectangles on the abscissa (running time). Frequency of discharge constant between the two parallel horizontal lines. 0, number of impulses (imp/sec) during the periods of recurrent inhibition. Inset: original records at moments marked 1, 2 and 3 in the diagram. Note that, when delayed recovery after recurrent inhibition begins, discharge frequency fails to reach its original level (at this rate of repetition of the antidromic stimulation periods). Discharge stopped for good with the last period of stimulation, having been five times temporarily silenced (Granit and Rutledge 1960).
inhibition was inserted. The points 1, 2 and 3 refer to moments cut out from the film and given in the inset. Consider now our equation. Because the impulse frequency is constant, the depolarizing pressure is also constant. Recurrent inhibition must therefore remove a constant fraction of it, as, indeed, it will often do, provided that there isenough excitatory drive. But in this situation the effect of the recurrent inhibition increases along the curve. In the original paper (Granit and Rutledge 1960) we have given the evidence for our conclusion that there must be a certain surplus or margin of excitatory drive before a neurone can withstand inhibition and fully recover afterwards. If the term be forgiven me, I could call it presynaptic excitation. We have taken it to express the idea that a constant depolarizing pressure can be maintained with a large or a small margin of support. When the margin decreases too much, recovery from the inhibition at first is delayed and later becomes difficult, and finally impossible. This is seen in the inset of Fig. 6, which gives the moments 1 , 2 and 3 in the discharge. It is clear from Fig. 6 that, whenever the excitatory surplus is low, the cell falls an easy prey to recurrent inhibition and we have reason to assume that this at first happens with the feebly supported neurones in the fringe. Thus the mechanism here described will contribute to the enhancement of motor contrast as established by the algebraical rules for mutual inhibition described in the previous paragraphs. So far we have been reasoning as if the recurrent inhibition were an automatic circuit which operates, in spite of its intercalated internuncial, the Renshaw cell, as the lateral ramifications of the Limulus optic nerve fibres do. But the spinal cord is like a References 0 . 34-35
30
R . GRANIT
grand piano which requires a virtuoso to handle its layed-out connexions. We must therefore proceed to give the player in the brain a chance of showing what he can do. This we did by studying not only individual Renshaw cells, but also the full recurrent circuit. The first line of approach differed but little from the one hitherto described. It was part of a set of experiments devoted to the study of the frequency limitation in the motoneurones (Granit et al. 1960). The second line was the approach with conventional microcapillaries (Haase and Van der Meulen 1961). The technique used in the first approach emerges from Fig. 7. I t looks very much like Fig. 1, but differs from it in that the discharge from the single fibre S is connected through an amplifier to the stimulator Stim. for the rest of the ventral root. This means that we make an extra recurrent circuit emphasize the inhibitory effect in proportion to the rate of firing of the single cell selected. Below is shown an original record of' a stretch reflex (I), and in 2 the same reflex is repeated with the recurrent circuit connected. It is seen that the discharge is slowly strangled, inhibition increasing in a cumulative fashion. In the experiments it proved an advantage to cut the muscle nerve and to use electrical stimulation of the muscular afferents instead of pull on a muscle
B
Gostroc- soleus
stretch
. 05 sec
Fig. 7 Arrangement differing somewhat from that of Fig. 1 . Here the amplified spike S from the single, tonic ventral horn cell is connected to the stimulator Stim., which therefore fires antidromically into the rest of the ventral root, VR, at the rate of discharge of S. Below, two records of stretching with a length-recorder, indicating the change in the muscle. The upper one is the control; in the lower one the antidromic shock is triggered by S in the manner described and the spike is silenced (Granit and Rutledge 1960).
RECURRENT INHIBITION AS A MECHANISM OF CONTROL
31
It is then easier to obtain the amount of surplus excitation needed to make the recurrent effect analyzable. If again we plot F6 against Fn with the shock triggering the recurrent feedback, results such as those shown in Fig. 8 are obtained. The relations are linear, but they vary a little in slope because of the variation in the firing frequency of the triggering cell.
20-
10
D
f
/' I
/
/ I
I
/
",
////
I
I
/
/
I
l
l
32
R. GRANIT
Fig. 9 Afferent stimulation maintained for 50 sec, during which time the antidromic shock was locked to the discharging spike for brief intervals only, as tests. Values from such tests were averaged over before and after one series during stimulation four consecutive 10 sec periods from control runs (o), ( 0 )of the point in the frontal part of the anterior cerebellum illustrated as inset. This well isolated spike had a n exceptionally high initial frequency of discharge of around 40/sec, but fell, during maintained afferent stimulation, to values between 20-25 impulses/sec. The point stimulated proved to be in the anterior lobe of the cerebellum (electrocoagulated), as is shown by the inset, in which it is marked by a pointer. The stimulus frequency during the experiment was 300/sec, with a strength of 0.45 V through the tip of the thin coated needle against the ground. By increasing the stimulus strength the discharging spike itself was ultimately inhibited. The effect of stimulation reduces the recurrent inhibition by 4.2 impulses/sec. Co-ordinates scaled in impulses/sec (Granit et al. 1960).
Finally let us look briefly at some of the results obtained by Haase and Van der Meulen (1961), who supplemented this work with a study of single Renshaw cells and used conventional microcapillaries to locate them. The method of choice proved to be to maintain an automatically repeated antidromic test shock while Renshaw cells, identified as belonging to the gastrocnemius nerve, were studied. We see the characteristic burst of such cells in the two figures taken from the paper (Fig. 10 and 11). For definition the latent period of the burst had to be 0.6-0.8 msec and the discharge had to be steady, in order to provide a reliable background for the effect of the stimulation of either the cerebellum or else the ventro-medial area of the reticular formation in a region extending from the superior colliculi to the medulla oblongata. Fig. 10, A and C (upper) show the effect of a weak antidromic shock eliciting 1-2 spikes, A before, C after, the record B that was taken during cerebellar stimulation of the point marked by a square in the section. It should be noted that cerebellar conditioning in B was by a single shock. Facilitation of the test response was considerable. In the lower part of the same figure A is the control and in B recording had been preceded by a tetanus, which is repeated also in the interval between B and C. Tests take place at a frequency of I/sec and the records are from successive sweeps.
RECURRENT INHIBITION AS A MECHANISM OF CONTROL
A
C
33
D
100 cl sec
Fig. 10 Deafferented decerebrate cat. Single Renshaw cell. Upper p a r t : A , the antidromic test shock to the gastrocnemius nerves was adjusted so that only one or two discharges appeared in control. B , cerebellar conditioning by a single shock to point marked by a square in section, results in strong facilitation. C, after cessation of the cerebellar conditioning the original rate of discharge re-appeared. Lower part: successive sweeps at rate l/sec; the same experiment continued with tetanic stimulation at 48/sec for 0.5 sec in intervals A-B and B-C. Thus A and D are controls, but the effect lingers on in D (Haase and Van der Meulen 1961).
It is clear that the tetanus has mobilized a facilitation of the Renshaw cell which is maximal in B and lingers on in C after cessation of stimulation. Fig. 11 shows in a similar manner inhibition of a submaximal discharge by tetani to the medial point marked by the medial filled circle in the inset. In these experiments spontaneous activity was continuously followed by another channel which wrote on a stationary spot on moving paper, but there was surprisingly little of it, the explanaion probably being that the spinal cord of these decerebrate animals was de-afferented
Fig. 11 As Fig. 10, but another experiment. Tetani, as in the lower part of Fig. 1. These were delivered to the medial point in inset. A and C, controls. B, test preceded by tetanus; discharge is inhibited (Haase and Van der Meulen 1961). References
p.
34-35
34
R. GRANIT
over the lumbar and sacral segments. Only occasional and highly irregular bursts were seen. It is important to note that none of the experiments showed the spontaneous activity influenced by the supraspinal stimuli used, nor was it ever possible to detect a supraspinal effect on Renshaw cells, unless they were actually tested antidromically. The effects were therefore never in themselves supraliminal. All these facts seem to demonstrate conclusively that the supraspinal effects on the Renshaw cells were not indirect in the sense that they presupposed an activation of the motoneurones. We therefore conclude that the recurrent circuit is at the disposal of the supraspinal centres. The limitation of the study to Renshaw cells of the ankle extensors and the central regions mentioned also gives some indication of the general functional significance of the supraspinal mechanism in the control of the mechanisms of tone. An interesting finding was that one of the Renshaw cells, which could be driven both antidromically and orthodromically (from the dorsal root), so that it thus allowed a comparison of both the tests, was, when it was driven orthodromically, facilitated only from a point in the ventromedial part of the reticular formation. This finding, and the absence of direct effects on the Renshaw cells, suggests that the supraspinal mechanisms of control operate through fairly specific internuncial organizations. REFERENCES BROOKS, V. B. and WILSON,V. J. Recurrent inhibition in the cat’s spinal cord. J. Physiol. (Lond.), 1959, 146: 380-391. DENNY-BROWN, D. B. On the nature of postural reflexes. Proc. roy. SOC.B, 1929, 104: 252-301. ECCLES, J. C., ECCLES, R. M., IGGO,A. and LUNDBERG, A. Electrophysiological studies on gamma motoneurones. Acta physiol. scand., 1960a, 50: 32-40. ECCLES, J. C., FA-IT,P. and KOKETSU, K. Cholinergic and inhibitory synapses in a pathway from motor-axon collaterals to motoneurones. J. Physiol. (Lond.), 1954,126: 524-562. ECCLES, R. M., IGGO,A. and ITO,M. The distribution of recurrent inhibition among motoneurones. J. Physiol. (Lond.), 1960b, 153: 49-5OP. FRANK,K. A. and FUORTES, M. G. F. Accommodation of spinal motoneurones of cats. Arch. ital. Biol., 1960,98: 165-170. FUORTES, M. G . F. Electrical activity of cells in the eye of Limulus. Amer. J. Ophthal., 1958, 46: 210-223. FUORTES, M. G. F. Initiation of impulses in visual cells of Limulus. J. Physiol. (Lond.), 1959, 148: 14-28. GRANIT,R. Neuromuscular interaction in postural tone of the cat’s isometric soleus muscle. J. Physiol. (Lond.), 1958, 143: 3 8 7 4 2 . GRANIT,R. AND RENKIN,B. Net depolarization and discharge rate of motoneurones, as measured by recurrent inhibition. J. Physiol. (Lond.), 1961,158: 461-475 GRANIT,R. and RUTLEDGE, L. T. Surplus excitation in reflex action of motoneurones as measured by recurrent inhibition. J. Physiol. (Lond.), 1960,154: 288-307. GRANIT, R., HAASE,J. and RUTLEDGE, L. T. Recurrent inhibition in relation to frequency of firing and limitation of discharge rate of extensor motoneurones. J. Physiol. (Lond.), 1960,154: 308-328. GRANIT,R., PASCOE, J. E. and STEG,G. The behaviour of tonic alpha and gamma motoneurones during stimulation of recurrent collaterals. J. Physiol. (Lond.), 1957, 138: 381-400. HAASE, J. and VANDER MEULEN, J. P. Effects of supraspinal stimulation on Renshaw cells belonging to extensor motoneurones. J. Neurophysiol., 1961, 24: 510-520. HARTLINE, H. K. Inhibition of visual receptors by illuminating nearby retinal areas in the Limulus eye. Fed. Proc., 1949, 8 : 69. HARTLINE, H. K. and RATLIFF,F. Inhibitory interaction of receptor units in the eye of Limulus. J. gen. Physiol., 1956,40: 357-376. KATZ,B. Depolarization of sensory terminals and the initiation of impulses in the muscle spindles. J. Physiol. (Lond.), 1950. 1 1 1 : 261-282.
RECURRENT INHIBITION AS A MECHANISM OF CONTROL
35
KUNO,M. Excitability following antidromic activation in spinal motoneurones supplying red muscles. J. Physiol. (Lond.), 1959,149:374393. LOEWENSTEIN, W. R. Mechanisms of nerve impulse initiation in a pressure receptor (Lorenzinian ampulla). Nature, 1960,188: 1034-1035. MACNICHOL, E. F. Visual receptors as biological transducers. In Molecular structure and functional activity of nerve cells. Amer. Inst. biol. Sci., Washington D.C., 1956,Publ. No. I . MATTHEWS, P. B. C. The dependence of tension upon extension in the stretch reflex of the soleus muscle of the decerebrate cat. J. Physiol. (Lond.), 1959,147:521-546. RENSHAW, B. Influence of discharge of motoneurons upon excitation of neighboring motoneurons. J. Neurophysiol., 1941,4 : 167-183. RENSHAW, B. Central effects of centripetal impulses in axons of spinal ventral roots. J. Neurophysiol., 1946,9:191-204. WILSON,V. J. Recurrent facilitation of spinal reflexes. J. gen. Physiol., 1959,42: 703-713. WILSON,V. J., TALBOT,W. H. and DIECKE, F. P. J. Distribution of recurrent facilitation and inhibition in cat spinal cord. J. Neurophysiol., 1960,23: 144-153. WOLBARSHT, M. L. Electrical characteristics of insect mechanoreceptors. J. gen. Physiol., 1960,44: 105-122.
DISCUSSION J. C. ECCLES: It is very encouraging indeed to see how the negative feed-back through Renshaw cells, that was initially developed on a purely qualitative basis, has, in Prof. Granit’s hands, become a subject for quantitative investigation, and moreover has given the opportunity for such an elegant theoretical development. I would also like to say that at first I was very much surprised by the evidence that cerebellar stimulation can cause changes, either an increase or decrease, in the discharge which antidromic volleys produce in Renshaw cells. However, on further consideration, it seems that this could occur by action on the interneurones that have now been shown by Curtis et al. (1961) to excite Renshaw cells by non-cholinergic synapses. In their experiments these interneurones were probably excited by high threshold muscle afferents acting through interneuronal pathways. A likely explanation of the effects of cerebellar stimulation would be that the descending volley from the cerebellum acted through these interneurones, so potentiating or inhibiting the Renshaw cell discharge. The potentiation is thus readily explicable. The inhibition would arise if the volley inhibited a tonic excitatory action of these interneurones on the Renshaw cell. I think these results of Prof. Granit are of great interest for they show for the first time that this homeostatic control of motorneuronal activity is susceptible to influences from higher centres. CURTIS, D. R., PHILLIS,J. W. and WATKINS,J. C. Cholinergic and non-cholinergic transmission in the mammalian spinal cord. J. Physiol (Lond.), 1961, 158: 296323. F. BREMER:
The observation on activation of the Renshaw type of spinal interneurons by impulses of cerebellar origin seems to me particularly interesting. While it emphasizes the difficult problem of the mediator, it suggests that the inhibition of spinal motor neurons produced by the stimulation of the anterior lobe of the cerebellum or of the “descending” reticular system could include a component of post-synaptic inhibition. This component of direct inhibition explains the fact that, although fundamentally strychnine-resistant, inhibition of cerebellar origin may be slightly reduced by the convulsive alkaloid. This explanation is not the only possible one, however, for when - as Terzuolo did - we examine the effect of stimulation of the anterior lobe of thecerebellum or ofthe inhibitory reticular formation on the rhythmic spinal potentials of strychnine tetanus in curarized cats, we observe not a trace of attenuation of the central inhibitory process.
36
R. GRANIT
TERZUOLO, C . Influences supraspinales sur le tetanos strychnique de la moelle Bpiniere. Arch. int. Physiol., 1954, 62: 179-196.
H. H. JASPER: This beautiful quantitative demonstration and theoretical treatment of the algebraic summation of excitation and inhibition in the control of tonic motoneurones of the cord, determining their rate of continuous discharge, is a model which must be applicable also to motoneurones of the cerebral cortex, if it could be tested there. In the studies of cortical motoneurones in the monkey during conditioned movements (to be mentioned later from experiments by Drs. Ricci and Doane) I have had the impression that one might make a distinction between “tonic” and “phasic” neurones also in the motor cortex. I would like to ask if there is further evidence that this may be true. However the patterns of cortical neuronal firing are often so complex that their firing rates must be determined by other than simple summation of excitatory and inhibitory states, unless one reduces this principle to a very brief time scale. It would seem that the time constants of neuronal nets and circuits and peripherally imposed temporal patterns may be of greater importance in most instances. One may well ask whether the “tonic” action of the ascending reticular formation may, however, fit into this theory, especially since ascending inhibitory effects seem to be as important as facilitatory effects from the reticular system. JASPER, H., RICCI,G. and DOANE, B. Microelectrode analysis of cortical cell discharge during avoidance conditioning in the monkey. Electroenceph. clin. Neurophysiol., 1960, Suppl. 13: 137-155.
J. C. ECCLES: According to the suggestions which I made earlier the inhibitory action of the cerebellar stimulation on the Renshaw cell response can be regarded as a dis-excitation, for it was postulated that it removed a tonic excitatory action evoked by a background of interneuronal activation.
R. JUNG: Prof. Granit has mentioned that there is a contrast mechanism in the motor system analogous to sensory contrast. This mechanism is so important in nearly all sensory systems that one would also suspect it to be at work in sensory-motor integration. The well known mechanism of lateral inhibition may explain most of these contrast phenomena together with reciprocal inhibition, as we have found in the visual system.
R. JUNGto H . H . Jasper:
I am very glad that Prof. Jasper has brought cortical neurones into this discussion. They provide good examples of these contrasting and reciprocal inhibitions. As Hering, Froehlich and Ebbecke used Sherrington’s spinal mechanisms for an explanation of sensory contrast phenomena we might be allowed here to proceed in a reverse manner and use cortical neurons to illustrate these spinal mechanisms. All who have worked with microelectrodes in the cortex have seen phenomena similar to the ones Prof. Jasper has ,just described: Inhibition of one neuron simultaneous with the excitation of neighbouring neurones. It seems to be common that these reciprocal neurones lie rather near together so that you get them with the same microelectrode, pushed some microns further. This may also be due to their mutual inhibitory synaptic relation. However, we have not yet any good examples of recurrent inhibition in cortical neurones except in the Ammons horn and no Renshaw mechanism has been demonstrated in the cortex, although there must be plenty of negative feed back mechanisms to prevent a convulsive explosion in the immense cortical neuronal apparatus and its synaptic powder barrel.
RECURRENT INHIBITION AS A MECHANISM OF CONTROL
37
G. GRANIT’s replies To J . C . Eccles There is little more to add to Sir John’s comments than the admission that his suggestions seem entirely plausible and deserve to be tested by experimentation. To F. Bremer Has cerebellar inhibition a component run by the Renshaw cells? This may or may not be the case. There are places in the cerebellum that excite, and others that inhibit these cells. No comparison has yet been made with the effects from corresponding cerebellar sites on motoneurones controlled by a definite set of Renshaw cells.
To H . H . Jasper The first question regarding tonic and phasic neurones in the cortex I cannot reply to, even though I am inclined to suspect that this is a classification that applies to a large number of organised aggregates of neurones. It seems such a sensible subdivision of elementary tasks. Phillips has found evidence for recurrent action in the motor cortex but the spinal cord is unique in providing a situation that can be handled in quantitative terms. This proviso is essential because from it alone has it become possible to deduce from the experiments that recurrent inhibition definitely is a mechanism serving discrimination by what, for short, we may call contrasf, motor or sensory, as the case may be. To R. Jung This very point is raised by Professor Jung’s remarks. I entirely agree that similar mechanisms. wherever they be found, must serve the same purpose, sc. contrast. The recurrent or lateral inhibition in Limulus does it, although this preparation is an invertebrate eye lacking also the internuncial cells of the motoneurones in the spinal cord. In fact, any recurrent inhibition, randomly distributed within an aggregate of neurones devoted to the same function, should behave as in the two examples mentioned, because its action is tied to frequency of firing and is mutual with respect to adjacent neurones in possession of recurrent collaterals. A very attractive feature of the findings is that they can be simply expressed in terms of impulse frequency which, with some justification, may be called the fundamental quantity of the code in which nervous messages are clothed. This has also been emphasized by Hartline and his co-workers.
H. K., RATLIFF,F. and MILLER,W. H. Inhibitory interaction in the retina and its significHARTLINE, ance in vision. In E. FLOREY (Editor). Nervous Inhibition. Pergamon Press, New York, 1961: 241-284.
PHILLIPS,C . G. Actions of antidroniic pyramidal volleys on single Betz cells in the cat. Quart. J. exp. Physiol., 1959, 44: 1-25.
Studies of the Integrative Function of the Motor Neurone * JOHN M. BROOKHART AND KISOU KUBOTA Department of Physiology, University of Oregon Medical School, Portland, Oreg. (U.S.A.)
INTRODUCTION
Sherrington (1904) long ago, in the development of the concept of the final common path, emphasized the key role played by the motor neurone in the overt expression of the functions of the central nervous system. The complexity of the higher organisms demands a mechanism for building a unified behavior out of multitudes of influences. Since that behavior can be manifest only through activity initiated in motor neurones, it seems appropriate, in a conference devoted to sensory-motor integrations, to devote some time to the question of the degree to which the motor neurone itself is capable of abstracting information from the spatio-temporal pattern of incoming impulses and converting this information into a self-engendered frequency modulated code suitable for the controlled activation of effector cells. We may rephrase this question by asking about the conditions under which the motor neurone may be forced to respond as a slave to each presynaptic volley, and, alternatively, about the conditions under which the motor neurone may behave as though it had its own independent pacemaker whose frequency alone is subject to modulation by incoming impulses. It is only in this latter condition that the motor neurones could be considered to have any integrative capabilities -that is to say, any ability to reflect the total of all of the influences brought to bear upon them from a variety of presynaptic sources. The origins of the notion that neurones are endowed with a pacemaker function are familiar to all of us (Barron and Matthews 1938; Gesell 1940; Alanis 1953; Bishop 1956; Morrell et al. 1956) and have been modified more recently by Granit and Rutledge (1960). The area in which we lack information and insight concerns the different synaptic organizations which may underly differences in input-output relationships. One hypothesis which warrants consideration is that significant differences in postsynaptic behavior may result from differences in the placement of presynaptic terminals on the neurone, those being placed close to the presumptive “trigger zone” having different potentialities than those making contact at sites more distant. The validity of this hypothesis is clearly predictable on purely physical grounds, taking into consideration the characteristics of electrotonic phenomena and the dimensions *This investigation was supported by Research Grant B-385, National lnslitutc of Neurological Diseases and Blindness, US. Public Health Service, and by the Medical Research Foundation of Oregon.
INTEGRATIVE FUNCTION OF THE MOTOR NEURONE
39
of the dendritic expansions of nerve cells. In addition to this, however, the possibility must be admitted that the neuronal membrane may have different functional capabilities in its different parts. DeLorenzo (1961) has demonstrated structural and histochemical differences between axo-somatic and axo-dendritic synapses. We are all familiar with the concept of electrical inexcitability of dendritic membrane which has been developed by Grundfest (1957) and which was reviewed critically by Eccles (1959) at the Amsterdam conference in 1959. Whether or not we admit that the available evidence is adequate to prove the validity of this concept, the fact remains that no evidence has been offered which would force us to discard this possibility. One of the serious drawbacks to this hypothesis has been the difficulty of subjecting it to a critical test. Until recently it has not been possible to examine a neuronal system in which the site of termination of presynaptic endings could be predicted with an acceptable degree of confidence. In studies of spinal reflex activities, the question has not been raised. In studies of cerebral and cerebellar cortical neurones, the evidence has been cloudy and the inferences indirect. However, there is one such neuronal system, with which it has been our pleasure to work for the past few years : the spinal cord of the frog offers us opportunities to compare the results of activation of motor neurones through axo-somatic and through axo-dendritic synapses. I should like to spend the first few minutes reviewing quickly the evidence upon which this statement is based. I shall then tell you of some of the observations that we have been making recently using repetitious excitation of these two varieties of input to motor neurones. SYNAPTIC ARRANGEMENTS IN THE FROG SPINAL CORD
Our interest in the frog’s spinal cord as an appropriate object for study was first aroused by the anatomical observations of Liu and Chambers (1957). Using axon degeneration techniques, these investigators observed that dorsal root fibers terminated solely in the dorsal horn in the vicinity of internuncial cells and on the long, dorsal dendrites of motor neurones such as are illustrated in Fig. 1 . Liu and Chambers (1957) also observed that a system of fibers descended several segments within the lateral column of white matter to terminate in the vicinity of the motor neurone cluster in the ventral horn. Before turning your attention away from this reproduction of a Golgi preparation, I should like to point out that intramedullary collaterals from motor axons have been seen in the amphibian. Sala y Pons (1892) shows two of them here in the upper left portion of Fig. 1. Chambers (personal communication) has also seen them wander off to become lost in the marginal dendritic plexus. 1 shall return to this subject briefly later on. In our hands, stimulation of the lateral column input and of the dorsal root input to the motor neurones gives rise to ipsilateral ventral root responses such as those illustrated at the left of Fig. 1. ln the case of the lateral column stimulation, the discharge occurs after a brief latency, is highly synchronized and lasts for only a brief period; the response to dorsal root stimulation is the same as that which has been studied so extensively by Bremer and his colleagues (Bonnet and Bremer 1952), Relerences P. 60-61
40
J. M. BKOOKHART, K. KUBOTA
Fig. 1 Structure and function of anuran spinal cord. Right: Golgi preparation of toad spinal cord showing motor neurones ( A ) and their complex dendrites. Note long dendrite extending into dorsal horn. Note also short motor axon collaterals ( K , U ) . Left: ipsilateral ventral root responses to stimulation of descending fibers in lateral column of white matter (above) and to stimulation of dorsal root (below) of frog’s spinal cord. (From Sala y Pons 1892.)
consisting of a somewhat delayed, asynchronous discharge superimposed upon an electrotonically propagated post-synaptic potential. When these two inputs to the motor neurones were excited repetitiously (Fig. 2), the ipsilateral ventral root discharges differed even more dramatically. Dorsal root excitation evoked an asynchronous discharge which was poorly maintained for the duration of the stimulation ; lateral column stimulation evoked a response which retained a high degree of synchrony and exhibited extensive recruitment (Brookhart et al. 1959). In order to gain further information about the functional capacities of these two inputs, we have devised a system whereby the entire spinal cord can be removed from the frog and placed, lateral surface up, in an artificial controlled environment (Fig. 3). With the dorsal and ventral roots extended to opposite sides of the cord and the pia carefully dissected away from the area of the ellipse shown on the lateral surface, the preparation may be examined using either gross or microelectrodes. When the motor neurone responses are examined with microelectrodes recording from within the cell body (Fig. 4), we characteristically observe that the EPSP evoked by dorsal root excitation is complex and multiple whereas that evoked by lateral column stimulation is simple and unitary (Machne et a/. 1959). The multisynaptic nature of dorsal root excitation may be avoided and the EPSP simplified by addition of pentobarbital to the bathing fluid. It has been through the study of monosynaptically induced EPSP’s that we have been able to conclude that monosynaptic connections
INTEGRATIVE FUNCTION OF THE MOTOR NEURONE STIY. DRK)
41
STIM. CORD
I 30
40
A
i
200yv
Fig. 2 Responses to repetitive stimulation. lpsilateral ventral root slow waves and motor neuronal discharges evoked in response to repetitious excitation of dorsal root (left) and lateral column (right) in two preparations. Stimulus frequencies indicated between pairs of records in cycles per second. Upper left: time calibration for all records. Upper right: response of amplifier system to rectangular puke of same duration as stimulus train (time constant 1.0 sec). At the recording sensitivity required to reveal the slow wave contours, much of the detailed evidence of the asynchronous impulse discharge has been sacrificed. The differences between slow waves evoked by the two forms of stimulation are clearly seen. Note the recruitment and persistently synchronous nature of the discharge evoked by lateral column stimulation. Note also the differences in recording sensitivity for the two sets of records (From Brookhart et of. 1959. Courtesy of Archives ltaliennes de Biologie.) References
P. 60-61
42
J . M. BROOKHART, K. KUBOTA
CORD STlM
I
'I
REFLEX STlM
Fig. 3 Arrangement for studying isolated spinal cord of frog. The spinal cord is removed and mounted in a grooved chamber which carries a continuous flow of Ringer-glucose (Oa 95%, COz 5%). The lateral surface is uppermost with roots extended to either side for stimulation and recording. The upper pair of electrodes is used to activate descending fibers in the lateral column. Pia is removed from the area of the ellipse after tryptic digestion using a dissecting microscope in order to prepare for the penetration by intracellular recording electrode. Damage to the marginal dendritic plexus during dissection may seriously depress the responsiveness of the preparation.
Fig. 4 Intracellular responses of single frog motor neurone. Upper records: membrane potential changes during three varieties of response from the same neurone. Lower records: time derivative of potential changes recorded simultaneously. A : antidromic activation. D R : activation over dorsal root fibers. Note latency of response and contour of EPSP. LC: activation over lateral column fibers. Compare to D R response with regard to latency and prepotential. Sensitivities as indicated. Time: 0.5 msec. D R and LC responses superimposed on RMP baseline. (From Machne et al. 1959. Courtesy of C. C. Thomas, Publisher.)
INTEGRATIVE FUNCTION OF THE MOTOR NEURONE
43
are made from the dorsal roots to the dendrites of these motor neurones, whereas monosynaptic connections are made from the lateral column fibers directly to the cell bodies of the neurones. Dr. Fadiga is primarily responsible for the careful and penetrating analysis of focal and intracellular potentials which has led to this conclusion (Brookhart and Fadiga 1960; Fadiga and Brookhart 1960). I cannot review all of those data for you here. I will simply point out that as indicated by the shaded areas in Fig. 5, the distribution of focal negativity accompanying monosynaptically induced EPSP’s differs for the two forms of excitation, being confined to the dorsal horn with dorsal root excitation and to the region of the motor nucleus with lateral column excitation. The upper two records in Fig. 6 reveal the wave forms of the focal potential initiated by dorsal root excitation on the left and by lateral column excitation on the right. Despite the similarity of wave forms in these two foci, the electrotonically propagated ventral root slow waves differ to a considerable degree. The slow wave from dorsal root excitation, shown on the lower left, is delayed both in rise and decay times to a greater degree
Topographical
distribution
of
focal
potentials
B Stim.
lat.
col.
Stim.
dors.
root
Fig. 5 Topographical distribution of focal potentials. Focal potentials resulting from monosynaptically induced EPSP’s were generated by lateral column stimulation (to the left) and dorsal root stimulation (to the right). The heavy black lines indicate the margins of the field delimited by a reduction to 20% of maximal amplitude. The shaded zones indicate areas of particularly intense focal negativity. The lines calibrated in microns below the surface denote representative electrode tracks. The penetrations originated near the ventral root and coursed through the motor nucleus 400-6OOp from the surface. Note that focal negativity from dorsal root stimulation was confined to the dorsal horn while that from lateral column stimulation was confined to the ventral horn. (From Brookhart and Fadiga 1960. Courtesy of the Journal of Physiology.) References P. 60-61
44
J. M. BROOKHART, K. KUBOTA
.A,
2 -
-J-
rnsec
Fig. 6 Comparison of focal and radicular responses. A , A1: simultaneous records from dorsal horn focus ( A ) and ventral root (A1) following monosynaptic dorsal root excitation. B, B1: simultaneous records from ventral horn focus (B) and ventral root (El) following monosynaptic lateral column excitation. All records from one preparation. Note greater delay in both rise and decay times of radicular response to DR excitation even though the focal negativities at zones of origin are identical in the two cases. This is regarded as evidence of greater distance of electrotonic propagation of monosynaptically induced postsynaptic response to DR excitation. (From Brookhart and Fadiga 1960. Courtesy of the Journal of Physiology.)
than that initiated by lateral column excitation as shown on the lower right, suggesting its more distant origin. This difference in form was also noted when the EPSP’s induced by the two forms of excitation were compared. In order to carry out this comparison properly, it was necessary to correct for the distorting effects of focal potentials which were superimposed on the changes in membrane potential. This was done by recording from the region just outside of a given cell after its death and summating the two responses algebraically. The lower records in Fig. 7 illustrate corrected responses from the same cell. The greater delay in rise time of the EPSP evoked by dorsal root excitation is obvious. The EPSP decay curves also differed significantly in a manner illustrated by the semilogarithmic plots in Fig. 8. The initial portion of the response to lateral column stimulation decayed significantly more rapidly than the response to dorsal root excitation. All of these findings are consistent with the inference that monosynaptic excitation of frog motor neurones occurs via axo-dendritic synapses from dorsal root fibers and over axo-somatic synapses from lateral column fibers. REPETITIOUS EXCITATION
I should now like to turn to the results which we have been obtaining during the last six months. During this time, we have been examining the intracellular potential changes initiated by repetitious excitation of these two systems of presynaptic fibers. Our particular interest has been in the character of the motor neuronal discharges evoked, and the relation of the discharges to the form of presynaptic excitation and subsequent membrane potential changes. In order to establish a background against which these results must be viewed, I must call attention first to certain special features of these preparations of isolated frog spinal cord.
INTEGRATIVE FUNCTION OF THE MOTOR NEURONE
45
Fig. I Intracellular EPSP’s and distorting focal potentials algebraically summated. In each of the five examples, the upper record (A-E) shows a monosynaptically induced EPSP. Immediately below (AI-EI) is shown the focal potential remaining after the death of the cell. The corrected tracing ( A z - E ~ is ) the result of algebraic summation of the other two records. A , B, and C were derived from different neurones; D and E from the same neurone. A , B, and D resulted from lateral column stimulation; C and E from dorsal root stimulation. The prolongation of the corrected EPSP resulting from dorsal root excitation is evidence of its origin by electrotonic propagation from a distant portion of the cell membrane. (From Fadiga and Brookhart 1960. Courtesy of the American Journal of Physiology.)
Limiting considerations. Frog motor neurones apparently differ in certain respects from cat motor neurones particularly in the form of their after-potentials. We have seen three major classes of after-potentials following antidromic excitation (Fig. 9). These post-spike changes are all brief and none of them break over into the phase of prolonged after-hyperpolarization which characterizes cat motor neurones (Eccles et al. 1957). We have not inquired into the reasons for these differences between frog and cat motor neurones. We can state that this late portion of the response is sensitive to changes in external K+concentration, and we may suspect that the antidromic focal potential distorts the true picture of membrane potential changes.
J. M. BROOKHART, K. KUBOTA
46
t
b = -0.1586
LC-EPSP
b=-0.1084
25
t
0.3
i 2
4
MILLISECONDS
8
AFTER
10
12
PEAK
Fig. 8 Corrected EPSP decay curves. Decay phases of monosynaptic EPSP's of ten units were corrected for focal distortions and mean values plotted on semilogarithmic coordinates. Responses to DR excitation above; to LC excitation below. The time constant of decay for D R responses is estimated to be 10.3 msec; that for the somatic component of LC responses is estimated to be 6.5 msec. The prolongation of the decay phase of the EPSP initiated by dorsal root excitation is interpreted as the result of electrotonic propagation from dendritic sites of origin. The fast component of the decay phase of the EPSP initiated by lateral column stimulation is considered to be characteristic of somatically induced responses. (From Fadiga and Brookhart 1960. Courtesy of the American Journal of Physiology.)
Valietii of antldromic splkes
5msec
Varieties of antidromic spikes. Representative examples of three classes of antidromic responses which differ in the detail of the post-spike potential changes. The reasons for these differences are unknown. Unlike cat motor neurones, frog motor neurones do not exhibit a prolonged phase of after-hyperpolarization. Sensitivity calibration : 50 mV.
47
INTEGRATIVE FUNCTION OF THE MOTOR NEURONE
Another consideration of some importance to the general characteristics of this preparation is the fact that we have seen no sign of inhibitory post-synaptic potentials (IPSP's). Stimulation of dorsal roots, dorsal root filaments and lateral column fibers has evoked only excitatory depolarization. We have, on rare occasions, also noted depolarizing responses evoked by stimulation of the ventral root adjacent to the root to which a given motor neurone contributed its axon. In the absence of critical controls we are not able to say surely that such responses are the result of antidromic recurrent facilitation (Washizu 1960; Wilson and Burgess 1961). We have not seen any responses which might be regarded as evidence of recurrent inhibition. To accept the absence of synaptically induced hyperpolarization as evidence that inhibition of frog spinal motor neurones does not occur would be most unwise in the light of demonstrated inhibition of reflexly induced muscular contraction (Veszi 1910). We must therefore conclude that either we have not utilized the proper selection of inputs to initiate inhibitory reactions or that inhibition occurs in a manner which does not reflect itself in the form of an IPSP (Frank and Fuortes 1957; Dudel and Kuffler 1961). The search for inhibitory reactions is still in progress. Antidromic responses. The ability of these motor neurones to generate repetitious discharges has been tested using antidromic shocks (Fig. 10). The responses of two
ANTIDROMIC
ORTHODROMIC .r
I
1
I
I T
B' DR
.
'I! !
1
I
.
1
j , ,, , . , .A 5msec Fig. 10 Repetitious spike formation. Characteristics of spiking behavior during antidromic and orthodromic excitation at 50 c/sec. A and A' from one neurone, orthodromically excited over lateral column fibers. B and B' from another neurone, orthodromicallyexcited over dorsal root fibers. Note stability of spike amplitude during antidromic responses compared to diminution in spike amplitude during orthodromic responses. Compare also the changes in firing levels in the two conditions. ".,.....-..
. I,, =..
48
J. M. BROOKHART, K. KUBOTA
neurones to such stimulation at 50 c/sec are illustrated to the left. Disturbances of spike amplitude and contour during the period of stimulation were not remarkable. This was in contrast to the diminution of spike amplitude which characterized the initial phases of repetitious responses initiated orthodromically in the same neurones. We regard this initial reduction in amplitude, at least in part, as the effect of shortcircuiting of the spike-forming membrane during the underlying EPSP (Fatt and Katz 1951; Fadiga and Brookhart 1960). We have attempted to summarize the frequencyfollowing ability of the neurones in a small population by preparing a graph relating firing index to frequency of antidromic stimulation (Fig. 11). Some of the neurones Antidromic stimulation 17 units
1
5
10
50
100
200
Frequency- clsec
Fig. 11 Performance characteristics of antidromically activated neurones. The graph relates firing index (No. of responses/No. of stimuli) to frequency of stimulation for seventeen randomly chosen cells excited antidromically. Each line represents the behavior of one cell. The majority of the cells followed stimulus frequency up to 50 c/sec. Response failure was often due to failure to invade the initial segment of the neurone. This graph roughly describes the ability of frog neurones to follow repetitious antidromic stimulation.
in this group were able to follow antidromic frequencies to values of 100 c/sec while others underwent a reduction of firing index at much Iower frequencies. Thus, seven of this group of seventeen cells had reduced their firing index to 0.5 or less by the time the frequency had reached 50 c/sec. When antidromic invasion failed, it often failed in the initial segment, leaving only a small complex without an IS spike (Initial Segment, Eccles 1957). Even when the firing index was reduced by conduction failure in the initial segment, the spikes which signalled successful invasion were normal
49
INTEGRATIVE FUNCTION OF THE MOTOR NEURONE
in amplitude and contour. Thus, it would appear that the soma membranes were not significantly affected by repetitious activity at frequencies which exceeded the critical values for more peripherally located portions of the cell. Orthodromic responses to lateral column stimulation. Orthodromic activation of the motor neurones over the lateral column fibers which make direct axo-somatic connections has been examined both in untreated spinal cords and in those in which internuncial activity was depressed with pentobarbital. Under conditions of repetitious stimulation, many interesting changes have been seen to occur in relation to cell firing thresholds, spike amplitudes and spike contours. Since these alterations are not considered pertinent to our major objectives in this presentation, I am going to confine your attention here primarily to the firing patterns and to the alterations in postsynaptic membrane potentials. In the untreated preparations (Fig. 12), the size of the Slim ht.-
LATERAL COLUMN (Unlre.ll*dl
s
EXCITATION
10
Slim I n l -
‘
>
25
d
I
*
I
I
‘
1
1
)
( 1
I
d
d
d
0 ~
( ( ‘ ( I 1!1 1 , I 1
& 100 JA&..L-
l
I
L
~
,
-
l
,
,
l
~
~
~
,
I( I I
~ J J J J S J J l S J J k J d
I l l
eooniJp..4----
I
%
I
50~~~~ljcJ~ddljjlrlvJGzz_l_
>-L-L I
50
10
- ,
10
I I ( I I
~ _ _
-
Il/I
addh-4wL-
-
sec
LAUUL
I
50mV
Fig. 12 Responses to repetitious lateral column stimulation. Intracellular potential records from one cell in a preparation which was not treated with pentobarbital. Relative stimulus intensity indicated above each column. Stimulus frequency indicated at the start of each record in cycles per second. Resting potential level indicated at the left of some records by straight line. Time and sensitivity applicable to all records. With the weaker stimulus, the single and summated EPSP’s were smaller and only a low firing index was attained. Nevertheless, when it did fire, the unit was always “triggered”. With the stronger stimulation, “driving” was achieved following the fourth stimulus at 25 c/sec; the cell was “driven” at 50 c/sec, initially “driven” and later “triggered” at 100 c/sec.
presynaptic volley may be made so small that the firing index never approaches unity. Nevertheless, when the cells did fire they were “triggered” in the sense used by Granit and Phillips (1957) with a uniform latency following each effective shock. When the size of the presynaptic volley was increased, the size of the individual EPSP’s was greater and a firing index of unity was readily achieved after summation of EPSP‘s had progressed to an adequate degree. With still larger presynaptic volleys a firing index of unity could be achieved without the necessity for summation. It is noteworthy that, as was the case with antidromic stimulation, failure of spike generation appeared to occur first in the initial segment. Only occasionally have we seen IS spikes generated without subsequent SD spikes (Soma-Dendrite, Eccles 1957). This form of stimulation characteristically induced a rapid summation of membrane depolarization to a
50
J. M. BROOKHART, K . KUBOTA
stable plateau value. At the end of the period of stimulation the membrane potential change gradually subsided over a period of 50-100 msec without any sign of the summating hyperpolarization which is so prominent in some of the records obtained from the cat (Curtis and Eccles 1960; Eccles and Rall 1951). Elimination of internuncial activity by treatment with pentobarbital did not alter these behavioral characteristics appreciably (Fig. 13). The pattern of firing and the L A T E R A L COLUMN EXCITATION ( Tr e o I e d )
I/
10
f
/t
-
!
2
9l
L
l
J
10 rn sec
4
U
-
1
I
L
L
Fig. 13 Responses to repetitious lateral column stimulation. Similar to Fig. 12 but from a different neurone in a spinal cord treated with pentobarbital to eliminate internuncial activity. Note similarities to behavior of untreated neurones.
pattern of membrane potential changes in this treated neurone was in all respects identical to that seen in Fig. 12. This similarity in firing behavior in the treated and untreated conditions is emphasized when we compare the relationship between firing index and frequency of stimulation for a larger population of cells (Fig. 14). In both situations, those cells which fired with low frequency stimulation behaved very much as though they were being fired antidromically. Those cells which fired only after summation of EPSP’s had been achieved by repetitious excitation approached their maximum firing indices at frequencies between 30 and 50 clsec and were usually “driven”. Taken together, these data permit the conclusion that internuncial influences do not contribute significantly to the behavior of neurones responding to axo-somatic excitation. We get the impression of a powerful synaptic relationship readily capable of “driving” (Granit and Phillips 1957) motor neurones to the limit of their ability to respond, capable of maintaining a steady discharge frequency, and with an ability to induce sustained summated depolarization which may reflect the prolonged duration of transmitter action (Fadiga and Brookhart 1960). In short, this is a synaptic
51
INTEGRATIVE FUNCTION OF THE MOTOR NEURONE
L A T E R A L COLUMN S T I M U L A T I O N U N T R E A T E D - I 3 Units
T R E A T E D - I I Units
1.0
z
3
z
,$0.5 i
;;
0
-
Frequency c / sec
5
10
50
100
200
Frequency- clsec
Fig. 14 Performance characteristics of neurones orthodromically excited over lateral column fibers. Firing index related to frequency of stimulation as in Fig. 11 for untreated (left) and treated (right) neurones. Note similarity of behavior of antidromicallyexcited neurones and those which exhibited firing index of 1.0 at low frequencies. Firing index of those units unresponsive to low frequency stimulation approached unity when summationof EPSP’s was achieved by repetitivestimuli. Those units requiring summation generally responded in a “driven” fashion in the 50-100 cycle range.
relationship which would not allow the motor neurone to exhibit whatever integrative capabilities it may have, but which would force it to reflect the presynaptic frequency. Orthodromic responses to dorsal root excitation. The results of dorsal root excitation have also been examined in the presence of and in the absence of depression of internuncial activity. In the untreated state (Fig. 15), rather wide variations in discharge pattern have been noted in different preparations. On the right are records from one cell in which it was impossible to achieve a firing index of unity, and in which summation of post-synaptic depolarization was not pronounced. On the left are records from another cell in which the firing index was greater than unity at low frequencies because each shock induced the generation of multiple discharges. At a stimulus frequency of 40 c/sec this cell settled down to a firing index of unity, behaving in a “driven” fashion (Granit and Phillips 1957). At a stimulus frequency of greater than 40 c/sec the ultimate behavior is described by a firing index of far less than unity and an average frequency of approximately 30 c/sec. In this condition the cell was neither “triggered” nor “driven” but initiated discharges at variable intervals following shocks. Its behavior resembled closely the type of activity sometimes seen in cortical cells to which I have heard Dr. Jasper apply the term “idling discharge”. From our past experience with this preparation, we can be quite confident that the activity initiated in the untreated cord by dorsal root excitation involves internuncial cells to an important degree. Indeed, when preparations treated with pentobarbital were examined, it was difficult to secure any discharge from the motor neurones. The frequency distribution diagrams presented here (Fig. 16) summarize our experience with 97 cells subjected to activation over dorsal root fibers. The horizontal axis refers to the number of impulses evoked from each cell by each single dorsal root stimulus. References P. 60-61
52
J. M. BROOKHART, K. KUBOTA
b,1
1
DORSAL ROOT [Untreoled 1
EXCITATION
’I
T
Fig. 15 Responses to repetitious dorsal root excitation. Intracellular potential changes initiated in two untreated neurones from two different preparations. Time and sensitivity calibrations (upper right) apply to all records. Frequency of stimulation shown at the beginning of each record in cycles per second. Wide variations have been encountered in ability to secure firing, associated with differences in EPSP’s. Note differences in amount of “synaptic noise” between spikes in these two units. The unit on the right never achieved a firing index of 1.0. The unit on the left showed a firing index greater than 1.0 at 1, 10 and 40 cjsec. After the initial high frequency burst it fired in a “driven” fashion at 40 c/sec. At 100 c/sec the firing index was quite low, and the rhythmic firing at approximately 30 c/sec occurred with random relation to individual dorsal root stimuli.
When internuncial participation was depressed by pentobarbital, none of the cells fired repetitiously and single discharge could be secured in only 28 per cent of the treated population. The individual EPSP’s were smaller in amplitude and the degree of depolarization achieved by repetitious stimulation was considerably less than seen during lateral column stimulation (Fig. 17). Within the limits of concentration of pentobarbital which the preparation could tolerate it has been impossible to eliminate completely all signs of internuncial participation. The characteristic membrane potential changes showed a ready ascent to a peak value. During this phase the EPSP’s were complex and obviously not solely monosynaptic in origin. After the initial burst of internuncial reinforcement, the level of membrane depolarization subsided to a plateau value which was of the order of 50 per cent of the peak value. The cessation of stimulation was followed by a return to resting membrane potential values along a time course resembling that of a unitary EPSP. Further appreciation of the importance of internuncial reinforcement of dorsal root impulses is gained from a comparison of the relations between firing index and frequency in the treated and untreated states (Fig. 18). As is evident from these graphs, dorsal root excitation was not capable of evoking a sustained high firing index when
53
INTEGRATIVE FUNCTION OF THE MOTOR NEURONE Repetitive spike generation (single DR stimuli)
...... ...... ...... ...... ...... ...... ...... ......
rmm
47 Units
1
6
1
7
1
8
1
9
1
1
0
U
J
t
I-
50
...... ...... ...... ...... ...... - ...... ......
50 Units
Fig. 16 Multiple discharge following single dorsal root stimuli. Frequency distributions of number of discharges to single dorsal root shocks in 47 untreated neurones (above) and 50 neurones treated with pentobarbital (below). The horizontal axis indicates the number of discharges. The vertical axis indicates the percentage of the respective populations. The differences between the treated and untreated groups are ascribed to differences in interneuronal behavior.
frequency was elevated. Even the firing index of those cells which fired repetitiously with low frequency stimulation dropped to unity or less at 50 c/sec. In no case was the summation of EPSP’s attendant upon repetitious excitation capable of increasing the firing index of those cells which could not be fired by a single stimulus. The depression of internuncial activity even further compromised the ability of these cells to engage in repetitious response with the result that firing indices were not sustained at unity even with low frequencies of 5 and 10 clsec. From responses such as these which I have just described, we get the impression that the monosynaptic axo-dendritic pathway is relatively ineffective in terms of initiation and sustenance of discharge unless supported by internuncial reinforcement. Many of our records lead us to believe that frog interneuronesdischarge repetitiously as do many such neurones in the cat spinal cord (Kolmodin 1957; Hunt and Kuno 1959; Wall 1959; Eccles ef al. 1960).It is quite reasonable to conclude that this internuncial activitycreates a brief wave of depolarization which engenders a short burst of high frequency impulses from the motor neurone. Araki and Otani (1959) express the opinion that the early cessaReferences P. 60-61
54
J . M. BROOKHART, K. KUBOTA DORSAL
ROOT EXCITATION ( Trea t ed)
20
Fig. 17 Responses to repetitious dorsal root excitation. Intracellular potential changes in a single neurone after treatment with pentobarbital. Frequency of dorsal root excitation indicated at the start of each record in cycles per second. Resting membrane potential baseline indicated at the right in some records. Note the complexity of the EPSP's in the early responses of the train suggesting brief burst ofinterneuroneactivity. Coincident with the appearance of unitary, monosynapticEPSP's the degree of depolarization diminishes. Note increasing level of steady-state depolarization with increasing frequency. The last record in the series illustrates the effects of re-stimulation at 197 c/sec after an abbreviated rest period of less than 0.5 sec. The response appears to be devoid of the effects of internuncial reinforcement in its early phases.
tion of firing followingdorsal root excitation is theresult ofaccommodation ofthe trigger zone. We may infer from the observed changes in firing thresholds of both segments of the spikes that accommodation does enter into the picture; but it appears to do so to an equal degree with both forms of stimulation. Consequently, we are inclined to believe that withdrawal of internuncial reinforcement plays a more important role in the abbreviation of the discharge evoked by dorsal root excitation. In summary, we have observed two different kinds of discharge pattern. Impulses delivered to the motor neurone over axo-somatic synapses are capable of driving these motor neurones at presynaptic frequencies, forcing them to reflect the level of their input in terms of output frequency. Axo-dendritic synapses, plus internuncial pathways, seem to afford a mechanism whereby the motor neurone has the opportunity to differentiate increases in its input level and to signal, by its output frequency, the rate of change of input rather than its absolute level.
INTEGRATIVE FUNCTION OF THE MOTOR NEURONE
55
DORSAL ROOT S T I M U L A T I O N TREATED- 14 Unifs
I.(
, 0.: C Frequency -clsec
Frequency- clsec
Fig. 18 Performance characteristics of neurones orthodromically excited over dorsal root fibers. Similar to Fig. 11 and 14. The influence of depression of internuncial reinforcement by treatment with pentobarbital is seen in the reduction of firing index at low frequencies and in the rapid diminution of firing index of treated neurones as frequency is raised. The treated group includes all of the neurones indicated in Fig. 16 which fired once in response to single dorsal root shocks.
The role of accommodation. We may look in several directions for reasons which might explain these different modes of behavior. The studies of Araki and Otani (1959) and Araki (1960) have emphasized the accommodative properties of toad motor neurones and have called attention to their ability to develop responses of somatic origin under certain circumstances. Although the records available to us at present are not adequate for critical analysis, the changes in IS and SD thresholds during both forms of orthodromic stimulation appear to be essentially identical. For the time being, we are inclined to the opinion that accommodation during depoIarization induced from distant sites on dendrites is not different from accommodation during direct somatic depolarization. Thus, we are not hopeful that differences in this property of the post-synaptic membrane will correlate with the observed differences in behavior. Thesignijkance of post-synapticmembrane potential changes. We have also examined the post-synaptic membrane potential changes which follow both forms of stimulation in the search for cues concerning the release of transmitter substance. The graphs of Fig. 19 represent the results of measurement of the amplitudes of individual EPSP’s averaged over the first 300 msec of stimulation and related to the frequency of References p , 60-61
J. M. BROOKHART, K . KUBOTA
56
LATERAL C O L U M N STIMULATION
7r
DORSAL ROOT S T I M U L A T I O N
12 Units
I
5
10
50
Frequency-clsec
100
I
200 I
5
10
50
100
200
Frequency-clsec
Fig. 19 Single EPSP amplitudes as a function of frequency of stimulation. All preparations treated with pentobarbital. Orthodrornic excitation over lateral column fibers (left) and dorsal root fibers (right). Mean values of single EPSP’s determined for first five to ten responses at lower frequencies and for responses during first 300 msec of stimulation at higher frequencies. Note that, within the frequency limits of motor neurones (see Fig. 11, 14), EPSP’s induced by lateral column stimulation either remained constant or rose slightly in amplitude, whereas EPSP’s induced by dorsal root excitation underwent progressive reduction with increasing frequency.
excitation. The amplitudes of single EPSP’s generated by single dorsal root and single lateral column stimuli fall into the same range of values. However, even though all of the preparations used in this measurement were treated with pentobarbital, the EPSP’s generated by dorsal root stimulation probably reflected some internuncial activity. In spite of this help, the average value for single dorsal root EPSP’s at low frequencies is obviously smaller than that resulting from lateral column stimulation. With increasing frequency of stimulation, the size of the individual EPSP’s generated by lateral column stimulation was quite constant; indeed, in some cases the size of the EPSP was actually increased with increasing frequency. The EPSP’s generated by dorsal root excitation, on the other hand, underwent a steady and progressive diminution as the frequency of excitation increased. It is impossible to ascribe this diminution specifically to either pre- or post-synaptic changes. It may reflect a change in the amount of transmitter released per impulse, or it may reflect desensitization of the post-synaptic membrane (Curtis and Eccles 1960; Thesleff 1959), an action which would be enhanced by the relatively long duration of transmitter action in this species (Fadiga and Brookhart 1960). Whichever the case, the direct axo-somatic terminals and somatic membrane apear to be more capable of inducing continued large increments of depolarization even though the frequency of excitation rises. If we were to design a system which would be capable of forcing or driving the post-synaptic element to respond to input frequency, we would make arrangements so that trans-
INTEGRATIVE FUNCTION OF THE MOTOR NEURONE
57
mitter action could be sustained at a powerful value at all frequencies of operation that fell within the capabilities of the membrane to respond. Such seems to be the case with the axo-somatic synapses of lateral column fibers. If we were to design a system for the purpose of responding only during periods when the level of input was increasing, we would make arrangements for a rapid subsidence of transmitter action after the achievement of a steady state of input. This appears to be the manner in which the dorsal root fibers and internuncial pathways behave in this preparation. We have already seen that both forms of excitation are capable of inducing sustained membrane depolarization during repetitious excitation. The graphs in Fig. 20 illustrate how the steady-state level of depolarization, measured late in the stimulus train, varied with changing frequency. In all cases the preparations were treated with pentobarbital. The late measurement, and the unitary contours of the EPSP’s allow us to look upon these responses as the reflection of monosynaptic excitation. For both
LATERAL COLUMN STIMULATION
Frequency- cfsec
DORSAL ROOT STIMULATION
Frequency-c/ 5ec
Fig. 20 Steady-state depolarization as a function of frequency of stimulation. All preparations treated with pentobarbital. Inserts above: representative intracellular potential changes induced by stimulation of lateral column fibers (left) and dorsal root fibers (right) at 30 c/sec. Measurements of steady-state depolarization made from resting membrane potential level to troughs between individual EPSP’s during last responses in train. See discussion in text.
58
J. M. BROOKHART, K. KUBOTA
varieties of excitation, the steady-state level rose as frequency increased. The slopes of the lateral column curves, and the final level reached, are such as to insure that the motor neurones were brought close to their firing levels. In this condition, the large individual EPSP’s would be more apt to secure firing even in the face of accommodation. We must not allow ourselves to be misled by the apparent weakness of the steadystate depolarization which is induced by repetitious axo-dendritic excitation. It will be recalled that Fadiga and Brookhart (1960) estimated that the degree of decrement associated with electrotonic propagation of an EPSP from its dendritic origin to the point of recording at the cell body was such that a depolarization of between 20-30 mV at the dorsal dendrites would be reflected at the cell body by an EPSP of only 2 mV. The failure of these cells to fire, in spite of the extensive dendritic depolarization which must have been induced, may be regarded as evidence that propagated action potentials are either not induced in these dendrites or fail to propagate the entire distance to the cell. The simplicity of the EPSP and its exponential decay argue against the latter possibility. It may be predicted that even though extensive dendritic depolarization is only able to shift the membrane potential of the cell through a few millivolts, this small shift should have its influence on excitability. Measurements of responsiveness to intracellular stimulation which are now being carried out will give us the answer to this question shortly. A FUNCTION FOR DENDRITES
It seems to us that this ability of dendrites to undergo extensive depolarization without propagated impulses is fundamental to the role played by axo-dendritic synapses in the integrative function of motor neurones. If this were not true -if dendrites were able to propagate impulses to the cell body and beyond -then axodendritic synapses would be equal to axo-somatic synapses in their power to secure “driving” of the motor neurone. Under such conditions, the motor neurone would be capable of functioning only as a simple relay and all sensorimotor integration would perforce be shifted upstream in neuronal systems. The ability of dendritic processes and membrane to receive extensive depolarizing input, and thereby to shift the somatic membrane potential without firing the cell establishes a mechanism whereby presynaptic activity delivered to dendrites can modulate excitability throughout an extensive range below the firing threshold. In proposing that dendritic extensions which do not propagate all-or-none impulses express their function by modulating excitability of the trigger zone without firing it, we are not challenging the validity of some inferences that dendrites may propagate action potentials (Cragg and Hamlyn 1955; Spencer and Kandel 1961). It does not seem probable to us that the transition from somatic membrane characteristics to dendritic membrane characteristics occurs abruptly a t a specific boundary somewhere on the cell processes. It seems more reasonable to suppose that the transition from electrically inexcitable to electrically excitable membrane occurs gradually, and may do so at different distances from the cell in different varieties of neurones. The data
INTEGRATIVE FUNCTION OF THE MOTOR NEURONE
59
of Spencer and Kandel (1961) indicate that the development of a dendritic action potential in the proximal portions of the dendrites of hippocampal pyramids adds an abrupt increment to synaptically induced depolarization and shifts the membrane potential past the firing level of the trigger zone. Thus it would appear that when the level of depolarization of dendritic membrane is reached at a portion of the dendrite which can produce a propagated response, cell firing is thereby ensured. Spencer and Kandel (1961) observed the signs of dendritic spikes in about 25 per cent of the cells they studied. They propose that the proximal dendritic spikes act as a booster for “otherwise ineffectual dendritic synapses”. By using the term “ineffectual” we are sure that Spencer and Kandel(l961) did not mean to imply that dendritic synapses were without any influence on cell functions. We would suggest that in some cells the structural characteristics, which are not uniform from cell to cell, were conducive to the extension of excitable membrane farther out on the dendritic stumps where it could function as a booster zone. In the majority of cells in which a booster zone was not present, impulses delivered to dendrites could still modulate excitability without producing discharge because of the high threshold of dendritic membrane. Thus, even though dendritic synapses might be ineffectual in securing discharge they could still be important to the integrative behavior of the motor neurone. We do not consider the disposition of synaptic terminals on distal dendrites to be the only mechanism whereby the motor neurones might be permitted to engage in integrative behavior. It seems to us highly probable that wide variations occur in the potency of individual synaptic terminals. Certainly the variations in synaptic structure must have some meaning. It is generally accepted that small, relatively impotent terminals on a soma might be incapable of inducing a threshold depolarization. If a sufficient number of such terminals were simultaneously active, integrative summation and firing would occur at strictly somatic loci. However, to accept this as the sole and only mechanism whereby integration could occur leads us into two difficulties. First, the somatically induced EPSP is brief compared to the dendritically induced EPSP delayed by electrotonic propagation. If each presynaptic impulse had only a brief influence on somatic membrane potential, then the frequency required for the production of a given level of summation would be higher and the limitations imposed on presynaptic systems would be greater than would otherwise be the case. Second, the membrane area of the somatic portion of the neurone has been estimated to be only a small fraction of the total area of the cell (Sholl 1956). If integration at somatic levels were the only possibility, the small area of the somatic membrane would operate to restrict the number of synaptic contacts to one-tenth or onetwentieth the number which could be made on dendrites and soma together. Such a serious restriction in number would reduce the variety of inputs capable of influencing motor neuronal behavior and would leave the dendritic synapses without any role to play in the regulation of neuronal functions. In summary, these experiments offer support for the validity of the suggestion made on theoretical grounds by Coombs et al. (1957) that the high threshold of dendritic portions of the motor neuronal membrane is a characteristic which is of considerable importance to the ability of motor neurones to exhibit integrative behavior. The Rc1crenc.s P. 6 0 4 1
60
J. M. BROOKHART, K. KUBOTA
difference between the point of view expressed here, and that so frequently encountered, is primarily one of attitude. Whereas others regard the characteristics of dendrites as detrimental to their ability to control impulse initiation, we are inclined to view these characteristics as an attribute by virtue of the part they play in support of integrative behavior. I cannot leave this subject without pointing out that this line of thinking leads us into a problem area of a more philosophical nature. We are dealing with the notion that the spatial arrangement of synapses on a neurone and its processes is of importance in determining the manner in which the neurone responds to its input. In terms of the behavior of the whole organism, this means that synaptic contacts cannot be made at random. It suggests that synaptic contacts destined to drive the recipient neurone as a relay should be made on or near the cell body; whereas synaptic contacts intended to modulate excitability and permit neuronal integration should be made on dendritic membrane distant from the cell. If this is an acceptable notion, then we must some day address our attention to the organizing force, the guiding hand, which leads the proper presynaptic terminals to the proper points on the neurone. I must admit that no experimental approaches to these problems have suggested themselves as yet. REFERENCES ALANIS,J. Effects of direct current on motor neurons. J. Physiol. (Lond.), 1953, 120: 569-578. ARAKI,T. Effects of electrotonus on the electrical activities of spinal motoneurons of the toad. Jap. J. Physiol., 1960, 10: 518-532. ARAKI,R. and OTANI,T. Accommodation and local response in motoneurons of frog’s spinal cord. Jap. J. Physiol., 1959, 9 : 69-83. BARRON, T. H. and MATTHEWS, B. H. C. The interpretation of potential changes in the spinal cord. J. Physiol. (Lond.), 1938, 92: 276-321. BISHOP,G. H. Natural history of the nerve impulse. Physiol. Rev., 1956, 36: 376-399. BONNET,V. et BREMER, F. Les potentiels synaptiques et la transmission nerveuse centrale. Arch. int. Physiol., 1952, 60: 33-93. BROOKHART, J. M. and FADIGA, E. Potential fields initiated during monosynaptic activation of frog motor neurons. J. Physiol. (Lond.), 1960, 150: 633-655. BROOKHART, J. M., MACHNE, X. and FADIGA,E, Patterns of motor neuron discharge in the frog. Arch. ital. Biol., 1959, 97: 53-67. COOMBS, J. S., CURTIS, D. R. and ECCLES,J. C. The generation of impulses by motoneurones. J. Physiol. (Lond.), 1957, 139: 232-249. CRAGG,B. G . and HAMLYN, L. H. Action potentials of pyramidal neurons in the hippocampus of the rabbit. J. Physiol. (Lond.),1955, 129: 608-627. CURTIS,D. R. and ECCLES, J. C. Synaptic action during and after repetitive stimulation. J. Physiol. (Lond.), 1960, 150: 374-398. DELORENZO, A. J. Electron microscopy of the cerebral cortex: 1. The ultrastructure and histochemistry of synaptic junctions. Bull. Johns Hopk. Hosp., 1961, 108: 258-279. DUDEL, J. and KUFFLER, S. W. Presynaptic inhibition at the crayfish neuromuscular junction. J. Physiol. (Lond.), 1961, 150: 543-562. ECCLES, J. C. The physiology of nerve cells. Johns Hopkins Press, Baltimore, 1957, 270 pp. ECCLES,J. C. The properties of dendrites. In D. B. TOWERand J. P. SCHADB(Editors), Structure and function of the cerebral cortex. Elsevier, Amsterdam, 1960, 192-203. ECCLES,J. C., ECCLES,R. M. and LUNDBERG, A. Durations of after-hyperpolarization of motor neurons supplying fast and slow muscles. Nature, 1957, 179: 866-868. ECCLES, J. C., ECCLES, R. M. and LUNDBERG, A. Types of neurone in and around the intermediate nucleus of the lumbosacral cord. J. Physiol. (Lond.), 1960, 154: 89-114.
INTEGRATIVE FUNCTION OF THE MOTOR NEURONE
61
ECCLES, J. C. and RALL,W. Repetitive monosynaptic activation of motoneurons. Proc. roy. SOC.3, 1951, 138: 4 7 5 4 9 8 . FADIGA,E. and BROOKHART, J. M. Synaptic activation of different portions of the motor neuron membrane. Amer. J. Physiol., 1960,198: 693-703. FATT,P. and KATZ,B. An analysis of the endplate potential recorded with an intracellular electrode. J. Physiol. (Lond.), 1951,115: 320-370. FRANK, K. and FUORTES, M. G. F. Presynaptic and postsynaptic inhibition of monosynaptic reflexes. Fed. Proc., 1957,16: 3 9 4 0 . GESELL, R. A neurophysiological interpretation of the respiratory act. Ergebn. Physiol., 1940,43: 477-639. GRANIT, R. and PHILLIPS,C. G. Effects on Purkinje cells of surface’stimulation of the cerebellum. J. Physiol. (Lond.), 1957,135: 13-92. GRANIT, R. and RUTLEDGE, L. T. Surplus excitation in reflex action of motoneurones as measured by recurrent inhibition. J. Physiol. (Lond.), 1960,154: 288-307. GRUNDFEST, H. Electrical inexcitability of synapses and some consequences in the central nervous system. Physiol. Rev., 1957,37: 337-361. HUNT,C. C. and KUNO,M. Background discharge and evoked responses of spinal interneurones. J. Physiol. (Lond.), 1959,147: 364-384. KOLMODIN, G. M. Integrative processes in spinal interneurones with proprioceptive connections. Acta physiol. scand., 1957,40: 1-89. LIU, C. N. and CHAMBERS, W. W. Experimental study of anatomical organization of frog’s spinal cord. Anar. Rec., 1957,127: 326. MACHNE, X., FADIGA,E. and BROOKHART, J. M. Antidromic and synaptic activation of frog motor neurons, J. Neurophysiol., 1959,22: 483-503. MORRELL, R. M., FRANK, K. and FUORTES, M. G. F. Site of origin of motoneurone rhythms. XXth internarional physiological congress, Brussels, 1956,pp. 660-661. SALAy PONS,C. Estructura de la medula espinal de 10s batracios. In Trabajos del Laboratorio de Znvestigacion Biologku. Barcelona, 1892,22 pp. SHERRINGTON, C . S. The correlation of reflexes and the principle of the final common path. A.R. Brir. Ass., 1904,7 4 : 728-741. SHOLL,D. A. The organization of the cerebral cortex. J. Wiley and Sons, New York, 1956, xvi, 125 pp. E. R. Electrophysiology of hippocampal neurons. IV. Fast prepoSPENCER, W. A. and KANDEL, tentials. J. Neurophysiol., 1961,24: 272-285. THESLEFF, S. Motor endplate “desensitization” by repetitive nerve stimuli. J. Physiol. (Lond.), 1959,148: 659-664. VESZI,J. Der einfachste Reflexbogen im Riickenmark. Z. allg. Physiol., 1910,II:168-175. WALL,P.D. Repetitive discharge of neurons. J. Neurophysiol., 1959,22: 305-320. WASHIZU, Y. Single motoneurons excitable from two different antidromic pathways. Jap. J. Physiol., 1960,10: 121-131. WILSON,V. J. and BURGESS, P. R. Intracelfular study of recurrent facilitation. Science, 1961, 134: 337-338.
62
J. M. BROOKHART, K . KUBOTA
DISCUSSION
J. C. ECCLES: (1) I was particularly interested in Professor Brookhart’s statement that he has not been able to find any postsynaptic inhibitory potentials in frog motoneurones. It occurs to me that presynaptic inhibition may provide the principal inhibitory mechanism in the frog. The large dorsal root potentials show that presynaptic depolarisation is very large. Furthermore, the negative feed-back of motoneuronal discharge does not occur via Renshaw cells, but by presynaptic inhibition, as is revealed by the dorsal root potential produced by ventral root volleys. (2) I assume from the records showing a very clear IS-SD separation of the motoneuronal spike potential that, just as with mammalian motoneurones, impulse generation occurs in the initial segment, both with dorsal root and lateral column stimulation. The large SD spike shows that a considerable part of the soma-dendritic membrane is electrically excitable. However the large depression of the spikes during repetitive synaptic activation shows that impulse propagation progressively fails over the SD membrane as it becomes more and more depolarised by the cumulative synaptic action. This would accord well with your suggestion that there is normally a very poor safety factor for this propagation.
F. BREMER: The peculiar histological structure of the spinal cord in batracians invites those interested in the physiological significance of neuron dendrites to institute experimental investigations which our colleague Brookhart must certainly have considered. I should like to ask him whether it would be technically possible using the elegant method which he describes - to eliminate from spinal functioning the dendritic arborizations of the motor neurons which, in frogs, extend into the whitc matter, and to see whether an elementary spinal reflex would subsequently occur. In the case of superficial dendritic arborizations of the neocortex, my experiments in destruction (by controlled thermocoagulation, paralysis or by topical application of depressor agents) have had no significant effect on the voltage and the form of the exosomatic component of the elementary evoked potential, nor on the facilitating effect of the reticular influx of arousal on this component. 1 believe that these negative results are a warning for those inclined to accept the currently popular conception which attaches fundamental importance to the dendrites in the regulation of elementary neuronic functions.
D. ALEE-FESSARD: I should like to mention an anatomical finding which was recently communicated to me by Dr. Bowsher of Liverpool. Using the Glees method in studying synaptic contacts at the level of thc primary somaesthetic thalamic relay on the one hand, and converging nuclei on the other hand, he found that the connections in the primary relay are both axo-dendritic and axo-somatic (the former being predominant), whereas the contacts in the converging nuclei are essentially axosomatic. I n the thalamic relay, therefore, there exists a duplicate system of connections reminiscent of the features which you described for the cord. It is interesting that this type of connection is found in a relay which must transmit information while at the same time maintaining good point-by-point correspondence between the periphery and the cortical projection. The axo-dendritic afferents, which are least well placed to engender the start of impulses, are probably efficacious only by virtue of a degree of integration of impulses of
INTEGRATIVE FUNCTION OF THE MOTOR NEURONE
63
the same origin; this is probably effected in the primary relay, and it preserves a certain somatotopic organization. In heterogeneous converging nuclei, where the density of afferents of the same category is only slight, the axo-somatic disposition of synaptic contacts ensures at least new integration. 0. POMPEIANO:
The findings of Prof. Brookhart again underline the importance of the concept that synapses on dendrites are not functionally equivalent to synapses on the perikarya. It has also been shown in mammals that there are different types of synaptic contact among afferents to the same population of neurones. One typical example is given by the lateral vestibular nucleus in the cat. Recent anatomical studies made with the silver techniques (Brodal et al. 1961) have shown that the primary vestibular fibres to Deiters’ nucleus have mainly axo-dendritic synapses on the small cells. The spinal afferents, on the other hand, synapse only with perikarya of giant Deitersian cells. There are also particularly striking differences between the cortical cerebello-vestibular and the fastigio-vestibular fibres with regard to their pattern of termination in Deiters’nucleus. The former end on the dendrites of giant cells while the majority of the fastigio-vestibular fibers synapse with the perikarya and dendrites of the small cells. A knowledge of the synaptic arrangements in the Deiters’ nucleus, as well as in other structures, is of great functional interest. BRODAL, A., POMPEIANO, 0. and WALBERG, F. The vestibular nuclei and their connections. Anatomy and functional correlations. Oliver and Boyd, Edinburgh, 1961. A. FESSARD:
I am wondering if the differences you showed to exist in effectiveness-that is in amplitudes, latencies, etc. - between activations either through a dorsal root or through the lateral column, may not be due, at least in part, to differences in afferent impulse densities characterizing these pathways; in other words, if the number of fibers involved in either case is equivalent or not; and, if not, if one cannot make this an important determining factor for the differences observed, assuming the occurrence of spatial summation at synapses, as is most likely.
BROOKHART’s replies To J. C. Eccles
With regard to the absence of inhibitory responses, I should say that it is only in the intracellular studies that this remark is applicable. We have only recently learned how to evoke reciprocal inhibition of ventral root reflex discharges through the stimulation of specific nerve trunks innervating specific muscles. The experiments now in progress should enable us to examine the intracellular events accompanying this kind of inhibition as well as that variety described by Bremer following contralateral dorsal root stimulation. I am not ready yet to accept the notion that the subsidence of excitatory depolarization during repetitious D R stimulation is due to presynaptic inhibition. When the sign of the afferent volley is recorded focally during this kind of stimulation, it undergoes no change even though the postsynaptic response undergoes diminution. If my understanding is correct, presynaptic inhibition should be reflected by a reduction of this sign of activity in parallel with the reduction of postsynaptic response. To F. Bremer Prof. Bremer’s kind remarks give me the opportunity to re-state the pattern of my own thinking. I would expect that the electrotonic prolongation of dendritically induced EPSP’s is conducive to their smooth summation so that a continuous asynchronous arrival of presynaptic impulses could readily lead to a sustained tonic effect on neuronal excitability. The example which Prof. Bremer cites, wherein no modification of surface-positive portions of an evoked cortical response occurs in the absence of any surface negativity assignable to dendrites
J. M. BROOKHART, K . KUBOTA
is a very interesting one. However, the synaptology involved in such a phenomenon is complex and poorly understood. 1 would not venture to try to explain the observation without a great deal more anatomical information about the presynaptic systems involved. We have not directed our attention to a specific test of this sort of situation in the frog. However, early in the series of experiments involving intracellular recording, we became convinced that the microdissection of the pial investment of the cord, necessary to permit penetration by the micropipette, had to be done with extreme delicacy. It is still our firm impression that good reflex and lateral column responses from a preparation can be seriously affected by removal of the pia in such a manner as to damage the underlying superficial marginal plexus of dendrites. Consequently, although we are only at the level of gross observation, and have no rigid analysis of our impressions, we are convinced that damage to this marginal plexus has a deleterious effect upon the excitability of the related motor neurones.
Joint reply to D . Albe-Fessard and to 0 .Pompeiano There are indeed interesting and important anatomical observations. It is important, in my estimation, to devise experiments which will clearly reveal the differences between the capabilities of these two varieties of presynaptic connections. To D . Albe-Fessard No comment on anatomic observation of VPL. See joint response to Albe-Fessard and to Pompeiano.
To A . Fessard Prof. Fessard has put his finger on a soft point in this whole structure. However, there are fairly good reasons for discounting the importance of spatial summation in the case of the lateral column responses. In the first place, we have the impression that the lateral column bundle is quite small. It is necessary to search carefully with small penetrating electrodes in order to find the proper focus for the elicitation of a clean lateral column response of short latency and free of after-discharge. Consequently, it seems to us that only a relatively small number of descending fibers is involved. Secondly, once the appropriate stimulus site has been found, very fine control of stimulus intensity is required in order to evoke only an EPSP without firing the motor neurones. Minimal increases from an intensity which is adequate to initiate a threshold EPSP will evoke a discharge. Consequently, while we have a strong impression that this system does not require much in the way of spatial summation, we still lack the critical numerical data. When records are made from appropriate foci, the focal signs of the presynaptic volley can be compared in terms of amplitude. The dorsal root presynaptic volleys are always seen to be much larger than those aroused by lateral column stimulation even when the postsynaptic responses are equivalent in magnitude. This is an additional reason for believing that spatial summation is not of great importance in establishing the powerful control of discharge exerted by lateral column fibers. Finally, in studies of EPSP’s alone, of a value subthreshold for discharge, we have found that the patterns of summation initiated over these two pathways are independent of the magnitude of the EPSP initiated by a single volley. The differences in patterns of LC- and DR-EPSP summation remain the same for submaximal through maximal EPSP values.
The Plasticity of Human Withdrawal Reflexes to Noxious Skin Stimuli in Lower Limbs K.-E. HAGBARTH
AND
B. L. FINER
Department of Clinical Neurophysiology, University of Uppsala, Uppsala (Sweden)
INTRODUCTION
A suddenly applied noxious skin stimulus in man is apt to elicit, not only a sensation of pain accompanied by startle and autonomic responses, but also a specifically patterned movement which tends to cause a withdrawal from the offending object. It is well-known that nociceptive withdrawal responses occur also in spinal preparations, the ipsilateral flexion and the crossed extension reflex serving as classical examples of such spinal defence mechanisms. Little is known, however, about the way in which such spinal nociceptive reflexes are normally integrated into the total adaptive defence behaviour of an intact human being. The electromyographic recording technique has proved to be a valuable tool in studies devoted to this problem, because it not only informs us about the strength and the temporo-spatial patterns of the muscular responses to the noxious stimuli, but it also allows accurate latency measurements, which may tell us something about the length of the reflex arcs involved. Recent electromyographic (EMG) studies of the human abdominal skin reflexes led to the conclusion that, even though these reflexes are spinal, they possess a certain degree of functional adaptability, which implies a type of learning in some of the supposedly most stable pathways of the neuraxis (Kugelberg and Hagbarth 1958; Hagbarth and Kugelberg 1958; cf. also Teasdall and Magladery 1959). We have now continued to study the adaptive behaviour of nociceptive cutaneous reflexes in man, but this time we have focussed our attention on ipsilateral hind limb reflexes. Previous EMG studies of these reflexes have shown that their latencies and patterns largely correspond to those of spinal flexion reflexes (Kugelberg 1948; Hoffniann et al. 1948; Dodt and Koehler 1950; Pedersen 1954). More recently, however, the receptive fields for individual muscles have been systematically analysed (Hagbarth 1960 b; CJ also Kugelberg et al. 1960) and a short review of these previous results is necessary, because they constitute a basis for the present study. Fig. 1 shows how the vastus medialis (an extensor of the knee) reacts when noxious electrical skin stimuli of short duration (20 msec) are applied to different areas of the limb. In order to visualize, not only excitatory, but also inhibitory, responses, a slight voluntary background contraction was maintained in the muscle. It can be seen that stimuli applied to the leg or the back of the thigh caused an initial inhibition followed References P. 77-78
66
K.-E. MAGUARTH, B. L. FINER
Fig. 1 Reflex responses in the vastus medialis (shadowed) to noxious skin stimuli within various regions of the limb (as indicated by radiating lines). Black bars under records show the duration of the stimuli. The lower middle record shows the control background activity in the muscle. Three to five sweeps are superimposed in each record. Time: 100 c/sec. (Hagbarth 1960 b).
by a rebound, while stimuli on the ventral aspect of the thigh caused an initial reflex discharge followed by a silent period. Occasionally, the cyclic activity persisted for a longer time and involved several alternating periods of excitation and inhibition, but attention was paid only to the initial reflex effect, which indicated by its short latency (60-80 msec) that it was a primary spinal response to the signal. In the drawings of Fig. 2, the skin areas which yielded primary excitation are indicated by and those which yielded inhibition are indicated by -. The first three drawings show the distribution of the positive and negative skin areas for a hip-, kneeand ankle-extensor respectively and the last drawing shows how the flexor muscle, the tibialis anticus, an antagonist to the gastrocnemius-soleus, responds to the stimuli. It should be noted that the tibialis anticus responds in a reciprocal way to the gastrocnemius-soleus: it is activated from those skin areas which inhibit the extensor and vice versa.
+
Gluteus maximus
Vastus medialis
Gastrocsoleus
Tibialls anticus
Fig. 2 Diagrams showing the approximate extent of inhibitory and excitatory skin areas for the muscles investigated. (Hagbarth 1960 b).
PLASTICITY OF HUMAN WITHDRAWAL REFLEXES
67
These results in man agree fairly well with those obtaiiled in spinal cats (Hagbarth 1952) and they can be summarized by stating that extensor muscles are inhibited from most parts of the limb, but are activated by stimuli on the skin which covers the muscle itself, while antagonistic flexor muscles respond in a reciprocal manner. This rule does not hold in a strictly anatomical sense but it may serve as an approximation in attempts to analyze what the patterns mean in terms of function. It is probable that the main biological function of these ipsilateral nociceptive reflexes is to cause a rapid movement away from an offending object and a closer study of the drawings in Fig. 2 reveals that the patterns suit this purpose amazingly well. The flexion movements which regularly occur at joints proximal to the stimulus represent the classical flexion reflex, which has a definite avoidance capacity; the protective function of extension movements in joints distal to the stimulus becomes equally evident if it is assumed that the subject is, at the moment of the stimulus, standing on his limb with the joints in slight flexion. In this situation, extension of the hip causes withdrawal from a gluteal stimulus and in a similar way knee extension causes withdrawal from an offending object on the front of the thigh, while ankle extension removes the leg from a calf stimulus. Because the avoidance capacity of a motor pattern depends on the initial support and position of the limb, the question arises as to whether the withdrawal patterns are plastic enough to adapt themselves to different positions. It was noted that the spinal responses described were not easily readjusted in this way, so that, for certain positions, they seemed definitely inappropriate. On the other hand, the joint movements which were actually observed in response to the stimuli seemed to change their direction in an appropriate way according to the position of the limb, so that they always resulted in an effective escap: from the stimulus. This indicated that there is, in addition to the initial spinal component of the withdrawal reaction, another later component which has a more plastic motor pattern. The present study confirmed this supposition. The opportunity was taken to compare the spinal withdrawal reflexes with the later, probably cerebral, reactions and to see how they differ in respect to functional adaptability. The main project was to analyze to what extent the nociceptive reactions can be modified or reorganized by experience and training and to what extent they can be affected by hypnotically induced analgesia or hyperalgesia of the skin. TECHNIQUE
Some of the experiments were performed on ourselves, others on medical students or adult patients who showed no obvious signs of organic neurological disease. Noxious electrical stimuli were applied somewhere on the skin of one of the lower limbs, the limb being held for each trial in a certain, predetermined position. The stimuli were obtained by letting an electrical current (square wave pulses at 500-1000/sec) pass through the sharp tips of a small bipolar needle electrode held in contact with the skin. A current of 5-10 mA was usually sufficient to cause an intense burning sensation of the skin and a withdrawal reaction of the extremity. Sometimes the stimuli had Rcferencrs p . 77-78
68
K.-E. HAGBARTH, B. L. FINER
a fixed duration of 20-40 msec, but in some of the training experiments the stimulating current was continuous and it was the withdrawal movement itself which brought the stimulus to an end by removing the limb from the electrodes, or alternatively a countermovement was required to push a switch which turned off the stimulating current in the electrodes attached to the limb. For the sake of simplicity, responses were recorded only from one-joint muscles, such as the soleus or vastus medialis, and the skin stimuli were restricted to the calf or the front of the leg. Surface electrodes or bipolar needle electrodes were used for recording and after adequate amplification the muscle potentials were displayed on the beam of a cathode ray oscilloscope with a sweep triggered by the electrical skin stimulus. During the hypnosis, which was conducted by B.L.F., the subject was told that the limb was either analgetic, so that the electric stimuli could hardly be felt, or alternatively that the limb was hyperalgetic, so that each stimulus was extremely painful. The subjects used in these experiments had been hypnotized earlier and it was known that in this state they were able to develop subjective analgesia or hyperalgesia to pin prick. RESULTS
Responses to noxious stimuli of short, fixed duration In these experiments, the avoidance reaction itself did not affect the duration of the stimulus which was fixed at 20-40 msec. The initial reflex responses to the stimuli were often followed by alternating periods of excitation and inhibition and it was often hard to distinguish any separate late component of the response which might be of cerebral origin. The study in these experiments was therefore restricted to the initial, spinal reflex effects. A constant finding was that the amplitude of the responses varied with the intensity of the voluntary background activity in the muscles. Reflex muscular discharges were facilitated by a moderate voluntary contraction of the muscle itself and inhibited by voluntary tonic activation of an antagonist. Such reflex changes due to variations in the excitability of the motor system were avoided in the present study. A constant voluntary background activity in the muscle was taken as a criterion that the excitability of the motoneurones did not vary during the experiment. One of the basic features of the spinal abdominal skin reflexes is that they easily become habituated to repeated stimuli (Hagbarth and Kugelberg 1958) and a response decline of this kind was also frequently found in the present study. Fig. 3, A , is derived from a subject who showed marked habituation of the extension reflexes appearing in the soleus in response to moderately intense calf stimuli. After 24 stimuli, applied at an interval of about 10 sec, the response had declined to a small fraction of its initial amplitude. At this stage, a single stimulus of higher intensity was applied without disrupting the continuity in the stimulus sequence. It caused a brisk reflex and, in addition, a dishabituation which was shown by the large responses to the succeeding, moderate stimuli.
PLASTICITY OF HUMAN WITHDRAWAL REFLEXES
69
Fig. 3 A : Habituation of reflex discharges in the soleus evoked by moderate skin stimuli on the calf, which
were regularly repeated at 6/min. A single intense stimulus in the 25th trial causes dishabituation, as is shown by the large responses to the succeeding moderate stimuli (numbers corresponding to trials after the intense stimulus). B: The same reflex when the subject himself elicits the stimuli (upper record), and when this is done by another person (lower record). Four sweeps superimposed in each record. See Text. Time: 50 c/sec.
A
+
B
Control Hysterical analgesia
r*c(p)-wk.Control
T-
vol. contraction
o n slgnal
Fig. 4 A : Hypnotically induced changes in the responses of the soleus to moderate calf stimuli. The early reflex discharges were almost abolished during suggested analgesia and increased during suggested hyperalgesia, as compared to the controls during pre- and posthypnotic wakefulness. The lowest record shows latency of rapid voluntary contraction on signal (weak calf stimulus). B : Soleus reflexes to calf stimuli in an hysteric patient with leftsided decreased sensitivity to pin-pricking. Upper record: reflex in right (normal) limb. Lower record: stimulus of similar intensity causes no reflex in analgetic limb. Time: 50 c/sec. References p. 77-78
70
K.-E. HAGBAKTII, B. L. FINER
It was also noted that the amplitude of the reflexes depended on the subject’s state of attention and his apprehension of the stimuli. In the experiment illustrated by Fig. 3, B, the subject himself was allowed to push the button which elicited the stimuli (the resulting responses are superimposed on theupper record). In the next four trials the stimuli were elicited by another person, and the subject did not know exactly when to expect them, but he knew that the intensity might be increased (lower record). Even though the intensity remained constant, the responses in the second trial were much larger than those in the first and the constant background activity showed that this change was not correlated with any alteration in the motor tension. Fig. 4, A , is derived from a subject who was easily hypnotized and who rapidly developed subjective analgesia or hyperalgesia to hypnotic suggestion. The hypnosis and the suggestions were tried on many other subjects, who also reported subjective changes in perception of this kind during the experiments, but most of them showed no systematic changes in the amplitude of the spinal reflexes in correspondence to the perceptual change. Ln this particular subject, and in a few others, however, such a correlation was observed. The reflexes investigated were the initial discharges appearing in the soleus in response to moderate calf stimuli. Each record in Fig. 4, A , is composed of four superimposed sweeps and the records show that the responses were, in comparison with the control responses before and after the hypnosis, decreased during hypnotic analgesia and increased during hypnotic hyperalgesia. Again, there was no major change in motor tension to account for the results which, in this instance, were obtained with great regularity each time the experiment was repeated. In a final test, the intensity of the stimulus was decreased so that it was just perceptible and the subject was instructed that on the signal he should contract his calf muscles with the shortest possible delay. The latency of the discharge in this reaction time experiment was about 200 msec as compared with the short latency of the spinal reflex, which was about 70 msec. In other subjects, the voluntary reaction time was shorter, but it was never seen to fall below 120 msec. A 61-year old woman, a patient from the neurological clinic, was the subject in one experiment. She reported decreased sensibility to skin stimuli of all modalities on the left side of her body, a disorder which was interpreted to be of a functional nature. The upper record in Fig. 4, B, shows her soleus response to a calf stimulus in the right limb, where she reported normal sensibility and the lower record shows that no response appeared in her left soleus when a skin stimulus of similar intensity was applied to the calf of this anaesthetic limb. On this side, a marked increase of the stimulus strength was required in order to get a reflex comparable to that in her right leg.
Responses to noxious stimuli with a duration determined by the defence reaction itself It has been mentioned in the introduction that the spinal nociceptive reflexes do not easily change their patterns according to the initial position of the limb and that therefore they do not always serve as appropriate withdrawal responses. The drawing in Fig. 5 shows a position of the limb in which a stimulus on the calf must cause it contraction of the knee extensors in order to induce a leg movement away from the stimulus. The initial reflex effect of a brief calf stimulus is, however, an inhibition of
PLASTICITY OF HUMAN WITHDRAWAL REFLEXES
71
Control
Fig. 5 In the limb position indicated, a short calf stimulus of fixed duration (30 msec) causes early reflex inhibition in the vastus medialis (upper record). A long stimulus, the duration of which is determined by the withdrawal movement, causes early inhibition, succeeded by heavy discharge (middle record). During hypnotic analgesia, the early inhibition remains, while the late discharge is abolished. Time: 50 clsec.
the activity in the vastus medialis, which apparently tends to cause a knee-flexion and a leg movement towards the stimulus (upper record). This early effect of the calf stimulus was also the same in those experiments in which the stimulating current was continuous and the subject could avoid it only by removing the leg from the electrodes (which initially were in contact with the skin, but not attached to it). When such a stimulus occurred, however, no flexion at the knee joint could be observed, but instead the subject rapidly removed this leg from the electrodes by an extension of the knee (as well as flexion of the hip), and the only electromyographic sign of this extension was a late, heavy discharge in the vastus medialis, which succeeded the early inhibition and had a latency of about 120 msec (middle record). Rec. n.
NVVVVVVVV\h
Fig. 6 Responses by the vastus medialis to noxious skin stimuli on the calf and the front of the leg. Duration of stimuli (outlasting the sweep) determined by the withdrawal movements. Positional changes cause adaptive sign reversal of the late responses, while early inhibition remains. Time: 50 cisec. Keferences o 77-78
72
K.-E. HAGBARTH, B. L. FINER
This late discharge could not be regarded as being a rebound, because its amplitude showed no correlation with the intensity of the preceding inhibition. The records of Fig. 5 were derived from a subject who easily developed subjective analgesia to hypnotic suggestion. During his analgetic state the late discharge in the vastus medialis had disappeared and he showed little tendency to remove his leg from the stimulus. The early inhibition in the vastus medialis remained, however, apparently uninfluenced by the suggestion (lower record). This may serve to exemplify the observation that, as a rule, the initial spinal reflex effect was more resistant to hypnotic suggestion than the late component of the avoidance reaction was. The latter showed amplitude variations which run more in parallel with the changes in the subjective estimation of the stimulus intensity. Fig. 6 shows how skin stimuli both on the calf and on the front of the leg evoke an early inhibition in the vastus medialis (cf. Fig. 2) and that this early effect (which has a duration of at least 20 msec) is largely independent of the initial position of the limb. This early tendency to flex the knee can apparently serve as a withdrawal response to stimuli on the front of the leg in the sitting position (upper left record), or to stimuli on the calf in the standing position (lower right record), but in the other two situations (illustrated by the upper right and lower left record) an initial discharge in the vastus medialis (extension of the knee) would better serve a withdrawal function. In contrast to the initial spinal effect, the late response (which has a latency of about 120 msec and a duration of 70-90 msec) changes its sign in the appropriate way according to the position of the limb: the late inhibition, which adds to the early inhibition in the upper left and lower right records, is reversed to a discharge in the upper right and lower left records, where it seems to compensate for the maladjusted initial response. Repeated application of stimuli to the calf in the sitting position did not result in a reversal of the initial effect. The early inhibition persisted, even thoughit occasionally showed a tendency to become habituated. The training experiment shown in Fig. 7 gives another example of how the late avoidance reaction adapts itself to a new situation more easily than the spinal withdrawal response does. In this experiment the stimulating electrodes were firmly attached to the front of the leg in the sitting position and the recording electrodes were inserted into the vastus medialis. The stimulating current was continuous but could be turned off by a switch held in front of the tibia, so that the subject, in order to end the stimulus, had to extend his knee or, in other words, make the opposite movement to a withdrawal. The stimuli were quite strong and sufficiently unpleasant to make the subject anxious to turn them off as soon as possible. They were repeated at jntervals of about 15 sec. The response to the first stimulus had about the same shape as the upper left record in Fig. 6: there was the characteristic early inhibition, succeeded by a second inhibitory phase which indicated flexion of the knee. However, a very late discharge seemed to appear at the end of the record. For each of the succeeding stimuli, the delay of this late discharge became gradually shorter. The discharge impinged upon the second inhibitory phase until, in the fourth trial, a substitution was completed. Now, the response had the same shape as the upper right
PLASTICITY OF HUMAN WITHDRAWAL REFLEXES
+.
73
4
250-253
Fig. 7 Responses by the vastus medialis to noxious skin stimuli on the front of the leg, which can be interrupted, as the drawing shows, by knee extension. After four trials the sign of the late response is reversed to fit the new situation, while the early inhibition remains after more than 250 trials, although it becomes less distinct. Time: 50 c/sec.
record in Fig. 6 and showed the normal avoidance response to a calfstimulus in the sitting position. This experiment was continued in serial sessions during the course of a month, during which time more than 500 stimuli were applied. This was done in order to find out whether the initial inhibition could also be converted by intense training into discharge. This did not, however, occur. The latency of the late discharge never became shorter than 110 msec. The early inhibition became progressively less distinct and in some of the late trials it was hardly visible; but the effect was not reversed (lowest record). Similar results were obtained in training experiments in which dorsiflexion of the ankle was required to break electrical stimuli on the calf. In this instance, the initial discharge in the ankle extensors remained; it showed a response decline, but it was not substituted by inhibition. DISCUSSION
The results show that the defence reactions to nociceptive skin stimuli in the human lower limbs involve two kinds of responses, which differ both in respect to latency and to functional adaptability. The latency of the early response is short enough to indicate a spinal reflex arc (cf. Kugelberg et al. 1960; Hagbarth 1960 b), while the latency of the late response equals that of a rapid voluntary reaction. It seems, however, inappropriate and rather meaningless to say that it is a voluntary reaction : the responses were automatic in the sense that the subjects themselves were unable to suppress them, especially when the stimuli were strong. It is probably more fruitful to regard them as conditioned, nociceptive cerebral reactions with an adaptive motor References p . 77-78
74
K.-E. HAGBARTH, B. L. FINER
pattern, which has developed largely through experience: an inappropriate or “faulty” defence reaction to a noxious stimulus will automatically receive “punishment” i n that the stimulus tends to remain, while a correct reaction is rewarded by a more rapid end to the stimulus. The results here recorded agree with the presumption that the motor patterns of the late responses are organized on such a functional basis. They appeared especially in those experimental situations in which the duration of the stimulus was determined by the response and in which their patterns changed according to the position of the limb, so that the final result of the motor act was always an interruption ofthe stimulus. Usually this motor act implied a movement which removed the leg from the electrodes (flexion or extension, depending on the initial position); but, if a counter-movement towards the electrodes was required to end the stimulus, the sign of the late response was reversed after a few trials to fit the new situation. Consequently, it is not a simple withdrawal reflex in the true sense of the word, but a highly organized defetice reaction with a sensory-motor pattern determined largely by returning sensory information concerning its avoidance capacity (cf: Anokhin 1961). If it is true that these nociceptive defence reactions are learned, it must be expected that they are absent in infants without experience of noxious stimuli. This is supported by the experiments of Melzack and Scott (1957), which showed that dogs reared in isolation with restricted sensory experience are unable to perform proper defence reactions to nose-burning and pin-pricking. The latency of the late response does not allow any conclusions as to whether it is mediated at a cortical or subcortical level, but the marked postural dominance indicates that its central arc has intimate functional correlations with midbrain and cerebellar structures (see Dow and Moruzzi 1958). Pavlov (1927) showed that, when he consistently presented food to a dog after each electric shock to a paw, the original (“unconditioned”) withdrawal responses were almost completely suppressed and substituted by conditioned salivary responses. It now seems that the original withdrawal responses in such experiments are rather complex motor acts, which may involve both a spinal reflex and a conditioned cerebral response, between which it is hard to distinguish without the aid of the electroniyographic technique. Classical conditioned avoidance responses to non-noxious stimuli represent an additional kind of defence reaction. Such conditioned withdrawal occasionally appeared in the present experiments, sometimes in response to an accidental noise which preceded the electric stimulus. These responses differ from the nociceptive reactions in that they do not serve to interrupt, but rather to prevent, noxious stimuli and it seems likely that in this case also the learning is largely based on instrumental conditioning in the sense that the pattern of the conditioned response is refined with the aid of returning sensory information concerning its defensive capacity (notwithstanding that conditioning may occur also without the aid of feed-back from the response; see Beck and Doty 1957; Knapp et al. 1958). We may infer that in an adult human being with great experience of those events which usually precede and succeed a noxious stimulus, there are a t least three kinds
PLASTICITY OF HUMAN WITHDRAWAL REFLEXES
75
of reflexes, which serve to protect him from an offending object which approaches and hits the skin : 1. a conditioned avoidance response (e.g., to a visual stimulus), 2. a nociceptive spinal withdrawal reflex, and 3. a nociceptive cerebral defence reaction. Now the question is: do the signals mediated by the spinal cord impinge upon the motor neurones according to an inborn fixed pattern, and independently of that experience and all those sensory cues which determine the patterns of the cerebral responses? If this is so, it seems inevitable that the spinal and the cerebral reflexes will be often in conflict and will counteract each other. The experiment illustrated in Fig. 6 indicates that such conflicts may occur because the spinal response is not adapted to the position of the limb in a way similar to the cerebral reaction. A conflict also occurred in those training experiments in which the cerebral reaction showed, in contrast to the spinal response, an adaptive sign reversal (Fig. 7). These results merely show, however, that in adults the spinal reflex patterns are less adaptable than the cerebral patterns are. Further experiments are needed to find out whether adaptive spinal reversals may possibly occur during infancy. It is certainly plausible to assume that the spinal patterns are innate, but it is hard to explain their remarkable functional precision in evoking withdrawal movements orientated to the stimulus, without assuming that in some way they are established by the aid of experience, which is either an experience acquired by the individual or is common to all members of the race ( r j Gerard 1961). The fact that the spinal limb reflexes did not show adaptive sign reversal does not imply that spinal transmission occurred in a non-adaptive way, independently of training and experience. On the contrary, the intensity of the reflexes showed funciional adaptability of a kind similar to that described earlier in the study of the abdominal skin reflexes (Hagbarth and Kugelberg 1958). Their habituation, their dependence on the subject’s state of attention and their occasional susceptibility to hypnotic suggestion, indicate a cerebral control of internuncial transmission in the spinal cord. There is anatomical, physiological and clinical evidence for such a supraspinal control of the interneurones in the cord, a control which checks transmission, not only in the spinal polysynaptic reflex arcs, but also in the spinal ascending pathways (see Hernindez-Pe6n 1959; Livingston 1958; Hagbarth 1960 a). It is conceivable that the centrifugal system primarily serves to establish a cerebral control of internuncial reflex transmission and, as Holmqvist et al. (1960) have suggested, the changes reflected in ascending paths may possibly convey feed-back information concerning this control. Functional readjustments which primarily occur at high levels of the neuraxis may be secondarily reflected by such centrifugal systems in subcortical reflex centres, so that harmonious co-operation between functionally allied reflex paths in widely separate levels of the neuraxis is allowed. It is well-known that removal of the control of the higher centres often gives rise to a released activity in the lower centres or, as Denny-Brown (1960) pointed out, “removal of one of the competing factors in the control of movement at a n y level results in overaction of the others” (cj: Konorski 1960). Denny-Brown and Chambers (1958) showed that parietal lobe lesions may enhance tactile and nociceptive avoidance reactions which were delicately orientated to the skin stimuli, while frontal lesions References D . 77-78
76
K.-E. HAGBARTH, B. L. FINER
caused a release of grasp reflexes and an abolition of avoidance. Kugelberg et al. (1960) emphasized that the pathological flexion reflex, including the Babinski response, is due to an imbalance between those flexion and extension reflexes which constitute the normal spinal withdrawal pattern. They concluded that the reversal of the normal plantar response is the result of a change in the normal cerebral governing of the spinal reflex patterns. In the present experiments we did not succeed in reversing the sign of spinal reflexes by physiological means, but in this connection it is of interest to note some reports which indicate that adult subjects, during hypnotic age regression to an infantile level, may exhibit Babinski reflexes (Gidro-Frank and Bowersbuch 1948; Mc Cranie et al. 1955). We have not tried to confirm these results, but we often noticed that the normal plantar response to stroking of the sole was abolished during hypnotically induced analgesia of the foot. Hernindez-Pe6n et al. (1960) recently showed that a forearm skin reflex in man can be modified during hypnotically induced anesthesia and hyperesthesia. They presumed that the reflex was spinal, but did not specify its latency. The results recorded above confirm that spinal skin reflexes are occasionally modified by hypnotic suggestions, but that they are not modified to the same extent as the late avoidance reactions are. It must also be emphasized that variations in muscle tone are apt to occur during the experiments, and such changes must be recognized and avoided in order to exclude reflex variations which merely depend on excitability changes in the motoneurone pool. The neurophysiological mechanisms underlying hypnotic phenomena are largely unknown, but it must be presumed that the perceptual changes which occur in response to hypnotic suggestions are associated with some readjustments in cerebral neural activity, and that the cerebral avoidance reactions, as well as the spinal withdrawal reflexes, can apparently be tuned to accord with the cerebral perceptual change. The dual nature of the withdrawal reactions may possibly explain some of the apparently contradictory results obtained in experiments in which nerve crossing and muscle transplantation is done in order to study central reorganization (see Sperry 1958). The results of the experiments here described suggest that the late component of the nociceptive withdrawal response is readjusted to fit the new situation and that there is in adults no reversal of the initial spinal reflex effects. SUMMARY
Ipsilateral defence reflexes to noxious electric skin stimuli in the human lower limbs were studied by the electromyographic recording technique and it was found that two types of responses are involved which differ both in respect to latency and functional adaptability . The early responses had a latency short enough to indicate spinal reflex arcs (60-80 msec) and their patterns seemed well adapted to induce withdrawal from stimuli applied at various sites of a limb which was supporting the body weight (flexion in all joints proximal to the stimulus - flexion reflex - and extension in the joint just distal to the stimulus, provided that it was applied on the extensor side).
PLASTICITY OF HUMAN WITHDRAWAL REFLEXES
77
These patterns were rigid in the sense that they were largely independent of the initial support and position of the limb; and the reflexes also appeared when they did not serve any protective withdrawal function (k.,in response to stimuli of short fixed duration). The intensity of the reflexes varied, however, with the attentive or apprehensive state of the subject; they often showed response decline during monotonous stimulation (habituation) and they were occasionally reduced by hypnotically induced analgesia and increased by hypnotic hyperalgesia. The late responses, which had a latency of about 120 msec, appeared predominantly in those experimental situations in which they served a protective function (i.e., in response to stimuli with a duration determined by the response). Their motor patterns were not fixed, but varied with the initial position of the limb, so that the result of the motor act was always an interruption of the stimulus. Usually this motor act implied a movement which removed the limb from the stimulating electrodes, but, in experiments in which a counter-movement towards the electrodes was required to end the stimulus, the sign of the late response was (in contrast to the early response) reversed after a few trials to fit the new situation. The amplitude of the late responses was easily influenced by hypnotically induced analgesia or by hyperalgesia. It is concluded that the late components of the ipsilateral withdrawal responses represent highly organized, conditioned, defence reflexes which are probably of cerebral origin, while the early components represent spinal reflexes mediated by interneurones which are under cerebral control.
REFERENCES ANOKHIN, P. K. A new conception of the physiological architecture of conditioned reflex. In J. F. DELAFRESNAYE (Editor), Brain mechanisms and learning. Blackwell, Oxford, 1961, pp. 189-227. BECK,E. C. and DOTY,R. W. Conditioned flexion reflexes acquired during combined catalepsy and de-efferentation. J. romp. physiol. Psychol., 1957, 50: 21 1-216. DENNY-BROWN, D. Motor mechanisms-introduction: the general principles of motor integration. In J. FIELD,H. W. MAGOUN and V. E. HALL(Editors), Handbook of'physiology, See. I, vol. II. American Physiological Society, Washington, 1960, pp. 781-796. DENNY-BROWN, D. and CHAMBERS, R. A. The parietal lobe and behavior. Res. Publ. Ass. nerv. ment. Dis., 1958, 36: 35-117. DODT,E. und KOEHLER, B. Uber das receptive Feld des Beugereflexes beini Menschen. Pfliig. Arch. ges. Physiol., 1950, 252: 362-368. Dow, R. S. and MORUZZI,G. The phvsiology and pathology of the cerebellum. Univ. of Minnesota Press, Minneapolis, 1958, pp. 368-374 GERARD, R. W. The fixation of experience. In J. F. DELAFRESNAYE (Editor), Brain mechanisms and learning. Blackwell, Oxford, 1961, pp. 21-32. GIDRO-FRANK, L. and BOWERSBUCH, M. K. A study of the plantar responses in hypnotic age regression. J . nerv. ment. Dis., 1948, 107: 443. HAGBARTH, K. E. Excitatory and inhibitory skin areas for flexor and extensor motoneurones. Acta physiol. scand., 1952,26, Suppl. 94: 58 pp. HAGBARTH, K. E. Centrifugal mechanisms of sensory control. Ergebn. Biol., 1960 a, 22: 47-66. HAGBARTH, K . E. Spinal withdrawal reflexes in the human lower limbs. J. Neurol. Neurosurg. Psychiat., 1960 b, 23: 222-227. HAGBARTH, K. E. and KUGELBERG, E. Plasticity of the human abdominal skin reflex. Brain, 1958,81: 305-3 18. HERNANDEZ-PEON, R. Centrifugal control of sensory inflow to the brain and sensory perception. Acta neurol. 1at.-amer., 1959, 5 : 279-298.
78
K.-E. HAGBARTH. B. L. FINER
HERNANDEZ-PEON, R., DlTrBORN, J., BORLONE, M. and DAVIDOVICH, A. Modifications of a forearm skin reflex during hypnotically induced anesthesia and hyperesthesia. Acfaneurol. luf.-amer.,1960, 6 : 32-42. HOFFMANN, P., SCHENCK, E. und TONNIES, J. F. Uber den Beugereflex des normalen Menschen. Pflug. Arch. ges. Physiol., 1948, 250: 724-732. HOLMQVIST, B., LUNDBERG, A. and OSCARSSON, 0. Supraspinal inhibitory control of transmission to three ascending spinal pathways influenced by the flexion reflex afferents. Arch. ital. Biol., 1960, 98: 60-80. KNAPP,H. D., TAUB,E. and BERMAN, A. J. Effect of deafferentation on a conditioned avoidance response. Science, 1958, 128: 842-843. KONORSKI, J. The cortical “representation” of unconditioned reflexes. In H. H. JASPER and G. D. SMIRNOV (Editors), The Moscow colloquium. Electroenceph. clin. Neurophysiol., Suppl. 13, 1960, pp. 81-89. KUGELBERG, E. Demonstration of A and C fibre components in the Babinski plantar response and the pathological flexion reflex. Brain, 1948, 71: 304319. KUGELBERG, E. and HAGBARTH, K. E. Spinal mechanism of the abdominal and erector spinae skin reflexes. Brain, 1958,81: 290-304. KUGELBERG, E., EKLUND, K. and GRIMBY, L. An electromyographic study of the nociceptive reflexes of the lower limb. Mechanism of the plantar responses. Brain, 1960, 83: 394-410. LIVINGSTON, R. B. Central control of afferent activity. In H. H. JASPER, L. D. PROCTOR, R. S. KNIGHTON,W. C. NOSHAY and R. T. COSTELLO (Editors). Henry Ford Hospital Symposium on the retirular formation of the brain. Little, Brown and Co., Boston, 1958, p. 177. M c CRANIE, E. J., CRASILNECK, H. B. and TETER, H. R. The electro-encephalogram in hypnotic age regression. Psychiat. Quart., 1955, 29: 85-88. MELZACK, R. and SCOTT,T. H. The effects of early experience on the response to pain. J. comp. physiol. Psychol., 1957, 50: 155-161. PAVLOV, J . P. Conditioned reflexes. Translated by G.V. Anrep. Oxford Univ. Press, London, 1927, pp. 289-290. PEDERSEN, E. Studies on the central pathway of the flexion reflex in man and animal. Acta psychiat. scand., 1954, Suppl. 88: 81 pp. SPERRY, R. W. Physiologicalplasticity and brain circuit theory. In H. F. HARLOW and C. N. WOOLSEY, (Editors). Biological and biochemical bases of’ behavior. Univ. Wisconsin Press, 1958, pp. 401-424. TEASDALL, R. E. and MAGLADERY, J. W. Superficial abdominal reflexes in man. A . M . A . Arch. Neurol. Psychiat., 1959, 81 : 28-36.
DISCUSSION F. BREMER:
The experiments with hypnotic suggestion carried out by our colleague Hagbarth remind me of previous observations made by Titeca on subjects showing hysterical hemianaesthesia and his unpublished observations on volunteers (nurses) in whom unilateral amaurosis or even homonymous hemianopsia was caused by hypnotic suggestion. In all these cases, the stimulus applied to the cutaneous or retinal region which was “anaesthetized”, never caused an EEG arrest reaction, although the same cutaneous or light stimulus was very effective when applied to “normal” regions. The determinism of these phenomena is a challenge to the neurophysiologist! TITECA, J. Unpublished observations. M. A. B. BRAZIER:
Following Professor Brenier’s remarks on the effect of hypnosis on the electrical activity of the brain, may 1 add some observations we have made. Using averaging techniques to emphasize the wave form of the train of oscillations that follow the primary response to flash in man, we have noted
DISCUSSION
79
that it is these late oscillations that are disturbed when hypnotic blindness is induced, and not the primary response. In other words, it is those electrical signs usually associated with the “nonspecific” system (rather than the specific ones) which are involved in the neurophysiological mechanisms brought into play by hypnosis. W. GREYWALTER:
I can corroborate Dr. Brazier’s observations on the comparative susceptibility of the various components of the evoked responses in man to modification by suggestion, attention, hypnosis and so forth. Both intra-cerebral and extra-cerebral recordings have shown the true primary response to random flashes, for example, to be invariant with changes in mental state; on the other hand the late and often protracted after-rhythms, together with the non-specific (anterior) responses, are very easily modified by changes in attention, by hypnosis and by hysterical conditions. However, to demonstrate this differentiation between primary and secondary effects, some sort of averaging technique is essential to provide statistically valid information. May I ask Dr. Hagbarth whether he thinks the primary inhibitions which he demonstrated are specific to cutaneous nociceptive stimuli, or are general responses to any rather startling stimulus, as a preliminary cancelling instruction preparatory to a new set?
In addition to Dr. Walter’s remarks I would like to say that Hoffmann and his co-workers found widespread primary inhibition in antagonistic muscles of the arm when they first described this kind of protective exteroceptive reflex in 1947/1948. Has Dr. Hagbarth also found this in the upper extremity, which, in contrast to the leg, has no supportive function in the human?
P., SCHENCK, E. und TONNIES,J. F. uber den Beugereflex des normalen Menschen. HOFFMANN, Pjiigers Arch. ges. Physiol., 1948, 250: 724-732. D. ALEE-FESSARD: I should like to ask Dr. Hagbarth whether it does not seem possible to him that the effects of suppression he observed on the nociceptive reflex in the course of various behavioral patterns, could be interpreted on the basis of a descending inhibitory action on the spinal relays of nociceptive afferents. This effect resembles that which we have observed a t the intralaminar thalamic level for afferents of the protopathic type (centre median). In this case, to explain both Dr. Hagbarth’s results and ours, it should be accepted that truly inhibitory actions occur at the first spinal relay - a hypothesis already suggested by R. Hernandez-Peon. H. W. MAGOUN: In support of Hagbarth’s generalisation that significant learning can occur at segmental levels in the cord, reference should be made to the recent work of Jennifer Buchwald and Eldred in Los Angeles who have shown that the gamma-efferent discharge to nociceptive stimulation can easily be conditioned to sound stimuli. This conditioning occurs after a very few trials and with stimulus intensities below those evoking alpha-motor neuron discharge. This learned behaviour of gamma motor neurons, which Buchwald and Eldred are analysing with single fibre recording, can be shown to display all the features of classical Pavlovian conditioning - generalisation, differentiation, extinction, etc. Their work will be published shortly; 1 merely wish to insert a reference to it at this point. BUCHWALD,JENNIFER S. and ELDRED,E. Conditioned responses in the gamma efferent system. J . nerv. ment. Dis., 1961, 132 (2): 146-152. BUCHWALD, JENNIFER S., BEATTY,D. and ELDRED,E. Conditioned responses of gamma and alpha motoneurons in the cat trained to conditioned avoidance. Exp. Neurol., 1961, 4 : 91-105.
80
K.-E. HAGBARTH, B. L. FINER
BUCHWALD, JENNIFERS. and ELDRED,E. Relations between gamma efferent discharge and cortical activity. Electroenceph. clin. Neurophysiol., 1961, 13: 243-247.
K.-E. HAGBARTH’s replies To F. Bremer Yes, it certainly seems as if hypnotically induced analgesia or anaesthesia may be accompanied by functional readjustments not only in spinal and supraspinal reflex centres but also in ascending sensory systems mediating cortical arousal responses. Both the suppression of the reflexes and the blocking of the cortical arousal response can possibly be explained by assuming a centrifugal inhibitory influence acting on the primary sensory relays in the cord, But there is no valid reason to suppose that these primary sensory relays are more susceptible to centrifugal control than other succeeding relays in the reflex arcs or in the afferent paths mediating the cortical arousal response. We saw in our experiments how the conditioned avoidance responses of long latency were more easily suppressed by hypnotic suggestion than the initial spinal withdrawal reflexes and I suppose that the number of synapses involved in the mediation of a response may be an important determining factor in its degree of plasticity. To M . A . B. Brazier and W . Grey Walter Dr. Brazier and Dr. Walter both emphasize that, contrary to the non-specific secondary responses, the primary evoked potentials in the cortex cannot be modified by hypnotic suggestion and changes in attention. Other workers who have not used your averaging technique, however, describe effects on the primary evoked potentials as well and I am just wondering which technique is the most reliable in this case. I suppose we agree that a centrifugal influence on ascending spinal paths need not necessarily be reflected, at the cortical level, as a change in amplitude of the primary specific responses. Even if we suppose that the spinal sensory units involved project to the specific rather than the non-specific systems, it cannot be excluded that changes may occur in a specific sensory message which cannot be detected by measuring the amplitude of the primary evoked potential in the cortex. Single unit recordings from sensory relays a t various levels of the ascending paths can probably give us a great deal of additional information about the control of the sensory systems. No, Dr. Walter, the early inhibitions I showed are not generalised, non-specific responses of a preparatory type. We never saw any generalised early inhibitory responses affecting the muscles of the limb in a non-reciprocal way. Some of the slides showed how a noxious stimulus to the foot or the limb caused an early inhibition of the knee-extensor vastus medialis. This inhibition, which has a latency short enough to indicate a spinal reflex arc, is usually accompanied by a simultaneous discharge in the knee-flexors and 1 think it is safe to conclude that these responses represent the classical flexion reflex. If the noxious stimulus is applied, not to the leg but to the ventral aspect of the thigh, it usually brings about the opposite effect: an early discharge in vastus medialis accompanied by an inhibition in the knee-flexors. So, far from being non-specific and generalised, these reflexes are organised and orientated according to a very specific pattern. And this pattern, illustrated in Fig. 2, is such that the reflexes seem well adapted to serve as protective withdrawal responses in a limb supporting the body weight. I was glad you asked the question however, because 1 think it is important to realise that not only specific withdrawal movements but also postural readjustments and non-specific startle reactions may occur in response to noxious skin stimuli. Responses of that kind as well as second order reflexes induced by the initial response are likely to occur in our experiments. Possibly responses of these various types contribute to the rebounds and oscillations which so frequently succeed the primary reflexes. To R. Jung We have not as yet recorded any exteroceptive reflexes in the upper extremities but in the lower extremities we regularly find reciprocal responses in antagonistic one-joint muscles such as soleus and tibialis anticus. The problem as to whether the innervation is reciprocal or not gets much more complex if recordings are made from two-joint muscles or muscles engaged not only in flexion and extension but also in ab- or adduction, rotation etc. Many of these muscles can act as synergists in one movement but as antagonists in another and simultaneous discharges appearing in muscles on opposite sides of an extremity do not necessarily mean that antagonists respond in a non-reciprocal manner.
DISCUSSION
81
To D . Albe-Fessard I believe that the suppression of the spinal withdrawal reflexes is due to a descending inhibitory influence on those spinal interneurons which mediate the reflexes. It is not possible to tell whether in these experiments there is also a simultaneous suppression of ascending signals in the cord but this possibility cannot be excluded and it would certainly be interesting to know whether the suppression of the nociceptive afferent responses which you observe at the thalamic level is due to a reduction of the ascending volley at a much lower level of the neuraxis. It is true that in the present reflex study we regularly used nociceptive skin stimuli but we have no evidence which suggests that only nociceptive relays are susceptible to control. In the study of the abdominal skin reflexes, which as a rule have a lower threshold than the limb reflexes, we sometimes used stimuli which were so weak that they could hardly be classified as nociceptive, and spinal reflexes to such non-noxious stimuli also showed adaptability of the kind described.
To H. W. Magoun Thank you, Dr. Magoun, for mentioning these interesting observations. I look forward to reading the paper when it is published.
Reticular Homeostasis and Critical Reactivity* P. DELL Laboratoire de Neurophysiologie, H6pital Henri-Rousselle, Paris (France)
It is the aim of this paper to attempt to link together three sets of data obtained during the past few years in the field of reticular physiology. These data have shown : first, that the brain stem reticular activating system is controlled by a whole series of mechanisms which adjust continuously and precisely its level of activity; secondly, that this system is responsible not only for multiple facilitatory, but also for inhibitory effects ; and thirdly, that the various groups of interneurons and motoneurons affected by reticular activity display a differential sensitivity to reticular discharges. Ln the course of the description of the data obtained it will become apparent that the adjustment of reticular activity to a given level by way of homeostatic controls is the essential integrative mechanism by which the diverse “final common paths” on the motor side are selectively placed at the disposal of one of the various afferent systems or cortico-spinal projections capable of activating them. In the waking animal a whole pattern of reciprocal associations of non-specific systems and sensorimotor mechanisms of highly specific function determines the fine adjustments of activity and the differentiation of motor responses - in brief, the critical (adaptive) reactivity. Nearly half a century ago Henry Head coined the concept of vigilance, which was meant to indicate at one and the same time a level of activity and the quality of the possible responses. The views presented here, which are based on a much more accurate experimental analysis than was possible in Head’s time, illustrate his farsighted idea. A. RETICULAR HOMEOSTASIS
As a point of departure for this study of reticular homeostasis we have grouped together in Figs. I , 2 and 3 a series of observations on the evolution of various somatic and autonomic activities under the effects of prolonged rnesencephalic reticular stimulation. In all cases the initial variable period of intensification or inhibition of test responses is followed by an attenuation or disappearance of the effects of stimulation, even though reticular stimulation is constantly continued. In Fig. I the cortical arousal activity diminishes after the 5th second of reticular -.
*
The research reported in this document has been sponsored in part by The Air Force Office of Scientific Research OAR through the European Office, Aerospace Research, United States Air Force, under contract No. A F 61 (052)-229.
83
RETICULAR HOMEOSTASIS AND CRITICAL REACTIVITY
stimulation. The upper tracing of Fig. 2 shows the rapid attenuation of the monosynaptic reflex of the masseter muscle. In the center is a recording of the motoneuron discharge of the digastric muscle, produced by direct intracerebral electrical stimulation of the masticator nucleus. This discharge is likewise facilitated by reticular stimulation and decreases progressively in amplitude. The polysynaptic jaw opening
APRES
NOVOCAINISATION DU
,, ,.. .
*
,
.,,.. . . ,.I
’ $TIM.
.,I
BULBE
,
.
RET. MEJ.
I.
.
-
.
,
4V
w
w
w w!Hw+www i*$*v
w#&wwdlF ‘ S T I M . ti MEOIAN 0.8 Y
1 5
Fig. 1 Comparison of the effects on the electrocorticograni of prolonged stimulation of the mesencephalic reticular formation and median nerve, before and after the injection of novocaine into the medial part of the bulb (Bonvallet and Bloch, in preparation).
reflex (bottom of Fig. 2), a test response which is inhibited by stimulation of the so-called “facilitatory” reticular formation, is also gradually attenuated. The same example is shown again at the top of Fig. 3 (R.O.), together with (second line Phr.) a recording of electrically integrated phrenic nerve activity: reticular stimulation increases the amplitude of respiratory discharges but, here again, the facilitatory effect is rapidly diminished. All of these observations were made on “enciphale isole” preparations (section at C1-C2 or section at C3 for phrenic nerve recording). Two other examples, taken from intact animals, are shown in Fig. 3: on the left is an illustration of the changes in cortical frequency (F. cortex), arterial pressure (P.A.) and the monosynaptic masseter reflex (R.M.) during reticular stimulation. After an initial intensification, each of these activities diminishes during reticuiar stimulation with its own characteristic time constant. The bottom of Fig. 3 shows the electrodermogram of the plantar foot Rrferencei P. 102-103
84
P. DELL
2 v
Fig. 2 Progressive reduction of the effects of mesencephalic reticular stimulation on the monosynaptic masseter reflex (top), on the direct response of digastric motoneurons (center), on the polysynaptic reflex (bottom) (Hugelin and Bonvallet 1957a; Hugelin and Dumont 1961). RO
SOL .~ .
___
,.-
. ..
.
ECQ
.
140
.-..
..I
.I'..,.
. -. ,. ,.
.-
sic)
R.M.
h
St. Rot.3V
Fig. 3 Progressive reduction of the effects of mesencephalic reticular stimulation. Lefr, lop to bottom, analysis of cortical frequencies, arterial pressure, amplitude of the monosynaptic masseter reflex (Bonvallet 1962). Upper right, top r5 bottom, jaw opening reflex, integrated discharge of the phrenic nerve, electrocorticogram (Cohen ei ul., in preparation). Lower righr, superimposed recordings of the corticograrn and the psychogalvanic responses to two successive stimulations of the mesencephalic reticular formation (Bonvallet and Bloch, in preparation).
RETICULAR HOMEOSTASIS AND CRITICAL REACTIVITY
85
pad of a cat superimposed on a corticogram: a first brief reticular stimulus (first arrow) produces a marked psychogalvanic response, a second stimulus (second arrow), identical to the first but given a few seconds later, has only a very attenuated effect. Reticular stimulation in non-anesthetized preparations always produces the same effects whether the responses studied are somatic or autonomic. To sum up, in non-anesthetized preparations, repetitive reticular stimulation of constant voltage has always the same effects on somatic activity (monosynaptic and polysynaptic reflexes, direct response of a group of motoneurons, corticographic activity) as well as upon autonomic activity (respiration, arterial pressure, electroderma1 or so-called psychogalvanic responses). After an initial phase of marked facilitation (or inhibition) there follows a second phase during which the response returns more or less rapidly to its original level or stabilizes at a somewhat higher level. The development and time course of this second “recuperation” or “recovery” phase varies with the test used. This depends on two factors. One is the intrinsic organization of the effector system (compare a monosynaptic reflex with the arterial pressure which depends on a slowly reacting sympathetic effector and on humoral factors). The other is the synaptic organization of reticular projections on the different types of effector systems (see below, section 11). Two types of hypotheses can be advanced as explanations of such observations : (a) It is conceivable that the diminution and even the suppression of reticular effects are the result of “fatigue”, “adaptation”, or an “escaping” at one or several points of the complex neuronal circuits which have been activated. In short, the attenuation of the reticular effects may be viewed as a purely passive phenomenon. (b) It is also possible that there are structures and mechanisms which can exert an influence on reticular activity and it becomes then conceivable that the observed effects are a result of active intervention of intra- or extra-reticular control systems. If that is so, then one must localize these structures, study their activation and analyze the dynamic properties of these control circuits. With this aim in mind, we will successively examine the intervention and role of bulbar, cortical, and extra-cerebral mechanisms ; finally the effects of these homeostatic regulatory mechanisms upon the properties of the reticular systems will be emphasized.
I. The role of the medulla in reticular homeostasis The observations reported and discussed in this section deal with bulbar structures which have “inhibitory” effects, some of which have already been studied and analyzed by different authors. The bulbar origin of parasympathetic tonus and its antagonism to sympathetic effects on the circulatory system are well known; Magoun and Rhines (1946) showed that inhibition of somatic motor activity is produced by bulbar stimulation; Moruzzi et al. (cf. Moruzzi 1960) demonstrated an ascending tonic action of bulbar origin which opposes the ascending reticular activating effects. Finally, the depressor effects on the EEG and motor activity which arise from vagal References p. 102-103
86
P. DELL
and baroceptor afferents which have a bulbar reflex center, suggest that the bulbar inhibitory structures might be activated reflexly (Bonvallet et al. 1954; Dell et al. 1954). The interest of the data here reported, which is based on various papers published by Bonvallet and Bloch, lies i n the fact that this bulbar mechanism is activated phasically either by somesthetic afferents or by the mesencephalic reticular formation. Hence, close functional interrelationship between these bulbar inhibitory mechanisms and the activating system have been demonstrated. The effects of these bulbar structures on autonomic responses (arterial pressure, psychogalvanic reflex) and on the corticogram have been studied under identical experimental conditions in these investigations. This makes possible the combination of phenomena which, up till now, have been considered in physiology under widely different headings. 1. When arterial pressure is used as the test in a diencephalic non-anesthetized preparation, repetitive stimulation of a nerve from the paw, which activated nociceptive afferents, will produce an increase in arterial pressure which is maintained at that level for the duration of the stimulation. If the stimulation is gradually increased for one minute to the voltage which was applied suddenly in the previous test, then arterial pressure increases gradually with the stimulation till it reaches the same value at the end as it did during the preceding test. On the other hand, in a pre-bulbar preparation the same stimulations result either in a slight, but rapidly suppressed, increase, or in an immediate reduction, or in an absence of any pressor effect. It must be concluded that the somesthetic afferents activate bulbar structures which can counteract centrally or peripherally the effects of sympathetic discharge (Bonvallet 1962). 2. The evidence is even clearer if one uses as one’s test the electrodermal response. In a diencephalic preparation brief nociceptive stimulation of an afferent nerve causes the classical electrodermal response (psychogalvanic reflex). If now this stimulation is repeated several secondsafter the first one, one obtains a new response which is as large as the first one. The results are completely different in the pre-bulbar animal. The second stimulation produces only a minimal response if it comes 10 sec after the first one. To be fully effective and to produce a response equal in amplitude to the first one, the second stimulus must be applied at least 40 sec later. After an injection of novocaine into the medial part of the bulb two consecutive afferent stimuli will each produce a psychogalvanic response of the same amplitude. These results are shown in the form of graphs in Fig. 4. Consequently, an afferent nociceptive stimulus activates, on the one hand, mesencephalo-hypothalamic sympathetic structures and lateral bulbar structures which produce a sympathetic discharge that can be repeated with the same intensity at short intervals (5 sec for example); and, on the other hand, such a stimulus also activates bulbar structures which suppress or attenuate these discharges and lead to a considerable refractoriness of the system. Similar observations have been made when the electrodermal response was produced by short repetitive stimulation, instead of afferent stimulation, of the mesencephalic or the lateral bulbar part of the reticular formation. In the diencephalic preparation the first stimulus causes a response without any refractory phenomena,
87
RETICULAR HOMEOSTASIS AND CRITICAL REACTIVITY
but in the pre-bulbar preparation a response of the same amplitude can be obtained only if the second stimulus is given 40 sec after the first. In this instance it appears difficult not to conceive of a dual effect of the stimulus: it causes a sympathetic response and it activates structures which can inhibit these sympathetic discharges for about 40 sec. The absence of this inhibition after restricted administration of novocaine into the posterior medial part of the medulla suggests that the central mechanism responsible for the inhibition is situated at that level (Bloch and Bonvallet 1960 a, b).
....
O1
10
20
30
l n t e r v a l l e e n t r e les 2 S t i m u l i
a0 sec
.,
Fig. 4 Recovery curves of the psychogalvanic response to stimulation of an afferent nerve, indicating the inhibitory effect of bulbar structures activated by the first stimulation (Bonvallet and Bloch in preparation). 0, A, without cortex or narcosis; 0, B, without cortex and chloralose; C , without cortex and chloralose, but with novocaine injected into the bulb; A,D, prebulbar without narcosis.
3. In parallel with these descending medullary effects one can also observe ascending effects which are particularly clear at the level of the cerebral cortex. They are activated by the same afferent influx and under the same experimental conditions. Indeed, their existence was suggested by a chance observation during an experiment in which the electrodermogram and electrocorticogram were simultaneously recorded. Brief, repetitive afferent or reticular stimulation caused a psychogalvanic response and short-lasting cortical arousal; the same stimulus applied about ten seconds later caused only very attenuated cortical and electrodermal responses. After a novocaine injection into the medial part of the bulb, however, the second cortical arousal X d r r e n c e r I'
102-103
88
P. DELL
reaction and the psychogalvanic response were identical and even greater than the initial one. This parallelism suggested that activation of the same bulbar structures is responsible for the attenuation of the autonomic and cortical responses. A variety of experiments have confirmed this hypothesis and have demonstrated that any arousing stimulus (reticular stimulation or afferent stimulation) activates, not only the ascending reticular activating system, but also bulbar structures which can attenuate or suppress the reticular arousal effects at the cortical level. Different reasons, given below, indicate that this result is a consequence of reciprocal effects between the bulbar and mesencephalic reticular structures, and that it is legitimate to postulate a functional mesencephalo-bulbo-mesencephalicintra-reticular circuit. The demonstration of the phasic activation of the inhibitory bulbar structures (Fig. 1) is clearest in an “enckphale isolC” preparation treated with a small dose of nembutal (section of the spinal cord at D2 to provide an afferent inflow from the forelimb; 5-8 mg nembutal). Prolonged repetitive stimulation of the median nerve or mesencephalic reticular formation then produces generalized cortical arousal which is gradually attenuated in all the cortical regions to the extent that stimulation is prolonged. On the other hand, after a median section of the cerebral trunk at the bulbo-pontine junction, or after novocaine injection in the posterior median part of the bulb, the same stimulation leads to an arousal which persists at the same intensity for the entire length of the stimulation. Such an arousal effect can be maintained without any attenuation for more than 4 min (Bonvallet and Bloch 1960, 1961; Bloch and Bonvallet 1961). We shall briefly discuss these results and attempt to determine their functional significance. It has been shown that the bulbar inhibitory actions are exerted on sympathetic responses (arterial pressure, the so-called psychogalvanic reflex), on spinal somatic motor activity and on the electrical activity of the cerebral cortex. These effects are activated in a phasic manner by a great variety of afferent stimuli, arising from: 1. somesthetic afferent fibers and, in particular, nociceptive afferents which are most probably situated in tracts already described by Bechterev (1900) and lately restudied by different authors in numerous animal species (bibliography in Mehler et al. 1960), pathways which are now known as the spino-reticular tract, which expands markedly at the level of the bulbar reticular formation; 2. labyrinthine afferents, especially those coming from the semi-circular canals, which are known to carry powerful arousing impulses, the effects of which on ocular motor activity vary greatly according to the presence or destruction of the posterior bulb (Dumont-TyC and Dell 1962); 3. interoceptive afferents of the vagus nerve (Grastyan et al. 1952) and, in particular, those from baroreceptors (Bonvallet et al. 1954; Dell et al. 1954), which are relayed at the level of the nucleus of the solitary tract, the region in which slow repetitive stimulation leads to deactivation of the corticogram and the appearance of spindle-like rhythms (Magnes, Moruzzi and Pompeiano 1961); 4. projections from the mesencephalic reticular formation. The observations of Bloch and Bonvallet reported above strongly suggest the existence of a reticulo-bulbo-
RETICULAR HOMEOSTASIS A N D CRITICAL REACTIVITY
89
reticular negative feed-back circuit, by means of which mesencephalic reticular structures activate a bulbar mechanism which, in turn, attenuates the activating mechanism itself. This suggests that each mesencephalic reticular activation brings into play a bulbar inhibitory mechanism. This mechanism is activated rapidly, either because of intracerebral mesencephalo-bulbo-mesencephaliccircuits, or because the afferent systems themselves have mesencephalic and bulbar projections (somesthetic, vestibular). In addition there is a second and slower, but highly effective, mode of stimulating this inhibitory mechanism which results from mesencephalic reticular activation which has led to an increase in arterial pressure, and thus has intensified sino-carotid baroreceptor discharges. This latter mode of action seems particularly helpful in clarifying the part played by the bulbar inhibitory regions. It is well known that somatic activity or an intense muscular effort produces a circulatory reaction which is associated with an increase in arterial pressure. The bulbar region is then activated by way of the sino-carotid receptors and the motor facilitation is counteracted (centrally and peripherally), so that an adjustment of the motor output to the circulatory capacity is provided. This “load adjustment” (recalling the “Entlastungsreflex” of Hess) affects not only circulatory regulation but also, simultaneously, the level of motor activity. Finally, the question may be raised whether one of the characteristics of this bulbar system is that it functions “cumulatively”. We observe, for instance, that the cortical effects of sino-carotid distention are prolonged for tens of seconds after the end of the distention and that the psychogalvanic reflex remains inhibited for long periods (40-50 sec) in pre-bulbar animals. This suggests that there may be successive additions to the residual inhibition left by preceding stimuli. This would lead to a gradual and prolonged “unloading”. These bulbar mechanisms and especially their tonic effects, have been studied by Moruzzi and his group (discussion in Moruzzi 1960), who used the electrocorticogram as an index, and related it to the state of wakefulness or sleep of their preparations. If one accepts some of the suggestions here made, then sleep would be only an extreme state resulting from an accumulation of bulbar inhibitory effects. If bulbar activity attains a certain level, and if, at the same time, nothing has disturbed the deactivation of the mesencephalic reticular formation which it has produced, then sleep is induced and is self-maintained. I l . The part played by the cerebral cortex in reticular homeostasis
In a series of publications, Hugelin and Bonvallet (1957 a, b, c, 1958) have demonstrated the existence, and have analyzed the properties of a reticulo-cortico-reticular regulatory circuit. Phasic intensification of reticular activity leads to classical cortical “arousal” and this, in turn, leads to a cortico-reticular bombardment from the entire cortical mantle, so that the reticular activity is thus inhibited. We refer the reader to the original articles for descriptions of the various experiments which have made it possible to determine the anatomy of this circuit and which have defined the conditions necessary for its activation. Certain aspects of these observations were recently
90
P. DELL
criticized by Bremer (1960); in this volume will be found a detailed reply to these objections by Hugelin (cj: p. 105). In the context of this study of reticular homeostasis we shall merely point out some of the dynamic properties of this feed-back circuit, properties which identify it as the essential governing mechanism for the adjustment of reticular activity to a given level in mammals with a well-developed cerebral cortex. Briefly, the experiments consisted of the simultaneous recording of electrocortical activity and one other test of reticular activity in “encephale isolt” preparations, e.g., the facilitatory action of the mesencephalic reticular formation on a monosynaptic reflex, or the inhibitory action of this same structure on certain polysynaptic reflexes. As the various examples shown in Figs. 1, 2 and 3 clearly indicate, the facilitatory reticular effect (or the inhibitory one) is rapidly attenuated and suppressed. This phenomenon of “recuperation” coincides with cortical “arousal” and is due to the regulatory action of the reticulo-cortico-reticular loop. From a dynamic point of view, and borrowing the terminology of servo-mechanisms, the essential functional characteristics of this system are as follows : 1. The reticular formation is a controlled system. The cerebral cortex is the feedback controller. Any external disturbance which alters the activity of the controlled system is detected and amplified (cortical arousal) by the feed-back controller. The cortical “arousal” induces an actuating signal which can oppose the variations existing at the level of the controlled system (reticular formation). 2. In such a servo-mechanism the alteration in reticular activity resulting from the disturbance is, usually, not fully corrected. A “system error” persists which is proportional to the disturbance. This error is greatly amplified at the level of the feed-back controller, the cerebral cortex. Consequently, one may speak of the gain of the feedback controller i.e., the cortical gain. 3. It is possible to evaluate this error and the corrective power of the system under study. An example is shown in Fig. 5 taken from an experiment in which the behavior of the reticulo-cortico-reticular loop was analyzed during different stages of induced hypoxia (6.57” 0 2 in Nz) using three test responses: (a) at the cortical level (electrocorticogram), (b) at the reticular level (reticular unit recorded by extra-cellular microelectrode) and (c) by means of a test of reticular facilitation (monosynaptic reflex of the masseter). It is known that such a hypoxia developes in three stages: first, a reticular activation due to carotid glorni discharges stimulated by the reduced oxygen pressure, and associated cortical arousal (Dell and Bonvallet 1954; Hugelin et al. 1959; Dell et al. 1961); secondly, a disappearance of cortical activity after a certain time, which is due to its sensitivity to the lack of oxygen; and thirdly, a post-anoxic phase of intense cortical desynchronization upon restoration of ventilation with normal air. Within the present context, several interesting facts are illustrated by Fig. 5: (a) The frequency of the reticular cell discharges increases only moderately during the onset and intensification of cortical arousal (graphs marked 10 sec and 1 min after onset of hypoxia). The reticular activation is much less than might be supposed from the intense cortical desynchronization. Then, exactly at the moment when the cortical
91
RETICULAR HOMEOSTASIS AND CRITICAL REACTIVITY
tracing changes from desynchronization to slow wave activity, prior to the hypoxic flattening of the record, the reticular cell suddenly begins to discharge at a rapid rate (tracing marked I min 30 sec after the onset of hypoxia). Finally, during the postanoxic phase, there is an even more marked cortical desynchronization, whereas the reticular cell is silent (tracing marked 2 min 30 after onset of hypoxia). FR ~. A
l
l
~
.
k
l
l
A
A
d
d
A
l
~
l
A
r
l 1
i&WW~.’-1L.”YI-;Y*.,..uu..,.,,,o.\
L.
-
Lingual stirnubs
A s s . - ~ ~ , ~ ~ ~ ~ - ~ l ~ ~ n - , ~ : ~ ~ * ~ .
*
I sec A
A
~
~
~
A
.
L
1 . 2 5 . - : : % : ~ + & & ? . W ? ! d . i i ~ ~
..
A
1
r
A !i
’
A ’
-
.
A
~
.
L
-
1A 1 11
A r
A
h 1A 1
145_.__-hC___-------
1Jl 205
1
-
A111114
,
I
~_
~
~
.
++=-rt=ttlH+tH~W~~
L
3
026e’. in N2-
~
A
~
~
‘JJslks
20
5oSec
.w-M\**v
10soctWIHHHWHHl I
A
m
i
n
% ”- -
m --Auq&9hdLbi
i l d l l l
u b d b m q w -
2rnh30-
w
~min~+-+ 0.5 sec
%
v
1w c
Fig. 5 The effects of progressive, acute hypoxia on several tests of reticular activity. Left, comparative effects on the corticogram and monosynaptic masseter reflex. The chemoreceptive origin of the cortical arousal is apparent after 14 min of hypoxia (6% OZ),as well as the appearance of motor facilitation, which coincides with the absence of electrocortical activity (Bonvallet er al. 1961). Right, simultaneous recordings of an “activating” reticular cell and of the corticogram. The frequency of the reticular discharge doubles at the exact moment of the electrocortical disorganization (Hugelin et a!. 1959).
(b) The amplitude of the monosynaptic reflex does not increase during the entire period of cortical arousal, so that facilitation of the monosynaptic reflex is not apparent when reticular activation occurs gradually. The amplitude of this reflex increases suddenly and marked facilitation occurs at the exact moment of the disappearance of cortical arousal and its replacement by a flat record. All the phenomena here briefly described have been analyzed in detail by Hugelin (1959) and Bonvallet and Hugelin (1961). These results provide the data for an approximate measure of the system error and the corrective power of the reticulo-cortico-reticular feed-back. (a) This error is proportional to the difference between the frequency of the reticular cell discharge at rest and the moment of maximal cortical “arousal” which just precedes the disappearance of this “arousal”. (b) The corrective power of the cortex in its capacity as the feed-back controller may be evaluated by comparing the condition of thecontrolled system - the reticular formation - at the moment of maximal control, when cortical “arousal” is very intense, and when the feed-back controller is suddenly destroyed by cortical anoxia, References P. 102-103
92
P. DELL
i.e. at the onset of slow wave activity. The corrective power is proportional to the
difference in the frequency of discharge of the reticular cell at these two moments. (c) The error of correction, as indicated by the slight increase in reticular discharge, is not reflected in the same fashion by the various tests of reticular activity which are available to us. If, for a moment, we look at the cortical activity, not as the feed-back control introduced into the regulatory system, but simply as a test of reticular activity, we can see that at this level the error shows up as a marked increase in electrocortical activity. The reason for this is that the reticular discharges are, in addition, markedly amplified by cortical networks. In contrast to this, there is no such amplification at the reticule-spinal level. The slight increase of reticular discharge does not lead to a facilitation of monosynaptic reflexes. Furthermore, other examples will show that, at the level of the connections between the reticular formation and the interneurons of polysynaptic reflexes, there is a somewhat greater amplification. Hence, partial inhibition of these reflexes can develop under conditions of mild reticular activation which remain incapable of facilitating a simultaneously recorded monosynaptic reflex. 4 . Another basic feature of such a system is an inherent inertia. The corrections are not made instantaneously at the level of the corrected system, the reticular formation. The delay is obviously equal to the sum of the delays intervening at each level of the two arms of the reticulo-cortico-reticular loop : the reticulo-cortical time and the time for cortical neuron recruitment on the one hand, and the cortico-reticular transmission time on the other hand. These delays have been measured (see, for example, Figs. 1 and 2 in Hugelin and Bonvallet 1957b). Under optimal conditions the feed-back effect on the reticular formation can begin 100 msec after the onset of the reticular stimulation and it increases exponentially for the next 200 to 300 msec. Thus, the reticular disturbance can be corrected after 400 msec, and only the error defined above persists. This delay depends upon the condition of the preparation and, in particular, upon the rapidity with which the cortical controlling system builds up an intense cortical “arousal”. Under optimal conditions the delay is of the order of 300 to 400 msec; in other instances it can last several seconds. The recovery phase is terminated when reticular activity has stabilized at a new level. The delay thus determined was obtained under artificial conditions because, in these experiments, the reticular disturbance was produced by electrical stimulation of the reticular formation. Under natural conditions one must take into consideration the fact that a certain period is required to produce the reticular disturbance itself, e.g., the time needed to recruit reticular neurons by extra-lemniscal collaterals or spino-reticular projections (this additional delay is of the order of 30 msec). In summary, these are the essential dynamic properties of the reticulo-corticoreticular loop. They make possible fine adjustments of reticular activity, and provide a supplementary non-specific central action for numerous extra-cerebral factors which, instead of affecting reticular activity directly, can act on the reticular monitor, the cerebral cortex.
RETICULAR HOMEOSTASIS AND CRITICAL REACTIVITY
93
111. The role of extra-cerebralfeed-back circuits in reticular homeostasis
We have shown so far how the dynamic interrelations between the reticular formation of the mesencephalon, the cerebral cortex and bulbar structures regulate any disturbance of reticular activity by controlling this activity and adjusting it to the vicinity of its initial level. It now remains necessary to replace the reticular system and its intra-cerebral control systems in their normal context, i.e., within the organism, and to study the effects of natural disturbances, i.e., disturbances due to changes of the internal or external “milieu”. Consequently there arise three series of new problems which are concerned with reticular homeostasis : 1. Disturbances due to sensory excitation or to a shift in one component of the internal “milieu” require a certain amount of time. A physically defined stimulus can, however, have very different reticular effects, because first, its reception by the sense organs may be modified, secondly, because the sensory message may be amplified or inhibited at the different relays of the sensory path and finally, because its entry and transmission at the reticular level may be altered, so that its effectiveness as a reticular activator is modified. 2. The same disturbance, in addition to its reticular effect, may have an effect at the level of one of the two systems so far examined which control reticular activity and thus increase or reduce the control gain and modify the dynamics of the intra-cerebral feed-back systems. 3. Finally, it is well known that any variation of reticular activity engenders autonomic changes, respiratory, circulatory, humoral and others; in turn, these can secondarily act upon the reticular formation or its control systems. The use of preparations which have been stabilized by anesthetics or sections of the neuraxis eliminate these effects during experimentation. It is, however, obvious that an animal’s behavior under natural conditions consists of hunting, fighting, and aggressive acts of all sorts. The reticular centers control the autonomic adjustments necessary to the effective execution of somatic activities and the resulting humoral changes react, in return at the central level, in the manner described in the preceding paragraph. We cannot fully deal with these three aspects of reticular control by extra-cerebral circuits and we shall simply give some recently studied examples which illustrate each of the above listed categories. (a) The reticular formation modifies sensory reception. Local orienting reflexes (eye movements, movements of the pinnae), pupillary diameter, mechanical sound transmission in the middle ear depend intimately on reticular activity. Changes in peripheral receptor sensitivity and in the transmission capacity of the first sensory relays have been attributed to reticular influences. Consequently, the sensory message produced by a well defined physical stimulus and arriving at the brain stem over classical sensory pathways may be highly variable, depending upon the circumstances. This aspect of the problem cannot be dealt with here. (b) More important, for reticular homeostasis, is the fact that a given message does not always have the same activating power over reticular activity but that this depends on the functional state of the reticulo-cortico-reticular system at any given moment. References n 102-103
94
P. DELL
Hugelin and Bonvallet (1958) have shown that the same standard peripheral stimulus led to a marked momentary facilitation of a monosynaptic reflex when the cortex showed spindle activity or slow waves, but that it produced only an insignificant facilitation when cortical activity showed wakefulness and desynchronization. Therefore, the capacity of a message to produce reticular activation is under the control of the cortical activity at the given moment. This mechanism also explains why a stimulus which is quite capable of evoking cortical “arousal” and motor facilitation in a resting animal is incapable of producing motor facilitation if it follows a short time after a prior stimulus has affected another sensory field and disturbed the reticulocortico-reticular equilibrium. (c) The momentary condition of the circulatory and respiratory constants plays a major part in the adjustment of the level of reticular activity. Years ago, we gave a number of examples by showing the opposite effects, direct and reflex, of arterial pressure, circulating adrenaline, and sino-carotid baroreceptor discharges on the reticular formation (Bonvallet et al. 1954; Dell el al. 1954). It wasshown that thestudy of the respiratory constants ($02 and p 0 2 ) were even more interesting, because it was possible to consider, not only their direct and reflex effects at reticular levels, but also their effects on the amplification capacity of the cortical feed-back controller. We have already analyzed one aspect of the dynamism of the reticulo-corticoreticular circuit by means of variations in p0z during hypoxia. The differential effects of hypoxia at cortical and reticular levels demonstrated, above all, the presence and power of the cortical “braking” effects. This was indicated by the spectacular increase in reticular discharge at the moment of suppression of this “braking” effect, as was shown by the disappearance of cortical electrical activity. Even more striking results were obtained by analyzing the effects of variations of pC02 on the equilibrium of the reticulo-cortico-reticular circuit (Gautier 1961). By slowly varying the p C 0 z all the intermediary phases of cortico-reticular interrelations can be demonstrated. Fig. 6 shows that for pC0z slightly below normal (light hypocapnia) the cortical record is not very active and the monosynaptic reflex, taken as test of facilitatory reticular bombardment, is very marked. If p C 0 z is increased by making the “encCphale isolC” preparation inhale a mixture of gradually increasing COz, one can note a progressive activation of the cortical record and a reduction in amplitude, and then a complete disappearance of the monosynaptic reflex. It goes without saying that the same experiment has been repeated on decorticated animals and that the effects of hypo- and hypercapnia have also been studied on a monosynaptic reflex after its disconnection from the reticular formation. These results show that, because of the greater sensitivity of the cerebral cortex to the effects of COZ,the cortical feed-back controller rapidly passes into a state of intense arousal. It greatly amplifies the small surplus of activity produced at reticular levels by the direct and reflex effects (chemoreceptors) of COz. Because of the direct action of COZ on the cortex, the intense cortical activation leads to reticular inhibitory effects which are indicated by the total disappearance of the reticular facilitation of the monosynaptic reflex. Such observations made on simplified preparations and by artificially varying certain autonomic constants cannot be considered to be mere artefacts or
95
RETlCULAR HOMEOSTASIS AND CRITICAL REACTIVITY
experimental curiosities. Indeed, if one examines the tables and curves of variation of circulatory and respiratory constants during muscular effort, one realizes that all somatic activity is associated with large fluctuations beyond the normal values of arterial pressure, pCO2, p02, etc. Furthermore, some of these changes cannot be attributed to peripheral factors; some of them stem from reticular effects on the
d
s
.
&
w c--(
1 SQC
Fig. 6 Comparative effects of the changes in alveolar pressure of carbon dioxide on the corticogram and the monosynaptic masseter reflex (the upper tracing of each record shows the deflections of the amplitude analyzer recorded on the polygraph). Observe that cortical deactivation goes hand in hand with an increase of the masseter reflex (release of facilitatory reticular influences), whereas corticographic arousal corresponds to its depression (inhibition of the facilitatory tonus of reticular origin) (Gautier 1961).
central regulatory mechanisms of these autonomic activities, such as the bulbar respiratory center (cf. Cohen and Hugelin 1961). In the normal, active animal all the regulatory links which are here artificially isolated, constitute a series of positive and negative feed-back circuits. This has been shown diagrammatically in Fig. 7 (see legend for descriptive details), and it is these circuits on which the level of reticular activity depends in the end.
96
Y. DELL
u (t)
CI
Fig. I Diagram showing certain functional relationships between the activating reticular formation and the structures sensitive to changes in blood composition. An afferent stimulation excites siniultaneously the cerebral cortex and the respiratory centers. This leads to aloweringof p C 0 z andincreased pOe By way of a chemo-reflex, the lowering of p C 0 z acts on the activating system and on the respiratory centers, the activity of which diminishes and, by a humoral path, it acts on the cerebral cortex. The excitability of the latter is reduced and the consequent absence of the cortical braking action increases the activity of the reticular system.
( d ) Finally, these circuits are not functionally equivalent. They possess highly variable time constants. Some regulating mechanisms permit relatively rapid phasic adjustments, others include humoral intermediaries and lead to rather slow variations. The rapid phasic adjustments are superimposed at any one moment on the more slowly developing changes. There exists, however, a fundamental difference between the neural and humoral factors and this suggests, necessarily, a hierarchical classification in terms of the contingent or unavoidable aspects of certain actions. Thedirecteffects of the humoral factors at the level of the reticular formation and its cortical control system are unavoidable and, hence, prepotent. Most of the effects transmitted neurally (of interoceptive or exteroceptive origin) pass, on the other hand, through relays which have filters and gain controls ;they can be preferentially suppressed or intensified. These various facts briefly reviewed here show that many neural and humoral factors, which have been regarded up till now as activating or deactivating agents of the mesencephalic reticular formation, must be regarded as links in complex extracerebral chains which emanate from, and lead to, reticular structures and are designed for the homeostatic control of this activity. IV. The properties conferred on the reticular systems by homeostatic mechanisms
In terminating this chapter we would like to show how the observations stated above lead, on the one hand, to modifications of the original and traditional image of the
RETICULAR HOMEOSTASIS AND CRITICAL REACTIVITY
97
reticular formation and, on the other hand, to the concept of a reticular system, which is endowed with homeostatic properties, and represents the fundamental regulator of somatic and autonomic activities. (a) The reticular activating system has been described as the site of convergence of heterogeneous somatic and autonomic afferent activity and to it has been attributed the part of integrator of the different spheres of activity of the organism. This view can be maintained, provided that one takes into account the fact that numerous central actions impinge on the reticular system and its control mechanisms at the same time, or that they may act on the cortical and bulbar control mechanisms and only indirectly on reticular activity. (6) There is, apart from artificial, experimental conditions, no reticular stimulus which can be physically defined and repeated. The reticular activation produced by a natural stimulus depends, first, on a regulation (by the reticular formation or its control mechanisms) of the incoming pathways from the periphery to their reticular point of entry. Secondly it depends on the rapidly acting reticular controls activated by the stimulus. The efficacy of this regulation reflects to a great extent the state of the organism at that given moment. The appearance, intensity, and duration of the facilitatory reticular discharges (ascending and descending) are the necessary conditions for the initiation of most sensory and motor activities. These conditions depend upon the selection of information which can produce a reticular disturbance the intensity of which is determined by the state of the control systems. (c) The intra- and extra-cerebral control mechanisms, are the guarantors of the ,functional stability of the reticular systems. Several authors have already emphasized the part played by local inhibitory circuits in the stabilization of motoneurons or sensory relays (cf. for example, Brooks 1959) and in this volume (cf. p. 23) Granit discusses the Renshaw cells from a similar point of view. Masses of neurons sensitive to humoral factors and subjected to bombardment by many afferent systems (which themselves may frequently produce continuous tonic discharges) could react only in a disorganized manner. Reticular stabilization is possible only because these disturbances are eliminated at their origin or, if they have already taken place, because of a rapid rectification of their effects. ( d ) The data at present available to us are purely qualitative and it is therefore difficult to determine whether reticular activity adopts different levels by constantly and gradually passing from one to the other, or whether there are preferential “set points” at which it becomes momentarily stabilized and fixed. ( e ) The major concept which, in our opinion, is to be derived from these observations is that of a power reserve of reticular activity. Most of the regulatory actions which we have described are “braking” actions. Released, left to themselves, and subjected to the same humoral conditions, the reticular neurons display a discharge rhythm which is much greater than the one they are forced to adopt under the stabilizing intra- and extra-cerebral influences. A first consequence was pointed out by Granit, when he wrote: “The purpose of spontaneous activity is to keep the neuron appropriately biased and responsive to signals of either sign”. It is, moreover, well known that a system can reach a higher level of activity much more rapidly and References P. 102-103
P. DELL
98
easily if it is not necessary to drive that system to the new level, but merely to release a brake. B. CRITICAL
REACTIVITY*
Wakefulness is characterized, not only by more or less intense sensory and motor activity, but also by the fact that the organism, exposed to a multitude of different stimuli, selects and responds in an adaptive manner to one of them. in this second chapter it thus becomes necessary to analyze how variations in the level of reticular activity condition the preferential dispatch of one afferent excitation among many others to a given group of motoneurons, this “final common path” being at the same time cleared of all other afferent impulses which might innervate it. Two mechanisms are involved : ( a ) The reticular formation is capable of selectively inhibiting certain groups of interneurons, and thus, while it blocks a limited part of the circuit, it leaves the circuit beyond that point available for response to impulses arriving over other pathways. (b) There is a differential sensitivity of the various relays of polysynaptic or corticospinal reflex pathways to reticular influences. This is due to differences in amplification of reticular effects at the level of the various relays. These two propositions will be illustrated by examples taken from two recent papers published by Hugelin and Dumont (1961) and Hugelin (1961). 1 . The differential action of reticular discharges The polysynaptic jaw opening reflex after stimulation of the lingual nerve and recorded from the nerve of the digastric muscle is depressed or suppressed by mesencephalic reticular stimulation. To analyze the mechanism of this inhibition, the first known relay of this reflex at the level of the substantia gelatinosa of Rolando, was stimulated by an intracerebral bipolar electrode (Fig. 8). In response to a single electric shock one can record from the digastric nerve a series of events. The first corresponds to a motoneuron discharge due to direct stimulation of the motoneuron pool by current spread; the second represents a motoneuron discharge due to stimulation by diffusion of interneuron group 12, adjacent to the motor nucleus. The third and the lesser, subsequent events represent motoneuron discharges due to the direct stimulation of the interneuron group 11 of the substance of Rolando (first relay of this reflex arc) which immediately surrounds the stimulating electrode. At the onset of reticular stimulation (second right-hand picture of Fig. 8) the motoneuron discharge ( I st event) is facilitated, the motoneuron response to the discharge of IZ (second event) is slightly inhibited and the motoneuron response to the discharge of 11 (subsequent deflections) is completely inhibited. Consequently, as ..
*
~.
Critical reactivity is an expression introduced by Pieron. It qualifies the capacity to respond with “judgment” to external stimulations. The word “critical” is used in its etymological sense ( X F L Y E L V , to judge), which it has preserved in French, German and Italian to qualify that part of Logic which deals with judgment. In this context, the title expresses the idea of motor responses adapted to the prevailing conditions.
RETlCULAR HOMEOSTASlS AND CRITICAL REACTIVITY
99
a result of relatively intense reticular discharge, each of the three principal relays of this reflex arc responds differently to the reticular stimulus. After several seconds of reticular stimulation (right-hand image 3 of Fig. 8) the cortical correcting circuit for reticular activity reduces its discharge. The reticular action at the level of these three relays is now altered: direct motoneuron discharge is no longer facilitated, whereas the discharge corresponding to activity in interneuron group I2 is no longer inhibited. Only the discharge due to activation of the first relay 11 remains inhibited, though not completely.
Fig. 8 The differential action of mesencephalic reticular stimulation on the three relays of the polysynaptic digastricjaw opening reflex. Explanationin the text (Dumont et al. 1961).
From these different observations one may conclude that the complete disappearance of a natural reflex response following reticular stimulation, is due to a selective inhibition of the interneurons of the first relay in the substance of Rolando. A partial disappearance, i.e., a reflex response of reduced amplitude, must be attributed to two different mechanisms, depending upon the circumstances: (a) At the onset of the reduction of the intense reticular stimulation reflex amplitude is the result of a direct reticular facilitation of the motoneurons; because of this facilitation the motoneurons can still be activated by reflexogenic volleys which are already subliminal due to the inhibition of group 11. (b) During the recovery phase, induced by the cortical feedback control, the reflex response reappears after complete suppression, because the reticular discharges are no longer sufficiently intense to completely block the interneurons 11 of the first relay. References P. 102-103
100
P. DELL
2. Switching in the agonist-antagonist system If one provokes at the same time a monosynaptic reflex of the masseter nerve (closing of the jaws) and a polysynaptic reflex of the digastric nerve (opening of the jaws) and if, during a given period, the two shocks succeed each other so that the digastric response (response equivalent to a flexor reflex activated by group I1 and 111 afferents of the lingual nerve) precedes the masseter response, then one observes the classical phenomenon of reciprocal inhibition in a resting “enciphale isolC” preparation: the monosynaptic masseter response (equivalent of an extensor reflex) is entirely inhibited. Mesencephalic reticular stimulation has different effects in this agonist-antagonist system, according to the intensity of the stimulation : (a) Relatively weak reticular stimulation, which cannot directly facilitate the masseter motoneurons, will nevertheless result in the appearance of the masseter response, because the interneurons I1 of the reflex arc of the digastric antagonist are inhibited. As a result the digastric response is suppressed, the antagonistic masseter reflex is released from reciprocal inhibition and the masseter response attains its normal amplitude. (b) Intense reticular stimulation produces the suppression of reciprocal inhibition described above and, at the same time, a direct facilitation of the masseter motoneurons. The amplitude of the masseter response is, therefore, much greater than normal. (c) If intense reticular stimulation is prolonged, the cortical control circuit is activated and the reticular discharge is gradually regulated. There is a decrease in the amplitude of the masseter response as compared to the level reached at the beginning of intense reticular stimulation (preceding case); it returns to a normal and then even a sub-normal amplitude. In this example the various motor responses result from a differential sensitivity to the effects of reticular stimulation of the interneurons of the first relay of the flexor pathway as compared to effects upon the extensor motoneurons. 3. Shift in the ,functionally active connections to the final common path (a) The afferent groups 11 and I l l of the lingual nerve used in the preceding example are not the only means by which opening of the jaw can be produced. Tactile stimulation of the posterior part of the tongue has the same effect and the afferent pathway is then made up of fibers of larger caliber. This type of reflex can be elicited by weak repetitive electrical stimulation of the chorda tympani (cf. top line of Fig. 9). Mesencephalic reticular stimulation does not inhibit but, on the contrary, facilitates, this jaw opening reflex. In this instance the increased amplitude of the reflex is due t9 a direct facilitatory reticular action on the digastric motoneurons (flexors) without inhibition of upstream located interneurons. Thus, the reticular effect differs, according to whether the same final common path (the digastric flexor motoneurons) is activated by one afferent pathway (instance 5 1) or by another (the present instance). The interneuronsof these tworeflex pathways are differentially sensitive to reticular influences. (6) Alternate activation of the final common path of the digastric motoneurons,
101
RETICULAR HOMEOSTASIS AND CRITICAL REACTIVITY
by a volley passing through the classic segmental reflex arc analyzed in 9 1, and through the cortico-motor path descending from the cortical masticatory area, shows that reticular stimulation depresses or suppresses the segmental response and markedly facilitates the cortico-motor response. This phenomenon was analyzed in detail by progressively increasing the intensity of reticular stimulation (cJ bottom line of Fig. 9). It could be shown that the progressive depression of the segmental response in favor of the cortical response is due to the relatively different sensitivities to reticular
Chocs uniques
3.4 v
34v Chocs r i p & &
52c
CORDE D U TYMPAN
uLUJ
.............................. -dilmg
L
L
L
U
..............................
-h
STlH. RET. VOLTAGE PROGRESSIF 6 V 3 sec
Fig. 9 The commutator-like action of reticular discharge on the digastric final common path. Top, on the left, stimulation of pain fibers in the chorda tympani provokes the digastric polysynaptic reflex, which is inhibited by an arousing stimulation. On the right, during the same experiment, repetitive stimulation of the fibers in the chorda tympani at lower threshold produces, in the drowsy animal, a small response which is facilitated at the moment of awakening. Bottom, reticular stimulation with maximum voltage given immediately or gradually, inhibits the nociceptive digastric reflex (stimulation of lingual nerve at the dots) and facilitates the cortico-motor response of the same motoneurons (stimulation of masticator area of cortex at open circles). , lingual stimulation; 0,cortical stimulation. (Hugelin, 1961).
discharges of the three major components of that response : the digastric motoneurons themselves, the interneurons of the segmental path, the interneurons of the corticomotor path. An outline of the neuronal organization which is at the basis of these changes in active connections of the final common path can be found in the discussion by Hugelin (cf. this volume, p. 105). This series of examples shows that when several reflex systems of different functional significance converge upon a final common path together with cortico-motor pathReferences A ! 102-103
I02
P. DELL
ways, the predominance of one of these afferent systems depends essentially on the momentary level of reticular activity. The change in the active connections of the final common paths is a phenomenon closely linked to variations in vigilance. In a sleeping or resting animal a painful stimulus produces a withdrawal movement the local protective function of which is obvious. This segmental and locally useful reflex is also seen in the waking animal, but in this it is no longer predominant. Reticular inhibition of the polysynaptic flexion reflex releases the antigravity muscles from reciprocal inhibition ; this creates the necessary conditions for more general activity, reactions of support, posture, and quadruped walking. Finally, when an incident sensory stimulation accentuates the waking state, there is facilitation of the corticomotor discharges and the so-called “spontaneous” movements. The final common path, cleared by inhibition of the interneurons of the segmental polysynaptic paths, is now entirely available to the pyramidal tract: diversified and adaptive motor behavior now becomes possible. It appears that, by the interplay of these various mechanisms, a system, which has been appropriately named non-specific because of its heterogeneous afferent pathways and effects on many activities, can lead to the differentiation of a whole range of motor functions. At the same time it can create the conditions which are necessary if higher nervous commands are to dispose freely of the motor effectors. These are the very conditions of any critical reactivity.
REFERENCES BECHTEREW, W. VON. Les voies de conduction du cerveau et de la moelle. Lyon, 1900. Quoted from BECHTEREW, W. VON. Die Funktionen der Nervencentra, G . Fischer, Jena, 1909-1911, vol. 3. BLOCH,V. et BONVALLET, M. Le declenchement des reponses electrodermales a partir du systeme reticulaire facilitateur. J. Physiol. (Puris), 1960a, 52: 25-26. M. Le contrde inhibiteur bulbaire des reponses electrodermales. C.R. BLOCH,V. et BONVALLET, SOC.Biol. (Paris), 1960b, 154: 4 2 4 5 . BLOCH,V. et BONVALLET, M . Interactions des formations reticulaires mesencephalique et bulbaire. J. Physiol. (Paris), 1961,53: 280-281. M. 1962, in preparation. BONVALLET, BONVALLET, M. et BLOCH,V. Le contrBle bulbaire des activations corticales et sa mise en jeu. C.R. Soc. Biol. (Paris), 1960, 154: 1428-1431. BONVALLET, M., and BLOCH,V. Bulbar control of cortical arousal. Science, 1961,133: 1133-1 134. BONVALLET, M., and BLOCH,V. 1962, in preparation. A. Influence de la formation reticulaire et du cortex cerebral sur BONVALLET, M. et HUCELIN, l’excitabilite motrice au cours de I’hypoxie. Electroenceph. d i n . Neurophysiol., 1961, 13: 270-284. BONVALLET, M., DELL,P. et HIEBEL, G . Tonus sympathique et activite electrique corticale. Electroenceph. clin. Neurophysiol., 1954, 6: 119-144. BREMER, F. Les regulations nerveuses de I’activite corticale. Arch. Suisses Neurol. Neurochir. Psychiat., 1960,86: 3448. BROOKS, V. B. Contrast and stability in the nervous system. Trans. N . Y. Acad. Sci., 1959, 21: 387-394. COHEN,M . I., et HUGELIN, A. Excitation reticulaire et activite du nerf phrenique. J . Physiol. (Pari.c), 1961, 53: 303-304. DELL, P., e t BONVALLET, M. Contrble direct et reflexe de I’activite du systtrne reticule activateur ascendant du tronc cerebral par l’oxygtne et le gaz carbonique du sang. C.R. SOC.Biol. (Paris), 1954, 148: 855-858. DELL,P., BONVALLET, M. et HUGELIN, A. Tonus sympathique, adrenaline et contrBle reticulaire de la motricite spinale. Electroenceph. d i n . Neurophysiol., 1954, 6 : 599-6 18.
RETICULAR HOMEOSTASIS AND CRITICAL REACTIVITY
I03
DELL,P., HUGELIN, A., and BONVALLET, M. Effects of hypoxia on the reticular and cortical diffuse (Editors), Cerebral anoxia and the Elecfroencephalograr, systems. In J. S. MEYERand H. GASTAUT C.C. Thomas, Springfield, Ill., 617 pp., 1961: 46-58. DUMONT-TYC, S. et DELL,P. Composantes facilitatrices et inhibitrices du reflexe vestibulooculaire. J. Physiol. (Paris), 1962, 5 4 : 331-332. GAUTIER,H. Technique de prelevement et de mesure continue de la concentration fractionnaire du gaz carbonique alvtolaire chez le chat ventile artificiellement. T h h e mid., Paris, 1961, 60 pp. E., HASZNOS, T., LISSAK,K., MOLNAR, L., and RIZSONYI,Z. Activation of the brain stem GRASTYAN, activating system by vegetative afferents. Actaphysiol. Acad. Sci. hung., 1952,3: 103-1 22. HUGELIN,A. Integration motrice et vigilance chez I’encephale isole. 11. ContrBle reticulaire des voies finales communes d’ouverture et de fermeture dela gueule. Arch. ital. Biol., 1961,99: 244-269. HUGELIN,A. et BONVALLET, M. ContrBle telencephalique de l’excitabilite des motoneurones alpha lors de I’excitation rtticulaire en l’absence d’anesthesie. J. Physiol. (Paris), 1957, 49: 212-214. HUGELIN, A. et BONVALLET, M. Tonus cortical et contrBle de la facilitation motrice d’origine reticulaire J . Physiol. (Paris), 1957a, 49: 1171-1200. HUCELIN,A. et BONVALLET, M. h d e experimentale des interrelations reticdo-corticales. Proposition d’une theorie de I’asservissement reticulaire a un systime diffus cortical. J. Physiol. (Paris), 1957b, 49: 1201-1223. M. Analyse des post-dkcharges rkticulaires et corticales engendrees HUGELIN,A. et BONVALLET, par des stimulations electriques reticulaires. J. Physiol. (Paris), 1957c, 49: 1225-1234. HUCELIN,A. et BONVALLET, M. Effets moteurs et corticaux d’origine reticulaire au cours des stimulations somesthbiques. RBle des interactions cortico-reticulaires dans le determinisme du reveil. J. Physiol. (Paris), 1958, 50: 951-977. HUCELIN,A., et DUMONT, S. Integration motrice et vigilance chez l’encephale isole. I. Inhibition reticulaire du reflexe d’ouverture de la gueule. Arch. ifal. Biol., 1961,99: 219-243. HUGELIN, A., BONVALLET, M. et DELL,P. Activation reticulaire et corticale d’origine chemoceptive au cours de I’hypoxie. Electroenceph. clirr. Neurophysiol., 1959, I1 : 325-340. MAGNES,J., MORUZZI,G., and POMPEIANO, 0. Synchronization of the EEG produced by lowfrequency electrical stimulation of the region of the solitary tract. Arch. ifal.Biol., 1961,99: 33-67. MAGOUN,H. W., and RHINES,R. An inhibitory mechanism in the bulbar reticular formation. J. Neurophysiol., 1946, 9 : 165-171. MEHLER,W. H., FEFERMAN, E., and NAUTA,W. J. H. Ascending axon degeneration following anterolateral cordotomy. Brain, 1960, 83: 718-750. MORUZZI,G. Synchronizing influences of the brain stem and the inhibitory mechanisms underlying the production of sleep by sensory stimulation. Electroenceph. elin. Neurophysiol., 1960, SUPPI,13:232-252.
DISCUSSION W. R. ADEY: Dr. Dell’s very elegant paper has indicated the existence of certain corticofugal influences on the reticular formation, and that these influences exert predominantly inhibitory effects on motor mechanisms. We have studied the role of diencephalic areas in the control of more caudal structures in midbrain and pontine reticular formation. In particular, we have focussed attention on the subthalamic region, which, from our studies, appears to exert a predominantly excitatory influence on the midbrain reticular substance. It was to this region that Morison (In Brain Mechanisms and Consciousness, Blackwell, Oxford, 1954, Ref. p. 15) directed attention as possessing a capacity for arousal of the deeply anaesthetized animal, and he suggested that influences arising in it or passing through it might produce this effect independently of the more dorsally situated intralaminar thalamic system. Our initial studies (Adey and Lindsley 1959) indicated that relatively large lesions in this region in cats and in monkeys were associated with major behavioural changes, including adoption of unusual postures, diminished movement and staring behaviour, and loss of spontaneous feeding, although food placed in the mouth might be eaten ravenously. There was gross interference with a previously learned conditioned avoidance response, with a slow and incomplete post-operative
104
P. DELL
recovery of this response. Exploration of the rostra1 midbrain reticular formation following both acute subthalamic lesions indicated profound changes in responsiveness to peripheral somatic stimuli. With chronic subthalamic lesions, there appeared to be a great reduction in the number of spontaneously firing neurons and a reduced responsiveness to sciatic nerve stimulation. Evoked potentials in the same midbrain zones to sciatic stimulation were also much reduced in acute experiments, but could be restored to their original level by a brief tetanisation at the recording site. This restoration of local excitability by tetanisation decayed over 10 to 15 minutes with the typical curve of a posttetanic potentiation. A variety of other diencephalic lesions produced no comparable changes in midbrain excitability. In later experiments, we have implanted a series of cats with chronic recording and coagulating electrodes, in order to study in a more sophisticated way the effects of subthalamic lesions on both previously learned discriminative performance and simultaneous EEG records from the hippocampal system, the midbrain reticular formation and primary sensory cortical areas (Adey et a/. 1961; Lindsley and Adey 1961). With smaller subthalamic lesions than in the initial experiments, unilateral damage was followed by a gross defect in discriminative performances in both maze and delayed response tests, when the reward was placed in the contralateral visual field. Recovery occurred progressively over 7 to 10 days. A second lesion in the opposite subthalamus was again followed by a temporary inability to discriminate objects in the opposite half of the environment, even when the initial placement was carefully observed by the cat in the delayed response test. Simultaneous EEG records showed a major reduction in the regular trains of 5-6 cycles per second slow waves in the hippocampal system normally accompanying the discriminative performance. This was obvious following a unilateral lesion and was profound after a bilateral lesion. Similar trains of slow waves usually appearing in the midbrain reticular formation and sensorimotor cortex concurrently with the hippocampal trains were abolished. With recovery of the learned performance in the two weeks following the lesions, there was a progressive return of most aspects of the initial slow wave patterns in both hippocampal and extra-hippocampal structures. Similar lesions in more dorsal thalaniic structures were not accompanied by such behavioural defects. Similar sensory defects confined to the opposite half of the environment have recently been described in the cat by Sprague et a/. (1960), and have been attributed by them to interference with specific afferent pathways, including the leniniscal systems. These pathways were not interrupted in our experiments. Our experiments suggest that these defects may arise from an interference with normal interrelations between the hippocampal system and subcortical structures, and particularly with those midbrain areas which our studies first showed to receive descending fluxes from the hippocampal system (Adey et al. 1956). ADEY,W. R . and LINDSLEY, D. F. On the role of subthalamic areas in the maintenance of brainstem reticular excitability. Exp. Neurol., 1959, I : 407426. Amy, W. R., MERRILLEES, N. C. R. and SUNDERLAND, S. The entorhinal area; behavioural, evoked potential and histological studies of its interrelationships with brainstem regions. Brain, 1956, 79: 414439. ADEY,W. R., WALTER, D. 0. and LINDSLEY, D. F. Effects of subthalamic lesions on learned behavior and correjated hippocampal and subcortical slow-wave activity. A.M.A. Arch. Neurol., 1962, 6 : 194-207. LINDSLEY, D. F. and ADEY,W. R. Availability of peripheral input to the midbrain reticular formation. Exp. Neurol., 1961, 4 : 358-316. J. M., CHAMBERS, W. W. and STELLAR, E. Attentive, affective and adaptive behavior in SPRAGUE, the cat. Science, 1961, 133: 165-173. A. ARDUINI:
I should like to ask Prof. Dell whether the behavioural changes produced by arousing stimuli disappear together with the reversal of the EEG changes. W. GREYWALTER:
I want to raise the general question of whether we can safely interpret EEG records in terms of “arousal” on subjective criteria of pattern changes such as Dell presents. Both in human EEG studies
DISCUSSION
and in animal experiments there seem to be conditions of behavioural lethargy, if not sleep, in which the electrical activity is asynchronous and of low amplitude. In normal human subjects the very first stages of drowsiness are usually associated with disappearance of alpha rhythms, before other features such as theta or delta rhythms appear and specific chronic de-afferentation (as in blind people) is usually followed by absence of the characteristic local rhythms. Secondly I should like to ask for more details about the action of COz, which I understand “activates” cortex but suppresses monosynaptic reflexes. Is the cortical activation a direct effect on the neuronic systems or an indirect one through vasodilatation or through the ascending reticular system? We have found that in man the administration of 7% COZcan suppress delta activity associated with midline disturbances, but it has no effect on focal delta activity due to local organic lesions, nor does it suppress the delta activity of deep sleep, even when this latter can be instantly abolished by sensory arousal, and the COZconcentration in the inspired air is high enough to induce a market dyspnoea in the sleeping subject. This may suggest that sleep involves a cumulative suppression of cortical activity by sensory occlusion and cortico-reticular inhibition. In relation to the latter, is the modulation of reticular function by cortex derived from the whole neocortical system or from paleo-cortex or only from certain limited regions?
F. BREMER: Our colleagues Hugelin and Mme. Bonvallet are familiar with the objections which I have made to their conception of a reticulo-cortico-reticular feed-back loop. These objections in no way concern the validity of the experimental findings nor their functional importance. The point to be considered concerns the target of the corticofugal influences which constitute the counterreaction to the reticular arousal of the cortex. Hugelin and Bonvallet believe they have demonstrated that these influences return to the ascending reticular system which emitted the corticopetal impulses, and that they inactivate these by neuronic inhibition. Yet the only known effects of cortico-reticular discharges consist of activation rather than inhibition of the reticular arousal system. On the other hand, the amplitude of the monosynaptic trigeminal reflex could not be considered, in my opinion, as an indication of tonic activity of the reticular arousal system unless it were demonstrated that the neurons of this system are identical with those of the facilitating descending reticular formation. This identity has hitherto been based only on the features of Golgi sections described by the Scheibels. Moreover the reflex, after its initial phase of facilitation in reticular arousal, may show a reduction of amplitude relative to its initial value (tracings presented by Hugelin and Bonvallet). There is therefore not only defacilitation in this case, but active inhibition. The question can finally be raised as to whether the effect of anoxia on the spinal activity might not be explained by a direct action on the neurons of the cord, which are known to be more sensitive to all convulsive agents than are those of the cortex.
HUGELIN,A. et BONVALLET, M. Tonus cortical et contrBle de la facilitation motrice d’origine reticulaire. J . Physiol. (Paris), 1957, 49: 1171-1223. SCHEIBEL, M. E. and SCHEIBEL, A. B. Structural substrates for integrative patterns in the brain stem reticular core. In: Reticular Formarion ofthe Brain. Little, Brown and Co., Boston, 1958: 31-55. A. HUGELIN: In his report this morning Professor Dell has shown the existence of several mechanisms controlling reticular activity. Among these, he has described the functioning of an operational reticulo-corticoreticular system. I should like to indicate the possible role of such a system in the integration of information at the level of the final common pathways. But first 1 should want to answer some questions raised by Professor Bremer. Professor Bremer asked whether the corticofugal inhibition which opposes the reticular facilitation of the motor neurons might not act a t the level of the reflex circuit used in testing, rather than at the level of the reticular cells, as we suggested. To my knowledge there are no arguments immediately applicable either in favour of or against this theory, with the exception of anatomical studies by Walberg and Kuypers, mentioned by Dr. Pompeiano, which are against Professor Bremer’s hypothesis.
106
P. DELL
On the contrary, there are several indirect arguments against this hypothesis. I n his presentation this morning, Professor Dell first showed that arousing stimuli give rise not only to corticographic activation and facilitation of motor neurons, but also to a series of somatic and vegetative manifestations. He then showed that all these effects of reticular origin are subject to the inhibitory control of the cerebral cortex. If we accept the hypothesis of a specific corticofugal pathway of motor neuron inhibition, we must also postulate the existence of numerous specific corticofugal pathways, each opposed to the effects of the reticular activating system at the level of the different specific centers upon which this system is projected. The specific pathways would facilitate the first relays of flexion reflex afferents, and inhibit the interneurons of the pyramidal system, the cells of the respiratory centers, those of the vasomotor and cardio-accelerator centers, the effector neurons of the psychogalvanic reflexes, the pupil-dilating neurons, etc. Such a complication has not been excluded but does seem rather unlikely. On the other hand, there are numerous arguments in favour of inhibition of reticular cells of corticofugal origin. We know, first of all, that the corticofugal inhibitory pathways pass through the internal capsule, emerging from corticospinal fibres at the level of the junction of the mesencephalon and diencephalon and bordering on the dorsal surface of the mesencephalic tegmentum. Direct electrical stimulation is no longer capable of effecting motor inhibition below this level; in view of this fact S. Tower (1936) maintained that the corticofugal inhibitory pathways inhi bit the facilitating extrapyramidal structures localized at the level of the upper brain stem. In this respect 1 should like to answer parenthetically a question raised by Professor Bremer. Professor Bremer remarked that, in some of our figures, the monosynaptic reflex which tests reticular excitability is not only restored to its initial level but obviously inhibited. We believe that this type of effect of direct electrical stimulation of the mesencephalon is due to simultaneous stimulation of the facilitating neurons and the corticofugal inhibitory pathways. This explanation is corroborated by two facts: I . We have never obtained secondary depression of the monosynaptic reflex by stimulating regions other than the dorsocaudal part of the lateral hypothalamic area and the rostrodorsal part of the mesencephalon. 2. This effect is not observed when reticular neurons are excited by the extralemniscal route. Therefore, I do not think that the secondary depression of the monosynaptic reflex could be used as an objection (in any case not absolute) to the theory of servo-control. There are other arguments in favour of corticofugal inhibition of reticular cells. All those authors who have recorded from isolated mesencephalic reticular cells have been struck by the fact that spontaneous unitary activity, or evoked activity, often develops in a direction which is the reverse of that expected in a purely passive system. Schlag (1958) believes that this may be caused by the inhibitory activity of the corticofugal system. BureS and BureSova (1961) observed a considerable increase in activity of numerous mesencephalic reticular cells at the time of onset of “spreading depression”; these authors believe that the increase in corticofugal inhibition affords a probable explanation for their results. We ourselves, with Bonvallet and Dell, observed the same phenomenon when cortical activity was arrested following anoxia (Hugelin, Bonvallet and Dell 1959; Bonvallet and Hugelin 1961). But Professor Bremer criticizes the use of anoxia and suspects that errors are introduced into our experiments because the oxygen deficiency acts on the central nervous system via several mechanisms. I should like to make an attempt to reassure him, showing why these experiments are conclusive. It is possible to demonstrate 3 different mechanisms jointly effecting an increase in reticular cellular activity during anoxic anoxia, viz. : ( I ) excitation of chemoreceptor origin; (2) liberation of a corticofugal inhibitory influence and (3) direct humoral excitation. When complete, immediate anoxia is produced, these 3 mechanisms act almost simultaneously. But because their thresholds are different, they can be dissociated during the progressive increase in hypoxia. This has been done in our experiments by having the subjects inhale a mixture of nitrogen and 6.5% oxygen. Activation of the chemoreceptors is apparent when the PaOz reaches 85 mm Hg and increases to the extent to which the anoxia progresses. Cortical electrical activity is disorganized and abruptly disappears within 1-5 seconds ~ 60 mm Hg. It is only a t this moment that the reticular cells begin to be directly when the P A Oreaches excited (by an unknown mechanism), and this excitation progressively increases until the P A O ~ reaches 45 mm Hg. To answer another question by Professor Bremer, our experiments were made in “enckphale isole” animals, and it was simultaneously verified that the spontaneous discharge of motor neurons and facilitation of spinal reflexes below the section of the cord were not changed as a result of hypoxia
DISCUSSION
107
except at P A O values ~ near or below 45 mm Hg, which is well after the onset of reticular facilitation. The control of reticular cells by corticofugal inhibition is demonstrable by comparing the effects of standard progressive hypoxia in preparations with and without cortex. When the effects of the cerebral cortex have been eliminated by diencephalic section of the brain stem, acute progressive hypoxia - by the chemoreflex route - provokes progressive facilitation of the cranial monosynaptic reflex used as test, and simultaneously a regularly increasing activation of the reticular cells, the frequency of discharge of which finally doubles; these two phenomena begin about 90 sec after the start of the experiment and gradually attain their maximum after I50 sec. The results are very different in preparations with an intact cerebral cortex; at 90 sec there is only a slight increase in frequency of certain reticular cells, chosen because they are excited by several kinds of peripheral stimuli and because their excitation always coincides with cortical desynchronization. No facilitation of motor neurons is observed at that time. After 120-150 sec, a t the exact moment of cessation of cortical activity because of lack of oxygen, motor facilitation abruptly appears, reaching its maximum in 1-5 sec. The frequency of reticular activating units doubles at that moment. The coincidence of the immediate motor facilitation and the massive increase in the frequency of discharge of the reticular cells with the disappearance of cortical activity was observed in all cases, without exception: this leaves no doubt as to the correlation which exists between the two types of phenomena. In our opinion this establishes with certainty the existence of a control of reticular activating cells and facilitating cells by inhibition of cortical origin. The fact that the corticofugal inhibition takes place at the level of the reticular cells can be regarded as established; two hypotheses can therefore be envisaged. The first is that the reticular ascending activating system (RAAS) and the reticular descending facilitating system (RDES) arise from the level of the same reticular neurons, and that inhibition of cortical origin takes place precisely at the level of these neurons. The second is that the RAAS and the RDFS are functionally distinct mechanisms, and that corticofugal inhibition acts on the neurons of RDFS only. If the RAAS and the RDFS are functionally distinct, they may be brought into action separately under natural conditions. It should then be possible, after excitation of the RDFS, to observe motor facilitation without cortical activation. 1. In favour of this hypothesis, it seems, is the fact that apparently spontaneous generalized activations of the corticogram are often accompanied by a depression of monosynaptic responses of the motor neurons, and that the same phenomenon can be observed after certain arousal stimulations have been immediately interrupted; we know, however, that in the former case there may be generalized cortical activation of humoral and extrareticular origin (cf. Desmedt and La Crutta 1957; Hugelin, Bonvallet and Dell 1959). We also know that the latter case may be explained by prolonged post-discharge of the cortex (Hugelin and Bonvallet 1957~). 2. The following major objection to this hypothesis can be made. Taking the precaution to test the excitability of motor neurons at intervals sufficiently short to ensure that a facilitation of a few tenths of a msec cannot escape observation, it has never been possible to obtain, by stimulation of a peripheral nerve, cortical arousal or isolated motor facilitation (Hugelin and Bonvallet 1958). 3. The effects of direct reticular stimulations are also in disagreement with the hypothesis of duality of the ascending activating and descending facilitating systems. In fact it is impossible to obtain motor facilitation without activation by varying the points of excitation and the parameters of stimulation and, although cortical activation without facilitation is readily obtained during a stimulation, this is never accompanied by motor inhibition except when the region of the corticofugal inhibitory pathways is stimulated. The hypothesis of functional unity of the RAAS and the RDFS seems more in accordance with the experimental results obtained (Hugelin and Bonvallet 1957b) but it is more difficult to understand because it rests, not only on anatomical notions but on principles of cybernetics. In this hypothesis it is accepted that ascending and descending messages are simultaneously emitted by the neurons of the same network, and that cortical activation is the origin of an inhibitory discharge which acts on the reticular neurons proper. According to this theory, the network of reticular neurons must be a controlled system, and the inhibitory cortical system a controlling system. Any change in activity of the controlled system is transmitted to the controlling system which then opposes the change. When repetitive stimulation has been established, motor facilitation is a t the same time effected as a consequence of the increase in the descending discharge; cortical activation is effected because the ascending discharge is increased; this in turn gives rise to the corticofugal discharge which opposes the excitation of the reticular cells. After an interval which corresponds with the delay in cortical action, both the ascending and the
108
P. DELL
descending effects must be reduced; this is in fact what is observed but, if facilitation ceases and if the system is stable, the cortical tonus remains higher than before stimulation started (cf. Dell report, p. 92). This last point can be understood if it is borne in mind that, in order to be effective, a controlling system must be more sensitive to disturbances than the system it controls. Because they undergo more considerable amplification in the cortex than in the motor neurons, the efferent reticular signals respond to this condition. The hypothesis that the operational system studied acts as a servo-mechanism is discussed in detail in a publication which we can only mention briefly. It suffices to add that, from a cybernetic point of view, the behaviour of such a system can be logically predicted. It has been possible to design an electronic model of a control system, taking into account the time of conduction and the inertia of the multineuronic systems. The responses obtained on the basis of such an analogous system (analogous to the response to disturbances) are comparable with the curve of excitation of motor neurons obtained during sustained reticular stimulations. However, it is probably more appropriate to ask how a control system consisting of nerve cells functions. In Fig. 10, an attempt has been made at depictinga simplified neuronic circuit capable of behaving in a manner similar to that of a reticulo-cortico-reticular system. In accordance with conclusions based on experimental findings, it was decided that, in principle: (1) the schema should include a single activating system simultaneously exciting the cortex and the motor neuron; (2) the cortex should be more sensitive to the reticular discharge than the motor neuron; (3) the cortex should inhibit the activating cells themselves; (4) the conduction time in the counter-reaction loop should be longer than the conduction time in the reticulospinal system.
Fig. 10 The schema consists of elements functioning on the “all or nothing” principle. It is agreed that the threshold of a neuron is reached when 4 excitatory synaptic terminations (black points) are simultaneously activated; that one inhibitory termination (open circle) annuls the effect of an excitatory termination; and that the conduction time is the same in each element, and equals the unit. The schema comprises several cell types. In the centre 3 neurons (reticular neurons R1-3) receive an indefinite number of excitatory terminations. A stimulation of intensity S causes discharge of a single
DISCUSSION
109
reticular neuron; an impulse of 2 s intensity causes discharge of RI and R2, and a shock of 3s intensity causes discharge of the 3 reticular neurons. Each reticular cell (of the type described by Scheibel and Scheibel 1958) issues an axon which bifurcates in T ; the ascending branch terminates at the level of a cortical cell (CI-3) in numerous excitatory terminations; the descending branch ends on a motor neuron MN but has only feeble excitatory power. The cortical cells in their turn excite an inhibitory reticular neuron IR, the connections of which are organized in such a way that it may be activated either by two cortical cells or by a cortical cell and an interneuron of a recurrent circuit of re-excitation. Two cases may be considered. That of repetitive stimulation, the impulses of which at once have a maximum voltage, and that of successive impulses of increasing intensity. (a) In the case of stimulation with impulses generally reaching an intensity of 3S, the 3 reticular cells R1, R2 and R3 discharge at time t. At time t+ 1, the motor neuron can discharge if it is simultaneously excited by fibre la. The 3 cortical cells also discharge. At time tf2, IR is excited. If a second impulse of intensity 3s is sent at time t+ 3, R1 and R3 may be the only reticular cells discharging, while R2 is inhibited by the discharge of IR; at time t f 4 , MN cannot discharge. When C1 and C3 are again excited, the system at time t+6, t+9, etc. is found in the same condition as at time tf3. Conclusion: after a phase of disequilibrium due to inertia of the counter-reaction loop, the state of reticular excitation reaches a stable phase which is less high than if the control loop did not exist; this level is sufficient to maintain cortical activation but insufficient for facilitation of motor neurons. (b) Supposing the intensity of impulses were only 2s. RI and R2 discharge at time t; at time t+ 1, MN is not facilitated but CI and C2 are excited. As a result, IR discharges at time t + 2 and only R1 is excitable at time t+ 3. The activity of the loop of re-excitation of IR ensures that only R l is excited
Fig. 11 at times t+6, t+9, etc. In this case, after cortical control, the state of reticular excitation and that of cortical excitation are less than they would be if the impulses of stimulation had 3s intensity. The comparison of results obtained with stimulations of different intensities shows that the state of reticular excitation is less with than without the control loop, but it remains proportional to the intensity of stimulation. (c) Because of the existence of the reticulo-cortico-reticular control loop, the motor effects obtained are different if stimulation is established at once at 3s or if it attains this intensity gradually. When 1s stimulation is applied at time t, R1 discharges; C1 is excited at time t + l , but IR does
110
P. DELL
not discharge, Consequently the second impulse of 2s intensity, applied at time t+ 3, caused discharge of CI and C2 at time t+ 4,-and IR at time t+ 5. At t+ 6, R2 is inhibited and an impulse of 3s intensity applied at that time, cannot discharge anything but R1 and R3. In this case, the inertia of the counterreaction system has been eliminated and M N is never facilitated. The physiological importance of this control mechanism can be readily understood when bearing in mind the difference in sensitivity of the specific mechanisms of reticular discharge. Fig. I 1 was obtained on the basis of the results presented this morning by Professor Dell, which show that feeble reticular stimulations strongly inhibit the flexion reflex afferents I1 and 111, that reticular stimulation of medium intensity produces a long lasting facilitation of corticomotor responses, and that only abrupt stimulation of high intensity transiently facilitates the monosynaptic reflexes (Hugelin 1961). Thelowerpart ofFig. 11 shows thesegmental control of the final common path to flexor and extensor muscles, by the afferent fibres of groups la, I1 and 111; on the right side of the scheme a cortical neuron of the pyramidal system is represented, and an interneuron of the same system. A study of this scheme shows that the polysynaptic flexion reflex is inhibited when a single reticular cell is excited; that the MN/FI can discharge and respond to corticomotor excitation when two reticular cells are excited; and that MN/Fl and MN/Ex are facilitated when three reticular cells discharge. (a) Let us consider the case of stimulation of 3s intensity. R1, R2 and R3 discharge at time t. The flexogenicafferences 11 and 111 are blocked at time t+ 1, At the same time, MN/FI - facilitated by the reticular discharge - can discharge in response to stimulation of the Ia flexor fibre; MNjEx is released from inhibition exerted by the flexion reflex afferences and, because also facilitated, can respond to stimulation Ia’Ext. MN/Fl can also respond to the pyramidal discharge at time t + l or t+2. The entire operation can be regarded as a transfer of active moto-neuron connections by selection of information. But at time t+3, when IR is discharged, R2 is no longer excitable; the result is that, at t+4, the flexogenic afferences are always inhibited; the pyramidal movements are still facilitated but the monosynaptic reflexes no longer are. (b) Now let us consider the case of stimulation by a train of impulses progressively increasing in intensity. At time t, R1 discharges in response to a shock of IS intensity. At time t+ I , only theflexion reflex afferents are altered by the reticular discharge. At t+3, R1 and R2 discharge in response to a 2s stimulation; the flexogenic afferents are inhibited at t f 4 and the corticomotor response is facilitated at t+ 5. At t+6, when IR discharges, a 3s impulse can only excite R1 and R3 in that, because the flexion reflex afferents are inhibited and the corticomotor responses facilitated, the monosynaptic responses are unchanged. Thus it becomes clear that the intervention of a negative feed-back loop modulates the effects of reticular stimulation and is one of the factors of differentiation in the motor neuron behaviour. The reticular discharge effects a transfer of connection by selection of information and the control loop renders this selection sequential. Our conclusion is that the first objection which comes to mind when envisaging the possibility of a unique activating system projecting in both ascending and descending directions onto numerous structures, is that this hypothesis is untenable because such a single system must operate “en bloc”. The introduction of the idea of differential sensitivity of the effectors to the discharge of this structure, however, indicates that this system may be a very flexible one. It is entirely conceivable that the functional flexibility of this unique activating system may be explained by control of its activity by a servo-mechanism. This would automatically and economically effect differentiations in the level of reticular activity thus, in the final analysis, determining the various kinds of possible behaviour. BONVALLET, M. et HUGELIN, A. Influence de la formation reticulaire et du cortex cerebral sur l’excitabilite motrice au cours de l’hypoxie. Elecrroenceph. clin. Neurophysiol., 1961, 13: 270-284. BURES,J., BURESOVA, 0. and FIFKOVA, E. The effect of cortical and hippocampal spreading depression of activity of bulbopontine reticular units in the rat. Arch. ital. Biof., 1961, 9 9 : 23-32. J. E. and LA GRUTTA, G. The effect of selective inhibition of pseudocholinesterase on the DESMEDT, spontaneous and evoked activity of the cat’s cerebral cortex. J. Neurophysiol., 1957, 136: 2040. HUGELIN, A. Integrations motrices et vigilance chez l’encephale isole. 11: ContrBle reticulaire des voies finales communes d’ouverture et de fermeture de la gueule. Arch. iral. Biol., 1961, 99: 244-269. HUGELIN, A. et BONVALLET, M. Tonus cortical et contrdle de la facilitation motrice d’origine reticulaire. J . Physiol. (Paris), 1957a, 49: 1171-1200. HUGELIN, A. et BONVALLET, M. Etude experimentale des interrelations reticulo-corticales. Proposition d’une theorie de I’asservissement reticulaire ii un systeme diffus cortical. J . Physiol. (Paris), 1957b, 49: 1201-1223.
DISCUSSION
111
HUGELIN, A. et BONVALLET, M. Analyse des post-decharges reticulaires et corticales engendrees par des stimulations Blectriques reticulaires. J. Physiol. (Paris), 1957c, 49: 1225-1234. HUGELIN, A. et BONVALLET. M. Effets moteurs et corticaux d’origine reticulaire au cours des stiniulations somesthetiques. R61e des interactions cortico-rkticulaires dans le determinisme du reveil. J. Physiol. (Paris), 1958a, 50: 951-977. M. Mise en evidence d’un contr6le cortical de l’etat d’excitation reticuHUGELIN, A. et BONVALLET, laire. C.R. Acad. Sci. (Paris), 195813,246: 1738-1741. HUGELIN, A., BONVALLET. M. et DELL,P. Activation reticulaire et corticale d’origine chemoreceptive au cows de l’hypoxie. Electroenceph. clin. Neurophysiol., 1959, I1 : 325-340. M. E. and SCHEIBEL, A. B. Structural substrates for integrative patterns in the brain stem SCHEIBEL, reticular core, in Reticular Formation of the brain, Little, Brown and Co., Boston, 1958: 31-55. SCHLAG, J. L‘activite‘ spontanee des cellules du systPme nerveux central. Arscia, Bruxelles, 1959 : 185 pp. TOWER,S. S. Extrapyramidal action from the cat’s cerebral cortex: motor and inhibitory. Brain, 1936, 59: 408444. W. GREYWALTER: With reference to the scheme of Hugelin we should realise that the change of input signal form could be accomplished by a much simpler passive network, and the time-lag by a delay-line. To prove that an active feed-back pathway really exists -or, in the case of the scheme, is necessary and sufficient is not so easy. However, one of the most satisfactory ways - assuming that we are dealing with a nearly “black box” - would be to alter the time constants of the delay circuit, to see if a condition could be obtained in which the feed-back would become positive with enough gain to establish oscillations. These might be identified with the spindle bursts or another apparently autochthonous rhythm in the brain, in conditions when the cortico-reticular relations are upset. reply to Grey- Walter HUGELIN, It is true that all known oscillatory systems are feed-back systems, but not all feed-back systems are oscillatory.
H. GASTAUT: Dr. Hugelin has offered several physiological and cybernetic data proving the existence of a corticoreticular inhibitory system. For my part I should like to present the following clinical evidence. Any cerebral anoxia - pathological or experimental, in man or in animals and regardless of its cause (anoxic anoxia, toxic, haemorrhagic, ischaemic anoxia, etc.) - first causes slowing and then silence of the cortical electrical activity. Simultaneously with the electrical silence of the cerebral cortex, we always see generalized contraction of the trunk and the limbs, evoking the rigidity of decerebration. A long time ago, well before the experiments of Hugelin, Noell and Donibrowski demonstrated that this rigidity depends on activation of tonigenic reticulospinal structures, electrically translated into activation of electrogenesis in the reticular formation of the brain stem. Also long ago I have reported - on the basis of the experiments and conceptions of H. Jackson - on a reticular “liberation”, by exclusion of a corticoreticular brake, of tonic spasms accompanying all cortical anoxias. I should like to point out once again that convulsions are only bilateral if the cerebral anoxia is bilateral (as in generalized cerebral ischaemia responsible for syncopes). When the anoxia affects only one cerebral hemisphere (as in sylvian cerebral ischaemia associated with some types of circulatory insufficiency), the convulsions are confined to the contralateral side. This suggests that corticoreticular inhibition is crossed - a suggestion which should be verified by animal experiments. NOELL,W. and DOMBROWSKI, E. Cerebral localisation and classification of convulsions produced by severe oxygen lack. School of Aviation Medicine, Randolph Field, Texas, 1947. 0. POMPEIANO:
Your conclusion that stimulation of the mesencephalic reticular formation is able to drive secondarily
112
P. DELL
a phasic ascending bulbar inhibitory control is grounded on the evidence that the cortical arousal elicited by mesencephalic stimulation is more intense and longer lasting after a medial prebulbar section of the brain stem. Judging from your illustrations, the spontaneous EEG background may have changed under the new experimental conditions; in fact the pattern of low voltage fast activity became more prominent after the prebulbar section. Since this section was made acutely, the EEG background changes induced by the lesion indicate that the higher degree of cortical desynchronisation might be the consequence of irritative phenomena, which could account for the more intense phasic arousal obtained after the section. It is true that you reproduced the same phenomenon after local application of novocaine in the caudal medulla, however one should not disregard the possibility that even then a puncture of the medulla together with the local vascular damage which often results from this procedure may cause some irritation. This objection could possibly be avoided by injecting the anaesthetic through a chronically implanted needle in the medulla. A parallel control should also be made by injecting the same volume of Ringer solution into the medulla and studying the effects of this injection on the EEG response elicited by stimulation of the mesencephalic reticular formation. Professor Bremer has suggested that a possible direct connection between the cortex and the motor nuclei of the cranial nerves might explain the corticofugal influence on the monosynaptic masseter reflex. It should be recalled that anatomical evidence of such a direct cortical projection to the cranial motor nuclei is lacking, at least in the cat (Brodal and Walberg 1960). Therefore the impulses from the cerebral cortex to the peripheral motor neurons of the cranial nerve nuclei must be assumed to reach them by way of intercalated neurons of the reticular formation, as suggested by Professor Dell. Anatomical observations have already shown that the reticular formation, which receives abundant corticofugal fibres, contains some cells which give off axons or collaterals to the peripheral motor neurons of some cranial nerve nuclei (Scheibel and Scheibel 1958).
F. Anatomical studies of some corticofugal connections to the brain BRODAL,A. and WALBERG, stem. In: D. B. TOWER and J. P. SCHADB(Editors) Structure and function of the cerebral cortex. Proceedings of the Second International Meeting of Neurobiologists, Amsterdam, 1959. Elsevier Publishing Co., Amsterdam, 1960: 116-123. M. E. and SCHEIBEL, A. B. In: Reticular Formation of the Brain, Intern. Symp. Henry SCHEIBEL, Ford Hosp. Little, Brown and Co., Boston, 1958. W. R. ADEY: The evidence from our experiments does not suggest that the modification of EEG rhythmic patterns in hippocampal structures following subthalamic lesions results from a lack of “arousal value” in the stimulus, due to some “hemianopic” interference with visual pathways as such. The strongest evidence is undoubtedly that after such lesions, either unilateral or bilateral, there appeared to be no interference with classical Pavlovian reflexes, but only with the discriminative aspects of the learned performance. Even at the height of the defect induced by the lesions, these animals unfailingly performed the motor acts initiating the approach whenever the test situation was presented, but failed grossly in the discriminative performance. Moreover, following a unilateral subthalamic lesion, the interference with hippocampal rhythmic patterns was bilateral and symmetrical. Additionally as I have already indicated, there was clear evidence of “downstream” effects following these lesions in midbrain reticular zones which our own and other studies have shown to receive physiological inputs from both amygdaloid and hippocampal areas of the rhinencephalon (Adey 1958; Gloor 1955). In our experiments, chronic subthalamic lesions reduced excitability of midbrain reticular neurons to peripheral stimuli, and evoked potentials in these same areas from sciatic nerve stimulation were similarly reduced in acute experiments. The regular trains of slow waves normally appearing in the midbrain reticular formation during the discriminative performance, concurrently with the theta wave train in the hippocampal system, were abolished by the subthalamic lesions (Adey, Walter and Lindsley, 1962). ADEY,W. R. The organization of the rhinencephalon. In: The Reticular Formation of the Brain. Intern. Symp. Henry Ford Hosp. Little, Brown, Boston, 1958: 621-645. D. F. Effects of subthalamic lesions on learned behavior ADEY,W. R., WALTER,D. 0. and LINDSLEY, and correlated hippocampal and subcortical slow-wave activity. A.M.A. Neurol., 1962,6 : 194-207. GLOOR,P. Electrophysiological studies on the connections of the amygdaloid nucleus in the cat. Electroenceph. clin. Neurophysiol., 1955, 7 : 223-242.
DISCUSSI 0N
113
H. GASTAUT: If I have understood Dr. Adey correctly he suggests that a subthalamic lesion interferes with the activities of the hippocampus and the reticular formation of the brain stem. I should very much like to know precisely what kind of interference he envisages. Is it a matter of a direct mechanism represented by the interruption of a hippocampo-reticular pathway via the subthalamus, a pathway for which I know of no anatomical proof, or perhaps an indirect mechanism represented by action at a distance from the subthalamic lesion mediated by a perifocal oedema or some other factor?
W. R. ADEYto H . Gastaut: The arrangement of the connections between the hippocampal system and subcortical structures appears from physiological experiments to be somewhat more extensive than the neuroanatomically defined connections which Dr. Gastaut has indicated. It is, however, necessary to think of these connections in terms of both afferent and efferent pathways which reciprocally interrelate hippocampal and subcortical structures. As our previous studies have indicated (Adey, Dunlop and Sunderland 1958), there is convergence from widespread diencephalic and mesencephalic areas on the hippocanipal system. The efferent pathways from the hippocampus descending to mesencephalic levels appear more restricted and form two main streams, one descending through epithalamic and dorsal thalamic structures, and the other running caudally in a more ventral position in the junctional zone between thalamus and hypothalamus. The latter run largely through the subthalamic and adjacent areas damaged in our studies. Amygdaloid influences reaching the globus pallidus (Adey 1959) may also pass through this region to more caudal levels (Johnston and Clemente 1959). It isalso relevant that Eidelberg, White and Brazier (1959) found major alterations in hippocampal evoked potentials following lesions in the nucleus centrum medianum. Our previous studies have provided the first information on the convergence of activity from the hippocampal cortex on the dorsal tegmental areas of the rostra1 midbrain, adjacent to the periaqueductal gfey matter (Adey, Merrillees and Sunderland 1956; Adey 1957). ADEY,W. R. In: Reticular Formation of the Brain. Intern. Symp. Henry Ford Hosp. Little, Brown and Co., Boston, 1957. ADEY,W. R. Recent studies of the rhinencephalon in relation to temporal lobe epilepsy and behavior disorders. lntern. Rev. Neurobiol., 1959, I : 1 4 6 . S. A survey of rhinencephalic interconnections with ADEY,W. R., DUNLOP,C. W. and SUNDERLAND, the brain stem. J. comp. Neurol., 1958,110: 173-204. ADEY,W. R., MERRILLEES, N. C. R. and SUNDERLAND, S. The entorhinal area; behavioural, evoked potential and histological studies of its interrelationships with brainstem regions. Brain, 1956, 79: 414439. E., WHITE,J. C. and BRAZIER,M. A. B. The hippocampal arousal pattern in rabbits. EIDELBERG, Exper. Neurol., 1959, 1 ; 483-490. C. D. An experimental study of the fiber connections between the JOHNSON,T. N. and CLEMENTE, putamen, globus pallidus, ventral thalamus and midbrain tegmentum in cat. J. comp. Neurol., 1959, 113: 83-101. JOHNSON, T. N. and CLEMENTE, C . D. An experimental study of the fiber connections between the putamen, globus pallidus, ventral thalamus and midbrain tegmentum in cat. J. comp. Neurol., 1959,113: 83-98.
P. DELL'S replies
To W. R . Adey Is there not an alternative explanation for your interesting results? If I am right there are incidental observations by H a s and also more recent work by Sprague (personal communication) showing that after lesions located at the niesodiencephalic level, well outside the classical visual pathways, the animal behaves as though it were blind or hemianoptic and does not respond to certain visual stimuli. To be integrated and provoke a response, stimuli (we exclude of course habituated ones) must not only affect the cortical analysers but must also have a reticular effect.
I14
1’. DELL
To A . Arduini All these experiments were performed on “encephale isole” preparations or on preparations with brain stem section, We have no information on these studied reflexes in free-running animals in which the mechanisms just described surely play a role but might be masked; you will remember that numerous postural and righting reflexes are easy to demonstrate after adequate brain stem section but difficult to observe in normal preparations. Even so you would probably agree that they are an integral component of the postural behaviour of the species studied. Your second question, or rather your suggestion, is not in agreement with known anatomical facts which stress the branching of the pontine reticular neurons and therefore their ability to produce ascending and descending effects simultaneously. Also may I refer you to the original discussions of Hugelin and BonvalIet (1957) where numerous experimental facts are given in favour of the statement that facilitating descending effects and cortical arousal effects arise from the same brain stem structures. A. et BONVALLET, M. etude experimentale des interrelations reticdo-corticales. Proposition HUGELIN, d’une theorie de l’asservissement reticulaire a un s y s t h e diffus cortical. J . Physiol. (Paris), 1957, 49: 1291-1293. To W . Grey Walter (1) We all agree that the word “arousal” may be disturbing when it is applied to immobilised preparations with section of the brain stem or the spinal cord but this word has now been widely used in this sense for many years. In all the experiments just reported this “arousal” was provoked by reticular or sensory stimulations and therefore we believe that the desynchronisation patterns which appeared were not associated with deep sleep but with awakening. To obtain a more quantitativeevaluation of theintensity of cortical arousal we count the rapid waves of the EEG (30 to 70 c/sec); it is a tedious procedure which however was very useful in many instances. (2) Changes in the partial pressure of COZ in the blood affect various central structures and the final picture depends on the mutual interactions of these structures (e.g., reticular formation and the cerebral cortex). COz, as has been shown in our laboratory, activates the brain stem reticular formation, first by its stimulating effects on the carotid chemoreceptors, and also by its direct action on reticular cells and consequently we observe cortical arousal. Moreover, COZintervenes in cortical activity by a direct effect on cortical cells; this has been shown by studying the changes in amplitude of a monosynaptic reflex in animals with and without cortex (Hugelin and Bonvallet 1957). Not much is known about the mechanism of this COZeffect on the cortex. However from the work of Ingvar and Soderberg (1956) it is clear that cortical vaso-dilatation can appear as a consequence of the cortical arousal; in no instance have these authors been able to observe cortical arousal as a consequence of a previous cortical vasodilatation although they carefully investigated this possibility. (3) The braking action of the aroused cortex arises from a11 cortical areas; it is a diffuse effect and does not arise from limited cortical areas projecting to the reticular formation.
A. et BONVALLET, M. etude experimentale des interrelations reticulo-corticales. Proposition HUGELIN, d’une theorie de I’asservissement reticulaire 9 un systeme diffus cortical. J . Physiol. (Paris), 1957, 49: 1291-1293. INCVAR, D. V. and SODERBERG, U. A. A new method for measuring cerebral blood flow in relation to the electroencephalogram. Electroenceph. clin. Neurophysiol., 1956, 8 : 40341 5. To 0. Pompeiuno It seems to me that an injection of novocaine in a brain structure has few chances of inducing irritative phenomena. In most cases, after the injection of novocaine into the bulbar structures, there was no conspicuous change in the cortical tracing as judged by the interval between spindle bursts and the duration and shape of each burst; in a few cases however the cortical tracing was more active. It would be interesting to repeat these experiments in chronic preparations though 1 am afraid that such injections would impair the activity of the respiratory centers. I thank you for reminding us that there are no direct connections from cortical areas to brain stem motoneurone pools. As explained by Hugelin in another part of this discussion in reply to Professor Bremer’s objections, these corticofugal inhibitory effects not only act on monosynaptic reflex arcs but on numerous other responses also, e.g., polysynaptic reflexes, respiratory activity, sympathetic discharges. SO YOU would have to postulate connections between cortex and most of the effector niechanisms; a more economical way of obtaining the same result is that cortex inhibits reticular activity; numerous published experiments substantiate the existence of this cortico-reticular control.
Thalamic Integrations and their Consequences at the Telencephalic Level* D. ALBE-FESSARD
AND
A. FESSARD
Centre d’budes de Physiologie Nerveuse, Centre National rle la Recherche Scientifque, Paris (France)
It has been traditionally assumed for a long time that the associative areas of the cerebral cortex must be the principal, if not the unique, site for central integration of sensory messages. We realize to-day, however, that an important proportion of these integrations are due to the convergence of impulses taking place at the level of the thalamus, or at lower levels. We shall designate this process as a “projected convergence”, when we recognize its characteristic features above the level at which it actually takes place. We propose to describe here some recent results and to present some new ideas on the organization of certain of these systems of afferent integration. Although we have considered the general problem of multisensory integration, we have specially treated here the case of the somesthetic projections evoked by their spinal afferents. These belong to two systems, which are anatomically distinct, the lemniscal system and the extralemniscal system, named also, for their functional properties, discriminative and convergent, a nomenclature which is, in fact, only a manner of reviving the old distinction of Head of “epicritic” and “protopathic”. In a first section (A), we shall examine some integrative aspects presented by diverse, non-primary, telencephalic projections and we shall suggest that they are not necessarily related to primary activities and may depend upon lower integrations. The second section (B) is devoted to the integrations of somatic afferents in the thalamus, as well as to subjacent integrations in the brain stem, which thus give rise to “projected convergence” towards the diencephalon, and then towards the telencephalon. The third section ( C ) considers the thalamo-telencephalic projections of the extra-lemniscal system and finally the interactions which take place between the two systems of afferents at the level of the primary somatic area. A. ASPECTS OF INTEGRATION
AT THE TELENCEPHALIC LEVEL
THE ELECTROPHYSIOLOGICAL EVIDENCE
1. The integrative capacity attributed to the associative cortex has been accepted for a long time as a principle although it had not been demonstrated physiologically. ~
*
~___
The research reported in this document has been sponsored in part by The Air Force Office of Scientific Research OAR through the European Office, Aerospace Research, United States Air Force. References p . 145-148
116
D. ALBE-FESSARD, A. FESSARD
The first clear indications of a regional convergence (local or projected) of peripherally induced impulses projected upon a common cortical zone were those for area S 11, which is known to receive somatic and auditory afferents; but the true demonstration that common neurons were also implicated dates from occlusion experiments by Bremer, Bonnet and Terzuolo (1954); and for another associative territory, from those of Amassian (1954), who used microelectrodes to record the elementary impulses and showed that in the cat, the cells in the anterior marginal gyrus receive convergent messages which come from diverse regions of the body and also from visual and auditory receptors. To these associative territories, there have been added in rapid succession two suprasylvian zones (anterior and posterior), and also the anterior region of the sigmoid gyrus (the motor area), in which, at first, convergence of somatic afferents
Fig. 1 Cut anesthetized with chlorufose. Hemicortical representation of the principal regions (S I1 excepted) in which activity can be evoked by electrical stimulation of the two anterior limbs (contralateral, a c ; homolateral, ah). [Taken from Albe-Fessard and Rougeul (1 9581, and Jankowska and Albe-Fessard (1961)J.Monopolar recording, indifferent electrode in the frontal sinus. In this and the following figures, a downward deflection corresponds to a positivity of the active electrode.
(Albe-Fessard and Rougeul 1955, 1958) (Fig. l), and later of visual (Fig. 2) and auditory afferents were demonstrated (Buser and Borenstein 1956, 1959; Buser and lmbert 1961 ; Libouban and Jutier 1961). Moreover Bruner and Buser (1960) have described analogous multisensory responses from the medial surface of the cat’s brain. Recently, Jankowska and Albe-Fessard (1961) have provided evidence for another zone of convergence at the S I level near to the primary projection focus of the anterior limb (Figs. 1 and 18). In the monkey, an analogous system of zones of convergence has been described in the frontal and the parietal cortices (Albe-Fessard, Rocha-Miranda and OswaldoCruz 1959). Areas of convergence involving somatic, visual, auditory and olfactory afferents
THALAMIC INTEGRATIONS
vis
1
+
100msec U
117
ac
3
ac
Fig. 2 Total occlusion of a somatic response (first and last trace) by a response to visual stimulation (noiseless white flash), observed at the level of the anterior marginal gyrus (Libouban and Jutier 1961).
upon common neurons have been demonstrated in other telencephalic structures of the cat (Fig. 3) : the caudate nucleus and the putamen (Albe-Fessard, OswaldoCruz and Rocha-Miranda 1960; Albe-Fessard, Rocha-Miranda and Oswaldo-Cruz 1960), the claustrum (Segundo and Machne 1956; Albe-Fessard et al. 1960), and the amygdaloid nuclei and the prepiriform cortex (Segundo and Machne 1956; Wendt and Albe-Fessard 1961). 2. These investigations were made possible by the use of chloralose as an anesthetic. This agent augments the potentials evoked from zones of convergence, whereas the barbiturates reduce them, increase their latency greatly or suppress them entirely. The use of chloralose has provoked some criticism, and the question has been raised as to whether the results obtained under chloralose correspond to a normal
.
100 msec
Fig. 3 Cut anesthetized with chloralose. Convergence recorded in the caudate nucleus at the level of two cells. A, responses of the same element to somatic, visual and auditorystimulation. B, the samecell responds to stimulation of the four limbs and to a cortical stimulation. References D . 145-148
118
D. ALDE-FESSARD, A. FESSARD
function. Responses revealing convergence have been obtained in the awake immobilized animal, without general anesthesia as has been demonstrated by Buser and collaborators (1956, 1958) and Hirsch and collaborators (1961); but the small amplitude of the “associative” evoked potentials and their wide extension indicate a distinct difference between the findings of these authors and the responses observed under chloralose. With several collaborators, we have endeavoured to show that, although chloralose augments associative responses, it is not itself directly responsible for these, but acts only by creating certain favorable conditions which can also be produced by other procedures : (a) At first, because it was suspected that chloralose acted by means of its well known convulsant effect, we have tried other anesthetics. Ether and fluothane have, like the barbiturates, a depressive action. On the other hand, another drug, viadril,* which has no convulsant effects, favors the appearance of associative responses to the same extent as chloralose does (Fig. 4).
v hL aC
?r-?$f at
20msec
ah
Fig. 4 Cur mesfhetized wirh viudril. Responses recorded at the level of the anterior marginal gyrus ( M A )
and of the posterior suprasylvian gyrus (SSp); stimulation of the two anterior limbs, contra- and homolateral (ac, ah). The results are to be compared with Fig. 1.
(b) Experiments done on the chronic animal have dispelled any remaining doubts, although certain procedural precautions are necessary. The implantation of the electrodes is performed under chloralose, in order to place them exactly at the point of maximal response. Once the animal has recovered, it is observed under unrestrained conditions and in a state of wakefulness. The evoked potentials obtained during the aninzal’s periods of inaftention were of large amplitude in the anterior marginal gyrus (Fig. 5), and in the regions of convergence of the suprasylvian area and of S I. The meagre success of previous attempts to obtain such responses in the nonanesthetized preparation, always maintained under stressful conditions, is certainly explained by the fact that under such conditions it is rare to obtain somnolence OJ simple inattention. Thus, chloralose has been only a convenient means for revealing integrated activities, the physiological reality of which can no longer be doubted. 3. The survey, at the telencephalic level, of regions ofconvergence (local or project* 21- hydroxypregnane-3,20-dionesodium hemisuccinate (kindly provided by Laboratories Pfizer Ltd).
119
THALAMIC INTEGRATIONS
1
1
2
v
V 5pmSec
7
9
10
8
12
v [ 200pv
Fig. 5 Uwestrained, awake cat. Implanted cortical electrodes, stimulation of the contralateral superficial radial nerve by permanently placed electrodes. Upper trace, primary somatic cortex, low parietal region, relatively silent. Lower trace, anterior region of the marginal gyrus (MA). The electrode was previously placed, under chloralose, at the point at which a maximal convergent potential could be observed. 1 to 5, inattentive animal; 6,7,8,9, 10, note the abolition of the anterior marginal response during the period of alertness which follows a call; 11, 12, the animal is again inattentive.
ed), and the study of their properties, are certainly not yet complete, but we realize that each region possesses certain characteristics which distinguish it from other regions.* This great complexity induced us to limit ourselves here to a consideration of somesthetic integration about which more information is available on the various pathways and relays, from the receptors to the cortex. Some of the integrations which involve the visual or auditory pathways will be considered in this symposium by other authors. We will not discuss the somatic inputs which are distributed to the primary auditory and visual areas. On the other hand, we lack results about the trigeminal inflow and will only briefly allude to the heterotopic convergence which is revealed by the evoked potentials of area S II. In short, rather than attempt a detailed and complete survey of the aspects of integration at the telencephalic level, we shall be concerned with some favorable examples (convergent systems, predominantly somatic, in the suprasylvian and anterior marginal gyrus, as well as the area S I) in order to analyze the modes of realization of this integration, and the relays and pathways implicated. We believe that the principles of this analysis can be applied to the interpretation of integrative mechanisms belonging to other systems. * For instance, as Bruner and Ruser (1961) have recently shown, there are two rather distinct types of non-primary visual projections to the suprasylvian associative cortex of the cat: one corresponding to the two foci of non-somatotopic somesthetic projection previously described by one of us (Albe-Fessard and Rouged 1958); the other to two neighboring foci, more specifically visual, and not subject to occlusion by reticular stimulation, as is the projection of the formei type. Ki’fcwmws p 145-148
120
D. ALBE-FESSARD, A. FESSARD
4. According to previous conceptions, the sensory messages which ascend to and converge in the cortical zones of association would have their origins mainly in the primary projection areas. Yet we now know that the responses to peripheral stimulation evoked in the convergent regions of the cerebral cortex, or in the basal ganglia, are not, in general, necessarily dependent upon the integrity of the primary areas. This major fact was first demonstrated in the somesthetic systems by the finding that the ablation of area S I and the motor area in the cat does not suppress, and may even increase, the associative responses (Albe-Fessard and Rougeul 1955, 1958). Ablation of the homolateral visual and auditory areas does not prevent the activation by light or sound of the “associative points” (Buser and Borenstein 1956) and likewise, an extensive cortical ablation does not suppress the multisensory responses of the caudate nucleus (Albe-Fessard, Oswaldo-Cruz and Rocha-Miranda 1960).* These responses of the convergent type are nevertheless subject to a certain control from primary areas, as is demonstrated, for example, by the administration of strychnine or by cooling of these areas (Amassian 1954; Buser and Borenstein 1956), or by responses or occlusion observed in the associative areas by stimulation of a primary area. However, the latency of these responses is too long for a direct transcortical connection. A more reasonable hypothesis is that of a cortical-subcortical-cortical pathway (Albe-Fessard and Gillett 1958). In any event, one is obliged to include the thalamus in these systems of integrative projections. We shall recall first (B) which regions of the thalamus are capable of functioning as relays for these projections and show that these regions are already centers of convergence, direct or projected. Finally (C) we shall consider their efferent projections towards the cortex and basal ganglia.
B. ASPECTS OF INTEGRATION AT THE DIENCEPHALIC LEVEL ANATOMICAL AND ELECTROPHYSIOLOGICAL EVIDENCE
1. The classical anatomical results lead us to conceive of two systems of integrative structures at the thalamic level, the associative nuclei, on the one hand, thus termed because they project to the association areas, after having received their afferents from other nuclei of the thalamus, and, on the other hand, the nuclei of the difSuse projection (or non-specific) system, comprising the intralaminar nuclei and certain midline nuclei. The afferents and efferents of the aspecific nuclei have not yet been determined in a precise way by anatomical methods, but it is commonly admitted that their situation destines them to integrative functions. The same is true for the association nuclei (after the terminology of Walker 1938), sometimes called “integrative nuclei” (terminology of Hassler 1955) or “elaborative systems” (Jasper and Ajmone Marsan 1952). In fact, we shall see that the classical anatomical divisions give us rather incomplete information about the actual distribution of the integrative properties of the thalamus
* At least one exception exists: the somesthetic responses o f the amygdala and the piriform lobe, which are of the convergent type, are suppressed permanently by ablation of S 11 or reversibly by cooling of this area (according to the recent work of Wendt and Albe-Fessard 1961).
THALAMIC INTEGRATIONS
121
and that a systematic electrophysiological exploration results in a more highly differentiated functional topography. If one is content with experiments in the field of somatic sensation, and if one then takes the absence of somatotopy as a sign of the capacity for integration (AlbeFessard and Rougeul 1958; Kruger and Albe-Fessard 1960), one can list for the cat the principal nuclei in which heterotopic convergence is observed. The ventro-basal complex retains a somatotopic organization even under chloralose. No responses are obtained from the nuclei of the anterior group, those of the lateral group, the pulvinar and the n. medialis dorsalis. On the other hand the centre median-parafascicular complex, the regions of the nucleus ventralis lateralis (VL) near to the VB complex, the region of the central commissural system which comprises the nucleus centralis medialis, and certain portions of the nucleus reticularis, have response characteristics which recall very closely those of the cortical zones whose responses are of the convergent type. These are, as we shall see later, acceptable candidates for the “integrative” relays for the associative cortical projections. We shall exclude from this report the responses of the system designated the “posterior group”, studied in detail by Poggio and Mountcastle (1960), which, as the stimulation experiments of Knighton (1950)” suggest, seems to serve as the relay for area S IT. We have also observed the same responses in the chloralosed cat; and analogous multisensory responses have been recorded in the zona incerta, the subthalamic nucleus and the red nucleus. Although the current procedure for demonstrating that one structure projects upon another is to stimulate electrically the first and to record the response of the second, we prefer primarily to draw our evidence from a comparison of the properties of a tested thalamic relay with those of the associative regions of the cortex when the two structures are activated by means of peripheral stimulation. These are more natural conditions of activation. Furthermore, direct electrical stimulation of the nuclei would introduce here the complication of the recruiting response, which is of a different nature. 2. Properties common to thalamic nuclei with convergent somatic afferents and association areas of the cortex (a) Latencies The response latencies of these nuclei to cutaneous electric stimulation of a fixed point are always of the same order, e.g., in the experiments of Kruger and AlbeFessard (1960) they are 9.9 to 12.8 msec for stimulation of the extremity of the contralateral or homolateral anterior limb (somewhat longer, 13 - 13.7 msec, for the n. centralis medialis). These values are well above those found for the primary relay (average 4.8 msec). However, they are always less than the response latencies of the associative cortex, these being 13-16 msec for the suprasylvian gyrus under * Actually Knighton concluded that the thalamic source of S I1 is exclusively the posterior segment of the VB complex, but this interpretation is not accepted by Rose and Woolsey (1958), Poggio and Mountcastle (1960). References g. 145-148
122
D. ALDE-FESSARD. A . FESSARD
identical conditions (16.8-18.3 msec for the caudate nucleus, connected with nucleus centralis medialis). Obviously it was necessary first to establish this fact before considering these nuclei as possible relays. (b) Multiserzsory convergence
Most of these nuclei were initially studied for their responses to somatic stimulation although, like the cortical associative areas, they also respond to visual and auditory stimulation. Their exploration by means of microelectrodes (Fig. 6) shows CM
.
VL
183t2.0
ac
ah
ah
18.521.3
CM
CM
T
200msec
Vis.
Som.
Fig. 6 Car onesthetized wirh chloralose. Above: convergent responses observed at the level of a cellular
element in the n. CM (left) and in the n. VL (right). Below: double response of a cell in the n. CM to a brief, natural visual stimulus (Vis) and to a brief cutaneous stimulation (Som).
that the multisensory or heterotopic convergence takes place at the level of the nerve cell with a predominance of convergence of afferent impulses of somatic origin. Of 221 cells studied in the centre median (Albe-Fessard and Kruger 1962) only four were somatotopic.
( c )Theparallel eflects of anesthetics These effects are the same on the responses of the nonspecific thalamic nuclei and on those of the convergent regions of the telencephalon (cortex and striatum) ;they are enhanced by chloralose and viadril (Fig. 7) and depressed by barbiturates and
ir”CM Ev
OC
q-L+ oc
50msec
oc
Fig. 7 Responses of the centre median tocontralateral stimulation (forelimb) in the awake animal prepared under ether, immediately before the injection of viadril (Ev), immediately after the injection of 60 nigikg of it and one hour after this injection (Ev).
123
THALAMIC INTEGRATIONS
ether. The recovery cycles established for thalamic and telencephalic convergent structures with chloralose are superimposable (Fig. 8) and differ from those of the specific relays by their longer duration.
01
I
100
200
300
400
500
600
790
m a c e r BOO
Fig. 8 Cuts under chloralose anesthesin. Recovery curves for responses (in percentages of amplitude) at the
level of the anterior marginal cortex (CoMA) and the primary somesthetic cortex (CoSI, the two phases, positive and negative, are treated separately), of the centre median (CM), and of the primary relay (VPL). Note the identical development of the curves CoMA and CM, in contrast to the much more rapid development in the VPL and for the positive phase of the response a t S I. Note also the different development of its negative phase, which approaches that of the slow recovery curves characterizing integrative activities.
( d ) The parallel effects of alertrzess in the unrestrained animal We have seen above that large cortical responses, indicating the effects of convergence, can be recorded in the unrestrained waking animal during phases of non-alertness (inattention). Fig. 9 shows that the thalamic responses with convergent features, recorded in CM and VL, have the same correlation with the state of alertness, monitored by a continuous tracing from the suprasylvian gyrus. Fig. 9 also shows that the response of the specific thalamic relay does not undergo such changes. The hypothesis is thus justified that the amplitude fluctuations of the evoked responses of the associative cortex according to the state of alertness of the animal may be only reflections of amplitude variations in the convergent nuclei of the thalamus. (e)Parallels in theformation of conditioned responses The parallel continues up to the complex operations involved in conditioning. Some experiments along this line have been carried out at our laboratory by G. Lelord with chronically implanted electrodes on cats which were conditioned to sound-electric shock (painless) combinations. 1 t is known that evoked cortical potentials can appear, or can increase in amplitude, in response to the conditioned stimulus alone, after a series of paired stimuli. These electro-cortical conditioned responses provide the striking evidence that associative processes have occurred. Several authors Rrferenres P 145-1 48
124
D. ALBE-FESSARD, A. FESSARD
Fig. 9 Awake cats. Recordings obtained in three animals with chronically implanted electrodes, one in the centre median (CM), one a t the boundary between the n. VPL and the n. VL (VL), one in the VPL itself. Each of these animals had in addition two cortical electrodes. The discontinuous traces correspond to the responses of the deepstructures. Thevertical traces (to be read from below upwards) are from the suprasylvian cortex and make apparent the transition from an inattentive state to a state of alertness. Note that the response of then. C M and the second response (with convergent properties) of the n. VL disappear or are strongly reduced during the state of alertness provoked by a call to a previously unalerted animal. By contrast the first response at the level of the n. VL and the response of the n. VPL seem rather to undergo a slight increase after these calls.
have already shown that the thalamus is implicated in the formation of new connections (Yoshii 1957; Morrell 1960). In connection with our studies of the properties of the various thalamic nuclei, Lelord has shown that whereas the specific relay nucleus (VPL) never acquires the ability to respond to the sound stimulus, the neighboring region in the nucleus ventralis lateralis (VL) is easily conditioned (Fig. 10). Here again, therefore, the cortical indication of “circuit closure” could be only the reflection of convergence at a lower level. The site for this integration is not necessarily thalamic, but may equally well be situated at lower, mesencephalic or bulbar levels (see later).
(f) The parallel effects of posterior column section (see below, 3a) 3. The somatic sensory afferents of thalaniic structures of convergence ( a ) Limiting ourselves to the ascending pathways of the spinal cord, we first wish to emphasize a point of major importance which has been amply confirmed by expe-
THALAMIC INTEGRATIONS
125
Fig. 10 Unrestrained awake animal. Electrodes implanted: at the level of the convergence area of the primary somatic cortex (SIC), upper trace for all records; at the level of the n. VL in the immediate vicinity of the VPL, lower trace for all records at the left; at the level of the n. VPL (region of the anterior limb), lower trace for all records at the right. The animal is conditioned to two paired, heterosensory shocks: 1, sound; 2, a non-painful electric shock applied to the superficial radial nerve by permanently placed electrodes. Lef: the two traces of the upper record: first association. The two following traces: in the course of the conditioning. The two lower traces: the conditioning is established, only the sound is presented. Note the occurrence after conditioning ofanample response to sound in the n. VL. Right: in the same animal, after extinction of the conditioning, the thalamic electrode has been lowered into the VPL. The successive traces correspond here also to the first association, to associations in the course of conditioning, and to the application of sound alone, once conditioning has been established. Note that, after conditioning, no response to sound alone appears at the n. VPL level, although this response is clearly visible in the cortical record.
riments in our laboratory: the agerent impulses going to the primary thalamic relay nucleus, and those which reach the structures of convergence of the thalamus, follow separate pathways from the spinal cords upwards. Total transsection of tlie posterior columns does not modify the response of the centre median-parafascicular complex whereas the responses of the ventro-basal complex are considerably reduced (Fig. 11). An experiment analogous to the preceding one has shown that, at the level of the associative cortex, the responses are maintained despite the total disappearance of the primary relay response (Fig. 11). The response of the caudate nucleus is likewise not modified. From this, one can deduce that the integrative centers must receive their somatic inflow essentially by way of the antero-lateral pathways of the cord (commonly termed the spino-thalamic tracts). However, stimulation of the posterior columns provokes responses in the convergence structures. It therefore became necessary to examine this fact more closely. Mallart (1961) was able to show that the response to stimulation of the transsected dorsal columns, raised and placed on electrodes, was not suppressed by another complete transection approximately five segments more rostral. On the other hand, References P. 145-148
126
D . ALBE-FESSARD, A. FESSARD
it was suppressed by a lateral cut of the spinal cord. Hence the fibers found in the dorsal columns which lead to the convergence structures do not ascend in these columns for more than a few spinal segments, at which point they join the anterolateral columns. These fibers were, by the way, described long ago by anatomists (Ram6n y Cajall909; Karplus and Kreidl1914).
Sect. co1.d
-
s.
P=
Fig. I 1 Cut onesthetized with chlorulose. Responses observed in the thalamus at the level of the n. VPL, and then the centrum medianuni and at the cortical level (anterior marginal gyrus), before (records above) and after (four records below) section of the dorsal columns (Sect. col. dors.). Note that the response of the n. VPL has disappeared, whereas the responses of the cortex are unchanged and the response of the n. C M (presented here after the section only) is normal; the middle and lower records were taken 10 niin and 30 min after the section.
Such a separation of the path of convergence and the lemniscal path renders doubtful the existence of abundant collaterals passing from the latter to the multisensory and heterotopic fields of the brain stem (reticular formation) and diencephalon. Indeed, for limb and body afferents (the face being disregarded) this notion of lemniscal collaterals does not seem to have received anatomical support either. The fibers of the dorsal columns do not give off collaterals to the spinal reticular formation (Glees and Soler 1951) or the brain stem (Torvik 1956). The fibers of the medial lemnicus also seem to lack them (Matzke 1951, Marchi method in the cat; Bowsher 1958, silver method in the macaque; see also Scheibel and Scheibel 1958, p.33, who used the Golgi method). With regard to the face, two distinct afferent systems may be also reasonably assumed. i t is, however, difficult to establish evidence for their morphological separation, because of a bifurcation of most of the primary trigeminal fibers, which results in a supply both to the principal nucleus and to the spinal nucleus of the trigeminal nerve. A special investigation of this has yet to be made. It is known that in neurosurgery the spinal fibers are severed to reduce trigeminal pain, but the presence of specifically tactile fibers has also been demonstrated in this bundle (Kruger, Siminoff and Witkovsky 1961). (b) Against the view that there is an absolute separation of the pathways one may cite the anatomical observation that a certain number of fibers which originate in
THALAMIC INTEGRATIONS
127
the anterolateral tracts join the lemniscal system without forming synapses in the gracilis and cuneatus nuclei. They project somatotopically on to the specific relay. These fibers, together with the cells in the posterior horns of the spinal cord from which they arise, are generally crossed, but included among them there is a small proportion of direct, ipsilateral fibers (White and Sweet 1955) and also some ipsilateral fibers arising from a redecussation at the bulbar level (Mehler et al. 1960; Bowsher 1961) and in the posterior commissure (Quensel 1898; Bowsher 1957). This system has been called the rzeospinothalamic system (see Mehler 1957), because it is much reduced in marsupials (Clezy et a/. 1961), and is, even in cats, poorly developed, though it develops rapidly in primates. However, even in man, in whom it attains its fullest development, its importance is limited in comparison with that of the medial lemniscal fibers which relay the fibers passing up via the dorsal columns (about 1500 at pontine level, according to Glees and Bailey 1951, in contrast to the large number of lemniscal fibers). Undoubtedly, this is the pathway along which somatotopic ipsilateral impulses fire the ventro-basal complex (Gaze and Gordon 1954, 1955; Poggio and Mountcastle 1960; Perl and Whitlock 1961; Whitlock and Perl 1961). The truly independent afferent system which supplies the convergence nuclei of the thalamus must be clearly distinguished from the one just described. In contradistinction it is now termed the paleospinothalamic system (Mehler 1957). It increases only slightly in the ascending scale of mammals. It thus seems to represent a more primitive projection of somatic sensibility. The fibers of this system synapse in the grey matter of the spinal cord, ascend as crossed and uncrossed fibers and reach the intralaminar nuclei: parafascicularis, centralis lateralis (Mehler 1957; Mehler et al. 1960) and the nucleus reticularis (Getz 1952; Bowsher 1957, 1961). Only a portion of the fibers of the antero-lateral columns directly reach the thalamus, and the majority relay below it. The shortest ones are spino-spinal, but the majority are spino-reticular : they terminate at different levels of the ascending reticular system. The localization of these levels is a task undertaken only fairly recently by anatomists, assisted by electrophysiologists. The latter must also determine whether the relays found at these points are likewise centers of convergence and whether they project to the non-specific thalamic nuclei. If they do, at least some of the integrative action observed at the thalamic level would be only a reflection of properties belonging to a structure at a lower level (projected convergence). The question is of importance for our subject. To answer it, we already have some data which seem sufficiently well established to provide a general idea of the organization of this system (see below, e). Particularly interesting is the microphysiological study made by Wall (1961) of two groups of cells: one in the nucleus gracilis and the other in the dorsal horn of the lumbar region of the spinal cord. Wall showed that the cells at these levels already act as points of convergence for fibers of different types: those in n. gracilis respond only to the largest fibers in the cutaneous nerves; the dorsal horn cells respond as though all types of cutaneous fibers terminated on them. (c) The anatomists have stated that a large part of the fibers with reticular terminaRrfcrencrr n 145-148
128
D. ALBE-FESSARD, A. FESSARD
tions stop at the bulbar level. There they synapse with the cells of the nucleus reticularis gigantocellularis (see Olzewski 1954; and Rossi and Brodal 1957). A smaller number reach the reticular formation of the pons and others finally reach the mesencephalic reticular formation. The first electrophysiological explorations of the brain stem reticular regions with microelectrodes (Amassian and De Vito 1954; Scheibel et al. 1955) immediately revealed the non-specificity of their neurons, the majority responding to afferent impulses of various origins; but it was particularly important to explore the nucleus gigantocellularis of the reticular formation. With Mallart and Bowsher, we have recently done this, using macrophysiological and microphysiological techniques. It then became quite apparent (Fig. 12) that the gigantocellularis is a nucleus which
4{f ,/
ad
Fig. 12 Cut anesthetized with chloralose. Responses recorded at the level of the nucleus gigantocellularis of the bulbar reticular formation (Ret. bulb.) and in the centre median (CM) to stimulation of the four limbs (Bowsher and Mallart, unpublished).
possesses all the convergent properties described by us for the associative regions of the.cortex. It responds to the stimulation of all parts of the body, and probably also to visual and auditory stimulation, and it exhibits the fatigue characteristics which all Convergent structures show. Moreover, it projects to the thalamus.
?
St.Ret.bulb.+ a C
?
St. Ret bulb.+
ah
Fig. 13 Cat anesthetized with chloralose. Recording in the centre median. Above: stimulation of the contraand homolateral anterior limbs. Below: the same, but preceded by single shock stimulation (0.5 msec, 4 V) of the bulbar reticular formation.
THALAMIC INTEGRATIONS
129
Nauta and Kuypers (1958) have described fibers originating in this bulbar region and terminating in the centre median-parafascicular complex. Stimulation of this nucleus results in marked responses in the centre median (Fig. 13). Lastly, localized cooling of the nucleus selectively suppresses the evoked potentials in the centre median without altering those of VPL. It thus seems to be well established that the bulbar reticular formation constitutes the principal relay of the extra-lemniscal path way leading to the centre median-parafascicular complex. Other relays undoubtedly exist, particularly in the mesencephalic reticular formation. Here our evidence is not so precise. The relay zones have not been adequately identified and the mesencephalo-thalamic projection fibers have not been definitely established anatomically. It is as yet uncertain whether they lie in the reticulo-thalamic or the tegmento-thalamic bundles or, as in Nauta and Kuypers’ experiments (19581, in a diffuse projection “blending inseparably with the massive reticular projections in Forel’s tractus fasciculorum”. Preliminary experiments using stimulation procedures suggest that the nucleus centralis medialis and the convergent zones of VL receive their afferent impulses from the mesencephalic reticular formation. ( d ) It is obvious that thalamic integrations are not limited to those due to inputs of sensory origin. Every convergent nucleus receives information from other parts of both the brain and the thalamus itself and integrates it with that derived from the sensory input. In particular there are important relations from cortex to centre median. These can be demonstrated by cortical stimulation; in the cat, the lowest threshold zones are the motor areas, S I1 and the inferior temporal region (AlbeFessard and Gillett 1961). What are the pathways for this cortical control, and what part does it play? A prolonged discussion on this subject would take us beyond the scope of this report. It should, however, be pointed out that: (i) the distribution of the most active cortical zones resembles that of the zones linked with the mesencephalic reticular formation, in which, because of this linkage, inhibitory effects are exerted on the ascending reticular activating system (Hugelin and Bonvallet 1958); ( i i ) the response to cortical stimulation may be followed by inhibitory effects in the centre median-parafascicular complex (Albe-Fessard and Gillett 1961); (iii) anything that reduces cortical activity contributes to the enhancement of the responses in this nucleus. This last point merits special consideration and other evidence should be considered in addition to the data already presented (effects of chloralose and effects of inattention; see above). Massion and Meulders (1961) were able to show that the evoked potentials in the centre median became as large as under chloralose after extensive ablation of the cerebral cortex and part of the corpus striatum. The same effect is obtained by cooling of the cortex (Massion and Meulders 1961) or by elimination of the reticular formation (“cerveau isolp” preparation). Hence, a predominantly inhibitory control is exerted by the telencephalon on certain thalamic nuclei. Most probably this is not due to a direct connection, but to a retroactive chain, the elements of which remain to be determined. Reference3 p , 145-148
130
D. ALBE-FESSARD, A. FESSARD
The exact intrathalamic organization of the system of convergent nuclei, together with its interconnections remains, likewise, to be determined. Some anatomists, like Mehler, consider the centre median as an integrating nucleus of the second order, at least in its parvo-cellular part (which is important only in the higher mammals). Others (McLardy 1951; Johnson 1961) describe it especially in terms of its distributive connections which bring it into relationship with the associative and non-sensory relay nuclei. (e) Fiber and receptor types. None of the thalamic structures receive messages conveyed by group I fibers, which stem from spindle and tendon receptors. This observation made by Mountcastle et al. (1952) for the primary relays has also been made for the centre median by Mallart (1961). By comparing the action potentials of peripheral nerves with responses in the CM, the latter author was able to show that group I1 fibers from muscle spindles are, likewise, not implicated in the afferents to the CM. On the other hand the rapid ixfibers of cutaneous origin (sometimes called cutaneous group 11) supply the C M as well as the VPL, although the CM response requires a certain degree of convergence. As to the fine fibers of group 111 from muscles or those of cutaneous origin (O), they supply jointly the CM and the VPL (Fig. 14).
5msec
Fig. 14 Animal anestherized wirh cklorulose. Lefr: Responses recorded from the centre median (CM) and the sural nerve (N) to stimulation of the same nerve (of purely cutaneous origin). Note the two peaks of the nerve response corresponding to fiber groups a and 8 (also called cutaneous groups 11 and Ill). Right: The responses of the C M are still present, although the fibers of t h ea group have been blocked by a train of shocks (method of Laporte and Bessou 1959).Different time base for the n. CM (20 msec) and the nerve ( 5 msec) (after Mallart, unpublished).
With regard to the nature of the receptors implicated in the activation of neurons belonging to the different thalamic nuclei and, for each receptor type, the corresponding receptor field, one must first mention the systematic investigations of Poggio and Mountcastle (1960) and, more recently, those of Per1 and Whitlock (1961). These deal essentially with the properties of the ventro-basal complex and the posterior group of nuclei. We have, for our part, attempted to distinguish the properties of the centre-median-parafascicular system and those of the ventro-lateral nucleus from those of the specific relays. To summarise the results, one may say that the majority of the cells of the VB complex are activated by very localized cutaneous stimuli produced by hair bending, light touch or joint movement. On the other hand, the CM and V L neurons respond to stimulation of a n y part of the body; the effective stimulations are pricking and sharp pressure applied to cutaneous or subcutaneous tissue, in short, stimuli which frequently deserve the designation “noxious”.
THALAMIC INTEGRATIONS
131
The unit responses of the ventro-basal complex show rapid or slow adaptation which depend on either hair bending or joint movement respectively. In the convergent nuclei adaptation is always rapid ; moreover a single brief stimulus frequently gives rise to a double response which consists of two discharges separated by a long silent interval (See Fig. 6 and Albe-Fessard and Kruger 1962). Noteworthy is the fact that every integrating neuron is characterized by a definite “afferent pattern”, one particular type of stimulus being, as a rule, more effective. Per1 and Whitlock (1961) made the same observation during their study of the receptivity of neurons of the primary nucleus (ventro-basal and posterior group nuclei) activated by spinothalamic afferents. C. THE PROJECTION OF THE INTEGRATED ACTIVITIES OF THE THALAMUS
TOWARDS THE TELENCEPHALON
A considerable number of anatomical and electrophysiological studies have been devoted to the problem of the efferent connections of the non-specific thalamic nuclei. There is no question of discussing these here, in particular because a fundamental difference in methods separates the electrophysiological work of our group from that of most of the earlier authors. Since the classical work of Dempsey and Morison (1943), these authors have almost always used repetitive local stimulation of the non-specific nuclei, taking the so-called “recruiting response” of the cortex as their index; we, on the other hand, have attempted to activate these nuclei by the normal play of their afferent impulses in the hope of obtaining results in closer agreement with physiological conditions. Even when, in our investigations, analytical requirements demanded direct stimulation of a nucleus, this stimulation was always given by isolated shocks which give rise to evoked potentials, the characteristics of which differ markedly from the characteristics of recruiting responses. In fact repetitive stimulation of a convergent structure, such as the centre median, results at the same time in an evoked potential and a series of responses of the recruiting type in associative cortex. However, it is not difficult to realize how contrasting are the characteristics of the two types of responses in their latency, fatigability, sign and topographic distribution. The initial evoked potential of short latency, with a primary positive phase and rapid fatigability, is alone comparable, in terms of
I
-
0.5 sec Fig. 15
Animal anestherized with nenihutul. Repetitive stimulation of the n. CM, recording at the level of the
suprasylvian gyrus. The first shock gives in this instance under nembutal an ample positive response similar to those observed under chloralose, but after a longer latency; the response fatigues immediately and is replaced by recruiting responses, Rcfcrenrra D 147-148
132
D. ALBE-FESSARD, A. FESSARD
its characteristics and distribution of zones of maximum amplitude to the evoked potentials produced by peripheral stimulation (Fig. 15). Despite the numerous anatomical studies devoted to the problem of the non-specific thalamic efferents (see McLardy 1951; Nauta and Whitlock 1954; Hassler 1955; Nashold et al. 1955; and the review by Macchi and Arduini 1957) a great many uncertainties exist and the controversies remain lively. We cannot discuss them here, but only hope that the few electrophysiological observations cited below will help to clarify this difficult probIem. We shall report in succession our results: (1) on thalamo-striatal connections; (2) on projections to the suprasylvian and anterior marginal association areas; and (3) on integrative projections to the primary area S I. We shall be confronted at this level by a special case, that of the important problem of the integrative relationships between the somatotopic projections of the discriminative somatic sensitivity and the non-discriminative projections of the extra-lemniscal system. Let us recall here that Poggio and Mountcastle (1960) have discussed the possibility of a projection of neurons from the posterior group to S 11. Certain cells of the VPL and VPM are also said to project to S I1 (Macchi et al. 1959), a cortical area which Wendt and Albe-Fessard have recently (1961) shown to be a convergent relay for the piriform cortex and the amygdala. 1. Projections to the striatum
Many anatomists have emphasized the importance of the thalamo-striatal efferents especially those of centromedian origin (McLardy 1948; Drooglever Fortuyn and Stefens 1951; Hovde and Mettler 1953; Hassler 1955; Johnson 1961). The connections are said to be direct with the putamen and caudate nucleus. Our electrophysiological observations confirm the projection from CM to putamen with a very short latency; however the direct connection with the caudate nucleus does not arise in the CM but in the nucleus centralis medialis, in agreement with theanatomicalfindings of Cowan and Powell (1955) and Powell and Cowan (1956). The path between the CM and caudate nucleus is undoubtedly poly-synaptic (latencies of 8-10 msec). 2. Projections to the associative areas Local stimulation of the centre median with a single shock leads, after a short latency (2-4 msec) to bilateral responses in the same cortical regions as those which, after peripheral stimulation, develop an evoked association potential (Fig. 16) that is to say, an evoked potential which disappears after coagulation of the centre median. In recent experiments we were able to show that localized cooling of the CM by means of a special probe results in a selective and reversible disappearance of the associative responses in the posterior suprasylvian and anterior marginal zones without noticeable alteration in the response of S I (Fig. 17). The motor area and gyrus proreus are probably inactivated under these conditions, but the experiment remains to be done. We may conclude that at the level of the centre median-parafascicular complex (CM), either fibers of passage or a true relay assure an already integrated projection of information towards the cortical association areas which we have described.
THALAMIC INTEGRATIONS
133
Fig. 16 Cut anesthetized with chlorulose. Short latency cortical responses to stimulation of the centre median (0.5 msec, 6 V) as displayed by the various convergent territories of the cortex (after Albe-Fessard and Rouged 1958).
Fig. 17 Cut anesthetized with chlorulose. An illustration designed to show the effects of a reversible block of the centre median on the responses recorded at the level of the primary somatic area (SI) and the anterior marginal gyrus (MA) to stimulation of the extremities of the contra- and homolateral anterior limbs. A probe with a diameter of 1 mm provides, by the expansion of propane, a local cooling of the tissue surrounding the probe tip (Dondey et ul. 1962). It has been placed in the CM (A 7.5, L 3, H 0.5). The path for the approach of the probe is inclined and the brain is entered by the cerebellum, in order to avoid any cortical lesion. At the arrow (1) the cooling is applied and this ends after 140 sec. Note the disappearance of the marginal convergent response, and the persistence of the primary lemniscal response, but also of the second phase of the contralateral potential in S I, and of the principal part of the potential evoked by homolateral stimulation at the same point (see next figure).
+
References P. 143-148
I34
D. ALULI-FCSSAKD, A. FIlSSARD
The most reasonable hypothesis, in our opinion, would be that of a thalamic relay, the CM, which transmits to the cortex ascending messages already integrated (to a certain degree) at the bulbar level. However, we are here opposed by the opinion of the anatomists, who do not recognize a direct ascending linkage between the centre median and the cortex (however, see Nashold et al. 1955). It is true that the latency, although it is brief, is compatible with at least one intermediary relay. The contradictory anatomical findings do not allow us to localize definitely this relay. The nucleus reticularis and the nucleus ventralis anterior have been implicated, but these may contain fibers of passage. The existence of a direct anatomical connection with the corpus striatum (Drooglever Fortuyn and Stefens 1951) and the electrophysiological demonstration of a connection between the CM and the putamen make it necessary to consider whether a striatal detour is involved. 3. Iniegrativeprojection to S I We reported in the first section (A) that a convergent focus had recently been de-
Fig. 18 A and B (see next page) Cut mesfhetized with chloruloAe. Systematic mapping of activity evoked in area S I by stimulation of the contralateral (A) and homolateral forelimb (B). The mapped area is delimited on a diagram of the cortex seen in the lower left. The territories corresponding to the principal types of response are designated by the letters 'i, {j, >!, h, t . Note that as in y (which corresponds to the anterior marginal responses) one observes also in /land 0 bilateral responses which are those of the convergent type o f S 1 (Jankowska and Albe-Fessard, unpublished).
THALAMIC INTEGRATIONS
135
I
wI
I
y I
I
I
P
Fig. 18 B (legend p. 134)
monstrated at the level of S I (Jankowska and Albe-Fessard 1961). This property has complicated the organization of this area without, however, destroying its essentially somatotopic character. We will try to determine whether this convergence is local or is merely projected. If it is projected, we must find out which thalamic nucleus it arises from and also its possible relations with the primary somatotopic projection. ( a ) Projections of the lemniscal arid extra-lemniscal systems at the level o j S I. We began with the notion that the projections of the lemniscal system were entirely contralateral, while those of the extra-Iemniscal system were bilateral. First, we recorded, from the same point, the potentials evoked by stimulation of either the contralateral or the homolateral anterior limb; we then established and compared the distribution of activities corresponding to the two systems for a given peripheral region. We did not limit ourselves, as the classical charts do, to the points of maximal activity, but we made as complete an inventory as possible of the potentials evoked by stimulation of a given region (Fig. 18A and B). Apart from the classical evoked activity with its initial positive phase, its strict contra-laterality and its areal limitation, we observed under these conditions, widely References p . 145-146
136
D. ALBE-FESSARD, A. FESSARD
distributed projected activities which were bilateral in origin and overlapped with the first. Analogous results were obtained by stimulation of the posterior limbs, except for the shifted focus of primary activity. The field for convergent activities of heterotopic origin is stable and maximal at a point approximately 6 mm medial and at the same anterior level as the maximal somatotopic point of the anterior limb. A complete cartography would show a superimposition on this field of the somatotopic projections of the various body regions the localisation of which varies. Because of the intermingling of the primary and convergence zones, it is difficult to recognize the separate components of the two projection types in a potential evoked by contralateral peripheral stimulation. In order to separate the components of the complex which the evoked “primary” potential is in reality, we have had recourse to the method of heterotopic double shocks. Stimulation of the contralateral anterior limb is preceded by an identical but homolateral stimulation; the homolateral stimulation can then occlude, by its prior presence in the projection system, that phase of the “primary” evoked potential which belongs to the convergent system. Fig. 19 shows, first of all, as might be expected, a complete occlusion of the evoked potential in the anterior marginal gyrus (here called ‘‘y point”) which represents pure, non-primary activity. Occlusion is only partial in the zones in which the potential is due to activation of the primary projection. It is striking that the occlusion involves the negative phase of the evoked potential, which thus seems to have an origin
Fig. 19 Cat anesthetized with chloralose. At the points a, /?,y and 6, indicated in Fig. 18, experiments have been performed which result in an occlusion of a contralateral response preceded by a homolateral stimulation. In the left column, the response evoked by the contralateral stimulus is presented alone; in the right column, the same stimulation is preceded by a homolateral stimulation. Note that in these instances the elements which disappear are the negative phase in 01, a second diphasic positivenegative phase in 6 and the entire response in /? and y.
THALAMIC INTEGRATIONS
137
direrent from that of thefirst phase. It seems to indicate that it depends only on the convergentpathway,* whereas the primary origin of the initial, surface-positive phase is again confirmed. This result agrees with the long established fact (Adrian 1941) that, at the primary cortical level, the recovery time course of the two phases of the evoked potential is quite different. The negative phase recovers more slowly, its duration corresponding to the recovery of a response of the anterior marginal gyrus (Fig. 8). (b) Do the afferent activities of the two systems converge on the same cortical neurons or do they constitute a mosaic of diferent elements? Information obtained by macrophysiological exploration suggests a convergence of the two systems towards common neurons, because a sufficient reduction of the interval between the two shocks (homofollowed by contralateral) affects, not only the negative phase, but also the positive phase of the primary type potential (Fig. 20). It was, however, advisable to make a microphysiological exploration to determine whether it is an occlusion rather than an inhibition. This research was undertaken with the collaboration of Dumont-TyC and Jankowska (Albe-Fessard et al. 1961). Microelectrode penetrations were made as perpendicular to the cortical surface as possible at a point of maximum amplitude for the primary evoked potential of
ac
ahtac
'
ac
ah+ac
Fig. 20 Cat anesthetized with chloralose. Responses recorded in the area S I, stimulation of the contralateral forelimb preceded in the three last lines by a homolateral stimulation. Note the reduction of the positive phase of the contralateralevoked potential when a small interval separates the two shocks.
* This property belongs essentially to the response evoked in a (see Fig. 18). In a slightly posterior region, in E (Fig. 18), another negative component is not occluded. Its origin has not been determined so far. References P. 145-148
138
D. ALBE-FESSARD, A. FESSARD
the anterior or posterior limb (distal parts). Of 56 cells studied in detail for the origin of their afferent projections (mainly in the forelimb region), only two were purely somatotopic, i.e., solely dependent upon the lemniscal system. Of 54 other cells which responded to stimulation of all four limbs, 23 were activated by the extralemniscal system alone; the remaining 31 cells responded both to this and to lemniscal activity. This statistical evaiuation corresponds to conditions which maximize the effects of the extra-lemniscal input, i.e. in our experiments chloralose anesthesia. According to our observations, this is also what happens in an unanesthetized animal which is in a state of lack of alertness.
1
-
ac
ah
PC
Ph
20msec
Fig. 21 Cot unesfhetized with cli1oralo.w. Recording by glass microelectrodes (same conditions for Figs. 22-24). Responses of a cortical cell of the S 1 cortex to stimulation of the four limbs. Note that in ac the somatotopic slow wave at short latency (evoked here by contralateral stimulation) is surmounted by spikes.
To identify these two categories of cells we have relied upon the following criteria: Extracellular unit recordings during contralateral limb stimulation show a short latency negative wave which is a"mirror image" of the positive phase of the surface potential (Figs. 21 and 23). This slow wave represents an excitatory synaptic potential. At its peak several action potentials or spikes (Figs. 21 and 22) frequently appear and these are followed by other, later ones. The first ones do not appear, but the second ones are still present when the stimulation is not in somatotopiccorrespondence with the projection zone under study (e.g., homolateral stimulation). It is, therefore, logical to attribute a double representation, lemniscal and extra-lemniscal, to that neuron. On the other hand, if the contralateral stimulation, appropriately located, does not result in early spikes (Figs. 23 and 24), and the discharge occurs after the slow wave peak only, then the lemniscal activity is absent or insufficient (the recorded slow wave may also arise in neighboring neurons). It then becomes possible to analyze what happens if the effects of the two categories of afferent impulses are made to interact on the same cortical neuron. The duration of an occlusion of two coupled extra-lemniscal impulses lasts a minimum of
THALAMIC INTEGRATIONS
139
Fig. 22 A cell which receives its afferents both by the lemniscal and extra-lemniscal pathways (somatotopic projection in this case for the contralateral posterior limb). Left: interaction between two extralemniscal projections. Homo- and contralateral stimulation of the anterior limbs. Right: a response evoked by extralemniscal projections (ah) precedes a response evoked by the two types of projections ( p c ) . Note that the evolution of the occlusion is very different in the two cases; particularly the projection coming from the homolateral forelimb occludes only for a very brief time the lemniscal portion of the projection coming from the posterior contralateral limb.
200 msec, whereas a discharge of lemniscal origin strikingly recovers a few msec after the arrival of an extra-lemniscal impulse, a time course much more compatible with a normal refractory period. Obviously, the two types of convergences do not have the same origin. Because of its temporal properties, which suggest the intervention of an inhibitory process, the first type seems to be related to the type of convergences found in the integrative relays of the spinoreticulothalamic system, from the bulbar reticular formation onwards : it represents a “projected” property which the cortical neurons reflect without themselves participating. On the other hand, it is natural to think that the meeting of the two types of afferent impulses takes place essentially at the corticaI level. The observation of a marked difference between the temporal parameters can only serve to confirm this idea, which is reconsidered in the last paragraph of this paper. ReferQnrPr I?. 14.7-148
I40
D. ALBE-FESSARD, A. FESSARD
I\
?Omroc
ah+ac
ah+ pc
Fig. 23 Above: responses to stimulation of the four limbs from a cell with purely extra-lemniscal afferents. The short latency slow wave is evoked in this case by the posterior limb stimulation (pc). Note that this slow wave is not surmounted by spikes. Below: occlusions produced between responses ah and ac, and also between ah and p c are total for spikes in both cases.
ah+PC
ah+ac
ah+-pc
ahtac
Fig. 24 The same kind of cell as the one in Fig. 23 (i.e. purely extra-lemniscal); occlusions between ah and pc, and also between ah and ac have the same temporal course in both cases. Compare the figure with Fig. 22.
(c) Thalamic and mesencephalic relays of non-specific inputs projecting on S I. The idea immediately occurred to us that the centromedian-parafascicular complex might be also the relay for the converging projections to S I, just as it is for the
141
THALAMIC INTEGRATIONS
analogous projections to the association areas. This, however, is not so because stimulation of the centre median does not give cortical responses comparable to those evoked by extra-lemniscal inputs at the level of S I. Moreover, localized cooling of this nucleus does not suppress the primary cortical responses which result from homolateral stimulation. Further, localized stimulation of the VPL produces responses in S I which have all the characteristicsof potentials evoked by stimulation of the limbs (Fig. 25), including their partial occlusion by prior stimulation of a ho-
-"\/-?r"St.VPL
P
+
ah
oc
SbVPL
I
Fig. 25 Cat anesthetized with chloralose. Left: responses obtained at two points of the cortex S I (indicated as 01 and in Fig. 18 A and R) to gross stimulation of the VPL. Right: this same stimulation of the VPL is preceded by a stimulation of the homolateral anterior limb, producing a suppression of the components of the associative type.
-
50msec
1
,
ac
SiVP+3
oc
StVP+l
ac
SVP-1
ac
Fig. 26 Cat a~is~rhetized with chloralose. Recording in the 01- area of S I. Lefr: above, response to a stimulation of the contralateral forelimb (ac); below, this response is preceded by:a homolateral peripheral stimulation (ah). Right: the same stimulation of the contralateral limb is preceded by a stimulation of moderate and fixed intensity (1 msec, 3 V) delivered at the level of the ventro-lateral nucleus l S ) , at the VPL (marked VP 3), then at the boundary of this nucleus and the n. VPL (VP itself (VP l), and finally immediately below the VPL (VP - 1). Note that all of these antecedent stimuli suppress the negative phase of evoked potential, but that only the stimulation delivered to the true primary relay affects the positive phase.
+
+
References p . 145-148
+
I42
D. ALBE-FESSARD, A. FESSARD
molateral limb. It is apparently contradictory that stimulation of the specific relay nucleus results in simultaneous activation of both the lemniscal and the extra-lemniscal systems. Does this mean that their thalamic fields of action are coextensive? In fact, conservative stimulation of this region by means of a gradually lowered bipolar electrode showed that the zone corresponding to the associative phenomena is more extensive than the strictly somatotopic zone (see Fig. 26). At least part of the relay for the convergence responses of S I cortex is located in that part of the nucleus vetitralis lateralis ( V L ) adjacent to the VPL.
-&J+ ac
ah
ac
LI\-
ac
ah
'
"/is-!StRet
ac
StRet
ac
ac
Fig. 21 Cat anesthetized with chloralose. Comparison between the effects of stimulation of the mesencephalic reticular formation and of stimulation of the homolateral anterior limb (ah) observed a t two points ( a and a), of the S I cortex. Note that the shapes of the evoked responses are similar, as are the effects of partial occlusion on the two elements of responses to stimulation of the contralateral forelimb (ac), recorded a t the same cortical points.
In this zone and in the nucleus reticularis just below the VB complex, one can demonstrate evoked responses of the associative type, as already indicated ($B), in the chloralosed animal (see Fig. 6) and in the waking animal during periods of inattention (see Fig. 9). In accordance with our practice, we began to look for the lower structure which might constitute the relay for the afferent impulses going to this "associative VL" (our limits of VL being those defined by architectonic criteria and not by functional ones). Experiments in which we stimulated the mesencephalic reticular formation (between planes A 2 and A 4, and L 1 to L 3) enabled us to reproduce on S I all the effects obtained by homolateral limb stimulation. Thus, single shock stimulation of this region leads to the occlusion of all the non-primary evoked potentials and also to suppression of the negative phase of extra-lemniscal origin in the primary evoked potential (Fig. 27). This fact had already been demonstrated by Covian, Timo-Iaria and Marseillan (1961). It may be supposed that the spinoreticulothalamic impulses relaying at this level are the same as those which produce the non-primary activity
THALAMIC INTEGRATIONS
143
of the S I cortex. However, we cannot entirely disregard the possibility of stimulation of fibers of passage. Further studies are required to decide between these two possibilities. ( d ) The locus of interaction between the two afferent systems (lemniscal and extralemniscal). The general assumption - supported by so many facts - that the two systems are independent below their cortical projection obviously indicates that their interaction occurs in the cortex. However, Mallart et al. (1961) carried out a microphysiological study at the level of the specific thalamic relay in an attempt to determine whether certain cells might receive some extra-lemniscal afferents despite their undisputable general somatotopic character (see Fig. 28). Of 72 units studied, 59
Fig. 28 Cat under chloralose anesthesia. A schematic distribution of the type of responses from 72 units to natural stimulations condensed on the stereotaxic plane A 9.5. Recording within the n. VPL and in n. VL immediately above VPL. Note the clear spatial separation existing between the group of somatotopic units and that of units devoid of spatial specificity.
responded to specific activations and 55 of these were located in the VPL itself. The other units responded only to non-lemniscal activations and all of these were located in the zone of the VL just above, and adjacent to, the VPL (part of the LP may also be involved). Among the 55 somatotopic cells of the VPL there were 1 1 which also had extra-lemniscal afferents. It was thus shown that 80%of the VPL neurons belonged solely to the lemniscal system, whereas 20% had afferents of both systems converging upon them. In contrast to this, purely lemniscalunits seem to be extremely rare in the S I cortex, under chloralose anesthesia or equivalent normal conditions (see p. 118). Such experimental conditions are different from those in which barbiturate anesthesia is used, and one must not forget that even light doses of barbiturates can suppress responses of extra-lemniscal origin. This seems to be the reason why Mountcastle was justified in concluding from his experiments on S I in cats (in 1957) References P. 145-148
144
D. ALEE-FESSARD, A. FESSARD
and from those done by Mountcastle and Powell (1959) on monkeys, that this area possesses lemniscal properties “to an overwhelming extent” (Mountcastle 1961, p. 432) although he does not deny that it also possesses “a range of functional properties comprising those of both the lemniscal and the spinothalamic systems” (ibid. p. 433). It is thus at the cortical level that the great majority of interactions occur between the two types of afferent impulses into which the different modalities of somesthetic sensitivity are divided. There is reason to believe that an analogous organization exists for the other sensory modalities. Interactions have indeed already been demonstrated between different types of specific afferent impulses of peripheral origin and projections produced by local stimulation of the non-specific thalamus (Jasper and Ajmone Marsan 1952; Jasper et al. 1955; Li, Cullen and Jasper 1956; Akimoto and Creutzfeldt 1958; Jung 1958; Branch and Martin 1958).
Fig. 29 Unrestrained awake animal. Cortical electrodes implanted at the level of area S I and of the anterior marginal cortex. Stimulating electrodes chronically placed on the contralateral superficial radial nerve. The discontinuous horizontal traces represent responses of the primary area to stimulation of the nerve. The continuous vertical trace (to be read from below upwards) is from the anterior marginal gyms and serves to indicate the degree of alertness of the animal. Note the considerable reduction of the evoked potential during the period of alertness which follows a call.
THALAMIC INTEGRATIONS
145
One observation which struck us during our experiments on the effects of alertness on evoked potentials seems to be worth mentioning here. It concerns the marked modifications of the primary evoked potential, even of its positive phase (Fig. 29), which depends on lemniscal inputs, while the potentials recorded in the specific thalamic relay nucleus (VPL) did not show such changes (see above Fig. 9). This phenomenon of the attenuation or abolition of the primary evoked potential, already demonstrated by Hernhndez-Pebn and coworkers (see Hernhndez-Pe6n 1960) is therefore not due to a deficit in the lemniscal afferent activity. Presumably it depends on intra-cortical processes which are either occlusive or inhibitory, or both, the origin and deep site of which remain to be discovered. Our study of thalamic integrations has thus led us to a higher level of integrative interactions, taking place in the cortex proper, a subject which is beyond the scope of the present paper.
REFERENCES ADRIAN,E. D., Afferent discharges to the cerebral cortex from peripheral sense organs. J. Physiol. (Lond.), 1941,100: 159-191. AKIMOTO, H. und CREUTZFELDT, 0. Reaktionen von Neuronen des optischen Cortex nach elektrischer Reizung unspezifischer Thalamuskerne. Arch. Psychiat. Nervenkr., 1958, 196: 494519. ALBE-FESSARD, D. et GILLETT, E. Convergences d’afferences d’origines corticale et periphkrique vers le centre median du chat anesthesie ou eveille. Electroenceph. clin. Neurophysiol., 1961,13: 257-269. ALBE-FESSARD, D. and KRUGER, L. Duality of unit discharges from cat centrum medianum to natural and electrical stimulation. J. Neurophysiol., 1962, 25: 3-20. ALBE-FESSARD, D. et ROUGEUL, A. Activites bilatkrales tardives evoquees sur le cortex du chat sous chloralose par stimulation d’une voie somesthesique. J. Physiol. (Paris), 1955, 47: 69-72. ALBE-FESSARD, D. et ROUGEUL, A. Activites d’origine somesthesique evoquees sur le cortex nonspecifique du chat anesthesie au chloralose. R81e du centre median du thalamus. Electroenceph. din. Neurophysiol., 1958, 10: 131-152. ALBE-FESSARD, D., DUMONT-TYC, S. et JANKOWSKA, E. Activites neuroniques somatotopiques et associatives recueillies au niveau de I’aire somatique primaire. J. Physiol. (Paris), 1961, 53: 243-244. ALBE-FESSARD, D., OSWALDO-CRUZ, E. et ROCHA-MIRANDA, C. Activites evoqukes dans le noyau caude du chat en reponse A des types divers d’afferences. I, Etude macrophysiologique. Electroenceph. clin. Neurophysiol., 1960, 12: 405420. ALBE-FESSARD, D., ROCHA-MIRANDA, C. et OSWALDO-CRUZ, E. Activitts d’origine somesthesique evoquees au niveau du cortex non-spkifique et du centre median du thalamus chez le singe anesthesib au chloralose. Electroenceph. clin. Neurophysiol., 1959, I I : 777-787. ALBE-FESSARD, D., ROCHA-MIRANDA, C. et OSWALDO-CRUZ, E. Activites Cvoquees dans le noyau caudt5 du chat en reponse a des types divers d’afferences. 11, h d e microphysiologique. Electroenceph. elin. Neurophysiol., 1960, 12: 649-661. AMASSIAN, V. E. Studies on organization of a somesthetic association area, including a single unit analysis. J. Neurophysiol., 1954, 17: 39-58. AMASSIAN, V. E. and DE VITO,R. Unit activity in reticular formation and nearby structures. J. Neurophysiol., 1954, 17: 575-603. BOWSHER, D. Termination of the central pain pathway in man: the conscious appreciation of pain. Brain, 1957, 80: 606-621. BOWSHER, D. Projection of 1 he gracile and cuneate nuclei in Macaca rnulatta. An experimental degeneration study. 1. romp. Neurol., 1958, 110: 135-155. BOWSHER, D. The termination of secondary somatosensory neurons within the thalamus of Macacu mulatta. An experimental degeneration study. J. comp. Neurol., 1961,117: 213-228. BRANCH, C. L. and MARTIN, A. R. Inhibition of Betz cell activity by thalamic and cortical stimulation. J. Neurophysiol., 1958, 21: 380-390.
146
D. ALBE-FESSARD, A. FESSARD
BREMER,F., BONNET, V. et TERZUOLO, C. Etude electrophysiologique des aires auditives corticales du chat. Arch. ini. Physiol., 1954, 3 : 390428. BRUNER, J. Reponses visuelles et acoustiques au niveau de la face mkdiane anterieure du cortex chez le chat sous chloralose. J. Physiol. (Paris), 1960, 5 2 : 36. BRUNER, J. et BUSER,P. Projections somesthesiques sur la face mediane antkrieure du cortex chez le chat anesthesie au chloralose. C . R. SOC.Biol. (Paris), 1960, 154: 530-533. BUSER,P. et BORENSTEIN, P. Donnees sur la repartition des reponses sensorielles corticales (somesthesiques, visuelles, auditives) chez le chat curarise non anesthesie. J. Physiol. (Paris), 1956a, 4 8 : 4 19-42]. BUSER, P. et BORENSTEIN, P. Observations sur les reponses corticales visuelles recueillies dans le cortex associatif suprasylvien chez le chat sous chloralose. J. Physiol. [Paris), 1956b, 4 8 : 422424. BUSER,P. et BORENSTEIN, P. Reponses corticales secondaires & la stimulation sensorielle chez le chat curarise non anesthesie. In : Condiiionnemeni el ReacliviiP en EleciroencPphalographie. Electroenceph. elin. Neurophysiol., 1957, Suppl. 6 : 89-108. BUSER,P. et BORENSTEIN, P. Reponses somesthesiques, visuelles et auditives, recueillies au niveau du cortex associatif suprasylvien chez le chat curarise non anesthesie. Eleciroenceph. clin. Neurophysiol., 1959, 1 I : 285-304. M. Sensory projections to the motor cortex in cats, a microelectrode study. BUSER, P. and IMBERT, In: Sensory Communications, Symposium Boston, 1959. Wiley and Sons, London, 1961 : 607-626. CLEZY,J. K. A,, DENNIS, B. J. and KERR,D. I. B. A degeneration study of the somesthetic afferent systems in the marsupial Phalanger. Aust. J. exp. Biol. med. Sci., 1961, 39: 19-27. COVIAN, M. R., TIMO-IARIA, C. and MARSEILLAN, R. F. Changes in cortical evoked potentials by (Editor), Brain Mechanisms and Learning. previous reticular stimulation. In : J. F. DELAFRESNAYE C.Z.O.M.S. Symposium, Montevideo, 1959. Blackwell, Oxford, 1961: 433444. COWAN,W. M. and POWELL, T. P. S. The projection of the midline and intralaminar nuclei of the thalamus of the rabbit. J. Neurol. Neurosurg. Psychiat., 1955, 18: 266-279. DEMPSEY, E. W. and MORISON, R. S. The electrical activity of a thalamo-cortical relay system. Amer. J. Physiol., 1943, 138: 283-296. DONDEY, M., ALBE-FESSARD, D. et LE BEAU,J. Premiltres applications neurophysiologiques d’une methode permettant le blocage electif et reversible de structures centrales par rkfrigeration localisee. Eleciroenceph. clin. Neurophysiol., 1962, 14 : 758-763. DROOGLEVER FORTUYN, J. and STEFENS, R. On the anatomical relation of the intralaminar and midline cells of the thalamus. Electroenceph. clin. Neurophysiol., 1951, 3 : 393400. GAZE,R. M. and GORDON, G. The representation of cutaneous sense in the thalamus of the cat and monkey. Quart. J. exp. Physiol., 1954, 39: 279-304. GAZE,R. M. and GORDON, G. Some observations on the central pathway for cutaneous impulses in the cat. Quart. J. exp. Physiol., 1955, 40: 187-194. GETZ,B. The termination of spinothalamic fibres in the cat as studied by the method of terminal degeneration. Aeta anai. (Basel), 1952, 16: 271. GLEES, P. und BAILEY, R. A. Schichtung und Fasergrosse des Tractus spinothalamicus des Menschen. Mschr. Psychiat. Neurol., 1951, 122: 129-141. J. Fibre content of the posterior column and synaptic connections of nucleus GLEES,P. and SOLER, gracilis. Zischr. Zellforsch., 1951, 36: 381. HASSLER, R. Functional anatomy of the thalamus. Congress Latino-amer. Neurocir. VI (Montevideo), 1955 : 754787. R. Neurophysiological correlates of habituation and other manifestations of HERNANDEZ-PE~N, plastic inhibition. The Moscow Colloquium on EEG of Higher Nervous Activity. Eleciroenceph. din. Neurophysiol., 1960, Suppl. 13: 101-1 14. J., ANDERSON, R., CALVET, J. and SCHERRER, J. Short and long latency cortical responses HIRSCH, to somesthetic stimulation in the cat. Exp. Neurol., 1961, 4 : 562-583. HOVDE, C . A. and METTLER, F. A. Distant electrical potentials evoked by stimulation of the putamen. Proc. Amer. Assoc. Anat. Anat. Rec., 1953, 115: 324-325. HUGELIN, A. et BONVALLFT, M. Effets moteurs et corticaux d’origine reticulaire au cours des stiniulations somesthesiques. RBle des interactions cortico-reticulaires dans le determinisme du reveil. J. Physiol. (Paris), 1958, 5 0 : 951-977. JANKOWSKA,E. et ALBE-FESSARD, D. Sur l’origine et I’interpretation de la seconde phase du potentiel evoque priniaire de I’aire somatique J. Physiol. (Paris), 1961, 5 3 : 374-375. JASPER, H. H. Electrical activity and mechanisms of the cerebral integration. In Biology of M e n i d
THALAMIC INTEGRATIONS
147
Health and Disease (Milbank Memorial Fund), Paul B. Hoeber, Inc., New-York, 1952: 226-243. JASPER,H. H. and AJMONEMARSAN,C. Thalamocortical integrating mechanisms. In Patterns of Organisation in the Central Nervous System, Res. Publ. Ass. nerv. ment. dis., 1952, 30: 493-512. JASPER, H. H., NAQUET, R. and KING,E. E. Thalamocortical recruiting responses in sensory receiving areas in the cat. Electroenceph. clin. Neurophysiol., 1955, 7: 99-1 14. JOHNSON, T. N. Fiber connections between the dorsal thalamus and corpus striatum in the cat. Exp. Neurol., 1961, 3: 556-569. JUNG,R. Coordination of specific and nonspecific afferent impulses at single neurons of the visual cortex. In Reticular Formation of the Brain. Henry Ford Symposium, Detroit, 1957, Little, Brown and Co, Boston, 1958: 423434. KARPLUS, J. P. und KREIDL,A. Ein Beitrag zur Kenntniss der Schmerzleitung im Ruckenmark, nach gleichzeitiger Durchschneidung beider Ruckenmarkshalften in verschiedenen Hohen bei Katzen. Pfliigers Arch. ges. Physiol., 1914, 158: 275-287. KNIGHTON, R. S. Thalamic relay nucleus for the second somatic receiving area in the cerebral cortex of the cat. J. comp. Neurol., 1950,97: 183-185. KRUGER, L. and ALBE-FESSARD, D. Distribution of responses to somatic afferent stimuli in the diencephalon of the cat under chloralose anesthesia. Exp. Neurol., 1960, 2 : 442467. KRUGER L., SIMINOFF, R. and WITKOVSKY, P. Single neuron analysis of dorsal column nuclei and spinal nucleus of trigeminal nerve in cat. J. Neurophysiol., 1961, 24: 333-349. LAPORTE, Y. et BESSOU, P. Modifications d’excitabilitede motoneurones homonymes provoquees par I’activation physiologique de fibres afferentes d’origine musculaire du groupe 11. J.Physiol. (Paris), 1959, 51: 897-908. 11, C. L., CULLEN, C. and JASPER,H. H. Laminar microelectrode analysis of cortical unspecific recruiting responses and spontaneous rhythm. J. Neurophysiol., 1956, 19: 131-143. LIBOUBAN, S. et JUTIER,P. Mise en evidence chez Ie chat de rkponses de type associatif dans la partie anterieure du gyrus marginalis par stimulations naturelles ou electriques de la voie optique. J. Physiol. (Paris), 1961, 53: 412413. MACCHI,G. e ARDUINI,A. Organizzazione anatomofunzionale dei rapporti talamo-corticali. Atti. Soc. Ital. Anat., XVIII Convegno sociale, Parma, 1957: 25-151. MACCHI,G., ANGELERI, F. and GUAZZI,G. Thalamocortical connections of the first and second somatic sensory areas in the cat. J. comp. Neurol., 1959, 3: 387405. MALLART, A. Donnees sur les types de fibres conduisant les afferences somatiques au noyau centre median du thalamus. J. PhysioZ. (Paris), 1961, 53: 422423. MALLART, A. Aferencias sensoriales a1 Nucleo Centro Mediano del Talamo. Thesis, Madrid, 1961. MALLART, A., MARTINOYA, C. et ALBE-FESSARD, D. Comparaison entre les types de stimulations naturelles qui evoquent des reponses dans les relais spkifiques ou les relais associatifs. J. Physiol. (Paris), 1961, 53: 421422. MASSION, J. et MEULDERS, M. Influence de la decortication sur les reponses d’origine somesthesique. J. Physiol. (Paris), 1960, 52: 172-175. MASSION, J. et MEULDERS, M. Les potentiels Cvoques visuels et auditifs du centre median et leurs modifications aprks dkortication. Arch. int. Physiol., 1961,69: 26-29. MATZKE, H. A. Course of the fibers arising from nucleus gracilis and cuneatus of the cat. Anat. Rec., 1949, 103: 487. MCLARDY, T. Projection of the centromedian nucleus of the human thalamus. Brain, 1948, 70: 290-303. MCLARDY, T. Diffuse thalamic projection to cortex: an anatomical critique. Electroenceph. clin. Neurophysiol., 1951, 3: 183-188. MEHLER, W. R. The mammalian “pain tract” in phylogeny. Anat. Rec., 1957, 127: 332. MEHLER, W. R., FEFERMAN, M. E. and NAUTA, W. J. Ascending axon degeneration following anterolateral chordotomy. An experimental study in the monkey. Brain, 1960, 83: 718-750. MORRELL, F. Microelectrodes and steady potential study suggesting a dendritic locus in closure. The Moscow Colloquium on EEG of Higher Nervous Activity. Electroenceph. elin. Neurophysiol., 1960, Suppl, 13: 65-80. MOUNTCASTLE, V. B. Modality and topographic properties of single neurons of cat’s somatic sensory cortex. J. Neurophysiof., 1957, 20: 408434. MOUNTCASTLE, V. B. Some functional properties of the somatic afferent system In Sensory Communications Symposium, Boston, 1959. Wiley and Sons, New York, 1961 : 844 pp.
148
D. ALBE-FESSARD, A. FESSARD
MOUNTCASTLE, V. B., COVIAN, M. R. and HARRISON, C. R. The central representation of some forms of deep sensibility. In Patterns of Organization in the Central Nervous System. Res. Publ. Ass. nerv. ment. Dis., 1952, 30: 339-368. MOUNTCASTLE, V. B. and POWELL, T. P. S. Neural mechanisms subserving cutaneous sensibility, with special reference to the role of afferent inhibition in sensory perception and discrimination. Bull. Johns Hopk. Hosp., 1959, 105: 201-232. MOUNTCASTLE, V. B., DAVIES,P. W. and BERMAN, A. L. Response properties of neurons of cat's somatic sensory cortex to peripheral stimuli. J. Neurophysiol., 1957, 20: 374-407. NASHOLD, B . S., HANBERRY, J. and OLSZEWSKY, I. Observations on the diffuse thalamic projections. Electroenceph. clin. Neurophysiol., 1955, 7 : 609-620. NAUTA,W. 3. H. and KUYPERS, H. G. Some ascending pathways in the brain stem reticular formation. In Reticular Formation of the Brain, Henry Ford Symposium, Detroit, 1957. Little, Brown, Co, Boston, 1958: 3-30. NAUTA, W. J. H. and WHITLOCK, D. G. An anatomical analysis of the non-specific thalamic projection system. In Brain Mechanisms and Consciousness. C.I.O.M.S. Symposium. Blackwell, Oxford, 1954: 81-117. PERL,E. R. and WHITLOCK, D. G . Somatic stimuli exciting spinothalamic projections to thalamic neurons in cat and monkey. Exp. Neurol., 1961, 3 : 256-296. POGGIO, G. F. and MOUNTCASTLE, V. B. A study of the functional contributions of the lemniscal and spinothalamic systems to somatic sensibility. Bull. Johns Hopk. Hosp., 1960, 106: 266-3 16. POWELL, T. P. S . and COWAN, W. M. A study of thalamo-striate relations in the monkey. Brain, 1956, 7 9 : 364-390. POWELL. T. P. S. and MOUNTCASTLE, V. B. Some aspects of the functional organization of the cortex of the postcentral gyrus of the monkey: a correlation of findings obtained in a single unit analysis with cytoarchitecture. Bull. Johns Hopk. Hosp., 1959, 105: 133-162. QUENSEL, F. Ein Fall von Sarcom der Dura spinalis. Neurol. Zentrbl., 1898, 1 7 : 482492. R A M ~YNCAIAL,S. Histologie du Systerne nerveux de I'Homme et des Vertgbris, Cons. Sup. Invest. cient., Madrid, 1909, Edit. 1952: 986 pp. C. N. Cortical connections and functional organization of the thalamic ROSE,J. E. and WOOLSEY, auditory system of the cat. I n Biological and Biochemical Buses of Behavior. The Univ. of Wisconsin Press, 1958: 476 pp. R o w , G. F. and BRODAL, A. Terminal distribution of spinoreticular fibers in the cat. Arch. Neurol. Psychiat. (Chic.), 1957, 7 8 : 439453. SCHEIBEL M. E. and SCHEIBEL A. B. Structural substrates for integrative patterns in the brain stem reticular core. In Reticular Formation of the Brain, Henry Ford Symposium, Detroit, 1957. Little, Brown and Co, Boston, 1958: 31-55. SCHEIBEL, M. E., SCHEIBEL, A. B., MOLLICA, A. and MORUZZI,G. Convergence and interaction of afferent impulses on single units of reticular formation. J. Neurophysiol., 1955, 18: 309-33 1. SEGUNDO, J. P. and MACHNE, X. Unitary responses to afferent volleys in lenticular nucleus and claustrum. J. Neurophysiol., 1956, 19: 325-339. TORVIK, A. Afferent connections to the sensory trigeminal nuclei, the nucleus of the solitary tract, and adjacent structures. An experimental study in the rat. J. comp. Neurol., 1956, 106: 51-141. WALKER, A. E. The Primate Thalamus. The University of Chicago Press, Chicago, 1938: 321 pp. WALL,P. D. Two transmission systems for skin sensations. In Sensory Communications, Symposium, Boston 1959. Wiley and Sons, London, 1961: 844 pp. WENDT, R. and ALBE-FESSARD, D. Sensory response of the amygdala with special reference to somatic afferent pathway. In: Physiologie de I'Hippocampe, Monrpellier, 1961. C.N.R.S. Publ., Paris, 1962: 17 1-200. W. H. Pain, its Mechanisms and Neurosurgical Control. Ch. C. Thomas Publ., WHITE,J. C. and SWEET, Springfield, 1955. WHITLOCK, D. G. and PERL,E. R. Thalamic projections of spinothalamic pathways in monkey. Exp. Neurol., 1961, 3 : 240-255. YOSHII,N. Principes methodologiques de I'investigation electroenckphalographique du comportement conditionne. In Conditionnement et RPactivite' en EEG. Electroenceph. clin. Neurophysiol., 1957, SUPPI.6 : 75-88.
DISCUSSION
149
DISCUSSION J. M. BROOKHART:
I should like to congratulate Mme. Albe-Fessard on a very interesting work which was clearly presented. While you were in the process of “dissecting” the evoked response into its two components, one lemniscal in origin, and an extra-lemniscal negative phase, I was reminded of the conclusions arrived at by Spencer and myself. In the study of augmenting, recruiting and spindle waves, we concluded that there were two basically different mechanisms of cortical response. I wonder if you would care to comment on the possibility that there are parallelisms between the two sets of observations. M. CARRERAS:
I was very interested in the observation made by Mme. Albe-Fessard that neurones which receive a somatic input from ipsilateral or bilateral peripheral areas can be found in VPL. As far as I know, analyses carried out in the VPL unit in Dr. Mountcastle’s laboratory, either on slightly anaesthetized or on unanaesthetized animals, none showed such functional properties among many hundreds of neurones recorded. However, such neurones are quite common in the thalamic region lying posteriorly to VPL (posterior group of thalamic nuclei of Rose and Woolsey), a region which seems to project to the second somatic area. We have, in fact, shown that neurones related to ipsilateral or bilateral peripheral receptive fields can easily be found in this cortical area. I wonder if Mme. Albe-Fessard can offer any suggestion regarding this difference between her experiments and those of Mountcastle. This seems to me to be very important and it is possible that it is a disagreement at the strictly observational level. POGGIO, G. F. and MOUNCASTLE, V. B. A study of the functional contributions of the lemniscal and spinothalamic system to somatic sensibility. Central nervous mechanisms. Bull. Johns Hopk. Hosp., 1960, 106: 266-316
MOUNTCASTLE, V. B. Some functional properties of the somatic afferent system. In: W. ROSENBLITH (Editor), Sensory Communications. Wiley and Sons,London, 1961. C. AJMONE MARSAN: Dr. Albe-Fessard should be complimented on her very interesting paper. I would like to obtain some additional information on the following details : A. In her presentation, the nucleus VL of the thalamus has been included among those subcortical structures characterized by non-specific (extra-lemniscal) projections. In my experience, the identification of VL and differentiation between it and VA nucleus is rather difficult; stereotaxic coordinates and histological checking are generally inadequate. I wonder whether some functional criteria were also applied (i.e. delimitation of a specific cortical response produced by VL stimulation) before conclusions were drawn about the rather unexpected part played by a nucleus up to now considered to be only part of the cerebello-rubral and motor cortical mechanisms. B. The latency of the cortical response to CM stimulation was found by Dr. Albe-Fessard to be of the order of 2 4 msec. In our experience as well as in that of most investigators, such latency value has lasted much longer-actually 10times as long (15-45 msec). These long delays were found in both the anaesthetized and in the cerveau isol.4 preparation. The discrepancy appears to be too marked to be ignored. In view of the fact that almost all the present results were obtained in animals under chlora-
I50
D. ALBE-FESSARD. A. FESSARD
lose, it seems justified to raise the question of whether this particular type of anaesthesia may be responsible for these unusual findings. The possibility of stimulating fibers which pass through the CM or nearby structures and are characterized by a lower excitability threshold, should also be taken into consideration. C. Finally, in some experiments in which the behaviour of single unit firing was studied with twin stimuli, Dr. Albe-Fessard uses the term occlusion to describe the decrease in number of spikes observed in the test response. i wonder whether this is the correct term.
H. H. JASPER: This extensive series of carefully conducted and controlled experiments has contributed much of precise value to the subject of this symposium. Of particular interest to me was the demonstration of certain rather specific properties of ascending pathways passing through, or terminating in the nucleus centrum medianum, and the fact that this central ascending system is not made up of collaterals from the lemniscal system. This raises a number of questions about the detailed anatomy of the thalamic unspecific system itself and in relation to the results obtained from n. ventralis lateralis. The nucleus centrum medianum is a very complex heterogeneous structure of ill-defined anatomical boundaries, especially in its caudal extent. It contains many fibres de passage. Those of the rubrothalamic system passing through into the nucleus V L and those from the subthalamus to nucleus ventralis anterior are of particular importance to some of the findings just presented. A detailed microscopic study of thalamic electrode sites might do much to clarify the significance of some of these electrophysiological findings. i t is gratifying to find that “chloralose responses” can also be found in an attenuated form in animals with chronic implanted electrodes and without anaesthesia. It may be recalled that the early suppression of “unspecific” responses by barbiturate anaesthesia has been considered to be of particular significance in the attempt to explain the anaesthetic action of such drugs. I wonder if Dr. Albe-Fessard has any explanation for the mysterious form of anaesthetic action of chloralose, in spite of the enhancement of all forms of evoked potentials, specific and unspecific?
G . F. RICCI:
Some of Dr. Albe-Fessard’s interesting findings in the somatic cortex are somewhat similar to the results obtained by Dr. von Euler and myself while working on the auditory system (1958). It may be of interest to recall here that in these experiments it was shown that the shape and polarity of the responses recorded from a “focal” electrode on the primary auditory area can be made to vary merely by moving the stimulating electrode within the medial geniculate body. To us this finding seemed to suggest the existence of two different types of connections between the medial geniculate body and the primary auditory cortex. We know very little about the significance of these two types of connections. There is, however, evidence that surface negative responses evoked by stimulation of the various thalamic nuclei may have something to do with the so-called aspecific cortical mechanisms. If so, then “aspecific” connections may be also represented within “specific” systems, because surface-negative responses can be selectively obtained by stimulation of the specific thalamic nuclei.
EULER,C. VON and Riccr, G . F. Cortical evoked responses in auditory area and significance of apical dendrites. J . Neurophysiol., 1958,21: 231-246.
J. L.O’LEARY: I should like to make the following comment on Ajmone-Marsan’s question concerning latency from the centro-median to the cortex and the statement by Jasper concerning fibers of passage. I have observed, in Nauta preparations after lesions in CM, the radiation into the thalamus of dege-
DISCUSSION
151
nerated axons; perhaps these could in part be fibers of passage which have not originated from CM neurones. A. ZANCHETII:
It is certainly gratifying to see how well the physiological data presented by Mme. Fessard fit into the picture of ascending anatomical connections to the different thalamic nuclei. However, there is one set of data which I have found somewhat disturbing, in so far as they do not seem to fit into the anatomical frame of reference. This discrepancy refers to a report by Mme. Fessard and Kruger that electrical responses can be recovered from centrum medianum also upon dorsal column stimulation. This is a point upon which I would like to have some more information. With regard to the subject of centrum medianum responses, it should be recalled here that, according to Mehler, Feferman and Nauta (1960), what they call centrum medianum does not actually receive spino-thalamic afferents, while connections from the antero-lateral funiculus of the cord and from the brain stem reach neighboring regions, which are however, according to Jasper and Ajmone Marsan’s Atlas of the diencephalon, included in the centrum medianum. I would therefore like to ask Mme. Fessard whether in her investigations centrum medianurn has been defined following the more restricted or the broader of the above-mentioned definitions.
D. Distribution of responses to somatic afferent stimuli in the KRUGER,L. and ALEE-FESSARD, diencephalon of the cat under chloralose anesthesia. Exp. Neurol., 1960,2: 442-467. M. E. and NAUTA,W. S. H. Ascending axon degeneration following MEHLER,W. R., FEFERMAN, antero-lateral cordotomy. An experimental study in the monkey. Brain, 1960,83: 718-750. H. GASTAUT : I apologize for joining in so late the discussion of the remarkable report by Mme. and M. Fessard. I do so only to underline a technical detail concerning the use of chloralose in the study of evoked potentials. In a work published in 1951, wereported that cerebellar potentials evoked by a flash of light, noise or touch occurred at the same time as those evoked on the frontal cortex, only in chloralose anesthetized animals or in those receiving cardiazol after barbiturate anesthesia. According to the terminology used by us at that time, we viewed them as irradiated potentinis of the non-specific pathways rather than as evoked specific potentials. Likewise in 1951 we observed in our study of cortical excitability cycles that chloralose exerted an effect on the post-excitatory recokeiy time of neurons, which it increased as much as barbiturates did. It also increased the response amplitude as much as cardiazol did, so that paradoxically sleep and convulsions were caused and these were accompanied by irradiated potentials. We therefore concluded that the sensory responses seen in chloralosed animals were non-specific irradiated potentials and we wrote: “It is possible that ignorance of these facts has led many authors to make an erroneous interpretation of the observation made on animals under chloralose, especially with respect to the distribution of sensory potentials and cortico-cortical or cortico-subcortical relationships”. As a result of this, we stopped using chloralose in the study of evoked potentials. Now that Mme. and M. Fessard have shown that all the responses obtained under chloralose can also be obtained in the unanesthetized animal, I believe our fears areno longer justified. Nevertheless 1 would like to hear the opinion of these authors, because they are certainly the most qualified t o speak about this matter. GASTAUT, H., NAQUET,R., BAIXER M. et ROGER,A. Signification de la reponse cCrCbelleuse a la stimulation lumineuse chcz le chat. J. Physiol. (Paris), 1951,43: 737-740. GASTAUT, H., ROGER,Y . ,CORRIOL, J. et NAQUET, R. Etude Clectrocorticographique du cycle d’excitabilitt cortical. Electroenceph. d i n . Neurophysiol., 1951, 3 : 401-428. G . F. Rossi:
I would like to pose you the following two brief questions: 1. You suggested that impulses evoked by peripheral sensory stimulation reach the reticular forma-
152
D. ALBE-FESSARD, A. FESSARD
tion by passing along fibers of the lateral columns of the spinal cord (spino-reticular fibers or “paleospino-reticular” pathway). This view is strongly supported by anatomical findings. As you have pointed out, anatomical studies have shown that the bulk of the spino-reticular fibers end a t the bulbar level (Rossi and Brodal 1957). Have you any physiological evidence t o support this view? What is the latency of the bulbar reticular response to peripheral sensory stimulation? Is it much shorter than the latency of the response of the midbrain reticular formation? 2. Do you know whether there are differences between the electrical potential evoked in the cortex by stimulation of the bulbar reticular formation and those evoked by midbrain reticular stimulation?
Ross], G. F. and BRODAL, A. Terminal distribution of spinoreticular fibers in the cat. Arch. Neurol. Psychiat. (Chic.), 1957,78: 439453. J. C. ECCLES: I was particularly interested in the depressed responses that were observed a t quite long intervals after various conditioning stimuli. The depressions of some hundreds of milliseconds in duration could be due to presynaptic inhibition. This would certainly be likely when both the conditioning and the testing stimuli were applied to the skin of the ipsilateral and contralateral limbs.
H. W. MAGOUN: Has Mme. Albe-Fessard made any observations which might indicate that the thalamic units which discharge in such striking correspondence with potentials evoked by lemniscal and extralemniscal afferents, discharge also in relation to the later spindle-burst waves which are often triggered by afferent stimuli? Monnier has differentiated a non-specific thalamo-cortical system for the spindleburst mechanism with properties different from other thalamo-cortical systems. Do any of the thalamic units which fire during the evoked potentials which Albe-Fessard has studied so elegantly, fire also during spindle-burst waves?
F. BREMER: 1would have liked to have intervened earlier during the discussion of the report, so rich in experimental data, of our colleagues the Fessards. I was prevented from doing so by the inscriptions made in advance, which testifies in any event to the attention which their communication has provoked. Also, should I limit myself to comment on one fact that struck me particularly by reason of my interest in all the problems of cerebral wakefulness? This fact is the contrast between the depression of the sensory response of the centre median during this wakefulness, and the resistance of the specific relay nucleus of the thalamus. Everything happens, it seems to me, as if the centre median owed, as does the cortex, this sensitivity to wakefulness to the fact that it acts as an integrating structure in which non-specific inputs and disparate specific inputs converge and are mutually occluded.
0.POMPEIANO: Because you recorded in your experiments associative responses from the CM by applying a painful stimulation (a pin-prick), one may think that the volleys conducted by the fastest cutaneous afferents, which are not concerned with pain sensation, do not reach the associative centers. It would be surprising indeed if association occurred only for painful sensations. It may happen that even the largest cutaneous afferent fibers project to the associative centers, but spatial and/or temporal summation is required in order to activate the CM. A second question which I would like to ask you is whether, in addition to stimulating pure cutaneous nerves like the superficial radial nerve, you also tried to stimulate pure muscular nerves in order to elicit associative responses? If so, which type of muscular afferents was responsible for this evoked activity?
DISCUSSION
153
D. ALBE-FESSARD’s replies To J. M . Brookhart We have found cells in the cortex belonging to the primary projection system which also receive afferents from the non-specific nuclei of the thalamus. Almost half these neurones receive only afferents of this type, by a pathway whose characteristics differ from those ending in the true association cortex. Thus there are at least two different types of “associative” projections and I wonder whether they may not correspond to the distinction that you and Spencer have made between the “recruiting” and “augmenting” types of response.
To M . Carreras I would like first of all to point out that in the true specific thalamic nucleus (VPL : A 9.5-10.5, according to the Atlas of Jasper and Ajmone Marsan), we have found only an extremely small number of non-somatotopic neurones while we have, on the other hand, found a large number in the S I cortex. This difference between our results and those of Mountcastle appears to be essentially due to the fact that we employ chloralose anaesthesia, which augments the responses of the nonspecific systems. To C . Ajmone Marsan A. I wish first to explain that, when we speak of the position of an electrode in a nucleus, we do not identify this position by the stereotaxic coordinates of the electrode as these are noted in the course of the experiment, but in relation to the reference electrode tracks seen in frozen sections. Having made this clear, we believe that, even if VL is concerned in the transmission of non-specific impulses to the S I cortex, this is not to deny its importance in the relay of impulses from the cerebellum and red nucleus. In our laboratory, Massion and Lelord have recently been studying its activation by impulses relayed from the cerebellum by way of the nucleus interpositus. B. I personally believe that short latency ( 2 4 msec) responses obtained in the cortex after stimulation of CM are found only in small well-defined areas, and that they can easily be missed in unanaesthetised animals, unless they have previously been demarcated under chloralose. Recruiting or non-recruiting responses at longer latency can appear elsewhere, in widespread areas of the cortex, after stimulation of CM. At the moment we are not working on this latter system. We are aware of the possibility of stimulating fibres of passage, but this cannot be got over by using different anaesthetics, even when the responses are of the recruiting type. I should add that we have always applied low stimulating voltages (maximum 4-6 V for a single stimulus of 0.5 msec) and have frequently used concentric electrodes. C. As to the term occlusion, it is certainly not correct, depression and suppression being certainly better. We wish to avoid the term inhibition, because this implies a particular hypothesis with regard to the mechanism of suppression. However, we have reason to believe that these suppressions, when they are of long duration, are due to an inhibitory process (videinfra “To Eccles”).
To H. H . Jasper The site of the electrode tips has always been histologically verified, but we have not been able, for reasons of space, to include photographs of the sections. We may say, however, that the best responses in CM are obtained between the coordinates A 7-8, L 2.5-3, H+ 1 . 5 4 ; this corresponds to a cellular region, although it is, of course, impossible to eliminate the presence of fibres of passage. However, there are even more fibres beyond these limits of CM, but we do not find any response there. Attention should also be called to the fact that these responses, when they are maximal in the awake or thalamic animal, have the same amplitude as they have in animals under chloralose anaesthesia. Thus chloralose does not produce an abnormal amplification of the responses. Work which is now in the process of publication shows that chloralose removes an inhibitory effect of telencephalic origin at the level of CM (see Meulders, Massion, Colle and Albe-Fessard. Electroenceph. clin. Neurophysiol., 1963, I S : 29-38. To G. F. Ricci I have already alluded to the similarity of our results to those of Ricci and von Euler in an earlier publication (Actualitis neurophysiologiques. Masson, Paris, 1961). The region of the medial geniculate body which gives rise to the negative phase of the surface response is perhaps the homologue of that part of VL close to VP which acts as a relay centre for non-specific impulses.
I54
U . ALRE-FESSARD, A. FESSARD
To A . Zurrclieiii Your first question is most pertinent and gives me the opportunity to give you orally information which was not presented in my communication. In fact, afferents to CM passing through the dorsal columns appear always to rejoin the antero-lateral columns a t a higher level. As to your second question, about our definition of CM, 1 would like to make it clear that we have always used the term in its widest sense, namely as it is shown in the Atlas of Jasper and Ajmone Marsan. To be more exact, we would prefer to use the expression “centromedian-parafascicular complex”. To H. Gustuuf We believe in fact that, when we use chloralose, we cause an enhancement of the effects of those impulses which ascend in the antero-lateral columns. Without creating new pathways this anaesthetic permits one to study under optimal conditions a system which, in the awake animal, fluctuates in its activity.
To G . F. Rossi We have effectively shown that spino-reticular afferent impulses strongly activate Olszewski’s nucleus giganto-cellularis. The latency of the response (about 10 msec for the contralateral forelimb) is slightly less than that of CM for the same peripheral stimulation. The latency of the midbrain responses is of the same order. We have not yet systematically studied cortical responses to stimulation of the medullary reticular formation. To J. C. Eccles It is highly probable that the long-lasting depression which follows stimulation of every structure in the associative system is due to a true inhibition. We have in fact been able t o observe this depression by itself in the absence of preceding excitation, in unicellular recordings both in CM and in the cortex. But we have no indication of the mechanism of this inhibition.
To G. V. Magoun Although we have not made a systematic study of them, we can say, as a result of our observations, that late trains of rhythmic responses appear sometimes in CM and the associative part of VL, but not in VPL. But it is especially in the S 1 cortex that we have been able to make extra- and intracellular recordings which show that the same units respond, not only to the primary stimulus, but also to each wave of the consecutive train of impulses. To F. Bremer The remark of our colleague, Professor Bremer, is certainly a likely interpretation of our observations.
To 0 .Pompeiano 1 regret that, in the oral presentation, it was not possible to present the findings of our collaborator Mallart, but they are summarized in the text (see p. 130). He has clearly shown that from cutaneous nerves of both groups 11 and 111, impulses reach CM just as well as VPL. From muscular nerves, only those of group 111appear to send impulses to both structures. In the chronic animal, the stimulus which effectively activates CM does not appear to be necessarily unpleasant for the animal. It is therefore necessary, when we are considering the associative projections, to avoid the use of the term “painful sensation”.
Influence of Unspecific Impulses on the Responses of Sensory Cortex S. P. NARIKASHVILI Institute of Physiology, Georgian Academy of Sciences, Tbilisi (U.S.S.R . )
Since the diffuse cortical effects of stimulation of the medial thalamus (Morison, Dempsey and Morison 1941; Morison and Dempsey 1942; Dempsey and Morison 1942a; Morison, Finley and Lothrop 1943), hypothalamus (Murphy and Gellhorn 1945) and brain stem reticular formation were discovered (Moruzzi and Magoun 1949), many more studies have been made of the part played by so-called nonspecific impulses in the elaboration of cortical evoked reactions. A depressive influence of the thalamic nonspecific stimulation (Jasper and Ajmone Marsan 1952; Stoupel 1958; Steriade and Demetrescu 1960; Landau, Bishop and Clare 1961) or brain stem reticular stimulation upon the cortical responses has been observed*. On the other hand, a facilitatory action of nonspecific thalamic activation has also been described**. At the same time some authors failed to find any influence upon the cortical primary responses by stimulation of the thalamic nonspecific structurest or of the brain stem reticular formationtt. Such various and often contradictory, or at least not quite definite, results impelled us to study the question in detail. This was done by my collaborators Drs. E. S. Moniava, D. V. Kadjaia, S. M. Butkhusi and myself, as reported in the following pages.
*
Moruzzi and Magoun 1949; Bremer and Bonnet 1950;Gellhorn1953;Bremer 1953,1954;Gauthier Parma and Zanchetti 1956; Nakao and Koella 1956; Hernandez-Peon, Guzman-Flores, Alcaraz and Fernandez-Guardiola 1957; Desmedt and La Grutta 1957; BuserandBorenstein 1959; Schoolman and Evarts 1959; Appelberg, Kitchell and Landgren 1959; Long 1959; Horn and Blundell 1959; Bremer and Stoupel 1959; Horn 1960; Naquet, Regis, Fischer-Williams and FernandezGuardiola 1960. ** Thalamic activation (Jasper and Ajmone Marsan 1952; Jung and Baumgartner 1955; Creutzfeldt and Baumgartner 1955; Li 1956; Creutzfeldt, Baurngartner and Jung 1956; Jung 1958, 1960; Jung, Creutzfeldt and Grusser 1957,1958; Akimoto andcreutzfeldt 1957,1958;Baumgartner 1958; Creutzfeldt and Akimoto 1958; Grutzner, Grusser and Baumgartner 1958; Long 1959; Creutzfeldt and Griisser 1959; Bremer and Stoupel 1959; Steriade and Demetrescu 1960; Morillo 1961) and reticular or hypothalamic activation (Gellhorn, Koella and Ballin 1954, 1955; Nakao and Koella 1956; Dumont and Dell 1958,1960; Bremer and Stoupel 1959; Long 1959; Mancia, Meulders and Santibafiez 1959; Bremer, Stoupel and Van Reeth 1960; Steriade and Demetrescu 1960). f Dempsey and Morison 1942; Jasper 1949; Jasper and Ajmone Marsan 1952; Stoupel 1958. tt Morison, Finley and Lothrop 1943; Moruzzi and Magoun 1949; Moruzzi 1949; Gauthier, Parma and Zanchetti 1956. Xrfcrmces D , 180-183
156
S. P. NARIKASHVILI
TECHNIQUE
Experiments were performed on cats, with two kinds of preparations according to two experimental procedures. When the influence of thalamic nonspecific nuclei was studied, the experiments were performed on cats under light nembutal(20-25 mg/kg) or chloralose anesthesia. When the influence of reticular formation was investigated, curarized (d-tubocurarine) and “enciphale isoli” preparations were used. Experiments with thalamic nonspeciJc stimulation. The influence of the thalamic nonspecific nuclei upon primary responses (PR) elicited by low frequency (1 per sec, and less) stimuli applied to the skin of the contralateral forelimb or to the specific thalamic relay nucleus (n. ventralis postero-lateralis, VPL) was studied by means of two techniques : either the thalamic nonspecific structures (zona incerta, n. centralis medialis, n. ventralis anterior) were stimulated (0.2-0.5 msec, 5-15 V) at different frequencies (from 8-10 to 100-200 per sec); or the interaction of nonspecific and specific impulses was studied by using paired stimuli (the intervals between conditioning and testing stimuli varied within wide limits, 5-1000 msec). The responses were recorded by means of a mirror oscillograph. Experiments with reticular stimulation. In these experiments, responses of the visual system were used as a test-reaction. The reticular formation (RF) was activated either by direct electrical stimulation using bipolar electrodes inserted stereotaxically at collicular level (diameter of the noninsulated tips, 20-25 microns; interelectrode distance 0.5-1 mm) or by stimulation of the skin of a forelimb (5-20 V, 1 msec). In these experiments the influence of the reticular volleys upon the specific responses was also studied under different conditions: (1) on a background of responses to low frequency (1 per sec and less) or frequent (4-50 per sec) photic stimuli (I msec) acting on both eyes, shortlasting stimulations of the skin or of the mesencephalic reticular formation (MRF) were applied; (2) the shortlasting stimulation of the skin or M R F was added to low or high frequency stimuli applied to different parts of the visual system (optic chiasm (OCh), optic tract (OT), lateral geniculate body (GL), optic radiation (OR)); (3) the first shock of paired repetitive stimuli was applied to M R F or skin, and the second triggered the flash or was applied directly to different parts of the visual path. Square pulses (0.2-1 msec, 3-10 V) at different frequencies and bipolar stimulating electrodes were used. In most instances, the potentials of deep structures were recorded with monopolar derivation; in others, either bi- or monopolar derivations were used, the indifferent electrode being inserted in the frontal bone or in the neck muscles. For recording we used an inkwriter and a two beam cathode ray oscillograph.
INFLUENCE OF UNSPECIFIC IMPULSES
157
RESULTS A. THE INFLUENCE OF NONSPECIFIC THALAMIC STIMULATION UPON RESPONSES OF THE
SOMATOSENSORY SYSTEM
1. Repetitive stimulation of the thalamic nonspecijic structures. During nonspecific thalamic stimulation at frequencies of 8-1 %/set a marked facilitation of PR's was observed in favourable cases, and above all when the preparation was in good functional state. More precisely, the primary response, especially its negative component, increased during the waxing phase ( c j Moruzzi et al. 1950), while it decreased during the waning phase (Narikashvili 1957; Narikashvili and Moniava 1957). In Fig. 1 (tracings 2 and 3), it is clear that PR was enhanced during the waxing phase of the recruiting responses, especially in its negative component. During the waning phase, the negative component of PR was markedly decreased (the end of tracing 3). This phenomenon, in particular the enhancement of the amplitude of PR during the waxing phase, was most marked at the beginning of the thalamic stimulation, during the first and second waxing phases.
Fig. 1 The influence of frequent or rare repetitive thalamic nonspecific stimulations on primary responses (PRs). Tracing 1. Upper curve, potentials of somatosensory cortex (peripheral part); lower one, those of central region (region of greatest amplitude for primary responses, PR) during stimulation of the zona incerta (indicated by horizontal line; 10 V, 200/sec, 0.5 msec). Tracings 2 and 3. Same conditions. Through stimulation of zona incerta (5 V, lO/sec, 0.5 msec, indicated by an arrow) recruiting responses are evoked on the background of which PR (especially its negative component) is markedly facilitated. Here and in all following figures an upward deflection corresponds to negativity. Time calibration here and in following figures (except special indications): 20 msec.
When the thalamic nonspecific nuclei were stimulated at a higher frequency(20-40 per sec), in contradistinction to the effects of stimulation at a low frequency (8-12 per sec), the "recruiting" response proceeded without waxing and waning. The amplitude of the waves was maximal from the beginning of the stimulation. Then, during stimulation, they decreased quickly (the higher the stimulation frequency, the quicker the reduction) and disappeared, despite the lasting stimulation. Now, if PR was elicited References a. 180-183
158
S. P. NARIKASHVILI
at the beginning of the thalamic stimulation, when the waves were high in amplitude, PR was markedly increased; whereas it was reduced if it coincided with the following period, when the waves had become markedly decreased as a result of more or less long lasting repetitive stimulation. In Fig. 2, some examples are presented. Thus tracing 2 in that figure shows that PR was considerably enhanced initially (compare with tracing 1). In one instance the negative component was increased and in the other the positive one was increased. After some time, when the effect of the specific thalamic action had decreased or had almost completely disappeared (the end of tracing 3), PR became markedly depressed in comparison to the responses observed before the thalamic stimulation. This effect is better seen in tracings 4 and 5. At the beginning of the thalamic stimulation (tracing 4), the amplitude of P R was considerably increased. In the course of the stimulation the “recruiting” waves soon became smaller and the amplitude of PR was itself smaller (tracing 5). After the thalamic stimulation had ceased, P R remained reduced
-.
05mV
Fig. 2 The influence of repetitive stimulation of the n. centralis medialis on cortical PR’s. ln all tracings: upper curves, a peripheral region of somatosensory cortex; lower ones, central region. N. centralis medialis is stimulated a t 35-36 per sec (5 V, 1 msec). Tracing I . PR’s before the stimulationofthalamic nonspecific nucleus. Trucings 2-7. PR’s on the background of nonspecific thalamic reaction ( 3 , direct continuation of 2; 5, continuation of 4 (3 sec passed between them); 7, continuation of 6). Tracings 8-9. Same experiment as in tracing 4 and 5, but here the record is presented from beginning to end. Cessation of thalamic stimulation is seen by the end of stimulation artifacts. Details in the text. This experiment has been carried out on the same cat at different times of the experimental day.
INFLUENCE OF UNSPECIFIC IMPULSES
I59
for some time (the end of tracing 5). The same experiment is reproduced as a whole in tracings 8 and 9. One can see clearly how the amplitude of PR first increased and then diminished in accordance with the cortical reaction to the thalamic nonspecific action. The prolonged depressant after-effect is also evident. This period of consecutive depression coincides with that of the EEG desynchronization. When the slow activity reappeared (right half of tracing 9), the amplitude of PR recovered its initial size (compare with the beginning of tracing 8). Fig. 3 summarizes these results. The facilitatory influence of nonspecific thalamic stimulation upon P R is well seen. It is also clear that PR depends in some way on the phase in which it occurs on the waves elicited by the nonspecific thalamic influence. When PR takes place during the ascending part of a wave, the negative component of PR becomes markedly increased, while the positive component is abolished (tracings 2 and 3). When P R coincides with the crest of the recruiting wave, its negative component is abolished, while its positive phase is considerably increased (tracings 4
Fig. 3 The changes in the cortical PR’s under the influence of repetitive stimulation of zona incerta. In all tracings: upper curves, peripheral part of somatosensory cortex; lower ones, central region. Tracing f. PR’s before the stimulation of zona incerta. Tracings 2 and 3. PR coincides with ascending part of the recruiting wave evoked by frequent stimulation of zona incerta (35-36/sec, 1 msec, 8 V). The amplitudeof its negative component increases, while its positive phase is blocked. Tracings 4 and5. PR coincides with the crest of a recruiting wave: the amplitude of its positive component increases and the negative one is blocked. Calibration of the amplitude, 0.5 mV.
and 5). These effects cannot be explained by pure physical summation of potentials, although to some extent such an assumption could be put forward with respect to the augmentation of the negative component of PR when it coincides with the ascending part of the wave; but it would fail to explain the augmentation of the positive component of PR when the latter coincides with the crest of recruiting wave. 2. Single stimuli applied to thalamic nonspecific nuclei. Experiments with paired stimuli (Narikashvili and Moniava 1959) have shown that a marked facilitation of PR is observed when stimulation of the thalamic nonspecific nucleus (conditioning Referencer p . ian-183
160
S. P. NARIKASHVILI
stimulus) precedes stimulation of the skin by 40-50 msec. This is shown in Fig. 4A: when the time interval between the conditioning and testing stimuli is large (tracing l), the amplitude of PR remains unchanged; but, when the interval is reduced down to 50 msec (tracing 2), both components of PR become markedly augmented. When the interval between the stimuli applied to the thalamic nonspecific nucleus and to the skin is further reduced, PR is either facilitated (not always to an equal degree), or its
Fig. 4 The influence of single, nonspecific, thalamic, or cutaneous conditioning stimulations on cortical PR elicited by a cutaneous stimulation. Upper curves: potentials of middlesuprasylvian gyrus, lower ones : somatosensory cortex. A , conditioning stimuli are delivered to n. ventralis anterior (15 V, 1 msec), testing stimulus to the skin of the forepaw. In spite of some artifacts, more or less facilitated PR’s are well seen. At short intervals the response to the testing stimulus is not blocked. B, the same conditions of recording. Paired single stimuli are delivered to the contralateral forepaw. The same preparation as in A. From the interval of 50 msec down (tracing 4) the blocking of PR begins,
amplitude remains equal to that observed without thalamic conditioning stimulation. Even down to an interval of 5 msec no reduction of PR was observed in any of these experiments; this suggests that there is no block caused by the conditioning nonspecific volley. When both stimuli were applied to the skin of the same preparation, quite different results were observed (Fig. 4 B). Namely, when the interval between the two stimuli
161
INFLUENCE OF UNSPECIFIC IMPULSES
was small, some facilitation of PR (especially of its negative component) was observed (tracings 2 and 3); but, from the interval of 50 msec onwards, the negative component of PR began to diminish (tracing 4). The occurrence of different effects of the two kinds of conditioning stimuli (either nonspecific thalamic or cutaneous) upon the responses to the same testing stimulus (cutaneous) seems to support the idea that the nature of the synaptic connections of the nonspecific and specific fibres with the cortical neurones is different. 3. The part played by the thalami< level in the facilitation of cortical responses. The phenomena described above of the facilitation of cortical specific responses under the action of nonspecific thalamic volleys obviously occur at the cortical level. This follows from the fact that not only are the cortical responses evoked by stimulation of afferent nerves facilitated, but the responses evoked by stimulation of the thalamic relay nucleus are also facilitated. For instance, if stimulation of thalamic nonspecific nuclei is added to cortical PR’s evoked by low frequency stimulations of the n. ventralis postero-lateralis (VPL), the PR’s evoked are considerably altered during the waxing phase of the “recruiting” response (Moniava and Narikashvili 1960). This is seen in Fig. 5 A. 1
hhrw
m
Fig. 5 The modifications of the cortical PR’s under the influence of thalamic nonspecific stimulation ( A ) and under the influence of increasing the intensity of thalamic specific stimulation (B). A, the same conditions of recording as in Fig. 4. N. ventralis postero-lateralis is stimulated by 2 V (0.5 msec, 51 sec) and n. centralis medialis by 5 V (0.5 msec, 8/sec). Tracing 5. Modifications of PR during spontaneous bursts of spindles. B, same preparation. Potentials are recorded from somatosensory cortex. Tracing 1 : N.ventralis postero-lateralis stimulated by 2 V (0.5 msec, 7/sec); tracing2: 3 V, tracing 3: 5 V ; tracing4: 8 V. Details in the text. References P. 180-183
162
S. P. NARIKASHVILI
In each tracing, the upper records represent potentials from the suprasylvian gyrus, while the lower ones represent those from the somatosensory cortex. In tracing 1, the PR’s in the somatosensory cortex are elicited by supraliminal stimulations of the VPL. Responses consisting of a small positive phase followed by an ample negative potential are clearly seen. Tracing 2 shows the reaction of both cortical areas to low frequency (8 per sec) repetitive stimulation of the thalamic nonspecific nucleus. It is well seen that the waxing phase of the recruiting response occurred at first in the somatosensory cortex and then in suprasylvian gyrus. In the following tracings (3,4) stimulation of VPL as in 1 was tested, and specific responses appeared on the background of the recruiting responses. From these oscillograms it is clear that, some time before the onset of the waxing phase in the suprasylvian gyrus, i.e. simultaneously with the onset of this phase in the somatosensory cortex (compare with tracing 2 ) , the PR’s were doubled and an additional negative wave appeared. PR’s with double peaks were recorded during the whole waxing phase of the recruiting responses in the suprasylvian gyrus, and only at the end of this phase, when the recruiting waves markedly diminish, did they return to their shape, that of a single negative wave (end of tracing 4). Thus the onset phase of the responses with double peaks in the somatosensory cortex coincided precisely with the moment when in this area the waxing phase of the recruiting response was elicited during isolated nonspecific thalamic stimulationses (tracing 2). What is the meaning of such a modification of the responses to the stimulation of the thalamic relay nucleus? Fig. 5 B shows that here the complication of the cortical specific responses shows the facilitation, i.e. an increased reactivity of the cortical neurones to the specific volleys. This figure shows the cortical responses evoked by different stimulation intensities applied to the thalamic relay nucleus. In tracing 1 , a liminal stimulus is applied to the VPL. The response, as in fig. 5 A (tracing I), consisted of an initial weak positive component followed by a single negative potential. When the stimulus was increased to a certain degree (tracing 2), a weak additional negative wave appeared in the response. When the intensity of thestimulation wasconsiderablyincreased (tracing3), a well-marked secondary wave appeared. This response resembled the one evoked during the waxing phase of the recruiting response (Fig. 5 A, tracings 3,4). Finally, when the stimulus intensity was further increased, three waves appeared (tracing 4). Thus, the fact that, during the waxing phase of a recruiting wave, the cortical responses evoked by stimulation of a thalamic relay nucleus appeared in the form of double negative potentials indicates without any doubt that the excitability of the cortical neurones responding to thalamic specific volleys was enhanced. Although the data just mentioned, and also those reported by many other authors, speak in favour of an interaction between nonspecific and specific impulses at cortical level, the close localization of the corresponding nuclei at thalamic level suggests the possibility that interactions occur equally well at this level. To test this idea, experiments were carried out in which modifications of the responses in the thalamic relay nucleus during nonspecific thalamic stimulations were studied (Narikashvili and Moniava 1961).
INFLUENCE OF UNSPECIFIC IMPULSES
163
Fig. 6 shows the thalamic response in n. ventralis postero-lateralis (VPL), together with that in the somatosensory cortex, in order that the influence of the recruiting reaction upon these responses can be tested. In most instances we could not detect any marked and regular change of the responses in the thalamic relay nucleus (to the stimulation of the skin) under the influence of the recruiting response. Thus in Fig. 6 some instances (from one experiment) of the facilitation of cortical PR’s are presented without any marked modification of the thalamic responses. In tracing 1 responses before the nonspecific thalamic stimulation are shown; other tracings were obtained during the recruiting response elicited by stimulation of the n. centralis medialis. Approximately the same results were obtained with paired stimuli. 1
Y
r A r ~ n n ~ ~ r r r d
Fig. 6 The influence of recruiting response upon the PR’s of the somatosensory cortex and of the thalamic relay nucleus. Upper curves: potentials of the n. ventralis postero-lateralis, lower ones : those of somatosensory cortex. Responses are evoked by the stimulation of the skin of the contralateral forepaw. The recruiting response is evoked by the stimulation of n. centralis medialis (8 V, 11-12/ sec, 0.5 msec). In tracing 5, potentials of the suprasylvian gyrus are also recorded (the lowest curve). Below, signal lines indicate the stimulation of n. centralis medialis and time (20 msec).
In rare instances, however, a considerable increase of the response amplitude in the thalamic relay nucleus appeared under the influence of a conditioning stimulus applied to the nonspecific thalamic nucleus. Fig. 7A shows that, when the interval between the paired stimuli was progressively shortened, the response amplitude in the thalamic relay nucleus markedly increased (tracings 5-9). At the same time, when the interval between the stimuli was short, a well-marked additional positive wave appeared (tracings 7-9). If both the conditioning and the testing stimuli were applied to the skin, quite different results were obtained. The results of such an experiment are presented in Fig. 7 B. From this figure it is evident that, when the interval between two cutaneous stimuli was shortened, the response to the testing stimulus progressively diminished (as was expected from B2) and was finally blocked. At the same time, the testing stimulus was followed by a slow wave (tracings 6-9). The shorter the interval, the Rrlrrencsr P. 180-183
164
S. P. NARIKASHVILI
higher was the amplitude of this wave. In contradistinction to the experiments in which the stimulation of the thalamic nonspecific nucleus was conditioning, this wave was negative. It is difficult to assess the significance of these slow waves of opposite signs, but it is obvious that they reflect different processes.
Fig. 7 The influence of single nonspecific thalamic and cutaneous conditioning stimulations on responses of the n. ventralis postero-lateralis. A , n. centralis medialis (conditioning stimulation) is stimulated by 10 V (0.5 msec). B, paired single stimuli are delivered to the skin of forepaw. Same preparation. Details in the text.
To sum up, we may conclude that the facilitation of the cortical primary responses is mainly conditioned by the direct action of the thalamic nonspecific volleys upon the cortical neurones. The facilitation of thalamic responses observed in rare instances indicates the possibility of an action of nonspecific volleys at the thalamic level. But, when this is observed (in experiments with paired stimuli), the time interval was much shorter than it was for the facilitation of cortical responses. Therefore this action should not be of major importance for the facilitation of cortical responses. 11. THE INFLUENCE OF RETICULAR STIMULATION UPON RESPONSES OF THE VISUAL SYSTEM
1. Changes of responses evoked by light flushes
( a ) Low frequency lightyushes. If, during the action of low frequency light flashes
165
INFLUENCE OF UNSPECIFIC IMPULSES
(1 per sec and less), when evoked potentials are simultaneously recorded in the visual cortex and in GL and a shortlasting activation of R F is produced by tetanic stimulation of the skin, then responses in both structures become markedly depressed (Narikashvili, Moniava and Kadjaia 1960).This is well seen in Fig. 8, which shows the results of several experiments on different animals. The negative phase was mainly depressed, but the positive components were also markedly reduced. Similar results were observed during direct electrical stimulation of the mesencephalic reticular formation.
2
9 I
\
0.4 mV
% 4
Fig. 8 The influence of afferent activation of RF on responses evoked by rare light flashes. In each tracing (1-4) from top to bottom potentials of GL (l), middle (2) and posterior (3) parts of lateral gyrus are recorded (bipolarly). Signal line indicates nociceptive stimulation of the skin of contralateral forepaw. Details in the text.
A detailed analysis of the reticular effect reveals that in most instances the depression of the responses is more pronounced in the cortex than it is in GL (Figs. 8 and 9). Very often when the cortical responses were considerably depressed, the geniculate responses were slightly reduced or were even not changed at all (Fig. 8). Usually the responses of the optic tract (OT) remained quite unchanged. This latter fact excludes the participation of retinal or pupillary processes (Granit 1955; Hernandez-Peh, Scherrer and Velasco 1956; Hernhdez-Pe6n et at. 1957). Quite infrequently, other correlations have been observed. Cortical responses were, at times, more resistant to the depressant influence of the reticular formation than were those from GL. There were also instances in which the geniculate responses were increased, while the cortical responses became depressed. ( b ) Frequent tight flashes. While reticular stimulation superimposed on low References P. 180-183
I66
S. P. NARIKASHVILI
frequency light flashes ( 1 per sec and less) depressed cortical responses, it exerted an opposite effect when the light flashes were more frequent: the amplitude of the PR’s increased and the responses became more regular (Narikashvili, Butkhusi, Kadjaia and Moniava 1961). Fig. 9 thus shows that, when the flicker frequency was 3 per sec, the characteristic depressive action of the reticular stimulation upon the cortical responses was still well shown, whereas at higher frequencies (5-6 per sec and more) considerable facilitation occurred (tracings 2 and 3), which manifested itself by increased amplitude and regularity of the responses. First of all, the exceptional regularity of the responses was remarkable. The responses followed the frequency of the stimulation. In other words, under the influence of the reticular activation, the ability of the cortical neurones to follow relatively high frequencies of light stimulation
~ 2 m v] 3
6
d
Fig. 9 The influence of reticular stimulation on responses evoked by frequent light flashes. I n each tracing from top to bottom are recorded (monopolarly) potentials of: posterior ( I ) and middle (2) parts of lateral gyrus, G L (3). OT (4) and posterior part of suprasylvian gyrus ( 5 ) . Upper signal line, light flashes; lower one, stimulation ofmesencephalic reticular formation (10 V, 1 OOkec, 1 nisec). Trcrciw I , frequency of light flashes, 3/sec; 2, 6/sec; 3, I I/sec.
INFLUENCE OF UNSPECIFIC IMPULSES
167
was considerably increased. This occurred mainly at the cortical level, because, when the frequency of the light stimulation was progressively increased, the capacity of the cortex to follow the flicker frequency (during the reticular stimulation) often appeared to be higher than that of GL. While the cortical responses were facilitated, the geniculate potentials were not markedly changed or they were even depressed, particularly at the beginning of the reticular stimulation (tracings 2 and 3). In all these instances, the OT responses were not modified, the role of the eye in these phenomena being thus excluded. When the rate of the light flashes was not too high (5-6 per sec), the facilitatory effect of the reticular formation was equally marked during the whole period of the reticular stimulation (Fig. 9, tracing 2). However, at higher frequencies (10-15 per sec) the reticular facilitation was most marked during a short initial period of the reticular stimulation (Fig. 9, tracing 3). It must be pointed out that the reticular facilitation of the cortical responses evoked by frequent light flashes was observed during the entire duration of the light stimulation, but was especially well-marked some time after the beginning of the repetitive light flashes, when the responses became weaker and irregular due to habituation. The fact that the reticular depression of the responses was more marked in the cortex than in GL (during the low frequency light stimulation) and the fact that the cortical and thalamic responses were altered in different ways in different cases, indicates that the modification of the cortical responses is not conditioned by identical changes in the activity in the whole thalamo-cortical system. These facts seem to show that reticular action takes place separately at each level of the afferent system. At any rate, they indicate that the reticular influence upon the cortical neurones occurs regardless of the influence exerted upon the thalamic relay nuclei. All these data induced us to investigate in detail the reticular influence upon the responses evoked by the electrical stimulation of different parts of the visual pathway. 2. The changes in the responses to electrical stimulation of different parts of the visual pathway ( a ) Low frequency stimulation. If OT or GL are stimulated with low frequency single stimuli (1 per sec and less), and if, on this background, R F is activated (by means of its direct electrical stimulation or by means of nociceptive cutaneous stimulations), the cortical responses become considerably augmented in contradistinction to the effect on responses to infrequent light stimulations (Narikashvili, Moniava and Kadjaia 1960). This is shown by the results of the two experiments presented in Fig. 10. In the first experiment (tracings 1-5) stimulation to GL was submaximal. The cortical responses were evoked irregularly and their amplitude was variable. When nociceptive stimulation to skin was added, the responses became considerably augmented, regular and almost absolutely equal in amplitude. This facilitation outlasted the cutaneous stimulation. In tracing 6, the influence of reticular activation upon the response to subliminal stimulation of GL is shown. Before the nociceptive stimulation no responses were evoked in the visual cortex, or sometimes very weak potentials were observed in response to the stimulation of GL. But during the reticular
168
S. P. NARIKASHVILI
activation, regular responses appeared at each stimulation of GL. At the beginning of reticular activation, responses were evoked even in the posterior lateral gyrus, in which responses were not evoked in this case even by submaximal stimulation of GL. One of the most characteristic features of the reticular facilitation of the cortical responses evoked by low frequency electrical stimulation applied to the afferent
2
6 d.
1 sec c
Fig. 10 The influence of reticular activation on cortical responses evoked by rare stimulations of GL. Tracings 1-5 are in direct continuation; from top to bottom are recorded (monopolarly) potentials of: somatosensory cortex (I), middle part of lateral gyrus (2) and suprasylvian gyrus (3). Duration of reticular activation (stimulation of skin 9-l0/sec) is seen by artifacts and responses on a curve of somatosensory cortex. G L is stimulated by 5 V (0.5 msec). In tracing6, same order: suprasylvian gyrus, middle and posterior parts of lateral gyrus. G L is stimulated by 3 V (0.5 msec). Details in the text.
pathway is its longlasting after-effect. Under favourable experimental conditions the shortlasting reticuiar activation (by means of afferent or direct stimulation) gave rise to a longlasting facilitation of the cortical responses to low frequency stimulation of GL (Fig. 10; Fig. 11 C).
169
INFLUENCE OF UNSPECIFIC IMPULSES
The duration of the reticular facilitatory after-effects, and also the intensity of this facilitation (as shown by augmentation of amplitude), obviously depend upon the intensity of the reticular stimulation. Weak stimulation of the skin or of RF may not produce characteristic effects ; but, beyond a definite intensity, the fully developed effect was observed. However, most important is the fact that the reticulareffect (both the augmentation of amplitude and the duration of after-effect) depends to some extent also on the intensity of excitation in the afferent system which undergoes the facilitation (Kadjaia, Moniava and Narikashvili 1961).
-
.
._-_
, . ~, ., _-_.~.. " . . ., . 1~"_._,~ ' " " " " ' " ' " ' ' " " ' ~ " " ' . " . ' . ' ' ' ~ ~ ' ~ ~ ~ ~~
-.
Fig. 11 The duration of after-effect after reticular facilitation at different intensities of stimulation of the visual structure. A . Threshold stimulation. Tracing I : left ( I ) and right (2) middle lateral gyrus, left suprasylvian gyrus (3), left optic tract (4) are bipolarly recorded. Upper signal line indicates the stimulation (2 V, l/sec, 0.5 msec) of left GL, and lower one that of the skin. Tracing 2 : bipolarly are recorded: left (1) and right (2) middle lateral gyrus, left G L (3) and posterior half of left suprasylvian gyms (4). Upper signal line indicates stimulation (4 V, l/sec, 0.5 msec) of opticchiasma and lower one that of the skin. B. Supraliminal stimulation of GL. Tracing 2, direct continuation of tracing 1 ; left GL is stimulated by 4 V. C. Submaximal stirnulation. Tracing 2 and 3, continuation of tracing 1 ; left GL is stimulated by 6 V.
Thus, Fig. 11 shows the result of an experiment performed on the same preparation. GL, or the optic chiasm, was stimulated with stimuli of different intensities and on this background of responses the influence of a uniform optimal reticular activation was tested. If one compares points A, B, C in this figure, one can see that the responses to weak (threshold) stimuli (Fig. 11A) were insignificantly facilitated, an effect which disappeared very soon after cessation of the reticular stimulation. When the stimuli were stronger, the facilitation was more marked and it lasted longer (Fig. 11B). References p . 180-183
170
S . P. NARIKASHVILI
Finally, when some definite submaximal stimulation of GL was used, the facilitation became still greater and was maintained for a long time (Fig. 11C). But, of course, such a dependency does not go beyond certain limits. When the amplitude of the responses reached its maximum, the facilitatory influence of the reticular activation could not be revealed at all (see also Bremer and Stoupel 1959). From these examples, as well as from previous figures, it is evident that, during unilateral cutaneous or reticular stimulation, facilitation appeared and lasted for a long time, not only i n the central visual area, in which the responses were of maximal amplitude, but also in the peripheral area of the same hemisphere and also in the opposite hemisphere. The same figures show that the responses in GL were also facilitated, although considerably less. Thus, in contradistinction to the cortical potentials evoked in response to rare light flashes, the responses to the rare stimulation of different parts of the visual pathways were significantly facilitated for a long time under the influence of reticular activations. This was not caused by different functional states of the preparation or by any other experimental conditions. As Fig. 12 shows, the responses to light flashes weredepressed
a
2 I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I.
I
1
Fig. 12 The influence of reticular stimulation on cortical responses evoked by intermittent light stimulation of the eye and electrical stimulation of GL. From top to bottom, middle part of lateral gyrus (I), OT (2) and posterior suprasylvian gyrus (3) are monopolarly recorded. Stimulation of G L ( 5 V, 0.5 rnsec) is indicated by an upper signal line and stimulation of MRF(IOV, 200persec, 1 msec)bya lower line. Tracing 2, continuation of tracing 1 . Light stimulations are indicated by the responses of OT. Details in the text.
under the influence of the reticular activation, while the responses to the geniculate stimulation were facilitated when the light stimulation of the eye and the electrical stimulation of GL were intermittently produced. At the same time Fig. 12 shows that, after the cessation of reticular stimulation, the cortical responses to the geniculate stimulation remained regular and were facilitated for a long time, but gradually decreased to the initial state (tracing 2). At the same time the responses evoked by light flashes remained also depressed for a long time and gradually regained their
171
INFLUENCE OF UNSPECIFIC IMPULSES
initial amplitude. It is worth mentioning that, after the reticular stirnulation had ceased, the responses to light flashes, when the latter were produced shortly after the geniculate stimulation, appeared to have been enhanced (marked with crosses). But, during the reticular stimulation (tracing I), the responses to light stimulation were not enhanced, even when they were preceded at the same intervals by stimulation of GL (marked with short horizontal lines). In other words the specific facilitation was not apparent during reticular stimulation. (b) Frequent stimuli. Under the influence of reticular activation the responses to frequent stimuli, like those evoked by rare stimuli, became facilitated. This facilitation, like that observed when frequent light flashes were used, consisted in the enhancement of the amplitude of the responses, the latter becoming regular. However, the most characteristic feature of this reticular facilitation was the regularity of the responses. Notwithstanding the fact that sometimes the amplitude of the responses does not increase under the influence of the reticular activation (or may even decrease) it was shown that they begin to emerge regularly with a constant amplitude, and follow the stimulation rate. The results of an experiment with different frequencies of OT stimulation are presented in Fig. 13. At a frequency of 4 per sec (Fig. 13A, tracing 1) reticular activation facilitated the responses in both the visual cortex and in GL. Rut in the cortex
,
-
~
r
I
3
,
- . -.-*--
__ 2
1Sec Y
.
.
,
.
.
. _,. , .,,'"
,,
,
I
,
,
I
. I I
.
, , .
.I.
,
. ..
., .., .. .
, lsec , P
..
_c_
Fig. 13 Reticular facilitation of responses in the visual system evoked by the stimulation of the optic tract (OT). A , in each tracing (1 to 3) from top to bottom posterior ( I ) and middle (2) parts of lateral gyrus, GL (3) and posterior part of suprasylvian gyrus (4) are monopolarly recorded. Upper signal line, stimulation of OT (8 V, 0.5 msec), lower one, stimulation of MRF (10 V, 200/sec, 0.5 msec). B, conditions of recording and stimulation are the same. Details in the text. Referrncer P. 180-183
172
S. P. NARIKASHVILI
the facilitation was much more marked. At 8 per sec (tracing 2) the same phenomenon was observed, but the difference between the cortical and geniculate facilitatory effects was more significant -the facilitation in the thalamic nucleus being much weaker. At 18 per sec (tracing 3) the facilitation was weaker in all parts of the visual system but was well marked in the posterior suprasylvian gyrus. In addition to this, the predominant influence of the reticular activation upon the regularity of the responses was evident in this case (tracing 3, curve 4). This phenomenon is easily observed in the cortical area (posterior part of the suprasylvian gyrus), in which no responses were evoked before the reticular activation, or they were very weak and irregular. At the same time, the activity evoked in this area and that in the lateral gyrus (tracing 2) show clearly that the reticular facilitation was most marked at the beginning of the stimulation and then decreased gradually. At 30 per sec (Fig. 13B, tracing 1) and 50 per sec (tracing 2), the geniculate responses became more enhanced than the cortical ones, but they did not follow the frequency of stimulation in either the cortex or the GL. In the GL a new regular rhythm appeared (10 per sec) in the form of group discharges consisting of 3-4 waves. Thus, according to the frequency of stimulation, the degree of the reticular facilitation and the ability of neurones to follow the rate of stimulation were both changed. The interrelationships between the cortex and the thalamic relay nucleus were also
Fig. 14 The part played by reticular facilitation in the formation of augmenting responses .Tracing I , left (1) and right (3) middle lateral gyrus, right CL (2) are monopolarly recorded. Left G L is stimulated (5 V, d/sec, 0.5 msec), and stimulation of M R F (15 V, 100/sec, 1 msec) is added (lower signal line). Tracing 2, immediate continuation of tracing I . Details in the text.
INFLUENCE OF UNSPECIFIC IMPULSES
173
changed inasmuch as the degree of facilitation of the responses at both levels was different. Facilitation of reticular origin during the frequent electrical stimulation of the visual pathways, as welt as during frequent light flashes, was observed for all periods of the afferent repetitive stimulations. It was, however, most marked after some time with the repetitive stimulation of the visual pathways when, in consequence of habituation, the responses became smaller and irregular. The features just described of the reticular facilitation of the responses in the visual system (enhancement of the responses, their regular emergence and their ability to follow higher stimulation frequencies) undoubtedly indicated an enhanced excitability of the neuronal elements. Because the augmenting response was conditioned by a progressive recruiting of more and more cortical neurones (due to the progressive increase of excitability with successive stimulations), it may be suggested that the reticular stimulation will favour the onset of this reaction. In contradistinction to some observations (Gauthier, Parma and Zanchetti 1956; Landau, Bishop and Clare, 1961), although not always, this may indeed be observed. Fig. 14 illustrates such an observation. Before the reticular activation, waxing and waning P R s were evoked in response to a slow repetitive stimulation of GL (tracing 1). But, evidently because of an insufficient excitability level of the cortical neurones, the period of waxing is not accompanied by the onset of the augmenting responses (Dempsey and Morison 1943). After the introduction of reticular stimulation (lower signal line at the end of tracing l), an augmenting response became evident. However this response was devoid of its characteristic waxing and waning phases. Then, some time after the beginning of reticular stimulation (beginning of tracing 21, the effect disappeared. DISCUSSION
1. On the characteristic features of the nonspecific facilitation Because the reticular formation, in comparison with the thalamic nonspecific structures, exerts a more significant influence on the cortical neurones, it is appropriate to characterize the features of nonspecific facilitation by referring to the effects of reticular stimulation. ( a ) First of all, in all cases of reticular stimulation one observes some augmentation of the cortical responses evoked by electrical stimulation of the afferent pathways. This augmentation is obviously determined, first, by the higher frequency of discharge of the cortical neurones (Li 1956), which can result in an increase of the surface negative potentials ; and then, by the recruitment of new cortical neurones. This recruitment seems to be conditioned by an increased excitability of the cortical neurones (Li 1956), as well as by their excitation under the influence of reticular impulses (Akimoto und Creutzfeldt 1958 ; Creutzfeldt und Akimoto 1958 ; Jung 1958 and others). ( b ) Simultaneously with the increase of their amplitude, the responses become more regular (cJ Gellhorn, Koella and Ballin 1954, 1955; Li 1956). This is one of the most characteristic features of the reticular facilitation, because, even when the augReferences P. 180-183
174
S. P. NARIKASHVILI
mentation is not well manifested (during both the rare and the frequent stimulation of the afferent systems), the responses become regular. Sometimes during reticular facilitation the responses happen to increase in an initial period only and may then return to their initial amplitude, b u t their regularity is usually maintained throughout any long-lasting reticular stimulation and even after its cessation. Reticular volleys thus seem to create more favourable conditions for the synchronous activity of the cortical neurones. If, in the absence of reticular stimulation, the level of the excitability of the different neurones in the sensory cortex undergoes considerable fluctuations, which is reflected in those of the PR’s, during the reticular stimulation the excitability level is stabilized for a long time at a definite value. This is why the reticular facilitation is not observed (even with respect to amplitude) when, before reticular stimulation, the responses are already regular (cf. also Gellhorn, Koella and Ballin 1955). Thus, the most suitable background for the occurrence of reticular faciIitation is that offered by irregular responses, as they appear spontaneously or as the result of habituation after continuous rhythmical stimulation of the afferent system. (c) In all instances of reticular facilitation by low frequency stimulation of the afferent pathways, facilitation, although it gradually weakens, lasts a long time, sometimes for many seconds, after the stimulation has ceased. Usually, it lasts as long as the EEG desynchronization produced by reticular stimulation, i.e. the two events are parallel. However, sometimes the reticular facilitation fails, while the EEG desynchronization is observed. But we have never observed the opposite effect, i.e. facilitation of cortical responses without EEG desynchronization. In certain states of the preparation when cutaneous stimulations did not produce a well-marked EEG desynchronization, facilitation of the cortical responses evoked by stimulation of different sites of the afferent pathways never occurred, nor did their inhibition occur when rare light flashes were applied. We can therefore conclude on the basis of our experiments that both these reactions are closely interconnected, although facilitation of the cortical responses is a more vulnerable reaction than is the EEG desynchronization. Perhaps this is the manifestation of different mechanismsaccording to which the ascending activating system exerts its influences (Dumont and Dell 1958, 1960; Horn 1960). (d) If the effect of the reticular stimulation is tested during frequent (more than 15-20 per sec) light flashes or during frequent stimulations of the visual pathways, it appears that the ability of the cortical neurones to follow high frequencies of stimulation becomes greater than it was before the reticular stimulation (cf. also Steriade and Demetrescu 1960; Sislina and Novikova 1961). This is also well shown by microphysiological experiments (Jung 1958; Creutzfeldt and Grusser 1959). This may be caused by a better and more stable synchronization of neuronal discharges as well as by shorter duration of the responses (Bremer and Stoupel 1959). Unlike the facilitation of the responses evoked by low frequency stimulations of the visual pathway, responses to the frequent stimulation (as well as to frequent flashes, i.e. 10-15 per sec and higher) were not characterized by long-lasting after-effects. After the cessation of a brief reticular stimulation, the facilitated responses evoked by frequent stimulations returned almost immediately to the initial level. When reti-
INFLUENCE OF UNSPECIFIC IMPULSES
175
cular stimulation lasted a long time, the facilitation of the responses soon decreased, in spite of the maintenance of reticular stimulation. In this sense, one may agree with the conclusion of Dumont and Dell (1960) which limited the optimal effect of facilitation to the first hundred of msec, from the beginning of the reticular stimulation. This is also seen from the facilitating influence of reticular volleys on the development of augmenting responses, seen only at the beginning of the reticular stimulation. The fact that, during the short initial period of the reticular stimulation, the responses appeared or became enhanced in cortical regions (e.g. posterior suprasylvian gyrus), where before the reticular stimulation they were weak and irregular, or were even absent, is in favour of the stronger reticular influence at the beginning of stimulation. These data, together with others concerned with the action of thalamic nonspecific impulses, undoubtedly show that the influence of nonspecific impulses is greatest at the beginning of stimulation of nonspecific structures. However, as for rare stimulation (1 per sec) of GL or OT, responses can be quite equally facilitated throughout the very longlasting continuous reticular stimulation, and after the cessation of the stimulation as well. Sometimes, especially during low frequency repetitive paired stimulations (conditioning stimulus being applied to RF), some repeated stimulations during certain periods were necessary to obtain the facilitatory effect, which increased gradually. 2 . The significance of excitation intensity of the afferent system The fact that increasing the intensity of the reticular stimulation causes a clearer manifestation of the effects is well known. However, the degree of reticular facilitation of the cortical responses depends also to a certain extent upon the intensity of the specific afferent stimulation. This refers to the augmentation of the amplitude, as well as to the maintenance of facilitation for a long time during and after reticular stimulation. The responses evoked by subliminal or threshold stimulations of G L or OT are less augmented and have an after-effect of shorter duration (or no after-effect) than the responses evoked by submaximal stimulation. Because augmentation of the amplitude depends upon the quantity of new recruited neurones, it is clear that the stronger the afferent stimulation (within definite limits) the more new neurones will be recruited during the reticular stimulation and the later the responses will return to their initial amplitudes. The same results have been obtained by Gellhorn, Koella and Ballin (1955) in experiments on hypothalamic facilitation of local strychnine discharges provoked by acoustic stimuli.
3. On the levels of nonspecijic f a c i l ~ t a t ~ o n The ratio between the changes in cortical and subcortical responses may be quite different under various conditions (optic or electrical stimulation of the visual tract, frequency of stimulation, location of the electrical stimulation, etc.). Unlike facilitation of the responses to frequent light flashes, when the responses in the visual cortex and in GL are almost equally facilitated, during an electrical stimulation of the visual path, the degree of facilitation of the cortical and geniculate responses depends to some extent upon the intensity and site of the stimulation of the visual pathway. If the References I. 180-183
176
S. P. NARIKASHVILI
responses are evoked by a weak stimulation of OT, the facilitation is more apparent in GL than in the cortex. If, under the same conditions of reticular activation, GL is stimulated with low frequency stimuli, the cortical responses are more greatly facilitated than they are when OT is stimulated. The definite part may be played by the intensity factor at these two levels, these being of some importance, as has already been said, for the determination of the degree of reticular facilitation. In particular, during the stimulation of the optic tract, a greater quantity of neurones may be excited in GL than in the cortex, and because of this the facilitation may be better shown in GL than in the cortex. All these facts, as well as the frequent discrepancies between the degrees of reticular facilitation of the thalamic and the cortical responses, suggest that there is no correlation between the facilitatory influences occurring at either level. The facilitatory effect at each level is determined by local conditions, which may change independently in GL and in the cortex in relation to the nature and the site of stimulation of the afferent pathway, as well as with the frequency of its stimulation. The independence of the cortical and the thalamic levels as regards facilitation is also proved by the following fact, which is observed in all tracings of responses to frequent stimulation of the visual pathway: at the beginning of the reticular stimulation, only the cortical responses are considerably facilitated, while the thalamic responses may even be depressed ; facilitation of the thalamic responses occurs later, when the facilitation of the cortical responses diminishes to some extent. The fact that in the cortex the facilitatory effect is maintained, after the reticular stimulation has ceased, for a longer time than in GL, also indicates some independence of the reticular facilitation at thalamic and cortical levels (Narikashvili et al. 1961). To sum up, one may conclude that the reticular influence is exerted independently at the thalamic and the cortical level. The effects are more important and long-lasting in the cortex than in the thalamic relay nuclei. This is why, in accordance with the findings of other authors (Dumont and Dell 1958,1960; Bremer and Stoupel 1959 and others), it may be assumed that the modifications of the cortical responses are mainly due to processes which develop directly at the cortical level. This is proved by the fact that there is facilitation not only of both the responses evoked by the stimulation of GL and OT but also of the cortical responses to the stimulation of the optic radiation (cf. also Bremer and Stoupel 1959; Landau, Bishop and Clare 1961). 4. Characters of the synaptic connections of the specific and nonspecificjbres The data presented above indicate that the connections of the nonspecific fibres with the neurones, at both the cortical and the thalamic levels, which react to the specific impulses, must differ in character from those of the specific fibres with these same neurones (cf. also Brookhart et al. 1957; Smith and Purpura 1958). This is evident, for instance, from the comparison of the effects of these two kinds of volleys in the experiments with paired stimuli. In these experiments the effect of the conditioning stimulus depends on its being applied either to specific or to nonspecific structures. According to the assumption of some authors (Beritoff 1948, 1956; Chang 1956; Akimoto and Creutzfeldt 1958; Landau et al. 1961), one might suggest that the
INFLUENCE OF UNSPECIFIC IMPULSES
177
nonspecific fibres terminate only on dendrites (axodendritic synapses). However, apparently their connections are not thus limited, because, after local poisoning of the cortical surface with nembutal, which is supposed to depress only the dendritic activity, facilitation of the positive component of the cortical responses under the influence of nonspecific volleys is completely maintained and even enhanced (Bremer andstoupel 1959). In addition to this, it is known that, under the influence of reticular volleys, the dendritic potentials are usually depressed (Purpura and Grundfest 1957; Purpura 1956, 1958; Landau et al. 1961), although occasionally a light facilitation has been observed (Loeb et al. 1961). From this it follows that, either the negative component of the responses and the dendritic potential evoked by direct stimulation of the cortical surface do not represent quite identical events, or the reticular facilitation which manifests itself in most instances by an augmentation of the negative component of the cortical response, is not related to, and does not depend upon, the action of nonspecific impulses on the dendrites. Apparently, the second assumption and is more correct ( c - Bremer and Stoupel 1959). If it is, it would be possible to speak of axosomatic connections for the nonspecific fibres. The possibility of provoking the discharges of the neuroneundertheinfluence ofnonspecificimpulses(Jungetal. 1957), as well as the facilitation of the positive component of PR (Narikashvili 1957; Narikashvili and Moniava 1957; Dumont and Dell 1958, 1960; Bremer and Stoupel 1959), seem to favour this suggestion. However, some facts seem to contradict it (Li 1956): the relatively long and variable latency of the neuronal discharge (indicating a polysynaptic pathway) evoked by nonspecific stimulation (Creutzfeldt and Baumgartner 1955; Akimoto and Creutzfeldt 1958; Morillo 1961), different types of reactions of the cortical neurones to nonspecificand specific impulses (Jung et al. 1957), and also the weaker ability of the cortical neurones to reproduce the frequency of the nonspecific stimulation (Creutzfeldt and Akimoto 1958). We are thus led to think that the nonspecific fibres do not terminate directly on those neurones to which the specific impulses are addressed, i.e. on the pyramidal cells of HI-IV layers (cf. Li 1956; Jasper 1958; Bremerand Stoupel 1959). The nonspecific fibres, terminating in all the layers of the cortex (Lorente de N6 1951), are probably connected to internuncial neurones which receive the afferent impulses (neurones with short axons : the short pyramids, short spindle-cells, star cells - Lorente de N6 1951). Excitation of these interneurones, in relation to the degree of their activation as well as to the functional state of pyramidal neurones, can increase the excitability, o r under certain conditions can cause the pyramidal neurones to discharge. Apparently this may be accomplished also by electrotonic action. 5. The mechanism of nonspecijic facilitation and depression of cortical responses Interpretations of the facilitatory or depressive influences on different structures are usually based on an increase or decrease of the responses evoked. However, when potentials are recorded with macroelectrodes it is necessary to take into consideration different factors which can mask the true character of the action. An increase of the evoked potentials, particularly of their negative components, under the influence of nonspecific volleys, simultaneously with an acceleration of the neuronal discharges References P. 180-183
178
S. P. NARIKASHVI1.1
(Li 1956; Jung et al. 1957) is caused by the recruiting of new neurones (Li 1956; Bremer and Stoupel 1959), but the amplitude of the response also depends on the degree of synchrony of the unit activities taking place in the neurones under the macroelectrode. The coincidence of all the factors which determine the augmentation of the amplitude surely depends upon the level of excitability or, in other words, of “readiness” of all the cortical neurones on which the afferent fibres terminate. It is known that nearly one third of the neurones in the visual cortex are not excited in response to the specific afferent volleys (Jung et a/. 1957). Apparently, this is due to their low excitability, but some neurones may not, under other conditions, respond to the specific volleys, because they are under the influence of impulses from other sources, i.e. from other parts of the nervous system. Neurones, particularly those in the visual cortex, are not separated from other sources of influence and their response is not limited to responses to impulses coming from the retinae. Not all the sources of these impulses are yet known, but there are certain data which indicate that quite various nonspecific and specific impulses (of other origin than from the retina; Jung 1960) do act upon these neurones. Even diffuse illumination and cessation of the retinal activity of the dark-adapted eye may significantly increase the amplitude of cortical responses which are evoked by stimulation of subcortical regions of the visual pathway (Chang 1952; Dumont and Deli 1960; Posternak et al. 1959; Arduini and Hirao 1959,1960; Bremer et al. 1960). In other words, because such complex interactions of afferent impulses from different origins can markedly influence the “readiness” of cortical neurones to respond to adequate volleys, it is not easy to solve unmistakably the problem of the exact significance of the augmentation or reduction of evoked responses which we have analyzed here. Do they indicate real facilitation or inhibition of the cortical neurones? As an example, one may consider the influence of diffuse illumination of the eye on the cortical responses, as it is evoked by stimulation of subcortical parts of the visual path. The augmentation of the cortical responses during diffuse illumination of the eyes was considered as an effect of direct facilitation. It became clear, however, from the findings of Posternak et al. (1959), Arduini and Hirao (1959, 1960), that the augmentation of the responses is conditioned, not by some direct facilitatory influence of diffuse illumination, but by the elimination of an inhibitory or occlusive action, which normally occurs in the cortex under the influence of a steady bombardment from the dark-adapted eyes. Actually it is a case of facilitation, involving more neurones, but its mechanism is different from that which had been earlier proposed. The above considerations must be taken into account when we are trying to understand the opposite effects of reticular excitation on the responses evoked either by light flashes or by direct stimulation of different regions of the visual pathways. The suppression of the cortical responses evoked by rare light flashes is not the result of a direct depressive (inhibitory) action of reticular impulses on cortical neurones, because, in all other instances, even during the action of the same light flashes but at higher frequencies (more than 5 per sec), a facilitatory action is observed, which is manifested, not only by an augmentation of the responses, but also by other signs (regularity of the responses, more rapid succession of the different components of the
INFLUENCE OF UNSPECIFIC lMPULSES
179
response; Bremer and Stoupel 1959). How then is the reduction of amplitude of the responses evoked by rare light flashes to be interpreted? In this we share Bremer’s opinion that there is some masking effect, considering it as the most probable mechanism (Bremer and Stoupel 1959). But then how can one explain the facilitation of the responses in the case of frequent light stimulations? Bremer’s suggestion (196 1) seems to be the most suitable explanation of this also. If the volleys generated by low frequency flashes, cannot, because of their asynchrony, overcome the relative refractoriness of cortical neurones receiving reticular impulses, volleys elicited during frequent light flashes arrive at the cortex more synchronously and enable them to recruit more neurones which were “unready to respond”. However, there remain some difficulties regarding this problem which must not be disregarded. Thus, the frequent submaximal electric stimulation of the visual structures, which somehow imitates the reticular influence in the sense that it induces relative refractoriness in the cortical receiving neurones, does not depress the responses to light flashes. Only intensive and high frequency stimulation of the visual pathway exciting probably all the cortical receiving neurones depressed the responses to light flashes (Moniava et al. 1961). But this is not so during the reticular influence. Converging evidence shows that, under the influence of reticular stimulation, only the excitability of the cortical receiving neurones is changed (increased). Only a part of these can be excited (Morillo 1961). Thus the state most similar to the one established in the visual cortex during reticular stimulation (in the sense of refractoriness of receiving neurones) is that obtained during moderate strength and frequency of stimulation of the visual pathways. Under these conditions the cortical responses to light flashes can only be facilitated. However, the imitation of the reticular influence in the way mentioned above has not been quite successful. A different type of experimental method should be sought for a solution of this problem. SUMMARY
1. Under the influence of thalamic unspecific impulses, primary responses of the sensory cortex evoked by peripheral stimulation or stimulation of a thalamic relay nucleus become markedly facilitated. This is observed during the recording of primary responses on a background of recruiting responses, as well as after a single conditioning nonspecific stimulation. It is chiefly the negative component of primary responses that is increased, but the positive component can be increased as well. The facilitation is better manifested at the onset of the nonspecific thalamic stimulation. During Ionglasting stimulation it gradually decreased to complete disappearance, or even to depression. Facilitation takes place mainly at the cortical level. 2. Under the influence of reticular volleys, primary responses in the sensory cortex evoked by stimulation of different parts of the afferent pathways become markedly facilitated. Facilitation is observed during repetitive stimulation of the reticular formation, as well as after single shocks. Cortical responses evoked by frequent adequate stimulation of the receptor also become facilitated, while the responses evoked by rare adequate stimulation of the receptor become depressed. The facilitation is maniRrferrnccr D . 180-193
180
S. P. NARIKASHVILI
fested, not only by an augmentation of the amplitude of responses, but also by a better regularity in their occurrence. A longlasting facilitatory after-effect associated with EEG desynchronization is shown during stimulation of afferent structures at low frequencies. This is a central phenomenon which is not associated with modifications in the peripheral receptor. 3. The degree of reticular facilitation in the sensory cortex, as wellas theduration of the after-effect, depend to some extent upon the intensity of stimulation applied to the afferent pathway itself. 4 . The possible mechanisms and the level of the reticular formation involved in the development of the observed phenomena are discussed. REFERENCES AKIMOTO,H. und CREUTZFELDT, 0. Beeinflussung von Neuronenentladungen der Hirnrinde durch das unspezifische Aktivierungssystem des Thalamus. Klin. Wschr., 1957,35: 199. H. und CREUTZFELDT, 0. Reactionen von Neuronen des optischen Cortex nach elektrischer AKIMOTO, Reizung unspezifischer Thalamuskerne. Arch. Psychiat. Nervenkr., 1958,196: 494-520. S. Reticular influence upon thalamic and cortical APPELBERG, B., KITCHELL, R. L. and LANDCREN, potentials evoked by stimulation of the cat’s tongue. Actaphysiol. scand., 1959,45: 48-71. ARDUINI,A. and HIRAO,T. On the mechanism of the EEG sleep patterns elicited by acutevisual deafferentation. Arch. ital. Biol., 1959, 97: 140-155. ARDUINI, A. and HIRAO,T. Enhancement of evoked responses in the visual system during reversible retinal inactivation. Arch. ital. Biol., 1960,98: 182-205. BAUMGARTNER, G. Mechanisms of reaction of single neurones of the visual cortex in the cat. Electroenceph. clin. Neurophysiol., 1958, 10: 195-196. BERITOFF, J. S. Handhook of general physiology of muscle and nervous system, Vol. 2 . MoscowLeningrad, 1948. BERITOFF,J . S. Morphological and physiological bases of temporary connections in the cerebral cortex. Trans.Inst. physiol. Georgian Acad. Sci., 1956, 10: 3-72. BREMER F. Some problems in Neurophysiology. The Athlone Press, London, 1953. BREMERF. The neurophysiological problems of sleep. In: Brain mechanisms and consciousneJs.Ch. C. Thomas, Springfield, 1954: 137-158. BREMER,F. Personal communication. 1961. BREMER F. et BONNET, V. Interpretation des reactions rythmiques prolongkes des aires sensorielles de I’ecorce ctrebrale. Electroenceph. clin. Neurophysiol., 1950, 2 : 389400. BREMER, F. et STOUPEL, N. Facilitation et inhibition des potentielsevoqu~scorticauxdansl’eveilcerebral. Arch. int. Physiol. 1959a, 6 7 : 1-37. BREMER, F. et STOUPEL, N. Etude pharmacologique de la facilitation des reponses corticales dans l’eveil reticulaire. Arch. int. Pharmacodyn., 1959b, 122: 234-248. BREMER, F. et STOUPEL, N. Discussion du mecanisme de la facilitation reticulaire des potentiels evoques corticaux. J. Physiol. (Paris), I959c, 51: 420421. BREMER, F., STOUPEL, N. et VAN REETH,P. CH. Nouvelles recherches sur la facilitation et I’inhibition des potentiels evoques corticaux dans I’eveil reticulaire. Arch. rtal. Biol., 1960, 98: 229-247. BROOKHART, J. M., ARDUINI,A,, MANCIA,M. and MORUZZI,G. Mutual facilitation of cortical responses to thalamic stimulation. Arch. ital. Biol., 1957, 95: 149-164. BUSER,P. et BORENSTEIN,P. Reponses somesthesiques, visuelles et auditives, recueillies au niveau du cortex “associatif” suprasylvien chez le chat curarise. Electroenceph. clin. Neurophysiol., 1959, 1 I : 285-304. CHANG H.-T. Cortical response to stimulation of lateral geniculate body and the potentiation thereof by continuous illumination of retina. J. Neurophysiol., 1952, 15: 5-26. CHANGH.-T. Pericorpuscular and paradendritic excitation in the central nervous system. In :Problems of the modern physiology of /he nervous and muscle systems. Tbilisi, 1956: 43-50. CREUTZFELDT, 0. und AKIMOTO, H. Konvergenz und gegenseitige Beeinflussung von lmpulsen aus der Retina und den unspezifischen Thalamuskernen an einzelnen Neuronen des optischen Cortex. Arch. Psychiat. Nervenkr., 1958, 196: 520-538.
INFLUENCE OF UNSPECIFIC IMPULSES
181
CREUTZFELDT, 0. and BAUMGARTNER, G. Reactions of neurones in the occipital cortex to electrical stimuli applied to the intralaminar thalamus. Electroenceph. clin. Neurophysiol., 1955, 7: 664-665. CREUTZFELDT, 0.und GRUSSER, 0.-J. Beeinflussung der Flimmerreaktion einzelner corticaler Neurone durch elektrische Reize unspezifischer Thalarnuskerne. Proc. Zst Znternat. Congr. Neurol. Sci., 1959, 3: 349-355. CREUTZFELDT, O., BAUMGARTNER, G . and JUNG,R. Convergence of specific and unspecific afferent impulses on neurons of the visual cortex. Electroenceph. clin. Neurophysiol., 1956, 8 : 163-164. DEMPSEY, E. W. and MORISON, R. S. The production of rhythmically recurrent cortical potentials after localized thalamic stimulation. Amer. J. Physiol., 1942a, 135: 293-300. DEMPSEY, E. W. and MORISON, R. S. The interaction of certain spontaneous and induced cortical potentials. Amer. J . Physiol., 1942b, 135: 301-308. DEMPSEY, E. W. and MORISON, R. S. The electrical activity of a thalamocortical relay system. Amer. J. Physiol., 1943, 138: 283-296. DESMEDT, J. E. and LA GRUTTA, G. The effect of selective inhibition of pseudocholinesterase on the spontaneous and evoked activity of the cat’s cerebral cortex. J. Physiol. (London), 1957, 136: 2 0 4 0 . DUMONT, S. et DELL,P. Facilitations specifiques et non-specifiques des rkponses visuelles corticales. J. Physiol. (Paris), 1958, 50: 261-264. DUMONT, S. et DELL,P. Facilitation reticulaire des mkanismes visuels corticaux, Ekctroenceph. din. Neurophysiol., 1960, 12: 769-796. GAUTAIER, C., PARMA, M. and ZANCHETTI, A. Effect of electrocortical arousal upon development and configuration of specific evoked potentials. Electroenceph. clin. Neurophysiol., 1956, 8 : 237-243. GELLHORN. E. On the physiological action of carbon dioxide on cortex and hypothalamus, Electroenceph. clin. Neurophysiol., 1953,5: 401-413. GELLHORN, E., KOELLA, W. P. and BALLIN, H. M.1nteraction on cerebral cortex of acoustic or optic with nociceptive impulses: the problem of consciousness. J. Neurophysiol., 1954, 17: 14-21. GELLHORN, E., KOELLA, W. P. and BALLIN, H. M. The influence of hypothalamic stimulation on evoked cortical potentials. Amer. J. Psycho/., 1955, 39: 17-88. GRANIT, R. Centrifugal and antidromic effects on the ganglion cells of retina. J. Neurophysiol., 1955a, 18: 388411. GRANIT, R. Receptors and sensory perception. Yale Univ. Press, New Haven, 1955b. A., GRUSSER, 0.-J. und BAUMGARTNER, G. Reaktionen einzelner Neurone im optischen GRUTZNER, Cortex der Katze nach elektrischer Reizung des Nervus opticus. Arch. Psychiat. Nervenkr., 1958, 197: 377404. HERNANDEZ-PEON, R., SCHERRER, H. and VELASCO, M. Central influences on afferent conduction in the somatic and visual pathways. Acta neurol. latinoamer., 1956, 2: 8-22. HERNANDEZ-PEON, R., GUZMAN-FLORES, C., ALCARAZ, M. and FERNANDEZ-GUARDIOLA, A. Sensory transmission in visual pathway during “attention” in unanesthetized cats. Acta neurol. latinoamer., 1957,3: 1-8. HERNANDEZ-PEON, R., GUZMAN-FLORES, C., ALCARAZ, M. and FERNANDEZ-GUARDIOLA, A. Habituation in the visual pathway. Acta neurol. latinoamer., 1958, 4 : 121-129. HORN,G. Electrical activity of the cerebral cortex of the unanesthetized cat during attentive behaviour. Brain, 1960,83: 51-16. HORN,G. and BLUNDELL, J. Evoked potentials in visual cortex of the unanaesthetized cat. Nature, 1959, 184: 59-60. JASPER, H. H. Diffuse projection systems : the integrative action of the thalamic reticular system. Electroenceph. elin. Neurophysiol., 1949, 1: 40541 9. JASPER, H. H. Recent advances in our understanding of ascending activities of the reticular system. In: Reticular Formation of the Brain. Little, Brown and Co., Boston-Toronto, 1958: 319-331. JASPER,H. H. and AJMONEMARSAN, C. Tlialamo-cortical integrating mechanisms. Res. Publ. ASS. nerv. ment. Dis., 1952,30: 493-512. JUNG,R. Coordination of specific and nonspecific afferent impulses at single neurons of the visual cortex. In: Reticular Formation of the Brain. Little, Brown and Co., Boston-Toronto, 1958: 423-434. JUNG,R. Microphysiologie corticaler Neurone: ein Beitrag zur Koordination der Hirnrinde und des visualen Systems. In: Structure and Function of the cerebral cortex. Elsevier, Amsterdam-Princeton, 1960: 204-233. G. Hemmungsmechanismen und bremsende Stabilisierung an einzelnen JUNG,R. und BAUMGARTNER, Neuronen des optischen Cortex. Ein Beitrag zur Koordination corticaler Erregungsvorgange. Pflugers Arch. ges. Physiol., 1955,261; 434456.
I82
S. P. NARIKASHVILI
JUNG, R., CREUTZFELDT, 0. and GROSSER,0 . - J . The microphysiology of cortical neurons. Germ. med. Mth., 1958, 3 (9): 269-276. JUNG, R , CREUTZFELDT, 0. und GRUSSER, 0.-J., Die Microphysiologie corticaler Neurone und ihre Bedeutung fur die Sinnes- und Hirnfunktionen. Dtsch. med. Wschr., 1957, 82: 1050-1059. KADJAIA, D. V., MONIAVA, E. S. and NARIKASHVILI, S. P. On the role of intensity of the excitation of visual afferent system in reticular facilitation. Bull. Georgian Acad. Sci., 1961, 27: 85-92. LANDAU,W. M., BISHOP,G. H. and CLARE,M. H. The interactions of several varieties of evoked response in visual and association cortex of the cat. Electroenceph. clin. Neurophysiol., 1961, 13: 43-53.
LI, C.-L. The facilitatory effect of stimulation of an unspecific thalamic nucleus on cortical sensory neuronal responses. J . Physiul. (London), 1956,131: 115-124. LOEB,C., MASSAZZA, G. and STACCHINI, G . Effect of reticular arousal upon the direct cortical response. Electroenceph. elin. Neurophysiul., 1961, 13: 130. LONG,R. G . Modification of sensory mechanisms by subcortical structures. J. Neirruphysiul., 1959, 22: 412427.
LORENTE DE NO, In: Fultun's Physiology of the Nervuus System. Oxford Univ. Press, New York. 1951 : 228-3 12. MANCIA,M., MEULDERS, M. and SANTIBAI~EZ, G . Changes of photically evoked potentials in the visual pathway of the cerveau isole cat. Arch. ital. Biol., 1959, 97: 378-398. MONIAVA, E. S. and NARIKASHVILI, S. P. On the influence of thalamic nonspecific nuclei upon the cortical effects of specific nuclei. Bull. Georgian Acad. Sci., 1960, 25: 605-612. MONIAVA, E. S., KADJAIA, D. V. and NARIKASHVILI, S. P. On the mechanism of reticular influence upon the responses of visual cortex. Pavluv J. higher nerv. Activ. 1961. MORILLO, A. Microelectrode analysis of some functional characteristics and interrelationships of specific, association and non-specific thalamo-cortical systems. Electroenceph. clin. Neurophysiul., 1961,13: 9-20. MORISON,R. S. and DEMPSEY, E. W. A study of thalanio-cortical relations. Amer. 1. Physiul., 1942, 135: 281-292. MORISON, R. S., DEMPSEY, E. W. and MORISON, B. R. Cortical responses from electrical stimulation of the brain stem. Amer. J. Physiul., 1941, 131: 732-743. MORISON, R. S., FINLEY, K. H. and LOTHROP, G. N. Influence of basal forebrain areas on the electrocorticogram. Amer. J . Physiul., 1943, 139: 410416. MORUZZI,G . Influence of brain stem reticular formation on cortical electrical activity. Electroenceph. clin. Neurophysiul., 1949, I : 519. MORUZZI, G. and MAGOUN,H . W. Brainstem reticularformation andactivationof the EEG. Electroenceph. din. Neurophysiul., 1949, I : 455473. MORUZZI, G . , BROOKHART, J. M., NIEMER, W. T. and MAGOUN,H. W. Augmentation of evoked electrical activity during spindle bursts. Electruenceph. clin. Neuruphysiul., 1950, 2: 29-3 1. MURPHY,J. P. and GELLHORN, E. The influence of hypothalamic stimulation on cortically induced movements and on action potentials of the cortex. J. Neurophysiul., 1945,8: 341-364. NAKAO,H. and KOELLA, W. P. Influence of nociceptive stimuli on evoked subcortical and cortical potentials in cat. J . Neuruphysiul., 1956, 19: 187-195. NAQUET, R., REGIS,H., FISCHER-WILLIAMS, M. and FERNANDEZ-GUARDIOLA, A. Variations in the responses evoked by light along the specific pathways. Brain, 1960, 83: 52-56. NARIKASHVILI, S. P. Primary responses and spontaneous electrical activity of cerebrum. Sechenuv Physiul. J . U.S.S.R., I 957, 43 : 642-650. NARIKASHVILI, S. P. and MONIAVA, E. S. On the interaction between diffuse and specific thalamocortical projection systems. Bull. Georgian Acad. Sci., 1957, 19: 341-354. NARIKASHVILI, S. P. and MONIAVA, E. S. On the interaction between the thalamo-cortical projection systems. Pavlov J. higher nerv. Activ., 1959, 9: 461469. NARIKASHVILI, S. P. and MONIAVA, E. S. On the interaction between thalamic specific and nonspecific nuclei. Trans. Inst. physiul. Georgian Acad. Sci., 1961, 12: 55-67. NARIKASHVILI, S. P., MONIAVA, E. S. and KADJAIA, D. V. On the mechanism of interaction between analyzers. Dukl. Akad. Nauk SSSR (Report Acad. Sci. USSR), 1960,134:229-232. NARIKASHVILI, S. P., BUTKHUSI, S. M., KADJAIA, D. V. and MONIAVA, E. S. On some peculiarities of reticular facilitation of visual responses. Trans. Inst. physiol. Georgian Acad. Sci., 1961, 13: in press. POSTERNAK, J. M., FLEMING, R . C. and EVARTS, E. W. Effect of interruption of the visual pathway on the response to geniculate stimulation. Science, 1959, 129: 39-40.
INFLUENCE OF UNSPECIFIC IMPULSES
183
PURPURA, D . P. Observations on the cortical mechanisms of EEG activation accompanying behavior arousal. Science, 1956, 123: 804. PURPURA, D. P. Organization of excitatory and inhibitory synaptic electrogenesis in the cerebral cortex. In: Reticular Formation of the Brain. Little, Brown and Co.,Boston-Toronto, 1958:435457. PURPURA, D. P. and GRUNDFEST, H. Physiological and pharmacological consequences of different synaptic organizations in cerebral and cerebellar cortex of cat. J. Neurophysiol., 1957, 20 : 494-522. SCHOOLMAN, A. and EVARTS, E. V. Responses to lateral geniculate radiation stimulation in cats with implanted electrodes. J . Neurophysiol,, 1959, 22: 112-129. SISLINA, N. N. and NOVIKOVA, L. A. Study of the role of specific and non-specific afferent systemsin ability of cortex to follow frequent stimulation. Sechenov Physiol. f., U.S.S.R., 1961. SMITH, T. G., JR. and PURPURA, D. P. lnteraction patterns of evoked cortical responses and focal paroxysmal discharges. Electroenceph. d i n . Neurophysiol., 1958, 10: 363. STERIADE, M. and DEMETRESCU, M. Facilitation of photic driving by stimulation of the non-specific activating system. Electroenceph. elin. Neurophysiol., 1960, 12: 750. STERIADE, M. and DEMETRESCU, M. Levels of facilitation and occlusion of photic evoked potentials by stimulation of unspecific arousal system. Electroenceph. clin. Neurophysiol., 1960, 12: 750. STERIADE, M. et DEMETRESCU, M. Phenomenes dedynamogenese reticulaire aux diversniveauxdesvoies optiques pendant la stimulation lumineuse intermittente. J. Physrol. (Paris), 1960, 52, 224-225. STERIADE, M. and DEMETRESCU, M. Unspecific systems of inhibition and facilitation of potentials evoked by intermittent light. J . Neurophysiol., 1960, 23: 602-617. STOUPEL, N. Etude de l'interaction dans l'ecorce d'influx thalamo-corticaux specifiques et non-spkifiques. Acta neurol. belg., 1958, 58: 759-771.
The Tonic Discharge of the Retina and its Central Effects* A. ARDUINI Physiological Institute of the University ofPisa and Neurophysiological Centre of the National Research Council, Pisa (Italy)
INTRODUCTION
Foreword This paper deals with the so-called “spontaneous” activity of the retina and discusses the results of a series of observations made from 1958 to 1961. The research, which dealt with the cortical effects of the tonic retinal discharge, was begun in collaboration with Dr. T. Hirao (Arduini and Hirao 1959, 1960a, b); it was continued in collaboration with Dr. Goldstein on the problem of the retinofugal activity impinging upon the lateral geniculate body (Arduini and Goldstein 1960, 1961 ; Goldstein and Arduini 1961). These aspects were discussed in a review presented at the Symposium on the Physiology of the Visual System held in Freiburg in 1960 (Arduini 1961). Last year our attention was transferred to the eye, and an attempt was made to relate the peripheral aspects of the retinal discharge to the results obtained at the level of the visual pathways and to the effects upon the EEC. Our studies have been concerned with fibre recordings from the optic chiasma with gross electrodes in collaboration with Dr. Pinneo, and from the retinal papilla with semimicroelectrodes in collaboration with Dr. Cavaggioni. This investigation is still under way. However, enough information is now available (Arduini and Pinneo 1961 ; Arduini and Cavaggioni 1961 ; Arduini and Pinneo 1962a, b), to present a rough picture of the behaviour of the retina in dark adaptation and of the changes in the adaptation level brought about by steady illumination. Background of the problem For a full bibliographic account of the so-called “spontaneous” activity of the retina, monographs on this subject (Granit 1947, 1955) should be consulted. The aim of this paper is to deal only with certain aspects of the problem - some facts about which are partly already well known - whose importance in visual and extra-visual functions has been somewhat overlooked. In terms of the electrophysiological evidence (Adrian and Matthews 1927a, b, 1928)
* The researches reviewed in this article have been sponsored jointly by theoffice of Scientific Research of the Air Research and Development Command, United States Air Force, through its European Office,contract No. A F 61(052)-107 and by the Rockefeller Foundation.
CENTRAL EFFECTS OF TONIC RETINAL DISCHARGE
185
it seems to be fairly well established that, in vertebrate eyes at any rate, the retina is continuously active, whether or not the receptors receive light. The state of continuous activity raises two problems : 1. how this continuous “rest” discharge changes with light from dark adaptation; 2. how any particular visual message is relayed in the presence of continuous activity. Because microelectrode studies have drawn attention to the events in the first milliseconds following short stimuli, these questions have been mostly left unanswered, although the few observations quoted in the literature give a clue to the solution. It has been reported that synchronized rhythms appear during illumination in the optic nerve of the eel (Adrian and Matthews 1928) and in the optic ganglion of the water-beetle (Adrian 1937). These observations were somewhat misleading and contrasted with reports of the disappearance of the sychronous bursts in the frog and cat optic fibres during illumination (Barlow 1953a; Kuffler 1953). However, the extensive work of Granit, which systematized the observations on frog and cat retinas (Granit 1947, 1955), clearly provides evidence in favour of the major point to be discussed in this paper. To quote directly: “A large number, perhaps most, of the ganglion cells maintains a spontaneous discharge which fluctuates somewhat but in general tends to increase a great deal in dark adaptation” (Granit 1955, p. 84), and again: “It is far more common however, to find light to excite or inhibit the discharge, according to stimulus strength and the nature of the element, with an overall effect of considerable depression of the rate of spontaneous firing in the light adapted state. In light adaptation fewer elements seem to be spontaneously active than in dark adaptation” (Granit 1955, p. 85). Thus, the effects of light at the very onset are clearly stimulating, at least in a large proportion of the retinal elements, whereas sustained illumination, as in light adaptation, produces an over-all depression of the retinofugal discharge. Perhaps these effects would have appeared even more strikingly had observations been conducted for a longer time on each ganglion cell (Granit 1940). This was the principal aim of Kuffler et al. (1957), who followed the behaviour of isolated retinal elements during dark adaptation and under a long-lasting (30-45 min) illumination at different intensities. To these authors we owe the open statement, implicit in Granit’s words, that the behaviour of a given unit under steady illumination is independent of its being an on or an off element. A second finding of these authors is that continuous illumination did not depress the activity of all the retinal ganglion cells, because the firing rate of some elements was actually increasing. Their results are, therefore, a partial answer to the questions we asked before: 1.: in day-light the over-all retinal output is less than in darkness; 2. a number of channels (nerve fibres) is free, or at least less busy, so that more sensory messages can be relayed. Stated in this way the door is open to another series of problems, related to signalto-noise discrimination, absolute and differential thresholds, the origin of tonic activity, dark light, and other psychophysiologically determined phenomena. These problems will be discussed in part in the last section of this paper, together with our own findings. Though these results are helpful, an important aspect of knowledge about retinal function is still missing, namely the aspect concerned with the behaviour of the whole References 8. 201-203
186
A . ARDUlNl
retinal output, that is to say, the output of the ensemble of ganglion cells in dark and light adaptation. Nor have we any quantitative information about this phenomenon, which would permit us to compare and make predictions about different conditions of illumination. The only quantitative approach to the retinal output is the statistical treatment of the ganglion cell discharge made by Kuffler et a/. (1957) and by Fitzhugh (1957). Even these were made on small samples. In reviewing the exhaustive work on retinal physiology, one is immediately aware of the fact that the technique of using fine microelectrode recording, which has been generally employed, did not allow one to gather information on the distribution of the activity among the different cell populations. To obtain data on the behaviour of a large population of retinal neurones, other methods must be applied, which provide reliable measures of the activity of a large number of retinal elements in each sample, without completely destroying the information related to the isolated components. This has been part of our task. However, before describing our methods, we must introduce the working hypothesis we have tried to test. On terminology
A comparison can be made with the afferents of the muscular system, from which the attribute “tonic” is borrowed. Signals are continuously discharged from many sensory receptors, apparently in a random fashion, and it follows that a specific mode of excitation of the receptors brings about a “patterning” of the already existing outflow, reproducing in a suitable code the energy patterns acting on the receptors. Available evidence gives the same picture for the retina as for other receptor systems. Together with the apparent randomness of the discharge, this has contributed to the widespread belief that the retinal tonic activity, in the absence of receptor stimulation, is to be considered as the same form of nuisance that communication engineers are faced with when they are dealing with the transmission of information, that is to say, pure random noise in the channel of communication. In this view “noise” has the connotation “unwanted information”, implying really that the physiological meaning of the continuous discharge is utterly unknown. We will try to show in the next sections that what has been regarded as noise is actually live information, and that it is not unwanted, but is needed for the actual act of vision. (a) The concept of tonic activity. When one is dealing with continuous activity in darkness, one is tempted to use the term “spontaneous”, which is so widely applied (see Granit 1955; Kuffler 1953). However, in this paper this term has been purposely avoided because 1. a more specific word is preferred, and 2. the term “spontaneous” is ambiguous; it includes, not only the activity in dark adaptation, but, paradoxically, also the activity going on during long-term steady illumination. Instead of the term “spontaneous” we shall use the word “tonic”, giving it the specific meaning which we shall now define. The introduction of this term was dictated by something more than a simple problem of finding a new label. It would appear that the word “spontaneous” can be applied safely to the activity of the retina in dark adaptation, that is to say, to the dark discharge (Arduini and Hirao 1959). The term would lose its meaning if it were applied to the continuous activity of the retina during steady
CENTRAL EFFECTS OF TONIC RETINAL DISCHARGE
187
illumination, because then a stimulus is present. However, during steady illumination, the modifications of the retinofugal discharge can be easily separated from the changes occurring at the onset of the stimulus. The problem is exactly the same when one considers the modification of the retinofugal activity during dark adaptation. In the course of this process, the discharge of the ganglion cells undergoes changes which cannot be mistaken for those which constitute the of effect. The two situations, light and dark adaptation, have in common the property of being beyond the transient state induced by the onset or the offset of illumination. Therefore, we shall use the term “tonic” to mean every state of the retinofugal discharge which is not phasic. (b) Tonic and phasic. The attribute “tonic” has always been applied in opposition to “phasic” and it implies a non-transient condition. One may immediately ask the question: when does the activity cease to be phasic (that is, transient), and acquire the character of tonic (that is, non-transient). It can be said that a given activity is phasic when it is referred to some particular event, and conversely, that an activity is tonic, when it does not have a time relationship with any particular event. It is better at this stage to abandon for a while the word “activity” and to use a more operational definition. Because by “activity” one usually means the ensemble of the impulse discharges and their mutual time relationships, we prefer to speak directly in terms of impulses, and to consider the process of firing as a stochastic, or random process, so that it can be treated in terms of probability theory. Therefore, we shall account for the time relationship of the impulses in terms of probability of firing, probability distribution functions and probability density functions. We must, thus, reword our definitions as follows. Phasic is the state in which the curves of probability distribution functions of the variable X cannot be shifted along the time axis because the form of the curve would change. This corresponds to saying that the impulses are time-locked to some particular event (e.g., a stimulus), which was our first definition. On the other hand, tonic is the state in which the curves of probability distribution functions of the variable X can have any translations on the time axis, with the curves retaining their form. Translated into terms of impulses, this implies that the discharges are not time-locked to any particular event and that they may be randomly distributed in time.
Measures A time-invariancy of this process is implicit in the fact that the form of the distribution functions does not change when the origin is shifted along a time axis. However the concept is conveyed also by the observation that this would constitute some kind of a neural “steady state”, that is to say, it remains invariant also from a thermodynamic standpoint. Thus, the problem is clearly one of finding a criterion for measurement. What measures are characteristic of the processes of our system, in such a way that, when a time-invariant condition is reached, we can safely assume that the system is in a steady state? Thermodynamically, a steady state is defined by the time-independency of the values of the macroscopic parameters at every point of a given system. This definition, of course, applies to non-living, as well as to living, systems. From a thermodynamic standpoint, measures on macroscopic parameters of a system have References P. 201-203
188
A. ARDUINI
only a statistical meaning, inasmuch as they are average measures taken over large time or space intervals. One reaches the same conclusion when one is dealing with measures on time-invariant random processes, because they are defined in terms of probability functions and “a statistical study of random processes is not a study of each function.. . belonging to the group of functions characterizing the process, but a study of the properties of the entire ensemble, by means of averaging the properties of the functions entering it” (Solodovnikov 1960, p. 86). ( a ) Approaches to statistical measurements in the nervous system. The problems related to the statistical measurements have been outlined recently in connection with both micro- (Gerstein and Kiang 1960; see also Communication Biophysics Group and Siebert 1959) and macroelectrode recordings (Communication Biophysics Group and Siebert 1959). The statistical methods have been applied mainly to evoked activity and to surface recordings. There are basically two procedures: 1. statistical handling of the microelectrode data (amplitude, frequency, interval and probability of firing measurements) ; 2. statistically meaningful transformations of gross electrode recordings (averages, auto- and cross-correlations). I n our particular task both microand gross electrode recordings have been used. Some comments must be made on both types of procedure. Because a statistically meaningful number of retinal output elements had to be treated, a large number of isolated unit recordings was also needed. This was achieved easily by using the electrode which had the largest diameter and still allowed a recording of isolated unitary discharges. Such an electrode picks up discharges from several different units at a time. This introduces some difficulties in the handling of the data and these are best overcome by using electronic computers for the analysis of the amplitude and time characteristics of the discharges. As far as gross electrodes are concerned, these are not generally accepted as a means of recording the ongoing activity, especially that recorded from fibre bundles. However, when statistical information is needed, as in our work, recording from fibre bundles has the great advantage that it deals with only one class of phenomena, the “all-ornothing” axonal events which are not polluted by cellular graded potentials. Even so, it would remain extremely difficult to evaluate quantitatively the amount of activity going on in the territory around the electrodes, because “spontaneously” there are no distinguishable waveforms to measure; they are generally locked to stimuli. As has been stated before (Solodovnikov 1960, p. 86), statistical measures must be related to the entire process by averaging the properties of the functions entering it. Our measures will be simple averages, even if these simple averages “do not in general completely describe the process. But, by analogy with the first few terms of the power series expansion of a function, the limited description can often be very good for some purpose, and there are cases of great importance.. . in which a small number of averages completely specifies a random process.” (Communication Biophysics Group and Siebert 1959, p. 78). (b) Methods. The method applied in our research has been discussed separately (Arduini and Pinneo 1962a). It consists in the measurement of the effective voltage (the root-mean-square, or square root of the mean of the instantaneous squared voltage), or of the continuous recording of a quantity proportional to the average
CENTRAL EFFECTS OF TONIC RETINAL DISCHARGE
189
energy of the potentials picked up by the gross electrodes. It has been demonstrated (Rice 1954; Goldstein 1960; Arduini and Pinneo 1962a) that these averages, when they are calculated from voltage recordings taken from bundles of fibres (or, more generally, when they deal with strictly “all-or-nothing’’ phenomena) are directly related to the number of impulses contributing to these voltages. These averages are not only directly, but also linearly, related to the number of impulses, provided that the impulses are randomly distributed in time in the different fibres. These average measures are therefore to be regarded as the “expectation” (statistical average, or ensemble average) of the random variable voltage which, in our experiments, is a continuous random variable. If the probability distribution functions of our variable represent a time-invariant (stationary) random process, our averages taken on the random variable will not show changes with respect to time. Conversely, when the measures we are taking on averages of voltage recordings reach steady values (Le., are invariant with time), we may infer that our random variable voltage has time-invariant probability distribution functions, which we have taken before as the definition for a tonic state of activity. These principles form the basis for our recording and data transforming procedures, and represent our first answer to the problem of the measures to be utilized. EXPERIMENTAL EVIDENCE
1. The tonic discharge in the optic$bres of the dark adapted eye
Obviously, the first question is: is there really such a phenomenon as the tonic activity, as we have defined it above? We will now consider the interpretation of the experimental data. The experiments (Arduini and Pinneo 1961, 1962b) were performed mainly on two types of preparations: 1. cats under barbital anaesthesia, with intact brains; 2. nonanaesthetized midpontine pretrigeminal preparations. Control experiments were also performed on precollicular decerebrate preparations with exposure of the chiasma. In these experiments the electrical activity was led from the optic chiasma by concentric electrodes. The MS (mean-square voltage) continuous curve as a function of time was integrated with readings of the RMS (root-mean-square voltage) directly from a panel instrument. MS values are more directly related to the actual ni.mber of active units (Rice 1954; Goldstein 1960; Arduini and Pinneo 1962a), but the RMS transform gives essentially the same information. Parallel experiments (Arduini and Cavaggioni 1961) were also performed on anaesthetized preparations in which semimicroelectrodes (35 p) recorded multiple unitary discharges from the fibres in the optic papilla of the opened eye. The following results refer, unless otherwise specified, to the anaesthetized preparations. The experiments were started after an arbitrary period of dark adaptation of 30 min. Taking the non-changing level of the RMS or MS curve as an index of the steady state of the retinal output, it was found that this state was reached with a different time-lag in different animals. In most of the preparations there was a sudden surge of the output discharge when the light in the animal room was turned off, to commence References P . 201-203
I90
A . ARDUINI
dark adaptation. The RMS value remained very high for a few minutes, reaching a steady level after 10 min or so of slow descent. Occasionally, in other preparations, the ofleffect was followed by a slow increment of the RMS value until a stable level was obtained, again in a time of several minutes. After that no major oscillations occurred, although minor fluctuations around the mean values were always present. In some experiments the of effect was characterised by the slow ascent of the level of activity following a pattern of synchronous bursts, with a dominant rate of 2-3 bursts/sec, but not infrequently with rates up to 40, or more, per second. These patterns generally vanished when the level reached the steady value and only rarely persisted during the dark period. This particular type of discharge has been named “pulsing” and it was found to be associated with lower levels of retinofugal activity (Arduini and Pinneo 1962b). In e;ery case the level of the retinal output, after it had reached steady values, was maintained as long as darkness was maintained and underwent only minor fluctuation. Summing up, records show that a state is actually attained which should meet our requirements for a steady state, because: 1. statistical measures indicated a nonchanging level; 2. its behaviour, after it reached a non-changing level, showed timeindependency from the stimulus (off). It would seem, therefore, that the retinal neurones, after a period of dark adaptation, settle down at a mean regular pace. However, the very presence of fluctuations of the RMS and MS levels indicates fluctuations, at the level of the isolated elements, in the total number of discharges. That this is so is demonstrated by the experiments with recordings by semimicroelectrodes from the optic fibres in the papilla (Arduini and Cavaggioni 1961). Individual units discharged at rates ranging from less than 1 to about 20/sec, very rarely higher, but also rarely at regular rates over long periods of time. Units might also stop discharging for several seconds and might resume firing quite unpredictably, so that the total number of firings counted in the unit of time underwent wide variations. It seems that the distribution of the outgoing activity in the retinal neurones has some kind of a rotational character (Granit 1940, 1955), although simple observation or counting of the discharges cannot offer any indication of real patterning. 2. The tonic discharge during continuous illumination The phenomena occurring at the onset of a long-lasting stimulation (5-45 min) did not concern us here. These are the well-known on effects. The on discharge produces, of course, a large increase of the level of activity in the chiasma (Fig. 1,2). Generally, however, this is not maintained for more than a few seconds, being followed by a decrease below the level of dark adaptation. Also this level is not sustained for more than 10 sec, and a new, somewhat higher level is quickly attained, always lower than the dark adaptation level, and is maintained with only minor fluctuations for the whole duration of the illumination. That is to say, a new steady state is reached, following the previous definition, which should, as such, be characterised by the stabilising of unit firing at a lower pace. This lower pace is due to the facts that: 1. the silent intervals occurring between the outbursts are longer; 2. possibly the number of impulses in each outburst is smaller; 3. some of the previously active units are now silent.
CENTRAL EFFECTS OF TONIC RETINAL DISCHARGE
191
t D
olff
E
7 set
Fig. 1 Anaesthetized preparation (nembutal 50 mg/kg). Concentric electrodes in the chiasma. CRO recordings. A : dark adaptation. At the arrow the light is turned on. B : 5 min after A ; C: 15 min after A - a t the arrow the light is turned off; D: 15 min after C ; E: during retinal ischaemia; F: noise from the electrodes in the chiasma at the death of the animal (compare with E). (From Arduini and Pinneo 1962b.)
The microelectrode recordings (Arduini and Cavaggioni 1961) show that these predictions are correct and that the results reproduce those of previous investigators (Granit 1955; Kuffler et al. 1957). There is, in fact,: 1. a decrease of the firings per unit of time for each unit; 2. some of the elements stop discharging altogether, while 3. a few others increase their firing rate, without compensating for the total loss of number of discharges, because 4. the total number of firings over all the units is lower than it is in dark adaptation. After a period of adjustment, the total number of discharges settles at a considerably stable level with only minor fluctuations. A larger fluctuation of the firing rate and of the number of discharges per unit time is found, however, at the level of the individual units. In other words, although a steady state can be seen also at the unit level, it is recognised more easily when one counts the total number of impulses fired by an entire population rather than the number of References D. 201-203
192
A. ARDUINI
c
f
E!
K
0
lo-
51
t
t
Light on
5
Light off 10
15
20
25
30 Time ( m i n )
35
Fig. 2 Schematic diagram of the time course of the events produced by illumination and the return to dark adaptation. (From Arduini and Pinneo 1962b.)
discharges of a given, or a few, individual units. In fact, the total number of impulses is kept fairly constant throughout the period of light adaptation, although the single elements change both the intensity and patterns of their discharge over a wide range, quite frequently in opposite directions. This point will be resumed in a moment. 3 . The relation of the level of tonic activity to the intensity of light during steadv illumination It was possible to vary the intensity of the stimulating light in logarithmic steps by means of neutral filters. It was found (Arduini and Pinneo 1961, 1962b) that the level of activity stabilised at progressively lower values when the light intensity was increased. The change of level from dark to light (i.e,, the actual amount of level decrease) is, at a first approximation, a power function of the intensity of the light (Fig. 3, A,B). Photographic controls of the pupils showed that during barbital anaesthesia the diameter of the pupils did not change when passing from dark adapted conditions to light. These results affirm that tonic activity, far from being pure “noise”, has a distinct information content. In this case, it is actually a continuous flow of live information about the intensity of the light falling on the retina. This fact was predicted in a previous paper, in which the possible functional significance of the dark discharge was discussed (Arduini and Hirao 1960a). This point will be discussed further in the last section of this paper. A striking feature of the behaviour of the tonic activity during continuous illumination at different intensities is the fact that, while the total number of discharges per unit time seems to be in a steady state, this is not true for individual units. They often show a considerable degree of fluctuation of firing, and can even have opposite signs of reaction to light (Arduini and Cavaggioni 1961). This finding means that tonic activity is, as a mean level, the parameter which has information content, whereas the behaviour of each individual unit does not necessarily contain the same information. This is the basis for Schubert’s (1958) hypothesis discussed elsewhere
CENTRAL EFFECTS OF TONIC RETINAL DISCHARGE
193
pVsMs = 16.2 I -'.
I
,
I
20
,
.
-
40
4
60
.
. . 10.0 . 120. . 140 .
80
I
*
160
Level of. illumination (Lurnens/rn2)
B
/
/*
-0.5
Log
A ~v,
= 0.17 log I +0.614
ao
0.5 1.0 1.5 2.0 2.5 3.0 Level of illurninatlon (log tumens/rnz)
Fig. 3 Plots of the relationship between change of Volt RMS in chiasrna and level of illumination in the anaesthetizedpreparation. A : normal scale; B. log RMS vs. log intensity. (From Arduini and Pinneo 1962b.)
(Arduini and Hirao 1960a). The observation is even more striking when it is compared with Granit's words (1955, see above) that light excites or inhibits according to the nature of the unit and to the intensity of the light. In the experiments reported here, the finding of an over-all decrease of the discharges in the tonic state is confirmed, and this must be independent of the nature of elemental firing (see also Kuffler et al. 1957). It would seem, therefore, that there is a real difference in the mechanisms (not in the units involved) underlying the phasic and the tonic reactions of the retina. 4. Relation of the change elicited by steady illumination to the level of tonic activity in
darkness Because the total number of firings per unit of time seems to be the factor involved in the attainment of the steady state, it seemed advisable to take into consideration References P. 201-203
194
A . ARDUINI
the value of the level of tonic activity in darkness and to test one and the same light intensity against different backgrounds of dark discharge. This can be done, however, only on RMS or MS recordings from gross electrodes (Arduini and Pinneo 1961, 1962b), because: 1. only gross electrodes offer a sample of the output sufficiently representative of the total; 2. microelectrode samples are subject to quantitative changes within too wide a range. A means for the quantitative comparison of different experiments was needed and it was found in the signal-to-noise ratio (S/N) principle (Arduini and Pinneo 1962a), the noise being that level of activity recorded by the same electrodes in the same position in the dead animal; everything recorded above the noise level was considered as the signal. The same principle is used in communication engineering, where signal-to-noise ratios are measured generally as RMS or, more frequently, as power ratios. Both measures have been used in these experiments. Because the MS values are directly related to energy, provided only that the circuit resistance is constant, the resistance of the measuring circuit was carefully controlled, not only at the end, but also during the experiments. This could be done by inducing a reversibIe total retinal ischaemia (Bornschein 1958), which completely suppressed all retinal outflow. With this procedure only the noise should be present. Accordingly, the MS values with ischaemia always fell consistently to the level found in the dead animal (Fig. 1). This observation shows, incidentally, that, if any centrifugal discharge was present in these experiments, its intensity was within the limits of error of the measurements. The result was that continuous illumination had different effects when it fell on retinas with different levels of tonic outflow in dark adaptation. The mean level of the S/N measured in RMS in the dark adapted anaesthetized preparation was 1.72, with values ranging from 1.45-2.75. Results different from those described in the previous sections appeared when the S/N was lower than 1.45. It was not infrequent to find such low values; some of the animals also had lower S/N’s, as low as 1.10, although many of them seemed to be otherwise in good condition. The main point is that steady illumination, when applied to retinas having tonic activity with S/N’s less than 1.45, characteristically brought about an increase of the level of the tonic activity. At the onset of stimulation the on discharge generally continued, sometimes with a minor pause, in a steady plateau; this plateau was maintained as long as the illumination was continued. The reaction when the light was turned off, after a short of discharge, was sometimes an abrupt decrease of the activity to the starting level; at other times a slow ascent to a new and higher dark adaptation level, which was reached within 5-10 min. If the new level of tonic activity in darkness still showed a S/N lower than 1.45, a further stimulation with the same intensity of light still brought about an enhancement of activity, which was, however, smaller than the previous one. This procedure could be repeated until the steady state of activity during dark adaptation had an S/N ratio higher than 1.45. The whole phenomenon can be compared to a “staircase” effect. When the S/N had become higher than 1.45, the same continuous illumination, with the same intensity of light, brought about the decrease in over-all level of retino-
CENTRAL EFFECTS OF TONIC RETINAL DISCHARGE
195
fugal activity as that described above. The value of 1.45 in the anaesthetized preparation can be considered, therefore, to be the reversal threshold for the effects of steady illumination. Within the range of S/N from 1.45 to 2.75 (in the anaesthetized preparations), the same intensity of illumination produces effects which are progressively stronger; that is to say, the higher the level of tonic activity in the dark adapted state, the larger the effects of steady illumination. Thus, the greater the distance of the S/N ratio in the dark from the reversal threshold, the greater the effect of steady illumination upon retinal activity, whether there is an increase or decrease in the activity. All these results were obtained consistently in different preparations, and were occasionally tested in the same preparation when the “staircase” effect was present. In some of the microelectrode recordings (Arduini and Cavaggioni 1961) parallel phenomena could be observed at the unit level, although quantitative comparison of experiments was not possible. A reliable comparison can be made only for different values of the total number of discharges in the same preparation, when other causes leading to changes of the total number of the discharges can be eliminated (such as increased pressure of the electrode, displacement etc.). In the same preparation it could be shown that continuous illumination had decreasing effects when the total number of discharges per unit time was high, while an enhancement was obtained when the number was lower. This constitutes the counterpart at unit level of the effects observed with gross electrodes (Arduini and Pinneo 1961, 1962b). In order to find a clear relationship, the total number of discharges had to be determined over a period of at least 2 min. However, when the firing of an isolated element was counted even for a period of 2 min, one failed, usually, to show any appreciable relationship between the number of discharges and the amount of decrease during illumination. The conclusion is, once more, that only the behaviour of the entire population has information content.
5 . The increase of the tonic level of activity with illumination The fact that the retinofugal discharge is enhanced by steady illumination only when the S/N of the dark discharge is less than 1.45, required a more rigorous examination of this apparent inversion of the effects of continuous illumination. It was assumed (Arduini and Pinneo 1961; 1962b) that, under those conditions, the output of the retina was sustained only, or predominantly, by those elements whose activity was enhanced by illumination (see section 2, the results of microelectrode recordings). The assumption was justified, because, on the few occasions when a retina with low RMS output during darkness was investigated by microelectrode recording (Arduini and Cavaggioni 1961), it was found that most of the units increased their firing rate during illumination; a few others, a clear minority but always present in these instances, were regularly inhibited. It was by no means clear, however, why, in some of the animals which were otherwise in good condition, the retinal activity was sustained only by the light-enhanced elements. As a working hypothesis the assumption was made (Arduini and Pinneo 1961, 1962b) that the low level of activity of the light-depressed retinal units was due References
I).201-203
196
A, ARDUINI
to an active inhibition exerted upon them, directly or indirectly, through some intraretinal mechanism. In other words, we postulated the existence of a mechanism controlling the tonic retinal output in darkness and therefore controlling in this way the information sent through the optic fibres with the “tonic” discharges in darkness and light. A series of experiments has been devised to test the existence of such a mechanism. It had been observed (Arduini and Goldstein 1961) that, after recovery from shortlasting retinal ischaemia, continuous illumination had a stronger enhancing effect on the evoked responses of visual cortex to lateral geniculate shocks. Because the enhancing effect of steady illumination (Chang 1952) had been attributed to the overall decrease of the dark discharge (Arduini and Hirao 1960a), the possibility was considered that after short-lasting retinal ischaemia the recovery of the retina led to an increase of the dark discharge and therefore to an enhancement of its depression by steady light. 6. Effects of short-lasting retinal ischaemia on the level of retinofugal activity It is known that any increase of intraocular pressure well above that of the arterial blood produces ischaemia of the retina which results in the total suppression of all retinal output in less than I min (Bornschein 1958; Arduini and Hirao 1959, 1960a, b). Because it was possible to perform this experiment only on the unopened eye, only the results of gross electrode recordings from the chiasma are reported (Arduini and Pinneo 1961, 1962). As soon as pressure is applied to the anterior chamber of the eye, a strong discharge is recorded, which is possibly an irritative one ; and immediately afterwards, the level of the retinofugal activity falls dramatically and reaches in about 40 sec the level found in the chiasma of the dead animal (Bornschein 1958). Under these conditions, the retina remains silent as long as the pressure is maintained. When the pressure is released (after no more than 10 min), the activity is slowly resumed and it takes from 10-20 min to return to the previous dark adapted level. It was observed, however, that after recovery from short-lasting ischaemia, the dark adaptation level of the retinal activity tended to overshoot the pre-ischaemic values. If, before ischaemia, the S/N of the tonic activity was below the reversal threshold (S/N less than 1.45), and therefore showed the light-enhanced type of reaction, after ischaemia the S/N might rise to higher values and therefore show the light-depressed type of reaction. These results would support the hypothesis of an inhibitory mechanism which depresses the tonic discharge of those retinal elements which are generally depressed by continuous illumination. A short-lasting ischaemia might, in fact, produce (more or less selectively) some kind of lasting damage of the hypothetical intraretinal mechanism which checks the discharge of the light-depressed elements; their activity would thus be released, with the visible consequence of an increase of the overall level of retinofugal activity. The effects of ischaemia could be explained also by assuming that the retinal neurones, when they recover from anoxia, increase their firing rate through some irritative mechanism (due, e.g., to accumulation of metabolites). This possibility is, however, not probable because the after-effect of a short-lasting ischae-
CENTRAL EFFECTS OF TONIC RETINAL DISCHARGE
197
mia on the level of retinofugal activity persists for hours. One may argue that barbital anaesthesia is not suited to reveal a mechanism which is supposed to be inhibitory, because it is known that this drug exerts a general depressant effect on retinal discharge (Kuffler 1953). The same consideration holds also for all the effects described in the preceding sections. Other experiments were, therefore, carried out on unanaesthetized pretrigeminal preparations. 7. The midpontine pretrigeminal preparation: effects of ischaemia and of nembutal As far as reactions to continuous illumination are concerned, the tonic retinofugal activity of the pretrigeminal preparation showed the same patterns of response as those shown by the anaesthetized cats (Arduini and Pinneo 1961, 1962b). The only differences were quantitative. The overall level of tonic activity of the chiasma in dark adaptation was consistently higher than it was in the anaesthetized cat, the mean figure of S/N being 3.15 (from 2.45-4.80). The reversal threshold for the effects of continuous illumination was, accordingly, somewhat higher, around 2.50. This means that all the values found in the anaesthetized preparation can be multiplied by a factor of about 1.75 to obtain the values for the unanaesthetized cat. In this preparation, the relationship between the background level of retinofugal activity in darkness and its decrease by continuous illumination is similar to that of the barbital preparation. Retinal ischaemia of a few minutes duration also showed the same effects. Another test was then devised, in view of the fact that nembutal depresses the activity of nerve cells. Small amounts of the drug were injected intravenously in the hope that they might suppress, more or less selectively, the hypothetical intraretinal mechanism mentioned above, which would inhibit the retinofugal discharge of the light-depressed elements; i.e., the same structures which have been assumed to be more easily disorganized by a short-lasting ischaemia. When 3-5 mg/kg of the drug (a dosage much lower than that used in anaesthesia) were injected into a pretrigeminal preparation, the level of the tonic output slowly increased, and became stable at a new, higher level, after an ascent lasting about 5 min. The increase of the level of the tonic retinal output was followed by manifestations similar to those which occurred when the enhancement was produced by short-lasting retinal ischaemia, as discussed in the previous section. The same intensities of steady illumination induced a depressing effect on the retinofugal discharge, which was stronger the higher the level reached during dark adaptation (see section 3). When the drug was injected when the S/N was below the reversal threshold, the enhancement of the activity might overshoot the critical values, so that the same intensity of steady illumination which previously increased the retinal output, later had a typical depressing effect. The potentiating influences exerted by small amounts of nembutal on the tonic activity of the retina are in fact much stronger than the effects of ischaemia, which are, in their turn, greater than the enhancing effects of light (see section 4). These three procedures may or may not be comparable as far as their intrinsic mechanisms are concerned, but the results are, as far as the general trend of the changes induced is concerned, the same. As far as the effects of nembutal are concerned, one may object that the increase References p . 201-203
198
A. ARDUINI
of the level of the retinofugal activity might be due to collateral effects of the drug, such as, for instance, the relative anoxia brought about by the fall of blood pressure that occurs when the drug is injected intravenously. It can be demonstrated, however, that this fall of blood pressure, which occurs also with small amounts of the drug, is always associated with a lowering of neuronal activity, not only in the retinofugal paths, but also on the cerebral cortex. The fall of blood pressure, moreover, can be avoided simply by injecting the drug very slowly. The only alternative hypothesis is, therefore, that small doses of nembutal increase the retinal dark discharge through a release mechanism similar to that which occurs after short-lasting ischaemia. Complex results are obtained when higher doses of nembutal are injected into the pretrigeminal preparation. For doses above 10-15 mg/kg the overall retinofugal level is usually depressed, and may fall below the reversal threshold of the S/N value. However, even when the S/N is below the reversal values, the light may still maintain its inhibitory power on the tonic level of the retinal output. It is sometimes necessary to reach lower S/N’s, even as low as 1.10 (by adding more of the drug), in order to see an increasing type of effect for the same intensity of illumination. The effects of nembutal may be tentatively interpreted as follows: the drug, when it is injected in small amounts, would interfere almost selectively with the hypothetical controlling mechanism, thus releasing the activity of the retinal elements, whose discharge is reduced by steady illumination. This would lead to an increase of the overall retinal output and to the appearance (or the enhancement) of the depressing effect of steady illumination. As more of the drug is introduced, the generally depressant action of nembutal would involve all the ganglion cells, thus decreasing the overall tonic output, but not changing the sign of the effects induced by illumination. Finally, only the light-enhanced elements would remain active, and the effect of steady light would be reversed. This, however, is merely a working hypothesis, and the concept of an intraretinal mechanism checking (or inhibiting) the discharge of the retinofugal elements should be controlled with further experiments. The explanation we have offered of the effect of short-lasting ischaemia and intravenous nembutal rests on the assumption that the mechanism controlling the tonic output is within the retina itself. An alternative hypothesis would be that either low doses of nembutal or short-lasting ischaemia would impair, at central or retinal levels, a centrifugal control on the retinal dark discharge. A series of experiments carried out on precollicular decerebrate preparations shows, however, that this cannot be regarded as being the only explanation of these findings. I n these preparations the nervous tissue was removed, leaving only a thin layer about 1 mm thick on the top of the chiasma, in order to spare its blood supply. Under these conditions, which interrupted both tracts, the injection of small amounts of nembutal had the same increasing effects 011 the level of the retinal dark discharge as they had in the midpontine pretrigeminal preparation. We will now discuss the problem of whether light-enhanced and light-inhibited retinal elements belong to intrinsically different populations of ganglion cells. A lightevoked discharge from a ganglion cell is obtained when photons are absorbed by the corresponding receptors, thus giving rise to an excitatory process in the straight chain
CENTRAL EFFECTS OF TONIC RETlNAL DISCHARGE
199
receptor-bipolar-ganglion cell. It is generally not admitted that inhibition plays any role in this chain, when the other neighbouring chains are disregarded. What is usually understood is that inhibition appears as a “lateral” influence (Kuffler 1953; Barlow 1953b; Barlow et al. 1957) exerted by some neurones of the excited chain on other ganglion cells whose receptors are either not affected, or are less strongly excited. According to this assumption each ganglion cell may be either inhibited or excited by light, depending upon the area of the retina which is stimulated. The experimental conditions under which the lateral inhibition has been studied are much different from those of the present study. Lateral inhibition appears as a suppression ring around an excited area, and it has been shown by means of short duration flashes. Geometrical conditions in all the experiments reported here did not permit focal excitation. The diameter of the beam, the absence of focussing lenses, and the intraocular scattering allowed a fairly diffuse illumination of large portions of the retina (as could be seen by direct inspection through dilated pupils). Moreover, no flash was ever utilized, but continuous illumination during 5-45 min was always used. There is no evidence, therefore, that the decrease of the retinal output observed under conditions of steady illumination utilises the same mechanism of lateral inhibition. It might be interesting to see whether the interneurones, which are likely to mediate lateral inhibition, are also affected, more or less selectively, by short-lasting ischaemia or by small doses of nembutal. GENERAL CONSlDERATlONS
Three main facts are evident from the results reported above. I. The influence of steady light on the tonic activity of the retina is the resultant of the algebraic sum of two opposite effects on the retinal ganglion cells (Arduini and Cavaggioni 1961). 2. The inhibitory effect overwhelms the excitatory one whenever the background of tonic activity represented by the retinal dark discharge is above a critical level (Arduini and Pinneo 1961, 1962b). 3. The amount of the inhibitory influence is directly related to the intensity of illumination (Arduini and Pinneo 1961, 1962b). The following comments will centre around the specific effects of the tonic activity on visual function. These facts must fit in with what is already known of the physiology of vision. In the short bibliographic account given at the beginning, a number of questions, mostly related to psychophysiology of vision, were left untreated. The following considerations will be concerned mostly with the psychophysiological problems. It is in this domain that the tonic outflow of the retina has been thought to condition the “response” to visual stimuli. The signalling of steady illumination The very presence of a continuous retinal outflow has intrigued physiologists: it could not be tied to any practical necessity for visual function, and it did not fit into the scheme of image formation or patterned vision. At best it was considered, as in Granit’s anticipation (Granit 1955), as an energizing bombardment needed to maintain a tonus central in the visual centres. One problem has always been overlooked, Relereneer D. 201-203
200
A. ARDUINI
i.e., the one related to the neural and peripheral mechanisms underlying the perception of a constant background illumination, as, for instance, when one looks at a cloudless sky or an evenly illuminated white cloth covering the whole visual field. A tacit assumption seems to have been that information on the intensity of a non-changing, evenly-illuminated field was carried through an increase of the firing rate of the on units, as is certainly true for very short flashes of light (see Granit 1947, 1955). However, the on effect (even if the rate of the on discharge is strictly related to the intensity of the flash) can hardly be used to explain the perception of a steady photic stimulus. The intensity of steady illumination may be signalled only by the tonic activity of the retina, provided that the intensity of this discharge is in some way related to the strength of the photic stimulus. It is, in fact, likely that the tonic activity of the retina represents such a mechanism, and has the required properties, namely, constancy against time and adaptability. The only fact that could not be expected is that the retinofugal discharge maintains an inverse relationship with the intensity of steady light, so that the brighter the background of illumination the lower the overall level of the retinofugal discharge. (The limits of the working range of this mechanism have not yet been deterinined with accuracy.)
Threshold during dark adaptation, visual acuity and signal-to-noise ratio It is not possible in this short review to discuss all the psychophysiological phenomena in which the tonic retinal discharge may be involved. We shall consider here only two problems: 1. threshold during dark adaptation; 2. visual acuity. Minimum visibility is extremely low in the human dark adapted eye, i.e., in a condition in which the intensity of the tonic retinal discharge presents its greatest intensity. Calculations of about one quantum for each rod have been reported (Hecht 1945; Pirenne 1951 ; Pirenne and Marriott 1959). On the other hand, visual acuity improves when the test figure is placed on a background of steady illumination, i.e., when the tonic discharge of the retina is likely to be decreased. If we accept the evidence that the retina is continuously active even when it is not directly stimulated, it is necessary to explain how, in experiments on minimum visibility, what is likely to be an extremely small phasic discharge may be detected on a strong background of tonic retinofugal activity. a whispered spoken sentence in a noisy telephone! The similarity with physical (that is non-living) situations cannot be carried too far. In no nerve fibre is there actually an amplitude superposition of noise and signal as in a telephone line. The actual situation in the nerve fibre is that of an impulse, timelocked to the stimulus, which occurs at some point of a sequence of randomly generated impulses. It is, therefore, a matter of measurement of the time-dependency of the spike to a particular event. In this respect only it might be considered a special case of the general problem of detection of signals against background noise of the same amplitude. The theoretical treatment of such a problem has found its most successful practical application in radar (see Woodward 1957), and has been later iitilised with reference to auditory messages (McPherson 1957; Green 1958; Birdsall 1960; Tanner 1960). The particular situation for a visual message would be that of an unknown signal whose time of occurrence is specified in terms of its probability
CENTRAL EFFECTS OF TONIC RETINAL DISCHARGE
20 1
distribution functions. Because the test stimulus is allowed to act for only a short time (milliseconds), and on one or very few receptors, one must actually consider the case of a phasic change concentrated on a small area. In this situation one is most likely to be confronted with a true phenomenon of lateral inhibition, which creates a ring of depressed firing around a focus of activity. Both in visual and auditory systems two assumptions have usually been made: 1. that the biological “noise” (what we have defined as the tonic retinal output) is not participating in the process of detection of the signals; 2. that stimulation with a continuous light or a continuous tone increases the output of the organ by adding a continuous stream of impulses to the biological “noise” existing in darkness. On these bases a thorough discussion has been developed on discrimination threshold experiments with an attempt at a quantitative appraisal of the “noise” (the “dark light”) in terms of quanta of light per unit area of the retina, which would give rise to a discharge of the same strength (Denton and Pirenne 1954; Barlow 1957). According to the results reported above these assumptions should be revised, in view of the fact that an increase of the intensity of continuous visual stimulation actually decreases the overall tonic retinal output. In other words the “noise” decreases with the increase of the background illumination. The decrease of the overall level of retinal output with increasing intensity of the background illumination, everything else remaining constant, brings about a larger average spacing of the random impulses (Arduini and Cavaggioni 1961). Both the tonic activity and the response to a stimulus (the “noise” and the “signal”) are determined uniquely by their probability distribution functions. Therefore, the increase of the average interval of the random events constituting the “noise” would decrease the ambiguity in the detection of the “signal”, much as the decrease of the bandwidth of the noise would decrease the ambiguity of detection of the signals in a radar receiver (Woodward 1957). This is equivalent to saying that the signal-to-noise ratio is increased. Accepting these assumptions, the improvement of visual acuity with an increase of illumination of the background field can be easily explained. The signalto-noise ratio reduction is likely to be of major importance when the visual acuity is studied within a single range of light intensity, either photopic or scotopic; at the borderline the well known anatomical factors (rods vs. cones) have of course the great importance that is usually attributed to them. I wish to thank Dr. M. H. Goldstein of the Massachusetts lnstitute of Technology for his criticisms and suggestions in the preparation of the manuscript. REFERENCES ADRIAN,E. D. Synchronized reaction in the optic ganglion of Dytiscus. J. Physiol. (Lond.), 1937, 91 : 66-89. ADRIAN, E. D. and MATTHEWS, R. The action of light on the eye. I : The discharge of impulses in the optic nerve and its relations to the electric changes in the retina. J . Physiol. (Lond.), 1927a, 63: 378414. ADRIAN,E. D. and MATTHEWS, R. The action of light on the eye. 11: The processes involved in retinal excitation. J. Physiol. (Lond.), 1927b, 64: 279-301. E. D. and MATTHEWS, R. The action of light on the eye. 111: The interaction of retinal ADRIAN, neurones. J. Physiol. (Lond.), 1928,65: 273-298.
202
A . ARDUlNl
ARDUINI, A. Influence of visual deatrerentation and of continuous retinal illumination on the excitabdity of geniculate neurons. In R. JUNGund H. KORNHUBER (Editors), Neurophysiologie und Psychophysik des visuellerz Systems. Springer, Goettingen, 1961: 117-125. ARDUINI, A. and CAVAGGIONI, A. Attivita tonica della retina registrata con semimicroelettrodi. Boll. SOC.ital. Biol. sper., 1961, 37: 1393-1395. ARDUINI,A. e GOLDSTEIN, M. H. Localizzazione e meccanismo dell’effetto Chang. Boll. SOC. ital. Biol. sper., 1960, 36: 1530-1532. ARDUINI, A. e GOLDSTEIN, M. H. Enhancement of cortical responses to shocks delivered to lateral geniculate body. Localization and mechanism of the effects. Arch. ital. Biol., 1961. YY: 397412. ARDUINI, A. and HIRAO,T. On the mechanism of the EEG sleep patterns elicited by acute visual deafferentation. Arch. ital. Biol.,1959,97: 140-155. ARDUINI, A. and HIRAO,T. Enhancement of evoked responses in the visual system during reversible retinal inactivation. Arch. ital. Biol.,1960a, 98: 182-205. ARDUINI,A. and HIRAO,T. EEG synchronization elicited by light. Arch. i d . Biol., 1960b, 98: 275-292. A. e PINNEO, L. Attivita nel nervo ottico e nel genicolato laterale nell’oscurith e durante ARDUINI, l’illuniinazione continua. Boll. SOC.ifal. Biol. sper., 1961,37: 430432. ARDUINI, A. and PINNEO,L. A method for the quantification of activity in the central nervous system. Arch. ital. Biol., 1962a, I O U : 415424. ARDUINI, A. and PINNEO, L. Properties of the retina in response to steady illumination. Arch. ital. Biol., 1962b, IUO, 425-448. BARLOW, H. B. Action potentials from the frog’s retina. J. Physiol. (Lond.), 1953a, IIY: 58-68. BARLOW, H. B. Summation and inhibition in the frog’s retina. J. Physiol. (Lond.), 1953b, IZY: 69-88. BARLOW, H. B. Increment thresholds at low intensities considered as signal/noise discriminations. J. Physiol. (Lond.), 1957, 136: 469488. H. B., FITZHUGH, R. and KUFFLER, S. W. Change of organization in the receptive fields BARLOW, of the cat’s retina during dark adaptation. J . Physioi (Lond.), 1957, 137: 338-354. BIRDSALL, T. S. Detection of signals specified exactly with a noisy stored reference signal. J. acoust. SOC.Amer., 1960,32: 1038-1045. BORNSCHEIN, H. Spontan und Belichtungsaktivitat in Einzelfasern des N. opticus der Katze. I. Der Einfluss kurzdauernder retinaler Ischaemie. 2. Biol., 1958, IIO: 210-222. CHANG,H. T. Cortical responses to stimulation of lateral geniculate body and the potentiation thereof by continuous illumination of retina. J . Neurophysiol., 1952, 15: 5-26. COMMUNICATION BIOPHYSICS GROUP OF RESEARCH LABORATORY OF ELECTRONICS and SIEBERT, W. M. Processing neuroelectric data. Mass. Inst. Technol., 1959, Tech. Rep. 351; 121 p. DENTON, E. J. and PIRENNE, M. H . The absolute sensitivity and functional stability of the human eye. J . Physiol. (Lond.), 1954, 123: 417442. FITZHUGH, R. The statistical detection of threshold signals in the retina. J. gen. Physiol., 1957, 40: 925-948. GERSTEIN, G. L. and KIANG,N. Y . S. An approach to the quantitative analysis of electrophysiological data from single neurons. Biophys. J., 1960, 1 : 15-28. GOLDSTEIN, M. H. A statistical model for interpreting neuroelectric responses. Inform. a. Control, 1960,3: 1-17. GOLDSTEIN, M. H. and ARDUINI, A. Cortical responses to shocks delivered to lateral and medial geniculate bodies under differing retinal conditions. Res. Lab. Electron., Mass. Inst. Technol., 1961, Quarf. Rep. No. 60: 215-221. GRANIT, R. Rotation of activity and spontaneous rhythms in the retina. A c f a physiol. scand., 1940, I : 370-379. GRANIT, R. Sensory mechanism of the retina. Oxford Univ. Press, London, 1947,412 p. GRANIT, R. Receptors and sensory perception. Yale Univ. Press, New Haven, 1955, XI1 369 p. GREEN, D. M. Detection of multiple component signals in noise. J . acoust. SOC.Amer., 1958, 30: 904-91 1 . (Editor), Science in progress. IV. Yale Univ. Press, HECHT,S . Energy and vision. In G. A. BAITSELL New Haven, 1945, Series I V : 75-97. KUFFLER, S. W. Discharge patterns and functional organization of mammalian retina. J. Neurophysiol., 1953, 16: 37-68.
+
CENTRAL EFFECTS OF TONIC RETINAL DISCHARGE
203
KUFFLER, S. W., FITZHUGH, R. and BARLOW, H. B. Maintained activity in cat’s retina in light and darkness. J. gen. Physiol., 1957,40: 683-702. Mc PHERSON, R. Inapplicability of the threshold concept to detection of signals in noise. J. acoust. SOC.Amer., 1957, 29 1 15 I . PIRENNE, M. H. Quantum physics of vision. Theoretical discussion. In J. A. V. BUTLERand J. T. RANDALL (Editors), Progress in biophysics and biophysical chemistry. Pergamon Press, London, 1951, YoZ. 2: 193-223. PIRENNE, M. H. and MARRIOTT, F. H. C. The quantum theory of light and the psychophysiology of vision. In S. KOCH(Editor), Psychology: a study of a science. Mc Craw-Hill, New York, 1959: 288-361. RICE,S. 0. Mathematical analysis of random noise. In N. WAX(Editor), Selectedpapers on noise and stochastic processes. Dover, New York, 1954: 133-294. SCHUBERT, G. Neurophysiologie des Simultankontrastes. Albrecht Y. Graefes Arch. Ophthal., 1958, 160 :94-97. SOLODOVNIKOV, V. V. Introduction to the statistical dynamics of automatic control systems. Dover, New York, 1960, I X 305 p. TANNER, W. P. Theory of signal detectability as an interpretative tool for psychophysical data. J. acoust. SOC.Amer., 1960, 32: 1140-1147. WOODWARD, P. M. Probability and information theory with application to radar. Pergamon Press, London, 1957, X 128 p.
+
+
DISCUSSLON R. JUNG: With regard to the prevalence of the dark-discharge from the retina, did 1 understand correctly that the barbiturized preparation is in principle similar to the mesencephalic preparation? J. M. BROOKHART: 1 should like to raise a question concerning the functional meaning of the dark discharge. He indicated his opinion that the dark discharge has an informational content. What sort of information do you think this dark discharge carries?
G. MORUZZI: There are two aspects of Dr. Arduini’s report which in my opinion deserve to be discussed from the angle of the integrative physiology of the central nervous system: ( I ) The steady decrease of the retinal dark discharge elicited by continuous illumination, and the significance of this phenomenon for visual perception, with particular regard to the evaluation of the intensity of diffuse, non-patterned visual stimuli. (2) The retinal dark discharge as a source of a tonic reflex control of central nervous structures which is not necessarily correlated with visual perception. Since Dr. Arduini has concentrated on the first problem, I shall report two experiments closely related to the second aspect of his report. (1) Drs. Berlucchi and Strata have investigated the ocular behavior of the owl, placed in a dark, sound-proof room. The animal had been blinded unilaterally by (i) intraocular compression lasting up to 6 h ; or (ii) by photocoagulation of the pecten and the underlying papilla. Using a Sniperscope or photographs taken with flashes much shorter in duration (2.5 msec) than the latent time of any palpebral response to light, they showed that, during the process of falling asleep, the drop of the lids constantly occurred earlier on the blind eye. Because the animal was in complete darkness, the palpebral asymmetry was due to the retinal dark discharge of the intact eye, whichobviously supports reflexly the central tone of the ipsilateral palpebral motoneurons.
A. ARDUINI
(2) Arduini and Hirao (1959) have shown that the low voltage fast activity of the midpontine pretrigeminal preparation is replaced by EEG synchronization during bilateral ischemic blackout of the retina. Because this effect is present in darkness, and inasmuch as pain is completely eliminated by the pretrigeniinal level of the section, the conclusion has been drawn that the activating structures responsible for the peculiar behavior of the midpontine pretrigeminal cat are supported reflexly by the retinal dark discharge. After post-collicular transection of the brain stem, the isolated cerebrum behaved like Bremer’s cervem isole‘: the EEG was dominated by trains of spindles separated by prolonged interspindle lulls. Bizzi and Spencer have shown, however, that even in these experimental conditions tonic activating structures, driven reflexly by the retinal dark discharge, are still at work. They are simply overwhelmed by the synchronizing structures, but they are still able to restrain synchrony. This conclusion is supported by the observation that a short-lasting (5-6 min), bilateral retinal black-out performed in complete darkness will produce, in a thoroughly reversible and predictable manner: ( i ) an increase in both the repetition rate and duration of the spindle bursts and ( i i ) the appearance of slow waves during the interspindle lulls.
ARDUINI, A,, and HIRAO,T. On the mechanism of the EEG sleep patterns elicited by acute visual deafferentation. Arch. ital. Biol., 1959,97: 140-1 55. BERLUCCHI, G . and STRATA,P. G . Palpebral asymmetry in the dark adapted owl (athena noctua) following unilateral irreversible visual deafferentation. Arch. ital. Biol., 1962, 100: 248-258. BIZZI,E. and SPENCER, W. A. Enhancement of EEG synchrony in the acute “cerveau isol6”. Arch. ital. Biol., 1962, 100: 234-242.
R. JUNG: Will you accept the explanation that the resting discharge in light and darkness represents the effect of a balanced system, set at different levels by the light input? Then you should admit that two reciprocal systems, the B, or on-center system and the D, or off-center system, are both active and the results cannot be explained by a single system in terms of a signal to noise ratio. The advantage of such a device could be that it can work in two directions, both forwards excitation and inhibition and still remain in relative balance of the total output. 1 have to modify Prof. Jasper’s remark concerning contrast mechanisms. These are just knocked out during dark adaptation, as Kuffler and his group have shown, by the disappearance of lateral inhibition. This results in increased convergence from receptors and bipolars on each ganglion cell. This increased convergence might amplify the dark discharge.
S. W., FITZHUGH, R. and BARLOW,H. B. Maintained activity in cat’s retina in light and KUFFLER, darkness. J. gen. Physiol., 1957,40,683-702. R. GRANIT: Pointed out, in summarizing the discussion, that it had followed two main lines. On the one hand it was necessary to understand what purpose the tonic discharge might serve, a question especially considered by Professor Moruzzi. On the other hand, one wanted to understand the nature of the retinal mechanism responsible for the tonic light and the tonic dark-discharge. Granit’s view was that one should first consider the possibility of Arduini’s index differentiating within a small fibre system emanating from cones and a large-fibre system emanating from rods before one proceeds to discuss retinal mechanisms of inhibition. A. ARDUINI’s replies
To R. Jung The differences found in the anesthetized and non-anesthetized (pretrigeminal) preparations are only quantitative and not qualitative ones. In the pretrigeminal cat the dark-discharge is on the average stronger than during barbital anesthesia by a factor of little less than 2. All the phenomena just described in the anesthetized preparations have also been observed in the pretrigeminal animals.
DISCUSSION
205
To J. M . Brookhart As we have shown, the decrease of the activity of the retina brought about by continuous illumination is a function of the intensity of the light. We suggest, therefore, that the information carried is concerned with the average brightness of the visual field. As far as other functional aspects of this tonic activity during light and dark adaptation are concerned a distinction should be made between specific and “aspecific” effects. A typical specific effect is represented by the increase of visual acuity with the intensity of retinal illumination which occurs both for rod and cone function. For a given range (either scotopic or photopic) the reduction of the tonic retinal activity operated by any increase of continuous illumination would increase the ratio of the signal to the biological “noise”. An example of non-specific function, implicit in Granit’s concept of the “energizers”, is given by our experiments with Dr. Hirao (1959, 1960). In the pretrigeminal preparation the tonic activity during dark adaptation is sufficient to maintain EEG activation (presumably by sustaining the central tone of the ascending reticular system) and the reduction of this tonic activity, during steady illumination of moderate intensity is able to induce EEG patterns of sleep. ARDUINI,A. and HIRAO,T. On the mechanism of the EEG sleep patterns elicited by acute visual deafferentation. Arch. ital. Biol.,1959,97: 140-155. ARDUINI, A. and HIRAO,T. EEG synchronization elicited by light. Arch. ital. Biol., 1960, 98:275292.
To R. Jung What professor Jung is proposing is an alternative between one-population and two-population models. There is little doubt about the validity of the two-population model, when we are concerned with the phasic phenomena elicited by changes in the intensity of retinal illumination. Jung’s distinction between B, or brightness system, which is activated by any increase of illumination, and D, or darkness system, which is activated by any decrease of illumination, is of great value for explaining such phenomena as simultaneous contrast and, more generally, patterned vision. The problem we are concerned with here is whether the same reciprocal organization is involved in the transmission of information under steady light, when the visual field is homogeneous and therefore patterned vision is absent. The main difficulty, in our opinion, is to explain how time-locked discharges, relatively short in duration, such as those characterizing the behavior of B and D systems, may convey information during continuous illumination of the whole visual field. Undoubtedly the same retinal units, whose time-locked activity corresponds to either B or D patterns, may present a tonic discharge (see e.g. Kuffler et al. 1957). According to professor Jung the level of the tonic discharge of the B and D systems should be responsible for the information received during darkness (“central gray”) and during steady illumination (Jung 1961). We have no crucial evidence for either proving or disproving the two-population hypothesis. It is true that when the level of the dark discharge was high, we were only rarely able to find units increased by steady illumination, thus behaving as B units (Arduini and Cavaggioni 1961). We realize that this could be due to the fact that our electrodes recorded predominantly a given group of optic fibers e.g. in relation to their diameter. It remains to be explained, however, why the same electrodes recorded a behavior similar to that of the B units whenever the level of the dark discharge was low. The point we want to make is that, when the retinal dark discharge is intense, only inhibitory effects were observed during steady illumination with our recording techniques, or these were predominant; and it is the reduction of the overall discharge which appears to be closely related to the intensity of steady illumination. This constant and close relation has not been found by the authors who were interested in the tonic behavior of single B and D units. This is shown by the fact that the maintained discharge was characterized by a high degree of variability even for a given unit under constant illumination (cf. Kuffler etal. 1957 and Arduini and Cavaggioni 1961). For this reason we suggest that the information value of the tonic retinal discharge-for a steady, uniform illumination of the visual field-is given by thesum of all the individual discharges. When the retinal dark discharge is intense, and when the intensity of steady light is moderate, inhibition of tonic activity appears to be the significant factor. Our work should be regarded as a preliminary attempt to integrate the retinal output under these simplified experimental conditions. S . W., FITZHUGH R. and BARLOW H. B. Maintained activity in cat’s retina in light and KUFFLER darkness. J. gen. Physiol., 1957,40: 683-702.
206
A. ARDUINI
JUNC R. Korrelationen von Neuronentatigkeit und Sehen. In: R. JUNG und H. KORNHUBER (Editors) Neurophysiologie und Psychophysik des visuellen Systems. Springer, Gottingen, 1961, 410435. ARDUINI, A. e CAVAGGIONI, A. Attivita tonica della retina registrata con semimicroelettrodi. Boll. Soc. iial. Biol. sper., 1961, 37: 1393-1395. To R. Granit I thank prof. Granit for summarizing our results. As 1 pointed out in answering Prof. Brookhart’s questions, the purpose of the tonic discharge from the retina must be sought both in the specific visual domain and in the non-specific sphere. The motor asymmetries described by Prof. Moruzzi must be explained in both terms. During the experiments and during the preparation of this paper we have always considered that the dual system of fibers of both small and large diameter, might be differentially involved in our effects. The quantitative index that we have utilized so far is unable to differentiate between the two systems. But this is the object of one line of our future researches and we hope that more data will be available soon.
Multisensory Convergence on Cortical Neurons Neuronal Effects of Visual, Acoustic and Vestibular Stimuli in the Superior Convolutions of the Cat’s Cortex R. JUNG, H. H. KORNHUBER
AND
J. S. DA FONSECA*
Department of Clinicnl Neurophysiology, University of Freiburg, FreiburglBr. (Cermnny)
Multisensory information is an essential function of the brain as is shown by studies of both behaviour and psychologic experience. The early experiments of Moruzzi and his group (Baumgarten and Mollica 1954; Moruzzi 1954; Scheibel et al. 1955; Mancia et al. 1957; Rossi and Zanchetti 1957) have well demonstrated the convergence of different sense modalities on neurons of the reticular formation. Other studies have shown such a convergence on neurons of the putamen (Segundo and Machne 1956), the caudate nucleus (Albe-Fessard et al. 1960), the centromedial (Albe-Fessard et Gillet 1961) and lateral thalamus (Borenstein et af. 1959), the amygdala (Machne and Segundo 1956) and the hippocampus (Green and Machne 1955). The neuronal basis of such multisensory information in the isocortex is, however, not yet well known. The experimental study of the convergence of several sensory modalities on the same cortical neurons is only just beginning (Kornhuber and Da Fonseca 1961a). Apart from our work on the visual cortex (Jung 1958, 1961a), Buser and Imbert (1 961) have collected a neuronal population with multisensory input from the motor cortex under curare or chloralose. We have explored multimodality afferents in the “enckphale isole” preparation without anesthesia in the hope of defining some so-called nonspecific cortical afferents more precisely in relation to their sensory origin. From experiments on the convergence of specific, unspecific and vestibular afferents at neurons of area 17, presented elsewhere (Jung 1958, 1961a, b), it became clear that, even in a primary receiving area such as the visual cortex, convergence of visual input with other modalities must play an important part. It was of interest to investigate also the surrounding areas for their afferents from visual and other modalities, especially for their vestibular input. After Grusser’s demonstrations (Griisser und Grusser-Cornehls 1960; Griisser et al. 1959) o f vestibular responses in area 17, Kornhuber and Da Fonseca (1961b) showed a much wider distribution of vestibular input and a less specific vestibular preponderance in the visual cortex than Griisser had assumed. ~~
*
Present address: Hospital Julio de Matos, Av. do Brasil, 53, Lisbon (Portugal).
R e f t v n c r s P 231-234
208
R. JUNG
et al.
From Albe-Fessard’s (Albe-Fessard et al. 1960; Albe-Fessard and Gillet 1961) and Buser’s work (Buser 1960; Buser and Imbert 1961) we were aware of the difficulties in showing these multisensory effects in preparations without chloralose. Records were therefore taken from single cortical neurons in the “enctphale-isolt” cat without anesthesia. Of course the “enctphale isolC” has other limitations, especially the exclusion of somatic afferents from the spinal cord. We have therefore restricted our experiments to optic, vestibular and acoustic sensory modalities. It is the purpose of this paper to describe some preliminary results from these multisensory effects on 450 cortical neurons. They will be discussed in relation to other results and with some speculation about the significance of these convergent neuronal mechanisms. It is hoped that these efforts may lead to a more precise investigation of this multimodal interaction. MATERIAL
About 450 neurons from various areas in the superior convolutions of the cortex of the cat were recorded and tested for their response to optic, acoustic and vestibular stimuli. In addition 67 neurons of the lateral geniculate body were recorded during optic, vestibular and reticular stimulation and the old material of Grusser et al. (1959) of 70 neurons of the visual cortex was again tested for optic and vestibular stimuli. The statistical analysis of this material and the correlation of the various response types has not yet been completed. It will be reported later together with the results of further experiments on the interrelation (facilitation or inhibition) of multisensory stimuli. The following incomplete results, illustrated by Figs. 1-6, are selected as prominent examples of multisensory convergence. They are intended to give a preliminary survey of the main findings, summarized in Fig. 8. No complete cytoarchitectonic investigation of the cat’s cortex with its individual variations is available, except the recent study of the visual areas 17, 18 and 19 by Otsuka and Hassler (1962). We therefore use the conventional mixed terms for the various cortical areas as determined by the evoked potential method, mainly by Woolsey et al. (1961), together with old anatomical terms and the location in the gyri. METHODS
Glass capillaries of the Ling-Gerard type with tips 0.5 to 3 microns in diameter, filled by the method of Tasaki et al. (1954) were used to record extracellularly single neuron activity from the cortex of cats in Bremer’s “enctphale isolc” preparation with artificial respiration. Visual stimulation consisted of diffuse white light directed onto closed lids: continuous illumination and intermittent stimulation with light and dark phases of equal duration at 500 lux. The lids were sutured to avoid any pattern stimulation which might be caused by the movements of the eye during labyrinthine stimulation.
MULTISENSORY CONVERGENCE ON CORTICAL NEURONS
209
Vestibular stimulation. Because it was necessary, during vestibular stimulation to avoid activation of other receptors and movements of the animal, labyrinthine polarization was provided at the round window of the ear by weak galvanic currents of 0.05 to 1.2 mA. The technique was similar to that described by Griisser (1959), but more obtuse electrodes with silver ball tips were used. In addition, caloric stimulation of the ipsi- or contralateral labyrinth was tested in a number of neurons, mainly from the vestibular area. Acoustic stimuli were hand claps or whistles and sometimes clicks. A six-beam cathode ray oscillograph, built by J. F. Tonnies, was used to record simultaneously several neuronal spikes, electrocorticograms and a photocell record of light stimulation. In the experiments on the lateral geniculate body presynaptic optic nerve fibres had to be excluded. Therefore only those spikes were selected for recording which showed the n-,!?-component described by Griisser-Cornehls and Grusser (1960). In the recordings from cortical neurons no such distinction between axons and cell bodies was made. We assumed that, among the millions of corticalcells and fibres, the few specific subcortical afferent axons which could be picked up by the microelectrode in the cortical layers would be negligible. e
RESULTS
1. Responses of cortical neurons to visual, auditory, and vestibular afferent impulses Cortical unit responses to afferent stimulation could be divided into two general classes or types as follows :
Type 1. Immediate responses (“specijk”) These responses have a short and constant latency with “d’emble2’ activation or inhibition. They maintain a good temporal relation to stimulus duration, with little or no continuation after its cessation (e.g. Figs. 3b and 3e) except for rebound (Fig. 3a) following inhibition. They correspond to what is usually considered to be a specijic response. One variety of these immediate responses may be distinguished from the others as follows : Type la. Initial brief transient responses These responses are characterized by an initial brief transient burst with little or no sustained discharge. They occur mostly in trisensory neurons (Figs. 5 and 6). Type 2. DeIayed activation (b‘unspecijic”j These are long and variable latency responses, generally with recruiting activation and without inhibition. Their duration is prolonged after the cessation of the stimulus (Figs. lc, 2b and 2d). These responses seem to be related to “arousal” effects presumably generated in the “unspecific” ascending activating system (Magoun 1958). References p. 231-234
210
K. JUNG
et a / .
2, Topographic distribution of responses in cortical areas of the superior convolutions (see Fig. 8). Distribution of neurons from the following corticalareas were recorded and tested for the three sense modalities: Visual area I, gyrus lateralis posterior medialis (mainly parasagittal), area 17. Most neurons in this area gave immediate responses, Type I , to light or darkness. The majority of responses to labyrinthine polarization were of the delayed activation type 2 (Fig. 2 ) . Vestibular responses of type I were rare. Only one trisensory neuron (Fig. 5 ) with type 1 responses to visual, auditory and vestibular stimuli was encountered among about 30 neurons, tested extensively with all three modalities. Few acoustic responses of type 2 occurred. Paravisual areas, gyrus suprasylvlus posterior and medialis (visual area 11 and surroundings including area 18 and 19). In these areas neuronal responses were very similar to the primary visual area 17. Most units showed type 1 responses to light or dark. About half of them show responses to labyrinthine polarization, mostly of the delayed type 2. Vestibular or acoustic responses of type I were rare. Some combinations of visual type I and type 2 responses were obtained. spontaneous activity- In darkness
-
-.
light on
l i e on
-,
cathodic polarization contralot. labyrinth
C
Fig. I Visuo-vestibular convergence in a neuron of the para-auditory corfex (Ep. gyrus ectosylvius posterior). The neuron responds to light on with type 1 (trace b), to labyrinthine polarization with type 2 (trace c). The light response is similar to B-neurons of the visual cortex. (Neuron FK 67-2).
Para-auditory area, gyrus ectosylvius posterior (Ep). In this area neuronal responses (Fig. 1) were very similar to those in paravisual areas, mostly visual type I , vestibular type 2. Auditory responses of type 1 or 2 were exceptional. Auditory area I, gyrus ectosylvius medialis and anterior. In this area most neurons showed immediate type I responses to acoustic stimuli, one fourth were also activated by labyrinthine polarization (some with type I). No visual responses were encountered.
MULTISENSOKY CONVERGENCE ON CORTICAL NEURONS
21 1
Vestibular area I, gyrus ectosylvius anterior and posterior bank of the gyrus suprasylvius anterior: most neurons in this area showed responses of type 1 to labyrinthine polarization; many also gave type 1 responses to acoustic stimuli (Fig. 3). Visual responses type I were found in 10% of the neurons tested. Three true trisensory neurons (Fig. 6) were found among 41 completely tested units. Somatic area I , posterior gyrus sigmoideus. The majority of these neurons show responses of type 2 to labyrinthine polarization. Vestibular, visual, and auditory responses of type 1 occurred rarely. One trisensory neuron was encountered among 3 1 units sufficiently tested. Association area, gyrus lateralis anterior: about one third of these units were influenced by labyrinthine polarization. Half of them responded with initial inhibition (Fig. 4), which was diminished after repeated stimulation. Auditory responses of type 1 were rare. No visual responses of type 1, but few of type 2 were found. Motor area, gyrus sigmoideus anterior. Only a few neurons were investigated. In Fig. 8, most of the data for the motor cortex are taken from Buser and Imbert's
--
llghl off
on Ilpht
Fig. 2 Convergence and interaction of visual (type 1 ) and vestibular (type 2 ) responses at a D-neuron ojthe primary visual cortex. Simultaneous interaction of visual and vestibular stimuli results in delayed activation during light inhibition and in occlusion during dark activation. (a) activation (type I ) by light off, inhibition following light on (diffuse illumination through closed lids). (b-d) Type 2 responses to onset or cessation of cathodic and anodic labyrinthine polarization duringinhibition by illumination. (e) Occlusion of the effect of vestibular stimulation (anodic) by specific dark activation. (f) Even 6 seconds after onset of darkness, another labyrinthine polarization causes no definite additional activation. (9) 9 sec after onset of another dark stimulation, the cessation of cathodic labyrinthine stimulus produces a slight type 2 response in addition t o receding dark activation. 0.3 mA constant in all stimuli (Neuron FK 108-2). Strength of polarizing current
-
References p. 231-234
212
R. JUNG et al.
study (1961) of acoustic and visual stimuli. Investigation of vestibular input is therefore incomplete. Anatomical data indicate that vestibular responses are to be expected.
3. Neuronal responses according to sense modalities
(a) Responses to optic stimuli Diffuse illumination (light on and off, flicker) elicited neuronal responses of type 1 not only in area 17 (Fig. 2), but also in many neurons of the paravisual and para-auditory areas (Fig. 1) and in a few neurons of the vestibular cortex (see Figs. 6 and 8). The visual type 1 responses in paravisual areas showed similar distinguished and reciprocal patterns of excitation and inhibition to those in the primary visual cortex. These response patterns, A-E, described in other papers (Jung, 1961 a,b) were mainly responses to light on (B-type), to light off (D-type), as well as on-off-responses (E-type). In contrast to area 17 some neurons of the surrounding areas showed greater variations in the intensity of the response with either increase or more rapid decrease after repeated stimuli. Most on-off-neurons from the paravisual areas, like E-neurons in the primary visual system, also had shorter latencies and stronger off-responses than on-responses. They might therefore be classed, according to Baumgartner’s system (1961), as on-off D-neurons belonging to the D-system which probably signals relative darkness (Jung, 1961a,b). The latencies of these light responsive neurons outside the visual area 1 were significantly longer than those of neurons in the primary visual cortex, Under our conditions of dark adaptation the latencies were 68 msec in visual area 1 and 80 msec in the paravisual areas. It is difficult to evaluate these latencies, because stimuli through closed lids were given after dark adaptation, which causes longer retinal delays. In some neurons of the paravisual areas, including Ep, combination of type 1 and 2 responses occurred which showed recruiting effects after repetitive stimulation. Pure type 2 responses to light were observed in some neurons of the vestibular and association cortex with long lasting and recruiting activation. (b)Responses to acoustic stimuli
Clapping, whistles or clicks elicited type 1 responses mostly in the auditory cortex, but also in about one third of the neurons of the vestibular area (see Fig. 3). Such responses occurred only in a few neurons in the visual, somato-sensory and associative areas. Type 1 responses consisted chiefly of brief groups of spikes with short latencies of about 15 - 40 msec both in the auditory and the vestibular cortex. Type 2 responses consisted mostly of prolonged moderate acceleration of the discharge with a long and varying latency, which lasted for several seconds. They were mainly observed in the visual and paravisual cortex in all types of light responsive neurons. These responses were more clearly seen after repeated whistling than after single clicks or clapping.
(c) Responses to vestibular stimuli I. Labyrinthine polarization. Cathodic and anodic polarization of the round window with threshold currents (0.05-0.15 mA) elicited responses of type 1 mainly in the
MULTISENSORY CONVERGENCE ON CORTICAL NEURONS
213
vestibular area. Higher currents up to 1 mA, however, caused, in all the cortical regions investigated, widespread neuronal responses of type 2 with prolonged activation at onset and cessation of polarization (d-type of Grusser et a/. 1959). Type 1 vestibular responses were localized chiefly in the vestibular receiving area of Walzl and Mountcastle (1949) (Figs. 3 and 8). These responses occurred also in about one fourth of the tested neurons in the auditory cortex and rarely in other cortical areas, as is shown by the circles in Fig. 8. The latency was always very short (4-25 msec) and the intensity of the response (activation or inhibition) was maximal in
Fig. 3 Vestibular and acoustic type 1 responses of a neuron in the vestibular cortex. Alteration of direction of specific responses (a, b) into strong on-off-responses in both directions by abnormal stimulus strength (c, d). (a) Direction-dependent inhibition by weak cathodic contralateral labyrinthine polarization (about 0.1 mA). (b) Short latency activation during anodic polarization of the same labyrinth. (c) Cathodic inhibition disappeared and was changed into short latency activation by stimuli of about tenfold current strength. (d) With this higher stimulus strength the silent period following anodic threshold polarization was changed into activation. (e) Shows the type 1 acoustic response of this neuron (Neuron FK 88-1).
the vestibular cortex. Type 1 showed different response types, which have been described in detail in another paper (Kornhuber and Da Fonseca 1961b). The most common responses were of the following two types : (a) activation at onset and cessation of polarization, (b) reciprocal effects with activation at onset and inhibition after cessation of polarization and vice versa after reversed direction of the polarization. In these neurons a reversal of the current direction always caused reversal of the activating or inhibiting responses. These responses in the vestibular area were obtained as well by ipsilateral as by contralateral polarization, Some neurons of the vestibular and association cortex (see Fig. 4) were inhibited at the onset and cessation of the polarization in both anodic and cathodic directions. The inhibitory responses in direction-dependent neurons were obtained only with References
I).
231-234
R. J U N G et ul.
214
weak currents of polarization and were changed into activation by stronger currents (about tenfold threshold strength as in Fig. 3c and d). There was no relation between neuronal discharge and the rhythm of vestibular nystagmus, except in one neuron, described elsewhere (Kornhuber and Da Fonseca 1961b), and when the eyes were open. A special inhibitory response not dependent on direction or weakness of the current was found in the association area (see Fig. 4 and p. 21 l).
rt-
labyrinthine polarization anodic
I
0
500
1000
1500
2000
2504
3000 0
500
1000
1500
2000
2500
3000rnsec
Fig. 4 Vestibular inhibition of u neuron in the ussociation cortex
(gyrus lateralis anterior) by onset as well as cessation of labyrinthine polarization in both anodic and cathodic current direction, plotted as neuronal discharge rate per second. Analysis includes only the first two stimuli in either current direction because this inhibition disappeared gradually after repetitive stimulation (Neuron FK 135).
Type 2 vestibular responses (Figs. 1 and 2) were common in all the cortical regions investigated, usually with higher intensities of labyrinthine polarization from 0.31.2 mA. They consisted of long latency (30-400), and delayed activation of several seconds duration following the onset and cessation of polarization. These responses were less common in the associative cortex of the gyrus lateralis anterior. In contrast to Grusser et al. (1959) we have found also type 2 activation in the geniculate neurons, but only when they were recorded stereotaxically when the cortex was intact. In “open” experiments, when the posterior cortex and white matter were removed, this activation was lacking as it was in Grusser’s experiments (1959). Therefore a recurrent activation from the cortex must be considered for geniculate neurons as demonstrated by WidCn and Ajmone Marsan (1960). This would also explain similar results obtained by Hubel (1960) in the cat and Arden and Soderberg (1961) in the rabbit for arousal and acoustic stimuli. [I. Thermic stimulation. To control polarization effects, the homolateral labyrinth was stimulated by hot or cold water. This usually elicited neuronal activation or inhibition after hot or cold calorization. The temporal relations of these effects to the eye movements and to nystagmus were remarkably close. They are described elsewhere (Kornhuber and Da Fonseca 1961b).
MULTISENSORY CONVERGENCE ON CORTICAL NEURONS
215
The convergence of several modalities Activation or inhibition by specific afferent input of one modality is combined, in most neurons, with some responses to other stimulus modalities. This additional multisensory effect is, however, chiefly of the “non-specific” type 2 with late prolonged activation (Figs. 1 and 2). Type 1 “specific” multisensory convergence on single neurons with constant short latency activation was observed as bisensory convergence, mostly o f the acoustic and vestibular afferents in the vestibular and auditory areas. Trisensory convergence of the three modalities tested was rare. Bisensory neurons of type I Bisensory effects showed all combinations, visuo-acoustic, visuo-vestibular and vestibulo-acoustic convergence at the same neuron. The latter convergence of the acoustic and vestibular afferents (Fig. 3) was most often observed in the two respective areas. In other areas bisensory convergence also occurred in a few neurons only.
vest ibuiar
___ __ ..___ 100 610 800 I,_/ .. 1 lobyr. P O l O r l i Y l i O i l I
0
.10
1000
1:oo
,h?,
liic
1.
Fig. 5 Trisensory neuron ofthe primary visual cortex, showing type I responses. Responses to acoustic, visual and vestibular stimuli (anodic polarization of the ipsilateral labyrinth). Discharge rate plotted per sec on a logarithmic scale. Average of 14 visual, 7 acoustic and 6 vestibular stimuli. (Neuron FK 110).
Usually the acoustic effects prevailed in the auditory, and the vestibular effects in the vestibular area. This specific prevalence and the observations of vestibulo-acoustic convergence also with weak polarization currents near threshold, make it improbable that the convergence was a pseudo-effect of peripheral origin by combined inadequate electric stimulation of both receptor mechanisms in the labyrinth. Verification by rotatory stimulation remains to be done. In a few neurons caloric stimulation showed the specific vestibular origin of the convergence effects in the vestibular area. Kcferencm p , 231-234
R. JUNG et al.
216
Trisensory neurons of type I Neurons responding with short and constant latencies to visual, acoustic and vestibular stimuli were rare, They were most often found in the vestibular area. All these neurons responded from a low spontaneous discharge rate with a primary brief burst and little sustained activation (type 1 a) to visual and acoustic stimuli and with somewhat longer activation or inhibition to vestibular stimuli (Figs. 5 and 6 ) . 200 100
50 L
"a 20
$ 10 0
5 5
._ D - 2
:
1.0
3
2
0.:
i
-olop
150
Convergence of type I responses with type 2 from other modaliries All combinations of this very common convergence of type 1 and type 2 occur on cortical cells, usually from two sense modalities. Convergence of type 1 visual with type 2 auditory and vestibular responses (Figs. 1 and 2) are common in the visual and paravisual areas. The modality of type 1 is usually determined by the cortical area and its main
MULTISENSORY CONVERGENCE ON CORTICAL NEURONS
217
specific input. Visual (type 1) and vestibular (type 2) or acoustic (type 2) responses are mostly found in neurons of the visual and paravisual areas. Acoustic (type 1) with vestibular (type 2) responses as well as vestibular (type 1) with acoustic (type 2) responses are mostly obtained in neurons of the acoustic and vestibular areas, rarely also in the primary somato-sensory area. Visual type 2 responses have not yet been investigated in detail, because slight accelerations of discharge required statistical treatment for their evaluation. They occur mostly in the paravisual areas as increasing acceleration with repeated stimuli, but also in the vestibular, association and other areas. In the paravisual area they were mostly combined with type 1 responses to light and type 2 responses after vestibular and auditory stimuli. Multimodality interaction Interaction of multisensory stimuli has not yet been studied systematically in all these neurons. The following results are preliminary. As Griisser (Griisser and Griisser-Cornehls 1960) found in his study of interaction of visual type 1 and vestibular type 2 responses in the visual cortex, facilitation of visual responses after labyrinthine polarization was observed both for type 2 and type 1 vestibular responses in neurons of the visual, paravisual and vestibular areas. However, occlusion was a more common interaction than ‘facilitation (Fig. 2). Occlusion between visual and vestibular responses was often observed : although vestibular stimulation had a clear activation effect on the neurons showing type 2 responses to onset and cessation of labyrinthine polarization, simultaneous light stimuli often received little or no facilitation by vestibular stimulation. This facilitation could reappear when the effects of light stimulation had declined. Visuo-vestibular facilitation seemed to occur mostly, when the neuron showed less spontaneous or evoked activity in the course of the experiment, either by adaptation, “fatigue” or inhibition (compare Figs. 2d, e, and g for dark activation and light inhibition). I n this stage vestibular type 2 responses were acting similar to “arousal” or “non-specific activation”. The critical flicker frequency for single neuronsJCFF, the maximum frequency at which neurons can follow flickering light) could be raised by vestibular stimulation, as was first described by Griisser and Griisser-Cornehls (1960). Also this flicker facilitation occurred mostly in states of diminished neuronal activity and was less apparent when the neuron showed strong light responses. Facilitation of responses to illumination by other sensory stimuli in light responsive neurons of the visual and paravisual areas chiefly concerned the sustained discharge. The high frequency primary discharge was mostly unchanged. This was true for both reciprocal types of light responses : for activation by illumination in neurons responding with the B-type or for activation by darkness in neurons responding with the D-type. Specific inhibition of these neurons was not abolished initially by type 2 delayed activation and any facilitation from other modalities. But the duration of this inhibition was usually shortened; and in longer inhibitory stages delayed activation from other modalities would break through (Fig. 2 b-d). Thus, the specific response patterns of cortical neurons, which were best studied for visual stimulation, were not changed References P. 231-234
218
R. J U N G et a/.
by other modalities of stimulation. Alteration of the D-response (pure off) into an E-response (on-off with preexcitatory inhibition to light-on), which was sometimes seen and was also described by Griisser (Griisser und Griisser-Cornehls 1960), is not a real change of response type, because most neurons of the E-type belong to the D-system (Baumgartner 1961 ; Jung et al. 1952). The interaction between different modalities was variable in the few true trisensory neurons, showing type 1 responses to all the modalities which we have found in the visual and vestibular area. These 5 neurons had little spontaneous activity and in contrast to type 1-type 2 visuo-vestibular interaction described above, multi-modality stimuli also did not change the background discharge. Ln one neuron from the vestibular cortex interaction was limited to the duration of stimulation without prolonged after-effects. During vestibular stimulation facilitation of light responses occurred, and the type of light response changed from off- to on-off-type. Vestibular stimulation facilitated also the acoustic responses to clapping. However, another trisensory neuron of the vestibular cortex, shown in Fig. 6, showed a more prolonged after-effect of vestibular stimulation (calorization and polarization). The visual and acoustic responses, which were less apparent before vestibular stimulation, became marked and constant after the first calorization, although spontaneous activity was not increased significantly. In another trisensory neuron of D-type in the visual cortex, shown in Fig.5,interaction of visual and vestibular stimulation, usually caused an elevation of C F F during flickering light. This enhancing effect of vestibular stimuli was most marked when the neuron showed diminished responses to light, as it was also observed in most neurons showing type 2 vestibular response. With the diminution of light responses, C F F decreases also, and additional labyrinthine polarization could raise the C F F to its previous level. DISCUSSION
Before discussing the results of multisensory convergence in various cortical areas we should like to mention briefly the convergence of specific and unspecific afferents in the visual cortex, which has been investigated extensively in our laboratory during the last 10 years. The results have been described in other papers (Akimoto and Creutzfeldt 1957/58; Jung 1958; Jung 1961b) and are summarized in Fig. 7. This hypothetical neuronal scheme shows possible cortical and subcortical pathways to explain these convergences and does not pretend to give an exact picture of the number of synaptic connections. I n addition to this convergence of retinal, thalamo-reticular and vestibular afferents the effect of caudate stimulation on visual neurons has been studied. After we had found constant inhibitory effects of caudate stimulation in nearly all recorded neurons of the motor cortex (Spehlmann et a/. 1960), Lehmann and Koukkou showed that neurons of the visual area are also influenced by caudate stimulation (unpublished experiments). About half of these neurons were inhibited, whilst the other half showed facilitatory convergence with specific visual afferents. After a tetanus in the caudate nucleus, these neurons fire with higher frequency during their sustained
MULTISENSORY CONVERGENCE ON CORTICAL NEURONS
219
discharge in response to visual stimuli. This facilitation is effective for visual responses of both reciprocal neuronal systems B and D, activated by light or darkness respectively. Visual
svstem
Vestibular system
Fig. I Scheme of the convergence of visual, vestibular and reticulo-thalamic systems in the cat’s visual cortex (from JUNG, 1961). The connections are inferred from the results of neuronal recordings in the cat’s cortex, but do not represent exact investigations of synaptic delays from subcortical structures. Most neurons of the visual cortex receive specific reticulo-geniculate afferents together with reticulothalamic and vestibular impulses. The long latencies of reticulo-thalamic and vestibular afferents indicate multisynaptic relays which are probably located chiefly in the brain stem. As reticular neurons already receive the convergence of several sensory modalities, a subcortical coniponent of multiniodality input is very probable. Vestibular and visual convergence is certain in the paramedian reticular formation from anatoniical and physiological evidence. Many neurons of the reticular formation seem to receive a rather specific afferent supply and therefore should not usually be considered as being non-specific. The intercortical connections are conjectural and must be investigated further. Crossing of fibres is only drawn for the optic nerve, and not for other afferents. Multisensory convergence in cortical neurons and its relation to subcortical structures
The diagram given in Fig. 8 and the examples described of multisensoryconvergence on single cortical neurons show that multisensory convergence is common in many cortical regions. The occurrence of neurons with multisensory input even in the primary receiving areas was not unexpected. Recordings of evoked potentials with macroelectrodes (Buser et Borenstein 1959; Buser et a / . 1959; Marshall et al. 1943; Maruyama and Kanno 1961; Thompson and Sindberg 1960; Woolsey 1961) have already shown a wide distribution of specific afferent impulses, which extends, not only to the socalled second sensory areas, but also to various other cortical fields. Buser’s experiments on irradiation of auditory and visual responses (Buser 1960; Buser et al. 1959) although they were usually done under chloralose, suggested such a convergence Rrf‘wm e i
I?
2.71-234
220
R. JUNG et at.
and demonstrated it also at single neurons of the sensorimotor cortex (Buser and Imbert 1961). Our investigations in Freiburg on the convergence of various afferents in the visual cortex (Griisser 1959; Jung 1961a) have shown that labyrinthine impulses participate in this convergence, as was demonstrated by the experiments of Grusser and coworkers (1959, 1960). Griisser’s vestibular responses in the visual cortex were however, all of the long latency type 2. These type 2 responses showing only delayed activation and no inhibition are similar to other unspecific activations of the cortex. They are probably conducted over the brain stem by way of the reticular formation Neuronol r e s p o n s e s t o d i f f e r e n t
sense modalities
IIIIII
visual(light-dark)
... :...*. ::::
auditory (click, clapping)
:$o:
$:2
vest 4 b u I or I y pe 1
1
labyrinth
-_----_ ---_- vestibular
J
polar12
type2
Fig. 8 Topographic distribiition of neuronal responses to three modalities in the cat’s cortex. Visual, acoustic and vestibular responses (type 1) and vestibular responses (type 2) in the superior convolutions of the cortex and in the lateral geniculate body of the cat. In the motor cortex our results were combined with those of Buser and Imbert (1961).
and the thalamus. The reticular formation is the common premotor coordination center for vestibulo-ocular responses, such as nystagmus (Lorente de N6 1928-1 933; Szentiigothai 1943; Shanzer and Bender 1959). In the reticular system, visual and vestibular impulses converge for the regulation of eye movements (Szenthgothai 1943; Jung 1958; Duensing 1961). It seems probable that the long latency responses to labyrinthine polarization in cortical neurons are precoordinated in these brain stem structures as indicated in Fig. 7. Visuo-vestibular convergence resulting in specific optovestibular coordination of body and eye movements should not be separated too rigidly from non-specific activation: “Attention” which is often correlated with non-specific reticular activation contains many specific mechanisms of attentive behavior which are special sensorimotor functions coordinated in cortical and subcortical centers (e.g. adversive movement and nystagmus). Duensing and Schaefer’s experiments (Duensing and Schaefer 1957; Duensing 1961) have shown that two types of neurons occur in the reticular formation, one showing an exact time relation to the nystagmic phases and the other showing diffuse activation. The first type with rhythmic alternation of activation and inhibition is exceptional in the cortex as mentioned by Kornhuber and Fonseca (1961b). A further important integration center of multisensory and sensorimotor coordination is found in the thalamus as shown by Albe-Fessard (Albe-Fessard and Gillet 1961) and Fessard (Fessard 1961). These centro-thalamic nuclei may represent
MULTISENSORY CONVERGENCE ON CORTICAL NEURONS
22 1
relay stations for sensory convergence in the cortex and may be independent of the reticular formation. Buser (Buser et al. 1959) has shown that electrolytic destruction of the mesencephalic reticular formation does not affect the evoked potentials following visual and auditory stimuli in the suprasylvian gyrus, conducted from the thalamus, Our experiments do not allow any conclusions about the pathways of multisensory convergence. But similar activations of neurons in the reticular (Baumgarten and Mollica 1954; Moruzzi 1954; Rossi and Zanchetti 1957; Scheibel et al. 1955), thalamic (Albe-Fessard and Gillett 1961), caudate (Albe-Fessard et al. 1960), claustral (Segundo and Machne 1956) nuclei and the hippocampus (Green and Machne 1955) by various afferents and our earlier experiments on visual and reticulo-thalamic convergence in the visual cortex (Sung 1958, 1961a; Creutzfeldt et al. 1961) seem to justify the assumption that both subcortical and intracortical (interareal) connections contribute to convergence at the cortical level. These neuronal systems seem to be so complex that it is impossible to propose a clear anatomical scheme for their convergence phenomena. By analogy with the visual system from which we have the most complete data on neuronal connections (Fig. 7), similar multiple and convergent pathways might be assumed also for other sensory systems and for the association areas. Parallels with other neuronal studies In spite of a multitude of papers on evoked potentials recorded with macroelectrodes after various sensory stimuli, there are very few microelectrode studies of sensory convergence in cortical neurons. Some authors such as Mountcastle (1957/1961) in his extensive neuronal recordings from the somato-sensory cortex have found specific representation of different modalities without convergence. We have found only five groups of studies on sensory convergence in cortical neurons, four in the cat and one in the rabbit: 1. Amassian (19541, in addition to his studies of interspatial and intermodality interactions by evoked potentials, first mentioned single unit recordings after sensory and acoustic stimuli in association area: he found in the same neuron convergence of somato-sensory stimuli from four limbs and from auditory stimulation. The latencies varied between 28 and 111 msec for nerve stimulations. Visual stimulation was not tested. 2. Landgren (1957) was the first to describe convergence of tactile, thermal and gustatory impulses from the tongue in neurons of the sensory area. His comparison of thalamic and cortical neurons in the afferent systems showed that bimodal or multimodal neurons occurred more frequently in cortical than in thalamic regions. 3. In our laboratory Griisser et al. (1959/60) described long latencyneuronal activation in the visual cortex, following labyrinthine polarization (Griisser et al. 1959), convergent with visual afferents (Griisser and Griisser-Cornehls 1960). In the same experiments a few neurons in the vestibular cortex were recorded with short latency responses, but visual and acoustic convergence was not investigated in these neurons. 4. Buser and Imbert (1960/1961) described polysensory neurons in the cat’s sensory motor cortex, activated by visual, auditory and somatic stimuli. Most of these References p . 231-234
222
R. JUNG et al.
neurons were recorded under chloralose, but some were also recorded from unanesthetized animals treated with curare. 5. In the rabbit’s visual cortex Lonio and Mollica (1959) found activation of some neurons after acoustic stimuli with a long latency of 200 msec. They were not always associated with attentive and orienting behavior. These acoustic neuronal activations and other responses to pain seem to be analogues of non-specific effects and may correspond to our type 2 responses in visual neurons, which show long aftereffects. The results of Buser and lmbert (1961) confirm Hassler’s anatomical postulate (Hassler 1949) that the pyramidal motor system is the final integrative cortical pathway of many sensory systems. They correspond to the experimental results of Jasper et al. (1960) on the integration of conditioning processes in neurons of the motor cortex (Ricci et al. 1957). Our own investigations have shown that such multisensory convergence occurs at many neurons in various cortical fields, including the primary receiving areas of the visual, auditory and somatosensory system. A prevalence of their specific modality input is always apparent in the primary receiving areas with type 1 responses to their own and type 2 to other modalities in the majority of neurons. Trisensory type 1 neurons however, are also found in these receiving areas (Figs. 5 and 6) Extrastriate projection of visual neuronal responses The wide distribution of visual type 1 responses in many cortical cells outside the primary visual cortex was a surprise in our experiments. Although recordings of evoked potentials (Marshall et al. 1943; Doty 1958, 1961 ;Vastola 19611, had indicated the possibility, we did not expect to find similar average latencies and response patterns in neurons of the suprasylvian and posterior ectosylvian gyrus as we had found in B, D, and E neurons of area 17. However, the long retinal delay of scotopic conditions in our experiments with dark adaptation and closed eyes might obscure some areal differences. The latency comparison should be reinvestigated under photopic conditions with open eyes. Since the first observations of Marshall et al. (1943) on evoked potentials in the suprasylvian gyrus it has been assumed that a visualprojection to the cortex which is not dependent on the geniculo-striate system remains after ablation of the primary visual area. It is believed, that their visual afferents in this projection are not secondary responses from the primary visual area and that they travel over other thalamic nuclei (Borenstein et al. 1959). But the evidence for this comes only from physiological experiments and cortical connections of extrageniculate visual pathways are not recognized anatomically. Anatomical and electrophysiological results are conflicting. Since von Gudden’s old observations (1889) on “seeing” rabbits without a visual cortex and with a degenerated geniculate nucleus, Doty (1 961) has recently demonstrated the persistance of visual evoked responses and behavioral vision in cats without a geniculo-striate system after early extirpation of area 17 at birth. On the other hand, Vastola (1961) has just shown that these responses remain after isolation of the geniculate nucleus from the pretectal and medial thalamic structures and by these and
MULTISENSORY CONVERGENCE ON CORTICAL NEURONS
223
other criteria he believed that they are conducted over the ipsilateral geniculate body. He estimated the conduction velocity of these visual afferents to be between 10 and 30 msec and called them “extrastriate secondary radiation fibres” which acted as a direct pathway from the geniculate to association cortex. Changes of visual pattern discrimination were seen after lesions in these cortical areas (Gualtierotti 1961). Ubiquity of delayed activation (type 2) after labyrinthine polarization Vestibular type 2 responses were, as is shown in our scheme (Fig. S), found in all the cortical regions investigated. These delayed responses to labyrinthine polarization may correspond, at least in part, to non-specific ‘‘arousal” activation of the cortex, or they may be explained by the peculiarities of our method of electrical stimulation of the labyrinth. This is supported, not only by the similarity to the general EEGresponses observed after vestibular and reticular stimulation and by the anatomical identity of the vestibular brain stem system with the medial reticular formation, but also by the following experimental observations on type I and 2 responses after labyrinthine polarization. “Specific” type I vestibular responses, which showed either short latency activation or inhibition, were usually elicited with weak polarization currents, whereas type 2 responses usually needed stronger currents, which were often 3 times or more as strong as those which produced type 1 effects. Such strong current polarization, which is often necessary for type 2 effects, may exceed the physiological limits of receptor excitation during this inadequate stimulus. One may assume that during adequate stimulations of normal receptor organs such abnormally strong discharges will be prevented by various regulation mechanisms. Stimulus strength was not altered in our optic and acoustic stimulation. Therefore nonphysiological stimulus strength did not occur in the two other sensory systems, vision and audition. Thus more caution should be exercised in evaluating vestibular responses to polarization currents which are well above threshold than is necessary in the case of auditory and visual response. Comparisons of these widespread vestibular delayed activations in the cortex with the distribution of visual and auditory responses may not be justified. A similar explanation of activation by nonphysiological stimulus parameters may be given for a combination of type 1 into type 2 responses with stronger stimuli. Neurons showing type 1 vestibular responses with inhibition in one direction of polarization and reciprocal activation in the other at low intensities could be converted into strong activations by polarization on and off by increasing the current strength several times over the threshold stimulus for type 1 (see Fig. 3). These apparently nonphysiological activations after strong currents show a mixture of type 1 and type 2 responses. Activation begins in these neurons, as it does with type I responses, with short latencies, but shows considerable prolongation of discharge following stimulus cessation as in type 2 responses. These responses are probably combinations of type 1 and 2 effects, due to unphysiologically strong stimulation. The relatively high incidence of responses to labyrinthine polarization in the auditory cortex raises the additional problem of vestibular specificity of this polarization. Only at weak current polarization does it seem certain that only vestibular and not References P. 231-234
224
R . JUNG
et a / .
cochlear receptors are stimulated. But the possibility cannot be ruled out that, with very strong current polarization, inadequate excitation of the acoustic system in the cochlea also occurs. This may then stimulate type 1 auditory responses in the auditory cortex, when higher intensities are used, which excite acoustic receptors. These elicit afferent impulses to the auditory cortex, resulting in short latency responses of its neurons even though these might be independent of vestibular input. Special features of trisensory neurons It is remarkable that all 5 trisensory neurons of type 1 had an extremely low spontaneousfiring rate of about 1 per sec and a very brief transient response (type 1 a) to sensory stimuli. These neurons were silent during long periods and showed brief high frequency responses to various afferent stimuli, as is shown in Figs. 5 and 6. This low resting discharge distinguished the trisensory neurons from other more active sensory elements in the cortex, which are activated only by a single modality. A low discharge rate seemed to be present also in Buser and Imbert’s (1961) multisensory neurons of the motor and sensory cortex as their figures show. They also mention a maximum of response with low spontaneous activity or grouped discharge. Grusser-Cornehls’s “true binocular” neurons, found in the visual cortex, also showed this low spontaneous activity, although these neurons receive a different kind of mutually facilitating convergence without response to single eye stimulation (Grusser-Cornehls und Grusser 1961). Phenomena of occlusion and facilitation occurred also in these trisensory neurons. However, a true trisensory facilitative convergence in neurons, which did not respond to one modality alone, was not encountered in our experiments. It should be added that we have not looked for it specially. Although intermodal interaction with simultaneous stimulation was variable, there seemed to be some preponderance of the responses to the modality, specific for the area in which these neurons were located. The small number of neurons in which we have recorded this trisensory response type l a (about 3% of all fully tested neurons) precludes any generalizations and for further confirmation we must await future work. Mixed convergence with type 1 and type 2 responses of diflerent modalities The commonest occurrence of multisensory convergence on single neurons is type 1 response from one, and type 2 response from another modality. This was the typical bimodality pattern of convergence, first described by Grusser and coworkers (Grusser und Grusser-Cornehls 1960; Grusser et al. 1959) in the visual cortex for visual stimuli causing specific responses of type 1 (mostly B, D or E responses) to illumination and darkness and for vestibular stimuli causing responses of type 2 (mostly &activation) to onset and cessation of labyrinthine polarization (Grusser et al. 1959). In view of many observations on arousal activation of cortical neurons over the reticular and thalamic nonspecific nuclei, it seems possible that the type 2 response is also a similar nonspecific activation. But no direct proof for the pathway of these responses over the reticulo-thalamic system is available. For the very common vesti-
MULTISENSORY CONVERGENCE ON CORTICAL NEURONS
225
bular type 2 response the evidence is indirect. We know from anatomical (Lorente de N b 1928; Szentagothai 1943) and physiological (Lorente de N b 1933; Duensing and Schaeffer 1957; Duensing 1961) evidence that the projection from the vestibular to the reticular neurons is the main relay station of ascending vestibular connections. Thus we have to modify Griisser’s first interpretation (Griisser und Griisser-Cornehls 1960) of the convergence of visual type I with vestibular type 2 responses which he considered to be mainly a specific regulation of the constancy and stability of the visual world by optovestibular coordination. As a similar convergence with vestibular type 2 responses is found in all other cortical regions investigated, their nonspecific role seems to be more important. In contrast, the direction-specific short latency responses which occur mainly in the vestibular area, could be more suitable for the coordination with visual and somatosensory functions. However, the oculomotor coordination is mainly a subcortical mechanism which brings the eyes into the adequate position and thus sends its final integrated results up to the cortex. Multimodality interaction The phenomena ofconvergent interaction of simultaneously applied stimuli of different modalities usually show occlusion and not summation. Facilitation seems to occur mainly when the dominant modality-specific responses corresponding to the cortical area are in some way decreased, be it by adaptation or by other factors. Both type 1 and type 2 vestibular responses showed the elevation of critical flicker frequency in light responsive neurons by other modality stimulation. For type 2 responses, the visual facilitation seems to correspond to the unspecific effects after thalamic or reticular stimulation, which have been described by Creutzfeldt and Griisser (1959) for flicker and by Akimoto and Creutzfeldt (1957/1958) for prolonged visual stimulation. Even if all the necessary experiments with various conditioning and test stimuli in different temporal sequences were performed for many neurons from these sense modalities, the analysis would require much time and rather complicated computing techniques. The exact investigation of these multimodality interaction effects seems difficult for two reasons: I . Analysis of responses representing 3 modalities with interacting temporal relations and interacting effects of stimulus variations will require a rather elaborate statistical treatment of data derived from many trials. 2. The results of these many trials might be influenced by conditioning processes, which might take effect during repeated exposures to combined sensory stimuli. Thus the normal neuronaf relations in the cortex could be contaminated by learning processes which were themselves influenced by the experimental procedures employed. At this preliminary stage of our investigations we would prefer not to classify the various interactions into enhancing, occluding or inhibiting types. How far they can be modified by arousal or conditioning remains an open question. It also is questionable whether we can apply to these unitary responses the terms used for behavioral responses of the whole animal, such as “arousal”, “alerting” or “conditioning”. In some neurons the responses to various stimuli were often small or References p . 231-234
2 26
K. J U N G et
al.
were initially absent, although they became gradually stronger with repeated stimulation. This delayed activation or potentiation which is the opposite of habituation, may occur with or without increased spontaneous activity. When it occurs with spontaneous activity, it may correspond to an arousal mechanism, which however seems to be independent of the arousal characteristics in the EEG. Such an activation could only be called unspecific, if all the modalities tested were facilitated. This is rarely so. In other neurons, the activation may be modality-specific. For the brief, transient type l a responses of trisensory neurons a similar facilitation of multisensory responses occurs without an increase of spontaneous discharge. This facilitation in some neurons depended on the duration of the stimulus, but did not depend on this in another neuron (described on p. 218) which showed a prolonged aftereffect of vestibular stimuli which caused type 1 responses. Although this facilitation did not correspond to other specific or nonspecific facilitations, it was also different from the classical conditioning experiments with combinations of stimuli of different modalities. However, the initial brief transient discharge of these trisensory neurons may be similar to the discharge of signalling neurons, described by Jasper et al. (1960) in their conditioning experiments, in that it occurs only at the onset of the conditioned stimulus. We leave the open questions whether this similarity, and also the delayed activation and facilitation phenomena, may not correspond to conditioning processes and whether conditioning may have some influence on multisensory interaction. With the exception of flicker stimulation, we have avoided any regularity in the succession of three-modality stimuli, which might provide a basis for the occurrence of conditioning. We have not yet found neurons which respond only to the combination of trimodality stimuli and which might correspond to the “decision units” postulated by Bullock (1961), but there has not yet been a sufficient search for such true trisensory units. All these correlations are speculative. They are preliminary and tentative conjectures, mentioned here only to stimulate discussion.
General signijicance of sensory convergence Although we should not jump too quickly from neuronal discharges to their significance in behavior, or to perceptual experience in man, a few words may be said about the possible significance of this multisensory integration. The existence of a neuronal system in the brain for receiving and storingmultisensory information is an obvious postulate of behavior studies and of psychological experience. If we look for parallels in animal behavior, we find that mu!tisensory stimuli are by far the most effective afferent inputs. Sensory messages from only one modality often remain without behavioral effect. But a combination of visual and acoustic stimuli may also not be sufficient to arouse attentive behavior responses i n animals and the addition of olfactory stimuli may be necessary to activate the response. It is well known that theabsence of the associated contribution from one sense organ (for example, the olfactory organ) makes an otherwise interesting and vivid sequence of visual and auditory perceptions quite uninteresting to an animal; thus a cat pays little attention to performance on a sound film or to images in a mirror. Our brain nearly always receives various kinds of information about the same thing
MULTISENSORY CONVERGENCE ON CORTICAL NEURONS
227
from different sense organs. This multisensory input is integrated by the brain, and it seems probable that it is influenced by attentive mechanisms. Different sense modalities can affect the same cortical neuron as we receive multisensory perceptions in our daily life. Of course, such an integration of sensory afferents at a few neurons can not yet be regarded as the neurophysiological basis of multisensory experience. The relation of attention to the unspecific reticular thalamic system is also still obscure. Moruzzi (1954), Von Baumgarten and Mollica (1954) and Amassian and De Vito (1954) were the first to describe similar integrations of different modalities and brain structures at neurons of the reticular formation. The so-called nonspecific ascending impulses that reach the cortex from the reticular thalamic system therefore contain multisensory components in their integrated messages. But this does not explain the organisation and differentiation of these afferent integrations in the various cortical areas. It is possible that only a part of the cortical multisensory neuronal effects travel over nonspecific afferents from the reticular formation or basal ganglia and this may be a source of the long latency type 2 responses. The short latency responses apparently reach the cortex over the long afferent pathways with relays in the specific thalamus, which probably have intrathalamic and interareal connections. Another possible significance of multisensory information might concern the temporal sequence of sensory events. Although the neurophysiological basis of the remarkable faculty of our perception and memory for a temporal order is still obscure, this order must have some neuronal equivalents. Multisensory inputs ending in various receiving areas at different neurons must be coordinated by common neuronal systems which are capable of temporal multisensory integration representing simultaneous and successive experiences. Experiments with conditioned reflexes have shown that temporary connections from different sense modalities can be made, not only by cortical, but also by subcortical, neuronal systems. It is therefore not surprising that these multisensory coordinations occur in subcortical as well as in cortical neurons. Filtering processes of multisensory information were mentioned by HernandezPe6n (1961) in relation to his experiments on the effects of attention and habituation in brain stem nuclei. The general significance of intersensory communication and plurivalent neurons with heterosensory convergence has been discussed extensively by Fessard (1961).
Preselective mechanisms in sensory attention and multirnodal redundancy : von Holst’s models of reafferenceand cybernetics Multisensory information and coordination may be important, not only for feedback mechanisms, but also for preselecting mechanisms in the regulation of sensory attention. This is not feedback, but rather “feed-forward”, as MacKay (1961) has remarked: it means “preselecting of activity in response to signs of forthcoming demand”. Such a mechanism would work in some way opposite to von Holst’s principle of reafference (von Holst and Mittelstaedt 1950). Otherwise MacKay’s matching processes are similar to von Holst’s model of reafference, but were apparently conceived independently of the latter. References I. 231-134
228
R. JUNG et
al.
Some of the most simple and useful models for certain feedback systems made on the basis of physiological results, were those of von Holst and Mittelstaedt, (Von Holst and Mittelstaedt 1950; Mittelstaedt 1960), which demonstrated the principle of Reufferenz. I was surprised that nobody at this conference mentioned von Holst’s sensorimotor coordination models with reafference. To elucidate the basic phenomena of sensorimotor and multisensory integration, von Holst’s models may be used advantageously, especially for optokinetic processes and visuo-vestibular integration, Possibly they might also be adapted to cybernetic electronic devices, although we are still ignorant of the physiological nature of the corollary discharge in von H o b ’ s Efferenzkopie. Von Holst’s models have the advantage that people who are relatively untrained mathematically can understand them. I myself prefer a clear Anschuuung or a visualized picture of physiological processes rather than abstract mathematical formulations or the statistical results of electronic computers. Our conceptions of the physiological processes with which we are dealing experimentally are in themselves a product of multisensory perceptions and of past experience. Unfortunately we can not simulate such a perceptive process without using complex machines. It seems probable that we shall need them too for processing experimental data from multisensory stimuli. But this does not mean that the nervous system uses the same cybernetics. Von Holst’s models or metaphors are nearer to Anschauung than to abstract mathematical formulations. These mechanisms are designed to match the expected sensory input by the cancellation of reafferent against corollary messages caused by efferent discharges. Thus von Holst’s mechanisms reduce the redundancy, but multisensory convergence may increase redundancy. Sensory information which results in final perception must have gone through many physiological processes. Some of these processes may be of the von Holst kind, others may be different. But it is reasonably sure that not all of them follow the rules of communication engineers. Multisensory information is redundant and this redundancy may be very useful in sensory experience for the selection of significant sensory data. Our results have shown that such multisensory convergence is effective, not only in subcortical centers, but also in the neuronal systems of the cortex. In contrast to my younger collaborators (H.K. and J.F.) I am not optimisticenough to expect that application of cybernetic principles will explain the functions of the brain. The superficial analogy of digital principles to the all-or-none response in neurons could only be applied to axonal conduction. I am not sure that analogous models would not be more helpful i n understanding the function of neuronal systems. All these comparisons do not provide enough evidence to confirm the assumption that the brain works as a digital or analogue computer. This does not mean that 1 am against the application of cybernetic models. But we have to realize their limitations and their basis in other than physiological principles. Since many of the very simple devices of mechanic and electronic techniques are never used by organisms it seems less likely that the brain uses mechanisms similar to the more elaborate apparatus of electronic engineers. With the increasing complexity of their computers, loosely called electronic brains, the probability decreases that
MULTISENSORY CONVERGENCE ON CORTICAL NEURONS
229
such complicated technical systems would correspond to the complex neuronal circuits. However, the logical principles implied by computers at the present time, or more adequate logics developed in the future might give a better solution. It might well be that such logistics would be somewhat similar to those present in the brain. Such general correlations might be independent of the essential differences between brains and computers. The brain with its spontaneous electrical and metabolic activities and their chemical mechanisms has many other features which are entirely foreign to computers. In spite of this scepticism I gladly accept computer machines for the analysis, processing, and integration of our experimental data. But this slave work of the machines will not produce new concepts. However, 1 think it is useful for neurophysiology to build models of neuronal circuits and to observe experimentally what they can do. Multisensory convergence and sensorimotor coordination It seems evident that multisensory integration is important for motor and behavioral regulations. But in contrast to the successful correlations, found between vision and neuronal activity (Jung 1961a,b), we can not yet correlate such complex sensorimotor processes with the elementary events at the neuronal level. Similar findings of sensory convergence in subcortical structures, e.g. in the reticular formation (Von Baumgarten und Mollica 1954; Moruzzi 1954; Scheibel et al. 1955), thalamic (Albe-Fessard and Gillett 1961) and caudate (Albe-Fessard et al. 1960) nuclei may indicate the direction of further investigations. They suggest that integrations of several modalities occur first at lower levels and are not peculiar to cortical neurons. Some of these lower multisensory convergence effects may be conducted in a simpler integrated form to cortical neurons as activation or inhibition. The latter seems to be the main integrative effect of the multisensory input to the caudate nucleus (AlbeFessard et al. 1960), which in its turn inhibits the neurons of the motor cortex (Spehlmann et al. 1960). Similar transformations of multisensory channels may occur in the thalamus and may be the basis of thalamic and sensory influence on cortico-motor neurons recorded by Li (1956, 1959). It seems premature to use these few results of neuronal recordings with multimodal convergence as a basis for a general hypothesis of brain functions and also for support of localistic or holistic views. These two aspects are not mutually exclusive, but they do supplement each other. This is well shown by the results of cortical neuronal physiology which demonstrate both point-to-point projection in the primary receiving areas (Mountcastle 1957; Hubel and Wiesel 1959) and convergence of several afferents in the neurons of these and other cortical fields (Amassian 1954; Jung 1958; Lomo and Mollica 1959; Buser and Imbert 1961). Before we can really understand the significance of multisensory and sensorimotor coordination in the brain, many more experiments on these complex neuronal systems in the cortex and subcortex must be done.
References P. 231-234
230
R. JUNG
et a/.
SUM MARY
1. In the superior convolutions of the cat’s cortex the responses of 450 neurons to optic, vestibular and acoustic stimuli were recorded. The optic stimuli were diffuse illumination through closed lids to avoid the effects of pattern stimulation during eye movements. Acoustic stimuli were whistles, clicks and clapping. Vestibular stimuli were round window polarization or calorization. 2. Two main types of neuronal responses to these sensory stimuli were found: Type I: immediate responses which showed, according to the duration of the stimulus, a brief, relatively constant latency with d’emblke-activation or inhibition. Type 2: delayed activation responses which showed a long and inconstant latency and varying intensity, mostly with recruiting activation and prolonged duration after stimulus cessation. 3. A wide distribution of visual type 1 responses was found. In the parastriate and paraacoustic areas of the medial and posterior suprasylvian and posterior ectosylvian gyrus, the various response patterns to illumination were similar to those in the primary visual cortex, mostly B (on), D (off) and E (on-off). Type 1 acoustic responses, although they were most common in the auditory area, were also encountered in other areas investigated, mainly in the vestibular cortex. 4. Type 1 responses to labyrinthine polarization of the shortest latency (4-20 msec) were concentrated in the primary vestibular area. They occurred also with slightly longer latencies in the auditory area. Type 2 delayed activations to labyrinthine polarization were distributed over all cortical regions when investigated with higher current intensity. Their relation to nonphysiological stimulus intensity and nonspecific activation is discussed. The vestibular responses of the association area consisted of initial inhibition of discharge in about half of the neurons; they were independent of current direction in the labyrinth and diminished with repeated stimulation. 5. Vestibular inhibition of cortical neurons was seen after weak cathodic or anodic labyrinthine polarization, mainly in the vestibular and “association” cortex, but not in the visual area. This neuronal inhibition may be reversed to activation by stronger polarization currents of 3-10 times threshold. 6. Convergence from several sense modalities on the same neurons was found in all cortical areas investigated. Bisensory responses were common, trisensory responses were rare. Neuronal convergence occurred also in primary receiving areas with characteristic prevalence of their specific modality, which showed type 1 responses, mostly combined with type 2 responses to one or two other modalities. 7. Trisensory neurons of type 1 responses to three modalities were found in about 11{, of the recorded neurons or 30/, of the neurons fully tested. They occurred in primary vestibular, somatic and visual areas. The 5 neurons of this type showed little spontaneous activity, but did show brief transient high frequency responses to optic, acoustic and vestibular stimuli. 8. Combination of visual type 1 responses with type 2 activation to labyrinthine polarization was most common in the visual and paravisual areas. Vestibular responses of type 1 in the vestibular and auditory areas were more often combined with auditory
MULTISENSORY CONVERGENCE ON CORTICAL NEURONS
23 1
responses of type 1 , than with visual type 1. or type 2 responses. Auditory type I responses were mostly combined with vestibular type 1 or type 2 responses in the auditory area. 9. The interaction of different modality stimulation, applied simultaneously, was studied only in a small number of neurons. So far as it was recorded, it showed a preponderance of the type 1 response. Type 2 activation often enhanced the sustained discharge caused by other modalities, but had little effect on primary specific activation. Type 2 activation does not abolish specific inhibition of type 1, e.g. initial inhibition by light in D-neurons, but is able to activate these neurons in later stages of light inhibition. Occlusion during multisensory stimulation is more common than is facilitation or inhibition. ZO. The results are discussed in relation to the nonspecific brain stem system, to multisensory influence in behavior and to subjective experience. Some speculations are added on the possible significance for mechanisms of attention, temporal sequence of sensory events, and sensorimotor coordination, including conditioned responses. REFERENCES AKIMOTO, H. und CREUTZFELDT, 0. Reaktionen von Neuronen des optischen Cortex nach elektrischer Reizung unspezifischer Thalamuskerne. Arch. Psychiat. Nervenkr., 1957158, 196: 494-519. ALBE-FESSARD, D. et GILLETT, E. Convergences d’afferences d’origines corticale et peripherique vers le centre median du chat anesthesie ou BveillB. Electroenceph. clin. Neurophysiol., 1961, 13: 257-269. ALBE-FESSARD, D., ROCHA-MIRANDA, C. et OSWALDO-CRUZ, E. Activites evoquees dans le noyau caude du chat en reponse a des types divers d’afferences. 11. Etude microphysiologique. Electroenceph. clin. Neurophysiol., 1960, 12: 649-661. AMASSIAN, V. E. Studies on organization of a soniesthetic association area, including a single unit analysis. J. Neurophysiol., 1954, 17: 39-58. V. E. and D E VITO, R. V. Unit activity in reticular formation and nearby structures. AMASSIAN, J. Neurophysiol., 1954, 17: 515-603. ARDEN,G. B. and SODERBERG, U. The transfer of optic information through the lateral geniculate body of the rabbit. In W, A. ROSENBLITH (Editor), Sensory Communication, Symposion. The M.I.T. Press, John Wiley Sons, New York, London, 1961: 521-544. BAUMGARTEN, R. VON und MOLLICA, A. Der Einfluss sensibler Reizung auf die Entladungsfrequenz kleinhirnabhangiger Reticulariszellen. Pflugers Arch. ges. Physiol., 1954, 259: 79-96. BAUMGARTNER, G. Die Reaktionen der Neurone des zentralen visuellen Systems der Katze im (Herausgeber), Neurophysiologie simultanen Helligkeitskontrast. In R. JUNGund H. KORNHUBER und Psychophysik des visuellen Systems, Symposion. Springer, Berlin, Gottingen, Heidelberg, 1961 : 296-3 1 I . BORENSTEIN, P., BRUNER, J. and BUSER, P. Organisation neuronique et convergence heterosensorielles dans le complexe lateral posterieur “associatif” du thalamus chez le chat. J. Physiol. (Paris), 1959, 5 1 : 413-414. BULLOCK, T. H. The problem of recognition in an analyzer made of neurons. In W. A. ROSENBLITH (Editor), Sensory Communication, Symposion. The M.I.T. Press, John Wiley & Sons, New York, London, 1961 : 717-724. BUSER,P. Observations sur I’organisation fonctionnelle du cortex moteur chez le chat. Bull. Acad suisse Sci. mdd., 1960, Fasc. 5. 355-397. BUSER, P. et BORENSTEIN, P. Riponses somesthesiques, visuelles et auditives, recueillies au niveau du cortex “associatif” suprasylvien chez le chat curarise non anesthesii. Electroenceph. clin. Neurophyiol., 1959, I 1 : 285-304. BUSER,P., BORENSTEIN P. et BRUNER, J. Etude des systemes “associatifs” visuels et auditifs chez le chat anesthesie au chloralose. Electroenceph. elin. Neurophysiol., 1959, I I : 305-342.
232
R. JUNG et al.
BUSER,P. and IMBERT, M. Sensory projections to the motor cortex in cats: a microlectrode study. In W. A. ROSENBLITH (Editor), Sensory Communication, Symposion. The M.I.T. Press, John Wiley Sons, New York, London, 1961: 597-626. 0. und GTUSSER, 0.-J. Beeinflussung der Flimmerreaktion einzelner corticaler CREUTZFELDT, Neurone durch elektrische Reize unspezifischer Thalamuskerne. In : Proc. 1st. Znt. Congr. neurol. Sci., Bruxelles, Vol. III, Clinical Neurophysiology and Epilepsy. Pergamon, London, 1959: 349-355. CREUTZFELDT, O., SPEHLMANN, R. und LEHMANN, D. Veranderung der Neuronaktivitat des visuellen Cortex durch Reizung der Substantia reticularis mesencephali. In R. JUNCund H. KORNHUBER (Herausgeber), Neurophysiologie und Psychophysik des visuellen Systems, Symposion. Springer, Berlin, Gottingen, Heidelberg, 1961 : 351-363. DOTY,R. W. Potentials evoked in cat cerebral cortex by diffuse and by punctiform photic stimuli. J . Neurophysiol., 1958, 21 : 437464. DOTY,R. W. Functional significance of the topographical aspects of the retino-cortical projection. In R. JUNGund H. KORNHUBER (Herausgeber): Neurophysiologie und Psychophysik des visuellen Systems, Symposion Freiburg. Springer, Berlin, Gottingen, Heidelberg, 1961 : 228-247. DUENSING, F. Die Erregungskonstellationen im Rautenhirn des Kaninchens bei den Labyrinthstellreflexen (Magnus). Ntrturwissenschaften, 1961, 48: 681-690. DUENSINC, F. und SCHAEFER, K.-P. Die Neuronenaktivitat in der Forniatio reticularis des Rhombencephalons beim vestibularen Nystagmus. Arch. Psychiat. Nervenkr., 1957, 196: 265-290. A. The role of neuronal networks in sensory communications within the brain. In W. A. FESSARD, ROSENBLITH (Editor). Sensory Communication, Symposion. The M.I.T. Press, John Wiley Sons, New York, London, 1961: 585-606. X. Unit activity of rabbit hippocampus. Amer. J. Physiol., 1955, 181: GREEN,J. D. and MACHNE, 2 19-224. GROSSER, 0 . J. und GRUSSER-CORNEHLS, U. Mikroelektrodenuntersuchungen zur Konvergenz vestibularer und retinaler Afferenzen an einzelnen Neuronen des optischen Cortex der Katze. Pfugers Arch. ges. Physiol., 1960, 2 7 0 : 227-238. GROSSER,0. J. und GROSSER-CORNEHLS, U. Reaktionsmuster einzelner Neurone im Geniculatum laterale und visuellen Cortex der Katze bei Reizung mit optokinetischen Streifenmustern. In R. JUNGund H. KORNHUBER (Herausgeber), Neurophysiologie und Psychophysik des visuellen Systems, Symposion. Springer, Berlin, Gottingen, Heidelberg, 1961 : 313-326. GROSSER,0. J., GRUSSER-CORNEHLS, U. und SAUR,G. Reaktionen einzelner Neurone im optischen Cortex der Katze nach elektrischer Polarisation des Labyrinths. Pfugers Arch. ges. Physiol., 1959, 269: 593-612. GROSSER-CORNEHLS, U. und GROSSER,0.J. Mikroelektrodenuntersuchungen am Geniculatum laterale der Katze: Nervenzell- und Axonentladungen nach elektrischer Opticusreizung. Pfliigers Arch. ges. Physiol., 1960, 271: 50-63. GROSSER-CORNEHLS, U. und GROSSER,0. J. Reaktionsmuster der Neurone im zentralen visuellen System von Fischen, Kaninchen und Katzen auf monoculare und binoculare Lichtreize. In R. JUNG und H. KORNHUBER (Herausgeber), Neurophysiologie und Psychophysik des visuellen Systems, Symposion. Springer, Berlin, Gottingen, Heidelberg, 1961 : 275-286. GUALTIEROTTI, T. Changes of visual pattern discrimination in adult cats after removal of an aspecitic area of cerebral cortex. Amer. J . Physiol., 1961, 200: 1215-1218. GUDDEN, B. VON. Gesammelte und hintedassene AbIiuf7d~~~~igen. J. F. Bergmann, Wiesbaden, 1889. HASSLER, R. uber die afferenten Bahnen und Thalamuskerne des motorischen Systems des Grosshirns. I. Mitteilung. Bindearm und Fasciculus thalamicus. Arch. Psychiat. Nervetpkr., 1949, 182: 759-785. HERNANDEZ-PE~N, R. Reticular mechanisms of sensory control. In W. A. ROSENBLITH (Editor), Sensory Communicution, Symposium. The M.I.T. Press, John Wiley & Sons, New York, London, 1961 : 497-520. HERNANDEZ-PE~N, R. and HAGBARTH, K.-E. Interaction between afferent and cortically induced reticular responses. J. Neurophysiol., 1955, 18: 44-55. HOLST,E. VON und MITTELSTAEDT, H. Das Reafferenzprinzip. Nuturwissenschufen, 1950,37: 256-272. HUBEL,D. H. Single unit activity in lateral geniculate body and optic tract of unrestrained cats. J . Physiol. (London), 1960, 150: 91-104. HUBEL,D. H. and WIESEL,T. N. Receptive fields of single neurones in the cat striate cortex. J . Physiol. (London), 1959, 148: 574-591.
MULTISENSORY CONVERGENCE ON CORTICAL NEURONS
233
JASPER, H. H. Recent advances in our understanding of ascending activities of the reticular system. In H. H. JASPER, L. D. PROCTOR, R. S. KNIGHTON, W. C. NOSHAY and R. T. COSTELLO (Editors), Reticular formation of the brain. Internat. Symposion, Little, Brown and Co., Boston, Toronto, 1958: 319-331. JASPER,H. H., RICCI, G. and DOANE,B. Microelectrode analysis of cortical cell discharge during avoidance conditioning in the monkey. In H. H. JASPER and G. D. SMIRNOV (Editors), The Moscow Colloquium on electroencephalography of higher nervous activity. Electroenceph. elin. Neurophysiol., 1960, SUPPI.13: 137-155. JUNG,R. Coordination of specific and nonspecific afferent impulses at single neurons of the visual cortex. In H. H. JASPERet a/. (Editors), Reticular formation of the brain. Little, Brown & Co., Boston 1958: 423-434. JUNC,R. Neuronal integration in the visual cortex and its significance for visual information. In W. ROSENBLITH (Editor), Sensory communication. The M.I.T. Press, John Wiley & Sons, New York, London, 1961a: 627-674. JUNC,R. Korrelationen von Neuronentatigkeit und Sehen. In R. JUNCund H. KORNHUBER (Herausgeber), Neurophysiologie und Psychophysik des visuellen Systems, Symposion. Springer, Heidelberg, 1961b: 410434. JUNG,R., VON BAUMGARTEN, R. und BAUMCARTNER. G. Mikroableitungen von einzelnen Nervenzellen im optischen Cortex: die lichtaktivierten B-Neurone. Arch. Psychiat. Nervenkr., 1952, 189: 52 1-529. KORNHUBER, H. H. and DAFONSECA, J. S. Convergence of vestibular, visual and auditory afferents at single neurons of the cat’s cortex. Fifth International Congress of Electroencephalography and Clinical Neurophysiology, Rome, Sept. 1961. Excerpta Medica, Intern. Congr. Ser. No. 37, 1961a: 14-15. KORNHUBER, H. H. and DA FONSECA, J. S. Opto-vestibular integration in the cat’s cortex. In M. BENDER (Editor), Symposium on the oculo-motor system. P. Hoeber, New York, 1961b. In press. S. Convergence of tactile, thermal and gustatory impulses on single cortical cells. Acta LANDGREN, physiol. scand., 1957, 40: 210-221. LANDGREN, S. The response of thalamic and cortical neurons to electrical and physiological stimulation of the cat’s tongue. In W. A. ROSENBLITH (Editor), Sensory Communication. The M.I.T. Press, John Wiley Sons, New York, London, 1961: 437-453. LEHMANN, D. and KOLJKKOU, M. Unpublished experiments. LI, C.-L. The inhibitory effect of stimulation of a thalamic nucleus on neuronal activity in the motor cortex. J. Physiol. (London), 1956, 133: 40-53. Lr, C.-L. Some properties of pyramidal neurons in motor cortex with particular reference to sensory stimulation. J. Neurophysiol., 1959, 22: 385-394. LOMO,T. e MOLLICA, A. Attivita di singole uniti della corteccia ottica primaria durante stimolazioni luminose, acustiche, olfattive e dolorifiche, nel coniglio senza narcosi. Boll. SOC.ital. Biol. sper., 1959, 35: 1879-1882. LORENTE DE NO, R. Die Labyrinthreflexe auf die Augenmuskeln nach einseitiger Labyrinthextirpation nebst einer kurzen Angabe iiber den Nervenmechanismus der vestibularen Augenbewegungen. Urban und Schwarzenberg, Wien, 1928. LORENTE DE NO, R. Vestibulo-ocular reflex arc. Arch. Neurol. Psychiat., 1933,30: 245-291. MACHNE, X. and SECUNDO, J. P. Unitary responses to afferent volleys in amygdaloid complex. J. Neurophysiol., 1956, 19: 232-240. MAC KAY,D. M. The science of communication. Inaugural lecture. KeeIe, University College of North Staffordshire, 1961. MAGOUN, H. W. The waking brain. Charles C. Thomas, Springfield, Ill., 1958. A. Microelectrode recording from midbrain reticular MANCIA,M., MECHELSE, K. and MOLLICA, formation in the decerebrate cat. Arch. ital. Biol.,1957, 95: 110-1 19. MARSHALL, W. H., TALBOT S. A. and ADES,H. W. Cortical response of the anaesthetised cat to gross photic and electrical afferent stimulation. J. Neurophysiol., 1943, 6 : 1-5. MARUYAMA, N. and KANNO,Y. Experimental study on functional compensation after bilateral removal of auditory cortex in cats. J. Neurophysiol., 1961, 2 4 : 193-202. MITTELSTAEDT, H. The analysis of behavior in terms of control systems. 5th Conference on group processes, 1958. Conference transaclions, Josiah Macy Jr., 1960. MORUZZI, G. The physiological properties of the brain stem reticular system. In J. F. DELAFRESNAYE (Editor), Brain mechanisms and consciousness, Blackwell, Oxford, 1954 : 21-48.
234
K. JUNG et al.
MomTcAsTLE, V. B. Modality and topographic properties of single neurons of cat’s somatic sensory cortex. J. Neurophysiol., 1957, 20: 408434. MOUNTCASTLE, V. B. Some functional properties of the somatic afferent system. In W. A. ROSENBLITH (Editor), Sensory Communication, Synrposion. The M.I.T. Press, John Wiley & Sons, New York, London, 1961: 403436. R. Uber Aufbau und Gliederung der corticalen Sehsphare bei der Katze. OTSUKA, R. und HASSLER, Arch. Psychiut. Nervenkr., 1962, 203: 212-234. B. and JASPER,H. H. Microelectrode studies of conditioning: technique and Riccr, G., DOANE, preliminary results. Congr. int. Sci. neurol., Bruxelles, 1957: 401415. A. The brain stem reticular formation. Anatomy and physiology. ROSSI,G. F. and ZANCHETTI, Arch. i d . Biol., 1957, 95, Fasc. 3-4: 199438. A. B., MOLLICA,A. and MORUZZI,G. Convergence and interaction of SCHEIBEL, M. E., SCHEIBEL, afferent impulses on single units of reticular formation. J. Neurophysiol., 1955, 18: 309-331. J. P. and MACHNE,X. Unitary responses to afferent volleys in lenticular nucleus and SEGUNDO, claustrum. J. Neurophysiol., 1956, 19: 325-339. S. and BENDER, M. B. Oculomotor responses on vestibular stimulation of monkeys with SHANZER, lesions of the brain-stem. Bruin, 1959, 82: 669-682. SPEHLMANN, R., CREUTZFELDT, 0. D. und JUNG, R. Neuronale Henimung im motorischen Cortex nach elektrischer Reizung des Caudatum. Arch. Psychiut. Nervenkr., 1960, 201 : 332-354. SZENTAGOTHAI, J. Die zentrale Innervation der Augenbewegungen. Arch. Psychiut. Nervenkr., 1943, 116: 721-760. TASAKI, I., POLLEY, E. H. and ORREGO, F. Action potentials from individual elements in cat geniculate and striate cortex. J. Neurophysiol., 1954, 17: 454-474. THOMPSON, R. F. and SINDBEKG, R. M. Auditory response fields in association and motor cortex. J. Neurophysiol., 1960, 23: 87-105. VASTOLA,E. F. A direct pathway from lateral geniculate body to association cortex. J. Neurophysiol., 1961, 2 4 : 469487. WALZL,E. M. and MOUNTCASTLE, V. Projection of vestibular nerve to cerebral cortex of the cat. Amer. J . Physiol., 1949, 159: 595-603. WIDBN,L. and AJMONEMARSAN, C. Effects of corticopetal and corticofugal impulses upon single elements of the dorsolateral geniculate nucleus. Exp. Neurol., 1960, 2 : 468-502. C . N. Organization of cortical auditory system. In W. A. ROSENBLITH (Editor), Sensory WOOLSEY, Comrnunica/ion, Symposion. The M.I.T. Press, John Wiley & Sons, New York, London, 1961: 235-258.
DISCUSSION 0. POMPEIANO:
Since the existence of a specific cortical vestibular area is well known in the literature, it seems to me that one of the most interesting aspects of your work is the demonstration of non-specific vestibulocortical projection. This fact seems to agree with the observations of those authors (Spiegel 1934; Price and Spiegell937; Gerebtzoff 1939, 1940), who found that labyrinthine stimulation may change the electrical activity of the whole cortex. Since galvanic stimulation of the labyrinth as used in your experiments does not give any information about the types of receptors stimulated, it would be of interest to study how the non-specific vestibulo-cortical projection is influenced by the individual branches of the vestibular nerve. These branches might be selectively stimulated by the technique of Anderson and Gernandt (1954). ANDERSSON, B. and GERNANDT, B. E. Cortical projection of vestibular nerve in cat. Actu oto-laryng (Stockh.), 1954, SUPPI. 116: 10-18. GEREBTZOFF, M. A. Des effets de la stimulation labyrinthique sur I’activite electrique de I’ecorce cerebrale. C.R. SOC.Biol. (Paris), 1939, 131: 807-813. GEREBTZOFF, M. A. Recherches sur la projection corticale du labyrinthe. I. Deseffets de la stimulation labyrinthique sur I’activite electrique de I’ecorce ctrebrale. Arch. int. Physiol., 1940, 50: 59-99.
DISCUSSION
235
PRICE,J. B. and SPIEGEL, E. A. Vestibulocerebral pathways. A contribution to the central mechanism of vertigo. Arch. Otolaryng. (Chicago), 1937, 26 : 658-667. E. A. Labyrinth and cortex. The electroencephalogram of the cortex in stimulation of SPIEGEL, the labyrinth. Arch. Neurol. Psychiat. (Chicago), 1934, 3 I : 469482.
P. DELL: Can you explain the increase in the frequency of the electrical activity in the contralateral cortex after an anodic labyrinthine polarisation? We have done similar experiments with Mrs. Dumont and observed that progressive cathodic polarisation produced an intense arousal in both cortices, while a progressive anodic polarisation produced a bilateral cortical arousal, although this occurred after a much longer delay and only when the polarizing current reached 1 mA or more. We looked for an inhibition of the visual evoked potential in response to stimulation of the chiasma during the polarisation of the labyrinth, but the results have never been well defined; the non-specific cortical arousal and its overwhelming facilitatory effect mask the possible inhibitory effects of the labyrinthine stimulation.
G. MORUZZI: 1 should like first of all to congratulate Professor Jung on his fine presentation and on the beautiful records we have seen. I would ask him whether pupillary changes were observed during labyrinthine stimulation. If so, I think that a distinction should be made between the responses recorded in complete darkness, which obviously were elicited directly by labyrinthine stimulation, and the effects of the convergence of visual and vestibular impulses, which may be influenced by changes in the pupillary diameter. Such a distinction was also made by Lomo and Mollica (1962), who investigated the effects of photic, auditory and painful stimulations on the units of rabbit’s visual cortex (Lomo and Mollicd 1959). Perhaps the effect of the convergence of visual and vestibular impulses on cortical neurones might be approached with the technique of the darkened contact lenses, provided with a sagittal fissure in the middle which simulates a fixed myotic pupil. Local atropine has several physiological and pharmacological drawbacks and may be advantageously replaced by severance of the ciliary nerves (see Affanni et al., 1962). AFFANNI, J., MANCIA, M. and MARCHIAFAVA, P. L. Role of the pupil in changes in evoked responses along the visual pathways. Arch. ital. Biol., 1962, 100: 287-296. LOMO, T. and MOLLICA, A. Attivita di singole unita della corteccia ottica primaria durante stimolazioni luminose, acustiche, olfattive e dolorifiche, nel coniglio senza narcosi. Boll. Soc. iral. Biol. sper., 1959, 35: 1879-1882. LOMO,T. and MOLLICA, A. Activity of single units in the primary optic cortex in the unanaesthetized rabbit during visual, acustic, olfactory and painful stimulation. Arch. ital. Biol., 1962, 100: 86-120.
P. BUSER: 1 should like to make the following points about the topographical distribution of non-primary visual responses on the cortex. They have resulted from our experiments, done in collaboration with Drs. Bruner and Sindberg, which have a relationship to the excellent presentation of Professor Jung. 1. On the curarized and unanesthetized preparation responses to visual stimulation are obtained from a very large number of cortical points in associative or motor areas. Considering the variations in form of these responses as a function of their point of origin, we have had the impression, that several distinct categories of projection exist, but we have not been able yet to prove this. 2. Analogous topographical studies carried out under chloralose anesthesia have confirmed this point of view. 3. Under chloralose, topographical analysis is made much easier by the fact that the form of the responses is particularly stereotyped; we have been able to distinguish two categories of visual response in the suprasylvian region. The first type is localized in two regions, one anterior and the
236
R. JUNG et al.
other posterior, which are located along the line of demarcation between the suprasylvian and ectosylvian gyri. The latency of these responses is relatively short, although it is longer than that of the primary evoked potential. The form of these secondary responses resembles precisely that of the primary evoked potentials. A second type of visual response can be recorded from the median portion of the suprasylvian gyrus; but this response has a long latency, it is relatively monophasic (surface positive) and closely resembles the responses from the motor cortex which are extremely fragile and sensitive to fatigue. It seems to us more than probable that these two categories of response in the suprasylvian region represent, from the anatomical point of view, two distinct systems of projection.
W. GREYWALTER: The evidence presented by Prof. Jung of the subtle interaction between B and D units in temporal and spatial transitions from light to dark reminds me of the EEG effect described by Evans and Gastaut, now called “lambda waves”. These are blunt monophasic waves that appear in random sequence in the posterior regions when a subject scans an intricate pattern. They are associated with saccadic eye movements, but not directly so, and it was suggested some years ago that they might be signs of rapid excitation-inhibition switching in the visual system as a blanking effect to prevent blurring of the field during scanning movements of the eyes. That some such process must exist is suggested by the simple experiment of trying to see your own eyes move when you look from one eye t o the other in your reflection in a mirror - you can’t. GASTAUT, Y . , Un signe electroencephalographique peu connu : les pointes occipitales survenant pendant I’ouverture des yeux. Rev. Neurol., 1951, 84: 640-693.
H. GASTAUT: Returning to the preceding remark of Grey Walter, 1 would like to recall, in connection with Professor Jung’s very excellent report, the existence in man of lambda waves. I have had the occasion to demonstrate, first with my wife in 1951, then with Alvim-Costa in 1957, that these waves always represent evoked potentials from the occipital cortex in response to afferent impulses of retinal origin which depend upon a saccadic movement of the globes displacing the macula on a strongly contrasting image. In this manner a cinematographic projection in black and white on a bright screen provokes lambda waves in all subjects (Gastaut and Bert 1954) when the picture presented maintains the sustained attention of the subject and also a rapid movement of the eyes (for example, when a film sequence showing soldiers marching is projected and causes a true optokinetic nystagmus). It is remarkable that, when a similar optokinetic nystagmus is caused by watching the motion of a cylinder provided with vertical dark and light bands similar, lambda waves are not provoked in the same subject, a fact which brings out clearly the part played by visualattention in the induction of these evoked potentials to light in man. 1 would like to draw attention again to the part played by the occipital and pre-occipital cortex in the elaboration of optokinetic nystagmus, with reference to the oculo-clonic attacks described by me in 1953, which occur in certain subjects during epileptic discharges in the occipital region and which represent a veritable “epileptic nystagmus.” GASTAUT, H., ALVIM-COSTA, C., GASTAUT, Y. et ALVIM-COSTA, M. R. MCcanisme des potentielsoccipitaux evoqu6s par les mouvements saccades des yeux. Acta physiof. pharmacol. neerf., 1957, 6 : 515-525.
GASTAUT, H . et BERT, J., EEG changes during cinematographic presentation. Electroenceph. d i n . Neurophysiol., 1954, 6 : 433444.
H. H. JASPER: Dr. Jung and his coworkers have presented us with a beautiful picture of sensory interaction a t the single neuronal level, which shows that so called “specific” cortical areas are truly areas in which
DISCUSSION
237
much convergence of impulses, specific and non-specific, does occur. I would like to raise the question as to what degree this interaction is excitatory or inhibitory in character.
D. ALBE-FESSARD:
I would like to draw Professor Jung’s attention to some results from our laboratories which confirm the findings which he has described concerning the inhibition of evoked activities in the motor cortex by stimulation of the caudate nucleus. G. Krauthamer, in my laboratory, using macroelectrode derivations, has shown that a short train of stimuli applied to the caudate nucleus suppresses electively (under chloralose anesthesia) the evoked responses which arrive over the extralemniscal pathway, but leaves intact the responses which arrive over lemniscal pathways. A. ARDUINI:
In the Symposium held in Freiburg last year it was suggested that the “off”-discharges might have an inhibitory or a generally suppressory effect, wiping out whatever activity the “on”-discharges had caused, that is to say, it was suggested that they had a kind of preparatory function and cleared the circuits for the reception of further incoming information. I wonder whether Prof. Jung has more comments to make about this problem of the significance of the “ofF”-responses.
F. BREMER: The research from the Freiburg laboratory, which our colleague Professor Jung has just summarized in such an excellent manner, has demonstrated the surprising generality of the phenomena of neuronal convergence at the cortical level. Without doubt some of these convergent phenomena are satisfactorily explained by reference to non-specific impulses arriving from the ascending reticular system. But this explanation is not valid when the converging sensory impulses elicit responses of very short latency, as happens notably in the area of polysensory projection to the anterior ectosylvian gyrus of the cat, which we have, with Bonnet and Terzuolo, hypothetically identified with auditory area 111 in the dog. On the other hand, non-specific vestibular projections have not, in microphysiological records, as well as in the macroelectrode records of Gerebtzoff, that general diffusion to the entire convexity of the cortex which characterizes activations of reticular origin. To the polysensory projection areas already known, which are situated outside the primary classical receiving areas, should be added the area which has just been discovered in the cat by Desmedt (and confirmed by Woolsey) in the anterior sylvian region. This area, the physiological significance of which remains to be studied, receives convergent afferent impulses from both auditory and visual stimuli and gives rise to potentials of very short latency. The more one studies the functional organization of the cerebral cortex, the more one is impressed and confused by the richness of the sensory information which is integrated within its neuronal circuits. BREMER, F., BONNET, V. et TERZUOLO, C . Etude electrophysiologique des aires auditives corticales du chat. Arch. int. Physiol., 1954, 62: 390-428. DESMEDT, J. E. MPm. Acad. roy. M i d . Belg., 1959, 4 : 133. GEREBTZOFF, M. A. Recherches sur la projection corticale du labyrinthe. I. Ses effets de la stimulation labyrinthique sur I’activite electrique de l’ecorce cerebrale. Arch. int. Physiol., 1940, 50: 59-99.
R. JUNG’s replies May I answer these questions one after the other? To 0 .Pompeiano The general alterations in cortical potentials, described by Spiegel and others, seem to be correlated with the widespread type 2 neuronal responses to labyrinthine stimuli, which cause activation in all
238
R. JUNG et
al.
the cortical areas which have been investigated. Dr. Pompeiano’s suggested use of Gernandt’s technique for the stimulation of various vestibular branches and of correlating it with the somewhat different locations of the semicircular canals on the cortex provides a good programme for further neuronal studies.
To P . Dell Dell’s observations on the prevailing activation would correspond to our type 2 vestibular responses. We have not found consistent differences between anodic and cathodic polarisation, except that cathodic currents are usually effective at a lower threshold. This difference is only observed in type 2 responses. Type 1 responses, with, direction specific effects, may also show low threshold effects for anodic polarisation. In some neurones the anodic current activates and the cathodic inhibits neuronal discharges and vice versa (Fig. 3). We agree, that some of the inhibitory effects found in type 1 responses might be masked by the prevailing activation caused either by arousal or by additional type 2 responses, especially when strong polarisation currents are used. Such a reversal of inhibition into activation by high intensity stimuli is shown in Fig. 3. However, it was usually found that neuronal inhibition from other modalities is not wiped out even by strong background excitation during such type 2 vestibular activation. To G. Moruzzi You are quite right in saying that pupillary dilatation may accompany vestibular stimulation, but we could not observe it in most of our experiments, which were donewith closed eyes. Even under these scotopic conditions an interaction with a pupil effect might still be possible, although it does not seem to be a very probable explanation of the facilitation and occlusion phenomena of multisensory interaction. Systematic investigations with the fixed pupil technique, proposed by Moruzzi, are advisable. However, as most of our experiments investigated the effect of visual and vestibular stimuli separately at different times on the same neurons, this kind of neuronal convergence is certainly independent of pupil changes as well as the prevailing occlusion effects of simultaneous stimuli. To P . Buser Our results agree well with those of you and your co-workers, although it remains difficult to correlate single neurone responses with the results of the evoked potential technique with and without chloralose. Your term “non-primary” seems to be appropriate for the responses in extraspecific areas. We have not used it, because we included the primary receiving areas in our studies. Neuronal responses to illumination in the medial suprasylvian gyrus are very similar to those in area 17. Their latencies cannot yet be cornpared, because in our experiments there is a longer retinal delay due to scotopic vision when the eyes are closed. I t has not yet been decided whether these non-primary responses are secondary effects mediated by the geniculate system or by other projections of the visual system. It seems quite possible that two distinct systems are represented, both in your responses of short and long latencies and also in our type 1 and type 2 responses. We have recorded from only a few neurones in the motor cortex after trisensory stimuli. Therefore we have relied for this region on the results of Buser and lmbert (cited in our paper) for visual and acoustic stimuli. Our original scheme did not include the motor-cortex proper and we expect to find that there are also the specific vestibular responses which are indicated by the anatomical connections. To W . Grey Walter and H . Gastaut The lambda waves associated with eye movements cannot yet be correlated with neuronal effects, but it is quite possible that they are an expression of the central regulation of movement perception and contrast stimulation. If I may be allowed to speculate, I would correlate the lambda waves with certain inhibitory phenomena. Afferent impulses during eye movements might not give the full story of these curious waves, because such impulses certainly provide little information during the quick phase of nystagmus. Recording from some cortical neurones which responded to nystagmus, we have found inhibition during the quick phase. Lambda waves are in complex coordination with eye movements and I wonder whether they may be explained by the Von Holst model (1951). If optokinetic nystagmus elicited by a drum does not provoke lambda waves, the correlation with attention would seem to be still more complex. Certainly optokinetic nystagmus is connected closely with attention. Although there is, as was shown by Zumpfe (1960/61), no complete blanking of all
DISCUSSION
239
visual impulses during the quick phase of nystagmus, there must be some inhibition of movement perception during this phase. Gastaut’s important observation of epileptic nystagmus in man agrees well with experiments on animals in which the occipital area was electrically stimulated. Such observations do not disprove the conception of inhibition, which may be concealed in cortical waves or even in seizure waves. VON HOLST, E. Zentralnervensystem und Peripherie in ihrem gegenseitigen Verhaltnis. Klin. Wschr., 1951,291 97-105. ZUMPFE,G. Der optokinetische Nystagmus bei Gesunden und die Wahrnehmung unbewegter Lichtreize wahrend des optokinetischen Nystagmus. Pjlug. Arch. ges. Physiol., 1960/61, 272: 78.
To H . H. Jasper Our analysis of the interaction of sensory convergence is just beginning. So far as we can a t present say, both facilitation and inhibition are observed in intermodal convergence as well as in the more common occlusion. Inhibitory actions are very marked in the type 1 vestibular responses and in the reciprocal visual responses of the B- and D-neuronal systems. Pure excitatory actions are represented in type 2 vestibular responses with delayed activation, but excitation seems to prevail also in other responses. “Unspecific” ascending afferents seem, after reticular or thalamic stimulation to be mainly excitatory, but they do not seem to be exclusively so, as they are in the delayed activation of type 2 vestibular responses. These are the only afferent mechanisms found by us in the cortex which show pure excitation without inhibition. The explanation may be that only the final results of interaction between activation and inhibition at lower levels are carried up to the cortex as excitation. Otherwise I think that inhibitory actions at cortical neurones are the most important mechanisms of coordination in the cortex. Constant “pure” inhibition was found only in neurones of the motor cortex after caudate stimulation. In the visual cortex, which is the only other area that we have investigated, inhibition after caudate stimuli is only apparent in half of the neurones. The importance of inhibition is shown clearly by neuronal recordings. Also single unit recordings do not reveal all the inhibitory events. Inhibition may occur in presynaptic pathways or may remain hidden in the background of activity. Inhibition cannot be shown by evoked potentials, in which all these complex interactions, including inhibition, are concealed. It is clearly seen only in neuronal recordings. These defects of the evoked response method cannot be compensated, even by the best averaging computer techniques. However, our results are derived only from a few samples of the millions of cortical neurones. They are therefore still very far from giving a complete picture of neuronal integration of excitation and inhibition in the cerebral cortex. To D . Albe-Fessard I am glad to hear that you have found such a selective suppression of extralemniscal evoked potentials by caudate stimulation. This would confirm our impression that the effects of caudate stimulation are closely connected with the thalamo-reticular system and that they may even use the same pathways as those used by the non-specific ascending actions. It is, however, difficult, from recordings of evoked potentials alone, to draw conclusions about the inhibitory mechanism of this suppression. We suspect that caudate stimulation may involve a presynaptic inhibition at the cortical level which is similar to the presynaptic mechanisms revealed by Eccles (see p. 8) in the spinal cord, but as yet we have no proof of this assumption.
To A . Arduini The significance of the “off”-discharges in the visual D-system might well be interpreted as a reciprocal inhibitory action which cancels “on”-excitation in the antagonistic B-system. This cancellation clears the blackboard for further information. However, this clearing action is not the only rather negative aspect of the D-system. Both antagonistic systems, the B-(brightness or “on”)-neurones and the D-(darkness or “off”)-neurones, should always be considered together. Darkness information may also be a relevant positive signal and in contrast vision it seems to be of equal importance to brightness information. In lower animals, the shadow reactions of the D-system may even be more important biologically as a means of signalling danger. This biological significance, in addition to the clearing action and the mechanisms of contrast vision, may explain why such elements with “off ”responses have developed so extensively and universally in the visual system, although they are lack-
240
R. JUNG et
al.
ing in other sense modalities. As yet we have not found any special significance of this “oK”-system for multisensory convergence. I To F. Bremer Although earlier studies from our laboratory had shown the frequent convergence of unspecific and specific afferents at neurones of the visual cortex, we were surprised by the abundance of multisensory convergence on cortical neurones in many other areas. 1 quite agree with Prof. Bremer that one should not confuse all the widespread vestibular effects with unspecific reticular activation. I am very interested in Desmedt’s new results on auditory and visual convergence in the anterior sylvian area, which seem to fit in well with Kornhuber’s and Fonseca’s neuronal findings. We are still far from understanding the physiological basis for the rich integration of sensory information by cortical neurones. Thus our experiments are a rather preliminary trial to inaugurate this wide field of research a t the neuronal level of the cerebral cortex.
The Direct Cortical Response Associated Events in Pyramid and Muscle During Development of Movement and After-Discharge* SALVATORE MINGRINO**, WILLIAM S. COXE, RICHARD KATZ*** AND SIDNEY GOLDRING Presented by JAMES L. O’LEARY Divisions of Neurology and Neurosurgery and Beaumont-May Institute of Neurology, Washington University School of Medicine, Saint Louis, Mo. (U.S.A.)
In Adrian’s (1936) description of the direct cortical response (DCR) a 20 msec surfacenegative potential was observed to follow each shock ofa repetitive train when stimulus intensity was weak. Early in such a stimulus course shocks of strong intensity also evoked negative potentials, but as stimulation continued they disappeared to be replaced by positive ones of somewhat longer duration. Muscle activity and afterdischarge did not occur during stimulation unless the positive polarity potentials appeared. Adrian concluded that the change to positive polarity signified the activity of pyramidal cells situated in the cortical depth, and that its appearance was a necessary prelude to the development of either muscle contraction or cortical afterdischarge. Since then, other DCR components have been identified (Chang 1951; Brooks and Enger 1959; Caspers 1959; Goldring et al. 1961). However, full appraisal of the DCR has required single and repetitive stimuli of graded intensity and d.c. as well as r.c. recording. Lately we have re-investigated the changes in these DCR components as they relate to initiation of muscle contraction and of cortical after-discharge. Simultaneous examination of pyramid activity helped significantly in interpreting the changes observed in the cortical record. Our observations differ in several respects from those of Adrian. This is perhaps attributable to interim perfection of recording and amplifying equipment and to differences in anesthesia. METHOD
Twenty-five monkeys (14 squirrel and 11 macaque) and an equal number of cats were used. Monkeys were prepared and studied under nembutal anesthesia, using 25-30 mg/kg injected i.p. as an initial dose. The amount was supplemented periodically with
* Aided by grants from US. Public Health Service (B-1517) and the Allen P. and Josephine B. Green Foundation.
** Now at Istituto di Neurochirurgia, Universita di Padova, Italy. *** Now at Upstate Medical Center, Department of Surgery, Syracuse, New York. References P. 256-257
242
J . L. O'LEARY et al.
smaller i.p. injections as the experiment proceeded. Cats were prepared under ether anesthesia. Thereafter all wound surfaces were heavily infiltrated with 1 per cent procaine and the animals were immobilized with Flaxedil (gallaminktriethyliodide) and carried on artificial respiration. Medullary pyramid and muscle activity were studied in conjunction with recording of the DCR in the monkey. In the cat, the suprasylvian cortex was probed with a recording electrode during after-discharge. Similar observations were made upon the precentral motor cortex of the monkey. The medullary pyramids of the monkey were exposed through a retropharyngeal approach. First the trachea and esophagus were transected and a cannula inserted into the distal tracheal segment. The proximal portions of trachea and esophagus were then retracted rostrally, the muscles stripped from the basi-occiput, and a small opening made in the bone. Upon incising the dura, the pyramids were exposed at the medullary level just rostra1 to their decussation. Activity of the pyramids was recorded monopolarly. A stainless steel needle with a 20-50 r c exposure was inserted into the pyramid and led against a larger needle (tip 300 p) placed upon adjoining dura. In the earlier experiments the position of the tip in the medullary pyramid was verified histologically. Later this proved unnecessary. Muscle responses were recorded between a needle which penetrated either the first dorsal interosseus or thumb flexor muscles and another which was inserted into the periosteum overlying the styloid process. The cortical stimulating electrode was a tripolar one consisting of 200 p steel wires insulated except for their tips which were placed 1 mm apart. A Grass stimulator was connected to the stimulating electrode via a circuit for balancing shock artifact and a step-up (1 : 4) isolation transformer. Stimuli were 5-15 V square waves of 0.01-0.2 msec pulse duration. Recording electrodes were calomel half cells. When recording surface cortical responses, one calomel electrode was connected to a saline soaked cotton wick (0.5 mm diameter) which rested on the cortical surface in the center of the tripolar stimulating arrangement. The other tip was led to frontal periosteum via a larger saline soaked cotton wick. For intracortical responses transcortical recording was used, and the calomel electrodes then made contact with the recording points via 0.9 per cent NaCl filled capillary pipettes. The probe had a tip diameter of 50-75 ,u ; the reference, which was inserted into subcortical white matter, had a tip diameter of 200 i'c. A dual beam oscilloscope was used. DCR was recorded on one beam after d.c. amplification. The other beam, activated by an r.c. amplifier, recorded simultaneously either medullary pyramid or muscle potentials. Records were photographed with a Grass camera. RESULTS
1. Relationship of DCR components to the medullary pyramid response
Fig. 1 illustrates DCR's of precentral motor cortex of squirrel monkey evoked by single stimuli of increasing strength. Similar responses occur in macaque and cat. The left column shows responses recorded on a time line sufficiently fast to resolve the
ASPECTS OF DIRECT CORTICAL RESPONSE
243
3 A
-
v 1
R hy thrnic waves
-
10 rn sec
1rnvI
1mvI
-
100 rn sec
Fig. 1 DCR’s to stimuli of increasing intensity in squirrel monkey. Records from precentral motor cortex. I : primary negative potential; 3: after-positivity ; 4: slow negativity. Second negative wave labelled 2 in previous communications (Goldring et al. 1961 ; Weinstein et al. 1961), is not evident in these responses. A, response to weakest stimulus; B through D, responses to stimuli of progressively stronger intensity. Left time marker is for responses in left column, the other for responses in right column. Larger voltage calibration signal is for responses in A through C, smaller one for responses in D. In this and all subsequent figures straight lines like those in right column are baseline from which slow potential changes are read. Upward directed potentials are positive.
detail of the initial part of the response. In the right column, responses were recorded on a slower time line to demonstrate the longer duration response components. Upon the slow time line, a very weak stimulus ( A ) evokes a 20 msec negative deflection, called priniary potential, which is followed by a lower amplitude and longer-lasting positive one called after-positivity. As the stimulus strength is increased ( B and C ) , the primary potential grows in amplitude while the after-positivity at first decreases and is then replaced by a slow, long-lasting, negative potential (circa 250 msec) called slow negativity. At these stimulus strengths a sequence of rhythmic waves may follow the slow negativity components. At intensities well above slow negativity threshold fast spikes of positive poiarity and brief duration may appear on the rising phase of the primary (negative) potential. With strong stimuli (D)one or more such spikes may even precede the start of the primary potential. Such spikes were observed in the DCR of the rabbit by Suzuki and Taira (1958), appearing in those responses recorded from postcentral and striate cortex but being absent in the one from retrosplenial cortex. We have observed them in References p . 156-257
244
et al.
J. L. Q’LEARY
DCR’s of primary sensory and motor projection areas of cat, monkey and man. They have not been detected in corresponding responses derived from association areas. I n our experience they are best seen in the squirrel monkey (Coxe et al., in preparation). During simultaneous recording from cortex and medullary pyramid, neither the cortical primary potential nor the still slower components which follow it have been observed to relate to pyramidal activity ;and the thresholds for the medullary pyramid responses (Patton and Amassian 1954) and for these cortical components also differ, being significantly higher for the former. By contrast, the more intense stimuli required t o evoke the fast cortical spikes of the DCR, produce medullary pyramid responses. A response was never recorded from the medullary pyramid unless the fast spikes also appeared in the cortical record, and the number of I waves (Patton and Amassian 1954) of the pyramidal response corresponded to the number of positive spikes in DCR. In Fig. 2, A , B and C observe that at stimulus strengths below threshold for the fast positive spikes of the cortical record no activity was recorded from the medullary pyramid. When the stimulus intensity was at threshold for the fast A
B
’
-r
-w !
-v G
Di7 - I
500pV
I
I
250 pV 500 pV 1 0 m s e c
250pV
I
Fig. 2 Siniultaneously recorded direct cortical and pyramidal responses in squirrel monkey. Stimulus site is hand area of precentral motor cortex. Upper trace of each pair ( A through H) is pyramidal record. In A stimulus intensity is threshold for DCR. Intensity is increased progressively in A through H, except in E which is same as D . Stimulus strength in latter was near threshold for DCR spikes. Note the absence of pyramidal response when stimulus fails to evoke DCR spikes ( E ) but appearance of pyramidal response when stimulus is effective (D).Smaller D in H i s direct response; I : indirect ones (after Patton and Amassian 1954). Short oblique lines over DCR in H identify the positive spikes. Voltage calibrations on left are for cortical responses, those on right for pyramidal ones; the smaller one of each pair is for responses in H, the larger ones are for all others. Positive is up.
ASPECTS OF DIRECT CORTICAL RESPONSE
245
cortical spikes, the record from the medullary pyramid showed a small response (D). Also, at threshold the fast cortical spikes were inconstant in the responses to successive stimuli, and when they failed to appear no response was detected in the medullary pyramid either (E). At stronger stimulus intensities the fast cortical spikes became more conspicuous; and the responses recorded from the medullary pyramids grew commensurately, the number of spikes in the pyramidal volley coming to correspond to the number of DCR spikes (F, G, H ) . The fast cortical spikes are reminiscent of similar potentials seen in conventional evoked responses of primary cortical projection areas (Clare and Bishop 1956; Bremer 1958). There the spikes have been interpreted as signalling all-or-none discharge of cell bodies. If one interprets the fast spikes of DCR similarly, the response of the medullary pyramid may be considered to be the conducted axon potentials of the cortical neurones. The start of each such cortical spike precedes the onset of its corresponding pyramidal I wave by 1-3 msec (Fig. 2 H ) . Such latencies are compatible with known conduction times along corticospinal axons to the medullary pyramid (Patton and Amassian 1960). The close relationship between DCR spikes and pyramidal responses is also evident during repetitive stimulation at 15/sec. At a stimulus just above threshold both for suchaspikeand for the medullary pyramid response, both grow together in amplitude as stimulation continues (Fig. 3, A and B). Concomitantly, the primary potential of DCR
-
20 rn sec
Fig. 3 Simultaneously recorded direct cortical and pyramidal responses in squirrel monkey. Stimulus frequency 15/sec. Stimulus site is hand area of precentral motor cortex. Upper traces in A , B and C are cortical records, lower are pyramidal ones. 2.5 sec and 0.25 sec of record are omitted between A and B, and B and C , respectively. Difference in duration between first and second responses in A as compared to interval separating all others is due to momentary failure of stimulator. In this particular case first evoked pyramidal response is larger than second. Usually this was not the case, first and second responses showing no amplitude differences. 250 ,uV calibration is for pyramidal trace; 1 mV signal for cortical one. Positive is up for cortical record, down for pyramidal one. References
I). 256-257
246
J . L. O'LEARY
et at.
decreases progressively in amplitudeandvirtually disappears. This also supports a close relationship between the fast spikes and the response of the medullary pyramid(Fig. 3 C). A similar linkage is observed when DCR and medullary pyramid responses are recorded simultaneously during cortical spreading depression (SD). At the height of the negative cortical d.c. shift which accompanies spreading depression (Fig. 4), single stimuli delivered periodically continue to evoke responses in the medullary pyramid and these show little difference from those elicited prior to SD. In the DCR recorded simultaneously fast spikes also remain evident during SD, although the primary (negative) potential of DCR disappears entirely. C
B
A
D
Fig. 4 Simultaneously evoked direct cortical and pyramidal responses during spreading cortical depression in squirrel monkey. Uppermost record in each column is pyramidal response; lower one, DCR. A : control; B and C : responses recorded during SD identified by the negative slow voltage change (the latter is not shown here). Note that the pyramidal response is not significantly changed. However, primary negative potential disappears from the DCR, only positive spikes remaining. D : responses after end of SD; primary negative potential re-appears and records are like those prior to SD. 250 ,ILVcalibration is for pyramid; 1 mV signal for DCR. Positive is up.
2. The DCR evoked by a stimulus suflciently strong to produce a contraction of muscle A single cortical stimulus is capable of evoking a muscle twitch but the stimulus intensity required is considerably above that necessary to elicit a response in the medullary pyramid (Fig. 5). With repetitive shocks to the cortex a significantly weaker A
7+ C
B
: + !
c--(
250pVI
2mV[
10 m sec
Fig. 5 A : simultaneously evoked direct cortical and pyramidal responses in macaque. Stimulus site is in
hand area of precentral motor cortex. Upper response is from pyramid, lower one from cortex. B and C : same as A , but upper records are from first dorsal interosseous muscle. In B, stimulus intensity which is significantly stronger than one used to evoke pyramidal response in A , fails to elicit muscle response. C : appearance of muscle response when stimulus intensity is increased to 1.5 times that used in B. 250 p V calibration is for pyramidal and muscle responses. 2 mV signal i s for the DCR. Positive is up.
241
ASPECTS OF DIRECT CORTICAL RESPONSE
cortical stimulus is required. For Fig. 6, the stimulus intensity was such that a muscle twitch did not follow the initial shock of a series. The subsequent few stimuli also failed to excite the muscle, but as stimulation continued a muscle twitch did follow each shock. It is important that in the DCR's recorded simultaneously with the muscle response it is a negative and not a positive polarity potential (Fig. 5 C and Fig. 6) that follows the shock (see Adrian 1936).
1rnV
L,
100 rn sec
100 rn sec
Fig. 6 Simultaneously recorded direct cortical and muscle (first interosseous) responses in macaque. Stimulus frequency 15/sec. Stimulus site is hand area in precentral motor cortex. Upper trace, muscle; lower one, cortex. Cortical record with d.c. amplifiers; muscle taken with r.c. amplifier, also serves as baseline for cortical trace. Right calibration, muscle; left, cortex. Positive is up.
3. D C R , muscle andpyramidal activity incident to repetitive cortical stimulation leading to cortical after-discharge ( a ) D C R and muscle activity. During strong repetitive stimulation of sufficient intensity to evoke the fast spikes of DCR, the negative primary potentials show a progressive amplitude decrease and steady potential shifts negatively, the latter resulting from addition between the slow negativities of the individual responses to the successive shocks (Goldring et al. 1958; Fig. 7 A ) . As stimulation continues,
1rnV 100 rn sec
l m V L 100 rn sec
Fig. 7 Changes in cortex and muscle incident to repetitive (15jsec) stimulation leading to cortical afterdischarge. Stimulus site is hand area in precentral motor cortex of macaque. Muscle records from first dorsal interosseous. Upper traces in A through P:muscle; lower ones: cortex. Cortical record taken with d.c. amplifier; muscle taken with r.c. amplifier also serves as baseline for cortical record. Fine upwardly directed lines at end of cortical trace in A, and in B, C and beginning of D consist of shock artifact and positive spikes. This detail cannot be resolved on time line used to derive this record (see text and Fig. 8). Left calibration, cortex; right, muscle. Positive is up. References P. 256-2.77
J. L. O’LEARY et al.
248
primary negative potentials disappear and a positive potential of longer duration comes to follow each shock; meanwhile, steady potential remains shifted negatively (Fig. 7 B and C ) . The positive potentials which follow each of a succession of shocks grow in amplitude, and if the stimulus is stopped as the amplitude builds up, afterdischarge may occur (Fig. 7 D).The cortical waves which comprise after-discharge bursts are also positively directed components and in form resemble the evoked positive potentials seen near the end of a stimulus period. Steady potential (SP) shifts negatively during stimulation but begins to recover towards the prestimulatory baseline as soon as stimulation is over. This recovery, which is a gradual one, is interrupted by periodic bursts of after-discharge, each riding upon a short-lived positively directed SP shift (Fig. 7, D,E and F). Restitution of steady potential to the prestimulatory baseline usually coincides with the end of after-discharge. These features of after-discharge and the accompanying SP shift are typical of the paroxysmal state as produced under our conditions of experiment (see also Fig. 9 and 10). As noted above and shown in Fig. 7 A , muscle activity can be elicited long before positive potentials replace the primary negative ones in the DCR. In the poststimulatory period muscle responses occur only during each after-discharge burst (Fig. 7, D and E). However, cortical after-discharge does not invariably occasion muscle activity (Fig. 7 F). (6) Nature of positive potentials that appear during strong repetitive stimulation. These represent a fusion of two separate processes as becomes evident if the responses are photographed on paper moving at a fast speed. Fig. 8 A shows responses to a
A
1rnV L 100 rn 5ec
1 mV-
20 m sec
Fig. 8 A : repetitively (15lsec) evoked DCR’s to strong stimuli. B, C and D : a similarly evoked series of responses, but photographed on paper moving at a faster speed. 1 sec of record is omitted between B and C , 2 sec between C and D. Solid arrow: positive spikes. Broken arrow: positive wave. Left calibration is for A ; right one for B. C and D. Positive is up.
ASPECTS OF DIRECT CORTICAL RESPONSE
249
repetitive (15/sec) train of stimuli recorded on paper moving at 10 cmjsec. That was the speed used to register the records shown in Fig. 7 where replacement of primary (negative) potentials by positive components is evident in the latter half of the stimulus period. In Fig. 8 B, C and D an identical response was photographed upon paper moving at 50 cm/sec. As stimulation continues, the fast positive spikes grow in amplitude, the primary (negative) potentials disappearing as they do so. Meanwhile, a low amplitude slow positive wave comes to follow the spikes (Fig. 8 C), and it also grows in amplitude and fuses with them (Fig. 8 0).On the slower-moving paper, resolution of this change in pattern in the course of stimulation is not possible. There, the fast spikes become lost in the shock artifact and only the fused product of spikes and trailing positivity is seen following each shock (Fig. 8 A ) .
P I 250pV
1mV
I
-
200 rn set
Fig. 9 Changes in cortex and pyramid incident to repetitive stimulation (1 5jsec) leading to cortical afterdischarge in macaque. Stimulus site is hand area of precentral motor cortex. Cortical record is upper trace in each row except C ; itwas takenwithdx. amplifier. Brokenlineindicatesrestinglevelofcortical SP. 0.2 sec, 0.5 sec, 3 sec and 0.5 sec are omitted between A and B, B and C, C and D,and D and E, respectively. E and F a r e continuous and 5 sec are omitted between F and G, and G and H, respectively. Note that the latencies of the positive potential changes in D and Eincrease by comparison with those shown in C. This is an inconstant feature in our records and not necessary for the development of after-discharge. 250 pV calibration: pyramid; 1 mV signal: cortex. Positive is up for cortical record, down for pyramidal one. Additional details discussed in text. References D. 256-257
250
J . L. O’LEARY
et al.
(c) DCR andpyramidal activity. The record of the medullary pyramid shows two significant changes during strong intensity, repetitive stimulation such as leads to the occurrence of after-discharge. First, “unitary” or tonic activity appears, coinciding with growth of the positive potentials in the cortical record (Fig. 9, D and E). Second, sequentially evoked pyramid responses show a marked increase in amplitude and duration (compare responses in Fig. 9 A with those in 9 0).Tonic firing in the pyramid continues into the immediate poststimulatory period preceding the appearance of paroxysm in the cortical record. Then (following the cessation of stimulation) the cortical record is quiet except for the continuance of the steady potential shift which had developed during stimulation (Fig. 9, E and F). The tonic barrage in the record from the medullary pyramid continues uninterrupted during the tonic phase of cortical after-discharge (Fig. 9 G). Thereafter, as the cortical paroxysm becomes clonic in character, the discharge in the medullary pyramid also tends to become discontinuous, clusters of tonic discharge coinciding there with each burst of cortical after-discharge (Fig. 9 H ; Fig. 10 and 11;see also Adrian and Moruzzi 1939). ( d ) Origin of’ after-discharge in the cortical thickness. Using a surface cortical stimulus, the after-discharge occupies the same superficial cortical layers as yield the primary (negative) potential of DCR. Records taken at successive depths as a critical recording electrode is raised from white matter to surface show that no significant after-discharge is recorded deeper than 1-1.5 mm below the cortical surface and that reversal points for after-discharge and primary potential are approximately the same. Not infrequently, however, the reversal point of after-discharge is somewhat more superficial than that of primary potential, the latter being centered approximately 0.2 mm below the cortical surface. Fig. 12 illustrates such a case. The left column shows primary (negative) potentials recorded at successively more superficial levels from white matter ( A ) to cortical surface ( F ) . In the right column, the trace labelled P opposite each primary response is the activity recorded during after-discharge at the same level. Presence of after-discharge was monitored by a surface electrode and the record thus derived is shown in the traces marked 5’. Observe that in strips A , B and C the primary potential is “positive” in deflection ;and after-discharge, which first becomes clearly evident in strips B and C, is opposite in polarity to that recorded at the surface. In D the trace to the left which registers primary potential is virtually isoelectric, indicating the reversal point for that potential. However, after-discharge is still recorded at this depth and continues to show opposite polarity as compared to the surface record. At a slightly more superficial level the primary potential reappears, but now as the conventional negatively directed deflection. At this level the trace recording the after-discharge is nearly isoelectric. Upon raising the probe to the surface a well-developed primary potential is recorded. There the after-discharge shows opposite polarity, the same as is recorded by the surface monitor. The fact that no significant after-discharge appears in the deepest cortical layers and that there is but one reversal point for the after-discharge makes it improbable that the positively directed potentials which comprise it reflect activity arising in the cortical depth. Such sinks, if they existed, would be expected to draw current from deeper cortical layers as well as from the more superficial ones; and under such
25 1
ASPECTS OF DIRECT CORTICAL RESPONSE
1 20m sec
toopv[
2mV1
lmVI
Fig. 10 Pyramidal tonic firing during bursts of cortical after-discharge. Upper traces in A, B and C: cortex; lower one: pyramid. Broken lines indicate cortical SP prior to stimulus initiating after-discharge. A and B from same animal, but B recorded on faster time line to show detail. C from another experiment. Pyramidal firing occurs with each positively directed shift of cortical SP; upon each shift is written a series of positively directed oscillations. 100 pV calibration: pyramid; 2 mV: A and B; 1 mV: C, cortex. Positive is up for cortex, down for pyramid.
B-
E T ImV
I
20m sec Fig. 11 Pyramidal “unitary” firing during cortical after-discharge. Upper traces in A, Band C: cortex; lower one: pyramid. A, B and C show consecutive slow cortical waves and associated pyramidal activity occurring at end of cortical after-discharge. Each positively directed wave is associated with appearance of a single spike in pyramidal record. References D . 256-257
252
J . L. O’LEARY
et al.
S
1mV
1
-
100 msec
2 m ~ 1Z
~
IV
Fig. 12 Comparison of reversal points of DCR and cortical after-discharge as a recording probe is raised from white matter to cortical surface. Squirrel monkey. Left column: DCR; tight: after-discharge. A : white matter. F : cortical surface. B through E are successively superficial levels. P:record derived by recording probe. S, after-discharge monitored by surface recording electrode. Broken line, resting level of cortical SP for surface electrode. Positive is up. I mV calibration: DCR; large 2 mV calibration for A ; small 2 mV calibration for B through F. Refer to text for details.
circumstances polarity of after-discharge would be required to reverse twice instead of once as a probe electrode is withdrawn from white matter to surface. 4. Pyramidal activiry during spreading cortical depression ( S D )
During the course of this study SD was produced not infrequently in the squirrel monkey. In our experience, it can be evoked in this animal almost as easily as in the rabbit. lncident to the negative slow voltage change which accompanies SD (Lead 1947), a tonic discharge develops in the record of activity of the medullary pyramid, becoming most intense at the height of the negative shift. In some animals the disc’iarge was observed to continue at a lesser intensity as SP recovers and shifts positively, all tonic activity eventually disappearing when SP resumed its resting state (Fig. 13).
253
ASPECTS OF DIRECT CORTICAL RESPONSE
I I
A
c ---
ID II I
I
I
I
---
I I
I
I
u
20 m sec
Fig. 13 Pyramidal activity during spreading cortical depression (SD). Upper trace in A and B and straight lines in C , D and F are cortical record. Cortical record is not shown in E because it had shifted negatively off of the oscilloscope face. A and beginning of B show DCR’s to repetitive stimulus which evoked the SD. Broken line in B through F is cortical SP prior to onset of SD. 5 sec, 6 sec, 6.5 sec, and 8 sec of record are omitted between A and B, B and C, C and D,and D and E, respectively. F : several min after E. Note appearance of pyramidal firing in C as cortical SP shifts negatively and increase in firing rate as negative shift increases in D and E (cortical record shifted negatively off tube face in E ) . F: return of cortical SP to resting level and absence of firing in pyramidal record. 100 pV calibration: pyramid; 2 mV signal: cortex.
DlSCUSSlON
Our findings indicate that for movement to result from strong repetitive surface cortical stimulation primary (negative) potentials of DCR need not be replaced by positive ones; nor does the positive potential appear to simply reflect activity of pyramidal cells in the deeper layers of cortex (Adrian 1936). The positive potential is actually a fusion of two processes: 1. the fast positive spikes which initiate the DCR as evoked by strong stimulation, and 2. a slower positive wave. The reversal point of the fast positive spikes is in the deeper cortical layers (unpublished data) and their configuration is identical to the positive spikes which characterize evoked responses of primary projection areas. In evoked responses of visual or sensorimotor cortex, the spikes which initiate the surface-positive potential
254
J . L. O’LEARY
et al.
are attributed to the all-or-none discharges of cell bodies (Clare and Bishop 1956; Bremer 1958). Our findings support the view that the positive spikes we describe arise in the cortical neurones as all-or-none impulses which conduct centrifugally to be recorded as pyramidal activity (Fig. 2, 3 and 4). We view the slower positive wave as a potentially paroxysmal process. I t has the same superficial reversal point as does the after-discharge (unpublished data) and in appearance strongly resembles the individual positive potentials which make up the bursts of after-discharge. In addition, after-discharge occurs sometimes when positive potentials do not appear during the end of the stimulus period, primary negative potentials of the DCR’s evoked by successive shocks transforming instead into notched negative or diphasic potentials. Under such circumstances the paroxysmal potentials of the after-discharge bursts mimic whatever process was being evoked before cessation of the stimulus. Thus, during stimulation leading to after-discharge, the unique form of the altered trace is not as important as the establishment of a potential capable of replication even in the absence of a stimulus. Finally we believe it is important that neither the positive potential which appears during stimulation as a substitute for the usual primary (negative) one nor the spontaneously appearing burst potential of the after-discharge, signify a “positive excursion” of the neuronal membrane with respect to the resting steady potential of the prestimulatory record. Rather, such oscillations are seen to arise from a baseline which had already shifted negatively during stimulation, and each such positive excursion identified in the d.c. record signifies a temporary membrane surge towards the prestimulatory state. When the prestimulatory state is restored after-discharge ceases. When strong intensity repetitive stimulation occasions cortical after-discharge tonic firing develops in the pyramid near the end of the stimulus period. At this time positive potentials are observed to follow each of the succession of shocks and SP is shifted negatively. Such pyramidal discharge continues into the poststimulatory period that precedes the first after-discharge burst. In that short interval the negatively shifted SP is the only sign that a change has occurred in cortical excitability. Thereafter, during after-discharge, clusters of pyramidal firing may coincide with each cortical paroxysmal burst. Such bursts also coincide with the somewhat gradual return of SP to its prestimulatory value; and each burst appearing before that value is reached consists of several rhythmic oscillations of positive polarity surmounting a transient positive SP surge. Thus, several SP events take place in cortex at the time that tonic pyramidal firing appears and later becomes clonic. We view these d.c. changes as related manifestations of the effect repetitive stimulation has in depolarizing cortical neurones, and rendering their membranes unstable for the time. It has been shown, for example, that membrane potential will oscillate during excessive depolarization (Li et al. 1961); and Kandel and Spencer (1961) have noted (intracellular recording) that changes in membrane depolarization lasting for as long as seconds occur in the pyramidal neurones of hippocampus following stimulation suficiently intense t o lead to paroxysm. During periods of membrane instability momentary spontaneous oscillations occur in which membrane potential may reach the critical
ASPECTS OF,DIRECT CORTICAL RESPONSE
255
firing level of the neurone, accounting for periodic accessions in pyramidal firing which develop during cortical paroxysmal bursts. The fact that the after-discharge occupies the same superficial cortical layers (upper 0.2 mm) as the primary (negative) potential of DCR (Fig. 12) suggests that the changes in membrane potential take place in the superficial dendritic plexus and that the membrane changes which occur there influence cell body or axon hillock by electrotonic propagation. Thus, what takes place in the superficial dendritic plexus remote from the perikarya could contribute to the firing rate of the neurone. Such a situation is perhaps analogous to the changes proved to occur in the stretch receptors of crustacea (Kuffler 1960), a far simpler synaptic situation. Excessive depolarization could inactivate an axon spike generator altogether and one can visualize membrane fluctuations as effecting accentuation and repression of firing. It is perhaps the rate of change in the membrane toward depolarization which causes it to abruptly exceed the critical firing level. Such factors as were elucidated by Kandel and Spencer (1961) are perhaps applicable to the analysis of certain features of our records. For example, maximal pyramidal discharge occurs when the negative SP shift is returning to the baseline and not when it is maximal (compare Fig. 9 F with 0). Again, the clusters in the medullary pyramid which coincide with the clonic phase of the cortical after-discharge must signify marked change in membrane excitability of the neurones, for the cortical after-discharge is made up of rhythmic waves surmounting a transient positive shift. This, in turn, occurs upon a baseline fast recovering from the preceding negative SP shift. The limitations of extrapolating from intracellular microelectrode studies (Kuffler 1960; Li et al. 1961; Kandel and Spencer 1961) to interpret our records derived extracellularly with gross electrodes are obvious. Nevertheless, at present we believe that our findings permit an interpretation based on current theory of excitation in single neurones. The occurrence of pyramidal discharge during the negative slow voltage change which accompanies Leiio’s spreading depression indirectly confirms Grafstein (1956) who observed cortical “unitary” firing during SD. We believe that the pyramidal firing results from massive depolarization of cortical neurones. Thus it is not dissimilar to the firing of a cell which has undergone injury or cathodal polarization. Tonic activity similar to that described herein was observed by Burei (1959) to arise in the midbrain reticular substance during spreading depression. Evidently, therefore, the neuronal discharge which takes place during the SD cortical negative d.c. shift is massive, non-specific and invades many paths besides the corticospinal one. SUMMARY
This study reports correlations between direct cortical response and that of medullary pyramid (or muscle), recorded simultaneously in macaque, squirrel monkey or cat. I . The medullary pyramidal response relates to the several fast positive spikes that initiate the direct cortical response of sensorimotor cortex to a strong stimulus and not to the primary negative component of that response. The latter, together with K c f e r e n m 9. 2 56-2 57
256
J . L. O’LEARY
et al.
the subsequent after-positivity and slow negativity components, do not relate to the medullary pyramidal potential. 2. The replacement of the primary negative component of direct cortical response by a positive one during repetitive stimulation is not essential to the development of activity in the related musculature, the latter becoming evident when cortical DCR still shows its primary negative component. 3. The positive replacement potential which occurs in DCR during strong repetitive stimulation results from fusion of two separate processes, the fast spikes and a trailing positive wave. The latter grows in amplitude during continuing stimulation. The fast spikes have a deep cortical origin ; however, the subsequent positive potential arises superficially as is demonstrated by probe experiments. 4. In repetitive cortical stimulation leading to after-discharge cortical SP shifts negatively and positive transients replace the primary negative ones. The negative d.c. shift continues into the immediately poststimulatory period. Preceding the onset of after-discharge the negative d.c. shift is the only remaining sign of a temporary change in the state of cortical activity. After-discharge bursts appear as the d.c. trace returns to its prestimulatory level. Replacement of repetitively evoked primary negative potentials of the direct cortical response by positive ones is usually a necessary prelude to the appearance of after-discharge. 5. When repetitive cortical stimulation leads to after-discharge, tonic activity appears in the record from the medullary pyramid as the cortical d.c. shift develops and positive potentials come to replace the negative ones there. The tonic discharge in medullary pyramid continues into the immediate poststimulatory period when the only change in the cortical record is the negatively shifted SP. It persists during the after-discharge, being continuous if the after-discharge is tonic. It is discontinuous and accompanies each paroxysmal burst if paroxysm is clonic. 6. Such cortical after-discharge arises in the same superficial cortical layers (upper 0.2 mm) as does the primary negative component of the direct cortical response. 7. In spreading depression tonic activity becomes evident in the medullary pyramid during the characteristic negative cortical d.c. shift. 8. Results are interpreted in the light of present day theory concerning membrane changes which lead to all-or-none firing.
REFERENCES ADRIAN,E. D. The spread of activity in the cerebral cortex. J . Physiol. (Lond.), 1936, 88: 127-161. ADRIAN,E. D. and MORUZZI, G . Impulses in the pyramidal tract. J. Physiol. (Land.), 1939, 97: 153-199. BREMER, F . Cerebral and cerebellar potentials. Physiol. Rev., 1958, 38: 357-380. BROOKS, V. B. and ENGER,P. S. Spread of directly evoked responses in rat’s cerebral cortex. J . gen. Physiol., 1959, 42: 761-777. BURES,J. Reversible decortication and behavior. In M. A. B. BRAZIER (Editor), The central nervous system and behavior. Second conference. Josiah Macy Jr. Foundation, Madison Printing Co., Madison, N. J., 1959: 207-248. CASPERS, H. Uber die Beziehungen zwischen Dendritenpotential und Gleichspannung an der Hirnrinde. Pjliig. Arch. ges. Physiol., 1959, 269: 157-181.
ASPECTS OF DIRECT CORTICAL RESPONSE
257
CHANG,H. T. Dendritic potential of the cortical neurone produced by direct electrical stimulation of the cerebral cortex. J. Neurophysiol., 1951, 14: 1-23. CLARE,M. H. and BISHOP, G. H. Potential wave mechanisms in cat cortex. Electroenceph. clin. Neurophysiol., 1956, 8 : 583-602. COXE,W. S., MINGRINO, S., GOLDRING, S. and O’LEARY, J. L. Investigation of the direct response in cerebral cortex of man. Second International Congress of Neurological Surgery, Washington. D C. ( U . S . A ) , 1961. Excerpta Medica,Amsterdam, No. 36: E 65 GOLDRING, S., O’LEARY,J. L. and HUANG,S. H. Experimental modification of dendritic and recruiting processes and their D C after-effects. Electroenceph. din. Neurophysiol., 1958, 10 : 663-676. GOLDRING, S., JERVA,M. J., HOLMES, T. G., O’LEARY,J. L. and SHIELDS, J. R. Direct response of human cerebral cortex. A . M . A . Arch. Neurol., 1961, 4 : 590-598. GRAFSTEIN, B. Mechanism of spreading cortical depression. J . Neurophysiol., 1956, I 9 : 157-172. KANDEL,E. R. and SPENCER, W. A. Electrophysiology of hippocampal neurons. 11. After potentials and repetitive firing. J. Neurophysiol., 1961, 24: 243-260. KANDEL,E. R. and SPENCER,W. A. Excitation and inhibition of single pyramidal cells during hippocampal seizure. Exp. Neurol., 1961, 4 : 162-179 KUFFLER, S. Excitation and inhibition in single nerve cells. Harvey Lect., 1958-1959: 176-217. LEXO,A. A. P. Further observations on the spreading depression of activity in the cerebral cortex. J. Neurophysiol., 1947, 10: 409-415. LI, CHOH-Lu,CHOW,S. N. and HOWARD,S. Y. Basic mechanisms of single cell discharge in the cerebral cortex. Epilepsia (Amst.), 1961, 2 : 13-22. PATTON,H. D. and AMASSIAN, V. E. Single and multiple unit analysis of cortical stage of pyramidal tract activation. J . Neurophysiol., 1954, 17: 345-364. PATTON, H. D. and AMAsstAN, V. E. The pyramidal tract: its excitationand functions. In J. FIELD, H. W. MAGOUNand V. E. HALL(Editors), Handbook ofphisiology, Vol. 1,Sec. I . Williams and Wilkins, Baltimore, Md., 1960: 837-863. SUZUKI,H. and TAIRA,N. Regional differences of direct cortical response. Jap. J. Physiol., 1958,8: 365-377. WEINSTEIN, W., KENDIG,J. H., GOLDRING, S., O’LEARY,J. L. and LOURIE, H.Hypothermia andelectrical activity of cerebral cortex. A.M.A. Arch. Neurol., 1961, 4 : 441-448.
Brief survey of direct current potentials of the cortex J . L. O’LEARY
The bioelectric potentials described here are those passed only by d.c. amplifiers, although we include with them such longer latency and slower evoked response components as can be shown to summate temporally during courses of repetitive stimulation. Electrode stability is an important factor in recording such voltage changes. We use calomel half-cells for the purpose. There is a mistaken impression that those are difficult to make and to maintain in a stable condition. In fact they can be constructed to suit a variety of purposes. Grossman and Gumnit (1960) described an adaptation suitable for use as an implanted electrode, and Rowland (1961) offers a corresponding one built around a coil of chlorided silver wire. The gamut of such potentials might include (1) slower components of evoked responses (i.e.,after-effects) and their temporal summations during repetitive stimulation; (2) the enduring slow potential change which accompanies Leio’s spreading depression (Leso 1951); (3) corresponding long-term potentials which develop during seizure discharge or at its cessation; (4) slow fluctuations in electrical activity of cortex which appear as an experimental preparation becomes unstable due to administration of pharmacological agents or as a result of repeated courses of stimulation; ( 5 ) quasi-steady visual and auditory cortical currents induced by appropriate sensory stimulation (Kohler et nl. 1955, 1957); (6) infra-slow spontaneous activity (Aladjalova 1957); (7) potentials which accompany local injury or anoxia; (8) slow “temperature” potentials of the hypothalamus (von Euler 1950). A majority of workers have been willing to accept a neuronal origin for slow potentials. However, such phemomena have also been ascribed to the blood-brain barrier (Tschirgi and Taylor 1958), and Hild, Chang and Tasaki (1958) have recorded “responses” from astrocytic glia stimulated in vitru in tissue culture. Slow potentials of the type under discussion here have been studied most intensively in cerebral cortex. There the cortical pyramids have been postulated to show a resting potential gradient which extends downwards along the dendritic shaft from the superficial dendritic plexus to the origin of the axon from the soma. Some composite of the gradients of individual pyramids is presumed to give rise to a difference of potential between surface and white matter. This, plus whatever slow potentials could arise from blood vessels or glial cells, produces a relatively steady baseline from which even prolonged changes in cortical activity can be recorded reliably. This baseline has been called “steady potential”. Actually the term means only that an optimally prepared experiment left undisturbed for several hours will show no more than 500 pV potential drift over the period. From the viewpoint of potentials which take their origins from cell layers it is important that when the laminae of the lateral geniculate nucleus respond to volleys from the optic nerve the potential field distribution is the same for slow potential shifts which arise during repetitive stimulation as it is for unit responses to single volleys (Vastola 1956). This suggests identical neuronal origins for slow as well as for conventionally described components of responses to stimulation. Slow potential components of several classes of evoked potential have been studied under conditions of single and repetitive stimulation and using stimuli of graded intensity. This is an important source of information, since the usually recorded components of direct cortical and direct hippocampal response, cortical recruiting response, and visual, auditory or somesthetic cortical evoked potential can serve as monitors or controls upon the more difficultly examined slow potential aftermaths of the same potentials. The response recorded from cerebral cortex immediately adjacent to a surface stimulating electrode (i.e., direct cortical response) is a graded one and has much to offer t o the
DIRECT CURRENT POTENTIALS
259
study of slow potential phenomena. It is possible to control the stimulus precisely, findings in experimental animals with respect to anesthesia, etc., can be verified in man during the course of neurosurgical operations, and (in animals) details of the change in the several response components which result from topical or intravenous administration of a variety of pharmacological agents are readily determined. It is important to utilize both single and repetitive stimuli, and to secure records a t graded stimulus intensities above threshold. Above threshold, the direct response to a sing!e shock consists of a 15-20 msec negative spike followed by an after-positivity which lasts 150 msec or more. With further increase in stimulus strength the after-positivity is gradually replaced by a potential of opposite sign having a duration of 250-300 msec. A corresponding direct cerebellar response shows no after-positivity (Rhoton et al. 1960). We call this late process slow negativity. During repetitive stimulation at 6 and 20/sec slow negativity summates temporally to produce a significant negative shift in the baseline of the record (Goldring, OLeary and Huang 1958; Caspers 1959). Of these components of direct response the primary process and the after-positivity reverse in the upper few tenths mm of cortex. Slow negativity needs not show a true reversal with increasing depth of a penetrating recording electrode. I n either event it disappears in the upper 0.4 mm. Thus all these response components appear to arise in the superficial cortical layers. The same components can be shown for the recruiting response as activated from midline thalamus (Goldring and OLeary 1957). The primary process and slow negativity components respond differently to pharmacological agents applied topically or given intravenously, and differences also exist for the same components between the direct cortical and recruiting response (Goldring et al. 1959, 1960; Gerber 1961). A few examples suffice to indicate how pharmacological data, support the distinctions we make between these several components. Procaine (given intravenously) abolishes slow negativity while leaving unchanged the primary potentials of both direct and recruiting responses. For the direct response barbiturates abolish the primary negative process and enhance slow negativity; for the recruiting response slow negativity is abolished by barbiturates leaving primary potential unaffected. The different effects upon slow negativity of direct cortical and recruiting responses may result from the multisynaptic path traversed between midline thalamus and surface cortex in the latter as contrasted with the oligosynaptic path of the former. Agents such as intravenous thiopental abolish the slow negativity of the recruiting response while leaving behind primary spike and after-positivity. As a result the usual negative shift during repetitive stimulation a t 6-30/sec is replaced by a minor positive one. The “positive” remainder is the residual after-positivity uncovered by the abolition of the opposing slow negativity. Our evidence suggests that the after-positivity of the direct cortical response is comparable to the after-hyperpolarization described by Eccles and others for spinal motoneurones. Slow potential components of the direct cortical and recruiting responses have been compared in the baby rabbit (Do Carmo 1960): infantile responses being characterized by rapid decay of the slow potential summation which occurs during repetitive stimulation. The salient features of the difference between infantile and adult direct cortical response have been confirmed for man (Goldring et al. 1961). After-effects which correspond to those of the direct cortical response have also been detected for the somatic and visual evoked responses. In rabbit visual response (Pearlman ei af. 1960) one such after-effect has been identified with the slow negativity of the direct cortical response; in that situation slow negativity has a threshold even lower than that of the short-latency primary one which is activated by light flash or electrical stimulus to the optic nerve. With repetitive stimulation the slow negativity of visual evoked response has been shown to sum temporally, just as does that of the direct cortical response. Gumnit (1960) has recorded surface-negative shifts in auditory cortex during continuous sounds of 1-10 sec duration. Such slow potential shifts have been induced in cortex by repetitive stimulation of both medial and lateral thalamus (Brookhart et al. 1958) and of brain stem reticular substance (Arduini et al. 1957). Similar long-lasting responses have been reported to result from direct hippocampal stimulation (von Euler et al. 1957; Kandel et al. 1960). Strychnineand veratrinespikes, induced in cortex by topical applicationoftheseagents, often showprolongedafter-effects (Goldring and O’Leary 1954), although their sure identification with after-positivity or slow negativity of direct cortical response awaits further study. It is important that temporal summation of after-effects occurs during spike clusters. This finding links the temporal summations of slower components of direct cortical and recruiting responses during repetitive stimulation to the sustained d.c. shifts which develop during continuous seizure discharges of several kinds. Strychnine spikes are chiefly negative in polarity, as is their after-effect summation during spike clusters. To the contrary, veratrine spikes are principally positive and their after-effect summation is also positive. l ~ L ’ ~ o c 0~. x261-242 c ~
260
J . L. O’LEARY
et al.
Sustained seizure discharges, induced in the cortex of the experimental animal by repetitive surface stimulation, by similar stimulation in the related thalamic nucleus, or subsequent to topical application of strychnine, is accompanied by a sustained shift of significant magnitude. Commonly, this is surface-positive in polarity and may amount to 1-2 mV(Go1dringand O’Leary 1951b). In different situations it can arise out of the preceding baseline or occur subsequent to a preliminary negative shift which of itself amounts to 1 mV. In the latter instance, if the electrical ictus only amounts to a brief after-discharge, the surface-negative shift may not show recovery before its termination. But in instances of prolonged discharge the cortex shifts positively past its original baseline (by 1-2 mv) before seizure termination. Such an ictus ends in an abrupt surface-negative shift which ushers in the post-ictal period. This, again, can carry the d.c. level significantly across the original baseline of the record. Recent evidence, based upon examination of those brief after-discharge bursts which develop in a cortex unstabilized by a period of intense surface stimulation, suggests that significant recovery from the maximum surface-negativity evident during the stimulus period is critical for the initiation of paroxysm, for after-discharge never appears until the potential nears its pre-stimulatory level. Those convulsant drugs which can be given intravenously in minimal doses to activate a seizure after a latent period (caffeine, picrotoxin, metrazol, strychnine, thiocarbohydrazide, dl-methioninedl-sulfoxamine) produce important support for the neural significance of the slow potential changes which develop during seizure. All such drugs unstabilize the previously steady potential in the latent period preceding the onset of seizure. In the case of thiocarbohydrazide a slowly developing negative d.c. shift precedes the onset of seizure, and in the case of three others (caffeine, picrotoxin and metrazol) a surface-negative change which may amount to 0.5 mV develops rapidly immediately preceding the onset of seizure discharge. We have likened that sudden change to a massive membrane depolarization leading to oscillation of neuronal membrane about the critical firing level of the neurones. dl-methionine-dl-sulfoxanine is unique in producing cycles o f negative-positive shift in the trace which long precede the appearance of any seizure activity. In this respect its effect is similar to that of LeBo’s spreading depression. I t is interesting that the seizures produced by the different convulsants are dissimilar early in their course but become similar in pattern in late phases of discharge. It is also important that during seizure (in the case of metrazol, at least) the cortex becomes inaccessible to direct stimulation, the direct cortical response then being obliterated. This refractoriness to stirnulation continues until the post-ictal negativity has receded to the baseline. All of these points illustrate the important role slow potential phenomena play in seizure origin, course and termination. We believe them to arise largely in membrane changes in cortical neurones as recorded extracellularly by macro-electrodes. Certainly, many parallels can be drawn between observations upon intracellular recording and what we have found in these envelope records of slow activity. What we define here as steady potential is very sensitive to anoxia and other regressive metabolic states. For example, Goldring and O’Leary (1951b) showed that clamping the tracheal airway at room temperature results in a surfacepositivecortical shift of 1-3 mV which develops within 1 rnin following cessation of respiration. This in turn is followed by a much larger negative shift (5-10 mV) which reaches its maximum 2 - 3 min later. The heart ceases to beat at the height of the negative shift. Under hypothermia (25”; Weinstein er a/. 1961) the positive shift does not start until 6-12 rnin after suspension of respiration and the subsequent negative one can be delayed as much as 5-10 min after the start of the positive one. Intravenous malononitrile, injected continuously (Goldring et a/. 1953), which produces intracellular anoxia, similarly affects steady potential, only it takes longer, permitting evaluation of evoked response alterations. During the positive shift positive evoked response after-effects are exaggerated, whereas during the subsequent negative shift it is the negative ones which are increased. Of course, during the negative shift evoked response disappears completely. Sodium hyposulfite introduced before the end of the positive shift, can reverse the process and restore the pre-injection stable state of steady potential. In monkeys we have shown that hypoglycemia produces a similar change; sufficiently early, that change can be reversed by glucose injection. Besides regressive metabolic situations, mechanical injury and vascular arrest are also capable of producing a significant steady potential change. Kempinsky (1954) showed that when the middle cerebral artery was clipped near its source the maximum slow potential source was in the corona radiata and not in overlying cortex. It results from a demarcation potential associated with ischemic change in the masses of myelinated segments found there. Thus certain slow potentials may have no counterpart in normal functioning and are not akin to those described previously. In conclusion I would leave you with the idea that there are three orders of electrical activity of cerebral cortex. First, the activity recorded by macro-electrodes using capacity coupled amplifiers, including, of course, evoked and spontaneous activity and the various epileptic rhythms recorded
DIRECT CURRENT POTENTIALS
26 1
by the electroencephalograph. Formerly we spoke of these as envelope summations of all-or-none unit discharges, visualizing the envelope as arising from cyclic increases and decreases in rate and quantity of unit firing. Today, following Bishop, we talk of dendritic potential, graded response, electrotonic spread, and decremental conduction to describe the electrical change which occurs in the membrane of origin. Spontaneous activity, for example, is sometimes spoken of as an oscillation in membrane potential which occurs at a sub-discharge level for the issue of axonalimpulses. However, we are still unable to state the source of such potentials in a way that would satisfy all investigators. The second order of potentials includes unit activity of all-or-none character as recorded by micro-electrodes. Thanks to studies of Eccles and others we are on firmer ground here when we speak of “potential sources”. However, correlation between presence or absence of clusters of unit potentials as they appear in the micro-electrode trace and the form of the conventionally recorded potentials referred to above presents many complications. The other order includes that activity which due to its slowness is screened out or at the least markedly distorted by capacity coupled amplifiers. After dissecting away such slow potential components as arise out of regressive metabolic states ’ or injury, a significant amount of data support the existence of slow phenomena of indisputably neuronal origin. A final full-bodied knowledge of cortical activity will rest upon the integration of all these three signs of cortical functioning. In our opinion it does little good to assess unit activity with expensive computers if in the programming of data the possibility of slow potential changes during the course of an experiment are ignored. Nor is it probable that meaningful correlates with behavior will arise from consideration of but one of these orders of activity, to the exclusion of the others. We would encourage, for example, more attention being paid to monitoring slow potential changes during the conduct of experiments undertaken for other purposes. Clues leading to significant observations could thus develop. In micro-electrode records from retina (as Brown and Tasaki 1961 ; cat retina) slow potentials of unit dimension have been recorded as responses to light flashes. If the neurones which produce them have counterparts in the central nervous system, it can be postulated that summations of such unit activity could well produce some of the d.c. changes described herein, thus relating more directly the phenomena under discussion to states of excitation and inhibition. REFERENCES ALADJALOVA, N. D. Infra-slow rhythmic oscillations of the steady potential of the cerebral cortex. Nature, 1957, 179: 957. K. Slow potential changes elicited in cerebral cortex ARDUINI,A., MANCIA,M. and MECHELSE, by sensory and reticular stimulation. Arch. ital. Biol., 1957, 95: 127-138. J . M., ARDUINI, A,, MANCIA,M. and MORUZZI, G. Thalamocortical relations as revealed BROOKHART, by i n d x e i slow potential changes. J . Neurophysiol., 1958, 21: 499-525. BROWN,K. T. and TASAKI,K. Localization of electrical activity in the cat retina by an electrode marking method. J . Physiol. (Lond.), 1961, 158: 281-295. CASPERS,H. Ober die Beziehungen zwischen Dendritenpotential und Gleichspannung an der Hirnrinde. Pflugers Arch. ges. Physiol., 1959, 269: 157-181. Do CARMO,R. J. Direct cortical and recruiting responses in postnatal rabbit. J. Neurophysiol., 1960, 23: 496-504. EULER,C. VON Slow “temperature potentials” in the hypothalamus. J. cell. comp. Physiol., 1950, 36: 333-350. EULER,C. VON, GREEN,J. D. and Rlcci, G. The role of hippocampal dendrites in evoked responses and after-discharges. Acta physiol. scand., 1957, 42: 87-1 11. GERBER, C. J. Effect of selected excitant and depressant agents on the cortical response to midline thalamic stimulation in the rabbit. Electroenceph. clin. Neurophysiol., 1961, 13: 345-364. GOLDRING, S. and O’LEARY,J. L. Summations of certain enduring sequelae of cortical activations in the rabbit. Electroenceph. clin. Neurophysiol., 1951a, 3 : 329-340. S. and O’LEARY,J. L. Experimentally derived correlates between ECG and steady cortical GOLDRING, potentials. J . Neurophysiol., 1951b, 14: 275-288. S. and O’LEARYJ. L. Correlation between steady transcortical potential and evoked GOLDRING, response. Electroenceph. elin. Neurophysiol., 1954, 6 : 189-212. S . and O’LEARY,J . L. Cortical d.c. changes incident to midline thalamic stimulation. GOLDRING, Electroenceph. din. Neurophysiol., 1951, 9 : 577-584. S. and O’LEARY,J. L. Pharmacological dissolution of evoked cortical potentials. Fed. GOLDRING, Proc., 1960, 19: 612-618.
262
J . L. O’LEARY
et al.
GOLDRING, S., O’LEARY, J. 1.and HUANC,S. H . Experimental modificationof dendriticandrecruiting processes and their d.c. aftereffects. Electroenceph. din. Neurophysiol., 1958, 10: 663-676. GOLDRING, S . , O’LEARY,J. L. and LAM,R. L. Effect of malononitrile upon the electrocorticogram of the rabbit. Electroenceph. d i n . Neurophysiol., 1953, 5 : 3 9 5 4 0 0 . GOLDRING, S., METCALF, J . S., HUANG,S. H., SHIELDS, J. and O’LEARY,J. L. Pharmacological selectivity manifested by agents acting upon the cortical dendritic spike and its slow aftereffects. J . new. ment. Dis., 1959, 1281 1-1 I . GOLDRING, S., JERVA,M . J. HOLMES, T. G., O’LEARY,J. L. and SHIELDS,J. R., Direct response of human cerebral cortex. Arch. Neurol. (Chic.), 1961, 4 : 590-598. GROSSMAN, R. G. and GUMNIT,R. T. A method of recording from the cortex while maintaining physiological conditions. Elecfroenceph. clin. Neurophysiol., 1960, 12: 920-921. GUMNIT,R. J. d.c. potential changes from auditory cortex of cat. J . Neurophysiol., 1960, 23: 667-675. HILD,W., CHANG J. J. and TASAKI, I. Electrical responses of astrocytic glia from the mammalian nervous system cultivated in vifro. Experienfia, 1958, 14: 220-221. KEMPINSKY, W. H. Steady potential gradients in experimental cerebral vascular occlusion. Electroenceph. din. Neurophysiol., 1954, 6 : 375-388. KEMPINSKY, W. H. Experimental study of distant effects of acute focal brain injury. Arch. Neurol. Psychiat. (Chic.), 1958, 79: 376-389. KOHLER, W., NEFF,W. D. and WEGENER, J . Currents of the auditory cortex in the cat, and currents of the human auditory cortex. J . cell. cot77p. Physiol., 1955, 4.5 (Suppl.): 1-54. KOHLER, W. and O’CONNELL, D. N. Currents of the visual cortex in the cat. J. cell. romp. Ph.vsiol., 1957, 49 (Suppl. 2): 1 4 3 . LEAo, A. A. P. The slow voltage variation of cortical spreading depression of activity. Electroenceph. din. Neurophysiol., 1951, 3 : 315-321. O’LEARY, J. L., KERR,F. W. L. and GOLDRING, S. The relation between spino-reticular and ascending cephalic systems. Reticular formation of the brain, Little, Brown Co., Boston, 1957. PEARLMAN, A. L., GOLDRING, S. and O’LEARY,J. L. Visually evoked slow negativity in rabbitcortex. Proc. Soc. exp. Biol. (N.Y.), 1960, 103: 600-603. RHOTON, A., GOLDRING, S. and O’LEARY,J. L. Comparison of direct cerebral and cerebellar cortical response in the cat. Amer. J . Physiol., 1960, 199: 677-682. ROWLAND,V. Simple non-polarizable electrode for chronic implantation. Elecfroenceph. d i n . Neurophysiol., 1961, 13: 290-291. TSCHIRGI, R. D. and TAYLOR, J. 1.Slowly changing bioelectric potentials associated with the bloodbrain barrier. Amer. J . Phjjsiol., 1958, 195: 7-22. VANASUPA, P., GOLDRING, S. and O’LEARY, J. L. Seizure discharges effected by intravenously adniinistered convulsant drugs. Electroenceph. clin. Neurophysiol., 1959, I I : 93-106. VANASUPA, P., GOLDRING, S. O’LEARY, J . L. and WINTER, D. Steady potential changes duringcortical activation. 1. Neurophysiol., 1959, 22: 273-284. VASTOLA, E. F. Steady potential responses in the lateral geniculate body. Electroenceph. clin. Nriirophysiol., 1956, 7 : 557-567. WEINSTEIN, W., KENDIG,J. H., GOLDRING, S., O’LEARY,J . L. and LOURIE,H . Hypothermia and electrical activity of cerebral cortex. Arch. Neurol. (Chic.), 1961, 4 : 441-448. W. R. ADEY: Dr. O’Leary has raised some most fascinating questions in his studies of cortical d.c. potentials. We have also been interested in the possibility of recording slow changes in cerebral activity, and we have developed a new technique for the measurement of impedance in any small volume of cerebral tissue in chronically implanted, freely moving animals. This technique has indicated that it is possible to record both fast and slow changes in impedance in cerebral structures, and to show rhythmic changes in impedance which, in most instances, are independent of concurrent electrophysiological rhythms in:the conventional sense. With our technique, developed colkaboratively with my colleagues R. T. Kado and J. Didio, thc impedance measurements are made with exceedingly small measuring currents, of the order of l O - l : j amperes per square micron of exposed electrode surface. In this way, we have mitigated the possibility of stimulation of tissue in which the measurements are made. The coaxial electrodes used have an external diameter of 0.5 mm, and measurements on an electrode model have shown that most of
DIRECT CURRENT POTENTIALS
263
the measuring current is distributed within a distance of half the diameter of the electrode. All measurements have been made at 1000 cycles per second, using a resistive Wheatstone bridge, and the imbalance signal from the bridge amplified through an amplifier with low-noise input and highly differential characteristics. The sensitivity of the system is limited only by the noise input level of the amplifier and it can readily resolve a 1% shift in baseline impedance. By the use of a pulsesampling technique, we have obtained an integrated readout of the impedance. The readout has a bandpass from d.c. to 30 cycles per second, and in appearance resembles an EEG record. Our measurements have been made principally in the dendritic layer of the pyramidal cells of the dorsal hippocampus with the dipole of the coaxial electrode oriented vertically, approximately in the long axis of the dendritic trees of these cells. Measurements have also been made in septal and amygdaloid areas. We have found that all physiological stimuli tested, including somatic, visual, auditory and olfactory, are associated with a rapid transient drop in impedance, lasting for a variable period and followed by a return to the original baseline. These impedance changes frequently outlast discernible EEG changes. In normal sleep there is a gradual drift in the impedance baseline towards higher levels with the drowsy state, and establishment of a high level plateau during sleep, with return to the original level on waking. There are striking rhythmicities in septal impedance records during normal sleep, with trains of very slow waves around 3 cycles per minute. In nembutal anaesthesia there is a progressive rise in impedance, and at the deepest anaesthetic level, a loss of transient responses to painful skin stimuli. In lightening anaesthesia, these transient responses gradually return, with increasingly abrupt rising phases. By contrast, hallucinogenic cyclohexamine drugs are associated with a prolonged drop in impedance in hippocampal dendritic structures. During the influence of these drugs, sudden stimuli, which produce only transient EEG changes, are associated with a very abrupt fall in dendritic impedance and an extremely slow return to the original level over a period of many minutes. During hippocampal seizures, induced by septal stimulation, there are irregular, rapid perturbations in impedance during the seizure, but during the period of EEG after-discharge there are long trains of rhythmic waves in the hippocampal dendritic impedance records at about 1 cycle per second, occurring on a slowly rising baseline. Repeated seizures may shift the baseline impedance to progressively higher levels. It is our hope that extension of these methods will permit us to elucidate some of the factors, such as ionic shifts and metabolic changes, occurring concurrently or causally in relation to the electrogenesisof the slow wave process. We hope to examine the relative participation of intraneuronal, intraglial and extracellular compartments in this process. We are considering the possibility that glial tissue may provide a non-linear impedance load, in both space and time, to the intrinsic electronic generators within the dendritic tree. We hope further to extend these investigations by the use of electromagnetic waves in the millimetric range.
ADEY,W. R., KADO,R. T. and DIDIO,J. Impedance measurements in brain tissue of chronic animals using microvolt signals. Exp. Neuro!., 1962, 5 : 47-66.
J. C. ECCLES: In considering the mode of generation of the potentials recorded by O’Leary from the surface and depth of the cerebral cortex, it is important to realise that there are two types of neural element that are oriented vertically to the cortical slab of tissue: the pyramidal cells with their apical dendrites and their axons passing deeply into the white matter; and the afferent fibres from the white matter entering the grey matter and usually passing up to the most superficial layer. Neurones with dendrites and axons randomly orientated within the grey matter will not contribute appreciably to the cortical potential. The original suggestion was that the brief negative potential was due to the synaptic depolarisation of the apical dendrites of pyramidal cells. I still think that this is the most likely explanation, conforming, as it does, with the time course, the size, and the diminution and reversal of the potential at progressively greater depths of the cortex. However that may be, the very slow negative wave is now of particular interest in relation to the possibility that there may bepresynaptic inhibition in the cerebral cortex. Thus if the surface stimulation evokes activity in an interneuronal system which has depolarizing synapses on the afferent fibres near their synaptic terminals, almost all of Dr. O’Leary’s observations are readily explained. Both the build up of the slow depolarisation during repetitive stimulation and its long duration correspond closely to that observed with primary
264
J . L. O’LEARY
et at.
afferent fibres. Furthermore, there is at the same time a reduction of the brief negative waves, which suggests that the synaptic excitatory action of the afferent fibres is reduced, which is the expected effect of the presynaptic depolarisation. Finally there is the effect of nembutal in potentiating and prolonging the slow negative wave, which corresponds to its action on presynaptic depolarisation in the spinal cord. However the action of picrotoxin inalso increasing the slow negative wave would be against its identification as a presynaptic depolarisation agent. I would like to ask O’Leary at what depth does the slow negative wave reverse to a positive wave?
R. JUNG: First I should like to congratulate O’Leary for his fine paper and especially for avoiding the notorious term “dendritic potentials”. Although the dendritic vogue started in St. Louis, O’Leary’s term “direct cortical response” (d.c.r.) seems to be more appropriate. Secondly 1 would like to continue Eccles’s remarks about the sources of cortical potentials. There must be a definite geometrical arrangement of the potential field, perpendicular to cortical surface for the generators of d.c. potentials and cortical brain waves, because all investigators have found a dipole distribution. Mainly on the basis of Goldring’s and O’Leary’s findings 1 have postulated since 1953 that steady potentials and brain waves can be explained by two opposite dipoles, one sirface-negative and one surfare-positive. The various brain waves and d.c. alterations should be the result of a regulated balance of two oppositely directed d.c. potential sources oscillating normally around a medium level. In 1958 I proposed the scheme of Fig. 1 to explain by this conception the various brain wave changes with electrical and pharmacological alterations. Although some details may have to be corrected 1 still believe this scheme to be a good working hypothesis for further investigations of the neurophysiological basis of the EEG. It would be in good agreement with Bishop’s,
-1
Polarization surface-negative
f
- (Catelectrotonus) e r a t r i n , KCI, GAB e t i c u l o - t halam.-stim. o r t i c a l stirn. series l o r m a l level ‘olarization urface -positive Anelectrotonus)
tf
+!
b
+
Spec thalam. s t i m . convuIsions,onset of anoxia, Malononitrile
Fie. ” 1. Hypothetical scheme of brain potentials with two opposite d.c. sources, altered by chemical or electrical influence (From Jung 1958). a,Spontaneous rhythms of alpha waves at different d.c. levels; 6, evoked potentials of primary receiving areas (following afferent nerve stimulation) at different d.c. levels; c, direct cortical responses after cortical stimulation (so-called “dendritic potentials”) at different d.c. levels. The first positive deflection of evoked potentials is less influenced than the negative deflection, most prominent in “dendritic potentials”. After large d.c. deviations from the normal steady-state in either positive or negative direction, the normal brain waves (-) may be converted into large convulsive potentials (- - -), tending to sweep to the normal midline level. The scheme is based on the assumption that brain waves may be the result of a regulated balance of two oppositely directed d.c. potentials oscillating normally around a relatively steady medium level. Alterations of the level of this balance may be electrophysical (caused by external polarisation), physiological (by non-specific afferences), chemical (by drug application), metabolic (by anoxia) or pathological (convulsions, spreading depression). Alteration of d.c. level may be caused either by increase of one d.c. component in the same direction or by diminution of the oppositely directed d.c. component (indicated by length of the arrows).
DIRECT CURRENT POTENTIALS
265
OLeary’s and Goldring’s findings and with the results which Creutzfeldt and others have found a t the neuronal level after surface polarisation of the cortex. Although it agrees also with Caspers’s results (1959), it is opposed to Caspers’s explanation, who admits only a single unidirectional source of “Dendritenpotentiale” for his d.c. potentials, which are modulated to brain waves. I think it fits better with observed facts if we admit two opposed dipoles in the cortex, located perpendicularly. How can this conception of a double potential source be combined with Eccles’s findings? If you allow me some simplification, I would propose the following: The surface-negative component might be produced in the main presynaptic fibres which run to the upper layers of the cortex and there receive presynaptic inhibition similar to the primary afferent fibres in the spinal cord. Eccles has shown that these root fibres generate the dorsal root potentials by their presynaptic depolarisation. Similarly cortical afferent fibres may also generate cortical slow potentials although they need not be their only source. An argument for the participation of afferent fibres would be the remarkable electrotonic spread which brain potentials show in the white matter. The surface positive component might be caused mainly by apical dendrites and their potential gradients to the cell bodies. This gradient might be diminished or augmented by postsynaptic depolarisation or hyperpolarisation at the dendritic membrane. The variations in presynaptic inhibition in the afferent fibres and of postsynaptic inhibition or excitation at the dendrites could then well explain the d.c. sources of oscillating brain waves, shown in Fig. 1 . This hypothesis would explain why brain waves are often connected with inhibition. It fits also the hypothesis of the “Bremswelle” derived from the diminution and reversal of direct cortical response following repeated cortical stimulation, leading to convulsive discharges (Jung and Tonnies 1950). The same phenomenon was also shown in O’Leary’s records. O’Leary and Caspers denionstrated clearly the d.c. component in these alterations of the direct cortical response which we could not at that time (in 1950) record exactly. 1 would be interested to know whether O’Leary would agree with theassumptionof two opposite steady potentials in the cortex which was first suggested by his own work.
BISHOP,G. H. and O’LEARY,J. L. The effects of polarizing currents on cell potentials and their significance in the interpretation of central nervous system activity. Electroenceph. din. Neurophysiol., 1950, 2: 401-416 CASPERS, H. Uber die Beziehungen zwischen Dendritenpotential und Gleichspannung an der Hirnrinde. Pjiig. Arch. ges. Physiol., 1959, 269: 157-181. FROMM, G. H., KAPP,H. und CREUTZFELDT, 0. D. Beziehungen zwischen corticaler Gleichspannung und Neuronenaktivitat. Pflug.Arch. ges. Physiol., 1960, 272: 51. GOLDRING, S. and O’LEARY,J. L., Correlation between steady transcortical potential and evoked response. 1. Alterations in somatic receiving area induced by veratrine, strychnine, KCI and novocaine. 11. Effect of veratrine and strychnine upon the responsiveness of visual cortex. Electroenceph. clin. Neurophysiol., 1954, 6 : 189-200; 201-212. JUNG,R, Neuropharmakologie: Zentrale Wirkungsmechanismen chemischer Substanzen und ihre neurophysiologische Grundlagen. Klin. Wschr., 1958, 36: 1 153-1 167. J. F. Hirnelektrische Untersuchungen uber Entstehung und Erhaltung von JUNG,R. und TONNIES, . Krampfentladungen: Die Vorgange am Reizort und die Bremsfahigkeit des Gehirns. Arch. Psychiat. Nrrvenkr., 1950, 185: 701-735.
W. GREYWALTER: Has O’Leary any information on the effects of COz on the slow potential changes? After all, the COz or carbonic acid content of the blood is more likely to vary than the 0 2 , and is known to have marked effects on excitability and stability in many parts of the brain as well as on the cerebral blood vessels. Also, is it not possible that since his recording conditions include a fairly large standing potential difference (which he balances o u t to some extent) he might be recording resistance changes as well as potential differences? These might be of great interest, as suggested by Adey-a sort of cortical psycho-galvanic response-but they would confuse interpretation of the true slow cortical potentials.
266
J . L. O’LEARY et
ul.
D. ALBE-FESSARI): It seems to me that the results presented by O’Leary contradict the interpretation of Purpura and Grundfest regarding the first surface-positive, subsequently surface-negative evoked potentials. These authors believe that the surface-negative response obtained by local stimulation corresponds with a phase of excitation; this, suppressed by the GABA, unmasks a surface-positive phase of supposedly inhibitory character. Yet you have shown (and to me this seems true for the evoked potential proper) that excitation is always linked up with a surface-positive phase (primary or secondary) rather than a negative phase, which develops without a concomitant sign of excitation. PURPURA, D. P. and GRUNDFEST, H. Physiological and pharmacological consequences of different synaptic organizations in cerebral and cerebellar cortex of cat. J. Neurophysiol., 1957, 2 0 : 494-522.
A. FESSARD: O’Leary, you said that you doubt that steady electric fields, as produced in central nervous structures, can really have a functional role. It is certainly wise toavoid speculation on this point. However, when there is a certain congruence between the spatial order existing within the nervous tissue and that displayed by the total electric field produced by the active elements of this tissue, 1 have little doubt that some mutual influences of an electrical nature play a significant role, namely in inducing synchrony of spontaneous activities. The regularity of architectonics in some structures, as in the cerebral or cerebellar cortex, hippocampus, etc. seems to make these structures sites of election for such electric field actions, which thus may represent a factor of integration. I would like to mention here recent experiments by Calvet andScherrer (196l)showing unitdischarges from pyramidal cells stopped each time a slow surface-negative wave appears during spontaneous spindling. This can be imitated by applying an artificial transcortical field of the same direction and of the same order of magnitude. One is tempted to interpret the influence of the surface-negative spindles on the spike discharges in the following way: two populations of cortical neurons, one producing surface-negative spindles, the other sending off spike discharges are assumed to be tightly interspersed, so that the transient fields produced by the former exert depressive actions upon the trigger zones of the latter. J . Relation des decharges unitaires avec les ondes cerebrales spontsnCes CALVET,J. et SCHERRER, et la polarisation corticale. C . R . Acau‘. Sci. (Paris), 1961, 252: 2297-2299.
F. BREMER: Local humoral and circulatory factors should certainly contribute to the complex determinism of d.c. potentials. Polygraphic registrations recently published by Meyer and Gotch (1961) showed that there is an interaction between cerebral haemodynamics and metabolism which must have its repercussion on the neuronic membrane potentials. However, the principal factor in these slow changes of surface potential of the cortex seems to be, a t least initially, represented by postsynaptic reactive depolarization of the superficial dendrites of the neocortex. The observations of Bonnet (1 957) seem to me very illustrative in this respect. She demonstrated that the amplitude and stability of the slow negative potential produced by repeated direct stimulation of the cortex are considerably increased following local application of physostigmine, and also of calcium ions which, as shown by studies in Katz’s laboratory, facilitate ejection of the acetylcholine mediator at the neuromuscular junction. These experiments --like those of Maclntosh and Oborin - suggest a cholinergic innervation of the superficial cortical dendrites. MEYER and GOTCH,Neurology (Minneap.), 1961, 11: 46. BONNET,V. La transmission synaptique d’influx au niveau des dendrites superficials de I’ecorce ckrebrale. Arch. int. Physiol., 1957, 65: 506-51 1 . P. E. Abstr. XlXlntern. Physiol. Congr., 1953. Montreal, 1953:580-581. MACINTOSH, F. C. and OBORIN,
DIRECT CURRENT POTENTIALS
267
O’LEARY’s replies To W. R. Adey Adey has developed a most interesting method which is worth exploiting with respect to the possibility of d.c. correlates. Historically, LeZo reported that impedance increases in spreading depression and that finding was confirmed by Freygang and Landau (1955) who passed a constant current pulse normal to the cortical surface and recorded the IR drop between two electrodes in the tissue. They found that impedance increased 20% with spreading depression and was down a few per cent with a locally evoked potential. Later Van Harreveld (1956) claimed to have shown that dendrites swell in spreading depression, presumably associated with the impedance change. That work is perhaps important with respect to Adey’s measurements in the stratum radiale of the dorsal hippocampus which of course contains numerous dendrites. Besides d.c. potentials the importance of correlations with pH and 0 2 electrode measurements needs mention, particularly since Adey mentions ionic shifts and metabolic changes as possibly related to impedance changes.
FREYGANG, W. H. and LANDAU, W. M. Some relations between resistivity and electrical activity in the cerebral cortex of cat. J. cell. comp. Physiol., 1955, 45: 371-392. VAN HARREVELD, A. and OCHS, S . Cerebral impedance changes after circulatory arrest. Amer. J. Physiol., 1956, 187: 180-192. To J. C. Eccles Like Eccles we recognize the likelihood of a dendritic origin for the direct cortical response and agree that other explanations, however remote, are possible. For example, the dendrites are embedded in a dense tangle of fine axons, the terminals of which could also contribute to the depolarization process. Also many such axons arise from short axon cells and Bullock (1959) has made the point that axons of short axon cells might differ from those of pyramids in retaining graded response characteristics. I am not sure but that Eccles himself did not entertain that notion also at some time in his past. Nevertheless it seems to me that the studies of Kandel and Spencer (1961), together with those of Green (1960), Eidelberg (1961) and others upon hippocampal pyramids have brought forward a very strong case for a dendritic origin of at least the primary component of the direct hippocampal response. Kandel and Spencer even monitored intracellular records with surface ones and concluded that (with exceptions) the surface record reflects fairly well the intracellular one. The suggestion Eccles put forward concerning the very slow negative wave is a very intriguing one. 1 know of no electron microscopic evidence for depolarizing synapses (or any synapses) upon afferent fibers near their synaptic terminals; and from the viewpoint of ultrastructure Gray, de Lorenzo and others have studied cortical synaptology rather thoroughly. However, such contact points would be extremely minute and could be missed of course. It is important that some of our results appear to coincide with Eccles’ finding upon presynaptic inhibition in the cord (nembutal, for instance). 1 do not take seriously the discrepancy with respect to picrotoxin since our results were not obtained with the direct cortical response in that situation, or with anything corresponding to close arterial injection. Rather, the drug was given intravenously and the allied recruiting response was in use. There, the augmentation of the slow negative wave could have been affected at the stimulus source in the midline thalamus as well as at the cortical end-station. When 1 return home we shall repeat our picrotoxin experiments with the direct cortical response and topical application. BULLOCK, T. H. 1959, see References, Reply to Dr. Jung E. Hippocampal “dendritic” response in rabbits. J. Neurophysiol., 1961, 24: 521-533. EIDELBERG, D. S., SCHINDLER, W. J. and STUMPF, C. Rabbit EEG “theta” rhythm: GREEN,J. D., MAXWELL, its anatomical source and relation to activity in single neurons. J. Nerrroghjsiol., 1960,23: 403420. W. A. 1961, see References, Reply to Dr. Jung. KANDEL, E. R. and SPENCER, To R.Jung Professor Jung’s theory encompasses oppositely directed dipoles at work in the cortex, a surfacenegative one in presynaptic fibers and a surface-positive one in dendrites. What Bishop termed “dendritic” potentials are those aroused indirectly by stimulation at a distance in thalamus or cortex and recorded from the most superficial layer of cortex where few cell bodies are found. They were obviously post-synaptic graded responses. In St. Louis a distinction is still made between such
268
J. L. O’LEARY
et at.
responses and ‘&directcortical” responses, those aroused by direct stiniulation at the cortical surface and recorded at a distance of not over 0.5 mm from the point stimulated. These two patterns of response are different in many details, not all of which are at present understood. In one particular especially, the direct cortical response must be the more complex one since it may include direct as well as post-synaptically excited potentials. Support for the theory that the direct cortical response includes post-synaptic responses of apical dendrites has come from various laboratories, and of late particularly from the beautiful work of Kandel and Spencer on the hippocampus. incidentally those workers while supporting dendritic origins follow our nomenclature in describing a direct hippocampal response which arises near the stimulated point in the region of basal (opposite to apical) dendrites of the hippocampal pyramids. At one time I worked over the Golgi architecture of several visual cortices rather methodically. In addition to the forest of dendritic shafts one sees an amazingly intricate plexus of fine fibers, disposed principally vertically. This plexus is densest in the layer of superficial pyramids and contains many short axon arborizations. Bishop has recently been working upon electron microscopic preparations of cortex and underlying white matter. He finds that many very fine fibers all of which are myelinated in the white matter, become unmyelinated in their course upwards and contribute unmyelinated terminals to the superficial layer of neuropil. Their activation in white matter results typically in a surface-negative postsynaptic dendritic potential, and they must also be stimulated by directly applied shocks at the cortical surface. When stimulated at a distance from the recording locus (as in his “dendritic potential” work) these presynaptic potentials are usually too low in amplitude to be recognized above the ever present spontaneous cortical noise. Their potentials probably contribute relatively little, however, as compared to the postsynaptic responses in any type of cortical record. Reverting to the fine fiber plexus between the dendritic shafts and the matter of direct stimulation, Bullock has been daring enough to propose that short axons as well as dendrites might be graded response elements. Now this relatively superficial snarl of axons must provide thousands of synapses upon the dendritic shafts and terminal ramifications of the pyramids. So 1 have some difficulty in dissociating the possibility of a pre- from a postsynaptic origin for all components of the direct cortical response, and that is why 1 take seriously the possibility that our late slow negativity component could be a manifestation of presynaptic inhibition as conjectured earlier by Eccles. The fact that the main components of the direct cortical response are surface-negative rather than surface-positive does not in my opinion offer too much evidence of the actual dipoles involved. The stimulus is applied to the cortical surface close upon the site of recording and creates there a sink, the principal sources for which are deep to the stimulus point and fade away from its margins. It is important to us that the initial 15-20 msec negative potential is followed by a trailing surfacepositivity. Recently we went to much trouble to prove that this after-positivity is real and reverses just as close to the surface as does the preceding negative component. Thus it is not evidence for a sink of deep origin such as is inferred in Adrian’s classical study of 1936; and we believe that it is truly the same potential as that of after-hyperpolarization recorded intracellularly from spinal niotoneurones by Eccles. What unit potential work we have done indicates that it is of the class of postexcitatory depressions. Now Bishop and Clare working with dendritic and recruiting potentials also found evidence of a post-excitatory depression following those negative components, and Jung will remember that Bishop and Bartley employed the term early in cortical neurophysiology in relating evoked response phenomena to alpha rhythm. Whether one calls this inhibition seems to be a matter of taste. For our trouble in showing that there is an indubitably positive potential that reverses very close to the surface, I got the dry comment from my colleagues on the J . Neurophysiol. staff that we had gone to a good deal of trouble to prove it. I am sure that Eccles would have been much more lenient with us, for this is perhaps a link in the evidence relating cortical potentials to membrane physiology as revealed in spinal motoneurones and crustacean stretch receptor neurones. We did not simply lift terms bodily from an intracellular environment and levitate them into an extracellular one, nor did we recognize a specific inhibiting process buried beneath simultaneously existent excitatory postsynaptic potentials. The late higher threshold slow component which swamps the after-positivity of the primary component as it grows with increasing stimulus strength has worried us if viewed as having a dendritic origin such as Bishop hypothesizes for the early component. Could a dendrite with no refractory period and graded response characteristics produce two negative response components to a single stimulus, one early and relatively rapid, the other late and significantly slow? Perhaps so, but that thinking would require more evidence than has come our way.
DIRECT CURRENT POTENTIALS
269
Finally, to return to Jung’s discussion of his theory which set off this effort to explain our position, I think thatif one considers the possible neural elements that could be involved, including the sources postulated by Jung, the cortex is certainly complex enough for his two dipole theory. ADRIAN,E. D. The spread of activity in the cerebral cortex. J. Physiol. (Lond.), 1936,88: 127-161. BISHOP,G. H. The interpretation of cortical potentials. Cold Spring Harbor Symposia on Quantitative Biology, 1936, Vol. IV: 305-319. BISHOP,G. H. Personal communication. BULLOCK, T. H. Neuron doctrine and electrophysiology. Science, 1959, 129: 997-1002. CLARE,M. A. and BISHOP,G. H. Properties of dendrites; apical dendrites of the cat cortex. Electroenceph. clin. Neurophysiol., 1955, 7 : 85-98. KANDEL, E. R. and SPENCER, W. A. Electrophysiology of hippocampal neurons. 11. After-potentials and repetitive firing. 111. Firing level and time constant. IV. Fast prepotentials. J. Neurophysiol., 1961, 24 : 243-259 ; 260-27 1 ;272-285. KANDEL, E. R., SPENCER, W. A. and BRINLEY, F. J., JR. Electrophysiology of hippocampal neurons. I. Sequential invasion and synaptic organization. J. Neurophysiol., 1961, 24: 225-242.
To W. Grey Walter In cats we have tried the inhalation of 10% carbon dioxide and 90%oxygen, using artificial respiration and flaxedil for immobilization. Two or three minutes after starting the COZ blood pressure rises and pulse rate is increased. At the time we saw a definite decrease in amplitude of both repetitively evoked primary potentials and of the underlying summed slow negativity which provides the associated d.c. shift. If such ventilation was continued for 3 to 5 min all trace of response disappeared. With 5% COz the same changes developed but took a longer time. In several experiments blood pH was measured before and during COZadministration. It fell from 7.4 to 7.1 associated with the elevation in blood pressure and diminution in response. We tried to repeat these experiments on human cortex using usual anesthetic equipment but vascular pulsation developed vitiating our recording situation. With respect to the other question relative to the extent to which we are measuring resistance changes, it is my opinion that in our circuit we are measuring principally, if not exclusively, potential differences. To D. Albe-Fessard As Dr. Albe-Fessard says, there is a difference in the strata in which specific and non-specificafferents terminate. The non-specific axonal component terminating throughout, the specific plexus being evident in the middle layers. The reversals of potential during probing of the cortical thickness localize the origin of the direct cortical response to the region superficial to the specific plexus. In further reply to Dr. Albe-Fessard our results upon GABA do differ from those published by Purpura and Grundfest. The fact that GABA reverses the polarity of the 20-msec primary DCR potential from negative to positive led these workers to view GABA as abolishing negative polarity EPSPs, leaving unopposed positive polarity IPSPs behind. In our work the primary negative component of either direct cortical or recruiting response was also shown to be reversed by GABA. However, the situation is more complicated than Purpura and Grundfest had supposed. The positivity after GABA has a duration longer than that of the primary negative component of the normal DCR and actually represents a fusion of the “reversed” initial component with a still unaffected after-positivity which can be shown to follow the initial potential which existed prior to GABA application. Besides, GABA strikingly augments the later slow negativity component of both direct cortical and recruiting responses. As a result the surface-negative d.c. shift which develops during repetitive stimulation, as illustrated today, is markedly augmented. Thus in the records of repetitive stimulation one sees after GABA the paradox of positive polarity transients rising out of a considerably augmented negative shift. Furthermore, the initial part of the GABA induced positivity was found to be a source for a deep sink of current flow, for we always found a negativity, 0.2 to 0.4 mm below the cortical surface corresponding in time with the surface positivity produced by GABA application. Putting this together we interpret the reversal of potential by GABA as resulting from depression of the cortical elements in the superficial plexus. Additional support for such a view was that surface application of non-specific depressants such as heat and veratrine produce a change similar to the one evoked by GABA. Finally our findings on GABA application to cerebellar cortex also differ from those reported
270
J . L. O’LEARY
et al.
by Purpura and Grundfest. They observed that GABA did not reverse the primary negative potential of the cerebellar DCR to positive and interpreted that observation as indicative of absence of lPSPs in cerebellar cortex. We found that GABA did reverse the primary negative potential to a positive deflection. With regard to the cerebellum it is also of interest that Granit and Phillips doing single unit recording from Purklnje cells have shown inhibition associated with both hyper- and depolarizing membrane shifts.
R. and PHILLIPS, C. G. Excitatory and inhibitory processes acting upon individual Purkinje GRANIT, cells of the cerebellum in the cat. J. Physiol. (Lond.), 1956, 133: 520-547. PURPURA, D. P. and GRUNDFEST, H . Physiological and pharmacological consequences of different synaptic organizations in cerebral and cerebellar cortex of cat. J . Neurophjaiol., 1957, 20: 494-522. PURPURA, D. P., GIRADO,M. and GRUNDFEST, H. Selective blockade of excitatory synapses in the cat’s brain by gamma aminobutyric acid (GABA). Science, 1957, 125: 1200-1202. To A. Fessard Perhaps 1 spoke too dogmatically in denying a functional role to steady electrical fields. 1 am perfectly willing to entertain the proposition that a poised neurone might have its activity influenced extrasynaptically by a weak electrical field in which it is situated. In fact 1 rather favor it. However, for its wholesome development the area of d.c. potentials requires the same kind of careful scientific approach that has made possible the advances in motoneurone and invertebrate electrophysiology. Too free speculation could ring the death knell upon a promising development by encouraging the dilettante to enter before the bulwark of evidence is sufficient to negate his errors. 1have been much interested in the experiments of Calvet and Scherrer showing that unit discharges from pyramidal cells stop each time a slow negative wave appears during spontaneous spindling. Upon whether or not there are two populations of neurones I have no evidence and no valid present opinion. Once Bishop and I proposed that there were two such interlocked populations of neurones one of which supported spontaneous and the other evoked activity. This had some general support for a while but was replaced by the theory that a single population did the whole job. It is important in support of your thesis that single units show such differences in their responses to extrinsic stimuli -take, for example, the work of Jung and of Hubel on visual cortical units. It would seem that some master system of neurones would be necessary to effectively integrate such differently acting components. J. Relation des decharges unitaires avec les ondes certbrales spontanees CALVET, J. et SCHERRER, et la polarisation corticale. C.R.Acut/. Sci. (Purik), 1961, 252: 2297-2299. HUBEL,D. B. Single unit activity in striate cortex of unrestrained cats. J . Physiol. (Lond.), 1959, 147: 226-238. JUNG,R. Excitation, inhibition and coordination of cortical neurones. Exp. Cell Res., 1958, Suppl. 5 : 262-271. O’LEARY,J. L. The role of architectonics in deciphering the electrical activity of the cortex. In P. Bucv (Editor), Precentral Motor Cortex. University of Illinois Press, Urbana, Ill., 1949, Chapt. 111: 84-110.
To F. Bremer We would certainly agree that cortical d.c. potentials have a composite origin which includes the operations of humoral and circulatory factors and of injury. Over the past ten years we have examined the effects of a variety of agents in altering the d.c. components. Anoxia, for example, has been studied by clamping the airway and producing intracellular anoxia through the introduction of cyanide. Such drastic interferences radically alter the steady potential levels and at the same time produce marked regressive changes in evoked potential and spontaneous activity components. We have worked to distinguish between those effects of anoxia which directly affect excitability and those which result from irreversible injury. In the same way we have examined theeffect of application of several gases and gaseous anesthetics, and of introducing agents such as barbiturates known to fix membrane potential. The sum of our experiences has given us the confidence that in limited situations, at the least, we are able to separate the d.c. changes which can be said to originate in cell membrane as excitatory or depressive processes of functional significance, from those which reflect
DIRECT CURRENT POTENTIALS
27 1
the consequences of changes in the internal milieu or are to be interpreted as deterioration in the preparation. With reference to the question of whether or not the superficial synapses of the cortex are cholinergic, we have no evidence of our own since we have not studied the effects of local application of physostigmine or of Ca2+ to the cortex. However, in a recent study of the ultrastructure of cortical synapses by electron microscopy de Lorenzo has inferred that a difference exists in osmiophilia between axodendritic (apical dendritic system) and axosomatic synapses, the former perhaps being cholinergic. DE LORENZO, A. J.
Electron microscopy of the cerebral cortex. I . The ultrastructure and histochemistry of synaptic functions. Bull. Johns Hopk. Hosp., 1961, 108: 258-279.
Studies of Non-Specific Effects upon Electrical Responses in Sensory Systems HERBERT H. JASPER The Montreal Neurological Institute of McCill University, Montreal (Canada)
INTRODUCTION
Electrical responses in specific sensory systems show wide variations due apparently to the modulating action of “extrasensory” or “unspecific” mechanisms. These variations, and their physiological and psychological significance, have been a major interest of our Honorary President, Professor Bremer (1960), and his associates during recent years, as well as that of many other participants in this Colloquium (HernandezPe6n et a!. 1956; Jouvet and Hernandez-Pe6n 1957; Dumont and Dell 1958; Bremer and Stoupel 1959; Mancia et at. 1959; Buser and Borenstein 1957; Dell 1960). In the present communication is presented a summary of some additional observations made with colleagues and co-workers bearing upon this subject. They pertain particularly to the physiological significance of changes in evoked potentials in relation to unitary activity occurring with “arousal” or alertness, habituation and attention, with considerations of their functional significance in behavoir. As a point of departure, a few observations reported briefly with Drs. Ricci and Doane at the Moscow Colloquium in 1958 (Jasper et al. 1960) will be reviewed to recall the consistent changes in occipital evoked potentials during motor responses conditioned to an intermittent light stimulus in the monkey. These changes have been further analyzed by Dr. Ricci in Rome, and in some work still in progress with Drs. Redding and Siegfried in Montreal, with the aid of microelectrodes. Two additional series of experiments will be summarized for their contribution to an understanding of the significance of changes in sensory evoked potentials in relation to attention and behavior. 1 . Studies of changes in acoustic evoked potentials from the primary auditory cortex in the cat as affected by electrical stimulation of the amygdaloid nucleus, carried out with Dr. Yamamoto (unpublished data). 2. Studies of the effects of distraction upon evoked potentials simultaneously recorded from visual and auditory systems in response to combined click and flash stimulation in the unanesthetized cat, carried out with Dr. John Jane and Dr. G. D. Smirnov (Jane, Smirnov and Jasper, in press).
ELECTRICAL RESPONSES IN SENSORY SYSTEMS
273
1. Decremental reduction in occipital evoked potentials during conditioned motor responses In the experiments with Ricci and Doane, previously reported in part, it was noted that the most consistent change in the electrical activity of the brain associated with a conditioned motor avoidance response in the monkey, was the reduction in surface cortical evoked responses to the conditioning stimulus (CS), namely intermittent brief flashes of light at about 5/sec. Such changes were even more consistently observed than were changes in surface electrical activity in the sensory or motor cortex corresponding to the conditioned movement (i.e., the arm area contralateral to the arm used in conditioned withdrawal). It was indeed unexpected to find that depression in the surface cortical evoked potentials from the primary receiving area of the conditioning stimulus, was so closely correlated with the conditioned motor response. The functional significance of this depression was obscure. It should be pointed out at the start that occipital evoked potentials were depressed not only during conditioned movements, but also during unconditioned motor responses, as shown in the first record of Fig. 1. In this example, taken from unpublished records of the experiments with Ricci and Doane (as for other tracings in this figure), the occipital evoked potentials (No. 1, third line) show some incremental change during the first three flashes of light; then, after a decrement, they become quite stable in amplitude until the electrical shock is delivered to the hand causing a rapid and strong withdrawal movement. During this movement it will be noted that the occipital evoked potentials are reduced in amplitude in a manner similar to that which occurred regularly in the conditioned withdrawal movements which were not preceded by the shock, as shown in the second tracing of Fig. 1 (No. 2, third line). In the third record of Fig. 1 is shown another example of initially increasing evoked potentials (“recruiting”) with a marked decrement just preceding the conditioned motor response in a fairly typical sequence. It is to be noted that both early and late components of the evoked potential complex are affected, though not equally. The upper line in each of the three tracings in Fig. 1 shows examples of changes in the firing pattern of single cells, recorded with microelectrodes, in the motor (1 and 2) and sensory (3) cortex for the arm involved in the movement. The second line in each sample is the surface electrical activity recorded with a gross electrode placed adjacent to the site of insertion of the microelectrode in motor and sensory cortices respectively. It is to be noted that the surface electrical activity from sensory and motor cortex shows no changes of significance during either the unconditioned or the conditioned motor responses, the animal being in a high degree of alertness prior to the occurrence of the conditioning stimulus. The microelectrode tracings shown in Fig. 1 give examples of unitary responses in motor and sensory cortex. The first tracing shows the type of cell in the motor cortex which characteristically responds only to the initiation of the series of conditioning stimuli, but does not participate in the motor movement. In the second example a cell from the motor cortex is shown to accelerate its rate of firing before, during and following the conditioned movement. In the third example, a cell from the sensory cortex fails to respond to the conditioning visual stimulus, but responds (apparently) References p . 285-286
274
H. H. JASPEK
CS
SR
Fig. 1 Changes in primary occipital surface evoked potentials in response to brief flashes of light repeated at S/sec used as a conditioning stimulus for a conditioned (shock) avoidance response in the monkey, according to the technique described by Jasper et al. (1960). In each sample record the upper line shows the activity of single neurones as recorded extracellularly with a tungsten microelectrode in the motor (1 and 2) and somato-sensory (3) cortices in the area contralateral to the upper limb representation which was conditioned to the withdrawal movement. A short time constant amplifier was used in this channel. The second line in each sample represents surface electrical activity recorded by means of a small silver ball electrode (about 0.5 mm in diameter) placed adjacent to the site of insertion of the microelectrode over motor (1 and 2 ) and sensory (3) arm areas. In the third line of each sample are shown the evoked potentials from the primary visual cortex recorded with a small silver ball electrode, with a reference electrode screwed into the skull.Downward deflections are surface positive. The monkey was only partially conditioned in the first sample (60 trials) and was making frequent errors which resulted in a shock and withdrawal movement (SR), the unconditioned response. Note that occipital evoked potentials remained quite constant up to the time of the shock (SR), and then they were reduced in amplitude during the unconditioned motor response. In the second example (2) after further conditioning (100 trials) the monkey was successfully avoiding the shock; the occipital evoked potentials were reduced in amplitude prior to the conditioned motor response (CR). I n the third sample (3), after further conditioning (145 trials) occipital evoked potentials showed an initial incremental change, augmenting with the first three flashes of light, then they decreased rapidly in anticipation of the CR. This decrement in visual evoked potentials occurred in anticipation of all conditioned motor responses, in these experiments. The EMG is superimposed upon the signal of motor responses in the lower line.
to the movement itself in a complex manner, first accelerating and then decreasing its rate of discharge. These complex changes in unit activity could not have been predicted from the record taken with a gross (about 0.5 mm) electrode placed on the cortical surface adjacent to the point of insertion of the microelectrode, although marked desynchronization of the surface tracing does occur, together with unit activation.
2. Microelectrode analysis of the SigniJicance of changes in surface occipital evoked potentials during conditioning Our analysis of unitary cellular discharge in the occipital cortex corresponding to the evoked potential changes during conditioning is not completed, due (among other
ELECTRICAL RESPONSES IN SENSORY SYSTEMS
275
things) to technical difficulties and the extraordinary complexity of the changes observed, some of which will require instrumental analysis before their true character can be adequately described or understood. Only a few examples in the form of a preliminary report can be given at the present time. In Fig. 2 is shown an example of a unit from the visual cortex which responded repetitively (3-4 times) to each flash of light at the beginning of the CS with initial suppression of interflash discharge, even when the surface evoked potential was first attenuated. Then the unit became dispersed in time, and did not follow so closely the light stimulus, during the conditioned response itself.
Fig. 2 Changes in pattern of unit response (upper line) recorded from a single neurone in the right occipital cortex of the monkey during a conditioned withdrawal movement of the left hand. This unit tended to grouped firing even before the light stimulus (CS). The stimulus produced repetitive responses to each flash of light which tended to become dispersed, with more discharge in the interflash interval, during the motor response (CR). The second line shows the surface electrical activity from the motor cortex, and the third line the surface electrical activity from the primary visual cortex, as in Fig. I . The EMG is superimposed upon the signal of motor responses (R) as in Fig. 1.
Another example shown in Fig. 3, from another series of experiments, shows a somewhat different unit behavior. This cell seemed to cease firing in response to the light stimulus during the conditioned response, but to fire spontaneously independent of the stimulus.
CR 15ec
Fig. 3 Example of a decrease in control of unit discharge in the visual cortex in response to the conditioning light stimulus during the conditioned motor response (CR) in the monkey. The second line in this tracing was the surface electrical activity from the parietal cortex, the third line the surface activity from the visual cortex. See text for further explanation.
A more complex change is shown in the occipital unit sampled in Fig. 4. There was a marked facilitation of this unit discharge in response to the CS just preceding the motor response, even with a decrease in the surface evoked response. This was followed by complete inhibition of unit activity during the actual motor response itself. This biReferencer P. 285-286
276
H. H. JASPER
-CR
0.5sec
Fig. 4 Inhibition or arrest of unit responses in visual cortex of the monkey during a CR as in previous figures. The animal was highly aroused during this trial, with absence of slow waves from the parietal cortex (2) and almost complete absence of surface evoked potentials from the visual cortex (3) just before, during, and just following the motor response (CR).
modal effect was not closely related to changes in surface evoked potentials. The firing rate of repetitive discharge was initially about 150/sec and was accelerated to about 250/sec just prior to the cessation of discharge during the motor response. In another example, shown in Fig. 5, there was apparent inhibition of unit discharge during the interval between flashes, giving a better grouping of discharges in relation to the flash, and less background “noise” between. This is reminiscent of the increase in signal-to-noise ratio described by Evarts (1960) as characteristic of the effects of general alertness upon the response of units in the visual cortex of the cat.
-CR
0 5 sec
Fig. 5 Inhibition of interstimulus firing of visual cortical units (1) in the monkey (experiments as above) just before and during a conditioned motor response (CR). This produced increased “contrast” between evoked and interstimulus “spontaneous” activity. The animal was quite alert during this trial.
Occasionally, during recent experiments, we have observed inhibition of occipital unit discharge during a motor response to be accompanied by rapid firing of a small adjacent cell picked up by the same extracellular microelectrode. An example is presented in Fig. 6, first during an unconditioned withdrawal response to an electric shock to the hand, and then, the same units recorded during a conditioned response to intermittent flashes of light at 5/sec. It is tempting to conclude that such small rapid unit discharges associated with the inhibition of larger units represent the firing of inhibitory interneurones of a type comparable to Renshaw cells in the cord. It may well point to collateral inhibition in the cortex for which there is a growing body of evidence. There can be no doubt that these preliminary studies of visual cortical unit discharges during motor responses confirm the conclusions reached from surface evoked
ELECTRICAL RESPONSES IN SENSORY SYSTEMS
277
SR
CR
0 5 sec
Fig. 6 Single units from visual cortex in the alert monkey firing spontaneously, not in response to light stimulus (S), with one unit inhibited and a small unit activated reciprocally during an unconditioned motor response to shock (SR) in upper record, and during conditioned response to the light (CR) in lower record. Lines 2 and 3 represent surface electrical activity from motor and visual cortices respectively.
potential studies that important modifications occur in the response of the cells of the sensory cortex when the animal performs a movement which has been conditioned to that stimulus, or even perhaps during unconditioned movements. One cannot escape the conclusion that motor activities may, therefore, modify sensory impressions. The nature of this modification is most complex, as judged by unitary studies. Both inhibitory and facilitatory processes are involved, even when surface evoked potentials are uniformly decreased in amplitude. Inhibition of interstimulus “spontaneous” activity, with or without concurrent facilitation of unit response to each stimulus, would result in increased contrast or “signal-to-noise ratio”. With further activation, however, some units begin firing more rapidly independent of the light stimulus. This would have the opposite effect, namely an increase in background “noise”. The non-visual activation of these units would occlude them by refractoriness to the incoming volleys of visual impulses. Modification of the incoming volleys themselves at the geniculate or retinal level has not been ruled out in these experiments, and probably does occur, as judged by other experimental studies. The high intensity of the light flashes used, together with controls showing similar effects after paralysis of the pupil with atropine, would eliminate the effects of the pupillary reflex in these results. 3. Unitary and surface evoked potential studies of the response of the auditory cortex in the cat as modified by electrical stimulation of the amygdaloid nucleus with hippocampal after-discharge The results of these studies, carried out with Dr. Shinjiro Yamamoto ofthe University of Kanazawa, Japan, will be summarized only briefly since they are to be published elsewhere in detail. It is well known that epileptiform seizure discharges arising in the amygdala, and conducted to the hippocampus and brain stem, often result in desynchronization of the electrical activity from the lateral surface of the temporal cortex at the onset of References P. 285-286
278
H. H . JASPER
the seizure. Similar results have been reproduced by electrical stimulation of the amygdala in man (Penfield and Jasper 1954). During this “flattening” of surface cortical electrical activity, the patient stares blankly and is usually unresponsive to commands. He is amnesic for the period of the attack. I n order to gain some insight into the mechanism of this amnesia and decreased responsiveness we have undertaken to study the effect of such minimal amygdaloid seizures upon the evoked electrical activity in the auditory cortex of the cat. Cats were mounted in a stereotaxic instrument and their temporal cortex exposed under ether anesthesia. Pressure points were injected with novocaine. They were then immobilized with flaxedil, or by spinal section at Cl (encdpphale isold) and maintained on artificial respiration. Small amounts of thiopental sodium (pentothal) were injected (4mg/kg) from time to time to prevent the generalization of epileptic seizures resulting from amygdaloid stimulation, and to avoid persistent desynchronization of the surface cortical electrical activity. Bipolar stimulating electrodes were placed in the amygdaloid nucleus, preferably in the lateral nuclear complex. Recording electrodes were placed in the hippocampus in order to record after-discharge conducted from the amygdala. Spontaneous surface electrical activity was recorded from the primary auditory cortex simultaneously. Glass micropipette electrodes were used for extracellular recording of single units within the auditory cortex adjacent to the surface recording electrodes. With careful adjustment of the stimulus strength it was possible to limit the afterdischarge to the hippocampus, so that it would not progress to involve the temporal cortex in seizure activity. It then produced only a desynchronization of spontaneous electrical rhythms (the “activation” effect) with marked suppression of surface cortical evoked potentials in response to click stimuli. This is illustrated in Fig. 7.
14 0 0 p V 15eC I
I
I
L
I
I
37- 1
Fig. 7 Inkwriting oscillograph tracings from the surface of the primary auditory cortex in the cat (bipolar silver ball electrodes about 1.5 mm separation) (Aud), and from a pair of needle electrodes inserted stereotaxically into the ventral hippocampus (Hip). Auditory (click) stimuli were administered a t 2/sec, then the amygdaloid nucleus was stimulated to produce a brief after-discharge detected in the hippocampus. Note the depression in auditory cortical evoked potentials during the amygdaloid stimulation and for about 10 sec after, about half of which time was not occupied by hippocampal after-discharge. Encihale is016 preparation with minimal dose of pentobarbital prior to stirnulation.
ELECTRICAL RESPONSES IN SENSORY SYSTEMS
279
Accompanying the consistent depression in surface evoked potentials in response to click stimuli, there were several forms of change in unitary cell discharge within the auditory cortex. Some units, which were responding regularly to the clicks prior to amygdaloid stimulation, showed marked activation and occlusion, failing to continue to respond to the clicks. In others, after a brief period of activation independent of the clicks, there would result a facilitated response to each click, as illustrated in Fig. 8, even though the evoked potential was reduced in amplitude. In others an initial period
175
10 m sec
205
40s
200 p v
Fig. 8 Oscilloscope sweeps of surface primary auditory cortical evoked responses to click stimuli in the enciphale isoli cat (upper lines) and extracellular unit responses recorded with a capillary pipette microelectrode inserted into the auditory cortex between the pair of surface electrodes. Before amygdaloid stimulation this unit was responding regularly to each click, usually with a single discharge as shown (BEF), and occasionally with a double discharge. This same unit was first activated by the amygdaloid stimulation to fire independently of the auditory stimulus (2S), and then showed a facilitated response to the click 5 sec later (5s). The surface evoked potential was depressed during this facilitated unit response. Facilitation of unit response continued for over 20 sec (20s) during which time the surface evoked potential recovered. Single unit responses returned after about 40 sec (40s). From unpublished experiments with S. Yamamoto.
of inhibition would be followed by facilitation, as shown in Fig. 9. Statistically a facilitatory effect upon unit responses was more common than an inhibitory effect, though frequently, facilitation became supraliminal and resulted in excessive discharge with occlusion of apparent responsiveness to the auditory stimuli. The similarity of these results to those obtained from the visual cortex during “activation” of the animal to conditioned avoidance movements is striking, and may illustrate a general principle involving a bimodal effect of activation depending upon the prior state of the cortex, and upon the degree of activation. 4. Changes in geniculate and cortical evoked potentials in response to combined visual
and auditory stimulation during visual or auditory distraction (experiments by J. Jane and G . Smirnov) Small stainless steel recording electrodes, insulated except the tip which was 0.1 mm in diameter, were placed on the primary visual and auditory cortex, and inserted stereotaxically into the medial and lateral geniculate bodies in six cats under nembutal References p . 285-286
280
H. H . JASPER
Fig. 9
Experimental set-up the same as in Fig. 8 in another cat showing depression in surface auditory evoked potentials following amygdaloid stimulation (4s)as compared to regular responses obtained before (BEF), accompanied by inhibition of both early and late unit responses. Early unit responses returned first (in 12 sec, 12s) and late responses returned in 20 sec (20s). Surface evoked potentials did not recover their control value until 40 sec (40s) after the end of amygdaloid stimulation. From unpublished experiments with S. Yamamoto.
anesthesia. About one week after recovery from the operation evoked potentials were recorded simultaneously from the geniculate bodies and visual and auditory cortices in response to combined and synchronized visual and auditory stimuli, repeated regularly at 1jsec. The visual stimulus consisted of a very brief and intense stroboscopic flash of white light directed into a cage with reflecting walls so that an intense stimulus would occur regardless of the direction of the cat’s gaze. Responses were obtained of almost the same magnitude with this light with the eyes closed or opened, and with the direction of gaze to any side of the cage. Controls with the pupils paralyzed with atropine were also carried out in some experiments. The auditory stimulus was a loud click (estimated at about 80-100 db) generated by a square wave output of the Grass Stimulator connected to a loudspeaker adjacent to the cage containing the cat. The flash and click were triggered by the same frequency generator, the click being delayed 10-20 msec following the flash. After the animals had become accustomed to the situation, and somewhat habituated to the test stimuli, recording of evoked potentials was carried out with an Offner Type T inkwriting apparatus, with controls taken on a multichannel oscilloscope either with single or superimposed sweeps. The combined stimuli were not administered continually. A series of 50 stimuli was administered during a control period without distraction, then, without interruption, distraction was introduced during the second series of 50 stimuli, and the test stimulation was then continued for another 50 stimuli. The total duration of the stimulus period was, therefore, 150 sec. After a period of rest, lasting several minutes, the entire series was repeated using a different form of distraction. Distracting stimuli consisted of introducing a live rat into the cage with the cat
ELECTRICAL RESPONSES IN SENSORY SYSTEMS
28 1
(visual)* and an intermittent squeaking noise (auditory), simulating that made by the rat, or actually made by the rat when its tail was pinched. The sound was delivered from a magnetic tape recorder. The object of these experiments was to determine if the different forms of distraction, visual or auditory, would have a selective action upon visual or auditory evoked potentials from either geniculate or cortical levels, when recorded simultaneously. The overall amplitude of evoked potentials was measured from the ink-writer tracing as obtained before, during and after auditory or visual distraction. In some experiments photographs were taken of superimposed sweeps of an oscilloscope before, during and after the distraction. Without consideration of detailed results to be published elsewhere, the principal conclusions from these experiments can be summarized as follows : 1. In three of the animals, without prolonged prior habituation to the test stimuli, evoked potentials were extremely variable to the series of test stimuli without distraction, but became much more constant during either visual or auditory distraction. Effects were similar on both auditory and visual systems, without obvious differential action depending upon the form of the distraction. Similar regularization occurred at geniculate and cortical levels. 2. With prolonged stimulation the evoked potentials at both geniculate and cortical levels were reduced in amplitude. They were then increased in both amplitude and regularity by either visual or auditory distraction, as shown in Fig. 10, for auditory and visual cortical evoked potentials during visual distraction. 3. Animals showing a higher degree of alertness prior to distraction, witha relatively desynchronized resting EEG, showed a marked reduction in evoked potential amplitude during either visual or auditory distraction, as shown in Fig. 11. 4. In superimposed oscilloscope tracings there was observed a decrease in both temporal and amplitude dispersion of evoked potentials during distraction in some animals. In others, those with reduced amplitude during distraction, there was increased temporal dispersion as well as depression in amplitude. Joy 1 AC
-
Vlsual R 1oopv Fig. 10 Inkwriting oscillograph (Offner Type T) records from implanted (stainless steel bipolar) electrodes over the primary auditory (AC) and visual (VC) cortices in the awake freely moving cat in response to combined visual (flash) and auditory (click) stimuli (the flash preceding the click by 10 msec) repeated at I/sec. Responses had become habituated by several hundred repetitions over a period of 3 days. They recovered promptly during visual distraction caused by introducing a rat (enclosed in a plastic box) into the cage with the cat. Controls with pupillary paralysis with atropine have shown that this effect was not due to pupillary dilatation (from Jane, Smirnov and Jasper 1962).
15ec
*
The rat was enclosed in a plastic transparent box in order to eliminate olfactory stimulation.
References P. 285-286
H. H. JASPER
282 v15 Cortex
SY- R
Audltory R
I 100 pv
visual R 15eC
Fig. 11 Surface (bipolar) evoked potentials from implanted electrodes in the primary visual cortex of the freely moving cat (inkwriter records as in Fig. 10). The animal was quite alert before distraction was introduced. Combined auditory and visual st imuli were being administered at 8/sec. Evoked potentials were diminished during either auditory or visual distraction. From Jane, Smirnov and Jasper (1962) as described in text.
Some of the quantitative aspects of these results in two animals, are shown in the following Table (Fig. 12) giving the mean, the range, and the standard deviation or measure of variance in overall amplitude of evoked responses to test stimuli before, during and following either auditory or visual distraction. The mean amplitude was significantly increased together with a marked decrease in variability (Sd) in these animals during either visual or auditory distraction. There was no significant difference in visual and auditory systems depending upon the form of distraction. It will be noted also from this Table that the changes induced in evoked potentials from the geniculate bodies was equally marked, if not greater, than that induced in evoked potentials from
(x)
-'
MG
153
70-240 35
173
130-210 16
153
70-240
LG
147
7%50
178
150-225 2 0
180
WO-250 35
167
12o-Mo 18
175
125-225 24
90
75-117
45
35
MG
68
20-210
44
LG
M
25-xx)
40
AC
72
0-165
39
0-257
62
119
57-157
20
107
0-186
47
71
0-133
29
98
75-158
16
97
0-175
31
b7
0-157
48
124
%-ZOO
21
131
70-200 29
7 .-('
vDIAC LVCl
I
I
I
16
I
172
110-240 29
58
125-275 43
78
0-125
I
33
1
Fig. 12 Table of over: I peak to peak amplitudes, in pV, of evoked potentials recordeL - y means o an inkwriting oscillograph a s i n Fig. 10 and 11, from bipolar implanted electrodes in the freely moving waking cat. Records from two animals showing increase in auditory and visual evoked potentials at both geniculate (MG and LG) and cortical level (AC and VC) are shown. The mean amplitudes, were obtained from 50 successive responses to combined auditory and visual stimuli a t l/sec, before, during, and after auditory (AD) or visual (VD) distraction as described in text. The range of the measurements (r) and a measure of variability, standard deviation (Sd) is also given. Note particularly the decrease in variability (Sd) during distraction in all instances.
x,
ELECTRICAL RESPONSES IN SENSORY SYSTEMS
283
the auditory and visual cortices, regardless of whether the distraction was visual or auditory. It should be noted that the range of evoked potential amplitudes before distraction overlapped considerably the mean value during distraction, i.e., there were some responses before distraction which exceeded the mean of those recorded during distraction, the most striking change being the decrease in variance at a relatively high mean response level. DISCUSSION
The foregoing brief summary of some additional observations concerning the intimate mechanisms regulating the electrical responses of sensory systems in relation to alerting or attentive responses of animals, and in relation to conditioned or unconditioned motor activities, serves to emphasize the inadequacy of simple formulations of these processes in terms of “activation”, “facilitation”, “arousal”, “excitation”, or “inhibition”. In the Moscow Colloquium we pointed out that “activation” of a monkey to perform a simple conditioned motor response was associated with complex changes in the firing patterns of neurones in frontal, motor, sensory, and parietal cortex. The number of units showing increased firing was statistically greater in sensory and motor cortices, but inhibition of unit discharge was predominant in frontal and parietal cortex. Changes in temporal and spatial pattern of firing of individual units, with inhibition of some units simultaneously with facilitation of others, seemed of greater significance than was any process which could be adequately described in general terms for the population of neurones as a whole. Prof. Jung and his co-workers have shown, as reported in the present colloquium, that inhibitory effects upon cortical neurones are more common than excitatory effects during “activation” of the animal by sensory stimulation. Likewise Evarts (1960) has shown that more units in the visual cortex are “inhibited” than excited when an animal is aroused from natural sleep. He pointed out, however, that inhibition of “spontaneous” activity lowered the background “noise” level during arousal so that unitary responses to visual stimulation stood out more clearly, i.e., there was an increase in signal-to-noise ratio or improved contrast. Huttenlocher has recently shown that spontaneously firing units recorded from the midbrain reticular formation were more commonly inhibited during arousal of the cats from natural (slow wave) sleep if their discharge rate was low, while units showing a relatively high rate of discharge during sleep were activated to a further increase in rate during waking. During sleep with low voltage fast cortical electrical activity (Dement 1958; Jouvet et al. 1959), all units sampled showed a marked increase in spontaneous discharge rate (more than double). It is known that during this form of sleep the animals are less easily aroused than during “slow wave” sleep, proving that the form of deep sleep with a desynchronized EEG is an active process in many cells of the mid-brain reticular formation, though these cells may not belong to “the ascending activating system”. Unit evoked responses to acoustic stimuli (clicks) were reduced in both types of sleep. This reduction in evoked response tended to be associated with an increase in spontaneous unit activity. Huttenlocher concluded that the reduction both References p . 285-286
284
H. H. JASPER
in unit responses and in evoked potentials during sleep was probably due to occlusion by increased “spontaneous” activity during sleep. Unitary analysis (carried out with Redding and Siegfried, unpublished) of the significance of the consistent changes in occipital surface evoked potentials, i.e., the incremental initial change followed by a marked decrement in evoked potentials just preceding a conditioned motor response (Jasper et al. l960), has shown that four unitary processes may underlie these changes in gross surface potential waves. There may be: 1. facilitation of unit response to individual flashes of the conditioning stimulus; 2. inhibition of spontaneous activity during the interflash interval (increased contrast) ; 3. inhibition of both unit responses and spontaneous interflash discharge; 4. activation of units to greater spontaneous activity blurring or occluding their response to individual stimuli (decreased contrast). The latter effect was more common in the extreme forms of “activation” such as that produced by stimulation of the amygdala to minimal seizure discharge in the experiments with Yamamoto. The increased regularity of relatively high amplitude evoked potentials, with decrease in variance and “sharpening” of form with attention, in the experiments with Jane and Smirnov could be explained by inhibition of spontaneous activity while at the same time facilitating specific responses to incoming sensory volleys. A further increase in nonspecific activation might then cause supraliminal facilitation, activating cells and resulting in diminished specific sensory response due to occlusion. It will be recalled that, in the experiments of Dumont and Dell (1958), the marked facilitation of visual cortical evoked potentials in response to direct electrical stimulation of the optic nerve during arousal did not occur if the animal showed an “aroused” EEG pattern before, and even some decrease in evoked potentials would then occur with further “arousal”. It is well known that decrease in amplitude of evoked potentials with increased alertness in the waking animal is more commonly observed under the usual experimental conditions and before habituation. Nonspecific activation appears therefore, to have a bimodal effect upon specific cortical sensory responses, first increasing their effectiveness by a combination of inhibited spontaneous activity and facilitation of specific response apparently due to an increase in the subliminal excitatory fringe. This would greatly enhance sensory contrast. With further increase in nonspecific activation, specific responses become blurred by occlusion, decreasing sensory contrast. However, it cannot be assumed, in all instances, that these changes in unitary activity will be faithfully reflected in alterations in surface evoked potentials, for reduction in evoked potentials has been observed not uncommonly together with inhibition of spontaneous activity and with facilitation or more concentrated grouping of individual unit responses to a given stimulus in individual cells. Such complex effects of unspecific activation can be understood if we abandon the hypothesis that “activation” is a relatively undifferentiated, unstructured, nondirectional, or unspecific general energizing process, turning up the “volume control” so to speak. If it is conceived of as a highly integrated process of control or “setting” of neuronal systems for coordinated functions, inhibition playing at least as important a role as facilitation, then the complex changes in unitary behavior would be expected.
ELECTRICAL RESPONSES IN SENSORY SYSTEMS
285
The integrative functions might, however, be lost or diminished during excessive massive supraliminal activation, as in states of extreme emotional stress. Studies of the efficiency of motor coordinations during the performance of complex tasks in animals and in human subjects under different amounts of “motivation”, tension, or general activation has led Malmo to propose a bimodal effect of activation upon the efficiency of coordinated visuo-motor tasks (Malmo 1959; Stennett 1957). Speed and accuracy increase steadily with increased attention and motivation up to a peak, and then decline with further increase in motivation or “tension” beyond a certain optimum level for a given individual. Such behavioral studies would be consistent with the interpretation given above to the changes observed in electrical activity in specific sensory systems at different levels of “unspecific” activation. Activation within the range of increased efficiency can hardly be called unspecific, however, since it is probably a highly integrated process directed toward a definite structured objective in behavior. Only in the extreme state which might be exemplified by a stereotyped startle response, or the frozen immobility of extreme panic could activation be considered truly “unspecific”. For this reason it is probably futile to expect to be able to detect electrophysiological correlates of attention or goal directed activation of behavior by studies of evoked potentials from large populations of neurones. Within the usual ranges of sensory stimulation attention is governed by the quality, novelty, and significance of stimuli, rather than by their intensity. This requires a selective facilitation of highly complicated temporospatial patterns of impulses which one could not expect to be revealed by the gross evoked potential technique. New electrophysiological methods and theories will be needed to make a more realistic attack on this problem. It would be naive indeed to assume that the brain is less complicated, considering the nature of the functions it performs. Adequate utilization of information theory and similar mathematical formulations may lead the way to a more rational view of cerebral mechanisms of sensory perception as affected by level of vigilance, and selective attention (Cherry 1956; Broadbent 1958). When the activity of single cells is recorded with microelectrodes one does not find separate and distinct areas or neuronal systems devoted solely to “excitation” or “inhibition”. One finds excitatory-inhibitory patterns of activity in all systems, specific and unspecific. Also one finds interactions between systems of neurones and convergence and divergence of impulses within and without specific sensory pathways throughout the brain in a highly organized fashion. Integrative mechanisms do not seem to be confined to the more or less “unspecific” neuronal systems, though it is in the interneuronal or interconnecting networks that one would expect to find the highest level integrations ;these may not be in anatomically distinct areas of the brain - cortical or subcortical. REFERENCES BREMER, F. Analyse des processus corticaux de l’eveil. Electroenceph. clin. Neurophysiol., 1960, SUPPI.13: 125-136. BREMER, F. et STOUPEL,N. Facilitation et inhibition des potentiels evoques corticaux dans l’eveil cerebral. Arch. int. Physiol., 1959, 67: 240-275.
286
H . H. JASPER
BROADBENT, D. E. Perception and communication. Pergamon Press, London, 1958,338 pp. BUSER, P. et BORENSTEIN, P. Suppression elective de reponses “associatives” par stimulation reticulaire chez le chat sous anesthesie profonde au chloralose. J. Physiol. (Paris), 1957, 49: 86-89. CHERRY, C. (Editor). Third London Symposiuni on Information Theory. Butterworths, London, 1956. DELL,P. Discussion in H. H. JASPERand G. D. SMIRNOV (Editors), The Moscow colloquium on electroencephalography of higher nervous activity. Electroenceph. elin. Neurophysiol., 1960, Suppl. 13: 134. DEMENT, W. The occurrence of low voltage fast EEG patterns during behavioural sleep in the cat. Electroenceph. clin. Neurophysiol., 1958, 10: 291-296. DUMONT, S. et DELL,P. Facilitations specifiques et non-spkcifiques des reponses visuelles corticales. J. Physiol. (Paris), 1958, 50: 261-264. EVARTS, E. V. Effects of sleep and waking on spontaneous and evoked discharge of single units in visual cortex. Fed. Proc., 1960, 19: 808-837. GAUTHIER, C., PARMA,M. and ZANCHETTI, A. Effect of electrocortical arousal upon development and configuration of specific evoked potentials. Electroenceph. din. Neurophysiol., 1956,8 :237-243. GELLHORN, E., KOELLA, W. P. and BALLIN,H. M. lnteraction on cerebral cortex of acoustic or optic with nociceptive impulses: The problem of consciousness. J. Neurophysiol., 1954, 17: 14-21. HAGBARTH, K. E. and FEX,J. Centrifugal influences on single unit activity in spinal sensory paths. J. Neurophysiol., 1959,22: 321-338. HERNANDEZ-PEON, R., SCHERRER, H. and VELASCO, M. Central influences on afferent conduction in the somatic and visual pathways. Acta neurol. 1at.-amer., 1956, 2 : 8-22. HORN,G. Electrical activity of the cerebral cortex of the unanaesthetized cat during attentive behaviour. Brain, 1960, 83: 57-76. HUTTENLOCHER, P. R. Evoked and spontaneous activity in single units of medial brain stem during natural sleep and waking. J. Neurophysiol., 1961, 24: 451468. JASPER,H., RICCI, G. and DOANE,B. Microelectrode analysis of cortical cell discharge during avoidance conditioning in the monkey. Electroenceph. clin. Neurophysiol., 1960, Suppl. 13: 137-1 55. JOUVET,M. et HERNANDEZ-PE~N, R. Mecanismes neurophysiologiques concernant I’habituation, I’attention et le conditionnement. Electroenceph. clin. Neurophysiol., 1957, Suppl. 6: 39-49. JOUVET, M., MICHEL,F. et COURJON, J. Sur un stade d’activitk electrique cerebrale rapide au coiirs du sommeil physiologique. C. R. SOC.Biol. (Paris), 1959, 153: 1024-1028. MALMO, R. Activation: A neuropsychological dimension. Psychol. Rev., 1959, 66: 367-386. MANCLA, M., MEULDERS, M. and SANTIBAREZ, H. G. Changes of photically evoked potentials in the visual pathway of the “cerveau isol6” cat. Arch. ital. Biol., 1959,97: 378-398. W. G. and JASPER, H. H. Epilepsy and the functional anatomy of the human brain. Little, PENFIELD, Brown and Co., Boston, Mass., 1954, 896 p. STENNETT, R. C. The relationship of performance level to level of arousal. J. exp. Psychol., 1957, 54: 54-61.
DISCUSSION G. F. RICCI: Prof. Jasper’s very interesting paper has given us a summary of the observations he has been making during recent years on evoked responses during attention and conditioning. After leaving his Department at the Montreal Neurological Institute 1 have continued to work on similar problems in association with Dr. F. Valassi, using a different approach, and I would like to show a few slides and comment on some points. In our experiments*, monkeys were conditioned to avoid a painful stimulus applied to the hand upon presentation of either a continuous light or a sound (CS). A single electric shock to the thala-
* The research reported in this document has been sponsored in part by The Air Force Office of Scientific Research OAR, through the European Office, Aerospace Research, United States Air Force.
287
DISCUSSION
A
mec IS
35
70
MO
250
(SO
850
1000
1500
25w
. .
larn=
IU101080
~
0
0420-200
190-1w
80-10
CONDITIONED RESPONSE
t .
A
,
."
10-80
100-190
300-390
800-870
1120-1550
Fig. 1 Facilitation of the responses evoked in the visual cortex by electric stimuli to the lateral geniculate body during conditioning (CS = Sound). In this, as in the following figures, each evoked response was elicited during one conditioned trial. In A , groups of responses were obtained by superimposing all evoked responses of an experiment according to the various delays after the onset of CS. In B, groups of responses were obtained by re-arranging and superimposing the responses in A that fell at approximately the same time intervals before or after the onset of the conditioned response, marked by the arrow in the figure. Bipolar surface to white matter recordings. (From Ricci and Valassi, in preparation).
mus or a single sensory stimulus were given at various moments of the conditioned trial, so that cortical evoked responses could be obtained throughout it. By this method, modifications of the responses at various moments of the conditioned trial and their relations to the conditioned responses could be tested. The results can be summarized as follows: (1) In the visual cortex, a facilitation of the responses to electric shocks to the lateral geniculate body, beginning 100 to 200 msec after the onset of the CS was consistently found (Fig. 1A). In some experiments, evoked responses were enhanced throughout the conditioned trial; in the majority of the experiments, however, the duration of the facilitatory period was found to last 200 to 500 msec. The relations between facilitation of the cortical evoked responses in the visual area and the conditioned response (CR) are shown in Figs. 1B and 2: the enhancement of amplitude of the responses constantly precedes the onset of the conditioned movement; in most experiments the evoked responses had returned to control amplitude when the CR began.
288
H. H. JASPER
Facilitation of the responses evoked in the visual cortex does not depend on the type of conditioning stimulus used, since it was obtained when the monkeys were responding to either light or sound (Figs. 1,2). It is also independent of potentiating effects of the light at the retinal level, such as those described by Chang (1952), since during the experiments in which sound was used the animals were kept in the dark. In the visual cortex, facilitation of the responses to subcortical electric stimuli has been shown to accompany the onset of EEG arousal (Dumont and Dell 1958, 1960; Bremer and Stoupel 1958, 1959 a,b; Bremer 1960). From the results of our experiments it may therefore be assumed that EEG arousal caused by the conditioned stimulus in the visual cortex precedes the onset of the conditioned response. (2) In the somatic area, responses evoked either by single electric stimuli to the thalamus or by shocks to the hand were not modified before the onset of the conditioned response. When the conditioned movement begins, responses to thalamic stimuli in the somatic cortex contralateral to the hand which performs it decrease in amplitude (Fig. 3), while responses to shocks t o the hand disappear entirely (Fig. 4). In the ipsilateral somatic cortex, on the contrary, responses to thalamic stimuli are not modified throughout the conditioned trial (Fig. 5 ) . Similar results have been obtained also with responses recorded from the motor area. In the sensory motor cortex, electrocortical arousal has been shown to be accompanied by a decrease of responses to peripheral stimuli (Gauthier, Parma and Zanchetti 1956). Although to a lesser degree, the primary responses evoked by electric stimulation of the thalamus are also decreased during EEG arousal (see controls in Fig. 3). Since during conditioning a similar modification of the evoked responses was found to begin only at the start of the conditioned response, it seems as though these cortical areas “wake up” only at the moment of the movement. In conclusion, our results have shown that during conditioned behaviour activating impulses arrive at different times at different cortical districts. These impulses come from a common subcortical structure (such as the reticular formation of the brain), whose function is probably that of activating mechanisms responsible for the integration of sensory information with past experience and mechanisms responsible for the initiation of the motor response. As a working hypothesis, it is possible to think that sensory integration begins at the onset of the cortical facilitation observed in the visual cortex and that it takes place during the interval occurring between the onset of the cortical facilitation and the beginning of the conditioned response. Local cortical mechanisms which are activated during this interval are probably important in reinforcing the activity of the reticular formation, so that excitation which builds up in it finally reaches a level sufficient t o cause activation of the cortico-spinal neurons of the motor area, which in turn causes the initiation of the conditioned motor activity.
F. Analyse des processus corticaux de l’eveil. Moscow Colloquium on EEC of higher nervous BREMER, activity, 1958. Electroenceph. clin. Neurophysiol., 1960, Suppl. 13: 125-1 36. BREMER, F. et STOUPEL,N. De la modification des reponses sensorielles corticales dans l’eveil reticulaire. Acta neurol. belg., 1958,58: 401403. BREMER, F. et STOUPEL,N. Facilitation et inhibition des potentiels evoquks corticaux dans I’eveil cerebrale. Arch. int. Physiol., 1959a,67: 240-275. BREMER, F. et STOUPEL,N. Etude pharmacologique de la facilitation des reponses corticales dans l’eveil reticulaire. Arch. int. Pharmacodyn., 1959b, 122: 234-248. CHANG, H. T. Cortical response to geniculate stimulation and the potentiation thereof by continuous illumination of the retina. J. Neurophysiol., 1952,15: 5-26. DUMONT, S. et DELL,P. Facilitations specifiques et non-specifiques des reponses visuelles corticales. J . Physiol. (Paris), 1958, 50: 261-264. DUMONT, S. et DELL,P. Facilitation reticulaire des mecanismes visuels corticaux. Electroenceph. clin. Neurophysiol., 1960, 12: 169-196. GAUTHIER, C., PARMA, M . and ZANCHETTI, A. Effect of electrocortical arousal upon development and configuration of specific evoked potentials. Electroenceph. clin. Neurophysiol., 1956, 8 : 231-243. W. GREYWALTER: The effect of temporal lobe activation, either in clinical seizures or following electrical stimulation may have profound local and systemic effects and I think these should be taken into account when interpreting their action on evoked responses. We have been able to record both local intracerebral
289
DISCUSSION
rnspc 700-650 550-420 380-300 250-200
CONOlTlONEO RESPONSE
100:120
40-20
CONTROLS
10-50
90-110
&k
.... 120OPV
A
10 rnsec
8
Fig. 2 Facilitation of visual evoked responses to lateral geniculate stimulation during conditioning (CS= Light). Evoked responses to thalaniic stimulation are grouped according to the time intervals before or after the onset of the conditioned response. Controls in the lower part of the figure were obtained during periods of EEG slow activity (A) and during spontaneous EEG fast activity (B). Bipolar surface to white matter recordings (From Ricci and Valassi, in preparation). CON DlTlONED RESPONSE
msec 260-110
90-10
25-75
300-320
675-750
1500-1600
.I . . 200 p v
CONTROLS
10 msec
A
B
Fig. 3 Modifications of responses evoked in the sensorimotor cortex by stimulation of the thalamus (N. ventralis postero-lateralis), during conditioning. Evoked responses are grouped as in Figs. 1B and 2. Prior to the onset of the conditioned response no enhancement of amplitude of the evoked response is seen similar to that of the visual cortex. Amplitude of the evoked responses decreases during the performance of the conditioned response. Controls in the lower part of the figure were obtained during EEG slow activity (A), and during spontaneous periods of EEG fast activity (B). (From Ricci and Valassi, in preparation).
290
H. H . JASPER
mssc beforv 160-120
LO-20
m58c after
0 -20
110-190
Conlrols
LOO.LB0
1150-1250
1500.1820
Fig. 4 Disappearance of responses evoked in the somatic cortex by single shocks to the hand during performance of the conditioned response. Evoked responses are grouped as in Figs. 1B, 2 and 3. Monopolar recordings (From Ricci and Valassi, in preparation). A
msec 640-590 ~~
‘
190-100
~~~
fr“ tflprd CONOITIONEO RESPONSE
0
580-500
160-100
20-80
220-260
P. 120?y
.*
10 mrrc
550-600
B lZOOpV
..**
10 mrrc
Fig. 5 Specificity of modifications of responses to thalamic (N. ventralis postero-lateralis) stimuli in the sensorimotor cortex during conditioning in relation to the hand performing the conditioned response. In A , responses in the left somatic cortex are not modified when the monkey is responding with the left hand (ipsilateral). In B, when the animal responds with the right hand (contralateral), responses evoked after the onset of the conditioned response are reduced. Bipolar surface to white matter recordings. (From Ricci and Valassi, in preparation). EEG and oxygen availability changes in temporal lobe seizures and in one caseparticularly the most consistent effect was a spectacular tachycardia followed by bradycardia and violent fluctuations in local 0 2 during the phase of hippocampal theta suppression. The attack involved limb extension, unresponsiveness and dreaming, and the evoked responses were considerably modified during the ictal phase. How these cardiac or vascular changes are related to the seizure process and disturbances of sensorimotor integration is still obscure, but I think they should be borne in mind. Next, I wonder whether visual impressions of evoked response changes against a varying background of intrinsic activity should not be supplemented by an averaging device. Transient events are bound to look less impressive against a “synchronized” background than a desynchronized one and their amplitude
DISCUSSION
29 1
is bound to vary more - in appearance - when they are masked by hippocampal or other rhythms. Last: does Dr. Jasper think that the motivational-activational optimum effect might be related to the autonomic accompaniments of learning which, as I mentioned yesterday, seem essential for the initiation of adaptive change but, in excess, may impede its consummation?
F. BREMER: In the well-documentedpresentation we have heard, one is faced with the necessity of making a choice. Our colleague Jasper has made an attempt, by his microphysiological observations, to verify the hypothesis of occlusion, put forward by us to explain the effect of depression that reticular arousal exercises on cortical evoked potentials. The difficulty of verification lies in the probability that the occlusive convergence between non-specific impulses of arousal and specific impulses has already taken place in cortical interneurons. If this hypothesis is correct, the occlusive convergence may be manifested by transient suppression (not by activation) of the discharge of the unit of which the potential is registered, while at the same time the interneuron discharge is being recorded. One of the tracings presented could be interpreted on the basis of this hypothesis. It shows the disappearance of the discharge of a high-voltage unit coincident with the appearance of the rapid discharge of a very low-voltage unit. But such observations do not permit the exclusion of a hypothesis of true presynaptic or postsynaptic inhibition. The significant fact is that stimulation of the reticular arousal system has a dynamogenic effect -sometimes spectacular - on the cortical evoked potential every time the afferentvolleywhich gives rise to this potential, is very synchronized. In the conditions of our experiments with Stoupel, and those of Madame Dumont and Dell, this synchronization was effected by application of a test stimulus to the visual pathway or to a thalamic relay nucleus. Again, by an effect of synchronization of the sensory influx, we can explain the fact demonstrated by Steriade and Demetrescu and confirmed by us (as you see from the tracing), that the responses of the visual area to light stimuli are regularly facilitated when the frequency of intermittent stimulation exceeds 6 flashes/second ; at lower frequencies the depression phenomenon (“masking”) occurs; synchronization of the impulses in the corticopetal volleys is clearly demonstrated by the brevity and the greater amplitude of the evoked potentials of the geniculate body at flash frequencies exceeding 6/sec. Reticular stimulation further increases synchronization at the level of the relay nucleus. It seems to me that the strongest argument in favour of the explanation of the phenomenon of masking by occlusion at the cortical level, is the fact that masking is also clearly seen following an in.izction of physostigmine - which activates the electrocortical tracing without arousing the animal. Concerning another aspect of this subject, we were disappointed to see that intensification of attention, either due to light or to auditory stimulation, was not manifested in the experiments of Jasper and his colleagues by differential modification of sensory evoked potentials simultaneously registered. If I have observed correctly, the two potentials were amplified. Such a result is puzzling because, in the experiments of Hernandez-Peon (1961), sensory distraction (negative aspect of attention) generally had an effect of depression rather than facilitation on the response to the stimulation suffering the distraction because of the attention to another stimulus or to a more “interesting” stimulus of the same sensory modality. BREMER, F. et STOUPEL, N. De la modification des reponses sensorielles corticales dans I’eveil reticulaire. Acra neurol. belg., 1958, 58: 401-403 ; Facilitation et inhibition des potentiels evoques corticaux dans I’eveil cCrebral. Arch. int. Physiol., 1959, 67: 240-275. DUMONT, S. et DELL,P. Facilitations specifiques et non-sptcifiques des reponses visuelles corticales. J . Physiol. (Paris), 1958, 50: 261-264; Facilitation reticulaire des mecanismes visuels corticaux. Electroenceph. elin. Neurophysiol., 1960, 12: 769-796. STERIADE, M. et DEMETRESCU, M. PhCnornenes de dynamogenksereticulaire aux divers niveaux des voies optiques pendant la stimulation lumineuse intermittente. J. Physiol. (Paris), 1960, 5 2 : 224-225. HERNANDEZ-PE~N, R. Reticular mechanisms of sensory control. In: W. A. ROSENBLITH (Editor), Sensory Communicat;on. M.I.T. Press, John Wiley, New York, 1961: 497-520. W. R. ADEY: My question to Dr. Jasper concerns the general thesis of using evoked responses to a train of brief stimuli repeated at regular intervals as an index of the functional changes in cerebral system in the
H. H. JASPER
course of learning, and in the performance of a learned response. It would seem t o me that, even when these brief recurrent stimuli are given in the sensory modality essential to the conditioned response, they are no more than intercurrent events in the influx of information on which the learned performance will be based. The transactional mechanisms on which the performance will be based are essentially continuous, and priorities would seem to attach to the handling of information directly concerned with, and essential to, the conditioned response. This information also arrives continuously in cerebral sensory system, and appears fundamentally different in most experimental designs from the addition of recurrent clicks or flashes, which are intercurrent transient sensory phenomena in the learned performance. It would seem, then, that the most that we might expect from such observations of trains of evoked potentials would be evidence of modulation of these evoked responses by the true transactional mechanisms occurring simultaneously and continuously in cerebral sensory systems. For these reasons, we have examined aspects of possible informational content of continuing EEG wave processes in learned performances in our own studies. A most challenging facet of this problem comes from the work of Griffin in the bat, where the frequency modulated trains of supersonic pulses emitted for navigational purposes remain effective in the face of broad-band white noise 20 decibels louder than the bats own cries. Can Dr. Jasper indicate ways in which he considers the evoked potentials in his studies may or may not relate to the transactional mechanisms in cerebral sensory systems? GRIFFIN, D. R. Listening in the Dark: the Acoustic Orientation of Bats and Men. Yale University Press, New Haven, 1958: 413 pp. A. FESSARD:
To record spikes from single units in the brain of an awake animal assuredly represents a remarkable achievement which we owe for the first time t o Dr. Jasper and his collaborators. My question is: why have you limited these records to those from a specific area, either motor or somato-sensory? A priori, it seems to me that places where sensory messages of different origins meet and interact - that is associative zones or nuclei - should give even more significant correlations with the effects of conditioning procedures.
H. H. JASPER’S replies: To G . F. Ricci 1 am very pleased to see the elegant results that Dr. Ricci and his coworkers have obtained in their studies of evoked potential changes in conditioning. Interpretation of these changes is rather difficult, however, if one uses only the simple conceptions of “arousal”, or perhaps I should say that this makes interpretation too easy! Dr. Ricci concludes that activating impulses come from “a common subcortical structure”, the reticular system, even though he has shown that they arrive a t different times at different local cortical areas. In well-established conditioned responses of a simple character the sensorimotor connections may become so stereotyped that the RS may not be involved to any significant extent. We have shown that habituation does occur with prolonged conditioning. We must not overlook the fact that local cortical activation can arise from specific thalamo-cortical or transcortical systems of neurones. To W. Grey Walter Dr. Grey Walter has brought up the question of autonomic responses to amygdaloid stimulation as being possibly related to changes in cortical evoked potentials in the experiments with Yamamoto. This is an important point, and one which concerned us very much during these experiments since changes in respiration and in blood pressure are not uncommonly found to follow stimulation of the amygdala. We were able to show, however, that the changes in auditory evoked potentials shown with minimal after-discharge in the hippocampus were not secondary to changes in blood pressure or respiration. I do not think that an averaging device would be useful in such experiments since individual responses are clearly seen in the oscilloscope tracings and should be studied individually, not only as an average of many, especially when they show continuous changes which would be masked in the average record.
DISCUSSION
293
Dr. Walter’s suggestions that the optimum activation may be related to autonomic accompaniments - it is more than likely true, that it may not be due to any particular part of the autonomic system, though this would bear further study. To F. Bremer I would like first to thank Prof. Bremer for his penetrating comments and for raising the important question of the selectivity of sampling in microelectrode records of cortical neuronal populations. It is certainly true that arrest of one unit discharge may be due to activation of a distant neurone: either an inhibitory neurone or one which is blocked by occlusion. With extracellular records it is not possible to determine the intimate mechanism of the arrest of discharge in a given neurone. However, the fact that “spontaneous” and evoked unit discharge may be decreased or increased when surface evoked potentials are decreased must be taken into consideration in the attempt to interpret the significance of surface evoked potential changes. We were also deceived by the lack of clear evidence for reciprocal changes in evoked potentials during visual or auditory distraction but, to my knowledge, such changes have not been demonstrated during simultaneous visual and auditory (test) stimulation. We intend to repeat these experiments using more refined techniques and weaker stimuli. To W. R. Adey 1 am not sure what Dr. Adey means by “transactional mechanisms in cerebral sensory systems”. If he means the temporo-spatial patterns of impulses which have “significance” in a given situation as judged by their determination of appropriate patterns of motor response or behaviour, regardless of their localization, specific or unspecific, then I would agree with him. Evoked potentials can be, at best, only a very crude sampling of some of these multiple events taking place simultaneously in the ongoing continuous pattern of activity in multiple and widespread circuits participating in the processing of sensory information and its translation into coordinated motor responses. The example of the bat who is able to discriminate his own radar signals from white noise of higher intensity is a beautiful illustration of the importance of specific patterns of impulses in a background of noise; the evoked potentials may, in some instances, represent only “noise” rather than the stimulus patterns of critical significance in a given situation. To A . Fessard Prof. Fessard has asked why we should direct so much of our attention to primary sensory receiving areas of the cortex when one would expect to find more of interest to the problems of conditioning in the association areas where there is evidence for convergence. This is precisely what we thought in the beginning of our studies, and we have found some very interesting results in the parietal area of the monkey, as I may show in an additional slide. This shows a single cell in the parietal cortex which has been conditioned to be facilitated in response to a conditioning light stimulus at one frequency of repetition, while being inhibited to the same flashes of light delivered at another frequency, the frequency of the differential unreenforced stimulus. When the animal made an error, responding to the differential stimulus, the same unit was facilitated instead of inhibited. However, studies of evoked potential changes and changes in unitary responses in the primary receiving area of the conditioning stimulus have also shown remarkable and constant relationships to conditioned motor responses. We feel that such changes require further analysis before proceeding to more complete studies of less “specific” neuronal systems.
Aspects of Sensorimotor Reverberation to Acoustic and Visual Stimuli The Role of Primary Specific Cortical Areas* P. BUSER, P. ASCHER, J. BRUNER, D. JASSIK-GERSCHENFELD** AND R. SINDBERG*** Laborafoire de NeurophysioloKie compare‘e, FaculfC des Sciences, Paris (France)
INTRODUCTION
Integrated motor behaviour depends, to a large extent, under normal conditions, upon responses to non-somaesthetic stimuli, as, for instance, stimuli of a visual or acoustic nature. From this it may be inferred that in the analysis of mechanisms of motor integration, an understanding of the topological and functional conditions of the release of centrifugal discharges by such non-somatic stimuli may be able to contribute t o our understanding of sensorimotor eraborations. Even if we do not include in our considerations the problem of the efficacy of these discharges at the segmental reflexlevel, it is not easy, as yet, to gain a precise insight into these central mechanisms. In view of these facts, an analysis of the various operations of motor conditioning is the most reliable method. By means of this method, it is possible to study how an integrated motor response is triggered by a “teleceptive” (visual or auditory) stimulus, whether the response is an orientation-reflex, a vegetative or an elaborated gestural activity. In the present state of our knowledge, however, it is hard to avoid the excessively schematic conclusion to which such studies often lead us. Conditional reflex experiments and, to a lesser degree, those concerned with other simple acts of learning, confront us with the problem of the role of subcortical structures, relative to that of the cerebral cortex (Buser and Roger 1957; Fessard and Gastaut 1958). The concept developed by Pavlov and his immediate successors, according to which cortical areas are of paramount importance, has received support from some recent observations (BureS 1959; Jouvet 1961 ; Pinto-Hamuy 1961). Other studies, especially those based on oscillographic analyses, tend to substitute for this view the conception of a subcortical “non-specific” substrate for temporary and acquired connections (Gastaut and
* The research reported in this document has been sponsored in part by the Air Force Office of Scientific Research, OAR, through the European Office, Aerospace Research, United States Air Force. ** Facultad de Medicina, Buenos Aires, Argentine. *** School of Medicine, University of Wisconsin, Madison, U.S.A.
THE ROLE OF PRIMARY SPECIFIC CORTICAL AREAS
295
Roger 1960; Hernkndez-Pe6n and Brust-Carmona 1961). This concept gains support by the very organization of these structures and by the fact that they are characterized by the convergence of neuronal impulses from many sources (Fessard 1954, 1960). In view of problems thus raised by studies of the mechanisms of conditioning and, especially, by the participation of neocortical areas, an attempt has been made to analyse as far as possible on “acute” experiments the process of release of centrifugal discharges that may accompany visual or acoustic stimulation. First data obtained are summarized herein. It soon became clear to us, as might have been expected, that in acute experiments the induction of centrifugal motor responses to “teleceptive” (visual or acoustic) stimuli is uncertain, except under special conditions such as chloralose anaesthesia. Therefore, initially at least, an analysis of these mechanisms has required the use of such experimental conditions. It should be noted however that partial confirmation of the results has later been obtained by experiments on animals prepared with only local anaesthesia and immobilized with curare. Finally the observations reported here aim at developing the outline of an hypothesis concerning the function of the primary sensory cortical areas in mechanisms of “sensori-motor integration”. In this particular case, it is proposed that these cortical areas possess a mode of action different from that recognized traditionally for the cortex, namely different from an inhibitory control on subcortical mechanisms. I. PYRAMIDAL A N D PERIPHERAL DISCHARGES TO AUDITORY A N D VISUAL STIMULI
It is well known that under deep chloralose anaesthesia visual or acoustic stimuli activate, in a more or less stereotyped manner, reverberating mechanisms which lead to efferent discharges. These discharges can be recorded from the pyramidal pathway and from the ventral spinal roots, two regions where their purely efferent nature is certain.
Fig. 1 Increase in amplitude of the pyramidal reflex discharges after application of strychnine to the motor cortex. Upper row: pyramidal responses to weak sensory stimuli (shock applied to a leg, click, flash, respectively). Lower row: after application on the motor cortex of a filter-paper soaked in a 0.5 O/OO strychnine solution. Responses to sensory stimulations of the same intensity as before show a marked increase in amplitude. Preparation under deep chloralose anaesthesia. Except when stated otherwise, all other figures refer to similar preparations. Time: 50 c/sec; amplitude: 100 ,uV (From Buser and Ascher 1960. Reproduced by permission from the Archives italiennes de Biologie). Xeferences P. J18-322
296
P. BUSER et
I
Lea
Click
al. Liaht
I
I
Fig. 2 Local cooling of the motor cortex depresses, then abolishes simultaneously its evoked potential to stimulation of the ipsilateral hind-leg, to a click, and to a flash (upper trace) and the corresponding pyramidal discharge (lower trace). The return to control was only partial in this experiment. C: control. At R, cooling for 4 min. Temperature reached: 10"C. During recovery, controls taken later at the 10th and 30th min with respective temperatures 25" and 36" C. Time scale: 20 msec.
A. Pyramidal discharges Bipolar electrodes were introduced stereotactically into one or both pyramidal tracts in the pontine region, by means of a technique permitting physiological control (Ascher and Buser 1958; Patton and Amassian 1960). Short stimuli, either visual (flash) or acoustic (clicks) evoked mass responses, whose characteristics are well known (Adrian and Moruzzi 1939; Ascher and Buser 1958; Patton and Amassian 1960). That these responses originate in the motor cortex* is proved by their total disappearance after its removal. Moreover, any experimental alteration of the reactivity of the motor cortex is immediately reflected in the amplitude of these
* This term refers to the pericrucial region, anterior sigmoid gyrus, and proximal part of the posterior sigmoid gyrus. The question of the exact extent of the motor area will not be discussed here [see Lance and Manning 1954; Patton and Amassian 1960; Buser 1960).
THE ROLE OF PRIMARY SPECIFIC CORTICAL AREAS
297
efferent discharges. They are amplified by local application of strychnine to the motor cortex (Fig. 1) and reduced or temporarily eliminated during transitory depression of the same area by drugs or local cooling (Fig. 2). These discharges obviously result from the arrival of afferent volleys (of visual or acoustic origin) to the ipsilateral motor area*. The existence of such afferent projections to the motor cortex has been well established (Gastaut and Hunter 1950; Wall et al. 1953; Hunter and Ingvar 1955; Feng et al. 1956; Buser and Ascher 1960; Jasper et al. 1960; Thompson and Sindberg 1960). Under chloralose, these evoked responses to light and sound (Fig. 2) are particularly large, and differ from primary potentials by their longer duration and greater latency; moreover, a comparison of the motor cortex response and the pyramidal discharges reveals their quasi-concomitance and suggests that the cortical activity is the cause of the pyramidal discharge**. Finally, relatively extensive processes of intermodal or heterosensorial facilitation may be observed, especially under chloralose. Successive application of a photic and an acoustic stimulus at a certain time interval (of the order of 20 msec) can release a pyramidal discharge of far greater amplitude than the sum of the discharges to each stimulation separately (Fig. 3). Such facilitations might take place at subcortical levels, at least in part (see p. 316).
I
Click
Flash
m3 3 1 Flash
Click
m Flash+Click
Fig. 3 A : discharges to auditory (click) and to visual (brief flash) stimulations recorded from the two pyramids. B : application of threshold stimuli: marked facilitation by pairing these two stimuli. Time scale: 50 cisec.
B. Peripheral discharges Under chloralose, efferent discharges to auditory and visual stimulation may also be obtained from the ventral lumbar roots (LG, L7 or Sl) or from various peripheral nerves (Alvord and Fuortes 1954). These efferent activities appear relatively constant in time and amplitude and thus may be subjected to detailed analysis. Their average latency is about 30 msec for sound and 70 msec for light; under conditions of stimula~~
* **
Recording was made anterior to the pyramidal decussation. Studying the pyramidal tract activity in free animals with implanted electrodes has confirmed our opinion that pyramidal responses to brief stimuli are also present and thus may not strictly be the result of the administration of chloralose (see below). References p . 318-322
298
P. BUSER
et a/.
tion used here, they are always bilateral. Moreover, these responses are susceptible to fatigue and are sensitive to repetition of the stimuli. When recording from the nerves supplying either flexors or extensors, discharges are only seen in those innervating flexors whereas those innervating extensors either do not react at all or undergo a transitory inhibition of their preexisting spontaneous activity during and after the flexor discharge. One may suppose that in this case the activation of the flexor neurones is accompanied by a reciprocal inhibition of the extensor neurones. In rare instances however, discharges appeared concurrently in the two groups of nerves; no explanation is found for such differences. These observations are in accordance with earlier data obtained by Moruzzi (1944) on myoclonus in animals under chloralose. Somaesthetic stimulations also elicit discharges that are very similar to those produced by light and sound (Fig. 3) (Ascher et al. 1961). Stimulation of one fore-leg produces a response in both L7 roots with a latency of about 20 msec. Stimulation of one hind-leg produces a more complex pattern: in the ipsilateral root, it consists of the short latency reflex discharge (7 msec) followed by a long latency response. In the contralateral root, only the late response is observed. Various facts suggest that these late responses to somaesthetic stimulations are closely related, in their mechanisms, to the root discharges to light and sound (Fig. 4a). L
PAG
S
P PG
Fig. 4
Discharges recorded from the two ventral roots L7 and elicited by a brief light-stimulus (L), by a click (S ) and by a shock applied to the left fore-leg (PAC) and to the left hind-leg (PPC). a : intact animal; 6: same animal after total ablation of the neocortex. Time calibration: 10 msec.
Finally, heterosensory reinforcement-phenomena may be observed, very similar to those obtained in records from the pyramids. Here also, facilitation results when the two stimuli (light and sound) are applied in the order flash-click and with an interval of approximately 10 msec (Fig. 5). In curarized animals with only local anaesthesia, efferent responses to isolated stimuli are not easily observed except in rare instances. Heterosensory reinforcements may be of use to increase their amplitude or even to elicit them (Fig. 6). Such activities can hardly be subjected to systematic analysis, due to their variability. They may however have an essential theoretical significance, since they diminish the tendency to consider phenomena observed under chloralose as being of an artificial nature.
THE ROLE OF PRIMARY SPECIFIC CORTICAL AREAS
299
Fig. 5 Heterosensory facilitations of efferent discharges recorded from the right and left L7 roots. Visual (L) and acoustic (S) stimulations, and stimulations of the right hind-leg (P)were each adjusted to subliminal values for the late efferent discharge (1st and 2nd row). The combination of any two of these stimuli, at a suitable interval, elicits a marked bilateral discharge.
Fig. 6 Discharges recorded from the semi-tendinous nerve in a non-anaesthetized curarized preparation. A : examples of heterosensory reinforcements of the discharges by combination of a sound (S) and a weak stimulation of one leg (P),both being of little or no efficiency when applied in isolation (anterior (A), posterior (P),homolateral ( H ) or contralateral ( C ) leg). B : each somatic stimulus, having been increased, elicits a late discharge.
C . Non-identity of the two categories of eferent discharges One could have been inclined to relate the ventral root discharges to those observed in the ponto-bulbar pyramid, especially in view of their respective latencies. However, a bilateral section of the pyramidal tract (performed under visual control, via a palatine approach) was never followed by a suppression of the spinal efferent responses. It may therefore be concluded that these reflect the activity of descending pathways distinct from the pyramidal tract, i.e., of “extrapyramidal” routes. For References P. 318-322
300
P. BUSER et
al.
reasons that remain to be stated (see below) the grouped pyramidal volley produced by a sudden sensory stimulus does not bring the spinal motoneurones to discharge. It may be concluded that, when exploring either the ponto-bulbar pyramid or the peripheral lumbar pathways, we are dealing operationally with two different categories of efferent activities. 11. EXTRAPYRAMIDAL AND PYRAMIDAL ACTIVATION FROM PRIMARY SENSORY AREAS
Characteristics of the responses evoked either in the pyramids or the peripheral fibres by electrical stimulation of various regions of the cortex* will now be considered. A. Cortical “efferent topography” based upon peripheral discharges The following observations were made (Ascher and Jassik-Gerschenfeld 1960). 1. In general, local stimulation of a cortical point by an isolated shock evokes practically identical discharges in the two symmetrical lumbar roots. 2. A systematic exploration of the cortical convexity shows that such responses to an isolated stimulus are not only elicited from the motor cortex but also from primary sensory areas, visual as well as acoustic. Moreover, the thresholds are practically identical for these various areas.
@?$ I
,,
Fig. I Discharges recorded from the left and right L7 roots to stimulation of the right cortex. Same gain on both channels, for all pictures. Note precession of the contralateral response to stimulation of the motor cortex.
3. There is no essential difference between the pattern of discharges from a primary area or from the motor area, except that in one given root the responses to stimulation of the contralateral motor cortex often slightly precede those evoked from all other cortical regions (Fig. 7).
*
In these experiments electrical stimuli are locally applied to the cortex by close bipolarelectrodes. The parameters of these stimulations (frequency, intensity and duration) are controlled. Records are taken from either the ventral L7 roots, from peripheral nerves or from the pontine pyramids.
THE ROLE OF PRIMARY SPECIFIC CORTICAL AREAS
30 1
Fig. 8 Discharges recorded from both L7 roots to stimulation of the right cortex, after bilateral pyramidotomy. Same gain on both channels in all pictures.
4. In the majority of the experiments, stimulation of the primary areas or the motor cortex appeared to be much more effective than that of association areas (medial or posterior suprasylvian gyrus and anterior lateral gyrus), as expressed by a very unequal density of cortical points with low threshold. However, the existence of corticifugal foci in the suprasylvian gyrus - especially at its latero-inferior limit (border of the ectosylvian gyrus) - should not be excluded, in as much as specific movements of the eyes and the pinna have been obtained from these regions (Claes 1938). After bilateral interruption of the pyramidal pathway at the pontine level, an exploration of the efferent topography gave the following results (Fig. 8): 1. the primary as well as the motor cortex remained effective; 2. the discharges obtained were unchanged, in their general pattern; the only modification consisted in a disappearance of the first component of shorter latency, emanating from the contralateral motor cortex, as mentioned above (3.). We may thus conclude that, except for that first component, which is probably of pyramidal origin, all other discharges elicited in the L, roots by an isolated cortical stimulus are of extrapyramidal origin, and typically bilateral. These results agree with those obtained by Colle and Massion (1958) and Gernandt and Gilman (1960). It seems moreover that the primary areas, visual or acoustic, possess an important corticofugal projection system of extrapyramidal nature; an abrupt and massive excitation of these pathways is likely to activate a certain number of spinal motoneurones. On the other hand, on application of repetitive shocks (a short train or a prolonged tetanization at 40 or 50 c/sec), definite differences are observed for a given root between stimulation of the contralateral motor cortex and that of all other regions. The early discharges coming from the contralateral motor cortex are increased on tetanization, while a rapid disappearance of all very fatigable extrapyramidal responses is noted after successive stimuli. At the same time, the response from the contralateral References P. 318-322
302
P. RUSER et al.
Fig. 9 Map of the cortical points eliciting pyramidal discharges through single electrical stimulation. Unanaesthetized curarized preparations. For each point, the value (in arbitrary units) is indicated of the stimulation threshold necessary for release of an indirect pyramidal discharge (see text). Filled circles represent the most active regions (thresholds below or equal to 5); open circles indicate points with higher threshold. Note a fairly good location of the active zones within the primary areas. (From Buser and Ascher 1960. Reproduced by permission from the Archives italiennes de Biologie).
motor area is increased. After pyramidal interruption this response as well as the early phase disappear. Obviously, they represent an activation of the motoneurones through the pyramidal system. These observations might explain the relative inability of a short duration pyramidal volley elicited by a sensory stimulus, to activate supraliminally the lumbar motoneurones. It is known moreover, that, at least in cats, pyramidal tract endings are located in the posterior horn, and never directly reach the motoneurones (Lloyd 1941; Chambers and Liu 1957).
B. Cortical efferent topography on the basis of pyramidal discharges When the pyramidal tract is explored instead of a spinal root, stimulation by isolated shocks at various points of the ipsilateral cortex yields the following results. 1. The discharge elicited from the motor cortex is of short duration, of very short latency (0.5 msec) and usually consists of two successive components. This is a wellknown phenomenon which needs no further comment (see Patton and Amassian 1960). 2. Stimulations of other cortical regions distinct from the area of origin of the pyramidal tract (Fig. 9) yield different results (Fig. 10A). Usually such stimulations, for instance applied to a primary sensory area evoke a discharge of a quite longer latency (5-1 5 msec). These discharges originate partly at least from an indirect activation of the motor cortex via a cortico-subcortico-corticalroute (Buser and Ascher 1960). Namely: (a) they are always accompanied by a potential in the motor cortex; the integrity of the motor area is indispensable for their appearance; (6) their long latency excludes any participation of direct pyramidal projections from the posterior cortex, the existence of which has been postulated (Walberg and Brodal), but
THE ROLE OF PRIMARY SPECIFIC CORTICAL AREAS
303
also disputed (Chambers and Liu 1957);
(c) deep coronary incisions between the motor area and the posterior cortex do not suppress the discharge to stimulation of the latter (Fig. 10B). Such a mechanism has been postulated before, viz. by lngvar and Hunter (1955), to explain the frontal “irradiation” of visual responses under cardiazol. However, these authors did not rule out the possible intervention of a second connection, purely cortico-cortical in nature, a mechanism which might also logically be considered. We are not able from our experiments either to prove or disprove the existence of such a mechanism; as yet, most of our observations have found a logical explanation in the hypothesis of a complex circuit, as postulated.
Fig. 10 Pyramidal and cortical responses to stimulation of the primary visual and acoustic cortex. A : responses from one pyramid (upper curve) and from the ipsilateral motor cortex to stimulation of ipsilateral visual (Cxv) and acoustic (Cxa) cortex. B : same activities recorded in another experiment, before and after deep bilateral coronary incision between the motor cortex and the posterior zones (see text). Time scale: 50 c/sec. Amplitude: 150 pV (pyramid) and 300 pV (cortex). (From Buser and Ascher 1960. Reproduced by permission from the Archives italiennes de Biologie).
3. All regions of the convexity do not possess the same efficacy. A systematic)opographical study has revealed that, as is the case with the extrapyramidal discharges, the primary visual and acoustic areas have the lowest thresholds, or, with a constant intensity, give the strongest discharges, in contrast to the association areas (Fig. 9). In summary, it thus seems that the primary visual and acoustic areas initiate a complex cortico-subcortico-cortical circuit likely to influence the neurones of origin of the pyramidal tract. The pathways involved in this circuit remain to be identified. 111. CONTROL OF THE EXTRAPYRAMIDAL A N D PYRAMIDAL REACTIVITY TO S O U N D A N D
LIGHT BY THE PRIMARY AREAS
With reference to the facts just described, we were faced with an alternative: we could either admit the independence of the process of sensorimotor reverberation from References P. 318-322
304
P. BUSER
et al.
the sensory cortical areas, or, on the contrary, attribute a functional value to the corticofugal pathways originating from these areas, and capable of activating both the extrapyramidal systems and the cortico-pyramidal pathway. Experiments to be described now were undertaken to clarify these points. A. Primary areas and extrapyramidal discharges This study yielded two sets of results, which were apparently contradictory. 1. Total elimination of the neocortex was usually first followed by a disappearance of the efferent responses to sensory stimuli, especially those to light, the acoustic responses being relatively more resistant. However, this disappearance of the reverberated responses proved to be transitory; usually, they reappeared later on, at least after some hours (Fig. 4b). One might thus conclude that the neocortical areas, either the sensory or the motor regions, are not indispensable for the activation of the extrapyramidal systems through visual or acoustic stimuli. 2. Other experiments were carried out, however, with the cerebral cortex being left intact. The aim of these experiments was to detect whether changes in the excitability of the primary cortical areas had an effect on the amplitude of the motor discharge. For this purpose, several techniques were used. On the one hand, a given sensory area was treated locally (by filter-paper) with a 2.5 O/o0 solution of strychnine. One might expect this treatment (which increases the primary potential to the corresponding primary stimulus) to have a facilitating influence on a possible “reverberation” process elicited at this level by the afferent volley. On the other hand, transitory depressions of the cortical reactivity were obtained by various procedures : local application of concentrated KCl ( 3 M ) , of ethylenechloride or of nembutal(6.5 O/,-,O) or local controlled coolings of the cortex. The latter method consisted of the application of a flat cooling element to the cortical surface. A liquid brought to adequate low temperature circulated through the cooling element during the period necessary for cooling and maintenance oflow temperature. Measurements with thermistors have shown that a nearly complete local reversible depression of the cortex is obtained in the area immediately below and around the cooling element, when temperatures of the order of 10°C are reached. At the same time regions at a distance of some 10 mm from the element varied by only a few degrees during the experiment. Using the various methods just described, it has been shown that, despite the apparent independence of the “reflex” mechanisms from the cortex, as postulated in 1, changes in the cortical excitability have an immediate repercussion on the amplitude of the efferent responses to sensory stimuli. Let us consider the details. 2a. Visual area. ( a ) Following a local strychninization of one visual area, the amplification of its evoked primary response is accompanied by a marked increase of the preexisting extrapyramidal discharge to light, sometimes together with a shortening of its latency. In experiments where no discharge to light was observed before, this procedure could cause such a discharge to appear. There is a rather good concomitance between the increase of the cortical responses and that of the peripheral
THE ROLE OF PRIMARY SPECIFIC CORTICAL AREAS
L
S
PAG
305
PPG
Fig. 1 1 Increase of the efferent discharges to light (recorded from right L7) after local application ofstrychnine to the visual cortex. SI: controls; stimulation by light (L), by sound (S) and by electrical shocks to the left fore- and left hind-legs (PAC and PPG). SIT: after local application of strychnine. Only the discharge to light is significantly modified. Time scale: each division represents 4 msec.
discharge (Fig. 11). This effect is bilateral, the increase involving discharges in the two symmetrical roots*. (b) On depression of one visual area by drugs or physical procedure, the converse occurs: as its primary potential decreases, there is a parallel reduction or even a complete disappearance of the motor discharge evoked by the corresponding stimulus (Fig. 12).
Fig. 12 Specific effect of a depression of the visual cortex (by application of ethyl-chloride) on the discharge to visual stimulation recorded from L7. The discharges to acoustic stimuli were not significantly changed in this experiment. For each picture: upper trace: L7; lower trace: visual cortex.
(c) These effects are strictly specific for the corresponding modality. Namely, treatment of the visual area influences only the root discharges evoked by light and has no effect on the responses to clicks or on the late responses to somaesthetic stimulations. ( d ) These changes are rapid and usually reversible, provided that the action on the cortex has not been too strong or prolonged. (e) Quite similar observations were made in animals after bilateral pyramidotomy. This confirms the independence of these mechanisms from the cortico-spinal pathway. 2b. The acoustic area. The results obtained by treatment of the acoustic area do not differ from those mentioned above; namely increases or decreases of peripheral
* The effects of strychninization have not been considered when local application gave rise to an epileptic focus. Refe?enccs p. 318-322
306
I>. BUSER
et a/.
Fig. 1 3 Specific effect of a local depression of the acoustic cortex on discharges elicited by clicks (S). Upper row: control; responses to sound (S),to light (L), to stimulation of the homolateral fore- and hindlegs (PAHand PPH). Lower row: immediately after local application of ethyl-chloride on the acoustic cortex. In all pictures, simultaneous recordings from L7 (upper trace) and acoustic cortex (lower trace).
discharges are observable under the same circumstances as for the visual area (Fig. 13). The only difference lies in the degree of specificity of these effects: in a certain number of our experiments, treatment of the acoustic areas” has not only altered the root discharges to clicks, but also those to light; in other experiments, the action has been strictly specific (Fig. 14). But in no case have the late somaesthetic responses shown any alteration during the same period. Thus it seems likely that the acoustic area is capable of influencing, at least sometimes, the two “teleceptive” modalities. Possibly, this might be related to the existence, in some preparations at least, of short latency visual responses in the acoustic area.
Fig. 14 Animal prepared under chloralose, having had a bilateral ablation of its sensorimotor cortex 4 months before. Application of strychnine to the visual area still has a relatively specific effect on the discharges produced by light (L). Discharges to sound ( S ) are unaltered, those to somatic stimulation ( P P H ) are slightly increased. Records are taken from the severed ventral root SIand from the visual cortex ( C V ) .
* Most of our studies were concerned with the area AH-Ep. Recent observations appear to indicate that the results also hold true for area At.
THE ROLE OF PRIMARY SPECIFIC CORTICAL AREAS
307
B. Primary areas and pyramidal discharges Our next problem was to determine whether primary sensory areas also exert an influence on the reactivity of the pyramidal motor system to sensory stimulations. As will be seen, such a control actually exists and the principles of its action are similar to those mentioned above for the extrapyramidal systems. However, the mechanism is here more complex, since it involves a circuit from the primary sensory area to the motor cortex, through subcortical relays. The major observations may be summarized as follows. 1. After complete elimination of the cortical sensory areas, or even of the entire convexity except the motor cortex, sensory stimuli still elicit responses on the motor cortex and in the corresponding pyramidal tract (Fig. 15). This, of course, only after
Fig. 15 Responses recorded from the right pontine pyramidal tract in an animal having had, 3 months before, an extensive bilateral ablation of its neocortex, leaving only the temporal pole, the insular zone and the motor cortex (see diagram). Chloralose anaesthesia. L, S: pyramidal responses elicited by, respectively, a flash and a click.
the depression caused by the ablation has subsided. Such facts indicate that at least a part of the afferent projections to the motor cortex are, as anatomical pathways, independent of the primary sensory areas, and that they may leave the main afferent systems at a subcortical level. This statement is in agreement with the opinion of most of the authors listed above (p. 297). 2. As a first indication of an influence of the primary areas on the pyramidal system, it was stated above that their stimulation by single electric shock can elicit a pyramidal discharge through a cortico-subcortico-cortical route (Fig. lo). 3. Our next step was to demonstrate that the pyramidal discharges to light are subjected to specific facilitations or depressions following appropriate treatments of the primary visual area. This was shown to be the case (Fig. 16). Treatments of the Referelices I). 318-322
308
P. BUSER
I
KclCxAc
et al.
-
-
40mr
Fig. 16 Effect on the pyramidal responses to a click of a local application of strychnine (Str C x Ac) and, subsequently, of KCl (KCI Cx Ac) on the acoustic cortex (area A 11-Ep). Records taken from one pyramid (Pyr) and from the ipsilateral acoustic cortex (Cx Ac).
acoustic area acted most often specifically on response to clicks, whereas in some cases they did not. In conclusion of this third section it may be stated that under the experimental conditions adopted here, the primary visual and acoustic sensory areas appear to exert a control on the mechanisms by which light or sound stimuli elicit centrifugal discharges. This control involves the extrapyramidal systems as well as the motorcortex and pyramidal tract. It is relatively or strictly specific for the sensory modality of the area considered; it is always specific with regard to the effects upon responses to somaesthetic stimulations. IV. SUBCORTICAL MECHANISMS
Observations to be reported now concern the deep substrates which may possibly be involved in the two types of cortical control, one acting upon extrapyramidal mechanisms, the other on the pyramidal system. A. Cortico-tegmental influences From what is known about the extrapyramidal motor mechanisms in general, various subcortical structures could be assumed to form the connection between the afferents, visual or acoustic, and the descending systems. Results obtained by other authors provide some useful indications as to their identification. Arguments of various kinds first suggest that the tegmental regions of the brainstem (mainly the reticular structures) could constitute one substrate for such mechanisms. The fact that the reticular formation receives various afferent inputs, and is, moreover, a zone of intermodal convergence is well-known and needs no further mention here (see Rossi and Zanchetti 1957). Furthermore, descending reticulospinal pathways originate from the same regions of the brain-stem. Their action on the motoneurones is demonstrated by the well-known facilitatory or inhibitory influences on reflex activity; however, a supraliminal action may also be possible, as efferent discharges have been obtained after stimulation of these reticular structures (Sasaki
THE ROLE OF PRIMARY SPECIFIC CORTICAL AREAS
309
et al. 1960a). The role of the tegmental regions (including reticular formation and also
red nucleus) in sensorimotor reflex mechanisms or integrations has often been suggested. Strauss (1929) for instance regards them as “the substrate of the startle reaction”. It seems also probable that they contribute to the elaboration of the orientation-reaction, in its electrographic and behavioural manifestations (Segundo et af. 1955a). Finally, the reticular region of the mesencephalon, according to Hess, forms a final zone of integration of postural mechanisms related to “specific directional” movements (see Hassler 1956; Jung and Hassler 1960). Therefore, it was rather logical to expect influences of the cortex on the reticular formation which would parallel those demonstrated to exist for peripheral discharges. Our experimental data, which thus far only concern the mesencephalic level, appear to be suggestive in this respect. 1. The topography of the corticofugal projections towards the mesencephalic reticular substance has again been considered, as a preliminary approach. The existence of such projections is well-known and a definite role has been ascribed to them, especially in the arousal mechanism (Jasper et af. 1952; Hugelin et af. 1953; Bremer and Terzuolo 1953, 1954; Von Baumgarten et al. 1954; French et af.1955; Hugelin and Bonvallet 1957a,b). In our experiments, isolated shocks were applied to various regions of the cortex. On stimulation of some of these areas, a response with short latency (2-8 msec) but slow rising phase was obtained in the reticular areas. The topographical distribution of the “active” cortical points appeared to be very constant from one experiment to another; actually the largest responses (or the lowest thresholds) were obtained in the sensorimotor cortex, the primary acoustic (especially the region AII-Ep) and the primary visual areas. (see also, for the latter area, Shanzer and Dumont-Tyc 1961.) On the other hand, the association areas of the convexity (lateral, anterior and suprasylvian gyri) have a rather limited effectiveness. It thus became clear, once more, that the cortical convexity is not homogeneous with respect to its projections towards the mesencephalic reticular formation. There is a striking density of these projections in the primary receiving areas; one might well wonder, in connection with the preceding results, what their functional significance may be (Fig. 17). 2. On the other hand, it must be accepted as an established fact that the sensory projections to the reticular area do not depend on the neocortex (Starzl et af. 1951 ; Dell 1952; see Rossi and Zanchetti 1957). 3. Experiments similar to those performed on the spinal or pyramidal discharges, i.e., including local changes in the excitability of the cortex, have yielded similar results, if not even more indicative of a strict modal specificity of the effects (Borenstein and Buser 1960). Namely, amplifications or depressions of reticular sensory responses of a given modality are observed on treatment of the corresponding specific cortical areas. Moreover, such modifications may be recorded a t various anteroposterior levels of the mesencephalic tegmentum (Fig. 18 and 19). It should be added that these observations did not, this time, depend upon chloralose anaesthesia, as they were also carried out on curarized animals prepared with only local anaesthesia. Such influences of the cortex on the reticular level, as well as others, References P. 318-322
310
P. BUSER
et al.
Fig. 17 Responses elicited in the mesencephalic reticular formation by cortical electric stimulation: mapping of cortical points, either evoking tegmental responses to a single electric shock (solid dots) or relatively inefficient and only active under intense stimuli (circles).
Fig. 18 Effect of an activation or a depression of the primary cortex on the acoustic or visual responses in the mesencephalic reticular formation. 1. Treatment of the visual cortex with control of the acoustic system. 11. Treatment, later during the same experiment, of the acoustic cortex with control of the visual system. Column A c : application of clicks; simultaneous recording from reticular formation (upper curve) and from ipsilateral acoustic cortex (lower curve). Column Vis: application of light stimuli; reticular recording unchanged (upper curve); lower curve: ipsilateral visual cortex. C : initial control; reticular and cortical responses to light and sound. Str I : 2 min after local application of a 0.5 O/OO strychnine solution to the visual area. Increase of the cortical and of the tegmental responses to light; responses to sound unchanged. K 1 : immediately after local cooling at the same site. Depression of visual responses. C1: control 30 min after cooling of the visual area. Accidental (?) depression of the acoustic system. Str 2: immediately after CI, local application of strychnine to the acoustic cortex. Transitory reappearance of a tegmental response to sound. No alterations in the reactivity of the visual system. K 2 : after local cooling of the same acoustic cortex. Tim-: 20 msec. Same calibration for all channels: 200 pv.
THE ROLE OF PRIMARY SPECIFIC CORTICAL AREAS
c
31 1
Str
Cx A
s L.
S A.
Fig. 19 Changes in the responses to light (S.L.) and to sound (S.A.) at two levels of the mesencephalic reticular formation (Ra, at A3 Horsley-Clarke plane, Rp at P2 plane). C : control. Str C x A c : local application of strychnine to the posterior acoustic cortex. R : return to control. Curarized preparation under local anaesthesia.
somewhat similar, recently described (Adey et al. 1957; BureB et al. 1961 ; Weiss and Fifkovh 1961b) must be attributed to the cortico-tegmental pathways. However, the detailed mechanism of their action remains to be clarified. The participation of other subcortical levels in the extrapyramidal sensorimotor reverberation should also be considered. Actually it is not impossible that the tegmental zone is located below the site of “reverberation” which, in such case, could be for instance striatal, as has been postulated by Moruzzi (1944). In the caudate nucleus, a specific cortical influence on the sensory reactivity is actually also observable, similar to that demonstrated for the reticular mesencephalic region (Encabo et al. 1961). However, most of the data obtained so far suggest an inhibitory action on the pathways descending from the caudate nucleus, the putamen and even the globus pallidus (Akert and Anderson 1951; Sasaki et al. 1960 b ; Spiegel and Szekely 1961). Only a few authors consider them to possess a facilitatory influence (Segundo et al. 1958). Thus it remains for further studies to determine the role of this and other structures.
B. Cor t icothalamic influences An influence of the cortex on structures of the median thalamus, especially on the centre me‘dian and the immediately adjacent structures also exists, as will be shown now. This was to be expected, when considering the role these non-specific structures probably play in the mechanisms of sensorimotor integration. 1. Projections have been demonstrated from various cortical areas to the median References P. 318-322
312
P. BUSER
et al.
regions of the posterior thalamus, but more particularly from the frontal zones (Jasper et a/. 1952; Albe-Fessard and Gillett 1961). Our immediate concern has been oriented towards the existence and characteristics of such projections from the primary visual as well as acoustic areas (Fig. 22C). 2. Previous studies had shown alterations in the median thalamus during a spreading depression of the cortex, which involved either the spontaneous activity (Sloan and Jasper 1950) or the sensory responses (BureB et a/. 1960; Weiss and FikovB 1961b). Our experiments showed again that changes in the excitability of the primary cortex - visual or acoustic - may influence the median thalamic responses, in the same way and with the same specificity as in the mesencephalic reticular formation. Figs. 20, 21 and 22 illustrate these points. Finally it should be pointed out that also in this case, quite similar results have been obtained without general anaesthesia. These results may be of interest, in view of the fact that some nuclei belonging to the postero-medial, “non-specific” thalamus (centre mkdian and adjacent regions) probably include the visual and acoustic pathways to the motor cortex.
Rec
4mn
30mn
-
100 ms
Fig. 20 Effects of local cooling of the visual cortex (Fd. Cx V.) on the response of the medial thalamus to light stimulation (SL). C : control. Rec: progressive recovery: three successive records taken after 4 min; final control at 30 min. Right column: responses t o stimulation of the ipsilateral hindleg (SP)remain practically unchanged during the whole experiment. On the microphotograph, the point explored, at the bottom of the electrode track is underlined by an arrow (frontal plane A6,S). Gain unchanged through the experiment. Upper trace of each record, medial thalamus; lower trace, visual cortex.
THE ROLE OF PRIMARY SPECIFIC CORTICAL AREAS
313
Fig. 21 Specific effects of an application of strychnine to the visual cortex (Stv CxV) on the pyramidal discharge and on the visual response of the median thalamus (SL).Controls for acoustic stimulations (SA). Records from the visual cortex (CxV),from the acoustic cortex (Cx.ac), from the medial thalamus (7%) and the ipsilateral pyramid (Pyr).
Let us summarize the facts related to this statement. (a) Obviously, these posterior ventromedian levels receive projections belonging to
the two sensory modalities considered here (Dell 1952; Ingvar and Hunter 1955; Liu and Shen 1958; Albe-Fessard and Mallart 1960; Buser and Bruner 1960; Meulders and Massion 1961). Moreover, intermodal facilitation processes to combinations of light and sound at short interval may be observed at this thalamic level, which in all respects are similar to those observed on the motor cortex and in the pyramidal tract (Fig. 22A). Furthermore, parallel variations in the amplitude of the thalamic and “cortico-motor” responses are often seen when changes occur in the experimental conditions (intensity of the stimuli, transitory slight depression not affecting the responses from the primary pathway, etc.). All these facts suggest a close relationship between the two categories of responses, thalamic and cortical. (6) The zone considered above belongs to a group of thalamic structures which can evoke, when stimulated electrically, short latency positive-negative responses* in the ipsilateral motor-cortex (Fig. 22B) ; their positive phase is usually accompanied by a typical pyramidal discharge** (Imbert et a[. 1959).
* The question of the direct or relayed character of these various projections will not be dealt with here. ** It might be pointed out that other “non-specific” thalamic nuclei (intralaminar or of the midline) as well as the caudate nucleus, either have no activating effect to the pyramidal tract following electrical stimulation (Brookhart and Zanchetti 1956; Purpura et al. 1958; Buser et al. 1961) or trigger off pyramidal discharges with a much longer latency (Purpura and Housepian 1961). References P. 318-322
Fig. 22A. Localization of thalamic points responding to sound and to light (taken from twelve experiments). Examples are shown of responses to weak sound (A), to a flash (V) and to a pairing of such stimuli ( V A) (upper curve of each picture). Marked parallel facilitation in the thalamus and on the motor cortex (lower curve). Time 20 msec. The probable lateral contour of the centre mPLliun has been underlined (Horsley-Clarke plane A 7).
+
Fig. 22B. Points of median thalamus eliciting pyramidal responses when stimulated electrically (with single shocks). Records taken from the contralateral (upper trace) and ipsilateral (lower trace) pyramidal tract.
THE ROLE OF PRIMARY SPECIFIC CORTICAL AREAS
315
Fig. 22C. Thalamic points responding to isolated stimulation of the visual or acoustic cortex.
Fig. 22D. Localization of some thalamic points which have been influenced by strychninization (S) or by depression (D) on a specific cortex, either visual (CxV) or acoustic (CxA). References
I).
318-322
316
P.
nusm et al.
(c) Specific alterations in the medial thalamic responses appear to be accompanied by parallel changes in the amplitude of the corresponding pyramidal discharges. The relationship between these two groups of modifications is rather strict (Fig. 21). Finally, regardless of whether the median thalamic structures are the only ones likely to form the pathway towards the motor cortex, all the facts could justify a search for signs of cortical control at this level as well as at the more caudal mesencephalic one. In addition, these observations seem to indicate that the influence of the sensory areas on the reactivity of the pyramidal system, as described above, may actually be exerted upon the input to the motor cortex itself, i.e., at the thalamic level. The conditioning of the efferent discharge would then only be a consequence of an action at the diencephalic level. This, however, remains merely an hypothesis. In summary of this fourth part of our study, we may conclude that repercussions of a cortical control are detectable at two subcortical levels - the reticular formation and the medioposterior thalamus. The first structure is likely to participate in the extrapyramidal elaborations, whereas the second group possibly intervenes in the activation of the pyramidal system. This might at least represent a partial and provisional explanation for the cortical actions on the efferent discharges. DISCUSSION
Since some specific points have already been discussed above, this discussion will be limited to a consideration of some genera1 problems raised by the various facts described. 1. One explanation of the corticofugal effects, which could not be excluded apriori, was that of a cortical action on the sensory influx at the input level (Desmedt and Mechelse 1959). One observation allows us to reject this assumption, namely that the amplitude of the afferent volley in the specific thalamic relay nucleus is not significantly affected, under the same experimental conditions, either by strychninization or by depression of the corresponding primary cortex. This fact agrees with recent observations by Weiss and Fifkova (1961a). 2. An essential point concerns the general mechanism of control by the cortex. Except for the purely motor effects, the general tendency is to attribute to the cortex an inhibitory influence upon subcortical mechanisms (the term “inhibition” being used here in a rather general and not strictly defined meaning). Various facts support this view: the spontaneous activity or intrinsic conduction velocity in the mesencephalic reticular formation is decreased under the influence of the neocortex (HernandezPe6n and Hagbarth 1955; Adey et al. 1957; BureS et af. 1961); also at the thalamic level, sensory responses in the centre mkdian are amplified after decortication in unanaesthetized animals (Massion and Meulders 1960, 1961) ; likewise, units in the lateral geniculate body may be blocked by electrical stimulation of the visual cortex (Buser and Segundo 1959; WidCn and Ajmone Marsan 1960); finally, the results obtained by Hugelin and Bonvallet (1957a,b) are especially suggestive of an inhibitory action of the cortex on the reticular system.
THE ROLE OF PRIMARY SPECIFIC CORTICAL AREAS
317
More rarely, reverse effects have been obtained, which have been called “dynamogenic”, “facilitatory” or “activating” influences (these terms being not sufficiently defined either). In cats (Bremer and Terzuolo 1952, 1954) and monkeys (Segundo et al. 1955b), stimulation of the cortico-reticular pathways may result in a general awakening; increased latency of the visual diencephalic responses has been observed after chronic ablation of the visual cortex (Ingvar and Hunter 1955). Finally, and in agreement with our results, reductions have recently been noticed to occur for the evoked reticular potentials, during a cortical spreading depression (Weiss and Fifkovh 1961b). The cortical influences reported here fall in the second category, the one in favour of an activating role of the cortex. One might ask what the reasons are for the differences between our observations and most of the results which suggest an inhibitory cortical influence eliminated by decortication. One remark to be made is that in almost no cases, has there been a precise topographical location of the inhibitory neocortical actions, in contrast to our own data. One thus feels inclined to think that both mechanisms are not necessarily incompatible ; further studies will perhaps yield informations on the conditions of the respective importance and mutual interplay of “local facilitatory” and “general inhibitory” influences. 3. Another question to be solved is that of the elementary mechanisms participating in the cortical effects described here. Let us only outline the problem: these mechanisms may be either phasic, that is to say, they may depend on the instantaneous density of the efferent influxes triggered off by the afferent volley (and thus would constitute a true “cortical reflex”), or they may have a sustained action, consequently a tonic one. This problem remains to be solved. 4. Finally it might be essential to stress again the specificity - regarding the sensory spheres - of these cortical actions*. Doubtlessly, our data represent a rough approach to the question. However, this notion of specific influences may deserve special attention in its theoretical aspects because of the scheme which it implies. According to our hypothesis, each primary area, i.e., the specific final stage of a sensory modality, would control exclusively for its own modality, the reactivity of the regions which we know to be zones of convergence (reticular formation, median thalamus, etc.. .). In such a way, it may be possible to reconcile two aspects of the sensorimotor integration processes which seem contradictory: on one side the existence of mechanisms of sensory convergences, on the other side, the fact that one category of stimuli may become selectively effective, in contrast to all others.
.
SUMMARY
In animals under chloralose anaesthesia, application of visual and acoustic stimuli leads to a reflex-activation of efferent systems. A description is given of the bilateral responses recorded at the pontine level of the pyramidal tract and from the ventral roots (L6 to Sl) or from various nerves of the hind-leg. It is also shown that these
* One will recall that, in experiments of a rather different principle, on disturbancesin “sensory” conditioning produced by epileptogenic cortical foci, Morrell et al. (1956) also stressed such a specificity in cortical effects. References P. 318-322
318
P. UUSER
et d.
two categories of efferents are not identical since pyramidotomy does not suppress the peripheral discharges. Therefore it is most likely that the latter originate from activation of “extrapyramidal” systems. This study is mainly concerned with the influence of the primary visual and acoustic cortices on the mechanism of these extrapyramidal and pyramidal “reflex” activations. Peripheral discharges can be obtained by electrical stimulation of the primary visual and acoustic areas; the association areas are not very efficient. Since these discharges showed no changes after pyramidotomy, it is concluded that such a process activates extrapyramidal corticofugal systems. Then it is demonstrated that there may be an “activating” control of these primary areas on the extrapyramidal reverberations elicited by sound and light. Total ablation of the neocortex does not permanently suppress these reflex discharges. However, in the presence of the neocortex, each primary area seems to be capable of affecting the amplitude of the discharges triggered off by stimulations of the corresponding sensory modality, and exclusively the latter. This is suggested by the effect of local application of strychnine to the cortex, which leads to an amplification of the peripheral responses, or that of local depression, which reduces or temporarily suppresses these responses. A similar mechanism is demonstrated to exist for the pyramidal activity. It is shown that pyramidal discharges can be obtained by electric stimulation of the primary areas, the responsible pathway being of the cortico-subcortico-cortical type. Again the sensory cortices (visual and acoustic) are apparently not obligatory for provoking these discharges evoked by sound and light; nevertheless amplifications or suppressions of the pyramidal discharge to a certain modality are observable, which closely parallel changes produced in the reactivity of the corresponding primary area. In a search of the subcortical structures possibly involved in this cortical control, it is shown that specific influences of the same type are exerted by the cortex on the response of “non-specific” structures, midbrain reticular formation and posteromedial thalamus (centre mPdian and internal medullary lamina). These changes in the sensory reactivity of the “integrative” subcortical structures might help to explain the cortical control on the extrapyramidal as well as on the pyramidal discharges.
REFERENCES ADEY,W. R., SEGUNDO, J. P. and LIVINGSTON, R. B. Corticofugal influences on intrinsic brain stem conduction in cat and monkey. J. Neuruphysid., 1957, 20: 1-16. ADRIAN,E. D. and MORUZZI,G. Impulses in the pyramidal tract. J . Physiol. (Lond.), 1939, Y7: 153-199. AKERT,K . und ANDERSON,B. Experimenteller Beitrag zur Physiologie des Nucleus caudatus. Acta physiol. scand., 1951, 22: 281-298. ALBE-FESSARD, D. et GILLETT,E. Convergences vers le centre median. Electroenceph. din. Neurophysid., 1961,13: 257-269. ALEE-FESSARD, D. et MALLART, A. Existence de reponses d’orgines visuelle et auditive dans le centre median du thalamus du chat anesthesie au chloralose. C.R . Acad. Sci. (Paris), 1960, 251 : 1040-1 042.
THE ROLE OF PRIMARY SPECIFIC CORTICAL AREAS
319
ALVORD,JR., E. C. and FUORTES, M. G. F. A comparison of generalized reflex myoclonic reactions elicitable in cats under chloralose anesthesia and under strychnine. Amer. J. Physiol., 1954, 176: 253-261. ASCHER,P. et BUSER, P. Modalites de mise en jeu de la voie pyramidale chez le chat anesthesie au chloralose. J . Physiol. (Paris), 1958, 50: 129-132. D. Modalites d’obtention de decharges peripheriques par ASCHER,P. et JASSIK-GERSCHENFELD, stimulation corticale chez le chat. J. Physiol. (Paris), 1960, 52: 7-8. ASCHER,P., JASSIK-GERSCHENFELD, D. et FANJUL-MOLES, M. L. R6le des aires corticales somatiques dans le contr6le de certaines decharges motrices somethesiques. J . Physiol. (Paris), 1961, 53: 254-255. P. et BUSER,P. Observations sur les projections du cortex dans la formation reticulee BORENSTEIN, mesencephalique chez le chat. C.R. Soc. Biol. (Paris), 1960, 154: 3 8 4 2 . BREMER, F. et TERZUOLO, C. R61e de I’ecorce ctrebrale dans le processus du reveil. Arch. inf. Physiol., 1952,60 : 228-23 1. BREMER, F. et TERZUOLO, C. Interaction de I’ecorce cerebrale et de la formation reticulee du tronc cerebral dans le mecanisme de I’eveil et du maintien de l’etat vigile. J. Physiol. (Paris), 1953, 45: 56. BREMER,F. et TERZUOLO, C. Contribution a I’etude des mecanismes physiologiques du maintien de I’activite vigile du cerveau. Interaction de la formation reticulee et de I’ecorce ckrebrale dans le processus du reveil. Arch. int. Physiol., 1954, 62: 157-178. BROOKHART, J. M. and ZANCHETTI, A. The relation between electrocortical waves and responsiveness of the corticospinal system. Electroenceph. clin. Neurophysiol., 1956, 8 : 427444. BURE~ J., Reversible decortication and behavior. In M. A. B. BRAZIER(Editor), The central nervous system and behavior. Second conference. Josiah Macy Foundation, Madison Printing Cy., Madison, N.J., 1959: 207-248. BUREB,J., BURESOVA, 0.and FIFKOVA, E. The effect of cortical and hippocampal spreading depression on activity of bulbopontine reticular units in the rat. Arch. ital. Biol., 1961, 99: 23-32. O., FIFKOVA,E. and OLDS,5. Vliv funktni dekortikace na vrzuiivost hypoBUREB,J., BURESOVA, thalamicke a tegmentalni oblasti. &. Fysiol., 1960, 9 : 405. BUSER,P. Observations sur l’organisation fonctionnelle du cortex moteur chez le chat. Bull. Schweiz. Akad. med. Wiss., 1960, 16: 355-397. BUSER,P. et ASCHER,P. Mise en jeu de reflexes du systeme pyramidal chez le chat. Arch. ital. Biol., 1960, 98: 123-164. BUSER,P. et BRUNER,J. Reponses visuelles et acoustiques au niveau du complexe ventromddian posterieur du thalamus chez le chat. C.R. Acad. Sci. (Paris), 1960, 251: 1238-1240. BUSER,P. et ROGER,A. Interpretation du conditionnement sur la base des donnkes electroencephalographiques. Premier congr2s international des sciences neurologiques. Acta Medica Belgica, Bruxelles, 1957: 417444. BUSER,P. et SEGUNDO, J. Influences reticulaires, somesthesiques et corticales au niveau du corps genouille lateral du thalamus chez le chat. C.R. Acad. Sci. (Paris), 1959, 249: 571-573. BUSER,P., ENCABO, H. et BORENSTEIN, P. Action suppressive du noyau caude sur la reactivite reflexe du systeme pyramidal chez le chat. C.R. Acad. Sci. (‘Paris), 1961, 253: 538-540. CHAMBERS, W. W. and LIU, C. N. Corticospinal tract of the cat. J. romp. Neurol., 1957, 108: 23-55. CLAES,E. Etude des relations fonctionelles des cortex sensitif, visuel et auditif avec les regions oculomotrices corticales. C.R. SOC.Biol. (Paris), 1938, 127: 1116. COLLE,J. et MASSION,J. Effet de la stimulation du cortex moteur sur l’activite electrique des nerfs phreniques et medians. Arch. int. Physiol., 1958, 66: 496-514. DELL,P. Correlations entre le systeme vegetatif et le systime de la vie de relation. Mesencephale, diencephale et cortex drebral. J. Physiol. (Paris), 1952, 44: 471-557. DESMEDT, J. E. and MECHELSE, K. Corticofugal projections from temporal lobe in cat and their possible role in acoustic discrimination. J. Physiol. (Loncl.), 1959, 147: 17-1 8. H., BORENSTEIN, P. et BUSER,P. Observations relatives aux projections du neocortex vers le ENCABO, noyau caude. J . Physiol. (Paris), 1961, 53: 434-435. FENG, T. P., LIU, Y . M. and SHEN,E. Pathways mediating irradiation of auditory and visual impulses to the sensorimotor cortex. Proc. XXth int. physiol. Congr. Brussels, 1956: 997.
320
P. BUSER
et al.
FESSARD, A. Mechanisms of nervous integration and conscious experience. In J. F. DELAFRESNAYE (Editor), Brain mechanisms and consciousness, Blackwell, Oxford, 1954: 200-236. FESSARD, A. Le conditionnement considere & I’echelle du neurone. Electroenceph. clin. Neurophysiol., 1960, SUPPI.13: 157-184. FESSARD, A. et GASTAUT, H. Correlations neurophysiologiques de la formation des reflexes conditionnels. In Le conditionnement et l’apprentissage. P.U.F., Paris, 1958: 15-90. FRENCH, J. D., HERNANDEZ-PE~N, R. and LIVINGSTON, R. Projections from cortex to cephalic brain stem (reticular formation). J. Neurophysiol, 1955, 18: 74-95. GASTAUT, H. and HUNTER,J. An experimental study of the mechanism of photic activation in idiopathic epilepsy. Electroenceph. clin. Neurophysiol., 1950, 2: 263-287. GASTAUT, H. et ROGER,A. Les mecanismes de l’activite nerveuse supbrieure envisages au niveau des grandes structures fonctionnelles du cerveau. Electroenceph. clin. Neurophysiol., 1960, Suppl. 13: 13-38. GERNANDT, B. E. and GILMAN, S. Interactions between vestibular, pyramidal, and cortically evoked extrapyramidal activities. J. Neurophysiol., 1960, 23: 516-533. HASSLER, R. Die zentralen Apparate der Wendebewegungen. Arch. Psychiat. Nervenkr., 1956, 194: 481-516. HERNANDEZ-PE~N, R. and BRUST-CARMONA, H. Functional role of subcortical structures in habituation and conditioning. In J. F. DELAFRESNAYE (Editor), Brain mechanisms and learning. Blackwell, Oxford, 1961: 393412. HERNANDEZ-PE~N, R. and HAGBARTH, K. E. Interaction between afferent and cortically induced reticular responses. J. Neurophysiol., 1955, 18: 44-55. HUGELIN, A. et BONVALLET, M. Tonus cortical et contrBle de la facilitation motrice d’origine rtticulaire. J. Physiol. (Paris), 1957a, 49: 1171-1200. HUGELIN, A. et BONVALLET, M. Etude experimentale des interrelations reticdo-corticales. Proposition d’une theorie de I’asservissement reticulaire & un systkme diffus cortical. J. Physiol. (Paris), 1957b, 49: 1201-1223. HUGELIN, A., BONVALLET, M. et DELL,P. Topographie des projections corticomotrices au niveau du tdlenckphale, du dienckphale, du tronc cerebral et du cervelet chez le chat. Rev. neurol., 1953, 89: 419-425. HUNTER, J. and INGVAR, D. H. Pathways mediating metrazol induced irradiation of visual impulses. Electroenceph. clin. Neurophysiol., 1955, 7: 39-60. IMBERT, M., ROGER,A. et BUSER,P. Le cortex pyramidal du chat. Organisation unitaire et systkme d’afferents rkflexogknes. J. Physiol. (Paris), 1959, 51 : 482483. TNGVAR, D. H. and HUNTER, J. Influence of visual cortex on light impulses in the brain stem of the unanesthetized cat. Acta physiol. scand., 1955, 33: 1-25. JASPER,H., AJMONE MARSAN, C. and STOLL,J. Corticofugal projections to the brain stem. Arch. Neurol. Psychiat. (Chicago), 1952,67: 155-171. JASPER,H., Rrccr, G. and DOANE,B. Microelectrode analysis of cortical cell discharge during avoidance conditioning in the monkey. Electroenceph. din. Neurophysiol., 1960, 12: 137-1 55. JOUVET,M. Recherches sur les mecanismes neurophysiologiques du sommeil et de I’apprentissage negatif. In J. F. DELAFRESNAYE (Editor), Brain mechanisms and learning. Blackwell, Oxford, 1961: 445-479. JUNG,R. and HASSLER, R. The extrapyramidal motor system. In J. FIELD,H. W. MAGOUN and V. E. HALL(Editors), Handbook ofphysiology : Neurophysiology. American Physiological Society, Washington, 1960, ZI: 863-928. LANCE,J. W. and MANNING, R. L. Origin of the pyramidal tract in the cat. J. Phjwiol. (Lond.), 1954, 124: 385-399. LIU, Y . M. and SHEN,E. Pathways mediating the electrical response of the motor cortex to brief auditory and visual stimuli in cat. Acta physiol. sin., 1958, 22: 104118. LLOYD,D. P. C. The spinal mechanism of the pyramidal system in cats. J. Neurophysiol., 1941, 4 : 525-546. MASSION, J. et MEULDERS, M. lnfluence de la dkortication sur les rdponses d’origine somesthesique au niveau du centre median. J. Physiol. (Paris), 1960, 52: 172-173. MASSION, J. et MEULDERS, M. Les potentiels Cvoques visuels et auditifs du centre median et leurs modifications aprks decortication. Arch. int. Physiol., 1961, 69: 26-29.
DISCUSSION
32 1
MEULDERS, M. et MASSION, J. Effet facilitateur de la lumikre continue sur les potentiels evoquds au niveau du centre median par la stimulation electrique de structures voisines du corps genouille lateral. Arch. int. Physiol., 1961, 69: 4 0 7 4 9 . MORRELL, F., ROBERTS, L. and JASPER, H. H. Effect of focal epileptogenic lesions and their ablation upon conditioned electrical responses of the brain in the monkey. Electroenceph. clin. Neurophysiol., 1956, 8 : 217-236. MORUZZI,G . Un nuovo riflesso interessante il sistema extrapiramidale. Arch. Fisiol., 1944, 44: 86-108. NIEMER, W. I. and JIMENEZ-CASTELLANOS, J. Corticothalamic connections in the cat as revealed by physiological neuronography. J. comp. Neurol., 1950, 93: 101-123. PATTON,H. D. and AMASSIAN, V. E. The pyramidal tract: its excitation and functions. In J. FIELD, H. W. MAGOUN and V. E. HALL(Editors), Handbook of physiology: Neurophysiology. American Physiological Society, Washington, 1960, 11: 837-862. PINTO-HAMUY, T. Role of the cerebral cortex in the learning of an instrumental conditional response. In J. F. DELAFRESNAYE (Editor), Brain mechanisms and learning. Blackwell, Oxford, 1961: 589-601. PURPURA, D. P. and HOUSEPIAN, E. M. Alterations in corticospinal neuron activity associated with thalamocortical recruiting responses. Electroenceph. clin. Neurophysiol., 1961, 13: 365-381. PURPURA, D. P., HOUSEPIAN, E. M. and GRUNDFEST, H. Analysis of caudate cortical connections in neuraxially intact and “telencephale isole” cats. Arch. ital. Biol., 1958, 96: 145-176. ROW, G . F. and ZANCHETTI, A. The brain stem reticular formation. Anatomy and physiology. Arch. ital. Biol., 1957, 95: 199-435. SASAKI,K., NAMIKAWA, A. and HASHIRAMOTO, S. The effect of midbrain stimulation upon alpha motoneurones in lumbar spinal cord of the cat. Jap. J. Physiol., 1960a, 10: 303-316. SASAKI, K., NAMIKAWA, A. and MATSUNAGA, M. Effects of stimulations of the pyramidal tract and striate body upon spinal motoneurones. Jap. J. Physiol., 1960b, 10: 403-413. SEGUNDO, J. P., ARANA,R. and FRENCH,J. D. Behavioral arousal by stimulation of the brain in the monkey. J. Neurosurg., 1955a, 12: 601-613. SEGUNDO, J. P., MIGLIARO, E. F. and ROIG,J. A. Effect of striatal and claustral stimulation upon spinal reflex and strychnine activity. J. Neurophysiol., 1958, 21: 391-399. SEGUNDO, J. P., NAQUET, R. and BUSER,P. Effects of cortical stimulation on electrocortical activity in monkeys. J. Neurophysiol., 1955b, 18: 236-245. SHANZER, S. et DUMONT-TYC, S. Etude Clectrophysiologiquedes projections du cortex visuel sur le tronc cerebral. J. Physiol. (Paris), 1961, 53: 473474. SLOAN,N. and JASPER,H. The identity of spreading depression and “suppression”. Electroenceph. clin. Neurophysiol., 1950, 2 : 59-78. SPIEGEL, E. A. and SZEKELY, E. G. Prolonged stimulation of the head of the caudate nucleus. A.M.A. Arch, Neurol., 1961, 4 : 55-65. STARZL, T. E., TAYLOR, C. and MAGOUN, H. W. Collateral afferent excitation of reticular formation of brain stem. J. Neurophysiol., 1951, 14: 479. STRAUSS,H. Das Zusammenschrecken: Experimentell kinematographische Studie zur Physiologic und Pathophysiologie der Reaktivbewegungen. J. Psychol. Neurol. (Lpz.) , 1928, 39 : 1 11-23 1. THOMPSON, R. and SINDBERG, R. Auditory response fields in association and motor cortex of cat. J. Neurophysiof., 1960, 23: 87-105. VON BAUMGARTEN, R., MOLLICA,A. und MORUZZI,G. Modulierung der Entladungsfrequenz einzelner Zellen der Substantia reticularis durch corticofugale und cerebellare Impulse. Pflugers Arch. ges. Physiol., 1954, 259: 56-78. WALBERG, F. and BRODAL, A. L. Pyramidal tract fibres from temporal and occipital lobes. Brain, 1953, 76: 491-508. WALL,P., RBMOND, A. G. and DOESON,R. L. Studies on the mechanism of the action of visual afferents on motor cortex excitability. Electroenceph. clin. Neurophysiol., 1953, 5 : 385-393. WEISS,T. and FIFKOVA, E. The effect of spreading cortical depression on activity of subcortical relay nuclei of specific afferent pathways. Arch. int. Physiol., 1961a, 69: 69-78. E. Evoked response in the mesencephalictegmentum during cortical spreading WEISS,T. and FIFKOVA, depression. Physiol. bohemoslov., 1961b, 10: 117-121.
322
P. BUSER et
al.
WIDBN,L. and AJMONE MARSAN,C. Effects of corticopetal and corticofugal impulses upon single elements of the dorso-lateral geniculate nucleus. Exp. Neurol., 1960, 2: 468-502. WINOKUR, G. L., TRUFANT, S. A , , KING,R. B. and O’LEARY, J. L. Thalamocortical activity during spreading depression. Electroenceph. clin. Neurophysiol., 1950, 2: 79-90.
DISCUSSION A. ZANCHETTI : 1 would like to comment onDr. Buser’s interesting finding that reflex pyramidal responses to light or sound can still be obtained after ablation of the specific sensory areas, though being markedly influenced by either excitation or depression of the specific sensory cortices. Does the persistence of a pyramidal discharge to photic or sound stimuli after ablation of the primary sensory areas implicate that higher intensity flashes or clicks have to be used in these conditions to discharge the motor cortex? If this is the case, a study of stimulus intensity-response amplitude curves with and without the sensory cortices in place might offer more quantitative information on the actual importance of the role played by the specific sensory cortices in the reflex activation of the motor cortex.
C. AJMONE MARSAN: I wish Dr. Buser would outline in greater detail which among his very interesting findings were observed only in the chloralose preparation and which could also be obtained in non-chloralose preparations. In the course of his presentation he has stated that most of the differences between the two types of preparation were of a purely quantitative nature but most of the illustrations were from experiments performed on animals under chloralose anaesthesia. A clearer statement on the matter appears to me rather important, particularly in order to make it possible to determine which phenomena can be considered of purely physiological significance and which should instead be considered to be of a pathological nature. For instance, it is well known that in certain human subjects (affected by epilepsy) a flash of light can produce violent jerking movements. In analogy one would still accept all Dr. Buser’s findings even if they were obtained exclusively i n chloralose animals but one would be cautious in using them in the interpretation of normal sensory-motor mechanisms and only consider them as a very valuable tool for the understanding of the pathophysiology of, for instance, reflex epilepsy.
W. GREYWALTER: Could Dr. Buser clarify whether he observed an afferent discharge in ventral horn cells evoked by cranial nerve stimulation after total neo-corticectomy? If so, this might recall the myoclonic jerks seen in certain stages of sub-acute encephalitis, in which neuropathological studies have shown marked cell loss in the cortex. Is it possible that chloralose acts as a de-corticant in itself? I feel that pharmacological analysis of “amplification” should be considered as a preliminary to instrumental analysis, which often does need some a priori assumption about what sort of phenomena are to be expected. Finally I was not able to estimate the latencies of these effects from the slides but, from our observations on human intracerebral responses, the time relations and latencies might provide valuable information about the course and nature of the pathways.
R. GRANIT: Has Dr. Buser considered the visuo-motor path discovered by RBmond and Wall?
DISCUSSION
323
With regard to chloralose the work by Haase and Van Der Meulen a t the Nobel Institute has shown this drug to weaken recurrent inhibition which may, in part at least, account for the curious state in which the chloralose animals are both anaesthetised and excitable at one and the same time. A. G. and DOBSON, R. L. Studies on the mechanism of the action of visual WALL,P. D., REMOND, afferents on motor cortex excitability. EEG Clin. Neurophysiol., 1953, 5 : 385-393. J. and VANDERMEULEN, J. P. Effects of supraspinal stimulation on Renshaw cells belonging HAASE, to extensor motoneurones. J . Neurophysiol., 1961, 24: 5 10-520.
P. BUSER’s replies To A. Zanchetti It has not yet been possible, in our experiments, to determine to what extent the threshold for pyramidal responses t o visual or auditory stimulation is changed after complete recovery from a posterior decortication. In observations on animals chronically deprived of these areas, it is true that no significant diminution of sensitivity has occurred. However, owing to the extreme fragility and variability of these processes of reverberating activation of the pyramidal tract, such an approach, although of incontestable theoretical value, has necessarily to take account of contingent parameters, which may be capable of masking the systematic variations Dr. Zanchetti is interested in.
To C . Ajmone Marson Dr. Ajmone Marsan has undoubtedly stressed out one of the major criticisms which may be raised by our conclusion. One might of course wonder (1) if results obtained under chloralose anaesthesia can be extended to other experimental conditions; (2) how far our experiments may actually represent a proper way of studying “sensorimotor integration”. (1) As to the first point, it might be stressed that a certain number of data reported here have been as well obtained on curarized animals, with only local anesthesia. There are following: Pyramidal and radicular discharges to sound and light, pyramidal and radicular discharges to cortical stimulation; specific cortical effects on sensory responses in medial thalamus and reticular formation; finally, cortical influences on pyramidal discharges to sensory stimulations. The only group of facts which have not yet been confirmed is the cortical specific influence on radicular discharges (which means practically, the reinforcement or appearance of such discharges by strychninization of one sensory area). Thus, except this latter case, there is a striking and even unexpected qualitative similarity of result obtained with chloralose or on unanaesthetized animals. (2) Regarding the second point - assimilation of clonic discharges to a normal process of sensorimotor reverberation, I am perfectly ready to consider that our study is merely that of a model. The question may be then raised, whether our data could be applied t o a more physiological phenomenon than myoclonia, namely the startle reactions, which also represent general motor responses to abrupt stimuli. To W. Grey Walter In animals deprived of their neocortex, we could indeed occasionally observe efferent discharges to cephalic stimulations. We know very little about the actual mode of action of chloralose. From our data, it does not seem that jerking could be attributed to a decorticant effect of the drug, but rather to a more general “facilitatory” (or desinhibitory) action upon neuronal systems a t several levels (including cortex). One could assume that in the primary cortical fields “desinhibition” could well lead t o increase or even to determine the triggering of a descending outflow by the ascending sensory volley. But this is merely hypothetical. To R . Granit Much could be said on the problem of the pathway for light to the motor cortex, whether this regards the results to which Dr. Granit just alluded (Wall et 01. 1953) or many others (Gastaut and Hunter, 1951; Hunter and lngvar 1955). All seem to agree (in accordance with our results) as to the possibility of a cortico-subcortico-corticalcircuit. Only Hunter and Ingvar postulate in addition the
324
P. BUSER et a[.
existence of a purely cortico-cortical route (a point which our results neither exclude nor demonstrate). As regards the subcortical relay of the afferent pathways to the motor cortex, Wall et al. suggested participation of the pretectal area, in accordance with the anatomical findings (Barris et al. 1935) whereas Hunter and Ingvar insisted on the importance of medioventral thalamic structures. Our own experimental data tend to support the latter suggestion. BARRIS, R. W., INGRAM, W. R. and RANSON, S. W. Optic connections of the diencephalon and midbrain of the cat. J. comp. Neurol., 1935, 62: 117-153. H. and HUNTER, J. An experimental study of the mechanism of photic activation in idioGASTAUT, pathic epilepsy. Electroenceph. clin. Neurophysiol., 1950, 2 : 263-287. HUNTER, J. and INGVAR, D. H. Pathways mediating metrazol induced irradiation of visual impulses. Electroenceph. clin. Neurophysiol., 1955, 7 : 39-60. A. G. and DOBSON, R. L. Studies on the mechanism of the action of visual WALL,P. D., REMOND, afferents on motor cortex excitability. Electroenceph. clin. Neurophysiol., 1953, 5 : 385-393.
New Data on the Specific Character of Ascending Activations P. K. ANOKHIN Moscow (U.S.S.R.)
As soon as the physiological peculiarities of the reticular structure of the brain stem were discovered, neurophysiologists were forced to examine a number of new problems of cortical-subcortical relationship (Moruzzi and Magoun 1949 ; Magoun 1950; Jasper 1949). The most important of these problems was how to qualify the new form of “arousal” spreading through the centrencephalic system, the physiological properties of which greatly differed from the arousal effect spreading through the classical lemniscal system. There were weighty reasons for setting these problems. Our established concepts on the spreading of excitations through the central nervous system, formed before the discovery of the physiological properties of the reticular structure, proved inconsistent. We supposed that excitations, generated in response to external or internal stimuli, extend from this initial point, frequently according to a linear principle, over an ever expanding territory. It seemed evident that the sources of energy required for this spreading process were drawn directly from successing excitations in every consecutive point of the central iiervous system. This concept, according to which primary excitation is self-supplied with energy along the whole line of its extension, has been proved to be inadequate. The discovery of the physiological specificity of the reticular structure of the brain stem has had a particularly important influence on our conception of the associative activity of the cerebral cortex. According to this conception the presence in the cortex of “two points of excitation” caused by an excitation of the lemniscal system, seemed by itself to be quite sufficient to make the associative connection between them. However, the simple fact that evoked potentials were obtained in the state of narcosis (Derbyshire et al. 1936) was the first actual proof of the inadequacy of this conception, which was founded on the idea of the existence of one lemniscal-thalamic system of excitation. This deduction could have been made even a t that early stage, but it was not then made. The research conducted by Moruzzi and Magoun and the systematic research later made by Jasper and other scientists showed that the cortical effect of external stimulation is a more complicated process than neurophysiologists had considered it to be. This was the origin of the first classification of ascending activations into “specific” References p . 338-339
P. K . ANOKHIN
326
and “nonspecific” (Magoun 1950, 1958). The most characteristic symptoms of nonspecific activations were: (a) lack of a specific single sensory modality, (b) activation of the EEG (desynchronization) and (c) generalized spreading over all the cortex. This first classification was based on the assumption that the “nonspecific” nature of activation is a universal and homogeneous property of every ascending activation of the cortex. In short, it was assumed that nonspecific activation, produced by the ascending excitations from the reticular structure, is the same for a11 kinds of cortical activation manifested by the desynchronization of its electrical activity. And all the types of generalized desynchronization of cortical activity were naturally admitted to be also equal, not differing from each other in any of their physiological qualities. The systematic research carried out in our laboratory showed that this concept of a single type of activating influence on the cortex was likewise inadequate and required considerable broadening. The very first experiments made by Agafonov, who applied pain activation under urethane narcosis, discovered certain new aspects of the ascending activating influences on the cortex. He obtained, by nociceptive stimulation of the hind limb of a rabbit several unexpected results (Agafonov 1956). Although the animal was actually in a narcotic sleep, its electric activity showed a sharp activation (desynchronization) of cortical electrical activity (Fig. 1). b
a
I
_-
..
-L C
H S.M
0.
Fig. 1 The desynchronizing effect of the nociceptive stimulus under urethane. (a) Sleep state under urethane; (b) the nociceptive stimulation (hot water) of a hind limb; (c) absence of desynchronizationfollowing chlorpromazine (aminazine) injection. Abbr. : SM, sensorimotor cortex; 0, occipital cortex.
It was evident to us that this was the first indication that urethane, as a narcotic, possesses special properties. This proved to be the chance observation which made us doubt, as early as 1953, the physiological truth of the conception that ascending activation has a certain universality and a single form. I shall try to reproduce the logical thread of the argument which later led to a series of specially directed experiments.
SPECIFIC CHARACTER OF ASCENDING ACTIVATIONS
321
In the first place we were forced to admit, on the strength of numerous experiments that had been by that time published, that the state of wakefulness is, in itself, an activated state of the cortex, which gets this activation from the rostra1 part of the reticular structure of the neuraxis (Magoun 1950). At the same time, every new stimulus which exerts its influence in the wakeful state immediately causes supplementary activation in the form of a sharply defined generalized desynchronization of electrical activity in the wakeful state (orienting-investigatory reaction). But, in an animal or man who is awake, pain stimulation also causes the sharp desynchronization of cortical electrical activity, because of the ascending activation which reaches it (Fig. 2). We have, therefore, three types of ascending activating influence on the cortex, all of them manifested by the generalized desynchronization of cortical electrical activity, and all differing only in the degree of activation. If all these three types of activation had the same physiological quality and the same nervous substratum, we ought to infer that urethane must, as a narcotic, also block all these types of activation. But, as we have seen, it depresses the activation of wakefulness and the activation of tentative-experimental reaction, but leaves the activation by painful stimulation intact. So that we had, in this case, a diyerent chemical specificity of the nervous substratum of three different activating influences on the cerebral cortex.
SmR
Siren1
4 sec
1
I
Fig. 2a Rcfercncer p . 338-339
I
328
P. K. ANOKHIN
TC'
Mth
Bell L
Fig. 2b
(a) The absence of desynchronization to a strong auditory stimulus under urethane. Abbr. : SML,
left sensorimotor cortex; OL, left occipital cortex; SMR, right sensorimotor cortex; OR, right occipital cortex; CMR, right centrum medianum. (b) The orienting-investigatory response under aminazine (chlorpromazine). A marked desynchronization of electrical activity of the cerebral cortex is seen within 1 sec after application of the conditioned alimentary stimulus (bell). Abbr.: SM, sensorimotor cortex; TC, temporal cortex; OC, occipital cortex; Mth, intralaminar thalamic nuclei; LTh, lateral thalamic nuclei; RF, reticular formation of the brain stem, ECG, electrocardiogram.
What was the actual substance of this selective action of urethane on the different activating mechanisms of the subcortical apparatus? The most likely supposition seemed to be that this selective sensitivity to urethane was related to the special chemical nature of the diflerent biological complexes of the subcortical structures: the hypothalamus and the reticular structure. On the strength of this conclusion we turned our attention to comparative studies of ascending activations of the cortex with definitely different biological origins. Such activities, differing as to their biological qualities, are well known to research workers. They are: the conditioned reflexes to food and defence against pain (Fig. 3). These experiments have been reported at several international conferences (Anokhin 1959, 1960) and in a special report delivered at the Pavlov session devoted to higher nervous activity, which was arranged jointly by the Academy of Medical Sciences of the U.S.S.R. and the New York Academy of Sciences (Anokhin 1961). The numerous experiments of my colleagues, A. Shumilina, V. GavliEek, I. Zatchinayeva, Y. Makarov and others, have shown that chlorpromazine (aminazine) in definite doses has a selective blocking effect only on the defensive states and defensive conditioned reflexes of the rabbit, leaving the conditioned food reflexes practically unchanged. Chlorpromazine correspondingly blocks defensive ascending activation and leaves
SPECIFIC CHARACTER OF ASCENDING ACTIVATIONS
329
food activation free to manifest itself (Shumilina 1956; GavliEek 1958, 1959; Zatchinayeva 1960; Makarov 1960). So that here we are faced, in principle, with the same phenomenon of selective influence as that observed when urethane is used, but in a demonstratively reciprocal aspect : chlorpromazine blocks selectively the mechanisms of ascending painful activation, leaving the mechanisms of wakefulness and food activation untouched.
s cs us T
Fig. 3 The selective effect of aminazine on the conditioned defensive reflex. (a) The animal’s response to the conditioned alimentary stimulus in the room in which the stimulus has always been reinforced by food. (b) The same animal’s response to the conditioned defensive stimulus in the room in which it has always been reinforced by electrical stimulation. (c) The same animal’s response to the same conditioned defensive stimulus 30 min after chlorpromazine injection. Abbr. : Sen. Mot. C.r., right sensorimotor cortex: Sen. Mot C.I., left sensorimotor cortex; Oc. Cr., right occipital cortex; Oc. C.1., left occipital cortex; Tem. C.r., right temporal cortex; Tem. C.1., left temporal cortex. Signals, from above downwards: S, salivation in drops; CS, conditioned stimulus; US, unconditioned stimulus; T, time in seconds.
We conducted experiments to ascertain the level at which chlorpromazine begins to block pain activation selectively. We adopted, as a characteristic symptom of pain activation of the cortex, the rise of blood pressure which always accompanies it and sharp changes of the respiratory rhythm. The experiments were made on decerebrated cats (according to Sherrington). The rise of blood pressure and quickened breathing References I. 338-339
330
P. K . ANOKHIN
due to painful stimuli disappeared immediately after an injection of chlorpromazine in the decerebrated cat (Fig. 4). These experiments showed that, from a specific and chemical point of view, the substratum blocked by chlorpromazine lies in the rostra1 part of the reticular structure of the brain stem. a
b
Af.I...I. rerg.
saliv.
n.ischiad.
-
- 1 sac
Fig. 4 The blocking effect of aminazine (chlorpromazine) on the painful sciatic stimulation of the decerebrate animal. ( a ) The change in the respiratory curve before the injection of aminazine; (b) the absence of any change in response to the stimulus of the same strength after the injection of aminazine. Abbr. : n. ischiad. sciatic stimulation; n. ling., lingual stimulation; resp., respiration: saliv., salivation.
The absence of conditioned defence activation (desynchronization) on the cortical surface, after the injection of chlorpromazine, must therefore be interpreted as the result of loss of ascending activation from some substratum which is sensitive to chlorpromazine and insensitive to urethane. All these facts, as well as the conclusions to be drawn from them, have led us to formulate the conception that every activating influence on the cortex is specific. However this activation is speciJc in relation to the biological quality of the given activity in general (food, defence, etc.) and not in relation to some sensory modality. (Fig. 5 ) . This specificity is particularly apparent in the pharmacological dissociation of different types of biological activity in animals. We recently obtained an additional demonstrative example of the biological specificity of ascending activating influences on the cortex in the experiments of our colleague, V. Sudakov. It iswell known that in a cat under narcosis (withurethane or nembutal) all the areas of the cortex show the slow electrical activity typical of the somnolent state. But if the cat has been deprived of food for some days before the experiment, the picture of electrical activity of its cortex undergoes a marked change.
33 1
SPECIFIC CHARACTER OF ASCENDING ACTlVATlONS C
i ~
~~
Decerebrate cat 5-transection of t h e spinal cord
Fig. 5 The illustration of the inhibitory effect of the injection of arninazine (chlorpromazine) on reciprocal interaction on the spinal cord level. (a)Reciprocal relationships before arninazine (chlorpromazine) injection. (6) Reciprocal relationships following after the injection of aminazine. Inhibition of both phases of reciprocal relationships is shown. (c) The recovery of reciprocal relationships after the subbulbar transection. The experiment shows that aminazine has disinhibited the inhibitory centres of the medulla by blocking the adrenergic substratum of the brain stem. The specific sensitivity of the brain stem formations to aminazine is thus emphasized. Abbr.: A, indicates aminazine (chlorpromazine) injection; S, subbulbar transection.
While the temporal, parietal and occipital areas of the cortex show the usual slow activity, the frontal areas of the cortex, on the contrary, show a clearly different activity of the “arousal” or desynchronized rapid type (Fig. 6). FR..
4
7 - - -w dub*y
..
FL A‘7,-
“\:.Z
-0.
.
.
Fig. 6 The selective activating effect of hunger on the electrical activity of the cortical frontal lobes. (a) Electrical activity of the cerebral cortex of the cat fed before the experiment. Slow activity in the frontal lobes is displayed. (b) Cortical electrical activity of the cat deprived of food for two days before the experiments. Referencas 0. 338-339
332
P. K. ANOKHlN
The fact that this special activity of the frontal area is connected with the state of food deprivation and is not blocked by a narcotic substance, i.e. is manifested in the somnolent state, turns our attention to this activity which requires a thorough experimental analysis. We argue as follows: if this activity is the result of excitation of the food centre of the hypothalamus by “hungry blood” (Anand and Brobeck 1960; Anderson 1960), then the change of this “hungry blood” into “satiated blood” must lead to the appearance of the usual slow activity in the frontal areas. “Satiation” of a starved cat under urethane narcosis was accomplished by two methods: the introduction of milk through the mouthand oesophagus into the stomach, with particularly marked irrigation of the mouth cavity, and the intravenous injection of a glucose solution. These methods were intended to imitate as nearly as possible the natural process of satiation which is, as we have shown, composed of several factors (Anokhin 1962). Our expectations proved to be correct. Immediately after the process of “satiation”, the activated state of the frontal areas of the cortex changed into the slow electrical activity characteristic of the somnolent state. Although this slow activity was not entirely similar to the electrical activity of the other areas of the cortex, it differed radically from the activity corresponding to the condition of food deprivation (Fig. 7).
Background activity ilntroduction of food
‘
(POMinutes la;‘&
I 60Minutes later
Fig. I Recovery of slow electrical activity of the frontal lobes after artificial “satiation” of the cat.
We must stress that the whole process of “satiation”, i.e. the process of replacing the electrical activity of the starved cat by the electrical activity of the “satiated” cat, developed under narcosis, on the background of a deeply somnolent state. These experiments leave no doubt that the independence of ascending activations of hunger in the frontal areas of the cortex from the somnolent state and narcotic blockade has a profound biological meaning. It is probably this physiological peculiarity of the activation described above that wakes the hungry animal from sleep and sends it in search of food. We evidently see it also in the behaviour of the infant who wakes only to receive food. Now we have interesting additional facts. As is well known from our previous experiments, urethane anaesthesia does not block the painful activation of the cortex. On the contrary, chlorpromazine blocks that activation even under urethane anaes-
333
SPECIFIC CHARACTER OF ASCENDING ACTIVATIONS
FR
-
-
q-%+-Jv%~*%vq~q~$~
FL
--
-
J
Sm R
0.4
*
~
J
~
$
~
!
* i,
b , p & & J ~ M ~ ~ 7:&! 4f4hJqJp)/ I
Srn L
Fig. 8 (a) Background electrical activity of the cerebral cortex of the hungry animal. (b) Recovery of slow
electrical activity of the frontal lobes following anodic polarization of the hypothalamic regions. Abbr. (Figs. 6, 7, 8): FR, right frontal lobe; FL, left frontal lobe; OR, right occipital cortex; OL, left occipital cortex; SmR, right sensorimotor cortex; SmL, left sensorimotor cortex; PR, right parietal cortex; PL, left parietal cortex; HpMR, right medial hypothalamus; HpML, left medial hypothalamus.
Our recent experiments have shown that chlorpromazine blocks selectively painful activation only, but does not at all affect hunger activation of the frontal lobes under urethane anaesthesia (Fig. 9). This new evidence shows that in this particular instance we also have several special qualities of ascending activations and their selective action on the cortical cells. It is reasonable to assume that all the ascending influences mentioned above spreading from different subcortical energetic points of different functional systems converge to the same cortical neuron, but only through different and specific synaptic References P. 338-339
334
B. K . ANOKHIN
Sm.R. % ------
Srn.L.
- , ‘I
1 #
1 soc
FR
d
C
” QL
1sec
0 3 ~
Fig. 9 Arguments for assuming a different chemical nature of the subcortical substrate which provides mobilization of either defensive or alimentary cortical connections. (a) The electroencephalogram in a hungry state. Activation of the cortical frontal lobes. (b) Theactivation effect of the nociceptive stimulus on all the cortical areas before aminazine (chlorpromazine) injection. (c) The blocking and activating effects of aminazine on the nociceptive stimulation (0.3 V). (d) Activation of the frontal lobes under hungry state is not blocked by aminazine. Abbr.: FR, right frontal lobe; FL, left frontal lobe; SmR, right sensorimotor cortex; SmL, left sensorimotor cortex; PR, right parietal cortex; PL, left parietal cortex; OR, right occipital cortex; OL, left occipital cortex.
SPECIFIC CHARACTER OF ASCENDING ACTIVATIONS
335
connections. However, this particular problem remains to be studied at some future time. The example of local hunger activation of the cortex stresses the physiological nature of the ascending influences on the cerebral cortex and confirms our conception of specific and selective ascending influences on the cortex. From this point of view, any complicated form of activity of the whole organism which has a biological quality (the “functional system” according to our terminology) forms a single restricted cortical-subcortical system, which is specific for that given pattern of activity. We are convinced by experimental material that the biological quality and purely energetic unity of the whole cortical-subcortical system is maintained by the subcortical structures. In other words: it is a system whose individual links have a heterogeneous physiological importance. Our experiments provide sufficient ground for concluding that every corticalsubcortical system of that kind is also heterogeneous in respect of the selective chemical sensitivity of its ind‘vidual links in relation to the activity of different pharmacological substances. The fact that two decisive physiologicalproperties of each of these systems of activation of cortical activity and highly selective sensitivity to chemical agents, belong to the same subcortical links of the system, is a remarkable feature of these cortical-subcortical systems. This was clearly shown, for example, in the selective blocking of cortical pain desynchronization by the intravenous injection of chlorpromazine or in the selective activation by the hypothalamus of the frontal areas of the cortex i.e. the state of hunger. Chlorpromazine blocks the activity of the subcortical links of the extensive cortical-subcortical system, which have an activating importance for all the remaining parts of the system, so that, naturally, the whole system of cortical-subcortical interaction which maintains the state of fear is disintegrated or depressed. It follows that, to eliminate the activity of any widely branched functional system of the organism, integrated on the level of cortical-subcortical interaction, it is not at all necessary to block all the links or all the components of the system. The very nature of the brain’s integrative activity is such that it is itself sufficient to eliminate (block, depress) the links of the system which maintain the energetic unity of the functional system ; and this functional system, with all its definite physiological or biological qualities, will then cease to exist as a whole. There is much experimental evidence for this point of view (see Fig. 8). Moreover, any general conception of the nature of a phenomenon is always more acceptable if it explains in the most plausible way the factual material collected at that time. Our point of view on the physiological peculiarities of specific and selective ascending influences allows us to explain several physiological phenomena. We can take, as an example, the selective therapeutic effect of “psychotropic” substances, especially the “tranquillizers”. There are abundant experimental and clinical data to prove that substances having a psychotropic effect have an exceptional specific and selective action on definite forms of psychical diseases (Himwich 1957; References P. 338-339
336
P. K . ANOKHIN
Uhr and Miller 1960). We are trying to work out an explanation of the physiological mechanism of their action on various psychic states. It is hardly likely, for instance, that the basic point of this explanation must be the effect of these substances on the elementaryprocesses ofexcitation and inhibition in the nervous system(Marrazzi 1961). There is no doubt that any drugs introduced into the bloodstream will eventually, influence the processes of excitation and inhibition. But this alone does not explain why these substances should influence both these processes developing, for example, in a given psychic state, nor why that psychic state is eliminated selectively as a whole in all its numerous components and systems of excitation and inhibition. As I consider the factual evidence relative to this problem, I become more and more convinced that all psychopharmacological effects can be explained in a much more acceptable way on the basis of the architectural principle, i.e. by the properties of a widely branched functional system with components which are heterogeneous from a chemical and physiological point of view. From this point of view we may consider that the state of anxiety or fear is caused by the existence of a selectively organized system of nervous connections between the cortex and the subcortex. This system is built in such a way that a great many associative connections occur, mainly in the cortex. On the other hand, energetic facilitation, which insures the selective contact between many elements of the cortex, that is to say, practically the maintenance of this interconnected branching system in the integrated and dominant state, is undoubtedly caused by the corresponding subcortical structures (“Emotions, the source of power of the cortical cells”, according to Pavlov). It is therefore quite natural that every dominating system of relationship disappears as soon as its chemically most vulnerable link is blocked. It is vulnerable because of its maximal energetic potencies and consequently its intense metabolic processes. Figuratively speaking, by means of the selective blockade of the most energetic and vulnerable links of the functional system, the tranquillizer “pulls out” the given functional system from numerous integrative systems of the whole brain, leaving other functional systems in a more or less normal state. In several special cases these last non-blocked functional systems may appear in very accentuated expression, for instance, in the appearance of greedy eating after the injection of chlorpromazine and after the disappearance of permanent anxiety (“release phenomenon”) (GavliEek 1958). The psychopharmacological effect is not the only phenomenon in the activity of the whole brain which can be easily explained on the strength of our conceptions of the specific character of the ascending activating influences on the cortex. The differentiation of the conditioned reflexes of a different biological nature are undoubtedly formed and defined on the same principles. Suffice it to say, that the complete elimination of the defence system and its cortical superstructures and the simultaneous maintenance of clearly manifested food reactions after the injection of chlorpromazine is a sufficient example of this law (GavliEek 1958; Anokhin 1961). There is one point in the conception mentioned above which cannot yet be precisely
SPECiFIC CHARACTER OF ASCENDING ACTIVATIONS
337
established. On it depends the further understanding of the mechanisms of realization of ascending activating influences at the level of the cortical cells. In examining the mechanisms of the selective spreading of the excitations over all the cortical synapses, we always bore in mind the well established fact that the primary selective chemical effect of pharmacological drugs is exerted on the subcortical structures phylogenetically differentiated to chemical sensitivity. We admit that the excitation originated here by specific chemical trigger mechanisms spreads in an ascending direction up to and including the cortical nervous elements. However, an important question arises: is there any chemical difference in the synaptic endings of the cortical cells and in the subsynaptic membranes of cells responsible for the realization of the ascending activations of diflerent biological origin? And is there no correspondence between the chemical qualities of the synapses of the cortical and subcortical level in the same functional system? We cannot answer these questions yet; but the answer must certainly be very important for the understanding of the integrative activity of the whole brain. It is possible to note only that a considerable part of the ascending synaptic contacts on the cortical level are undoubtedly formed in the prenatal or early postnatal period on purely morphological principles, consolidated by heredity (Purpura 1960; Ata-Muradova 1960). Their chemical properties are undoubtedly determined also by heredity and we may therefore admit, on phylogenetic principles, a certain chemical peculiarity of these synaptic contacts. The great variety of these chemical processes in the synapses has been recently shown in a comparative physiological aspect by H. Grundfest (1961). Very detailed material has recently been provided by Bullock (1959), who showed that the membrane of the same cell is not wholly homogeneous, On the contrary, it has, both chemically and physically, very varied properties in different areas of the cell body (Bullock 1959). For example, some sections of the membrane have a very low threshold of sensitivity and can therefore be the generators of nervous impulses, while other points d o not possess these properties. There are also grounds for speaking of the different chemical properties of different synapses. The division of synapses into “depolarizing” and “hyperpolarizing” is only a first step towards this chemical differentiation. In this connection I should like to draw your attention to a very interesting observation made in our laboratory and closely connected with the question under discussion. It is thought that gamma aminobutyric acid (GABA) blocks the depolarizing synaptic connections and leaves unchanged the hyperpolarizing ones. This classification is based on the purely physical signs of the nervous processes. It is assumed that physical features are only one sign of the state of the nerve cell membrane. However,some observations on the action of GABA during the ontogenesis lead us to doubt whether all the depolarizing synapses have the same chemical properties. We have shown that the usual negative component of the primary evoked potential after application of GABA really disappears, but that a new negative component with longer latency immediately appears (Fig. 10). Thus actually we have two negative potentials which are identical from the electrical References P 338-339
338
P. K . ANOKHIN
point of view only, but are quite different in their metabolic basis. That is probably because the first negative potential is blocked by GABA, while the second one escapes from its chemical action. This problem will, in any event, be studied experimentally with the greatest care. Should it appear that individual synapses of the cortex have a chemical variety which
lmin
15min
Fig. 10 An illustration of a different metabolic basis for evoked potentials having the same negative sign. (a) The evoked potential before the administration of GABA. Apredominant negative component of the primary evoked potential is seen. (b) Depression of the negative component of the primary potential and the appearance of a new negative component with longer latency (f) after the administration of GABA. (c) The relationship of the primary and secondary negative potentials 1 min after the administration of GABA. A remarkable augmentation of the secondary negative potential with depression of the primary negative potential is visible. ( d ) After 15 min, the secondary negative potential has increased.
corresponds in some degree to the chemical properties of the subcortical synapses, the whole study of the problem of specificity and selectiveness of ascending activations would lead us in an extremely interesting direction. REFERENCES AGAFONOV, V. G. Inhibitory action of chlorpromazine on effects of painful stimulation. Zh. Nevropat. Psikhiat., 1956, 56: 94103. ANAND, B. K. and BROBECK, J. R. Yule i.Bioi. Med., 1951,24: 123. Cited by E. STELLAR, Drive and Motivation. In: Handbook of Physiology, Vol. III, Sect. I (Neurophysiology), 1960: 1501-1 527. ANDERSON, B. and MCCANN, S. M. Acta physiol. scand., 1955, 35: 191, 333. Cited by E. STELLAR, Drive and Motivation. In: Handbook of Physiology, Vol. I l l , Sect. Z (Neurophysiology), 1960: 1501-1527. ANOKHIN, P. K. Electroencephalographic analysis of cortico-subcortical relations in positive and negative conditioned reactions. Pavlovian Conference on Higher Nervous Activity. Ann. N. Y. Acad. Sci., 1961,92: 899-938. ANOKHIN, P. K. Zh. vyssh. nerv. Deyut. (J. Higher Nervous Activity), 1962: in press. ATA-MURADOVA, F. On development of the activating effect of the reticular formation in the postnatal period. In : Evolution of the Physiological Functions. U.S.S.R. Academy of Medical Sciences, Moscow, 1960: 122-129. BULLOCK, T. Reviews of modern physics. Biophysical Sciences, 1959,31: 259. DERBYSHIRE, A. D., RAMPEL, B., FORBES, A. and LAMBERT, E. F. The effect of anesthetics on action potential in cerebral cortex of the cat. Amer. J. Physiol., 1936,116: 577-592. GAVLICEK, V. Electroencephalographic characteristics of conditioned defensive dominant state. J. Physiol. U.S.S.R., 1958, 44: 305-316.
SPECIFIC CHARACTER OF ASCENDING ACTIVATIONS
339
G A V L I ~ EV.K ,Effect of aminazine upon dominance of conditioned defensive attitude. J. Physiol. U.S.S.R., 1959, 45: 938-946. GRUNDFEST, H. Pavlovian Conference on Higher Nervous Activity. Part 111, Deviance and Drugs, Discussion. Ann. N . Y. Acad. Sci., 1961, 92. HIMWICH, H. Tranquillizing Drugs. Symposium Amer. Ass. for the Advancement of Science. Washington, 1957. JASPER, H. H. Diffuse projection systems. The integrative action of the thalamic reticular system. Electroenceph. elin. Neurophysiol., 1949,l: 405420. MAGOUN, H. W. Caudal and cephalic influences of the brain stem reticular formation. Physiol. Revs., 1950, 30: 459414. MAGOUN, H. W. The waking Brain. Thomas, Springfield, Ill., 1958. MAKAROV, Y. A. The role of the reticular formation in establishment of the defensive alimentary dominant. Proceedings First Scientific ConJ on the Physiology, Morphology and Pharmacology of the Reticular Formation. Moscow, 1960: 15-16. MARRAZZI,A. Inhibition as a determinant of synaptic and behavioural patterns. Pavlovain Conference on Higher Nervous Activity. Ann. N.Y. Acad. Sci. 1961,92: 990-1003. MORUZZI, G. and MAGOUN, H. W. Brain stem reticular formation and activation of EEG. Electroenceph. clin Neurophysiol., 1949, I : 455-413. PURPURA, D. P. Morphological basis of elementary evoked response patterns in the neocortex of the new-born cat. Pavlovian Conference on Higher Nervous Activity, Ann. N.Y. Acad. Sci., 1961, 92: 849-859. PURPURA, D. P., CARMICHAEL, M. W. and HOUSPIAN, E. M. Physiological and anatomical studies of development of superficial axodendritic synaptic pathways in neocortex. Exp. Neurol., 1960,2: 324. SHUMILINA, A. I. Characteristics of conditioned reflex activity of the dog under aminazine. Zh. Nevropat. Psikhiat., 1956,56: 116-121. UHR,L. and MILLER,J. Drugs and Behaviour. John Wiley and Sons, New York, London, 1960. ZATCHINAYEVA, I. A. The electroencephalographic characteristics of establishment of the conditioned alimentary positive and inhibitory reflexes against the background of aminazine application. Proceedings First Scientific C o d . on the Physiology, Morphology and Pharmacology of the Reticular Formation. Moscow, 1960: 50-52.
The Characteristics and Origin of Voluntary Movements in Higher Vertebrates I. S. BERITASHVILI (BERITOFF) Institute of Physiology, Georgian Academy of Sciences, Tbilisi (U. S.S.R.)
A correct comprehension of voluntary movements is one of the most interesting and most complicated problems in biology and medicine. This problem is a psychophysiological one. Its solution can be achieved only by a combination of physiological and psychological methods of investigation. Until recently this problem has been studied in men by psychologists. But I intend to solve it by means of observations and experiments on higher vertebrates, such as cats and dogs. I am just as sure as I. M. Sechenov was many years ago, that the mind, in the form of images or recollections, is the main controlling mechanism of purposeful behaviour in higher vertebrates and children (1903). Therefore, not only in men, but also in higher vertebrates, adaptation to the environment is primarily voluntary in character. For a long time I have been engaged in studying the psycho-nervous activity of higher vertebrates as it is revealed in alimentary and defensive behaviour directed by images of vitally important objects of the environment. Strictly speaking these behavioural acts directed by images may be called voluntary acts. In adult men voluntary movements are not merely directed by images of vitally important objects of the environment, but also produced as results of definite conscious aims planned to satisfy personal or social needs. This is accomplished by means of verbal planning of work activities. These conscious aims, which arise in the form of images or ideas, direct each active behavioural act of men. Such purely human voluntary movements are not considered in this paper. Before discussing our main problem -the characteristics and origin of voluntary movements -I will consider briefly our point of view about the origin of the image of the environment, its structural and psychophysiological basis, and the production of different behavioural acts according to images. On the basis of known structural, physiological and clinical data concerning the cerebral cortex, it has been suggested that the stellate neurones with pericellular axon nets, situated in the 1V layer of the primary focal area of the analyzer, may be the main perceiving elements of the cortex. When they are excited they may evoke elementary differentiated sensations of light, colour, sound, touch, pressure, warmth, cold, etc. (Beritoff 1960). According to this hypothesis, the perception of external objects and the creation of images is due to the functional integration of the large complex of sensory stellate
VOLUNTARY MOVEMENTS I N HIGHER VERTEBRATES
341
cells by means of internuncial and association pyramidal neurones which are mainly located in and between the secondary areas. Reproduction of the image of a given object is evoked by the action of any component of the object or of the environment where it has been perceived. When the image is projected into the environment, the animal, in response to the image, reacts just as it would react to the action of the external world itself. It is proposed that the functional system which reproduces images of the external world consists of sensory and non-sensory stellate and internuncial pyramidal neurones connected with cortical projection neurones, which serve as the starting mechanism of orientating movements of the eyes and the head. When this system is excited, it produces the image and orientating reaction towards the imagined object, thus regulating the behaviour of the animal in the presence of corresponding emotional excitation in a manner similar to that which takes place when the external world is directly perceived (Beritoff 1960). We may state that the image of the environment, with all its vitally important objects, is a substitute for the real environment. This is its main biological significance. With this introduction, we may begin to analyze voluntary movements. In higher vertebrates voluntary movement consists of purposeful movements of the animal in space from one object to another and of so-called instrumental movements of the forepaw for the purpose of getting food or repulsing the enemy. In all these cases the movements are directed in the familiar environment by images of vitally important objects projected into this environment. The image of the food location reproduced under the influence of the familiar environment causes the hungry animal to run immediately to the appropriate place. If this place is far away or is hidden and is therefore not seen, the animal reaches this place just the same, avoiding all obstacles on its way. The animal goes to the vitally important object, the food, not only according to the image of the food location, but also according to the image of the qualities of the food. If images of bread in one place and of meat in another place are involved, at first the animal runs to the place where the meat is, because it prefers meat as food; only then does it go to the place where the bread is. It follows from this fact that the voluntary movements directed by images of food are governed, not only by a projection into the environment of images of food as such, but also by the projection of the qualities of the food (Beritoff et al. 1934, 1935). I have studied different voluntary behavioural acts of various animals under different experimental conditions. But in order to be more concise in this paper, we must consider only some of the results obtained during the investigation of food behaviour in dogs. In these experiments the vision of the dogs was eliminated by means of a light-proof mask put on the head. Vision was eliminated because it was easier to investigate the behavioural acts directed by images in the absence of permanent visual perception of the environment. A blindfolded dog in a familiar experimental situation comes out of its cage in response to a food signal and goes straight towards a definite food-box, puts its forepaws on it, pushes its head through a hole and eats a given amount of food. Then it turns around and goes back into the cage. It runs to and from the food-box by the References P. 348
342
I . S. BERITASHVILI
shortest route, even in the absence of the food signal, that is to say, spontaneously, when the food-box was closed. Moreover, the blindfolded dog may go to the place where food is from any point of the given familiar room and then return into the cage or go to another place where food is. It is evident that this food behaviour is ofan orientating nature and that it is purposefully directed to definite objects. Each new dog, when it is led into a given experimental room for the first time, behaves differently. It goes round all persons and objects in the room, touches them, sniffs at them and even licks them. This is the usual inborn behaviour of dogs in an unusual environment. It is directed to the search of vitally important objects. When we studied the alimentary behaviour of dogs in an environment familiar to them, we concluded that this behaviour is based on the creation of the images of food and of its location. These images seem to be produced in the dog, not only when the eyes are open, but even when the animal is blindfolded, i.e. in the absence of vision. We have established that during the movements of the animal the vestibular receptors are stimulated. The impulses originating there arrive in the cortex and activate the vestibular analyzer (the anterior third of the ectosylvian gyrus). During each movement of the head a definite group of stellate neurones with a pericellular axonal net becomes excited under the action of the afferent volleys originating from the vestibular receptors. We think that these stellate neurones may be the sensory cells which produce sensations, i.e. subjective experiences. In the vestibular analyzer these are revealed as sensations of to and fro, up and down movements and as sensations of turning to the right and to the left. During the movements of the animal which result from the integration of these sensations, the whole route covered is perceived with all its turning points precisely projected into the environment (Beritoff 1953, 1959). The perception of the route covered leaves its traces after the cessation of the vestibular stimulation. The traces are retained for a long time in the form of images projected into the environment. Thus, if the blindfolded animal is led forwards for 3 or 4 meters and is then turned to the right at a right angle and led for 2 meters to the place where food is and if, after it has eaten the food, the animal is returned by the same route, the image of the whole covered route of the food itself and of its location arises in the animal. This is evident from the fact that the animal can when it is blindfolded, go to the place where the food is, not only by the same route, but even by the shortest route, although it has never taken this shortest route (Beritoff 1959). Obviously, the perception of the experimental situation has become associated with the perception of the food location in the form of an integrated image. We propose that such an image may be created because of the integration of the sensory stellate neurones excited by the entire experimental situation and then by the vestibular stimulation produced during the movement of the animal in the given situation. If the animal moving in the room touches some objects or sniffs at them, appropriate tactile and olfactory images of these objects arise and are appropriately projected into space; they are projected to those points where they were perceived by various receptors when the animal moved.
VOLUNTARY MOVEMENTS IN HIGHER VERTEBRATES
343
When it leaves the familiar place in the given environment, the aniinal can determine its location in this environment with the help of the vestibular perception. When, therefore, the blindfolded animal moves in the familiar room, it turns its head towards the food-box when the image of this food-box is reproduced. Such an important role of the vestibular perception is confirmed by the fact that, after bilateral labyrinthectomy, the blindfolded animal cannot create the image of the route covered; it cannot localize and project into the environment the images of the objects perceived on the way (Beritoff 1953, 1959). Once it has been created, the image of the spatial location of the objects remains for a long time, for many days or even months. This image is easily reproduced both under the influence of the food signal and under the influence of different stimuli originating from the familiar food situation. The reproduction of the image of the food location in the hungry animal evokes the orienting movements of the head towards the food-box. The animal moves towards it even though its eyes are covered. The animal moves to the place where the food is as if it sees this place. The drive to the food is an inborn property of the nervous mechanism of the hungry animal. Evidently the purposeful movement of the hungry animal to the place where the food is, is a function of the following: in the hungry animal the chemical composition of the blood is changed in such a way that the excitability of motor mechanisms is enhanced, so that they become readily activated under the influence of various external and internal stimuli. The animal begins to produce searching movements and to sniff at all objects. The direction of the movement is always determined by the direction of orienting head movement. When the animal sees the food, or when the image of the food location is reproduced, it turns its head towards the location. The searching movements are also directed towards it. When the animal becomes satiated, the composition of the blood changes in the opposite direction. The excitability of the motor mechanisms decreases. The reduction of the excitability of the motor mechanisms results in the cessation of the searching movements. The animal lies down and begins to doze. It is obvious that, when this happens, the sight of the place where the food is, or the reproduction of its image, do not produce the movement towards it. Moreover, even the orienting head movements may not appear. Thus, the reproduction of the food location evokes voluntary movement towards it only when the animal is hungry, i.e. when the animal needs the food and performs the searching movements in order to get it. The movements resulting from the reproduced image of the food location are not the only voluntary ones. All other movements of the animal may be voluntary. Thus, when the animal reaches the food-box, it lifts its forepaws in order to put them on the food-box; but the blindfolded animal also does the same when it approaches the projected food-box and when the food-box has been removed. This means that the animal senses the distance covered and thus determines its own location in relation to the food-box. Because the animal at this moment lowers its head and sniffs at the food-box as if it sees the food-box, we may be sure that under the influence of vestibular volleys of a definite character and duration the same psycho-nervous processes References P. 348
344
I. S . BERITASHVILI
are reproduced as those which were evoked when the animal was led for the first time to the food-box and was fed there. When the food-box is closed after eating, the normal dog turns back. This happens because the image of the closed or empty food-box arises. Simultaneously the psychonervous process which projects the location of the cage is reproduced, because, after the food-box has been closed, the dog has been usually led to the cage. If the dog comes to the food-box and it does not open, it stays there for a while and then turns back. In this instance, also, the animal turns back because the image of the closed food-box arises and simultaneously the image of the cage is reproduced. Turning back, the animal stretches its neck towards the projected image of the cage. This is followed by the movement of the whole animal by means of the same physiological mechanism which has led the animal to the food-box. Usually the normal dog when it comes back, even when it is blindfolded, passes the door of the cage. But, if the dog walks past the cage, it begins immediately to perform searching movements with its head and sniffs at the objects. As soon as it touches the cage, it finds the door and enters the cage. Evidently, this touching reproduces thepsycho-nervous processes which project thelocation of the door of the given cage. All the movements of the blindfolded animal mentioned above, i.e. forward movements towards the food-box, lifting of the forepaws, lowering of the head, search for the opening of the food-box, returning at last to the cage and finding its door, all these movements are directed by the images of the route, of the food-box and of the cage, projected to definite points. These movements are therefore voluntary acts. As has been mentioned in our previous investigations the food behaviour is voluntary in the beginning; it is accomplished by means of the above-mentioned psycho-nervous processes. But the psycho-nervous behaviour repeated in the same situation acquires the characteristics of a conditioned chain reflex, in which each previous segment of the behaviour - strictly speaking the whole complex of external and internal stimuli acting in this segment of the behaviour - becomes a conditional signal for the following segment of the behaviour. The beginning of this chain reflex is set up by the food signal which evokes the conditioned orienting movement of the head towards the food-box and then the conditioned coming out of the cage. Of course, at this time the image of the food-box, placed in the given room: is also reproduced. But in the conditioned chain reflex it is not of great importance, because this reflex is the quickest and the most economical act of the organism (Beritoff 1947, 1959). Consequently such a habitual movement is not voluntary, because it proceeds according to the “stimulus-response” principle. The involuntary nature of the automatized behaviour becomes apparent when the food-box is moved in the sight of the dog. In response to the food signal, the dog runs to the place where the food-box usually is, sniffs there at the floor and only then does it go to the new situation in which the food-box has been placed. This may happen even if the food-box is moved together with the signal (as a rule, the signal, the bell, is near this food-box). All the same, in response ot the first trial of the signal, the dog runs to the place where the food-box usually is (Beritoff et al. 1934). Obviously, the dog runs to the food-box irrespective of the image of the food-box, because the
VOLUNTARY MOVEMENTS IN HIGHER VERTEBRATES
345
food-box has been moved in its sight and therefore its image should be projected to the new place. The dog performs the voluntary movement according to the image, when it has not found the food in the usual place: it turns away and goes to a new place. Before the automatization of the food behaviour the dog would act otherwise. If the food signal has been moved, the dog would go not to the food-box, but towards the sound. It would approach the sounding bell, sniff at it and even lick it (Beritoff et al. 1934). This is because the image of the food is projected to the sound source. Until the food behaviour becomes automatized in a form of conditioned chain reflex, the dog moves according to this image. When the conditioned chain reflex cannot be carried out because the environmental situation has been changed, the reproduction of the image of the food-box turns the automatized involuntary behaviour into a voluntary one. Thus, if an obstacle is placed in the way of the given automatized behaviour, e.g. if the food-box is screened, the conditioned chain reflex is arrested. The animal stops for a moment and then begins to move. This movement is voluntary. It corresponds to the image of the food-box projected behind the obstacle. Directed by this image, the animal passes round the obstacle and then turns correctly to the food-box. The roundabout movement itself is a habit acquired by the animal duringits past life, a result of individual experience. The blindfolded animal, too, running against an obstacle, passes round it. At first it goes back some steps and turns to one or another side at an angle of 90". Then it goes along the obstacle turning to it now and again and touching it. Finally, this turning leads the animal to the opposite side of the obstacle. But all the time the animal perceives its spatial location by means of the vestibular analyzer and projects into the room the image of the location of the food-box. Therefore, when it passes round the obstacle, the animal turns its head correctly towards the given food-box and moves to it. The mechanical stimulation of the skin of the head and the vestibular stimulation play the main part in this roundabout movement. The skin stimulation serves as a conditioned signal of moving away from the obstacle and turning at an angle of 90°, while the labyrinthine impulses help to determine the distance between the animal and the obstacle (Beritoff 1959). But in general this passing round the obstacle in order to reach the food-box is a voluntary movement, because it is directed by the image of the location of the food-box. Passing round the obstacle is not the only method used by the animal to overcome the obstacle. Walking along the obstacle, the animal tries to get through apertures in it and, if the obstacle is low enough, it jumps over it. All these movements are voluntary (Beritoff et al. 1934). If the obstacle is not low enough and if there are no appropriate apertures, the animal walks round the obstacle. The animals use all these habitual movements for overcoming the obstacle, because the image of the food is projected behind the obstacle. If the image of the food were not reproduced behind the obstacle, the animal would walk along it without trying to overcome it. If the blindfolded animal has several times met the same obstacle at the same point on its way to the food-box, it may freely pass round the obstacle, touching it once or even without touching it at all. The animal begins to swing off the direct way at some References
p.
348
346
I. S. BERITASHVLLI
distance from the obstacle; it then walks round it for some distance and goes directly towards the food-box. This roundabout movement is evoked by the image of the food location, created by means of the vestibular perception of the distance covered to the obstacle and the turning poitits around it. It permits the animal to imagine the location of the obstacle and that of the food-box, as well as its own location in the room. If the obstacle is imperceptibly removed, the animal performs the same roundabout movement. This shows that passing round the obstacle at some distance depends, mainly, upon the reproduction of the image of its location, which is created on the basis of the vestibular perception. As we have established, blindfolded labyrinthectomized animals are incapable of walking round the obstacle at some distance (Beritoff 1953). However, the animal may pass round a new obstacle at some distance without running against it. This happens when the animal reaches the obstacle slowly. At approximately a distance of 20-30 cm the animal begins to walk round it. But this happens if the obstacle is solid and if its height reaches the level of the head. If the new obstacle is not solid, e.g. if it is a grating, the dogs and the cats do not walk round it at a distance; they do this only if they touch the obstacle. We have established that the dogs and the cats perceive the solid objects because the air waves reflected from such objects stimulate the auditory receptors. Each time the blindfolded animal runs against the obstacle, it feels pain. Subsequently the animal begins to walk round the obstacle. When the animal comes near to the obstacle, the action of the reflected air waves is experienced. This action is subliminal. But as Gershuni et a/. (1947, 1948) have shown by experiments performed on men, the conditioned reflexes may be easily established in response to such subliminal auditory stimuli. Therefore, if the animal has run against the obstacles several times such reflected waves become a conditional signal for walking round the obstacle (Beritoff 1959). These conditioned reflexes, as well as many others established during individual life, are used by the animal in its voluntary movements, i.e. in the movements directed by the images of vitally important objects. We have considered in detail the origin of the voluntary movements in relation to images of vitally important objects which are projected into the environment on the basis of the vestibular perception. But it is clear that the origin of the voluntary movement based on visual images is just the same. If a piece of meat is shown to an animal locked in a cage, and if this piece of meat is hidden behind a screen, and if a piece of bread is shown from another place and is hidden there behind a screen, the visual image of food and its location undoubtedly arises in the animal. If after some time the cage is unlocked, the dog runs at first to the place of the meat, because it prefers meat, and only then does it go towards the place where the bread is. This happens even when the food is removed immediately after it has been shown. Evidently the given movement of the dog towards the place where the food is, is voluntary and is a function of the psycho-nervous activity which produces the visual images of meat and bread and of their location. Very often our dogs and cats attempt to pull in the food with their paws. If the animal is placed on one side of a grating and the food is placed at a distance on the other
VOLUNTARY MOVEMENTS IN HIGHER VERTEBRATES
347
side, the animal tries to push its forepaws through the grating and to bring the food nearer. If the food-box is closed when the dog or cat comes t o it, the animal tries to open the box with its forepaws. These so-called instrumental movements are produced even if the animal is blindfolded. The blindfolded animal also performs these instrumental movements when it smells food placed behind the grating, or when it touches the food-box. Undoubtedly, these movements are voluntary. They serve the purpose of getting the food according to the image projected not far from the animal. When the animal sees the food, or according to its earlier experience, projects the image of it to a distance of more than 1 meter, it does not perform the instrumental movements, but runs to the place where the food is and when it is near this place it begins to use its forepaws. The instrumental movements themselves, as well as the locomotor ones, are inborn reflexes. The difference between the two movements lies in the fact that the sensorimotor cortex is necessaryforthe instrumental movements, but is not essential for locomotion. During the visual or the olfactory perception of near-by food, our dogs and cats perform instumental movements according to the principle of conditioned reflexes. The appearance and the smell of the near-by food evoke these movements of the forepaws. But, even in this case the image of the food plays an important part: it is projected to a definite point in the environment and therefore the paws move purposefully to it. Actually, in adult animals, instrumental movements are always voluntary, because the animals recognize food and its qualitative peculiarities and that is why they bring it closer with their forepaws. CONCLUSION
Under the term “voluntary movements” we imply behavioural acts which are directed by images of vitally important objects in the environment. All the receptors participate in the creation of these images, but their projection into the environment is due to visual, vestibular and auditory perception. Head turning and movements towards the food-box when the image of the presence of food in the food-box is reproduced, as well as moving away from the food-box because of the arousal of the image of the empty food-box when the food has been already eaten, are such voluntary acts. Returning to the usual place (e.g. the cage) from any point of the familiar environment in accordance with the reproduced image of this usual place, is also voluntary. The behavioural act of refusal to go to such a food-box or food location at which the animal has just eaten the whole food is voluntary, too, because in this instance the image of the empty location is reproduced. All the so-called instrumental movements of the forepaws, by means of which the animal gets the food according to the reproduced image of the near-by food, are voluntary as well. The image of the environment, with all its vitally important objects, substitutes for the real environment. This is its main biological significance. The animal can produce expedient movements towards the vitally important objects, not only because of the Kefercncer p . 348
348
1. S . BERITASHVILI
immediate action of these objects upon its receptors, but also because of the reproduced images of these objects. During each act of voluntary behaviour directed by images, the animal utilizes both the inborn and the conditioned reflex nervous mechanisms. After repetition of the same voluntary behaviour in the same environment it becomes a habitual automatized act which proceeds more economically according to the principle of conditioned chain reflexes. This means that each segment of this behaviour is evoked as a conditioned reflex, because all previous external and internal stimuli have become conditioned signals for this segment. During automatized conditioned reflex behaviour, the image of appropriate vitally important objects is also produced, but it directs behaviour only in such cases when the environment is suddenly changed and the usual course of the conditioned reflex behaviour is disturbed.
REFERENCES BERITOFF, I. S. Characteristic forms of behaviour of higher vertebrates. Sovetsk. Psichonevrol. (MOSCOW), 1933, 9 : 55-67. BERITOFF, 1. S . Basic forms of nervous and psychonervous activity. Acad. Sci. U.S.S.R., MoscowLeningrad, 1947. BERITOFF, I. S. The role of vestibular and kinesthetic stimuli for the orientation of animals in the environment. Trans. Inst. Physiol., Georgian Acad. Sci.,1953, 9 : 3-24. BERITOFF, I. S. Spatial projections of percepted objects in external environment by means of labyrinthine receptors. J. Physiol. U.S.S.R., 1957, 43: 600-610. BERITOFF (BERITAsHvm), I. S. Nervous mechanisms of spatial orientation in higher vertebrates. Acad. Sci. Georgian S.S.R.,Tbilisi, 1959. BERITOFF, I. S. Physiological significance of neurones of the cerebral cortex. Arkh. Anat. Gistol. Embriol., 1960, 39: 3-38. BERITOFF, I. S. On the psychical activity of the cerebral cortex of the higher vertebrates. Stud. Cercef. Neurol., 1960, 4 : 567-578. BERITOFF, I. S . et a/. A study of individual behaviour in dogs. 10 communications. J. Physiol. U.S.S.R., 1934,17: 176-183, 912-920, 1186-1197; 1935,18: 385-397; 19: 41-51,496-507. GERSHUNI, G . V. A study of subsensory reactions during the activity of sensory organs. J. Physiol. U.S.S.R., 1947, 33: 393412. GERSHUNI, G . V., KOZHEVNIKOVA, V. A., MARUSEVA, A. M. and CHISTOVICH, A. M. On the peculiarities of formation of temporary connections upon imperceptible sound stimuli in man. Byull. iksp. Biol. Med., 1948,26: 206-209. SECHENOV, I. M. Elements of thought. Izbr. filosof. proizv., Moscow, 1947: 398-537.
Responses in Non-Specific Systems as Studied by Averaging Techniques* MARY A. B. BRAZIER Brain Research Institute, University of California, Los Angeles and Massachusetts Institute of Technology, Cambridge, Mass. (U.S.A.)
Within the wider framework of an interest in what the apparent multiplicity of afferent systems in the brain could signify in terms of information-carrying function, we began some years ago to explore the electrophysiological routes through the midline thalamus. The word “information” is being used here, not in its vernacular form, meaning descriptive content, but as used in Information Theory. Information is conveyed by the selection of an item by virtue of its probability of occurrence when compared with a multitude of other possible items. Because of a suspicion that, in the unanesthetized animal, many electric signs of transmission were going undetected among on-going activity of higher amplitude, we resorted to computer techniques that increase the signal-to-noise ratio, i.e., that emphasize the potential changes evoked by and time-locked to a stimulus, the unrelated background activity becoming averaged out by the process. These techniques have been described in earlier publications (Brazier 1960b, 1961b; Brazier et al. 1961) and will therefore be outlined only briefly here. In this report, discussion will be restricted to some of the results that have emerged from studies in the midline region close to the transition between midbrain and thalamus. In some early experiments in which we were using a computer to detect evoked responses among unrelated activity of higher amplitude, we noticed that electrodes, in the region usually delineated as the centre median (CM), recorded a bimodal response to flash. This work was reported at the symposium on the reticular formation held in Detroit in 1957, and at that time we suggested that the double response might imply that impulses from the single flash have reached this nucleus by more than one pathway (Brazier 1958). That the CM responded to electric shock to optic nerve was shown in 1952 by Prench et al. in the monkey. Although in the unanesthetized animal these responses are of low amplitude compared with the background activity, these authors were able to detect their presence in the tracings from an inkwriting oscillograph in animals acutely prepared under ether and recorded from when unanesthetized and immobilized with syncurine. Latencies and waveforms could not, of course, be defined at this
* The work reported here was supported by USPHS Grant B-3160 and Contract No. 233(69) from the Office of Naval Research, U.S. Navy. References P. 364-366
350
M. A . B. BRAZIER
recording speed. lngvar and Hunter in 1955 mapped responses to flash in many parts of the midbrain including the CM. Inflow from the auditory system was detected electrophysiologically by French et al. in 1953. Typically the responses are of long duration in contrast with the classical responses of the specific thalamic nuclei. The recording of interaction between auditory and visual responses in the C M was reported last year by Albe-Fessard and Mallart (1960a,b) including interaction with somatic and corticofugally induced responses. The work of Meulders and Massion (1961a,b) and Massion and Meulders (1961) is also important in this context. The characteristics of the somatic afferent inflow to the CM have been known since the work of Magoun and McKinley (1942) and have been studied by several workers since (e.g., Collins and O'Leary 1954), and most recently in great detail by Albe-Fessard and her collaborators (AlbeFessard and Rougeul 1958; Albe-Fessard and Gillett 1958, 1961; Albe-Fessard 1960; Kruger and Albe-Fessard 1960; Albe-Fessard et al. 1961). METHOD
The animals used in our experiments were cats in which the operative procedures were carried out under deep pentobarbital anesthesia. Electrodes were inserted stereotactically and no recordings were made for 3-4 weeks after implantation. As a general rule each cat had two nine-pin plugs, i.e., eighteen recording points giving bipolar information from nine loci in any one cat. One cat in the series had only one plug inserted, one cat had three. For determination of latencies, each of these eighteen or more points was referred to an electrode pasted temporarily to the shaved skin at the back of the neck. A reference in the frontal sinus could not be used since the computer analysis always revealed a pick-up from the electroretinogram. The electrodes used for these chronic implantations were enameled stainless steel wires 1/100 in. in diameter, glued together with insulating material in pairs, side by side, to give rigidity when directed by the stereotactic instrument. Recording points either 0.5 mm, 1 mm, or 1.5 mm apart in the vertical direction were used according to the goal of the particular experiment, and these were freed of insulation at the tips. The animals were tested in a sound-reducing box (actually an old refrigerator) lined with reflecting metal foil and with one end replaced by transparent lucite. For all tests their pupils were atropinized. The stroboscope was outside the lucite wall and was itself encased in a sound-reducing box with a second diffusing screen forming an airlock in front of the light, in order to reduce faint sound made by each flash. Control computer tests for responses to the faint click of the stroboscope, audible outside the testing box, revealed none in the auditory cortex of these cats when the light was obscured by a dark cloth. A standard calibrated level of white noise was maintained at all times as the background against which both flash and click stimuli were given. The clicks, also standardized for intensity, were delivered through a small loudspeaker fixed inside the test box. The original recordings were made on frequency modulated tape and simultaneously monitored on an inkwriting oscillograph. This monitoring is essential for ensuring
351
RESPONSES IN NON-SPECIFIC SYSTEMS
that the state of the animal is being maintained at a constant level of alertness, since drowsiness and sleep introduce changes in response. Photographs of single responses were taken with a Polaroid camera from a cathode ray oscilloscope. On one channel of a 7-channel magnetic tape was recorded a pulse, synchronous with the incidence of the stimulus that was being delivered to the animal. The computers, of which three types have been used in this work, have been described by their originators (Barlow 1957; Mitchell and Olsen 1957; Clark 1958; Communications Biophysics Group and Siebert 1959), and all achieve the same general goal, namely: by a process whichcan be regarded as a special case of crosscorrelation, the average waveform of those potential changes that are time-locked to the stimulus is increased at the expense of those that have no such temporal relationship. The stimuli used in the experiments to be reported were flash and click, for it seemed desirable to avoid the even more unphysiological stimulation by electric shock that, by synchronizing the afferent volley, may well force synapses that are impenetrable in the normal state of the brain. RESULTS
When the region generally outlined as the centre mCdian was explored, it was found that not all points within the usually depicted limits gave the multiple response to flash noticed in our earlier work, the stereotactic placement at which it was most prominent corresponding to the Horsley-Clarke co-ordinates F7, L3 and H 0.5 to H 1 (according to the Jasper-Ajmone Marsan Atlas, 1954). At this frontal plane, in the unanesthetized animal, the response to flash (which is illustrated in Fig. 1) usually consists of: 1. a relatively small negative-positive component followed by: 2. a large negative deflection usually beginning about 25-37 msec after the flash and rising to a plateau having a duration of about 40 msec; 3. a second large negative deflection appearing about 85-90 msec after the flash and also of long duration, the center of its plateau usually being about 100-125 msec from the stimulus. For convenience in referring to these components small numbers have been inserted in the curve for Cat B, by which they can be identified. The recordings illustrated are unipolar in order to facilitate assessment of the time course. The relative sizes of components No. 2 and No. 3 vary with electrode placement and from cat to cat. A small wave appearing between these two major negative deflections is sometimes seen (as, for example, in Cat C in Fig. I). With electrodes placed more anteriorly, i.e., in the region stimulation of which optimally evoked recruiting responses in the cortex, no response either to light or sound was found. Only one cat in a series of seventeen was an exception to this general finding. A possible explanation of this singular finding is that one point of the bipolar electrode may have been in one functional zone and the second in the other. The responses shown in Fig. 1 are from three different cats with electrodes at the optimal site described above. The wavelike form of the major components is suggestive of some serial barrage of inflowing impulses followed later by another such barrage. The suggestion also occurred to us that the second wave of activity in the
+
References p . 364-366
+
352
M. A. B. BRAZIER
centre mtdian (CM) might be corticofugal in origin, and we have followed with interest the work of Albe-Fessard and Gillett (1958, 1961) on responses evoked in the CM by cortical stimulation. In the course of some studies on the influence of several anesthetic agents on the RESPONSE TO FLASH IN CENTRE MEDIAN 240 RESPONSES AVERAGED ARC-I COMPUTER
CAT A I
L-r--t
,
,
I
I
20
40
60
80 100
2OOmsec
25
50
75
I00
250msac
25
I
1
50
75
100
0
I
I
CAT 8
I
0 CAI C
I
0
I 250msec
Fig. 1 Average of 240 responses to a l/sec light flash in three different cats. Note timescale for top graph differs from the other two. Three components are discussed in the text: 1. a small initial negativepositive diphasic component; 2. a large wavelike deflection, about 40 msec in duration; 3. a subsequent large wavelike deflection, also of long duration, appearing 85-90 msec after the flash. Negativity at the exploring electrode is recorded in this and all subsequent figures as an upward deflection. The recordings reproduced here are unipolar to assist assessment of latencies.
various subcortical centers, we have found striking differential effects and have suggested that light doses of these agents may be used to distinguish vulnerable pathways from the more resistant ones. It seemed therefore of interest to determine whether this technique could help to elucidate the afferent routes into the CM for impulses initiated by the retina, and whether these had differential sensitivities. In zones of overlapping function, where neurones serving one system may be scattered among others of different function, the electrolytic lesion cannot make the differentiation. In such cases, the pharmacological “lesion” may help to make the discrimination. Since the CM receives connections from the reticular formation (McLardy 1956; Nauta and Kuypers 1958) which is well known to be vulnerable to barbiturates, one of the drugs used was pentobarbital. A striking change was found with this drug. There was a marked augmentation of the previously rather inconspicuous diphasic first component (No. 1) and a very great augmentation of the first of the two slow negative waves. In contrast, there was complete disappearance of component No. 3, the late slow wave which, in the record from the unanesthetized animal, had its peak
RESPONSES IN NON-SPECIFIC SYSTEMS
353
at about 110 msec after the flash. This effect is illustrated in Fig. 2, both of which records are from a cat with implanted electrodes. So great was the increase in amplitude under pentobarbital that the amplification had to be reduced by half that used for the record taken without anesthesia. EFFECT OF PENTOBARBITAL ON RESPONSE TO FLASH IN CENTRE MEDIAN 240 REPONSES A V E R A G E D . ARC-I COMPUTER
0
25
50
75
100
250msec
0
25
50
75
100
250mrec
Fig. 2 Effect of pentobarbital on response to flash in the centre me'diun. The drug essentially abolished component No. 3 (the second major negative wave) but augmented the earlier negative wave so greatly that the amplification used had to be half that of the upper record. Component No. 1, the initial diphasic wave is also markedly augmented, both in its negative phase and in its subsequent positivity. Upper record: unanesthetized; lower record : with pentobarbital.
The lower record shown in Fig. 2 was run 1 h 45 min after a 20 mg/kg i.p. dose of pentobarbital when the level of anesthesia had become quite light. We have remarked in several earlier studies (Brazier 1954a, b, 1960a, 1961a, 1963) that the effect of this drug in very light doses is an apparent unleashing of a normally restrained response - a process suggestive of a removal by the drug of an inhibitory influence. Fig. 3 illustrates the same change of waveform in another cat whose late response, when unanesthetized, was less conspicuous. This result forcefully suggests that the two major negative responses (components No. 2 and No. 3) evoked by flash, travel by pathways differentially responsive to barbiturates. For comparison of time relationships, the CM response in the unanesthetized animal is shown in Fig. 4 together with the averaged unipolar responses recorded simultaneously from other sites. In this unanesthetized cat, the latencies are, when averaged: visual cortex 12 msec, superior colliculus 26 msec, CM 30 msec, reticular formation 45 msec. It should perhaps be noted that latencies to flash are not, of course, the same for all positions in the midbrain reticular formation. At histology, the electrodes in this cat, whose records are illustrated in Fig. 4, were found at sites corresponding to the Horsley-Clarke co-ordinates F3, L2, H-2.5, and -1.5, which is References P. 364-366
3 54
M . A. B. BRAZIER
EFFECT OF PENTOBARBITAL DFI RESPONSE TO FLASH IN CENTRE MEDIAN ?.W RLBWNSES AVCRABEO &Reel COMPUTE8
I
PENTOBARBITAL
Fig. 3 Another example of the effect of pentobarbital in a different cat whose late response when unanesthetized was less conspicuous than that shown in Fig. 2. The same change in waveform is apparent.
S I M U L T A N E O U S RESPONSES TO F L A S H IN V A R I O U S BRAIN CEFJTERS
JISLAL CORTEX
CENTRE MEDIAN
SUPERIOR CoLLlCULUe .-
-.
...
-_r' 200msec
RETICULAR FORMATION 1
0
1
(
1
7
1
20
40
60
80
100
Fig. 4 Simultaneous recordings, from four electrode positions, of averaged responses to flash in a n unanesthetized cat. Recordings are unipolar to facilitate comparison of time courses.
355
RESPONSES 1N NON-SPECIdIC SYSTEMS
caudal to the red nucleus and about 5 mm below the lower edge of the superior colliculus. The superior colliculus has a timing comparable to the earlier negative wave in the CM (i.e., component No. 2), but has no later negativity and hence would appear unrelated to the third component of the CM. The visual cortex, as can be seen in the same Fig. has, of course, experienced its primary potential before any of the subcortical centers pictured here have begun to respond. The later events in the visual cortex (Le., the complex of surface-negative waves, occurring between 30 and 80 msec after the flash) have a time course that might perhaps implicate them as a source for late corticofugal effects in the C M where evidence of a late and scattered barrage can be seen. As noted earlier, the late response in the CM becomes less conspicuous with more caudal positions of the electrodes. Electrodes placed just at the posterior edge of the C M register a markedly different response. This is illustrated in Fig. 5. In the unanesthetized cat the latency to the first wavelike deflection has been found to be 16-18 msec which coincides with the early diphasic component No. 1 seen more rostrally. This response is 8 or more msec earlier than that of the superior colliculus.
FA RESPONSE TO FLASH
240 Responses average
12.5 25 FLASH
I
I
50
75
I
ARC- I
100msec
Fig. 5 Response to flash at the caudal border of the CM.
Histological identification of the electrode tip for the record shown placed this at a position corresponding to co-ordinates F6.5, L3 and H 1.5. It thus seems reasonable to consider this first potential the sign of activity of pretectal fibers on the route of the pupillary reflex. It would not be surprising, then, that its latency bears no relationship to that of the superior colliculus which plays no part in this reflex (Ranson and Magoun 1933; Magoun 1935; Magoun et al. 1936). The second hump of the record shown in Fig. 5 has a similar time course to that of component No. 2 of the CM response recorded more rostrally, but the late wave (No. 3) is lacking. Fig. 6 illustrates, from a different cat, another response at this electrode position together with an example of the augmenting effect of a-chloralose. The amplification
+
Referencrr D. 364-360
356
M. A . B. BRAZIER
for the three graphs shown in Fig. 6 is the same for all. The response in the unanesthetized animal was less than 10 pV in amplitude and could not be detected except by averaging. Even when averaged, the response in the absence of the drug is so small that one cannot with certainty identify the components discussed so far. With chloralose, however, the early diphasic component No. 1 with its latency of 16 msec (in these bipolar recordings) is strikingly augmented, as is also component No. 2 (the wavelike negativity extending from 35-65 msec after the flash). AUGMENTATION OF RESPONSE BY Q CHLORALOSE I
I=
IOpV
CONTROL
INCREASING DEPTH OF ANESTHESIA
Fig. 6 Augmentation of response to flash by a-chloralose. Recording site at caudal border of CM. In the unanesthetized animal the response was less than 10 pV in amplitude and not detectable in the single trace. Bipolar recordings.
Responses at this electrode site are also augmented by pentobarbital as can be seen in Fig. 7, which depicts simultaneous recordings from visual cortex and midbrain and is introduced to illustrate two further points. One is the confirmation that component No. 1, whether or not one may identify it with pretectal activity, is extremely resistant to light barbiturate anesthesia and is, in fact, augmented by it. The second point is that the late Forbes response, so prominent in the visual cortex at this level of barbiturate anesthesia, cannot apparently be held responsible for late events in the CM, since these are abolished by the drug. A different computer has been used for the analyses in Fig. 7 - the Evoked Response Detector of Barlow (1957) -but again, the average waveform is depicted by the envelope of the pen deflections. The flash occurred at the first of the continuous series of deflections which are plotted here at 1 msec intervals. On moving to the experiments in which click was used as a stimulus, the finding
357
RESPONSES IN NON-SPECIFIC SYSTEMS
RESPONSES TO FLASH IN MIDBRAIN AND VISUAL CORTEX PENTOBARBITAL ANESTHESIA 150 RESPONSES AVERAGED ERD COMPUTEh
VISUAL CORTEX
I
4.-
M I DBRAlN I
..
,
O
,,
2'0 40
i
,
I
60
60
160msec
Fig. 7 Resistance of early components of the flash response (No. 1 and No. 2 ) to pentobarbital (lower record) at a level that abolishes the later wave in the CM and evokes a Forbes secondary discharge in the visual cortex {upper record). (ERD computer.)
by French et al. (1953) of a response in the CM was easily confirmed. What was perhaps unexpected was the very short latency - rarely longer than 8 msec, or about 4-5 msec after the response of the inferior colliculus. Our data do not enable us to say at this time whether the impulses have reached the CM from the inferior colliculus. Some examples of this are seen in Fig. 8, where unipolar recordings have been reproduced in order to indicate the time course. RESPONSES TO C L I C K I N CENTRE M E D I A N
I N 3 UNANESTHETIZED CATS (AVERAGE RESPONSES)
0 I 0 20
b
Ib 20
0 10 2 0 m s e c
Fig. 8 Examples, from three cats, of the response to click recorded at the same site in the CM that gives a multiple response to flash. References P. 364-366
358
M . A . B. BRAZIER
The time-relationships of the response to click in the CM to those in the inferior colliculus and auditory cortex (recorded simultaneously) are shown in Fig. 9. The very early compact response of the inferior colliculus is followed only a few milliseconds later by the beginning of the longer-drawn-out response of the CM. This is an unanesthetized cat and its response at the auditory cortex can be seen to have a latency of 8 msec. CLICK RESPONSES IN THE BRAIN AUDITORY CORTEX
240 RESPONSES AVERASED. ARGl COMPUTER
CENTRE MEDIAN
r
INFERIOR COLLICULUS
Ib
t
25
1
$0
f5
100
250 mmec
Fig. 9 Simultaneous recordings, from three sites, of the response to a click stimulus.
Because the same electrode placements record responses to both flash and click, though at greatly different latencies, it seemed of interest to see whether pairing these stimuli would indicate any interaction (Brazier 1961b). At the present stage of the work, only stimuli paired synchronously (externally to the animal) have been tried. Some typical results are shown in Fig. 10 and 11. In Fig. 10 with flash alone, the small diphasic first component precedes a large wavelike second component (most conspicuous between 30 and 65 msec) and has a long drawn-out third component which does not return to the baseline until 125 msec after the flash. The first diphasic component was obliterated by the early arrival in the CM of a click externally synchronous with flash. The amplitude of the first slow wave was also markedly cut, although the later components, lasting for about 65-1 10 msec after the flash, maintained their previous level of activity. In other words, such interaction as there was seemed to be with the components designated here as No. 1 and No. 2 of the CM response, and not with the later events.
359
RESPONSES IN NON-SPECIFIC SYSTEMS
EFFECT OF P A I R I N G C L I C K W I T H FLASH. CENTRE MEDIAN 2 4 0 RESPONSES AVERAGED
I 1. ,nu"
PAIRED C FLASH LICK
0
25
50
75
100 msec
Fig. 10 Effect of pairing click with flash recorded in the CM. The introduction of the click obliterates the small diphasic first component of the visual response and reduces the amplitude of the more conspicuous first slow wave, but leaves unchanged the level of the later long drawn-out component that does not return to the baseline until 125 msec after the flash.
In Fig. 11, from a different experiment, the response to click before pairing is shown, as well as the response to flash and the effect of pairing. Again, the late components (coming 65-150 msec after the stimuli) seem unaffected in comparison with the earlier response. E F F E C T O F PAlRlNG C L I C K AND F L A i H CENTRE M E D I A N
FLASH ALONE
d l PAIRED C L I C K A N D F L A S H
0
25
50
75
100
z5r
m
~
e
~
Fig. 11 Another example of the interaction of a click with the earlier components only of the CM response to flash. Hrferrnces p. 364-366
360
M. A. B. BRAZIER
The results at this present stage of our study would lead us to think, as mentioned above, that only a part of the region generally described as the centre me‘dian receives agerents from the visual and auditory systems. And, as others including Jasper et al. (1955) have found before us, only the anterior portion of the CM appears to have efferents which evoke a recruiting response at the cortex. In our experiments it is from this effective “recruiting” zone that we have failed to find responses to flash. There do seem, therefore, to be regions within the usually defined limits of the CM which have special properties that differentiate them functionally from their neighbours. Needless to say, we have been extremely interested in the recent reports from Albe-Fessard with her fine recordings of a duality of unit responses, both to light and to sound, in the CM (Albe-Fessard and Mallart 1960a,b). We all recognize the complexity of relationship and sometimes lack of relationship between unit responses and evoked responses recorded by a gross electrode. When one moves even further to the averaged response, the complexity is even more confounding. Moreover the unit discharge represents “output” of the neurone in its purest form, whereas the evoked response is a complex of “input” potentials (some of them electrotonic in nature), on a background of fluctuating excitatory level that may or may not result in effective output (i.e., in the discharge of cells). Our goal was not an anatomical study but, as stated at the opening of this paper, one that is groping towards an understanding of the information carried by functioning pathways within the brain that are ancillary to the classically known specific ones. The approach to developing concepts of what the brain may regard as information can be made in other ways. One may work on the assumption that the codes lie in the behavior of the single unit and that some massive computational analysis of these individual reports is made by the brain. One may also work from the assumption that it is the profile of activity in a population of neurones that is the determining factor. It is an exploration of the latter approach that is being made here. If the question: “How does the brain get its information?” is asked in the framework of Information Theory, it would be the dissimilar, the novel, that would carry the message. One would then conceive the pertinent question to be: “How, on the average, does the activity currently in the brain differ from what it has just been experiencing?” In reference to the above questions we have carried out experiments into this field of inquiry. We find that if we fractionate our analyses and, instead of taking the average of 240 1-per-second flashes, as in most of the examples shown so far, we take the average of the first 60 and of the second 60 and so on seriatim, we can detect a serial change of considerable significance in the responses. This is illustrated in Fig. 12 for the CM of a cat whose eyes have been atropinized. The serial change in the late component of the multiple response is very marked and, whatever the mechanism for this depressed responsiveness may prove to be, it is tempting to propose that this signals the information that the stimulus is still being received but is unchanged, i.e., the message is that the identical situation has been experienced before. Paripassu with the decrease of the late component is an augmentation and a conspicuous waxing and waning of the earlier events. Fig. 13 is another
RESPONSES IN NON-SPECIFIC SYSTEMS
361
example of this phenomenon. It will be noted that the time course of these changes in average waveform differs considerably from the much longer exposures generally used to detect habituation. As noted above, the waveform of the CM response to flash changes profoundly if the animal is barbiturized. Another feature of the anesthetized animal is that serial
Fig. 12 Failure of late wave response to flash in CM on continued stimulation. Note augmentation, with waxing and waning, of components No. 1 and No. 2 (the early diphasic and the first of the large negative waves).
analyses no longer show a seriatim change (Fig. 14). Information that the identical stimulus has been experienced before is no longer being carried, for each response or group of responses, is identical with previous ones. This procedure for conveying this category of information would thus appear to be differentiable even within the non-specific system since not all components of the CM react in a similar way. Non-specific systems are known to be more sensitive and more flexible in this respect than the classical specific ones. For example, in the specific References p . 364-366
362
M. A . B. BRAZIER
EFFECT OF CONTINUED IiSEC FLASH ON RESPONSE I N CENTRE MEDIAN AVERAGE OFI sf 60 RESPONSES
2nd 6 0
3 r d 60
4 t h 60
5 t h 60
6 t h 60
7th 60
8 t h 60
Fig. 13 Another example of the change in response recorded in the CM on continued stimulation by I/sec flash.
visual projection cortex it is the complex of waves that follows the initial component of the primary evoked potential that begins to fail on repetition (Fig. 15). These are the components that are dependent on the integrity of the midline systems that are so vulnerable to barbiturates. As can be seen from the averaged responses of the same cat when anesthetized with pentobarbital (Fig. 15, right-hand side), the message which may be proposed as conveying “change from the previous average situation” is no longer carried. Not only does the negativity of the primary response (increased in amplitude by release from inhibition at this level of barbiturate), show no change on repetition, but the late “secondary response” enhanced by the drug, also fails to denote that this is a recurring stimulus. Under the anesthetic, the information received is the same for the new experience, averaged over the first to the 60th flash, as for the 180th to the 240th.
RESPONSES IN NON-SPECIFIC SYSTEMS
363
RESPONSE OF CENTRE MEDIAN TO CONTINUOUS I/SEC FLASH: EFFECT OF BARBITURATE UNANESTHE' tZED
PENTOBARBITAL 2
AVERAGE OF RESPONSES
2 n d 60
3 r d 60
Fig. 14 The elimination by pentobarbital of the sequential changes in the CM induced by continued l/sec flash. Sixty responses averaged in each sample.
OF V I S U A L CORTEX TO C O N T l N U O U S I/SEC F L A S H . E F F E C T OF B A R B I T U R A T E
RESPONSE
UNANESTHETIZED
PENTOBARBITAL
AVERAGE OF RESPONSES 1st 60
2 n d 60
3 r d 60
4th 60
I FLASH Fig. 15 Left: Sequential changes in the late components of the visual cortex reponse to continued l/sec flash. Right: stereotyped response from same electrode site a t a level of pentobarbital anesthesia that releases the Forbes secondary discharge. Sixty responses averaged in each sample shown. References p . 364-366
3 64
M . A . B. BRAZIER
The results that are reported here are some that have been obtained by the approach mentioned above, namely an examination of the proposal that the profile of activity in an aggregate of neurones, by signalling their average behavior, carries information in the brain. It is not suggested that profile analysis can substitute for the analysis of units; it is the latter (and not only their discharges but also their electronic fluctuations) that form the underlying mechanism that we would all like to understand. The suggestion is that this probabilistic model may be complementary to a deterministic one as the hypothetical basis from which to design experiments, for the extremely subtle codes of firing pattern in single neurones have been shown by several workers to be intimately related to behavior (as studied by conditioning techniques). The remaining question would appear to be: how does the brain act on these individual reports? Does it average them to assess the probability that a significant number of them are giving an approximately similar report? Hopefully some answer may come from a combination of these two methods of attack on the problem of coding in the brain. ACKNOWLEDGEMENTS
Thanks are due to many who have helped in the work here. In particular, Miss Ruth Carpenter for her help with the experimental work, Mr. Mishell Stucki for operating the computers and Miss Margaret Carroll for the histological examination of the brains. SUMMARY
This report is concerned with part of a continuing program designed to explore the information-carrying functions of pathways within the brain that are ancillary to the classically known specific ones. The approach to developing concepts of what the brain may regard as information can be made in many ways. One may work on the assumption that the codes lie in the behavior of the single unit and that some massive computational analysis of these individual reports is made by the brain. One may also work from the assumption that the profile of activity in a population of neurones is a determining factor. It is an exploration of the latter approach that has been made in the work covered in this report. In order to obtain this profile of neuronal behavior, the raw experimental data have all been processed through computers for obtaining the average activity. The part of the non-specific system reported in this section of the work is the centre mkdiun of the thalamus, explored under different conditions in cats with implanted electrodes. Use has been made of certain drugs known to have differential effects on various brain pathways for a kind of “pharmacological dissection” of their function. REFERENCES ALBE-FESSARD, D. Rtponses globales ou unitaires observks dans le centre median du thalamus chez le chat kveillk. C. R. Acad. Sci. (Puris), 1960, 250: 2618-2620. ALBE-FESSARD, D. et GILLETT,E. Interactions au niveau du centre median entre les influx d’origine somesthtsique et d’origine corticale. Activitks unitaires. J. Physiol. (Paris), 1958, 50: 108-111.
RESPONSES IN NON-SPECIFIC SYSTEMS
365
ALBE-FESSARD, D. et GILLETT, E. Convergences vers le centre median. Electroenceph. clin. Neurophysiol., 1961, 13: 257-269. ALBE-FESSARD, D. et MALLART, A. Dualite des reponses du centre median a la stimulation visuelle. C. R. Acad. Sci. (Paris), 1960a, 251: 1191-1193. ALBE-FESSARD, D. et MALLART, A. Existence de reponses d’origine visuelle et auditive dans le centre median du thalamus du chat anesthesie au chloralose. C . R. Acad. Sci. (Paris), 1960b, 251: 1040-1042. ALBE-FESSARD, D. et ROUGEUL, A. Activites d’origine somesthesique Cvoquees sur le cortex nonspecifique du chat anesthesie au chloralose: r d e du centre median du thalamus. Electroenceph. clin. Neurophysiol., 1958, 10: 131-1 52. ALEE-FESSARD, D., MALLART, A. et ALEONARD, P. Reduction au cours du comportement attentif de l’amplitude des reponses evoquees dans le centre median du thalamus chez le chat tveille libre, porteur d‘electrodes a demeure. C . R. Acad. Sci. (Paris), 1961, 252: 187-189. BARLOW, J. S . An electronic method for detecting evoked responses of the brain and for reproducing their average waveform. Electroenceph. clin. Neurophysiol., 1957, 9 : 340-343. BRAZIER, M. A. B. Studies of electrical activity of the brain in relation to anesthesia. In H. ABRAHAMSON (Editor), Neuropharmacology. Josiah Macy Foundation, New York, 1954a: 107-144. BRAZIER, M. A. B. The action of anesthetics on the nervous system. In J. F. DELAFRESNAYE (Editor), Brain mechanisms and consciousness. Blackwell, Oxford, 1954b: 163-1 69. BRAZIER, M. A. B. Studies of responses evoked by flash in man and cat. In H. H. JASPER,L. D. PROCTOR, R. S. KNIGHTON, W. C. NOSHAY and R. T. COSTELLO (Editors), Reticular formation of the brain. Little, Brown and Co., Boston, 1958: 151-167. BRAZIER, M. A. B. Some actions of anesthetics on the nervous system. Fed. Proc., 1960a, 19: 626-628. BRAZIER, M. A. B. Some uses of computers in experimental neurology. Exp. Neurol., 1960b, 2: 123-143. BRAZIER, M. A. B. Some effects of anaesthesia on the brain. Brit. J. Anaesth., 1961a, 3 3 : 194-204. BRAZIER, M. A. B. Paired sensory modality stimulation studied by computer analysis. In Pavlovian Conference. Ann. N . Y. Acad. Sci., 1961b: 1054-1063. BRAZIER, M. A. B. The electrophysiological effects of barbiturates on the brain. In W. S.ROOT (Editors), Physiological pharmacology. Academic Press, New York, 1963. and E. G. HOFFMANN BRAZIER, M. A. B., KILLAM, K. F. and HANCE, A. J. The reactivity of the nervous system in the light of the past history of the organism. In: W. A. ROSENBLITH (Editor), Sensory communication. M. I. T. Press, J. Wiley and Sons, New York, 1961: 699-716. CLARK, W. A., BROWN, R. M., GOLDSTEIN, M. H., MOLNAR, C. E., OBRIEN,D. F. and ZEIMAN, H. E. The average response computer (ARC): A digital device for computing averages and amplitude and time histograms of electrophysiological response. IRE Trans. Biomed. Electronics, 1961, 8 : 46-5 1. COLLINSW. F. and O’LEARY, J. L. Study of a somatic evoked response of midbrain reticular substance. Electroenceph. clin. Neurophysiol., 1954, 6 : 619-628. COMMUNICATIONS BIOPHYSICS GROUPand SIEBERT, W. M. Processing neuroelectric data. Res. Lab. Electronics, Mass. Inst. Technol., Tech. Rep. 351, 1959, 121 p. FRENCH,J. D., VAN AMERONGEN, F. K. and MAGOUN, H. W. An activating system in brain stem of monkey. A.M.A. Arch. Neurol. Psychiat., 1952, 68: 577-590. FRENCH, J. D., VERZEANO, M. and MAGOUN, H. W. A neural basis for the anesthetic state. A.M.A. Arch. Neurol. Psychiat., 1953, 69: 519-529. INGVAR, D . H. and HUNTER, J. Influence of visual cortex on light impulses in the brain stem of the unanesthetized cat. Acta physiol. scand., 1955, 33: 194-218. JASPER, H . H. and AJMONEMARSAN, C. A stereotaxic atlas of the diencephalon of the cat. National Research Council, Ottawa, 1954. JASPER, H. H., NAQUET, R. and KING,E. E. Thalamocortical recruiting responses in sensory receiving areas in the cat. Electroenceph. clin. Neurophysiol., 1955, 7 : 99-1 14. KRUGER, L. and ALBE-FESSARD, D. Distribution of responses to somatic afferent stimuli in the diencephalon of the cat under chloralose anesthesia. Exp. Neurol., 1960, 2 : 442467. MAGOUN, H. W. Maintenance of the light reflex after destruction of the superior colliculus in the cat Amer. J. Physiol., 1935, 111: 91-98. MAGOUN, H. W. and MCKINLEY, W. A. The termination of ascending trigeminal and spinal tracts in the thalamus of the cat. Amer. J. Physiol., 1942, 137: 409416. MAGOUN, H . W., ATLAS,D., HARE,W. K. and RANSON, S . W. The afferent path of the pupillary light reflex in the monkey. Brain, 1936, 59: 234-249.
366
M. A. B. BRAZIER
MASSION,J. et MEULDERS, M. Les potentiels kvoques visuels et auditifs du centre mCdian et leurs modifications aprbs decortication. Arch. int. Physiol., 1961, 69: 26-29. MCLARDY, T. Some midbrain-thalamic long-fiber connections in the monkey. I n J. ARIENSKAPPERS (Editor), Progress in neurobiology. Elsevier, Amsterdam, 1956: 282-288. MEULDERS, M. et MASSION, J. Comparaison de I’amplitude des potentiels evoquks au niveau du centre median par une stimulation lumineuse chez le chat “cerveau isolc” et le chat “mkdiopontin pretrigiminal”. C . R. Acud. Sci. (Paris), 1961a, 252: 1209-1211. MEULDERS, M. et MASSION,J. Effet facilitateur de la lumiere continue sur les potentiels evoquCs au niveau du centre mkdian par la stimulation electrique de structures voisines du corps genouille latkral. Arch. int. Physiol., 1961b, 69: 407-409. MITCHELL, J. L. and OLSEN,K. H. TX-0, a transistor computer with a 256 X 256 memory. Proceedings Eastern joint Computer Conference, IRE, 1957: 93-101. NAUTA,W. J. H. and KUYPERS,G. J. M. Some ascending pathways in the brain stem reticular formation. In H. H. JASPER,L. D. PROCTOR,R. S. KNIGHTON,W. C. NOSHAYand R. T. COSTELLO (Editors), Reticular formution of the bruin. Little, Brown and Co., Boston, 1958: 1-30. RANSON,S. W. and MAGOUN,H. W. The central path of the pupilloconstrictor reflex in response to light. Arch. Neurol. Psychiut. (Chicugo), 1933, 30: I 193-1 204.
DISCUSSION D. ALBE-FESSARD : 1 am happy to see to what extent the results obtained by Dr. Brazier confirm our own findings in chronic animal preparations. 1. A striking feature: our responses are maximal at the same point on the stereotaxic instrument (anterior 7.5; lateral 3; H between 0 and 1). We always use these coordinates to ensure optimal responses to visual, somatic and auditory stimuli in chronic animals {without anaesthesia) or in animals under chloralose anaesthesia. 2. We ourselves have observed the responses of the centre me‘diun in chronic animals. These responses are marked only in inattentive animals, as clearly shown in Fig. 16 below, and in Fig. 9 of our report in this symposium. This fact explains why the average amplitudes of C M responses in aroused animals are not always as high as those obtained under chloralose anaesthesia. 3. In one respect our results differ: our frequency of stimulation was less high (one stimulus per 5 seconds in general). Undoubtedly this is one of the reasons for the lower amplitude of the average responses observed by you. 4. The second response which you observe, and which is also clearly visible in alert animals (see Fig. 1) might well correspond, as you suggest, with a cortical detour of afferents reaching cells not yet activated by the direct route. This second response is rare under chloralose (for with this anaesthetic the responses are maximal from the first afferent volley), but it is often seen under other anaesthetics.
M. A. B. BRAZIERTO D. ALBE-FESSARD: I would like, if I may, to ask a question of Mine. Albe-Fessard. In her experience has she found a site within the centre n@diunwhich gives responses to visual stimuli and also evokes recruitment on stimulation? D. ALBE-FESSARD TO M. A. B. BRAZIER: Yes, a great proportion of widespread units in the centre rn&liun respond to visual and auditory stimuli, as was shown by A . Mallart and myself. On the, other hand, we know that rhythmic stimulation of the C M evokes recruiting responses in the associative cortex (supra, Albe-Fessard and Fessard, p. 131, Fig. 15), and the best location for the stimulating electrode seems to be the anterior part of the CM. At any rate, there appears to be a zone of overlap for both properties.
DISCUSSION
367
ALBE-FESSARD, D. et MALLART, A. Existence de reponses d'origine visuelle et auditive dans le centre median du thalamus du Chat anesthesie au chloralose. C . R . Acud. Sci (Paris), 1960, 251: 10401042. W. R. ADEY:
Those of us who have been interested in problems of average response computing and automatic data analysis have all wondered about the possibility that certain difficulties may occur in the interpretation of the computed results. Since the brain is a volume conductor, there is the possibility that disturbances occurring at foci remote from the recording electrode may appear in the computed read-out. This possibility will increase in average-response computation as the number of averaged responses is raised. It is also more likely to occur with monopolar than with bipolar recording. Can Dr. Brazier indicate the steps taken in her experiments to assess these difficulties? F. BREMER: Our colleague attributes to the contribution of non-specific afferences the polyphasic character of the wellknown type of the response of the visual area to light. Without denying the possibility of this contribution for the very late phases of the response, it must be pointed out that a polyphasic outline very much like that of an evoked cortical potential already characterizes the response to
Fig. 16. median cortex 6s) Free. alert cat with in-dwelling - receotor . electrodes. roo tracing- of the suorasvlvian . (bipolar lead); bottom tracing in the centre me'diun (CM). Weak and invariable stimulation of the superficial radial nerve of the contralateral foreleg via in-dwellingelectrodes. The tracings (1-8) are samples obtained by continuous registration while the animal passed from wakefulness to drowsiness. Note the increase of the response evoked in the centre median, followed by the occurrence of slow waves in the cortical and thalamic regions studied. At the end of 8, slight arousal cwses re-appearance of waking activities (9, 10).
368
M. A . B. BRAZIER
light registered in the optic tract. This complexity of elementary response of the optic pathways is apparently explained by the fact that the retina itself is already a miniature brain (as expressed by Granit).
A. FESSARD :
I think I heard correctly that you offer two alternative solutions as to how information may be conveyed and utilised within the CNS: frequency modulation in single unit discharges or changes in shape of average profiles, the latter view having the advantage of being closer to the formal concept of information, which, as is well known, is related to the degree of “unexpectedness” of events which emerge from the average level. 1 am asking if you are not in fact disregarding a third solution, the one that would give efficiency to “patterns” of activated - or temporarily inactivated - units in the population of neurons which are involved, either at the afferent, or associative, or finally motor efferent level. This corresponds to my own schematic representation of how the brain works when utilising information for the execution of an adapted act (or readjustment of a posture). I conceive of this operation as a succession of “patternings”, from the receptive periphery (including the participation of interoceptors) to the final pathways of motor action. It seems t o me that some evidence for this being so can be found in the experiments presented some time ago by Jasper and his collaborators when they made it obvious that a sensory-motor conditioning in monkeys was accompanied by a reorganisation of the discharge patterns in the motor cortex. On the other hand, I wonder if you can really explain how and where such an operation as an “averaging” can actually be carried out within the CNS. It seems to me that a risk of confusion does exist concerning the kind and amount of information either brains or computers have to deal with. “Information” is an ambiguous term which must be related to the properties of the system that deciphers and utilises the message. What the computer is presented with - a set of simultaneous EEG tape records, for instance - is only an incomplete and distorted by-product of the intimate nervous activities from which the brain elaborates its patterns of action.
P. BUSER: With regard to the distribution of visual responses in the median thalamus our experiments (jointly with J. Bruner) have revealed visual as well as acoustic responses not only in the centre midian but also in directly lateral or posterolateral regions - the suprageniculate nucleus and the nucleus limitans. BUSER,P. et BRUNER,J. RCponses visuelles et acoustiques a u niveau du complexe ventromedian posterieur du thalamus chez le chat. C.R. Acad. Sci.(Paris), 1960, 251: 1238-1240.
C . AJMONEMARSAN:
Dr. Brazier, with regard to the interpretation of some of the changes observed in the various components of the response under different conditions, and, in particular, of the changes during barbiturate anaesthesia - you have stated that the prominent increase of the first component is probably the result of “removal of inhibitory influences”. While your suggestion is quite acceptable I wonder if you have any evidence to support it and if one should not also consider the possibility that such an increase might only be an apparent one: in fact, with the averaging technique, a simple “stabilisation” of the latency would probably produce similar results (i.e. increase in average amplitude).
W. GREYWALTER: One thing Dr. Brazier did not mention about the average technique is that the “noise penetration” or clarification factor is a function of the square root of the number of samples, so that the likelihood of picking up too much, as Dr. Adey suggests is not as great as it may seem. With a “white noise”
DISCUSSI 0N
369
background the improvement in signal-noise ratio would be about 3.5 per 12 samples, about 8 for 60 and about 16 for 240, and these smaller figures are the ones we should consider in assessing the resolution of the system. Added to this, is the inverse square law governing the physical electric field spread, which also limits the volume of likely derivation.
H. GASTAUT: I was very much interested to hear that, in the course of the habituation of visual evoked potentials in the centre median, only the third component progressivly attenuates and even disappears, while the second component persists and even seems strengthened in the course of repetitions. This is the more interesting to me because, in the course of similar experiments made by Dr. Beck in my laboratory, all the late components of evoked potentials diminished in amplitude, regardless of the site of registration, and particularly in the centre median. Yet the second component described by Dr. Brazier in the centre median is rather late (about 75 msec). J should very much like to know, therefore, whether Dr. Brazier has some explanation to offer for its extraordinary persistence in the course of habituation. BECK,E. and GASTAUT, H. The effects of habituation and reinforcement on evoked potentials to light in the cat. Medical Electronics and biological Engineering, in press.
H. H. JASPER: Dr. Brazier has posed two important questions for discussion : (a) the relationship between evoked potentials recorded from large assemblies of nerve cells and their processes and information processing, and (b) the validity of summated records which exclude all except the constant components of a long series of evoked potentials as an indication of “integrated” information processing in the brain. Studies of relationships between the firing patterns of individual cells within a population of neurones from which evoked potentials are recorded simultaneously have revealed a rich variety of individual neurone behaviour in response to sensory stimulation which is not always readily related to changes in evoked potentials. Considering the highly complicated discriminative and integrative functions carried out by the normal waking brain, it would seem to me much more likely that the coded pattern of information processing would be found in relation to that of individual cells and their interrelations in networks or circuits rather than to the gross summated evoked potential, and perhaps even less in the stable components of a long series of evoked potentials. It seems now well established that evoked potentials are more of the nature of summated synaptic potentials and not summated spike potential discharges of individual neurones. Their significance for information processing would be, therefore, different from that which might be obtained even if it were possible to record simultaneously from the thousands of individual neurones which must take part in the production of a single evoked potential. Relations between unit discharge and evoked potentials seem to be most readily demonstrated when “information processing” is impaired, such as under the effect of chloralose anaesthesia or epileptic discharge. In the alert unanaesthetized animal, unit discharge patterns in sensory receiving areas of the cortex in response to brief stimuli show a remarkable degree of individuality, often not clearly related to changes in surface evoked potentials. It would seem to me highly improbable that one would find isomorphism between the manner in which synaptic potentials are synchronized and summated in a single evoked potential and then electrotonically summated in series and the way the brain actually carries out such processes, but it is an interesting hypothesis which Dr. Brazier has shown much ingenuity in putting to a test. F. BREMER TO H. H. JASPER: I cannot agree with our colleague Jasper’s scepticism as to the functional significance of cortical evoked potentials. First of all for a reason of principle. The evoked potential represents the integration of synchronized responses of innumerable cortical cells, while the unitary potential represents that of the cortical
370
M. A . B. BRAZIER
cell near the exploratory micro-electrode. It seems obvious that the instructive value of the integrated response - in the case of variance - prevails over that of the unitary response, which may not be representative. On the other hand, there is a definite instance of congruence between evoked and unitary potentials, i.e. that of reticular facilitation of the response of the visual area in cats in response to an afferent volley. This congruence - which is revealed by comparison between the micro-physiological observations of Griisser and Creutzfeldt and the macrophysiological observations of Dumont and Dell and ourselves (with Stoupel) -is confirmed by the simultaneous registration of two types of responses as carried out recently by Creutzfeldt in the laboratory of our colleague Jung. This was done in a study of the effects of stimulation of the reticular activating system on evoked potentials of the visual area in “encephale isol6” cats. CREUTZFELDT, O., SPEHLMANN, R. und LEHMANN, D. Veranderungen der Neuronaktivitat des visuellen Cortex durch Reizung der Substantia reticularis mesencephali. In: R. JUNGund H. KORNHUBER (Herausgeber): Neurophysiologie und Psychophysik des visuellen Systems. Springer, Berlin-Gottingen-Heidelberg, 1961 : 351-363. BREMER, F. et STOUPEL N. Facilitation et inhibition des potentiels evoques corticaux dans l’eveil cerebral. Arch. int. Physiol., 1959, 67: 240-275. DUMONT, S. et DELL,P. Facilitations specifiques et non-specifiques des reponses visuelles corticales. J. Physiol. (Paris), 1958, 50: 261-264. GRUSSER, 0.-J. and GRUSSER-CORNEHLS, U. Neuronal discharge and evoked potential in the primary visual cortex of cats. Fifth Internat. Congress of Electroencephalography and Clinical Neurophysiology, Rome, Sepf. 1961. Excerpta Medica, Intern. Congr. Ser. No. 37, 1961.
R. JUNGTO F. BREMER: I quite agree with Professor Bremer about the good correlation between surface potentials and neuronal discharge in the special conditions he mentions. This is true for the facilitation of the evoked potentials following reticular stimulation or acoustic arousal as Creutzfeldt and his colleagues found at the neuronal level in confirmation of Prof. Bremer’s and Dell’s findings and also for Grusser’s analysis of evoked potentials and spikes. Although we encountered good correlations between visual neuronal patterns in the cat and vision in man I would be more cautious about a general correlation between brain waves and neurons. In the more complex conditions which Prof. Jasper mentioned and in less artificial arrangements there may indeed be wide discrepancies between microelectrode recordings and brain waves, O., SPEHLMANN, R. und LEHMANN, D. Veranderungen der Neuronaktivitat des visuellen CREUTZFELDT, Cortex durch Reizung der Substantia reticularis mesencephali. In: R. JUNGund H. KORNHUBER (Herausgeber). Neurophysiologie und Psychophysik des visuellen Systems. Springer, Berlin-Gottingen-Heidelberg, 1961 : 351-363. BREMER,F. et STOUPEL, N. Facilitation et inhibition des potentiels evoques corticaux dans l’eveil cerebral. Arch. int. Physiol., 1959, 67: 240-275. DUMONT, S. et DELL,P. Facilitations specifiques et non-spkcifiques des reponses visuelles corticales. J. Physiol. (Paris), 1958,50: 261-264. GROSSER, 0.-J. and GROSSER-CORNEHLS, U. Neuronal discharge and evoked potential in the primary visual cortex of cats. Fifth Internat. Congress of Electroencephalography and Clinical Neurophysiology, Rome, Sept., 1961. Excerpta Medica, Intern. Congr. Ser. No. 37, 1961. W. R. ADEY:
My remarks are concerned mainly with what Dr. Jasper has said concerning the evoked potential, and the problem of considering it as a profile of neural activity, and also with the question raised by Professor Fessard, concerning the relation of cortical integration to patterns of neural activity. In discussing the means by which information is handled in cerebral systems, it is necessary that any model, either physiological or mathematical, or both, should obviously be both appropriate and adequate. Most investigators would agree that the evoked potential is not the envelope of neuronal firing, and that in its major aspects, it is constituted of graded phenomena, such as post-synaptic
DISCUSSION
371
potentials, and dipoles appearing electrotonically in dendritic mechanisms. However, it would seem that these processes are indeed at the core of the “transactional” mechanisms of the cortex, and may be intimately related to the integration and to the storage of information, either temporarily or permanently, as has been pointed out elsewhere. These phenomena would essentially precede the initiation of a pattern of neuronal firing in response to a particular stimulus, or the modification in a predictable way of an existing firing pattern. It is here that we face a problem of fundamental importance. We must decide whether our cerebral model will have a deterministic or probabilistic basis. From many studies of cerebral unit activity, and from morphological considerations, it would seem that our model should be probabilistic, with the deterministic inputs in specific sensory systems superimposed on the inherently probabilistic aspects of cortical activity. It would seen1 that even the organization of afferent influxes in specific pathways may have probabilistic aspects at thalamocortical levels. On this basis, the pattern of neuronal firing which represents the cortical response to a particular stimulus might have consistent aspects in spatio-temporal organization, when viewed across the whole population of responding neurons, but would not necessarily involve the same individuals in the neuronal population in responses to a sequence of identical stimuli. It is necessary that we pay careful attention to this probabilistic aspect of responsiveness,for there is great redundancy in cerebral systems, and this redundancy must presumably play its part in the establishment of any pattern of neuronal firing in response to a stimulus. The informational content of a particular spatio-temporal firing pattern might well be essentially identical in response to successive stimuli, but the individuals within the neuronal population so involved need by no means be the same. It would seem equally important that, in our quest for solutions to the coding of neuronal firing in cerebral systems, our attention should not be diverted from the problems of possible informational transactions occurring in the preceding and concurrent graded events which constitute a major aspect of the evoked potential. A. ARDUINI: The effects of barbiturate on the size and form of evoked potentials (flashes) could be explained, on the basis of our experience, through modifications of the output discharge of the retina and not necessarily of central phenomena. On the other hand the barbiturate might interfere with the mechanism for the reduction of the redundancy of the information, preventing the reduction of the message to its essentials and thereby permitting averaging of larger waveforms. Other objections related to the mechanism of information processing, for instance, pattern analysis versus space and time averaging, are left for a general discussion of the subject. H. W. MAGOUN: Earlier, Gastaut sought an explication for the divergent changes in amplitude of the two components of the response of the centre midian which Brazier records upon repeated flashes of light, in which the earlier wave becomes increased as the later one is reduced. Possibly AIbe-Fessard‘s observation of an increase in amplitude of the earlier wave during inattention may be related to Brazier’s proposal that reduction of the later wave signals lack of stimulus novelty. Monotonous stimuli customarily provoke inattention, in relation to which the earlier wave may become augmented.
M. A, B. BRAZIERS replies To D. Albe-Fessard Mme. Albe-Fessard asks whether we have looked for later responses in the CM (those occurring more than 500 msec after the flash). We have not seen them in the averaged response, but it is my impression that their latencies are very variable and not closely time-locked to the flash. Hence they average out in the cross-correlation with the stimulus. Perhaps if we program the computer to average a greater number of responses, we may find them.
372
M. A. B. BRAZIER
I would like to add an explanation of our reason for using about 240 responses as a general rule. The latency, size and waveform of the responses one records depend very greatly on the state of the animal. This is why we monitor every experiment with a continuous ink record. In our experience, about 4 rnin is the optimal period for maintaining an unanesthetized cat at the same level of alertness while it is being flashed repetitively. Flashing at 1 per sec for 4 min gives us 240 responses. An additional reason is a purely technical one related to the computer we were using for these analyses. Perhaps if we succeed in maintaining this level of alertness for a longer period, we may find Mme. Albe-Fessard’s long-latency responses in the larger sample. To W . R . Adey Theoretically this may be a hazard, but in practice we find it negligible, for we have tested specifically for it. In any case, as I mentioned in my paper, we always run bipolar recordings as a check. The unipolar records are used only for determination of latencies. As Grey Walter has pointed out, even theoretically this cannot be expected to be a very potent factor for two reasons: ( I ) the signal to noise discrimination increases only as the square root of the number of samples, and (2) action at a distance follows the inverse square law. Both these factors act to minimize the effect that you propose.
To F . Bremer Professor Bremer is, of course, correct that multiple responses are recorded in the optic nerve, but as Bishop has so clearly shown, the fibers of the optic nerve go to multiple destinations. The timing and interval analyses of these multiple discharges in the compound nerve do not correlate, in our experience, with the oscillations that follow the primary response in the visual cortex. BISHOP,G. H. and CLARE.M . H. Organization and distribution of fibers in the optic tract of the cat, J . comp. Neurol., 1955, 103: 269-304. To A . Fessard I did not wish to suggest that there were only two possible models that we could use as hypotheses about brain function. Both (and others) may be fruitful for the design of new experiments. The model I have been exploring here is a probabilistic one: namely that action on the part of the brain results when the probability of events occurring within it exceeds that to be expected by chance. For such an estimate of probability, a comparison would have to be made by the brain between average activity before the sensory message was received with the average of that evoked by the sensory inflow. To P . Buser I was most interested to hear Dr. Buser’s experiences, all of which strengthen the concept that the centre median, as usually delineated, cannot be considered as homogeneous in function. In our experiments we have not yet explored the more lateral nuclei. To C. Ajmone Marsnn A decrease in variance will, as you say, give an increase in the amplitude of the median and a wider distribution around the median. We have explored and published computer analyses of this variance that confirm its decrease under barbiturate anesthesia. However, this alone cannot explain the results presented here, for (as I mentioned in my paper) we monitored all our recordings with cathode ray traces of single sweeps. The augmentation caused by barbiturates in light doses is so great that responses, previously undetectable in the background activity, become conspicuous even in the single trace. BRAZIER, M. A. B. Some uses of computers in experimental neurology, Exp. Neurol., 1960,2: 123-143.
To H . Gastaut Dr. Gastaut’s remark has been answered by Dr. Magoun, and I have nothing more to add. To H . H. Jasper I believe that what is disturbing Dr. Jasper is that, in this work on brain signals, I am exploring a probabilistic hypothesis rather than a deterministic one. T o analyze the myriad complexities of
DISCUSSION
313
the brain’s function by non-statistical description is too gigantic a task to be conceived, but exploration in terms of probability theory is both practical and rational. In characterizing nervous activity, one would not therefore attempt the precise definition that arithmetic demands but would seek the statistical characteristics of the phenomena that appear to be relevant. The margin of safety that the brain has for “appropriate reaction” is thus much greater than a deterministic, arithmetically precise operation would impose. Chaos would result from the least slip-up of the latter, whereas only a major divergence from the mean would disturb a system working on a probability basis. The rigidity of arithmetic is not for the brain, and a search for a deterministic code based on arithmetical precision is surely doomed to disappointment. To F. Bremer Of course I agree with Professor Bremer that pentobarbital has a well-known effect on the retina. However, depression of the retinal neurones should affect all components of the CM response, whereas (as the slides show) there is a differential effect, only some components being depressed and others even being augmented.
To A . Arduini The depressant effect of barbiturate on the retina is well-known, but I do not see how the differential effects on the various components of the CM can be explained in this way. As the slides have shown, some components of the multiple response are very vulnerable to barbiturates, others very resistant.
A Transcranial Chronographic and Topographic Study of Cerebral Potentials evoked by Photic Stimulation in Man* H. GASTAUT with the collaboration of E. BEEK, J. FAIDHERBE, G. FRANCK, J. FRESSY, A. REMOND, C. SMITH AND P. WERRE Unite de Recherches Neurobiologiques, Institut National d’Hygit?ne, Maneille (France)
A grant from the U.S. Department of Health, Education and Welfare, for international research on cerebral mechanisms of behaviour, has made possible a new series of investigations in the Neurobiological Research Department of the National Institute of Hygiene at Marseilles, which have been undertaken with a view to the use of cerebral potentials in man, evoked by visual stimulation. The aim of these studies is partly the medical one that they may help in the diagnosis of lesions which affect the visual pathways, and partly the psycho-sociological one: they may make possible the classification of individuals on the basis of the manner of reception, transmission and integration of visual messages in the brain. In order to achieve rapid practical results it was decided to use certain special techniques which permit the automatic analysis of evoked potentials from the total electrical activity recorded across the scalp. The project was made possible by an agreement with the European Office of U.S.A.F. Aerospace Research, which placed at our disposal an apparatus designed and constructed by Dr. RCmond, which has been given the name of “phasotron” (RCmond and Ripoche 1958; RCmond 1961). Two groups of workers have studied the characteristics of visually-evoked potentials in man, and in particular the variations of form, amplitude, latency, and location, as functions of the four following groups of variables: (a)Thephysicalcharacteristics of the stimulus (its intensity, frequency of stimulation, etc.); (6) the affective state of the subject (attention, emotion, etc.); (c) the repetition of visual stimulation or its combination with other types of stimulus, in order to demonstrate the phenomena of habituation and conditioning; ( d ) the alteration or dysfunction of the visual pathways (hemianopsia, photogenic epilepsy, etc.). A brief description of the methods employed precedes the tabulation of the results so far obtained.
* This work has been carried out at Marseilles, in the Neurobiological Research Department of the National Institute of Hygiene, with the help of Grant No. M-3258 of the U.S.P.H.S. and Contract No. AF61(052)-20 of the U.S.A.F.
CEREBRAL POTENTIALS EVOKED BY PHOTIC STIMULI
375
METHOD
Visually-evoked potentials, provided by an Alvar Vartclat, have been studied in 50 subjects in 100 experiments. The tube and reflector of this stroboscope were directed either at the eyes of the subject, at a distance of several centimetres, or at the inner surface of a white hemisphere 1 m in diameter, in the centre of which the subject is placed. The first method has the advantage of yielding the maximum brilliance of the flashes, the second of giving uniform illumination of the retina, irrespective of the position of the eyes. The potentials were recorded from scalp electrodes. These were usually bipolar, being at a distance of 3.5 or 7.0 cm from each other. Some monopolar recordings were also done, but we are not yet entirely satisfied with this method, because we have reservations about the stability of the so-called indifferent electrode. Each stimulation is represented by a series of 20, 50, 100 or 200 flashes, rhythmically repeated. Different durations of the inter-stimulus interval were tried before those of the order of 1000,90 and 70 msec were adopted*. These respond to frequencies of about 1, 11, and 14 c/sec, which are particularly interesting for a number ofreasons. The phasotron, connected in parallel with the electroencephalograph, comprises a holding system to which have access 8 electronic shutters, which open for 1 msec at any moment chosen after each flash. These 8 shutters can be opened simultaneously on 8 different EEG channels, so that the phasotron delivers, after the last flash, a histogram of 8 columns, each of which represents the mean amplitude of the signal recorded on each channel during the millisecond following the predetermined delay after each flash. In this manner a “topogram” is obtained which gives the mean amplitude of a brief fraction (at any given interval after the flash) of the evoked potentials in the 8 corresponding derivations. By performing a sufficient number of topograms with different delays, one can reconstruct chronograms, the contours of which represent the mean evoked potentials simultaneously recorded in each of the 8 channels. If only a single chronogram from a single channel is required, it can be obtained by opening the 8 shutters in succession; when this is done the delay between the stimulus and the opening of the first shutter, and the interval between the 8 elevations recorded from a single pair of electrodes are chosen. Thus the phasotron will give a chronographic histogram in which each column represents the mean amplitude of the information retained at 8 separate times rhythmically spaced after the initial delay. The histograms are easily transformed into chronographic or topographic curves by graphic interpolation of the centres of the different columns. It is possible to effect a linear interpolation by opposing neighbouring values in straight lines. However, a secondary interpolation is preferable, To obtain this, three successive values are joined by a parabola. By choosing appropriate delays and intervals as functions of the duration of the
* For preference, stimulus durations close t o prime numbers were chosen (e.g. 989, 89 and 73 msec); this makes it possible t o avoid interference with the responses by the a. e. frequency (or its harmonics), which complicates the interpretation of the results. References I. 392
376
H. GASTAUT
et al.
potential to be studied and the periodicity of the stimulus, knowledge is gained of the distribution in time and on the scalp of the potentials evoked by S.L.I. The responses obtained are a measure of the mean responses (Mean Evoked Potential = PEM), corresponding to the sum of the isolated responses repeated at short intervals of time under identical conditions and in sufficient number to reduce the variation. A complementary apparatus provides 8 isochronic curves which represent the simultaneous state of charge of the integrated circuit corresponding to each shutter, so that information is obtained about the manner in which summation proceeds until the final mean potential is obtained. Basic information about the temporal elements of cerebral responses appearing simultaneously from a number of regions demands many and careful manipulations if it is to be presented as a consecutive record. The cerebral responses and synchronizing signals are registered on a magnetic tape with 8 tracks, which can be topographically re-read as often as is required, with a different delay each time, in order to obtain the values of a group of histograms which is as representative as one could wish. The interpolative calculations of the second or third order are performed by a digital calculator which supplies directly, according to the programme with which it is fed, and in the form of a diagram : (1) The group of evoked potentials as a function of time (one per electrode); (2) the group of instantaneous topograms (one per delay); (3) the points of projection on the scalp of the rough values of the potential of each histogram, presented as maps of the level of spatio-temporal organisation of the responses.
Characteristics of the mean evoked potential produced in man by low frequency photic stimulation* Although both teams working on man have used low frequency stimulation at about Ic/sec, we are presenting here the results obtained by the three authors cited, who have made a special study of their effects on 20 normal adult subjects. TECHNIQUE
This was identical with that already described; the details are as follows: Flashes were applied indirectly, by means of the hemisphere; bipolar recording with short (3.5 cm) inter-electrode spacing, arranged around the inion (vide schema in figures); utilisation of bursts of 50 or 100 flashes, separated by intervals of 983 msec in most a~stances. RESULTS
The mean evoked potentials (PEM) recorded by this method closely resemble those obtained by the use of different methods of summation and integration by Cobb and Dawson (1960) and Ciginek (1961); they show, among other things, the same
* By P. F. Werre, C. J. Smith and E. C . Beek.
CEREBRAL POTENTIALS EVOKED BY PHOTIC STIMULI
377
variations as those described by these authors from one subject to another, from one time to another, and from one place to another. It has always appeared to us that this variability is so much more important than has been supposed by the authors quoted above that it is difficult, if not impossible, to put forward the plan of a hypothetical evoked potential of this type to serve as a reference for all others. Thus it is to the variations themselves that we have devoted our attention.
( I ) Variations in the elaboration of PEM It is well known that an evoked potential recorded under standard conditions varies from one moment to another, notably as a result of spontaneous cerebral activity; this is one of the difficulties which integrative methods seek to reduce. The fact that we have decided to utilise the average of at least 50 responses is precisely because such a number is statistically necessary to obviate the effects of sporadic fluctuation. In this way it is possible to obtain stable responses, identical from one moment to another in any given experimental situation. In all instances, examination of isochronic curves (see Methods) shows that the manner in which this mean curve is elaborated varies according to the individual and the time, and can be roughly averaged by : (a) Continuous summation, which levels out small variations in the amplitude of successively integrated evoked potentials; (b) Discontinuous summation, which smooths out large variations which may in their turn be regular or irregular. (2) The variation of PEA4 from one subject to another
It is sufficient to compare the PEM’s recorded at 8 homologous points in the two subjects GR and GU (Fig. 1) to see that they possess common characteristics which permit an overall description analogous to that already given by Cobb and Dawson, and by Ciganek. Thus the PEM of subject GU, recorded on the first line between two midline electrodes 3.5 and 7.0 cm respectively above the inion, shows a series of very low amplitude oscillations between the 20th and the 130th msec after stimulation. This corresponds to the complex response described in the same locus by Cobb and Dawson, because the following characteristics can be observed successively under the upper electrode : An initial surface-positive response of 0.5 pV amplitude, 20 msec latency, and with a peak at 25 msec (latency 20-25 msec and peak at 25-30 msec, according to Cobb and Dawson). Three subsequent waves: (a) one negative, of 1 p V , with peak at 51 msec (35-45 msec, Cobb and Dawson); (b) one positive, of 0.5 p V , with peak at 63 msec (60-70 msec, Cobb and Dawson); (c) one negative, of 3 p V , with peak at 111 msec (90- 100 msec, Cobb and Dawson). References P. 392
H. GASTAUT et al.
378
5 G R
- - - 5 GU
Fig. I Variabiliiy of averaged PEM's from subject lo siibject and area to area. The PEM's of subjects GR and G U have clearly many points in common. Even the after-discharges are seen to be frequently in phase. Comparison of the different areas shows that the after-discharge is least marked and least regular in leads 20-21, 20-5 and, to a lesser extent, in leads 20-15. This is true for both subjects. A similar trend is seen in the case of subject OL (Fig. 3). Perhaps, the after-discharge is more a property of the associative areas. Clarification: Stimulation parameters are noted at the top of each figure: p(eriod) 741,983, etc., indicates that flashes followed each other by 741,983 etc., msec; during the first 150 msec 100 responses were averaged, therafter 50. Electrode arrangement: 20 just above the inion, 21 and 22 posterior midline, separated by 35 mm; 5 and 15,35 mm laterally right and left of 21 ;4 and 8 laterally right, 14 and 18 laterally left of 22, all separated by 35 mm. Polarity convention: an upward deflection indicates that the first electrode listed in a pair has become relatively negative. Data points can be deduced, since they occur at the angles of the straight lines connecting them. Lead 1-2, in some figures, was included for purposes of control, but is not discussed here.
After this group, a rhythmic after-discharge, with a periodicity of about 100 msec, of the type described by Cighnek, is observed. In the same way, the PEM of subject GR, recorded on the fourth line between two electrodes one of which is situated on the inion and the other 3.5 cm laterally, shows a response of the same type, but this is even closer to the values described by CigAnek, because the summits of the 7 successive deflections, of alternate positive and negative polarity (with respect to the median electrode, No. 20), are situated at 27, 39, 51, 75, 99, I I 1 and 170 msec, compared with 39, 53, 73, 94, I14 and 135 msec as measured by him when he used a statistical procedure which is open to question. The fact that two such different responses are found in five subjects is evidence of a remarkable inter-individual variation. This variability is perhaps more marked during
3 79
CEREBRAL POTENTIALS EVOKED BY PHOTIC STIMULI
the first part of the response than during the subsequent rhythmic after-discharge, in which the different elements more frequently coincide in different subjects. It might
-.-----APERIODIC
FLASH
Fig. 2 The independence of averaged PEM'Ato the periodicity of the flashes. In both the subjects illustrated, BE showing no after-discharge and GA a small and regular after-discharge, responses to the periodic and aperiodic flash, delivered at a rate of about one per sec, are seen to be similar. The periodicity in question apparently does not influence the responses. Stimulation parameters, etc., as for Fig. 1 .
:se]
(: Fig. 3 The variability of averaged PEM's from subject to subject and area to area. The differences between subject BA and TA (comparing homologous areas) are rather large. The form of the curves, one excepted, shows great dissimilarity. Amplitudes are very different; note that at 115 msec in lead one, one response IS four times the other. Comparison of different areas within each subject, shows that in BA all the curves are broadly similar, with the exception of the second one. In TA this similarity is certainly less. Stimulation parameters, etc., as for Fig. I , References P 392
H. GASTAUT et
380
al.
be supposed that the similarity of the after-discharges is related to the frequency of the flashes which provoke it, but this is, in fact, not so, because we have been able to show that aperiodic as well as periodic flash stimuli given at frequencies of about 1 cjsec produce that same rhythmic after-discharge (Fig. 2). These differences become even more evident when (Fig. 3) the PEM’s of the live subjects BA and TA are compared. Thus, for example, in subject BA, the PEM’s recorded on the first two lines, between the pairs of electrodes 21-22 and 20-21, show a reversal of phases at about 120 msec, while this phenomenon is entirely lacking in subject TA, who shows, at the same point in time, waves perfectly in phase. On the other hand, 80 msec later, it is TA who shows, unlike BA, phase reversal between the PEM’s recorded on lines 3 and 4. These phase reversals have often been observed in one or several components of the PEM’s and it is not possible to draw positive conclusions from them. Nevertheless, it is possible to state that they are most often seen between the two median electrodes, from 100 to 150 msec after the stimulus, and that the point at which the inversion occurs is situated in the neighbourhood of
-
W..LI
______I-
........W.W
sa
Fig. 4 The variability ofaveraged PEM’s from time to time and area to area. Several responses of subject OL obtained on the first experimental day are seen to differ markedly from those obtained 2 and 3 weeks later (e.g., leads 20-21 and 20-15); others are very similar (e.g.,leads 5-8 and 21-4). The differences are especially pronounced in the leads which have electrode 20 in common. It is possible that slight changes in the position of the electrode can cause the deviations observed, but, because electrode 20, just above the inion, is the one with the best anatomical reference point, this seems unlikely. With respect to the right-left differences, leads 5-8 and 15-18 show similar responses. The more median leads 20-5 and 20-1 5 show, on the contrary, some dissimilarity. The responses to flashes with a period of 83 msec do not differ much. Stimulation parameters, etc., as for Fig. 1.
CEREBRAL POTENTIALS EVOKED BY PHOTIC STIMULI
38 1
the common electrode (No. 21), approximately 40 mm above the inion; this is in accordance with the opinions of Cobb and Dawson. Despite the inadequacy of our apparatus in detecting rapid rhythms, one can also see in Fig. 3 that at the beginning of the PEM shown on the first line, there is a rhythm at about 100 c/sec in subject BA, identical to that described by Cobb and Dawson (1960). This rapid rhythm was entirely lacking in subject TA. (3) Variations with the passage of time
The PEM's recorded in a single subject from one day or one week to another show many points of similarity, but sometimes the differences are even greater than those between different individuals. Thus, in Fig. 4, second line, subject OL shows, between 75 and 125 msec after stimulation, opposite signs between the first recorded PEM and those recorded two and three weeks later. In the fifth line opposite signs can be seen very clearly in parts of the after-discharge recorded at an interval of several weeks. On the other hand, the PEM's recorded on the other lines remain remarkably similar from week to week. In Fig. 5 also (subject ZA) it can be seen that the PEM's are particularly inconstant from week to week.
-, O - L - n a
200 1 , 303 1 1 4Dl 1 1 503 1 1 603 1 1
._-. -2'
6 W?
I ~ P O (
w 0
msec
5 ZA
Fig. 5 The variability of averaged PEM'sfrom time to time and area to area. In contrast to subject OL(Fig. 3), subject ZA possesses an irregular after-discharge.Here especially the discrepancies between the three trials, separated by 2 and I weeks, is pronounced. The first parts of the responses do not show great variability. The responses to flashes with a period of 83 msec do not deviate extensively. Stimulation parameters, etc., as for Fig. 1 . References P. 392
382
H. GASTAUT
et al.
( 4 ) Extra-individual spatial variations Despite an overall similarity, there are notable differences between the PEM’s recorded from different occipital and parieto-occipital sites. It was found, particularly in the initial response, that the PEM’s recorded from the midline were more complex in form than those recorded more laterally. On the other hand the rhythmic afterdischarges tended to be less clear near the inion than further laterally. So far as the symmetry of the response is concerned, it can be seen from Fig. 4 that it is greater in laterally-recorded PEM’s (electrodes 5-8 and 15- 18) than in those recorded near the mid-sagittal plane (20-5 and 20- 15). Such a finding may depend on the fact that the electric field is generated in a truly asymmetrical fashion on the occipital poles of the hemispheres, but it seems more probable to turn upon the fact that the common electrode (20) is not always exactly on the inion and that this itself does not necessarily mark the inter-hemispheric fissure. ( 5 ) The variations due to affective state
These variations have not been systematically investigated, but were imposed by chance in the course of some recordings. Thus it was noted (Fig. 6):
Fig. 6 Variation in the PEM due to changes in the psychological state of the subject (explanation in text).
(a) When subject BA was in a state of alertness, a response was recorded at 35 rnsec which disappeared when the subject was relaxed. The later responses were at 54, 84 and 133 msec in the alert state, but when the subject was relaxed, they were amplified and retarded (63, 91 and 140 msec). (6) In subject PE during sleep there was an elimination of amplitude of the late responses, particularly i n that at 175 msec.
CEREBRAL POTENTIALS EVOKED BY PHOTIC STIMULI
383
(c) In subject NA, a response at about 70 msec changed the polarity during passage from the alert to the drowsy or sleeping state. ( d ) In subject AE also there was an inversion of the polarity at about 28 msec during his transition from the attentive to the inattentive state. It is noteworthy that in all subjects these variations are marked in the later part of the response, as had been remarked by Cobb and Dawson. The failure of the initial response to change negates the idea of a modification of the specific afferent cortical volley suggested by Jouvet and Courjon, but suggests rather a modification of intrinsic cortical excitability, particularly in cortical regions remote from the specific area, in which former the later part of the response predominates. CONCLUSIONS
When photic stimuli at low frequency (1 c/sec) are used, in order to obviate interference with the responses, the phasotron extracts from the activity of the occipital cortex a PEM which corresponds closely with that previously described by Cobb and Dawson. Comparing the results obtained by the team responsible for the present findings (Werre, Smith and Beek) and those of other groups using the phasotron, it appears that the first response usually recorded is initially positive, with a latency of 20-25 msec and a peak at 25-30 msec. This is in accord with the findings of Cobb and Dawson. The first response described by Cigtinek is a negative one a t 39 msec and appears to correspond with the second wave of Cobb and Dawson. The most consistent components of the PEM are the later elements, the appearance of which in the more lateral parieto-occipital regions and variations with state of attention permit them to be considered as non-specific. On the other hand the initial part of the PEM may be considered specific. It is often unrecordable or is only very slightly evident because of its low amplitude (0.5-2 p V ) ; it predominates in the midline, above the inion and shows very little intra-individual variation with changes in the psychological state. The more or less complete PEM’s recorded under these conditions with the phasotron vary considerably both inter- and intraindividually, which makes comparative studies difficult. For this reason we have decided not to use this method for diagnostic or psychological studies. We prefer to utilise more constant and stable responses such as those we have found under other conditions of stimulation. Characteristics of the mean evoked potential obtained by high frequency intermittent photic stimulation in man*
Because there are considerable differences according to whether flashes are administered at a rate greater or less than 12-13 per sec, we are presenting our results separately, according to whether they were obtained by a stimulus of 11 flashes per sec or 14 flashes per sec.
* By J.
Faidherbe, G. Franck, J. Fressy and H. Gastaut.
Referenccs p . 3Y2
384
H. GASTAUT
L.HEM. EYES CLOSED
R.HEM.
et al.
L.HEM.
R. HEM.
L.HEM.
R.HEM.
ANT.
CHRONOGRAPHY
TOPOGRAPHY
I
22rn;ec' I
I
TOPOGRAPHY
I
75.77 mse
Fig. 7 Chronograms and topograms (bipolar recordings) of potentials evoked by trains of 200 photic stimuli repeated at 11 flashes per second, in the occipital regions and along two symmetrical rows of electrodes running longitudinally across the scalp. Positivity of the most posterior electrode causes a downward deflection. Note the similarity of these curves and their symmetry between the two hemispheres in a single subject.
In both instances, we shall consider first the results obtained by chronographic studies of the P.E.M.'s recorded in the occipital region, and then those obtained by topographic studies of a line of electrodes arranged from back to front in an occipitofrontal plane in one or both hemispheres (for the electrode positions, see Figs. 7, 8 and 9). A . Results obtained with aflash frequency of 11 per second 1. Chronographic study. Nine subjects were studied, 3 of them twice. The PEM's are shown in Figs. 7 and 8. Their form is very different from that seen after lowfrequency stimulation. The wave is diphasic, the first part being positive. Its peak is
385
CEREBRAL POTENTIALS EVOKED BY PHOTIC STIMULI
DE5 EYES CLOSED I I s/s
fli' q l *\
f LY J.
'
DON.
SEP I
FRE....-
,
,%p--
m m
C H RO NOGR A PHY
f
I
f
ti
TOPOGR.
B
60msec E(
C
75.8msec E(
Fig. 8 Chronograms (bipolar recording) and topograms (monopolar recording with reference at the mastoid process) of potentials evoked by trains of 200 photic stimuli delivered a t the rate of 11 flashes per second, in the occipital regions and along a row of parasagittal electrodes. Positivity of the upper longitudinal row of electrodes (active electrodes) causes a downward deflection. A, time interval between stimulus and sample = 25 or 30 msec; B, time interval between stimulus and sample = 60 msec; C, time interval between stimulus and sample = 75 msec.
found on average at 22 msec (15-30) and its mean amplitude is $12.5 pV (+3.5 to +15). The wave inverts at 46 msec (40-55) and the second part is negative, with a much more irregular form than that of the positive initial component. A fairly constant peak is found at 77 msec (77-80), which has a mean amplitude of - 1 1 pV (-2 to -18). It is sometimes preceded by an earlier negative peak, almost always of lower amplitude, which supervenes between 50 and 60 msec. We found that this chronographic response was symmetrical in 3 out of 5 subjects examined in this respect. In the two cases in which the curves were not quite symmetrical, careful study led to the conclusion that this was due to asymmetrical placing of the electrodes. References p . 392
386
H . GASTAUT
DE5-
EYES CLOSED 14 F.S
et al.
D O N
CHRONOGRAPHY I
EO
-%
A 14 rn sec 8-8-61
B 14 rn sec 12-8-61 -5pi
C 45 rn sec 8-8-61
D 45 rnsec 12-8-61
Fig. 9 Chronographic contours are shown in the upper part of the figure, topographic contours are shown in the lower part. In the subject DES two records were taken. During the topographic recording shown in the solid curve, the same subject, o n command, made an effort at mental concentration. A , topographic samples were taken 14 msec after the stimulus; different contours are shown corresponding t o the following psychological states: (1) control record with no instruction given t o the subject, who was in a state of relaxation; (2) after the subject had been asked t o concentrate on the stimulation; (3) after a period of rest when the subject was asked to relax; (4) record taken while the subject was making a maximal effort at concentration, extracting square roots and ( 5 ) during the subsequent period of relaxation. B, time interval between stimulus and sample 14 msec as in A; ( I ) control record with no instructions; (2) after a period of relaxation, subject having been asked to relax; (3) control as in 1. C, time interval 45 msec (this curve is the inverse of contour A). (I) Control record with no instructions; (2) while subject was making effort at mental concentration. D,time interval 45 msec (this curve is the inverse of contour B).
2. Topographical study. Eight subjects have been examined, one of them twice. Two types of recording have been carried out: (a) Bipolar recording with short inter-electrode spacing (Fig. 7). ( i ) Topography with a "stimulus-response" dehy of 22 msec. The maximal positive voltage, averaging 11 pV (+9.5 to 12.5) is usually recorded between the inion
+
+
387
CEREBRAL POTENTIALS EVOKED BY PHOTIC STIMULI
and the posterior parietal electrode. In one subject the topographical curve was entirely positive, but in all the others a negative phase was seen. The point of inversion is posterior, between electrodes 1 and 2 or 2 and 3 (Fig. 7), the average voltage of the peak of the negative wave was -5 pV (-4.5 to -5.5), and was found in the parietal region. In two subjects, the anterior region showed positive curves ( + I to 1.5 pV), similar to the posterior areas. (ii) Topography with a “stimulus-response” delay of 75 or 77 msec. The curves obtained with this order of delay represent the mirror-image of those described after a 22 msec delay. This time the posterior areas are negative, with an average of - 11.7pV (-7.5 to -20) at the maximum. In two subjects, the curve was entirely negative, but the other two presented a positive phase. In one of these it was found anteriorly, in the other in the middle, the anterior like the posterior regions being negative. The symmetry of the topographical curves at these two delay times (22 and 75-77 msec respectively) is remarkable. (b) Monopolar recording with long inter-electrode distance (Fig. S), with a common reference electrode on the mastoid process. (i) Topography with a “stimulus-response” interval of 25-30 msec. The maximal positive voltage, average +7 pV (+2.5 to +11) is found on the line between the inion and the reference electrode. In all cases a negative phase is found, distributed over the whole of the anterior regions; the point of inversion is posterior. The average maximum of the negative voltage is -3 pV (- 1 to -4). (ii) Topography with a “stimulus-response” interval of 75-80 msec. The maximal negative voltage is always found between the inion and the mastoid and it has a mean of -4pV (-1.5 to -8). In one subject (SEP) the whole wave was negative, but in the three others a positive phase was found in the middle regions, with an average maximum of +3 pV. The anterior areas remained positive in two of these cases, while in the third (DES) a negative voltage of - 1 pV was found anteriorly.
+
B. Results obtained with a stimulation of 14j7ashes per second 1. Chronographic study. Three subjects were examined, 1 of them twice (Fig. 9, upper part). The response is diphasic, of a more regular form than those obtained at I 1 per sec, and it appears to be almost sinusoidal. Contrary to what we have observed at 11 per sec, the first part of the wave is negative. i n all cases, the peak is at the 14th msec after the stimulus, with a mean amplitude of -3.8 pV (-7 to -2.2). The point of inversion is at the 30th (26-33) msec. The second part of the wave at this flash rate is positive and it is again more regular in form than that obtained at 11 flashes per sec. Its peak is seen at about 45 (41 -50) msec after the stimulus, and has a mean amplitude of f3.35 (2.2-5.8) pV. 2. Topographical study. Only cases recorded with widely spaced, bipolar electrodes which had a common reference at the mastoid process have been reproduced (lower part of Fig. 9). (i) Topography with a “stimulus-response” interval of 14 msec. The maximal negative voltage is always found between the inion and the reference electrode, and has a mean of -2.8 pV (--1.8 to -4.5). References D. 392
388
H. GASTAUT
et al.
The curve showed a positive phase in 2 subjects with a posterior point of inflexion. In 1 of the subjects (DES), the maximal positive voltage was found at the third electrode, in front of which the curve showed a second inflexion, the anterior regions being negative again. In the other subject (FRE), the maximal positive voltage was recorded at the sixth electrode and persisted at this level as far forward as the eighth. ,4 third subject (DON) showed negative voltages in the anterior and posterior regions, while the middle areas showed no difference of potential with reference to the mastoid. (ii) Topography with a “stimulus-response” delay of 47 msec. The maximal positive response was recorded at the inion in 3 subjects, and had a mean value of 3.6 pV (f2.6 to 4.5). In the two subjects (DES and FRE) who showed diphasic waves at 14 msec, the same is also the case at 41 msec. The point of inflexion was situated at the 3rd electrode or just behind it. In one of these cases, maximal negativity (-2.2 p V ) was recorded at the 3rd electrode, in the other at the 5th (- 1.7 pV). In both of them, negativity decreased progressively in an anterior direction. In the third case (DON), all potential difference disappeared anterior to the fifth electrode.
-+
DISCUSSION
When photic stimuli a t or above 10 per sec are used, the response from occipital cerebral activity extracted by the phasotron is entirely different from that obtained at lower frequencies (about 7 per sec). This difference is found in the wave form and in the latency of response; it is more stable from one subject to another, from one moment to another, and also between symmetrical areas of the two sides. 1. The stability of the response This is demonstrated by the fact that, following repeated trains of 200 photic stimuli delivered at 11 flashes per second every 40 sec, the form of successive chronographic (Fig. 10) or topographic curves is practically invariable in normal subjects, given that the physical parameters of the stimulus and the psychic state of the subject do not change. Only the levels of the different points of the contours vary, and these within narrow limits. On the other hand, if the psychological state of the subject changes, for example during concentration on mental arithmetic, it is possible to see, in particularly co-operative subjects, a significant change in the level of different points in the recorded waves (Fig. 9). Moreover, during periods of relaxation following such mental effort, the curves tend to resume their initial form, with, however, persistence of some changes. 2. The form of the response evoked by intermittent photic stimulation at I1 flashes per second This is a diphasic wave, initially positive. Its first part resembles the initial response obtained by Cobb and Dawson (1960) with flashes at long intervals; but its amplitude is greater (10-15 pV instead of 0.5-2 p V ) and the peak latency shorter (20 msec instead of 25 msec). The second phase is constituted by a negative wave lasting about 50 msec, with an amplitude of about 10 pV, which has a fairly variable and irregular
CEREBRAL POTENTIALS EVOKED BY PHOTIC STIMULI
-
CHRONOGRAPHY D.IO.I.!0.
EYES CLOSED
I I F/S
389
D E L
-
Fig. 10 Chronograms (bipolar recordings) evoked in the occipital area by successive trains of 200 photic stimuli delivered at the rate of 1 I flashes per second. A burst of flashes is given every 40 sec. Note the stability of the form of the evoked potentials; only the level of different points of the contour varies within narrow limits.
form; the peak is at about 75 msec. This wave does not correspond to any of those described by Cobb and Dawson (1960) or Ciganek (1958; 1961). The similarities between the initial part of the response at 11 flashes per sec and the PEM at 1 flash per sec make it possible to draw an analogy between the two, and to consider the first as a modified form of the second. The response to photic stimuli at very low frequency (Cobb and Dawson, 1961; Werre, Smith and Beek, passim) may be regarded as lasting several hundred msec. Thus, when the stimulus frequency is raised to 11 per sec, a flash arrives before the response evoked by the preceding one has terminated. This would bring about the phenomena of summation and facilitation, deforming the evoked potential and conferring new characteristics upon it. References p . 392
390
H. GASTAUT
et al.
(a) The researches of Van Hoff (1960) may be quoted in favour of the phenomenon of summation. This author was able, by a mathematical technique, to predict the different forms taken by the PEM in response to different stimulus frequencies, and he demonstrated the identity of theoretically derived curves. Van Hoff’s calculation for PEM’s evoked at 11 flashes per sec are sufficiently near to those obtained by us to permit the conclusion that summation does in fact play a role in this production. (b) In support of the idea of facilitation, Gastaut et al. (1952) described an excitatory cycle in the occipital cortex of man, with a facilitatory phase extending from the 70th to the 150th msec after the flash. It would appear reasonable to assume that the very large amplitude of the initial phase of the PEM evoked at 11 flashes per sec, is due to the fact that each new flash arrives during the period of facilitation provoked by its predecessor. 3. The response evoked by.flashes ut the rate of 14 per second The form is further modified; it is a more or less sinusoidaldiphasic wave with a negative initial phase. This aspect corresponds to that described by Cighnek (1958) at a flash frequency of 16 per sec, except that the peak latencies vary with the frequency of stimulation. Such a response, in which the polarity is the opposite of that obtained at a stimulus frequency of 11 per sec, and is of smaller amplitude, cannot be considered as a facilitation effect of the PEM evoked by stimulation at 1 flash per sec, even though the periodicity of the flashes used (73 msec) coincides, even more precisely than at a frequency of 11 flashes per sec, with the phase of facilitation of cortical excitability by Gastaut et al. (1952). On the other hand, it is possible that this response represents the algebraic sum of partiaIly superimposed potentials according to the theory of Van Hoff; but we cannot actually confirm this hypothesis, because Van Hoff has not studied the form of the responses evoked at a frequency higher than 11 flashes per sec. Consideration must be given to the possibility of a new phenomenon, unrelated to those occurring at lower stimulus frequencies. This is the view taken by Cighek (1961), which is based on the fact that, at a critical frequency of 16 flashes per sec photosensitive epileptic subjects show a discharge of diencephalic wave and spike activity. This author assigns a thalamic origin to the responses recorded from the scalp when stimulation is applied at such a frequency. We do not agree with this view, because : (a) We have obtained this initially negative sinusoidal response at other stimulus frequencies higher than 14 flashes per sec, and not only at 16 flashes per sec. (b) Epileptic fits are often induced by stimulation at 10 flashes per sec as well as 16 per sec, whereas the PEM’s evoked by the two stimulus frequencies are entirely different. 4. The general topographical distribution on the scalp of PEM’s at all frequencies It appears to us that the generalized distribution of responses is not a physiological phenomenon which occurs at the surface. That is to say, we do not believe that the
CEREBRAL POTENTIALS EVOKED BY PHOTIC STIMULI
39 1
potentials recorded from different parts of the scalp have been transmitted there by conduction along non-specific nervous pathways. It would seem rather to be due to a physical phenomenon. A study of the distribution of evoked potentials (at 11 flashes per sec, for example) on the posterior part of the skull leads us to consider the possibility of a sagittally-oriented dipole, whose point of maximum positivity is very close to the inion, while its negative fraction begins in the posterior parietal region. Such a hypothesis takes account of the inversion of topographical contours from the second or third electrode in bipolar recording with wide inter-electrode spacing and a reference electrode on the mastoid process. So far as the anterior regions are concerned, it is difficult to compare the results obtained by closely-spaced bipolar electrodes and those obtained by widely-spaced electrodes with a reference point at the mastoid. In fact, recordings made at a stimulus frequency of 11 flashes per sec, with a common reference electrode only reveal (at an interval of 25 or 30 msec) negativity at the second or third electrode. But two cases recorded with closely-spaced bipolar electrodes showed a secondary positivity in the frontal region. In order to explain this anterior positivity, CigAnek assumed the superimposition of the electric field of the eye itself, which is positive in frontal areas and negative in parietal ones. Considering the stimulus frequencies used, it appears doubtful to us that ocular movements at the same frequency as the flashes would be induced, as they are in the arousal reaction (Gastaut 1953; Zang-Zao 1953). The topographical contours of the second wave present the same difficulties. In some subjects, this contour may be simply explained by the existence of a dipole, the posterior extremity of which is negative[DON and FRE, recorded witha common reference electrode (Fig. 8); MAR with closely-spaced electrodes (Fig. 7)]. In others, the phenomenon is more complex, the most anterior regions showing a negative voltage 75 msec after the flash, in common with the posterior part of the scalp, while the parietal areas are positive. It is difficult to imagine that, in the 50 msec which separate the peaks of the first and second waves, the eyes could have moved to such an extent that their electric field in the anterior part of the scalp would have reversed its sign. This argument may be added to that stated in the preceding paragraph against the possible interference of the ocular electric field with the PEM at the stimulus frequencies employed by us. It is, however, possible that the electroretinogram may have an effect on the voltage recorded on the anterior part of the scalp, and we are at present performing experiments to test this hypothesis. CONCLUSIONS
We should point out that the difficulties with which we are faced in the interpretation of responses to flash frequencies of 11 and 14 per second are due only to a variation of the order of 1 microvolt, which supervenes in regions furthest from those in which the visual pathways terminate. On the other hand it can be stated (with the minor reservations given above) that the responses recorded in the posterior and middle regions (particularly at a stimulus References P. 392
392
H . GASTAUT et
al.
frequency of 11 flashes per sec) correspond to PEM’s which, because of the phenomena of summation and facilitation, are particularly large, symmetrical and stable. Because of this, their use may be suggested in psychology and diagnosis, as has already been suggested by R h o n d (1958; 1961). We have already noted that mental concentration can bring about significant changes in the level of different points in the PEM. We are at the moment studying the variations produced in the PEM by: various psychological conditions, such as a highly unpleasant environment; habituation and conditioning; and central lesions causing unilateral interruption of the visual pathways.
REFERENCES CIGANEK,L. Potentiels corticaux chez l’homme, evoques par des stimuli photiques. Rev. Neurol., 1958,9Y: 194-196. CIG~NEK, L. Die elektroencephulographische Lichtreizantwort der menschlichen Hirnrinde. Verlag der Slowakischen Akademie der Wissenschaften, Bratislava, 1961, I52 pp. COBS,W.A. and DAWSON, G.D. The latency and form in man of the occipital potential evoked by bright flashes. J. Physiol. (Lond.),1960,152: 108-121. GASTAUT, H., CORRIOL,J. et ROGER,A. Le cycle d’excitabilite des systBmes afferents corticaux chez l’homme. Electroenceph. elin. Neurophysiol., 1952,4: 235. S . Les pointes negativesevoquees sur le vertex. Rev. Neurol., 1953,89: 382-399. GASTAUT, R ~ M O N A. D , Recherche sur des renseignements significatifs dans les enregistrements psysiologiques et mecanisation possible. En : A.M.MONNIER(Redacteur), Actualitis neurophysiologiques, Masson, Paris, 1961,2: 167-21 1. R~MOND A., et RIPOCHE,A. Technique topographique: sur un compteur d‘integration, exemples d’utilisation. Rev. Neurol., 1958, Y9: 179-186. VAN HOFF,W. M. The relation between the cortical responses to flash and t o flicker in man. Acta ghysiol. pharmacol. need., 1960,9: 210-224. ZANG,ZAO.Le champ electrique de l’oeil en electroencephalographie. Arch. Neuro-psiquiaf. ( S . Paulo), 1 9 5 3 , I l : 1-16.
DISCUSSION C. AJMONEMARSAN: With regard to the different types of response observed with single flash stimulation on the one hand and with 14/sec photic stimuli on the other it should be possible to account, at least partially, for the difference itself. If 1 remember correctly, with single flashes one gets a positive deflection with its peak a t about 35 msec followed by a prominent negative deflection whose peak occurs about 65-75 msec after the stimulus. When the flashes are repeated at 14/sec (i.e. about every 71 msec) it is not surprising to see an initial negative potential practically without latency because each subsequent flash would fall almost exactly at the peak of the second (negative) deflection. W. GREYWALTER: (1) This fascinating contribution deserves extensive discussion, but I should like to put three questions of a simple technical nature. First, what was the intensity of the flash stimuli used and was the effect of varying it investigated? Second, the choice of electrode derivation system is of crucial importance
CEREBRAL POTENTIALS EVOKED BY PHOTIC STIMULI
393
in this type of study - were comparisons made between common reference, average reference and bipolar derivations? Third, to what extent were the subjects “pre-habituated” by testing and preliminary runs? We have found the very first stimulus of all, and certainly the first dozen, of unique importance. (2) The nature of the stimulus should certainly be considered more carefully by us all and I think we should investigate the effects of modulated continuous light as described by Van der Tweel. This might be more instructive and easier to analyse in its effects than the brutal all-or-none flash. TWEEL,L. H. VAN DER, SEM-JACOBSEN, C. W., KAMP,A., STORMVAN LEEUWEN, W. and VERINGA, F. T. H. Objective determination of response to modulated light. Acra physiol. pharmacol. neerl., 1958, 7 : 528.
A. FESSARD: Presumably, the modifications in visual evoked potentials as you observed by changing the flicker rate of illumination are accompanied by definite variations in the induced visual perception. Such changes have often been a theme for systematic studies by psychologists. I wonder if you have questioned your subjects on this matter.
H. GASTAUT’s replies: To C. Ajmone Marsan Van Hoff suggests (in the study previously mentioned) that phenomena of summation are responsible for the changes in average potential evoked as a function of the frequency of stimulation. According to Ciganek, the general form of the primary response is not changed when the frequency of stimulation is less than 10 flashes/second. Only the amplitude varies. We believe that phenomena of facilitation or inhibition (according t o the frequencies used) must intervene to modify the form of this response. Stimulating a t 15-16 flashes/second, we have seen responses very much like those obtained a t 14 flashes/second (the amplitude is diminished). The mechanism of modification of the average potential evoked as a function of the frequency of stimulation is certainly not unequivocal; at present we are trying to determine what role is played in this phenomenon on the one hand by processes of summation, and on the other hand by facilitation.
VANHOFF,W. M. The relation between the cortical responses to flash and to flicker in man. Acta physiol. pharmacol. neerl., 1960, 9 : 21 0-224. CIGANEK,L. Die elektroencephalographische Lichtreizarztwort der menschlichen Hirnrinde. Verlag der Slowakischen Akademie der Wissenschaften, Bratislava, 1961 : 152 pp. To W. Grey Walter First question (a) Unfortunately we are unable just now to furnish all the technical characteristics of the stroboscope used. The instant energy delivered during each flash is 0.3 J, and each flash lasts 50 psec. The stroboscope was a “VarCclat”, manufactured by the firm of Alvar and used under the same conditions as in the EEG laboratory for diagnosing photosensitive epilepsy; the tube about 12 cm from the eyes. (b) The question of the type of registration is of crucial importance. In the part of the paper concerning man, we describe results obtained by bipolar registration a t short distance and long distance bipolar registration with a common mastoid electrode of reference (monopolar if preferred). We are at present trying out several methods of derivation, with particular emphasis on finding a truly inactive site, relative to the phenomenon studied, for placing the electrode of reference. You know how long and delicate is the solution of this problem! (c) Some individuals are in fact so surprised by the first chain of flashes, even when they have watched the phenomenon being applied to another subject, that the first results are too artificial to be useful. The development of the response must be studied to the extent to which the subject passes from this alert, uncomfortable state to one of relaxation, which is relatively rapidly attained. At the moment we prefer to work only with subjects of good emotional stability, ignoring the results until the subject has become accustomed to the 4itiiation ( c j Fig. 10). Second question The investigations of Van der Tweel are certainly interesting; however, we have not used this technique for methodological reasons,
394
H. GASTAUT
et al.
We are at present attempting to define the mechanism by which intermittent light stimulation activates photosensitive epileptics. We must therefore use a stimulation which is effective in that sense. We believe that Van der Tweel’s apparatus is too complex and delicate for this work. Undoubtedly, however, the method of Van der Tweel is more refined than the flickering used in EEG laboratories, and more useful for the purely physiological study of mechanisms of vision.
To A . Fessard During the investigation, the results of which we presented, we have not especially questioned subjects as to the quality of their subjective experience. In this connection, as Professor Fessard suggests, there is room for very interesting experimentation. Cobb and Dawson have demonstrated that, at very low frequencies of stimulation, there are correlations between the real intensity of illumination (and the degree of attention) and the form of the average potential evoked. We intend to take up this investigation of the frequencies we use, considering not only the real but also the subjective intensities. COBB,W. and DAWSON, G. The latency and form in man of the occipital potential evoked by bright flashes. J. Physiol. (Lond.), 1960, 152: 108-121.
Specific and Non-Specific Cerebral Responses and Autonomic Mechanisms in Human Subjects during Conditioning W. GREY WALTER Burden Neurological Institute, Bristol (Great Britain)
Compared with the wealth of information about the non-specific pathways and responses in domestic animals, very little seems to be known about humans. The chief difficulty is that conventional scalp records are extremely difficult to interpret; in suitable experimental conditions the responses themselves are usually small often smaller than the “noise” or intrinsic activity - and furthermore they represent the compound activity of many regions. In particular, the responses to visual stimuli as seen in scalp derivations are only rarely specific in the sense of originating entirely in the visual projection areas, which are usually far from the scalp in the medial occipital cortex. The “non-specific” responses (evoked by stimulation in several modalities) are also peculiar in their topology, seeming to arise mainly at the vertex, but in fact involving wide regions of frontal and temporal cortex, including medial and orbital zones. Both sensory convergence and wide distribution are criteria of non-specificity . Special methods of recording and analysis are necessary to identify and measure these responses, and the experimental conditions must be specified and designed very carefully to provide large enough samples for statistical analysis while allowing for the effects of “habituation” or extinction due to monotonous repetition of stimuli. Since investigations in this domain are intended to reveal the relation between “significance” of stimuli and cerebral activity, strictly speaking no experiment can be repeated, for no two stimuli can have precisely the same significance. In the simplest case, the very first of a series of light flashes often evokes a particularly large response, while subsequent flashes at random intervals produce a steadily diminishing effect. But the very same stimuli, presented with or’before a powerful stimulus in another modality which acts as an unconditional signal for defence or gratification, may evoke steadily greater responses as the significance of the light flashes increases with accumulation of significance by invariable association. Even the nature and the distribution of the “contingent” response usually changes with repetition and with the details of the contingency; the inclusion of an operant response by the subject, for example, considerably modifies many features of contingent responses. With these considerations in mind several series of experiments have been made with human subjects, both normal volunteers, and patients referred for special Reference8 II. 401
396
W. GREY WALTER
investigation, firstly to find out the simplest way of studying the relation between conditioning, cerebral activity and autonomic mechanisms. The first series of investigations followed the lines of those already described (Walter 1960). The situation was essentially that of a Type I1 operant conditioned reflex (CR) in which two types of conditional stimuli were provided in both auditory and visual modalities and the unconditional stimulus was an alarming noise which the subjects could abbreviate or avoid by appropriate lever movement. All 58 subjects (37 normals, 21 patients) were studied completely in this way, and a further 255 experiments were made to investigate particular aspects of the EEG and autonomic effects encountered in the major experiments. In all these studies the EEG was analysed both with automatic frequency analysers and with Toposcopic displays specially adapted for the revelation of very small responses to rhythmic pattern stimulation. The resulting records were of great complexity and presented difficulties in data reduction and correlation, since four separate records had to be perused and co-ordinated. The results showed conclusively that the individual differences, which are familiar to all experimenters, cannot be ignored and must be accepted as a part of the experimental problem. The importance of these personal idiosyncrasies is indicated by their correlation with operational effectiveness; for example certain features in the time relations of cerebral and autonomic response were found to be closely associated with total incompetence of the subjects to establish conditional responses in these experiments. The basic hypothesis underlying these experiments was that the brain can act as a sequential computer of contingency when provided with signals from two or more sources and one of the objectives was to discover how, and if possible where, the cerebral mechanisms establish a suitable value for the level of confidence or probability on which external action may be based. In these terms the results of these experiments suggest: 1. that the intrinsic adaptive capacity of a human brain does in fact depend on its accurate computation of contingent association in real time; 2. that the manner in which the cerebral computation is performed varies considerably in different brains, particularly in the extent to which information from different modalities is combined and correlated ; 3. that interaction between neo-cortical and paleo-cortical systems (including autonomic functions) is essential for the initiation of the computation - but may impede its completion; 4. that the inter-personal distribution of the variables determining the above may be such as to justify a pragmatic, operational typology of human behaviour patterns. The information obtained from these experiments - together with the difficulty of appreciating and describing such elaborate records - suggested that a simplified version of this arrangement might provide adequate measurements particularly with the use of the recently developed Electronic Average Response Indicator. This instrument provides means for extracting average responses simultaneously from two channels (Cooper and Warren 1961) and can be arranged to accumulate 6, 12,24 or 48 samples at random or regular intervals, correlated with two or more stimuli, and covering durations of 10 msec-2 sec according to the specified situation. The averaged
CEREBRAL RESPONSES AND AUTONOMIC MECHANISMS
397
outputs are written out directly on a conventional ink recorder and thus form part of the normal record which includes registration of signals, PGR, EKG and respiration as well as EEG. The great gain in noise penetration with such a system has already been described (Dawson 1953; Brazier this Colloquium). Features which are quite invisible in the primary records are displayed unequivocally with only twelve samples, although this number provides a theoretical penetration gain of only about 3.5. The reason for the much greater operational efficacyof the averaging system over its expected performance is that the “noise” in these conditions has frequency and waveform characteristics very like those of the wanted signals which it therefore masks very confusingly. In practical application it has been found possible to measure the average of twelve samples of evoked responses, each equivalent to about 2 pV, masked by noise over the band 1-100 clsec equivalent to about 20 p V , peak-peak. With this system the specific responses in the occiput to visual stimuli can be detected without difficulty and distinguished from the non-specific responses to visual and auditory stimuli in the anterior regions, even when the stimulus intensity is quite low and the individual stimuli are separated by long and variable intervals. Records obtained by averaging from the scalp have been compared with derivations from deep within the brain, using chronic electrode implants in patients referred for special investigation and intracerebral therapy. In this way the strictly local responses within the brain can be related to their appearance as projected on the scalp. The experimental situation provides for stimulation with isolated or repetitive light flashes of any intensity, clicks, tones of various intensities and instrumental responses on the part of the subjects, either arbitrarily requested or in operant control of one or other of the stimuli. In summary, the results obtained show that: ( a ) The truly specific occipital responses to visual stimulation show little or no habituation and may include four or five phases within the primary response. These components are represented separately within the visual cortex. (b) The parieto-occipital responses on the scalp include both the specific and parietal non-specific responses, with long lasting after-effects related to phase-locked synchronisation of one component of the intrinsic rhythms (Fig. 1). This effect also is strictly localised within the occipital cortex and appears only in subjects with prominent intrinsic alpha rhythms. The synchronised after-rhythm and the nonspecific parietal components show marked habituation ; they are also diminished by distraction, drowsiness and by overbreathing, but are augmented by attention. association with specific stimuli and by inhalation of carbon dioxide (Walter 1962). (c) The anterior (fronto-vertical) non-specific responses to visual stimuli also contain at least four components. They have a latency little, if at all, longer than the occipital responses. They show habituation to a degree varying from subject to subject. When combined with an unconditional (auditory noise) stimulus they are extended in distribution and often prolonged, and this effect is even more marked with the addition of a noise-linked operant response, during the period of significance establishment, that is for 20-150 trials. References P. 401
398
W . GREY WALTER M.M. age 13 s/r2 years L.Occipita - L.Mastold
Average period 2 seconds Sample size N-12
3
R
Flash
B
5
W
I
I 1
Tone
Flash 2 sec
-
Fig. 1 Coherent phase-locked after rhythms in a girl aged 13. Scalp electrodes: left occipital to mastoid. Each tracing is the electronic average of twelve samples covering a period of 2 sec at random intervals of 0.5-5.0 sec. A : three averages of twelve responses each evoked by single flashes at random intervals, showing the coherence of the after-rhythm from trial to trial following the evoked response. B : the association of flash with tone augments the after-rhythm, which persists for the whole 2 sec even during the presentation of the auditory stimulus. Note the constant phase-relations in the three separate runs, each of which is the average of twelve trials; there is almost no phase-slip although over 600 separate waves are represented in these tracings.
(d) The non-specific (anterior) responses to auditory stimulation, though similar in their general time relations to the visual responses, are somewhat shorter in latency and have a distinctive pattern. These responses also show marked habituation and contingent recruitment when classed as conditional responses. Their most striking feature, however, in these experiments, is that when evoked by a loud noise (which acts as a specific stimulus for an instrumental or operant response) they show progressive attenuation as compared with the conditional (flash-evoked) response. When this has occurred, withdrawal of the warning, conditional flash results in immediate restoration of the non-specific auditory responses to a size and extent even greater than their original values (Fig. 2). This progressive attenuation and restoration can be repeated many times, and appears to some extent in all the adult subjects so far, though the number of trials needed for the process varies very considerably, at least from 12-60.
399
CEREBRAL RESPONSES AND AUTONOMIC MECHANISMS F3
-N-12
A Mld.Occ
-4 ‘
Dm.
110P
Flash only -2rc-
Q
1
I
Flesh t Noise
-
-+NtB+C
@
F+NtBtC
@
Fig. 2 Non-specific (C2ZC3) and specific (Mid Occ-T5) responses to visual and auditory stimuli in a normal man aged 20 at various stages during a conditioning experiment. Each record is the electronic average of twelve trials. A : non-specific (vertical) and specific (occipital) responses to twelve single flashes at random intervals of about 2 sec, preceded by 24 similar stimuli. Whereas the occipital response is a single positive wave, the vertical one contains several components. B : average of twelve sets of flash followed by loud noise. The non-specific vertical responses to flash are modified by the association and those to the noise are widely dispersed in time and space. C : “habituation” of non-specific responses after three presentations of the associated visual and auditory responses together with an instrumental response by the subject. The non-specific auditory response has almost disappeared. D:average of twelve responses to flash alone after 150 associations of flash with noise; the pattern is indistinguishable from that in A before association. E: restoration of the habituated non-specific auditory response by withdrawal of the conditional flash (compare with C and F ) . There is a marked amplification with less temporal and spatial dispersion. The subject’s instrumental response interrupted the noise. F: re-habituation of the non-specific auditory response by restoration of the warning flash. The nonspecific visual response at this final stage shows clear division into two components.
( e ) Correlated with these contingent variations in the non-specific anterior responses are changes in the pattern of autonomic variables. Apart from typical personal variations, the usual development is for the phase of novelty (orientation) of the specific (unconditional) auditory stimulus to be associated with tachycardia, dyspnoea and psycho-galvanic responses (PGR) of high amplitude and short (1.5-1.7 sec) latency (Walter 1960). In contrast, the effects of the visual (neutral or conditional) stimuli on the cardiorespiratory rhythms are minimal and they disappear by habituation after six to twelve trials. When the visual and auditory stimuli are associated, however, the PGR to flash reappears after a few presentations but with the short latency of the auditory responses. As the conditioning procedure is continued the latency of both conditional and unconditional responses gradually rises till both are at theconditional level of about 2.0 sec. References
8. 401
400
W. GREY WALTER
In general the phase of conditional PGR latency abbreviation and tachycardia corresponds with the period during which the significance of the light-noise contingency is being established, and the magnitude of both conditional (visual) and unconditional (auditory) non-specific responses is large. The phase of conditional autonomic response extinction corresponds with the state of unconditional non-specific attenuation. This suggests that the long latency of conditional PGR is due to its delay by the mechanism of cerebral computation and that the autonomic responses are an
Flash only
Flash with noise
51
em =Wr” A
I’ ECG
1
/
-
,
15eC
A-----+
4-
EMG I t hand
Fig. 3 Contingent augmentation of “non-specific” conditional responses recorded simultaneously from two implanted electrodes in human subject during conditioning. Electrode 6 is in the posterior orbital cortex about 15 mm from the midline on the left. Electrode 5 1 is in the medial frontal cortex about 15 mm from the midline on the right. Each record is the electronic average of twelve trials. The subject was a chronic psychoneurotic female patient aged 32 whose symptoms had recently been relieved by intracerebral polarisation at other electrodes. A : average of responses to twelve flashes presented alone at random intervals of about 2 sec. This series followed a period of similar stimulation leading to “habituation”. The responses are insignificant at electrode 6 and diminutive at electrode 51. B : average of responses to twelve combinations of (flash followed by loud noise) I sec later. Both electrodes show a clear response to both modalities. The response to flash (the conditional stimulus) is about twice as large as in A . C : average of twelve responses to (flash followed by noise and instrumental interruption of noise by subject). The responses to flash are further augmented at both electrodes by the double association and instrumental response but the response to the noise (the unconditional stimulus) is attenuated at the orbital electrode 6, while it persists at the medial electrode 51.
CEREBRAL RESPONSES AND AUTONOMIC MECHANISMS
401
essential component of the preliminary stages of conditional association or contingency computation. Intra-cerebral location studies suggest that the attenuation of the unconditional non-specific responses with repetition is more marked and stable in the orbital and other limbic regions than in the medial and lateral frontal cortex (Fig. 3). In colloquial terms these observations (which are still in progress) suggest that in most normal human brains presentation of a novel stimulus in any modality evokes two types of widely dispersed cerebral responses, one type associated with alarm and the other with attention. The alarm responses in limbic cortex evoke rapid and general autonomic action, and show rapid and stable habituation. The attention responses, mainly in medial and lateral frontal cortex, also show habituation but can be progressively augmented by significant association with unconditional stimuli even when these latter have lost their alarming or novel character by repetition. The novelty or “orienting” response is a compound of these two mechanisms. In abnormal mental states, both the tendency to habituation and to contingent amplification are liable to disturbance leading to states of panic and/or adaptive inertia.
REFERENCES COOPER, R. and WARREN, W. J. The use of barrier grid storage tubes type 951 IA for extraction of average evoked responses from the EEG. . I Physiol. . (Lond.), 1961, 157: 38. WALTER, W. GREY.A statistical approach to the theory of conditioning. Electroenceph. clin. Neurophysiol., 1960, Suppl. 13: 311-391. WALTER, W. GREY.Spontaneous oscillatory systems and alterations in stability. In: R. G. GRENELL (Editor), Progress in neurobiology. Vol. V : Neural Physiopathology. Hoeber, New York. 1962, VoZ. 5 : Chapter 1.
DlSCU SS ION H. GASTAUT: I asked to be the first to speak after Grey Walter’s paper in order to be able to congratulate him while I am still under the spell of the enthusiasm which his well-documented report aroused in me. In his magnificent presentation 1 believe I see formal proof of the necessity of knowledge of the overall electrical signs of cerebral electrical activity, to the same extent as partial signs and sometimes even in the same way. Certainly, regional and local niacrophysiology and microphysiology, which were the concern of nearly all the papers presented here, are indispensable aids in solving the problems of the overall electrophysiology of the brain; but it is nevertheless necessary to pose these problems, and Grey Walter can only be congratulated for having done this in such a masterful manner. W. R. ADEY: The evidence presented by Dr. Grey Walter impressively supports the value of automatic analytical methods which reveal the nature of phase patterns in EEG waves recorded in cerebral systems during the learning process. As he has indicated, we have used auto-correlation, cross-correlation. crossspectral, coherence and phase-modulation analyses of our EEG data from the cat in discriminative training procedures (Adey et a / . 1960,1961;Adey 1961a). These studies have indicated in auto-correlograms the appearance of “specific” hippocampal
402
W . GREY WALTER
rhythmic bursts in the course of discriminative motor performances, at an essentially single frequency of about 5.5 cjsec. Similar, but less regular, rhythmic bursts were seen simultaneously in extrahippocampal structures, including the midbrain reticular formation and the primary visual and sensorimotor cortical areas. Using cross-correlation analysis, we have observed consistent shifts in phase patterns between different hippocampal regions in the course of learning discriminative motor performances. In the fully trained animal, it was found with this analytical method that there were consistent differences in phase patterns within the hippocampal system in the course of performance of correct and incorrect discriminative motor acts. We have initiated the use of a more comprehensive form of EEG analysis called cross-spectral analysis, which permits, inter nlin, assessment of phase, amplitude and coherence relations between two simultaneous records across a continuous spectrum of frequencies. These studies have confirmed earlier cross-correlation analyses which had indicated radically different phase patterns between correct and incorrect responses in the fully trained animal. Recently, with the engineering assistance of our colleagues Blum, Brown and Weinberg, and the computing facilities of the Space Technology Laboratories, we have initiated examination of small amounts of phase modulation occurring in an apparently regular burst of slow waves at an apparently single frequency. This study has used a mathematical analysis devised by Goodman ( I 960). It has been successfully applied to the detection of very small amounts of phase modulation in magnetometer readings from earth-orbiting satellites, occasioned by variations in the earth’s magnetic field. Our own application of this method has indicated the occurrence of small amounts of rhythmic phase modulation in the 5.5 ckec bursts of hippocampal slow waves at moments of maximal attention. We have also noticed sudden momentary disruptions of regular phase patterns, suggesting that perhaps there are sudden changes in attention, or, more speculatively, that the process of consciousness itself may be discontinuous. These studies have been aimed at testing the possibility that the conveyance of information in cerebral systems may be related to the wave processes themselves, with neurons functioning as phasecomparators for these graded, analogous wave processes. These problems have been discussed in detail elsewhere (Adey 1960) .There is evidence that the exceedingly regular hippocampal wave trains during the discriminative performance may act as “pacemakers” for comparison with similar but less regular bursts occurring simultaneously in reticular and primary sensory pathways. ADEY,W. R. Brain Mechanism and the Learning ProcesA. C . JudJon Herrick Memorial Symposium, Chicago, 1960. Fed. Proc., 1961a. ADEY,W. R. Biomedical Engineering Symposium. Institute of Radio Engineers. San Diego Naval Hospital, 1961b. D. Unpublished observations. ADEY,W. R.,BROWN,D. and WEINBERG, C. W. and HENDRIX, C. E. Hippocampal slow waves; distribution and phase ADEY,W. R., DUNLOP, relations in the course of approach learning. A.M.A. Arch. Neurul., 1960, 3: 7490. C. E. Computer techniques in correlation and spectral ADEY,W. R., WALTER,D. 0. and HENDRIX, analyses of cerebral slow waves during discriminative behavior. Exp. Neurol., 1961, 3: 501-524. GOODMAN, N. R. Measuring amplitude and phase. J . Franklin Inst., 1960, 270: 437450.
reply to W. R . Adey W. GREYWALTER’S
The observations reported by Adey are most encouraging as evidence for functional significance of intrinsic brain rhythms. The effects of quite small phase shifts could easily be large enough to act as signal carriers, demodulated by neuron groups acting as coincidence indicators. My colleagues Cooper and Mundy Castle (1960) have described very similar abrupt shifts in human records analysed with our toposcope. They called these “instantaneous phase changes” and noted that they were more frequent at the beginning of a defensive CR experiment, and were associated with the development phase of a motor response. There is one technical point I should like to raise. Cooper (1959) has shown that with bipolar recording the time relations of an electric disturbance moving past an electrode pair may be quite deceptive, since the vectors in effect rotate abruptly. The records Adey describes were taken, I believe, with closely spaced bipolar implanted electrodes, so the phase shifts might be ambiguous as indicators of propagation rate. Inevitably bipolar recording indicates the spnce derivative of the electric fields in the immediate neighbourhood of the electrodes, and it would be valuable to know what the methods
CEREBRAL RESPONSES AND AUTONOMIC MECHANISMS
403
Adey describes would reveal for records taken with a common reference or average reference system. However, we should not discount the value of close bipolar derivation; the nervous structures themselves probably “see” the local fields in rather the same way, as space-time differences of potential and the exaggerated time-differences thus produced may be essential for their function. COOPER, R . An ambiguity of bipolar recording. Electroenceph. clin. Neurophysiol., 1959,II: 819820. COOPER, R. and MUNDY-CASTLE, A. C. Spatial and temporal characteristics of the alpha rhythm: a toposcopic analysis. Electroenceph. d i n . Neurophysiol., 1960, 12: 153-1 65.
SLEEP MECHANISMS Chnirman: J. M. BROOKHART
A Study of the Signs of Sleep in the Cat * GlAN FRANC0 ROSS1 Clinica Neurochirurgica dell'Universitri, Cenova (Italia)
I would like to report the results of researches carried out by Drs. Candia, Favale, Giussani and myself on the blood pressure changes occurring during sleep. During the last few years we have been engaged in the study of the mechanisms of sleep in intact, unanaesthetized and unrestrained cats. Obviously, such a study can yield reliable results only if unmistakable and clear signs of both the occurrence and the degree of depth of sleep are available. When we began our researches, information on the sleep-wakefulness rhythm was provided by the EEG (see Rossi and Zanchetti 1957 for references) and by the EMG recorded from the posterior cervical muscles (see Jouvet 1961 for references). However, several observations indicated that, in some instances, both these tests could be either misleading or incomplete. These observations prompted us to search for a new test which might perfect and integrate the evidence provided by the EEG and the EMG in evaluating the depth of sleep. Recordings of the heart rate and the respiration gave variable results. Quite satisfactory results were obtained, however, by studying the level of the systemic blood pressure. The blood pressure was recorded through a polythene cannula permanently inserted into either the femoral or the carotid artery. Blood clotting was avoided by filling the cannula with a heparin solution and by keeping the animal heparinized. The changes in the blood pressure during sleep were analysed and compared with the changes occurring in the electrocortical rhythms and in the electrical activity of the posterior cervical muscles recorded through permanently implanted electrodes. The results obtained may be summarized as follows. 1. There is usually no significant change in the blood pressure at the very beginning of sleep, when the first bursts of 8-15 cjsec waves appear on the EEG and when the
* The experiments reported here have been sponsored in part by the Air Force Office of Scientific Research through the European Office, Aerospace Research, United States Air Force, under Contract AF 61(052)-461.
SIGNS OF SLEEP I N THE CAT
405
EMG activity is of moderate amplitude (Fig. 1). A slight fall in blood pressure may occur when synchronization of the electrocortical rhythms is fully developed and slow 2-4 c/sec waves are present in the EEG recording. A constant and marked fall in blood pressure is observed (ranging from 15 to 35 mm Hg) during deep sleep, when the EEG rhythms are desynchronized and the EMG becomes completely flat (Fig. 1). 2. The level of the blood pressure may show oscillations of variable amplitude during deep sleep in spite of the uniformity of the EEG and the EMG patterns. The threshold of behavioural arousal following reticular or sensory stimulation is not constant during deep sleep. The threshold variations are related to the changes in the blood pressure level, the highest thresholds observed coinciding with the lowest arterial pressures and vice versa.
Fig. 1 EMG, EEG and blood pressure changes from relaxed wakefulness to deep sleep. In this and in the following figure bipolar recording from the posterior cervical muscles (EMG), bipolar EEG recordings from the left fronto-temporal (LFT) and temporo-occipital (LTO) regions, and blood pressure recording from the right common carotid artery (BP) are reproduced. No changes in the blood pressure during the transition from relaxed wakefulness to light sleep. After 3 min marked fall in blood pressure during the transition from light to deep sleep. (From 0. Candia, E. Favale, A. Giussani, and G. F. Rossi, unpublished figure.)
3. As previously reported (Jouvet 1961; Rossi et al. 1961), high frequency stimulation of the reticular formation, when carried out during sleep with EEG synchro-
Fig. 2 Induction of EMG, EEG and blood pressure patterns of deep sleep by bipolar stimulation of the midbrain reticular formation (Ret.) with 1 msec, 300/sec, 2.8 V rectangular pulses. (Modified from Candia et al., 1962.) References P . 406
406
SLEEP MECHANISMS
nization, may be followed by the appearance of the typical EEG and EMG patterns of deep sleep. In this case, too, the blood pressure falls to the level reached during the spontaneous episodes of deep sleep (Fig. 2). Controls have shown that the same reticular stimulation does not affect the EMG and the blood pressure when it is performed during the waking state. These results indicate that the study of the blood pressure may provide useful information on the depth of sleep. A general relationship between the two phenomena has in fact been found: the deeper the sleep, the lower the level of the arterial pressure. However, the most constant and important finding has been the critical occurrence of a marked fall in blood pressure coinciding with the disappearance of any EEG synchronization and with complete muscular relaxation. We believe that this coincidence is likely to have some physiological significance. These findings strongly support the view of Jouvet (1961) that sleep should be subdivided into two fundamental phases : light sleep, characterized by EEG synchronization, moderate muscular tonus and basal blood pressure level; deep sleep, characterized by EEG desynchronization, EMG flattening and marked decrease of the blood pressure. The recognition of this sharp subdivision of sleep into two different stages will probably be of help in the study of the mechanisms underlying sleep itself. REFERENCES O., FAVALE, E., GIUSSANI, A. and ROSSI,G. F. Blood pressure during natural sleep and CANUIA, during sleep induced by electrical stimulation of the brain stem reticular formation. Arch. ital. Biol., 1962, 100: 216-233. JOUVET,M . Telencephalic and rhombencephalic sleep in the cat. In G. E. W. WOLSTENHOLME and M. O'CONNOR (Editors), The nature of sleep. A Ciba Foundation Symposium. Churchill, London, 1961 : 188-206. Rossr, G. F., FAVALE, E., HARA,T., GIUSSANI, A. and SACCO,G . Researches on the nervous mechanisms underlying deep sleep in the cat. Arch. ifal.Biol., 1961, YY: 270-292. Rossr, G. F. and ZANCHETTI, A. The brain stem reticular formation. Anatomy and physiology. Arch. ital. Biol., 1957, 95: 199-435.
The Rhombencephalic Phase of Sleep * M. JOUVET Institut de Physiologie, FacultP de Midecine, Universite' cle Lyon (France)
The EEG in chronic cats is characterized by three principal types of activity, viz., rapid low-voltage activity, synchronized spindle activity and high-voltage slow waves.
* This investigation was made possible by the support of The Air Force Office of Scientific Research through the European Office, Aerospace Research, United States Air Force, under contracts A F 61 (052)109 and A F 61 (052)472.
407
RHOMBENCEPHALIC PHASE OF SLEEP
Corresponding with these three types are the wide range of different levels of wakefulness, between attention and deep sleep. In the present state of our methods of analysing EEG activity, it is not surprising that similar cortical activities may correspond with widely different levels of wakefulness. Initially described in the course of neuropharmacological investigations (slow activity with wakeful behaviour after injection of atropine -Wikler 1952; rapid cortical activity during ether anaesthesia Adrian and Matthews 1934, Bremer 1936b), these dissociations between the EEG and the state of wakefulness were subsequentIy observed under physiological conditions. The most paradoxical is the rapid cortical activity in the course of physiological sleep, described by Derbyshire et al. (1936) and Dement (1958), who showed that this activity occurs in cycles during sleep and is accompanied by ocular movements. In the past few years it has become evident that the cortical EEG activity considered in isolation - is an unreliable index of the various levels of wakefulness. Consequently our attempts to correlate the EEG with behaviour require more exact methods of analysis. Among these, a few seem to be particularly adequate, viz.: (a) a study of correlations between the EEG and certain somatovegetative phenomena
SMC
A
B
C
Fig. 3 The two EEG aspects of sleep in intact animals (cats). A : wakefulness; rapid cortical-subcortical activity. B : “slow wave phase of sleep”; cortical spindles and slow waves, hippocampal high-voltage spikes, reticular slow waves, slight diminution of the EMG of the nuchal muscles, increase in the plethysmographic index. C : “RPS”; rapid cortical activity identical with that of wakefulness. Slow rhythmic activity of 5 c/sec at the level of the ventral hippocampus and the n. RPC (PRF), where spindle activity also occurs. Complete disappearance of the nuchal EMG activity, ocular movements. Deceleration of the cardiac rhythm (EKC) and increased plethysmographic index; respiratory irregularity. SMC: sensorimotor cortex; E X ’ : ectosylvian cortex; MRF: mesencephalicreticular formation. Reference$ p . 423-424
408
SLEEP MECHANISMS
which are pathognomic of a state of wakefulness; (6) recording of subcortical activity in order to find characteristic regional electrical patterns; (L') an attempt at dissociation, by means of selective sections through the brainstem, of various corticopetal afferences responsible for the same cortical activity; ( d ) finally, micro-electrode analyses of the various neuronal populations in the cortex responsible for similar EEG activity. The results to be discussed in this paper were obtained by the first mentioned three methods in the course of investigations into the nerve structures responsible for physiological sleep (Jouvet et al. 1958, 1960; Jouvet 1961, 1962a). They made it possible to define a pontolimbic system responsible for the phase of rapid cortical activity during sleep, and to distinguish this from the ascending reticular activating system (ARAS) (Moruzzi and Magoun 1949) responsible for the cortical arousal reaction which has similar EEG features. METHODS
Experiments were carried out on 68 intact chronic cats, decorticate preparations, decerebellate preparations, or preparations with total dissection or limited coagulation of the brain stem. The animals carried chronically implanted cortical and subcortical electrodes. The electromyographic activity of the nuchal muscles was systematically recorded with the bipolar electrodes in situ. Ocular movements, the ECG, the plethysmogram of the foreleg, and the respiration were also recorded. The animals were studied as a rule for at least 6 days, and EEG's were obtained in the course of different phases of physiological sleep. The electrode positions and the topography of the lesions were verified in all cases by serial sections, stained with Nissl, Loyez and Luxol Fast Blue. RESULTS
I. The two EEG aspects of physiological sleep in the intact cats (Fig. 3) A. Slow sleep (60-70:;, of the duration of behavioural sleep) is characterized by the appearance of spindle activity of 12-18 cjsec at the level of the cortex, thalamus, and mesencephalic and pontine reticular formations (RF). Slow waves (2-4 cjsec) subsequently appear at the level of these same structures. During this time the activity of the hippocampus is characterized by high-voltage spikes (500-800 p V ) . Throughout this stage there is always EMG activity from the neck muscles and the threshold of arousal by reticular stimulation is raised from 20 to 30 per cent relative to the stage of dropping off to sleep. This slow wave phase of sleep seems to be the expression of a corticofugal activity, for it is no longer possible to record spindles or slow waves from various subcortical structures in completely decorticate animals, or behind a section of the brain stem at the mesodiencephalic border. Arguments in favour of a telencephalic origin of the slow wave phase of sleep have been presented elsewhere (Jouvet and Michel 1958; Jouvet 1962a). B. Rapid wave phase of sleep or rhombencephalic phase of sleep (RPS). This
409
RHOMBENCEPHALIC PHASE OF SLEEP
phase always follows a slow wave phase and never occurs immediately after awakening. From the point of view of behaviour, it is characterized by two types of phenomena. On the one hand there is total atonia of the animal, characterized by an abrupt dropping of the head. Parallel with this, the EMG activity of the nuchnl muscles always disappears completely. On the other hand, there appears a pattern of highly characteristic paroxysmal phenomena. The most important of these consists of ocular movements while the palpebral fissure opens slightly, revealing the completely contracted nictitating membrane and filiform myosis. These movements are rapid and explosive, lateral or vertical, and conjugated. At the same time there are jerky movements of the vibrissae, less often brief twitches of the ears, jaws, tail or limbs. Finally there appear constant changes in the vegetative system : the respiratory rhythm becomes irregular, more superficial and especially more rapid than during the slow wave phase of sleep, while the cardiac rhythm may either show distinct acceleration or, more often, a deceleration. From an EEG point of view, the RPS is characterized by cortical, diencephalic and mesencephalic rapid, low-voltage activity similar to that seen in the waking state. This activity coincides with disappearance of the nuchal EMG or it follows or precedes this disappearance by a few seconds. The ventral and dorsal hippocampus consequently presents an almost continuous, very regular theta-rhythm of 5 c/sec. At the same time there appear a t the level of the n. reticularis pontis caudalis (RPC) high-voltage waves at 8 c/sec, in the form of spindles. These are sometimes associated with a 5 c/sec rhythm similar to that of the hippocampus. An identical activity has been recorded at A
1
2
3
4
9
E
Fig. 4 Periodicity of the RPS: schematic representation of a 4 h session of recording in five different chronic preparations. Above the horizontal line: cortical EEG activity. Vertical hatching: resting stage. Oblique hatching: slow sleep. White rectangles: activation during the RPS. Black strokes: arousal. Below the line, the black rectangles represent disappearance of the EMG of the neck and the rhythmic activity at the level of the n. RPC, corresponding with the periodic appearance of spontaneous RPS. A : intact animal. B : total decortication. C : total section of the mesencephalon. D : coagulation of the n. RPC. E : retropontine section. Time marks in hours, References p 423-424
410
SLEEP MECHANISMS
the level of the peri-aqueductal grey matter, the interpeduncular nucleus and the posterior hypothalamus. During this rapid phase, cortical or reticular auditory evoked responses are of lower amplitude than in the waking state or in the slow wave phase of sleep. The rapid wave phase of sleep, with an average duration of 10-15 min, is periodically repeated at intervals of 10-30 min during behavioural sleep. Thus it corresponds to 30-40 per cent of the total duration of sleep. During the rapid phase, the threshold of arousal by reticular or sensory stimulation is appreciably raised as compared with that in the slow wave phase. This characteristic, which is associated with subcortical EEG signs, complete lack of muscular tonus and
2&C
Fig. 5 RPS in completely decorticate cats. A : the animal’s initial behaviour is quiet, but there is no subcortical slow activity. Suddenly the head drops (disappearance of nuchal EMG) and rapid ocular movements appear. At the level of the dorsal hippocampus (H) and the ventral part of the ARF (level of interpeduncular nucleus), a slow rhythmic activity of 5 c/sec occurs. The pontine electrode ( P R F ) is localized at the brachium conjunctivum. In B 40 sec after A : the black stroke represents an acoustic stimulus of great intensity. The animal awakens. Calibration: SO pV.
RHOMBENCEPHALIC PHASE OF SLEEP
41 1
somato-vegetative phenomena, makes it possible to detect the periodicappearanceofthe RPS in animals following section or ablations of the neuraxis at various levels (Fig. 4). 11. Structures responsible for the RPS
A. Ablations of the cerebellum, the neocortex, and successive sections of the brain stem were carried out to determine the structures required and sufficient for the periodic appearance of the RPS. 1. The cerebellum, which could be safely presumed to play a r81e in the atonia which is so characteristic of the RPS, is not involved in this respect, because episodes of RPS, typical from an EEG and behavioural point of view, continue to occur in animals after total cerebellectomy. 2. In animals with total neodecortication, the RPS appears very clearly (Fig. 5). In the EEG there are high-voltage spindles at the level of the RPC. The hippocampal activity develops a 5 c/sec rhythm, while the thalamus and the other subcortical structures continue to show the same constant rapid rhythm of very low voltage. From a behavioural point of view, there is a complete disappearance of muscular tonus, and disappearance of EMG activity. Eye movements are similar to those in normal animals. The duration of the RPS is 10-15 min. 3. Section o f t h e brain stem (Fig. 6). Total sections of the mesencephalon do not
Fig. 6 Delimitation of structures triggering the RPS. Mediosagittal section of the brain stem and coordinates of the Horsley-Clarke planes A10, A0 and P10. Total brain stem sections in seven animals are indicated by black strokes. All sections but Gsuppress the cortical activation during wakefulness and the RPS. The sections A , Band Cleave intact the behavioural RPS with some ocular movements. After sections D, E and F, the ocular movements disappear during the RPS, but there is still disappearance of muscular tonus and a slow rhythmic activity at the level of the n. RPC. After section G, passing on the dorsal side of the knee of the facial nerve, and on the ventral side of the trapezoid bodies (TB), there is no longer a periodic disappearance of hypertonia, and all the peripheral phenomena of RPS are suppressed. In black: lesion of the n. RPC suppressing the RPS. M C : mamillary corpuscles; R N : red nucleus; n.rr1: nucleus of the third nerve; I P N ; interpeduncular nucleus; pyr: pyramids. References P. 423-424
412
SLEEP MECHANISMS
prevent the periodic occurrence of the RPS. A typical case is that of an animal in which the entire encephalon in front of the pons was resected. During the 9 days' survival, two distinct behavioural states were seen (Fig. 7). The first was characterized by considerable rigidity and hyperextension of the front legs. At this stage, the subcortical activity recorded at the nucleus reticularis pontis oralis (RPO) and RPC was rapid and of low voltage, while the electrical activity in the nuchal muscles was very considerable. The second state was characterized by RPS. The periodicity and duration were similar to those in intact animals and their aspects identical with the picture of
RPCV
ECG
RESP
1 s
Fig. I RPS in pontine animals (section E of Fig. 6 ) . Top: wakeful state - rapid low voltage activity at the level of the nucleus RPO and the RPC. Important EMG activity in nuchal region (hypertonia of decerebration). Bottom: RPS - rhythmic activity of 3 c/sec in the brain stem, particularly distinct at the level of the derivations of the nucleus RPC, where there is also paroxysmal activity resembling that of spindles. Disappearance of EMG activity. Cardiac deceleration and respiratory acceleration. RPC Vand D : ventral and dorsal derivation of the n. RPC. Calibration: 50 pV.
RHOMBENCEPHALIC PHASE OF SLEEP
413
cataplexia, described by Bard and Macht (1958) in mesencephalic and chronic pontine preparations. They were characterized by complete flaccidity of the animal and disappearance of the nuchal EMG. There were no eye movements but only some palpebral blinking, deceleration of the cardiac rhythm and respiratory acceleration. From an EEG point of view there appeared rhythmic 3 c/sec activity at the level of the n. RPC, resembling that recorded from the hippocampus in normal animals. This activity was mixed high-voltage spindle activity, synchronized with respiratory movements. It thus seemed established that the nerve structures responsible for the RPS are localized caudal to the mesencephalon. Determination of the posterior delimitation of these structures was a much more delicate problem because it is extremely difficult to keep animals alive for several days after section below the pons. Two preparations survived for a week after total ablation of the cerebellum and total section of the brain stem passing dorsally at the level of the posterior two-thirds of the n. RPC, and ventrally terminating at the junction of the pons and the trapezoid bodies. The phenomena which depend upon structures localized above the section can be divided into three EEG and behavioural stages, the interpretation of which is not always easy. (a) A stage of obvious cerebral wakefulness: the pupils are dilated 2-3 mm; the nictitating membranes are not visible and the eyeballs show spontaneous vertical movements. The eyes may consistently follow an object moving vertically in the field of vision. During this stage there is a rapid low-voltage corticogram. This aspect corresponds exactly with that described in midpontine specimens by Batini et al. (1959a,b,c). (b) A stage of cerebral sleep: the pupils are myotic and there are still some rare spontaneous movements of the eyeballs. But no movement of fixation is seen when a moving object traverses the visual field. Spindles occur at the cortical level. The spindle bursts are very short during the first few days (15-200/, of the duration of recording); they subsequently become protracted and, by the 7th day, constitute 50 per cent of the total time of recording. (c) Apart from these two stages, which are relatively readily interpreted in terms of behaviour, there is a “stage of difficult interpretation”: the pupils are very tightly contracted; there may be some spontaneous ocular movements but the animal is no longer capable of following a moving object. However, cerebral electrical activity is rapid. When an olfactory stimulus is applied at this stage, slight mydriasis is seen and the animal may then follow an object although cerebral activity does not change. On the other hand, stimulation of the hippocampus during this period causes the appearance of cortical spindles. The cerebral structures localized behind the brain stem section are responsible, on the other hand, for a state of permanent wakefulness characterized by considerable EMG activity in the neck, with never a period of atonia associated with disappearance of EMG activity. Analysis of the cardiac rhythm and]the respiratory rhythm has not revealed a periodic variation analogous to a RPS. After retropontine section there is a distinct predominance of rapid cortical activity. Kelerences 8. 423-424
414
SLEEP MECHANISMS
The periods of rapid cortical activity are as a rule associated with a waking cerebral behavioural state but, within these periods of rapid activity, there also seem to exist phases of cephalic sleep. In the absence of concomitant peripheral indices, however, it is impossible to determine whether these phases correspond with the analogue of a cerebral RPS; new investigations must be made if this problem is to be solved. On the other hand, behind the section there is a “state of wakefulness” which is lasting and which is expressed in continuous muscular activity. The nerve structures responsible for the RPS, therefore, are localized anterior to a retropontine section and posterior to a prepontine section. It is for this reason that circumscribed lesions were made at the level of the pontine R F with the object of selective suppression of the EEG and behavioural aspects of the RPS (Fig. 8). These experiments have shown that destruction of the n. RPC brings about complete disappearance of the cortical and peripheral phenomena of the RPS. In six animals, the lesion affected exclusively four-fifths of or the entire n. RPC, leaving intact the gigantocellular nucleus. It included the posterior quarter and the inferior quarter of the n. RPO. Ln a fifth animal, the lesion involved the entire n. RPC and the median and anterior three-quarters of the gigantocellular nucleus. The lesions left intact the entire n. RPO. The lesion common to all these animals, therefore, was that of the nucleus RPC. These cats were able to keep upright and capable of a staggering gait. They could eat. All showed periodic paroxysmal changes in behaviour from the 3rd to 4th day on. Periodically (2-3 times/h) they fixed their gaze, lifted the head with dilated pupils
Fig. 8 Destruction of the nucleus RPC (Loyez). This animal has not shown a RPS for more than a month. There was electrical arousal activity and normal slow sleep.
RHOMBENCEPHALIC PHASE OF SLEEP
415
and ceaselessly tried to reach an imaginary object with a front paw. These reactions of a “hallucinatory” type were associated with rapid cortical and rhythmic hippocampal activity. From an EEG point of view, the records of the first 2 days consisted almost entirely of a rapid low-voltage corticogram and hippocampal theta activity associated with wakeful behaviour. In the course of subsequent days, however, slow wave phases of sleep with cortical spindles and both cortical as well as subcortical slow waves reappeared during 50-60 per cent of the duration of recording. In other animals the slow wave phases of sleep re-appeared on the day following the intervention. In all animals the EEG and behavioural arousal reaction were normal. At no time was a return of the RPS observed, although observations covered a continuous period of 12 h/day and sometimes the fuIl 24 h. Although the total duration of recording in some animals amounted to over 200 h, no rapid activity during behavioural sleep, no relaxation of muscular tonus with disappearance of the nuchal EMG were observed. Towards the 15th day, however, the RPS returned in four animals. The phases were very brief and never exceeded 2-3 min. They occurred with a periodicity identical to that in the preoperative animals, but their duration did not even amount to 1 per cent of the duration of recording. Destruction of the nucleus RPC, therefore, entails the complete disappearance of EEG and peripheral phenomena characteristic of the RPS. For this reason, we suggest that the phase of rapid cortical activity during sleep be named the Rhombencephalic Phase of Sleep; this designation would seem to be more explicit than the term “paradoxical phase” which we previously suggested (Jouvet et al. 1959).
B. The nerve pathways responsible for the EEG phenomena during the RPS Once the structures triggering the RPS had been localized at the level of the n. RPC, circumscribed lesions were made in the brain stem, in front of the nucleus, in an attempt at selective suppression of the neocortical activation and the rhythmic hippocampal activity characteristic of this phase of sleep. 1. Lesions of the lateral parts of the brain stem, aflecting the specific ascending pathways and leaving intact the mesencephalic R F and the ventral part of the brain stem, did not affect the neocortical and paleocortical EEG signs of the RPS, slow sleep and wakefulness, and their somatovegetative correlations. They did not, there[ore, affect the pathways of ascending projection of the system responsible for the “cortical activation” during sleep. 2. Lesions involving the mesencephalic tegmentum and leaving intact the ventral one-third of the mesencephalon, suppress the cortical activity during arousal by nociceptive stimulation or by stimulation of the RF. Such lesions, however, do not suppress possible rapid cortical activity during the RPS. Animals thus treated, therefore, show rapid cortical activity only during the most profound phase of sleep. This fact indicates that the pathways responsible for the rapid cortical activity during the RPS do not generally make use of the mesencephalic RF, at least not entirely. 3. Lesions were made at the following levels (Fig. 9): median part of the n. RPO, n. centralis superior of Bechterew, the region of the interpeduncular nucleus and the References Y. 423-424
416
SLEEP MECHANISMS
peri-ependymal grey substance, subthalamic region (median hypothalamus, lateral hypothalamus, medial forebrain bundle) and the septum. This produced partial or complete suppression of the cortical activation and rhythmic hippocampal activity during the RPS. But these lesions did not affect the cortical activation during the waking state. During the RPS, tracings can sometimes show slight activation relative to the slow phase of sleep, but spindles persist and rhythmic hippocampal activity is absent (Fig. 10). The other symptoms of the RPS, however, are complete. Thus, lesions of the ascending part of the limbic midbrain circuit (Nauta and Kuypers 1957; Nauta 1958) prevent the appearance of EEG signs of the neocortical and paleocortical type, characteristic of the RPS. i t is therefore likely that the ascending pathways issuing from the n. RPC make at least partial use of such a system.
A9
PO
Fig. 9 Lesions suppressing cortical activation during the RPS. Top. Left: coagulation of the subthalamic region and part of the thalamic intralaminar nuclei. Right: complete destruction of the septum. Absence of activation for 5 days, followed by recovery. Cenfre. Left: destruction of the interpeduncular region and the peri-ependymal grey matter. Right: destruction of the nuclei of Rapht at the level of the n. RPO. Bottom. Left: destruction of the n. centralis superior of Bechterew. Right: coagulation of the median part of the n. RPO and n. RPC. All animals showed rapid cortical activity during the wakeful state.
41 7
RHOMBENCEPHALIC PHASE OF SLEEP
The periodic occurrence of the RPS is governed by an active mechanism. It is possible to trigger RPS periods - with typical EEG and behavioural features lasting 10-15 min - by stimulating the pontine RF or the mesencephalic R F for a few seconds during the slow wave phase of sleep. This phenomenon has also been attained in decorticate and chronic mesencephalic animals. The triggering of the RPS by stimulation of the brain stem is governed by welldefined temporal conditions, for the spontaneous or provoked RPS is nearly always
,
,
, . ,
/.
...
..
,
7-
,
'I'
"
,
cIc-c)-
.
... . , .
. . .
.
1-
'
"
--I
-t
Fig. 10 Absence of cortical activation during the RPS after lesion of the ascending part of the limbic midbrain circuit. A: wakeful state; rapid cortical activity (SMC), hippocampus ( H ) and level of the MRF. B: slow wave phase of sleep; slow cortical and reticular activity. Hippocampal spikes. C: RPS; discrete activation of the cortex with persistence of a few spindles; absence of hippocampal rhythmic activity. Complete disappearance of nuchal EMG activity. Calibrations: 2 sec, 50 pV. References B . 423-424
418
SLEEP MECHANISMS
followed by a refractory phase of 10-15 min during which stimulations are either ineffective or provoke arousal. The duration of these refractory periods and the triggering mechanism of the RPS for several minutes after brief stimulations, suggest that a neurohumoral agent released at the level of the n. RPC could be responsible for this phase of sleep. Among all the drugs tested, two proved to be of particular interest. Atropine at a dosage of 0.2 mg/kg suppresses the RPS (both EEG and behavioural features) in normal or mesencephalic animals. These results are an argument in favour of the existence of a neurohumoral mechanism of the cholinergic type (antagonistic action of atropine; very tightly contracted pupils, complete relaxation ofthe nictitating membranes, deceleration of the cardiac rhythm during the RPS). Chlorpromazine (5-10 mg/kg), however, does not affect the peripheral phenomena of the RPS but suppresses the rapid cortical activity and rhythmic limbic activity, replacing them by slow sleep activity. Meanwhile, behavioural and cortical arousal reactions remain possible. This drug therefore seems to exert a selective influence on the ascending pathways of projection of the pontolimbic system, while not affecting the n. RPC.
LV. Ontogenesis of the RPS (Jouvet et al. 1961) A study of sleep in kittens has revealed that the RPS is the first form of sleep observed after birth and has confirmed the duality of the systems responsible for the rapid cortical activity in the course of the RPS and arousal. During the first 8 days of life (Fig. 1 IA), behavioural RPS (with complete disappearance of the nuchal EMG and ocular movements) constitute about 40 per cent of the total duration of recording; they are not associated with any change in the cortical tracing, which remains identical regardless of the state of wakefulness (low-voltage
I
Fig. 11A EEG activity in kittens 30 h after birth. I : behavioural wakeful state. 11: RPS; disappearance of EMG activity: 2; ocular movements: 3 ; respiratory irregularity: 4;cortical activity: 1 is the same a s during the wakeful state.
RHOMBENCEPHALIC PHASE OF SLEEP
419
tracing with occasional spindles of 12-15 cjsec). These RPS constitute almost theentirety of behavioural sleep during this period. Towards the end of the 2nd week of life there is, during the RPS, very distinct EEG differentiation, characterized by rapid lowvoltage activity. In this stage the cortical activity of arousal (6-8 cjsec) is not yet very rapid (Fig. 11B). There are brief periods of slow sleep (4-5 cjsec) of a special type because they are accompanied by complete disappearance of the E M G and absence of ocular movements. Towards the 20th day there appear brief periods of rapid cortical activity during wakefulness, while the slow sleep phases are increasingly accompanied by muscular activity in the neck. Only at the end of the 2nd month does the EEG in kittens begin to resemble that in adult cats, with three distinct stages : wakefulness, slow sleep with muscular activity, and the RPS. Thus, as the organization of the complexity of the synapses at the level of the cortical neuronesproceeds, and is parallel with the neuroglial development, the cortical electrical activity is first of all subjected to the periodic influence of the RPC nucleus, which is transmitted to it by the pontolimbic system, and later is subjected only to the influence of the ascending reticular activating system (ARAS); while a phase of slow sleep without muscular activity (found exceptionally in the adult animal) evolves before giving place to the telencephalic phase of sleep (Fig. 12). These findings show that: 1. the “cortical activation” during the RPS precedes, by several days, the cortical activation of wakefulness, via the ARAS; 2. the RPS is dependent upon an ontogenetically earlier system than that responsible for the slow wave phase of sleep. To the extent to which ontogenetic development can be compared with phylogenetic development, therefore, it seems permissible to compare the RPS with “archeo-sleep” (Jouvet et al. 1960).
Fig. I 1 B Same kitten, 21st day. I l l : behavioural wakeful state, slow cortical activity of 7 c/sec. 1V: RPS; rapid low-voltage cortical activity; numerous jerks of the body during the RPS which, from a behavioural point of view, resembles that observed in decorticate animals. Calibrations: 50 pV. References D . 423-424
420
r r SLEEP MECHANISMS
1
111/111/111/111//111/lllllllllllllllllllll!ll\ llllllllllllllllllllllllllllllllllllllllllll lllllllllllllllllllllllllllllllllllllllllll
Fig. 12 Schematic development of the different phases of sleep during the first 2 months in kittens. Ordinate: time in weeks. Abscissa: wake-sleep cycle (calculated on the basis of 5-6 h recording). Vertical hatching: behavioural RPS - dashes indicate absence of rapid cortical activity which appears towards the 12th day. Oblique hatching: behavioural wakefulness - dashes indicate the absence of waking cortical activity, which occurs only at about the 21st day. In white: phase of sleep with slow activity (5 cisec), absence of nuchal EMG, absence of ocular movements. Dotted: phase of slow sleep with nuchal EMG activity.
DISCUSSION
The results discussed above would seem to warrant the conclusion that the RPS expresses the activity of a system, the organization of which can be schematically described as follows (Fig. 13). A triggering “centre”, localized at the level of the n. RPC, is connected with the cortex via the intermediary of distinct ascending pathways in the ARAS, making use of the ventral part of the mesencephalon, and probably of the limbic midbrain circuit. The descending pathways governing the somatovegetative phenomena are still insufficiently known. It seems likely that the inhibitory pontobulbar RF (Magoun and Rhines 1946; Magoun 1950) is involved because disappearance of tonus is seen in the alpha and gamma hypertonia which follows decerebellation and decerebration. The exact limits of this system, the mechanisms of action, and its functions, still raise numerous problems - some of which have been discussed elsewhere (Jouvet 1962a). 1. One fact seems to have been established. The RPS represents the most “profound” state of sleep because the thresholds of arousal by stimulation of the ARAS
RHOMBENCEPHALIC PHASE OF SLEEP
42 1
and sensory stimulation are higher than those during the slow sleep phase (Jouvet el al. 1959; Benoit and Bloch 1960; Hara et al. 1960; Hubel 1960; Huttenlocher 1960; Rossi et al. 1961). Observation of very considerable unit activity within the R F (Huttenlocher 1961) during the RPS, is not readily reconciled with the increase in the threshold of arousal, but it might explain the diminution of reticular evoked responses by a phenomenon of occlusion. The diminution of cochlear responses which is some-
Fig. 13 Structures responsible for the RPS. Very schematic representation. Dotted: nucleus RPC - where are situated the structures responsible for triggering the RPS. Hatched: ARAS. In black: pontolimbic pathways making use of part of the limbic midbrain circuit. Posteriorly the inhibitory bulbar RF.
times observed for several minutesduring the RPS, seems to indicate that certain afferent control mechanisms may also be involved, although not constantly (Jouvet 1962a). 2. The rapid cortical activity of the RPS is similar to that of the state of wakefulness, from which it is not readily distinguishable at a glance. However, it is more “monotonous”, of slightly lower voltage and interspersed with slight spindles of low voltage during ocular movements ; the “experienced eye” can recognize it. The neurone mechanisms on which it is based can be differentiated from those of the arousal reaction because the ontogenetic study of sleep and the selective brain stem sections make it possible to distinguish the responsible structures. The diminution of cortical evoked responses, disappearance of the response by recruitment provoked by stimulation of the diffuse thalamic system (Rossi et ul. 1961), observed in the course of this rapid activity, resemble phenomena observed in the course of intensive arousal reactions ; without more highly perfected analysis, it is impossible formally to differentiate the cortical neuronic mechanisms responsible for the RPS and arousal. A microelectrode analysis would therefore be most valuable. The establishment of the important role of limbic influences in the genesis of such activity is an argument in favour of rhinencephalic control of neocortical activity, whether directly or via the thalamus (Nauta 1956). 3. The diminution of muscular tonus is in very direct relation with the appearance of rapid cortical electrical activity, but the two phenomena are not reciprocally related Rcferenccr P. 423-424
422
SLEEP MECHANISMS
in the way of a hypothetical feedback. Disappearance of EMG activity and cortical activation do not always coincide in time; they may be separated by a few seconds. Chlorpromazine can suppress the rapid cortical activity without causing loss of tonus. In curarized animals, abolition of tonus as such does not cause the appearance of permanent RPS. Observations made in patients with tetanus and curarized for several days have shown that such subjects may be awake and conscious and are capable of describing, afterwards, complex stimuli which were presented to them while they were curarized. An “isolated encephalic” preparation, disconnected for the most part from its muscular afferents, may show behavioural wakefulness (Bremer 1936a). It seems therefore more logical to consider that, under the influence of a triggering cause still unknown, the n. RPC can command both the appearance of rapid cortical activity and a diminution in muscular tonus, although the two phenomena are not necessarily interrelated. 4. The cause of the periodic appearance of the RPS remains obscure. It is based on a mechanism which exists at birth, which resists both hypothermia (for RPS have been observed at rectal temperatures of 30°C in mesencephalic animals; Jouvet 1962a), and small doses of nembutal(l0-I 5 mg/kg). The problem is to establish whether this cause is intrinsically cerebral (localized at the level of the n. RPC) or peripheral. The former hypothesis is plausible when it is admitted that certain periods of rapid cortical activity in retropontine preparations belong to the RPS, while on the other hand somatic, vegetative or humoral afferents could be responsible for triggering the RPS. 5. The physiological importance of the periodic appearance of RPS during sleep is not yet known. The function does not seem to be vital because animals have survived for several weeks when deprived of the RPS after a lesion at the level of the n. RPC. A study of their behaviour, however, has made it possible to establish disturbances in two different spheres, depending on both the ascending and the descending pathways issuing from the n. RPC. Periodical disturbances of behaviour, of the hallucinatory type (interpretation of which is of course very subjective) show the existence of dysfunction in the sphere of perception. It is interesting to compare this state with that observed in man deprived of sleep for protracted periods (Katz and Laudis 1935; Bliss et al. 1959); this state is also accompanied by somaesthetic, auditory and visual hallucinations after an identical lapse of time. On the other hand, numerous facts indicate that, in man, the oneiric occurs during the RPS (Dement and Kleitman 1957; Jouvet 1962a,b). Although the functions of the dream remain obscure, their intervention in psychic life is obvious (Freud 1926; Ey 1952). The probable role of the limbic system in some phenomena of memory (Grastyan et al. 1959; Adey et al. 1960) and in the determinism of rapid cortical activity during the RPS, indicates that, periodically during sleep, certain memory processes undergo integration at “a higher level”. The constancy of the vegetative changes which accompany the RPS (even when concomitant cerebral phenomena are suppressed, as in mesencephalic animals) also indicates that certain peripheral aspects of the RPS are within the realm of homoe-
RHOMBENCEPHALIC PHASE OF SLEEP
423
ostatic regulations of the “milieu inrkrieur”. The progressive tachycardia in animals with a lesion of the n. RPC expresses this disorder in the vegetative system (Jouvet 1962a). It is possible that the determination of a specific neurohumoral agent of the RPS will enable us to understand the intimate nature of the mechanisms relating the oneiric phenomena to homoeostatic regulations and thus participating in the integrative unity of the organism. Figs. 3,4,6, 10 and 13 from Jouvet 1962a are reproduced by courtesy of the Editor of Archives italiennes de biologie.
REFERENCES ADEY,W. R., WALTER, D. 0. and HENDRIX, C. E. Hippocampal slow waves, distribution and phase relations in the course of approach learning. A.M.A. Arch. Neurol., 1960,3: 74-90. ADRIAN,E. D. and MATTHEWS, B. H. C. The interpretation of potential waves in the cortex. J. Physiol. (Lond.) ,1934,81: 440471. BARD,P. and MACHT,M. B. The behaviour of chronically decerebrate cats. In G. E. W. WOLSTENHOLME and M. O’CONNOR (Editors), Neurological basis of behaviour. A Ciba Foundation Symposium. Churchill, London, 1958: 55-71. BATINI,C., PALFSTINI, M., RON, G. F. and ZANCHETTI, A. EEG activation patterns in the midpontine pretrigeminal cat following sensory deafferentation. Arch. iful. Biol., 1959a, 97: 26-32. BATINI,C., MAGNI,F., PALESTINI, M., Rossi, G. F. and ZANCHETTI, A. Neural mechanisms underlying the enduring EEG and behavioral activation in the midpontine pretrigeminal cat. Arch. ital. Biol., 1959b, 97: 13-25. BATINI,C., MORUZZI, G., PALESTINI, M., ROW, G. F. and ZANCHETTI A. Effects of complete pontine transections on the sleep-wakefulness rhythm: the midpontine pretrigeminal preparation. Arch. ital. Biol., 1959c, 97: 1-12. BENOIT,0. et BLOCH,V. Seuil d’excitabilite reticulaire et sommeil profond chez le chat. J . Physiol. (Paris), 1960,52: 17-18. BLISS,E. L., CLARK, L. D. and WEST,C. D. Studies of sleep deprivation -relationship to schizophrenia. A . M.A. Arch. Neurol. Psychiat., 1959,81: 348-359. BREMER, F. Activite dlectrique du cortex cerebral dans les etats de sommeil et de veille chez le chat. C.R. SOC.Biol. (Paris), 1936a, 122: 464-467. BREMER, F. Action de differents narcotiques sur les activites electriques spontanees et reflexes du cortex citrebral. C.R. SOC.Biol. (Paris), 1936b, 121: 861-866. DEMENT, W. The occurrence of low voltage, fast electroencephalogram patterns during behavioural sleep in the cat. Electroenceph. d i n . Neurophysiol., 1958, 10: 291-296. DEMENT, W. and KLEITMAN, N. Cyclic variations of EEG during sleep and their relation to eye movements, body motility and dreaming. Electroenceph. clin. Neurophysiol., 1957,9: 673-690. DERBYSHIRE, A. J., REMPEL, B., FORBES, A. and LAMBERT, E. F. The effects of anesthetics on action potentials in the cerebral cortex of the cat. Amer. J. Physiol., 1936, 116: 577-596. EY, H. Le r&ve“fait primordial” de la psychopathologie. Dans H. EY(Redacteur), Etudes psychiatriques. Dksclee de Brouwer, Paris, 1952,296 p. FREUD, S. La science des rtves (traduction franCaise), 7e Cd., P.U.F., Paris, 1926,212 p. GRASTYAN, E., LISSAK,K., MADARASZ, I. and DONHOFFER, H. Hippocampal electrical activity during the development of conditioned reflexes. Electroenceph. din. Neurophysiol., 1959, I1 : 409-430. HARA,T., FAVALE, E., Rossr, G. F. et SACCO, G. Ricerche sull’attivita elettrica cerebrale durante il sonno nel gatto. Riv. Neurol., 1960,30: 448-460. HUBEL, D. H. Electrocorticograms in cats during natural sleep. Arch. ital. Biol., 1960,98: 171-181. HUTTENLOCHER, P. R. Effects of state of arousal on click responses in the mesencephalic reticular formation. Electroenceph. clin. Neurophysiol., 1960, 12: 819-827. HUTTENLOCHER, P. R. Evoked and spontaneous activity in single units of medial brain stem during natural sleep and waking. J. Neurophysiol., 1961,24: 451468. JOUVET, D., VALATX, J. L. et JOUVET,M. Etude polygraphique du sommeil du chaton. C.R. Sor. Biol. (Paris), 1961,155: 1660-1664. JOUVET, M. Telencephalic and rhombencephalic sleep in the cat. In G. E. W. WOLSTENHOLME and
424
SLEEP MECHANISMS
M. O’CONNOK (Editors), The nature of sleep. A Ciba Foundation Symposium. Churchill, London, 1961: 188-208. JOUVET,M. Recherches sur les structures nerveuses et les mecanismes responsables des differentes phases du sommeil physiologique. Arch. ifaf.Biol., 1962a, 100: 125-206. JOUVET,M. Sur l’existence d’un systeme hypnique ponto-limbique; ses rapports avec I’activite onirique. Dans Physiologie de I’hippocampe. Colloque C .N.R.S. Paris, 1962b: 297-330. JOUVET,M. et MICHEL,F. Recherches sur l’activite electrique cerebrale au cows du sommeil. C.R. Soc. Biol. (Paris), 1958, 152: 1 1 67-1 170. J. et MICHEL,F. Etude des mecanisnies du sommeil physiologique. t y o n JOUVET,M., DECHAUME, d d . , 1960,204: 479-521. JOUVET,M., MICHEL,F. et COURJON, J. Sur un stade d’activite electrique cerkbrale rapide au cows du sommeil physiologique. C.R. Soc. Biol. (Paris), 1959, 153: 1024-1028. KATZ,S. E. and LANDIS,C. Psychologic and physiologic phenomena during a prolonged vigil. Arch. Neurol. Psychiat. (Chicago), 1935,34: 307-317. MAGOUN,H. W. Caudal and cephalic influences of the brain stem reticular formation. Physiol. Rev., 1950,30: 459474. MAGOUN,H. W. and RHINES,R. An inhibitory mechanism in the bulbar reticular formation. J. Neurophysiol., 1946,9: 165-171. MORUZZI,G . and MAGOUN,H. W. Brain stem reticular formation and activation of the EEG. Electroenceph. clin. Neurophysiol., 1949, 1 : 455-473. NAUTA,W. J. H. An experimental study of the fornix system in the rat. J. romp. Neurol., 1956, 104: 247-273. NAUTA,W. J. H. Hippocampal projections and related neural pathways to the midbrain in the cat. Brain, 1958, 81: 319-340. H. G. J. M. Some ascending pathways in the brain stem reticular NAUTA,W. J. H. and KUYPERS, R. S. KNIGHTON,W. C. NOSHAY and R. T. COSTELLO formation. In H. H. JASPER,L. D. PROCTOR, (Editors), Reticular formation of the brain. Henry Ford Hospital Symposium. Little, Brown and Co., Boston, 1957: 3-30. E., HARA,T., GIUSSANI, A. and SACCO,G. Researches o n the nervous mechaRoss], G. F., FAVALE, nisms underlying deep sleep in the cat. Arch. ital. Biol., 1961,99: 270-292. WIKLER,A. Pharmacologic dissociation of behaviour and EEG “sleep patterns” in dogs: morphine, n-allylmorphine and atropine. Proc. Soc. exp. Biol. ( N . Y . ) , 1952, 79: 261-265.
A Study of the So-called “Paradoxical Phase” of Sleep in Cats K.
LISSAK, G. KARMOS AND E. GRASTYAN
Physiological Institute, University of PPcs, PPcs (Hungary)
According to observations made on humans, physiological sleep is characterized by periods of low voltage fast activity (paradoxical phase: PP) which appears on the EEG recording without any sign of behavioural arousal (Aserinsky and Kleitman 1955; Dement and Kleitman 1957). Similar observations have been made in animal experiments (Rimbaud et al. 1955; Dement 1958; Jouvet and Michel 1959; Grastyan 1959; Hubel 1960; Grastyin and Karmos 1961). Several authors have also established that electrical stimulation of various brain
425
PARADOXICAL PHASE OF SLEEP
structures may artificially elicit (Grastyan 1959; Jouvet and Michel 1959; Rossi 1961) as well as abolish the PP (Grastybn 1959). The present report will deal briefly with the problem of muscle tone as well as with observations made by stimulating the mesencephalic reticular formation and the hippocampus in connection with the so-called paradoxical phase of sleep. The investigations were made with permanently implanted electrodes in cats during behavioural sleep. Decrease or disappearance of muscle tone is one of the most important somatic manifestations of the PP. However, our more recent observations do not entirely
BHIP BMC -
** 0
-
w h , + -#p~
I
I
I
JAc
MY
*
:.*-
-I
loop”
1 sec
Fig. 14 Three different phases of behavioural sleep in the cat. First group: the phase characterized by spindles. Note the presence of persistent tone in the neck muscles. Second group: the same phase as before but the tone of the neck muscles begins to disappear. Third group: the so-called “paradoxical phase”. A continuous theta rhythm in the hippocampus, the activity of neocortex begins to get desynchronized. Muscle tone has disappeared. Abbreviations: B Hip : left hippocampus; BMC: left motor cortex; JAC: right acoustic cortex; MY: myogram of neck muscles.
5 50pV
5 0 pV
50pV 50yV Fig. 15 Stimulation of the mesencephalic reticular formation in deep behavioural sleep elicited the peculiar electrical pattern of the “paradoxical phase”. Continuous theta rhythm in both hippocampi, desynchronization in motor and sensory cortices. No behavioural arousal. Abbreviations see Fig. 1. Sens C: sensory cortex. References II.428
426
SLEEP MECHANISMS
agree with former data gathered by ourselves and by others: namely, in the majority of cases it has now been found that the decrease of muscle tone precedes the characteristic electrical manifestations of the brain (Fig. 14). We shall return to this later. Stimulation of the mesencephalic reticular formation during a stage of sleep characterized by spindles and slow waves in the EEG can provoke the peculiar signs of the PP. The activity ofthe neocortex becomes desynchronized and a continuous theta rhythm appears in the hippocampus without any somatic sign of behavioural arousal (Fig. IS). This continuous hippocampal theta rhythm is also one of the most conspicuous electrical manifestations of the spontaneously developing PP (Jouvet and Michel 1959; Grastyan 1959). If the mesencephalic reticular formation is stimulated with voltages higher than are necessary for the elicitation of the PP, behavioural
F Ret 2.3V, 150cyclesl sec, 31-17 sec Fig. 16 Stimulation of the mesencephalic reticular formation in deep sleep with a higher voltage elicited a real behavioural awakening. Desynchronization in all the structures recorded. Note absence of theta waves in the hippocampus. Abbreviations: H : hippocampus; M C : motor cortex; A C : acoustic cortex.
AC
d &
oc
0
C
I loo$ X
CI 1 o o p v
F Ret 2.6V,150cycles lsec ,0.5 rn sec Fig. 17 Stimulation of the mesencephalic reticular formation during the "paradoxical phase" enhanced the electrical signs of this phase. The hippocampal theta rhythm became somewhat more frequent and regular. No behavioural arousal resulted. Abbreviations: J H i p m, B Hip m: right and left monopolar hippocampal recordings; OC: occipital cortex.
427
PARADOXICAL PHASE OF SLEEP
arousal will follow (Fig. 16). It is worth considering, however, that in the latter case no theta rhythm appears in the hippocampus but desynchronization. If, however, stimulation is carried out during the PP an accentuation of the theta activity can be observed in the hippocampus but without somatic manifestations (Fig. 17, 18). HIP
-
rloo~v
4 hoopv RC
I
I
F Ret 2 V , 150cycles Isec,l m sec
1 sec
Fig. 18 Stimulation of the mesencephalic reticular formation during the “paradoxical phase” first enhanced the theta rhythm and then abruptly ,desynchronizedhippocampal activity. The arrow marks the time when behavioural arousal resulted.
Stimulation of the hippocampus in the presence of slow waves has, in most cases, no consequences whatsoever (Lissitk et al. 1957). If stimulation is applied during the PP, sleep spindles and slow waves appear immediately in the neocortex and persist long after the cessation of stimulation (Fig. 19). Muscle tone is not influenced by stimulation alone. In some cases it reappears slowly after stirnulation, sometimes only with behavioural arousal. On the basis of the facts reported here the following interpretation of the PP is proposed. The observation that a PP can be provoked by electrical stimulation of the reticular formation suggests that an increased function of this structure may be I
I I I JAC - MY
vI
‘ -
-
J HIP 7V, 70cycles /sec,O.5 m 5ec
-
- J loop”
1 set
Fie. 19 Stimulation of the hippocampus during a spontaneous “paradoxical phase” resulted in the reappearance of sleep spindles; at the same time hippocampal theta waves disappeared. No change in muscle tone. Abbreviations: ESE: left sensory cortex; EF. RET. : left reticular formation; EMC: left motor cortex; J A C : right acoustic cortex. Reference3 p . 428
428
SLhEP MECHANISMS
responsible, also under physiological conditions, for inducing this phenomenon. This increased function can be considered as a release function of the mesencephalic reticular formation. This assumption is supported by the observation of Batini et a/. (1958) according to which a synchronizing structure is located in the pons, which is assumed to act by influencing rostra1 portions of the activating system. According to the observations of Jouvet and Michel (1 959) sleep spindles appear in the pons synchronously with the disappearance of muscle tone. Should this synchronized activity reflect a decreased pontine function, it might result in a functional release of the mesencephalic reticular formation. On the other hand, however, depressed function of the pons may be caused - a t least partly - by lack of an afferent inflow from the muscles. This fits in well with our observation that decrease of muscle tone occurs before the characteristic electrical changes in the brain. The validity of the assumption that the PP is a released function of the activating systems (prepontine structures) is further supported by the observation that stimulation of the hippocampus restores the synchronized pattern of sleep without increasing muscle tone. REFERENCES N. Two types of ocular motility occurring in sleep. J. uppl. Physiol., ASERINSKY, E. and KLEITMAN, 1955,8: 1-10.
BATINI,C., MORUZZI,G., PALESTINI, M., ROSSI,G. F. and ZANCHETTI, A. Persistent patterns of wakefulness in the pretrigeminal midpontine preparation. Science, 1958,128: 30-32. DEMENT, W. The occurrence of low voltage, fast electroencephalogram patterns during behavioural sleep in the cat. Electroenceph. clin. Neurophysiol., 1958,lO: 291-296. DEMENT, W. and KLEITMAN, N. Cyclic variations in EEG during sleep and their relation to eye movements, body motility, and dreaming. Electroenceph. din. Neurophysiol., 1957,9: 673-690. GRASTYAN, E. In M. A. B. BRAZIER(Editor), The central nervous system and behavior. Second conference. Josiah Macy Foundation, Madison Printing Cy., Madison, N.J., 1959: 119-205. GRASTYAN, E. and KARMOS, G. A study of a possible “dreaming” mechanism in the cat. Actuphysiol. Acad. Sci. hung., 1961,2O: 41-50. HUBEL,D. H. Electrocorticograms in cats during natural sleep. Arch. iral. Biol., 1960, 98: 171-181. JOUVET,M. et MICHEL,F. Correlations electromyographiques du sommeil chez le chat decortique et mesencbphalique chronique. C.R. SOC.Biol. (Paris), 1959, 153: 422425. LISSAK,K . , GRASTYAN, E., CSANAKY, A., KEKESI,F. and VEREBY, GY.A study of hippocampal function in the waking and sleeping animal with chronically implanted electrodes. Acta physiol. pharmacol. need., 1957,6: 451-459. RIMRAUD, L., PASSOUANT, P. et CADILHAC, J. Participation de I’hippocampe a la regulation des etats de veille et de sommeil. Rev. neurol., 1955, 93: 303-308. Rossr, G . F., FAVALE, E., HARA,T., GIUSSANI, A., and SACCO,G. Researches on the nervous mechanisms underlying deep sleep in the cat. Arch. ital. Biol.,1961,9Y: 270-292.
EEG Synchronization Induced by Peripheral Nerve Stimulation * 0. POMPEIANO
Istituto di Fisiologia rkell’Universita rli Pisa e Centro di Neurofisiologia del C.N.R., Sezione di Pisa, Pisa (Italia)
The experiments I am going to report were made in collaboration with Dr. J. E. Swett. They show that under appropriate conditions EEG synchronization can be actively induced, in normal cats, by stimulation of a well-defined group of afferent fibres. The experiments were performed on freely moving intact cats, in which 6-10 cortical, 1-2 EMG (placed in the neck extensor muscles) recording electrodes, and 1-3 nerve stimulating electrodes had been chronically implanted. After a two-day recovery period the animals were periodically subjected to recording sessions over a period of 1-3 weeks. The unrestrained animals were placed in a large glass-walled box, which permitted continuous observation of their reactions to the stimuli. The superficial radial, musculo-cutaneous, sciatic, femoral, gastrocnemius, hamstring and saphenous nerves were tested. Provided that cortical activity had a tendency to be slightly synchronized, it was possible to produce rhythmic high voltage, low frequency oscillations over wide regions of the cortex with trains lasting 2-4 sec of low frequency (3-8/sec), low intensity rectangular pulses (0.1-0.5 msec) (see Fig. 20 A ) . The stimulus intensity was increased to reach the values giving a repeatable, mild EEG arousal reaction. This is defined as the threshold for arousal (Fig. 20 B). Stimulus intensities between 0.4-0.9 times this value gave, almost constantly, marked generalized EEG synchronization, which was sometimes accompanied by closure of the eyes at the onset of the stimulus and by decrease of the EMG activity of the skeletal musculature. The waxing and waning of the large cortical waves, so often seen when the EEG is spontaneously synchronized, was often observed during induced synchronization. The phenomenon had a tendency to outlast the duration of the stimulus (Fig. 20 A ) . The frequency of the synchronous waves followed that of the electrical pulses, provided the rate of stimulation was between 3-8/sec. Single shocks caused short bursts of synchronization (spindle bursts) to appear after a short latency. Increasing the frequency above 8/sec brought about a gradual reduction of the amplitude of the synchronized EEG patterns until a true arousal occurred at frequencies higher than 12-18/sec. The optimum stimulus parameters for obtaining EEG synchronization were unable to reproduce this effect on an EEG background of low voltage, fast activity, either ‘This investigation was supported by PHS research grant B-2990 from the National Institute Neurological Diseases and Blindness, N.I.H., Public Health Service, U.S.A.
Of
430
SLEEP MECHANISMS
C
rnm Hg
4 SEC
I
400 yV
Fig. 20 EEG and blood pressure changes induced by low frequency stimulation of a cutaneous nerve with increasing stimulus intensity. Intact unanaesthetized cat. Experiment made 36 h after the implantation of the electrodes and the cannulation of the left femoral artery. Stimulation of the left superficial radial nerve with rectangular pulses a t 5/sec and 0.5 msec pulse duration. Bipolar records. 1 : stimulus marker; 2: left fronto-frontal; 3 : left parieto-temporal; 4: left temporo-occipital; 5 : right fronto-frontal; 6 : right parieto-temporal; 7 : right temporo-occipital; 8 : left neck EMG; 9: blood pressure recorded from the left femoral artery. A : stimulation with 0.30 V causes generalized EEG synchronization outlasting the stimulus. Note waxing and waning of the response and the large amConrinucd on p o w 4 3 l
ACTIVELY INDUCED EEG SYNCHRONIZATION
43 1
during behavioural arousal or during deep sleep. However, the desynchronized EEG patterns in these two extreme conditions were different in that the primary evoked potential was much larger during deep sleep than during arousal. No change was noted in the heart rate and in the arterial blood pressure (recorded electromanometrically from a chronically implanted femoral cannula), when EEG synchronization was induced by peripheral nerve stimulation (Fig. 20 A ) . A slight decrease in the blood pressure (from 6-12 mm Hg) did occur when the threshold for arousal was reached (Fig. 20 B,C). Although cutaneous nerves, such as the superficial radial nerve, gave the best synchronization, this effect could also be obtained with low frequency stimulation of purely muscular nerves. However EEG synchronization was then obtained with threshold stimuli for fibres in the alpha range, and the effect was noticed only when muscular movements appeared. When muscular contraction was prevented by ligature of the nerve distally to the stimulating electrode, induced EEG synchronization was abolished but the arousal reaction could still be elicited. To obtain such an arousal it was necessary to use stronger stimulus intensity than for cutaneous nerves. These findings demonstrate that the EEG synchronization elicited by stimulation of an intact muscular nerve, is due to the activation of skin and/or joint receptors as a consequence of the muscular movements. Analysis of the peripheral nerve fibres conducting the volleys that are responsible for the EEG changes was carried out by utilizing animals whose thresholds for EEG synchronization and arousal were well substantiated in several trials. As a terminal experiment the cats were subjected to barbiturate anaesthesia, and the neurogram of the superficial radial nerve was recorded from the posterior division of the brachial plexus (Fig. 21). The neurogram of the hamstring nerve was recorded from L7-S1 dorsal roots. EEG synchronization elicited by low frequency stimulation of the superficial radial nerve was obtained when the strength of current was liminal for most of the group I1 cutaneous afferent fibres. The same group of fibres was responsible for the arousal reaction when the frequency of stimulation was increased above 12-18/sec. Group I11 fibres were responsible )n all cases for the arousal reaction elicited by low frequency stimulation of a cutaneous nerve (Fig. 21). Group I and I1 ~~~
~~~
~~
-.
Confrnued from FIE 20.
plitude of the primary evoked potentials recorded from the right postcentral gyrus. There is a slight tendency for the synchronization to follow the rate of stimulation. No change in blood pressure is observed. B: stimulation with 0.37 V causes EEG synchronization in the parietal, temporal and occipital leads, followed by EEG arousal which begins before the end of the stimulus. The rhythm of synchronization follows the rate of stimulation. Arousal is greater in the frontal leads. There is a weak inhibition of the EMG activity during induced EEG synchronization. In B the temporary decrease of EMG activity is followed by a typical recruitment of motor unit activity. During arousal there is a slight drop in arterial pressure. C: stimulation with 0.50 V causes a strong EEG arousal with increase of EMG activity. There is also a marked fall of arterial pressure and a slight increase of the heart rate. In this preparation the threshold for the primary evoked potentials was 0.17 V, the threshold for EEG synchronization 0.19 V, and that for arousal 0.37 V. (From 0. Pompeiano and J. E. Swett, Arch. i f d . Biol., 1962, 100: 343-380.)
432
SLEEP MECHANISMS
muscular afferent fibres produced no noticeable effect on the ECoG, but as for the cutaneous nerves, group 111 fibres proved to be responsible for the arousal elicited by low frequency stimulation of the hamstring nerve. When high frequencies of stimulation were used the threshold for the arousal decreased slightly. In these exper-
1111111111-v Fig. 21 Compound action potentials recorded from a bundle of the brachial plexus initiated by electrical stimulation of the superficial radial nerve. Monopolar recording. Stimulation of the left superficial radial nerve with rectangular pulses 0.5 msec in duration, at increasing strengths of current. Five sweeps for each record. Nerve length 1 1 8 mm. Voltage calibration 100 pV. In a the stimulus intensity is threshold for the primary evoked potentials, in h threshold for EEG synchronization, in dthreshold for the arousal. Note the selective activation of group I1 fibres in a and b and the progressive development of group 111 wave from c to$ The current used i n f , which produces maximal activation of group 111 fibres, was never applied to the normal intact animal.
iments it was not possible to exclude that some of the higher threshold group I1 muscular afferent fibres contributed to the arousal response. The stimulus intensities used for producing EEG arousal by stimulating both the cutaneous and muscular afferent fibres, in our chronic preparations, were never sufficient to activate group IV (C group) fibres. DISCUSSION H.W. MAGOUN: I should like to make reference to several recent studies in which induced cessation of behavior, resembling Pavlovian internal inhibition, has been found to be associated with synchronization of the electrical activity of the brain, like that in drowsiness or the induction of sleep. In their work on the neuroendocrine control of sexual activity, Sawyer and Kawakami (1959) have observed that during a period of several minutes following coitus, the female rabbit displays a languid relaxation and torpor, leading to sleep, during which its EEG displays cortical spindle bursts and a prodigious hypersynchronization in limbic structures. Similar patterns of electrical activity, associated with drowsiness
DISCUSSION
433
and sleep, have been induced by Sterman and Clemente (1961) by direct bilateral stimulation of the preoptic region. In another category of innate behavior, Anokhin told us yesterday of the pronounced synchrony of cortical and hypothalamic electrical activity induced in hungry dogs by tubing of food directly into their stomachs and injecting glucose into their blood streams. Innate behavior is customarily described as being terminated by satiety and, in showing how this may be associated with the induction of synchrony and spindle bursts in the EEG, these findings suggest that satiety may be the expression of a type of internal inhibition which, after consummation is achieved, serves importantly in bringing innate behavior to an end. Striking effects upon behavior have also been obtained by Buchwald and his associates (1961) by driving the EEG synchronizing thalamo-cortical system through stimulating the caudate nucleus, in the cat. Low-frequency caudate shocks which triggered spindle bursting in the EEG were found to block lever-pressing activity and greatly prolong reaction time to visual signals. In recent electrophysiological studies of higher nervous activity, increased synchronization and spindle bursting in the EEG has been encountered in most of the categories of evoked cessation of behavior which Pavlov described as internal inhibition. Such alterations have been reported during differentiation of conditional reflexes (Kogan 1960), during the inhibition of delay and conditioned inhibition (Cluck and Rowland 1959), and during extinction (Roitbak 1960; Hernandez-Peon 1960). Many contributions thus point to the existence of a non-specific thalamo-cortical system, the lowfrequency excitation of which evokes large slow waves, as well as recruiting responses and spindle bursts in the EEG. It seems likely from the work which Pompeiano and Swett have just reported that they may be driving this mechanism by low-frequency peripheral afferent stimulation. Although sometimes dissociated from behavior, these patterns of electrical activity of the brain are characteristically associated with internal inhibition, behavioral drowsiness and sleep, probably more by gating neuronal discharge into ineffective firing patterns than by true pre- or post-synaptic inhibition of neuronal activity. Differentiable components must exist in this thalamo-cortical system, for its subdivisions are capable of being driven more or less independently from different parts of the brain. Involvement of this system from bulbo-pontile sources, as the Pisa group and Dell et al. (1961) have proposed, may be designed to effect a general reduction of visceral processes. Its excitation from hypothalamic and limbic structures appears to serve a negative feedback control of pituitary secretion and to provide a means of terminating innate behavior by satiety. When activated from the basal ganglia and cortex, this system appears to manage all the Pavlovian categories of internal inhibition of higher nervous activity, including that of sleep itself. If these inferences are correct, it is now possible to identify a thalamo-cortical mechanism for internal inhibition, capable of modifying activity of the brain partially or globally, so as to bring behavior to a halt. The actions of this mechanism are opposite to those of the reticular activating system, with capacities for focal or generalized excitation of the brain. The Sherringtonian principles of reciprocal innervation appear relevant to the manner in which these antagonistic mechanisms determine the alternating patterns of brain activity manifest behaviorally as wakefulness and sleep. BUCHWALD, N. A., WYERS,E. J., LAUPRECHT, C. W. and HEUSER, G. The “caudate-spindle” IV. A behavioral index of caudate-induced inhibition. Electroenceph. d i n . Neurophysiol., 1961, 13: 531-537. DELL,P., BONVALLET, M. and HUGELIN, A. Mechanism of reticular deactivation. In G . E. W. WOLSTENHOLME and M. O’CONNOR (Editors), The nature af sleep. A Ciba Foundation Symposium. Churchill, London, 1961: 86-103. CLUCK,H. and ROWLAND, V. Defensive conditioning of electrographic arousal with delayed and differentiated auditory stimuli. Electroenceph. din. Neurophysiol., 1959, I 1 : 485-496. HERNANDEZ-PEON, R. Neurophysiological correlates of habituation and other manifestations of plastic inhibition. Electroenceph. clin. Neurophysiol., 1960, Suppl. 13: 101-1 12. KOOAN, A. B. The manifestations of processes of higher nervous activity in the electrical potentials of the cortex during free behaviour of animals. Electroenceph. clin. Neurophysiol., 1960, Suppl. 13 : 51-61. MORUZZI, G. Synchronizing influences of the brain-stem and the inhibitory mechanisms underlying the production of sleep by sensory stimulation. Electroenceph. din. Neurophysiol., 1960, Suppl. 13: 231-253. POMPEIANO, 0. and SWETT,J. E. EEG synchronization produced by peripheral nerve stimulation. Experientia (Basel), 1961, 17: 323.
434
SLEEP MECHANISMS
ROITBAK, A. I . Electrical phenomena in the cerebral cortex during extinction of orientation andconditioned reflexes. Electroenceph. d i n . Neurophysiol., 1960, Suppl. 13: 91-98. SAWYER, C. H. and KAWAKAMI, M. Characteristic of behavioral and electro-encephdlographic afterreactions to copulation and vaginal stimulation in the female rabbit. Endocrinology, 1959, 65 : 622-610. STERMAN, M. B. and CLEMENTE, C. D. Cortical recruitment and behavioral sleep induced by basal forebrain stimulation. Fed. Proc., 1961, 20: 334.
R. NAQUET: Although, as Jouvet, and after him Rossi have shown, the synchronization of the electrocorticographic rhythms cannot be compared to sleep, but is solely one of the necessary preliminary stages of the occurrence of sleep in the normal cat, nevertheless, it would seem of interest to present a report today on the results of some experiments which we carried out last year, and which were concerned with the phenomena of hemispheric synchronization, and particularly with the precipitation of spindles in a single hemisphere. We have actually found it possible to obtain variations of the rhythms of a single hemisphere by two different methods used at more or less the same time in two laboratories, one in Paris, the other in Marseilles. Other authors before us have mentioned the possibility of provoking spindles in a single hemisphere, without change in the activity of the other hemisphere, with the aid of small lesions of the reticular formations of the brain stem (Rothballer 1956; Cordeau and Mancia 1959). Once the asymmetry had appeared it persisted, and as long as the experiment lasted the resting rhythms never returned to their initial form, except during stimulation of the reticular formation or during nociceptive stimulation, which reestablished transiently a symmetrical waking record. 1. In order to obtain a reversible cerebral synchronization we have applied, in collaboration with Denavit, ALbe-Fessard and Le Beau, the refrigeration-tube technique recently used by Dondey, Albe-Fessard and Le Beau (1961). We have decided to explore the reticular formation and particularly those regions where the coagulation provokes unilateral or bilateral spindles. We have used two sorts of gas: butane and propane. With butane, which cools the end of the tube to approximately 1O"C, we have never been able to provoke the appearance of spindles in the corresponding hemisphere behind the A3 plane of Horsley-
Fig. 22 Successive monopolar recordings of two homologous cortical regions (right and left suprasylvian cortex). Tube inserted, stereotactically, on the right, obliquely from the posterior to the anterior in the Horsley-Clarke plane: 3 Lat., 3 M : -1. Note the appearance of the spindles in the cortex 90 sec after the application of the cold.
43 5
DISCUSSION
Clarke. The only effect that could sometimes be seen was an acceleration of the activity observed previously (accentuation of the rapid frequencies); we have, on the other hand, been able to provoke the appearance of unilateral spindles by gently cooling certain parts of the reticular formation in the vicinity of anterior plane 4. With propane, the temperatures at the tip of the tube can be decreased to - 10 to - 15°C. Between the anterior planes 2 and 3, as soon as a temperature of - 10°C is reached, after a generalized waking reaction, in the hemisphere subjected to the cooling we observed the appearance of fusiform bursts. These spindles appear 1-2 min after the required temperature is reached, and disappear when the cooling is discontinued. When the temperature has risen again to 0°C or higher, this spindle phase is followed by an intense, bilateral waking reaction (Fig. 22). The spindles are as a rule observed only in
-
T - ~ H
~
73 +
+
14
,,i+wwu*Ir.L-LY.ICLrl-.I--L1*.*.YI
later
Injection flaxedil
~
q
w
A +
p~
! ~ ~ ; & , .
.
~:v+&+w+d:q;,l~%~,k-
~
,
~
~
,
~
w
+
~
\
4
~
~
-n--+A-.-
~
:
~
:w 2 ,* ~ m ~**& ~ ~ ~~* r n~! \ : h; w ?~ *
~
Fig. 23 Non-anaesthetized cat, immobilized with flaxedil. Monopolar recording referred to an indifferent electrode. 1, 2, 3, 4: recording with the aid of silver electrodes applied directly to the cortex. 11, 12, 13, 14: recording through screws inserted into the bone. T = 0: recording at the beginning of the experiment; note the perfect symmetry between the two hemispheres. T = 2 h later: after a rapid intravenous injection of flaxedil, a generalized slowing of the record appeared. The spindles nevertheless had a larger amplitude on the exposed side. The two hemispheres responded very well to pinching.
~
~
~
436
SLEEP MECHANISMS
one hemisphere, and when occasionally they encroach on the contralateral hemisphere, they always remain far greater on the cooled side. These preliminary results will be assessed and discussed later on. 2. With the aid of another technique, in collaboration with Tiberin and Rhodes (1961), we were also able to produce asymmetries between the two hemispheres. In one hemisphere, the activity remained rapid, with small amplitudes, whereas spindles appeared in the other; during nociceptive stimulation, however, the activity in both hemispheres became the same (rapid, with small amplitudes). This asymmetry may occur after modification of the functional condition of a hemicortex of a nonanaesthetized cat which has been immobilized with flaxedil. One of the hemispheres was recorded without craniotomy, by means of screw electrodes implanted into the bone; the other was recorded with the aid of silver electrodes applied directly to the exposed cortex without protection with paraffin, after wide craniotomy and opening of the dura mater. In the early stage of the experiment (apart from the variations of the amplitude due to the different recording techniques) the activity of the two hemispheres was the same. After 3 or 4 h of recording under these conditions, the exposed hemisphere began to develop spindles which were precisely the same as those observed after lesions of the reticular formation (Fig. 23 and 24). Under the experimental conditions used, the precipitation of these spindles must be ascribed to the most superficial structures of the neo-cortex. Actually the only difference between the two hemispheres is that in the exposed cortex there was the possibility of micro-traumata, or of minimal circulatory disturbances. On the basis of these results we may conclude that the changes of the functional condition of one hemicortex or of the two cortices may make possible the precipitation of a cortical synchronization analogous to that provoked with reticulo-thalamic stimulations or lesions. These data confirm the hypotheses recently advanced concerning the importance of the cortex in the formation of the spindles (Jouvet and Michel 1959; Naquet and Fernandez-Guardiola 1960; Naquet et al. 1960) and demonstrate that, at any rate under certain circumstances, the cortex plays an active and not merely a passive role in causing cortical synchronization. CORDEAU, J. P. and MANCIA, M. Evidence for the existence of an electroencephalographic synchronization mechanism originating in the lower brain stem. Electroenceph. din. Neurophysiol., 1959, I 1 : 551-564. M., ALBE-FESSARD, D. et LE BEAU,J. Refroidissement temporake et localise de structures DONDEY, cerebrales profondes. Resultats physiologiques preliminaires. Rev. neurol., 1961, 105: 186-1 87. JOUVET,M. et MICHEL,F. Correlations electroenctphalographiques du sommeil chez le chat decortiqueet mesencephalique chronique. C.R. SOC.Biol. (Paris), 1959,153: 422-425. A. Effets de diffkrents types d’anoxie sur l’activite electrograNAQUET, R. et FERNANDEZ-GUARDIOLA, phique cerebrale spontanee et evoquee chez le chat. C.R. SOC.Biol. (Paris), 1960, J . Physiol. 52: 885902. NAQUET, R., LANOIR,J. et RHODES,J. A propos du “waxing and waning” de la reponse par recrutement. C.R. Soc. Biol. (Paris), 1960,154: 2351. ROTHBALLER, A. B. Studies on the adrenaline-sensitive component of the reticular activating system. Electroenceph. d i n . Neurophysiol., 1956, 8 : 603-621. TIBERIN, P., RHODES, J. et NAQUET,R. Les hemisynchronisations corticales a point de depart cortical. C.R. SOC.Bid. (Paris), 1961,155: 1346-1349. R. JUNG:
I should like to present some data illustrating the different patterns of cortical neurones during sleep as seen in multiple recordings. The experiments, performed by Drs. Creutzfeldt, Lehmann and Koukkou in my laboratory, showed different neuronal sleep patterns in different cortical regions. Simultaneous recordings from several neurones were obtained mostly from the motor and visual cortex and from some regions between these two areas. The experiments on the motor cortex are published in the Ciba sleep symposium (Creutzfeldt and Jung 1961). Since then, Dr. Lehmann and Dr. Koukkou have collected a larger material of about 250 neurones in enciphale isolk cats during sleep. Special attention was paid to the spindle stage and to the relation between EEG spindles and neuronal discharge. A preliminary communication of neurones of the visual cortex was given a t the Rome EEG Congress (Lehmann and Koukkou 1961). With the onset of sleep the pattern of cortical neuronal discharge is changed: the more or less
431
DISCUSSION
+
d
w
w
W
w
%
,
m
Fig. 24 Continuation of the previous figure. T = 4 h later: long bursts of spindles appeared spontaneously in the exposed hemisphere, while on the other side at this time there was only a moderate slowing of the rapid activity. Spontaneously, and during nociceptive stimulation (pinching) rapid activity appeared in the two hemispheres. This activity was symmetrical.
continuous random waking activity is altered into periodic grouped discharges, separated by longer pauses. These periods contain either short high frequency bursts of 3-7 spikes (usually found in posterior cortical areas) or low frequency trains of 5-20 spikes (most common in anterior areas). In the motor cortex the periodic neuronal trains occur mostly during EEG spindles, simultaneously in several neurones. These periodic trains with various spike intervals of 15-200 msec may contain doublets and, very exceptionally, high frequency bursts,
438
SLEEP MECHANISMS
In contrast, the neurones of the visual cortex mostly show periods of high frequency bursts with spike intervals of 2-5 msec during the spindle stage of sleep. Usually these bursts precede the spindle waves and are not synchronized in several neurones, although they appear together for longer time periods. During the spindle the neuronal burst pattern is dissolved into random single spike discharges (with lower average frequency). All types of visual neurones defined by their responses to light (A, B, C, D, E) change their pattern in a similar way before and during spindles. The reciprocal type of neuronal activity in the B- and D-system characteristic for light and dark stimuli is less marked during sleep but may also appear in alternating bursts. Following arousal the neurones are either inhibited or activated and return to their random single spike pattern. The average frequency of discharge is usually higher for random spikes during arousal than for similar random patterns during sleep or for burst activity during sleep but neuronal inhibition during arousal may also occur (Creutzfeldt and Jung 1961). The different patterns of neuronal activity during sleep in both the motor and visual cortex show the complex nature of neuronal sleep mechanisms. Before these can be understood the neurophysiological mechanisms of spindle waves, the significance of burst discharges and their relation to long lasting inhibition of neurones has to be elucidated further. Because generalizations from the findings in enciphale is016 cats may be hazardous, more experiments are needed in chronic unrestrained cats. 0. and JUNG,R. Neuronal discharge in the cat’s motor cortex during sleep and CREUTZFELDT, and M. O’CONNOR(Editors), The nature of sleep. A Ciba arousal. In G. E. W. WOLSTENHOLME Foundation Symposium. Churchill, London, 1961 : 131-170. LEHMANN, D. and KOUKKOU, M. Neuronal discharge patterns and spontaneous EEG spindles in the visual cortex of enciphale isoli cats. Vth int. Congr. Electroenceph. d i n . Neurophysiol., Rome, 1961. Excerpta med. ( A m s t . ) , Int. Congr. Series, 1961, No. 37.
G . MORUZZLto M . Jouvet: First of all 1 want to congratulate Dr. Jouvet for the important results he has reported. I was particularly struck by the possibility of abolishing selectively, with appropriate midbrain lesions, either the EEG arousal or the episodes of sleep with low voltage fast activity. These observations represent, in my opinion, a promising approach to one of the major tasks which confront the neurophysiologist, who is obviously eager to know more about the differences between true EEG activation and desynchronized sleep. I have already discussed this point a t the Ciba Symposium on sleep (Moruzzi 1961) and I prefer to concentrate on the brain stem mechanisms of the phenomenon. The experimental evidence on sleep without EEG synchronization, a problem to which Dr. Jouvet has contributed so much, is already so impressive that it could hardly be reviewed in a short discussion. I shall not emphasize the factual data, which are well known and have been confirmed repeatedly, but shall confine my remarks to their interpretation. I suggest that a distinction should be made between (a) those data which lead directly, and I would say unavoidably, to a given interpretation; (b) those which support a hypothesis without proving it, and finally (c) those whose significance is not yet clear. (a) There is one group of observations which, in my opinion, can no longer be a subject of controversy as far as their interpretation is concerned. The experiments of Jouvet (1962), Hubel (1960), Rossi et a / . (1961) and Candia et al. (1962), have shown quite conclusively, for instance, that sleep is deeper during the episodes of low voltage fast activity. Another important observation is that the episodes characterized by the disappearance of the spindle trains never occur in the acute cerveau isold or during barbital anaesthesia at the surgical level. In these conditions the EEG patterns are consistently similar to those of the synchronized stage of sleep. Nobody would be prepared to say that the episodes of sleep without synchronization occurring in the normal cat are related to a depression of the cerebrum which is never attained during the coma produced by a midbrain transection or during barbital anaesthesia. Undoubtedly the difference between any phase of physiological sleep and coma or barbital anaesthesia is not purely a quantitative one. However since all the neural phenomena leading to the appearance of spindle trains are usually attributed to the decrease (physiological sleep) or to the withdrawal (coma) of an ascending reticular influence, it is well to point out that the functional deafferentation of a normal brain can hardly become a t any moment more extensive than the anatomical deafferentation of the cerveau isolt!. Hence I do not see how we
DISCUSSION
439
can escape from the conclusion that the episodes of sleep without EEG synchronization are actively produced by neural structures, which are prevented from acting on the cerebrum by brain stem transection or by the action of barbital. This conclusion does not imply, necessarily, that synchronized sleep is also produced actively, but we must at least concede that the deafferentation theory is unable to explain sleep without EEG synchronization. (b) There is a second group of phenomena which strongly suggest the explanatory hypotheses put forward by Jouvet, but which (in my opinion) do not provide crucial evidence in their favour. There is little doubt, e.g., that the postural manifestations of deep sleep described by Jouvet and his colleagues in the normal or in the erebellectomized cat resemble the cataplexic episodes found by Bard and Macht (1958) in the decerebrate cat. Jouvet’s hypothesis of a rhombencephalic phase of sleep is undoubtedly useful, since it may lead to new experiments. It would be interesting to see, e.g., whether the fall in blood pressure observed by Rossi and his colleagues (Candia et al., 1962) during the episodes of sleep without synchronization is also present during the cataplexic episodes of Bard and Macht. There is little doubt, nevertheless, that the conclusion that the episodes of sleep without EEG synchronization are due to ascending volleys arising from structures lying caudal to the midbrain does not necessarily imply that the primum mobile of the phenomenon is to be found in the rhombencephalon. The cataplexic episodes of Bard and Macht (1958) are undoubtedly rhombencephalic in origin. However, to extend this conclusion to the episodes of deep sleep of the normal cat is a stimulating but as yet not a decisively proven hypothesis. Some questions may be relevant in this connection: we may ask why does sleep never start, in the normal cat, with a typical episode of deep, desynchronized sleep? This should be the case at least in some instances, if the deep phase of sleep really arose in the rhombencephalon as an autochtonous phenomenon. Actually as far as we know the deep phase of sleep is always heralded by a lighter stage of sleep characterized by EEG synchronization. This, in my opinion, would suggest, that the ascending rhombencephalic discharge is related in some way to the neural mechanisms underlying synchronized sleep, or that the pontine nuclei are only one link in the chain of neural and humoral events leading to sleep without synchronization. It might well turn out to be a rewarding task for future experimenters to investigate the relationships between these pontine structures, those responsible for synchronization and the activating reticular system. These interrelations, anyway, are likely to be destroyed or severely altered by any midbrain transection. (c) Those phenomena whose significance is yet hard to assess I would put in the third category. Jouvet and his colleagues have reported that trains of spindles may be led from the pons during the episodes of low voltage fast sleep. However, in my opinion, it has not as yet been proved that this electrical activity is the cause of sleep without synchronization. Might it not be a contingent phenomenon related to the ocular movements that occur during the same period of time? We know that lateral nystagmus is due to the activity of the pontine nuclei. I think that it would be interesting to record, simultaneously, the electrical activity of the pons and the EMG of the facial muscles and of the extrinsic eye muscles.
BARD,P. and MACHT,M. B. The behaviour of chronically decerebrate cats. In G. E. W. WOLSTENHOLME and M. O’CONNOR (Editors), Neurological basis of behaviour. A Ciba Foundation Symposium. Churchill, London, 1958: 55-75. CANDIA, O., FAVALE, E., GIUSSANI, A. and ROW, G. F. Blood pressure during natural sleep and during sleep induced by electrical stimulation of the brain stem reticular formation. Arch. ital. Biol., 1962,100: 216-233. HUBEL,D. M. Electrocorticograms in cats during natural sleep. Arch. ital. Biol., 1960,98: 171-181. JOUVET, M. Recherches sur les structures nerveuses et les mecanismes responsables des diffirentes phases du sommeil physiologique. Arch. ital. Biol., 1962, 100: 125-206 MORUZZI, G. General discussion in G. E. W. WOLSTENHOLME and M. O’CONNOR (Editors). The nature of sleep. A Ciba Foundation Symposium. Churchill, London, 1961: 392-393. ROSSI,G. F., FAVALE, E., HARA,T., GIUSSANI, A. and SACCO, G. Researches on the nervous mechanisms underlying deep sleep in the cat. Arch. ital. Biol., 1961, 99: 270-292.
G.F. ROW to M . Jouvet:
I havetwobriefremarkswhicharerelated to Dr. Jouvet’s presentation. 1 quite agree with Dr. Jouvet’s
SLEEP MECHANISMS
conclusions that the structures responsible for deep sleep, or a t least the most important structures, are located at the level of the pons. However, I d o not think that they should be identified exclusively with the n. reticularis pontis caudalis. According to the appearances in the slides projected by Dr. Jouvet himself and according to my own personal experience, I think that the most important structures in the pons having a sleep inducing function are at the level of the rostral part of then. reticularis pontis caudalis and at the level of the caudal part of the n. reticularis pontis oralis. The object of my second remark is to draw your attention to the surprising and, 1 think, interesting conclusion which emerges on considering together the results of Dr. Jouvet and those obtained three years ago by Batini et al. (1959a,b) in the Physiological Institute of Pisa. Let us briefly summarize these findings. 1. Transection of the middle pons (midpontine pretrigeminal preparation) is followed by persistent EEG desynchronization and by behavioural manifestations of wakefulness. 2. The most rostral brain stem transection which prevents the appearance of the episodes of muscular relaxation which are characteristic of deep sleep is located at the same niidpontine level. 3. A brain stem transection between the midbrain and the pons (cerveau isold) or at the level of the upper pons (rostropontine pretrigeminal preparation - Batini et al. 1959b) js followed by the appearance of enduring behavioural sleep; persistent EEG synchronized rhythms characterize these preparations. 4. Episodes of muscular relaxation similar to those observed during deep sleep still occur in the cerveau isole‘ (Jouvet 1961). 5. Extensive destruction of the pontine reticular formation abolishes the EEG and EMG nianifestations of deep sleep (Jouvet 1961). From these data it appears, as stated by Dr. Jouvet, that the influence responsible for the EEG and EMG signs of deep sleep might originate from the pons, mainly from its rostral part. On the other hand, as previously suggested by Batini e t a / . (1959 a,b), at the same level there are also located structures which appear to be very important for the maintenance of the wakeful state. Anatomical considerations indicate that both these structures should be identified with neurones of the reticular formation. The final conclusion reached is therefore that neurones of decisive importance for both arousal and deep sleep are intermingled in the pontine reticular regions. BATINI, c.,MAGNI,F., PALESTINI, M., RON, G . F. and ZANCHETTI, A. Neural mechanisms underlying the enduring EEG and behavioral activation of the midpontine pretrigeminal cat. Arch. irul. Biol., 1959a, Y7: 13-25. BATINI, c.,MORUZZI, G., PALESTINI, M., Ross[, G . F. and ZANCHETTI, A. Effects of complete pontine transections on the sleep-wakefulness rhythm: the midpontine pretrigeminal preparation. Arch. ital. Biol., 1959b, 97: 1-12. JOUVET,M. Telencephalic and rhombencephalic sleep in the cat. In G. E. W. WOLSTENHOLME and M. O’CONNOR (Editors), The nature o f sleep. A Ciba Foundution Symposium. Churchill, London, 1961 : 188-208. A. ZANCHETTI :
There has been somedivergence of opinion in recent years on the interpretation of some findings presented by our group in Pisa, and on the significance of the prolonged EEG activation patterns of the midpontine preparation. We suggested (Batini et a / . 1958, 1959a,b) that these resulted from the suppression of a synchronizing influence originating from the caudal brain stem ; an interpretation which has received some support from subsequent work by Dr. Dell’s group (Bonvallet and Bloch 1961; Dell et a / . 1961), and from the recent finding that EEG synchronization can be induced by direct electrical stimulation of the lower brain stem tegnientuni (Magnes et al. 1961; Favale et al. 1961). On the other hand, Dr. Jouvet (1961) has advanced the hypothesis that at least part of the prolonged EEG activation patterns of the midpontine preparation represents a sign of so-called “activated” sleep rather than wakefulness. Although our results were not based upon EEG data alone, but were supported by behavioural observations on purposeful ocular movements, there is no doubt that Jouvet’s hypothesis should be examined more critically. For this reason, I have been quite interested in the report by Dr. Jouvet that the EEG manifestations of “activated” sleep can be abolished by interruption of the mammillary peduncular system. If this is confirmed, then an investigation of the EEG patterns of a cat with a lesion in the mammillary peduncular system and
DISCUSSION
44 1
with subsequent midpontine transection might well provide a means of definitely settling the discrepancies as interpreted by Jouvet and our group. Dr. Jouvet’s report has also aroused my interest from another point of view. As you may recall from my presentation two days ago, my coworkers and I have recently been working on the influence exerted by the mammillary peduncular system on the diencephalic behaviour of sham rage. We have seen that sham rage is not prevented by section of the mammillary peduncle, which means that this system has not an essential activating influence upon one of the regions on which it projects. If, somewhat arbitrarily, we were to extend these conclusions so as to include also those other targets of the niammillary peduncle which in some way wouId be responsible for the EEG patterns of “activated” sleep, it might then be suggested, tentatively, that the functions of the mammillary peduncular system might be inhibitory in nature. This suggestion would fit in well with the explanation as advanced by Rossi et al. (1961) and by Moruzzi (1961), that the activated patterns of deep sleep were due to inhibition of the synchronizing mechanisms. Finally, I should like to support Prof. Moruzzi’s suggestion that the spindle-like bursts of slow waves recorded by Jouvet from the pontine reticular formation during “activated” sleep are likely to represent events related to the ocular jerks observed during this phase of sleep. Indeed, Mikiten et al. (1961) have recently reported similar waves in the lateral geniculate nucleus during “activated” sleep, and it is reasonable that they represent on-off responses due to ocular movements.
M., ROSSI,G. F. and ZANCHETTI, A. Neural mechanisms underlying BATINI,C., MAGNI,F., PALESTINI, the enduring EEG and behavioral activation in the midpontine pretrigemind cat. Arch. ital. Biol., 1959a, 97: 13-25. BATINI,C., MORLJZZI,G., PALESTINI, M., RON, G . F. and ZANCHETTI,A. Persistent patterns of wakefulness in the pretrigeminal midpontine preparation. Science, 1958, 128: 30-32. G., PALESTINI, M., ROSSI,G. F. and ZANCHETTI, A. Effects of complete pontine BATINI,C., MORLJZZI, transections on the sleep-wakefulness rhythm: the midpontine pretrigeminal preparation. Arch. ital. Biol., 1959b, 97: 1-12. BONVALLET, M. and BLOCH,V. Bulbar control of cortical arousal. Science, 1961, 133: 1133-1134. DELL,P., BONVALLET, M . and HUGELIN, A. Mechanisms of reticular deactivation. In G. E. W. WOLSTENHOLME and M. O’CONNOR[Editors), The nature of sleep. A Ciba Foundation Symposium. Churchill, London, 96 : 86-102. FAVALE, E., LOEB,C., ROW, G. F. and SACCO,G. EEG synchronization and behavioral signs of sleep following low frequency stimulation of the brain stem reticular formation. Arch. ital. Biol., 1961,Y9: 1-22. JOUVET,M. Telencephalic and rhonibencephalic sleep in the cat. In G . E. W. WOLSTENHOLME and M. O’CONNOR(Editors), The nature of sleep. A Ciba Foundation Symposium. Churchill, London, 1961: 188-206. G . and POMPEIANO, 0.Electroencephalogram synchronizing structures in the MAGNES,J., MORLJZZI, and M. O’CONNOR[Editors), The nature of sleep. lower brain stem. In G. E. W. WOLSTENHOLME A Ciba Foundation Symposium. Churchill, London, 1961: 57-78. C. D. EEG desynchronization during behavioral sleep MIKITEN,T. M., NIEBYL,P. H. and HENDLEY, associated with spike discharges from the thalamus of the cat. Fed. Proc., 1961, 20: 327. MORUZZI,G. Recenti sviluppi nello studio elettrofisiologico del sonno. Progr. med. (Napoli), 1960,16: 781-786. Ross~,G. F., FAVALE, E., HARA,T., GILJSSANI, A. and SACCO,G. Researches on the nervous mechanisms underlying deep sleep in the cat. Arch. ital. Biol., 1961, YY: 270-292.
R. GRANITto 0. Pompeiano: Dr. Granit expressed his intcrest in Pompeiano’s discovery of sleep-inducing skin receptors in the medium range of conduction velocities from slow-frequency stimulation of the radial nerve. Surely such receptors are likely to operate spontaneously, and so we encounter the problem of whether the central regions which shift the EEG-pattern from one type to another may not be foci from which particular receptor-systems are controlled centrifugally or directly.
442
SLEEP MECHANISMS
F. BKEMEK to 0 .Fotttpeiano: The experiment which Dr. Pompeiano has just described to us is indeed a surprising one. There is no doubt that when we stroke a cat that knows us well, this may help to render the animal drowsy. But in the rabbit, a brief touching of the hair, which must be subjectively pleasant to the animal, always has a waking effect, which is most clearly demonstrated by the regularization of the potentials of the hippocampus. Can Dr. Pornpeiano give us an explanation of the effect of sleep seen in his experiments with brief low-frequency electrical stimulation of the cutaneous nerves?
A . ARDUINI to 0. Ponrpeiano: I should like to point out that the fact that sleep follows low frequency stimulation of sensory nerves does not mean that the sleep is actively induced. We know, from our experiences, that sleep can be produced during steady retinal illumination. This has been interpreted, in a paper in collaboration with Dr. Hirao, as passive sleep (reduction of the tonic retinal bombardment upon the reticular formation). Together with Dr. Pinneo we have demonstrated that the output of the retina is reduced during light adaptation. We would not be very surprised if during low frequency stimulation of sensory nerves the input to the reticular system would also be found to be reduced, through some kind of inhibitory mechanism.
ARDuiNi, A. and HIRAO,T. On the mechanism of the EEG sleep patterns elicited by acute visual deafferentation. Arch. ital. Biol., 1959, 97: 140-155.
G . MORUZZIto K . Lissak The problem raised by Professor Lissak of the temporal relations between peripheral phenomena and electrophysiological manifestations occurring in the central nervous system during the episodes of desynchronized sleep is of the greatest importance. The relationships between clonic movements and cervical atonia on one side and pontine electrical activity should be investigated.
M. JOUVET's replies: To G . Moriizzi I . Concerning the problem of the identification of the deep sleep phase with desynchronized EEG activity and with the periodic hypotonia observed in mesencephalic cats (which is similar to that observed by Bard and Mach), we could observe every intermediate degree of activity in all our preparations. In the intact animal one observes periodic hypotonia, eye movements, and fast cortical activity. In the decorticate cat there are only periodic hypotonia and eye movements. The pontine cat shows the muscular phenomenon only. The periodicity and the time ratio (about 20-25% of the recording time) are similar in all preparations. 2. The neural structures acting upon the n. reticularis pontis caudalis are not well known. There is not as yet sufficient evidence to decide whether it is an intrinsiccerebral mechanism or an externally triggered mechanism; however it appears that the disappearance of EMG activity is not, per se, responsible for deep sleep. Patients may be conscious under curare. 3. There are two types of activity in the pons during the rhombencephalic phase of sleep. (a) The first one is an S-lO/sec, irregular, burst of spindle-like activity. It almost always coincides with eye movements and twitching of the vibrissae. It is recorded bipolarly from a region near the Vlth or Vllth nerve and is probably associated with discharges in these structures. (b) The second one is a rhythmic 4-5/sec, regular activity, similar to the hippocampal theta activity. It is recorded on a small field in the middle of the pontine reticular formation and is often masked by some fast activity. Such an activity is almost continuous during deep sleep and does not seem to be related to eye movements.
DISCUSSION
443
To G.F. Rossi Our most caudal brain stem transection, in which there is no periodic hypotonia is a retropontine transection. This leaves about 314 of the n. reticularis pontis caudalis in front of the transection. This corresponds with our coagulation experiments in which most of the n. reticularis pontis caudalis was destroyed. I certainly think that the neural structures triggering off deep sleep must be relatively diffuse and may also be found in theposterior part of then. reticularis pontis oralis. But they certainly lie in the pontine reticular formation just caudal to the activating system.
To A . Zanchetti Your suggestion is very elegant. My suggestion would be to destroy the region of the interpeduncular nucleus in the midpontine preparation. This would yield the best results, as it is the origin of the ascending corticopetal pathways which may be more diffuse rostrally. 0. POMPEIANO’s replies: To A . Arduini The possibility that the EEG synchronization induced by low frequency stimulation of group 11 cutaneous afferent fibres is due to inhibition of the ascending activating system is a hypothesis that should be tested, but there is at least another possibility which should be examined. When recording the EEG responses of the cortex to single shock stimulation of group 11 cutaneous fibres, we found that the latency of the first synchronous wave was constant and of the same order as “associative” responses investigated by other authors (Buser 1957). This suggests that the same mechanism is probably driven by the afferent volleys. BUSER,P. Activites de projection et d’association du neocortex cerebral des mammiferes. J. Physiol. (Paris), 1957,49: 589-656.
To F. Bremer It is quite likely, as you have suggested, that the sleep induced by stroking the cat’s fur may involve mechanisms similar to those responsible for EEG synchronization elicited by low frequency stimulation of group 11 cutaneous afferent volleys. K. LlSSAK’s replies:
To G. Moruzzi Regarding Professor Moruzzi‘s remark may I point out that the main reason for modifying our earlier explanation was the fact that the decrease in muscle tone, in the majority of cases, preceded the characteristic electrical manifestations of the brain, namely, the theta rhythm in the hippocampus and the desynchronized activity of the neocortex. I perfectly agree with professor Moruzzi that the next most important step to be made would be a very precise and simultaneous analysis of motor phenomena and pontine electrical activity. This might shed some more light on the mode of triggering of the paradoxical phase.
GENERAL DISCUSSION ON INHIBITION Cliuirmun: F. BREMER
Introduction F. BREMER Luboruiory of Genera/ Pathology, Universit.v of Brussels (Belgium)
Moruzzi and 1 thought that it might be of interest to devote the general discussion of this afternoon to an inventory and analysis of central inhibitory processes localized in supraspinal structures. In this exploration, application of the conception of presynaptic inhibition will probably prove fruitful, but it will not be easy. It is a fact, at any rate, that we possess as yet only very meagre data regarding the role of direct inhibition in the cerebral cortex, inhibition which has been identified as postsynaptic on the basis of the demonstration of modifications in the membrane of the inhibited neurones. There are obvious technical reasons for this lack of data, not the least being the relatively small size of the nerve cells of the cortex, which renders it very difficult to insert a micro-electrode. In publications such as those of Branch and Martin (1958), in which such alterations in nerve membranes were studied it is reported not to have been found regularly. This has given these authors the idea of a presynaptic determinism of the observed suppression of spontaneous unitary discharge. In other instances, the membrane hyperpolarization obviously followed an initial highfrequency discharge of the nerve cell and therefore could not be accounted for by a genuine inhibition. There still remains the possibility of occlusion by competitive convergence on interneurones. Even though some reserve is indicated concerning information obtained i n a purely pharmacological manner (the example of the evoked cortical potentials and their modification by y-aminobutyric acid appears to me to be significant in this respect!) the strychnine test may nevertheless be very useful for determining, with some probability, the inhibitory mechanism involved. The studies of Desmedt and Monaco ( 1 960) on inhibition of the cochlear response by stimulation of the olivocochlear bundle indicate this possibility very clearly. BRANCH,C. L. and MARTIN, A. R . Inhibition of Betz cell activity by thalaniic and cortical stimulation. J. Neurophysiol., 1958,21: 381-390.
DESMEDT, J. E. et MONACO, P. Suppression par la strychnine de I’effet inhibiteur centrifuge exerce par le faisceau olivo-cochleaire. Arch. int. Phurnmcodyn., 1960, 129: 244-248.
INTRODUCTORY REPORTS
Cerebral Inhibitory Phenomena R.JUNG Department of’Clinicat Neurop/~ysiology,University of Fre~burg13r.(Germany)
In introducing this discussion on inhibition 1 should like to point out the main features of presynaptic inhibition as outlined in the account given by Sir John Eccles. These characteristics seem to me to be of general importance and specially noteworthy when considering inhibitory phenomena at higher levels in the CNS. Although presynaptic inhibition has been recognized only at the synaptic relays of primary afferent fibres, it seems likely that it also occurs at other synapses, for example in the cerebral cortex and at subcortical levels. The following three points appear to be clearly established with regard to presynaptic inhibition: 1. long duration of inhibition of 100-300 msec, with a maximum at 200 msec; 2. association with slow potentials having a comparable duration and a dipolar configuration; 3. participation of interneurones along the pathway responsible for inhibition by presynaptic depolarization. Two further important points: 4. the nature of the postulated axon-axonal synapses; 5. their neuropharmacological differences from postsynaptic inhibition have still to be investigated more precisely morphologically and experimentally. If we look for parallels at supraspinal levels, we find similar features in the cortex : long inhibitions, slow waves and interneuronal participation. This opens up new prospects for the interpretation and investigation of cortical potentials. I can give three examples of similar inhibitions that have been found in our laboratory in investigations on single cortical neurones. The three conditions mentioned for presynaptic inhibition are fulfilled in these three situations: ( a ) the long neuronal inhibition following epicortical electrical stimuli (Creutzfeldt et al. 1956); (b) callosal inhibition following stimulation of the contralateral cortex in neurones of the sensorimotor cortex described in the same paper; (c) inhibition in the motor cortex following caudate stimuli (Spehlmann et al. 1961). All these long inhibitions may be preceded by brief excitation and the first two seem to be unaffected by the application of strychnine. In the strychninized cortex long inhibitory pauses of single neurones follow strychnine spikes as we have shown here in Pisa six years ago (Baumgartner and Jung 1955). Other characteristics of presynaptic inhibition have still to be investigated by testing the depolarization of presynaptic terminals, which seems to be invariably associated with this inhibition. A further possible example of presynaptic inhibition may be the constant inhibitory mechanisms which are active in the visual system from the retina to the cortex Rrfercnrsl
n. 447
446
INHIBITION
and which form the basis of contrast phenomena: reciprocal inhibition of antagonistic neurones and lateral inhibition of synergistic neurones. These two inhibitions have similar time courses, they cause slow potentials and involve interneurones. They may also reveal spinal presynaptic inhibitory functions as a contrast mechanism besides being a simple negative feed-back. Presynaptic inhibition may also contribute substantially to the production of electrical potentials in the cerebral cortex. This is suggested by the marked electrotonic spread which most cortical potentials show in the fibres of the white matter. We all agree about the dipolar nature of cortical potentials and their reversal in the middle layers. However, it is not generally accepted that these dipoles are complex balance phenomena, resulting from two oppositely directed potential fields as I had proposed in 1953 and 1958. Most people accept apical dendrites as one source of brain waves and it seems likely that the depolarization of presynaptic inhibition is another of the slow cortical potentials. The analogy with spinal cord potentials is of course not complete; for example the potential fields may be in the opposite direction from the surface positive phase, observed in the spinal cord. However, this surface negativity on the cortex would be expected if presynaptic inhibition acted near the synaptic endings which are mostly situated i n the upper layers of the cortex. Of course it would be equally unjustifiable to correlate all slow waves with presynaptic inhibition as it would be to call them all dendritic potentials. But for reasons of cortical geometry these two structures, presynaptic fibres and dendrites, are the most likely sources of brain potentials, and we may raise this point again in the discussions on d.c. potentials. The slow components of spinal presynaptic inhibition may revive the hypothesis discussed by Tonnies and myself (l950), that cortical potentials are related to inhibition and may be restraining, anti-convulsive mechanisms. We then called the cortical response “Bremswelle” or braking wave and showed that a diminution or reversal of surface negative components occurs after repeated stimulation and thus announces the seizure discharge. The fatigability of these waves, however, following repeated stimulation seems to be at variance with the findings in the spinal cord and may point to a more complex nature of these cortical potentials, as do also the negative d.c. shifts during stimulation. But 1 refrain from discussing the functional meaning of brain potentials and presynaptic inhibition which would lead us into far too wide a field of speculation. All these points may be superficial analogies which need more experimental investigation. But it seems promising to use some of the tools which have been developed in connection with investigations on presynaptic inhibition i n the spinal cord: electrical testing, strychnine insensitivity. and possibly picrotoxin depression and barbiturate enhancement. Before ascending to the height of the brain in our discussion 1 propose at first to return to the spinal level and to some applications in the frog’s spinal cord where we find such large dorsal root potentials even after antidromic stimulation of the motor neurones.
INTRODUCTORY REPORTS
447
REFERENCES G. und JUNG, R. Hemmungsphanomene an einzelnen corticalen Neuronen und BAUMGARTNER, ihre Bedeutung fur die Bremsung convulsiver Entladungen. Arch. Sci. biol. (Bologna), 1955, 39: 474-486. O., BAUMGARTNER, G. und SCHOEN, L. Reaktionen einzelner Neurone des sensoCREUTZFELDT, motorkchen Cortex nach elektrischen Reizen: 1 . Hemmung und Erregung nach direkten und kontralateralen Einzelreizen. Arch. Psychiat. Nervenkr., 1956, 194: 597-619. JUNG, R. und TONNIES,J. F. Hirnelektrische Untersuchungen uber Entstehung und Erhaltung von Krampfentladungen: Die Vorgange am Reizort und die Bremsfahigkeit des Gehirns. Arch. Psychiat. Nervenkr., 1950,185: 701-735. SPEHLMANN, R., CREUTZFELDT, 0. und JUNG, R. Neuronale Hemmung im motorischen Cortex nach electrischer Reizung des Caudatums. Arch. Psychiat. Nervenkr., 1961, 201: 332-354.
A. FESSARD
Following the initial work of Baumgarten and Jung (1952), many authors have described the interruption of the unitary autorhythmic activity which follows brief excitations of neurons in the cerebral cortex. It was found that in intra-cellular derivations, this interruption is accompanied by an internally negative wave of long duration (more than 100 or even 200 msec) (Albe-Fessard and Buser 1953, 1955). This phenomenon was rediscovered by several authors in various preparations. I t was observed by Albe-Fessard and her collaborators (1960) in different parts of the cerebral cortex, and also in the hippocampus, the red nucleus, and the centrum medianum of the thalamus. Is it the sign of a true inhibitory process, and, if so, what is the nature of this process? One must at any rate exclude the idea of its being necessarily a post-reactional process, for it is sometimes observed without any previous excitation (Albe-Fessard 1960). Two hypotheses can be considered. It might represent an hyperpolarization due to IPSP’s elicited,for instance, by a neuronal circuit analogous to that of thc Renshaw cells, an assumption made by Phillips (1959) for the pyramidal neurons of the cortex. The somewhat surprisingly great amplitude of the wave might be explain. ed by an abnormal state of cell depolarization due to the implantation of the electrode into the cell. We might also have to deal with a simple return towards the normal polarization level, caused by a momentary arrest of the continuous barrage that maintains the autorhythmicity. This arrest may be conceived as being due to a post-synaptic inhibitory process taking place in a pool of interneurons, but the new explanation is now offered that there may possibly be a pre-synaptic inhibition. If it is there the intense depolarization assumed to be present in the afferent terminals might be the cause of the electrical negativity as it is recorded intracellularly in the same way as it would appear outside the cell: thus there would not actually be a real increase of the cellular polarization. In order to choose between these hypotheses, several obvious criteria could be used. They have not been applied so far, because pre-synaptic inhibition is still for most of us quite a recent discovery which we greet with the feeling that it can open new avenues for explanation. In favor of this mechanism, we can only mention now References II. 44&
448
IN HIBITION
the significant fact that the slow polarization wave accompanying the pause is enhanced rather than depressed by local strychnine poisoning. REFERENCES ALBE-FESSARD, D. Sur l’origine des ondes lentes observees en derivation intracellulaire dans diverses structures cerebrales. C . R. Soc. Biol. (Piiris), 1960, 154: 11. ALBE-FESSARD, D. et BUSER,P. Explorations de certaines activites du cortex moteur du chat par niicroelectrodes: derivations endo-soniatiques. J . Physiol. (Paris), 1953, 45: 14. ALBE-FESSARD, D. et BUSER,P. Activites intracellulaires recueillies dans le cortex sigiiio’ide du chat : participation des neurones pyramidaux au “potentiel evoque” somesthesique. J. Physiol. (Paris) 1955,47: 67-69. BAUMGARTEN, R. and JUNG,R. Microelectrode studies on the visual cortex. Rev. neurol., 1952, 87: 151-155.
PHILLIPS,C. G . Actions of antidroniic pyramidal volleys on single Betz cells in the cat. Qucirf. J . exp. Physiol., 1959,154: 25.
DISCUSSION J.C. ECCLESto F. Brenier: There is a very remarkable correlation between the electron-niicroscopic observations of Dr. Engstroni and the physiological investigations of Drs. Desmedt and Monaco on the olivo-cochlear endings on the hair cells of the organ of Corti. He has shown that there are two quite distinct types of nerve endings around the bases of the hair cells. First are nerve endings with a very clear cytoplasm that make very intimate contact with the hair cells. Secondly there are endings filled with synaptic vesicles that make intimate contact both with the clear terminals and with the hair cells. In every respect these contacts are typical qynapses, the vesiculated fibers being the presynaptic components. Preliminary degeneration experiments accord with the postulate that these fibres are the synaptic endings of the olivo-cochlear bundle, while the clear fibres are the receptor terminals of the cochlear nerve. E N G S T R ~ H. M , Electron micrographic studies of the receptor cells of the organ of Corti. In G. L. RASMUSSEN and W. WINDLE(Editors), Neural inechanisni~of the auditory and ves/ibulur systems. C . C. Thomas, Springfield, Ill., 1961. DESMEDT, J. E. et MONACO, P. Suppression par la strychnine de I’effet inhibiteur centrifuge exerce par le faisceau olivo-cochleaire. Arch. inf. flmmiacodyn., 1960, 129: 244-248. J.C. ECCLESto A . Fessard: 1 would suggest that two tests could be used to determine whether these prolonged hyperpolarizations are genuine postsynaptic potentials. One test is to see if they are reversed by intracellular injection of chloride ions; the other test would be to see if they are greatly depressed after the intravenous injection of a subconvulsive dose of strychnine. An alternative interpretation could be that the hyperpolarization is due to the removal of a tonic excitatory bombardment. However, 1 would like to add that postsynaptic inhibitory potentials with durations of up to 100 msec can be generated in spinal niotoneurones by single cutaneous volleys. Even longer durations are produced by single ventral root volleys when the Renshaw cell pathway is operating in the presence of anticholinesterases.
C . AJMONE MAKSAN: The phenomenon described and illustrated by Dr. Fessard has also been commonly observed by us. In our experience, it is the characteristic behaviour of “injured” cells in response to various
DISCUSSION
449
types of stimuli. For this reason I am tempted to generalize and suggest that such a particular behaviour could be essentially pathological, actually expressing abnormal conditions of the cell (even when its firing pattern is not typical of injury). K . LISSAK:
Regarding the humoral side of inhibition may I ask Professor Eccles whether he has any experience of differential effects of inhibitory substance (substance I, Florey) and GABA with the use of picrotoxin. Also can he say whether the effect of these inhibitory substances is to act presynaptically or postsynaptically. J. C. ECCLES to K . Lissak:
Drs. Curtis, Phillis and Watkins in the Canberra Laboratory have, as you know, produced very good evidence that GABA and its related substances d o not act on the motoneurone in the same way as the inhibitory transmitter substance. Using a co-axial microelectrode they record from the motoneurone intracellularly with the inner barrel and electrophoretically inject GABA or 8-alanine into its external environment. In contrast to the action of the inhibitory transmitter, the membrane potential is not changed, but there are large decreases in the excitability of the membrane and in the sizes of the testing excitatory and inhibitory postsynaptic potentials. Also the spike potential of the cell is depressed. The diminution of the excitatory postsynaptic potential is much greater than would be expected from the depressed excitability of the postsynaptic membrane, which suggests that GABA may be acting on the excitatory synapses and diminishing their output of transmitter. It is thus even possible that a substance related to GABA could be the transmitter responsible for presynaptic inhibition. CURTIS, D. R., PHILLIS, J. W. and WATKINS,J. C. The depression of spinal neurones by y-amino-nbutyric acid and ,5’-alanine. J . Physiol. (Lond.), 1946, 146: 185-203. H. W. MAGOUN: It has been very exciting for me to have been present at both the conferences at which Eccles has presented, with such articulate clarity and enthusiasm, the two concepts of inhibition with which his name can now be associated. I recall with great pleasure the Ciba Symposium in London, celebrating the Sherrington Centenary, at which he elaborated the concept of inhibition depending upon hyperpolarization of the post-synaptic membrane, induced by a Renshaw cell inhibitory transmitter substance, for which strychnine formed the curare-like blocking agent. It would be of great interest to know whether the inhibition of spinal motor activity induced by cerebellar or bulbo-reticular stimulation depended upon such a mechanism of post-synaptic inhibition. With regard to the criterion of strychnine susceptibility, Terzuolo and Gernandt (1 956) utilized the neurophysiological preparation which could be most saturated with this agent and observed the maintenance of the inhibition of the electrical signs of spinal strychnine tetanus upon stimulation of the cerebellar or bulbo-ventral inhibitory system. Subsequently, Terzuolo (1959) explored the criterion of hyperpolarization of the post-synaptic membrane and, with intracellular recording, observed cerebellar and bulbo-reticular inhibition of spinal anterior horn cell discharge without hyperpolarization of the post-synaptic membrane. These findings indicated that something other than post-synaptic hyperpolarization was responsible for cerebellar and reticule-spinal inhibition, as was proved to be the case also with GABA. Along the lines of Eccles’ proposal, Purpura an Grundfest (1956) have suggested an alternative inactivation of excitatory transmitter substance, I n his work with Brazier and William (Adey et al. 1960), to which reference was made earlier, Adey was too modest to mention that they had also specified a presynaptic inhibition. This mechanism of presynaptic inhibition, which Eccles has elaborated so articulately, would at present seem to account best for the experimental data available and it is hoped that additional evidence may be forthcoming in its support. In the case of cerebellar inhibition, a rebound facilitation or excitation is a characteristic and prominent feature. Does Eccles have any suggestion that might relate rebound to presynaptic inhibition?
INHIBITION
ADEY,W. R. DUNLOP, C. W. WILLIAM, K. F. and BRAZIER, M. A. B. Investigations of the action of thiosemicarbazide on the cerebellar cortex of the cat. In E. ROBERTS (Editor), fnhibitiotz iir the tiervous systeni arid y-arnbiobutyric acid. Perganion Press, Oxford, 1960; 3 17-323. GRUNDFEST, H. Central inhibition and its mechanisms. In E. ROBERTS(Editor), Inhibition in the nervous system and y-aminohtyric acid. Pergamon Press, Oxford, 1960: 47-65. PURPURA, D. P. and GRUNDFEST, H. Nature of dendritic potentials and synaptic mechanisms in cerebral cortex of cat. J . Neurophysiol., 1956, 19: 513-595. TERZUOLO, C. A. Cerebellar inhibitory and excitatory actions upon spinal extensor motorneurons. Arch. ital. Biol., 1959,97: 316-339. TERZUOLO, C. A. and GERNANDT, B. E. Spinal unit activity during synchronization of a convulsivc type. Ameu. J . Physiol., 1956,186: 262-270.
J. ECCLESto H . W . Magoun: I regret that I have nothing to add to our attempts to understand the rebound that follows inhibition. Sherrington’s explanation was that inhibition suppressed a concomitant excitation, which appeared as a rebound response after the inhibitory suppression had ceased. 0. POMPEIANO: Considering the problem of cerebellar inhibition in terms of electrical changes occurring at the level of the extensor motoneurons, one should discuss the possibility, based upon both anatomical and physiological evidence, that stimulation of the vermal cortex of the cerebellar anterior lobe mainly causes activation of the inhibitory portion of the reticular formation and also inhibition of neuronal activity in Deiters’ nucleus. It is well known that this structure exerts a facilitatory influence on the extensor rnotoneurons. These facts are mentioned merely because in the cat vestibulospinalfibers’originatingfrom Deiters’nucleus are said to make monosynapticconnections with the motoneurons (Schimert 1938), while the reticula-spinal fibers seem to make synapses with interneurons exclusively. Therefore in order to analyze cerebellar inhibition one should experimentally exclude either the cerebello-reticular or the cerebello-vestibular path and if possible study the influence of the pure anatomical system which is left on the membrane potentials of the extensor motoneurons.
SCHIMERT, J. S. Die Endigungsweise des Tractus vestibulospinalis. Z . Auat. Entwickl. Gesch., 1938,108: 761-167. G . MORUZZI: I am sure that several among us will be happy if Prof. Eccles would develop his views on the functional significance of presynaptic inhibition. I have two questions in mind. First of all, presynaptic inhibition should be without effect in the absence of synaptic excitation, i.e. when we are dealing with autochthonous activities, to use Sherrington’s terminology. Hence only postsynaptic inhibition may account, e.g., for the vagal inhibition of the heart pacemaker. I would ask Prof. Eccles whether he thinks that some inhibitory mechanisms are specially related to postsynaptic or presynaptic inhibition. 111 the central nervous system it is impossible to say whether a background of “spontaneous” activity is partially or totally autochthonous in nature, or whether it is maintained by a tonic afferent inflow, or by reverberating circuits. It would then be important-and this is my second question-to have some criteria for differentiating between pre- and postsynaptic inhibition in the central nervous system. J.C. ECCLESto G . Moruzzi: When we consider the control of neuronal activity, we have many possibilities. First, it is not excluded that neurones have autochthonous activity as suggested by Prof. Moruzzi. Such activity of course cannot be acted on by presynaptic inhibition, but it is susceptible to suppression by postsynaptic inhibition. Apart from such autochthonous activity, nerve cells are caused to discharge impulses by excitatory synaptic action, and of course this excitation is susceptible to depression by
DISCUSSION
45 1
either presynaptic or postsynaptic inhibition. I would like to emphasize that so far presynaptic inhibition has been demonstrated only with the primary afferent fibres of the spinal cord. With intracellular recording and other less specific methods of testing it has been shown that virtually all medullated afferent fibres are subjected to presynaptic inhibition. All synaptic excitatory actions of these fibres are depressed in this manner. It is of paramount importance to find out whether presynaptic inhibition is exerted elsewhere in the central nervous system and on other axonal endings besides those of primary afferent fibres. I would suggest that this can be tested most simply by excitability measurements on axons close to their synaptic termination. We employ for this purpose a rather coarse glass microelectrode (about I M o h m resistance) and filled with 4 M N a C I . This stimulating electrode is placed extracellularly in the regions of the synaptic terminals of the axons under examination, and the recording electrode would be at some millimetres distance in the tract made by these axons. The site of the stimulating electrode would be chosen so that graded stimulation gave a nicely graded size of conducted spike potential at the recording site. Under such conditions any conditioning stimulus that depolarized axonal terminals would cause a large conducted spike response; hence a presynaptic inhibitory action could be presumed.
F. BREMER to G . Moruzzi: An evolutionary hypothesis could be envisaged in relation to the question raised by our colleague Moruzzi. Central inhibition of the indirect or presynaptic type could have first appeared in vertebrates. It seems, in any case, that this type of inhibition is of the greatest functional importance in the frog. Its strength was shown by the socalled inhibition of Richet, in which all spinal reflexes are abolished by the tonic discharges of descending inhibitory impulses from the brain stem. This inhibition is not abolished by strychnine.
F. BREMER to J.C. Eccles: I believe that it is dangerous to make the convulsive action of a drug depend upon its ability to suppress postsynaptic inhibition. Eccles himself has shown that convulsive agents, like metrazol and picrotoxin, do not have a paralyzing effect upon postsynaptic inhibition at the spinal level. Desmedt and Monaco have made the same observation for inhibition of cochlear responses caused by stimulation of the olivo-cochlear bundle.
DESMEDT, J. E. et MONACOP. Suppression par la strychnine de I’effet inhibiteur centrifuge exercC par le faisceau olivo-cochleaire. Arch. int. Pharmacodyn., 1960, 129: 244-248.
J.C. ECCLESto F. Bretner: I would not go so far as to postulate that there is no postsynaptic inhibition in Amphibia. I would, however, like to suggest that there may be, in comparison with mammals, a transitional situation. Certainly presynaptic depolarization is very large in Amphibia. Moreover, the negative feed-back from motoneuronal discharge occurs, not by the postsynaptic Renshaw inhibitory mechanism of mammals, but by inducing depolarization of the dorsal root fibres with the consequent presynaptic inhibition. However, I think that the very effective action of strychnine in producing convulsions in Amphibia certainly suggests that there are well developed postsynaptic inhibitory actions also.
G. MORUZZI to J.C. Eccles: Presynaptic inhibition is characterized by depolarization of presynaptic fibres, leading to diminution of the EPSP. The following questions are concerned with the physiological requirement of a strict selectivity in several processes of central inhibition: (i) How may the long lasting mediator of presynaptic inhibition selectively affect only a given presynaptic fiber, without influencing the polarization of neighbouring unmyelinated presynaptic endings?
452
IN HIBITION
( i i ) How, and at what distance from the endings, does the hypothetical axon-axonal synapse produce the depolarization of the presynaptic fiber? Some prediction about the localization of the hypothetical axon-axonal synapses might be of great help to those who wish to study the problem of presynaptic inhibition by the technique of electronmicroscopy.
J.C. ECCLESto G. Moriizzi. Although presynaptic inhibition in the mammalian spinal cord is fairly general in its action, and displays less selectivity than do some types of postsynaptic inhibitory action, this need not be a characteristic of its field of action at other sites. It could, in fact, be even more specific than postsynaptic inhibition, because it could depress selectively one type of excitatory synapse on a neurone and not another type, whereas postsynaptic inhibition necessarily operates against all the excitatory synaptic actions on a neurone. Because it is postulated that the depolarization of presynaptic terminals is produced by a chemical transmitting synapse, there is no more difficulty in explaining selectivity of action than with postsynaptic excitation or inhivtion. One can assume that the synapses are of the conventional type with apposition across a 200 A synaptic cleft, and that on diffusion of the transmitter out of this cleft it is so effectively dissipated that it is unable to have an appreciable effect elsewhere. It is postulated that these depolarizing synapses on the presynaptic fibres are located very close to the synaptic terminals. Otherwise their effectiveness would be reduced by losses caused by electrotonic transmission, and this has been shown diagrammatically in papers being published in The Journal of Physiology. Addendum (February 1962): Since the discussion at Pisa, there have been reports by electron-niicroscopists of synapses located on the synaptic knobs, both in the spinal cord (E.G. Gray, Noture (Lond.), 1962, 193: 82. and in the lateral geniculate body (J. Szentagothai, personal communication). These findings correspond precisely to the structural relationship previously postulated on physiological grounds.
R . JUNG: As Prof. Moruzzi casts some doubt upon the differential mechanism of presynaptic inhibition,
I would like to point out that this inhibition acts by a very specific mechanism: by a x o n - u x o d synupses. Although these have not yet been described very often in the mammalian nervous system, they are well known in lower animals. Physiological evidence given by Sir John Eceles has proved their effectiveness at least in the primary afferent fibres of the spinal cord. Such axon-axonal synapses are not at all coarse mechanisms of control. They may be very specific and may work just from one fibre exclusively on a few other fibres, thus producing a selective depression of specific afferents on certain synapses. If you allow me to speculate further at the risk of being heretical, I would even say: there may also be dendrito-axonal synapses causing presynaptic inhibition. With Tonnies I assumed in 1947 that such a recurrent pathway, called “Ruckmeldung”, might be effective in motoneurons. Later this was shown to be the Renshaw mechanism. But 1 still think that, in the frog’s spinalcord, there may be such a dendrito-axonal inhibition from its motoneurons for the following reason. Motoneuron dendrites are very large there and can be traced back to the dorsal roots, where they set up, after antidromic stimulation, rather large dorsal root potentials. A similar mechanism may exist in other parts of the nervous system especially in the brain. JUNG, R. und TONNIES,.I.F. Hirnelektrische Untersuchungen uber Entstehung und Erhaltung von Krampfentladungen : Die Vorgange am Reizort und die Bremsfahigkeit des Gehirns. Arch. Psychiat. i i . Z. Neurol., 1950,185: 701-735.
FINAL DISCUSSION
Introductory Remarks G. MORUZZf
Problems of general interest were debated on several occasions during the discussions of the reports and were the themes of two sessions held during the last days of the Symposium. The Editors have thought that a deviation from the rule of publishing all discussions in their chronological order was justified on this occasion, in an attempt to avoid dispersion of related topics into different parts of the book and to facilitate as far as possible continuity of thought among our readers. DISCUSSION ON Following these considerations, the last part of the GENERAL INHIBITION contains comments and replies which had actually been made during the FINAL DISCUSSION, which will remain therefore concentrated upon the notions of SPECIFICITY and INTEGRATION. The same considerations apply to the presentations of Dr. Zanchetti and of Professors Anokhin and Lissak. They were actually made during the discussion of Professor Anokhin’s report, but they appear so closely related to the questions raised by Professors Brookhart and Grey Walter on the problems of specific and unspecific mechanisms, that their inclusion in the FINALDISCUSSION seems appropriate. The concept of integration had been touched upon several times, as may have been expected in a Symposium devoted to sensory-motor integration, but it has seemed advisable to go deeply into this subject, which is of fundamental importance for the physiology of the central nervous system, and to examine thoroughly the development of Sherrington’s classical notion which has been made possible by modern electrophysiological investigations on brain stem and the cerebrum. Professor Fessard has kindly accepted my request to close the Symposium with a Commentary centered around the notion of integration.
A. ZANCHETTI
Specificity of the Reticular Activating System* The question of the specificity of the reticular activating system, already discussed by Professor Anokhin, seems worthy of some further comment. This mechanism was initially defined, perhaps with a certain naivetk, but undoubtedly with a strong impact on subsequent research, as an unspecifically organized brain energizer, while for the last few years an increasing number of authors have speculated upon its specific participation in several subtle processes and psycho-physiological phenomena. In our opinion there are two aspects of the problem which should be studied further, before we commit ourselves to either of these interpretations, or perhaps to both. The first aspect is that of the specificity of the reticular influence, i.e. the dependency of different neural phenomena upon the activity of the same or of different reticular and extrareticular mechanisms. The second aspect is that of the specificity of the intrinsic reticular organization, in terms of size and convergence of receptive fields, and responsiveness to various afferent stimuli. So far as the specificity of the reticular influence is concerned, a preliminary qualification is made necessary by a number of recent experiments (for detailed references, see Zanchetti, 1962) which show that several different reticular mechanisms contribute to sleep-wakefulness regulation: but our discussion now will refer only to the properties of the classical reticular activating system. We have thought that a study of the ascending mechanisms which maintain the sham rage behavior of the decorticate cat might shed some light also on the question of the specificity of the reticular activating system. Indeed, on the one hand, sham rage behavior has some similarity to wakefulness, or at least to arousal, in so far as both can be considered as being instances of mass behavior, while, on the other hand, the diencephalic regions shown by Bard (1928) to be indispensable for rage behavior appear to receive anatomical connexions from the brain stem (the ascending component of the limbic-midbrain circuit of Nauta, 1958) other than those which are held responsible for maintaining electrocorticographic activation. Our study (Apelbaum, Bizzi, Malliani and Zanchetti, 1961) has taken advantage of the partial segregation of several ascending pathways at the midbrain level, where the classical lemniscal and spinothalamic systems are laterally placed; the reticular pathways to the subthalamus and thalamus pass mediodorsally, while hypothalamopetal pathways run medioventrally (mammillary peduncle system) and through the periaqueductal grey matter (Schutz’s dorsal longitudinal fasciculus). 60 acute de-
* The researches referred to in this article have been sponsored by the Aeronautical Systems Division, AFSC, and the Air Force Office of Scientific Research, OAR, through the European Office, Aerospace Research, United States Air Force, under Contract No. AF 61 (052)-253, and by Consiglio Nazionale delle Ricerche.
SPECIFICITY OF THE
u s
455
corticate cats were subjected to different types of stereotactic electrolytic lesions of the midbrain which aimed at interrupting discretely each of the ascending afferent systems. Sham rage outbursts upon direct stimulation of the hypothalamus provided evidence that mesencephalic injury had not impaired efferent outflow from the hypothalamus. As shown in Fig. 1, neither a lesion of the rostral periaqueductal grey matter (including the ascending fibers of Schiitz’s dorsal longitudinal fasciculus), nor a wide interruption of the pontine component of the mammillary peduncle system, nor a more complete interruption of both the mesencephalic and pontine components of the same system, could prevent either the spontaneous outbursts of rage or those evoked exteroceptively. Nor was this type of diencephalic behavior impaired by
Fig. 1. Midbrain lesions which d o not prevent sham rage manifestations in the decorticate cat. Top row, lefr : anatomical scheme showing the ascending connexions to the hypothalamus (Schiitz’s bundle from the central grey to the posterior and periventricular hypothalamus and intralaminar thalamus, mammillary peduncle from Gudden’s nuclei and ventral tegmental area to the mammillary bodies and lateral hypothalamus). Top row, right: interruption of the medial lemnisci and classical spinothalamic tracts. Middle row, left: destruction of the rostral pole of the mesencephalic central grey with interruption of Schiitz’s bundle; right: medial tegmental lesion. Bottom row, left: destruction of the interpeduncular nucleus and the mammillary peduncles in the caudal mesencephalon; right: destruction of the ventral tegmental area and of the mammillary peduncles in the rostral mesencephalon. Anatomical abbreviations for Figs. 1 and 2, as follows. AVT: area tegmentalis ventralis; CI: colliculus inferior; CM: corpus mammillare; CO: chiasma opticum; CS: colliculus superior; GC: substantia grisea centralis; G M : nucleus geniculatus medialis; HL: hypothalamus lateralis; HPV: hypothalamus periventricularis; IL: intralaminar thalamus; IP: nucleus interpeduncularis; LM : lemniscus medialis; NGD: nucleus dorsalis of Gudden; NGP: nucleus profundus of Gudden; NR: nwleus ruber; N 111: third nerve; PM: pedunculus mammillaris; PP: pes pedunculi; SN: substantia nigra.
456
FINAL DISCUSSION
laterally placed lesions, which interrupted the medial lemnisci and the classical spinothalamic tracts, or by lesions restricted to the medial portion of the midbrain reticular formation. Some further acute experiments, in which lesions were inflicted on intact preparations that were subsequently decorticated, have shown that the reticular lesions which did not modify sham rage behavior were likewise incapable of changing the normally desynchronized electrocorticogram to a permanently synchronous pattern (Fig. 2). Only large tegmental destructions, including lateral and midline regions, sufficient to synchronize the electroencephalogram, constantly prevented the occurrence of spontaneous and evoked rage behavior after decortication. However, the rather severe, although by no means complete, impairment of hypothalamic efferent conduction, as well as the easy deterioration of the decorticate animal, call for caution in interpreting causes of the abolition of sham rage in acute preparations. Before results obtained in chronic preparations become available, we would simply emphasize that sham rage behavior does not appear to depend on the specific afferent pathways ascending to the posterior and lateral hypothalamus, although this conclu-
B
__ -“llr--
-
F”
b p
RLMG
f.*t
& !
“ E t
Fig. 2. The effect of medial tegmental lesions on the electroencephalogram, and upon sham rage development after subsequent decortication. Top: anatomical drawing showing the extent of the lesion. A : after the lesion, EEG from left (L.)and right (R.) frontal (Fr.) and parietal (Par.) leads still displays activated patterns. B: after subsequent decortication outbursts of rage are evident whichoccur spontaneously, or are easily elicited by tactile (tact. st.) or noxious stimulation (nox. st.), and by direct excitation of the hypothalamus (3 V). R respiration, BParterial pressure, L. EMG and R. EMG electromyograms from muscles of left and right forelimbs (From Malliani et al. 1963).
SPECIFICITY OF THE
u s
457
sion does not rule out the possibility that either the mammillary peduncle or Schiitz’s bundle system may be somehow engaged in the control (cg.,inhibitory in sign) of hypothalamic’ functions. However, the indication that sham rage behavior, like the waking state, may be maintained by tonic and/or phasic impulses running through the ascending reticular activating system, gives some support to the thesis that, at least in this field, the action of the reticular activating mechanism might be that of a rather widespread and indiscriminate energizer. The second aspect of our problem, that of the type of organization within the reticular system, has also been taken up recently by our group (Rudomin, Borlone, Malliani and Zanchetti 1961), by means of simultaneous recordings from the midbrain tegmentum and from what is considered to be one of the rostra1 projections of the reticular formation, i.e. the basal diencephalon. Only a very limited aspect of these researches, which are still in progress, will be reported here. It refers to the quite different picture of evoked potential distribution in the thalamus, subthalamus and hypothalamus which results from using various recording arrangements. The basal diencephalon of intact immobilized cats under local anesthesia was systematically explored with stereotactically oriented electrodes, and the electrical responses to single pulse stimulation of somatic afferent nerves (radial, sciatic, infraorbital) were displayed on a cathode ray oscilloscope. All electrode placements were verified histologically. Continuous monitoring of EEG, reticular evoked activity and arterial pressure ensured a steady normal state of the preparation. Fig. 3a-d shows the quite different distribution of somatically evoked potentials when either coarse monopolar exploring electrodes (180 y tip diameter) or concentric bipolar electrodes (1 mm interelectrode distance) were employed. With the former leading arrangement the diencephalic area responsive to somatic stimuli was very large and included almost all of the regions explored, i.e. the midline and intralaminar thalamic nuclei, the subthalamus, and the whole hypothalamus. Potentials picked up from several neighboring structures were rather similar in shape, polarity and latency. The contribution from the indifferent electrode was ruled out by using several reference points, including the lateral geniculate nucleus. Recordings with concentric bipolar electrodes gave a quite different and much more selective picture. The size of the somatically activated region was still fairly large, but highly responsive, scarcely responsive, and unresponsive regions could be clearly distinguished, as well as differences in the latency, duration and amplitude of the evoked potentials. Only slight inconstant responses could be recorded from the anterior and the ventromedial hypothalamic nuclei and from the preoptic region, while consistent, often high amplitude potentials were usually found in the zona incerta, H1 and H2 fields of Fore1 of the subthalamus, in the whole extent of the lateral hypothalamic area, and along the mammillo-thalamic system. Also the dorsal hypothalamic area and the posterior hypothalamus were often responsive to somatic stimuli. Recording with close bipolar electrodes made it possible also to outline rather distinct potential fields produced by somatic inflow to the diencephalon. Inversion of the polarity of the response was usually observed when the electrode tips passed through different active regions, e.g. from the median centre complex to the subReferences D. 461
458
FINAL DISCUSSION
thalamus, and from the subthalainus to the lateral hypothalamus, or from active to inactive areas, e.g. from the dorsal to the anterior and tuberal hypothalamus. As for the interpretation of these findings, it might be argued with some reason that a rather coarse electrode leading against a distant reference is likely to record not only local potentials, but also activity originating at some distance and conveyed to the exploring electrode by electrotonic spread. Such activity would not be picked up by concentric bipolar electrodes, either because of their higher resistance or because of cancellation. Of course, such an assuniption should be substantiated by M
B
.vhc--\
B
a
200 msec
M
-
B
*h
50 mrec
50
-\
\
5Omsec
+-/ M
B
200
B
B
//++--
b
%--J h d J I O 6 G 200 msec
50 msec
\
PV 200msec
5Omsec
Fig. 3. a-d The different patterns of distribution in the diencephalon of potentials, evoked somatically, when they are recorded by either monopolar or bipolar leading arrangement. The data represented in these figures have all been obtained in the same experiment. All the sweeps are responses to single shocks (0.1 msec, 5 V) to the contralateral sciatic nerve. The fast sweeps are distinguished from the slower ones by the time calibrations at the bottom of each column, where amplitude calibrations are also indicated. Columns marked M and B show responses led with the monopolar and the bipolar arrangements, respectively. The following anatomical abbreviations have been used: a H D : area hypothalamica dorsalis; CA : conimissura anterior; C C :corpus callosum; Cd: nucleus caudatus; Ch: chiasma opticuni; C1: capsula interna; CL: nucleus centralis lateralis; CS: colliculus superior; Fx: fornix; G C : substantia grisea
SPECIFICITY OF THE
RAS
459
identification of the distant source of the evoked activity. Another explanation we would like to suggest tentatively is that the potentials led unipolarly might also represent local activity, and that in large areas of the diencephaIon, as well as of the reticular formation, two types of evoked activity might co-exist. the former type being too diffuse and uniform to be picked up by close bipolar electrodes, while the latter would be organized in definite and restricted potential fields better delimited by the concentric leading arrangement. Of course, we have no decisive data to prove either of the hypotheses mentioned
M
-
B
C
\ 9LJ“-3 --31oorv
50msec
50msec
2OOmsec
B
M
,% Jk.
\
--
B
B
d
iec
SDmsec
200msec
50msec
centralis; GP: globus pallidus; H 1, H 2: Forel’s fields; Ha: hypothalamus anterior; HL: hypothalamus lateralis; Hp: hypothalamus posterior; Hvm: hypothalamus ventromedialis; IP: nucleus interpeduncularis; LD : nucleus lateralis dorsalis; LME: lamina medullaris externa; LP: nucleus lateralis posterior; MD: nucleus medialis dorsalis; Mm: corpus mammillare; NCM : nucleus centralis medialis; Ped. : pedunculus cerebralis; PVH: nucleus paraventricularis hypothalami pars dorsalis; R. : nucleus reticularis; RE: nucleus reuniens; RPO: regio praeoptica; SO : nucleus supraopticus; SMx : commissura supramammillaris; Spt: area septalis; Sth: nucleus subthalarnicus; TO: tractus opticus; TTC: tractus tegmentalis centralis; VL: nucleus ventralis lateralis; VM : nucleus ventralis medialis; VPL: nucleus ventralis posterolateralis; VPM : nucleus ventralis posteromedialis: ZI: zona incerta (From Rudomin et al. in preparation). References p . 461
460
FINAL DISCUSSION
above, but the possibility of the co-existence of a more diffuse and a more restricted evoked activity in the diencephalon should be given careful consideration. Recording with tungsten microelectrodes 1-2 p in tip diameter is presently used to show what hypothalamic areas respond with slow potentials and unit discharges to somatic, visual and auditory stimuli. Further information can be derived from the effect of localized lesions around the tip of the recording electrode. As Fig. 4 shows, for a somatically evoked response in the posterior hypothalamus, a very limited tissue electrolysis through the tip of the recording microelectrode changes the positive-negative configuration of the potential to a purely positive deflection, an observation which suggests that the abolition of focal activity has made the recording region a source of adjacent electrical activity. This approach is extended to responses monopolarly recorded from the anterior hypothalamus and the preoptic region.
A 1 4
82 4
c2 4
D3 .4
4 10 msec
Fig. 4. The influence of localized tissue destruction on evoked potential in the posterior hypothalamus. Tungsten microelectrode recordings. Single shock stimulation to the superficialradial nerve. Simultaneous recordings from the diencephalon and the midbrain reticular formation are shown on the left of the figure. From A through D the midbrain electrode is kept in the same place for reference (4), while the diencephalic electrode is moved from CL (A 1) to Hp (B 2) and from upper part of Hp to lower part of the same nucleus (D 3). Between B and C there is no movement of the diencephalic microelectrode, but a small tissue electrolysis is performed through the tip of the microelectrode. Note the change of the positive-negative configuration of the potential to a purely positive deflection, while the previous diphasic configuration returns as soon as the electrode is lowered a few tenths of a mm, thus attesting the limited size of the lesion. The anatomical drawings on the right show the exact localization of the microelectrodes. For abbreviations, see the legends of the preceding figures (From Malliani et al. in preparation).
SPECIFICITY OF THE
46 1
In view of the lack of more numerous and crucial data, we would not like to commit ourselves to any of the possible interpretations on the intrinsic reticular organization. We would stress, however, that the present trend to conceive the reticular activating system as a highly and specifically organized mechanism might be justified, but only in part, because some specificity in its functioning does not at all contradict its functioning in other conditions more massively as an energizer of cortical and subcortical functions.
REFERENCES J., BIZZI,E., MALLIANI, A. and Z A N C H EA.~ ,Ascending afferentpathways maintaining APELBAUM, sham rage behavior in the acute decorticate cat. Proc. 5th Znt. Congr. Electroenceph. elin. Neurophysiol., Rome, 1961: 190-191. BARD,P., A diencephalic mechanism for the expression of rage with special reference to the sympathetic nervous system. Amer. J. Physiol., 1928,84: 490-515. NAUTA,W. J. H., Hippocampal projections and related neural pathways to the midbrain in the cat. Brain, 1958,81: 319-340. RUDOMIN, P., BORLONE, M., MALLIANI, A., and ZANCHETTI, A., Evoked electrical responses to somatic stimuli in the basal diencephalon of the cat. Proc. 5th Int. Congr. Electroenceph. clin. Neurophysiol. Rome, 1961: 217. ZANCHETTI, A., Somatic functions of the nervous system. Ann. Rev. Physiol., 1962, 24: 287-324.
P. K. ANOKHIN
I am glad to note that Prof. Zanchetti’s point of view inclines to the admission of the specific properties of ascending activation. The object chosen by Prof. Zanchetti for his research: fury (rage) is especially suited to this kind of research. It is known that the specific features of ascending activation were discovered in our laboratory particularly on biological integral acts of behaviour (pain and hunger activation). But I must remark that we must, at this time, draw a most careful line between the differentiated structure and differentiated functions of the brain stem area in general and biologically specijk ascending activations to the cortex. The latter must now become the subject of most intensive research. The inadequacy of the general classification of the ascending excitations created at the time when the physiological properties of the reticular formation of the brain stem were first discovered is becoming more and more evident. But I would not say, with Professor Zanchetti, that the existing classification of excitations into “specific” and “unspecific” showed “naivete” in the manner of setting the problem. Such is the law of scientific progress: all that is new is always lifted to a higher level from which the defects of the past are visible, but these defects usually have a lawful origin. Wherein lies the inadequacy of this classification of excitations into “specific” and “unspecific”? First of all, in the vagueness of our position in relation to the criterion for this classification. Every classification of phenomena must have a strict criterion; otherwise the same phenomena may fall into quite opposite categories as a result of the wrong choice of criteria.
462
FINAL DISCUSSION
It has been assumed that specific afferent excitation always has a definite sensory modality and that unspecific excitation does not have such modality specificity. This is actually true. But it is also true that the ascending excitations of the reticular system produced a generalized activity which led to their wrong classification as “diffuse”. I f sensory modality were the only property of ascending excitations, the existing classification would be quite satisfactory. The data collected within the last few years have provided growing evidence of the existence of other criteria for the classification of ascending activations. Of the criteria obtained in our laboratory the biological quality of ascending activation is most important. As you have seen from my report, this conception has developed gradually, beginning with the selective influence of narcotic substances. As a result, we have now approached the construction of a new classification, which would embrace the most important parameters of the classified excitations. I am giving below a table which can serve as a model for the classification of the most important criteria of ascending activation.
TABLE I -
No.
__ Criterion of classification
~-.
Source of ascending influences
Lemniscal system Thalamic system Brain stem Hypothalamus
~
~
Generalization
Sensory modality
Biological quality
local semi-generalized generalized (a) generalized (b) semi-generalized
specific unspecific unspecific
unspecific specific specific
~
~
unspecific ~
~.
specific
-~
~~
The three criteria in the above table : generalization, sensory modality and biological quality, are most certainly not the only criteria for the differentiation of ascending activations. Further research will undoubtedly disclose other peculiarities of these excitations, as yet unknown to us. But, even at this stage, it can be said with certainty that no classification can be adequate if it is founded on the direct stimulation of the reticular structure of the brain stem, because when this is done, it is impossible for us to recognise any biological quality in the desynchronization of the cortical electrical activity obtained. On the contrary, research correlated with some form of behavioural reactions will always yield more adequate results, because the actual part played by ascending excitation has developed by evolution in connection with the formation of different acts of an adaptive nature. This has been clearly shown in the latest experiments in our laboratory, which have proved that the ascending “hunger” activation of the frontal areas of the cerebral cortex is not blocked by anaesthetics ( K . Sudakov 1961).
STIMULATION OF THE CENTRUM MEDIANUM
463
It is interesting to note that chlorpromazine, which blocks ascending pain activation, has, as we have described, no influence on hunger activation in a hungry cat. In addition to what has been said about ascending activations of a generalized character, we can indicate other kinds of ascending influences, those which arise, for example, in the form of primary and secondary discharges in response to stimulation of some receptor surface. In phenomena of this kind observed by us, we always had some electrical summation of many individual ascending discharges, which probably had different physiological significances. It seems that we can now form a definite idea of the prospects of future research on the characteristics of specific but selectively-generalized ascending activations. If every ascending activating influence on the cortex has a selective effect on definite synapses of the corticalcells, this suggests several very important questions of principle. The first among them is the problem of chemical specificity of the synapses on which the ascending activations of a different biological quality are projected. How do these synapses differ from each other? What is their relationship? This opens up a wide prospect for new research, which must combine, on an analytical and synthetic level, the study of the mechanisms of the brain and of behaviour.
K. LlSSAK
A Dual Behavioural Effect from Stimulating the Region of the Centrum Medianum Direct electrical stimulation of various brain structures is, despite its highly unphysiological character, still one of the most important methods giving some information about the special processes by which definite parts of the brain contribute to the intricate mechanism of sensory integration and the learning process. Different varieties of this method have been widely used also in chronic experimentation in the last few years. We have in our laboratories also carried out, since 1954, a series of systematic stimulations of different subcortical structures on a background of different conditional reflexes. During the last year the diffuse thalamic projection system was more thoroughly investigated. It was found that the stimulation of different subcortical structures can influence conditional reflexes in a characteristic manner. I should like to compare very briefly the important differences which have been found between the influences of the diffuse thalamic projection system and those of lower parts of the ascending activating system. This might give some insight into how this system works under natural conditions. As a general rule it was found that stimulation of various points of the hypothalamus or mesencephalic reticular formation always influenced antagonistic conditional reflexes in a reciprocal manner. Stated more precisely this means that, when stimulaReferences P 465
464
FINAL DISCUSSION
tion of a given point activated the avoidance response, it inhibited the conditional alimentary response; and vice versa. The same reciprocal effects manifested themselves in the after-effects of the stimulation. A remarkable difference was found by Kopa, Szab6 and Grastyin when they stimulated the region of the centrum medianum of the thalamus in an avoidance situation. In an avoidance situation there are two, essentially different, moments when the influence of the stimulation can be checked. The first is the moment when the animal sits on the metal grid from which it gets the electrical shock, the second when it sits on the place (in our case it is a little bench) where it takes flight after the electric shock, or after the conditioned stimulus. It turned out that stimulation of the region of the centrum medianum in these two situations resulted in two absolutely different behavioural effects. If stimulation was applied on the grid, it induced manifestations of fear and promptly activated the avoidance response. If stimulation with exactly the same parameters (generally 3 V, 100 c/sec, 1 msec) was applied on the bench, a relaxing or even sleep-like effect was elicited (Fig. 5). Both effects could be invariably repeated. When stimulation of the same points was checked in an alimentary situation, it was found that the same parameters also activated the alimentary reflexes. In two cases, in fully satiated animals, stimulation elicited a reaction of turning away from the food tray. These reactions were, however, not so conspicuous as those observed in the avoidance situation. The same dual phenomenon was found in 5 cats, but exclusively in the region of the centrum medianum. Other parts of the diffuse thalamic projection system have given similar effects to those of the hypothalamus and reticular formation.
8 +72
1'
[;-
3I'
Fig. 5. A dual behavioural effect from stimulating the region of the centrum medianum. A . Stimulation on the grid floor of the cage (where the animal got shocks during the establishment of the conditional reflex) elicits the characteristic avoidance response. B. Stimulation on the bench (where the animal never got shocks) elicits a relaxing effect. Speed of recording 24 pictureslsec. Numbers on the top denote the picture selected and copied graphically. Numbers below the pictures show the time in seconds after the application of stimulus. Note the different time courses of the two effects.
CONSIDERATIONS
465
In an attempt to explain these observations it was supposed, on the basis of electrophysiological data (Jasper 1949 and Monnier et al. 1950, 1959) that two reciprocally interconnected and overlapping systems of the centrum medianum might be responsible for the dual effect. Tt may be assumed that the excitability of the two systems can be influenced selectively or can even be blocked by the sensory influence coming from the actual environment. On the basis of our observations this general switching mechanism seems to be one of the most specific functions of the diffuse thalamic projection system. REFERENCES KOPA,J., SZABO, 1. and GRASTYAN, E., A dual behavioural effect from stimulating the same thalamic point with identical stimulus parameters in different conditional reflex situations. AcfuPhysiol.Hung., 1962, 21: 207-214. JASPER,H . H. Diffuse projection systems: the integrative action of the thalamic reticular system. Electroenceph. clin. Neurophysiol., 1949, I : 405420. MONNIER, M. Action de la stimulation Blectrique du centre somnoghe sur l'klectro-corticogramme chez le chat. Rev. Neurol., 1950, 83: 561-563. TISSOT,R. et MONNIER,M. Dualit6 du systeme thalamique de projection diffuse. Electroenceph. clin. Neurophysiol., 1959, 11: 615-686.
J. M. BROOKHART
I have no direct questions to put to any of the speakers at this colloquium. T am going, instead, to assume the role of the Devil's Advocate and address a general proposition to the group. In the past few years we have been operating on the general proposition that input to the central nervous system is achieved through the activation of both specific and unspecific pathways, and it is possible to differentiate the two on anatomical and functional grounds. T would propose that this dichotomy is no longer meaningful. During the past week we have seen numerous examples of extensive spatial as well as modality convergence on single units and populations of neurons at several levels of organization. These sorts of observations make sense to me, for the unanesthetized nervous system is a vibrant and tremendously busy collection of cells, operating as an integrated whole. Consequently, it might be predicted that an event in one portion of the structure would be quite likely to produce perturbations and repercussions in many other portions, independent of the fact that the activity produced might occupy only a specific anatomical pathway during one stage of its transmission. 1think this is particularly true when you consider that the problems we present to the nervous system consist, for the most part, of abnormal patterns initiated by electrical shock, or events so insignificant as to be entirely lacking in or of questionable meaning for the preparation. T wonder, therefore, if it is not appropriate to consider discarding the concept of specific and unspecific systems and to look for a more workable framework upon which to construct hypotheses, upon which to design experiments in which more sophisticated questions are put to the central nervous system in such a fashion as to produce a n integrated output which can then be subjected to analytical dissection.
466
FINAL DISCUSSION
W. GREY WALTER
Like Dr. Brookhart, I am sometimes discouraged by the apparent anarchy that has been revealed to us, as if anything could go anywhere in the CNS. If the system were really random on the scale of the human brain, this would be an engineers nightmare and the neurophysiologists funeral. Our epitaph would be: “They looked for everything everywhere and discovered nothing: Omnia ubique nihil”. But the CNS is not a random network and has in fact an elegant anatomy peculiar t o each species, or at least each genus. We should, I think, consider inter-specific, specific and unspecific mechanisms more carefully, and also the phylogeny of sensorymotor integration. At the same time we should make a great effort to study the ontogeny, the maturation of these mechanisms, particularly in man. This is one of the problems I am planning to tackle during the coming years and we already have some very odd information about the three categories of unspecific effects in children. Some children seem to behave more like cats than humans, but the personal variations are baffling still. Lastly, great use is made in animal neurophysiology of pharmacological dissection and amplification, but we cannot use many of these components in manchloralose and strychnine for instance. Could not some of you investigate the action of the drugs we can use in man? Alcohol, for instance, has not been mentioned, but weall know that it has an action on some of these mechanisms; what action precisely? People still talk of it “dis-inhibiting”, but what sort of inhibition and where? Pavlov, of course, used caffeine a great deal, and this again nearly everyone takes freely everyday. Why are these two drugs of almost universal, mild addiction? Perhaps because they slightly simplify the system that we have been saying is “really very complicated”.
A. FESSARD
Integration: A Commentary on the Pisa Symposium
Lord Adrian (1949) wrote: “It is difficult to resist speculating about the integrative processes of the brain because the whole of human achievement depends on them.” The contributions to this Symposium have once more underlined the fact that it is impossible to discuss either neuronal mechanisms or the functions of the central nervous system as a whole without having constant recourse to the concept of Integration. In principle, the theme of our meeting has been such as to emphasize this fundamental aspect of nervous activity, for we have been invited to consider “Sensori-motor Integration”, i.e. the mechanisms which confer dynamic unity on the temporal sequence: Sensory Excitation + Central Elaboration + Specific Motor Response (A + M + X, Fig. 6). This unity is well revealed in the observation of innate or conditioned reflexes.
INTEGRATION : A COMMENTARY
467
In fact, although the papers of Jasper, Buser, and Hagbarth are in line with this general thesis, almost all, if not all the others, may be said to be only partially concerned. Their authors have indeed worked within the general framework, but they have dealt with other forms of integration; collectively, these partial studies greatly contribute to a profound analysis of sensori-motor mechanisms and to a better comprehension of their functional consequences. What in fact are the forms of integration which have been considered in the course of this Colloquium? A rapid survey of the contributions suffices to reveal such a diversity of aspects of neural operations and mechanisms as to entail a possible risk of confusion. The work presented at this meeting furnishes an excellent opportunity to reflect on this concept of “integration”-so general, so widely invoked in neurophysiology, and yet so ill-defined.
Fig. 6.
The general theme of the meeting especially invited the consideration of the integration of “specific” and “unspecific” activities, and this subject was indeed dealt with by about a third of the participants. Among the other contributions together with or apart from the theme of sensori-motor integration itself, are found attempts to conjoin: excitation and inhibition (cf- Eccles, Granit, Albe-Fessard, Jung, Jasper) ; interacting multisensory projections (cf. Albe-Fessard, Jung, Buser) ; the somatic and the vegetative (cf. Dell, Anokhin, Grey Walter); successive segments of activity (cf. O’Leary, Gastaut, Brazier), etc . . . . Beside this, there must also be considered certain anatomical relationships existing between various nervous territories, implicated in a single behavioural or viscero-regulatory function, and thus qualifying for the appellation of “integrative”. Finally, in addition to this spatial integration, various forms of temporal integration must also be taken into account, for example in conditioned reflexes, where, so to speak, traces of the past come to be integrated with the present state (cf. Anokhin, Beritashvili, Brazier, Jasper). Once again, this multiplicity of aspects of the problem stresses the need to compare and classify the various points of view. The simple definition of Bullock (1957) may be taken as a starting point. According to him, integration is an operation “to put parts together into a whole”. But there are many ways in which the nervous system, and the products of its activity, can be divided into parts; and what may, under different circumstances, be called a “whole” depends on the numerous ways in which may be considered the dynamic entities which are presented within the two domains of physiology and psychology. The problems of integration are rightly bound, in their diversity, to the different choices which present themselves, either in the processes of analysis, or in the cohesive dynamic syntheses which we call “functions”; and in these choices, there is a greater or lesser degree of convention, of relativity and even of arbitrariness. Thus it may be said that the study of nervous integration is presented under two aspects, distinct Relerences P. 474
468
FINAL DISCUSSION
and complementary, which have been constantly intermixed in these discussions : on the one hand, it is a question of studying for their own sakes the mechanisms of cohesion between elements which the necessities of analytic research force us to separate artificially-elements which may be anatomical (nuclei, neurons, or parts of neurones according to the level at which the experiment is conducted), or dynamic (action potentials, potential gradients, chemical transmitters, etc . . .); on the other hand, and obversely, it is the observation of a global operation set in motion by the nervous system which serves as the point of departure, in so far as such an operation appears, outside of any reference to nervous mechanisms and according to varying criteria of greater or lesser validity, as forming a “whole”, an integrated dynamic unity; thus guided from the beginning, we shall be led to seek in the central nervous system for coherent anatomo-functional systems capable of forming the neurophysiological substrate of, for example, alimentary behaviour, a reaction of choice (i.e. the problem of decision), or the perception of a form (Gestalt), or yet again the varying degrees of wakefulness, etc . . . Likewise, the feeling which we all have of the unity of our conscious personality can be ascribed to the most closely integrated processes of higher nervous activity, as has been thought by most authors who have considered this difficult problem (cf. Fessard 1954). In all these generalisations can be recognised some of the particular themes which have been developed at this meeting, either in the one direction, or in the other. The two approaches must obviously come together in working to a common end, which is to explain co-ordinated behaviour, somato-visceral regulations, and even mental operations by “the integrative action of the nervous system”. This reminder of the title of a famous book which we have all studied and admired provides an opportunity to quote from it some relevant passages. Having noted in its introduction that the properties of nervous tissue are in the first instance those of ordinary cellular metabolism, subsequently becoming specialized for the transmission, at great speed, of states of excitation (nervous impulses), Sherrington adds: “. . . a third aspect which nervous reactions offer to the physiologist is the integrative. In the multicellular animal, especially for those higher reactions which constitute its behaviour as a social unit in the natural economy, it is nervous reaction which par excellence integrates it, welds it together from its components, and constitutes it from a mere collection of organs an animal individual. This integrative action in virtue of which the nervous system unifies from separate organs an animal possessing solidarity, an individual, is the problem before us in these lectures . . .”. In this text, integration is viewed as the result of the co-ordinating action of the central nervous system, an action which is brought to bear on a multiplicity of effectors and which ensures the synergy of somatic or visceral behaviour. This is the operation shown schematically in Fig. 7(2) and it is one of the ways of conceiving the integration of elements of action such as X, Y, Z. But Sherrington’s great conspectus did not overlook other aspects of the problem, beginning with sensori-motor integration (Fig. 6), when he states (p. 7, edition of 1947) ,,. . . The unit reaction in nervous integration is the reflex. . .”; and at a higher level of complexity, speaking of the spatial and temporal co-ordination
INTEGRATION : A COMMENTARY
469
of multiple reflexes: “this compounding of reflexes with orderliness of coadjustment and of sequence. . .” (p. 8, edition of 1947); and finally, in the elaboration of the “final common pathway”, this other aspect of integration which is expressed in the convergence of afferent impulses (Fig. 70)). It is noteworthy that it is this last aspect which has for the most part been considered at this Symposium. For most of us, nervous integration consists essentially of this concentration of messages of peripheral or central origin on discrete areas of projection, association, or of efferent outflow. Modern microphysiological research, in the van of which stands this distinguished School of Pisa, has made it possible to study these various modes of convergence at the level of the single neurone. The systematic researches of Albe-Fessard, of Jung, and their collaborators, have demonstrated precise examples of this at the present meeting. They have dealt at the same time with the convergence of messages of different sensory modality or different topical origin, and, from another point of view, of the interactions which take place between specific messages and non-specific afferent impulses.
Once again, these studies have underlined the widespread disposition of structures showing convergence, which are present at all levels of the neuraxis: in the spinal cord, the brain stem reticular formation, the intralaminar and midline nuclei of the thalamus, the caudate nucleus, the hippocampus, and the association cortex. The motor cortex, the highest level of integration for neurones which discharge into pyramidal and extrapyramidal pathways, has been studied by Buser and his collaborators; in general, it may be said that the systematic investigation of the higher centres of integration has been one of the most successful fields exploited by contemporary neurophysiology. Moreover, the spinal motoneurone, prototype of integrating centres, has not been overlooked, and has been usefully discussed by Eccles, Granit, Brookhart and Kubota, Hagbarth and Finer. Can all forms of integration be said to have been mentioned? Certainly not; the most basic type-at least, that type which conditions all others-remains to be considered. This is the integration which is produced within nervous structures themselves, and may be described as intrinsic integration. This results from interactions between the fractions which have been artificially separated by the analytic procedure. The central nervous system is not only an integrator in the double sense shown in Fig. 7(2) and (3); it is itself integrated. It is because of this property that it can exert its co-ordinating function on the structures placed under its control; to the extent to which this property may be lacking (immaturity, trauma, pathological change, References p . 474
470
FINAL DISCUSSION
senescence), behaviour, organic regulation, mental activity may appear less coherent, sometimes even totally disorganised. On the other hand, structures showing convergence, even though they create conditions favourable to integration, do not explain its intimate mechanism. This latter should be sought in the interactions which exist between the structural or dynamic elements, which are concerned in each case, according to the level at which the experiment is performed. At the level of gross structures, the areas concerned in a functional operation are interconnected by bundles of nerve fibres. This is the first stage of the study of the inside of the “black box” (M of Fig. 6), now imagined open. Firstly, one can observe interactions according to linear sequences, as in Fig. 6; but the term integration should be avoided in the case of a simple relay (or chain of relays), which simply transmits signals 1 : I (cf. Bullock 1957). In actual fact, it is doubtful whether such a system ever occurs in the central nervous system of Vertebrates. It is almost always a case of intricate and complex interconnections, which are at the same time convergent, divergent, bidirectional and retrograde (Fig. 8). It is precisely these last connections which, when they are of “negative feedback”-type, ensure internal coordination in the whole system concerned, and great stability in its external manifestations. Such properties plainly justify the qualifications of “well integrated” for such systems. A good example of this can be seen in the report of Dell, whose collaborators have demonstrated two great self-regulating circuits : one between the reticular activating system and the cortex (Hugelin, Bonvallet, reported by Dell), and the other between the medulla oblongata and the midbrain (Bloch and Bonvallet, ibk,!.). These circuits unite their effects with other (extracerebral) mechanisms in such a way as to ensure for the Reticular Formation a powerful homeostatic r81e. For its part, the group working with Albe-Fessard has been investigating certain properties
t
exc.
1 inh.
of thalamo-cortical circuits-two-way circuits which, it appears, play a part in the arousal and alerting reactions. The thalamus may be cited as a typical example of a cerebral organ which is capable of all types of integration. It is, in the first place, a relay in the sensori-motor pathway; it is also, in certain of its nuclei, a centre of polymodal convergence, and subsequently it is the point of departure of divergent projections which control various areas of cortex. Furthermore, it contains complex internal interrelations between its various nuclei; they have been and are the object of numerous controversies. Altogether, it forms, with the cortex, a powerfully integrated functional system.
INTEGRATION : A COMMENTARY
47 1
There is a strong temptation to have recourse to cybernetic schematisations in order to represent such organisation as that just described. Hugelin has done so with enthusiasm and not without profit. Jung has made allusion to the models of von Holst and to the concepts of McKay. In such schemata, it is frequently useful to think at neuronal level, and we would do the same in the last part of this summary. The neurone itself is a microcosm in which can be found all the fundamental functions which have been confounded under the single name of integration. In the first instance, in the normal sequence of its segmentary activities, from the dendrites to the termination of the axon, the neurone is a small-scale model of sensori-motor liaison, and this, in so far as only one individual cell is concerned, in the most “integrated” fashion possible. On its synaptic receptive surfaces there converge hundreds, sometimes thousands, of afferent terminals, often of various origins; and on this surface, between neighbouring synaptic contacts, there occur electrotonic and perhaps chemical interactions. Subsequently, there takes place the operation which perhaps best deserves to be called “integration”, which strictly defined is “decisive”, and which happens on the initial segment of the axon where propagated action potentials originate, when the critical threshold has been achieved by spatial or temporal summation. Finally by distributing its excitatory or inhibitory signal to all its axonal terminations, the neurone exerts an integrative action on all the excitable elements with which it is connected. The integration pertaining to the individual neurone has been here presented under several forms. In the paper of Brookhart we have been shown a striking example of the summative properties of dendritic arborisations, properties which they owe above all to their relative or absolute inability to respond by an all-or-none discharge. Electrotonus, whether reinforced or not by decremental conduction, appears as the variable factor responsible for this integration. This latter, whether resulting from spatial or temporal summation of post-synaptic effects, does not necessarily result in the discharge of an impulse. I n most cases, the neurone displays “spontaneous” activity, and the result of this post-synaptic integration is expressed as a modulation of the frequency of rhythmic discharges. Such is the most general mechanism of neuronal integration, according to modern concepts. In the paper of O’Leary and his collaborators, we were introduced to a new mechanism of interaction, the stable potential gradient whose function cannot be understood unless it be considered together with other forms of nervous electrogenesis, i.e. spikes and non-propagated slow potentials; for it depends on them and they depend on it, in such a way that it would be impossible to interpret one of these phenomena without knowledge of the other two. But the most significant form of integration, that which is richest in its functional consequences, is undoubtedly that which takes places between excitatory and inhibitory processes. Sherrington, in his masterly fashion, blazed the trail. Today microelectrode techniques allow us to study this interaction at unitary level (as shown in several of the present contributions), and sometimes even to investigate it quantitatively, as has been done by Granit for a type of recurrent inhibition. The process of integration in this case can be symbolically expressed in simple mathematical form, Refermcrr
n. 474
472
FINAL DISCUSSION
as the algebraic sum of generator potentials (their amplitude being deduced from the frequency of impulses discharged by the neurone). We have now to take account of a new dynamic entity, “presynaptic inhibition”, whose properties have been described and presented by Eccles. Henceforth this should be included among the other processes of basic central activity. In the last analysis, in the synaptic microcosm in which originate the essential properties of the nervous system, the morphological substrate where the cohesion of the elements is stronger than anywhere else is formed by the terminal segment of the axon and its corresponding sub-synaptic zone. Histochemical procedures, electron microscopy, and the analysis of its dynamic properties, show it to be more complex, and richer in its potentialities, than was at first imagined. In consequence of its biochemical specificity (whose nature and diversity themselves raise many further problems), it is indeed the site of processes of temporal integration richer in their functional significance than would be the case were it only a question of the simple electrotonic processes of the summation of local potentials. While this is a subject which has not been touched upon at this Symposium, it nevertheless deserves to be mentioned because of its implications in sensori-motor integration. The short-term conservation of traces of activity, such as is seen in the phenomenon of “post-tetanic potentiation” has perhaps greater functional significance than is commonly believed; and if the somewhat daring hypothesis of a stock-piling of information, coded and integrated in macromolecules, turns out to have an organic basis, it is at this fundamental synaptic level that it will be necessary to locate the operations which determine such processes. Inter-neuronal integration, finally, represents the intermediate stage of unified organization which, starting from synaptic integration, enables us to begin to understand the integration which mutually co-ordinate gross nervous structures. At this level of “neurone pools”, functional cohesion is equally the concern of interconnections (axo-dendritic or axo-somatic), showing the various arrangements which have already been mentioned. In the complicated and compact network of convergent, divergent, intercurrent and recurrent connections, showing an extraordinary “power of connectivity”, each neurone has a strictly defined relationship with its neighbours. A synchronising interneuronal influence, exerted by the resultant electric field, may also contribute, in certain cases, to the making of ensembles of great internal cohesion within these pools. However, the definition of their state of integration poses a difficult problem, which can only be resolved by micro-electrode sampling of neurone pools participating in one and the same function, and by appropriate data processing. Macroelectrode recording of evoked or spontaneous potentials provides a sort of experimental electrical “integration”, derived from a multitude of unitary activities occurring within a more or less defined space. In certain cases, it appears not improbable that the instrumental integration reflects in some way a similar process of interneuronal integration, but it would be imprudent to generalize from such an assumption. Arduini has demonstrated an instance, that of the “dark discharge” of the retina, whose function in the visual perceptive system seems clear enough: the signalling of stable illumination, the contrasting effect of lateral inhibition, and a factor
INTEGRATION : A COMMENTARY
473
which augments visual acuity as a function of illumination, by increasing the signal/ noise ratio. This obviously presupposes a central integration of the tonic discharge, involving all the active elements and with an extended period of activity. In the same way, in investigations utilising evoked potentials for studying the formation of conditioned reflexes (cf. Jasper, supra; discussion by Ricci), it is tempting to see in this cortical evoked potential something more than a simple sign, but rather perhaps the partial reflection of the evolution of an integrative process whose localisation is not necessarily cortical. The temporal summation of successive evoked potentials, obtained by repetitive stimulation, may seem to be no more than an instrumental integration, largely practised to day to increase the signal/noise ratio (see Brazier, Gastaut, Grey Walter, supra). But Brazier, after comparing a large number of such mean profiles under various conditions, has been led to suggest that the profile of activity of an aggregate of neurones may well exist as the product of elaboration of certain nervous centres and so constitute an aspect of the treatment of information by the brain. This would signify, for the brain, the means of having a criterion for identifying the dissimilar, the new, or the improbable; that is to say, to be “informed” in the Information Theory sense of the word. The phenomenon of habituation, and, by contrast, the reactions of alerting and surprise, lend some credence to this hypothesis; but, as the author herself stresses, this “probability” model, if it corresponds to some anatomo-physiological reality, should not replace suggested models in which every sensori-motor connection depends on the configuration of determined unitary neuronal activities. At this level of neuronal aggregates, it appears necessary not to limit ourselves to the statistical aspects (which do not result from a true integration), but to consider the c‘shape’’of organised neuronal activities. As Jasper expresses it in his contribution, it is necessary to consider the metamorphoses of the temporal and spatial configurations formed by individual unitary discharges - some neurones being excited, others inhibited. These metamorphoses take place in series at different levels of the neuraxis, from the primary sensory relays to the ultimate effectors, in a regular sensori-motor sequence. Microelectrode sampling techniques, in demonstrating the variety of assortment of afferents of neighbouring neurones in an aggregate (particularly in associative regions) strongly suggest the reality of the intricacy of functional neuronal configurations. It is reasonable to define their specificity, as does Anokhin, as that of the reaction pattern rather than that of a sensory modality. But what is it that is responsible for the functional unity of each of the neuronal aggregates in network-like structures, where the interconnections are jumbled up without apparent structural differences? Biochemical specificity (cf. Anokhin) determined by heredity? Synaptic differentiation, gradually acquired by experience? Or the existence of organised assemblies capable of responding only to a certain afferent pattern, like a lock to a key? None of these hypotheses should at present be rejected. The last of them recently inspired Bullock (1961) to some interesting reflections : “All relevant input and central predisposition”, he said, ‘<mustbe read and finally evaluated by a single integrator”. Taking as examples acts as varied as the sudden take-off of a fly, the quick recognition References P. 474
474
FINAL DISCUSSION
of a friend’s face, and the decisive choice between two forms of behaviour, he attributes their triggering-off to “recognition-of-criterion units” and suggests three possible mechanisms for such integration : specific neurones, a mass of neurones randomly interconnected and discharging by virtue of a critical proportion of its undifferentiated units (model of Beurle), or a metastable system with specific connections and with positive feedback (as suggested by McKay). Although these last considerations have their place in the theme of sensori-motor integration, it would seem dangerous to use them as generalisation outside of the particular case in which the passage from sensation to action (motor or mental) is sharply determined by a mechanism of binary option (“all or nothing” or “either-or”), implicating a critical threshold. Even in this case, integration would appear to depend on the cohesion of the ensemble of nervous structures involved rather than on a final concentration of signals on one specific neurone; Bullock himself concludes by saying that the third possible mechanism (metastable system with positive feedback) appears to him to be the most likely, at least for “higher level” decisions. But in fact, a system is not well integrated unless it is stabilised by negative feedback, and the operations of “decision” considered by Bullock should be rather considered as the sudden transition from one state of integration to another. To conclude, it seems that in the general case of a multineuronal assembly engaged in a specific sensori-motor operation, the question of a special mechanism destined to confer upon it in isolation the quality of an “entity”, does not really exist. It can only be called “integrated” in relation to all the congruent assemblies, at all levels of the neuraxis, which participate in this particular operation. This must be the last word in this essay on sensorimotor integration, in all its many aspects, to close a Colloquium which must have given us all much food for thought. I wish to acknowledge Dr. David Bowsher’s kind assistance in the English translation of the French written version of the present essay. ADRIAN, E. D. Sensory integration. The Sherrington Lectures. University Press of Liverpool, 1949: 1-20, T. H. Neuronal Integrative Mechanism. In Recent Advances in Invertebrate Physiology. BULLOCK, University of Oregon Publications, 1957: 1-20. BULLOCK, T. H. The Problem of Recognition in an Analyzer made of neurons. In W. A. ROSENBLITH (Editor), Sensory Communication. M. I. T. Press. 5. Wiley and Sons, New-York, 1961: 717-724. FESSARD, A. Mechanisms of Nervous Integration and Conscious Experience. In J. F. DELAFRESNAYE (Editor), Brain Mechanisms and Consciousness. Blackwell, Oxford, 1954: 200-236. FESSARD, A. The role of Neuronal Networks in Sensory Communications within the Brain. In W. A. ROSENBLITH (Editor), Sensory Communication, M. I. T. Press, J. Wiley and Sons, New-York, 1961: 585-606. SHERRINGTON, C. S. The Integrafive acrion ofthe Nervous System (1906). New ed., Cambridge University Press, 1947, 433 pp.
Author Index Italics indicate the pages on which the paper of the author in this volume is printed. Ades, H. W., 219, 222 Adey, W. R., 262-263, 311, 316, 401, 402, 422, 449 Adrian, E. D., 137, 184, 185, 241, 247, 250, 253, 269,296,407,466 Agafonov, V. G., 326 Ajmone Marsan, C., 120, 144, 155, 214, 309, 312, 316, 351 Akert, K., 31 1 Akimoto, H., 144, 155, 218, 225 Aladjalova, N. D., 258 Alanis, J., 38 Albe-Fessard, D., 115-145, 207, 208, 220, 221, 227,266, 312, 315, 350, 352, 360, 368, 434 Alvord, Jr., E. C., 297 Amassian, V. E., 116, 120, 128, 221, 227,244246, 302 Anderson, B., 31 1 Anokhin, P. K., 74, 325-338, 461-463 Apelbaum, J., 454 Appelberg, B., 155 Araki, T.,153, 55 k a n a , R., 309 Arden, G. B., 214 Arduini, A., 132, 185-201, 259, 442 Ascher, P.,294-324 Ata-Muradova, F., 337 Atlas, D., 355 Azerinsky, E., 424 Bailey, R. A., 127 Ballin, H. M., 155 Bard, P., 413,439,454 Barlow, H. B., 185, 186, 191, 193, 199, 201, 350, 356 Barris, R. W., 323 Barron, D. H., 1, 13, 38 Batini, C., 413, 425, 440 Baumgarten, R. von, 207, 221, 226, 309, 447 Baumgartner, G., 155, 212, 218,445 Bechterew, W. von, 88 Beck, E. C., 74, 369 Beek, E., 374-394 Bender, M. B., 220 Benoit, O., 421 Beritashvili, I. S., 340-348 Bessou, P., 16, 130 Birdsall, T. S., 200 Bimschein, H., 194, 196
Bishop, G. H., 38, 155, 173, 245, 254, 265, 269, 372 Bizzi, E., 454 Bless, E., 422 Bloch, V., 83, 84, 87, 88, 421, 440 Blundell, J., 155 Bonnet, V., 2, 13, 39, 116, 155, 266 Bonvallet, M., 83, 84, 86-92, 94, 129, 309, 316, 433, 440 Borenstein, P., 116, 120, 155, 207, 219, 221, 272, 309, 311 Borlone, M., 458 Bowersbuch, M. K., 76 Bowsher, D., 126, 127,474 Branch, C. L., 144, 444 Brazier, M. A. B., 349-373, 449 Bremer, F., 2, 13, 39, 90, 116, 155, 170, 245, 254, 266,272,288,291,309,317,370,407,422,444 Brinley, F. J., 269 Broadbent, D. E., 285 Brodal, A., 128, 302 Brookhart, J. M., 38-60, 259, 465 Brooks, V. B., 23, 97, 241 Brown, D., 401 Brown, K. T., 261 Bruner, J., 116, 119, 207,219, 221,294-324, 368 Brust-Carmona, H., 295 Buchwald, N. A., 433 Bullock, T. H., 226, 267, 269, 337, 467, 470, 473 Burgess, P. R., 47 Buser, P., 116, 118-120, 155, 207, 21 1, 219-222, 224, 272, 294-324, 368, 443, 447 BureS, J., 255, 294, 311, 312, 316 BureSova, O., 311, 312, 316 Butkhusi, S. M., 155, 166 Cadilhac, J., 424 Calvet, J., 266, 270 Candia, O., 405, 439 Caspers, H., 241, 259, 265 Cavaggioni, A., 184, 189-192, 195, 199,201 Chambers, W. W., 39, 75, 302, 303 Chang, H. T., 196, 241 Chang, J. J., 258, 288 Cherry, C., 285 Chistovich, A. M., 346 Chow, S. N., 254, 255 Ciganek, L., 377, 389, 390, 393 Claes, E., 301
476
AUTHOR INDEX
Clare, M. H., 155, 173, 244, 254, 269, 372 Clark, L. D., 422 Clark, W. F., 351 Clemente, C. D., 433 Clezy, J. K. A,, 127 Cobb, W. A., 377, 381, 389, 394 Cohen, M. I., 84,95 Colle, J., 301 Collins, W. F., 350 Communication Biophysics Group, 188, 351 Coombs, J. C., 59 Cooper, R., 397, 402 Cooper, S., 8 Cordeau, J. P., 434 Courjon, J., 283, 415, 421 Covian, M. R., 142 Cowan, W. M., 132 Coxe, W. S., 241-256 Cragg, B. G., 58 Creed, R. S., 6, 8 Creutzfeldt, O., 144,155,218,221,225,265,370, 436,438,445 Cullen, C., 144 Curtis, D. R., 50, 56, 449 Da Fonseca, J. S . , 207-231 Dawson, G . D., 377, 381, 394, 397 Dechaume, J., 419 Dell, P., 82-102, 155, 272, 284, 288, 291, 309, 315, 370,433,440 DeLorenzo, A. J., 39, 271 Dement, W., 283,407, 422, 424 Demetrescu, M., 155, 291 Dempsey, E. W., 131, 155, 173 Denny-Brown, D., 3, 27, 75 Denton, E. J., 201 Derbyshire, A. J., 407 Desmedt. J. E., 155, 316,444, 448, 451 De Vito, R., 128, 227 Didio, J., 262 Doane, B., 222,226,272, 274, 284, 297 Dobson, R. L., 297, 322 Do Carmo, R. J., 259 Dodt, E., 65 Dondey, M., 133,434 Donhogger, H., 422 Doty, R. W., 74, 222 Dow, R. S., 74 Drooglever Fortuyn, J., 132, 134 Dudel, J., 47 Duensing, F., 220,225 Dumont-Tyd, S., 84, 88, 98, 99, 137, 155, 272, 284, 288, 291, 309, 370 Dunlop, C. W., 401, 449 Eccles, J. C., 1-18, 23, 24, 39, 45, 48-50, 53, 56, 263-264 Eccles, R. M., 1, 3-6, 8, 10, 11, 13, 23 Eidelberg, E., 267
Encabo, H., 311 Enger, P. S., 241 Engstrom, H., 448 Euler, C. von, 258, 259 Evarts, E. V., 155, 276, 283 Ey, H., 422 Fadiga, E., 4346, 48, 50, 56, 58 Faidherbe, J., 374-394 Fanjul-Moles, M. L., 298 Fatt, P., 2, 48 Favale, E., 405, 421, 439, 441 Fernandez-Guardiola, A., 155, 436 Fessard, A., 2, 115-145, 266, 294, 295, 447-448, 466-474 Fifkova, E., 311, 312, 316, 317 Finer, B. L., 65-77 Finley, K. H., 155 Fischer-Williams, M., 155 Fitzhugh, R., 185, 186, 191, 193, 199 Forbes, A., 325,407 Franck, G., 374-394 Frank, K., 1, 6, 23, 47 French, J. D., 309, 350, 357 Fressy, J., 374-394 Freud, S., 422 Freygang, W. H., 267 Fromm, G. H., 265 Fuortes, M. G. F., I, 6, 23, 47, 297 Gasser, H. S., 1, 2 Gastaut, H., 294,297, 323, 369, 374-394 Gauthier, C., 288 Gautier, H., 94, 95, 155, 173 GavliEek, V., 329, 336 Gaze, R. M., 127 Gellhorn, E., 155 Gerard, R. W., 75 Gerber, C. J., 259 Gernandt, B. E., 301, 449 Gershuni, G. V., 346 Gerstein, G. L., 188 Gesell, R. A., 38 Getz, B., 127 Gidro-Frank, L., 76 Gillett, E., 120, 129, 207, 208,220, 221, 312, 350, 352 Gilman, S., 301 Girado, M., 270 Giussani, T., 405, 421,439, 441 Glees, P., 127 Gluck, H., 433 Goldring, S., 241-256, 259, 260, 265 Goldstein, M. H., 184, 189, 196 Goodman, N. R., 402 Gordon, G., 127 Gotch, 266 Graham, H. T., 1
AUTHOR INDEX
Granit, R., 3, 5, 23-34, 38,49-51, 165, 184-186, 190, 191, 193, 199, 200, 270 Grafstein, B., 255 Grastyhn, E., 88,422,424-428, 464 Green, D. M., 200, 207, 221, 259, 267 Griffin, D. R., 292 Grossman, R. G., 258 Grundfest, H., 39, 266, 270, 337, 449 Griisser, 0. J., 155, 207-209, 213, 214, 217, 218, 220, 221,224,225, 370 Griisser-Cornehls, U., 207-209, 213, 214, 217, 218, 221,224,225, 370 Griitzner, A., 155 Gualtierotti, T., 223 Gudden, B. von, 222 Gumnit, R. T., 258, 259 Guzrnhn-Flores, C., 155 Haase, J., 30, 32, 33, 322 Hagbarth, K.-E., 6, 65-77, 316 Hagiwara, S., 2 Hamlyn, L. H., 58 Hance, A. J.: 349 Hara, T., 405, 421, 441 Hare, W. K., 355 Harreveld, A. van, 267 Hartline, H. K., 23, 27 Harvey, R. J., 16 Hashiramoto, S., 309 Hassler, R., 120, 132, 208, 222, 309 Hecht, S., 200 Hendley, C. D., 441 Hendrix, C. E., 401, 422 Hernhndez-Peon, R., 75, 76, 145, 155, 165, 227, 272, 291,295, 309, 316,433 Heuser, G., 433 Hild, W., 258 Himwich, H., 355 Hirao, T., 184, 186, 192, 193, 196,442 Hirsch, J., 118 Hoffmann, P., 65 Holrnes, T. G., 241, 243, 259 Holmquist, B., 11, 75 Holst, E. von, 227 Horn, G., 155 Hovde, C. A., 132 Howard, S. Y.,254, 255 Huang, S. H., 247, 259 Hubbard, J. I., 5 Hubel, D. H.,214,270,421,424 Hugelin, A., 84, 89-92, 94, 95, 98, 99, 129, 309, 316,440,433 Hughes, J., 1 Hunt, C. C., 6, 53 Hunter, J., 297, 303, 315, 317, 323, 350 Huttenlocher, P. R., 421 I g w , A., 3 Imbert, M., 116, 207, 211, 220-222, 224, 315
477
Ingram, W. R., 323 Ingvar, D. H., 297, 303, 315, 317, 323, 350 Ito, M., 3 Jankowska, E., 116, 134, 135, 137 Jasper, H. H., 120, 144, 155,222, 226, 272-293, 297, 309, 312, 325, 351, 360, 465 Jassik-Gerschenfeld, D., 294-324 Jerva, M. J., 241, 243, 259 Johnson, T. N., 130, 132 Jouvet, M., 272,283,294,404,405,406424,425, 426,428, 434,436, 438, 440 Jung, R., 144, 155, 207-231, 264, 265, 270, 309, 436,445-447,452 Jutier, P., 116 Kadjaia, D. V., 155, 165-167, 169 Kado, R. T., 262 Kapp, H., 265 Karnp, A., 393 Kandel, E. R., 58, 59, 254, 255, 259, 267, 269 Kanno, Y., 219 Karrnos, G., 424428 Karplus, J. P., 126 Katz, B., 23, 48 Katz, R., 241-256 Katz, S. E., 422 Kawatami, M., 432 Kernpinsky, W. H., 260 Kendig, J. H., 243, 260 Kiang, N. Y.S., 188 Killam, K. F., 349 King, E. E., 360 Kitchell, R. L., 155 Kleitman, N., 422, 424 Knapp, H. D., 74 Knighton, R. S., 121 Koehler, B., 65 Koella, W. P., 155 Kohler, W., 258 Koketsu, K., 2, 11 Kolmodin, G. M., 53 Konorski, J., 75 Kogan, A. B., 433 Kopa, J., 464 Kornhuber, H. H., 207-231 Kostyuk, P. G., 1, 2, 6, 11-15 Koukkou, M., 436 Kozak, W., 6 Kozhevnikova, V. A., 346 Kreidl, A., 126 KrnjeviC, K., 11 Kruger, L., 121, 122, 126, 131, 350 Kubota, K., 38-60 Kuffler, S. W., 47, 185, 186, 191, 193, 197, 199, 255 Kugelberg, E., 65, 68, 73, 75, 76 Kuno, M., 23, 53 Kuypers, H. G., 129, 352, 416
478
AUTHOR INDEX
La Grutta, G., 155 Laporte, Y., 5, 16, 130 Lam, R. L., 260 Lambert, E. F., 407 Landau, W. M., 155, 173, 267 Landgren, S., 155, 221 Landis, C., 422 Lanoir, J., 436 Lauprecht, C. W., 433 Leio, A. A. P., 252, 258 Le Beau, J., 434 Lehniann, D., 221, 370, 436 Li, C. L., 144, 155, 254, 255 Libouban, S., 116 Liddell, E. G . T., 8, 9 Lissak, K., 422, 424-428, 463-465 Liu, C . N., 39, 302, 303, 315 Livingston, R. B., 75, 309, 31 1, 316 Lloyd, D. P. C., 2, 4-6, 302 Loeb, C., 440 Loewenstein, W. R., 23 Lonio, T., 222 Long, R. G., 155 Lorente de NO, R., 225 Lothrop, G . N., 155 Lourie, H., 243 Lundberg, A., 4-6, I I , 13 Macchi, G., 132 Machne, X., 40, 42, 117, 207, 221 Macht, M. B., 413, 439 Maclntosh, F. C., 266 MacKay, D. M., 227 MacNichol, E. F., 23 Madazlisz, I., 422 Magnes, J., 88, 440 Magni, F., I , 6, 8, 10, 413, 440 Magoun, H. W., 85, 155,309,325-327,350,355, 357, 408, 420, 432 Magladery, J. W., 65 Makarov, Y. A,, 329 Mallart,A., 125, 128, 130, 143, 315, 350, 360, 369 Malliani, A,, 454, 458 Malrno, R., 285 Mancia, M., 155, 207, 259, 272, 434 Mark, R. F., 6 Marriott, F. H. C., 200 Marseillan, R. F., 142 Marshall, W. H., 219, 222 Martin, A. R., 144, 444 Maruseva, A. M., 346 Maruyama, N., 219 Massion, J., 129, 301, 315, 316, 350 Matsunaga, M., 31 I Matthews, B. H. C., I , 2, 13, 16, 27, 28, 38, 184, 185, 407 Matzke, H. A,, 126 Maxwell, D. S., 267 McCranie, E. J., 76
McIntyre, A. K., 6 McKinley, W. A,, 350 McLardy, T., 130, 132, 352 McPherson, R., 200 Mechelsc, K., 207, 259, 316 Mehler, W. H., 88, 127, 130 Melzack, R., 74 Mettler, F. A., 132 Meulders, M., 129, 155, 272, 315, 316, 350 Meulen, J. P. van der, 30, 32, 33, 322 Meyer, 266 Michel, F., 283,408,415,419,421,424426,428, 436 Migliaro, E. F., 31 I Mikiten, T. M., 441 Miller, J., 336 Mingrino, S., 241-256 Mitchell, J. L., 351 Mittelstaedt, H., 227 Mollica, A,, 207, 221, 222, 227, 309 Monaco, P., 444, 448, 451 Moniava,E. S., 155, 157, 159, 161, 162, 165-167, 169 Monnier, M., 464 Morillo, A., 155 Morrell, R. M., 38, 124 Moruzzi, G., 74, 85, 88, 89, 155, 157, 207, 221, 227,250,259,296,298,309,31 I , 325,408,413, 428 440, 44 I , 453 Mountcastle,V. B., 121, 127, 130, 132, 144, 213, 22 I Mundy-Castle, A. C . , 402 Murphy, J. P., 155 Nakao, H., 155 Naniikawa, A,, 309, 31 I Naquet, R., 155, 317, 360,436 Narikashvili, S. P., 156-183 Nashold. B. S.. 132. I ” . Nauta, W. J. H., 129, 132, 352, 416, 454 Neff. W. D.. 258 Niebyl, P. H.,441 Oborin, P. E., 266 Ochs, S., 267 O’Connell, D. N., 258 Olds, J., 312 O’Leary, J. L., 241-27/, 350 Olsen, K. H., 351 Orrego, F., 208 Oscarsson, O., 5, 16 Oswalda-Cruz, E., I 16, I 17, 120, 208 Otani, T., 53, 55 Otsuka, R., 208 Paintal, A. S., 16 Palestini, M., 413, 424, 440 Parnia, M., 155, 173, 288 Passouant, P., 424
AUTHOR INDEX
Patton, H. D., 244, 245, 296, 302 Pavlov, J. P., 74 Pearlnian, A. L., 259 Pedersen, E., 65 Penfield, W. G., 278 Perl, W. R., 127, 130, 131 Phillips, C. G., 49-51, 270, 447 Phillis, 3. W., 449 Pinnio, L., 184, 188-1 90, 192, 194, 195-1 97, I99 Pinto-Hamuy, T., 294 Pirenne, M. H., 200, 201 Poggio, G. F., 121, 127, 130, 132 PolIey, E. H., 208 Pompeiano, O., 88, 429-432,433,440 Powell, T. P. S., 132, 144 Purpura, D. P., 266, 270, 337, 449 Quensel, F., 127 Rall, W., 50 Rambn y Cajal, S., 126 Ranson, S. W., 323, 355 Ratliff, F., 23, 27 Reeth, P. Ch. van, 155 Regis, H., 155 RBmond, A. G., 297, 322, 323,374-394 Rempel, B., 407 Renkin, B., 24, 26-28 Renshaw, B., 2, 3,23, 27 Rhines, R., 85, 420 Rhodes, J., 436 Rhoton, A., 259 Ricci, G., 222, 226, 259, 272, 274, 284, 297 Rice, S. O., 189 Rimbaud, L., 424 Ripoche, A,, 374 Rocha-Miranda, C., 116, 117, 120, 208 Roger, A., 294, 295, 315 Roig, .I.A,, 311 Roitbak, A,, 433 Rossi, G. F., 128, 207, 221, 308, 309, 404406, 413,421,425,428,434,439,440,441 Rothballer, A. B., 434 Rougeul, A., 116, 119-121, 133, 350 Rowland, V., 258,433 Rudoniin, P., 458 Rutledge, L. T., 29, 30, 38 Sacco, G., 405, 421, 440, 441 Sala y Pons, C., 39, 40 Santibahez, G., 155, 272 Sasaki, K., 309, 31 I Saur, G., 207,208,213,214,221,224 Sawyer, C. H., 432 Schaefer, K. P., 220, 225 Scheibel A. B., 126, 207, 221 Scheibel, M. E., 126, 128, 207, 221 Schemer, H., 165, 266, 270, 272 Schimert, J. S., 450
479
Schindler, W. J., 267 Schmidt, R. F., 1, 2, 7, 1G15 Schoen, L., 445 Schoolman, A,, 155 Schubert, G., 192 Scott, T. H., 74 Sechenov, I. M., 340 Segundo, J. P., 117, 207, 221, 309, 311, 316, 317 Shanzer, S., 220, 309 Shen, E., 315 Sherrington, S. S., 2, 8, 9, 38, 468 Shields, J. R., 241, 243, 259 Sholl, I).A,, 59 Shumilina, A. I., 329 Siebert, W. M., 188, 351 Siminoff, R., 126 Sindberg, R. M., 219, 294-324 Sloan, N., 312 Smith, C., 374-394 Soderberg, U., 214 Soler, J., 126 Solodovnikov, V. V., 188 Spehlmann, R., 218, 221, 370,445 Spencer, W. A,, 58, 59, 254, 255, 267, 269 Sperry, R. W., 76 Spiegel, E. A., 31 1 Starzl, T. E., 309 Stefens, R., 132, 143 Steiner, J., 6 Stennett, R. C., 285 Steriade, M., 155, 291 Sterman, M. B., 433 Stoll, J., 309, 312 Storm van Leeuwen, W., 393 Stoupel, N., 155, 170, 272, 288, 291, 370 Straws, H., 309 Stunipf, C., 267 Sem-Jacobsen, C. W., 393 Suzuki, H., 243 Sweet, W. H., 127 Swett, J. E., 431, 433 Szibo, I., 464 Szekely, E. G., 311 Szentagothai, J., 220, 225 Taira, N., 243 Takenchi, A., 2 Takenchi, N., 2 Talbot, S. A., 219, 222 Tanner, W. P., 200 Tasaki, I., 2, 208, 258, 261 Taylor, C., 309 Taylor, J. L., 258 Teasdall, R. E., 65 Terzuolo, C., 116, 309, 317, 449 Thesleff, F., 56 Thompson, R. F., 219, 297 Tiberin, P., 436 Timo-Jaria, C., 142
480 Tissot, R., 464 Tonnies, J. F., 11, 265, 446, 452 Torvik, A., 126 Tshirgi, R. D., 258 Tweel, L. H. van der, 393 Uhr, L., 336 Van Hof, W. M., 390, 393 Vastola, E. F., 222, 258 Velasco, M., 165, 272 Veringa, F. T. H., 393 Verzeano, M., 350, 357 Veszi, J., 47 Walberg, F., 302 Walker, A. E., 120 Wall, P. D., 16, 53, 127, 297, 322, 323 Walter, D. O., 401, 422 Walter, W. Grey, 265, 395-403, 466 Walzl, E. M., 213 Warren, W. J., 396 Washizu, Y., 47 Watkins, J. C., 449 Wegener, J., 258 Weinberg, D., 401
AUTHOR INDEX
Weinstein, W., 243, 260 Weiss, T., 311, 312, 316, 317 Wendt, R., 117, 132 Werre, P., 374-394 West, C. D., 422 White, J. C., 127 Whitlock, D. G., 127, 130-132 Widen, L., 214, 316 Wikler, A., 407 William, K. F., 449 Willis, W. D., 1, 5-7, 10, 16 Winsbury, G. J., 5 Wilson, V. J., 23, 47 Witkovsky, P., 126 Wohlbarsht, M. L., 23 Woodward, P. M., 200, 201 Woolsey, C. N., 208, 219 Wyers, E. J., 433 Yoshii, N., 124 Zanchetti, A., 155, 173, 207, 221, 288, 308, 404, 440,454-461 Zang, Z., 391 Zatchinayeva, I. A., 329
Subject Index
Ablation, and chlorpromazine, 421 of cerebellum, 41 1 retinofugal, 184 of cerebral cortex, 129 decrease of, 194 of corpus striatum, 129 spontaneous, of neocortex, 41 1 infra-slow, 258 of neuraxis, 41 1 interstimulus, inhibition of, 277 Accommodation, tonic, and dorsal root reflex, 6 concept of, 186 role of, 55 level, and light intensity, 192, 193, 195 Act, origin of, 185 behavioral, of man, 340 retinal, 186 Action, Acuity, facilitatory, 155 visual, 200 thalamic non-specific, cortical reaction to, 159 Adaptability, Action potential, functional, 73 and extracellular unit recordings, 138 Adaptation, dendritic, 58 dark, 185, 200 of peripheral nerves, 130 light, 185 Activation, Afferent fibers, “arousal”, 223 cutaneous, in spinal cord, 11 of cortex, 223 group Ib, 10, 15 of cortical neurons, 224 inflow from muscles, 428 ascending, Afferents, homogeneous, 326 auditory, 116 influences of, 337 extralemniscal, 143 of different biological origins, 337 flexor reflex, 6, 11 reticular system, 408, 41 9 interoceptive, of vagus nerve, 88 specific character of, 325 labyrinthine, 88 at onset and cessation of polarization, 213 muscular, 30 by non-physiological stimulation, 223 sensory, somatic, 124 conditioned defence, 330 somatic, 116 delayed (“unspecific”), 209 in thalamus, 115 d’emblee, 230 spinal, somesthetic projections of, 1 15 effect, 278 visual, 116 extrapyramidal, 300 After-discharge, of extra-lemniscal systems, 142 cortical, 241 of lernniscal systems, 142 hippocampaI, 277 of monkey, 283 After-effect, pain, under urethane narcosis, 326 facilitatory, with EEG desynchronization, I80 pyramidal, 300 reticular facilitatory, 169 reticular, 97 After-hyperpolarization, supraliminal, excessive massive, 285 prolonged, 45 Activity, After-potential, neocortical, rhinencephalic control of, 421 positive, 1 of pretectal fibers, 355 prolonged, 45 of retinal ganglion cells, 185 Alerting, rapid control, 414 behavioral responses, 225 and chlorpromazine, 418 Alertness, rhythmic limbic, 418 effect on evoked potential, 145
482
SUBJECT INDEX
State Of, 123 Amplitude, fluctuation of, 123 for evoked potential, I37 reduction of, in EEG, 429 variations, 123 Amygdala, relay for, 132 Aniygdaloid nucleus, electrical stimulation of, 272 with hippocampal after-discharge, 277 Analgesia, hypnotic, 70 subjective, 70 to hypnotic suggestion, 72 Analysis, computational, massive, 360 micro-electrode, of changes in surface occipital evoked potentials during conditioning, 272 Anoxia, cortical, 91 Area, acoustic, 305, 306 association, 21 I, 309 auditory, 210 primary, 303, 304, 307 sensory, 300 specific cortical, 293 skin, effect of stimuli, 66 somatic, 21 1 vestibular, 21 I visual, 210 Arousal, 325 behavioral, 427 responses, 225 reticular, of sensory stimulation and, 405, 410 cortical, associated, 90 generalized, 88 short-lasting, 87 reticular, effects, 88 threshold for, 429 Association area, 21 I, 214 Associative territories, I I6 Attention, perceptual, I4 sensory, 227 Attenuation, progressive, 398 unconditional non-specific, 400 Auditory cortex, anatomical area I, 210 Autonomic nervous system, cerebral niechanisms during conditioning, 395-404 conditional response, 400
Axon, motor, 39 Barbiturate, anesthesia, and neurograni of superficial radial nerve, 431 Bechterew-nucleus, 41 5 Behavior, adaptive defence, 65 alimentary, of dogs, 342 arousal, 427 Blood pressure, at beginning of sleep, 404 Brain function, general hypothesis of, 229 Brain steni, conduction over, 226 heterotopic fields of, 126 niultisensory fields of, 126 reticular stimulation of, 155 stimulation of, 417 successive sections of, 41 1 Cancellation, of reafferent, 228 Capacity, avoidance, 67, 74 integrative, of associative cortex, 1 1 S Capillaries, niicro-, 30 Cardiazol, visual responses under, 303 Cataplexia, 41 3 Caudate nucleus, 207 Cells, internuncial, 39, 51 rate of firing of, 30 sensory stellate, 341 Centre median, 311, 349, 352, 360 of thalamus, 364 stimulation, 464 Cerebral cortex, ablation of, 129 activated state of, 327 after-discharge, 247 association, 21 3 association areas, 115, 122, 132 integrative capacity of, 113 as feed-back controller, 90 auditory response of, in cat, 277 cells of, level of, 337 direct current potentials of, 258 efferent topography, 300, 302 evoked potentials, 279 in reticular homeostasis, 89 neurons in, 219 piriform, I32 potentials evoked by photic stimulation in, 374-395
SUBJECT INDEX
prepiriform, 117 primary specific areas, 294-324 rapid activity of, 414 and chlorproniazine, 418, 422 sensory, neurons in, excitability of, 174 peripheral stimulation of, responses of, 179 unspecific impulses and, 155-180 sornatosensory, potentials from, 162 specific visual projection, 362 stimulation, repetitive, 247 subcortical systems, interaction, 335 suprasylvian, 242 vestibular, 213 Chain, conditioned, reflex, 345 Chiasm, optic, stimuli to, 156 Chloralose, effect on potentials, 117 Chlorpromazine, and ascending activation, 329 and rapid cortical activity, 418, 422 Circuits, local inhibitory, 97 mesencephalo-bulbo-mesencephalic, intrareticular, 88 reticulo-bulbo-reticular, feed-back, 89 reticulo-cortico-reticular equilibrium of, 94 regulatory, 89 Claustrum, telencephalic structure of, 117 Cochlear nucleus, response, diminution of, 421 Coefficient, regression, 25 Colliculus, superior, 32, 355 Component, negative-positive, 35 1 Complex, biological, of substructural structures, 328 centre mCdian-parafascicular, 129 of surface negative waves, 355 Computer, sequential, of contingency, 396 to detect evoked responses, 349 Condition, non-transient, 187 Concept, of tonic activity, 186 Conditioning, and autonomic mechanisms, 395-404 and changes in surface occipital evoked potentials, 274 and non-specific cerebral responses, 395-404 and specific cerebral responses, 395-404 behavioral responses, 225
48 3
cerebellar, 32 stimuli, 156 thalamic, stimulation, 160 Conduction, decreniental, 261 over brain stem, 220 Connections, axosoniatic, 49 for non-specific fibers, 177 intracortical (interareal), 221 monosynaptic, 43 subcortical, 221 thalamo-striatal, 132 Contingency, sequential computer of, 396 Contribution, associated, absence of, 226 Convergence, bisensory, 21 5 bisensory type 1, 2 I5 in subcortical structures, 229 mixed with type 1 and type 11 responses, 224 multimodal, 229 niultisensory, 122, 207, 219, 229 of gustatory impulses from tongue, 221 of impulses, 1 I5 of several modalities, 215 of tactile impulses from tongue, 221 of thermal impulses from tongue, 221 of type I responses with type I[ responses, 216 reticulo-thalamic, experiments on, 221 thalamic structures of, somatic sensory afferents of, 124 trisensory, 21 5 trisensory type I, 216 vestibulo-acoustic, 21 5 visual, experiments on, 221 visuo-acoustic, 21 5 visuo-vestibular, 21 5 Convolution, superior, of cat’s cortex, 207 Coordination, optovestibular, 220 sensorimotor, 229 Correlation, between EEG and somatovegetativity, 407 Corpus striatum, ablation of, 129 Current, auditory cortical, 258 depolarizing, 24 quasi-steady visual, 258 Curves, of probability distribution functions, 187 Dark, adaptation, 185, 200 level of, 190 discharge, retinal, 198
484
SUBJECT INDEX
Decrease, of retina, 184-206 in variability, 282 retinal, cortical effects of, 184 Decision units, 226 transient, of trkensory neurones, 226 Deflection, type “pulsing”, 190 negative, 351 Discrimination, of alternate positive and negative polarity, 378 signal-to-noise, 185 Dendrites, Distraction, of hippocampal pyramids, 58 auditory, 279 Depolarization, effects on evoked potentials, 272 degree of, 52 visual, 279 excitatory, 47 Distribution, increments of, 56 topographical, on scalp of PEM’s, 390 membrane, 49,57 postsynaptic, 51 Eel presynaptic, slow time course, 6 optic nerve of, 185 summated, 50 Effect, threshold, 59 activation, 278 wave of, 53 bimodal, 276 Depression, facilitatory reticular, 90 consecutive, EEG desynchronization, 159 nonspecific upon electrical responses in senof cortical responses, 177 sory systems, 276 of responses, I65 reticular arousal, 88 of retinofugal discharge, I85 Efferents, spreading, Lefo’s, 258 nonspecific thalamic, 132 Desynchronization, thalamo-striatal, 132 conditioned defensive activation, 330 Electrode, cortical, 90 intracellular, 11, 24 EEG, 174 Electrodermogram, consecutive depression, 159 of plantar foodpad of cat, 83 of cortical electrical activity, 326 Electroencephalogram, Diminution, desynchronization, 174 of cochlear response, 421 with facilitatory after-effect, 180 of reticular evoked response, 421 synchronized, reduction of amplitude in, 429 Discharge, Elimination, after, 241 of neocortex, 304 all-or-none, 245 Enckphale isole, baroreceptor, 94 multimodality afferents in, 207 cortical, 247 with artificial respiration, 208 cortico-motor, facilitation of, 102 Ending, efferent, 299 presynaptic, 39 extrapyramidal, 304 Energy, monosynaptic, 13 average, of gross electrodes potential, 189 motor neuronal, 44 Environment, of ganglion cells, image of, 341 changes, 187 real, 341 statistical treatment of, 186 EPSP, of signalling neurons, 226 amplitude of, 52, 55 peripheral, 295, 297, 300 dendritically induced, 59 pyramidal, 295, 296, 302, 307 depolarization, 52 reticular, 94 dorsal root, 56 reticular cell, 90 electrotonic propagation of, 58 retinal dark, I98 monosynaptic, 6, 13, 43 retinofugal, 187 size of, 49 depression of, 187 somatically induced, 59 rhythm, and reticular neurons, 97 summation of, 49, 50 sinocarotid, 94 Excitability, tonic, changes, in motoneuron pool, 75 during continuous illumination, 190 level, during reticular stimulation, 174 of dark adapted eye, 189 of cortical neurons, 162
SUBJECT INDEX
of neuronal elements, 173 of neurons in sensory cortex, 174 Excitation, antidroniic, 45 axodendritic, 58 axosomatic, 50 dorsal root, 43, 51, 56 monosynaptic, 57 of interneurons, 177 presynaptic, 29, 44 recurrent, 23 synaptic, 13 Eye, dark adapted, reticular activity of, 178 tonic discharge of, 189
485
Frequency, critical flicker, 217 of stimulation, 175 Function, brain, general hypothesis of, 229 information carrying, 349 integrative, 285 Ganglion cells, chain receptor-bipolar-, 199 discharge, statistical treatment of, 186 retinal, activity of, 185 Geniculate body, lateral, 185 stimuli to, 156 Golgi tendon organs, of muscle, 5 Gyrus, anterior, 309 ectosylvian, 342 anterior third of, 342 lateral, 309 posterior, ectosylvian, 230 suprasylvian, 230 suprasylvian, 123, 222 potentials from, 123
Facilitation, a t cortical level, 179 hypothalamic, of local strychnine discharges, 175 in thalaniic nucleus, I72 nonspecific, 173 of cortical neurons, 178 of cortical responses, 161, 174 of cortical primary responses, 164 of test response, 32 of thalaniic responses, 164 reticular, 173 Hippocampus, 207 at cortical level, 176 Homeostasis, at thalamic level, 176 reticular, of cortical responses, 168, 175 role of cerebral cortex in, 89 Feed-back, role of extracerebral feed-back circuits, 93 circuits, extracerebral, in reticular homeorole of medulla in, 85 stasis, 93 Hyperalgesia, controller, cerebral cortex as, 90 hypnotic, 70 effect on reticular formation, 92 subjective, 70 information, convey of, 75 Hypercapnia, negative, 2 effect of, 94 recurrent, 31 Hyperpolariza tion, reticdo-cortico-reticular, 91 summating, 50 system, synaptically induced, 47 models for, 228 Hypocapnia, negative, 13 effect of, 94 Fiber, Hypothalamus, a-fibers, of cutaneous origin, 130 posterior, 410 ascending tract, 13 stimulation of, 155 excitatory, presynaptic, 2 Hypothesis, lateral column, 44 general, of brain function, 229 axosomatic synapses, 57 Hypoxia, group 11, from muscle spindles, 130 at cortical level, 94 nonspecific, at reticular level, 94 on dendrites, 177 induced, 90 synaptic connections of, 176 presynaptic, 44 Image, pretectal, 355 of environment, 341 primary trigeminal, 126 of location, 345 specific, Impulse, synaptic connections of, 176 afferent,
486
SUBJECT INDEX
action on cortical neurons, 138 to primary thalamic relay nucleus, 125 convergence of, 115 from muscle, group La, 5 from tongue, 221 nonspecific, 155-1 80 ascending, 227 of cortical evoked reaction, 155 presynaptic, 59 spino-reticulo-thalamic, 142 Inflow, afferent, from muscles, 428 somatic, 350 Influence, cortico-tegmental, 308 cortico-thalamic, 31 I internuncial, 50 Information, multisensory, filtering processes, 227 in isocortex, 207 Inhibition, 444-452 early, 72 central, in spinal batracians, 19 initial, 65 lateral, 27, 199 mediate lateral, I99 mutual, 29 of cortical neurons, 178 of interstimulus “spontaneous” activity, 277 postsynaptic, 1-22 presynaptic, 1-22 action of strychnine, 7 rebound, 66 reciprocal, suppression of, 100 recurrent, 23-35, 47 Input, extralemniscal, I4 I level, 316 Integration, central, of sensory messages, I 15 inter neuronal, 472 multisensory, 1 15, 220 sensorimotor, 220, 295 thalaniic, and telencephalon, 1 15-148 visuo-vestibular, 228 Intensity, of light, and level of tonic activity, 192 Interaction, between lemniscal and extralemniscal systems, 143 cortical-subcortical, 335 intermodal, 224 interspatial and intermodality, 221 multimodality, 217, 335 of vestibular type 11, 217 of visual type I, 217 Interneuron,
and ischemia, 199 collateral branches of, 1 D-type, synaptic connections, 12 of corticomotor path, 101 of frog, 53 of segmental path, 101 of segmental polysynaptic paths, inhibition of, 102 selective inhibition of, 99 supraspinal control of, 75 Interstimulus, spontaneous “activity”, inhibition of, 277 I-waves, 244 IPSP, 47 Irradiation, frontal, 303 Ischemia, of midpontine pretrigeminal preparation, 197 retinal, 194, 196 Isocortex, multisensory information in, 207 Isolation, of geniculate nucleus from thalaniic structure, 222 Level, anteroposterior, 309 cortical, 74 facilitation of, 179 reticular facilitation of, 176 of cortical cells, 337 of input, 316 pontine, 3 17 thalaniic, reticular facilitation at, I76 Light, adaptation, 185 flashes, responses evoked by, 164 Limbic system, activity, rhythmic, and chlorproniazine, 418 Litnitlus, eye, 23 optic nerve fibers, 29 LOOP, dynamic properties of, 92 reticulo-cortico-reticular,regulatory action of, 90 Mechanism, autonomic, 395404 control, extracerebral, 97 intracerebral, 97 homeostatic, 96 inhibitory, 196 neurophysiological, underlying hypnotic phenomena, 76 of realization of ascending activating influences, 337 preselective, 227
SUBJECT INDEX
servo-, 90 sleep, 404-406 spinal defence, 65 subcortical, 308 Measurements, statistical, in nervous system, 188 Medial lemniscus system, bulbar inhibitory, 86 relation with activating system, 86 Medulla oblongata, role in reticular homeostasis, 85 Membrane potential, 44,45, 59 postsynaptic changes, 55 Membranes, dendritic, 58 inexcitability of, 39 depolarization, 49, 67 postsynaptic, 2 desensitization of, 56 somatic, 58, 59 Mesencephalon, reticular formation spindle activity, 408 reticular region of, 309 total sections of, 41 1 Message, corollary, 228 Milieu, interieur, homeostatic regulations of, 423 Modalities, convergence of, 21 5 Motoneurons, accommodative properties of, in toad, 55 and Renshaw cells, 2 dendrites of, 39, 43 direct reticular facilitation of, 99 discharge of, 51, 83 firing rate of, 24, 26 frog, 45 inhibition of, 2 integrative function of, 38-61 membrane of, 23 pool, excitability changes in, 76 stabilization of, 23 tonic, 23 Motor activity, behavior, 278 Motor area, 21 I Movements, instrumental, 347 voluntary, in higher vertebrates, 340 Muscle, afferent inflow of, 428 flexor, 67 spindle, 23 annulospiral endings, 4 tension, 16 tone, disappearance in PP, 425, 428 tonic, 3 Negativity,
slow, 242 Neocortex, elimination of, 304 Neodecortication, total, 41 1 Nerve, afferent, stimulation of, 161 femoral, 429 gastrocnemius, 32, 429 musculocutaneous, 429 peripheral, action potential, 130 stimulation of, and EEG, 429433 saphenous, 429 sciatic, 429 superficial radial, 429, 43 I trigeminal, principal nucleus of, 126 spinal nucleus of, 126 Nerve cells, dendritic expansion of, 39 depression of and nembutal, 197 Neurograni, of hamstring, 43 1 of superficial radial nerve, 431 Neuronal responses, to sense modalities, 212 Neurons, cortical, 137 afferent impulses of, action on, 138 cerebellar, 39 cerebral, 39 excitability of, 162 facilitation of, 178 influence of reticular formation, 173 influence of thalaniic structures, 173 inhibition of, 178 recruiting of, 173 extra-lemniscal representation to, 138 in sensory cortex, excitability of, 174 internuncial, excitation of, 177 in visual cortex, I 7 8 leniniscal representation to, 138 polysensory, in cat, 221 pyramidal, functional state of, 177 refractoriness of, 179 reticular, discharge rhythm, 97 signalling, 226 stellate, non-sensory, 341 trisensory, 218, 224, 226 Nucleus, amygdaloid, 1 17 electrical stimulation of, 272 caudate, 117, 122 responses of, 125 synapses in, 127 centralis superior of Bechterew, 415 convergent, intrathalamic organization of, 130
487
488
SUBJECT INDEX
of thalamus, amplitude variations, 123 gigantocellular, 128, 414 gracilis, 16 synapses in, 127 interpeduncularis, 410, 415 intralaminar, 127 principal, of trigeminal nerve, 126 reticularis, 127 gigantocellular is, 128 pontis oralis, 412 spinal, of trigeminal nerve, 126 thalamic, Facilitation of, 172 nonspecific, 127, I56 responses of, 122 relay, primary, afferent impulses to, 125 Nystagmus, vestibular, discharge and rhythm of, 214 Occlusion, between visual and vestibular responses, 21 7 of evoked potential, 136 Onset, of augmenting responses, 173 of labyrinthine polarization, 217 Organization, intrathalamic, of system of convergent nuclei, I30 synaptic, 38 Orientation, reaction, elaboration of, 309 Origin, cutaneous, u-fibers of, 130 of tonic activity, 185 of voluntary movements in higher vertebrates, 340 output, overall level, decrease of, 201 overall tonic, 198 tonic retinal, 201 Pacemaker, of motoneuron, 38 Parameters, macroscopic, time-independency of, 187 Path, final common, 82 lemniscal, 126 multisynaptic, 259 oligosynaptic, 259 Pathway, afferent, electrical stimulation of, 173 antero-lateral, 125 ascending, of spinal cord, 124 internuncial, 54 monosynaptic axodendritic, 53 spinal ascending, 75 visual, stimulation of, 179 Pattern,
of firing of individual units, 283 sensory-motor, 74 P.E.M., mean evoked potential, 376-394 Phase, second inhibitory, 72 waxing, of “recruiting” response, 161 Phasotron, 374 Phenomenon, electrotonic, 38 heterosensory reinforcement, 298 hypnotic, neurophysiological mechanisms underlying, 76 release, 336 somatovegetative, 407 Photostimulation, intermittent, response evoked by, 388 low frequency, in man, 376 Picrotoxin, action on pre- and postsynaptic intribition, 21 Plethysmogram, of foreleg, 408 Polarization, contralateral, 213 ipsilateral, 213 labyrinthine, 212, 217, 223 cessation of, 213, 217 onset of, 217 Pons, depressed function of, 428 reticular formation, spindle activity, 408 sleep spindles in, 428 Positivity, after-, 242 Postsynaptic potentials, electrotonically propagated, 40 inhibitory, 47 Posture, maintenance of, 3 Potential, antidroniic, 45 associative, 132 cerebral, evoked by photic stimulation, 374395 cortical evoked, 279 dendritic, 261 evoked by direct stimulation, 177 diphasic, 254 direct current, of cortex, 258-271 dorsal root, 11 evoked, 131 associative, 132 by stimulation of limbs, 141 effect of alertness on, 145 in visual cortex, 165 occlusion of, 136 primary, abolition of, 145 amplitude for, 137
SUBJECT INDEX
attenuation of, 145 from somatosensory cortex, 162 from suprasylvian gyrus, 162 generator, 11 geniculate, 279 intracellular, 43, 44 occipital evoked, decremental reduction in, 173 during conditioning, 274 positive slow, 1 primary, 242 slow, change of, 23 steady, 248 surface negative, 173 synaptic excitatory, 138 Preparation, diencephalic, 86 “enckphale isole”, 83, 88 nerve-muscle, 27 Pressure, arterial, on stimulation of reticular formation, 86 Principle, of “Reafferenz”, 228 “stimulus-responses”, 344 Probability, distribution functions, 189 curves of, 187 Process, in tracortical, inhibitory, 145 occlusive, 145 optokinetic, 228 pupillary, 165 retinal, 165 Projection, corticospinal, 82 extrastriate of visual neuronal responses, 222 from vestibular to reticular neurons, 225 non-discriminative, of extralemniscal system, 132 non-primary, telencephalic, 115 of extralemniscal system, 135 of lemniscal system, 135 of thalamus, to telencephalon, 131 reticular, synaptic organization, 85 somatotopic, of discriminative somatic sensitivity, 132 somesthetic, evoked by spinal afferents, 115 spino-reticular, 92 thalamo-telencephalic, of extralemniscal system, 115 to association areas, 141 to striatum, 132 Propagation, electrotonic, 255 Property, universal and homogeneous, of ascending acti-
489
vation, 326 Putamen, 117, 207 Pyramid, and muscle, associated events, 242 medullary response, 242, 246 Pyramidal tract, at pontine level, 317 discharge, 302 Reaction, avoidance, 72 cerebral, 75 cortical, to thalamic nonspecific action, 159 cortical evoked, nonspecific impulses of, 55 defence, 74,75 nociceptive, 75 time, 70 Reactivity, extrapyramida 1, 303 pyramidal, 303 sensory, of integrative subcortical struct res, 318 Receptors, auditory, 116 excitation of, physiological limits, 223 spindle, 130 static, 16 tendon, 130 visual, 116 Recording, electromyographic, 65 extracellular, 139 Recruiting response, elicited by stimulating nucleus centralis medialis, 163 waxing phase of, 161 Recruitment, extensive, after lateral column stimulation, 40 of cortical neurons, 173 Reduction, decremental, in occipital evoked potentials, 273 Redundancy, and von Holst’s mechanism, 227 multimodal, 227 Reflex, abdominal skin, 65 arc, spinal, 73 spinal potysynaptic, 75 conditioned chain, 345 dorsal root, 6 extension, 65 flexion, 67 monosynaptic, 2 attenuation of, 83 facilitation of, 94 masseter, during reticular stimulation, 83 muscle stretch, 8
490
SUBJECT INDEX
nociceptive, 65-77 pathways, corticospinal, 98 polysynaptic, 98 psychogalvanic, 89 after novocaine injection, 87 on stimulation of reticular formation, 86 pupillary, 355 spinal, 75 amplitude of, 70 latency of, 70 stretch, 8, 30 subcortical centers, 75 withdrawal, 65-77 Refractoriness, of cortical neurons, 179 Reinforcement, internuncial, 53, 54 phenomena, heterosensory, 297 Relationship, input-output, 38 Relay, integrative, 139 mesencephalic, 140 specific thalamic, 143 synaptic, 14, 16 thalamic, 140 Renshaw cells, 2, 27, 29, 32 facilitation of, 33 Representation, extralemniscal, to neuron, 138 lemniscal, to neuron, I38 Response, amplitude fluctuations of, 123 antidromic, 47 augmenting, 173 development of, 175 onset of, 173 autonomic, attenuation of, 88 sensation of pain, 65 avoidance, classical conditioned, 74, 75 behavioral, 225 bilateral, 317 cochlear, diminution of, 421 conditional autonomic, 400 conditioned, 283 electro-cortical, 123 parallel in formation of, 123 contingent, 395 cortical, 141, 241-257 attenuation of 88 depression of, 177 direct, and pyramidal activity, 250 evoked by stimulating of afferent pathways, 177 facilitation of, 161, 174 primary, 155 facilitation of, 164
reticular facilitation of, 168 delayed, type I i activation, 230 depression of, 165 diphasic, 387 electrical, nonspecific effects upon, in sensory systems, 272-293 electrodermal, 86 electronic average, indicator, 396 ecoked by flashes, 390 evoked by intermittent photic stimulation, 388 evoked by light flashes, 164 graded, 261 immediate specific, 209 immediate type I, 230 inhibitory, 65 initial brief transient, 209 in visual cortex, 171, 175 reticular activation of, 171 medullary pyramids, 242, 246 multisensory, of cat’s brain, 116 nociceptive, 76 nonspecific fronto-vertical, 397, 398 nonspecific to visual and auditory stimuli, 397 of auditory cortex in cat, 277 of caudate nucleus, 125 of nonspecific thalamic nuclei, 122 of optic tract, 165 of sensory cortex, 155-1 83 orthodromic, 48, 51 primary, of sensory cortex, 179 recruiting, 131, 163 waning phase of, 157 waxing phase of, 157 salivary, conditioned, 74 short-lasting positive-negative, 3 13 sinusoidal, 387 specific and nonspecific cerebral, during conditioning, 395404 specific and parietal nonspecific, 397 specific occipital, 397 stability of,388 startle, sensation of pain, 65 subcortical, 175 surface positive, 377 thalamic, I63 facilitation of, 164 to acoustic stimuli, 212 to optic stimuli, 212 to vestibular stimuli, 212 type 1 and 11 of different modalities, 224 type 1 vestibular, 213 type 11 vestibular, 214 visual, 303 withdrawal, 72, 76 Reticular, activating system (RAS), specificity, 45446I Reticular formation, 32, 126, 220, 348 and autonomic activity, 85 and somatic activity, 85
SUBJECT INDEX
brain stem, stimulation of, 155 bulbar, 88 critical reactivity, 82-102 discharges, sinocarotid baroreceptor, 94 evoked response, diminution of, 421 facilitation of, 82-102 feed-back effect on, 92 homeostasis and, 82-102 homeostatic role, 470 inhibition of, 82-102 inhibitory pontobulbar, 420 mesencephalic, 86, 88, 128, 129 projections from, 88 midbrain, 283, 353 physiology, 82-102 stimulation, 427 arterial pressure, 46 electrodermal response, 86 psychogalvanic reflex, 86 Retina, spontaneous activity, 184-206 tonic discharge, central effects of, 184-206 Reverberation, sensorimotor, aspects of, 294-324 Root, dorsal, 40 ventral, 40, 47, 317 Section, retropontine, 413 Sensation, pain, accompanied by startle responses, 65 accompanied by autonomic responses, 65 Sensitivity, discriminative somatic, somatotopic projections of, 132 somesthetic, modalities of, 144 Shock, antidromic, 32 lateral geniculate, 196 repetitive, 301 Signal, -to-noise discrimination, 185 -to-noise ratio, 194, 200, 276, 349 Signalling, neurons, 226 of steady illumination, 199 Skin, stimulus, 65-77 Sleep, and EEG-pathways, 415 archeo-, 419 blood pressure at beginning of, 404 cerebral, 413 deep, characterization of, 406 definition, 415 mechanisms of, 404406 paradoxical phase and disappearance of mus-
49 I
cle tone, 425 paradoxical phase of, in cat, 424-429 physiological, and EEG, 408 rhombencephalic phase of, 406424 signs of, in cat, 404-406 slow, 408 spindles, in pons, 428 telencephalic phase of, 419 -wakefulness rhythm, 404 Spinal cord 1-18 antero-lateral pathways of, 125 grey matter of, 127 input into, 16 isolated, frog, 44 of cat, 53 of decerebrate animals, 33 of frog, 39 pathways, ascending, 124 posterior horns of, 127 postsynaptic inhibition, 1-18 presynaptic inhibition, 1-18 Spread, electronic, 261 Stimulation, antidromic, 23, 26, 48 auditory, 122 caloric, 2 I5 caudate, on visual neurons, 218 cerebellar, 32 conditioning, 160 direct, dendritic potentials evoked by, 177 electrical, 277, 100 direct, 156 of afferent pathways, 173 of amygdaloid nucleus, 272 of mesencephalic reticular formation, I65 of visual structures, 179 of visual tract, 175 frequency of, 175 lateral column, 52 low frequency, photic, in man, 376 mesencephalic reticular, 100 of afferent nerves, 161 of brain stem, 41 7 reticular formation, 155 of hypothalamus, 155 of limbs, potentials evoked by, 141 of medial thalamus, 155 of mesencephalic reticular formation, 156 of reticular formation, 427 of thalamic relay nucleus, 161 of visual pathway, 176, 179 optic, excitability level during, 174 of afferent pathways, cortical responses, 173 of mesencephalic reticular formation, 165 of visual tract, 175 peripheral, nerve and EEG, 429-433
492
SUBJeCT INDEX
responses of sensory cortex, 179 photic, in man, 374 repetitive, 258 habituation, 68 reticular, experiments with, 156 prolonged mesencephalic, 82 somesthetic, 298 supraspinal, 3 1 tactile, of tongue, 100 thalamic nonspecific, 156 thermic, 214 vestibular, 209 visual, 122, 208, 279 Stimulus, acoustic, 207, 209, 294 hypothalamic facilitation of, 175 auditory, 295 in anterior region, 397 response principle, 344 conditioning, 156, 161, 273 multimodality, 218 multisensory, interrelation of, 208 non-physiological, activation by, 223 recticular, psychogalvanic response, 85 skin, 65-77 teleceptive, 294 testing, 156 unconditional, 397 vestibular, 207 visual, 207, 294 specific response in occiput, 397 Striatum, projections to, 132 Structure, cerebellar, 74 convergent, telencephalic, 123 thalamic, 123 midbrain, 74 reticular, 328 subcortical, 328 integrative, 318 relation to multisensory convergence, 219 synaptic, 59 telencephalic, of cat, 116 thalamic, nonspecific, 155 visual, electric stimulation, 179 Strychnine, action on pre- and postsynaptic inhibition, 7, 21 discharges, by acoustic stimuli, hypothalamic facilitation of, 175 Suggestion, hypnotic, 70 subjective analgesia to, 72 Summation, discontinuous, 377 continuous, 377
Synapses, arrangement of, 59 axodendritic, 39, 44, 54, 58, 177 axosomatic, 39, 44,54 of lateral column fibers, 57 dendritic, 59 excitatory, 2, 14 in cuneatus nuclei, 127 in gracilis nuclei, 127 inhibitory, 2 Synchronization, of neuronal discharges, 174 System, activating ascending reticular, 88 ascending activating, 283 averaging, 349 centrifugal, in subcortical reflex centers, 75 cortical-subcortical, 335 extralemniscal, activation of, 142 non-discriminative projections, 132 projections of, I35 thalamo-telencephalic projections of, 1 15 functional, 335 gamma-loop, of extensors, 10 homeostatic, 8 interconnected branching, 336 lemniscal. 115, 144 activation of, 142 projections of, 135 nonspecific, 349 pontolimbic, 408, 419 reticular, 408, 419 ascending, 127, 129 thalamic, 227 sensory nonspecific effect upon electrical responses, 273 somatosensory, response, I57 spinoreticulothalamic integrative relays of, 139 spinothalamic, 144 visual, 156 Tegmentum, mesencephalic, 309, 41 5 Telencephalon, and thalamic integration, 115-148 convergent regions of, 122 projections of thalamus to, 13I Terminals, axosomatic, 56 synaptic, 59 Thalamus, afferents of, somatic, 115 centromedial, 207 lateral, 207 medial, stimulation of, 155 midline, 349 nuclei, convergent, amplitude variations, 123
SUBJECT INDEX
nonspecific, 155-180 postero-medial, nonspecific, 312 projections, to telencephalon, 131 spindle activity, 408 stimulation of, 155-180 Threshold, depolarization of, 59 during dark adaptation, 200 for arousal, 429 reversal for effects of continuous illumination, 197 Tongue, tactile, thermal and gustatory impulses from, 22 1 Topography, cortical efferent, 300, 302 of corticofugal projections, 309 with stimulus-response delay, 387, 388 with stimulus-response interval, 387 Tract (path, bundle), optic, response of, 165 stimuli to, 156 pyramidal, 102 spinocerebellar, dorsal and ventral, 5 spino-reticular, 88, 127 spino-thalamic, I25 neo-, 127 paleo-, 127 visual, electrical stimulation of, 175 optic stimulation of, 175 Transection, of posterior column, 125 Transmission, internuncial, cerebral control of, 75 time, cortico-reticular, 92
493
Transmitter, action, 50, 57 amount of, 56 substance, release of, 55 Unit, “decision”, 226 extracellular, recording of, 138 individual, pattern of firing, 283 Variability, decrease in, 282 Variation, due to affective state, 382 extra-individual spatial, in cerebral evoked potentials, 382 Vestibular system, area I, 21 1 response in area, 17, 207 stimuli, neuronal effects of, 207 Viadril, associative responses, 118 Visual system, area I, 210, 304 Visual cortex, stimuli, neuronal effects of, 207 Voltage, mean-square, 189 root-mean-square, 189 Volume, control, 284 Wakefulness, cerebral, 413 Waves, I-waves, 244 pyramidal, 245
PRINTED IN T H E NETHERLANDS B Y N E D E I< L A N DS 11 B O E li D I< U I i I N R I C H T 1 NG N. V., ' S -H E R T O G E N B O S C H
