Physiology of Spinal Neurons

PROGRESS I N BRAIN RESEARCH V O L U M E 12 PHYSIOLOGY OF SPINAL N E U R O N S

PROGRESS IN BRAIN RESEARCH

ADVISORY BOARD W. Bargmann H. T. Chang

E. De Robertis J. C. Eccles J. D. French

H. Hydkn J. Ariens Kappers

S. A. Sarkisov J. P. Schadt

F. 0. Schmitt

Kiel Shanghai Buenos Aires Canberra Los Angeles Goteborg Amsterdam Moscow Amsterdam Cambridge (Mass.)

T. Tokizane

Tokyo

H. Waelsch

New York

J. Z. Young

London

PROGRESS I N BRAIN RESEQiRCH V O L U M E 12

PHYSIOLOGY OF SPINAL NEURONS EDITED B Y

J . C . ECCLES The John Curtin School of Medical Research, Department of Physiology, Canberra City AND

J. P. S C H A D E Central Institute for Brain Research, Amsterdam

ELSEVIER P U B L I S H I N G C O M P A N Y AMSTERDAM

/

LONDON

1964

/

NEW Y O R K

ELSEVIER P U B L I S H I N G COMPANY

335

J A N V A N G A L E N S T R A A T , P.O. B O X

211,

AMSTERDAM

AMERICAN ELSEVIER P U B L I S H I N G COMPANY, INC.

52

V A N D E R B I L T A V E N U E , N E W Y O R K N.Y.

10017

ELSEVIER P U B L I S H I N G COMPANY LIMITED 1 2 B , RIPPLESIDE COMMERCIAL ESTATE R I P P L E R O A D , B A R K I N G , ESSE X

This volunie contains a series of lectures delivered during a workshop on PHYSIOLOGY O F SPINAL NEURONS

which was held as part of the first International Summer School of Brain Research. at the Royal Academy of Sciences, Amsterdam (The Netherlands) from 15-26 July, 1963 This meeting was organized by the Central Institute for Brain Research and sponsored by the Netherlands Government and the NATO Advanced Study Institute Program

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

199

ILLUSTRATIONS A N D

3

64-18506

TABLES

ALL RIGHTS RESERVED T H I S B O O K OR A N Y P A R T T H E R E O F M A Y N O T BE R E P R O D U C E D I N A N Y FORM, I N C L U D I N G P H O T O S T A T I C OR M I C R O F I L M FORM, W I T H O U T W R I T T E N PERMISSION FROM T H E PUBLISHERS

List of Contributors

J. C. ECCLES,The John Curtin School of Medical Research, Department of Physiology, Canberra. I. ENGBERG, Department of Physiology, University of Goteborg, Goteborg (Sweden). R. GRANIT, The Nobel Institute for Neurophysiology, Karolinska Institutet, Stockholm. D. KERNELL,The Nobel Institute for Neurophysiology, Karolinska Institutet, Stockholm. A. LUNDBERG,Department of Physiology, University of Goteborg, Goteborg (Sweden). F. MAGNI,Istituto di Fisiologia dell’ Universita di Pisa e Centro di Neurofisiologia del C.N.R., Sezione di Pisa, Pisa (Italy). 0. OSCARSSON, Institute of Physiology, University of Lund, Lund (Sweden). C. G. PHILLIPS, University Laboratory of Physiology, Oxford (Great Britain). R. PORTER,University Laboratory of Physiology, Oxford (Great Britain). R. F. SCHMIDT, Institut fur Allgemeine Physiologie, Universitat Heidelberg, Heidelberg (Deutschland). T. A. SEARS,The John Curtin School of Medical Research, The Australian National University, Canberra. A. VAN HARREVELD, Kerckhoff Laboratories of the Biological Sciences, California Institute of Technology, Pasadena, Calif. (U.S.A.). P. D. WALL, Department of Biology and Research Laboratory of Electronics, Massachusetts Institute of Technology, Cambridge, Mass. (U.S.A.). W. D. WILLIS,lstituto di Fisiologia dell’ Universita di Pisa e Centro di Neurofisiologia del C.N.R., Sezione di Pisa, Pisa (Italy).

Other volumes in this series:

Volume 1 : Brain Mechanisms Specific and Unspecific Mechanisms of Sensory Motor Integration Edited by G. Moruzzi, A. Fessard and H. H. Jasper

Volume 2 : Nerve, Brain and Memory Models Edited by Norbert Wiener and J. P. Schade Volume 3 : The Rhinencephalon and Related Structures Edited by W. Bargmann and J. P. Schade Volume 4: Growth and Maturation of the Brain Edited by D. P. Purpura and J. P. Schadt Volume 5 : Lectures on the Diencephalon Edited by W. Bargmann and J. P. Schade Volume 6 : Topics in Basic Neurology Edited by W. Bargmann and J. P. Schadt Volume I : Slow Electrical Processes in the Brain b y N . A. Aladjalova Volume 8 : Biogenic Amines Edited b y Harold E. Himwich and Williamina A. Himwich

Volume 9: The Developing Brain Edited b y Williamina A. Himwich and Harold E. Himwich Volume 10: Structure and Function of the Epiphysis Cerebri Edited by J . Ariens Kappers and J. P. Schadt Volume 11 : Organization of the Spinal Cord Edited by J. C . Eccles and J. P. Schad6 Volume 13 : Mechanisms of Neural Regeneration Edited by M . Singer and J. P. Schadk Volume 14 : Degeneration Patterns in the Nervous System Edited b y M . Singer and J. P. Schadb

The remarkable progress in basic neurophysiology over the last few yearBisIillustrated by the fine collection of reviews in this volume. Gathered for this week of lectures and discussions were representatives from many of the leading schools of research into the properties of nerve cells and of their functional organization in the spinal cord, which has long been regarded as the simplest level of the central nervous system. However, after reading the complexities of neuronal interconnection here described, one may well wonder if this is an illusion! A tremendous amount of integration is carried out at the spinal level - far more than Sherrington conceived in his classic book ‘The integrative action of the nervous system’ -but, of course, this development in our concepts of spinal integration would have delighted him. One can predict that much more complexity of behaviour will be revealed as methods of investigation become more refined and are pursued with that systematic intensity that characterizes so much of present neurophysiology. Our ultimate hope undoubtedly is that, with increasing knowledge of neuronal interconnections, there will emerge clear ideas on basic patterns of neuronal organization, the same general type of pattern being employed in integrating the many different modalities of input. Of great importance are the many modes of control exercised by the higher centres onto the spinal mechanisms. Though the wealth of descending pathways had long been revealed by anatomical investigations, there have until recently been only relatively crude concepts of the mode of operation of these pathways, both on the local mechanisms in the spinal cord, and on the relay of impulses up the ascending tracts to the brain. More than one third of this volume is devoted to the descending and ascending pathways, and a wealth of new information and ideas will be found in these important papers. We can anticipate many new developments in these attempts to understand the physiology of communication up and down the spinal cord. I am going to be rash enough to predict that the centre of interest in the nervous system is now moving from the investigation of properties of the individual neurones and of the individual synapses to the much wider concepts of the patterns of functional organization, which give ultimate meaning to the individual neuronal and synaptic properties and subsume all these into the various levels of organization. In the context of these ideas we can appreciate that far more work is required on the way in which the nervous system handles inputs produced by carefully controlled adequate stimulation. The cruder physiological methods of electrical stimulation are, of course, needed in order to define the patterns of connection and the modes of operation; but this understanding must be given functional meaning by rigorous investigations with adequate inputs and into the way in which these act and interact. J. C . ECCLES

This Page Intentionally Left Blank

Con tents .................................

V

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

VII

List of contributors Preface

The excitatory responses of spinal neurones J. C. Eccles (Canberra City) . . . . . .

. . . . . . . . . . . . . . . . . . . . . .

Maintained firing of motoneurones during transmembrane stimulation R. Granit (Stockholm). . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . .

35

. . . . . . . . . . . . . . . .

42

. . . . . . . . . . . . . . . . . . . . . . .

56

. . . . . . . . . . . . . . . . . . . . . . . .

65

The delayed depolarization in cat and rat motoneurones D. Kernell (Stockholm) . . . . . . . . . . . . . . The properties of reticulo-spinal neurons W. D. Willis and F. Magni (Pisa, Italy) Presynaptic inhibition in the spinal cord J . C. Eccles (Canberra City) . . . .

1

Presynaptic control of impulses at the first central synapse in the cutaneous pathway P. D. Wall(Cambridge, Mass.) . . . . . . . . . . . . . . . . . . . . . . . . The pharmacology of presynaptic inhibition R. F. Schmidt (Heidelberg, Germany) . .

. . .

92

......................

119

Ascending spinal hindlimb pathways in the cat A. Lundberg (Goteborg, Sweden) . . . . .

. . . . . . . . . . . . . . . . . . . . .

135

Differential course and organization of uncrossed and crossed long ascending spinal tracts 0. Oscarsson (Lund, Sweden) . . . . . . . . . . . . . . . . . . . . . . . . . . .

164

Three ascending tiacts activated from Group I afferents in forelimb nerves of the cat 0. Oscarsson (Lund, Sweden) . . . . . . . . . . . . . . . . . . . . . . . .

. . .

179

Supraspinal control of transmission in reflex paths to motoneurones and primary afferents A. Lundberg (Goteborg, Sweden) . . . . . . . . . . . . . . . . . . . . . . . . . .

197

The pyramidal projection to motoneurones of some muscle groups of the Baboon’s forelimb C. G. Phillips and R. Porter (Oxford, Great Britain). . . . . . . . . . . . . . . . . .222 Afferent connections t o reticulo-spinal neurons F. Magni and W. D. Willis (Pisa, Italy) . .

. . . . . . . . . . . . . . . . . . . . .

Investigations on respiratory motoneurones of the thoracic spinal cord T. A. Sears (Canberra). . . . . . . . . . . . . . . . . . . . . .

246

. . . . . . . . .

259

. . . . . . . . . . . . . . . . . . . . . .

274

Effects of spinal cord asphyxiation A. Van Harreveld (Pasadena, Calif.).

280

Author index.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

308

Reflexes to toe muscles in the cat’s hindlimb I. Engberg (Goteborg, Sweden) . . . . .

Subject index

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

3 I2

This Page Intentionally Left Blank

1

The Excitatory Responses of Spinal Neurones J. C . ECCLES The John Curtin School of Medical Research, Department of Physiology, Canberra City

The body and dendrites of neurones are specialized for the reception and integration of information from other nerve cells, which of course occurs via the encrusting synaptic structures. During normal activity any one neurone is continuously bombarded by impulses impinging at its synapses, some of which are excitatory and tend to make this neurone in turn fire an impulse while some have an inhibitory action tending to prevent the initiation of impulses. My present task is to give an account of the way in which excitatory synaptic action causes spinal neurones to discharge impulses. An initial brief reference must be made to the ionic composition of nerve cells and to the electrical charges on their surfaces because these are essentially concerned both in nerve impulse transmission and in synaptic action. As shown in Table I the surface TABLE I I O N I C C O N C E N T R A T I O N S A N D E Q U I L I B R I U M P O T E N T I A L S FOR C A T M O T O N E U R O N E S

Outside m M

Inside mM

Na 150 K 5.5 CI 125

about 15 150

9

Equilibriuni potential (according to ihe Nernst equation) in mV

about $60 -90 -70

membrane of a nerve cell separates two aqueous solutions that have very different compositions. Within the cell sodium and chloride ions are at a lower concentration than outside, whereas with potassium there is an even greater disparity - almost 30-fold - in the reverse direction. Under resting conditions potassium and chloride ions move through the membrane much more readily than sodium. Necessarily the electrical charge across the membrane influences the rates of diffusion of charged particles in both directions between the interior and exterior of the cell. The potential across the surface membrane is normally about -70 mV, the minus sign signifying inside negativity. As seen from the table the equilibrium potential for chloride ions is approximately Refrrmcrs p . 29-31

2

J. C . E C C L E S

the same as the resting potential, which signifies that under such conditions the inward and outward diffusion of chloride ions approximately balance. On the other hand the large electrochemical potential difference for sodium ions (1 30 mV) will cause the diffusion of sodium inwards to be more than 100 times faster than outwards. Fortunately the resting membrane is much less permeable t o sodium than to potassium and chloride ions: but of course there must be some other factor concerned in balancing sodium transport across the membrane. Hence we have the postulate that there is built into the membrane a kind of pump that uses metabolic energy to force sodium ions uphill (up the electro-chemical gradient) and so outward through the cell membrane, as is diagrammatically shown in Fig. 1. This diagram further shows that there E X T E R 10R

I

SURFACE

MEMBRANE

K+

IONIC

FLUXES

V K '

-

I

INTERIOR

1

DIFFUSIONAL

N 0 Y . METABOLIC

FLUXES

DRIVE

IONIC FLUXES

ONAL

Fig. 1. Diagrammatic representation of K+ and Na' fluxes through the surface membrane in the resting state. The slopes in the flux channels across the membrane represent the respective electrochemical gradients. At the resting membrane potential (-70 mV) the electro-chemical gradients, as drawn for the K i and Na+ ions, correspond respectively to potentials which are 20 mV more negative and about 130 mV more positive than the equilibrium potentials (note the potential scale). The fluxes due to diffusion and the operation of the pump are distinguished by the direction of hatching. The outward diffusional flux of Na+ ions would be less than 1 of the inward and so is too insignificant to be indicated as a separate channel in this diagram, because the magnitudes of the fluxes are indicated by the widths of the respective channels. (From Eccles, 1957.)

is an excess of diffusion outwards of potassium down the electro-chemical gradient of about 20 mV, and again the transport of potassium ions is balanced by an inward pump. In fact, as shown, the sodium and potassium pumps are loosely coupled together and driven by the same metabolic process, which is now fairly well defined (Hodgkin, 1958; Caldwell et a]., 1960a,b).

EXCITATORY RESPONSES OF S P I N A L N E U R O N E S

3

THE SPIKE POTENTIALS OF SPINAL NEURONES

Under resting conditions the surface membrane of the nerve cell and its axon resembles a leaky condenser charged at a potential of about -70 mV. If this charge is suddenly diminishe'd, by about 20 mV, i.e. t o -50 mV, it initiates a n intense regenerative process adcling t o the depolarization and reversing the membrane potential as may be seen in Fig. 2A at -77 mV. It is postulated that this regenerative process is due to a high sodiurn permeability which occurs for only a fraction of a millisecond and is followed by a rapid development of a high potassium permeability with recharging of the membrane by outward movement of potassium ions (cf. Hodgkin,

-A&-600 - - - -

A-

-78

A-

//f

A

-80

-

8

k

>

. I mSEC

Fig. 2. (A). Intracellular responses evoked by an antidromic impulse, indicating stages of blockage of the antidromic spike in relation to the initial level of membrane potential. Initial membrane potential (indicated to the left of each record) was controlled by the application of extrinsic currents. Resting potential was at -80 mV. The lowest record was taken after the amplification had been increased 4.5 times and the stimulus had been decreased until it was just at threshold for exciting the axon of t h e motoneurone. (From Coombs et a / . , 1955a.) (B). Schematic drawing of a motoneurone showing dendrites (only one drawn with terminal branches), the soma, the initial segment of axon (IS) and the medullated axon (M) with two nodes, at one of which there is an axon collateral. The three arrows indicate the regions where delay or blockage of an antidromic impulse is likely to occur. The regions producing the M, IS, and SD spikes are indicated approximately by the labelled brackets. (From Eccles, 1957.) References p. 29-31

4

I. C . E C C L E S

1958). As a consequence the membrane potential rapidly is restored to normal. The potential change lasts for less than one thousandth of a second, its brevity earning it the name, spike potential. This general statement may serve as an introduction t o a more detailed treatment of neuronal spike responses. The complex morphology of a neurone is associated with a corresponding complexity in its excitatory responses. For example in Fig. 2B an antidromic impulse propagates in the direction of the arrow up t o the motoneurone in which has been inserted a microelectrode, as shown diagrammatically. The full sequence of potential changes is illustrated in Fig. 2A in which current applied through one barrel of a double microelectrode changed the membrane potential from the resting level of -80 mV either up as far as -87 mV or down as low as -60 mV. This procedure (Coombs et a]., 1955a) shows that the antidromic spike potential has three distinct components, each of all-or-nothing character. There is first the very small spike (about 5 mV) that is seen alone sometimes at -82 mV and always at -87 mV, and which is shown by threshold differentiation (lowest record of Fig. 2A) t o be generated by a n impulse in the motor axon of the impaled motoneurone. Second there is the larger spike of about 40 mV that is always set up at membrane potentials of -80 mV or less, and also sometimes at -82 mV. Finally, the full-sized spike is superimposed at membrane potentials of -77 mV or less, and rarely at -78 mV. Fig. 2B shows the antidromic pathway together with the regions of the motoneurone (M, IS and SD) in which the three components of the spike potentials are believed to be generated. This identification was made originally (Brock et al., 1953; Coombs et a/., 1955a) on the grounds that the large spike must be generated in that part of the motoneuronal membrane that is most closely related to the intracellular electrode, that is in the membrane of the soma and adjacent dendritic regions; and it was supported by an analysis of the extracellular field potentials generated by antidromic invasion of a single motoneurone (Fatt, 1957a) and of the action of an antidromic impulse on excitatory synaptic potentials. The conclusions from these earlier arguments have been fully confirmed by the very rigorous investigations of Terzuolo and Araki (1961) and Araki and Terzuolo (1962). The diagrammatic assignment in Fig. 2B of specific regions of the motoneurone to the three types of spike potentials in Fig. 2A, can therefore be regarded as firmly established, and it will be convenient to use the terms IS and SD spikes as indicated in Fig. 2. By means of double microelectrodes Terzuolo and Araki (1961) recorded simultaneously from inside a motoneuronal soma and just outside it. The intracellular antidromic responses in Fig. 3A and B resemble either the large composite IS-SD spikes of Fig. 2A, with an inflection on the rising phase, or else the smaller IS spike, as with some responses evoked by the second antidromic impulse in Fig. 3B. The composite IS-SD spike is seen to be associated with a complex extracellular potential: firstly a negligible upward deflection (positivity), then a double downward negative wave, and a final large positivity. The second negative wave and the final positivity are shown in Fig. 3B to be associated with the SD spike, because, in the superimposed traces for the second antidromic response, these two waves are eliminated when the SD

EXCITATORY RESPONSES OF S P I N A L NEURONES

5

ETm" -1

I msec

msec

-~._

I msec

Fig. 3. (A, B). Potential changes recorded simultaneously, inside and outside the soma of spinal motoneuroiies with parallel microelectrodes. In A the stimulus was adjusted to activate the axon in the ventral root in approximately half of the superimposed traces of antidromic spike responses. In B the stimuli were applied to the ventral root at a short time interval, so that in some of the trials the soma-dendrite complex was not invaded. ( C ) and (D) show simultaneous intracellular traces of dendritic and soma spike potentials from the same motoneurone. In (E) there was simultaneous intracellular and extracellular recording as in A, but the extracellular recording was from a remote dendrite. (From Terzuolo and Araki, 1961.)

invasion fails. Furthermore, the onset of the second extracellular negative wave always occurs at the inflection between the IS and SD spike potentials. Terzuolo and Araki (1961) point out that in interpreting these potential changes it is essential to recognize that the extensive radiating dendritic arborization of a motoneurone gives a closed field organization of extracellular potentials (cfi Lorente de N6, 1947, 1953), as soon as the antidromic impulse invades the initial segment of the axon and the soma, which both lie approximately central to the dendritic arborization. The dendrites form the dominant source for extracellular currents flowing radially inwards first to the activated initial segment, then into the activated soma; hence an extracellular electrode i n close proximity to the soma, and thus near to the centre of the closed field, would be expected to record negativity for both of these activations, as is seen to occur in Fig. 3A and B. The terminal positive wave only occurs after an SD spike potential, which is illustrated in Fig. 3B. This association indicates that the spike potential propagates from the soma out along the dendrites so that the soma becomes a source for extracellular current flowing into the sinks on the dendrites. Conclusive evidence for impulse propagation along dendrites is provided by Fig. 3C-E. C and D give two examples in which simultaneously recorded spike potentials from the same motoneurone had quite different time courses, one was the typical intrasomatic potential as in Fig. 3A, the other had a slower rise and a much slower decline. In C the two spike potentials were about the same height, but there was 0.3 msec between summits, which must be attributed to dendritic conduction time. Rrfrrmcrs p . 29-31

6

J. C . E C C L E S

In D the dendritic conduction time was almost 0.5 msec and the very small IS spike also demonstrated the remoteness of that intracellular electrode from the soma. A dendritic conduction time of about 0.4 msec for 0.3 mm propagation was observed by Fatt (1957a) in his elegant mapping of the extracellular field potential generated by antidromic invasion of a single motoneurone. In the extracellular record of Fig. 3E the second negative wave was much later than in Fig. 3A and B, the onset being simultaneous with the SD spike summit, and there was no subsequent positive wave. Evidently this extracellular recording must be from a region of the neurone that is rather more than 0.2 msec conduction time from the soma, and even so far out on the dendrites that the further dendritic invasion does not give opportunity for a reverse current flow in the outward direction. It can be concluded that the experiments here described, together with those of Fatt (1957a) and Lorente de N6 (1947) prove that an antidromic impulse propagates some hundreds of microns along the dendrites of motoneurones at a velocity of the order of 1 m/sec, but do not allow any statement about impulse conduction in fine dendritic branches. At least with the large dendrites we are justified in assuming that the surface membrane has the same excitatory properties as the soma membrane. It was originally believed that an initial IS spike was observed only as a stage of

A A-

d

L 200~/sec

Fig. 4. Spike potentials evoked in a motoneurone with a membrane potential of -70 mV by three different modes of stimulation; A, antidromic; B, monosynaptic; C , by a depolarizing pulse through one barrel of the double microelcctrode. The lower traces show the electrically differentiated records. D and E are tracings of A and B with perpendicular lines from the origins of thc IS and SD spikes, the horizontal lines giving the respective thresholds. In F the lines of current flow are drawn as described in the text. (From Coombs et al., 1957a.)

EXCITATORY RESPONSES OF SPINAL NEURONES

7

progressive invasion of a motoneurone by an antidromic impulse propagating sequentially from the medullated axon to the initial segment of the axon, then to the soma and dendrites. This explanation was refuted by the discovery that, with synaptic and direct electrical stimulation of a motoneurone, the IS spike also precedes the SD spike (Araki andOtani, 1955; Fatt, 1957b; Fuortesetal., 1957; Coombset ul., 1957a,b). The electrically differentiated records in the lower row of Fig. 4A, B and C show indubitably that with antidromic, synaptic and direct electrical excitation of the motoneurone, its first response is an IS spike. In B and C the stimulus applies the depolarization more directly to the soma than to the initial segment, where the spike is initiated; hence it must be concluded that the initial segment has a much lower threshold. Comprehensive analytical investigations have led to the conclusion that with motoneurones in good condition the threshold depolarization of the initial segment is always less than half of that for the soma, the respective ranges being 6-18 mV and 20-37 mV with mean values of 10 mV and 27 mV respectively (Coombs et al., 1957b). In Fig. 4 A-C potentials in the upper traces give virtually a record of the changes in soma membrane potential and show that the activation of the initial segment adds very effectively to the depolarization of the soma membrane so that its threshold is attained and the SD spike generated. The lines of current flow into the activated initial segment from the soma are shown in Fig. 4F. The levels of depolarization for synaptic generation primarily of an IS spike and secondarily of an SD spike are given in the drawings of Fig. 4E. As would be expected the IS spike generates an SD spike at practically the same level of depolarization for antidromic (D) and synaptic activation (E) of the motoneurone. When SD impulse generation fails, the IS spike attains a peak of only about 40 mV and then declines as seen in Fig. 2A. It will be appreciated that it i s merely the electrotonic extension of the IS spike to the soma that has this small value. With intracellular recording from the initial segment the spike potentials may be in excess of 80 mV (Coombs et al., 1957a; Terzuolo and Araki, 1961). The relationship of depolarization to IS and SD spike generation is well shown in the voltage-clamp recordings of Fig. 5 (Araki and Terzuolo, 1962). By means of a feedback device the membrane potential is displaced to a desired level and held there by current applied through one barrel

C V

1 msec

0.4

Fig. 5 . Threshold difference between axon and soma. (A-E). Upper beam, membrane current; lower beam, membrane potential. (F). Relation between cathodal displacement of the membrane potential from the resting level (abscissa) and peak inward current (ordinate). Full explanation in the text. (From Araki and Terzuolo, 1962.) References p . 29-3I

8

J. C. E C C L E S

of a double microelectrode assemblage. In Fig. 5 B a depolarization of about 10 mV was applied by the voltage-clamp, and, after the initial displacing current, a later inward current by the voltage-clamp (C trace) showed that the membrane depolarization had resulted in a spike-like inward current across it. Much larger depolarizations in C and D also resulted in the same brief inward current, but with a shorter latency. However, with a still further increase in depolarization (E), there was second large component of inward current. The plotted points of Fig. 5F show the step-like in-

n

DSCT

4 B

/

A

V

CUT l50m"

-

A V

v

A

r

rnsec

L

A

A

h

rnsec

Fig. 6. (A-D). Tntracellularly recorded responses of a neurone of Clarke's column with the electrically differentiated records immediately below. D is an antidromic spike potential evoked by an impulse descending the dorsal spino-cerebellar tract. A, B and C are responses evoked by progressively larger afferent volleys in the nerve to quadriceps muscle. The neurone was discharging spontaneously, hence the sloping base lines. (From Curtis, Eccles and Lundberg, unpublished observations.) (E-G). Intracellularly recorded responses of a neurone of the dorsal horn in the L7 segment, with the electrically differentiated records immediately below. G, the antidromic spike potential evoked by an impulse descending the lateral column on the ipsilateral side, the extracellular potential generated by the descending volley being seen in the lowest trace. E and F are responses evoked by a small and a large afferent volley from the superficial peroneal nerve. (From Eccles et al., 1960.)

crement in inward current that occurred with a depolarization of 30 mV. There can be no doubt that the two components of motoneuronal excitability revealed in Fig. 5 are the IS and SD components of Figs. 2, 3 and 4. Of course under voltage-clamp conditions the inward current of the IS response cannot add to the depolarization of the soma membrane as in Fig. 4F. The SD membrane can be activated only when the applied voltage attains threshold level. The SD current in Fig. 5E had a later onset than the IS current solely because the applied voltage was just above the SD threshold, a long latency being observed under comparable conditions for the IS spike in Fig. 5B. With larger voltages the IS and SD spikes were synchronous(Araki andTerzuolo, 1962). In the spinal cord there is much variation with respect to the threshold discrimination between the IS and SD regions of nerve cells. For example, the same large discrimination as with motoneurones is seen in the differentiated records from the

EXCITATORY RESPONSES OF S P I N A L NEURONES

9

cells in the dorsal horn on which large cutaneous fibres make synaptic contacts (Fig. 6E-G; Eccles et al., 1960), while there is no sign of IS-SD separation with the intermediate neurones that relay the group Ia and Ib afferent impulses from muscle (Fig. 71,J; Eccles et a/., 1960). Possibly these are amongst the interneurones in which

A

C

D J-

msec

-----???-

,

,

I

Fig. 7. Intracellular recording from a type A interneurone in the L7 segmental level at a depth of 2.0 mm from the cord dorsum. Membrane potential, about -70 mV throughout. A-D are superimposed traces of EPSP's (above) and dorsal root action potentials (below) evoked by progressively increasing stimuli, applied to the combined FDL-PI nerves. E-H are single traces at faster sweep and lower amplification, the evoking stimuli being about same strength as in C for E and F, and in D for G and H. Lower trace of 1 and upper of J are electrically differentiated records of the EPSP and superimposed spike evoked by a volley from the posterior tibia1 nerve, which supplies flexor digitorum brevis and the various toe muscles, there being two superimposed traces in I. K shows, at slow sweep speed, a record of potentials evoked in response to a stimulus of same strength as in C , but at lower amplification. The EPSP was just at threshold, evoking a spike in one of the two traces. Potential and time scales are given at appropriate places, one of the potential scales obtaining for A-D and the other for E-K, as shown. Note very fast sweep for I and J.

Hunt and Kuno (1959) could find no IS-SD separation. On the contrary there is good IS-SD discrimination with the cells of the ventral spino-cerebellar tract (Eccles et al., 1961b) and the cells of Clarke's column (Fig. 6B-D; Curtis et al., 1958). An important functional consequence of the much lower threshold of the IS component of a neurone is that it acts as a far better integrator of the whole synaptic excitatory and inhibitory bombardment than would be the case if impulses were generated anywhere over the whole soma-dendritic membrane. If these latter conditions were obtained, a special strategic grouping of excitatory synapses (cf. Lorente de N6, 1938) could initiate an impulse despite a relative paucity of the total excitatory synaptic bombardment and a considerable inhibitory bombardment of areas remote from this focus. As it is, both excitatory and inhibitory synaptic action are effective only in so far as they affect the membrane potential of the initial segment. It is here Rrferenrr.~p . 29-31

10

J. C. E C C L E S

C

B

A

57 m V (A) 5 rnV (B, C)

\

5 sec

210 sec

I

15 set

Amsec @b *w 620 sec

1

600sec

Lr 1 1 1 I 1 I I I 1 1 1 1 I I t ---I10 msec 10 msec Fig. 8. Effect of the injection of Na’ ions on the antidromic spike and after-hyperpolarization of a motoneurone, the microelectrode being filled with Naps04 (1.2 equiv. per litre). After obtaining the top record in A, the N a k content of the motoneurone was increased by about 25 p. equiv. by applying a depolarizing current of 4 x lO-*A for 120 sec, and the further records in A were obtained at the approximate times indicated following the injection. Complete recovery had occurred by the time of the last record in A, and the top record of B was then taken. This was followed by the injection of approx. 25 p. equiv. Na+ (applying 5 x 10-sA for 150 sec) and the remaining records in B and C were taken at the indicated times after the injection. All records are formed by superposition of about 40 faint traces. In B and C, a full action potential has been set up in each sweep, although the spike is not shown with the slow sweep and high amplification used to display the after-hyperpolarization. Voltage scale applies to different parts of the figure as indicated. (From Coombs et al., 1955a.)

EXCITATORY RESPONSES OF SPINAL NEURONES

11

that the conflict between excitation and inhibition is joined, not generally over the motoneuronal surface, as was envisaged by Sherrington in his concept of algebraic summation (Sherrington, 1925; Eccles and Sherrington, 1931). It is possible to alter the ionic composition of a motoneurone by passing a current through a microelectrode filled with an appropriate salt. For example, if a current is passed into a motoneurone through a microelectrode filled with a concentrated sodium salt, it is largely carried into the neurone by Na+ ions and leaves across the surface membrane of the neurone mainly by the outward passage of K + ions. In Fig. 8A the current of 4 x IOPsA for 120 sec will add about 25 p. equiv. of Nab to the cell and at the same time deplete its potassium by about 20 p. equiv. Since a standard motoneurone contsins about 35 p. equiv. of Kf ions, at least half will be removed by the current. It is seen typically in Fig. 8A that immediately after the passage of this current an antidromic impulse i n the motor axon failed to invade the soma and dendrites of the neurone. There was merely an IS spike which was diminished in size and had an abnormally long time course. Invasion was first observed after about 20 sec, but the spike potential was then small and very prolonged. Thereafter progressive recovery occurred, so that a normal spike potential was observed about 300 sec later. All of the experimental observations on the neuronal spike potential are explained satisfactorily by the hypothesis stated above (cfi Hodgkin, 1958) that the spike potential arises because of a brief high permeability first to Naf ions and then to K+ ions. The steep rising phase of the spike with reversal of the merrbrane potential is attributable to the intense net inward flux of Na+ ions, which causes the membrane potential to approach the equilibrium potential for Na+ ions. The intracellular injection of 25 p. equiv. of Naf ions would cause a very large diminution of the concentration difference across the membrane, and hence account for the much slower rising phase of the spikes at 25 sec in Fig. 8A. Its lower voltage can be explained both by the lowered equilibrium potential for Na+ ions resulting from this change in concentration and by the diminution of membrane potential which resulted from the Na+ injection. Similarly, the slower falling phase of the spike in Fig. 8A is attributable to the lowered internal concentration of K+ ions with the consequent diminution in the outward flux of K f ions during the phase of high K + ion permeability. It is significant that the slow rising phase seen in Fig. 8A is a prominent feature only when there is an incr:ased internal Na + concentration, while the slow falling phase occurs whenever the internal Kf concentration is diminished. In Fig. 8A the slow recovery back to normal indicates that there has been an extrusion of the excess of Nak ions and a replacement of the lost K + ions. The extrusion of the Naf ions occurs against the electro-chemical gradient and hence must be due to the operation of the sodium pump (c$ Fig. 1). The absorption of K + ions could be due in part to diffusion along the electro-chemical gradient, but in part the potassium pump must be concerned because the normal internal potassium concentration is about double the equilibrium concentration (Coombs et al., 1955a). Under certain specified conditions the maximum slopes of the rising and falling phases of the spikes can be used as measures of the internal Nai- and K+ ion concenReferences p . 29-3/

12

J. C. E C C L E S

trations respectively. Using these criteria Ito and Oshima (1964) have shown that, after an injection of Na+ ions as in Fig. 8, the normal concentrations of Na+ and K are restored with an exponential time course, For example in the differentiated records of Fig. 9A the maximum slopes for the SD spike are given by VZ and V4. When the +

I C

I \

I. B

500 400

2

nA

300

50

200

-2 100

50t.

-3

\

-?$I, ,

AV2 0

-A

RESTING POTENTIAL ,

min

- 90

Ava

rnin

-100

2

,

/

'

'

,

'

'

'

I

Fig. 9. (A). The sizes of the four peaks in the rising and falling slopes of the differentiated antidromic spike are shown by VI, VZ,V3 and V4. (B). Differences of VZ and V4 from their respective values at full recovery (VZ and V4) are plotted on a semilogarithmic scale against the time after an injection of sodium ions by a depolarizing current of 5 x lo-* A for 60 sec into a PBST motoneurone. ( C ) . Changes in the after-hyperpolarization following the indicated sodium injection from a NaCI-filled single microelectrode inserted into gastrocnemius-soleus motoneurone. The measurements were made at the points indicated by the arrows in D and E. The interrupted line through the plotted points is an exponential curve with a time constant of 104 sec. (D, E). Specimen records of the membrane potential (upper traces) obtained at 36 (D) and 342 sec (E) after a sodium injection.

differences from the fully recovered values are plotted on a semi-logarithmic scale in B, both VZ and V4 lie on the same straight line, the time constant of these exponential decays being about 90 sec. The identity of these two time courses of recovery indicates that as in other cells (cf. Caldwell et al., 1960a, b) there is the loose coupling of the Na+ and K+ pumps, that is illustrated in Fig. 1. THE AFTER-HY PERPOLARIZATION OF S P I N A L NEURONES

The SD spike potential of motoneurones is always followed by a prolonged afterhyperpolarization (AHP). It is important to discriminate between two motoneuronal hyperpolarizations that are generated by impulses in motor axons. One is the true after-hyperpolarization which is a sequel of the SD spike per se. The other is an inhibitory postsynaptic potential that is generated by impulses in motor axons operating through a pathway from motor-axon collaterals to Renshaw cells (Eccles et al., 1954). With antidromic activation the discrimination is secured easily by employing

13

EXCITATORY RESPONSES OF SPINAL NEURONES

a stimulus that is just at threshold for the axon of the motoneurone under observation. Thus as in Fig. IOA, two sets of records are obtained according to whether the axon is or is not excited, the difference between them being attributable to the true afterhyperpolarization. This method of discrimination depends on the experimental finding

-61

= A. -

AFTER-HY P E R P O LARIZATION

.----

c

B

/

U

-70

-5c

m -

-70

10

20

I

I

I

30

40

50

I

I

60 msec

-85 & = 4 / -IIC

,

'

I

,

10 msec

Fig. 10. (A). After-hyperpolarizations of a motoneurone, occurring at various levels of membrane potential as controlled by extrinsic current. For each record, the stimulus applied to the ventral root was adjusted to the critical strength at which the axon of the particular motoneurone was sometimes excited and other times it was not. The membrane potentials in mV at which the action potentials were evoked are given alongside each record. The resting potential varied from -76 mV to -79 mV. The spike component of the antidromic action potential does not appear in these records, the amplification being too high and the sweep too slow to display it satisfactorily. (From Coombs er al. 1955a.) (B). Diagrammatic representation of K+ ion fluxes during the after-hyperpolarization that occurs when the membrane potential has been preset at three different levels in addition to the resting potential of -70 mV. The relative diffusional fluxes inward and outward are calculated according to the Nernst equation, and are shown as equal at -90 mV (the equilibrium potential) and with the outward flux double the inward at -70 mV. (From Eccles, 1957.)

that an impulse in the axon of a motoneurone has a negligible inhibitory effect on that motoneurone (Eccles et al., 1954). It is of interest that the IS spike of a motoneurone is followed by a negligible after-hyperpolarization (Brock et a/., 1953). This is a further example of the difference between the IS and SD surface membranes. The distribution of synapses is also a distinguishing feature; many on the SD membrane, few on the IS membrane. With a normal neurone the SD spike potential does not immediately reverse to give References p . 29-31

14

J. C. E C C L E S

the AHP, but continues as a declining depolarization for several milliseconds (Figs. 2A, at -77 mV; 8A). The subsequent AHP increases to reach a maximum of about 5 mV at about 10 msec, and thereafter it gradually declines so that it can no longer be detected after about 100 msec (Figs. 8B, C lowest records; IOA). The motoneurones innervating fast contracting muscles have AH P’s of about 70-80 msec duration, while with those innervating slowly contracting muscles the AHP’s have about double that duration (Eccles et al., 1958). This differentiation is of functional significance because the AHP controls the frequency of repetitive discharge of a motoneurone. Thus the motoneurones supplying slow muscles discharge at the low frequencies appropriate for such muscles. There is good correlation between the duration of the AHP’s and the differentiation of motoneurones into tonic and phasic types (Granit et al., 1956, 1957). With cutaneous relay cells (cf. Fig. 6E-G) the spike potential is followed by an AHP that is comparable with that of a motoneurone both in size and duration (Eccles et al., 1960). Long AHP’s have also been observed for the spinal interneurones relaying Group la (Fig. 7K) and Group Ib impulses (Eccles e t a / . , 1960), and for the cells of origin of the ventral spino-cerebellar tract, where the AHP usually has a duration of 30-70 msec (Eccles et a/., 1961b). As shown in Figs. 8B, C and 9D, E the AHP is abolished when a considerable fraction of the intracellular potassium is replaced by sodium. The subsequent recovery follows much the same time course as the recovery of the spike potential. The A H P is similarly changed when depletion of intracellular potassium is coupled with injection of a cation other than sodium, for example tetramethylammonium or lithium ions. Since injection of a wide variety of anions into the neurone is without significant effect on the AHP, it may be concluded that it is not associated with an increased permeability to some anions. It therefore stands in sharp contrast to another hyperpolarizing potential - the inhibitory postsynaptic potential. It thus appears that the AHP is entirely due to the increase in membrane charge produced by the net outward movement of K+ ions. Depletion of intracellular potassium reduces this net movement and may even reverse it, as shown by the inversion of the AHP in Fig.9D. The recovery illustrated in Fig. 8B a n d C provides a good means of evaluating the rate of replacement of the lost intracellular potassium. In Fig. 9C an injection of Na+ ions by a current of 5 x lo-@A for 120 sec resulted i n reversal of the AHP (Fig. 9D) and recovery followed an exponential curve plotted in Fig. 9C. The time constant of this curve is 104 sec, so there is excellent agreement between the time courses of recovery indicated by the two measures of intracellular potassium concentration, the steepness of the declining slope of the spike in Fig. 9B (open circles) and the size of the AHP (Ito and Oshima, 1964). If the A H P is thus due solely to the net outward movement of K + ions, the effect produced by changing the membrane potential becomes of great significance. If the movement of the K + ions is due t o the operation of a pump, it is unlikely that it would be affected greatly by variation in the membrane potential. At least the sodium pump in giant axons is not affected appreciably by such conditions (Hodgkin and Keynes, 1955). On the other hand, if the movement of the K + ions is occurring along

EXCITATORY RESPONSES OF SPINAL NEURONES

15

their electro-chemical gradient, it should be changed very effectively, and even reversed, by varying the potential. Fig. 1OA shows that the size, but not the time course, of the AHP is greatly changed when the membrane potential is varied by an extrinsic current through one barrel of a double microelectrode, being greatly increased by depolarization and diminished by hyperpolarization. It was impossible to extend the series of Fig. 10A beyond a hyperpolarization to -87 mV, because the antidromic impulse then failed to invade the motoneurone. However, extrapolation suggests that beyond a membrane potential of about -90 mV the AHP should reverse to an after-depolarization. These observations indicate that the AHP is due to the net outward diffusional movement of Kf ions and not to the operation of a potassium pump. Furthermore, they show that the potassium equilibrium potential across the neuronal membrane is about -90 mV. Thus, following the spike potential there is a prolonged phase (about 100 msec) of increased potassium permeability of the neuronal membrane. In Fig. 10B there is illustrated the manner in which the AHP is produced by this increased K+ ion permeability and is affected by the level of the membrane potential. The unbalanced Kf ion fluxes, derived by calculation, are shown to be directly related to the size and sign of the after-potential. As revealed by the extent of the compensation, the additional potassium permeability causes an increase of only about 40% above the normal level of membrane conductance. It is probable that such an increase does not require even a doubling of the normal potassium permeability. SYNAPTIC EXCITATORY ACTION

Investigations with single stimulation The simplest example of synaptic action is illustrated in Fig. 11, where a single synchronous synaptic bombardment diminishes the electric charge on the cell membrane. There is a rapid rise to the summit and a slower, approximately exponential decay. This depolarization becomes progressively larger in A-C as the number of activated synapses increases. There is in fact a simple summation of the depolarizations produced by each individual synapse. In the much faster records of D-G it is seen that, when above a critical size, the synaptic depolarization evokes the discharge of an impulse, just as occurs in peripheral nerve, there being the explosive increase in sodium permeability at the double arrows in E-G. The only effect of strengthening the synaptic stimulus in E-G is the earlier generation of the impulse, which in every case arises when the depolarization reaches 18 mV. The depolarizing potentials that excitatory synapses produce in the postsynaptic membrane are called excitatory postsynaptic potentials (EPSP’s). The generation of impulses by excitatory synapses is entirely attributable to the EPSP’s, as may be seen in Fig. 1 IH-K, where an EPSP that failed to generate an impulse was caused to do so when superimposed on an applied depolarizing current, which in I, J, K was 4, 10 and 18 m p A respectively, and so contributed progressively more of the requisite 18 mV depolarization. There has now been extensive investigation of a wide variety of nerve cells in the central nervous system, and in every case synaptic transmission of impulses is due to References p 29-31

16

J . C. E C C L E S V

C

-

msec

Y

A

A‘ msec

Fig. 1 I . (A-C). EPSP’s obtained in a biceps-semitendinosus motoneurone with afferent volleys of different size. Inset records a t the left of the main records show the afferent volley recorded near the entry of the dorsal nerve roots into the spinal cord. They are taken with negativity downward and at a constant amplification for which no scale is given. Records of EPSP are taken at an amplification that decreases in steps from A to C as the response increases. Separate vertical scales are given for each record of EPSP. All records are formed by superposition of about 40 faint traces. (D-G). Intracellularly recorded potentials of a gastrocnemius motoneurone (resting membrane potential, -70 mV) evoked by a monosynaptic activation that was progressively increased from D to G. The lower traces are the electrically differentiated records, the double-headed arrows indicating the onsets of the IS spikes in E-G. (H-K). Intracellular records evoked by monosynaptic activation that was applied at 12.0 msec after the onset of a depolarizing pulse whose strength is indicated in mpA. A pulse of 20 mpA was just below threshold for generating a spike. H shows control EPSP in the absence of a depolarizing pulse. Lower traces give electrically differentiated records. Note that the spikes are truncated. (From Coombs et al., 1957b.)

this same process of the production of EPSP’s, which in turn generate impulse discharge when attaining a critical level of depolarization. For example Fig. 6A, B and C shows the effect of progressively larger synaptic excitation on a cell of the dorsal spino-cerebellar tract. In Fig. 6E and F the EPSP can be seen rising to the level at which a spike is initiated in a neurone in the dorsal horn that is monosynaptically excited from cutaneous afferent fibres. Interneurones of the intermediate nucleus also show graded EPSP responses as in Fig. 7A-D, while the faster traces of Fig. 7 E-H show the generation of an impulse when the EPSP is above a critical level. Where the IS threshold is much lower than the S D threshold, as for example in the motoneurone, the EPSP produced by the activation of synapses covering the soma and dendrites is effective not by generating an impulse in these regions, but by electro-

EXCITATORY RESPONSES OF S P I N A L NEURONES

17

tonic spread of the depolarization of the initial segment, which would occur by lines of current flow which are the reverse of those indicated in Fig. 4F. By recording the impulse discharged along the motor nerve fibre in the ventral root, it is found that usually this impulse starts to propagate down the medullated axon about 0.05 msec after the initiation of the IS spike. For example, it is calculated from the conduction time for the antidromic impulse in Fig. 12A and B that the impulses in D, E, G-J were ANTIDROMIC

DIRECT

0 RTHODROMIC

msec

Fig. 12. Upper traces are intracellularly recorded spike potentials evoked in a biceps-semitendinosus motoneurone (resting membrane potential, -60 mV) by an antidromic impulse (A, B), by a depolarizing pulse that began at the artifact and continued throughout the traces (C-E) and by monosynaptic activation by an afferent volley from the nerve to biceps-semitendinosus (F-J). In A, C and F the lower trace is an electrically differentiated record of the upFer trace. In D, E, G-J the lower trace is recorded monophasically from an isolated ventral root filament of L7. In the lower trace of B the electrode that records monophasically is used to record the antidromic volley relative to an indifferent earth lead, negativity being downwards. Arrows in D-J indicate time of initiation of the impulse in the medullated axon, as calculated from the spike in the ventral root, the measured antidromic conduction time after allowance of 0.07 msec for the M-IS interval. Same voltage scale for all intracellular records, and time scale obtains for all records. A compensatory circuit was employed with the depolarizing pulses of C-E. (From Coombs et al., 1957b.)

set up in the medullated axon at the times of the arrows, i.e. at a time 0.05 msec after the onset of the IS spike. Invariably, in normal motoneurones, synaptic excitatory action generates an SD spike, not directly by its depolarizing action, but indirectly through the mediation of an IS spike which lifts the depolarization of the SD membrane to threshold by currents that flow in the direction shown in Fig. 4F (Coombs et al., 1957b; Terzuolo and Araki, 1961; Araki and Terzuolo, 1962). When we investigate the way in which excitatory synapses produce the characteristic time course of the EPSP (Fig. 11 A-C), it is found that this depolarization of the postsynaptic membrane is due to currents flowing as shown in Fig. 13B from the membrane in through the cleft and so into the postsynaptic cell. Actually, analysis of the EPSP by means of the electric time constant of the membrane indicates that these References p . 29-31

18

J. C . E C C L E S

currents have the very brief duration shown by the broken line in Fig. 13A (Curtis and Eccles, 1959). By the voltage-clamping technique Araki and Terzuolo (1962) have obtained direct

1

I S

3.10'F

lo*

0s

5, I O n ~

6 i

7 -

--- - 7 h - 7 - 4

TllllillrlTTlllllrllllrrrm msec

I

m rec

mrec

msec

< -

msec

Fig. 13. (A). The continuous line is the mean of several monosynaptic EPSP's, while the broken line shows the time course of the subsynaptic current required to generate this potential change. (B). Diagram showing an activated cxcitatory knob and the postsynaptic membrane. As indicated by the scales for distance, the synaptic cleft is shown at 10 times the scale for width as against length. The currcnt generating the EPSP passes in through the cleft and inward across the activated subsynaptic membrane. (C).Formal electrical diagram of the membrane of a motoneurone with, on the right side, the circuit through the subsynaptic areas of the membrane that are activated in producing the monosynaptk EPSP. Maximum activation of these areas would be indicated symbolically by closing the switch. (D-I). EPSP's of neurone; recorded intracellularly as described in the text.

A

B

D

I

;

2 msec

E

-2 msec Fig. 14. Voltage-clamp method of determining the excitatory synaptic current flow. In A and D lower trace shows monosynaptic EPSP with a later spike in D. Under voltage-clamp conditions at the resting membrane potential the same monosynaptic excitation evokes the currents shown in B and E respectively. In C are plotted the current ( I ) and the voltage (2) shown in B and A respectively. (From Araki and Terzuolo, 1962.)

19

EXCITATORY RESPONSES OF S P I N A L NEURONES

records of the current that flows through the subsynaptic membrane when an EPSP (Fig. 14A) is produced by a single presynaptic volley. The membrane potential is clamped at the resting level in Fig. 14B and the current flow required for this clamping during synaptic activation can be assumed t o be an actual record of the flow of current across the activated subsynaptic membrane (Fig. 13B). This technique is necessarily imperfect because of the synapses on dendrites remote from the application of the clamp; but it is certainly a more reliable method than the analysis

--+30.y

-%-

L

-0

;

L

7

8

--+20

?<

$9

+3

IrC

-14

20 -100

-80

-60

-40 -20

Membrane potential (mV)

-&

-102 -84

7 r

Fig. 15. (A). EPSP's set up in a biceps-semitendinosus motoneurone at various levels of membranc potential indicated in mV to the left of each record. The resting potential was at -66 mV; the other potentials were obtained by the application of an extrinsic current through one barrel of a double microelectrode. The records are each formed by superimposing 15-20 sweeps. Spike potentials are evoked by the EPSP a t membrane potentials of -42 mV and -60 mV. (B). Plotting of maximum rate of rise of EPSPagainst initial level of membrane potential for series partly shown in A. ( C ) .Plot of peak amplitude of EPSP (open circles) and point of maximum curvature at start of action potential (filled circles) against initial level of membrane potential. (D). Plot of steady level of potential recorded inside the motoneurone (filled circles) and after withdrawing from the motoneurone (open circles) against applied current. (From Coombs ef al., 1955b.3 Referenres p. 29-81

20

J. C. E C C L E S

employed inderiving Fig. 13A. It shows that there is an initial intense current (Fig. 14C) that is very similar to the calculated curve (Fig. 13A). It is seen in Fig. 14E that the same synaptic current flow is recorded when the voltage-clamp is applied to an EPSP that generates a spike potential (D). This observation can be readily understood because there could be no generation of the spike under voltage-clamp conditions. When the membrane potential is displaced in the depolarizing direction by a steady current, the EPSP is reduced in size, and eventually when the membrane potential is at about zero potential, the EPSP is also at zero. A still larger background current reverses both the membrane potential and the EPSP (Fig. 15A, B and C). Hence it has been concluded that the currents producing the EPSP are caused to flow because under the excited synapses there is a virtual short-circuit of the membrane potential, which is illustrated diagrammatically in Fig. 13C, where the normal membrane with its potential of -70 mV would be depolarized by closing the switch in the component labelled E synapses, which is seen to be merely a resistance through which current would tend to reduce the membrane potential to zero. The equilibrium potential for excitatory synaptic action may therefore be said to be at about zero membrane potential. The existence of residual transmitter action after the initial peak as in Figs. 13A and 14C is exhibited indubitably at several types of excitatory synapse on spinal neurones, though the actual time course of the synaptic depolarizing action has not yet been determined by the voltage-clamping or analytical techniques. For example, the EPSP produced by the synchronous synaptic bombardment of a Renshaw cell has a brief intense phase that declines after about 2 msec onto a large residual potential that persists for as long as 50 msec (Fig. 131). Normally this residual EPSP is responsible for the repetitive discharges that are such a remarkable feature of Renshaw cell responses (Renshaw, 1946; Eccles et al., 1954, 1961a). When the enzymic destruction of synaptic transmitter is prevented by a large dose of eserine, a single synchronous synaptic activation can produce a repetitive discharge that persists for as long as 2 sec (Eccles et al., 1956a). In Fig. 13F-H are EPSP’s produced in the cells of origin of the dorsal spinocerebellar tract by afferent volleys in Group la and Ib fibres (Eccles et al., 1961~).The neurones were deteriorated, and hence there was an initial rapid decline of the EPSP, as in Fig. 131, but thereafter there was a slowly declining component indicative of a prolonged transmitter action. Before deterioration had set in this prolonged transmitter action was adequate to evoke a second spike discharge. The existence of a small residual transmitter action is indicated also by the monosynaptic EPSP of a deteriorated motoneurone (Fig. 13E). Other instances of prolonged synaptic action are exhibited in the many situations where single afferent volleys produce repetitivedischarges of cells in the spinal cord ;for example by a Group Ta volley acting on intermediate neurones (Eccles et a/., 1956b), and by a cutaneous afferent volley acting on cells in the dorsal horn (Figs. 6F and 16) (McIntyre et al., 1956; Hunt and Kuno, 1959; Haapanen et al., 1958; Wall, 1959; Eccles et al., 1960, 1962a). Intracellular recording, as in Fig. 13 E-I, provides eviderce that this repetitive discharge is due to a depolarization that smoothly follows on from

EXCITATORY RESPONSES OF SPINAL NEURONES

21

the initial spike potential, and often not to later synaptic bombardments via interneuronal relays. However, in Fig. 16 the repetitive discharges undoubtedly are in part due t o the delayed synaptic bombardments. In all these types of neurones there has been no direct measurement of the electric time constant, so it is not possible to calculate the approximate time course of synaptic action as in Fig. 13A. Nevertheless 1.OT 3

1.05T

1.15TJ

1.2T

1,

1.35TI

. I

SP

1 msec

b

Fig. 16. Monosynaptic excitatory post-synaptic potentials (EPSP’s) and spikes of a C type interneurone (depth 2.3 mm) produced by single volleys of increasing size in SP nerve. The lower traces show the intracellular records, the upper ones the cord dorsum potentials recorded at L7 segmental level. Stimulus strength (relative to threshold as unity) is indicated for each record. Voltage calibration is for intracellular recording only. The resting potential was -60 mV. Each record is formed by the superposition of several faint traces. (From Eccles et al., 1962a.)

the general picture emerges that with most types of excitatory synapse a presynaptic impulse acts on the subsynaptic membrane causing it to produce an initial intense flow of depolarizing current which subsides in a very few milliseconds to a low residual current that may persist for 10 or more msec. When well developed, this residual synaptic action causes a single presynaptic volley to evoke a repetitive discharge, as is most strikingly shown with Renshaw cells. There is relatively little evidence relating to the ionic permeabilities that are responsible for the currents generating the EPSP’s of nerve cells, i.e. for the current flow through the E-synapse component of Fig. 13C. By electrophoretic injection through an intracellular electrode, large changes can be made in the ionic composition of motoneurones (Coombs et al., 1955b; Araki et al., 1961, 1962). These injections of anions or cations may cause a considerable decrease in the membrane potential, and the EPSP is then diminished correspondingly. Otherwise there is no appreciable change in voltage and time course, which contrasts with the large changes simultaneously observed in the spike potential, in the after-hyperpolarization following a spike Figs. 8 and 9) and in the inhibitory postsynaptic potential. Hence it has beenconcluded that the transmitter substance causes the subsynaptic membrane to become permeable to all these types of injected ions, the largest of which have diameters (in the hydrated state) more than double that of the hydrated K+ and CI- ions. In the light of the recent evidence (Takeuchi and Takeuchi, 1960) that anions are not appreciably concerned in generation of the end-plate potential (EPP), further investigation is desirable, particularly with isolated preparations where it will be possible to change the exlracellular ions in the way that was done in the investigations on the EPP. References p-29-31

22

J . C’. E C C L E S

lnvestigutions with repetitive stirwulution Repetitive synaptic action has been studied in detail by intracellular recording of the excitatory postsynaptic potentials (EPSP’s) generated monosynaptically in motoneurones, which provide a direct and quantitative measure of synaptic efficacy (Curtis and Eccles, 1960). Fig. 17A shows monosynaptic EPSP’s set up in a motoneurone by

Fig. 17. (A). Repetitive monosynaptic EPSP’s recorded intracellularly from a cat gastrocnemius motoneurone (cf Fig. 13) with a d.c. amplifier, the frequency of the maximum group I volleys from gastrocnemius nerve being shown to the left of each record. (B). Repetitive monosynaptic EPSP’s formed by superimposed traces at the indicated frequencies (cisec) when the EPSP had attained a steady state. (C). The EPSP’s partly illustrated in B are expressed as fractions of the mean size obtaining at 0.4 c/sec or slower, and plotted against the respective stimulus frequencies on a logarithmic abscissa1 scale. Above the frequency scale, the corresponding stimulus intervals are shown in rnsec. (From Curtis and Eccles, 1960.)

repetitive stimulation. It is seen that a steady state is attained after the first few EPSP’s, even over a wide range of frequencies; and it is found that this steady state is maintained for hundreds of responses. Evidently, there is an initial phase of adjustment of the EPSP to the steady state size characteristic of that frequency. By a superposition technique it is possible to measure accurately the sizes of EPSP’s during this steady state, as in the specimen records of Fig. 17B. When the sizes of the superimposed EPSP’s for the series that is partly illustrated are plotted against the frequency or the volley interval (scaled logarithmically), the points on the extreme right of Fig. 17C reveal that there is no appreciable change in the size of the EPSP’s until the frequency is in excess of 0.4/sec. There is a progressive depression as the frequency is raised to 5-l0/sec. With further increase in frequency, the EPSP increases to a maximum at about 50/:ec, being then almost as large as at the lowest frequencies. All the motoneurones so investigated have exhibited a trough at 5-20/sec and an increased EPSP as the frequency is raised to 50-100/sec, though usually it remains below the level observed at very low frequencies. Above 100/sec each successive EPSP is superimposed on the tail of the preceding responses so that its size has to be determined by a subtraction technique. It is shown in this way (Curtis and Eccles, 1960) that above 100/sec the size of the added EPSP

EXCITATORY RESPONSES OF SPINAL NEURONES

23

declines with increasing frequency, particularly above 200/sec. It may be assumed that the size of the EPSP is approximately proportional to the amount of transmitter that is acting on the neurone, so it is possible to obtain an approximate measure of the amount of transmitter that is produced by repetitive stimulation in unit time by multiplying the size of the EPSP during the steady state by the frequency. As so measured, the rate of liberation of transmitter increases as the frequency is raised to 300/sec; but above that frequency a plateau is attained, the rate of liberation of transmitter being then almost three times greater than for stimulation at 100/sec. Evidently, high frequencies of stimulation result in a remarkable process of mobilization of transmitter, whereby fresh supplies become available for release at a rate that prevents serious depletion. Since the potentiation during repetitive stimulation is dependent on increased mobilization of transmitter in the presynaptic terminals, it is not unexpected that large differences sometimes are observed between the potentiation for different types of synapses on the same neurone. The neurones of the ventral spino-cerebellar tract (Eccles et al., 1961b) regularly exhibit a large absolute potentiation of the EPSP’s monosynaptically produced by Group Ib afferent volleys (up to 200% at 100/sec in Fig. 18A), whereas with Group la afferent volleys (Fig. I8B) there is merely a relative potentiation at frequencies of about SO/sec which is even less well developed than in Fig. 17C. With some interneurones in the intermediate nucleus there is also monosynaptic innervation by Group Ia and Ib afferent volleys, as shown i n the threshold series of Fig. 18C. In Fig. 18D there is at high frequencies a large potentiation of the later Ib component of the combined EPSP and little change i n the earlier l a component. It is tempting to conclude that Ib afferent fibres differ from l a in having synapses with a much more efficient mobilization process at high frequencies. However, when Ia and Ib fibres converge onto the same neurone of the dorsal spino-cerebellar tract, there is little if any more potentiation for the EPSP’s set up by the Ib afferent volleys than for the Ia, both giving curves like Fig. 17C (Eccles et al., 1961~). An observation even more disturbing to any attempt at generalization is that the monosynaptic Group l a synapses on respiratory motoneurones of the thoracic cord regularly exhibit a large frequency potentiation (Sears, 1963), which in Fig. 18E is as large as for the Ib synapses in Fig. 18A. Recently Fadiga and Brookhart (1962) have described another example of differing frequency-potentiation for two monosynaptic pathways onto the same motoneurone. Frequency-potentiation is very highly developed for the pyramidal monosynaptic pathway onto motoneurones of the forelimb muscles of the baboon, but it is not known if the monosynaptic Group la synapses have poorly developed potentiation as in Fig. 18B (Landgren et al., 1962). Evidently, frequency-potentiation is as yet an unpredictable phenomenon, though there can be no doubt about its functional importance in enabling synapses to respond effectively to the prolonged repetitive stimulation that must be occurring under conditions of normal activity. Post-activation potentiation The synapses concerned in monosynaptic activation of motoneurones exhibit very Rrfrrences p . 29-31

h,

P

PBST

C D

111msec 11111111 Fig. 18. Monosynaptic EPSP's recorded at the steady state during a wide range of frequencies, as indicated in c/sec for each series of superimposed traces (A, B, D and E). A and B give respectively group Ib and Ia activation of the same cell of the VSC tract. The lower traces show the afferent volleys entering the spinal cord through L7 dorsal root. (From Eccles er al., 1961b.) C shows that graded stimulation at the strengths indicated relative to threshold gives evidence of the converging Ia and Ib activation of a neurone in the intermediate nucleus of the spinal cord with stimulation of the posterior biceps-semitendinosus nerves. In D maximum Ia Ib volleys are applied at the frequencies indicated. (From Eccles, Kostyuk and Schmidt, unpublished records.) In E are monosynaptic EPSP's of a thoracic motoneurone innervating the external intercostal muscle. (From Sears, 1963.)

EXCITATORY RESPONSES OF SPINAL NEURONES

25

Fig. 19. (A-C). Curves showing early post-tctanic potentiation following brief conditioning tetani. The points plot the potentiation of the steepest part of the rising slope of the EPSP's as a fraction of the mean control. Intracellular records from the biceps-semitendinosus motoneurone (resting potential -60 mV) are shown as insets, the first being the response at about 0.2 sec post-tetanically, the second the control. Note spike origin in A and B. Conditioning tetani of 40 impulses, the respective frequencies being 640/sec, 400/sec and 100/sec for A, B and C and the durations being indicated by the initial hatched blocks. (D-G). Post-tetanic potentiation of monosynaptic EPSP's (see inset records of potentiated and control EPSP's and their first differentials) following long conditioning tetani (3200 volleys). The respective conditioning tetani for D, E, F, G were 640/sec for 5 sec, 400/sec for 8 sec, 200/sec for 16 sec and 100/sec for 32 sec, as shown by the hatched blocks. Thus each horizontal row AD, BE and CG has the same frequency of the conditioning tetanus. Note the large difference in time scales (about 100) between A-C on the one hand and D-G on the other. (From Curtis and Eccles, 1960.)

interesting post-activation responses. In the series (Curtis and Eccles, 1960) plotted in Fig. 19A, B and C and illustrated in the inset records, the conditioning tetani of 40 volleys were at 640, 400 and 100/sec respectively and ceased at zero on the time scale. Only one testing stimulus was applied after each conditioning tetanus, and the sizes of the EPSP's were measured as the maximum slopes of the rising phases, as illustrated in the inset records. It will be seen that a conditioning tetanus of 400/sec is as effective as that at 640/sec, but after a tetanus at 100/sec there is no potentiation relative to the control EPSP. Since the total duration of the conditioning tetanus was only 0.4 sec in Fig. 19C, and since the potentiation in A and B had declined very little by 0.4 sec after the tetanus, it is evident that the increased temporal spread of the conditioning References p . 29-31

26

J. C. E C C L E S

volleys is not responsible for the absence of potentiation in Fig. 19C. A conditioning tetanus of 200/sec gives a small potentiation, whereas 100/sec is often followed by depression. Thus it is frequency rather than number of impulses that is of particular significance in producing post-tetanic potentiation after brief conditioning tetani. These potentiations or depressions of the EPSP after brief conditioning tetani correspond to the post-tetanic potentiations or depressions of monosynaptic reflexes under comparable conditions (Eccles and Rall, I95 1 ; Lloyd, 1952). It seems that there is a correlation between the rate of transmitter liberation during the steady state of responses to repetitive stimulation on the one hand (Figs. 17 and 18), and the potentiations or depressions that occur i n the testing EPSP just after the cessation of these tetani on the other (Fig. 19). In discussing this correlation it is convenient to use the concept of ‘available transmitter’, a measure of which is provided by the amount of transmitter that is liberated by a given size of presynaptic impulse. For example, on termination of repetitive synaptic stimulation at 300/sec or higher, there is a sudden cessation of the liberation of transmitter which has been running at the maximum attainable rate per unit time; albeit with an output per impulse much below the level for a single impulse under resting conditions. It would therefore be expected that, immediately after cessation of this maximum drain of transmitter from the presynaptic terminal, the unchecked mobilization process would result in an accumulation of its available transmitter above the resting level; as a consequence, a testing impulse would cause an increased liberation of transmitter, so giving the potentiated EPSP’s seen in Fig. 19A and B. The time course of this early post-tetanic potentiation indicates that the available transmitter goes on accumulating for as long as 200 msec after repetitive synaptic stimulation ceases, and that it subsequently declines slowly over several seconds. During lower frequencies of synaptic stimulation (e.g. 200/sec), there is a lower rate of liberation of transmitter; and corresponding to this lower rate of mobilization there is less potentiation after cessation of stimulation. At frequencies below 100/sec there is a prolonged phase of depressed EPSP’s, which presumably is related t o the depression observed during low frequency tetani and which is an index of depletion of available transmitter (Fig. 17B and C). After a brief conditioning tetanus at 100/sec there is apparently an approximate balance between the potentiation and the depressant action (Fig. 19C). After long conditioning tetani the potentiation of synaptic transmission, as indicated by the size of a testing EPSP, rises much more slowly and is much more prolonged than after the brief tetani, its total duration being measured in minutes instead of seconds. In this series the potentiation was much the same after 3200 impulses at 640 or 400/sec (Fig. 19D, E), corresponding to the similar potentiations following 40 impulses at these two frequencies (Fig. 19A and B). Contrary to the findings with brief tetani (Fig. 19C), there was a considerable potentiation after 3200 impulses at 100/sec (Fig. 19G), while following tetanization at 200/sec there was an intermediate level of potentiation (Fig. 19F). In all these respects the observations with EPSP’s correspond closely with previous investigations employing the monosynaptic reflex as a test of potentiation (CA Lloyd, 1949; Eccles and Rall, 1951). However, a difference arises when the relative amounts of potentiation are compared. Usually reflexes display

EXCITATORY RESPONSES OF SPINAL NEURONES

27

much larger potentiations because the changes in synaptic efficacy are sampled by the responses of a population of motoneurones of which many are brought into the discharge zone by a relatively small increase of EPSP. On the other hand the size of the EPSP of a single unit can be assumed to be proportional to the synaptic efficacy, which is thus shown to be potentiated to no more than 1.5-2.0 times the control. The most likely explanation for the potentiation that follows a long high-frequency tetanus is that under such conditions there is a prolonged after-hyperpolarization of the presynaptic fibres, with as a consequence an increase in the presynaptic spike potential (Wall and Johnson, 1958; Eccles and KrnjeviC, 1959a, b); such an increased presynaptic spike can be expected to have an increased synaptic excitatory action (Lloyd, 1949; Eccles and Rall, 1951; Takeuchi and Takeuchi, 1962; Eccles et al., 1962b; Hubbard and Schmidt, 1963), and the increase in the monosynaptic EPSP of motoneurones has a comparable time course (Eccles et a/., 1959; Eccles and KrnjeviC, 1959b). The alternative explanation is that there is an increased availability of transmitter, which is strongly suggested for the neuromuscular junction by the greatly increased frequency of miniature end-plate potentials that is observed post-tetanically (Liley, 1956; Brooks, 1956; Hubbard, 1963). This latter explanation is, of course, specially applicable to the post-tetanic potentiation after brief tetani, such as 40 impulses at 600/sec, because under such conditions there is unlikely to be an appreciable after-hyperpolarization of the nerve terminals (CJ Fig. 7A of Eccles and KrnjeviC, 1959a).

PRESYNAPTIC

I’RESY NAPTlC MEMBRANE MEMBRANE

SYNAPTIC CLEFT

. *... *

. . * *

.

*

Trn*.IC.IIT

TRANSMITTER TER MOLECULES JLES

‘

POST SYNAPTIC^^^^^^^^^ MEMBRANE

POSTSYNAPTIC

Fig. 20. Schematic representation of a portion of a synaptic cleft with synaptic vesicles in close proximity in the presynaptic terminal, and one actually discharging the transmitter mdecules into the syna; tic cleft. Some of these mo!ecules are shown combined with receptor sites on the postsynaptic membrane with the consequent opening u p of pores through that membrane.

Fig. 20 shows diagrammatically the detailed events which are presumed to occur when an impulse reaches a presynaptic terminal, and which we would expect to see if electron microscopy can be developed to have sufficient resolving power. Some of the synaptic vesicles are in close contact with the membrane and one or more are caused by the impulse to eject their contained transmitter substance into the synaptic cleft. Diffusion across and along the cleft, as shown, would occur in a few microseconds for distances of a few hundred Angstroms. Some of the transmitter becomes Rrferenc.es p . 29-31

28

J. C. E C C L E S

momentarily attached to the specific receptor sites on the postsynaptic membrane with the consequence that there is an opening up of fine channels across this membrane, i.e. the postsynaptic membrane fronting the synaptic cleft momentarily assumes a sieve-like character. The cations, sodium and potassium move across the membrane thousands of times more readily than normally; and this intense ionic flux gives the current that depolarizes the cell membrane and produces the EPSP. This current flows for only about one thousandth of a second because the transmitter is eliminated from the synaptic cleft by diffusion into the surrounding extracellular spaces and by enzymic destruction, at least when acetylcholine is the transmitter substance. It will be appreciated that the structural and functional arrangements account for transmission across a synapse being in one direction only. This diagram serves to illustrate further fundamental problems of synaptic activation that are being investigated, but it is only possible to make brief mention of them in serial order. There is firstly the identification of the specific transmitter substance, which is as yet unknown for the vast majority of synapses in the central nervous system. A second major problem concerns the mechanism by which a presynaptic nerve impulse causes the ejection of packets of transmitter into the synaptic cleft. A third concerns the way in which other synaptic vesicles with their contained transmitter are moved up into the firing line to replace the emptied vesicles. Probably the emptied vesicles recharge themselves with transmitter by virtue of their contained enzyme systems. The urgency of these processes of mobilization and manufacture of transmitter will be appreciated when it is realized that the total transmitter i n synaptic terminals is sufficient for only a few minutes of synaptic activity at a normal intensity. The efficiency of these processes was briefly considered above. On the postsynaptic side there are problems relating to the nature of the receptor sites for attachment of the transmitter molecules and to the way in which the ionic channels are opened up in the membrane. Finally, there are very important problems relating to the accommodation of the postsynaptic membrane and its ability to discharge repetitively in response to a prolonged depolarization when produced by applied current and by asynchronous synaptic bombardment. These various fields come within the purview of other contributors to this symposium. SUMMARY

An initial brief reference t o the extracellular and intracellular ionic composition for motoneurones serves as a basis for an account of their excitatory responses, both the spike potential and the excitatory postsynaptic potential. The spike potential is discussed in relation to the complex morphology of the motoneurone. Recent investigations have corroborated the original investigations in which the 2 main spike compo nents were attributed on the one hand to the initial segment of the axon and the axon hillock, the IS spike, and, on the other hand to the soma and dendrites, the SD spike. The excitatory properties of these 2 components of the motoneurone are discussed, as also is the propagation of impulses along dendrites. Alterations of the internal sodium and Fotassium concentrations affect the rising and falling phases of the spike

E X C I T A T O R Y R E S P O N S E S OF S P I N A L N E U R O N E S

29

potential exactly as would be predicted from the ionic theory of Hodgkin and Huxley. Following the spike potential of motoneurones and many other types of neurone there is a prolonged after-hyperpolarization. Recent investigations have fully corroborated the original suggestion that this potential is entirely attributable to a raised potassium conductance of the cell membrane, particularly of the soma-dendritic membrane. Excitatory synapses act by liberating a chemical transmitter that induces in the subsynaptic membrane a brief inward current that depolarizes the postsynaptic membrane to give the excitatory postsynaptic potential. A spike potential is initiated if the depolarization exceeds a critical level. With many synapses the initial brief synaptic current is followed by a prolonged low current attributable to a residual transmitter action. The ionic mechanism generating the postsynaptic current has an equilibrium potential at about zero membrane potential. Possibly on analogy with the neuromuscular junction the ionic mechanism of synaptic excitation is restricted to a raised membrane conductance to the cations sodium and potassium. With repetitive activation of synapses it is shown that there tends to be a depression of transmitter output at low frequencies (0.5-l0/sec), but at higher frequencies there is an increased effectiveness, frequency-potentiation, which may be very powerful with some types of synapse. The level of frequency-potentiation is a measure of the effectiveness of transmitter mobilization in the activated presynaptic terminals. Following a brief repetitive stimulation there is a post-tetanic increase in synaptic action, which is attributable to the residual effects of transmitter mobilization. Following long repetitive stimulation the post-tetanic potentiation is attributable largely to the increased size of the presynaptic impulses during the prolonged phase of after-hyperpolarization of the tetanized presynaptic fibres.

REFERENCES

P. G., OSCARSSON, O., A N D OSHIMA, T., (1962); Injection of alkaline ARAKI,T., ITO, M . , KOSTYUK, cations into cat spinal motoneurones. Nature (Lond.), 196, 1319-1 320. O., (1961); Anion permeability of the synaptic and non-synaptic ARAKI,T., ITO,M., AND OSCARSSON, motoneurone membrane. J . Physiol. (Lond.), 159, 410435. ARAKI,T., A N D OTANI,T., (1955); Response of single motoneurons to direct stimulation in toad’s spinal cord. J . Neiivophysiol., 18, 472485. C. A., (1962); Membrane currents in spinal motoneurons associated ARAKI,T., AND TERZUOLO, with the action potential and synaptic activity. J . Neuvophysiol., 25, 772-789. J. S., AND ECCLES,J. C., (1953); Intracellular recording from antidromically BROCK,L. G., COOMBS, activated motoneurones. J . Physiol. (Lond.), 122, 429461. V. B., (1956); An intracellular study of the action of repetitive nerve volleys and of botulinum BROOKS, toxin on miniature end-plate potentials. J. Physiol. (Lond.), 134, 264-277. CALDWELL, P.G., HODGKIN, A. L., KEYNES, R. D., AND SHAW,T. I., (1960a); The effects of injecting ‘energy-rich’ phosphate compounds on the active transport of ions in the giant axons of Loligo. J . Physiol. (Lond,), 152, 561-590. P. C., HODGKIN, A. L., KEYNES,R. D., AND SHAW,T. I., (1960b); Partial inhibition of CALDWELL, the active transport of cations in the giant axons of Loligo. J . Physiol. (Lond.), 152, 591-600. J. S., CURTIS,D. R., AND ECCLES,J . C., (1957a); The interpretation of spike potentials of COOMBS, motoneurones. J . Physiol. (Lond.), 139, 198-23 1. COOMBS, J. S., CURTIS,D. R., AND ECCLES,J. C., (19S7b); The generation of impulses in motoneurones. J. Physiol. (Lond.), 139, 232-249.

30

J. C . E C C L E S

COOMBS, J. S., ECCLES,J. C., A N D FATT, P., (1955a); The electrical properties of the motoneurone membrane. J . fhysiol. (Lond.), 130, 291-325. COOMBS, J. S., ECCLES,J . C., A N D FATT,P., (1955b); Excitatory synaptic action in motoneurones. J . fhysiol. (Lond.), 130, 374-395. CURTIS,D. R., AND ECCLES, J. C., (1959); The time courses of excitatory and inhibitory synaptic actions. J . fhysiol. (Lon(/.), 145, 529-546. CURTIS,D. R., AND ECCLES,J. c.,(1960); Synaptic action during and after repetitive stimulation. J . fhysiol. (Lond.), 150, 374-398. CURTIS, D. R., ECCLES, J. C., AND LUNDBERG, A., (1958); Intracellular recording from cells in Clarke’s column. Acta physiol. scand., 43, 303-314. ECCLES,J. C., (1957); The Physiology of’Nerve Cells. Baltimore. Johns Hopkins Press. ECCLES,J . C., (1961); The mechanism of synaptic transmission. Ergehn. fhysiol., 51, 299-430. ECCLES,J. C., ECCLES,R. M., AND FATT,P., (1956a); Pharmacological investigations on a central synapse operated by acetylcholine. J . Physiol. (Lond.), 131, 154-169. ECCLES,J. C., ECCLES,R. M., IGGO, A., AND LUNDBERG, A., (1961a); Electrophysiological investigations on Renshaw cells. J . fhysiol. (Lunrl.), 159, 461-478. ECCLES, J. C., ECCLES,R. M., A N D LUNDBERG, A., (1958); The action potentials oftheu-motoneurones supplying fast and slow muscles. J. fhysiol. (Lond.), 142, 275-291. ECCLES, J. C., ECCLES,R. M., A N D LUNDBERG, A., (1960); Types of neurone in and around the intermediate nucleus of the lumbo-sacral cord. J . fhysiol. (Lond.), 154, 89-1 14. ECCLES, J. C., FATT,P., A N D KOKETSU, K., (1954); Cholinergic and inhibitory synapses in a pathway from motor-axon collaterals to motoneurones. J . fhysiol. (Lond.) , 216, 524-562. ECCLES,J. C., FATT,P., A N D LANDGREN, S., (195613); The central pathway for the direct inhibitory action of impulses in the largest afferent nerve fibres to muscle. J . Neurophysiol., 19, 75-98. ECCLES, J. C., HUBBARD, J. I., AND OSCARSSON, 0..(1961b); Intracellular recording from cells of the ventral spino-cerebellar tract. J . Physiol. (Lond.), 158, 486-5 16. ECCLES,J. C., KOSTYUK, P. G., AND SCHMIDT, R. F., (1962a); Central pathways responsible for depolarization of primary afferent fibres. J . fhysiol. (Lond.), 161, 237-257. ECCLES,J. C . , KOSTYUK, P. G., A N D S C H M I ~R. T , F., (1962b); The effect of electric polarization of the spinal cord on central afferent fibres and on their excitatory synaptic action. J . fhysiol. (Lon(/.), 162, 138-150. ECCLES,J . C., A N D KRNJEVIC, K., (1959a); Potential changes recorded inside primary afferent fibres within the spinal cord. J . fhysiol. (Lond.), 149, 250-273. ECCLES,J. C., A N D KRNJEVIC, K., (1959b); Presynaptic changes associated with post-telanic potentiation in the spinal cord. J . fhysiol. (Lon(/.), 149, 274-287. ECCLES,J. C., KRNJEviC, K . , AND MILEDI,R., (1959); Delayed effects of peripheral severance of afferent nerve fibres on the efficacy of their central synapses. J . fhysiol. (Lon(/.), 145, 204-220. ECCLES,J. C., OSCARSSON, O., AND WILLIS,W. D., (1961~); Synaptic action of Group I and I 1 afferent fibres of muscle on the cells of the dorsal spino-cerebellar tract. J . fhy.siu1. (Lonil.), 158, 5 17-543. ECCLES,J. C., AUD RALL,W., (1951); Repetitive monosynaptic activation of niotoneurones. Proc. roy. SOC.B, 138, 475498. ECCLES, J. C., A N D SHERRINCTON, C. S., (1931); Studies on the flexor reflex. VI. Inhibition. froc. m y . SOC.B, 109, 91-1 13. FADIGA, E., A N D BROOKHART, J. M., (1962); Interactions of excitatory postsynaptic potentials generated at different sites on the frog motoneuron. J . Neurophysiol., 25, 790-804. FATT,P., (1957a); Electric potentials occurring around a neurone during its antidromic activation. J. Neurophysiol., 20, 27-60. FATT, P., (l957b); Sequence of events in synaptic activation of a motoneurone. J . Neiiruphysiul., 20, 61-80. FUORTES, M. G . F., FRANK, K., AND BECKER, M. C., (1957); Steps in the production of niotoneuron spikes. J . gen. Physiol., 40, 735-752. GRANIT, R., HENATSCH, H. D., A N D STEG,G., (1956); Tonic and phasic ventral horn cellsdifferentiated by post-tetanic potcntiation in cat extensors. Acta physiol. scand., 37, 114-126. GRANIT,R., PHILLIPS, C. G., SKOCLUND, S., AND STEG,G., (1957); Differentiation of tonic from phasic alpha ventral horn cells by stretch, pinna and crossed extensor reflexes. J . Neiiruphysiol., 20, 470481. HAAPANEN, L., KOLMODIN, G. M., AND SKOGLUND, C. R., (1958); Membrane and action potentials of spinal interneurones in the cat. Actn physiol. scand., 43, 315-348.

EXCITATORY RESPONSES OF S P I N A L N E U R O N E S

31

HODCKIN,A. L., (1958); Ionic movements and electrical activity in giant nerve fibres. Proc. roy. SOC.B, 148, 1-37. HODGKIN, A. L., AND KEYNES, R. D., (1955); Active transport of cations in giant axons from Sepia and Loligo. J . Physiol. (Lond.), 128, 28-60. HUBBARD, J. I., (1963); Repetitive stimulation at the mammalian neuromuscular junction and the mobilization of transmitter. J. Physiol. (Lond.), 169, 641-662. HUBBARD, J. I., AND SCHMIDT, R. F., (1963); An electrophysiological investigation of mammalian motor nerve terminals. J . Physiol. (Lond.).. 166,. 145-167. HUNT,C. C . , AND KUNO,M.,-(1959); Properties of spinal interneurones. J . Physiol. (Lond.), 147, 346-363. ]TO, M., A N D OSHIMA, T., (1964); The extrusion of sodium from the spinal motoneurones of the cat. In course of publication. S., PHILLIPS, C. G . , AND PORTER,R., (1962); Minimal synaptic actions of pyramidal imLANDGREN, pulses on some u-motoneurones of the baboon’s hand and forearm. J . Physiol. (Lond.), 161, 91-111. LILEY,A. W., (1956); An investigation of spontaneous activity at the neuromuscular junction of the rat. J . Physiol. (Lond.), 132, 650-666. LLOYD,D. P. C., (1949); Post-tetanic potentiation of response in monosynaptic reflex pathways of the suinal cord. J . pen. Phvsiol.. . , 33. , 147-170. LLOYD;D.P. C., (1952); Electrotonus in dorsal nerve roots. Cold Spr. Harb. Syrnp. qiranf. Biol., 17, 203-219. DE NO, R., (1938); Synaptic stimulation of motoneurons as a local process. J . Nertrophysiol., LORENTE 1, 195-206. LOXENTE DE NO, R., (1947); Action potential of the motoneurons of the hypoglossus nucleus. J . cell. comp. Physiol., 29, 207-288. LORENTE D E N6, R., (1953); Conduction of impulses in the neurons of the oculomotor nucleus. The Spinal Cord. Ciba Foundation Symposium. London. Churchill (p. 132-173). MCINTYRE, A. K., MARK,R. F., AND STEINER, J., (1956); Multiple firing at central synapses. Nature (Lonrl.), 178, 302-304. RENSHAW, B., (1946); Central effects of centripetal impulses in axons of spinal ventral roots. J . Neurophysiol., 9, 191-204. SEARS, T. A., (1963); The properties and distribution of the synapses made by monosynaptic afferents in the intercostal nerves. J . Physiol. (Lond.),in course of publication. C. S., (1925); Remarks on some aspects of reflex inhibition. Proc. roy. SOC.B, 97, SHERRINGTON, 519-545. TAKEUCHI, A., A N D TAKEUCHI, N., (1960); On the permeability of the end-plate membrane during the action of transmitter. J . Phjwio/. (Lond), 151, 52-67. N., (1962); Electrical changes in pre- and postsynaptic axons of the TAKEUCHI, A., AND TAKEUCHI, giant synapse of Loligo. J . gen. Physiol., 45, 1181-1 193. TERZUOLO, C . A., A N D ARAKI,T., (1961); An analysis of intra- versus extracellular potential changes associated with activity of single spinal motoneurons. Awz. N . Y. Acad. Sci., 91, 547-558. WALL,P. D., (1959); Repetitive discharge of neurons. J . Neurophysiol., 22, 305-320. A. R., (1958); Changes associated with post-tetanic potentiation of a WALL,P. D., AND JOHNSON, monosynaptic reflex. J. Neiirophysiol., 21, 149-158. DISCUSSION

VERHEY:1 would disagree with your following description: ‘current out from the inside of the nerve cell through the membrane to the outside which depolarises the membrane’. Would not it be better to describe depolarization as the result of unloading of the membrane as a capacitor? ECCLES:I am afraid I do not understand the basis of your disagreement. With reference to the formal electrical diagram of a nerve membrane this shows that adequate

32

DISCUSSION

current flows both from the resistance elemcnt and through the capacitor which are in parallel. The flow of current through both of these channels serves t o diminish the potential across the membrane. GELFAN: I think Sir John has left the impression that the ionic hypothesis is something that is accepted by everyone, and that it is the only way in which you could explain either the resting membrane potential or the genesis of nerve impulses. There are quite a number of investigators throughout the world who d o not explain either the resting membrane potential or the impulse itself on a basis of the inward flow of sodium and the outward flow of potassium. I a m sure he must be familiar with the literature, that there is a number of controversial points regarding the initial segment. There is some evidence that the initial segment would not see the sodium o r potassium ions at all. It would be seeing something else, and it may be rather indifferent t o either the sodium or the potassium. I don’t believe it is necessary to cite the literature about this. In connection with the inward flow of sodium a t the time of the initiation of the impulse I might remind a good number here, that not in the spinal cord but in peripheral nerve, another mechanism works. Nerves exposed to a complete lack of outside sodium, the sodium being replaced by ammonium salts, were conducting very well. The main problem with the use of micro-electrodes, although they have been a tremendous advance for us, is that we d o not see actually where the micro-electrode is. ECCLES:I chose to neglect this literature with which I am perfectly familiar because I think it is an unnecessary complication and is not helpful in my attempts to explain the mode of operation of synapses. I found that with the basis of the ionic theory of Hodgkin and Huxley I can build a coherent theory of synaptic action that has stood up to all experimental tests so far applied. N o one has produced a comparable theory of synaptic action that is based upon some other theory of the resting and action potentials of cell membranes. With regard t o your statement that we d o not actually see where the micro-electrode is, I reply that we can by much circumstantial evidence be sure that the micro-electrode is actually implanted in a nerve cell which we can identify. We are in fact just using more subtle ways of seeing.

JENIK: Terzuolo and Araki have observed considerably prolonged spike potentials in the dendrites. How might these be explained? By an increase of membrane time constant o r by influence of sub-synaptic membrane?

ECCLES:I d o not think that either of your suggestions give the likely explanation of the slow decline of spike potentials in dendrites. My own suggestion is that there is in dendrites a lower level of potassium conductance during the falling phase of the spike. However, before we take these slow potentials too seriously 1 would like to have many more experimental observations made. 1 think that Terzuolo and Araki actually have seen this phenomenon only twice.

EXCITATORY RESPONSES OF S P I N A L NEURONES

33

HORSTFEHR: If the axonic zone of the motoneuron (the IS-zone in your drawing) fires first because of its threshold difference, is it not possible that in this zone there is, under physiological conditions, a higher anion (e-)-concentration than in the dendritic trees, so that one could regard the neuron as an electrical dipole in which the initial segment of the axon is the negative pole and the dendritic end the positive one? Now my question is whether, for the neuron, the concept of dipole (or negentropic system) is right. ECCLES:If I have understood your question correctly I can answer in the negative to the effect that our observations show that there is no appreciable resting depolarization in the nerve cell between dendrites and the initial segment of the axon. Certainly the much lower threshold of the initial segment is not explicable on this basis because it is restricted to a quite sharply defined zone. The membranes of the initial segment and the adjacent soma membrane behave exactly like the dendritic membrane. WIESENDANGER: Renshaw and you have reported that the initial segment can be re-excited by the soma-dendritic depolarization. In this respect I would like to ask: is the interpretation still valid that only those axons which convey the impulse can be re-excited ? ECCLES:You have raised quite an interesting point here, and all I can say in reply is that 1 believe the initial explanation of this current discharge to be still valid. The phenomenon is observed in only about 2% of the neurons and the explanation is that i n these neurons there is such a long delay between the initial segment spike and the soma dendritic spike that the initial segment has recovered excitability while the soma dendritic spike is still fully developed. As a consequence the initial segment is depolarized by current flow into the soma dendritic membrane and an impulse is generated that runs antidromically down the axon. : Could the difference in electrical properties between conducting and CORNER sub-synaptic membranes be explained on the basis of the absence in the latter of a bound substance needed to produce the permeability changes, thus requiring that it be supplied by the pre-synaptic volley?

ECCLES:1 fear I do not understand this question. Possibly you mean that the actual chemical mechanism responsible for producing the high sodium conductance of the spike is not present in the sub-synaptic membrane. This would certainly be the case for electrically inexcitable membrane, but the presynaptic volley does not in fact influence spike generation by supp!ementing some of this postulated sodium conductance mechanism. All it does is to depolarize the post-synaptic membrane so that an impulse is generated at the lowest threshold zone. ROMANES:What is the volume of extracellular space required for your hypothesis of ionic balance? We find only 200 A or so between the cell membrane and the sur-

34

DISCUSSION

rounding glia in electronmicroscopical pictures and it seems that you cannot assume the cells to lie in a structureless conducting medium. Has this any significance?

ECCLES:I agree with your statement that as an electron microscopist there is only a lattice of 200 spaces between nerve cells and between nerve cells and glia. The electrical conductance of these lattice spaces is adequate for providing return circuits for all synaptically induced currents. Possibly the special glia cells, astrocytes, may have very good electrical conductance which would enable them to supplement the conductance of the fine lattice of extracellular spaces. No doubt the extracellular space may have a conductance as low as 1/10 of a structureless extracellular medium, but such a relatively low conductance is still perfectly adequate.

35

Maintained Firing of Motoneurones during Transmembrane Stimulation R A G N A R GRANIT The Nobel Institute fur Neurophysiology, Karolinska lnnstitittet, Stockholm

When the confluence zones of subsynaptic currents stimulate motoneurones to discharge repetitively, they will do so at a frequency which is some function of current strength. If rate of discharge varies but little with current strength, the motoneurone has little power of amplification, if, on the other hand, slight changes of current strength greatly influence firing rate the motoneurone will serve as an efficient amplifier of input. Motoneurones have been studied a great deal but this function has curiously been neglected. Yet knowledge of it is fundamental for integrative work. A fair imitation of this property of motoneurones can be obtained by stimulating them from the inside by the tip of the micro-capillary in the manner introduced by Araki and Otani (1955). The present paper deals with this problem and is based on recent work in our laboratory by a team consisting of Granit, Kernell, Shortess and Smith (see references). Special references will have to be looked up in the original papers. The work was begun at a time when there was a temporary shortage of cats in the laboratory and so it was decided to use the rat as our standard preparation. The findings were in the end also verified on a number of cats. Bradley and Somjen (1961) have recently shown that for many purposes the rat is quite serviceable. Especially we have found this animal useful when both dorsal and ventral roots for the hind limb segments are cut because these segments in the rat are supplied by blood vessels from the thoracic end and not by the roots, as are the corresponding segments in the cat. Our preparations were anaesthetized with pentobarbitone, 55 mg/kg, and curarized. They were maintained on artificial respiration by a mixture of oxygen with 1 % carbon dioxide. Technical details are given in our first paper (Granit et al., 1963c) and i n the work of Bradley and Somjen. The stimulating circuit was that of Araki and Otani (1955) and KCI micro-capillaries were used. The rat as a preparation proved to be of interest also because most intracellularly recorded spikes, elicited by the antidromic route, were practically without afterhyperpolarization (Fig. Id) differing in this respect from the majority of cat motoneurones, so well known from the work of Eccles and his colleagues (see e.g., Eccles, 1957). Partly this was due to the prominent delayed depolarization which followed the action potential both when it was fired antidromically (Fig. 1 a-d) and when it was elicited by an intracellular shock (Fig. 1 e-g). We did not succeed in eliciting References p . 41

36

R. GRANI'I

delayed depolarization separately. It merely occurred in sequence to a full-size spike. The IS-spike was never followed by delayed depolarization. A considerable number of experiments in our first paper (Granit et al., 1963~)were devoted to a study of delayed depolarization. The outcome of this work was ultimately summarized in the hypothesis that delayed depolarization represents activation of the dendrites. I shall not review these experiments because Kernell has since tried to investigate this hypothesis, and in the discussion he will concentrate on this aspect of our findings and present his own later work. a

h

b

C

I

d

k

, 0 . 2 5ec , Fig. 1. Intracellular records from rat motoneurones. a-d, antidromic discharge of spike of 81 mV at different amplifications. Times in msec for a-c, calibrations 20 mV; note, that spikes b-d go beyond range of oscillograph deflexion. e-g, another spike stimulated from the inside by shock of 0.5 msec (e) and 0.12 msec(f), the former repeated in g with time (msec)replacing current recorder. h-I, beginning of near-threshold repetitive discharge as in I, of which h-k show variations of initial frequency (time in msec) in different trials. Note development of after-hyperpolarization (Granit et al., 1963c.)

Delayed depolarization is taken up in this connexion because it has the interesting property of disappearing when repetitive firing is elicited by maintained intracellular stimulation. This is shown in Fig. 1 h-1. Repetitive firing is here elicited at stimulus strengths near threshold so that the initial frequency of discharge varies a little from case to case. The initial spikes were separately recorded on the sweep circuit in records h to k, and on the expanded time scale it is seen that during transmembrane stimulation delayed depolarization is gradually replaced by after-hyperpolarization. In fact, the rat spike, fired repetitively from the inside by transmembrane currents, now behaves like (according to the Canberra group) most cat spikes seem to do howsoever fired, i.e. they now show good after-hyperpolarization. This made us look at the cat motoneurones with renewed interest and, actually, among them, too, a considerable number failed to give after-hyperpolarization when fired antidromically. Kernell has since made further observations dealing with the relative frequency and properties of cat motoneurones that discharge antidromic spikes succeeded by delayed depolarization.

M A I N T A I N E D F I R I N G OF MOTONEURONES

37

It is clear that the employment of the firing mechanism must be different when the motoneurone discharges repetitively as compared with how it operates in response to an antidromic invasion. The simplest explanation seems to be that the rhythmically discharging motoneurone employs a very much enlarged firing area and that it, only when it operates in this manner, can deliver (in rats) the phase of after-hyperpolarization which develops early during transmembrane stimulation. The reason why many, if not most, cat motoneurones are in this state before rhythmic firing has begun, remains obscure. When the transmembrane current is allowed to act for several seconds a number

Fig. 2. Intracellular stimulation of rat motoneurone to discharge repetitively by current strengths

marked on the right. Note change of sensitivity of current recording beam of oscillograph at 7.7 nA. For the strongest current spike has diminished. On the left, initial spikes on sweep and time in msec (Granit et al., 1963b).

of motoneurones respond repetitively throughout the period of stimulation. The majority merely responds phasically and when current strength is increased the rates of discharge in the burst merely increases without the latter expanding much in duration. The discharge of a tonically firing motoneurone is shown in Fig. 2. The initial spikes have also been recorded on the fast sweep circuit. These are shown on the left in the figure, and the increase of the after-hyperpolarization early in transmembrane stimulation is conspicuous also in these records. The question of why all of the motoneurones could not be forced to respond tonically is difficult to answer. Intracellular recording is beset with pitfalls and it is easy to ascribe failure of lasting repetitive firing to technical faults. On the other hand, we had a number of penetrations which to all appearance were excellent in the sense that the spikes were large and well-maintained. When such motoneurones refuse to respond tonically t o long-lasting intracellular stimulation we hesitate to ascribe this to a faulty technique. It seems more reasonable to us to accept the best findings at References p. 41

38

R. G R A N I T

their face value and to conclude that there actually are motoneurones which ‘adapt’ so quickly that maintained firing is quickly suppressed. This brings us to the general problem of ‘adaptation’ of the motoneurone membrane. A good introduction t o this question is the detailed analysis of the records of Fig. 2,

Fig. 3. Frequency of discharge plotted against current strength for motoneurone of Fig. 2. Curve I , slope ccn;tant 5.9 imp/sec/nA, dcrives from 1/3 sec following first interval; curve 2, same after 1.3 sec; curve 3, same after 2.6 sec but measuring timc extend:d to 1/2 sec. Slope constant of 2 and 3 is 4.1 imp/sec/nA (Granit et a/., 1963b).

presented in Fig. 3. Impulse frequency is here plotted against current strength and the three curves refer to the moments 1/3, 1.3 and 2.6 sec after initiation of transmembrane stimulation. Leaving out the first pair of spikes, all of our experiments show that the relation between current strength and impulse frequency is linear over a range which exceeds the normal firing rate of the motoneurones. When such curves are plotted for different moments during the rhythmic discharge, adaptation presents itself as a decrease i n the slope of the curves, as in Fig. 3. In most cases the adaptive process is completed within half a second. Occasionally it is a little slower. When the slope decreases, the range within which the linear relation between current strength and firing is maintained also tends to decrease (Fig. 3). It has become customary to speak of ‘accommodation’ in a rather loose manner and so, perhaps, it is best to emphasize at the outset that adaptation by no means can be correctly referred to as ‘accommodation’. The motoneurones which fire tonically are non-accommodative, or they would not fire in this manner. A general discussion of accommodation and adaptation has been given in our second paper (Granit et al., 1963a). The adaptive process is very prominent in the phasic motoneurones. Fig. 4 shows the response of such a motoneurone analyzed as in Fig. 3. It never acquired a steady

39

MAINTAINED FIRING OF MOTONEURONES

state of maintained firing. The slopes, however, are linear there, too, and rapidly diminish as a function of duration of stimulation. The initial slopes of many motoneurones are very high and spread over a considerable range of values. The ‘adapted’ slopes, however, do not vary very much. Selecting our 16 best tonically firing motoneurones, 6 of which were from cats, the average adapted slope constants were of the order of 2.5 imp/sec/nA with a range of from 4.1 to 1.3 imp/sec/nA. Since phasic motoneurones only utilize the initial discharge, their slope constants will always be high and vary from case to case. Tonic firing thus takes place in a manner which makes the motoneurones relatively independent of membrane current, provided the rhythmic threshold has been reached. Some of them were stimulated for very long times (Granit et al., 1963b) and the adapted

4 sec

80-

lnitiai -I

60 -

3

LO -

-1 _ _2 3

20 -

1-1; 11-12 3

O-

3

2 -1

5L-

3

If-2SeC

Fig. 4. Adaptation of phasic rat motoneurone, spike size 75 rnV, in terms of curves relating frequency of discharge to current strength for the moments marked in the figure. First interval not included in initial count (Granit et a/., 1963b).

slope constants then remained constant although the absolute frequency of discharge slowly diminished and the rhythmic threshold rose. The best known limb reflex in which tonic firing occurs is the stretch reflex. From our results would follow that in this reflex the neurones would have to fire at slow rates and, in the steady state, be relatively independent of the degree of extension. This, of course, is well known to be the case (Denny-Brown, 1929; Adrian and Bronk, 1929; Granit, 1958). It is now clear from the present work that an essential factor in the stabilization of frequency of discharge is the adaptation of the motoneurone membrane itself. To this add themselves the after-hyperpolarization and the recurrent inhibition as further stabilizing influences. Whether the motoneurones that can be fired repetitively over seconds from the inside are the very ones which are tonic with respect to muscular afferents, cannot be decided on the basis of the results presented here. More essential is at the moment that if motoneurones can fire tonically, their membranes are so designed that adapted slope constants are small. Essentially the motoneurones behave like the sense organs, which start firing at a high rate and adapt quickly. This behaviour is common if they are organs provided with a tonic discharge. Clearly, the properties of the motoneurones are matched to those of the sense organs. Many reflexes also open in (what Sherrington used to Reteremas p . 41

40

R. G R A N I T

call) d’emblbe fashion, that is, abruptly with a rapid burst of impulses quickly diminishing i n frequency i n the manner required by the properties ofthe cell membranes. On the other hand, many internuncial cells are known to fire at fast rates for some time. This is generally ascribed to the properties of the transmitter substance (Eccles, 1957) but before the slope constants of such neurones have been measured one should be cautious in ascribing their mode of firing exclusively to the chemical mechanism. It is very likely that such neurones have large slope constants and perhaps also much adaptation. If this were so their function would be to provide amplification of input in addition to what else they might do. This approach to the problems is still an untilled field. We do not yet possess the data that would enable us to distinguish ‘amplifying’ cells from ‘stabilizing’ cells on the basis of neural adaptation. But the suggested corollaries to our experiments all represent as many problems capable of an experimental solution. (Granit and Kernell have since confirmed this prediction. Added in proof.) Of special interest is the strict validity of the linear relation between current strength and impulse frequency. It can be discussed from many points of view. I t means, for instance, that i n maintained firing the product of current intensity and spike interval is a constant. This constant is the level from which the motoneurone fires and the stronger the current, the sooner it is reached. The linear rule also means that in the steady state postsynaptic inhibition and excitation are added up algebraically, as shown by Granit and Renkin (1962). Inhibition merely implies sliding ‘downwards’ on the same curve along which excitation slides ‘upwards’. The quantitative implications of this finding are extensive. SUMMARY

1. The aim of this work was to study the relation between transmembrane current strength and discharge frequency of rat and cat motoneurones. 2. Comparison of maintained transmembrane stimulation with single antidromic shocks disclosed that the delayed depolarization, seen after the antidromic spike, disappears during intracellular stimulation of a cell to repetitive activity. This transformation of the cell’s mode of firing is prominent i n most rat cells and i n several cat cells (cf. p. 52 Kernell in the Discussion). 3. The motoneurones which are capable of firing repetitively to maintained transmembrane stimulation do so along curves relating current strength linearly to firing rate. 4. The slope constants of such curves are initially steep, but after about 0.5 sec they acquire a fixed or ‘adapted’ value and are then of the order of merely 1.5-4 imp/sec/nA. 5. A process of ‘adaptation’ is defined by the decline in the firing rate and the decrease of the slope constants. 6. Some implications of these results for integrative work on motor control are discussed.

MAINTAINED FIRING O F MOTONEURONES

41

REFERENCES

ADRIAN,E, D., AND BRONK, D. W., (1929); The dischaige of impulses in motor nerve fibres. Part 11. The frequency of discharge i n reflex and voluntary contractions. J . Physiol. (Lond.), 67, 119-151. ARAKI,T., A N D OTANI,T., (1955); Response of single motoneurons to direct stimulation in toad’s spinal cord. J . Neuroplzyhiol., 18, 472-485. BRADLEY, K., A N D SOMJEN, G. G., (1961); Accommodation in motoneurones of the rat and the cat. J. Physiol. (Lond.), 156,75-92. D. B., (1929); On the nature of postural reflexes. Proc. m y . SOC.B., 104, 252-301. DENNY-BROWN, ECCLES, J. C., (1957); The Physiology of Nerve Cells. Baltimore. The Johns Hopkins Press. GRANIT, R., (1958); Neuromuscular interaction in postural tone of the cat’s isometric soleus muscle. J . Physiol. (Lond.), 143, 387-402. GRANIT,R., KERNELL, D., A N D Sr
42

The Delayed Depolarization in Cat and Rat Motoneurones D. K E R N E L L The Nobel Institute for Neurophysiology, Karolinska Institutet, Stockholm

From numerous studies of motoneurones from different species using the intracellular recording technique it is known that the antidromic spike is generally followed by a short phase of depolarization of varying shape preceding the onset of afterhyperpolarization. Previous studies of this event are not extensive (Brock et a / . , 1952; Araki and Otani, 1955; Eccles, 1957; Eccles et al., 1958a; Machne et al., 1959; Araki, 1960), and it is mostly described as a ‘true’ after-potential homologous with the well-known after-potentials succeeding the spikes in muscle and nerve fibres. This identification may or may not be legitimate and for the time being we have therefore preferred the neutral designation ‘delayed depolarization’. As Prof. Granit already mentioned (this Volume, p. 35) we have found that in the rat motoneurones delayed depolarization is very prominent and generally has the shape of a hump taking off from the falling phase of the spike. Similar observations have been reported in work on amphibian motoneurones. Granit et al. ( 1 963) devoted special attention to this phenomenon and thought that it possibly might be a sign of dendritic activation. The aim of the present investigation was to extend the analysis of delayed depolarization. MATERIAL A N D METHODS

In the present work some rats were used, but most of the experiments were performed on cats, partly because the properties of cat lumbar motoneurones are so well-known in other respects. The microelectrodes used were filled with 2 M potassium citrate. Otherwise the technique was as in our earlier work (Granit, this Volume, p. 35; Granit et al., 1963). The delayed depolarization seemed to behave in a very similar way in rats and cats. The reported results refer to cat motoneurones when not otherwise stated. RESULTS

General characteristics of the delayed depolarization The records topping Fig. 1 show antidromic spikes from six different cat motoneurones. As is seen, the delayed depolarization could assume quite different forms in different cells of this animal. In some cells (records on the left) the delayed de-

DELAYED DEPOLARIZATION I N MOTONEURONES

43

polarization had the shap:: of a more or less prominent hump. Nearly all of the antidromic spikes from rat motoneurones (Granit et al., 1963), and also those reported from frogs and toads (Araki and Otani, 1955; Machne et al., 1959; Araki, 1960), possess a delayed depolarization in the shape of a hump. In the cat, however, many spikes were followed by a smoothly decaying delayed depolarization in which no hump could be seen (Fig. I , middle records). There existed several intermediary forms between delayed depolarizations with and without a definite hump. Finally, there were also a few units (Fig. 1, records on the right) practically lacking delayed depolarization.

z 60 mV

c 60 mV

30L

Fig. 1. Top: antidromic spikes from six different cat motoneurones. All calibrations 10 mV. Spike amplitudes: left row (type I) upper 66 mV, lower 71 mV; middle row (type 11) upper 69 mV, lower 87 mV; right row (type 111) upper 64 mV, lower 47 mV. The time mark is 1000 c/sec here and in all the following pictures. Bottom: diagram showing the occurrence of type I, I1 and 111 units in cat. Cells with spike amplitudes above 60 mV are plotted upwards from the horizontal line. Black areas indicate units with a duration of 70 msec or more of the after-hyperpolarization.

The diagram (Fig. I ) shows the relative occurrence of these three different kinds of antidromic spikes in a material consisting of 95 normal appearing units. About half of all the cells and 60% of those with a spike amplitude exceeding 60 mV had a delayed depolarization in the shape of a hump. Delayed depolarization, smoothly declining or in the shape of a hump, occurred in cells with high membrane and spike potentials from which stable recordings were obtained, sometimes for several hours. Both varieties of delayed depolarization also occurred irrespectively of moderate variations in temperature or depth of the anaesthesia. Only 12 out of the 95 spikes were not followed by any clearly visible delayed depolarization. However, most of these spikes were of fairly low amplitude, and many of them on the verge of spontaneous activity, indicating that the cells were somewhat depolarized. It should be Rcferences p. S l j S Z

44

D. K E R N E L L

emphasized that all the results reported below were more readily obtained from the cells with large spike amplitudes than from those with low ones. The black areas in the diagram represent units with durations of after-hyperpolarization exceeding 70 msec. Delayed depolarization with a definite hump had a tendency to be more common in cells with short after-hyperpolarizations than in those with long ones. However, both shapes of delayed depolarization could occur together with an after-hyperpolarization of any duration. The amplitudes of delayed depolarization ranged between 1 and 12 mV and were mostly around 3-7 mV. The durations measured from the onset of the spike varied between 2 and I3 msec and generally were of the order of 2.5-6.0 msec. The range of variation in amplitude or duration was about the same for delayed depolarizations with a hump as for those without it. The delayed depolarization, when present, was always in sequence to a SD-spike and was never seen following an 1s-spike. I n cases of varying IS-SD-latency the delayed depolarization preserved a constant time relation to the SD-spike. A delayed depolarization of the same form and size as the one following an antidromic spike was also generally seen following a spike elicited by a short depolarizing pulse through the microelectrode or a threshold monosynaptic excitatory postsynaptic potential. Increase in strength of the ventral root stimulus was never seen to augment the delayed depolarization. Evidently the delayed depolarization is therefore directly or indirectly caused by the SD-spike and does not represent any kind of postsynaptic phenomenon. In all these respects my observations on cats agree with those of Granit e t a / . (1963) on rats. The effects of preceding antidromic spikes In our earlier work on the rat (Granit et al., 1963) we found that the delayed depolarization was often greatly diminished after the second of two antidromic spikes elicited with a short interval, i.e. one of the factors responsible for the delayed depolarization seemed to have a refractory period. As will be shown below, this was the case also with cat motoneurones if the intervals between antidromic spikes were made short enough. At longer intervals, however, the delayed depolarization following the second of two antidromic spikes was larger than the one following the first spike, as is demonstrated in Fig. 2A. In the records is also seen that the delayed depolarization of the second spike has attained a more prominent hump. In practically all cases in which this experiment was carried out an increase in amplitude of the second delayed depolarization was seen. In about half of the cases the increase was accompanied by an accentuated hump or by the appearance of a hump if there had been none from the beginning. The same result was obtained if a spike elicited by a short depolarizing pulse was used as the conditioning stimulus instead of an antidromic spike. This increase in amplitude of the delayed depolarization was often maximal at an interval of about 5 msec. With longer intervals this effect of the conditioning spike subsided in rough proportion to the lengthening of the interval. The duration of the effect was rather variable, ranging as a rule from 20 to more than 55 msec after the

D E L A Y E D D E P 0 L A R I Z A T I 0 N IN MO T O N E U R O N E S

45

onset of the conditioning spike. The time course of the increase in the delayed depolarization showed no relation to the time course of the after-hyperpolarization.

- -

10 rnsec

10 msec

1

c

2

10 rnsec

Fig. 2. Cat motoneurone. Spike amplitude 66 mV. All spikes antidromic. Further explanations in text.

In Fig. 2C:l is illustrated what happens at extremely short intervals. At the first interval of 1.9 msec the delayed depolarization following the second spike is increased in amplitude but has completely lost its hump and is now of the smoothly decaying variety. In Fig. 2C:2 the interval between the first two spikes is 2.8 msec, and here the delayed depolarization presents a prominent hump. The intervals at which the hump of the delayed depolarization in different units disappeared or was greatly diminished, ranged between 1.9 and 4.0 msec. When the hump had disappeared there always persisted a smoothly decaying depolarization which was of larger amplitude than the unconditioned delayed depolarization. When the delayed depolarization had a smoothly decaying shape from the beginning it was never clearly diminished in amplitude at these very brief intervals, although its rate of fall towards the base-line was often increased. When the antidromic repetition rate was very fast as in Fig. 2B:2 and C:2 the hump was accentuated in the second spike but then gradually disappeared leaving a smoothly declining delayed depolarization. In spite of the change in shape of the delayed depolarization its amplitude was in this case rather constant throughout. It almost looks as if there were a lengthening of the refractory pxiod for the hump under these conditions suggesting that it and the smoothly decaying potential may be different entities. In Fig. 2B and C, presenting short trains of antidromic spikes, it can be seen that practically no further increase in amplitude of the delayed depolarization occurs beyond the second spike for any frequency of stimulation (the case illustrated here References p . 51/52

46

D. K E R N E L L

was the only one seen where the third delayed depolarization actually was a little larger than the second). This was generally the case irrespectively of whether the delayed depolarization was smoothly decaying or had the shape of a hump. In this respect the delayed depolarization seems to differ from the negative after-potential seen in the squid nerve, the cause of which is assumed to be an accumulation of potassium ions at the membrane (Frankenhaeuser and Hodgkin, 1956). When the steady state membrane potential altered spontaneously or was changed by the application of polarizing currents, the difference between the unconditioned delayed depolarization and the one conditioned by a preceding antidromic spike became less and less at more depolarized levels. In the illustrated case (Fig. 3) the 3

2

1

4

Fig. 3. Same cat motoneurone as in Fig. 2. All spikes antidromic. Transient spontaneous change in the resting membrane potential from 1-4. The time mark also represents a reference level for comparison of membrane potentials.

change in steady state membrane potential was spontaneous. The unconditioned delayed depolarization is seen to have been transformed from one with practically no hump (Fig. 3:l) to one with a very prominent hump in the depolarized state (Fig. 3:4), and this hump is not very much affected by a preceding antidromic spike. The effects of polarizing currents The typical changes in size and form of a delayed depolarization provided with a hump which were obtained with polarizing currents are shown in Fig. 4, A-B. All Control

Hyperpol 83

Control

83

-

u

10 msec Control

Hyperpol.

23

11

Depot. 37 nA

10 msec

63

101 n A

Fig. 4. All spikes antidromic and elicited 1.06 sec after the onset of the indicated polarizing currents. B is close to the firing threshold of the cell. C is from a cat motoneurone with spike amplitude 42 mV. Note: In this picture the relation to the time mark gives no measure ofthe relative membrane potentials. A and B from the same cat motoneurone with spike amplitude 65 mV. The depolarization in

spikes are antidromic. Fig. 4A shows a case of strong hyperpolarizing current and B what happens to the same cell during a depolarizing current. The effects are somewhat similar to what was seen in Fig. 3. With stronger depolarizing currents it was never possible to obtain any reversal of the hump; it merely went below the baseline in a

DELAYED DEPOLARIZATION I N MOTONEURONES

47

manner similar to what is seen after the control spike of Fig. 4C:1, which seemed to be depolarized from the beginning. With very strong depolarizing currents the hump sometimes disappeared. The smoothly decaying delayed depolarization on the other hand could be reversed by depolarizing currents (cf. Eccles, 1957). When no hump existed on the delayed depolarization from the beginning it was sometimes possible to make a small one appear by applying depolarizing currents. With hyperpolarizing currents the hump was sometimes diminished, but only in the cell illustrated under Fig. 4C could it be made to disappear completely. As in our previous work (Granit et al., 1963) the amplitude of the delayed depolarization varied with the applied current in a linear manner. The efjrects of synaptic activity When the membrane potential was changed by using tetanic orthodromic stimulation the effects on the shape and size of the delayed depolarization were as a rule much more prominent than when transmembrane currents were applied. They could also be different from those seen with polarizing currents. In Fig. 5A there is a 2 sec interval between each antidromic spike. When a tetanic stimulus to the brain stem is given (A: 3-6) which decreases the membrane potential, the delayed depolarization attains a very prominent hump and has a long duration compared to what is seen in the preceding controls (A: 1-2). This happened in spite of the fact that the membrane conductance was increased during the tetanic stimulus. During the subsequent slow return of the membrane potential towards th: control value it is seen that the amplitudes of the delayed depolarizations A:7-8 are about the same as those in A : 1-2 although the membrane potentials preceding the spikes A: 7-8 are lower than those Brain stem t e t .

1

2

3

7

8

9

10

5

6

11

12

Fig. 5. Cat motoneurone with spike amplitude 60 mV. Stable recordings were obtained from this cell for more than 2 h. (A). 2.12 sec interval between each illustrated antidromic spike. The tetanic stimulation is given to the ipsilateral brain stem at the level of the superior colliculus. The time mark also represents a reference level for comparison of membrane potentials. (B). Antidromic spikes fired during the application of different depolarizing currents with strengths as indicated.

preceding the spikes A : 1-2. That these synaptic actions on the delayed depo1ar;zation are not any simple functions of the changes in membrane potential occurring in the soma is further stressed by comparison with Fig. 5B. Here, between two experiments such as the one in Fig. 5A, the same antidromic spike was fired into the cell during References p . 51/52

48

D. K E R N E L L

different strengths of depolarizing currents, and now there is almost no accentuation of the hump. A synaptic stimulus which hyperpolarized the membrane, i.e. an inhibitory stimulus, had effects which were opposite to the typical excitatory actions described above. In Fig. 6 there is again a 2-sec interval between the antidromic spikes. About 7 sec of -

2

1

Tet. Pero.

3

5-6

10 msec

Fig. 6 . Cat motoneurone with spike amplitude-68 mV."2.12 sec interval between each illustrated antidromic spike. 7.4 sec of controls left out between A and B. The same tetanic stimulation is given to the peroneal nerve in A and B. The broken line represents a reference level for comparison of membrane potentials.

repeated controls are cut out between A and B. The same inhibitory stimulus is given in A and B. When the inhibition is maximal (A:2-3, B:2) the spikes do not seem to possess any delayed depolarization whatever. When the hyperpolarizing effect of the stimulus is weaker (A:4-5, B:3-5) there is a gradual return of a smoothly decaying delayed depolarization, whereas the hump does not reappear clearly until later, in the controls. B:5 is here actually elicited at a membrane potential which is but a little lower than that of A:6. In spite of this the hump is virtually absent from B:5, which is subject to inhibitory stimulation, while it is well-developed in A:6 which is a control spike. The typical diminution or abolition of the hump in the delayed depolarization during postsynaptic inhibition can thus take place in relative independence of changes of steady-state membrane potentials recordable in the soma. This conclusion i s Brain stem tet.

,

I

8

9

10

11

12

Fig. 7. Rat motoneurone with spike amplitude 68 mV. 1.06 sec interval between each antidromic spike. The tetanic stimulation to the brain stem is given ipsilaterally at the level of the superior colliculus. The continuous line represents a reference level for comparison of membrane potentials.

further underlined by the behaviour of the cell in Fig. 7. This is a motoneurone from a rat in which an antidromic spike is elicited about once every second. The slow tetanus given to the brain stem has a weak depolarizing effect on the membrane in

DELAYED DEPOLARIZATION I N MOTONEURONES

49

the beginning, and yet it influences the delayed depolarization in a manner otherwise typically seen with synaptic stimuli which hyperpolarize the membrane. When looking at the membrane potentials it is seen that for instance the spikes 3 and 10 lacking the hump are preceded by membrane potentials lower than those in the controls 1-2 and higher than in control 11-12. DISCUSSION

The delayed depolarization in motoneurones from rats and amphibians generally has the shape of a rather pronounced hump (Machne efal.,1959; Araki, 1960; Granit et al., 1963). In the cat the delayed depolarization has thk shape only in about half the number of units, being otherwise of a smoothly decaying variety. There was no fundamental difference in amplitude between delayed depolarizations with and without a hump. However, there were several ways of affecting the shape of a delayed depolarization. The hump could be abolished so that the shape of the delayed depolarization became smoothly decaying. It waq also possible to make a smoothly decaying delayed depolarization acquire the shape of one with a hump. One possible explanation for these changes in the shape of the delayed depolarization may be that the hump reflects a current of different origin from the one giving rise to the smoothly decaying delayed depolarization. The hump had certain properties in common with spike processes: it could not be reversed by polarizing currents, it had a refractory period, it could be accentuated or sometimes made to appear with depolarizing currents, and it could also be diminished, in one case even abolished, with hyperpolarizing currents. Typically synaptic stimuli had effects in the same general direction as the corresponding polarizing currents, but their influence was, as a rule, very much stronger. The synaptic effects could thus rarely be imitated by polarizing currents. Sometimes the typical synaptic actions also occurred independently of changes of membrane potential as recorded in the soma. If the synaptic effects also in this case were dependent upon the actions of the synaptic currents on the membrane potentials, these findings would tend to locate at least the main factors responsible for the delayed depolarization to a membrane provided with synapses but situated far from the recording site in the soma, i.e. in the dendrites. The hypotheris that the smoothly declining delayed depolarization and the hump may represent different phenomena receives some support from the facts that a smoothly declining delayed depolarization contrary to the hump, showed no clear signs of refractoriness and that it could be readily reversed by polarizing currents, As earlier pointed out by Eccles (1957) it bears some resemblance to the afterpotential following the spike in the IS-segment and in certain jntraspinal fibres. The smoothly declining delayed depolarization may thus well represent a ‘true’ afterpotential directly connected with the SD-spike, but it seems probable from the results of Terzuolo and Araki (1961) that this depolarizing potential is very much longer and larger when the spike is recorded in the proximal part of the dendrites than when it is obtained in the soma. Such a long-lasting depolarization in sequence to an References p . 51/52

50

D. K E R N E L L

SD-spike in the dendrites may well be capable of displaying the increase in amplitude and the accentuation of the hump that was seen when the delayed depolarization followed the second of two antidromic spikes. Actually a conditioning antidromic spike changed the shape of the delayed depolarjzation in a manner very similar to what was seen with depolarizing currents or excitatory postsynaptic activity. The fact that the time course of these effects of a conditioning antidromic spike showed no relation to the time course of the after-hyperpolarization also tends to locate them to the dendrites, or at least to a site outside the soma. Summarizing the situation, it thus seems probable, both from the experiments with synaptic stimulation as well as from those with a conditioning antidromic spike, that important factors responsible for the delayed depolarization are located to the dendrites. In several situations the hump behaved differently from, and rather independently of the smoothly declining delayed depolarization, indicating that these two phenomena may represent factors of different origin. The hump has certain properties in common with spike processes whereas the smoothly declining potential seems to reflect a large depolarizing after-potential or even an extremely slow falling phase of the SD-spike which is propagated out into the proximal part of the dendrites (Terzuolo and Araki, 1961). On this view the hump would reflect a spike discharge further out in the dendrites occurring possibly after some delay as a sequence to an SD-spike propagated out into thc proximal part of the dendrites. As there are many dendrites in the motoneurone the hump could reflect spike discharges from many different sources which would account for differences in amplitude and temporal dispersion of the hump. Both the amplitude and the shape of the hump as well as its latency would also be modified to a large extent by the rather rapidly changing membrane conductance in the soma membrane following its activation by the SD-spike. Further, these parameters of the delayed depolarization would also be influenced by an electronic spread of the potentials. If the hump on the delayed depolarization really represents spikes elicited further away in the dendrites it should also be possible in normal motoneurones to elicit such spikes in isolation by synaptic stimulation at the site of its origin. Small spike discharges elicited from the dendrites by stimulation and not directly propagating into the soma have previously been reported to occur in chromatolysed motoneurones (Eccles et al., 1958b) and in the pyramidal cells of the hippocampu? (Spencer and Kandel, 1961; and others). However, it should not be forgotten that a most important alternative explanation of the delayed depolarization has not been ruled out, namely that it may be an afterpotential which for unknown reasons can change its shape when subjected to various modifying influences. Fully to exclude this alternative would require an experimental situation in which different parts of the neurone membrane could be investigated separately. SUMMARY

1. The investigation was concerned with the depolarizing potential which follows the intracellularly recorded antidromic spike in lumbar niotoneurones, i.e. the delayed

DELAYED DEPOLARIZATION IN MOTONEURONES

51

depolarization. Most of the experiments were performed in cats. Some rats were also used. 2. Cat motoneurones in good condition were generally followed by a delayed depolarization of varying size. The delayed depolarization in these cells had the shape of a hump in about half the number of units, being otherwise of a smoothly decaying variety. 3. It was confirmed that the delayed depolarization is directly or indirectly caused by the SD-spike and that it is not a postsynaptic phenomenon. 4. When the delayed depolarization had the shape of a hump and two antidromic spikes were elicited at short intervals (19-40 msec) the hump of the second delayed depolarization disappeared. With the conditioning antidromic spike at longer intervals (5 up to more than 55 msec) the delayed depolarization was increased in amplitude and in about half the number of cases it presented a more prominent hump. 5. Depolarizing currents diminished the amplitude of the delayed depolarization and made it acquire a more prominent hump. The hump was not reversed in direction by depolarizing currents. Hyperpolarizing currents increased the amplitude of the delayed depolarization and made its hump less prominent. 6. Excitatory and inhibitory synaptic stimuli typically had effects on the delayed depolarization in the same general direction as the corresponding polarizing currents, but their influence was, as a rule, very much stronger. Synaptic stimuli had especially marked effects on the shape of the delayed depolarization. The synaptic effects could rarely be simulated by polarizing currents. Sometimes the typical synaptic actions on the delayed depolarization occurred independently of changes of membrane potential as recorded within the soma. 7. Some possible interpretations of the delayed depolarization are discussed. The material seems to lend some support to the hypothesis that important events responsible for the delayed depolarization take place in the dendrites. It should be stressed, however, that the evidence presented is not conclusive at this point. REFERENCES ARAKI,T., (1960); Effects of electrotonus on the electrical activities of spinal motoneurons of the toad. Jap. J. Physiol., 10, 518-532. ARAKI,T., A N D OTANI,T., (1955); Responses of single motoneurons to direct stimulation in toad’s spinal cord. J. Neurophysiol., 18, 472485. BROCK,L. G., COOMBS, J. S., AND ECCLES,J. C., (1952); The recording of potentials from motoneurones with an intracellular electrode. J. Physiol. (Lond.), 117, 431460. ECCLES,J. C., (1957); The Physiology of Nerve Cells. Baltimore. The Johns Hopkins Press. A., (1958a); The action potentials of the alpha rnotoECCLES,J. C., ECCLES,R. M., AND LUNDBERG, neurones supplying fast and slow muscles. J. Physiol. (Lond.), 142, 275-291. R. R., (195813); The behaviour of chromatolysed motoneurones ECCLES,J. C., LIBET,B., AND YOUNG, studied by intracellular recording. J . Physiol. (Lond.), 143, 11-40. FRANKENHAEUSER, B., A N D HODGKIN, A . L., (1956); The after-effects of impulses in the giant nerve fibres of Loligo. J. Physiol. (Lond.), 131, 341-376. GRANIT,R., KERNELL,D., A N D SMITH,R. S., (1963); Delayed depolarization and the repetitive response to intracellular stimulation of mammalian motoneurones. J. Physiol. (Lond.), 168, 890-9 10. MACHNE, X., FADIGA, E., AND BROOKHART, J. M., (1959); Antidromic and synaptic activation of frog motor neurons. J. Neurophysiol., 22, 483-503.

52

DISCUSSION

SPENCER, W. A., AND KANDEL, E. R., (1961); Electrophysiology of hippocampal neurons. 1V. Fast prepotentials. J. Neuuophysiol., 24, 272-285. TERZUOLO, C., AND ARAKI,T., (1961); An analysis of intra- versus extracellular potential changes associated with activity of single spinal motoneurons. Ann. N . Y . Acad. Sci., 94, 547-558. DISCUSSION

ECCLES : There are many interesting questions concerning the delayed depolarization that you have so thoroughly investigated. If the hump-like component is produced by spikes generated at remote regions of the dendrites, this spike generation would be much later than the latest spikes recorded by Terzuolo and Araki in their intracellular records from dendrites. However, the very slow decline they observed for dendritic spikes certainly would contribute to the smoothly declining phase of the delayed depolarization. I have hitherto assumed that the delayed depolarization of motoneurons was simply homologous with the after-depolarization of nerve fibers, where under conditions of background depolarization the spike declines to a brief notch that rises again to a later after-depolarization. I wish to ask if you have ever recorded hump-like potentials superimposed upon EPSP’s, and that could be attributed to the production of local responses generated i n dendrites by strategic concentrations of excitatory synapses. With chromatolysed motoneurons these local responses are observed, but so far as 1 remember our observations with chromatolysed motoneurons did not reveal any special tendency for the antidromic spike potential to be followed by hump-like delayed depolarizations. Have you investigated chromatolysed motoneurons? KERNELL : Actually I have occasionally seen such hump-like potentials superimposed upon excitatory postsynaptic potentials, but 1 am not yet prepared to commit myself as to their nature. I have not investigated chromatolysed motoneurons. WALL: 1 would like to ask about an alternative hypothesis. Introducing a microelectrode from the dorsal direction while the ventral root is being fired, and stopping as soon as the record shows a successful intracellular antidromicvolley, I would assume that the most likely place to hit first is a dendrite rather than a cell body. Now if that is the case, let me suggest that your type 1 with the hump represents an intradendritic recording with a partial block of transmission along the dendrite, your type I1 represents an intracellular recording, and your type I11 represents a fully blocked dendrite. One test of that would be to ask what would happen if you were recording of your type 1 with the hump and simply wait on, pushing the electrode in a little further. Does it convert to the type I l l ? I would appreciate having your comments about this.

KERNELL: On your theory one would expect to find IS-spikes to vary in size with the size and shape of the delayed depolarization. On scrutiny of the records this has not been observed. Further, since the size and shape of the hump could be altered rather easily and even abolished or made to appear i n a reversible fashion in any one unit, as shown above, your hypothesis does not seem very probable as judged from the present material.

DELAYED DEPOLARIZATION I N MOTONEURONES

53

GELFAN:It was not clear to me what percentage of the motor neurons which you impale fire repetitively, and what percentage just once. GRANIT:While many motoneurons can be made to fire a few spikes to intracellular stimulation, a small number only will be found to fire tonically in the way illustrated above. Undoubtedly the large majority of the penetrated motoneurons will fail to fire tonically for technical reasons beyond control and so calculations of percentages are meaningless. We have, however, emphasized that some of the - to all appearance - good penetrations giving large spikes may well represent motoneurons in which the adaptive process, described above, is so well-developed that they for this reason respond phasically. As a matter of course, this statement represents a hypothesis.

ECCLES:It is remarkable that we have had to wait till now for the first systematic study of the effect of prolonged steady depolarization on motoneurons. Yet this experimental test provides a good replica of the action of a steady level of excitatory synaptic action; and this is the simplest condition of physiological stimulation of motoneurons. It will be appreciated that the synapses will then be applying a steady depolarizing current to motoneurons exactly as in the experiments with steady current application. The linear relationship of applied current to frequency is thus directly applicable to studies of excitatory synaptic action. In part the accommodation with slowing of frequency during a steady current can be attributed to the fall in potential that occurs after the initial summit, as has been found by Araki, Ito and Oshima. This summit is attained in about 20 msec with an approximately exponential decay governed by the time constant of the membrane, and the subsequent decline continues for about 100 msec to a steady plateau at about 70% of the summit. Hyperpolarizing currents give mirror image effects, and there is a reverse sequence of comparable potential changes on cessation of the currents. The changes in membrane excitability as measured by test current pulses exactly parallel the potential changes. GRANIT:There is little to add to these statements except that the general problem of accommodation has been discussed in our original paper no. 2 (Granit, Kernell, Shortess). Our belief is that, in view of the change of the type of response during intracellular stimulation, accommodation does not suffice to explain our findings. Furthermore, adaptation varies during tonic and, hence, non-accommodative firing. For this reason we have preferred a neutral descriptive term such as ‘adaptation’. LUNDBERG:I am in favour of the hypothesis that the delayed depolarization in motoneurons is a sign of dendritic invasion. For the interpretation we also have to consider findings on other types of nerve cells. Kuffler was the first to suggest that the delayed depolarization in the crustacean stretch receptor cell is due to dendritic events. Grampp (Actaphysiol. scand., suppl. 213, 59 (1963)) has made a detailed investigation of this phenomenon with intracellular and extracellular recording from the initial segment, the soma and the dendrites. He has given strong evidence that the large delayed depolarization in the crustacean stretch receptor cells is caused by a dendritic action potential that electrotonically spreads to the soma. Furthermore, Grampp has References p . 55

54

DISCUSSION

shown that under certain defined conditions this delayed depolarization may evoke multiple discharges. There is the possibility that in the CNS synaptic actionsexerted on the dendrites may regulate the degree of dendritic invasion and in this way influence the ability of a cell to respond with multiple spike discharges. KERNELL: Thank you for this interesting comment. WILLIS:Another type of neuron showing a delayed depolarization after the spike potential is the reticular formation neuron. Dr. Magni and I have seen several instances of this phenomenon, and in some the threshold was low enough that a second spike arose from the depolarization. KEKNLLL: Thank you for the interesting information. It is interesting to compare the antidromic responses reported from different kinds of nerve cells because it seems like cells lacking dendrites, as for instance spinal ganglion cells (Ito, 1957), giant cells from Aplysiu (Tauc, 1957) or sympathetic ganglion cells from the frog (Nishi and Koketsu, 1960), generally have no or a smoothly decaying delayed depolarization whereas cells with extensive dendrite trees, like the pyramidal cells of the hippocampus (Kandel and Spencer, 1960) or the retinal ganglion cells (Tomita et a/., 1961) often display a delayed depolarization with the shape of a hump.

SEARS:Has Dr. Granit made any measurements of motoneuron discharge frequency in terms of membrane depolarization? The reason I ask this, is that in experiments on respiratory motoneurons using repetitive stimulation of ‘Ia’ type afferents in the intercostal nerves, I have found that there is a linear relationship between synaptically induced depolarization and discharge frequency. GRANIT: No, for the time being only in terms of current as described. CREUTZFELDT: In addition to Prof. Granit’s report on intracellular stimulation of motoneurons and to Sir John’s remark on accommodation of motoneurons to intracellularly applied currents I should like to mention shortly our own experience with cortical nerve cells (experiments with Dr. Lux and Dr. Nacimiento). The threshold of nerve cells in the cats motor cortex (Betz cells and unidentified cells) is about 10 times below that of motoneurons probably because of their smaller size. Injured cells with a decreased membrane potential have a still lower threshold. The time constant measured with the strength-latency method lies in different cells between 5 and 12 msec. Using long supra-threshold pulses of about 0.5 sec, the initial high frequency discharge drops down within 70-100 msec to a steady state. The initial frequency as well as the steady state discharge are linearly related to the strength of the applied current. The steady discharge stays unchanged for several seconds and only with just supra-threshold stimuli it finally drops down to the spontaneous level. Using slowly rising currents with slopes of even more than I-sec duration no sign of accommodation could be observed, i.e. the neuron begins to discharge always if the current reaches the same absolute threshold value. These observations show that the phasically discharging cortical cells react to long depolarizing intracellular currents in the same way as spinal motoneurons do.

DELAYED DEPOLARIZATION I N MOTONEURONES

55

GRANIT:I can merely repeat what I have said above, namely that the concept of accommodation provides little aid for the understanding of tonic firing which by definition is non-accommodative. Slowly rising currents may provide information about accommodation when the preparation consists of a single node i n a nerve fiber, the adjacent nodes having been cocainized, but not when the index used is the spike fired by the soma of a motoneuroii. I do not understand why the time constant of motoneurons should necessarily be related to repetitive firing. As to the explanation of tonic firing in terms of the ionic theory, I am aware of work, soon to be published, which will deal with this question in detail. SEARS:1 have just one other point that seems relevant. The after-hyperpolarization of the motoneuron has this specific action of wiping out the pre-existent EPSP. So even if there is a static barrage on the motoneurons this is converted to a postdepolarization. GELFAN:If you are going to treat the spinal neurons as sense organs, as indeed they are, with such large receptive surface as represented by dendrites, the adaptation to conduct stimulation will vary, as it does in peripheral sense organs. I should imagine that ‘phasic’ motoneurons adapt rapidly; tonic motoneurons less so, and spontaneously discharging interneurons not at all. This implies different physical characteristics of neurons. Phasic motoneurons with relatively long subnormality, for example, are not apt to fire repetitively at high frequency under normal conditions. KERNELL: When talking of tonic and phasic motoneurons I would like to mention that I have made some further investigations on the rhythmic properties of cat motoneurons as related to the duration of the after-hyperpolarization. Both neurons with very short and very long after-hyperpolarization could give well-maintained rhythmic discharges to depolarizing currents, and the duration of the after-hyperpolarization could not be correlated to the ability of rhythmic firing. Also it is interesting that differences in adapted slope constants show no evident relation to the duration of the after-hyperpolarization. The frequency range, however, seems to be related to the duration of the after-hyperpolarization. REFERENCES

ITO, M., (1957); The electrical activity of spinal ganglion cells investigated with intracellular microelectrodes. Jap. J. Physiol., 7,297-323. KANDEL, E. R., AND SPENCER, W. A,, (1961); Electrophysiology of hippocampal ncurones. 11. After-potentials and repetitive firing. J. Neuvophysiol., 24, 243-259. NISHI,S., AND KOKETSU, K., (1960); Electrical properties and activities of single sympathetic neurons in frogs. J . cell. comp. Physiol., 55, 15-30. TAUC,L., (1957); Stimulation du soma neuronique de I’aplysie par voie antidromique. J. Physiol. (Paris),49, 973-986. TOMITA, T., MURAKAMI, M., HASHIMOTO, Y., AND SASAKI, Y., (1961); Electrical activity of single neurons in the frog’s retina. The Visiral System: Neirvophysiology and Psychophysics. BerlinGottingen-Heidelberg, Springer Verlag, pp. 24-3 I .

56

The Properties of Reticulo- Spinal Neurons W. D. W I L L I S *

AND

F. M A G N I

Istituto di Fisiologia dell’ Universita di Pisa e Centro di Neurofisiologia del C.N.R.. Sezione di Pisa, Pisa (Italy)

Although specific information is available about the histology of reticulo-spinal neurons (e.g. Torvik and Brodal, 1957; Scheibel and Scheibel, 1958), our knowledge of their physiological properties is largely inferential. Activity within the reticular formation has been studied by extracellular recording of potentials from many or single units (see Rossi and Zanchetti, 1957, for references) and by intracellular recording from the cell bodies or axons of reticular neurons (Kostyuk and Limansky, 1961; Limansky, 1961). However, the types of units investigated were not identified, and there is no assurance that the behavior of reticular neurons in general holds true for reticulo-spinal neurons in particular. The present work was undertaken to provide a method for identifying reticular neurons by antidromic activation of their axons. We shall be concerned here only with the properties of reticular neurons having axons descending into the spinal cord. Full reports of the work, including the study of reticular neurons with ascending axons, may be found elsewhere (Magni and Willis, 1963, 1964a, b). METHODS

Cats were used, either anesthetized with nembutal(35 mg/kg), decerebrate or ‘pyramidal’ (Whitlock et d., 1953). The cerebellum was removed, so that glass microelectrodes could be inserted into the medial part of the pontine and medullary reticular formation through the floor of the fourth ventricle. Reticular neurons were impaled by the microelectrodes, and reticulo-spinal neurons were identified by intracellular recording of the antidromically conducted action potential that resulted from stimulation of the spinal cord. For this purpose, a pair of electrodes was placed beneath the cord, generally with the cathode at L1. The roots and denticulate ligaments at L l and L2 were cut, so the cord could be lifted gently by the electrodes; further protection against shunting of the stimulus was provided by a plastic sheet placed beneath the cord. The volley produced by cord stimulation was monitored by a ball-tipped elec-8 trode placed on the cord dorsum, usually one segment rost;al to the stimulating

* Postdoctoral Research Fellow, National Institute of Neurological Diseases and Blindness, U.S. Public Health Service.

THE PROPERTIES OF RETICULO-SPINAL N E U R O N S

57

electrode, or on the floor of the fourth ventricle. Afferent volleys to the reticulospinal neurons were provided by stimulation of the cerebral cortex, the region of the central tegmental tract of the mesencephalon, or peripheral nerves. RESULTS

Characteristics of antidromic action potentials of reticulo-spinal neurons Antidromic action potentials recorded intracellularly from reticulo-spinal neurons resembled those of other types of central neurons, such as spinal cord motoneurons (Brock et al., 1953), in arising abruptly from the baseline at a constant latency. However, their configuration differed in certain respects from that of motoneuronal action potentials. The spike potential of most reticulo-spinal neurons had a simple form, at least while the resting membrane potential was hgh. For instance, Fig. 1 (A and B) shows

Fig. 1. Reticulo-spinal neuron which developed an IS-SD delay and block as the membrane potential declined. In this and succeeding figures showing activity of reticulo-spinal neurons, the upper traces are the intracellularly recorded potentials (depolarization upward) and the lower traces the cord volley. When three traces are shown, as in C and D this figure, the lowermost one is the field potential recorded after the microelectrode was withdrawn to a just extracellular position. The 1 msec time scale below record B applies to A-C, while that below H applies to E-H. The 10 msec scale is for D only. The 50 mV potential scale refers t o A, B and E-H, while the 2 mV scale is for C and D. (From Magni and Willis, 1963.)

the action potential of a reticule-spinal neuron in response to stimulation of the spinal cord at L1. There is no inflection on the rising phase of the spike potential comparable to the initial segment-soma dendritic (IS-SD) delay seen invariably in motoneuronal spike potentials. Nor could such a delay be produced in reticulo-spinal neurons OF this type by using two consecutive stimuli at a close interval (see Fig. 3H). However, when the resting potential declined, there often (but not always) appeared such a delay and even a block, as in Fig. 1 (E-H). The delay usually occurred on the References p . 64

58

W. D. W I L L I S A N D F. M A G N l

rising phase of the spike potential, although sometimes it appeared at the summit of the spike or even on the descending limb. Although the actual sites of delay could not be determined, it seems reasonable to assume that most occur, as in motoneurons, at the initial segment-soma dendritic junction. A few reticulo-spinal neurons had an IS-SD delay comparable to that of motoneurons even when their resting potential was high. One such may be seen in Fig. 2.

Fig. 2. Reticulo-spinal neuron with spike potential having IS-SD delay. The msec time scale applies to all records. The 50 mV potential scale is for A, B and E, while the 2 mV scale is for the remaining records. (From Magni and Willis, 1963.)

The arrow in Fig. 2A indicates the point of inflection. The second of two spike potentials, separated by 2 msec, displays a more pronounced delay, with intermittent blockage of the SD component (Fig. 2E). Other features of the action potentials of reticulo-spinal neurons which could be studied using the technique of antidromic activation were the after-potential sequence and the refractory periods. Typical records are shown in Fig. 3. The spike potential of this reticulo-spinal neuron was simple in form (Fig. 3A), even when a double stimulus was employed (Fig. 3H). The after-depolarization was brief (about 2 msec), and it was followed by a short-lasting (about 30 msec), low amplitude after-hyperpolarization (Fig. 3 D, E). The after-hyperpolarization increased in size with repetitivc stimulation, but this did not prevent the neuron from discharging at a rate approaching 700/sec (Fig. 3G). The absolutely refractory period lasted 1 msec (Fig.3H). The second of two spike potentials was reduced at intervals up to about 5 msec (Fig. 31), which presumably corresponds to the duration of the relatively refractory period. Since most reticulo-spinal neurons have action potentials that lack an IS-SD delay and that have small after-potentials, it was necessary to exercise care before concluding that recording was from the cell body region rather than the axon of a given neuron. This was crucial in these experiments, since axons from many sources traverse

THE P R O P E R T I E S O F R E T I C U L O - S P I N A L N E U R O N S

59

the reticular formation, while we wished to study only neurons belonging to the reticular formation. The chief criterion used to show that recording was from the soma-dendritic region was the presence of a postsynaptic potential from some source.

Fig. 3. Reticulo-spinal neuron with typical action potential. The msec time scale below C refers to A-C, F and I. The time scale below G-H is 10 msec when applied to D, E and G and 1 msec when applied to H. The 50 mV potential scale is for A, D and G-I, while the 3 mV scale is for B, C, E and F. (From Magni and Willis, 1963.)

For example, the reticulo-spinal neurons of Figs. 1-3 all received EPSP’s from stimulation of a pathway in the region of the central tegmental tract of the mesencephalon (Figs. ID, 2F and 3F). The neurons of Figs. 2 and 3 also had EPSP’s from spinal cord stimulation, as seen in the records taken at threshold for the axons (Figs. 2D and 3C). In addition, the neurons of Figs. I and 2 received EPSP’s from the left and right sensorimotor cerebral cortex (Fig. 2G, H ; not illustrated for the neuron of Fig. I). Psopesties of the axons of reticulo-spinal neusons Antidromic activation proved to be a useful technique not only for thc investigation of the action potentials of reticulo-spinal neurons, but also for the study of properties of the axons of these neurons. The conduction velocities of the axons of 100 consecutive reticulo-spinal neurons were measured, with the result shown by the histogram of Fig. 4. About 90% of these axons’had conduction velocities between 90 and 130 m/sec, suggesting that that part of the reticulo-spinal tract which could be studied with our experimental approach is a fairly homogeneous group of large diameter fibers. In keeping with the rapid conduction velocity, the axons of reticulo-spinal neurons References p . 64

60

W. D. W I L L I S A N D F. M A G N I

24 22 20 18

P3 a c

n 5 z

16

14 12 10

8 6 4 2

50

90

70

130

110

rn/sec

Fig. 4. Histogram showing conduction velocities of 100 reticule-spinal axons. The conduction velocities are grouped into categories of 10 m/sec, as shown on the abscissa, and the numbers of axons in each category are indicated on the ordinate.

80

2:. 40 20

@

0

2

4

6

8 I

10

.

.

3

12

14

.

16

. .

Time threshold , .

.

Fig. 5. Relationship between strength of stimulation of spinal cord and size of antidromic reticulospinal volley. (A). Sample records of the field potentials recorded from the floor of the fourth ventricle in response to stimulation of the spinal cord at L1 using the indicated stimulus strengths. The values are given as multiples of the strength required to excite the lowest threshold fibers of the cord. The amplitudes of the field potentials from the records illustrated, as well as others, are plotted as a percentage of maximum against the stimulus strength in the graph of B.

THE P R O P E R T I E S OF R E T I C U L O - S P I N A L N E U R O N S

61

had thresholds for electrical excitation which tended to be among the lowest in the spinal cord. Fig. 5 shows the relationship between the strength of stimulation of the cord and the size of the potential recorded from the floor of the fourth ventricle in association with the antidromic volley in reticulo-spinal neurons. The sample records in Fig. 5A illustrate the potentials evoked by stimuli of the indicated strengths relative

Fig. 6. Caudal extent of projection of reticulo-spinal axons. The msec time scale applies to all the records. The 200 pV potential scale is for A-D, the 50 mV scale for E, F and the 2 mV scale for G, H. Note that the gain of the cord volley record was increased for D (upper trace) and G (middle trace). (From Magni and Willis, 1963.)

to that required to excite the lowest threshold fibers of the cord. The graph of Fig. 5B shows that the field potential was maximal when a strength of about 10 T was employed. In another experiment, seven reticulo-spinal neurons were impaled, and the threshold stimulus strengths for their axons ranged from 1.5-7 T. Although one might have expected axons with such large conduction velocities to have even lower thresholds than those observed, it is necessary to keep in mind that the entire spinal cord was being stimulated, with the attendant likelihood of an asymmetric distribution of stimulating current. The extent and course of projection of the axons was also studied. By shfting the stimulating electrodes along the spinal cord, it was possible to show that although most reticulo-spinal axons terminate rostra1 to L3, some end at S1 or more caudally. In Fig. 6(A-D) are shown the field potentials which could be recorded from the floor of the fourth ventricle after stimulation of the axons of reticulo-spinal neurons at the indicated levels of the spinal cord. The size of the potential was markedly reduced when the stimulated site was moved from L2 to L3, but a small potential was still visible when stimuli were applied at S l . The records in Fig. 6(E-H) illustrate a reticulo-spinal neuron whose axon could be excited either at L2 (Fig. 6E) or S1 References p . 64

62

W. D. W l L L l S A N D F. M A G N I

(Fig. 6F, G). It received an EPSP from the right sensorimotor cortex (Fig. 6H). Experiments were performed in which lesions were placed in either the dorsum or ventrum of the spinal cord between the stimulating electrodes and the brain stem to determine in which part of the cord the reticulo-spinal axons pass. The results suggest that most of the axons lie in the ventral part of the cord. Reticulo-spinal neurons were found which had at least two axons, one descending into the spinal cord and one ascending to the region of the central tegmental tract of the mesencephalon. Such a neuron is shown in Fig. 7. The antidromic action potential

Fig. 7. Reticular neuron with axons projecting both rostrally and caudally. The time scale applies to A-F. The 50 mV potential scale is for A, B, D and E, while the 2 mV scale is for C and F. (From Magni and Willis, 1963.)

from spinal cord stimulation is shown in Fig. 7(A-C), while that from mesencephalic stimulation is in Fig. 7(D, E). EPSP’s from spinal cord (Fig. 7C) and mesencephalic (Fig. 7E, F) stimulation are also seen. The site of mesencephalic stimulation was determined by placing a lesion through the stimulating electrode, as shown by the blackened area in Fig. 7G. The approximate location of the neuron is indicated by the dot in Fig. 7H.

THE PROPERTIES OF R E T I C U L O - S P I N A L N E U R O N S

63

DISCUSSION

Reticulo-spinal neurons appear to be well adapted for discharging at high rates. Indeed, Amassian and DeVito (1954), as well as other workers, have observed unitary reticular discharges at frequencies as high as 1000/sec. The brief refractory period and the small and short-lasting after-hyperpolarization contribute to this ability. Another factor is the absence of recurrent inhibition. Thus, reticule-spinal neurons are able t o differ markedly from motoneurons in their discharge behavior. The observation concerning the conduction velocity of reticulo-spinal axons is in agreement with previous work (Lloyd, 1941), as is the evidence that the axons traverse the ventral part of the cord (Torvik and Brodal, 1957). The finding that reticulo-spinal axons project as far caudally as SI conflicts with the work of Torvik and Brodal (l957), but both the anatomic and the physiologic approaches agree in suggesting that the bulk of the axons terminate at or rostra1 to the upper lumbar cord. The demonstration of reticular neurons with axons descending into the spinal cord and ascending into the midbrain confirms the histologic evidence for the presence of such neurons (Brodal, 1957; Scheibel and Scheibel, 1958). The function of the reticulo-spinal neurons investigated in this work is as yet uncertain. The iicurons with axons projecting both into the cord and rostrally into the midbrain presumably coordinate activity at higher and lower levels of the nervous system. Although little can be said at present about what these neurons do at the spinal cord level, it is possible at least to suggest that they do not participate in one activity attributed to reticulo-spinal discharges. This is the tonic descending inhibition of interneurons in pathways of Group Ib and flexor reflex afferent fibers which has been described by Lundberg and his co-workers (R. M. Eccles and Lundberg, 1959; Holmqvist and Lundberg, 1959, 1961; Holmqvist et a/., 1960). The axons of the tonic inhibitory pathway descend in the dorsal part of the cord and have thresholds more than ten times the threshold of the most excitable cord fibers. Although these properties rule out the participation of the reticulo-spinal neurons under study in this work, it is possible that reticulo-spinal neurons of the lower medulla or other regions not explored in this work might have axons of suitable diameter and course to meet the requirements of this pathway. It is also possible that such neurons are interspersed among those investigated here and that the stimulus strength applied to the cord was too low to excite them antidromically or that their cell bodies are too small to sustain impalement. It is interesting to speculate whether the reticulo-spinal neurons studied in this work might be responsible for the presynaptic inhibitory action produced in pathways of Group Ia, Ib and flexor reflex afferent fibers upon stimulation within the dorsomedial part of the caudal pons and the medulla (Carpenter et al., 1962). The pathway involved passes through the ventral part of the cord, and apparently originates from the region in which the reticulo-spinal neurons described here are located. SUMMARY

Reticulo-spinal neurons are identified by intracellular recording of the antidromic References p . 64

64

W . D . W I L L I S A N D F. M A G N I

action potentials resulting from stimulation of the spinal cord. The characteristics of their action potentials are described. The axons of reticulo-spinal cells are found to conduct at rates up to 130 mjsec and to have relatively low thresholds for electrical stimulation. The axons project as far caudally as S1 and travel in the ventral part of the cord. Reticular neurons with axons projecting both caudally into the cord and rostrally to the midbrain are described. The functional role of the reticulo-spinal neurons studied is discussed. A C K N 0 W LEDGE MEN T S

This research has been sponsored jointly by the Office of Scientific Research, OAR, through the Enropean Office, Aerospace Research, United States Air Force, under Grant EOAR 62-9, any the Rockefeller Foundation. REFERENCES AMASSIAN, V. E., AND DEVITO,R. V., (1954); Unit activity in reticular formation and nearby structures. J . Neurophysiol., 17, 575-603. J. S., AND ECCLES,J. C., (1953); Intracellular recording from antidromically BROCK,L. G . ,COOMBS, activated motoneurones. J . Physiol. (Lond.), 122, 429-461. BRODAL, A., (1957); The Reticular Forma/ion of the Brain Stem. Edinburgh, O h e r and Boyd. D., ENGBERG, I., A N D LUNDBERG, A,, (1962); Presynaptic inhibition in the lumbar cord CARPENTER, evoked from the brain stem. Experientia (Basel), 18, 450. A., (1959); Supraspinal control of interneurones mediating spinal ECCLES,R. M., AND LUNDBERG, reflexes. J . Physiol. (Lond.), 147, 565-584. A., (1959); On the organization of the supraspinal inhibitory control HOLMQVIST, B., AND LUNDBERG, of interneurones of various spinal reflex arcs. Arch. ital. Biol., 97, 340-356. B., A N D LUNDBERG, A,, (1 961) ; Differential supraspinal control of synaptic actions HOLMQVIST, evoked by volleys in the flexion reflex afferents in alpha motoneurones. Acta physiol. scand., 51, Suppl. 186. O . , (1960); Supraspinal inhibitory control of HOLMQVIST, B., LUNDBERG, A., AND OSCARSSON, transmission to three ascending spinal pathways influenced by the flexion reflex afferents. Arch. i t d . Biol., 98, 60-80. KOSTYUK, P. G., AND LIMANSKY, Y. P., (1961); Intracellular potential recording from medullary reticular neurones. Abstr. Fijth Int. Congr. Electroenceph. clin. Neurophysiol., Rome (pp. 15-16). Y . P., (1961); Intracellular derivation of the action potentials of individual neurons in LIMANSKY, the reticular formation of the medulla. Sechenov physiol. J . USSR, 47, 671-677. LLOYD,D. P. C., (1941); Activity in neurons of the bulbospinal correlation system. J . Neurophysiol., 4, 115-134. MAGNI,F., AND WILLIS,W. D., (1963); Identification of reticular formation neurons by intracellular recording. Arch. ital. Biol., 101, 681-702. MAGNI,F., AND WILLIS,W. D., (1964a); Cortical control of brain stem reticular neurons. Arch. ital. Biol.,in the press. MAGNI,F., AND WILLIS,W. D., (1964b); Subcortical and pcripheral control of brain stem reticular neurons. Arch. ital. Biol., in the press. ROSSI,G. F., AND ZANCHETTI,A., (1957); The brain stem reticular formation. Arch. ital. Biol., 95, 199435. A. B., (1958); Structural substrates for integrative patterns in the SCHEIBEL, M. E., AND SCHEIBEL, brain stem reticular core. Reticular Formation of the Brain. H. H. Jasper et al., Editors. Henry Ford Hospital Symposium. Boston, Little, Brown and Co. (p. 31-55). TORVIK, A., AND BRODAL,A., (1957); The origin of reticulo-spinal fibers in the cat. An experimental study. Anat. Rec., 128, 113-135. WHITLOCK,D. G., ARDUINI,A., AND MORUZZI,G., (1953); Microelectrode analysis of pyramidal system during transition from sleep to wakefulness. J . Neurophysiol., 16, 414429.

65

Presynaptic Inhibition in the Spinal Cord J. C. E C C L E S The John Cuvtin School of Medical Research, Department of Physiology, Canberra City

A. I N T R O D U C T I O N

Barron and Matthews (1938) found that dorsal root volleys gave rise to a depolarization that spread electrotonically along the same or adjacent dorsal roots and postulated that this dorsal root potential (DRP) was produced by the same potential generator that gave the P wave of the cord dorsum (Gasser and Graham, 1933); and this identification has been accepted by all subsequent investigators (Bremer and Bonnet, 1942; Bernhard, 1952; Koketsu, 1956; Eccles, Magni et al., 1962; Eccles et al., 1962a; Eccles et a/., 1963a, 1963b). Barron and Matthews (1938) further postulated that the potential generator for the P wave and the DRP had an inhibitory action in the spinal cord because it gave rise to electric currents that caused blockage of conduction in the collateral branches of interneurones. Subsequently there has been additional evidence that inhibitory action in the spinal cord is due to block or depression of presynaptic excitatory impulses (Renshaw, 1946; Brooks et al., 1948; Howland et al., 1955). Particularly convincing evidence of presynaptic inhibition was reported by Frank and Fuortes (1957) and Frank (1959) who showed that muscle afferent volleys produce inhibition by diminishing the size of the monosynaptic EPSP of motoneurones without having any other demonstrable action on those motoneurones. There is no associated change in motoneuronal excitability or in membrane potential either at the normal resting potential or when the membrane potential is altered by a background depolarizing or hyperpolarizing current (Frank, personal communication); hence, there can be no change in the ionic permeability of the postsynaptic membrane, such as invariably occurs with postsynaptic inhibition. An intensive investigation during the last three years has revealed that exactly the same type of inhibition is widely distributed throughout the mammalian central nervous system, and that it is more powerful than postsynaptic inhibition in depressing the central excitatory actions of almost all primary afferent fibres. Investigation of the presynaptic inhibitory action on a wide variety of afferent fibres of limb nerves has shown that three major types can be distinguished in relation to the identity of the recipient afferent fibres, but no topographical pattern has been discerned. For example the Group Ib fibres of a muscle receive presynaptic inhibitory action from the Group Ib afferent fibres of all muscles of that limb regardless of such functional relationships as synergism or antagonism. Special significance therefore attaches to the subdiviR(,fcrences p. 88\89

66

J . C. E C C L E S

sion into the three categories defined by the recipient afferent fibre; the la type, the Ib type, and the flexor reflex afferent type, which has been chiefly investigated in relation to large cutaneous afferent fibres. B. P R E S Y N A P T I C I N H I B I T I O N O N I A A F F E R E N T F I B R E S

By several different experimental procedures it has been shown that volleys in Group Ta and Ib afferent fibres of flexor muscles depolarize the Group l a fibres of both extensor and flexor muscles, and hence exert a presynaptic inhibitory action. Since presynaptic inhibition of the monosynaptic activation of motoneurones by impulses in Group Ia afferent fibres can be more easily and accurately investigated than any other variety, it will provide all of the initial illustrative material. However, as a result of a fairly comprehensive investigation, it has been established that the other varieties have similar properties, so it can be presumed that they are all examples of precisely the same type of central inhibition. The initial description will cover sequentially three aspects of presynaptic inhibitory action : firstly, the depression of monosynaptic excitatory action ; secondly, the consequent depression of reflex discharge; and finally, the depolarization of the presynaptic fibres that is postulated to be responsible for presynaptic inhibition. This depolarization of presynaptic fibres is measured directly by an intrafibre microelectrode, and also indirectly by their

0

70

60

o

0

4 PBST-PL

-

'

u

20° I

40 I

60

80

100

120

140

160

I

I

I

I

I

8

MSEC 180 200 I

I

Fig. I . Depression of monosynaptic EPSP by presynaptic inhibition. In A the EPSP (CON) in a plantaris motoneurone is seen to be depressed by four Group I conditioning volleys in the nervc to the knee flexcra, posterior biceps plus semitendinosus (PBST). The timing of the conditioning and testing afferent volleys is shown in the upper traces (positivity upwards in both traces). In B the time course of the EPSP depression (expressed as percentage of control) is shown for the series illustrated in A. In C the control EPSP (CON) of another experiment is seen to be greatly depressed both at 5 and 83 msec after a conditioning tetanus of 22 Group I volleys (Eccles et al., 1961).

PRESYNAPTIC INHIBITION I N THE SPINAL CORD

67

increased excitability; and of course it gives rise to the dorsal root potential and to a characteristic field potential in the spinal cord with ventral negativity and dorsal positivity, which includes the P wave on the dorsum of the cord first reported by Gasser and Graha:n (1933). The most direct evidence of presynaptic inhibition is illustrated in Fig. I A , where four conditioning Group I volleys from the knee flexor muscle, posterior bicepssemitendinosus, depress the monosynaptic EPSP of a gastrocnemius-soleus motoneurone relative to the control (CON) response, but do not produce any appreciable postsynaptic potential, either EPSP or IPSP. The time course of this depression of the EPSP is given by the plotted points of Fig. 1 B, there being a latency of about 5 msec, a maximum at about 20 msec and a total duration in excess of 200 msec. Several

A

, c o d lOmV

j 7y4 PBST+GS

0 100

w

100

200

300

100

200

300

400

rnscc

500

-

75

-

50

-

25

-

Oh 0

I

rnsec

400

Fig. 2. Presynaptic inhibition of monosynaptically generated reflex discharges. In A the superimposed control EPSPs (CON) are seen always t o generate the discharge of an impulse at the arrow, whereas conditioning presynaptic inhibition causes impulse generation to fail or be delayed. B and C are comparable series of presynaptic inhibition of monosynaptic reflex discharges into the ventral root, B being from an animal anaesthetized by nembutal and with the spinal cord severed, while C is from a decerebrate unanaesthetized preparation. Specimen records of the reflex discharges are shown, C being control and the others at the indicated intervals after the conditioning tetanus of four PBST volleys (Eccles, Schmidt ef al., 1962). Ri>fi.rmcesp' 88/89

68

J. C . E C C L E S

conditioning volleys at 200 to 300/sec are usually employed as in Fig. IA, in order to increase the size of the presynaptic inhibition by temporal facilitation (Eccles et a/., 1963b). For example in the series of Fig. IA, one conditioning volley produced a maximum depression of less than 10 %, and with two it was less than 20 %. However, the time course of the depression resembled that in Fig. 18, except that it was virtually over by 200 msec (Eccles et al., 1961). Frank and Fuortes (1957) and Frank (1959) reported that with presynaptic inhibition the depression of the monosynaptic EPSP was not associated with any alteration in its time course, which also appears to be the case in Fig. IA; but there is a more convincing demonstration in Fig. lC, where the monosynaptic EPSP of a gastrocnemius motoneurone is greatly diminished in size following conditioning by 22 PBST volleys at 210/sec. The EPSPs at the various test intervals after the tetanus are depressed in size even below 25 %, but have the same time course as the control. The specimen records of Fig. 2A show that, as would be expected, the simple diminution of the EPSP may prevent it from generating the discharge of an impulse, T

A

D 0

-o-o0

20

40

60

80

100

150

-0 -, -o 0

200

0

-0-

0

I

250msec300

Fig. 3. A shows the experimental arrangement for obtaining the responses B-C. The microelectrode is inseited 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 differences between the corresponding intracellular and extracellular potentials, as plotted in C for 1, 2 and 4 volleys, upward deflections signalling depolarization. D shows the time course of the excitability changes which a single PBST volley produces in Group l a afferent fibres of gastrocnemius nerve at the region of their synaptic terminals in the ventral horn (cf. Eccles, Magni et al., 1962). Ordinates show excitability values as percentages of the control, while abscissae give test intervals (Eccles, 1963).

rn M

F

4 SMAB

I

0

z J

1 PBST

7 Fig. 4. Intracellular recording of primary afferent depolarization. As shown by the spike responses to a quadriceps afferent volley in A, B (lower traces) the microelectrode was inserted into a quadriceps afferent fibre at the upper L6 segmental level. With all other records (C-M) the upper trace was recorded in the same way as that of A and B, and signalled the afferent volleys. The second trace in these recoids is the intracellular potential recorded with an amplifier having a time constant of 1 sec, while the lowest trace gives the potential produced by an identical series of nerve volleys, but recorded after withdrawal of the microelectrode to a just extracellular position. Any change of potential across the fibre would be registered as the difference between the intracellular and extracellular traces when the initial parts of their traces were superimposed. Note that depolarizations of the fibre occurred in C, K, L, M, but that there was virtually no membrane change in all other records. Four volleys at 280/sec in the nerves indicated by symbols were employed in C-I, and with J-M various numbers of PBST volleys at the same frequency were employed. Note higher amplification for J-M and same time scale for C-M. Muscle afferent volleys were maximum for Group I, while cutaneous volleys were generated by stimuli three to four times threshold (Eccles, Magni et al., 1962).

0

0

w

0

0. \o

70

3 . C. E C C L E S

or at least delay the discharge. Presynaptic inhibition of monosynaptic reflex discharge can be observed for a population of motoneurones, as in the specimen records of ventral root discharges in Fig. 2B, which are a part of the plotted series. For the initial 20 msec or so there may be postsynaptic inhibitory or excitatory actions superimposed, but thereafter the depression of the reflexes can be assumed to be entirely due to presynaptic inhibition (Eccles, Schmidt et a/., 1962). Such powerful and prolonged presynaptic inhibitions are often observed in preparations under barbiturate anaesthesia (Eccles, Schmidt et al., 1962, 1963d; Eccles and Willis, 1963). In decerebrate unanaesthetized animals the inhibition is always less (Fig. 2C), but nevertheless is much longer than postsynaptic inhibitions in the spinal cord. As shown in Fig. 3A intracellular recording from Group Ia primary afferent fibres has been possible only in the dorsal region of the cord, where the fibres are rather coarse. Since field potentials are relatively large, it is essential to record potentials both inside the fibre and just outside (Fig. 3B) the actual membrane potential change

Fig. 5 . Primary afferent depolarization by Group Ia and Ib afferent volleys of PBST nerve. In upper row of A are specimen records of PADs recorded by a microelectrode in a FDHL-PL Group l a fibre (membrane potential, -64 mV) and of P waves of the cord dorsum (middle row) evoked by four PBST volleys at the indicated strengths relative to threshold. In the lowest row of A specimen records at fast speed from the cord dorsum showing testing of the Ia-Ib composition of the afferent volleys by the double stimulus technique, the results being plotted in C as percentages of maximum. The strengths of the first stimuli are shown above relative to threshold with the exception of the control spike potential (CON) which is for a maximum Group I volley. In B the sizes of the PADs are plotted against stimulus strengths. Note that the specimen records at 1.43 T were at maximum for Group Ia and barely above threshold for Ib, while 1.14 T and 1.72 T gave approximately 50% excitation of Group Ia and Ib fibres, respectively. Voltage scale in A is for PAD records only, while the slow time scale is for both the PAD and the P waves (Eccles et nl., 1963b).

P R E S Y N A P T I C I N H I B I T I O N IN THE S P I N A L C O R D

71

being obtained by subtraction, as has been done for the plotted records of Fig. 3C with the depolarizations produced by 1, 2 and 4 volleys in the PBST nerve. The time course of this presynaptic depolarization is in excellent agreement with t k t of the depression of the EPSP in Fig. I B , there being a latent period of about 4 msec, a summit at about 20 msec and a duration far in excess of 100 msec and probably at least 200 msec. Comparable depolarizations of Group la afferent fibres are regularly produced by volleys in the Group I afferent fibres of flexor muscles (Eccles et al., 1961 ; Eccles et al., 1963b). In Fig. 4 a Group la afferent fibre from quadriceps muscle is seen to be depolarized by Group I volleys from the knee flexor muscles, posterior biceps-semitendinosus (PBST in J to M) and the pretibial flexors (PDP in C), but virtually not at all by volleys in nerves of the various extensor muscles (D, E, F) or by cutaneous volleys (G, H, I). By a finer threshold discrimination it can be shown as in Fig. 5 that both the Ia and

80

0

1.0

1.2

1.4

1.6

1.8

2.0

2.2

2.4 2.8 3.2 3.6 4.0

STIM. STRENGTHS

I 100

%80 60

40 20

0

Fig. 6. Types of muscle afferent fibres responsible for EPSP depression. I n B-D a conditioning PBST volley was evoked by stimuli of the indicated strengths relative to threshold and followed at a fixed interval (9.0 msec) by a monosynaptic EPSP produced by a maximum Group la gastrocnemius volley. The points in F are plotted as in Fig. 5C. Note that a considerable EPSP depression was produced by a Group Ia afferent volley, i.e. with stimulus strengths up to 1.6 T. Body temperatures 34.5”.In E the larger plotted points up to 2 . 0 T are means of two or three records (Eccles et al., 1961). References p . 88189

72

J. C. E C C L E S

Ib afferent impulses from flexor muscles are effective in depolarizing Group la fibres from any type muscle. By a comparable threshold discrimination it can be shown (Fig. 6) that both Ia and Ib afferent impulses from a flexor nerve contribute to the depression of the monosynaptic EPSP (cf. Fig. 1). This parallel provides further evidence for a causal relation between the presynaptic depolarization and presynaptic inhibition. It should be noted that the knee extensor muscle, quadriceps, is a partial exception to the rule of exclusive action by flexor afferents in that both Ia and Ib quadriceps volleys exert a small depolarizing action on Ia afferent fibres. An alternative method of displaying depolarization of primary afferent fibres is to test their excitability by brief current pulses applied through a coarse extracellular microelectrode, which is a technique notably developed by Wall ( I 958). Depolarized fibres are more readily excitable; consequently more fibres are excited by a given submaximal stimulus and there is a larger spike potential antidromically transmitted along the peripheral afferent nerve under observation. The percentage change in excitability is determined by reference to the responses evoked by application of a calibrating range of pulse strengths through the microelectrode (Eccles, Magni et a / . , 1962). For example the points obtained in this way are plotted in Fig. 3D to give the time course of excitability increase produced by a single PBST volley. If suitably scaled this curve would represent the average time course of depolarization of the Group la fibres from gastrocnemius musle. The latency is about 4 msec, the summit is at I5 msec and the total duration is in excess of 300 msec. Two conditioning PBST volleys caused an excitability increase of almost double size, but with a comparable time course. Evidently this method of testing for the depolarization of a population of Group l a afferent fibres gives results in good agreement with the depolarizations directly observed by intrafibre recording (Fig. 3C). The advantage of the method of excitability testing is that it can be applied throughout the whole course of the Group Ia fibres within the spinal cord, and is not restricted to the dorsal segment, as is the case with intracellular recording. This full range of excitability testing is important because it shows that the increase in excitability of Group la fibres near their terminals in the motoneuronal nucleus is at least three times greater than at the region where intracellular recording is possible (Eccles et a/., 1963b). For example in Fig. 7B and C the maximum increase in excitability of both PBST and CS afferent fibres is at a depth of about 3.5 mm, which is in the motoneuronal nucleus (see Fig. 7A). At this location the excitability testing of primary afferent fibres is restricted to Group Ia fibres. In both Fig. 7B and C there tends to be a subsidiary peak of excitability in the region of the intermediate nucleus at about 2.5 mm depth. Both the Ia and Ib fibres of the muscle nerves under test would be contributing to this excitability, and both types have numerous synaptic endings in the intermediate nucleus (Eccles et al., 1960). Superficially thereto, there is a rapid decline in the excitability increments. It can therefore be anticipated that the depolarization at the motoneuronal terminals of the fibre would be correspondingly larger than in Figs. 3C and 4C and K-M. Since the increase in excitability always shows that the depolarization extends along the whole intramedullary course of the primary afferent fibres, this depolarization

73

P R E S Y N A P T I C I N H I B I T I O N IN THE S P I N A L C O R D

must be generated in these fibres by some type of depolarizing synapse, and cannot be secondary to an applied current that is generated in some other structures (Eccles, Magni et al., 1962). The excitability tests such as Fig. 7B, C indicate that the site of action of depolarizing synapses on primary afferent fibres is in the region of their synaptic endings; and this is also indicated by the field potentials that arise secondary

D

rnm

1 4PDP->

O 2mm

2 .o

4PDP

* V

3 0

PBST

3 2 100

I C

3 4

a$"-

110 100 0

1

2

3

4

Fig. 7. Locations of depolarizing foci on primary afferent fibres. B and C are excitability increases measured as in Fig. 3D, and plotted against the depth in mm along a microelectrode track rather more medial than that in A, where the depths in mm are marked. The conditioning stimuli and the fibre terminals on which they were tested are indicated for each series. In D the series of focal potential records with three or four superimposed traces weie at the indicated depths, in mm, along the track shown in A and were evoked by four PDP Group I volleys at 300/sec. A is a traced enlargement of the section through the spinal cord with the microelectrode in situ. Upward deflections in D signal negativity. Same potential and time scales for all traces (Eccles et a/., 1963b).

to the primary afferent depolarization. For example in Fig. 7D the negative field potentials produced by depolarization of Group Ia fibres are larger and sharper at a depth of 3.4 mm than at more superficial or deeper locations. It will be noted that the potential on the surface is positive (the P wave) with a similar time course to the negative wave at depths of 2 mm and more. Reversal occurs at a depth of about 1.6 mm (cf. Eccles, Magni et a/., 1962; Eccles et al., 1963a). The postulated synapses upon synapses have been found by Gray (1962, 1963; personal communication) to be very numerous in the spinal cord at a depth corresponding approximately to the intermediate nucleus. We can therefore have confidence that primary afferent depolarization is produced by depolarizing synapses exhibiting the standard structural features such as synaptic vesicles and the active zones of References p. 88/89

74

J. C. E C C L E S

contact that are distinguished by vesicle accumulation and more densely staining membranes on either side of the synaptic cleft. The latency of synaptic action is now known not to exceed 0.5 msec in the mammalian central nervous system (Eccles, 1961b), so it is likely that most of the central delay of presynaptic inhibition is occupied in transmission along an interneuronal chain. It has been postulated that this central pathway includes at least two serially arranged interneurones ; and interneurones with appropriate properties have been discovered (Fig. 8; Eccles et al., 1962a). Under the most diverse experimental conditions the primary afferent depolarization in the spinal cord invariably displays a prolonged phase (10 to 25 msec) of increase to a rounded summit, which is much more prolonged than is generally observed for postsynaptic actions i n the spinal cord. The simplest explanation is

,

1.2T

A

2T

7

6T

L-

% -.-.-

--,

PBST B

2.7 T

L

..__._: " .

1 5 m~

j 100 msec

, PBST 3.2 T

SMAB 3.0 T

frf-rmrrn lo GS 2.5 T

FDHLPL 5 T

PDP 3.5 T

C

D 25 T

16 T

su Fig. 8. Extracel!ular recording (upper traces) from a D type interneurone (depth 2.15 mm). The lower traces are the cord dorsum potentials at L7 segmental level. A shows the response to tetanic stimulation (265/sec) of PBST nerve with stimulus sttengths given relative to threshold. In B the stimulus strength to PBST was kept constant at 2.15 T, just maximal for Group I, and various tetanus frequencies were employed: numbers show frequencies in cis. C shows the responses to prolonged tetanic stimulation (300/sec) of different muscle nerves at a much slower sweep speed. The nerve and the stimulus strength relative to threshold are indicated on each record. D, another D type interneurone. The nerves stimulated arc indicated below each column. The stimulus strengths relative to threshold strength are givcn for each record, there being increasing strength from below upwaids. Voltage calibration is for upper traces only, which give extracellular spike potentials. Lower traces give cord dorsum potentials (Eccles et al., 1962a).

PRESYNAPTIC INHIBITION IN THE SPINAL CORD

75

that there is a prolonged repetitive bombardmeiit of the synapses depolarizing the presynaptic fibres, just as occurs for example with the inhibitory action of Renshaw cells o!i motoiieurones; and interneurones showing this postulated repetitive discharge have been frequently observed in and around the intermediate nucleus (Fig. 8 D ; Eccles et a/., 1962a). If it be postulated that the primary afferent depolarization (PAD) is produced by the action of a chemical transmitter, the long duration (Fig. 3C, D ; Fig. 4 K-M) can be attributed either to a prolonged action of the transmitter or to a slow passive decay of the depolarization because of the long electric time constant of the membrane. The passively decaying component of the PAD should be erased by an impulse propagating down the afferent fibre to its central terminals, exactly as occurs with an antidromic impulse and the EPSP of motoneurones (Curtis and Eccles, 1959; Eccles, 1961b). IPBST-SMAB

B+d

msec

Fig. 9. Effect of an impulse in a fibre on the primary afferent depolarization (PAD) of that fibre. As shown by the upper traces, the PAD produced by four PBST volleys in a SMAB affcrent fibre was subjected to the action of an impulse in that SMAB fibre at the times marked by the arrows in A to E while F shows the response to the impulses alone, only the small after-depolarization (ADP) being seen because the spike was too large and brief at that amplification and sweep speed. Lower traces in A to F give the cord dorsum potentials. Voltage calibration is for the upper traces only, but time scale is common to both. In G are shown traces of the PAD and of the interacting impulse at the four positions of A to D. The control trace of the PAD is the mean of several records and the various other traces in G were adjusted to this record up to the times of the interpolated spikes that are indicated by the short vertical lines. H shows mean potential produced by antidromic volley alone (F) and plotted on the same coordinates as G (Eccles et al., 1963b).

In Fig. 9 A-E are specimen records showing the effect of superimposing an action potential at various times during the PAD of a nerve fibre. A small after-depolarization (ADP) and later after-hyperpolarization follows the spike potential in the control records, F, H, but this single volley produces no detectable PAD. In the superimposed tracings (G) the ADP is reduced in size when superimposed on the depolarization of the PAD, and is even reversed towards the summit of the PAD. The PAD fails to recover to its control level at all intervals after the interpolated action potential, the Rrferenres p . SXjS9

76

J. C. E C C L E S

deficit being larger with interactions close to the summit of the PAD. However, the level of depolarization after the interaction is always larger than for the ADP and the subsequent hyperpolarization in Fig. 9F and H, so it can be presumed that a considerable proportion of the PAD either survives the propagation of an impulse into the fibre terminals or is rebuilt afterwards. Somewhat comparable observations have been reported for the action of an interpolated afferent volley in partly destroying the DRPs recorded from a frog dorsal root (Eccles and Malcolm, 1946), but the situation there was more complicated because the interpolated volley itself produced a large DRP. The simplest hypothesis is that the impulse destroys all the PAD that is preformed i n that fibre, and that subsequently the lingering transmitter rebuilds much of the depolarization. As would be expected, the rebuilding is seen to be very effective after the earliest interpolation of the impulse in Fig. 9G, and to be progressively less effective at longer intervals. Thus it would be envisaged that the transmitter continues to act throughout the whole duration of the PAD; consequently it is not necessary to postulate in addition that a large area of the surface membrane of the primary afferent fibre has a time constant hundreds of times longer than in peripheral medullated nerve fibres. The production of the PAD can therefore be attributed to the prolonged action of a chemical transmitter substance which operates in a manner comparable with other depolarizing transmitters, namely, by effecting a high permeability to ions. The attempt to evaluate the equilibrium potential for the presumed ionic depolarizing process has not been successful. All that can be stated is that it is likely to be at least 30 mV depolarized in relation to the resting membrane potential (Eccles et a/., 1963c). There has been excellent agreement in every respect between the observed depolarizations of the primary afferent fibres and the depression of their synaptic excitatory action (Eccles et al., 1961; Eccles, Magni et al., 1962; Eccles, Schmidt et al., 1962, 1963a); yet, it is important to realize that this agreement is only qualitative. This qualitative relationship also obtains for the demonstration that passive depolarization of primary afferent fibres diminishes their ability to produce EPSP (Eccles et al., 1962a). For example in Fig. 10 the passage of polarizing currents through the spinal cord, as in the diagram, causes a decrease in the monosynaptic EPSP produced by a maximum volley when the presynaptic terminals are depolarized, and an increase when they are hyperpolarized. The currents would of course also polarize the motoneuronal membrane and so change the size of the EPSP (Coombs et al., 1955); but this effect would be expected to be the reverse from that actually observed, the motoneuronal spike potentials of Fig. 10, being changed in the opposite direction to the EPSPs. It can therefore be assumed that, just as with the squid giant synapse (Takeuchi and Takeuchi, 1962) presynaptic depolarization diminishes the size of the presynaptic spike potential and so diminishes the EPSP it produces. However, it has not been shown that with presynaptic inhibition the diminution in spike potential is sufficient to account for the large diminutions observed for the synaptic excitatory action. When superimposed on PADS, the observed diminutions of spike potentials of afferent nerve fibres are no more than a few per cent (Eccles et al., 1963c), which may seem inadequate to account for an EPSP depression that may

77

PRESYNAPTIC INHIBITION IN THE SPINAL CORD

A

T’O rnV

-

1 rnsec

0

180 I

I

I

t0.4

t0.2

0

1

+0.6

I -0.2

I

I

-0.4 K A -0.6

Fig. 10. Changes produced in monosynaptic EPSPs by polarizing current across the cord. Tntracellular recording from a motoneuronc supplying the anterior biceps muscle, the membrane potential being -70 mV. A shows specimen records of EPSPs under the influence of increasing currents in both directions (applied as in inset), as indicated (in mA) for each record, CON being the control value. Each record consists of many superimposed faint traces. The upper traces are the intracellular records, which are differentiated in the lower traces. In B the amplitudes of the EPSPs (in mV) for the series partly shown in A are plotted against the direction and strength of the polarizing currents. C shows the amplitudes of antidromically evoked SD-spikes measured under the influence of the same currents. In A, B and C ( I-) and (-) indicate the polarity of thr: dorsal electrode. Voltage calibration in A is for intracellular recording only (Eccles ef al., 1962a).

amount to more than 50% (Fig. IC). Yet, investigations on the neuromuscular junction have led to the postulate that the liberation of transmitter has a very steep relationship to the size of the spike potential in the synaptic terminals (Liley, 1956); and Katz (1962) has suggested that, as a consequence, the relatively small presynaptic depolarization produced in presynaptic inhibition may nevertheless have a large depressant action on transmitter liberation and so on the EPSP. For example, if every 15 mV depolirization gives a 10-fold increase in transmitter output, a diminution of the spike potential by 5 mV would reduce the EPSP to less than 50%. It is important also to allow for the fact that in all investigations on the depressant action of the P A D 011 the presynaptic spike potential the intrafibre recording has been remote from the presynaptic terminals; and the excitability testing at these terminals (Fig. 7B, C) suggests that the depolarizations are at least three times larger. I n addition, there would be a further factor besides the mere depression of spike potential by an amount equivalent to the PAD. Since it can be assumed that the P A D References p. SSjX9

78

J. C. E C C L E S

is generated by the flow of ionic currents with an equilibrium potential well below that of the spike potential, these ionic currents would be effective in reducing the absolute potential level of the spike summit, as for example occurs with the endplate potential (Katz, 1962) and this factor would be particularly effective when the presynaptic inhibitory synapses are superimposed directly upon the excitatory synaptic

-

Fig. 1 1. 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 Ia and I b 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.

knobs, which appears to be the usual location (Gray, 1962,1963). Thus, it is postulated that the EPSP depression is produced by the combined influence of two factors in diminishing the size of the presynaptic spike, the depolarization per se, and the localized depression of the spike potential by the ionic currents that give rise to the depolarization. Fig. 9 provides evidence that these ionic currents continue to flow throughout the whole duration of the PAD. Fig. 11 will serve to summarize the simplest pathways that are responsible for depolarizing the central terminals of Group la primary afferent fibres. Both the la and Ib afferent fibres from the knee flexor are shown making excitatory synaptic connections with interneurones in the dorsal horn, which strictly should be shown located in the intermediate nucleus (Eccles et a/., 1960). These interneurones in turn excite other interneurones (shown in black) which have been designated D-cells (Eccles et a/., 1962a), and which make the axo-axonic synapses on l a synaptic knobs

P R E S Y N A P T I C I N H I B I T I O N I N T H E S P I N A L CORD

79

on a motoneurone. The diagram is deficient in many respects. Some pathways would have many excitatory interneurones in serial arrangement. In fact the minimal central latency for the onset of the PAD of Group la fibres is so long - about 4 msec -that even the simplest paths may have several serial interneurones. Also the diagram does not show that the same flexor afferent fibres are just as effective in depolarizing the monosynaptic Ta fibres ending on flexor motoneurones. C. P R E S Y N A P T I C I N H I B I T I O N O N I B A F F E R E N T F I B R E S

The same methods of investigation have shown that Ib afferent fibres from muscle are just as effectively depolarized by a presynaptic inhibitory action, but they have a distinctive receptive field. For example the intracellular recording from the l b fibre in Fig. 12 shows that this fibre is depolarized by Group I volleys from extensor muscles

10 msec

Fig. 12. Tntracellular records from a Q Group Ib fibre. The Q spike potential is shown after a stimulus maximal for Group I in A and at threshold for the impaled fibre in B. Slow depolarizing potentials are recorded in C-I as described for Fig. 4. The four muscle afferent volleys (220/sec) in C-G were at a strength maximum for Group I. The single cutaneous volleys were at a strength 4 times threshold. Note separate time and potential scales for the spike potential in A and B and for the slow potentials in C-I (Eccles et al., 1963a).

even more effectively than by flexor Group 1 volleys. In addition cutaneous volleys exert a considerable depolarization in Fig. 12, but this does not occur with some Ib fibres. In Fig. 13 the depolarization of the Ib primary afferent fibres was tested by the resulting increase in excitability. Ia afferent impulses were entirely without effect on Ib fibres (Fig. 13 C, D), which contrasts with their considerable depolarizing power on la fibres (cf. Fig. 5). Group 11 and 111 volleys from muscles also have an appreciable depolarizing influence on Ib fibres. In summary it can be stated that Ib afferent fibres receive by far the largest part of their presynaptic inhibition from I b impulses of either flexor or extensor muscles. The probable pathways are shown diagrammatically in Fig. 14 where Ib fibres from the Golgi tendon organs of both flexor and extensor muscle are seen to act via interneuronal pathways in order to depolarize l b afferent fibres and so depress their References p . 88\89

D

... Fig. 13. Afferent fibres responsible for increasing the excitability of Group Ib fibres of an extensor ncrve. In A is shown the Group Ib antidromic spike potential (CON) recorded in the SMAB nerve after the centrally evoked Group I volley had collided with a Group Ia hamstring volley. This Ib spike was conditioned at an interval of 36 mscc by 4 PBST volleys (300/sec) produced by strengths of stimulation indicated relative to the threshold strcngth in the specimen records of A. The amplitude of the conditioned spike relative to the control size is plotted in C against the strengths of the stimuli setting up the conditioning volleys. A double stimulus scrics was employed to determine the Ia-Ib composition of the PBST volleys. Specimen records of this series are given in B and the series is plotted in D (Eccles, et at., 1963a).

synaptic excitatory power. This influence thus exerts a negative feedback control on the central actions mediated by Ib afferent fibres. D. P R E S Y N A P T I C I N H I B I T O R Y A C T I O N O N C U T A N E O U S A F F E R E N T F I B R E S

lnvestigations have been restricted to the large cutaneous afferent fibres, the agroup of Hunt and McIntyre (1960). Intracellular recording from such fibres reveals that they receive powerful and prolonged depolarizing influences from all cutaneous nerves of the same limb (SU, SP and PT in Fig. 15). The Group 1 volleys from all of the four muscle nerves (second row in Fig. 15) also produced an appreciable depolarization. When the stimuli were increased so as to excite all of Group 11 and most of the Group 111fibres (third row), there was a considerably larger depolarization in every instance. Fig. 16 illustrates an experimental test in which the excitability of cutaneous fibres was employed as a measure of the depolarizing influence exerted on them by Group I

PRESYNAPTIC INHIBITION I N THE SPINAL CORD

81

muscle impulses over the whole threshold range. The calibration series in row A was employed in assessing the increased excitability produced by 4 conditioning PBST volleys (row B) set up by the indicated stimulus strengths. Depolarization of the cutaneous fibres was produced only when the PBST stimuli were above 1.6 T (curve D),

(Extensor

\cb f i b r e

Fig. 14. Diagram of probable pathways for presynaptic inhibition by Group Ib impulses. Ib fibres from both flexor and extensor muscles are seen to end on interneurones whose axon collaterals converge onto a presynaptic inhibitory neurone that sends its axonal terminals to the synapses of I b primary afferent fibres.

and comparison with the Ia-Ib compositions plotted in E (from the series partly illustrated in C) shows a very good correspondence between the size of the Ib volleys i n the PBST nerve and the depolarization of the cutaneous fibres, while even a maximal Ia volley was without effect. In the plotted curves of Fig. 17, the time course of the depolarization of cutaneous fibres, as measured by the increased excitability, has the characteristic long duration of presynaptic inhibition for the four different types of afferent input indicated by the symbols. This was also illustrated by the intracellular records of Fig. 15. Furthermore, in Fig. 17 Groups 1, I1 and I11 of muscle each have a presynaptic inhibitory effect on cutaneous afferent fibres, as is also shown in Fig. 15, though at a level much below that exerted by a cutaneous volley. Fig. 18 shows diagrammatically the postulated pathways for the presynaptic inhibitory influences on cutaneous primary afferent fibres. The principal action is by cutaneous volleys, but the pathways for actions by Ib, 11 and 111 muscle afferent Rc'feriwces p. 88/89

82

J. C. E C C L E S

-

PDP 18T

2 PBST

06

FDHLPL

Fig. 15. Intracellular recording of the primary afferent depolarization (PAD) of a fibre of the PT nerve. The PT record in the upper right hand corner shows in the upper trace the intracellular spike and in the lower the cord dorsum potential. The spike was photographed at the end of the experimental series when the resting potential had somewhat declined from the initial value of -50 mV. All other records were at a higher amplification and a t the much slower sweep speed in order to display the depolarizations produced by afferent volleys in various cutaneous and muscle nerves of the hind limb, as indicated by the symbols. The upper traces are the intracellular records, depolarization being upwards. The middle traces are the field potentials similarly recorded, but with the microelectrode just outside the fibre; and the lower traces are the cord dorsum potentials, but with upward deflexion negative. All records are formed by the superposition of several traces, usually four. Subtraction of the extracellular fields from the intracellular potentials gives the PAD’S, which are shown in mV for each record. The cutaneous nerves in the first row have been stimulated with single shocks of 4 times threshold strength. All other nerves were stimulated with 4 shocks at 300/sec and the stimulus strength is indicated on each record relative to the threshold (T). The 1 mV calibration is for the intracellular and extracellular records. The 10 msec timer is for all except the spike record, which was recorded at the faster sweep speed (Eccles et al., 1963~).

fibres are also shown. As with Ib presynaptic inhibition, there is the basic performance of a negative feedback system ;cutaneous impulses produce presynaptic depolarization of cutaneous fibres including those conveying the impulse, with the consequence that the effectiveness of later impulses in these fibres is depressed. In this way the presynaptic inhibition resulting from a continuously acting input will automatically turn itself down to some steady level. The presynaptic inhibitory action on cutaneous afferent fibres has been demonstrated by the depression of the reflex discharges evoked by afferent volleys in cutaneous fibres. The flexor reflex exhibits the prolonged depression characteristic of presynaptic

P R E S Y N A P T I C I N H I B I T I O N I N THE S P I N A L C O R D

83

CALIB.4 SP

2.157

130%

120

110

T

D

-

PBST-90V-SP

o

I

I

1.2

14

W

20

xT

2.2 2 . 3 2 . 4

2 0

XT

22 2 3 2 4

---------

100

50

12

10

14

16

I8

IJ.

Fig. 16. Excitability changes produced in SP fibres by Group Ia and Ib afferent volleys from PBST nerve. The excitability changes were measured as in Fig. 3D relative to a calibration series of antidromically conducted spikes in SP nerve that were produced by current pulses driven by a range of voltages as illustrated in the specimen series (CALIB) in the upper row (A). In the series of B the pulse voltage was constant at 90 V and was preceded at a fixed interval of 35 msec by 4 afferent volleys in PBST nerve at 220/sec, as is indicated by the record at slow sweep to the extreme right of B. The stimulus strength for each of the conditioning series of PBST volleys is indicated. In the specimen records in C the double stimulus technique was used to evaluate the Ia-Ib composition of the fourth volley of the tetanic train in the PBST nerve for the stimulus strengths indicated relative to threshold. These spike potentials were recorded at the entry of L7 dorsal root into the cord, the spike to the left being for the last of the train of four volleys, that to the right being testing volley, which alone is shown as CON. The percentage Ia and Ib compositions of the PBST afferent volleys at each stimulus strength are plotted in E on the same abscissal scale of stimulus strengths (relative to the PBST threshold) that was used in plotting in D the percentage increases in excitability of SP afferent fibres that were derived from records such as those of B. Voltage calibrations in A are for A and B only. Same time scale for A, B and C except for extreme right of B. Note broken abscissal scales to right of D and E (Eccles er al., 1963~).

inhibition. There is a similar prolonged inhibition of the discharges evoked both monosynaptically and polysynaptically in the ipsilateral tract of the dorsolateral column. E. G E N E R A L D I S C U S S I O N O F P R E S Y N A P T I C I N H I B I T I O N

Tables I and I1 serve to illustrate the differences between the fields from which Ia and References p . &&I89

84

J. C. E C C L E S

IPT(8T)+SP

0

4PBST(I)+SP 4PBST(I+II)+SP 4PBST(l+ll+lll) +SP

A

90

,1111

100

I

200

I

300

I msec

Fig. 17. Excitability changes produced by muscle (PBST) and cutaneous (PT) volleys in cutaneous (SP) fibres. A microelectrode filled with 4 M NaCl (resistance 1.5 MR) was inserted from the cord dorsum at L7 segmental level to a depth of 1.5 mm. The excitability was tested by single pulses of 0.2 msec duration applied by a Grass stimulator, the electrode being negative to the indifferent electrode. Note that each of the three curves for PBST was determined for volleys at 220/sec at the times indicated by the short vertical lines of the abscissa (Eccles et al., 1963~).

TABLE I EXCITABILITIES OF G R O U P

I

F I B R E S R E S U L T I N G FROM

CONDITIONING

BY T H E V O L L E Y S

A S I N D I C A T E D A N D MEASURED BY T H E T E C H N I Q U E I L L U S T R A T E D I N FIGS. 3 D ,

4 4 4 S M A B GS FDHLPL

4 PBST

4

4

PDP

Q

136

118

108

101

102

102

142

131

106

100

102

102

124

114

127

117

125

115

Meansforall Iafibres in intermediate nucleus Means for l a fibres in motor nucleus Means for all Ib fibres

I

l

SP

SU

101

99

13 l PT 100

106 100 115

T A B L E I1 MEAN

la Ib

M I L L I V O L T S O F D E P O L A R I Z A T I O N S (PAD) P R O D U C E D B Y A C T I O N O F A F F E R E N T VOLLEYS ON P R I M A R Y A F F E R E N T F I B R E S O F TYPES Ia, I b

VALUES IN

4 PBST

4 PDP

Q

4 SMAB

4 CS

4 FDHLPL

I SP

mV 0.50 (34) 0.46 (28)

mV 0.36 (41) 0.28 (28)

mV 0.12 (14) 0.40 (13)

mV 0.04 (39) 0.42 (22)

mV 0.02 (42) 0.27 (24)

mV 0.02 (39) 0.25

0.03 (39)

4

(21)

mV 0.10 (22)

1

su mV 0.02 (40) 0.05 (21)

I PI mV 0.06

(41) 0.12 (22)

85

P R E S Y N A P T I C I N H I B l T I O N I N THE S P l N A L C O R D

2mm

,--\ .-d #----

I

I I

/

/

.-.

/

Fig. 18. Diagram illustrating the pathways for presynaptic inhibitory actions on cutaneous primary afferent fibres. Three cutaneous afferent fibres ( C ) and single Ib, 11 and 111 muscle afferent fibres are shown with monosynaptic endings on the interneurones and also there are shown the secondary interneurones (D) that are postulated o n the presynaptic inhibitory pathways. A separate pathway is shown for Ib presynaptic inhibition. The locus of the presynaptic depolarization is shown at a depth of 1.5 to 1.75 mm (Eccles et al., 1963~).

Ib fibres receive their PADs. Both in the intermediate and motoneuronal nuclei depolarization of Group Ia fibres is produced almost exclusively by the Group I volleys from the flexor muscles PBST and PDP, the only other appreciable action being by quadriceps volleys. Afferent volleys from other extensor muscles or from cutaneous nerves have an insignificant action. The excitability values for Group Ib fibres in Table I present a quite different pattern. Group I afferent volleys from all types of muscle have approximately the same effectiveness of action, and cutaneous volleys usually have an appreciable, though smaller, action. Table I1 shows that the mean values of the PADs are in good agreement with the excitability increments. In Fig. 19 there are shown diagrammatically the pathways responsible for the three types of presynaptic inhibition as specified by the recipient fibres, Ia, Ib and cutaneous. The thicknesses of the pathways are proportional to the mean depolarizations produced in the intracellular records from almost one hundred primary afferent fibres. This diagram serves to summarize the distinctive characters of the pathways concerned in the three types of inhibition. It also illustrates the negative feedback character of presynaptic inhibition, particularly the Ib and cutaneous types, where the dominant action was from Ib onto Ib and from cutaneous onto cutaneous. The effectiveness of this negative feedback control is demonstrated by the dorsal root potentials during a prolonged repetitive stimulation. For example in Fig. 20 are shown the DRPs produced by repetitive stimulation of the muscle nerves at 220/sec References p . 88189

J. C. E C C L E S

86

FIBRES RECEIVING

FIBRES GIVING la

FLEXOR Ib

la

EXTENSOR

Ib

II Ill

CUTANEOUS

Fig. 19.Diagram showing types of afferent fibres depolarizing Group I fibres of muscle and a-cutaneous fibres. The types of afferent fibres which produced the depolarizations are listed to the left. The widths of the arrows from each of these types to each of the three recipient types give approximate measures of the amounts of depolarization (PAD) that are produced, being averages of measurements on a large number of fibres (Eccles et al., 1963~).

DRPs

lOOmsec

v

r r r r r lOOmsec

&US

-

-

1;L ii,

s8C

~

- r r r r r

lOOmsec

lOOmsec

sec

-

-rrrrrr

58c

1

-m-rrr s8C

Fig. 20. Dorsal root potentials produced by single and by prolonged repetitive stimulation of muscle and cutaneous nerves. The first responses to PT and SU nerves were produced by a single stimulus. All other responses were produced by repetitive stimulation (4 T strengths at 220/sec) at the indicated durations. Note the changes in sweep speed. D.C. amplification throughout, but there is a lower amplification for the PT and SU records on the extreme right (Eccles et al., 1963a and c).

PRESYNAPTIC INHIBITION IN THE SPINAL CORD

87

for durations even in excess of 3 sec. The DRP rapidly declined from the initial maximum, but, after about 0.8 sec, a plateau was attained which declined very little during the subsequent period of tetanization, and was maintained at a level about 35 to 50% of the initial maximum. In the two lowest rows there are similar series for tetanization of the cutaneous nerves. The smaller DRPs evoked by S U ncrve decline relatively more before attaining a steady state. These observations are attributable to the negative feedback on the afferent pathway that produces the presynaptic depolarizations. It will be appreciated that there would be a similar development of negative feedback in response to prolonged afferent input from receptor organs. SUMMARY

An initial historical introduction describes early investigations on the depolarization of the central terminals of primary afferent fibres; and the principal part of the paper gives an account of recent investigations which establish that this depolarization is responsible for the depressed synaptic effectiveness of these primary afferent fibres. This account is largely devoted to investigations on presynaptic inhibition of the Group Ia afferent fibres from muscle, but it is also shown that there is a comparable presynaptic inhibition for Group Ib and for cutaneous afferent fibres. Volleys in Group Ia afferents from flexor muscles are shown to produce a prolonged (about 200 msec) depression both of the monosynaptic EPSP evoked by Ia afferent volleys and of the monosynaptic reflex discharge. Concurrently there is a depolarization of these Group Ia primary afferent fibres, as is shown by several tests: intracellular recording from these fibres; increased excitability to test stimuli; electrotonic propagation to give the dorsal root potential ; field potentials in the spinal cord, including the P-wave from the cord dorsum. It is postulated that this depolarization is due to special chemically transmitting synapses that are located on or close to the synaptic terminals of the Group Ia fibres; and this postulate is corroborated by electron-microscopic observations. It is further postulated that the depolarization of the presynaptic terminals diminishes the spike potential in these terminals and reduces the output of transmitter, thus effecting presynaptic inhibition. Presynaptic inhibitory action on Group Ia afferent fibres is produced almost entirely by Group I volleys from flexor muscles. With Group Ib afferent fibres the effective inputs are Group l b and I1 volleys from all muscle nerves and also to a less extent from cutaneous volleys. Presynaptic inhibition of cutaneous primary afferent fibres is exerted predominantly by cutaneous volleys, and to a lesser extent by Group Ib and I1 volleys from muscle. In all cases there appear to be at least 2 serially arranged interneurones on the presynaptic inhibitory pathway, so accounting for the central latency of 2-4 msec and for the prolonged rising phase of about 20 msec, which is atrributable to the repetitive discharge of the 2nd interneurone. The prolonged falling phase of presynaptic inhibition (about 200 msec) appears to be due to a prolonged action of the chemical transmitter. RrJwcnrrs p . 88/89

88

J . C. E C C L E S

REFERENCES BARRON,D. H., AND MATTHEWS, B. H. C., (1938); The interpretation of potential changes in the spinal cord. J . Physiol (Lond.), 92, 276-321. BERNHARD, C. G., (1952); The cord dorsum potentials in relation to peripheral source of afferent stimulation. Cold Spr. Harb. Sytnp. quant. Biol., 17, 221-232. BREMER, F., AND BONNET,V., (1942); Contribution a I’etude de la physiologie gherale des centres nerveux. 11. L‘inhibition reflexe. Arch. int. Physiol., 52, 153-194. BROOKS,C. McC., ECCLES,J. C., AND MALCOLM, J. L., (1948); Synaptic potentials of inhibited motoneurones. J . Neuropliysiol., 11, 417430. COOMBS,J. S., ECCLES,J. C., AND FATT, P., (1955); Excitatory synaptic action in motoneurones J. Physiol. (Lond.), 130, 374-395. CURTIS,D. R., AND ECCLES,J. C., (1959); The time courses of excitatory and inhibitory synaptic actions. J . Physiol. (Lond.), 145, 529-546. ECCLES,J. C., (1961a); The nature of central inhibition. Proc. Roy. Soc. B, 153, 445-476. ECCLES,J. C., (1961b); The mechanisms of synaptic transmission. Ergebn. Physiol., 51, 299-430. ECCLES,J. C., (1963); Presynaptic and postsynaptic inhibitory actions in the spinal cord. G. Moruzzi, A. Fessard and H. H . Jasper, Editors. Brain Mechanisms. Progress in Brain Research. Vol.1. Amsterdam, New York, Elsevier (pp. 1-18). ECCLES,J. C., ECCLES,R. M., AND LUNDBERG, A., (1960); Types of neurone in and around the intermediate nucleus of the lumbo-sacral cord. J. Physiol. (Lon
w.

P R E S Y N A P T I C I N H I B I T I O N I N T H E S P I N A L CORD

89

KATZ,B., (1962); The transmission of impulses from nerve to muscle, and the subcellular unit of synaptic action. Proc. Roy. SOC.B, 155, 455-479. KOKETSU, K., (1956); Intracellular potential changes of primary afferent nerve fibers in spinal cord of cats. J . Neurophysiol., 19, 375-392. LILEY,A. W., (1956); The effects of presynaptic polarization in the spontaneous activity at the mammalian neuromuscular junction. J . Physiol. (Lond.), 131, 427-443. RENSHAW, B., (1946); Observations on interaction of nerve impulses in the gray matter and on the nature of central inhibition. Amer. J . Physiol., 146, 443448. TAKEUCHI, A., AND TAKEUCHI, N., (1962); Electrical changes in pre- and postsynaptic axons of the giant synapse of Loligo. J . gen. Physiol., 45, 1181-1 193. WALL,P. D., (1958); Excitability changes in afferent fibre terminations and their relation to slow potentials. J . Physiol. (Lonci.), 142, 1-21.

D1 S C U S S I O N

WALL:We all profit by the tremendous richness of the experimental data that you present. Could I take up just two points? One is this apparent argument between us about the location of the origin of this potential. Speaking of the cutaneous afferents, I place the origin somewhat more dorsally than you do. There is probably a very simple explanation of this. We published in the J . Neurophysiol. in 1955 a plot of where we found the cells of lamina IV and also, using a monopolar micro-electrode for stimulation, where we found the sural afferents. Those two plots d o not match. The afferents appear to be located somewhat more ventral than the cell bodies. 1 think the reason for this is the sort of anatomy that Dr. Szenthgothai presented. The large a-fibres come up, turn round, and swoop back again. What one is actually stimulating with this technique is the myelinated part of the axon and not the unmyelinated terminal arborization, therefore it is not very surprising that one would seem to get the maximum in a more ventral position versus the technique of looking for the actual sinks which seem to be generating it. A much more fundamental point that you brought up here was the explanation for the very long duration. I would suggest an alternative explanation i.e. that there is repetitive firing going on throughout the disturbance. Just on general principles I am biased against the idea of anything hanging around in a system working this rapidly for 200 msec. A powerful argument against a passive drift, and in favor of this being generated actively the whole time, is the fact that if you are firing a blast of C-fibres you can turn the mechanism off within 10 msec and reverse the whole procedure. In order to turn off the transmitter which simply happens to be there, you have t o assume a whole series of antagonistic transmitters. It is a much simpler hypothesis t o assume that there are cells producing a transmitter with an ordinary type of time course, and transmitters produced by cells which can be either excited o r inhibited. ECCLES:The first point that I would like t o make in reply is that we are not plotting the absolute excitability of the afferent fibres, but merely the percentage change in excitability at each particular depth. It is this percentage change that we find to be largest at 1.5 mm depth, which is very close to the synaptic terminals of the cutaneous fibres on the neurons of lamina 1V. Secondly the field potential studies also show the maximum negativity of the appropriate long duration at a depth of about 1.5 mm.

90

DISCUSSION

With regard to the question of duration of transmitter action my comment is that there is n o general principle that transmitter should be short lasting. For example, with the activation of Renshaw cells by impulses i n axonal collaterals of motoneurons, the transmitter, acetylcholine, even persists for more than a second when acetylcholinesterase is inactivated. Evidently there must be in this situation perisynaptic barriers greatly impeding the diffusion of the transmitter. We would postulate the existence of similar barriers around the presynaptic inhibitory synapses. HORSTFEHR: Could you say something about the electrophysiological aspects of the relationship between neurons and glia in the spinal cord? ECCLES:1 presume that glia in part form the extracellular space for the neurons and conduct the return circuits of the synaptically generated currents, but all evidence is against any active role of glia in relation t o neuronal responses. Intracellular investigations have shown that the surface membranes of glial cells have properties that fit them merely to be passive with respect to the activity of neurons.

GRANIT: We were inclined to some extent to discard the concept of feedback. Far more important in my opinion is the whole problem of central control. The whole complex of inputs is coming into the spinal cord and there must be some mechanism of selection. The central mechanism of control of inputs and outputs must be the dominating feature here altogether. I myself a m more interested in integration than in potentials. I hope that somebody will start using this sort of experiments with primarily natural stimuli and look for the central control mechanisms. ECCLES:I agree that it is highly desirable to use natural stimulation that shows the actual significance of presynaptic inhibition. I also agree that negative feedback is just one of the ingredients in the very complex central mechanism of integration. There is no doubt that the higher centres have the dominating influence on the spinal cord responses by the continual barrage produced by the powerful descending pathways. LUNDBERG: In judging the functional significance of the presynaptic inhibition of the central actions from l a afferents the main difficulty lies in the finding that the action is evoked by both volleys in la and Ib afferents. Do you think the actions from those 2 afferent systems may be mediated by different pathways? This is important because of the supraspinal control systems that may affect the actions from Ia and Ib afferents differentially. ECCLES:It is an important point that you have raised with respect t o the interneurons relaying the l a and Ib afferent impulses. In our experience these interneurons tend to be rather specific for the first relay, but we believe that the presynaptic inhibitory pathway contains 2 serially arranged interneurons, and on the second one probably there is convergence from the Ia and Ib pathways. However, I would like to state that much more investigation of these interneurons is required before we can give a definitive account of these pathways.

PRESYNAPTIC INHIBITION I N THE S P I N A L CORD

91

SPRAGUE:Do the interneurons which mediate the presynaptic inhibition on Ia fibres lie in the same area as those interneurons which mediate the postsynaptic inhibition on Ia fibres? Do you find by micro-electrode techniques of this area, that the characteristics of these 2 interneurons are similar or different? I am thinking of course of interneurons which will discharge in such a manner as to fulfil the exponential nature of the postsynaptic inhibition. ECCLES:It is difficult to answer this question. In our investigations all we have been able to do is to find interneurons with properties that would make them appropriate for mediating this or that central action. In this way we claim to have found interneurons appropriate for presynaptic inhibition and others for mediating postsynaptic inhibition. However, we certainly postulate that different interneurons are concerned in mediating the presynaptic and postsynaptic inhibitory actions that are exerted by Ia afferent fibres. Our reason for doing this is that in Ia postsynaptic inhibition there is only 1 interneuron on the pathway, and it presumably has purely inhibitory action which is blocked by strychnine. On the other hand the first relay cell on the presynaptic inhibitory pathway has an excitatory synaptic action which is not blocked by strychnine.

92

Presynaptic Control of Impulses at the First Central Synapse in the Cutaneous Pathway P A T R I C K D. W A L L Department of Biology and Research Laboratory of Electronics, Massachusetts Institute of Technology, Cambridge, Mass. (U.S.A.)

When nerve impulses travel out of the parent axons and cross the junction region to produce a response in the cell on which they impinge, there are many hazards and opportunities. The journey can fail or succeed at a number of points, and these points represent areas at which control can be exerted on the overall input-output relations of the system. I n the vertebrates, the motor neurone in the spinal cord has apparently been the most suitable cell in which t o study the postsynaptic events which control the cell’s response. In contrast, I believe, that the cells in the most dorsal part of the dorsal horn may turn out t o be the most suitable place i n which to study the details and mechanisms of presynaptic control of transmission. The advantages of these cells lie in their relative simplicity, their anatomical arrangement, and the experimenter’s ability t o manipulate their inputs and outputs. The cell bodies lie in a horizontal lamina, the fourth lamina of Rexed (1952). Within this lamina there is a topographic organization with more distal parts of the body projecting onto the medial part of the lamina and more proximal parts being represented laterally. If one could look down through the spinal cord over its whole length, one would see on each side a continuous lamina running from the coccygeal segments up into the medulla, where it would fuse with the nuclei of the trigeminal nerve. At all levels the same organization applies with the feet and ventral parts of the body being represented medially and the more dorsal or proximal parts of the skin projecting onto the lateral side of the lamina. Some of the axons of these cell bodies run into the dorsolateral white matter of the cord where they run close to the surface. These axons are of large diameter and can easily be recorded. The dendrites of these cells in lamina IV extend dorsally into the region of the substantia gelatinosa, and it is in this region that they contact the terminal arborizations of the entering cutaneous afferent fibers. It should be noticed (Fig. 1) that the larger afferent fibers reach the cells by an indirect course and terminate by running dorsally. In contrast, the smaller fibers enter by a much more direct path and end in the same general region but with their terminals pointing ventrally, instead of the dorsal course of the terminals of the larger fibers. The details of this anatomy are presented by a number

PRESYNAPTIC CONTROL OF IMPULSES

93

Fig. 1. The upper drawing shows a cross section of the dorsal quadrant of cat lumbar cord. The lower diagram shows the three main components of the cutaneous afferent system in the upper dorsal horn. The large diameter cutaneous peripheral fibers are shown as thick lines running from dorsal root to terminate in the region of the substantia gelatinosa. The finer peripheral fibers are shown as dashed lines running directly into the same region. The large cells of lamina 1V of Rexed on which cutaneous afferents terminate are shown as large black spheres with their dendrites extending dorsally. The small cells represent the cells of substantia gelatinosa. Their axons are not shown, but interconnect the cells of substantia gelatinosa and also run in the Lissauer tract which is shown as the most lateral structure.

of other authors in Progress in Brain Research, Volume 1 I , particularly by Dr. A. Szentagothai. The cells of substantia gelatinosa itself lie packed among the terminal arbors of the afferent fibers and the dendrites of the cells in lamina IV. These small cells interconnect with each other by short axons and also by longer axons which run laterally and make up the bulk of the tract of Lissauer. The cells of substantia gelatinosa have not been shown t o send axons which run outside this region, and we must therefore assume at once that they must be expressing themselves on the only other two structures in the region. These two structures are the arriving afferent fibers and the cells of lamina IV. This region of the spinal cord is entirely concerned with the reception of afferent impulses from the skin. We have not seen in the cat any signs of afferents from muscle or joints ending here. The physiology of cutaneous afferent myelinated fibers has been studied extensively in the periphery by many workers, and recently a most detailed study has been published (Hunt and Mclntyre, 1960). We also know something of the physiology of the unmyelinated fibers, particularly from the work of Douglas and Ritchie (1962) and of Iggo (1959). This fortunate combination of intensive anatomical and physiological work means that the peripheral input to the region of interest is unusually well understood. The physiology of the cells in lamina IV has also been studied, both t o natural and artificial stimuli (Frank and Fuortes, Ri,ferenci,s p. 114/115

94

P. D. W A L L

1956; Kolmodin, 1957; Wall, 1959, 1960). Each cell has a definite receptive field to light touch of the skin. The size of the receptive field varies from a few square millimeters on the toes to somewhat larger on the legs. On the legs the average size of the receptive field was 63 x 32 mm, with the smallest being 30 x 15 mm, and the largest 150 x 35 mm. Recently Taub (1963) has reported that if light pressure stimuli are used it can be shown that the main excitatory receptive field is accompanied by inhibitory fields, stimulation of which affects the excitatory receptive field. These cells respond to all the types of skin stimuli which are known to produce impulses in peripheral nerves. The response of the cell goes on increasing over the full range of pressures from the slightest hair movement or skin touch to intense pressures. In addition, all these cells respond briskly to cooling the skin and to chemical stimuli. The apparent fusing of the specific properties of individual afferent fibers in the response of these cells is their most striking characteristic, and it is concluded that probably all types of peripheral cutaneous fibers converge onto these cells. The method of convergence is of interest because it was shown that all the peripheral fibers which will produce a response in the cell on pressure to the skin run in a very small ‘microbundle’ which enters the spinal cord within a single rootlet. In contrast to this very restricted pathway of convergence of fibers from the pressure receptive field, it is possible to produce responses to electrical stimuli, that is to say to the arrival of synchronized volleys, from a very wide area and from fibers which must enter the spinal cord some segments away from the location of the cell body. It is apparent that one form of convergent organization is revealed by the use of light touch stimuli and another much more diffuse one by the use of electrical stimuli to nerve trunks. Impulses in cutaneous fibers converge onto the cells of the dorsal horn and produce the discharge of impulses in the axons of these cells. The success of this process of transfer from afferent to efferent axons can be controlled at a number of points. The processes which can be affected are: 1. lmpulse transmission in the incoming axon and its arborization A. Blockade B. Depolarization C. Hyperpolarization 2. Repetitive firing in the terminal arborization 3. Chemical transmission at junction points A. Synthesis B. Release C. Transfer D. Reception 4. Electrical transmission at junction points A. Release - presynaptic membrane properties B. Transfer - extracellular impedance C. Reception - postsynaptic membrane properties 5. Propagation from point of contact to point of impulse initiation 6. Threshold variation at point of impulse initiation 7. Repetitive firing

PRESYNAPTIC CONTROL OF IMPULSES

95

A. Repetitive firing of external origin B. Repetitive firing of internal origin In our present state of ignorance it would not seem wise to neglect any of these possibilities, nor would it seem likely that nature would not have established many different types of control over such an important event as the invasion of the central nervous system by impulses from the periphery. Furthermore we do not know enough of the functioning of these control mechanisms to speak of their relative importance in the actual life of the animal so that it would seem rash, as some authors have done, to disregard some of these control points or to relegate their activity to rare pathological or artificial states, into which the nervous system can be forced. We should examine each point of the pathway to discover what evidence exists for the control of the input-output relations of the system at each point. 1.

CONTROL OF PRESYNAPTIC IMPULSE TRANSMISSION

I A . Control by blockade

The first evidence for the failure of transmission in incoming afferent fibers was presented in a remarkable paper by Barron and Matthews in 1935 entitled Intermittent conduction in the spinal cord. They showed that impulses entering the cord sometimes fail to continue up into the dorsal column fibers attached to the peripheral axons. Second, they showed that impulses may run antidromically back out of the afferent fibers. They had simultaneously attacked two basic tenets of neurophysiology, and the natural reaction of physiologists seems to have been to ignore the paper. This process was aided by the investigation of the antidromic impulses, the dorsal root reflex, by Toennies (1939) who stressed that the phenomenon was exaggerated by cooling and synchronous volleys. While this is quite true, the fact remains that some signs of this phenomenon continue with asynchronous volleys at normal temperatures. The possible role of the dorsal root reflex in blocking incoming impulses by collision has nothing to do with the problem of intermittent orthodromic conduction, as Barron and Matthews were careful to point out. They stressed that the blocks lasted for such a long time that there would have been plenty of time to observe the dorsal root reflex impulses which were producing the block if such were the case. Instead, they proposed that some influence from the grey matter was spreading along the penetrating collaterals and affecting transmission in the large myelinated axons as they entered the spinal cord. In 1955 we re-examined this phenomenon (Wall et al., 1955). We found that for a prolonged period after the arrival of a volley of nerve impulses over one dorsal root the passage of impulses from that dorsal root or a neighbor into the dorsal columns was partially blocked. The height of the blockade occurred 15-20 msec after the first volley, and the duration of the blockade was greatly extended by barbiturate anesthesia. By high frequency firing and by overheating, we could eliminate the dorsal root reflex so that we could show that the block was not a consequence of collision between entering and leaving impulses. An intriguing aspect of the block was that it worked only on orthodromic impulses while impulses fired antidromically from dorsal columns to Referencrs p. 114/115

96

P. D . W A L L

dorsal roots were unaffected. We see therefore that there is good evidence that impulses can be blocked even in the large entering fibers whether by collision with the dorsal root reflex, o r by a process which we shall now discuss spreading from the collaterals ending in the grey matter. If conduction could be made to fail in the large fibers, it would not be surprising if this phenomenon were even more obvious in the fine fibers of the terminal arborization where the safety factor of transmission would be expected to be much lower. This was suggested by Barron and Matthews in 1938; and subsequently, the suspicion rose in the minds of other authors. Renshaw (1946) found that the monosynaptic reflex could be inhibited at a time when he could find no change in motoneurone properties, and presumed that some presynaptic process must be responsible. In 1948 Brooks et al. discovered that inhibition associated with very large volleys was accompanied by a decrease in the height of the spike potential recorded in the cord, which is the sign of impulses in the arriving fibers. They attributed part of this to block by collision and part to block by some other process. By 1953, we decided t o make a direct investigation to see if we could discover where the impulses were failing to traverse a reflex path, and we chose the well-known inhibitory effect of one dorsal root volley on another. For physical reasons, electrical records from roots and from the surface of the cord d o not determine uniquely

Fig. 2. Righl, distribution of isopotentials within spinal cord 20 msec after conditioning root L6 alone had been stimulated. This time was found to be that a t which stimulation of L6 exerted its maximum inhibitory effect on the reflex which followed stimulation of the test root L5. The map came from 162 recording points. Lrfr, distribution of sinks (dots) representing the location of active depolarization and sources (diagonal lines) within the spinal cord 20 msec after the conditioning volley. These pictures show that at the height of the dorsal root potential and inhibition, themajor region of activity lies in the substantiagelatinosa. See also Fig. 8. (Reprinted from Howland rt a/., 1955, Fig. 2).

P R E S Y N A P T I C C O N T R O L OF I M P U L S E S

97

Fig. 3. Top l e f , isopotential map 1.2 msec after test root L5 had been stimulated. Bottom left, the sink (dot) and source (diagonal line) distribution derived from the potential distribution. This shows the pattern of the earliest disturbance recorded in the region after test stimulation. The impulses are mainly still within the axons in the white matter of the dorsal column. Top right, isopotential map of the entering volley 1.2 msec after the test stimulus to L5 when it had been preceded 20 msec before by a stimulus to the conditioning root L6. Bottom right, the source-sink distribution. The conditioning volley alone produces the disturbance shown in Fig. 2. It is evident that the impulses from the test volley are inhibited presynaptically by the conditioning stimulus. This figure is similar to the one published in Howland et al., 1955. References p . 1141115

98

P. D. W A L L

the site of electrical events within the cord. A grid of successive stations of microelectrodes was used to produce a series of potential maps each of effectively simultaneous values showing the fields of the inhibitory volley remaining at the time of maximum inhibition, and those engendered by the test volley alone and by the inhibited volley. These maps indicate the potentials in the extracellular medium at enough points to permit calculation of the amount of current flowing into or out of the neurons in small regions. Thereby, we were able to estimate to what extent those small regions are the sites of impulses where current is absorbed from the medium (sinks) or sites contributing current (sources). Fig. 2 shows such a map derived from 162 recording points at the height of the inhibition produced by the conditioning volley. Fig. 3 shows the expected early arrival of impulses on the left and the failure of these impulses to arrive in the inhibited state. The details of this experiment are to be found in Howland et al. (1955). The location of the failure is clearly within the white matter and at a time preceding any postsynaptic activity produced by the test volley. If under extreme conditions a block can occur in this location, it would seem very likely that less severe inhibition might be associated with block occurring further into the terminal arborization which raises the most interesting possibility of block in individual terminal twigs. IB. Control by depolarization There is no doubt that the membrane potential of the terminal arbors of cutaneow afferents can be varied. It is in fact this variation which generates the dorsal root potential. The dorsal root potential is measured by placing a cut dorsal root on electrodes with one electrode very close to, but not touching, the cord and the other electrode on the cut end. We shall be discussing the various ways of generating dorsal root potentials below. We examined in particular the dorsal root potential generated by a single afferent volley i n the manner first shown by Barron and Matthews (1938) and carefully studied by Lloyd and McIntyre (1949). The method of examination was to test the excitability of the fine terminal fibers since threshold varies with membrane potential. In this method (Wall, 1958), a microelectrode was placed among the terminals of afferent fibers in the cord. The number of fibers fired by a fixed strength of stimulus was measured by recording the size of antidromic volleys which appeared out on peripheral nerves. It was found that after the arrival of an orthodromic volley in neighboring fibers, cutaneous nerve fibers underwent a severe depolarization, while large muscle sensory fibers from a muscle such as gastrocnemius underwent only minor short-lived depolarizations. The time course of the increase of excitability i n the cutaneous fibers exactly matched the time course of the dorsal root potential (Fig. 4). It was shown that this was not isolated to the spinal cord since similar changes could be observed in endings in nucleus gracilis in the medulla. Having shown that the depolarization was not found in all types of fibers, it was of interest to ask where in the terminal arbor the depolarization occurred. This was done by localized microelectrode stimulation at many points throughout the distribution of an entering fiber. The results we show in Fig. 5. It will be seen that the closer the test point approached the actual terminals,

-. c 4 c

(n

>

.

d

2

0 0 0

z

1

P

0

r 0 Ln

1

1

1

1

1

1

1

1

1

1

1

I

I

I

I

I

I

I

I

I

I

I

Fig. 4. Comparison of the time course of the passive dorsal root potentials (on the left) and the excitability changes in the terminal arborization of passive sural nerve fibers (on the right). The dorsal root potential shows first the short positive deflection (DR IV), then the dorsal root reflex and last the prolonged negative deflection (DR V). The dorsal root potentials and excitability changes are recorded on two different time scales marked in 10 msec. The excitability of the sural nerve endings was measured by placing a stimulating microelectrode among them and by recording the size of the antidromic volley evoked on the nerve in the periphery: as the excitability increased more nerves were fired and the size of the compound action potential recorded peripherally increased. An increase of excitability is taken to mean a decrease of membrane potential. (Reprinted from Wall, 1958.)

\D \D

100

P. D. W A L L

Fig. 5. Time course of cxcitability changcs in sural nerve fibers after the arrival of an orthodromic volley from a neighboring dorsal root. Each record is derived from a different stimulus point along the course of penetration of the sural nerve fibers. The positions of the stimulating microclectrodes are marked in the diagram which shows the dorsal quadrant of the L7 segment of spinal cord. The orthodromic volley was set off at the time marked by the line below each record. (Reprinted from

Wall, 1958.)

PRESYNAPTIC CONTROL OF IMPULSES

101

the greater was the change of excitability; therefore, presumably, the greater was the change of membrane potential. The very extensive recent work of Dr. Eccles and his group on this subject began with Eccles et al. (1961) and has been summarized by Sir John Eccles in this Volume (p. 65). They followed up particularly the work of Frank and Fuortes (1957) who had shown that there was a type of inhibition of motoneurones in which the properties of the motoneurone were unchanged, and yet a monosynaptic excitatory postsynaptic potential, EPSP, was diminished. Eccles et a]. (1961) showed that the diminution of the EPSP was associated with a depolarization of the primary afferents. It should be pointed out that this work was only a more elegant way of showing the presynaptic inhibition, which had already been shown by the workers cited above including Dr. Eccles himself in Brooks et al. (1948). However, they added a quite specific hypothesis which coupled the membrane potential of the presynaptic element to its excitatory effect on the postsynaptic cell. They postulated that presynaptic depolarization results in EPSP depression because it depresses the size of presynaptic impulses and hence decreases the liberation of excitatory transmitter substance. What is the evidence for this direct coupling of size of action potential to its postsynaptic excitatory effect? The only direct evidence comes from the very curious giant synapse in the squid stellate ganglion (Hagiwara and Tasaki, 1958; Takeuchi and Takeuchi, 1962). All other evidence that has been proposed (Eccles et al., 1962) remains indirect and leaves open at least one other possibility, that is, the possibility of block. The present data have been collected from too great a distance from the various processes of transmission to make a definite statement abolit the actual cause of presynaptic inhibition. This is the reason for the caution of Frank (1959) and for my own reticence to accept one theory over the alternatives. Both block of individual fine terminal twigs and Eccles’ suggestion of direct coupling between height of action potential and amount of transmitter would result in the same variation EPSP when observed from the distance of the cell body. The EPSP could just as well be varied by the number of terminal boutons invaded as by a direct coupling of hypothetical transmitter release to membrane potential. The recent remarkable work of Kuno (1963) might suggest a way to resolve the issue in certain selected locations. He has examined a motoneuron bombarded by a single I A afferent fiber from the triceps surae muscle nerve. He fired into the cord a single impulse at 2-sec intervals and observed the EPSP generated in the motoneuron. With no apparent variation in the state of the cord, he recorded a considerable variation in the height of the EPSP. Sometimes there was no effect at all; and at other times EPSP‘s were generated, but he noted that the EPSP’s were generated in a series of unit steps, being about 0.1 3-0.24 mV. If we examine the list of points of hazard for transmission, we can see that this fractionation of the excitatory effect of a single arriving impulse could begin with variation of block in individual terminal twigs and could proceed to variation of ‘packets’ of transmitter released and so on down the line to the recording point. It may be that one can devise experiments in this simple case to differentiate between References p . 114!115

102

P. D. W A L L

the possibilities. Until this has been done, I feel strongly that one should not arrive at definite conclusions about the manner in which presynaptic depolarization results in a decreased excitatory effect. 1C. Control by hyperpolarization The best known and most carefully studied example of this is post-tetanic potentiation where a rapid burst of impulses is followed by a period in which the fibers which have carried the burst are more effective in stimulating the postsynaptic cells. This effect was analyzed by Lloyd (1949), who suggested that it was due to the summation of positive after-potentials. We tested this theory by direct test of the preand postsynaptic excitability changes (Wall and Johnson, 1958) and showed that the change was limited to a hyperpolarization of the active fibers. Post-tetanic potentiation of transmission across the first cutaneous synapse is not as striking as that of the extensor muscle monosynaptic reflex because it is counteracted by a simultaneous potentiation of inhibition. We also owe to Lloyd the discovery of a second type of hyperpolarization of the afferent terminals. This was called DRP VI (Lloyd, 1952) and is seen both in active fibers and in their passive neighbors following the long period of depolarization, which marks the most striking aspect of the dorsal root potential which we discussed in the previous section. The hyperpolarization is seen in the terminals of cutaneous fibers after they have been depolarized by the arrival of a single afferent volley in their vicinity, and it is completely abolished by small doses of barbiturate anesthetic. Before proceeding, it is necessary to introduce a diversion at this point to consider the meanings of the terms hyperpolarization and depolarization. These words were originally used to discuss variations about a resting potential in peripheral nerves. The resting membrane potential is an idealized concept used to denote the undisturbed consequences of the action of cell membrane and cytoplasm on the ionic medium in which the cell normally rests. While it may be convenient to consider a peripheral axon as though isolated from all other nerve cells and from other parts of its own cell, this is a completely artificial concept for any part of a nerve cell in the central nervous system. It is true, of course, that there is a theoretical resting membrane potential which a terminal axon would reach if it were possible to isolate it without damage from all surrounding tissue: But, in fact, as we find a cell in situ, it is under the continuous influence of the activity of surrounding cells. One of these influences on terminal arbors of cutaneous fibers is the mechanism tending to produce depolarization and presynaptic inhibition. A marked dorsal root potential can be produced by very small afferent volleys. In the extreme case, Fessard and Matthews (reported in Barron and Matthews, 1938) could see a dorsal root potential evoked by a single impulse in the frog. If this is the case, we must take into account the steady barrage of impulses which enters the cord in cutaneous as well as proprioceptive fibers in the absence of any phasic stimuli. We have recently reexamined this problem by recording the steady potential on a dorsal root (Fig. 6). Wa have consistently found that, for a rootlet of the L7 dorsal root in unanesthetizep spinal cats, light pressure applied on the foot results in a steady-state depolarization

P R E S Y N A P T I C C O N T R O L OF l M P U L S E S

103

Fig. 6. Recording of potential between cut periphery and central end of small rootlet in L6 segment of cord in spinal cat. D.C. recording. Duration of trace, 90 sec. In lower trace, the foot was gently squeezed and a depolarization was observed throughout period of pressure change. In the upper trace, the peroneal region was squeezed and a hyperpolarization was observed after the stimulus was removed.

of the root. Furthermore, pressure applied in the peroneal region often resulted in an increase in the polarization of the root. The rootlet from which we were recording was cut and therefore isolated from the periphery, and yet the membrane potential of its terminals, as measured by the dorsal root potential, was dependent on the steady state of pressure to the skin. Therefore, we suggest that the terminals of afferent cutaneous fibers are continually held in a partially depolarized state by the action of the continuous barrage of impulses from the periphery. One consequence of this is to make the explanation of any increase of membrane potential ambiguous. An observed hyperpolarization from the state in which we first encountered the nerve might be due to turning down the steady tendency of other elements to depolarize the nerve, or it might be the consequence of some internal tendency to increase the membrane potential. Having encountered signs of a steady-state control of the membrane potential of the terminal arbor, we began to wonder what role the small diameter peripheral nerve fibers might play in this. It seemed, a priori, that the very slow conduction velocity of these fibers would rule out their role in fast phasic actions, particularly in large animals. We wondered if they might be exerting their influence on some tonic mechanism such as the one just described, which can gxerate a steady dorsal root potential. RpferrnwS p . 1 / 4 / 1 1 5

P. D. W A L L

104

I

1

I

I

I

I

I

I

I

I

Fig. 7. Recording of passive dorsal root potential on cut rootlet of L7 in spinal cat. Stimulus in upper trace maximal to lateral popliteal nerve evoked an afferent volley in all types of fibers and resulted in the usual sequence of potential changes. In the lower trace, the same stimulus was applied to the peripheral nerve, but the large diameter fibers had been blocked by anodal polarization, allowing only impulses in fine fibers to arrive at the cord. This volley results in a hyperpolarization. Time marks, 20 msec.

We therefore set about blocking the action of large diameter fibers by anodal polarization of a peripheral nerve (Fig. 7). In the top trace, a single volley of all types of impulses arrives and evokes the expected passive dorsal root potential, consisting primarily of a prolonged depolarization. When we had blocked the large myelinated fibers, the same stimulus which now produced only impulses in fine fibers at the cord resulted in a purely positive dorsal root potential. As the blocking current is increased, the depolarization phase can be seen to decrease while the hyperpolarization phase increases. This suggests the exciting possibility that the large and small fibers from a peripheral nerve have antagonistic effects on the polarization of central terminals, with the larger fibers tending to depolarize, and the finer to hyperpolarize - thus providing a dynamic control mechanism of the level of excitability of the system. We do not know yet the relation of this phenomenon to the DRP VI of Lloyd (1952) or to a phenomenon observed by Lundberg (this Volume, p. 135). Lundberg reports that if the dorsal root potential is held severely depolarized by descending impulses from the head, he has occasionally been able to decrease this depolarization by a volley sent in from the periphery. He relates this relative hyperpolarization to

PRESYNAPTIC CONTROL OF IMPULSES

105

other results (Lundberg and Vyklicky, 1953) where it is believed there must be a way of controlling the generatio!i of dorsal root potentials by some means other than by presynaptic inhibition of the impulses which are producing it. It is too early to decide if Lundberg and Vyklicky are in fact proposing an alternative hypothesis because so far we have been unable to repeat some of the phenomena on which their paper is based. The mechanism by which hyperpolarization leads to increased efficiency of synaptic transmission is just as difficult to decide as the mechanism for presynaptic inhibition. The variation of membrane potential may vary the effect by many means, including variable transmitter release and variable block. Whatever the detailed mechanism, the fact remains that the input-output function appears to be set by the continuous action of two opposing processes so that presynaptic control seems to be exerted by a combination of relative inhibition and facilitation. The facilitation may turn out to be disinhibition. Lastly we should point out a consequence of the presence of a tonic component of the dorsal root potential. A phasic dorsal root potential is non-linearly related to the size of an afferent volley evoking it. This, in combination with the controlled tonic potential, means that variations in the height of a phasic dorsal root potential may be related either to changes of the base line or to changes of the actual maximum attained.

2.

REPETITIVE FIRING I N THE TERMINAL ARBORIZATION

We know that if depolarization of the terminal arbor of cutaneous fibers with respect to their parent axon reaches sufficient intensity, impulses will be generated. These impulses will run antidromically out of the depths and be recorded on dorsal roots as the dorsal root reflex. They occur both in fibers which have recently carried an orthodromic volley and in their passive neighbors. They may occur as a high frequency burst lasting as long as 15 msec and on other occasions may be seen emitted as a continuous low frequency barrage. The particular timing of individual impulses is determined by factors within the fiber carrying them, but their ultimate cause lies outside the fiber (Wall, 1959). Under the usual crude circumstances in which they have been observed, these impulses have been considered as having an inhibitory action by colliding with any impulses which might be coming in following the volley which originally initiated them. Their effect would then be to exaggerate the presynaptic inhibitory action which blocks neighboring inputs and closes down any active line after it has transmitted a large volley. But let us consider a much more subtle situation. Suppose that an entering fiber splits into two separate terminal arbors and further suppose that a dorsal root reflex is generated in only one of these arbors. Then while one of these regions is held blocked, the other is subject to a repetitive barrage. Transmission across one set of synaptic contacts would be reduced, while across the other there would be amplification. No phenomena of this type have yet been observed, but at the same time the use of sudden massive synchronized inputs and the method References p . 1141115

106

P. D. W A L L

of observation has prohibited such observations. This section was introduced only to emphasize that we have the anatomy and physiological properties of terminal arborizations which would permit switching within them, and we should be watching for the appearance of such axonal switching in the vertebrate. 3-7.

CONTROL B Y OTHER METHODS

These were mentioned for reasons of caution and will not be reviewed in detail. There are more possible points of control than the present simple classification into pre- and postsynaptic control mechanisms. There is a strong tendency in the literature to assign any observed change in input-output characteristics of a system to one or the other of today’s popular methods of control. We have just been through a long struggle to establish the existence of a presynaptic control, and so we remain sensitive to denials of the many other possibilities. Only a few of the obvious ones are mentioned in this list. At a time when the method by which dendrites communicate with the axon of a cell is unknown, it would be foolhardy to deny the possibility that dendrites, too, may be under control so that transmission from the tips of dendrites to the final point of impulse initiation may also be modulated with individual dendrites being turned off or amplified. Finally, we should not forget that the output of a cell which affects the cells on which it projects, must include its resting level of discharge and its tendency to fire repetitively. If one examines repetitive firing (Wall, 1959), one finds clear cut evidence that some aspects of repetitive firing are organized within the cell and other aspects are impressed on the cell from the outside. This allows outside structures to control the rate of rise of repetitive discharge and rate of decline and to interlock in various ways the firing pattern of one cell with the response pattern of another. In this way, the polysynaptic modulation of a synaptic region may become more important than the existence of some monosynaptic transmission. The origin of presynaptic control Since the first discovery of dorsal root and dorsal cord potentials, two interrelated arguments have been going on. The first asked what structures were involved i n the generation of the effect, and the second asked what mechanism these structures used to generate the effect. It was clear that the terminals of arriving afferent fibers could be held depolarized, but it was not clear what was doing it or how it was done. One school maintained that terminal arborizations could interact directly without the intervention of interneurones (Barron and Matthews, 1938; Dun, 1941; Brooks and Fuortes, 1952). In opposition, such authors as Hughes and Gasser (1934), Bonnet and Bremer (l938), Eccles and Malcolm (1946), and Lloyd and Mclntyre (1949), maintained that interneurones must be involved. The basic problem, described most carefully by Lloyd and McTntyre (1949), was to explain how a very prolonged depolarization, which appeared in the fibers which had carried the afferent volley, could also appear with equal intensity in passive neighbors.

P R E S Y N A P T I C C O N T R O L OF IMPULSES

107

Barron and Matthews (1938) had proposed an entirely new process to explain the spreid. They suggested that closely packed terminals were not ionically insulated from each other so that massive depolarization of one group could spread to others. Later authors seem to me to have dismissed this suggestion in an entirely arbitrary manner and turned instead to more familiar mechanisms which involved postulated actions of interneurones. A response of interneurones, which received both active and passive fibers, was assumed to be leaking back across the synapse and affecting the terminals of both sets of afferents. One issue above all confuses the interpretation of these earlier papers. It was not recognized that, under the conditions of the experiments, the great bulk of the phenomenon was limited to the cutaneous afferents and very little if any contribution was being made by the low threshold muscle afferents. The considerable difference between these two systems was emphasized by Koketsu (1956) and Wall (1958), and is apparent in the work of the Canberra group (Eccles et a/., 1962) although our emphasis differs. This differentiation of what had been considered a general phenomenon has been continued in great detail by Dr. Eccles and colleagues who now separate three groups: extensor group I afferents, flexor group I afferents, and flexor reflex afferents. The huge difference between the dorsal root potential produced by a single volley in a skin nerve, cutaneous superficial peroneal, and a train of volleys in IA fibers from biceps semitendinosus can be seen, for example, in Fig. 1 of Lundberg and Vyklicky (1963). This large difference led me to concentrate on cutaneous afferents (Wall, 1958, 1962) while Dr. Eccles’ group, because of the much greater knowledge of the details of proprioceptive reflexes and their tremendous experience of motoneuron recording, have concentrated on muscle afferents. We have arrived at apparently contradictory views on the origin of presynaptic inhibition and dorsal root potentials (Wall, 1962; Eccles et a]., 1962). Neglecting the possibility that we may very well both be wrong, it is of interest to examine our two positions. Briefly they attribute the depolarization of primary afferent fibers to the action of deep interneurons, while we suggest the substantia gelatinosa as the origin, at least for the cutaneous fibers. Furthermore, we imply that the action is brought about by quite different mechanisms. Before proceeding with a detailed analysis we must first decide if both groups were dealing with the same phenomenon because there is an obvious difference of procedure where they use a far more refined manipulation of various specific inputs. A single low voltage shock to a dorsal root produces a large dorsal root potential which must be attributed primarily to a cutaneous afferents and since the results of Eccles et al. (1962) specifically apply to these afferents we can assume that we have both been dealing with the same potential. They suggest the presence of a second related mechanism responsible for the action of la and b fibers from flexors on la afferents, but since a single shock to this system produces only a minor disturbance, I have neglected it in my analysis. Their results on group I1 muscle afferents would not be involved in my experiments since stimulus strengths necessary to fire these fibers were never used. Therefore we can conclude that the two sets of experiments should be comparable in so far as they deal with low threshold flexor reflex afferents, particularly a cutaneous fibers. References p . 114/115

108

P. D. W A L L

First, we can ask where does the disturbance actually lie during generation of the depolarizstion? We know that the terminals are being depolarized by something, but we want to know how activity is distributed at this time. Fig. 2 in this paper and Fig. 3 in Eccles et a/. (1 962) show the results of such an attempt. The results, as far as they go, are roughly the same, but differences of method and attitude lead to our different interpretations. Their map unfortunately does not cover the lateral third of the cord where we believe the most interesting events are taking place. They have mapped on a 250 EL grid while we felt it necessary to use a 100 p grid and sometimes even more recording points. They have located the position of only one electrode and assumed that the others lie parallel t o it, while we have never been able to trust this procedure because the electrodes are sufficiently flexible never to penetrate on exactly parallel courses. Finally and most important, we believe for the reasons stated in Howland et al. (1955) that the meaning of these voltage contour maps can only be derived from calculating the source-sink distribution. When this is done, it will be seen that the major sinks lie in the distribution of substantia gelatinosa and at the same time there are some smaller scattered sinks in the depths of the dorsal horn. Their interpretation is quite different since they say ‘the contour diagram conforms with a field potential produced by an active depolarization of the cutaneous fibers in the spinal cord at a depth of 1.2-2.0 mm’. We cannot agree with this interpretation because the distribution of active regions, sinks, does not conform with any known

Fig. 8. Source-sink current map as shown in Figs. 2 and 3 with dots representing the sinks where activity is present. The dorsal columns and substantia gelatinosa had been completely sectioned between L6 and L7, but Lissauer’s tract and other white matter was intact. An afferent volley was fired into the intact segment L7 10 msec (A), 20 msec (B) and 30 msec (C) before the current distribution was analyzed in the deafferented segment L6. 10 msec after the stimulus, a dense group of current sinks were located along the border of the lateral white matter and in the region of Lissauer’s tract. In B and C the more ventral activity fades, while the sinks in the region of the Lissauer tract extend medially into the region of the substantia gelatinosa. The height of the dorsal root potential is reached at C and subsequent maps show a slow fading of this pattern without further changes of location. Section of the Lissauer tract between L6 and L7 abolishes the dorsal root potential and the activity in the substantia gelatinosa, but leaves the more ventral activity unaffected. (Reprinted from Wall, 1962.)

PRESYNAPTIC CONTROL OF IMPULSES

I09

anatomy of terminal arbors. The situation becomes much clearer if the inputs to the rzgion are greatly simplified. In order to examine the role of active afferent fibers in the generation of the dorsal root potential an area of spinal cord was prepared to which no active afferent fibers ran. This was done by sectioning the dorsal column. It was found that this segment, isolated from any actual entering A fibers, generated a perfectly good dorsal root potential. The distribution is shown in Fig. 8. It will be seen that initially there is activity in the region of the Lissauer tract and also deeper in the grey matter. The Lissauer activity spreads medially to involve the whole region of substantia gelatinosa while the deeper disturbance dies. The activity in the region of substantia gelatinosa remains for the whole duration of the slow potential. If the Lissauer tract is cut between the segment where the volley enters and the recording segment, there is a complete abolition of the dorsal root potential and of the substantia gelatinosa activity. In contrast, the deep activity continues and is presumably relayed by propriospinal fibers in the lateral white matter and has nothing to do with the generation of the presynaptic depolarization. The time course and location of this deep activity fits well with the D cells recorded by Eccles et al. (1962) and which they suggest are the actual generators of the dorsal root potential. In other respects the experiments have concentrated on different directions since while ours was directed to finding the distribution of sources and sinks specifically related with the generation of presynaptic inhibition, theirs was directed in a search for interneurons fired by those inputs which were known to produce it. They have succeeded in this quest in the discovery of D cells, but the firing pattern of the D cells implies a mechanism by which they might generate the presynaptic effect and we must now discuss the mechanism. The very small cells of substantia gelatinosa and the small cells elsewhere in the cord have so far defied attempts at unit analysis so that WE cannot ask them if they also fulfill the criteria of responding to the right inputs. We can only look at their apparent activity en nzasse and suggest that they do. The D cells are polysynaptic cells with respect to the input, whereas the substantia gelatinosa cells may well be fired monosynaptically by the input. The D cells fire for a short high frequency burst during the rising phase of the depolarization and therefore the implication is that they secrete a burst of depolarizing transmitter substancc onto the presynaptic terminals during this phase and the remaining prolonged depolarization is taken up with the removal of this transmitter. There is a general disadvantage to a mechanism of this sort and that is that once generated there is no way of turning it off except by over-riding it and we know of no control system in biology which works i n such a one-sided fashion. There are two specific objections to their suggested mechanism. First, it has been known for a long time that barbiturate increases the amplitude and extent of the dorsal root potential and at the same time decreases the firing of the large cells in the dorsal horn. This unfortunate combination of effects has made it necessary for them now to add the additional hypothesis that barbiturate must somehow be potentiating their depolarizing transmitter substance, while at the same time reducing the amount produced. The second specific objection concerns the interpretation of the positive dorsal root potential. We have shown above (Fig. 5) that a volley arriving at the cord, carried only by fine fibers, generates Referenrrs p . 1 141I I5

110

P. D . W A L L

a positive dorsal root potential. They would have two possible ways of explaining this: either there is a hyperpolarizing transmitter system or more simply, the relative positive shift is the consequence of a turn off of a continually active generator of depolarization. If the latter is the case the decline of the recorded potential could not have a time course faster than the decline of the negative dorsal root potential since both movements would be the consequence of an arrest of transmitter release and its subsequent removal. In fact we find that the time constant of the onset of the positive dorsal root potential is considerably faster than that of the decline of the negative potential. This suggests that the simplest explanation is that the membrane potential is controlled by a mechanism which can be turned both on and off fairly rapidly. Therefore in summary we present the following tentative hypotheses which suggest further experiments and which fit the known data on the origin of presynaptic control. HYPOTHESES

Hyypothesis I : The membrane potential of cutaneous aflerent terminals is controlled by substantia gelatinosa cells. This part of the hypothesis rests on two types of experiments. The first is the analysis of the source-sink distribution of current within the spinal cord during the tail of dorsal root potentials and presynaptic inhibition. The only significant areas of activity which can be detected lie in the region of the substantia gelatinosa. The only other likely structures in this region are the terminal arborizations of the afferent fibers themselves, but we know that their distribution is much more extensive than the limited distribution of the sinks of activity. The second experiment involves the generation of a dorsal root potential in a segment where no active afferent fibers end. I n this case, as we have described above, the dorsal root potential is completely abolished by section of the Lissauer tract. This tract is, according to Szentiigothai and described i n Volume 11 (p. 155), concerned with interconnecting one part of substantia gelatinosa with another. He also believes, with Earle (1952), that its more medial part contains some C fiber collaterals, but we are not concerned with these fibers here because the peripheral stimulus was adequate only for the A u-fibers. Transecting the substantia gelatinosa and Lissauer tract isolates one part of substantia gelatinosa from that part which received the afferent signal. A consequence of this isolation is that no dorsal root potential or terminal depolarization occurs although many deeper cells are still responding. Additional evidence in support of the substantia gelatinosa as a generator comes from an analysis of the various different cord segments. The size of the substantia gelatinosa and of the evoked dorsal root potential r:mains remarkably constant over the whole range of segments from upper lumbar to coccygeal while the size of all other components varies over a very wide range (Wall, 1962). The weakness of this part of the hypothesis is that we have not been able to study any small cells individually and cannot test them for the required properties of convergence as was done for the D cells of Eccles et al. (1962). I n the absence of

PRESYNAPTIC CONTROL O F IMPULSES

111

techniques to study in detail the fine terminals, fine dendrites and small cells, any hypothesis which attributes a phenomenon to one of these components remains weak. For this reason it is always possible to invent special unknown properties of the other components of the region which would explain the phenomenon, but it is apparent that the firing of the substantia gelatinosa cells remains the simplest suggested mechanism since it requires no new properties except that their firing must somehow be able to influence the afferent fibers. This requires some sort of axon-axonal contact, but the anatomy of the region is not yet sufficiently well studied to discuss the anatomical basis for the interaction. Hypothesis 2: Subsiantia gelatinosa cells are excited monosynapiically by large afJerents . There is no direct physiological evidence for this suggestion, but anatomically it is clear that the endings of the large cutaneous afferents end directly in the region of the dendrites and cell bodies of the small cells. It is also certain that a small fraction of the A a-fibers in a root can evoke a maximal dorsal root potential in the region of their entry. If the first hypothesis is accepted, then it is not necessary to involve cells other than the arriving afferents and the small dorsal cells. It is true, of course, that the A afferents certainly also produce monosynaptic discharges in the large cells in Rexed’s layer IV. However, we do not want to involve these cells in the generation of the cutaneous dorsal root potential for one of the same reasons for which we do not want to involve the D cells of Eccles et al. The large cells are sensitive to barbiturate and can be given such heavy doses that their firing fails completely at a time when some dorsal root potential is still generated. Furthermore, their repetitive firing increases as all A components are recruited to their afferent bombardment while the dorsal root potential rapidly saturates. Finally, post-tetanic potentiation has a minor effect on these cells and a major one on the dorsal root potential. In other words, the input-output characteristics of the large cells do not follow that of the dorsal root potential. It is now known that many descending volleys can generate a dorsal root potential (McCulloch et al., 1952; Hagbarth and Kerr, 1954; Andersen et al., 1962; Carpenter etal., 1962; Taub, 1963; Lundberg, this Volume, p. 135, etc.). In one case (Taub, 1963) it WJS shown that a descending volley from the mid brain reticulum close to the red nucleus will evoke a dorsal root potential which occludes with that produced from a dorsal root and is therefore presumably generated by the same mechanism. In the frog, even an antidromic ventral root volley will generate a dorsal root potential (Barron and Matthews, 1938) although it should be remembered that in these animals the motoneurone dendrites extend up to the dorsal root entry region. It is too early to say if all dorsal root potentials are generated by the same mechanism and how such arriving volleys might affect the activity of substantia gelatinosa. Any candidates for the generation of presynaptic inhibition must be shown to respond not only to the relevant peripheral inputs, but also to the intersegmental influences and to descending volleys. We should also remember that substantia gelatinosa does not have a monopoly of small cells of unknown function. Similar small cells are scattered throughout Rrfercnces

p . 1141115

112

P . D. W A L L

the nervous system including the cortical Golgi type I1 cells. If it turns out that part of the dorsal root potential is generated by substantia gelatinosa and other parts by deeper cells, then it will be necessary to investigate the function of the deep small cells.

Hypothesis 3: Substantia gelatinosa is excited by other parts of substantia gelatinosa. It is suggested that after the excitation of one group of cells by an arriving cutaneous volley, the excited cells bombard their neighbors and a massive interdependent discharge grows up inthe region, spreads to neighboring regions and then slowly dies down. The anatomy for such a discharge is known and includes the massive short interconnections between the cells and the longer ones by way of Lissauer’s tract. The spread of the dorsal root potential from one side of the cord to the other has always been a mystery which is solved if one may consider the substantia gelatinosa as a continuous structure crossing the midline. In the sacral segments it is obviously a single lamina. In the lumbar and more rostra1 regions it is split by the wedge of the dorsal columns but there are clear signs that it retains connections from one side to the other. The spatial distribution and time course of the dorsal root potential fit well with the suggested mutual bombardment of neighboring small cells. Thus the dorsal root potentials in one root depend on the afferent signals that have arrived in that segment and on the activity that has spread from neighboring segments. The evidence for this is the slowness of the onset of the DRP if the segment is isolated from arriving afferents and the abolition of the slow DRP by Lissauer tract section. The contralateral DRP has all the properties of the indirect DRP evoked by the Lissauer tract. Hypothesis 4: Substantia gelatinosa is inhibited by fine afferents. The evidence for this is that a volley in fine fibers which is not accompanied by a volley of A fibers generates a purely positive dorsal root potential. It is suggested that this is generated by the turning off of a steadily generated negative dorsal root potential which is produced by the steady activity of substantia gelatinosa cells which are maintained in their state of activity by the steady afferent barrage i n the excitatory fibers. This mechanism for the generation of the positive DRP is at present purely speculative, but it is tempting to point to the anatomy shown in the diagram Fig. 1. It will be noted that the terminal arborizations of the large fibers approach the small cells from the ventral direction while the fine fibers penetrate directly from the dorsal direction. This opposing direction of terminal arborizations together with the opposing effect of their volleys on the dorsal root potential is suggestive. It is also reminiscent of the situation in the accessory olive where afferents from one cochlear nucleus end on one end of the cell and excite while afferents from the other side end on the other end of the cell and inhibit. We have mentioned the other example of a positive DRP which always follows a large negative DRP (DR VI, Lloyd, 1952). It would seem reasonable to guess that this represents the postexcitatory depression which so often characterizes the end of an after-discharge and is the sign of the inhibitory process which has cut short the after-discharge.

PRESYNAPTIC CONTROL OF IMPULSES

113

Hypothesis 5: Barbiturate preferentially blocks the inhibition. We have mentioned several times above that one of the most troublesome phenomena to explain for the dorsal root potential is that it is prolonged and increased by barbiturate. We have now shown that the purely positive dorsal root potential evoked by fine fibers is completely abolished by barbiturate. It would seem reasonable to suggest then, that the DRP normally observed is produced by a balance between excitatory and inhibitory effects and that barbiturate blocks the latter. It is further necessary to postulate on this hypothesis that the excitation of the cells of substantia gelatinosa is relatively insensitive to barbiturate. It is not too unreasonable to suggest a special pharmacology for these cells. It is certain that their chemistry must be different from other nerve cells since they resist staining with all ordinary nerve cell stains. It is reasonable that barbiturate should have this effect since the consequences of its action will be to raise the level of presynaptic inhibition to the full level which can be maintained by the ongoing generators. Finally, let me suggest that perhaps none of us who have been looking at presynaptic inhibition and the dorsal root potentials have been doing the right experiments to demonstrate its functional importance to the free living animal. First, we should remember that it is a very slow mechanism by comparison with the fast reflex circuits and relay systems. It comes on slowly and remains on for long periods of time. Perhaps we should not be examining it with sudden synchronized inputs, but rather studying it with slow tonic variations of input. Secondly, it has appeared in the past to be a one-sided control system only turning off the input sluggishly and for long periods. Perhaps now that a movement in the opposite direction has been detected we can look for signs of its action as a dynamic control of some aspect of the input. Thirdly, the most unlikely and unbiological aspect of this presynaptic mechanism is its indiscriminate all-pervading action on almost all inputs. This apparent result may very well be due to the crudity of stimuli and recording methods. It is to be hoped that by paying attention to these three aspects of presynaptic inhibition and the experimental technique used to investigate it, it will become an even more interesting phenomenon than it has been in the past. SUMMARY

We have shown that nerve impulses entering the spinal cord in large cutaneous nerve fibers come under the influence of a control mechanism while the impulses are still in the terminal arborization of the afferent fiber. There is evidence for a tonic presynaptic control mechanism which determines the effectiveness of entering impulses on central cells. The control is exerted on the membrane potential of the terminal arbor. Depolarization may produce blockade of the terminal arbor and hyperpolarization may increase the effectiveness of synaptic transmission. We show that impulses in large fibers result in a depolarization of the terminal arborization, while impulses in fine fibers result in a relative hyperpolarization. There is evidence for continuous action of this control mechanism so that phasic events modulate the References p . 1141115

114

P. D. WALL

position of the control mechanism and therefore, the sensory posture. We present the evidence that the cells of substantia gelatinosa mediate this control over entering sensory information. ACKNOWLEDGEMENTS

The author is greatly indebted to Dr. J. Y. Lettvin, Dr. W. S. McCulloch and W. H. Pitts for their continuous help throughout this work and for their initiation of the emphasis here on presynaptic properties. The work was supported by the U. S. Army Signal Corps, the Air Force Office of Scientific Research, and the Office of Naval Research; in part by The Teagle Foundation, Inc. and Bell Telephone Laboratories, Inc.; and in part by the National Institutes of Health, Bethesda, Md. (Grant B-1865(C3) and Grant MH-04737-03) and the National Science Foundation (Grant G16526). REFERENCES ANDERSEN, P., ECCLES,J. C., AND SEARS,T. A., (1962); Presynaptic inhibitory action of cerebral cortex on spinal cord. Nature (Land.), 194, 740-743. BARRON, D. H., AND MATTHEWS, B. H. C., (1935); Intermittent conduction in the spinal cord. J. Physiol. (Land.), 85, 73-103. BARRON, D. H., AND MATTHEWS, B. H. C., (1938); The interpretation of potential changes in the spinal cord. J. Physiol. (Land.), 92, 276-321. BONNET,J., ET BREMER, F., (1938); Etude des potentiels electriques de la moelle Bpiniere chez la grenouille spinale. C . R . Sac. Biol. (Paris), 127, 806-812. BROOKS,C. McC., ECCLES,J. C., AND MALCOLM, J. L., (1948); Synaptic potentials of inhibited motoneurones. J. Neurophysiol., 11, 417430. BROOKS, C. McC., AND FUORTES, M. G. F., (1952); The relation of dorsal and ventral root potentials to reflex activity in mammals. J. Physiol. (Land.), 116, 380-394. CARPENTER, D., ENGBERG, Q., AND LUNDBERG, A., (1962); Presynaptic inhibition in the lumbar cord evoked from the brain stem. Experientia (Basel), 18, 450. W. W., AND RITCHIE,J. M., (1962); Mammalian nonmyelinated nerve fibers. Physiol. DOUGLAS, Rev., 42, 297-334. DUN,F. T., (1941); The latency and conduction of potentials in the spinal cord of the frog. J. Physiol. (Land.),100, 283-298. EARLE,K. M., (1952); The tract of Lissauer and its possible relationship to the pain pathway. J. camp. Neural., 96, 93-1 1 1 . ECCLES,J. C., (1961); The mechanism of synaptic transmission. Ergebn. Physiol., 51, 299430. ECCLES,J. C., ECCLES,R. M., AND MAGNI,F., (1961); Central inhibitory action attributable to presynaptic depolarization produced by muscle afferent volleys. J. Physiol. (Land.), 159, 147-1 66. ECCLES, J. C., KOSTYUK, P. G., AND SCHMIDT, R. F., (1962); Central pathways responsible for depolarization of primary afferent fibers. J. Physiol. (Lond.), 161, 237-257. ECCLES, J. C., AND MALCOLM, J. L., (1946); Dorsal root potentials of the spinal cord. J. Neurophysiol., 9, 139-160. FRANK, K., (1959); Basic mechanisms of synaptic transmission in the central nervous system. I. R. E. Tram. med. Electron., ME-6, 85-88. FRANK,K., AND FUORTES,M. G. F., (1956); Unitary activity of spinal interneurons in cats. J. Physiol. (Land.), 131, 424435. FRANK,K., AND FUORTES, M. G. F., (1957); Presynaptic and postsynaptic inhibition of monosynaptic reflexes. Fed. Proc., 16, 3940. HAGBARTH, K. E., AND KERR,D. I. B., (1954); Central influences on spinal afferent conduction. J. Neurophysiol., 17, 295-307. HAGIWARA, S., AND TASAKI,I., (1958); A study of the mechanisms of impulse transmission across the giant synapse of the squid. J. Physiol. (Land.), 143, 114-137.

PRESYNAPTIC CONTROL OF IMPULSES

115

HOWLAND, B., LETTVIN, J. Y . , MCCULLOCH, W. S., PITTS,w. H., AND WALL,P. D., (1955); Reflex inhibition by dorsal root interaction. J. Neurophysiol., 18, 1-17. HUGHES,J., AND GASSER, H. S., (1934); Some properties of the cord potentials evoked by a single afferent volley. Amer. J. Physiol., 108, 295-306. HUNT,C. C., AND MCINTYRE, A. K., (1960); An analysis of fiber diameter and receptor characteristics of myelinated cutaneous afferent fibers in the cat. J. Physiol. (Lond.), 153, 99-112. IGGO,A., (1959); A single unit analysis of cutaneous receptors with C afferent fibers. Pain and Irch nervous Mechanisms. Y . Zotterman, Editor. London. Churchill (pp. 41-56). KOKETSU, K., (1956); Intracellular potential changes of primary afferent fibers in spinal cords of cats. J. Neurophysiol., 19, 375-392. KOLMODIN, G. M., (1957); Integrative processes in single spinal interneurones. Acra physiol. scand., 139, SUPPI.40, 1-89. KUNO,M., (1963); The quanta1 nature of monosynaptic transmission in spinal motoneurons of the cat. Physiologist, 6, 219. LLOYD, D. P. C., (1949); Post-tetanic potentiation of response in monosynaptic reflex pathways of the spinal cord. J. gen. Physiol., 33, 147-170. LLOYD, D. P. C., (1952); Electronics in dorsal roots. Cold Spr. Harb. Symp. quant. Biol.,17, 203-219. LLOYD,D. P. C., AND MCINTYRE, A. K., (1949); On the origin of dorsal root potentials. J. gen. Physiol., 32, 409443. LUNDBERG, A., AND VYKLICKY, L., (1963); Inhibitory interaction between spinal reflexes to primary afferents. Experieniia (Basel), 19, 247, 1 4 . MCCULLOCH, W. S., LETTVIN, J. Y., PITTS,W. H., AND DELL, P. c . , (1952); An electrical hypothesis of central inhibition and facilitation. Res. Publ. Ass. nerv. men?. Dis.,30, 87-97. RENSHAW, B., (1946); Observations on the interactions of nerve impulses in the gray matter and the nature of central inhibition. Amer. J. Physiol., 146, 443448. REXED, B., (1952); The cytoarchitectonic organization of the spinal cord in the cat. J. comp. Neurol., 96, 415466. TAKEUCHI, A., AND TAKEUCHI, N., (1962); Electrical changes in pre- and postsynaptic axons of the giant synapse of Loligo. J. gen. Physiol., 45, 1181-1 193. TAUB,A., (1963); Local segmental and supraspinal interaction with a dorsolateral spinal cutaneous afferent system in the cat. Thesis. Massachusetts Institute of Technology. TOENNIES, J. F., (1939); Conditioning of afferent impulses by reflex discharges over the dorsal roots. J. Neurophysiol., 3,5 15-525. WALL,P. D., (1958); Excitability changes in afferent fibre terminations and their relation to slow potentials. J. Physiol. (Lond.), 142, 1-21. WALL,P. D., (1959); Repetitive discharge of neurons. J. Neurophysiol., 22, 305-320. WALL,P. D., (1960); Cord cells responding to touch damage and temperature of skin. J. Neurophysiol., 23, 197-210. WALL,P. D., (1962); The origin of a spinal cord slow potential. J. Physiol. (Lond.), 164, 508-526. WALL,P. D., AND JOHNSON, A. R., (1958); Changes associated with post tetanic potentiation of a monosynaptic reflex. J. Neurophysiol., 21, 148-1 58. WALL,P. D., MCCULLOCH, W. S., LETTVIN, J. Y., AND PITTS,W. H., (1955); Factors limiting the maximum impulse transmitting ability of an afferent system of nerve fibres. 3rd London Symposium on Information Theory. London. Butterworth (pp. 329-344). DISCUSSION

LUNDBERG: We have also found positive dorsal root potentials in our case on stimulation of the brain stem. I think it can be assumed that they are caused by primary afferent hyperpolarization. We cannot exclude that there may be a synaptic hyperpolarizing action, but a more likely explanation is inhibition of reflex paths and cessation of a steady depolarization caused by spinal reflex action. With respect to your findings I think it is important to realize that you are recording not only from cutaneous afferents but also from muscle afferents. Dr. Vycklickf and I have recently found that volleys in the FRA inhibit the pathway from l a to Ib

116

DISCUSSION

afferents. If the Ia fibers are slightly depolarized by synaptic bombardment a volley in the FRA could be expected to evoke primary afferent hyperpolarization in Ia fibers. WALL:1 am very happy that you have also seen a positive dorsal root potential. We must be careful, when speaking of phasic effects on a tonic mechanism, to define the state of the tonic mechanism before the observation is made. In the example you have shown us, the presynaptic inhibition is already fully established and saturated so that the only possible effect of a mixed volley is inhibition of the inhibition.

ECCLES:1 have several difficulties with your story that the generation of the dorsal root potential takes place i n the substantia gelatinosa. Firstly, 1 am concerned with the detailed mechanism by which the interneurons are supposed to depolarize the presynaptic terminals. Thus George Gray has informed me that he finds axo-axonic synapses to be very rare in the substantia gelatinosa, whereas they are extremely numerous deeper in the dorsal horn in laminae 4 to 6. Secondly 1 would point out that a considerable fraction of the dorsal root potentials produced by dorsal root volleys are generated in the large muscle afferent fibers. So far as I am aware the large muscle afferents do not send collaterals into the substantia gelatinosa, so at least subsidiary zones for primary afferent depolarization have to be envisaged. Actually the depolarization of these large afferent fibers occurs at their terminals in lamina 6 for Ia and Ib fibers, and also in the motor nucleus for the group Ia fibers. WALL:It is open season and the hunting is good with the electron microscope. You choose only to see synapses where the bubbles appear, but 1 can’t believe that there is only one type of anatomical synapse in the whole animal kingdom. There is no physiological evidence which allows one to decide that large areas of intimate cellcell contact are functionless because they do not have presynaptic vesicles or desmosomes. I think we need to know a lot more physiology and detailed anatomy before we can inspect electron micrographs and state categorically that one cell does not affect its neighbor. When the electron microscopists themselves cannot agree on the criteria for the identification of an anatomical synapse, it is premature for physiologists to take only one criterion. Dr. Robertson, for example, (Am. N . Y . Acad. Med., 94 (1961) 339-389) shows a synapse without presynaptic vesicles or desmosomes. On the second part of your question, I have spoken here only about the cutaneous pathway and 1 specifically do not want to generalize to include the muscle afferents. 1 had chosen the cutaneous system because the effect is very large for a single volley as compared with the proprioceptive effect and because the anatomy within the cord is easier to analyze. However, if you will accept the story I have told for the cutaneous afferents, one can speculate with three alternate hypotheses : the proprioceptive afferents might affect the substantia gelatinosa by indirect pathways; or there may be scattered small cells lying about the endings of the proprioceptive afferents similar i n property to the substantia gelatinosa cells but not organized into a specialized lamina ; or, lastly, there may be a quite different mechanism as you have suggested. We are

PRESYNAPTIC CONTROL OF IMPULSES

I17

dealing here with the classical problem of the function of a neuropil. We should not be satisfied until we have assigned a function to all the anatomical structures within these general junction regions. We should not try to force a simple telephone switch board model onto a structure which obviously has far subtler potentialities given by its anatomy. I am suggesting that the very small nerve cells in neuropil may play a modulating role on transmission. GELFAN: It sounds as if there is some fear of saying that one neuron can influence another by means other than synapses, as if that is heresy to mention that. In this respect I might remind you that the first explanation proposed by Renshaw for inhibition was the effect of current flow from the activated motoneurons upon their neighbouring ones. Now that is just about the situation we have when we are artificially stimulating nerves in our laboratory. I don’t know why we should be afraid to say that axons, dendrites or cell bodies can influence each other due to the enormous amount of current flow. I don’t know of any reason why the current flow of one neuron cannot influence the other, depending upon the polarity that can shift the membrane equilibrium either down or upwards as to whether it is going to be excitation or inhibition. WALL: I quite agree with you that we should consider many different types of synaptic interaction. However, I would like to make quite clear that while our experiments have often measured fields of electric current in the central nervous system, we have made those measurements only in order to locate the activity at any one time and there is absolutely no implication in our experiments that the field which we measure is itself the stimulus to other cells. Our experiments were not designed to investigate the nature of synaptic transmission, but only to discuss which structures were involved and the nature of the control system.

JUKES: I would like to make a point regarding the mechanism you talked about. It seems to me that your explanation does not apply to all synapses in the central sensory pathways. The inhibition which one can see in the dorsal column nuclei seems to be totally different. It has a different latency, it only occurs in certain cells, and it can be evoked with very small mechanical stimuli. It is rather often very difficult to see it with electrical stimulation because it seems to jam the mechanism if you stimulate the nerves electrically. I would be pleased to know your comments about this. WALL:As I said to Dr. Eccles, it would be foolish to generalize from the system I have studied to any other, but you raise a most interesting point I would like to discuss. In 1958, I showed that there was a prolonged depolarization of terminals in the nucleus gracilis following an afferent volley. This depolarization had the same general properties as the one in the dorsal horn. There is, therefore, a crude indication of a similar presynaptic control mechanism as that seen in the cord, but as you and your colleagues have shown, it has some very different properties. Recently, Taub (1963)

1 I8

DISCUSSION

has shown in the dorsal horn cells that there are inhibitions from discrete areas of skin on receptive fields of the cells in the lamina, four which remind one somewhat of the type of result you have seen in the dorsal column nuclei. I believe that there may be some massive general interactions demonstrated by crude electrical stimulation, but far more interesting and specific interactions shown by limited natural stimuli. I would further guess that the apparent subdivisions of the laminar structures in the dorsal cord may form the anatomical substrate for these intimate interactions. I think the crude experiments I have described are useful only in showing a new possibility in the repertoire of cell-cell interaction but that the actual role of this mechanism requires far more refined experiments.

I19

The Pharmacology of Presynaptic Inh i bition R. F. S C H M I D T Institut Jur Allgemeine Physiologie, Universitat Heidelberg, Heidelberg (Germany)

Although it seems well established that transmission at the synapses of the vertebrate central nervous system is mainly brought about by transmitter substances liberated from the presynaptic endings only little is known about the possible chemical structure of these transmitters (for review see Perry, 1956; Paton, 1958; Curtis, 1961, 1963; Curtis et al., 1961a; Florey, 1961; McLennan, 1961; Eccles, 1962). There is one exception: it has been shown that the synaptic endings of the axon collaterals of motoneurones at Renshaw cells liberate acetylcholine on activation (Eccles et al., 1954, 1956; Curtis and Eccles, 1958; Longo et al., 1960; Curtis et al., 1961a). It is interesting to note that the evidence for this statement is entirely pharmacological for it demonstrates that pharmacological studies may add considerably to our understanding of the function of the vertebrate spinal cord. This review is concerned with the pharmacology of presynaptic inhibition. Our interest will mainly be focussed on a comparison of the pharmacological properties of pre- and postsynaptic inhibition and further on the action of various anaesthetics on presynaptic inhibition. The evidence relating to the synaptic mechanism and the functional organisation of presynaptic inhibition has been reviewed by Eccles (this Volume, p. 65). A study of Eccles’s report will facilitate the understanding of this paper. Cats and toads were used for the experiments. The cats were decerebrate or lightly anaesthetized with pentobarbital sodium, and the spinal cords were severed at the second lumbar segment. Decerebration was done by jntercollicular section in animals anaesthetized by ether. Peripheral nerves of the hind limb were set up for stimulation (for details see Eccles et al., 1963d). The toad spinal cords were isolated and kept in a perspex chamber of small volume. The lumbar segments with their dorsal and ventral roots were used for stimulating and recording (for details see Schmidt, 1963). Fig. 1 shows diagrammatically the main postulated features of the presynaptic inhibitory pathway together with the principal recording techniques which have been used during these studies. The diagram indicates only one of the several systems of presynaptic inhibition, namely the pathway from Group I afferents to Group Ia afferent fibres. If a volley in a flexor nerve of the hind limb, such as the nerve to anterior-biceps-semitendinosus muscles, enters the spinal cord it will - among many other actions - activate a series of at least two interneurones on the pathway to the inhibitory synapse upon the terminals of Ia afferent fibres near the point where they Rpferences p. 1301131

120

R. F. S C H M I D T

make synaptic contact with motoneurones. Liberation of transmitter at this presynaptic synapse will depolarize the primary afferent fibre at its terminal and this depolarization, henceforth called ‘primary afferent depolarization’, can be recorded as a positive potential wave from the cord dorsum (P-wave) or as the negative going dorsal root potential (DRP) from a dorsal rootlet (Barron and Matthews, 1938;

Fig. 1. Schematic drawing of a spinal cord section showing the postulated pathway for presynaptic inhibitory action on a Group l a primary afferent fibre. Three Group I afferents with their monosynaptic connections on interneurones and on a motoneurone are shown. The interneurone that has a presynaptic connection on a Group Ia fibre can be activated by Group Ia afferents from several flexor muscles. The principal recording techniques used for studying pharmacological actions on presynaptic inhibition are also indicated. For further explanation see text.

Eccles et al., 1962a, 1963b, c). The depolarization of the primary afferent fibre will reduce the size of the spike in the presynaptic terminal and thereby decrease the transmitter output. In a given pool of motoneurones, reduction of excitatory inflow will result in a smaller number of motoneurones being excited, and this can be detected by recording the size of a monosynaptic reflex test spike (VRR, cf. Eccles et al., 1962c) from the ventral roots. There are several other methods of investigating presynaptic inhibitory actions such as the intracellular recording from motoneurones (Frank and Fuortes, 1957; Eccles et a/., 1961) or primary afferent fibres (Eccles et al., 1962b, 1963a, b, c) and the testing of the increased electrical excitability of the presynaptic fibres during depolarization (Wall, 1958; Eccles et al., 1962b). However, the recording of the

121

THE PHARMACOLOGY OF PRESYNAPTIC INHIBITION

dorsal root potential and the testing of presynaptic inhibitory action by a monosynaptic reflex are the only procedures that are sufficiently stable over the long periods of time that are required for the full development of a drug action and even for its decline. In the isolated toad spinal cord, the same technique of recording the primary afferent depolarization as the dorsal root potential has bsen used, but neighbouring dorsal or ventral roots were stimulated instead of pzripheral nerves. The pathway of presynaptic inhibition is probably short, including only a few interneurones (Eccles et a/., 1962a). There are, however, still numerous sites at which a pharmacological substance could act. Since we are registering only the endproduct of a chain of events our ability to identify the exact point of action of a particular drug is rather limited. Nevertheless, a great deal of information has been obtained by our tests about the pharmacological effects on primary afferent depolarization and on presynaptic inhibition. To facilitate the understanding of the individual results it should first be stated that the presynaptic depolarization produced in the cat by several varieties of input: Group I flexor volleys, Group I extensor volleys, and cutaneous volleys, and the primary afferent depolarization produced in the isolated toad spinal cord by dorsal root stimulation all were similarly affected by all of the drugs tested. Furthermore, it should be made clear, that the physiological significance of the primary afferent depolarization in the toad is not yet known (cf: Schmidt, 1963). But, as a result of these experiments, it can at least be stated that it is produced by 120 r

A

3 PBST +GS

I 0 CONT. t O.lrng STRYCH

0

PBST

loo--

% 80

2of

Q+

70 0

I

I

100

200

80

-

60

-

40

-

/kg

20 r

1

0

msec300

0

C O . l m g STRYCH /kg

L

I

2

4

I

msec

6

Fig. 2. Action of strychnine on the presynaptic and postsynaptic inhibition of monosynaptic reflexes. Cat spinal cord in situ. In A the presynaptic inhibitory action of three posterior-biceps-semitendinosus (PBST) volleys (maximum for Group I and at 300/sec) was tested by monosynaptic reflexes evoked by a gastrocnemius (GS) volley and recorded monophasically in the S1 ventral root. The sizes of the test reflexes were calculated as percentages of the control reflex and plotted against the testing intervals (first posterior-biceps-semitendinosus volley to the gastrocnemius volley) t o give the inhibitory curve. At each testing interval there were several superimposed traces and the means were plotted. At the same time the postsynaptic inhibitory curves were determined in B, one quadriceps (Q) Group I afferent volley inhibiting the monosynaptic reflex produced by a posterior-bicepssemitendinosus volley. Abscissae give intervals between the entry of the quadriceps volley into the cord and the onset of the testing reflex spike. In both A and B the open circles plot inhibitory curves in the decerebrate unanaesthetized preparation, the filled circles the curves after the same i.v. injection of 0.1 mg/kg strychnine. The small late facilitation in B is presumably due to the effect of strychnine in enhancing a polysynaptic excitatory action on the posterior-biceps-semitendinosus motoneurones. (From Eccles et al., 1963d.) Rpferences p. 130/131

122

R. F. S C H M I D T

a mechanism quite similar to that of the primary afferent depolarization produced in the cat spinal cord and that in all probability it has the same physiological significance. Convulsants Postsynaptic inhibition is depressed by some convulsant drugs, notably strychnine, and not by others and it has been suggested that the convulsant activity of strychnine can be explained by the depression of postsynaptic inhibition (Bradley et al., 1953; 120

-

A 4PDP

010

80

-+

120

I

PBST1,2

Q la

0

+ PBST

-

40 -

C O.lmg STRYCH./kg

50

0

B

Va

60

20

-

100 I

msec

60

-

40

-

20

-

150 I

0

.

0 CONT. 4- O.lmg STRYCH./kg 1

2

I

I

3

4

I

1

msec

5

L

Fig. 3. Action of strychnine on the presynaptic and postsynaptic inhibition of monosynaptic reflexes and on dorsal root potentials. Cat spinal cord in situ. The presynaptic (A) and postsynaptic (B) curves are plotted as in Fig. 2 before and after the i.v. injection of 0.1 mg/kg strychnine. In C the dorsal root potentials were produced in the conventional manner from the most caudal rootlet of L6 dorsal root and evoked by the various volleys as indicated. PBST = posterior-biceps-semitendinosus; PDP = peroneal-deep peroneal; FDHL-PL = flexor digitorum-hallucis longus-plantaris; SU = suralis; SP = superficial peroneal. The muscle afferent volleys were set up by four stimuli a t 300/sec and just above maximal for Group I, while the single stimuli evoking the cutaneous afferent volleys were a t 4 x threshold strength so as to excite all of the a-group (Hunt and McTntyre, 1960). Time constant of thc amplifier 1 sec. The first row shows the control dorsal root potentials, the second row was obtained after the injection of 0.1 mg/kg strychnine and the third row after a further injection. The effect of the same injection of strychnine is shown in A, B and the second row of C. Same voltage and time scale for all records. (From Eccles et al., 1963d.)

THE P H A R M A C O L O G Y O F P R E S Y N A P T I C I N H I B I T I O N

123

Eccles et al., 1954; Fatt, 1954; Eccles, 1957; Curtis, 1959, 1963). It was evidently important to determine if strychnine also depresses presynaptic inhibition. Fig. 2 shows how strychnine acts on the depression of monosynaptic reflexes by post- and presynaptic inhibition. In Fig. 2B a monosynaptic reflex evoked by a volley in the posterior-biceps-semitendinosus nerve is inhibited by a preceding volley in the antagonistic quadriceps nerve. This postsynaptic inhibition is nearly abolished by i.v. injection of 0.1 mg/kg strychnine. In A the presynaptic inhibition exerted by three posterior-biceps-semitendinosus volleys onto a monosynaptic reflex evoked by stimulation of the gastrocnemius-soleus nerve is practically unaltered by the same strychnine injection. Fig. 3 shows in A and B a similar experiment. Here the presynaptic inhibition is even increased after the strychnine injection. In C dorsal root potentials produced by stimulation of various muscle and cutaneous nerves of the hind limb before (first row) and after two successive injections of strychnine (second and third row) are reproduced. There is an increase in the amplitude of the dorsal root potentials after

A

Strychnine

-

u

J

500 rnsec

J

sec

Fig. 4. The effect of strychnine on the ventral and dorsal root responses of an isolated toad spinal cord. The upper traces in A-D and E are dorsal root potentials recorded in the 9th dorsal root. Time constant of the amplifier 10 sec. The lower traces in A-D are the e!ectrically integrated ventral root reflexes. A is the control record. B was taken 10 min after the end of a 15-min application of 0.01 mg/100 ml strychnine. C-E were recorded 2 h after the strychnine application. A-C were evoked by stimulation of a neighbouring dorsal root, D occurred without previous stimulation. E was also recorded in the absence of stimulation and shows rhythmically generated ‘spontaneous’ dorsal root p3tentials.Voltage calibration is for dorsal root potentials only and is in B for A, B and in D for C-E. Time calibration in A is for A, B and in C for C , D. The gain of the ventral root reflex trace is twice as high in A, B than in C, D . (From Schmidt, 1963.) References p . 130/13/

124

R. F. S C H M I D T

the strychnine injection but no appreciable change in time course. The increased dorsal root potential amplitude corresponds well with the increase in presynaptic inhibitory action shown in Fig. 3A. Apparently strychnine increased the excitability of interneuroiies on the presynaptic inhibitory pathways so that a given afferent input produced a stronger presynaptic inhibitory effect. In in vitro preparations, such as the isolated toad spinal cord, the action of still heavier doses of strychnine can be observed. Such an experiment is illustrated in Fig. 4. The dorsal root potentials and the electrically integrated records of the prolonged ventral root reflexes recorded after stimulation of a neighbouring dorsal root are shown before (A) and 10 and 120 mill (B, C) after heavy poisoning of the preparation with strychnine. In B the dorsal root potentials are increased and their time course is prolonged. In C the time course is prolonged still further but the size is reduced by nearly half. At this time dorsal root potentials and ventral root reflexes occurred also in the absence of stimulation (D, E). On isolated frog spinal cords observations similar to those shown in Fig 4 (A-C) have already been described by Umrath ( 1 933), Dun and Feng (1 944) and Eccles and Malcolm (1946). What do these experiments tell us? The important fact is that presynaptic inhibition is not affected by strychnine at all in concentrations which lead already to convulsions in spinalized animals. And any changes in presynaptic inhibition which we observe with high strychnine concentrations - particularly the prolonging of the declining phase of the primary afferent depolarization - are most probably due to the removal of postsynaptic inhibition at the interneurones of the presynaptic inhibitory pathway. Some of the central inhibitions resistant to strychnine (for review see Eccles et a/., 1963d) may now be recognized as examples of presynaptic inhibition. For instance, the inhbition of stretch receptors in the hamstring muscle, which is not affected by injection of 0.1 mg/kg strychnine or even larger doses, is very likely a case of 4PBST+GS,GS,

80

DRP

BPBST

-

P I C - l o x , O.I%-, 2

' 0 CONT.

m

i

n

k

30min

+

'Oh

100

200

,"L_

SMAB

--

L-h-

L

4

PT

L

-'

A- -L --LA,

I

I

0

GS

&

C P I C - l o x . 0.1% 0 2Omin I O I ~ C3Omin

20 -0

0

0

PDP

miec

300

Fig. 5. Action of picrotoxin on presynaptic inhibition of monosynaptic reflexes and on dorsal root potentials. Cat spinal cord in silu. The presynaptic inhibitory curves in A were obtained as in Fig. 2, and show the effect of direct application of picrotoxin to the spinal cord. Open circles give initial curves, and conditions for other curves are indicated in the key to the symbols. In B there are shown the dorsal root potentials produced by the indicated afferent volleys (SMAB = semimembranosusanterior biceps; for the other abbreviations see legend of Fig. 3) before and 20 and 50 min after the application of picrotoxin, and finally in the fourth row 10 h after a brief initial period of repeated washing of the cord with warm Ringer-Locke solution. Same time and potential scale for all records. (From Eccles et al., 1963d.)

THE P H A R M A C O L O G Y OF P R E S Y N A P T I C I N H I B I T I O N

125

presynaptic inhibition (Liddell and Sherrington, 1925; Cooper and Creed, 1927). There are several other convulsants whch do not seem to affect postsynaptic inhibition and it may be asked whether some of them may act by impeding presynaptic inhibition. Picrotoxin is the most potent member of this group and was most thoroughly studied both in the cat and in the toad. Fig. 5 shows an experiment in which picrotoxin was applied topically to the cat spinal cord. The left hand diagram shows the depression of a gastrocnemius monosynaptic reflex by four volleys in the posterior-biceps-semitendinosus nerve before and 20 and 50 min after the application of picrotoxin. Clearly, the inhibition of the monosynaptic reflex is reduced after the picrotoxin application. This change must represent a genuine reduction in presynaptic inhibitory action because the size of the testing monosynaptic reflex (not shown in Fig. 5) was hardly affected, being increased 10-1 5 %. Simultaneously, as shown in Fig. 5B, the dorsal root potentials produced by different types of afferent input are depressed, indicating a smaller primary afferent depolarization. Ten hours later the dorsal root potentials have recovered their original size. Such doses of picrotoxin and even higher ones, up to 2 mgikg i.v., had little or no depressant action on postsynaptic inhibition. A possible explanation of the depressant action of picrotoxin on presynaptic inhibition would be that it exerts a powerful background excitatory action on the interneuronal presynaptic pathways, which consequently are not as fully available for activation by conditioning volleys, i.e. that there is an occlusive action. But it has been found, on the contrary, that the increase in the neurone activation produced by strychnine actually increases presynaptic inhibition, apparently by facilitating the activation of interiieurones on the presynaptic inhibitory pathway. Furthermore, picrotoxin depresses presynaptic inhibition at a dosage below that causingconvulsions. It may therefore be suggested that picrotoxin produces a genuine depression of the presynaptic inhibitory synapses. On analogy with the pharmicological blockage of cholinergic synapses on heart and skeletal muscle, in sympathetic ganglia and on Renshaw cells it may be assumed that this selective depression is attributable to a competitive occupation of the receptor sites for the presynaptic inhibitory transmitter substance. Such a mechanism has already been proposed to account for the blocking action of picrotoxin at the inhibitory synapse of the crayfish neuromuscular junction (Robbins and Van der Kloot, 1958; Grundfest et al., 1959; Grundfest and Reuben, 1961). Certainly, the sharp differentiation between the inhibitory depressant actions of strychnine and picrotoxin suggests that either different transmitter substances or quite different receptor sites or both are present at postsynaptic and presynaptic synapses. As mentioned above, the convulsant action of strychnine seems entirely due to its depression of postsynaptic inhibition. Similarly it may be suggested that the depressant action of picrotoxin on presynaptic inhibition is sufficient to account for its convulsant action. It is, however, appreciated that the depression of presynaptic inhibition by a convulsant dose of picrotoxin is much less striking than the depression of postsynaptic inhibition by a convulsant dose of strychnine. Furthermore, picrotoxin acts at all levels of the central nervous system and it has yet to be shown that presynaptic inhibition acts on other synapses than those made by primary afferent fibres. Refermres p . 13Oll31

126

R. F. S C H M I D T

Anaesthetics

The investigation of the action of various groups of anaesthetics, barbiturates, volatile and aliphatic anaesthetics and urethane, showed that all anaesthetics led either in the beginning or throughout their application to a shortening of primary afferent depolarization, usually accompanied by a n increase in amplitude. Fig. 6 illustrates this at first puzzling behaviour. The dorsal root potentials and the electrically integrated ventral root reflexes of lumbar roots of a n isolated toad spinal cord are shown. Throughout their application all ether concentrations lead to an increase and a shortening of the dorsal root potentials although the ventral root reflexes are reduced and nearly suppressed with the higher ether concentrations, as would be expected from the effect of an anaesthetic. Similar observations have been made with chloralhydrate, paraldehyde and urethane. Thus, anaesthetics seem to act much less on presynaptic inhibition than on motor reflexes of the spinal cord and this relative increase in inhibition might partly explain their action.

Fig. 6. The effcct of diethyl-ether on the dorsal root potentials (upper traces) and ventral root reflexes (integrated records, lower traces) produced and recorded as in Fig. 4. CON were the control records. The ether concentration was 50 mg/100 ml in A, 150 mg/100 ml in B and 300 mg/100ml in C and D. The numbers indicate the time in seconds after the application of ether. A, B, C were taken consecutively with 15 min intervals in normal Ringer between each application, D 1 h after C . Voltage calibration is for the dorsal root potentials only. (From Schmidt, 1963.)

The barbiturates and chloralose have an additional effect on primary afferent depolarization. This is demonstrated in Fig. 7 which is again from an isolated toad spinal cord and similar to the previous one. Instead of ether pentobarbital sodium (nembutal) has been added to the Ringer solution. After a brief initial period in which the dorsal root potential is shortened its time course is prolonged. For nembutal this has already been reported by Eccles and Malcolm (1946) for dorsal root potentials in the frog spinal cord and by Lloyd (1952) for the dorsal root potentials in the cat spinal cord. Similar prolongations have now been obtained with phenobarbital (luminal) and pentobarbitone sodium (surital) and with chloralose. In Fig. 7 there are also shown the dorsal root potentials which can be evoked by

THE PHARMACOLOGY OF PRESYNAPTIC INHIBITION

1

127

0 DRP size 0 VR-DRP size

NEMBUTALo 01%

0 DRP holf decoy 0 VR-DRP half decay

%-

- - -

- - - -

E

- -

VRR ( i n t e g r . )

l

I

I

0

10

20

I

30

L

I

40

min

50

Fig. 7. The effect of nembutal on the dorsal root responses evoked by dorsal root (DRP) or ventral root (VR-DRP) stimulation. Isolated toad spinal cord. Time constant of the amplifier 10 sec. (A). Specimen records of dorsal root potentials (upper traces) recorded in the 9th dorsal root and produced by a single afferent volley of 4 x threshold in the 8th dorsal root. The lower traces show the integrated reflex discharge recorded from the 8th ventral root. The numbers indicate the time in minutes after the start of a 0.01 % nembutal application, CON being the csntrol record. B corresponds to A but the potentials were evoked by a supramaximal stimulus to the 9th ventral root. In C , D, E various param-ters (see symbols) of the DRP, the VR-DRP and the ventral root reflex are plotted against th- duration of the nembutal application. Potential scale in A is for the dorsal root potentials only. Note the different time scales for A and B. (From Schmidt, 1963.)

stimulation of ventral roots (VR-DRP). They are always much shorter and smaller than the dorsal root potentials evoked by dorsal root stimulation and much more sensitive against all types of drugs tested. Fig. 7C shows that VR-DRPs are already completely abolished when the dorsal root potentials evoked by dorsal root stimulation are still two thirds of their control size. In cats the prolonged dorsal root potential during barbiturate anaesthesia is combined with a prolonged presynaptic inhibition (Fig. 8). With light anaesthesia (as in Fig. 8) there is also often an increase in presynaptic inhibition. This increase is always accompanied by increased dorsal root potentials. Deepening of anaesthesia leads to a decline of both the dorsal root potential and the presynaptic inhibitory action together with a further prolongation References p . 130/131

R. F. S C H M I D T

128

%I

_ _ - - - - - _ _ - _ _ _ _ _ _ I _ 4PBST

100

+

GS-GS

- - ----

80

60

.-.

+

- -

20

/ x + -/ + x -0 O

0

100

C 2 0 r n s NEMB /ka

i '

X I

I

I

200

30 0

400

msec

1

500

Fig. 8. Effect of nembutal on presynaptic inhibition of a monosynaptic reflex. The presynaptic inhibitory curves were determined as in Fig. 2, but with four posterior-biceps-semitendinosus volleys inhibiting a monosynaptic reflex evoked by two gastrocnemius volleys at 1.5 msec interval. The first gastrocnemius volley itself evoked no reflex, but merely facilitated the reflex evoked by the second volley that was maximum for Group I. Open circlesgive initial curves in decerebrate unanaesthetized cat, and conditions for other curves are indicated in the key for the symbols. (From Eccles eral., 1963d.)

of their time course. Thus the increased inhibition under light anaesthesia seems to be a consequence of the increased primary afferent depolarization. A supplementary cause would be that the monosynaptic reflex normally decreases under anaesthesia and is probably more readily inhibited. What are the possible reasons for the increase and the shortening of the primary afferent depolarization in the beginning of or throughout anaesthesia and why do barbiturates lead to a prolongation in the later stages? The most likely explanation of the shortening is that the anaesthetic curtails the after-discharge of the interneurones in the pathway responsible for the primary afferent depolarization. The increased amplitude of the primary afferent depolarization is probably caused by the removal of spontaneous background activity of the same interneurones. An overall reduction of background primary afferent depolarization would increase the efficacy of volleys in primary afferent fibres producing the inhibition because the synaptic endings of the fibres which activate the presynaptic inhibitory interneurones (cf. Eccles et al., 1962a) would be less depolarized. Furthermore, decreased background primary afferent depolarization would hyperpolarize the fibres receiving the inhibitory action. The hyperpolarization will increase the potential difference between the resting potential of the fibrz and the equilibrium potential of the depolarization thus enhancing the synaptic drive during activation. There remains the question of why the barbiturates and chloralose prolong the declining phase of the primary afferent depolarization i n the later stages of anaesthesia. In a systematic study, it has been found (Eccles et a/., 1963d) that Neinbutal application invariably decreased the rate of discharge of interneurones firing spontaneously or after orthodromic activation. So it seems likely that barbiturates and chloralose prolong the primary afferent depolarization either by the inhibition of an enzyme that destroys the

THE PHARMACOLOGY O F PRESYNAPTIC INHIBITION

129

transmitter substance or that these anaesthetics act by impeding the diffusional spread of the transmitter substance from its site of action. The present experiments do not allow to distinguish between these or any other possibilities. Cholinergic drugs A wide variety of cholinergic drugs, dihydro-P-erythroidine, gallamine thriethiodide (Flaxedil), atropine, nicotine, tetraethyl- pyrophosphate and eserine, have been injected intravenously and applied topically to the spinal cord of cats and toads. In general these drugs do not have any action whatsoever in concentrations highly active on known cholinergic synapses. There is one exception: The dorsal root potential generated in the toad spinal cord by stimulation of ventral roots can be inhibited by dihydro-B-erythroidine and to a smaller extent by curare, an observation first made by Kiraly and Phillis (1961). This indicates that the pathway of this system contains a cholinergic synapse. Possibly this synapse corresponds to the synapse made by the motor axon collaterals at the Renshaw cell in the cat spinal cord; conceivably it could also be at some later synapse on the polysynaptic pathway.

Amino acids When applied topically acidic amino acids (such as glutamic acid) as well as neutral amino acids (such as GABA and y-amino-propanesulfonic acid) both depressed the dorsal root potential in the cat and toad spinal cord although they have antagonistic effects on postsynaptic spinal elements (Curtis et al., 1960, 1961b; Curtis and Watkins, 1960, 1961; Curtis, 1961, 1963). Attempts have therefore been made to determine the depolarizing potency of these drugs on primary afferent fibres by testing the excitability increase after the amino acid application. Glutamic acid appeared to be the most potent substance and on analogy with results on motoneurones (Curtis et al., 1961b) it may be assumed that this amino acid directly depolarizes primary afferent fibres. Rather large concentrations are required to obtain appreciable excitability increases and it seems therefore unlikely that any of the tested amino acids is closely related to the transmitter substance. C O N C L U SI ON

The research into the pharmacology of presynaptic inhibition has just begun and the results obtained so far are still of a very preliminary nature. At least they indicate that presynaptic inhibition is not only in its physiological (see Eccles, 1963, this Volume, p. 65) but also in its pharmacological properties quite distinct from any of the postsynaptic inhibitons we have known so far. It is hoped that because of these differences pharmacological tests are going to provide a powerful tool for analyzing the presynaptic or postsynaptic nature of inhibitory processes.

References p. I30/131

130

R. F. S C H M I D T REFERENCES

B. H. C., (1938); The interpretation of potential changes in the BARRON, D. H., AND MATTHEWS, spinal cord. J. Physiol. (Lond.), 92, 276-321. BRADLEY, K., EASTON, D. M., AND ECCLES,J. C., (1953); An investigation of primary or direct inhibition. J . Physiol. (Lond.), 122, 474-488. COOPER, S., AND CREED, R. S., (1927); More reflex effects of active muscular contraction. J . Physrol. (Lond.), 64, 199-214. CURTIS, D. R., (1959); Pharmacological investigations upon inhibition of spinal motoneurones. J. Physiol. (Lond.), 145, 175-192. CURTIS,D. R., (1961); The identification of mammalian inhibitory transmitter. Nervous Inhibition. E. Florey, Editor. Oxford, Pergamon Press (p. 342-349). CURTIS,D. R., (1963); The pharmacology of central and peripheral inhibition. Pharmacol. Rev., 15, 333-364. CURTIS,D. R., AND ECCLES,R. M., (1958); The excitation of Renshaw cells by pharmacological agents applied electrophoretically. J . Physiol. (Lond.), 141, 435-445. CURTIS, D. R., PHILLIS,J. W., AND WATKINS, J. C., (1960); The chemical excitation of spinal neurones by certain acidic amino acids. J. Physiol. (Lond.), 150, 656-682. CURTIS,D. R., PHILLIS, J. W., AND WATKINS, J. C., (1961a); Cholinergic and non-cholinergic transmission in the spinal cord of the toad. J. Physiol. (Lond.), 158,296-323. CURTIS,D. R., PHILLIS, J. W., AND WATKINS, J. C., (1961b); Actions of amino acids on the isolated hemisected spinal cord. Brit. J. Pharmacol., 16, 262-283. CURTIS,D. R., AND WATKINS, J. C., (1960); The excitation and depression of spinal neurones by structurally related amino acids. J. Neurochem., 6, 117-141. CURTIS,D. R., AND WATKINS, J. C., (1961); Analogues of glutamic and y-amino-n-butyric acids having potent actions on mammalian neurones. Nature (Lond.), 191, 1010-101 1. DUN,F. T., AND FENG,T. P., (1944); A note on the two components of the dorsal root potential. J. Neurophysiol., 7 , 327-329. ECCLES, J. C., (1957); The Physiology of Nerve Cells. Baltimore, Johns Hopkins Press. ECCLES,J. C., (I 962); Spinal neurones : Synaptic connexions in relation to chemical transmitters and pharmacological responses. Proc. First Int. Pharmacol. Meeting, Vol. 8. Oxford, London, New York, Paris, Pergamon Press (p. 157-182). J. C., (1964) ; Presynaptic inhibition in the spinal cord. Physiology of spinal Neurons. ProECCLES, gress in Brain Research, Vol. 12. J. C. Eccles and J. P. Schadk, Editors. Amsterdam, Elsevier (p. 65). ECCLES,J. C., ECCLES, R. M., AND FATT,P., (1956); Pharmacological investigations on a central synapse operated by acetylcholine. J . Physiol. (Lond.), 131, 154-169. ECCLES,J. C., ECCLES,R. M., AND MAGNI,F., (1961); Central inhibitory action attributable to presynaptic depolarization produced by muscle afferent volleys. J . Physiol. (Lond.), 159, 147-166. ECCLES, J. C., FATT,P., AND KOKETSU, K., (1954); Cholinergic and inhibitory synapses in a pathway from motor-axon collaterals to motoneurones. J. Physiol. (Lond.), 216, 524-562. ECCLES,J. C., KOSTYUK, P. G., AND SCHMIDT, R. F., (1962a); Central pathways responsible for depolarization of primary afferent fibres. J. Physiol. (Lond.), 161, 237-257. ECCLES, J. C., MAGNI,F., AND WILLIS,W. D., (1962b); Depolarization of central terminals of Group I afferent fibres from muscle. J . Physiol. (Lond.), 160, 62-93. ECCLES, J. C., AND MALCOLM, J. L., (1946); Dorsal root potentials of the spinal cord. J . Neurophysiol., 9, 139-160. ECCLES, J. C., SCHMIDT, R. F., AND WILLIS,W. D., (1962~);Presynaptic inhibition of the spinal monosynaptic reflex pathway. J . Physiol. (Lond.), 161, 282-297. ECCLES, J. C., SCHMIDT, R. F., AND WILLIS,W. D., (1963a); Depolarization of central terminals of Group Ib afferent fibers from muscle. J. Neurophysiol., 26, 1-27. ECCLES, J. C., SCHMIDT, R. F., AND WILLIS,W. D., (1963b); The location and the mode of action of the presynaptic inhibitory pathway on to Group I afferent fibers from muscle. J . Neurophysrol., 26, 506-522. ECCLES, J. C., SCHMIDT, R. F., AND WILLIS,W. D., (1963~);Depolarization of the central terminals of cutaneous afferent fibers. J. Neurophysiol., 26, 646-661. ECCLES, J. C., SCHMIDT, R. F., AND WILLIS,.W D., (1963d); Pharmacological studies on presynaptic inhibition. J. Physiol. (Lond.), 168, 500-530. FATT,P., (1 954); Biophysics of junctional transmission. Physiol. Rev., 34, 674-710. FLOREY, E., (1961); Comparative physiology: Transmittersubstances. Ann. Rev. Physiol., 23,501-528.

T H E P H A R M A C O L O G Y OF P R E S Y N A P T I C I N H I B I T I O N

131

FRANK, K., AND FUORTES, M. G. F., (1957); Presynaptic and postsynaptic inhibition of monosynaptic reflexes. Fed. Proc., 16, 39-40. GRUNDFEST, H., AND REUBEN, J. P., (1961); Neuromuscular synaptic activity in lobster. Nervous Znhibition. E.Florey, Editor. Oxford, Pergamon Press (p. 92-104). GRUNDFEST, H., REUBEN, J. P., AND RICKLES, N. H., (1959); The electrophysiology and pharmacology of lobster neuromuscular synapse. J . gen. Physiol., 42, 1301-1324. HUNT,C. C., AND MCINTYRE, A. K., (1960); An analysis of fibre diameter and receptor characteristics of myelinated cutaneous afferent fibres in cat. J . Physiol. (Lond.), 153, 99-112. J. K., AND PHILLIS,J. W., (1961); Action of some drugs on the dorsal root potentials of the KIRALY, isolated toad spinal cord. Brit. J . Pharmacol., 17, 224-231. C. S., (1925); Further observations on myostatic reflexes. LIDDELL,E. G. T., AND SHERRINGTON, Proc. roy. SOC.B, 97, 267-283. LLOYD, D. P. C., (1952); Electrotonus in dorsal nerve roots. Cold Spr. Harb. Syinp. quant. Biol., 17, 203-219. LONGO,V. G., MARTIN,W. R., AND UNNA,K. R., (1960); A pharmacological study of the Renshaw cell. J . Pharmacol. exp. Ther., 129, 61-68. H., (1961); Inhibitory transmitters - a review. Nervous Inhibition. E. Florey, Editor. MCLENNAN, Oxford, Pergamon Press (p. 350-368). PATON,W. D. M., (1958); Central and synaptic transmission in the nervous system (pharmacological aspects). Ann. Rev. Physiol., 20, 431-470. PERRY,W. L. M., (1956); Central and synaptic transmission (pharmacological aspects). Ann. Rev. Physiol., 18, 279-308. ROBBINS, J., A N D VANDER KLOOT, W. G . , (1958); The effect of picrotoxin on peripheral inhibition in the crayfish. J . Physiol. (Lond.), 143, 541-552. SCHMIDT, R. F., (1963); Pharmacological studies on the primary afferent depolarization of the toad spinal cord. Pjiig. Arch. ges. Physiol., 277, 325-346. WALL,P. D., (1958); Excitability changes in afferent fibre terminations and their relation to slow potentials. J . Physiol. (Lond.), 142, 1-21. K., (1933); Der Erregungsvorgang in den Motoneuronen von Rana esculenta. Pjiig. Arch. UMRATH, ges. Physiol., 233, 357-370.

DISCUSSION

PHILLIPS: Liddell and Sherrington (Proc. voy. SOC.B., 97 (1925) 267-283) noticed that strychnine did not interfere with inhibition of tonic stretch by pulling on antagonist muscle. An early example of presynaptic inhibition. SCHMIDT: Yes, indeed. They 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 strychnine, or even larger doses. In addition, the time course of this inhibitory action was precisely the time course that would be expected for presynaptic inhibition. SZENTAGOTHAI: Your findings on the strychnine effects on the two types of inhibition fit well with the strychnine effects experienced on deplanted groups of nerve cells, made according to the Paul Weiss’ technique. Strychnine effects are experienced on deplants containing only around 10 nerve cells in loose completely haphazard distribution. It would be extremely difficult to imagine that such a complicated arrangement as required by the presynaptic inhibitions, i.e. one synaptic knob on top of the other, would be preserved in such a preparation. The postsynaptic inhibitory mechanisms with their postulated specific chemical transmitter actions could be expected to be

132

DISCUSSION

much less affected by such circumstances. As shown by the strong strychnine effect in such preparations this appears to hold true. LUNDBERG: In recent experiments we have found that L-DOPA (the precursor of the catecholamines) inhibits the primary afferent depolarization evoked from the FRA but not the one from Group I muscle afferents (AndCn, Lundberg, Rosengrin and Vycklickf, Experientia (Basel), 19 (1963) 654). Our experiments were prompted by the histochemical work of Carlsson, Falck, Fuxe and Hillarp (Acta physiol. scand., in the press). With a fluorescence method they have demonstrated that there are noradrenergic nerve terminals in the spinal cord, which disappear after chronic transection of the cord. In the acute spinal animal DOPA gives a larger increase of the flexor reflex. Although DOPA depresses also the paths from the FRA to motoneurons there is a more pronounced effect on the paths to primary afferents. The increased reflex may, therefore, be due to abolishment of presynaptic inhibition. It is possible that DOPA inhibits reflex arcs by inducing synthesis and overflow of noradrenaline from adrenergic nerve terminals with inhibitory action on reflex paths from the FRA. GELFAN:1 would like to come back to the question of strychnine. How do you distinguish between a subconvulsive dose, presuming you removed postsynaptic inhibition, and convulsive dose of strychnine? SCHMIDT:In our laboratory the animals are not curarized at all, they are just spinalized. With large doses of strychnine they will convulse without any obvious stimulus or tapping on their feet will lead to convulsions. In subconvulsive doses they will not do that or just show very slight movement after peripheral noxious stimulation. There is, of course, no qualitative difference in the action of subconvulsive and convulsive doses of strychnine. It is just a matter of the strychnine concentration. GELFAN: What is happening to the discharge pattern of the motoneurons? SCHMIDT:Prolonged firing of rather high frequency will occur because all types of postsynaptic inhibition have disappeared. You cannot record IPSPs anymore. GELFAN: This than does mean that you have removed all inhibition and that the neurons discharge spontaneously? SCHMIDT: Due to the lack of inhibition the central nervous system starts to oscillate by any trivial afferent input which the cat will have all the time. This is not really a ‘spontaneous’ process. 1 was certainly very interested in those remarkable findings of Dr. LundECCLES: berg where it seems at first sight that he distinguishes from a pharmacological point of view between two types of primary afferent depolarization. That prompts me to ask the question: where does he think it is acting, and could it not be acting upon

THE PHARMACOLOGY OF PRESYNAPTIC INHIBITION

133

interneurons rather than upon the terminals on the presynaptic fibers? All our other evidence indicates that these terminals work by the same transmitter. But of course there are interneurons on the pathway which could give you discriminative pharmacology and that might be an explanation of DOPA. And I ask also: what happens to the dorsal root potentials that you can generate from the cortex? Are they also affected by DOPA selectively? And, finally: have you looked at interneurons and seen what it does to the discharges? LUNDBERC: I have to apologize that I did not make myself very clear. The possibility that I outlined was not that it was working on the reflex arc. The experiments we have done so far concerned transected spinal cords. The degenerating descending fibers are seen presynaptic, postsynaptic, everywhere. We did find that L-DOPA inhibits the primary afferent depolarization evoked from the FRA. We have only this indication and we have not worked with it on the dorsal root potentials evoked from other sources than from primary afferents and we have not yet recorded from interneurons. GRANIT:Dr. Schmidt, can you tell me now why the retinal action potential gets so big under nembutal? SCHMIDT:No, I am afraid I can’t. But I may point out that Kidd (J. Anat. (Lond.), 96 (1962) 179-187) has recently described axo-axonal synapses in the retina of the cat and the pigeon. It might be speculated that the pharmacology of these synapses is similar to that of the presynaptic synapses in the spinal cord. If this is so, their prolonged synaptic actions during nembutal anaesthesia may well be responsible for the increase of the retinal action potential which you observe. SEARS:Have you looked for any differentiation of ether and nembutal on the type of interneuron you think that may be mediated by presynaptic inhibition? SCHMIDT:No, we have not. We have looked for these interneurons during nembutal anaesthesia only. We were very much worried about the results of Dr. Wall, indicating that you can have dorsal root potential without any interneurons firing at all. We did not know that meanwhile Dr. Wall is suggesting that interneurons are as a matter of fact making the dorsal root potentials and maintaining it throughout. I actually would very much like to hear Dr. Wall’s comment on this point of shortening of interneuronal discharge together with a prolonged decline of the dorsal root potentials under nembutal and other barbiturates. WALL:My comment of course is that you are looking at the wrong interneurons. Everything you do to these interneurons you are looking at goes in the wrong direction. Furthermore there is another point that I brought up yesterday, that one can now actively turn off a steady DRP. This suggests to me that it is being continually pro-

134

DISCUSSION

duced by an active process, and the active process is located by a source-sink mapping up i n the dorsal regions with very small cells that one does not pick up with this type of micro-electrode.

ECCLES: There is a very good correlation indeed between the interneuronal discharges we have observed and the dorsal root potential during progressively deeper anaesthesia. The DRP declines in size with increase in anaesthesia, which corresponds well with the decline in number and frequency of discharges of interneurons. The only difference is in the duration of the declining phase ofthe DRP, which is prolonged, whereas the interneuronal discharge occurs only during the brief rising phase of the DRP. Our inference is that this slowed decline is most probably due to the inhibition of an enzyme that inactivates the transmitter responsible for presynaptic inhibition. This postulate is analogous to the increased duration of Renshaw cell activation by a single antidromic volley when the cholinesterase is inactivated, and which we attribute to the longer duration of ACh action. It is important to realize that the principal results with our pharmacological experiments were on the presynaptic inhibition of Group Ia fibers, which is of course exerted on monosynaptic terminals or motoneurons in the ventral horn and has nothing to do with the substantia gelatinosa. But even with the cutaneous presynaptic inhibition the main effect is at least exerted on the synapses of cutaneous fibers on the large cutaneous cells in lamina IV and not with the substantia gelatinosa. WILLIS:I would like to ask Dr. Wall if his source-sink maps can distinguish between prolonged activity in interneurons and a prolonged action of a transmitter substance depolarizing afferent terminals. WALL:I think so. JUKES:It seems to me that if you go on giving bigger and bigger doses of nembutal you get a lowering of the blood pressure and the effect might be just simply due to a decreasing washing out of the transmitter.

ECCLES : The suggestion that anaesthetics might prolong the D R P because they lower the blood pressure and that the consequent poor circulation results in slower removal of the transmitter is of course answered by the similar effects of anaesthetics in prolonging the DRPs of the isolated spinal cord of the frog and toad.

135

Ascending Spinal Hindlimb Pathways in the Cat A. L U N D B E R G

Department of Physiolog.v, University of Ccteborg, CoteborR (Sweden)

Modern electrophysiological methods have vastly increased our knowledge concerning ascending spinal pathways, a field in which the classical textbook knowledge has been based entirely on anatomical and clinical investigations. As in the analysis of spinal reflex function the advance has depended not only on new recording techniques but also on better knowledge of receptors and their afferents. In this review, start will be made with the discharges that can be recorded in the dorsal part of the lateral funicle. The pioneering efforts of Grundfest and Campbell (1942) and Lloyd and McIntyre (1950) were concerned with these discharges, and they were also the object of our first study in this field. In the first section these results are presented in order to introduce the methods used throughout these investigations. Later sections deal with the functional organization of the different pathways and the last section is concerned with certain aspects on the supraspinal control. (1)

ASCENDING DISCHARGES I N THE D O R S A L P A R T OF T H E L A T E R A L F U N I C L E

Recording of mass discharge: In a study of the properties of nerve fibres in the spinal cord Rudin and Eisenman (195 1) recorded from dissected fascicles. In collaboration with Laporte this technique was modified in order to record the activity in ascending spinal pathways in vivo (Laporte and Lundberg, 1955; Laporte et al., 1956a; McIntyre and Mark, 1960). In Fig. 1 records A-D were obtained from a dissected spinal cord fascicle comprising the hatched and black area in the drawing below the records. In record A the ipsilateral superficial peroneal nerve (cutaneous) was stimulated and in B the ipsilateral quadriceps nerve. It was proved that the discharges in the dissected dorsal part of the lateral funicle (DLF) were postsynaptic and monosynaptically transmitted. Recordings at different spinal cord levels revealed a conduction velocity of about 100 m/sec for the fastest fibres activated from muscle or from skin (cJ also Lloyd and McIntyre, 1950). Records C and D were taken after a lesion (black area in drawing) in the known regions of the dorsal spinocerebellar tract (DSCT). The monosynaptic discharge evoked from muscle afferents has almost disappeared, but a significant part of the discharge from cutaneous afferent remains, which is of interest in relation to the fact that the lesion did spare the most dorsomedial part of the lateral funicle (cf: section 3). Another type of recording was used in E and F, Fig. 1. References p. 160-163

136

A. L U N D B E R G

fiBL J i > SP

Q

rC

G

D

H

-w@+-c (

20 msec

15 m s e c

-

10msec

Fig. 1 . Ipsilateral mass discharges in the dissected dorsal part of the lateral funicle (DLF) and from the spinal half. Recording was made in the mid-thoracic region, in A-D from a dissected fascicle of the extent shown by the hatched and black area in the left spinal cord section and in E-H from the spinal half (except dorsal column). The superficial peroneal nerve (SP) was stimulated in A, C, F and H and the quadriceps nerve (Q) in B, D, E and G. Lower records were taken after the spinal cord lesion shown in black in corresponding sections below the records. (A-D, E and G from Laporte et al., 1956a; F and H from Oscarsson, 1958).

Ia

-

10

0

I b

10

Fig. 2. Contribution from Ia and Ib afferents to the mass discharge in the dissected DLF. Recording was made as in Fig. 1 (lower traces with negativity downwards) and simultaneously from a dorsal root filament in L5 (upper traces) and triphasically from the L5 dorsal root entry zone. The quadriceps nerve was stimulated at increasing strength. In the curve (from another experiment) 100% on the ordinate is the discharge evoked in the DLF by a maximal group I volley (Lundberg and Oscarsson, 1956).

137

ASCENDING SPINAL PATHWAYS

The ipsilateral dissected cord, except the dorsal column, was dissected free and placed on electrodes. Record F was taken after the dorsal lesion shown in the drawing below the records ; the early monosynaptic component has disappeared but part of the late mass discharge remains. In Fig. 2 the separation of the group I volley in Ia and Ib components (Bradley and Eccles, 1953; Eccles et a/., 1957; Laporte and Bessou, 1957) 10 msec

A

B

I

Fig. 3. Late mass discharges evoked from high threshold muscle afferents. Recording as in Fig. 1 from the dissected DLF (lower traces with negativity downwards) and from a dorsal root filament in L.5 (upper traces). The quadriceps nerve was stimulated. Group I and group I1 afferents were stimulated in A, the additional late discharge in B was evoked from group I11 afferents (Laporte et al., 1956a).

B

A

I m sec

C

D

E

I

Fig. 4. Mass discharges evoked from cutaneous afferents. Recording as in Fig. 1 from the dissected DLF (upper traces) and from the saphenous nerve (lower traces). The saphenous nerve was stimulated at increasing strength, that in E being maximal for 6-fibres. Observe that upper and lower traces were recorded at different speeds and that a slow speed was used in F-H (Laporte et al., 1956a).

was used to investigate the effect from muscle spindle and Golgi tendon organ afferents. The records and the curve clearly show that both Ia and Ib afferents contribute. Fig. 3, at slower sweep speed shows that there is a late mass discharge when high threshold muscle afferents are stimulated. The late effect in A is evoked mainly from group I1 (cf. also Fig. 5 , E and F) and the additional effect in B from group 111 muscle afferents. With respect to the discharge evoked from cutaneous afferents Fig. 4 shows that the early monosynaptic discharge is evoked from very low threshold cutaneous afferents (A) and that the higher threshold afferents in the &range contribute a late component. The effects from C fibres have never been investigated. References p . 160-163

138

A. L U N D B E R G

Unit recording: The mass discharges described above can be accounted for by activity in large axons (Laporte et al., 1956b). Many neurones can be activated from low threshold group la fibres and in many of these fibres an additional (probably monosynaptic) spike is evoked from group I1 afferents (Fig. 5). Other neurones are activated by group Ib afferents as is illustrated in Fig. 6, in which triphasic recording from the G A h -

H

B

1 .

D

L

50msec

1

2 msec Fig. 5 . Neurones activated from muscle spindles. In A-F intra-axonal recording from the DLF (right vertical trace), from the dissected DLF (left vertical trace) and from a dorsal root filamcnt in L5 (horizontal trace). Records E-F were obtained with increasing strength of stimulation of the quadriceps nerve. Observe that the group I spike displays separation in l a and Ib and that the first spike appears with a very low threshold Ia volley, but the second spike in the axon only whcn group I1 fibres are stimulated. Records G I were obtained in another experiment in which the ner\es were in intact connection with the muscles. The axon in the D L F was monosynaptically activated from low threshold group I afferents in the G-S nerve. The neuronc discharged on stretch of the muscle ( G ) and paused during isometric contraction (H and I). Record K shows the discharge in another neuroce on stretch of extensor digitorum longus for the duration shown by the black line. (A-F from Laporte et al., 1956b; G-I from Lundberg and Oscarsson, 1956; and K from Holniqvist et nl., 1956).

dorsal root entry zone and recording from a dorsal root filament was used to differentiate between group Ib and I1 afferents. Experiments with adequate stimulation revealed that, correspondingly, the discharge in some units pauses (Fig. 5) and in other units (Fig. 6) accelerates during contraction as would be expected from units activated from muscle spindle and a Golgi tendon organ afferents respectively (Laporte

ASCENDING SPINAL PATHWAYS

139

and Lundberg, 1956; Lundberg and Oscarsson, 1956; McIntyre and Mark, 1960). A third type of neurone does not receive excitatory action from group I muscle afferents but a train of impulses is evoked from high threshold muscle afferents (Fig. 7). In all likelihood the action is polysynaptic but this is difficult to establish with certainty because of the slow conduction in the primary afferents. There is a

E

F

G

H

L(

100 msec

Fig. 6. Neurone activated from tendon organ afferents. In each record simultaneous recording from above, downwards from L5 dorsal root filament, triphasically from dorsal root entry zone, intraaxonally from fibre in the D L F and from the dissected DLF. Stimulation of the quadriceps nerve. Observe that the axonal discharge in C is evoked at a strength below threshold for group I1 afferents. E-H were obtained from a neuron that discharged during contraction of soleus (Lundberg and Oscarsson, 1956).

remarkable degree of convergence from ipsilateral nerves on to these neurones. Not only are they activated from antagonist muscle but from practically all muscles in the hindlimb and also from high threshold joint afferents and very effectively from cutaneous afferents (Laporte et al., 1956b; Holmqvist et al., 1956, 1960a, b ; Oscarsson, 1958). I n many of these units a single cutaneous volley evokes a train of 10-20 impulses of which the first impulse often is monosynaptically evoked. Later discharges may be polysynaptic. On adequate stimulation of skin and muscles (Fig. 7, F) there may be a long-lasting discharge after cessation of stimulation. These afferents give rise to the flexor reflex and have been denoted flexor reflex afferents (FRA). As will be described in the following sections the FRA influence many ascending spinal pathways. A fourth type of unit was monosynaptically activated by low threshold cutaneous afferents but did not receive excitation from high threshold muscle afferents. Adequately some of these units were activated on light pressure of the pad and others by tactile stimuli from very small receptive field. In section 3 it will be described that units of the latter type do not belong to the dorsal spinocerebellar tract. The exReferences p. 160-163

140

A. L U N D B E R G

2 msec Fig. 7. As in Fig. 5 but recording from a neurone that is activated from high threshold muscle afferents and not from group I muscle afferents. The quadriceps nerve was stimulated a t increasing strength in A-E. Volleys in group I1 afferents evoke the impulses in C and D, the additional impulses in I are evoked from group 111 afferents. Record F was obtained in another experiment from a neurone of this type. There is a resting discharge of l/sec lower part of the record, and a load of 500 g was attached to the quadriceps tendon for the duration of the black line. There is a discharge that for a long time outlasts the stretch. Such a discharge was evoked from all six muscles tested. (Laporte et al., 1956b; and F from Holmqvist et al., 1956).

periments described above served as a basis for the analysis of the dorsal spinocerebellar tract (DSCT) and the spinocervical tract as will be described in the following sections.

(2)

THE DORSAL SPINOCEREBELLAR TRACT

From recording of evoked cerebellar potentials, it could be concluded that group 1 activated neurones described in the last section belong to the DSCT (cf. Grundfest and Campbell, 1942; Laporte et a/., 1956a). A close analysis did reveal that also on stimulation of cutaneous afferents an evoked potential can be recorded from the anterior cerebellar cortex and that this potential can be ascribed to monosynaptic activation of DSCT neurones (Lundberg and Oscarsson, 1960, their Fig. 5 and detailed discussion p. 365).

141

A S C E N D I N G S P I N A L P A T H WAYS

For a further analysis DSCT units were identified by antidromic activation from the anterior cerebellar cortex as shown in Fig. 8 (Lundberg and Oscarsson, 1960). As shown in the diagram intra-axonal recording was made in the thoracic region and, i n addition to cerebellar stimulation, the lateral funicle could be stimulated in the mid-thoracic region and in L5. The unit in Fig. 8 was monosynaptically activated A

2"

DSCT

D

L5

i

2 msec

Fig. 8. Identification of DSCT axon by antidromic stimulation from the cerebellar cortex. Microelectrode recording from DSCT axon (upper traces) and in the lower traces of A, D and E from the contralateral spinal half (except the dorsal column). The discharge in A was evoked from the left gastrocnemius-soleus nerve ( G ) . The left intermediate cortex of the anterior cerebellar lobe was stimulated in B and the left lateral funicle at the indicated levels in C-E. The diagram t o the right gives termination of 57 DSCT axons as indicated by the low threshold foci for antidromic activation of individual axons. Vertical lines denote border of the intemediate regions, horizontal lines sulci (Lundberg and Oscarsson, 1960).

from group I muscle afferents (A) and could be antidromically activated on stimulation of the lateral funicle in Th8 (C) and from the intermediate region of the ipsilateral anterior cerebellar cortex at low threshold from a very circumscribed field. By contrast to the VSCT (cf. section 4) the DSCT axons could at threshold stimulation only be activated from a very small cortical area indicating that the individual DSCT fibres have only a small terminal area. The diagram to the right in Fig. 8 shows the terminal area for DSCT axons. The main termination is in the ipsilateral intermediate rzgion, though there is probably also a termination in the most lateral part of vermis. The lateral distribution of the DSCT has not been decided with anatomical methods in the cat. With respect to the longitudinal distribution it has been shown with anatomical methods (Grant, 1962) that the anterior terminal area of the DSCT comprises Larsell's lobules I-IV. The terminal area in Fig. 8 is the posterior part of this area. The rostral part of the anterior cerebellum in opposition to the lower corpora quadrigemina was not exposed for stimulation. A number of fibres were activated at higher strengths from the anterior margin of the exposed cortex and probably terminated in the more rostral not exposed part of the anterior cerebellum. It must be noted that of the group I activated neurones only 76 % could be antidromically stimulated from the anterior cerebellar cortex. The most likely possibility is that the remaining axons terminated References p . 160-163

142

A. L U N D B E R G

in the rostra1 buried area discussed above; another alternative would be the posterior lobe where a DSCT termination occurs in pyramis and its adjoining folia (Beck, 1927; Brodal and Jansen, 1941 ;Anderson, 1943; Grant, 1962). There is, however, the possibility that the axons of some group I activated neurones do not reach the cerebellar cortex; a spinovestibular connection should, in particular, be looked for (Lorente de N6, 1924; Pompeiano and Brodal, 1957). In the following discussion of the functional organization it will be assumed that all group I activated axons in the DLF belong to the DSCT. There was in Fig. 8 no activation by a strong stimulus in LV (D) but when the electrode was moved proximally to upper LV this axon could be stimulated. Of more than 200 identified DSCT axons, none could be stimulated from L5. This is in agreement with the fact that Clarke’s column in cat nerves extends beyond the fourth lumbar segment (Rexed, 1954). It is well established with anatomical methods that the DSCT originates from Clarke’s column (Flechsig, 1876) and experiments with intracellular recording from Clarke’s column cells have given confirmatory evidence (Curtis et al., 1958; Eccles, Oscarsson et al., 1961). The DSCT axons are uncrossed i n the spinal cord (a few exceptions have been found) and their mean velocity is 78 m/sec, ranging from 110 to 40 m/sec. The following main subgroups of DSCT neurones were identified by antidromic stimulation from the cerebellar cortex : (I) Neurones, monosynaptically activated by Ia(and11) muscle afferents (cJ Fig. 5 ) ; (2) Neurones, monosynaptically activated by Ib muscle afferents (cf. Fig. 6); (3) Neurones, monosynaptically activated from the pad by light pressure ; (4) Neurones, monosynaptically activated exclusively from skin. These units are adequately effectively activated by tactile stimuli, from a relatively restricted skin field, but additional activation is given by pressure and pinching from a larger area; (5) Neurones, activated from skin as (4) in addition from high threshold afferents of many muscles (cf. Fig. 7). It is an important principle that an ascending spinal pathway may have subdivisions with widely different function. In considering the functional significance of the message forwarded by these subgroups of DSCT it is pertinent to start with the proprioceptive carried by group I activated neurones. Both with the neurones activated by spindle afferents and those activated by tendon organ afferents there is a restricted receptive field, often only one muscle but sometimes a few (Laporte et a/., 1956b; Holmqvist et al., 1956; Curtis et al., 1958; Lundberg and Winsbury, 1960; Eccles, Oscarsson et al., 1961). The Ia activated DSCT neurones is the only pathway known that gives la information to higher centres from the hindlimbs and the Ib channel is exclusive in informing higher centres of the tension in single muscles. The linkage from group I afferents is often very strong; some neurones can follow frequencies of afferent stimulation of more than 500/sec (Holmqvist et al., 1956). On repetitive stimulation as well as with adequate stimulation the neurones usually respond effectively only from one muscle (Fig. 9). Attention has been given to the problem whether Ia and Ib afferents converge onto the same DSCT neurones. In experiments with adequate stimulation there is little or no indication of such convergence (Lundberg and Oscarsson, 1956; Lundberg and

143

ASCENDING SPINAL PATHWAYS

Winsbury, 1960). With electrical stimulation and recording from axons it was noted that in many neurones, excited by Ia, a second postsynaptic spike appeared with a Ib volley generated in the refractory period of the preceding la volley. This indicated convergence of Ia and Ib, and in recent experiments with intracellular recording

U

10 msec

10 rn sec c (

5 msec c (

"1111

* Fig. 9. Group I convergence on DSCT neurone. Recording from fibre in the DLF and in A, B, E, F also from the dorsal root entry zone in L7 and from the dissected DLF (middle traces). The nerve to gastrocnemius-soleus (G-S) was stimulated in A and B, the nerve to quadriceps in E and F. In C and D the G-S nerve was stimulated at different frequencies. The quadriceps nerve was stimulated in G and the plantaris nerve in H. The test with repetitive stimulation has shown that G-S is the main receptive field of this neurone (Holrnqvist et al., 1956).

---

12 m V

1 msec

-LLLLL"-

F

1.08

A G

H

d

1.42

P

f

1.58

i

9 -

Fig. 10. Convergence of group Ia, Ib and I1 afferent impulses on DSCT cell. The upper traces are intracellular from a DSCT cell and the lower are from the dorsal root entry zone. The nerves to posterior biceps-semitendinosus (PBSt) were stimulated in A-E and the nerves to anterior bicepssemimembranosus (AbSm) were stimulated in F-J. For both nerves the group I volley displayed the separation in Ia and Ib components. Observe that the effect from PBST is evoked from Ia and that there is additional EPSP from group I1 (record E, arrow). The effect from AbSm, on the other hand is evoked mainly from Ib afferents (H-J). Stimulus strengths are indicated in multiples of threshold strengths (Eccles, Oscarsson et a[., 1961). References p . 160-163

144

A. L U N D B E R G

(Eccles, Oscarsson et al., 1961) have given decisive evidence that this may occur (Fig. lo). In addition it was found that the group I1 connection is not exclusively with the neurones receiving la excitatory action but that there are also neurones receiving excitatory action only from Ib and group I1 afferents. However, the majority of the cells received effect either from l a or from Ib afferents. As regards forwarding of information to the cerebellum, it seems possible that the convergence of Ib on the la activated and of group I1 afferents on the Ib activated DSCT neurones are aberrant connections without much functional significance. There is convergence of many primary afferents on DSCT neurones and reason to assume that effective synaptic activation during adequate stimulation requires considerable summation. When considering the proprioceptive information in DSCT, attention must also be given to inhibition of transmission from group I muscle afferents. This was first described by Grundfest and Campbell (1942) and has been repeatedly confirmed with axonal recording. Inhibition by group 1 volleys from antagonists and from other muscles in the hindlimb and also from muscles providing excitation has been reported (Laporte eta/., 1956b; Laporte and Lundberg, 1956; Lundberg and Oscarsson, 1956; Holmqvist et a/., 1956). With intracellular recording disynaptic IPSPs have been found (Curtis et al., 1958; Eccles, Oscarsson et al., 1961). There is also evidence of polysynaptic IPSPs from group I muscle afferents (Eccles, Oscarsson et al., 1961). Eccles, Oscarsson et al. (1961) found that many of the DSCT cells did not receive IPSPs and suggest that the inhibition of discharges discussed above, is due to presynaptic inhibition of transmission from group I muscle afferents (cf. Eccles, Oscarsson et a/., 1961; Eccles et al., 196313). This presynaptic inhibition differs from that observed of group l a action on motoneurones in that the Ia terminals to DSCT receive primary afferent depolarization not only from group 1 afferents of flexor muscles but also from extensors. On the other hand Ib afferents to the DSCT receive their main primary afferent depolarization from extensors, whereas in the segmental connection flexors and extensors contribute equally to the primary afferent depolarization of Ib terminals (Eccles et a/., 1963a). Another difference is that volleys in cutaneous afferents probably also evoke presynaptic inhibition of transmission to DSCT (Eccles et a f . , 1963b, their Fig. 3). A further study of the inhibition to DSCT may provide a clue to a possible integration in the Clarke’s column relay. With respect to the inhibitory connection (particularly postsynaptic) it must be emphasized that the DSCT neurones have a resting activity (about lO/sec) and that this resting activity is not depending on primary afferent activity (Holmqvist et a/., 1956). There is the possibility that the DSCT can relay inhibition against a background of resting discharge. However, in those group I activated DSCT which are very effectively activated from spindle or tendon organ afferents of one muscle this is such a dominating event that it seems doubtful if inhibition can serve any other function than eliminating stray excitation from other muscles. However, in some neurones the adequate activation is not very effective, and in those inhibition may play a more important role. There is the possibility that some DSCT neurones forward a highly integrated message, at present poorly understood. There is no evidence for a parallelism with the segmental actions from group I afferents.

ASCENDING SPINAL PATHWAYS

145

Also the exteroceptive subgroups of the DSCT are of interest. One of these channels informs exclusively about pressure on the pad and each neurone is usually activated from one or part of one pad. It is easy to understand that this information is of importance in cerebellar integration of posture and movements. Neurones of the other exteroceptive channels are characterized by the lack of modality specificity in that they can be activated by touch, pressure and pinching. The significance of this convergence is not understood, but in any case the anterior cerebellar gets reasonably good spatial information regarding cutaneous events through this channel. The fifth subgroup finally is activated from the FRA (cf. Fig. 7) and cannot be classified either as exteroceptive or proprioceptive. The action from skin is often monosynaptic and from skin these neurones, as the fourth subgroup, respond on touch, pressure and pinching, with larger receptive field, for the different actions in the same order. In addition these units are activated by high threshold afferents from many muscles. The significance of the FRA influence will be discussed separately in section 7 in connection with the supraspinal control of transmission from the FRA. Under certain conditions of supraspinal control this DSCT channel may function like the fourth group as an exteroceptive channel. Some of the DSCT neurones that are activated from group I muscle afferents may in addition be influenced from the FRA, some are excited, others inhibited (Lundberg and Oscarsson, 1960; Lundberg et al., 1963). There is the possibility that these admixtures represent aberrancies in the connection to the two DSCT channels. In recent A

6

-

-

10 rnsec

2 rnsec

Fig. 11. Contribution of pathways not belonging to the DSCT to the mass discharge in the DLF. As in Fig. 1 recording from the left dissected DLF (ThlO); in each record simultaneous recording at two sweep speeds. Supramaximal stimulation of the left hamstring nerve in A and B and of the left superficial peroneal nerve at a strength of 1.03 times threshold in C and D and at 20 times threshold in E and F. The right records B, D and F were taken after a section of the dorsal part of the lateral funicle in L5. Observe the marked decrease of the discharge evoked from cutaneous afferents but that a small monosynaptic discharge remains in F. The discharge evoked from group I muscle afferents is not changed by the lesion but thereis a reduction of the latemass discharge (LundbergandOscarsson, 1961).

experiments we have seen these effects more frequently and it cannot be excluded that there are special subgroups of DSCT neurones with these mixed actions and that they may forward significant information. Rrferences p. 160-163

146

A. L U N D B E R G

(3)

THE SPINOCERVICAL TRACT

In the literature there has been a tendency to ascribe all discharges that can be recorded from the dorsal part of the lateral funicle to the DSCT. Anatomical investigations give good evidence for another pathway (for references see Busch, 1961) and the physiological investigations have made it possible to define a new pathway; the spinocervical tract. It is not known if this pathway is identical with the spino-olivary tract of Grundfest and Carter (1954). Figs. 11 and 12 illustrate the contribution that this pathway makes to the mass discharges in the dissected DLF, discussed in relation to Fig. I . Fig. 11 shows that after a lesion in L5 caudal to Clarke's column there is, as would be expected, no change in the group I evoked discharge which is in the DSCT, but a dramatic change in the mass discharge evoked from cutaneous afferents, the early part of the discharge a large part of which appears at very low strength of stimulation (C) cannot be evoked after this lesion. There remains a cutaneous change (F) having a latency 0.3 msec longer than that in E before the lesion. The discharge in F is in the DSCT (cf. section 2 ) but the early monosynaptic discharge in E must be

-

A m e d . fasc.

Hamstr.

sup. P

A

p"

Fig. 12. There was simultaneous recording from two dissected fascicles as shown in the drawing. In each record the upper trace is from the medial and the lower from the lateral funicle. The ipsilateral hamstring nerve was stimulated at a strength of 2 times threshold in A and 30 times threshold in C . The ipsilateral superficial peroneal nerve was stimulated in B and D at strengths of 1.05 and 20 times threshold respectively. The discharge from group I muscle afferents is in the lateral funicle. Most of the early monosynaptic discharge was evoked from cutaneous afferents particularly a t the lower threshold in the medial fascicle but there was a small component also in the lateral (arrow in D) (Lundberg and Oscarsson, 1961).

in a pathway having cells of origin caudal to Clarke's column. This pathway is more effectively activated than DSCT at low threshold stimulation of cutaneous nerves, but the differences in latency in E and F is probably caused by the slower conduction velocity that the cutaneous afferents have in the dorsal column where they pass up to Clarke's column (Lloyd and McIntyre, 1950). Fig. 12 illustrates that the DSCT and the pathway having origin caudal to Clarke's column have different location in the lateral funicle. Two fascicles were dissected as shown in the figure. Whereas the lateral fascicle contains the group I evoked DSCT discharge, the early cutaneous discharge is largely in the medial fascicle. It is, however, noteworthy that on stimulation of muscles afferents there is a late mass discharge evoked from high threshold muscle afferents also in both the medial and the lateral fascicle (record C). Likewise there is in Fig. 11 (A and B) after the lesion of the lateral funicle in L5 a reduction in the late mass discharge. With recording from single

ASCENDING SPINAL PATHWAYS

147

fibres two types of neurones were encountered, which were identified as not belonging to DSCT in that they could not be antidromically activated from the anterior cerebellar cortex and in that these axons could be stimulated in the lateral funicle in L5: (1) These are the most numerous and are monosynaptically activated by the most low threshold cutaneous afferents from the ipsilateral side. They are adequately activated by light touch from a very restricted cutaneous field but do not receive additional activation on pressure and pinching of the skin.

Tib

5 msec

Fig. 13. Identification of spinocervical tract axon. Microelectrode recording from axon in:the most dorsomedial part of the DLF in the lower thoracic region and in C also from the contralateral spinal half (except dorsal column). The neurone was monosynaptically activated from the tibia1 nerve and adequately exclusively by tactile stimulation of one toe. The lateral funicle was stimulated at the indicated levels in A and B. The axon could not be stimulated when the cervical electrode was moved to CI. The distance from the site of microelectrode recording to the stimulating electrode in C2 was 150 rnm and to the stimulating electrode in C5 62 mm (Lundberg and Norrsell, unpublished).

(2) Neurones, activated by ipsilateral cutaneous and high threshold muscle afferents. Most of these units could be activated by tactile stimuli from a receptive field, that could be relatively restricted. Additional activation from a larger area is evoked by pressures and pinching of the skin. These neurones resemble closely in their activation the FRA activated DSCT neurones but differ in that the latter receive excitatory action from the sensorimotor cortex (Lundberg et al., 1963). The axons of these two types of neurones conduct at 100-40 m/sec. Intracellular recording has been made from the cells of origin of these axons (Eccles et al., 1960). They are located close to the entry zone of the afferents in the dorsal horn. It has been shown that the axons of both types of neurones terminate ipsilaterally i n the upper cervical region (Fig. 13). The experimental arrangement is shown in the diagram. The axons could be stimulated with a movable electrode in C2 but not when the electrode was moved to C1 and not at any strength from a grid of electrodes in the Referencesp. 160-163

148

A. L U N D B E R G

lower medullary region. In the same experiments all DSCT axons could be traced up to corpus restiforme by electrical stimulation. Detailed studies have shown that these neurones forward information to the sensory areas of the cerebral cortex (Norrsell andvoorhoeve, 1962; Anderson, 1962b), and hence is the spinal pathways of Morin’s path to the cerebral cortex, relaying in the lateral cervical nucleus and after crossing, in thalamus (Morin, 1955; Catalan0 and Lamarche, 1957). Gordon and Jukes (1963) have recorded from cells of the lateral cervical nucleus and found responses of two kinds which correspond to the response of the two types of neurones in the spinocervical tract. Morin (1955) originally suggested that this path to the cerebral cortex is supplied by collaterals from the DSCT. This idea has been directly refuted by Norrsell and Voorhoeve (1962) and by Anderson’s demonstration (1962b) that after section of the dorsal column and the spinocervical tract no short-latency evoked potential can be evoked in the sensory area via the direct spinocerebellar tracts. Morin et al. (1963) have also recorded from cells in the lateral cervical nucleus. Despite the evidence for a special spinocervical tract they persist in assuming that the activation is through collaterals from the DSCT (cJtheir Fig. 8). Testing against the resting discharge Lundberg and Oscarsson (196 1) were never able to observe inhibitory action from the periphery on the tactile units of the spinocervical tract. However, presynaptic inhibition is not disclosed by this test and Eccles et al. (1962) have shown that transmission to the spinocervical tract can be presynaptically inhibited by reflex actions from the periphery. We have confirmed this finding also with unit recording. In recording the effect transmitted to cortical cells via the spinocervical pathway Anderson (1962b) did not find surround inhibition as was found with cells activated via the other pathways supplying cortex, the dorsal column system, and that would be expected from the findings of Gordon and Paine (1960) in the gracilis relay. Further attention must be given to this problem, but it seems possible that the presynaptic action cannot very markedly counteract the effective activation that is evoked in the spinocervical tract on adequate activation of the skin. The surround inhibition in the gracilis nucleus, being spatially restricted, is of another order of effectiveness. It is possibly evoked through a different mechanism. Anderson (1962b) has pointed out that the occurrence of inhibition in the dorsal column relay while sharpening the spatial discriminatory power may decrease the synaptic security. Morin’s pathway supplied by the spinocervical tract may mediate information about cutaneous events quickly and with higher degree of safety. Lundberg and Norrsell (1960) found that the tactile placing reaction in the cats hindlimb disappeared after a small lesion interrupting the spinocervical tract. However, this finding cannot be taken to indicate that the spinocervical tract is the afferent path in cortical reflex giving tactile placing because, when Morin’s path was transected in the contralateral ventral funicle after the relay in the lateral cervical nucleus, tactile placing remained. (4)

THE VENTRAL SPINOCEREBELLAR TRACT

Detailed electrophysiological investigations have given information not only of the

149

ASCENDING SPINAL PATHWAYS

function of this pathway but also of previously unknown anatomical data, as the location of the cells of origin and of its termination. These investigations started with recording from the superior cerebellar peduncles (SCP) where the VSCT is known to form a superficial layer. Volleys in contralateral group I muscle afferents evoke in the SCP a monosynaptically transmitted discharge A

B

C

loo]

J

0

. i 0

Ib

10

n

Fig. 14. Activation of VSCT from Ib afferents. Recording was made with a surface electrode in contact with the superior cerebellar peduncle (SCP) (upper traces). The contralateral hamstring nerve was stimulated at increasing strength. The lower traces show the incoming volley recorded at the dorsal root entry zone. Observe that the discharge in the SCP appears with the Ib volley. This is further shown in the curve in which 100% on the ordinate represents the unconditioned VSCT discharge (Holmqvist and Oscarsson, 1956).

(Fig. 14) that is caused entirely by group Ib impulses (Holmqvist and Oscarsson, 1956; Oscarsson, 1956). This discharge remains after contralateral transection of the cord and after ipsilateral section of the dorsal part of the lateral funicle, but disappears after a superficial lesion in the ipsilateral part of the lateral funicle (Oscarsson, 1956). The further analysis has been made with recording from the spinal cord in the midthoracic level. Stimulation of contralateral muscle afferents give rise to a mass discharge consisting of a monosynaptically transmitted early spike discharge which is caused by Ib impulses and a later mass discharge (Fig. 22a) caused by impulses in high threshold afferents (Oscarsson, 1957, 1958). The late mass discharge is not to any significant extent in the VSCT but in other pathways, that will be dealt with in section 5 and 6 . With recording from single fibres, Oscarsson (1957) was able to show that all neurones that could be monosynaptically activated from contralateral group I muscle afferents belong to VSCT, because they could be antidromically activated References p . 160-163

150

A. L U N D B E R G

from the SCP. In further experiments (Lundberg and Oscarsson, 1962a) these units have also been identified by antidromic stimulation from the cerebellar cortex. A VSCT unit is shown i n Fig. 15 and in G is shown that this unit had a bilateral terminal field in the anterior cerebellar cortex. The diagram to the right shows that the VSCT fibres terminate in longitudinal zones consisting in a medial strip of the intermediate cortex and a lateral strip of the vermal cortex. The majority of the VSCT fibres

Fig. 15. Antidromic identification of VSCT axon from the anterior cerebellar cortex. Lower traces are records from a VSCT axon on the left side and the upper traces are mass discharges from the right dissected ventral quadrant. The neurone was activated monosynaptically from the right hamstring nerve (H, record E with simultaneous recording at two speeds). A and B show antidromic spikes elicited from the two cerebellar termination areas (left and right, Cbl and Cbr), that are indicated in diagram G . Note slight latency differencies. C and D show antidromic spikes elicited from the superior cerebellar peduncle (SCP) and Th 8 respectively. The diagram to the right shows the terminal areas of VSCT axons. Vertical lines indicate border of the intermediate regions, horizontal lines sulci. Observe that recording was made from axons on the left side and that the main terminal area is on the right side (Lundberg and Oscarsson, 1962a).

terminate contralaterally, but some ipsilaterally and some, after branching, both contralaterally and ipsilaterally as shown for the unit in Fig. 15. There was not as for DSCT axons a distinct terminal field, but slightly above threshold the fibres could be activated from fairly large areas. This suggests profuse terminal branching and indicates that single fibres make synaptic contact with cells scattered in large areas of the cortex. The lateral extent of this termination agrees well with that found by Grant ( I 962). The VSCT cell bodies are located in the lateral part of the intermediate zone and base and neck of the dorsal horn (Hubbard and Oscarsson, 1961). After crossing the VSCT ascends in the lateral part of the lateral funicle and enters the SCP on the same side. As shown above the main terminal area is contralateral to that of the tract; in other words, there is a double crossing and the main termination is ipsilateral to the side supplying the Ib monosynaptic excitation. VSCT has a higher conduction velocity than DSCT, the mean value being 92 and 78 m/sec and the fastest 120 and 110 m/sec respectively (Oscarsson, 1956, 1957; Lundberg and Oscarsson, 1960, 1962a, b). The convergence of Ib effects on VSCT neurones have been studied with recording from axons (Oscarsson, 1957) and with intracellular recording from VSCT cells (Eccles, Hubbard et al., 1961). By contrast to the DSCT there is a much more extensive convergence of monosynaptic excitation from contralateral nerves. This con-

ASCENDING SPINAL PATHWAYS

151

vergence is not limited to Ib afferents of synergic muscles. Most of the VSCT neurones belong to two groups, one activated from hip extensors, knee flexors, ankle and toe extensors (Fig. 16), the other group mainly from knee, ankle and toe extensors. Oscarsson (1960) suggests that the VSCT neurones carry information not of changes of tension in single muscles of synergic muscles acting at one joint but concerning

EJ/ PBST

Nk

Q 0

GI

msec

5 msec

Fig. 16. Patterns of excitatory and inhibitory convergence on two VSCT cells obtained with intracellular recording from the cell bodies (upper traces). Lower traces in A-G and I-Q record the incoming volley at the L7 dorsal root entry zone about 2.5 cm caudal to the cells. Records A-H are from one cell, H showing the antidromic spike on stimulation of the contralateral dissected spinal half. Records I-Q are from another VSCT cell. Antidromic invasion was observed, but the cell was lost before the records had been taken. The various nerves stimulated are indicated on each record. Abbreviations: S, sural; CP, common peroneal nerve; FDHL, flexor digitorum and hallucis longus; GS, gastrocnemius-soleus; PBSt, posterior biceps-semitendinosus; Q, quadriceps; SP, superficial peroneal; DP, deep peroneal; AbSm, anterior biceps-semimembranosus. SA in record I stands for sacral roots 2 and 3. The stimulus strength was supramaximal for group I but not for group I1 afferents (Eccles, Hubbard et al., 1961).

stages of movements or position of the whole limb. It is in agreement with this hypothesis that VSCT neurones in their adequate activation from tendon organ require the participation of many muscles (Oscarsson, 1960). VSCT neurones are also characterized by having strong effects from the FRA. References p . 160-163

152

A. L U N D B E R G

Oscarsson (1957) has divided VSCT into two groups, the I-neurones receiving inhibitory actions from the FRA and the E-neurones receiving excitatory actions from the FRA. The former neurones are more common. Fig. 17 shows the inhibitory action of conditioning volleys in muscle and joint afferents on the test VSCT discharges recorded in the spinal half. The inhibition lasting for about 85 msec is followed by a RIGHT HAMSTRING NERVE

-IVSCT

(H)

% IOC

50

0

50

100

RIGHT P KNEE JOINT

0

50

100

150 NERVE

150

ZOO

ti +SCT (0 GS)

---\ % ' a-

250

-

VSCT(H)

ZOO

250

3OOmsrc 5c

--

3OOmaec

\

"\

Multi les of +hres(old for H nerve

Fig. 17. Inhibition of VSCT from high threshold muscle and joint afferents. In the curves loOD,; on the ordinate is the unconditioned VSCT discharge recorded in the dissected spinal half and evoked from the contralateral hamstring (H) nerve. The abscissa gives the interval between conditioning and testing volleys. A conditioning group I volley in the contralateral H nerve (same as used for testing) gave no inhibition ( X ) but when high threshold afferents were stimulated (0) a large inhibition resulted. The effect in the lower left curve was evoked from contralateral high threshold joint afferents. In the right curve the abscissa (logarithmic scale) gives the strength of conditioning stimulation in multiples of threshold strength. The outer ordinate is the amplitude of conditioned VSCT discharge in per cent of the unconditioned. Inner ordinate represents amplitude of incoming volleys in per cent of the maximal group I volley. Conditioning volleys were given in the contralateral hamstring nerve and the test VSCT discharge was evoked on combined stimulation of the contralateral quadriceps and gastrocnemius-soleus nerves. The interval between conditioning and testing volleys was 18 msec. Observe that the inhibitory action is evoked from group I1 and I11 muscle afferents (Oscarsson, 1957).

phase of facilitation. The curve to the right in Fig. 17 shows that the inhibition from muscle nerve is evoked by group 11 and 111 afferents. The joint afferents responsible are the high threshold ones activated at a strength above 3 times threshold. A very effective inhibition is also exerted from low and high threshold cutaneous afferents. These inhibitory actions have also been found with intracellular recording (Fig. 16) (Eccles, Hubbard et al., 1961). Inhibition is not, however, evoked exclusively from the FRA. Oscarsson (1957) showed that Ib impulses sometimes can evoke inhibitory actions and this has been confirmed by Eccles, Hubbard et al. (1961). This inhibition is often tri- or polysynaptic and in addition there may be a disynaptic inhibition from Ia muscle afferents. There is probably also presynaptic inhibition of transmission from Ib to VSCT (Eccles et al., 1963b). With adequate stimulation in the spinal cat the VSCT is dominated by the FRA which give inhibition of a resting discharge in I units and excitatory action in E units (Oscarsson, 1957). There are clear effects from muscles evoked both on stretch and contraction and high threshold afferents are responsible for these effects. On con-

153

ASCENDING SPINAL PATHWAYS

traction the inhibitory effect from the FRA conceals the facilitatory Ib action in the spinal state. Marked effects are also obtained from skin particularly from the contralateral side. The I units are inhibited but do usually receive excitation from the ipsilateral limb and from the proximal part of the contralateral limb. The E units receive excitatory effects from both hindlimbs. Effects are evoked on light touch but increase with stronger stimulation. The significance of the FRA effects to VSCT will be discussed in section 6 in connection with the supraspinal control of pathways. The VSCT neurones discussed above are those with monosynaptic excitatory action from contralateral Ib afferents. It is not certain that all VSCT neurones do receive this action of the VSCT axons identified with antidromic stimulation from the cerebellar cortex, 46% could not be activated monosynaptically from group I muscle afferents. These neurones may constitute a group of their own but another possibility is that they received monosynaptic excitation from undissected nerves.

(5)

A VENTRAL SPINOBULBAR TRACT

On stimulation of contralateral nerves a late mass discharge (Fig. 18, lower traces in records F-J) can be recorded from the spinal half (Oscarsson, 1958). It is caused

Q

Contra stim

SM

PBSt

-

" -UiL

S

sc

IJU- ++JwaL

'Ivc"

3iilscc

Fig. 18. Convergence onto a bVFRT neuron. Recording from the dissected spinal half (lower trace) and from an axon ascending in the ventral part of the lateral funicle. Supramaximal stimulation of the quadriceps (Q), posterior biceps-semitendinosus (PBSt), semimembranosus (SM), saphenus (S) and sciatic (except hamstring) nerves (Sc). Upper records show the effect from ipsilateral and lower from contralateral nerves. Lower right record is the antidromic spike appearing on stimulation of the dissected spinal half (Oscarsson, 1958).

by activity in a ventral pathway and it has already been illustrated in Fig. 1 that such a late discharge is also evoked in the ipsilateral ventral cord. Oscarsson found that these discharges result on stimulation of group I1 and 111 muscle afferents, high threshold joint afferents and cutaneous afferents, i.e. the FRA that provide inhibition to the VSCT. Axonal recording from the ventral part of the lateral funicle revealed axons of neurones that could be activated from these afferents of a bilateral receptive field as shown in Fig. 18 (Oscarsson, 1958; Lundberg and Oscarsson, 1962b). The axons of these neurones are denoted the bilateral ventral flexor reflex tract (bVFRT) and the mean conduction velocity of the axons is 84 mjsec ranging from 50 to 100 mjsec. Cell bodies of the bVFRT are probably located in the ventromedial part of the grey matter (Oscarsson, unpublished). References p . 160-163

I54

A. L U N D B E R G

Investigations regarding the termination of bVFRT (or at least part of it, see next section), has depended on a monosynaptic connection (Fig. 19) that neurones of this type receive from higher centres (Holmqvist et al., 1960b). This descending path takes its origin from the brain stem at the level of the Deiters’ nucleus, but it is not known if the cells are in the vestibular nucleus or in the lateral reticular formation. The descending axons conduct at 100 m/sec and they are extremely effective in activating bVFRT neurones, many of which can follow discharge frequencies of more than 500/sec (Holmqvist et al., 1960b; Lundberg and Oscarsson, 1962b). This descending pathway makes monosynaptic connections also with the VSCT (Oscarsson, 1957; Eccles, Hubbard et al., 1961 ; Lundberg and Oscarsson, 1962a) but in VSCT neurones the action is weaker and does not often give discharges. The monosynaptic discharge evoked from this descending tract has been traced in the brain stem in order to find the termination of the bVFRT (Lundberg and Oscarsson, 1962b). In the lower medullary region the action can be recorded just medially to the VSCT discharge (Fig. 20) (cf. also Bohm, 1953). At a somewhat more rostra1 level of the superior cerebellar peduncle the discharge is smaller but found more dorsally. More rostrally in the nervous system there was only once a trace of action at the level of the inferior colliculi as shown I Fig. 20. Neuroanatomical investigations have revealed termi-

5 msec

7 7

Fig. 19. Descending monosynaptic activation of bVFRT neurone. The upper traces in A-C and E are mass discharges recorded from the dissected spinal half. The lower traces in A-E are from a bVFRT axon ascending in the ventral part of the lateral funicle. Stimulation of the brain stem (BS) through a needle electrode (inserted through the cerebellum to the site shown in the right diagram) was made in record E, the upper trace shows that a single stimulus evokes a monosynaptic mass discharge in the contralateral spinal half. The lower trace with axonal recording shows that the bVFRT neurone was activated from the brain stem. The discharges in A-C are evoked as indicated from the left hamstring nerve (IH), the left superficial peroneal (ISP) nerve and the right tibia1 nerve (rTib). The antidromic spike in D is evoked from the spinal half (Holmqvist et a/., 1960).

nation of ventral ascending spinal pathways in the lateral reticular nucleus (Brodal, 1949) as well as in other reticular nuclei (Rossi and Brodal, 1957; Nauta and Kuypers, 1958) and there is also a spinotectal tract (Morin et a/., 1951 ; Anderson and Berry, 1959). It is not possible to decide if the bVFRT is one of these pathways but it is only tentatively suggested that it is a spinoreticular pathway possibly with a spinotectal component.

ASCENDING SPINAL PATHWAYS

155

Fig. 20. Termination of the spinobulbar pathway. The left dissected spinal half (except dorsal column) was stimulated in a descending direction in order to activate the descending pathway making monosynaptic connection with bVFRT neurone. The discharge is recorded in the right brain stem with a needle electrode (A and B). For comparison the VSCT discharges evoked from Ib afferents of the left hamstring nerve are also recorded (C and D). The left diagram shows the location of VSCT and bVFRT discharges in the brain stem at a level 3 mm rostra1 to obex. In A and C recording is made from the right inferior colliculus a t the level of the commissure (diagram E), in B and D just caudal to the inferior colliculus (diagram F) (Lundberg and Oscarsson, 1962b).

On adequate activation the bVFRT neurones are excited by the same stimuli that give excitation to the dorsal FRA pathways and inhibition to VSCT. There is the characteristic feature of activation from skin both on touch, pressure and pinching (Lundberg and Oscarsson, 1962b). It is also possible to inhibit this pathway from the periphery. All adequate stimuli of muscle and skin that can give excitation can also give inhibition to this pathway and there is often inhibition and excitation to the same cells from different regions and sometimes even from the same region of the receptive field (Lundberg and Oscarsson, 1962b; Landgren, Lundberg and Vyklick$, unpusblished). Anderson (1962a) and Anderson et a/. (1964) have shown that activity in this pathway can give synchronization or desynchronization of slow waves in the cerebral cortex, but the pathway may have other functions as well (cf. section 7). (6)

OTHER ASCENDING SPINAL PATHWAYS

Other pathways have been recorded but not been identified either with respect to origin or termination. Mention should first be made of the subgroup of bVFRT neurones similar in their activation pattern from the FRA to the spinobulbar tract of section 5, but not receiving any monosynaptic descending excitation from the brain stem. These neurones may be subgroups of the ventral spinobulbar tract, but since the termination of the latter pathway was investigated on basis of the descending monosynaptic activation it is obvious that special experiments are required in order to decide if the two types of bVFRT neurones have the same termination. Lundberg and Oscarsson (1962b) described one ventral pathway activated exclusively from the contralateral FRA, the latency of activation was so brief that monosynaptic actions from the FRA could not be excluded. The question was raised if these fibres could be spinothalamic (cf. Anderson and Berry, 1959) but in further experiments in which antidromic stimulation was tried from the mesencephalon it was Refcrenres p.:16O-i63

156

A. L U N D B E R G

not possible to stimulate these axons antidromically (Landgren, Lundberg and Vyklicki, unpusblished). Reference should finally be made to a third type of neurones found by Lundberg and Oscarsson (1961, 1962b). The axoiis are located in both the dorsal and ventral parts of the cord and characterized by being discharged from the FRA after a long latency. In these experiments part of the spinal cord was in intact connection in higher centres in order to allow antidromic identification from higher centres. With these long latencies it was therefore difficult to exclude that the discharges were recorded from descending axons, that could be activated in the brain stem. However, axons with a similar type of activation pattern have been found also in spinal cats and it is possible that there is an ascending spinal pathway with this characteristics. This is, however, by no means proved since, for example, VSCT neurones can respond in this fashion. In further experiments (Landgren, Lundberg and Vyklickf, unpublished) with stimulation from the brain stem it has been found that these axons can be stimulated from medial regions of the brain stem but in no case was it possible to decide ifit was an ascending axon or the axon of a descending reticulospinal neurone activated from the FRA as described by Magni and Willis (1963). (7)

S I G N I F I C A N C E O F THE EFFECTS FROM THE

FRA

ON ASCENDING SPINAL

PATHWAY

With some of the pathways discussed in the preceding sections it has been rather easy to formulate a hypothesis regarding their function without considering the supraspinal control. This holds true for most of the subdivisions of the DSCT and of the spinocervical tract. With many of the pathways influenced from the FRA the situation is different. It has been a striking finding that so many pathways are influenced by the FRA. This holds true for at least one and possibly three subdivisions of the DSCT (section 2), for one subdivision of the spinocervical tract (section 3) and for all neurones of the VSCT (section 4) and of the spinobulbar pathway, described in section 5. Since both cutaneous and muscle afferents contribute to the action the information can neither be classified as proprioceptive nor as exteroceptive and, at least in the latter two pathways the action from the FRA can only give very crude spatial information. It is, however, noteworthy that these afferents are those evoking the flexor reflex (Eccles and Lundberg, 1959b), the receptive field of which likewise is very large. These similarities raised the question if the FRA message of the ascending pathways in some way could be informative of flexor reflex events in the spinal cord (Eccles and Lundberg, 1959b; Lundberg, 1959; Holmqvist et al., 1960a). This possibility was particularly interesting because of the descending supraspinal control of the flexor reflex paths. Transmission from the FRA to motoneurones can be tonically inhibited from the brain stem (Eccles and Lundberg, 1959a; Holmqvist and Lundberg, 1959) and facilitated from the corticospinal tract (Lundberg and Voorhoeve, 1962). Furthermore there is evidence of alternative paths from the FRA to motoneurones, the descending paths presumably select which path that should be open i n a given situation

ASCENDING SPINAL PATHWAYS

157

(Holmqvist and Lundberg, I961 ; Holmqvist, 1961). Information regarding transmittsbility in flexor reflex paths would undoubtedly be valuable information for higher centres and the FRA information of ascending pathways could at least partly have this function. PER CENT 140 6

Fig. 21. Supraspinal inhibitory control of transmission from the FRA to VSCT and bVFRT neurones. In the left curves 100% on the ordinate is the unconditioned VSCT monosynaptic test discharge evoked from the contralateral left gastrocnemius-soleus and posterior biceps-semitendinosus nerves (G-S -1 PBSt) and recorded from the right dissected ventral quadrant. The effect of a conditioning volley in the nerve to the left flexor digitorum longus and in the left sural nerve are shown in the upper and lower curves respectively. The experiment was made on a decerebrate cat in which both the ventral quadrants had been sectioned and the effects were examined before ( 0 ) and after ( Y ) cold blockage of the intact dorsal cord. After rewarming the decerebrate inhibition of transmission from the FRA returned. Conditioning stimulus strength is expressed in multiples of threshold strengths and the initial facilitatory effect in the upper curve was evoked from group I afferents. The right records A-F are recorded from the right spinal half (except dorsal column) in a decerebrate cat. The left hamstring nerve (IH) is stimulated and there is a release of the late mass discharge, evoked from high threshold muscle afferent, after transection of the remaining dorsal part of the cord. As has been found for the action from the FRA to motoneurones (Holmqvist and Lundberg, 1959) there is no further release in E and F after section of the ventral quadrant (Holmqvist et al., 1960a).

From this point of view it is of considerable interest that the supraspinal control systems influence transmission from the FRA t o ascending spinal pathways in the same way as to motoneurones. Fig. 21 illustrates this for the descending inhibitory control from the brain stem that is tonically active in the decerebrate state and probably exerted at an interneuronal level (Holmqvist et a/., 1960a; Oscarsson, 1960). Fig. 22 shows the effect from the sensorimotor cortex; the action from cortex parallels those from the FRA and is caused by excitation from the corticospinal tract of interneurones transmitting actions from the FRA (Magni and Oscarsson, 1961 ; Lundberg et al., 1963). The connections to the spinobulbar pathway are summarized in Fig. 23. It has been described in section 5 that adequate stimulation can evoke either excitatory or inhibitory effects in these iieurones and there is evidence that the inhibition is postReferences p . 160-163

158

A. L U N D B E R G C

S*H

A D cortextH

Fig. 22. Effects from the sensorimotor cortex on VSCT and bVFRT neurones. Upper traces were recorded from a VSCT axon (A-D) and from a bVFRT axon (E-J) and lower traces from the dissected right ventral quadrant. The left hamstring nerve was stimulated in A-D (10 superimposed traces). In A at supramaximal strength and in B-D at a slightly submaximal group I strength which unconditioned regularly elicited a spike in the neurone. In C there is inhibition of the spike and the mass discharge by a conditioning volley in the left sural nerve and the inhibition in D was produced by repetitive stimulation of the right sensorimotor cortex. In the bVFRT neurone the discharges in E, F, G and I were evoked as indicated on stimulation of the right and left hamstring nerves (RH and LH) and of the right and left sural nerves (RS and LS). The mass discharge and axonal discharge in G and J were evoked from the sensorimotor cortex on the left and right side respectively (Magni and Oscarsson, 1961).

IPSIL. V T EXC ITAT ION

BIL. PYR. TRACT EXC I TAT ION

L

BIL. RET.-SPINAL IN H I BIT ION

BIL. RET. -SPI NAL INHIBITION

1 P

BIL. FRA EXCl TAT ION

?

BIL. FRA INHIBITION

Fig. 23. Organization at the segmental level of actions from the FRA and of three supraspinal control systems influencing the transmission to bVFRT neurones. Ecxitatory internewones and synaptic knobs are indicated by open circles, inhibitory internewones and synaptic knobs by filled circles Polysynaptic connections are drawn as disynaptic. Ipsil. vt excitation is provided from the tract originating in the brain stem and descending in ventral quadrant to make monosynaptic contact with the neurones. Bil. ret.-spinal inhibition is provided from the centres giving tonic inhibition of transmission from the FRA in the decerebrate cat. The action is drawn to be exerted at an inteineuronal level but this has not been proved. Bil. pyr. tract excitation is the excitatory action exerted from the corticospinal tract on interneurones transmitting effects from the FRA. There is probably also an action from the corticospinal tract on the inhibitory interneurones ttansmitting effect from the FRA (Lundberg and Oscarsson, 1962b).

ASCENDING SPINAL PATHWAYS

159

synaptic (Carpenter et al., 1964), hence the connection from the FRA through excitatory and inhibitory interneurones. It is postulated that the same afferents supply both paths. Through the reticulospinal inhibitory system there can be suppression of both the excitatory and inhibitory paths. It is suggested that the spinobulbar pathway can forward two messages: (1) Excitation which is dominating in the spinal state; (2) Inhibition (through an alternative path from the same afferents) forwarded as cessation of a discharge provided from the very effective descending monosynaptic connection drawn in Fig. 23 and already discussed in section 5. The alternative excitatory and inhibitory pathways from the FRA resemble the situation with motoneurones where this also has been found for paths to ipsilateral flexor and contralateral extensor and flexor motoneurones (Holmqvist and Lundberg, 1961; Holmqvist, 1961). The parallelism is still more striking if the release from the tonic decerebrate supraspinal control is considered. A low pontine lesion that releases the inhibitory paths from the FRA to motoneurones does also release the inhibitory path to the spinobulbar tract. Likewise the excitatory path to the spinobulbar tract and t o motoneurones are released in parallel by a more caudal medullary lesion (Carpenter et al., 1964). These similarities may indicate that the inhibitory and excitatory paths to ascending pathways are related to the corresponding paths to motoneuroncs. The connections to the VSCT have been summarized by Magni and Oscarsson BIL.RET.-SPINAL INHIBITION

PYR. TRACT

EXTRAPYR.

IPSIL.VT EXCITATION

I

I b

F R A f r o m periphery of receptive field F R A f r o m centre of receptive field

Fig. 24. Organization of three ascending and four descending pathways to VSCT. As in Fig. 23 but in addition there is extrapyramidal facilitation that can be evoked from an anterior area of the cerebral cortex and (possibly via the rubrospinal tract) give short latency facilitation of VSCT neurones (Magni and Oscarsson, 1961). References p . 160-163

160

A. L U N D B E R G

(1961) and Fig. 24 is from their paper. Apart from the two control systems discussed above, there is also a weak monosynaptic excitation from the descending system discussed above (Oscarsson, 1957) and also extrapyramidal facilitation from the cerebral cortex, possibly via the rubrospinal tract(Magni and Oscarsson,l961). The diagram shows that this extrapyramidal facilitation can be excluded by collateral inhibition from inhibitory interneurones transmitting effects from the FRA. Magni and Oscarsson (1961) have suggested that when the reticulospinal inhibitory control of transmission from the FRA is operating the VSCT transmits monosynaptic I b actions, whereas otherwise an FRA message is forwarded as inhibition of the discharge provided from the two descending excitatory paths. The FRA actions forwarded by the direct spinocerebellar tracts, particularly the strong effects on all VSCT neurones, are of particular interest with regard to the hypothesis that the ascending FRA effects are related to spinal reflex action and of importance in motor regulation. The ideas discussed above may seem very complex, but I do believe, that they give more fruitful basis for further work than, for example, speculations that informations from the different afferent systems acting on these pathways can be coded by higher centres from the discharge patterns. It should not necessarily be assumed that all FRA actions to ascending pathways should be related to reflex events, particularly the FRA subdivision of the spinocervical tract may inform more directly about receptor events.

SUMMARY

This review deals with ascending spinal pathways activated from hindlimb afferents. There is a discussion of the functional organization of the following pathways: (a) the dorsal spinocerebellar tract; (b) the spinocervical tract; (c) the ventral spinocerebellar tract; (d) a spinobulbar tract. The last section deals with the significance of the actions from the flexor reflex afferents on ascending spinal pathways. REFERENCES ANDERSON, F. D., AND BERRY, C. M., (1959); Degeneration studies of long ascending fibre systems in the cat brain stem. J . con?p. Neurol., 111, 2. 195-222. ANDERSON, R. F., (1943); Cerebellar distribution of the dorsal and ventral spinoccrebellar tracts in the white rat. J . c o t p . Neurol., 97,415423. ANDERSON,S. A,, (1962a); Cortical effects by activity in a ventral ascending spinal pathway. Med. Exp., 6 , 21-24. ANDERSON,S. A,, (1962b); Projection of different spinal pathways to the second somaticsensoryarea in cat. Acraphysiol. scand., 56, Suppl. 194, 1-74. ANDERSON,S. A., NORRSELL, U., AND WOLPOW,E. R., (1964); Cortical synchronization and desynchronization via spinal pathways. Acta physiol. scand., in the press. BECK,G . M., (1927); The cerebellar terminations of the spinocerebellar fibres of the lower lumbar and sacral segments of the cat. Brain, 50, 60-98. BOHM,E., (1953); An electro-physiological study of the ascending spinal anterolateral fibre system connected to coarse cutaneous afferents. Acta physiol. scand., 29, Suppl. 106, 106-1 37.

ASCENDING SPINAL PATHWAYS

161

BRADLEY, K., AND ECCLES,J. C., (1953); Analysis of the fast afferent impulses from thigh muscles. J . PhyJiol. (Lond.), 122, 462-473. BRODAL, A., (1949); Spinal afferents to the lateral reticular nucleus of the medulla oblongata. J. comp. Neurol., 91, 259-295. BRODAL, A., AND JANSEN, J., (1941); Beitrag zur Kenntnis der Spino-Cerebellaren Bahnen beim Menschen. Anat. Anz., 91, 185-195. BUSCH,H. F. M., (1961); An anatomical Analysis of the White Matter in the Brain Stem of the Cat. Assen, Van Gorcum. CARPENTER, D., ENGBERG, I., AND LUNDBERG, A., (1964); Decerebrate control of inhibitory and excitatory actions from the FRA to ascending pathways, to be published. CATALANO, J. V., AND LAMARCHE, G., (1957); Central pathway for cutaneous impulses in the cat. Amer. J. Physiol., 189, 141-144. CURTIS,D. R., ECCLES,J. C., AND LUNDBERG, A., (1958); Intracellular recording from cells in Clarke’s column. Acta physiol. scand., 43, 303-314. ECCLES, J. C. ECCLES, R. M., AND LUNDBERG, A., (1 957); Synaptic actions on motoneurones in relation to the two components of the group I muscle afferent volley. J . Physiol. (Lond.), 136, 527-546. ECCLES,J. C., ECCLES,R. M., AND LUNDBERG, A., (1960); Types of neurone in and around the intermediate nucleus of the lumbosacral cord. J. Physiol. (Lond.), 154, 89-1 14. ECCLES, J. C., HUBBARD, J. I., AND OSCARSSON, O., (1961); Intracellular recording from cells of the ventral spino-cerebellar tract. J . Physiol. (Lond.), 158, 486-516. ECCLES,J. C., KOSTYUK, P. G., AND SCHMIDT, R. F., (1962); Presynaptic inhibition of the central actions of flexor reflex afferents. J. Physiol. (Lond.), 161, 258-281. ECCLES, J. C., OSCARSSON, O., AND WILLIS,W. D., (1961); Synaptic action of group I and I1 afferent fibres of muscle on the cells of the dorsal spinocerebellar tract. J . Physiol. (Lond.), 158, 517-543. ECCLES, J. C., SCHMIDT, R. F., AND WILLIS,W. D., (1963a); Depolarization of central terminals of group Ib afferent fibres of muscle. J . Neurophysiol., 26, 1-27. ECCLES, J . C., SCHMIDT, R. F., AND WILLIS,W. D., (1963b); Inhibition of discharges into the dorsal and ventral spinocerebellar tracts. J . Neurophysiol., 26, 635-645. ECCLES, R. M., AND LUNDBERG, A., (1959a); Supraspinal control of interneurones mediating spinal reflexes. J . Physiol. (Lond.), 147, 565-584. ECCLES, R. M., AND LUNDBERG, A., (1959b); Synaptic actions in motoneurones by afferents which may evoke the flexion reflex. Arch. ital. Biol.,97, 199-221. FLECHSIG, P., (1876); Die Leitungsbahnen im Gehirn und Ruckenmark des Menschen. Leipzig, Engelmann (p. 382). GORDON, G., AND JUKES,M. C. M., (1963); An investigation of cells in the lateral cervical nucleus of the cat which respond to stimulation of the skin. J. Physiol. (Lond.), 169, 28P. GORDON, G., AND PAINE,C. H., (1960); Functional organization in nucleus gracilis of the cat. J . Physiol. (Lond.), 153, 331-349. GRANT,G., (1962); Spinal course and somatotopically localized termination of the spinocerebellar tracts. Acta physiol. scand., 56, Suppl. 193, 1 4 5 . GRUNDFEST, H., AND CAMPBELL, B., (1942); Origin, conduction, and termination of impulses in the dorsal spino-cerebellar tract of cat. J. Neurophysiol., 5, 275-294. GRUNDFEST, H., AND CARTER, W. B., (1954); Afferent relations of inferior olivary nucleus, I. Electrophysiological demonstration of dorsal spino-olivary tract in cat. J . Neurophy~iol.,17, 72-91. HOLMQVIST, B., (1961); Crossed spinal reflex actions evoked by volleys in somatic afferents. Acta physiol. scand., 52, Suppl. 181, 1-67. HOLMQVIST, B., A N D LUNDBERG, A., (1959); On the organization of the supraspinal inhibitory control of interneurones of various spinal reflex arcs. Arch. ital. Biol., 97, 340-356. HOLMQVIST, B., AND LUNDBERG, A., (1961); Differential supraspinal control of synaptic actions evoked by volleys in the flexion reflex afferents in a-motoneurones. Actaphysiol. scand., 54, Suppl. 186, 1-51. HOLMQVIST, B., LUNDBERG, A., AND OSCARSSON, O., (1956); Functional organization of the dorsal spino-cerebellar tract in the cat. V. Further experiments on convergence of excitatory and inhibitory actions. Acta physiol. Jcancl., 38, 77-90. HOLMQVIST, B., LUNDBERG, A., AND OSCARSSON, O., (1960a); Supraspinal inhibitory control of transmission to three ascending pathways influenced by the flexion reflex afferents. Arch. ital. Biol., 98, 60-80. HOLMQVIST, B., LUNDBERG, A., AND OSCARSSON, O., (1960b); A supraspinal control system monosynaptically connected with an ascending spinal pathway. Arch. ital. Biol.,98, 402-422.

162

A. L U N D B E R G

HOLMQVIST, B., AND OSCARSSON, O., (1956); Synaptic connections of Ib muscle afferents to ventral spinocerebellar tract neurons. Experientia (Basel), 1218,296. HUBBARD, J. I., AND OSCARSSON, O., (1961); Localization of the cell bodies of the ventral spinocerebellar tract in lumbar segments of the cat. J . comp. Neurol., 118, 199-204. LAPORTE, Y . ,AND BESSOU,P., (1957); Etude des sous-groupes lents et rapides du groupe I (fibres afferents d’origine musculaire de grand diamktre) chez le chat. J . Physiol. (Lond.), 49, 1025-1037. LAPORTE,Y . , AND LUNDBERG,A., (1955); Analyse du faisceau spinocerebelleux dorsal chez le chat a I‘aide de microelectrodes intra-axonales. Microphysiol. comp. Pltfments excitables. 67, 435-457. LAPORTE, Y., AND LUNDBERG, A., (1956); Functional organization of the dorsal spino-cerebellar tract in the cat. 111. Single fibre recording in Flechsig’s fasciculus on adequate stimulation of primary afferent neurones. Acta physiol. scand., 36,205-218. LAPORTE, Y.,LUNDBERC,A., AND OSCARSSON,O., (1956a); Functional organization of the dorsal spinocerebellar tract in the cat. I. Recording of mass discharge in dissected Flechsig’s fasciculus. Acta physiol. scand., 36, 175-187. LAPORTE, Y.,LUNDBERC,A., AND OSCARSSON, O., (1956b); Functional organization of the dorsal spino-cerebellar tract in the cat. 11. Single fibre recording in Flechsig’s fasciculus on electrical stimulation of various peripheral nerves. Acta physiol. scand., 36, 188-201. LLOYD,D. P. C., AND MCINTYRE, A. K., (1950); Dorsal column conduction of group I muscle afferent impulses and their relay through Clarke’s column. J . Neurophysiol., 13,39-54. LORENTE D E NO, R., (1924); Etudes sur le cerveau posterieur. Trav. Lab. Rech. biol. Univ. Madrid, 22, 51-65. LUNDBERG, A., (1959); Integrative significance of patterns of connections made by muscle afferents in the spinal cord. Symp. X X I Congr. int. Ciencias Fisiol. Buenos Aires, 100-105. LUNDBERC, A., AND NORRSELL, U., (1960); Spinal afferent pathway of the tactile placing reaction. Experientia (Basel), 1613, 123. LUNDBERG, A., NORRSELL, u., AND VOORHOEVE, P., (1963); Pyramidal effects on ascending spinal pathways. Acta physiol. scand., 59,462473. LUNDBERG, A., AND OSCARSSON, O., (1956); Functional organization of the dorsal spino-cerebellar tract in the cat. IV. Synaptic connections of afferents from Golgi tendon organs and muscle spindles. Acta physiol. scand., 38,53-75, LUNDBERG, A., AND OSCARSSON, O., (1960); Functional organization of the dorsal spino-cerebellar tract in the cat. VII. Identification of units by antidromic activation from the cerebellar cortex with recognition of five functional subdivisions. Acta physiol. scand., 50, 356-374. LUNDBERC, A., AND OSCARSSON, O., (1961); Three ascending spinal pathways in the dorsal part of the lateral funiculus. Acta physiol. scand., 51, 1-16. LUNDBERG, A., AND OSCARSSON, O., (1962a); Functional organization of the ventral spinocerebellar tract in the cat. IV. Identification of units by antidromic activation from the cerebellar cortex. Acta physiol. scand., 54,252-269. LUNDBERG, A., AND OSCARSSON, O., (1962b); Two ascending spinal pathways in the ventral part of the cord. Acta physiol. scand., 54,270-286. LUNDBERG, A., AND VOORHOEVE, P., (1962); Effects from the pyramidal tract on spinal reflex arcs. Acta physiol. scand., 56,201-219. LUNDBERG, A., AND WINSBURY, G., (1960); Functional organization of the dorsal spino-cerebellar tract in the cat. VI. Further experiments on excitation from tendon organ and muscle spindle afferents. Acta physiol. scand., 49, 165-170. MACNI, F., AND OSCARSSON,O., (1961); Cerebral control of transmission to the ventral spinocerebellar tract. Arch. ital. Biol., 99, 369-396. MAGNI,F.. AND WILLIS,w . D., (1963); Identification of reticular formation neurons by intracellular recording. Arch. ital. Biol., 101,681-702. MCINTYRE, A. K., AND MARK,R. F., (1960); Synaptic linkage between afferent fibres of the cat’s hind limb and ascending fibres in the dorsolateral funiculus. J . Physiol. (Lond.), 153,306-330. MORIN,F., (1955); A new spinal pathway for cutaneous impulses. Amer. J, Physiol., 183,245-252. MORIN,F., KITAI,S. T., PORTNOV, H., AND DEMIRIJAN, c . , (1963); Afferent projections to the lateral cervical nucleus : a microelectrode study. Amer. J. Physiol., 204, 667-672. MORIN,F., SCHWARTZ, H. G., AND O’LEARY,J. L., (1951); Experimental study of the spinothalamic and related tracts. Acta psychiat. neurol. scand., 26, 371-396. NAUTA, W. J. H., AND KUYPERS, H. G. J. M., (1958); Some ascending pathways in the brain stem reticular formation. Reticular Formation ofthe Brain. London, Churchill.

ASCENDING SPINAL PATHWAYS

163

NORRSELL, U., AND VOORHOEVE, P., (1962); Tactile pathways from the hindlimb to the cerebral cortex in cat. Acta physiol. scand., 54, 9-17. OSCARSSON, O., (1956); Functional organization of the ventral spinocerebellar tract in the cat. I. Electrophysiological identification of the tract. Acta physiol. scand., 38, 145-1 65. OSCARSSON, O., (1957) ; Functional organization of the ventral spino-cerebellar tract in the cat. 11. Connections with muscle, joint, and skin nerve afferents and effects on adequate stimulation of various receptors. Acta physiul. scand., 42, Suppl. 146. 1-107. OSCARSSON, O., (1958); Further observations on ascending spinal tracts activated from muscle, joint, and skin nerves. Arch. ital. B i d , 96, 199-215. OSCARSSON, O., (1960) ; Functional organization of the ventral spino-cerebellar tract in the cat. 111. Supraspinal control of VSCT units of I-type. Acta physiol. scand., 49, 171-183. POMPEIANO, O., AND BRODAL, A., (1957); Spino-vestibular fibres in the cat. An experimental study. J . comp. Neurol., 108, 353-378. REXED,B., (1954); A cytoarchitectonic atlas of the spinal cord in the cat. J. comp. Neurul., 100, 297-379. ROSSI,G. F., AND BRODAL, A., (1957); Terminal distribution of spinoreticular fibres in the cat. Arch. Neirrol. Psychiat., 78, 439453. RUDIN,D. O., AND EISENMAN, G . , (1951); A method for dissection and electrical study in vitro of mammalian central nervous tissue. Science, 114, 300-302.

164

Differential Course and Organization of Uncrossed and Crossed Long Ascending Spinal Tracts 0. OSCARSSON Institute of Physiology, University of Lund, Lund (Sweden)

Recent investigations on ascending spinal tracts with electrophysiological technique suggest that uncrossed and crossed tracts differ in two respects (Magni and Oscarsson, 1962; Holmqvist and Oscarsson, 1963): (1) uncrossed tracts are located dorsally of the crossed tracts; (2) uncrossed tracts are polysynaptically activated only from ipsilateral nerves and crossed tracts both from contralateral and ipsilateral nerves. Furthermore, some evidence suggests that the cell bodies of uncrossed and crossed tracts occupy different areas in the grey matter. The identification of uncrossed and crossed tracts has been based on the hlstological finding that, in most spinal cord segments, primary afferents terminate exclusively, or almost exclusively, on the ipsilateral side (Schimert, 1939; Escolar, 1948; Liu, 1956; Sprague, 1958). Hence, tracts activated monosynaptically from ipsilateral afferents are uncrossed and tracts activated monosynaptically from contralateral afferents, crossed at the spinal level. METHODS

The results were obtained on electrical stimulation of nerves or dorsal roots. The activity evoked in ascending tracts was recorded either as a mass discharge led from dissected fascicles of the spinal cord or studied by intra-axonal recording from single fibres. Of these two methods the former warrants a more detailed description. The technique of recording from isolated strands of the cord was devised by Rudin and Eisenman (1951) for the study of properties of fibres in the cord. It was developed further and used extensively for studying the functional organization of various ascending tracts by Laporte et al. (1956), Lundberg and Oscarsson (1961), and Holmqvist and Oscarsson (1963). The fascicles are prepared as follows. The spinal cord is transected and one pair of roots severed caudally of the transection. The dorsal funiculi are stripped off for a distance of about 2 cm in caudal direction. The remaining part of the cord is divided into the midline and the ‘cord-halves’ (except dorsal funiculi) split longitudinally into subdivisions of various sizes, here called ‘fascicles’. The subdividing is done by cutting with a pair of fine scissors. In the cat, four or five fascicles can be made on each side

LONG ASCENDING S P I N A L TRACTS

165

without difficulty. The dissected fascicles are mounted on electrodes for monophasic recording of discharges in ascending tracts, one electrode being in contact with the severed end, the other on the dissected part close to its site of separation from the intact cord. The method can be used for determining the exact position of individual tracts with known functional properties: there is remarkably little destruction of ascending fibres during the dissection and the size of the fascicles can be assessed afterwards by histological methods. The main limitation of the method is that potentials due to activity in unmyelinated and thin, myelinated fibres are too small to be detected. The same limitation holds for the other method which has been used: intra-axonal recording from single fibres. Successful impalement occurs very seldom with fibres having a conduction velocity lower than 25-30 mjsec which would correspond to a diameter of 4-5 ,u (assuming that the Hursh factor of 6 is valid in the CNS). RESULTS

Recording from dissected fascicles in mammalian species

The records in Fig. 1 illustrate the differential characteristics of mass discharges evoked in dorsally and ventrally located tracts. Two fascicles were dissected at the upper lumbar level of the spinal cord in a monkey. The transectional areas of the CONTRAL.

IPSIL. MUSCLE

MUSCLE

G

H

I

100

wv

Fig. 1. Discharges in ascending spinal tracts evoked by volleys in muscle and skin afferents of ipsilateral and contralateral hindlimb nerves. Monkey. The muscle (hamstring) and skin (lateral sural) nerves were stimulated at a strength of about 20 times threshold. Records E-L were obtained from the dissected fascicles (i) and (ii) as indicated. The upper and lower traces show the discharge recorded simultaneously at two speeds. The ingoing volleys (A-D) were recorded from the dorsal roots 4.2 cm below the cord dissection at mid-L2. The fast time scale applies to A-D and upper traces in E-L. Voltage scale applies to E-L. (From Oscarsson el al., 1963b). References p . 1751176

166

0. O S C A R S S O N

fascicles are shown in the diagram. The upper and lower traces in Fig. 1, E-L show the discharges recorded at two speeds and led from the fascicles as indicated. The top traces (A-D) show the ingoing primary afferent volleys recorded triphasically from the dorsal roots about 4 cm more caudally. Tracts ascending in the dorsal fascicle were activated by volleys in muscle (E) and skin (F) afferents of ipsilateral nerves, whereas no trace of activity was evoked from contralateral nerves (G, H). The ventral fascicle contained tracts activated from contralateral as well as ipsilateral nerves. The latency of the initial component of the mass discharges evoked from ipsilateral nerves in the dorsal fascicle (E, F) and from contralateral nerves in the ventral fascicle (K, L) was 1.O-1.4 msec when measured relatively to the ingoing volley. This short latency proves that the transmission was monosynaptic. It can be concluded that the dorsal fascicle contains uncrossed tracts and the ventral fascicle, crossed tracts. On the other hand, there was no evidence for monosynaptic excitation from ipsilateral nerves to tracts in the ventral fascicle. The initial part of the discharges evoked from ipsilateral nerves in this fascicle (I, J) was related to group I1 muscle afferents and low threshold cutaneous afferents. The long latency indicates that the transmission was:poly sy naptic. I PSIL.

CONTRAL.

6 MUSCLE

SKIN

MUSCLE

SKIN

c-D-l

HINDLIMB TRACTS

Fig. 2. Discharges recorded at the third cervical segment from tracts activated by stimulation of ipsilateral and contralateral muscle (hamstring) and skin (superficial peroneal) nerves in the hindlimbs. Cat. The records were obtained from fascicles i-iii as indicated. The upper and lower traces show the discharges recorded simultaneously on a fast and slow time base. Middle traces in A-D show ascending volleys recorded from the dissected dorsal funiculi at C3. Time scales in msec. Distance from stimulating electrode on hamstring nerve to C3 about 37 cm. (From Holmqvist and Oscarsson, 1963.)

Similar observations have been made on recording from fascicles dissected in the upper lumbar region of the phalanger, rabbit, cat, dog, and monkey (Magni and Oscarsson, 1962; Holmqvist and Oscarsson, 1963, Oscarsson et al., 1963b). In all these species uncrossed tracts occur in the dorsal half of the lateral funiculus and crossed tracts in the area ventrally thereof. There is little overlap of the areas containing uncrossed and crossed tracts and the boundary between them corresponds approximately to a horizontal line going through the central canal. Recording from fascicles dissected at the cervical level has disclosed that the borderline between uncrossed and crossed tracts varies according to the spinal cord

LONG ASCENDING S P I N A L TRACTS

167

level and the segmental origin of the tracts. In the experiment of Fig. 2, made on the cat, three fascicles (i-iii) were dissected at the level of the third cervical segment. Muscle and skin nerves in the hindlimbs were stimulated. The arrangement of the records corresponds to that in the previous figure, except that the middle traces in A-D show the afferent volleys led from the dorsal funiculi dissected for recording at C3. In the dorsal fascicle (i) ipsilateral but not contralateral nerves evoked monoand polysynaptic discharges (A-D). The intermediate and ventral fascicles (ii and iii) contained tracts which received strong excitatory effects from contralateral nerves and weaker effects from ipsilateral nerves. A large monosynaptic discharge was evoked from contralateral group I afferents in the intermediate fascicle (G). Volleys in skin and high threshold muscle afferents in ipsilateral and contralateral nerves evoked a late activity with a latency suggesting polysynaptic excitation (E-L). These observations show that the borderline between uncrossed and crossed ‘hindlimb tracts’ has shifted dorsally at the cervical level when compared with the lumbar level (Holmqvist and Oscarsson, 1963). IPSIL.. MUSCLE

CONTRAL SKIN

MUSCLE

SKIN

FORELIMB TRACTS

-Eclrr Fig. 3. Discharges recorded at the third cervical segment from tracts activated by stimulation of ipsilateral and contralateral muscle (deep radial) and skin (superficial radial) nerves in the forelimbs. Cat. The records were obtained from the dissected fascicles i-iii as indicated. The upper and lower traces show the discharges on a fast and slow time base. Middle traces in A-D show ascending volleys recorded from the dissected dorsal funiculi at C3. Time scales in msec. Distances: stimulating electrodes on the nerves - C7 dorsal root entrance, 11.5 cm; C7 dorsal root entrance - recording p!ace, 4.5 cm. (From Holmqvist et al., 1963.)

The records in Fig. 3 were obtained in the same experiment but on stimulation of forelimb nerves. Tracts in the dorsal and intermediate fascicles were activated monosynaptically only from ipsilateral nerves. Stimulation of contralateral nerves produced no discharge in the dorsal fascicle and only a small discharge in the intermediate fascicle. Presumably this discharge was partly, at least, due to inclusion of ventrally located, crossed tracts. The ventral fascicle contained tracts which were monosynaptically activated from contralateral nerves. Stimulation of contralateral nerves evoked large, and stimulation of ipsilateral n:rves small polysynaptic discharges in this fascicle (Holmqvist et al., 1963). Rderences p . 1751176

168

0. O S C A R S S O N

The results of the experiment illustrated in Figs. 2 and 3, together with other experiments, indicate that uncrossed tracts, at the cervical level, occupy approximately the dorsal third of the lateral funiculus when originating from the lumbar intumescence and approximately the dorsal two thirds when originating from the cervical intumescence. Tracts activated f r o m afferents in sacral and caudal roots Afferents belonging to sacral roots differ from afferents in most other segments by terminating not only ipsilaterally but also contralaterally in the grey matter. This has been shown histologically by Sprague (1 958) and, correspondingly, motoneurones in the sacral cord have been observed to receive monosynaptic excitation from contralateral afferents (Curtis et al., 1958; Frank and Sprague, 1959). IPSIL. L7

-A -+

CONTRAL. L7 33

93 C

L

D

MSEC

1

M

-

eosc

N

Y

Fig. 4. Discharges recorded at the first lumbar segment from tracts activated by stimulation of ipsilateral and contralateral L7 and S3 dorsal roots. Cat. The records were obtained from fascicles i-iv as indicated. The ingoing volley was recorded triphasically from the dorsal funiculus at the L7 level (upper traces in A-D). The pairs of traces show the discharges recorded simultaneously on a fast and slow time base. The distance between the two recording sites was 6.5 cm. (From Holmqvist and Oscarsson, 1963.)

The discharges evoked in ascending spinal tracts by stimulation of lumbar, sacral, and caudal roots have been investigated in the cat (Holmqvist and Oscarsson, 1963). Fig. 4 shows an experiment in which four fascicles were dissected at the upper lumbar level and the L7 and S3 dorsal roots prepared for stimulation. A volley in the L7 root evoked discharges with a pattern conforming to that produced by stimulation of hindlimb nerves. Large discharges were evoked by ipsilateral volleys in the two dorsal fascicles (A, E), whereas contralateral volleys were largely ineffective (C, G). In the two ventral fascicles monosynaptic discharges were evoked from contralateral nerves (K, 0).The small monosynaptic discharge (I) evoked from the ipsilateral root in fascicle (iii) was due to some uncrossed fibres included in this fascicle. Similar

LONG ASCENDING SPINAL TRACTS

169

observations were made on stimulation of other lumbar roots and of the sacral roots down to, and sometimes including S2. On the other hand, stimulation of the lower sacral and the caudal roots produced, in the dorsal fascicles, small contralateral discharges with distinct monosynaptic components. This is shown for the S3 dorsal root in Fig. 4, D and H. It is reasonable to explain these discharges as due to activation of uncrossed tracts from contralaterally terminating primary afferents. Recording from dissected fascicles in birds and amphibians

The organization of ascending tracts described above seems to apply generally to the mammalian cord (Magni and Oscarsson, 1962). Experiments made on birds and amphibians suggest that a similar organization exists in all higher vertebrates. The records in Fig. 5 were obtained from three fascicles dissected at the cervical IPSIL. SCIATIC

RADIAL

CONTRAL. SCIATIC

RADIAL

Fig. 5. Discharges evoked in tracts ascending in the fascicles (i-iii) indicated in the diagram on stimulation of ipsilateral and contralateral sciatic and radial nerves. Duck, mid-cervical level. Upper and lower traces were taken simultaneously at different speeds. Time scales in msec. Distances: stimulating electrode on sciatic nerve - spinal cord, 7 cm; spinal cord - recording site, 29 cm; stimulating electrode on radial nerve - spinal cord, 6.5 cm; spinal cord - recording site, 16 cm. (From Oscarsson et at., 1963a.)

level in a duck. Leg (sciatic) and wing (radial) nerves were stimulated as indicated. In the dorsomedial fascicle (i) stimulation of ipsilateral nerves evoked large discharges, whereas stimulation of contralateral nerves produced no trace of activity. The discharge evoked from the wing nerve had a distinct monosynaptic component. In the dorsolateral fascicle (ii) large mono- and polysynaptic discharges were evoked from contralateral nerves. Ipsilateral nerves evoked polysynaptic activity and a trace of a monosynaptic response from the radial nerve. In the large ventral fascicle (iii) only small discharges were observed. However, in other experiments with recording from the ventral quadrant of the cord distinct monosynaptic responses were evoked from contralateral, but not from ipsilateral nerves. The results indicate a similar References p . 17511 76

170

0. O S C A R S S O N

organization as in mammals but the uncrossed tracts occupy only a small dorsal part of the lateral funiculus (Oscarsson et a/., 1963a). The records shown in Fig. 6 were obtained from the thoracic cord of the frog. Records A and B were obtained from the dissected 'cord-half' (except dorsal funiculus) and show that stimulation of ipsilateral as wcll as contralateral nerves evoked disIPSILda

C 0NTR A L.

IPSIL.

CONTRAL.

..

msec

mmwnmnm

msec

Fig. 6 . Distribution of ipsilateral and contralateral discharges in the lateral and ventral funiculi. Frog. (A-D) Discharges recorded from the dissected cord-half (except dorsal funiculus, see diagram) at mid-thoracic level on stimulation of ipsilateral (A, C) and contralateral (B, D) sciatic nerve a t 18 times threshold. Upper and lower traces show the ascending discharge a t different speeds, middle traces show the incoming volley recorded, at the fast speed, from the dorsal roots 15 mm more caudally. A and B were obtained before and C and D, after the lesion shown in the diagram (hatched). The voltage scale applies to the lower traces. E-H illustrate a different experiment. The records show discharges recorded from the dissected ventral quadrant of the cord (see diagram) at the upper thoracic level on stimulation of the ipsilateral and contralateral sciatic nerve a t 20 times threshold. A and B were obtained beforc and G and H after the lesion shown in the diagram (hatched). Conventions as in A-D. (From Oscarsson and Rosen, 1963.)

charges. These discharges were initiated by monosynaptic components. After a lesion that destroyed the ventral quadrant of the cord, only stimulation of ipsilateral nerves evoked a discharge ( C , D). Records E-H are from a different experiment. The ventral quadrant was dissected for recording. Volleys in ipsilateral and contralateral nerves evoked mono- and polysynaptic discharges. Following the lateral lesion indicated in the diagram, only contralateral nerves were effective (G, H). These and other experiments suggest that uncrossed tracts in the frog spinal cord occupy the whole lateral funiculus and crossed tracts approximately the ventral quadrant. This organization is essentially the same as that in mammals and birds, but the uncrossed and crossed tracts occupy largely overlapping areas (Oscarsson and Rosin, 1963). Unit discharge in ascending tracts

Ascending spinal tracts activated from hindlimb afferents have been investigated extensively with microelectrode recording from single fibres in the lateral and ventral

LONG ASCENDING SPINAL TRACTS

171

funiculi of the cat. The observations made during these investigations confirm those made on mass discharge recording and give some additional information. A . Distribution of monosynaptic excitation Readily observable monosynaptic activation occurs only in some ascending tracts and it is possible that monosynaptic excitation from primary afferents is lacking in other tracts (cf. Lundberg and Oscarsson, 1960, 1961, 1962a, b). Among monosynaptically activated tracts the dorsal and ventral spino-cerebellar tracts (DSCT and VSCT) are especially well known. Units belonging to the DSCT and VSCT can be distinguished by their connections with primary afferents and by their mode of termination in the cerebellar cortex (Lundberg and Oscarsson, 1962a). Though DSCT and VSCT fibres largely occupy separate areas in the dorsal and ventral part of the white matter, there is some intermingling in a border zone and a few fibres of either tract may be found displaced deep into the area occupied by the other tract (Oscarsson, 1957; Lundberg and Oscarsson, 1962a). Hence the location of the fibres in the white matter is unsuitable as a criterion for distinguishing DSCT and VSCT axons. Several hundreds of DCST units identified by their connections with primary afferents and by their mode of termination in the cerebellum have been studied. Among these units only two were monosynaptically activated from contralateral instead of ipsilateral nerves (Lundberg and Oscarsson, 1960).Comparable observations have been made with VSCT units (Lundberg and Oscarsson, 1962a). For example, in one investigation 2 out of 61 units were monosynaptically activated from ipsilateral instead of contralateral nerves. These observations indicate that the vast majority of the DSCT neurones have uncrossed axons and the vast majority of VSCT neurones, crossed axons. However, exceptionally a DSCT axon may cross to the other side of the cord, or a VSCT axon ascend without crossing. Presumably the latter cases should be regarded as aberrancies with little functional significance. They give, however, information about the effectiveness of the mechanisms that during the development guide the growth of axons along certain paths. The information concerning units in other tracts is less detailed. However, the spino-cervical tract which ascends in the lateral funiculus dorsally of the DSCT and terminates i n the lateral cervical nucleus, is monosynaptically activated by ipsilateral but not contralateral cutaneous afferents as shown b3th on mass discharge and unit recording (Lundbxg and Oscarsson, 1961 ; Holmqvist and Oscarsson, 1963). Very recently a spino-cerebellar tract (RSCT) activated from group I afferents in ipsilateral forelimb nerves has been discovered (Holmqvist et al., 1963; Oscarsson and Uddenberg, unpublished). There is no trace of monosynaptic mass discharge on stimulation of group I afferents in contralateral nerves. Tentatively, it might be hypothesized that individual tracts consist of either uncrossed or crossed units but not of both.

B. Distribution of polysynaptic excitation Units bslonging to dorsally located tracts are, as a rule, polysynaptically activated References p. 17511 76

172

0. O S C A R S S O N

only from ipsilateral nerves. No exceptions have been noted with units in the DSCT or spino-cervical tract but some fibres belonging to a tract with unknown termination sometimes discharge a few impulses 10-30 msec following stimulation of contralateral nerves (Lundberg and Oscarsson, 1960, 1961). This late discharge is presumably due to weak excitation exerted through long chains of interneurones. Recording from units in ventrally located tracts has presented a more varied picture. Three groups of ascending axons with different functional characteristics have been recognized (Lundborg and Oscarsson, 1962a, b). Units belonging to all three groups receive strong polysynaptic excitation or inhibition from contralateral afferents which may be connected with the assumed contralateral location of the cell bodies. Polysynaptic activation from ipsilateral nerves has been observed in all the groups. In the VSCT excitation and inhibition from ipsilateral nerves is weaker than from contralateral nerves (Oscarsson, 1957; Lundberg and Oscarsson, 1962a). Of the other two pathways, provisionally denoted the bilateral and the contralateral ventral flexor reflex tracts (bVFRT and cVFRT), the former receives equally strong excitation from ipsilateral and contralateral nerves, whereas units belonging to the latter tract are either weakly or not at all activated from ipsilateral nerves (Lundberg and Oscarsson, 1962b). DISCUSSION

The main findings described in this paper can be summarized as follows: (1) Tracts in the dorsal part of the ventrolateral white matter are mono- and polysynaptically activated only from ipsilateral nerves. (2) Tracts in the ventral part of the ventrolateral white matter are monosynaptically activated only from contralateral nerves and polysynaptically, both from ipsilateral and contralateral nerves. In most spinal segments primary afferents terminate almost exclusively on the ipsilateral side. This has been shown in histological investigations on the mammalian cord (Schimert, 1939; Escolar, 1948; Liu, 1956; Sprague, 1958) and recently also in investigations on the amphibian cord (W. W. Chambers and C.-N. Liu, personal communication). Hence the findings described under (1) and (2) suggest that dorsal tracts originate from ipsilateral cell bodies and ventral tracts, from contralateral cell bodies, i.e. they are uncrossed and crossed tracts respectively. Similar findings have been made in mammalian, avian, and amphibian species suggesting a basically similar organization in all higher vertebrates. The spinal cord sectors containing uncrossed and crossed tracts vary at different levels of the cord and in different groups of animals, as is illustrated in Fig. 7. The vertically hatched areas contain uncrossed tracts and the horizontally hatched areas, crossed tracts. There is little overlapping of the areas containing uncrossed and crossed tracts arising from the same segmental level in mammals and birds. Some overlapping at the lumbar level has been observed in the cat (Holmqvist and Oscarsson, 1963) and may also occur at other levels, but this overlapping is much smaller than that found in the frog. The uncrossed and crossed tracts are distinguished not only by their differential location in the white matter but also by their polysynaptic connections with primary

LONG ASCENDING SPINAL TRACTS

173

MAMMAL FORELIMB

HINDLIMB

BIRD LEG

FROG WING

HINDLIMB

Fig. 7. Spinal cord sectors containing uncrossed (vertical hatching) and crossed (horizontal hatching) ascending tracts at indicated segmental levels. The various diagrams refer to tracts activated from hindlimb and forelimb nerves as indicated.

afferents: the uncrossed tracts are polysynaptically activated only from ipsilateral afferentr and the crossed tracts both from contralateral and ipsilateral afferents. This might suggest that the cell bodies of uncrossed and crossed tracts are located in different regions of the grey matter. Tlus assumption receives some support from the location of the cell columns that have so far been identified as the origin of some individual tracts. The cell columns of uncrossed tracts are vertically hatched and those of crossed tracts horizontally hatched in Fig. 8.4. The dorsal spino-cerebellar tract (DSCT) originates from cells in Clarke’s column (e.g. Jansen and Brodal, 1958) and the spino-cervical tract (SCT) from cells in the head of the dorsal horn (Eccles et ~7/., 1960; Wall, 1960; Lundberg and Oscarsson, 1961). The cell bodies of the ventral spino-cerebellar tract (VSCT) occur in the lateral part of the intermediate zone and the lateral parts of the base and neck of the dorsal horn (Hubbard and Oscarsson, 1962). These cells are presumably distinct from the border cells of Cooper and Sherrington (1940) which occur in ‘the ventrolateral fringe of the spinal grey matter’ (cf. Hubbard and Oscarsson, 1962). Axons of the spinal border cells ascend in the contralateral ventral quadrant of the cord and their function is unknown (cf., however, Sprague, 1953). Our observations suggest that ascending spinal tracts are organized as shown schematically in Fig. 8B. The figure refers to the lumbar region of the mammalian cord but the organization would be essentially the same at other levels of the cord and in other classes of higher vertebrates. The uncrossed tracts ascend in the dorsal part of the lateral funiculus and the crossed tracts ventrally thereof. The cell bodies of uncrossed tracts occur in the dorsomedial part of the grey matter and those of crossed tracts in the ventrolateral part: the borderline is tentatively drawn as suggested References p . 1751176

174

0. O S C A R S S O N

Fig. 8. (A). Location of cell columns giving rise to uncrossed (vertical hatching) and crossed (horizontal hatching) ascending tracts. To the former tracts belong the spino-cervical tract (SCT) and the dorsal spino-cerebellar tract (DSCT) and to the latter, the ventral spino-cerebellar tract (VSCT) and the tract of unknown function and termination originating from the spinal border cells (BC) of Cooper and Sherrington (1940). (B). Organization of uncrossed and crossed tracts according to the hypothesis outlined in the text. Tracts ascending in the dorsal part of the lateral funiculus originate from ipsilateral cells in a dorsomedial region of the grey matter. These cells receive terminals from ipsilateral primary afferents and ipsilateral interneurones. Tracts ascending in the ventral part of the cord originate from contralateral cells in a ventrolateral region of the grey matter. These cells receive terminals from ipsilateral primary afferents and have connections with ipsilateral as well as contralateral interneurones. The borderline between the two regions in the grey matter is tentatively drawn (broken line) as suggested by the cell collumns shown in A. (Modified from Magni and Oscarsson, 1962.)

by the location of the cell columns shown in Fig. 8A. Polysynaptic paths from primary afferents to tract cells are drawn as disynaptic. Only the cells of crossed tracts are innervated by interneurones conveying excitation from contralateral afferents. There are no observations in our experiments that contradict the hypothesis illustrated in Fig. 8B. However, some limitations of the evidence should be noted. 1. The recording methods select pathways containing relatively coarse fibres. It seems, however, unlikely that long, thin-fibred pathways would have a different organization, the more so as the tracts investigated constitute a variety of pathways with diverse function and termination. 2. Our identification of uncrossed and crossed tracts depends on the demonstration of monosynaptic connections with primary afferents. Some tracts do not receive any appreciable monosynaptic excitation and can not be identified as crossed or uncrossed with our method. However, in all except one case, these tracts receive polysynaptic excitation predominantly or exclusively from primary afferents entering the cord ipsilaterally of the assumed cell bodies. The exceptional tract receives equally strong excitation from ipsilateral and contralateral afferents (Oscarsson, 1958 ; Lundberg and Oscarsson, 1962b). It is, at present, impossible to say if this ventrally located tract has its cell bodies on the contralateral side in conformity with the present hypothesis. 3. The observation that crossed tracts have a bilateral receptive field might have exceptions. Clinical observations on chordotomy cases suggest that the spino-thalamic

LONG ASCENDING SPINAL TRACTS

175

tract has a purely contralateral receptive field (e.g. Hyndman and Wolkin, 1943; Stookey, 1943). Fibres of this tract might have been missed in our experiments on the cat because of their small size or there might be species differences. 4. The suggestion that the cell bodies of uncrossed and crossed tracts occupy different regions of the grey matter obviously needs anatomical confirmation. SUMMARY

Recent investigations have disclosed that the long ascending spinal tracts in mammals, birds, and presumably also amphibians are organized as follows: 1. Tracts in the dorsal part of the ventrolateral white matter are mono- and polysynaptically activated only from ipsilateral nerves. 2. Tracts in the ventral part of the ventrolateral whte matter are monosynaptically activated only from contralateral nerves and polysynaptically, both from ipsilateral and contralateral nerves. These observations and histological evidence that primary afferents, in most spinal segments, terminate almost exclusively on the ipsilateral side show that the former tracts are uncrossed and the latter crossed at the spinal level. The boundary between uncrossed and crossed tracts is usually sharp but its position varies at different levels of the cord (Fig. 7). It is suggested that the differential organization of uncrossed and crossed tracts is related to a differential location of the cell bodies in the grey matter of the cord. The uncrossed tracts are assumed to originate from cells in the dorsomedial, and the crossed tracts from cells in the ventrolateral part of the grey matter (Fig. 8A, B). Only the cells of the crossed tracts are innervated by interneurones conveying excitation from contralateral afferents (Fig. 8B). REFERENCES COOPER,S., AND SHERRINGTON, CH. S., (1940); Gower’s tract and spinal border cells. Brain, 63, 123-134. CURTIS,D. R., KRNJEVIC, K., AND MILEDI,R., (1958); Crossed inhibition of sacral motoneurones. J . Neurophysiol., 21, 3 19-326. ECCLES,J. C., ECCLES,R. M., AND LUNDBERG, A., (1960); Types of neurones in and around the intermediate nucleus of the lumbo-sacral cord. J . Physiol. (Lond.), 154, 89-1 14. ESCOLAR, J., (1948); The afferent connections of the lst, 2nd, and 3rd cervical nerves in the cat. J . comp. Neurol., 89, 79-92. FRANK, K., AND SPRAGUE, J. M., (1959); Direct contralateral inhibition in the lower sacral spinal cord. Exp. Neurol., 1, 2843. HOLMQVIST, B., AND OSCARSSON, O., (1 963); Location, course, and characteristics of uncrossed and crossed ascending spinal tracts in the cat. Acta physiol. scand., 58, 57-67. HOLMQVIST, B., OSCARSSON, O., AND UDDENBERG, N., (1963); Organization of ascending spinal tracts activated from forelimb afferents in the cat. Acta physiol. scand., 58, 68-76. HUBBARD, J. I., AND OSCARSSON, O., (1962); Localization of the cell bodies of the ventral spinocerebellar tract in lumbar segments of the cat. J . cornp. Neurol., 118, 199-204. HYNDMAN, 0. R., AND WOLKIN,J., (1943); Anterior chordotomy; further observations on physiologic results and optimum manner of performance. Arch. Neurol. Psychiat. (Chic.), 50, 129-148. JANSEN, J., UND BRODAL, A., (1958); Handbuch der mikroskopischen Anatomie des Menschen. IVj8. Das Kleinhirn. Berlin, Springer-Verlag.

176

0. O S C A R S S O N

LAPORTE, Y . , LUNDBERG, A., AND OSCARSSON, O., (1956); Functional organization of the dorsal spino-cerebellar tract in the cat. I. Recording of mass discharge in dissected Flechsig’s fasciculus. Acta physiol. scand., 36, 115-187. Lru, C.-N., (1956); Afferent nerves to Clarke’s column and the lateral cuneate nuclei in the cat. Arch. Neurol. Psychiat. (Chic.), 15, 61-17. LUNDBERG, A., AND OSCARSSON, O., (1960); Functional organization of the dorsal spino-cerebellar tract in the cat. VII. Identification of units by antidromic activation from the cerebellar cortex with recognition of five functional subdivisions. Acta physiol. scand., 50, 356-374. LUNDBERG, A., AND OSCARSSON, O., (1961); Three ascending spinal pathways in the dorsal part of the lateral funiculus. Acta physiol. scand., 51, 1-16. LUNDBERG, A., AND OSCARSSON, O., (1962a); Functional organization of the ventral spino-cerebellar tract in the cat. IV. Identification of units by antidromic activation from the cerebellar cortex. Acta physiol. scand., 51, 252-269. LUNDBERG, A., AND OSCARSSON, O., (1962b); Two ascending spinal pathways in the ventral part of the cord. Acta physiol. scand., 51,270-286. MACNI,F., AND OSCARSSON, O., (1962); Principal organization of coarse-fibred ascending spinal tracts in phalanger, rabbit, and cat. Acta physiol. scand., 51, 53-64. OSCARSSON, O., (1957); Functional organization of the ventral spino-cerebellar tract in the cat. 11. Connections with muscle, joint, and skin nerve afferents and effects on adequate stimulation of various receptors. Acta physiol. scanrl., 42, Suppl. 146, 1-107. OSCARSSON, O., (1958) ; Further observations on ascending spinal tracts activated from muscle, joint, and skin nerves. Arch. iral. Biol., 96, 199-215. OSCARSSON, O., AND ROSEN,I., (1963); Organization of ascending tracts in the spinal cord of the frog. Acta physiol. scand. 59, 154-160. OSCARSSON, O., R O S ~ NI., , AND UDDENBERG, N., (1963a); Organization of ascending tracts in the spinal cord of the duck. Acta physiol. scand., 59, 143-153. OSCARSSON, O., ROSEN,I., AND UDDENBERG, N., (1963b); A comparative study of ascending spinal tracts activated from hindlimb afferents in monkey and dog. Arch. ital. Bid., in press. RUDIN,D. O., AND EISENMAN, G . , (1951); A method for dissection and electrical study in vitro of mammalian central nervous tissue. Science, 114, 300-302. SCHIMERT, J., (1939); Das Verhalten der Hinterwurzelkollateralen im Riickenmark. Z. Anat. Entwickl. Cesch., 109, 665-687. SPRAGUE, J. M., (1953); Spinal ‘border cells’ and their role in postural mechanism. J . Neurophysiol., 16, 464474. SPRAGUE, J. M., (1958); The distribution of dorsal root fibres on motor cells in the lumbosacral spinal cord of the cat, and the site of excitatory and inhibitory terminals in monosynaptic pathways. Proc. roy. SOC.B, 149, 534-556. STOOKEY, B., (1943); The management of intractable pain by chordotomy. Ass. Res. nerv. Dis. Proc., 23, 416433. WALL,P. D., (1960); Cord cells responding to touch, damage, and temperature o f skin. J . Neurophy~iol.,23, 197-210.

DISCUSSION

NIEUWENHUYS: In the terminology of Cajal your findings can be summarized 1 think as follows: Funicular cells are located in the dorsal part of the gray matter, whereas the commissural occupy the ventral part of the gray matter. I think this thesis fits in well with the anatomical evidence now available. However, there is one exception. In all vertebrates there has been described a spino-bulbar, spino-mesencephalic, resp. spino-thalamic tract, originating from cells in the dorsal horn and constituting the so-called ventral arcuate system (‘Bogenfasern’ of His). I think there are two possibilities: (a) This system consists of thin fibers, and (b) this system does not constitute a long ascending tract, or even: it may not exist at all in the adult stage.

LONG ASCENDING S P I N A L TRACTS

177

SZENTAGOTHAI : The only explanation from anatomical viewpoint, that I could think of, is that there is no such thing as the classical concept of a spino-thalamic tract arising from the dorsal horn and immediately crossing in the anterior white commissure. One can cut away horizontally the whole dorsal horn at the level of the dorsal commissure, or place small lesions into the dorsal horn, without getting the slightest signs of degeneration in the same or the next upper segments of fibers within the anterior commissure. Whenever the lesion reaches the intermediate gray matter, i.e. the lamina V1 of Rexed, one immediately gets degenerated fibers in the anterior white commissure. This fits exactly with the location by Oscarsson of cells giving rise to VSCT fibers. After lesions placed into laminae VI and VII we were able to trace degenerated fibers in the cerebellum in the projection area of the VSCT terminating as mossy fibers. Similar lesions give also rise to terminal degeneration of few fibers in the ventrolateral basal nucleus of the contralateral thalamus. Thus there are real spino-thalamic fibers coming from lower lumbar or upper sacral segments, but they arise not from the dorsal horn, but from the intermediate region and central parts of the ventral horn. KUYPERS:I feel that in regard t o the spinal cord and the ascending pathways your paper has been a revolutionary one. The problem of the descending influences on ascending conduction is now more easy to understand. It has always been claimed that the reticular formation has an important influence upon the ascending conduction. Assuming that the ascending fibers came primarily from the dorsal horn, I could not find anatomically any reasonable pathway which would be able to influence the ascending conduction. However, we made the restriction that if this descending influence upon the ascending conduction was exerted, it had to be exerted first and foremost by cells located in the medial part of the intermediate zone and a part of the ventral horn, for this was the place where the prime determination of descending fibers from the reticular formation took place. It is now most gratifying to see that instead of having the origin of the crossed ascending pathways in the posterior horn, you are placing it precisely in the area which is so open to influences of the reticular formation. I think this has cleared the issue, at least for my feeling, considerably. SPRAGUE: I agree with Dr. Oscarsson that the cells giving rise to the ventral spinocerebellar tract are indeed different from the border cells, i n contrast to the original supposition of Cooper and Sherrington. Just what the border cells give rise to, what sort of a tract, is not clear to me although I could make some comments that might be instructive. Firstly: in Nauta-preparations large neurons in the same area as you have placed the origin of the crossed ventral spino-cerebellar tract receive a rich dorsal root input which would accord with the monosynaptic activation of that pathway. However these particular cells are receiving the input from the ipsilateral dorsal root which does not accord so well. Is that correct?

OSCARSSON : This is exactly in accordance with our observations: primary afferents

178

DISCUSSION

make synaptic contacts only with ipsilateral cell bodies. It was in relation to the crossed VSCT fibers that the monosynaptic excitation was described as contralateral. SPRAGUE:The second point is that cells, lying in the area of the border cells in the cat do not receive any appreciable dorsal root input. They receive input from the reticulospinal pathways, as Dr. Kuypers has described, but this area receives very little dorsal root input.

OSCARSSON: We don’t know which tract these cells give origin to. However, some ascending tracts receive excitation from primary afferents mainly or exclusively through segmental interneurones, as for example the bVFRT and possibly the cVFRT described by Lundberg and Oscarsson (Actaphysiol. scand., 54 (1962) 270). The transmission to these tracts is strongly influenced by descending systems originating in the medulla oblongata.

179

Three Ascending Tracts Activated from Group I Afferents in Forelimb Nerves of the Cat 0. O S C A R S S O N Institute of Physiology, University of Lund, Lund (Sweden)

The ascending projections of group I afferents from hindlimb muscles are well established. These afferents ascend in the dorsal funiculus for several segments but do not continue above the lower thoracic levels (Lloyd and McIntyre, 1950). The group I afferents make synaptic contacts with the cell bodies of the dorsal and ventral spino-cerebellar tracts (DSCT and VSCT) (Lloyd and McIntyre, 1950; Oscarsson, 1957a). The DSCT is uncrossed, ascends in the dorsal part of the lateral funiculus, and reaches the cerebAluni through the restiform body. It contains components activated monosynaptically from Ia and Ib muscle afferents (Lundberg and Oscarsson, 1956, 1960). The VSCT is crossed, ascends ventrally of the DSCT, and reaches the cerebellum through the brachium conjunctivuin. It receives monosynaptic excitation exclusively, or almost exclusively, from Ib afferents (Oscarsson, 1957b; Eccles et al., 1961a). There is no evidence that hindlimb group I afferents activate other ascending tracts or that collaterals of the DSCT and VSCT activate other structures than the cerebellum. Experiments made during the last two years have revealed that group I afferents in forelimb nerves activate three ascending tracts. One tract originates from cell bodies located at, or slightly above, the level of the dorsal root entrance, ascends in the middle third of the lateral funiculus, and terminates in the cerebellum. This tract i? anatomically and functionally distinct from the DSCT and VSCT and will be denoted the rostra1 spino-cerebellar tract (RSCT). The other two pathways are activated from group I afferents ascending in the dorsal funiculi to the cuneate nuclei (cf. Rexed and Strom, 1952). One of them originates from cells in the external cuneate nucleus and reaches the cerebellum through the ipsilateral restiform body as a component of the cuneo-cerebellar tract. The other pathway originates from the main cuneate nucleus and projects, via the crossed medial lemniscus, to the postcruciate gyrus of the cerebral cortex. THE ROSTRAL SPINO-CEREBELLAR TRACT

Ascending spinal tracts activated from forelimb nerves were recently investigated by recording from fascicles of the cord dissected at the third cervical segment. It was References p . 193-195

180

0. O S C A R S S O N

shown that group I afferents in these nerves activate only one ascending tract. Some anatomical and functional characteristics of this tract were described and it was concluded that it is distinct from the spino-cerebellar tracts activated from hindlimb afferents (Holmqvist et al., 1963b). HINDLIMB IPSIL.

CONTRAL.

FORELIMB IPSIL.

CONTRAL. J

Fig. 1. Discharges recorded at the third cervical segment from ascending spinal tracts on stimulation of ipsilateral and contralateral muscle nerves in the hindlimb (hamstring) and forelimb (deep radial). The stimulus strength was about 20 times the nerve threshold. The records were obtained from fascicles i-iii as indicated. The upper and lower traces show the discharges recorded simultaneously on a fast and slow time base. Middle traces in A-D show ascending volleys recorded from the dissected dorsal funiculi at C3. Time scales in msec. (Modified from Holmqvist and Oscarsson, 1963, and Holmqvist ef al., 1963b.)

The records in Fig. 1 show mass discharges in ascending tracts led from fascicles dissected as indicated on the diagram. Muscle nerves in the hind- and forelimbs were stimulated. The monosynaptic discharges evoked by impulses in group I afferents of hindlimb nerves are readily recognized. The DSCT discharge (A) is recorded from the ipsilateral, dorsal fascicle and the VSCT discharge (E) from the contralateral, intermediate fascicle. Only one tract is activated from group I afferents in forelimb nerves. This tract is ipsilateral and ascends in the intermediate fascicle. The discharge (H) is

Fig. 2. Relation between size of afferent volley and discharge evoked in the group 1 activated forelimb tract. Upper and lower traces show, at two speeds, the discharge evokcd by stimulation of the deep radial nerve at indicated strengths (multiples of nerve threshold). Middle traces show the ingoing volley recorded from the dissected dorsal funiculi at C3. From the experiment also illustrated in Fig. I . (Modified from Holmqvist et al., 1963b).

G R O U P I AFFERENTS IN FORELIMB NERVES

181

comparable in size with the VSCT discharge and has a monosynaptic latency. The small early discharges recorded from the ipsilateral, dorsal fascicle (G) and from the contralateral, ventral fascicle (L) appeared only when the stimulus strength was raised to activate group I1 afferents. %

I

' C4

I C5

'

C 6 I C 7 IC8ITI

Fig. 3. (A). Location of the group I activated forelimb tract at the level of the third cervical segment. The spinal cord sector containing this tract is indicated by vertical hatching. For comparison, the sectors are shown which contain the dorsal (horizontal hatching) and ventral (stippled) spino-cerebellar tracts. (B). Two experiments made to determine the segmental level of the cell bodies of the group I activated forelimb tract. In one experiment the dorsal funiculi were transected at successively more caudal levels (open circles), in the other the lateral funiculus was transected at successively more rostral levels (triangles), while the mass discharge was recorded from the dissected lateral funiculus at C3 on stimulation of the ipsilateral deep radial nerve. Ordinate: amplitude of monosynaptic discharge in per cent of control value. Abscissa: segmental level of transection after which the mass discharge was recorded and measured. The fourth cervical to second thoracic segments are indicated on the horizontal scale. See text. (Modified from Holmqvist et al., 1963b.)

The relation between the size of the ingoing volley and the mass discharge evoked in the forelimb tract is shown in Fig. 2. The discharge appeared at a strength of 1.3 (A, B), and grew to a maximum at 1.9 times threshold (D). I n other experiments the threshold varied between 1.2 and 1.4 and the maximum was reached at, or slightly below, maximum for the group I volley. The threshold for evoking the discharge is similar to the threshold of Ib (tendon organ) afferents in hindlimb nerves. The forelimb tract is presumably activated mainly or exclusively from Ib afferents, just as the VSCT. By recording from variously dissected fascicles it was shown that the group I activated forelimb tract ascends jn the middle third of the lateral funiculus at the C3 level. Its location relative to the DSCT and VSCT is shown in Fig. 3A. The forelimb tract (vertical hatching) is ventral of, but overlaps partly, the VSCT (stippled). The level of the synaptic relay in the cord was determined by transection of the dorsal funiculus (interruption of presynaptic fibres) at successively more caudal levels (circles) and by transection of the lateral funiculus (interruption of postsynaptic axons) at successively more rostral levels (triangles), while the reduction of the mass discharge was watched by recording from the lateral funiculus (Fig. 3B). The experiments show that the relay occurs at, or slightly above, the level of the dorsal root References p. 193-195

182

0. O S C A R S S O N

entrance. These data show that the forelimb tract is anatomically distinct from DSCT and VSCT: it differs from DSCT in arising from cell bodies located rostrally of Clarke’s column and in having a ventral position in the cord, and from VSCT in being uncrossed. On the other hand, the group I activated forelimb tract resembles the VSCT in its termination and functional organization, as has been shown on recording from single units (Oscarsson and Uddenberg, unpublished). In the third cervical segment intraaxonal recording was made from fibres ascending in the lateral funiculus. Units that could be discharged from any of the dissected forelimb nerves were tested for antidromic activation from the cerebellar cortex, as had previously been done with units in the DSCT and VSCT (Lundberg and Oscarsson, 1960, 1962). The great majority of the units that were monosynaptically activated from group I afferents in ipsilateral forelimb nerves could be antidromically activated from the cerebellum. The group I activated tract was denoted the rostral spino-cerebellar tract (RSCT) as it terminates in the cerebellum and originates from cell bodies in the rostral part of the cord. A typical RSCT unit is shown in Fig. 4. In records A-F and I the deep radial (DR) nerve was stimulated at increasing strengths. A spike appeared irregularly at 1.3 A

D

Y

DR

1.3

2.7 h

Fig. 4. Recording from RSCT axon (lower traces) and recording from the surface of the dorsal funiculi in C3 (upper traces). The stimulating and recording arrangements are shown in diagram P. The following nerves were stimulated: ipsilateral deep radial nerve (DR), ipsilateral nerve to long head of triceps (LHT), ipsilateral nerve to biceps (B), ipsilateral median nerve (M), ipsilateral superficial radial nerve (SR), and contralateral radial nerve (r. R). A-F were obtained at the stimulus strengths indicated in multiples of the nerve threshold and G-L at about 15 times threshold. M and N show antidromic spikes elicited from the cerebellar cortex on stimulation of the points indicated in the diapram of the explored part of the anterior cerebellar lobe (0)(cf. Fig. 6). Larsell’s lobules IV and V are indicated. The interrupted line separates hind- and forelimb areas. Records M and N were obtained at faster speed than A-L. (From unpublished observations of Oscarsson and Udden berg.)

G R O U P I AFFERENTS I N FORELIMB NERVES

I83

times threshold when the ingoing volley (upper traces) was about one third maximal. The spike came regularly at higher strengths and a second spike was evoked by impulses in group I1 afferents when the strength was raised to 2.7 times threshold. Further increase of stimulus strength produced further spikes. Stimulation of the ipsilateral biceps nerve (B) was ineffective as was stimulation of the contralateral radial nerve (r. R). However, late spikes appeared with stimulation of low threshold afferents in the superficial radial nerve (SR) and of high threshold afferents in the nerve to the long head of triceps (LHT) and the median nerve (M). Records M and N show antidromic spikes evoked by weak electrical stimulation of the points indicated in the diagram of the anterior lobe (0). Similar observations were made with the other group I activated units. A monosynaptically evoked spike appeared on stimulation of high threshold group I afferents in one or more nerves. In similarity with the VSCT there was very often convergence of group I excitation from muscle groups working at d.ifferent joints: one group of RSCT units was activated both from extensors of the hand and extensors of the forearm and another group from flexors of the hand as well as extensors of the forearm. As in VSCT, no convergence was observed from antagonists working at the same joint (Eccles et al., 1961a). Presumably the RSCT units forward information concerning stages of movement or position of the whole limb rather than information of increased tension in individual muscles, just as has been suggested for the VSCT (Oscarsson, 1960). The group I activated units usually received polysynaptic excitation from the flexor reflex afferents in one or several ipsilateral nerves. Inhibitory action from the flexor reflex afferents was sometimes noted. Though the RSCT units are polysynaptically influenced from the flexor reflex afferents in similarity with the VSCT units (Oscarsson, 1957b; Eccles et al., I961a), they differ from the latter in that the synaptic actions of these afferents are not predominantly inhibitory as in VSCT. This is also

I

0 RSCT

A

VSCT

X DSCT

O ] 0

10

I

20

30

MSEC

40

Fig. 5 . Effect produced by a conditioning cutaneous voiley on mass discharges in DSCT, VSCT and RSCT. The tract discharges were recorded from the dissected lateral funiculus at the C3 level on stimulation of group I afferents in the ipsilateral (DSCT) and contralateral (VSCT) hamstring nerve and on combined stimulation of group I afferents in the deep radial nerve and the nerve to the long head of triceps (RSCT). The conditioning volleys were obtained by stimulation of the ipsilateral sural (DSCT), controlateral sural (VSCT), and ipsilateral superficial radial (RSCT) nerve at a strength of about 10 times threshold. Abscissa: volley interval in msec. Ordinate: amplitude of conditioned discharge in per cent of control value. 0-0 = RSCT; A-A = VSCT; X - x = DSCT. (From unpublished observations of Oscarsson and Uddenberg.) References p . 193-195

184

0. O S C A R S S O N

demonstrated in Fig. 5. In the same preparation the monosynaptic discharges in DSCT, VSCT and RSCT were conditioned by a preceding volley in cutaneous afferents from the same limb that supplied the monosynaptic excitation. The VSCT was strongly and the DSCT weakly inhibited (Oscarsson, 1957b), whereas the RSCT was facilitated. In other experiments the RSCT mass discharge was either facilitated or weakly inhibited by conditioning volleys in the flexor reflex afferents. The termination areas of the three spino-cerebellar tracts are shown in Fig. 6. Only DSCT

VSCT

RSCT

Fig. 6 . Cerebellar termination of left DSCT, VSCT, and RSCT. The diagram refers to the culmen of the anterior cerebellar lobe with Larsell’s lobules IV and V indicated. The curved line represents the rostra1 border of the exposed part of the anterior lobe. The interrupted line separates hindlimb and forelimb areas. Vertical lines indicate borders of intermediate cortices, horizontal lines sulci. (Data compiled from Lundberg and Oscarsson, 1960, 1962, and from unpublished observations of Oscarsson and Uddenberg.)

the culmen of the anterior lobe was explored in detail. Larsell’s lobules IV and V are indicated and the interrupted line shows the boundary between the hind- and forelimb areas according to anatomical (Grant, 1962b) and physiological investigations (Grundfest and Campbell, 1942; Snider and Stowell, 1944; Carrea and Grundfest, 1954; Combs, 1954). The dots indicate points from which individual units could be activated antidromically at a low stimulus strength. The DSCT terminates almost exclusively in the ipsilateral intermediate cortex, whereas the VSCT and RSCT terminate bilaterally in longitudinal zones consisting of a medial strip of the intermediate cortex and a lateral strip of the vermal cortex. Another similarity of the VSCT and RSCT units is that many of them can be activated antidromically from more than one point, often one ipsilateral and one contralateral, indicating branching of single fibres. Contralateral termination is most common with the VSCT units, whereas ipsilateral termination occurs more often with the RSCT units. This might be connected with the contralateral location of the VSCT cells and the ipsilateral location of the RSCT cells. The DSCT and VSCT terminate almost exclusively in the hindlimb area, whereas the RSCT units terminate approximately equally often in the hindlimb area as in the forelimb area. The bilateral termination areas of the VSCT and RSCT might indicate that the information conveyed by these tracts is used in the motor coordination of ipdateral

G R O U P I AFFERENTS I N FORELIMB NERVES

185

and contralateral limbs. The termination of the RSCT in both hindlimb and forelimb areas might correspondingly indicate that the information is used for the coordination of fore- and hindlimb movements. This tallies well with the hypothesis that the information mediated by these tracts concerns stages of movement or position of the whole limb. On the other hand, the DSCT units have small receptive fields and terminate ipsilaterally. It is possible that the information conveyed by these units is used mainly for the adjustment and coordination of movements within the ipsilateral limb. In this context it is interesting that the DSCT seems to terminate in the intermediate cortex, whereas the VSCT and RSCT have a more medial termination including a lateral strip of the vermis. According to Chambers and Sprague (1955a, b) each intermediate cortex regulates ‘the spatially organized and skilled movements and the tone and posture associated with these movements of the ipsilateral limbs’ and each vermal cortex, ‘tone, posture, locomotion, and equilibrium of the entire body’ (cf. also Pompeiano, 1958). THE CUNEO-CEREBELLAR TRACT

The cuneo-cerebellar tract originates from cells in the external cuneate nucleus and reaches cerebellum through the ipsilateral restiform body (Ferraro and Barrera, 1935). The external nucleus receives afferents from the dorsal funiculus but from no other known sources. The cuneo-cerebellar tract has, mainly on anatomical grounds, been assumed to be a forelimb homologue of the DSCT (Blumenau, 1891 ; Sherrington, 1890, 1893; Ferraro and Barrera, 1935; Brodal, 1941 ; Grant, 1962a). This hypothesis has now been substantiated by results obtained in an electrophysiological investigation of the tract (Holmqvist et al., 1963a). The mass discharge in the cuneo-cerebellar tract was recorded monophasically from the ‘dissected restiform body’ prepared as described by Holmqvist et al. (1963a). The spinal cord was almost completely transected sparing only the dorsal funiculi so as to exclude interference by activity in other ascending tracts. Records A-F in Fig. 7 show the mass discharge evoked by stimulation of the ipsilateral deep radial nerve at the indicated strengths. The first spike-like discharge appeared before any ingoing volley was discernible (A), showing that transmission through the external cuneate nucleus occurs with very little need for spatial summation. The first two spikes of the mass discharge were entirely due to excitation produced by the group I volley (A-D) ; the third spike was partly due to group I and partly to group I1 activation. At higher strengths of stimulation a late, prolonged discharge was added (E, F). Records G-L show the discharge evoked by stimulation of cutaneous afferents in the superficial radial nerve. Contralateral nerves were always ineffective (M, N). The units contributing to the mass discharge in the cuneo-cerebellar tract were analysed by intra-axonal recording from fibres in the region indicated by the dotted line in the diagram of Fig. 7. The results show that the tract contains one proprioceptive and one exteroceptive subdivision. The former consists of units activated monosynaptically from group I muscle afferents. One unit activated from the deep radial nerve is shown in Fig. 8. One or two impulses were discharged at a strength References p . 193-195

---0.O S C A R S S O N

186

--AA & A1)

A IM 1.01 B

1.06 C

1.35 D

G IS 1.01H

1.21 I

1.43

J

E

12 F

21 K

7.2 L

1.7

30

.IIIL-+--v-

M

CM N

cs 4Yec

-.-

$&#*YP4$

.-‘pCwskkwac

I5ocIv IOGec

Fig. 7. Mass discharges recorded from the dissected restiform body on stimulation of ipsilateral and contralateral muscle (deep radial) and skin (superficial radial) nerves in the forelimbs (IM, CM, IS, CS). The spinal cord was transected at C3 except for the dorsal funiculi. Upper two traces in each set of records show, on a fast time base, the ingoing volley recorded triphasically from the dorsal funiculi at C3 and the mass discharge in the restiform body. Lower trace shows the mass discharge on a slow time base. Stimulus strengths, in multiples of threshold for evoking a mass discharge, are indicated in A-L. The stimulus strength in M and N was at 20 times threshold. Dots mark stimulus artefacts. Voltage scale refers to mass discharge recording. The diagram describes the recording conditions. The ‘dissected restiform body’ was prepared as follows. The cerebellum was sucked away leaving the peduncles and adjacent white matter intact. The brachium conjunctivum and brachium rontis were cut through along the interrupted line. A loop tied to the peduncles was hooked into one c f the recording electrodes and used for lifting the ‘dissected restiform body’ from underlying tissue (not shown). The other recording electrode was placed against the ‘dissected restiform body’ where it was in continuity with the brain stem at the rostra1 border of the eighth nerve. Axonal recording was performed within the area surrounded by the dotted line. (Modified from Holmqvist et al., 1963a.)

evoking no perceptible ingoing volley (A) and four or five spikes appeared at a strength of about 1.2 times threshold (C). No additional impulses were elicited when the stimulus was increased to supramaximal for group I afferents (F). Stimulation of the ipsilateral skin nerve (G) and contralateral nerves (H, I) evoked no activity. The other units activated from group I afferents were similar. Ths first impulse appeared at a very low stimulus strength and a large group I volley almost always produced a repetitive response. There was never additional activation from group I I and I11 muscle afferents or from skin afferents. The response to stretch of musclc was a slowly adapting discharge and the receptive field was one or a few adjaccnt muscles. The group I activated units in the DSCT have similar characteristics (Laporte et al., 1956; Lundberg and Oscarsson, 1956; Holmqvist et al., 1956; Lundberg and Winsbury, 1960; Eccles et al., 1961b). They are monosynaptically activated from group I muscle

187

G R O U P I AFFERENTS I N FORELIMB N E R V E S

afferents in one or a few muscle nerves and the response is often repetitive. The DSCT units do not reczive additional excitation from group IT1 muscle afferents and cutaneous afferents but they are, presumably in contrast with the cuneo-cerebdlar units, sometimes activated from group TI muscle afferents. The DSCT contains one functional subgroup activated from Ia and another from Ib afferents. It is unknown if there are two corresponding subgroups in the cuneo-cerebellar tract. The exteroceptive subdivison in the cuneo-cerebellar tract will not be described in detail. It consists of units activated from cutaneous afferents and often also from high threshold (group TI and 111) muscle afferents. There are several subgroups which

-

&&L lldmmuw

WItmMnw-.

1

. !4w3pwwY .

Fig. 8. Cuneo-cerebellar unit activated from group I muscle afferents. Upper two traces in each set of records show, on a fast time base, microelectrode recording from the axon and recording from surface of the dorsal funiculi at C3; lower two traces show, on a slow time base, microelectrode recording from the axon and recording from dissected restiform body. A-F were obtained on stimulation of the ipsilateral muscle (deep radial) nerve (IM) at indicated strengths in multiples of mass discharge threshold. G I show that no discharge was elicited by stimulation (at 20 times threshold) of the ipsilateral skin (superficial radial) nerve (IS) or of the contralateral muscle and skin nerves (CM, CS). Superposed sweeps. Time scales in msec. (From Holmqvist et al., 1963a.)

correspond largely to those described previously for the exteroceptive subdivision in the DSCT (Lundberg and Oscarsson, 1960). However, the exteroceptive units in the cuneo-cerebellar tract differ, in one respect, conspicuously from those in the DSCT. The DSCT units are monosynaptically activated from cutaneous afferents, whereas the corresponding cuneo-cerebellar units are disynaptically activated from these afferents as well as from the sometimes converging high threshold muscle afferents. CEREBRAL PROJECTION OF GROUP 1 AFFERENTS

It is generally conccded that group I afferents from stretch receptors in hindlimb muscles do not project to the cerebral cortex (Mountcastle et al., 1952; McIntyre, 1962). However, a cerebral projection of forelimb group I afferents was suggested by References p. 193-195

188

0. O S C A R S S O N

th: findings of Amassian and Berlin (1958). These observations have now been confirmed and extended (Oscarsson and Roskn, 1963a, b).

-* SI

k

1.01

8 I

.

O

S

L

--

18

A

3 . O H P

$& 12.0

Ew

-5 MSEC

q5 MSEC 2 O

12.0H

4

6 . , l L 4 J

o

-

SII

F

V

4 q

Fig. 9. Evoked potentials recorded from the somatic area I (SI) and 11 (SII) on stimulation of the contralateral deep radial nerve. The potentials were simultaneously recorded on a fast (left traces) and slow time base. The inset records (upper left traces) show, on the fast time base, the primary afferent volley triphasically recorded from the dorsal funiculus at the C3 level immediately after the cortical recording. Stimulus strength in multiples of nerve threshold is indicated on each set of records. Positivity is signalled upwards. Voltage scale applies to cortical potentials (right traces). (From Oscarsson and RosCn, 1963a.)

Records A-E in Fig. 9 show surface-positive potentials evoked in the first somatic area (SI) on stimulation of group 1 muscle afferents in the contralateral deep radial nerve. The cortical potentials appeared at a strength producing a hardly visible ingoing volley (upper traces, A) and grew to a maximum with the group I volley (B-D). Additional activation of high threshold afferents did not increase the amplitude further but caused some increase of the following negative potential (E). On the other hand, in the second somatic area potentials appeared only when the strength was increased to activate group 11 afferents (F-J). Similar observations were made on stimulation of nerves to single muscles, for example the nerves to extensor carpi radialis, extensor digitorum communis, biceps, and the long head of triceps. The potentials evoked from group 1 afferents were limited to a small part of the forelimb region of SI, as defined by Woolsey and collaborators (Woolsey, 1947, 1959) (Fig. 10A). The group I potentials occurred only in the rostra1 part of this region, in the area between the postcruciate dimple and the cruciate sulcus. The responsive area was denoted R-SI and is shown schematically in Fig. 1OC. On the other hand, the potentials evoked from cutaneous and high threshold muscle afferents had two maxima, one in R-SI and the other in an area caudally of the dimple denoted C-SI (Fig. 1OC). These observations indicate that the forelimb region of S1 is differentiated

GROUP I AFFERENTS I N FORELIMB NERVES

189

Fig. 10. Dorsal view of the rostra1 pole of the cerebral hemisphere to show the forelimb region of the first somatic area (SI). The ansate, coronal, and cruciate sulci and the postcruciate dimple are indicated in A. (A). The area enclosed by the interrupted line shows the forelimb region of SI according to Woolsey (1947, 1959). (B). The area enclosed by the interrupted line shows the cortex responding to tactile stimulation of the forelimb and the hatched area, the cortex from which forelimb motor responses could be elicited in the experiments of Livingston and Phillips (1957). (C). Location of the responsive areas, C-SI and R-SI, discussed in the present paper. (From Oscarsson and Rosen, 1963b.)

into two parts with different function. It is tempting to associate R-SI with the motor and C-SI with the sensory cortex. Livingston and Phillips (1957) mapped, in the same experiments, the cortical areas responsive to tactile stimulation and those from which movements could be elicited. The forelimb regions of these areas are shown in Fig. 10B: the interrupted line encloses the responsive area and the hatched field shows the ‘motor’ area. The latter corresponds remarkably well to R-SI of the present investigation, though it extends more rostrally. Presumably the group 1 projection to the cerebral cortex represents a feedback channel used for adjusting the motor output from the cortex. The receptors of the group I afferents projecting to the cerebral cortex were identified by natural stimulation. A discharge from the appropriate receptors could be recognized by the depression of the cortical potential evoked on electrical stimulation

150

T.

100

50

0

Fig. 11. Change in amplitude of cortical potential evoked by a volley in Group I afferents of the nerve to extensor digitorum communis (EDC), on loading the tendon with various weights (A) and after close arterial injection of succinylcholine chloride (B). Ordinate: amplitude of the surface-positive cortical potential in per cent of control value obtained before loading (A) and injection (B). Abscissa: time in seconds after beginning of loading (A); time in minutes after injection (B). Interrupted vertical line in A indicates release from loading. Filled circles in B show absence of effect after transection of the EDC nerve. (From Oscarsson and Rosen, 1963b.) References p. 193-195

I90

0. O S C A R S S O N

of the intact nerve. This depression may be correlated with the marked reduction of the cortical potential that occurs at even low repetition rates of stimulation in the anaesthetized preparation. Fig. 11A shows the reduction of the potential evoked by stimulation of the nerve to extensor digitorum communis, on loading the muscle tendon with 10, 20 and 100 g. 10 g was sufficient to cause a definite depression and larger loads produced a marked reduction of the cortical potential. There was an initial phase of strong depression followed by a moderate one that remained until the end of the loading. Presumably the time course of the depression reflects the

Fig. 12. Relation between size of afferent volley and discharge evoked in medial IemnisLus and cuneo-cerebellar tract. The deep radial nerve was stimulated a t indicated strengths (multiples of nerve threshold). The traces show from above downwards: the lemniscal discharge and the ingoing volley at a fast speed, and the lemniscal discharge and the discharge in the cuneo-cerebellar tract at a slow speed. Positivity is signalled upwards. The lemniscal discharge was recorded with a steel needle electrode in the region of the lemniscus at a low pons level. The discharge in the cuneo-cerebellar tract was recorded from the dissected restiform body as described in Fig. 7. The ingoing volley was recorded triphasically from the dorsal funiculi at C3. Time scales in msec. (Partly unpublished records by Oscarsson and Rosen.)

adaptation of the receptors. Fig. 1 IB shows the depression of the cortical potential that occurred after close arterial injection of succinylcholine which is known to evoke a discharge in muscle spindle, but not in tendon organ afferents (Granit et a/., 1953). Following transection of the nerve the effect was abolished (filled circles). It was concluded from these and other experiments that group I afferents projecting to the cerebral cortex originate from muscle spindles, but an additional projection from tendon organs can not be excluded. The group I projection path belongs to the dorsal funiculus-medial lemniscus system. The cortical potentials evoked from group I afferents disappeared after a lesion in the dorsal funiculi but remained after almost complete transection of the cord sparing these funiculi. Additional confirmation was obtained on recording from the medial lemniscus at the pons level: the expected discharge appeared on stimulation of group I afferents in contralateral forelimb nerves (Fig. 12). The findings described in this and the previous section indicate that group 1 afferents ascending in the dorsal funiculi activate neurones in the external as well as in the main cuneate nucleus. The neurones of the external nucleus give origin to the uncrossed cuneo-cerebellar tract and the neurones of the main nucleus to the crossed medial lemniscus. The properties of the group I relays in the two nuclei were compared by simultaneous recording from the two tracts on stimulation of the deep radial nerve. The traces in Fig. 12 show from above downwards: the lemniscal discharge and the ingoing volley at a fast speed, and the lemniscal discharge and the discharge

191

G R O U P I AFFERENTS I N FORELIMB NERVES

in the cuneo-cerebellar tract at a slow speed. A discharge with a monosynaptic latency appeared in both tracts at a very low stimulus strength producing a hardly visible ingoing volley (A). The initial part of the discharge grew to a maximum with the group I volley (B-D) and additional activation of high threshold afferents prolonged the activity. An input-output curve obtained from a similar experiment is shown in Fig. 13A. Triangles represent the amplitude of the lemniscal discharge

X

b

50 GROUP I VOLLEY

100

( X MAX.)

I

CUNEOCEREBELLAR T R A C T

2

5

10

20

50

100 200

500

STIMULI / S E C

Fig. 13. (A). Input-output curve for transmission of impulses from group I afferents through the main and external cuneate nuclei. The deep radial nerve was stimulated. Abscissa: amplitude of ingoing volley recorded triphasically from the dorsal funiculus at C3. Ordinate: amplitude of mass discharge in medial lernniscus (triangles) and cuneo-cerebellar tract (crosses). (B). Effect of frequency on transmission through the main and external cuneate nuclei. The deep radial nerve was stimulated. Abscissa: stimulation frequency (log scale). Ordinate : amplitude of mass discharge in medial lemniscus (triangles) and cuneo-cerebellar tract (crosses) in per cent of value at a frequency of l/sec. Each point was obtained from superposed records when a steady state was attained, e.g. after 10-20 stimuli at the high frequencies. (Modified from Oscarsson and Rosen, 1963b.)

and crosses, the amplitude of the discharge in the cuneo-cerebellar tract. The curve fits both sets of points suggesting that, at both relays, the same types of afferents were responsible for the postsynaptic discharge and that transmission through both nuclei occurs with little need for spatial summation. On the other hand, the two group I relays differ in their ability to transmit impulses at high frequencies. The mass discharge in the lemniscus decreased at frequencies above 10, and that in the cuneo-cerebellar tract only at frequencies above 75/sec (Fig. 13B). In its ability to follow high frequencies the group I relay in the external cuneate nucleus is similar to the group I relays of the DSCT and VSCT (Holmqvist et al., 1956; Oscarsson, 1957b), whereas the relay in the main nucleus is intermediate between these relays and the monosynaptic connections between group I afferents and motoneurones (Adrian and Bronk, 1929; Lloyd and Wilson, 1957). It seems likely that the ability to follow high frequencies is related to the properties of the next synaptic relays. There is evidence suggesting that the thalamic relay has a pronounced recurrent inhibition (Andersen and Eccles, 1962) which presumably limits the transmission at high frequencies, whereas at least some neurones in the cerebellar cortex may follow very high frequencies of orthodromic stimulation (Granit and Phillips, 1956). COMMENTS

Group I afferents in hindlimb nerves have two ascending projections, the dorsal and the ventral spino-cerebellar tract. It must now be recognized that group J afferents in References p . 193-195

192

0. O S C A R S S O N

forelimb nerves have three projections, two terminating in the cerebellar cortex and one in the cerebral cortex. It is of interest to compare the hind- and forelimb projections and to discuss possible reasons for differences between them. The cuneo-cerebellar tract is obviously a forelimb homologue of the DSCT. The two tracts are largely equivalent as channels for proprioceptive and exteroceptive information. Ln both tracts the proprioceptive subdivision consists of units monosynaptically activated from group I muscle afferents and carrying information with a high degree of spatial discrimination. The exteroceptive subdivision has a similar functional organization in the two tracts but is monosynaptically activated from cutaneous afferents in the DSCT and disynaptically, in the cuneo-cerebellar tract. The reason for this discrepancy is unknown but might be connected with an earlier synaptic relay in the main cuneate nucleus (Holmqvist et al., 1963b). The rostral spino-cerebellar tract is anatomically distinct from the DSCT and VSCT but resembles functionally the latter tract. It can be regarded as a functional forelimb equivalent of the VSCT. RSCT units are monosynaptically activated from lugh threshold group I afferents, presumably identical with tendon organ (Ib) afferents. There is very often convergence of group I excitation from muscles working at different joints suggesting that the RSCT conveys information about stages of movement or position of the whole limb, as has previously been suggested for the VSCT (Oscarsson, 1960). The RSCT differs from the VSCT in that the polysynaptic effects from the flexor reflex afferents are not predominantly inhibitory as in the latter tract. The functional significance of this difference is obscure. Another difference relates to the termination of the two tracts. The VSCT terminates almost exclusively in the hindlimb area of the cerebellar cortex, whereas the RSCT terminates in the forelimb as well as the hindlimb area. Possibly the information carried by RSCT is more directly utilized in the coordination of fore- and hindlimb movements. The projection of group I afferents to the cerebral cortex through the dorsal funiculus-medial lemniscus system has no hindlimb equivalent. The forelimbs are used not only in locomotion but also in a wide variety of other movements, such as handling and perhaps exploration of the environment. This might necessitate a direct feedback channel for the control of movements elicited from the cortex, whereas hindlimb movements might be controlled mainly through reflex mechanisms at lower levels. The group I projection presumably represents such a feedback system and it is significant that it terminates in a rostral part of the postcruciate gyrus which probably is a motor area in the cat. Our findings demonstrate marked differences between ascending pathways related to the h n d - and forelimb levels respectively. Group I afferents project to the cerebellar cortex through two ‘forelimb’ and two functionally equivalent ‘hindlimb tracts’. The functional organization of each forelimb tract is similar to, but not identical with, the functional organization of the corresponding hindlimb tract and the anatomical organization is different. The cerebral projection of Group I afferents is related exclusively to the forelimb level. Obviously, these results call for considerable caution when transferring observations on tracts originating from one segmental level of the body to those originating from another level.

G R O U P I AFFERENTS I N FORELIMB NERVES

193

SUMMARY

Group I muscle afferents in hindlimb nerves activate two ascending tracts, the dorsal and the ventral spino-cerebellar tract (DSCT and VSCT). Recent investigations show that group I afferents in forelimb nerves activate three ascending pathways. 1. The rostra1 spino-cerebellar tract (RSCT) originates from cell bodies at, or slightly above, the level of the dorsal root entrance, ascends ipsilaterally in the middle third of the lateral funiculus, and terminates in a characteristic manner in the anterior cerebellar lobe. It receives monosynaptic excitation from high threshold group I muscle afferents, presumably identical with tendon organ afferents. Convergence of group 1 excitation from muscles working at different joints is common suggesting that the RSCT forwards information concerning stages of movement or position of the whole limb rather than information about increased tension in individual muscles. It is suggested that the RSCT is a functional forelimb homologue of the VSCT. 2. The cuneo-cerebellar tract contains one component that is monosynaptically activated from very low threshold group I muscle afferents, presumably identical with muscle spindle afferents. The response to a single volley in group I afferents is repetitive and transmission can occur at very high frequencies. The receptive field is small; it often consists of a single muscle. Other components of the cuneo-cerebellar tract are disynaptically activated from cutaneous afferents. It is concluded that the cuneocerebellar tract is a forelimb equivalent to the DSCT. 3. The third pathway is a projection to the cerebral cortex of large muscle spindle afferents. The group I afferents ascend in the dorsal funiculus and activate monosynaptically cells in the main cuneate nucleus which give origin to a component of the medial lemniscus. After a presumed thalamic relay the projection terminates in a small cortical area between the cruciate sulcus and the postcruciate dimple. It is suggested that this area has a motor function in the cat. The group I projection is presumably a feedback system used for the control of movements elicited from the cortex. The cerebral projection of forelimb group I afferents has no hindlimb equivalent. REFERENCES ADRIAN, E. D., A N D BRONK,D. W., (1929); The discharge of impulses in motor nerve fibres. Part 11. The frequeniy of discharge in reflex and voluntary contractions. J. Physiol. (Lond.), 67, 119-151. AMANAN,V. E., AND BERLIN,L., (1958); Early cortical projection of Group I afferents in the forelimb muscle nerves of cat. J . Physiol. (Lond.), 148, 61P. ANDERSEN, P., AND ECCLES,J. C., (1962); Inhibitory phasing of neuronal discharge. Nature (Lond.), 196, 645-647. L., (1891); Ueber den ausseren Kern des Keilstranges im verlangerten Mark. Neurol. BLUMENAU, Cbl., 10, 226-232. BRODAL, A., (1 941); Die Verbindungen des Nucleus cuneatus externus mit dem Kleinhirn beim Kaninchen und bei der Katze. Experimentelle Untersuchungen. 2.ges. Neurol. Psychiut., 171, 167-199. R. M. E., AND GRUNDFEST, H., (1954); Electrophysiological studies of cerebellar inflow. CARREA, 1. Origin, conduction and termination of ventral spino-cerebellar tract in monkey and cat. J . Neurophysiol., 17, 208-23 8. CHAMBERS, W. W., AND SPRAGUE, J. M., (1955a); Functional localization in the cerebellum. I.

194

0. O S C A R S S O N

Organization in longitudinal corticonuclear zones and their contribution to the control of posture, both extrapyramidal and pyramidal. J. cump. Neurol., 103, 105-129. CHAMBERS, W. W., AND SPRAGUE, J. M., (1955b); Functional localization in the cerebellum. 11. Somatotopic organization in cortex and nuclei. Arch. Neurol. Psychiat. (Chic.), 74, 653-680. COMBS,C. M., (1954); Electro-anatomical study of cerebellar localization. Stimulation of various afferents. J . Neurophysiol., 17, 123-143. ECCLES, J. C., HUBBARD, J. I., AND OSCARSSON, O., (1961a); Intracellular recording from cells of the ventral spino-cerebellar tract. J . Physiol. (Lond.), 158, 486-516. ECCLES, J. C., OSCARSSON, O., AND WILLIS, W. D., (1961b); Synaptic action of group I and I1 afferent fibres of muscle on the cells of the dorsal spino-cerebellar tract. J. Physiol. (Lond.), 158, 517-543. FERRARO, A., AND BARRERA, S. E., (1935); The nuclei of the posterior funiculi in Macacus rhesus. An anatomic and experimental investigation. Arch. Neurol. Psychiat. (Chic.),33, 262-275. GRANIT, R., AND PHILLIPS, C. G., (1956); Excitatory and inhibitory processes acting upon individual Purkinje cells of the cerebellum in cats. J . Physiol. (Lond.), 133, 520-547. GRANIT,R., SKOGLUND, S., AND THESLEFF, S., (1953); Activation of muscle spindles by succinylcholine and decamethonium. The effects of curare. Acta physiol. scand., 28, 134-151. GRANT,G., (1962a); Projection of the external cuneate nucleus onto the cerebellum in the cat: An experimental study using silver methods. Exp. Neurol., 5, 179-195. GRANT,G., (196213); Spinal course and somatotopically localized termination of the spinocerebellar tracts. An experimental study in the cat. Actaphysiol. scand., 56, Suppl. 193, 1 4 5 . GRUNDFEST, H., AND CAMPBELL, B., (1942); Origin, conduction and termination of impulses in the dorsal spinocerebellar tracts of cats. J . Neurophysiol., 5, 275-294. HOLMQVIST, B., LUNDBERG, A., AND OSCARSSON, O., (1956); Functional organization of the dorsal spino-cerebellar tract in the cat. V. Further experiments on convergence of excitatory and inhibitory actions. Acta physiol. scand., 38, 76-90. HOLMQVIST, B., OSCARSSON, o., AND ROSBN,I., (1963a); Functional organization of the cuneocerebellar tract in the cat. Acfa physiol. scand., 58, 216-235. HOLMQVIST, B., OSCARSSON, O., AND UDDENBERG, N., (1963b); Organization of ascending spinal tracts activated from forelimb afferents in the cat. Acta physiol. scand., 58, 68-76. LAPORTE, Y . , LUNDBERG, A., AND OSCARSSON, o., (1956); Functional organization of the dorsal spino-cerebellar tract in the cat. 11. Single fibre recording in Flechsig’s fasciculus on electrical stimulation of various peripheral nerves. Acta physiol. scand., 36, 188-203. LIVINGSTON, A., AND PHILLIPS, C. G., (1957); Maps and thresholds for the sensorimotor cortex of the cat. Quart. J . exp. Physiol., 42, 190-205. LLOYD,D. P. C., AND MCINTYRE, A. K . , (1950); Dorsal column conduction of group I muscle afferent impulses and their relay through Clarke’s column. J. Neurophysiol., 13, 39-54. LLOYD,D. P. C., AND WILSON,V. J., (1957); Reflex depression in rhythmically active monosynaptic reflex pathways. J . gen. Physiol., 40, 409426. LUNDBERG, A., AND OSCARSSON,o.,(1956); Functional organization of the dorsal spino-cerebellar tract in the cat. 1V. Synaptic connections of afferents from Golgi tendon organs and muscle spindles. Acta physiol. scand., 38, 53-75. LUNDBERG, A., AND OSCARSSON, o., (1960); Functional organization of the dorsal spino-cerebellar tract in the cat. VII. Identification of units by antidromic activation from the cerebellar cortex with recognition of five functional subdivisions. Acta physiol. scand., 50, 356-374. LUNDBERG, A., AND OSCARSSON, o.,(1962); Functional organization of the ventral spino-cerebellar tract in the cat. IV. Identification of units by antidromic activation from the cerebellar cortex. Acta physiol. scand,, 51, 252-269. LUNDBERG, A., AND WINSEURY, G., (1960); Functional organization of the dorsal spino-cerebellar tract. VI. Further experiments on excitation from tendon organ and muscle spindle afferents. Acta physiol. scand., 49, 165-170. MCINTYRE, A. K., (1962); Central projection of impulses from receptors activated by muscle stretch. Symposium on MuJcle Receptors. D. Barker, Editor. Hong Kong, University Press (p. 19-30). MOUNTCASTLE, V. B., COVIAN,M. R., AND HARRISON, C. R. (1952); The central representations of some forms of deep sensibility. Ass. Res. nerv. Dis.Proc., 30, 339-370. OSCARSSON, O., (1957a); Primary afferent collaterals and spinal relays of the dorsal and ventral spino-cerebellar tracts. Acta physiol. scand., 40, 222-23 1. OSCARSSON, O., (1957b); Functional organization of the ventral spino-cerebellar tract in the cat. 11. Connections with muscle, joint, and skin nerve afferents and effects on adequate stimulation of various receptors. Acta physiol. scand., 42, Suppl. 146, 1-107.

GROUP I AFFERENTS I N FORELIMB NERVES

195

OSCARSSON, O., (1960); Functional organization of the ventral spino-cerebellar tract in the cat. 111. Supraspinal control of VSCT units of I-type. Actaphysiol. scand., 49, 171-183. OSCARSSON, O., AND ROSBN,I., (1963a); Cerebral projection of Group 1 afferents in forelimb muscle nerves of cat. Experientia (Basel), 19, 206. O., AND ROSEN,I., (1963b); Projection to cerebral cortex of large muscle spindle afferents OSCARSSON, in forelimb nerves of the cat. J. Physiol. (Lond.), 169, 924-945. POMPEIANO, O., (1958); Responses to electrical stimulation of the intermediate part of the cerebellar anterior lobe in the decerebrate cat. Arch. ital. Biol., 96, 330-360. REXED,B., A N D STROM,G., (1952); Afferent nervous connexions of the lateral cervical nucleus. Acta physiol. scand., 25, 219-229, SHERRINGTON, CH. S., (1890); On out-lying nerve-cells in the mammalian spinal cord. Phil. Trans. B, 181, 3348. SHERRINGTON, CH. S., (1893); Note on the spinal portion of some ascending degeneration. J . Physiol. (Lond.), 14, 255-302. A., (1944); Receiving areas of the tactile, auditory and visual systems SNIDER, R. S., AND STOWELL, in the cerebellum. J . Neurophysiol., 7 , 331-357. WOOLSEY, C. N., (1947); Patterns of sensory representation in the cerebral cortex. Fed. Proc., 6, 437441. C. N., (1959); Some observations on brain fissuration in relation to cortical localization WODLSEY, of function. Structure and Function of the Cerebral Cortex. D. B. Tower and J. P. Schade, Editors. Proc. 2nd Int. Meet. Neurobiol. Amsterdam, Elsevier (p. 64-68).

DISCUSSION

GELFAN:The cortical projection of afferent outflow from muscle, particularly in group la fibers, has interested us for some time, as it has other neurophysiologists. Since hitherto experiments on animals failed to demonstrate any such projection of impulses from spindle annulo-spiral endings, Dr. Sylvester Carter, a hand surgeon, and I decided to test it in man. So far we have tried it only on 4 suitable surgical cases, in which the tendons at the wrist were exposed under local anaesthesia, limited to the skin, and with no pre-operative medication. When a tendon was pulled so as to stretch the muscle, as for example the palmaris longus, the sensation was never referable to the muscle. The patient would sometimes report that the skin over the muscle area was being pulled. When the tendons were pulled so as to move the fingers, the reports accurately identified the specific finger movement, i.e. joint movement was readily recognized and appreciated. But there has so far been no evidence of any coiiscious recognition of changes in length of muscles. OSCARSSON: Our experiments have, so far, been limited to the cat and we don’t know anything about the organization in other species. However, the information reaching the cerebral cortex from large muscle spindle afferents in the cat might not enter the ‘consciousness’ of the animal. This is suggested by the recent investigation of Giaquinto, Pompeiano and Swett (Arch. ital. Biol., 101 (1963) 133-148). These authors showed that repetitive stimulation of group I afferents in the deep radial nerve did not influence the EEG and behaviour of sleeping or waking cats. The group I projection terminates in what is presumably the motor cortex of the cat and might represent a feedback system adjusting movements elicited from this cortical area. Thus the cerebral cortex seems to contain mechanisms as unrelated to consciousness as the motor regulating mechanisms in the cerebellum.

196

DISCUSSION

CREUTZFELDT: I should like to ask Dr. Oscarsson whether he has an idea in which thalamic nucleus the corticopetal Ia afferents may be relayed. lntracellular recordings from Betz cells in the cat motor cortex performed in our lab (Lux, Nacimiento and myself) have consistently shown short latency primary EPSPs after stimulation of the VPL nucleus of the thalamus. This monosynaptic connection between VPL and Betz cells may represent the next link in a la-spino-corticospinal reflex pathway whose centrifugal path would then be the corticospinal tract. The relatively low following frequency of cortical responses may suggest a frequency limiting recurrent inhibition in the thalamic relay as assumed also in other thalamo-cortical systems. OSCARSSON: We have not investigated the site of the presumed thalamic relay. KUYPERS: The data you presented are of extreme interest to me since we have been looking at the descending pathways to the cuneate and gracilis nucleus. At least in monkey there is some precentral projection to the cuneate and gracilis. Peculiarly enough it was always thought that the cerebral projection was first and foremost postcentral and only in exceptional cases did the Woolsey school feel that there was any precentral contribution. Dr. Creutzfeldt has pointed out that the VPL projects to the precentral gyrus or an area comparable to the precentral gyrus. This would mean that the precentral gyrus is part of a VPL-circuit and this would facilitate the explanation of a projection to the cuneate and gracilis nucleus.

197

Supraspinal Control of Transmission in Reflex Paths to Motoneurones and Primary Afferents A. L U N D B E R G

Department of Physiology, University of Goteborg, Giiteborg (Sweden)

In the introduction to ‘The integrative action of the nervous system’ Sherrington (1906) states that the reflex is ‘the unit reaction in nervous function’. In Sherrington’s opinion integration is at least in part a compounding of reflexes. It was this integrative aspect that prompted us to investigate the descending control of reflexes. Electrophysiological work of the last decades has greatly advanced our knowledge of spinal reflexes from different types of receptors, but the problem of how different reflexes can co-function harmoniously has been given little attention. It is now clear that the supraspinal control of reflex paths is one important aspect of this problem. Spinal reflexes in the cat can be modified from higher centres in different ways. When the excitability of the a-motoneurones is changed, there is also an action on the reflexes but this effect is as far as we know indiscriminate, affecting all reflexes alike. Another type of descending control of reflexes depends on the well-known efferent y-control of the muscle spindle (Leksell, 1945; Granit, 1955). A selective effect on reflexes from muscle spindles can be exerted by a change in the y-bias but this mechanism will not be discussed further. This report deals with the descending control of transmission from primary afferents to motoneurones and also with the control of transmission in the recently disclosed reflex paths to primary afferents giving primary afferent depolarization and thereby presynaptic inhibition (Frank and Fuortes, 1957; Eccles et al., 1961). There are both facilitatory and inhibitory influences, but a start will be made with the facilitatory action from the sensorimotor cortex, which in some respects is less complex than the inhibitory control from the brain stem discussed in the second part. I. F A C I L I T A T O R Y E F F E C T F R O M T H E S E N S O R I M O T O R C O R T E X

There may be several descending systems facilitating transmission in reflex paths but so far the only one disclosed is the corticospinal tract. Facilitatory effects on transmission in reflex paths to motoneurones and to primary afferents will be described in (a) and (b) respectively. Effects on interneurones are described in section (c). (a) Reflex paths to motoneurones

Facilitation from the pyramidal tract of spinal reflexes at an interneuronal level was References p . 217-219

198

A. L U N D B E R G

first demonstrated by Lloyd (1941) (cf also Lindblom and Ottosson, 1956; 1957) and has recently been systematically investigated with respect to reflex paths from different afferent systems (Lundberg and Voorhoeve, 1962). Fig. 1 illustrates facilitation from the sensorimotor cortex of the Ta inhibitory pathway to a-motoneurones. Intracellular recording was made from a gastrocnemius motoneurone. A l a volley from the A

DP 1.30

cortex

-

7

C

cortex+DP

50x p

I PER C E N T INHIBITION

100

o

-

20 msec

5c-

corlex+DP+G-S

o~D-o-o--o-o

h

P

& ~ . ‘ . r X x

I

i

‘

Fig. 1. Facilitation t t t 5 from the contralateral sensorimotor cortex of the Ia inhibitory path to moto-

neurones. Records A-C are intracellular records from a gastrocnemius-soleus (G-S) motoneurone and internal positivity (depolarization) is signalled upwards. The lower traces in A-C are from the L7 dorsal root entry zone and a downwards deflection signals negativity. The left and right traces in each record were taken simultaneously at different speeds. In A and C are shown the responses evoked by a la volley in the deep peroneal nerve (DP). In B the sensorimotor cortex was stimulated and in C there is spatial facilitation on combined stimulation of the cortex and the DP nerve. The graph shows the effect of cortical stimulation on the reciprocal Ia inhibition from the DP nerve of the G-S monosynaptic test reflex. The effect of 3 stimuli from the sensorimotor cortex is shown by circles, whereas crosses show the inhibition without cortical stimulation. Intervals on the abscissa are between the first cortical stimulus and the inhibited monosynaptic test reflex. The drawing t o the right shows the cortical areas from which there were facilitation of the reciprocal Ia inhibitory paths. The pathway tested in the hindlimb was from DP to G-S and in the forelimb from biceps to the lateral and medial divisions of triceps (Lundberg and Voorhoeve, 1962).

antagonist pretibial flexors did not evoke any IPSP in A. Record C shows that when the same Ia volley is preceded by stimulation of the sensorimotor cortex a large la TPSP is evoked. Hence there is a powerful facilitation of the Ia inhibitory pathway at a strength of cortical stimulation that in itself does not evoke any synaptic actions in the motoneurones (B). Presumably the effect is due to excitatory action from the sensorimotor cortex on the inhibitory interneurone now known to exist in the Ia inhibitory pathway (Eccles eta/., 1956; Eccles and Lundberg, 1958; Araki eta/., 1960; Eide et al., 1961). The graph in Fig. 1 gives the time course of the facilitatory effects; the cortical areas from which minimal action can be evoked on the Ia inhibitory paths in hind- and forelimb are also shown in Fig. 1. Effects from the sensory and motor regions cannot be differentiated with respect to hindlimb actions but the forelimb effect is from the motor region (cJ Livingston and Phillips, 1957) hence it is postulated that the action on the reflex path, as could be expected, is from the motor cortex.

CONTROL OF TRANSMISSION I N REFLEX PATHS

199

The same facilitatory action from the sensorimotor cortex is found on the excitatory and inhibitory paths from Ib afferents and from the FRA (flexor reflex afferents = group I1 and 111 muscle afferents, cutaneous afferents and high threshold joint afferents). The records in Fig. 2 reveal facilitation of excitatory and inhibitory paths from cutaneous afferents in a flexor and an extensor motoneurone respectively. Similar findings were made for the flexor reflex actions evoked from group 11and 111muscle afferents and from high threshold joint afferents. In flexormotoneurones, volleys in the FRA can evoke either excitation or inhibition (cf. Eccles and Lundberg, 1959a; Paintal, 1961 ; Holmqvist and Lundberg, 1961); the inhibitory path, normally not open in the spinal state, can be facilitated from the cortex, and the same holds true for the excitatory path from cutaneous afferents to extensor motoneurones (cf. Iagbarth, 1952).

A

cortex

Sur

AD

'

E Sur

cortex

F cortex+Sur

Fig. 2. Facilitation from the contralateral sensorimotor cortex of reflex paths from cutaneous afferent to motoneurones. Intracellular recording from a posterior biceps-semitendinosus motoneurone in A-C and from a gastrocnemius-soleus motoneurone in D-F. The lower traces were recorded from the L7 dorsal root entry zone and negativity is signalled by a downwards deflection. The sural nerve was stimulated in A and D, B and E show the effect of cortical stimulation alone and C and F of combined stimulation of cortex and the sural nerve (Lundberg and Voorhoeve, 1962).

Recently Engberg (1963b) has found strong facilitation from the cortex of the excitatory path from the pad to toe extensor motoneurones. In summary : all polysynaptic reflex paths to a-motoneurones hitherto investigated can be facilitated from the cortex. Experiments with transection of the pyramid and with transection of the brain stem sparing the pyramid, have shown that effects on reflex paths are mediated by the corticospinal tract. To reveal the facilitatory effects discussed above, the strength of cortical stimulation was subliminal or liminal for actions on the inotoneurones investigated. At higher strength, stimulation of the cortex evoked synaptic potentials. Although there was often evidence of mixed inhibitory and excitatory action in individual motoneurones (Lundbergand Voorhoeve, 1962, Figs. 18 and 19) it is of special interest that excitatory actions often dominated in flexor motoneurones, and inhibitory in extensor motoneurones. This is illustrated by the intracellular record i n Fig., 3 A and B and by the curves obtained with the monosynaptic test method in the graph of Fig. 3. This References p . 2 I 7-2 I9

200

A. L U N D B E R G

distribution of synaptic actions to flexor and extensor motoneurones should be correlated with actions from the cortex on spinal reflex paths. The dominating spinal reflex is the flexor reflex; the expected result of an activation of interneurones of ipsilateral reflex paths would therefore be inhibition of extensor and facilitation of flexor motoneurones. There are several other findings supporting the hypothesis that

fl\l

0

I

10

Cortex-, PBSt Cortex-,G-S

I

I

I

20

30

40

.*

msec

IOrnsec

Fig. 3. Effects from the contralateral sensorimotor cortex on a-motoneurones. The graph to the left shows the effect of cortical stimulation ( 5 stimuli) on the monosynaptic test reflex from gastrocnemiussoleus, G-S, ( a ) and from posterior biceps-semitendinosus, PBSt, (x), 100% on the ordinate represents the unconditioned amplitude of the test reflex. Contitioned test response, expressed as percentage of control amplitude, is plotted as a function of interval between the first cortical stimulus and the monosynaptic test reflex. The same strength of cortical stimulation was used to obtain the two curves. A and B are intracellular recordings (lower trace in A and upper trace in B) from a G-S and a PBSt motoneurone as indicated in the records. Upper trace in A and the lower trace in B are from the dorsal root entry zone in L7 and a downwards deflection signals negativity (Lundberg and Voorhoeve, 1962 and unpublished results),

the synaptic action from the sensorimotor cortex is secondary to the activation of reflex paths. For example in some flexor motoneurones, inhibition was the dominating effect from the cortex, and these motoneurones also received mainly IPSPs from the FRA. Toe extensors (in particular flexor digitorum brevis) differ from other extensor motoneurones in that they receive strong excitatory action from the cortex, and this should be correlated with the existence of an effective excitatory action from the pad reaching these motor nuclei exclusively (Engberg, 1963a, b). The finding that many motoneurones receive mixed excitatory and inhibitory actions from the sensorimotor cortex is in agreement with the fact that such a variety of reflex paths can be mobilized from the cortex. It is also likely that the corticospinal tract has effects on crossed reflex paths. The discussion above refers to the dominating ipsilateral actions. (b) Reflex paths to primary aflerents

It has been shown independently by two groups that stimulation of the sensorimotor cortex evokes primary afferent depolarization (Andersen et al., 1962; Carpenter,

CONTROL OF TRANSMISSION I N REFLEX P A T H S

20 1

-- - v

Lundberg et al., 1962, 1963). I n Fig. 4 dorsal root potentials (DRP) were recorded bilaterally and the cord was hemisectioned as shown in the diagram. In lumbar roots the threshold action was from the hindlimb area. After section of the pyramid, the

A

B

left h

rlkl*lrrcp

-w

right

Fig. 4. Primary afferent depolarization evoked from the sensorimotor cortex. The dorsal root potentials in A and C (upper traces) were evoked from the right sensorimotor cortex and recorded from the most caudal dorsal rootlet in L7 on the left and right side as indicated. The lower traces were all rzcorded with one electrode placed on the dorsal column and negativity is signalled by a n upwards deflection. The corresponding records B and D were obtained after section of the right pyramid (cf. lower right diagram). The shaded area in the lower left diagram is the region from which actions could be obtained at threshold stimulation (Carpenter, Lundberg et al., 1963).

--A

B

cortex 0.3mA I0.5m/"

0.6

C

F

j

dissected

10 msec Fig. 5 . The upper traces are dorsal root potentials recorded from the most caudal dorsal rootlet in L6, the lower traces were recorded from the dorsal root entry zone in L7. The contralateral sensorimotor cortex was stimulated in A-E and the strength is indicated in each record. Record F was obtained in another experiment in which the contralateral pyramid was dissected and stimulated in a descending direction. Observe that there is initial negativity in the cord dorsum recording. Even at strong stimulation there was no reversal to an initial positivity as in D and E. References p . 217-219

202

A. L U N D B E R G

2500

00

.

:

0 0

09% 08

O

- 0

%-

n

%%

8

oSP--Sural cortex+Sural

o * a 0 m

(A)

0

I5O-

"0 0

"i 100-

.

o :

O

.o'

0*

0

0 .

I

I

I

50

25

I

150

100

.

200msec

3000 0 0

250-

0

Q

g

O

0

0

ooo

200-

(B)

0

o:o

0

O

co

0

eb8,

o

0 0

. e 0

**.*

*.

8 0

50

I

0

100

0

0,"

o

150-

loo-

PBSt 4 G-S Ia cortex+G-SIa

80'8

.

.O

o

o

0

0 0

*.

*

*

0

8

200

300

400

500msac

Fig. 6. (A) Effects from the contralateral sensorimotor cortex on the excitability of terminals of primary afferents. To obtain the upper graph the testing stimulus was delivered through a microelectrode inserted into the dorsal horn at the site of the maximal N1 potential and the discharge was recorded in the sural nerve. There is a facilitatory effect of a conditioning single volley in the cutaneous superficial peroneal nerve, SP, ( 0 ) and of a train of cortical stimuli ( ). (B) The lower graph is from an experiment in which the excitability of presynaptic terminals of Ia fibres was tested. The testing stimulus was delivered through a microelectrode inserted into the motor nucleus of gastrocnemius-soleus (G-S) at the site of the maximal Ia focal potential. Observe that conditioninggroup 1 volleys from the PBSt nerve give a large increase of excitability. By contrast there is no effect from the ccntralateral cortex. (Carpenter, Lundberg et al., 1963).

CONTROL OF TRANSMISSION IN REFLEX PATHS

203

DRP, the associated P wave, as well as the preceding negative cord dorsum potential (CDP) disappeared, showing that the corticospinal tract is responsible (B and D). However, effects can be conducted outside this tract. Fig. 5 shows the DRPs and CDPs evoked at different strengths of cortical stimulation. At weak stimulation there is a negative DRP and superimposed on it at somewhat stronger strength the positive CDP that is associated with the DRP (cf. Barron and Matthews, 1938; Eccles et a/., 1962~).At still stronger stimulation, the initial negative CDP reverses to a positivity and the DRP increases (D and E). Of these effects only those associated with an initial negative DRP can be ascribed to the corticospinal tract. This action disappears after transection of the pyramid and, furthermore, with descending stimulation of the dissected pyramid only this effect is evoked (record F), never the initial positive CDP. On strong cortical stimulation the initial positive CDP can be evoked also after transection of the pyramid. Experiments with excitability measurements from the terminals of primary afferents ad modum Wall (1958) and with intracellular recording from primary afferents have shown that the DRP evoked from the cortex represents a primary afferent depolarization in cutaneous afferents (Fig. 6A), in group Ib and TI muscle afferents, but not in l a afferents (Fig. 6B) (Andersen et a[., 1962; Carpenter, Lundberg et a/., 1962). Our previous findings with respect to actions from the cortex on motoneurones, raised the question if the action on primary afferents could be due to an excitatory action on interneurones of the reflex paths to primary afferents. Experiments showing spatial facilitation from the two sources indicate that this is the case. Fig. 7 shows

---

A

PBSt

B

cortex

C

cortexrPBSt

- v

cortex

Sural

Fig. 7. Spatial facilitation between pathways to primary afferents. The lower traces in A to C and the upper traces in D to F were recorded from the most caudal rootlet in L6. The other traces are from the dorsal root entry zone in L6. A shows the effect of a single group I volley in the nerve from PBSt. In B the contralateral sensorimotor cortex was stimulated at a strength liminal for evoking a dorsal root potential. The facilitatory effect of a combined stimulation of PBSt and the cortex is shown in C . In the corresponding lower records, D-F, the sural nerve was stimulated. It should be noted that there is facilitation from cortex of both component I and component I1 of the DRP (Carpenter, Lundbcrg et al., 1963).

that cortical stimulation can facilitate the DRPs evoked from group I muscle afferents (A-C) and from cutaneous afferents (D-F) and the same was found for the DRPs evoked by volleys in high threshold muscle and joint afferents. Further experiments w!th intra-axonal recordings from fibres of different types have revealed facilitation References p . 217-219

204

A. L U N D B E R G

from the cortex of the following reflex paths to primary afferents (Carpenter, Lundberg et al., 1963): Ib to Ib (cf. Eccles et al., 1962a, b, 1963a), FRA to FRA (cf. Eccles et al., 1963b), cutaneous to cutaneous (Carpenter, Engberg et al., 1963). There was on the other hand never any indication that the pathway to la afferents could be facilitated. This pathway can nevertheless be influenced from the sensorimotor cortex, but the effect is inhibition. A similar inhibitory effect of the path from la to Ia afferents is evoked in the spinal cat by volleys in the FRA, presumably by presynaptic inhibition at an interneuronal level (Lundberg and Vyklickf, 1963a). Hence also in this particular case the cortical effect can be explained by activation of interneurones of a spinal path from the FRA. ( c ) EfSect on interneurones

The results described above led to the postulate that interneurones of many reflex paths receive excitatory action from the corticospinal tract. This was confirmed with intracellular recordings from interneurones activated by different types of afferents (Lundberg et al., 1962). The interneurone in Fig. 8 receives monosynaptic excitatory

B Y r L +,

cortex

,

,at,

IlmV

.

U

5msec cortex ext race11.

Fig. 8. Excitatory action from the contralateral sensorimotor cortex on an interneurone activated by group I muscle afferents. Intracellular recording (upper traces) was made from a cell receiving monosynaptic excitatory action from the nerves to gastrocnemius-soleus, G-S (record A) and from the nerve to plantaris, PL (record B). Volleys in other peripheral nerve tested has no action on this cell. The effect of cortical stimulation is shown in C. Record D was obtained after withdrawal of the microelectrode to a just extracellular position. The lower traces were recorded from the L7 dorsal root entry zone and downwards deflection signals negativity (Lundberg et al., 1962).

action from group I muscle afferents (A and B) and there is excitatory action also from the sensorimotor cortex (C). The interneurone in Fig. 9 receives excitatory action from the FRA ;there is the characteristic convergence from a very large receptive field. Repetitive cortical stimulation evokes an EPSP with steps (record C) and record N, at higher amplification, shows that a single cortical stimulus evokes an EPSP which to judge from its time course may very well be a monosynaptic action, although the latency is too long in relation to the onset of the corticospinal discharge 0 to make

205

C O N T R O L O F T R A N S M I S S I O N I N REFLEX P A T H S

this certain. On a few occasions it was observed that monosynaptic EPSPs could be evoked in interneurones (Lundberg et al., 1962, Fig. 4) but with the longer latencies usually found, it is not possible to say if the action is monosynaptic or not. Effects like those in Fig. 8 are probably polysynaptic. From intracellular recordings of 3 1 interneurones, 27 received mainly excitatory action, but in the remaining 4 interneurones, cortical stimulation evoked mainly IPSPs. These interneurones were also inhibited by volleys in the FRA (Fig. 10) and the probable explanation of the inhibitory effect from the cortex is excitation of

----+C

-

ESt

ABSm

JllP-

H

2.3

P

h

M

L 11t1

N

CORTEX

A 2Omsrc

_zzc

5 rnscc

Fig. 9. Effect from the sensorimotor cortex on an interneurone activated from the FRA. These intracellular recordings were obtained from an interneurone located at a depth of 1.9 mm from the cord dorsum, probably in the dorsal part of the intermediate region. The upper traces were recorded at the L7 dorsal root entry zone. Records A--K were obtained on stimulation of the nerves indicated, the abbreviations being: Sur, sural; G-S, gastrocnemius-soleus; BSt, biceps-semitendinosus; ABSm, anterior biceps-semimembranosus; FDL, flexor digitorum and hallucis longus 4- the interosseus nerve; Q, quadriceps; Joint, posterior nerve to the knee joint. The peroneal (P) nerve was stimulated in H-K. Stimulus strengths are given in multiples of threshold strengths for the nerves. In L a short train of stimuli was given to the sensorimotor cortex, a t a strength just threshold for effect and in M at a somewhat higher strength but still submaximal with respect to the negative dorsal horn potential that could be evoked from cortex. Record N shows theEPSP evoked by a single cortical stimulus of the same strength that was used in L. The left and right traces in records K-N were obtained simultaneously at two sweep speeds, the slow speed to the left and the fast speed to the right below record L. A-J were taken at the slow speed indicated below L. Record 0 was obtained a t the end of the experiment and shows the discharge evoked by a single cortical stimulus in the dissected contralateral dorsal half of the lateral funicle in L6 (Lundberg el al., 1962).

interneurones, inhibitory with respect to the interneurone from which the recording was made. This explanation is supported by the finding of spatial facilitation between the inhibitory paths from the periphery and the cortex (records F-H). It must be emphasized that at present we cannot decide to which path a particular interneurone belongs; the investigations of synaptic actions on interneurones nevertheless virtually prove the hypothesis that the corticospinal tract mobilizes the spinal reflex path by evoking excitatory action in their interneurones, Rrfrrences p . 217-219

206

A. L U N D B E R G

A

B

Sur

cortex

-

v

.....

-

10 msec

50 msec

C

Joint

D

ABSm

v w

-Iu+-=-

cortex

v v .....

SP

L

I-@--

F

E

G

Joint

H cortex+Joint

,....

Fig. 10. Inhibitory action from the sensorimotor cortex in an interneurone receiving inhibition from the FRA. The intracellular recordings (upper traces) were obtained from a cell in the dorsal horn of L7. Lower traces in A and F-H and middle traces in B-E were recorded from the dorsal root entry zone in L7. The lower traces in B-E are microelectrode recordings obtained after withdrawal to a just extracellular position. This interneurone was monosynaptically activated by large afferents in thc sural nerve (record B). The dominating effect of cortical stimulation was an IPSP (record B). C-E shows that lPSPs were evoked also from peripheral nerves, the abbreviations being: joint, posterior nerve to the knee joint; ABSm, anterior biceps-semimembranosus; SP, superficial peroneal nerbe. The effects in C and D were evoked from high threshold afferents. F-H illustrates spatial facilitation between the paths from the cortex and the periphery. Liminal stimulation was used in F and G and when combined in H the action is larger than the sum of the effects in F and G. Calibration betwecn A and B refers to record B and H. The lower amplification in A was not recorded (Lundberg and Norrsell, unpublished results).

Comments There is now strong evidence that the reflex is the unit reaction upon which the corticospinal tract operates in the cat. In the primate there is the additional monosynaptic connection that presumably subserves fine movements (Bernhard et a/., 1953; Landgren eta/., 1962a, b). It seems likely that this mechanism is superimposed on the phylogenetically older mobilization of reflexes. At least Sherrington (1906) emphasized that the movements evoked by stimulations of the motor cortex in the monkey resemble that of the spinal reflex. The activation from the cortex of reflex arcs to motoneurones and to primary afferents would tend to have opposite results, the former facilitating and the latter inhibiting transmission of reflexes to skeletal muscles. Tower (1935) has shown that after transection of the pyramid the threshold for the flexion reflex increases markedly suggesting that under these conditions the dominating effect is the elimination of facilitatory action on the reflex paths to motoneurones. Presumably the same explanation holds true for the appearance of the Babinski sign in humans with diseased corticospinal tracts (cf. Kugelberg et a/., 1960).

CONTROL OF TRANSMISSION I N REFLEX P A T H S

207

It is not certain that the effects from the cortex on primary afferents should be taken to indicate a distribution of presynaptic inhibition from the sensorimotor cortex. The mobilization of the reflex to primary afferents may be the main factor. With respect to the reflex depolarization of Ib afferents and of the FRA it has been pointed out that in both cases presynaptic inhibition constitutes a negative feedback serving to reject stray excitation and hence contributing to the local sign of a reflex (Eccles et al., 1963a, b). It seems likely that with the facilitation from the cortex of a reflex path from a primary afferent system to motoneurones, there is also a mobilization of the negative feedback subserving this reflex. From this point of view it is interesting to note that the path to Ia fibres cannot be facilitated from the sensorimotor cortex and in this case the presynaptic inhibition cannot be described as a negative feedback (cf. Eccles et al., 1962~).The hypothesis that the effects on the paths to primary afferents are accessory to those on the paths to motoneurones rather presupposes that it is an effect of the motor cortex. Kuypers (1960) has shown that both the sensory and the motor cortical regions contribute to the corticospinal tract in the monkey. This has also been found in the cat, the fibres from the sensory cortex having a more medial termination of the dorsal horn than those from the motor cortex (Nyberg-Hansen and Brodal, 1963). A further analysis of the relative contribution of effects to primary afferents from motor and sensory areas is obviously of interest. There may be other mechanisms whereby facilitation of reflex paths can be evoked from higher centres. The foremost possibility would be through hyperpolarization of the terminals of primary afferents whereby the central actions evoked from these afferents would be increased. In the next section it will be shown that a positive DRP representing a primary afferent hyperpolarization can be evoked from the brain stem (Lundberg and Vyklicki, 1963b). The possibility that there may be a hyperpolarizing synaptic action in the terminals of primary afferents cannot be excluded, but a more likely alternative is probably inhibition of spinal reflex paths to primary afferents and a cessation of a tonic reflex depolarization. Through operation of this mechanism, reflex paths to motoneurones may be facilitated. Likewise the Ia pathways to motoneurones may be facilitated from the corticospinal tract through removal of primary afferent depolarization in the central terminals of Ia afferents (Lundberg and Vyklickf, 1963a). 11. I N H I B I T I O N O F R E F L E X P A T H S T O M O T O N E U R O N E S

A N D PRIMARY AFFERENTS

It is well known that in decerebrate cats the flexor reflex increases markedly after transection of the spinal cord (Sherrington and Sowton, 1915; Forbes et al., 1923). This holds true also for the inhibitory component of the reflex (Ballif et al., 1925). Fulton (1926) was the first to suggest that the release was due to disappearance of inhibitory action on interneurones mediating the reflex. Electrophysiological investigations with stimulation of the brain stem revealed that, independent of changes in motoneuronal excitability, polysynaptic transmission to motoneurones can be inhibited (Kleyntjens et al., 1955; Hugelin, 1955; Lindblom and Ottosson, 1955). Following References p . 217-219

208

A. L U N D B E R G

the investigations by Job (1953) our own approach, as reported in section (a) has been to compare synaptic actions in motoneurones and primary afferents of decerebrate and spinal cats. I n section (b) it will be reported that stimulation of the brain stem can evoke primary afferent depolarization and hence presynaptic inhibition of transmission in reflex paths. Finally in section (c) it will be shown that transmission in reflex paths can be inhibited from the brain stem also at an interneuronal level. ( a ) Tonic decerebrate inhibition

In experiments with conditioning of monosynaptic reflexes and with intracellular recording it was shown that there is a very effective tonic inhibition in the decerebrate state of polysynaptic transmission to motoneurones from some primary afferent A

1.07

1.18

&

C d

I4e 15.3

-rrnrnrrrrnrnr msec Fig. 11. Decerebrate inhibition of transmission from high threshold muscle afferent to motoneurones. Intracellular recording (upper traces) was performed from many PBSt motoneurones in a decerebrate cat before and after a spinal transection. Lower traces are from the dorsal root entry zone. The PBSt nerve was stimulated and the strengths are indicated in the records in multiples of threshold strengths from the nerve. Records A and H were obtained before transection of the spinal cord and volleys in high threshold muscle afferent do not evoke an EPSP (F-H). Records 1-0 were obtained from another PBSt rnotoneurone after transection of the cord and illustrate the characteristic excitatory action evoked from high threshold afferents in the spinal state (L-P) (Eccles and Lundberg, 1959b).

system (Eccles and Lundberg, 1959b). This holds true for the FRA as is illustrated for the excitatory actions from high threshold muscle afferents in Fig. 11. The inhibitory paths from the FRA are likewise very effectively inhibited. There is a similar tonic inhibition of the Ib excitatory and inhibitory pathways but not of the inhibitory pathway from Ia afferents. Kuno and Per1 (1960) have confirmed these results and

C O N T R O L O F T R A N S M I S S I O N I N REFLEX P A T H S

209

have raised the question if this descending inhibition could be an occlusion due to maximal descending activation of the reflex paths to motoneurones. This explanation is excluded by the intracellular recordings from motoneurones showing that the resting synaptic bombardment is smaller in the decerebrate than in the spinal state (Eccles and Lundberg, 1959b). Experiments with spinal cord lesions revealed that the responsible descending paths are located in the dorsal part of the lateral funicle, DLF, (Fig. 12).The tonic control is maintained as long as either DLF is intact(Ho1mqvist and Lundberg, 1959); a bilateral effect is exerted from each side. Thecentres responsible for this control are located to the medial ventral part of the medullary and lower pontine brain stem (Holmqvist and Lundberg 1961 ; Carpenter et al., unpublished). A very detailed study has been made of the release after brain stem lesions of different levels - a rostra1 lesion can give an almost complete release of the inhibitory ipsilateral and contralateral paths from the FRA with maintained control of the excitatory paths (Holmqvist and Lundberg, 1961 ; Holmqvist, 1961). The main implication of

Fig. 12. Location of descending pathways responsible for the decerebrate inhibition of reflex arcs. The curves were obtained in two experiments in which the monosynaptic test reflex from gastrocnemius-soleus (G-S) was conditioned by volleys in the nerve from flexor digitorum longus (FDL). Conditioned stimulus strengths are indicated in each graph and expressed in multiples of threshold strengths for the FDL nerve. The cat was decerebrate and the actions were investigated after the spinal lesions indicated. Observe that in the right graph there is no release from the decerebrate inhibition when the ipsilateral dorsal part of the lateral funicle is intact. In the left graph an ipsilateral hemisection does not give any release but following transection of the dorsal part of the lateral funicle on the contralateral side there is a complex release (Holmqvist and Lundberg 1959).

these findings has been with regard to the organization of the reflex paths from the FRA, which as disclosed by these experiments have both inhibitory and excitatory connections to the same motoneurones. It seems likely that the supraspinal control systems discussed can select whether volleys in the FRA shall evoke inhibition or References p . 217-219

210

A. L U N D B E R G

excitation in a given motor nucleus. Further attention should also be given to the possibility of a differential control of the reciprocal actions from the FRA to flexor and extensor motoneurones (cf. Holmqvist and Lundberg, 196 I). It is known that the dorsal root reflex can be inhibited from higher centres (Hagbarth and Kerr, 1954; Kleyntjens et a/., 1955). In recent experiments the DRPs have also been compared in the decerebrate and spinal states, the aim being to find out if there is similar decerebrate control of transmission to primary afferents (Carpenter, Engberg et al., 1963). Following transection of the cord there is no increase of the DRP that can be evoked from group Ia and Ib and from group Ib of extensors. However, there is no evidence of decerebrate inhibitory control of the paths from group I afferents to Ia, Ib and to cutaneous afferents (cf. Eccles et al., 1962c; Eccles et a[., 1963a, b). On the other hand with the DRPs evoked from the FRA there is a pronounced change after transection of the spinal cord as is illustrated for the effect from high threshold muscle afferents in Fig. 13. In the decerebrate state there is very

A

0ecereb.-

-

F

Spinal.

I .7

G

3.0

H

6.1

10.5rnV

I

15

10.5~1~ 60

50msec

Fig. 13. Dorsal root potentials evoked from high threshold muscle afferents in decerebrate and spinal cats. The upper traces were recorded from the most caudal dorsal rootlet in L6 and the lower traces from the dorsal root entry zone in L7. Single stimuli were given to the gastrocnemius-soleus nerve before and after spinal transection. Corresponding records in the upper and lower row wete obtained at the same stimulus strength, which is indicated between corresponding records and expressed in multiples of threshold strengths (Carpenter, Engberg ef a[., 1963).

little action (record D) the characteristic large DRPs appear only after transection of the cord. With the DRP evoked from cutaneous afferents the situation is more complex because it has two components (Fig. 14). The component I has short latency and a restricted distribution. It is large in rootlets adjacent to the zone of entry of the afferents evoking the effects, but in one segment rostra1 or caudal to this zone the magnitude is less than 25% of the maximal. Component I is of the same size in the decerebrate and spinal state. Component 11 on the other hand cannot be evoked in the decerebrate state, but is regularly found after transection of the cord (record D). It has a longer latency than component I and a much wider distribution along the cord, as was found by Bernhard (1953) for the deep P wave that is associated with component I1 and released with it. If a large component I is evoked in a rootlet there is very little additional component 11 (record C, Fig. 14). It is postulated that component I is evoked through a short-latency path supplied exclusively by cutaneous afferents (and reaching mainly cutaneous afferents). This path is apparently not subject to decerebrate inhibition. Component 11, on the other hand, is part of the flexor reflex actions and this path from the FRA to the FRA is subject to a very effective tonic inhibitory control in the decerebrate state. With respect to the cord

CONTROL OF TRANSMISSION I N REFLEX PATHS

A

B

mid L7

21 1

uppers1

Decereb.

mid L7

1.8 upper

SI

.#

-."

.I--.

.----._ 43

Fig. 14. Dorsal root potentials (DRP) evoked from cutaneous afferents in decerebrate and spinal cats. The DRPs (upper traces) were recorded from a dorsal rootlet in mid L7 (A and C) and upper Sl ( B and D). In all records the cord dorsum potentials (lower traces) were recorded from the dorsal root entry zone in L7. The superficial peroneal nerve (SP) was stimulated at a strength of 3.1 times threshold. The SP fibres enter the cord in upper L7 and the early component of the dorsal root potential that is identical in the decerebrate and spinal states, is large close to the zone of entry. In the lower diagram decerebrate and spinal DRPs are superimposed and shown at three different strengths of stimulation, which are indicated in multiples of threshold strength for the SP nerve (Carpenter, Engberg ef al., 1963).

dorsum potentials (CDP) there is following spinal transection not only an increase of the P waves that are associated with the DRPs evoked from FRA but also an increase of the negative CDP evoked from high threshold muscles and joint afferents. Of the effects evoked from cutaneous afferents the same holds true for the NZCDP but not for the N1 potential. These findings have been discussed by Carpenter, Engberg et al. (1963). What is the mechanism of this very powerful inhibition of some of the reflex paths to motoneurones and to primary afferents? Kleyntjens et al. (1953) as well as Eccles and Lundberg (1959b) postulated an inhibition of interneurones. With the demonstration that presynaptic inhibition is a physiological mechanism (Eccles, 1961) there is also the possibility of descending primary afferent depolarization to consider, and i n the next section it will be demonstrated that such an action can be evoked from the brain stem. (b) Primary afferent depolarization evoked from the brain stem Large DRPs can be evoked in the lumbar cord on stimulation of different regions of References p. 217-219

212

A. L U N D B E R G

A

Brain stem

B

C

PBSt

D

Sural

Fig. 15. Primary afferent depolarization evoked from the brain stem. The upper traces were recorded from a dorsal root filament in lower L6 and the lower traces from the dorsal root entry zone in lower L7. In A the brain stem was stimulated at the site shown in B. Observe that the positive potential recorded from the cord dorsum has two waves and that only the later one can be associated with the dorsal root potential. For comparison record C and D show the dorsal root potentials evoked by group I volleys in the nerve to PBSt and by a single volley in the sural nerve (Carpenter, Engberg et al., 1962).

% o f test 1751

125-

Brain stern+G-S o PBSt+G-S Ia

0

-

Ia

0

o

0

(L 0". 0

0

0

0 0.4 0

0

e

.

0

00

. 0

0 0.

loo-2.2

,

0 I

1

I

. I

Fig. 16. Effect of conditioning volleys from the brain stem of the presynaptic terminals of Ia fibres. The testing stimulus was delivered through a microelectrode inserted into the motor nucleus of gastrocnemius-soleus (G-S) to the site where the maximal Ia focal potential could be recorded. The test discharge was recorded in the nerve to G-S. The curves show the effect of group I volleys in the nerve fromposterior biceps-semitendinosus, PBSt, (0) and of stimulation of the brain stem ( 0 )at the site shown in Fig. 15 (Carpenter, Engberg et al., 1962).

C O N T R O L OF T R A N S M I S S I O N I N REFLEX P A T H S

213

the brain stem (Carpenter, Engberg et al., 1962). The DRP in Fig. 15 was elicited by a train of weak stimuli dorsally in the midline as marked in the Fig. Experiments with excitability measurements from the terminals of primary afferents revealed that a depolarization was elicited not only in Ib and cutaneous afferents but also in Ia fibres (Fig. 16). Central actions by impulses in these afferents can be presynaptically inhibited

B

A

-

100 msec

-

200 msec

Fig. 17. Dorsal root potentials (DRP) evoked from the brain stem. The DRPs (upper traces) were recorded (d.c. amplifier) from the most caudal dorsal rootlet in L6. The cord dorsum potentials (lower traces) from the L7 dorsal root entry zone (a.c. amplifier with 0.8 sec time constant). The site of brain stem stimulation is indicated for each record in the right diagram. In record C there is a wellmaintained plateau. A and B were obtained from the same site of stimulation and with higher strength of stimulation in B the depolarization decreases and a depolarizing rebound appears. The frequency of stimulation was 500/sec in A-C. In D the frequency of stimulation was decreased to 1lO/sec and the strength was decreased with the result that a positive DRP appears. Voltage calibration below A refers to the DRP. Time calibration below C is for A-B; record D was taken at the slower speed (Lundberg and Vyklickq, 1963b).

from the brain stem. The effects in Fig. 16 and 17 are produced through a pathway descending in the ventral quadrant of the spinal cord. At somewhat higher strength of stimulation large DRPs can also be evoked practically from everywhere in the brain stem. It is difficult to exclude synaptic activation of the descending systems described above, but presumably there is at least one other system descending from the brain stem that causes primary afferent depolarization in Ib and in cutaneous afferents but not in Ia afferents. These latter effects are mediated by pathways descending in dorsal as well as ventral parts in the ventrolateral funicles. Primary afferent depolarization can in the decerebrate cat be evoked from the most lateral strip of the intermediate region of the anterior cerebellum and also from the vestibular nerve (Carpenter et al., unpublished). The presynaptic inhibition evoked from the brain stem may have a more general physiological-significance than regulation of transmission through reflex arcs - there is in particular the possibility of a general regulation of the sensory input. It is neverReferences p:2I 7-21 9

214

A. L U N D B E R G

theless obvious that in any experimental analysis dealing with the control of transmission in reflex arcs it is of paramount importance to consider the actions on primary afferents. The presynaptic inhibition of reflex arcs that can be evoked by descending primary afferent depolarization (PAD) is powerful, but it is in my opinion not as effective as the decerebrate inhibition. It has been suggested by Andersen et a/. (1962) that the decerebrate inhibition is due to tonic primary afferent depolarization but the available evidence is in my opinion against this explanation. For example, the N1 CDP that presumably represents the monosynaptic excitatory action in second order cells has the same size and the same rate of rise in the decerebrate and spinal states. Hughes and Gasser (1934) found that the N1 is decreased during the P wave and had there been a tonic descending PAD in the decerebrate state the expected result of a spinal transection would have been an increase of the N1 potential (for further discussion of these problems cf. Carpenter, Engberg et a]., 1963). However, this evidence is indirect but in the next section it will be shown that inhibition of reflex paths not caused by primary afferent depolarization can be evoked from higher centres. (c)

Inhibition of reflex paths not caused by primary afferent depolarization

The first indication of an inhibition not being caused by primary afferent depolarization was found during the study of the PAD evoked from the brain stem. If more longlasting stimulation is employed than that used to obtain record A in Fig. 15 it is sometimes possible to obtain a fairly well-maintained steady depolarization as is shown in Fig. 17, C. From other regions of the brain stem the depolarization is less well-maintained during stimulation (record A) and when the strength of stimulation is increased (B) the amount of absolute depolarization actually decreases (records A and B were obtained from the same site of stimulation). After cessation of stimulation there is a depolarizing rebound. A comparison of records A and B leaves the impression that with the increased strength of stimulation a positive potential has been superimposed on the depolarizing negative. With changed parameters of stimulation, i n particular lower frequency, it was often possible to obtain a hyperpolarizing DRP without any sign of depolarization (record D). It is postulated that the positive DRP is caused by a primary afferent hyperpolarization. A synaptic action hyperpolarizing the membrane of the terminals cannot be excluded. Another explanation could be inhibition of reflex paths to primary afferents and cessation of a steady depolarization caused by spinal reflex action. In favour of the latter explanation is the finding that this stimulation of the brain stem did decrease the DRPs evoked from various afferent sources. In Fig. 18 the strength of brain stem stimulation was decreased so as to give no potential change in the dorsal root and cord dorsum recording (A and D). Stimulation of the brain stem gives a marked decrease of DRP evoked by , a volley in the sural nerve (rf. E and F). Recording was from the lowest L6 filament and it is mainly component I1 that undergoes a decrease; records B and C show that component I evoked from the-superficial peroneal nerve is not so much changed. There was likewise a very marked inhibition of the DRPs evoked from high threshold muscle afferents. It can therefore be concluded that the path from the FRA to the

21 5

CONTROL OF TRANSMISSION I N REFLEX P A T H S

A BS

200 msec

D BS

BS+sural

sural

-_

c--LccII

h h

BS*Sur DRP %

Of

lest

40;

0

Illm loohec

ibo

260

3dn

4bO

560

Bbmrac

Fig. 18. Inhibition from the brain stem of reflex paths to primary afferents. The recording and stimulating conditions are described in the legend of Fig. 17. Stimulation of the brain stem (BS) was adjusted with respect to frequency and strength so as to give no dorsal root potential (DRP) or cord dorsum potential (record A and D). B and E show the effect of the single volley in the superficial peroneal nerve, SP, and in the sural nerve respectively. Conditioning stimulation of the brain stem decreases markedly the DRP from the sural nerve (F) but not so much the DRP from the SP nerve (record C). The inhibitory effect is exerted mainly on component 11. The graph shows the time cours of the DRP depression (Lundberg and Vyklicky, 1963b).

FRA is effectively inhibited by this descending system. Since component I and the FRA action to a large extent are evoked in the same fibres, it seems extremely unlikely that the effect can be exerted on the receptive fibres by some mechanism not changing the membrane potential. Hence it is postulated that there is an inhibition at an interneuronal level. Fig. 19 shows that also the DRP evoked from volleys in Ia afferents may be inhibited. This DRP represents the PAD in Ia afferents (Eccles et al., 1962c) and the result is taken to indicate that also the pathway to Ia afferents can be inhibited. The same holds true for the pathway from Ib afferents of extensor muscles (E and F) which give PAD to Ib afferents and to cutaneous afferents (Eccles el al., 1963a, b). These experiments will be supplemented with studies of the effect on the presynaptic inhibition of central actions from the different afferents.

Comments As already pointed out in the introduction the inhibitory control of reflex paths presents a more complex picture than the facilitatory. One reason is that inhibition can be evoked both by primary afferent depolarization and at an interneuronal level. By the former mechanism, brain stem centres can give presynaptic inhibition not only of central actions from cutaneous and Ib afferents but also of these from the l a References p . 21 7-219

216

A. L U N D B E R G

A

BS It

I ' 0.2rnV

C PBSt-al

2 8s PBSt

4

D

E

F

I msec

200 msec

Fig. 19. Inhibition from the brain stem of reflex paths from group I muscle afferent to primary afferents. These records were obtained in the same experiments as those in Fig. 18. The PBSt nerve displayed separation in Ia and Ib volleys (record D). The dorsal root potential (DRP) in B was evoked by submaximal stimulation of Ia fibres and in C this D R P is depressed during stimulation of the brain stem. In the corresponding lower records A and F the effect on the D R P evoked by a train of maximal group I volleys in the nerve from gastrocnemius-soleus (G-S) is shown (Lundberg and Vyklickf, 1963b).

afferents. It can be assumed that this inhibition pertains both to the reflex paths to primary afferents and to motoneurones - it is a general adjustment of the input. A fine inhibitory regulation would require an action on interneurones whereby a particular reflex path can be acted upon in isolation. It is therefore of some principle interest that on stimulation of the brain stem it is possible to evoke an inhibition of reflex paths that is not due to primary afferent depolarization but presumably is exerted at an interneuronal level (Figs. 18 and 19). So far this has been strictly demonstrated only for the paths to primary afferents. By this mechanism presynaptic reflex inhibition of central action from different afferent systems, among them la, can be regulated. The existence of this regulatory mechanism gives emphasis to the functional importance of reflex paths to primary afferents. In these investigations our starting point was the decerebrate tonic inhibition of reflex paths to motoneurones. This descending inhibition is among other reasons of interest for its great effectiveness but is, as has been discussed above, difficult to investigate in detail with respect to its mechanism. The available evidence suggests that it is not caused by a primary afferent depolarization. In my opinion descending systems evoking postsynaptic inhibition in interneurones (cf. Lundberg and Voorhoeve, 1961 and Fig. 10) or presynaptic inhibition at an interneuronal level (cf. Lundberg and Vyklickf, 1963a) should be looked for. Further investigations along these lines must also take into account the finding of a differential release of excitatory and inhibitory paths to motoneurones from the decerebrate supraspinal control (Holmqvist and Lundberg, 1961).

C O N T R O L OF T R A N S M I S S I O N I N REFLEX P A T H S

217

A final comment concerns the functional significance of reflex paths to primary afferents and to motoneurones. In the electrophysiological analysis of these paths and of their control from higher centers it has been an advantage to consider them as separate entities with different final common paths. It should, however, be kept in mind that in reflex regulation of movement the actions on primary afferents are subsidiary to the actions on motoneurones. A certain degree of parallelism has been found both with respect to the facilitatory and inhibitory control of these reflex paths but it is obvious that with either of these descending actions, the effects exerted on the paths to primary afferents and on the paths to motoneurones have an entirely different physiological significance. SUMMARY

The report deals with the supraspinal control of transmission in reflex paths to motoneurones and to primary afferents. Part I : (a and b) Transmission from different afferent systems to motoneurones and primary agerents can be facilitated from the corticospinal tract. (c) The effect is caused by an excitatory action on interneurones of these spinal reflex paths. Part 11: (a) There is tonic inhibition in the decerebrate state of transmission from some primary afferent systems to motoneurones and to primary afferents. (b) Electtical stimulation of the brain stem can inhibit spinal reflex paths by evoking primary afferent depolarization. (c) Stimulation of the brain stem can inhibit transmission in spinal reflex paths to primary afferents at an interneuronal level. REFERENCES P., ECCLES,J. C., AND SEARS,T. A., (1962); Presynaptic inhibitory action of cerebral ANDERSEN, cortex on the spinal cord. Nature (Lond.), 194, 740-741. ARAKI,T., ECCLES, J. C., AND ITO, M., (1960); Correlation of the inhibitory postsynaptic potential of motoneurones with the latency and time course of inhibition of monosynaptic reflexes. J . Physiol. (Lond.), 154, 354-377. BALLIF,L., FULTON,J. F., AND LIDDELL, E. G., (1925); Observations on spinal and decerebrate knee-jerks with special reference to their inhibition by single break-shocks. Proc. roy. Soc., 98, 589-607. BARRON, D. H., AND MATTHEWS, B. H. C., (1938); The interpretation of potential changes in the spinal cord. J. Physiol. (Lond.), 92,276-321. BERNHARD, C. G., (1953); The spinal cord potentials in leads from the cord dorsum in relation to peripheral source of afferent stimulation. Acta physiol. scand., 29, Suppl. 106, 1-29. BERNHARD, C. G., BOHM,E., AND PETERSEN, I., (1953); Investigation oil the organization of the corticospinal system in monkeys. Acta physiol. scand., 29, Suppl. 106, 79-135. CARPENTER, D., ENGBERG, I., FUNKENSTEIN, H., AND LUNDBERG, A., (1963); Decerebrate control of reflexes to primary afferents. Acta physiol. scand., 57, In the press. CARPENTER, D., ENGBERG, I., AND LUNDBERG, A., (1962); Presynaptic inhibition in the lumbar cord evoked from the brain stem. Experientia (Basel), 18, 450-451. CARPENTER, D., LUNDBERG, A., AND NORRSELL, U., (1962); Effects from the pyramidal tract on primary afferents and on spinal reflex actions to primary afferents. Experientia (Basel), 18,337-338. CARPENTER, D., LUNDBERG, A., AND NORRSELL, U., (I 963); Primary afferent depolarization evoked from the sensorimotox cortex. Acta physiol. scand., 59, 126-142.

218

A. L U N D B E R G

ECCLES, J . C., (1961); The mechanism of synaptic transmission. Ergbn. fhysiol., 51, 299-430. ECCLES,J. C., ECCLES,R. M., AND MAGNI,F., (1961); Central inhibitory action attributable to presynaptic depolarization produced by muscle afferent volleys. J . fhysiol. (Lond.), 159, 147-166. ECCLES,J. C., FATT,P., A N D LANDGREN, S., (1956); The central pathway for the direct inhibitory action of impulses in the largest afferent nerve fibres to muscle. J . Neurophysiol., 19, 75-98. ECCLES,J. C., KOSTYUK, P. G., A N D SCHMIDT, R. F., (1962a); Central pathways responsible for depolarization of primary afferent fibres. J . fhysiol. (Lond.), 161, 237-257. ECCLES, J. C., KOSTYUK,P. G., AND SCHMIDT, R. F., (1962b); Presynaptic inhibition of the central actions of flexor reflex afferents. J . Physiol. (Lond.), 161, 258-281. ECCLES, J. C., MAGNI,F., A N D WILLIS,W. D., (1962~);Depolarization of central terminals of group I afferent fibres from muscle. J . Physiol. (Lond.), 160, 62-93. ECCLES,J. C., SCHMIDT, R. F., AND WILLIS,W. D., (1963a); Depolarization of central terminals of group Ib afferent fibres of muscle. J . Neurophysiol., 26, 1-27. ECCLES, J. C., SCHMIDT, R. F., AND WILLIS, W. D., (1963b); Depolarization of the central terminals of cutaneous afferent fibers. J. Neurophysiol., 26, 646-661. ECCLES, R. M . , AND LUNDBERG, A., (1958); The synaptic linkage of ‘direct’ inhibition. Actaphysiol. scand., 43, 204-21 5. ECCLES,R. M., AND LUNDBERG, A., (1959a); Synaptic actions in motoneurones by afferents which may evoke the flexion reflex. Arch. ital. Biol.,97, 199-221. ECCLES, R. M., A N D LUNDBERC, A,, (1959b); Supraspinal control of interneurones mediating spinal reflexes. J . Physiol. (Lond.), 147, 565-584. EIDE,E., LUNDBERG, A., AND VOORHOEVE, P., (1961); Monosynaptically evoked inhibitory postsynaptic potentials in motoneurones. Acta physiol. scand., 53, 185-195. ENGBERC, I . , (1963a); Effects from the pyramidal tract on plantar reflexes in the cat. Actapliysiol. scand., 59, Suppl. 213, 38. ENGBERG, I., (1963b); Plantar reflexes in cats. Experientia (Basel), 19, 487-488. FORBES, A., COBB,S., A N D CATTELL, H., (1923); Electrical studies in mammalian reflexes. 111. Immediate changes in the flexion reflex after spinal transection. Amer. J . fhysiol., 63, 30-44. FRANK, K., A N D FUORTES, M. G. F., (1957); Presynaptic and postsynaptic inhibition of monosynaptic reflexes. Fed. Proc., 16, 39-40. FULTON, J. F., (1926); Muscular Contraction andthe Reflex Controlof Movenient. Baltimore, Williams and Wilkins p. 644. GRANIT,R., (1955); Receptors and Sensory Perception. New Haven, Yale University Press. HAGEARTH, K . E., (1952); Excitatory and inhibitory skin areas for flexor and extensor motoneurones. Acta physiol. scancl., 26, Suppl. 94. 1-58. HAGBARTH, K. E., A N D KERR,D. I. B., (1954); Central influences on spinal afferent conduction. J . Netrropliysiol., 17, 295-307. HOLMQVIST, B., (1961); Crossed spinal reflex actions evoked by volleys in somatic afferents. Acta physiol. scand., 52, Suppl. 181. 1-67. HOLMQVIST, B., AND LUNDBERG, A., (1959); On the organization of the supraspinal inhibitory control of interneurones of various spinal reflex arcs. Arch. ital. Biol., 97, 340-356. HOLMQVIST, B., AND LUNDBERG, A,, (1961); Differential supraspinal control of synaptic actions evoked by volleys in the flexion reflex afferents in a-motoneurones. Acta physiol. scand., 54, SUPPI. 186. 1-51. HUGELIN, A,, (1955); Analyse de l’inhibition d’un reflexe nociceptif (rkflexe linguomaxillaire) lors de I’action du systeme reticulo-spinal dit ‘facilitateur’. C. R. SOC.Biol. (Paris), CXLIX, Novembre, p. 1893. HUGHES,J., AND GASSER, H. S., (1934); The response of the spinal cord to two afferent volleys. Anier. J. Physiol., 108, 307-321. Joe, C., (1953); Uber autogene Inhibition und Reflexumkehr bei spinalisierten und decerebrierten Katzen. Pfliigers Arch. ges. Physiol., 256, 406-418. KLEYNTJENS,F., KOIZUMI, K., AND BROOKS, C. McC., (1955); Stimulation of suprabulbar reticular formation. Arch. Neurol. Psychiat., 73, 425438. KUCELBERG, E., EKLUND, K., AND GRIMBY, L., (1960); An electromyographic study of the nociceptive reflexes of the lower limb. Mechanism of the plantar responses. Brain. 83, 394-410. KUNO,M., A N D PERL,E. R., (1960); Alteration of spinal rcflexes by interaction with suprasegmcntal and dorsal root activity. J . Physiol. (Lond.), 151, 103-122. KUYPERS, H. G. J. M., (1960); Central cortical projections to motor and somatosensory cell groups. Brain, 83, 161-1 84.

CONTROL OF TRANSMISSION I N REFLEX PATHS

219

LANDGREN, S., PHILLIPS, C. G., AND PORTER, R., (1962a); Minimal synaptic actions of pyramidal impulses on some a-motoneurones of the baboon’s hand and forearm. J . Physiol. (Lonrl.), 161, 91-111. LANDGREN, S., PimLIPs, C. G., A N D PORTER, R., (1962b); Cortical fields of origin of the monosynaptic pyramidal pathways to some a-motoneurones of the baboon’s hand and forearm. J . Physiol. (Lond.), 161, 112-125. LEKSELL, L., (1945); The action potential and excitatory effects of the small ventral root fibres to skeletal muscle. Acta physiol. scand., 10, Suppl. 31, 1-84. LINDBLOM, U. F., AND OTTOSSON, J. O., (1955): Bulbar influence on spinal cord doisum potentials and ventral root reflexes. Acta physiol. scanrl., 35,203-214. LINDBLOM, U. F., A N D OTTOSSON, J.O., (1956; 1957); Influence of pyramidal stimulation upon the relay of coarse cutaneous afferents in the dorsal horn. Acta physio/. scand., 38, 309-318. LIVINGSTON, A,, AND PHILLIPS, C. G., (1957); Maps and thresholds for the sensorimotor cortex of the cat. Quart. J . exp. Physiol., 42, 190-205. LLOYD,D. P. C., (1941); The spinal mechanism of the pyramidal system in cats. J . Neurophysiol.,4, 525-546. LUNDBERG, A., NORRSELL, U., A N D VOORHOEVE, P., (1962); Pyramidal effects on lumbo-sacral interneurones activated by somatic afferents. Acta physiol. scand., 56, 220-229. LUNDBERG, A,, AND VOORHOEVE, P., (1961); Actions on interneurones a t activation of supraspinal systems controlling the transmittability of spinal reflex arcs in the cat. Acta physiol. pharmacol. need., 10, 11-35. LUNDBERG, A., AND VOORHOEVE, P., (1962); Effects from the pyramidal tract on spinal reflex arcs. Acta physiol. scand., 56, 201-219. LUNDBERG, A,, A N D V Y K L I C KL., ~ , (1963a); Inhibitory interaction between spinal reflexes to primary afferents. Experientia (Basel), 19, 247-248. LUNDBERG, A., AND V Y K L I C KL., ~ ,(1963b); Brain stem control of reflex paths to primary afferents. Acta physiol. scand., 59, Suppl. 21 3 , 9 I . NYBERG-HANSEN, R., AND BRODAL, A,, (1963); Sites of termination of corticospinal fibers in the cat. An experimental study with silver impregnation methods. J . comp. Neurol., 120, 369-391. PAINTAL,A. S., (1961); Participation by pressure-pain receptors of mammalian muscles in the flexion reflex. J . Physiol. ( L o n d . ) , 156, 498-514. SHERRINGTON, C. S., (1947); The integrative Action of the nervous System. 2nd ed., Cambridge University Press pp. 290-291. SHERRINGTON, C. S., AND SOWTON,S. C. M., (1915); Observations on reflex responses to single break-shocks. J. Physiol. (Lond.), 49, 331-348. TOWER,S . S., (1935); The dissociation of cortical excitation from cortical inhibition by pyramid section and the syndrome of that lesion in the cat. Brain, 53,238-254. WALL,P. D., (1958); Excitability changes in afferent fibre terminations and their relation to slow potentials. J. Physiol. (Lond.), 142, 1-21. DISCUSSION

WALL: When you were speaking about the pathways necessary to maintain the decerebrate state and you were somewhat puzzled about the importance of the dorsal lateral columns, do you think that there is any possibility that also afferent pathways are involved in maintaining the decerebrate state?

LUNDBERG:Our experiments with spinal cord lesions have demonstrated that the descending paths are located in the dorsal part of the lateral funicles. It is probable that activity in ascending pathways regulates the descending control of reflex paths in the intact animal: the tracts activated from the FRA are of particular interest in this respect. We do not have decisive evidence with regard to your question if the descending decerebrate control depends on ascending activity, but the available evidence does not support this hypothesis. The control remains after transection of the

220

DISCUSSION

ventral half of the spinal cord and also after ablation of the cerebellum. This suggests that activity in ventral ascending spinal tracts and in direct spinocerebellar tracts is not required in maintenance of the decerebrate state. KUYPERS : It is extremely difficult to look at your results and try to translate them into an anatomical language. I would suggest that in your experiments you are dealing again with a descending pathway which in its course hooks onto a propriospinal system, located in the lateral funiculus. This seems to me the most likely explanation of your physiological experiments. LUNDBERG: This explanation is attractive in view of the fact that anatomical investigations have failed to show a reticulospinal tract in the dorsal part of the lateral funicle. However, it has recently been shown that axons from nucleus raphe magnus descend in this part of the cord (Brodal, Taber and Walberg, 1960). The centres responsible for the tonic decerebrate control of transmission in reflex paths are located in the ventromedial part of the caudal brain stem, a region which contains the nucleus raphe magnus. WIESENDANGER: You showed that facilitation of the flexor reflex disappeared completely after section of the pyramids. This may be somewhat contrary to what Dr. Kuypers showed, that there are two systems, Bl and B2 operating with about the same effect. Could your results be explained by the fact that the investigations were done shortly after the operation and that some time is needed until the other systems come in operation? LUNDBERG: In the anaesthetized preparation the effects that can be evoked from the sensorimotor cortex at a moderate strength of stimulation seem to be transmitted entirely by the cortico-spinal tract. At higher strength of stimulation effects are mediated to the spinal cord also via other routes. So far our analysis has been confined to the effects evoked via the cortico-spinal tract. It is an interesting suggestion that the extrapyramidal route may be more effectively activated from the cortex some time after pyramidectomy. If so, this preparation could be utilized for a systematic analysis of the extrapyramidal actions. WALL: I would like to ask about this quite dramatic picture where yolv showed the dorsal root potential being evoked by a cutaneous volley, not being changed by spinalization in the segment, but a quite interesting change in neighbouring segments. I think the dorsal root potential in the neighbouring segments is produced by two different mechanisms, one by afferents which get into neighbouring segments and produce the fast component, and then a later component which comes in by spread through interneuronal pathways. A good demonstration of that is to look on the opposite side of the cord where you have only in the higher segments a slow component and not the fast one. What I wanted to ask is: this mechanism, whatever it produces, that is one that is

C O N T R O L OF T R A N S M I S S I O N I N REFLEX PATHS

22 1

very easily saturated. With 10% of the A input in the root you can produce the maximum dorsal root potential. Could it have been that the reason why you did not see a change in the root, in the segment that you were firing, was that it was already saturated, that you could not produce a larger dorsal root potential?

LUNDBERG:I think the situation is very different when you stimulate peripheral nerves. Component I of the dorsal root potential evoked from cutaneous nerves reaches its maximum only when the strength of stimulation is 2-4 times threshold for the nerve. Hence most of the low threshold cutaneous afferents do contribute to it. Furthermore, at some distance from the zone of maximal afferent entry component I is very small and in this case there cannot be the question of a saturated line. ECCLES:I wish to ask about the cortical areas from which you have been able to elicit dorsal root potentials in the spinal cord. We find that both somatosensory areas I and I1 are effective, but that I is mostly contralateral whereas 11 has much the same effectiveness on both sides. I would like particularly to know if you have studied the pre- and post-cruciate areas because it would be good to have corroboration of our finding that the pre-cruciate area is effective by way of the post-cruciate, being inactive when the post-cruciate is ablated. LUNDBERG: I agree with your finding that dorsal root potentials also can be evoked from the pre-cruciate area but in our experience only if the strength of stimulation is raised above the required to evoke an effect from post-cruciate region of the somatosensory area I. The area in the latter region from which dorsal root potentials can be evoked at threshold stimulation is that from which reflex paths to motoneurons can be facilitated. We have made no systematic investigation of effects from the somatosensory area I1 and have not made ablations in the cortex. In our experience bilateral effects can be produced also from area I. Fig. 4 of the manuscript shows that each pyramidal tract has bilateral effect in the lumbar cord.

222

The Pyramidal Projection to Motoneurones of Some Muscle Groups of the Baboon’s Forelimb C. G. P H I L L I P S

AND

R. P O R T E R

University Laboratory of Pliysiology, Oxford (Great Britain)

INTRODUCTION

We shall describe experiments on colonies of cortico-spinal neurones which form synapses upon a motoneurones in the opposite side of the baboon’s cervical spinal cord. Abundant evidence for the existence of such synapses in the primates has been furnished by neuro-anatomical investigation (see Kuypers’ article in Vol. 11 of the Progress in Brain Research series) and by physiological experiment both on the forelimb (Cooper and Denny-Brown, 1927; Bernhard and Bohm, 1954), and on the hindlimb (Bernhard er al., 1953; Preston and Whitlock, 1960, 1961). Bernhard aiid Bohm (1 954) state that the monosynaptic cortico-motoneuronal response was ‘generally more pronounced’ in the forelimb than in the hind. A colony is defined as all those pyramidal neurones making monosynaptic connexion with a single u motoneuroiie (Landgren et a/., 1962b). Thus defined, the colony is, for the purpose of the experimentalist, a convenient abstraction from the complex totality of the cortico-spinal system, which also controls the iiiterneurones of reflex arcs, the fusimotor neurones, aiid the neurones of ascending spinal systems. It should be assumed, indeed, in the absence of evidence to the contrary, that the axons of the colony may send branches to some or all of these other cell-types, aiid also to other ( I motoneurones. Yet the justification for the concept of the colony does go beyond the mere methodological convenience that such colonies can be readily isolated by intracellular recording from the somas of (1 motoneurones. For the reflex management of muscle is in terms of the motor unit (Sherrington, 1931), and if the brain manages muscle in similar fashion, the colonies of pyramidal cells controlling the motor units become significant functional groupings with which intracortical mechanisms, lying upstream of the corticofugal pyramidal neurones, can operate. Hern et a/. (1962) and Landgren et al. (1962a,b) confined their experiments to colonies projecting to sample motoneurones of the hand and forearm, because the range and precision of hand movement in the baboon led them to expect a powerful cortical command of these final common paths. They found that pyramidal neurones could be selectively and directly stimulated by a focal surface anode (see also Phillips and Porter, 1962), and they used brief (0.2 msec), weak (0.4-3.1 mA) pulses to evoke well-synchronised pyramidal discharges. These discharges gave rise to minimal

THE PYRAMIDAL PROJECTION TO MOTONEURONES

223

monosynaptic excitatory actions (EPSP) which they recorded intracellularly from the test motoneurones. These were followed in some cases by inhibitory synaptic actions (IPSP) beginning 1.2-1.4 msec after the start of the EPSP. Preston and Whitlock (1 961), the pioneers of intracellular recording of pyramidal synaptic actions, had found the same excitatory-inhibitory sequence in lumbar motoneurones. Landgren et al. (1962b) assumed that similar-sized cortical shocks excited similarsized populations of pyramidal cells. But they found that the colonies, or parts of colonies, contained within these populations commanded different quantities of monosynaptic excitatory action on different test motoneurones (Landgren et al.. 1962b, Fig. 2). This was evidence of different densities of monosynaptic connexion; but since the stimuli were very weak, there was no certainty that the populations were large enough to include whole colonies. The new experiments we now publish have used stronger stimuli in order to try to measure the maximum quantity of monosynaptic excitatory action commanded by the colony projecting to each test motoneurone, and in order to compare the quantities for different colonies. We have also looked at motoneurones of proximal muscle groups in order to compare their colonies with those of the forearm and hand. We have made intracellular recordings from 266 motoneurones in 26 baboons, using techniques described in detail by Hern et al. (1 962). The motoneurones were identified and classified by antidromic stimulation of the ulnar, median, posterior interosseous, musculocutaneous and triceps nerves, which innervate groups of muscles which, in terms of function, are sufficiently differentiated not to require further subdivision for the purposes of a first comparison of quantities of cortico-motoneuronal control. We assume that the method is capable of detecting the smallest synaptic action on the motoneurone membrane. The colonies projecting to test motoneurones cannot be stimulated in isolation. Cortical shocks excite populations of corticofugal neurones; these populations include parts and wholes of colonies, but must also include neurones projecting to other cortical areas and to basal structures. The first task is to assess the relative sizes, and the spatial extents (horizontally and in the depth of the Rolandic fissure), of the populations of cortico-spinal neurones excited by surface-anodal shocks of measured strengths. Such a n assessment is needed as a guide to the probable spatial extents of the colonies. Evidence that the shocks have been strong enough to reach cells in the depth of the Rolandic fissure is also necessary if one is to assert in any case that no monosynaptic connexion from the precentral gyrus to a test motoneurone exists. STIMULATION OF CORTICO-SPINAL POPULATIONS

Fig. 1 shows that the dorsolateral white matter of the baboon’s cervical cord contains enough cortico-spinal fibres of homogeneous conduction velocity to give a sharp ‘tract wave’ (Landgren et al., 1962a) in response to a brief shock to the cortical arm area. The histogram, which shows the time of arrival of impulses in 68 single corticospinal fibres at the same level (C 5-6) in six baboons of similar size, makes it clear Reference7 p . 242

224

C. G. P H I L L I P S A N D R. P O R T E R

that the majority of the impulses reach this level at the time of the negative crest of the tract wave (negative deflexions are downwards in all records). The tract wave can be recorded from a fine enamelled silver wire resting lightly on the dorsolateral surface of the cord, and Fig. 2 shows its amplitude as a function of cortical stimulus strength. At strengths above about 4.5 mA, the tract wave is followed by 'I waves' (Patton and Amassian, 1954). The plotted curve shows that the tract wave reaches a

0

I

2

3

msec

Fig. 1. Above: wave recorded from lateral cortico-spinal tract of baboon at C5-6 level with silverfilled microelectrode. Positive deflexion upwards; about 20 superimposed traces. Cortical stimulation by surface anode, 0.2 msec, 2.8 mA. Below: times of arrival of impulses in 68 single cortico-spinal axons in response to surface-anodal stimuli in six experiments. Time scale applies to both parts of figure. (From Landgren et al., 1962a.)

maximum amplitude at strengths over 6.0 mA. Even the strongest of these shocks (repeated at 2 cis) caused no visible or palpable muscular response in any part of

Fig. 2. Right of figure shows increasing cortico-spinal discharge in response to increasing surfaceanodal stimulation of a point on the right precentral gyrus, midway between the superior and inferior precentral fissures. Superimposed records made from bare tip of enamelled 80 p silver wire applied to dorsolateral surface of left side of cord, caudal to lamina of C2 vertebra. Positivity upwards. Time in msec. For measurements see left of figure; amplitude of tract wave (ordinates) plotted against strength of 0.2 msec current pulses to cortex.

THE PYRAMIDAL PROJECTION TO MOTONEURONES

225

the body. The measured amplitudes of thc tract wave are a function of the size of the responding cortico-spinal population; these amplitudes remained remarkably constant from the same point on the cord surface throughout many hours, provided that the shunting layer of cerebrospinal fluid was removed from under the covering paraffin pool at the times of recording. The growth of the population with increasing stimulation depends on physical

II

..

.. .. . . : ; : r

I

. *.

1

I0

:.

.

:

:

.

:

O-

IFaceward b' ' I

I

I

5

I

I

'

I

0

I

' ' I

' '

Legward

I

5

'

I

rnm

Fig. 3. Spread of surface-anodal stimuli horizontally in the cortex. Curves of cortico-spinal fibre thresholds (ordinates) plotted against distances from lowest-threshold points on cortex: (A) at right angles to central fissure; (B) along precentral gyrus. Horizontal lines show extent of outline maps drawn at the indicated strengths, for other fibres for which curves were not plotted. (Modified from Landgren et al., 1962b.) Refrrences p . 242

226

C. G. P H I L L I P S A N D R. P O R T E R

spread of stimulating current both horizontally along the precentral gyrus, and in depth down the anterior Rolandic wall. Fig. 3 illustrates horizontal spread, and measures its extent at strengths up to 3.0 mA. The curves were made by measuring the threshold for single cortico-spinal units at points along lines ‘parallel’ and at right angles to the Rolandic fissure. The latency of response was the same at every point, indicating physical spread (for comparison with physiological spread, cf. Phillips and Porter, 1962). Spread of stimulus in depth is illustrated in Figs. 4 and 5. For these experiments sharpened enamelled silver wires (diameter 0.2 mm) were thrust free-hand through the post-central gyrus and into buried precentral cortex. A bead of sealing-wax limited the depth of penetration. The wire was free of external attachments which

I

I

Fig. 4. Spread of surface-anodal stimuli in anterior wall of Rolandic fissure. Right: tracing of forrnalinhardened slice of cortex, cutting right Rolandic fissure at right angles, between superior and inferior precentral fissures. Scale (mm) makes no allowance for shrinkage. Bare, pointed tip of 200 ,u diameter enamelled silver needle lies in buried precentral cortex, having been inserted through postcentral gyms during life. Dotted line marks boundary between white and grey matter. Needle moves with pulsating brain; a bead of sealing wax prevents deeper penetration. Buried cortex can be stimulated by applying spring-mounted contact to bare wire on top of bead. Lefr: superimposed records of responses of single axon in left lateral cortico-spinal tract to just-supraliminal cortical stimuli (0.2 msec pulses): 2.5 mA to surface anode (see diagram), 0.28 mA anodal pulse at buried electrode, 0.36 mA cathodal pulse at buried electrode. Note tract wave in response to surface stimulation, its 0.28 rnA and small size at -0.36 mA. virtual absence at

+

might drag on it and damage the cortex, and it moved freely with the pulsating brain. Stimulation at its bare tip could be effected by applying a spring-mounted stigmatic electrode to the bare wire above the bead. In one type of experiment, a single cortico-spinal axon was found, by laborious probing of the dorsolateral cervical cord, which responded to ‘buried’ stimulation at a very low threshold (Fig. 4). The strength needed at a point on the convexity of the precentral gyrus was then determined. The threshold was nearly ten times higher at the surface than at the buried point; the upper record shows that the surface shock excited a good tract wave and that the responding population included the unit whose low threshold at the ‘buried’ point attested its nearness thereto. An experiment with populations instead of with single units was also made in the experiment whose geometry is drawn in Fig. 4. The first step was to establish that a

THE P Y R A M I D A L PROJECTION T O MOTONEURONES

227

population could be stimulated by the buried electrode at such low strength as to make it likely that this population was very near to that electrode. Fig. 5 4 b, c illustrates the response to 0.4 mA cathodal (led from the surface of the cord), and measures the

a 1

b

fl-

1 .o 1.7

C

2001

1.8

P

3 .O

d

f

uv 3.2

Fig. 5. Spread of surface-anodal stimuli in anterior wall of Rolandic fissure. For geometry see Fig. 4. Tract waves recorded as in Fig. 2. Left: a, b, c: Paired 0.2 msec stimuli to buried cathode, 0.4 mA. The responding population of axons was refractory at 0.57 (interval between shocks in cis 0.53 msec). Time in msec. d, e , f : First shock to surface anode, 1.8 mA; second to buried cathode, 0.45 mA. The ‘buried’ population is not made refractory by the surface stimulus. Right: surface-anodal shock (S +) precedes ‘buried’ cathodal shock at fixed interval of 0.65 msec. Increasing strengths of S + noted in margin. Top record is control response to ‘buried’ stimulus; bottom record is control response to S + , 3.2 mA. Surface stimulation at this strength makes the ‘buried’ population refractory.

population’s refractory period. The second step is to excite a similar-sized population by the surface anode (1.8 mA in Fig. 5d) and to show that the ‘buried’ population is not the same population, since the surface stimulus does not make the buried population refractory (Fig. 5d, e, f ) . This step proves that the ‘buried’ population is not merely the nearby axons of the ‘surface’ population. The final step is to increase the strength of the surface shock until the ‘buried’ population is stimulated by it, as shown by refractoriness. Fig. 5 (right) shows that the critical surface-anodal current in this experiment was 3.2 mA. The rate of conduction of the cortico-spinal volley through the cervical enlargement is measured in Fig. 6. In five baboons the slopes were of the order of 60 m/sec (legend, Fig. 6). The latency of EPSPs of cervical motoneurones will show that these are obviously related to this sharp, early component of the cortico-spinal volley. References p . 242

228

C. G . PHILLIPS AND R. PORTER

'9"" t

P

'V

B ?

/

Y P

40

Fig. 6 . Conduction of cortico-spinal volley through cervical enlargement. S t stimuli, 0.2 msec, 4.5 mA, to right precentral gyrus midway between superior and inferior precentral fissures. Right: superimposed records taken from contralateral dorsolateral surface of cord, at headward end of laminectomy and at points 4.75, 1 I .8, 17.1, 22.8, 27.4 and 32.5 mm caudal. Positivity upwards. Time: msec. Left: plot of conduction velocity. Time of arrival at each point was rr.easured from start of shock to start of negativity. Slope 59 m/sec. Slopes in other experiments were 63.4, 62.5, 56.5 and 58.7 m/sec. Dotted line originates 75 mm from rostra1 recording point. Shortest distance between this point and the stimulated point in these specimens was 77-84 mm.

0.8

-. -

w - - - - - - -

/

\

1

1 mV

1.7

2.2 ----

---------.

Fig. 7. Superimposed intracellular records of monosynaptic excitatory action of cortico-spinal volleys on a median nerve motoneurone at C7-8 level. Depolarization signalled by upward deflexion. Membrane potential - 74 mV. Strengths of 0.2 msec S + pulses noted in margin (in mA). Final record : extracellular control taken after microelectrode withdrawn from cell. Calibration of recording system: 1 mV for responses to 0.8 and 1.7 mA; 5 mV for remaining records. Time: msec.

229

THE P Y R A M I D A L PROJECTION TO M O T O N E U R O N E S

INVESTIGATION OF CORTICO-MOTONEURONAL COLONIES

Median nerve (radial flexors of wrist, pronators, flexors of proximal phalanges of all digits, flexors of distal phalanges of three radial digits, radial lumbricals, some intrinsic muscles of thumb). Records from the median motoneurone which received the largest quantity of monosynaptic cortico-spinal excitatory action are reproduced as Fig. 7. It was situated at C7-8 level. The earliest pyramidal impulses reached t h s level 1.74 msec after the start of the cortical shock (time of positive peak of tract wave), and the majority of impulses at 2.08 msec (time of negative peak of tract wave). The synaptic potential rose abruptly at 2.4 msec after the start of the shock. The delay 0

II 3.

20;o 0

2

2.

0.

I.

0 c

0

0

0

c 0

1.1

1//-Y-

0

3 O 0

I

2

3

4

5mA

o

- d Id - W2 O -

r

3

4

5mA

Fig. 9. Fig. 8. Fig. 8. Quantities of monosynaptic excitatory action (ordinates) at different strengths of cortical stimulation with single S + , 0.2 msec pulses (abscissae) in 94 motoneurones of the median nerve. Eleven other motoneurones gave no monosynaptic excitatory response a t stimulus strengths of 4.5 mA. For three motoneurones, typical curves show growth of monosynaptic action with increasing volleys (filled circles). For remaining motoneurones, open circles show measurements a t a single strength (the largest strength, or the only strength, used). Fig. 9. Results, plotted as in Fig. 8, from 45 motoneurones of the ulnar nerve. References p . 242

230

C. G . P H I L L I P S A N D R . P O R T E R

is thus 0.66 msec if measured from the arrival of the earliest impulses, or 0.32 msec if measured from the arrival of the majority of impulses at this segment. Since some of this time would be consumed in slowed conduction in tapering presynaptic arborisations, these measurements prove that the EPSP is monosynaptic. The second synaptic impact at strengths greater than 2.2 mA may be due to repetitive firing of pyramidal cells, or to the discharge of interneurones. It comes late enough not to interfere with the measurement of maximal monosynaptic amplitude. The absence of evidence of mV

0

1

Z

3

4

5

6

7

8

9

1

0

m

A

Fig. 10. Results, plotted as in Fig. 8, from 20 rnotoneuronesof the posterior interosseous nerve.

inhibitory synaptic action is not significant, since the membrane potential was -74 mV, and KC1-filled microelectrodes were used in these experiments. Growth of monosynaptic action as a function of cortical stimulus strength is plotted in Fig. 8 (tallest curve). The maximum quantity was not reached until 3.0 mA shocks were given. At this strength the pyramidal cell population would extend nearly to the depth of the Rolandic fissure (Figs. 4, 5) and up to 10 mm in the medial and lateral directions from the stimulated point (Fig. 3). We cannot escape the conclusion that the cells of this median motoneurone’s colony occupy a cortical territory which has about this extent (cJ Landgren et al., 1962b). That they d o not occupy a wider territory is shown by the flattening of the curve at a strength at which the population would still not have reached its maximum extent (Fig. 2).

THE P Y R A M I D A L PROJECTION T O MOTONEURONES

23 1

For simplicity three curves only are plotted in Fig. 8 (filled circles). These are tending to different maxima. The open circles show, for the remaining median motoneurones, either the largest monosynaptic action obtained, or the action obtained at a single stimulus strength in cases in which one strength only was used. Ulnar nerve (ulnar flexors of wrist, flexors of distal phalanges of ulnar fingers, ulnar lumbricals, interossei, hypothenar muscles, some thenar muscles). In Fig. 9 the curves for four ulnar motoneurones are plotted in full. These show different maxima, but their plateaux are reached at strengths of stimulation which indicate that the extent of the cortical territory occupied by their colonies is similar to that of the median motoneurone of Figs. 7 and 8. Posterior interosseous nerve (elbow flexors, supinators, dorsiflexors of wrist and digits, abductors of thumb). The results are plotted in Fig. 10. Musculocutaneous nerve (elbow flexion, supination). Fig. 11 shows the results. The curves indicate more extensive colonies : maximum monosynaptic action needs stronger stimuli than in Figs. 8-10. mV

2.01

"/

0 0

Fig. 1 1 . Results plotted as in Fig. 8, from 47 motoneurones of the musculocutaneous nerve.

Triceps nerves (adduction at shoulder, elbow extension). In preliminary experiments made with the whole musculospiral nerve trunk, (and not included in the present total of experiments), we found that 7/23 motoneurones showed no detectable monosynaptic action with cortical shocks of 5 mA. This was then the maximum output of our stimulator. We therefore increased its output before embarking on experiments on nerve branches to the triceps, which had to be laboriously dissected off the main musculospiral trunk in the musculospiral groove. About half of these motoneurones (28/49) showed no monosynaptic action, and therefore lacked cortical pyramidal colonies as defined. There is no doubt that the strongest shocks reached the depth of the fissure, so that no colony would have been missed. The results for the remaining 21 triceps motoneurones are plotted in Fig. 12. References p . 242

232

C . G. P H I L L I P S A N D R. P O R T E R

Inspection of Figs. 8-12 makes it obvious that motoneurones of the distal groups receive, in general, greater quantities of monosynaptic action than those of proximal groups. It is legitimate to compare these quantities, since few membrane potentials were below 60 mV and none was below 55 mV (cf. Coombs et al., 1955, Fig. 2B). Detailed comparison is facilitated by an arbitrary division of the graphs into regions and by counting the numbers of curves and isolated responses falling within each region. Take first the ragion lying to the left of the tallest curve in Fig. 8. This contains about 30% of the median motoneurones (31/105), about 40% of the ulnars (18/45) and 60 % of the posterior interosseous motoneurones (12/20), but it contains no musculocutaneous or triceps motoneurones. Take then the region lying below the lowest curve of Fig. 11. This contains 73% of the triceps motoneurones (36/49); 57% (28/49) showed no monosynaptic response. This region contains 32% of the musculocutaneous motoneurones (1 5/47), but only 18 % of the median (19/105), 11 % of the ulnar (5/45), and about 10 % of the posterior interosseous (2/20). We now publish illustrative intracellular records of synaptic potentials from mV

LO

;o8

a5 -

00 00

I

I

0

1

2

I

3

I

4

S

I

b

I

l

I

8

I

9

l

1

O

0 0

1

m

A

Fig. 12. Results, plotted as in Fig. 8, from 21 motoneurones of the triceps nerves. No monosynaptic excitatory responses were obtained from 28 further triceps motoneurones. Of these, 15 gave early inhibitory responses.

musculocutaneous and triceps motoneurones of baboons, to supplement our previously published records from ulnar, median and posterior interosseous motoneurones (Hern et al., 1962; Landgren et al., 1962a). Fig. 13 (A) is from a musculocutaneous motoneurone with a membrane potential of -75 mV. It may be compared with Fig. 7, in which, also, the monosynaptic EPSP showed no evidence of inhibitory action, but was followed by a second excitatory action. Fig. 13 (B) illustrates the findings in another musculocutaneous motoneurone whose membrane potential was -61 to -63 mV. Such a value may signify some injury to the membrane, but it may also be due, in part, to depolarization by background synaptic activity associated with light anaesthesia (nitrous oxide, oxygen and 0.5 % chloroform with minimal barbiturate). This sequence of monosynaptic excitation followed by early inhibition was seen in several musculocutaneous moto-

T H E P Y R A M I D A L P R O J E C T I O N TO MOTONEURONES

233

neurones. The time between start of monosynaptic excitation and start of inhibition was 1.5 msec. Similar values were found by Preston and Whitlock (1960, 1961) in the lumbar enlargement of monkeys and by Landgren et al. (1 962a) in the cervical enlargement of baboons. Larger values would be expected in the lumbar enlargement A

2.5

3.6

B

-

4.5

2.4

Fig. 13. Superimposed intracellular records from two motoneurones of the musculocutaneous nerve. (A) Rostra1 C5 level. Membrane potential -75 mV. Strengths of 0.2 msec stimulating pulses noted in margin in mA. Bottom record is extracellular control. (B) Caudal C4 level. Membrane potential -61 to -63 mV. Bottom record is extracellular control. Time in msec. Calibration: 1 mV for all records.

if different components of the cortico-spinal volley were responsible for the excitatory and inhibitory effects, since the temporal dispersion of such components would be greater at the longer conduction distance. This consideration supports Eccles’ suggestion (1957, p. 179) that the inhibition is mediated by local interneurones. Fig. 14 shows a musculocutaneous motoneurone with minimal EPSP and more conspicuous TPSP, which becomes complicated in response to the stronger shocks. Fig. 15 shows that an apparent absence of inhibitory synaptic action need not mean an absence of inhibitory synapses. The membrane potential of this musculocutaneous motoneurone was -72 mV. After the recording of EPSPs had been completed, it was deliberately damaged by repeated stabs with the microelectrode. This reduced the membrane potential to -41 mV, and unmasked a biphasic inhibitory action. It is obvious that any systematic study of cortico-spinal inhibition will require the References p . 242

234

C.G. PHILLIPS AND R. PORTER

5.6

0.6

0.8 7.2

1 .o

1.8

8.8 [A

3.9

mV

8.8

Fig. 14. Superimposed intracellular records from another motoneurone of the musculocutaneous nerve, to show minimal monosynaptic excitatory action followed by early inhibitory synaptic action, and complex responses to stronger 0.2 msec S stimuli (strengths in margin). Last record is extracellular control. Calibration: 1 mV. Time in msec.

+

0-

mV

5.3

-72-

[.

7.2 1 .Q

mV

9.6 1mv]

3.2 1.9

4.2

9.4

9.6

Fig. 15. Intracellular recording from another motoneurone of the musculocutaneous nerve, membrane potential -72 mV. First record is of surperimposed antidromic spikes. Series of excitatory synaptic potentials set up by S - t cortical shocks ( I .O-9.6 mA) show inhibitory notchings. Membrane potential was then reduced to -41 mV by several deliberate stabs. Note inhibitory synaptic potentials in response to 3.2 and 9.4 mA stimuli. Last record is extracellular control. Calibrations: 0 and -72 mV for antidromic spike; 1 niV for responses to 1.0 and 1.9 mA; 5 mV for remaining responses. Time in msec.

THE PYRAMIDAL PROJECTION TO MOTONEURONES A

235

B

3.2 2.9

4 .O

5.2

8.8

&id+-.o

1

7

3.1

rnv

9.5

8.8

9.5

Fig. 16. Intracellular records from two triceps motoneurones. (A) Membrane potential -68 mV. After responses to cortical stimuli of 0.7-8.8 mA had been recorded, cell was deliberately injured by stabbing; membrane potential -43 mV. Early inhibitory response unmasked. Bottom record is extracellular control. (B) Another motoneurone, membrane potential -62 mV. Threshold for inhibitory synaptic action (1.0 mA) is lower than that for monosynaptic excitatory action (>1.0, < 1.8 mA). Bottom record is extracellular control. Gain : 1 mV for (A), 0.7,0.9, 2.9 and 3.1, and (B), 1.0, 1.8, 3.2, 5.2; 5 mV for remaining records and extracellular controls. Time: msec.

use of electrodes and of depolarizing currents to reveal inhibition by shifting the membrane potential away from its equilibrium level. Fig. 16 (A) shows recordings from the triceps motoneurone from whlch the uppermost curve of Fig. 12 was plotted. Its membrane potential remained at -68 mV, until it was deliberately injured to unmask an IPSP. Fig. 16 (B) shows records from a second triceps motoneurone, in which the cortical threshold for IPSP (1.0 mA) was lower than that for the monosynaptic EPSP. Fig. 17 shows records from another triceps motoneurone. The presence of minimal monosynaptic action in the first record at stimulus strength 1.7 mA is made obvious in the second record by repetitive stimulation at 200/sec, in which condition the transmitting potency of pyramidal synapses is enhanced (Landgren et al., 1962a). As single-shock stimulation is strengthened, the initial monosynaptic response merges into a large, steadily-rising wave of depolarization, on which, exceptionally, one can no longer discern the early peak which permitted the measurement of References p . 242

236

C . G. PHILLIPS A N D

1.7

1.7

K. P O R T E R

A 8 A [. 8.0

m~

10.0

Fig. 17. Intracellular records from a motoneurone of the nerve to long head of triceps. Membrane potential -77 mV. A small monosynaptic excitatory component at 1.7 mA is made more obvious when the shocks are repeated at 200/sec. As stimulation is strengthened, depolarisation continues in a slowly rising curve. Calibrations: 1 mV for records on left; 5 mV for synaptic potentials on right and extracellular control; 0 and -77mV for antidromic spike. Time: msec.

maximum monosynaptic action in most of our records ( e . g . Figs. 13-1 6). The measurement of quantity in this cell was possible only at strength 1.7 mA, and it is so plotted on Fig. 12. Fig. I8 illustrates pure inhibitory action on another triceps motoneurone. The latency of the first IPSP is 3.7 msec, and of the second, 8.0 msec. N o monosynaptic excitatory action could be detected in the triceps motoneurone whose records are shown in Fig. 19, even by repetitive stimulation at 200/sec (cf.

1.o

6.1

1.8

[I mv

7.1

5.1

w l

9.8

Fig. 18. Intracellular records from a motoneurone of one of the nerve branches to distal part of triceps. Membrane potential -63 mV. N o monosynaptic excitatory action, but early inhibitory action. Final record is extracellular control. Calibrations: 1 mV. Time: msec.

THE PYRAMIDAL PROJECTION TO MOTONEURONES

237

Fig. 17). This cell is interesting because stimulation at one cortical point (B) gave pure inhibitory effects whose latencies, at strength 6.8 mA, were 3.0 and 8.0 msec, whereas stimulation at point A, 9 mm medial to B, gave excitatory depolarization only. Membrane potential was the same (-64 mV) throughout both sets of recordings. A

3.1

B 4.8

5.1

6.0

7.0

7.7

[I m"

9.8

10.0

9.6

9.8

Fig. 19. Intracellular records from a motoneurone of a nerve branch to distal part of triceps, to show different actions of stimulation at point A (excitatory, with no monosynaptic component; this was also checked by repetitive stimulation, cJ Fig. 16), and point B, (inhibitory), 9 mm lateral t o point A. Membrane potential-64 mV. Separate extracellular control records for points A and B. Calibrations: 1 mV. Time: msec.

Since the muscles supplied by the inusculocutaneous nerve are antagonists of the triceps in the movement of elbow extension, and since Bernhard and Rohm (1954, Fig. 11) have published a map showing reciprocal localization of areas for monosynaptic excitation of biceps and triceps in the precentral cortex of the macaque, we have carefully reviewed our own experiments from this point of view. We felt obliged to reject all those motoneurones for whch no 'best point' (Landgren et al., 1962b) could be mapped, the synaptic actions being evoked only with stronger cortical stimuli, and then with equal ease from all parts of a wide area; we were then left with 31 musculocutaneous (Fig. 20, bi) and 31 triceps motoneurones (Fig. 20, tr) in eight preparations, in only one of which, unfortunately, did motoneurones of both types satisfy our criteria. Filled circles in Fig. 20 denote monosynaptic excitation; half-filled circles, monosynaptic excitation followed by inhibition (e.g. Fig. 13, B ) ; and open circles, pure inhibition (e.g. Fig. 18). Scrutiny of these maps has failed to convince us that any obvious reciprocal monosynaptic localization exists in the baboon. The suggestion contained in the bottom right-hand map is contradicted by the lower three maps on the left, which show bi points in positions lateral to the tr Refel emes p. 242

238

C. G. P H I L L I P S A N D R . P O R T E R

bi

f:I:-

tr

Fig. 20. Localization of ‘best points’ for 31 motoneurones of the musculocutaneous nerve (bi) and 31 motoneurones of the triceps nerves (tr) in eight brains. Each map shows the brachio-facial genu of the right Rolandic fissure (above) and the superior and inferior precentral fissures (below) ; the occipital pole lies above and the frontal pole below. These motoneurones were the only examples from these muscle groups for which ‘best points’ could be convincingly localized with 0.2 msec S stimuli of strength 2.0 mA or less. Unfortunately, ‘antagonistic’ motoneurones satisfying this criterion were collected from the same preparation in one experiment only (bottom right). Filled circles : monosynaptic excitatory action. Open circles: early inhibition only; no monosynaptic excitatory action. Half-filled circles : monosynaptic excitatory action followed by inhibitory action. There is no obvious grouping to suggest reciprocal localization. This part of the precentral gyrus also contained ‘best points’ for distal muscles in these and in other baboons’ brains.

+

point of the first-mentioned map. It is to be remembered that this region of the precentral gyrus also contained the colonies for motoneurones of the distal muscle groups in these and in our other brains. It is possible that the discrepancy is to be explained by the different experimental

T H E P Y R A M I D A L P R O J E C T I O N TO M O T O N E U R O N E S

239

methods used by ourselves and by Bernhard and Bohm (1954). Their methods were the best available at that time. They stimulated the cortex at 25/sec. After 1 sec of stimulation, discharges appeared in the triceps and biceps nerves at monosynaptic latency. A background of facilitation was thus necessary to reveal these monosynaptic actions, and the observed reciprocity may have been due to localization in the ‘facilitatory’ system and not to localization in the monosynaptic pathway. It is interesting that Bernhard and Bohm’s biceps field lies to the lateral side of their triceps field. In our Fig. 19 the inhibitory region for a triceps motoneurone does lie lateral to the excitatory field, but this was a motoneurone on which no monosynaptic excitatory action could be detected. The lack of obvious reciprocal localization at the headward end of the monosynaptic cortico-motoneuronal pathway in the baboon has prompted us to wonder whether reciprocal control need necessarily imply reciprocal localization in the cortex. Such doubts gather force when one reflects upon the extraordinary mobility of the forelimb. The muscles acting at the elbow joint have complex functional relationships, not a simple antagonistic relationshp. Thus in man, whose forelimb resembles in general structure that of the baboon, biceps is a powerful supinator of the forearm. In the movement of supination, unwanted elbow flexion is prevented by co-contraction of triceps (Beevor, 1904, pp. 23-24). Again, triceps (long head) is an adductor of the arm. In the movement of adduction, unwanted extension of the elbow is prevented by co-contraction of biceps (Beevor, 1904, p. 34). Only in the movements of elbow flexion and extension are biceps and triceps related reciprocally. In the shoulder and in the hand, where joint mobility is so much greater, the reciprocal relationships of the muscles may be almost infinitely labile. Cortical control of the pyramidal colonies which innervate their motor units should be correspondingly subtle. CONCLUSION

Monosynaptic cortico-spinal excitation of motoneurones of the baboon’s forelimb is brought about by intermingled colonies of corticofugal pyramidal cells. These colonies thus provide for direct cortical control of the motor units of the final common path, which are the ultimate units of motor function. We have neglected the many important ‘indirect’ pathways which control the motor units through fusimotor neurones and through interneurones of the brain stem and cord. Nor have we here concerned ourselves with the intracortical synaptic mechanisms which play selectively on the corticofugal pyramidal cells in natural motor activity. To isolate the cortico-motoneuronal colonies for experimental study we have used the simplest type of electrical stimulation. We have deliberately avoided the complexities of classical repetitive stimulation, which may, and especially when focal cathodal or bifocal electrodes are applied to the cortex, engage synaptic systems upstream of the corticofugal cells (cf. Hern et al., 1962; Phillips and Porter, 1962), and may conceivably excite or inhibit the corticofugal cells in varying patterns of localization. Our results have shown that many of the colonies connected to distal motoneurones command larger quantities of monosynaptic excitatory action than are commanded References p . 242

240

C. G. P H I L L I P S A N D R. P O R T E R

by any of the colonies connected to proximal motoneurones, and the significance of this finding must now be considered. The rate of rise of the cortically-evoked EPSP is comparable to the rate of rise of the EPSP evoked by stimulation of the Group Ia afferent fibres (Landgren et a/., 1962a, Fig. 2). This shows that these cortico-spinal synapses are not far from the soma of the motoneurone. But it will be noticed that even the largest quantity ofmonosynaptic action evoked by a single cortico-spinal volley in an ulnar motoneurone (Fig. 9, 3.4 mV) is too small to reach the firing level, and it might be supposed that the cortico-motoneuronal connexion would therefore be insignificant in normal function, at least without a background depolarization of the motoneurone sufficient to bring its membrane potential within 1 or 2 mV of the firing level. This difficulty is to some extent resolved by a special property of pyramidal synapses, illustrated in Fig. 21. The experiment begins by choosing strengths of stimulation of

1

200

I 1 1 1 1 1

P" 5

mV

Fig. 21. Superimposed intracellular recordings from a motoneurone of the median nerve, comparing its responses to repetitive Group la afferent volleys and to repetitive cortico-spinal volleys, both at 200/sec. Upper pair: simultaneous recording of submaximal Group Ia volleys from dorsal root entry zone (upper trace) and of the monosynaptic EPSP's evoked by these (lower trace). Lower pair: response to repeated S pulses (0.2 msec, 2.4 mA), the first of which evokes a quantity of monosynaptic action comparable to that evoked by the first Group Ia volley. Upper frace: cortico-spinal volleys, recorded from pointed silver wire projecting 80 ,u from 15 p Pyrex capillary, inserted into dorsolateral white matter 22 mm rostra1 to the motoneurone. Lower trace: monosynaptic EPSPs. Time scale given by shock artefacts at 5 msec intervals.

+

Group la axons (upper records) and of corticofugal axons (lower records) which evoke monosynaptic EPSPs of approximately equal size on the same test motoneurone. Six of these stimuli are then delivercd at 200/sec. The upper records show that the size of the six incoming Group l a volleys does not change, and the resulting monosynaptic potentials do not increase in amplitude (Curtis and Eccles, 1960). It is otherwise when the cortico-spinal volleys are repeatcd. There is a progressive growth of monosynaptic transmitting potency, without obvious change in the size of the six cortico-spinal volleys (Landgren et al., 1962a). Thus the high-frequency discharges

THE P Y R A M I D A L PROJECTION TO M O T O N E U R O N E S

24 1

of which pyramidal neurones are known to be capable would be specially effective in depolarjzing the motoneurones over the monosynaptic pathway. Since monosynaptic connexion is most strongly developed in relation to distal motoneurones, attention turns naturally to its probable significance in relation to movements of the hand. Wood Jones (1920, p. 235) discussed the relative lack of morphological specialization in the hands of primates, and concluded that the essential differences lay in the central nervous organization. ‘The difference between the hand of a man and the hand of a monkey lies not so much in the movements which the arrangement of muscles, bones and joints makes it possible for either animal to perform, but in the purposive volitional movements which under ordinary circumstances the animal habitually exercises.’ Kuypers’ contribution (Vol. 1 1 of the Progress in Brain Research series) shows that the monosynaptic pathway is denser in chimpanzee than in macaque. It is probable that the baboon, whose brain is more convoluted than the macaque’s, occupies an intermediate position. Part at least of the increased nervous control postulated by Wood Jones may depend on the increase in monosynaptic control. Such an increase would confer greater possibilities of fractionating items of movement (Leyton and Sherrington, 1917). The powerful effect of iterative corticospinal activity should provide for the rapid initiation of movement. ACKNOWLEDGEMENT

We thank Mr. C. H. Carr for technical assistance. SUMMARY

I n trying to understand the working of the corticospinal motor systems it is rational to begin at the level of the corticofugal pyramidal neurones. One can begin by asking a fundamental question about the nature of motor localization: how large are the cortical areas contributing axons to different spinal destinations, and are these areas separate or do they overlap? We have confined ourselves to the simplest of several possible corticospinal control systems - the localization in the cortical arm area of the baboon of colonies of pyramidal neurones which project monosynaptically to sample motoneurones belonging to different muscle groups of the upper limb. Such colonies form significant functional groupings since the motor units to which they project are the ultimate elements in the organization of muscular contraction. The monosynaptic actions of the corticospinal discharges have been recorded from the sampled motoneurones with intracellular microelectrodes. We have avoided ‘classical’ repetitive stimulation, which could conceivably select pyramidal neurones in different combinations by activating synaptic mechanisms lying upstream of them in the cortex, and thereby obscure the fundamental pattern of corticofugal localization. We have used single, brief surface-anodal shocks which directly excite populations of corticospinal cells whose axons conduct impulses at about 60 mlsec. We have found that the cortico-motoneuronal colonies for motor units of different muscles overlap, and that they differ in their spatial extent and in the quantity of References p. 242

242

C. G. P H I L L I P S A N D R. P O R T E R

monosynaptic excitatory action they command on the test motoneurones. The colonies projecting to motoneurones controlling distal muscles tend to occupy smaller cortical areas, and to command larger monosynaptic actions; the colonies projecting to the motoneurones of proximal muscles tend to occupy larger areas, but to command smaller monosynaptic actions. About half of the triceps motoneurones received no monosynaptic projection, i.e. possessed no corticofugal colonies as defined. They received polysynaptic excitatory control and also inhibitory synaptic control from the cortex. The hands of primates are morphologically primitive, and depend on a refined neural control for their wide range of usefulness. It is probable that such versatile but precise control depends in part on a special development of the monosynaptic corticospinal pathway to the motoneurones of the distal muscles of the upper limb. The directness of this pathway should increase the accessibility of hand motor units to the complex intracortical neuronal systems lying upstream of the corticofugal pyramidal neurones. REFERENCES BEEVOR, C. E., (1901); The Croonian Lectures on Muscular Movements and their Representation in the Central Nervous System. London, Adlard. BERNHARD, C. G., A N D BOHM,E., (1954); Cortical representation and functional significance of the corticomotoneuronal system. Arch. Neural. Psychiat. (Chic.), 72, 473-502. BERNHARD, C. G., BOHM,E., AND PETERS~N, I., (1953); Investigations on the organization of the cortico-spinal system in monkeys. Acta physiol. scand., 29, Suppl. 106, 79-105. COOMBS, J. S., ECCLES,J. C., AND FATT,P., (1955); Excitatory synaptic action in motoneurones. J. Physiol. (Land.), 130, 374-395. COOPER,S., AND DENNY-BROWN, D., (1927); Responses to stimulation of the motor area of the cerebral cortex. Proc. roy. Sac. B, 102, 222-236. CURTIS,D. R., AND ECCLES, J. C., (1960); Synaptic action during and after repetitive stimulation. J. Physiol. (Land.), 150, 374-398. ECCLES, J. C., (1957); The PhysioZogy of Nerve Cells. Baltimore, The Johns Hopkins Press. HERN,J. E. C., LANDGREN, S., PHILLIPS, C. G., AND PORTER, R., (1962); Selective excitation of corticofugal neurones by surface-anodal stimulation of the baboon’s motor cortex. J. Physiol. (Lond.), 161, 73-90. LANDGREN, S., PHILLIPS,C. G., AND PORTER, R., (1962a); Minimal synaptic actions of pyramidal impulses on some a-motoneurones of the baboon’s hand and forearm. J . Physiol. (Land.), 161, 91-111. LANDGREN, S., PHILLIPS, C. G., AND PORTER, R., (1962b); Cortical fields of origin of the monosynaptic pyramidal pathways to some a-motoneurones of the baboon’s hand and forearm. J. Physiol. (Land.), 161, 112-125. LEYTON, A. S. F., AND SHERRINGTON, C. S., (1917); Observations on the excitable cortex of the chimpanzee, orang-utan and gorilla. Quart. J . exp. Physiol., 11, 135-222. PATTON, H. D., AND AMASSIAN, V. E., (1954); Single- and multiple-unit analysis of cortical stage of pyramidal tract activation. J . Neurophysiol., 17, 345-363. PHILLIPS,C. G., AND PORTER,R., (1962); Unifocal and bifocal stimulation of the motor cortex. J. Physiol, (Land.), 162, 532-538. PRESTON, J. B., AND WHITLOCK, D. G., (1960); Precentral facilitation and inhibition of spinal motoneurons. J. Neurophysiol., 23, 154-170. PRESTON, J. B., AND WHITLOCK, D. G., (1961); Intracellular potentials recorded from motoneurons following precentral gyms stimulation in primate. J. Neurophysiol., 24, 91-100. SHERRINGTON, C. S., (1931); Quantitative management of contraction in lowest level co-ordination. Brain, 54, 1-28. WOODJONES,F., (1920); The Principles of Anatomy as seen in the Hand. London, Churchill.

T H E P Y R A M I D A L P R O J E C T I O N TO M O T O N E U R O N E S

243

DISCUSSION

ECCLES:I am greatly interested by the remarkable finding with repetitive cortical stimulation. The potentiation of the successive EPSPs undoubtedly gives great power to the corticospinal action. This potentiation is so large that I would expect the synapses on motoneurons to look different from the group Ia synapses. With electron microscopy I would expect different arrangements of synaptic vesicles in synapses that exhibit little potentiation with repetitive stimulation and in synapses that exhibit such great potentiation. Possible selective degeneration would enable these two types of synapses to be identified. I would like to ask whether there has been any study of brief and delayed post-tetanic potentiation with these pyramidal tract synapses on motoneurons. 1 would certainly expect the brief post-tetanic potentiation to be very large. PHILLIPS: We have not investigated delayed post-tetanic potentiation in this system. With regard to the early effect: the superimposed sweeps of Fig. 21 were recurring at I-sec intervals. It is obvious that the potentiating effect of each train of 6 pyramidal volleys at 200/sec had subsided completely during this interval of time. Further work is needed on the effects of larger numbers of presynaptic volleys. LUNDBERG: You did mention that a number of triceps brachii motoneurons do not receive EPSP’s from the motor cortex. How do motoneurons of fast and slow muscles compare in this respect? PHILLIPS:We have not measured the contraction times of the different muscle groups. Something might be learned from the conduction velocities of the motor axons, but we recorded the antidromic impulses on slow sweeps, and the stimulus artefacts were so small that I am afraid we may not be able to compare the conduction velocities of the axons of those motoneurons which did, or did not, exhibit monosynaptic pyramidal actions. WILLIS:I would like to make two points. With regard to the question of reciprocal innervation in the cervical cord, Dr. Schmidt and I have evidence that the pattern of monosynaptic connections among cervical motoneurons does not always fit a simple myotatic unit relationship. The other point is that the monosynaptic corticofugal connections to reticulospinal neurons of the cat also show a striking potentiation with repetitive stimulation, as will be shown tomorrow. PHILLIPS:This is most interesting. It appears then that at the level of reflex control also, reciprocal innervation of the forelimb is a bonajde physiological affair, and not the mere consequence of rigidly stereotyped anatomical arrangements. It is very interesting that yet another corticofugal system shows an increasing monosynaptic action of repeated impulses. Dr. Lundberg has also seen such an increasing action of pyramidal synapses on lumbar interneurons in the cat. It may be a general property of the synapses formed by corticofugal pyramidal neurons.

244

DISCUSSION

SZENTAGOTHAI : The non-existence of any specific topographic relation between pyramidal cells connected with the motoneurons of several partly antagonistic muscles, so beautifully shown in your experiments, might give rise to speculations concerning the developmental aspects. It would show that also if there apparently is a general tendency of corticospinal axons, arising from a certain region, to contact with certain segments of the spinal cord, there is no predetermined specific attraction between certain pyramidal axons and a fixed set of motoneurons. This would again mean that inside of a large group of axons entering a spinal segment the connections finally stabilized must be governed to considerable extent by chance and the pyramidal neurons that have gained connection with a given motoneuron are identified later on by some ‘learning process’ which builds them in into the functional pattern or connection system of the cortex. PHILLIPS : There are fascinating possibilities here. GELFAN: What was the particular reason to investigate the motoneurons to the forelimbs and not the cells in the lumbosacral cord? PHILLIPS: It was part of a general desire to get away from the much-studied hindlimb, which is largely a postural and locomotor organ; by contrast, the forelimb of primates, especially in its distal segments, is primarily an explorer and manipulator of the environment, and is therefore especially challenging to the student of cerebral motor control. GRANIT: Seeing this build-up in an animal which is so lightly anaesthetized, raises the question whether there might not be the same thing also from the peripheral muscles and in the monosynaptic paths.

PHILLIPS:In describing Fig. 21 I perhaps did not sufficiently emphasize that the records were from one and the same motoneuron, showing, under identical conditions of light anaesthesia, that these different synapses have these different effects. ECCLES:We have published records with the DSCT cells, where one and the same DSCT cell showed quite different potentiations to repetition to Ia fibers, which were negligible, and to the Ibfibers on the samecell which more than doubled with repetition. First we thought that it was perhaps the Ib synapse being different from la, perhaps with a different chemical transmitter, and, therefore, different relationships to frequencies of stimulation. But it turns out that it was within the la group. There are rules there, that we are gradually fumbling towards, about frequency but certainly not at all understood yet, but they seem of the greatest physiological importance. SPRAGUE:You have made a few remarks that possibly implicated the dendrites in this mechanism. Would you be willing to speculate further on possible anatomical connections that the pathway might have on the motor neurons in achieving these effects?

THE PYRAMIDAL PROJECTION TO MOTONEURONES

245

PHILLIPS: We know from Dr. Szentigothai’s work that the synapses of group Ia afferents are applied to the basal dendrites and soma of the motoneuron. It may be inferred that those pyramidal synapses which evoke monosynaptic EPSPs whose time course is comparable to that of the group l a E P S P , are also applied so close to the soma t!iat their E P S P ’ s do not suffer electrotonic distortion when recorded from the soma. What I was speculating about was whether any part of the slowly-rising synaptic depolarization (e.g. Fig. 17) could possibly be due to large monosynaptic pyramidal actions up011 outlying dendrites. Such depolarization could of course be due to repetitive asynchronous bombardment of the motoneuron by pyramidallyactivated interneurons, and this would be the orthodox, and perhaps the more likely, explanation. But one can play with the idea that it may be the function of some synapses (those nearest the soma) to produce steeply-rising actions which can quickly interfere with the prevailing level of activity; and that other synapses (more remote from the soma) may bring about more smoothly-graded depolarizing pressures which affect the threshold and firing of the motoneuron. It seems improbable that the synapses applied to outlying dendrites can exert no action on the impulse-generating mechanism of the initial segment of the axon. But evidence will be hard to get. G R A N I T : Did you also see a delayed depolarization of the kind that we have been discussing?

PHILLIPS:We have not looked at the records of the antidromic impulses of these cervical motoneurons systematically to see if they ever show your delayed depolarization.

246

Afferent Connections to Reticulo-Spinal Neurons F. M A G N I

AND

W. D. W I L L I S *

Istituto di Fisiologia dell’ Universitd di Pisa e Centro rli Neurojisiologia del C.N.R., Sezione di Pisa, Pisa (Italy)

In our previous paper (this Volume, p. 56), a method has been described for identifying reticulo-spinal neurons by intracellular recording of antidromic action potentials resulting from stimulation of their axons within the spinal cord. Neurons thus identified could then be investigated with respect to their afferent connections. Attention has so far been restricted to a study of the action of volleys from the cerebral cortex, from a pathway passing through the region of the central tegmental tract in the mesencephalon, and from peripheral nerves of the forelimb. METHODS

A description of most of the techniques employed in these experiments has already been given (Willis and Magni, this Volume, p. 56). Further details may be found in the text and in the full reports of the work (Magni and Willis, 1963, 1964a, b). RESULTS

Effects of corticofugal volleys upon reticulo-spinal neurons Since it would have been difficult to explore point by point the cerebral cortex with stimulating electrodes while recording intracellularly from a single neuron, no attempt was made to map the active and inactive cortical areas. Instead, a number of fixed bipolar electrodes were applied to various cortical regions, and stimuli were applied through these in succession during the time of impalement of each neuron. The strength of stimulation was kept subthreshold for movements. The areas stimulated are indicated by the rectangular shaded areas in the diagram of Fig. 1 (I). The full complement of electrodes was used only on one hemisphere, but additional electrodes were placed on the pre- and postcruciate regions of the opposite side to allow comparison of the effectiveness of the two sides. The neuron illustrated in Fig. 1 received excitatory postsynaptic potentials (EPSP’s) from all the cortical regions examined (Fig. I , A-G). The strongest action was from

* Postdoctoral Research Fellow, National Institute of Neurological Diseases and Blindness, U.S. Public Health Service.

AFFERENT CONNECTIONS TO RETICULO-SPINAL N E U R O N S

247

Fig. 1. EPSP’s from various cerebral cortical regions. In this and succeeding figures, sample records illustrating activity in reticulo-spinal neurons are arranged as follows. The upper traces are the intracellularly recorded potentials (depolarization upwards) and the lower traces the cord volley (usually recorded at T12). When three traces are shown, the lowermost one is the field potential recorded after the microelectrode was withdrawn to a just extracellular position. (A-G). EPSP’s recorded in a reticulo-spinal neuron after stimulation of the indicated cortical areas. (H). Antidromic action potential from cord stimulation at L1. The 10 msec timer and the 2 mV potential scale apply to A-G, while the msec and 50 mV scales are for H. The diagram in I shows the cortical areas stimulated. Abbreviations: pr., precruciate; PO., postcruciate; s.a., anterior suprasylvian gyrus; t., temporal; o., occipital. The letters I. or r. refer to the side of the body, left or right. (From Magni and Willis, 1964a.)

a temporal region which is considered to be part of the primary auditory cortex (Fig. IF; CJ Ades, 1959). A considerable effect was also obtained from an occipital region known to be part of the primary visual cortex (Fig. 1G; cJ Doty, 1958). The neuron received EPSP’s of about equal size from the pre- and postcruciate regions of both sides (Fig. 1, A-D). That this was widespread convergence of cortico-reticular impulses is suggested by the strikingly different patterns of postsynaptic activation observed in neighbouring reticulo-spinal neurons (Fig. 2). The excitatory actions shown in Fig. 2, among the most powerful observed in these experiments, were obtained from another unit of the same animal, within a short period of time. Many of the cortical regions had such a strong linkage to this reticulospinal neuron that it was fired repetitively (Fig. 2, C-F). However there was no EPSP from either the temporal or occipital area (Fig. 2G, H), although the neuron was found only a short distance from that illustrated in Fig. 1. The cell body was located on the left side of the brain stem, and the neuron received a slightly stronger excitation from the left than from the right precruciate area (c$ Fig. 2C, D). References p . 256

248

F. M A G N I A N D W. D. W I L L I S

The latencies of the EPSP’s in Fig. 2, C-F were all less than 1 msec. Therefore, these were undoubtedly monosynaptic, at least initially. The latencies of the EPSP’s in Fig. 1, however, were 2-4 msec. This longer delay may indicate a relay between the cortex and the reticular formation, a relay within the cortex, or simply the need

Fig. 2. Monosynaptic EPSP’s from various cortical areas. (A). The antidromic action potential from cord stimulation at L1. (B). An EPSP and orthodromic action potential from a volley in a pathway traversing the central tegmental tract region of the mesencephalon. (C-H). The effects of stimulation of the indicated cortical areas. The msec time scale applies to all the records. The 50 mV potential scale is for A, and the 2 mV scale for the other records. Note that the extracellular field potential in A was taken at the higher gain (2 mV scale). (From Magni and Willis, 1964a.)

for summation of the effects of slowly conducting corticofugal fibers with those of rapidly conducting ones before there is a detectable potential change. Other reticulospinal neurons receiving undoubted monosynaptic EPSP’s from the cerebral cortex are illustrated in later figures (eg. Fig. 3 shows an EPSP with a latency of 1.4 msec). In general, the most powerful excitatory effects have been produced by volleys from the pre- and postcruciate regions. No consistent differences have been found in the potency of ipsilateral and contralateral stimulation of comparable pericruciate sites. The other cortical areas explored were less effective, although individual reticulo-spinal neurons might receive a larger EPSP from one of these than from the sensorimotor cortex (e.g. Fig. 1). The least effective area of those tried was the occipital region. Inhibitory postsynaptic potentials (IPSP’s) have only occasionally been observed after cortical stimulation, and then only following an EPSP. Examples are shown in Figs. 5G and 9G. The only cortical areas tried which have produced 1PSP’s were in the pericruciate region. The effect of variation of stimulus parameters upon cortically evoked EPSP’s has

AFFERENT CONNECTIONS TO RETICULO-SPINAL NEURONS

249

been studied in some reticulo-spinal neurons. Fig. 3 shows the result of chinging the strength of stimulation upon the size of an EPSP. Th: site of stimulation was the anterior suprasylvian gyrus. The EPSP increased in amplitude as the stimulus strength was increased (0.5, 1, 3, 5 and 10 V in Fig. 3, A-E, respectively). The EPSP rcachd

Fig. 3. Effect of varying stimulus strengths upon EPSP size. A-E show the EPSP‘s and, in D and E, the orthodromic action potentials produced by a variety of strengths of stimulation applied to the cerebral cortex (see text for voltages). The record in F shows the antidromic action potential from cord stimulation at Ll ; note small EPSP in traces without spike. The 10 msec time scale applies to A-E, while the msec scale is for F. The 2 mV potential scale is for all records. (From Magni and Willis, 1964a.)

threshold height for initiating a spike discharge in many sweeps when the stimulus strength was 5V (Fig. 3D), while higher stimulus strengths resulted in firing of the neuron in each sweep (Fig. 3E). Sometimes single shock stimulation was insufficient to produce a large EPSP. When this occurred, it was often possible to evoke large summed EPSP’s by repetitive cortical stimulation. An example is shown in Fig. 4. The small EPSP’s from single shock stimulation of the left and right sensorimotor cortex are seen in Fig. 4E and G, respxtively. Brief tetanic trains of stimuli at 400/sec were employcd to produce the summed EPSP’s from the same cortical regions shown in Fig. 4F, H. Cortically evoked EPSP’s were also observed in neurons which had axons projecting both caudally into the spinal cord and rostrally at least to the mesencephalon. Fig. 5 gives an example of such a case. The antidromic action potentials from spinal cord (Fig. 5, A-C) and mesencephalic (Fig. 5D) stimulation are shown, as are EPSP’s from the left (Fig. 5E, G) and right (Fig. 5F, H) sensorimotor cortex. There is a small IPSP after the EPSP from left cortical stimulation (Fig. 5G). References p. 256

250

F. M A G N I A N D W. D. W I L L I S

Fig. 4. Effect of repetitive stimulation upon EPSP’s. A and B are the antidromic action potentials from cord stimulation at L l . C, E and G are EPSP’s produced by single shock stimulation of a pathway in the region of the central tegmental tract of the mesencephalon, the left and right sensorimotor cortex, respectively. Repetitive stimulation of the same areas (5 stimuli at 400/sec) produced the summed EPSP’s in D, F and H. The msec timer is for A-C, E and G, while the 10 msec timer is for D, F and H. The 50 mV potential scale is for A, while the 2 mV scale applies to the other records. (From Magni and Willis, 1964a.)

Fig. 5. Cortical connections to neuron with ascending and descending axons. (A-C). Antidromic action potentials from cord stimulation at LI . (D). Antidromic action potential from stimulation within the mesencephalon in the region of the central tegmental tract. Postsynaptic potentials from volleys evoked from the left sensorimotor cortex are seen in E and G and from the right sensorimotor cortex in F and H. The msec time scale applies to A-F, and the 10 msec scale to G and H. The 50 mV potential scale is for A, and the 2 mV scale for the remaining records.

A F F E R E N T C O N N E C T I O N S T 0 R E T I C U LO- S P I N A L N E U R O N S

25 1

Efects of stimulation of a pathway in the region of the central tegmental tract Examples of EPSP’s produced in reticulo-spinal neurons by stimulation in the region of the central tegmental tract of the mesencephalon have bcen given in the preceding paper (this Volume, p. 56) and in Figs. 2B and 4C, D. Another instance of this action is shown in Fig. 6. The EPSP produced by various strengths of stimulation (2, 3, 7 and 10 V; 0.2 msec duration) is seen in Fig. 6, E-H. Its latency was 0.5 msec,

Fig. 6 . EPSP’s from region of central tegmental tract. A-D are the antidromic action potentials produced by stimulation of the spinal cord at L l . EPSP’s evoked by a variety of stimulus strengths (see text for voltages) applied through a concentric electrode in the rcgion of the central tegmental tract in the mesencephalon are shown in E-H. The location of a lesion made through the stimulating electrode at the end of the experiment is indicated by the darkened area in 1. The msec time sca!e is for A-H. The 50 mV potential scale is for A and B, while the 2 mV sca!e applies to C-H. (From Magni and Willis, 1361b.)

and so it was monosynaptic. The point of stimulation is indicatcd by the darkened area in the diagram of Fig. 6 (I), which shQws thc extent of a lesion made through the stimulating electrode at the conclusion of the experiment. In a few cats, mesencephalic stimulation was found to produce IPSP’s in some reticulo-spinal neurons. For instance, stimulation at the site indicated in the diagram of Fig. 7E produced both a small EPSP and an IPSP in the neuron illustrated (Fig. 7C). The latency of the EPSP was about 0.5 msec, while that of the IPSP was 0.5-1 msec greater. Several other reticulo-spinal neurons received EPSP-JPSP sequences, a few just IPSP’s ,while many showed only EPSP’s or no effect, after stimulation near this mesencephalic site. References p . 256

252

F. M A G N l A N D W. D. W I L L I S

Fig. 7. EPSP-IPSPsequence from region of central tegmental tract. (A, B). Antidromic action potentials from stimulation of the cord at LI. The change in form of the after-potential accompanied the alteration in spike height as the rccording conditions varied. (C). A small EPSP followed by an IP3P as the result of stimulation in the region of the central tegmental tract in the mesencephalon. (0). An EPSP from stimulation of the left sensorimotor cortex.(E). Diagram showing the location of a lesion placed at the mesencephalic stimulation site. The msec timer applies to A-D. The 50 mV pstential scale is for Aand B, and the 2 mV scale for C and D. (From Magni and Willis, 1964b.)

Fig. 8. Effects of peripheral nerve stimulation upon a reticulo-spinal neuron. A and E show the antidromic action potential produced by a volley set up at the LI level of the spinal cord. The postsynaptic potentials produced by various strengths of stimulation of the superficial radial nerve are shown in B-D and of the deep radial nerve in F-H. See text for details. The monitoring records in this and the following figures were made from the cord dorsum at C1. The msec timer is for A and E, and the 10 msec timer for the remaining records. The 50 mV potential scale is for E, while the 2 mV scale is for the others. (From Magni and Willis, 1964b.)

A F F E R E N T C O N N E C T I O N S TO R E T I C U L O - S P I N A L N E U R O N S

253

Effects of peripheral nerve stimulation The study of peripheral nerve actions upon reticulo-spinal neurons has been limited with respect both to the nerves stimulated and to the neurons from which records have been made. Only two forelimb nerves, the superficial radial (cutaneous) and the deep radial (muscular) nerves, have been employed. The reticulo-spinal neurons studied were only those sending their axons caudally as far as L1. The animals in this series were all pyramidal cats (Whitlock et al., 1953) except for one decerebrate, immobilized by curare. A large proportion of the reticulo-spinal neurons, about half, were unaffected by stimulation of the superficial and deep radial nerves. The majority of the ones affected received either a simple EPSP or IPSP, usually an EPSP. A large number of neurons, however, received a more complex sequence of potentials IPSP-EPSP, EPSP-IPSP, or IPSP-EPSP-IPSP. The actions of the two nerves were generally identical. Fig. 8 shows an example of a reticulo-spinal neuron which received an IPSP-EPSP sequence from each of the two peripheral nerves. The effect of different stimulus strengths is seen, both for the action of the superficial radial (Fig. 8, B-D) and the deep radial (Fig. 8, F-H) nerves. The smallest detectable potential as the strength of stimulation was decreased was an EPSP from the superficial radial nerve (Fig. 8B; strength 1.8 times threshold for the most excitable fibers of the nerve) and an IPSP from the deep radial nerve (Fig. 8F; strength 3T). Raising the stimulus strength from 9 and 12T (Fig. 8C,G) to 30 and 38T (Fig. 8D, H) had little effect on the size of the potentials. The latencies of the IPSP’s were about 7-7.5 msec from the shock artifacts and of the EPSP’e about 19-20 msec.

Fig. 9. Complex response from peripheral nerve volleys. A shows the antidromic action potential and EPSP’s from cord stimulation at L l , B and E are postsynaptic potentials from stimulation of the superficial radial nerve, while C and F are from deep radial nerve volleys. D is from stimulation of the left sensorimotor cortex. The msec time scale is for A-C, and the 10 msec scale is for D-F. The 2 mV potential scale is for all the records. (From Magni and Willis, 1964b.) References p . 256

254

F. M A C N I A N D W. D. W I L L I S

A neuron receiving an IPSP-EPSP-IPSP sequence is illustrated in Fig. 9, the effects of the superficial radial volley being seen in Fig. 9E and that of the deep radial volley in Fig. 9F. The onset of these potentials is seen on a faster sweep in Fig. 9B, C. The neuron also received an EPSP followed by an IPSP from the left (Fig. 9D) and right (not illustrated) sensorimotor cortex. The latencies of the initial 1PSP’s from the peripheral nerve volleys were 8-9 msec from the shock artifacts; the EPSP’s began by 22-23.5 msec; and the late IPSP’s were apparent by 65-70 msec. It is interesting that stimulation of the spinal cord at LI (Fig. 9A) produced only EPSP’s, besides the antidromic action potential. While still recording from the same neuron as in Fig. 9, a study was made of the interaction of the initial IPSP from stimulation of the superficial radial nerve with the EPSP from the left sensorimotor cortex. The sample records in Fig. IOA, B are

s.r. 1. pr. s.r. + 1. p,: predicted. Fig. 10. Interaction of IPSP and EPSP. (A). The first IPSP and the EPSP produced by stimulation of the superficial radial nerve in the same reticulo-spinal neuron that was illustrated in Fig. 9. (B). The EPSP from a volley set up in the left sensorimotor cortex. (C). The effect of interacting the two. The tracings of D are explained in the text. The msec timer and the 2 mV potential scale are for A-C. (Fom Magni and Willis, 1964b.)

the controls, while Fig. 1OC shows the potential resulting from timing the cortical stimulus to fall at the beginning of the IPSP. The tracings of Fig. 10D represent the following: the approximate averages of the potentials produced by nerve and cortical stimulation; the potential predicted if these effects summed algebraically; the potential actually observed during the interaction. The observed potential is less than that predicted for the first 15 msec, while it is greater thereafter. When the cortical stimulus

AFFERENT CONNECTIONS TO RETICULO-SPINAL NEURONS

255

was timed to fall near the omet of the EPSP from peripheral nerve stimulation, the resultant potential was an algebraic summation (not illustrated). The simplest explanation for the interaction pattern observed in Fig. 10 would be as follows. The EPSP may have been reduced in height initially because of the high membrane conductance which occurs during the generation of IPSP’s (Eccles, 1961). The later facilitation could be the result of the convergence of cortico-spinal impulses and peripheral nerve impulses upon the same system of flexor reflex interneurons at the spinal cord level (cf. Lundberg and Voorhoeve, 1962; Lundberg et a/., 1962). DISCUSSION

There is ample evidence, both anatomic and physiologic, of a direct pathway from various areas of the cerebral cortex to the brain stem reticular formation (for references, see Brodal, 1957; Rossi and Zanchetti, 1957; Haartsen, 1962). The present work confirms this and demonstrates that reticulo-spinal neurons are among the reticular neurons receiving direct cortical control. The cortico-reticulo-spinal pathway would presumably be involved in ‘extrapyramidal’ activities. It would be of considerable interest to know how much of the cerebral cortex is capable of producing at least small effects in some reticular neurons. Previous investigations may not have used sensitive enough tests to provide a final answer to this question. Since several of the cortical regions studied in this work are consider:d to be sensory receiving areas, one wonders if corticofugal activity from these areas might have something to do with sensory control. This possibility comes particularly to mind when one considers that reticular neurons send axon collaterals into all the brain stem sensory nuclei (Scheibel and Scheibel, 1958). Little can be said at present concerning the excitatory and inhibitory potentials produced in reticulo-spinal neurons by stimulation in the region of the central tegmental tract. A number of pathways traverse the area, any of which could be responsible. The often complex actions evoked by peripheral nerve stimulation are difficult to explain. Since the cerebellum was removed and since a similar pattern could be observed in a decerebrate, cerebellectomized animal, a complicated supraspinal loop appears to be ruled out. However, it is possible that a lumbosacral cord loop could be involved. Alternatively, the complex actions may take origin in segmental mechanisms or in pathways linking cervical cord and brain stem. SUMMARY

Reticulo-spinal neurons of the cat brain stem have been shown to receive excitatory connections from a wide area of the cerebral cortex. The strength of the connections varies both with the cortical area stimulated and with the particular neuron investigated. The effects of changing stimulus parameters are described. Examples are given of excitatory and inhibitory postsynaptic potentials evoked in References p . 256

256

F. MAG"

A N D W . D . WILLIS

reticulo-spinal neurons by stimulation of a pathway traversing the region of the central tegmental tract of the mesencephalon. The action of peripheral nerves of the forelimb upon reticulo-spinal neurons is shown in many cases to be complex.

A C K N O W LEDCEMENT

This research has been sponsored jointly by the Office of Scientific Research, OAR, through the European Office, Aerospace Research, United States Air Force, under Grant EOAR 62-9, and by the Rockefeller Foundation.

REFERENCES ADES,H. W.,(1959); Ccntral auditory mechanisms. Handbook ofPhysrology, Sect. I , Neurophysiology, Vol. 1 . J. Field et al., Editors. Washington, American Physiological Society (p. 585-613). BRODAL,A., (1957); The Reticular Formation of the Brain Stem. Edinburgh, Oliver and Boyd. DOTY,R. W., (1958); Potentials evoked in cat cerebral cortex by diffuse and by punctiform photic stimuli. J . Neurophysiol., 21, 437464. ECCLES,J. C., (1961); The mechanism of synaptic transmission. Ergebn. Physiol., 51, 299-430. HAARTSEN, A. B., (1962); Cortical Projections to Mesencephalon, Pons, Medulla Oblongata and Spinal Cord. Leiden, Eduard IJdo. A., NORRSELL, U., AND VOORHOEVE, P., (1962); Pyramidal effects on lumbosacral interLUNDBERG, neurones activated by somatic afferents. Acta physiol. scand., 56, 220-229. LUNDBERC, A., AND VOORHOEVE, P., (1962); Effects from the pyramidal tract on spinal reflex arcs. Acta physiol. scand., 56, 201-219. MAGNI,F.,AND WILLIS,W. D., (1963); Identification of reticular formation neurons by intracellular recording. Arch. ital. Biol., 101, 681-702. MAGNI,F.,AND WILLIS,W. D., (1964a); Cerebral cortical control of brain stem reticular neurons. Arch. ital. Biol., submitted for publication. MAGNI,F., A N D WILLIS,W. D., (1964b); Subcortical and peripheral control of brain stem reticular neurons. Arch. ital Biol., submitted for publication. RON, G . F., AND ZANCHETTI, A., (1957); The brain stem reticular formation. Arch. iral. Biol., 95, 199435. SCHEIBEL, M. E., A N D SCHEIBEL, A. B., (1958); Structural substrates for integrative patterns in the brain stem reticular core. Reticular Formation of the Brain. H. H . Jasper et al., Editors. Henry Ford Hospital Symposium. Boston, Little, Brown and Co. (p. 31-55). WHITLOCK, D. G., ARDUINI,A., AND MORUZZI,G., (1953); Microelectrode analysis of pyramidal system during transition from sleep to wakefulness. J. Neurophysiol., 16, 414429.

DISCUSSION PHrLLiPs: I can't help drawing attention to the increasing transmitting potency of pyramidal synapses on the reticular spinal neurons. Dr. Lundberg showed the same phenomenon for spinal interneurons. We now have this phenomenon in two species of animal and on three different types of neurons. These synapses really need morphological investigation to see if there is anything special about the vesicles or the mitochondria or anything else.

WILLIS : The enhanced transmitter potency which we observed in reticulo-spinal neurons in response to repetitive cortical stimulation was strikingly similar to that

A F F E R E N T C O N N E C T I O N S TO R E T I C U L O - S P I N A L N E U R O N S

257

seen at other synapses made by corticofugal fibers. We would agree with Dr. Phillips that this is an important property of these synapses and that this deserves fuller investigation. CREUTZFELDT: 1 should like to ask Drs. Magni and Willis how they explain the very short latencies of reticular neurons to cortical stimulation. From the pictures one gets the impression as if the latencies are between 0.8 and 1.5 msec even for stimulation in the visual and auditory area. A monosynaptic relay between the different cortical areas and the reticular nerve cells through the ‘pyramidal tract’ would indeed represent a very remarkable finding as clear antidromic effects in cortical nerve cells have not yet been found after stimulation in the mesencephalic reticular substance. Finally I should like to ask the authors whether they found in their intracellular records a long lasting after-positivity like the humps described by Prof. Granit and collaborators as being indicative for dendritic activity. Such a dendritic action potential should be visible especially in reticular neurons with their large and long dendrites. WILLIS: The latencies of the excitatory potentials produced in reticulo-spinal neurons by stimulation of the visual and auditory cortex were generally longer than those of the excitatory potential from stimulation of the sensorimotor cortex. Whereas the excitatory potentials having the shortest latencies (between just under 1 msec to I .5 msec) were no doubt monosynaptic, those having somewhat longer latencies may or may not have been monosynaptic. We feel they probably were monosynaptic, the impulses being conducted in relatively fine fibers, but we have no proof for this at present. An explanation for the negative results in attempts to fire cortical neurons antidromically from the mesencephalic reticular formation might be that, according to Brodal’s group, there are relatively few cortico-reticular endings at this level. Most cortico-reticular fibers end in the pons and medulla, at least in the cat. We did see ‘hump-like’ depolarizations after the spike potentials in some reticular neurons. We did not study these in detail, and so we could not say whether or not they had the same properties as the ones described by Prof. Granit and his co-workers. It is interesting that in some reticular neurons having low resting potentials a second discharge occurred at the peak of the ‘hump’. LUNDBERG: First of all I would like to congratulate you on your record in velocities. There is one specific point I would like to go into and that is the action of the ascending pathways. It seems to me that a FRA pathway influences these reticulospinal neurons, and this of course is extremely pleasing to us in view of our own results. KUYPERS : Your methods are extremely sophisticated ways of approaching the different connections in the reticular formation. A real problem arises in regard to the cortical projections to the cells in the reticular

258

DISCUSSION

formation. We don’t have that much detailed knowledge of the cat, but in the monkey the temporal and occipital lobe do not give rise to very many pyramidal fibers. Another point concerns the preparation of a brain stem with only a pyramidal tract. I would think that it is very difficult to transect the total brain stem without touching the pyramidal tract or vice versa transecting the pyramidal tract and leaving everything else. According to work done in Brodal’s laboratory, there are cortico-reticular connections in the cat from the temporal and occipital lobes. Although these are less numerous than the ones from the sensorimotor cortex, we presume them to be responsible for the effects we have observed. Our ‘pyramidal’ cats had partial transections in the rostral mesencephalon, leaving little more than the pyramidal tracts and the substantia nigra intact. We do not have any evidence about the route followed by our corticofugal volleys below the rostral mesencephalon, although we feel that the collaterals known to enter the reticular formation of the pons and medulla from the pyramidal tracts are important in this regard.

259

Investigations on Respiratory Motoneurones of the Thoracic Spinal Cord T. A. S E A R S * The John Curtin School of Medical Research, The Australian National University, Canberra

As a fresh approach towards obtaining a better understanding of the mechanism of respiration, intracellular recording has been made from respiratory motoneurones of the spontaneously breathing, anaesthetised cat. In this way, it has been possible to determine some characteristics of the synaptic drives which cause the periodic and alternate discharge of inspiratory and expiratory motoneurones in the thoracic spinal cord, and to establish the nature of some of the segmental reflexes acting on these motoneurones. A preliminary account of these investigations has appeared elsewhere (Eccles et a/., 1962). In related studies I have also investigated some of the factors influencing the discharge of the fusimotor neurones innervating the intercostal muscle spindles (Sears, 1962; 1963). In this lecture I shall briefly describe the results of these investigations, and then discuss them with particular reference to the central control of the proprioceptive reflexes of the respiratory muscles. METHODS

The anatomy and dissection of the intercostal nerves, the method of fixation of the animal, and the electrical methods used for intracellular recording from thoracic motoneurones have been described previously (Eccles et al., 1962). In other experiments (Sears, 1962; 1963) recordings were made of the efferent discharge occurring during spontaneous respiration in thin, naturally occurring filaments of the intercostal nerves. The filaments were freed just prior to the level at which they take an intramuscular course in the intercostal muscles. Thus the terminal destination of each filament in either the external (inspiratory) or the internal (expiratory) intercostal muscle was known with complete certainty. These dissections confirmed that the external and internal intercostal muscles receive their innervation from the two nerves which previously were described as the ‘nerve to the external intercostal muscle’ and the ‘main intercostal nerve’ respectively (unpublished observations mentioned

* Visiting Wellcome Research Fellow. Present address: The Institute of Neurology, The National Hospital, Queen Square, London., W.C.1. Rrfermces p . 2711272

260

T. A. S E A R S

in Eccles et al., 1962). Here I shall describe these nerves as the ‘external’ and ‘internal intercostal nerve’ respectively, since the terms connate both the anatomical relationship of the two nerves, and the intercostal muscle which each innervates. RESULTS

Eferent discharges in the intercostal nerves Patterns of activity in spontaneous respiration. It is appropriate to give first an account of the patterns of efferent discharges which may be recorded from the intercostal nerves in the lightly anaesthetised, spontaneously breathing cat. Such an account will also facilitate the understanding of the intracellular recordings to be described later. Although there were variations in the patterns of activity in different animals, according to the particular filament from which recording was made, and to the depth of anaesthesia, these patterns consistently showed certain distinctive features. Typical

Fig. 1. Efferent discharges in intercostal nerves of the spontaneously breathing, anaesthetised cat (sodium pentobarbital). Upper traces show monophasic recordings from expiratory nerve filaments in A, B and C, and from an inspiratory nerve filament in D. Lower traces, diaphragm EMG. Time scale 1 sec, voltage calibration 100 pV (a-spikes retouched). From Sears, 1963. (With courtesy of Nature.)

RESPIRATORY MOTONEURONES

26 1

recordings from expiratory nerve filaments of three different animals are shown in the upper traces of Fig. IA, B and C, and from an inspiratory nerve filament in Fig. 1D. The lower traces show the diaphragm electromyogram to denote the phase of inspiration. The activity in expiratory and inspiratory filaments consists of spikes of two distinct sizes. The small spikes of the expiratory filament of A, fired throughout each expiratory pause, and they were inactive during inspiration. The small spikes of B and C fired continuously throughout the respiratory cycle but their discharge frequency was modulated with a respiratory periodicity. This frequency was lowest at the height of inspiration (about 8 cjsec in B and 50 c/sec in C) but it accelerated the instant inspiration ceased and continued progressively to increase to reach a maximum towards the end of the expiratory pause (24 c/sec in B and 100 cjsec in C ) . C also showed a progressive recruitment in the number of small spikes, a phenomenon which was very striking in some filaments. During inspiration, there was a decrease both in the discharge frequency and in the number of active units in these expiratory nerve filaments. It should be noted that this decrease occurred before the onset of activity in the diaphragm EMG. In contrast, the large spikes of the expiratory nerve filaments fired at much lower frequencies, not usually greater than 8-10 c/sec in quiet respiration, and they did not commence firing until the latter half or two thirds of the expiratory pause (cf. the timing of the ‘active’ phase of the expiratory pause described by Sears, 1958); they also ceased firing, prior to the onset of inspiration, earlier than the small spikes. Small and large spikes discharging with a respiratory periodicity were also present in inspiratory nerve filaments, as illustrated in D, but their discharge occurred during inspiration. When the small spikes of inspiratory nerve filaments fired throughout the respiratory cycle, as often they might, their frequency accelerated at the onset of, and continued to increase during, the inspiratory phase of respiration. As in the case of the recordings from expiratory nerve filaments, the acceleration in frequency of the small spikes was apparent before the onset of large spike activity. Very often, especially in expiratory nerve filaments, there were only small spikes present during spontaneous respiration, such activity being phased during inspiration or expiration according to the filament. The subsequent occurrence of large spikes, whether spontaneously, or whether they were evoked such as by the Hering Breuer reflexes (Sears, 1963), confirmed that the small spikes had been correctly identified. In recordings made from 39 nerve filaments innervating the external intercostal muscle, the activity was either wholly confined to the inspiratory phase of respiration, or, the discharge frequency and the number of active units was then greatest. Conversely, the activity of 64 filaments innervating the internal intercostal muscle was always maximal during the expiratory phase. This phasing of the activity in the two types of filaments was the rule for all thoracic segments. Exceptions were always shown to be due to ‘false’ leading from nearby active muscle fibres. The constancy of these findings provides a physiological basis to justify the use of the terms inspiratory and expiratory nerve filaments. Function ofjbres giving rise to small and large spikes. Hunt (1951) showed on limb muscle nerves that spike potential size was an adequate criterion for identifying the References p . 2711272

262

T. A. S E A R S

activity of a- and fusimotor fibres. The large and small spikes were so well differentiated in the present work that th3y may be assigned with assurance to groups of large and small diameter fibres (cf. Blair and Erlanger, 1933; Gasser and Grundfest, 1939). The presence of such grouping of the motor fibres in intercostal nerves has been established histologically (Sears, 1963). By utilising combined mechanical and electrical recording Andersen and Sears (1963) have shown that excitation of the group of high threshold motor fibres in the intercostal nerves does not lead to the development of external muscle tension so that a fusimotor fibre function may be attributed to these fibres (cf. Kuffler et al., 1951). Time does not permit me to cite the other evidence in detail, which derived from the use of electromyography, but there is therefore strong justification for concluding that the small spikes arise in fusimotor fibres and the large spikes in a-motor fibres (Sears, 1962, 1963). Ekland, et al., (1963) have also recorded large and small spike activity from nerve filaments innervating the external intercostal muscles. This finding confirmed the conclusion reached earlier by Critchlow and Von Euler (1962), on the basis of recordings they made of the afferent discharges from intercostal muscle spindles, that the intercostal fusimotor neurones are subjected to a rhythmic control. Some factors influencing intercostal a- and fusimotor neurone activity. It has been shown that the activity of fusimotor neurones innervating inspiratory and expiratory intercostal muscle spindles is phased during inspiration and expiration respectively. This phased activity characteristic of spontaneous respiration is central in origin since it persists in the flaxedil-paralysed animal, and it is not abolished by section of the dorsal roots (Sears, 1963). According to Ekland et al. (1963), the activity of the fusimotor neurones innervating the external intercostal muscle spindles may also be driven through a spinal reflex. However, they did not state to what extent in their experiments this spinal mechanism contributed towards the periodic fusimotor neurone discharge present during spontaneous respiration. The effect of hypercapnia, hypoxia or the pharmacological excitation of peripheral

Fig. 2. Effect of inhaling 5 % COZ in air on efferent discharges in the intercostal nerves. The upper traces show recordings from an expiratory nerve filament, and the lower traces from an inspiratory nerve filament. A = breathing air; only fusimotor discharges present in the expiratory nerve filament. B = during inhalation of 5 "/u COZin air; note inhibition of expiratory fusimotorneurone discharge during inspiration.

RESPIRATORY MOTONEURONES

263

chemoreceptors, was generally to augment the activity of a-and fusimotor neurones during the relevant phases of the respiratory cycle. This augmentation consisted of an increase in the discharge frequency of already active units, and the recruitment of fresh units into activity. However, as inspiratory activity increased, correspondingly there was greater inhibition of any expiratory fusimotor neurone activity occurring during inspiration, as illustrated in Fig. 2. It is well known that hyperventilation caused by artificial respiration abolishes spontaneous respiration. During an apnoea so induced the periodic activity of a- and fusimotor neurones was abolished, as seen in the record D of Fig. 3, although the

Fig. 3. Effect of hyperventilation by artificial respiration on efferent discharges in an expiratory nerve filament. Upper traces, recordings from filament; lower traces, diaphragm EMG. A = control during spontaneous respiration. B and C = during artificial respiration when the animal was apnoeic (C at increased amplification), D = during the apnoea shortly after stopping the pump. E = approximately 20 sec after D when the animal first began to breathe. Voltage calibration 50 pV.

fusimotor neurones continued to discharge tonically and asynchronously. This record was taken from an expiratory nerve filament at the beginning of an apnoea which lasted for about 20 sec, with the respiratory pump stopped. At the first sign of renewed inspiratory activity, the fusimotor neurone discharge in the expiratory nerve filament was inhibited (Fig. 3E); it then reappeared during the first expiratory pause (note the absence of a-motoneurone activity) and subsequently with each expiration. References p . 2711272

264

T. A. S E A R S

a-Expiratory motoneurone activity did not resume in this filament until 4 more breaths were taken. When artificial respiration was applied while the animal was apnoeic, as in Fig. 3B, there then occurred in the expiratory nerve filament a periodic burst of repetitive firing of an a-spike with each inflation of the lungs and chest (Fig. 3B and C ) . The faster rhythm of the pump (30 strokes per min) compared to the rhythm of spontaneous respiration seen in Fig. 3A (about 15 per min), is clearly evident. By increasing the stroke volume of the pump, the discharge frequency was increased; by reducing the stroke volume to below a critical value the discharge was abolished. Since it was found that such activity was abolished by sectioning the dorsal roots in the same and adjacent segments, it is concluded that the response is due to a segmental reflex initiated from proprioceptors, presumably muscle spindles (Huber, 1902) which are excited by chest inflation. This reflex is the inflation reflex of expiratory muscles previously demonstrated by Sears ( I 958) using electromyography. It is probable that the spinal mechanism of this reflex is the same as that subserving the response of the intercostal muscle to stretch described by Ramos and Mendoza (1959). INTRACELLULAR

RECORDING

FROM

RESPIRATORY

MOTONEURONES

OF

THE THORACIC S P I N A L CORD

‘Central respiratory drive potentials’. Glass microelectrodes (resistance 5-10 M Q), filled with 3 M potassium chloride or 2 M potassium citrate, were pushed into the spinal cord just medial to the root entry zone through a small hole made in the pia. The regions containing motoneurones were located by searching for the motoneuronal field potentials evoked by stimulating the ipsi-segmental intercostal nerves. Impaled cells were identified as motoneurones by their ability to produce antidromic somadendritic spikes (Brock et al., 1953). In the spontaneously breathing, lightly anaesthetised animal the membrane potentials of different motoneurones were subjected to slow, rhythmic fluctuations having a respiratory periodicity as illustrated in Fig. 4. The motoneurones designated as inspiratory and expiratory were so identified according to whether an antidromic soma-dendritic spike was evoked in them by stimulation of the external or internal intercostal nerves respectively. These records were obtained within 15 min of each other during which time the rhythm of breathing was essentially unaltered. The records have, therefore. been arranged one above the other as if they had been taken simultaneously. In the inspiratory motoneurone, the depolarising phase of its slow potential occurred during inspiration as registered by the diaphragm electromyogram. On the other hand, the depolarising phase of the slow potential in the expiratory motoneurone occurred during the expiratory pause. The maximum amplitudes of such slow potentials were invariably greater in inspiratory than in expiratory motoneurones impaled in the same animal. Slow potentials such as those illustrated are normal occurrences in thoracic respiratory motoneurones of the spontaneously breathing animal. Since the periodic firing of respiratory motoneurones is causally dependent on

RESPIRATORY MOTONEURONES

A

B

-42

265

EX P I R A T 0 R Y MOTON EURO N E

INSPIRATORY MOTONEURONE

t SEC

Fig. 4. 'Central respiratory drive potentials'. Upper traces, intracellular d.c. recordings from thoracic respiratory motoneurones. B = recorded approximately 15 min after A; records aligned above each other according to the diaphragm EMG (lower traces).

C

D

1 1 1 1 1 1 se c

Fig. 5 . A and C intracellular d.c. recordings from an inspiratory motoneurone, C recorded 10 sec after A. D same as C , recorded at low gain with a C.C. amplifier (time constant of 0.02 sec). B = diaphragm EMG, recorded simultaneously with A. From Eccleset al., 1962. (With courtesy of Nature.) :

References p. 271J.272

266

T. A. S E A R S

these rhythmic slow potentials it has been suggested that they be called 'central respiratory drive potentials', abbreviated to CRDPs (Eccles et al., 1962). The intracellular recordings shown in Fig. 5, illustrate how the periodic repetitive discharge of an inspiratory motoneurone is related to the depolarising phase of its CRDP. Repetitive discharge occurs throughout that period of the CRDP when the membrane potential is lower than the critical firing threshold the actual discharge frequency being determined by the magnitude of the depolarisation below this level ; accommodation may occur. The recording from the expiratory motoneurone illustrated in Fig. 6 is of double interest. Firstly, it shows the transition from a phase of re-

i'l"

A

- 60

B I,ll"llol

.4

..,

C " " ~ ' " * l

sec

Fig. 6. A = intracellular d.c. recording from an expiratory motoneurone. Note firing during the depolarising phases of the first two cycles of the CRDP. B - diaphragm EMC. From Eccles et al., 1962. (With courtesy of Nature.)

petitive discharge throughout the latter half of the expiratory pause, to firing late in the pause, and finally, to a phase of CRDPs alone, due to a steady increase in the average membrane potential and a decrease in the amplitudes of successive cycles of the CRDP. Secondly, it illustrates a different form of the CRDP in that the membrane potential progressively diminished during the expiratory pause, whereas the membrane potential of the expiratory motoneurone of Fig. 4 remained relatively constant throughout the pause. These different forms of the CRDPs and the different discharge patterns they would be expected to evoke, clearly provide an explanation of the different 'activity patterns' of respiratory motoneurone discharge described by Gesell et al., (1940). Some minutes after respiratory motoneurones were impaled with potassium chloride filled electrodes, it was observed repeatedly that the hyperpolarising phases of their

RESPIRATORY MOTONEURONES

267

CRDPs were converted to phases of depolarisation so that there were then two phases of depolarisation for each respiratory cycle. Reversals of this sort were also induced by passing hyperpolarising currents through the microelectrode; they were not seen with potassium citrate filled electrodes. From the work of Coombs et al., (1955) on the ionic mechanism of inhibition it can be concluded that the hyperpolarising phase of the CRDP is, in fact, a phase of postsynaptic inhibition causcd by the release of inhibitory transmitter. Evidently, in spontaneous respiration, respiratory motoneurones are subjected alternately to periodic barrages of excitatory and inhibitory impulses. I n some expiratory motoneurones, it appeared from records taken after reversal had occurred, that the major part of the CRDP was due to the periodic inhibitory bombardment. The significance of the inhibitory phase of the CRDP is discussed later. Monosynaptic excitation of thoracic respiratory motoneurones. A synaptic potential of brief latency, similar in form to the monosynaptic excitatory postsynaptic potential (EPSP) of lumbosacral motoneurones, was evoked in thoracic motoneurones by stimulation of low threshold afferent fibres in the intercostal nerves (Fig. 7). This

w Fig. 7. Repetitive activation of monosynaptic pathway. Upper traces, monosynaptic EPSP evoked at the frequencies indicated in cjsec. Lower traces, afferent volley recorded from cord dorsum.

synaptic potential increased in amplitude, without significant changes in form, when the stimulus intensity was increased and it was usually maximal, or within 10 to 20 % of maximal, at the axon threshold of the impaled motoneurone. Since the onset of the synaptic potential occurred within 0.6 to 0.8 msec of the arrival of the afferent volley at the cord dorsum, so allowing time for only one synaptic delay (Brock et al., 1952), the existence of a monosynaptic pathway to thoracic motoneurones was thereby demonstrated (Eccles et al., 1962). The convergence of monosynaptic excitation on respiratory motoneurones was determined by stimulating several intercostal nerves in turn and measuring the amplitudes of the monosynaptic EPSPs so evoked (cf. Eccles et al., 1957). The respiratory motoneurones of all segments examined (T.5 to T. 11) received monosynaptic excitation from the ipsi-segmental internal intercostal nerve, and from one or other, or both, of the juxta-segmental internal intercostal nerves. On the other hand, about 30 % of inspiratory motoneurones did not receive any monosynaptic excitation from the ipsi-segmental external intercostal nerve. Some of these cells comprised the 70 % of inspiratory motoneurones which unexpectedly were found to receive monosynaptic excitation from the ipsi-segmental internal intercostal nerve. References p . 2711272

268

T. A . S E A R S

Repetitive activation of the monosynaptic pathway. In about 65 % of motoneurones examined, the amplitude of the monosynaptic EPSP showed a striking potentiation with increasing frequeiicy of repetition. For example, in the recordings from an expiratory motoneurone shown in Fig. 7, increasing the repetition frequency from 1.0 cjsec to 43.0 cjsec caused more than a two-fold increase in the amplitude of the EPSP. This absolute potentiation of the monosynaptic EPSP in many cells (and the absence of significant depression in the remainder) may be contrasted to the absolute depression of the monosynaptic EPSP observed in most lumbosacral motoneurones (Curtis and Eccles, 1960). One consequence of this potentiation and the lack of depression, is that during repetitive activation of the monosynaptic pathway, successive EPSPs summate to produce a sustained depolarisation which may be several-fold greater in amplitude than that of single EPSP. Such a depolarisation can cause repetitive firing of the motoneurone as illustrated in Fig. 8. This expiratory motoneurone showed a CRDP,

Fig. 8. Summation of repetitively evoked monosynaptic EPSPs with CRDP in an expiratory motoneurone. A = antidromic somadendritic spike. B = superimposed EPSPs evoked at 9, 33 and 50 c/sec to show absence of depression (as. recording). C, D, E and F = d.c. intracellular recordings to show a t the left side of each trace in CRDP alone. The right side of each trace shows the CRDP in summation with the depolarisation evoked by repetitive stimulation of the monosynaptic pathway at the frequencies indicated in c/sec. Time scale 1 sec.

one cycle of which is shown as control at the left side of the traces in C , D, E and F. When the ipsi-segmental internal intercostal nerve was stimulated at 50 cjsec and 80 cjsec the evoked depolarisation summed with the concurrent CRDP without causing firing of the motoneurone. However, at 100 cjsec and 125 cjsec, the extra

RESPIRATORY MOTONEURONES

269

depolarisation was adequate to cause repetitive firing but only when in summation with the depolarising phase of the CRDP. Such recordings illustrate the physiological effectiveness of the segmental input. In fact, discharge frequencies as high as 40 impulses per sec have been evoked in expiratory motoneurones in this way. It may be noted, that the effectiveness of the intercostal monosynaptic pathway can only truly be judged when the pathway is activated repetitively. DISCUSSION

In spontaneous respiration thoracic respiratory motoneurones show slow, rhythmic fluctuations of their membrane potentials. These slow potentials have been named ‘central respiratory drive potentials’ on account of the causal relationship they clearly bear to the periodic discharge of respiratory motoneurones. The CRDPs consist of phases of depolarisation and hyperpolarisation in alternating succession, the phase of hyperpolarisation being a phase of synaptically induced inhibition, It is not intended to consider here how these CRDPs come about, for example, by associating them with any particular hypothesis concerning the functional organisation of the respiratory centres. To the best of my knowledge, however, no hypothesis has been formulated to include the possibility that spinal respiratory motoneurones are subjected alternately to excitatory and inhibitory impulses. The phased inhibition is of considerable functional significance since it provides one means by which the central nervous mechanism of respiration exerts a control over the segmental proprioceptive reflexes of respiratory muscles. This problem will now be discussed with reference to the ‘inflation’ reflex of expiratory muscles. If spontaneous respiration is abolished by hyperventilation or by spinalisation, it may be demonstrated that when the chest is inflated, activity is evoked in expiratory muscles due to th.: operation of a spinal reflex. The effective stimulus is almost certainly the stretchmg of the intercostal muscles, and the monosynaptic pathway is presumed to provide the afferent limb of the reflex arc. As I pointed out previously (Sears, 1958), when the animal is breathing spontaneously, the ‘inflation’ reflex must be inhibited during inspiration since expiratory muscles are then either inactive or least active even though the expiratory muscles are stretched and the reflex pathway excited. Such inhibition is clearly provided for by the hyperpolarising phase of the CRDP, the magnitude of which appears to be related to the amount of inspiratory activity (i.e., the depth of inspiration). On the other hand, when the activity of the central nervous mechanism of respiration is abolished by hyperventilation or by spinalisation, so abolishing the CRDPs, the inflation reflex is dis-inhibited. Since a monosynaptic pathway to inspiratory motoneurones from inspiratory muscles has also been demonstrated, the phased inhibition of inspiratory motoneurones is presumably required to inhibit their reflex activation which would otherwise occur during expiration. The question now arises, under what circumstances does the monosynaptic pathway become effective in controlling respiratory motoneurone discharge? The answer to this question resides in the nature of the rhythmic control exerted over the inRefrrrnces p 2711272

270

T. A. S E A R S

tercostal fusimotor neurones. Intercostal fusimotor neurone discharge leads in time on a-motoneurone dischargc, and so leads on muscular contraction. It is evident, therefore, that the afferent discharge from the relevant intercostal muscle spindles must be driven in the manner conceived by Eldred et a/. (1953) in their ‘follow-up s m o ’ theory for the nervous control of muscular contraction (see also Granit, 1955; Hammond et al., 1956). In this theory, by causing the intrafusal muscle fibres to contract, S O extending the sensory portion of the muscle spindle, fusimotor neurone dischargc leads to an increased afferent discharge from the muscle spindle as if the muscle itself had been stretched. As a consequence, the stretch reflex pathway is activated and the resulting a-motoneurone discharge causes the muscle to shorten to a length proportional to the shortening of the spindles. Ln following the length of the muscle spindle (the latter measuring thc misalignment between the spindles and the muscle), the changing length of the muscle satisfies the ‘demand’ for a certain movement to occur, and this ‘demand’ is seen to be conveyed initially through the fusimotor fibre system, the so-called ‘indirect route of muscle activation’. The ‘demand’ in the situation considered here, appears to be related to the ‘demand’ for a certain pulmonary ventilation, since hypocapnia abolishes, and hypxcapnia and hypoxia augment, the periodic discharge of the iiispiratory and expiratory fusimotor neurones. More specifically, the system responds as if the time course of inspiratory and expiratory fusimotor neurone discharge represents the ‘demanded’ time course of inspiration and expiration appropriate to the prevailing chemical drive to respiration. Thus any mechanical factor which causes the shortening of the extra fusal fibres to lag on the shortening of the intrafusal fibres, such as an increased resistance to air flow to and from the lungs, would lead to an increased discharge from the inuscle spindles with the consequence that the intercostal motoneurones would be subjected to an increased excitatory drive. The extra depolarisation so evoked would summate with the concurrent central respiratory drive potential to increase the discharge frequency of active motoneurones and to recruit others into activity (cJFig. 8). The inspiratory and expiratory phasing of the fusimotor neurone discharge to inspiratory and expiratory muscle spindles, ensures that the information concerning the degree of misalignment between the intra- and extrafusal muscle fibres is fed back from the appropriate muscle according to the phase of the respiratory cycle under way. The possibility that the ‘indirect’ route of muscle activation is concerned in the regulation of respiratory movements, was suggestcd previously by Nathan and Sears (1960) on the basis of the paralysis of respiratory muscles which they observed following restricted dorsal root section in man. This possibility has also been considered i n some detail by Campbell and Howell (1962, 1963). The mechanism discussed above clearly provides a means by which the force and time course of respiratory movements can be controlled, either subtly in the moment to moment adjustments characteristic of normal breathing, voluntarily as in spcech, and during the maximal respiratory efforts required in coughing and straining. The importance of the proprioceptive reflexes of respiratory muscles, and of their central control, is thus clearly established.

RESPIRATORY MOTONEURONES

27 1

SUMMARY

An account has been given of some recent investigations on respiratory motoneuroiies of the thoracic spinal cord. Intracellular recording shows that during spontaneous respiration the membrane potentials of respiratory motoneurones are subjected to slow, periodic changes which have been called ‘central respiratory drive potentials’ (CRDPs). The CRDP consists of a phase of depolarisation during which the cell may discharge repetitively, followed by a phase of hyperpolarisation when the cell is inhibited; these changes occur reciprocally in inspiratory and expiratory motoneurones. Some features of the monosynaptic pathway to thoracic respiratory rnotoneurones are also described. Particular emphasis is placed on the potentiation of the monosynaptic EPSP when the pathway is activated repetitively, since the resulting summed depolarisation can cause repetitive discharge of the inotoneurone over a wide range of frequencies. Other experiments show that the activity of fusimotor neuroiies innervating inspiratory and expiratory muscle spindles, like the activity of the corresponding u-motoneurones, is phased during inspiration and expiration respectively. The characteristics of the fusimotor neurone discharge lead to the suggestion that the intercostal muscle spindles serve as misalignment detectors in a ‘follow-up length servo’ mechanism importantly concerned in the regulation of respiratory movements. REFERENCES ANDERSEN, P., AND SEARS,T. A., (1963); Submitted for publication. BLAIR,E. A., AND ERLANGER, J., (1933); A comparison of the characteristics of axons through their individual electrical responses. Amer. J. Physiol., 106, 524-570. BROCK,L. G., COOMBS, J. S., AND ECCLES,J. c.,(1952); The recording of potentials from motoneurones with an intracellular electrode. J. Physiol. (Lond.), 117, 431460. BROCK, L. G., COOMBS, J. S., AND ECCLES, J. C., (1953); Intracellular recording from antidromically activated motoneurones. J. Physiol. (Lond.), 122, 429461. CAMPBELL, E. J. M., AND HOWELL, J. B. L., (1962); Proprioceptive control of breathing. Ciba Found. Symp. Pulmonary Structure and Function. London, Churchill (p. 2 9 4 5 ) . CAMPBELL, E. J. M., AND HOWELL, J. B. L., (1963); The sensation of breathlessness. Brit. med. Bull., 19, 36-39. COOMBS, J. S., ECCLES,J. C., AND FATT,P., (1955); The specific ionic conductances and the ionic movements across the motoneuronal membrane that produce the inhibitory postsynaptic potential. J. Physiol. (Lond.),130, 326-373. CRITCHLOW, V., AND VON EULER,C., (1962); Rhythmic control of intercostal muscle spindles. Experientia (Basel), 18, 426427. CURTIS,D. R., AND ECCLES,J. C., (1960); Synaptic action during and after repetitive stimulation. J. Physiol. (Lond.), 150, 374-398. ECCLES,J. C., ECCLES,R. M., AND LUNDBERG, A., (1957); The convergence of monosynaptic excitatory afferents onto many different species of a-motoneurone. J. Physiol. (Lond.), 137, 22-50. ECCLES,R. M., SEARS,T. A., AND SHEALY, C. N., (1962); Intracellular recording from respiratory motoneurones of the thoracic spinal cord. Nature (Lond.), 193, 844-846. EKLAND, G., VONEULER,C., AND RUTKOWSKI, S., (1963); Intercostal y-motor activity. Ac/aphgsio/. scand., 57, 48 1 4 8 2 . ELDRED, E., GRANIT,R., AND MERTON, P. A., (1953); Supraspinal control of the muscle spindles and its significance. J . Physiol. (Lond.), 122, 498-523. GASSER, H. S., AND GRUNDFEST, H., (1939); Axon diameters in relation to the spike dimensions and the conduction velocity in mammalian A-fibres. Amer. J . Physiol., 127, 393-414.

272

DISCUSSION

GESELL, R., MAGEE,C., AND BRICKER, J. W., (1940); Activity patterns of the respiratory neurons and muscles. Amer. J . Physiol., 128, 615-628. GRANIT, R., (1 955); Receptors and Sensory Perception. New Haven, Yale University Press; London, Cumberlege. HAMMOND, P. H., MERTON,P. A., AND SUTTON,G. G., (1956); Nervous graduation of muscular contraction. (Physiology of voluntary muscle). Brit. med. Birll., 12, 214-218. HUEER, G. C., (1902); Neuro-muscular spindles in the intercostal muscles of the cat. Amer. J. Anat., 1, 520-521. HUNT,C. C., (1951); The reflex activity of mammalian small-nerve fibres. J . Physiol. (Lond.), 115, 456469, KUFFLER, S. W., HUNT,C. C., A N D QUILLIAM, J. P., (1951); Function of medullated small nerve fibres in mammalian ventral roots: Efferent muscle spindlc innervation. J . NeurophyJiol., 14,29-54. NATHAN,P. W., AND SEARS, T. A., (1960); Effect of posterior root section on thc activity of some musclcs in man. J . Neurol. Neiiromrg. Psychiat., 23, 10-22. RAMOS,J . G., AND MENDOZA, E. L., (1959); On the integration of respiratory movements 11. The integration at spinal level. Acta physiol. laf.-amer., 9, 257-266. SEARS,T. A., (1958); Electrical activity in expiratory muscles of the cat during inflation of the chest. J. Physiol. (Lonrl.), 142, 35P. SEARS,T. A., (1962); The activity of the small motor fibres system innervating respiratory muscles of the cat. A u f . J . Sci., 25, 102. SEARS,T. A., (1963); Activity of fusimotor fibres innervating muscle spindles in the intercostal muscles of the cat. Nature (Lond.), 197, 1013-1014.

DISCUSSION

LUNDBERG: I think we must state very clcarly that thanks to the systematic and extremely fine work of the Canberra group these types of reflexes are now being regarded as important physiological mechanisms. GELFAN:Regarding the distribution of a- and fusiform fibers I suspect that for the intercostal fibers they may be half and half. Do you have any quantitative data on this fiber spectrum? SEARS:I have determined the fiber calibre spectra of motor and sensory fibers in the intercostal nerves, In chronically de-afferented nerves the calibre spectrum of the motor fibers is bi-modal. There is a well-defined peak at 4 to 6 p, intermediate sized fibers from 6 to 8 p, and a broad range of large diameter fibers extending from 8 t o 20 p. The largest diameter fibers are thus seen t o be as large as any found in limb muscle nerves. The broad range of large diameter fibers gives this end of the spectrum the form which would be obtained by combining the calibre spectra of limb nerves t o ‘slow’ and ‘fast’ muscles such as the soleus and gastrocnemius muscles. In fact, Dr. Andersen and I (submitted for publication) have found in agreement with Glebovskii (1961), but contrary to Biscoe (1961, 1962),* that the intercostal muscles are comprised of a mixture of fast and slow motor units, although the latter are nothing like as slow as the fibers of the soleus muscle. GRANIT:1 merely want to put on record that similar work has been carried out in Stockholm at the Nobel Institute for Neurophysiology by Prof. Curt Von Euler. I am happy to find that inasmuch as the two overlap, they are in essential agreement.

*

See list of references at the end of the main paper.

RESPIRATORY MOTONEURONES

273

In the main, however, they supplement each other and in both cases there is clear evidence for the great importance of the gamma loop in respiration.

SZENTAGOTHAI : D o you have convergence of primary sensory afferents from the second neighbouring segments? Already in 1939 I was amazed at the large number of primary sensory endings on motoneurons in the thoracic cord, as found by the aid of degeneration. Did you find any respiratory drive on motoneurons that send their axons into the dorsal branches of thoracic nerves? SEARS:Individual motoneurons showed considerable variation in the pattern of the monosynaptic innervation they received from the adjacent segments. Several motoneurons were impaled which received a small amount of monosynaptic excitation from two segments away. These were all expiratory motoneurons and this finding probably reflects the synergism to be expected between the internal intercostal muscles of adjacent segments, and between these muscles and the abdominal oblique muscle since both have an expiratory function. In answer to your second question, the motoneurons with axons in the dorsal spinal rami did not show ‘central respiratory drive potentials’, at least under the conditions of quiet, unstimulated respiration which were essential to the success of intracellular recording in the spontaneously breathing animal.

JUKES: I have one or two points I would like to make. Anaesthetics alter the sensitivity of the respiratory center on the whole respiratory apparatus, both to PC02 and anoxia, and may do so in different directions so that patterns of respiration you get with activity in the a- and other fibers may be different with different anaesthetics. There is another point which I would like to mention. If you cut all the dorsal roots in a cat the cat still goes on breathing, but unfortunately there are no studies on the relationship in such cats between PC02, anoxia and ventilation. If you cut the vagus you get an enormous increase in depth and a slowing in rate, but they have a normal response to PC02 in terms of total ventilation. Do you have any experience with micro-electrodes in spinal animals? SEARS:Thank you for your comments. I did of course mention at the beginning of the lecture the influence of changing levels of anaesthesia on the pattern of discharge of a- and fusimotor fibers, but this whole question must be gone into in much greater detail. In the few experiments I have done on spinal animals I have seen nothing comparable to the ‘central respiratory drive potentials’ present during spontaneous respiration. In such experiments it is necessary to exclude the possibility of segmental reflex activation of the motoneurons, but again, further experimentation is required before these questions can be properly answered.

274

Reflexes to Toe Muscles in the Cat’s Hindlimb I. E N G B E R G Departtneni of Pliysiology, University of Giiirhorg, Ciiiehorg (Sweden)

The dominating spinal reflex response to activation of several afferent systems from the extremities is the flexion reflex. This has mainly been regarded as a single unitary reaction with the same type of activation of all muscles that shorten the limb whether the reflex is evoked by skin afferents, high threshold joint afferents or group I 1 and I I I muscle afferents. However, recent experiments have shown that there may be a more complex situation with different reflex actions from different skin areas (Hagbarth, 1952; Kugelberg, 1962). Already the demonstration ofthe extensor thrust (Sherrington, 1905) gave an example of a highly specialized reflex from the plantar. This was elicited by an innocuous pressure, and he suggested that it could be of some importance i n locomotion. Sherrington (1910) also compared muscle activation in the general flexion reflex with the reflex stepping and he based his classification of muscles into flexors and extensors both on their participation in the flexion reflex and on their activity during stepping. Thus, muscles like flexor digitorum longus and plantaris, extending the ankle and bending the toes in plantar direction were grouped as extensors, and toe dorsiflexors as flexors. The intrinsic muscles of the foot were not studied in this respect, and their reflex connections have not been subject to much investigation. In the following experiments on cats, information of a rather selective reflex from the plantar cushion to some of these muscles and to other toe extensors was found (Engberg, 1963). If one applies a moderate pressure, for example with the tip of a finger, to the central pad of the hindfoot (in an acute spinalized cat) one is able to see and feel activity in muscles plantar flexing the toes. In Fig. I the activity is led off from flexor digitorum brevis during a continuous gentle squeezing of the pad between two fingers. Attempts have been made to eliminate as far as possible stimulation of proprioceptors in the foot - they seem, however, not to be of any importance for this muscle activation, because any manipulation with the rest of the foot or the toes, leading to about the same dislocations as the pad squeezing, is ineffective in itself. As shown in Fig. 1, discharge in the muscle is continued as long as the pressure is kept up, at least within the tested periods of about 10-30 sec. It stops as soon as the stimulus is withdrawn. The minimum stimulus needed varied somewhat from one preparation to another. Sometimes there has been a slight efiect already upon a very gentle pressure by a finger tip, sometimes more pressure applied to the whole surface of the pad is

275

R E F L E X E S TO TOE M U S C L E S

required. In most cases, a pressure of the same type and magnitude as the cat exerts when standing on the leg, will have the effect illustrated in Fig. 1. Weak electrical stimulation with single shocks to the pad (see legend of Fig. 2 for parameters), produces a strongly increased excitability in the motor nuclei of flexor

20 msec

pad pessed

Fig. I . Electromyography from FDB showing excitatory effects from the central pad of the hind-foot. Several superimposed records with monosynaptic test reflexes (evoked by stimulation of the intact tibial nerve) are taken before, during and after the application of a continuous, very light pressure on the pad. The interval between each set of records is about 15 sec. (A reflex discharge is evoked from the pad besides the facilitation of the test). A

FDB

0 FDL 0 PL

0

10

20

30

LO

50 msec

Fig. 2. Facilitation of monosynaptic reflexes to FDB, F D L and PL obtained by electrical stimulation of the central pad with single shocks (condensor discharges of 12 V, half decay time 25 psec). 100 % on the ordinate represents the height of the unconditioned test reflexes. The time is measured between the conditioning and testing stimuli. The drawing indicates the areas from which the same muscles are activated on adequate stimulation. At the central pad only gentle pressure is required, in the dotted area stronger pinching.

digitorum longus (FDL), plantaris (PL) and flexor digitorum brevis (FDB), lasting about 30 msec. This is tested in Fig. 2 by conditioning monosynaptic reflexes evoked in the nerves to these muscles with pad stimuli at different time intervals. The excitatory effects from the pad are much more pronounced to flexor digitorum brevis than to flexor digitorum longus and plantaris. If the monosynaptic test reflex is evoked in the lower part of the tibial nerve supplying all the intrinsic muscles of the foot (except extensor digitorum brevis on the foot dorsum) and recorded in the S2 ventral root, there will be the same strong facilitation from the pad as for flexor digitorum brevis alone. As flexor digitorum brevis forms only a minor part of the References p . 2781279

216

I. ENGBERG

muscle mass in the foot, this indicates that the motor nuclei of several other of these muscles receive a corresponding effect from the pad. If the electric stimulus is applied in the same way to one of the small toe pads, this is not the case. A monosynaptic test reflex recorded by electromyography in flexor digitorum brevis is then facilitated though not as effectively as from the pad, but the ‘mixed’ test reflex of all the foot muscles is inhibited. The adequate stimulus to the toes needed to evoke this inhibition seems to be of a more nociceptive nature. Firm squeezing of the end phalanx or pinching of the skin is necessary, and the inhibition still goes on for a few seconds after the withdrawal of such a strong stimulus. Flexor digitorum longus and plantaris motor nuclei likewise receive inhibition from the toes. There is on the other hand some excitatory effect to all the mentioned motor nuclei evoked by pinching of the plantar skin in the neighbourhood of the central pad (stippled area in Fig. 2).

Pad

L

B Joint

hyperpol

P depol

Pad

I&%--

M

C G -S

Sur

0

ii#hm!b

Joint

0

D

PS

N

Pad

j\p,p

R Joint

-

20ms

I 5 m ~

Fig. 3. The upper tracings in each set are intracellular recordings of synaptic effects in motoneurones evoked by volleys in high threshold joint afferents (Joint), group I1 and I11 afferents from gastrocnemius-soleus (G-S), skin afferents in n. suralis (Sur) and n. peroneus superficialis (PS) and by electrical stimulation of the central pad (Pad). A-E are recorded in an FDB motoneurone, F-J in an interosseus motoneurone, K-P and Q-T in two plantaris motoneurones. In record K the maximal la EPSP in the first plantaris motoneurone is taken with a faster sweep speed. The lower traces are triphasic recordings of the incoming volleys at the dorsal root entry zones; some are taken with a short time constant.

Intracellular recordings from motoneurones in the spinal cord confirm these findings. Excitatory postsynaptic potentials are set up in all the four motoneurones of Fig. 3 on stimulation of the pad. The central delay of this action is 3.5-4 msec (incoming volleys are not clear in the figure). The additional depolarizing synaptic potential seen in the plantaris neurones when they are hyperpolarized (N,S) reveals

R E F L E X E S TO T O E M U S C L E S

277

that some inhibitory action is also evoked by the pad stimulation. The effect to FDL has sometimes shown a gradual change from an excitatory action towards a dominating inhibition when the stimulus strength was increased to a clearly nociceptive one. On the other hand there has been no sign of an inhibitory contribution to FDB or interosseus motor nuclei. In contrast to this excitatory action from the pad, are the dominating inhibitory synaptic potentials evoked by stimulation of flexion reflex afferents from different sources. Volleys in high threshold knee joint afferents (BGMR), group I1 and I11 muscle afferents (CH), and skin afferents in the sural nerve (DI), all give the IPSP’s typical of extensor motoneurones in the general flexion reflex. (There are some exceptions to this, in that FDB and probably some of the other small muscles in the foot receive excitation from skin branches of the superficial peroneal nerve (Fig. 3E) and sometimes from branches of the sural nerve that enter the foot.) It thus seems that these intrinsic foot muscles should be classified together with plantaris and FDL as extensor muscles and this also fits very well with their activity during normal locomotion. Electromyographic investigations have shown that they are activated in the extension phase of the step in a manner that is very similar to that of other extensors of the limb. Their special reflex activation from the pad is, however, not seen to go to any other extensors of the hindlimb. Therefore it would seem to be a reflex function suited to assist in the regulation of muscle tonus necessary to stabilize the metatarsals and toes when the cat puts its weight on the foot. It might be of interest in the discussion of supraspinal control of reflex arcs to see the cortical influence upon this plantar reflex. The mechanism behind the well-known influence of a pyramidal lesion upon the plantar reflexes in man has been elucidated in a recent work by Kugelberg et al. (1960). They concluded that the r61e of the pyramidal tract is to mediate a suprasegmental control of the spinal reflex systems responsible for the different plantar responses! It was postulated by Lundberg and Voorhoeve (1962) that in cats the pyramidal tract exerts influence on motoneurones by exciting interneurones of various spinal reflex arcs. Stimulation of the sensorimotor C o r t e x 4 FOB

Cortex+ short foot muscles

E

A intact

-Av-.

C pyramidal A-v lesion

D -Av-

15 msec

G

H

h f7imrh I“”Q A

+

Fig. 4. The effects of a conditioning stimulation of the sensorimotor cortex on the monosynaptic test reflexes of the short foot muscles. A-D are recorded in the FDB muscle (EMG) and E-H in the S1 S2 ventral roots. The test stimulus was given to the distal part of the intact tibia1 nerve. The cortical stimulation consisted of trains of square wave pulses each of 0.2 msec duration and a strength of 1 mA. A, C , E, and G are test stimuli alone, B, D, F, and H are the same tests conditioned from cortex. The lower row of records are taken after section of the contralateral pyramid. The extent of the lesion is shown in the drawing. (The lower traces in E-H are recorded at the dorsal root entry zone of L7.)

+

Rekrences p . 278/279

278

I. E N G B E K G

cortex thus results in an inhibitory action on the motor nuclei of extensor muscles in general, in parallel with the strong inhibitory effect of the flexion reflex. In contrast to this, a dominating excitatory influence is exerted by the sensorimotor cortex on the motoneurones of the plantar nerves as seen in Fig. 4. A-D show the action in FDB motoneurones and E-H the summation of effects to the whole group. The effect is mediated via the pyramidal tract as shown by the control records taken after section of the pyramid. It can be demonstrated that there is spatial facilitation between the excitatory paths from cortex and from the pad. This is done in Fig. 5 by adjusting each conditioning

A

test

6

pad +test

C

cortex+test

D cortex+pad+test

-

20msec

Fig. 5. Spatial facilitation between the effects from cortex and from pad on the same monosynaptic test reflex that was used in the experiment of Fig. 4. The conditioning stimulations of pad and cortex are adjusted to give minimal effects when given alone (B, C) and then combined (D).(Lower traces recorded as in Fig. 3 and 4.)

stimulus for minimal effects i n itself upon the testing monosynaptic reflex, and then combining the two conditionings. From this it may be concluded that impulses in the pyramidal tract excite interneurones of the reflex paths from:the pad to motor nuclei of short foot muscles. SUMMARY

A description is given of a spinal reflex from the plantar, evoked by innocuous pressure particularly of the central pad. The intrinsic muscles of the foot are activated in the reflex together with flexor digitorum longus and plantaris. The reflex connections from other sources to the intrinsic foot muscles have been studied, partly by intracellular recordings from the motoneurones, partly by monosynaptic testing. I n general they receive inhibition from afferents giving rise to the flexion reflex, and it is concluded that they belong to the extensor group of hindlimb muscles. This is supported by some investigations of their activity during locomotion. Cortical excitatory action upon the motor nuclei of the short foot muscles is shown to be mediated via the pyramidal tract. There is spatial facilitation between the excitatory paths to these motor nuclei from cortex and from the pad and it is concluded that impulses in the pyramidal tract excite interneurones in the reflex path from the plantar. REFERENCES I., (1963); Plantar reflexes in cat. Experientia (Busel), 19, 487-488. ENGBERG, HAGBARTH, K.-E., (1952); Excitatory and inhibitory skin areas for flexor and extensor motoneurones. Acta physiol. scand., 26, Suppl. 94.

R E F L E X E S TO TOE M U S C L E S

279

KUCELBERG, E., (1962); Polysynaptic reflexes of clinical importance. Electroenceph. din. Neuuophysiol., Suppl. 22, 103-1 1 I . KUCELBERG, E., EKLUND,K., A N D C R I M B YL., , ( I 960); An electromyographic study of the nociceptive reflexes of the lower limb. Mechanism of the plantar responses. Brain, 83, 394-410. LUNDBERG, A., AND VOORHOEVE, P., (1962); Effects from the pyramidal tract on spinal reflex arcs. Acta physiol. scad.., 56, 201-219. SHERRINCTON, C. S., (1905); On reciprocal innervation of antagonistic muscles. Proc. uoy. SOC.B, 76, 161-269. SHERRINCTON, C. S., (1 910); Flexion-reflex of the limb, crossed extension-reflex, and reflex stepping and standing. J . Physiol. (Lond.), 40, 28-121.

280

Effects of Spinal Cord Asphyxiation A. V A N H A R R E V E L D Kerckhoff Laboratories of the Biological Sciences, California Institute of Technology, Pasadena, Calq. ( U.S.A .)

Asphyxiation of the spinal cord arrests reflex activity and produces potential and impedance changes within minutes. These changes are readily reversible on reoxygenation of the tissue. Longer periods of asphyxiation cause death of nerve cells resulting in permanent functional alterations. The effects of asphyxiation on the spinal cord can therefore be divided in acute and chronic ones which will be discussed separately. A C U T E EFFECTS OF CORD A S P H Y X I A T I O N

Asphyxia1 survival of spinal reflexes Several authors noted that the arrest of spinal reflex activity by cord asphyxiation is preceded by a period in which the reflex responses are enhanced. The asphyxia1 survival time varies from about 0.5 to 4 min depending on the reflex activity used as the criterion of synaptic conduction. The most sensitive are tendon reflexes elicited by natural stimuli which after a period of enhancement disappear 35 to 45 sec after complete circulatory arrest. Flexion reflexes elicited by stimulation of the n. peroneus superficialis survive about 15 to 20 sec longer (Van Harreveld, 1944b). However, the reflex action potentials in ventral roots elicited by dorsal root stimulation survive much longer, up to 3 to 4 min (Van Harreveld, 1941). Asphyxial potentials In addition to the arrest of reflex activity a number of other changes occur in the spinal cord during the first minutes of asphyxiation. A potential develops between the gray matter and a n indifferent electrode (Van Harreveld and Hawes, 1946). Fig. I shows 3 such potentials caused by arrest of the artificial respiration (A), by ventilation of the lungs with nitrogen (B) and by arresting the circulation in the cord by clamping the aorta (C). A downward deflection indicates negativity of the gray matter with respect to the indifferent electrode. The potentials are quite similar, differing only in the latencies which had means of 50 sec, 21 sec and 8.5 sec respectively for the 3 methods of oxygen deprivation used (Van Harreveld and Hawes, 1946). These differences can be explained by the different oxygen reserves available t o the spinal

EFFECTS OF S P I N A L CORD A S P H Y X I A T I O N

28 I

Fig. 1. (A) Depolarization potential during 2 min respiratory arrest, (B) during nitrogen ventilation of the lungs for 1.5 min and ( C )during clamping of the aorta for 5 min. In the latter record a small adjustment of the galvanometer position was made at the arrow. A downward deflection indicates negativity of the gray matter. Beginning and end of the procedure causing the potential are indicated. The vertical calibration lines indicate 1 mV, the time signal minutes. (From Amer. J . Plzysiol., 147 (1946) 671 .)

cord after the start of the procedures. The potentials reach a maximum in 2 to 2.5 min and are readily reversible on reoxygenation of the cord. The similarity of the potentials elicited by ventilation with nitrogen (anoxia) and those caused by arrest of the respiration and circulation indicates that oxygen lack is the cause of the potentials and not accumulation of metabolites like carbon dioxide. Although the potentials shown in Fig. 1 are thus strictly speaking anoxic potentials they have usually been produced by asphyxial procedures and have been indicated as asphyxial potentials. The potential field in the spinal cord produced by asphyxiation was investigated by ventral direction through the cJrd, the bipolar electrode recorded negativity of the ventral tip in the dorsal horn, positivity i n the ventral horn. A reversal of the potential and Biersteker, 1964). When the electrode pair was advanced in a dorsoventral direction through the cord, the bipolar electrode recorded negativity o f the ventral tip in the dorsal horn, positivity in the ventral horn. A reversal o f the potential occurred somewhat dorsal of the level of the central canal. The monopolar recording showed that the asphyxial potential was largest at the spot where the reversal in direction of the bipolar potential occurred. The potential field can be conceived as consisting of two dipoles, one i n the dorsal and one in the ventral horn o f which the negative poles meet in the region o f the central canal. The monopolarly recorded potential is very large at this spot. Potentials up to 25 mV have been recorded in this location.

Fig. 2. Spinal conductivity changes produced by clamping the thoracic aorta for 5 and 15 min. At the arrow pointing up the aorta clamp was applied, at the arrow pointing down the blood flow was released. On the ordinate is plotted the conductivity in mhos x lofi,on the abscissa time in minutes. (From Amer. J . Physiol., 206 (1964) 8-14.) References p . 302-304

282

A. V A N H A R R E V E L D

Aspliyxicd itnpedance changes

Simultaneously with the development of the asphyxial potential the spinal cord impedance, measured with electrodes placed i n the gray matter, increases (Van Harreveld and Biersteker, 1964). Fig. 2 shows the changes in electrical conductivity (reciprocal of the impedance) during 5 and 15 min of circulatory arrest by clamping the aorta. The latency is short (about 10 sec). The conductivity drops rapidly for 5 to 6 min, then the rate decreases. During the initial rapid impedance increase 15 to 20% of the tissue conductivity is lost. The conductivity drop is like the asphyxial potential readily reversible by reoxygenation of the cord. Since blood has a lower impedance than the central nervous tissue proper, the emptying of blood vessels resulting from the circulatory arrest can be expected to cause a slight increase in tissue impedance (Van Harreveld and Ochs, 1956). However, ventilation of the lungs with nitrogen, which during the first minutes tends to cause an increase in blood pressure, results in a similar drop in conductivity as observed after circulatory arrest (Fig. 3). Since a rise in blood pressure has been shown t o cause a 130t

Fig. 3. Spinal conductivity changes caused by feeding nitrogen into the apparatus for artificial respiration. Nitrogen was administered during the period between t h e arrows. Conductivity in mhos x lo6 on the ordinate, time in minutes on the abscissa. (From Ameu. J . fhy.\ioI., 206 (1964) 8-14.)

decrease of the impedance of gray matter (Van Harreveld and Schade, 1962a), the d r o p in conductivity actually observed cannot be ascribed to vascular changes. The latency of the conductivity drop caused by ventilating the preparation with nitrogen was 20 to 30 sec. The mean latency of the anoxic potential in such experiments was 21 sec (Van Harreveld and Hawes, 1946). The similarity in the latency of the impedance drop and asphyxial potential after circulatory arrest and after ventilation with nitrogen suggests that these phenomena are related.

The nature of the changes in tissue impedance Similar asphyxial potentials and conductivity drops as observed in the spinal cord have been found in the cerebral and cerebellar cortex (see e.g. Fifkova et a/., 1961 ; Leilo, 1947; Leilo and Ferreira, 1953; Van Harreveld, 1961, 1962; Van Harreveld and Ochs, 1956; Van Harreveld and Stamm, 1953 a n d Van Harreveld and Tachibana, 1962).

EFFECTS OF SPINAL CORD ASPHYXIATION

283

The impedance of a tissue is an important parameter since it can be considered as a measure for the quantity of extracellular electrolytes. The intracellular electrolytes, which are surrounded by high resistance cell membranes, can contribute only little to the transport of the low frequency current used for the impedance measurement which therefore, is carried almost exclusively by the extracellular electrolytes (Van Harreveld and Ochs, 1956). A major change in the tissue impedance thus indicates a change in the quantity of extracellular electrolytes. Evidence for this relationship between tissue impedance and extracellular electrolytes has been demonstrated in many tissues (Van Harreveld and Biber, 1962). For instance, the submaxillary salivary gland loses a mean of 40% of its conductivity during the first 4 to 5 min of secretion (Van Harreveld, Potter et al., 1961). This would in the concept developed above be indicative of a marked loss of extracellular sodium chloride during secretion. The chloride distribution in the gland was examined with a histochemical method for chloride which consisted in rapid freezing of the tissue followed by substitution fixation in an alcoholic silver nitrate solution at -25" (Van Harreveld and Potter, 1961). As the ice in the tissue is dissolved by the alcohol the chloride is precipitated by the silver ions. The silver chloride can then be made visible by reduction to a colored subhalide. Fig. 4 shows sections of glands treated with this histochemical

Fig. 4. Control (A) and experimental (B) submaxillary gland (rabbit) treated with a histochemical method for chloride after rapid freezing. The control gland was removed before, the experimental gland 6.3 rnin after an injection of pilocarpine. During this time the tissue lost 54% of its conductivity. Calibration line indicates 100 / t . (From Amev. J . PhyJiol., 201 (1961) 1005.)

method for chloride: (A) is of a non-secreting gland and (B) of a gland which had secreted for 6.3 min after an injection of pilocarpine. Chloride can be seen in the connective tissue of both glands. Much chloride is present in the spaces between the tubules of the non-secreting gland, but most of this material has disappeared from References p . 302-304

284

A. V A N H A R R E V E L D

the intertubular spaces of the secreting gland. Apparently the gland secretes the extracellular electrolytes faster than they can be replaced from the blood. Such observations support the concept that changes in tissue impedance indicate variation? in the amount of extracellular electrolytes. Because of the relationship between tissue impedance and extracellular electrolytes the specific impedance of gray matter is a parameter of considerable interest. This value has been determined in 3 laboratoria with different methods for the ccrebral cortex of cats and rabbits (Freygang and Landau, 1955; Ranck, 1963 and Van Harreveld, Murphy et a/., 1963). Mean values of 208 to 230 f2 cm were found. These values which are 4 t o 4.5 times those of a tyrode solution a t 38" are quite low and would indicate the presence of considerable amounts of extracellular electrolytes in the cortical gray matter. Assuming that the extracellular material is isotonic with blood and cerebrospinal fluid this would be indicative of an appreciable extracellular space in central nervous tissue (Van Harreveld, 1962; Van Harreveld and SchadC, 1960). Such a conclusion is contradicted, however, by the paucity of extracellular space i n electron micrographs of central nervous tissue reported by many authors. Attempts have been made to reconcile this discrepancy. The possibility has been considered that part or all the glia would have special features which would give it properties simulating an extracellular space (Katzman, 1961 ; Van Harreveld, 1962; Van Harreveld and SchadC, 1960). Glia elements would contain most of the sodium chloride found in gray matter, and would be surrounded by cell mEmbranes which hardly impede ion movements. An alternate explanation assumes the presence of an appreciable conventional extracellular space in the living tissue, which after the arrest of the circulation and the procedures necessary to prepare the tissue for electron microscopy is taken up by the intracellular compartment (Van Harreveld, 1962; Van Harreveld and SchadC, 1960).

Asphy.uia1 chloride transport Whatever the nature of the extracellular space in gray matter may be an asphyxial transport of chloride in certain cellular elements has been found to accompany the asphyxial conductivity drop. In the cerebral cortex such a transport has been observed into the apical dendrites of the pyramidal cells (Van Harreveld and SchadC, 1959). In the cerebellar cortex an asphyxial chloride movement was found into the dendrites of Purkinje cells and into the fibers of Bergmann (glial elements) which run through the entire thickness of the molecular layer (Van Harreveld, 1961). Fig. 5 shows sections of the cerebellar cortex treated with the histochemical method for chloride. (A) is a photomicrograph of a preparation frozen while the circulation was intact and (B) 8 min after arrest of the circulation. In the oxygenated cortex the chloride distribution i n the molecular layer is rather diffuse. In the asphyxiated tissue chloride is concentrated in large dendrites which can be seen to arise from Purkinje cells and in fibers of Bergmann, which show the chloride accumulation especially clearly close to the cerebellar surface. The findings in the cerebellar cortex are of special interest since they show that the asphyxial chloride transport is not restricted

EFFECTS OF S P I N A L CORD ASPHYXIATION

Fig. 5. Photomicrographs of cerebellar cortex of rats treated with a histochemical method for chloride after rapid freezing. (A) shows a preparation of cerebellum frozen while the circulation was intact; (B) after the asphyxial impedance change had occurred. The horizontal calibration line indicates 100 / r . (From J. cellular comp. Physiol., 57 (1961) 104.)

to neural elements but that under asphyxial conditions this ion can also move into glial structures, like the fibers of Bergmann. The asphyxial impedance increase in the spinal cord suggested that a similar chloride transport as observed in the cerebral and cerebellar cortex also occurs into spinal cellular elements. Although the conditions for the application of the histochemical method for chloride are less favorable for the spinal cord of the cat than for superficial structures like the cortex, an asphyxial chloride tra~isportcould be demonstrated into the dendrites of dorsal horn somas of rats (Van Harreveld and Biersteker, 1964). Fig. 6 shows two photomicrographs of the ventral part of the dorsal horn of rat spinal cord. (A) is of a cord frozen while the circulation was intact and (B) was frozen 8 min after arrest of the circulation. Dendrites are faintly visible in the oxygenated preparation. In the asphyxiated cord they contain more black material indicative of a chloride transport. Also the dendrites in the asphyxiated preparation appear to be thicker than in the oxygenated control. A swelling of the dendrites during asphyxiation could be demonstrated by measuring the diameters of dendrites in oxygenated and asphyxiated preparations. Fig. 7 shows two histograms of the dendritic diameters in control (plain histogram) and asphyxiated (hatched histogram) preparations. A similar swelling of the elements into which chloride is transported has been observed in the cerebral (Van Harreveld, 1957) and cerebellar cortex (Van Harreveld, 1961). This swelling has been ascribed to a transport of water, which must accompany the electrolytes moving into the cellular elements to maintain osmotic equilibrium. No evidence of a chloride or water movement has been found into the somas which give rise to the dendrites showing this transport. Rcfirmces p. 302-304

286

A. V A N H A R R E V E L D

Fig. 6. Photomicrographs of the ventral portion of the dorsal horn. The spinal cords were treated with a histochemical method for chloride after rapid freezing. The preparation sfown i n (A) was frozen while the circulation wds intact; ( B ) was frozen 8 min after circulatory arrest. The calibration line indicates 10 / I . (From Amer. J . PhyJ/o/., 206 (1964) 8-14.)

100

50

0 2 .

5

10

15

20

Fig. 7. Histograms of the diameters of dendrites i n the ventral portion of the dorsal horn. The distribution of dendritic diameters in preparations frozen while the circulation was intact is represented by the histograms without cross-hatching, the distribution of diameters i n preparations frozen 8 min after circulatory arrest by the cross-hatched histogram. On the abscissa the classes are plotted (arbitrary units), on the ordinate the numbers of fibers in each class. (From A n w . J . Physrol., 206 (1964) 8-14.)

The changes i n the spinal cord and in the cerebral and cerebellar cortex after circulatory arrest, consisting of the development of an asphyxia1 potential, an impedance increase and a transport of electrolytes and water into certain cellular elements, are quite similar. All these changes can be explained by an increase in permeability of the cell membrane for inorganic ions, especially for sodium ions (Van Harreveld, 1962; Van Harreveld and Ochs, 1956). A markedly increased sodium permeability will result in depolarization of the cellular elements involved. If other

EFFECTS O F SPINAL C O R D A S P H Y X I A T I O N

287

sections of the cell remain polarized, sources and sinks are created which may produce the asphyxial potentials. Assuming furthermore, that the membrane remains impermeable for the large organic intracellular anions then the increased ion permeability elicits conditions in which Donnan forces will transport extracellular electrolytes into the cellular compartment, accounting for the conductivity drop of the tissue, and the transport of chloride accompanied by water into the cellular elements in which the membrane changes occurred.

The origin oj the asphyxial potential The histochemical findings in oxygenated and asphyxiated spinal cords support the concept that the asphyxial potential is caused by the depolarization of certain parts of nerve cells while in other parts of the cell the membrane potential is maintained. A chloride transport was observed into dendrites of dorsal horn neurons, whereas such a transport could not be demonstrated into the somas. The latter (and probably the axons) would then represent the sources of the potential developing during cord asphyxiation, whereas the dendrites would act as the sinks. The assumption that the asphyxial potential of the spinal cord is a neural phenomenon is supported by the observation that the destruction of the spinal neurons by asphyxiation prevents the development of the potential (Van Harreveld and Hawes, 1946). The following experiments demonstrate that the soma potentials of spinal neurons are quite resistant to oxygen lack (Van Harreveld and Biersteker, 1963). Microelectrodes were passed through the spinal cord of rats in a dorso-ventral direction. Each potential jump on entering a cellular element was recorded. Although part of these jumps may have been caused by entering glia cells, it is likely that the majority represent soma potentials since penetration results in general in single or repetitive discharges. The largest potentials (about 70 mV) were recorded in the ventral part of the cord. These may represent the membrane potential of motor cells. After circulatory arrest the maximum potentials dropped 10 to 15 mV during the first few minutes, followed by a slow decline over the ensuing 60 to 90 min. Kolmodin and Skoglund (1959), who also examined the soma potentials in cats with microelectrodes, observed a similar decline during the first minutes of asphyxiation, but Nelson and Frank (1959) found a smaller initial depolarization if any. It is tempting to ascribe the initial drop in the soma potential to dendritic depolarization, electrotonically conducted to the cell body. The membrane potentials could be restored promptly after an hour of asphyxiation by re-establishing the circulation. The great resistivity to oxygen lack of the soma potential seems to be a special property of spinal neurons. When the same experiment was performed on the rat's cortex a drop of the soma potential was noted starting 2 to 3 min after arrest of the circulation, when the cortical asphyxial potential began to develop. The cortical somas depolarized much faster than the spinal cells. A low value of the membrane potential was reached after 10 to 15 min of oxygen deprivation.'Also the depolarization of the cortical somas was reversible even after relatively long periods of asphyxiation (45 min). References p. 302-304

288

A.

VAN HARREVELD

These findings suggested an experiment to test the mechanism of the asphyxial potential proposed above. If the source of this potential is the soma membrane potential, then the course of the asphyxial potential after complete depolarization of the dendrites should be determined by the polarization state of the somas. The spinal asphyxial potential can, therefore, be expected to decline considerably slower after having reached its maximum than the cortical potential. This was investigated by leading the asphyxial potentials off with bipolar electrodes to minimize electrode potentials due to differences i n temperature and chloride concentration at the electrode tips (Van Harreveld and Biersteker, 1964). Fig. 8 (A) shows that the spinal asphyxial

/ - - -

0

0

4

5

7

24

12

40

16

E0

20 25

Fig. 8. (A) spinal asphyxial potential during prolonged asphyxiation. The aorta was clamped at time zero. The development of the asphyxial potential is shown during the first 4 min. The figures under the black bars indicate minutes of asphyxiation. The aorta is released after 60 min. (B) cortical asphyxial potential during prolonged asphyxiation. The vessels arising from the aortic arch were clamped at time zero. Development of the asphyxial potential during the first 8 rnin shown. The figures under the bars indicate minutes of asphyxiation. The vessels are released after 25 rnin of asphyxiation. The vertical calibration lines indicate 5mV. (From Arner. J . Physiol., 206 (1964) 8-14.)

potential had not declined completely after 60 min of oxygen deprivation, whereas the cortical potential (B) reached a low value about 15 min after the potential had reached its maximum. These experiments support the concept that the asphyxial potential is caused by the differential depolarization of dendrites.

The mechanism of the asphyxia1 arrest of reflex activity The differences in asphyxial survival of reflex activity mentioned above may be considered in the light of the asphyxial changes described above. The cells with the highest membrane potentials which were encountered in the ventral horn and which probably were motor nerve cells, depolarized 10 to 15 mV during the first minutes of asphyxiation. If, as suggested, this is due to dendritic depolarization conducted electrotonically to the soma then it can be expected that dendritic synapses become unoperational shortly after circulatory arrest. When reflexes are elicited by natural stimuli which activate a limited percentage of the synapses ending on the motoneurone, then

EFFECTS OF SPINAL CORD ASPHYXIATION

289

dendritic depolarization may arrest the reflex by restricting the effective synapses to those connecting with the somas. This may explain the early arrest (after 35 to 60 sec of asphyxiation) of tendon reflexes and of flexion reflexes elicited by stimulation of the superficial peroneal nerve. Stimulation of the dorsal root may be able to overcome this adverse effect by the activation of a larger percentage of the somatic synapses, explaining the materially longer (3 to 4 min) survival time of this form of reflex activity. The arrest of the reflex action potentials elicited by dorsal root stimulation may be caused by other mechanisms either pre- or postsynaptically as discussed by Brooks and Eccles (1947) and by Lloyd (1953). C H R O N I C EFFECTS O F C O R D A S P H Y X I A T I O N

Reflex activity after recovery f r o m cord asphyxiation After recovery from the acute effects of asphyxiation of up to 20 to 25 rnin duration the spinal reflex activity of cats is usually not strikingly changed. Occasionally the reflexes are somewhat more vivid than in normal animals. Asphyxiation between 30 and 35 min changes the reflex activity of the preparation profoundly, however. One of the remarkable features of such preparations is the hypertone, rigidity or spasticity which has been observed by many authors (Biersteker and Van Harreveld, 1963; Haggqvist, 1938; Hochberg and HydCn, 1949; Gelfan and Tarlov, 1959; Kabat and Knapp, 1944; Kosman et al., 1951; Krogh, 1950; Rexed, 1940; Tureen, 1936; Van Harreveld and Marmont, 1939). It has not been generally realized, however, that distinct periods of tone develop which, as will be discussed, are apparently caused by different mechanisms. The

Fig. 9. Course of tone of the quadriceps muscle after 65 rnin asphyxiation. The quadriceps tendon was tapped every 2 rnin. The figures indicate the interval after the end of asphyxiation in minutes. The initial tone developed after 8 min, reached a maximum after 12-14 min and had disappeared after 20 min. At 60 min the cord was subjected to a renewed asphyxiation of 20 sec duration which caused a strong extensor contraction. After 85 min secondary tone developed which reached a maximum and disappeared 150 rnin after the end of asphyxiation. (From Amer. J. Physiol., 139 (1963) 619.)

following description is based on observations of cats in which the cord was asphyxiated by increasing the dural pressure above the blood pressure for 30 to 65 min (Van Harreveld, 1943, 1944a; Van Harreveld and Marmont, 1939). Very soon (within 6 to 20 min) after the end of the circulatory arrest a slight tone tends to develop, which is References p . 302-304

290

A. V A N H A R R E V E L D

transient and disappears again 5 to 15 min later (Van Harreveld, 1943, 1944a). This tone which has been called the ‘initial’ tone is so weak that it can be demonstrated only by recording the muscle tension (Fig. 9). This is followed by a period of one to several’ hours during which the limb is flaccid. The initial tone, which, as will be discussed below, must be considered as a reflex tone, is of considerable interest in view of the speedy repolarization of motoneurons on reoxygenation mentioned above. It would appear that as soon as the neurons are repolarized synaptic conduction becomes possible, resulting in tone. Of interest also is the transient nature of the initial tone. The disappearance of postasphyxial reflex activity as in the case of the initial tone was ascribed by Lloyd (1953) to postanoxic hyperpolarization of neural elements. During the flaccid period after the initial tone muscle contractions can be elicited by a short period of renewed cord asphyxiation (Fig. 9). The resulting dendritic depolarization could make conduction again possible by counteracting the postulated hyperpolarization. Indeed elements of myotatic reflex activity could be demonstrated in the contractions elicited in this way (Van Harreveld, 1943). After the flaccid period which lasts from one to several hours a much more pronounced ‘secondary’ tone develops (Fig. 9). This causes an extension of the hind legs and is often so strong that the knee and ankle can only be bent by exerting considerable force. During this time tendon reflexes are present which, when tone is slight, are distinctly hyperactive. At the height of tone the tendon reflexes are difficult to examine, but still appear brisk. Clonus can usually be elicited by bringing the leg in certain positions. A flexion reflex is sometimes observed at that time. The secondary tone is most obvious in the extensor muscles but can also be demonstrated in muscles with flexor function like the anterior tibia1 muscle (Van Harreveld, 1944a). The secondary tone tends to remain high during the first 24 to 48 h but then diminishes. It may even disappear completely, resulting in a second flaccid period during which no reflexes can be elicited. However, in preparations asphyxiated for relatively short periods (30 to 35 min) tone increases again, or may in the flaccid preparations redevelop about 1 week to 10 days after asphyxiation. This third period of rigidity, the ‘late’ tone, reaches a maximum several days to 1 week later and then remains unabated until the end of the animal’s life (Biersteker and Van Harreveld, 1963; Gelfan and Tarlov, 1959; Van Harreveld and Marmont, 1939). Brisk but small tendon reflexes and clonus are in general present during the late tone, and also small flexion reflexes can often be elicited. Gelfan and Tarlov (1959) showed that i n preparations with late tone a myostatic contracture of the musculature tends to develop, which hampers the evaluation of active muscle contractions. After long asphyxiations (up to about 65 min) the initial and secondary tone develop as described above. However, the longer the spinal cord has been asphyxiated the shorter the secondary tone tends to be. After 60 to 65 min asphyxiation the secondary tone may be present for not more than a few hours (Fig. 9). Then the preparation becomes flaccid, and since no late tone develops in the legs of such preparations they remain without tone for the duration of the animal’s life. After asphyxiations of 45 to 55 min there may develop a late tone in the tail, and slight movements on pinching the tail are usually observed.

EFFECTS OF S P I N A L CORD ASPHYXIATION

29 1

Nerve cell destruction by asphyxiation

The anatomical changes produced in the spinal cord by asphyxiations of 30 min and longer are profound. Preparations of spinal cords asphyxiated by increasing the pressure in the dural cavity above the blood pressure were compared with those of normal cords. In 4 normal cats the volumes were determined (SchadC and Van Harreveld, 1961) of a 25% sample of the nerve cells in the peroneus-tibialis neuron pool in the 7th lumbar segment (peroneus nucleus, dorsolateral and central tibialis nuclei, and ventro-lateral nucleus, which correspond to cell columns 4, 5 and 6 of Romanes, 195I). The plain histogram shown in Fig. 10 which represents the mean cell volume

-

v ( 1000,1A ~

Fig. 10. Histogram (non-hatched) of the volume distribution of a 25 % sample of the nerve cells in the peroneus-tibialis neuron pool of L7. The figure represents the mean of 4 spinal cords. The hatched areas represent cells showing retrograde degeneration after severing the sciatic nerve. On the abscissa are plotted the classes (units indicate volume of 1000 p 3 ) ;on the ordinate the number of cells in each class. (From J. comp. Neurol., 117 (1961) 396.)

distribution in these preparations has two maxima, one at 2000 to 6000 p3 the other at 28,000 to 32,000 p3.The mean number of neurons in the 25 %sample of this nucleus was 73 1, of which 79 % had a volume smaller than 16,000 p3,2 1 % a larger volume. In preparations in which the sciatic nerve had been cut an appropriate time before the fixation of the spinal cord, motor cells could be recognized by the resulting retrograde degeneration. The (hatched) histogram of these cells shows that the group of cells larger than 16,000 p3 consisted mostly of motoneurons, the group of smaller cells mostly of interneurons, although about 10 % of this group showed retrograde degeneration. The latter cells may represent the motor neurons of y-efferents. Asphyxiations up to 20 min in duration did not significantly change the total number of nerve cells in the peroneus-tibialis nucleus. In preparations asphyxiated for 30 to 35 min and showing the late tone, extensive destruction of nerve cells had taken place; 93 % of the cells normally present in the peroneus-tibialis nucleus had disappeared (Van Harreveld and Schadt, 1962b). The (hatched) histogram of Fig. 11 shows the mean volume distribution in 5 spinal cords of the remaining cells in the peroneus-tibialis nucleus. The distribution had changed considerably as compared with the controls: only 38 % of the cells had a volume smaller than 16,000 p3 against References p . 302-304

292

A. V A N H A R R E V E L D

’

LiA

20 0 0

4

8

12

16 16 24

32

40 40 56

64

---+ ~(~100op3) Fig. 11. The hatched histogram shows the mean cell volume distribution of a 25% sample of the peroneus-tibialisneuronpool in L7 of 5 spinal cords asphyxiated for 28 to 35 min. The non-hatched histogram is the volume distribution in the normal cord. Units as in Fig. 10. (From J . Neuropathnl. exp. Neurol., 21, ( I 962) 4 18.)

79 in the controls, whereas 62 had a larger volume against 21 % in the controls. The destruction of interneurons thus was very severe and although the destruction of the motor cells was considerable, a larger percentage of them had survived. In the rest of the cord the destruction of neuronal elements was also severe. In the area around the spinal canal there was almost complete destruction of neurons. Considerable cell death was observed in the dorsal horn, and extensive neuronal destruction in the ventral horn especially i n its central medial region was found. Gelfan and Tarlov (1959, 1962) and Tarlov and Gelfan (1960) who asphyxiated the cord by clamping the thoracic aorta, also described an extensive destruction of interneurons in rigid dogs, but observed at least in part of their preparations with (late) tone hardly any destruction of motor cells. The above observations seemed to indicate a relationship between the size of a nerve cell and its sensitivity to oxygen deprivation. However, histograms of the diameters of the fibers in the ventral roots of asphyxiated preparations (Fig. 12) showed almost the same distribution in a- and y-groups as found in normal control cats (Biersteker and Van Harreveld, 1963). Assuming that thin axons take origin from small motcneurons, and thick motor fibers from large nerve cells, the relationship between cell size and sensitivity does not seem to hold. The sensitivity to asphyxiation may, therefore, depend more on the function of the cell than on its size, leading to the suggestions that interneurons are more sensitive to oxygen lack than motoiieurons and that a considerable percentage of the surviving cells in the group smaller than 16,000 p3 in the peroneus-tibialis nucleus of asphyxiated preparations may be cells of origin of y-efferents. There is an interesting contrast between the prompt reversibility of the spinal asphyxiat potential (Van Harreveld and Biersteker, 1964) and of the asphyxial soma depolarization (Van Harreveld and Biersteker, 1963) after long periods of asphyxiation and the very severe ultimate destruction of spinal neurons. This seems to indicate that the ion pumps which have to restore the concentration gradients resulting in the membrane potential are not very sensitive to oxygen lack and can

EFFECTS O F S P I N A L C O R D A S P H Y X I A T I O N

293

resume their function soon after oxidative energy becomes available again (Van Harreveld and Tachibana, 1962). What seems to be sensitive to oxygen lack are the structures which form the enzyme systems necessary for the maintenance of the life of the neuron. This difference in sensitivity to oxygen deprivation explains the often 700-

500-

-

600

400500-

400

300-

-

300200100-

loot

rr-L A

1;11

100

Fig. 12. Histograms of the fiber diameters in the ventral root of L7 of a normal preparation, (A) and of 3 preparations, (B), ( C ) and (D), asphyxiated for 30 to 35 min and showing destruction of motoneurons which varies in seriousness. On the abscissa the classes are plotted in p, on the ordinate the number of fibers in each class. (From J . Physiol. (London), 166 (1963) 385.)

pronounced but transient secondary tone observed in preparations asphyxiated for 50 to 65 min from which it is known that the ultimate destruction of the spinal neurons is very severe (Van Harreveld and Marmont, 1939). It has to be assumed that neurons damaged so severely that they will be destroyed in the end are able to function temporarily as shown by the secondary tone. Electrophysiological features of preparations with secondary and late tone The electrophysiological properties of preparations with secondary and late tone were compared with those of acute and chronic spinal control cats (Van Harreveld and Spinelli, 1964). Reflex action potentials elicited by stimulation of the dorsal root of the first sacral segment (DSl) and of the gastrocnemic nerve with conditioning and test shocks supramaximal for the monosynaptic response were recorded from the first sacral ventral root (VS1). Tnhibition of monosynaptic responses elicited by stimulation of the gastrocnemic nerve by shocks to the sural nerve, and retrograde inhibition of responses to DS1 stimulation by shocks to VSl were examined. FurtherReferences p.-302-304

294

A. V A N H A R R E V E L D

more, the responses in VS1 elicited by repetitive stimulation of DSl and of the gastrocnemius nerve were recorded. Fig. 13 (A) shows the recovery cycles of 6 acute spinal preparations obtained by plotting the heights of the monosynaptic test responses to stimulation of DSI (ex-

Fig. 13. (A) recovery cycles of the monosynaptic spike in the ventral root of SI , elicited by stimulation of the dorsal root at the same segmental level with 2 supramaximal stimuli in 6 acute spinal control preparations. On the ordinate the heights of the monosynaptic responses to the test stimuli are plotted as a percentage of the height of the conditioning spike; on the abscissa the stimulus interval in msec. (B) in the same preparations the heights of the first 20 monosynaptic responses at a stimulus frequency of 50/sec are plotted as a percentage of the response to the first stimulus in the series. The potentials were led off from the ventral root of SI. The dorsal root was stimulated with shocks which were supramaximal for the monosynaptic response. (From Topics in Basic Neurology, Vol. 6, Progress in Brain Research).

pressed as a percentage ofthe spike caused by the preceding supramaximal conditioning shock), against the stimulus interval. These recovery cycles are characterized by varying degrees of direct facilitation caused by summation of excitatory postsynaptic potentials, and by a pronounced postactivation depression from which the preparation has often not recovered 400 msec after the conditioning shock. The responses elicited by stimulation of the gastrocnemic nerve showed a similar cycle. Comparable results were obtained in chronic spinal control preparations. The marked and long lasting depression observed in the recovery cycle of the control preparations by stimulation of DS1 is a complex phenomenon. It includes the effects of the positive after-potentials of motoneurons (Brooks et a]., 1950), of indirect inhibition, both autogenic and by stimulation of cutaneous pathways, and of antidromic inhibition by the activation

295

EFFECTS OF S P I N A L CORD ASPHYXlATION

of Renshaw cells. The long duration of the postactivation depression suggests that the later part of the cycle is a presynaptic effect which may be identified with the presynaptic inhibition described by Eccles et 01. (1962a, b). In all these mechanisms resulting in the depression of reflex activity interneurons are involved except in the effect of positive after-potentials. Sural stimulation caused marked inhibition of the monosynaptic spike produced by gastrocnemic stimulation, and the depressing effect of antidromic stimulation of VS1 was pronounced. The reflex arcs in spinal control preparations carry impulse trains elicited by stimulation of the dorsal root at 50/sec poorly. I n Fig. 1 (B) the responses to the first 20 shocks of such a repetitive stimulus are plotted as a percentage of the monosynaptic spike on the first shock. The incomplete transmission of the impulse trains can be ascribed to the mechanisms which cau3e the postactivation depression. Preparations with secondary tone yielded different recovery cycles. Fig. 14 (A)

400

\ II

I

I

I

50

I

I

I

I

I

lo0

I

,

2m

I

400

Fig. 14. Recovery cycles (A) and responses to stimulation at 50/sec (B) in preparations with secondary tone, 24 h after 30 to 35 min asphyxiation of the cord. See Fig. 13 for explanation of the graphs. (From Topics in Basic Neurology, Vol. 6,Progress in Brain Research).

shows the cycles of 5 preparations produced by stimulation of DS1 with two supramaximal stimuli. There was evidence of direct facilitation at intervals of 2 to 3 msec. However, the facilitation in most of the preparations was of longer duration than would be consistent with the direct facilitation due to summation of postsynaptic potentials. In one preparation the response remained facilitated for 150 msec. The long durationoffacilitation indicates that it is due to interneuronal, indirect facilitation. The postactivation depression in 4 of the preparations was slight or even absent. The References p . 302-304

296

A. V A N H A R R E V E L D

recovery cycle produced by stimulation of the gastrocnemic nerve showed less indirect facilitation, but was characterized by a very short (20 nisec) postactivation depression. The preparations showed antidromic inhibition although this was less pronounced than in the controls. Sural stimulation did not inhibit the monosynaptic response elicited by gastrocnemic stimulation. The attenuation of antidromic inhibition and the absence of indirect inhibition suggest a mechanism for the pronounced indirect facilitation. In normal control preparations indirect facilitation may not be able to manifest itself because of the inhibitory mechanisms responsible for the postactivation depression which include antidromic and indirect inhibition. By attenuation of inhibition as demonstrated in preparations with secondary tone indirect facilitation may then be unmasked as one of the prominent features of these preparations. In agreement with the lack of pronounced postactivation depression these preparations were able to transmit trains of impulses with much less attenuation than the controls (Fig. 14, B). In one preparation the responses even became larger during repetitive stimulation which may have been the result of indirect facilitation and potentiation. The reflex action potentials elicited by DSI stimulation in preparations with late tone were several times larger than the potentials in the control preparations, notwithstanding the severe destruction of motor cells reported above. An unusual niagnitude of reflex action potentials was also observed in similar preparations by Gelfan and Tarlov (1959). In Fig. 15 (A) are shown the recovery cycles produced by stimulating

OF -i? ---.h -

-

-

L

10

20

Fig. 15. Recovery cycles (A) and responses to stimulation at 50/sec (B) in preparations with late tone, 2 weeks after 30 to 35 min asphyxiation of the cord. See Fig. 13 for explanation of the graphs. (From Topics in Basic Neurology, Vol. 6 , Progress in Brain Research).

DSI with two shocks in 6 preparations examined 2 weeks after 30 to 35 min asphyxiation. Direct facilitation is not marked, although the maxima at an interval of 2 to 3 msec suggest that this is not entirely lacking. The postactivation depression is not pronounced in 3 of the preparations. These recovery cycles were of cats with the most pronounced late tone. Some of the recovery cycles produced by gastrocnemic nerve

EFFECTS O F S P I N A L C O R D A S P H Y X I A T I O N

297

stimulation also did not show a marked postactivation depression. Antidromic inhibition was present. Also indirect inhibition was observed in all but one preparation. Three preparations with pronounced late tone transmitted trains of impulses rather well (Fig. 15, B). These were the same preparations with pronounced late tone which showed little postactivation depression. Gelfan and Tarlov (1959) suggested that the excitability of motoneurons in preparations with (late) tone is markedly increased due to the death of a large percentage of the interneurons (denervation hypersensibility). This is supported by the unusual magnitude of the reflex action potentials on DSl stimulation in such preparations, indicating that a large part of the reduced pool of motoneurons is activated by single volleys. Also the near absence of direct facilitation, suggesting that the subliminal fringe of the motoneuron pool is small, favors the assumption of an abnormally high excitability of the motoneurons. This increased excitability of the motoneurons can be expected to counteract the mechanisms causing the postactivation depression, which as mentioned above is not marked in preparations with a pronounced late tone. The preparations with late tone gave evidence of interneuronal activities nothwithstanding the extensive destruction of interneurons observed. Many of the preparations showed a small (multisynaptic) flexion reflex. Also the reflex action potentials elicited by DSI stimulation showed some late activity. Some postactivation depression was always present in the recovery cycles. Most of the preparations exhibited indirect and antidromic inhibition. Denervation hypersensitivity of the motoneurons to the excitatory transmitter compound might make up for the loss of interneurons, accounting for multisynaptic excitatory responses in these preparations. The assumption that the motoneurons become hypersensitive not only for the excitatory but also for the inhibitory transmitter compound might in the same way explain inhibitory activities notwithstanding the major loss of interneurons. The features of the reflex activity in preparations with secondary and late tone described above suggest a mechanism for the rigidity observed in these animals. In order to produce reflex tone the monosynaptic arc must be able to transmit impulse trains from the muscle stretch receptors at a frequency which can maintain sustained contractions. In the spinal control prepartions impulse trains at 50/sec, elicited by stimulation of the gastrocnemius muscle nerve or DS1, were transmitted poorly. Preparations with secondary tone transmitted impulse trains at this frequency much more efficiently. It has been suggested that this is due to the inactivation of inhibitory mechanisms responsible for the postactivation depression which results in the unmasking of indirect facilitation. The rigidity observed may be based on this improved transmission of impulse trains. The monosynaptic arcs in preparations showing pronounced late tone also tended to transmit impulse trains more efficiently than in spinal controls. As mentioned above the motoneurons in these preparations have become hyperexcitable which will reduce postactivation depression and promote the transmission of impulse trains. The enhanced motoneuron excitability will furthermore result in the activation of a major portion of the motoneuron pool on single stimuli. Both these effects of the enhanced motoneuron excitability may account for the late tone. References p . 302-304

298

A. V A N H A R R E V E L D

Little i s known of the activity of the y-efferent system in rigid preparations. Gelfan and Tarlov (1959) demonstrated action potentials i n y-efferents of preparations with (late) tone. There is no doubt, however, that in the preparations with late tone described above a considerable percentage of the y-efferents are destroyed (Biersteker and Van Harreveld, 1963; Van Harreveld and SchadC, 1962b). The electrophysiological differences between preparations with secondary and late tone described above support the distinction of these two periods of tone as separate entities. This is of importance, since occasionally the secondary tone in the cat continues into the late tone without a period of flaccidity or even a marked decrease in rigidity. The r e j e x nature of the asphyxial rigidity

In the preceding discussion exaggerated myotatic activity was assumed to be the basis of the asphyxial rigidity. This concept is supported by the following observations. Monosynaptic reflexes are electrophysiologically functional in preparations with secondary and late tone as shown above. The unusual magnitude of the monosynaptic potentials in cats with late tone even indicates that in such preparations myotatic reflexes are hyperactive. The myotatic nature of the rigidity is furthermore supported by myograms led off during stretch and relaxation of muscles in preparations with asphyxial tone. Figs. 16 and 17 show that during initial, secondary and late tone

B I

0 Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

C Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

' 0 Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

D Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

FE Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Fig. 16. Effects of bending and stretching the knee on the electromyogram of the quadriceps muscle. The start of flexion and extension is indicated by dots. (A) 70 min after 50 min cord asphyxiation when secondary tone starts to develop flexion of the knee increases the electrical activity. (B) 20 min later the effect of flexion is enhanced. ( C ) stretch of the knee arrests the electrical activity. (D) 8 min after a 50 min cord asphyxiation when initial tone starts to develop, flexion of the knee causes activity in the electromyograni. (E) 10 min later activity is present which can be arrested by extension of the knee. Time signal indicates seconds. (From Anlev. J . Physiol., 139 (1963) 622.)

s fi$

A

-

t

h

1

t rn

C

1 0 -n

Fig. 17. Electromyograms of the quadriceps muscle in preparations with late tone, 3 months after 30 to 35 min asphyxiation. Preparation (A) had no pronounced tone, (B) and ( C ) showed the late tone. At the arrows pointing up the knee is flexed; at the arrows pointing down it is extended again. The horizontal calibration line indicates 1 sec, the vertical 100 pV. (From J. Physiol. (London), 166 (1963) 386.

*L.

.

C0Y

2 2

*r

-I+

Fig. 18. Electromyograms of the quadriceps muscles on the operated (upper traces) and normally innervated (lower traces) sides of a unilaterally deafferented preparation. In record (A) the knee on the innervated side is flexed at the arrow pointing upward and extended again at the arrow pointing down. In record (B) the bending and stretching is repeated on the deafferented side. The horizontal calibration line indicates 1 sec, the vertical lines 100 p V for the records of normal and operated sides. Note the difference in amplification used for the two records. (From J . Physiol. (London), 166, (1963) 387.)

h, \D

W

300

A. V A N H A R R E V E L D

stretch of the muscle enhances the activity in the myogram, whereas relaxation decreascs it (Biersteker and Van Harreveld, 1963 ; Krogh, 1950; Van Harreve!d, 1943, 1944a). Another concept of the asphyxia1 rigidity has been developed lately by Gelfan and Tarlov (1959), according to which the rigidity would be caused by the spontaneous discharge of motoneurons made hypersensitive by the destruction of interneurons. This concept was mainly based on the observation that deafferentiation produced only a temporary relaxation of the asphyxial rigidity. However, deafferentiation does more than eliminate the sensory influx. It apparently causes an additional increase in the excitability of the motor cells i n rigid preparations as shown by the following observation (Biersteker and Van Harreveld, 1963). Stretch of the quadriceps muscle in a n animal with late tone causes sometimes a slight increase in the electrical activity of the heterolateral muscle. Several days after unilateral deafferentiation this reaction may become quite marked and stretch of the quadriceps muscle with intact sensory innervation may cause a pronounced tonic contraction of the heterolateral muscle and a marked increase of its electrical activity (Fig. 18). Such observations would seem to indicate that deafferentiation of rigid preparations enhances the excitability of the motoneurons. The destruction of their monosynaptic innervation may enhance the denervation hypersensitivity already present i n these preparations. In experiments i n which preparations with late tone were bilaterally deafferentiated the excitability became apparently so high that even a small intact rootlet was able to maintain rigidity (Biersteker and Van Harreveld, 1963). However, in cats in which the deafferentiation had been complete, and in which by appropriate sections of the spinal cord the flow of impulses from other segments had been prevented, no evidence of active muscle rigidity was found. I n such preparations a light pressure directly on the cord, made possible by the laminectomy necessary for the deafferentiation. resulted in muscular activity, proving that functional motoneurons were present. The concept first proposed by Gelfan and Tarlov (1959), that motoneurons in rigid preparations become hypersensitive due to the loss of a substantial part of their neural connections, has been very fruitful, and explains many of the features of preparations with late tone as discussed above. This hypersensitivity could conceivably lead to the spontaneous discharge of motoneurons, although no evidence was found for such a mechanism of rigidity in preparations with secondary and late tone. (Biersteker and Van Harreveld, 1963). The prolonged asphyxial survival t i m e in preparations with late tone

A remarkable feature of the preparations with late tone is their enhanced resistance to renewed asphyxiation (Van Harreveld, 1941). The attention was first directed to this phenomenon by the slow relaxation of the rigidity after sacrificing such preparations. This is not a reliable criterion for the survival of reflex tone, however, since the active muscle tone is complicated by the myostatic contractures which tend to develop in such preparations. The asphyxial survival of the reflex action potentials in a ventral root elicited by dorsal root stimulation is a more useful indicator of the

EFFECTS OF SPINAL CORD ASPHYXIATION

30 1

sensitivity of the reflex arc to oxygen deprivation. The survival time of these potentials in acute and chronic spinal control preparations after circulatory arrest by severing the aorta or by ventilation of the lung with nitrogen was between 2 min 25 sec and 4 min 35 sec. In preparations asphyxiated for 35 rnin and showing two weeks later the late tone, the survival time varied between 7 rnin 35 sec and 13 rnin 40 sec after cutting the aorta, and between 13 rnin 30 sec and 17 rnin 45 sec after ventilating the preparation with nitrogen. In 4 animals asphyxiated for 35 rnin and examined two days later the survival time varied between 1 min 10 sec and 4 min 25 sec. In 3 preparations maintained for 3 to 4 days the survival time was 3 rnin 20 sec to 4 min, and in 3 cats kept for 6 days it was between 6 rnin 40 sec and 8 rnin 5 sec. The greater resistivity to oxygen lack is thus not present during the first days after asphyxiation but starts to develop after about 1 week. The possibility has been considered that the death of a large percentage of the spinal neurons would decrease the metabolism of the cord to such an extent that the oxygen stored in the blood and in the nervous tissue would suffice for the needs of the remaining structures for a materially longer period than in the controls. However, when 2 weeks after asphyxiation the preparations show the long asphyxia1 survival time, the glia has proliferated markedly. The metabolism of these newly formed glia cells will diminish the expected decrease of the oxygen consumption of the cord due to the neuronal death. The oxygen uptake of the spinal cord 2 weeks after 35 min asphyxiation was found to be about half that of control cords (Van Harreveld and Tyler, 1942) which is insufficient to explain the marked increase in survival time. Furthermore the long survival time after ventilating the preparations with nitrogen which washes the oxygen out of the blood (and tissues) disagrees with this thesis. The enhanced survival time in preparations with late tone is for the time being unexplained. It shows, however, that the anoxic survival of synaptic conduction can under unusual conditions be of the same order as that of mammalian peripheral nerve. The short survival of spinal reflexes in normal control cats is, therefore, probably not an intrinsic property of the reflex arc, but may for instance be caused by the release of compounds which arrest reflex activity. The changes produced in the cord by asphyxiation might interfere with such a mechanism. SUMMARY

Acute asphyxiation of the spinal cord causes negativity of the gray matter with respect to an indifferent electrode. Also an increase of the impedance of the gray matter was observed. The latency of these changes is quite short (about 10 sec) and they are quickly reversible on reoxygenation of the cord. The impedance change is believed to be due to a transport of electrolytes from an extracellular space into cellular elements of the cord. This explanation is supported by the observation of an asphyxial chloride transport into dendrites of dorsal horn cells. The transport would be caused by an increase in ion permeability of the cellular membrane, creating Donnan forces responsible for the electrolyte movements. The increased ion permeability of the dendritic membrane, which will result in the depolarization of the cellular elements Rcfermces p . 302-304

302

A.

VAN HARREVELD

involved, may also account for the asphyxia1 potential, since the soma membrane potential of spinal neurons was found to be quite resistant to 0 2 deprivation. One of the more enduring effects of spinal cord asphyxiation for 30 to 35 rnin is the development of extensor tone in the hind legs. Three periods of tone were distinguished. A slight and fleeting ‘initial’ tone developed 6 to 20 min after the end of asphyxiation and disappeared again 5-15 min later. This was followed, after a period of flaccidity of one to several hours duration, by a marked ‘secondary’ tone. This tone declined or even disappeared after 1 to 3 days. However, a few days later tone developed (or increased) again. This third period of rigidity, the ‘late’ tone, which may be quite pronounced remained to the end of the animal’s life. Evidence was found in all 3 periods of tone for the involvement of myotatic reflex activity. A marked cell destruction especially of interneurons was found in preparations with late tone. The preparations with secondary tone showed less postactivation depression in the recovery cycle than spinal control preparations. Also these preparations conducted trains of reflex impulses with less attenuation than the controls. In preparations with late tone signs of hyperexcitability of the motoneurons were found. These preparations also showed less postactivation depression and conducted trains of reflex impulses rather well. It was suggested that the ability of the asphyxiated spinal cords to conduct impulse trains is a mechanism underlying the hypertone in these preparations. Preparations with late tone are relatively insensitive to renewed asphyxiation. Reflex action potentials have been observed in such preparations up to 10 to 15 min after asphyxiation of the cord. ACKNOWLEDGEMENTS

The investigations discussed in this paper were supported in part by research grants from the United States Public Health Service, from the National Science Foundation and from the Office of Naval Research. REFERENCES

P. A., A N D VAN HARREVELD, A,, (1963); The nature of the rigidity caused by spinal cord BIERSTEKER, asphyxiation. J . Physiol. (London), 166, 382-394. BROOKS, C. McC., DOWNMAN, C. B. B., AND ECCLES,J. C., (1950); After-potentials and excitability of spinal motoneurones following orthodromic activation. J . Neurophysiol., 13, 157-176. BROOKS,C. McC., AND ECCLES,J. C., (1947); A study of the effects of anaesthesia and asphyxia on the monosynaptic pathway through the spinal cord. J . Neurophysiol., 10, 349-360. ECCLES,J. C., KOSTYUK,P. G., AND SCHMIDT,R. F., (1962a); Presynaptic inhibition of the central actions of flexor reflex afferents. J . Physiol. (London), 161, 258-281. ECCLES,J. C., SCHMIDT,R. F., AND WILLIS,W. D., (1962b); Presynaptic inhibition of the spinal monosynaptic reflex pathway. J . Physiol. (London), 161, 282-297. FIFKOVA, E., BURES,J., KOSHTOYANTS, 0. KH., KRIVANEK, J., A N D WEISS, T., (1961); LeBo’s spreading depression in the cerebellum of rat. Experientia (Busel), 17, 572-573. FREYGANG JR., W. H., AND LANDAU, W. M., (1955); Some relations between resistivity and electrical activity in the cerebral cortex of the cat. J . cellular comp. Phys/ol., 45, 377-392. GELFAN,S., AND TARLOV,1. M., (1959); lnterneurones and rigidity of spinal origin. J . Physiol. (London), 146, 594-617. GELFAN, S., AND TARLOV, 1. M., (1962); Neuronal population and caliber spectra in entire L7 segment of normal and experimental rigid dogs. Fed. Proc., 21, 368.

EFFECTS O F S P I N A L C O R D ASPHYXIATION

303

HAGGQVIST, G., (1938); Die tonische Innervation der Skeletmuskeln. Z . mikr. anat. Forsch., 49, 169-186. HOCHBERG, I., AND HYDEN,H., (1949); The cytochemical correlate of motor nerve cells in spastic paralysis. Acta physiol. scand., Suppl. 60, 17, 5-63. KABAT,H., AND KNAPP,M. E., (1944); The mechanism of muscle spasm in poliomyelitis. J . Pediut., 24, 123-137. KATZMAN, R., (1961); Electrolyte distribution in mammalian central nervous system. Neurology, 11, 27-36. KOLMODIN, G. M., A N D SKOGLUND, C. R., (1959); Influence of asphyxia on membrane potential level and action potentials of spinal moto- and interneurons. Actaphysiol. scand., 45, 1-18. KOSMAN, A. J., HILL,J., AND SNIDER, R. S., (1951); Electromyographic and histological studies on animals made spastic by spinal cord ischemia. Fed. Proc., 10, 75-76. KROGH,E., (1950); The effect of acute hypoxia on the motor cells of the spinal cord. Acta physiol. scand., 20,263-292. LEAo, A. A. P., (1947); Further observations on the spreading depression of activity in the cerebral cortex. J. Neurophysiol., 10, 409414. LEAo, A. A. P., AND FERREIRA, H. M., (1953); AltraCBo da impedancia eletrica no decurso da depress20 alastrante de atividade do cortex cerebral. An. acad. bras. Ciencius, 25, 259-266. LLOYD,D. P. C., (1953); Influence of asphyxia upon the responses of spinal motoneurons. J . gen. Physiol., 36, 613-702. NELSON, P. G., AND FRANK, K., (1959); Effects of anoxia and asphyxia on cat spinal motoneurons. Physiologist, 2, 88-89. RANCKJR., J. B., (1963); Specific impedance of rabbit cerebral cortex. Exp. Neurol., 7, 144-152. REXED, B., (1940); Some observations on the effect of compression of short duration of the abdominal aorta in the rabbit. Actapsychiat. scand., 15, 365-398. ROMANES, G. J., (1951); The motor cell columns of the lumbo-sacral spinal cord of the cat. J. comp. Neurol., 94, 313-363. SCHADE,J. P., AND VAN HARREVELD, A., (1961); Volume distribution of moto- and interneurons in the peroneus-tibialis neuron pool of the cat. J . comp. Neurol., 117, 387-398. TARLOV, I. M., AND GELFAN, S., (1960); Rigidity from spinal interneurone destruction: Histological study. Trans. Amer. neurol. Ass., 120-122. TUREEN,L. L., (1936); Effects of experimental temporary vascular occlusion of the spinal cord. Arch. Neurol. Psychiat. (Chicago), 35, 189-807. VAN HARREVELD, A., (1941); The resistance of central synaptic conduction to asphyxiation. Amer. J. Physiol., 133, 572-581. VANHARREVELD, A., (1943); Tone and tendon reflexes after asphyxiation of the spinal cord. Amer. J. Physiol., 139, 617-625. VANHARREVELD, A., (1944a); Reflexes in the anterior tibia1 muscle after cord asphyxiation. Amer. J . Physiol., 142, 428434. VANHARREVELD, A., (194413); Survival of reflex contractions and inhibition during cord asphyxiation. Amer. J . Physiol., 141, 97-101. VAN HARREVELD, A., (1957); Changes in volume of cortical neuronal elements during asphyxiation. Amer. J . Physiol., 191, 233-242. VANHARREVELD, A., (1961); Asphyxial changes in the cerebellar cortex. J. cell. comp. Physiol., 57, 101-1 10. VANHARREVELD, A., ( I 962); Water and electrolyte distribution in central nervous tissue. Fed. Proc., 21, 659-664. VANHARREVELD, A., AND BIBER, M. P., (1962); Conductivity changes in some organs after circulatory arrest. Amer. J. Physiol., 203, 609-614. VANHARREVELD, A., AND BIERSTEKER, P. A., (1963); Asphyxial depolarization of spinal and cortical cells. Fed. Proc., 22, 280. VANHARREVELD, A., AND BIERSTEKER, P. A., (1964); Acute asphyxiation of the spinal cord and of other sections of the nervous system. Amer. J . Physiol., 206, 8-14. VANHARREVELD, A., AND HAWES, R. C., (1946); Asphyxia1 depolarization in the spinal cord. Amer. J. Physiol., 147, 669-684. VAN HARREVELD, A., AND MARMONT, G., (1939); The course of recovery of the spinal cord from asphyxia. J . Neurophysiol., 2, 101-1 11. VANHARREVELD, A., MURPHY, T., AND NOBEL, K. W., (1963); Specific impedance of rabbit's cortical tissue. Amer. J . Physiol., 205, 203-207.

304

DISCUSSION

VAN HARREVELD, A., AND OCHS,S., (1956); Cercbral impedance changes after circulatory arrest. Anier. J . Physiol., 187, 180-192. VAN HARREVELD, A,, A N D POTTER, R. L., (1961); Histochemical differentiation of chloride from other ions precipitated by silver nitrate i n freeze-substitution fixation. Slain Techno/., 36, 185-1 93. VAN HARREVELD, A,, POTTER, R. L., A N D SLOSS,L. J., (1961); Electrical conductivity and clectrolytc distribution in a secreting salivary gland. Amer. J . Physiol., 201, 1002-1006. VAN HARREVELD, A,, AND SCHADE, J. P., (1959); Chloride movements in cerebral cortex after circulatory arrest and during spreading depression. J . cell. comp. Physiol., 51, 65-77. VAN HARREVELD, A., AND SCHADE, J. P., (1960); On thc distribution and movements of water and electrolytes in the cerebral cortex. Sirirciure anti Function of [he Cerebral Cortex. D. 0 . TOWER A N D J. P. SCHADE, Editors. Proceedings of the Second International Meeting of Neurobiologists, Amsterdam. Amsterdam. Elsevier (p. 239-256). VAN HARREVELD, A., AND SCHADE, J. P., (1962a); Changes in the electrical conductivity of cercbral cortex during seizure activity. Exp. Neurol., 5, 383-400. VAN HARREVELD, A., A N D SCHADE, J. P., ( I 962b); Nerve cell destruction by asphyxiation of the spinal cord. J . Neuropathol. exp. Neurol., 21, 410-423. VAN HARREVELD, A,, A N D SPINELLI, D., (1964); Mechanisms of the extensor rigidity caused by spinal cord asphyxiation. Topics in Basic Neirrology, Vol. 6 , Progress in Brain Research. W. Bargmann and J. P. Schadi., Editors. Proceedings of the Third International Meeting of Neurobiologists. Amsterdam. Elsevier (p. 174-179). VAN HARREVELD, A., AND STAMM, J. S., (1953); Cerebral asphyxiation and spreading cortical depression. Atner. J . Physiol., 173, 171-175. VAN HARREVELD, A., A N D TACHIBANA, S., (1962); Recovery of cerebral cortex from asphyxiation. Amer. J . Physiol., 202, 59-65. VAN H A R R ~ V EA,, L D A, N D TYLER, D. B., (1942); Metabolism of asphyxiated spinal cord. Anier. J . Physiol., 138, 140-148.

DISCUSSION

GRANIT:Would there be any sprouting (a) of afferent terminals, and (b) of muscular terminals? VAN HARREVELD: In answering this question one has to distinguish between the secondary tone and the late tone. In the mechanism of the secondary tone which is present only during the first 2 to 3 days axon sprouting can hardly be involved. The development of the late tone between 10 and 14 days after asphyxiation might be accompanied by axon sprouting, however. Eccles, Eccles and Shealy did not find physiological evidence for this after cutting the dorsal roots. However, as remarked by Prof. Eccles, the deafferentiation in the rigid preparations is much more severe and axon sprouting might be a factor. It is likely that during the late tone sprouting of the muscle terminals occurs, but we have n o direct evidence for this. G R A N I T With : hyperexcitability one would also have to consider the possibility of loss o f tonic inhibitions. Perhaps one should try various methods of assessment: (a) reflexes; (h) strychnine; (c) picrotoxine. : As mentioned during secondary tone there are indications that VAN HARREVELD inhibition is markedly depressed. During the late tone this is less obvious. No exhaustive investigations of the effects of strychnine and picrotoxine in rigid preparations have been carried out.

EFFECTS OF S P I N A L CORD A S P H Y X I A T I O N

305

ECCLES:I have several comments and questions. Firstly, I would suggest that Dr. Stavraky’s evidence that cutting of dorsal roots causes development of hyperexcitability of motoneurons be regarded with caution until it can be corroborated by modern testing procedures. Secondly, I am uncertain how the recording procedures could have discriminated between asphyxia1 changes in membrane potentials of the somas and dendrites of neurons. I would regard it as more probable that the axons were less depolarized by asphyxia and consequently the sources for the soma-dendritic sinks. Thirdly, I would like to enquire if the changes in membrane potentials of neuronal elements were associated with changes in spike potentials. VAN HARREVELD: The typical discharges on penetration of the soma membrane, and the observation of antidromic potentials with the inflection indicating the presence of the action potential of the initial segment and of the soma-dendritic potential make us believe that the potentials are not axon, but soma potentials. The spike potentials during asphyxiation are now being investigated. GELFAN:I should -also -like to reply] to Prof. Granit’s question about sprouting terminals. This question was raised by Dr. McCough at the end of my report to the American Physiological Society in 1957 on experimental hind-limb rigidity in dogs. I asked him in turn about the latency for the onset of the reflex exaggeration and ‘spasticity’ in spinal animals, which he proposed as due to replacement of degenerated terminals of descending tracts by sprouts from neighbouring afferent fibres. The latency is 2 or more weeks. In view of the undisputed fact that dogs and cats, after temporary aortic occlusion, promptly exhibit the rigidity after recovery from anaesthesia, there is no possibility of dorsal root fiber sprouting, if it occurs, playing any role in this rigidity. We discussed this matter in our publication (Gelfan and Tarlov, 1959). Van Harreveld and Schade (1962) for the same reason, also did not consider terminal sprouting as a factor in this rigidity. The 7 days latency for the ‘final’ stage of rigidity whichVan Harreveld described is still lj2 of the latent period for functioning terminal sprouts. In dogs (Gelfan and Tarlov, 1959) and cats (Murayama and Smith, 1961) made rigid by temporary occlusion of thoracic aorta, the rigidity is usually maintained uninterruptedly from the beginning. Dr. Van Harreveld does not dispute the rigidity of our dogs whose lumbosacral dorsal roots had been cut extradurally 30 days before the aortic occlusion. In these there can obviously be no question about dorsal root fiber sprouting. Finally, the evidence which I presented about the reduction of synaptic density on spinal neurons of rigid dogs to about 1/4 of normal cannot support any contention about terminal sprouting in such preparations. We have also recorded electromyographically from unanaesthetized chronically rigid dogs; animals with intact dorsal roots and whose cords had not been transected. The effect of attempts to flex the rigid knees, or turning of head and neck, on the ‘spontaneous’ motoneuron discharges, as recorded with concentric needle electrodes in quadriceps and hamstring muscles, varied from animal to animal, reflecting no doubt the differences in degree of neuronal destruction. Usually very little or no

306

DISCUSSION

effect could be directly ascribed to these maneuvers since similar variations in number of discharging motor units and frequency of discharge occurred spontaneously. This is similar to the observations made on a human case of rigidity of the arms of spinal origin by Rusworth et al. (1961). The Oxford group found very little evidence of a true stretch reflex in the muscles involved. As previously pointed out, the unremitting pillar-like rigidity of the dog’s hind limbs cannot be an exaggerated stretch reflex (Gelfan andTarlov, 1959). It is comparable to an ‘u’-rigidity, to use Granit’s terminology, since it is not abolished or prevented by dorsal root section. The extensive destruction of interneurons must also interrupt the y-loop. If it is still felt necessary to consider this hind-limb rigidity in reflex terms there can now be hardly any doubt that it is the excitability of the motoneurons which is exaggerated. The contention that really complete deafferentiation of motoneurons is not possible is essentially only a formal one. As 1 showed, the extensive destruction of lumbosacral interneurons reduces the synaptic density on cell bodies and dendrites in the rigid preparations to about 1/4 of normal. This degree of denervation of motoneurons appears to be enough to alter motoneuron characteristics in such preparations. The immediate external environment of the soma membrane is altered by this removal of synaptic covering over the surface. As already suggested (Gelfan, 1963), the denervated motoneuron is also deprived of a stabilizing influence normally provided by input terminals in addition to dying impulses. These negative factors may constitute the ‘stimuli’ for the oscillating membrane potentials which, when large enough, induce spontaneous discharges. There can hardly be a pacemaker for these discharges since they are asynchronous. Each motoneuron is its own pacemaker. As emphasized in my presentation, most or even all of the motoneurons in our rigid dogs may survive when 75% or more of the interneurons in the entire L7 spinal gray is destroyed. The intense pillar-like rigidity of the hind limbs in our dogs also reflects the high motoneuron survival rate. Dr. Van Harreveld used a different method of producing temporary ischemia of the cat lumbrosacral cord, previously isolated by transection. The motoneuron survival, at least in the peroneus-tibialis nucleus of his cats, was only about 15 %and he characterizes the hind-limbs as having ‘extensor’ or ‘high extensor tone’. One wonders about the additional effect of dorsal root section on such a small percentage of surviving motoneurons. Sherrington had already noted more than half a century ago the profound and permanent effect of rhizotomy. The effect of cutting dorsal columns, on the other hand, appears to be temporary. Ransom also claimed that it is not possible to cut dorsal roots without damaging the cord. VAN HARREVELD: Dr. Gelfan’s insistance that the rigidity of animals with asphyxiated cord is due to a spontaneous discharge of deafferented motor cells is in contradiction to the monosynaptic action potentials which can be recorded from such preparations. As found, also by Gelfan and Tarlov, these action potentials are of unusual magnitude, notwithstanding an appreciable loss of motoneurons. This indicates that a large percentage of the motor neuron pool is activated which supports the very fruitful suggestion of Gelfan and Tarlov that the excitability of the moto-

EFFECTS O F S P I N A L CORD A S P H Y X I A T I O N

307

neurons in the rigid preparations i s enhanced. It would seem logical that this enhanced monosynaptic activity is the basis for the high extensor tone of the asphyxiated preparations. The enhanced activity in the myogram during stretch of muscles in rigid preparations supports this mechanism. An ultimate increase in excitability might result in spontaneous discharges of the motoneuron, however. Dr. Gelfan questioned the possibility that the surviving motoneurons in our preparations could account for the hypertone observed. It would seem, however, that the surviving lj5 of the motoneurons when all activated by volleys of impulses from muscle spindles during attempts to bend the joint would be able to produce a contraction of considerable force. Furthermore it is possible that by terminal branching a materially larger part of the muscle becomes innervated than the percentage of surviving motoneurons would indicate. HUGHES:May I mention a human case that is relevant to the experimental rigidity produced in cats by the speaker, and in dogs by Prof. Gelfan. This case, a man with a cervical cord tumour, developed intensepgidity of the upper limbs. At necropsy isolated motoneurons were demonstrated and considered to be responsible by their functional isolation for the unusual rigidity. The case was published (Rusworth et al. (1961).

VAN HARREVELD: I am familiar with this interesting observation. A destruction of interneurons by tumor growth could well account for the rigidity observed in this patient. WIESENDANGER: The problem seems to remain whether this rigidity is caused by the y- or the a-fibers. As you said, cutting of the dorsal roots is not a criterium to decide this question because the deafferentiation as such may result in a rigidity. I have investigated electromyographically this deafferentiation rigidity in a group of animals with section of the roots intradurally and extradurally. In the intradurally deafferentiated animals the rigidity developed only after 3 weeks or later and there was no rigidity in the extradural group. So this may be a way to approach this problem. And if you say that this rigidity came a few days after the deafferentiation again I would suggest that this was a real a-rigidity and not due to hypersensitivity of the a-motoneurons. VAN HARREVELD: There is no doubt that a large percentage of the y-efferents are destroyed in asphyxiated preparations. However, it would seem possible that the muscle contracture which tends to develop also includes intrafusal fibers promoting the sensory discharge of muscle spindles. Although an enhanced y-efferent discharge can therefore not be excluded, we believe that the increased excitability of the amotoneurons is the main factor in the late tone. REFERENCES RUSWORTH, G . , LISHMAN, W. A,, HUGHES, J. T., AND OPPENHEIMER, D. R., (1961); Intense rigidity of the arms due to isolation of motoneurons by a spinal tumour. J . Neurol. Neurosurg. Psychiat., 24, 132-142.

308

Author Index * Ades, H. W., 247 Adrian, E. D., 39, 191 Amassian, V. E., 63, 188, 224 Andersen, P., 111, 191, 200, 203, 214, 262 Anderson, F. D., 154 Anderson, R. F., 142 Anderson, S. A. ,148, 155 Araki, T., 4, 5, 7, 8, 17, 18,21, 35,42, 43,49, 198 Arduini, A., 56, 253 Ballif, L., 207 Barrera, S. E., 185 Barron, D. H., 65, 95, 96, 98, 102, 106, 107, 111, 120,203

Beck, G. M., 142 Becker, M. C., 7 Beevor, C. E., 239 Berlin, L., 188 Bernhard, C. G., 65, 206, 210,222, 237, 239 Berry, C. M., 154, 155 Bessou, P., 137 Biber, M. P., 283 Biersteker, P. A., 281,282,285,287-290,298,300 Blair, E. A., 262 Blumenau, L., 185 Bohrn, E., 154,206, 222, 237, 239 Bonnet, V., 65, 106 Bradley, K., 35, 122, 137 Bremer, F., 65, 106 Bricker, J. W., 266 Brock, L. G., 4, 13, 42, 57, 264, 267 Brodal, A., 56, 63, 142, 154, 173, 185, 207, 255 Bronk, D. W., 39, 191 Brookhart, J. M . , 23, 42, 43, 49 Brooks, C. McC., 65,96, 101, 106,207,210,21 I , 289, 294

Brooks, V. B., 27 BureS, J., 282 Busch, H . F. M., 146 Caldwell, P. C., 2, 12 Campbell, B., 135, 140, 144, 184, 270 Carpenter, D., 63, I l l , 159, 201-204, 210-214 Carrea, R. M. E., 184 Carter, W. B., 146 Catalano, J. V., 149 Cattell, H., 207

*

Chambers, W. W., 185 Cobb, S., 207 Combs, C. M., 184 Coombs, J . S., 3,4,6,7, 10, 1 I , 13, 17, 19,21,42, 57, 76, 232, 264,267

Cooper, S., 125, 173, 222 Covian, M. R., 187 Creed, R. S., 125 Critchlow, V., 262 Curtis, D. R., 9, 16-18, 22, 25, 75, 119, 123, 129, 142, 144, 168, 240, 268

Dell, P. C., 111 Demirijan, C., 148 Denny-Brown, D . B., 39, 222 DeVito, R. V., 63 Doty, R. W., 247 Douglas, W. W., 93 Downman, C. 9. B., 294 Dun, F. T., 106, 124 Earle, K. M., 110 Easton, D. M., 122 Eccles, J. C., 1-34, 35, 42, 47, 49, 50, 57, 65-91, 96, 101, 106, 108, 109, 119-124, 126, 128, 129, 137, 142-144, 147,150-152,154,156, 173, 179, 183, 186,191, 197-200,203,204,207-209,211, 214,215,232,233,240,264-268,289,294,295 Eccles, R. M., 8, 9, 13, 14, 20, 42, 63, 66, 68, 72, 101, 111, 119, 120, 137, 147, 173, 197, 259, 260, 267 Eide, E., 198 Eisenman, G., 135, 164 Ekland, G., 262 Eklund, K., 206, 277 Eldred, E., 270 Engberg,I., 63, 111, 159, 199, 200, 210-214, 274-279 Erlanger, J., 262 Escolar, J., 164, 172

Fadiga, E., 23, 42, 43, 49 Fatt, P., 3, 4, 6, 7, 10-13, 19--21, 76, 119, 123, 198, 232, 267

Feng, T. P., 124 Ferraro, A., 185 Ferreira, H. M., 283

Italics indicate the pages on which the paper of the author in these proceedings is printed.

AUTHOR INDEX

Fifkovri, E., 282 Flechsig, P., 142 Florey, E., I19 Forbes, A., 207 Frank, K., 7, 65, 68, 93, 101, 120, 168, 197, 287 Frankenhaeuser, B., 46 Freygang, Jr., W. H., 284 Fulton, J. F., 207 Funkenstein, H., 210-214 Fuortes, M. G. F., 7, 65, 68, 93, 101, 120, 197 Gasser, H. S., 65, 67, 106, 262 Gelfan, S., 289, 290,292, 296-298, 300 Gesell, R., 266 Gordon, G., 149 Graham, H. T., 65, 67 Granit, R., 14, 35-4/, 42-44, 47, 49, 191, 197, 270

Grant, G., 141, 142, 151, 184, 185, 190 Gray, E. G., 73, 78 Grimby, L., 206, 277 Grundfest, H., 125, 135, 140, 144, 146, 184, 262 Haapanen, L., 20 Haartsen, A, B., 255 Hagbarth, K. E., 11I, 199, 210, 274 Haggqvist, G., 289 Hagiwara, S., 101 Hammond, P. H., 270 Harrison, C. R., 187 Hashimoto, Y . , 54 Hawes, R. C., 280, 282, 287 Henatsch, H. D., 14 Hern, J. E. C., 222, 223, 232, 239 Hill, J., 289 Hochberg, I., 280 Hodgkin, A. L., 2, 3, 11, 12, 14, 46 Holmqvist, B., 63, 138-140, 142-144, 150, 154, 156, 157, 164, 166-168, 171, 172, 180, 181, 185-187, 191, 192, 199, 209, 210, 216 Howell, J. B. L., 270 Howland, B., 65, 96-98, I08 Hubbard, J. I., 9, 14, 21, 24, 27, 149-152, 154, 173, 179, 183 Huber, G. C., 264 Hugelin, A., 207 Hughes, J., 106 Hunt, C. C., 9, 20, 80, 93, 122, 261 Hyden, H., 289 Hyndman, 0. R., 175

309

Kabat, H., 289 Kandel, E. R., 50 Katz, B., 77, 78 Katzman, R., 284 Kernell, D., 35-39, 42-55 Kerr, D. I. B., 1 11, 210 Keynes, R. D., 2, 12, 14 Kiraly, J. K., 129 Kitai, S. T., 148 Kleyntjens, F., 207, 210, 21 1 Knapp, M. E., 289 Koizumi, K., 207, 210, 21 I Koketsu, K., 12, 13, 20, 54, 64, 107, 119, 123 Kolmodin, G. M., 20, 94, 287 Koshtoyants, 0. Kh., 282 Kosman, A. J., 289 Kostyuk, P. G., 20,21,27, 56, 65, 74, 77, 78, 101, 107-1 10, 120, 121, 204, 294, 295

Krivanek, J., 282 KrnjeviC, K., 27, 168 Krogh, E., 289, 300 Kuffler, S. W., 262 Kugelberg, E., 206, 274, 277 Kuno, M., 9, 20, 101, 208 Kuypers, H. G. J. M., 154, 207 Lamarche, G., 149 Landau, W. M., 284 Landgren, S., 23, 155, 198, 206, 222-225, 230, 232, 235, 237, 239, 240

Laporte, Y . , 135-140, 142, 144, 164, 186. Lelo, A. A. P., 282 Leksell, L., 197 Lettvin, J. Y . ,65, 95-98, 108, 111 Leyton, A. S. F., 241 Libet, B., 50 Liddell, E. G . T., 125, 207 Liley, A. W., 27, 77 Limanski, Y . P., 56 Lindblom, U. F., 198, 207 Liu, C.-N., 164, 172 Livingston, A,, 189, 198 Lloyd, D. P. C., 26, 27, 63, 98, 102, 104, 106, 112, 126, 135, 146, 179, 191, 198, 289, 290

Longo, V. G., 119 Lorente de Nb, R., 5, 6, 9, 142 Lundberg, A., 8,9, 13, 14,20,42,63, 72, 78, 105, 107, 111, /35-163, 164, 171-174, 179, 182, 184, 186, 187, 191, 192, 197-221, 255, 267, 277

Iggo, A., 20, 93 Ito, M., 12, 14, 21, 54, 198

Machne, X., 42, 43, 49 Magee, C., 266 Magni, F., 56-64, 65, 66, 68, 69, 71-73, 76, 101,

Jansen, J., 142, 173, Job, C., 208 Johnson, A. R., 27, 102 Jukes. M. G. M.. 149

120, 156-160, 164, 166, 169, 174, 197, 203, 207, 210, 215, 246-258 Malcolm, J. L., 65, 76, 96, 101, 106, 124, 126 Mark, R. F., 20, 135, 139 Marmont, G., 289, 290, 292

310

AUTHOR INDEX

Martin, W. R., 119 Matthews, B. H. C., 65, 95, 96, 98, 102, 106, 107, 111, 120, 203 McCulloch, W. S., 65, 95-98, 108, 111 Mclntyre, A. K., 20, 80, 93, 98, 106, 122, 135, 139, 146, 179, 187, McLennan, H., I 1 9 Mendoza, E. L., 264 Merton, P. A., 270 Miledi, R., 27, 168 Morin, F., 148, 149, 154 Moruzzi, G., 56, 253 Mountcastle, V. B., 187 Murakami, M., 54 Murphy, T., 284 Nathan, P. W., 270 Nauta, W. J . H., 154 Nelson, P. G., 287 Nishi, S., 54 Nobel, K. W., 284 Norrsell, U., 145, 147-149, 155, 157, 201 -205, 255 Nyberg-Hansen, R., 207 Ochs, S., 282, 283, 286 O’Leary, J. L., 154 Oscarsson, O., 9,14,20-24,63,135-146,149-160, 164-178, 179-196 Oshima, T., 12, 14, 21 Otani, T., 7, 35,42,43 Ottosson, J. O., 198, 207 Paine, C. H., 149 Paintal, A. S., 199 Paton, W. D. M., 119 Patton, H. D., 223 Perl, E. R., 208 Perry, W. L. M., 119 Petersen, I., 206, 222 Phillips, C. G., 14,23, 189, 191, 198,206,222-245 Phillis, J. W., 119, 129 Pitts, W., 65, 95-98, 108, 111 Pompeiano, O., 142, 185 Porter, R., 23, 206, 222-245 Portnov, H., 148 Potter, R. L., 282 Preston, J. B., 222, 223, 233 Quilliam, J . P., 262 Rall, W., 26, 27 Ramos, J. G., 264 Ranck, Jr., J. B., 284 Renkin, B., 40 Renshaw, B., 20, 65, 96 Reuben, J. P., 125 Rexed, B., 92, 142, 179, 289 Rickles, N. H., 125

Ritchie, J . M., 93 Robbins, J., 125 Romdnes, G. J., 291 Rostn, I., 165, 166, 169, 170, 185-191 Rossi, G. F., 56, 154, 255 Rudin, D. C., 135, 164 Rutkowski, S., 262 Sasaki, Y . , 54 Schade, J . P., 282, 284, 291, 298 Scheibel, A . B., 56, 63, 255 Scheibel, M. E., 56, 63, 255 Schimert, J., 164, 172 Schmidt, R. F., 20, 21, 27, 65, 67, 68, 70-80, 82-86, 101, 107-109, 119-134, 144, 152, 204, 207, 210, 215, 295 Schwartz, H. G., 154 Sears, T. A,, 23, 24, 11 I , 200, 203, 214, 259-273 Shaw, T. I., 2, 12 Shealy, C. N., 259, 260, 265-267 Sherrington, C . S., 11, 125, 173, 185, 197, 206, 207, 222, 241, 274 Shortess, G. K., 35, 37-39 Skoglund, C. R . , 20, 287 Skoglund, S., 14, 20, 190 Sloss, L. J., 283 Smith, R. S., 35, 36,42-44, 47,49 Snider, R. S., 184, 289 Somjen, G. G., 35 Sowton, S. C. M., 207 Spencer, W. A., 50 Spinelli, D., 293 Sprague, J. M., 164, 168, 172, 173, 185 Stamm, J. S., 282 Steg, G., 14 Steiner, J . , 20 Stookey, B., 175 Stowell, A,, 184 Strom, G., 179 Sutton, G. G., 270 Tachibana, S., 282, 293 Takeuchi, A., 21, 27, 76, 101 Takeuchi, N., 21, 27, 76, 101 Tarlov, I. M., 289, 290, 292, 296-298, 300 Tasaki, I., 101 Taub, A., 94 Tauc, L., 54 Terzuolo, C . A., 4, 5, 7, 8, 17, 18, 49 Thesleff, S., 190 Toennies, J. F., 95 Tomita, T., 54 Torvik, A,, 56, 63 Tower, S. S., 207 Tureen, L. L., 289 Tyler, D. B., 301 Uddenberg, N., 165-167, 169-171, 180-184 Urnrath, K., 124

A U T H O R INDEX

Unna, K. R., I19 Van Der Kloot, W. G., 125 Van Harreveld, A., 280-307 Von Euler, C., 262, 272 Voorhoeve, P., 145, 147, 149, 156, 157, 198-200, 204, 205, 216, 255, 277 Vyklicky, L., 105, 107, 207, 213, 215, 216 Wall, P. D., 20, 27, 65, 72, 92-118, 120, 173, 203 Watkins, J. C., 119, 129 Weiss, T., 282 Whitlock, D. G., 56, 222, 223, 233, 253

31 1

Willis, W. D., 20, 23, 56-64, 65, 67-73, 75, 76, 79, 80, 82-86, 119-122, 124, 128, 142-144, 152,156,l86,203,201,207,210,215,246-258, 295 Wilson, V. J., 191 Winsbury, G., 142, 186 Wolkin, J., 175 Wolpow, E. R., 156 Wood Jones, F., 241 Woolsey, C. N., 188 Young, R. R., 50 Zanchetti, A., 56, 255

312

Subject Index Accommodation, asynchronous synaptic bombardment, 28 relation to adaptation, 38 Adaptation, cortical potential depression, I90 frequency of discharge, motoneuron, 186 motoneuron membrane, 38, 39 relation to accommodation, 38 relation to after-hyperpolarization, 55 Afferent fibers, Ia from sural muscle, relation to EPSP, 101 connections to reticulo-spinal neurons, 246-258 cutaneous, depolarization and presynaptic inhibition, 102 inhibition, 80-83, 85 membrane potential, 103 post-tetanic potentiation, 102 effect of synaptic depolarization, 54 flexor reflex, reticulo-spinal neuron, 63, 107 presynaptic, inhibition of Ia, 66, 79, 107 inhibition of Ib, 79-81, 107 localization of, 121 relation to spinal interneurons, 147 After-hyperpolarization (see also Hyperpolarization) action on EPSP, 55 cutaneous relay cells, 14 effect on adaptation, 39 effect on potassium-sodium balance. 14 primary afferent depolarization, 75 properties of spinal neurons, 12-15 relation to delayed depolarization, 44 relation to SD spike, 12 time course, relation to delayed depolarization, 45 transmembrane stimulation, 36 Amino acids, effect of dorsal root potential, 129 Anaesthetics, effect, on blood pressure, DRP, I34 on motoneuron, 244 on sensorimotor cortex, 220 Antidromic impulse, activity, identification of DSCT, 141 identification of VSCT, 149, 150 after-potential and refractory period, 58

delayed depolarization, 43, 45 evoking intracellular responses, 3-6, 35 identification of reticule-spinal neurons, 246 inhibition of dendrites, 1 I relation to Na-injection, 10 relation to presynaptic potentials, 75 repetitive effect, 45 slope characterization, 12 Apnoea, effect on motoneurons, 263 Asphyxiation, cell membrane permeability, 286, 287 chloride transport, 284-287 effect on metabolism, 301 effect on spinal cord, 280-307 mechanism, arrest of reflex activity, 288, 289 myotatic rigidity, 298, 300, 306 nerve cell destruction, 29 I , 292 survival time, 300, 301 Barbiturates, effect on presynaptic inhibition, 70 primary afferent depolarization, 126 sensitivity of substantia gelatinosa, 1 1 3 Brain stem, control centers of inhibition, 209 effect on delayed depolarization, 44 effect on membrane potential, 48, 49 primary afferent depolarization, 21 I , 2!2 supraspinal control, motoneurons, 197-221 Cerebellar cortex, asphyxial chloride transport, 284-287 projection forelimb nerves, 192 projection group I afferents, 181-191 stimulation, antidromic activity, 141, 142 Cerebral cortex, areas stimulated, effect on reticulo-spinal neurons, 246,255 postcruciate, 247 primary auditory, 247, 257 primary visual, 247, 257 sensorimotor, 248 asphyxial chloride transport, 284-287 circulatory arrest, 286 effect on motoneurons, 206, 207 evoked potential, somatic area, 188, I89 facilitation of reflexes, 197-207 localization, facilitatory system, 239 motor control, 244

SUBJECT INDEX

projection forelimb nerves, 192, 195 sensorimotor, conditioning system, 277, 278 Cerebrospinal fluid, electrolytes during asphyxiation, 284 Cholinergic drug, effect on dorsal root potential, 129 Conduction, blockade, orthodromic versus antidromic, 95 cortico-spinal, 227, 228 Control, cortical, on reticulo-spinal neurons, 255 decerebrate inhibition, 210 inhibitory, reflex pathways, 215-217 presynaptic impulse transmission, blockade, 95-98 depolarization, 98-1 02 hyperpolarization, 102-105 supraspinal, motoneurons, 197-221 plantar reflex, 277 Convulsan ts, picrotoxin, inhibitory effects, 125 strychnine, inhibitory effects, 122-124 Cuneate nucleus, descending pathways, 196 relation to cuneo-cerebellar tract, 185 relation to group I afferents, 179, 193 Cuneo-cerebellar tract, characterization of activation, 185-1 87 discharge pattern, 190 properties of neurons, 190 relation to cuneate nucleus, I85 subdivision, 185, 187 Current, action on glial cells, 90 determination by voltage-clamp method, 18 effect on delayed depolarization, 46-49 electrotonic spread, 16, 17 extracellular relation to dendrites, 5 hyperpolarizing effect on depolarization. 47, 49, 51 polarizing, effect on monosynaptic activity, 76. 77 Decerebrate animal, descending control, 219 effect on motoneurons, 208 tonic inhibition, 208-211, 216 Dendrite, asphyxial arrest of reflex activity, 288, 289 asphyxial potential, 287 chloride transport, spinal cord, 285 control mechanism for transmission, 106 delayed depolarization, 36 crustacean stretch receptors, 53 depolarization, recovery after asphyxiation, 290 measure of conduction time, 5, 6 reticulo-spinal neurons, 257

313

Depolarization, afferent, relation to depolarizing synapses, 73, 76, 79, 80, 83, cutaneous afferent fibers, 98, 113 delayed, chromatolysed motoneurons, 50 effect of antidromic invasion, 35, 44 effect of polarizing currents, 46, 47 general characteristics, 42-44 relation to somadendritic membrane, 44 reticular formation, 54 time course, 45 dendritic, asphyxial arrest, 288, 289 excitability test, 72 presynaptic fiber, 75 primary afferent, effect of anaesthetics, 128 methods of recording, 120 properties of initial segment, 36 reticular neurons, 257 threshold level, 7 voltage-clamp recording, 7 Discharge, adaptive, I86 ascending tracts, 180 asphyxiation, motoneuron, 305 massive, Ia and I b afferents, 136-138 ascending tract, 164, 165, 181 effect in cuneo-cerebellar tract, 185 initial component, 166 medial lemniscus, 191 monosynaptic, 191 repetitive, during central respiratory drive, 266 effect from subsynaptic currents, 35 effect on membrane, 106 relation to antidromic invasion, 37 reticulo-spinal neurons, 63 DOPA, inhibition of primary depolarization, 132, 133 Dorsal spino-cerebellar tract (DSCT), conditioning cutaneous stimulation, 183, 184 discharge pattern, 135 monosynaptic activation, 171 properties of neurons, 140-146, 173, 179 convergence, summation, 143, 144 cutaneous afferents, 140 propriocertive information, 144 termination, 140, 142 relation to cuneo-cerebellar tract, 187 termination area, 184 DSCT (see Dorsal spino-cerebellar tract) Electromyogram, during expiration and inspiration, 261 quadriceps muscles, 298-300 EPSP (see Excitatory postsynaptic potential)

314

SUBJECT INDEX

Eserine, inhibitory effect on transmitter, 20 Excitation, cortical stimulation, 205 flexor reflex afferents, 183 orthodromic stimuli, 100 pre- and postcruciate cortex, 248 relation to inhibition, 40 relation to initial segment, 9 reticulo-spinal neurons, from cortex, 255 sensorimotor cortex, effect on afferents, 198-200 Excitatory postsynaptic potential (EPSP), central tegmental tract, 251, 252 cortically evoked, 240 current flow, 17 depression, 66-79, 101 influencing factors, 78 relation to presynaptic depolarization, 71, 101 effect of membrane potential, 20 ionic flux, 28 monosynaptic, 18 repetitive stimulation, 22, 23 relation to post-tetanic potentiation, 26 relation to presynaptic inhibition, 67 Facilitation, convergence of cortico-spinal nerves, 255 extrapyramidal, 160 flexor reflex, 220 recovery response, dorsal root, 234, 295 sensorimotor cortex, 197-207 temporal presynaptic inhibition, 68 Feedback, forelimb information, 192, 195 relation t o spinal control mechanism, 90 Glutamic acid, depolarization of primary afferent fibers, 129 effect on massive discharge, 137 relation to presynaptic inhibition, 79 Hippocampus, depolarization of pyramids, SO, 54 Hyperpolarization (see also After-hyperpolarization), peripheral nerves, 104 primary afferents, 201 relation to cortical depolarization, 102-104, 1 I3 respiratory drive potential, 266, 267 Impedance, nature of extracellular electrolytes, 282-284 registration in spinal cord, 280, 301 relation to chloride transport, 285

Inhibition (see also Presynaptic inhibition), afferent fibers, 65, 66 cortical stimulation, 205 hyperpolarization of synaptic stimuli, 48 interneuronal pathways to motoneurons, 21 1 pathway to la afferents, 203, 204 potential generator, 65 presence in spinal cord, 98 recurrent, effect on motoneuron adaptation, 39 reflex paths to motoneuron and primary afferents, 201-215 relation to facilitation, 40 relation to initial segment, 9 thalamic relay, I91 tonic decerebrate inhibition, 208-21 1 Inhibitory postsynaptic potential (IPSP), antidromic activity, 12, 13 occurrence in gray substance, 223 presence in spinal cord tract, 144 Interneurons, decerebrate tonic inhibition, 216 denervation hypersensitivity, 297 destruction after asphyxiation, 292, 302, 306 effect of anaesthetics, 128 excitability, relation to strychnine, 124 extrapyramidal facilitation, 160 inhibitory regulation, 216 local inhibition, 233 nembutal, rate of discharge, 128 picrotoxin, depression of presynaptic inhibition, 125 rate of discharge, relation to dorsal root potential, 134 sensorimotor influence on, 204-206 source of primary afferent depolarization, 78, 79, 81, 85, 107, 119, 121 Ionic composition, alteration by current flow, 1I distribution on neuronal membrane, 1 effect on EPSP, 21 relation to spinal cord asphyxiation, 283 IPSP (see Inhibitory postsynaptic potential) IS spike (see Potential) Metabolic pump, influence of electrolytes, 2, 11, 12, 14, 15 spinal asphyxia1 potential, 292, 293 Motoneurons, adaptation membrane, 38, 39 asphyxiation membrane, 286, 287 effect of anaesthetics, 244 effect of recurrent inhibition, 39 influence of decerebration, 208 inhibitory pathways, 201-215 properties of respiratory, 23 relation to interneuronal pathways, 21 1 stimulation cortex. potential changes, 206,207 supraspinal control, brain stem, 197-221

SUBJECT INDEX

Nembutal, effect, on blood pressure, discussion, 134 on dorsal root potential, 126 on retina, discussion, I33 interneuron, decreased rate of discharge, 128 Nerve, intercostal, discharge, 260, 261, 272 fusimotor fiber function, 262, 272 Neuroglia, asphyxiation, survival time, 301 electrolytes during asphyxiation, 284 electrophysiological relationships to neurons, 90 Neuron, activation from muscle spindle, 138 adaptation, frequency of discharge, 39 amplifier, function, 35 dipole model, 33 effect of hypoxia, 262-264 massive discharge, 139 measurement of cell volume, 291 recurrent inhibition, I I rebistance to oxygen lack, 287 size, 292 Pathway, la afferents, presynaptic inhibition, 78 ascending spinal, cat, 135-160 cortico-reticulo-spinal, 255 cutaneous, presynaptic control, 92-1 18 input-output relations, 95 presynaptic inhibitory, 119, 120 sensorimotor to motoneurons, 199, 200 supraspinal control, 153, 156, 157 Peduncle, superior cerebellar, relation to VSCT, 149 Permeability, after circulatory arrest, 286 change during postsynaptic, 65 ionic, relation to currents in EPSP, 21 potassium, during spike potential, 3 relation to spike potential, 11 relation to transmitter, 76 Picrotoxin, presynaptic inhibitory effect. 124, 125 Post-tetanic potentiation, control depolarization, 102 pyramidal tract synapse, 243 relation to EPSP, 26 Potassium, permeability, spike potential, 3 Potential, asphyxial, 280, 281 origin, 287, 288 central respiratory drive, 264-267 dorsal root, 65, 67, 99 cutaneous afferent activation, 220 @

315

effect, of amino acids, 129 of brain stem, 115 of cholinergic drugs, 129 of strychnine injection, 123 evoked by ventral roots, 126, 127 hyperpolarization, 102, 214 tonic activated, 105 electrochemical, 2 endplate, 21 equilibrium, electrolytes, I , 20 extracellular complex, 4-6 interneuronal rate of discharge, 134 membrane, 3, 4 depolarization afferent fibers, 98 displacement by current, 20 effect from voltage-clamp method, 19 effect of tetanic stimulation, 47 level, effect on depolarization, 46 negative after-potential, 46 postsynaptic, reticulo-spinal neurons, 59 relation to depolarization, 50, 52 relation to presynaptic inhibition, 107 spike, of spinal neurons, 3-12 IS, antidromic stimulation, 4 recurrent inhibition, 1I relation to EPSP, 16 relation to SD, 13 reticulo-spinal neurons, 57-59 threshold depolarization, 7 SD, antidromic stimulation, 4 relation to IS, 9 reticulo-spinal neurons, 57-59 Potentiation, conditioning tetanic, 25, 26 EPSP, cortical stimulation, 243 post-activation, 23 repetitive stimulation, 23 respiratory motoneurons, 268, 271 Presynaptic inhibition (see also Inhibition), afferent neurons, producing, 121 antidromic and orthodromic, 105 blockage of spinal cord conditioning, 96 depolarization, 66, 76, 87 effect from prolonged asphyxiation, 295 factors, depolarization and local current flow, 78 ionic flow origin, 78 magnitude, relation to EPSP, 77, 101 pharmacological aspects, 119-1 34 regulation sensory input, 213, 214 relation to dorsal root potential, 107, 1I5 relation to superimposed action potential, 75, 76

reticulo-spinal neurons, 63 spinal cord, 65-89, 197

3 16

S U B J E C T INDEX

Receptors, exteroceptors, information storage, 145 projection to cerebral cortex, 189 proprioceptors, information storage, 144 skin, afferent neurons to cortex, 93 Recruitment, a-fusimotor neurons, 263 inspiration, 261 Reflex, control, 243 y-control, 197 cortico-spinal tract, facilitation, 197-200 depression, presynaptic inhibition, 69, 70 dorsal root, 95, 105, 210, 21 1 y-efferent system, 298 flexor afferent, 139, 145, 147, 152-154, 156 effect on ascending spinal pathway, 156160, 199 excitatory inhibition to motoneurons, 209 influence on reticulo-spinal neurons, 257 inhibitory action on interneurons, 206 polysynaptic effect, 192 tonic decerebrate inhibition, 208 mechanism of asphyxia1 arrest, 288, 289 monosynaptic test, 275, 276 plantar, supraspinal control, 277, 278 presynaptic inhibition, 66 sensorimotor, excitatory and inhibitory, 199,200 stretch, relation to tonic firing, 39 stretch, respiration, 270 Refractory period, antidromic invasion, 58, 63 delayed depolarization, 44, 49 Renshaw cells, comparison to synaptic depolarization, 75 relation to dorsal root potential, 134 relation to IPSP, 12 synchronous synaptic bombardment, 20 time of transmitter action, 90 transmitter function, acetylcholine, I 19 Respiration, activity pattern, motoneurons, 264 central drive potential, 264-267, 269, 271 hyperventilation, 263 inflation reflex, 264, 269 influence on stroke volume, 264 recruitment of respiratory neurons, 261 Reticular formation, afferent connections, 246-258 delayed depolarization, 54 descending and ascending axons, 59-62 influence on ascending conduction, 177 influence on border cells, 177 neurons, properties, 56-64 presynaptic inhibition, 63 properties of axon$, 59-62 relation to EPSP, 59

reticulo-spinal inhibiting system, 159, 160 Reticulo-spinal neurons, afferent connections, 246-258 central tegmental tract, 255 cortical control, 255 mesencephalic stimulation, 251 monosynaptic cortical connection, 248

SD spike (see Potential) Spinal cord, Ia afferents, 198 ascending hindlimb pathways, cat, 135-1 60 ascending tracts, course and organization, 164-1 78 asphyxiation, effects from, 280-307 conductivity after aorta clamping, 281 cortico-spinal inhibition, 233, 234 cortico-spinal population, 223-228, 241 descending influence, dorsal root potential, 104 descending influence on ascending conduction, I77 distribution of synaptic activity, 174, 175 dorsal tracts, polysynaptic, 171 effect of circulatory arrest, 286 fascicles, dorsal, 166, 167 intermediate, I67 lateral, 170 ventral, 166, 167, 170 funicular cells, location, 176 horizontal lamina, presynaptic control, 92 hypersensitivity, 300 intermediate nucleus, 73, 75, 78 location of respiratory motoneurons, 259-273 nerve cell destruction, asphyxiation, 291, 292 oxygen lack compared t o cortex, 287 primary afferent depolarization, 21 3 pyramidal synapses, 256 reflex facilitation, sensorimotor cortex, 197207 respiratory motoneurons, 259-273 reticulo-spinal tract, 220, 243 structure of cuneo-cerebellar tract, I79 supraspinal control, motoneurons and primary afferents, 197-221 thalamic relay, 191 voltage contour maps, 108 Spino-thalamic tract, characterization, I75 contralateral receptive field, 175 Strychnine, action on pre- and postsynaptic inhibition, 91, 131, 121-124 convulsive effect, 124, 132 effect on tonic stretch, 131 Substantia gelatinosa, dorsal root potential and inhibition, 96, 116 relation to lamina IV, 93, 134

SUBJECT INDEX

source of primary afferent depolarization, 107 Succinyl choline. effect on cortical potential, 190 Synapse, activity, effect on delayed depolarization, 47, 49, 51 axo-axonic, in intermediate nucleus, 78,79 cortico-spinal, 240 distribution, IS and SD membranes, 13 excitatory, 15-28 action from antidromic impulse, 4 mechanism of, 21 time course of EPSP, 17 hyperpolarizing stimuli, 48 inhibitory stimuli, 48, 277 monosynaptic, characteristics, frog, 170 cortico-spinal excitation, 239, 242 distribution, ascetlding tract, 171, 172, 175 dorsal fascicle, 166, 167 EPSP, cortex, 248 excitatory, 204 forelimb nerves stimulated, 167 presynaptic inhibition, 66 respiratory excitation, 267 sacral caudal roots, 168 polysynaptic, distribution, ascending tracts, 171-1 73 dorsal fascicle, 167 reflex paths to a-motoneurons, 199 post-activation potentiation, 23-26 potentials, 232 potentiation, 23

317

presynaptic spikes from after-hyperpolarization, 27 pyramidal, 256 respiratory, 259 survival time, 301 vesicles, content of transmitter, 28 Thalamus, corticopefal afferent fiber, discussion, 196 relay function, 191 Transmission, influence of hyperpolarization, 105 relation between block and presynaptic inhibition, 101 Transmitter, acetylcholine, 28 action time, 90 concentration, relation to EPSP, 23, 101 in central respiratory drive potential, 267 interneurons, 40 potentiation, effect on, 27 rate of liberation, repetitive stimulation, 26 relation to afferent depolarization, 75, 76 Ventral spino-cerebellar tract, characterization, 148-1 53, 173, 179 conditioning cutaneous stimulation, 183, 184 connection, 159 monosynaptic activation, 171 relation to border cells, 177 terminal area, 184 Vesicles, relation to afferent depolarization, 73

PRINTED I N T H E NETHERLANDS