Renshaw cell activity in man1 - BMJ · Renshaw(1946)first describedthechangeswhich occur in the excitability ofspinal motoneurones as a result ofstimulation ofventral roots in the

Journal of Neurology, Neurosurgery, and Psychiatry, 1973, 36, 674-683

Renshaw cell activity in man1

J. L. VEALE2 AND SANDRA REES

From the Van CleefFoundation Laboratory, Alfred Hospital,and Department ofPhysiology, Monash University, Clayton, Victoria, Australia

SUMMARY The H-reflex elicited in triceps surae by percutaneous stimulation of the posterior tibialnerve was conditioned by stimuli applied through the same electrode. The differential sensitivity ofmotor and sensory fibres to duration of the stimulus pulse made it possible to condition the H-reflexwith either a motor or a sensory stimulus. With both types of conditioning, the H-reflex was in-hibited at conditioning-test intervals of 2-3 msec and was then facilitated, the peak of facilitationoccurring at 5-8 msec with motor conditioning and 6-10 msec with sensory conditioning. The phaseof facilitation was followed by further inhibition. We have concluded (1) that the effects of motorconditioning on the H-reflex result from the discharge of Renshaw cells activated by the antidromicvolley in the motor axons, and (2) that the effects of sensory conditioning (at the times used inthese experiments) are largely due to the activation of Renshaw cells secondary to the discharge ofalpha motoneurones by the conditioning volley.

Renshaw (1946) first described the changes whichoccur in the excitability of spinal motoneuronesas a result of stimulation of ventral roots in thespinal cord of the cat. Since then considerableinformation has accumulated concerning thepharmacology and electrophysiology of the re-current effects of motoneurone discharge. Theseeffects are thought to be due to the synapticexcitation of interneurones (Renshaw cells)which are activated by cholinergic recurrent col-laterals of spinal motoneurones. They exert anipsilateral post-synaptic inhibition upon moto-neurones, interneurones, and also other Ren-shaw cells (Ryall, 1970; Ryall, Piercey, andPolosa, 1971). The inhibition of motoneuroneswhich results from ventral root stimulation isbrought about by the direct action of Renshawcells on motoneurones, while the facilitationwhich can also occur is a result of the inhibitionof inhibitory interneurones-that is, disinhibi-tion (Wilson and Burgess, 1962: Hultborn,Jankowska, and Lindstrom, 1971a, b, c).

1 This work was funded by the Van Cleef Foundation and grant71/308 from the National Health and Medical Research Council ofAustralia.2 Van Cleef Foundation Research Fellow. Address for reprints:Professor J. L. Veale, Department of Human Physiology and Pharma-cology, University of Adelaide, Adelaide, South Australia, 5001,Australia.

674

Experimental work to date has been largelyon the cat and monkey, and there have been noaccounts of Renshaw-like recurrent effects afterthe stimulation of motor axons in man. We havepreviously demonstrated (Veale, Mark, and Rees,1972) that there is a differential sensitivity ofmotor and sensory fibres to stimulus pulse dura-tion. With long duration pulses (1 msec or more)sensory fibres have a lower threshold than motorfibres, whereas with short duration pulses (lessthan 200 tsec) the motor fibres have the lowerthreshold. With this technique, a Hoffman (H)test reflex can be conditioned by stimuli deliveredto motor axons. These will conduct action poten-tials centrally leading to antidromic invasion ofthe spinal cord via the ventral roots. Thus a con-ditioning and test situation comparable withthat used by Renshaw can be investigated. In thisway we have demonstrated both inhibitory andfacilitatory effects in man, and believe these tobe secondary to the excitation of Renshaw cells.Aspects of this work have been reported previ-ously (Veale, Rees, and Mark, 1972).

METHODS

The subjects were 14 normal adults, with ages rang-ing from 20 to 40 years. They lay supine on a couch

Protected by copyright.

on February 13, 2021 by guest.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.36.4.674 on 1 August 1973. D

ownloaded from

http://jnnp.bmj.com/

Renshaw cell activity in man

DIFFERENTIALE -M.G.

FIG. 1. Diagram of arrangement of stitrecording electrodes.

with the relaxed left leg supported undand ankle (Fig. 1). Surface recording ele(applied over triceps surae (gastrocnemiumostly), and a copper sheet, about 3 cmapplied to the knee cap as the stimulaelectrode. The cathodal electrode was a idome about 1 cm in diameter, coveredsoaked cloth, and mounted on an inwhose position could be freely adjusteclamped. This electrode was applied oNterior tibial nerve in the popliteal fossa.skin areas for stimulating and recordinprepared by mild abrasion, washing, anwith methanol. Electrode resistances c10 kQ were obtained and Cambridge el(was employed at all stimulating and rec

Stimuli were generated by two Devicstimulators connected in series. One gwave electrical pulses of 1 msec duraticwas always used for eliciting the test Hother stimulator was modified to prodown to 5 ,usec duration, with a rise tin1 ,usec. This stimulator was used for the cpulse, and generally was set at 50 ,usec. Thvoltage used was 90 V. A Devices digtrolled the timing of the stimuli and the gaveraging device.The potentials were recorded differenti

fled and displayed on a 565 TektronixAveraging of responses was done usiIMnemotron Computer of Average Transmodel 400B), or a Radiation Instrumerment Laboratory (RIDL) analyser syst34-27), and outputs recorded on eitherX-Y plotter or a Heath Servorecorder.trains of pulses that advance the memaveraging device, one sweep of the devi(split into four or five segments withpauses between them. In a single run, thof procedures could be repeated up to 5(result of each procedure being consiste

and averaged. It was considered that more reliableTO ANODE results were obtained in this way, since random

variations could be expected to affect all proceduresin the single run equally.

In previous work (Veale, Mark, and Rees, 1972)we demonstrated that there is a differential sensitivityof motor and sensory fibres to stimulus pulse dura-tion. The popliteal fossa was explored with the stimu-lating electrode to find a spot where test stimuli (1msec duration) yielded an H-reflex without any direct

nulating and motor (M) response, and motor conditioning stimuli(20-50 ,usec duration) yielded a small M responsewith no H-reflex. Typical records are shown in Fig. 2.It was required furthermore that no H-reflex resulted

ler the thighctrodes wereIs and soleussquare, wastting anodalround metalby a saline- 1 msec. 22 voltssulated rodd and thenver the pos-The relevantig were firstid swabbingf less thanectrode jelly /ording sites. 100#sec 62 voltses Mark IVave square- Am)n, and this[-reflex. Theduce pulsesne of about 20 msecsonditioning l _Ie maximumitimer con- N.H.(24)ating of the FIG. 2. Two myograms, each an average of 20 repe-ially, ampli- titions, elicited by stimulation of the posterior tibialla.lly, ampli- nerve. The upper trace shows an H-reflex with no Mns oitheso response, the lower an M response with no H-reflex.sgeiteCAT The duration and strength of the stimuli applied aretients ( Al shown above the stimulus artefacts, which have beenteD (odvel truncated for clarity. Calibrations apply to bothem(model amsa Moseley myograms.By gating

iory of thece could be from motor stimulation during a medium to strongsignificant voluntary contraction of the triceps surae. In some

lis sequence experiments a sensory conditioning stimulus (1 msec0 times, the duration) was used. In this case the stimulus strengthbntly stored was adjusted to elicit a small H-reflex without a

675



http://jnnp.bmj.com

/J N

eurol Neurosurg P


ownloaded from


J. L. Veale and Sandra Rees

direct motor response. Care was taken to ensure thatthe contractions after the test or conditioning stimuliwere occurring in triceps surae, and not in the lateralperonei or the deep toe flexor muscles.When undertaking repeated sequences with test

and conditioning stimuli, the interval between suc-cessive test (H) reflexes was between one and twoseconds. This interval is obviously a compromisebetween having it so short that the H-reflex is extin-guished, and having it so long that the experimentis impossibly tedious for the subject.Two sequences of five stimulating procedures were

commonly employed:1. A test stimulus alone in the first segment, then,

for the remaining four segments, a test stimulus con-ditioned at four different time intervals. These con-ditioning intervals could be negative as well as posi-tive. This sequence was commonly employed whenbuilding up the graphs for motor and sensory con-

IO sec ' 1 sec

I 1 m\/I

ILl

ditioning. Frequently two of the conditioning timeintervals were kept constant throughout many runs,while the remaining two were changed. In this wayone had a check on the overall stability of the results.The test stimulus alone gave the standard againstwhich the conditioned responses were expressed.

2. In this sequence, the first, second, and fourthsegments were the test stimulus alone, and the thirdand fifth were test stimuli conditioned at someselected time interval (the same for both). The firstand second segments gave a check that the stimulusrepetition rate was not significantly altering the sizeof the H-reflex. Each of the conditioned responses(in the third and fifth segments) was preceded andfollowed by a test alone segment. (The first segmentfollows the fifth, of course.) Thus any prolongedafter-effects following the conditioning of a responsecould be discovered. The average of the precedingand following test responses gave the standardagainst which the conditioned responses were ex-

1 2 sec 3 sec

HI ~~H | H

100 msec

Test alone

III/V I F I

I I I

Test, 0 msec Conditioning. 0 msec,

Conditioning, 2 msec, Test. 4 msec

(negative I

conditioning) I

Conditioning. 0 msec

Test, 8 msec

50 Sweeps, motor conditioning L. K. (19)FIG. 3. Averages of myograms with 50 sweeps of the CAT, with each sweep split into four segments,separated by the vertical dashed lines. Each segment of averaging lasts 100 msec, and begins 1 sec afterthe start of the previous segment, as shown by the timings marked at the top of the dashed lines.In the first segment, a test stimulus only was delivered at the start of the segment. In the secondsegment the test stimulus was delivered at the start of the segment and the motor conditioningstimulus 2 msec later (negative conditioning). The direct M response can be seen, as well as the absenceofsignificant effect on the test reflex. In the third andfourth segmsnts the motor stimulus was given afewmilliseconds after the start of the segment, and the test stimulus follows by 4 and 8 msec respectively.The inhibition at + 4 msec andfacilitation at + 8 msec can be clearly seen. Stimulus artefacts have beentruncated for clarity. Calibrations apply to all segments.

IIIr,I

II

I

676



http://jnnp.bmj.com

/J N

eurol Neurosurg P


ownloaded from



0 sec

[1A 11

4m-[ TEST ALONE0I4mV

I .1aB ----- 11- c

80 msec

:1-5 sec 3 sec

CONDITIONED.TEST ALONE + 2 msec

:CONDITIONED.i + 10 msec

:4 5 sec 6 sec

CONDITIONED,TEST ALONE +2 msec

l

CONDITIONED,

+ 10 msec

10 SWEEPS, MOTOR CONDITIONINGD.H. (33)

FIG. 4. Two averages ofmyograms, each with 10 sweeps ofthe RIDL, and each with sweep splitinto five segments, separated by vertical dashed lines. Each segment ofaveraging lasts 80 msec,and begins 15 sec after the start of the previous segment, as shown by the timings marked atthe top ofthe dashed lines. In all segments a test stimulus was delivered 30 msec after the start ofthe segment. In segments three and five it was preceded by a conditioning stimulus, 2 msecbeforehand in A, 10 msec in B. The contrasting effects on the H-reflex can be clearly seen. Stimu-lus artefacts have been truncated for clarity. Calibrations apply to all records.

pressed. Control of the pulses for the various se-quences was aided by a Ledex stepping switch.

Stimulation was begun 10 to 20 seconds beforeaveraging of the responses was started. This elimin-ated some of the early changes in an H-reflex thatoccur after a rest period and improved the stabilityof the results. Subjects were encouraged to relaxfully throughout the recording session of one to twohours, but were not permitted to sleep.

Measurements of the greatest peak-to-peak ampli-tude of the triphasic H-reflexes were made, andconditioned responses expressed as a percentage ofunconditioned responses.

RESULTS

MOTOR CONDITIONING OF H-REFLEX Condition-ing stimuli applied to motor axons will causeaction potentials to travel in both directions awayfrom the stimulating electrode. Those travellingantidromically will enter the spinal cord via theventral roots and will exert effects on the testreflex, which reach the spinal cord in spindleafferent nerve fibres via the dorsal nerve roots.An example of the experimental records,

shown in Fig. 3, illustrates the effects exerted on

677



http://jnnp.bmj.com

/J N

eurol Neurosurg P


ownloaded from



the H-reflex by motor conditioning stimuli ap-plied at -2, + 4, and + 8 msec. The completeresults from four of the 14 experiments are pre-sented graphically in Fig. 6 below. The graphsshow the percentage changes in H-reflex ampli-tude (in terms of the unconditioned reflex re-sponse) with various intervals between motorconditioning and test stimuli.Motor conditioning at 0 msec could not be

investigated since the two stimuli were appliedthrough the same electrode. Inhibition of the H-reflex was evident at 1 msec and a point of maxi-mal inhibition was reached at 2-3 msec. Occa-sionally the H-reflex fell to 10% of the uncon-ditioned response but was more usually 3050%O.This inhibitory period was immediately followedby a return of the response back to near the con-trol level (80-120%) or into a phase of facilita-tion-that is, up to 200-250%. This is clearlyseen in Fig. 4. The peak of facilitation occurredat 5-8 msec. This was followed by further in-hibition which persisted in varying degrees forup to 30 msec, this being the limit of our testingperiod. The actual extent of inhibition and facili-tation could vary quite widely between subjectsand in the same subject on different days (seesubject LK in Fig. 6), although the pattern ofthe response as described remained the same. It isour impression that this variation partly dependsupon the relative and absolute size of the test

and conditioning volleys, although we did nottest this systematically and the relationship iscomplex.The use of negative conditioning times enabled

us to determine whether or not the conditioningand test stimuli were exciting two separate frac-tions of the triceps surae motoneurone pool.The afferent volley giving rise to the test reflexwas initiated 2-10 msec before the conditioningstimulus to the motor axons (negative condition-ing). The test reflex would therefore have en-gaged the spinal motoneurones and would bedescending in the ventral root axons before theantidromic volley had reached the spinal cord.If a reduction in the size of the recorded H-reflexhad occurred it would have indicated that theantidromic motor action potentials evoked bythe conditioning stimulus had collided with theorthodromic motor action potentials elicited bythe test stimulus (occlusion, Fig. 5). As there wasno significant reduction in the H-reflex withnegative conditioning (Figs 3 and 6) there was nocollision in motor axons. Motor (conditioning)and sensory (test) stimuli were therefore excitingseparate portions of the triceps surae moto-neurone pool. This conclusion is consistent withthe visual observation that the test and con-ditioning volleys elicited contractions in differentparts of the triceps surae muscle.

FIG. 5. Diagram to illustratecollision in motor axons withnegative conditioning intervals(motor stimulus later thansensory stimulus). In theexample illustrated, one of themotor axons carrying anorthodromic action potentialthat otherwise would elicit anH-reflex is also carrying an

= antidromic action potentialfrom the motor conditioningstimulus: the H-reflex in thiscase will be halved. If negativeconditioning is without effecton the H-reflex, then there isno commonality between motoraxons carrying action poten-tials for the H-reflex responseand those stimulated by themotor conditioning stimulus.

678



http://jnnp.bmj.com

/J N

eurol Neurosurg P


ownloaded from



MOTOR CONDITIONING

DH (33)

I I 1 I I I I I1

NHt23)

__-

<I t I I I I 1 1 U- I I I I I. Ir

SENSORY CONDITIONING

- TS(34)

_

_ I I I I-_ 2 §I-

_ JV(28)

_1_ I i-5 0 10 20 -10 0 10 20 3011 - 0 10 20 30

CONDITIONING INTERVAL (ms.c)FIG. 6. Four examples (to the left of the double vertical lines) of motor conditioning of the H-reflex,and (to the right) two examples of sensory conditioning. The ordinates are all percentage changes in theH-reflex. The abscissae give the delay in milliseconds between the conditioning and subsequent teststimuli. The graphs shown cover the spread of results obtained.

FIG. 7. Graph (filled circles)ofpercentage change in theconditioned H-reflex at + 2msec conditioning interval,plotted against the strength (involts) of the conditioningmotor stimulus. Also plottedin open circles is the size ofthe direct M wave generatedby the conditioning stimulus,and its magnitude is indicatedby the vertical scale on theright.

LK(21)

awz 3000-

Z 200-0az

U. 1000 _

u, 0-U) 0, 300-Lnz _0a.> 200-CU'0z 100-iF

zo n.

I'I'

LKz (14)

l'11il

120

100

80

a

U) Zz O0~UO 0

ZlY o

oZzO IL

Z $0o E4)tl

-40

-),0

0z

z

C)

m52 56 60 64 68 72 76 80

STIMULUS VOLTAGE

679

I

---i11

i



http://jnnp.bmj.com

/J N

eurol Neurosurg P


ownloaded from



EFFECT OF VARIATION IN STRENGTH OF MOTOR

CONDITIONING STIMULUS In this type of experi-ment a conditioning interval of +2 msec was

selected and the effect of changing the strengthof the motor conditioning stimulus investigated.At this time interval there is significant inhibitionof the H-reflex. As the strength of the condition-ing stimulus was reduced, the extent of the in-hibition of the H-reflex was correspondingly re-

duced, eventually reaching the control level of100%. At this point there was still a significantrecordable M wave. Presumably, some centralsummation is necessary before inhibitory effectsoccur. The results of one such investigation are

shown in Fig. 7. Over the range of 56 to 78 V,the conditioning stimulus elicited only an Mwave whose magnitudes are plotted in the Figure.Strengths of conditioning stimuli immediatelyhigher than 78 V also generated an M wave andno H-reflex, but could generate an H-reflex whenthe subject voluntarily contracted triceps surae.

At such stimulus levels stimulation of type Ia

nerve fibres is obviously occurring. However,below this level of 78 V it is unlikely that anysignificant stimulation of la fibres is taking place.

SENSORY CONDITIONING OF H-REFLEX Condition-ing stimuli applied to sensory fibres will cause

action potentials to travel orthodromically to thespinal cord in the Ia afferent fibres and to exerttheir effects on the test reflex which will followin the same fibres. Results from two of the seven

experiments are presented in Fig. 6.

Sensory conditioning at 0-1 msec could notbe investigated, since the duration of the con-

ditioning stimulus was 1 msec and test and con-

ditioning stimuli were applied through the same

electrode.In all of the experiments, complete extinction

of the H-reflex was evident at 2-3 msec and intwo experiments lasted until 4 msec. Rapid re-

covery of the H-reflex followed this period ofextinction and there was usually a period offacilitation of the response to 200-250%, al-though in two of the experiments the responsereturned to only 65% and 85%. The peak of thisfacilitation, which occurred at 6-10 msec, was

followed quite abruptly by further inhibition ofthe response. In six of the experiments markedsuppression had occurred by about 20-30 msec,

although it occurred as early as 10 msec in oneexperiment.

OTHER ASPECTS Tetanic stinulations In threeexperiments tetanus trains of four motor or foursensory conditioning stimuli were applied, in-stead of the usual single stimulus. With suchtetanic motor conditioning, various levels ofinhibition were evident at test intervals of up to30 msec. With tetanic sensory conditioning theH-reflex was completely extinguished for inter-vals of up to 30 msec. In neither case was post-tetanic facilitation evident.

Absence of inhibition In a few subjects it wasnoted that motor conditioning stimuli failed toproduce initial inhibition of the H-reflex. Onvisual examination of the muscle groups con-tracting in these subjects, it was evident that themotor conditioning stimuli were stimulating (in-advertently) the deep toe flexor muscles ratherthan the gastrocnemius/soleus muscles. Correc-tion of this error by repositioning the electrodeyielded typical results.

Motor conditioning stimuli in peroneal nerve Intwo experiments the motor conditioning stimuliwere applied to the peroneal nerve, while thetest stimuli were applied through a second elec-trode to the posterior tibial nerve, both electrodesbeing positioned over the popliteal fossa. Fortest intervals of up to 30 msec the H-reflex wasusually inhibited to 50-75o of the uncondi-tioned response. Facilitation of the H-reflex didnot occur at any of the test intervals.

Possible interactions at stimulating site It ispossible that the demonstrated effects of motorconditioning arise because the conditioningstimulus has partially depolarized some sensoryfibres. It could then be envisaged that this effectpersists at the stimulating site sufficiently longfor the excitability to be different when the teststimulus is delivered. Thus the change in sizeof the conditioned reflex may reflect change insize of the test volley travelling to the cord. It istechnically difficult to record neurograms fromthe posterior tibial nerve, so we undertook aninvestigation using the ulnar nerve, selected forease of recording. Identical paired stimuli wereapplied at the wrist, and the evoked neurograms

680



http://jnnp.bmj.com

/J N

eurol Neurosurg P


ownloaded from



recorded at the elbow. The time interval betweenthe stimuli was varied, and the effect upon thesecond response observed. No change in theheight of the second response occurred, even

when the separation between stimuli was as

short as 1 msec. It was accordingly consideredunlikely that such a mechanism would be inoperation when stimulating the posterior tibialnerve.

Possible cutaneous stimulation It is possible thatthe demonstrated conditioning effects arise be-cause of stimulation of skin afferent nerve fibresby the conditioning stimulus. This was elimin-ated by using two stimulating electrodes in thepopliteal fossa. One was adjusted to produce a

small test H-reflex. The other was made to pro-duce an easily perceptible stimulation of the skinclose to the test electrode, but no muscular con-

traction (direct or reflex). Using this latter stimu-lus as a conditioning stimulus, it was demon-strated that it was without effect on the testH-reflex.

Possible muscular influences Although the con-

ditioning and test volleys are exciting substan-tially different regions of triceps surae, it may beargued that there is enough commonality for theconditioning effects to be explicable in terms offacilitation of the contractile mechanism itself.Since the direct M wave occurs some 5-6 msec

after the conditioning stimulus is given, and thetest H response some 35 msec or more after thetest stimulus, the region of interest is from 30msec and later. Double stimuli, each elicitingdirect M waves were delivered over a range of10-40 msec separation, and the second (con-ditioned) response compared with the first. Novariations in amplitude occurred.

DISCUSSION

We have concluded (1) that the effects of motorconditioning on the H-reflex result from the dis-charge of Renshaw cells, activated by the anti-dromic volley in the motor axons, and (2) thatthe early excitability changes that follow sensoryconditioning of an H-reflex are largely due to theactivation of Renshaw cells secondary to thedischarge of alpha motoneurones by the con-

ditioning volley.

In order to justify the first conclusion, it isnecessary to exclude several important possi-bilities:

1. The conditioning pulse has stimulatedmotor axons, and caused a direct motor con-traction. This in turn has caused an afferentdischarge responsible for the changes in the testreflex. After delivery of a stimulus to motoraxons in the popliteal fossa, it would take about6 msec for conduction and initiation of the directmotor (M) contraction, and another 6 msecfor the afferent discharge to travel back to thepopliteal fossa, taking a total of approximately12 msec. It is obvious then that such a mechan-ism could not explain any reflex changes thatoccur earlier than this. Our testing intervals weretaken to 30 msec, but there were no suddeneffects occurring in the 12-30 msec range, onlya steady decline from the early peak of facilita-tion.

2. The conditioning M contraction has alteredsome peripheral aspect of the nerve-musclemechanism, rendering it more sensitive to theH volley. This is excluded by the double M ex-periments, and also is most unlikely, since theM and H waves were exciting different regionsof the muscle.

3. The antidromic action potentials have in-vaded the cell bodies of the motoneurones, andinduced complex excitability changes respon-sible for the changes in the test reflex. This iseffectively excluded by the demonstration, usingnegative conditioning, that the conditioning andtest stimuli are exciting different fractions of themotoneurone pool of triceps surae. The anti-dromically activated motoneurones are not thoseexcited by the testing H-reflex.

4. There is a remote possibility of the effectsbeing based upon some dendro-dendritic inter-action. That is, the conditioning volley leads toan antidromic depolarization of the dendrites ofsome of the motoneurones. This gives rise to achange in the excitability of dendrites of othermotoneurones in the pool, explaining the facili-tation phase of the conditioned H-reflex. Nelson(1966) has demonstrated that such an interactiontakes place between motoneurones in the mam-malian spinal cord (cat). If the information givenby Nelson is applicable to humans, then thisinteraction is most unlikely to be the explanationof our results. It required large antidromic vol-

681



http://jnnp.bmj.com

/J N

eurol Neurosurg P


ownloaded from



leys to a major part of a ventral root in order todemonstrate slight facilitation of short duration,about 1 msec. Our conditioning volleys werevery much smaller, and the resulting facilitationlarger and of longer duration. We conclude thatdendrodendritic interaction is not a plausibleexplanation for our phase of facilitation.

5. The possibility of skin afferent fibres beingexcited by the conditioning stimuli and causingthe changes has been excluded.We therefore concluded that the conditioning

motor action potentials invaded the collateralbranches of the motor axons and stimulatedRenshaw cells, whose activity was responsiblefor the subsequent effects.The general pattern of recurrent inhibition and

facilitation resulting from antidromic motor con-ditioning of motoneurones in man is similar tothe pattern found in experimental animals byRenshaw (1946) and later elaborated upon byEccles, Fatt, and Koketsu (1954), Wilson (1959)and others. There was a period of inhibition ofthe motoneurones (as shown in our experimentsby a depression of the H-reflex) at short condi-tioning test intervals followed by a period offacilitation at longer intervals. In both situationsthe motoneurones were found to be maximallyinhibited at a conditioning-test interval of 2-3msec but the peak of facilitation was reachedmuch earlier and was of a shorter duration inman than it was in the experimental animals(Renshaw, 1946; Wilson, 1959). Wilson (1959)did, in fact, mention that the onset of facilitationoccurred at a conditioning-test interval ofapproximately 3 msec and this concurs with ourresults. As similar experiments have not beenperformed on primates, it is impossible to knowwhether differences in the results indicate a truespecies difference or are merely a reflection ofthe different experimental conditions.

It seems most likely that the brief period of in-creased excitability is due to recurrent disinhibi-tion of motoneurones. This would result fromRenshaw cells inhibiting interneurones which aretonically inhibitory to motoneurones. The evi-dence from cats (Hultborn et al., 1971a, b, c)indicates that the inhibitory interneurones in-clude those on the disynaptic inhibitory pathwayfrom the muscle spindles of antagonist muscles.

In considering the second conclusion-that

the early excitability changes that follow sensoryconditioning may be largely due to recurrenteffects-there are a number of factors to con-sider. The early (2-3 msec) suppression and sub-sequent facilitation (5-15 msec) of the H-reflexhave been previously described in man (Paillard,1955; McLeod, 1969). Paillard (1955) ascribedthe early suppression to refractoriness in nervefibres. Without direct monitoring of the con-ditioning and test volleys at these short intervals,one cannot be certain of the constancy of thetesting volley reaching the cord, but it is possiblethat refractoriness of nerve fibres contributes.Furthermore, the absolute and relative refractoryperiods of the muscle may influence the results.The phase of facilitation of the H-reflex is ex-plicable in terms of activation of the subliminalfringe by the conditioning volley as well as bydisinhibition of motoneurones via Renshawcells. However, the pattern of inhibition andfacilitation with sensory conditioning is so com-parable with that obtained with motor condition-ing that we accordingly concluded that recurrenteffects subsequent to activation of the alphamotoneurones with sensory conditioning con-tributed significantly to the observed changes.There has as yet been no anatomical demon-

stration in animals of cells in the spinal cord withthe necessary location and connections to identi-fy them unequivocally as Renshaw cells. Scheibeland Scheibel (1971) have given a thorough ac-count of the difficulties relating to their identifi-cation. They are careful to make it plain thatthey are not disputing the demonstrated electro-physiological and pharmacological effects thatfollow stimulation of ventral roots (see Willis,1971, for a review). This problem of anatomicalidentification is not of direct importance to thepresent investigation.

We acknowledge the considerable contribution ofDr. R. F. Mark, who took a substantial part in theinitial planning and development of this project. Heactively participated in several of the early experi-ments, has contributed much to our discussions ofthis work, and has read the manuscript. We thankMiss Louise Keegan for her help in the laboratory,and for the preparation of the illustrations. We aregrateful to Professor A. K. McIntyre for the use ofthe CAT, and to Dr. W. R. Webster (Department ofPsychology) for the use of the RIDL. We are gratefulto Professor R. Porter for reading the manuscript.

682



http://jnnp.bmj.com

/J N

eurol Neurosurg P


ownloaded from



REFERENCES

Eccles, J. C., Fatt, P., and Koketsu, K. (1954). Cholinergicand inhibitory synapses in a pathway from motor-axoncollaterals to motoneurones. Journal of Physiology, 126,524-562.

Hultborn, H., Jankowska, E., and Lindstr6m, S. (1971a).Recurrent inhibition from motor axon collaterals of trans-mission in the Ia inhibitory pathway to motoneurones.Journal of Physiology, 215, 591-612.

Hultborn, H., Jankowska, E., and Lindstr6m, S. (1971b).Recurrent inhibition of interneurones monosynapticallyactivated from group Ia afferents. Journal of Physiology,215, 613-636.

Hultborn, H., Jankowska, E., and Lindstrbm, S. (1971c).Relative contribution from different nerves to recurrentdepression of Ia IPSPs in motoneurones. Journal ofPhysiology, 215, 637-664.

McLeod, J. G. (1969). H reflex studies in patients with cere-bellar disorders. Journal of Neurology, Neurosurgery, andPsychiatry, 32, 21-27.

Nelson, P. G. (1966). Interaction between spinal motoneu-rones of the cat. Journal of Neurophysiology, 29, 275-287.

Paillard, J. (1955). Reflexes et Regulations d'Origine Proprio-ceptive chez l'Homme. Arnette: Paris.

Renshaw, B. (1946). Central effects of centripetal impulses in

axons of spinal ventral roots. Journal of Neurophysiology,9, 190-204.

Ryall, R. W. (1970). Renshaw cell mediated inhibition ofRenshaw cells: patterns of excitation and inhibition fromimpulses in motor axon collaterals. Journal of Neurophysi-ology, 33, 257-270.

Ryall, R. W., Piercey, M. F., and Polosa, C. (1971). Inter-segmental and intrasegmental distribution of mutual in-hibition of Renshaw cells. Journal of Neurophysiology, 34,700-707.

Scheibel, M. E., and Scheibel, A. B. (1971). Inhibition andthe Renshaw cell. A structural critique. Brain, Behaviorand Evolution, 4, 53-93.

Veale, J. L., Rees, S., and Mark, R. F. (1972). Renshaw cellactivity in normal and spastic man. New Developments inElectromyography and Clinical Neurophysiology, 3, 250-261.

Veale, J. L., Mark, R. F., and Rees, S. (1972). Differentialsensitivity of motor and sensory fibres in human ulnarnerve. Journal of Neurology, Neurosurgery, and Psychiatry,36, 75-86.

Willis, W. D. (1971). The case for the Renshaw cell. Brain,Behavior and Evolution, 4, 5-52.

Wilson, V. J. (1959). Recurrent facilitation of spinal reflexes.Journal of General Physiology, 42, 703-713.

Wilson, V. J., and Burgess, P. R. (1962). Disinhibition in thecat spinal cord. Journal of Neurophysiology, 25, 392-404.

683



http://jnnp.bmj.com

/J N

eurol Neurosurg P


ownloaded from


Documents

Renshaw cell activity in man1 - BMJ · Renshaw(1946)first describedthechangeswhich occur in the excitability ofspinal motoneurones as a result ofstimulation ofventral roots in the