Journal of Neurology, Neurosurgery, and Psychiatry, 1975, 38, 855-864

Cutaneous afferent activity in median and radialnerve fascicles: a microelectrode study'

DAVID BURKE, RODERICK A. MACKENZIE2,NEVELL F. SKUSE, AND A. KEITH LETHLEAN

From the Division ofNeurology, The Prince Henry Hospital, and the School of Medicine,University ofNew South Wales, Little Bay, N.S. W. 2036, Australia

SYNOPSIS Cutaneous afferent activity was recorded in fascicles of the median and radial nerves ofnormal subjects using percutaneous microelectrodes. Multi-unit fascicular responses were dominatedby activity in large myelinated fibres. Easily tolerated electrical stimuli evoked the full spectrum offast and slow myelinated fibre activity but more intense levels were required to activate unmyelinatedfibres. Correlation of the evoked potentials and the sensations reported by the subject suggested thatfast myelinated fibres mediate tactile sensations. Pricking pain appeared with the activation of slowmyelinated fibres. The only sensations induced by electrical stimuli were tactile or painful.

Sensory fascicles of peripheral nerves containlarge and small myelinated fibres, unmyelinatedafferent fibres, and unmyelinated sympatheticefferent fibres. The different modalities of normalcutaneous sensation may be attributed to varyingpatterns of impulse transmission in afferent fibregroups, and the variable symptoms of peripheralnerve lesions may reasonably be attributed tovarying patterns of fibre involvement. An under-standing of the function of individual fibregroups in sensory nerves is therefore germanenot only to normal sensation but also to patho-logical states.

Routine neurophysiological studies of peri-pheral nerve function are necessarily restrictedto large myelinated fibres. The function of otherfibres must be inferred from clinical data or fromhistopathological findings on biopsy specimens.Using meticulous techniques, potentials fromsmall myelinated fibres can be recorded byneedle macroelectrodes inserted close to therelevant nerve (Buchthal and Rosenfalck, 1966),but even with this method potentials conductedat 17 m/s are the slowest that can be recordedreliably in normal man (Buchthal, 1973). SinceI Supported by the National Health and Medical Research Councilof Australia.2 Commonwealth Postgraduate Scholar and Edwin and Daisy StreetResearch Fellow.(Accepted 13 February 1975.)

855

action potential amplitude is a function of thesquare of fibre diameter, it is not unexpectedthat difficulty is encountered in recording poten-tials from small fibres even with averaging tech-niques, particularly since such potentials aremore dispersed and must be picked up throughthe insulating tissue that surrounds the fascicle.The microelectrode technique described by

Vallbo and Hagbarth (1968) allows recording ofneural responses from within the fascicle, inwhich site fascicular insulation becomes abenefit rather than a handicap. The techniquehas proved suitable in normal subjects for study-ing mass neural responses to physiologicalstimuli (Vallbo and Hagbarth, 1968; Hagbarthet al., 1970) and the evoked compound actionpotential (Hallin and Torebjork, 1973; Torebjorkand Halin, 1973). Even the activity of singlesensory fibre axons may be recorded, be theymyelinated (Hagbarth et al., 1970) or unmyelin-ated (Torebjork, 1974; Torebjork and Hallin,1974).The studies to be reported were undertaken

with the aim of developing the technique as adiagnostic tool. Of necessity, the findingscomplement and amplify the earlier studies fromHagbarth's department. Particular attention hasbeen devoted to the spectrum of conductionvelocity that can be recorded reliably in normal

group.bmj.com on February 11, 2018 - Published by http://jnnp.bmj.com/Downloaded from

David Burke, Roderick A. Mackenzie, Nevell F. Skuse, and A. Keith Lethlean

sensory fascicles and the stimulus levels re-quired to activate these fibres. Attempts havebeen made to identify the fibres responsible forthe sensations evoked by electric stimuli of vary-ing strength. In the accompanying paper(Mackenzie et al., 1975), the technique has beenused to monitor the development of preferentialfibre blocks in order to elucidate the roles ofdifferent fibres in sensations evoked by morephysiological stimuli. A preliminary report ofsome of this work has been presented to theAustralian Association of Neurologists (Burkeet al., 1974).

METHODS

Sensory fascicles of the median nerve were studied in20 experimental sessions in six normal subjects, aged24 to 45 years, the microelectrodes being insertedjust proximal to the wrist. Fascicles of the super-ficial branch of the radial nerve were studied in 12experimental sessions in three of these subjects, themicroelectrodes being inserted 5-6 cm proximal tothe wrist.The experimental technique was essentially similar

to that of Hagbarth et al. (1970). Insulated tungstenmicroelectrodes tapered to a tip of diameter 1-5 [±mwere inserted manually through the skin into therelevant nerve. Penetration of an appropriatefascicle was facilitated by stimulating through themicroelectrode tip at a stimulus level sufficient toproduce paraesthesiae in the distribution of thefascicle. Using the severity of the paraesthesiae as aguide, the electrode was advanced until the voltagenecessary to induce paraesthesiae was minimal,always less than 1.0 V and usually close to 0.5 V. Atthese levels the microelectrode was invariably intra-fascicular. The electrical stimuli were delivered at1/s but at times more precise adjustments could bemade using frequencies of 5-10/s. When the micro-electrode was intrafascicular, minute adjustmentsof position could be made by recording through themicroelectrode and monitoring on a loudspeaker theintensity of the neural activity induced by scrapingthe skin of the fascicular innervation zone.

Recordings were 'monopolar', the indifferent elec-trode being a similar microelectrode with a bared tipof length 2 mm inserted subcutaneously at a trans-verse distance of 2 cm from the active electrode. Twodifferent amplification systems were used. The mostsatisfactory consisted of a preamplifier of fixed gainof 1 000 coupled to an amplifier of variable gain,usually 50, occasionally 100. In some experimentsthe first stage amplifier consisted of a PAR low noiseamplifier model CR-4 coupled to a modified

Tektronix 122 preamplifier, the total gain of thesystem being 33 000 or 330 000. This latter systemproved less satisfactory for short conduction dis-tances due to the recovery time of the initial stage.For both amplification systems the bandwidth(-3 dB) was set at 0.2 or 0.3 to 3 kHz. The relativelyhigh low-frequency filter was found necessary tominimize movement artefact and 50 Hz interference,which sometimes proved a problem when averagingmultiple sweeps and using high gains. Even so, insome experiments an earthed metal shield wasnecessary to eliminate vestiges of 50 Hz interference.The filtering system did not affect the results signifi-cantly, since the shape of individual action potentialswas not a factor under study.The amplified neural activity was monitored on a

loudspeaker and a Tektronix 564 storage oscillo-scope and was recorded on tape (Electrodata Model6300-2). An amplitude discriminator (cf. Hagbarthet al., 1970) was used in some experiments to im-prove signal-to-noise ratio of gross multi-unitresponses to physiological stimuli and additionallythe treated multi-unit neurogram was then integratedusing a R-C low pass filter with a time constant of0.1 s. The evoked compound action potential waselicited using electrical stimuli usually of 0.4 msduration, although in some experiments, particularlyin the radial nerve, the duration of 0.1 ms was pre-ferred because this reduced stimulus artefact forshort conduction distances. Stimuli were delivered bya Devices Stimulus Isolation Unit through a pair ofuninsulated needle electrodes inserted subcutaneouslyinto the richest site in the fascicular innervationzone. The evoked activity was amplified, monitored,recorded on tape, and averaged on a fixed programmeaveraging computer (ND 801 Enhancetron 1024).The amplitude discriminator was not used in studiesof the evoked compound action potential.The stimulus voltage delivered through the elec-

trodes in the fascicular innervation zone wasadjusted until just perceived by the subject-that is,perceptual threshold, Tp. At different stimulusintensities, expressed as multiples of Tp, the re-sponses to 100, occasionally 200, consecutive stimuliwere averaged in order to highlight low amplitudeslowly conducting potentials. Stimulus current wasmonitored in control experiments using a TektronixType P6016 Current Probe and, under the experi-mental conditions, a linear relationship was foundbetween stimulus current and applied voltage.During the recording of A fibre potentials stimuliwere delivered at 1.0/s, but, for C fibre responses, astimulus repetition rate of one every 3 s was used.The slower stimulation rate was found necessarybecause C fibres sometimes failed to respond to thehigher rate and also because more intense stimuli

856

group.bmj.com on February 11, 2018 - Published by http://jnnp.bmj.com/Downloaded from

Cutaneous afferent activity in median and radial nerve fascicles

were usually necessary to activate unmyelinatedfibres. For experiments in which the perceptualcorrelates accompanying repetitive stimulation werestudied, stimuli were delivered at frequencies of 50-500 Hz. In control experiments, it was found thatstimulus current did not change significantly duringbrief trains at these frequencies.Throughout an experiment, frequent checks were

made to ensure that the microelectrode remained inthe same site in the fascicle, and additionally thatperceptual threshold remained constant. It wasconsistently found that perceptual threshold forsingle shocks was lowered after trains of stimuli atfast rates. As a routine the responses to single shocksat the slow standard rates were studied fully beforestimuli at fast rates were given. All experiments wereperformed in an air-conditioned laboratory. Skintemperature was measured using an Ellab ElectronicThermometer and maintained at 32°C using radiantheat when necessary.

In all experiments appropriate antiseptic pre-cautions were taken. Untoward reactions have beenlimited to transient paraesthesiae in the distribution

t. -i~~~~~~2- 1 A

FIG. 1 Multi-unit responses to cutaneous stimuli(median nerve). a. Responses to moderately heavytouch stimuli. Note the 'on' and 'off' responses.b. Responses to scraping the skin with sandpaper.Lower traces illustrate the raw neurogram and upper

traces the 'integrated' neurogram. In a the amplitudediscriminator has been used to eliminate most of thenoise, while in b the neurographic activity is not so

treated. As in subsequent figures the neural responses

have been retouched.

of the penetrated fascicle on percussion over thepenetration site and occasional dysaesthesiae onstroking the fascicular innervation zone. Thesesymptoms commonly developed within 48 hours ofthe experiment and usually lasted a few days, nevermore than two weeks. There have been no permanentsequelae. In most cases six weeks elapsed betweenexperiments on the same nerve but occasionally thiswas reduced to three weeks for the radial nerve.Conduction velocities were calculated by dividing

the distance between stimulating and recordingelectrodes by the latencies of the recorded potentials.The values obtained are approximate only, since inall probability they underestimate the true conduc-tion velocity because of the inaccuracies of surfacemeasurement of fibre length. This is particularly sowhen stimulating small terminal ramifications ofsensory axons. For this reason, the myelinated Afibre spectrum has been divided into fast and slowconducting components, above and below 15-20 m/srespectively. The latter group contains the so-calledA-delta fibres of other authors (Collins et al., 1960;Dyck et al., 1972).

RESULTS

RESPONSES TO PHYSIOLOGICAL STIMULI Afterpenetration of a sensory fascicle, stroking of thefascicular innervation zone evoked a profusemulti-unit discharge accompanied by a character-istic sound on the loudspeaker (Fig. 1). Suchactivity could not be recorded when the micro-electrode was extrafascicular even if it wasimmediately adjacent to the fascicle. In bothmedian and radial nerves the fascicular area wasusually confined to that of a digital nerve, withcorresponding extension onto palm or dorsum ofthe hand. In many experiments the full extent ofthe fascicular area could not be mapped, since itappeared that the microelectrode picked upactivity from only part of the fascicle. Minoradjustments of the microelectrode at timesbrought other parts of the fascicular area intobetter focus. With median nerve fascicles digitalpulp skin usually provided the richest neuralactivity.

Fascicular responses to a range of physio-logical stimuli were tested, but only tactilestimuli produced a consistent response. Themulti-unit neural discharge was most intensewhen an abrasive stimulus was used, particularlyif applied transversely across the dermal ridges.Thermal sensations induced by cold or warm

857

group.bmj.com on February 11, 2018 - Published by http://jnnp.bmj.com/Downloaded from

David Burke, Roderick A. Mackenzie, Nevell F. Skuse, and A. Keith Lethlean

20 Tp

FIG. 2 Myelinated A fibre activit(median nierve). The range of m velin-ated fibre potentials recorded in amedian nerve fascicle with a reasonablyintense stimuluis. Input gain 50 000.Average of 200 sweeps.

60 40 20 10 5rns

metal and pain sensations induced by a pinprick produced no greater discharge than couldbe achieved with a similar tactile stimulus.Cooling produced by the evaporation of etherand warmth, heat and pain induced by radiantheat failed to evoke any consistent discharge.Typically, the multi-unit response to a touch

5S0 Tp

b6040 20I

b

10 5msr

10 0 Tp

30,uV

c

40 20 10m/s

FIG. 3 The effects of fascicular insulation (nmedianinerve). a. The neurographic response to a stimulusvoltage of 5.0 Tp showing a complex potential withcomponents conductinig dowlu to 10 mls. b. The re-

sponse to the same stimulus after withdrawal of themicroelectrode to a site which was just extrafascicular.The response consists of a simple triphasic potential.Average of 200 sweeps at each site.

FIG. 4 Myelinated A fibre activity (radial nerve).a. Five sweeps have beeni superimposed. A few slowpotentials can be discerned but signal to noise ratio ispoor. b. Average of 100 sweeps using input gain of33 000. Slow myelinated potentials are more obvious.c. Average of 100 sweeps using input gain of 330 000.Multiple slow potentials are visible but the voltage ofthe fast potentials exceeds the input levels of theaverager so that their details are lost.

858

a

a)

5

al.

group.bmj.com on February 11, 2018 - Published by http://jnnp.bmj.com/Downloaded from

Cutaneous afferent activity in median and radial nerve fascicles

stimulus contained dynamic and static elements(Fig. 1). The former consisted of distinctive 'on'and 'off' discharges to application and removalof the stimulus. The latter was of lower ampli-tude but became more intense the heavier thetouch.

EVOKED COMPOUND ACTION POTENTIAL Electricalstimuli delivered through needle electrodes in therichest site of the fascicular innervation zone

evoked a polyphasic response, the amplitude andcomplexity of which depended on the stimuluslevel (Fig. 2). Stimuli applied outside the innerva-tion zone produced on a few occasions a smalltriphasic action potential best seen after averag-ing multiple stimuli. The slower and more

dispersed components were never so recorded.In eight experiments the evoked compoundaction potential recorded from within thefascicle was compared with that recorded whenthe microelectrode was withdrawn to a sitewhich was just extrafascicular (Fig. 3). Theintrafascicular response consisted of a complexof many potentials depending on the stimulusintensity, but the potential recorded at theextrafascicular site consisted of a simple tri-phasic wave due to activity in only the fastestmyelinated fibres. Potentials due to more

slowly conducting fibres were lost. When themicroelectrode was just extrafascicular thevoltage that had to be applied through it to pro-duce paraesthesiae in the distribution of thefascicle rose to 2-3 V.With intense stimuli, the large-amplitude early

group of potentials due to synchronized activityfrom fast myelinated fibres was always followedby a variable number of slower potentials due todispersion of activity in slow myelinated fibres.Usually these slow potentials were only poorlyvisible in single sweeps, requiring averaging foraccurate identification (Fig. 4a, b), but this was

not always so. The degree of amplification re-

quired for slow A fibre potentials was sometimesso great that the voltage produced by the fast Afibres exceeded the input limits of the averager-with resultant blocking (Fig. 4b, c).Very slow potentials from unmyelinated fibres

were commonly recorded with intense stimuli,conduction velocities being 0.4-1.4 m/s (Fig. 5).Such potentials were found with ease in radialnerve fascicles, but in the median nerve un-

b

_IL_.-

.8 4 2 0o6m/s

FIG. 5 Unmyelinated C fibre potentials (radialnerve). a. Single sweep. A cluster of potentials withvelocity up to 1.4 m/s. b. Average of 50 consecutivesweeps. Note that the potentials are not identical in aand b because of latency changes and, occasionally,because offailure to activate individual fibres duringrepetitive stimulation. Stimulus was delivered usingsurface electrodes, the stimulus being approximately7-10.0 Tp. Such profuse C fibre discharges were notfound in median nerve fascicles.

myelinated fibres appeared more sparse and werefound, often with difficulty, in only approxi-mately 50°o of fascicular penetrations. However,a systematic search for C fibre potentials bystimulating at multiple sites within the innerva-tion zone was not attempted. As a general rule,when unmyelinated fibre potentials were re-

corded, the microelectrode was in a suitable sitefor recording sympathetic efferent activity.These latter potentials were heard after un-

expected or intense stimuli and produced a

blowing sound audible at a long and variablelatency of 0.5-1.0 s after the stimulus. Theintensity of this sympathetic discharge variedwith the subject's state of preparedness, couldbe elicited by strong stimuli delivered to theopposite limb, and could be accentuated byalterations in the rate and depth of respirationindependent of stimuli. An unexpectedly severestimulus evoked a shower of sympathetic dis-charges.The unmyelinated fibre potentials could have

arisen from antidromically activated sympa-thetic efferent fibres as well as from afferent Cfibres (Hallin and Torebjork, 1973), but no

20puV,L-- - - L,.].Ih L

. IT" or

|-

859

A im, 11 isTV r

group.bmj.com on February 11, 2018 - Published by http://jnnp.bmj.com/Downloaded from

David Burke, Roderick A. Mackenzie, Nevell F. Skuse, and A. Keith Lethblean

attempt was made to distiniguish between thesepossibilities. Potentials time-locked to the stimu-lus with these velocities will be consideredpredominately afferent C fibre activity.The results obtained from stimulation through

needle electrodes were compared with thoseobtained with surface electrodes (a pair of leaddiscs of 1.0 cm diameter). Such stimuli weremore readily tolerated but were felt to be morediffuse and heavy. Presumably because of thelarger fascicular area stimulated with surfaceelectrodes, the evoked neurogram often con-tained more intense neural activity for acomparably painful stimulus. The less discretestimulus proved unsuitable for studies of per-ception. Stimulation of the radial nerve trunkwas attempted in addition to stimulation of theskin of the fascicular innervation zone, butproved unsatisfactory because it resulted in avery large fast myelinated fibre response whichmasked the slower components. Stimulation ofthe nerve trunk was therefore limited to studiesinvolving long conduction distances.

.I*!i

PERCEPTUAL CORRELATES Detailed correlatioinsof stimulus voltage (expressed in multiples ofperceptual threshold, Tp), the resultant neuro-graphic activity and the sensation reported bythe subject were made for the median nerve.Since a brief electric pulse represents a most un-physiological stimulus, the sensations elicited byrepetitive stimulation at frequencies within thephysiological range (50-500 Hz) were alsochecked. Few perceptual differences were notedat the different frequencies in this range.The subthreshold stimulus of 0.75 Tp evoked

a low amplitude response due to activity in fastmyelinated fibres (Fig. 6). This neural activitywas not perceived, despite maximum mentalconcentration and the provision of visual andauditory cues. With repetitive stimulation at thislevel a light fluttering sensation was perceived.When the stimulus voltage was decreased to alevel at which repetitive shocks were just im-perceptible, an evoked compound action poten-tial could not be recorded despite averaging.Thus, although these low threshold potentialswere not perceived when activated in an un-

0 75 Tp

3 0 Tp

1.0 Tp

1I5 Tp

p'50 Tp

tvtrt

2-0 Tp

U

60 40 (020

60 40 20 10

FIG. 6 Evoked compound action potential at stimulusintenisities up to 2.0 Tp (median nerve). Up to 2.0 Tpthe evoked activity lies within the fast myelinated Afibre range. Note the subthreshold activity at 0.75 Tp.Each response is the average of 200 sweeps.

m/S

FIG. 7 Evoked compound action potential at 3.0 Tpand 5.0 Tp (median nerve). Slow myelinated A fibrepotentials can just be made out at 3.0 Tp but are moreobvious at 5.0 Tp. There is simultaneous growth of thefast myelinated A fibre component. Each response isthe average of 200 sweeps.

d-- -'V

1.

860

0.-

.I

group.bmj.com on February 11, 2018 - Published by http://jnnp.bmj.com/Downloaded from

Cutaneous afferent activity in median and radial nerve fascicles

physiological way, repetitive firing reachedconsciousness.

Single stimuli up to twice perceptual thresholdproduced a discrete light tactile sensation usuallyreported as a 'tap', a 'flutter', or a 'pulse'.Repetitive stimulation produced a vibratingsensation which increased in intensity with thestimulus. Pain was never reported. The com-ponents of the evoked compound action poten-tial remained in the fast myelinated fibre range(Fig. 6). At 3.0 Tp, single stimuli were sometimesreported to evoke a pricking sensation, sharp butdefinitely not painful, and at this level slowmyelinated fibre potentials began to appear(Fig. 7). Repetitive stimulation at 3.0 Tp wasinvariably reported as painful. Even at thefastest stimulus frequencies, subjects reportedthat the sensation was always maximal early inthe stimulus train but eased over the ensuingseconds.At 5.0 Tp, single stimuli were felt to be sharp

5 0 Tp

20 10

II

rn/s

S

10 0 Tp

~LEILbIkdi_ vJLJA .a.L..IllF h F Am- -

IIa

20 10M/s

FIG. 8 Evoked compound action potentials at 5.0 Tpand 10.0 Tp (median nerve). Different fascicle fromFigs 6 and 7. A slightly greater slow myelinated fibreresponse is seen at 10.0 Tp. Each response is average of200 sweeps.

and pricking, sometimes mildly painful. Repeti-tive stimulation was more painful than at 3.0 Tpbut was qualitatively similar. Potentials conduct-ing at around 10 m/s were more apparent in theneurogram (Fig. 7). Single stimuli at 7.0 Tp and10.0 Tp were invariably felt as painful prickingsensations of greater severity, and repetitivestimulation was intolerable to all but well-motivated subjects. The sensation was mostintense early in the train but subsequently eased.This stimulus level activated more slow myelin-ated fibres, the slowest potentials in the evokedneurogram conducting at 4-7 m/s (Fig. 8). Inonly two experiments have further increases instimulus voltage resulted in additional com-ponents conducted within the myelinated fibrerange. Thus, stimulus levels of 10.0 Tp areadequate to activate the full spectrum of myelin-ated fibres in cutaneous fascicles, and with a well-positioned microelectrode a full spectrum ofmyelinated fibres has always been recorded.

At 15.0 Tp, single stimuli were reported toproduce a heavy sharp jabbing pain, unlike a pinprick. This sensation increased in severity up to50.0 Tp. An aching lingering 'after-pain' per-sisted between stimuli, particularly at the higherstimulus levels. Over 50.0 Tp, the perceivedsensation reached a plateau. Further increases instimulus voltage produced little change, so thatif 50.0 Tp could be tolerated so could 100.0 Tp.Repetitive stimulation at 20.0 Tp on threeoccasions was intolerable. Afferent C fibrepotentials were recorded with these stimuluslevels in approximately 5000 of fascicular pene-trations (Fig. 9). Such activity usually appearedat 15.0-20.0 Tp but on two occasions C fibrepotentials were recorded at 10.0 Tp. They havenot been recorded in median nerve fascicles atlower stimulus levels.

Less detailed observations on the radial nervewere essentially similar, there being a similarorder of recruitment of myelinated fibres withincreasing stimulus level. Occasionally, a lessintense sensation was reported than expectedfrom the stimulus level and the recorded neuralactivity. In radial nerve fascicles, afferent C fibrepotentials were recorded regularly at 5.0-7.0 Tpand occasionally the sensation reported with thisactivity was sharp but not painful. Repetitivestimulation at these levels always proved painful.

861

5

&ALI&lab& j

group.bmj.com on February 11, 2018 - Published by http://jnnp.bmj.com/Downloaded from

David Burke, Roderick A. Mackenzie, Nevell F. Skuse, and A. Keith Lethleani

Stimulus

20

100 0 Tp

16 II ' -VP -, I''s

to0 0075mrs

05

FIG. 9 Unmyelinated C fibre activity (mediani nerve). A cluster ofpotentials conducting with velocity 0.7-1.2 m/s in response to avery intense stimulus. The averaging sweep was triggered 100 msafter the shock was delivered. Average of 200 sweeps. Comparewith the multiple potentials obtained at a low stimulus level, notrequiring averaging in the radial nerve (Fig. 5).

DISCUSSION

The findings of the present study largely confirmthe earlier observations of Vallbo and Hagbarth(1968), Hagbarth et al. (1970), and Hallin andTorebjork (1973) on the nature of the multiunitresponses recordable in cutaneous fascicles. Amajor purpose of the present study was toreassess the degree of reliability with which thedifferent components of the evoked compoundaction potential could be recorded in normalnerve fascicles and to attempt to correlate fibreactivity with stimulus level as a prelude tostudies in patients with peripheral neuropathy.Potentials from afferent A fibres covering thefull spectrum of fibre size can be recordedreliably with a well-positioned intrafascicularmicroelectrode. The required stimulus levels are

tolerable. Afferent C fibre potentials presentmore of a problem, at least in the median nerve,

because high intensity stimulation is usually re-

quired and C fibres may be located in the fascicleonly after some searching. Thus pain couldprove a limitation to investigations of C fibrefunction in unsophisticated patients. A furtherlimitation to the microelectrode technique isthat quantitative observations on fibre groupscannot be made. The components of the evokedneurogram depend on the site of the micro-electrode within the penetrated fascicle. Evenwith high gains, the recording field of the micro-electrode is limited, particularly for low ampli-tude potentials from small fibres. The field varieswith the size of the fibre action potential, andthus the number and size of the different com-

ponents of the averaged evoked neurogram do

not reflect the absolute numbers or proportionsof different fibres in the fascicle.A good correlation was found between activity

in fibres of different size and the evoked sensa-tion. Similar correlations have been reported byCollins et al. (1960) who recorded from theexposed sural nerve of patients who had beenallowed to wake up fully from general anaes-thesia during the course of a cordotomy opera-tion for intractable pain. Dyck et al. (1972)studied sensation with quantitative methods innormal subjects and in patients with poly-neuropathy, and then related these measure-ments to the evoked compound action potentialobtained in biopsy specimens of sural nerve. Todate the only non-traumatic in vivo humanstudies to relate perception and the differentcomponents of the evoked neurogram have beenthose of Hallin and Torebjork (1973) andTorebjork and Hallin (1973). The primaryemphasis of these papers, however, was to con-trast the function of myelinated and un-myelinated fibres. Relatively little attention wasdevoted to differences within the myelinatedfibre spectrum, and 'no serious attempts weremade to get a precise measure of the stimulusstrength used' (Hallin and Torebjork, 1973). Acorrelation of stimulus level with different fibrethresholds was not made. Repetitive stimulationwas studied only in the radial nerve using lowfrequencies of up to 5 Hz.

That stimuli just below perceptual thresholdmay evoke a low amplitude neural response hasbeen reported previously (Buchthal and Rosen-falck, 1966), although no definite function was

862

aW# liv"Am,,- ,,8-

group.bmj.com on February 11, 2018 - Published by http://jnnp.bmj.com/Downloaded from

Cutaneous afferent activity in median and radial nerve fascicles

attributed to the responsible fibres. Singleelectric shocks represent most unphysiologicalstimuli, but repetitive stimulation probablyparallels normal circumstances a little moreclosely. In the present experiments, a light tactilesensation was reported when such 'subthreshold'stimuli were delivered repetitively. There was noevidence of truly subthreshold neural activity-of potentials which when activated repetitivelyfailed to evoke any sensation.

Tactile sensations appear to be mediated byfast conducting myelinated fibres of low thresh-old. Pricking pain is reported with the recruit-ment of slow myelinated fibres. Hallin andTorebjork (1973) reported that C fibre potentialswere recorded in some experiments at stimuluslevels which evoked the sensation of prickingpain. This was noticed in a couple of the presentexperiments at 10.0 Tp, a stimulus level whichevoked a reasonably strong pricking pain butwhich only activated a few C fibres. It mighthave been noticed more often if a systematicsearch had been made within each fascicle for Cfibres, as was done by Hallin and Torebjork.However their conclusions were also based ondata from the radial nerve, in which C fibresappear to be more abundant and of lowerthreshold. In the present paper, the results ofrepetitive stimulation support the belief thatpricking pain arises from activity in smallmyelinated fibres irrespective of the activity inunmyelinated fibres. When delivered repetitively,a stimulus previously felt to be only mildlypricking evokes a definite pain sensation, thestimulus level being far below threshold for un-myelinated fibres. The severity of pain evoked byrepetitive stimulation at 7.0-10.0 Tp attests tothe potency of this myelinated fibre pain. Aqualitatively and quantitatively different painsensation was reported with single shocks whichconsistently activated unmyelinated fibres. Thustwo different pain sensations may be felt onactivation of small myelinated and unmyelinatedafferent fibres. Which pain is of greater import-ance in normal human experience remains un-certain. The unmyelinated fibre pain is presum-ably responsible for causalgic and burning painsbut, certainly, activation of small myelinatedfibres may be accompanied by a pain of at leastequal severity.

Only tactile and painful sensations have beenevoked by electrical stimulation, despite activa-tion of afferent fibres which under differentconditions produce thermal sensations. Thepresent experimental method must be consideredinadequate for more detailed studies of suchcutaneous sensations because the limitations ofelectrical stimulation require that activation ofhigh threshold afferent fibres be accompanied byactivation of low threshold afferent fibres.

The authors would like to thank Mr and Mrs EdwinStreet for their wholehearted support of this project andProfessor Karl-Erik Hagbarth for help in setting up themicroelectrode technique. They are also grateful to DrDavid Gillies for assistance with equipment and toProfessor James W. Lance for support and criticismthroughout the study. Figures were photographed by theDepartment of Medical Illustration, University of NewSouth Wales.

REFERENCES

Buchthal, F. (1973). Sensory and motor conduction in poly-neuropathies. In New Developments in Electromyographyand Clinical Neurophysiology, vol. 2, pp. 259-271. Editedby J. E. Desmedt. Karger: Basel.

Buchthal, F., and Rosenfalck, A. (1966). Evoked actionpotentials and conduction velocity in human sensorynerves. Brain Research, 3, 1-122.

Burke, D., Skuse, N. F., and Lethlean, A. K. (1974). Sensoryfunction of the median nerve; preliminary studies usingmicro-electrode techniques in man. Proceedings of theAustralian Association of Neurologists, 11, 89-95.

Collins, W. F., Nulsen, F. E., and Randt, C. T. (1960).Relation of peripheral nerve fiber size and sensation inman. Archives of Neurology, 3, 381-385.

Dyck, P. J., Lambert, E. H., and Nichols, P. C. (1972).Quantitative measurement of sensation related to com-

pound action potential and number and sizes of myelinatedand unmyelinated fibers of sural nerve in health, Fried-reich's ataxia, hereditary sensory neuropathy and tabesdorsalis. Handbook of Electroencephalography and ClinicalNeurophysiology, 9, 83-118.

Hagbarth, K.-E., Hongell, A., Hallin, R. G., and Torebjork,H. E. (1970). Afferent impulses in median nerve fasciclesevoked by tactile stimuli of the human hand. BrainResearch, 24, 423-442.

Hallin, R. G., and Torebjork, H. E. (1973). Electricallyinduced A and C fibre responses in intact human skinnerves. Experimental Brain Research, 16, 309-320.

Mackenzie, R. A., Burke, D., Skuse, N. F., and Lethlean,A. K. (1975). Fibre function and perception duringcutaneous nerve block. Journal ofNeurology, Neurosurgery,and Psychiatry, 38, 865-873.

Torebj6rk, H. E. (1974). Afferent C units responding to

863

group.bmj.com on February 11, 2018 - Published by http://jnnp.bmj.com/Downloaded from

David Burke, Roderick A. Mackenzie, Nevell F. Skuse, and A. Keith Lethlean

mechanical, thermal and chemical stimuli in human non-glabrous skin. Acta Physiologica Scandinavica, 92, 374-390.

Torebjork, H. E., and Hallin, R. G. (1973). Perceptualchanges accompanying controlled preferential blocking ofA and C fibre responses in intact human skin nerves.Experimental Brain Research, 16, 321-332.

Torebjork, H. E., and Hallin, R. G. (1974). Identification ofafferent C units in intact human skin nerves. BrainResearch, 67, 387-403.

Vallbo, A. B., and Hagbarth, K.-E. (1968). Activity from skinmechanoreceptors recorded percutaneously in awakehuman subjects. Experimental Neurology, 21, 270-289.

864

group.bmj.com on February 11, 2018 - Published by http://jnnp.bmj.com/Downloaded from

study.fascicles: a microelectrodemedian and radial nerve Cutaneous afferent activity in

LethleanD Burke, R A Mackenzie, N F Skuse and A K

doi: 10.1136/jnnp.38.9.8551975 38: 855-864 J Neurol Neurosurg Psychiatry

http://jnnp.bmj.com/content/38/9/855Updated information and services can be found at:

These include:

serviceEmail alerting

corner of the online article. this article. Sign up in the box at the top right Receive free email alerts when new articles cite

Notes

http://group.bmj.com/group/rights-licensing/permissionsTo request permissions go to:

http://journals.bmj.com/cgi/reprintformTo order reprints go to:

http://group.bmj.com/subscribe/To subscribe to BMJ go to:

group.bmj.com on February 11, 2018 - Published by http://jnnp.bmj.com/Downloaded from