Journal of Neurology, Neurosurgery, and Psychiatry 1982;45:501-506

Hair follicle discrimination dysfunction in multiplesclerosis patientsRICHARD J SCHNEIDER, RONALD BURKE

From the Laboratory ofNeuroscience, Maryland Institute for Emergency Medical Service Systems, Baltimore,Maryland, USA

SUMMARY A method was developed of assessing somatosensory deficits quantitatively using hairfollicle displacement as a stimulus within a psychophysical signal detection task paradigm. Multi-ple sclerosis patients with and without somatosensory disturbances could be differentiated andcompared with normal subjects. This method may distinguish patients with somatosensory dys-function, and dorsal funiculus neuropathology may underlie this distinction.

Concern with the difficulty of assessing sensory per-ception is common to physiology, psychology andneurology. However, the outcome of studies of sen-sory function in these separate but related scienceshave seldom been brought together. Advances inthe neurophysiology of single cells and fibres in thecutaneous sensory system of the body (somatosen-sory electrophysiology) and the psychology ofstimulus-response relationships (psychophysics)may provide answers to questions concerning thedetection of somatosensory dysfunction.The general problem is how to identify and meas-

ure differences in sensory acuity relating tophysiological organisation (for example, somatosen-sory pathway anatomy and function), psychophysi-cal factors (for example, the cognitive significance ofresponses), and neurological pathology relating todisease or injury. For example, both in man andmonkeys, researchers have had difficulty in deter-mining deficits ensuing from lesions to selectedpathways, such as, the dorsal funiculus.-''0Although some researchers were able to show sus-tained sensory deficits in monkeys,' 1-13 a precise,quantifiable method of demonstrating the tradition-ally taught pattern of dysfunction eluded them.Moreover, the attempt to show deficits with simplepassively applied stimuli led to failures which stimu-lated speculation on differing roles for the dorsalfuniculus pathway'4-'6 and to explanations involvingredundancy of spinal sensory pathways."' Mean-while, other evidence was coming to light whichAddress for reprint requests: Dr RJ Schneider, Laboratory ofNeuropsychology, Building 9, Room IN107, National Institute ofMental Health, Bethesda, MD 20205, USA.Received 21 June 1981 and in revised form 13 November 1981Accepted 12 December 1981

could account for these failures and which suggestedfurther investigation that might demonstrate thesesensory losses.'8-20 Thus, we showed thatinformation from hair follicle stimuli wastransmitted to the primary somatosensory cortexuniquely via the dorsal funiculus in Macaca mulattamonkeys.'9 20 21 These studies on the monkeysuggested that neurological deficits related to dorsalcolumn dysfunction would be detected especiallywell by tests of hair follicle displacementdiscrimination. Further, psychophysical methodshad evolved which enabled us to control formisleading results deriving from psychophysicalfactors (for example response bias) in grossneurological examination.2' 23 This study combinedour insights on the transmission of hair follicledisplacement sensation in monkeys withpsychophysical methods for analysis of sensoryacuity. The purpose of this present study has been tocreate an objective, quantifiable and sensitive meansof detecting somatosensory deficits in patients withspinal cord neuropathology. We have investigatedpatients with sensory pathology caused by thedemyelination or plaque formation associated withmultiple sclerosis.

Methods

Multiple sclerosis patients were solicited from a groupreferred to the Department of Hyperbaric Medicine at theMaryland Institute for Emergency Medical Services Sys-tems for experimental treatment with hyperbaric oxygen.All had been diagnosed as having multiple sclerosis by atleast two physicians. Their sensory symptoms were evalu-ated from medical records, from interviews and from aneurological examination prior to the testing sessions. Fivepatients were selected for study. Three of them-males

501

by copyright. on M

ay 9, 2022 by guest. Protected

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.45.6.501 on 1 June 1982. Dow

nloaded from

Schneider, Burke

aged 35, 38 and 48-had sensory symptoms which had notresolved at the time of the study. They all complained ofsensory loss or paraesthesiae-numbness, tingling, temp-erature sensations-both transiently in the past and at thetime of testing. All could walk, but they had difficulty withbalance, motor control and coordination of the legs.Two-a male aged 39 and a female aged 28-had no sen-sory symptoms at the time of the study. The female patienthad experienced sensory paraesthesiae in the past whichhad resolved. The male had never complained of sensorydisturbances, but had a great deal of motor difficulty. Thefour normal control subjects were drawn from the popula-tion of the Institute's employees; they were two femalesand two males aged 28, 32, 26 and 34 years, respectively.Female subjects refrained from shaving their legs for twoweeks prior to testing.By means of electromechanical and logic circuitry a

sequence of events constituting a psychophysical yes/nodiscrimination task was presented to subjects in a mannerconsistent with signal detection theory (TSD).21 The pro-cedure and the associated series of events as applied by usare described below. These events included an alertinglight, a mechnical hair displacement stimulus and a feed-back light. The lights were standard green and white 7-5

OX.

watt bulbs. The mechanical stimulus was delivered by agalvanometer (MFE, model R-4-154) which oscillated at afrequency of 10 Hz. A 1-5 mm long brass rod was attachedat a right angle to the longitudinal axis of the galvanometershaft (fig). This rod contacted and displaced hair follicles1 mm above the skin surface of the subject's leg. Hair folli-cle displacement was at either 9-45 mm (S+) or 6-20 mm(S-) measured at the tip of the oscillating rod. Thus, thedifference in hair displacement to be discriminated at thetip of the rod was 3-25 mm. A blind prevented the subjectfrom viewing the stimulus delivery. A trough restrained legmovements while comfortably supporting the leg of theseated subject (fig). A holder for the galvanometer per-mitted omnidirectional placement while damping vibra-tion; it avoided cues being transmitted to the subject viathe leg restraint trough.The sequence of events was as follows. The central pro-

gramming equipment initiated a discrete trial every eightseconds, the start of the trial being denoted by a greenlamp. Two seconds later, either of the two amplitudes ofhair displacement was presented with equal probability(p(S+) = p(S-) = 0.5) in a quasirandom sequence. Thesubject identified which of the two stimuli was being pre-sented either by pressing or refraining from pressing a

Figure Hair follicle displacement stimulus arrangement. The subject sits with his or her legresting in an adjustable, plexiglass trough. The hair follicle stimulator is held by a vibrationdamping, omnidirectional supporting stand. The stimulator oscillates a brass rod whichdisplaces hair follicles both rostrally and caudally (depending on the phase of the oscillation)during each stimulus presentation. The rod oscillates along the long axis ofthe leg. Holes in thetrough permit access to medial, lateral andposterior dermatomes as well as anterior shown herebeing stimulated. Inset: Close-up ofstimulator and holder.

502

by copyright. on M

ay 9, 2022 by guest. Protected

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.45.6.501 on 1 June 1982. Dow

nloaded from

Hair follicle discrimination dysfunction in multiple sclerosis patients

hand held button. A response when S+ was presented wasdefined as a correct identification. In all cases a stimuluswas presented for a maximum of 2 seconds. If a buttonpress response did not occur during this observation inter-val, the subject was considered to have refrained frompressing. The alerting lamp and the hair follicle displace-ment stimulus (either S+ or S-) were extinguishedimmediately following the button press response or theelapse of the two second observation interval. If a correctidentification was recorded, which could either be pressingthe button when S+ was presented or refraining from pres-sing the button when S- was presented, the white lamp wasilluminated for 0-5 seconds. Incorrect identification did notcause this (feedback) lamp to be illuminated. Thus, thesubject was presented with a series of discrete trial succes-sive discriminations. He identified which stimulus had beenpresented during each discrete trial by either pressing ornot pressing a push button. Correct identifications werefollowed by the illumination of a feedback lamp and incor-rect identifications were not. This procedure is analogousto a yes/no signal detection theory paradigm where thealerting interval is two seconds and the observation andresponse intervals are concurrent and two seconds inlength.The purpose and methods of the experiment were

explained to the subject. The operation of the logic equip-ment and the sequence of stimulus events was demon-strated and explained. The subject's right leg wasrestrained in an elevated position (in our sample the sen-sory disturbance had always been worse on this side) andthe area to be stimulated was exposed. The galvanometerwas positioned and the subject began a series of practicetrials. During initial testing of a subject, practice mightinvolve verbal coaching by the experimenter or allowingthe subject to view the stimuli as they were presented.With repeated testing sessions the subject gained under-standing of the experiment and practice might involve asfew as twenty trials. In all cases practice was continued atleast until the subject expressed the desire to proceed tothe actual experimental task. The results of practice werediscarded. During testing and data collection the gal-vanometer was screened from the subject's view as men-

tioned above, and the experimenter was not in the roomwith the subject. Testing continued until a series of 180discrete trials had been presented. Each subject was testedrepeatedly, various numbers of times at irregular intervals.The results of the experiment were analysed in

accordance with signal detection theory. Research hasdemonstrated the reduced variability inherent independent variables based on the assumptions of signaldetection theory.24 Preliminary research in this laboratoryhas confirmed this result25-27 and demonstrated theindependence of the signal detection theory sensitivitymeasure d' and the associated criterion or bias measurebeta (/3). These results recommend d' as a superiordependent variable to employ in experiments of this type.*

Results

Gross neurological sensory examination of the legsof the multiple sclerosis patients revealed noobvious deficits in two-point discrimination,vibratory sensation (tuning fork at 128 Hz), lighttouch sensation with cotton wool, pin-pricksensation, thermal sensation, proprioception (toemovement) or, in particular, hair displacementsensation produced by bending individual hairs withforceps. After subjective testing, these patients weretested on the objective, psychophysical paradigm.

Following the series of 180 discrete trials, d' was

computed for each subject. For analytic purposes,mean d' (Md') was computed for each subject overall blocks of discrete trials presented to that subject.Each subject's Md' was then based on n x 180discrete trials. The figures in the table present Md'

*As applied here, d' is a quantitative measure of the capacity of thesomatosensory system to distinguish one stimulus from another.Subjects who distinguish two given stimuli more accurately willhave a higher d' than those who are less accurate. If somatosensorycapacity is diminished, stimuli are less accurately distinguished andd' is reduced.

Table

Normal subjectsSubjectSexMd'SDN'

male2-2050-57813

2

male2-5670-4483

Multiple sclerosis patients (with sensory symptoms)Subject 1 2Sex male maleMd' 1-788 1-333SD 0-478 0 700N' 13 12

Multiple sclerosis patients (without sensory symptoms)Subject 1 2Sex female maleMd' 2-550 2-343SD 0 99J7 0-668N' 5 4

3female2-3450-7934

3male1-4140 3574

4female2-0370-2543

All subiects

2-2890-224

23

All subjects

1-5120-24329

All subjects

2-4470-1469

N' represents the number of times a subject was presented a block of trials where each block of trials consisted of 180 discrete trials.

503

by copyright. on M

ay 9, 2022 by guest. Protected

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.45.6.501 on 1 June 1982. Dow

nloaded from

Schneider, Burke

and the standard deviation of d' (SD) for eachsubject and for each group of subjects.The group differentiation, normal, multiple

sclerosis with sensory symptoms and multiplesclerosis without sensory symptoms, was based onthe assumption that a diagnosis of multiple sclerosiswas not fundamental to sensory dysfunction.Rather, a history of sensory dysfunction insomatosensation (for example, paraesthesiae) wasindicative, and therefore no difference in Md' was tobe expected between the normal and multiplesclerosis without sensory symptom groups. Thevalidity of this assumption was evaluated bycomparing Md' between these latter two groupsusing the Sheffe method. The resulting differencewas found to be insignificant (difference = 0-158,degrees of freedom = 2/6, probability <non-significant).Based on the equivalence of the normal and

multiple sclerosis without sensory symptoms groups,these subjects were combined to form a single group(Md' = 2-341, SD = 0*185, n = 6). A confidenceinterval was then computed to estimate the limitswithin which Md' for this combined group could beexpected to fall. The result of this computation gaveus the interval P (2-092 < u < 2.590) = 0.999.Comparison of Md' for the multiple sclerosis withsensory symptoms group, Md' = 1-512, to the rangeof this confidence interval strongly suggests thatthese d' values were sampled from differentpopulations. In other words, the score of the groupof multiple sclerosis patients with sensory symptomslike paraesthesiae in the legs, by beingcomparatively low in sensitivity, falls outside therange of scores which can be expected for thesensory symptom free, combined group. Themagnitude of the effect of the experimentaldifferentiation multiple sclerosis patient versusnormal subject accounts for 80% of the samplevariance and 70% of the population variance.

Discussion

Our goal was to determine whether we couldobjectively evaluate sensory dysfunction in multiplesclerosis patients with psychophysical methods,analysing the data by signal detection theory. Theresults showed that this was so. However, weacknowledge that the small sample size and unequalnumber of subjects (cell frequencies) in each grouplimit the generality of our analysis. But thehomogeneity of variance and the large number ofdiscrete trials presented to each subject suggest thatthese data are highly reliable. The greater thenumber of trials presented to a subject, the smallerthe error of estimating d' for that subject.2'

Concerning the magnitude of the effect, one guideto it is the proportion of variance accounted for bythe experimental differentiation. When that propor-tion is about 30%, it is conventionally considered asmall to medium effect size. In the present case, themagnitude of the effect accounted for more than70% of the variance. An effect this size would betermed strong. If we compare the two groups with-out sensory symptoms-normal subjects and thosewith multiple sclerosis-to the multiple sclerosisgroup with sensory symptoms, we have samples oftwo populations-one without sensory symptoms,the other with them: these were estimated to differby 3-5 standard deviation units. An effect thisstrong, the repeated measures design and the rela-tively small variability in the data should leave littledoubt that a valid difference between these twosamples of subjects was demonstrated.

In the absence of histologically confirmed pathol-ogy, we are assuming that a history of sensory lossand paraesthesiae indicate sensory tractneuropathology. This may or may not be so, but thenormal scores of the multiple sclerosis group with-out sensory symptoms suggest that there is nothingintrinsic about multiple sclerosis which causes lowerscores. Recent studies28 in the monkey indicate thatparaesthesiae may result from anterolateral whitematter damage without involvement of other spinalafferent pathways like the dorsal funiculus.Nevertheless, our patients had no gross diminutionof sensation to pin-prick, warmth or cold, which areall sensations which require an intact spinothalamictract. This may indicate the difficulty in comparingsensations inferred in animals with sensations inhumans. On the other hand, it might be that a highresolution examination of the kinds of sensations inhumans dependent on conduction in the spino-thalamic tract with proper stimuli and psychophysi-cal methods like the ones used here would uncover asensory deficit not apparent on gross examination.Whatever the case, we believe the assumption ofsensory tract neuropathology is valid.The origin of this study was the observation that

the hair follicle projection to the primarysomatosensory cortex in Macaca mulatta was com-pletely eliminated by a dorsal funicular tractotomy.Our later studies with behaving Macaca mulattamonkeys showed that sensory loss on a hair folliclediscrimination task indicated dorsal funiculusdamage.2-27 In humans we have found such resultsare independent of hair density, sex, age and othernon-nervous factors. In humans, we cannot knowthe precise pathways involved nor the extent of con-tribution of each. Multiple sclerosis, of course, pre-sents disseminated neuropathology. We have noreason to believe, however, that the pathways are

504

by copyright. on M

ay 9, 2022 by guest. Protected

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.45.6.501 on 1 June 1982. Dow

nloaded from

Hair follicle discrimination dysfunction in multiple sclerosis patients

markedly different in humans from those inmonkeys. The development of the dorsal funiculuscoincided with the evolution of mammals.29 Theappearance of body hair instead of reptilian scales isa notable morphological change in mammals. It isreasonable to assume that evolving spinal sensorypathways would transmit information from thenewly evolved hair follicles. In monkeys and man,the dorsal funiculus is the dominant, recentlydeveloped, spinal sensory pathway. Consequently,experimental results on monkeys may be more per-tinent to the human nervous system than those onother animals.We believe that failures to document deficits in

man and monkeys following dorsal funicular lesionshave resulted from choices of stimuli transmittedredundantly by several pathways," from not takinginto account such processes as fibre sorting,'82030and from methodological weaknesses in the mannerof applying sensory testing or psychophysics toanimals.3'-33 Some researchers have chosen stimuliwhich are related to neurological tests forsomatosensation. Directional, moving touch wouldseem to mimic the components of graphaesthesia,for example. By the use of such a stimulus, a sus-tained deficit following dorsal funiculus tractotomyin monkeys has been found.'2 Others have chosenstereognostic type tasks wherein the animal seems torequire the dorsal funiculus to make an accurate dis-crimination." Our earlier electrophysiologicalresults'922 show that several stimulus submodalitycomplexes, that is, hair follicle displacement, distalglabrous skin touch and distal proprioception, aretransmitted to the primary somatosensory cortex inMacaca mulatta uniquely via the dorsal funiculus.Distal proprioception is already widely used byneurologists to evaluate-dorsal funiculus dysfunc-tion. We chose another of these types of passivestimulation to assess suspected somatosensory dys-function. When this stimulation is combined with anappropriate sensitive methodological approach toevaluating sensation taken from psychophysics,2' 23we are able to quantify sensory deficits in humans. Itis easier to quantify our task than it would be forgraphaesthetic or stereognostic ones, and suchquantification clearly shows that passive lossescoexist with active ones. We believe psychophysicaltesting with the appropriate stimuli allows us towork out the location (tract) and the extent of aspinal lesion. As such, it may be developed into apowerful tool for neurologists.

We thank Dr RA Cowley for his support of thisstudy and Dr Roy Myers for his assistance in provid-ing a population of multiple sclerosis patients. Sup-

ported by grant number RG-1207-A-1 from theNational Multiple Sclerosis Society.

References

'Rabiner AM, Browder EJ. Concerning conduction oftouch and deep sensibilities through the spinal cord.Trans Am Neurol Ass 1948;73:137.

2 Mettler FA, Liss H. Functional recovery in primatesafter large subtotal spinal cord lesions. J NeuropathExp Neurol 1959;18:509-16.

3Cristiansen J. Neurological observations of macaqueswith spinal cord lesions. Anat Rec 1966;154:330.

4Cook AW, Browder EJ. Functions of posterior columnsin man. Arch Neurol 1966;12:72-9.

Levitt M, Schwartzman R. Spinal sensory tracts andtwo-point tactile sensitivity. Anat Rec 1966;154:436.

6 Vierck CJ Jr. Spinal pathways mediating limb positionsense. Anat Rec 1966;154:436.

Eidelberg E, Kreinick CJ, Langeschied C. On thepossible functional role of afferent pathways in skinsensation. Exp Neurol 1975;47:419-32.

8 Eidelberg E, Rick C. Lack of effect of partial spinal cordsections upon thermal discrimination in the monkey.Appl Neurophysiol 1975;38: 145-52.

Schwartzman RJ, Bogdonoff MD. Behavioral andanatomical analysis of vibration sensitivity. ExpNeurol 1968;20:43-5 1.

Schwartzman RJ, Bogdonoff MD. Proprioception andvibration sensibility discrimination in the absence ofthe posterior columns. Arch Neurol 1969;20:349-53.

"Azulay A, Schwartz AS. The role of the dorsal funiculusof the primate in tactile discrimination. Exp Neurol1975;46:315-32.

12 Vierck CJ Jr. Tactile movement detection and discrimi-nation following dorsal column lesions in monkeys.Exp Brain Res 1974;20:331-46.

Vierck CJ Jr. Proprioceptive deficits after dorsal columnlesions in monkey. In: Kornhuber HH, ed. TheSomatosensory System. Stuttgart: Georg Thieme Ver-lag, 1975:310-7.

' Semmes J. Protopathic and epicritic sensation: A reap-praisal. In: Benton AL, ed. Contributions to ClinicalNeuropsychology. Chicago: Aldine, 1969:142-71.

'5 Wall PD. The sensory and motor role of impulses travel-ing in the dorsal columns towards cerebral cortex.Brain 1970;93:505-24.

16 Wall PD, Noordenbos W. Sensory functions whichremain in man after complete transection of dorsalcolumns. Brain 1977;100:641-53.

7Eidelberg E, Woodbury C. Apparent redundancy in thesomatosensory system in monkeys. Exp Neurol1972;37:573-81.

'8 Whitsel BL, Petruocelli LM, Sapiro G, Ha H. Fiber sort-ing in the fasciculus gracilis of squirrel monkeys. ExpNeurol 1970;29:227-42.

'9 Schneider RJ. The effects of lesions of the posteriorfuniculus of Macaca mulatta. Ph.D. Dissertation, UnivPittsburgh, 1972.

20 Dreyer DA, Schneider RJ, Metz C, Whitsel BL. Dif-ferential contributions of spinal pathways to the body

505

by copyright. on M

ay 9, 2022 by guest. Protected

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.45.6.501 on 1 June 1982. Dow

nloaded from

Schneider, Burke

representation in the postcentral gyrus. JNeurophysiol 1974;37:119-45.

21 Green DM, Swets JA. Signal Detection Theory andPsychophysics. New York: Krieger, 1974.

22 Schneider RJ, Kulics AT, Ducker TB. Proprioceptivepathways of the spinal cord. J Neurol NeurosurgPsychiatry 1977;40:417-33.

23 Egan JP. Signal Detection and ROC Analysis. NewYork: Academic Press, 1975.

24 Pollack I, Hsieh RH. Sampling variability of the areaunder the ROC curve and of d'. Psych Bull'1969;71: 161-73.

25 Schneider RJ, Burke RF. Detection of sub-clinical spinaltract sensory dysfunction. Soc Neurosci Abstr1979;5:729.

26 Schneider RJ, Burke RF. Detection of sensory spinaltract dysfunc.tion with signal detection theory. SocNeurosci Abstr 1980;6:727.

27 Schneider RJ, Burke RF. Psychophysical evaluation ofsomatosensory impairment implicating dorsal (col-umn) funiculus involvement. Soc Neurosci Abstr1981;7:612.

28 Levitt M, Levitt JH. The deafferentation syndrome inmonkeys: dysesthesias of spinal origin. Pain1981;10:129-47.

29 Ariens-Kappers CU, Huber GHC, Crosby EC. TheComparative Anatomy of the Nervous System of theVertebrates Including Man. New York: Macmillan,1936.

30Horch KW, Burgess PR, Whitehorn D. Ascending col-laterals of cutaneous neurons in the fasciculus gracilisof the cat. Brain Res 1976;117:1-17.

31 Blough DS. The study of animal sensory processes byoperant methods. In: Honig WK, ed. OperantBehavior: Areas of Research and Application. NewYork: Appleton-Century-Crofts, 1966:345-79.

32 Stebbins WC. Principles of animal psychophysics. In:Stebbins WC, ed. Animal Psychophysics: The Designand Conduct of Sensory Experiments. New York:Appleton-Century-Crofts, 1970:1-19.

3 Blough DS, Blough P. Animal psychophysics. In: HonigWK and Staddon JER, eds. Handbook of OperantBehavior. New Jersey: Prentice Hall, 1977:514-39.

506

by copyright. on M

ay 9, 2022 by guest. Protected

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.45.6.501 on 1 June 1982. Dow

nloaded from