testingthe gate-control theory ofpainin man - jnnp.bmj.com · of pain, which they called the...

Journal of Neurology, Neurosurgery, and Psychiatry, 1974, 37, 1366-1372

Testing the gate-control theory of pain in manP. W. NATHAN' AND P. RUDGE

From the National Hospital, Queen Square, London

SYNOPSIS According to the gate-control theory of pain, the electrical stimulation of large nervefibres should stop the pain induced when only C fibres are active. This kind of pain was induced bypressure, repeated pinprick, cold and heat in the ischaemic limb. The peripheral nerves were electric-ally stimulated in the same way as is done by patients treating their chronic pain by electricalstimulation. There was no change in the quantity nor the quality of the C fibre pain. In other experi-ments, electrical stimulation of the peripheral nerves induced no change in pain threshold to a heatstimulus when only C fibres were conducting, nor when the whole spectrum of fibres were conducting.Although many experiments have been reported that are consistent with the gate-control theory, theexperiments reported here, and others mentioned, are inconsistent with the theory.

In 1965 Melzack and Wall put forward a theoryof pain, which they called the gate-controltheory. They based this theory upon work pre-viously done by Mendell and Wall (1964), inwhich it was shown that stimulation of smallafferent nerve fibres resulted in a positive dorsalroot potential, whereas stimulation of largeafferent fibres caused a negative dorsal rootpotential. These observations were interpreted asevidence of presynaptic facilitation and pre-synaptic inhibition of the cutaneous input.According to the gate-control theory, the outputof a group of first central transmission (T) cellsin the posterior horn depends upon the balanceof large and small afferent fibre activity of theposterior roots. If this balance is such that thesmall fibre activity predominates, there will bepresynaptic facilitation and a large increase in Tcell activity; this results in pain phenomena (gateopen). The converse applies if large fibre activitypredominates (gate closed). This is a quantita-tive theory of pain; the number of impulses in agiven time reaching the relevant parts of thebrain determines the occurrence of pain. Theimplication of the theory, as Melzack and Wallpointed out, is that exciting the large afferentfibres will cause a reduction of the local input tothe spinal cord and thus pain will be reduced orstopped.1 Member of the external scientific staff of the Medical ResearchCouncil.

1366

Cases have been published by Wall and Sweet(1967) and Sweet and Wepsic (1968) in whichchronic pain was abolished during the period ofelectrical stimulation of the large nerve fibres.Since that time, the technique of stimulatinglarge fibres has been developed for the treatmentof many painful conditions.Although the theory has led to the successful

treatment of chronic pain, this does not neces-sarily mean that it is correct. Certain implica-tions of the theory can be examined in normalsubjects. According to the theory, the excitationof large peripheral nerve fibres should be par-ticularly effective in stopping or greatly reducingthe pain induced when only the non-myelinatedor C fibres are active. It should also be effectivein raising the pain threshold to every form ofnoxious stimulus. The aim of the investigationreported here was to test these aspects of thetheory. The experiments were designed to findout if stimulation of large nerve fibres of peri-pheral nerves of the upper limb with the appara-tus used by patients causes any changes in theperception of induced pain or in the pain thresh-old.

MATERIAL

Seven males between the ages of 28 and 60 yearswere the subjects of these experiments.

guest. Protected by copyright.

on 15 July 2019 byhttp://jnnp.bm

j.com/

J Neurol N

eurosurg Psychiatry: first published as 10.1136/jnnp.37.12.1366 on 1 D

ecember 1974. D

ownloaded from

http://jnnp.bmj.com/

Testing the gate-control theory ofpain in man

METHODS

In most experiments, pain was induced by heatingthe dorsal surface of the blackened terminal phalanxof a finger by means of radiant heat; the source wasa mounted motor-car cigar lighter. The distancebetween the heater and the skin was adjusted so thatthe pain occurred when the terminal phalanx hadbeen heated for five to 12 seconds. This distance waskept constant throughout the experiment. The painthreshold was measured as the length of time ofheating. The subject was given a stopwatch and toldto start it when the stimulus was swivelled intoposition over the blackened digit. He was notallowed to look at the digit being heated, or the stop-watch. When the subject felt pain, he stopped thewatch and the stimulus was moved away. The painstimulus was applied at one minute intervals. At firstwe expected that there would be two end-points, apain perception threshold when warmth becamepain, and a pain tolerance threshold, when the painbecame so severe that the subject involuntarily with-drew his hand. But in fact these thresholds turnedout to be very close.

In some experiments, the stimulus to cause painwas squeezing the terminal phalanges, spraying themwith ethyl chloride, and single and repeated pin-pricks. In most series of experiments, pain wasinduced in the upper limb with the circulationoccluded. The circulation to the distal part of thelimb was blocked by a sphygmomanometer cuff at apressure of 220 mmHg, placed on the lowerarm.

U

0 36

00E 28-

0 20

circulation intact

PRELIMINARY TRIALS ON PAIN THRESHOLD TO RADIANTHEAT As our method of inducing pain was by heat-ing a small area of skin, a preliminary series ofexperiments was carried out to examine the relationbetween the initial skin temperature of the area andthe pain threshold. The time taken to reach painthreshold was recorded at various skin tempera-tures; this was done both with and without occlu-sion of the circulation. The temperature of theblackened skin was taken by a thermocoupleimmediately before each application of the heatstimulus. It was found that the initial temperature ofthe skin of the terminal phalanx affected the painthreshold if the skin temperature was below 29-30CG.In the experiments with the circulation to the limboccluded, the skin temperature always fell below thisfigure when the cuff had been on the limb for about30 minutes. The application of the stimulus once aminute always raised the skin temperature. Anexample is shown in Fig. 1. With the circulation tothe limb intact, the temperature of the skin remainedconstant around* 32°C. There was no consistentchange in the pain threshold. With the circulationoccluded, the temperature of the skin rose pro-gressively with repeated application of the heatstimulus. It rose from 25°C to 32°C with the firstfour applications of the heat stimulus, and thenremained constant. For the first three applications ofthe stimulus, the pain threshold fell; after this, therewas no consistent change. For this reason, no elec-trical stimulation was performed until the skintemperature had reached a constant plateau and twoconsecutive threshold readings were similar.

circulation occluded

I I a- II_~

10 -U,

Q 00-6&%-O

0 2 4 6 8 10 0 2 4 6 8 10t i m e in minutes

FIG. 1. Diagram showing relationship between skin temperature of terminal phalanx andpain threshold with aradiant heat stimulus. On the left, the circulation to the limb is unimpeded; on the right, the circulation to thelimb has been occludedfor 20 minutes. Temperature °C (0) indicated on vertical axis on left; pain threshold inseconds (0) indicated on vertical axis on right. Horizontal scale, time_in minutes.

1367

I 1-



j.com/

J Neurol N


ecember 1974. D

ownloaded from


P. W. Nathan and P. Rudge

METHOD OF ELECTRICAL STIMULATION Electricalstimulation of the peripheral nerves was carried outwith a Stimtech EPC Stimulator; this is the portablestimulator issued to patients with chronic pain. Thisstimulator is designed to give a constant current out-put of 0-50 mA for an electrode impedance into 0 to5,000 D. It delivered modified rectangular pulses, theduration of which could be varied between 50 and500 microseconds at a frequency between 15 and 180pulses/second. An analysis of the output showed thatit was linear over the range used in these experi-ments. The electrodes used were those supplied withthe apparatus. They were two discs 4 5 cm diameter;the surface was Gelfoam and the backing was stain-less steel, mounted in a plastic framework. Thedampened electrodes were placed firmly on the skinof the upper arm so that they stimulated the medianor the ulnar nerve. They were proximal to thesphygmomanometer cuff in experiments in whichthe circulation was occluded. Voltage and pulsefrequency were adjusted so that the subject feltdefinite but not painful paraesthesiae in the littlefinger in the case of the ulnar nerve, and in the middleor index fingers in the case of the median nerve. Boththe electrodes and the stimulation parameters wereadjusted so that they did not cause local painbeneath the electrodes. It was emphasized to the sub-jects that these paraesthesiae were not to be painful,but should be the same as patients would be likelyto experience when they treated themselves by trans-cutaneous electrical stimulation. In some experi-ments, the voltage was increased so that the electricalstimulation was painful. The electrodes were fixedfirmly by strapping and a crepe bandage. Throughoutan experiment, care was taken to see that the amountof paraesthesiae due to electrical stimulation was thesame as it had been during the control period. Minoradjustments of voltage, usually a decrease, weresometimes needed to obtain this.

PROCEDURE The electrodes were placed on the sub-ject's arm and the parameters needed for both non-painful and painful paraesthesiae were noted. Thepain threshold without electrical stimulation of thenerves was then determined at intervals of oneminute; 10 observations were made. After this, theeffect of electrical stimulation on the pain thresholdwas determined. The procedure consisted of givingthe heat stimulus alternately with and withoutelectrical stimulation. This procedure was followedthroughout, in order to eliminate any possible driftof threshold. The electrical stimulation of thenerves was applied just before the heat stimulus,and removed after the pain threshold had beenobtained.

RESULTS

FIRST SERIES In the first series of experiments,the effect of electrical stimulation of large nervefibres on the kind of pain experienced when onlyC fibres are conducting was examined. For thesake of convenience this will be referred to as Cfibre pain. After the cuff had been on the upperlimb for 20 to 25 minutes, the sensibility of theends of the fingers was examined to find out ifC fibre pain was present. When this point hadbeen reached, this pain was induced by squeez-ing, repeated pinpricks, radiant heat, and ethylchloride spray. In all five subjects, electricalstimulation of the relevant nerves in the arm, at astrength that caused non-painful paraesthesiae,had no effect on C fibre pain.As the amount of inhibition said to occur

from excitation of the large afferent fibres mightbe insufficient to have an effect on the perceptionof the pain, this stimulation was increased sothat it became painful. The result differed amongthe subjects. In two subjects, the induced painwas abolished; in one, the pain was lessened; intwo, it had no effect.As an increase in the intensity of electrical

stimulation above that used by patients appearedto have an effect on C fibre pain in some sub-jects, it did appear that the relationship betweenelectrical stimulation and the pain might well bequantitative. It was therefore necessary to havea measure of the amount of pain. In all subse-quent series of experiments, pain was induced bythe radiant heat stimulus, and the effect ofelectrical stimulation on the pain threshold wasstudied.

SECOND SERIES As electrical stimulation iseffective in preventing pain in patients, many ofwhom have nerve fibres of all groups conducting,a series of experiments was performed without acuff blocking the circulation to the limb. In thisseries, pain was induced by the radiant heatstimulus, the duration of heating being a functionof pain threshold.

In none of the five subjects did an electricalstimulation that caused non-painful paraes-thesiae have any effect on the threshold of paininduced by radiant heat. When the intensity ofelectrical stimulation was increased to a painfuldegree, in three subjects there was again nochange in the pain threshold. In the other two

1368



j.com/

J Neurol N


ecember 1974. D

ownloaded from



A 4

29 31 33

B

35

FIG. 2. Pain threshold in limbwith circulation occluded, withand without electrical stimula-tion. In A, to left of thearrow, intensity of electricalstimulation caused comfortableparaesthesiae. In B, to right ofarrow, electrical stimulationwas painful. Black columns,with electrical stimulation;white columns, no electricalstimulation. Vertical axis, painthreshold in seconds; horizontalaxis, duration of occlusion ofcirculation in minutes. No

m, observation was made at 30537 minutes.

of ischaemia (min )

subjects, there was a significant increase in thethreshold.

THIRD SERIES As the gate theory would predictthat the electrical stimulation of large nerve

fibres would reduce C fibre pain, and as this wasnot found in the first series of experiments, itwas important to find out if stimulation never-

theless changes its threshold. In this series ofexperiments, pain was induced by the radiantheat stimulus in the ischaemic limb.

Electrical stimulation both at non-painful andpainful intensities had no effect on the painthreshold in all five subjects. The results of a

typical experiment are shown in Fig. 2. Figure2, A, shows the results with non-painful electricalstimulation of the nerve, and Fig. 2, B-to theright of the arrow-shows those with painfulelectrical stimulation. The black columns showpain threshold during electrical stimulation andthe white columns show threshold withoutstimulation. It is clear that the thresholds are

similar with and without stimulation. It will alsobe seen in this example how the first pain thresh-olds are slightly longer than all subsequent ones.

As was mentioned in the Methods sections, thisis due to the fingers having a lower temperaturebefore any heat has been applied. But as the pain

thresholds were always determined in alternatingpairs with and without electrical stimulation, anychanges that occurred in the background tem-perature of the skin would not have obscuredpossible effects of electrical stimulation.

In no experiments did electrical stimulation ofthe opposite limb have any effect on painmeasured in the other limb.An incidental finding in all series of experi-

ments was that the voltage of stimulation mighthave to be reduced for the subject to feel thesame amount of paraesthesiae later in the experi-ment as he had done at the start of the experi-ment; the voltage never had to be raised.

DISCUSSION

The gate-control theory of pain supposes thatthe effect of any cutaneous input to the spinalcord is determined by the preceding balance ofthe activity of large afferent nerve fibres on theone hand and small afferent fibres on the other.The theory has been used by Melzack et al.(1963), by Sullivan (1968), by Higgins et al.(1971), and by Satran and Goldstein (1973) toaccount for the results they obtained on theperception of pain when it was associated withvarious forms of concurrent or preceding

8-

6

4-

2-

0

27

duration

II-.L II _u

1369



j.com/

J Neurol N


ecember 1974. D

ownloaded from



stimulation. The theory has also been used toexplain the pain and hyperaesthesia of manyforms of neuropathy, in which there may beselective damage to the large afferent fibres.Electrical stimulation of the large afferent fibre isnow used for treating many painful conditions;this is one of the results of the gate-controltheory of pain. But although this form of treat-ment is helpful in a minority of patients withchronic pain, this theory is not the only possibleexplanation of its effectiveness.

According to the theory, the electricalstimulation of large afferent fibres carried outjust as patients do for relieving their pain shouldhave abolished or diminished the pain producedby a variety of stimuli in the ischaemic limb,when only the small nerve fibres are conducting.It did not do so. Nor did it reduce the pain whenall nerve fibres were conducting. It might be con-sidered that the pain caused by the radiant heatstimulus was so intense that any small change inthreshold would not be detected. This explana-tion is unlikely to be correct for two reasons.First, it was sometimes seen that if a subjectturned his attention away from the stimulation ofhis finger, the threshold for pain went up con-siderably; yet it never went up with electricalstimulation of the nerves. Secondly, electricalstimulation had no effect on the quantity andquality of C fibre pain induced by stimuli otherthan radiant heat. We conclude that electricalstimulation as used by patients has no effect oncutaneous pain induced in normal subjects.The question one is asking is not whether

inputs arriving in afferent fibres of various sizesconverge at some place in the central nervoussystem; it is whether the diminution or oblitera-tion of pain caused by electrical stimulation oflarge afferent fibres is due to the mechanismoccurring at the entrance to the spinal cord pro-posed by Melzack and Wall (1965).To disprove the gate theory, it is necessary to

demonstrate facts that are inconsistent with it.The facts presented here-that large fibrestimulation did not alter the quantity or qualityof radiant heat skin pain-are inconsistent withthe theory. If the gate theory merely proposedthat there is an interaction between the effects ofarriving impulses, all will accept it. There is aconvergence of afferent nerve impulses at manyplaces in the central nervous system; perhaps the

first place is in lamina v, as has been shown byWall and many other physiologists. There aremany facts that fit this conception better thanthe gate hypothesis. For instance, one of theauthors (P.W.N.) found that electrical stimula-tion in patients with post-herpetic neuralgiaraised the threshold to the tactile sensationcaused by von Frey hairs; and the thresholdremained raised for minutes after stoppingelectrical stimulation. This is not accounted forby the gate theory; it is a different phenomenonfrom the kind of interaction proposed in thattheory. There are the experiments reported bySatran and Goldstein (1973). In these experi-ments the increase in pain tolerance lasted for 20minutes after stopping electrical stimulation;this is surely not the kind of presynaptic inhibi-tion envisaged in the gate theory. Contralateralelectrical stimulation was also effective in raisingthe threshold; this was also not envisaged by thegate theory.

It is our opinion that deep pain is differentfrom the rapidly rising, severe pain we inducedin the digits. Because the word pain is used forthe aching pain induced by certain stimuli to thefascia, tendons, and muscles, for the burningpain induced in the skin, for referred pain andfor various pains due to lesions of neuralstructures, one is tempted to believe that theyall have the same mechanism and make use ofthe same neurones and neural pathways. Weprefer not to assume this. Indeed we do find thatthe same electrical stimulation as we have usedhere reduces and tends to obliterate the paininduced by strong squeezing ofthe tendo Achillis.

Other explanations of the effects of non-painful stimulation of large afferent fibres on thetreatment of painful states in patients might bebased on the phenomena of attention and dis-traction. In our experiments, when the electricalstimulation itself was painful, subjects agreedthat what they felt was related to attention. Onintrospection, one has the feeling that one candirect attention either to the primary painfulstimulus or to the painful electrical stimulation.If the electrical stimulation is on one upper limband the painful stimulus is on the other, we foundthat it is easy to direct ones attention to eitherthe one or the other. It is more difficult, but notimpossible, to concentrate on either the painfulstimulus or the peripheral electrical stimulus

1370



j.com/

J Neurol N


ecember 1974. D

ownloaded from



when they are both in the same nerve territory.Similarly, as has been mentioned, if a subject hadhis attention distracted, the threshold wasraised. This shows that attention is important inthe set-up of these experiments and also that ithad more effect on pain threshold than had theelectrical stimulation.These experiments have shown that what the

gate-control theory forecast was wrong; andthat the successful treatment of pain by electricalstimulation does not work on account of thepremises of the gate-control theory, althoughthese premises were the instigation and therationale of this form of therapy. Furthermore,the experiments are one more example of howlaboratory pain and pain in clinical medicine aretwo different things. It is the merit of the gate-control theory that it has once more awokeninterest in the interplay of two or more inputs tothe central nervous system and in phenomenasuch as counter-stimulation and referred pain.The gate-control theory gives a special role to

the ongoing activity ofsome afferent nerve fibres.It requires that certain afferent fibres are con-tinuously active and that their activity keeps the'T-cells' adequately facilitated and ready tosend massive volleys of impulses to the brain.These volleys are believed to cause pain and thereactions associated with pain. Some clinicalevidence that this ongoing activity exists wasrecently put forward by Nathan et al. (1973).They had electrodes on peripheral nerves,proximal to a cuff obliterating the circulation, ina set-up similar to that reported in the presentwork. They showed that it was necessary toincrease the strength of the electrical stimulationgradually to obtain a constant amount of par-aesthesiae during the production of the ischaemicblock. From this, they concluded that the on-going activity in the peripheral nerves had beenremoved by the ischaemia, and that this was thereason that the amount of electrical stimulationhad to be increased. In the present series ofexperiments, this was not found; indeed, therewas a tendency for the strength to require reduc-tion. This difference can be explained by thelarger electrodes (45 mm) used in these experi-ments compared with the small (7 5 mm) onesused before. With large electrodes, the currentapplied to the nerves is less dependent on theposition of the electrodes than with small ones.

Since in the previous work the small electrodeswere placed in the optimum position for par-aesthesiae, any change in position induced by thecuff would have resulted in an increase in thecurrent needed to obtain the same amount ofparaesthesiae. In the present experiments withlarge electrodes, such movements were of lessimportance; and in these experiments no changesin the voltage were needed. If the explanation iscorrect it appears that the previous evidence ofNathan et al. (1973) is invalid and was an artefactof the method. The same method of stimulating alarge peripheral nerve proximal to an occludingcuff was used in the present experiments; nochange was needed in the intensity of electricalstimulation to maintain a constant amount ofparaesthesiae; and so no evidence for the occur-rence of ongoing activity has been obtained.The gate-control theory has resulted in the

treatment of many chronic painful conditions bycontinuous electrical stimulation of large afferentfibres and in about one-third of cases this form oftreatment of chronic pain is successful. In manypatients in whom it is successful, the pain is lessfor an hour or more after stimulation of theperipheral nerves has stopped; and in somepatients, the whole course of the painful con-dition is greatly improved. These beneficial re-sults are not readily explained on the gate-control theory. It also appears from the experi-ments reported here on normal subjects thatcertain essential parts of the theory are wrong.Needless to say, that in no way lessens the valueof electrical stimulation of large nerve fibres forthe treatment of chronic pain.

REFERENCES

Higgins, J. D., Tursky, B., and Schwartz, G. E. (1971).Shock-elicited pain and its reduction by concurrent tactilestimulation. Science, 172, 866-867.

Melzack, R., and Wall, P. D. (1965). Pain mechanisms: a newtheory. Science, 150, 971-979.

Melzack, R., Wall, P. D., and Weisz, A. Z. (1963). Maskingand metacontrast phenomena in the skin sensory system.Experimental Neurology, 8, 35-46.

Mendell, L. M., and Wall, P. D. (1964). Presynaptic hyper-polarization: a role for fine afferent fibres. Journal ofPhysiology, 172, 274-294.

Nathan, P. W., Noordenbos, W., and Wall, P. D. (1973).Ongoing activity in peripheral nerve: interactions betweenelectrical stimulation and ongoing activity. ExperimentalNeutrology, 38, 90-98.

Satran, R., and Goldstein, M. N. (1973). Pain perception:

1371



j.com/

J Neurol N


ecember 1974. D

ownloaded from



modification of threshold of intolerance and corticalpotentials by cutaneous stimulation. Science, 180, 1201-1202.

Sullivan, R. (1968). Effect of different frequencies of vibra-tion on pain-threshold detection. Experimental Neurology,20, 135-142.

Sweet, W. H., and Wepsic, J. G. (1968). Treatment ofchronic pain by stimulation of fibers of primary afferentneuron. Transactions of the American Neurological Associa-tion, 93, 103-107.

Wall, P. D., and Sweet, W. H. (1967). Temporary abolition ofpain in man. Science, 155, 108-109.

1372



j.com/

J Neurol N


ecember 1974. D

ownloaded from


testingthe gate-control theory ofpainin man - jnnp.bmj.com · of pain, which they called the...

Documents