Control and postural thixotropyitself up". Asthe thixotropic effect is present in ana-esthetised and paralysed patients at normal temperatures5 it cannot be neurally determined and

Journal of Neurology, Neurosurgery, and Psychiatry 1986;49:69-76

Control and postural thixotropy of the forearmmuscles: changes caused by coldM LAKIE, EG WALSH, GW WRIGHT

From the Department of Physiology, University Medical School, Edinburgh, Scotland

SUMMARY The forearm was cooled in water at 5-10°C while wrist biodynamics were investigated.Pronounced loosening following a perturbation (thixotropy) was no longer seen. The wrist becamestiffer for large or moderate but not small movements; EMG activity did not increase. Cooling thewrist alone, or opposite forearm, was without effect. The ability to make rapid reciprocatingmovements was reduced and muscle relaxation time was increased. Single movements were notaffected.

It has long been a belief of physiologists that coldincreases muscle tone, and people who are cold oftenexperience a sensation of stiffness. In the classicaldescription of a victim of hypothermia (rectal tem-perature 18°C), Laufman1 observed that the neckmuscles were rigid and the elbows could be flexed onlywith great force. The influence oftemperature on mus-cle contraction and energy metabolism has beenextensively investigated2 3 but there appears to be lessinformation about the degree to which coolingchanges the properties of relaxed muscles and aboutthe mechanisms which underlie these changes. Coldalso has severe deleterious effects on manual per-formance.4

Muscle tone, the resistance of the limbs to passivemovement, is often assessed by manipulation, but amore accurate and revealing investigation is to subjectthe limb to mechanically produced forces. By using atorque motor to oscillate the wrist the stiffness anddamping of the forearm muscles can be estimated.'Postural thixotropy, where the motion produced by arhythmic torque depends on the history ofmovement,can be revealed in this way. We have now investigatedthe behaviour of the wrist when the forearm is cooled.

Methods

The forearm was placed in a metal tank; the wrist and handprotruded through a flanged opening at one end (fig 1).A water-tight seal was provided by a surgical rubber glovewith the fingers removed; the cut end ofthe glove was everted

Address for reprint requests: Dr EG Walsh, Department of Phys-iology, University Medical School, Teviot Place, Edinburgh EH89AG, UK.

Received 5 October 1984 and in revised form 13 March 1985.Accepted 23 March 1985

and fixed over the flange. Water was pumped continuouslythrough the tank from reservoirs at a known temperature;the forearm was completely submerged. The wrist was con-centric with the axis of rotation of the torque motor whichproduced flexion/extension movements of the joint in thehorizontal plane. The motor was attached to the hand by alight crank and Velcro strapping. A G6M4 printed motor(Printed Motors Ltd., Borden, Hants), was modified toreduce friction in a manner similar to that described byMarsden et al;6 it was fitted with a concentric lowfriction/low inertia precision plastic potentiometer to recordthe. movement. Some experiments employed a very low fric-tion basket wound motor (Philips MOlO) with an integraltachometer which signalled angular velocity. The motorswere driven by a class B amplifier which employed currentfeed-back to ensure that the output current (and thus thetorque from the motor) was a faithful reproduction of theinput wave form. The necessary waveforms were generatedby an analogue circuit. The applied torque (from the currentsignal), the displacement (from the potentiometer) and thevelocity (from the tachometer, or obtained by electronic

Fig 1 Apparatus. The arrangement of the cooling waterbath (A), the wrist seal (D), and printed motor (B) isshown. The potentiometer (C) is beneath the motor. Waterenters at (E) and exits at (F). The hand is secured to thehandle with light Velcro strapping. The Philips motor issimilar, but it has an integral tachometer and is not fittedwith a potentiometer.

69

Protected by copyright.

on February 19, 2020 by guest.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.49.1.69 on 1 January 1986. Dow

nloaded from

70 Lakie, Walsh, Wright

Torque

14ftftfftft AAAMAMMAN10.25 Nm

Displacement

IO1 rad

Torque

o 1< IO~~~~~~~~~~~~~~~~025 Nm

.,1 Ileil AARARAAADisplacement0 1 rod

lsFig 2 Thixotropy. The (upper) control records show the large increase in movement that follows a short lasting larger

perturbation of 2, 3 and 4 cycles respectively. The increase is the same in each case; the stiffness has been decreased by

prior movements. Cooling (40 min at 8°C, lower records) reduces the degree of loosening that follows a perturbation but

does not change the stiffness that obtains before the perturbation. The size of the movement caused by the perturbation is

much less following cooling.

differentiation of the position signal) were recorded on a

multichannel rectilinear inkwriter. For the experiments on

thixotropy a bistable was used to control electronic gates so

that one or more cycles of the sinusoidal torque could bereplaced by a larger rectangular force. The circuitry, cali-brating and recording system have been previouslydescribed.5

Other experiments were performed using a "hanginghand" tremorograph.78 The hand was slung from an

induction generator supplied with direct current on one of itstwo windings. This instrument was almost frictionless andhad negligible inertia; the non-energised winding provided a

signal corresponding to acceleration.8For the tracking experiments the hand was coupled to an

identical induction generator and a concentric potentiometerin tandem. Position and acceleration could be recordedsimultaneously, and the movement, which was

extension/flexion in the horizontal plane, was loaded to a

negligible extent. In these experiments a waveform generatorcontrolled one spot on a large screen oscilloscope. A secondspot was controlled by movements of the subject's wrist, theperson being instructed to keep his spot aligned with theother as it moved horizontally. The position of the two spotswas recorded on an inkwriter. The error signal was computedby an operational amplifier and the zero was inserted bymultiplexing at 22 Hz. The resulting trace was a blackenedarea with a height proportional to the mismatch between thetwo spots.

Surface recordings of EMG were made in most of theexperiments. Suction cup EMG electrodes placed over theforearm flexor and extensor muscles worked statisfactorilydespite immersion. Intramuscular fine wires and needle elec-trodes were occasionally used.

In experiments involving electrical stimulation of the fore-arm muscles stimuli (duration 100 ps) at up to 400 volts wereapplied to the motor point of extensor digitorum communisusing suction cup electrodes. A Devices clinical stimulator(Type 3072) was used.

All measurements of temperature of the skin and waterwere made with a multi-channel telethermometer (YellowSprings Instrument Corp) and small surface thermistorprobes. The room temperature was warm, in the range25-30°C. Control observations were made with water at35°C which was the approximate temperature of the skin ofthe forearm measured at the elbow; subjectively it was nei-ther warm nor cold. There was no difference between mea-surements in air and in water at this temperature. The coldwater was normally in the range 5-10°C. Under these condi-tions the experiments were not uncomfortable and the initialsense of chill faded rapidly. More discomfort was caused bycooling a hand in water at a similar temperature; movementsof abduction and adduction of the fingers were then affectedmuch more than those offlexion and extension which dependprimarily on the muscles of the forearm rather than theintrinsic muscles of the hand. There was never any shiveringor general feeling or cold, but the immersed skin developedgooseflesh and redness sharply demarcated at the water level.The skin temperature of the hand was measured and it wasabout 5°C cooler than the control side. To prevent this thehand was sometimes warmed by an infra-red lamp or insu-lated by a glove. There was no obvious numbness, diminu-tion of touch sensitivity or loss of proprioceptive sense in thefingers suggesting that the conduction of the sensory neu-rones passing through the cooled forearm was not seriouslydisturbed. At times temperatures of 1-3'C were used but thisinevitably caused some persistent discomfort.

Protected by copyright.

on February 19, 2020 by guest.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.49.1.69 on 1 January 1986. Dow

nloaded from

Control and postural thixotropy of the forearm muscles: changes caused by cold

Some observations were made during rewarming; effectsthat took over 30 minutes to appear in cold water disap-peared after only 10 minutes rewarming (35°C). On occasionmeasurements were made at one wrist as the other wascooled; no changes were observed. The subjects were volun-teers of both sexes; information about the number, age andsex of the subjects is contained in the results section. Theinformed consent of each subject was obtained.

Results

Changes in thixotropy and stiffnessThirteen experiments were performed on eight sub-jects, six were male. The rmean age ofthe male subjectswas 33, and of the females, 23. In this experiment, asin the ones described later, similar changes wereobserved in each subject.Under certain conditions the motion of a limb

resulting from rhythmic sinusoidal torques of low fre-quency (for example 1-3-5 Hz) depends strikingly onthe past history of movement. We have called thisproperty postural thixotropy.S Thus, if the wrist isdriven by a torque appropriate in frequency (forexample 3 Hz) and amplitude and a transient per-turbation is applied, a large and self-perpetuatingincrease in the amplitude of movement occurs. Theperturbation can be applied by the driving motor, buta voluntary movement of the wrist is equally effective.If the motion is arrested for about 2 s and the rhyth-mic force reapplied the resulting amplitude of move-ment will again be small.5 The amplitude of move-ment when the small rhythmic torque was first appliedwas not altered by cooling, but a perturbation, even iflarge or repeated, had a much reduced effect. Thereduction became noticeable after about 15 min inwater at 5°C, and after 40 min the effect was strikinglydecreased (fig 2). Even ifmany active or passive move-ments were made the system would not fully "loosenitself up". As the thixotropic effect is present in ana-esthetised and paralysed patients at normaltemperatures5 it cannot be neurally determined andthe failure to loosen up following cooling must pre-sumably be due to a direct effect on the muscle itself.During these observations the EMG was silent.

In 18 experiments (four male subjects mean age 22yr, 12 female subjects, mean age 21 yr) the motorapplied low frequency (for example 1 Hz) alternatingtorques to the hand pushing it into flexion and exten-sion in turn. The torque was modulated so that succes-sive cycles became gradually larger, attained a max-imum, and then the programme repeated. In this waythe stiffness of the limb could be monitored as themovement increased. At normal temperatures it wasalways found that the relaxed wrist was dis-proportionately stiff for small movements but becamemore compliant as the size of the movement increasedand eventually started to become stiffer again with the

largest forces as the anatomical limits of the joint wereapproached (fig 3A). The torque and the resultingdisplacement could be recorded on magnetic tape andthen played back at lower speed onto a X-Y plotter.By jointing up the points where the system was sta-tionary a composite compliance curve for the wristcould be obtained. In fig 3B the data from fig 3A hasbeen plotted in this way. The stationary points havenot been jointed together to permit visualisation ofhow the curve is constructed. The three regions: ini-tially very stiff, then compliant and finally becomingstiffer again are apparent. It was always found that thecompliance was much less below a certain criticaltorque size. This corresponded to the increase instiffness that is seen when small sinusoidal torques are

Torque

I Nm

Displacement

I 0.5 rad

I -lnI5s0 orque

-r

*.

<'''"''''''s','-,'..........f' \............

0 5 Nm

0.5 radDisplacement

Fig 3 (A) The ramped rectangular driving torque and theresulting displacement of the wrist are shown. The size ofthe torque increases in linear steps but the displacement isvery non-linear, being disproportionally small when thetorque is low. In (B) the displacement has been plotteddirectly against torque using the X-Y plotter. The locus ofthe stationary points is the composite compliance curve forthe wrist. The disproportionate stiffness for small move-ments is again apparent. This figure also shows (heavy line)the composite compliance curve of the same subject follow-ing cooling of the forearm in water at 10°C for 30 min. Thestiffness for smallforces is unaltered but becomes muchgreater as the forces increase.

71

TC

Protected by copyright.

on February 19, 2020 by guest.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.49.1.69 on 1 January 1986. Dow

nloaded from

72Control Cooled

Velocity

[lOrad s1

0.5 sFig 4 Transients. A briefpulse of current (75 ms)through the motor applied a sharp tap to the hand. Thevelocity was recorded. The resulting transient oscillations ofthe wrist are shown. These are tracings from an oscilloscopeof 6 superimposed taps. Cooling (40 min at 6°C) does notchange the frequency of these transients; thus the stiffnessfor small movements has not altered. Records of acceler-ation (obtained with the tremorograph) were similar.

applied.5 In fig 3B the composite compliance curve forthe same subject following cooling has been super-imposed. In this case to preserve clarity only the curvehas been drawn in and the points from which it wasconstructed are not shown. After cooling, the stiffnessfor small forces was unchanged, but the loosening thatnormally occurred with large forces was less, thus thedisplacement produced by the largest torques was rel-atively smaller. As in the previous experiments theeffect of cooling has not been to increase the stiffnessfor small movements, but to reduce the ability of themuscles to become more compliant with larger move-ments. When larger forces were applied manually inan attempt to overcome the increased stiffness, dis-comfort in the stretched musculature was experienced.When the wrist and hand only were cooled byimmersion in cold water for up to 40 min there was nochange in compliance.

In nine of these experiments the surface EMG fromthe flexor carpi ulnaris and extensor digitorum com-munis was recorded during the above investigations.In most subjects the movements did not cause observ-able discharges in the control periods or when cooled.The method was sensitive enough to detect a smallvoluntary movement of a finger. Some subjectsshowed shortening reactions during the controlperiod, that is, the extensors became active as thehand was pushed into extension by the motor, and theflexors showed activity during flexion. After coolingthese reactions became less prominent or disappeared,either because there was now less movement orbecause the sensitivity of the reflex arc that producedthem may have been decreased by cooling. At no timewas an increase of EMG produced by cooling; thiswas confirmed by integrating the EMG waveforms.Most observations on the stiffness for small forces

were made using the tremorograph (seven experi-ments on four male subjects, mean age 40). With this

Lakie, Walsh, Wright

*instrument the hand was free to flex-or extend in thehorizontal plane. Small taps delivered manually or bya motor driven brush initiated a decrementing seriesof oscillatory transients at about 8 Hz. Similar results

_,were obtained using the Philips motor driven by abrief current pulse; the resultant velocity transientswere recorded by the integral tachometer. Mea-surements made on anaesthetised and paralysedpatients have shown that the stiffness producing thesetransients resides in the muscular system and is notdependent on neuromuscular control.5 In six experi-ments on three male subjects (mean age 46) weinvestigated transients before and after cooling theforearm. The recordings were not altered by cooling(fig 4) and the EMG was silent throughout. As thefrequency of the transients was unchanged, thestiffness of the muscles for these small movementscannot have altered. This finding verifies the resultsdescribed above. The decrement of the transients wassimilar in the cooled and control states indicating thatthe damping of the system had not changed.

Voluntary movements and trackingVoluntary movements were studied in the course ofthe above experiments, but additional informationcame from tracking tasks (eight experiments, five sub-jects, one female, mean age 28) and from intra-muscular recording (two male subjects, mean age 47).During voluntary wrist movements activity could berecorded from several single units by a concentricneedle electrode. As cooling progressed the number ofunits that could be seen decreased and the duration ofthe remainder was increased by about 30%. Thesechanges reversed on rewarming. These results may becompared with those of Stalberg9 who noted a reduc-tion in the propagation velocity of single human mus-cle fibres in biceps brachialis with cooling to 23°C.Below this temperature he could not elicit actionpotentials.

Control CooedVelocity

[5 rad s-1

lsFig 5 Voluntary movements. The subject was instructed tomake reciprocating movements as fast as possible; he wasable to keep the size of the movements nearly constant bymonitoring the amplitude on a large screen oscilloscope. Forlarge movements (top traces) cooling (30 min at 8°C)reduced the rate from 6 Hz to less than I Hz. Smallermovements (bottom trace) were much less affected.

Protected by copyright.

on February 19, 2020 by guest.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.49.1.69 on 1 January 1986. Dow

nloaded from

Control and postural thixotropy of the forearm muscles: changes caused by cold 7

Control Cooled

Target Displacement

Error dL tI 4 r'r'i'd

2sFig 6 Tracking. The movement of the target and subject'sspot are shown, and the difference between them (error).Cooling (35 min at 7°C) does not impair voluntary move-

ments of this type.

Localised cooling causes striking changes in thecontrol of voluntary movements. At normal tem-peratures, alternating flexion and extension move-ments of the wrist with an amplitude of about 0 5 radmay be executed at a frequency of 5 or 6 Hz. Aftercooling the forearm in water at 10°C for 30 min themaximum rate was less than 1 5 Hz. With colder watera rate of 0 5 Hz became a struggle. When cooled, verysmall reciprocating movements could however still bemade with reasonable facility (fig 5). Single move-ments of flexion or extension separated by a few sec-onds could be made almost as well as normal. Wehave called this selective inability to performreciprocating movements "cold adiadokokinesia"''When rapid reciprocating movements were attemptedfollowing cooling considerable voluntary effort was

required for a laboured result, and there was oftenrecruitment of motor activity in other parts of thebody. Direct observation of the contours of the fore-arm revealed that with repeated movements reciprocalaction was not occurring normally, for the flexorswere still tense during the extension phase and vice-versa. The EMG record following cooling showed aconsiderable increase in activity required to generatethese alternating movements, and this was at timewhen fewer motor units remained operational.

Control Cool4

These changes are reflected in records obtained dur-ing tracking tasks. There were two categories of targetmovement. In the first the subject was required tofollow the abrupt jump of the spot from one side ofthe screen to the other. The jumps occurred at unpre-dictable intervals, but they were always separated byat least one second. This timing ensured that there wastime for the muscles to relax before a movement in theopposite direction was required. After prolongedcooling at 9°C there was no deterioration in the per-formance of this task (fig 6). The second target move-ment was a sine wave which progressively increased infrequency. Above about 2 Hz true tracking is nolonger possible because of the time delays which existbetween eye and arm. At normal temperatures thesetime delays limit performance, for similar movementscan be made at about three times this rate when nottracking. However, with practice, movements phaselocked to the target were possible up to about 6 Hz.After moderate cooling it became impossible to followa spot at a frequency of more than 1 Hz, with moreprolonged cooling 0-5 Hz became marginal. The per-formance was not improved by long periods of prac-tice or by limbering up.

It was noted that when the target speed was low (forexample 0 1 Hz) the performance at normal tem-peratures was imperfect as the motion generated wasdiscontinuous. These imperfections were well seen inthe acceleration record; they represented a phys-iological action tremor. After cooling these discon-tinuities were invariably greatly reduced (fig 7). In thissense, cooling might be said to improve tracking per-formance. We ascribe this improvement to the pro-longed twitch time of the motor units so that the forceinput to the wrist became less spiky. A similarimprovement in performance could be obtained atnormal temperatures by attaching to the apparatus avane immersed in treacle.

In four experiments (four male subjects, mean age

led

TargetDisplacement

Ii Nm

Acceleration_"W%-- *,-- --IP -I

Ilrads-1

5sFig 7 Tracking tremor. The discontinuities that are present when tracking (seen in the acceleration record) are muchreduced by cooling. The degree of cooling is the same as for.fig 6.

73

Protected by copyright.

on February 19, 2020 by guest.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.49.1.69 on 1 January 1986. Dow

nloaded from

74

Control Cooled

IlsFig 8 Electrically evoked twitches tracedfrorecorder. The twitches were isotonic against a

levelforce from the motor. Cooling (40 min areduces the size of the twitch; the time to pea)is not much increased, but the relaxation timeand the overshoots that are seen at normal terabolished.

42) the muscles of the forearm were stelectric shocks before and after cooling.1coupled to the printed motor which wasswith a steady current in order to stretchmuscles being stimulated. The resulttwitch was recorded by the potentiometmethods have traditionally been usedtwitch times; the advantage of the isotused here is that the conditions are the s.other experiments. A typical record is shThe duration of the twitch was greatlycooling, and the greatest prolongationrelaxation phase. In contrast, the latenicontraction time were increased only sliwas also some decrease in the size of theovershoots that were seen at normal Iwere abolished.

Discussion

Effectiveness of coolingWe measured the temperature of the watebut not that of the forearm muscles. Thebeen a gradient of temperature from thethe deeper tissues and determinationswould have had little significance. EEdholm" sampled the deep temperatureforearm; with a bath temperature of 12perature in the deepest parts was 18°Cfor two hours. Clarke, et al'2 found t]perature of the forearm muscle at a depwas about 20°C after an immersion of 40at 1°C. Accordingly, the deeper forearrour experiments will not have been near

the circulating water.

Passive propertiesThe stiffness of the forearm can be grea

by localised cooling. Cooling of the wris

Lakie, Walsh, Wrightother forearm is without effect. The increase ofstiffness is not accompanied by an increased EMG;

Disp acement consequently we conclude that the changes producedDtspkacemnent by cooling are a direct effect on the muscle tissue itself.Muscular rigidity has been described as a symptom of

[0.1rod accidental hypothermia"3 and has sometimes beenattributed to overactivity of spinal reflexes.

In our experiments changes in stiffness can beobserved after about 15 min in water at8°C and are

'm a chart well developed after 30 min. The change is not a sim-constant low ple increase in spring stiffness, but is produced by a

50C) non-neurally mediated change in the complex thix-k contraction otropic properties of the muscles. Thus, for smallis prolonged, forces, the stiffness of the limb is not increased bymperatures are cooling; this is shown by the lack of change in the

transients induced by a tap, and the unchangedresponse to small sinusoidal and rectangular torques.

timulated by When larger forces are applied the situation isrhe hand was different. The amount of movement is now much lesssupplied only than it was before cooling and the ability of a per-the extensor turbation to loosen the system is lost. The stiffness ofing isotonic a tissue is inversely related to the elongation that;er. Isometric results from an applied force. In a linear (Hookian)to measure system the same value of stiffness would be obtained

onic method at all force levels. The wrist, however, is very non-ame as in the linear, being considerably stiffer for small movements.own in fig 8. It has been known for some time that relaxed animalincreased by muscle is stiffer for small rather than large move-was in the ments;`4-6 the "preliminary rigidity" of Denny-period and Brown.'4 Our hypothesis' is that when stationary, or

ightly. There with minimal agitation the musculature undergoes a: twitch. The transition from a state resembling a sol to a statetemperatures resembling a gel. With movements of greater size the

system reverts to the sol state and becomes more com-pliant. Thus the muscles will always be stiff abouttheir resting position. Thixotropic effects are due to asol-gel transformation. Weak bonds (between thethick and thin filaments) are broken by agitation and,

r in the bath, once disrupted, take time to reform. The stiffness isre must have therefore dependent on the previous history of move-superficial to ment. It might be expected that cooling would reduceat one site the disruptive effects of Brownian motion on the weak

larcroft and physico-chemical bonds involved and favour the gelof the cooled rather than the sol state. A similar mechanism has3°C the tem- been observed by Edwards et al. in active muscle;3after cooling warming the human quadriceps muscle was shown tohat the tem- increase the metabolic cost of maintaining a con-)th of 15 cm traction and cooling decreased the relaxation rate fol-|min in water lowing a twitch. These authors have suggested that atm muscles in low temperatures the bonding of actin and myosinrly as cold as becomes greater. We believe that it is the inability of

the cooled muscles to loosen up which underlies theseresults and that they can be best explained on the basisof cold induced changes in muscle thixotropy. Thus,

ttly increased the ability to "limber up" is lost when the muscles are;t itself or the cooled.

Protected by copyright.

on February 19, 2020 by guest.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.49.1.69 on 1 January 1986. Dow

nloaded from

Control and postural thixotropy of the forearm muscles: changes caused by cold

The muscle relaxing effect of a soak in a hot bath orother forms of heat treatment may be explicable onthis basis. However, some physiotherapists use topicalcryotherapy in an attempt to reduce the muscularhypertonicity of spasticity. 7 The results of local cool-ing of spastic muscles may be different. This effect maybe due to changes in reflex control. It has been shownthat primary and secondary muscle spindles and ten-don organs fire at reduced rates when cooled andsubjected to tensions greater than that required tosustain a steady discharge.18

Active propertiesIt has been long established that cooling of mam-malian skeletal muscle changes its twitch time.Tuttle"9 working with the cooled human calf musclesunder isotonic conditions, showed that the latencyand contraction time were slightly increased, but therelaxation time was more than doubled. DK Hill'6found that the isometric twitch of the rat soleusbecame remarkably prolonged at low temperatures.The retardation with cooling was principally in therelaxation phase. Cooling the forearm muscles in ourexperiments decreased the twitch size and greatlyincreased the time for relaxation. The inability to relaxpromptly and the decrease in twitch size can explainthe difficulty in making rapid reciprocating move-ments. AV Hill20 suggested that an athlete would runfaster if his muscles were warmed; certainly rapidalternating movements cannot be made if the musclesare cold. There is a report of a swimmer who wasunable to move and had muscular rigidity after 33 minin water at 1 6'C.2' Under conditions where forces actagainst muscles that are imperfectly relaxed greaterthan normal stresses may arise; this may predispose toinjury of muscles or tendons in cold conditions.

Metabolic effectsMuscle tone has long been associated with the pro-duction and regulation of body heat. We have else-where considered resting muscle tone and have foundno evidence that nervous discharges are responsible.'As resting muscle tone results from a passive processonly trivial amounts of heat may be expected to begenerated. Working with the relaxed human forearm,Holling22 estimated that the oxygen consumption wasas low as 007 ml/g/h. If this figure can be applied toall the muscles of the body it represents only about14% of the BMR.

Since the last century there have been reports thatthe metabolic rate increases as the body is cooled andbefore shivering commences.23 This "pre-shiveringhypertonia" has been attributed to a neurally medi-ated increased in muscle tone. There appears to havebeen little interest in this subject since satisfactoryEMG recordings became possible. We have traced

only a note24 and a short paper.2" The present resultssuggest that at least in part any increase in muscle tonethat occurs following cooling may be a direct effect onthe muscle tissues; this is unlikely to have significantmetabolic consequences. Some workers now believethat even in adults cold induced thermogenesis may bedue to brown fat. The phenomenon has been shownto persist in patients following neuromuscular block-ade.26 In infants,27 and no doubt in some adults, acold environment may induce restlessness which mayraise the metabolic rate by intermittent muscularactivity.

The support of the Royal Society and the MedicalResearch Council is gratefully acknowledged.

References

'Laufman H. Profound accidental hypothermia. JAMA195 1;147:1201-12

2Bergh U, Ekblom B. Influence of muscle temperature onmaximal muscle strength and power output in humanskeletal muscles. Acta physiol Scand 1976;107:33-37.

3Edwards RHT, Harris RC, Hultman E, Kaijser L, Koh D,Nordensjo L-O. Effect of temperature on muscleenergy, metabolism and endurance during successiveisotonic contractions, sustained to fatigue, of the quad-riceps muscle in man. J Physiol (Lond)1972;222:335-52.

4Provins KA, Clarke RS. The effect of cold on manualperformance. J Occup Med 1968;2: 169-76.

sLakie M, Walsh EG, Wright GW. Resonance at the wristdemonstrated by the use of a torque motor, an instru-mental analysis ofmuscle tone in man. J Physiol (Lond)1984;353:265-85.

6Marsden CE, Merton PA, Morton HB. Servo action in thehuman thumb. J Physiol (Lond) 1976;257: 1-44.

'Walsh EG, Wright GW. A hanging-hand tremorograph. JPhysiol (Lond) 1982;329:1-2P.

8Lakie M, Scott DB, Walsh EG, Wright GW. Resonance atthe wrist in anaesthetised subjects. J Physiol (Lond)1983;338:32P.

9Stalberg E. Propagation velocity in human muscle fibres insitu. Acta Physiol Scand 1966;70:287-399.

°Lakie M, Walsh EG. Cold adiadokokinesia. J Physiol(Lond) 1982;328:41-42P.

Barcroft H, Edholm OG. The effect of temperature onblood flow and deep temperature in the human forearm.J Physiol (Lond) 1943;102:5-20.

12Clarke RSJ, Helon RF, Lind AR. Vascular reactions ofthe human forearm to cold. Clin Sci 1958;17:165-79.

13 Simpson JA. On the muscular rigidity and hyperreflexiadue to hypothermia in man with observations on- theaccommodation of peripheral nerve. J Neurol Neu-rosurg Psychiatry 1955;18:191-5.

"Denny-Brown DE. On the nature of postural reflexes.Proc R Soc (Biol) 1929;104:252-301.

s Nichols TR, Houk JC. Improvement in linearity and reg-ulation of stiffness that results from actions of the

75

Protected by copyright.

on February 19, 2020 by guest.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.49.1.69 on 1 January 1986. Dow

nloaded from

76stretch reflex. J Neurophysiol 1976;39.1: 119-42.

16 DK. Resting tension and the form of the twitch of ratskeletal muscle at low temperature. J Physiol (Lond)1972;221: 161-7 1.

Knutson E. Topical cryotherapy in spasticity. Scand JRehab Med 1970;2:159-63.

18Eldred E, Lindsley DF, Buchwald JS. The effect ofcoolingon mammalian muscle spindles. Exp Neurol1960;2: 144-57.

9 Tuttle WW. The effects of decreased temperature on theactivity of intact muscle. J Lab Clin Med1941;26:1913-9.

20Hill AV. Trails and Trials in Physiology. 1965 London:Arnold.

21Pugh LGC, Edholm OG. The physiology of channelswimmers. Lancet 1955;2:761-8.

22Holling HE. Observations on the oxygen content ofvenous blood from the arm vein and the oxygen con-

Lakie, Walsh, Wrightsulnption of resting human muscle. Clin Sci1939;4:103-1l1.

23Cannon WB, Querido A, Britton SW, Bright EM. Studiesin the conditions of activity in endocrine glands XXIThe role of adrenal secretion in the chemical control ofbody temperature. Am J Physiol 1927;79:466-507.

Burton AC, Bronk DW. The motor mechanism of shiv-ering and of thermal muscle tone. Am J Physiol1937;119:284.

2SPetajan JH, Williams DD. Behaviour ofsingle motor unitsduring pre-shivering tone and shivering tremor. Am JPhys Med 1972;51:16-22.

26 Jessen K, Rabol A, Winkler K. Total body and splanchnicthermogenesis in curarised man during a short exposureto cold. Acta Anaesthesiol Scand 1980;24:339-44.

27 Hey EN. The relation between environmental temperatureand oxygen consumption in the new-born baby. J Phys-iol (Lond) 1969;200:589-603.

Protected by copyright.

on February 19, 2020 by guest.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.49.1.69 on 1 January 1986. Dow

nloaded from