Visual - BMJ

Journal of Neurology, Neurosurgery, and PsychiatrY 1988;51:934-943

Visual and mechanical control of postural and kinetictremor in cerebellar system disordersJEROME N SANES,* PETER A LEWITTt KARL-HEINZ MAURITZ$

From the Human Motor Control Section, Medical Neurology Branch, National Institute ol Neurological anctCommunicative Disorders and Stroke, National Institutes of Health, Bethesda,* Lola vette Clinic, Detroit,USAt and Rehabilitationszentrum, der Universitdt zu Koln, Federal Republic of German

SUMMARY The influence of alterations in visual guidance and somaesthetic sensory inputs wasstudied in five patients with kinetic and postural tremor characteristic of cerebellar impairment. Thepatients performed wrist flexion-extension movements or movements about the shoulder with orwithout visual guidance. Different types of mechanical loads were applied to oppose the wristmovements. The tremor was greatest when the patients used visual cues to guide movements.Kinetic tremor was substantially suppressed during performance of similar movements that werenot guided directly by vision. Viscous loads suppressed the tremor nearly linearly, whereas constantloads opposing extension enhanced the tremor. The postural tremor was not observed during iso-metric contractions. These results support the view that processing of visual information con-tributes to the impairment of movement in disorders with cerebellar-type tremor and that certainsomaesthetic inputs can selectively influence the generation of postural tremor.

The cerebellum receives an abundant influx of somaticsensory and visual information.'`3 Although, cere-bellar dysfunction is not typically accompanied byimpairments in basic discrimination tests, 4 5 deficitsdue to cerebellar disease are generally evident whenvisual and somaesthetic sensory stimuli need to beintegrated into involuntary or voluntary movements.For example, long-latency stretch reflexes areenhanced in certain types of cerebellar deficits,6 andarm and eye step and pursuit movements areabnormal in cerebellar dysfunction.7-I Theseresults, and additional work on animals and humans,support a prevalent view that the cerebellar systemcontributes to integration of sensory informationbefore and during the performance of movements.'2One issue ofcontinuing interest with respect to cere-

bellar dysfunction is the contribution of visual andsomaesthetic inputs to the production of cerebellarpostural and kinetic tremor. Damage at a number ofsites in the cerebellum or its outflow pathways canproduce characteristic patterns of tremor. One early

Address for reprint requests: Dr Jerome N Sanes, NINCDS,Building 10, Room 5N-226, Bethesda, MD 20892, USA.

Received 13 October 1987 and in revised form 23 February 1988.Accepted 7 March 1988

view4 13 14 for the origin of tremor is that the damageto cerebellar circuits permits a series of over-corrections that develop into oscillatory errors inperformance. While some experimental evidence sup-ports this concept of a primarily central origin ofkinetic and postural tremor, 15-18 other evidence isinconsistent with this model. With experimentallyinduced cerebellar dysfunction in non-human pri-mates, the external mechanical conditions affecting alimb have been shown to influence the production ofcerebellar kinetic and postural tremor.'9 Addition-ally, and in contrast to the early view,4 3 '4it has beenreported that alterations in visual guidance do notaffect the magnitude of cerebellar postural tremor.20However, it appears that manipulations of the visualenvironment influence the amount of dysmetria inmovements performed by patients with cerebellarataxia.7The present study investigated the influence of

visual guidance on human cerebellar-type tremor andexamined the effect of various types of mechanicalloading on cerebellar postural and kinetic tremor. Inaddition to demonstrating major influences of sensoryinputs on the production of cerebellar tremor, theseobservations offer additional practical insights intorehabilitative strategies that can be offered to patientsdisabled with cerebellar-type tremor.

934

Protected by copyright.

on Decem

ber 27, 2021 by guest.http://jnnp.bm

j.com/

J Neurol N

eurosurg Psychiatry: first published as 10.1136/jnnp.51.7.934 on 1 July 1988. D

ownloaded from

http://jnnp.bmj.com/

Visual and mechanical control ofpostural and kinetic tremor in cerebellar system disordersMaterials and methods

Five patients with postural and kinetic tremor were studied.In each case, the tremor was characterised by rhythmic, 3 to5 Hz, to and fro movements in an upper extremity. Theneurological diagnosis of the patients included structuraland degenerative disease (table). The patient with a rednucleus tumour contralateral to the limbs with tremor,though without a cerebellar lesion, had typical clinicalfeatures found in cerebellar outflow disorders and wasstudied for comparison.AccelerometryPatients were seated and a triaxial accelerometer (WilcoxonResearch Inc., Model 139) was securely strapped to thedorsal surface of the right hand between the meta-carpophalangeal and wrist joints. The accelerometer hadthree piezoelectric strips mounted orthogonally withinprotective shields to detect motion in the x (lateral), y(vertical), and z (sagittal) planes. The complete apparatusweighed 22-5 g. The sensitivity of the piezoelectric elementswas approximately 60 mV/g. The voltages proportional todisplacement of the piezoelectric elements were amplifiedwith a band pass frequency of 2 to 40 Hz. The amplifierrolloff was 6 dB per octave below 2 Hz and 18 dB per octaveabove 40 Hz. After amplification and filtering, data fromeach axis of movement were led to separate channels of a10-bit A/D converter of a PDP 11/03 minicomputer andsampled at 100 Hz.

Involuntary movement was evaluated for three activatedmovements or postures of the arm. These movements were(1) an intended constant posture at 90° of horizontal armflexion and 900 of forward elevation of the arm, (2) a repeti-tive vertical movement in the sagittal plane from shoulderlevel to the knee, and (3) a repetitive horizontal movementwith an excursion of the approximate distance between theshoulders performed at a level slightly below the shoulder.For all tasks the elbow was extended fully. Patients wereallowed to choose a desired speed; they typically performedabout one movement cycle per second. When patients per-formed under the visually targeted condition, the end-pointof the movements or the target for postural maintenance wasprovided by the outstretched hands of the examiner. ForTable Patient description

Patient AgeNo. Sex (yr) Diagnosis

COM* Male 58 Unilateral cerebellar infarctin the region of thedentate nucleus andsuperior cerebellarpeduncle of thenondominant side

C02* Female 65 Olivopontocerebellardegeneration

C03t Female 35 Multiple sclerosis, cerebellarand brainstem lesions

C04* Female 70 Vascular tumour in theregion of the red nucleus,nondominant side

C23$ Male 55 Olivopontocerebellardegeneration

* Examined in accelerometry and torque motor studies.t Examined only with accelerometry studies.+ Examined only with torque motor studies.

each posture or movement, three 20 second samples ofmove-ment were digitised. The postures and movements were per-formed with and without visual control. For trials withoutvisual control, patients were asked either to close their eyesor look elsewhere and to execute movements similar to thesize of those that were performed just before with a visualtarget.Data were analysed offline with a PDP-I I computer.

Acceleration data from each of the x, y, and z axes were firstanalysed independently and then combined, after digitalfiltering and integration, to derive a statistic of the totaldistance travelled in each 20s test epoch. For each axis, theRMS value of acceleration was derived; this value was thenused for integration and summation across the three axes toderive the total distance travelled. A spectral analysis wasdone with standard fast-Fourier analysis procedures and theproportion and absolute values of power in 0-5 Hz bandwidths between 1-5 and 25 Hz were derived.

Torque motor studiesPatients were seated and the forearm was stabilised betweentwo padded plates. The extended hand was placed betweentwo padded plates of a handle that was attached to the axleof a low-friction, brushless DC torque motor (Aeroflex TQ-64). The hand was hidden from the patient's view, and thewrist was positioned directly over the axle of the torquemotor. The patient viewed a visual display that showed aposition and a target cursor. The location of the target cursorwas computer controlled, and the position cursor corre-sponded to the location of the hand as controlled by rotationof the torque motor axle. Patients were instructed to orientthe torque motor handle so that the position cursor and thetarget cursor were aligned. The size of the target cursor wasvaried from 6° to 20° to accommodate the tremor amplitudeof each patient. The cursor alignments had to be maintainedfor 15 to 2 5s, whereupon the target jumped to a newlocation. Patients were instructed then to move the hand attheir own pace to reestablish the alignment between the cur-sors. The required movement size was 30° with an optimalstart position of 150 of extension and a final position of 15°of flexion. A series of 25 movements was performed for eachsensory condition (see below). The handle position was dig-itised at 200 Hz. Muscle activity was recorded with surfaceelectrodes from the forearm flexor and extensor musclegroups. The signals were amplified with high input imped-ance AC-coupled devices (Grass Instrument Co. or BakElectronics, filter settings at 30 Hz and 3 KHz) and thenfull-wave rectified and low pass filtered (4-pole Bessel filterdesign, -3dB at 50 Hz) and then digitised at 200 Hz.Data were analysed offline by inspecting each trial and

marking, with an electronic cursor, periods of time after thepatient had reached a final hand position (fig 1). The rootmean square of the velocity record was measured in thisperiod.Visually guided movements For all series of movements,both the position and the target cursors were initially shownon the visual display. At the end of the alignment period, abrief sound, a double ringing of the computer terminal bellsignified the cue to begin movement. On trials in which thetarget cursor remained visible, the target jumped to a newlocation on the display screen, thereby giving an additionalmovement cue to the patients. Simultaneous with the occur-

935


on Decem


j.com/

J Neurol N


ownloaded from


Sanes, LeWitt, Mauritz

C01 Position

Flexor EMG

Extensor EMG

A A_de k1 sec

C02

4 +

C04

1 sec

C23t

~~KAAAANkAAJkMJLk)~~~~~~~

Fig 1 Wrist position and flexor and extensor muscle activity in each of the four patients examined. Kinematic and muscleresponses were observed when patients performed voluntary wrist movements towardflexion with the torque motorapparatus. In most instances, tremor in the hand was accompanied by reciprocal bursting in antagonist muscles, althoughother muscle activity patterns were evident (see C02 and C23). The arrows demarcate the times between which the amountof tremor was calculated. Position calibration 30° in all records.

rence of the sound a change in the visual display sometimesoccurred so as to alter the visual guidance condition. Foursuch conditions were tested. These were the permutation ofboth the target and position cursors being present or absent.Each condition (for example, the target cursor present andthe position cursor absent) was tested for 25 consecutivetrials. Then a new condition was imposed until all four weretested. The complete visual guidance condition (that is,target and position cursors present) was given first, and theorder of the remaining conditions varied in a counter-balanced fashion across the subject group.Mechanical Influences These experiments examined howchanges in mechanical loading conditions affected cerebellarpostural and kinetic tremor and whether the tremor would beevident when isometric contractions were performed. For allmovements, both the target and the position cursors werecontinuously available for visual guidance. Movements wereopposed or assisted by a constant load, a viscous load, or aninertial load. The constant loads were generated by applyinga continuous voltage to a control circuit and amplifier systemcoupled to the torque motor. Constant loads between 0-64

Nm opposing flexion and 0-64 Nm opposing extension wereused. The viscous load was generated by applying negativevelocity feedback to the torque motor control circuit. Theamount of viscosity was determined by the gain of thenegative velocity feedback. Inertial loads were applied byattaching weights of 250, 500, or 1000 g to the handle of themotor. The resulting moments of inertia were 3 03 kg m210- 3, 6-05 kg m2 10-3 and 12-1 kg m2 10- 3 for the differentloads. The patients were not informed as to the type ormagnitude of the applied load. A series of 25 trials waspresented for each of the viscous, constant, and inertialloads. For isometric contractions about the wrist joint, thehand was inserted between two lightly padded plates of aheavy aluminium frame. The forearm was secured as before.The aluminium frame had strain gauges (in a bridgeconfiguration) that measured deformation of the metalframe caused by wrist flexion or extension. Patients wererequired to perform separate series of 25 isometric con-tractions between 4 Nm of flexion to 2 Nm ofextension. Forloaded movements and isometric contractions, the data wereanalysed as for visually guided movements.

936

I

I

A

./ ^AA


on Decem


j.com/

J Neurol N


ownloaded from


937Visual and mechanical control ofpostural and kinetic tremor in cerebellar system disorders

Targeted No targetPostural maintenance

I/AAAV\V\AAAA0 ~~~~~~~~~~5

20

Fig 2 Tremor during postural maintenance and voluntary arm movements. The acceleration records arefrom patient C03during intended fixation of the arm and voluntary movements in the horizontal plane. These records are taken from thevertical axis for fixation andfrom the horizontal axis for the arm movements and are shown during performance with(Targeted) and without (No Target) visual guidance. The numbers in the upper right panel refer to the elapsed time inseconds; each tracing is 5s record, in sequence from top to bottom. The frequency of the repetitive voluntary limb movement(about I Hz) can be seen in the lower right panel. Note the diminution of tremor in the No Target condition.

Results

Visual guidanceAccelerometry Figure 2 illustrates the accelerationdata from one patient during postural maintenanceand during horizontal movements performed withand without visual targeting. Postural tremor wasgreatest when the patient attempted to maintain a

constant arm position of 900 of both horizontal andforward flexion directed at a visual target. Duringhorizontal movements, tremor was nearly alwaysgreater when visual targets were present than withoutexternal visual targets. For the group of patients, themean total tremor in the 20s test was larger during thevisually targeted movements or posture than duringthe untargeted movements or posture (p < 0-05, fig 3).

Horizontal movement


on Decem


j.com/

J Neurol N


ownloaded from


0 Target

* No target

O

Posture--1-

Vertical Horizontal-I

Movement condition

Fig 3 Average tremor expressed as distance travelled.Summary of accelerometry data for postural maintenanceand the vertical and horizontal arm movements for multiple20s samples. The mean, SEM data for the group ofpatientsare shown. Note that tremor was reduced when visualguidance was eliminated (* = p < 0-01, ** = p < 0-025).

The average amount of tremor suppression in thenon-visually guided conditions was approximately thesame for the three types of movements (1-2 m),though the relative proportion of suppression was

greatest (60% to 80%) for the horizontal movementsand the constant arm position. There was no evidencethat the frequency of tremor changed in relation to thetype of visual targeting.Torque motors Figure 4 shows the hand positiondata from one patient during performance of two ofthe four visual targeting conditions, and fig 5 illus-trates the average results from three patients for allvisual conditions. Withdrawal of visual guidance, byremoving either the target or the position cursor (orboth), reduced the amount of tremor. This effect wasobserved on most trials (see fig 4, bottom). The mosteffective visual guidance condition for reducingtremor was when the position cursor was absent butthe target cursor was seen; all patients tested showedreduction of tremor at the end of movements per-

formed with this type of visual guidance condition.Complete withdrawal of both the position and thetarget cursors was most effective for reducing tremorin one patient, but ineffective for the other twopatients. Removing only the target cursor reduced theamount of tremor for one patient.

Mechanical influencesViscous loads Figure 6a illustrates the results fromthe experiments in which viscous loads of varyingintensities opposed wrist movements. The heavierviscous loads suppressed tremor more than the lighterviscous loads. Linear regression analyses yielded

Sanes, Le Witt, Mauritz

correlations between 0-81 and 0 96 relating theamount of viscosity and tremor in individual patients.When considered independently, only one patient(C04) exhibited a significant linear relationshipbetween the magnitude of the viscous load and theamount of tremor (R = 0 96, df = 2, p < 0-05).However, pooling the data from all patients revealeda significant linear correlation between the amount ofthe viscous load and the tremor magnitude (R = 0 83,df = 16, p < 0 01). Fitting the individual patient datawith second-order polynomials resulted in significantcorrelations (p < 0 05) between the amount of vis-cosity and the magnitude of postural tremor for all

Target present - position present

* * * * I * !* II+ IHI I

Target present- position absent

3O00

1300

i i ii i i i i!I Iif wii ll IIi I :II 1f ! !1

is

Fig 4 Tremor and hand movements. The instantaneoushand position for successive trials performed with (top) or

without (bottom) complete visual guidance is shown forpatient C02. Note the suppression of tremor whenmovements were performed when the position cursor was

blanked simultaneously with the cue to begin movement.

938

10-

if 8--

E 6-0E

4-

2-


on Decem


j.com/

J Neurol N


ownloaded from


Visual and mechanical control ofpostural and kinetic tremor in cerebellar system disorderspatients. The nonlinearity in tremor reduction withrespect to the intensity of the viscous load wasapparent insofar as movements performed against thelightest viscous load, although of a different magni-tude for different patients, resulted in the greatestdiminution of tremor. The lightest viscous loadreduced tremor by 38% to 56% (mean, SEM = 48,4 32%) compared with the tremor observed duringmovements performed without viscous resistance. Incontrast, the heavier loads reduced tremor only anadditional 2% to 23% (mean, SEM = 14, 2-56%).These differences were significant (t = 7-615, df = 12,p < 0-0001).Opposing loads The results concerned with posturaltremor when constant loads opposed or assisted wristflexion are shown in fig 6b. In general, loads thatrequired patients to activate the wrist extensormuscles in order to maintain an intended position inincreased postural tremor in comparison withunloaded movements. For two of the three patientsstudied with this method, the load that requiredpatients to activate the wrist flexor muscles so as tomaintain the intended position decreased the posturaltremor. For the remaining patient (C23), although thetremor when flexor loads were opposed seemedgreater than during unloaded movements, only thelightest flexor load resulted in enhancement of pos-tural tremor (p < 0-025). Additionally, for patientC23 the postural tremor observed when extensorloads were applied was greater than the tremor

l co0El C02

120|C04

100

.580 *

~60 -

40-

204

TaPp TpPa Toai TpPpVisual guidance condition

Fig 5 Average tremor for the visual guidance conditions.Summary of the torque motor data. The average data,derivedfrom the RMS ofhand velocity, are shown for allpatients. (* = p < 0 05, ** = p< 0-005). The amount oftremor is expressed relative to tremor magnitude observed inthe target cursor present-position cursor present (TpPp)condition. T = target cursor; P = position-cursor,a = absent; p = present.

observed when flexor loads were opposed (p < 0-01).Inertial loads The effect of inertial loads on posturaltremor was tested on four patients (fig 6c). For threeof the patients, inertial loads of either 500 or lOOOgapplied to the torque motor handle decreased theamount of tremor. The patient with the rubral tumourshowed a slight increase in tremor; this wassignificantly greater than the no-load condition onlywhen the 500 g load was applied (p < 0-05). Thepalliative effects of inertial loads on tremor were notnecessarily linear, especially as one patient showed anincrease in tremor with an increase of load. Indeed,the tremor of only one patient (C23) was linearlyrelated to the magnitude of the inertial load (R =0-99, p < 0-01); the tremor of the remaining patientsshowed nonlinear responses when increasing inertialloads opposed movement.Isometric contractions Figure 7 shows the resultsfrom the patient who was separately evaluated duringboth isometric and isotonic contractions. This patient(C23) exhibited substantial tremor during isotoniccontractions when either a flexor or an extensor (notshown) load of 0-64 Nm opposed movement. In con-trast, during an isometric contraction that required astep change in torque from 0 to I to 4 Nm towardflexion or extension (not shown), there was no evi-dence of rhythmicity in the torque record. Instead,most trials demonstrated substantial amounts ofinstability, resembling the qualities of dysmetricmovements typically observed in cerebellar disease.

Discussion

The results of this study indicate that both visual andsomaesthetic sensory inputs influence the magnitudeof cerebellar kinetic and postural tremor in humans.In particular, patients with cerebellar dysfunctionexhibited enhancement of tremor when movementswere visually targeted. Additionally, viscous loads,inertial loads, and constant loads opposing flexion alltended to reduce tremor. Finally, there was noevidence of rhythmic involuntary movements whena patient performed voluntary muscle contractionsisometrically. This last result indicates that sensoryinputs related to muscle stretch and joint rotation maybe important in initiating or maintaining tremor incerebellar patients.Of particular significance to the interpretation of

these data is the nature of postural and kinetic tremorin voluntary movement. Both types of tremor may bepresent in cerebellar and outflow pathway disorders,as in each of the patients studied, although dependingon the particular pathology one or the other patternmay predominate. Each type of tremor needs to bedistinguished from the incoordination and decom-

939


on Decem


j.com/

J Neurol N


ownloaded from


* co1El C02i C048 C23

0.0Viscosity

0.64 0.32 0.16F 0 0-16E 0.32 0.64Opposing load (Nm)

250 500Inertial load (gm)

Sanes, Le Witt, Mauritzposition of movement also found with cerebellardysfunction. While distinctions have been drawnbetween the clinical characteristics of postural andkinetic tremors in cerebellar disease,21 22 our datasuggest no significant difference between either type intheir sensitivity to manipulation of visual guidance.Hence, it seems possible that postural and kinetictremors due to cerebellar dysfunction could originatefrom similar neural pathways.An interpretation of the dynamics of kinetic tremor

in cerebellar disease is that there is a defect in avisually based corrective mechanism that controls theaccuracy of voluntary movements.4 15 23 This wouldbe similar to recent interpretations of reduced dys-metria in the absence of visual guidance in patientswith cerebellar ataxia.7 However, this notion has beenchallenged by the demonstration that the lack ofvisual guidance during performance of voluntarymovement did not alter the amount of kinetic tremorin monkeys with dysfunction of the deep cerebellarnuclei.20 24 Instead, cerebellar tremor was thought toreflect disorders in short and long-latency reflex path-ways.2425 For example, a simple alteration in thestretch reflex feedback system, such as latency or gainchanges, could cause oscillations in an otherwisestable system.25

Several observations from the present data argueagainst cerebellar kinetic and postural tremor asreflecting only a corrective mechanism. First,although the tremor was suppressed when the visualfeedback loop was opened, some tremor remained.Thus, it was apparent that neither postural nor kinetictremor were initiated as a direct consequence of intentper se, since the tremors can be dramatically reduced,or even abolished, when the same movements areperformed voluntarily but without precise visual guid-ance. Second, during the isometric contraction task,there was no evidence of rhythmic oscillations in thetorque records in spite of the fact that these musclecontractions were performed under closed-loop visualcontrol. Instead of tremor, the isometric step con-tractions were accompanied by a movement disorderakin to serial dysmetria. In its kinematic profile, serialdysmetria is considerably different from tremor.22 Athird rationale for invoking more than correctivemechanisms in the generation of cerebellar tremor isthat the mechanical loads affected the magnitude of

Fig 6 Tremor and mechanical loads. In all sections, thetremor is expressed as the RMS of hand velocity and isplotted relative to the magnitude of the RMS velocity whenno load opposed movement. a. Viscous loads. The tremorwas reducedfor all loads (p < 0 001). b. Opposing loads.Tremor magnitude is displayedfbr both loads opposingflexion (to the left of the 0 opposing load) and loadsopposing extension. (* = p < 005; ** = p < 0-01).c. Inertial loads. (* p < 005; **= p < 001).

940

'00

0z

.0,

0

200

X 150a00

o

50

0

120

100

X 80

Z 60z;2 40

20

0


on Decem


j.com/

J Neurol N


ownloaded from


Visual and mechanical control ofpostural and kinetic tremor in cerebellar system disordersIsometric Isotonic

I.........iIi ..........I.. .1.., 1.... . I .. I.........1 1........I .........I1., ... ... .. .,.IIIIIII.,II IIIIII.,...... I I ... .... .... .....

Fig 7 Isometric performance. A series of isometric (left) and isotonic (right) contractions performed separately by patientC23. The constant opposing load was 0 64 Nm, and the target isometric torque for the step contraction was 1 Nm. Note theabsence of rhythmic oscillations in the torque records during the isometric contractions. Calibration for the isometriccontractions is I Nm andfor the isotonic contractions is 300. The amount of tremor was derivedfrom the RMS ofhandvelocity (isotonic) orfrom the RMSfirst derivative of wrist torque (isometric).the tremor, sometimes abolishing it, even though thepatients performed limb movements with visualguidance. Although some loads might be expected tosuppress tremor (for example, viscous loads), it hasbeen shown that some aspects of cerebellar tremor arenot susceptible to mechanical loading.26 27 Finally,the frequency of the postural and kinetic tremor washigher (2-5-4 Hz) than the frequencies (- 20-2 5 Hz,allowing for 200s minima each in reaction time andcorrection time) possible to consider that a voluntarycorrective mechanism solely generated the disorderedmovements. Despite, our rejection of the notion thatcerebellar tremor is generated only by visually basedcorrective mechanisms, it was nevertheless clear thatvisual input had some impact on movement control inpatients with cerebellar tremor as it does on cerebellardysmetria.7The present results fail to support the view that

visually based corrective mechanisms are the onlycause of tremor, but they also conflict with the resultsin monkeys showing that withdrawal of visualguidance was ineffective in reducing cerebellartremor.'9 20 24 It is likely that differences in metho-dology can explain the divergent results. The mostsignificant difference may be that in contrast to ourpatients the monkeys tended not to have tremor wheninvoluntary perturbations or voluntary movementswere initiated. Additionally, in an earlier study,'9tremor was examined after a mechanical pertur-bation, so that it is probable that the mechanicalcharacteristics of the limb, rather than voluntary

mechanisms, were being measured. In more recentstudies,2024 even though the voluntary movementswere examined in relation to tremor, the movementsperformed by the monkeys were more rapid than themovements performed by the patients in the currentstudy. Although rapid movements are affected byperipheral inputs,2829 it is likely that slower move-ments are more influenced by visual control.30

The influence of constant loads and isometric per-formance on tremor magnitude may provide someadditional insights into the mechanisms of cerebellarpostural tremor. In particular, the tremor was selec-tively enhanced when extensor muscles were activatedvoluntarily but absent during voluntary isometriccontractions. A previous report3' demonstrated thatfinger extension enhanced cerebellar tremor. Thus, itis possible that voluntary activation of extensormuscles contributes to enhancement of posturaltremor in patients with cerebellar disease. In view ofthe absence of tremor during isometric contractions,the contribution of muscle spindle afferents to themaintenance of tremor might be considered. Duringvoluntary ramp displacements of the hand, thedischarge pattern of neurons in the cerebellarinterpositus deep nucleus was remarkably similar tothat of muscle spindle afferents.32 In another study,33cellular activity in the interpositus nucleus and musclespindle activity were correlated to both the hand posi-tion and EMG associated with the action tremor of anexperimental monkey. The contribution of musclespindles to cerebellar tremor is likely related to the

941

I


on Decem


j.com/

J Neurol N


ownloaded from


942passive length changes of muscles (that is, reflexmechanisms) rather than to mechanisms of alpha-gamma co-activation. Muscle spindle afferents areactivated during both isotonic and isometric volun-tary muscle contractions.34 Therefore the reflexivecontributions associated with voluntary movementwould seem to be a factor in the appearance of tremorduring isotonic contractions and the absence oftremor during isometric performance. The potentialmuscle spindle, and thus reflex, contribution tocerebellar postural tremor could be supported byobservations that stretch reflexes are larger duringisotonic contractions than during isometric con-tractions, even when the initial conditions of torqueand position are held constant.3637There are only limited pharmacological options

for symptomatic control of cerebellar tremor.3839However, the contribution of visual input to intentiontremor may have important consequences for rehabi-litative approaches to the disabilities of cerebellardisease. For example, tremor amplitude may be less-ened by directing a patient to execute a goal-directedmovement without the use of visual guidance (that is,guiding the movement from memory as to theintended target position). Similarly, physical methodsto reduce the severity of kinetic tremor might take intoaccount the ways in which a particular patient'stremor is influenced by mechanical loads. Previousapproaches for attenuating tremor by physical meanshave made use of weights4041 or viscous dampingdevices attached to body appendages. More idealmethods to lessen tremor oscillations can be devel-oped along these principles to apply the particularmechanical load factors observed to decrease tremor.

References

1 Gilman S, Bloedel JR, Lechtenberg R. Disorders of the Cere-bellum, FA Davis Co., 1981

2 Kim JH, Ebner TJ, Bloedel JR. Comparison of response proper-ties of dorsal and ventral spinocerebellar tract neurons to aphysiological stimulus. Brain Res 1986;369:125-35.

3 Glickstein M, May JG, III, Mercier BE. Corticopontineprojection in the macaque: the distribution of labelled corticalcells after large injections of horseradish peroxidase in thepontine nuclei. J Comp Neurol 1985;235:343-59.

4 Holmes G. The symptoms of acute cerebellar injuries due to gun-shot injuries. Brain 1917;40:461-535.

5 Angel RW. Barognosis in a patient with hemiataxia. Ann Neurol1980;7:73-7

6 Friedemann H-H, Noth J, Diener HC, Bacher M. Long latencyEMG responses in hand and leg muscles: cerebellar disorders.J Neurol Neurosurg Psychiatry 1987;50:71-7

7 Beppu H, Nagaoka M, Tanaka R. Analysis of cerebellar motordisorders by visually guided elbow tracking movement. 2.Contribution of the visual cues on slow ramp pursuit. Brain1987;110:1-18.

8 Beppu H, Suda M, Tanaka R. Analysis of cerebellar motor disor-ders by visually guided elbow tracking movement. Brain1984;107:787-809.

Sanes, LeWitt, Mauritz9 Miall RC, Weir DJ, Stein JF. Visuo-motor tracking during

reversible inactivation of the cerebellum. Exp Brain Res1987;65:455-64.

10 Vilis T, Hore J. A comparison of disorders in saccades and in fastand accurate elbow flexions during cerebellar dysfunction.Prog Brain Res 1986;64:207-15.

11 Optican LM, Robinson DA. Cerebellar-dependent adaptivecontrol of the primate saccadic system. J Neurophysiol1980;44: 1058-76.

12 Brooks VB, Thach WT. Cerebellar control of posture and move-ment. In: Brooks VB ed. Handbook of Physiology, Section !,the Nervous System, Vol. 11, Motor Control, Bethesda, Amer-ican Physiological Society, 1981: 877-946.

13 Holmes G. Clinical symptoms of cerebellar disease and theirinterpretation. The Croonian Lectures. II. Lancet 1922:i;1231-7.

14 Holmes G. The cerebellum of man. Brain 1939;62:1-30.15 Growdon JH, Chambers WW, Liu CN. An experimental study

of cerebellar dyskinesia in the rhesus monkey. Brain1967;90:603-32.

16 Liu CN, Chambers WW. A study of cerebellar dyskinesia in thebilateral deafferented forelimbs of the monkey (Macacamulatta and Macaca speciosa). Acta Neurobiol Exp Warsaw1971;31:262-89.

17 Murphey JT, Kwan HC, MacKay WA, Wong YC. Physiologicalbasis of cerebellar dysmetria. Can J Neurol Sci 1975;38:279-84.

18 Gilman S, Carr D, Hollenberg J. Kinematic effects ofdeafferentation and cerebellar ablation. Brain 1976;99:311-30.

19 Vilis T, Hore J. Effects of changes in mechanical state of limb oncerebellar intention tremor. J Neuropkvsiol 1977;40:1214-24.

20 Flament D, Vilis T, Hore J. Dependence of cerebellar tremor onproprioceptive but not visual feedback. Exp Neurol 1984;84:314-25.

21 Fahn S. Cerebellar tremor: clinical aspects. In: Findley LJ,Capildeo R, eds. Movement Disorders, New York, OxfordUniversity Press, 1984:355-63.

22 Sabra AF, Hallett M. Action tremor with alternating activity inantagonist muscles. Neurology 1984;34:151-6.

23 Sutton GG, Sykes K. The effect of withdrawal of visualpresentation of errors upon the frequency spectrum of tremorin a manual task. J Physiol (Lond) 1967;190:281-93.

24 Hore J, Flament D. Evidence that a disordered servo-like mech-anism contributes to tremor in movements during cerebellardysfunction. J Neurophysiol 1986;56:123-36.

25 Mauritz KH, Schmitt C, Dichgans J. Delayed and enhanced longlatency reflexes as the possible cause of postural tremor in latecerebellar atrophy. Brain 1981;104:97-116.

26 Homberg V, Hefter H, Reiners K, Freund H-J. Differentialeffects of changes in mechanical limb properties on phys-iological and pathological tremor. J Neurol NeurosurgPsychiatry 1987;50:568-79.

27 Lee RG, Stein RB. Resetting of tremor by mechanical per-turbations: a comparison of essential tremor and parkinsoniantremor. Ann Neurol 1981;10:523-31.

28 Pelisson D, Prablanc C, Goodale MA, Jeannerod M. Visualcontrol of reaching movements without vision of the limb.11. Evidence of fast unconscious processes correcting thetrajectory of the hand to the final position of a double-stepstimulus. Exp Brain Res 1986;62:303-1 1.

29 Sanes JN. Kinematics and end-point control of arm movementsare modified by unexpected changes in viscous loading. J Neu-rosci 1986;6:3120-7.

30 Woodworth RS. The accuracy of voluntary movements. PsycholRev Monogr 1899;3:1-114.

31 Chase RA, Cullen JK, Sullivan SA. Modification of intentiontremor in man. Nature 1965;206:485-7.

32 Thach WT, Schieber MH, Mink J, Kane S, Home M. Cerebellarrelation to muscle spindles in hand tracking. Prog Brain Res1986;64:217-24.


on Decem


j.com/

J Neurol N


ownloaded from


Visual and mechanical control ofpostural and kinetic tremor in cerebellar system disorders33 Elble RJ, Schieber MH, Thach WT. Activity of muscle spindles,

motor cortex and cerebellar nuclei during action tremor. BrainRes 1984;323:330-4.

34 Burke D, Hagbarth, K-E, Skuse NF. Recruitment order ofhuman spindle endings in isometric voluntary contractions. JPhysiol (Lond) 1978;285:101-12.

35 Hulliger M, Nordh E, Vallbo AB. Direction in muscle spindleafferents related to direction of slow precision movements inman. J Physiol (Lond) 1985;362:437-53.

36 Akazawa K, Milner TE, Stein RB. Modulation of reflex EMGand stiffness in response to stretch of human finger muscle. JNeurophysiol 1983;49:16-27.

37 Kanosue K, Akazawa K, Fujii K. Modulation of reflex activity ofmotor units in response to stretch of a human finger muscle.

Jap J Physiol 1983;33:995-1009.38 Legg NJ. Treatment of cerebellar tremor. In: Findley LJ,

Capildeo R, eds. Movement Disorders, New York, OxfordUniversity Press, 1984;377-86.

39 Sabra A, Hallett M, Sudarsky L, Mullally W. Treatment ofaction tremor in multiple sclerosis with isoniazid. Neurology1982;32:912-1 3.

40 Hewer RL, Cooper R, Morgan MH. An investigation into thevalue of treating intention tremor by weighting the affectedlimb. Brain 1972;95:579-90.

41 Morgan MH, Hewer RL, Cooper R. Application of an objectivemethod of assessing intention tremor-a further study on theuse of weights to reduce intention tremor. J Neurol NeurosurgPsychiatry 1975;38:259-64.

943


on Decem


j.com/

J Neurol N


ownloaded from


Documents

Visual - BMJ