18ournal ofNeurology, Neurosurgery, and Psychiatny 1993;56:1078-1-084

The influence of external timing cues upon therhythm of voluntary movements in Parkinson'sdisease

J S Freeman, FW J Cody, W Schady

Department ofPhysiologicalSciences, UniversityofManchester,Manchester M13 9PTFW J CodyJ S FreemanDepartent ofNeurology,Manchester RoyalInfirmary,Manchester M13 9WLW SchadyCorrespondence to:Dr F W J Cody, Departmentof Physiological Sciences,Stopford Building,University of Manchester,Manchester M13 9PT.Received 19 May 1992and in revised form8 January 1993.Accepted 19 January 1993

AbstractThe ability of patients with Parkinson'sdisease (PD) and healthy subjects to syn-chronise finger tapping, produced byrhythmic wrist movements, with audi-tory signals of target frequencies (range1-5 Hz) and to sustain such rhythms fol-lowing sudden withdrawal of auditorycues was studied. Healthy subjects wereable, in the presence of auditory cues, toduplicate target frequencies accuratelyover the range investigated both in termsofmean tapping rate and in regularity oftapping. PD patients were less accurateunder these conditions and on averagetended to tap too rapidly at the lower(1-3 Hz) target frequencies and tooslowly at the highest (5 Hz) target fre-quency. In addition, the variability oftheir tapping rhythms was generallygreater. Healthy subjects were able tosustain tapping rhythms well followingsuppression of auditory signals. By con-trast, withdrawal of external timing cuesresulted in marked impairment of thepatients' rhythm generation. At lowerfrequency targets (1-3 Hz) patients' tap-ping rates increased over rates whichwere already elevated in the presence ofexternal cues. Conversely, at higher tar-get frequencies (4-5 Hz), the averagetapping rate tended to decline furtherfrom previously depressed levels. Theaccuracy of almost all patients fell out-side the normal range. Two patterns oftapping errors were found. The first washastening of tapping which was most evi-dent at intermediate target frequencies.The second was faltering which occurredmainly at the higher target frequencies.These forms of behaviour may resultfrom inherent abnormalities of internalrhythm generation since they occurredboth in the presence and absence ofexternal dtming signals. Overall, ourfindings are consistent with the view thatthe basal ganglia have a role in the inter-nal cueing of repetitive voluntary move-ments.

(_7 Neurol Neurosurg Psychiatry 1993;56: 1078-1084)

A current theory of basal ganglia functionattributes these motor centres with a primaryrole in the internal initiation and timekeepingof movements. This view is, in part, foundedon the common clinical observation that

many patients with Parkinson's disease (PD)experience difficulty in performing repetitivevoluntary movements.' An extreme exampleof this type of movement deficit is the wellknown "freezing" phenomenon of PD, inwhich patients exhibit an inability to start orto continue cyclic motor activities such aswalking, speech, or handwriting.234 In addi-tion, there is much anecdotal evidence thatPD freezing may be both triggered and termi-nated by external sensory inputs. For exam-ple, periodic visual inputs from a stripedpattern on the floor or a staircase can helpPD patients to sustain locomotion. Thus,suppression of freezing by external cues indi-cates an increased dependence of the rhythmgenerator upon extrinsic reinforcement for itscontinued operation. As a corollary, freezingmay represent a basic failure of internalrhythm generation by the basal ganglia.

Detailed measurements of the influence ofexternal sensory information on the ability ofPD patients accurately to produce rhythmicmovements are, therefore, highly relevant inassessing the possible role of the basal gangliain internal rhythm generation. There is, asyet, a dearth of such quantitative data.Nakamura et al 5 6 7 8have reported abnormali-ties in the ability of PD patients to performrepetitive finger tapping in response to audi-tory signals and have proposed that character-istic deficits in rhythm formation exist in thisdisorder.

In the present experiments we havefocused upon a quite different aspect ofrhythm formation in PD, namely the relianceof patients on external timing cues to gener-ate and maintain rhythms. To this end, wecompared the effects of withdrawal of audi-tory cues upon finger tapping performance inPD patients and healthy subjects. We alsoinvestigated whether abnormal rhythm pro-duction in PD was related to the occurrenceof freezing in everyday activities and to theco-existence of tremor.

MethodsSUBJECTSNine patients (five men, four women) aged60-8 (6 7) years, mean (SD), and 12 healthysubjects (five men, seven women) aged 63A4(7 9) years were studied. All subjects partici-pated with informed consent and the proto-cols were approved by the local ethicalcommittee.The diagnosis of PD was made by a con-

sultant neurologist on the basis of the classic

1078 on M

arch 18, 2020 by guest. Protected by copyright.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.56.10.1078 on 1 October 1993. D

ownloaded from

The influence of external timing cues upon the rhythm of voluntary movements in Parkinson's disease

triad of tremor, rigidity, and bradykinesis andthe absence of any atypical signs or symp-toms. Patients were investigated while ontheir routine therapy, which in all casesincluded standard levodopa formulations andin some cases additional medication. No dis-tinction was made between the "on" and"off' phases. The average duration of thepatients' disease at the time of testing was 6-4(3 0) years. Patients were assessed for rigid-ity, bradykinesia, and gait, each scored onfour point scales9 in which 0 = normal orabsent and 3 = severely disturbed, immedi-ately before experimental sessions. Rigidity:eight patients were graded 1 and one patientgraded 2. Bradykinesia: five patients weregraded 1 and four patients were graded 2.Gait: two patients were graded 0, twopatients were graded 1, three patients weregraded 2, and two patients were graded 3.Four of the patients had an asymmetrical pat-tern of Parkinsonian signs and of these twowere more severely affected on the left sideand two on the right. Postural tremor wasmeasured accelerometrically at each wrist.Mean peak frequencies for the left and rightwrist were, respectively, 6-0 (1.5) Hz and 6-1(1 5) Hz with corresponding ranges of4*3-8-7 Hz and 4-0-8-2 Hz. The incidence offreezing of gait during everyday activities wasalso scored for individual patients accordingto a four point scale: 0, never; 1, several timesper month; 2, several times per week; 3,daily.

FINGER TAPPING TESTSubjects sat in a chair with one forearm rest-ing comfortably upon a table placed in frontof them. A flexible metal loop was fittedsnugly to the index finger just proximal to theterminal interphalangeal joint. The hand wasplaced palm downwards over a wooden boardto which was attached a metal contact plate.Subjects produced tapping of the index fingerby making rapid, alternating flexion andextension movements of the wrist. They wereinstructed to make small but distinct move-ments, raising the finger about 5 mm abovethe contact plate between strikes, so as tominimise any effects of bradykinesia in thepatients. Each time the metal loop on theindex finger struck the contact plate it com-pleted an electrical circuit and generated abrief voltage pulse.

Auditory cues of target tapping frequencieswere played through a loudspeaker assequences of "clicks". These signals wereproduced by an electronic signal generatorfeeding into an audioamplifier. Trains of reg-ularly spaced cue pulses at 1, 2, 3, 4, and 5Hz were used. Blocks of trials in whichdifferent target frequency cues were pre-sented were interspersed in a pseudorandommanner.Two main protocols were employed. In

protocol 1, tapping performance wasrecorded over 30 second periods, throughouteach of which an auditory cue signal at agiven preselected frequency was presentedcontinuously. Subjects were instructed to tap

in rhythm with the auditory "clicks". Suchrecording periods were separated by gaps ofabout 1 minute to allow subjects to rest. Inprotocol 2, tapping performance was againrecorded over a series of 30 second periods.However, in this case the auditory cues werepresented for only the first 10 seconds, afterwhich the "clicks" were abruptly turned off.Subjects were instructed to tap in rhythmwith the cue signal during the initial 10 sec-ond phase and then to continue to tap at thesame rhythm during the remaining 20 secondphase of the trial when the cues were absent.Again, 1 minute recovery periods wereallowed between trials. Tapping performanceof each hand was tested: protocol 1 was firstapplied to both hands and then both handswere tested with protocol 2.

DATA ANALYSISA computer was used to sample, store, andanalyse tapping and tremor data. Voltagepulse trains corresponding to the occurrenceof, respectively, finger taps and auditory cuesignals were sampled at 100 Hz by separatechannels of the A/D. Any tap pulse which fol-lowed its predecessor by an interval of <50ms was rejected to eliminate counting artefac-tual contacts. Mean (SD) frequencies andplots of instantaneous frequencies (recipro-cals of successive intertap intervals) of tap-ping pulses were calculated for each samplingperiod. Accelerometric tremor recordingswere sampled at 50 Hz by one channel of theA/D and analysed by Fourier transform todetermine the power spectrum, peak fre-quency and overall power.

STATISTICSMultivariate analyses of variance(MANOVA) were applied to test whetheroverall differences existed between-firstly,the mean rates and variability of tapping per-formance of the PD and control groups in thepresence of auditory cues, and secondly, thetapping performance within each group in thepresence and absence of cues. Unpaired ttests (corrected for unequal variance) andpaired t tests were used to identify those sig-nal frequencies at which differences existedbetween the mean tapping rates, respectively,of the PD and control groups and of individ-ual subjects in the presence and absence ofcues. These comparisons each involved fivesets of dependent tests. In this situation thereis an increased risk of false positives (type Ierrors) and a compensatory a adjustment pro-cedure (for example, Bonferroni) may beapplied. In the present analyses, Bonferroniadjustment dictates that a p value of 0-01should be interpreted at the a = 0 05 level ofsignificance. However, many statisticians (seeCohen'0 for review) consider the Bonferroniadjustment to be excessively conservative andthat its use leads to an unacceptable lossof power (increase in type II errors-thatis, false negatives). Therefore, precise, non-adjusted p values are given throughout thetext so that readers can decide significancelevels.

1079 on M

arch 18, 2020 by guest. Protected by copyright.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.56.10.1078 on 1 October 1993. D

ownloaded from

Freeman, Cody, Schady

Normal

A

TflT I 1 l FF F

0

PD 5-0

B

9 (s)

C

0 9 (s)

D

111111l111111 11111111111111ITITHUllllll

N

I 4.0

l) 3-0 -

a)

2-0-C:._.c, 1.0-

0o0 -

-A Left

o Normal subjectsA PD Patients

0.0 1.0 20 30 4.0 5.0Signal frequency (Hz)

510l B

0 9 (s) 0

E

5 HzAuditorysignals

Taps

0

9 (s)

F

9 (s) 0

Figure 1 Finger tapping performance ofa normal subject and PD patient in the preseof auditory signals of targetfrequency. Separate series ofpulses representing the occurreiof, respectively, auditory signals andfinger taps show the relative timing of these two seevents in 9 second records obtained during 30 second periods of tapping at 2 Hz, 3 Hz,and S Hz targets in the normal (A, C, and E) and PD (B, D, and F) subjects.

-

I 40 -

a) 3-0 -

a'a)20

.a_

Q. 1*0 -aH

Right

0o0 J

0.0 1.0 20 3.0 4.0 5.0Signal frequency (Hz)

Figure 2 The relationships between finger tapping9 (s) frequency and the frequency of auditory cue signals innce healthy subjects and PD patients. Mean (SD) tappingnce frequencies are shown. Datafor left (non-preferred) handtseof are plotted in (A) andfor the right hand in (B). Thets of dotted line, with unity slope, indicates exact correspondence

of tapping and cue frequencies. The mean tapping rates ofthe PD group, across the range of signalfrequencies,differed significantly from those of the control group(p = 0 01, MANOVA).

ResultsEXTERNALLY SIGNALLED RHYTHMGENERATION IN PD AND HEALTHY SUBJECTSFigure 1 compares the abilities of a healthysubject and a representative PD patient toduplicate the rhythm of sequences of auditorycues, presented at varying frequencies, bymaking voluntary wrist movements to pro-duce tapping of the index finger.

Segments of recordings made from, respec-tively, the healthy subject (fig 1A, C, and E)and the PD patient (fig 1B, D, and F) during2 Hz, 3 Hz and 5 Hz trains of equally spacedauditory "clicks" show characteristic differ-ences in performance. The pattern of pulsesproduced by the tapping movements of thehealthy subject reproduced far more accu-rately the rhythm of the auditory cues, bothregarding mean frequency and regularity,than those of the patient. In particular, thepatient tapped too rapidly at lower (1-3 Hz)signal frequencies and too slowly at higher(4-5 Hz) signal frequencies. The lack of con-sistency in the relative timing of tap and cuepulses in figure 1 suggests that neither the PDpatient nor the healthy subject respondeddirectly to each individual "click" in the trainof auditory signals at any of the cue frequen-cies. Interspersed in the patient's tappingpulse trains are occasional very short intervals(about 110-125 ms), corresponding toinstantaneous tapping rates of 8-9 Hz, which

are not present in the records of the healthysubject. However, equally short intertapintervals appeared in the records of healthysubjects striving to tap at their maximum rate(typically averaging about 6 Hz for a 10second period), when their rhythm becamefar more variable. The mechanism responsi-ble for these sporadic, closely spaced pulsepairs is uncertain. However, none had instan-taneous frequencies more than about 10 Hzand their occurrence had no definiteperiodicity.

Plots of group data (fig 2) indicate that PDpatients were, in general, less exact thanhealthy subjects in replicating the cue fre-quency and in producing an even pace.Results from the left (fig 2A; non-preferred inall subjects tested) and right (fig 2B) handsare illustrated. The relatively larger SD valuesof the patient group suggest that the tappingperformance of the patients, at each signalfrequency, was more variable than that of thecontrols and this was confirmed by statisticalanalysis (p = 0-01, MANOVA, note that non-Bonferroni adjusted p values are giventhroughout, see Methods). The mean tappingrates, considered across the overall range oftarget frequencies, of the patient and controlgroups also differed (p = 0-01, MANOVA).Figure 2 shows that the PD patients as awhole tended to tap more rapidly thanhealthy subjects at low intermediate signal

2 HzAuditorysignals

Taps

3 HzAuditorysignals

Taps

J U . L J L L L L J I

.. I. .. " ., , ff I. ... ... .. I...-.. ........ oIII

1080 on M

arch 18, 2020 by guest. Protected by copyright.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.56.10.1078 on 1 October 1993. D

ownloaded from

The influence of external timing cues upon the rhythm of voluntary movements in Parkinson's disease

A Normal

B

4

2

0) (s)

PD

C Normal

10

PD

10

D10o

5

30 (s) 0

Figure 3 Plots of instantaneous frequency offinger tapping in a healthy subject andpatient in the continuous presence of3 Hz auditory signals andfollowing removal of i

Each point represents the occurrence ofa single tap. The height ofeach point indicateinstantaneous frequency in relation to the immediately preceding tap and was calculathe reciprocal of the intertap interval. In (A) and (B) the tapping performances of,respectively, the healthy subject andPD patient in the continuous presence of cues areshown. The solid horizontal lines indicate cuefrequency. In (C) and (D) the auditorsignal was played, respectively, to the healthy subject and patient during the first IOseach plot. At the times indicated by the arrows the auditory signals were turned off antapping during the remaining 20s periods was in the absence of auditory cues. The soihorizontal lines in the initial 10 second periods of the records show thefrequency of thauditory signals. These levels are continued as dashed lines in the later 20 second periThe scales of the vertical axes differ in (D) from those in (A) (B) and (C).

frequencies and less rapidly at the higheHz) frequency investigated. The mean

hand tapping rates of the patient grouphigher than those of the control group fo:frequencies of 1 Hz and 2 Hz and lowethe 5 Hz signals (p values, respecti0-004, 0-009, and 0-036, t test). Essensimilar findings were obtained for right Iferred) hand performance, although onlyHz did the difference approach a signifilevel (p = 0 049). In neither the patientcontrol groups were there significant dences between the mean tapping rates oversus right hands at any of the targetquencies.

Detailed investigation of the behavioLindividual patients revealed that two dissubclasses of tapping abnormalitiestributed to the impaired accuracy, in ternmean rates, of the PD group as a w]Abnormal performance was defined as

ping rates which fell, at two or more t;frequencies, outside the range of v.obtained from healthy subjects. Data forParkinsonian hand (18 hands frompatients) were analysed independently.first pattern of abnormality, termed ha:ing, comprised an increase and the sectermed faltering, comprised a reductiomean tapping frequencies. Instances oftening (four hands) were most commo3 Hz and 4 Hz targets while falteringhands) was only displayed at the higher (

and 5 Hz) target frequencies. At the 5target, some hasteners tapped in excess o

signal frequency. Only one patient shcabnormal performance with both handsshe displayed contrasting patterns on thesides.

RHYTHM GENERATION IN THE ABSENCE OFEXTERNAL SIGNALS IN PD AND HEALTHY

;',, SUBJECTSFigure 3 presents instantaneous frequencyplots of the tapping performance of represen-tative healthy and PD individuals during 30second periods in which a 3 Hz auditory cue

30 (s) signal was either present throughout (A andB) or removed after 10 seconds (C and D).The plots obtained during continuous

auditory signals show two main features.Firstly, the healthy subject (fig 3A) and PDpatient (fig 3B) were both able to establishalmost immediately a tapping rhythm approx-imating to the cue frequency upon onset ofthe auditory signal. Secondly, in neither case

30 (s) is there any definite sign of a transition in per-tPD formance such as might result from loss ofcues concentration or tiredness. Comparison ofists the tapping performance of individual sub-ted as jects during the first and last 10 second peri-

ods of the 30 second tests confirmed that iny neither patients nor controls was there a sys-of tematic deterioration in accuracy as trials pro-id gressed. Similar findings were obtained for alle target frequencies.rods. Figure 3C shows that the tapping perfor-

mance of the healthy subject was relativelyunaffected by withdrawal of the auditory sig-nals. By contrast, the performance of thepatient (fig 3D, note different frequency

st (5 scale) underwent a clear alteration followingleft removal of auditory cues. There was an

were increase in average tapping frequency and ther cue rhythm became far more irregular. Similarr for trends were noted for the 3 Hz target fre-ively, quency in the patient group as a whole (seetially below). These changes seem certain to have(pre- resulted from withdrawal of cue signals, sinceat 3 they did not occur in the presence of auditoryicant cues (fig 3B).t nor The plots in figure 4 compare the tappingiffer- performance of healthy and PD groups in theWfleft presence of auditory signals and followingfre- their sudden removal once the rhythm had

been well established. Results for the leftar of (non preferred) hand are illustrated.;tinct Figure 4 shows that the withdrawal ofcon- external cues had a pronounced influence onns of the performnance of the patients (fig 4B)hole. whereas it had little effect on that of thetap- healthy subjects (fig 4A). Statistical analysisarget confirmed that the mean tapping rates of thealues PD group, across the whole range of signaleach frequencies, differed significantly in the pres-nine ence and absence of cues (p = 0 01,The MANOVA) while those of the control groupsten- did not. Cue suppression resulted in the:ond, patients' mean tapping frequencies increasingIn in at the two lowest target frequencies (1 Hzhas- and 2 Hz) and declining at the 4 Hz target (pin at values, respectively, 0-038, 0-027, and 0-033,(four paired t test). Similar findings were obtained4 Hz for their right (preferred) hand tapping per-5 Hz formance. Withdrawal of cue signals causedof the increases in mean tapping rates at targets of)wed 1, 2, and 3 Hz and a reduction at 5 Hz (p val-and ues, respectively, 0-009, 0 035, 0-028, andtwo 0 047, paired t test). In all cases, these

changes represented a further reduction in

0 30

C.c;0)

r-)

CD._

aCl

o

F-

._aa

0

6

06

4 #I .11 . '. .,. .

'k. . . .1. ,. 6% 0 t, 0.

'I, , -.1" o' $ $0

1081

A

1

2

; -n-V ;- re-N - 0., ; , 0 0 - . * e - ,

. . .0 . .1.4 04,

on March 18, 2020 by guest. P

rotected by copyright.http://jnnp.bm

j.com/

J Neurol N

eurosurg Psychiatry: first published as 10.1136/jnnp.56.10.1078 on 1 O

ctober 1993. Dow

nloaded from

Freeman, Cody, Schady

Figure 4 Plots comparingthe tapping performance ofsubjects in the presence andabsence of auditory signalsfor the group of healthysubjects (A) and group ofPD patients (B). Mean(SD) values for the left(non-preferred) hand aregiven. Dotted lines indicateunity slope. The meantapping rates of the PDgroup, across the range ofsignalfrequencies, differedsignificantly in the presenceand absence of auditorycues (p = 0.01,MANO VA). Those of thecontrol group did not.

5 0 -

I 4-0-

0

CTU) 3 0-

0)U)$" 2.0-

CL._m 1-0-

0-0 -

5-0 -

NI 4-0-

0CU) 3-0 -

CD"", 2 0 -

0.Q

CD 1-0-

0-0 -

A

0.0 1-0 2-0 3-0 4Signal frequency (Hz)

B

0.0

PD

1-0 2-0 3-0 4-0Signal frequency (Hz),

. excluded: one showed different abnormalities¶in the two hands and one exhibited an alter-

ation in pattern between the test conditions.There was no evidence of any differencesbetween these two patient categories concern-ing scores of clinical bradykinesia. However,there was a tendency for the patients in thehastener group to show more frequent freez-ing episodes. All four of these patients hadfreezing ratings of 2 or 3, whereas each of thethree patients in the falterer group had ratingsof 0 or 1.

5-0

DiscussionT The present findings demonstrate that the

ability of PD patients to generate simple,4' rhythmic voluntary movements is impaired in

two distinct respects. First, and in agreementw-ith previous studies,567 patients are less ablethan healthy subjects to synchronise accu-rately their movements to extrinsic timingcues. Secondly, the patients exhibit a greaterreliance on external cues for rhythm forma-tion. Thus, in PD there appears to be adeficit of external guidance co-existing with aprobably more fundamental derangement of

5-0 internal cueing for stereotyped, repetitive vol-untary movements.

accuracy. Although these effects of cue sup-pression, at individual signal frequencies, are

often at borderline levels of statistical signifi-cance (see Methods), the fact that they wereobserved bilaterally strongly supports theirgeneral validity.

Analysis of the performance of individualpatients' hands demonstrated that instancesof the two subclasses of abnormal perfor-mance, hastening and faltering, were foundfollowing cue withdrawal. Hastening was nowthe predominant pattern (six hands out of 18possible instances) and was most commonlynoted for 2-4 Hz targets while faltering (twohands) was observed at 3-5 Hz targets.

RELATIONS OF ABNORMAL RHYTHMGENERATION IN PD TO TREMOR,

BRADYKINESIA, AND FREEZINGTremor characteristics were comparedbetween PD hands showing different types oftapping abnormalities. PD hands were

grouped according to hastening, faltering, or

"within normal" behaviour as defined above.Neither the mean peak frequency nor thepower of postural tremor differed significantly(t test) between these groups. Individualinstances of abnormal tapping performancewere noted in the absence of appreciabletremor, while in other instances tapping accu-

racy was "within normal limits" despite thepresence of pronounced, low frequencytremor.To analyse the relation of abnormal tap-

ping performance to bradykinesia and freez-ing, patients were categorised as hasteners or

falterers depending on whether they displayedexclusively hastening or faltering performanceas previously defined. Two patients were

ABNORMALIMES OF SYNCHRONISATION OFMOVEMENTS TO EXTERNAL CUES IN PDOur findings generally confirm earlier obser-vations of Nakamura and colleagues-7 forfinger tapping and of Logigian et al,1' forrepetitive isometric contractions of fingermuscles that PD patients show two separatetypes of synchronisation abnormalities. Thefirst is an elevation of tempo as targets exceed2-3 Hz and the second a depression of tempoat targets approaching 5 Hz.A straightforward explanation of the latter

form of behaviour, since slowness of move-ment is a common feature of PD, is thatpatients cannot produce sufficiently rapidmovements. We cannot exclude the possibil-ity that an inherent inability of some of ourpatients to maintain an adequate pace ofmovement contributed to their synchronisa-tion errors for the 5 Hz target. However, anumber of patients with extreme disturbancesof rhythm formation, the hasteners, actuallyachieved mean tapping rates in excess of 5Hz. Thus, slowness of movement was not auniversal limiting factor in the PD group anda more basic synchronisation deficit wasprobably responsible.

This view is supported by the occurrencein patients' records of sporadic groups of twoto three taps at instantaneous frequencies of8-9 Hz. Such closely spaced taps, which con-tribute to the irregularity of rhythms in PD,are unlikely to have been produced by indi-vidually planned, separate movements. Thefact that none occurred at instantaneous ratesexceeding about 10 Hz suggests that theywere not simply artefactual contacts; if so,rates up to 20 Hz (for 50 ms rejectioninterval) would be anticipated. Although aninteraction between Parkinsonian tremor and

1082

4-0

on March 18, 2020 by guest. P

rotected by copyright.http://jnnp.bm

j.com/

J Neurol N

eurosurg Psychiatry: first published as 10.1136/jnnp.56.10.1078 on 1 O

ctober 1993. Dow

nloaded from

The influence of external timing cues upon the rhythm of voluntary movements in Parkinson's disease

voluntary activity has been reported,"I thelack of any obvious periodicity in the occur-rence of these high frequency bursts arguesagainst such a mechanism having been thesole cause. A clue to their likely origin is pro-vided by the appearance of similar 8-9 Hzpulse pairs in the records of healthy subjectsattempting to tap at maximal rate when inter-tap variability increases dramatically andapproaches Parkinsonian levels. Present mod-els'2 of repetitive movement propose that acentral neural oscillator sets the tempo of therhythm. Thus, the increased variability oftapping of healthy subjects at maximal ratesmay be due to destabilisation of such a neuraloscillator. If so, the occurrence of ananalagous oscillator instability, but atrelatively lower frequencies, may underliePD patients' difficulties in synthesisingsequences of regular and appropriately pacedmovements.

DEPENDENCE OF PD PATIENTS ON EXTERNALCUES FOR RHYTHM FORMATIONThe main new finding of the present study isthat the tapping performance of PD patientsdeteriorates following withdrawal of externaltiming cues. Tapping rates increased for lowintermediate frequencies and decreased forhigh target frequencies. These changes corre-spond to an exaggeration of pre-existing pat-terns of inaccuracies found in the presence oftiming cues.

Models of rhythmic movement generationby the nervous system comprise two concep-tually distinct elements, namely an internaltimekeeper or oscillator component and animplementation or effector component."3Although movement implementation itself isabnormal in PD, inaccuracies of repetitivemovement arising from this source are likelyto occur equally in both the presence andabsence of external timing cues. Theincreased dependence of PD patients onexternal timing cues for rhythm formation,therefore, argues strongly in favour of adeficit of internal timekeeping. Since healthysubjects are able to generate motorprogrammes for ongoing repetitive move-ments by extrapolating temporal parametersfrom previous similar movements,'4 thiscapacity must be assumed to be impaired inPD.Nakamura et al'7 have, on quite separate

grounds, also proposed a disturbance of inter-nal rhythm formation in PD. These authorsplaced considerable emphasis on a hasteningphenomenon in PD which was characterisedby a clear cut transition frequency of 2-3 Hz,at which the tapping rate suddenly jumped toa new plateau level of 5-6 Hz. The latter fre-quency was thought to represent an intrinsicoscillation occurring within the healthy ner-vous system which is unmasked in PD. In thisrespect, however, the present results differfrom those of Nakamura et al.57 The tappingrhythms of our PD patients very rarelyshowed an obvious transition frequency orany sign of convergence to a singlefrequency.

RELATION OF ABNORMALITIES OF RHYTHMGENERATION TO TREMOR, BRADYKINESIA, ANDFREEZING IN PDTremor is a common feature of PD. It occurspredominantly at frequencies of 4-6 Hz'5 andis believed to result from oscillatory dischargeof neurons in the ventrolateral thalamus.'6 '7Logigian et al" have proposed that in PD theneural oscillators for repetitive voluntarymovement and for tremor become synchro-nised. The general tendency of our patients'tapping frequency to shift towards the PDtremor band is consistent with this "attrac-tor" theory. However, none of the patientsshowed a dominant frequency in their tap-ping behaviour such as would be expected if apowerful entrainment occurred. More impor-tantly, the incidence of the different types ofabnormal tapping performance (hastening orfaltering) among our PD sample correlatedwith neither the peak frequency nor thepower of the recorded postural tremor. Inaddition, clear instances of tapping deficitswere seen in patients lacking appreciabletremor. This echoes the finding ofNarabavashi and Nakamura3 that distur-bances of repetitive tapping persist in PDpatients following alleviation of tremor (andrigidity and akinesia) by medication or surgi-cal thalamotomy. Therefore, while the tap-ping performance of our patients waspresumably influenced to some extent by co-existing tremor (and its pathological neuralgenerator), the balance of evidence favours amore fundamental disruption of an indepen-dent internal oscillator for repetitive voluntarymovement.As discussed earlier, slowness of voluntary

movement does not appear to be responsiblefor inaccuracies of rhythmic tapping move-ments in PD. Equally, there was no system-atic association between the presence ofhastening or faltering tapping behaviour andthe clinical score of bradykinesia.Nakamura et al36 regarded the hastening

phenomenon they observed in rhythmic fin-ger tapping as closely related to PD freezing.A relationship between hastening and freezingis also suggested by the observation that PDpatients often show a freeze-release-hasten-freeze cycle gait and speech in which, forexample, freezing episodes during walking areimmediately preceded by a quickening ofpace and shortening of step.'8 In keeping withthis notion, we found that all of our PDpatients whose sole tapping disorder was anabnormally increased rate reported relativelyhigh incidences of freezing episodes duringeveryday activities. Therefore, some commonpathological processes probably predispose todisturbances of rhythmic hand movementsand freezing in PD, although it would bepremature to assume a single causativemechanism.

Overall, the present finding that ourpatients were abnormally dependent uponexternal timing cues to regulate the tempo offinger tapping indicates that in PD deficits ofcentral programming are not confined to thesequencing of relatively complex movements

1083 on M

arch 18, 2020 by guest. Protected by copyright.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.56.10.1078 on 1 October 1993. D

ownloaded from

Freeman, Cody, Schady

(see Refs 19-22) but also apply to the rhythmgeneration which underlies extremely simple,stereotyped, and repetitive voluntary move-ments.The work was supported by the MRC. JSF was in receipt of aSERC Studentship. We wish thank Ms Frances Culshaw fordevelopment of computer software and tapping apparatus andDrs RG Lascelles and D Neary for allowing us to studypatients under their care.

1 Marsden CD. The mysterious motor function of the basalganglia: the Robert Wartenberg lecture. Neurology 1982;32:514-39.

2 Andrews CJ. Influence of dystonia on the response tolong-term L-dopa therapy in Parkinson's disease. J7Neurol Neurosurg Psychiatry 1973;36:630-6.

3 Narabayashi H, Nakamura R. Clinical neurophysiology offreezing in parkinsonism. In: Delwaide PJ, Agnoli A,eds. Clinical neurophysiology in Parkinsonism. New York:Elsevier, 1985:49-57.

4 Kanazawa I. Clinical pathophysiology of basal ganglia dis-ease. In: Vinken PJ, Bruyn GW, Kiawans HI, eds.Handbook of clinical neurology. Vol 5. Extrapyramidal dis-orders. New York: Elsevier, 1986:65-91.

5 Nakamura R, Nagasaki H, Narabayashi H. Arrhythmo-kinesia in parkinsonism. In: Birkmayer W,Homykiewicz 0, eds. Advances in Parkinsonism. Basle:Roche, 1976:258-68.

6 Nakamura R, Nagasaki H, Narabayashi H. Disturbancesof rhythm formation in patients with Parkinson's dis-ease: Part I. Characteristics of tapping responses to theperiodic signals. Percept Mot Skills 1978;46:63-75.

7 Nagasaki H, Nakamura R, Taniguchi R. Disturbances ofrhythm formation in patients with Parkinson's disease:Part II. A forced oscillation model. Percept Mot Skills1978:46:79-87.

8 Nagasaki H, Nakamura R. Rhythm formation and its dis-

turbances: a study based upon periodic response of amotor output system. J Hum Ergology 1982;11:127-42.

9 Webster DD. Critical analysis of the disability inParkinson's disease. Modem Treatment 1968;5:257-82.

10 Cohen J. Things I have leamed (so far). Am Psychol1990;45:1304-12.

11 Logigian E, Hefter H, Reiners K, Freund H-J. Doestremor pace repetitive voluntary behavior in Parkinson'sdisease? Ann Neurol 1991;30:172-9.

12 von Galen G, Wing A. In: Smith M, Wing A, eds. Thepsychology of human movement, London: Academic Press,1984:135-51.

13 Wing A, Kristofferson A. Response delays and the timingof discrete motor responses. Percept Psychophys 1973;14:5-12.

14 Schmidt RA, Zelaznick HN, Hawkins B, Frank JS, QuinnJT. Motor output variability: a theory for the accuracyof rapid motor acts. Psychol Rev 1979;86:415-51.

15 Findley LJ, Gresty MA, Halmagyi GM. Tremor, the cog-wheel phenomenon and clonus in Parkinson's disease. JNeurol Neurosurg Psychiatry 1981;44:534-46.

16 Alberts W. A simple view of parkinsonian tremor.Electrical stimulation of cortex adjacent to the rolandicfissure in awake man. Brain Res 1972;44:357-69.

17 Lee R, Stein RB. Resting tremor reset by mechanical per-turbations: a comparison of essential and Parkinsoniantremor. Ann Neurol 1981;10:523-31.

18 Ward CD. In: Swash M, Oxbury J, eds. Clinical neurology.Edinburgh: Churchill Livingstone, 1991:1395-474.

19 Flowers KA. Lack of prediction in the motor behaviour ofparkinsonism. Brain 1978;101:35-52.

20 Benecke R, Rothwell JC, Dick JPR, Day BL, MarsdenCD. Simple and complex movements off and on treat-ment in patients with Parkinson's disease. J NeurolNeurosurg Psychiatry 1987;50:296-303.

21 Stelmach GE, Teulings H-L. Temporal and spatial char-acteristics of repetitive movement. Int Jf Neurosci 1987;35:51-8.

22 Harrington DL, Haaland KY. Sequencing in Parkinson'sdisease: abnormalities in programming and controllingmovement. Brain 1991;114:99-115.

Neurological stamp

Luigi Galvani (1737-98)

Luigi Galvani, the anatomist, physician and physiologistwho discovered 'animal electricity', came from Bologna.The galvanometer, which was invented by Andre Ampere(1775-1836), was named after Galvani as was theprocess of covering steel with a layer of zinc (galvanism).

Galvani observed that static electricity that was storedin a Leyden jar caused dissected frogs' legs to twitch.This occurred if they were placed on metal during athunderstorm. He also noted that when dissected frogs'legs were hung from brass hooks on an iron railing, themuscles contracted when they came into contact with theiron. Galvani concluded that the source of the electricitywas in the muscles and nerves of the animals. His find-ings were later disproved by Alessandro Volta who by1800 had constructed electric batteries consisting of twodifferent metals in an electrolytic salt solution. Voltaestablished that the source of the electricity in Galvani'sexperiment had been two different metals with the ani-mals' body fluids acting as the conducting medium.Galvani's observations were, however, the starting pointof electrophysiology.

Galvani was honoured with this Italian stamp in 1934on the occasion of the First International Congress ofElectro-Radio-Biology (Stanley Gibbons 423, Scott 330).

L F HAAS

POSTFi CENT.

'I 4o'D'st>;iELETTRRADIO -Bgs t .t LwXGrrsA_

4aAntAkiYlk Am .A.Am3Am A LJkAk AM Ak A

.:~~~~~~~*:.f;~~~~~~s*5,S. as~~~~~rr L '~~~~~~~

.;.;INAZ..

DIELTR

IL-. .- . . -0 .. . -M

I I

1084 on M

arch 18, 2020 by guest. Protected by copyright.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.56.10.1078 on 1 October 1993. D

ownloaded from