velocity modulation and rhythmic synchronization of gait in huntington's disease

Velocity Modulation and Rhythmic Synchronization of Gait inHuntington’s Disease

*Michael H. Thaut, PhD, †Regina Miltner, PT, †Herwig W. Lange, MD, *Corene P. Hurt, MS, and†Volker Hoemberg, MD

*Center for Research in Neurologic Rehabilitation, Colorado State University, Fort Collins, Colorado, U.S.A.; and†Neurological Therapy Center, Heinrich-Heine University, Dusseldorf, Germany

Summary: This study analyzed the ability of patients withHuntington’s disease (HD) to modulate gait velocity withoutexternal sensory cues and in response to an auditory rhythmiccue within a frequency entrainment design. Uncued gait pat-terns of 27 patients were first assessed during normal, slower,and faster self-paced walking. During rhythmic trials, metro-nome and musical beat patterns were delivered at rates 10%slower and 10–20% faster than baseline cadence to cue gaitpatterns. After the rhythmic trials, patients were retested atnormal gait speed without rhythm. Gait velocities in the pa-tients with HD were below normal reference values in allranges. Patients were able to significantly (p <0.05) modulatetheir gait velocity during self-paced and rhythmic metronomecueing but not during music. The ability to modulate gait ve-locity was retained regardless of the severity of the disease.Gait velocity declined with an increase in disability and choreascore. The disability score differentiated better between gait

velocity of moderately and severe patients than chorea score.Slowness of gait was significantly correlated only with disabil-ity score and not with chorea. Patients had more difficultyproducing adequate step rates than stride lengths during normaland fast walking speeds. After the rhythmic trials, unpaced gaitvelocity remained significantly (p <0.05) higher than baseline.This carry-over effect was not seen after the uncued trials.Synchronization ability was deficient in all patients, deterio-rated with severity of disease, and was already compromised inpatients with soft disease signs. Rhythmic tracking of musicdeclined more with severity of disease than metronome track-ing. In summary, patients were able to modulate velocity withand without external cues. Velocity adaptations to the externalrhythm in music and metronome were achieved without exactsynchronization between step cadence and rhythmic stimulus.Key Words: Huntington’s disease—Gait—Auditory rhythm—Sensorimotor facilitation.

Huntington’s disease (HD) is an inherited neurodegen-erative disorder transmitted by a dominant autosomalgene with high penetrance on the short arm of chromo-some 4 which shows a lengthening of the (CAG)n-sequence from 10–30 triplets in the normal population toat least 37 in HD.1 The main site of morphologic changesin HD is within the striatum affecting primarily gabaer-gic output neurons but also cholinergic intrinsic neu-rons.2 Motor disturbances in HD characteristically in-volve hyperkinetic and/or dystonic abnormalities. Thesedisturbances share certain features with Parkinson’s dis-ease (PD), a basal ganglia disorder involving the dopa-minergic nigrostriatal system, and are distinct from PD inothers.3 Hyperkinetic features in persons with HD that

are distinct from PD include movement chorea affectingthe whole body, whereas patients with PD show a pre-dominance of rhythmic tremors mainly in the upperlimbs. Shared features between HD and PD includeslowed execution of movements in the upper limbs.3,4

This shared observation of movement impairment sup-ports the importance of the basal ganglia in reference tospeed control as well as initiation and maintenance ofsequential movements irrespective of the neuropatho-logic differences between HD and PD. Several authorshave proposed that slowed movement is a fundamentalmotor deficit in both HD and PD which may be the resultof abnormal pallidocortical projections activating thepremotor cortex and supplementary motor areas4,5 and/orabnormal projections to central pattern generators, forexample, for gait, in the brain stem altering preferredmovement frequencies.6,7

Dysharmonic and hyperkinetic choreic movementsand bradykinesia coexist in patients with HD.4 Several

Received September 4, 1997; revisions received September 21, 1998and May 11, 1999. Accepted May 12, 1999.

Address correspondence and reprint requests to Michael H. Thaut,PhD, Center for Research in Neurologic Rehabilitation, 226 AmmonsHall, Colorado State University, Fort Collins, CO 80523, U.S.A.

Movement DisordersVol. 14, No. 5, 1999, pp. 808–819© 1999 Movement Disorder Society

808

studies have suggested that slowed execution of move-ment is more debilitating than chorea and also appears tohave more predictive value in describing the functionalprogression of the disease.8,9 Indeed, in the more ad-vanced stages of HD, chorea may abate and brady- orakinetic symptoms are exacerbated.10 It has also beenreported that pharmacologic suppression of chorea doesnot lead to functional improvement of movement, andthat neuroleptic treatment does not alleviate slowness ofmovement.11 In contrast to PD, gait has not been ad-dressed in sufficient detail in HD. Koller et al11 reporteddecreased gait velocity in patients with HD which wasattributable to reduced stride length and step cadence. Inaddition to slowness, alterations in HD gait include largelateral sway, wide base of foot position, large variabilityin swing and stance phase durations, difficulty in speedmodulation, and difficulties in initiation of steps. There-fore, it appears that gait abnormalities are a characteristicfeature of the illness which may lead to marked func-tional impairment and disability.

In previous research with patients with PD, our re-search group was able to use rhythmic auditory stimula-tion (RAS) to improve gait patterns in a frequency en-trainment design12 as well as during a 3-week home-based training program.13 In the entrainment design, bothmedicated and unmedicated patients were able to rhyth-mically synchronize their gait patterns with RAS at dif-ferent frequencies and achieve higher gait velocities.This finding is in agreement with findings by Freund andHefter6 that patients with basal ganglia disorder can ac-cess rhythmic stimuli as long as central oscillator fre-quencies critical to tremor production are avoided. In thetraining study, patients with PD improved gait velocity,stride length, cadence, and variability of muscle activa-tion patterns14 significantly over training efforts withoutrhythmic sensory cueing and over untrained control sub-jects. Several researchers have suggested that sensorycues can replace deficient pallidal–cortical projectionsand serve as a signal to the supplementary motor area toactivate the next movement, thus improving bradykinesiaand akinesia in PD.15 Based on current theory thatslowed movement execution in HD is caused by similarbasal ganglia mechanisms as in PD, we sought to inves-tigate the ability of patients with HD to modulate gaitvelocities, and to test if RAS could be an efficient ex-ternal sensory stimulus to cue gait velocities throughrhythmic entrainment.

METHODS

SubjectsA total of 27 patients (13 men and 14 women) with a

genetic diagnosis of HD were examined. All patient data

are summarized in Table 1. Twenty-three patients had adefinite HD diagnosis, four patients were diagnosed asHD gene carriers with three of them showing soft neu-rologic signs (questionable chorea) and one with no neu-rologic signs. The mean age of the study sample was 47± 10.7 years. Triplet scores for CAG-normal sequencesand CAG-HD sequences were known for 24 (normalscores) and 25 (HD scores) patients and are reported inTable 1. Degree of chorea and functional disability werescored on a 0–3 scale based on a modified version16 ofthe Shoulson-Fahn score.17 The mean duration of diseasewas 7.3 ± 3.3 years. The mean chorea score was 1.37 ±0.61, and the mean disability score was 1.28 ± 0.65. Onepatient had no overt choreitic movement (score of 0),three had soft neurologic signs (0.5), and 23 patients hadmild to more severe forms of chorea (1–2.5). Six patientshad no or only slight forms of disability (scores of0–0.75), 10 patients had mild to moderate (0.75–1.5),and 11 had moderate to more pronounced degrees ofdisability (1.5–2.5). In the latter group, three patients hadsevere disability (>2.0) needing almost complete assis-tance in all activities of daily living. Eight patients didnot take any drugs. Nineteen patients were on some formof medication (Table 1).

Experimental Procedure

All patients walked on a 26-m walkway under differ-ent pacing conditions. The experiments were conductedin the gait training area of a neurologic rehabilitationcenter familiar to the patients. Data for gait analysis wereonly used from the middle 20 m of the walkway to allowfor acceleration and deceleration. The first four gait trialswere conducted under the following conditions withoutrhythmic cueing: (a1) normal speed as pretest baseline;(b) slower than baseline; (c) faster than baseline; and (a2)normal speed first posttest. Patients were not given anytempo cues during these trials but asked to walk com-fortably at their preferred normal, fast, and slow walkingpace. In the second set of gait trials, patients walked withRAS and were asked to synchronize their step patternswith the rhythmic cue during the first three trials. Besidesthe instructions, there was no specific training given tothe patients to walk to RAS. The following four condi-tions were implemented: (br) RAS metronome set 10%slower than baseline; (c1r) RAS metronome set 20%faster than baseline (15% or 10%, respectively, if thepatient could not attain the higher step rate); (c2r) RASmusic set faster (like c1r) than baseline; and (a3r) secondposttest at normal speed without RAS. The three mostseverely disabled patients walked with handheld assis-tance from a physiotherapist. All other patients walked

GAIT IN HUNTINGTON’S DISEASE 809

Movement Disorders, Vol. 14, No. 5, 1999

independently. Informed consent was obtained from allpatients prior to the experiment.

RAS Materials

For single pulse pacing, an electronic metronome wasused in which beats per minute (b/min) setting wasmatched to the desired step cadence. For music pacing,prerecorded audiotapes of the instrumental version of afolk song were used. The music had been precomposedand recorded on a synthesizer/sequencer module prior tothe experiment in digital MIDI programming. Musictapes at the exact b/min setting matching the desired gaitcadence were recorded from the sequencer/synthesizerand used during the gait trials. The rhythmic structure ofthe music consisted of accentuated on-beats in 2/4 meterwhich determined the b/min frequency to cue step pat-

terns and additional rhythmic patterns in between theon-beats which were marked by melodic or rhythmicevents subdividing the basic meter in low-integer ratiosof 1:2 and 1:4.

Data Recording and Analysis

Stride parameters were recorded with a computerizedstride analysis system (Infotronic CDG, Eindhoven, TheNetherlands). The system consisted of sensors embeddedin shoe inserts, a portable microprocessor to record dataduring walking, a computer, analysis software, and a datadownloading interface between the microprocessor andthe computer. The software calculated gait velocitybased on optoelectronic start/stop coupling of the micro-processor for 20 m of walking, cadence based on sensorsurface contacts, stride length as a product of velocity

TABLE 1. Patient description (N = 27)

Subject HD parentAge(yrs) Sex

Diseaseduration

(yrs) CAG-HD CAG-N Chorea Disability Medication*

1 Unknown 50 M 12 46 22 2.5 2.33 22, 32, 332 Father 32 M — — — 1.5 2.33 15, 31, 33, 423 Father 31 F 11 57 21 2.0 2.25 11, 12, 434 Mother 43 M 8 46 16 2.0 2.00 22, 33, 425 Father 66 F 11 43 21 1.5 1.83 226 Unknown 52 F 7 44 21 2.0 1.83 13, 237 Father 62 M 9 41 19 1.0 1.67 238 Mother 37 F — 55 29 2.0 1.67 —9 Mother 60 F 9 43 24 2.0 1.58 22, 2310 Mother 59 M 8 2 21 1.5 1.58 2211 Father 60 M 3 42 19 1.0 1.58 —12 Mother 47 F 9 41 18 2.0 1.42 13, 22, 23, 3213 Mother 43 F 10 51 19 1.0 1.42 —14 Mother 40 F 8 51 21 1.0 1.33 2315 Mother 46 M 6 45 17 2.0 1.33 1216 Unknown 68 M 10 42 21 1.5 1.17 13, 14, 2117 Father 35 F 3 54 18 1.0 1.17 4118 Mother 48 F 11 42 17 1.0 1.00 2319 Mother 46 M 11 41 20 2.0 1.00 —20 Mother 47 F 9 45 20 1.0 1.00 1621 Mother 46 F 6 45 27 1.5 1.00 1622 Mother 48 F 8 — — 1.0 0.75 11, 2323 Father 58 F 3 42 19 1.5 0.67 2224 Father 35 M 4 45 18 0.5 0.33 —25 Mother 41 M 1 46 17 0.5 0.25 —26 Father 32 F 0 42 — 0.0 0.00 —27 Mother 38 M 6 44 22 0.5 0.00 —

* Medication Code KeySedatives11 4 Alprazolam12 4 Oxazepam13 4 Lorazepam14 4 Zopiclon15 4 Valproate16 4 Memantine (Akatinol)Benzamides21 4 Cisapride22 4 Tiapride23 4 Sulpiride

Neuroleptics31 4 Levopromazine32 4 Perphenazine33 4 HaloperidolAntidepressants41 4 Hypericin42 4 Amitriptyline43 4 Fluvoxamine

M. H. THAUT ET AL.810


and cadence, and swing symmetry. Furthermore, per-centage deviations between RAS frequency and cadenceas a measure of asynchrony were calculated off-line.Swing symmetry was computed off-line by dividing themean of the shorter swing time over the longer swingtime. Statistical analysis used repeated measures analysisof variance (ANOVA) procedures with planned compari-sons on the change scores between all walking condi-tions compared with the baseline condition. Data werealso analyzed for sample subgroups divided by severityof chorea and disability. Correlations were computed be-tween stride length and cadence, cadence and RAS, cho-rea and disability scores, and chorea and disability scoreswith different gait trials. All descriptive statistics werecomputed as means and standard errors of the mean.

RESULTS

Gait Trials at Baseline Velocity

During the baseline trials, all stride parameters showedabnormal scores below normal age-matched values(Table 2). Normal values reported in the literature (av-eraged between values for men and women in the agerange 40–49 years) are 77.4 ± 7 m/min for velocity, 1.22± .07 m for stride length, and 124 ± 7 steps/min forcadence.18

Patients were divided into three groups by chorea se-verity (0–0.54 I, n 4 4; 1–1.54 II, n 4 14; >1.54III, n 4 9) and three groups by disability severity (0–0.754 I, n 4 6; >0.75–1.54 II, n 4 10; >1.54 III,n 4 11) for an assessment of the effect of severity ofdisease on gait performance. Results summarized inTable 3 indicated that gait velocities decreased with se-verity of the disease. Although no inferential statisticswere applied as a result of a highly unequal distributionof patients across disease groups, the group means illus-trate that the disability score differentiated better be-tween the gait velocity of moderate and severe patientgroups than the chorea score.

The stronger association between disability and gaitwas also evidenced by the fact that chorea score and gait

velocity were not significantly correlated in our studysample (r4 −0.23; p >0.05), whereas disability scoreand velocity showed a significant inverse correlation (r4 −0.56; p <0.05).

Gait Trials at Slower Velocity

During slow self-paced trials, 25 of 27 patients wereable to decrease their gait velocity compared with base-line. No velocity changes were noted in the other twopatients both of whom had severe disability scores. Themean decrease for the whole study sample was 25.5%.Velocity decreases were driven by an average 13% re-duction in cadence and a 12% reduction in stride length.Symmetry ratios worsened slightly during the slow gaittrial (0.88).

During slow metronome trials, 21 of 27 patients wereable to decelerate their gait velocity. The metronometimekeeper at 10% below baseline cadence yielded amean cadence deceleration of exactly 10% accompaniedby a 7% shortened stride length. These reductions re-sulted in a mean velocity decrease of 20% for the wholestudy sample. Unlike during self-paced trials, symmetryratios remained stable at baseline values during slowrhythmic cueing.

When separated by chorea and disability score, thedata showed, for both self-paced and cued trials, that thepatients on all three severity levels of the disease wereable to slow their gait velocity within similar ranges(Table 4).

ANOVA with planned comparisons showed statisti-cally significant mean reductions in velocity (slow self-pacedD[ifference] 4 13.1 ± 2.5 m/min, p4 0.0001/

TABLE 2. Descriptive statistics of gait parameters (means and standard errors)

Velocity (m/min) Stride length (m) Cadence (steps/min) Symmetry (ratio)

Normal baseline 51.1 ± 3.87 1.02 ± .05 95.3 ± 3.7 .91 ± .017Slow self-paced 37.1 ± 3.20* 0.89 ± .06 81.9 ± 3.6 .88 ± .024Fast self-paced 63.0 ± 4.32* 1.17 ± .06 105.3 ± 4.2 .95 ± 0.10Normal retest (posttest 1) 52.1 ± 3.84 1.06 ± .05 98.3 ± 2.9 .92 ± .018Metronome slow 40.9 ± 3.47* 0.95 ± .06 86.1 ± 3.9 .91 ± .015Metronome fast 62.7 ± 4.34* 1.18 ± .06 104.2 ± 3.9 .93 ± .014Music 54.4 ± 4.96 1.07 ± .07 101.2 ± 3.2 .93 ± .013Normal retest (posttest 2) 55.9 ± 4.13* 1.05 ± .06 101.8 ± 3.8 .93 ± .018

* Significantly different at p <0.05 from normal baseline

TABLE 3. Means and standard errors of gait velocity bychorea and disability

Disability Chorea

I. n4 6 65.0 m/min ± 15.4 I. n4 4 75.0 m/min. ± 7.3II. n410 58.2 m/min ± 18.0 II. n414 48.5 m/min ± 18.3III. n411 37.2 m/min ± 22.3 III. n4 9 47.5 m/min ± 18.2



slow metronomeD 4 9.8 ± 1.9 m/min, p4 0.0001) forself-paced and metronome cued trials.

Gait Trials at Faster Velocity

During the faster, self-paced gait trials, 19 patientswere able to increase their gait speed over baseline. Theaverage gain in velocity for the whole study sample was24.3%. Eight patients showed no change; five of thosewere severely disabled and three were moderately dis-abled. The increase in velocity showed a higher contri-bution from lengthened stride (+14%) than cadence

(+10%). Symmetry ratios improved slightly over base-line to 0.95.

An accelerated metronome timekeeper improved ve-locity in 23 of 27 patients over baseline. Two of thenonchanging patients were from disability group III andtwo from disability group II. The mean gain was 15% forstride length and 10% for cadence with a mean velocitygain of 26.0% for all patients. Symmetry ratios improvedslightly over baseline values to 0.93.

With music, 17 of 27 patients were able to increasetheir walking speed over baseline. Of the 10 nonchang-

TABLE 4. Means and standard errors of gait velocity changes by disability and chorea (in percent)

Slow self-paced Fast self-paced Slow metronome Fast metronome Music

I. Disability (n 4 6) −24.0 ± 8.1 27.0 ± 13.9 −13.5 ± 5.8 23.0 ± 16.0 11.4 ± 10.5Chorea (4) −26.0 ± 10.8 20.0 ± 13.1 −13.5 ± 6.6 21.7 ± 18.2 7.3 ± 11.4

II. Disability (10) −23.6 ± 13.5 18.6 ± 16.2 −17.7 ± 18.9 25.5 ± 12.9 15.9 ± 12.3Chorea (14) −15.0 ± 11.3 21.2 ± 15.3 −14.3 ± 19.1 27.9 ± 17.3 8.3 ± 12.5

III. Disability (11) −28.2 ± 20.2 21.6 ± 12.3 −19.7 ± 16.5 28.1 ± 21.6 −0.4 ± 10.9Chorea (9) −37.1 ± 12.8 22.6 ± 17.9 −23.7 ± 13.6 25.1 ± 17.7 8.5 ± 6.1

FIG. 1. Frequency entrainment (means and standard errors of percentage deviations of step cadence from set RAS rate).



ing patients, seven had a severe disability score, twomoderate, and one mild. Seven patients actually de-creased and three patients showed no change from base-line velocity. Musical pacing resulted in considerablylower velocity gains than self-paced and metronome cue-ing. The mean gain was 9.3% with a much larger con-

tribution from cadence gains (+7.5%) than stride length-ening (+1.9%). The symmetry ratio in music also im-proved slightly over baseline to 0.93.

Similar to the slow gait trials, the patients on all threeseverity levels of the disease were able to accelerate theirgait velocity within comparable ranges during the self-

FIG. 2. (A, B, and C) Scatterplots ofRAS (beats/min) versus cadence (steps/min) for slow metronome, fast metro-nome, and music cueing.



paced and metronome trials. During music, the ratesof velocity increase were also similar across dis-ease groups yet smaller compared with the self-pacedand the metronome condition. Unlike all the other con-ditions, however, patients in disability group III were notable to increase their gait velocity with music at all(Table 4).

ANOVA with planned comparisons showed signifi-cant velocity gains for self-paced and metronome cuedtrials (self-pacedD 4 11.9 ± 1.8 m/min, p4 0.0001/metronomeD 4 12.7 ± 1.9 m/min, p4 0.0001). Ve-locity gains for music were nonsignificant (D 4 5.3 ±2.6 m/min).

Short-Term TrainingShort-term training effects were tested by comparing

pre- versus posttest data between (1) baseline and thefirst posttest after the self-paced gait trials and (2) base-line and the second posttest after the rhythmically cuedgait trials. It is important to note that both posttestswere performed after the fast self-paced and the fastrhythmically cued walking conditions, respectively, toequalize potential carry-over effects. ANOVA withplanned comparisons showed no significant differencebetween baseline scores and the first posttest after theuncued gait trials (velocity:D 4 +0.96 ± 1.1 m/min, p4 0.3879; stride length:D 4 +0.01 ± 0.02 m, p40.7111; cadence:D 4 +2.83 ± 2.2 steps/min, p40.2137). However, significant differences were found for

velocity between baseline and second posttest after theRAS trials (D 4 4.8 ± 1.9 m/min, p4 0.0243). Theaverage improvement rate for velocity was 9.5%. Im-provements in stride length and cadence after RAS werehigher than after the uncued gait trials yet statistically notsignificant (stride length:D 4 +0.03 ± 0.03 m, p40.4332; cadence:D 4 +6.52 ± 3.8 steps/min, p40.1055).

When separated by disability and chorea score, groupsbenefitted differently from RAS cueing at posttest. Di-vided by disability, group I improved by an averageof 9.5 ± 2.5%, group II by 15 ± 3.9%, and group IIIby only 4.2 ± 2.7%. When separated by chorea score,group I improved by 12 ± 5.9%, the moderate group(II) by 9.9 ± 4.5%, and the severe chorea group (III)by only 2.25 ± 1.8%. Grouped by chorea, velocity post-test improvements were seen in four of four level I pa-tients, 10 of 14 level II patients, and four of nine level IIIpatients. For disability, five of six level I patients, sevenof 10 level II patients, and six of 11 level III patientsimproved.

Frequency EntrainmentStrength of frequency entrainment was measured as

percentage deviations of step cadence from RAS rate(Fig. 1). Complete 1:1 synchronization between step andRAS rate would result in a 0% deviation score. In aprevious rhythmic gait synchronization study, asyn-chrony scores were 0% for healthy elderly, 0.47% for

FIG. 2. (Continued).



patients with PD on medication, and 2.7% for patientswith PD off medication.12 In the level I groups with softsigns of chorea and disability, music provided the small-est asynchrony score (chorea: 1.1 ± 0.5%; disability: 0.8

± 0.7%). In all other groups for all cueing conditions,frequency deviations were higher and asynchrony scoresgot progressively worse with severity of disease. Musicgenerated a larger asynchrony rate than metronome in

FIG. 3. (A, B, and C) Scatterplots ofstride length (m) versus cadence (steps/min) during normal, slow, and fast un-cued gait trials.



the most severe disability group showing a reversal ofthe initial advantage of music for the level I patients.Asynchrony scores were similar between chorea anddisability groups across severity levels with two excep-tions on level III: during slow metronome cueing, severechorea resulted in higher asynchrony than severe dis-ability, and during music cueing, the high disability pa-tients were more asynchronous than the severely choreicpatients.

Although actual asynchrony scores were high, the cor-relation coefficents between RAS and cadence werehighly significant indicating a strong positive relation-ship between change in cueing frequency and step fre-quency: r4 0.78 (p4 0.001) for slow metronome, r40.79 (p4 0.001) for fast metronome, and r4 0.71 (p40.001) for music (Fig. 2). Furthermore, the patients’ abil-ity to synchronize their cadence to the metronomeshowed a strong significant correlation with subsequentimprovement in gait velocity during the posttest afterRAS cueing (r4 −0.60; p4 0.025). Synchronizationability to music showed a small nonsignificant correla-tion with subsequent gait improvement (r4 −0.15).

Stride and Cadence Analysis

Scatterplots of stride length and step cadence undernormal, slow, and fast conditions are depicted in Figure3. Coefficients for normal and fast conditions were sta-tistically significant (normal: r4 0.511, p4 0.009; fast:

r 4 0.493, p4 0.012). The slow walking condition didnot yield a significant relationship between deceleratedcadence and reduced stride length (r4 0.191; p 40.370).

DISCUSSION

The data in this study confirm previous research11

showing that slow gait parameters are a prominent fea-ture of HD gait. The mean velocity of 51.1 m/min in ourstudy sample was slightly higher than the mean of 45.6m/min reported previously.11 However, our group alsoincluded several patients with soft to mild disease signsto assess a wide range of gait deficits relative to diseaseprogression.

Our study extends previous gait data in HD relative togait parameters during velocity modulation and rhythmicsynchronization. First, most patients were able to ad-equately modulate gait velocity with and without sensorycueing. The percentage range of velocity modulation wassimilar across all levels of severity of disability and cho-rea. However, among the patients who could not modu-late gait speeds, there was a higher representation of thesevere disability group. Overall, adapted slow and fastwalking speeds stayed below normal age-expected val-ues as reported in the literature.18 Compared with healthysubjects, our study sample walked on the average 33.9%slower during normal walking speed, 28.2% slower dur-ing slow walking, and 31.4% slower during fast walking.

FIG. 3. (Continued).



Differential contributions of stride length and cadencemodulation were observed during walking speeds. Dur-ing normal walking speed, cadence was proportionallymore below normal reference values (−24%) than stridelength (−16%). During slow walking, both parameterswere at a similar rate of decrease compared with normaldata (normal stride length: −16%; slow stride length:−15%). During fast walking, the decrease in cadencebelow normal values (−26%) was again more pro-nounced than the shortened stride length (−10%), indi-cating a greater difficulty in producing adequate steprates than adequate stride lengths at normal and acceler-ated speeds. These findings are somewhat in contrast topatients with PD whose gait bradykinesia is commonlycharacterized by stride lengths that are proportionallymore reduced than step rates.

Second, disease progression had a clear impact on gaitparameters. Mean values ranged from 75 m/min and 65m/min, respectively, for the milder groups to 47 m/minfor most severe chorea patients and 35 m/min for thepatients with the highest disability scores. The disabilityscore differentiated better between gait performance ofthe moderate and more severe groups than the choreascore. This observation was further substantiated by asignificant correlation between disability score and gaitvelocity and a nonsignificant correlation coefficient forchorea score and velocity. Thus, slower gait appeared tobe more associated with a functional progression of thedisease than with chorea in our study sample. The non-significant correlation between chorea and gait velocityis in line with earlier research findings.11 The significantcorrelation between disability and gait velocity is in con-trast with other recent research19 which found gait ve-locity and disease severity statistically unrelated. How-ever, the respective research findings may not be imme-diately comparable because assessment instruments andpatient diagnostics (for example, disease duration, distri-bution of patient disabilities across the spectrum of dis-ease severity) were different between the two studysamples.

Third, stride length and cadence, which are highlycorrelated in normal gait,20 showed smaller than normalyet statistically significant values for normal and fastwalking speeds. During slow walking, the directionalrelationship between reduction in stride length and de-celerated cadence disappeared in our study sample. Wemay conclude that the ability to proportionally scalestride length and cadence was still functionally evident inour patients in the normal and fast but not the slowranges.

The lower performance with music cueing across alllevels of disease severity may indicate that complex au-

ditory pattern perception as motor timing cues may bealready affected early on in the disease process. Thecomplete unresponsiveness of the most severe disabilitygroup to music cueing further suggests that disease pro-gression leads to an increasing deterioration of musicprocessing as motor timing cue. Although direct mea-sures of cognition were not available for this studygroup, a strong cognitive component in the progressiveunresponsiveness of patients with HD to music may beimplied by the fact that high correlations between dis-ability and dementia with prolonged and exacerbated dis-ease states have been reported in the research litera-ture.21,22The fact that the most severe chorea group didnot show the same negative response to music furthersupports the argument for a role of cognitive deteriora-tion in the lower performance with music. Cognitivedeficits as part of the neuropathology in HD have beenattributed to degenerative processes in the nigrostriataldopaminergic system and pathologic changes in the fron-tostriatal system.23–25Unlike patients with HD, patientswith PD can usually use rhythmic–musical cues well formotor timing and movement synchronization.12–14,26

Thus, the inability by the patients with HD to extractuseful motor timing cues from highly complex musicalpatterns may be the result of the more severe cognitiveinvolvement of patients with HD resulting from patho-logic changes in frontostriatal circuitry.27

An important finding is a carry-over effect of en-hanced gait velocity during the uncued posttest only aftermetronome cueing and not after self-paced velocitymodulation. Strong entrainment effects of auditoryrhythm on gait performance have been reported previ-ously in PD.12,26A possible explanation emerged in ourstudy group by the significant correlation between met-ronome synchronization ability and higher gait velocityafter metronome cueing. The strong association betweenexternal rhythmic entrainment and temporary mainte-nance of higher gait velocities after removal of the ex-ternal rhythm may suggest a mechanism in which exter-nal step-to-beat frequency coupling, mediated by the“magnet effect” of the metronome, may have temporarilycreated stronger time traces than uncued walking to resetintrinsic gait oscillators.

Tempo and time recall have been shown to be surpris-ingly stable in music and rhythm.28 In a previous studywith patients with PD, we also saw accurate gait temporeproduction 24 hours after rhythmic entrainment.14 Inthis study, the intrinsic “resetting” of gait frequenciesafter brief rhythmic entrainment is remarkable becausethe patients were simply asked to walk at their normalpreferred speed. Simple carry-over from accelerated gait



tempo cannot explain this performance because pa-tients also walked faster immediately preceding the firstposttest.

However, unlike the case in patients with PD, the highvariability in frequency entrainment indicated that exactphase and period matching to the external cue was highlyimpaired in patients with HD. Using previously estab-lished oscillatory entrainment models, the data are actu-ally characteristic of a weakly coupled oscillator systemwith a strong presence of noise factors in the couplingprocess.29–31This inability was already noticeable in pa-tients with soft and mild disease signs and deterioratedsignificantly with disease progression. The observationof timing deficits is further emphasized by the fact thatthe person without any neurologic signs (genetic diag-nosis only) already showed high asynchrony scores (met-ronome slow: 3.1%, metronome fast: 6.9%, music:0.9%). Thus, impaired performance in sensorimotor syn-chronization tasks may be a disease sign before overtneurologic signs become manifest. However, most pa-tients with HD could modulate their walking speed in thecued directions, especially with the metronome, whichindicates that, although precise timing of synchronizationmay already be disturbed at an early disease stage, thegeneral mechanisms of rhythmic entrainment remainintact.

The results of this study provide further insight intospecific gait deficits associated with HD as well as firstevidence that rhythmic facilitation can improve locomo-tor function in patients with HD after a short trainingperiod contingent on stimulus condition and diseasestate. However, the long-term feasibility of RAS as arehabilitative technique to facilitate gait ability in HDremains to be determined.

Acknowledgment: This research was supported in part by agrant from the Deutsche Forschungsgesellschaft, Sonderforsch-ungsbereich 194 to Drs. Thaut and Hoemberg (DFG, SpecialResearch Section 194).

REFERENCES

1. The Huntington’s Disease Collaborative Research Group. A novelgene containing a trinucleotide repeat that is expanded and un-stable on Huntington’s disease chromosomes.Cell 1993;72:963–971.

2. Lange HW. Quantitative changes of telencephalon, diencephalon,and mesencephalon in Huntington’s chorea, postencephalitic, andidiopathic parkinsonism.Verhandlungen der Anatomischen Ge-sellschaft1981;75:923–925.

3. Hefter H, Hoemberg V, Lange HW, Freund HJ. Impairment ofrapid movement in Huntington’s disease.Brain 1987;110:585–612.

4. Thompson PD, Berardelli JC, Rothwell BL, Day JPR, Benecke R,Marsden CD. The coexistence of bradykinesia and chorea in Hun-

tington’s disease and implications for theories of basal gangliacontrol of movement.Brain 1988;111:223–244.

5. Priori A, Berardelli A, Inghilleri M, Polidori L, Manfredi M. Elec-tromyographic silent period after transcranial brain stimulation inHuntington’s disease.Mov Disord1994;9:178–182.

6. Freund JH, Hefter H. The role of basal ganglia in rhythmic move-ment.Adv Neurol1993;60:88–92.

7. Marsden CD. What do the basal ganglia tell premotor corticalareas? In:CIBA Foundation Symposium 132.Chichester: Wiley,1987:282–300.

8. Bradshaw JL, Phillips JG, Dennis C, et al. Initiation and executionof movement sequences in those suffering from and at-risk ofdeveloping Huntington’s disease.J Clin Exp Neuropsychol1992;14:179–192.

9. Girotti F, Marano R, Soliveri P, Geminiani G, Scigliano G. Rela-tionship between motor and cognitive deficits in Huntington’s dis-ease.J Neurol1988;235:454–457.

10. Hayden MR.Huntington’s Chorea.Berlin: Springer, 1981.11. Koller WC, Trimble J. The gait abnormality of Huntington’s dis-

ease.Neurology1985;35:1450–1454.12. McIntosh GC, Brown SH, Rice RR, Thaut MH. Rhythmic audito-

ry–motor facilitation of gait patterns in patients with Parkinson’sdisease.J Neurol Neurosurg Psychiatry1997;62:22–26.

13. Thaut MH, McIntosh GC, Rice RR, Miller RA, Rathbun J, BraultJM. Rhythmic auditory stimulation in gait training for Parkinson’sdisease patients.Mov Disord1996;11:193–200.

14. Miller RA, Thaut MH, McIntosh GC, Rice RR. Components ofEMG symmetry and variability in parkinsonian and healthy elderlygait. Electroencephalogr Clin Neurophysiol1996;101:1–7.

15. Cunnington R, Iansek R, Bradshaw JL, Phillips JG. Movement-related potentials in Parkinson’s disease: presence and predictabil-ity of temporal and spatial cues.Brain 1995;118:935–950.

16. Lange HW, Strauss W, Hassel PC, Woeller W, Tegeler J. Lang-zeittherapie bei Huntington-Kranken.Psycho1983;5:286–290.

17. Shoulson I, Fahn S. Huntington’s disease: clinical care and evalu-ation.Neurology1979;29:1–3.

18. Oeberg T, Karsznia A, Oeberg K. Basic gait parameters: referencedata for normal subjects, 10–79 years of age.J Rehabil Res Dev1993;30:210–223.

19. Hausdorff JM, Cudkowicz ME, Firtion R, Wei JY, Goldberger AL.Gait variability and basal ganglia disorders: stride-to-stride varia-tions of gait cycle timing in Parkinson’s disease and Huntington’sdisease.Mov Disord1998;13:428–437.

20. Murray MP. Gait as a total pattern of movement.Am J Phys MedRehabil1967;9:290–333.

21. Hoemberg V, Hefter H, Granseyer G, Strauss W, Lange HW,Hennerici M. Event-related potentials in patients with Hunting-ton’s disease and relatives at risk in relation to detailed psychom-etry. Electroencephalogr Clin Neurophysiol1986;63:552–569.

22. Brandt J. Cognitive impairment in Huntongton’s disease: insightsinto the neuropsychology of the striatum. In: Boller F, Grafman J,eds.Handbook of Neuropsychology,vol 5. Amsterdam: Elsevier,1991.

23. Ginovart N, Lundin A, Farde L, et al. PET study of the pre- andpost-synaptic dopaminergic markers for the neurodegenerativeprocess in Huntington’s disease.Brain 1997;120:503–514.

24. Hanes KR, Andrewes DG, Pantelis C. Cognitive flexibility andcomplex integration in Parkinson’s disease, Huntington’s disease,and schizophrenia.J Int Neuropsychol Soc1995;1:545–553.

25. Sprengelmeyer R, Canavan AGM, Lange HW, Hoemberg V. As-sociative learning in degenerative neostriatal disorders: contrasts inexplicit and implicit remembering between Parkinson’s and Hun-tington’s diseases.Mov Disord1995;10:51–65.

26. Richards CL, Malouin F, Bedard PJ, Cioni M. Changes induced byL-dopa and sensory cues on the gait of parkinsonian patients. In:Woollacott M, Horak F, eds.Posture and Gait: Control Mecha-nisms,vol II. Eugene, OR: University of Oregon Books, 1992:126–129.

27. Baeckman L, Robins-Wahlin TB, Lundin A, Ginovart N, Farde L.



Cognitive deficits in Huntington’s disease are predicted by dopa-minergic PET markers and brain volumes.Brain 1997;120:2207–2217.

28. Levitin DJ, Cook PR. Memory for musical tempo: additional evi-dence that auditory memory is absolute.Perception & Psychophys-ics 1996;58:927–935.

29. Thaut MH, Kenyon GP, Schauer ML, McIntosh GC. The connec-tion between rhythmicity and brain function: implications for

therapy of movement disorders.IEEE Eng Med Biol Mag1999;18:101–108.

30. Thaut MH, Miller RA, Schauer ML. Multiple synchronizationstrategies in rhythmic sensorimotor tasks: period vs phase correc-tions.Biol Cybern1998;79:241–250.

31. Thaut MH, Tian B, Azimi MR. Rhythmic finger tapping to cosinewave modulated metronome sequences: evidence for subliminalentrainment.Human Movement Science1998;17:839–863.



velocity modulation and rhythmic synchronization of gait in huntington's disease

Documents