Journal of Neurology, Neurosurgery, and Psychiatry 1991;54:1093-1098

Articulatory deficits in Parkinsonian dysarthria:an acoustic analysis

H Ackermann, W Ziegler

AbstractTwelve patients with idiopathic Parkin-son's disease had acoustic speech analysisof sentence utterances to provide infor-mation on speech tempo and accuracy ofarticulation. As a measure of rate ofspeech the duration of opening-closingmovements during articulation was

determined from speech wave variables.The intensity of sound emission duringarticulatory closure as required for stopconsonant production, for example, /p/,/t/, /k/, was used as an index of the degreeof closure. Speech tempo was not signifi-cantly different from normal. The patien-ts, however, had a reduced capacity ofcompleting articulatory occlusion. Thiswas interpreted as reflecting a reductionin movement amplitude of thearticulators. Articulatory "undershoot"was not uniform but influenced by lin-guistic demands in that the closuresassociated with a stressed syllable were

performed at the expense of unstressedones. Furthermore, switching betweenopening and closing movements of thearticulators in sentence productionseemed undisturbed. These resultsindicate that motor planning of speechdiffers from arm movement control.

Department ofNeurology, Universityof Tiibingen, GermanyH AckermannClinicalNeuropsychologyGroup, MunicipalHospital,Bogenhausen, Munich,GermanyW ZieglerCorrespondence to:Dr Ackermann, Departmentof Neurology, University ofTubingen, Hoppe-Seyler-Str3, 7400 Tubingen, GermanyReceived 19 November 1990and in final revised form10 April 1991.Accepted 26 April 1991

The motor deficits of hypokinesia (reducedrange of simple limb movements with con-

sequent target undershooting) and bradykin-esia (slowness ofmovements) are characteristicof Parkinson's disease (PD).' In speech, per-ceptual as well as acoustic signs of hypokinesiain articulation have been widely presumed.2"Moreover, kinematic analyses have provideddirect evidence of reduced and slowedarticulatory movements particularly of the lipsand the jaw in patients with PD.'2-'7 Studies oflimb motor control have shown not only slow-

ness of single movements but also a delay inexecuting complex motor sequences due toproblems in "switching" from one motorprogramme to another. The impairment inrunning motor programmes seems to be closelyrelated to the degree of clinical bradykinesia.'Whether this principle applies to speech con-

trol as well is not known.Prosodic features such as linguistic or

emotional stress influence speech movementsin sentence production. In normal speech theproduction of stressed syllables is associated

with increased articulatory effort."8 This raisesthe question of how far patients with PD areable to produce such extra effort, suggestingthat sentence utterances rather than simplerepetitive oral movements ought to be used ininvestigating Parkinsonian dysarthria.We used acoustic data for the description of

hypokinesia and bradykinesia within thespeech motor system of patients with PD.Acoustic analysis was based on variablescharacterising the sound wave emitted duringspeaking and allows limited inferences aboutarticulatory features. A sentence repetitionparadigm was used to approximate most closelythe conditions ofnatural speech to ascertain theeffectiveness of certain linguistic constraints,that is, sentence accent, on speech movementsin Parkinsonian dysarthria.

Subjects and methodsSubjectsTwelve patients with idiopathic Parkinson'sdisease (seven men, five women, 47 to 76 years,median 62 years) participated in this study.The clinical data are presented in the table. Allpatients were on drug therapy; none hadpositive signs of dementia at psychopatho-logical examination. The range ofoverall motorimpairment at the time of speech evaluationextended from slight akinetic-rigid states tosevere disability. Two patients suffered fromlevodopa induced on/off fluctuations; in both,the speech recordings were obtained during theoffstate. The control group included 12 "youngnormals" (seven men, five women, 19 to 35years, median 29 years) and 12 "elderly nor-mals" (eight men, four women, 42 to 64 years,median 53 years).

Speech examinationAll subjects underwent speech examination in asound attenuated room. Recordings were madeon a Nagra IV S tape recorder with a Senn-heiser microphone positioned about 20 cm infront of the subject. Speech materials includedsentences of the form "Ich habe /geCVCe/gehort" ("I have heard . . .") with C = /p/, /t/,/k/ and V = /i/, /y/, /u/, /a/. The formula/geCVCe/ together with the lists of consonantsand vowels to substitute for C and V enabledthe use of 12 target words-for example,"getite", "gepype", "gekuke". These "words"represent regular phonological forms of stan-dard German, so that to a certain extent thesentences fulfil natural speech demands. The

1093

on August 16, 2020 by guest. P

rotected by copyright.http://jnnp.bm

j.com/

J Neurol N

eurosurg Psychiatry: first published as 10.1136/jnnp.54.12.1093 on 1 D

ecember 1991. D

ownloaded from

Ackermann, Ziegler

Table Clinical data ofpatients with Parkinson's disease

Duration Motor signs (limbs)of disease CT-scan

Patient Age Sex (years) akinesia rigidity tremor Depression (atrophy) Medication

1 47 m 12 (+) - - - - Id,am2 60 m 4 (+) - + - (+) ld,ac3 69 f 1 + (+) - - - Id4 63 m 5 (+) (I)- (±) - ld,am5 73 f 10 ( ) (+) - - (+) Id,am,br6 58 f 10 (+) (+) (+) - - ld,br,ac7 57 m 9 + + (+) (I)(+) Id,br8 71 m 2 + (+) + - - ac9 66 m 12 + (+) - - - Id,br10 52 f 8 + (+) - - - ld,br11 60 f 6 + + + - - - ld,br,ac12 76 m 13 + + + + (+) (+) ld,br

Id = levodopa, br = bromocriptine, am amantadine, ac = anticholinergic drugs- = missing, (±) = slight, + = moderate,+ + = distinct.

test sentences were read by the examiner in a

quasi-randomised order and repeated by thesubjects at a suitable rate and loudness level.The entire test, consisting of 12 test sentences,was run twice by all speakers.

Perceptual evaluationTo obtain perceptual judgments the patients'recordings were evaluated by five experiencedspeech therapists. The evaluation protocolincluded seven-point ratings of variables suchas a patient's overall articulatory and vocalimpairment and his or her speech tempo.

Acoustic evaluationSound pressure level (SPL) contours Speechsignals were digitised at a sampling rate of20 kHz after low-pass filtering at 9 kHz andwere fed into a LSI 11/73 for further process-ing. Sound pressure level (SPL) contours werecalculated for each of the 24 sentences. TheSPL was determined every 3-2 ms over a 1-2-6ms window and the resulting contour wassmoothed by a 30 Hz low-pass filter. Figure 1

provides examples for the test sentence "Ichhabe gepipe gehort" as spoken by a normalsubject and a patient with PD. During com-plete closure (as required for the stop conson-ant sequence /b/, /g/, /p/, /p/, /g/ of the testsentence of figure 1) intensity is normally

Figure 1 Sound pressurelevel (SPL) contours oftest sentence "Ich habegepipe gehort" obtainedfrom normal speaker(upper half) andfrompatient with Parkinson'sdisease (lower half).Trace at top showsoscillogram of test sentenceas spoken by normalsubject. Vertical linesmark syllable boundaries,triangle indicates onset ofstressed vowel. Aboveoscillogram part of testsentence "ich habe gepipegehort" is depicted.Reversed "e" representsphonetic symbol denotingshort and centralised e-sound.

SPL (dB)

beig ap i peg

Time (s) *.

reduced to the level of the tape noise, whereasvocal tract opening for a vowel is associatedwith an intensity peak. Incomplete closureresulting from reduced extension of man-dibular, labial, or lingual movements or fromreduced occlusive force may be expected toresult in an increase of sound pressure duringstop realisation (as seen in the lower half offigure 1) as acoustic energy is still emitted fromthe mouth.'9 SPL contours allow calculation ofsyllable durations and of the amount of soundemission during closure, and moreover reflectdynamic aspects of sound pressure changeduring closing and opening movements.Mean syllable duration (MSD) The dura-tions of four syllables of each test utterancewere measured: /be/, /ge/, /CV/, /Ce/ (such as/be/, /ge/, /pi/, /pe/ in the example of figure 1).Syllable duration was determined by measur-ing the time period ofa syllabic peak in the SPLcontour. Specifically the duration of a syllablewas defined as the time interval between the twopoints where the SPL contour crossed the 5%threshold between the SPL maximum corres-ponding to the syllabic peak and the minimaleft and right of it.'9 The four syllable durationsof a sentence were averaged and the medianvalue ofthe 24 test utterances of the subject wascalculated. We called this measure the meansyllable duration (MSD). As the syllables usedin the computation ofMSD consisted of a stopconsonant and a vowel each, MSD may beconsidered here as reflecting the time interval ofan opening-closing cycle. By virtue of itsdefinition MSD provides an index of speechtempo which is unaffected by pausing, syllableiteration, or by difficulties in word-finding orspeech initiation problems that may also bepresent in PD.20Intensity during closure (IDC) For the twostop consonants preceding and following the(stressed) target vowel of a test word (forinstance the two /p/s in the example of figure 1)intensity minima were computed. As men-tioned above this variable measures the acous-tic quality of stop consonant production and isparticularly sensitive to incompleteness ofclosure (undershooting). Median values ofintensity minima were determined over all 24sentences of a subject for each of the pre- andpost-stressed closures, respectively. We calledthe resulting measure the intensity duringclosure (IDC).

1094

on August 16, 2020 by guest. P

rotected by copyright.http://jnnp.bm

j.com/

J Neurol N

eurosurg Psychiatry: first published as 10.1136/jnnp.54.12.1093 on 1 D

ecember 1991. D

ownloaded from

Articulatory deficits in Parkinsonian dysarthria: an acoustic analysis

Figure 2 Perceptualjudgments offive speechpathologists for patients'overall impairment ofvoice (abscissa) andarticulation (ordinate)given in mean ratingscores.

c0

.3

4-

2-

0-

Mean rating scores (5 Listeners)

0 = Normal6 = Severe

0

0

0

*0 00

*0

0

0

0

2Voice

4

Transition quotient (TQ) The period of fullopening during vowel production may withsome justification be considered a steady state,characterised by temporarily stable acousticconditions, in contrast with the intervals ofopening and closing where acoustic conditionsare subject to fast changes. The relativedurations of these periods within a consonant-vowel-consonant cycle contain dynamic infor-mation on the underlying movements. Inparticular, the relative duration of the steadystate interval may be considered an index ofthetime period when the switch between theopening and closing gestures occurs. We con-sidered the period between the 95% thresholdsof the rising and falling SPL contour thesteady-state portion and the remainder of thefull syllabic cycle (the two 5%-95% slopes) asthe movement portion. A transition quotient(TQ) was computed by dividing the movementportion of a syllable by the full syllable dura-tion. This quantity was determined for thetarget CV syllable in each sentence and againthe median value over all utterances ofa subjectwas used to characterise his or her perfor-mance.

considered normal and three were more severe-ly dysarthric (numbers 10, 11, 12). Impairmentof vocal variables-for example, vocal inten-sity, pitch, and intonation-was judged moresevere than articulatory deficits, although thedifference was only marginally significant (Wil-coxon, p < 0.05). Considering articulationalone, three patients were judged as undistur-bed and two further patients obtained meanscores lower than 1.Speech tempo The mean syllable durationsobtained in the two normal groups and in thepatients with PD respectively are shown infigure 3. Nonparametric testing (Mann-Whitney) revealed that the young normalgroup spoke significantly faster than the elderlynormal group (p < 0-01). No significant dif-ferences were obtained between the patientswith PD, however, and either the young nor-mals or the elderly normals (p > 0-1). The twoclinically most impaired patients (numbers 11and 12 in the table) presented with increasedMSD, whereas three further patients wereslightly faster than the elderly normals, show-ing mean syllable durations within the lowerrange ofthe young normals. In accordance withthese measurements, perceptual evaluationindicated slowed speech in the former two andaccelerated speech in the latter three patients.No correlation between degree of severity asjudged by perceptual evaluation and meansyllable duration could be detected (Spearman,r = 026, p > 0-1).Articulatory precision Unlike mean syllableduration, intensity during closure (IDC) wassubstantially disturbed in the patients as agroup. Figure 4 shows that the IDC values ofthe post-stressed consonant were significantlyincreased in the patients compared with eitherofthe two normal groups. A significant correla-tion between perceived severity of overallarticulatory impairment and IDC was found(Spearman, r = 0 64, p < 0-05). Whereas thepatients who were judged as having little or no

ResultsPerceptual evaluation Figure 2 presents theresults of perceptual judgments of the patients'overall impairment of voice (abscissa) andarticulation (ordinate). The data reveal thatmost cases were judged to be of mild tomoderate severity. One patient (number 1) was

Figure 3 Mean syllabledurations (96 syllables ineach subject) ofyoungnormals (YN), elderlynormals (EN), andpatients with Parkinson'sdisease (PD). Durationsare shown on logarithmicscale.

NS

NS

0

** (p<0*O1)

0iYN EN

Group

L-

3C

,-

c

0

*** (P<0-001)

** (p<0-01)I

00

0

NS

I

00

o8

PDENr:rmnI

a Figure 4 Intensity during closure (IDC)-that is,O minimal sound pressure level (SPL) during stop

consonant production-in young normals ( YN), elderly§3 normals (EN), and patients with Parkinson's disease(PD). For each subject median value determinedfrom

,---- post-stressed closures over all 24 sentences is depicted,PD reflecting articulatory accuracy of stop consonant

production.

400-

0(0cn- 300-0

.2

100

1095

on August 16, 2020 by guest. P

rotected by copyright.http://jnnp.bm

j.com/

J Neurol N

eurosurg Psychiatry: first published as 10.1136/jnnp.54.12.1093 on 1 D

ecember 1991. D

ownloaded from

Ackermann, Ziegler

Figure 5 Difference ofintensity during closure(IDC) between stopspreceding andfollowingstressed target vowel inyoung normals (YN),elderly normals (EN),and patients withParkinson's disease (PD).Median values over all 24sentences of each subjectare depicted.

40-

30

-o

'o 20-

C20

._AX 1

-: O

-10.

NS

--I

YN

articulatory impairmer

within the normal ra

increased values were o

and severe cases.

Production of sentence a

the IDC difference betof the post-stressed ai

syllable in the three esalmost all normal speakhigher in the post-stresposition. This corresp

the realisation of stressbe associated with an

activity.2' In the patiedifferences were positistence productions follpattern. The resultingsignificantly enhanced ithe normal subjects, whefforts to produce sent

expense of an overproIof the post-stressed stoTransition quotienttransition quotient (TQdifferences between pai

mal speakers (figure 6).'enhanced movementnormals and patientsyounger normals. Theswere not significant. 'complement ofTQ ma)of the faculty of sv

movement directions (cresults show that the 1had no particular probl

Figure 6 Transitionquotients (TQ) -that is,movement portion ofsyllable divided by fullsyllable duration-inyoung normals (YN),elderly normals (EN),and patients withParkinson's disease (PD).Median values over allutterances of each subjectare given.

0.9-

c -

0

c 0-8-.2

-

0-7-

YN

Discussion** (p<0-o1) Speech motor deficits can be described on the

basis of perceptual, acoustic, or physiological**(p<O(EMG, dynamic, or kinematic) measures. Per-

o ceptual assessment of dysarthria may cover8 those aspects of a patient's speech impairment

that are most relevant from the viewpoint of his0

or her communication needs. Descriptions8 based on auditory parameters, however, are

generally not reliable22 and are not easily inter-9 8 pretable in terms of underlying pathological9 o mechanisms.23 In contrast, physiological

assessment techniques such as the x raymicrobeam, 12-14 strain-gauge measure-ments,'5 16 or electromagnetic articulography,24

EN PD yield direct observations of speech motor pat-Group terns. Yet data obtained by these techniques are

inherently difficult to interpret as betweenit assumed IDC valuess2 As an exam-mge,nd cnsideablysubject variation iS considerable.Asaexmnge, and considerably ple, the data reported by Forrest et all7 may be

mentioned. Their measurements of jaw open-ing amplitudes showed that their normal:ccent Figure 5 depicts subjects varied between values characteristic of

ween the stop consonant their most severe examples of Parkinsonismnd that of the stressed and values two times as high as the maximumxcperimental groups. In obtained in their patients, although jaw;ers the IDC:values wereseds tha InCthuessrese movement amplitudes were among themeasures that showed the highest group dif-onds to expectations as ferences between the patients and the normalis generally assumed to group in their study. Not unexpectedly there-increase in muscular fore there is a notoriously low correlationbetween movement data and perceived degreesve, suggesting their sen- ** * 26

[owedasuggstin te en- of dysarthric impairment. Moreover, thelowed a normal accent

difference values were recording of speech movements by physio-the atients relative to logical techniques usually requires a high

amount of cooperation of the subject if reliable:eich maymean trat their data are to be produced, a precondition thateortincaccentwerieratith often cannot be met in many subjects withortsonante.deteriation neurological illness. Further, these techniques

Consutationat ofthepotentially disturb natural speech production,Computatednosioniofich which imposes particular limitations on thetievealednot signiant analysis of more complex utterances like sen-Tiens wih PD andeynor- tences. This may be one reason why parametricthmerewsao tendenyto studies of articulation in Parkinsonism have sotm bothard winheder far focused on elementary speech tasks-for

compafferedes, howitt example, single word production or rapid syl-sedifferences, however,..ro the extent that the lable repetition (diadochokinesis). Although

diadochokinesis seems to predict the overallybe considered anindex2adequacy of speech in dysarthria fairly well,2vitching between two .itching betwing)tou the syllable repetition tasks commonlypatienig exan ing) ouere employed in the measurement of diadochokin-patsienthsrexaminedhe etic rates have the disadvantage of differing

from natural speech in terms of the underlyingmotor organisation.

NS A way to circumvent the limitations ofNS physiological assessment techniques and, at

the same time, obtain objective and quanti-tative data, is provided by acoustic analyses

8 0 of the speech wave. Although a one-to-one0 correspondence between acoustic measures of

a I3 speech sounds and the corresponding speech0 movements does not exist, certain inferences

8 concerning disturbed motor patterns mayo nevertheless be drawn from acoustic data if

speech tasks are appropriately designed. Mosto 0 importantly, acoustically verified deviationsfrom a given target may be considered to reflectrelevant aspects of disturbed speech rather

EN PD than functionally irrelevant idiosyncracies ofGroup articulation.

1096

on August 16, 2020 by guest. P

rotected by copyright.http://jnnp.bm

j.com/

J Neurol N

eurosurg Psychiatry: first published as 10.1136/jnnp.54.12.1093 on 1 D

ecember 1991. D

ownloaded from

Articulatory deficits in Parkinsonian dysarthria: an acoustic analysis

In accordance with earlier studies' ourpatients displayed on average slightly morepronounced phonatory than articulatory dis-turbances, as far as clinical-perceptual ratingsare concerned. Logemann and Fisher sugges-ted a progressive involvement of speech organsin PD, beginning at the laryngeal level andproceeding in the oral direction.7 An alternativeexplanation would be that the laryngealapparatus is more vulnerable to the patho-physiological processes underlying PD, mean-ing that even very mild motor disturbanceshave clearly audible consequences. Finally,impairment of respiratory support may con-tribute to phonatory problems in patients withPD.Our main acoustic finding was a relative

preservation of speech tempo and, at the sametime, a reduction of articulatory precision instop consonant production. A normal or evenincreased rate ofspeech has often been reportedin PD27"' and is considered specific for thisdisorder by some authors.529 At first sight thefinding of a normal or even increased speechtempo seems to be incompatible with theconcept of bradykinesia, which is a well knownfeature of limb akinesia in PD' and has beenfound in the speech motor system as well."7 Yet,the observation of normal or reduced syllabledurations does not necessarily mean that theunderlying articulatory movements were of anormal or even increased velocity. The fact thatarticulatory closure for stop consonant produc-tion was obviously incomplete in some of thepatients suggests that these patients performedarticulatory movements of reduced amplitude.Despite possible bradykinesia they seem tohave produced normal syllabic rates at theexpense of movement excursion. This is inagreement with the findings of Forrest et al ofconsistent reductions in both peak velocity andamplitude of mandibular and labial openinggestures at normal movement times.'6 17

This trade off between tempo andarticulatory precision is made possible by thefact that reduced articulatory precision maystill be compatible with the requirements ofintelligible speech. In this respect speech isdifferent from other motor acts, like grasping,where a similar trade off cannot be made with-out compromising the requirements of themotor task. A task more similar to speechwould be handwriting, with the typical symp-tom of micrographia being observable in PD.Micrographia resembles hypokinetic speech inthat smaller movement excursions compensatefor the inability to execute high velocitystrokes.0 Thus the results of speech analysis-are not incompatible with the concept ofbradykinesia as outlined for limb motorcontrol, emphasising the slowing down ofindividual movements due to inadequate"energising" and "scaling" of initial agonistmuscle activity in relation to functionaldemands.3' 32The tempo-precision trade off is-well known

from normal speech, where an increase inspeech rate is usually achieved by a reduction ofmovement amplitudes rather than by speedingup the movements themselves." In dysarthria

the maintenance of a normal rate at the expenseof articulatory precision is far from universal.On the contrary, most dysarthric syndromesare associated with slowed speech and, at thesame time, undershooting articulation.5 Thefaculty of maintaining a normal syllabic ratetherefore seems characteristic of the Parkin-sonian type of dysarthria.

Yet, following our data and also our earlierobservations, the syndrome of Parkinsoniandysarthria is by no means homogeneous withrespect to speech rate. Seguier et al observedslowed speech in 5% of their patients.29 In oursample, two patients had increased syllabledurations, though the observed increase mustbe considered fairly mild when compared tospastic dysarthria of similar severity.34 Bothsubgroups with either increased or decreasedMSD presented with hypometria in terms ofincompleteness of closure. Possibly the twodifferent types reflect consecutive stages in thedevelopment of Parkinsonian dysarthria ordifferent degrees of impairment, respectively.According to the clinical impressions and per-ceptual judgments of experts the two patientswith increased MSD were most severely affec-ted, with severe phonatory deficits in additionto their articulation problems. At this severitythe maintenance of a normal syllabic rate at theexpense of articulatory precision might haveled to completely unintelligible articulation,causing the patients to slow their speech. Analternative explanation would be that acritically reduced velocity of laryngeal abduc-tions and adductions forced the patients tolower their syllabic rate.

Articulatory imprecision was not uniformlydistributed among the stop consonants of asentence but rather varied systematically withthe sentence accent. Comparison of closures ofthe stressed and the post-stressed syllable of asentence showed normal or almost normalperformance in the stressed position but anoverproportionately reduced realisation in thepost-accent position. In other words, hypo-kinesia affects successive articulatory gesturesofan utterance to different degrees. Obviously,the patients with PD were forced by thelinguistic status of the stressed CV-syllable toraise more "effort" in the closing movement.Thus inappropriate "energising" of agonistmuscle activity and inadequate "scaling" ofagonist burst size to task requirements3' 32 arenot universal features of Parkinsonian motordeficits. With respect to complex speech motorpatterns patients with PD seem to be capable ofappropriate "energising" and "scaling" of oneelementary movement at the expense of others,dependent upon linguistic demands.

Clinical and experimental data indicate thatpatients with PD have difficulties in "running"a sequence of motor programmes linked to-gether within a complex motor plan-that is, in"switching" from one motor programme to thenext one."' For instance, when performing asequence of two simple movements patientsshow a prolongation ofthe interval between thetwo elementary- actions. Thus in complexmovements, there is an extra delay in additionto the increased slowness of the individual

1097

on August 16, 2020 by guest. P

rotected by copyright.http://jnnp.bm

j.com/

J Neurol N

eurosurg Psychiatry: first published as 10.1136/jnnp.54.12.1093 on 1 D

ecember 1991. D

ownloaded from

Ackermann, Ziegler

movements.6'" Similarly, when drawing tri-angles and squares the pauses at the angles areincreased.8 Connor et al showed that isolatedsyllables and syllables produced within repeti-tion trains do not differ significantly in terms ofacoustic variables.'0 As repetitive syllableproduction should have greater demands onmotor planning than isolated syllables thesefindings seem contrary to planning deficits inspeech motor control of patients with PD.Syllable repetitions, however, may differ fromconnected speech insofar as an iteration of themotor programme or highly automatisedprocesses may take over. Our analysis of theopen close transitions in sentence productionshowed no deviations for the utterances of thepatients as transition quotients were withinnormal limits. One might consider this result asbeing in contrast with the observations ofBenecke et al36" and Berardelli et al8 In theseparadigms, however, the single movements areobviously governed by independent motorprogrammes.39 In contrast, there is some lin-guistic evidence that complex articulationpatterns underlying sentence utterances areunder the control of a single "articulatoryplan" structured by suprasegmental-that is,prosodic-features."' Therefore, the concaten-ation of elementary articulations might not beafflicted by the pathophysiological mechanismsoutlined by Benecke et al' " and Berardelliet al.38

All acoustic work was carried out by the Neurophonetics Groupat the Max Planck Institute for Psychiatry, Munich, under theguidance ofW Ziegler. We thank E Hartmann, P Hoole, and IWiesner for their cooperation. Most of the patients were kindlyreferred to us by W H Oertel, Department of Neurology,Klinikum Grofihadern, University (LMU) of Munich. Thiswork was supported by a BMFT grant.

1 Marsden CD. Slowness ofmovement in Parkinson's disease.Movement Dis 1989;4:S26-S37 (suppl 1).

2 Canter GJ. Speech characteristics of patients with Parkin-son's disease. III. Articulation, diadochokinesis, andoverall speech adequacy. J Speech Hear Dis 1965;30:217-24.

3 Darley FL, Aronson AE, Brown JR. Differential diagnosticpatterns of dysarthria. J Speech Hear Res 1969;12:246-69.

4 Darley FL, Aronson AE, Brown JR. Clusters of deviantspeech dimensions in the dysarthrias. J Speech Hear Res1969;12:462-96.

5 Darley FL, Aronson AE, Brown JR. Motor speech disorders.London: Saunders, 1975.

6 Logemann JA, Fisher HB, Boshes B, Blonsky ER.Frequency and cooccurrence ofvocal tract dysfunctions inthe speech ofa large sample ofParkinson patients. JSpeechHear Dis 1978;43:47-57.

7 Logemann JA, Fisher HB. Vocal tract control in Parkinson'sdisease: phonetic feature analysis of misarticulation. JSpeech Hear Dis 1981;46:348-52.

8 Ludlow CL, Bassich CJ. Relationships between perceptualratings and acoustic measures of hypokinetic speech. In:McNeil MR, Rosenbek JC, Aronson AE, eds. The dysar-thrias: physiology, acoustics, perception, management. SanDiego: College Hill, 1984:163-95.

9 Ludlow CL, Bassich CJ. The results of acoustic andperceptual assessment of two types of dysarthria. In:McNeil MR, Rosenbek JC, Aronson AE, eds. The dysar-thrias: physiology, acoustics, perception, management. SanDiego: College Hill, 1984:121-53.

10 Connor NP, Ludlow CL, Schulz GM. Stop consonantproduction in isolated and repeated syllables in Parkin-son's disease. Neuropsychologia 1989;27:829-38.

11 Weismer G. Articulatory characteristics of parkinsoniandysarthria: segmental and phrase-level timing, spirantiza-

tion, and glottal-supraglottal coordination. In: McNeilMR, Rosenbek JC, Aronson AE, eds. The dysarthrias:physiology, acoustics, perception, management. San Diego:College Hill, 1984:101-30.

12 Hirose H, Kiritani S, Ushijima T, Yoshioka H, SawashimaM. Patterns of dysarthric movements in patients withParkinsonism. Folia phoniat 1981;33:204-15.

13 Hirose H, Kiritani S, Sawashimna M. Velocity ofarticulatorymovements in normal and dysarthric subjects. Foliaphoniat 1982;34:210-5.

14 Hirose H. Pathophysiology of motor speech disorders(dysarthria). Folia phoniat 1986;38:61-88.

15 Hunker CJ, Abbs JH, Barlow SM. The relationship betweenparkinsonian rigidity and hypokinesia in the orofacialsystem: a quantitative analysis. Neurology 1982;32:749-54.

16 Connor NP, Abbs JH, Cole KJ, Gracco VL. Parkinsoniandeficits in serial multiarticulate movements for speech.Brain 1989;112:997-1009.

17 Forrest K, Weismer G, Turner GS. Kinematic, acoustic,and perceptual analyses of connected speech produced byParkinsonian and normal geriatric adults. J Acoust Soc Am1989;85:2608-22.

18 Sussman HM, MacNeilage PF. Motor unit correlates ofstress: preliminary observations. J Acoust Soc Am 1978;64:338-40.

19 Ziegler W, von Cramon D. Spastic dysarthria after acquiredbrain injury: an acoustic study. Brit J Dis Comm 1986;21:173-87.

20 Ackermann H, Ziegler W. Die Dysarthrophonie des Parkin-son-Syndroms. Fortschr Neurol Psychiat 1989;57:149-60.

21 Ostry DJ, Keller E, Parush A. Similarities in the control ofthe speech articulators and the limbs: kinematics oftonguedorsum movement in speech. J Exp Psychol HumanPercept Perform 1983;9:622-36.

22 Zyski BJ, Weisiger BE. Identification of dysarthria typesbased on perceptual analysis. J Commun Disord 1987;20:367-78.

23 Kent R, Netsell R. A case study of an ataxic dysarthric:cineradiographic and spectrographic observations. JSpeech Hear Dis 1975;40:115-34.

24 Schonle PW, Grabe K, Wenig P, Hohne J, Schrader J,Conrad B. Electromagnetic articulography: use of alter-nating magnetic fields for tracking movements ofmultiplepoints inside and outside the vocal tract. Brain Lang1987;31:26-35.

25 Forrest K, Weismer G, Adams S. Statistical comparison ofmovement amplitudes from groupings of normal geriatricspeakers. J Speech Hear Res 1990;33:386-9.

26 Caligiuri M. The influence of speaking rate on articulatoryhypokinesia in parkinsonian dysarthria. Brain Lang 1989;36:493-502.

27 Alajouanine T, Sabouraud 0, Gremy F. Etude oscillo-graphique de la parole dans la maladie de Parkinson: effetsprecoces de la coagulation pallidale. Rev Franc Etudes ClinBiol 1964;9:41 1-9.

28 Ludlow CL, Connor NP, Bassich CJ. Speech timing inParkinson's and Huntington's disease. Brain Lang 1987;32:195-214.

29 Seguier N, Spira A, Dordain M, Lazar P, Chevrie-Muller C.Etude de les relations entre les troubles de la parole etles autres manifestations cliniques dans la maladie deParkinson. Folia phoniat 1974;26:108-26.

30 Sanes JN, Evarts EV. Psychomotor performance in Parkin-son's disease. In: Delwaide PJ, Agnoli A, eds. Clinicalneurophysiology in parkinsonism. Amsterdam: Elsevier,1985:117-32.

31 Hallett M, Khoshbin S. A physiological mechanism ofbradykinesia. Brain 1980;103:301-14.

32 Berardelli A, Dick JPR, Rothwell JC, Day BL, MarsdenCD. Scaling of the size of the first agonist EMG burstduring rapid wrist movements in patients with Parkin-son's disease. J Neurol Neurosurg Psychiat 1986;49:1273-9.

33 Kent RD. The segmental organization of speech. In:McNeilage PF, ed. The production of speech. New York:Springer, 1983.

34 Ziegler W, Hartmann E, Hoole P, von Cramon D. AkustischeDimensionen dysarthrischer St6rungsmerkmale. Munchen:GSF, 1990.

35 Marsden CD. The mysterious motor function of the basalganglia. Neurology 1982;32:514-39.

36 Benecke R, Rothwell JC, Dick JPR, Day BL, Marsden CD.Disturbance of sequential movements in patients withParkinson's disease. Brain 1987;110:361-79.

37 Benecke R, Rothwell JC, Dick JPR, Day BL, Marsden CD.Simple and complex movements off and on treatment inpatients with Parkinson's disease. J Neurol NeurosurgPsychiat 1987;50:296-303.

38 Berardelli A, Accornero N, Argenta M, Meco G, ManfrediM. Fast complex arm movements in Parkinson's disease.JNeurol Neurosurg Psychiat 1986;49:1146-9.

39 Benecke R, Rothwell JC, Day BL, Dick JPR, Marsden CD.Motor strategies involved in the performanceofsequentialmovements. Exp Brain Res 1986;63:585-95.

40 Levelt WJM. Speaking. From intention to articulation.London: MIT Press, 1989.

1098

on August 16, 2020 by guest. P

rotected by copyright.http://jnnp.bm

j.com/

J Neurol N

eurosurg Psychiatry: first published as 10.1136/jnnp.54.12.1093 on 1 D

ecember 1991. D

ownloaded from