research article changes in vowel articulation with...

Research ArticleChanges in Vowel Articulation withSubthalamic Nucleus Deep Brain Stimulation inDysarthric Speakers with Parkinsonrsquos Disease

Vincent Martel Sauvageau1 Joeumll Macoir1 Meacutelanie Langlois2

Michel PrudrsquoHomme2 Leacuteo Cantin2 and Johanna-Pascale Roy3

1 Centre de Recherche Institut Universitaire en Sante Mentale de Quebec Laboratoire Langage et CognitionQuebec QC Canada G1J 2G3

2 CHU de Quebec-Enfant-Jesus Departement des Sciences Neurologiques Quebec QC Canada G1J 1Z43Universite Laval Centre Interdisciplinaire de Recherche sur Les Activites Langagieres Quebec QC Canada G1V 0A6

Correspondence should be addressed to Vincent Martel Sauvageau vincentmartel-sauvageau1ulavalca

Received 11 July 2014 Accepted 7 October 2014 Published 21 October 2014

Academic Editor Nobutaka Hattori

Copyright copy 2014 Vincent Martel Sauvageau et al This is an open access article distributed under the Creative CommonsAttribution License which permits unrestricted use distribution and reproduction in any medium provided the original work isproperly cited

Purpose To investigate changes in vowel articulation with the electrical deep brain stimulation (DBS) of the subthalamic nucleus(STN) in dysarthric speakers with Parkinsonrsquos disease (PD) Methods Eight Quebec-French speakers diagnosed with idiopathicPD who had undergone STN DBS were evaluated ON-stimulation and OFF-stimulation (1 hour after DBS was turned off) Vowelarticulation was compared ON-simulation versus OFF-stimulation using acoustic vowel space and formant centralization ratiocalculated with the first (1198651) and second formant (1198652) of the vowels i u and aThe impact of the preceding consonant contexton articulation which represents a measure of coarticulation was also analyzed as a function of the stimulation state ResultsMaximum vowel articulation increased during ON-stimulation Analyses also indicate that vowel articulation was modulated bythe consonant context but this relationship did not change with STNDBS Conclusions Results suggest that STNDBSmay improvearticulation in dysarthric speakers with PD in terms of range of movement Optimization of the electrical parameters for eachpatient is important andmay lead to improvement in speech finemotor control However the impact on overall speech intelligibilitymay still be small Clinical considerations are discussed and new research avenues are suggested

1 Introduction

Parkinsonrsquos disease (PD) has traditionally been consideredstrictly as a motor disorder with symptoms such as tremormuscle rigidity and bradykinesia In recent decades manyresearchers have investigated other clinical manifestations ofPD such as mood changes language impairment cognitionand sleep disorders Today PD is commonly viewed as amul-tisystemic degenerative disorder [1] Alongside these symp-toms up to 90 of people with PD develop speech disordersover the course of the disease [2] In many studies authorsinvestigated the impairment of speech systems in PD from amotor acoustic or perceptual point of view Studies exam-ining physiological changes in the speech systems of people

with PD reported impaired respiratory [3] laryngeal [4]and orofacial [5 6] functions which have an impact onthe acoustic signal of speech Studies investigating suchchanges reported reduced intensity level [7] and fundamentalfrequency (119891

0) range [8] altered phonation quality [9] and

inaccurate and reduced articulation [10ndash12] All these changesaffect listenersrsquo perceptions such as perceived softer speechreduced voice quality and poor articulation As a resultimpaired intelligibility is very common in PD [13]

Various pharmacological and surgical techniques are nowavailable to help manage the different motor symptoms ofpatients with PD One of these surgical techniques is deepbrain stimulation of the subthalamic nucleus (STN DBS)Even though it has been demonstrated that STN DBS can

Hindawi Publishing CorporationParkinsonrsquos DiseaseVolume 2014 Article ID 487035 9 pageshttpdxdoiorg1011552014487035

2 Parkinsonrsquos Disease

drastically reduce motor symptoms and improve patientsrsquoquality of life it has been associated with relatively smallchanges in dysarthria severity levels and intelligibility Inmost studies the impact of STNDBSondysarthriawas at bestmixed and was more often than not minor andor poor [14]For a review of the impact of STNDBS on speech systems see[15]

The articulation of speech sounds requires fine motorcontrol Speech units can be characterized in terms of artic-ulatory gestures (range of movement) or in terms of acousticdistinctiveness Acoustic distinctiveness refers to the fact thattwo speech units that are acoustically differentiated are moreeasily identified by our perceptual system [16] Using thisparadigm a category of soundsmdashthe vowelsmdashcan be acous-tically described and differentiated from one another by theiracoustic characteristicsThe principal acoustic characteristicsused to describe vowels are their formants The first twoformants of a vowel 1198651 and 1198652 are spectral values that allowcategorization of the phoneme 1198651 and 1198652 respectively serveas indicators of the open-close and front-back position of thearticulators (jaw tongue) during the production of speech[17] The articulation of vowels is very important for speechintelligibility and reduced acoustic distinctiveness of vowelshas been reported in studies of dysarthric speakers includingpeople with PD [18] In a study analyzing different metrics toevaluate the effects of STN DBS on speech production Weis-mer et al [19] suggested that vowel production has promisingresults regarding articulation measurements A few studies[20 21] investigated the changes in vowel production thatcould occur with STN DBS in PD However these studieslooked at speech components other than articulation such asvoice quality or speech rate

Coarticulation is another important acoustic factor forspeech intelligibility [22 23] Coarticulation can be definedas the influence speech units have on each other in connectedspeech [24] Coarticulation effects fall into two categories andare described as anticipatory (ldquoright to leftrdquo) or carry-over(ldquoleft to rightrdquo) Anticipatory coarticulation is generally con-sidered a consequence of motor planning because the posi-tions of the articulators for a given speech gesture are modi-fied in prediction of a following gesture Carry-over coartic-ulation on the other hand is considered a consequence ofarticulatory inertia and is influenced by the physical char-acteristics of the articulators [25] Initially investigated innormal speakers more and more studies looked at possiblechanges in coarticulation patterns in dysarthric speakers [2627] Coarticulation is amarker of speech finemotor control Itcan be acoustically measured with different methods includ-ing analyzing the influence of a given context on the articu-lation of the following phoneme (carry-over coarticulation)In some studies authors used this type of measurement tocharacterize coarticulation in normal speakers [28ndash30]Theyshowed that in a consonant-vowel (C-V) sequence the placeof articulation of the consonantmodulates the formant valuesof the vowel Some studies showed that STN DBS can inducechanges in fine motor control during speech production [15]One could therefore presume that coarticulation could beinfluenced by STN DBS However this hypothesis has notbeen specifically investigated in any previous studies

Even though articulatory function is very important forspeech intelligibility particularly for vowels only a few stud-ies examined the impact of STN DBS on articulation in PDin terms of range of movement Moreover to the best of ourknowledge no studies have investigated the coarticulationchanges that could occur with STN DBS Therefore the pri-mary objective of this study was to examine the effects of STNDBS on speech articulation in dysarthric speakers with PDusing acoustic measurements of vowelsThe second objectivewas to examine the coarticulation changes with STNDBS byanalyzing the impact of the preceding consonant context onvowel articulation which represents a measure of carry-overcoarticulation

2 Methods

21 Participants

211 Demographic Characteristics The study was approvedby the local institutional ethics committee for the safety ofhuman subjects and written informed consent was obtainedfrom all participants Eight participants (5women and 3men)aged 53ndash72 years with idiopathic PD diagnosed 9 to 25 yearsprior to the study were recruited in an outpatient clinic Allparticipants were taking antiparkinsonian medications withL-dopa equivalent dose from 300 to 1500mg per day andno changes were made to their medication during the studyAll of them were native speakers of Quebec French who hadalways lived in the province of Quebec Although no formalhearing evaluationwas conducted all participants were func-tional in conversation and none reported any hearing impair-ment General cognitive functions were measured using theMontreal Cognitive Assessment (MoCA) [31] and noparticipant fell below the cut-off score according to age andeducation level [32] Each participant with PD was also askedto complete the French version of the Voice Handicap Index(VHI) [33] which is one of the most common instrumentsdesigned to measure quality of life with respect to dysarthriain PD Higher values indicate greater voice handicapCharacteristics of each participant are reported in Table 1

212 Speech Characteristics Two speech-language pathol-ogists were asked to reach an agreement on dysarthriadiagnosis dysarthria severity and acoustic characteristics foreach participant during ON-stimulation Their evaluationwas based on audio recordings of each participant reading ldquoLabise et le soleilrdquo [34] a standardized French text commonlyused in French phonetic experiments to study normal orpathological speech [35] This text is considered the Frenchequivalent of the English textTheRainbow Passage [36] Eachrecording was presented to the two speech-language pathol-ogists independently in a quiet room via open-air speakersDysarthria severity ranged from mild to severe with alteredphonation reduced articulation and abnormal speech ratebeing the main acoustic features identified in most of therecordings All participants were diagnosed with hypokineticdysarthria which is consistent with the speech profile gener-ally observed in people with PD [13] Although the presence

Parkinsonrsquos Disease 3

Table 1 Characteristics of participants

Participant Age Sex Education MoCAa Years post-Dx L-Dopa dose VHI scoreb

PD1 66 M 9 25 25 1500mg 83PD2 66 F 11 26 16 600mg 34PD3 71 M 6 25 11 500mg 49PD4 72 F 19 26 19 300mg 56PD5 53 F 11 29 12 300mg 34PD6 65 F 12 27 14 500mg 51PD7 72 F 19 25 19 500mg 48PD8 69 M 13 24 9 300mg 42Notes aMoCA = Montreal Cognitive AssessmentbVHI = Voice Handicap IndexYears post-DX = Years after the diagnosis has been made

Table 2 Individual DBS parameters

Participant Yearspost-op

Frequency (Hz) Voltage (V) Pulse width (120583s) Electrode configuration(ldquo0rdquo = off ldquo+rdquo = anode ldquominusrdquo = cathode)

Left Right Left Right Left Right Left contacts Right contacts0 1 2 3 IPGc 0 1 2 3 IPGc

PD1 4 60 60 55 60 60 60 0 0 0 minus + 0 0 0 minus +PD2 4 145 145 35 38 90 90 + 0 minus minus 0 + 0 minus minus 0PD3 4 185 185 33 33 60 60 0 0 minus minus + 0 0 minus minus +PD4 5 185 185 35 35 90 90 0 0 minus minus + 0 0 minus minus +PD5 4 185 185 37 37 90 90 + 0 minus minus 0 + 0 minus minus 0PD6 3 145 145 35 35 60 60 0 + minus 0 0 0 0 minus + 0PD7 3 180 180 30 37 60 60 0 0 minus minus + 0 0 minus 0 +PD8 2 180 180 38 38 60 60 0 0 minus minus + 0 0 minus minus +Notes IPGc = Internalized Pulse Generator Case

of a speech disorder was not a criterion to be eligible for thisstudy all participants reported speech difficulties

213 Deep Brain Stimulation Characteristics The partici-pants in this study had undergone bilateral DBS of the STNsurgery 2 to 5 years prior to the study All of them hadbeen operated on by the same neurosurgeons (LC and MP)and were regularly followed by the same neurologist (ML)Surgery was done under local anaesthesia and sedation withthe CRW stereotactic frame The day before surgery patientshad high-resolutionT2-weightedMRIs (30-T unit Siemens)These images were fused with a T1-Gadolinium (double dose15-T unit Siemens) acquired with the localization frame onthe day of surgery Neuronavigation (Stealth System fromMedtronic Inc Minneapolis MN) was used to plan thesurgery and fuse the images The target was the STN andwas calculated from themid-commissural pointThe indirectcoordinates were 3mm behind the mid-commissural point12mm lateral and 4mm inferior The target was confirmedby microrecording and microstimulation then a quadripolarlead was implanted (Model 3387 Medtronic Inc Minneapo-lis MN) Surgery was done on both sides on the same day

One to 3 days later the neurostimulator was implanted(Activa System Medtronic Inc Minneapolis MN)

At the time of the study the electrical parameters ofthe DBS had been stable for a period of at least 2 monthsFor 7 of the 8 participants stimulations were delivered ata frequency of 145ndash185Hz and a voltage of 30ndash38 V Forone participant (PD1) the stimulations were delivered at alower frequency (60Hz) but higher voltage (55 V left 60 Vright)The configuration of the electrical parameters for eachparticipant was selected and adjusted by their neurologistbased on observed and reported clinical symptoms of MPsuch as tremor rigidity speech difficulty and dyskinesiaIndividual DBS parameters for each participant are reportedin Table 2 No postoperative measurements of the electrodesrsquolocation were available at the time of recruitment

22 Evaluation Sessions Two evaluation sessions for eachparticipant took place on different days Participants PD4PD5 and PD8 were first recorded in the ON-stimulationstate and then at least 6 weeks later in the OFF-stimulationstate For the other participants OFF-stimulation recordingstook place first and ON-stimulation recordings were made


afterward This design was chosen to counterbalance thepossible habituation effects of the task between the two stim-ulation conditions ON-stimulation recordings were made athome with no change to the stimulation settings used in theparticipantrsquos everyday life OFF-stimulation recordings weremade at the hospital under medical surveillance and tookplace one hour after the stimulator was turned OFF All thesesessions took place in a quiet room at the same time of theday for each participant tominimize variations inmedicationcycles All participants took their antiparkinsonian medica-tion at least one hour prior to the sessions and were in ON-state medication during the recordingsThe evaluations wereadministered by the first author of this paper (VMS) with thecollaboration of a graduate student specializing in speech andlanguage disorders in PD

221 Neurological Assessment At both evaluation sessionsthe severity of motor symptoms was measured using themotor section of Unified Parkinsonrsquos Disease Rating Scale(UPDRS-III) [37] Disease stage was estimated using theadapted Hoehn and Yahr scale [38] This neurological assess-ment was done for each participant to document the impactof their antiparkinsonianmedication as well as the long-termeffects of STN DBS on their motor symptoms

222 Speech Assessment All recordings were made using aShure 510A head-mounted microphone and a Zoom H4naudio recorder at a sampling rate of 441 kHz Mouth-to-microphone distance was approximately 4 to 8 cm for eachparticipant but remained constant throughout the recordingsession Vowel articulation was measured with a reading-aloud task of spoken ldquoconsonant-vowel-consonant-vowelrdquo(CVCV) tokens with the target vowels i u and a andthe consonants p t k b d and g (plosives) andf s and ⎰ (ldquochrdquo) (fricatives) The use of these threevowels provided a way to measure maximum vowel spaceacoustically occupied by each participant These consonantswere selected because they represent a variety of acceptedphonemic contexts in French The plosive contexts alsoenabled us to investigate vowel articulation as a function ofthe preceding consonantal context (p and b for labial con-text t and d for alveolar context and k and g for velarcontext) which represents a measure of coarticulationThesetokens were embedded in the carrier phrase ldquoJe pense CVCVcette foisrdquo (ldquoI think CVCV this timerdquo) in order to standardizethe prosody and accentuation of productions The task wasrepeated twice in each recording session (ON- and OFF-stimulation) and each individual token occurred twice ineach repetition A total of 108 productions (9 consonants times3 vowels times 2 token occurrences times 2 task repetitions) werethus recorded per participant per stimulation condition Theorder of presentation of the tokens was randomized but thissequence remained the same between all participants andthroughout all recording sessions

23 Acoustic Analyses Acoustic analyses were done by atrained phonetician using Praat software v5330 [39] runningon Windows OS Acoustic segmentations were conducted

using different visual criteria on a spectrogram and oscillo-gramMultiple scripting procedures were implemented in theanalyses when no manual intervention was required

231 General Vowel Articulation Vowel articulation wasmeasured by analyzing 1198651 and 1198652 formant frequencies of iu and a of the last vowel of the CVCV tokens in a 500-msanalysis window Vowel duration was also measured forcovariance analyses Vowel onset was first determined by theappearance of stable formant frequencies on the spectrogramand the offset was determined by the last glottal pulse visibleon the oscillogram With these values two variables werecalculatedThe first variable is the acoustic vowel space (AVS)which is the surface of the triangle formed by the 1198651 and1198652 formant values of the vowels i u and a HigherAVS values correspond to increased vowel articulation Thisvariable is calculated using the following formula

AVS =10038161003816100381610038161003816100381610038161003816

1198652i (1198651uminus1198651a)+1198652u (1198651aminus1198651i)+1198652a (1198651iminus1198651u)2

10038161003816100381610038161003816100381610038161003816

(1)

The second variable is the formant centralization ratio (FCR)which is a coefficient that represents the magnitude ofcentralization of the formants 1198651 and 1198652 for vowels i uand a This metric was developed by Sapir and colleagues[10] and has been used in other studies on vowel articulationin PD Higher FCR values represent higher formantcentralization and consequently reduced vowel articulationThis variable is calculated using the following formula

FCR = 1198652u + 1198652a + 1198651i + 1198651u1198652i + 1198651a

(2)

232 Vowel Articulation by Consonant Context Vowel artic-ulation was also measured as a function of the precedingconsonant context by measuring 1198651 and 1198652 only for vowelsin labial (p and b) alveolar (t and d) and velar (kand g) contexts and then calculating AVS for each contextrepresenting a place of articulation The fricative contextswere excluded from these analyses because their place ofarticulation is not comparable to those of the plosive contextsSegmentation criteria for this variable were the same as thoseused for general vowel articulation

24 Reliability of Acoustic Data Due to the nature of thespeech samples acquired in this study the formant detectionalgorithms used by the Praat software can produce outlierartefacts [40] To minimize the number of these erroneousformant values in our data pool a statistical multivariate out-lier detection procedure was applied to the measured 1198651 and1198652 frequencies for individual vowels and for each participantUsing leverage values standardized residual scores andinfluence factors [41] data differentiated at 119875 lt 001 wereexcluded from further analysis With this procedure 23 datapoints (13) of the entire data pool were rejected Also dueto the recording methods and participantsrsquo fluctuating voicequality (eg oversaturation of the microphone voice breaksand no discernible glottal pulse) 41 data points (24) could


not be analyzed Finally one participant (PD2) was not ableto complete the second repetition of the task during OFF-stimulation due to fatigue Of the 54 tokens in the secondrepetition 43 are missing

3 Results

31 Data Pooling Because the speech task was administeredtwice per experimental condition preliminary analyses weremade to verify differences between the two repetitions underboth conditions On all variables no statistical difference wasfound between both repetitions during ON- and OFF-stim-ulation Therefore data obtained in both repetitions duringON- and OFF-stimulation were pooled for subsequent anal-yses without considering the occurrence of the repetition

32 Statistical Analyses The effect of the electrical stimula-tion condition (ON-stimulation versus OFF-stimulation) oneach variable under study was analyzed using a mixed modelanalysis of variance (ANOVA) Repeated measure factor(occurrence of the repetition) was entered in the modeland was based on an unstructured covariance matrix whichallows for unequal variance between each repetition [41]Participantswere entered in themodel as a random factor andwere based on a scaled identity covariance matrix Finallyall dependent variables were analyzed using the stimulationcondition (ON versus OFF) as a fixed factor Subsequentanalyses were conducted by entering the consonantal context(labial versus alveolar versus velar) as another fixed factorAll statistical analyses were conducted using SPSS v 20 [42]This procedure follows the guidelines suggested in [43] forphonetic research in order to avoid statistical problems likepseudo-replication of data

33 Neurological Assessment Table 3 reportsmeans and stan-dard deviations of the total UPDRS-III score and Hoehnand Yahr stage in both OFF- and ON-stimulation condi-tions Statistical analyses indicate a significantly lower totalUPDRS-III score during ON-stimulation which indicatesthat the electrical stimulations of DBS reduce the severity ofmotor symptoms For the Hoehn and Yahr stage a statisticaltendency (119875 = 007) was obtained (average ON score 226average OFF score 263) The score on the single speechitem (item 18) of the UPDRS-III was also analyzed separatelyScoring for this item reflects the degree to which speechis impaired that is considered not impaired (0) mildlyimpaired (1) moderately impaired (2) severely impaired(3) or mostly unintelligible (4) Analyses indicate that theaverage score on this item was slightly lower during ON-stimulation (averageON score 206 averageOFF score 219)but the difference failed to reach statistical significance Inboth conditions none of the participants showed signs ofdyskinesia or reported other side effects

34 General Vowel Articulation Table 4 reports means andstandard deviations for 1198651 and 1198652 formant frequencies (Hz)for the vowels i u and a during OFF-stimulation andON-stimulation Descriptive data for11986511198652AVS (Hz2) FCR

Table 3 Scores on motor subscale of the UPDRS (no symptoms =0) Hoehn and Yahr (HampY mild = 1) staging during ON- and OFF-stimulation means (standard deviations) 119905 and P values

UPDRS-III(total) HampYa UPDRS-IIIb

(speech item)

OFF-stimulation 483 263 219(167) (074) (088)

ON-stimulation 297 226 206(918) (104) (062)

119905 439 205 068119875 lt001lowast 007 gt005lowastSignificant effectaHampY = Hoehn and YahrbUPDRS-III = Unified Parkinsonrsquos Disease Rating Scale

i u

a

200

300

400

500

600

700

800

5001000150020002500

ON-stimulationOFF-stimulation

F2 (Hz)

F1

(Hz)

Figure 1 Acoustic space of average 1198651 and 1198652 values of the vowelsi u and a during OFF-stimulation and ON-stimulation

and vowel duration (msec) are also reported for the OFFand ON conditions Statistical analyses indicate a significantincrease in AVS (+75723Hz2) and decrease in FCR (minus0098)in the ON-stimulation condition No significant changewas found in vowel duration These results indicate thatmaximum vowel articulation increased with the electricalstimulations and that this change was not related to theduration of the vowels Figure 1 displays the average 1198651 and1198652 values for i u and a during OFF-stimulation andON-stimulation in a standard 11986521198651 acoustic space Visualanalysis of the maximum acoustic space indicates that themain changes in vowel articulation that occur with electricalstimulations are on 1198652 for i and u and on 1198651 for a

35 Vowel Articulation by Consonant Context Figure 2 dis-plays acoustic vowel space values (Hz2) calculated from 1198651and 1198652 formant frequencies of the vowels i u and a pro-duced in labial alveolar and velar consonant contexts duringOFF-stimulation and ON-stimulation Statistical analyses foreach context indicate a significant effect of the stimulationcondition for each consonant context with an increase invowel articulation in the ON-stimulation condition labial


Table 4 Formant frequencies acoustic vowel space formant centralization ratio and vowel duration at OFF- and ON-stimulation means(standard deviations) 119865 and 119875 values

i u a Acoustic vowelspace (Hz2)

Formantcentralization ratio

Vowel duration(ms)1198651 (Hz) 1198652 (Hz) 1198651 (Hz) 1198652 (Hz) 1198651 (Hz) 1198652 (Hz)

OFF-stimulation 3115 22331 3292 11866 5833 17363 152036 1240 1302(544) (2190) (666) (2831) (1276) (2635) (14251) (0035) (88)

ON-stimulation 3076 23457 3285 10770 6801 17530 227759 1142 1384(451) (2644) (613) (2501) (911) (2680) (17905) (0031) (105)

Change +75723 minus0098 +82(11504) (020) (51)

119865 43327 24482 2665119875 lt0000lowast lt001lowast gt005

lowast

Significant effect

0

5

10

15

20

25

30

35

ON-stimulation OFF-stimulation

LabialVelarAlveolar

Acou

stic v

owel

spac

e (H

z2)times10000

Figure 2 Acoustic vowel space values for vowels in labial velar andalveolar context in both stimulation conditions

context 119865(1 68) = 1971 119875 lt 001 velar context 119865(1 69) =1346 119875 lt 001 alveolar context 119865(1 70) = 3231 119875 lt001 Furthermore a mixed model ANOVA analysis wasperformed on data from all contexts with stimulation (OFFversus ON) and consonant context (labial versus alveolarversus velar) entered as fixed effects A significant effectof the stimulation was found 119865(1 9433) = 2255 119875 lt0000 with an increase in vowel articulation in the ON-stimulation condition A significant effect of the consonantcontext was also found 119865(2 380) = 3197 119875 lt 0000 wherevowel articulation increased following the contexts alveolarltvelar lt labial On the other hand no statistical interactioneffect between the stimulation condition and consonantcontext was found 119865(2 380) = 0621 119875 gt 005These resultsindicate that vowel articulation is influenced by the precedingconsonant context but that this effect is not modulated byDBS whether the stimulation is OFF or ON

4 Discussion

This study reports results regarding the impact of the electri-cal stimulations of bilateral STN DBS on vowel articulationin 8 individuals with PD With respect to motor symptomswhen STNDBS was turned OFF the severity of motor symp-tomsmeasuredwith theUPDRS-III was significantly greaterand the Hoehn and Yahr stage marginally increased Otherstudies investigating the impact of STNDBS onmotor symp-toms in PD concluded that bilateral implantation inducesgreater improvement in motor symptoms in people present-ing more severe motor symptoms [44] Our results concern-ing motor symptoms are consistent with these previous stud-ies It is also important to note that none of the participantsreported side effects directly related to STNDBS in our studySpecifically many side effects (such as stimulation-induceddyskinesia stimulation-induced hypotonia or apraxia ofeyelid opening) are frequently associated with nonoptimalstimulation settings [45] The absence of such side effectsas well as the reduction in motor symptoms measured bythe UPDRS-III suggests that the electrical settings wereoptimally configured for each participant in our study

With respect to speech the impact of STN DBS on vowelarticulation was analyzed using the first two formants of thetarget vowels i u and a and then by calculating twoacoustic factors acoustic vowel space and formant central-ization ratio These two measures serve as acoustic markersof vowel articulation and represent the range of articulatorymovement during the production of vowels for each partic-ipant Our results indicate that STN DBS increased vowelarticulation while vowel duration remained unchangedShorter vowels are usually related to vocalic undershoot inmany studies [46] Our result is therefore important since itcan be assumed that the change in articulation we observedwith STN DBS is due to a change in the articulatory range ofmovement per se rather than a change in speech rate or vowelduration

Further analyses of formant values for each vowel enablea more precise description of the acoustic changes that occurwith STNDBS and by extension its impact on speech motorcontrol For i and u the main changes with STN DBS


occurred on 1198652 (increase for i and decrease for u) while1198651 remained stable 1198651 is usually associated with the widthof the resonance cavity (aperture) while 1198652 represents thelength of the cavity (front-back distinction) Therefore ourresults indicate that the electrical stimulations of STN DBSimprove the range of anteroposteriormovement of the tonguedorsum For the vowel a STN DBS led to an increase on1198651 while 1198652 remained stable 1198651 is the acoustic factor thatrepresents vowel aperture controlled by the jaw Thereforethis result indicates that STN DBS may improve the rangeof movement in the jaw opening Among the three vowelswe analyzed a showed the broader change with STN DBSThis result indicates that different articulatory gestures inspeech production may be differently affected by STN DBSStudies investigating speech motor control with STN DBSin PD using articulatory measurements are scarce Furtherstudies should examine the impact of STN DBS in PD onthe distinct jawtonguelips articulatory processes in termsof range target and velocity movements

In the present study we observed significant changes invowel articulation as a function of the preceding consonantcontext which represents an acoustic marker of coarticu-lation More specifically we showed that articulation wassignificantly reduced when vowels were produced in an alve-olar context as compared to velar and labial contexts Eventhough the consonant context influences vowel articulationthis coarticulatory phenomenon did not vary as a function ofthe STN DBS in our study To our knowledge our study isthe first to look at the impact of STN DBS on coarticulationThe acoustic metric used in this study investigated the impactof the preceding consonantal context on vowel articulationin its stable portion which is a measure of carry-overcoarticulation Even though our results indicate that this typeof coarticulation is not sensitive to STN DBS the impact ofthe stimulation on speech coarticulation should be exploredin future studies because (1) it has been shown to be altered insome studies on dysarthria in PD [26] and (2) it is essential tospeech intelligibility [47] In this regard future studies couldalso investigate the impact of STN DBS on anticipatory coar-ticulation because it is a marker of motor programming [25]and implies mechanisms complementary to those implied incarry-over coarticulation

Studies investigating articulatory changes with STN DBSin PD are almost nonexistent In a preliminary reportDromey and Bjarnason [48] studied the impact of electri-cal stimulations on speech and language characteristics inpeople with PD including vowel articulation Of their sixpatients two showed increased vowel articulation in the ON-stimulation condition three showed decreased articulationand one did not varyHowever the authors did not specify theindividual electrical parameters of the stimulations Stimula-tion frequency and voltage were associated with other alteredspeech systems in some studies [49] so that interpretation oftheir mixed results is difficult The lateralization of the stim-ulations may also have an impact on articulation Wang andcolleagues [50] analyzed the impact of left- versus right-sideSTN DBS on different speech mechanisms including artic-ulation accuracy They concluded that left-side stimulationaltered articulation accuracy while it remained unchanged

or even improved with right-side stimulation In our studyall the patients had undergone bilateral STN DBS whichmay have better effects on speech than unilateral stimulationparticularly compared to left-side stimulation

The intrinsic characteristic of the pathology (idiopathicPD versus young-onset PD severity of the motor symptoms)is another factor that may account for the differing results Inour study all patientswere diagnosedwith idiopathic PDwithage at onset greater than 40 years which may explain whyour results are homogeneous Surprisingly to our knowledgespeech characteristics and dysarthria profiles in PD in regardto onset of the disease have not been investigated Futurestudies should examine this aspect more closely

41 Limitations of the Study Some limitations must be takeninto consideration when interpreting the findings of thisstudy First this investigation was conducted with only asmall number of participants Generalization of these resultsmust therefore be viewedwith caution and the large inter- andintravariability in speech disorders and clinical presentationin people with PD must be kept in mind Another limitationof this study is that our speechmeasurements were taken onlyin reading tasks which may cast doubt on the naturalness ofthe speech This type of limitation is commonly recognizedin phonetic studies [10 51] but it is a methodological choicemade to control the phonemic syntactic and prosodiccontexts around the target sounds

With the exception of item 18 of the UPDRS-III (whichdid not vary significantly with STN DBS) speech intelligibil-ity was not formally assessed here Even though vowel artic-ulation is a strong factor associated with intelligibility theimpact of the changes in vowel articulation on overall speechintelligibility must be viewed with caution because additionalacoustic factors are responsible for reduced intelligibility inPD such as phonation or prosodic disturbances

5 Conclusions

This study is one of the first to investigate changes in vowelarticulation and coarticulation that occur with STN DBS inPD in terms of maximum range Using acoustic measure-ments we found that bilateral STN DBS improves articula-tion in terms of anteroposterior tongue movements and jawopening during speech production Coarticulation did notchange as a function of the stimulations Previous studiesthat investigated the impact of STN DBS on different speechsystems in PD had mixed results Differences in stimulationparameters and configuration criteria may explain thesedifferences According to our results dysarthria is a symptomof PD that is sensitive to STN DBS Therefore speech qualityshould also be considered when optimizing stimulationparameters for patientsWhen optimized stimulation shouldnot negatively affect speech and could in fact improve someaspects such as articulation

Conflict of Interests

The authors declare that there is no conflict of interestsregarding the publication of this paper


Acknowledgments

This study was supported by a doctoral grant from the Fondsde Recherche Societe et Culture du Quebec The authors thankChristina Tremblay and Lise Mailloux for their help withthe evaluation sessions and Noemie Auclair-Ouellet for thedysarthria diagnoses

References

[1] N Archibald N Miller and L Rochester ldquoNeurorehabilitationin Parkinson diseaserdquo in Handbook of Clinical Neurology M PBarnes and D C Good Eds vol 110 pp 435ndash442 ElsevierAmsterdam The Netherlands 3rd edition 2013

[2] R Pahwa K E Lyons andWC KullerHandbook of ParkinsonrsquosDisease Informa Healthcare New York NY USA 4th edition2007

[3] P B Mueller ldquoParkinsonrsquos disease motor-speech behavior in aselected group of patientsrdquo Folia Phoniatrica vol 23 no 5 pp333ndash346 1971

[4] S Gallena P J Smith T Zeffiro and C L Ludlow ldquoEffects oflevodopa on laryngeal muscle activity for voice onset and offsetin Parkinson diseaserdquo Journal of Speech Language and HearingResearch vol 44 no 6 pp 1284ndash1299 2001

[5] H Ackermann J Konczak and I Hertrich ldquoThe temporalcontrol of repetitive articulatory movements in Parkinsonrsquosdiseaserdquo Brain and Language vol 56 no 2 pp 312ndash319 1997

[6] Y Yunusova G Weismer J R Westbury and M J Lind-strom ldquoArticulator movements during vowels in speakers withdysarthria and healthy controlsrdquo Journal of Speech Languageand Hearing Research vol 51 no 3 pp 596ndash611 2008

[7] C M Fox and L O Ramig ldquoVocal sound pressure level andself-perception of speech and voice in men and women withidiopathic Parkinson diseaserdquoThe American Journal of Speech-Language Pathology vol 6 no 2 pp 85ndash94 1997

[8] R D Kent H K Vorperian J F Kent and J R DuffyldquoVoice dysfunction in dysarthria application of the Multi-Dimensional Voice Programrdquo Journal of Communication Dis-orders vol 36 no 4 pp 281ndash306 2003

[9] L A Ramig I R Titze R C Scherer and S P Ringel ldquoAcousticanalysis of voices of patients with neurologic disease rationaleand preliminary datardquo Annals of Otology Rhinology and Laryn-gology vol 97 no 2 pp 164ndash172 1988

[10] S Sapir J L Spielman B H Story and C Fox ldquoEffects ofintensive voice treatment (the Lee Silverman Voice Treatment[LSVT]) on vowel articulation in dysarthric individuals withidiopathic Parkinson disease acoustic and perceptual findingsrdquoJournal of Speech Language and Hearing Research vol 50 no 4pp 899ndash912 2007

[11] E Baudelle J Vaissiere J L Renard B Roubeau and CChevrie-Muller ldquoCaracteristiques vocaliques intrinseques etco-intrinseques dans les dysarthries cerebelleuse et parkinsoni-ennerdquo Folia Phoniatrica et Logopaedica vol 55 pp 137ndash1462003

[12] Y-I Bang K Min Y H Sohn and S-R Cho ldquoAcoustic char-acteristics of vowel sounds in patients with Parkinson diseaserdquoNeuroRehabilitation vol 32 no 3 pp 649ndash654 2013

[13] J R DuffyMotor SpeechDisorders Substrates Differential Diag-nosis andManagement Elsevier Health Sciences St Louis MoUSA 3rd edition 2012

[14] G M Schulz and M K Grant ldquoEffects of speech therapy andpharmacologic and surgical treatments on voice and speechin Parkinsonrsquos disease a review of the literaturerdquo Journal ofCommunication Disorders vol 33 no 1 pp 59ndash88 2000

[15] S Skodda ldquoEffect of deep brain stimulation on speech per-formance in Parkinsonrsquos diseaserdquo Parkinsonrsquos Disease vol 2012Article ID 850596 10 pages 2012

[16] R D Kent and C ReadThe Acoustic Analysis of Speech Singu-lar San Diego Calif USA 1992

[17] G Fant Acoustic Theory of Speech Production Mouton TheHague the Netherlands 1960

[18] K Tjaden andG EWilding ldquoRate and loudnessmanipulationsin dysarthria acoustic and perceptual findingsrdquo Journal ofSpeech Language and Hearing Research vol 47 no 4 pp 766ndash783 2004

[19] GWeismer Y Yunusova and K Bunton ldquoMeasures to evaluatethe effects of DBS on speech productionrdquo Journal of Neurolin-guistics vol 25 no 2 pp 74ndash94 2012

[20] V S Lee X P Zhou D A Rahn III E Q Wang and J JJiang ldquoPerturbation and nonlinear dynamic analysis of acousticphonatory signal in Parkinsonian patients receiving deep brainstimulationrdquo Journal of Communication Disorders vol 41 no 6pp 485ndash500 2008

[21] F Karlsson E Unger S Wahlgren et al ldquoDeep brain stim-ulation of caudal zona incerta and subthalamic nucleus inpatients with Parkinsonrsquos disease effects on diadochokineticraterdquo Parkinsonrsquos Disease vol 2011 Article ID 605607 10 pages2011

[22] A-P Benguerel and T Ulkrainetz Mcfadden ldquoThe effect ofcoarticulation on the role of transitions in vowel perceptionrdquoPhonetica vol 46 no 1ndash3 pp 80ndash96 1989

[23] D J Sharf and H Ostreicher ldquoEffect of forward and backwardcoarticulation on the identification of speech soundsrdquo Languageand speech vol 16 no 3 pp 196ndash206 1973

[24] R D Kent and F D Minifie ldquoCoarticulation in recent speechproduction modelsrdquo Journal of Phonetics vol 5 pp 115ndash1331977

[25] E Farnetani and D Recasens ldquoCoarticulation models in recentspeech production theoriesrdquo in CoarticulationmdashTheory Dataand Techniques W Hardcastle and N Hewlett Eds pp 31ndash67Cambrige University Press Cambridge UK 1st edition 2006

[26] K Tjaden ldquoAn acoustic study of coarticulation in dysarthricspeakers with Parkinson diseaserdquo Journal of Speech Languageand Hearing Research vol 43 no 6 pp 1466ndash1480 2000

[27] K Tjaden ldquoAnticipatory coarticulation inmultiple sclerosis andParkinsonrsquos diseaserdquo Journal of Speech Language and HearingResearch vol 46 no 4 pp 990ndash1008 2003

[28] Y Kondo ldquoPhonetic underspecification in schwardquo in Proceed-ings of the International Conference on Spoken Language pp 311ndash314 1994

[29] K N Stevens and A S House ldquoPerturbation of vowel articu-lations by consonantal context an acoustical studyrdquo Journal ofspeech and hearing research vol 6 pp 111ndash128 1963

[30] M Chafcouloff A Marchal and T Benkirane ldquoCoarticulatorypatterns in stop sequencesrdquo in Proceedings of the 1st Euro-pean Conference on Speech Communication and Technology(EUROSPEECH rsquo89) vol 2 pp 304ndash307 Paris France 1989

[31] Z S Nasreddine N A Phillips V Bedirian et al ldquoThe Mon-treal Cognitive Assessment MoCA a brief screening tool formild cognitive impairmentrdquo Journal of the American GeriatricsSociety vol 53 no 4 pp 695ndash699 2005


[32] H C Rossetti L H Lacritz C Munro Cullum and M FWeiner ldquoNormative data for the Montreal Cognitive Assess-ment (MoCA) in a population-based samplerdquo Neurology vol77 no 13 pp 1272ndash1275 2011

[33] V Woisard S Bodin and M Puech ldquoThe Voice HandicapIndex impact of the translation in French on the validationrdquoRevue de Laryngologie Otologie Rhinologie vol 125 no 5 pp307ndash312 2004

[34] International Phonetic Association The Principles of the Inter-national Phonetic Association Department of Phonetics Uni-versity College London UK 1949

[35] J-P Roy J MacOir VMartel-Sauvageau and C-A BoudreaultldquoTwo French-speaking cases of foreign accent syndrome anacousticphonetic analysisrdquo Clinical Linguistics and Phoneticsvol 26 no 11-12 pp 934ndash945 2012

[36] G Fairbanks Voice and Articulation Drillbook Harper amp RowNew York NY USA 1960

[37] C G Goetz B C Tilley S R Shaftman et al ldquoMovementdisorder society-sponsored revision of the unified Parkinsonrsquosdisease rating scale (MDS-UPDRS) scale presentation andclinimetric testing resultsrdquoMovement Disorders vol 23 no 15pp 2129ndash2170 2008

[38] B K Scanlon H L Katzen B E Levin C Singer and SPapapetropoulos ldquoA formula for the conversion of UPDRS-III scores to Hoehn and Yahr stagerdquo Parkinsonism and RelatedDisorders vol 14 no 4 pp 379ndash380 2008

[39] P Boersma and D Weenink ldquoPraat doing phonetics by com-puterrdquo 2013

[40] C Gendrot andM Adda-Decker ldquoImpact of duration on F1F2formant values of oral vowels an automatic analysis of largebroadcast news corpora in French and Germanrdquo in Proceedingsof the 9th European Conference on Speech Communicationand Technology (INTERSPEECH rsquo05) pp 2453ndash2456 LisbonPortugal September 2005

[41] B G Tabachnik and L S Fidell Using Multivariate StatisticsPearson Education New York NY USA 6th edition 2013

[42] IBM Corporation IBM SPSS Statistics for Windows IBMCorporation Armonk NY USA 2011

[43] L Max and P Onghena ldquoSome issues in the statistical analysisof completely randomized and repeated measures designs forspeech language and hearing researchrdquo Journal of SpeechLanguage and Hearing Research vol 42 no 2 pp 261ndash2701999

[44] H A Taba S S Wu K D Foote et al ldquoA closer look atunilateral versus bilateral deep brain stimulation results of theNational Institutes of Health COMPARE cohortrdquo Journal ofNeurosurgery vol 113 no 6 pp 1224ndash1229 2010

[45] P Krack V Fraix A Mendes A-L Benabid and P PollakldquoPostoperative management of subthalamic nucleus stimu-lation for parkinsonrsquos diseaserdquo Movement Disorders vol 17supplement 3 pp S188ndashS197 2002

[46] B Lindblom On Vowel Reduction The Royal Institute ofTechnology Stockholm Sweden 1963

[47] D P Kuehn and K L Moll ldquoPerceptual effects of forwardcoarticulationrdquo Journal of Speech and Hearing Research vol 15no 3 pp 654ndash664 1972

[48] C Dromey and S Bjarnason ldquoA preliminary report on disor-dered speech with deep brain stimulation in individuals withParkinsonrsquos diseaserdquo Parkinsonrsquos Disease vol 2011 Article ID796205 11 pages 2011

[49] A L Tornqvist L Schalen and S Rehncrona ldquoEffects of differ-ent electrical parameter settings on the intelligibility of speechin patients with Parkinsonrsquos disease treated with subthalamicdeep brain stimulationrdquoMovement Disorders vol 20 no 4 pp416ndash423 2005

[50] E Q Wang L V Metman R A E Bakay J Arzbaecher BBernard and D M Corcos ldquoHemisphere-specific effects ofbubthalamic nucleus deep brain stimulation on speaking rateand articulatory accuracy of syllable repetitions in Parkinsonrsquosdiseaserdquo Journal of Medical Speech-Language Pathology vol 14no 4 pp 323ndash334 2006

[51] K Tjaden and G E Wilding ldquoEffects of speaking task onintelligibility in Parkinsonrsquos diseaserdquo Clinical Linguistics andPhonetics vol 25 no 2 pp 155ndash168 2011

Submit your manuscripts athttpwwwhindawicom

Stem CellsInternational

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014


MEDIATORSINFLAMMATION

of


Behavioural Neurology

EndocrinologyInternational Journal of



Disease Markers


BioMed Research International

OncologyJournal of



Oxidative Medicine and Cellular Longevity


PPAR Research

The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014

Immunology ResearchHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Journal of

ObesityJournal of



Computational and Mathematical Methods in Medicine

OphthalmologyJournal of


Diabetes ResearchJournal of



Research and TreatmentAIDS


Gastroenterology Research and Practice


Parkinsonrsquos Disease

Evidence-Based Complementary and Alternative Medicine

Volume 2014Hindawi Publishing Corporationhttpwwwhindawicom


drastically reduce motor symptoms and improve patientsrsquoquality of life it has been associated with relatively smallchanges in dysarthria severity levels and intelligibility Inmost studies the impact of STNDBSondysarthriawas at bestmixed and was more often than not minor andor poor [14]For a review of the impact of STNDBS on speech systems see[15]

The articulation of speech sounds requires fine motorcontrol Speech units can be characterized in terms of artic-ulatory gestures (range of movement) or in terms of acousticdistinctiveness Acoustic distinctiveness refers to the fact thattwo speech units that are acoustically differentiated are moreeasily identified by our perceptual system [16] Using thisparadigm a category of soundsmdashthe vowelsmdashcan be acous-tically described and differentiated from one another by theiracoustic characteristicsThe principal acoustic characteristicsused to describe vowels are their formants The first twoformants of a vowel 1198651 and 1198652 are spectral values that allowcategorization of the phoneme 1198651 and 1198652 respectively serveas indicators of the open-close and front-back position of thearticulators (jaw tongue) during the production of speech[17] The articulation of vowels is very important for speechintelligibility and reduced acoustic distinctiveness of vowelshas been reported in studies of dysarthric speakers includingpeople with PD [18] In a study analyzing different metrics toevaluate the effects of STN DBS on speech production Weis-mer et al [19] suggested that vowel production has promisingresults regarding articulation measurements A few studies[20 21] investigated the changes in vowel production thatcould occur with STN DBS in PD However these studieslooked at speech components other than articulation such asvoice quality or speech rate

Coarticulation is another important acoustic factor forspeech intelligibility [22 23] Coarticulation can be definedas the influence speech units have on each other in connectedspeech [24] Coarticulation effects fall into two categories andare described as anticipatory (ldquoright to leftrdquo) or carry-over(ldquoleft to rightrdquo) Anticipatory coarticulation is generally con-sidered a consequence of motor planning because the posi-tions of the articulators for a given speech gesture are modi-fied in prediction of a following gesture Carry-over coartic-ulation on the other hand is considered a consequence ofarticulatory inertia and is influenced by the physical char-acteristics of the articulators [25] Initially investigated innormal speakers more and more studies looked at possiblechanges in coarticulation patterns in dysarthric speakers [2627] Coarticulation is amarker of speech finemotor control Itcan be acoustically measured with different methods includ-ing analyzing the influence of a given context on the articu-lation of the following phoneme (carry-over coarticulation)In some studies authors used this type of measurement tocharacterize coarticulation in normal speakers [28ndash30]Theyshowed that in a consonant-vowel (C-V) sequence the placeof articulation of the consonantmodulates the formant valuesof the vowel Some studies showed that STN DBS can inducechanges in fine motor control during speech production [15]One could therefore presume that coarticulation could beinfluenced by STN DBS However this hypothesis has notbeen specifically investigated in any previous studies

Even though articulatory function is very important forspeech intelligibility particularly for vowels only a few stud-ies examined the impact of STN DBS on articulation in PDin terms of range of movement Moreover to the best of ourknowledge no studies have investigated the coarticulationchanges that could occur with STN DBS Therefore the pri-mary objective of this study was to examine the effects of STNDBS on speech articulation in dysarthric speakers with PDusing acoustic measurements of vowelsThe second objectivewas to examine the coarticulation changes with STNDBS byanalyzing the impact of the preceding consonant context onvowel articulation which represents a measure of carry-overcoarticulation

2 Methods

21 Participants

211 Demographic Characteristics The study was approvedby the local institutional ethics committee for the safety ofhuman subjects and written informed consent was obtainedfrom all participants Eight participants (5women and 3men)aged 53ndash72 years with idiopathic PD diagnosed 9 to 25 yearsprior to the study were recruited in an outpatient clinic Allparticipants were taking antiparkinsonian medications withL-dopa equivalent dose from 300 to 1500mg per day andno changes were made to their medication during the studyAll of them were native speakers of Quebec French who hadalways lived in the province of Quebec Although no formalhearing evaluationwas conducted all participants were func-tional in conversation and none reported any hearing impair-ment General cognitive functions were measured using theMontreal Cognitive Assessment (MoCA) [31] and noparticipant fell below the cut-off score according to age andeducation level [32] Each participant with PD was also askedto complete the French version of the Voice Handicap Index(VHI) [33] which is one of the most common instrumentsdesigned to measure quality of life with respect to dysarthriain PD Higher values indicate greater voice handicapCharacteristics of each participant are reported in Table 1

212 Speech Characteristics Two speech-language pathol-ogists were asked to reach an agreement on dysarthriadiagnosis dysarthria severity and acoustic characteristics foreach participant during ON-stimulation Their evaluationwas based on audio recordings of each participant reading ldquoLabise et le soleilrdquo [34] a standardized French text commonlyused in French phonetic experiments to study normal orpathological speech [35] This text is considered the Frenchequivalent of the English textTheRainbow Passage [36] Eachrecording was presented to the two speech-language pathol-ogists independently in a quiet room via open-air speakersDysarthria severity ranged from mild to severe with alteredphonation reduced articulation and abnormal speech ratebeing the main acoustic features identified in most of therecordings All participants were diagnosed with hypokineticdysarthria which is consistent with the speech profile gener-ally observed in people with PD [13] Although the presence






















AVS =10038161003816100381610038161003816100381610038161003816


10038161003816100381610038161003816100381610038161003816

(1)


FCR = 1198652u + 1198652a + 1198651i + 1198651u1198652i + 1198651a

(2)





3 Results







(speech item)

OFF-stimulation 483 263 219(167) (074) (088)

ON-stimulation 297 226 206(918) (104) (062)


i u

a

200

300

400

500

600

700

800

5001000150020002500


F2 (Hz)

F1

(Hz)









OFF-stimulation 3115 22331 3292 11866 5833 17363 152036 1240 1302(544) (2190) (666) (2831) (1276) (2635) (14251) (0035) (88)

ON-stimulation 3076 23457 3285 10770 6801 17530 227759 1142 1384(451) (2644) (613) (2501) (911) (2680) (17905) (0031) (105)

Change +75723 minus0098 +82(11504) (020) (51)


lowast

Significant effect

0

5

10

15

20

25

30

35


LabialVelarAlveolar

Acou

stic v

owel

spac

e (H

z2)times10000



4 Discussion












5 Conclusions





Acknowledgments


References


























































of






Disease Markers



OncologyJournal of





PPAR Research



Journal of

ObesityJournal of





































AVS =10038161003816100381610038161003816100381610038161003816


10038161003816100381610038161003816100381610038161003816

(1)


FCR = 1198652u + 1198652a + 1198651i + 1198651u1198652i + 1198651a

(2)





3 Results







(speech item)

OFF-stimulation 483 263 219(167) (074) (088)

ON-stimulation 297 226 206(918) (104) (062)


i u

a

200

300

400

500

600

700

800

5001000150020002500


F2 (Hz)

F1

(Hz)









OFF-stimulation 3115 22331 3292 11866 5833 17363 152036 1240 1302(544) (2190) (666) (2831) (1276) (2635) (14251) (0035) (88)

ON-stimulation 3076 23457 3285 10770 6801 17530 227759 1142 1384(451) (2644) (613) (2501) (911) (2680) (17905) (0031) (105)

Change +75723 minus0098 +82(11504) (020) (51)


lowast

Significant effect

0

5

10

15

20

25

30

35


LabialVelarAlveolar

Acou

stic v

owel

spac

e (H

z2)times10000



4 Discussion












5 Conclusions





Acknowledgments


References


























































of






Disease Markers



OncologyJournal of





PPAR Research



Journal of

ObesityJournal of


















3 Results







(speech item)

OFF-stimulation 483 263 219(167) (074) (088)

ON-stimulation 297 226 206(918) (104) (062)


i u

a

200

300

400

500

600

700

800

5001000150020002500


F2 (Hz)

F1

(Hz)









OFF-stimulation 3115 22331 3292 11866 5833 17363 152036 1240 1302(544) (2190) (666) (2831) (1276) (2635) (14251) (0035) (88)

ON-stimulation 3076 23457 3285 10770 6801 17530 227759 1142 1384(451) (2644) (613) (2501) (911) (2680) (17905) (0031) (105)

Change +75723 minus0098 +82(11504) (020) (51)


lowast

Significant effect

0

5

10

15

20

25

30

35


LabialVelarAlveolar

Acou

stic v

owel

spac

e (H

z2)times10000



4 Discussion












5 Conclusions





Acknowledgments


References


























































of






Disease Markers



OncologyJournal of





PPAR Research



Journal of

ObesityJournal of





















OFF-stimulation 3115 22331 3292 11866 5833 17363 152036 1240 1302(544) (2190) (666) (2831) (1276) (2635) (14251) (0035) (88)

ON-stimulation 3076 23457 3285 10770 6801 17530 227759 1142 1384(451) (2644) (613) (2501) (911) (2680) (17905) (0031) (105)

Change +75723 minus0098 +82(11504) (020) (51)


lowast

Significant effect

0

5

10

15

20

25

30

35


LabialVelarAlveolar

Acou

stic v

owel

spac

e (H

z2)times10000



4 Discussion












5 Conclusions





Acknowledgments


References


























































of






Disease Markers



OncologyJournal of





PPAR Research



Journal of

ObesityJournal of
























5 Conclusions





Acknowledgments


References


























































of






Disease Markers



OncologyJournal of





PPAR Research



Journal of

ObesityJournal of

















Acknowledgments


References


























































of






Disease Markers



OncologyJournal of





PPAR Research



Journal of

ObesityJournal of










































of






Disease Markers



OncologyJournal of





PPAR Research



Journal of

ObesityJournal of
















research article changes in vowel articulation with...

Documents