psychology article summariess

Brain & Language 165 (2017) 1–9

Contents lists available at ScienceDirect

Brain & Language

journal homepage: www.elsevier.com/locate/b&l

Distinct patterns of imprecise consonant articulation among Parkinson’s disease, progressive supranuclear palsy and multiple system atrophy

http://dx.doi.org/10.1016/j.bandl.2016.11.005 0093-934X/� 2016 Elsevier Inc. All rights reserved.

⇑ Corresponding author at: Department of Circuit Theory, Faculty of Electrical Engineering, Czech Technical University in Prague, Technicka 2, 166 27 Praha 6, Czech Republic and Department of Neurology and Centre of Clinical Neuroscience, First Faculty of Medicine, Charles University in Prague, Katerinska 30, 120 00 Prague 2, Czech Republic.

E-mail address: [email protected] (J. Rusz).

Tereza Tykalova a, Jan Rusz a,b,⇑, Jiri Klempir b,c, Roman Cmejla a, Evzen Ruzicka b a Department of Circuit Theory, Faculty of Electrical Engineering, Czech Technical University in Prague, Technicka 2, 166 27 Prague 6, Czech Republic b Department of Neurology and Centre of Clinical Neuroscience, First Faculty of Medicine, Charles University in Prague, Katerinska 30, 120 00 Prague 2, Czech Republic c Institute of Anatomy, First Faculty of Medicine, Charles University in Prague, U nemocnice 3, 128 00 Prague 2, Czech Republic

a r t i c l e i n f o

Article history: Received 14 July 2016 Revised 15 November 2016 Accepted 18 November 2016

Keywords: Parkinson’s disease Atypical Parkinsonian syndromes Voice onset time Dysarthria Speech disorders Acoustic analysis Perceptual assessment

a b s t r a c t

Distinct speech characteristics that may aid in differentiation between Parkinson’s disease (PD), progres- sive supranuclear palsy (PSP) and multiple system atrophy (MSA) remain tremendously under-explored. Here, the patterns and degree of consonant articulation deficits across voiced and voiceless stop plosives in 16 PD, 16 PSP, 16 MSA and 16 healthy control speakers were evaluated using acoustic and perceptual methods. Imprecise consonant articulation was observed across all Parkinsonian groups. Voice onset time of voiceless plosives was more prolonged in both PSP and MSA compared to PD, presumably due to greater severity of dysarthria and slower articulation rate. Voice onset time of voiced plosives was signif- icantly shorter only in MSA, likely as a consequence of damage to cerebellar structures. In agreement with the reduction of pre-voicing, MSA manifested increased number of voiced plosives misclassified as voice- less at perceptual evaluation. Timing of articulatory movements may provide important clues about the pathophysiology of underlying disease.

� 2016 Elsevier Inc. All rights reserved.

1. Introduction

Idiopathic Parkinson’s disease (PD) is a common neurological disorder characterized by the progressive loss of dopaminergic neurons in the substantia nigra pars compacta, affecting 1.6% of persons over the age of 65 years (deRijk et al., 1997). Dopamine concentrations have been shown to be significantly reduced before distinct motor deficits become apparent (Hornykiewicz, 1998). The cardinal signs of PD, often referred to as Parkinsonism, include resting tremor, bradykinesia, muscular rigidity and postural insta- bility. Other neurodegenerative diseases that go beyond the signs and symptoms of Parkinsonism are known as atypical Parkinsonian syndromes (APS). Progressive supranuclear palsy (PSP) and multi- ple system atrophy (MSA) are the most common APS, with an esti- mated prevalence of 30–40 per 100,000 among persons older than 65 years (Schrag, Ben-Shlomo, & Quinn, 1999). Characteristic clin- ical features of PSP include supranuclear gaze palsy, frequent falls, bradykinesia, axial rigidity, cognitive decline and communication

disorders (Nath, Ben-Shlomo, Thomson, Lees, & Burn, 2003; Steele, Olszewski, & Richardson, 1964), reflecting widespread neu- rodegeneration involving the midbrain as well as the globus pal- lidus, striatum, hypothalamic nucleus, pons, superior cerebellar peduncle and cerebellar dentate nucleus (Nath et al., 2003). Con- versely, MSA manifests by various combinations of autonomic, cerebellar and Parkinsonian features (Wenning, Colosimo, Geser, & Poewe, 2004), corresponding to degeneration of the cerebellum, middle cerebellar peduncle, striatum, substantia nigra, inferior oli- vary nucleus and pons (Gilman et al., 2008). APS differ from PD by poor levodopa response and more rapid disease progression result- ing in shorter life expectancy (O’Sullivan et al., 2008; Wenning, Litvan, & Tolosa, 2011). Furthermore, the underlying pathophysiol- ogy differs as PD and MSA are a-synucleinopathies while PSP is a tauopathy. However, the differentiation between PD and both PSP and MSA can be challenging as the initial signs are frequently nonspecific and overlap those of PD (Osaki et al., 2004; Schrag et al., 1999).

1.1. Speech impairment in PD, PSP and MSA

Dysarthria is a well-recognized clinical manifestation of Parkin- sonian disorders, developing in 90–100% of patients with PD, PSP and MSA during the course of the disease (Ho, Iansek, Marigliani, Bradshaw, & Gates, 1998; Kluin, Foster, Berent, & Gilman, 1993;

2 T. Tykalova et al. / Brain & Language 165 (2017) 1–9

Kluin, Gilman, Lohman, & Junck, 1996; Muller et al., 2001; Rusz et al., 2015). Speech impairment is an early and prominent mani- festation that can contribute primarily to the diagnosis of PSP (Goetz, Leurgans, Lang, & Litvan, 2003; Kim & McCann, 2015; Wenning et al., 2011), but has also been largely documented in the early stages of PD and MSA (Huh et al., 2015; Kim, Kent, Kent, & Duffy, 2010; Rusz, Cmejla, Ruzickova, & Ruzicka, 2011).

Due to dysfunction of the basal ganglia, the majority of PD patients manifest hypokinetic dysarthria characterized by mono- pitch, monoloudness, reduced stress, variable rate, imprecise artic- ulation, harsh voice quality, speech dysfluencies and inappropriate silence (Darley, Aronson, & Brown, 1969b; Ho et al., 1998). Con- versely, PSP and MSA patients typically evolve mixed dysarthria with a combination of hypokinesia, ataxia and spasticity as a result of more widespread neuronal atrophy (Kluin et al., 1993, 1996; Rusz et al., 2015). Indeed, previous studies (Kluin et al., 1993, 1996) investigating 46 MSA and 44 PSP patients using oral motor and perceptual speech analysis have reported mixed dysarthria with combinations of all hypokinetic, spastic and ataxic compo- nents in two-thirds of APS patients. Hypokinetic components fol- lowed by ataxic components were predominant in MSA patients, while spastic components were mostly present in PSP patients (Kluin et al., 1993, 1996).

Considering individual speech aspects, only the occurrence of stuttering-like behaviour has been reported to be distinctive for PSP as compared to MSA (Kluin et al., 1993, 1996; Rusz et al., 2015). A small number of studies have also focused on an objective description of the dysarthria profile in APS in comparison to PD (Huh et al., 2015; Kim et al., 2010; Rusz et al., 2015; Sachin et al., 2008; Skodda, Visser, & Schlegel, 2011). In general, these studies have shown that the impairment of specific speech dimensions is more pronounced in APS than in PD (Huh et al., 2015; Rusz et al., 2015; Skodda et al., 2011). Dysarthria in PSP has been reported to be characterized by stuttering-like behaviour, reduced speech rate, decreased intonation variability, prolonged pauses, articula- tion imprecision and poor quality of voice (Rusz et al., 2015; Skodda et al., 2011), whereas MSA patients have been said to man- ifest with excess pitch fluctuations, excess intensity variations, increased voice pitch, reduced speech rate, prolonged phonemes, vocal tremor, voice perturbations and slow variable alternating motion rates (Huh et al., 2015; Kim et al., 2010; Rusz et al., 2015; Saxena, Behari, Kumaran, Goyal, & Narang, 2014). However, little effort has been made to investigate consonant articulation in APS.

1.2. Consonant articulation in PD, PSP and MSA

The description of disturbed consonant articulation in various diseases has typically been based on perceptual assessment in sub- groups of patients defined by dysarthria subtype such as spastic, ataxic or hypokinetic, rather than by disease aetiology (i.e., PD, PSP or MSA; (Chakraborty, Roy, Hazra, Biswas, & Bhattacharya, 2008; Darley, Aronson, & Brown, 1969a; Hartelius, Gustavsson, Astrand, & Holmberg, 2006; Logemann & Fisher, 1981; Weismer, 1984). Furthermore, previous studies were limited primarily to documenting the occurrence of articulation deficits and did not describe specific features characterizing imprecise consonants (Chakraborty et al., 2008; Darley et al., 1969a; Hartelius et al., 2006). In particular, in the classic study by Darley et al. (1969b), imprecise consonant articulation was perceptually found to be one of the most deviant speech dimensions in PD. The presence of imprecise consonant articulation has also been perceptually revealed in a cohort of MSA and PSP patients (Hartelius et al., 2006). Interestingly, although in general speech deviation of greater severity was found in PSP, consonant articulation was more severely affected in MSA (Hartelius et al., 2006).

With regard to acoustic analyses, several measurements can be used for description of consonants including various measures of duration, formant transitions, spectral moments or energy-based measures (Kent & Read, 1992). Among them, voice onset time (VOT) determined for stop plosives is perhaps the most frequently used parameter and a relatively large amount of data has been published on VOT in PD patients. Unfortunately, previous studies have provided rather contradictory findings. While some research- ers have reported increased VOT duration (Forrest, Weismer, & Turner, 1989; Novotny, Rusz, Cmejla, & Ruzicka, 2014), others have observed unchanged (Fischer & Goberman, 2010; Ravizza, 2003) or even decreased VOT (Flint, Black, Campbelltaylor, Gailey, & Levinton, 1992) in PD subjects. It has been suggested that these discrepancies may be due to the fact that the measurement of VOT is dependent on speaking rate (Volaitis & Miller, 1992); how- ever, VOT ratio, a rate-independent variation of VOT, did not clarify these ambiguous findings (Fischer & Goberman, 2010; Novotny et al., 2014).

Only one previous study has focused on the acoustic investiga- tion of consonant characteristics for five categories of plosives in PD, PSP, and MSA in comparison to controls (Saxena et al., 2014). However, this study provided rather inconsistent findings across various consonant categories and speaker groups (Saxena et al., 2014). In particular, the authors revealed no significant alterations of VOT duration in dentals across all groups, but observed increased VOT duration of velars in PD, palatals in PSP, bilabials in MSA, PSP and PD, and of retroflexes in PSP and MSA (Saxena et al., 2014). However, a direct comparison of consonant articula- tion between PD, PSP and MSA has never been performed.

1.3. Aim of the present study

The aim of the current study was therefore to investigate the patterns and degree of consonant articulation deficits across differ- ent voiceless and voiced stop plosives in PD, PSP, MSA and healthy speakers using objective acoustic measures to help elucidate dis- tinct speech characteristics that could aid in the differentiation between various forms of Parkinsonism. In addition, perceptual examination of phonetic contrast between voiceless and voiced plosives was performed to determine if consonant imprecision was a notable feature of dysarthria in PD, PSP and MSA. Addition- ally, the relationships between speech performances and clinical manifestations were explored to provide greater insight into the pathophysiology of speech production in PD, PSP and MSA.

2. Methods

2.1. Participants

From 2011 to 2015, a total of 48 consecutive patients including 16 fulfilling the diagnostic criteria for idiopathic PD (5 men and 11 women), 16 with a diagnosis of probable PSP (11 men and 5 women) and 16 with a diagnosis of probable MSA (5 men and 11 women) were recruited. Among APS, hereafter hypernym for the MSA and PSP subgroups, 13 PSP patients were diagnosed with PSP-Richardson syndrome, 2 with PSP-Parkinsonism and 1 with PSP-pure akinesia with gait freezing, whereas MSA patients were diagnosed with the MSA-Parkinsonian subtype in 14 cases and the MSA-cerebellar subtype in 2 cases. The clinical diagnoses of all patients were established by a specialist in movement disorders (JK) according to the UK Parkinson’s Disease Society Bank Criteria for PD (Hughes, Daniel, Kilford, & Lees, 1992), the NINDS-PSP clin- ical diagnostic criteria for PSP (Litvan et al., 1996) or the consensus diagnostic criteria for MSA (Gilman et al., 2008). At the time of the examination, all patients treated pharmacologically were on stable

T. Tykalova et al. / Brain & Language 165 (2017) 1–9 3

medication for at least 4 weeks, consisting of various doses of levo- dopa alone or in combination with different dopamine agonists and/or amantadine. None of the patients received antipsychotic drugs. Disease duration was estimated based on the self-reported occurrence of first motor symptoms. APS patients were further scored according to the natural history and neuroprotection in Parkinson plus syndromes–Parkinson plus scale (NNIPPS; range from 0 to 346, where a higher score indicates more severe disabil- ity (Payan et al., 2011), while PD patients were rated according to the Unified Parkinson’s Disease Rating Scale motor subscore (UPDRS III; range from 0 to 108, where a higher score indicates more severe motor disability). Item 18 of the UPDRS III was used for the perceptual description of overall dysarthria severity. No patients reported a history of speech, language or hearing disorders unrelated to their Parkinsonism symptoms. No statistically signif- icant differences between the PSP and MSA groups were found for symptom duration, medication dose, motor severity or speech severity (Mann-Whitney U test: p = 0.14–0.98). Patient demo- graphic characteristics are summarized in Table 1.

The majority of investigated PD, PSP, and MSA subjects had also participated in a former study focused on the detailed assessment of severity and patterns of dysarthria (Rusz et al., 2015). In general, dysarthria was uniformly present in all Parkinsonian patients and ranged from mild to severe. The PD group manifested mild hypoki- netic dysarthria with dominant monopitch, imprecise vowels and inappropriate silences. The PSP group showed mild to moderate hypokinetic-spastic dysarthria with dominant dysfluency, slow speech rate, inappropriate silences, imprecise vowel articulation, monopitch and harsh voice quality. The MSA group presented mild to moderate ataxic-hypokinetic dysarthria dominated by excessive pitch and loudness variations, prolonged phonemes, vocal tremor and strained-strangled voice quality. However, characteristics related to imprecise consonant articulation were not previously reported.

The healthy control (HC) group consisted of 16 subjects (5 men and 11 women) of comparable age (mean age 62.8, SD 7.3, range 53–74 years). No significant differences in age distribution were detected between the HC, PD, PSP and MSA groups (analysis of variance: p = 0.30). Healthy individuals reported no history of neu- rological disease or any disorder that may affect speech, language or hearing.

All participants were Czech native speakers and were able to fully cooperate during the procedure. No subjects had signs of major depression or cognitive deficits that could interfere with

Table 1 Patient demographics.

PD (n = 16; 5 men, 11 women) Mean/SD (range)

Age (years) 63.3/7.3 (49–74) Symptom duration (years) 6.3/4.1 (2–14) Levodopa equivalent (mg/day) 781/471 (0–1680) Amantadine (mg/day) 100/136 (0–300) NNIPPS total – UPDRS III total 17.3/9.6 (7–38) UPDRS III speech item 0.7/0.8 (0–2) Tremor subscore 2.1/1.4 (0–5)a

Rigidity subscore 3.1/1.9 (1–7)a

Bradykinesia subscore 7.8/5.0 (2–18)a

Bulbar/pseudobulbar subscore – Cerebellar subscore –

PD, Parkinson’s disease; PSP, progressive supranuclear palsy; MSA, multiple system a Parkinson plus syndromes-Parkinson plus scale; UPDRS, unified Parkinson disease ratin

a UPDRS III subscore. b NNIPPS subscore. c 7 of 16 MSA patients manifested non-zero cerebellar subscore.

the measurements. The study was approved by the Ethics Commit- tee of the General University Hospital in Prague, Czech Republic, and all participants provided written, informed consent for the neurological examination and recording procedure.

2.2. Speech material

A series of tokens designed as ‘‘CVtka” were used for the assess- ment of consonant articulation, where C represented a consonant and V corresponded to a corner vowel. Specifically, six stop plo- sives were covered including /p/, /t/, /k/, /b/, /d/ and /g/. The vowels consisted of /a/, /ɪ/ and /u/. These 3 vowels were chosen as they are a representative sample with respect to vowel height, as vowel height has been reported to have an effect on some consonant characteristics such as VOT (Fischer & Goberman, 2010). From a phonetic point of view, /b/, /d/ and /g/ in the CV context are usually pronounced as prevoiced in Czech (i.e., voiced during closure), while /p/, /t/ and /k/ are pronounced as voiceless and unaspirated (Simackova, Podlipsky, & Chladkova, 2012). Furthermore, the stress is always on the first syllable in two-syllable words used without prepositions. The suffix /tka/ was added to evoke more naturally sounding tokens. Indeed, from 18 created stimuli, 7 were existing Czech words whereas the remaining 11 were meaningless. No analyses were performed on the suffix.

2.3. Recording procedure

The audio data were recorded in a quiet room with a low level of ambient noise using a head-mounted condenser microphone (Bayerdynamic Opus 55, Heilbronn, Germany) placed approxi- mately 5 cm from the subject’s lips. The speech signals were sam- pled at 48 kHz with 16-bit resolution. The recordings were collected during one session with a speech specialist who con- veyed instructions to the subjects. Each participant had to com- plete a series of speaking tasks as part of a larger protocol lasting approximately 20 min. There were no time limits during the recordings. All participants were asked to repeat their performance at any time if they or the examiner were not fully satisfied with their initial attempt. The performance of the task focused on con- sonant articulation was selected for further investigation. During the task, the participants were instructed to read the words pre- sented by the examiner on paper cards. The subjects were further warned not to be surprised as some of the words would be mean- ingless and to simply read what they saw. As Czech is a language

PSP MSA (n = 16; 11 men, 5 women) (n = 16; 5 men, 11 women) Mean/SD (range) Mean/SD (range)

66.1/5.1 (54–72) 61.9/5.9 (52–72) 4.3/2.4 (1–11) 3.8/1.3 (2–6) 447/538 (0–1500) 567/573 (0–1700) 169/178 (0–400) 131/145 (0–400) 71.9/28.5 (19–132) 71.8/22.5 (43–123) – – 2.1/0.7 (0–3) 1.7/0.7 (0–3) 2.2/2.6 (0–7)b 1.4/2.4 (0–9)b

3.6/3.1 (0–11)b 3.9/3.3 (0–11)b

22.1/10.9 (4–45)b 25.1/8.3 (12–39)b

9.2/4.2 (3–18)b 7.5/2.2 (4–12)b

0.0/0.0 (0–0)b 4.1/6.4 (0–22)b,c

trophy; SD, standard deviation; NNIPPS, natural history and neuroprotection on g scale.

4 T. Tykalova et al. / Brain & Language 165 (2017) 1–9

with fixed unambiguous pronunciation rules, no training of read- ing was performed. Tokens were printed in a large font on separate cards to provide optimal breathing patterns during reading and to minimize the effect of individual speech rate. Each of the 18 cards included one stimulus. The cards were presented at a stable pace, approximately one card per two seconds in quasi-randomized order. The entire task was repeated twice. As a result, a total of 36 tokens were obtained from each participant.

2.4. Acoustic analysis

Audio samples were analyzed using specialized speech software PRAAT� (Boersma, 2002). As previous research has shown that altered VOT is among the most powerful indicators of speech dis- order in PD (Novotny et al., 2014) and subjects with cerebellar atrophy present with a reduced categorical separation of the VOT of voiceless and voiced stop consonants (Ackermann, Graber, Hertrich, & Daum, 1999; Ackermann & Hertrich, 1997), three acoustic variables including VOT, VOT ratio and vowel duration were investigated. These acoustic variables were measured from the first syllable of each token by hand using both wide-band spec- trogram and oscillographic sound pressure signal displayed on screen. VOT was determined as the interval between the articula- tory release of stop and the onset of vocal fold vibration (Kent & Read, 1992). Both positive and negative values of VOT were allowed. The negative values of VOT refer to voicing lead or pre- voicing (voicing begins before the stop release) characteristic for voiced plosives (Kent & Read, 1992). If multiple bursts occurred, the initial burst was used to measure VOT (Fischer & Goberman, 2010). The VOT ratio was defined as VOT divided by the duration of whole syllable (Fischer & Goberman, 2010; Novotny et al., 2014).

All data were analyzed by one investigator (TT) as the selection of CV boundaries is a simple and well-defined task, as also docu- mented by previous researchers reporting high intra- and inter- rater reliability (Fischer & Goberman, 2010; Flint et al., 1992). The assigned values for each participant obtained from two vocal task runs were averaged. Subsequently, the values for each stop plosive were averaged across all three corner vowels. For statistical comparisons, the acoustic parameters were further assessed for a subset of voiceless (defined as the average of /p/, /t/ and /k/) and voiced (defined as the average of /b/, /d/ and /g/) consonants sepa- rately. No statistically significant differences between the speech performances of male and female participants across any acoustic variable or speakers group were found (Mann-Whitney U test: p = 0.12–0.86).

2.5. Non-measurable data

The acoustic variables were not obtained from a small propor- tion of data due to methodological constraints. In particular, the amount of omitted tokens was 6.1% in MSA, 3.3% in PSP and less than 1% in PD and HC speakers. The main reason for classification of stimuli as non-measurable was the absence of burst in the PSP group (52% of omitted tokens) and the absence of voicing, i.e., missing vowel or vowel lasting less than 4 pitch periods in the MSA group (29% of omitted tokens). As the missing data were likely to be distributed randomly, at least 3 tokens for each stop conso- nant were always available for further analysis.

2.6. Perceptual assessment

A total of 2240 recorded tokens from all four groups of partici- pants were used for perceptual analyses of phonetic contrast between voiceless and voiced plosives. The remaining 64 (2.8%) tokens were discarded as they were not used in acoustic analyses due to methodological constraints such as incomplete stop conso-

nant production, devoicing of vowels or the presence of dysfluen- cies. All tokens were randomized using a computer algorithm and subsequently separated to sequences containing a maximum of 100 tokens; each lasting approximately 10 min. An interval of at least 5 s duration with no speech production was included between two contiguous tokens to ensure enough time for listen- ers to note the perceived item. Three speech specialists (TT, JR, RC) performed the perceptual assessment using the same standard over-the-ear headphones (ATH-T500, Audio-technica) adjusted to comfortable volume. The listeners were instructed to replay each sequence of recordings and note the initial consonant of each token. At least 10 min break was required between two sequences. As a measure of perceived voicing contrast the percentage of /p/, /t/ or /k/ misassigned as /b/, /d/ or /g/ (Err/voiceless/) and vice versa (Err/voiced/) was determined. The final score was calculated by averaging perceptual ratings obtained across three raters. Re- analysis of 20% of all data was performed by the same investigator (TT) that performed the original set of measures. Based upon a two-way random single measures intra-class correlation, the esti- mated inter-rater reliability across three raters was 0.95 (p < 0.001) whereas intra-rater reliability across the same rater was 0.97 (p < 0.001).

2.7. Statistical analysis

All acoustic variables were normally distributed whereas per- ceptual metrics were not found to be normally distributed (Kolmogorov-Smirnov test). Group differences were calculated using analysis of variance for acoustic parameters and the Kruskal-Wallis test for perceptual measures. Post-hoc significance was assessed by the Fisher least-squares difference. Pearson and Spearman correlations were applied to test for significant relation- ships between normally and non-normally distributed data, respectively. Due to the exploratory nature of the study, adjust- ment for multiple comparisons was not performed and the level of significance was set to p < 0.05.

3. Results

Table 2 provides acoustic data for consonant articulation across all groups through six plosives including /p/, /b/, /t/, /d/, /k/ and /g/. Fig. 1 shows a representative example of stop consonant duration for the voiceless plosive /t/ and voiced plosive /d/ depicted for HC, PD, PSP and MSA subjects.

Fig. 2A depicts comparison of articulation performances among PD, PSP, MSA and HC subjects for subsets of voiceless plosives. VOT was found to be the best parameter for differentiating between groups [F(3,60) = 16.7, p < 0.001, g2 = 0.45]. Post hoc comparisons revealed significantly longer VOT in both PSP and MSA compared to PD or HC individuals (both p < 0.001). VOT ratio significantly dis- criminated between speaker groups [F(3,60) = 5.8, p = 0.002, g2 = 0.22], as the HC group manifested significantly smaller VOT ratio than all patient groups including PD (p < 0.05), PSP (p < 0.01) and MSA (p < 0.001). The vowel duration slightly varied among groups [F(3,60) = 2.9, p = 0.04, g2 = 0.13] due to the PSP group, which showed significantly longer vowel length compared to both HC (p < 0.05) and PD (p < 0.01).

Fig. 2B illustrates comparison of articulation performances among PD, PSP, MSA and HC subjects for subsets of voiced plosives. Group differences were found for both VOT [F(3,60) = 9.3, p < 0.001, g2 = 0.32] and VOT ratio [F(3,60) = 14.5, p < 0.001, g2 = 0.42]. Post hoc comparisons revealed significantly shorter neg- ative VOT in MSA as compared to all groups including HC (p < 0.001), PD (p < 0.01) and PSP (p < 0.001). Accordingly, a smaller negative VOT ratio was found in MSA as compared to HC, PD and

Table 2 Results of acoustic speech analyses for six plosives including /p/, /b/, /t/, /d/, /k/ and /g/. Bold numbers indicate patient values that significantly differ from the HC group (p < 0.05).

Bilabial plosives Alveolar plosives Velar plosives

/p/ /b/ /t/ /d/ /k/ /g/ Mean/SD Mean/SD Mean/SD Mean/SD Mean/SD Mean/SD

VOT (ms) HC 14/4 �100/37 20/5 �102/42 29/10 �109/32 PD 18/6 �69/23 24/10 �71/30 31/10 �76/30 PSP 28/8 �90/54 40/16 �94/53 51/19 �83/49 MSA 28/11 �43/33 39/10 �41/50 43/11 �22/60 VOT ratio (-) HC 0.17/0.04 �0.54/0.09 0.22/0.05 �0.53/0.09 0.30/0.07 �0.54/0.08 PD 0.22/0.06 �0.46/0.11 0.27/0.09 �0.44/0.11 0.34/0.08 �0.44/0.12 PSP 0.26/0.06 �0.42/0.17 0.31/0.09 �0.43/0.16 0.36/0.07 �0.37/0.14 MSA 0.26/0.08 �0.25/0.22 0.34/0.07 �0.19/0.29 0.35/0.12 �0.10/0.34 Vowel duration (ms) HC 69/13 78/17 73/15 84/18 67/16 88/22 PD 63/11 73/12 65/12 77/11 60/10 82/13 PSP 84/42 99/47 90/38 104/41 89/36 107/40 MSA 77/25 84/23 78/35 91/33 77/28 90/24

VOT, voice onset time; HC, healthy controls; PD, Parkinson’s disease; PSP, progressive supranuclear palsy; MSA, multiple system atrophy; SD, standard deviation.

Fig. 1. Representative example of stop consonant duration of voiceless plosive /t/ and voiced plosive /d/ depicted for HC, PD, PSP and MSA subjects. HC, healthy controls; PD, Parkinson’s disease; PSP, progressive supranuclear palsy; MSA, multiple system atrophy.

T. Tykalova et al. / Brain & Language 165 (2017) 1–9 5

Fig. 2. Acoustic measures of VOT, VOT ratio, and vowel duration across HC, PD, PSP and MSA speakers for a subset of (A) voiceless plosives and (B) voiced plosives. Symbols represent mean values and error bars represent SD values. Statistical comparison between groups: *p < 0.05, **p < 0.01, ***p < 0.001. VOT, voice onset time; HC, healthy controls; PD, Parkinson’s disease; PSP, progressive supranuclear palsy; MSA, multiple system atrophy.

6 T. Tykalova et al. / Brain & Language 165 (2017) 1–9

PSP (p < 0.001). Additionally, when compared to HC speakers, PD individuals exhibited shorter negative VOT (p < 0.05) and PSP patients smaller negative VOT ratio (p < 0.05). The vowel duration also varied among groups [F(3,60) = 2.8, p < 0.05, g2 = 0.12] due to the PSP group, which exhibited slightly longer vowel length com- pared to both HC (p < 0.05) and PD (p < 0.01).

Fig. 3 displays results of perceptual analyses of phonetic con- trast between voiceless and voiced plosives among PD, PSP, MSA and HC groups. No statistically significant difference between groups for Err/voiceless/ was observed [v2(3,60) = 3.9, p = 0.27, g2 = 0.06]. In contrast, Err/voiced/ varied significantly among groups [v2(3,60) = 17.1, p < 0.001, g2 = 0.27] due to the MSA group, which showed more misassignment errors than HC (p < 0.001), PD (p < 0.01), or PSP (p < 0.05).

The cerebellar motor impairment in APS patients correlated with a reduction of pre-voicing (cerebellar NNIPPS subscore vs. voiced VOT: r = 0.59, p < 0.001; cerebellar NNIPPS subscore vs. voiced VOT ratio: r = 0.56, p = 0.001) as well as with the percentage of voiced plosives perceptually misassigned as voiceless plosives (cerebellar NNIPPS subscore vs. Err/voiced/: r = 0.58, p < 0.001). No other significant correlations were detected between speech parameters and non-speech motor subscores in Parkinsonian groups. Additionally, overall dysarthria severity in all patients sub- groups correlated with a prolonged duration of VOT of voiceless plosives (UPDRS III speech item vs. voiceless VOT: r = 0.50, p < 0.001). Significant correlation was also observed between the reduction of pre-voicing and percentage of voiced plosives percep- tually misassigned as voiceless plosives (voiced VOT vs. Err/ voiced/: r = 0.65, p < 0.001; voiced VOT ratio vs. Err/voiced/: r = 0.73, p < 0.001).

4. Discussion

Although disturbed consonant articulation is considered to be among the most deviant speech dimensions in all dysarthria sub- types (Chakraborty et al., 2008; Darley et al., 1969b), the present study represents the first attempt to clarify potential differences in consonant articulation deficits of both voiced and voiceless plo- sives in PD, PSP and MSA. We observed divergent patterns of artic- ulation abnormalities for a subset of voiceless and voiced stop consonants in PD, PSP and MSA groups. Voice onset time of voice- less plosives was found to be more prolonged in both PSP and MSA compared to PD, likely due to the greater severity of dysarthria and slower articulation rate in APS. Voice onset time of voiced plosives was revealed to be significantly shorter only in MSA presumably as a consequence of damage to cerebellar structures. This finding was further supported by perceptual evaluation where only MSA patient target words with initial voiced plosive were misassigned as those with initial voiceless plosive. Acoustic analysis demon- strated that a slight deterioration of consonant articulation for both voiced and voiceless plosives was observed in PD as compared to HC. The observed trends with respect to imprecise articulation in PD, PSP and MSA were consistent among individual consonants as well as for averaged groups of voiceless and voiced consonants. The current method based on reading isolated two-syllabic words appears to be clinically feasible for assessing consonant articula- tion in Parkinsonism as the mean rate of measurable tokens was found to be greater than 93% for APS and 99% for PD or HC, and in this regard seems to be even better suited than the method based upon rhythmic syllable repetitions, where the mean rate of measurable VOT in dysarthrias was 85% (Ozsancak, Auzou, Jan, & Hannequin, 2001).

Fig. 3. Perceptual evaluation of voicing contrast based on Err/voiceless/ and Err/ voiced/ measures across HC, PD, PSP and MSA speakers depicted using boxplots. Statistical comparison between groups: *p < 0.05, **p < 0.01, ***p < 0.001. Err/voice- less/, percentage of voiceless plosives perceptually misassigned as voiced plosives; Err/voiced/, percentage of voiced plosives perceptually misassigned as voiceless plosives; HC, healthy controls; PD, Parkinson’s disease; PSP, progressive supranu- clear palsy; MSA, multiple system atrophy.

T. Tykalova et al. / Brain & Language 165 (2017) 1–9 7

Regarding voiceless plosives, significantly longer VOT was observed in both PSP and MSA subjects in comparison to PD, likely as a result of greater speech impairment in APS. Indeed, we observed significant relationship between overall dysarthria sever- ity and voiceless VOT. The observed differences between PD and APS also appear to be at least partially attributed to alterations in speaking rate as they were not preserved using the measurement of VOT ratio. Slow speaking rate is well documented in APS (Huh et al., 2015; Rusz et al., 2015; Skodda et al., 2011) and is primarily due to the presence of spastic components of dysarthria in APS patients (Rusz et al., 2015). In line with this observation, prolonged vowel duration associated, to a certain extent, with decreased speech rate was observed only in PSP patients, which commonly manifest spastic elements of dysarthria (Kluin et al., 1993).

On the contrary, the articulation of voiced plosives was mark- edly impaired only in MSA patients while nearly intact in PD and PSP subjects. In particular, the voiced plosives in MSA were charac- terized by the shortening of negative VOT duration until the voic- ing lead completely vanished and only short burst remained. In other words, the voiced plosives were occasionally pronounced as their voiceless cognates. Notably, contrary to voiceless plosives, these trends in voiced plosives were not suppressed using VOT ratio and thus cannot be interpreted as a simple effect of decreased speaking rate. Moreover, voiced VOT was not affected in PSP despite a similar overall severity of motor and speech impairment

as compared to MSA. Consequently, we hypothesize that the dis- ruption of voiced plosives in MSA reflects underlying cerebellar neurodegeneration. This hypothesis is further supported by the correlation between the severity of cerebellar involvement in APS and the disruption of voiced plosives as reflected by VOT and VOT ratio.

In agreement with the reduction of pre-voicing, the present study revealed a reduced perceptual contrast between voiced and voiceless plosives in MSA. This observation is in accordance with previous studies where reduced phonological contrast between / t/ and /d/ was documented in patients suffering from cerebellar atrophy (Ackermann & Hertrich, 1997; Ackermann et al., 1999), and generally confirms that VOT plays a primary role in perceptual determination between voiceless and voiced plosive cognates (Auzou et al., 2000). As perceptual judgment appears to be suffi- cient to reveal effects of cerebellar patterns on production of the voicing contrast in mild to moderate dysarthria, the current find- ings may have wider clinical implications in the evaluation of patients who may have MSA or other cerebellar degenerations.

Our further results indicate disturbed coordination between laryngeal and supralaryngeal articulators in PD patients that man- ifested a slightly reduced negative VOT as well as greater positive VOT ratio compared to HC. These findings are in agreement with previous research showing affected consonant articulation in PD (Auzou et al., 2000; Flint et al., 1992; Novotny et al., 2014; Saxena et al., 2014). However, previous results related to altered voice onset time in hypokinetic dysarthria secondary to PD are rather contradictory (Auzou et al., 2000; Fischer & Goberman, 2010; Saxena et al., 2014). Notably, our PD subjects had a tendency to prolong the duration of VOT in plosives with a naturally short VOT length, while shorten the duration of VOT in plosives with a naturally long VOT length. Although these trends may appear inconsistent, they can be interpreted with respect to knowledge about VOT categories in normal healthy speakers (Auzou et al., 2000). In most languages, VOT values for voiced and voiceless stops are in discrete duration ranges that correspond to one of three voicing categories including long negative VOT, short VOT and long positive VOT (Auzou et al., 2000). Thus, it can be assumed that in PD subjects, the plosives with short VOT duration are likely to be unchanged or extended beyond normal, while the plosives with both positive and negative long VOT are biased to be reduced. Indeed, previous studies in dysarthrias investigating stops with obviously long VOT duration reported shortening (Flint et al., 1992; Morris, 1989) while those examining naturally short VOT found increased or unchanged duration (Fischer & Goberman, 2010; Forrest et al., 1989; Novotny et al., 2014; Ravizza, 2003). However, other factors such as the effect of different speaking tasks or speaking rates cannot be excluded.

It thus remains to be elucidated by what mechanism the impaired programming of movements due to basal ganglia and cerebellar control circuit involvement contributes to imprecise articulation of voiced and voiceless stop consonants. Converging evidence from neuroimaging, limb control and neuropsychological studies suggests that patients with PD are unable to maintain a programmed response or to rapidly switch between responses, whereas patients with cerebellar disease have a reduced ability to program movement sequences in advance of movement onset (Spencer & Rogers, 2005). The accurate production of stop conso- nants requires close coordination between the larynx and the artic- ulators (i.e., lips, tongue and jaw). Production of word-initial voiceless plosives involves a period of articulatory closure during which the vocal folds are maintained in a relatively open position without glottal pulsing, whereas word-initial voiced plosives are characterized by voicing lead at the beginning followed by a period of articulatory closure. To achieve precise coordination of the glot- tal opening gesture and articulatory closure release, it has been

8 T. Tykalova et al. / Brain & Language 165 (2017) 1–9

suggested that the control of speech movement timing occurs over aggregates rather than individual articulators (Lofqvist & Lindblom, 1994; Munhall, Lofqvist, & Kelso, 1994). Difficulty in ini- tiating articulation due to a reduced ability to maintain the speech motor program, which is characteristic for speakers with hypoki- netic dysarthria (Gurd, Bessell, Watson, & Coleman, 1998; Svensson, Henningson, & Karlsson, 1993), appears to contribute to prolonged positive VOT in voiceless plosives and slightly short- ened negative VOT in voiced plosives in PD, PSP as well as MSA patients, as observed in the present study. In MSA, among deficits aggravated due to the presence of hypokinetic dysarthria, insuffi- cient programming in advance of speech onset due to ataxic dysar- thria may lead to additional disruption of coordination between the larynx and the articulators, presumably resulting in substan- tially shortened or completely missing negative VOT in voiced plosives.

There are certain limitations to the present study. Due to lim- ited opportunities in recruiting a larger number of participants we were not able to balance patient groups for gender, particularly in the PSP group where male participants predominated. We there- fore cannot exclude the possibility that the observed changes in consonant articulation are partially affected by gender-specific aspects of speech. Nevertheless, no evident gender-specific differ- ences for VOT have been reported in healthy adults in other lan- guages (Lundeborg, Larsson, Wiman, & McAllister, 2012; Morris, McCrea, & Herring, 2008). Furthermore, as articulation of voiced consonants is not characterized by voicing lead in all languages (Auzou et al., 2000), our findings related to voiced stops are likely to be language-specific and may not be easily generalized to all languages. One potential limitation is that the test material was comprised of two-syllabic single-words only and thus the token- initial position of the target consonants precluded the evaluation of the consonantal closing gesture. In addition, our speech material was designed to minimize the effect of speaking rate and thus cur- rent findings may differ from those obtained from more complex speech utterances. Indeed, previous research assumed that the motor control of sentence utterances differs from that of syllable repetition (Ackermann & Hertrich, 1993). In particular, while there is clear evidence that VOT extracted from single syllables is sensi- tive to changes due to cerebellar atrophy (Ackermann & Hertrich, 1997; Ackermann et al., 1999), no alterations were observed in VOT obtained from sentence utterances in Friedreich ataxia (Ackermann & Hertrich, 1993), which can be considered a model of afferent cerebellar dysfunction. Thus, the cerebellum might not be a prerequisite for the initiation of articulatory gestures within the framework of sentence utterances (Ackermann & Hertrich, 1993), and single-word material may be a more feasible task than longer sentence utterances to observe distinctive pat- terns of imprecise articulation among PD and APS.

5. Conclusions

In conclusion, our results confirm the distinctive critical role of basal ganglia and cerebellar control circuit involvement in articula- tory undershoot of voiced and voiceless stop consonants. In partic- ular, duration of VOT in voiced plosives was revealed to be shorter only in MSA, while nearly intact in PD and PSP subjects. Consider- ing that the only distinctive speech feature currently known for APS is the occurrence of stuttering-like behaviour in PSP (Kluin et al., 1993, 1996; Rusz et al., 2015), the alterations in VOT of voiced plosives may represent a novel marker of cerebellar dys- function in MSA. Further studies are needed to elaborate our find- ings in other languages, through various kind of speech material, in larger populations of PD and APS, as well as across different types of neurodegeneration with cerebellar atrophy.

Acknowledgement

This study was supported by the Ministry of Health of the Czech Republic, grant nr. 15-28038A. All rights reserved.

References

Ackermann, H., Graber, S., Hertrich, I., & Daum, I. (1999). Phonemic vowel length contrasts in cerebellar disorders. Brain and Language, 67(2), 95–109. http://dx. doi.org/10.1006/brln.1998.2044.

Ackermann, H., & Hertrich, I. (1993). Dysarthria in Friedreich’s ataxia: Timing of speech segments. Clinical Linguistics & Phonetics, 7(1), 75–91.

Ackermann, H., & Hertrich, I. (1997). Voice onset time in ataxic dysarthria. Brain and Language, 56(3), 321–333. http://dx.doi.org/10.1006/brln.1997.1740.

Auzou, P., Ozsancak, C., Morris, R. J., Jan, M., Eustache, F., & Hannequin, D. (2000). Voice onset time in aphasia, apraxia of speech and dysarthria: A review. Clinical Linguistics & Phonetics, 14(2), 131–150.

Boersma, P. (2002). Praat, a system for doing phonetics by computer. Glot International, 5(9/10), 341–345.

Chakraborty, N., Roy, T., Hazra, A., Biswas, A., & Bhattacharya, K. (2008). Dysarthric Bengali speech: A neurolinguistic study. Journal of Postgraduate Medicine, 54(4), 268–272.

Darley, F. L., Aronson, A. E., & Brown, J. R. (1969a). Clusters of deviant speech dimensions in the dysarthrias. Journal of Speech and Hearing Research, 12(3), 462–496.

Darley, F. L., Aronson, A. E., & Brown, J. R. (1969b). Differential diagnostic patterns of dysarthria. Journal of Speech and Hearing Research, 12(2), 246.

deRijk, M. C., Tzourio, C., Breteler, M. M. B., Dartigues, J. F., Amaducci, L., LopezPousa, S., ... Rocca, W. A. (1997). Prevalence of Parkinsonism and Parkinson’s disease in Europe: The EUROPARKINSON collaborative study. Journal of Neurology Neurosurgery and Psychiatry, 62(1), 10–15. http://dx.doi.org/10.1136/ Jnnp.62.1.10.

Fischer, E., & Goberman, A. M. (2010). Voice onset time in Parkinson disease. Journal of Communication Disorders, 43(1), 21–34. http://dx.doi.org/10.1016/j. jcomdis.2009.07.004.

Flint, A. J., Black, S. E., Campbelltaylor, I., Gailey, G. F., & Levinton, C. (1992). Acoustic analysis in the differentiation of Parkinsons-disease and major depression. Journal of Psycholinguistic Research, 21(5), 383–399. http://dx.doi.org/10.1007/ Bf01067922.

Forrest, K., Weismer, G., & Turner, G. S. (1989). Kinematic, acoustic, and perceptual analyses of connected speech produced by Parkinsonian and normal geriatric adults. Journal of the Acoustical Society of America, 85(6), 2608–2622. http://dx. doi.org/10.1121/1.397755.

Gilman, S., Wenning, G. K., Low, P. A., Brooks, D. J., Mathias, C. J., Trojanowski, J. Q., ... Vidailhet, M. (2008). Second consensus statement on the diagnosis of multiple system atrophy. Neurology, 71(9), 670–676. http://dx.doi.org/10.1212/01. wnl.0000324625.00404.15.

Goetz, C. G., Leurgans, S., Lang, A. E., & Litvan, I. (2003). Progression of gait, speech and swallowing deficits in progressive supranuclear palsy. Neurology, 60(6), 917–922.

Gurd, J. M., Bessell, N., Watson, I., & Coleman, J. (1998). Motor speech versus digit control in Parkinson’s disease: A cognitive neuropsychology investigation. Clinical Linguistics & Phonetics, 12(5), 357–378. http://dx.doi.org/10.3109/ 02699209808985231.

Hartelius, L., Gustavsson, H., Astrand, M., & Holmberg, B. (2006). Perceptual analysis of speech in multiple system atrophy and progressive supranuclear palsy. Journal of Medical Speech-Language Pathology, 14(4), 241–247.

Ho, A. K., Iansek, R., Marigliani, C., Bradshaw, J. L., & Gates, S. (1998). Speech impairment in a large sample of patients with Parkinson’s disease. Behavioural Neurology, 11(3), 131–137.

Hornykiewicz, O. (1998). Biochemical aspects of Parkinson’s disease. Neurology, 51 (2), S2–S9.

Hughes, A. J., Daniel, S. E., Kilford, L., & Lees, A. J. (1992). Accuracy of clinical- diagnosis of idiopathic Parkinson’s-disease – A clinicopathological study of 100 cases. Journal of Neurology Neurosurgery and Psychiatry, 55(3), 181–184. http:// dx.doi.org/10.1136/jnnp.55.3.181.

Huh, Y. E., Park, J., Suh, M. K., Lee, S. E., Kim, J., Jeong, Y., ... Cho, J. W. (2015). Differences in early speech patterns between Parkinson variant of multiple system atrophy and Parkinson’s disease. Brain and Language, 147, 14–20. http:// dx.doi.org/10.1016/j.bandl.2015.04.007.

Kent, R. D., & Read, C. (1992). The acoustic analysis of speech. San Diego, Calif.: Singular Pub. Group.

Kim, Y., Kent, R. D., Kent, J. F., & Duffy, J. R. (2010). Perceptual and acoustic features of dysarthria in multiple system atrophy. Journal of Medical Speech-Language Pathology, 18(4), 66–70.

Kim, J. H., & McCann, C. M. (2015). Communication impairments in people with progressive supranuclear palsy: A tutorial. Journal of Communication Disorders, 56, 76–87. http://dx.doi.org/10.1016/j.jcomdis.2015.06.002.

Kluin, K. J., Foster, N. L., Berent, S., & Gilman, S. (1993). Perceptual analysis of speech disorders in progressive supranuclear palsy. Neurology, 43(3), 563–566.

Kluin, K. J., Gilman, S., Lohman, M., & Junck, L. (1996). Characteristics of the dysarthria of multiple system atrophy. Archives of Neurology, 53(6), 545–548.

Litvan, I., Agid, Y., Calne, D., Campbell, G., Dubois, B., Duvoisin, R. C., ... Verny, M. (1996). Clinical research criteria for the diagnosis of progressive supranuclear

T. Tykalova et al. / Brain & Language 165 (2017) 1–9 9

palsy (Steele-Richardson-Olszewski syndrome): Report of the NINDS-SPSP international workshop. Neurology, 47(1), 1–9.

Lofqvist, A., & Lindblom, B. (1994). Speech motor control. Current Opinion in Neurobiology, 4(6), 823–826.

Logemann, J. A., & Fisher, H. B. (1981). Vocal-tract control in Parkinsons-disease – Phonetic feature analysis of misarticulations. Journal of Speech and Hearing Disorders, 46(4), 348–352.

Lundeborg, I., Larsson, M., Wiman, S., & McAllister, A. M. (2012). Voice onset time in Swedish children and adults. Logopedics Phoniatrics Vocology, 37(3), 117–122. http://dx.doi.org/10.3109/14015439.2012.664654.

Morris, R. J. (1989). VOT and dysarthria: A descriptive study. Journal of Communication Disorders, 22(1), 23–33.

Morris, R. J., McCrea, C. R., & Herring, K. D. (2008). Voice onset time differences between adult males and females: Isolated syllables. Journal of Phonetics, 36(2), 308–317.

Muller, J., Wenning, G. K., Verny, M., McKee, A., Chaudhuri, K. R., Jellinger, K., ... Litvan, I. (2001). Progression of dysarthria and dysphagia in postmortem- confirmed Parkinsonian disorders. Archives of Neurology, 58(2), 259–264. http:// dx.doi.org/10.1001/archneur.58.2.259.

Munhall, K. G., Lofqvist, A., & Kelso, J. A. S. (1994). Lip-larynx coordination in speech – Effects of mechanical perturbations to the lower lip. Journal of the Acoustical Society of America, 95(6), 3605–3616. http://dx.doi.org/10.1121/1.409929.

Nath, U., Ben-Shlomo, Y., Thomson, R. G., Lees, A. J., & Burn, D. J. (2003). Clinical features and natural history of progressive supranuclear palsy – A clinical cohort study. Neurology, 60(6), 910–916.

Novotny, M., Rusz, J., Cmejla, R., & Ruzicka, E. (2014). Automatic evaluation of articulatory disorders in Parkinson’s disease. IEEE-ACM Transactions on Audio Speech and Language Processing, 22(9), 1366–1378. http://dx.doi.org/10.1109/ Taslp.2014.2329734.

Osaki, Y., Ben-Shlomo, Y., Lees, A. J., Daniel, S. E., Colosimo, C., Wenning, G. K., & Quinn, N. (2004). Accuracy of clinical diagnosis of progressive supranuclear palsy. Movement Disorders, 19(2), 181–189. http://dx.doi.org/10.1002/ mds.10680.

O’Sullivan, S. S., Massey, L. A., Williams, D. R., Silveira-Moriyama, L., Kempster, P. A., Holton, J. L., ... Lees, A. J. (2008). Clinical outcomes of progressive supranuclear palsy and multiple system atrophy. Brain, 131, 1362–1372. http://dx.doi.org/ 10.1093/brain/awn065.

Ozsancak, C., Auzou, P., Jan, M., & Hannequin, E. (2001). Measurement of voice onset time in dysarthric patients: Methodological considerations. Folia Phoniatrica Et Logopaedica, 53(1), 48–57. http://dx.doi.org/10.1159/000052653.

Payan, C. A. M., Viallet, F., Landwehrmeyer, B. G., Bonnet, A. M., Borg, M., Durif, F., ... Grp, N. S. (2011). Disease severity and progression in progressive supranuclear palsy and multiple system atrophy: Validation of the NNIPPS - PARKINSON PLUS SCALE. PLoS ONE, 6(8). http://dx.doi.org/10.1371/journal.pone.0022293. Doi: ARTN e22293.

Ravizza, S. M. (2003). Dissociating the performance of cortical and subcortical patients on phonemic tasks. Brain and Cognition, 53(2), 301–310. http://dx.doi. org/10.1016/S0278-2626(03)00131-3.

Rusz, J., Bonnet, C., Klempir, J., Tykalova, T., Baborova, E., Novotny, M., ... Ruzicka, E. (2015). Speech disorders reflect differing pathophysiology in Parkinson’s

disease, progressive supranuclear palsy and multiple system atrophy. Journal of Neurology, 262(4), 992–1001. http://dx.doi.org/10.1007/s00415-015-7671-1.

Rusz, J., Cmejla, R., Ruzickova, H., & Ruzicka, E. (2011). Quantitative acoustic measurements for characterization of speech and voice disorders in early untreated Parkinson’s disease. Journal of the Acoustical Society of America, 129(1), 350–367. http://dx.doi.org/10.1121/1.3514381.

Sachin, S., Shukla, G., Goyal, V., Singh, S., Aggarwal, V., Gureshkumar & Behari, M. (2008). Clinical speech impairment in Parkinson’s disease, progressive supranuclear palsy, and multiple system atrophy. Neurology India, 56(2), 122–126.

Saxena, M., Behari, M., Kumaran, S. S., Goyal, V., & Narang, V. (2014). Assessing speech dysfunction using BOLD and acoustic analysis in Parkinsonism. Parkinsonism & Related Disorders, 20(8), 855–861. http://dx.doi.org/10.1016/ j.parkreldis.2014.04.024.

Schrag, A., Ben-Shlomo, Y., & Quinn, N. P. (1999). Prevalence of progressive supranuclear palsy and multiple system atrophy: A cross-sectional study. Lancet, 354(9192), 1771–1775.

Simackova, S., Podlipsky, V. J., & Chladkova, K. (2012). Czech spoken in Bohemia and Moravia. Journal of the International Phonetic Association, 42(2), 225–232. http:// dx.doi.org/10.1017/S0025100312000102.

Skodda, S., Visser, W., & Schlegel, U. (2011). Acoustical analysis of speech in progressive supranuclear palsy. Journal of Voice, 25(6), 725–731. http://dx.doi. org/10.1016/j.jvoice.2010.01.002.

Spencer, K. A., & Rogers, M. A. (2005). Speech motor programming in hypokinetic and ataxic dysarthria. Brain and Language, 94(3), 347–366. http://dx.doi.org/ 10.1016/j.bandl.2005.01.008.

Steele, J. C., Olszewski, J., & Richardson, J. C. (1964). Progressive supranuclear palsy – Heterogeneous degeneration involving brain stem basal ganglia + cerebellum with vertical gaze + pseudobulbar palsy nuchal dystonia + dementia. Archives of Neurology, 10(4), 333.

Svensson, P., Henningson, C., & Karlsson, S. (1993). Speech motor control in Parkinson’s disease: A comparison between a clinical assessment protocol and a quantitative analysis of mandibular movements. Folia Phoniatrica et Logopaedica (Basel), 45(4), 157–164.

Volaitis, L. E., & Miller, J. L. (1992). Phonetic prototypes – Influence of place of articulation and speaking rate on the internal structure of voicing categories. Journal of the Acoustical Society of America, 92(2), 723–735. http://dx.doi.org/ 10.1121/1.403997.

Weismer, G. (1984). Articulatory characteristics of Parkinsonian dysarthria: Segmental and phrase-level timing, spirantization, and glottal-supraglottal coordination. In M. R. McNeil, J. C. Rosenbek, & A. E. Aronson (Eds.), The dysarthrias. Austin, Texas: Pro-Ed.

Wenning, G. K., Colosimo, C., Geser, F., & Poewe, W. (2004). Multiple system atrophy. Lancet Neurology, 3(2), 93–103. http://dx.doi.org/10.1016/S1474-4422 (03)00662-8.

Wenning, G. K., Litvan, I., & Tolosa, E. (2011). Milestones in atypical and secondary Parkinsonisms. Movement Disorders, 26(6), 1083–1095. http://dx.doi.org/ 10.1002/mds.23713.