Overenunciate this: Clear speech an acoustic study in Parkinson's Disease
MetadataShow full item record
Purpose: Research Question #1 sought to investigate how different definitions of Clear speech effect segmental and suprasegmental acoustic measures of speech in speakers with Parkinson's Disease (PD) and an age and sex-matched neurologically healthy control group. Research Question #2 sought to investigate the relationship between subjective ratings of Speaking Effort and acoustic measures associated with Clear speech adjustments. Methods: 28 speakers (14 PD; 14 Control), aged 55 to 81 years served as speakers. Each speaker produced 18 different sentences selected from the Sentence Intelligibility Test (SIT; Yorkston & Beukelman, 1996). Sentences were selected to have three to five occurrences of monophthongs (I, [varepsilon], υ, [wedge], i, æ, u, α), consonants ( t, k, s, [long esh]) and two to four occurrences of diphthongs (α I, eI). Speakers were instructed to produce stimuli in the following conditions: Habitual, Clear, Overenunciate, and Hearing Impaired. Two recordings of the Habitual condition were obtained to address potential influences of condition order, stimulus familiarity and/or fatigue. The Habitual condition was recorded twice, once at the beginning of the experiment (Pre-Habitual) and once at the end of the experiment (Post-Habitual). For the Habitual conditions, speakers were instructed to "say the following sentences". For the Clear condition, speakers were instructed to "speak clearly." In the Overenunciate and Hearing Impaired conditions, speakers were instructed to "overenunciate each word" and "say the following sentences while speaking to someone with a hearing impairment", respectively. For Research Question #1 a variety of acoustic measures were of interest. Segmental acoustic measures included tense and lax vowel space area, tense and lax vowel duration, fricative duration and first moment (M1) coefficient difference measures for stop and fricative pairs. The dynamic measure of F2 slope was included for two diphthongs. Suprasegmental measures included fundamental frequency (F0), Sound Pressure Level (SPL), articulation rate, speaking rate, pause frequency and pause duration. Data were analyzed using descriptive and parametric analyses. For Research Question #2 speakers used a visual analog scale to make judgements of self-perceived "Speaking Effort" in every condition. A subset of acoustic measures from those listed above was selected to investigate the relationship between Speaking Effort and clear speech adjustments. Data were analyzed using descriptive and parametric analyses. Results and Discussion: Results from Research Question #1 revealed that for a majority of acoustic adjustments, all three variants of Clear speech instruction differed from the Habitual condition. In general, the Overenunciate condition tended to elicit the greatest magnitude of change for segmental measures (vowel space area, vowel durations) as well as the slowest articulation and speaking rates. Moreover, the Hearing Impaired condition elicited the greatest fricative durations and suprasegmental adjustments in mean F0, F0 inter-quartile range (F0 IQR) and SPL. Although group differences were only observed for a few acoustic measures, group trends were in the expected direction consistent with acoustic differences reported in the PD literature. Finally, findings from Research Question #2 revealed a modest relationship between Speaking Effort and various Clear speech acoustic adjustments. For Control speakers, greater ratings of Speaking Effort were associated with greater acoustic adjustments in vowel space area, F2 slope of /eI/, F0 IQR, SPL and articulation rate. For the PD group, Speaking Effort was only associated with two acoustic measures, first moment coefficient difference measures for fricative pairs and F0 SD. Results have both theoretical and clinical implications. Theoretically, findings have implications for a model of speech production both in normal speakers as well as for speakers with dysarthria. Clinically, findings suggest that particular clear speech instructions may target distinct speech subsystems.