Objective: Examine the temporal stability (i.e., repeatability) and sensitivity (i.e., change) of acoustic features of speech in people with Parkinson’s Disease (PWP).

Background: Voice and speech impairments often occur in PWP. While there is evidence to support the utility of acoustic features of speech for early detection of Parkinson’s Disease (PD), its validity for monitoring change over the disease course still requires systematic evaluation.

Method: Speech samples were acquired from 27 individuals with a confirmed diagnosis of PD, at 3-month and 12-month periods following a baseline assessment. This included sustaining the vowel “a” for a prolonged period, reading a short passage, and delivering a brief unprepared monologue. Eighty-three acoustic features capturing speech timing, vocal intensity and phonation were extracted. Stability of each acoustic feature was assessed used Cronbach α coefficients (α≥ 0.70 deemed stable) while sensitivity was established using paired T-tests (p <0.05 deemed sensitive).

Results: Of the acoustic features investigated, 63.9% and 55.4% were stable at 3-months and 12-months respectively (α≥ 0.70). Features encompassing speech timing from the reading and monologue tasks were most stable. Conversely, 7.2% and 14.5% of acoustic features were sensitive to change at 3-months and 12-months periods (p<0.05). Acoustic features related to speech timing (pause duration and total speech time) and vocal intensity (intensity range) were most sensitive to change at both 3-month and 12-month periods. Pause duration from monologue speech negatively correlated with UPDRS total scores ((r)=-0.43, p=0.05). No other significant relationships were found between acoustic features and PD disease severity, perceptual ratings of speech or voice-rated quality of life

Conclusion: Overall, many acoustic features remained stable over time irrespective of PD disease course, however, others varied widely, indicating their inappropriateness for PD disease course monitoring. This highlights the need for more rigorous systematic evaluation of not only acoustic features of speech, but also the type of speech stimuli collected if we are to utilize speech as a reliable surrogate endpoint in monitoring PWP.

« Back to MDS Virtual Congress 2021

MDS Abstracts - https://www.mdsabstracts.org/abstract/evaluation-of-acoustic-features-of-speech-for-monitoring-parkinsons-disease-progression/