Acoustic analysis of sustained vowels in Parkinson’s disease: new insights into the differences of UK- and US-English speaking participants from the Parkinson’s voice initiative

Sustained vowels have often been used to clinically assess vocal performance and infer symptoms in Parkinson’s disease, with most studies focusing on cohorts from a single linguistic background. Arguably, sustained vowels are generic and language-independent, however it is not clear how findings might generalize across cohorts of people from different linguistic backgrounds. In this study, we aimed to compare phonations from UK- and US-English speaking people with Parkinson’s disease using the largest known speech-Parkinson’s database collected using a standard telephone network, the Parkinson’s Voice Initiative (PVI). We processed 1988 sustained vowel /a/ phonations from the US-cohort and 525 phonations from the UK-cohort. We stratified data according to gender and computed the fundamental frequency (F0) as a function of age and characterized phonations using 307 acoustic measures that we have used in previous related work. There was generally very good agreement between UK- and US-English speakers in terms of F0 characteristics and traditional acoustic measures such as jitter and shimmer. However, we find pronounced cohort differences with a few of the complex nonlinear acoustic measures. These findings provide useful insights into the acoustic differences between two English speaking cohorts, which should be taken into account when generalizing findings.