Tabea Thies, Felix Dörr, Louisa Schwed, Johannes Tröger, Tom Hänhel, Florin Gandor
*Presented at the International MSA Congress, May 2025
Abstract
Objective: Parkinson’s disease (PD) and multiple system atrophy (MSA) are neurodegenerative disorders that share overlapping motor and non-motor symptoms, often complicating early and accurate diagnosis. Speech characteristics offer a promising avenue for differentiation through objective, non-invasive means. This study evaluates an automatic speech analysis method to distinguish between PD and MSA.
Methods: Speech Assessment: Quiet environment in the clinic, Olympus LS-P4 Audio Recorder, Reading text (two repetitions); Motor Assessment: Part III of the MDS Unified Parkinson’s Disease Rating Scale (MDS-UPDRS III)
Data Processing: By using ki:elements signal processing pipeline SIGMA, acoustic features were extracted and the ki: SB-M intelligibility score was calculated from the recordings by comparing generated scripts to the corresponding reference text.
Results: Kruskal-Wallis test revealed significant differences in 16 of 70 features across groups. See poster for table. Individuals with MSA present with reduced speech loudness, slower speech tempo, breathier voice quality and overall reduced phonatory control. Intelligibility was reduced in the MSA group and declines with increasing motor impairment.
Conclusion: Speech characteristics differed between MSA and PD. In MSA, reduced intelligibility was linked to greater motor impairment. These results suggest that automatic speech analysis might be able to distinguish between the two disorders.
