Nicklas Linz, Simona Schäfer, Louisa Schwed, Felix Doerr, & Johannes Tröger
* Poster presented at the CNS Summit 2024, Boston (USA)
Abstract
Objective: Speech problems often accompany motor disorders like Parkinson’s Disease (PD) and Huntington’s Disease (HD) significantly affecting patients’ daily functioning. Objective measurements of speech intelligibility could serve as new outcome measures in clinical trials. This study uses the ki speech biomarker for motor speech symptoms (SB-M) to quantify intelligibility in PD and HD patients.
Design: Speech data was collected from 93 PD patients, 40 HD patients, and 100 controls, speaking three different languages. The SB-M intelligibility measure was assessed by comparing the word accuracy rate of automatically recognized text to the presented text.
Results: SB-M intelligibility was significantly lower in patients than controls: 1) Czech HD data: H = 37.432, p < .01, Cohen’s d = 1.872, 2) Czech PD data: H = 6.221, p < .05, Cohen’s d = 0.675, 3) Colombian PD data: H = 16.266, p < .01, Cohen’s d = 0.859. Intelligibility scores correlated significantly with dysarthria items of the UPDRS in PD patients and UHDRS in HD patients: 1) Czech HD data: r = -.37, p < .05, 2) Czech PD data: r = -.328, p < .05, 3) Colombian PD data: r = -.403, p < .01.
Conclusion: SB-M measured intelligibility differed significantly between controls and patients with HD or PD in all datasets. Scores also correlated with dysarthria items in the UPDRS and UHDRS. These findings suggest that automatic intelligibility measures can be effective across diseases and languages. Future research could investigate combining additional objective measures like Vowel Space Area to enhance detection of impaired intelligibility.
