Speech as a digital biomarker in multiple sclerosis: Automatic analysis of speech metrics using a multi-speech-task protocol in a cross-sectional MS cohort study

Area of Science:

• Neurology
• Speech-Language Pathology
• Biomedical Engineering

Background:

• Speech abnormalities are common in people with multiple sclerosis (pwMS).
• Previous research primarily analyzed acoustic speech features, overlooking linguistic aspects.
• This study aims to analyze both acoustic and linguistic speech features and their correlation with cardinal symptoms of multiple sclerosis (MS).

Purpose of the Study:

• To investigate the relationship between acoustic and linguistic speech features and cardinal symptoms in people with multiple sclerosis (pwMS).
• To identify specific speech abnormalities associated with disease severity and neurological impairment in MS.
• To explore the influence of cognitive load on speech production in pwMS.

Main Methods:

• A prospective study involving 282 native German speakers (137 pwMS, 145 healthy controls).
• Speech recording using a comprehensive protocol including narrative, articulatory, and phonatory tasks.
• Speech analysis utilizing the SIGMA speech features library for acoustic and linguistic feature extraction.

Main Results:

• People with MS (pwMS) exhibited significant phonation and articulation instabilities (e.g., jitter, loudness variations) during cognitively demanding narrative tasks compared to healthy controls.
• Speech abnormalities, including jitter and loudness peaks, showed significant correlations with the Expanded Disability Status Scale, indicating a link to disease severity.
• Phonatory instabilities, pitch and loudness variations, and reduced speech duration were associated with MS disease severity, motor, and cognitive symptoms.

Conclusions:

• Phonatory instabilities and altered speech characteristics are associated with multiple sclerosis (MS) disease severity and neurological impairment.
• Cognitively demanding tasks exacerbate speech motor control deficits in pwMS, suggesting an interaction between cognitive and motor functions in speech production.
• Acoustic and linguistic speech analysis provides valuable biomarkers for assessing disease progression and functional impact in MS.