Detecting fatigue in multiple sclerosis through automatic speech analysis.

Area of Science:

• Neurology
• Biomedical Engineering
• Speech Pathology

Background:

• Multiple sclerosis (MS) is a chronic neuroinflammatory disease impacting the central nervous system.
• Fatigue is a prevalent and debilitating symptom in MS patients, significantly affecting daily life.
• Current methods for measuring MS-related fatigue are challenging and lack objectivity.

Purpose of the Study:

• To evaluate the efficacy of automated speech analysis in detecting fatigue in individuals with MS.
• To explore the potential of specific speech patterns as objective biomarkers for MS fatigue.

Main Methods:

• MS patients underwent clinical assessments and a comprehensive speech protocol involving free speech tasks.
• A support vector machine model was trained using speech features and a cognition score.
• Speech features from picture description tasks were analyzed for their correlation with fatigue.

Main Results:

• The machine learning model achieved an Area Under the Curve (AUC) of 0.74 in detecting fatigue using speech features and a cognition score.
• Speech features from a picture description task alone yielded an AUC of 0.68, indicating the utility of specific speech patterns.
• Cognitive fatigue showed a significant association with a lower speech ratio in free speech (ρ = -0.283, p = 0.001).

Conclusions:

• Automated speech analysis, particularly from a single narrative free speech task, provides an objective, ecologically valid, and low-burden method for fatigue assessment in MS.
• Speech analysis tools show promise for clinical applications in monitoring and managing fatigue in multiple sclerosis patients.