Automatically measured speech intelligibility models bulbar-specific disease severity and progression in Amyotrophic Lateral Sclerosis.

Amyotrophic lateral sclerosis (ALS) is a progressive neurodegenerative disease that leads to widespread motor deterioration, including significant motor speech impairments. Speech intelligibility is a crucial component of communication affected in ALS, requiring objective, scalable assessment methods as an indicator of disease progression and treatment efficacy. Objective: This study investigates whether speech and bulbar function in ALS could be evaluated and monitored utilizing an automated digital measure of speech intelligibility derived from naturalistic picture descriptions. Methods: Speech recordings from 44 patients living with ALS (plwALS) and 49 matched healthy controls (HC) were analyzed and processed utilizing an automated speech analysis pipeline to extract an intelligibility score. These were part of a cross-sectional and longitudinal study involving two assessments. Results: The findings confirmed that speech intelligibility is significantly reduced in plwALS compared to HC. Those with bulbar-onset ALS have lower intelligibility than those with spinal-onset ALS, and the intelligibility of individuals with bulbar symptoms-regardless of the onset type-is lower than in plwALS without bulbar symptoms. Declining ALS-related speech scores correspond with worsening intelligibility in longitudinal assessments. Intelligibility correlates strongly with bulbar-specific clinical measures but not with global scores, highlighting its role in tracking bulbar progression. In some plwALS, we were able to demonstrate that automated speech analyses are more effective in detecting worsening in intelligibility earlier than standard clinical scoring. Conclusion: Our findings highlight that automated speech intelligibility assessments can be a valuable marker to improve clinical monitoring and facilitate earlier intervention in ALS as a supplement to standard assessments.