Remote Assessment for ALS using Multimodal Dialog Agents: Data Quality, Feasibility and Task Compliance.

Area of Science:

• Neurology
• Speech-Language Pathology
• Human-Computer Interaction

Background:

• Amyotrophic Lateral Sclerosis (ALS) progressively affects motor neurons, impacting speech.
• Remote assessment methods are crucial for managing ALS, especially for individuals with mobility challenges.
• Existing remote assessment tools may lack comprehensive speech function evaluation.

Purpose of the Study:

• To evaluate the feasibility, task compliance, and audiovisual data quality of a multimodal, dialog-based system for remote ALS assessment.
• To determine if speech severity or end devices impact the quality and compliance of remote recordings.
• To establish the utility of conversational agents for monitoring speech in ALS.

Main Methods:

• 53 individuals with ALS and 52 healthy controls interacted with a cloud-based conversational agent (Tina).
• Participants completed speech tasks designed to assess motor speech function.
• Audio and video recordings were collected and rated for quality and task compliance.

Main Results:

• Excellent participant compliance (98%) and high audio (95.2%) and visual (84.8%) quality rates were achieved.
• 80.8% of recordings were both compliant and of high quality.
• Recording quality and compliance were unaffected by speech severity or end devices.

Conclusions:

• Dialog-based systems are feasible and effective for remote speech assessment in ALS.
• The system demonstrates high data quality and participant compliance, supporting its use in remote monitoring.
• This approach offers a valuable tool for tracking speech changes in ALS patients remotely.