This research introduces a real-time conversational speech system for robots that detects cognitive impairment markers in people living with dementia.

• Integrates a WebSocket server with a fine-tuned LLM specifically trained on dementia patterns
• Processes speech and generates responses in under 1.5 seconds
• Analyzes speech biomarkers to identify potential cognitive decline
• Designed as a Progressive Web App for accessibility

This innovation has significant medical implications by enabling early detection of cognitive changes, potentially improving diagnosis timelines and care management for dementia patients. The system could serve as a scalable, non-invasive screening tool in various healthcare settings.

Developing Conversational Speech Systems for Robots to Detect Speech Biomarkers of Cognition in People Living with Dementia