AI-Enhanced Automatic Life Story Structuring for Reminiscence Therapy in Older Adults: Technical Feasibility Study

Area of Science:

• Geriatric Mental Health
• Artificial Intelligence in Healthcare
• Digital Therapeutics

Background:

• Storytelling interventions show promise for older adults' well-being but face challenges with disorganized narratives and high caregiver cognitive load.
• Current manual narrative organization limits scalability and consistency in storytelling interventions.
• Technological solutions for enhancing narrative processing in digital life story structuring remain underexplored.

Purpose of the Study:

• To design an event timeline generation algorithm to optimize the Story Mosaic system for processing older adults' life narratives.
• To enable automatic extraction and organization of life story events into structured timelines, preserving clinical context.
• To reduce manual intervention costs and increase treatment efficacy via AI-driven narrative structuring.

Main Methods:

• Developed the CARE event timeline (CARE-ET) algorithm, combining temporal attention and graph-based event relationship modeling.
• Utilized multifeature extraction to identify event clues from oral histories and a hierarchical attention mechanism for event element prioritization.
• Employed adaptive compression algorithms to reduce redundancy while maintaining narrative continuity, validated through event summary, timeline quality, and usability assessments.

Main Results:

• The CARE-ET algorithm demonstrated superior performance in narrative flow and temporal accuracy compared to baseline methods.
• The AI-optimized Story Mosaic system achieved an 'A' rating for usability in evaluations conducted by 10 caregivers.
• Experimental results confirm CARE-ET's effectiveness in structuring fragmented narratives, enhancing system usability for interventions.

Conclusions:

• The CARE-ET method facilitates structured extraction of event summaries, converting disorganized life stories into event timelines for interventions.
• This AI-driven approach supports caregiver-assisted well-being interventions for older adults by simplifying narrative processing.
• The study lays a foundation for intelligent assistive technologies in geriatric mental health, with future research focusing on cognitive preservation and integration with dementia care protocols.