Designing a Dementia Caregiver-Centered Recommendation and Context-Aware AI Media Analytical Application: A Comprehensive System Requirements Analysis

Abstract

During technology-assisted reminiscence therapy sessions, caregivers need to quickly assess and select suitable multimedia content to elicit positive memories and emotions in the participating person or persons living with dementia. Especially when relying on online content, caregivers encounter difficulties in identifying dementia-friendly media items due to the large number of search results and the uncertainty about dementia-related selection criteria (e.g., visual and thematic composition). This study seeks to develop an AI-based recommendation and media analysis application for reminiscence therapy by presenting participatory co-design results and outlines caregiver-driven application requirements. Collaborating with n = 14 caregivers from three care facilities, we defined user needs based on our initial definition of the minimum viable components that emerged from previous studies and from the research undertaken as part of this study, including profiling, media search, media storage, media rating, AI-based media analysis, and a recommendation system. Our main findings result in system requirements that correspond to the practical implementation of dementia friendly, digital media-based reminiscence sessions in daily practice. These findings also aid future research and the design of similar applications, benefiting researchers and developers.