LifeBio Memory: Development of a reminiscence therapy online platform with machine learning to increase engagement

Short Summary Statement This resubmission proposal (originally scored 36) aims to investigate the effectiveness of LifeBio Memory (LBM) technology, enhanced with Artificial Intelligence (AI) capabilities, in improving health-related quality of life, depression, relationship closeness, and loneliness among older adults with mild cognitive impairment (MCI) or at risk of MCI and their informal caregivers. The LBM technology has been shown to increase engagement with reminiscence therapy, but its efficacy across diverse linguistic and cultural backgrounds remains untested. To address this gap, we will extend our investigation to include participants from various racial and ethnic groups using a mixed-methods approach. A significant innovation of this study is the integration of AI-powered voice technology, enabling the LBM system to learn and replicate an individual's unique vocal characteristics, creating a deeply personal and immersive experience. The study consists of three aims: 1. Develop and refine LBM's multi-language support, including Spanish, Korean, and English for African-American populations, with a focus on incorporating AI-driven voice clone capabilities and culturally relevant presentation. 2. Conduct usability testing with 48 participants to assess the learnability, efficiency, error recovery, and subjective satisfaction of the LBM app and AI voice clone feature. 3. Test the feasibility, acceptability, and initial efficacy of the LBM intervention using a randomized waitlist-controlled trial among 84 diverse racial and ethnic dyads. We will evaluate outcomes at baseline, 6 weeks post-follow-up, and 10 weeks post-follow-up, assessing health-related quality of life, depression, relationship closeness, and loneliness using standardized measures. We will examine the impact of the AI-powered voice clone feature on user engagement, emotional response, and overall satisfaction with the LBM experience. This study has significant scientific merit, addressing a critical gap in literature by investigating the effectiveness of LBM among diverse populations and exploring the potential benefits of AI- driven voice clone technology. The results will inform the development of culturally sensitive interventions to support older adults with MCI or at risk of MCI and their caregivers, potentially leading to innovative, personalized therapies that improve health outcomes and reduce healthcare disparities. If successful, this intervention could offer invaluable non-pharmacological support, warranting widespread implementation through home and community-based services.