Associations of Wearable-Measured Sleep and Physical Activity With Memory Performance in Older Adults: Cross-Sectional Study With Actigraphy and MRI

Abstract

Background: Cognitive decline is a common aspect of aging, and identifying modifiable lifestyle factors, such as physical activity and sleep, is crucial for promoting healthy brain aging. While both are individually linked to cognition, few studies have simultaneously assessed their independent and combined effects using objective wearable-based data, particularly in older Asian populations. Objective: This study aimed to examine the independent and interactive effects of wearable-assessed sleep and physical activity parameters on memory performance in healthy older adults. We also explored whether age and hippocampal volume moderated these associations. Methods: This prospective cross-sectional analysis included 88 cognitively healthy community-dwelling adults (≥60 years of age) from the Integrating Systematic Data of Geriatric Medicine to Explore the Solution for Healthy Aging cohort in Taiwan. Participants underwent 12-day wrist-worn actigraphy, brain magnetic resonance imaging, and neuropsychological assessments. Light-intensity physical activity (LPA) and wake after sleep onset (WASO) were selected based on age-adjusted partial correlations with Consortium to Establish a Registry for Alzheimer’s Disease Neuropsychological Battery memory scores. Multivariate regressions, age-stratified models (cutoff=72 years), and PROCESS moderation and mediation analyses were conducted, adjusting for age, education, daytime sleepiness, and hippocampal volume. Results: Partial correlation analyses adjusting for age showed that higher LPA (r=0.260; P=.02) and lower WASO (r=–0.251; P=.02) were significantly associated with better memory scores. Age significantly moderated both effects: LPA was beneficial beyond 73.8 years of age, and WASO was detrimental beyond 71.1 years of age. Multivariate regression models confirmed that both WASO (β=–.044; P=.04) and LPA (β=.042; P=.01) were significant predictors of memory. In subgroup analyses (age ≥72 years), both LPA (β=.054; P=.04) and WASO (β=–.111; P=.01) remained significant predictors. Moderated mediation analyses showed that WASO was associated with reduced LPA (β=–.325; P=.03), but the indirect effect on memory via LPA was not significant. Instead, WASO exerted a direct and age-moderated effect on memory performance. Hippocampal volume moderated both associations, supporting the brain reserve hypothesis. Conclusions: Our findings highlight WASO and LPA, as measured by wearable devices, as modifiable behavioral factors linked to memory function in older adults. The impact of these factors intensifies with advancing age and may be influenced by hippocampal reserve. Promoting daily light physical activity and maintaining sleep continuity may serve as accessible, age-tailored strategies for preserving cognitive health in aging populations.