Abstract
STUDY OBJECTIVES: This study examines age-related variations in activity patterns using actigraphy data from the National Health and Nutrition Examination Survey (NHANES). By analyzing sleep onset, wake times, and daily activity levels across different age groups, we aim to uncover key changes in chronotype and physical engagement with aging. From a systems-biology perspective, minute-level rest-activity traces are emergent outputs of coupled circadian-homeostatic-behavioral networks. Treating actigraphy as a high-throughput phenotyping readout, we use NHANES to extract system-level markers (phase, amplitude, and transition dynamics) that reflect network organization across the lifespan. METHODS: Actigraphy data from NHANES (2011-2013) were analyzed using machine learning techniques to identify distinct activity clusters among four age groups (19-30, 31-50, 51-70, 71-80). We implemented an unsupervised machine learning pipeline that clustered average-day actigraphy profiles, enabling the identification of distinct, age-dependent rest-activity phenotypes from the NHANES dataset. Sleep-wake cycles, activity intensities, and circadian periodicities were assessed through clustering and statistical modeling. Key metrics, including winding down activity and time to alertness, were derived to evaluate age-related variations. RESULTS: Younger individuals exhibited delayed chronotypes with later sleep and wake times, whereas older adults showed advanced and more structured schedules. Winding down periods lengthened with age, and overall activity levels declined progressively. Time to alertness showed a strong correlation with wake time in younger groups but diminished with age, indicating a weakening circadian influence. CONCLUSION: Aging is associated with shifts in sleep-wake cycles and activity patterns, reflecting biological and behavioral adaptations. These findings highlight the importance of personalized interventions to support optimal activity and sleep alignment across the lifespan. Insights from actigraphy data can inform public health strategies and clinical approaches to aging-related changes in physical activity and circadian regulation. These age-stratified, interpretable "dynamical phenotypes" provide observables to calibrate and validate systems-level models of sleep-wake regulation and behavior-physiology coupling, supporting hypothesis generation and intervention design in systems biology.