JAMA Netw Open. 2026 Mar 2;9(3):e261194. doi: 10.1001/jamanetworkopen.2026.1194.
ABSTRACT
IMPORTANCE: Improving mental health through targeting behaviors like sleep and physical activity in treatment has been challenging, in part due to the challenge of measuring these factors in an accurate manner. Mobile technology can enable an understanding of the dynamic, complex associations between physical activity and sleep, but most prior mobile technology studies have had modest sample sizes and utilized cross-sectional, between-person designs, thus limiting their impact.
OBJECTIVE: To investigate the within-person associations between sleep, physical activity, and daily mood among individuals receiving mental health treatment to inform health behavior recommendations.
DESIGN, SETTING, AND PARTICIPANTS: This 12-month cohort study involved patients seeking mental health care at the University of Michigan academic medical center mental health care clinics. All participants were enrolled between May 13, 2020, and December 12, 2022. Data analysis was performed from September 2024 to June 2025.
EXPOSURES: Objective, wrist-based actigraphy measures of sleep (total sleep time, interrupted nighttime sleep, and napping) and physical activity (step count) were obtained.
MAIN OUTCOMES AND MEASURES: The primary outcome was participant-reported daily mood score, on a scale of 1 (worst mood) to 10 (best mood). Linear mixed-effects models were used to estimate associations among sleep, physical activity, and mood scores.
RESULTS: A total of 1476 participants (mean [SD] age, 36.5 [14.2] years; 1062 [72.0%] female) were included in the analysis. Sleep duration was associated with subsequent mood through an inverse U-shaped relationship, with both short and long sleep duration associated with poorer mood (quadratic term b = 0.027; 95% CI, -0.031 to -0.023). Notably, patients varied substantially in their optimal sleep duration (mean [SD], 6.8 [1.9] hours) for peak mood. Physical activity was positively associated with subsequent mood (linear term b = 0.160; 95% CI, 0.149 to 0.162; P < .001; quadratic term b =- 0.022; 95% CI, -0.027 to -0.017; P < .001), with diminishing associations at higher than usual activity for that individual. Conversely, daily mood was associated with subsequent sleep (linear term b = -1.377; 95% CI, -1.877 to -0.877; P < .001; quadratic term b = -0.394; 95% CI, -0.765 to -0.023; P = .037) and step count (linear term b = 0.020; 95% CI, 0.003 to 0.030; P = .02; quadratic term b = -0.010; 95% CI, -0.020 to -0.001; P = .03) in a wavelike manner.
CONCLUSIONS AND RELEVANCE: This cohort study of people seeking mental health care found complex, bidirectional associations between sleep, physical activity, and mood with individual variation in optimal sleep duration for mood scores. The findings advance progress toward effectively targeting health behaviors to improve mental health.
PMID:41801196 | DOI:10.1001/jamanetworkopen.2026.1194
