Associations of 24-hour movement behavior composition with depressive symptoms and problematic internet use in...

Although 24-hour movement behaviors (24-HMB), including sleep, moderate-to-vigorous physical activity (MVPA), and screen time, are associated with adolescent mental health, their compositional nature remains insufficiently examined. We investigated the associations of 24-HMB composition with depressive symptoms, problematic internet use (PIU), and their co-occurrence at one-year follow-up.

In this prospective cohort study, 12,316 adolescents from the Guangzhou Childhood Adolescent Cohort (2019-2025) were included. Baseline sleep duration, screen time, depressive symptoms, and PIU were assessed using questionnaires, and MVPA was estimated from the number of days per week with at least 60 min of activity. After 1 year of follow-up, outcome-specific analytic samples included 9887 adolescents for depressive symptoms, 11,622 for PIU, and 12,316 for co-occurrence, with 2169, 1149, and 845 events, respectively. Nonlinear associations were examined using restricted cubic splines, and stratified compositional logistic regression models and exploratory isotemporal substitution analyses were then conducted using outcome-specific sleep-duration thresholds identified from the fitted spline curves.

Sleep showed nonlinear associations, with the predicted odds reaching its nadir at 8.50 h/day for depressive symptoms, 8.81 h/day for PIU, and 8.68 h/day for co-occurrence. The associations of 24-HMB composition differed across sleep-duration strata for depressive symptoms (P for interaction = 0.014), PIU (P for interaction < 0.001), and co-occurrence (P for interaction = 0.004). Among adolescents sleeping below the outcome-specific thresholds, relatively longer sleep was associated with lower odds of depressive symptoms (OR = 0.53, 95% CI: 0.37-0.76), PIU (OR = 0.33, 95% CI: 0.22-0.50), and co-occurrence (OR = 0.50, 95% CI: 0.32-0.79), particularly when replacing screen time. In the lower-sleep strata, relatively longer screen time was associated with higher odds of these outcomes, whereas theoretical reallocations from screen time to the estimated MVPA component were associated with lower odds in exploratory substitution analyses.

Our findings highlight the importance of considering 24-HMB as an integrated composition. The associations of 24-HMB reallocations with mental health outcomes differed across sleep-duration strata. Reducing screen time, with reallocation to sleep or MVPA according to sleep-duration strata, may offer tailored strategies to improve adolescent mental health.