Daily prospective associations of structured and unstructured physical activity within 24-hour movement behaviour composition with next-day affect in adolescent athletes: a Bayesian multilevel compositional analysis

Daily 24-h behaviours, including sleep (SLP), sedentary behaviour (SB), light physical activity (LPA), structured moderate-to-vigorous physical activity (S-MVPA), and unstructured moderate-to-vigorous physical activity (US-MVPA), may influence adolescent athletes’ mental health through their associations with everyday affective experiences. However, these behaviours are co-dependent components of a finite 24-h day and have rarely been examined prospectively in youth athletes. Here, we investigated the daily, prospective associations between 24-h behaviour composition and next-day affect. This study used a repeated-measures prospective daily design. A total of 169 adolescent rowing and canoe/kayak athletes completed 14 consecutive days of accelerometer-based behavioural monitoring and ecological momentary affect assessment, yielding 2,366 daily observations and 2,197 lagged observations. Daily time use was classified into SLP, SB, LPA, S-MVPA, and US-MVPA. Four affective outcomes were assessed: high-arousal positive affect, low-arousal positive affect, high-arousal negative affect, and low-arousal negative affect. Bayesian multilevel compositional models were used to examine between-person and within-person associations, adjusting for individual- and day-level covariates. Pairwise compositional isotemporal substitution analyses estimated expected changes in next-day affect following 30-min reallocations between behaviours. Associations between 24-h behaviours and next-day affect were most evident at the within-person level. Relative to the remaining behaviours, more LPA was associated with higher high-arousal positive affect (posterior mean = 0.294 [95% Bayesian credible interval: 0.241, 0.346]) and low-arousal positive affect (posterior mean = 0.269 [0.218, 0.320]), as well as lower high-arousal negative affect (posterior mean = − 0.226 [− 0.260, − 0.192]) and low-arousal negative affect (posterior mean = − 0.198 [− 0.229, − 0.167]). In contrast, more SB was associated with lower high-arousal positive affect (posterior mean = − 0.092 [− 0.141, − 0.042]) and low-arousal positive affect (posterior mean = − 0.144 [− 0.194, − 0.094]), and higher high-arousal negative affect (posterior mean = 0.113 [0.081, 0.146]) and low-arousal negative affect (posterior mean = 0.076 [0.048, 0.104]). Further, 30-min reallocations from SLP, SB, and S-MVPA to LPA were associated with higher high-arousal positive affect (posterior mean = 0.034 [0.012, 0.062], 0.092 [0.045, 0.152], and 0.043 [0.017, 0.076], respectively) and low-arousal positive affect (posterior mean = 0.025 [0.004, 0.050], 0.067 [0.027, 0.117], and 0.032 [0.008, 0.062]). Reallocating 30 min from SB to LPA was also associated with lower high-arousal negative affect (posterior mean = − 0.057 [− 0.099, − 0.005]), whereas reallocating time from SB to US-MVPA showed no clear evidence of an association with affective outcomes. Among adolescent athletes, 24-h movement-behaviour compositions were prospectively associated with next-day affect. Greater relative time spent in LPA, particularly replacing SB, was associated with more favourable affective states, whereas SB to US-MVPA reallocations showed no clear association. These findings support considering the overall balance of daily movement behaviours in relation to athlete wellbeing.