physical activity, sleep, ambulatory assessment, multilevel model, young adults

Purpose: Observational evidence on sleep–physical activity associations is inconsistent (e.g., Atoui et al., 2021), likelyreflecting heterogeneity in samples, monitoring duration, measurement approaches, and analytic levels. Most priorstudies relied on brief monitoring periods, with few examining within-person day-to-day or mid-term dynamics. Thisstudy investigated these associations between sleep and next-day steps across four 14-day bursts over one year in alarge sample of young adults. We hypothesized that longer total sleep time (TST) and greater wake after sleep onset(WASO) would be associated with fewer steps, and higher subjective sleep quality (SSQ) with more steps.Methods: Daily step counts were modeled using three-level linear mixed-effects models (days nested within burstswithin persons). Data were collected between 2019 and 2022 from 388 Czech young adults (18–29 years; 48.2%women) who continuously wore Fitbit Charge 3/4 during four 14-day bursts across one year. Separate models wereestimated for TST, WASO, and SSQ as predictors of next-day steps. Models adjusted for age, sex, day type, season,burst number, and wear time, with random intercepts and slopes specified at both burst and person levels, and anAR(1) structure for day-to-day autocorrelation.Results: Sleep–activity associations varied by analytic level. At the within-person level, higher WASO predicted fewernext-day steps (β = –7.48, p = .027), whereas daily fluctuations in TST and SSQ were unrelated to steps. At the betweenburst level, longer TST (β = –14.82, p < .001) and greater WASO (β = –49.75, p < .001) were linked to lower step counts.Between-person associations between sleep indicators and steps were not significant. Age, season, burst number,and wear time consistently influenced step counts.Conclusion: These findings highlight the importance of extended monitoring to capture both day-to-day andmid-term within-person variability in sleep–activity relationships among young adults. Daily fluctuations in WASO,but not TST or SSQ, were associated with next-day activity, indicating that short-term sleep disruptions may coincidewith lower next-day activity. In contrast, between-burst associations for TST and WASO suggest that sleep–activityassociations emerge primarily across broader (mid-term) timescales, even after accounting for season, reflectinggradual behavioral shifts rather than short-term fluctuations.

Purpose: Observational evidence on sleep–physical activity associations is inconsistent (e.g., Atoui et al., 2021), likelyreflecting heterogeneity in samples, monitoring duration, measurement approaches, and analytic levels. Most priorstudies relied on brief monitoring periods, with few examining within-person day-to-day or mid-term dynamics. Thisstudy investigated these associations between sleep and next-day steps across four 14-day bursts over one year in alarge sample of young adults. We hypothesized that longer total sleep time (TST) and greater wake after sleep onset(WASO) would be associated with fewer steps, and higher subjective sleep quality (SSQ) with more steps.Methods: Daily step counts were modeled using three-level linear mixed-effects models (days nested within burstswithin persons). Data were collected between 2019 and 2022 from 388 Czech young adults (18–29 years; 48.2%women) who continuously wore Fitbit Charge 3/4 during four 14-day bursts across one year. Separate models wereestimated for TST, WASO, and SSQ as predictors of next-day steps. Models adjusted for age, sex, day type, season,burst number, and wear time, with random intercepts and slopes specified at both burst and person levels, and anAR(1) structure for day-to-day autocorrelation.Results: Sleep–activity associations varied by analytic level. At the within-person level, higher WASO predicted fewernext-day steps (β = –7.48, p = .027), whereas daily fluctuations in TST and SSQ were unrelated to steps. At the betweenburst level, longer TST (β = –14.82, p < .001) and greater WASO (β = –49.75, p < .001) were linked to lower step counts.Between-person associations between sleep indicators and steps were not significant. Age, season, burst number,and wear time consistently influenced step counts.Conclusion: These findings highlight the importance of extended monitoring to capture both day-to-day andmid-term within-person variability in sleep–activity relationships among young adults. Daily fluctuations in WASO,but not TST or SSQ, were associated with next-day activity, indicating that short-term sleep disruptions may coincidewith lower next-day activity. In contrast, between-burst associations for TST and WASO suggest that sleep–activityassociations emerge primarily across broader (mid-term) timescales, even after accounting for season, reflectinggradual behavioral shifts rather than short-term fluctuations.