Longitudinal associations between 24-hour movement behaviours and cognitive function in adults aged 55 and above

Background Age-related cognitive decline poses challenges to healthy ageing. Physical activity (PA), sedentary behaviour (SB) and sleep have been linked to cognitive health, yet much evidence is cross-sectional and fails to account for the interdependent nature of these 24-hour movement behaviours. This observational study applied a compositional approach to investigate longitudinal associations between 24-hour movement behaviours and cognition in healthy adults aged 55 and above. Methods Community-dwelling adults aged ≥ 55 years were assessed at three time points, each one year apart (baseline n = 233; 51.1% women; mean age 68.3 ± 7.7 years). Each time, 7-day wrist-worn accelerometery data (ActiGraph wGT3X-BT) were acquired and processed with GGIR to derive time spent in PA, SB and sleep. Additionally, cognition was measured using the Cambridge Neuropsychological Test Automated Battery (CANTAB), yielding composite z‑scores for executive function (EF), short‑term memory (STM), long‑term memory (LTM) and processing speed (PS). Linear mixed-effects models tested longitudinal associations between time-use composition (i.e. 24-hour movements behaviours expressed as isometric log-ratios) and cognition, including an interaction term between time-use composition and time, and were adjusted for sociodemographic and health covariates. Post‑hoc compositional isotemporal substitution estimated cognitive differences associated with time reallocations between behaviours. Results No significant associations were observed between time-use composition and STM, LTM or PS. After adjusting for age, sex, educational level and social isolation, EF was significantly associated with time-use composition ( p  = 0.005) and time ( p  < 0.001), with no significant interaction effect, suggesting a stable relationship over time. Post-hoc analyses indicated that moderate-to-vigorous PA (MVPA) was the primary driver of the EF association. While reallocating time from light PA (LPA) to SB showed some positive EF differences, reallocations from LPA or sleep to MVPA were associated with better EF z-scores. Conversely, reallocating time away from MVPA related to lower EF, underscoring its central role. Conclusions In adults aged ≥ 55, 24‑hour movement behaviours were associated with EF but not memory or PS, and these associations remained stable over time. Maintaining or increasing time in MVPA may support cognitive health, highlighting the need for intervention studies to confirm these findings. Future research should also examine the (cognitive) context of behaviours.