Background Studies exploring device-measured physical activity (PA) and cognition are lacking outside of older age, prior to prodromic cognitive decline. Further, existing studies in midlife do not account for the interdependence between the full spectrum of movement behaviours and primarily rely on self-reported PA measures. Limited evidence points towards moderate and vigorous PA (MVPA) as being important for cognition, with little evidence on light intensity PA (LIPA), sedentary behaviour (SB) and sleep. We aimed to investigate these associations in midlife using 24-hour movement data.

Methods The 1970 British Cohort Study is a prospective birth cohort study of UK-based participants, born within one week of each other. Of 8,581 participants at age-46 follow-up, 6,597 participants consented to wear a thigh-worn accelerometer device for one week to assess habitual patterns of MVPA, SB, LIPA and sleep. Cognitive assessments were concurrently taken including tests of executive function (verbal fluency and 2-letter cancellation), and memory (immediate and delayed recall). We derived a composite z-score of overall cognition. Compositional linear regression was used to assess associations between 24-hour movement patterns and cognition with adjustments for sex, marital status, education, socioeconomic class, body mass index, disability status and depressive symptoms. We modelled how different daily movement compositions are associated with cognition.

Results An analytic sample (N=4,046; 50% female) was derived with complete accelerometer, cognitive and covariate data. Greater time spent in MVPA, relative to other behaviours, was strongly associated with higher cognition when adjusting for sociodemographic factors but attenuated when adjusting for health status. When modelling reallocation of time (e.g. replacing time in one behaviour with another), replacing 15 minutes of MVPA time with 15 minutes more SB was associated with a -1.57% (95% CI: -3.10, -0.03) decrease in cognition score. Similarly, replacing 11 minutes of MVPA with LIPA was associated with a -1.47% (95% CI: -2.87, -0.05) reduction in cognition score. Lastly, replacing 12 minutes of MVPA with sleep was associated with a -1.48% (95% CI: -2.91, -0.06) reduction. There was little effect of reallocating time between SB, LIPA, and sleep.

Discussion Applying a compositional analysis approach, we found MVPA to be most critical for cognition in this mid-life sample of adults. With the simultaneous rise in sedentary behaviour in leisure and occupational settings, and increasing prevalence of cognitive decline in later life, reinforcing MVPA in place of other daily movements may have benefit. This warrants further investigation by applying similar methodologies to longitudinal data.