Associations between multidimensional sleep measures and domain-specific cognitive function across repeated...

Sleep is increasingly recognised as a potentially modifiable factor in cognitive ageing, but evidence is often limited by cross-sectional designs, single sleep measures and global cognitive outcomes. This study examined whether multiple sleep characteristics (device-based and self-report) were associated with domain-specific cognitive function across repeated observations in community-dwelling adults aged 55 years and older. Adults aged ≥ 55 years were assessed at three time points over 2 years (baseline n = 233; 51.1% women; mean age 68.3 ± 7.7 years). Sleep measures included device-measured sleep duration and sleep efficiency from 7-day wrist accelerometery, subjective sleep quality assessed using the Pittsburgh Sleep Quality Index and a multidimensional sleep health score combining device-measured and self-reported sleep dimensions. Cognitive outcomes were Cambridge Neuropsychological Test Automated Battery composite z-scores for short-term memory, long-term memory, executive function and processing speed. Linear mixed-effects models examined associations between time-varying sleep characteristics and cognitive outcomes and tested sleep-by-time interactions. Models were adjusted for age, sex, educational level, living situation and total physical activity. Sleep duration, subjective sleep quality and multidimensional sleep health were not significantly associated with any cognitive domain. Higher sleep efficiency was significantly associated with better executive function; sensitivity analyses using wake after sleep onset showed a comparable pattern. Sleep-by-time interactions were not significant, indicating that associations remained stable over the follow-up period. In adults aged ≥ 55 years, sleep characteristics were largely not associated with domain-specific cognitive function, except for a positive association between sleep efficiency and executive function. Associations did not differ across time points.