S. Goulding, S. Raeesi, J. A. E. Anderson, M. D. A. E. Oliver, L. L. Barnes, M. Estrella, M. Dadar, C. Morrison
INTRODUCTION: Cognitively stimulating activity (CSA) is associated with cognitive health, but whether associations with cognitive change differ by life stage and sex remains unclear. METHODS: Data were drawn from Rush cohorts. Participants were 2,614 adults aged 65 years and older (1,965 females; 649 males). CSA during childhood, middle adulthood, and late life was classified using stage-specific pooled medians. Linear mixed-effects models examined global cognition and five domains, adjusting for age, education, and cognitive diagnosis. Sex-engagement-time interactions were tested with false discovery rate correction. RESULTS: Childhood engagement showed limited associations; only the working-memory sex-engagement-time interaction survived correction ({beta} = -0.014, qFDR=.047). Middle adulthood showed the strongest sex-dependent associations, with significant interactions for global cognition, episodic memory, semantic memory, processing speed, and visuospatial ability (qFDR [≤].017). Higher engagement was associated with more favorable global and visuospatial trajectories among females but more negative trajectories across several domains among males. In late life, only working memory showed a significant sex interaction ({beta} = 0.017, qFDR=.018). High engagement in childhood and middle adulthood followed by low late-life engagement was associated with more favorable global cognitive trajectories among females but less favorable trajectories among males compared with both consistently low and consistently high engagement (sex-pattern-time qFDR<.001). DISCUSSION: Associations between CSA and cognitive change varied by life stage and sex and were not uniformly favorable. The observational design limits causal interpretation.