A systematic review and meta-analysis of the relationship between midlife physical activity and later-life cognition: Is sex an overlooked factor?

• Cohort review and meta -analysis across 15 studies, n = 33,295. • Higher PA associated with higher global cognition, memory, and processing speed. • No association between PA and executive functions or verbal fluency/naming. • Sex-specific effects remain unclear due to limited stratified data. Midlife (ages 40–59) may be a critical window to preserve cognitive health; however, the long-term impact of midlife physical activity (PA) on late-life cognition remains unclear. This systematic review synthesized cohort evidence of the relationship between midlife physical activity levels and later-life cognition, while exploring the potential moderating role of biological sex. The Medline, Embase, Web of Science, APA PsycINFO, CINAHL, and SPORTDiscus databases were searched. Cohort studies that assessed midlife physical activity and used standardized neuropsychological tests in cognitively unimpaired adults were included. Risk of bias was completed using the Risk of Bias in Non-randomized Studies–of Exposure tool. Random-effects meta -analyses (Cohen’s d ) compared none/low vs. moderate/high PA for global cognition, verbal episodic memory, executive functions, verbal fluency, and processing speed domains. Fifteen studies (n = 33,295; 66.1% female) were included in the qualitative synthesis, with eight used in the meta -analyses. All studies relied on self-reported PA measures, and only one study completed sex-stratified analyses. Higher midlife PA was associated with better later-life global cognition ( d = 0.22, 95% CI: 0.06–0.38), verbal episodic memory ( d = 0.19, 95% CI: 0.06–0.32), and processing speed ( d = 0.20, 95% CI: 0.03–0.36). Heterogeneity was substantial (I 2 72.2–84.7%). Pooled effects were not significant for executive functions ( d = 0.09, −0.16–0.34) or verbal fluency ( d = − 0.03, 95% CI: −0.21–0.16). Midlife PA showed modest, albeit population-meaningful associations with higher later-life global cognition, episodic memory, and processing speed. Future cohorts should standardize and incorporate objective PA metrics and ensure inclusion of sex-stratified estimates to clarify dose–response and sex-specific effects.