Abstract
This study investigates physical activity within dyadic relationships as a predictor of cognitive function using data from the Survey of Health, Aging, and Retirement in Europe (SHARE). A two-wave dyadic analysis was conducted for 33 opposite-sex couples (mean age 67.7 years). Physical activity was measured with a three-axial accelerometer, and cognitive function through verbal fluency at baseline and follow-up. Multilevel lagged dependent variable models were estimated using mixed effects in Stata Version 18/SE. Verbal fluency remained stable over time (β = .62, p < .001). Individuals' own physical activity (steps) predicted better cognitive function at follow-up (β = .30, p < .001). Spouses' physical activity also showed a positive association, though weaker (β = .25, p < .01). Individual activity was more strongly associated with cognitive outcomes than partner activity. However, partner effects remained significant after accounting for individuals' own activity, indicating that dyadic influences contribute uniquely beyond personal behavior. These findings underscore the relevance of dyadic processes for cognitive health in later life. Future research should examine underlying mechanisms and evaluate couple-based interventions.