Decomposing the within-person and between-person sources of variation in physical activity-cognition associations for low-active older adults

Objectives: To examine within-person and between-person sources of variation in the relationship between physical activity and cognition in older adults participating in a walking program. To explore whether demographic, health and fitness variables, and their interactions with activity, are significant predictors of cognition. Design: Brief longitudinal burst design. Participants: 118 participants (91 females, mean age = 72.81 + 5.24 years). Main Outcome Measures: Cognition, self-reported moderate-to-vigorous walking and self-reported moderate-to-vigorous physical activity were assessed at baseline and 6, 9, 12 and 16 weeks follow-up. Attendance at weekly walks was also recorded. Results: Within-persons, changes in physical activity were related to select measures of executive functioning, with increased activity predictive of better cognition (three of four cognitive functions, p<.05). Between-persons, activity was also associated with cognition (two of four cognitive functions, p<.05). Younger age and higher education were related to better cognition. Interactions of demographic, health, and fitness variables with changes in within-person activity were generally non-significant. Conclusion: The results highlight the importance of distinguishing within- from between-person effects in longitudinal analyses of the association between physical activity and cognition. This stringent within-person test of association underscores the potential value of simple physical activity interventions for improving cognitive function.