Aging is associated with declines in executive function leading to difficulties performing everyday tasks. To combat aging declines, many studies investigate varying rehabilitation interventions. Although improvement on task performance has been noted, improvement on other tasks remains inconsistent. The current study sought to replicate Miyake and Friedman’s (2012) Unity and Diversity (U&D) model including physical activity as a predictor across a sample of older adults. Two models were computed to estimate three latent variables: Common EF, Updating, and Shifting. Model 1 replicated Miyake and Friedman’s bifactor model. Model 2 was a structural model that included physical activity as a separate independent variable. Extracted factor scores were used to predict within network connectivity for several resting state networks. We didn’t find evidence for a Shifting factor or a relationship between EF factors, physical activity, and within-network connectivity. Future analyses will continue to examine this model and modified versions in older adults.