PROJECT SUMMARY Throughout life, we routinely make decisions that impact our physical and financial health (e.g., which life insurance plan to choose or whether to get a medical treatment). Our ability to make appropriate future decisions depend, at least in part, on accurate memory for past choices, including memory for the context in which those decisions were made. Older adults are impaired at these flexible, context-specific decisions. To what extent is this related to their memory? Memory precision declines with age; this is true both of memory for specific details of individual items (item memory) as well as memory for which items occur when and where (context memory), and also how individual items and concepts are related to one another (statistical learning). These declines are significantly more pronounced in individuals diagnosed with Alzheimer’s disease and related dementias. This R01 proposal from two New Investigators aims to understand how memory failures lead to decision failures. We will examine this question using a synergistic computational and neurobiological approach. Specifically, we will measure the influence of different kinds of memories – item, context, statistical learning - on choices, using novel computational frameworks we have developed, and variants of well-validated tasks tuned to test these formal hypotheses. We will relate this framework to measures in behavior, functional neuroimaging, and advanced diffusion imaging methods which we have also recently developed. Although numerous studies have attributed choice performance to the striatum, fewer have assessed the specific contributions of the medial temporal memory circuit. To address these limitations, cognitively normal older adults will complete our decision tasks while undergoing functional and diffusion magnetic resonance imaging (MRI) to assess contributions of the striatal and medial temporal circuits. Extending our recent work, we will test for unique contributions of recent