PROJECT SUMMARY/ABSTRACT With the rapid aging of the U.S. population, Alzheimer’s disease and other dementias have emerged as a serious public health problem. Identifying the risk/protective factors of dementia is crucial in designing evidence-based interventions to reduce dementia risk in late life. Marital relationships have long been recognized as one of the most important social relationships that affect individuals’ well-being throughout adulthood. However, little is known about whether and how marital relationships are related to dementia risk in late life. This study is designed to enrich our understanding of how marital biography influences the risk of dementia among adults ages 65 years and older in the United States. Guided by the life course perspective, the project has three specific aims. AIM 1. Examine the associations between marital biography components and risk of dementia in late life. We will examine whether and to what extent major components of marital biography influence dementia risk over the 18-year period in the United States. We hypothesize that marital instability and longer duration in unmarried states are associated with a higher risk of dementia. AIM 2: Identify the mechanisms linking marital biography and the risk of dementia in late life. We will test four potential mechanisms that could explain the associations between components of marital biography and the risk of dementia: 1) economic resources (e.g., income, wealth, and access to health care), 2) psychosocial factors (e.g., social integration and depression), 3) biomedical (e.g., cardiovascular disease, diabetes, and stroke), and 4) health behaviors (e.g., smoking, drinking, and exercise). We hypothesize that the combination of these four factors will account for a significant proportion of the differences in dementia risk among people with different marital biographies. Aim 3: Examine sex and race differences in the associations and mechanisms discussed in Aims 1 and 2. Data will be drawn from the Health and Retirement Study (1998-2016). Discrete-time hazard models will be applied to predict dementia risk. The proposed project will provide the first in- depth and systematic analysis of how marital biography influences the risk of dementia in late life.