Dementia and neurodegenerative diseases including Alzheimer’s dementia (AD), which increase in prevalence with age, greatly impact the health and quality of life of older adults, as well as their families and support systems. At a population level, as the “Baby Boomer” generation enters older age, both the number and the proportion of older adults will continue to grow and along with it, the number of people with dementia and Alzheimer’s disease. Some estimates project that by 2050, the prevalence of AD could grow three fold from levels in 2000 to as much as 13 million affected individuals. Identification of both preventable risk factors for dementia and Alzheimer’s disease and for early indicators of decline is critical in intervention planning and in preparing healthcare and care infrastructures. Both level of cognition and rate of cognitive decline in mid and late life are predictive of the development of dementia. Prior research has identified behavioral, environmental, and genetic, namely the APOE e4 allele, risk factors for cognitive function and decline and described pathophysiologic processes. Early indicators and many risk factors for cognitive decline are measured and represent temporally concurrent circumstances, but researchers have advocated for a life-course approach to studying cognitive aging. In this dissertation I used data from three cohorts to look at predictors of cognitive level, rate of decline, and of neuropathology associated with cognitive decline. Specifically, I aimed to 1) evaluate the association of actigraph and self-reported sleep with cognitive function and 5-year cognitive decline, 2) evaluate whether the genetic penetrance of the APOE e4 risk allele varies by early life environment operationalized as birth-year cohort in one study and Adverse Childhood Experiences (ACEs) in another. In the first aim, I found that in cross-sectional models, actigraph sleep disruption measures were associated with worse cognition. Actigraph sleep disruption measures were also associated with odds of 5-year cognitive decline (4 or more points), with wake after sleep onset having the strongest association. Longitudinal associations were generally stronger for men than women. Self-reported sleep showed little association with cognitive function. In the second aim, I found that decline due to age in a postwar birth-year cohort was very similar regardless of e4 status group in contrast to a prewar birth-year cohort, where those with an e4 allele had a much greater rate of decline over age compared to those without an e4 allele. In a complementary analysis in the third aim, I did not find that ACEs modified the association of the APOE e4 allele with cognitive function or decline nor that ACEs modified the relationship of the allele with primary AD pathology. However, I did observe that ACEs were independently associated with odds of gross chronic infarcts driven by those in the that had experienced emotional neglect having a higher odds of showing evidence of infarct, compared to those with few or no ACEs. In an interaction model, variation in penetrance of the allele by ACE was marginally significant.