Methodological Considerations for the Study of Cognitive Reserve and its Relationship to Cognitive Decline in Older Adults

Cognitive reserve (CR) is a key construct in promoting healthy cognitive aging. Methodological inconsistencies have limited the ability for researchers to make broad conclusions about relevant life experiences that may build CR and delay cognitive decline. This dissertation explored notable issues within the operationalization of CR based on proxy life experiences by applying a systematic approach toward understanding the use of CR, as well as an understanding of consequences of inconsistent operationalization. The project had three aims: review and explore the association between various CR proxies utilized in the literature and develop a composite score for CR, describe differences in the association between CR and cognitive decline under varying operationalization for CR, and examine the relationship between different measures of CR and 1) clinical diagnosis of mild cognitive impairment (MCI) and Alzheimer’s disease (AD), and 2) survival following diagnosis. We conducted a two-phase search of PsycInfo, PubMed, and Web of Science which resulted in 20 proxy categories that have been used by the published literature. A correlation analysis and a confirmatory factor analysis showed that few of the CR proxies were strongly associated with one another, leading to the hypothesis that the inconsistency in results in the literature stems from varying operationalization within proxies used to predict cognitive decline. We used generalized linear mixed effects models to examine the relationship between different CR proxies and cognitive decline in healthy adults, and Cox proportional hazards models to examine CR operationalization and time to MCI diagnosis, time to AD diagnosis, and survival following diagnosis. Both analyses showed that the relationship between CR and decline is sensitive to CR operationalization. Given the inconsistency in results based on CR operationalization, the dissertation provides avenues for future research to enhance operationalization of CR. As the most consistent proxies, education, leisure activities, cognitive activities, intelligence, socioeconomic status, social activities in late life, and perceived social isolation should be considered in future research. That said, there may be an inability in CR research to consistently rely on the same individual proxy. In an effort to address the relative limit of individual proxies, this dissertation also developed a cumulative lifecourse-based score through a confirmatory factor analysis to put forth an alternative operationalization. Both this score and a lifecourse score based on the Lifetime of Experiences Questionnaire were relatively consistent predictors of protection against decline. These findings imply the need for a lifecourse approach combining multiple proxies in CR operationalization. Collectively, attention to qualities of predictors, adherence to a lifecourse approach, and collaborative movement towards a consensus operationalization of CR are imperative shifts in the study of CR that are highlighted by this dissertation