Graduation Year

2025

Document Type

Dissertation

Degree

Ph.D.

Degree Name

Doctor of Philosophy (Ph.D.)

Degree Granting Department

School of Aging Studies

Major Professor

Debra Dobbs, Ph.D.

Co-Major Professor

Brent Small, Ph.D.

Committee Member

Jennifer Lister, Ph.D.

Committee Member

Ross Andel, Ph.D.

Keywords

Cognitive aging, Dementia, Longitudinal methods, Machine Learning, SuperAging

Abstract

Some level of cognitive decline is generally expected with advanced age. However, a subgroup of individuals – commonly labeled as ‘SuperAgers’ – maintain youthful memory abilities into their 80s and beyond. Relatively little is known about factors across the life course related to this phenotype. This dissertation used population-based data from the Health and Retirement Study (HRS) to investigate lifestyle factors related to the most youthful cognitive trajectory in the United States and assess how their cognition fares over time. Three research questions were addressed: (1) Can samples of youthful cognitive agers be identified longitudinally in a population-based dataset? (2) What lifestyle factors are associated with belonging to the youthful cognitive aging group? and (3) How stable are youthful cognitive agers over time and which lifestyle factors predict changes in performance?

The first research question employed growth mixture modeling to examine the cognitive aging trajectories of 8,054 older adults in the HRS over 12 years. Five distinct trajectory groups emerged with one standing out as the most optimal (N=623, 8.4%). They displayed consistently high total cognition scores, therefore being deemed ‘youthful’ cognitive agers. Additionally, memory trajectories were modeled over time to determine the overlap between youthful cognitive agers and youthful memory agers. Most individuals in the youthful cognition group were also in the youthful memory group.

The second research question used machine learning to select lifestyle correlates related to youthful cognitive aging. Machine learning identified 15 significant predictors including baseline age, gender, race, education, urbanicity, total wealth, chronic conditions, self-rated health, IADL limitations, childhood socioeconomic status (SES), depressive symptoms, vigorous physical activity, drinking days per week, self-rated hearing, and volunteering. Correlates were subsequently used in logistic regression models to compare youthful cognitive agers to the other trajectory groups and youthful cognitive agers to the lowest performing trajectory group. Regarding the first logistic regression model, various demographic (i.e., older age, female gender, White race), childhood (i.e., more education, higher SES in childhood), and late life factors (e.g., urbanity, higher total wealth, fewer depressive symptoms, better self-rated health, reporting volunteering) were related to higher odds of youthful cognitive aging. As for the second logistic regression model, there were some variable overlaps but notable differences. For example, more drinking was related to higher odds of youthful cognitive aging, though this finding should be interpreted with caution.

Finally, the third research question examined the stability of youthful cognitive agers over 10 years. Results indicate that even though youthful cognitive agers were the most optimal trajectory group, they were not immune to subsequent decline. Of the lifestyle variables included, only self-rated health was related to the rate of decline. Those with better self-rated health exhibited fewer declines. Findings from the second and third research questions suggest that the lifestyle factors associated with belonging to the youthful cognitive aging group differ from those that influence cognitive decline over time.

Taken together, this dissertation underscores the substantial variability in cognitive aging, as evidenced by the five distinct trajectory groups identified in our analysis. The results highlight the importance of incorporating lifestyle factors when studying preserved cognition in late life. Furthermore, using a life course approach is especially valuable, as both early life and late life factors were related to youthful cognitive aging. Additionally, self-rated health may serve as a key indicator of cognitive decline among those with preserved cognitive abilities. Future research should explore the interplay between lifestyle variables and neurological factors to uncover mechanisms underlying cognitive preservation and inform intervention strategies.

Download

Share

COinS