Abstract 4342417: Predictors of Cognitive Resilience in the Multi-Ethnic Study of Atherosclerosis

Background: Identifying modifiable factors of cognitive resilience (CR), an individual’s ability to sustain better-than-expected cognitive performance given the degree of brain pathology, may shed light on strategies to delay dementia onset. Objective: To quantify CR and identify its main predictors in a racially and ethnically diverse cohort in the United States. Methods: We selected 1,335 individuals enrolled in 2000-02 (Exam 1; age 45-84) in the Multi-Ethnic Study of Atherosclerosis (MESA). We focused on a composite score of global cognition based on three validated neuropsychological tests (Cognitive Abilities Screening Instrument, Digit Symbol Coding, and Digit Span) administered at Exams 5 (2010-12), 6 (2016-18), and 7 (2022-24). At Exam 6, neuroimaging biomarkers were assessed, including white matter fractional anisotropy, white matter hyperintensities, and total brain volume, adjusted for intracranial volume. First, we used linear mixed-effects models adjusted for age, sex, and race/ethnicity, to estimate CR as the random slopes of global cognition, capturing individual deviations from the (marginal) expected rate of cognitive decline given a specific neuroimaging biomarkers profile. CR estimates were z-scored to ease interpretation; positive values indicated better CR. Second, using general linear models, we assessed whether various modifiable social and health-related risk factors measured at Exam 1 were associated with CR a decade later. Each predictor’s contribution to CR was assessed by calculating the percentage change in the R-squared estimate by comparing the fully-adjusted model (with all predictors) to reduced models (individual predictors removed). Results: Study participants were on average 66 years-old (SD=8) and 46% male. In the fully-adjusted model, increased educational attainment and incremental family income were independently associated with higher CR (β=0.190; 95%CI=0.114, 0.267; β=0.119; 95%CI=0.064, 0.174, respectively). Conversely, more depressive symptoms and higher waist girth were independently associated with lower CR (β=-0.013; 95%CI=-0.021,-0.004; β=-0.010; 95%CI=-0.018,-0.002, respectively). The main contributors to CR were education and income (%R-squared Change=39% and 22%, respectively). Conclusion: These findings indicate that improving social (education and family income) and health-related factors (depressive symptoms and waist girth) early may promote CR decades later and help individuals preserve cognitive health.