Black older adults are more likely to develop dementia, and at a younger age, than White older adults, and cognitive performance differs between race as a function of age. Default mode network (DMN) function has been linked to normal and abnormal aging and may represent a mechanistic neuromarker of these effects. The current study evaluated the impact of age, DMN functional connectivity (FC), blood glucose (HbA1C), and race (as a proxy for social determinants of health) on verbal episodic memory task (Rey Auditory Verbal Learning Task – RAVLT) in a matched sample of Black and White older adults. We predicted that the RAVLT performance would be inversely associated with age and HbA1C, and positively associated with DMN FC. Finally, Black older adults were also expected to perform more poorly on the RAVLT. Participants included a subsample of the Human Connectome Project – Aging sample. Fifty Black adults over 50 were matched to 50 White participants on age, education, and sex. 14 participants and their matched pair were excluded during rigorous functional MRI (fMRI) quality control, reducing the sample 72. Demographic, serum, fMRI, and cognitive assessments were completed in one visit. FMRI data processing included a priori seed region FC of the DMN (i.e., mPFC, IPL, PCC, hippocampus). Analyses were performed with the full sample of 72 participants and a subsample of 50 with complete blood glucose data (i.e., blood glucose sample, n=50). A three-step hierarchical linear regression was conducted on the full sample with age, DMN FC, and race as independent variables and RAVLT total score as the dependent variable. None of these models significantly predicted RAVLT performance. A four-step hierarchical linear regression was performed in the blood glucose sample using age, HbA1C, DMN FC, and race as independent variables and RAVLT total score as the dependent variable. The first step of the regression with age as a predictor was significant; however, steps two, three, and four did not yield further significantly predictive utility. While results did not support a priori hypotheses, they provide a unique matched sample approach incorporating brain, health, and social factors into a single study.
