Exploring epigenetic aging and cognitive change in middle adulthood.

Greater age deviation (AgeDev), in which a person's biological age is older than their chronological age, is associated with worse cognitive performance. Most findings, however, rely on lab-administered cognitive measures and do not account for time-ordered effects of practice. The present study leveraged a measurement burst dataset to examine 6 distinct features of cognitive performance (peak performance, change in peak performance, intraindividual variability, forgetting, retest-related gains), and explored whether individual differences in these features can be predicted by AgeDev. In a sample of 139 adults aged 25-65, baseline AgeDev (Horvath1, Horvath2, Hannum, PhenoAge, GrimAge), DunedinPACE, and SystemsAge were derived from DNA methylation. Across 4 annual high-frequency assessment bursts, processing speed and working memory were assessed via a Symbol Search and N-Back task, respectively. Using Bayesian Double Exponential Models (BDEMs), we found that baseline peak performance was lower for individuals with greater AgeDev, DunedinPACE, and chronological age. Interestingly, individuals with greater AgeDev and faster DunedinPACE showed greater retest-related gains within and across bursts. Neither AgeDev nor chronological age predicted differences in cognitive change across time. Individuals with greater AgeDev showed more intraindividual variability in processing speed. Finally, greater chronological age (but not AgeDev) was linked with more forgetting between bursts in both tasks. These findings underscore the value of multidimensional indices of cognitive performance. Importantly, even in a sample spanning young adulthood through early old age, such indices may provide unique information regarding the nature of longitudinal patterns in cognition associated with chronological or biological aging.