C-24 Longitudinal Trajectories of Working Memory Performance in Typically Aging Older Adults

Abstract

Abstract Objective Cross-sectional studies have shown age-related differences in working memory (WM), but the trajectory is unclear due to the scarcity of longitudinal studies. Additional research is needed to better characterize the course of age-related changes in WM in older adults. The present study sought to address this gap in the literature by conducting serial assessments of WM in a longitudinally followed cohort of typically aging adults. We hypothesized a significant age × time interaction, such that WM would show pronounced declines with advancing age. Methods 640 functionally intact participants in an aging cohort (clinical dementia rating = 0; age range 52-99, mean age = 75) completed a computerized WM measure, Running Letter Memory (RLM), every ~15 months for up to 8.5 years (mean follow-up = 1.9 years). Longitudinal changes in RLM scores were analyzed using linear mixed effects models, allowing for random slopes and intercepts. All models were adjusted for sex and education. Results RLM performance did not significantly decline over time (b = -.14, p = .43). As hypothesized, there was a significant age × time interaction predicting RLM scores (b = -.08, p = .006). Specifically, RLM performance remained relatively stable (or slightly improved) until around age 75, beyond which increasingly precipitous declines were observed with advancing age. Conclusion The present results suggest that WM performance does not evidence declines until the mid-70s in typically aging adults, at which point increasingly steep decline trajectories are observed with advancing age. These findings highlight that cognitive aging does not occur at a constant rate in late life.