Abstract:Background: In many developed countries, cognitive functioning (as measured by neuropsychological tests) appears to be improving over time in the population at large, in parallel with the declining age-specific incidence of dementia. Here, we investigated cohort effects in the age-associated trajectories of verbal memory function in older adults. We sought to determine whether they varied by decade of birth and, if so, whether the change would be explained by increasing educational attainment. Methods: Pooling data from two prospective US population-based studies between 1987 and 2015, we identified four birth cohorts born 1902-1911, 1912-1921, 1922-1931, and 1932-1943. Among these cohorts, we compared age-associated trajectories both of performance and of practice effects on immediate and delayed recall of a 10-item Word List. We used mixed effects models, first including birth cohorts and cohort X age interaction terms, and then controlling for education and education X age interaction. Results: We observed significant cohort effects in performance (baseline and age-associated trajectories) in both immediate recall and delayed recall, with function improving between the earliest- and latest-born cohorts. For both tests, we also observed cohort effects on practice effects with the highest levels in the latest-born cohorts. Including education in the models did not attenuate these effects. Conclusions: In this longitudinal population study, across four decade-long birth cohorts, there were significant improvements in test performance and practice effects in verbal memory tests, not explained by education. Whether this reflects declining disease incidence or other secular trends awaits further investigation.

Cohort effects in verbal memory function and practice effects: a population-based study