The developmental trajectories of executive function from adolescence to old age

Abstract

Executive functions demonstrate variable developmental and aging profiles, with protracted development into early adulthood and declines in older age. However, relatively few studies have specifically included middle-aged adults in investigations of age-related differences in executive functions. This study explored the age-related differences in executive function from late childhood through to old age, allowing a more informed understanding of executive functions across the lifespan. Three hundred and fifty participants aged 10 to 86 years-old completed a battery of tasks assessing the specific roles of inhibitory control, working memory, cognitive flexibility, and planning. Results highlighted continued improvement in working memory capacity across adolescence and into young adulthood, followed by declines in both working memory and inhibitory control, beginning from as early as 30-40 years old and continuing into older age. Analyses of planning abilities showed continued improvement across adolescence and into young adulthood, followed by a decline in abilities across adulthood, with a small (positive) change in older age. Interestingly, a dissociation was found for cognitive flexibility; switch costs decreased, yet mixing costs increased across the lifespan. The results provide a description of the developmental differences in inhibitory control, working memory, cognitive flexibility and planning, above any effects of IQ or SES, and highlight the importance of including middle-aged adults in studies seeking to establish a more comprehensive picture of age-related differences in executive function.

Figure 1

Illustrations of the stimuli and procedure employed in each of the four EF tasks: (A) Stroop colour-word task; (B) operation span; (C) task switching; (D) Tower of Hanoi.

Figure 2

Relationship between age and executive function measures, adjusted for IQ and SES index. (A) log-transformed Stroop congruency effect, (B) OSpan partial score, (C) Tower of Hanoi score, (D) log-transformed Task Switching switch cost; and (E) log-transformed Task Switching mixing cost. The bold line indicates the best-fitting regression line and the dashed line indicates the 95% confidence intervals (CIs). Stroop congruency effect and Task Switching switch and mixing costs are reversed scored so that a higher value indicates better performance and all variables are z-scored for ease of comparison. Note: All measures were adjusted for IQ and SES to be comparable to the described regression models. To adjust for IQ and SES, the residuals were obtained from the regression line fit when fitting each executive function measure as a dependent variable in a linear model and IQ and SES index as predictor variables.