Comparison of working memory performance in athletes and non-athletes : a meta-analysis of behavioural studies

Abstract

The relationship between sports expertise and working memory (WM) has garnered increasing attention in experimental research. However, no meta-analysis has compared WM performance between athletes and non-athletes. This study addresses this gap by comparing WM performance between these groups and investigating potential moderators. A comprehensive literature search identified 21 studies involving 1455 participants from seven databases, including PubMed, Embase, and ProQuest. Athletes primarily engaged in basketball, football, and fencing, while non-athletes included some identified as sedentary. The risk of bias assessment indicated low risk across most domains. Publication bias, assessed through a funnel plot and statistical tests, showed no significant evidence of bias. The forest plot, using a random effects model, revealed moderate heterogeneity. The overall effect size indicated a statistically significant, albeit small, advantage for athletes over non-athletes (Hedges’ g = 0.30), persisting across sports types and performance levels. Notably, this advantage was more pronounced when athletes were contrasted with a sedentary population (Hedges’ g = 0.63), compared to the analysis where the sedentary population was excluded from the non-athlete reference group (Hedges’ g = 0.15). Our findings indicate a consistent link between sports expertise and improved WM performance, while sedentary lifestyles appear to be associated with WM disadvantages.