Psychological Bulletin, Vol 151(11), Nov 2025, 1382-1388; doi:10.1037/bul0000490
Recent reviews and meta-analyses suggest the cognitive benefits of physical exercise observed in primary studies may be inflated due to multiple sources of bias, including selection bias, placebo effects, regression to the mean, and publication bias. When these biases are accounted for, the evidence for the purported enhancements has been shown to be inconclusive. The recent meta-analysis by Mavilidi et al. (2025) makes a relevant contribution by pointing out the potential role of moderating variables and therefore the possibility that, under certain conditions, the effect may be more substantive. Yet, a critical evaluation of Mavilidi et al.’s methods reveals several issues in the statistical analyses and interpretation of publication bias analyses. These appear to have led Mavilidi et al. to conclude the presence of an overall cognitive benefit of physical activities. The present commentary provides a reanalysis of the data, applying appropriate methodological corrections. After an adequate analytical strategy, the final effect was reduced and yielded inconclusive evidence of an overall cognitive benefit. Particularly, publication bias methods highlight that the overall effect of chronic physical activity on cognition is likely smaller and therefore inconclusive. Yet, as in the original meta-analysis, the cognitive benefits were significantly larger for chronic interventions with an evidence-based delivery, programs with a clear cognitive component, or high cognitive demands compared to other physical exercise interventions. These results support the possibility that motor–cognitive training/sports games and holistic movement practices/martial arts may be effective activities to improve cognitive functioning. (PsycInfo Database Record (c) 2025 APA, all rights reserved)