Children’s cognitive benefits from exercise peak at moderate levels before declining, study finds

New research using machine-learning models reveals that children’s cognitive scores improve with exercise up to a certain threshold, beyond which benefits diminish, highlighting a nuanced relationship between physical activity and brain development.

Researchers writing in Frontiers in Psychiatry report that children’s cognitive scores may rise with more exercise, but only up to a point, after which the benefit appears to taper off. Using data from 8,623 children in the China Family Panel Studies, the team built three random forest models to examine how overall exercise dose, daily activity time and weekly activity frequency related to cognitive ability. The study used interpretable machine-learning tools, including SHAP values and partial dependence plots, to make the patterns easier to read.

The authors say the results point to a non-linear relationship rather than a simple “more is better” effect. Their models showed moderate predictive performance, and the patterns they identified suggested possible thresholds of around 720 minutes a week, 6 days a week and 85 minutes a day. Beyond those levels, cognitive ability scores appeared to fall, which the researchers describe as evidence of a potential inverted U-shaped association.

The new findings fit with a wider body of research suggesting that movement can support thinking skills in children and adolescents. A network meta-analysis indexed by PubMed found that exercise was linked with small-to-moderate gains in general, fluid and crystallised intelligence, with dual-task balance training standing out as especially effective. Other studies have also found associations between physical activity duration and cognitive development in Chinese schoolchildren, reinforcing the idea that regular exercise may help support brain development during key growth periods.

At the same time, the Frontiers team cautions that the study is observational, so it cannot prove cause and effect. That limitation matters because family background, school environment and other habits can also shape cognition. Related research has shown, for example, that dietary patterns and teacher support are both linked with cognitive outcomes in Chinese children, underlining how exercise sits within a broader web of developmental influences. The authors say the dose ranges they identified may offer a preliminary guide for future programmes, but they stress that longitudinal and experimental studies are still needed before any evidence-based exercise recommendation can be made.