Optimal dose and type of exercise to improve cognitive function in older adults: A systematic review and bayesian model-based network meta-analysis of RCTs

Ageing Res Rev. 2022 Apr:76:101591. doi: 10.1016/j.arr.2022.101591. Epub 2022 Feb 17.


Objective: To examine the dose-response relationship between overall and specific types of exercise with cognitive function in older adults.

Design: Systematic Review and Bayesian Model-Based Network Meta-Analysis.

Data sources: Systematic search of MEDLINE, Web of Science, Scopus, PsycINFO and SPORTDiscus.

Eligibility criteria for selecting studies: Randomized controlled trials of exercise interventions in participants aged 50 years or over, and that reported on at least one global cognition outcome.

Results: The search returned 1998 records, of which 44 studies (4793 participants; 102 different effect sizes) were included in this review with meta-analysis. There was a non-linear, dose-response association between overall exercise and cognition. We found no minimal threshold for the beneficial effect of exercise on cognition. The estimated minimal exercise dose associated with clinically relevant changes in cognition was 724 METs-min per week, and doses beyond 1200 METs-min per week provided less clear benefits. We also found that the dose-response association was exercise type dependent, and our results show that clinically important effects may occur at lower doses for many types of exercise. Our findings also highlighted the superior effects of resistance exercises over other modalities.

Conclusions: If provided with the most potent modalities, older adults can get clinical meaningful benefits with lower doses than the WHO guidelines. Findings support the WHO recommendations to emphasise resistance training as a critical component of interventions for older adults.

Keywords: Aging; Cognition; Dose-response relationship; Exercise; Health planning guidelines.

Publication types

  • Meta-Analysis
  • Research Support, Non-U.S. Gov't
  • Review
  • Systematic Review

MeSH terms

  • Aged
  • Bayes Theorem
  • Cognition*
  • Exercise Therapy
  • Exercise*
  • Humans
  • Network Meta-Analysis