Dose–response relationship of exercise intervention on cognitive function in patients with stroke: a systematic review and meta-analysis

OBJECTIVE: To systematically evaluate the effects of exercise interventions on cognitive function in stroke patients and to explore the dose-response relationship based on metabolic equivalent minutes per week (MET·min/week). METHODS: A comprehensive search was conducted from database inception through 25 August 2026, across five electronic databases (PubMed, Embase, Scopus, Web of Science, and the Cochrane Library), two clinical trial registries (ClinicalTrials.gov and the WHO International Clinical Trials Registry Platform [ICTRP]), and a gray literature database (ProQuest Dissertations & Theses Global). A nonlinear dose-response model was used to examine the association between weekly exercise dose (MET·min/week) and improvement in cognitive function after stroke. RESULTS: = 0.047), but not for exercise frequency, intervention duration, session duration, or time since stroke. The dose-response curve suggested a potentially nonlinear association between weekly exercise dose and cognitive effects within the observed dose range, with a relatively high model-predicted effect at approximately 465 MET·min/week (Hedges' g = 0.50, 95% CI: 0.26-0.73). CONCLUSIONS: Exercise interventions were associated with improvements in post-stroke cognitive function. Dose-response analysis suggested a potentially nonlinear association between weekly exercise dose and cognitive effects, with a relatively high effect estimate at approximately 465 MET·min/week. However, this finding was exploratory and requires further validation.