Exercise improves general cognition, memory and executive functions across all ages, with small to moderate effect sizes. However, in people with advanced dementia or specific subgroups with memory complaints, the benefit is not always demonstrable in individual large trials.
Exercise has a measurable positive effect on the brain. The strongest evidence comes from a massive umbrella analysis of 133 systematic reviews involving more than 258,000 participants and 2,724 randomised studies (PMID 40049759). The average improvement in general cognition was an SMD of 0.42, in memory 0.26, and in executive functions (planning, focusing, switching) 0.24. These are small to moderate effects, but they apply to all ages and even at low intensity.
For people over 50, the message is similarly positive. A meta-analysis of 36 randomised studies found a pooled effect size of 0.29 for cognitive functioning (PMID 28438770). Aerobic training, strength training, combined training and tai chi all worked, with sessions of 45 to 60 minutes at at least moderate intensity yielding the greatest gains. Other meta-analyses show that older adults benefit relatively the most from exercise, while in children and young people memory and executive functions improve particularly strongly (PMID 37924980, 40049759).
There are also positive results in neurodegenerative conditions. In a randomised trial involving 130 people with Parkinson's disease, six months of home cycling led to better brain connectivity, less brain atrophy and improved cognitive control compared with a stretching group (PMID 34951063). This suggests that exercise may also slow brain decline in the context of a serious brain disease.
The picture is not without nuance, however. A well-designed large RCT with 585 older adults aged 65 to 84 who had memory complaints but not yet dementia found that 18 months of intensive exercise (target of 300 minutes per week) did not significantly improve memory and executive functions more than the control group (PMID 36511926). People with advanced dementia also did not appear to benefit cognitively from 24 weeks of walking and strength training in another RCT of 91 patients with a mean age of 82, although they did walk somewhat better during the intensive phase (PMID 32192537).
Regarding types of training: aerobic training helps most for general cognition, strength training appears most effective for executive functions, and mind-body exercises such as tai chi or yoga seem to contribute most to memory (PMID 37924980). A small addition: strength exercises combined with cognitive tasks performed simultaneously, so-called dual-task training, appeared to help older adults with cognitive impairments somewhat more than standard strength training alone, but that evidence is still limited by small sample sizes (PMID 38658827). Exercise is safe and has few side effects, but it is not a miracle cure: the effect varies by person, age and stage of any cognitive decline.
Based on one large umbrella review (PMID 40049759, 133 systematic reviews, ~258,000 participants), multiple meta-analyses of RCTs (PMID 28438770, 37924980) and individual RCTs (PMID 34951063, 36511926, 32192537, 38658827). Two well-designed RCTs (PMID 36511926 and 32192537) show no significant effect in specific subgroups, which is an important caveat to the otherwise predominantly positive picture.