Exercise and Dementia Prevention: Which Type Works Best?

“Exercise is good for the brain” is well-established advice, but it understates an interesting detail in the research: not every form of exercise appears to affect cognitive decline equally. A network meta-analysis comparing different exercise types head-to-head found that one form in particular stood out — and it was not the one most people would guess.

Exercise habit changes and dementia risk after stroke

A 2023 study published in Scientific Reports followed ischaemic stroke patients over eight years and grouped them by how their exercise habits changed after their diagnosis: persistent non-exercisers, exercise dropouts, new exercisers, and persistent exercisers. The differences in dementia incidence by year eight were substantial:

The gap between persistent exercisers and persistent non-exercisers — roughly 11% versus 24% — is more than a twofold difference. Notably, the “new exercisers” group, who started exercising only after their stroke, still ended up meaningfully better off than those who never did, suggesting it is not too late to benefit even after a significant vascular event.

A dose-response relationship

A broader systematic review and meta-analysis of cohort studies, published in PLoS One in 2026, examined physical activity, sedentary behaviour, sleep and dementia together, and found a dose-response pattern: as physical activity level rose (measured in MET-minutes per week, a standard measure of activity intensity and duration combined), the adjusted hazard ratio for incident dementia dropped from a baseline of 1.0 down to roughly 0.7–0.8, before the benefit plateaued at around 800–1,000 MET-minutes per week. In practical terms, this points to a meaningful reduction in risk associated with regular activity, with the biggest gains achieved at moderate, sustainable levels of activity rather than requiring extreme volumes of exercise.

Some researchers in this field have gone as far as proposing that exercise be considered an important, complementary part of care planning at multiple stages of Alzheimer’s disease, from prevention through to established disease — though decisions about care for someone with a diagnosed condition should always be made with their treating doctor, taking their full clinical picture into account.

Which type of exercise works best?

Perhaps the most specific and useful finding in this area comes from a 2022 network meta-analysis published in the Journal of Sport and Health Science, which directly compared several categories of exercise — aerobic exercise, mind-body exercise (such as tai chi or yoga), multicomponent exercise, and resistance exercise — against a non-exercising control group, specifically for their effect on global cognition in people with mild cognitive impairment or dementia.

The result: resistance exercise had the highest probability of being the most effective exercise therapy for slowing cognitive decline, particularly for patients already experiencing dementia — outperforming aerobic exercise, mind-body exercise, and multicomponent programmes in the comparative rankings. This is a somewhat counterintuitive finding, since aerobic exercise (which improves cardiovascular fitness and blood flow) is often assumed to be the more “brain-relevant” form of activity. The data suggests resistance training — activities like weight-bearing exercises, resistance bands, or bodyweight strength training — deserves more attention in this context than it typically receives.

What this means in practice

None of this suggests aerobic exercise, walking, or multicomponent activity are without value — the Scientific Reports stroke study and the dose-response data both used general activity levels, not resistance training specifically, and still found strong protective associations. The more precise, actionable takeaway is that resistance training specifically appears worth including as part of an exercise routine aimed at supporting cognitive health, rather than defaulting to aerobic activity alone. A well-rounded routine that includes both aerobic and resistance components is a reasonable, evidence-aligned approach for most people, and any new exercise programme — particularly for older adults or those with existing health conditions — is worth discussing with a doctor first.

Frequently asked questions

Is it too late to start exercising for brain health if I’m already older or have had a stroke? The stroke-cohort research above found that even “new exercisers” who started after their stroke diagnosis had meaningfully lower dementia incidence than those who never exercised, suggesting starting later still provides a measurable benefit.

Which type of exercise is best specifically for memory, rather than overall cognition? Within the research summarised here, resistance exercise specifically showed a significant effect on memory function, while multicomponent exercise (combining several exercise types) was found most effective for slowing decline in global cognition and executive function more broadly.

How much exercise is needed to see a cognitive benefit? The dose-response data referenced above suggests measurable risk reduction begins well before extreme activity levels, with benefits plateauing around 800–1,000 MET-minutes per week — roughly consistent with standard public health activity guidelines rather than elite-level training volumes.

Can exercise replace medical treatment for someone already diagnosed with dementia? No. The research summarised here studies exercise as a factor associated with slower cognitive decline, not as a replacement for medical care. Any exercise programme for someone with a diagnosed cognitive condition should be discussed with their treating doctor.

This article is for general information only and does not constitute medical advice. If you have a diagnosed cognitive condition, cardiovascular condition, or are recovering from a stroke, consult your doctor before starting a new exercise programme.