Does physical activity improve cognitive function in older adults?

What the strongest evidence says: a small but real benefit

The single largest and most rigorous study on this question is a 2024 meta-analysis that pooled data from 104 observational studies with over 341,000 participants [1]. It found that older adults who were physically active had a 3% lower risk of developing cognitive impairment or decline (risk ratio 0.97). This means that for every 100 people who might otherwise develop cognitive problems, about 3 fewer would do so if they were active. The same study also found that physical activity was linked to better scores on tests of global cognition, episodic memory (remembering specific events), and verbal fluency (the ability to quickly recall words) [1]. However, the effect was weak, and the benefit disappeared in studies that followed people for more than 10 years, suggesting that the protective effect may be strongest in the shorter term.

A separate 2023 meta-analysis of 20 randomized controlled trials (the gold standard for proving cause and effect) in older adults with mild cognitive impairment found that exercise had a large and statistically significant positive effect on cognitive function (standardized mean difference of 1.25) [4]. This means that, on average, people who exercised scored about 1.25 standard deviations higher on cognitive tests than those who did not. The best results came from programs that combined multiple types of exercise (aerobic, strength, balance), lasted at least 16 weeks, and were done 3-5 times per week for 35-50 minutes per session at moderate intensity [4].

What kind of activity works best? Variety and consistency matter

The evidence suggests that the type of activity matters, and that a combination of different exercises is more effective than any single type. A 2023 meta-analysis of 20 randomized controlled trials found that multi-component exercise (combining aerobic, strength, and balance training) and mind-body exercises (like tai chi or yoga) both significantly improved global cognitive function in older adults with mild cognitive impairment [6]. Multi-component aerobic exercise improved scores on the Montreal Cognitive Assessment by an average of 5.74 points, while mind-body exercise improved them by 1.29 points [6]. Conventional aerobic exercise (like walking or cycling) also helped, but the effect was smaller and less reliable [6].

Even light physical activity—the kind that doesn't make you breathe hard—can make a difference. A 2022 study of 145 older adults in Taiwan found that those who engaged in at least 3 hours of light physical activity per day (like slow walking or light housework) had an 84% lower chance of cognitive impairment compared to those who did less than 3 hours (odds ratio 0.16) [5]. This effect held even after accounting for more vigorous activity, suggesting that simply moving more throughout the day is protective. Another study found that light activity was specifically linked to better executive function (the ability to plan and organize) [7].

The benefits also extend to people with existing cognitive problems. A 2022 trial of 99 older adults with mild cognitive impairment found that 12 weeks of exercise training (aerobic and resistance, twice a week) significantly improved scores on the Mini-Mental State Examination and also improved physical function, such as handgrip strength [10]. A meta-analysis of 24 trials in people with dementia found a small but significant benefit of exercise on cognitive function (effect size 0.18), suggesting that even after cognitive decline has begun, staying active can help [9].

How does it work? And what are the limitations?

Physical activity likely helps the brain in several ways. One key mechanism is improving cerebrovascular health—the health of blood vessels in the brain. A 2023 study comparing 13 older adults who had exercised regularly for years with 13 sedentary controls found that the active group had much better cerebrovascular responsiveness (the ability of brain blood vessels to expand and deliver more oxygen when needed), and this was strongly correlated with better cognitive scores [8]. Another study in patients with atrial fibrillation (a heart condition that raises stroke risk) found that regular exercise was associated with fewer brain lesions (like silent strokes) and larger brain volume, both of which are linked to better cognition [3].

However, the evidence is not all positive, and there are important caveats. A large 2022 randomized controlled trial of 585 older adults with subjective cognitive concerns (but not dementia) found that neither mindfulness training, exercise, nor a combination of both improved episodic memory or executive function at 6 or 18 months compared to a health education control group [2]. This is a high-quality trial that directly contradicts the observational evidence. The key difference may be that the participants in this trial already had cognitive concerns, and the exercise program (targeting 300 minutes per week) may not have been enough to reverse existing decline. This highlights that physical activity may be more effective for prevention than for treatment.

Another limitation is that the benefit, while real, is small. The largest meta-analysis [1] found that the effect of physical activity on cognitive change was a standardized regression coefficient of only 0.01, meaning it explains just 1% of the variation in cognitive decline. For context, that is a much weaker effect than, say, the effect of education or genetics. So while physical activity is a worthwhile part of a healthy lifestyle, it is not a magic bullet. The best approach is to combine regular, varied exercise with other brain-healthy habits like a good diet, social engagement, and managing cardiovascular risk factors.

About These Sources

This answer is built on 10 peer-reviewed studies — published from 2021 to 2024, 2 from 2024 or later, 6 in Q1 journals, collectively cited 403 times — selected as the most relevant from 15 studies that passed quality screening, drawn from 87 papers retrieved from a database of over 500 million.

Sources used in this answer

Related research questions