Aerobic Exercise and Executive Function: The Dose Matters

Your Brain Does Not Want One Exercise Prescription

If your goal is "better brain health," aerobic exercise helps. That part is old news. The newer and more useful point is narrower: the best exercise dose depends on which executive skill you want to protect. A 2024 systematic review and meta-analysis by Ye et al. pooled 42 randomized controlled trials in 2,881 healthy adults aged 45 and older and found that cognitive flexibility, working memory, inhibitory control, and planning do not respond the same way.

The differences were large enough to matter in real life. Working memory improved the most, with a pooled effect size of Hedge's g = 0.392 (p<0.001). Cognitive flexibility improved too, at g = 0.343 (p<0.001). Inhibitory control improved less, but still significantly, at g = 0.229 (p<0.001). Planning barely moved: g = 0.094, with p = 0.440. That is the opposite of a "just move more" message. It says the executive system is not one thing, and the brain does not reward every kind of effort equally.

What Actually Worked, and for What

The most practical part of the Ye analysis was its dose-response breakdown. The authors did not just ask whether aerobic exercise helped. They asked what kind, how long, how often, and at what intensity each executive subdomain seemed to respond best.

For cognitive flexibility, the strongest pattern came from mind-body aerobic exercise such as tai chi, yoga, and qigong, done for 13 to 24 weeks, in sessions lasting 46 to 60 minutes, 5 to 7 days per week. That adds up to roughly 230 to 420 minutes a week. For working memory, the apparent sweet spot was different: general aerobic exercise such as walking, running, or cycling at progression intensity, again for 13 to 24 weeks, but with shorter sessions of 20 to 45 minutes, 5 to 7 days per week, or about 100 to 315 minutes weekly. Inhibitory control had the oddest profile. It peaked with low-intensity general aerobic exercise, 20 to 45 minutes per session, only 3 to 4 days per week, for an estimated 60 to 180 minutes weekly.

That low-intensity result matters because it breaks the lazy assumption that harder is always better. Sometimes the best cognitive dose looks less like boot camp and more like a brisk, sustainable habit.

Planning Did Not Improve

The null result for planning is not a footnote. It is the boundary of the claim. Across the same 42 trials that found benefits for other executive functions, planning did not improve in a meaningful way. The pooled estimate, g = 0.094 with p = 0.440, does not prove a true effect of exactly zero. It does tell you that aerobic exercise alone did not produce clinically important gains on planning tasks in this literature.

That makes sense biologically and behaviorally. Planning tasks demand hierarchical goal setting, sequencing, and prospective simulation. Aerobic exercise may raise arousal, support plasticity, and improve blood flow, but that does not mean it trains complex multi-step reasoning. If planning is the target, the current evidence points toward adding explicit cognitive demands rather than hoping a treadmill will teach project management to your prefrontal cortex.

Why the Effects Might Differ

The simplest plausible mechanism is not a single mechanism. The evidence supports at least three pathways, all incomplete.

First, brain-derived neurotrophic factor, or BDNF, has decent support as a mediator. A 2025 meta-analysis by Cheng et al. found that walking, running, and cycling all significantly increased circulating BDNF in older adults. A separate 2025 meta-analysis by Gholami et al. found that chronic exercise training increased resting BDNF concentrations in older adults across randomized trials, which fits better with the 13-to-24-week intervention window in the Ye paper than with a one-off post-workout spike.

Second, aerobic training seems to change cerebral perfusion. In a 2024 randomized controlled trial by Tomoto et al., one year of progressive moderate-to-vigorous aerobic exercise increased cerebral blood flow in cognitively normal older adults. That is a biomarker outcome, not proof that every participant thought better afterward, but it is a plausible route by which exercise could support executive performance over time.

Third, the low-intensity advantage for inhibitory control may fit an inverted-U catecholamine model: too little stimulation does not help, too much may overshoot the prefrontal cortex's preferred operating range. That explanation is plausible, not settled. The data support the low-intensity pattern; the mechanism remains partly speculative.

The biggest mechanistic gap concerns cognitive flexibility. If general aerobic exercise already raises BDNF, why would mind-body exercise do better for set-shifting? One reasonable answer is that tai chi, yoga, and qigong are not just cardio with nicer lighting. They add motor sequencing, sustained attention, body-position monitoring, and in some settings social coordination. Those ingredients may matter. The problem is that the available trials do not cleanly isolate them.

The Evidence Gets Messier Once You Stop Averaging People Together

The strongest challenge to a universal recommendation comes from a 2023 randomized trial by Barha et al. In that study, aerobic exercise improved executive function in females without the BDNF Val66Met polymorphism, but not in males and not in Met-allele carriers. That is not a minor quirk. It suggests the pooled gains in meta-analyses may reflect a strong response in some groups and little or no response in others.

This does not overturn the Ye meta-analysis. It does limit how confidently you can apply its average effect sizes to every reader. If half the population responds differently, population means start to hide as much as they reveal.

Other studies complicate the picture in a different way by studying different populations. The 2025 EXERT trial tested moderate-high intensity aerobic training versus low-intensity stretching, balance, and range-of-motion work for 18 months in sedentary older adults with amnestic mild cognitive impairment. Intervention effects on cognition did not differ between groups at 12 months, even though there was no cognitive decline in the aerobic group. That result does not contradict the Ye paper directly because the population was not healthy older adults. Mild cognitive impairment is a different biological problem and should not be smuggled into a healthy-aging conclusion.

The same caution applies to a 2021 trial by Tomoto et al. showing that one year of aerobic exercise altered cerebral vasomotor reactivity in people with mild cognitive impairment, with vascular changes linked to cognitive outcomes. Useful mechanism data, yes. Direct proof for healthy adults, no.

A broader 2025 network meta-analysis by Zhang et al. supports the idea that exercise type matters for cognitive outcomes in older adults, which lines up with Ye et al.'s subdomain-specific findings. Still, network meta-analysis depends on indirect comparisons across heterogeneous trials, so it cannot rescue every detail of the dose-response story.

What to Do With This if You Are 45 or Older

The most defensible conclusion is not that everyone needs one magic exercise routine. It is that aerobic exercise is a useful tool for some executive functions, with different best bets for different goals.

• For working memory, the evidence favors general aerobic exercise such as walking, running, or cycling for 20 to 45 minutes, 5 to 7 days per week, over 13 to 24 weeks.
• For inhibitory control, lower intensity appears better than pushing hard, with 20 to 45 minutes, 3 to 4 days per week, over 13 to 24 weeks.
• For cognitive flexibility, mind-body formats such as tai chi, yoga, and qigong look better, but at a higher time cost: 46 to 60 minutes, 5 to 7 days per week, for 13 to 24 weeks.
• For planning, aerobic exercise alone is not the right tool based on current randomized evidence.

The useful message is narrower than "exercise boosts cognition" and more actionable: match the exercise dose to the cognitive function you care about, and do not expect the same payoff across all executive skills.

The Limits Matter

Before anyone starts prescribing 57 minutes of qigong for task-switching, the subgroup results need humility. The 42 trials in the Ye meta-analysis averaged about 69 participants each. Once the data were split into bins by duration, intensity, session length, and frequency, many subgroup cells were likely small. That makes the "optimal" doses informative but statistically fragile.

There are other limits. The mind-body category is broad and messy. Peripheral blood BDNF is an imperfect proxy for what happens inside the brain. The full risk-of-bias distribution across the Ye trials was not available from the abstract. Publication bias in exercise-cognition research is a real concern, and the abstract did not provide enough detail to quantify it. Even the planning null has a ceiling: a tiny true effect could exist and still be too small to matter.

So the evidence supports a real benefit for some executive functions in healthy adults over 45. It does not support a one-size-fits-all protocol, and it does not support the stronger fantasy that aerobic exercise tunes up every prefrontal skill if you just suffer hard enough. The brain, annoyingly, has preferences.