Is Exercise Good for the Brain? Checking the Effect Sizes

Exercise does help cognition. It just helps less than the marketing suggests. What improves are specific task scores such as executive function and memory, with effect sizes in the small-to-moderate range. This article walks through the numbers and what can honestly be claimed from them.

What the meta-analyses report

A network meta-analysis in Frontiers in Aging Neuroscience (2025) pooled 58 randomized controlled trials covering 4,349 cognitively healthy older adults and compared five exercise modalities. Effect sizes (SMD) on overall cognitive function came out as follows.

• Mind-body exercise (tai chi, yoga, and similar): 0.62
• Resistance training: 0.55
• Aerobic exercise: 0.49
• Multicomponent training: 0.38

An effect size around 0.5 is conventionally classed as moderate. Narrowed to executive function in healthy older adults, however, a separate three-level meta-analysis reported a pooled effect of g = 0.155 (95% CI 0.084–0.226), with working memory at g = 0.208 and cognitive flexibility at g = 0.150 — small.

Same topic, different populations and outcome tasks, different numbers. 'Exercise is good for the brain' does not pin down a claim unless it says which number it refers to.

Why the effect sizes are hard to interpret

Several things drive the heterogeneity: participants' age and baseline cognitive status, the type, intensity, and duration of the exercise, and which tasks were chosen as outcomes. In studies of people with mild cognitive impairment, aerobic exercise shows a large effect on global cognition (around SMD = 0.81) while producing no significant improvement in executive function or processing speed.

It is unsurprising that groups with more room to improve show larger effects. The corollary is that healthy adults with no cognitive complaint should not expect the same numbers.

Why 'exercise raises your IQ' overstates it

Improving a task score and raising IQ are different claims. Training reliably improves the trained task; it often carries to closely similar tasks (near transfer); and it rarely reaches intelligence in general (far transfer). That pattern has been confirmed repeatedly in cognitive training research.

Exercise research is no exception. A few points of improvement on an executive function task sits a long way from a composite index like IQ moving. We cover the same question in can you raise your IQ and hobbies that train the brain.

Why exercise is still worth recommending

A small effect size is not the same as no reason to bother. Exercise brings benefits alongside the cognitive ones.

• Sleep quality tends to improve, which feeds directly into the next day's performance (see sleep and cognitive function)
• Mood and anxiety measures tend to improve
• Blood pressure, blood glucose, and weight respond, which matters for long-run health risk
• Side effects are minimal and the cost is nothing

Treating the cognitive effect as a bonus keeps expectations and reality closer together.

A realistic way to build it in

1. Start at roughly three sessions of 30 minutes a week: close to the floor of most intervention designs

2. Give it 12 weeks before judging: do not try to measure an effect over a few days

3. Combine aerobic and resistance work: the case for picking only one is weak

4. Choose what you will keep doing: adherence differences dwarf effect-size differences

5. Do not cut sleep to exercise: that trade usually nets out negative

Track the change as a number

BrainRank's free IQ test scores 20 questions across spatial reasoning, pattern recognition, logical reasoning, and classification (about 10 minutes) using item response theory (IRT). Comparing before a habit change with a few months later, under matched conditions, gives you domain-level numbers instead of an impression. A single score carries measurement error, so a difference of a few points is safest treated as no difference. Results are estimates for entertainment and self-understanding, not a medical or diagnostic assessment.