Exercise, Cognition, and the Limits of What Research Can Actually Tell Us

This episode of Yoga Research and Beyond came from a listener request, which I always appreciate. It tells me people aren’t just consuming research headlines, they want help making sense of them.

The paper we discussed is titled “Effectiveness of Exercise for Improving Cognition, Memory and Executive Function: A Systematic Umbrella Review and Meta-Meta-Analysis,” published in 2025 in the British Journal of Sports Medicine. That’s a mouthful, but the scope matters. This paper asked a big question: Does exercise improve cognition, memory, and executive function across different ages, populations, and types of exercise?

Short answer: yes, but not in the way headlines tend to suggest.

What Kind of Paper Is This, Really?

This wasn’t a single experiment. It wasn’t even a typical systematic review. It was an *umbrella review*—a review of systematic reviews, all of which examined randomized controlled trials (RCTs).

The authors started with over 6,600 studies. After removing duplicates and excluding low-fit papers, they ended up with 133 systematic reviews, covering 2,724 RCTs and more than 258,000 participants.

That filtering process matters. It reduces cherry-picking and dampens the temptation to build big claims on a single flashy study. This is exactly why I like reviews like this: they force us to look at the body of evidence, not just the outliers.

That said, an umbrella review inherits the strengths and weaknesses of the reviews it includes. And in this case, that’s important.

What They Measured (and What They Didn’t)

The authors looked at three outcomes:

• General cognition
• Memory
• Executive function

These were measured using standardized cognitive tests—games, questionnaires, lab tasks, and assessments designed to capture changes from pre-intervention to post-intervention.

What they did not measure:

• Longevity
• Independence in daily living
• Reduced caregiving burden
• Delayed need for assisted living

That distinction matters more than most people realize.

The Results, Without the Hype

Across the board, exercise showed small to moderate positive effects on cognition.

Using standardized mean differences (effect sizes):

• General cognition: 0.42 (medium effect)
• Memory: 0.26 (small effect)
• Executive function: 0.24 (small effect)

Children and adolescents showed larger gains in memory and executive function. People with ADHD also benefited more than average. Light-to-moderate intensity exercise sometimes outperformed high-intensity exercise. Shorter programs often showed bigger effects than longer ones.

Mind-body practices, like yoga and tai chi, showed meaningful effects on memory. And the biggest effect on general cognition came from “exergames,” which combine movement with video-game-style cognitive challenges.

That result annoyed me a little. But it also made sense.

Why These Results Are Easy to Misuse

This is where I start pushing back.

The effect sizes here are not strong enough to support cause-and-effect claims like:

• “Yoga prevents dementia.”
• “Exercise will fix your memory.”
• “This practice will protect your brain as you age.”

That leap is where research gets weaponized by wellness marketing.

I have a close friend whose mother practiced yoga faithfully for decades. Multiple times a week, socially engaged, physically active. She still developed dementia, and her decline has been severe. To suggest that more yoga would have prevented that isn’t just inaccurate. It’s cruel.

Exercise may help slow the decline. It may help people perform better on specific cognitive tasks. It may help maintain engagement with life. But you cannot exercise your way out of genetics. And studies like this don’t tell us how these small, short-term effects translate into real-world outcomes over decades.

The External Validity Problem

Most of these studies measured performance on the same kinds of tasks participants practiced during the intervention. If you train in memory games, you get better at memory games. That’s not shocking.

What I want to know, and what this paper can’t answer, is whether those gains meaningfully change people’s lives.

• Do they live independently longer?
• Do families spend less on home health care?
• Do people manage daily tasks more effectively as they age?

Those are the outcomes that actually matter. And they’re much harder to study.

Important Limitations (That the Authors Acknowledged)

To their credit, the authors didn’t gloss over the flaws.

• 71% of the included reviews were rated “critically low” quality.
• Many reviews failed to pre-register protocols.
• Many didn’t report excluded studies or funding sources.
• Cognitive tests often weren’t sensitive enough for healthy adults.
• Shorter trials may look better due to novelty, weak control groups, or lower dropout rates.

This paper was pre-registered, which is a green flag. But it still sits atop a shaky pyramid.

So What Do We Do With This?

Here’s my take.

Exercise is good. Any kind of movement that keeps people engaged in their bodies and their lives is worth supporting. Mind-body practices belong in that conversation.

But this research doesn’t justify turning yoga, or any exercise, into a cognitive cure-all. It doesn’t support moralizing health outcomes. And it definitely doesn’t support blaming individuals when biology does what biology does.

Use this research to support participation, not to sell certainty.

Because in the end, we’re all headed in the same direction. Exercise might help us navigate the path with more engagement and resilience, but it doesn’t rewrite the destination.

And research literacy means being honest about that.