Exercise for ADHD Executive Function: Best Modality, Unclear Dose

The useful answer comes before the precise one

If you want the part you can use now, it is this: in children and adolescents with ADHD, exercise that makes the brain and body work at the same time beats plain aerobic exercise for executive function, especially inhibitory control and working memory. That conclusion has shown up in five separate meta-analyses from independent research groups, including Yang et al. (2026), Alhwaiti et al. (2025), Li et al. (2025), Ouyang et al. (2025), and Wang et al. (2026).

The harder question is dose. A 2025 Bayesian dose-response meta-analysis by Pan et al. tried to move the field past the usual yes-or-no verdict on exercise and estimate domain-specific sweet spots for cognitive flexibility, working memory, and inhibitory control across 33 studies. That is the right question. It is also where the evidence gets shaky. The abstract reports dose-dependent patterns and at least one acute optimal estimate, but the available full text did not match the abstract, so the paper's detailed MET-minute values, credible intervals, and curve shapes could not be independently checked. Precision is the headline. Verification is the problem.

What the literature agrees on

There is broad agreement that exercise helps executive function in ADHD youth. A 2025 meta-analysis of randomized controlled trials by Pengfei Wang et al. found that aerobic exercise alone improved all three core executive-function domains: inhibitory control, working memory, and cognitive flexibility. That matters because it sets the floor. Plain aerobic exercise is not useless. It works. It just does not look like the best option.

When researchers compare modalities rather than lumping them together, the ranking gets more interesting. Across network meta-analyses and a multilevel meta-analysis, combined cognitive-motor exercise repeatedly comes out on top for inhibitory control, with similar advantages for working memory. Ouyang et al. compared six exercise types across 26 randomized controlled trials and ranked combined cognitive-motor exercise highest across inhibitory control, broader executive function, and gross motor outcomes. Yang et al. and Li et al. reached the same general conclusion while also testing prescription variables. Different teams, different analytic methods, same winner. That is not proof of a universal law, but it is the most replicated pattern in this literature.

Trial-level evidence points the same way. He et al. (2026) ran a randomized controlled trial in children with ADHD and found that combined exercise outperformed an aerobic-only condition for working memory. Zhu et al. (2026) reported working-memory gains with integrated cognitive-motor exercise in a randomized clinical trial targeting both core ADHD symptoms and executive functions. Those studies do not settle the dose question, but they do strengthen the case that adding cognitive demands changes the result.

Why combined exercise probably wins

"Combined" is a clunky label for a simple idea: make the child move and think at the same time. That can mean martial arts, dance patterns, obstacle courses with navigation, sport drills with fast rule-switching, or running tasks that include memory and inhibition demands. The brain does not care whether the activity sounds elegant on a grant application.

A controlled trial by Huan Wang et al. (2025) gives that idea some biological support. The intervention was cognitively engaging running in children with ADHD, not just steady aerobic output. It improved inhibitory control and increased prefrontal activation measured by fNIRS. The same study also found that physical self-efficacy significantly moderated the benefit. Children who believed they could handle the activity did better. That is not a cute side note. It means the best-looking prescription on paper may underperform if the child hates it, avoids it, or feels incompetent doing it.

The dose story is less clean than the headline suggests

The prescription variable that shows up most consistently is not frequency by itself. It is the interaction between intensity and session duration, often expressed as MET-minutes. Yang et al. and Li et al. both identified the duration-intensity combination as the strongest predictor for inhibitory-control improvement. That is useful directional guidance. It tells parents and clinicians to care more about how hard and how long the session is than about chasing a magical number of weekly checkboxes.

What those studies did not do is deliver a defensible universal threshold. Both network meta-analyses ran into the same wall: between-study heterogeneity. Participants ranged from ages 6 to 18, which spans several neurodevelopmental stages. ADHD subtype reporting was inconsistent. Stimulant medication was often not controlled. Outcome measures for executive function varied across studies. Intervention length ranged from 4 to 24 weeks. Once you pile all that into one model, the estimate may look precise because software can always draw a curve. The child in front of you may not cooperate with the curve.

That concern is not theoretical. Khodaverdi et al. (2022), the only pre-Bayesian review here that directly asked whether a precise effective aerobic dose for executive function in ADHD youth could be established, concluded that the evidence was too heterogeneous to converge on a single threshold. The Bayesian framework is a smarter way to handle uncertainty than pretending it does not exist. It does not erase the underlying variance.

Measurement adds another problem

Even if there is a real neurological dose-response curve, behavioral tests may not capture it cleanly. In a controlled study of adolescents with ADHD, Van Riper et al. (2023) found that aerobic exercise effects were detected more reliably in prefrontal brain activity than in behavioral executive-function performance. That brain-behavior mismatch matters. If the field estimates precise dose thresholds from noisy behavioral tasks, some of the curve may reflect measurement limits rather than the true biology.

The medication issue makes this worse. Stimulants can improve inhibitory control and working memory on their own, and most randomized trials in these meta-analyses did not stratify cleanly by medication status. That leaves open the possibility that medicated and unmedicated children have different baselines, different responsiveness to exercise, and different ceiling effects. Any dose model built on mixed populations inherits that blur.

The real-world gap is not subtle

No major authority has issued ADHD-specific exercise guidelines for executive function. The only widely recognized benchmark is the World Health Organization recommendation of at least 60 minutes per day of moderate-to-vigorous physical activity for children ages 5 to 17, summarized in Hansen et al. (2022). That is a population guideline, not an ADHD prescription, and it says nothing about executive-function domains.

Meanwhile, many young people with ADHD do not meet even that baseline. Taylor et al. (2023) and Parnes et al. (2025) both found low adherence to 24-hour movement guidelines in ADHD youth, with better adherence associated with better cognitive or behavioral outcomes. That leaves the field in an awkward but familiar place: researchers are trying to fine-tune the optimal exercise dose for a population that often struggles to hit the generic minimum. It is like arguing over the best tire pressure while the bike is still missing a chain.

The practical priority is adherence first, precision second. A sustainable cognitive-motor program at moderate-to-vigorous effort is more plausible than a mathematically elegant prescription that dies in week three.

What parents, clinicians, and very motivated spreadsheet people can do with this

• Use exercise for ADHD executive function as a real intervention, not as wellness wallpaper. The evidence supports small-to-moderate benefits.
• Favor combined cognitive-motor activities over plain aerobic work when the goal is inhibitory control or working memory.
• Treat duration plus intensity as more important than frequency alone, but do not pretend the literature has nailed a universal MET-minute target.
• Expect individual variation. Self-efficacy, developmental stage, and probably medication status can change the response.
• Assume ongoing practice is required. The available evidence does not report post-intervention follow-up on whether gains persist after training stops.

The bottom line is refreshingly plain. The field has a strong modality signal and a weak precision signal. If you want the best evidence-based move today, choose exercise that forces coordination, attention, and movement at the same time, then make it doable enough to keep happening. The Bayesian dose map may eventually sharpen that advice. Right now, it is still penciled in.