ADHD and exercise on low-capacity days

What the research says, and what it does not

Articles about ADHD and exercise often sound more certain than the studies behind them, so it helps to know what is actually there.

A 2017 systematic review looked at exercise in children and adults with ADHD. Cardio exercise showed benefits on some executive functions, attention and behaviour, both straight after a session and over longer programmes. Most of the studies were in children, most were small, and their quality was mixed. Exercise that was not cardio had weaker support.

A 2025 meta-analysis looked only at adults. Across 14 small trials, 373 adults with ADHD in total, single sessions and longer programmes of exercise improved inhibitory control, the ability to hold back a response. The authors found a risk of bias and possible publication bias, and other outcomes were not assessed.

Put together, that is a small claim. Exercise may help some aspects of ADHD for some people. It has not been shown to treat ADHD, and nothing here says it can replace medication. There is also no tested ADHD dose: no study tells you how many minutes, at what intensity, will help your attention.

For general health, the WHO guidelines for adults suggest 150 to 300 minutes of moderate aerobic activity a week, or 75 to 150 minutes of vigorous activity, plus muscle strengthening. That is a guide for most adults, not an ADHD prescription.

Why the plan collapses after a week

The plan that fails is usually written on a good day. You feel clear, you read that exercise helps, and you commit to five runs a week. That plan assumes the clear day will repeat.

Capacity does not repeat. Sleep, stress, illness, work, caring for someone and how many decisions are still open all change what you can do today. The book's example is Dana, who kept designing a twelve-part morning routine on good Sundays. On Thursdays the plan became the standard she was failing. An exercise plan built the same way meets the same Thursday.

Then a missed session turns into a missed week, and the plan feels broken. Getting back means facing it again, which is harder than the run itself.

Build a Low Power version of movement

The book's Low Power idea fits here. Low Power is the smallest version of a routine that still protects what matters. You build it from the outcome you want to keep, not by cutting pieces off the ideal plan. (The full method is in a low-power routine for ADHD bad days.)

For movement, the protected outcome is usually modest: the habit stays alive, so the full version is still within reach next week. The Low Power version only has to keep that true. It will not meet the weekly guideline by itself. That is fine. It is there for the days the full version cannot happen.

If the goal was a run every morning, running stays the full version for good weeks. On a bad day it becomes shoes on and once around the block, which is enough to keep the habit alive.

Missed days are not debt

A missed session does not have to be made up. One study of habit formation, in 96 adults from the general population, found that missing a single chance to do a behaviour did not materially affect how the habit formed. It was not a study of ADHD or of exercise, so treat it as a hint. It does point the same way as the book: what matters is how fast you come back.

Doubling the next session to pay back the missed one makes returning harder. The book calls the time from noticing a system has lapsed to having it working again restart latency, and judges a system by it, not by a streak.

After a gap, the Low Power version is where you start again.

Where this goes wrong

The small version grows. If the Low Power walk turns into a plan with a route, a distance and a target pace, it is a second Full Power plan. Cut it back.

Exercise becomes the treatment plan. If you notice you are using movement to avoid a conversation with a doctor about ADHD, sleep or mood, that is the conversation to have.

Pain or illness gets pushed through. A low-capacity day caused by illness or injury may need rest. Moving through it is not the point of a small version.