Do exercise breaks improve concentration? Mostly, no

We sell a timer that puts exercise into your breaks. It would be commercially convenient to tell you the opposite. Here is the evidence instead.

What the question usually means

“Do exercise breaks improve concentration” is asked in two different ways, and they have different answers.

• Right now. You are stuck at 2pm, you do ten squats, and the question is whether the next twenty minutes of work are better. This is the acute question, and it is the one this article answers.

• Over months. You move more, sleep better, feel less wrecked at 5pm, and the question is whether your working life improves. That is a much longer causal chain, it is not what these studies tested, and the honest answer is that nobody has run that trial properly.

Almost every article that claims exercise snacks make you sharper is answering the second question with evidence from the first. That is where the overclaiming happens.

Three sets of researchers measured it. None of them found it

The studies below all interrupted sitting with movement and then tested cognition directly. They are the most relevant evidence there is, because the design matches what you would actually do at your desk.

Wennberg and colleagues, BMJ Open, 2016. Participants broke up sitting with three minutes of light walking every thirty minutes. Fatigue fell significantly at both the four-hour and seven-hour marks, which is a real and useful result. Executive function and memory did not move.

Duran and colleagues, Medicine & Science in Sports & Exercise, 2023. Eleven adults, eight hours of sitting, five different break protocols, run by a team at Columbia. Parts of the result are striking. The half-hourly five-minute walk took 58% off the glucose spike after a meal, and systolic blood pressure came down by four or five millimetres of mercury under every dose tested. Fatigue improved, mood improved, everywhere except the shortest hourly condition. Cognition, though: not one of the five protocols moved it.

Eleven people is a very small study, which cuts both ways. It is thin evidence for the glucose finding as well as for the null one.

A 2021 mapping review of the same literature reached the same place: breaking up sitting did not produce a demonstrable cognitive benefit.

Three separate looks, three times nothing. If exercise breaks sharpened thinking in a way you could feel, this is the design that would have caught it.

The number, and the bias underneath it

Zoom out from sitting breaks to acute exercise in general and the picture does not improve.

Writing in Communications Psychology in 2024, Garrett and colleagues went through 113 studies of acute exercise and cognition. Pooled, the answer was g=0.13, small enough to disappear into the noise of any single experiment. It also arrived with a warning attached. Selective publication had left strong marks on the literature.

Publication bias matters more here than the effect size does. It means the studies that found nothing were disproportionately never published, so the true average is likely to be even smaller than 0.13. When the correction for bias is bigger than the effect, the honest reading is that there may be no reliable effect at all.

Breaks in general do not do it either

Maybe the problem is exercise specifically, and any break would help? No.

Albulescu and colleagues pooled 22 samples covering 2,335 people in PLOS ONE in 2022, looking at micro-breaks of ten minutes or less. What they found:

• Vigour: up, d=0.36. Real.

• Fatigue: down, d=0.35. Real.

• Performance on the task: d=0.16, which did not reach significance at p=.116.

• Restricted to work that was cognitively demanding: d=−0.09. Nothing, tipped a fraction the wrong way.

Read those four lines together and you have the whole of this article. The feeling improves. The performance does not, and it least of all improves on exactly the kind of work you were hoping to protect.

One detail is worth keeping: whatever performance benefit showed up was concentrated in the breaks that ran longest. If you want it, five minutes is probably not the length to be arguing about.

The study everyone borrows, and what it actually showed

When a real experiment does get cited for the focus claim, it is usually Ariga and Lleras, Cognition, 2011. Participants who switched briefly to a different task at roughly ten-minute intervals held their performance steady across forty minutes of vigilance work. The groups that did not switch all fell away.

Genuine result, and an interesting one. It is also not what it is being used for. What was measured was sustained vigilance, what did the interrupting was another mental task rather than any kind of movement, and the session was forty minutes long. Extending that to “movement breaks make you better at writing code” is an analogy. So is leaning on Mackworth’s 1948 vigilance-decrement work, which was real and concerned a man watching a radar screen.

Analogies are fine. They are just not findings, and you should be told which one you are being sold.

What actually holds up

This is not an argument for sitting still. Three things survive scrutiny, and they are worth more than the claim being dropped.

Energy and mood, reliably. Reed and Ones pooled 158 studies covering 13,101 people and found acute aerobic exercise raises activated positive affect at d=0.47, an effect that lasts at least thirty minutes. Low-intensity exercise worked consistently, which means you do not have to destroy yourself for it. That is roughly three times the size of the disputed performance effect, and it is not disputed.

Fitness, from very little. A 2025 review in the British Journal of Sports Medicine, assembled by Ahmed and colleagues, drew on 11 trials and 414 adults classed as inactive. Exercise snacks under five minutes long, taken at least twice daily across four to twelve weeks, lifted cardiorespiratory fitness at moderate certainty, and nine in ten participants kept doing them. What did not move: body composition, blood pressure, blood lipids. Keep the expectations narrow.

Discomfort, and no bill attached. Six trials, 232 office workers, reviewed by Radwan and colleagues for Cogent Engineering in 2022. Breaks in which people moved beat breaks in which they simply stopped, measured on musculoskeletal discomfort. The part that matters commercially: neither typing speed nor the quality of the work went down.

That last point is the one to build on.

The honest claim: breaks do not cost you anything

The strongest thing you can say about movement breaks and work is negative, and it is still worth saying.

The British Journal of Sports Medicine carried a study in 2026 built on NPR’s Body Electric Challenge. Of 19,342 US adults recruited, 11,484 saw out fourteen days of five-minute movement breaks at intervals they picked themselves. Tiredness fell, low mood fell, good mood rose. Engagement with work climbed by 4 to 7%, and performance, as the participants rated it themselves, by 1 to 3%. Taking a break once an hour turned out to be the sweet spot, almost matching the half-hourly schedule while being far easier to sustain.

No control group, everything self-reported, so those percentages are soft and should never be quoted as hard effects. The direction is what settles the practical question. Nineteen thousand people interrupted themselves for a fortnight and their work held up.

So the case for exercise breaks is not “you will think better”. It is “you will feel better, you will be fitter, you will hurt less, and it will not cost you output”. That is a good deal. It just is not the deal being advertised.

How to read the next article that tells you otherwise

Four things to check, in order:

1. Did they measure cognition, or did they measure mood and call it focus? Vigour and fatigue are not concentration. They get swapped constantly.

2. Was the comparison a break, or no break? Feeling better than someone who sat still for four hours is a low bar.

3. Is the citation observational? Association is not effect. Look for the word “associated” in the original and check whether it survived the retelling.

4. How many people? A lot of the most-quoted numbers in this field, including the 58% glucose result, come from studies of eleven to twenty people.

If the vocabulary needs sorting out first, the terms are pulled apart in exercise snacks, movement snacks and deskercise, which shows that each one sits on a separate literature. The question of timing gets its own treatment in how often should you get up from your desk.

What we do with this

PomoFit puts a bodyweight exercise into every break. We think that is worth doing for energy, fitness and long-term health, and because the evidence says it will not slow you down. We do not think it will make you smarter, and you will not find that claim anywhere on the site.

To try it, the timer is here, and nothing needs registering to use the free tier. Choose an interval, do whatever appears on the break screen, skip the ones you do not fancy. If assembling your own sequence sounds like work, twelve ready-made plans already exist, and most people begin with either weekly brain fog killer or energy boost cardio. Pro runs $4.99 a month. The free version is a finished timer rather than a trial of one.