Exercise for Depression: What 218 Trials Found, and What They Did Not

The claim that exercise helps depression is old, vague, and easy to dismiss as the thing people say when they have nothing useful to offer. In 2024 it stopped being vague. A network meta-analysis in the BMJ pooled 218 randomised trials and 14,170 people, ranked the exercise types against each other and against antidepressants, and produced numbers specific enough to act on. It also rated its own confidence in most of those numbers as very low, which is the part worth understanding before you read the table.

What did the exercise and depression review actually do?

A network meta-analysis is not the same as a standard one. An ordinary meta-analysis pools trials of one treatment against one comparison. A network meta-analysis connects the trials into a web, so that treatments never directly compared in a single study can still be ranked against each other through their shared comparisons. It is how you get a league table out of a literature where nobody ran walking against yoga head to head.

The BMJ review gathered 218 unique studies, 495 treatment arms, and 14,170 participants with depression. It separated modalities rather than lumping them: walking and jogging, yoga, strength training, mixed aerobic exercise, tai chi and qigong, cycling, dance, and combinations with antidepressants or psychotherapy. Effects are reported as Hedges’ g against active control conditions, which means the comparison is usually something like a stretching group or usual care rather than nothing at all. That matters, because it makes the numbers harder to achieve than they would be against a waiting list.

Which exercise is best for depression?

Effect sizes are negative because the outcome being measured is depression severity, and a reduction is the goal. As a rough guide, 0.2 is small, 0.5 is moderate and 0.8 is large.

The line that gets quoted out of context is the last one. Exercise appearing to outperform antidepressants in a table is an irresistible headline and a misleading reading of what happened. The SSRI figure comes from the trials in this network that included a drug arm, compared against the same active controls; it is not a head-to-head verdict, and it sits on a wide confidence interval whose upper bound touches zero. The defensible statement is that exercise produced effects in the same range as standard treatments, not that it beat them.

The other line worth pausing on is dance, which produced the largest reductions of anything measured. The authors treat it cautiously because the trials were few and small, and a result built on a handful of studies is exactly the kind that shrinks when more data arrives. It is mentioned here because leaving it out would be selective, not because anyone should reorganise their week around it.

Which exercise works best for men with depression?

The review ran moderation analyses by sex, and the pattern was not uniform.

Strength training and cycling showed larger effects in women. In men, the stronger results came from yoga, tai chi and qigong, and from aerobic exercise combined with psychotherapy. Walking and jogging worked across both groups, which is one reason it tops the overall table.

That finding is worth sitting with, because it inverts the assumption most men bring to the subject. The intuition is that lifting is the male-coded intervention and yoga is the one for other people. The data in this review points the other way, and the mechanism people usually propose is plausible enough: the modalities that helped men most are the ones combining physical effort with sustained attention, breath control, or a therapeutic relationship, rather than pure physical output.

Two cautions. These were study-level moderation analyses, not individual patient data, which is a weaker form of evidence and more vulnerable to confounding between trials. And an average effect across a population tells you nothing about which activity you personally will still be doing in six weeks, which is the variable that actually determines the outcome.

How much exercise do you need for depression?

Intensity mattered. Vigorous exercise produced larger effects (g = -0.74, 95% CI -1.10 to -0.38) than light activity (g = -0.58, -0.82 to -0.33). The gap is real but modest, and both ends of the range were associated with meaningful reductions, which is the more useful message for anyone who is currently doing nothing.

There is a practical tension here that the paper cannot resolve. Depression reduces energy, motivation and the capacity to initiate activity, and it does so in proportion to severity. Prescribing the intervention with the highest effect size to the people least able to start it is a familiar failure mode. The honest reading is that vigorous exercise is somewhat better if you can do it, and that the intensity difference is small enough that starting with whatever you will actually repeat is the correct trade.

The trials in this literature typically ran supervised sessions two to three times a week for eight to twelve weeks. Very few followed anyone past a year, so the durability of the effect is unmeasured. That is a significant gap for a condition that recurs.

Why is the evidence rated low certainty?

The authors graded their own confidence as low for walking and jogging and very low for every other modality. This is not false modesty, and understanding why is the difference between reading the review correctly and misreading it.

The core problem is blinding. In a drug trial, a participant can take a pill without knowing whether it is the active one. There is no way to run someone through a twelve-week strength programme without their knowing. Everyone in an exercise trial knows which group they are in, and everyone has a prior belief about whether exercise helps mood. Since depression is measured almost entirely by self-report questionnaires, expectation bleeds directly into the outcome measure. Out of 218 studies, only one met the Cochrane criteria for low risk of bias.

Two further limitations. Many of the modality estimates rest on small trials, with a median of 17 participants per arm in the walking and jogging studies, which makes individual results unstable even when the pooled figure looks solid. And the sex, age and severity findings are study-level rather than individual-level, which means they describe differences between trials as much as differences between people.

None of this makes the conclusion wrong. The direction of effect is consistent across a very large literature, it is biologically plausible, and it agrees with clinical guidelines that already recommend exercise for mild to moderate depression. It does mean the precise ranking should be held loosely.

What should you actually do?

For mild to moderate depression, the evidence supports structured exercise as a genuine treatment rather than a supportive measure, and the most defensible version of the advice is unglamorous: walk or run, briskly, several times a week, for at least eight weeks before judging it.

For moderate to severe depression, the review does not support replacing anything. The strongest combination results in the whole analysis were exercise plus an SSRI and aerobic exercise plus psychotherapy, both landing at around -0.55. Exercise is additive here, and treating it as a substitute for care you are already receiving is not what this data says.

If you are a man deciding where to start, the sex-specific findings suggest trying the thing you were going to skip. Yoga and tai chi outperformed in men in this analysis, and the barrier for most men is not the difficulty of the activity.

And if symptoms are severe, if they have lasted more than a few weeks, or if you have thoughts of harming yourself, this is a conversation for a doctor rather than a training plan. In the United States, the 988 Suicide and Crisis Lifeline can be reached by call or text at any hour. Elsewhere, your local emergency number applies.

This article is for information only and is not medical advice. Depression is a treatable medical condition and exercise is one option among several. Do not stop or change prescribed treatment without speaking to the clinician who prescribed it.