"Exercise is good for mental health" is one of those statements everybody agrees with and nobody examines. The research is more interesting than the platitude, and knowing what it does and doesn't establish is genuinely useful.
What the evidence base looks like
There's a substantial body of randomised controlled trials examining exercise as an intervention for depression, and multiple meta-analyses pooling them.
The general finding is that exercise produces reductions in depressive symptoms compared with control conditions, with effect sizes that are clinically meaningful.
Several caveats are important and frequently omitted. Trials in this area are difficult to blind — participants know whether they're exercising — and blinding failure tends to inflate apparent effects. Many trials are small. And there's evidence of publication bias in the literature, with smaller positive studies more likely to be published.
Analyses accounting for these factors generally find smaller effects than the raw pooled estimates, and still find an effect. The honest summary is that exercise appears genuinely helpful for depressive symptoms, with the magnitude uncertain and probably smaller than the most enthusiastic claims.
How it compares
Some trials have compared exercise directly against antidepressant medication or psychotherapy.
The general finding in these comparisons is broadly similar effectiveness for mild to moderate depression, with substantial uncertainty given trial sizes.
That's frequently reported as "exercise works as well as antidepressants", which overstates the confidence available. What can reasonably be said is that exercise appears to be a legitimate treatment option, particularly for milder presentations, and that it's a reasonable adjunct alongside other treatments.
It is not a substitute for treatment in severe depression, and framing it that way is potentially harmful.
What kind and how much
The evidence on modality is less clear than people assume.
Aerobic exercise has the largest evidence base, simply because it's been studied most. Resistance training has a growing body of evidence with comparable effects. Mind-body practices like yoga have shown benefits in several trials.
What seems to matter more than modality is that it happens and that it's sustained. Adherence is the binding constraint in nearly every trial, and the intervention people will actually keep doing beats a theoretically superior one they abandon.
On dose, there's some evidence of a relationship between volume and effect, without a clearly established threshold. Meaningful benefits have been reported at levels well below official physical activity guidelines, which is encouraging for anyone starting from nothing.
Possible mechanisms
Several are proposed and none is established as the primary route.
Neurobiological. Effects on neurotrophic factors, on inflammation, and on neurotransmitter systems have been proposed and demonstrated in animal models. Translation to human clinical effect is less certain.
Behavioural activation. Depression involves withdrawal from activity, and structured activity is itself a component of effective psychological treatment. Exercise may work partly by being an activity that gets people out of the house and into a routine.
Self-efficacy. Completing something difficult and observing improvement provides evidence against the belief that nothing you do matters, which is a core feature of depressive thinking.
Social contact. Group-based exercise includes social interaction, which is itself protective. This confounds some trials.
Given that the mechanism is unclear, claims about specific biological pathways in popular coverage should be treated with caution.
The important caveat
The most common harm from this literature isn't clinical, it's how it's used.
Telling someone with depression that they should exercise is frequently unhelpful, because the illness itself impairs motivation, energy and initiation. Anhedonia and fatigue are symptoms, not choices, and being advised to do something the illness prevents can compound feelings of failure.
The trials that work generally involve supervised, structured programmes with support — someone expecting you, a scheduled session, assistance with getting started. That's quite different from advice to go for a run.
Which suggests the useful framing: exercise is a treatment that requires the same support any treatment requires, not a self-help instruction that removes the need for care.
Practical suggestions
For anyone considering this as part of managing their own mood.
Start considerably smaller than feels reasonable. Ten minutes of walking is a legitimate intervention and is achievable on a bad day, which a gym session may not be.
Attach it to something existing rather than creating a new commitment. Walking somewhere you were going anyway.
Prefer outdoor and social where possible, since both appear to add something, though the evidence there is weaker.
And treat it as one component rather than the answer. If symptoms are significant, persistent, or affecting your ability to function, that warrants professional assessment regardless of what else you're doing.
If you are experiencing symptoms of depression, please speak to a doctor or mental health professional. This article is general information and not a substitute for clinical care. If you are in crisis, contact emergency services or a crisis helpline in your country.