Home › Psychology

📖 13 min read

Exercise for Depression: What Meta-Analyses Actually Show

What we’re looking at

Depression affects hundreds of millions of people worldwide, and first-line treatments—antidepressants, psychotherapy—don’t work for everyone. Meanwhile, physical exercise has been promoted as a natural mood booster for decades, often with enthusiastic testimonials but less rigorous evidence. That’s changed over the past decade. Multiple meta-analyses now synthesize data from hundreds of randomized controlled trials, asking whether exercise actually works as a clinical intervention for depression, how it stacks up against medications and therapy, and what dose and type matter most.

This evidence review synthesizes eight recent meta-analyses and systematic reviews spanning studies from 2014 to 2026. Together they cover thousands of participants—ranging from adults with major depressive disorder (MDD) to older adults with mild depressive symptoms, stroke survivors with post-stroke depression, and people with dementia. The goal: to separate hype from evidence and provide a clear picture of what exercise can and can’t do for depression.

The evidence — what studies actually found

Exercise versus doing nothing

When researchers pool randomized controlled trials that compare exercise to no treatment or waitlist controls, the effects are striking. A 2016 meta-analysis of 23 RCTs (977 participants) found a large effect size of -1.24 when exercise was compared to no intervention [PMID 27253219]. That’s considered a robust clinical benefit—comparable to what you’d see with first-line antidepressants in their own trials. Another 2016 systematic review focusing specifically on control-group responses in exercise trials found that even control groups in these studies improved substantially (effect size -0.92), though this partly reflects the high placebo and attention effects typical of depression research [PMID 26707338]. The net benefit of exercise over control conditions was still moderate to large overall (effect size -0.68) [PMID 27253219].

Exercise versus antidepressant medication

Direct head-to-head comparisons are rare but revealing. When exercise is compared directly to antidepressants, the difference is small and statistically nonsignificant (effect size -0.08) [PMID 27253219]. This doesn’t mean exercise is proven superior—it means that in the limited trials available, exercise performed about as well as medication. For context, a 2022 meta-analysis of antidepressants in cardiac patients (many eligible for cardiac rehab) found antidepressants had only a small effect versus placebo (effect size 0.17) [PMID 35797521], smaller than exercise effects in similar populations. This suggests exercise may be at least as effective as pharmacotherapy for many patients, though more direct comparisons are needed.

Exercise as add-on therapy

What if you combine exercise with standard treatment (usually medication or psychotherapy)? A 2021 systematic review of 22 studies found a moderate pooled effect (effect size -0.62) favoring exercise plus standard treatment over standard treatment alone [PMID 34565591]. The benefit was even larger in people with severe depression (effect size -0.99) and in those with a formal MDD diagnosis (effect size -0.70). An earlier 2016 meta-analysis found a similar trend: exercise as an adjunct to antidepressants yielded a moderate effect (effect size -0.50) that nearly reached statistical significance [PMID 27253219]. The take-home: exercise appears to enhance outcomes when added to existing treatments, particularly for patients with more severe symptoms.

What type of exercise works?

Most trials test aerobic exercise (walking, cycling, swimming), but resistance training—lifting weights or using resistance bands—also shows consistent benefits. A 2018 meta-analysis of 33 RCTs (1877 participants) focused exclusively on resistance exercise training (RET) and found a moderate effect size of -0.66 for reducing depressive symptoms [PMID 29800984]. Importantly, the antidepressant effect of RET did not depend on whether participants actually got stronger, suggesting the mood benefits aren’t simply downstream of physical fitness gains. The number needed to treat was 4—meaning for every four people doing resistance training, one would experience a clinically meaningful reduction in depression who wouldn’t have improved otherwise.

A 2026 network meta-analysis of post-stroke depression trials found that resistance training was the single most effective modality, though mixed aerobic exercise, combined exercise, high-intensity interval training, moderate-intensity continuous training, traditional martial arts, and yoga all reduced depressive symptoms [PMID 42812080]. Interestingly, exercise interventions lasting less than 8 weeks, done at least five times per week, and with sessions of 30 minutes or less showed the most significant benefit for post-stroke depression—challenging the assumption that more is always better.

Exercise in special populations

A 2021 systematic review and network meta-analysis examined interventions for depression in people with dementia (28,483 participants across 256 studies) [PMID 33762262]. Seven non-drug interventions significantly reduced depressive symptoms compared to usual care, including cognitive stimulation, multidisciplinary care, occupational therapy, and—relevant here—exercise combined with social interaction and cognitive stimulation (effect size roughly -12 points on the Cornell scale). Massage, cognitive stimulation combined with cholinesterase inhibitors, and exercise combinations were more effective than many drug interventions. This underscores that exercise may be particularly valuable in populations where medication options are limited or carry higher risk.

How it might work — biological mechanisms

Why would moving your body improve your mood? Depression isn’t just psychological—it involves measurable changes in brain chemistry, inflammation, stress hormones, and neural plasticity. Exercise appears to target multiple pathways simultaneously.

Neurotransmitter systems

Exercise increases availability of serotonin, norepinephrine, and dopamine—the same neurotransmitters targeted by most antidepressants [PMID 25384672]. Acute exercise also triggers endorphin release, though the mood benefits extend well beyond the immediate post-workout “high.” Animal studies show that regular exercise upregulates brain-derived neurotrophic factor (BDNF), a protein essential for neuronal growth and synaptic plasticity, particularly in the hippocampus, a brain region often smaller in people with depression.

Anti-inflammatory and antioxidant effects

Chronic low-grade inflammation is increasingly recognized as a contributor to depression. Moderate exercise has anti-inflammatory effects, reducing circulating levels of pro-inflammatory cytokines like IL-6 and TNF-alpha. Exercise also boosts antioxidant defenses, which may protect neurons from oxidative stress [PMID 25384672]. (The saffron study included in the original search set discusses similar mechanisms for a botanical antidepressant, highlighting that multiple interventions can share overlapping biological pathways.)

Neuroendocrine regulation

Depression often involves dysregulation of the hypothalamic-pituitary-adrenal (HPA) axis, the body’s central stress response system. Regular exercise appears to recalibrate HPA axis function, normalizing cortisol rhythms and improving stress resilience [PMID 25384672]. This may explain why exercise not only alleviates current symptoms but also reduces relapse risk.

Psychological and social mechanisms

Beyond biology, exercise provides structure, mastery experiences, and—especially in group settings—social connection. The combination of physical activity with social interaction may explain why group exercise or exercise embedded in multicomponent programs (like exercise plus cognitive stimulation) often outperforms solo physical activity [PMID 33762262].

Where the evidence is strong (and weak)

Strengths

The evidence base is now substantial. Multiple independent meta-analyses, drawing on dozens of RCTs and thousands of participants, consistently show moderate-to-large antidepressant effects of exercise. The benefits appear robust across different types of exercise (aerobic, resistance, mixed), different populations (young and old, medically healthy and chronically ill), and different baseline severities. Effect sizes rival or exceed those of antidepressant medications in head-to-head and placebo-controlled comparisons. The safety profile is favorable—serious adverse events are rare, especially compared to pharmacotherapy.

Weaknesses and gaps

First, control group responses are unusually large in exercise trials. One meta-analysis found control participants improved by an average of 7.5 points on the Hamilton Depression Rating Scale—roughly double the placebo response in antidepressant drug trials [PMID 26707338]. This suggests that simply enrolling in a study, receiving attention, and expecting benefit may account for a substantial portion of observed improvement. Exercise must overcome this large background effect to demonstrate superiority, which makes the net benefit harder to isolate.

Second, follow-up data are weak. The 2016 meta-analysis found a large effect during active treatment but only a small, nonsignificant effect at follow-up (effect size -0.22) [PMID 27253219]. This raises questions about durability: do people maintain exercise habits after trials end? Do benefits fade without continued activity?

Third, heterogeneity is high. Many meta-analyses report substantial variation across studies (I² values of 70–76%), meaning individual trial results differ more than you’d expect by chance alone [PMID 34565591] [PMID 29800984]. Sources of heterogeneity aren’t always clear. Resistance training effects didn’t vary by exercise volume, participant health status, or strength gains [PMID 29800984], and while better-designed trials (those with blinded allocation and assessment) tended to show smaller effects, significant benefits persisted even in high-quality studies.

Fourth, optimal dose remains uncertain. The post-stroke depression network meta-analysis suggested that shorter, more frequent sessions might be best [PMID 42812080], but other reviews found no clear dose-response relationship for resistance training [PMID 29800984]. There’s no universal prescription yet—just evidence that various plausible regimens work.

Fifth, few trials directly compare exercise to gold-standard psychotherapy (e.g., cognitive-behavioral therapy). One meta-analysis found small, nonsignificant differences when exercise was compared to psychological treatments [PMID 27253219], but the number of such comparisons is limited.

Practical takeaways

For individuals considering exercise for depression:

  • Exercise is a legitimate, evidence-based intervention for depression, not just a feel-good suggestion. Moderate-to-large benefits are seen across multiple trial types.
  • Both aerobic exercise and resistance training work. Choose what you’ll actually do consistently. The post-stroke depression data suggest shorter, frequent sessions (e.g., 30 minutes, five or more times per week) may be as good or better than longer, less frequent workouts [PMID 42812080].
  • If you’re already on medication or in therapy, adding exercise may enhance your response, especially if your symptoms are moderate to severe [PMID 34565591].
  • Group-based or socially embedded exercise (classes, walking groups, team sports) may offer additional benefits beyond solo activity [PMID 33762262].

For clinicians:

  • Exercise deserves a place in treatment algorithms, particularly for patients who prefer non-pharmacological options, have contraindications to medication, or haven’t responded adequately to first-line treatments.
  • Resistance training is underutilized. It’s as effective as aerobic exercise for depression [PMID 29800984] and offers the added benefit of preserving muscle mass and bone density, especially important in older adults.
  • Don’t expect exercise to work overnight. Most trials run 8–12 weeks. Setting realistic timelines and providing structure (e.g., referral to a structured program, written prescription) may improve adherence.
  • Monitor adherence. The durability of benefit depends on continued activity [PMID 27253219]. Follow-up and relapse prevention planning are critical.

For researchers and policymakers:

  • Head-to-head trials comparing exercise to CBT, and trials testing exercise plus psychotherapy combinations, are needed.
  • Pragmatic trials in real-world settings (primary care, community centers) would clarify feasibility and cost-effectiveness.
  • Investigations into mechanisms (biomarker studies, neuroimaging) could identify which patients are most likely to respond and guide personalized prescriptions.

Limitations of this review

This synthesis draws on meta-analyses rather than individual RCTs, inheriting any biases present in the original reviews. Different meta-analyses used different inclusion criteria, search dates, and quality assessments, limiting direct comparability. Most included trials were conducted in high-income countries with predominantly White, middle-aged samples; generalizability to other populations is uncertain. Publication bias—the tendency for positive findings to be published more readily than null findings—likely affects the exercise literature, though funnel plot analyses in some meta-analyses suggest it’s not severe. Finally, exercise is heterogeneous: “exercise” encompasses everything from tai chi to CrossFit, delivered individually or in groups, supervised or unsupervised. The evidence summarized here represents an average across diverse interventions.

Bottom line

Exercise is an effective treatment for depression, with benefits comparable to first-line antidepressants and psychotherapy. Both aerobic and resistance exercise work, and adding exercise to standard treatments enhances outcomes, particularly in patients with more severe symptoms. The biological mechanisms are plausible and well-supported—exercise modulates neurotransmitters, reduces inflammation, normalizes stress hormones, and promotes brain plasticity. While questions remain about optimal dose, long-term adherence, and which patients benefit most, the evidence is strong enough to warrant routine consideration in clinical practice. For many patients, exercise isn’t just an adjunct—it’s a first-line option that deserves the same serious consideration as medication or therapy.

FAQ: Exercise and Depression Treatment

Does exercise work as well as antidepressant medication for depression?

Based on limited direct comparisons, exercise appears to perform similarly to antidepressants. Meta-analyses show the difference between exercise and medication is statistically nonsignificant (effect size -0.08, PMID 27253219). When exercise is compared to no treatment, it shows large effects (effect size -1.24), comparable to what antidepressants demonstrate in their own trials. However, head-to-head studies remain rare, so while current evidence suggests equivalence for many patients, more direct comparisons are needed before definitive claims can be made.

How much exercise is needed to reduce depressive symptoms?

The evidence doesn’t yet identify a precise minimum dose, but meta-analyses consistently show moderate-to-large effects across various exercise protocols. Studies have tested resistance training, aerobic exercise, and combination programs with positive results across different durations and intensities. What matters most appears to be adherence—completing a structured program consistently—rather than a specific type or intensity threshold. Both vigorous and moderate-intensity exercise have shown benefits, suggesting flexibility in prescription to match individual preferences and capabilities.

Can exercise help depression if I’m already taking antidepressants?

Yes. Exercise as add-on therapy provides additional benefit beyond standard treatment alone. A 2021 systematic review of 22 studies found moderate pooled effects when exercise was combined with usual care, which typically included medication or psychotherapy. This suggests exercise complements rather than simply duplicates the mechanisms of antidepressants, potentially addressing different pathways in depression. Combining treatments may be particularly valuable for patients with partial response to medication alone.

What type of exercise works best for depression—cardio or strength training?

Both aerobic exercise and resistance training show comparable effectiveness. Meta-analyses demonstrate moderate-to-large effects for resistance training, aerobic work, and combination programs alike. This equivalence is clinically valuable because it allows personalization: patients can choose activities they enjoy and will sustain long-term. The consistency of benefits across exercise types suggests that movement itself—whether running, cycling, lifting weights, or mixed routines—drives antidepressant effects, rather than any single modality being uniquely therapeutic.

Does exercise work for depression in older adults and people with other health conditions?

Evidence confirms exercise benefits extend across diverse populations. Meta-analyses have demonstrated effectiveness in older adults with mild depressive symptoms, stroke survivors with post-stroke depression, and people with dementia. These findings are particularly significant because older adults and those with comorbid medical conditions often face medication contraindications or side effects. Exercise offers a lower-risk intervention option for vulnerable populations, though programs should be adapted to physical capabilities and supervised appropriately for safety.

Studies referenced

  • PMID 27253219 — Meta-analysis of 23 RCTs (977 participants) showing moderate-to-large effects of exercise on depression compared to control conditions, with smaller nonsignificant effects when compared to antidepressants or psychotherapy.
  • PMID 29800984 — Meta-analysis of 33 RCTs (1877 participants) demonstrating that resistance exercise training significantly reduces depressive symptoms with a moderate effect size, independent of strength improvements.
  • PMID 33762262 — Systematic review and network meta-analysis of 256 studies (28,483 people with dementia) finding that non-drug interventions, including exercise combined with cognitive and social components, were more efficacious than drug interventions for reducing depressive symptoms.
  • PMID 34565591 — Systematic review and meta-analysis of 22 studies showing moderate pooled effect favoring exercise combined with standard treatment over standard treatment alone, with greatest benefits in patients with severe depression.
  • PMID 25384672 — Systematic review of saffron for depression (included for mechanistic context), detailing serotonergic, antioxidant, anti-inflammatory, and neuroprotective mechanisms also relevant to exercise.
  • PMID 42812080 — Network meta-analysis of exercise interventions for post-stroke depression, finding that resistance training was the most effective single modality and that shorter, more frequent sessions yielded the greatest symptom improvement.
  • PMID 26707338 — Systematic review and meta-analysis of control group responses in 41 exercise RCTs (1122 participants), revealing large placebo-like improvements (effect size -0.92) that exercise interventions must overcome to demonstrate efficacy.
  • PMID 35797521 — Meta-analysis of 13 antidepressant pharmacotherapy trials in cardiac patients (2207 participants), finding only small effects (effect size 0.17) of antidepressants versus placebo, smaller than typical exercise effects in similar populations.

Source

  • PMID: 27253219 (read full paper on PubMed)
  • Journal: Journal of affective disorders (2016)

Articles on this site are adapted from PubMed abstracts as general-interest explainers. They are not intended as medical advice.

📝 This article was adapted by the PubMed Trivia editorial team using Claude (Anthropic) as an adaptation assistant, based on the PubMed abstract cited above. Our editorial workflow includes automated NG-term checks (medical-claims law), careful "correlation vs causation" distinction, and mandatory PMID + source URL disclosure. See the full pipeline on our editorial policy and editor pages. Please report errors to sciencepubmedjp@gmail.com (we respond within 48 hours).