Exercise for Depression — Body Handbook

Psychology · §454

Exercise for Depression

For mild-to-moderate depression, the trial evidence puts exercise on roughly the same shelf as antidepressants and structured talk therapy — not as a wellness add-on, as a first-line treatment with a pooled effect across more than two hundred randomised trials. The catch is dose: under-prescribed, sporadic exercise doesn't deliver what supervised, progressive, three-to-five-times-a-week training does.

Do · Weekly Evidence Moderate Chapter Psychology

You're looking at one of the more evidence-backed treatments in the manual, with a side-benefit profile no antidepressant touches: better sleep, sharper thinking, more daily energy, a meaningfully lower risk of dying early. The catch — and it's a real one — is that depression makes the exact behaviour that helps feel impossible. Starting matters more than starting big.

Three things happen when you train that plausibly explain why mood follows. First, your brain produces more brain-derived neurotrophic factor. After a single hard session it surges; over weeks of training the baseline rises Szuhany et al. 2015. Second, the hippocampus — a memory and mood structure that shrinks in depression — grows back. In a one-year trial of older adults, aerobic walking added about two percent of volume to the anterior hippocampus while a stretching group lost ground Erickson et al. 2011. Third, the stress system itself recalibrates: the cortisol-spiking machinery that runs hot in depression cools down with regular training, sleep deepens, and chronic low-grade inflammation drops a notch Mikkelsen et al. 2017.

None of these alone has been proven the cause of mood improvement. The convergence is the point — it is hard to find a depression-relevant biological system that exercise doesn't touch.

What the trial picture actually says

The single largest synthesis is a 2024 BMJ network meta-analysis: 218 randomised trials covering 14,170 patients with diagnosed depression. Walking or jogging produced a standardised mean difference of -0.62 against usual care; yoga -0.55; strength training -0.49; mixed aerobic exercise -0.42. Higher intensities helped more. Both sexes, all adult ages, and people with chronic medical comorbidities benefited Noetel et al. 2024. For context, the same yardstick puts the typical antidepressant at around -0.30 against placebo across more than five hundred trials Cipriani et al. 2018.

The honest counter-argument: the strictest Cochrane review shrank the apparent effect to a small -0.18 once the analysis was restricted to the highest-quality trials, and a 2017 trial-sequential analysis judged the picture still inconclusive given small studies and unblinded designs Cooney et al. 2013, Krogh et al. 2017. Exercise trials cannot be double-blinded — participants know whether they are exercising. That ceiling on study quality won't move.

What does move is the convergence across multiple modern meta-analyses, the cohort-level prevention effect across more than two million person-years in the general population Pearce et al. 2022, and the head-to-head equivalence with drugs and therapy at twelve to sixteen weeks. The reasonable summary: a moderate effect at the conservative end, a large one at the optimistic end, and a first-line treatment in either case.

What you lose by not addressing this

You may be tracking your depression as an inner-life problem. The literature tracks it as also a body problem — sedentary depressed adults accumulate cardiovascular, metabolic, and cognitive decline faster than active depressed adults, and the years compound. Meeting standard activity guidelines is associated with roughly a quarter lower risk of developing depression in the first place; even half-dose activity confers about an eighteen percent reduction Pearce et al. 2022.

The flip side: the version of you that stays mostly still through this episode loses ground twice. Once on the depression itself, which has a high recurrence rate without active maintenance. Once on the cardiovascular and cognitive trajectory that sets the next two decades — the same trajectory your future partner, kids, or coworkers will be navigating with you. The Friday you didn't move for the seventh week in a row isn't loud about anything. The decade is.

The dose that moves the needle

The dose with the trial-grade effect is the same one general health guidelines use: 150 minutes of moderate aerobic activity per week, or 75 of vigorous, or some mix — split across three to five sessions WHO 2020. In the DOSE trial of mild-to-moderate depression, that public-health dose cut symptom scores by 47 percent, more than double a half-dose comparison; doing the same weekly minutes in three sessions versus five made no difference Dunn et al. 2005. Across modern meta-analyses, higher intensity at matched volume beats lower intensity Heissel et al. 2023. Strength training works on its own as well — about as well as aerobic — by a thirty-three-trial meta-analysis, independent of how much strength you actually gain Gordon et al. 2018.

The cheapest version is a walk you can do without changing clothes. Take that one. The trial evidence doesn't reward expensive equipment; it rewards minutes accumulated at adequate intensity, week after week.

Who should clear it with a clinician first

The narrow medical contraindications: a known cardiac condition that hasn't been cleared for moderate-to-vigorous activity, uncontrolled high blood pressure, a recent cardiac event, severe valvular disease. None of these put activity off-limits — they put unsupervised intensity off-limits. Your clinician sets the safe ceiling, and most of the trial-grade benefit lives at intensities a properly cleared patient can reach.

Why "I tried exercise and it didn't help" usually has a specific cause

Most people who say exercise didn't work for their depression didn't fail at exercise — they failed at dose. Three patterns drive most of the trial-to-life gap.

Intensity too low. The trials with the largest effects use heart rates in the seventy to eighty-five percent range, not casual strolls. The gentler you go, the further your dose drifts from what the literature is actually studying.
Frequency too sporadic. One session a week doesn't accumulate. Three is the floor; four or five is where the trial protocols sit.
No progression. Doing the same easy session for months produces less and less signal. Bump the difficulty every couple of weeks — faster, hillier, heavier, longer, in that rough order.

The fourth and largest failure mode is on the initiation side. The trials that work are supervised, scheduled, social, and accountability-structured. Self-prescribed exercise in a depressed person without those scaffolds quietly disappears within three weeks. If you have access to a gym buddy, a group class, a trainer, or a structured app — use it. The cost of the structure is worth less than the value of actually doing the thing. Stubbs et al. 2016 found about an eighteen percent dropout rate even inside supervised trials of depressed patients; outside trials, without the scaffolds, the rate is higher and unmeasured.

Two stories that keep getting in the way

The first is the wellness-influencer version: that exercise is a feel-good adjunct, a mood booster you might add on top of real treatment. The trial evidence puts it on the same shelf as the SSRIs and CBT, not below them Noetel et al. 2024. If a pill produced these effect sizes it would be the standard of care.

The second is its mirror image — the "just go for a walk" brush-off that depressed people get from non-depressed people, which lands as invalidating because it skips over the activation problem that is half the disease. Both versions miss the same thing: the studied dose is specific, the effect at that dose is real, and getting yourself to the studied dose when depression has shrunk your activation is the actual problem to solve. Treat it like a clinical intervention, not like advice.

Older adults, and people who haven't responded to medication

Adults over sixty-five with depression have a slightly different cost-benefit calculus. Antidepressants in this group carry higher fall risk, more drug interactions, and slower response, and the trial evidence here is among the cleanest in the literature. The original SMILE trial established aerobic equivalence with sertraline at this age Blumenthal et al. 1999, and the SEEDS trial found that adding aerobic exercise to sertraline outperformed sertraline alone, with the biggest gains on depressed mood and psychomotor slowing Belvederi Murri et al. 2015. For this group the question is less "exercise instead of medication" and more "how do I add the exercise that reduces my medication need."

If you've already tried two or more antidepressants without remission, the augmentation evidence is also worth knowing. In treatment-resistant patients on stable medication, adding moderate exercise five times a week produced twenty-one percent remission versus zero in medication alone over twelve weeks Mota-Pereira et al. 2011. In SSRI non-remitters, a supervised public-health dose added on top of unchanged medication produced a meaningful additional remission rate at a number-needed-to-treat under eight Trivedi et al. 2011. Both findings sit in small trials; both point the same direction. If medication has only half-worked, EPA-rich fish oil is another inexpensive, drug-free add-on some pair with the exercise at this stage.

How it sits against medication and therapy

The first-line alternatives are antidepressant medication — usually an SSRI or SNRI — and structured psychotherapy, primarily cognitive behavioural therapy, behavioural activation, or interpersonal therapy. At sixteen weeks the head-to-head trials don't separate the three; you'd pick on side-effect tolerance, access, and what your future self can sustain.

Exercise's distinctive advantages: no pharmacological side-effect burden, broader collateral benefits across cardiovascular fitness, sleep, body composition, and cognitive function, and a maintenance edge if you keep it up Babyak et al. 2000. Its disadvantages: the activation cost is highest in the population that needs it most, and it only delivers if you actually reach the studied dose. None of these are mutually exclusive — combining exercise with medication, or with therapy, or with both, is the strongest play for many people, and was the explicit design of the augmentation trials.

What changes over weeks, months, and a year

In a typical twelve-to-sixteen-week trial, mood improvement starts emerging between weeks two and four, deepens through weeks eight to twelve, and reaches full response by weeks twelve to sixteen Dunn et al. 2005, Blumenthal et al. 2007. Sleep is usually the first thing to shift. Within the first two to four weeks of consistent training, people fall asleep faster and sleep deeper; the trial literature shows reliable improvements in sleep onset latency, sleep efficiency, and slow-wave sleep that compound with the mood improvement rather than waiting for it Mikkelsen et al. 2017.

Daily energy moves in roughly the same window. The fatigue that made the afternoon feel like a wall lifts in stages — first because the cardiovascular adaptation lowers the effort cost of normal movement, then because the sleep is better, and finally because the mood lift takes some of the weight off. Resistance training reduces fatigue independent of any strength gain, which means the "I'm too tired to lift" loop breaks before the muscle does anything visible Gordon et al. 2018.

By month three, the secondary changes start showing up to people who aren't you. Posture shifts. You climb the stairs without thinking about it. The colleague who hadn't seen you in a quarter says you look different and can't say why. Cognitive performance also improves in this window — sharper attention, better working memory, faster decision-making — and in older adults a year of aerobic training adds measurable volume to the hippocampus, the memory and mood structure that shrinks in depression Erickson et al. 2011. The body-composition and skin-tone changes that come with steady training are slow but real over six-to-twelve months; the version of you a year in usually looks healthier in photographs without anyone being able to name what changed.

At a year out, if you're still going, the maintenance edge kicks in for mood. The SMILE ten-month follow-up tracked an eight percent relapse rate in the participants who kept exercising versus thirty-eight percent in those who stopped — a separation no antidepressant has shown without continued dosing Babyak et al. 2000. Five and ten years out, the cardiovascular and dementia-prevention literature compounds in your favour: meeting standard activity guidelines associates with roughly a twenty-to-thirty percent lower risk of dying early, with the same direction of effect across heart disease, type 2 diabetes, several cancers, and late-life dementia WHO 2020.

The most honest version of the long-run forecast: depression episodes still happen. Exercise is not a vaccine, and the recurrence rate of major depression is what it is. But the next episode tends to land softer, the floor underneath you is higher, and the recovery is faster, in the person who has kept a meaningful baseline of activity through adult life.

Adjacent topics worth knowing about: structured psychotherapy (the other behavioural-track first-line treatment), SSRI antidepressants (when medication belongs in the picture and when it doesn't), sleep and circadian alignment (whose mood effects compound with exercise's), and morning sunlight exposure (a small daily action that lifts the same systems). If exercise alone isn't enough, the live question is usually whether to add medication, add therapy, or both — not whether to drop the exercise.

Related in the handbook

Helps

— Resistance training carries an antidepressant-sized effect of its own, a strong option for women alongside cardio.
— Walking is one of the simplest forms the antidepressant effect of exercise can take.
— The dose that helps depression is mostly easy, conversational cardio — zone 2 is the unhurried way to get it.
— For depression only half-responding to medication, EPA-rich fish oil is a reasonable add-on to pair with exercise.

Alternatives

— Exercise and inner work are the two non-drug mood levers with the strongest evidence — they stack well.
— For mild-to-moderate depression, exercise is the free first move; ketamine is for the treatment-resistant end.
— Like exercise, meditation matches a starter antidepressant for many people — two drug-free levers on the same problem.
— Exercise works for depression on a par with therapy; many people do both, and the combination beats either alone.
— For the winter, seasonal version of low mood, morning bright light is the more specific tool alongside exercise.
— Exercise is for everyday depression; TMS is the escalation when pills and lifestyle haven't shifted a treatment-resistant case.

— Resistance training, not just cardio, carries an antidepressant-sized effect on mood.
— Cold plunges give a short dopamine-and-noradrenaline mood bump; the evidence is thinner than for exercise, so treat it as an add-on.
— Already lifting for your mood? A daily scoop of creatine adds a small extra antidepressant signal on top — cheap and worth stacking.
— Men with erectile dysfunction get depressed about three times as often; the exercise you'd do for mood treats the ED underneath it too.
— Part of why exercise lifts mood may run through the gut-brain axis, alongside its direct brain effects.
— For the depression that rides along with painful chronic disease like HS, exercise is a real lever.
— For men at risk, treating depression matters — and exercise is a drug-free, accessible place to start.
— For the monthly mood drop of PMDD, regular exercise is a real, drug-free piece of the plan.
— One of the few mood levers on par with exercise: regular time with people you're close to. Stack the two.
— Exercise for mood works better outdoors. Two hours a week in real green space adds its own measurable lift.
— Aerobic training is the engine behind the antidepressant effect, and VO2 max is what it's building.

1. Substance and claimed effects

The substance is structured physical exercise — predominantly aerobic (walking, jogging, cycling) and resistance training — prescribed as an intervention for unipolar depressive symptoms ranging from subclinical to major depressive disorder (MDD). The defining claim is that exercise reduces depressive symptom severity to a degree comparable with first-line pharmacotherapy and psychotherapy, with replicated effects across populations and modalities Noetel et al. 2024, Singh et al. 2023. Adjacent consequences relevant to scoring include broad mortality reduction, sleep architecture improvement, cognitive performance (especially executive function and hippocampal-dependent memory), daily energy / vitality, body composition and cardiovascular fitness with downstream visible effects, and prevention of incident depression in non-depressed populations Pearce et al. 2022, Schuch et al. 2018, Erickson et al. 2011. The substance is non-prescription, low-cost, and effectively dose-titratable; the central practical question is therapeutic dose, not access.

2. Evidence by addressing question

Mechanism

Several biological pathways converge plausibly. Acute and chronic aerobic exercise raises peripheral brain-derived neurotrophic factor (BDNF), a neurotrophin implicated in hippocampal neurogenesis and synaptic plasticity; the meta-analytic effect on serum BDNF is large after single bouts (g=0.46) and meaningful with repeated training, with attenuated magnitude in females Szuhany et al. 2015. Chronic aerobic training in older adults increases anterior hippocampal volume by ~2% over one year, reversing typical age-related atrophy and tracking improved memory performance — a hippocampal-volume increment of clinical interest because the hippocampus is structurally smaller in MDD Erickson et al. 2011. Additional plausible mechanisms include normalization of HPA-axis hyperactivity, reduction in chronic low-grade inflammation (CRP, IL-6, TNF-α), improved vagal tone and heart-rate variability, increased monoamine availability acutely, and downstream effects via improved sleep, body composition, and self-efficacy Mikkelsen et al. 2017. None of these mechanisms is decisively isolated as the cause of mood improvement, but the redundancy of the mechanistic case makes "no plausible pathway" untenable.

Evidence

The most recent and largest synthesis is a 2024 BMJ network meta-analysis of 218 RCTs covering 14,170 participants with clinical depression diagnoses. Compared with usual care, walking/jogging produced a standardised mean difference (SMD) of -0.62, yoga -0.55, strength training -0.49, mixed aerobic exercise -0.42, and tai chi/qigong -0.42. Dance produced the largest point estimate (-0.96) but on far fewer trials. Effects were larger at higher intensities and held in adults of both sexes and across ages, including with comorbid conditions Noetel et al. 2024. A 2023 BJSM umbrella review of 97 prior meta-analyses (over 1,000 RCTs, ~128,000 participants) reported a pooled median SMD of -0.43 for depression, with higher-intensity activity yielding larger effects and shorter (~12-week) programs outperforming longer ones Singh et al. 2023. A 2023 BJSM meta-analysis restricted to clinical depression diagnoses reported still larger pooled effects (overall SMD ~-0.95; strength training alone ~-1.04; combined aerobic + strength ~-0.46) Heissel et al. 2023. A meta-analysis adjusted for publication bias reported an overall SMD of ~-1.11 against non-active controls in MDD, attenuating but remaining significant after trim-and-fill Schuch et al. 2016. Resistance training has its own dedicated meta-analysis (33 RCTs, 1,877 participants): pooled SMD -0.66, independent of strength gains, total dose, or baseline depression status Gordon et al. 2018.

Head-to-head comparisons with first-line treatments are limited in number but consistent in direction. In the SMILE trial (n=156 MDD adults over 50), 16 weeks of supervised aerobic exercise (three 45-minute sessions weekly at 70–85% max heart rate) matched sertraline's remission rate; the exercise arm showed slightly faster initial response with sertraline and equivalent endpoint outcomes Blumenthal et al. 1999. Ten-month follow-up found that participants who continued exercising had lower relapse than the medication arm Babyak et al. 2000. SMILE-II (n=202) confirmed equivalence of supervised exercise, home-based exercise, and sertraline against placebo over 16 weeks, all three active arms producing similar remission rates above placebo Blumenthal et al. 2007. In Cochrane's 2013 pooled analysis (39 trials, 2,326 participants), exercise outperformed control at SMD -0.62; head-to-head, exercise was statistically indistinguishable from antidepressants (SMD -0.11) and from psychological therapy (SMD -0.03); when limited to the highest-quality trials, the exercise-vs-control effect shrank to a small SMD of ~-0.18 Cooney et al. 2013. The 2017 BMJ Open update applying trial sequential analysis concluded that the evidence remained inconclusive given small, high-risk-of-bias trials, with the most rigorous estimates compatible with a clinically trivial effect Krogh et al. 2017.

For context: in the largest network meta-analysis of antidepressants (522 trials, ~117,000 patients), the pooled SMD of antidepressants vs placebo for acute MDD treatment was ~-0.30 (odds ratio 1.66 for response) — i.e., exercise's pooled effect size in the 2024 BMJ review and in publication-bias-adjusted meta-analyses meets or exceeds typical drug-vs-placebo effects, though drug trials are uniformly blinded whereas exercise trials cannot be Cipriani et al. 2018. Pooled effect sizes for evidence-based psychotherapy vs control hover around SMD -0.7 against waitlist and ~-0.3 to -0.5 against active controls Cuijpers et al. 2023.

Exercise is also studied as augmentation. In TREAD (n=126 SSRI non-remitters), 12 weeks of supervised aerobic exercise at the "public health dose" (16 kcal/kg/week ≈ 150 min/week brisk walking) produced 28.3% remission vs 15.5% in a low-dose comparison arm (NNT ≈ 7.8) Trivedi et al. 2011. In treatment-resistant patients on stable adequate pharmacotherapy, adding 30–45 minutes of moderate exercise 5×/week produced 21% remission and 26% response vs 0% / 0% in pharmacotherapy alone at 12 weeks Mota-Pereira et al. 2011. SEEDS, in late-life depression, found that adding aerobic exercise to sertraline outperformed sertraline alone on affective symptoms, particularly depressed mood and psychomotor retardation, with the progressive-aerobic group benefiting most Belvederi Murri et al. 2015.

Prevention is a separate evidence stream. A 2018 prospective-cohort meta-analysis (49 studies, ~267,000 participants, median 7.4-year follow-up) found that meeting physical activity recommendations was associated with a 17% reduction in incident depression, consistent across age, sex, and continent Schuch et al. 2018. A 2022 JAMA Psychiatry dose-response meta-analysis (15 cohorts, 2+ million person-years) found a curvilinear inverse relationship: half the WHO-recommended dose (~8.8 mMET-h/week, equivalent to ~75 min/week brisk walking) was associated with an 18% lower risk of depression; meeting the full guideline dose conferred a 25% reduction; further increases yielded marginal returns Pearce et al. 2022.

Protocol

Across trials, the dose range with replicated efficacy is roughly 150 minutes/week moderate-intensity or 75 minutes/week vigorous-intensity aerobic exercise — the WHO/ACSM standard WHO 2020. The DOSE study (n=80 mild-to-moderate MDD) directly tested this: a "public health dose" of 17.5 kcal/kg/week reduced Hamilton Depression Rating Scale scores by 47%, more than double the 30% reduction in a low-dose arm of 7 kcal/kg/week; frequency (3× vs 5× per week) did not matter at matched weekly volume Dunn et al. 2005. Heissel's meta-regression and the 2024 BMJ network meta-analysis both find greater effects at higher intensities; Singh's umbrella review concurs Singh et al. 2023, Noetel et al. 2024, Heissel et al. 2023. Resistance training shows an antidepressant effect across a broad range of doses, with effect size independent of total volume in Gordon's meta-regression Gordon et al. 2018. The practical clinical protocol — supervised when possible, 3–5×/week, 30–60 minutes per session, progressive in intensity, ≥8 weeks duration, ideally outdoor or social — appears to be the consensus operating prescription across trials.

Contraindications

For aerobic and resistance exercise at the moderate-vigorous intensities studied, true medical contraindications are narrow: unstable cardiac disease, acute decompensated heart failure, severe uncontrolled hypertension, recent cardiac event without medical clearance, and active eating-disorder pathology where exercise functions as a compensatory behaviour. Pregnancy is not a contraindication to moderate exercise but warrants tailoring. The dominant practical "contraindication" is severe melancholic or psychotic depression where motivation collapse and psychomotor retardation make the intervention infeasible without intensive support — not a safety contraindication but a feasibility one. Stubbs's meta-analysis of dropout in 40 exercise-depression RCTs found a pooled dropout of 18.1% in the exercise arms (vs higher in control), with no excess attrition relative to typical psychotherapy or pharmacotherapy trials — meaning the feasibility ceiling is real but not catastrophic Stubbs et al. 2016.

Misconceptions

Three patterns dominate the gap between literature and lay belief. First: "exercise improves mood" gets shrunken in popular framing to a feel-good wellness adjunct, obscuring the clinical magnitude — exercise's effect size on diagnosed MDD is in the same ballpark as antidepressants and psychotherapy, not a soft supplement Noetel et al. 2024, Cipriani et al. 2018. Second: the public-health dose is widely under-prescribed in clinical settings — clinicians who recommend exercise rarely specify intensity, frequency, or duration, while the trials with the largest effects use supervised, progressive, intensity-titrated protocols Heissel et al. 2023. Third: aerobic and resistance training are both effective; the lay default of "cardio is for depression" ignores Gordon's meta-analysis showing resistance training's SMD -0.66, independent of strength gain Gordon et al. 2018.

Audience

Effects generalize broadly. The 2024 BMJ network analysis found that men and women, younger and older adults, and people with chronic comorbidities all benefited; women may benefit slightly more from strength training and men from yoga, but interaction effects are small Noetel et al. 2024. Older adults are well-studied (SMILE, SEEDS, Belvederi Murri's work) and respond similarly, with the practical advantage of fewer pharmacotherapy options due to polypharmacy and adverse-event risk; aerobic exercise in late-life depression also confers cognitive and disability benefits beyond mood Blumenthal et al. 1999, Belvederi Murri et al. 2015. Adolescents and perinatal populations are studied in smaller bodies of evidence with consistent direction. The population where exercise has the weakest signal is severe melancholic or psychotic depression — not because it doesn't work, but because feasibility collapses Stubbs et al. 2016.

Alternatives

The competing first-line treatments are antidepressant medication (SSRIs, SNRIs) and structured psychotherapy (cognitive behavioural therapy, behavioural activation, interpersonal therapy). The most rigorous head-to-head trials show statistical equivalence at 12–16 weeks across exercise, sertraline, and CBT Blumenthal et al. 1999, Blumenthal et al. 2007, Hallgren et al. 2015. Exercise's distinguishing advantages are non-prescription access, broad consequence profile (cardiovascular, metabolic, cognitive, sleep, body composition), absent pharmacological side-effect burden, and lower long-term relapse when continued Babyak et al. 2000. Its disadvantages are activation cost (requires effort the depressed reader has least of), the need for adequate dose and intensity, and the absence of a delivery system as scalable as a prescription.

Failure modes

The trial-real-world gap is the main failure mode. Trials use supervised, dose-monitored, social-context protocols with adherence support; self-prescribed exercise in a depressed person without scaffolding typically fails on initiation rather than on efficacy. Heissel's meta-regression finds that shorter (<12-week) supervised interventions outperform longer ones, suggesting that the supervised structure may matter more than the duration Heissel et al. 2023. Dropout in exercise arms of trials is ~18%; in unsupported self-initiated exercise, real-world dropout is higher and unmeasured Stubbs et al. 2016. The most common pattern of failure: too low intensity, too sporadic frequency, no progression, no accountability structure. Exercise that doesn't reach a target intensity may not engage the BDNF and HPA-axis pathways most plausibly responsible for the antidepressant effect.

Practicalities

Cost is near-zero for walking, jogging, or bodyweight resistance training; modest for gym membership ($300–$1,200/year); moderate for personal training or supervised group programs ($1,000–$5,000/year). Time burden is the dominant cost: 150 minutes/week is ~22 minutes/day. Equipment requirements range from running shoes to a small home setup. Insurance coverage for exercise prescription is rare in the US but expanding under value-based care contracts; the UK NHS has commissioned exercise referral schemes under NICE guidance NICE 2022. The largest practical friction in depressed populations is initiation: the activation-energy cost is highest in the population that needs the intervention most.

Stakes

Untreated depression carries substantial direct mortality (suicide risk, all-cause mortality elevation), occupational and relational impairment, and downstream physical health degradation. Sedentary behaviour, independent of depression, is a major contributor to cardiovascular and metabolic mortality. The compound case for not addressing low physical activity in a depressed adult is therefore double — losses to both mental health and cardiovascular trajectory. The cohort evidence shows clear dose-response between physical activity and depression risk in the general population: people meeting WHO guidelines have ~25% lower depression risk; even half-dose activity confers ~18% reduction Pearce et al. 2022.

Payoff

Onset of antidepressant effect in trials emerges between weeks 2 and 4 with meaningful symptom reductions visible by week 8 and full response by weeks 12–16 in most successful protocols Dunn et al. 2005, Blumenthal et al. 2007. Sleep onset, energy, and subjective vitality often shift within the first 2–4 weeks of consistent training; mood lift typically lags slightly. Long-term: continued exercise after acute response confers protection against relapse — the SMILE 10-month follow-up found that exercise-arm participants who kept exercising had a relapse rate of 8% vs 38% in those who stopped Babyak et al. 2000. Beyond mood, the consequence chain includes cardiovascular fitness gains, sleep quality, cognitive performance (executive function, hippocampal-dependent memory), reduced inflammation, improved body composition, and ~20–30% mortality reductions seen in physical-activity cohort literature at maintained dose Erickson et al. 2011, Mikkelsen et al. 2017.

History

Recommendations for movement as a treatment for melancholy predate modern psychiatry — found in Hippocratic texts, Galenic medicine, and 19th-century moral-treatment-era asylum practice (gardening, walking, agricultural labour as therapy). Systematic RCT investigation began in the 1980s; Blumenthal's SMILE trials in the 1990s established the head-to-head equivalence finding that has anchored the modern evidence base Blumenthal et al. 1999. Major guideline bodies have incorporated exercise as a first-line or adjunctive recommendation: NICE recommends group-based exercise for less-severe depression as one of several first-line options NICE 2022; APA recommends exercise for adults with MDD particularly for those preferring alternatives to medication or psychotherapy APA 2019.

3. The credibility range

The optimist case

The 2024 BMJ network meta-analysis is one of the largest and most methodologically rigorous syntheses of any depression intervention ever conducted — 218 RCTs, 14,170 patients with diagnosed depression — and it reports moderate-to-large pooled effects across multiple modalities, with the largest effects at higher intensities Noetel et al. 2024. Three independent contemporary meta-analyses (Heissel 2023, Schuch 2016, Singh 2023) converge on similar magnitudes after attempting to correct for publication bias and trial heterogeneity. Head-to-head trials of meaningful size (SMILE, SMILE-II, REGASSA, SEEDS, TREAD) show statistical equivalence with sertraline and CBT — and superior maintenance at long-term follow-up Babyak et al. 2000. The mechanistic case is multi-pathway and biologically plausible: BDNF, hippocampal neurogenesis, HPA-axis normalization, inflammation, monoamines, sleep, self-efficacy. Cohort epidemiology in the general population confirms a dose-response prevention effect — physical activity is causally protective against incident depression across millions of person-years Pearce et al. 2022, Schuch et al. 2018. Resistance training has its own dedicated meta-analytic base with effects roughly equivalent to aerobic, broadening the modality menu Gordon et al. 2018. The intervention has no pharmacological side effects, confers wide-spectrum non-mood benefits, and is accessible at near-zero cost. The optimist's strong claim: exercise is a first-line depression treatment whose effect size is competitive with — and sometimes superior to — antidepressants and structured psychotherapy, with a far better adverse-event profile and substantial collateral benefits.

The skeptic case

Exercise trials cannot be double-blinded. Participants randomized to exercise know they are exercising; placebo-controlled comparisons are limited to active-comparator designs (stretching, attention controls), all of which leak expectancy. The Cochrane 2013 review found that the SMD shrank from -0.62 to ~-0.18 when restricted to the highest-quality trials with adequate allocation concealment, blinded outcome assessment, and intention-to-treat analysis — a small effect, possibly clinically trivial Cooney et al. 2013. The 2017 BMJ Open trial sequential analysis concluded the question was not settled given the predominance of small, high-risk-of-bias trials and stated that the most rigorous estimates were compatible with no clinically meaningful effect Krogh et al. 2017. The publication-bias-adjusted estimate from Schuch 2016 (~-0.66 after trim-and-fill) is consistent with substantial small-study and reporting bias inflating the literature Schuch et al. 2016. Head-to-head equivalence with antidepressants is based on small trials with limited power to detect non-inferiority. Adherence in real-world dissemination is far lower than in supervised trials; the trial-to-clinic translation gap may erode much of the effect. Cohort epidemiology cannot resolve reverse causation — depressed people exercise less because they are depressed, not the other way around. The skeptic's strong claim: the true effect of self-administered exercise on diagnosed MDD is small once trial bias is properly controlled, and most of the apparent advantage comes from expectancy and supervision effects that don't transfer to a depressed person told to "go for a walk."

The author's call

The evidence base is large, the mechanisms are plausible and multiple, the modality and dose are flexible, and the head-to-head equivalence with antidepressants — even at the conservative end of the effect-size range — is more than sufficient to make exercise a first-line option. The blinding problem is real but cannot be solved (this is true of psychotherapy too, which is uncontroversially first-line), and the cohort-level prevention effect is independent of any blinding issue in treatment trials. Most of the high-quality-trial attenuation observed in Cochrane 2013 reflects strict methodology penalising the older trial set; the 2024 BMJ network analysis with much higher-quality trial inclusion still finds moderate-to-large effects. The strongest practical caveat is dose: under-dosed, unsupervised, sporadic exercise is the modal real-world delivery, and that version likely does not achieve the effect sizes seen in trials. The honest framing is that exercise is a well-evidenced treatment whose effect ceiling depends sharply on whether the reader actually reaches the studied dose — meeting it likely yields antidepressant-class benefit; falling well short likely doesn't. Evidence: 4. Controversy: 2 — disagreement exists at the methodological margin (true effect size after rigorous bias control), but no serious camp now claims "exercise doesn't help depression."

4. Stakeholder and incentive map

Pro-exercise. Sports medicine and exercise physiology research communities; integrative-health and lifestyle-medicine advocates; public-health bodies (WHO, ACSM, USPSTF endorsements for physical activity broadly); clinicians frustrated with medication side-effect burden and partial-response rates; patients seeking non-pharmacological options. Generally low commercial incentive — no industry profits from a person exercising more — though gyms, fitness apps, and wearables have a tangential financial stake.
Skeptic / counter-incentive. Methodologists rightly concerned about blinding and bias inflation in unblindable trials; some psychiatry voices wary of overselling exercise to patients with severe melancholic depression who cannot initiate; pharmaceutical industry has indirect interest in framing antidepressants as the first-line standard. Some clinicians worry that an "exercise prescription" without supervision functions as a brush-off for the patient.
Cultural. Wellness influencers and self-improvement communities oversell exercise as a near-panacea, eroding clinical credibility through hyperbole. The "just go for a walk" trope from non-depressed people to depressed people is a community-level invalidation that has soured many depressed patients on exercise advice altogether.
Guideline bodies. NICE (UK), APA (US), and CANMAT (Canada) all recommend exercise as a treatment option, though typically positioned as adjunct or alternative rather than first-line equal to pharmacotherapy/psychotherapy — guideline conservatism reflects the blinding issue and trial-quality concerns more than the effect-size question.

5. Population variability

Sex. Effects generalize across sexes; modality preference may differ (some evidence of larger BDNF response in men, larger strength-training benefit in women's MDD), but pooled effect sizes are similar Szuhany et al. 2015, Noetel et al. 2024.
Age. Late-life depression responds well (SMILE, SEEDS), with added cognitive and physical-function benefits. Adolescents and young adults respond with smaller bodies of evidence but in the same direction. Mid-life adults (the modal MDD population) are well-represented in the trial base.
Severity. Strongest evidence is in mild-to-moderate depression. Severe MDD with melancholic or psychotic features: feasibility limits effect; supervised programs and combination with pharmacotherapy are required.
Treatment-resistant depression. Augmentation evidence (TREAD, Mota-Pereira) supports adding exercise to ongoing pharmacotherapy in non-remitters, with meaningful remission rates achievable.
Comorbidities. Cardiovascular disease, type 2 diabetes, obesity — exercise carries additional benefit and is generally safe with standard pre-participation screening. Eating-disorder history: caution; exercise can function as a compensatory behaviour.
Baseline activity. Largest absolute mood gains are in sedentary populations; well-trained athletes plateau on the curve, though resistance training adds modest benefit even at high baseline aerobic fitness Gordon et al. 2018.

6. Knowledge gaps

True double-blind comparisons of exercise to placebo are impossible by design — this ceiling on study quality won't change. The question of how much of the trial effect transfers to unsupervised self-prescription remains poorly quantified; pragmatic trials in primary-care prescription settings (REGASSA-type) are needed at larger scale Hallgren et al. 2015. The minimum effective dose for clinical depression — versus the WHO general-population recommendation — is not precisely characterized; Dunn's dose study remains the cleanest direct comparison and is small Dunn et al. 2005. The relative efficacy of modalities (aerobic vs resistance vs yoga vs mixed) in head-to-head terms is still indirect, derived from network analyses rather than direct comparisons. Mechanism — whether BDNF, neurogenesis, HPA-axis, inflammation, or self-efficacy is the dominant pathway — remains unresolved; it may be that the multiplicity of effects is itself the mechanism. Whether exercise added to optimized pharmacotherapy adds meaningfully to remission rates in non-resistant patients, beyond TREAD's narrow non-remitter population, is under-studied. Whether exercise can replace maintenance pharmacotherapy after acute response — a critical practical question given Babyak's finding of low relapse with continued exercise — has not been adequately tested.

7. Per-dimension evidence summary

mood. Primary outcome in the entire trial literature; pooled SMDs of -0.4 to -0.95 across major meta-analyses, head-to-head equivalence with SSRIs and CBT, cohort-level prevention with dose-response. Score driver.
energy. Improvement in subjective vitality and fatigue is a near-universal trial finding within 2–4 weeks; cardiovascular adaptation lifts daily energy floor over 8–12 weeks; resistance training reduces fatigue independent of strength gains Gordon et al. 2018.
focus. Aerobic exercise improves executive function and processing speed; hippocampal volume gains track memory improvements; chronic training improves attention and working memory across age ranges Erickson et al. 2011, Mikkelsen et al. 2017.
sleep. Aerobic and resistance training improve sleep onset latency, sleep efficiency, and slow-wave sleep; effects emerge over weeks of consistent training. Bidirectional benefit with mood improvement.
longevity. Physical activity is one of the most robust mortality predictors in cohort epidemiology — meeting WHO guidelines confers ~20–30% all-cause mortality reductions; cardiovascular disease, type 2 diabetes, several cancers, dementia all show consistent inverse associations.
health_short_term. Cardiovascular fitness, blood pressure, insulin sensitivity, body composition, musculoskeletal capacity — all improve over 8–16 weeks at studied doses. Felt impact on daily functioning is substantial.
beauty_direct. Skin perfusion / post-exercise flush is real but transient. Not a primary effect within days–weeks.
beauty_cumulative. Body composition change, postural improvement, cardiovascular skin tone, and reduced skin glycation via improved metabolic health all contribute to a meaningful long-term aesthetic shift over months to years.
cost_burden. Walking and bodyweight work are free; gym membership and equipment trivial-to-minor; supervised programs minor-to-moderate. The dominant cost is time, not money.
effort_burden. 150 min/week moderate intensity is ~22 min/day; sustained over months is substantial willpower in a depressed person whose baseline activation is impaired. Initiation is the hardest cost.
evidence. Hundreds of RCTs across decades, multiple high-quality meta-analyses, cohort epidemiology in millions, multi-pathway mechanism, guideline endorsement — 4. Capped below 5 by the unsolvable blinding problem.
controversy. No serious camp denies efficacy. Methodological disagreement on true effect size after bias correction; clinical disagreement on positioning (first-line equal vs adjunct). 2.

Scoping. The brief named "aerobic and resistance exercise as interventions for depressive symptoms, including dose ranges and effect sizes relative to pharmacological and psychotherapeutic treatments." The article covers all four — aerobic and resistance get explicit modality treatment in protocol and evidence; the 150 min/week dose and the DOSE-trial dose-response anchor the protocol section; the SMILE / SMILE-II / Cochrane equivalence with sertraline and CBT carry the comparison. No narrowing relative to brief.

Holistic scoring vs framing tension. The framing of this entry is depression-treatment-specific, but per spec §1a the meta scores are holistic against the substance — exercise's longevity, energy, focus, sleep, and short-term health effects are real and non-zero whether or not the article centres on them. Those dimensions get paragraph-level coverage in mechanism and payoff so the article tracks the meta. A future "Exercise" entry (general movement / fitness rather than depression-specific) would re-cover this ground at greater depth; this entry stays in the mental-health lane.

Beauty (direct) at 0. Initially considered 1 for the post-exercise flush, but the dimension targets days-to-weeks topical effects and the flush is hourly. The cleaner call is 0; cumulative beauty captures the real long-run aesthetic shift.
Mood at 4, not 5. SMILE/SMILE-II head-to-head equivalence with sertraline and the 2024 BMJ network analysis arguably support 5 ("transformative; on the level of an effective psychiatric intervention"), but the Cochrane 2013 high-quality-trial attenuation to SMD -0.18 and the Krogh 2017 trial-sequential analysis create enough uncertainty at the rigorous end that 5 felt over-claimed. 4 fits the honest middle of the credibility range.
Evidence capped at 4. Exercise trials are unblindable by design; that's a hard ceiling on study-quality grading, independent of the size of the literature. Five would require placebo-comparable blinding, which can't exist here.
Controversy at 2. No camp denies efficacy; the live disagreement is about effect-size magnitude after methodological correction. That's margin-level pushback, not foundational disagreement.
Effort burden at 3, not 4. 22 minutes a day doesn't meet the "an hour+ daily" anchor for 4, even though depression makes the willpower cost feel larger than the time cost would suggest. Captured the activation-cost honestly in the pitch and in failure-modes rather than inflating the score.

Category. Placed in mental rather than exercise because the entry's framing and most useful entry point is for someone thinking about depression treatment. A reader browsing Movement & Exercise will be looking for general fitness coverage; this entry is the mental-health-specific lens on exercise.

Contraindications. Included cardiac-condition, uncontrolled-hypertension, and eating-disorder-history. Eating-disorder history is the non-obvious one — included because exercise can quietly function as a compensatory behaviour and worsen the underlying problem; flagged in the warning callout. Pregnancy intentionally not included — moderate exercise during pregnancy is recommended, not contraindicated, with standard tailoring; mild adjustment, not exclusion.

Future-link candidates (unwritten or potentially un-IDed). The related field assumes entries for general exercise, depression (the condition), psychotherapy, and ssri-antidepressants; some or all may not yet exist. Editor to wire these once they land. The out-of-scope section also references sleep / circadian alignment and morning sunlight, which likely live in sleep and light categories.

Separate-entry candidates surfaced during writing.

Behavioural activation — distinct from generic CBT and has its own evidence base in depression; would be a natural sibling.
Treatment-resistant depression: augmentation strategies — the TREAD / Mota-Pereira / SEEDS augmentation findings touched here belong in a broader treatment-resistance entry.
Late-life depression — SMILE and SEEDS specifically address the older-adult cohort with a different cost-benefit calculus than mid-life MDD.

Excluded by design. Depression itself as a diagnostic entity, the choice between SSRI / SNRI / atypical antidepressants, specific psychotherapy modality comparisons, and the programming detail of general exercise (Z2 vs threshold, periodisation, etc.) — all out of scope for the depression-specific lens. Pointed at in out-of-scope rather than expanded here.