RightRep EN · PT Open the app
BLOG

Can You Train With Anxiety or Depression? The Heavy-Mind Barbell Audit: The BMJ Network Meta Where 218 Trials and 14,170 Clinically Depressed Participants Put Strength Training at g −0.49 and Walking/Jogging at −0.62 Against Active Controls With Effects Proportional to Prescribed Intensity, the Duke SMILE File Where 16 Weeks of Exercise Matched Sertraline at P=.67 While the Drug Ran the Faster First Mile, the 10-Month Tail Pricing Self-Directed Training at OR 0.49 Against Depression Diagnosis, Gordon's 33-Trial Resistance Ledger at Effect 0.66 With NNT 4, the Withdrawal Trap That Raises Depressive Symptoms Inside a Median 7 Days of Stopping, the Overtraining-Mood Confound, HUNT's 12% of Future Depression Preventable at One Weekly Hour of Any Intensity, Chekroud's 1.2-Million Dose File at 45 Minutes 3–5×/Week With More Not Always Better, and São Paulo's Megacity Carrying 19.9% Twelve-Month Anxiety — the Program That Starts Before Motivation Arrives Is the Treatment Column

By RightRep AI Assistant · · 22 min read

Anxiety and depression are the conditions this library has already opened from the treatment side — does exercise help anxiety and does exercise help depression are the general files, and this door answers the diagnosed lifter's version: what happens to a barbell program when the mind runs heavy on its own schedule. The scale is the largest this library carries: mental disorders rose from 80.8 million to 125.3 million disability-adjusted life-years between 1990 and 2019 — now 4.9% of all global DALYs and 14.6% of every year lived with disability on the planet — with age-standardised rates highest in Australasia, high-income North America and Tropical Latin America, Brazil's own band [1]; in São Paulo, the megacity survey's 5,037 face-to-face clinical interviews found ~30% of adults carrying a twelve-month disorder, anxiety the most common group at 19.9% against mood at 11% — prevalence "at greater levels than similar surveys conducted in other areas of the world" — and only one-third of serious cases in any treatment [16]. The default scripts fail in opposite directions. The scared script reads the diagnosis as a fragility verdict and waits for motivation to arrive before the first session — which inverts the actual causal order, because in depression motivation is downstream of action, not upstream of it; meanwhile every week of inactivity is not neutral but actively priced, since experimentally withdrawing training raises depressive symptoms and anxiety within a median of 7 days, worse once withdrawal passes two weeks [9]. The reckless script trains to outrun the mind — the barbell as the only anesthetic, volume stacked on broken sleep, program decisions made at 2 a.m. — and never notices that the overtraining syndrome is itself defined by mood disturbance and fatigue [10], so a baseline-disturbed lifter cannot feel the difference between life and overreach until performance has already fallen. The reframe that carries the whole file is the Duke trial's own ten-month ledger: the exercise group's protection against relapse came not from the supervised program but from what patients did on their own afterwards — self-directed training during follow-up priced the odds of a depression diagnosis at 0.49 [6]. Which puts the fences in order. First: the diagnosis is a clinician's. This file assumes anxiety or depression diagnosed and managed by a mental-health professional — including the bipolar screen that must precede any "training out of it" plan, because a first hypomanic swing under a self-prescribed program is a psychiatric event. Second: the red-flag list outranks every training decision. Suicidal thoughts, a depression that is worsening rather than easing, thoughts that feel foreign or losing touch with reality — these are same-day medical events, and the program pauses without apology. Third: the drug shelf is the prescriber's. All 21 antidepressants beat placebo across 522 trials and 116,477 patients [13] — exercise is a column alongside that shelf, never a replacement for it, and the decision to start, change or stop belongs to the prescriber, not to a good month in the gym. With the fences standing, the operational file is unusually complete for a brain condition: the permission question has a 218-trial network meta-analysis [2], two Duke head-to-heads against sertraline [5][7], a resistance-specific ledger with a number-needed-to-treat of four [4], and the dose question has both a million-person association file [12] and a prevention cohort with a population-level fraction [11] — while the two-way street (training changes mood, stopping training changes it back) gives this door its own subject: how to build a program that survives bad weeks, which is the only kind of program a heavy mind ever runs.

The short answer: yes — treatment-grade, with the effort paradox to manage

Strip the question down and it splits into three checkable parts. One: does regular training treat diagnosed depression and anxiety? Yes — across 97 systematic reviews covering 1,039 trials and 128,119 participants, physical activity carried medium effects on depression (−0.43), anxiety (−0.42) and psychological distress (−0.60), with the largest benefits in people already diagnosed with depression [3]; the BMJ network meta restricted to trials meeting clinical cut-offs for major depression puts the barbell's own shelf at g −0.49 against active controls (643 patients, 22 comparisons), behind only walking/jogging at −0.62 and yoga at −0.55 [2]. Two: does it match medication? At the acute mile, no — sertraline answers faster [5]; at month four, statistically yes — Hamilton and Beck scores showed no difference between exercise and medication arms (P=.67) [5]; and at month ten the ledger flips, with significantly lower relapse in the exercise group (p=.01) and continued self-directed training pricing diagnosis at OR 0.49 [6]. Three: can a single session make things worse? The symptom itself is the risk factor — but the pooled dropout across 40 exercise trials in depressed cohorts ran at 18.1%, lower than the control arms at OR 0.642 [14], so the feared "it will be too much" is measurably rarer than the same fear in untreated comparison groups. The short answer stands on the first two questions and spends the rest of this file on the third, because the third — the effort paradox, where the treatment requires the exact resource the disease depletes — is the one the diagnosed lifter actually lives.

The reframe: action precedes motivation — the program as behavioral activation

Every other file in this library gets to assume the reader can start when they feel like it. This one cannot, and that inversion is the door's real subject. Depression's core attack is on initiation: the disease degrades precisely the machinery a program needs to launch. The clinical answer is behavioral activation — schedule the behavior, let motivation follow — and the training version writes itself: the session is an appointment, not a feeling. The Duke follow-up files give this its dose-response spine: at one year, remission had risen from 46% to 66% among available participants, and the association between ongoing exercise and remission was curvilinear, concentrated entirely between 0 and 180 minutes a week [7] — the first three weekly hours carry essentially the whole relationship, exactly the territory where "I only have energy for a short session today" lives. The worst week is not a program failure but the program's target demographic: the 7-day withdrawal ledger says missing training is not neutral (depressive symptoms and anxiety measurably rise [9]), so the design goal inverts from the usual one — not "how do I maximize this block" but "what is the version of this session that survives the week I don't want to get out of bed." That is why the minimum viable session (below) is a prescription with a citation, not a concession.

The permission file: the network meta and the Duke head-to-heads

The crown is the 2024 BMJ network meta-analysis, and its scale is the point: 218 unique studies, 495 arms, 14,170 participants, every trial requiring participants to meet clinical cut-offs for major depression [2]. Against active controls, the modality shelf ranked walking or jogging first at Hedges' g −0.62 (95% CrI −0.80 to −0.45; 1,210 patients), yoga second at −0.55, strength training third at −0.49 (−0.69 to −0.29; 643 patients), mixed aerobic at −0.43 and tai chi at −0.42 — with three findings a formulator actually uses: the effects were proportional to the intensity prescribed, strength training and yoga were the most acceptable modalities (fewest dropouts), and the authors' conclusion places exercise "alongside psychotherapy and antidepressants as core treatments" [2]. The resistance-specific file is the lane's own: 33 randomized trials, 1,877 participants, mean effect 0.66 (95% CI 0.48–0.83), number needed to treat four — and the moderators are the honest part: total prescribed volume, health status and even whether strength actually improved did not change the antidepressant effect, which was blunted only by the methodological features (blinded allocation and assessment) that make trials harder, not by anything about the lifter [4]. Then the Duke head-to-heads. SMILE-1 randomized 156 adults aged ≥50 with major depression to exercise, sertraline or both: at 16 weeks no statistical difference on Hamilton or Beck scores, P=.67, with medication simply faster out of the gate [5]. The 10-month follow-up is the file's most quotable ledger: relapse significantly lower in the exercise group (p=.01), and among all completers, self-directed exercise during follow-up carried OR 0.49 against a depression diagnosis at the end (p=.0009) [6]. The four-arm replication (supervised exercise, home exercise, sertraline, placebo; 202 sedentary adults with MDD) confirmed the acute tie and added the tail: at one year, remission 46%→66%, initial treatment arm and follow-up medication use predicted nothing — the exercise done during follow-up predicted both Hamilton scores and diagnosis [7]. The anxiety side is thinner but points the same direction: 15 trials, 675 patients, aerobic exercise against waitlist at −0.41 (−0.70 to −0.12), with high-intensity programmes outperforming low-intensity ones [8] — and the umbrella file's intensity gradient held across depression, anxiety and distress alike [3].

The two-way street: the withdrawal trap and the overtraining confound

Two mechanisms make this condition's program design genuinely different from every other file in this library, and both cut against the reader's instincts. The withdrawal trap: across 19 experimental studies (385 individuals in withdrawal conditions), stopping training consistently increased depressive symptoms and anxiety, on a median timeline of 7 days, with a range of 1–42 days and stronger effects once withdrawal exceeded two weeks [9]. Read that as a lifter: the two weeks after an injury, a move, a lost job or a hospitalization are not a neutral pause — they are a pharmacological washout of something that was working, happening precisely when the reserves to rebuild the habit are lowest. The program that takes this seriously builds its bad-week protocol before the bad week (minimum viable sessions, restart rules) the way a migraine program builds its trigger rules before the storm. The overtraining confound: the joint ECSS/ACSM consensus defines the overtraining syndrome by prolonged maladaptation — fatigue, performance decline and mood disturbances — arriving after excessive overload plus inadequate recovery, and it is an exclusion diagnosis: no single hormone, test or marker confirms it [10]. In a healthy lifter, rising irritability and falling enthusiasm are early overreach warnings. In this file's reader, those are Tuesday. The practical consequence is that the usual "listen to your body" advice is structurally broken here — the body's mood channel is already noisy — so the load-bearing instrument becomes the dated log: when mood falls but performance holds, it is life; when mood and performance fall together for more than a couple of weeks on a climbing program, the consensus file's differential (overreach, under-recovery, or the illness itself) is a conversation with both the coach's data and the clinician, not a toughness call [10]. And the million-person file puts a ceiling on the "more is better" instinct this population rarely has anyway: among 1,237,194 US adults, exercisers carried 1.49 (43.2%) fewer bad mental-health days per month, with the largest associations at 45-minute sessions, 3–5 times a week — and more exercise than that was not always better [12].

The dose file: 45 minutes, 3–5 a week, and the minimum viable session

Three dose files triangulate, and they agree on an unusually modest prescription. The association file: across 1,237,194 adults, every exercise type helped (reductions 11.8–22.3% in mental-health burden), team sports largest at 22.3%, cycling 21.6%, aerobic and gym activities 20.1%, with the sweet spot at 45 minutes and 3–5 sessions a week, and frequencies beyond that showing no further — sometimes less — association [12]. The treatment file: the network meta's effects scaled with intensity rather than volume [2], the umbrella found higher intensity buying greater improvement [3], and the Duke follow-up concentrated the whole remission association inside the first 180 weekly minutes [7]. The prevention file: 33,908 healthy adults followed 11 years — one hour a week of any intensity prevented an estimated 12% of future depression cases at population level, with most of the protection already present at low doses [11]. Stack them and the depressed lifter's template writes itself: 2–3 strength days plus 1–2 aerobic days inside 45-minute sessions, roughly 150–180 weekly minutes, intensity pushed where tolerable, social formats preferred — and under it, the file's own invention, the minimum viable session: on the days the disease wins the morning, the prescription is not zero but 10–20 minutes, one or two exercises, weights the body knows — because the withdrawal ledger prices zero at real cost [9], the curvilinear file prices the first minutes at the highest return [7], and the adherence file's dropout gradient says the program that forbids bad days is the program that gets abandoned (18.1% pooled dropout, climbing with baseline severity, cut by supervision and structure [14]).

The honest counterfile: what training does not do

Five honesties keep this file from becoming a pep talk. The evidence's own confidence grade is low: in the network meta, only one included trial met Cochrane criteria for low risk of bias, and CINeMA confidence was low for walking/jogging and very low for the rest — directionally robust, precision limited [2]. The novelty fades: the umbrella found effectiveness diminished with longer-duration interventions [3] — the program must change shape before it becomes wallpaper, exactly what deload logic and variation exist for. Medication answers faster in the acute mile [5], and the drug shelf is genuinely efficacious — all 21 antidepressants beat placebo across 522 trials and 116,477 patients, OR 1.37–2.13, with only agomelatine and fluoxetine dropping out less than placebo [13] — so "exercise or meds" is a false fork the evidence never supports; the Duke files themselves ran exercise as one arm of a treatment system [5][7]. Prevention and treatment are not symmetric across diagnoses: HUNT's 33,908-person protection ran against incident depression, not anxiety [11] — for anxiety, the evidence lives on the treatment side [8], and a lifter whose anxiety is the louder column should weight the acute-file rules, not the prevention math. And dropout is the real enemy: 18.1% pooled across 40 trials, predicted by higher baseline depressive symptoms and reduced by professional supervision [14] — which is the adherence argument for everything this file prescribes below: scheduled sessions, social formats, home fallbacks, and a program that plans for its own worst week instead of pretending there won't be one.

The bad-week file: sessions that survive the symptom

This is the section that separates this door from the treatment-side files. Every rule prices against the ledgers above:

  1. Schedule the session like a prescription, not a feeling. The behavioral-activation spine: the appointment exists before motivation does, and the 0–180-minute curve [7] means the first sessions of the week are the highest-yield minutes in the whole file.
  2. Keep the minimum viable session contract. Bad day → 10–20 minutes, one or two known movements, no PRs, no judgment. Zero is the only failing grade, because zero is the price the withdrawal ledger charges [9]. A home fallback kills the commute excuse the disease relies on.
  3. Never edit the drug shelf. Starting, tapering or stopping antidepressants because the gym is going well is the file's most dangerous move — the shelf outperformed placebo across 522 trials [13], and discontinuation decisions belong to the prescriber who ran the diagnosis.
  4. Defend sleep as training infrastructure. The insomnia-depression loop feeds on late sessions and 2 a.m. program decisions; the sleep file and the mood file want the same fixed wake time.
  5. Run the log as the disentangling instrument. Track mood and performance on the same page: mood-down-performance-held is life; both-falling-for-weeks on a climbing program is the overreach conversation with data in hand [10]. This is workout tracking doing clinical work.
  6. Bias social and scheduled formats. Team formats carried the largest association in the million-person file (22.3%) [12], supervised delivery cut dropout [14] — the training partner is a pharmacological asset, and the lonely 2 a.m. garage session is the format the disease picks when it gets a vote.
  7. Restart fast after gaps. The withdrawal effects begin inside days [9] — so after injury, travel or a bad month, the restart rule is "this week, at minimum-viable dose," not "Monday after things settle."

The fences, in order

  1. The diagnosis fence. Anxiety and depression are clinician diagnoses; the bipolar screen precedes any training-based plan, because energy-surge-driven "great months" are how bipolar first announces itself, and a training log is not the instrument that catches it.
  2. The red-flag fence. Suicidal ideation, rapidly worsening depression, dissociation or reality-testing problems are same-day psychiatric events. The program pauses; the fences never negotiate with the calendar.
  3. The drug fence. The prescriber owns the shelf [13]; the training log documents response, it does not titrate. Changes in sleep, appetite, energy or sexual function get reported, not endured silently or "trained through."
  4. The interplay fence. Mood disturbance is definitional to overtraining [10] and baseline to this population — so sustained dual decline (mood and performance) is a medical conversation with the log in hand, not a deload guess.
  5. The severity fence. Severe, psychotic, peripartum or treatment-resistant depression, and panic-level anxiety, mean specialist-led care with training as adjunct at most — the file's prescriptions are for the managed, outpatient reality of these conditions, not a replacement for the care team.

The Brazil anchors

The pt-BR lane's reader holds a specific position in this condition, and the Brazilian file is unusually complete. The São Paulo megacity survey — 5,037 face-to-face WHO-CIDI household interviews — found ~30% twelve-month prevalence of any DSM-IV disorder, anxiety disorders the most common group at 19.9% (mood 11%, impulse-control 4.3%, substance 3.6%), crime exposure associated with every disorder class, and only one-third of serious cases receiving any treatment — with the authors' own verdict that prevalence ran "at greater levels than similar surveys conducted in other areas of the world" [16]. GBD puts Brazil's region in the world's top band for mental-disorder burden [1]. And the prospective anchor is Brazilian: ELSA-Brasil's four-year follow-up of 12,709 adults found any volume of leisure-time physical activity protective against incident clinical depression — RR 0.65 (0.50–0.83) already at 1–149 minutes a week, holding at 0.61–0.67 through the higher bands, with 150–299 weekly minutes of total activity carrying RR 0.55 (0.40–0.76) — protection at doses this file's minimum-viable session can hit, in the exact population reading the pt-BR twin [15]. For the Brazilian lifter, the numbers are not imported: one in five adults in the gym's postcode runs the anxiety column, most of the serious cases are untreated, and the leisure-hour is measurably protective at any dose the calendar can give it [15][16].

So what do you actually do? The practical file

Seven rules, every one sourced from the files above:

  1. Get the diagnosis, the bipolar screen and a treatment plan first, then program — this file assumes the clinician is already in the loop [5][13].
  2. Anchor the week on 2–3 strength days and 1–2 aerobic days, ~45-minute sessions, inside ~180 weekly minutes [2][7][12] — the modality the network meta ranked most acceptable [2], on ordinary progressive overload and sustainable set budgets, with warm-ups honest but unheroic.
  3. Push intensity where tolerable — the effects scale with it [2][3][8] — while keeping the withdrawal ledger's respect for total volume: the million-person file's sweet spot is 3–5 sessions, 45 minutes, and more is not always better [12].
  4. Write the minimum-viable-session contract in advance: bad days get 10–20 minutes of known movements, and zero needs a reason [7][9][14].
  5. Run mood and performance in one log, and read the two slopes together — the instrument that separates life from overreach in a population where the mood channel is noisy by default [10].
  6. Choose formats with structure and humans in them — scheduled classes, training partners, supervised delivery: the association and adherence files both price them up [12][14].
  7. Keep the fences standing: the shelf is the prescriber's, red flags are same-day events, and the program runs with the treatment plan, never instead of it [13].

How RightRep fits into this

The heavy-mind program runs on one currency this app counts better than any notebook — the dated, dosed, camera-verified session:

  • Behavioral activation needs the appointment to be real. A session that the camera counts is a session that happened, on a day the memory would have argued otherwise — the ledger the 0–180-minute curve [7] says matters most.
  • The withdrawal rule needs dated evidence. The 7-day trap [9] makes this week's restart the operative question, and the streak of camera-counted sessions is the honest answer to "when did I last train."
  • The log-as-instrument needs both columns on one page. Reps counted, load climbed, and a mood note beside them — the two-slope read [10] this file prescribes is what the app's history already produces.
  • The home fallback needs to be real training. The minimum viable session survives the disease only if home sessions count as sessions — counted reps and honest load are what make 15 living-room minutes a column entry instead of a concession.
  • The clinician-visit artifact. "What have you actually been doing since the last visit?" — the question every follow-up asks is answered by the camera-counted log you hand over [5][13].

The bottom line

  • The crown: 218 trials, 14,170 clinically depressed participants — walking/jogging g −0.62, yoga −0.55, strength training −0.49 against active controls, effects proportional to prescribed intensity, strength and yoga best tolerated [2].
  • The head-to-heads: exercise matched sertraline at 16 weeks (P=.67, medication faster early) [5]; at 10 months relapse was lower in the exercise group (p=.01) and self-directed training during follow-up priced diagnosis at OR 0.49 [6]; at one year, follow-up exercise — not original treatment arm — predicted remission, concentrated in 0–180 weekly minutes [7].
  • The resistance file: 33 trials, 1,877 participants, effect 0.66, NNT 4 — independent of volume and strength gain [4]; anxiety: 15 trials, 675 patients, −0.41, high intensity beating low [8]; the umbrella: 1,039 trials, 128,119 participants, −0.43/−0.42/−0.60 across depression/anxiety/distress [3].
  • The two-way street: stopping training raises depressive symptoms and anxiety within a median 7 days, worse past two weeks [9]; mood disturbance defines overtraining, making the log the disentangling instrument [10].
  • The dose: 45 minutes, 3–5×/week, team/gym formats largest, more not always better — 1,237,194 adults, 1.49 (43.2%) fewer bad mental-health days [12]; prevention: one weekly hour of any intensity worth 12% of future depression cases [11].
  • The honest counterfile: CINeMA confidence low/very low, one trial at low risk of bias [2]; effects diminish in longer interventions [3]; all 21 antidepressants beat placebo across 522 trials — the shelf is real, exercise is the column beside it [13]; dropout 18.1%, climbing with baseline severity, cut by supervision [14].
  • Brazil: São Paulo 19.9% twelve-month anxiety with one-third of serious cases treated [16]; ELSA-Brasil: any leisure-time volume at RR 0.61–0.67 against incident clinical depression [15]; the region sits in GBD's top burden band [1].

Train like the appointment is the treatment — action arrives before motivation, the first 180 weekly minutes carry the curve, zero is the only failing grade, and the fences are the clinician's: at rightrep.app.

This post was written by the RightRep AI Assistant — the same AI that counts and scores your reps at rightrep.app.

Sources

  1. GBD 2019 Mental Disorders Collaborators. Global, regional, and national burden of 12 mental disorders in 204 countries and territories, 1990-2019: a systematic analysis for the Global Burden of Disease Study 2019. The Lancet Psychiatry, 2022;9(2):137–150. pubmed.ncbi.nlm.nih.gov
  2. Noetel M, Sanders T, Gallardo-Gómez D, et al. Effect of exercise for depression: systematic review and network meta-analysis of randomised controlled trials. BMJ, 2024;384:e075847. pubmed.ncbi.nlm.nih.gov
  3. Singh B, Olds T, Curtis R, et al. Effectiveness of physical activity interventions for improving depression, anxiety and distress: an overview of systematic reviews. British Journal of Sports Medicine, 2023;57(18):1203–1209. pubmed.ncbi.nlm.nih.gov
  4. Gordon BR, McDowell CP, Hallgren M, Meyer JD, Lyons M, Herring MP. Association of efficacy of resistance exercise training with depressive symptoms: meta-analysis and meta-regression analysis of randomized clinical trials. JAMA Psychiatry, 2018;75(6):566–576. pubmed.ncbi.nlm.nih.gov
  5. Blumenthal JA, Babyak MA, Moore KA, et al. Effects of exercise training on older patients with major depression. Archives of Internal Medicine, 1999;159(19):2349–2356. pubmed.ncbi.nlm.nih.gov
  6. Babyak M, Blumenthal JA, Herman S, et al. Exercise treatment for major depression: maintenance of therapeutic benefit at 10 months. Psychosomatic Medicine, 2000;62(5):633–638. pubmed.ncbi.nlm.nih.gov
  7. Hoffman BM, Babyak MA, Craighead WE, Sherwood A, Doraiswamy PM, Coons MJ, Blumenthal JA. Exercise and pharmacotherapy in patients with major depression: one-year follow-up of the SMILE study. Psychosomatic Medicine, 2011;73(2):127–133. pubmed.ncbi.nlm.nih.gov
  8. Aylett E, Small N, Bower P. Exercise in the treatment of clinical anxiety in general practice – a systematic review and meta-analysis. BMC Health Services Research, 2018;18(1):559. pubmed.ncbi.nlm.nih.gov
  9. Weinstein AA, Koehmstedt C, Kop WJ. Mental health consequences of exercise withdrawal: a systematic review. General Hospital Psychiatry, 2017;49:11–18. pubmed.ncbi.nlm.nih.gov
  10. Meeusen R, Duclos M, Foster C, et al. Prevention, diagnosis, and treatment of the overtraining syndrome: joint consensus statement of the European College of Sport Science and the American College of Sports Medicine. Medicine and Science in Sports and Exercise, 2013;45(1):186–205. pubmed.ncbi.nlm.nih.gov
  11. Harvey SB, Øverland S, Hatch SL, Wessely S, Mykletun A, Hotopf M. Exercise and the prevention of depression: results of the HUNT cohort study. American Journal of Psychiatry, 2018;175(1):28–36. pubmed.ncbi.nlm.nih.gov
  12. Chekroud SR, Gueorguieva R, Zheutlin AB, et al. Association between physical exercise and mental health in 1·2 million individuals in the USA between 2011 and 2015: a cross-sectional study. The Lancet Psychiatry, 2018;5(9):739–746. pubmed.ncbi.nlm.nih.gov
  13. Cipriani A, Furukawa TA, Salanti G, et al. Comparative efficacy and acceptability of 21 antidepressant drugs for the acute treatment of adults with major depressive disorder: a systematic review and network meta-analysis. The Lancet, 2018;391(10128):1357–1366. pubmed.ncbi.nlm.nih.gov
  14. Stubbs B, Vancampfort D, Rosenbaum S, et al. Dropout from exercise randomized controlled trials among people with depression: a meta-analysis and meta regression. Journal of Affective Disorders, 2016;190:457–466. pubmed.ncbi.nlm.nih.gov
  15. Werneck AO, Schuch FB, Vancampfort D, et al. Physical activity domains and incident clinical depression: a 4-year follow-up analysis from the ELSA-Brasil cohort. Journal of Affective Disorders, 2023;329:385–393. pubmed.ncbi.nlm.nih.gov
  16. Andrade LH, Wang YP, Andreoni S, et al. Mental disorders in megacities: findings from the São Paulo megacity mental health survey, Brazil. PLoS ONE, 2012;7(2):e31879. pubmed.ncbi.nlm.nih.gov