what workslast verified 2026-09-23

what works that costs nothing

most of the mental health evidence people hear about is for things somebody sells. this is the other list: the formulas from the same journals that cost nothing and need no appointment. each row is one thing you can start this week, the trial behind it, and the catch, because there is always a catch and a page that hides it is selling something too.

the eight, with the catch

1. walk, or jog, most days[1]

the formula
thirty minutes if you can, brisk enough to notice your breathing. harder counts for more.
the evidence
network meta-analysis of 218 trials, 14,170 people with major depression, BMJ 2024
the size
moderate: g −0.62 against usual care or a placebo tablet. yoga −0.55, strength training −0.49, tai chi −0.42. effects rose with intensity.
the catch
the authors say confidence is low: only one of 218 trials met the bar for low risk of bias, and nobody can blind a person to whether they exercised. the honest summary is “works in trial after trial, measured imperfectly.”

2. half the recommended dose is most of the benefit[2][3]

the formula
about 75 minutes of brisk walking a week. the first hour does the most work.
the evidence
15 prospective cohorts, 191,130 adults followed for years, JAMA Psychiatry 2022; and a genetic study of 611,583 adults, JAMA Psychiatry 2019
the size
18 percent lower risk of depression at half the recommended volume; 25 percent at the full volume; little more beyond that. the genetic study points the arrow the right way: measured activity lowers depression risk, not the reverse.
the catch
cohorts show association. the genetic study supports cause but only for activity measured by a wrist sensor, not self-report, which is the same thing as saying: what people say they do did not predict anything.

3. lift something[4]

the formula
two or three sessions a week. how much you lift, and whether you get stronger, did not change the result.
the evidence
33 randomised trials, 1,877 adults, JAMA Psychiatry 2018
the size
effect 0.66; number needed to treat 4, meaning one in four people got a meaningful improvement they would not otherwise have had.
the catch
the trials that blinded the assessor found smaller effects. take the 0.66 as an upper bound.

4. thirty minutes of bright light, early morning[5]

the formula
a 10,000-lux light box, 30 minutes soon after waking, for eight weeks. this is for ordinary, non-seasonal depression.
the evidence
double-blind, sham-controlled trial of 122 outpatients, JAMA Psychiatry 2016
the size
light alone: d 0.80 against sham; 44 percent in remission against 30 percent on placebo. light plus the pill: 59 percent.
the catch
one trial, 122 people. in this trial the antidepressant alone did not beat placebo (19 percent remission), which is one small trial’s result, not a verdict on the drug. the box costs money; a bright morning outdoors is the free version and was not what was tested.

5. fix sleep first[6][7]

the formula
a six-to-ten-week course of cognitive behavioural therapy for insomnia, which exists as free and low-cost digital programs: fixed wake time, out of bed when awake, no catching up.
the evidence
3,755 students with insomnia, Lancet Psychiatry 2017; 1,149 adults with insomnia and low mood, Lancet Psychiatry 2016
the size
insomnia fell by a large margin (d 1.11), and paranoia and hallucinatory experiences fell with it, through the sleep. in the second trial depression scores were lower at six weeks and six months.
the catch
the second trial did not prevent new major depression (9 versus 13 cases, not significant) and fewer than half finished follow-up. the effect is on symptoms, not on diagnoses.

6. schedule one thing you value, every day, and do it whether you feel like it or not[8][9][10]

the formula
this is behavioural activation. write down what you have stopped doing that used to matter. put one back on the calendar tomorrow. then the next.
the evidence
440 adults with major depression, Lancet 2016; 495 adults with moderately severe to severe depression in Goa, Lancet 2017; 2,452 teenagers, Nature Human Behaviour 2022
the size
delivered by junior workers it matched full cognitive behavioural therapy at one year (8.4 versus 8.4 on the PHQ-9). delivered by lay counsellors in six to eight sessions: remission 64 percent against 39 percent with usual care, fewer days off work, fewer suicidal thoughts. one online session for teenagers: a small effect, d 0.18, three months later.
the catch
the single session is small, and the authors say so. the big results came with a person, even an untrained one, checking in. the free version is the calendar and a friend who asks.

7. five minutes of sighing[11]

the formula
breathe in through the nose, a second smaller sip in, then a long slow breath out through the mouth. repeat for five minutes a day.
the evidence
a remote randomised study over one month, Cell Reports Medicine 2023
the size
better mood improvement and a lower breathing rate than five minutes of mindfulness meditation.
the catch
smaller journal, one month, effects modest. it is on this list because it takes five minutes and the comparison was a fair one, not because the evidence is the size of the rows above.

8. an eight-week mindfulness course, for staying well[12]

the formula
mindfulness-based cognitive therapy: eight weekly group sessions plus daily practice. for people who have had depression three or more times and want an alternative to staying on pills for years.
the evidence
424 adults on maintenance antidepressants, 95 UK practices, followed two years, Lancet 2015
the size
relapse over 24 months did not differ between the course (with a supported taper) and staying on the medication: hazard ratio 0.89, not significant. serious adverse events did not differ.
the catch
“not different” is not “better,” and the authors say there was no evidence of superiority. the taper in that trial was supervised by the person’s doctor. this row is a question to bring to a prescriber, not a plan to make from a page.

how to read the sizes

a “g” or “d” of 0.2 is small, 0.5 is medium, 0.8 is large; they say how far the treated group moved compared with the spread of scores. number needed to treat is how many people take the thing for one to get a benefit they would not otherwise have had; 4 is very good. hazard ratio 0.89 with a confidence interval crossing 1 means “we could not tell the two apart.” a percentage from a cohort is a risk difference between people who did and did not do the thing on their own, which is not the same as a trial.

what is not on the list, and why

supplements, because we have a page on them and the answer is mostly no. cold water, gratitude journals, and time in nature, because the trials are small, short, or not in these journals; they may be fine, and this page only carries what it can show. diet, because the best-known trial is small and we have not read it for this page. anything that costs money every month.

questions people ask

Does exercise really treat depression, or just help a little?

The largest analysis, 218 randomised trials in the BMJ in 2024, found moderate effects for walking or jogging, yoga and strength training against usual care or a placebo tablet, larger when the exercise was harder. The same paper says confidence is low because almost no trial could be blinded. So: consistently positive, imperfectly measured. It is listed there as a treatment to consider alongside therapy and medication, not instead of them.

How much exercise is enough for mental health?

Less than the guideline. In 15 cohorts with 191,130 adults, half the recommended volume, about 75 minutes of brisk walking a week, carried 18 percent lower risk of depression; the full volume 25 percent; more than that added little. The first hour does most of the work.

Can I stop my antidepressant and do these instead?

Not from this page. One trial on this list found an eight-week mindfulness course with a doctor-supervised taper matched maintenance medication for relapse over two years, and the authors say that is “not superior,” not “better.” Everything here can be added to a treatment today; replacing a treatment is a decision with your prescriber.

Which of these has the weakest evidence?

The five-minute breathing row: one month, a smaller journal, modest effects. It is on the list because the comparison was fair and the cost is five minutes. The single-session online intervention for teenagers is also small (d 0.18) and its authors say so.

sources

  1. 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. 218 trials, 14,170 participants meeting clinical cut-offs for major depression. Versus active controls: walking or jogging Hedges’ g −0.62 (95% credible interval −0.80 to −0.45), yoga −0.55, strength training −0.49, mixed aerobic −0.43, tai chi or qigong −0.42; effects proportional to prescribed intensity; equally effective with and without comorbidities and at different baseline severities. “Only one study met the Cochrane criteria for low risk of bias”; confidence low for walking or jogging and very low for the others. PMID 38355154. https://doi.org/10.1136/bmj-2023-075847
  2. Pearce M, Garcia L, Abbas A, et al. Association between physical activity and risk of depression: a systematic review and meta-analysis. JAMA Psychiatry 2022;79:550-559. 15 prospective cohorts, 191,130 adults, 2.1 million person-years. Half the recommended activity volume (4.4 marginal MET-hours a week) was associated with 18% (13% to 23%) lower risk of depression; the full recommended volume (8.8) with 25% (18% to 32%); diminishing benefit beyond that. Estimated 11.5% of depression cases preventable if less active adults met the recommendation. PMID 35416941. https://doi.org/10.1001/jamapsychiatry.2022.0609
  3. Choi KW, Chen CY, Stein MB, et al. Assessment of bidirectional relationships between physical activity and depression among adults: a 2-sample Mendelian randomization study. JAMA Psychiatry 2019;76:399-408. Genetic instruments from 611,583 adults: accelerometer-measured activity protective against major depression (OR 0.74 per SD, 95% CI 0.59 to 0.92); no evidence depression reduces activity; no relationship for self-reported activity. PMID 30673066. https://doi.org/10.1001/jamapsychiatry.2018.4175
  4. Gordon BR, McDowell CP, Hallgren M, et al. Association of efficacy of resistance exercise training with depressive symptoms: meta-analysis and meta-regression analysis of randomized clinical trials. JAMA Psychiatry 2018;75:566-576. 33 trials, 1,877 participants; mean effect 0.66 (95% CI 0.48 to 0.83); number needed to treat 4; effect not associated with total volume prescribed, health status, or whether strength improved; smaller effects in trials with blinded allocation or assessment. PMID 29800984. https://doi.org/10.1001/jamapsychiatry.2018.0572
  5. Lam RW, Levitt AJ, Levitan RD, et al. Efficacy of bright light treatment, fluoxetine, and the combination in patients with nonseasonal major depressive disorder: a randomized clinical trial. JAMA Psychiatry 2016;73:56-63. 122 outpatients, 8 weeks, double-blind, sham-controlled. 10,000-lux light for 30 minutes in the early morning. Light alone d 0.80 versus placebo; combination d 1.11; fluoxetine 20 mg alone d 0.24, not superior to placebo. Remission: placebo 30.0%, fluoxetine 19.4%, light 43.8%, combination 58.6%. PMID 26580307. https://doi.org/10.1001/jamapsychiatry.2015.2235
  6. Freeman D, Sheaves B, Goodwin GM, et al. The effects of improving sleep on mental health (OASIS): a randomised controlled trial with mediation analysis. Lancet Psychiatry 2017;4:749-758. 3,755 university students with insomnia at 26 UK universities; digital CBT for insomnia versus usual care. At 10 weeks: insomnia reduced (Cohen’s d 1.11), paranoia (d 0.19) and hallucinations (d 0.24); insomnia mediated the change; no adverse events. PMID 28888927. https://doi.org/10.1016/S2215-0366(17)30328-0
  7. Christensen H, Batterham PJ, Gosling JA, et al. Effectiveness of an online insomnia program (SHUTi) for prevention of depressive episodes (the GoodNight Study): a randomised controlled trial. Lancet Psychiatry 2016;3:333-341. 1,149 adults with insomnia and depression symptoms below the major-depression threshold; six-week online CBT-I versus an attention-matched placebo program. Depression symptoms (PHQ-9) significantly lower at 6 weeks and 6 months; new major depressive disorder at 6 months 9 versus 13 cases, not significantly different; 44% completed 6-month follow-up. PMID 26827250. https://doi.org/10.1016/S2215-0366(15)00536-2
  8. Richards DA, Ekers D, McMillan D, et al. Cost and outcome of behavioural activation versus cognitive behavioural therapy for depression (COBRA): a randomised, controlled, non-inferiority trial. Lancet 2016;388:871-880. 440 adults with major depression; BA from junior mental health workers versus CBT from psychological therapists. PHQ-9 at 12 months: 8.4 versus 8.4 (mean difference 0.1, 95% CI −1.3 to 1.5); non-inferior. PMID 27461440. https://doi.org/10.1016/S0140-6736(16)31140-0
  9. Patel V, Weobong B, Weiss HA, et al. The Healthy Activity Program (HAP), a lay counsellor-delivered brief psychological treatment for severe depression, in primary care in India: a randomised controlled trial. Lancet 2017;389:176-185. 495 adults with PHQ-9 above 14 in ten primary health centres in Goa; lay-counsellor behavioural activation plus usual care versus usual care. Remission (PHQ-9 below 10) at 3 months 64% versus 39%; fewer days out of work; fewer suicidal thoughts or attempts (prevalence ratio 0.61). PMID 27988143. https://doi.org/10.1016/S0140-6736(16)31589-6
  10. Schleider JL, Mullarkey MC, Fox KR, et al. A randomized trial of online single-session interventions for adolescent depression during COVID-19. Nature Human Behaviour 2022;6:258-268. 2,452 adolescents aged 13 to 16 from all 50 US states with elevated depression symptoms; one online session of behavioural activation or a growth-mindset session versus a supportive control. Three-month depressive symptoms reduced, Cohen’s d 0.18; hopelessness d 0.16 to 0.28. PMID 34887544. https://doi.org/10.1038/s41562-021-01235-0
  11. Balban MY, Neri E, Kogon MM, et al. Brief structured respiration practices enhance mood and reduce physiological arousal. Cell Reports Medicine 2023;4:100895. Remote randomised study of three daily 5-minute breathing exercises versus 5 minutes of mindfulness meditation for one month. Exhale-focused cyclic sighing produced greater improvement in mood and a larger reduction in respiratory rate than meditation (p < 0.05). Smaller journal and shorter study than the rest of this page; listed as such. PMID 36630953. https://doi.org/10.1016/j.xcrm.2022.100895
  12. Kuyken W, Hayes R, Barrett B, et al. Effectiveness and cost-effectiveness of mindfulness-based cognitive therapy compared with maintenance antidepressant treatment in the prevention of depressive relapse or recurrence (PREVENT): a randomised controlled trial. Lancet 2015;386:63-73. 424 adults with three or more previous episodes on maintenance antidepressants, from 95 UK general practices; MBCT with support to taper versus continued antidepressants. Time to relapse over 24 months did not differ (hazard ratio 0.89, 95% CI 0.67 to 1.18); serious adverse events did not differ; “no evidence that MBCT-TS is superior.” PMID 25907157. https://doi.org/10.1016/S0140-6736(14)62222-4

related: does trazodone work for sleep · supplements for mental health · does tms work for depression

adam