Your Brain Is Not One Simple Story.

What you'll learn

What the serotonin study found, and the conclusion people wrongly drew from it

What actually treats depression, including the option that gets left out of the argument

What 54 randomised trials found when they tested cannabis for mental health conditions

Why nearly half of dementia cases may be preventable or delayed

Why a frightening statistic about athletes' brains can be accurate and still tell you almost nothing

How SmartBar applies the deficiency framework

SmartBar's educational framework begins with two organizing terms, and applies them wherever the body has inputs and burdens. Deficiency means something the body needs and isn't getting. Toxicity means a burden the body must process and clear, not a poison.

On this page that framework describes contributors and supports. It does not describe causes, and it does not describe treatments. Depression and anxiety are not nutrient deficiencies. Dementia is not a lifestyle failure. Anyone who tells you otherwise is selling something, and we would rather draw that line ourselves than let you assume we mean more than we do.

This page is educational and is not medical advice. Nothing here is a reason to start, stop or change a medication. If you are in crisis or thinking about harming yourself, please contact emergency services or a crisis line in your country now, rather than finishing this article.

For about thirty years, most people were told a simple story. Depression is a chemical imbalance in the brain, serotonin is low, and antidepressants correct it.

It was never quite what researchers believed. It was what patients were told, because it was easy to explain and it took blame away from people who had been blamed enough.

Then a review pulled that story apart in public, and the reaction caused its own harm.

That pattern repeats everywhere in mental health. A partial truth gets sold as the whole truth. Evidence arrives showing it was too simple. The correction overshoots. And people make decisions about their own health in the wreckage.

What the serotonin study actually found

A team led by Joanna Moncrieff published an umbrella review in Molecular Psychiatry concluding there was no consistent evidence that depression is caused by low serotonin.[1] It was downloaded over a million times and made headlines worldwide.

The response was substantial. A Comment in the same journal, written by 36 researchers, argued the review presented no new analysis, applied its own quality criteria, interpreted findings differently from the original researchers, and departed from standard umbrella review methodology.[2] Further replies followed.

The most useful point came from the critics rather than the defenders. The serotonin theory was never proposed by researchers as "you are depressed because one chemical is low." No current theory of how antidepressants work claims they act only by correcting a prior chemical imbalance.[3]

So the review largely dismantled a version of the idea that specialists had already moved past, while the public still believed it, because that is how it had been explained to them.

Here is where real harm entered. The headlines became "depression is not a chemical imbalance," and readers drew a further conclusion the study did not support: that antidepressants therefore do not work.

That does not follow. Whether a drug helps is a separate question from whether a particular mechanism explains an illness. Aspirin worked for a century before anyone understood why. Whether antidepressants help, for whom, and how much is addressed by a different body of evidence entirely.[3]

If you take one thing from this page

Do not stop an antidepressant abruptly.

Discontinuation effects are real, and the size of the problem is itself disputed in a way worth seeing.

A systematic review in Lancet Psychiatry found discontinuation symptoms in 31% of people stopping antidepressants against 17% on placebo. After subtracting that non-specific effect, the authors estimated the true incidence at around 15%, roughly one in seven, with about one in 35 experiencing severe symptoms.[4]

Earlier reviews put the figure near 56%, with nearly half of cases severe. The Lancet Psychiatry authors argue those relied on surveys prone to selection bias. Researchers behind the deprescribing guidelines have published reanalyses disputing the lower estimate.[4]

What nobody disputes: discontinuation symptoms are real, they are more likely with shorter half-life drugs and longer treatment, and guidelines recommend tapering rather than stopping suddenly.

Even the researchers who challenged the serotonin story have raised concerns about withdrawal. There is no version of this where reading a headline and stopping your medication is the sensible response.

If you are taking an antidepressant and this article has made you reconsider it, that is a conversation to have with your prescriber or pharmacist, not a decision to make alone. They can advise on whether a change is appropriate and, if it is, how to taper safely. Nothing on this page is a reason to start, stop or change any medication.

So what does contribute?

Depression and anxiety do not have one cause. What the evidence supports is a set of contributors that vary between people: genetic vulnerability, adverse experiences particularly in childhood, chronic stress, physical illness, social isolation, poverty, sleep disruption, substance use, and life events. Hormonal and reproductive transitions matter too, and are frequently treated as an afterthought. Depression during pregnancy and after childbirth is common, while perimenopause appears to be a period of increased vulnerability: one meta-analysis found roughly 40% higher odds of depressive symptoms or diagnosis during perimenopause than before it.[5]

That is less satisfying than a chemical imbalance, and more accurate.

It also means the same diagnosis can arrive by different routes in different people, which is part of why treatment response varies so much. A framework that explains everything explains nothing, and this page will not pretend otherwise about food, screens, cannabis or anything else.

A note on anxiety

Anxiety and depression overlap heavily and are not the same thing. Anxiety disorders are not simply "too much stress," and they are not a milder version of depression. Some of the same supports recur across both, including psychotherapy, exercise and sleep, but the disorders and their treatments are not interchangeable.

Anxiety also has measurable physical correlates. A meta-analysis of 46 cohort studies covering more than two million people found anxiety associated with higher cardiovascular mortality, coronary disease and stroke.[6] That does not prove anxiety itself causes those outcomes. But it is another reminder that mental and physical health are not separate systems.

What actually treats depression

The serotonin argument tends to split the conversation into "medication works" against "medication does not work." That framing leaves out one of the largest evidence bases in the field.

Psychotherapy. Hundreds of randomised trials support psychological treatments for depression. Network meta-analyses covering cognitive behavioural therapy, interpersonal therapy, behavioural activation, problem-solving therapy and psychodynamic therapy find the established approaches effective, with relatively modest differences between them.[7]

This is not "talking about your problems." Behavioural activation systematically rebuilds activity when withdrawal has become self-reinforcing. Cognitive work targets specific thinking patterns. These are structured interventions with trial evidence, not conversation.

The American College of Physicians recommends either cognitive behavioural therapy or a second-generation antidepressant as initial treatment for moderate to severe major depression, with combination treatment also an option.[7] Choice depends on severity, previous response, side effect tolerance, availability and preference.

So the honest landscape is not medication against lifestyle. It is several treatments with real evidence, a genuine choice between them, and, as this page will show later, a global failure to make any of them available to most people who need them.

Food, mood, and the microbiome

This is where we have to be most careful, because a food transparency platform has the most to gain by overstating.

The landmark trial. SMILES randomised adults with moderate to severe depression to twelve weeks of dietary counselling or to matched social support. The dietary group improved significantly more, with a large effect size.[8]

That effect size is the problem. It was larger than most antidepressant trials produce, from a twelve week diet intervention in a small sample. Other researchers published a direct challenge asking whether recruitment practices explained it, and the trial authors responded in print.[8]

The wider evidence is more modest. Observational studies associate Mediterranean-style eating with roughly 30% lower depression risk, but intervention evidence is mixed.[8]

On the gut-brain axis: the biology is real, but commercial claims have outrun clinical evidence. Meta-analyses of probiotic trials now report positive pooled signals for depressive symptoms, but studies remain heterogeneous in strains, populations and designs, and the evidence does not establish a general "fix your microbiome and treat depression" prescription.[9]

So what we will say: diet quality is associated with depression risk in large observational datasets, and eating well is worth doing for reasons that do not depend on this literature holding up. What we will not say is that diet is an established standalone treatment for major depression. It is not, on current evidence.

What helps without buying a wellness product

Movement, which has more than support-level evidence. A network meta-analysis pooled 218 randomised trials with over 14,000 participants and found walking, jogging, yoga, strength training and dance all reduced depressive symptoms, with benefits proportional to intensity.[10] The same analysis rated confidence in most individual comparisons as low, largely because participants cannot be blinded to whether they are exercising. Broader umbrella analyses covering more than 79,000 participants find reductions in both depressive and anxiety symptoms, with effect sizes varying considerably by population.[10] Our Exercise page starts with walking.

Sleep and connection. Poor sleep is both a symptom of depression and a contributor to it, and treating insomnia can improve mood independently; our Sleep page covers why the first-line treatment is behavioural and rarely offered. Social isolation is consistently associated with worse outcomes across this library, and purpose with better ones, which our Lifestyle page covers.

These are not substitutes for individualised care. Exercise itself has randomised treatment evidence for depression. Sleep and social connection matter through several pathways. None of them is a reason to abandon an effective treatment.

Cannabis

Cannabis is now among the most widely promoted mental health remedies, for anxiety in particular. The randomised evidence has caught up, and it does not say what the marketing says.

A systematic review and meta-analysis in Lancet Psychiatry screened 5,774 studies and analysed 54 randomised controlled trials with 2,477 participants, spanning 1980 to 2025, testing cannabinoids as a primary treatment for mental disorders and substance use disorders.[11]

No significant benefit was found for anxiety, PTSD, psychotic disorders, anorexia nervosa or opioid use disorder. Evidence for OCD and several other conditions was too sparse to meta-analyse, and there were no randomised trials at all for depression, despite depression being among the most common reasons medical cannabis is prescribed in legalised markets.[11] Cannabinoids increased cocaine craving compared with placebo.

The quality caveats matter and cut both ways: 24 of the 54 trials were rated at high risk of bias, a fifth raised conflict of interest concerns, and the median trial enrolled around 31 participants. This is a thin evidence base being synthesised, which is itself the finding.

On the risk side the evidence is firmer. A systematic review found higher potency cannabis associated with psychosis and cannabis use disorder, while results for anxiety and depression were mixed and sometimes contradictory.[12] Case-control work found daily users had roughly three times the odds of a psychotic disorder, rising further with high potency products.[12] Experimental studies show a single administration of THC induces psychotic and other psychiatric symptoms with large effect sizes, while CBD does not.[13]

The honest summary. For anxiety, randomised evidence has not established benefit. For people vulnerable to psychosis, high-THC products carry a materially different risk profile.

Screens, young people, and an argument that is not settled

We are asked about this more than anything else, and the honest answer is that it is unresolved.

The prominent argument is that smartphones and social media rewired childhood and drove a rise in adolescent mental illness.

Leading researchers dispute it. Writing in Nature, Candice Odgers argued the central causal claim is not supported by the science and risks distracting from other causes of the youth mental health crisis.[14] Reanalyses of the underlying data thousands of different ways found the average association very small, and an analysis of 946,798 people across 72 countries found no evidence that Facebook adoption tracked widespread population-level psychological harm, while cautioning that the design cannot resolve individual causal effects.[15]

The emerging middle position is more useful than either camp. Researchers increasingly argue that "screen time" is too crude a variable: what someone is doing, with whom, at what developmental stage, and what the platform is amplifying may matter more than the clock.[14] Social comparison, exclusion, feedback sensitivity and exposure to harmful content are different exposures with different effects, and lumping them together as hours has held the field back.

Two things this does not mean. It does not mean screens are harmless, which even the critics do not claim. And it does not mean a parent's observation of their own child is wrong, because individual effects can be real where population averages are small.

The clearest concern is displacement, when it actually occurs: hours online that replace sleep, movement or face-to-face time. Screens can also be connection, so this is not automatic. But those three have strong evidence elsewhere in this library, and that argument does not require the addiction framing to hold.

The aging brain

Here the evidence gets stronger, and connects most directly to everything else on this platform.

Around 45% of dementia cases may potentially be prevented or delayed by addressing 14 modifiable risk factors across the life course, according to the 2024 report of the Lancet standing Commission.[16]

The list: low education, hearing loss, high LDL cholesterol, depression, traumatic brain injury, physical inactivity, diabetes, smoking, hypertension, obesity, excessive alcohol, social isolation, air pollution, and untreated vision loss.[16]

Nine of those already have their own article on this platform, which is not a coincidence: what protects your heart and your metabolism largely protects your brain.

Two honest boundaries. These are population attributable fractions, not individual guarantees. Eliminating a risk factor across a population is not the same as any one person avoiding dementia. And the Commission's data come largely from high-income countries; a 2025 analysis argues poverty, income inequality, wealth shocks and viral infection should be added for global applicability.[17]

Is Alzheimer's "type 3 diabetes"?

You may have seen the phrase. There is real science under it.

Post-mortem studies of Alzheimer's brains consistently show reduced insulin and insulin-like growth factor signalling, lower receptor density, and markers of insulin resistance. Researchers have argued since 2008 that Alzheimer's is fundamentally metabolic, and a 2025 systematic review examines the case seriously.[18]

Three things keep it a hypothesis. It is not a recognised diagnostic category anywhere. A published critical appraisal asks directly whether the label holds. And researchers themselves say the mechanisms remain largely unknown, with brain insulin resistance appearing at least partly independent of systemic type 2 diabetes.[18]

The metabolic link to cognitive decline is real and worth taking seriously. The renaming is not established. Both are true at once.

On cholesterol

High midlife LDL cholesterol was added as a dementia risk factor in 2024.[16] Claims that lowering circulating cholesterol necessarily deprives the brain oversimplify the physiology: the brain synthesises most of its cholesterol locally. The wider statin and cognition evidence is contested and deserves its own article, which it is getting.

Head injuries and contact sport

Here is a statistic you may have seen. In selected studies of donated brains from former contact sport athletes, CTE has been found in more than half of those examined — a 2025 meta-analysis of eight studies covering 1,000 athletes put the pooled prevalence at 53.7%.[19]

That figure is real, and close to meaningless without the next sentence.

Those were donated brains. Families donate when they suspect something was wrong. Screen an unselected brain bank instead and the picture inverts: of 636 cases from one brain bank, CTE appeared in 0.8%.[20]

Same condition. Two orders of magnitude apart. Who ends up in the sample changes the number dramatically.

The field genuinely disagrees about what that means. A systematic review of every reported case concluded the true incidence remains unknown and that case reporting is inherently subject to selection bias.[20] A 2023 paper in Acta Neuropathologica argues the opposite: over 97% of published cases had known repetitive head impact exposure, a dose-response relationship with years of play persists after accounting for selection bias, and no plausible alternative explanation has been advanced.[21]

What is not disputed is that traumatic brain injury appears on the Lancet Commission's list of modifiable dementia risk factors, and that repetitive head impacts are worth avoiding where avoidable.

This section is here because it teaches something the rest of the page depends on. A frightening statistic can be accurate and still describe almost nothing about your risk, if you do not know how the sample was chosen.

The gap that matters most

Across several articles now, this platform has run into the same finding. The first-line treatment for chronic insomnia is behavioural and is rarely offered. Medication for alcohol use disorder reaches a fraction of those who could benefit. Here it appears again, at its most severe.

Minimally adequate treatment for major depression ranged from 23% in high-income countries to 3% in low and lower-middle-income countries.[22]

Treatment gaps by region run to roughly 45% in Europe, 67% in Africa and 70% in the Eastern Mediterranean.[22]

So when someone is not getting better, the question is not only what is wrong with them. It is often whether anything effective was ever available. That is a systems failure, and no amount of individual effort closes it.

Key Takeaways

The chemical imbalance story was always a simplification. That is not evidence that antidepressants do not work, which is a separate question with separate evidence.

Never stop an antidepressant abruptly. Discontinuation effects are real, disputed in size, and tapering is recommended.

Psychotherapy belongs in the conversation. Guidelines place cognitive behavioural therapy alongside medication as an initial option for moderate to severe depression.

Cannabis does not have randomised evidence supporting its popular use for anxiety or depression. Across 54 trials there was no significant benefit for anxiety, PTSD or psychotic disorders, and no trials at all for depression.

Around 45% of dementia cases may be prevented or delayed by addressing 14 modifiable risk factors, most of which already have an article on this platform.

A frightening number can be accurate and still mislead if you do not know who was in the sample. The CTE prevalence figures are the clearest example here.

How to think about this in thirty seconds

Distrust single-cause explanations, whether the villain is serotonin, sugar, screens or cannabis. Contributors vary between people.
Separate "does this explain the illness" from "does this treatment help." Different questions, and the answers can differ.
Know that there is more than one treatment with evidence. Therapy, medication and their combination are all options, and the choice is legitimately yours to discuss.
Ask who was in the study before acting on a statistic. Donated brains, self-selected surveys and clinic populations all produce numbers that do not describe you.
If something is wrong, ask what is actually available. Too many people never receive minimally adequate care in the first place.

What SmartBar does with this

Very little directly, and it would be dishonest to imply otherwise.

SmartBar scores packaged food. It cannot assess your mental health, and there is no scan that helps here.

Where the platform genuinely touches this page is at the metabolic end. Diabetes, obesity, hypertension and high LDL cholesterol all appear on the dementia risk factor list, and all four are shaped in part by what is in the food you buy. Making that legible is a real contribution to one part of a much larger picture.

The rest of this page exists because a subject that affects our readers does not stop mattering because it cannot be routed through a barcode.

This article is educational and is not medical advice. Nothing here is a reason to start, stop or change any medication, and antidepressants in particular should not be stopped abruptly. If you are struggling, a GP or primary care doctor is a reasonable first contact, and most countries have free and confidential mental health helplines. If you are in crisis, having thoughts of harming yourself, or worried about someone who is, please contact emergency services or a crisis line immediately. Dementia, Parkinson's disease and other neurological conditions require assessment by qualified clinicians, and decisions about medications for them belong in that conversation rather than with a website.

Keep exploring: Lifestyle · Sleep · Exercise · Nutrition · Addiction · Obesity

Medical Disclaimer: Read our full Medical Disclaimer

References

Moncrieff, J., Cooper, R.E., Stockmann, T., Amendola, S., Hengartner, M.P., Horowitz, M.A. (2023). "The serotonin theory of depression: a systematic umbrella review of the evidence." Molecular Psychiatry 28:3243–3256. DOI: 10.1038/s41380-022-01661-0. (Published online 20 July 2022; issue dated 2023.)
Jauhar, S., Arnone, D., Baldwin, D.S., et al. (2023). "A leaky umbrella has little value: evidence clearly indicates the serotonin system is implicated in depression." Molecular Psychiatry 28:3149–3152. DOI: 10.1038/s41380-023-02095-y. Authored by 36 researchers; one of several contemporaneous responses rather than a single unified rebuttal. (Argues the umbrella review presented no new analyses, used its own criteria for assessing research quality, interpreted findings differently from the original research, and departed from conventional umbrella review methodology.)
Expert commentary and antidepressant efficacy evidence. Science Media Centre expert reaction (2022), noting that no current theory of antidepressant action asserts the drugs work only by correcting a prior chemical imbalance. Efficacy: Cipriani, A., Furukawa, T.A., Salanti, G., et al. (2018). "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 391(10128):1357–1366. DOI: 10.1016/S0140-6736(17)32802-7.
Henssler, J., Schmidt, Y., Schmidt, U., Schwarzer, G., Bschor, T., Baethge, C. (2024). "Incidence of antidepressant discontinuation symptoms: a systematic review and meta-analysis." Lancet Psychiatry 11(7):526–535. DOI: 10.1016/S2215-0366(24)00133-0. (31% versus 17% on placebo; estimated true incidence approximately 15%, roughly one in 35 severe. Contested: earlier reviews reported approximately 56% with nearly half severe (Davies & Read, 2019, Addictive Behaviors 97:111–121), which Henssler et al. argue relied on studies prone to selection bias; Horowitz and Taylor have published reanalyses disputing the lower figure. Both sides agree symptoms are real and tapering is recommended.)
Badawy, Y., Spector, A., Li, Z., Desai, R. (2024). "The risk of depression in the menopausal stages: A systematic review and meta-analysis." Journal of Affective Disorders 357:126–133. DOI: 10.1016/j.jad.2024.04.041. PMID: 38642901. (17 prospective cohort studies; OR 1.40 [95% CI 1.21–1.61] for depressive symptoms or diagnosis during perimenopause versus premenopause, with no significant increase after menopause.)
Anxiety and cardiovascular outcomes. Emdin, C.A., et al. (2016). "Meta-Analysis of Anxiety as a Risk Factor for Cardiovascular Disease." American Journal of Cardiology. PMID: 27324160. (46 cohort studies, 2,017,276 participants, 222,253 with anxiety. Cardiovascular mortality RR 1.41 [1.13–1.76]; coronary heart disease RR 1.41 [1.23–1.61]; stroke RR 1.71 [1.18–2.50]. The authors state that whether these associations are causal is unclear.) Roest, A.M., Martens, E.J., de Jonge, P., Denollet, J. (2010). "Anxiety and risk of incident coronary heart disease: a meta-analysis." Journal of the American College of Cardiology. PMID: 20620715. (20 prospective studies, 249,846 initially healthy people, mean follow-up 11.2 years; HR 1.26 [1.15–1.38] for incident CHD independent of traditional risk factors. A published critique notes that some analyses mixed cohorts including people with existing coronary disease at baseline, blurring aetiological and prognostic questions.)
Psychotherapy for depression. Cuijpers, P., Quero, S., Noma, H., et al. (2021). "Psychotherapies for depression: a network meta-analysis covering efficacy, acceptability and long-term outcomes of all main treatment types." World Psychiatry 20(2):283–293. DOI: 10.1002/wps.20860. (Established psychotherapies for adult depression found efficacious and acceptable, with few consistent differences between them.) Guideline: Qaseem, A., Owens, D.K., Etxeandia-Ikobaltzeta, I., et al. (2023). "Nonpharmacologic and Pharmacologic Treatments of Adults in the Acute Phase of Major Depressive Disorder: A Living Clinical Guideline From the American College of Physicians." Annals of Internal Medicine 176(2):239–252. DOI: 10.7326/M22-2056. (Recommends either cognitive behavioural therapy or a second-generation antidepressant as initial monotherapy for acute moderate to severe major depressive disorder, with combination therapy an option. Current version: Qaseem, A., Owens, D.K., Etxeandia-Ikobaltzeta, I., et al. (2026), "Nonpharmacologic and Pharmacologic Treatments of Adults in the Acute Phase of Major Depressive Disorder: A Living Clinical Guideline From the American College of Physicians (Version 1, Update Alert 4)," Annals of Internal Medicine, published online 10 March 2026, DOI: 10.7326/ANNALS-26-00548, PMID: 41802255.)
Jacka, F.N., O'Neil, A., Opie, R., et al. (2017). "A randomised controlled trial of dietary improvement for adults with major depression (the 'SMILES' trial)." BMC Medicine 15:23. DOI: 10.1186/s12916-017-0791-y. (12 weeks, single-blind, adjunctive dietary intervention versus matched social support; Cohen's d approximately 1.16. Challenged: Molendijk, M.L., Fried, E.I., van der Does, W. (2018), "The SMILES trial: do undisclosed recruitment practices explain the remarkably large effect?", BMC Medicine 16:243, DOI: 10.1186/s12916-018-1221-5, PMID: 30591065; Jacka and colleagues responded in print, PMID 30591059. Wider intervention evidence is mixed, including a lifestyle programme combining diet and exercise that found no difference from attention control, and a large Mediterranean-diet cardiovascular trial reporting a non-significant depression signal.)
Asad, A., Kirk, M., Zhu, S., Dong, X., Gao, M. (2025). "Effects of Prebiotics and Probiotics on Symptoms of Depression and Anxiety in Clinically Diagnosed Samples: Systematic Review and Meta-analysis of Randomized Controlled Trials." Nutrition Reviews 83(7):e1504–e1520. DOI: 10.1093/nutrit/nuae177. PMID: 39731509. PMCID: PMC12166186. (23 randomised trials, 1,401 clinically diagnosed patients; pooled reduction in depressive symptoms with probiotics, with high heterogeneity including variation associated with intervention duration and formulation. Newer syntheses point in the same broadly promising but unsettled direction. The body claim is deliberately limited to the absence of a general prescription.)
Noetel, M., Sanders, T., Gallardo-Gómez, D., et al. (2024). "Effect of exercise for depression: systematic review and network meta-analysis of randomised controlled trials." BMJ 384:e075847. DOI: 10.1136/bmj-2023-075847. PMID: 38355154. Correction: BMJ 2024;385:q1024. (218 studies, 14,170 participants; confidence in most individual comparisons rated low, largely reflecting the impossibility of blinding.) Umbrella evidence: Munro, N.R., Teague, S., Somoray, K., et al. (2026), "Effect of exercise on depression and anxiety symptoms: systematic umbrella review with meta-meta-analysis," British Journal of Sports Medicine 60(8):590–599, DOI: 10.1136/bjsports-2025-110301, PMID: 41667154 — 63 reviews, 81 meta-analyses, 1,079 component studies, 79,551 participants, reporting reductions in both depression and anxiety symptoms.
Wilson, J., Dobson, O., Langcake, A., et al. (2026). "The efficacy and safety of cannabinoids for the treatment of mental disorders and substance use disorders: a systematic review and meta-analysis." Lancet Psychiatry 13(4):304–315. DOI: 10.1016/S2215-0366(26)00015-5. (5,774 studies screened; 54 randomised controlled trials, 2,477 participants, published 1980 to May 2025. No significant effects for anxiety, anorexia nervosa, psychotic disorders, PTSD and opioid use disorder; insufficient data to meta-analyse OCD, ADHD, bipolar disorder and tobacco use disorder; no randomised trials identified for depression. Cannabinoids increased cocaine craving versus placebo. Quality limitations stated by the authors: 24 of 54 trials at high risk of bias, 20% raising conflict of interest concerns, median trial enrolment approximately 31.5 participants. The review also reported low-certainty positive signals for cannabis use disorder, insomnia, tic disorders and autistic traits. Correction status resolved: the publisher records that this online publication has been corrected, with the corrected version first appearing on 16 June 2026. The article page at the DOI above now serves the corrected version, and every figure cited here was taken from it. Two subsequent Lancet Psychiatry papers in the July 2026 issue cite it at 13:304–315, matching this citation. Competing interests declared by the authors: two received consultation fees from WHO, one received payment for expert testimony on the risks of cannabis use, and one is an appointed member of the Australian Medicinal Cannabis Expert Working Group and has received Therapeutic Goods Administration funding for independent evidence reviews on medicinal cannabis. Disclosed on the principle applied throughout this library, including where a paper's findings run against commercial interests rather than with them. Not to be confused with the earlier Black, Stockings et al. cannabinoid meta-analysis, Lancet Psychiatry 2019;6:995–1010, which carries its own separate correction from December 2019.)
Petrilli, K., Ofori, S., Hines, L., Taylor, G., Adams, S., Freeman, T.P. (2022). "Association of cannabis potency with mental ill health and addiction: a systematic review." Lancet Psychiatry 9:736–750. DOI: 10.1016/S2215-0366(22)00161-4. (20 studies; higher potency cannabis associated with psychosis and cannabis use disorder, while results for anxiety and depression were mixed and sometimes contradictory.) Di Forti, M., et al. (2019). "The contribution of cannabis use to variation in the incidence of psychotic disorder across Europe (EU-GEI): a multicentre case-control study." Lancet Psychiatry 6:427–436. (Daily use associated with approximately three times the odds of psychotic disorder, higher for high-potency products. Case-control design; a published correspondence notes cannabis is frequently mixed with tobacco in Europe, which the analysis must account for.)
Hindley, G., et al. (2020). "Psychiatric symptoms caused by cannabis constituents: a systematic review and meta-analysis." Lancet Psychiatry 7(4):344–353. DOI: 10.1016/S2215-0366(20)30074-2. (A single THC administration induced psychotic, negative and other psychiatric symptoms with large effect sizes; CBD did not induce symptoms, with inconclusive evidence on whether it moderates THC effects. Effects larger with intravenous than inhaled administration.)
Odgers, C.L. (2024). "The great rewiring: is social media really behind an epidemic of teenage mental illness?" Nature 628:29–30. DOI: 10.1038/d41586-024-00902-2. (A review and critique, not a primary study.) On the limits of "screen time" as a variable: Orben, A., Meier, A., Dalgleish, T., Blakemore, S.-J. (2024), "Mechanisms linking social media use to adolescent mental health vulnerability," Nature Reviews Psychology 3:407–423.
Vuorre, M., Przybylski, A.K. (2023). "Estimating the association between Facebook adoption and well-being in 72 countries." Royal Society Open Science 10(8):221451. DOI: 10.1098/rsos.221451. (946,798 people across 72 countries; no evidence that Facebook adoption tracked widespread population-level psychological harm. The authors explicitly caution that the design cannot resolve individual causal effects.) See also Orben, A., Przybylski, A.K. (2019), "The association between adolescent well-being and digital technology use," Nature Human Behaviour 3:173–182.
Livingston, G., Huntley, J., Liu, K.Y., Costafreda, S.G., Selbæk, G., Alladi, S., et al. (2024). "Dementia prevention, intervention, and care: 2024 report of the Lancet standing Commission." The Lancet 404:572–628. DOI: 10.1016/S0140-6736(24)01296-0. (14 modifiable risk factors; approximately 45% of dementia cases potentially prevented or delayed. High midlife LDL cholesterol added in 2024 based on cohort studies of over one million participants plus a Mendelian randomisation meta-analysis of 27 studies, contributing an estimated 7% population attributable fraction. Population attributable fractions are population-level estimates and do not translate into individual guarantees.)
Mostert, C.M., et al. (2025). "Broadening dementia risk models: building on the 2024 Lancet Commission report for a more inclusive global framework." PMID: 41004922.
Alzheimer's disease and brain insulin resistance. de la Monte, S.M., Wands, J.R. (2008). "Alzheimer's disease is type 3 diabetes — evidence reviewed." Journal of Diabetes Science and Technology 2(6):1101–1113. PMID: 19885299. See also "A systematic review on type 3 diabetes: bridging the gap between metabolic dysfunction and Alzheimer's disease" (2025), Diabetology & Metabolic Syndrome, DOI: 10.1186/s13098-025-01930-2, and Kandimalla, R., Thirumala, V., Reddy, P.H., "Is Alzheimer's disease a type 3 diabetes? A critical appraisal." ("Type 3 diabetes" is not a recognised diagnostic category in any classification system.)
Qi, et al. (2025). "Prevalence of Chronic Traumatic Encephalopathy in Athletes With Repetitive Head Impacts: A Systematic Review and Meta-Analysis." Scandinavian Journal of Medicine & Science in Sports. DOI: 10.1111/sms.70047. (Eight studies, 1,000 former athletes; pooled prevalence 53.7% [95% CI 37.6–69.5], I² = 93%. All cases derive from donated brains, a self-selected sample. The review also reports a 39.0% suicide rate among diagnosed cases with 95% CI 12.7–69.0 and I² = 94.4%; that estimate is too imprecise and too subject to selection bias to quote meaningfully, and is not used in the body.)
Maroon, J.C., Winkelman, R., Bost, J., Amos, A., Mathyssek, C., Miele, V. (2015). "Chronic Traumatic Encephalopathy in Contact Sports: A Systematic Review of All Reported Pathological Cases." PLOS One 10(2):e0117338. (Concluded incidence remains unknown owing to the absence of large longitudinal studies, and that case reporting is inherently subject to selection bias.) Unselected brain bank comparison: McCann, H., Bahar, A.Y., Burkhardt, K., Gardner, A.J., Halliday, G.M., Iverson, G.L., Shepherd, C.E. (2022). "Prevalence of chronic traumatic encephalopathy in the Sydney Brain Bank." Brain Communications 4(4):fcac189. DOI: 10.1093/braincomms/fcac189. PMID: 35950093. PMCID: PMC9356727. (636 brains screened; five cases identified, 0.79%.)
McKee, A.C., Stein, T.D., Huber, B.R., Crary, J.F., Bieniek, K., Dickson, D., Alvarez, V.E., Cherry, J.D., Farrell, K., Butler, M., Uretsky, M., Abdolmohammadi, B., Alosco, M.L., Tripodis, Y., Mez, J., Daneshvar, D.H. (2023). "Chronic traumatic encephalopathy (CTE): criteria for neuropathological diagnosis and relationship to repetitive head impacts." Acta Neuropathologica 145(4):371–394. DOI: 10.1007/s00401-023-02540-w. PMID: 36759368. PMCID: PMC10020327.
Moitra, M., Santomauro, D., Collins, P.Y., Vos, T., Whiteford, H., Saxena, S., Ferrari, A.J. (2022). "The global gap in treatment coverage for major depressive disorder in 84 countries from 2000–2019: A systematic review and Bayesian meta-regression analysis." PLOS Medicine 19(2):e1003901. DOI: 10.1371/journal.pmed.1003901. PMID: 35167593. PMCID: PMC8846511. (Minimally adequate treatment 23% [95% UI 2–55%] in high-income countries versus 3% [95% UI <1–25%] in low and lower-middle-income countries. The authors note insufficient data from low and lower-middle-income countries as a primary limitation.)
Mekonen, T., Chan, G.C.K., Connor, J.P., Hides, L., Leung, J. (2021). "Estimating the global treatment rates for depression: A systematic review and meta-analysis." Journal of Affective Disorders 295:1234–1242. DOI: 10.1016/j.jad.2021.09.038. PMID: 34665135. (65 studies, 1.1 million participants, 79 countries; pooled treatment rate 34.8%, comprising 48.3% in high-income, 21.4% in middle-income and 16.8% in low-income countries. Not used in the body, which relies on the more meaningful "minimally adequate treatment" measure in reference 22.)

Sourcing note: this article cites peer-reviewed studies, systematic reviews and international commission reports with DOI/PMID where available. Where a scientific question is unresolved, as with the serotonin hypothesis, screen use and youth mental health, the type 3 diabetes framing, and CTE prevalence, both positions are cited rather than the more convenient one. Where a figure depends heavily on how a sample was selected, that is stated in the body rather than left in the reference. Population attributable fractions are identified as population-level estimates. Industry funding is disclosed where a cited trial carries it. This article makes no treatment recommendation and provides no dosing or tapering information. Cross-link targets (#) resolve once the site structure is live. Reading time is measured, not asserted: body words excluding the reference list and the metadata header, counted at 225 words per minute and rounded to the nearest minute.