is depression a chemical imbalance? what the 2022 review, the replies, the scans and the labels actually say

somebody said it to you in a small room, probably kindly: depression is a chemical imbalance, the chemical is serotonin, and the tablet puts it back. it is one of the most widely believed sentences in medicine. is it true?

here is the record as of october 2026: the two 1960s papers where the idea started, the 2022 umbrella review that went looking for the evidence, the reply from thirty-five researchers in the same journal and the authors' answer, the newest brain scan and the depletion studies both sides argue over, what the trials say about whether the drugs work regardless, what the drug's own label says about mechanism, two experiments on what the explanation does to the people who hear it, and what the national institute and the psychiatrists' association say today, quoted and dated. both directions, every number with its source.

where did the idea come from?

from the drugs, and from two review papers. in 1965 joseph schildkraut published “the catecholamine hypothesis of affective disorders: a review of supporting evidence” in the american journal of psychiatry^[1]. the word in the title is hypothesis. it proposed that the catecholamines, noradrenaline among them, were involved in depression and mania, reasoning from how the first antidepressants acted rather than from a measurement in patients. in 1967 alec coppen's review in the british journal of psychiatry extended the idea to serotonin^[2]. his own abstract describes the evidence as “by no means conclusive” while arguing that biochemical changes are “most important in the aetiology of affective disorders.” neither paper measured a deficiency in a patient. both offered a hypothesis to be tested.

the 2022 review puts the timeline in one sentence: a link between lowered serotonin and depression “was first suggested in the 1960s, and widely publicised from the 1990s with the advent of the Selective Serotonin Reuptake Inhibitor (SSRI) antidepressants”^[3]. it cites surveys suggesting “80% or more of the general public now believe it is established that depression is caused by a ‘chemical imbalance’”^[3]. the order matters and it is the subject of did psychiatry work backwards from its drugs?: the drugs came first, the theory was fitted to them, and the public message came thirty years after the hypothesis.

what does the serotonin theory actually claim?

the claim worth testing is narrow: that depression is caused by too little serotonin, or too little serotonin activity, in the brain, and that reuptake inhibitors work by correcting it. the authors of the 2022 review describe it as “the idea that depression is caused by low levels or low activity of serotonin”^[5]. that is different from three weaker claims that are often swapped in when the strong one is challenged: that serotonin is involved in mood somehow; that drugs acting on serotonin can help; and that depression has a biology. the first two are not what anyone was told in the small room, and the third is not in dispute on this page. keep the strong claim in view, because it is the one the public believes and the one the evidence below is about.

what did the 2022 umbrella review find?

in july 2022 six researchers led by joanna moncrieff published a systematic umbrella review in molecular psychiatry, pre-registered with prospero, covering the main areas of serotonin research from the start of the literature to december 2020^[3]. it included 17 studies: 12 systematic reviews and meta-analyses, one collaborative meta-analysis, one meta-analysis of large cohort studies, one narrative synthesis, one genetic association study and one earlier umbrella review. the areas were serotonin and its metabolite 5-hiaa in body fluids, 5-ht1a receptor binding, serotonin transporter (sert) levels by imaging or at post-mortem, tryptophan depletion studies, sert gene associations, and sert gene-environment interactions^[3].

what it found, area by area, in its own words. 5-hiaa: two meta-analyses of overlapping studies “showed no association with depression (largest n = 1002).” plasma serotonin: one meta-analysis of cohort studies “showed no relationship with depression, and evidence that lowered serotonin concentration was associated with antidepressant use (n = 1869).” receptor and transporter binding: the 5-ht1a meta-analyses (largest n = 561) and three sert meta-analyses (largest n = 1845) “showed weak and inconsistent evidence of reduced binding in some areas, which would be consistent with increased synaptic availability of serotonin in people with depression, if this was the original, causal abnormaly” (their spelling), with the caveat that “effects of prior antidepressant use were not reliably excluded.” tryptophan depletion: “no effect in most healthy volunteers (n = 566), but weak evidence of an effect in those with a family history of depression (n = 75).” genes: the two largest studies, one of 115,257 people and one of 43,165, “revealed no evidence of an association with depression, or of an interaction between genotype, stress and depression”^[3].

the conclusion, verbatim from the abstract: “The main areas of serotonin research provide no consistent evidence of there being an association between serotonin and depression, and no support for the hypothesis that depression is caused by lowered serotonin activity or concentrations. Some evidence was consistent with the possibility that long-term antidepressant use reduces serotonin concentration”^[3]. the review had no specific funding. its declared interests belong on the page: one author co-founded a company that helps people stop antidepressants, and the lead author receives royalties on books about psychiatric drugs^[3].

what did the replies say?

in june 2023 molecular psychiatry published a reply signed by thirty-five researchers, titled “a leaky umbrella has little value”^[4]. its abstract says the review's conclusion is “overstated, including methodological weaknesses in the review process, selective reporting of data, over-simplification, and errors in the interpretation of neuropsychopharmacological findings.” its strongest points are specific. on tryptophan depletion, the reply says the review quoted a meta-analysis for its null result in healthy people but left out that same meta-analysis's estimate for depressed people not taking antidepressants, which from eight samples “was large (Hedge's g = −1.9 (95% CIs −3.02 to −0.78),” falling to −1.06 when an outlier is removed; the reply concedes the intervals were wide and several samples came from one group^[4]. it adds that plasma tryptophan, which the review did not cover, is lower in depression (g = −0.45, and −0.84 in unmedicated people). on imaging, it argues the review assumed 5-ht1a receptors are presynaptic autoreceptors when most are postsynaptic, so that lower binding could mean lower serotonin signalling rather than more; and that 149 of 364 people in one sert meta-analysis were drug-naive, so prior antidepressant use cannot explain that finding^[4].

its own conclusion: “acute tryptophan depletion and decreased plasma tryptophan in depression indicate a role for 5-HT in those vulnerable to or suffering from depression, and that molecular imaging suggests the system is perturbed. The proven efficacy of SSRIs in a proportion of people with depression lends credibility to this position”^[4]. notice what that sentence does and does not claim: a role, a perturbed system, vulnerability. it does not say depression is caused by low serotonin. several of the reply's authors declare honoraria or consultancy from drug manufacturers, and one lists a manufacturer as an affiliation; all are declared in the paper^[4].

the review's authors answered in the same issue^[5]. we could read only the first paragraph, because the rest is paywalled, so that is all we quote: “There are always further studies that can be presented to support the serotonin theory, and research on other possible biological mechanisms, which may or may not involve serotonin. However, we synthesised all existing studies from the main areas of research on serotonin and depression. The onus is on others to present a convincing case based on other types of research or alternative hypotheses, not on us to disprove every speculation”^[5]. that is the shape of the whole argument. one side says: we looked where the theory said to look and found nothing consistent. the other says: you looked in the wrong places and misread what you found. a reader does not have to pick; the two things both sides agree on are that the strong public claim is not established, and that whether the drugs work is a different question.

is there newer direct evidence?

one scan, small. in a study published online in october 2022 and in print in biological psychiatry in june 2023, david erritzoe and colleagues used a pet tracer sensitive to serotonin release, giving 17 antidepressant-free patients in a major depressive episode and 20 healthy controls an oral dose of d-amphetamine and measuring the change in tracer binding^[6]. in the controls binding fell by 15% (± 14%); in the patients by 6.5% (± 20%); p = .041. the authors call it “clear evidence for dysfunctional serotonergic neurotransmission in depression.” read the design before the adjective: 37 people, 31 of them men, controls a decade younger on average (32 against 44), one protocol, no replication yet, and a p-value that would not survive a second comparison^[6]. it is the first direct measurement of serotonin release in depression and it points the theory's way. it is also exactly the kind of result the 2022 review would have graded as weak until it is repeated. several of its authors signed the “leaky umbrella” reply^[4],[6].

the depletion studies are the other battleground, and the 2007 meta-analysis both sides cite is clear on what they show^[7]. it identified 73 tryptophan-depletion studies and pooled 45. “5-HT or NE/DA depletion did not decrease mood in healthy controls.” it “slightly lowered mood in healthy controls with a family history” of depression, produced “a moderate mood decrease” in drug-free patients in remission, and “induced relapse in patients with MDD in remission who used serotonergic antidepressants.” the authors' reading: the studies “fail to demonstrate a causal relation. They presumably clarify a vulnerability trait to become depressed”^[7]. in plain terms: taking serotonin away from a healthy person does not make them depressed; taking it away from someone whose remission depends on a serotonergic drug can bring the depression back. that is evidence that the drug's effect runs through serotonin. it is not evidence that the illness started there.

does it matter whether the theory is right if the drugs work?

for the question of whether to use them, no, and both sides say so. whether a drug beats placebo and why it does are separate questions answered by separate trials. on the first, the largest comparison is the 2018 lancet network meta-analysis of 522 double-blind randomised trials and 116,477 adults^[8]: “all antidepressants were more effective than placebo,” with odds ratios for response ranging from 2.13 for amitriptyline to 1.37 for reboxetine, over a median of about eight weeks. the same abstract rates 46 (9%) of the trials at high risk of bias and 380 (73%) at moderate, and the certainty of evidence “moderate to very low”^[8]. a 2008 meta-analysis of every trial submitted to fda for four newer antidepressants, published and unpublished, found that “drug-placebo differences in antidepressant efficacy increase as a function of baseline severity, but are relatively small even for severely depressed patients”^[9]. so: a real average benefit, modest in size, from short trials. none of it depends on serotonin being the cause.

the label agrees, by saying less than people assume. section 12.1 of the zoloft prescribing information, and of the generic sertraline label effective 27 august 2026, is one sentence: “Sertraline potentiates serotonergic activity in the central nervous system through inhibition of neuronal reuptake of serotonin (5-HT)”^[10]. that is the whole of the mechanism section. it describes what the molecule does to a transporter. it does not say the person taking it has too little serotonin, and it does not say the drug corrects an imbalance. the gap between that sentence and “you have a chemical imbalance” is the gap this page is about. what the label does say about warnings and side effects is on what the sertraline warnings mean.

what does the explanation do to the people who hear it?

the chemical-imbalance explanation was offered partly as kindness: if the fault is chemistry, it is not you. two experiments tested whether it works that way^[11],[12]. in 2014, kemp, lickel and deacon gave people who reported a past or current depressive episode the results of “a bogus but credible biological test” telling them their symptoms were, or were not, caused by a chemical imbalance^[11]. the result, from the abstract: “chemical imbalance test feedback failed to reduce self-blame, elicited worse prognostic pessimism and negative mood regulation expectancies, and led participants to view pharmacotherapy as more credible and effective than psychotherapy.” the authors call the effects “unhelpful and potentially iatrogenic.”

the same year, lebowitz and ahn published in pnas a series of studies in which us clinicians read descriptions of patients whose symptoms were explained either biologically or psychosocially^[12]. the abstract sets up the hoped-for direction and then reports the opposite: “Biological explanations have been thought to make patients appear less accountable for their disorders, which could increase clinicians' empathy. To the contrary, biological explanations evoked significantly less empathy.” the authors add that “the ongoing shift toward biomedical conceptualizations has many benefits” and that these are “unintended negative consequences.” both are single experiments with vignettes and a deception design, not consultations, and both abstracts report direction rather than size, so no effect sizes appear here. what they show is that the kindness is not free: in the people tested, the explanation did not lift blame, and it lowered hope.

what do the big institutions say today?

we read both pages on 12 october 2026 and quote them as printed. the national institute of mental health's depression page has one sentence on causes: “Research suggests that genetic, biological, environmental, and psychological factors play a role in depression”^[13]. there is no other causal sentence on the page and the words “chemical imbalance” and “serotonin” do not appear in it.

the american psychiatric association's public page lists four risk factors^[14]. the first: “Biochemical: Differences in certain chemicals in the brain (such as the neurotransmiters serotonin, dopamine and norepinephrine) may contribute to symptoms of depression.” then: “Genetic: Depression can run in families. For example, if one identical twin has depression, the other has a 70 percent chance of having the illness sometime in life.” “Personality: People with low self-esteem, who are easily overwhelmed by stress, or who are generally pessimistic may be more likely to experience depression.” “Environmental: Continuous exposure to violence, neglect, abuse or poverty can also pose risks for developing depression”^[14]. read the verbs. “may contribute to symptoms” is a long way from “is caused by.” the profession's own public page, in 2026, does not make the claim its patients were told. the pages carry no change log, so we cannot say when the wording became this careful; we can say what it is today.

what to ask

these are questions, not advice. they are the ones the record above makes it reasonable to ask.

• when you say this medication helps, what outcome are we expecting to see, and by when?
• what is the evidence that it works for my situation, separate from the explanation of how it works?
• if the explanation i was given was a chemical imbalance, what would you say instead now?
• what is the plan for reviewing whether i still need it, and how often?
• if i ever want to stop, what taper and check-ins would you want in place first?
• what else, besides medication, has evidence for someone in my position?

adam