The Story Everyone Knows
Almost everyone knows what causes depression: the brain runs low on serotonin, a chemical messenger, and the resulting “chemical imbalance” drags mood down with it, while antidepressants fix the imbalance the way insulin fixes diabetes by topping up what is missing. Surveys put public belief at 85 to 90 percent, and for decades the story lived in textbooks, on clinic walls, and in drug advertisements because it is simple, destigmatizing, and convenient for everyone involved.
There is just one problem. When researchers finally gathered every major line of evidence and checked whether depressed people actually have less serotonin, the answer came back empty.
The Review That Read Everything
In July 2022, Molecular Psychiatry published the reckoning: Joanna Moncrieff, a UCL psychiatry professor, and five colleagues had conducted an umbrella review, a systematic survey of the systematic reviews, with the protocol registered in advance, three databases searched through December 2020, and 361 publications screened to the 17 strongest syntheses. Six research domains and tens of thousands of participants, all aimed at one question: is depression associated with lowered serotonin?
Domain by domain, the signal failed to appear: two meta-analyses of the serotonin metabolite 5-HIAA in body fluids found no association with depression (largest n=1,002). A meta-analysis of cohort studies measuring blood plasma serotonin found no relationship either (n=1,869), and the one pattern it detected ran backward: lower serotonin tracked antidepressant use, not depression. Two meta-analyses of the 5-HT1A receptor (largest n=561) and three of the serotonin transporter SSRIs target (largest n=1,845) came back weak and inconsistent, and where receptor binding did look reduced, it implied more synaptic serotonin in depression, the opposite of the theory’s prediction. Tryptophan-depletion experiments failed to induce depression in healthy volunteers (n=566), a separate systematic review (n=342) plus ten later studies (n=407) agreed, and the genetic evidence was the most decisive of all: a study of 115,257 people and a collaborative meta-analysis of 43,165 found no link between the serotonin transporter gene and depression, and no interaction between the gene, stress, and depression.
Their conclusion was blunt. The main areas of serotonin research, they wrote, provide no consistent evidence linking serotonin to depression, and no support for the hypothesis that depression is caused by lowered serotonin activity or concentrations. They added an underreported twist: some evidence suggested long-term antidepressant use itself reduces serotonin concentration, raising the possibility that the drugs create the deficiency they supposedly correct.
The Arithmetic of Belief
The UCL press release noted that one in six adults in England is prescribed an antidepressant in a given year, which, against roughly 44 million adults, works out to about seven million people. If 85 percent hold the chemical-imbalance belief, as surveys indicate, then more than 6 million people in England alone take a drug whose advertised mechanism has no consistent evidence behind it. Prozac reached the market in 1987. Prescriptions climbed steeply through the 1990s and 2000s on the strength of this story, building thirty-five years of prescribing on a hypothesis that never cleared its evidence bar.
The Strongest Counterargument
Pushback was swift and heavyweight, and it deserves a full hearing. Their charges were methodological. In 2023, thirty-five researchers led by Sameer Jauhar published a formal rebuttal in the same journal, “A leaky umbrella has little value,” arguing that Moncrieff’s team had not conducted a conventional umbrella review at all, since imaging, genetic, and depletion studies cannot be judged under one verdict, that the team had excluded relevant studies, and that it had underappreciated the complexities of neuroscience. A separate reply argued serotonin remains a crucial but dynamic underpinning of depression that older methods may have been too crude to catch, while psychopharmacologist David Nutt publicly called for the paper’s retraction, which the journal declined to issue.
A second line of defense does not dispute the review so much as its target: when The New York Times covered the controversy in November 2022, leading psychiatrists called the chemical-imbalance story an “old theory” the field had discarded long ago, and said it hardly matters because the drugs work regardless. That claim has formidable evidence behind it. A 2018 Lancet network meta-analysis by Andrea Cipriani and colleagues pooled 522 double-blind trials with 116,477 patients, finding all 21 antidepressants more efficacious than placebo, with odds ratios between roughly one and a half and two. Writer and psychiatrist Awais Aftab added a sharper critique, arguing that Moncrieff, who co-chairs the Critical Psychiatry Network, had attacked a vague public misperception in order to undermine the drugs themselves, though the literature still suggests serotonin plays some role in mood regulation.
Both camps leave the table uncomfortable. Moncrieff’s side is right that the thirty-year deficiency story is not supported by the evidence, while the critics are right that this does not make serotonin irrelevant or the drugs ineffective. What neither side can supply is the mechanism. If depression is not low serotonin, nobody has a satisfying account of what 21 effective drugs are actually doing, making Moncrieff’s own sentence, “We do not understand what antidepressants are doing to the brain exactly,” the most honest in the debate.
What We Didn't Prove
An umbrella review cannot prove a negative with finality, as Moncrieff herself states, and absence of consistent evidence is not proof of absence, particularly when the review’s own quality ratings were sobering: eleven of seventeen included syntheses met fewer than half the methodological quality criteria. The review says nothing about whether antidepressants work. Cipriani’s 522-trial analysis says they do, modestly, and it does not rule out serotonin playing some role in mood regulation, only the simple deficiency story. Bipolar depression, postpartum depression, and depression with physical illness were excluded; the observational evidence cannot fully exclude confounding by prior antidepressant use; with allegiance cutting both ways between a prominent critic of mainstream psychiatry and thirty-five of its defenders.
The Bottom Line
The slogan is dead and the mystery is not. The most repeated explanation in psychiatry, that depression is a serotonin deficiency corrected by pills, has no consistent evidence behind it after decades of searching. The pills still beat placebo, and nobody can fully explain why. That gap between “it works” and “we know why” is where honest psychiatry now lives; patients deserve to hear it plainly instead of the insulin analogy.
What You Can Do
If you take an antidepressant, do not stop because of this article. The review’s authors explicitly warn against abrupt discontinuation, and withdrawal can be severe. Any change to medication goes through your prescriber.
At your next appointment, ask what the drug is believed to be doing, what the evidence for that mechanism is, and what the plan for eventually stopping looks like. Honest doctors will tell you the mechanism is uncertain, and that uncertainty is information, not failure.
Know the alternatives with trial evidence: structured psychotherapy and regular exercise both have credible support for depression, and neither requires believing a story about your serotonin levels. If you are a clinician, a teacher, or anyone who explains depression to others, retire the insulin analogy. The review cites evidence that believing in a fixed chemical defect predicts a more pessimistic outlook on recovery, which means the slogan may actively make the prognosis feel worse.