What Is This?
In July 2022, a comprehensive umbrella review published in Molecular Psychiatry — one of the highest-impact journals in the field — concluded there is no consistent evidence that depression is caused by low serotonin levels. The "chemical imbalance" theory that has underpinned psychiatric practice since the 1980s and justified the prescription of SSRIs to hundreds of millions of people was, the authors concluded, not supported by the evidence.^1
This was not a fringe paper. It was authored by Joanna Moncrieff and colleagues at University College London, and it reviewed every major line of evidence for the serotonin hypothesis: serotonin metabolite levels, serotonin receptor studies, gene association studies, precursor depletion experiments. None of them supported the hypothesis. The authors didn't argue that SSRIs don't work (in some patients, they do). They argued the theoretical basis for why they were supposed to work was wrong.
The vacuum this created is being filled by a radically different framework: metabolic psychiatry. The central claim, advanced primarily by Harvard psychiatrist Chris Palmer and Stanford's Shebani Sethi, is that many psychiatric conditions — including schizophrenia, bipolar disorder, and treatment-resistant depression — are disorders of brain metabolism, not neurotransmitter signalling. The brain is the most energy-hungry organ in the body, consuming 20% of total calories. If cellular energy metabolism is impaired — particularly in the mitochondria that power neurons — the consequences express as psychiatric symptoms.^2
The intervention at the frontier of this model: the ketogenic diet. Not as a lifestyle choice, but as a metabolic therapy targeting brain energy metabolism directly.
Why Does It Matter?
- The serotonin hypothesis collapse is bigger than most people realise. The "chemical imbalance" framing was not just a scientific theory — it was the entire cultural and commercial basis for modern psychiatry. It justified first-line SSRI prescriptions (now the most commonly prescribed drug class in the UK), built a $50B+ global antidepressant market, and shaped how millions of people understand their own minds. If the foundational theory is wrong, the treatment system built on it needs rethinking from the ground up.
- The clinical trial results from keto for psychosis are striking. A 2024 Stanford pilot study by Sethi and colleagues gave a ketogenic diet to patients with schizophrenia or bipolar disorder who had metabolic problems (a common side effect of antipsychotic drugs). After four months: 79% had clinically meaningful psychiatric improvement, 43% met criteria for "clinical recovery." These are patients for whom existing medications had already failed to produce adequate outcomes.^3
- The mitochondria connection explains the psychiatric-metabolic comorbidity. People with serious mental illness die on average 10–25 years earlier than the general population, primarily from metabolic diseases (heart disease, diabetes). The standard explanation was "poor lifestyle choices." The metabolic psychiatry explanation: both the psychiatric condition and the metabolic disease share a common cause — impaired mitochondrial function and cellular energy metabolism. They're not separate problems with lifestyle linking them; they're the same underlying problem expressing differently.
- It changes what "treatment" means. If depression is a metabolic disorder, then diet, exercise, sleep, and metabolic interventions (including, in extreme cases, ketogenic therapy) are first-order treatments — not adjuncts to medication. This doesn't mean medication is useless; it means the model that confines treatment to pharmacology is incomplete and potentially harmful when it crowds out metabolic interventions that may be more effective.
- The dietary data is no longer anecdotal. RCTs showing Mediterranean diet produces better depression outcomes than social support alone. Studies showing exercise matches SSRIs in mild-moderate depression. Keto trial results at Stanford and now multiple other centres. The evidence base has reached a point where ignoring diet as a psychiatric intervention requires active effort.
Key People & Players
Chris Palmer (Harvard Medical School) — Psychiatrist and neuroscientist who has spent 25+ years studying the intersection of metabolism and mental illness. Author of Brain Energy (2022), the first mainstream book to lay out the metabolic psychiatry thesis comprehensively. Has treated patients with schizophrenia, bipolar, and severe depression using ketogenic diet, with documented remissions. His public positioning is deliberately moderate — he doesn't claim keto cures everything, he claims metabolism is a lever that conventional psychiatry ignores.^4
Shebani Sethi (Stanford University) — Founding director of Stanford's Metabolic Psychiatry Clinic, the first dedicated metabolic psychiatry clinical programme at a major academic medical centre. Ran the pilot trial showing keto produces clinically meaningful improvement and recovery in schizophrenia and bipolar patients. Expanding to randomised controlled trials.^5
Joanna Moncrieff (UCL) — Lead author of the 2022 serotonin umbrella review. Her work didn't directly argue for metabolic psychiatry, but it cleared the ground by definitively challenging the theoretical basis of the dominant pharmacological model.^6
Georgia Ede (Harvard) — Psychiatrist and nutritional researcher, author of Change Your Diet, Change Your Mind. Has written the most accessible clinical account of dietary interventions for psychiatric conditions, grounded in the metabolic framework.^7
Nicholas Norwitz (Oxford/Harvard) — Young researcher who has been a collaborator on keto-psychiatry research and has documented his own use of ketogenic diet for bipolar disorder. Bridges the academic and public communication sides of the field.^8
The Current State
Metabolic psychiatry is still early-stage but moving fast. The evidence has crossed the threshold from anecdote to clinical trial. The resistance from mainstream psychiatry is substantial — paradigm shifts in medicine are slow, and an evidence base suggesting that dietary intervention rivals pharmaceutical treatment creates obvious commercial and institutional friction.
Where the evidence stands:
- Ketogenic diet for schizophrenia/bipolar: Multiple pilot trials showing meaningful psychiatric improvement. One RCT protocol registered (Frontiers in Nutrition, 2024). Not yet enough for clinical guidelines, but the signal is strong.^9
- Exercise for depression: More robust evidence base than keto. Multiple meta-analyses showing effects comparable to SSRIs in mild-moderate depression. Exercise has no commercial backer pushing it through clinical trials — which partly explains why the evidence is less prominent than its size warrants.
- Mediterranean diet RCT (SMILES trial, 2017): Diet intervention outperformed social support on depression scores and was more cost-effective. The first RCT showing diet change can treat depression. Since replicated.
- Mitochondrial dysfunction research: Growing evidence that mitochondrial abnormalities are present across multiple psychiatric diagnoses. Drugs that improve mitochondrial function (some already in trials for other conditions) are being investigated for psychiatric indications.
The most significant open question: why does ketosis specifically help some psychiatric patients? The leading hypotheses involve reduced neuroinflammation, alternative energy substrate for struggling neurons, stabilisation of neural excitability via GABA/glutamate balance, and improved mitochondrial function. The mechanism is not yet settled.
Practical implications right now:
Most of what's established is negative: ultra-processed food, alcohol, and inadequate sleep reliably worsen psychiatric outcomes. The positive interventions with the best evidence base (exercise, sleep quality, Mediterranean-style diet) are already well-established general health recommendations. For someone managing or preventing psychiatric conditions, these aren't controversial. The frontier (ketogenic therapy for serious psychiatric conditions) is where clinical guidance is still being established.
Best Resources to Learn More
- Brain Energy by Chris Palmer — The comprehensive case for the metabolic model of mental illness. Balanced, well-sourced, and careful about distinguishing established evidence from emerging hypothesis.^10
- Stanford Metabolic Psychiatry Clinic — The clinical programme running the trials. Research publications listed.^11
- Moncrieff et al. 2022 — Serotonin umbrella review — The paper that challenged the serotonin hypothesis. Read the abstract; the full paper is technical.^1
- NPR Health Shots: Keto's new frontier — mental illness (2024) — The best mainstream journalism on the Stanford trials.^12
- Change Your Diet, Change Your Mind by Georgia Ede — More accessible and dietary-focused than Palmer's book. Good companion.^13