Metabolic psychiatry is the name given to a line of research asking whether some mental illness involves how brain cells produce and use energy — and, if so, whether changing the fuel the brain runs on might make a difference.
It is a serious question being asked by serious people. It is also, at this stage, a hypothesis. Nothing below is established, and nothing below is a treatment.
The idea in plain terms
Your brain is expensive to run. It is roughly 2% of your body weight and uses something like 20% of your energy, and it normally runs on glucose.
When carbohydrate is scarce for long enough, the liver starts making ketones from fat, and the brain can run on those instead. This is ordinary physiology, not a trick — it happens during fasting, and it is the basis of the ketogenic diet that has been used in epilepsy since the 1920s.
The epilepsy connection is where this whole field comes from. A diet used for decades to reduce seizures turns out to change brain energy metabolism, and several psychiatric medicines — sodium valproate, lamotrigine — started life as anticonvulsants. That is the thread researchers are pulling.
What is actually proposed
Several things, at different levels of confidence:
- That mitochondria, the parts of cells that generate energy, may function differently in some psychiatric conditions.
- That ketones might provide a more efficient or more stable fuel for brains where glucose metabolism is impaired.
- That ketogenic eating affects inflammation, oxidative stress and neurotransmitter balance in ways that might be relevant.
Each of those is a reasonable research question. None of them is a finding you can rely on, and the distance between “this is biologically plausible” and “this helps patients” is where most promising medical ideas die.
What has been tested so far
Not much, and mostly not in depression.
The most substantial trials are in bipolar disorder and schizophrenia. A Stanford pilot put twenty-three people with those conditions on a ketogenic diet for four months and reported meaningful improvements — but everybody got the diet, so there is nothing to compare it against.
In depression specifically there is one randomised trial, published in early 2026, in people whose antidepressants had not worked. The ketogenic group did better at six weeks by about two points on a depression questionnaire — a result that only just reached statistical significance and had disappeared by twelve weeks. The authors described its clinical relevance as uncertain.
Beside that sits a case series describing three patients at a private clinic.
The evidence page sets all of this out properly, with what each study can and cannot tell you.
Why it attracts so much overstatement
Because the story is good. A diet, not a drug. An explanation that feels mechanical and concrete in a field where explanations are usually fuzzy. Striking numbers from small studies. And a steady supply of people willing to sell you something on the strength of it.
The honest position is that metabolic psychiatry is a legitimate and interesting research area that has not yet produced a treatment, and that anyone telling you otherwise has got ahead of the evidence.
Where to go next
- Why a bipolar finding is not a depression finding
- What this way of eating involves
- What the evidence shows
General education, not a diagnosis and not a treatment. If your mood has been low for weeks, your GP is the next step.
Explore the guides