From Epilepsy to Psychiatry: The Metabolic Bridge

Imagine a mechanic who is called out because a car keeps stalling at junctions. She fixes it. A few weeks later the same driver returns with a different complaint: the headlights flicker and the radio cuts out. Most people would treat these as three separate faults with three separate repairs. The mechanic checks the alternator. It turns out the stalling and the flickering were never separate problems at all — they were two visible symptoms of one unreliable power supply. The symptoms differed because they appeared in different systems. The cause was shared.

This is, in compressed form, the argument that carried the ketogenic diet from neurology into psychiatry. The classical ketogenic diet was formalized in the 1920s as a deliberate attempt to reproduce the metabolic effects of fasting, which physicians had observed could dramatically reduce seizure frequency in children whose epilepsy did not respond to available drugs. What the diet does is change the brain's dominant fuel. Instead of running primarily on glucose, the brain shifts toward ketone bodies — chiefly beta-hydroxybutyrate and acetoacetate — produced by the liver from fat. That single change in fuel turned out to have a cascade of downstream consequences.

Decades of epilepsy research have mapped those consequences reasonably well. Ketosis appears to stabilize neuronal membranes, promote mitochondrial biogenesis (the production of more, and more efficient, cellular power plants), reduce oxidative stress, and nudge neurotransmitter synthesis toward a more inhibitory, GABAergic balance. Each of these converges on the same endpoint: a neural network that is harder to destabilize. And here is the observation that opened the door to psychiatry. None of those mechanisms is specific to epileptic circuits. They are general properties of how neurons handle energy.

Three dimensions deserve a closer look. The first is metabolic flexibility — the ability of neurons to switch cleanly between fuels as circumstances change, rather like a hybrid engine that moves between power sources without hesitation. In several psychiatric conditions, researchers find that this flexibility is compromised: chronic hyperinsulinemia, rigid dependence on glucose, and low-grade neuroinflammation leave neurons less able to absorb metabolic stress. The second dimension is that ketones are not passive fuel. Beta-hydroxybutyrate inhibits histone deacetylases, enzymes that regulate which genes are read. In other words, the fuel also leaves instructions — it influences gene expression relating to neuronal survival, adaptation, and inflammation. The third dimension is the evidence gradient itself, which is steep. In epilepsy the diet is supported by controlled trials and a long clinical record. In psychiatry the landscape is mostly pilot studies, case series, and open-label work in depression, bipolar disorder, and schizophrenia. These early reports are frequently encouraging. They are not yet proof.

In ordinary working life, the metabolic lens tends to show up as a shift in the questions people ask. Someone who has always described their mid-afternoon fog, irritability, and flattened motivation as a mood or willpower issue starts noticing that these symptoms track with sleep, meals, and physical activity rather than with events. The relevant clinical background is that metabolic syndrome, insulin resistance, and markers of inflammation are unusually common in psychiatric populations. That does not tell you what is causing what in any individual, and it certainly does not diagnose anyone. It does suggest that energy is a variable worth measuring rather than assuming.

Several misreadings recur. The first is treating mechanistic overlap as clinical transfer: because ketosis stabilizes excitable networks in epilepsy, it must therefore treat depression. Psychiatric conditions are heterogeneous, and a plausible pathway is a hypothesis, not a result. The second is imagining the diet as a straightforward alternative to medication. Ketosis can alter fluid and electrolyte balance and interacts with the handling of mood stabilizers and antipsychotics, which is precisely why medication changes belong with the prescribing clinician and nowhere else. Ketogenic therapy is also unsuitable for a number of people, including those with certain inherited metabolic disorders, type 1 diabetes, a history of disordered eating, and those who are pregnant. The third misreading is expecting the transition to feel uniformly good. Rapid metabolic shifts can temporarily worsen anxiety or disrupt sleep — an argument for supervision, not for panic.

Here is the counterintuitive part. The most valuable export from epilepsy to psychiatry may not be the diet at all. It may be the infrastructure built around it: structured initiation, real-time monitoring of glucose and ketones, systematic tracking of side effects, caregiver involvement, and slow deliberate medication adjustment. Neurology learned to treat a dietary intervention with the same seriousness as a drug. Psychiatry, which has often filed nutrition under lifestyle advice, is now importing that discipline — and the discipline may prove more transformative than the menu.