Ketogenics

Notes on the ketogenic diet: what you eat, why it might work, and what you do alongside it. Sources are listed at the end.

The idea

The core move is to restrict carbohydrate below about 50 grams a day.The core move is to restrict dietary carbohydrate below about 50 grams a day — for perspective, less than one plain bagel — so the body's stored glycogenglycogen, the compact reserve of carbohydrate kept in the liver and muscles — a day's worth at most — runs low and it must find another fuel. In that state, called nutritional ketosis, the liver converts fat into ketone bodies.In that state, called nutritional ketosis, the liver converts fat into ketone bodies — chiefly beta-hydroxybutyratebeta-hydroxybutyrate, the main one, which powers brain and muscle and carries most of the claimed signaling benefits,, acetoacetateacetoacetate, the first-made ketone, which the body converts into the other two, and acetoneacetone, the volatile leftover that gives keto breath its fruity smell — which circulate in the blood and can be measured. Ketosis is not ketoacidosis.Ketosis is not ketoacidosis: nutritional ketosis holds blood ketones around 0.5 to 5 millimolarmillimolar, a unit of concentration; for scale, most meters call 0.5 the threshold of "in ketosis",, while diabetic ketoacidosisdiabetic ketoacidosis, seen almost only in type 1 diabetes, is a medical emergency — insulin is absent, so ketones and blood sugar spiral upward together.

What you eat

A ketogenic plate is built on fat, adequate protein, and green vegetables.A ketogenic plate is built on fat, adequate protein, and non-starchy green vegetables, with sugars, grains and most fruit excluded entirely.

Fat carries most of the calories.Fat carries most of the calories — roughly 70 to 80 percent of them: olive oil, butter, cream, avocados, nuts, seeds, and fatty cuts of meat and fish.
Protein stays moderate.Protein stays moderate, around 15 to 25 percent of calories — meat, poultry, fish, eggs, cheese — because excess protein can be converted to glucose in the liver via gluconeogenesisgluconeogenesis, the same process that keeps your blood sugar steady overnight.
Carbohydrate is nearly cut out.Carbohydrate is nearly cut out, under 20 to 50 net grams a day, coming mostly from leafy greens, broccoli, zucchini, mushrooms, and a handful of berries now and then.
Drinks are water, coffee and tea.Drinks are water, coffee and unsweetened tea; juice, soda and beer are out, and dry wine or spirits are the occasional exception.

The excluded list is everything sugary or starchy.The excluded list is everything sugary or starchy — bread, pasta, rice, potatoes, beans, most fruit, sweets, and sweetened drinks — plus anything labeled "low fat," which usually means sugar was added back.

There are several variants of the diet.There are several variants of the diet: the classic therapeutic version at 80 to 90 percent fat, used for epilepsy under strict medical supervision; the standard version around 70 to 75 percent; targeted keto, which adds carbs around workouts; and lazy keto, which skips the ratios and just counts carbs.

A fuller food list — what to eat, what to count, what to skip — has its own page.A fuller food list — what to eat, what to count, what to skip — has its own page: ketogenic foods, with the carb numbers behind each verdict.

Two Thai recipes built on the same slow-roasted lamb live under recipes.Two Thai recipes built on the same slow-roasted lamb live under recipes: stir-fried rice noodles with kale and fresh rice paper rolls — neither is keto-strict, since both lean on rice, but the lamb prep is shared and good.

The theories

The original theory came from fasting.The original theory came from fasting: fasting was known for millennia to control seizures, and in 1921 the Mayo Clinic designed a diet that mimics fasting's chemistry so patients could keep eating.

Modern theories cluster around fuel, signal, and appetite.Modern theories cluster around fuel, signal, and appetite — three ideas: ketones as an alternative fuel, ketones as a signaling molecule, and the diet's effect on appetite.

The fuel theory says some brains run better on ketones.The fuel theory says some brains run better on ketones than on glucose: in epilepsy, and possibly in some migraine, Alzheimer's and other conditions, brain cells that mishandle glucose can still burn beta-hydroxybutyrate.
The signal theory says ketones change gene expression.The signal theory says ketones change gene expression: beta-hydroxybutyrate is a signaling molecule, not just fuel — it inhibits histone deacetylaseshistone deacetylases, enzymes that clamp down on gene activity; blocking them lets anti-inflammatory and stress-resilience programs switch on — a mechanism shared with fasting and touches other pathways, which in lab work reduces inflammation and changes how stress-responsive genes are expressed.
The appetite theory says fat and protein keep you full.The appetite theory says fat and protein keep you full: high-fat, adequate-protein meals simply satisfy longer, so people spontaneously eat fewer calories — the most mundane but best-supported explanation for keto weight loss.

Mood and the mind

The diet's psychiatric turn is recent.The diet's psychiatric turn is recent: after a century in epilepsy, ketogenic therapy is now being trialed for bipolar disorder, schizophrenia and severe depression, on the idea that they share a metabolic weakness — unstable brain energy — that ketones might steady. Mood stabilisers came from anticonvulsants.Mood stabilisers came from anticonvulsants — drugs that calm the over-excitable neuron — and the diet's long anti-seizure record is part of why it reached psychiatry.

The psychiatric studies are all recent, and small.The psychiatric studies are all recent — pilots from 2023 and 2024, one randomised trial from 2025 — and small enough that each is described here on its own terms.

Bipolar has a pilot from Edinburgh.Bipolar has a pilot from Edinburgh: 27 euthymiceuthymic — in a stable phase, neither manic nor depressed volunteers followed a modified ketogenic diet for six to eight weeks and reported fewer symptoms of depression and anxiety, with higher ketone levels tracking lower impulsivity, in a design with no control group.
Schizophrenia and bipolar have a Stanford pilot.Schizophrenia and bipolar have a Stanford pilot: 21 adults on antipsychotic medication followed a 4-month ketogenic diet, improving psychiatric scores while reversing metabolic syndromemetabolic syndrome — the cluster of belly fat, high blood pressure and high blood sugar that antipsychotic drugs tend to build in most of them.
Treatment-resistant depression has a randomised trial.Treatment-resistant depression — depression that several drugs have failed to lift — has a randomised trial, the first of its kind: 88 UK patients, six weeks, where both diet groups improved but the ketogenic arm improved more.
A meta-analysis pools the smaller studies.A meta-analysis pools the smaller studies: the 2025 edition covered 50 studies and found small-to-moderate improvement in depressive symptoms across the randomised trials, with larger effects reported in the uncontrolled designs.

The psychiatric theories extend the epilepsy ones.The psychiatric theories extend the epilepsy ones — steadier fuel, calmer signalling — plus one new idea about metabolism itself.

The brain-energy hypothesis is the centerpiece.The brain-energy hypothesis is the centerpiece: brain scans in schizophrenia and bipolar show disrupted glucose handling and insulin signaling even before any medication, and ketones bypass that broken pathway entirely, feeding neurons a fuel it can still use.
Mitochondria may be the shared weakness.Mitochondria — the power plants inside every cell — may be the shared weakness: they run poorly in serious mental illness, and beta-hydroxybutyrate improves their output while dialing down oxidative stress, at least in lab models.
The antipsychotic tradeoff may be eased.The antipsychotic tradeoff may be eased: those medications work but drive weight gain and rising blood sugar, and the diet directly counteracts that damage while the psychiatric scores improve alongside.
GABA and glutamate return from the epilepsy story.GABA and glutamate return from the epilepsy story: the diet's anti-seizure chemistry rebalances the brain's main inhibitory transmitter, GABA, against its excitatory one, glutamate — and the same balance governs mood and anxiety.

For everyday brain function, the studies are few and short.For everyday brain function — focus, memory, mental clarity in otherwise healthy people — the controlled studies are few and short, mostly days to weeks; the widespread reports of mental clarity are anecdote, and the underlying observation is that ketones fuel the brain steadily rather than in glucose swings.

Medication interactions matter most here.Medication interactions matter most here: lithium levels can rise as the body sheds water and salt in the first weeks, so anyone on lithium needs blood monitoring; diabetes and blood-pressure drugs may need dose changes within days; and ketogenic therapy sits beside a psychiatrist, never instead of one.

Adjunctive practices

Electrolyte replacement is the standard companion practice.Electrolyte replacement is the standard companion practice: low insulin on keto makes the kidneys dump sodium, so practitioners deliberately add salt, potassium-rich foods like avocado and leafy greens, and magnesium, to avoid the headaches, cramps and fatigue of the so-called keto flu.

Exercise is encouraged, with a caveat.Exercise is encouraged, with a caveat: resistance training and low-intensity aerobic work fit keto fine, but performance in high-intensity glycolyticglycolytic — burning stored sugar rather than fat, the metabolism that sprints and hard intervals demand sport usually drops, because those efforts rely on muscle glycogen that the diet keeps low.

Endurance athletes are the interesting case.Endurance athletes are the interesting case: after several weeks of adaptation, fat oxidation rises and steady-state performance can be preserved, but most controlled studies still find no improvement, and some find a real cost in race-pace efforts.
Timing is one workaround.Timing is one workaround: targeted keto eats a small amount of carbohydrate just before or after training, refilling glycogen for hard sessions while staying ketogenic the rest of the day.

Fasting pairs naturally with keto.Fasting pairs naturally with keto, since both raise ketones and lower insulin; many practitioners combine 16:8 time-restricted eating with the diet, though the combination is optional and raises the risk of chronically under-eating.

Exogenous ketones and MCT oil exist as shortcuts.Exogenous ketones — lab-made ketones (esters and salts) that you swallow instead of make — and MCT oil, a fat from coconut oil that the liver turns into ketones unusually quickly, exist as shortcuts: they raise blood ketones within hours rather than days, but both are expensive and transient, and neither is shown to reproduce the diet's effects on its own.

You can test your ketones.You can test your ketones by blood meter, which is the most accurate; by breath acetone, which is cheap per test after the device; or by urine strips, which only work in the first week or two before the body starts reabsorbing acetoacetate.

What is out there

A map of the studies so far, by population, size and duration.A map of the studies so far, by population, size and duration, without verdicts — outside epilepsy, studies running a year or longer are few, so the family's review of the last thirty years of literature can sharpen this map.

Epilepsy has the longest record.Epilepsy has the longest record: continuous clinical use since 1921, randomised trials in children with drug-resistant seizures since the late 1990s, adult epilepsy programmes, and neurology guideline endorsements, with studies running from weeks to multi-year follow-up.
Weight loss has randomised trials at three horizons.Weight loss has randomised trials at three horizons — 6 and 12 months, repeatedly meta-analysed, with fewer studies following to 24 months — typically with dozens to a few hundred participants per arm.
Type 2 diabetes has supervised trials and one 2-year study.Type 2 diabetes has supervised trials — some inpatient, some in primary care — and one 2-year remote-care study of 349 people reporting medication reductions and weight loss sustained at two years.
Psychiatry has pilots from 2023 and 2024, and a trial from 2025.Psychiatry has pilots from 2023 and 2024 — Edinburgh in bipolar, Stanford in bipolar and schizophrenia, described above — and a randomised trial from 2025 in treatment-resistant depression, plus the 50-study meta-analysis; all small, all recent.
Healthy people have the shortest literature.Healthy people have the shortest literature: trials of days to weeks in athletes and in cognition experiments, plus observational cohorts that track low-carbohydrate eating over years without assigning it.
Safety studies past one year are few.Safety studies past one year are few: trials mostly run weeks to months, so what is known about long-term effects rests on a small number of longer studies and on registry data.

Talk to a doctor before starting.Talk to a doctor before starting if you take diabetes or blood-pressure medication, since doses may need adjusting within days; if you have any history of disordered eating; if you are pregnant or breastfeeding; or if you have a rare metabolic disorder that makes fat burning dangerous.

Sources

Written as stretchtext — Ted Nelson's format, from cc-stretch via stretchserve.

This page is stretchtext, served by stretchserve (live) — skim the page as it reads, and click a dotted phrase to grow it.