Diet & Nutrition for Mental Health | Dr. Chris Palmer

The episode opens on a framing problem rather than a finding. Huberman grants that nearly everyone has a working notion of mental illness and a working notion of nutrition, and that everyone has felt a food change how they feel within minutes. What he says almost nobody appreciates is the time scale: the interaction may be intimate, possibly causal, and slow. His formulation is that food eaten yesterday, the day before, or a decade ago may still be shaping how the brain and body produce present experience. That single sentence sets the evidentiary bar for everything that follows, because a claim about effects accumulating over years cannot be tested by how someone feels after lunch.

The introduction also plants the hypothesis the conversation keeps returning to: that obesity may in many cases be a consequence of brain dysfunction rather than a metabolic fault that later reaches the brain. Huberman pairs this with a preview of Palmer's mechanistic bet, that mitochondrial function and dysfunction are central to mental health and illness, and with a historical point that matters more than it first appears, that ketogenic diets were not invented for weight loss but as treatment for a neurological condition, epilepsy. The ordering is deliberate: mechanism, history, then case material.

Palmer's entry into this field is not a research programme but a personal history, and he tells it plainly. Childhood obsessive-compulsive disorder, a mother's psychotic break, a period of homelessness with her, then depression and suicidality. He recovered enough to finish medical school well, by his account winning an award as one of the top students, and to land a Harvard internship and residency in psychiatry. It was during that residency that he was diagnosed with metabolic syndrome: high blood pressure, what he calls horrible lipids, and pre-diabetes.

The detail that gives the story its force is that he believed he was already doing the recommended thing. He was on a low-fat diet and exercising regularly, and his doctor kept repeating the same instruction. His reported reply is the most quotable line of the section: he kept asking, "What diet? What exercise?" while his blood pressure climbed year over year. The endpoint was three prescriptions at once, for pre-diabetes, cholesterol and blood pressure, in a man in his twenties.

Huberman asks directly whether he was overweight, and the answer complicates the usual picture. Technically he was not, but he had abdominal fat, which he now reads as a sign of insulin resistance he did not recognise at the time. His physician's explanation was hereditary: diabetes, hypertension and overweight in the parents, therefore genes, therefore medication. Palmer, already a physician, drew the obvious inference about what three drugs in his twenties implied for his forties and fifties.

The conversion is described as reluctant and almost embarrassed. He had heard that the Atkins diet might improve cholesterol and pre-diabetes, did not believe it, and considered it dangerous and reckless because that is what his training had taught him. His reasoning for trying it anyway was not conviction but exhaustion: he had tried the dogma for years and it had failed, so this would be his last experiment before accepting medication. His version was modified, red meat still avoided out of fear, closer to what he calls a healthy version and probably nearer the South Beach diet, which did not yet exist. Within three months, by his account, the metabolic syndrome resolved completely; he lost roughly ten pounds, most visibly the abdominal fat, and his physician asked what he was doing.

Huberman redirects the story from cardiometabolic markers to mental state, and this is where the interview becomes more than a diet anecdote. Before the change Palmer was functioning at a high level while still carrying low-grade depression and OCD, and he was on medication. His account of treatment is unsparing: Prozac for a long period, sleep destroyed by it, then a suggestion of additional pills to fix the sleep, a sequence he says did not sit well with him even as a psychiatry resident being taught that this is how the disorders are treated. He also had extended psychotherapy, including psychoanalytically oriented work and a hospitalisation with lithium and imipramine, and he characterises the psychotherapy as essentially worthless to him and possibly harmful in the end.

What he says struck him was not the disappearance of metabolic syndrome, which was the goal, but the change in mood, energy, concentration and sleep. The concrete marker he offers is behavioural and oddly specific: for the first time in his life he began waking before his alarm and feeling rested, having previously engineered his mornings around snooze buttons to arrive on time. That kind of detail is worth noting precisely because it is not a lab value. It is also, of course, self-report from an n of one who knew what he was taking.

The reframing he draws from it is about temperament rather than biomarkers. He had assumed there were two kinds of people, the "haves" and the "have-nots", the second group being the tired, the flat, the ones for whom working hard exhausted the budget for playing hard. He had read the energetic as lucky or privileged, endowed with better genes or a better childhood. Huberman supplies the image of the kids with genuine smiles in the yearbooks, and Palmer agrees. The implication he wants heard is that the dividing line may be metabolic and therefore movable, which is a far larger claim than any single case can support.

Huberman's response is the right one for this material. He thanks Palmer for the disclosure and draws out the point that high functioning and accomplishment are compatible with a very rough internal road, and that the visible record does not show the iteration required to arrive at something that worked. He explicitly endorses the observation that many more people are suffering than is visible, and that the two-group model cannot be how it works.

The path from personal result to clinical practice went through family first. His sister and sister-in-law were annoyed at a Thanksgiving that he could sit through without wanting the pumpkin or apple pie, and his explanation was that he no longer craved it. They tried the diet and, by his account, saw large improvements in mood, energy and sleep. Within a few years he had formed the impression of a powerful antidepressant effect, though at this stage the sample is people who watched him succeed and were motivated to copy him, which is a textbook setup for expectancy.

His clinical setting sharpens the anecdote into something more interesting. As an attending at a tertiary care hospital he almost never sees a first episode; he inherits patients who have seen six or more psychiatrists, tried dozens of medications, spent decades in therapy and often had ECT, with nothing working. That is the population in which an experimental intervention is both most justified and most difficult to evaluate, because such patients differ from ordinary clinic populations in ways that make spontaneous change unlikely but also make attribution hard.

The first results were partial and he says so. Not everyone was helped, not everyone was interested or able. Among those who could do it, one woman became hypomanic within a month after more than five years of continuous depression, suicidality and repeated hospitalisation. Palmer reads the switch as evidence of a genuinely potent antidepressant effect, likening it to a medication that happened to work for her. A clinician reading the same case would also see a mood switch that requires monitoring, which is why the observation cuts both ways.

His decision to keep his head down until 2016 is the most credible thing in the section. He notes that there were then few trials of the Atkins diet for safety or efficacy even in weight loss and diabetes, let alone in mental disorders, that he felt himself to be on the fringe, and that this was not going to be met with praise. So he offered it quietly to patients and otherwise waited. That is not the behaviour of someone selling a protocol, and it explains why the claim arrives in this conversation attached to case material rather than to a trial.

Huberman presses on the operational details, which is where vague nutritional advice usually collapses. Palmer concedes he was winging it at first: early patients were told to try Atkins, and his stated aim was ketosis, measured at that time with urine strips because he was not aware of blood ketone monitors. Huberman notes that urine strips are less accurate but still useful as a rough guide, and Palmer agrees. The exchange matters because it establishes, without anyone underlining it, that the early cases were conducted at a low level of measurement precision.

The core clinical observation is stated as a pattern rather than a trial result. When patients attempted the diet without producing ketones, they often did not improve; the clinical benefit appeared once ketosis was reached. This is a dose-response shaped observation, which is more persuasive than a simple before-and-after, but it is still open-label, unblinded and confounded by adherence, since the patients who reach ketosis are by definition the ones executing the intervention most faithfully.

Huberman generalises the point: any eating pattern, including fasting, that raises ketones to the point where the patient is in ketosis would count as a step in the right direction, independent of the specific foods involved. Palmer confirms that fasting will do it. He describes using Atkins-style fat fasts, and notes that overweight and obese patients often reach ketosis easily because they have large fat stores, which is a physiological point rather than a therapeutic one. Huberman adds the familiar early effect of carbohydrate restriction, water loss and rapid scale movement that feels like fat loss, and Palmer agrees while noting that the field was simultaneously becoming more sophisticated, with blood ketone monitoring and multiple ketogenic variants.

The pivot of the episode is a single patient, and Palmer's telling is unusually concrete. A 33-year-old man with schizoaffective disorder, his patient for eight years, with daily auditory hallucinations and paranoid delusions involving powerful families possessing technology that could control and broadcast his thoughts. He could not go into public without terror, and he had built an explanation around something he had done at eleven years old. Seventeen medications had failed to touch the symptoms while causing substantial weight gain. The presentation is not ambiguous, and that is the point: this is not a mild or borderline case that could be explained by diet improving general wellbeing.

The motivation for the attempt was not psychiatric at all. At 340 pounds, the patient decided he would never get a girlfriend unless he lost weight, while simultaneously describing himself as a loser living with his father with nothing going for him, and asking to address at least one of his problems. Palmer is explicit that he expected nothing for the psychosis, because schizoaffective disorder is not depression and he regarded them as entirely different disorders. That stated expectation is what makes the subsequent report more than a hopeful anecdote: the clinician was not looking for the effect he later described.

The sequence he reports is two weeks of behavioural change with no medication adjustment, better eye contact, more smiling, more speech, a patient who seemed to have come to life; then, at six to eight weeks, an unprompted report that the voices were fading and that the belief about the families no longer held. Palmer quotes him arriving at the conclusion that the whole edifice sounded crazy in retrospect: "It probably never was." That sentence, spoken by the patient rather than the doctor, is doing a great deal of rhetorical work in this episode, because insight of that kind is not something antipsychotics reliably produce in a chronic, refractory case.

The functional outcome is what Palmer emphasises: 160 pounds lost and kept off, completion of a certificate programme, going out in public without paranoia, performing improv in front of a live audience, eventually moving out of his father's home to live independently. He says this blew his mind as a psychiatrist and sent him on a scientific search for what had happened. The honest reading is that the functional detail is more informative than the symptom ratings, because it is harder to produce by encouragement alone, and equally that a single case selected for a dramatic response cannot tell us the base rate of such responses.

One of the more useful moments is a straight error, corrected live. Huberman guesses that schizoaffective disorder is a low-grade schizophrenia, the kind of presentation that seems odd but not frightening. Palmer corrects him: schizoaffective disorder is essentially schizophrenia with superimposed mood episodes, and it can be more severe, in his phrase full-blown schizophrenia plus full-blown usually bipolar illness. The mild, eccentric, mildly paranoid category Huberman was reaching for is schizotypal. Huberman then states the correction in his own words for the audience and notes that it is now entirely clear who the clinician in the room is.