NIH Director Jay Bhattacharya, despite his many faults, did recommend a very accomplished and appropriate scientist to the post at the National Institute of Mental Health (NIMH), Vikaas Sohal. However, Science reports that this recommendation was rejected by RFK Jr., who, rather, favors Shebani Sethi, a “metabolic psychiatrist” who has never held an NIH grant and who was rejected by the search committee. The full story is disturbing, and represents RFK’s continued “Lysenkoisation” of the American health infrastructure.
I don’t need to repeat the excellent reporting in Science, so instead let’s take a look at metabolic psychiatry and why it is extremely dubious from an SBM perspective. Metabolic psychiatry is the notion that metabolic health is an important contributor to, and may even be the primary cause of, even severe psychiatric conditions. I actually encountered an apparent practitioner while helping a friend navigate the world of mental health for a family member. My experience was fairly representative of the problems.
The patient is someone with a challenging combination of conditions, including autism, depression (probably bipolar), and psychosis. Medications help, but balancing all the effects and side effects has been difficult. A recent consultant had what I found to be a surprising diagnosis—all of their symptoms were caused by metabolic disorders, specifically hypothyroidism and vitamin deficiencies. I found this surprising partly because the patient does not have hypothyroidism or the stated vitamin deficiencies (all of which can be confirmed with simple blood tests), and a subsequent visit with an endocrinologist confirmed this.
The psychiatrist sent a package of studies to back up their diagnosis and treatment plan, and I dutifully went through them. What I found was a very typical pattern—studies that absolutely do not establish that subclinical hypothyroidism (which is what he thought they had) can cause any of the patients’ symptoms. Some were pre-clinical studies of doubtful relevance. Others involved symptoms of severe hypothyroidism, not relevant to someone who is at worst subclinical (meaning they have no typical symptoms of hypothyroidism and don’t meet criteria on blood work). He was extrapolating wildly and unprofessionally from preliminary, preclinical, or not-relevant research all to support his preferred narrative.
This is all the problem with the field of metabolic psychiatry as a whole. This is not to say that metabolism is not important to mental and neurological health. It is. But that is not the same thing as metabolism being the cause of any specific mental or neurological disorder.
This does involve how we conceptualize mental illness, something which is an ongoing metadiscussion among researchers, and has played prominently at the NIMH. For anything affecting the function of the brain (whether the manifestations are psychiatric or neurological), there are several potential layers of cause. One is biology, meaning whether the cells of the brain are biologically healthy or are they displaying any pathology. If the brain is infected, for example, that is a biological problem.
Another layer is the function of neurons themselves, which may be suboptimal even if there is no definable pathology. This is often thought of as a “neurochemical” problem—are brain cells making too much or too little of a neurotransmitter, or is there something wrong with their basic signalling function?
The third layer is the “wiring” of the brain, which is an important determinant of brain function. Even if the cells themselves are all happy and healthy, and the neurons and glia are all functioning optimally, the pattern of networks within the brain may be causing issues. The simplest way to think of this is that some critical brain function is either hyperactive or hypoactive. But there may be more complex network issues that have to do with stability, self-regulation, balance, or other complex functions.
Psychiatric diagnoses are generally thought of as either network disorders or biochemical issues or some combination of both. Biological/pathological issues are generally considered to be handled by neurologists, and in fact should be ruled out before considering a symptom primarily psychiatric.
Metabolic psychiatry is therefore a radical idea—the notion that the biology of brain cells themselves may be the ultimate culprit in mental illness. I am not saying that the idea has no merit and should not be researched. Optimizing brain health is already considered to be very important to neurological and psychiatric care. Exercise, for example, is critical to brain health, as is good nutrition, adequate sleep, minimizing alcohol use, and controlling things like blood pressure. However, these factors are considered to be secondary, not primary, meaning that optimizing them can help mitigate a neuropsychological problem, but is not the primary cause and is therefore not a cure.
The criticism of metabolic psychiatry is not that we should not be looking to optimize biological factors when managing mental illness, and studying how best to do that. Rather, it is that proponents tend to go way beyond the current research, and are making claims for an entirely new explanatory paradigm of mental illness that is not science-based.
This is a common issue we address here. The fact that there may be some non-specific symptomatic effects of sticking people with needles does not establish that chi and meridians exist. Similarly, the fact that antidepressants increase serotonin does not mean that underproduction of serotonin is the cause of depression. Massage may make people feel better, but that is not enough evidence to justify the claim that it is squeezing toxins out of the tissues.
Psychiatry is unfortunately rife with this fallacy. A great example is EMDR (eye movement desensitization and reprocessing). The research shows that it has, at best, some non-specific benefit, but the studies are largely poorly controlled, and all of the benefit likely results from the therapeutic relationship and perhaps the effect of introducing a novel element into the treatment. The research absolutely does not establish the proposed mechanism of action of EMDR proponents.
Similarly, the studies of metabolic psychiatry are mostly preliminary, poorly controlled, and not designed to truly test efficacy. Most are correlational only. Nor are they capable of establishing any causal mechanism. This 2026 systematic review of the ketogenic diet for anxiety and depression is representative. The conclusion:
“In this systematic review and meta-analysis, KDs were associated with modest improvements in depressive symptoms, particularly with biochemical ketosis verification, while anxiety evidence was inconclusive. Given heterogeneity, comparators, and short follow-up, well-powered trials with standardized, verified protocols, structured support, and prespecified outcomes are needed to confirm efficacy and durability.”
This type of conclusion is all-too familiar to us. The quality of the research is poor, and the effects are dubious. This would all be fine if such evidence were viewed entirely as preliminary, and used to design more rigorous studies. But that is often not what we see. Instead, we often see extrapolation from such preliminary evidence to broad conclusions about efficacy and mechanism.
Sethi’s research has focused on the ketogenic diet, and is mostly the preliminary uncontrolled or poorly controlled type of data discussed above. But her public statements are more effusive:
“We’re seeing huge changes,” Sethi said. “Even if you’re on antipsychotic drugs, we can still reverse the obesity, the metabolic syndrome, the insulin resistance. I think that’s very encouraging for patients.”
We have discussed the problem with ketogenic diets for weight loss—the evidence shows short-term benefit only, but has not established long-term benefits for weight loss or metabolic improvements. So she is overstating the metabolic benefits of a ketogenic diet, then extending them to mental illness based on preliminary symptomatic evidence.
Sethi is not remarkable in that she is pursuing a somewhat reasonable hypothesis, but is overstating the evidence and is not being sufficiently skeptical of the causal interpretations—something unfortunately too common today. But she is being considered to head the NIMH, a position that requires a lot of research experience and a very critical eye for these meta-questions in psychiatric research. She seems underqualified, and the search committee seems to have come to the same conclusion. But her research aligns with RFK’s MAHA ideology, and that is apparently all that matters.
