Otomobiltamir
Biology & Genetics

Metabolic Psychiatry and Mental Illness: What the Evidence Does and Doesn’t Show

Published Oct 28, 2026 Reads 915 By Steven Novella

Metabolic health can affect brain function, but current evidence doesn’t show it is the primary cause of specific psychiatric disorders.

Metabolic Psychiatry and Mental Illness: What the Evidence Does and Doesn’t Show

Metabolic psychiatry makes a plausible starting point sound like a settled explanation: metabolism can affect brain health, so perhaps metabolic problems cause psychiatric illness. That distinction matters. The evidence described here supports studying metabolic health; it doesn’t establish that metabolic dysfunction is the primary cause of a particular person’s depression, psychosis, or other psychiatric condition.

The debate has taken on added significance at the National Institute of Mental Health (NIMH). NIH Director Jay Bhattacharya recommended Vikaas Sohal, a highly accomplished scientist, to lead the institute. Despite his many faults, that was an appropriate choice, the author argues. But Science reports that RFK Jr. rejected the recommendation and favors Shebani Sethi, a “metabolic psychiatrist.” Sethi has never held an NIH grant, and the search committee rejected her. The author sees the episode as another sign of RFK’s “Lysenkoisation” of American health infrastructure: research aligned with his MAHA ideology appears to count more than the experience and judgment expected of an NIMH director.

Where the hypothesis outruns the evidence

The concerns here aren’t an argument against studying metabolism and mental health. They come from a personal encounter with the way the theory can be applied. While helping a friend find care for a family member with autism, depression thought likely to be bipolar, and psychosis, the author encountered a consultant who attributed all of the patient’s symptoms to hypothyroidism and vitamin deficiencies. Yet the patient had neither. Simple blood tests could check those claims, and an endocrinologist later confirmed they were unfounded.

The consultant supplied research to support the diagnosis and treatment plan. On examination, the studies didn’t show that subclinical hypothyroidism causes the patient’s symptoms. Some were preliminary studies in animals or other preclinical settings; others concerned severe hypothyroidism, which is a poor match for someone without typical symptoms or qualifying blood-work results. The author’s judgment is blunt: the consultant had stretched weak or irrelevant research to fit a preferred explanation.

That case points to the larger problem. Metabolic health matters to mental and neurological health, but that doesn’t make it the cause of any specific disorder. Brain-related conditions can involve different levels of dysfunction: disease or injury in brain cells, problems with neuronal signaling, or changes in how brain networks are organized and regulate activity. Psychiatric diagnoses are generally understood in terms of network and biochemical function, alone or in combination. Neurologists typically investigate identifiable biological pathology, which should be ruled out before symptoms are treated as primarily psychiatric.

Metabolic psychiatry makes a more radical claim: that problems in the biology of brain cells may sit at the root of mental illness. There’s a reasonable research question there. Exercise, nutrition, sleep, limiting alcohol, and controlling blood pressure can all support brain health and help care for people with psychiatric or neurological conditions. But benefits from improving those factors don’t show that they caused the illness, or that changing them will cure it.

Here’s the thing: a treatment can help without its proposed explanation being right. Needles might produce nonspecific symptom relief without proving that chi or meridians exist. Antidepressants can raise serotonin without showing that depression results from producing too little serotonin. Massage may feel good without squeezing toxins from tissue.

EMDR, or eye movement desensitization and reprocessing, offers another example. Its research suggests, at most, some nonspecific benefit. The studies are largely poorly controlled, and any gains may come from the therapeutic relationship or from adding a novel feature to treatment. They don’t establish the mechanism proposed by EMDR’s proponents.

Modest findings, sweeping claims

Much of the metabolic psychiatry research has similar limitations: preliminary results, weak controls, correlational designs, and little ability to test whether a proposed mechanism actually causes a disorder. A 2026 systematic review of ketogenic diets for anxiety and depression captures the gap. Its conclusion was:

“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.”

That’s a cautious conclusion, and the evidence it describes is preliminary. Such findings can guide better trials. They can’t support confident claims about a treatment’s effectiveness or explain why it might work. The problem arises when early results are presented as proof of a broad new account of mental illness.

Sethi’s research has focused on ketogenic diets, largely through the uncontrolled or poorly controlled studies described above. Her public comments sound more certain:

“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.”

The evidence on ketogenic diets for weight loss points to short-term benefits; it hasn’t established lasting improvements in weight or metabolic health. Sethi appears to extend claims about those outcomes to mental illness, drawing on preliminary evidence about symptoms. That leap needs stronger support.

Sethi is pursuing a hypothesis that may be worth investigating. The concern is that she appears to overstate what the research shows and to accept causal explanations too readily—a familiar problem in psychiatry. Leading the NIMH demands substantial research experience and a sharp eye for these questions. Sethi seems underqualified, and the search committee reached the same conclusion. Yet her work aligns with RFK’s MAHA ideology. In this appointment, that alignment may be carrying more weight than the evidence or the committee’s assessment.

About the author

Steven Novella, MD, founded Science-Based Medicine and serves as its executive editor. He is an academic clinical neurologist at Yale University School of Medicine, hosts and produces the weekly podcast The Skeptics’ Guide to the Universe, and writes the NeuroLogicaBlog, which covers neuroscience, science, skepticism, critical thinking, and the media’s relationship with society. He has also created two courses for The Great Courses and written a book titled The Skeptics’ Guide to the Universe.

This article is filed under Neuroscience/Mental Health and Politics and Regulation. Its tags include Jay Bhattacharya, MAHA, metabolic psychiatry, National Institute of Mental Health, National Institutes of Health, NIH, NIMH, Podcast Jay, Robert F. Kennedy Jr., Shebani Sethi, and Vikaas Sohal.

Source: Steven Novella · sciencebasedmedicine.org

Discussion

Sign in to join the discussion.