The Stanford Ibogaine Study, and How It Was Reported

This is the study behind almost every favourable ibogaine headline since January 2024, and it is a good study of a kind that cannot answer the question people think it answers.

What the evidence shows and what it is worth is set out on ibogaine for PTSD, which appraises the papers. This page is about the study as a public event: who was in it, who paid, and why it keeps arriving as news.

What actually happened

Thirty male special operations veterans with traumatic brain injury, twenty-eight of them mild, from repeated blast and combat exposure, received a single course of ibogaine, given orally with intravenous magnesium, at a clinic in Mexico. The protocol has a name, MISTIC, for Magnesium-Ibogaine: the Stanford Traumatic Injury to the CNS. The results were published in Nature Medicine in January 2024, and they were large.

Three features of the design decide how much weight it can carry, and none of them is a criticism of the authors, who state all three plainly.

Nobody was recruited to treatment. The participants had independently scheduled their own treatment, after being approved for a grant by a veterans’ non-profit; the study recruited them afterwards, by referral. It observed people who had already decided to go.

There was no control group. No randomisation, no placebo, no comparison arm. A single-arm observation cannot separate the drug from expectation, from the journey, from the fact of being studied, or from the passage of time.

The sample was narrow by the authors’ own account. They describe it as highly homogeneous, mostly white men from elite military units who tended to be in above-average physical condition. That is not the population that appears in the fatality record.

The safety claim is narrower than the coverage

The study reported no bradycardia, no tachycardia, no haemodynamic instability and no clinically meaningful QT prolongation. Magnesium was given specifically to blunt the cardiac effect.

The qualifier is in the methods. The QTc was followed visually on a continuous five-lead monitor, and “clinically meaningful” is defined in the paper as qualitatively detectable on that monitor. Screening electrocardiograms were taken beforehand, but the paper reports no measured QTc value at any point after dosing, and no comparison against a threshold. The interval was watched, not recorded.

That matters because the Dutch study that did measure it found half its participants crossing 500 milliseconds with electrolytes confirmed normal beforehand. A visual read on a monitor and a measured interval are different instruments, and only one of them produces a number. What ibogaine does to a heart covers why.

One cohort, many announcements

This is the part that distorts the public record, and it is nobody’s fault in particular.

The January 2024 Nature Medicine paper was the first. Since then the same thirty men have been analysed again for brain imaging and electroencephalography, for mystical experience, and for the subjective quality of the experience. The mystical-experience paper appeared online in November 2025, in the Journal of Affective Disorders’s February 2026 issue. Stanford’s own news page, filed under a January 2024 URL, has since been updated to open its account with a July 2025 resting-state EEG paper in Nature Mental Health on the same cohort of thirty. Stanford’s page describes that paper as analysing “electroencephalography and MRI scans”, though the paper itself reports EEG only, and notes that it “expands on” the 2024 study.

Each is a legitimate paper. Together they produce something misleading, because each arrives with its own coverage and none of the coverage says these are the same thirty people. A reader meeting the sixth announcement reasonably concludes the evidence is accumulating. The participant count has not moved since 2024.

Who paid, and who benefits

The paper’s competing-interests statement is unusually full, and the site treats that as a credit rather than a mark against it. Everything below is declared in the paper.

Two authors are named on a patent application covering the safety of this protocol, two on a second covering ibogaine for disorders of brain ageing, and the three clinic shareholders on a third, also on safety during ibogaine therapy. Three are shareholders in the company that operates the clinic where the treatment took place, and one founded a company that sources and converts ibogaine precursors. Participants’ travel and treatment were funded by grants from a veterans’ non-profit.

The other half of the ledger belongs here too. The academic side of the study took no money from either the non-profit or the clinic. It was funded by private philanthropic gifts, and the paper records that the funders played no part in its design, execution, analysis or writing.

None of that makes the result wrong. It does mean the most widely reported favourable result in this field was not produced by disinterested parties, and that the clinic where it happened has a commercial interest in the finding. Who is developing ibogaine covers the wider picture of what companies in this field have actually delivered.

The study that points the other way

A second cohort exists and is rarely mentioned beside the first. Eighty-six special operations veterans treated at a clinical programme in Mexico were followed prospectively, and the analysis was published by a research group with no connection to the Stanford study, working from outcome surveys the treatment programme itself collects. Their improvements were real and much smaller: effect sizes between 0.27 and 0.41 at one month, against the Stanford cohort’s one-month figures above 2.

Those eighty-six received 5-MeO-DMT as well as ibogaine, so it is a different intervention and not a clean comparison. It is still the larger prospective series in this population, and it points away from the headline number.

What would settle it

A randomised controlled trial in a population that is not self-selected, with a measured QTc rather than an observed one, and participants who did not pay to be there. What the trials have actually run sets out the registry, and the short answer is that no such trial has reported.

Until one does, thirty men remain thirty men, however many times they are analysed.