The results of a study published on March 18, 2026 in JAMA Psychiatry were quite shocking to many psychedelic advocates. Its blunt conclusion was that psychedelic-assisted therapy is no more effective than traditional antidepressants for treating depression. Cue the headlines. Psychedelics Show No Edge Over Antidepressants. Scientists Pump the Brakes on Psychedelic Hype.
Before you file this away as another cycle of overpromise and disappointment, it’s worth reading what the study actually did and what it reveals about the lens through which mainstream medicine still sees the human mind.
What the Study Found
The research team conducted a systematic review and meta-analysis of 24 clinical trials. Eight covered psychedelic-assisted therapy (PAT), mostly psilocybin. Sixteen were open-label trials of traditional antidepressants (SSRIs and SNRIs). The combined data touched over 8,000 patients.
Their core insight was a methodological one. Psychedelic trials have a well-known blinding problem. The drugs are so experientially obvious that 90–95% of participants know they’ve received the active substance. Standard antidepressant trials, by contrast, use proper double-blind designs where participants can’t easily tell. The researchers argued this creates an unfair comparison as psychedelic results are always “open-label” by nature, so they should be compared against open-label antidepressant trials, not blinded ones.
When they made that comparison, the advantage of psychedelics disappeared. Both groups showed equivalent improvement. Roughly a 12-point drop on the Hamilton Depression Rating Scale. The difference between them was 0.3 points. Statistically meaningless.
Even Szigeti, the senior author, was disarmed. “What I wanted to show is that even if you compare psychedelics to open-label antidepressants, psychedelics are still much better,” he said in a press release. “Unfortunately, what we got is the opposite result.”
Systematic Bias
The study’s methodology has attracted sharp challenge from some of the most credentialed researchers in the field. Prof. Robin Carhart-Harris – the Ralph Metzner Distinguished Professor at UCSF and lead author of the landmark 2021 psilocybin-versus-escitalopram head-to-head trial – has responded in detail in an extensive public thread on X.
His critique goes well beyond a methodological quibble. He argues the paper’s conclusion is fundamentally invalid, because the antidepressant trials it relies on are riddled with systematic biases that artificially inflate their results. Biases that are entirely absent from the psychedelic trials they’re being compared against.
His first objection concerns washout periods. Many of the open-label antidepressant trials included in the meta-analysis allowed patients only a one-to-two week washout from their previous antidepressants before starting the study drug. This is a significant problem. When a patient stops an antidepressant abruptly, they often experience discontinuation effects that temporarily worsen their symptoms, pushing baseline scores artificially high. When they then restart a new antidepressant, symptoms drop rapidly, not because the new drug is especially effective, but because they’re rebounding from withdrawal.
Carhart-Harris calls this “systematic bias” and notes it is entirely absent in PAT trials, where prior psychedelic exposure is rare and no such rebound dynamic exists. He backed up this point with direct screenshots from the specific studies the meta-analysis cited.
His second objection concerns patient selection. Several of the antidepressant trials pre-screened participants and excluded those who had previously failed to respond to antidepressants. This cherry-picks a population more likely to respond, further inflating reported efficacy. Again, PAT trials have no equivalent mechanism and no equivalent advantage.
Third, Carhart-Harris points to what he calls the “pragmatic” design features of many open-label antidepressant studies. Unlike the fixed-dose, time-limited protocols used in psilocybin research, several of the TAD trials allowed flexible dose titration based on patient response, early discontinuation if the drug wasn’t working, and the continuation of outside psychotherapy. All of these features improve measured outcomes. None of them are present in PAT trials. He illustrated this specifically using data from the STARD citalopram evaluation – one of the trials the meta-analysis drew on.
His fourth objection returns to expectancy, but with added force. Open-label antidepressant trials are well-established to be inflated by expectancy effects as patients who know they’re receiving a drug and expect it to help tend to report greater benefit. Yet as Carhart-Harris has shown in his own research and in four psychedelic trials conducted by his group, the same dynamic simply does not apply to psilocybin. Pre-trial expectancy showed no correlation with actual clinical outcomes. If anything, as his 2021 trial found, patients with the highest expectations for psilocybin did worse, not better. The premise of “equal unblinding” therefore doesn’t hold. Expectancy still systematically favours antidepressants in ways it doesn’t favour psychedelics.
Taken together, Carhart-Harris describes the overall comparison as an “indefensible deck stack.” The antidepressant data is contaminated by legacy trial-design flaws; the psychedelic data is not. Comparing them as if they were equivalent, he argues, does not level the playing field. He goes further, suggesting that the paper’s publication in JAMA reflects a “contrarian bias” in the field, where research casting psychedelics in a negative light finds an unusually receptive audience.
Apples and Oranges
This whole drama reflects a deeper pattern that’s prevalent in modern health research.
The study treats antidepressants and psychedelic therapy as functionally interchangeable units of depression reduction. Feed in a pill, measure the output on a rating scale. Compare outputs. Declare a winner. It is, at its core, a machine model of the mind. Inputs, outputs, optimisation. The assumption is that healing is a single linear variable we can extract and rank.
But the body, and particularly the brain, does not work this way. SSRIs modulate serotonin reuptake on a continuous daily basis. Psilocybin triggers a single profound shift in the brain’s default mode network, often described by participants as among the most meaningful experiences of their lives, which then appears to enable lasting psychological reorganisation. These are not variations of the same intervention. They are different kinds of change entirely.
Carhart-Harris has long argued that the expectancy effects so central to SSRI response may not apply in the same way to psychedelics, where the experience itself seems to generate the therapeutic signal, independent of what patients hoped for going in. That is not a minor caveat. It describes a fundamentally different healing process.
Side Effects
And the comparison looks even starker when you hold the two treatments side by side on their own terms. SSRIs are taken daily, often indefinitely. Their side effect profile is well-documented: sexual dysfunction, emotional blunting, weight gain, and discontinuation effects that can make stopping them a project in itself.
Psilocybin-assisted therapy, by contrast, typically involves one or two sessions. Effects have been shown to persist for months, and in some cases much longer. A 2025 trial found a single psilocybin session delivered measurable depression relief in cancer patients two years later.
The emerging side-effect picture is, to put it mildly, different. Preclinical research published in 2024 and 2025 has shown psilocybin and its active metabolite psilocin suppressing pro-inflammatory cytokines like TNF-α and IL-6, suggesting immunomodulatory effects that go well beyond the brain. A 2025 study published in npj Aging found that psilocin extended cellular lifespan in human cells by more than 50% and improved survival in aged mice. These are early findings and they need replication in humans at scale. But they point in a direction that no one studying SSRIs is pointing.
This is not to say psilocybin is without risk. It carries real contraindications, requires careful clinical support, and is not appropriate for everyone. But the risk profile is categorically different, and the fact that a drug taken twice might do what a daily pill does indefinitely, while also showing signals of systemic anti-inflammatory and cellular benefits, should be part of the story.
This mirrors a tension we see everywhere in how modern systems approach biology. The industrial model wants standardisable, repeatable, scalable interventions that can be compared in a spreadsheet. The regenerative model recognises that living systems change through relationship, context, and transformation.
What This Tells Us
None of this is to say the Williams-Barnett-Szigeti paper is wrong or unimportant. Blinding bias in psychedelic research is a problem the field needs to reckon with. Overhyped claims do real damage. The study’s call for rigour is necessary.
But the conclusion that psychedelics are “no better” than SSRIs, served raw to a general audience, flattens a more complicated and more interesting story. As Carhart-Harris’s detailed rebuttal makes clear, the antidepressant trials used as the benchmark are themselves deeply flawed. Thet are products of an era of drug development that modern researchers would not consider acceptable. Comparing psilocybin to that baseline and calling it even is not rigour.
The debate ultimately exposes a field still operating on a mechanical model of the mind at the very moment our understanding of healing is maturing toward complex living systems. Psilocybin doesn’t slot neatly into that old framework, and the friction that creates isn’t a problem with the drug. It’s a problem with the framework.
That understanding is still being built. Studies like this one are part of it. So is the pushback they produce.
If you’re genuinely curious about the science, culture, and evolving story of fungi, The Spore Report is worth your time — a newsletter that covers the research and ideas without the hype.
Read The Spore Report →
