Psilocybin and Alcohol Use Disorder: What the NEJM Trial Found

Alcohol use disorder is one of the most common and most lethal psychiatric conditions in the world, and one of the most stubbornly resistant to treatment. Roughly 29.5 million Americans met diagnostic criteria for AUD in 2020, yet the three medications the FDA has approved — naltrexone, acamprosate, and disulfiram — produce only modest effects and depend on a daily adherence that most patients cannot sustain. Against that backdrop, a 2022 randomised controlled trial reported something that does not happen in addiction medicine: an 83% reduction in heavy drinking days from just two dosing sessions of psilocybin.

The trial, led by Michael Bogenschutz and published in JAMA Psychiatry, was the first adequately powered, double-blind test of psilocybin-assisted therapy for alcohol use disorder. Its results have reframed a question the field had largely abandoned: not whether a drug can blunt craving day by day, but whether a small number of profound experiences, properly supported, can change the neural architecture that sustains addiction itself. Understanding why requires understanding what alcohol does to the brain — and what psilocybin does to undo it.

The Disorder That Survives Abstinence

Alcohol use disorder is not simply heavy drinking. It is a chronic, relapsing brain disorder defined by compulsive use, loss of control over intake, and the emergence of negative emotional states — anxiety, dysphoria, restlessness — during abstinence. Chronic exposure to alcohol drives lasting neuroadaptations in the brain’s GABA and glutamate systems and degrades the prefrontal circuitry responsible for inhibitory control. Critically, these changes do not reverse the moment a person stops drinking. They persist for months, and they are a principal engine of relapse.

This is why conventional pharmacotherapy struggles. Naltrexone, acamprosate, and disulfiram each target a symptom — reward, craving, or aversive conditioning — without altering the underlying architecture that makes alcohol feel necessary to the self. They must be taken daily, indefinitely, by people whose disorder specifically erodes the capacity for sustained daily behaviour change. The result is a treatment ceiling the field has not meaningfully moved in decades.

How Psilocybin Works: 5-HT2A and the Entropic Brain

Psilocybin is a prodrug. After ingestion it is rapidly dephosphorylated to psilocin, which acts as a partial agonist at the serotonin 2A (5-HT2A) receptor — densely expressed on the pyramidal neurons of the prefrontal cortex and throughout the brain’s association networks. Activation of these receptors produces a transient but dramatic increase in cortical excitability and, with it, a measurable rise in the entropy of brain activity: the ordinarily tight, top-down constraints that the brain places on its own processing loosen.

Carhart-Harris and colleagues formalised this as the entropic brain account. Under psilocybin, the hierarchical predictive machinery that keeps cognition running along well-worn grooves is temporarily destabilised. For a healthy brain this manifests as the characteristic fluidity of the psychedelic state. For an addicted brain — one locked into rigid, self-perpetuating patterns of craving and identity — that same destabilisation opens a window in which those patterns can be revised rather than merely resisted.

The Default Mode Network and the Addicted Self

The default mode network — anchored in the medial prefrontal cortex, posterior cingulate cortex, and angular gyrus — is the brain’s self-referential system. It generates the continuous narrative of who we are: our autobiography, our habits, our sense of being a particular kind of person. In addiction, the DMN does something corrosive. It binds the identity of “drinker” into the self, runs the ruminative loops that precede relapse, and makes the prospect of abstinence feel less like a choice than like the loss of a self.

Carhart-Harris’s 2012 neuroimaging work showed that psilocybin acutely deactivates the posterior cingulate and medial prefrontal cortex, and that the degree of this deactivation tracks the intensity of ego dissolution — the experience of the ordinary self becoming porous or falling away. For someone with alcohol use disorder, this is not an abstraction. It is a few hours in which the fixed story of “I am someone who drinks” becomes, briefly, fluid. The therapeutic work done in and after that window — reframing, re-deciding, re-authoring — is what determines whether the change endures.

Bogenschutz 2015: The Proof of Concept

The modern psilocybin-for-AUD programme began with a small open-label pilot. In 2015, Bogenschutz and colleagues gave ten participants with DSM-IV alcohol dependence two psilocybin sessions embedded within a motivational enhancement therapy framework. Drinking fell sharply, with the largest reductions in the weeks immediately following the sessions — the neuroplasticity window — and abstinence held substantially above baseline at 36-week follow-up. Crucially, the intensity of the mystical experience predicted the magnitude of the drinking reduction. There were no serious adverse events. The pilot was uncontrolled and tiny, but it established both the signal and the mechanism worth testing at scale.

Bogenschutz 2022: The Randomised Controlled Trial

The definitive test followed in 2022. Ninety-three adults with alcohol use disorder were randomised, double-blind, to receive either psilocybin (25 mg/70 kg) or an active diphenhydramine placebo across two dosing sessions, with both arms receiving the same twelve weeks of psychotherapy. The primary outcome was the percentage of heavy drinking days across a 32-week follow-up.

The psilocybin group recorded an 83% reduction in heavy drinking days, compared with 51% in the placebo group. Forty-eight percent of psilocybin participants reported no heavy drinking at all during the follow-up window, versus 24% on placebo. Secondary measures — drinks per day, World Health Organization risk levels, craving — all favoured psilocybin, and no serious adverse events were attributed to the drug. For a field accustomed to single-digit percentage-point improvements, the magnitude and the durability of the effect were striking.

The Mystical Experience as Predictor

One of the most consistent and most provocative findings across the psilocybin literature is that how the session is experienced predicts the clinical outcome. Using the validated Mystical Experience Questionnaire, both the 2015 pilot and the 2022 trial found that participants who underwent more complete mystical-type experiences — characterised by unity, transcendence of time and space, and noetic insight — showed the greatest reductions in drinking.

This relationship recurs across psilocybin trials for depression, end-of-life distress, and tobacco dependence, and it points to a mechanism rather than a coincidence. The depth of the experience appears to index the depth of default mode network disruption — the degree to which a person genuinely encounters a self not organised around the addiction. The more complete that encounter, the more therapeutic work the window permits, and the more durably the new framing is encoded.

How Psilocybin Compares to Existing Medications

Naltrexone, the most effective approved AUD medication, reduces heavy drinking by roughly 20–25% relative to placebo in trials such as COMBINE (Anton, 2006) — and only while it is taken daily, with real-world adherence rarely exceeding half of patients at six months. Psilocybin, in the 2022 trial, produced an 83% reduction in heavy drinking days from two sessions, with the benefit sustained across eight months. The comparison is confounded by the intensive psychotherapy that surrounds psilocybin dosing, and the two approaches are not interchangeable. But the magnitude and persistence of the psilocybin effect are of a different order than anything a daily pill has achieved — and they fit a different theory of change: not symptom suppression, but a brief, structural reorganisation.

Limitations

The evidence is genuinely promising and genuinely preliminary. The most serious problem is blinding: diphenhydramine is not a convincing active control, and the unmistakable subjective effects of psilocybin mean participants and therapists can usually tell who received what, inflating effect sizes through expectancy. The 2022 sample, at 93 participants, is modest by the standards of registration trials. The protocol — high-dose sessions with two therapists and extensive preparation and integration — raises real questions about scalability and cost. Medically complex patients were excluded, the contribution of drug versus psychotherapy cannot be fully disentangled, and data beyond 32 weeks remain sparse. A larger replication with a more credible active comparator is the necessary next step.

What the Evidence Shows

Psilocybin is not a cure for alcohol use disorder, and it is not a pill to be swallowed each morning. It is a tool that creates a specific, time-limited neurobiological state: heightened cortical entropy, suppression of the self-referential default mode network, and a window of plasticity in which the rigid identity of the drinker can be loosened and revised. What is done within that window — the preparation, the experience, and above all the integration — is what heals.

For a disorder that has defied incremental progress for half a century, and that shortens and ends millions of lives, an 83% reduction in heavy drinking days from two carefully supported sessions is among the most important findings in modern addiction science. The regulatory and practical questions are real and unresolved. The biological case — that you can treat a chronic relapsing condition by changing the brain’s architecture rather than masking its symptoms — is now serious, and testable, and being tested.