Psilocybin and Alcohol Use Disorder: What the NEJM Trial Found

In 1985, a DEA scheduling petition triggered a series of hearings on a compound that had been quietly used in psychotherapy practices throughout the 1970s. Alexander Shulgin, the chemist who had re-synthesised it and introduced it to therapists, called it “the ideal drug for psychotherapy.” Administrative Law Judge Francis Young concluded that MDMA had “a currently accepted medical use in treatment in the United States.” The DEA overrode the ruling and placed MDMA in Schedule I.

Four decades later, two Phase 3 randomised controlled trials published in Nature Medicine have confirmed what those therapists were observing in their practices. MDMA-assisted therapy produces clinically significant reductions in PTSD severity in populations for whom every existing treatment had failed. Understanding why requires understanding what MDMA actually does to the brain — a story that begins not with serotonin alone but with a specific neurochemical combination that has no precedent in the existing pharmacopoeia.

MDMA is not a psychedelic. It does not produce visual distortions, ego dissolution, or the entropic disruption of predictive coding that characterises psilocybin. It is an empathogen — a compound that produces a state of emotional openness, social connection, and reduced defensive responding while leaving cognitive clarity largely intact. The wall between self and other does not dissolve because the self disappears. It dissolves because the circuits that maintain it — threat detection, emotional armour, the defensive scanning of social cues for danger — are temporarily quieted.

The Neurochemical Storm: Serotonin, Dopamine, Norepinephrine

MDMA’s pharmacological mechanism is fundamentally different from both classical psychedelics and conventional antidepressants. SSRIs block the reuptake transporter, preventing serotonin from being cleared from the synapse and increasing its dwell time. MDMA does something more aggressive: it enters the presynaptic terminal as a substrate at the serotonin transporter (SERT), dopamine transporter (DAT), and norepinephrine transporter (NET), and reverses the direction of transport.

Rather than blocking clearance, MDMA causes the transporters to run in reverse — actively pumping monoamines out of the presynaptic terminal into the synapse from intracellular stores. The result is a massive, rapid increase in synaptic serotonin, dopamine, and norepinephrine concentrations that dwarfs what any reuptake inhibitor can achieve. Serotonin release is approximately 3–5 times greater than dopamine release under MDMA, which distinguishes it pharmacologically from amphetamines (which preferentially release dopamine) and gives it its characteristic psychological profile.

The massive serotonin surge produces most of MDMA’s psychological effects: elevated mood, increased sociability, reduced anxiety, heightened empathy, and a sense of emotional openness that users reliably describe as removing the filter between what they feel and what they can express. The dopamine component adds motivational salience and reward valence to social interactions. The norepinephrine component contributes to the mild stimulant properties and some of the cardiovascular effects.

Critically, MDMA also triggers the release of serotonin from dense-core vesicles in the hypothalamus — neurons that co-release serotonin alongside oxytocin. This is where the empathogenic effect goes beyond simple mood elevation.

Oxytocin: The Bonding Signal

Oxytocin is a neuropeptide produced in the hypothalamus and released both peripherally (into the bloodstream) and centrally (into the brain). Its peripheral role in uterine contraction and lactation is well-established. Its central role is the subject of intense research: oxytocin appears to be a core mediator of social bonding, trust, and the capacity to interpret the intentions of others as benign rather than threatening.

Dumont et al. (2009) measured plasma oxytocin concentrations in healthy volunteers before and after MDMA administration in a controlled clinical setting. Oxytocin concentrations approximately doubled following MDMA ingestion. Crucially, the increase in oxytocin correlated with the subjective ratings of prosocial effects — feelings of closeness, openness, and social belonging — suggesting that the oxytocin release is not an epiphenomenon but a driver of MDMA’s empathogenic quality.

Thompson et al. (2007) provided pharmacological evidence linking the oxytocin mechanism to 5-HT1A receptor activation, suggesting that MDMA’s serotonin release drives oxytocin secretion via a specific receptor pathway. Later work would identify hypothalamic oxytocin neurons bearing serotonin receptors as the direct link between MDMA’s serotonergic mechanism and its prosocial effects.

The Amygdala and the Fear Extinction Window

The amygdala is the brain’s threat-detection engine. In PTSD, it is chronically over-activated — generating fear responses to cues associated with the original trauma long after any real threat has passed. The prefrontal cortex, which should exert inhibitory top-down control over the amygdala, is functionally impaired in PTSD patients, unable to execute the extinction of fear memories that ordinarily allows threat responses to extinguish when the threat is no longer present.

Neuroimaging studies of MDMA’s acute effects show a consistent finding: amygdala reactivity to threatening stimuli is reduced under MDMA, while reactivity to positive social stimuli is preserved or enhanced. The amygdala does not go offline — it shifts its responsiveness away from threat and toward connection. Simultaneously, prefrontal-amygdala connectivity increases, improving the top-down regulatory capacity that PTSD has compromised.

This creates a therapeutic window. The traumatic memory — normally inaccessible because approaching it triggers the amygdala into hyperarousal that floods the system and prevents cognitive processing — can now be approached with the therapist’s support. The fear response is dampened enough that the patient can stay within their window of tolerance. The therapeutic relationship, enhanced by the oxytocin-driven sense of trust and connection, provides the interpersonal container within which reprocessing can occur. MDMA does not erase the trauma. It creates a neurobiological state in which the brain can do the work that trauma has made impossible.

Nardou 2019: The Critical Period That MDMA Reopens

In 2019, Romain Nardou and colleagues at MIT and Harvard published a study in Nature that reframed how the field thinks about MDMA’s mechanism of therapeutic action. The finding was both elegant and deeply significant: MDMA reopens the critical period for social reward learning in the adult brain.

Critical periods are developmental windows during which specific types of learning occur with exceptional efficiency — the young brain’s extraordinary plasticity allowing it to acquire language, form social attachments, and lay down the fundamental neural architecture of the self. These windows normally close as the brain matures. The social reward learning critical period — the window during which social experiences have their most profound and lasting impact on the architecture of the social brain — closes in late adolescence under normal developmental conditions.

Nardou et al. demonstrated that MDMA, via oxytocin signalling acting on metabotropic glutamate receptor 5 (mGluR5) in the nucleus accumbens, transiently re-opens this critical period in adult mice. During the hours following MDMA administration, the adult social brain recovers something of the plasticity it had during development — a heightened capacity for social experiences to produce lasting changes in the neural circuitry of social reward and attachment.

This mechanism explains something that was previously difficult to account for: why MDMA-assisted therapy produces therapeutic benefits that persist for months and years after the drug has left the system, from only two or three sessions. The drug is not providing ongoing pharmacological support — it is opening a window of neuroplasticity during which the psychotherapy can produce changes that outlast the drug effect by orders of magnitude. What is learned in the critical period, whether in development or in its re-opened form, persists.

Mitchell 2021: The First Phase 3 Trial

The first Phase 3 randomised controlled trial of MDMA-assisted therapy for PTSD, led by Jennifer Mitchell at UCSF and conducted under the MAPS protocol, was published in Nature Medicine in 2021. The trial enrolled 90 adults with severe, chronic, treatment-resistant PTSD — a population who had failed multiple prior treatment attempts — at clinical sites across the United States, Canada, and Israel.

Participants were randomly assigned to receive MDMA-assisted therapy or placebo-assisted therapy. Both groups received identical psychotherapy support: preparatory sessions before the drug sessions, two or three experimental sessions (each 8 hours in duration with two therapists present), and integration sessions following each drug session. The primary outcome was change in CAPS-5 score — the Clinician-Administered PTSD Scale, the field’s gold-standard severity measure.

The results were the strongest ever reported in a Phase 3 PTSD trial. The mean CAPS-5 reduction was 24.4 points in the MDMA group versus 13.9 points in the placebo group — a clinically and statistically significant difference. Crucially, 67% of participants in the MDMA group no longer met diagnostic criteria for PTSD at the primary endpoint, compared with 32% in the placebo group. Functional impairment showed similarly large between-group differences. The effect size substantially exceeded anything previously reported with first-line PTSD pharmacotherapies including SSRIs and prazosin.

The safety profile was consistent with prior Phase 2 data. Adverse events were predominantly mild and transient — nausea, hyperhidrosis, and insomnia — and resolved without intervention. No serious adverse events attributable to MDMA were recorded.

Mitchell 2023: Replication at Scale

The second Phase 3 trial, also led by Mitchell and conducted across multiple international sites including in Europe and Australia, enrolled 104 adults with moderate-to-severe PTSD and was published in Nature Medicine in 2023. The protocol was identical to the 2021 study, with pre-specified primary and secondary endpoints registered before unblinding.

The replication held. At the primary endpoint, 71.2% of participants in the MDMA group met the PTSD treatment response criterion — defined as a 10-point or greater reduction on the CAPS-5 — compared with 47.6% in the placebo group. Mean CAPS-5 reduction was 23.7 points in the MDMA arm versus 14.8 in the placebo arm. Secondary outcomes including depression, anxiety, functional impairment, and self-reported PTSD severity all showed consistent superiority of MDMA over placebo.

Two large, independent, well-designed Phase 3 trials now point in the same direction. The effect sizes are large. The population is one of the most treatment-resistant in psychiatry. The durability of benefit — documented in follow-up assessments at 12 months — aligns with the Nardou critical period model: the therapeutic work done within the neuroplasticity window persists because it produces structural, not pharmacological, change.

The FDA Decision and the Regulatory Frontier

In June 2024, the FDA’s Psychopharmacologic Drugs Advisory Committee voted 9–2 against recommending approval of midomafetamine (MDMA) for PTSD, and the FDA issued a Complete Response Letter to MAPS in August 2024 declining approval and requesting additional trials. The decision shocked many in the field, given the strength of the Phase 3 evidence, but it reflected genuine regulatory concerns that the scientific community must address.

The primary concern was functional unblinding. MDMA’s subjective effects are unmistakable — participants in trials can readily determine whether they received active drug or placebo, and this knowledge inflates effect size estimates through expectancy effects to a degree that is difficult to quantify. The FDA also raised concerns about the difficulty of standardising and evaluating the therapy component separately from the drug component, about investigator conduct at some sites, and about the cardiovascular risk profile in populations not represented in the trials.

These are legitimate scientific concerns, not dismissals of the evidence. The correct response is not to conclude that MDMA-assisted therapy does not work — the effect sizes are too large and too consistently replicated to be attributable to expectancy alone. The correct response is to design the next generation of trials with expectancy control arms, active placebos that produce some subjective experience without MDMA’s specific neurochemical action, rigorous therapist fidelity monitoring, and broader, more medically complex populations. The regulatory question is not whether MDMA works. It is whether the current evidence base is sufficient to support approval — and what is needed to make it so.

Limitations

Beyond the blinding problem, the MDMA-assisted therapy literature has several limitations that merit honest acknowledgment. Sample sizes, though Phase 3 standard, are small relative to conventional psychiatric drug trials. The therapy protocol — two therapists, eight-hour sessions, intensive preparation and integration — is highly resource-intensive and not readily scalable in current healthcare systems. The populations studied have been relatively healthy adults with PTSD; those with significant cardiovascular disease, active suicidal ideation, or severe dissociative symptoms have been excluded. Long-term safety data for the therapeutic use protocol specifically — as distinct from recreational or chronic use — is limited. And the mechanism by which the two or three session treatment produces durable remission, while theoretically coherent via the critical period model, has not been directly confirmed in human clinical populations.

What the Evidence Shows

Across two Phase 3 trials, a consistent finding has emerged: MDMA-assisted therapy produces clinically significant reductions in PTSD severity — and in rates of diagnostic-criteria loss — that substantially exceed those achieved by any existing pharmacological or psychotherapeutic intervention for the same population. The neurochemical mechanism — massive serotonin release driving oxytocin surge, amygdala suppression, and the transient re-opening of the social reward learning critical period — provides a coherent explanation for why this works and why the effects last.

MDMA is not a magic molecule. It is a specific tool that creates a specific neurobiological state: reduced defensive responding, increased social trust, and a brief window of heightened emotional plasticity. What is done within that window — the therapeutic work, the relational repair, the careful reprocessing of what could not previously be approached — is what heals. The drug creates the conditions. The therapy does the work.

For PTSD — a disorder that has resisted every existing treatment for decades, and that destroys the lives of millions of people who carry trauma they cannot discharge through any available means — the emergence of a treatment with 67–71% response rates in the most treatment-resistant patients is among the most significant clinical findings in modern psychiatry. The regulatory path is not clear. The science is.