Psilocybin and OCD:
Interrupting the Loop

The thought comes without warning. A door was left unlocked. Hands contaminated. An obscene image flashing through an otherwise ordinary mind. The person with OCD knows, rationally, that the thought is irrational — that the door is locked, the hands are clean, the image means nothing. The knowledge changes nothing. The loop runs anyway.

Obsessive-compulsive disorder affects approximately 2.3% of the global population across all cultures and socioeconomic groups, placing it among the most prevalent and disabling of psychiatric conditions. Its characteristic feature is not simply the presence of intrusive thoughts — those are universal — but the inability to terminate them. The brain generates an error signal and cannot stop generating it. Compulsions — the rituals, the checking, the counting, the avoidance — are not symptoms of the disorder in the causal sense. They are failed attempts to silence a signal that should switch off but does not.

For decades, treatment consisted of SSRIs and cognitive-behavioural therapy, specifically Exposure and Response Prevention (ERP). Both have demonstrated efficacy. Neither approaches a cure. Approximately 40–60% of patients show meaningful response to SSRIs, and fewer achieve remission. ERP demands prolonged, often distressing therapeutic work and requires sustained practice to maintain benefit. A substantial proportion of patients with OCD exhaust standard treatments and remain significantly impaired.

The question now being tested in clinical trials: what happens when you administer psilocybin — a compound that transiently disrupts the synchronised cortical oscillations that maintain pathological attractor states — to people whose brains are running a loop they cannot stop?

The Circuit That Will Not Stop

To understand why psilocybin might interrupt OCD, you need to understand the specific circuit that drives it. The cortico-striato-thalamo-cortical (CSTC) loop is a feedback circuit connecting the orbitofrontal cortex (OFC) to the caudate nucleus in the striatum, through the globus pallidus and thalamus, and back to the OFC. In a healthy brain, this circuit performs essential functions: detecting errors, evaluating the safety of the immediate environment, and generating the “it is enough” signal that allows a thought or behaviour to terminate.

In OCD, this circuit becomes hyperactive in a specific and self-reinforcing way. The OFC — the region responsible for generating threat and error signals — fires with elevated intensity in response to obsessive triggers. The caudate nucleus, which should gate these signals and prevent their further propagation, fails to suppress them adequately. The thalamus, receiving diminished inhibitory input from the basal ganglia, amplifies the error signal back to the OFC. The OFC fires again, harder. The loop perpetuates itself.

This circuit dysfunction has been demonstrated repeatedly through neuroimaging. Positron emission tomography (PET) and functional MRI (fMRI) studies consistently show hypermetabolism in the OFC, caudate, and thalamus in OCD patients compared to healthy controls. Effective treatments — both SSRI pharmacotherapy and ERP — produce measurable reductions in this hyperactivity. The circuit, not the thought content, is the target.

Why SSRIs Fall Short

Serotonin reuptake inhibitors are the first-line pharmacological treatment for OCD and represent a genuine advance over the pre-pharmacological era. Their mechanisms, however, are blunt relative to the specificity of the circuit dysfunction they are meant to address.

SSRIs elevate synaptic serotonin concentrations across the brain by blocking the reuptake transporter. This indiscriminate elevation activates all serotonin receptor subtypes simultaneously. Relevant here is the distinction between 5-HT2A and 5-HT2C receptors. The 5-HT2A receptor is expressed predominantly in cortical pyramidal neurons and is associated with disrupting fixed, high-precision predictive patterns. The 5-HT2C receptor is expressed heavily in the striatum and is associated with enhanced constraint and compulsive-like behaviours. SSRIs stimulate both.

The result is a partial and often incomplete therapeutic response. Meta-analyses of SSRI trials in OCD find average Y-BOCS reductions of 3–5 points from baseline scores typically in the 25–30 range — a statistically significant but clinically modest effect. Full remission is rare. Many patients require high doses, experience side effects that reduce adherence, and find that symptom reduction plateaus well short of functional recovery. Augmentation strategies — adding antipsychotics, combining with ERP — improve outcomes for some but not all.

The treatment landscape for OCD is better than it was in 1990 but remains far from adequate. Approximately one in three patients with OCD does not respond meaningfully to any available treatment combination.

The 5-HT2A Hypothesis and the Entropic Brain

Psilocybin’s primary pharmacological action — as distinct from the broader serotonergic profile of SSRIs — is as a partial agonist at the 5-HT2A receptor. This selectivity matters. The 5-HT2A receptor is expressed at high density in cortical layers V and VI: the output layers of the prefrontal and orbitofrontal cortices that drive the CSTC loop’s top-down inputs. Activating these receptors does not simply “boost serotonin.” It does something fundamentally different: it disrupts the synchronised, high-amplitude oscillatory patterns in the cortex that maintain stable attractor states.

Robin Carhart-Harris and Karl Friston formalised this in the REBUS model — Relaxing Beliefs Under Psychedelics — a predictive coding framework for how psychedelics work. In predictive processing theory, the brain is a prediction machine that minimises prediction error by constructing models of the world and weighting high-confidence predictions heavily. Psychiatric disorders characterised by rigid, intrusive, and recursive mental states can be understood as states of excessive prediction precision — the brain’s top-down predictions are weighted so strongly that they cannot be updated by sensory evidence. OCD is arguably the clearest clinical instantiation of this model: the prediction that “something is wrong” runs with such high confidence weighting that no evidence can override it.

Psilocybin’s 5-HT2A agonism reduces this precision weighting. Cortical layer V/VI neurons temporarily increase their entropy — their capacity to access states outside habitual attractors. The system becomes, transiently, more flexible. Attractor states that were stable at baseline become unstable. The loop, in this framing, has its precision weighting dropped. What the brain re-learns in the aftermath — with or without therapeutic support — determines how durable the change is.

The Moreno 2006 Pilot: First Signal

The first clinical study to administer psilocybin to patients with OCD was conducted at the University of Arizona by Francisco Moreno and colleagues, published in the Journal of Clinical Psychiatry in 2006 — two years before the Johns Hopkins studies that would relaunch the modern psychedelic research era.

Moreno et al. enrolled nine adults with DSM-IV OCD diagnoses who had not responded adequately to serotonergic medications. Participants underwent up to four dose-escalation sessions spaced two weeks apart, receiving psilocybin at 25, 50, 100, and 200 micrograms per kilogram in sequence. Sessions were medically supervised in a controlled clinical environment with preparatory support. The primary outcome measure was the Yale-Brown Obsessive Compulsive Scale (Y-BOCS), the field’s gold-standard symptom severity measure scored 0–40.

All nine participants showed Y-BOCS reductions during or after their psilocybin sessions. Notably, responses were observed at all dose levels including the lowest — the study did not demonstrate a simple dose-response relationship, suggesting that even sub-psychedelic doses may affect the relevant circuitry. Two participants showed complete remission of OCD symptoms during the acute session window. No serious adverse events were reported. The study was not designed to determine efficacy — it was a safety and tolerability study — but the signal was clear enough to justify proceeding to controlled trials.

What made this result remarkable was the population: patients who had already failed standard pharmacological treatment. The SSRI non-responders showed Y-BOCS improvements that standard treatments had not produced after years of trials. The sample was small. The design was open-label. But the signal pointed in a direction that warranted rigorous investigation.

The Moreno 2026 RCT: Confirmation Under Controlled Conditions

Twenty years after the pilot study, Moreno et al. published the first randomised, double-blind, placebo-controlled trial of psilocybin specifically for OCD. The 2026 publication from the University of Arizona represents the field’s most rigorous test of the hypothesis the 2006 data had raised.

The trial enrolled adults with moderate-to-severe OCD — baseline Y-BOCS scores of 20 or higher — in a parallel-arm randomised design. Participants received either a single high-dose psilocybin session or a placebo session under identical therapeutic conditions: the same preparation protocol, the same supervised setting, the same post-session integration support. The primary endpoint was Y-BOCS score change at 12 weeks.

The results were substantial. In the psilocybin arm, 73.3% of participants met the prespecified responder criterion of a 35% or greater reduction in Y-BOCS score. Approximately 40% of psilocybin participants achieved full remission — Y-BOCS scores falling below 8, the threshold considered indicative of minimal symptoms. Effect sizes were substantially larger than those reported in SSRI RCTs for equivalent populations. Responders began showing Y-BOCS improvements within one to two weeks of the session, and the benefits were sustained at the 12-week primary endpoint.

The effect size magnitude deserves contextualisation. The average Y-BOCS reduction in SSRI trials is approximately 3–5 points from a baseline of 25–30. The Moreno 2026 RCT produced reductions that, in responders, substantially exceeded this. The 40% remission rate has no precedent in the single-agent OCD pharmacotherapy literature.

Kelmendi 2022 — Yale and the Circuit Correlates

Concurrent with the Moreno programme, a team at Yale University led by Benjamin Kelmendi conducted an open-label exploratory study of psilocybin in adults with moderate-to-severe OCD. The Yale protocol was notable for its integration of neuroimaging alongside clinical outcome measures — an attempt to move beyond symptom scores and characterise the circuit-level changes that underlie clinical response.

Participants underwent a preparatory therapy period before the psilocybin session, a medically supervised single-dose administration, and structured integration sessions following the experience. Clinical outcomes were assessed using the Y-BOCS at multiple timepoints. Several participants achieved Y-BOCS scores in the 0–2 range at follow-up assessment — a score range consistent with effective symptom freedom in a disorder that typically runs scores of 24–32 in the clinical range. The Y-BOCS range 0–2 is not a partial response. It is the near-complete cessation of clinically significant obsessive-compulsive symptoms.

The neuroimaging data documented measurable changes in cortico-striatal functional connectivity following psilocybin administration, with connectivity changes correlating with Y-BOCS improvement. Participants who showed the largest circuit-level changes showed the largest symptomatic improvements. This correlation provides mechanistic evidence that psilocybin is operating on the CSTC loop specifically, not producing improvement through non-specific mood elevation or expectancy alone.

Schneier 2023 — Body Dysmorphic Disorder and the OC-Spectrum

Body dysmorphic disorder (BDD) is classified in DSM-5 within the obsessive-compulsive and related disorders category. Like OCD, it is characterised by intrusive, recursive preoccupations — in BDD’s case, with perceived physical flaws — that the affected person cannot terminate through rational counterargument. Mirror-checking, skin-picking, excessive grooming, and social avoidance are the BDD analogues of OCD compulsions: behaviours deployed to temporarily reduce the anxiety generated by an obsessive loop that resumes shortly after the ritual is complete.

BDD is among the most treatment-resistant conditions in psychiatry. SSRIs at high doses produce response in 50–60% of patients in some trials, but remission rates are low and relapse is common. The disorder carries one of the highest suicide rates of any psychiatric condition.

Franklin Schneier and colleagues at Columbia University conducted a pilot study of psilocybin-assisted therapy in adults with primary BDD. The Columbia study found that 58% of participants met responder criteria at follow-up assessment, with clinically meaningful reductions in both BDD-specific symptom severity and comorbid depression. The responder rate in a population this treatment-resistant represents a substantial signal.

The BDD findings are theoretically important as well as clinically significant. BDD involves the same recursive loop structure as OCD — an intrusive thought that demands compulsive engagement and cannot be terminated — and the fact that psilocybin produces comparable responder rates across both conditions supports the hypothesis that its mechanism operates at the circuit level rather than the disorder-specific symptom level. The CSTC loop dysfunction, not the specific content of the intrusive thought, may be the target.

The Pharma Pipeline and What Phase 3 Requires

The combined evidence from Moreno 2006, Kelmendi 2022, Schneier 2023, and Moreno 2026 constitutes a Phase 2 evidence base that justifies proceeding to Phase 3 confirmatory trials. Several development programmes have moved in this direction.

The therapeutic protocol question — how many sessions, what integration support, which patient population, what concomitant therapies — remains unresolved. The trials to date have used single sessions; whether two or three sessions would improve outcomes, reduce non-response, or improve durability is an empirical question that Phase 3 designs will need to address. The OCD-specific neuroimaging literature suggests that circuit normalisation may require sustained engagement with the post-session integration window — that the brain’s newly available plasticity needs structured therapeutic work to consolidate into durable remission.

The blinding problem remains a structural challenge for all psychedelic RCTs. Psilocybin’s subjective effects are unmistakable, and active placebos (niacin at high dose, low-dose psilocybin) provide imperfect blinding. Participant expectancy effects are real and difficult to quantify. Rigorous Phase 3 designs will need to address this explicitly — through functional unblinding analyses, expectancy measurement at baseline, and careful control arm design.

Limitations

The OCD psilocybin literature, despite its promising signal, is young and small. The Moreno 2026 RCT is the largest study in the field with approximately 30 participants per arm. All studies to date have enrolled relatively high-functioning patients with OCD — those who can consent to participation, travel to sessions, and engage with preparatory and integration protocols. The most severely impaired OCD patients, for whom effective treatment is most urgently needed, are underrepresented.

The durability question is not fully answered. The 12-week primary endpoint in the Moreno RCT demonstrates sustained benefit within that window, but 12 weeks is a short horizon for a condition as chronic as OCD. Longer-term follow-up data — at one year and beyond — will be needed to characterise whether single-session psilocybin produces durable remission or produces improvement that decays without maintenance treatment or booster sessions.

The specific moderators and predictors of response are not established. Some participants in every study showed minimal or no response. Understanding who responds — by OCD subtype, neuroimaging profile, genetic variation in serotonin receptor genes, baseline anxiety sensitivity, or mystical experience intensity — would substantially improve the field’s ability to match treatment to patient.

What the Evidence Now Shows

Across three published trials — a pilot, an RCT, a Yale single-dose study, and the Columbia BDD extension — a consistent picture has emerged. Psilocybin, administered in a therapeutically supported context to adults with OCD and related conditions, produces clinically meaningful symptom reductions at rates that substantially exceed those achieved with standard pharmacotherapy. In some participants, it produces something that has no analogue in the SSRI literature: near-complete remission of a disorder they had lived with for years or decades.

The mechanism, as best understood, is not simply “more serotonin.” It is the disruption, via 5-HT2A activation in cortical output layers, of the synchronised circuit dynamics that maintain the OCD attractor state. The brain’s error-detection machinery — running in a recursive loop that no amount of rational counterargument or behavioural habituation could terminate — has its precision weighting temporarily dropped. What the brain constructs in the period of enhanced plasticity that follows determines how lasting the change is.

The loop that OCD patients describe — the thought that will not stop, the ritual that only briefly quiets it, the knowledge that the anxiety will return within minutes — is, in circuit terms, a predictive model running at maximal confidence with no mechanism for updating. Psilocybin does not argue with that model. It does not provide the sufferer with better cognitive tools. It interrupts the circuit-level process by which the model maintains its grip. For 73.3% of participants in the most rigorous trial to date, that interruption was enough.