Psilocybin for Eating Disorders: Evidence, Mechanism, and the Identity Layer

Psilocybin for eating disorders sits at the intersection of an unusually compelling mechanism and an unusually thin evidence base, with the additional complication that eating disorders are medically dangerous in a way that depression and OCD are not. An estimated 30 million Americans will experience an eating disorder in their lifetime, and anorexia nervosa carries a 5.6 percent lifetime mortality rate, according to NIMH eating disorder statistics and the meta-analysis by Arcelus and colleagues (Archives of General Psychiatry, 2011). That mortality figure is what makes preliminary psilocybin data worth taking seriously, and what makes it equally important not to overpromise.

The argument for psilocybin in eating disorders is not that it restores eating. The argument is that it acts on a brain network, the default mode network, that maintains the fixed self-concept on which anorexia in particular depends. Standard treatments reach the cognitive content of the disorder. They rarely reach the level at which the patient experiences the body and the control of intake as essential to who they are. Psilocybin acts on that level pharmacologically, in hours rather than years.

The honest version of this article is that psilocybin for eating disorders is a credible research direction with one published Phase 1 trial, one ongoing Phase 2 trial, and a coherent mechanistic story. It is not a treatment people can access outside of clinical research, and it is never an alternative to medical stabilization. What I can usefully add is what the identity-layer work for body-fixated clients actually looks like in integration, and why the standard psychedelic integration template often misses the substrate that eating disorders run on.

Why Do Eating Disorders Resist Standard Treatment?

Eating disorders show the lowest treatment response rates of any psychiatric category, with anorexia carrying the highest mortality rate at 5.6 percent lifetime, according to Arcelus and colleagues (Archives of General Psychiatry, 2011), a meta-analysis of 36 studies covering more than 17,000 patients. Much of that mortality is medical, from cardiac and metabolic complications of starvation, but a substantial portion is from suicide. The disorder is biologically and psychiatrically severe in a way that frequently exceeds what current treatment can reach.

The standard treatments, CBT-E for adults and family-based therapy for adolescents, have outcome data but with a strikingly persistent residual group. Roughly half of treated anorexia patients achieve full remission, a quarter improve partially, and a quarter remain chronically ill. SSRIs work poorly for anorexia in the underweight state and modestly for bulimia and binge-eating disorder. The disorder appears to have a structural quality that does not yield to symptom-targeted intervention the way depression often does.

The reason, in the clinical pattern I have observed and that the research literature increasingly converges on, is that eating disorders are not really food disorders. They are identity disorders that use food and body as the medium. The control of intake produces a felt sense of self that the patient cannot otherwise access. Asking them to relinquish the control without first reaching what the control is doing for them tends to feel, from inside the disorder, like asking them to relinquish their identity. That request is harder than it sounds.

How Does Psilocybin Act on the Eating Disorder Brain?

Neuroimaging of anorexia nervosa consistently shows hyperactivity and elevated coherence in the default mode network, the system that handles self-referential processing and the maintenance of stable self-concept, according to fMRI studies summarized in reviews in Trends in Cognitive Sciences. Psilocybin's 5-HT2A agonism acutely disrupts default mode network coherence, which on a mechanistic level looks like exactly the intervention the neurobiology of body-image rigidity suggests would be useful.

What the Default Mode Network Does

The default mode network is a set of interconnected brain regions, including the medial prefrontal cortex, posterior cingulate, and angular gyrus, that is most active when attention is internally directed. It handles autobiographical memory, self-referential thought, simulation of the self in past and future scenarios, and the construction of the stable self-concept. In eating disorders the network appears to run in a state of elevated, rigid coherence, with body and intake-control features dominating the self-representation in ways that resist update from contradicting information.

5-HT2A and DMN Disruption

The work of Carhart-Harris and colleagues (Frontiers in Human Neuroscience, 2014) on the entropic brain hypothesis describes how 5-HT2A agonism on cortical pyramidal neurons increases the diversity of accessible brain states and reduces the dominance of constrained, repetitive activity patterns. In eating disorders this maps directly to the DMN. The rigid self-concept is a constrained pattern. Psilocybin temporarily increases entropy in the system, which loosens the grip of the fixed self-representation and creates a window in which alternative self-concepts become accessible.

The downstream BDNF and TrkB signaling cascade then opens a two to four week period of elevated synaptic plasticity in cortical regions, including those most implicated in self-representation. This is the window in which the experience of a less body-fixated self can begin to consolidate into a new circuit-level baseline. The mechanism is not appetite normalization. It is identity-structure loosening plus a plasticity window for new self-organization.

The Spriggs Data on Body-Image Rigidity

Spriggs and colleagues (Journal of Psychopharmacology, 2021) documented reductions in the rigidity of body-related thoughts following psilocybin use in a non-clinical sample, with body-image flexibility increasing in the post-session window. The study was not in clinical eating disorder populations, and the inference to clinical relevance is partial. What it offers is mechanistic confirmation that psilocybin can move the body-related cognitive substrate in the direction that eating disorder treatment would want it to move, in healthy users where the baseline is not pathological.

What Does the Trial Evidence Actually Show?

The published clinical evidence for psilocybin in eating disorders consists of one Phase 1 trial in anorexia from Peck and colleagues (Nature Medicine, 2023) with 10 participants, plus an ongoing Phase 2 trial at UCSF (NCT06399263) that began recruitment in 2025. The Peck trial used a single 25 mg dose of psilocybin with psychological support, demonstrated safety and tolerability, and reported significant reductions in body-shape and weight concerns at three months in this medically fragile population.

Peck 2023: The First Anorexia Trial

The Peck study was small, open-label, and primarily designed to establish safety and feasibility rather than efficacy. All ten participants completed the protocol. There were no serious adverse events related to the drug, which mattered substantially given the cardiovascular risk profile of underweight anorexia patients. Secondary outcome measures included the Eating Disorder Examination Questionnaire and measures of body-related cognition. At the three-month follow-up the group showed statistically significant reductions in shape and weight concern subscales, with several participants reporting durable shifts in their relationship to body image.

The trial was not powered to establish efficacy and used no control condition. Its primary contribution was establishing that the protocol is safe in a medically fragile population and that the signal on body-image rigidity is consistent with the mechanistic prediction. That opened the path for the larger trials now in progress.

UCSF Phase 2 and the Current Research Landscape

The UCSF psilocybin trial registered under NCT06399263 is the most rigorous psilocybin-for-anorexia study currently in collection. It is a Phase 2 design with a substantially larger sample, control conditions, and longer follow-up, comparing single-dose psilocybin to active comparator in adults with anorexia nervosa. Results have not yet been published as of mid-2026 and the trial remains in data collection. Beyond UCSF, smaller groups at Johns Hopkins and at Imperial College London have been examining psilocybin in adjacent eating-disorder presentations, including binge eating disorder.

The combined picture is a coordinated research push that should produce substantially more data within the next two to three years. Until then, the responsible position is that the mechanism is well-mapped, the preliminary signal is consistent across the small samples examined, and the trial base is too thin to support clinical claims about routine treatment efficacy. For a parallel case where mechanism precedes a thin trial base, see psilocybin for OCD.

Do Different Eating Disorders Respond Differently?

Eating disorders are not a single condition, and the response profile to any intervention is likely to differ across anorexia nervosa, bulimia nervosa, binge eating disorder, and avoidant/restrictive food intake disorder (ARFID). The DSM-5 separates these conditions because they differ in presentation, mechanism, and treatment response. The current psilocybin research is concentrated in anorexia, which means inference to other subtypes is preliminary at best.

Anorexia Nervosa

Anorexia is the most studied target for psilocybin in eating disorders because the disorder's central feature, a rigid self-concept organized around control of intake and body, is the kind of feature psilocybin's mechanism is best positioned to reach. The Peck 2023 trial and the UCSF Phase 2 both focus on anorexia specifically. The clinical logic is that the identity-structure looseness produced by DMN disruption is the missing layer in standard anorexia treatment, which tends to reach cognitive content and behavior but not the underlying self-concept that sustains them.

Binge Eating Disorder and Bulimia

Binge eating disorder and bulimia have different mechanisms. Both involve cycles of dysregulated intake and, in bulimia, compensatory behaviors, but the underlying structure is closer to impulse-control and emotion-regulation pathology than to identity-rigidity pathology. Psilocybin may help here as well, but through a different angle. The plasticity window after a session could support new emotion-regulation strategies and disrupt the binge-purge or binge-restrict cycles that bulimia and BED run on. There is no published clinical trial of psilocybin in these conditions specifically, though some research interest exists.

ARFID and Other Presentations

ARFID, characterized by restriction of food intake without body-image disturbance, has yet a different mechanism. It often involves sensory aversion, fear of aversive consequences such as choking or vomiting, or low interest in eating. The DMN-disruption rationale that fits anorexia does not obviously fit ARFID. There is no current research on psilocybin in ARFID, and extrapolating from anorexia data would be inappropriate. The point worth keeping in mind is that "eating disorders" is a category that covers structurally different conditions.

What Does the Identity Layer Beneath Food Actually Look Like?

In integration practice with body-fixated clients, the recurring observation is that the eating behavior is the surface and the identity work is the substrate. Roughly one in twelve of the high-functioning clients I work with present with subclinical or clinical eating disorder features, often unrecognized as such because the patterns are work-shaped (the executive who eats once a day, the founder who controls intake as discipline). This is a clinical observation from my own practice rather than a population estimate, and the prevalence is likely higher than what surfaces in standard intake conversations.

The pattern that emerges in deeper integration work follows a recognizable structure. There is an early experience of overwhelm or chaos, often relational. The control of intake provides a felt sense of agency and predictability that the rest of life does not. The body becomes the one thing the person can reliably control, which makes the body's state the primary medium through which they experience their own continuity and identity. The eating behavior is downstream of this. Asking someone to eat differently without addressing the identity function is asking them to release a regulatory mechanism they have nothing to replace it with.

The session, when it works for these clients, shows them the identity function directly. They experience a self that is not organized around body or intake-control. The experience is usually accompanied by what they describe as relief, often followed by grief, and then by the question of what would actually organize identity if not control of body. That question is the integration work. It is not a question the session answers. It is a question the session makes possible to ask.

What I have observed across these cases is that the post-session plasticity window is where the identity-layer work either takes or does not take. If the client uses the two to four weeks after the session to actively rebuild day-to-day life around different organizing principles, even small ones such as how they decide when to eat, what they notice about hunger, how they relate to their reflection, the new pattern has time to consolidate. If they return immediately to the same routines, the plasticity window consolidates the old pattern. The session is not the intervention. The two to four weeks afterward are the intervention, and the session is what makes that window workable.

"Eating disorders do not feel like a disorder to the person living them. They feel like the one stable thing in a life that otherwise will not hold still. The session does not tell you that you can eat. It shows you that you can be a self without holding the body in place."

Why Is Eating Disorder Integration So Different From Standard Integration?

Integration for eating disorder presentations looks different from standard psychedelic integration because the disorder runs on a different substrate. Depression integration centers on meaning and reconnection, while eating disorder integration centers on rebuilding the self-concept on grounds other than body control, ideally during the two to four week window of elevated post-session plasticity. Using the wrong template wastes the window or, worse, creates new ritualized patterns. For the underlying framework, see psychedelic integration therapy.

The Self-Concept Update

The session, when it works for an eating disorder client, produces a window in which the self can be experienced without the body-control organizing scaffolding. This is not a cognitive insight. It is a felt experience of identity being separable from intake regulation. Integration work is essentially the deliberate, repeated practice of letting daily life happen without immediately re-organizing it around food and body. Each repetition during the plasticity window updates the underlying self-concept more efficiently than equivalent practice without the window.

Why Standard Integration Misfires on Eating Disorders

Standard psychedelic integration practices include journaling, meaning-making, and somatic awareness work. These are useful for many presentations. For eating disorder integration they can become problematic. Journaling can turn into a verification ritual: have I documented enough, have I processed enough. Somatic awareness can intensify body fixation rather than loosen it, especially if the practice is framed around noticing body sensations without first updating the relationship to the body itself. Skilled integration for eating disorders recognizes these risks and structures the work around identity reconstruction rather than around body monitoring. For somatic-specific cautions, see somatic psychedelic integration.

Preparation: The Medical Layer Comes First

Eating disorder preparation differs from other preparation work because the medical layer is non-negotiable and comes first. Underweight patients need medical clearance. Electrolyte status, cardiac function, and nutritional restoration need to be addressed before any psychedelic dosing is appropriate. The Peck 2023 protocol used medically stable participants. Outside of clinical research, the same standard applies and is more, not less, important. The session is not a shortcut around medical care. It is a possible adjunct after medical care has stabilized the body enough that an altered state is not itself a medical risk. For the broader preparation framework, see psilocybin therapy preparation.

Frequently Asked Questions