Psilocybin vs. Escitalopram: Does the Advantage Vanish?

Psilocybin therapy appears to help depression. The harder question is whether it helps more than a standard antidepressant such as escitalopram. A 2026 meta-analysis suggests the dramatic advantage often advertised for psychedelic-assisted therapy may shrink to almost nothing once researchers compare treatments under similarly unblinded conditions.

That does not make psilocybin a placebo. It changes what the evidence can support: a real therapeutic signal, but no demonstrated population-level superiority over conventional antidepressants on depression symptom scales.

The problem: participants can usually tell they took a psychedelic

Blinding is central to drug trials. If participants do not know whether they received the active treatment, expectations should affect both groups more evenly. Psychedelic therapy makes that unusually difficult. A profound alteration in perception is not a subtle side effect; it can function as an announcement from the drug itself.

This creates an uneven comparison when psychedelic-assisted therapy is tested against an inactive placebo. Participants who recognize that they received psilocybin may expect improvement. Those who recognize they received the control may become discouraged, reducing placebo response. An analysis by Hieronymus et al. found that control-group MADRS outcomes were systematically lower in psilocybin trials than in SSRI or esketamine trials, a pattern consistent with suppressed control-arm response. The authors also raised the possibility that unblinding could affect supervising staff, not only participants.

That is a design problem, not proof that every reported psychedelic benefit is illusory. But it means a large gap between psilocybin and blinded placebo cannot cleanly answer how psilocybin compares with an antidepressant taken openly in ordinary clinical care.

What happened when the comparison was leveled?

Williams et al. addressed that question by comparing psychedelic-assisted therapy (PAT) trials with open-label trials of traditional antidepressants (TADs), where neither treatment has a credible blind. Across 8 PAT trials involving 249 patients and 16 open-label TAD trials involving 7,921 patients, PAT had a 0.3-point advantage on the HDRS-17 depression scale. The 95% confidence interval ranged from −1.39 to 1.98 points, and the result was not statistically significant (p=.73).

The estimate does not establish that the treatments are equivalent. With only 249 patients in the PAT group, the confidence interval still allows a clinically meaningful difference in either direction. It does establish that this evidence set did not detect a meaningful advantage for PAT once expectancy conditions were made more comparable.

The same review compared blinded and open-label conditions within each treatment category. For traditional antidepressants, open-label administration produced outcomes about 1.3 HDRS-17 points better than blinded administration (95% CI, 0.07 to 2.51; p=.04), using 16 open-label trials with 7,921 participants and 144 blinded trials with 31,792 participants. For PAT, the corresponding difference was not statistically significant (d=0.67; 95% CI, −3.08 to 1.73; p=.58). Williams et al. interpreted this pattern as evidence that nominally blinded psychedelic trials effectively operate as open-label trials.

That interpretation fits the broader concern, although it has an important limitation: most PAT trials did not prospectively measure blinding integrity by asking participants to guess their assignment and reporting correct-guess rates. In a secondary analysis of the psilocybin-versus-escitalopram trial, Szigeti et al. found that pre-treatment expectancy was associated with response in the psilocybin arm. The study did not center on formal correct-guess rates, so it supports an expectancy concern without quantifying how often unblinding occurred.

The direct psilocybin–escitalopram trial still matters

The most recognizable head-to-head study remains the 2021 randomized trial by Carhart-Harris et al., which enrolled people with depression rather than a specifically treatment-resistant population. Psilocybin and escitalopram produced comparable results on the primary endpoint, the self-reported QIDS-SR. Several secondary measures favored psilocybin, including measures related to emotional processing and well-being.

Primary and secondary outcomes answer different questions. The trial supports the view that psilocybin therapy can perform in the same neighborhood as escitalopram for depression symptoms. It does not provide clean evidence that psilocybin is broadly superior, especially because the perceptual effects of psilocybin make treatment guessing a live concern.

Different biology, uncertain clinical consequences

Comparable symptom scores do not mean identical treatment effects. In a secondary neuroimaging analysis of that trial, Harding et al. analyzed usable fMRI data from 41 patients: 22 assigned to psilocybin therapy and 19 to escitalopram. The groups showed dissociable fMRI activation patterns while processing musical surprises after the six-week intervention. This is evidence of differing neural responses, not proof that one treatment repairs depression through a superior mechanism or produces better long-term patient outcomes.

Weiss et al. also reported different self-reported personality trajectories in the same six-week trial. Psilocybin therapy was associated at week 6 with lower neuroticism (B=−0.63), introversion (B=−0.38), disagreeableness (B=−0.47), and impulsivity (B=−0.40), plus higher absorption (B=0.32), conscientiousness (B=0.30), and openness (B=0.23). Lower neuroticism (B=−0.47) and disagreeableness (B=−0.41) remained at month 6. These results are interesting, but personality questionnaires invite expectancy and demand effects. A participant who knows they had psilocybin may sincerely report feeling more open; that does not isolate the drug effect from the meaning and expectation attached to the experience.

Why larger meta-analyses look more favorable

Several major reviews still find large antidepressant effects for psychedelic treatments against blinded placebo or non-psychoactive comparators. A 2024 BMJ network meta-analysis by Hsu et al. reported standardized mean differences of 1.37 to 3.12 across psilocybin, LSD, MDMA, ayahuasca, and escitalopram comparisons. A separate 2024 BMJ review by Metaxa et al. also reached broadly positive conclusions about psilocybin for depressive symptoms, while a 2025 living review by Højlund et al. found broadly supportive evidence across serotonergic psychedelics and MDMA.

These findings do not directly contradict Williams et al. They use different contrasts and, in Hsu et al.'s case, standardized mean differences rather than raw HDRS-17 points. The numbers cannot be converted into a simple "inflation ratio." The reviews show large effects versus blinded placebo; Williams et al. found no statistically significant advantage versus open-label traditional antidepressants. Hsu et al. also noted that differing placebo effects between active and control groups could bias results.

Where psychedelic therapy may still have the strongest case

The case for PAT may be more compelling in populations with fewer satisfying options than in general major depression. A 2024 Cochrane review examined psychedelic-assisted therapy for anxiety, depression, and existential distress in people with life-threatening illness. A 2025 phase 2b randomized trial by Ross et al. tested 25 mg psilocybin against 100 mg niacin, with psychotherapy, in adults with life-threatening illness and clinically significant depression and/or anxiety. These studies support continued research in serious, targeted settings; they do not erase the need to account for functional unblinding.

The same problem extends beyond psilocybin. Swieczkowski et al. reviewed 40 registered ketamine, esketamine, and arketamine protocols and found inconsistent approaches to blinding, expectancy bias, and regulatory alignment. Consciousness-altering psychiatric treatments need trial designs that acknowledge the obvious: a control cannot do its job if everyone knows who received the memorable intervention.

Verdict

Psilocybin-assisted therapy has a credible antidepressant signal. The evidence does not yet show that it beats open-label SSRIs or other traditional antidepressants for general depression. Its higher cost, specialist staffing, preparation, supervised dosing, and integration sessions therefore need a reason beyond headline effect sizes.

For people with treatment-resistant depression or life-threatening illness, the balance may differ. For someone who has not tried standard treatment, the current evidence does not justify choosing psilocybin solely because it is assumed to work better. Larger comparative trials should measure treatment guessing prospectively, pre-register how they will handle expectancy, and test whether psilocybin's distinct neural and psychological effects lead to outcomes patients can actually feel months later.