The effects of psilocybin therapy versus escitalopram on cognitive bias: A secondary analysis of a randomized controlled trial

Major Depressive Disorder (MDD) is a leading contributor to global disease burden and is closely linked to maladaptive cognitive schemas, such as pessimistic beliefs about the self and future, which contribute to the maintenance of depressive symptoms. Cognitive behavioural therapy (CBT) and selective serotonin reuptake inhibitors (SSRIs) are first-line treatments and target these dysfunctional attitudes. Previous research has shown that psilocybin, when combined with psychological support, can produce rapid and sustained reductions in depressive symptoms, and it has been hypothesised that 5-HT2A receptor agonism during the psychedelic state may enhance neural and psychological plasticity to promote revision of maladaptive cognitive biases. Henry and colleagues conducted a secondary analysis of a randomised controlled trial to compare the effects of two high-dose psilocybin sessions plus psychological support versus a six-week daily course of escitalopram on cognitive biases relevant to depression. The investigators focused on pessimistic versus optimistic future forecasting and dysfunctional attitudes, assessed at baseline and at a primary six-week endpoint, and hypothesised a greater shift away from pessimistic forecasting after psilocybin therapy. This study is presented as the first direct comparison of psilocybin therapy and an SSRI on behavioural indices of dysfunctional attitudes in patients with MDD.

The study was a two-arm, randomised controlled trial that enrolled fifty-nine patients with MDD who were matched for age, gender and education and randomised to psilocybin (n = 30) or escitalopram (n = 29). Participants in the psilocybin arm received two 25 mg doses of psilocybin and a daily inert oral placebo, whereas those in the escitalopram arm received two 1 mg doses of psilocybin alongside daily oral escitalopram (starting 10 mg/day after the first dosing session and escalating to 20 mg/day after the second). Both arms received an equivalent, standardised psychological support protocol that included preparatory and integrative therapy sessions, two therapeutic guides per participant and music during dosing. Baseline assessments occurred 7–10 days before the first psilocybin administration, with six additional study visits over six weeks and a final follow-up three weeks after the second psilocybin dose. For this secondary analysis the primary outcomes were changes from baseline to six weeks in cognitive biases as measured by three validated instruments: the 24-item Dysfunctional Attitudes Scale (DAS-24) with subscales for achievement, dependency and self-control; the Revised Life Orientation Test (LOT-R) assessing self-reported optimism versus pessimism; and the Prediction of Future Life Events task (POFLE), a behavioural forecasting task with 40 events split into desirable and undesirable items. POFLE had two balanced versions administered in counterbalanced order and participants were contacted 30 days after completing each version to record which forecasted events actually occurred. Secondary measures included depressive symptoms (Beck Depression Inventory 1A, BDI-1A) and psychological flourishing (Flourishing Scale, FS). Follow-up cognitive-bias assessments were not conducted beyond six weeks. Statistical analyses used t-tests, linear mixed effects models (LME) and two-way ANOVAs as appropriate, with Welch's correction when comparing to normative samples and Bonferroni correction for multiple post hoc comparisons. Correlations were used to explore relationships between cognitive-bias changes and clinical outcomes. Results are reported with 95% confidence intervals and Cohen's d effect sizes, and statistical significance was set at p ≤ 0.050. The extracted text does not clearly report some procedural specifics (for example full details of randomisation or masking success), though the investigators note challenges in blinding due to distinctive drug effects.

Fifty-nine participants were included in the analysis (psilocybin n = 30; escitalopram n = 29). Primary cognitive-bias outcomes at six weeks showed differential effects by treatment arm across measures. LOT-R (self-reported optimism): A two-way mixed ANOVA revealed a significant time × condition interaction [F(1,57) = 15.89, p = 0.0002]. Post-hoc testing indicated a large and significant increase in LOT-R scores in the psilocybin group at six weeks relative to baseline (reported effect d = 1.1). There were no between-group differences at baseline, but the psilocybin group had significantly greater optimism scores than the escitalopram group at six weeks. POFLE (behavioural forecasting): LME modelling found a significant main effect of time [F(1,99) = 20.86, p < 0.0001], but no significant main effect of condition [F(1,99) = 0.86, p = 0.355] and no overall interaction [F(1,99) = 2.78, p = 0.098]. Subsequent analyses splitting desirable and undesirable events indicated a significant time × condition interaction for desirable events [F(1,44) = 7.93, p = 0.007], although there was a baseline between-group difference (psilocybin group more pessimistic for desirable events). After adjusting for baseline via ANCOVA, change in forecasting for desirable events did not differ between groups. Within-group effects showed that psilocybin-treated patients became more optimistic about desirable events at six weeks (reported within-group M diff ≈ 0.16, p = 0.0002), whereas escitalopram-treated patients showed reduced pessimism for undesirable events (within-group M diff ≈ 0.07, p = 0.018). Overall, the psilocybin group's improvement in POFLE total scores appeared driven mainly by changes in forecasting of desirable events. DAS-24 (dysfunctional attitudes): A significant time × condition interaction was observed [F(1,57) = 8.285, p = 0.006]. There were no baseline differences, but at six weeks the psilocybin group had significantly lower (improved) DAS-24 total scores than the escitalopram group (M diff = 21.14, SE diff = 5.63, p = 0.0004, d = 1.0). Within-group analyses showed significant decreases in DAS-24 total scores for the psilocybin arm. Subscale analyses indicated that psilocybin produced significant improvements across all three subscales: achievement (M diff = 10.37, p < 0.0001, d = 1.0), dependency (M diff = 7.97, p < 0.0001, d = 0.9) and self-control (M diff = 6.40, p = 0.0006, d = 0.8). Escitalopram produced a significant improvement only on the achievement subscale (M diff = 4.10, p = 0.005, d = 0.6), with no significant changes on dependency or self-control. Depressive symptoms and well-being: BDI-1A scores showed a significant time × condition interaction [F(1,57) = 6.821, p = 0.011]. Both groups improved versus baseline: psilocybin M diff = 18.40 (p < 0.0001, d = 1.5) and escitalopram M diff = 10.83 (p < 0.0001, d = 1.1). Between-group comparison of change scores favoured psilocybin (M diff = -7.57, SE diff = 2.899, t(57) = 2.612, p = 0.012, d = 0.7). Flourishing Scale scores also showed a time × condition interaction [F(1,57) = 4.454, p = 0.039]; both groups increased versus baseline, but the psilocybin group improved more at six weeks (between-group M diff reported ≈ 5.35, p = 0.039). The extracted text includes an implausible effect-size value for flourishing (d = 10.0), and the presentation of some SE and CI details appears inconsistent in places; these anomalies are noted because the extraction is imperfect. Correlations: Changes in dysfunctional attitudes correlated with decreases in depressive symptoms in both arms (psilocybin: rs = 0.519, p = 0.003; escitalopram: rs = 0.579, p = 0.001). For psilocybin, increases in LOT-R scores correlated strongly with reductions in BDI-1A (rs = -0.758, p < 0.0001), while the correlation between POFLE change and BDI-1A did not reach significance (rs = -0.349, p = 0.074). Improvements in flourishing correlated with all three cognitive-bias outcomes following psilocybin (LOT-R rs = 0.753, p < 0.0001; POFLE rs = 0.448, p = 0.019; DAS-24 rs = -0.599, p < 0.0001). No comparable relationships were observed for optimism/pessimism measures after escitalopram.

Henry and colleagues interpret their findings as evidence that two high-dose psilocybin sessions with standardised psychological support produced broader and larger improvements in cognitive biases associated with depression than a six-week escitalopram regimen. They highlight robust increases in self-reported optimism (LOT-R) and decreases in dysfunctional attitudes across achievement, dependency and self-control domains after psilocybin, whereas escitalopram produced improvement primarily on the achievement domain and reduced pessimism about undesirable events on the behavioural forecasting task. The investigators situate these differential effects in relation to prior work showing psychedelic-associated increases in optimism and improved accuracy of future forecasting. They propose mechanistic links to serotonergic neuroscience, emphasising 5-HT2A receptor agonism as a putative promoter of neural and psychological plasticity, cognitive flexibility and reduced ‘‘precision-weighting’’ of maladaptive predictive models (a predictive processing account). This framework is used to explain how psychedelics might destabilise entrenched negative beliefs and thereby enable adaptive belief updating when combined with psychotherapy. The authors also contrast putative valence-specific actions of treatments, suggesting escitalopram may blunt affective responsivity and reduce pessimism for negative events while not enhancing positive expectancies, whereas psilocybin may preferentially increase optimism for desirable outcomes. The discussion acknowledges important limitations reported by the investigators. Effective blinding was likely compromised due to distinctive subjective effects of psilocybin and SSRI side effects, creating potential expectancy biases. The six-week time frame may have been insufficient for escitalopram to reach full therapeutic effect. The sample was predominantly Caucasian, highly educated males with moderate baseline depression, limiting generalisability. The authors also note that multiple cognitive metrics were analysed and that correction for multiple comparisons was applied within but not across measures, a choice they defend as acceptable for exploratory hypothesis generation. Finally, they underscore the critical role of psychotherapeutic context in psychedelic trials, note heterogeneity in psychological-support models across studies, and call for research to define essential components of therapy and to examine ethical and safety considerations associated with enhancing plasticity.

In this secondary analysis of a randomised trial, psilocybin therapy produced larger and more comprehensive improvements in measures of negative cognitive bias in patients with MDD at six weeks post-treatment than a six-week course of escitalopram, according to the authors. The findings are presented as supportive of further investigation of psilocybin therapy's mechanisms, safety and efficacy at larger scale, while recognising the need for careful consideration of psychological support, trial design and participant diversity.