The Relationship Between Participant Pretreatment Clinical Presentation and the Quality of Psilocybin Experience: A Retrospective Analysis

Earlier research indicates that the antidepressant effects of psilocybin are linked to the subjective, dose-dependent psychedelic experience, with dimensions such as oceanic boundlessness (a mystical-type experience) predictive of symptom reduction. Clinical trials typically standardise preparation and the treatment environment because 'set and setting' are believed to shape safety and therapeutic benefit. In the phase 2 COMP 001 trial, a single 25 mg dose of COMP360 psilocybin produced greater improvements in depressive symptoms than 1 mg, and dose-dependent subjective effects were associated with outcomes. However, considerable within-dose variation in the intensity and quality of psychedelic experience was observed and is not well understood. N. and colleagues set out to examine to what extent individual pretreatment clinical characteristics—across cognitive-affective, somatic, and functional impairment domains—contribute to acute subjective effects of COMP360 psilocybin, relative to the pharmacological dose. The objective was to quantify the incremental predictive value of these pretreatment variables for established measures of acute psychedelic experience (the 5D-ASC and the Emotional Breakthrough Inventory), using data from participants with treatment-resistant depression (TRD) who received 25 mg, 10 mg, or 1 mg in the COMP 001 trial.

This retrospective analysis used data from 233 participants with treatment-resistant major depressive disorder who had been randomised in COMP 001 to receive a single administration of COMP360 psilocybin at 25 mg (n = 79), 10 mg (n = 75), or 1 mg (n = 79). TRD was defined by DSM-5 criteria for moderate or worse major depressive episodes (HAMD-17 ≥ 18) and failure to respond to 2–4 adequate pharmacological treatments. Participants underwent a 3–6-week run-in during which prohibited CNS medications were tapered and discontinued at least 2 weeks before dosing. Administration occurred in a standardised supportive environment with trained therapists providing nondirective preparation, monitoring during the session, and integration support. Pretreatment clinical characteristics (independent variables) were measured one day before dosing and spanned cognitive-affective measures (Digit Symbol Substitution Test (DSST); two Adult ADHD Self-Report Scale items; HAMD-17 psychomotor retardation item; GAD-7; HAMD-17 anxiety item; PANAS positive and negative affect; MSI-BPD), somatic measures (HAMD-17 somatic and somatic–gastrointestinal items), and functional impairment (Sheehan Disability Scale and HAMD-17 work and activities item). Subjective psychedelic experience (dependent variables) was assessed by the Five Dimensions of Altered States of Consciousness questionnaire (5D-ASC) administered immediately after dosing—yielding oceanic boundlessness (OC), anxious ego dissolution (AED), visual restructuralization (VR), auditory alterations (AA), and reduction of vigilance (RV)—and by the Emotional Breakthrough Inventory (EBI) administered one day later. The authors used hierarchical multiple regression to test two models per outcome: model 1 included dose only; model 2 added all pretreatment predictors. Model comparisons used ANOVA F tests. Model assumptions (linearity, independence of errors, homoscedasticity, normality) were evaluated; Box–Cox transformations were applied where nonlinearity or heteroscedasticity were observed. Multicollinearity was assessed via variance inflation factors (VIF), with a VIF > 5 flagged as problematic. Minor missing data were left unimputed; analyses used available-case data, causing slight sample-size variation across models. Because six separate outcome models were tested, the authors set a corrected significance threshold of P = 0.0083 for the model-comparison tests (Bonferroni correction). Individual predictor significance was assessed at P = 0.05. Analyses were performed in R (v4.4.2) using standard packages for regression, Box–Cox transformation, and graphics.

Sample characteristics: The analysable sample comprised 233 participants (mean age 39.8 years; 52% female) randomised equally across 25 mg, 10 mg, and 1 mg dose groups; pretreatment demographics and clinical characteristics were reported as comparable across dose groups. Descriptive and correlation analyses were conducted (figures/tables referenced in the text). For each 5D-ASC dimension and for EBI, the authors report the variance explained by dose alone (model 1) and by dose plus pretreatment predictors (model 2), together with which individual pretreatment variables reached conventional significance within model 2. Oceanic boundlessness (OC): Model 1 (dose only) explained R2 = 0.31 (F1,200 = 89.34, P < 0.001). Adding pretreatment predictors (model 2) increased R2 to 0.42 (ΔR2 = 0.11; F14,186 = 2.56, P < 0.01), but this improvement did not survive the multiple-comparison threshold (P = 0.0083). In model 2, significant predictors included COMP360 dose (β = 1.45, P < 0.001), PANAS Positive Affect (β = 0.08, P < 0.001), HAMD-17 somatic–GI item (β = 0.55, P < 0.01), GAD-7 total score (β = −0.07, P < 0.05), and DSST score (β = 0.03, P < 0.05). Reduction of vigilance (RV): Model 1 R2 = 0.13 (F1,200 = 30.41, P < 0.001); model 2 R2 = 0.22 (ΔR2 = 0.09), but the model-improvement test was not significant (ΔR2 F14,186 = 1.55, P = 0.096). Within model 2, COMP360 dose (β = 2.31, P < 0.001), MSI-BPD score (β = 0.54, P < 0.05), and DSST (β = 0.07, P < 0.05) were significant predictors. Anxious ego dissolution (AED): Model 1 R2 = 0.21 (F1,200 = 52.95, P < 0.001); model 2 R2 = 0.30 (ΔR2 = 0.09; F14,186 = 1.79, P = 0.043), but again the improvement did not meet the corrected threshold. Significant predictors in model 2 were COMP360 dose (β = 0.49, P < 0.001), PANAS Positive Affect (β = 0.03, P < 0.01), and GAD-7 total score (β = −0.04, P < 0.05). Visual restructuralization (VR): Model 1 R2 = 0.34 (F1,200 = 101.80, P < 0.001); model 2 R2 = 0.42 (ΔR2 = 0.08; F14,186 = 1.78, P = 0.045), not surviving the multiple-comparison correction. Within model 2, significant predictors were COMP360 dose (β = 1.88, P < 0.001), PANAS Positive Affect (β = 0.08, P < 0.01), DSST (β = 0.04, P < 0.05), and HAMD-17 somatic–GI (β = 0.63, P < 0.01). Auditory alterations (AA): Model 1 R2 = 0.13 (F1,200 = 29.48, P < 0.001); model 2 R2 = 0.22 (ΔR2 = 0.09; F14,186 = 1.46, P = 0.12), not significant. In model 2, COMP360 dose (β = 0.16, P < 0.001) and PANAS Positive Affect (β = 0.01, P < 0.05) were significant predictors. Emotional breakthrough (EBI): Model 1 R2 = 0.15 (F1,200 = 35.29, P < 0.001); model 2 R2 = 0.22 (ΔR2 = 0.06; F14,186 = 1.15, P = 0.32), not a significant model improvement. Significant predictors in model 2 were COMP360 dose (β = 1.83, P < 0.001), PANAS Positive Affect (β = 0.13, P < 0.05), and HAMD-17 somatic–GI (β = 0.83, P < 0.05). Across outcomes, COMP360 dose was the most consistent and strongest predictor of subjective psychedelic experience, with effects typically larger than those of individual pretreatment variables. Although adding pretreatment characteristics produced statistically significant increases in explained variance for some outcomes before correction, none of the model-improvement tests met the prespecified multiple-comparison threshold. The authors report minor missing data that were not imputed and note that model diagnostics and Box–Cox transformations were used where necessary.

N. and colleagues interpret their findings as indicating that the acute subjective effects of psilocybin are driven predominantly by dose, rather than by the pretreatment clinical profile across cognitive-affective, somatic, and functional domains. While the addition of pretreatment characteristics increased explained variance for three 5D-ASC dimensions (oceanic boundlessness, visual restructuralization, and anxious ego dissolution), the effect sizes were small and the model-improvement tests did not survive correction for multiple comparisons. The authors therefore contend that pre-existing patient characteristics are not major determinants of the capacity to experience acute psilocybin effects. The authors situate their findings within earlier work showing that personality and affective traits can modulate psychedelic experiences in healthy volunteers, noting mixed results in the literature about the role of positive affect. In this clinical sample, higher trait positive affect was associated with greater oceanic boundlessness, anxious ego dissolution, visual restructuralization, and emotional breakthrough, suggesting a complex role for positive affect in shaping both positive and challenging aspects of the experience. Lower pretreatment anxiety was associated with more expansive mystical-type experiences and emotional breakthrough, in line with prior studies linking anxiety to more challenging experiences. The discussion also highlights associations between higher performance on the DSST (a measure reflecting processing speed and associative learning) and increased oceanic boundlessness and reduction of vigilance, and between somatic (GI) symptoms and greater emotional breakthrough, oceanic boundlessness, and visual restructuralization. The authors propose that heightened interoceptive sensitivity in those with somatic symptoms may facilitate engagement with bodily and affective material during psilocybin administration, potentially promoting emotional processing. Key limitations acknowledged by the authors include the constrained set of pretreatment variables available in the dataset (limiting generalisability to other demographic, personality, cognitive, and clinical factors), the conceptual and empirical interdependence of predictors, the inability of the retrospective, observational regression approach to establish causality, limited generalisability beyond this trial population, and the reliance on retrospective self-report measures that may be affected by memory or expectancy. They suggest future research should examine a broader range of predictors, employ longitudinal or experimental designs to test causality, include more diverse populations, and use real-time or objective measures of experience. The authors conclude that dose remains the dominant factor in determining acute psilocybin experience, challenging the notion that pretreatment clinical profiles are determinative of a patient's capacity to engage with or benefit from psilocybin treatment.