Examining mystical experiences as a predictor of psilocybin-assisted psychotherapy for treatment-resistant depression

Psilocybin-assisted psychotherapy (PAP) has shown promise for treatment-resistant depression and other psychiatric conditions, but the psychological and biological mechanisms that produce therapeutic benefit remain uncertain. Prior studies have identified acute subjective phenomena during psychedelic sessions—such as emotional breakthrough, ego dissolution, and mystical-type experiences—as candidate predictors or mediators of clinical improvement. However, evidence is mixed: some human studies find mystical-type experiences associated with enduring positive outcomes, while other trials have not observed a predictive relationship. Measurement and conceptual questions also persist, including whether mystical experience should be treated as a continuous dimension or dichotomised into a ‘complete mystical experience’. Brudner and colleagues sought to examine whether mystical experiences, as measured by the Mystical Experience Questionnaire-30 (MEQ-30), predict antidepressant response in a clinical sample with treatment-resistant depression (including major depressive disorder and Bipolar II disorder). Using data from a previously reported waitlist-controlled feasibility trial, the investigators tested whether MEQ-30 scores collected after dosing predicted clinician-rated depressive symptoms two weeks post-dose. The study additionally described frequencies of dichotomised complete mystical experiences across up to three fixed 25 mg psilocybin dosing sessions, since the prevalence of such experiences at this dose in a treatment-resistant clinical sample had not been clearly reported.

The study recruited participants with DSM-5 major depressive disorder or Bipolar II disorder who met criteria for treatment resistance (failure to respond to at least two guideline-concordant pharmacological trials for the current episode). Key exclusion criteria included age outside 18–75, recent moderate-to-severe alcohol or substance-use disorder, psychedelic use within the past 12 months, neuromodulation during the trial, and uncontrolled cardiovascular or major medical illness. Participants were asked to refrain from augmentation medications for the study duration; tapering decisions were made between participants, the study doctor, and the referring physician, and the study doctor could exclude individuals for whom tapering was unsafe. Design features included randomisation to immediate treatment or a two-week waitlist control arm for the parent feasibility trial. Synthetic psilocybin (25 mg) was administered in powder form dissolved in water. Participants received one, two, or three PAP dosing sessions depending on clinical criteria for retreatment (clinical benefit, tolerability, and signs of relapse). Each dosing session was embedded in a psychotherapeutic protocol comprising one preparatory session (1–2 h), a supportive dosing session (6–8 h) with a therapist dyad and a music playlist option, and two integration sessions (1–2 h each), totalling roughly 4.5 h of psychotherapy per dose. Therapists completed an intensive PAP training programme specific to the protocol. Primary clinical measurement used the clinician-rated Montgomery–Åsberg Depression Rating Scale (MADRS), administered at baseline and at two weeks after each dosing session by independent raters. The MEQ-30, a validated 30-item self-report instrument assessing mystical-type phenomena, was completed on the dosing day after acute effects resolved; total scores and the four subscales (Mystical, Positive Mood, Ineffability, Transcendence of Time and Space) were computed. A dichotomous ‘complete mystical experience’ variable was derived when participants scored at least 60% of possible points on each subscale, following precedent in prior studies. For statistical analysis, hierarchical (block) regressions were run separately for each dose to test whether MEQ-30 total scores incrementally predicted MADRS scores at two weeks beyond pre-dose MADRS. Pre-dose MADRS was entered in block 1 and MEQ-30 total scores for the contemporaneous dosing session were entered in block 2; standardised beta coefficients and p-values are reported. A repeated-measures ANOVA tested whether MEQ-30 total scores differed across doses. Analyses were conducted in Jamovi. The extracted text indicates that detailed procedural and training descriptions and some primary outcome results appear in the parent trial report.

Recruitment occurred from 1 November 2021 to 1 February 2023, with all visits completed by 26 July 2023. Thirty-one participants were enrolled, one of whom withdrew before receiving treatment due to difficulties tapering antidepressant medications; 30 participants thus received at least one psilocybin dose. Full 2-week post-dose one assessments were completed by 28 participants. Seventeen participants received and completed assessments after a second dose, and five participants received and completed assessments after a third dose. For the first dosing session, baseline (pre-dose) MADRS scores significantly predicted MADRS at two weeks, F(1, 26) = 8.03, p = 0.009. Adding Dose 1 MEQ-30 total scores in the second block produced a significant model, F(2, 25) = 7.52, p = 0.003, with block 2 explaining a significant additional proportion of variance (ΔR² = 0.14, F(1, 25) = 5.60, p = 0.026). The standardised beta for MEQ-30 in this model was −0.387, p = 0.026, indicating that a one standard deviation increase in MEQ-30 total score was associated with a 0.387 standard deviation decrease in MADRS score at two weeks post-dose 1. These dose 1 results are presented graphically in the paper. By contrast, hierarchical models for dose 2 and dose 3 were not significant. The dose 2 model yielded F(2, 14) = 1.89, p = 0.188 and block 2 did not explain additional variance (ΔR² = 0.04, F(1, 14) = 0.77, p = 0.395). The dose 3 model was similarly non-significant, F(2, 2) = 1.11, p = 0.475, and block 2 did not add explanatory variance (ΔR² = 0.40, F(1, 2) = 1.70, p = 0.322). A repeated-measures ANOVA found no statistically significant difference in MEQ-30 total scores across doses, F(2, 8) = 3.36, p = 0.087. Regarding the dichotomous ‘complete mystical experience’ outcome, the extracted text reports a dose 1 complete mystical experience rate of 28.6% (described in the Discussion). When considering only those who received at least two doses (n = 17), 8 participants (47.1%) had a complete mystical experience at dose 2, while 6 (35.3%) had one at dose 1; all six participants who had a complete mystical experience at dose 1 also had one at dose 2. Among the five participants who received three doses, 3 (60%) had a complete mystical experience at dose 2 and dose 3, while 1 (20%) had one at dose 1. The authors note that frequencies are summarised in a table in the paper.

Brudner and colleagues interpret their findings as providing preliminary support that mystical-type experiences, as measured by the MEQ-30, are associated with greater short-term antidepressant response after an initial 25 mg psilocybin session in a treatment-resistant clinical sample that included people with MDD and Bipolar II disorder. The significant association was observed for dose 1: higher MEQ-30 total scores predicted lower MADRS scores two weeks later after accounting for pre-dose depression severity. The absence of a significant predictive effect at the second and third doses is discussed in multiple, non-mutually exclusive ways. The investigators caution that small sample sizes for later doses (N = 17 for dose 2; N = 5 for dose 3) substantially reduce statistical power and may have obscured existing effects. They also raise the possibility of ceiling effects—participants who experienced strong benefits or intense mystical experiences at dose 1 may have had limited further variability in outcomes— and consider that mystical experiences might matter most at an initial, novel exposure. The authors acknowledge that a true null effect across later doses is also possible and that larger samples are required to resolve these alternatives. Several limitations are acknowledged. First, the exploratory analyses were constrained by the relatively small sample size and by attrition across subsequent doses. Second, the MEQ-30 captures only one facet of the acute subjective experience; other constructs such as emotional breakthrough, ego dissolution, music impact, visual imagery and culturally mediated expectations may also influence outcomes and were not modelled here. Third, no correction for multiple comparisons was applied across the three dose-level tests, increasing the possibility of a Type I error for the dose 1 finding. Fourth, the correlational design precludes causal inference; the authors recommend formal mediation analyses in appropriately powered trials to test causality. Fifth, aspects of the specific PAP protocol (for example, the particular music, presence of a therapist dyad, and therapy room setup) may influence both mystical experiences and outcomes, limiting generalisability. The authors also discuss broader conceptual and cultural considerations. They note challenges in defining and measuring mysticism, and urge that psychedelic research engage with the historical, religious and indigenous literatures on mystical practice to improve conceptualisation and measurement. Despite methodological constraints, the reported frequencies of complete mystical experiences at 25 mg provide hypothesis-generating data and motivate larger, better-controlled studies to clarify how mystical-type phenomena relate to therapeutic benefit across multi-dose PAP protocols.