How does psilocybin therapy work? An exploration of experiential avoidance as a putative mechanism of change

A growing literature indicates that psilocybin therapy—administration of psilocybin alongside psychological support—can improve a range of mental health outcomes including depressive and anxious symptoms, suicidal ideation, and subjective well-being. Prior qualitative and survey-based work has highlighted reductions in experiential avoidance (defined as unwillingness to remain in contact with distress and rigid efforts to minimise it) and increases in connectedness as recurrent themes reported by participants after psychedelic experiences. However, quantitative, controlled tests of experiential avoidance as a mechanism of change in psilocybin therapy have been lacking, and many prior studies relied on non-clinical convenience samples and measures with contested validity. Zeifman and colleagues set out to test whether reductions in experiential avoidance mediate improvements in mental health following psilocybin therapy, and whether this mediation differs from that seen with a standard antidepressant (escitalopram). The study also explored whether increases in connectedness might act as a serial mediator (i.e., reductions in experiential avoidance leading to greater connectedness, which then leads to better outcomes) and whether specific acute features of the psilocybin experience (ego dissolution, mystical-type experiences, emotional breakthrough, psychological insight) predict reductions in experiential avoidance.

The investigators used data from a double-blind randomised controlled trial (clinicaltrials.gov NCT03429075) in which 59 participants with Major Depressive Disorder were randomised to one of two arms. The psilocybin therapy condition received two high doses of psilocybin (25 mg each, administered three weeks apart) plus six weeks of daily inactive placebo capsules. The escitalopram condition received two very low doses of psilocybin (1 mg each) plus six weeks of daily escitalopram (10 mg for three weeks, then 20 mg for three weeks). All participants received a psychotherapeutic package that included preparation sessions, therapist-supported dosing sessions (two therapists present, eyeshade, music, encouragement to ‘‘turn inward’’), and integration sessions. Recruitment and eligibility involved clinician screening (video and in-person assessments), confirmation of MDD diagnosis, and medical checks (ECG, urine drug screen, blood and pregnancy tests). Primary outcome assessments occurred at baseline (pre-treatment) and at the 6-week primary endpoint (three weeks after the second dosing session). Acute psychedelic experience measures were collected one day after each dosing session, and perceived psychological insight was assessed at post-treatment. Key self-report and clinician-assessed measures included the Brief Experiential Avoidance Questionnaire (BEAQ), the Watts Connectedness Scale (WCS), the Warwick–Edinburgh Mental Wellbeing Scale (WEMWBS), the Montgomery–Åsberg Depression Rating Scale (MADRS), the QIDS-SR-16, the Suicidal Ideation Attributes Scale (SIDAS), and the STAI-Trait for trait anxiety. Acute-session measures comprised the Ego Dissolution Inventory (EDI), the MEQ-30 for mystical-type experiences, the Emotional Breakthrough Inventory (EBI), and the Psychological Insight Questionnaire (PIQ). Statistical analysis focused on mediation using structural equation modelling (SEM). Separate within-participant mediation models were run for each treatment condition and for each outcome (well-being, clinician-assessed depression, self-reported depression, suicidal ideation, trait anxiety). Change in experiential avoidance (BEAQ) was the mediator; the indirect effect was computed as the product of the ‘‘a’’ path (change in experiential avoidance over time) and the ‘‘b’’ path (predictive effect of that change on outcome change). Multiple-groups analysis compared indirect effects across conditions. Variables were grand-mean centred and standardised, and bootstrapping with 5,000 samples was used to estimate test statistics and standard errors. Exploratory serial mediation tested whether changes in experiential avoidance led to changes in connectedness (WCS), which then mediated outcome change; acute psilocybin measures were used in further exploratory analyses. Normality checks and appropriate correlation methods (Pearson or Spearman) were applied as needed.

Fifty-nine participants were randomised: 30 to the psilocybin therapy condition and 29 to escitalopram. The final sample had mean age 41.22 (SD = 10.91), was predominantly male (66.1%), white (88.1%), and highly educated (93.2% with university-level education). Mean duration of the current major depressive episode was 18.66 years. Some protocol deviations occurred: two psilocybin-condition and one escitalopram-condition participants missed the second dosing session due to COVID-19 restrictions; a small number discontinued daily capsules for suspected allocation or adverse events. Within the psilocybin therapy arm, there was a significant reduction in experiential avoidance over time (‘‘a’’ paths). Controlling for change in experiential avoidance, participants showed significant increases in well-being and decreases in clinician-assessed depression (MADRS), self-reported depression (QIDS-SR-16), and trait anxiety, but not a significant direct change in suicidal ideation (SIDAS) when controlling for the mediator (c' paths). Reductions in experiential avoidance predicted increased well-being and decreased MADRS, QIDS-SR-16, SIDAS, and STAI-Trait (b paths). The indirect (a × b) effects via reductions in experiential avoidance were significant for well-being (B = 4.51, SE = 1.98, p_boot < .001), MADRS (B = -3.32, SE = 1.49, p_boot = .010), QIDS-SR-16 (B = -1.82, SE = 0.94, p_boot = .015), SIDAS (B = -1.88, SE = 0.94, p_boot = .010), and STAI-Trait (B = -5.21, SE = 2.14, p_boot = .003). In contrast, the escitalopram condition did not show a significant reduction in experiential avoidance over time (no significant ‘‘a’’ paths). Changes in experiential avoidance did not predict changes in outcomes (non-significant ‘‘b’’ paths), and indirect effects via experiential avoidance were non-significant for well-being (B = 0.25, SE = 0.43, p_boot = .579), MADRS (B = -0.01, SE = 0.20, p_boot = .979), QIDS-SR-16 (B = -0.11, SE = 0.21, p_boot = .694), SIDAS (B = -0.08, SE = 0.21, p_boot = .738), and STAI-Trait (B = -0.38, SE = 0.62, p_boot = .542). Multiple-groups analyses testing differences in indirect effects between conditions found that the mediation via reductions in experiential avoidance was significantly larger in the psilocybin therapy arm versus escitalopram for well-being (B = 4.25, SE = 1.94, p_boot = .011), MADRS (B = -3.31, SE = 1.50, p_boot = .023), QIDS-SR-16 (B = -1.71, SE = 0.96, p_boot = .043), and STAI-Trait (B = -4.82, SE = 2.22, p_boot = .033). The difference for suicidal ideation approached but did not reach statistical significance (B = -1.80, SE = 0.96, p_boot = .058). Regarding predictors of reductions in experiential avoidance within the psilocybin condition, greater reported ego dissolution was associated with larger reductions in experiential avoidance (rs = -0.48, p = .007) as was greater psychological insight (rs = -0.38, p = .037). Mystical-type experience (rs = -0.30, p = .107) and emotional breakthrough (rs = 0.10, p = .592) were not significantly associated. No significant associations between acute-session measures and experiential avoidance changes were observed in the escitalopram arm. Exploratory serial mediation suggested that for several outcomes (except suicidal ideation), improvements via reductions in experiential avoidance were serially mediated through increases in connectedness, though the authors note caution given sample size and exploratory status.

Zeifman and colleagues interpret their findings as supporting reduced experiential avoidance as a putative mechanism by which psilocybin therapy produces improvements in a range of mental health outcomes among people with Major Depressive Disorder. They report that reductions in experiential avoidance mediated improvements in well-being, clinician- and self-rated depression severity, suicidal ideation, and trait anxiety following psilocybin therapy, whereas the same mediation pathway was not observed in the escitalopram arm. Exploratory analyses further indicated that, for most outcomes, the effect of reductions in experiential avoidance on clinical improvement occurred through increases in connectedness, suggesting a possible sequential relationship between decreased avoidance and enhanced connectedness. The authors situate these quantitative results within prior qualitative and survey-based research that described shifts from avoidance to acceptance and from disconnectedness to connectedness after psychedelic experiences. They also reference theoretical neuropharmacological distinctions between SSRIs and psychedelics, proposing that SSRIs may increase passive coping via 5-HT1A-related mechanisms while psilocybin promotes active coping and greater emotional responsiveness via 5-HT2A signalling, which could plausibly underlie differences in how each treatment affects experiential avoidance and connectedness. Several limitations are acknowledged. The reliance on self-report BEAQ at only pre- and post-treatment constrains causal inference about temporal sequencing; observer-based or ecological momentary assessments and biological or neuroimaging markers would strengthen claims about mechanism. The trial sample was modest in size and demographically limited, which reduces generalisability and limits power for complex or serial mediation models. Maintaining blinding in psychedelic trials remains challenging and expectancy or placebo effects could have influenced outcomes despite efforts to minimise them. The authors recommend future studies use larger, more diverse samples, active control drugs or methods to better preserve blinding, neuroimaging to identify treatment-specific neural changes, and designs that measure experiential avoidance more granularly. Clinically, the paper suggests emphasising therapeutic elements that reduce experiential avoidance and foster connectedness within psilocybin therapy, and exploring adjunctive combinations with interventions that target those processes (for example, elements of acceptance and commitment therapy) to potentially enhance efficacy and cost-effectiveness. The authors also note potential relevance of these mechanisms for other disorders characterised by high experiential avoidance, while cautioning that further research across different conditions is required. Overall, the study presents converging quantitative evidence that reductions in experiential avoidance—and increases in connectedness linked to certain acute experiences such as ego dissolution and psychological insight—may be important mechanisms of therapeutic change in psilocybin therapy.