Shame, guilt and psychedelic experience: Results from a prospective, longitudinal survey of real-world psilocybin use

Classic psychedelics such as psilocybin produce profound shifts in affect, perception and cognition, and individuals often characterise these shifts as deeply meaningful. Earlier literature has documented a spectrum of psychedelic experiences from intensely positive to psychologically challenging, and scholars have debated whether difficult experiences reflect modifiable contextual factors (set and setting) or user vulnerability. Despite growing attention to challenging affects such as fear and grief, comparatively little research has examined shame and guilt in the context of psychedelic experiences, even though these self-conscious emotions are implicated in a wide range of psychological disorders and have distinct appraisals and behavioural tendencies (shame involving a negative view of the self, guilt involving appraisal of a specific wrong). The introduction also distinguishes state-shame (acute, context-specific) from trait-shame (a more stable predisposition), noting that both may be relevant to outcomes after psychedelic use but have been underexplored empirically. Mathai and colleagues set out to address this gap using data from a larger prospective, longitudinal online survey of naturalistic psilocybin use. Their specific aims were to: 1) characterise acute shame and guilt experienced during psilocybin sessions and identify general predictors of those experiences; 2) test whether acute shame or guilt and the ability to process them predict wellbeing 2–4 weeks later; and 3) examine whether trait shame (measured with the External and Internal Shame Scale) changes after psilocybin use over a 2–3 month window. The authors framed the study as exploratory but focused on whether acute self-conscious emotions may be activated by psilocybin and whether their processing relates to longer-term psychological outcomes.

This was a prospective, longitudinal survey of adults (≥18 years) planning to use psilocybin outside clinical research. Recruitment occurred online between 22 July 2020 and 14 July 2022 and involved six automated Qualtrics surveys timed around a reference psilocybin experience: baseline (about 2 weeks prior), 1 day prior, 1–3 days after, 2–4 weeks after (primary endpoint), and 2–3 months after. Items assessing shame and guilt were added after the study had begun, so the sample available for those measures (SSGS) was smaller than the full cohort. The analytic sample for acute state shame/guilt analyses was n = 679 (participants with SSGS data); a complete-case sample for trait-shame longitudinal analyses (EISS at all three timepoints) was n = 215. Participants were predominantly US-resident, White, college-educated, male, in relationships, and with prior psychedelic experience; mean age ranged about 38.9–41 years. Baseline psychiatric history included substantial rates of anxiety and mood disorders. Typical reported dose form was dried mushrooms (mean dose = 3.1 grams) and motivations for use commonly included self-exploration and mental-health reasons. Setting and sitter presence varied in this naturalistic sample. Key measures included the 8-item External and Internal Shame Scale (EISS) for trait shame, the State-Trait Anxiety Inventory (STAI) for state and trait anxiety, a modified Beck Depression Inventory (BDI-II) excluding the suicidality item, the revised Adverse Childhood Experience scale (ACE), the Cognitive Flexibility Scale (CFS), and the 10-item State of Surrender scale (SOS). Acute post-session instruments were the Mystical Experience Questionnaire (MEQ), the Challenging Experience Questionnaire (CEQ), and an adapted 10-item State Shame and Guilt Scale (SSGS) with separate shame and guilt subscales scored 5–25. Participants who reported any shame or guilt (SSGS > 5) received follow-up categorical questions on duration, difficulty (1–5) and whether they were able to work through these feelings constructively (1–5); the latter two items were analysed as "shame/guilt difficulty" and "shame/guilt processing." Wellbeing at 2–4 weeks was assessed with a single 1–7 Likert item about perceived long-term changes to personal wellbeing. Statistical analyses were tailored to each aim. For Aim 1, descriptive statistics and two multivariable least-squares regression models predicted SSGS-Shame and SSGS-Guilt (n = 679) from candidate covariates (age, gender, race, education, marital status, religious belief, baseline trait shame, STAI-State/Trait, BDI-II, ACE, CFS, SOS, and sitter presence); covariates with p < 0.2 entered final models, with checks for multicollinearity (VIF threshold 4). Pearson correlations (Šidák correction for five comparisons) illustrated relationships between trait shame and other baseline psychometrics. For Aim 2, multiple linear regression tested predictors of wellbeing 2–4 weeks after use, including age, baseline STAI-Trait, SSGS-Shame, SSGS-Guilt, shame/guilt difficulty and processing, CEQ and MEQ; follow-up regressions examined shame-only and guilt-only subsamples. For Aim 3, repeated-measures one-way ANOVA (Geisser–Greenhouse correction) compared EISS across baseline, 2–4 weeks and 2–3 months (n = 215); Tukey post-hoc tests were used for pairwise comparisons. The paper notes some variables (for example, exact dose, concurrent substances or sitter qualifications) were not incorporated into predictive models.

Aim 1a — prevalence and characteristics: Among 679 participants with SSGS data, 463 (68.2%) reported some shame or guilt during their psilocybin session (SSGS subscale > 5). Specifically, 378 individuals (55.7%) reported experiences of shame and 352 (51.8%) reported guilt; 80 participants (11.8%) rated a shame- or guilt-related feeling at the maximal value. Reported durations clustered in the 10–60 minute category (categorical range: <10 minutes to entire session). Despite frequent acute shame/guilt, the vast majority of sessions were rated positively (89.7% rated the experience as positive). Aim 1b — predictors of acute shame/guilt: Pearson correlations showed baseline trait shame (EISS) was strongly associated with baseline psychometrics: STAI-State r = 0.56, STAI-Trait r = 0.77, BDI-II r = 0.70, ACE r = 0.31, and CFS r = -0.55 (all adjusted p < 0.001). However, baseline trait shame did not predict acute state shame (β = 0.005, p = 0.90) or guilt (β = -0.004, p = 0.94). In the final multivariable model for SSGS-Shame (covariates retained: age, race, STAI-Trait, CFS), only younger age (β = -0.04, p < 0.001) and higher baseline trait anxiety (STAI-Trait; β = 0.08, p < 0.001) significantly predicted higher acute shame scores. For SSGS-Guilt (covariates retained: age, SOS, sitter), only younger age predicted higher acute guilt (β = -0.05, p < 0.001). Aim 2 — relationship of acute shame/guilt and processing to wellbeing: In multiple regression predicting wellbeing at 2–4 weeks (n = 679), higher mystical experience (MEQ; β = 0.39, p < 0.001) and greater ability to process shame/guilt during the session (shame/guilt processing; β = 0.16, p = 0.01) were significant positive predictors. Acute shame (β = 0.02, p = 0.57), acute guilt (β = -0.02, p = 0.32), shame/guilt difficulty (β = 0.04, p = 0.55) and CEQ scores (β = 0.03, p = 0.96) were not significant predictors. Numerically, MEQ and wellbeing scores were highest among participants who reported the greatest constructive processing (level 5), even compared with participants who reported no shame/guilt. Subsample analyses isolating shame-only (n = 112) and guilt-only (n = 85) participants produced similar patterns: for shame-only, MEQ (β = 0.33, p = 0.004) and shame processing (β = 0.26, p = 0.04) predicted wellbeing; for guilt-only, MEQ (β = 0.44, p < 0.001) and guilt processing (β = 0.29, p = 0.01) predicted wellbeing. Aim 3 — change in trait shame over time: Among 215 participants with complete EISS data, repeated-measures ANOVA indicated a significant effect of psilocybin on trait shame [F(1.96, 418.4) = 18.18, p < 0.001]. Post-hoc comparisons showed a small but significant decrease in EISS at 2–4 weeks compared with baseline (Cohen's dz = 0.31; adjusted p < 0.001) and at 2–3 months compared with baseline (Cohen's dz = 0.37; adjusted p < 0.001), with no significant difference between the two post-session follow-ups (adjusted p = 0.36). At the individual level, 56.7% improved on EISS at 2–4 weeks, 13.5% were unchanged, and 29.8% showed increased trait shame.

Mathai and colleagues interpret these findings as evidence that acute experiences of shame and guilt are common during real-world psilocybin sessions and can vary widely in intensity. Younger age predicted both acute shame and guilt, and higher baseline trait anxiety predicted acute shame; otherwise, few baseline demographic or psychological variables (including trait shame itself) predicted state-shame or state-guilt. The authors highlight that state-shame/guilt scores observed here were higher than baseline means for healthy volunteers and comparable to shame-induction experimental paradigms, supporting the claim that psilocybin can uniquely activate self-conscious emotions. Crucially, the data indicate that the ability to constructively work through shame or guilt during the session — rather than the mere presence or intensity of those emotions — predicted improved wellbeing 2–4 weeks later. Higher mystical-type experience (MEQ) also predicted wellbeing, and those who processed shame/guilt most constructively tended to report the highest MEQ and wellbeing scores. The authors situate this pattern within literature on "emotional breakthrough" and models proposing that encountering and resolving difficult material can be part of therapeutic change in psychedelic-assisted contexts. At the same time, Mathai and colleagues emphasise a cautionary note: a notable minority exhibited increased trait shame after psilocybin, suggesting that heightened emotional plasticity without resolution may reinforce negative self-beliefs or retraumatise some users. They argue that context — preparation, support during dosing, and post-session integration — likely moderates whether shame-related activation leads to growth or harm, and recommend attention to these factors in clinical practice and consent processes. The authors acknowledge several limitations. Naturalistic design entailed minimal standardisation of dose, setting and concurrent substance use, limiting causal inference but reflecting real-world heterogeneity. The sample was predominantly White, educated and self-selected, reducing generalisability to more diverse populations. Measurement challenges included the brevity of the trait shame instrument, the late addition of shame/guilt items (reducing sample size), potential underreporting of shame, and longitudinal attrition. They call for replication in controlled research settings, exploration across clinical populations, and use of qualitative or mixed methods to unpack the phenomenology and mechanisms of shame/guilt processing with psychedelics. Finally, the authors recommend that clinicians disclose the possible emotional demands of psychedelic-assisted therapies and be trained to navigate complex emotions such as shame and guilt, while also working to reduce stigma for individuals who undergo challenging psychedelic journeys.