Perceived changes in mental health and social engagement attributed to a single psychedelic experience in autistic adults: results from an online survey

Autism spectrum disorder is characterised by lifelong differences in social communication and restricted or repetitive behaviours, and autistic people experience elevated rates of mental health problems, loneliness and reduced quality of life. Recent clinical and theoretical work on classic psychedelics (5-HT2A agonists such as LSD, psilocybin and DMT) and related drugs has emphasised rapid and sometimes durable improvements in mood and social connectedness in non-autistic samples. Proposed mechanisms include acute ‘pivotal’ or hyper‑plastic neural states that permit revision of rigid, maladaptive beliefs and promote psychological flexibility and social connectedness. However, contemporary empirical studies specifically examining psychedelic effects in autistic people are scarce, with only a very small MDMA trial and protocol-level work known to the authors. Stroud and colleagues therefore conducted an exploratory online survey to examine how autistic adults attribute perceived long‑term changes in mental health and social functioning to a single, self‑identified “most impactful” psychedelic experience. The study also investigated associations between these perceived changes and mechanistically relevant variables such as features of the acute psychedelic experience (mystical and challenging dimensions), changes in psychological flexibility, loneliness, camouflaging and autistic traits, with the stated aim of informing whether progression to controlled intervention studies in this population is warranted.

The study was a cross‑sectional, anonymous online survey approved by University College London ethics committee; participants provided electronic informed consent. Recruitment used convenience sampling via social media, psychedelic and autism forums, and collaboration with the Autistic Psychedelic Community. Inclusion criteria were age 18 or older, English fluency, reporting a formal autism diagnosis or self‑identifying as autistic, and at least one lifetime use of a psychedelic substance. Of 284 survey completers, the analysed sample was restricted to 233 participants who anchored their selected “most impactful” experience to a classic psychedelic (e.g., LSD, psilocybin, mescaline, DMT); 51 participants were excluded because their index experience involved MDMA, cannabis or ketamine. Participants completed the survey once (Qualtrics). They identified their most impactful psychedelic experience and rated its relative impact. Acute subjective experience during that episode was assessed with the 30‑item Mystical Experience Questionnaire (MEQ‑30) and the 26‑item Challenging Experience Questionnaire (CEQ); both were rescaled to a 0–100 range for interpretability. A suite of established measures was adapted to retrospectively assess perceived long‑term change attributed to the experience: the DASS‑21 for general psychological distress (and subscales for depression, anxiety, stress), the Social Phobia Inventory (SPIN) for social anxiety, the Acceptance and Action Questionnaire II (AAQ‑II) to index psychological (in)flexibility (note: decreases on the adapted AAQ‑II indicate increased psychological flexibility because items are negatively worded), the AQ‑Short (adapted) and AQ‑10 (current autistic traits), a shortened Camouflaging Autistic Traits Questionnaire (CAT‑Q; 9 items), the Social Connectedness Scale (adapted SCS), a social engagement measure (LSNS‑6 style with family and non‑kin factors), the Three‑Item Loneliness Scale (TILS), and two WHOQOL items on relationship and sexual satisfaction. Adapted outcome items asked participants whether their selected experience led to enduring changes (defined in the survey as lasting more than a few days), and response options ranged from −5 (decreased very much) to +5 (increased very much). The adapted measures showed acceptable internal consistency in this sample (Cronbach’s alpha ≥ 0.8) though they were not fully psychometrically validated. Statistical analysis used JASP 0.16.3. The principal inferential approach comprised two linear regressions predicting perceived change in psychological distress (adapted DASS‑21) and in social engagement (LSNS‑6 total) from predictors that included MEQ and CEQ scores and perceived changes on other adapted measures (AQ‑Short, TILS, SPIN, AAQ‑II, SCS, camouflaging), while controlling for age, gender identity, years of education and time since the indexed experience. Distributional assumptions, multicollinearity and homogeneity of variance were inspected graphically. Descriptive statistics, chi‑square and t‑tests were used to characterise a subgroup that reported increased distress; additional correlations are reported in the supplement.

The analysed sample comprised 233 participants, mean age 29.8 years (SD = 9.3), with the largest demographic being well‑educated men from North America or Europe. A majority (n = 136; 58.4%) reported a professional autism diagnosis; the remainder (n = 97) self‑identified as autistic. Mean AQ‑10 scores were above the screening cut‑off in both groups (formal diagnosis: 7.31; self‑identified: 7.29) and did not differ between them. Most respondents (206/230 with complete demographics; 89.7%) reported at least one additional neurodevelopmental or mental health condition, the commonest being ADHD (n = 129; 56.1%), depression (n = 81; 35.2%), generalised anxiety disorder (n = 80; 34.8%) and social anxiety disorder (n = 72; 31.3%). The selected psychedelic experience had occurred a mean of 3.99 years previously (SD = 6.39). Over half (n = 127; 54.7%) reported co‑use of multiple substances during the experience, most often a classic psychedelic plus cannabis (n = 67; 28.8%). On measures of acute subjective experience, mean MEQ score was 63.5 (SD = 22.7) and mean CEQ score was 22.68 (SD = 20.16), values broadly similar to prior trials; 34% of participants had MEQ scores exceeding 60% of the maximum (a proposed threshold for a ‘complete’ mystical experience). Seventy‑eight percent of respondents rated their indexed experience among the 10 most impactful of their lives. Across the adapted perceived‑change measures, the largest mean changes (absolute value exceeding ±1 on the −5 to +5 scale) were observed for non‑kin social engagement, social connectedness, relationship satisfaction, general psychological distress, depression, stress, social anxiety and psychological flexibility. Numerically, average perceived changes were modest. In the regression predicting change in psychological distress (DASS‑21), several predictors reached statistical significance, including MEQ (p = 0.048), change in loneliness and change in SPIN; however, the largest and most substantial predictor by standardised beta (Std B) was change in psychological inflexibility (AAQ‑II) with Std B = 0.43, indicating that greater increases in psychological flexibility were associated with larger reductions in distress (standardised beta provides an effect‑size measure for regression coefficients). The CEQ did not significantly predict changes in psychological distress (Std B ≤ 0.03, p ≥ 0.47). For change in social engagement (LSNS‑6 total) five predictors were significantly associated with LSNS‑6 change while controlling for covariates, but effects were modest (largest Std B ≤ 0.28). MEQ showed a small effect on DASS‑21 but no effect on LSNS‑6. Time since experience was also associated with LSNS‑6 scores (Std B = −0.16, p = 0.02), suggesting some attenuation with longer elapsed time. A subgroup of 26 participants reported a net increase in psychological distress on the adapted DASS‑21 (mean = 0.82, SD = 0.70). This subgroup was demographically similar to those reporting no change or improvement, but reported increased loneliness (TILS mean = 0.87), decreased connectedness (SCS mean = 0.30) and decreased psychological flexibility (AAQ‑II mean = 0.64) relative to the no‑worse group (all p < 0.001). The subgroup also reported higher CEQ scores (mean = 33.96) compared with the no‑worse group (mean = 20.66; p = 0.011); MEQ scores did not differ significantly between groups (p = 0.517).

Stroud and colleagues interpret their findings as indicating that many autistic respondents attribute enduring improvements in a range of mental health and social outcomes to a single, subjectively impactful psychedelic experience. Reported improvements included reductions in general psychological distress, social anxiety and loneliness, and increases in social engagement, psychological flexibility and social connectedness. The authors note parallels with non‑autistic samples where mystical‑type acute experiences have been linked to improved mental health, although in this dataset the MEQ’s effect on distress was small. By contrast, perceived increases in psychological flexibility emerged as the most substantial predictor of reductions in distress, consistent with theoretical models that cast flexibility as a transdiagnostic mediator of psychedelic‑related therapeutic change; the authors briefly define psychological flexibility as the capacity to respond non‑rigidly to distress while pursuing valued goals. The investigators emphasise caution. A meaningful minority of participants attributed enduring negative outcomes to their experience, including increased anxiety, loneliness and reduced connectedness, and some rated their indexed experience among the most negatively impactful of their lives. The authors suggest several explanations for adverse outcomes: non‑clinical contexts, impure or mixed substances, non‑standardised dosing and lack of trained facilitators may exacerbate challenging experiences and worsen long‑term outcomes; higher baseline rates of anxiety and comorbidity in autistic people could elevate risk; and autism‑specific neurodevelopmental differences may interact with psychedelic effects in unpredictable ways. The CEQ was higher in the subgroup with increased distress, but CEQ did not predict outcomes in the main regressions, which the authors attribute in part to the CEQ’s inability to distinguish resolved from unresolved challenging episodes; they recommend additional measures (e.g. Psychological Insight Scale, Emotional Breakthrough Inventory) in future work. Several limitations are acknowledged: the convenience, self‑selecting sample (including recruitment via psychedelics‑related groups) likely biases results toward positive experiences; reliance on cross‑sectional, retrospective reports risks recall bias and may capture short‑term afterglow rather than durable change; adapted questionnaires were not fully psychometrically validated and using modified measures as both predictors and outcomes complicates interpretation. The authors also note measurement overlap between autistic trait measures (AQ‑Short) and mental health measures (e.g., SPIN), which may explain the observed perceived reductions in autistic‑related social behaviour as downstream effects of reduced anxiety rather than changes in core neurodevelopmental traits. For future research, the authors recommend careful, ethical small‑scale experimental work with fully informed adult autistic participants, following contemporary safety guidelines, tight exclusion criteria for psychosis and epilepsy, systematic recording of adverse events and meaningful engagement with the autistic community. They stress that research should focus on improving wellbeing and mutual understanding rather than attempting to ‘treat’ or normalise autism, and that any intervention development should explicitly consider potential elevated vulnerability to adverse effects in this population.