Who Are You After Psychedelics? A Systematic Review and a Meta-Analysis of the Magnitude of Long-Term Effects of Serotonergic Psychedelics on Cognition/Creativity, Emotional Processing and Personality

The researchers conducted a systematic review and meta-analysis following Cochrane and PRISMA guidance; the protocol was registered on PROSPERO (Record ID=297742). Electronic searches of PubMed, Embase and PsycINFO were completed in July 2022 using MeSH and keyword combinations covering psilocybin, DMT/ayahuasca, mescaline, LSD and a broad range of cognition, creativity, emotion and personality terms. Manual reference checks of relevant reviews were also performed. Two independent reviewers screened records (title/abstract then full text) with a third reviewer resolving discrepancies. Inclusion criteria were: interventional (non-microdosing) randomized or non-randomized controlled trials in adults, documented doses of serotonergic psychedelics (LSD, psilocybin, mescaline, DMT), publication in peer-reviewed journals, and use of standardised psychological/neuropsychological measures with baseline and at least one-week post-intervention data. Non-human studies, case reports, reviews and conference abstracts were excluded; only English-language human studies were considered. From eligible full texts the researchers extracted study design, population, drug/dose regimen, participant numbers, outcome measures, follow-up periods and group-by-time effects. Risk of bias was assessed with RoB 2.0 for randomized and crossover trials and ROBINS-I for non-randomized studies. For meta-analysis, pre/post assessments or change scores were used; where necessary SDs were calculated from reported SEs. Standardised mean differences (SMDs, Hedges’ g) were computed for between-group post-treatment differences. Meta-analyses were restricted to constructs with data from at least three studies. Emotional processing was analysed separately for five emotion domains (anger, disgust, fear, happy, sad) using random-effects models in R with the 'meta' and 'metafor' packages. Between-study heterogeneity was quantified using I2, with thresholds of >25%, >50% and >75% indicating low, moderate and high heterogeneity respectively.

Study selection yielded ten controlled studies meeting eligibility criteria, representing 304 enrolled participants; data were obtainable for 236 participants. Six trials were randomized, double-blind and placebo-controlled; the remainder included open-label and participant-blinded crossover designs. Three trials were in clinical populations (social anxiety disorder, major depressive disorder, treatment-resistant depression) and seven in healthy volunteers. Follow-up ranged from one week to three months. Substances and participant counts across trials included ayahuasca (68 participants), psilocybin (60 participants) and LSD (10 participants). Psilocybin doses ranged from 10 mg to 30 mg per session (single or two-session regimens); one LSD study used 75 micrograms. Ayahuasca dosing and alkaloid concentrations varied substantially across studies (reported DMT 0.68–1.58 mg/ml; harmine, harmaline and tetrahydroharmine concentrations also variable). Cognition/creativity: Three studies contributed outcomes. Most reported comparisons between psychedelics and controls were not statistically significant. Notable positive findings included an ayahuasca-associated increase in the originality component of the Torrance Test of Creative Thinking, and a psilocybin-associated increase in the novelty component of the Alternative Uses Task. One psilocybin study reported a decrease in convergent thinking on the Picture Concept Task. For cognitive flexibility, a psilocybin trial reported a reduction in perseverative errors on the Penn Conditional Exclusion Test at 1 and 4 weeks; effect size calculations sometimes showed large effects but multiple comparisons were not controlled and many results were not statistically significant. Emotional processing: Three studies reported emotion-recognition outcomes (two ayahuasca, one psilocybin) using dynamic facial recognition tasks scored for accuracy and reaction time. Individually, the two ayahuasca trials found no statistically significant differences versus control at follow-ups (seven, 14, 21 days, one study to three months), while the psilocybin trial reported a significant improvement in overall reaction time at one month. Pooled reaction-time data at approximately one month from the three studies yielded statistically significant faster recognition times in the active groups for disgust (SMD = -0.63, 95% CI [-1.01 to -0.25], I2 = 65%) and sadness (SMD = -0.45, 95% CI [-0.85 to -0.06], I2 = 60%). There were no significant pooled differences for anger, happiness or fear. The authors note that one included study combined psychotherapy with the active drug, so combined effects cannot be excluded. Personality: Two eligible studies assessed personality. No significant changes in personality traits were observed 21 days after ayahuasca in one trial. A separate LSD trial reported a significant increase in openness at two weeks. Effect-size calculations showed non-significant large decreases in extraversion for the ayahuasca group and smaller decreases for the LSD group compared with controls.

The authors interpret the aggregated evidence as providing limited support for enduring effects of serotonergic psychedelics on cognition/creativity, emotional processing and personality when considering contemporary controlled trials. For cognition and creativity, most controlled findings were null; some isolated signals (increased divergent thinking after ayahuasca or psilocybin, altered convergent thinking and improvements in a set-shifting measure after psilocybin) were observed but are inconsistent across studies and hampered by small samples and multiple outcome testing without correction. For emotional processing, pooled reaction-time data suggested faster recognition of disgust and sadness in active treatment groups, which the authors note could reflect enhanced processing or alternatively a shift toward negative affective bias. They emphasise that one trial included additional psychotherapy only in the active arm, limiting attribution of effects purely to the drug. For personality, limited and inconsistent evidence was reported: an LSD trial found increased openness at two weeks, while ayahuasca trials did not show persistent trait change; prior pooled analyses and single trials not included here due to unavailable data have reported longer-term openness changes, but the current controlled-trial evidence is inconclusive. The authors acknowledge multiple limitations that temper confidence in conclusions: heterogeneity in design, timing and outcome measures; blinding and expectancy confounds; small sample sizes; unaccounted differences in demographics and prior psychedelic exposure; and restriction to English-language, post-1990s studies. They also describe strengths of the review, including its focus on controlled trials with documented doses and standardised measures. The authors recommend future studies with larger samples, better control conditions, standardised dosing and longer follow-up to confirm or refute the preliminary findings and to clarify therapeutic or mechanistic significance.

The authors conclude that contemporary controlled trials provide very limited evidence for lasting beneficial effects of serotonergic psychedelics on cognition/creativity, emotional processing and personality. Nonetheless, pooled analyses offer preliminary evidence that these drugs may reduce reaction times for recognising disgust and sadness. They recommend future trials with larger samples, improved controls, standardised validated measures and longer-term follow-ups to verify these tentative findings.