Can Psychedelic Use Benefit Meditation Practice? Examining Individual, Psychedelic, and Meditation-Related Factors

Meditation practices and psychedelic use both induce non-ordinary states of consciousness and have attracted growing scientific and public interest as potential means to improve psychological wellbeing. Earlier research has suggested overlapping phenomenology and neural targets between some meditation forms and psychedelics, for example changes in the Default Mode Network and experiences of ego dissolution that have been associated with longer-term wellbeing. Empirical work to date includes a small experimental trial administering psilocybin during a five-day meditation retreat that reported greater ego dissolution and psychosocial gains at follow-up, large cross-sectional and longitudinal surveys linking psychedelic exposure with higher levels of current mindfulness practice, and qualitative reports that many people perceive enhancements when combining psychedelics with meditation. Despite these signals, it remains unclear which specific person-level, psychedelic-related, or meditation-related factors are associated with meditators perceiving that psychedelics benefit their meditation practice. Jiwani and colleagues set out to address this question using a large sample of regular meditators who had used psychedelics. Their aim was to identify which variables across three domains—individual characteristics, psychedelic-related factors (including set and setting), and meditation-related factors—were most strongly associated with the self-reported influence of psychedelic experiences on the quality of regular meditation. To accommodate a wide range of correlated predictors, the investigators employed machine learning methods and developed tailored survey items to capture perceived effects of psychedelics on meditation practice alongside multidimensional measures of meditation background and psychedelic use.

The study used an online convenience sample recruited via social media and targeted advertising to contemplative communities between 29 October 2020 and 8 June 2021. Eligibility required adults (≥18 years) with a good understanding of English and a regular meditation practice defined as at least three sessions per week during the prior 12 months. Although the recruitment materials described the study focus on psychedelics and meditation, psychedelic exposure was not required; analyses reported here were restricted to participants who had both meditation and psychedelic exposure (N = 863). The sample had a mean age of 37.7 years (SD = 12.6), 79.4% were assigned male sex at birth, and mean education was 17.5 years (SD = 3.3). Data were collected via Qualtrics after written informed consent and ethical approval. The primary outcome was a single-item question asking participants to rate, on a 7-point Likert scale (1 = strong negative influence to 7 = strong positive influence), whether their psychedelic experiences had influenced the quality of their regular meditation practice. Predictor variables were organised into three domains. Individual factors included demographics, two life-satisfaction items, Big Five personality traits (BFI), and substance use (alcohol, tobacco, cannabis). Psychedelic-related factors covered frequency of use (excluding microdosing, MDMA/ecstasy, and ketamine), age of initiation, lifetime exposure to specific substances (psilocybin, LSD, ayahuasca, N,N-DMT, 5-MeO-DMT, mescaline, others), and single-item measures of set and setting such as intention setting, attention to safety, being with trusted people, and having an interpretative framework. Meditation-related factors captured frequency and years of practice, retreat experience (days and length), whether meditation or psychedelic use came first, adequacy of meditation framework, and a coded taxonomy of 17 practice types derived from participants’ free-text descriptions. Analyses were not preregistered; data and code are available on the Open Science Framework. Descriptive zero-order Pearson correlations between outcome and predictors were computed with false discovery rate correction. For multivariable modelling, three algorithms were compared: ordinary least squares (OLS) linear regression, elastic net linear regression (a regularised linear model that penalises coefficients to reduce overfitting), and random forest (a non-parametric ensemble tree method). Model selection and hyperparameter tuning used three repeats of 10-fold cross-validation (30 held-out folds), optimising mean R2 on held-out data. Feature engineering included winsorising predictors at ±2 standard deviations, k-nearest neighbour imputation for missing predictor data, dimension reduction via principal components for certain correlated groups, and Yeo-Johnson power transformation for quantitative predictors used in linear models. Feature importance for the selected best model was assessed with three complementary approaches: SHAP values (a game-theoretic explanation method), a Bayesian feature ablation procedure estimating incremental R2 contributions, and standardised coefficients from the elastic net model.

On the primary outcome, participants reported on average that psychedelic experiences improved the quality of their regular meditation practice (mean = 5.49 on a 1–7 scale, SD = 1.24, indicating between “slight” and “moderate” positive influence). Coding of meditation backgrounds found the most frequently reported practice types were eclectic/self-created practices (22.0%), Western-style Vipassana (19.1%), and teachings associated with Sam Harris (16.1%). Participants reported between 0 and 7 practice types with a median of two. Zero-order correlations between the outcome and predictors ranged from −0.17 to 0.34, and 28 predictors remained statistically significant after false discovery rate correction. The strongest bivariate associations were psychedelic-related: 12-month psychedelic use (r = 0.34, 95% CI [0.28 to 0.40], p < .001), frequency of setting intentions during psychedelic use (r = 0.30, 95% CI [0.23 to 0.36], p < .001), and lifetime exposure to N,N-DMT (r = 0.23, 95% CI [0.16 to 0.29], p < .001). Among individual factors, higher cannabis use (r = 0.18), openness to experience (r = 0.16), and agreeableness (r = 0.15) showed the largest positive associations. For meditation-related variables, less retreat experience correlated negatively with the outcome (r = −0.17) while engaging in eclectic/self-created practices had a small positive association (r = 0.11). In multivariable analyses, the elastic net linear regression had the best out-of-sample performance (mean R2 = 0.266, SE = 0.013) compared with random forest (R2 = 0.252, SE = 0.011) and OLS (R2 = 0.251, SE = 0.012). For the elastic net model the root mean squared error across held-out folds was 1.04 and mean absolute error was 0.848 on the 7-point outcome. Three complementary feature-importance methods converged on the same leading predictors. SHAP values identified lifetime/12-month psychedelic use and setting intentions during psychedelic use as the most influential features, with global SHAP values of approximately 0.25 and 0.15 respectively. The next most important features by SHAP were agreeableness, exposure to N,N-DMT, and longest retreat. Bayesian feature ablation quantified incremental contributions to model R2: the largest increases were for psychedelic use (ΔR2 = 0.048) and intention setting (ΔR2 = 0.019); agreeableness and N,N-DMT contributed more modestly (ΔR2 = 0.006 and 0.005, respectively). Standardised coefficients from the elastic net echoed these rankings: psychedelic use β = 0.30, setting intentions β = 0.18, agreeableness β = 0.09, and N,N-DMT β = 0.08.

Jiwani and colleagues interpret the findings as indicating that, within this sample of regular meditators who have used psychedelics, those who report greater and more regular psychedelic use and who report setting intentions before psychedelic sessions are most likely to perceive psychedelics as benefiting their meditation practice. Personality, specifically higher agreeableness, and exposure to N,N-DMT also showed consistent but smaller associations. The investigators suggest a profile of individuals who treat psychedelic use as a practice that may be integrated with meditation: regular engagement, intentional preparation, and exposure to substances that commonly occasion intense transcendence-like experiences may together foster perceived synergy. The authors place these results in relation to prior work, noting consistency with studies linking regular psychedelic use to wellbeing and mindfulness practice and with literature implicating intention setting in facilitating peak or mystical experiences. They also highlight a divergence from some previous findings in that openness to experience was not among the most important predictors in the present cross-sectional analysis, whereas experimental work has sometimes found openness to predict positive acute psychedelic experiences. Key limitations acknowledged by the authors include the cross-sectional design, which precludes causal inference about whether psychedelic use leads to improvements in meditation quality or whether individuals predisposed to perceive benefits are more likely to use psychedelics regularly. The authors further note that many variables showed significant zero-order associations even after correction, indicating a nuanced picture that may warrant targeted follow-up. They call for longitudinal designs and randomised trials that manipulate psychedelic exposure to establish causality and to identify safe and effective frequencies of use. Practical and clinical implications discussed by the investigators include the need to develop therapeutic models for integrating meditation and psychedelics, to temper expectations about single-session effects in favour of understanding longer-term practice, and to remain attentive to psychological risks associated with psychedelics such as challenging experiences, hallucinogen persistent perception disorder, and prolonged psychosis. Finally, the extracted Limitations section in the available text is incomplete; the cross-sectional limitation is explicitly stated but other acknowledgements of potential bias or generalisability constraints may be present in the full article beyond the provided extraction.