Effects of classical psychedelics on implicit and explicit emotional empathy and cognitive empathy: a meta-analysis of MET task

Empathy is a multifaceted social-cognitive capacity encompassing cognitive empathy (understanding another's mental states) and emotional empathy (sharing another's feelings), with the latter often divided into implicit (automatic arousal) and explicit (deliberate self-report) components. Earlier experimental studies reported mixed findings on how classical psychedelics—principally LSD, psilocybin and ayahuasca, all 5-HT2A receptor agonists—affect social cognition: some work found increases in emotional empathy and sociality, while effects on cognitive empathy have been inconsistent. No previous meta-analysis had synthesised results across a single standardised task to clarify the overall pattern of effects across these substances. In response to that gap, Olami and colleagues conducted a PRISMA-guided meta-analysis restricted to studies that used the Multifaceted Empathy Test (MET), a validated task that provides measures of cognitive empathy (accuracy in identifying emotions from photos), explicit emotional empathy (self-report ratings) and implicit emotional empathy (arousal ratings). By limiting synthesis to one task, the researchers aimed to improve comparability across studies that varied in drug type, dose, setting and timing of assessment, and to estimate the pooled effects of LSD, psilocybin and ayahuasca on the MET-derived empathy outcomes.

The authors searched PubMed (Medline), PsycINFO and Scopus from inception to November 2023 using keywords for classical psychedelics (LSD, psilocybin, ayahuasca, DMT and related terms) combined with empathy- and MET-related terms. Eligible studies were peer-reviewed, English-language, original articles on healthy human participants that administered at least one dose of LSD, psilocybin or ayahuasca and reported MET outcomes. Exclusions comprised reviews, non-human studies, retrospective self-report assessments of prior psychedelic use, and non–full-text items (letters, abstracts, conference papers). No exclusions were applied for participant age, sex, drug dose, administration route or study size. Data extracted included study identifiers, sample sizes and demographics, drug type and dose, and MET outcome means and standard deviations. When means and SDs were not reported, effect sizes were estimated by digitising figures, or calculated from reported F or p values. The MET outcomes analysed were: cognitive empathy (accuracy measured as percentage correct), explicit emotional empathy (self-report ratings), and implicit emotional empathy (arousal ratings). Positive and negative stimulus sub-analyses were performed where data permitted. Meta-analyses used Hedges' g (a small-sample bias–corrected standardised mean difference, interpreted with Cohen's D benchmarks) and were implemented using the metafor package in RStudio (version 2023.06.1) and Comprehensive Meta-Analysis (Version 3.3.070). A random-effects framework was applied given expected between-study variability. Heterogeneity was quantified with the Q statistic and I2 index, with I2 values of ~25%, 50% and 75% taken as low, moderate and high heterogeneity. Small-study bias was assessed with Egger's regression test and funnel-plot–based trim-and-fill analyses. Influence diagnostics and sensitivity analyses were used to identify outliers. Risk-of-bias assessment was performed in duplicate: the revised Cochrane RoB 2 tool for randomized trials and ROBINS-I for non-randomised intervention studies, with disagreements resolved by discussion.

Five studies contributed data for the primary MET analyses (combined overall sample n = 158). For cognitive empathy (accuracy), classical psychedelics showed no statistically significant pooled effect versus placebo or baseline (k = 5, Hedges' g = -0.1153, 95% CI -2.0316 to 1.8010, p = 0.9061, Z = -0.1179). Between-study heterogeneity was high and statistically significant (Q = 36.7194, df = 5, I2 = 87.34%, p < 0.0001). Explicit emotional empathy increased under classical psychedelics compared with placebo or baseline (k = 5, combined n = 158, Hedges' g = 0.7921, 95% CI 0.1786 to 1.4057, p = 0.0114, Z = 2.5304). Heterogeneity was negligible for this outcome (Q = 1.1867, df = 5, I2 = 0.00%, p = 0.9461). Tests for small-study bias were not significant (Egger's p = 0.7800), although a trim-and-fill analysis estimated two potentially missing studies on the right; influence diagnostics identified no overall outliers for the pooled explicit-empathy estimate. Implicit emotional empathy ratings were also elevated following classical psychedelic administration (k = 5, combined n = 158, Hedges' g = 0.7635, 95% CI -0.0191 to 1.5461, p = 0.0500, Z = 1.9121). Between-study heterogeneity was negligible (Q = 0.4442, df = 5, I2 = 0.00%, p = 0.9940). Egger's test did not indicate small-study bias (p = 0.6427), though trim-and-fill suggested three possibly missing studies on the right; influence diagnostics did not flag outliers for the overall implicit-empathy analysis. Analyses stratified by stimulus valence yielded the following summaries. For cognitive empathy, neither positive nor negative stimuli produced significant effects (positive: k = 3, n = 106, Hedges' g = 0.7794, 95% CI -0.4135 to 1.9722, p = 0.2003; negative: k = 3, n = 106, Hedges' g = 0.7432, 95% CI -0.4431 to 1.9295, p = 0.2195); heterogeneity for these sub-analyses was low (I2 = 0.00%). For explicit emotional empathy by valence, positive stimuli were associated with a large pooled effect (k = 3, n = 109, Hedges' g = 1.6016, 95% CI -1.0784 to 4.2815, p = 0.0415), but between-study heterogeneity was substantial and significant (Q = 8.4933, I2 = 78.58%, p = 0.0143) and Egger's test indicated possible small-study bias (p = 0.0391); influence diagnostics identified the Dolder et al. study as an influential outlier. For negative stimuli, explicit emotional empathy was also elevated (k = 3, n = 109, Hedges' g = 0.8142, 95% CI -0.3727 to 2.0010, p = 0.0078) with low heterogeneity (I2 = 0.00%) and no clear publication-bias signal. Implicit-empathy valence analyses showed increases for both positive (k = 4, n = 128, Hedges' g = 1.0934, 95% CI -0.0528 to 2.2397, p = 0.0415; I2 = 15.32%) and negative stimuli (k = 5, n = 158, Hedges' g = 0.9082, 95% CI -0.0189 to 1.8353, p = 0.0500; I2 = 0.00%). One influential outlier for implicit positive stimuli was again identified as Dolder et al. Publication-bias tests were generally non-significant across outcomes except for explicit positive stimuli, where Egger's test suggested bias. Risk-of-bias assessment found heterogeneity in study quality. Using RoB 2 on three randomized trials, two were judged low risk and one (Uthaug et al.) high risk due to deviations from intended interventions and reliance on facilitator-administration in retreat settings. ROBINS-I applied to two nonrandomised studies rated Mason et al. as low-to-moderate risk and Kiraga et al. as having a serious overall risk owing to confounding, selection bias, and missing-data concerns.

Olami and colleagues interpret their pooled findings as evidence that classical psychedelics tend to enhance emotional components of empathy—both explicit self-reported empathy and implicit arousal responses—while producing no consistent change in cognitive empathy as measured by the MET. They situate these results alongside prior experimental reports of increased emotional empathy following LSD, psilocybin and ayahuasca. The authors propose several plausible mechanisms to explain heightened emotional empathy. Perceptual alterations under psychedelics, including changes in bodily perception and reports of ego dissolution (a reduced sense of separateness from others), might blur self–other boundaries and thereby increase emotional resonance. Neurophysiological correlates invoked include reductions in posterior alpha power previously linked to heightened emotional empathy. Personality change provides another possible pathway: psychedelics have been associated with increases in openness, and higher openness correlates with greater empathy. Comparisons are drawn to MDMA, which also enhances emotional empathy and modulates serotonergic systems; the authors note that LSD's broad receptor profile (including dopaminergic binding) and ayahuasca's additional MAOI and β-carboline constituents (e.g. harmine) could produce distinct effects, even though the meta-analysis grouped substances on the basis of shared 5-HT2A agonism. They further cite a pharmacological observation that ketanserin (a 5-HT2A antagonist) did not fully block LSD-associated increases in emotional empathy, implying additional receptor mechanisms may be involved. The paper acknowledges important limitations that constrain interpretation. Chief among these is the small number of eligible studies (five), which reduces statistical power and the generalisability of pooled estimates. Heterogeneity in substances, doses, settings (laboratory versus retreat/ceremony), timing of assessment and study designs contributes further uncertainty. Risk-of-bias assessments showed variability across studies, from low to high/serious risk, particularly where interventions occurred in uncontrolled retreat contexts or where self-selection and missing-data issues were present. Measurement considerations are also noted: the MET may not capture all aspects of cognitive empathy, and other tasks (e.g. Reading the Mind in the Eyes Test, MASC, FERT) have produced inconsistent findings across studies. Given these caveats, the authors recommend future research employ more rigorous, standardised methodologies: placebo-controlled, properly randomised designs with consistent dosing and administration, objective measures and neuroimaging to probe neural mechanisms, and longitudinal and clinical-population studies to evaluate therapeutic potential. They also highlight the value of comparing different empathy measures and examining individual differences in subjective responses. Finally, the investigators highlight the meta-analysis' strengths—use of a single standardised task (MET) to improve comparability and inclusion of varied real-world settings—which they argue provides initial, ecologically valid evidence that classical psychedelics may selectively enhance emotional empathy and warrants further investigation.