The relation between naturalistic use of psychedelics and perception of emotional stimuli: An event-related potential study comparing non-users and experienced users of classic psychedelics

Psychedelics such as psilocybin, LSD, mescaline and DMT produce robust alterations of perception and consciousness, and a growing body of laboratory and clinical research has examined their acute neurophysiological and subjective effects. Earlier experimental studies report that psychedelics can acutely alter behavioural and neural responses to emotional stimuli: for example, reduced recognition of sad and fearful faces, lowered amplitude of face-selective ERP components (N170) to negative faces, reduced P300 responses to negatively valenced words, and decreased amygdala reactivity and connectivity when processing negative emotional cues. Some longitudinal and cross-sectional work further suggests longer-lasting increases in positive affect and decreases in negative affect after psychedelic sessions, but these findings derive mostly from controlled clinical or laboratory settings and self-report measures that are vulnerable to bias. Orłowski and colleagues set out to test whether similar alterations in emotional reactivity are associated with naturalistic use of classic psychedelics (i.e. use outside supervised clinical or lab contexts). The present pre-registered cross-sectional EEG study compared experienced naturalistic psychedelic users (≥15 lifetime experiences) with non-users while participants performed a gender discrimination task on faces showing neutral, angry, fearful or happy expressions. The main hypotheses concerned two pre-registered ERP indices of emotional processing (N170 and P300): users were expected to show attenuated N170 and P300 amplitudes to negative expressions (fear, anger) relative to non-users. Exploratory analyses targeted earlier attention/perception-related components (N200 and P200) to provide a fuller picture of processing from early perceptual stages to later cognitive evaluation.

This pre-registered, cross-sectional EEG study recruited native Polish speakers through online surveys distributed via social media profiles related to harm reduction and drug policy. Ethical approval was obtained and participants provided informed consent. Screening excluded individuals with recent heavy use of empathogens, stimulants or dissociatives, frequent benzodiazepine/synthetic cannabinoid/opioid use, psychiatric or neurological diagnoses, current psychoactive prescriptions, or elevated scores on a shortened AUDIT. The experimental group comprised individuals reporting 15 or more lifetime experiences with classical psychedelics (LSD, psilocybin, DMT variants, mescaline, ayahuasca and derivatives); the control group comprised people who had never used psychedelics but indicated future willingness to do so. Participants were asked to abstain from psychedelic use for at least 30 days prior to EEG recording. The final analysed sample pooled data from two laboratories and consisted of 56 psychedelic users and 55 non-users; data from three additional participants were excluded for technical problems. The sample size was not precalculated. Stimuli were 104 face images (13 female and 13 male models, four expressions each: angry, fearful, happy, neutral) drawn from the Warsaw Set of Emotional Facial Expression Pictures. Each experimental session comprised 416 trials (104 per expression), presented in pseudorandom order. Participants performed a gender discrimination task (button press) while maintaining fixation; faces were displayed for 100 ms following a jittered fixation interval. Behavioural responses (accuracy and reaction time) were collected. EEG was recorded from 64 scalp electrodes at both sites (Brain Products ActiCap in Warsaw; BioSemi ActiveTwo in Cracow), with additional earlobe and ocular channels. EEG preprocessing used EEGlab-based custom MATLAB scripts: data were high-pass (0.1 Hz) and low-pass (40 Hz) filtered, re-referenced to the mean scalp signal, down-sampled to 250 Hz, and noisy channels were identified and interpolated. Continuous data were epoched from −400 ms to 1200 ms around stimulus onset. Trials with incorrect or absent responses or implausibly fast responses were removed; automatic artifact rejection (FASTER) and ICA (with MARA for component rejection) were applied before re-referencing to earlobes. The numbers of accepted epochs did not differ significantly between groups or conditions (mean ~373 epochs per participant; ~93 per condition). ERP analyses focused on two pre-registered components and two exploratory components. N170 was measured at occipitoparietal electrodes (P7, TP7, P8, TP8) in the 132–200 ms window; P300 at centroparietal sites (CP1, CPz, CP2, P1, Pz, P2) in the 320–520 ms window. Exploratory components were a frontal N200 (F1, F2, Fz, FC1, FC2, FCz; 180–280 ms) and an occipitoparietal P200 (PO3, POz, PO4, O1, Oz, O2; 180–248 ms). Single-trial mean amplitudes within these spatiotemporal windows were entered into statistical models. Statistical analysis used linear mixed-effects models (lmerTest in R) with fixed effects of Group (users, non-users) and Facial expression (angry, fearful, happy, neutral) and a random intercept for participant. The effect of data-collection site was evaluated via model comparison using Bayesian Information Criterion. Because meditation practice differed between groups and can affect emotionality, the investigators applied weights in the main models to down-weight individuals with extensive lifetime meditation hours; the weighting used a Yeo–Johnson transform, centring and inversion to produce weights, and unweighted control analyses were also reported as qualitatively similar. Post hoc tests were Holm-corrected for multiple comparisons.

Behavioural performance showed ceiling accuracy and no between-group differences in accuracy or reaction times, indicating the gender discrimination task remained easy and emotion-irrelevant for both groups. The numbers of epochs retained for analysis were comparable across groups (users: 373.38 ± 18.89; non-users: 372.96 ± 18.58) and conditions (≈93 epochs per condition). N170: A robust main effect of facial expression on N170 amplitude was observed (p < 0.001). Post hoc tests showed fearful and angry faces were associated with lower (less negative) N170 amplitudes than happy or neutral faces. The Group main effect was not significant (p = 0.238), but there was a significant Emotion × Group interaction (p = 0.017). Within the psychedelic user group, N170 amplitude was significantly lower during perception of fearful faces than during neutral (z = 3.75, p = 0.012) or happy faces (z = 3.57, p = 0.023). The non-user group showed no significant within-group differences between emotions. Direct between-group contrasts within individual emotion conditions did not reach significance. P300: Facial expression also affected P300 amplitude (p < 0.001); fearful faces elicited greater (more positive) P300 amplitude than neutral (z = 4.72, p < 0.001) or happy faces (z = 2.93, p = 0.017). Neither a main effect of Group nor an Emotion × Group interaction was significant for P300. N200 (exploratory): A main effect of facial emotion was present (p < 0.001): angry faces produced lower N200 amplitude than other expressions. There was a main effect of Group on N200 amplitude (p = 0.022), with psychedelic users exhibiting lower amplitudes overall than non-users, and a significant Emotion × Group interaction (p = 0.004). Post hoc comparisons indicated that users showed lower N200 amplitude to fearful faces compared with non-users (z = 2.79, p = 0.021). Within the non-user group, N200 amplitude was lower for angry faces compared with fearful, happy and neutral faces (all p < 0.001), whereas the user group showed no significant within-group differences across emotions. P200 (exploratory): Facial expression influenced P200 amplitude (p < 0.001), with angry faces producing larger (more positive) P200 than the other expressions. No main effect of Group or Emotion × Group interaction was found for P200. Control analyses using unweighted models reportedly produced qualitatively similar results; model selection considered inclusion of data-collection site where appropriate.

Orłowski and colleagues interpret their findings as evidence that naturalistic, repeated use of classic psychedelics is associated with attenuated neural responses to negative emotional stimuli at early, automatic stages of processing. The key ERP results supporting this view were lower N170 amplitudes to fearful faces within psychedelic users and lower frontal N200 amplitudes to fearful faces in users relative to non-users. In contrast, later ERP components linked to attentional allocation and cognitive evaluation (P200, P300) did not show group differences, suggesting the effects may be specific to early perceptual and automatic evaluation stages rather than later, controlled processes. These electrophysiological results align with prior laboratory studies that reported acute reductions in N170 amplitude and decreased amygdala reactivity to negative stimuli following psychedelic administration, and they are consonant with the authors' earlier large questionnaire study that found greater positive and lower negative self-reported emotional reactivity among psychedelic users. The investigators emphasise that their design is correlational and cannot establish causality; they note several methodological caveats. The gender discrimination task rendered emotional content task-irrelevant, unlike many acute-administration studies that employed emotion-recognition tasks, though the presence of emotion effects in the ERPs indicates the paradigm was sensitive to emotional content. Despite recruitment efforts yielding about 5000 survey responses, the groups could not be matched on lifetime meditation practice or lifetime cannabis use—both of which can influence emotional processing—so the authors applied statistical weighting to reduce the influence of heavy meditators; inclusion of weights increased the observed interaction effects but could not fully substitute for experimental matching. The study also lacked detailed data on contextual factors (set and setting, intentions) surrounding participants' psychedelic use, and self-selection bias may have favoured individuals with more positive psychedelic histories. Finally, the authors note that while questionnaire data suggested greater positive emotional reactivity in users, no ERP evidence for enhanced processing of positive emotions emerged here, and they found no behavioural differences in task performance. Taken together, the authors conclude that naturalistic psychedelic use may be linked to reduced neural reactivity to negative emotional stimuli during early stages of perception, but they caution that longitudinal and more tightly controlled studies are required to disentangle causality and to address confounding variables such as meditation and cannabis use.

The study found that experienced users of classic psychedelics, recruited from naturalistic settings, showed attenuated early neural responses to negative facial expressions (notably in N170 and N200 components) compared with non-users, while later ERP indices of cognitive evaluation (P200, P300) did not differ between groups. The authors suggest these findings are consistent with a lasting reduction in automatic negative emotional reactivity associated with psychedelic use, but they stress that the cross-sectional design and uncontrolled confounds prevent causal inference and warrant cautious interpretation.