Naturalistic use of psychedelics does not modulate processing of self-related stimuli (but it might modulate attentional mechanisms): An event-related potentials study

Classic psychedelics (for example psilocybin, LSD, mescaline, DMT) are known to acutely alter consciousness and can induce ego dissolution, a transient blurring of boundaries between self and world. Neuroimaging work has linked acute ego-dissolution to reduced activation and altered connectivity within the default mode network and disruptions in salience-network function. While laboratory studies typically probe acute or short-term effects after a single controlled dose, naturalistic use is characterised by repetitive intake in uncontrolled settings, and it remains unclear whether repeated naturalistic use produces longer-lasting changes in self-processing. Orłowski and colleagues conducted a preregistered, cross-sectional EEG study to test whether experienced naturalistic users of classic psychedelics show persistent alterations in neural processing of self-related stimuli. The investigators compared experienced users (≥15 lifetime psychedelic experiences; N = 56) to non-users (N = 57) using each participant's own name (Self), an Other-name (passive control) and a Target-name (active control requiring a button press). The pre-registered primary outcome was P300 amplitude; an exploratory analysis examined stimulus-induced alpha-band desynchronisation. The a priori hypothesis was that users would show reduced P300 (and weaker alpha desynchronisation) to the Self-name, with no group differences for Other or Target stimuli.

This was a preregistered, cross-sectional EEG study recruiting participants via an online survey disseminated through Polish harm-reduction and psychedelics-related social media channels. Inclusion criteria for the users group required at least 15 lifetime experiences with classic psychedelics (explicitly excluding microdosing), while the non-users group comprised people without prior psychedelic use who reported an intention to use in the future. Participants were matched for gender, age, education and alcohol use across two data-collection sites (Warsaw and Cracow). The final sample comprised 56 users and 57 non-users; participants were asked to abstain from psychedelics for at least 30 days before testing. The extracted text does not clearly report additional exclusion criteria or clinical screening beyond these recruitment details. Stimuli and task: For each participant three names were used: Self, Other (gender- and length-matched), and Target (another Other-name). Each session presented 300 trials (100 per condition) in pseudo-random order. On Target trials participants pressed a response button with their right index finger; no response was required for Self or Other trials. Trials began with a fixation cross (700–1200 ms), followed by a 100 ms name presentation and a 1200 ms blank screen. A brief training of 8 trials preceded the main task; a mid-task break was permitted. The entire task lasted about 12 minutes. EEG recording and preprocessing: EEG was recorded from 64 channels at both sites (Brain Products ActiCap at 1000 Hz in Warsaw; BioSemi ActiveTwo at 2048 Hz in Cracow). Signals from earlobes and four EOG channels were also acquired. Preprocessing used EEGLAB-based custom scripts: data were band-pass filtered (0.1–40 Hz), re-referenced to the average of scalp electrodes, down-sampled to 250 Hz, and noisy channels were detected and interpolated (mean ~2.4 channels removed). Continuous data were epoched from −400 to 1200 ms around stimulus onset. Trials with misses, false positives or implausibly fast responses were excluded; automatic artifact rejection (FASTER) removed additional epochs (mean ~12.8). Independent component analysis (ICA) was applied (on 1 Hz high-pass filtered data) and artifact components were removed using MARA (mean ~40 components removed). Final signals were re-referenced to the earlobe mean. The number of retained epochs did not differ between groups or conditions (overall mean ~283 epochs per participant). ERP and time–frequency analyses: The pre-registered P300 analysis defined a centro-parietal electrode cluster (CP1, CPz, CP2, P1, Pz, P2) and a 200 ms time window centred on the grand-average peak, yielding 268–468 ms post-stimulus; single-trial mean amplitudes over this spatio-temporal window were used in statistical models. Alpha-band desynchronisation was an exploratory analysis: Morlet wavelets (3–30 Hz) produced time–frequency maps, from which the strongest alpha desynchronisation was identified in the 8–18 Hz band between 252–500 ms over 17 occipito-parietal electrodes; power values baseline-corrected to −200–0 ms were extracted for analysis. Statistical approach: Linear mixed models were fit in R (lmerTest), with participant as a random effect and fixed effects of Group (users, non-users) and Identity (Self, Other, Target). To mitigate confounding from meditation practice and lifetime cannabis use, the authors constructed weights based on transformed lifetime meditation hours and cannabis exposures (Yeo–Johnson transform, centred, inverted, then multiplied to yield a composite weight); weighted models are reported as primary, with unweighted models reported as qualitatively similar in supplementary material. Models with and without data-collection site as a fixed effect were compared using BIC; post-hoc tests were Holm-corrected. The data and analysis scripts were made available on OSF according to the authors.

Sample and data quality: The final sample included 56 psychedelic users and 57 non-users. After preprocessing the average number of epochs retained per participant was similar across groups (users: 283.14 ± 6.07; non-users: 283.86 ± 6.59) and across conditions (Self ~94, Other ~94, Target ~95), indicating comparable data quantity between groups. P300 results: A strong main effect of Identity was observed for P300 amplitude (F(2,31919) = 217.15; p < 0.001). Post-hoc comparisons showed the expected ordering: Target evoked the largest P300, Self was intermediate, and Other was smallest (Target > Self > Other; all pairwise p < 0.001). A main effect of Group indicated that users exhibited higher overall P300 amplitude than non-users (F(1,111) = 5.26; p = 0.024). Crucially, there was a significant Identity × Group interaction (F(2,31919) = 10.51; p < 0.001). Within each group the same pattern (Other < Self < Target) held with significant pairwise contrasts. Between-group contrasts by condition: In the Self condition users showed a numerically higher P300 but this difference did not reach significance after correction (z = 2.36, p = 0.055). For the Other condition users exhibited significantly larger P300 amplitudes than non-users (z = 2.91, p = 0.018). No significant group difference was found for the Target condition (z = 1.50, p = 0.151). Comparing between-condition differences across groups, non-users displayed a larger amplitude difference between Self and Target (z = −2.76, p = 0.023) and a larger Other–Target difference (z = 4.55, p < 0.001) than users. In other words, the increase in P300 for task-relevant Target stimuli (relative to passive Self/Other) was greater in non-users than in users. Alpha desynchronisation results: Identity significantly affected alpha desynchronisation strength (F(2,31919) = 94.51; p < 0.001), driven by stronger desynchronisation for Target than for Self (z = 11.57, p < 0.001) or Other (z = 12.20, p < 0.001). No difference emerged between Self and Other (z = 0.63, p = 0.527). There was no main effect of Group (F(1,111) = 0.02; p = 0.900) and no Identity × Group interaction (F(2,31919) = 1.72; p = 0.180). Thus the exploratory alpha measure did not provide evidence of altered self-related processing in users. The authors note that results from weighted and unweighted models were qualitatively similar.

Orłowski and colleagues interpret their findings as indicating that repeated naturalistic use of classic psychedelics is not associated with long-term attenuation of neural responsiveness to a robust self-related stimulus (own name). Both groups showed the canonical P300 ordering (Target > Self > Other), and the pre-registered hypothesis—reduced P300 to Self in users—was not supported. The exploratory alpha analysis likewise failed to show weaker self-related desynchronisation in users. The authors highlight an unexpected group-level difference in attentional modulation: non-users showed a larger increase in P300 amplitude for task-relevant Targets relative to passive Self/Other stimuli, whereas users showed a smaller Target-related increase. Two interpretations are advanced. One possibility is subtle, sustained differences in attentional mechanisms among users, potentially indicating reduced phasic allocation of attentional resources to task-relevant stimuli; however, the authors note that behavioural performance (accuracy and reaction time) did not differ between groups, and no prior evidence indicates chronic attentional deficits in psychedelic users. An alternative, speculative interpretation is that regular psychedelic use may produce a more even distribution or sustained allocation of attention across stimuli—an "attention equalizing" effect—leading to smaller relative increases for highly task-relevant items. The observed larger P300 to Other-names in users might reflect increased salience attribution to stimuli unrelated to the self, although this remains tentative. Several methodological limitations are acknowledged. The cross-sectional design precludes causal inference; unmeasured temperament or personality factors could explain group differences. The investigators were unable to match groups fully for lifetime meditation hours and cannabis use, both of which can affect self-related experience; to mitigate this they used statistical weighting but residual confounding may remain. The study did not obtain dose or direct ego-dissolution experience measures for each psychedelic episode, which constrains interpretation about whether participants actually experienced ego-dissolution. No pre-study power calculation was performed due to lack of prior effect-size estimates, although the sample of 113 is relatively large for an EEG study and the presence of significant group effects for some contrasts suggests adequate sensitivity. The authors also note that their focus on the P300 may miss frontal self-related activations commonly reported in fMRI studies; they therefore recommend future work using source localisation or multimodal EEG–fMRI to probe frontal contributions. Finally, self-selection bias may have led to overrepresentation of users with predominantly positive experiences. The authors propose several directions for future research: longitudinal or experimental designs to address causality, inclusion of personality and cognitive assessments, more detailed measures of psychedelic dose and subjective experiences (including ego-dissolution), matched groups for meditation and cannabis use, and replication of the attentional findings using dedicated cognitive paradigms and neural source-modelling. Overall, they emphasise that while acute psychedelic states produce robust self-boundary changes, their data do not support durable alterations in neural processing of an explicit self-related stimulus among frequent naturalistic users; instead, the primary novel signal concerns attentional modulation and warrants replication.

The authors conclude that repeated naturalistic use of classic psychedelics does not appear to produce persistent alterations in neural processing of a self-related stimulus (own name), as indexed by P300 amplitude and alpha desynchronisation. However, the study provides preliminary evidence of altered attentional mechanisms in users, manifested as a smaller P300 increase to task-relevant Target stimuli and larger P300 to Other-names. This unexpected finding requires replication and more targeted investigation. The authors stress the limitations of cross-sectional inference and call for longitudinal and mechanistic studies to clarify whether and how naturalistic psychedelic use affects attention and self-related processing.