EEG microstate dynamics during psilocybin intoxication relate to acute experience and persisting psychological changes

The authors note that psilocybin and other serotonergic psychedelics have renewed interest as potential treatments for psychiatric disorders, with clinical trials suggesting rapid and sometimes sustained benefits after only one or two administrations. However, they also emphasise that the neural mechanisms linking the acute psychedelic state to later psychological change remain incompletely understood. Earlier research has shown that the subjective quality of the psychedelic experience can predict longer-term outcomes, but relying on self-report alone limits mechanistic inference. At the neural level, prior work has focused more on spectral power, entropy, and large-scale network changes than on the fast spatiotemporal organisation of EEG activity. Against this background, the study set out to examine whether resting-state EEG microstate dynamics during psilocybin intoxication are altered in healthy volunteers, and whether those acute changes relate to both subjective experience and persisting psychological changes 28 days later. The researchers aimed to characterise psilocybin’s effects on microstate temporal organisation and to test whether individual differences in those dynamics track acute phenomenology and later self-reported change. They present this as an exploratory attempt to evaluate EEG microstates as candidate neural markers of psychedelic-induced changes in consciousness with possible therapeutic relevance.

The study analysed data from a double-blind, randomised, placebo-controlled crossover trial in healthy volunteers. Participants were screened carefully for neurological, psychiatric, substance-use, and major medical disorders; those with personal or first-degree family histories of psychotic illness were excluded. For the present analysis, the authors used the first 20 enrolled participants, but 5 were excluded because of missing sessions or poor-quality EEG, leaving 15 participants in the final dataset. The sessions were separated by at least 28 days. Participants received a weight-adjusted oral dose of pharmaceutical-grade psilocybin of approximately 0.26 mg/kg, or visually identical placebo capsules containing wheat starch only. Each person completed two experimental days, one with psilocybin and one with placebo, in counterbalanced order. During the sessions, they rested in a reclined position under supervision. EEG was recorded at five time points: a 10-minute pre-dose baseline and four 10-minute recordings at 50-60, 90-100, 180-190, and 360-370 minutes after ingestion, covering peak intoxication and recovery. Blood samples for psilocin levels were collected at baseline and at 1, 2, 4, and 6 hours after dosing. Subjective experience was measured with the Altered States of Consciousness questionnaire at the end of each session, acute clinical symptoms were rated repeatedly with the Brief Psychiatric Rating Scale during the session, and persisting effects were assessed 28 days later with the Persisting Effects Questionnaire. EEG was recorded from a 21-channel cap, sampled at 1000 Hz, and preprocessed using filtering, independent component analysis, artefact rejection, average re-referencing, and downsampling. The cleaned data were segmented into 2-second epochs, yielding 40 seconds of clean signal per participant per condition per time point. Microstate analysis was carried out in Python using GFP peaks, which identify time points with the highest signal-to-noise ratio, followed by hierarchical clustering to derive microstate maps. The authors analysed both a three-microstate solution and the canonical four-microstate solution, and repeated all analyses in two frequency bands: 1-40 Hz and 2-20 Hz. The three-microstate 1-40 Hz solution was treated as primary because the cross-validation criterion suggested three states, while the 2-20 Hz four-state solution was kept as a secondary, more traditional comparison. Outcome measures included microstate lifespan, coverage, frequency of occurrence, and transition probabilities, although transition probabilities were calculated for completeness and not analysed statistically. Repeated-measures ANOVA with Benjamini-Hochberg false discovery rate correction was used for GFP peak counts and microstate features, followed by paired t-tests where appropriate. Relationships between microstate measures and experience data were examined using Spearman correlations with false discovery rate correction. To reduce the number of correlations, the researchers aggregated experience variables and microstate features using principal component analysis and correlated the first principal component from each domain.

The psilocybin condition significantly increased GFP peak density compared with placebo, especially during peak intoxication. In the primary 1-40 Hz analysis, condition, time, and their interaction were significant, and the number of GFP peaks was higher at T2, T3, and T4 under psilocybin. In the 2-20 Hz analysis, condition was also significant, and the T3 psilocybin condition showed a higher number of GFP peaks than placebo. Cross-validation indicated that three microstates was the optimal solution across conditions and time points, although the authors also reported the canonical four-state solution. The canonical microstate maps resembled the expected template maps in both frequency bands, with group-level mean correlations above 0.8. For explained variance, the 2-20 Hz four-state solution showed no significant time effect, but in the 1-40 Hz analysis condition was significant: explained variance was slightly higher in placebo than in psilocybin, with an unbiased Cohen’s d of 0.414. The main temporal microstate effects were reductions in lifespan and increases in frequency of occurrence during psilocybin intoxication, most clearly at T2 and T3, when the subjective effects were strongest. These changes were seen in both frequency-band analyses, with strong effect sizes for some microstate-specific comparisons. Coverage was comparatively stable: no significant differences were found in the 1-40 Hz analysis, and in the 2-20 Hz analysis there was a single significant difference at T2. Overall, the authors describe coverage as largely preserved despite altered timing dynamics. Subjectively, all four Altered States of Consciousness dimensions were significantly elevated under psilocybin relative to placebo. Brief Psychiatric Rating Scale scores also increased during intoxication, particularly for thought disturbance and hallucinations, withdrawal and retardation, tension and excitement, and hostile suspiciousness at 70 minutes; anxiety and depression did not differ significantly. The authors also report that these subjective and observer-rated effects aligned with blood psilocin levels and dose, with correlations showing coherent clustering among the psychometric and pharmacokinetic measures. Finally, microstate measures during peak intoxication correlated with experiential outcomes. The aggregated microstate component was associated with acute subjective experience, and at T2 it also correlated with positive persisting effects on the Persisting Effects Questionnaire 28 days later. Negative persisting effects were low overall, with mean scores below 10% of the scale maximum, and were not significantly associated with microstate dynamics.

The authors interpret the findings as showing that psilocybin accelerates and destabilises resting-state EEG microstate dynamics during peak intoxication. They argue that the increased number of GFP peaks and the shortened microstate lifespan indicate more rapid switching between global brain states, consistent with earlier work showing that psychedelics increase neural entropy, signal diversity, and large-scale network reorganisation. In their view, microstate peak density may serve as a topography-based complement to complexity metrics such as Lempel-Ziv complexity. They emphasise that the temporal profile of these changes followed psilocybin pharmacokinetics, peaking at T2 and T3 alongside peak plasma psilocin levels and maximal subjective effects. The increased frequency of occurrence is presented as a necessary counterpart to shorter lifespans and as further evidence of faster state transitions. The authors link these observations to broader theories suggesting that psychedelics increase the rate at which the brain explores its state space. By contrast, microstate coverage was largely unchanged, which the authors interpret as suggesting that psilocybin mainly alters the sequencing and stability of brain states rather than the repertoire of states available. They note that this pattern may distinguish the psychedelic state from psychiatric conditions such as schizophrenia and depression, where altered coverage has been reported, although they caution that their small sample limits sensitivity to smaller coverage effects. The authors also see translational interest in the correlations between acute microstate dynamics, subjective intensity, and persisting positive effects 28 days later. They suggest that greater microstate destabilisation during the acute state was associated with stronger later changes in mood, well-being, and personal meaning. However, they stress that these are correlational findings and do not establish causality or therapeutic efficacy. They say that validation in clinical populations is still required before any treatment-related inference can be drawn. In discussing limitations, the authors highlight the small sample size (N=15), which reduces power for time-resolved and smaller effects, particularly in the secondary analysis. They also raise an alternative explanation for the EEG findings: psilocybin-related increases in heart rate, pupil dilation, and sympathetic arousal could potentially contribute to faster EEG fluctuations. Because physiological signals were not simultaneously recorded, they could not rule this out. The authors further state that future work should include concurrent physiological monitoring, clinical samples, multimodal imaging, pharmacological manipulations, and computational modelling to clarify what microstates reflect mechanistically. They conclude that microstate measures are promising pragmatic biomarkers of altered brain dynamics, but their functional interpretation remains incomplete.

The authors conclude that psilocybin markedly reorganised resting-state EEG microstate dynamics, increasing GFP peak density and microstate switching while shortening microstate lifespan, whereas microstate coverage was largely preserved. They state that these findings are consistent with an acceleration and destabilisation of large-scale brain dynamics rather than a major change in the set of accessible brain states. They also conclude that acute microstate differences were associated with both subjective psychedelic experience and persisting positive psychological effects, suggesting potential value as neural markers of psychedelic-associated changes in consciousness, while noting that mechanistic interpretation remains limited.