Human brain changes after first psilocybin use

Lyons and colleagues frame the study around the question of whether single high doses of psilocybin produce enduring anatomical and functional brain changes in humans. Previous work has established that psychedelics produce robust acute effects on brain function and can produce long-term psychological changes, and preclinical studies suggest 5-HT2A receptor agonism may be linked to structural plasticity. However, long-term in vivo human neuroimaging evidence is limited and inconsistent, especially in healthy, psychedelic-naive volunteers. To address this gap, the investigators performed a multi-modal within-subjects study combining electroencephalography (EEG), functional MRI (fMRI) and diffusion tensor imaging (DTI) in healthy volunteers receiving their first-ever high-dose (25 mg) oral psilocybin. A fixed-order repeated-measures design was used: participants first received a 1 mg psilocybin 'placebo' dose and, one month later, a 25 mg dose. Brain and behavioural outcomes were assessed at baseline, during dosing (EEG) and at one month post-dosing (MRI and behavioural measures), with the aim of identifying acute-to-longer-term neural changes and their relationships to psychological outcomes such as insight and well-being.

Twenty-eight healthy, psychedelic-naive volunteers (mean age 41, SD 8.7; 43% female) participated in a fixed-order, within-subjects study. Each participant received two oral doses of psilocybin one month apart: a low-dose control (1 mg) followed by a high dose (25 mg). EEG was recorded at baseline and at 1, 2 and 4.5 hours post-dose during each session. Structural and functional MRI (including diffusion-weighted imaging and an 8-minute eyes-closed resting-state run) were acquired at baseline and at one-month after each dosing session. Behavioural measures included subjective intensity ratings during dosing, a psychological insight scale (PIS) measured sub-acutely, and the Warwick-Edinburgh Mental Wellbeing Scale (WEMWBS) assessed with reference to two-week periods. Analytical approaches varied by modality. EEG time-series were analysed using linear mixed-effects models testing Dose (1 mg vs 25 mg) by Time (0–4.5 h) effects; informational entropy was quantified as Lempel–Ziv complexity (LZc) and spectral power was examined in canonical bands. DTI analyses used ANOVA across three timepoints (baseline, one-month post-1 mg, one-month post-25 mg) to test for Timepoint interactions with axial diffusivity (AD); significant effects were followed by post-hoc t-tests and hemisphere-specific tests with Bonferroni correction. Free-water correction was applied and fractional anisotropy (FA) was additionally reported. A tract-based spatial statistics (TBSS) analysis was also run and reported as non-significant. fMRI analyses included whole-brain BOLD contrasts to emotional faces (happy, fearful, neutral), region-of-interest (amygdala) analyses and generalized psychophysiological interaction (gPPI) connectivity. Resting-state functional connectivity (RSFC) used seed-based ANOVAs for predefined seeds (parahippocampus, amygdala, vmPFC, sgACC) plus hippocampus and dACC added after peer review; network modularity and mass-univariate network metrics were also examined. Correlational and cluster-corrected matrix methods assessed relationships among contemporaneous change scores. Predictive analyses used cross-validated sensor selection for EEG–behaviour regressions and linear mediation models to test whether sub-acute psychological insight mediated the relationship between acute entropy and one-month well-being.

EEG: Robust acute changes occurred after 25 mg psilocybin but not after 1 mg. Linear mixed-effects modelling showed significant increases in cortical informational entropy (LZc) at 1 and 2 hours post-25 mg (p < 0.001), accompanied by decreases in alpha power, an increase in gamma power at 2 hours, and decreases in theta power at 1 and 2 hours. No significant EEG changes were detected under 1 mg. DTI / anatomical: ANOVA revealed significant Timepoint-by-AD effects for two tracts after the high dose: a bilateral prefrontal cortex–striatum tract (PFC–STR; F(2,48)=10.30, p=0.005) and a PFC–thalamus tract (PFC–THA; F(2,48)=9.51, p=0.008). Post-hoc t-tests showed decreases in axial diffusivity one-month post-25 mg versus one-month post-1 mg for PFC–STR (t(24) = -3.72, p = 0.006) and PFC–THA (t(24) = -3.85, p = 0.005). Hemisphere-split tests indicated decreases in both hemispheres but, after Bonferroni correction, left-hemisphere effects remained significant. Free-water correction produced consistent AD effects and revealed significant FA changes; TBSS produced no significant clusters. fMRI — emotional faces and amygdala: Whole-brain BOLD contrasts showed a large effect for fearful faces in the key contrast (one-month post-25 mg vs one-month post-1 mg), with effect size reported as d = 0.9 (95% CI [0.33, 1.34]). An inclusive whole-brain interaction between face type and Timepoint was not significant. Exploratory amygdala ROI analyses and gPPI functional connectivity results (reported in the supplement) showed left amygdala effects and gPPI findings that survived false discovery rate correction. Resting-state functional connectivity and modularity: Seed-based RSFC analyses found a pronounced increase in amygdala RSFC one-month post-25 mg compared with one-month post-1 mg (mean difference = 3.68, SE = 0.64, p < 0.0001; d = 1.2) and when comparing post-25 mg to pre-dose baseline (M diff = 1.49, SE = 0.56, p = 0.046; d = 0.6, Bonferroni-corrected). Across-network measures showed no statistically significant group-average changes in network modularity or consistent within- or between-network RSFC changes. However, when controlling for pre-25 mg well-being, a significant negative correlation was observed between modularity change and well-being change for the key contrast (r = -0.40, p = 0.04), i.e., decreased modularity correlated with improved well-being. Modularity change also correlated with aggregated AD change (PFC–STR and PFC–THA merged). Subjective and psychological outcomes: Subjective intensity showed a strong Dose-by-Time interaction (F(8,197) = 57.73, p < 0.0001), with markedly greater intensity reported after 25 mg versus baseline and versus 1 mg; 94% of participants rated the 25 mg session as the most unusual conscious state of their lives. The 1 mg session produced no significant subjective-intensity elevation. Psychological insight (PIS) showed a robust Dose-by-Time interaction (F(2,44) = 14.05, p < 0.0001), with higher insight scores after 25 mg; details on exact sub-acute scores are provided in the supplement. Well-being (WEMWBS) analyses indicated significant improvements that were exclusive to the 25 mg condition, with authors noting robust one-month well-being increases compared with the 1 mg control. Predictive and correlational findings: Acute LZc under 25 mg predicted next-day psychological insight and one-month well-being. LZc increases in posterior electrodes (P3, P4, O1) at 2 hours post-25 mg predicted next-day insight (Spearman r_s = 0.59, p = 0.006). Whole-head LZc at post-dose timepoints predicted one-month well-being change (r_s = 0.66, p = 0.006). Cross-validated analyses using all sensors produced consistent results. A mediation analysis indicated that 2-hour LZc significantly predicted next-day insight (β = 197, SE = 67, p = 0.008, normalised β = 0.55); both LZc and insight predicted one-month well-being, but the direct effect of LZc on well-being became non-significant when insight was included (β = 29.6, SE = 17.2, p = 0.10), consistent with statistical mediation. A cluster-corrected correlation matrix of contemporaneous change scores revealed a non-spurious structure among brain and behavioural changes (cluster p = 0.006). All acute and enduring effects reported were exclusive to the high-dose (25 mg) condition.

Lyons and colleagues interpret their findings as evidence that a single high-dose psilocybin session in psychedelic-naive healthy volunteers produces measurable changes in brain function and microstructure from as early as one hour to at least one month post-dosing, changes which were not observed after a 1 mg control dose. They emphasise the predictive role of acute increases in cortical signal entropy (LZc) for later mental-health-relevant outcomes: higher entropy during the acute session predicted next-day psychological insight, which in turn statistically mediated improvements in well-being at one month. On anatomical findings, the authors note decreased axial diffusivity in prefrontal–subcortical tracts one month after high-dose psilocybin and discuss this cautiously as a possible marker of microstructural plasticity. They highlight several ambiguities: axial diffusivity can change for many biological reasons (neurofibril or glial growth, myelination, axon density, membrane permeability, extracellular fluid) and can also reflect pathology in other contexts. The DTI effects were robust to free-water correction and correlated with modularity changes, lending support to their potential functional relevance, but the authors call for multi-shell diffusion acquisitions in future work to disambiguate underlying tissue processes. Functional imaging effects were more modest than the EEG and DTI findings. Reduced alpha power and increased cortical entropy during the acute state are described as robust and dose-dependent markers of the psychedelic experience. fMRI changes at one month were subtler in this healthy sample than reported in clinical trials with depressed patients; nevertheless, reductions in amygdala/salience responses to fearful faces and increases in amygdala RSFC were observed and a modularity change correlated negatively with well-being change. The authors suggest functional brain changes may be easier to detect in clinical populations with greater baseline atypicality, though they also acknowledge that assay sensitivity, metric choice and parameters could account for variability across studies. Limitations acknowledged by the study team include interpretational uncertainty around DTI metrics, the fixed-order design (1 mg then 25 mg) which leaves open the possibility of order confounds despite the functional inactivity of the 1 mg dose, and modest sample size. The authors recommend between-subject confirmatory studies, larger samples, and more advanced diffusion imaging to clarify the biological underpinnings and replicability of the observed effects. They conclude that acute entropic brain activity and sub-acute psychological insight appear important to the mechanisms linking the psychedelic experience with subsequent well-being improvements, while cautioning against over-interpretation of the microstructural findings until further evidence is available.