Psilocybin induces time-dependent changes in global functional connectivity: Psi-induced changes in brain connectivity

Psilocybin (Psi) is a serotonergic psychedelic that has regained scientific interest because of promising clinical findings and its utility for probing consciousness, cognition and emotion. Molecular work shows that psilocin, Psi's active metabolite, binds to multiple serotonin receptor subtypes including 5-HT2A and 5-HT1A, and pharmacological blockade of 5-HT2A reduces Psi effects in humans. However, system-level neurobiology in humans remains incompletely characterised: prior resting-state BOLD studies used region-of-interest approaches or intravenous dosing with brief, rapidly evolving effects, limiting whole-brain and time-resolved inference about connectivity changes and their receptor-level correlates. This study aimed to map time-dependent, whole-brain changes in resting-state functional connectivity after oral psilocybin and to relate these effects to cortical gene expression of serotonin receptors. Anticipating parallels with previous LSD findings, Preller and colleagues hypothesised that psilocybin would produce a linear increase in connectivity changes from 20 to 70 minutes after administration, yielding hyper-connectivity in sensory regions and hypo-connectivity in associative networks at peak effect; they further predicted a positive spatial association with cortical HTR2A expression and a negative association with HTR1A expression. The work therefore combines a data-driven global brain connectivity (GBC) approach with time-resolved scanning and gene-expression mapping to link network-level effects to receptor topography.

Twenty-four participants were enrolled; one was excluded for an incomplete scan, leaving 23 healthy adults (12 male, 11 female; mean age 26.3 years, range 20–40; mean verbal IQ 105.48). A fully double-blind, randomised, placebo-controlled crossover design was used with two sessions separated by two weeks. On each occasion participants received either placebo (179 mg mannitol and 1 mg aerosil p.o.) or psilocybin 0.2 mg/kg p.o., and resting-state fMRI scans were acquired at three post-administration time points (T1 = 20 min, T2 = 40 min, T3 = 70 min). Neuroimaging preprocessing followed standard pipelines with an explicit note that mean gray matter signal regression (global signal regression, GSR) was applied in primary analyses to mitigate spatially pervasive artefacts such as respiration, but results are reported both with and without GSR because this step is contentious. Global Brain Connectivity (GBC) was computed as the average connectivity of each voxel/vertex to all others, providing a data-driven, whole-brain index of brain-wide integration. Second-level analyses entered subject-wise GBC maps into a 2 (Drug: Psi, Placebo) × 3 (Time: T1, T2, T3) repeated-measures ANOVA and used FSL's PALM tool with threshold-free cluster enhancement (TFCE) and 10,000 permutations to control type I error. To assess relationships between baseline connectivity and drug-induced change while avoiding mathematical coupling, Oldham's method was employed: the investigators correlated the average of Psi and Placebo scans (OH = (Psi+Pla)/2) with Psi–Placebo change scores in hyper-connected, hypo-connected and all affected regions. For receptor mapping, cortical gene-expression maps were derived from the Allen Human Brain Atlas (AHBA) for candidate receptors (HTR1A, HTR2A, HTR2C, HTR5A, HTR7), and unthresholded Z-maps of Psi>Placebo at each time point were correlated with these expression maps. Behavioural effects were measured with the 5D-ASC questionnaire and analysed with repeated-measures ANOVA; supplementary material contains further details of preprocessing and behavioural analyses.

Psilocybin produced widespread, time-dependent changes in global brain connectivity. Across time points (TFCE-corrected p<0.05, 10,000 permutations) Psi compared with placebo induced hypo-connectivity in subcortical regions and bilateral cortical associative networks including medial and lateral prefrontal cortex, cingulum, insula and temporoparietal junction. Concurrent hyper-connectivity was observed in sensory regions, most prominently bilateral occipital cortex, and additionally in right superior temporal gyrus, precuneus and left postcentral gyrus. The hypo/hyper pattern strengthened over time and was largest at T3 (70 minutes). Objective measures of the relationship between hypo- and hyper-connectivity showed strong negative correlations across participants at each time point (T1 r = -0.92, p<0.001; T2 r = -0.91, p<0.001; T3 r = -0.80, p<0.001), indicating linked, systems-level perturbations. Subjective effects assessed with the 5D-ASC increased under psilocybin: a repeated-measures ANOVA revealed main effects of Drug (F(1,22)=46.83, p<0.001) and Scale (F(10,220)=13.33, p<0.001) and a Drug×Scale interaction (F(10,220)=12.08, p<0.001). Bonferroni-corrected comparisons showed higher scores on all 5D-ASC scales for Psi versus placebo at T2 and T3 (all p<0.05); differences at T1 were not significant. Analyses of intra-individual stability indicated that connectivity maps in the placebo condition were highly correlated across time (mean r≈0.58), while intra-individual correlations between Psi time points decreased from T1→T2→T3 (highest between T1 and T2 r≈0.49; lowest between T1 and T3 r≈0.36), consistent with evolving drug effects. Using Oldham's method, baseline GBC (OH) showed no significant association with Psi–Placebo change at T1 and nonsignificant effects at T2 after correction, but at T3 OH correlated with Psi–Placebo change scores in both hyper- and hypo-connected areas (p<0.05, Bonferroni corrected), and absolute changes across all affected areas correlated positively with OH, indicating that baseline connectivity predicted the magnitude of peak psilocybin-induced alterations. Correlations with cortical gene-expression maps showed time-dependent relationships: psilocybin-induced GBC changes correlated positively with HTR2A expression and negatively with HTR1A expression, and a strong negative correlation with HTR7 expression was also observed. The difference in correlation strength between HTR2A and HTR1A increased over time (Δr T1 = 0.36; T3 = 0.75). Individual-level correlations between each participant’s Psi>Placebo map and HTR2A/HTR1A maps confirmed a significantly larger HTR2A–HTR1A difference at T3 versus T1 (p<0.05, Bonferroni corrected). Results were robust to Spearman rank correlations and were reported both with and without GSR (supplementary figures). The extracted text notes that HTR2C showed low correlations that increased over time, and cautions that inferring receptor densities from RNA maps requires care because of regional cytoarchitectural factors.

Preller and colleagues interpret their findings as filling several gaps in the human neurobiology and neuropharmacology of psilocybin. First, the data-driven GBC approach revealed a consistent pattern across individuals and time: increased brain-wide connectivity in sensory networks alongside decreased connectivity in associative networks. The investigators note that this pattern mirrors results previously reported after LSD and may represent a common systems-level signature of the psychedelic state, wherein heightened sensory synchronisation and reduced associative integration could underlie characteristic subjective experiences. Second, time-resolved scanning demonstrated that both subjective effects and connectivity changes evolve from 20 to 70 minutes after oral administration, with the strongest effects at 70 minutes; this underscores the importance of measurement timing and complements previous intravenous studies that capture rapid bolus effects. Third, intra-individual analyses showed that psilocybin-induced neural effects are more stable between scans closer in time, matching the temporal profile of subjective symptoms. Fourth, baseline GBC predicted the magnitude of peak connectivity changes, suggesting that pre-drug network organisation could serve as a marker for individual sensitivity to psilocybin and potentially inform personalised approaches to psychedelic therapy. Fifth, spatial correlations between psilocybin-induced GBC changes and cortical gene-expression maps implicate the 5-HT2A receptor system (positive association) and 5-HT1A system (negative association) as primary molecular contributors to observed network effects, with an additional negative relationship with HTR7 expression noted; the authors caution that gene-expression maps are an indirect proxy for receptor density and that regional protein–RNA relationships vary. Limitations acknowledged include modest sample size for correlational analyses (which likely reduced power to link subjective measures and connectivity after correction), sensitivity of some findings to global signal regression (analyses are reported both ways), and uncertainty about the functional role of HTR7 in psychedelic states. The authors recommend future studies addressing dose–response, intravenous versus oral administration comparisons, later time points during recovery, and direct molecular imaging to strengthen receptor-level inferences.

The study concludes that oral psilocybin produces time-dependent, whole-brain changes in global connectivity characterised by sensory network synchronisation and associative network disintegration, with effects peaking around 70 minutes. Baseline GBC predicted individual differences in response magnitude, and spatial correspondence between connectivity changes and cortical HTR2A and HTR1A expression supports a central role for these receptor systems in psilocybin’s action. The authors suggest these findings advance mechanistic understanding and point to potential targets and markers for the development and individualisation of psychedelic-based therapeutics.