Psilocybin acutely alters the functional connectivity of the claustrum with brain networks that support perception, memory, and attention

Classic serotonergic psychedelic drugs such as psilocybin acutely alter sensory perception and executive function and are being revisited for potential therapeutic value across neuropsychiatric disorders. Previous imaging studies indicate that psilocybin perturbs the integrity and coupling of large-scale networks such as the default mode network (DMN) and task-positive/executive networks, but the circuit-level mechanisms underlying these effects remain unclear. The claustrum, a subcortical telencephalic nucleus with high expression of 5-HT2A receptors and widespread glutamatergic projections to cortex, is a plausible target mediating psilocybin’s broad cortical effects; prior human imaging has linked the claustrum to task onset and network switching distinct from neighbouring structures like the insula and putamen. Barrett and colleagues therefore tested the hypothesis that psilocybin modulates claustrum function in humans. Using resting-state fMRI in a blinded, within-subject design, the study compared measures of claustrum activity (variance and amplitude of low-frequency BOLD fluctuations) and claustrum functional connectivity with canonical cortical networks after oral placebo and after a moderate dose of psilocybin (10 mg/70 kg). The investigators also examined relationships between claustrum metrics, network integrity/modularity, and contemporaneous subjective effects.

Design and participants. This was a blinded, placebo-controlled, within-subject study in which each participant underwent two resting-state fMRI sessions: one following placebo and one following 10 mg/70 kg oral psilocybin. Twenty healthy volunteers were enrolled, but five were excluded from analysis because of excessive head motion or normalization failures during one or both scans, leaving 15 participants (10 male, 5 female) in the final analysis. The extracted text reports a mean age of 51.3 years but does not clearly report the age range. All participants were medically and psychologically screened, reported long-term meditation practice, and had prior exposure to a high psilocybin dose in an earlier study. Exclusion criteria included clinically relevant cardiac abnormalities, family history of bipolar disorder or psychosis, substance use disorder, or recent psychedelic use. Procedures and subjective ratings. Participants received standard psychological support (preparation, monitoring during sessions, debrief) and were instructed not to meditate during scans. Scans began approximately 100 minutes after capsule administration, coinciding with expected peak subjective effects. The order was fixed: placebo first (8:00) and psilocybin second (12:00). Immediately after each resting-state scan participants rated subjective effects on an 11-point verbal scale (0–10), covering overall drug strength, several mindfulness-related items (now-ness, letting go, equanimity), mystical-experience items (e.g. ineffability, timelessness), and positive/negative emotional valence. MRI acquisition and preprocessing. Imaging used 3T Philips Achieva scanners. Resting-state EPI parameters included 3 mm isotropic voxels, TR 2.2 s, 210 volumes retained (~7 min 42 s analysed). A T1-weighted anatomical MPRAGE was also acquired. Preprocessing employed SPM12 and the CONN toolbox and included slice timing correction, realignment, coregistration, segmentation, normalization to MNI space, nuisance regression (motion regressors, scrubbing, CSF and white-matter principal components), detrending, despiking, and bandpass filtering (0.008–0.09 Hz). Global signal regression was not performed. Five participants were excluded for motion (>10% volumes flagged) or normalization errors; the remaining 15 did not differ significantly in motion between conditions. Claustrum isolation and ROI approach. The authors extracted 264 ROI time courses from the Power atlas (10 mm spheres) from unsmoothed whole-brain BOLD data. To isolate claustrum signal from nearby insula and putamen sources, a small-region confound correction (SRCC) was applied: mean claustrum timeseries were regressed on ipsilateral flanking ROI timeseries and nuisance regressors, and the residuals served as corrected left and right claustrum seed timecourses. Group-averaged, hand-drawn claustrum ROI templates (from high-resolution 7T data) were used and visually inspected for fit. Analyses. Subjective ratings were compared between drug conditions with paired t-tests and Bonferroni correction. For neural measures, the amplitude of low-frequency fluctuations (ALFF) and the variance of each claustrum ROI timeseries were computed and compared with paired t-tests. Functional connectivity was assessed by computing Pearson correlations (Fisher-transformed) between each claustrum seed and each Power-atlas ROI; these edges were submitted to mixed-effects general linear models with subject and ROI as random effects to test main effects and interactions of drug condition and network label. Planned network-wise comparisons focused on networks where claustrum connectivity changed significantly. Network integrity was defined as average within-network Fisher-transformed correlations; modularity was operationalised as the inverse of the participation coefficient (participation coefficient reflects inter-network connectivity of a node). Separate mixed-effects models tested associations between claustrum connectivity and network integrity/modularity. Specificity analyses compared claustrum findings to flanking insula and putamen ROIs using paired tests and mixed-effects ANOVAs. Relationships between subjective ratings and ALFF/variance in claustrum, insula, and putamen were tested with general linear models, first modelling overall strength of drug effect and then adding other subjective items with overall strength as a covariate. Multiple comparisons and family-wise error were controlled with Holm-Bonferroni adjustments.

Subjective effects and blinding. Participants reported substantially greater subjective effects during the psilocybin scan than during placebo (mean strength during psilocybin M = 8.1, SD = 2.31 versus placebo M = 1.35, SD = 2.08). Although more than half the participants gave non-zero strength ratings for the placebo scan and some believed they had received a low psilocybin dose then, group-level differences confirm robust subjective effects of the active dose. Claustrum activity (ALFF and variance). Psilocybin significantly decreased both the amplitude of low-frequency fluctuations (ALFF) and the variance of the BOLD signal in the left and right claustrum compared with placebo. These reductions were reported as specific to the claustrum when contrasted with flanking ROIs (left and right insula and putamen); paired tests for flanking regions did not show the same pattern. Claustrum functional connectivity. Mixed-effects models revealed main effects of network and significant interactions of drug condition with network label for both left and right claustrum. Network-wise planned comparisons showed that psilocybin: decreased left claustrum connectivity with the fronto-parietal task control network (FPTC); decreased right claustrum connectivity with the auditory network and with the DMN; and increased right claustrum connectivity with the FPTC. Post-hoc contrasts indicated that the psilocybin-induced change in left claustrum–FPTC connectivity was significantly larger than the corresponding change for the left insula and left putamen. For the DMN, the change associated with the right claustrum differed significantly from that of the right insula; comparisons involving the right putamen were unclear in the extracted text (the reported statistics appear inconsistent), so the precise relation to the putamen cannot be definitively reconstructed from the provided material. Other planned comparisons (e.g. right claustrum vs right insula/putamen for FPTC or auditory networks) were non-significant. Associations with network integrity and modularity. Right claustrum connectivity with the DMN was positively associated with DMN integrity when modelling both drug conditions together and when modelling psilocybin scans alone. Right claustrum connectivity with the FPTC showed a negative association with FPTC integrity when modelling both conditions, but this relationship was not significant when restricted to psilocybin scans only. Left claustrum connectivity with the FPTC was negatively associated with FPTC modularity (inverse participation coefficient) both across conditions and within psilocybin scans. No other significant associations between claustrum connectivity and network integrity/modularity were observed. Associations between subjective and neural effects. Regression analyses identified a significant negative association between ratings of overall drug strength and ALFF in the claustrum (t = 3.269, p[uncorrected] = 0.006, adjusted R2 = 0.409). No subjective–neural associations survived correction for multiple comparisons, but observed relationships were reported as modest-to-moderate and specific to the claustrum versus the insula or putamen. No serious adverse events were reported; no pharmacological, medical, or psychological interventions were needed.

Barrett and colleagues interpret their findings as evidence that psilocybin acutely reduces measures of claustrum activity and alters claustrum functional connectivity with networks supporting perception, memory, and attention. Reduced ALFF and variance in both claustral hemispheres occurred alongside network-specific connectivity changes: decreased right claustrum coupling with the DMN and auditory networks, increased right claustrum coupling with the fronto-parietal task control network, and decreased left claustrum coupling with the fronto-parietal network. The authors suggest these effects are consistent with a role for 5-HT2A receptor signalling in modulating claustrum function and, in turn, influencing the integrity and modularity of large-scale cortical networks during the psychedelic state. In relation to prior work, the study links claustrum changes to phenomena commonly reported under psychedelics—altered sensory experience (visual and auditory changes, synesthesia) and disruptions in attention or executive function, including ineffability. Right claustrum connectivity with the DMN tracked measures of DMN integrity, while left claustrum connectivity related to modularity of the fronto-parietal control network, supporting the idea that claustral modulation may contribute to psilocybin-induced network reorganisation. The authors acknowledge several limitations. Drug order was fixed with placebo always first, which reduces carry-over but introduces potential order or diurnal confounds; the sample was small, highly selected (long-term meditators, prior high-dose psilocybin exposure, older average age), and only a single, moderate psilocybin dose was tested. These factors limit generalisability and statistical power to detect brain–subjective associations. The extracted text also contained at least one ambiguous post-hoc comparison involving the putamen, which the authors did not resolve in the provided material. For future work they recommend task-based fMRI during psychedelic administration, exploration of laterality effects, comparison of psychedelic-naïve versus experienced participants, and larger pooled or multicentre datasets amenable to clustering or machine learning to characterise individual differences in claustrum connectivity.

The study provides pharmacological evidence that 5-HT2A receptor signalling significantly influences human claustrum function. Psilocybin at a moderate dose reduced claustrum activity and altered claustrum connectivity with networks implicated in sensory perception and cognitive control, and some claustrum changes related to contemporaneous subjective experiences. Given the claustrum’s broad cortical connectivity and reported structural differences in conditions such as depression and schizophrenia, the authors conclude that the claustrum warrants further investigation as a potential mediator of both the subjective and therapeutic effects of psychedelics.