Increased global functional connectivity correlates with LSD-induced ego dissolution

Tagliazucchi and colleagues situate their work within a long history of psychedelic research and note that, despite LSD's central place in that literature, there were no modern human neuroimaging studies of its acute effects. Earlier studies and theoretical work implicate serotonergic mechanisms—particularly the serotonin 2A (5-HT2A) receptor—in psychedelic effects, and prior imaging with related compounds (for example psilocybin) suggested altered functional connectivity of high-level association networks such as the default-mode, salience, and frontoparietal networks. The authors therefore identified a gap: the need for whole-brain, contemporary fMRI data on how LSD affects intrinsic functional organisation and how those changes relate to subjective experiences such as "ego dissolution." This study set out to characterise LSD-induced changes in global and local functional connectivity and to test whether changes in connectivity of high-level cortical areas correlate with participants' reports of ego dissolution. Using both exploratory whole-brain measures and hypothesis-driven analyses, the investigators aimed to map regions showing increased global coupling, assess how LSD affects modular organisation and hub connectivity, and relate these neurophysiological changes to receptor distributions (5-HT2A) and to subjective ratings obtained under LSD versus placebo. The work is presented as a first modern fMRI investigation of acute LSD effects in humans and as a test of hypotheses derived from prior psychedelic imaging and receptor biology.

The study employed a randomised, balanced within-subject design in which 15 healthy volunteers underwent intravenous LSD and placebo fMRI sessions; the extracted text indicates that a stricter motion-cleaning procedure reduced an original cohort of 20 to 15 datasets. Functional MRI data were analysed at rest. Subjective effects were quantified using visual analogue scales (VAS) including an item measuring "ego dissolution," and the altered state of consciousness (ASC) questionnaire is also mentioned. The investigators compared LSD versus placebo contrasts (often computed as LSD minus placebo) and correlated regional connectivity changes with the magnitude of subjective reports. For whole-brain connectivity, the brain was parcellated into 401 even-sized regions of interest (ROIs) covering cortical and subcortical grey matter. Functional connectivity density (FCD) was computed for each ROI as the average correlation between that ROI's BOLD time series and all other ROIs; high FCD therefore denotes regions strongly correlated with the rest of the brain. The authors performed permutation testing and multiple-comparison control (false-discovery rate, FDR) for statistical inference and used Bonferroni correction for some network-overlap tests. Seed-based regression analyses were performed using four seeds derived from the FCD difference map (a frontal seed, a parietal seed centred on the temporo-parietal junction/angular gyrus, the precuneus, and bilateral thalamus) to identify regions that became more coupled with these hubs under LSD. Network-level analyses assessed modularity (a measure of how well the brain parcellates into modules with dense within-module and sparse between-module connectivity) across a range of link densities, participation coefficient (quantifying how much a node communicates across modules versus within its own module), and the rich-club coefficient F(k) (measuring preferential connectivity among high-degree hub nodes, normalised by degree-preserving randomisation). The investigators also tested spatial overlap between FCD increases and independent resting-state network (RSN) maps derived from the Human Connectome Project, compared FCD increases with 5-HT2A receptor distribution maps from PET, and re-analysed previously acquired psilocybin fMRI data for comparison. Details on data acquisition, preprocessing and motion control are reported to be in the Supplemental Experimental Procedures.

Global functional connectivity increased under LSD. Using the 401-ROI FCD measure, Tagliazucchi et al. found a distributional shift under LSD characterised by a tail of highly coupled regions not present under placebo; overall FCD values were higher in the LSD condition. Spatially, FCD increases were observed predominantly in frontal, parietal and inferior temporal cortices, and in the bilateral thalamus. These FCD increases overlapped significantly with three high-level RSNs—the bilateral frontoparietal, default-mode, and salience networks—derived from independent component analysis of HCP data. A significant spatial overlap was also reported between the regions of increased FCD under LSD and the cortical distribution of 5-HT2A receptors (from PET), whereas no significant overlap was found with 5-HT1A or 5-HT1B receptor maps. Re-analysis of prior psilocybin data revealed FCD increases in similar regions, and LSD FCD increases also showed spatial correspondence with those psilocybin effects. Subjective reports of ego dissolution were selectively associated with regional FCD increases. Across ROIs, the only VAS item that survived multiple-comparison correction when correlated with regional FCD increases (LSD minus placebo) was ego dissolution; regions that survived correction included bilateral temporo-parietal junctions (angular gyri) and bilateral insular cortex. The authors further identified a subset of regions whose FCD correlated with ego dissolution and did not correlate with other VAS scores (reported at p < 0.05 uncorrected). Example scatterplots for left/right angular gyrus and insula illustrate the positive association between FCD and ego-dissolution ratings. Seed-based analyses showed that the regions with increased global connectivity under LSD became more engaged with sensory cortices. Regression using the four FCD-derived seeds revealed increased coupling with sensory areas; permutation testing indicated a significant overlap between these seed-derived difference maps and four HCP-derived RSNs: sensorimotor, auditory, and two visual RSNs (medial and lateral). Network topology changed under LSD in ways that indicate greater between-module integration and reduced hub-to-hub cohesion. Modularity (the separability of intrinsic brain modules) decreased under LSD across a range of network link densities, consistent with reduced segregation of networks. Participation coefficients increased in frontal and midline regions, indicating greater cross-module communication by those nodes. The rich-club coefficient F(k) was lower under LSD than placebo, suggesting decreased preferential connectivity among the brain's dominant hub regions; k-core analysis indicated that under placebo hubs were concentrated within a primary sensory module, whereas LSD reduced this preferential hub communication. The extracted text notes that the investigators attempted to control for head motion but that significant motion differences persisted between conditions; motion-control procedures and related information appear in the Supplemental Experimental Procedures.

Tagliazucchi and colleagues interpret their findings as evidence that LSD enhances global and between-module brain communication while diminishing the integrity of individual functional modules, effects that are particularly centred on high-level association cortices and the thalamus—regions with high 5-HT2A receptor density. They relate these results to previous psilocybin imaging work that reported decreases in within-network integrity and increases in between-network connectivity, and they report that re-analysis of earlier psilocybin data yields FCD increases similar to those observed here with LSD. The authors propose a plausible mechanistic link: 5-HT2A receptor agonism increases excitability of layer V pyramidal neurons, which may elevate metabolic demand and thus relate to the observed increases in functional connectivity; prior reports of increased glucose metabolism under psychedelics and correlations between metabolism and subjective ego-dissolution are invoked to support this chain of inference. Electrophysiological findings are integrated into the account: prior EEG and MEG studies report broadband reductions in oscillatory power (notably alpha decreases) under psychedelics, and multimodal work suggests an inverse relationship between global functional connectivity and alpha power. The authors therefore suggest that psychedelic-induced desynchronisation of oscillatory activity (via 5-HT2A-mediated effects) may be consistent with the observed expansion of global connectivity. Functionally, the regions showing increased global coupling perform diverse roles related to self-related processing: frontoparietal areas are implicated in conscious information access, precuneus activity links to self-reflection and autobiographical memory, temporo-parietal junctions to out-of-body-like phenomena, and insular cortex to self-awareness and emotional processing. The authors argue that increased cross-talk between these high-level networks and sensory systems could blur hierarchical distinctions between lower-level sensory processing and higher-level integrative systems, offering a neurobiological account for the subjective blurring of ego boundaries and ‘‘expanded awareness.’’ The discussion acknowledges important limitations reported by the investigators: residual head-motion differences despite mitigation efforts, a reduced sample size (15 clean datasets from an original 20), reliance on a single numeric ego-dissolution rating rather than a more extensive characterisation of subjective experience, and the fact that participants were experienced psychedelic users who may have been able to distinguish LSD from placebo. The authors note they found no correlation between past use and the reported effects in this dataset but suggest replication in psychedelic-naive participants would be informative. Finally, the authors position their results within a broader conceptual framework contrasting psychedelic states with unconscious states: whereas sleep and anaesthesia show reduced frontoparietal metabolism and increased modularity (loss of global integration), LSD produced increased frontoparietal FCD and decreased modularity. They update a prior entropy-based hypothesis of conscious states to suggest that a brain's modularity during a period of time may be predictive of the subjective quality of consciousness experienced then, and they conclude that their measures linking increased global integration in association cortices to ego dissolution advance understanding of both the psychedelic state and a core aspect of human selfhood. They also reiterate that, when conducted with appropriate safeguards, human psychedelic research can be safe and informative.