Finding the self by losing the self: Neural correlates of ego-dissolution under psilocybin

Lebedev and colleagues frame the self as a multifaceted construct—encompassing self-awareness, agency, ownership, multimodal integration and narrative identity—that can be disrupted in states such as acute psychosis, temporal lobe epilepsy auras and mystical experiences. Previous work implicates medial temporal lobe (MTL) circuitry (including the parahippocampal cortex, PHC), the default mode network (DMN) and the salience network in aspects of self-related processing, and earlier neuroimaging and electrophysiological studies have reported MTL abnormalities and altered DMN dynamics under psychedelics such as psilocybin. The authors note that ego-dissolution under psychedelics offers an experimentally tractable perturbation for probing the neural bases of selfhood and its disturbances, and they situate their approach within contemporary Bayesian/free-energy accounts of hierarchical brain function. The present study set out to identify large-scale network changes associated with psilocybin-induced ego-dissolution. Using graph-theoretical and connectivity methods applied to resting-state fMRI, the investigators tested specific hypotheses that parahippocampal–neocortical decoupling and altered DMN or salience network integrity would be related to subjective reports of ego-loss. They also explored whether baseline network metrics might predict individual susceptibility to ego-dissolution under psilocybin. The study therefore combines hypothesis-driven ROI analyses with whole-brain, hypothesis-free network approaches to examine both state changes and baseline predictors.

The study used a within-subject, counterbalanced, placebo-controlled design. Fifteen healthy volunteers (13 males, 2 females) with prior psychedelic experience were scanned on two occasions: one saline (placebo) session and one session during which 2 mg psilocybin was infused intravenously midway through a 12-minute, eyes-closed, task-free fMRI run. The extracted text does not clearly report the participants' mean age in an unambiguous format. Inclusion criteria included age over 21, absence of personal or immediate family history of major psychiatric disorder, no drug dependence or cardiovascular disease, and no prior significant adverse reaction to hallucinogens. Screening comprised physical examination, ECG, blood and urine tests and psychiatric assessment; participants completed anxiety and depression inventories. Subjective experience was measured after scanning using 22 visual analogue scale (VAS) items (0–100). Principal component analysis (PCA) reduced these ratings; the first principal component (dominantly loading on an item reading "I lost all sense of ego") was taken as the primary measure of ego-dissolution. Pearson correlations were used to relate this composite score to neuroimaging metrics after testing normality with the Shapiro–Wilk test. Imaging was acquired on a 3T Siemens Trio Tim with a 32-channel head coil. Anatomical MPRAGE and 240 T2*-weighted EPI volumes per session were obtained (TR 3000 ms, 3 mm isotropic voxels). Preprocessing combined SPM-8 and FSL-5.0 steps: realignment, slice-timing correction, linear and nonlinear registration to native and MNI space, regression of white matter and CSF signals and Friston's extended 24-parameter motion model, band-pass filtering, and 6 mm smoothing. Outlier detection on PCA scores and connectivity matrices found no outliers. For network analyses, the authors used Craddock parcellations (primarily a 200-ROI atlas, with replication using a 600-ROI version) and Yeo's 7-network scheme for within-network analyses. Adjacency matrices employed Pearson correlations between ROI time series and graph measures were extracted with the Brain Connectivity Toolbox. Community structure was estimated from placebo data (fine-tune modularity algorithm), yielding five communities that informed calculation of a Shannon entropy–based diversity coefficient for each ROI (a measure of how diffusely an ROI connects across communities). Analyses contrasted post-psilocybin, pre-psilocybin and placebo states, and included both hypothesis-driven ROI tests (notably anterior PHC and posterior medial cortex nodes) and exploratory whole-brain tests. Within-network integrity was assessed via mean clustering coefficient per Yeo network. Hypothesis-free, edgewise analyses used Network Based Statistics (NBS) with permutation testing to control family-wise error. Statistical analyses were performed in R (version 3.1) with correction for multiple comparisons using FDR for node-wise tests and FWE via NBS for edgewise tests. The investigators also assessed head motion and temporal signal-to-noise ratio (tSNR) to evaluate potential confounds.

Subjective measures: PCA of the 22 VAS items yielded a first principal component that explained 36.37% of variance and correlated strongly with the item "I lost all sense of ego" (r = 0.82). This component also correlated with items reflecting loss of separation from surroundings (r = 0.76), a sense of merging with the environment (r = 0.72), and with ratings of spiritual/mystical (r = 0.81) and supernatural qualities (r = 0.69). The second component explained 13.97% of variance and appeared related to emotional valence, with positive loadings on spiritual quality and negative loadings on muddled thinking and unusual bodily sensations. Diversity coefficients — hypothesis-driven analyses: A strong negative association was observed between ego-dissolution intensity (first PCA component) and the diversity coefficient of the anterior parahippocampal cortex (aPHC). In other words, higher ego-dissolution was associated with reduced functional diversity or decoupling of aPHC from cortical communities under psilocybin. An ROI-based connectivity check involving the hippocampal formation corroborated an association between ego-loss and reduced MTL–cortical connectivity. By contrast, the posterior medial cortex nodes (posterior cingulate cortex and retrosplenial cortex) did not show associations with the primary ego-dissolution measure; instead, the second PCA component (emotional valence) related strongly to retrosplenial diversity, with larger RSC diversity linked to more unpleasant experiences (muddled thinking, unusual bodily sensations and paranoid ideas). Diversity coefficients — hypothesis-free analyses: In baseline (pre-psilocybin) data, lower diversity in left frontoparietal regions (left inferior frontal gyrus, left middle frontal gyrus and superior parietal lobule, belonging to higher cognition/control communities) predicted greater ego-dissolution under psilocybin; these associations survived multiple-comparisons correction. Within-network integrity: Mean clustering coefficient analyses indicated that ego-dissolution was strongly associated with disintegration (reduced clustering) of the salience network. This result was supported when comparing placebo to post-psilocybin states. Whole-brain functional connectivity: Hypothesis-free edgewise NBS analyses linked ego-dissolution with reduced interhemispheric communication and disconnection of MTL regions from parietal lobes. The pattern persisted when negative edges were excluded. Confound checks: Head-motion analysis showed only one significant comparison — mean standard deviation of head motion was slightly higher in the post-psilocybin compared with the pre-psilocybin state (p_uncorr = 0.015) — but motion metrics did not significantly impact the network measures and did not correlate with the ego-dissolution composite. tSNR analyses indicated that, overall, SNR measures (placebo, pre-psilocybin, post-psilocybin and their differences) did not correlate with ego-dissolution; however, for filtered data only, reduced tSNR in the retrosplenial cortex was strongly negatively related to its diversity coefficient (r = -0.71, P < 0.005).

Lebedev and colleagues interpret their findings as indicating that psilocybin-induced ego-dissolution is associated with MTL–neocortical decoupling, reduced integrity of the salience network and diminished interhemispheric communication. They emphasise that the parahippocampal cortex—conceived as a multimodal hub mediating MTL–neocortical cross-talk—showed reduced diversity of cortical connections in association with ego-loss, consistent with the idea of a disrupted contextual "feed" into ongoing cognition. The authors note that, although prior work has implicated the DMN in self-related processing and they had predicted altered DMN integrity with ego-dissolution, the present results did not show a direct association between DMN disintegration and the ego-loss composite. They propose a functional distinction in which the DMN preferentially supports the "narrative self" (ego-identity), while the salience network contributes to aspects of the "minimal" or embodied self (ego-vitality, agency, demarcation). Supporting this, increased functional diversity of posterior DMN nodes (notably the retrosplenial cortex) correlated with unpleasant, psychosis-like content (second PCA component) rather than with the primary ego-loss measure. In discussing baseline predictors, the investigators report that lower diversity (i.e. tighter segregation) in executive frontoparietal regions at baseline predicted greater susceptibility to ego-dissolution. They suggest this may reflect that individuals with especially segregated network organisation at baseline possess firmer ego-boundaries and are therefore more sensitive to the ego-dissolving effects of psilocybin. The authors frame these results within Bayesian/free-energy perspectives, proposing that psychedelics may reduce confidence in hierarchical predictions and thereby alter inference about the world; decreased PHC–neocortex communication and salience network disruption are offered as potential mechanistic substrates for these effects. On clinical implications, the paper highlights parallels between psychedelic and psychotic states while noting differences—for example, increased DMN activity/connectivity is typically reported in acute endogenous psychosis whereas that pattern was not linked to ego-dissolution here. The researchers also point to possible relevance for addiction treatment, referencing historical work that used ego-dissolution approaches, and note that ego-dissolution is not intrinsically unpleasant and may underlie therapeutic or mystical-type experiences. Key limitations acknowledged by the authors include the small sample size, the fact that all participants had prior psychedelic experience (which may affect generalisability), and the use of a PCA-derived composite that captures the most variable aspects of experience rather than the most universal. They stress that the ego-dissolution construct requires further validation and that future studies should combine subjective and behavioural measures and apply standardised instruments (for example, ego-pathology inventories) to characterise different domains of ego-disturbance across states.

Lebedev and colleagues conclude that psilocybin-induced ego disturbance is associated with MTL–cortical decoupling, decreased salience network integrity and reduced interhemispheric communication. They suggest that maintenance of the perceptual sense of self or ego may depend on the normal functioning of these systems and call for standardised evaluation of ego-dissolution and related phenomena to better compare different altered states of consciousness and advance understanding of the neural bases of the self.