Me, myself, bye: regional alterations in glutamate and the experience of ego dissolution with psilocybin

Psychedelics induce pronounced alterations of consciousness, including transient distortions in self-experience commonly described as ego dissolution, in which self-referential awareness is reduced and self–world boundaries are blurred. Earlier research links the hallucinogenic effects of classical psychedelics such as psilocybin to agonism of cortical 5-HT2A receptors, and preclinical work further implicates glutamatergic signalling downstream of 5-HT2A activation as a potential pathway mediating acute neural and behavioural effects. Functional imaging studies have also associated alterations in default mode network (DMN) function with ego-dissolution experiences, and both the medial prefrontal cortex (mPFC) and hippocampus are candidate regions due to their 5-HT2A receptor density and roles in self-referential processing. Mason and colleagues set out to test, in humans, whether acute psilocybin alters glutamate concentrations in key brain regions and whether such changes relate to subjective ego-dissolution and resting-state network functional connectivity (FC). Using ultra-high field proton magnetic resonance spectroscopy (7T MRS) targeted at the mPFC and right hippocampus, together with resting-state fMRI and validated questionnaires of altered states and ego dissolution, the study aimed to characterise region-specific neurometabolic changes during peak drug effects and assess their association with both network-level connectivity and subjective experience.

The investigators employed a randomised, double-blind, placebo-controlled, parallel-group design. Sixty healthy adults with prior but not recent (within 3 months) psychedelic experience were allocated to oral psilocybin (0.17 mg/kg) or placebo, with groups matched on age, sex and education. Ethical approvals and informed consent procedures were completed in line with relevant regulations. Neuroimaging and subjective assessments were timed to capture peak effects. Structural MRI was acquired at 50 minutes post administration, single-voxel MRS at the mPFC (65 min) and right hippocampus (95 min), and resting-state fMRI at 102 min, all on a 7T scanner. MRS voxels measured glutamate, GABA, N-acetyl-aspartate (tNAA) and myo-inositol (mI) relative to total creatine (tCr) to correct for common acquisition variabilities. MRS data were analysed using LCModel. Resting-state data processing used the CONN toolbox with group independent component analysis (ICA) restricted to 20 components; within-network and between-network FC were assessed with voxel-wise and ROI-based approaches, respectively. Subjective state was measured retrospectively at 360 minutes post dosing using the 5-Dimensions of Altered States of Consciousness (5D-ASC) scale and the Ego Dissolution Inventory (EDI). Venous blood samples at 80, 150 and 360 minutes assessed psilocin concentrations. Statistical tests included non-parametric Mann–Whitney U tests for metabolite and questionnaire comparisons, two-sample t-tests for voxel-wise and ROI FC contrasts (voxel p < 0.001 uncorrected, cluster p < 0.05 FDR-corrected), and canonical correlation analysis to relate biological predictors (relative glutamate in mPFC and hippocampus plus extracted within-network FC metrics) to subjective measures of ego dissolution. Missing data were handled with iterative imputation where applicable, and significance was set at p < 0.05.

Demographics and pharmacokinetics: The psilocybin (n = 30) and placebo (n = 30) groups did not differ on reported demographic variables. Mean serum psilocin peaked at 80 minutes post dose (15.61 ± 1.66 ng/mL) and declined by 360 minutes (4.85 ± 0.54 ng/mL), consistent with the administered oral dose. Subjective effects: Psilocybin produced robust subjective effects. Scores increased on all 5D-ASC dimensions (U = 13.5–225; p ≤ 0.001; effect sizes 0.43–0.84) and on the EDI (U = 91.5, p < 0.001, effect size = 0.67), indicating both positively and negatively valenced ego-dissolution experiences relative to placebo. MRS findings: Spectral quality and exclusions were reported but specific numbers were given in tables referenced in the text. In the mPFC, relative glutamate (glutamate/tCr) was higher after psilocybin than placebo (psilocybin 1.23 ± 0.02 vs placebo 1.14 ± 0.02; U = 200.50, p = 0.01; effect size = 0.80). The mPFC also showed higher tNAA/tCr (1.41 ± 0.03 vs 1.31 ± 0.02; U = 210.0, p = 0.02; effect size = 0.72) and higher GABA/tCr (0.17 ± 0.01 vs 0.14 ± 0.01; U = 66.0, p = 0.01; effect size = 0.99) under psilocybin. By contrast, hippocampal glutamate/tCr was lower after psilocybin versus placebo (0.77 ± 0.03 vs 0.88 ± 0.03; U = 163.50, p = 0.03; effect size = 0.69). No other significant group differences in hippocampal GABA, tNAA, mI or total creatine were observed, and the authors note that reliable hippocampal GABA quantification was not achievable in this dataset. Resting-state networks: After quality control the rsfMRI sample was reduced to 22 psilocybin and 26 placebo participants. ICA identified multiple canonical networks, though the DMN and sensorimotor networks were split into subcomponents; anterior and posterior DMN components were included in analyses. Within-network coactivation was significantly reduced under psilocybin in visual networks 1 and 2, both DMN subcomponents (anterior and posterior), and the auditory network. Between-network FC showed widespread increases under psilocybin relative to placebo, affecting all investigated networks except the lateral motor network. Brain–behaviour relationships: A canonical correlation analysis examined four biological predictors (mPFC and hippocampal glutamate plus selected within-network FC measures) against three ego-dissolution variables (anxious ego dissolution (AED), oceanic boundlessness (OB), and EDI scores). The full model was significant (F(12,61.14) = 2.47, p = 0.008) and explained 65.9% of variance. The first two canonical functions were noteworthy (squared canonical correlations R_c^2 = 0.363 and 0.282). Function 1 was dominated by AED and was principally predicted by increased mPFC glutamate with a secondary contribution from anterior DMN FC, indicating that higher frontal glutamate most strongly related to negatively experienced ego dissolution. Function 2 was dominated by EDI and OB and was primarily predicted by decreased hippocampal glutamate with secondary contribution from posterior DMN FC, implicating hippocampal glutamate decreases in positively valenced ego-dissolution experiences.

Mason and colleagues interpret their findings as the first human demonstration that psilocybin acutely induces region-dependent alterations in glutamate, with increases in the mPFC and decreases in the hippocampus, and that these differential changes correlate with distinct dimensions of ego dissolution. The observed mPFC glutamatergic increase aligns with preclinical evidence that 5-HT2A receptor activation elevates cortical glutamate release and with human imaging reports of increased frontal cerebral blood flow after psychedelics. The concurrent mPFC increase in GABA and tNAA suggests a complex local response involving both excitatory and inhibitory elements and a potential change in neuronal metabolic state; tNAA is noted as a marker often interpreted as reflecting neuronal viability or function. Regarding the hippocampal glutamate decrease, the researchers discuss possible mechanisms including indirect inhibition via 5-HT2A receptor effects on interneurons and direct activation of inhibitory 5-HT1A receptors, which are relatively dense in limbic regions; however, the study cannot distinguish these mechanisms. The authors link hippocampal glutamate decrements and posterior DMN decoupling to hypotheses that loss of access to autobiographical semantic information underlies positively experienced ego dissolution, and they cite parallels with prior findings that decoupling between medial temporal lobe structures and other networks correlates with ego-dissolution ratings. The canonical correlation results are interpreted to suggest that increased mPFC glutamate is the strongest predictor of negatively valenced ego-dissolution (anxious, loss-of-control experiences), whereas reduced hippocampal glutamate predicts positively experienced ego-dissolution (feelings of unity and boundary dissolution). The authors note the apparent paradox that acute psychedelic administration can increase anxiety while longer-term clinical trials report reductions in anxiety, and they propose speculative mechanisms such as acute glutamate-driven hyperfrontality versus longer-term receptor downregulation and plasticity; they link the glutamatergic pathway to putative downstream increases in BDNF and neuroplasticity described in preclinical models. Limitations acknowledged by the study team include the moderate dose chosen to maintain participant tolerance in the scanner, which may not produce maximal ego dissolution, and technical constraints of ultra-high field fMRI such as increased geometric distortions and relatively short scan duration for test–retest reliability. The authors also highlight methodological limits in reliably quantifying hippocampal GABA in this dataset and the difficulty of maintaining blinding due to the overt subjective effects of psychedelics, suggesting active placebo or cross-psychotropic controls in future work. Finally, they emphasise that while the findings offer a neurochemical correlate for aspects of ego dissolution and a potential mechanistic link to therapeutic effects, causal pathways and long-term changes remain to be established in further research.