LSD-induced entropic brain activity predicts subsequent personality change

This was a balanced-order, placebo-controlled within-subjects neuroimaging study in which 19 healthy adults underwent two intravenous infusions—LSD (reported in the extraction as 75 mg, I.V.) and saline placebo—separated by at least two weeks. For each session three 7.5-minute eyes-closed resting-state BOLD fMRI scans were acquired in the following order: rest 1 (no music), rest 2 (music), and rest 3 (no music). Two ambient tracks were used for the music condition, presented through headphones and selected to be balanced for emotional potency. The infusion was administered over two minutes and occurred 115 minutes before the resting-state scans began. One subject stopped the MRI scanning prematurely due to anxiety; their behavioural/personality data were retained in the analyses. Eligibility screening included standard physical and psychiatric examinations, electrophysiology, urine/blood drug screens and pregnancy tests. The extracted text does not clearly report the exact lower age limit (it states "age 211 years" which appears garbled). Reported inclusion criteria otherwise comprised no personal or immediate family history of major psychiatric disorder, no cardiovascular disease, and no history of a significant adverse response to psychedelics. All participants had used classic psychedelics at least once previously but not within two weeks of their first study day. Image preprocessing used SPM12 via the DPARSFA pipeline in MATLAB. Preprocessing steps included slice-timing correction, realignment, registration to MNI space with a population template, regression of white matter and CSF signals, motion correction, detrending and band-pass filtering (0.01–0.1 Hz). Voxel-wise sample entropy (SamEn) was calculated using the "complexity" toolbox; SamEn measures the unpredictability of a time-series (higher sample entropy indicates lower predictability). Entropy maps were parcellated using Yeo's 17-network functional atlas and a multi-scale SamEn approach was applied across time scales 1–5 (equivalent to averaging non-overlapping windows of increasing length and then computing SamEn). Voxel maps were smoothed (6 mm FWHM) for voxelwise analyses. Behavioural measures included the NEO-PI-R (240 items) administered at baseline and two weeks after the LSD/placebo sessions to assess personality change, and an in-scanner visual analogue scale for ego-dissolution completed after each scan. Statistical analyses were implemented in R (version 3.2.2) with mixed-effects models ("nlme" package) treating subject as a random effect. Voxel-wise contrasts were followed up in SPM12 and family-wise error correction for clusters was estimated via Monte Carlo simulation (AlphaSim, 5,000 iterations; cluster-forming threshold P < 0.005). Models included covariates for motion and state-by-drug interactions (linear and quadratic terms) and a nuisance factor indicating whether subjects experienced headphone sound problems during the music scan. Primary contrasts of drug effects on brain dynamics used an FWE threshold of P FWE < 0.05; behavioural analyses used P uncorrected < 0.05. Additional analyses tested whether entropy changes under LSD predicted changes in openness (DEntropy → DPersonality), and whether these relationships interacted with scan state (music) and with ego-dissolution ratings, yielding interaction terms up to the four-way DEntropy × DPersonality × State × Ego-dissolution. Secondary checks included effects of drug order and prior psychedelic exposure (summarised via principal component scores for lifetime uses and days since last use).

Direct comparisons of LSD versus placebo revealed widespread increases in sample entropy. A global cortical effect was reported (T = 4.34, P < 0.001), with significant changes in 11 of the 17 functional networks including primary and secondary sensory and associative systems as well as higher-order divisions. Voxel-wise analyses produced FWE-corrected clusters of increased entropy in frontoparietal, medial occipital, posterior and dorsal cingulate regions. Multi-scale analyses showed that entropy increases under LSD were present primarily at the shorter time scales (scales 1–3) and were not significant at scales 4 and 5. Personality assessment focused on trait openness. Openness increased significantly after the LSD session (T = 1.95[19], P = 0.03; Cohen's d = 0.16), whereas no significant openness changes were observed after placebo. The extracted text notes that a fuller psychological dataset is reported elsewhere and that exploratory analyses of other traits were performed post hoc. Critically, acute LSD-induced increases in brain entropy predicted subsequent increases in openness. Global entropy increases predicted openness change (T = 2.3, P = 0.035), and a significant change-by-state interaction (T = 3.83, P = 0.0005) indicated that the predictive relationship was enhanced when music was played during the scan. At the network level, both main effects of LSD on entropy and change-by-state interactions were predictive of openness increases, with contributions from both sensory and higher cognitive networks. An anomalous subject showing an atypical decrease in openness post-LSD was identified as a suspected outlier; non-parametric Spearman correlations produced consistent results, suggesting robustness of the finding. When in-scanner ego-dissolution ratings were added to the model, a four-way interaction (DEntropy × DOpenness × State × Ego-dissolution) reached significance for two specific networks: the orbitofrontal network (N10; T = 2.07[27], P = 0.048) and the superior frontoparietal network (N13; T = 2.36[27], P = 0.026). This pattern indicates that participants who experienced greater ego-dissolution during the music scan and showed larger entropy increases in those networks also tended to exhibit the greatest increases in openness. The equivalent global four-way interaction did not reach significance (T = 1.71[27], P = 0.099). Post hoc inclusion of drug order (whether LSD was given in the first or second session) and composite measures of prior psychedelic experience did not eliminate the observed relationships between entropy changes and openness. The extraction also notes minor sound problems for four subjects during the music scan and that a nuisance factor for sound problems had little impact on the main results.

The researchers interpret their findings as supporting the entropic brain hypothesis: LSD produced pronounced increases in brain entropy across sensory and higher-order networks, predominantly at short time scales, consistent with a shift toward more random or less predictable neural dynamics. These acute changes in brain dynamics were predictive of enduring increases in trait openness measured two weeks later, and the link was strongest for scans during and after music listening and in participants who reported ego-dissolution. Lebedev and colleagues position these results alongside prior behavioural and animal work suggesting that classic psychedelics can enhance associative learning, cognitive flexibility and plasticity, and that peak mystical-type experiences are associated with persisting personality change. They discuss serotonin 2A receptor (5-HT2AR) agonism as a likely mechanistic substrate for the acute psychological effects and speculate that 5-HT2AR-related signalling may facilitate the neurobiological processes that enable major psychological change, pointing to animal findings of 5-HT2AR-mediated plasticity and the high developmental expression of this receptor. The authors highlight possible clinical relevance, suggesting that entropic shifts in cortical activity may underlie psychedelics' capacity to open individuals up to lasting change and that music and experiences of ego-dissolution could be important therapeutic mediators. They also note the concept of a psychedelic "afterglow," a transient post-acute period of increased psychological flexibility during which psychotherapy might be particularly effective. Several limitations acknowledged in the paper temper the conclusions. The music scan was not counter-balanced in order, preventing clear separation of music effects from pharmacodynamic timing; some participants had prior psychedelic experience although the reported openness increases appeared not to be attenuated by past use; and one participant discontinued scanning due to anxiety (their behavioural data were retained). The authors call for further work to delineate whether and how lasting changes in brain function or structure accompany the psychological changes, to explore the role of 5-HT2AR signalling in non-drug-related existential change and trauma, and to investigate how factors such as music and set-and-setting contribute to positive therapeutic outcomes with psychedelics.