Changes in global and thalamic brain connectivity in LSD-induced altered states of consciousness are attributable to the 5-HT2A receptor

Disorders of perception and thought contribute substantially to global disease burden, and psychedelic drugs provide pharmacologic probes for studying the neural mechanisms of consciousness. Preclinical and early human work indicate that lysergic acid diethylamide (LSD) acts at multiple serotonin (5-HT) and dopamine receptors and produces characteristic alterations in perception, mood and self-experience. However, human data linking specific receptor actions of LSD to large-scale functional brain network changes remain limited, and prior connectivity studies have been constrained by seed-based approaches and have not systematically addressed the influence of global signal (GS) artefacts. This study set out to characterise time-dependent, receptor-mediated effects of LSD on whole-brain functional connectivity using a data-driven Global Brain Connectivity (GBC) metric, and to relate those connectivity patterns to cortical gene expression maps for candidate receptors. Preller and colleagues hypothesised that LSD-induced GBC changes would spatially match cortical expression of the HTR2A gene (encoding the 5-HT2A receptor), that ketanserin (a 5-HT2A antagonist) would block LSD effects, and that artefact removal via global signal regression (GSR) would materially influence GBC findings. The work therefore aims both to specify the receptor basis of LSD-induced network changes and to evaluate methodological factors relevant to pharmacological resting-state fMRI.

Design and participants: The investigators conducted a double-blind, randomised, placebo-controlled, cross-over study in 24 healthy adults (19 male, 5 female; mean age 25.0, SD 3.6, range 20–34) recruited from university advertisements. Screening excluded current or past psychiatric disorder, major medical conditions and recent substance use. One of 25 initial participants was excluded for registration failure, leaving n = 24. The protocol was authorised by Swiss regulators and registered at ClinicalTrials.gov (NCT02451072). Interventions and timing: Each participant underwent three sessions, two weeks apart, receiving in counterbalanced order: (i) placebo pretreatment + placebo (Pla), (ii) placebo pretreatment + LSD (LSD; reported dose 100 mg po in the extracted text), and (iii) ketanserin pretreatment (40 mg po) + LSD (Ket+LSD). Pretreatment was given 60 minutes before LSD/placebo administration. Resting-state fMRI scans were acquired at two time points per session: session 1 at 75 minutes and session 2 at 300 minutes after treatment. Subjective effects were assessed with the full 5D-ASC at 720 minutes and a short 5D-ASC at 180, 250 and 360 minutes. Neuroimaging acquisition and preprocessing: MRI data were collected on a Philips 3.0T scanner with a 32-channel head coil. Resting-state EPI parameters included TR = 2,500 ms, TE = 27 ms, 45 slices, 240 volumes (10 minutes) per scan; high-resolution T1- and T2-weighted images were also acquired. Preprocessing followed Human Connectome Project pipelines: distortion correction, motion correction, registration to structural space, conversion to CIFTI grayordinate space and high-pass filtering (>0.008 Hz). Nuisance regressors included ventricular and white matter signals, motion parameters and their derivatives. Global signal regression (GSR) and its first derivative were performed in the primary pipeline, but parallel analyses were also run without GSR to examine its influence. Motion scrubbing used framewise displacement and RMS-intensity criteria; subjects with >50% flagged frames would be excluded (none were excluded on this basis in the final sample). Connectivity measures and additional analyses: Global brain connectivity (GBC) was computed for each grayordinate by correlating its time series with all other grayordinates, Fisher z-transforming correlations and averaging to yield a single GBC value per grayordinate. Complementary seed-based thalamic analyses were performed by extracting bilateral thalamus time series and computing correlation and covariance maps. To evaluate GS topography, a GS beta map was derived by fitting the GS as a regressor in a GLM at each grayordinate and entering the resulting beta weights into second-level analyses. All main analyses were conducted both with and without GSR. Statistical approach: Group-level GBC maps were entered into a 2 × 3 repeated-measures ANOVA (session × drug condition) and tested using permutation-based inference (FSL PALM) with threshold-free cluster enhancement (TFCE) and 10,000 permutations, family-wise error corrected at p<0.05. Behavioural 5D-ASC scores were analysed with repeated-measures ANOVA (treatment, scale, time where applicable). Correlations between imaging measures and subjective scores were Bonferroni-corrected. Gene expression analyses correlated unthresholded GBC Z-score maps with cortical gene expression maps from the Allen Human Brain Atlas (group-averaged left-hemisphere maps for DRD1, DRD2, HTR1A, HTR2A, HTR2C and HTR7) following a previously described surface-based interpolation procedure.

Sample and subjective effects: Twenty-four participants completed the study. Retrospective 5D-ASC scores showed a significant main effect of drug condition and scale and a drug × scale interaction (ANOVA F (2,46)=88.49, p<0.001 for drug condition). Post-hoc tests indicated increased ratings on nearly all 5D-ASC scales in the LSD condition compared to Pla and Ket+LSD (all p<0.05) except for spiritual experience and anxiety. Short-version 5D-ASC scores over time (180, 250, 360 min) similarly showed significant main effects and interactions; LSD effects peaked earlier and were abolished by ketanserin, with no differences between Pla and Ket+LSD at any time point. Global brain connectivity (with GSR): When GBC was computed after GSR, permutation-corrected analyses revealed widespread drug-related differences. Comparing LSD to the combined (Ket+LSD)+Pla contrast, LSD induced hyper-connectivity predominantly in sensory and somatomotor regions (occipital cortex, superior temporal gyrus, postcentral gyrus, precuneus) and hypo-connectivity in associative cortical areas and subcortex (medial/lateral prefrontal cortex, cingulum, insula, temporoparietal junction). Effects were bilateral. The LSD versus Pla and LSD versus Ket+LSD contrasts showed highly similar patterns, with the respective Z-maps correlating at r = 0.91 (p<0.001). Mean connectivity (Fz) did not differ between Pla and Ket+LSD in hyper- or hypo-connected areas. Across subjects, greater LSD-induced hyper-connectivity in sensory regions correlated with stronger hypo-connectivity in associative regions (r = −0.90, p<0.001), suggesting linked systems-level perturbations. Influence of global signal regression (no GSR): Repeating analyses without GSR yielded a different spatial pattern: predominantly left-hemispheric widespread differences, with LSD-related hypo-connectivity mainly in the right insula and hyper-connectivity in the cerebellum and left-hemispheric frontal/temporal regions. LSD>Pla and LSD>Ket+LSD Z-maps remained correlated (r = 0.81, p<0.001). However, the relationship between hyper- and hypo-connectivity flipped sign (r = 0.92, p<0.001), and combined GBC maps with and without GSR were uncorrelated within subjects (r = 0.003, p = 0.99). Analyses of local grayordinate-wise variance and the variance of the mean GS showed no substantial amplitude changes between conditions, but GS beta maps revealed a pronounced spatial topography shift under LSD that correlated with GBC changes: the LSD−Pla GS beta map correlated negatively with the GBC map after GSR (r = −0.65, p<0.001) and positively before GSR (r = 0.66, p<0.001). This demonstrates that shifts in GS topography substantially influence GBC interpretations. Thalamic seed analyses and robustness: To resolve directionality concerns, the investigators computed thalamic seed-based connectivity using both correlation and covariance and constructed a conjunction map robust to processing choices. Using this mask, LSD consistently decreased thalamic coupling with associative areas and increased coupling with sensory-somatomotor regions after GSR; these thalamic effects matched GBC results. Without GSR, thalamic and GBC outcomes were inconsistent. Individual differences in top- vs bottom-10% thalamic connections followed predicted negative relationships after GSR but not without GSR. Network-level and time-course results: Analyses across seven functionally defined networks showed significant main effects of drug for most networks when including GSR, with LSD differing from Pla and Ket+LSD in many networks. No significant session (scan 1 vs 2) differences were observed within Pla or LSD conditions, but within Ket+LSD there were session-dependent GBC changes (decreases in occipital areas and increases in cingulate, temporoparietal junction, thalamus, basal ganglia). Subjective effects peaked earlier (≈180 min) and decreased thereafter; ketanserin blocked subjective effects throughout the time course. Relationship to subjective experience: Change in somatomotor network GBC (LSD−Pla, session 2, with GSR) correlated strongly with the mean 5D-ASC short-version score at 250 min (r = 0.81, p<0.001, Bonferroni corrected). Correlations between subjective scores and the other six networks were non-significant. All five short-version scales (blissful state, disembodiment, changed meaning of percepts, elementary imagery, spiritual experience) correlated significantly with somatomotor Fz change. Gene expression correlations: The unthresholded GBC Z-score map for LSD vs (Ket+LSD)+Pla with GSR showed a significant positive spatial correlation with the cortical HTR2A expression map (r = 0.50, p<0.001), stronger than for other candidate receptor genes. The same correlation without GSR was weaker (r = 0.18, p<0.001) and significantly lower than with GSR. GBC Z-scores with GSR were also strongly negatively correlated with HTR7 expression (r = −0.63, p<0.001), consistent with a strong anti-correlation between HTR2A and HTR7 expression (r = −0.68, p<0.001). Correlations with gene expression maps were generally stronger after GSR.

Preller and colleagues interpret their findings as showing that LSD produces a characteristic pattern of brain-wide dysconnectivity: increased GBC across sensory and somatomotor networks alongside reduced GBC in associative networks. They argue that this pattern is consistent with an imbalance favouring enhanced sensory processing that is insufficiently integrated by associative systems, a configuration that may underlie psychedelic phenomenology. Ketanserin pretreatment normalised both subjective and neural LSD effects, leading the investigators to attribute these network changes primarily to stimulation of the 5-HT2A receptor. Methodological observations receive substantial emphasis. The authors show that GBC results are highly sensitive to global signal regression: analyses without GSR yielded different spatial patterns and altered relationships between hyper- and hypo-connectivity and subjective measures. GS beta-map analyses indicate that LSD alters the topography of the global signal itself, which in turn drives GBC changes when GS is not removed. To address directionality concerns, thalamic seed-based analyses that are robust to GS transformations replicated the post-GSR finding of decreased thalamo-associative coupling and increased thalamo-sensory coupling. The investigators therefore consider GSR-cleaned results—corroborated by thalamic seeds—to provide a coherent interpretation, while acknowledging that GSR may also remove neuronal signal and that the GS contains both neural and artefactual components. The spatial match between LSD-induced GBC changes and cortical HTR2A expression maps is presented as convergent evidence implicating the 5-HT2A receptor in LSD neuropharmacology. The negative correlation with HTR7 expression is related to the anti-correlated cortical expression of HTR2A and HTR7; the authors note that LSD's reported antagonism at 5-HT7 makes a major role for HTR7 less likely, but that further study is warranted. Time-dependent effects are discussed: ketanserin blocked subjective effects over the entire measured time course, but session-dependent GBC changes in the Ket+LSD condition suggest more complex, potentially multi-phasic pharmacology, raising the possibility of later involvement of other receptors such as dopamine D2; the authors stress this hypothesis requires further testing. Limitations and future directions acknowledged by the investigators include the controversy over GSR and the difficulty of disentangling artifactual from neural contributions to the GS, unknown effects of LSD on neurovascular coupling and haemodynamic responses, and the need to test additional receptor antagonists (and ketanserin alone) to map receptor contributions beyond 5-HT2A. They also propose cross-validation of GBC–gene expression correspondences with PET receptor maps and suggest animal studies combining neuronal recordings with haemodynamic measures to clarify neurovascular influences. Clinically, the authors highlight the amygdala hyper-connectivity and somatomotor system associations with subjective effects as potentially relevant for understanding therapeutic mechanisms, but they do not overstate translational claims and call for further mechanistic and dose-ranging studies.

Preller and colleagues conclude that LSD induces widespread alterations of global brain connectivity characterised by synchronisation of sensory and somatomotor networks and disintegration of associative networks, and that these neural and subjective effects are largely attributable to stimulation of the 5-HT2A receptor as shown by blockade with ketanserin. They emphasise that GSR materially influences GBC results and that cortical gene expression mapping provides convergent evidence linking in vivo pharmacological effects to HTR2A spatial distribution. The authors suggest that these findings advance understanding of psychedelic neurobiology, point to potential roles for other receptors (notably 5-HT7) that merit further study, and identify somatomotor network connectivity as closely related to the psychedelic experience, with implications for future therapeutic development and mechanistic research.