Differential contributions of serotonergic and dopaminergic functional connectivity to the phenomenology of LSD

Psychedelic compounds profoundly alter conscious experience, and lysergic acid diethylamide (LSD) is the prototypical example that has been the subject of renewed neuroscientific interest. Previous neuroimaging work has documented system-level changes under LSD such as increased blood flow in primary visual cortex, reduced network integrity, and altered dynamics of integration and segregation. However, pharmacological functional MRI alone provides limited insight into molecular mechanisms, and although 5-HT2A receptor agonism has been shown to be necessary for many LSD effects, LSD also has affinity and agonist activity at multiple other serotonergic and dopaminergic receptors (including 5-HT1A/1B and D1/D2), making it difficult to delineate their distinct contributions in vivo. Lawn and colleagues set out to investigate how LSD modulates functional connectivity (FC) when analysis is informed by the spatial distributions of specific receptor systems. The study applied receptor-enriched analysis of functional connectivity by targets (REACT) to an existing within-subject resting-state fMRI dataset (LSD versus placebo) and examined whether receptor-enriched FC maps for 5-HT1A, 5-HT1B, 5-HT2A, D1 and D2 relate to subjective phenomenology measured with the Altered States of Consciousness (ASC) questionnaire and visual analogue scales (VAS). The goal was to explore dissociable serotonergic and dopaminergic contributions to perceptual, self-related and cognitive aspects of the LSD experience.

This secondary analysis used a subset of a publicly available dataset comprising 15 healthy volunteers (mean age 30.5 ± 8.0 years; 4 male, 11 female). Participants completed two scanning sessions in a balanced-order, within-subjects design separated by at least two weeks, receiving an intravenous infusion of either LSD (75 μg in 10 mL saline administered over 2 minutes) or placebo (10 mL saline). After a mock-scanner acclimatisation of approximately 60 minutes, MRI scanning was carried out about 70 minutes post-dose, a time corresponding to near-peak subjective effects in this protocol. Imaging was acquired on a 3 T scanner. Anatomical T1-weighted images were 1-mm isotropic. Three eyes-closed resting-state BOLD runs were collected (TR = 2000 ms, TE = 35 ms, 3.4-mm isotropic voxels; each run ~7 minutes) but only runs 1 and 3 were analysed here because run 2 included music. The uploaded dataset had undergone prior preprocessing; the extracted text does not report the full preprocessing pipeline in detail. Subjective experience was characterised with the 11-dimension ASC questionnaire (administered at the end of the LSD day) and VAS ratings taken after each resting-state run. VAS items used in analyses included reports of simple patterns/colours, complex visual imagery and ego dissolution. To reduce dimensionality of the ASC sub-scores, principal components analysis (PCA) was performed after excluding anxiety (which did not differ between conditions). Suitability for PCA was tested using the Kaiser-Meyer-Olkin measure and Bartlett's test; components were retained using the eigenvalue-one criterion and varimax rotation was applied to produce orthogonal, interpretable components. Molecular information came from population-based PET atlases: high-resolution 5-HT1A, 5-HT1B and 5-HT2A maps derived from 210 subjects, a D1 atlas from 13 volunteers ([11C]SCH23390), and a D2 atlas from 6 subjects ([18F]Fallypride). All atlases were normalised to values between 0 and 1 and voxels used as references in the kinetic models (cerebellum) were excluded. REACT was applied in a two-step regression analogous to dual regression but using molecular templates as spatial regressors. In step one the molecular templates (all five simultaneously) were regressed into demeaned, masked resting-state data to estimate subject-specific time series reflecting dominant BOLD fluctuations within each receptor distribution. In step two those time series were used as temporal regressors to produce subject-specific receptor-enriched spatial maps of BOLD response for LSD and placebo, restricted to grey matter. Spatial maps from runs 1 and 3 were averaged for each participant, session and receptor. For group inference, paired-sample t-tests compared LSD versus placebo receptor-enriched maps. Voxelwise linear relationships between delta FC maps (LSD minus placebo) and subjective measures (three PCA-derived ASC components and VAS items) were tested using FSL-Randomise with 5,000 permutations and threshold-free cluster enhancement (TFCE). Variance smoothing was set to 6 mm for the between-condition comparison and clusters surviving family-wise error correction (pFWE < 0.05) were reported. Mean FC values were extracted from significant clusters and, where appropriate, post hoc Pearson correlations were computed (for example between cerebellar vermal lobule X clusters and ego-dissolution VAS).

Receptor-enriched FC maps were estimated for each participant, receptor and session and then compared between LSD and placebo. Significant LSD-induced increases in FC were observed in maps enriched for 5-HT1A, 5-HT1B, 5-HT2A, D1 and D2 (all pFWE < 0.05, TFCE corrected). No significant FC decreases were reported for 5-HT1A, D1 or D2 maps, but decreases were found in cerebellar vermal lobule X for both 5-HT1B- and 5-HT2A-enriched maps. Spatially, 5-HT1A-enriched FC increases were localised to the right lingual gyrus. 5-HT1B-enriched increases were widespread, including bilateral cerebellar hemispheric lobules I–IV and V/VI, bilateral thalami, supplementary motor areas, superior parietal lobules, posterior cingulate regions, temporal–occipital fusiform and parahippocampal areas, precuneus and ventromedial prefrontal cortex, among others. 5-HT2A-enriched increases were observed in the right lateral occipital cortex, right lingual gyrus, left paracingulate and bilateral precentral gyri. D1-enriched increases involved right hippocampus, right thalamus, superior parietal lobe, occipital pole and multiple occipito-temporal regions. D2-enriched increases were found in right middle/superior temporal gyri, left posterior cingulate, intracalcarine and precuneus regions and the right cuneus. Subjective measures: LSD increased scores on all ASC sub-scales and VAS items except anxiety. PCA on the 11 ASC sub-scores (excluding anxiety) yielded three components that together explained 83.1% of variance: PC1 (positive-valence items such as unity, spiritual experience, blissful state, insightfulness and changed meaning), PC2 (cognitive/negative-valence items with strong loadings for disembodiment and impaired control and cognition) and PC3 (perceptual items such as audio-visual synaesthesia and imagery). PCA suitability metrics were KMO > 0.5 and Bartlett's test p < 0.05. Voxelwise correlations between delta FC and the three PCA components identified two TFCE-significant relationships with PC2: D1-enriched FC in the paracentral lobule (R2 = 0.79) and D2-enriched FC in a cluster encompassing paracentral lobule, parietal opercular cortex, precuneus and primary motor cortex (R2 = 0.73). By contrast, 5-HT2A-enriched FC extracted from these same clusters did not correlate with disembodiment (R2 = 0.06 and R2 = 0.01 respectively). Exploratory analyses of individual ASC sub-scores are reported in supplementary materials (extracted text indicates their presence but does not detail them). Correlations with VAS items produced two TFCE-significant effects: a negative correlation between 5-HT1B-enriched FC in the right intraparietal sulcus and simple hallucinations (R2 = 0.79), and a negative correlation between 5-HT1A-enriched FC in the precuneus and complex imagery (R2 = 0.88). 5-HT2A-enriched FC in the same clusters showed similar directionality (R2 = 0.66 and R2 = 0.63) but did not reach significance. Neither D1 nor D2 FC from these clusters related to the visual experiences. In a post hoc test, 5-HT1B-enriched FC in cerebellar lobule X (R2 = 0.20) and 5-HT2A-enriched FC (R2 = 0.16) correlated with VAS-reported ego dissolution but not with the ASC disembodiment score.

Lawn and colleagues interpret these findings as evidence that LSD modulates functional networks tied to multiple serotonergic and dopaminergic receptors, and that these receptor-enriched FC changes relate differentially to phenomenology. Serotonergic systems were predominantly associated with perceptual alterations, while dopaminergic systems related more to aspects of perceived selfhood and cognition. The authors frame these observations within existing models of psychedelic action such as cortico-striato-thalamo-cortical (CSTC) and REBUS (relaxed beliefs under psychedelics), noting that both serotonin and dopamine modulate components of CSTC loops and that prior imaging work has shown increased thalamic FC under LSD especially with sensory and somatomotor regions. Neuroanatomically, the study found increased 5-HT1A-enriched FC in thalamic and sensory regions and widespread serotonergic effects in visual and parietal areas that correlated with visual phenomenology. The precuneus and superior parietal lobule, regions implicated in visuo-spatial imagery, episodic memory retrieval and self-referential processing, showed receptor-enriched FC relationships with hallucinations and imagery. The only consistent FC decrease was observed in cerebellar vermal lobule X for 5-HT1B and 5-HT2A maps; post hoc correlations suggest a modest link between altered vestibulocerebellar serotonergic FC and ego dissolution measured shortly after scanning, though retrospective ASC disembodiment scores did not show the same relationship. Regarding dopaminergic findings, D1- and D2-enriched FC in somatomotor, superior parietal and insular/opercular regions correlated negatively with PC2, which loaded on disembodiment and impaired control/cognition. The authors suggest these regions form a network implicated in body ownership, agency and multisensory integration, and that dopaminergic modulation of this network may contribute to LSD-induced disembodiment and cognitive perturbations. They also discuss temporal dynamics: prior PET work suggests D2/3 ligand displacement under LSD peaks later (~240 minutes), and the current resting scans at ~70 minutes may sample an earlier, predominantly serotonergic phase, so dopaminergic contributions might be time-dependent and potentially stronger later in the acute LSD response. The investigators acknowledge several limitations. Molecular atlases used for REACT were derived from independent PET datasets and are assumed to generalise to the present cohort; this cross-sample assumption may not hold universally. The sample was small (N = 15) and female-biased, which limits power—particularly for voxelwise correlations—and precludes robust assessment of sex differences. The extracted text also notes overlap between different serotonergic receptor distributions, which complicates interpretation of receptor-specific findings and cautions against over-attributing effects to single receptor types. Finally, the authors emphasise that their findings are exploratory and call for replication in larger, better-powered datasets, for time-resolved studies across the LSD time-course, and for pharmacological manipulations (e.g. receptor-specific antagonists) to provide causal tests of receptor contributions. Overall, the study team concludes that LSD's effects on brain network organisation involve multiple receptor systems beyond 5-HT2A, with a double dissociation pattern in which serotonergic receptor-enriched FC related to perceptual experiences and dopaminergic receptor-enriched FC related to selfhood and cognition. They recommend subsequent replication, direct pharmacological intervention and more fully powered designs to probe interactions between systems and potential clinical relevance.

LSD interacts with a complex set of serotonergic and dopaminergic targets beyond the canonical 5-HT2A receptor. Using a receptor-enriched FC approach, the study characterised LSD-induced changes in networks informed by 5-HT1A, 5-HT1B, 5-HT2A, D1 and D2 distributions and linked these to subjective phenomenology: dopaminergic systems were associated with alterations in perceived selfhood and cognition, while serotonergic systems were more closely related to hallucinations and visual imagery. The reported double dissociation—that the same relationships held within but not between the serotonergic and dopaminergic systems—supports the view that multiple receptor systems contribute in distinct ways to the psychedelic state. The authors stress the need for replication, better-powered datasets, time-resolved measurement across the LSD experience, and pharmacological manipulations to provide causal mechanistic insight and to inform translational applications.