TL;DR
This paper presents the first comprehensive "tripartite hierarchy" of psychedelic effects, linking molecular pharmacology, whole-brain neuroimaging, and subjective experience across DMT, LSD, and psilocybin. The researchers found that while these drugs share a common "neural fingerprint" of brain network disintegration, the relationship between receptor binding and the actual "trip" is surprisingly non-linear and complex.
What Problem Does This Paper Address?
Despite the "psychedelic renaissance" in clinical research, we still lack a unified model of how these substances work across different scales of biological organization. We know they hit serotonin receptors (pharmacology), we know they change brain connectivity (neuroimaging), and we know they induce altered states (phenomenology)—but how do these levels actually relate to one another?
The literature is currently a "Wild West" of fragmented data, inconsistent analysis techniques, and small sample sizes. This paper attempts to synthesize decades of disparate research into a single, cohesive framework to identify the specific "fingerprints" of classic psychedelics and determine if one level of the hierarchy can reliably predict the others.
Methods
The authors conducted three parallel systematic reviews and meta-analyses:
- Phenomenology: They pooled scores from the 5-Dimensional and 11-Dimensional Altered States of Consciousness (ASC) scales.
- Neuroimaging: They used GingerALE (Activation Likelihood Estimation) to find common clusters of BOLD activation and developed a novel in-house algorithm to aggregate pairwise functional connectivity (FC) data across different brain parcellations, mapping them all to the standard 7-network Yeo atlas.
- Pharmacology: They performed a meta-analysis of binding affinities ($K_i$) and functional activity ($EC_{50}$) at various serotonin and dopamine receptors in both humans and rodents.
Finally, they attempted to "align" these levels by weighting receptor maps and connectivity changes to see if they shared a common spatial structure.
Datasets
The study synthesized data from a massive swath of existing literature:
- Neuroimaging: 58 fMRI studies (though only 22 collected original data; the rest were secondary analyses).
- Phenomenology: 47 studies using ASC scales for DMT, LSD, and psilocybin.
- Pharmacology: 70 studies covering receptor affinities and functional assays.
- Reference Data: They used the Human Connectome Project (HCP) resting-state data as a null distribution and a state-of-the-art PET atlas of receptor densities to map pharmacological effects onto brain anatomy.
Key Insights and Ideas
- The Non-Linear Hierarchy: The most striking finding is the "weak one-to-one correspondence" between levels. You cannot simply look at a drug's receptor binding profile and perfectly predict the resulting brain activity or the subjective experience. The emergence of a "trip" is a highly non-linear process.
- LSD’s Unique Profile: LSD stands out for having binding affinities 1–2 orders of magnitude higher than DMT or psilocin across the board. Notably, its moderate affinity for dopamine receptors (which DMT and psilocybin lack) may explain why it produces significantly more intense "visionary restructuralisation" (visual hallucinations).
- Challenging the DMN "Meme": While popular science often focuses on psychedelics "shutting down" the Default Mode Network (DMN), the authors found that the story is more about disintegration and desegregation. Psychedelics don't just lower DMN activity; they blur the boundaries between all major networks, increasing connectivity between systems that usually stay separate.
- The "Visual" Bias: The meta-analysis revealed that psilocybin’s neuroimaging signature is heavily concentrated in the visual network. However, the authors warn this might be an artifact: most psilocybin fMRI studies involved visual tasks, whereas LSD studies were often resting-state.
Key Results
- Subjective Experience: LSD yielded significantly higher ratings for visual hallucinations than psilocybin ($p = 0.0126$).
- Brain Connectivity: Across all drugs, there was a significant increase in between-network connectivity and a decrease in within-network connectivity. The most dramatic shifts occurred in the Frontoparietal (FPN) and Visual networks.
- Pharmacology: LSD showed significantly higher potency in inducing inositol phosphate (IP) formation at the $5-HT_{2A}$ receptor compared to DMT.
- Data Reuse Warning: The authors highlighted a major "risk of bias" in the field: over a third of all psychedelic neuroimaging papers are secondary analyses of the same two primary datasets (Carhart-Harris et al., 2012 and 2016).
Limitations and Future Work
The primary limitation is the heterogeneity of the source data. Different labs use different brain "maps" (parcellations) and preprocessing pipelines, making direct comparison difficult. The authors also noted a lack of high-quality fMRI data for DMT compared to LSD and psilocybin.
Future Directions: * Standardization: The authors urge the field to adopt unified pipelines like fMRIPrep and standard parcellations (like Schaefer-232) to make future meta-analyses more robust. * Causal Modeling: Moving beyond correlations to use "whole-brain models" that can simulate how changing a single receptor's activity ripples up through the hierarchy to change consciousness. * Dose-Response Studies: Most current data comes from single, high doses. We need more research on how these hierarchical fingerprints change across a range of doses in the same individuals.
Summary updated