The Common Signature of a Psychedelic Brain: A Mega-Analysis Resolves a Decade of Contradiction

For a decade, the neuroscience of psychedelics has run ahead of its own evidence base. Influential single-site studies described the psychedelic brain as a system whose modular architecture briefly dissolves – networks "desegregate," the default mode network collapses, everything talks to everything. These images were compelling and clinically suggestive, but built on small samples scanned and analyzed in incompatible ways: a field rich in striking claims and poor in replication. This mega-analysis, led from the University of California San Francisco, was designed to test which claims survive when the raw data, not the published conclusions, are harmonized.

The design is the contribution. Rather than pooling effect sizes from papers, the authors gathered subject-level resting-state fMRI from 11 datasets covering five serotonergic psychedelics, reprocessed every brain through one uniform pipeline, and fitted a Bayesian hierarchical model that estimates a common effect while explicitly accounting for differences between drugs and sites. This is the same logic that recently disciplined the PTSD imaging literature: heterogeneity that once manufactured disagreement is absorbed into the model instead of being inherited from incompatible outputs.

The signal that emerged is more specific than the popular "everything connects to everything" picture. The robust, cross-drug effect was increased connectivity between transmodal association networks and unimodal sensory networks – the high-level systems carrying abstract, self-referential and contextual processing become more tightly coupled to raw sensory cortex. This is a structured reorganization, not noise: the brain's hierarchy is flattened in a particular direction. Subcortical regions were not bystanders. The thalamus – the great relay and gate of cortical traffic – along with the striatum and cerebellum, shifted their coupling with sensorimotor networks, implicating gating and prediction circuitry in the acute state.

Equally important is what the model deflated. The headline claim of large within-network disintegration did not hold up uniformly: within-network reductions were weak-to-moderate and selective, varying by drug and network. That is a quieter but more honest result, and it matters clinically. If psychedelic therapeutics work partly by transiently increasing communication between abstract self-models and sensory reality, then the target is a specific reconfiguration of cortical hierarchy, not a wholesale dissolution of brain organization. The probabilistic map this study provides lets the next wave of trials ask precise mechanistic questions instead of re-litigating whether an effect is real.

Why a probabilistic map changes the field

The deeper lesson is methodological. Psychedelic neuroimaging did not lack data; it lacked comparability, and a small literature with large degrees of analytic freedom is fertile ground for irreproducible findings. By reprocessing many cohorts through one pipeline and modeling uncertainty explicitly, the consortium produced something the field has lacked: a calibrated estimate of what is reliable, with the variability across drugs and sites quantified rather than hidden. It is a cornerstone against which future studies can be measured.

What this does and does not establish

This is acute resting-state pharmacology in mostly healthy volunteers, not a model of clinical response. The map describes how psychedelics reorganize large-scale connectivity in the hours of the drug experience; it does not show that this signature predicts who improves from psilocybin- or LSD-assisted treatment, nor does it isolate the durable changes that might mediate therapeutic benefit. For practice the value is interpretive: clinicians and prospective prescribers now have a defensible, cross-drug account of the acute neural mechanism – increased coupling between association and sensory systems with subcortical involvement – to replace the looser metaphors that have circulated, and a clearer hypothesis for what acute brain change a therapeutic protocol is actually inducing.