An international analysis combining brain imaging data from multiple independent studies has identified a common pattern in how psychedelic drugs alter communication between different brain networks. The researchers found that substances such as psilocybin and LSD reliably increase functional connections between brain regions responsible for sensory input and those involved in abstract, associative thought. The findings were published in the journal * Nature Medicine*.
In recent years, classic psychedelic drugs like psilocybin, LSD, and DMT have reentered mainstream psychiatric research as experimental treatments for conditions such as depression and anxiety. These substances all reliably bind to a specific type of serotonin receptor in the brain, initiating profound shifts in perception and consciousness.
To understand how these altered states physically manifest, researchers often use functional magnetic resonance imaging. This noninvasive scanning technique measures spontaneous, coordinated blood flow in the brain. When people simply rest inside a scanner, their brain activity naturally synchronizes into distinct, large scale networks. Some of these networks handle basic sensory and motor tasks, while others manage higher level cognitive functions like memory recall, self-reflection, and goal planning.
Past brain imaging studies of psychedelics have painted a fragmented picture of these network changes. Because collecting data on individuals under the influence of powerful psychoactive drugs is difficult and expensive, most studies rely on small sample sizes. Different research groups also apply different statistical methodologies to their individual datasets. This unchecked variability has caused conflicting reports, with some laboratories finding that specific brain connections increase while others observe opposite effects altogether.
To resolve these inconsistencies, a global team of scientists formed a collaborative consortium to pool existing brain imaging data into a single, standardized analysis. The project was led by neuroscientist Manesh Girn at the University of California, San Francisco and Danilo Bzdok at McGill University, who worked with dozens of independent researchers across three continents.
The research team gathered 11 independent datasets originally collected by different laboratories from five different countries. In total, the analysis included scans from 273 healthy adults who received one of five psychedelic substances: psilocybin, LSD, DMT, ayahuasca, or mescaline.
Instead of looking at each dataset in isolation, the researchers applied a uniform processing pipeline to all the brain scans. This standardization helped eliminate variations caused by different laboratories using disparate software programs to clean and prepare their raw imaging data.
The team then analyzed the structural data using a statistical framework known as Bayesian hierarchical modeling. Traditional frequentist statistics often rely on arbitrary numerical thresholds to declare a biological effect as either present or entirely absent. In contrast, the Bayesian approach calculates a continuous probability that a specific change occurred. This method directly accounts for the variability across different participants, drugs, and original study designs, allowing the scientists to pinpoint the most reliable brain changes while maintaining a graded measurement of uncertainty.
Add PsyPost to your preferred sources The pooled analysis identified a consistent brain signature operating across the different psychedelic substances. The researchers found robust increases in functional connectivity between the brain’s sensory networks and its association networks.
Sensory networks, sometimes called unimodal networks, handle direct, incoming information from the environment, such as visual processing and physical touch. Association networks, often referred to as transmodal networks, are systems like the default mode network and the frontoparietal network. These systems synthesize raw data to support complex thought, memory building, and the brain’s baseline resting state.
Under normal circumstances, these sensory and association systems operate with a heavy degree of separation, maintaining a strict processing hierarchy that keeps basic perception distinct from abstract thought. Under the influence of psychedelics, the lines of communication between them flatten out. The scans showed that these distinct networks synchronize and integrate much more freely during the drug experience.
The researchers also mapped changes deep within the brain’s subcortical structures. Specifically, they looked at the dorsal striatum, a region made up of the caudate and putamen. This area is primarily involved in action selection and linking sensory input to behavioral output. The analysis showed a high probability that the striatum strongly increases its communication with the sensory systems of the cortex during the psychedelic state.
In measuring different drugs against one another, the researchers noted that LSD and psilocybin displayed virtually identical brain network alterations. This overlap aligns with their comparable pharmacological properties and the large number of participants included for both substances in the pooled data. Mescaline exhibited a broadly similar pattern to LSD and psilocybin in terms of network merging.
The substance DMT caused similar architectural shifts but with even stronger network perturbations. Ayahuasca presented an idiosyncratic pattern in the statistical models, which the authors attribute to its complex pharmacology and the extremely small number of participants who were scanned using that exact substance.
The collective data also challenged a popular idea in the neuroimaging field. Previous single laboratory studies frequently reported that psychedelics cause the brain’s individual functional networks to break down internally, a phenomenon often described as within-network disintegration. When the researchers averaged the data across all the pooled studies, this effect proved incredibly weak. The Bayesian analysis revealed very little statistical certainty for reductions in connectivity within specific networks, putting those earlier network breakdown claims into question.
Expanding on the biological findings requires consideration of several limitations inherent in combining historical data. The original studies were conducted over a span of several years and used different magnetic resonance imaging scanners, which varied heavily in their magnetic field strengths, recording intervals, and technical specifications.
The participants across the different datasets also received varying drug dosages, encountered different administration methods, and were scanned at wildly different chronological points during their respective drug experiences. Some participants received intravenous injections, while others swallowed capsules. Some researchers began brain scans immediately after administration, while other scientists waited up to two hours for the subjective effects to peak.
Differences in simple study design play a role in interpretation as well. The vast majority of the datasets relied on trials that actively compared the drug experience to a placebo control session. However, one dataset lacked a placebo entirely, and another study used a fixed-order design in which the sequence of drug and placebo conditions was not randomized. Such differences can introduce small biases related to participant expectation and novelty.
Head motion remains a persistent challenge in this area of neuroscience. Individuals experiencing the subjective effects of psychedelics have a documented tendency to move around more inside the scanner than those sitting still after receiving a placebo. While the data processing pipelines were formulated to minimize the impact of participant movement on the results, residual visual noise in the imaging data remains unavoidable.
To build on this foundational map, the scientists suggest that future research should abandon retrospective pooling and move toward prospectively harmonized trials. In these future projects, multiple laboratories would agree to use identical protocols for drug dosing, participant selection parameters, and brain scanning machine configurations before any data is even gathered.
The study, “An international mega-analysis of psychedelic drug effects on brain circuit function,” was authored by Manesh Girn, Manoj K. Doss, Leor Roseman, Katrin H. Preller, Fernanda Palhano-Fontes, Lorenzo Pasquini, Frederick S. Barrett, Pablo Mallaroni, Natasha L. Mason, Christopher Timmermann, Drummond E. McCulloch, Patrick M. Fisher, Brian S. Winston, Flora Moujaes, Felix Muller, Matthias E. Liechti, Franz X. Vollenweider, Johannes G. Ramaekers, Kim Kuypers, Draulio B. Araujo, Olaf Sporns, Joshua Siegel, Nico Dosenbach, David J. Nutt, Robin L. Carhart-Harris, Emmanuel A. Stamatakis, and Danilo Bzdok.