Psilocybin Rewires Brain Networks Through Activity-Dependent Synaptic Plasticity

Depression is characterized by reduced excitatory synaptic density in the prefrontal cortex, and fast-acting treatments like ketamine and psilocybin are known to promote dendritic spine growth in frontal cortical neurons. However, identifying which presynaptic neurons actually wire into those new spines—and whether psilocybin changes the balance of brain network inputs—had remained unknown. This study addressed that gap directly, providing the first brain-wide map of psilocybin's effect on synaptic connectivity.

The team employed an engineered, EnvA-pseudotyped, glycoprotein-deleted rabies virus combined with Cre-dependent AAV helper viruses to perform monosynaptic retrograde tracing in two inducible mouse lines: Fezf2-2A-CreER (targeting subcortical-projecting pyramidal tract, PT neurons) and PlexinD1-2A-CreER (targeting intratelencephalic, IT neurons). Psilocybin or saline was given one day before rabies injection into the dorsal medial frontal cortex, so that labelled inputs would reflect drug-evoked—not pre-existing—connectivity. Brains were processed via tissue clearing, light-sheet fluorescence microscopy, and machine-learning nuclei detection, yielding counts across 316 Allen Brain Atlas regions, with ~500,000 input cells mapped per animal.

Psilocybin reorganized PT neuron inputs in a highly network-specific pattern. Inputs from sensorimotor, visual-auditory, and medial (default-mode homolog) networks increased substantially, while inputs from the lateral network, ventromedial prefrontal cortex (ILA, ORBm), lateral insular cortex, hippocampus, and mediodorsal thalamus decreased. This network selectivity was statistically robust (chi-squared test, P = 6×10⁻⁵). Strikingly, IT neurons showed the opposite pattern—gaining inputs from the ventromedial prefrontal and lateral networks, while losing inputs from sensorimotor sources—demonstrating cell-type-specific bidirectional reorganization within the same brain region.

To test whether psilocybin's acute effect on firing rates drives this structural rewiring, the researchers combined Neuropixels electrophysiology with chemogenetic silencing (DREADDs). Psilocybin increased spiking in regions whose input fraction subsequently rose, and decreased activity in regions whose inputs fell. Chemogenetically silencing the retrosplenial cortex (RSP)—which normally shows increased firing and increased input fraction after psilocybin—during drug administration specifically abolished the RSP input gain, without affecting other regions. This demonstrates that activity-dependent plasticity, triggered by acute drug-evoked firing, is the mechanistic driver of the network rewiring. Slice electrophysiology confirmed lasting changes in synaptic strength consistent with LTP-like mechanisms.

Therapeutically, the weakening of recurrent cortico-cortical loops is particularly compelling: hyperconnected default-mode and frontal-parietal recurrent circuits are implicated in rumination and depression. Psilocybin's selective disruption of these loops, combined with strengthening of sensory and medial network inputs, may underlie both the acute psychedelic experience and the lasting antidepressant benefit. The study also demonstrates that chemogenetic or other neuromodulatory interventions co-administered with psilocybin could be used to sculpt which circuits are strengthened or weakened, opening a path toward precision psychedelic therapy.