Psilocybin Improved Depression Without a Detectable Rise in SV2A Binding
A pre-registered PET sub-study has pushed the biologic question behind psilocybin one step closer to the clinic, and one step away from an easy synaptic explanation. In 35 participants with moderate-to-severe recurrent major depressive disorder, a single 25 mg dose of psilocybin produced the expected rapid antidepressant benefit, yet imaging with [11C]UCB-J did not show a group-level increase in SV2A binding in the hippocampus or frontal cortex when compared with 100 mg niacin placebo plus psychotherapeutic support.
The result matters because SV2A PET is widely used as a proxy for synaptic density. If psychedelic antidepressant effects depended on a broad, measurable increase in synapses over the first two weeks after dosing, this study should have detected it. Instead, the imaging signal stayed flat at the group level even as clinical scores improved. In practical terms, the antidepressant response appears to emerge without a detectable change in this synaptic marker, at least on the timeline and in the brain regions examined here.
The design was straightforward and clinically relevant. Participants were randomized 1:1, underwent [11C]UCB-J PET scans 2 to 7 days before dosing, and repeated imaging 15 ± 7 days afterward. The trial remained anchored to a recently published randomized study in which psilocybin had already shown rapid and sustained antidepressant effects, so this analysis focused on mechanism rather than efficacy alone. By keeping the imaging window close to the dosing period, the study tested a biologically plausible claim: if psilocybin rapidly remodels synapses in humans, the signal should be visible early.
It was not.
Baseline SV2A Binding Carried More Signal Than Change Over Time
The more interesting finding came from the baseline scan. Higher pretreatment VT was associated with greater subsequent clinical improvement in the psilocybin arm, and this association survived false discovery rate correction at Day 8 and Day 15, two of the three pre-registered timepoints. In contrast, change in VT showed associations with later symptom improvement only in unadjusted analyses; once baseline VT and baseline MADRS were included, those relationships weakened and did not survive correction for multiple comparisons.
That pattern points away from a simple “more synapses after treatment equals better response” model. It suggests that the pre-existing synaptic state may shape who benefits most from psilocybin. A higher baseline SV2A signal could mark a brain that still has enough structural reserve, circuit integrity, or plasticity potential to translate the psychedelic experience into durable symptom relief. In other words, the baseline substrate may matter more than a short-term post-dose rise that never materializes at the group level.
The biology is plausible. Depression has long been linked to synaptic loss and reduced connectivity in mood-related circuits, especially in regions such as the hippocampus and frontal cortex. Psychedelics are known from preclinical work to induce synaptic remodeling within 24 hours, which made synaptic density a credible human biomarker to test. Yet a failure to detect a group-level rise in SV2A binding does not exclude plasticity altogether. It may mean the relevant change is too subtle for the tracer, occurs in other circuits, unfolds outside the sampling window, or reflects receptor-level and network-level reorganization rather than a wholesale increase in synapse number.
What the Imaging Signal Can and Cannot Explain
The study leaves a clear message for translational psychiatry: symptomatic improvement after psilocybin does not require a measurable increase in SV2A binding on the scale captured here. That finding narrows the search for mechanism. The antidepressant effect may depend on rapid circuit recalibration, altered neuronal synchrony, or synaptic strengthening without enough change in total synaptic density to move [11C]UCB-J PET.
The clinical consequence is more immediate than the mechanistic uncertainty. If baseline SV2A binding eventually proves reproducible as a predictor of response, PET imaging could help identify which patients are most likely to benefit from psilocybin-assisted therapy. That would matter in a treatment model that is resource-intensive, psychologically demanding, and still under active study. A selection biomarker could reduce exposure for likely non-responders and focus treatment on patients with the greatest chance of meaningful improvement.
Still, the data remain preliminary. This was a preprint, not a peer-reviewed report, and the sample was small. The imaging was limited to hippocampus and frontal cortex, and the study cannot rule out region-specific or time-dependent synaptic changes elsewhere in the brain. The analysis also compared psilocybin with niacin placebo, which preserves blinding better than an inert control in psychedelic research, but does not erase all expectancy effects created by the psychotherapy-supported setting.
The strongest conclusion is also the most restrained: psilocybin can produce rapid antidepressant benefit without an accompanying group-level rise in SV2A binding, while baseline synaptic marker levels may carry prognostic information. Before that signal can be used at the bedside, it will need replication in larger cohorts, across additional brain regions, and with imaging schedules that can capture faster or more localized plasticity if it exists at all.