That psychedelics profoundly alter experience for a few hours is well established. The harder question, and the decisive one for a clinic, is different: does anything measurable remain, in the tissue, in the wiring of the brain, weeks after the substance itself is long gone? A new study from the Centre for Psychedelic Research at Imperial College London asks exactly that, in people who had never taken a psychedelic before. That naivety is what makes the sample valuable: what shows up is harder to pin on habituation or a rehearsed expectation.
The Design
28 healthy, psychedelic-naive participants. Each served as their own control and went through both conditions: once a high dose of psilocybin (25 mg), once a very low dose (1 mg) as placebo. Before and after the sessions the brain was measured with several methods: EEG for the dynamics of activity, functional MRI for connectivity, diffusion tensor imaging (DTI) for the microstructure of the white matter. The timeline ran from one hour to one month after dosing. A small, exploratory, but methodically dense setup.
What Showed Up After a Month
One month after the high dose, cognitive flexibility was measurably better. It was tested with an established rule-shift task, the IDED test: participants made fewer errors when the underlying rule changed, that is, they could change their minds more easily (p = 0.016). In parallel, psychological insight rose, measured with the Psychological Insight Scale, and so did well-being, measured with the Warwick-Edinburgh scale. Crucially, all of these changes appeared only after the 25 mg, not after the 1 mg placebo. The effect rests on the substance, not on expectation or setting alone.
A Trace in the White Matter
The finest anatomical trace appeared in two fiber tracts that connect the prefrontal cortex to deeper structures: the tract to the striatum and the tract to the thalamus. In both, axial diffusivity, a measure of how directionally water moves along nerve fibers, was reduced a month later. That both tracts share the same prefrontal origin points to a common source for the change. It should be read with caution: DTI is hard to interpret where fibers cross, and the authors themselves urge waiting for replication.
The Network Grows More Permeable
On functional MRI, so-called modularity decreased (p = 0.007). Modularity describes how strongly the brain divides into separate, self-contained networks. Lower modularity means the regions are a little less walled off, a little more globally integrated, more permeable to one another. The striking part is the direction of the relationship. The more modularity fell, the greater the gain in well-being (r = minus 0.40). This is exactly the pattern known from studies in depression. A rigid, over-segmented architecture loosens, and with it something in lived experience seems to loosen too.
Entropy Predicts the Benefit
The thread that may matter most for practice leads back to the acute phase. One to two hours after dosing, the complexity of brain activity on EEG rose, measured as Lempel-Ziv complexity, a measure of informational entropy, peaking at two hours, precisely when the subjective experience was most intense. And that acute richness predicted something: those who showed a more entropic brain in those hours reported more psychological insight the next day and more well-being a month later. Statistically it was next-day insight that carried the arc from acute brain dynamics to later benefit. It is not the trip as such that predicts the outcome, but how richly the brain works in the moment and what the person makes of it in the hours after.
The Most Honest Finding
This is where the study is strong precisely because it does not overstate. Enduring functional brain changes were largely absent. Most of what happens acutely returns to baseline. The authors even suspect that in healthy people there may not yet be a sufficiently sensitive assay to reliably detect lasting functional traces at all, and that such traces are likely clearer in people who are mentally unwell. What remains is subtle: the trace in the white matter and the loosened network architecture. Not a rebuilt machine, more a quietly shifted balance.
Limits, Named Honestly
The study is small and exploratory. These were healthy volunteers, not patients. The two conditions ran in a fixed order, which does not entirely rule out practice or habituation effects, even though the main effects were far larger than any plausible order effect. The primary outcomes were not pre-registered, and statistical correction was done within, not across, all measures. These are correlations over one month, not causal proof and not long-term data. The work points a direction, not a promise.
What We Take From It
For how we think about substance-assisted work, this confirms a principle that shapes how we proceed: the session is not the goal, it is the prompt. If next-day insight is what carries the arc to benefit, then the real work lies in preparation and above all in integration, in the days and weeks in which an experience becomes a change. That is also why we prefer to measure trajectories rather than moments, and work with the person, not only with the molecule.
Substance-assisted approaches belong strictly within a medically supervised, legally permitted setting. This text contextualizes research and is not an invitation to obtain or use any substance.
Source: According to PubMed. Lyons T, Spriggs M, Kerkelä L, Rosas FE, Roseman L, Mediano PAM, Timmermann C, Erritzoe D, Nutt DJ, Carhart-Harris RL et al. (2026). Human brain changes after first psilocybin use. Nature Communications 17(1). DOI: 10.1038/s41467-026-71962-3 (Open Access).