Psychedelic effects of psilocybin correlate with serotonin 2A receptor occupancy and plasma psilocin levels

Psilocybin is the principal psychoactive compound in magic mushrooms and a classic serotonergic psychedelic whose subjective effects resemble those of LSD and mescaline. Recent clinical trials have reported therapeutic potential for psilocybin across several neuropsychiatric conditions, including treatment-resistant major depressive disorder, cancer-related anxiety and depression, and substance use disorders. Pharmacological and preliminary imaging studies have implicated the serotonin 2A receptor (5-HT2AR) as the principal molecular mediator of psychedelic effects, but direct human evidence linking psilocybin's active metabolite (psilocin) plasma concentrations, cerebral 5-HT2AR target engagement and the subjective psychedelic experience was lacking. D. and colleagues set out to characterise, in humans, the relationships among plasma psilocin levels, cerebral 5-HT2AR occupancy and subjective intensity of psychedelic effects. Using the 5-HT2AR agonist PET radioligand [11C]Cimbi-36, the study aimed to quantify receptor occupancy after oral psilocybin and to model how occupancy and psilocin concentrations relate to concurrent self-reported intensity and to standard psychedelic experience questionnaires administered after the imaging sessions.

Design and participants: Eight healthy volunteers (three females, mean age 33.0 ± 7.1 years) were recruited for a within-subject neuroimaging study. Screening included psychiatric and medical assessment (Mini-International Neuropsychiatric Interview and physical/ECG/blood screening) and urine drug testing. Five participants reported prior serotonergic psychedelic use (median 1 occasion, range 0–55), with a median 42 months since last use. Participants met additional exclusion criteria reported in the supplementary material. Ethical approvals and informed consent were obtained. Imaging and procedural overview: Each participant completed a baseline [11C]Cimbi-36 PET scan (PET 0) and MRI prior to the psilocybin day (mean 49 ± 12 days earlier). On the intervention day participants ingested a single oral psilocybin dose (range 3–30 mg in 3 mg capsules) about one hour before the first post-drug PET scan (PET 1). Five participants had a second post-drug PET scan later the same day (PET 2; mean 344 ± 41 min after ingestion); three participants had only PET 1. During all PET scans a standardised music playlist was played and two psychologists provided interpersonal support. Each PET acquisition lasted 120 minutes. Pharmacology and blood sampling: Venous blood samples for plasma psilocin concentrations were taken at the time of radioligand injection and at 20-minute intervals throughout each PET session. Psilocin concentrations were determined by ultra performance liquid chromatography with tandem mass spectrometry; data are reported in μg/L (analysis conducted in μg/kg but presented as μg/L). For one participant (Subject 1) some concentrations were below the limit of quantification (LOQ) but above the limit of detection (LOD); values were handled as described in the analysis section. PET acquisition and outcome measures: High-resolution T1- and T2-weighted MRI were acquired for PET coregistration. [11C]Cimbi-36 PET data were acquired for 120 minutes on a high-resolution research tomography scanner. Regions of interest were defined automatically and neocortex (volume-weighted average of all cortical regions) was chosen a priori because of high 5-HT2AR expression. Cerebellum served as the reference region. Kinetic modelling used the simplified reference tissue model (SRTM) and nondisplaceable binding potential (BP_ND) was the primary PET outcome. The authors report that psilocybin did not affect [11C]Cimbi-36 radiometabolism or protein binding (supplementary data). Subjective measures and questionnaires: During scans, participants rated subjective psychedelic intensity on a 0–10 Likert scale at 20-minute intervals until effects waned. After the last scan session (mean 468 ± 80 min post-dose) participants completed Danish translations of the MEQ30 (mystical experiences questionnaire), the 11-dimension altered states of consciousness questionnaire (11D-ASC) and the Ego-Dissolution Inventory (EDI). Questionnaires were administered after imaging to avoid suggestive effects and because intensity ratings provided better temporal resolution. Data analysis and modelling: Within-scan mean psilocin concentration (psilocin AUC normalized by 120 min) was calculated using the trapezoid method in GraphPad Prism. The relation between neocortical BP_ND (expressed as percent occupancy relative to baseline) and mean psilocin concentration was modelled using a single-site binding function: Occ = Occ_max × C_P/(EC50 + C_P), fitted in GraphPad Prism. Associations between intensity ratings and occupancy were modelled with a quadratic function (Intensity = β1 × occupancy + β2 × occupancy^2), and intensity versus psilocin used a stimulus-response function analogous to the occupancy model (Intensity = Intensity_max × C_P/(EC50 + C_P)). Non-linear modelling yielded 95% confidence intervals; other statistical tests were conducted in R. For PET 2 when intensity recordings were discontinued, participants were asked whether intensity had changed; all reported no change, so the last recorded score was extrapolated for mean PET 2 intensity. The authors report additional post hoc and exploratory analyses, and voxel maps were generated for visualisation only.

Receptor occupancy: Psilocybin administration produced substantial, dose-related neocortical 5-HT2AR occupancies. At the first post-dose scan (PET 1) occupancies ranged from 43% to 72%. For PET 2 occupancies were generally lower (27–47%) except for one participant (Subject 1) whose PET 2 occupancy was 2%. Psilocin pharmacokinetics and occupancy model: There was notable inter-individual variability in plasma psilocin pharmacokinetics; reported C_max median [range] was 11.9 [2.3–19.3] μg/L. The relation between mean plasma psilocin concentration and neocortical 5-HT2AR occupancy fit a single-site binding model well. From the model Occ_max was 76.6% (95% CI 67.3–88.0%), EC50 was 1.95 μg/L (95% CI 1.17–3.15 μg/L), and the model R2 was 0.92. Subjective intensity and its associations: Time courses of mean subjective intensity ratings were qualitatively similar to plasma psilocin curves. A non-linear stimulus-response model for intensity versus psilocin yielded an Intensity_max of 10.8 and an EC50 of 4.5 μg/L (95% CI 2.1–9.8 μg/L), with R2 = 0.35. Intensity ratings were also positively associated with occupancy; this association was well described by a quadratic model (β1 = -0.02, 95% CI -0.13 to 0.10; β2 = 0.002, 95% CI 0.0006 to 0.003), with R2 = 0.81. Questionnaire outcomes: Psilocybin produced pronounced changes on psychedelic experience questionnaires. The MEQ30 total score median [range] was reported as 2.9 (post-hoc association with mean PET 1 intensity: p_unc. = 0.02, Bonferroni-adjusted p_FWE = 0.06, R2 = 0.61). Post-hoc linear regressions showed positive associations between mean PET 1 intensity and MEQ30, global 11D-ASC score and EDI score. Additional points and handling of data quirks: For Subject 1 some psilocin concentrations were below the LOQ (0.5 μg/kg) but above the LOD (0.1 μg/kg); the authors set those time points to a mean value of 0.3 μg/kg for modelling—sensitivity checks using LOQ/LOD yielded similar model parameters. Per-participant EC50 estimates averaged PET 1 = 4.5 ± 1.9 μg/L and PET 2 = 6.2 ± 6.0 μg/L; these did not differ statistically (paired t-test mean difference -1.7, 95% CI -10.2 to 6.7, p = 0.6). The authors report no systematic effect of psilocybin on radioligand metabolism or protein binding. Motion during PET did not differ in median terms from baseline, although two participants had maximum movements up to 35 mm and 20 mm during PET 1; model fits remained acceptable.

The authors interpret their findings as demonstrating that oral psilocybin (3–30 mg) produces dose-dependent occupancies of cerebral 5-HT2ARs measurable with an agonist PET radioligand, and that plasma psilocin levels and 5-HT2AR occupancy are closely associated with concurrent subjective intensity of the psychedelic experience. The occupancy–concentration relation conformed to a single-site binding model with an Occ_max of ~77% and an EC50 of 1.95 μg/L (≈10 nM), values the authors note are comparable to in vitro affinity estimates for 5-HT2AR agonists. The authors emphasise that using an agonist radioligand like [11C]Cimbi-36 may preferentially label receptors in a high-affinity state and could therefore provide a biologically relevant estimate of functionally engaged receptors. Their data show that plasma psilocin closely tracks both receptor occupancy and subjective intensity, suggesting that measuring psilocin concentrations could allow estimation of 5-HT2AR occupancy without PET. The observed inter-individual pharmacokinetic variability implies that clinical trials might benefit from relating clinical outcomes to psilocin levels or estimated occupancy rather than to absolute dose. The authors discuss implications for microdosing, noting that one participant who received 3 mg (0.05 mg/kg) experienced noticeable perceptual effects with 43% occupancy, suggesting that microdoses producing no perceptual effects would need to be smaller than 3 mg. They propose 0.5–2 mg as a reasonable range for potential microdosing studies based on their data, while acknowledging that definitive cut-offs are not established. Several limitations are acknowledged. The fitted Occ_max <100% (≈77%) could reflect kinetic-modeling violations, rapid receptor internalisation, psilocin-induced reductions in brain 5-HT, or radioligand sensitivity to receptor state, any of which could bias occupancy estimates. Psilocin also has weaker affinities at other serotonin receptors (5-HT2B, 5-HT2C, 5-HT1A) and low affinity for SERT; these off-target interactions could influence interpretation. The small sample size (n = 8), predominance of males, narrow age range and recruitment of participants specifically interested in a psilocybin neuroimaging study limit generalisability. The PET environment and the open-label nature of the intervention (no placebo control) may have influenced subjective ratings. Motion artefacts in two participants and the fact that some subjective ratings during PET 2 were extrapolated are noted as further constraints. Finally, the authors cannot exclude effects of psilocin metabolites or expectation-induced serotonin changes on occupancy estimates. Overall, D. and colleagues conclude that their human data strongly support 5-HT2AR stimulation by psilocin as a key determinant of psychedelic effects, and they propose that plasma psilocin measurement could serve as an objective proxy for estimating brain 5-HT2AR occupancy in future studies.