Acute psilocybin and ketanserin effects on cerebral blood flow: 5-HT2AR neuromodulation in healthy humans

Abstract

Psilocin, the active metabolite of psilocybin, is a psychedelic and agonist at the serotonin 2A receptor (5-HT2AR) that has shown positive therapeutic effects for brain disorders such as depression. To elucidate the brain effects of psilocybin, we directly compared the acute effects of 5-HT2AR agonist (psilocybin) and antagonist (ketanserin) on cerebral blood flow (CBF) using pseudo-continuous arterial spin labeling magnetic resonance imaging (MRI) in a single-blind, cross-over study in 28 healthy participants. We evaluated associations between plasma psilocin level (PPL) or subjective drug intensity (SDI) and CBF. We also evaluated drug effects on internal carotid artery (ICA) diameter using time-of-flight MRI angiography. PPL and SDI were significantly negatively associated with regional and global CBF (∼11.6% at peak drug effect, p < 0.0001). CBF did not significantly change following ketanserin (2.3%, p = 0.35). Psilocybin induced a significantly greater decrease in CBF compared to ketanserin in the parietal cortex (p_FWER < 0.0001). ICA diameter was significantly decreased following psilocybin (10.5%, p < 0.0001) but not ketanserin (−0.02%, p = 0.99). Our data support an asymmetric 5-HT2AR modulatory effect on CBF and provide the first in vivo human evidence that psilocybin constricts the ICA, which has important implications for understanding the neurophysiological mechanisms underlying its acute effects.

Introduction

Psilocybin has emerged as a promising therapeutic showing rapid and lasting reductions in clinical symptoms across several brain disorders, including major depression,^1–3 cancer related anxiety and depressive symptoms,^4,5 and cigarette and alcohol addiction.^6,7 A psychedelic dose of psilocybin induces subjective effects lasting 6–8 hours characterised by visual alterations, profound mood changes and shifts in the perception of space, time and selfhood.^8,9 Psilocybin has also been shown to produce lasting changes in self-reported measures including increased personality trait openness, decreased neuroticism, increased mindfulness and life-satisfaction in healthy individuals.^10–17 After intake, psilocybin is dephosphorylated to its active metabolite, psilocin (4-OH-N,N-dimethyltryptamine), ¹⁸ which is a partial agonist at several serotonergic receptors, including the serotonin 2A receptor (5-HT2AR). ¹⁹ Accumulating evidence shows 5-HT2AR stimulation underlies the acute psychedelic effects,^20–25 e.g., pre-treatment with ketanserin blocks the acute psychedelic effect of psilocybin, ²¹ and other classical psychedelics, ²² and ketanserin administration reverses acute psychedelic effects following lysergic acid diethylamide (LSD). ²⁶ Neuronal 5-HT2ARs are predominantly expressed on the apical dendrites of excitatory pyramidal neurons in the cortical layers III and V, GABAergic neurons, and monoaminergic terminals.^27–29 Notably, 5-HT2ARs are also expressed on smooth muscle cells in blood vessels, where they can induce vasoconstriction and consequently increase blood pressure.^30–32 In vitro studies confirm that 5-HT2AR-induced arterial contractions are abolished by antagonists like ketanserin, which acts as a vasodilator in the presence of serotonin.^33–37 Understanding the in vivo neurovascular effects of 5-HT2AR modulation could enhance our knowledge of how psychedelics affect brain function.

Recent blood oxygen level-dependent functional magnetic resonance imaging (BOLD fMRI) and electroencephalography (EEG) studies of human brain function show substantial acute effects of psychedelics on functional brain connectivity. Most consistently, resting-state BOLD fMRI scans during the acute experience increases global connectivity, particularly between transmodal resting-state networks.^38,39 Estimating cerebral blood flow (CBF) with arterial spin labelling (ASL) can provide complementary information related to neural signalling that can be estimated in absolute units (ml/100 g/min) and may inform neural correlates of psychedelic drug effects. ⁴⁰ Acute psilocybin effects on CBF have been previously reported in three studies using ASL fMRI.^41–43 Carhart-Harris and colleagues ⁴¹ reported decreased CBF up to 12 minutes following intravenous administration of 1–2 mg of psilocybin in 15 healthy participants. Lewis and colleagues ⁴² reported decreased CBF approximately 60 minutes following oral psilocybin administration of 0.16 or 0.215 mg/kg in 58 healthy participants, an effect echoed in a follow-up analysis including 12 additional participants. ⁴³ Lewis and colleagues also considered an alternative CBF quantification strategy using regional relative hyperperfusion, which they argued aligns with results following psychedelic administration using single photon emission computed tomography (SPECT) and positron emission tomography (PET) studies of CBF and glucose metabolism, respectively.^44,45 Contrasting acute psilocybin effects with a 5-HT2AR antagonist would contribute to our understanding of how 5-HT2AR modulation shapes neurovascular function.

In this randomised, single-blind, cross-over study in healthy participants, we contrasted the acute effects of a single oral dose of psilocybin or ketanserin on CBF with pseudo-continuous ASL (pcASL). We mapped changes in CBF at several time points after psilocybin administration. To more comprehensively characterise cerebrovascular effects, we examined acute drug effects on internal carotid artery (ICA) diameter. We evaluated the association between CBF and plasma psilocin levels (PPL) and subjective drug intensity (SDI), which are related to brain 5-HT2AR receptor occupancy and functional connectivity.^46,47 Together, these analyses help resolve acute 5-HT2AR related neurophysiological effects. Based on previous studies, we hypothesized that psilocybin would reduce CBF and induce constriction of the ICA, stemming from its 5-HT2AR agonist effects, whereas the 5-HT2AR antagonist ketanserin would have limited effects on CBF and ICA diameter.

Methods and materials

Participants

Twenty-eight healthy volunteers participated in this study (10 females, age (mean ± SD): 33 ± 8 years).

Participants were recruited from a database of individuals interested in participating in a brain imaging study involving the administration of psychedelics. BOLD fMRI data unrelated to the current study and from a subset of participants described herein has been reported in previous studies.^8,17,47–49 Here, we present the pcASL data for the first time. Recruitment began in September 2018 and ended in March 2021. Forty-four individuals were assessed for eligibility, of which 16 were excluded (see Figure 1). Full details of the exclusion criteria have been reported previously. ⁴⁶ Before inclusion, participants received written and verbal descriptions of the study protocol, including the expected effects of psilocybin and potential side effects of psilocybin and ketanserin; all participants provided written consent prior to study participation in accordance with The Declaration of Helsinki II. The data presented here is part of a broader study approved by the ethics committee for the Capital Region of Copenhagen (journal identifier: H-16028698, amendments: 56023, 56967, 57974, 59673, 60437, 62255) and Danish Medicines Agency (EudraCT identifier: 2016-004000-61, amendments: 2017014166, 2017082837, 2018023295).

Figure 1.

CONSORT flow diagram. This diagram shows the enrolment and allocation of participants in a study evaluating the effects of psilocybin and ketanserin on cerebral blood flow. 44 participants were assessed for eligibility, with 28 being randomised into the study. 17 participants received psilocybin first and 11 received ketanserin first, with a minimum 3 week washout period in between. All 28 participants completed the psilocybin intervention, whereas 4 participants failed to complete the ketanserin session.

Study design

This study used a single-blind, within-subject, cross-over design with two drug intervention sessions: (i) psilocybin and (ii) ketanserin. Seventeen participants were assigned to receive psilocybin first, 11 assigned to ketanserin first. The allocation was generated by a staff member who was not otherwise involved in the project. Drug sessions were separated by at least three weeks (mean: 31 days, median: days, range: [21–105]; time between sessions was more than 42 days for only one participant due to delay following the Covid-19 lockdown). Additional details about the overall study design have been described previously. ⁴⁷

Study procedures

Participants were asked to refrain from alcohol the night before intervention days, to eat a light breakfast and to refrain from caffeine intake on the day of intervention. Before drug administration and pcASL data acquisition, a urine sample was obtained and screened for amphetamines, opioids, benzodiazepines, barbiturates, tetrahydrocannabinol, cocaine, ketamine, phencyclidine, and gamma-hydroxybutyrate (Rapid Response, BTNX Inc). Participants met with two psychologists before the intervention day for preparation (approx. one week) and again the day after the psilocybin intervention for integration of the experience. On intervention days, between scans, participants retired to a comfortable and pleasant room adjacent to the MR scanner, where music was played. At least one psychologist was present with the participant to provide interpersonal support throughout the psilocybin intervention, including during MR scanning.

Participants completed a pre-drug MRI scan including structural, pcASL, and angiography scans (detailed below). The pre-drug scan was followed by a brief conversation with the psychologists, after which drug was administered. We administered psilocybin according to body weight in units of 3 mg capsules (0.26 ± 0.04 mg/kg, absolute dose 19.8 ± 3.7 mg, mean ± SD) and ketanserin as a single 20 mg tablet. These doses were selected to achieve a comparable peak 5-HT2AR occupancy 1–3 hours after administration of approximately 70% (see Supplementary Figure S1).^46,50,51 Drugs were given with a glass of water. On psilocybin days, pcASL scans were acquired at approximately 40, 80, 130, and 300 minutes post-drug to capture both peak drug effects as well as the subsequent decline. Blood samples for PPL and SDI assessments were taken after each scan. Participants were sent home once SDI levels dropped below 1 after the final scan. On ketanserin days, pcASL scans were acquired before and at approximately 80 minutes post-drug. While not in the scanner, participants were listening to music in the adjacent room. On ketanserin days, participants with an SDI < 1 after 130 minutes completed self-report questionnaires and were sent home following a short debriefing.

Plasma psilocin level

PPL was measured in venous blood samples drawn from a catheter in the antecubital vein after each pcASL scan. The blood samples were centrifuged and the plasma fraction frozen and briefly stored at −20°C before being transferred to −80°C until analysis. Plasma psilocin levels were measured using ultra performance liquid chromatography and tandem mass spectrometry, as previously described, ⁴⁶ reflecting free, unconjugated psilocin and not glucuronidated psilocin. ⁵² Plasma psilocin is reported in units µg psilocin per L plasma (µg/L).

Subjective drug intensity

Following each pcASL scan, an SDI rating was acquired by asking the participant: “How intense is your experience right now on a scale from 0 to 10?” (Likert scale, 0 = not intense at all, 10 = very intense).

Magnetic resonance imaging

Participants were scanned on one of two 3 Tesla Siemens Magnetom Prisma MRI scanners (Siemens AG, Erlangen, Germany) at Rigshospitalet, Copenhagen (MR1 and MR2). Each participant completed all their scans on only one MRI scanner. Fifteen participants completed scan sessions on MR1 and thirteen participants on MR2. All images were visually inspected for quality using the FSL image viewer, FSLeyes. ⁵³

Magnetic resonance imaging parameters

MR1

Related data from the participants scanned on this scanner has been reported previously.^47,48 Imaging data were acquired using a 64-channel head/neck coil. During the pre-drug scan we acquired a high-resolution, T1-weighted 3D MPRAGE structural image (inversion time = 900 ms, echo time = 2.58 ms, repetition time = 1900 ms, flip angle = 9°, in-plane matrix = 256 × 256, in-plane resolution = 0.9 mm ×0.9 mm, number of slices: 224, slice thickness = 0.9 mm, no gap). CBF was estimated using a 2D echo-planar imaging (EPI) gradient spin-echo sequence (scan time = 5.33 min, echo time = 12 ms, repetition time = 4000 ms, flip angle = 90°, label duration: 1508 ms, single post-labeling delay = 1500 ms, in-plane matrix =64 × 64, in-plane resolution = 3 mm × 3 mm, number of slices = 20, slice thickness = 5 mm no gap, number of control/label pairs = 40, slice readout duration =35 ms). To calibrate the pcASL signal, an M0 scan was acquired using the same imaging parameters as the pcASL except for a repetition time of 10000 ms. We acquired a gradient echo sequence (GRE) to correct for geometric distortions in the pcASL images (echo time 1 = 4.92 ms, echo time 2 = 7.38 ms, repetition time = 400 ms).

MR2

Imaging data were acquired using a 32-channel head coil. During the pre-drug scan, we acquired a high-resolution, T1-weighted 3D MPRAGE structural image (inversion time = 920 ms, echo time = 2.41 ms, repetition time = 1810 ms, flip angle = 9°, in-plane matrix = 288 × 288, in-plane resolution = 0.8 × 0.8 mm², slice thickness = 0.8 mm, number of slices = 224, no gap). CBF was estimated using a five post-label delay 3D turbo gradient spin echo sequence (scan time = 7.11 min, 12 label/control pairs, native voxel size: 2.5 × 2.5 × 3 mm³, image matrix = 96 × 96 × 40, label duration = 1508 ms, post-label delays = [500, 500, 1000, 1000, 1500, 1500, 2000, 2000, 2000, 2500, 2500, 2500] ms, echo time = 3.78 ms, repetition time = 4100 ms, flip angle: 120°), two background suppression pulses optimised for each PLD. The acquisition included an M0 scan with which to calibrate the pcASL signal without any background suppression. Spin-echo echo-planar acquisitions with opposite phase-encoded blips were acquired to correct for geometric distortions using topup implemented in FSL (FMRIB software library, www.fmrib.ox.ac.uk).^54,55 Notably, for the 13 participants scanned on MR2, we also acquired a time-of-flight (TOF) MR angiography sequence using a 3D flow-compensated gradient echo technique with multiple slabs (echo time = 3.42 ms, repetition time = 19 ms, flip angle = 18°, slice thickness = 1.3 mm, spacing between slices = 1.3 mm, base resolution = 256 × 256, phase resolution = 0.8, partial Fourier factor = 0.75).

Participant exclusion and missing data

During psilocybin intervention days, four participants did not complete all planned pcASL scans. On MR1, one participant missed the post-drug scan at +300-min due to technical issues, and another did not complete it due to nausea. On MR2, one participant did not complete any post-drug scans due to discomfort, while another missed the +130-min and +300-min scans due to nausea. For the ketanserin session, four participants were excluded: one due to the Covid-19 lockdown, and three due to drug availability issues. We analysed only the pre-drug and +80-min scans, which corresponds to periods of comparable 5-HT2AR occupancy between psilocybin and ketanserin (see Supplementary Figure S1).^46,50

CBF quantification

MR1 preprocessing

The FSL command fsl_motion_outliers was applied to label and control images from MR1 data; flagged label/control pairs were excluded from the scan-specific quantification of CBF (median [range] of flagged volumes: 4 [0–9] out of 40 pairs per scan). The pcASL data underwent intramodal registration dividing label and control images (40 pairs) before realigning them. The pcASL data were then unwarped to correct for B0 field distortions. Realignment and distortion correction was then performed using SPM12 (https://www.fil.ion.ucl.ac.uk/spm/software/spm12/). The high-resolution T1-weighted structural image was co-registered with the realigned and motion corrected pcASL data. We corrected for the interleaved slice acquisitions in the 2D EPI protocol using a delay in slice time acquisition of 35 ms.

MR2 preprocessing

For each of the PLDs, a label/control difference image was realigned and the high-resolution T1-weighted structural image was co-registered with the M0 image. Co-registration with the pcASL images like we did for the MR1 data was not possible without introducing transformation artefacts.

Two post-psilocybin scans from one MR1 participant and five scans from two MR2 participants—four scans from one and one scan from the other—were excluded due to poor image quality. On ketanserin days, four scans across two MR2 participants were excluded due to image artifacts, and three additional scans were excluded for falling outside the peak occupancy time window.

All pcASL images from both scanners were quantified to absolute CBF values using BASIL in FSL (FMRIB software library, www.fmrib.ox.ac.uk), ⁵⁶ which performs subtraction of label/control images and calibrates the signal to absolute CBF (ml/100g/min) incorporating the M0 scan and modelling the tracer kinetics using a single, well-mixed tissue compartment model. ⁵⁷ We used standard settings for the quantitative modelling described previously. ⁵⁷ The structural image was segmented in SPM12 into grey matter, white matter, and cerebrospinal fluid probability maps. The calibrated pcASL image was normalised into (MNI) standard space based on estimated warps from the segmentation step. Normalised pcASL images were spatially smoothed with a 3D Gaussian kernel of 8 mm Full-Width at Half-Maximum (FWHM).

CBF terminology

We report all CBF values in absolute physiological units (ml/100g/min). “Global CBF” refers to whole-brain CBF, whereas “regional CBF” describes absolute CBF values within specific regions of interest. We did not apply global signal regression when estimating drug effects because we view our repeated-measures design as enabling the assessment of changes in both regional and whole-brain CBF relative to baseline without additional normalization or regression, ensuring physiologically interpretable results.

Ica diameter quantification

We developed and applied an in-house, semi-automatic tool (GitHub link: https://github.com/kristian1801/ica-segmentation-tool) for ICA segmentation and diameter quantification. TOF images underwent preprocessing to enhance vascular structures, including background subtraction, anisotropic diffusion filtering, and top-hat filtering. User-guided segmentation was performed via a graphical interface, where the user manually selected the ICA. A 2D region-growing algorithm with dynamic intensity thresholding was then applied to segment the chosen vessel across all axial slices. The software reconstructed a 3D model of the segmented ICA and calculated its diameter by taking the median of measurements from each image slice.

Data analysis

Regions of interest (ROIs)

ROIs for CBF-associated analyses were defined using the Anatomical Automatic Labelling (AAL3) atlas. ⁵⁸ The AAL3 delineates 170 ROIs in MNI space but here we grouped adjacent ROIs to form 13 larger, bilateral ROIs for the analysis, based on previous studies of acute psilocybin effects on CBF^41,42: prefrontal cortex, temporal cortex, parietal cortex, occipital cortex, anterior cingulate cortex (ACC), posterior cingulate cortex (PCC), thalamus, amygdala, putamen, caudate nucleus, hippocampus, insula, orbitofrontal cortex (OFC). See Supplemental Material for a description of AAL3 regions combined to produce each ROI. We computed and evaluated grey-matter weighted regional CBF as defined by subject-specific grey-matter segmentation maps. We computed global grey-matter weighted CBF across all voxels within cortical and subcortical regions in the AAL3 atlas. The cerebellar ROI was excluded from analyses due to inconsistent field of view coverage across participants.

Associations of CBF and ICA diameter with PPL and SDI

A total of 124 pcASL scans were included from psilocybin intervention days (70 from MR1, 54 from MR2), and 49 pcASL scans were included from ketanserin days (30 from MR1, 19 from MR2). We used linear mixed-effects models (LME) to estimate associations between exposure variables (PPL or SDI) and brain measurements (global or regional CBF, or ICA diameter). We included participant ID as a random effect to account for correlations between repeated observations from the same participant; age, sex, and scanner were included as fixed effects. We fitted separate LMEs for each combination of outcome (global CBF, 13 regional CBF measurements, and ICA diameter) and exposure (PPL or SDI). The regression coefficient β quantifies the association, representing the average change in the brain measurement (Y) for a one-unit increase in the exposure variable (X), while keeping other characteristics constant. We assessed statistical significance using Wald tests. For the 13 regional CBF tests, we adjusted p-values using Dunnett’s method to control the family-wise error rate (p_FWER) ⁵⁹ as these examined spatially related measurements across brain regions. Global CBF and ICA diameter were treated as independent tests because they reflect a composite measure across all regions (global CBF) and a distinct metric from a separate sequence (ICA). We used a statistical significance threshold of p (or p_FWER) < 0.05. We also report the estimated maximum percent change $\Delta \text{\%}=100\cdot \frac{E\left[Y\text{|}{X}_{\mathit{max}},Z\right]-E\left[Y\text{|}{X}_{\mathit{min}},Z\right]}{E\left[Y\text{|}{X}_{\mathit{min}},Z\right]}$ , where E[Y|X, Z] denotes the expected value of the brain measurement Y, given the exposure X and other covariates Z, X_max is the maximum exposure value (19.3 µg/L for PPL, 10 for SDI) and X_min is the minimum exposure value (0 for both PPL and SDI). As an exploratory analysis, we estimated receptor occupancy using PPL values and the Emax model from our previous work ⁴⁶ (Supplementary Figure S2): Occupancy = (Occ_max* PPL)/(EC₅₀ + PPL), where EC₅₀ is the concentration at half-maximal effect. Occ_max was set as 76.6% and EC₅₀ as 1.96 µg/L.

Ketanserin effects and drug-x-time effects on CBF and ICA diameter

For ketanserin effects, each model included time (pre-drug vs. +80-min) as the independent variable, with sex, age, and scanner as covariates. Percent changes in CBF and ICA after ketanserin were calculated using: $\Delta \text{\%}=100\cdot \frac{E\left[Y\text{|}{X}_{\mathit{max}},Z\right]-E\left[Y\text{|}{X}_{\mathit{min}},Z\right]}{E\left[Y\text{|}{X}_{\mathit{min}},Z\right]}$ , where E[Y|X, Z] denotes the expected value of the brain measurement Y, given the exposure X and other covariates Z. X_max is the +80 min timepoint; X_min is pre-drug. Drug-x-time models included a drug-x-time interaction term to compare psilocybin and ketanserin effects over time. p-values were adjusted using Dunnett’s method (p_FWER < 0.05), except for global CBF and ICA diameter. See supplementary Figure S1 for occupancy estimates.

Visualisations, code and data availability

The MATLAB and R code used to generate the results presented in this study can be made available upon request. We used BrainNet Viewer ⁶⁰ (https://www.nitrc.org/projects/bnv/) to generate the brain colour maps presented in Figure 2(a). The datasets generated and/or analysed during the current study can be made available upon completion of a formal data-sharing agreement.

Figure 2.

Time course of cerebral blood flow (CBF), plasma psilocin level (PPL) and subjective drug intensity (SDI) before and after psilocybin intake. (a) Brain representations are colour maps created from the absolute voxel-wise CBF values and (b) A boxplot of global, grey-matter weighted CBF at pre-drug, +40-, +80-, +130- and +300-min following psilocybin administration. Asterisks mark the mean value. (c) and (d) temporal trajectories of PPL and SDI (group mean ± 1 SD) associated with imaging sessions.

Results

On psilocybin days, we observed ranges for PPL [0–27.6 µg/L], SDI [0–10], and CBF [22.9–77.0 ml/100g/min] throughout the time course of the intervention. On ketanserin days we observed ranges for CBF [34.5–84.2 ml/100g/min]. The time course of PPL and SDI over the experiment was similar, confirming both drug uptake and associated subjective effects (Figure 2(b) and (c)). All participants reported psychoactive effects following psilocybin administration, as evidenced by high ratings on the SDI scale. In contrast, ketanserin produced negligible subjective effects, and most participants rated their SDI at or near zero at the +80-min post-drug scan session (SDI score: mean = 0.23, median = 0, range = 0–3).

Global CBF decreases following psilocybin intake

We observed decreased global CBF following psilocybin administration (Figure 2(a)). There was a significant negative association between global CBF and both PPL (ß = −0.38, 95% CI = [−0.56, −0.20] ml/100g/min per µg/L of psilocin, p < 0.0001) and SDI (ß = −0.75, 95% CI = [−0.1, −0.41] ml/100g/min per unit SDI, p < 0.0001), see also Table 1 and Figure 3. This corresponds to a ∼11.6% decrease in global CBF at peak PPL or SDI.

Table 1.

Associations with plasma psilocin level and subjective drug intensity.

	Plasma psilocin level				Subjective drug intensity
MEASURE	ß	SE	PFWER	Δ%	ß	SE	PFWER	Δ%
Global CBF	−0.38	0.09	<0.0001***	−11.62	−0.75	0.13	<0.0001***	−11.56
Parietal	−0.53	0.10	<0.0001***	−15.47	−0.97	0.17	<0.0001***	−14.51
Occipital	−0.38	0.10	0.0007***	−13.40	−0.78	0.15	<0.0001***	−13.77
Prefrontal	−0.30	0.10	0.0167*	−8.22	−0.59	0.15	0.0008***	−8.25
Temporal	−0.36	0.09	0.0006***	−11.00	−0.77	0.16	<0.0001***	−11.86
OFC	−0.04	0.13	1.00	−1.20	−0.23	0.24	0.88	−3.67
ACC	−0.33	0.11	0.017*	−8.60	−0.70	0.19	0.0014**	−9.18
PCC	−0.60	0.16	0.003**	−13.90	−1.22	0.27	<0.0001***	−14.52
Insula	−0.22	0.13	0.43	−5.96	−0.57	0.24	0.10	−7.95
Putamen	−0.15	0.09	0.43	−5.60	−0.54	0.16	0.0077**	−9.83
Caudate	−0.11	0.10	0.79	−4.79	−0.35	0.16	0.14	−7.45
Thalamus	−0.03	0.14	1.00	−1.13	−0.40	0.25	0.44	−7.84
Hippocampus	0.03	0.11	1.00	1.03	−0.30	0.21	0.55	−5.17
Amygdala	−0.23	0.12	0.29	−9.29	−0.51	0.23	0.15	−10.57
ICA diameter	−0.02	0.003	<0.0001***	−10.51	−0.03	0.01	<0.0001***	−6.61

Expected change in outcome measure (cerebral blood flow or internal carotid daimeter) for a unit change in plasma psilocin level (PPL) or subjective drug intensity (SDI) as estimated by the linear mixed model. ß represents the estimated change. Δ% denotes the estimated percent change for PPL = 19.3 µg/L (max value based on previous study [49]) versus PPL = 0, and SDI = 10 (max value) versus SDI = 0. P_FWER indicates family-wise error rate corrected p-values, except for Global CBF and ICA diameter, which are uncorrected. CBF measures are reported for global and regional areas. Brain regions include OFC (Orbitofrontal cortex), ACC (Anterior cingulate cortex), and PCC (Posterior cingulate cortex). SE: standard error. Significance levels: *p < 0.05, **p < 0.01, ***p < 0.001.

Figure 3.

Associations between regional cerebral blood flow (CBF) and plasma psilocin level (PPL) and subjective drug intensity (SDI). (a), (b), (c) and (d) display the global, parietal, occipital and temporal region associations with PPL, respectively. Panels (e), (f), (g) and (h) display associations with SDI. Associations with CBF are plotted adjusting for age, sex and scanner. Each color represents a unique participant, with multiple dots of the same color indicating repeated measurements from the same individual across different timepoints. When multiple dots overlap at the same SDI rating, this indicates that a participant reported identical subjective intensity scores at different timepoints.

Associations between CBF and PPL or SDI

All 13 regions, except the orbitofrontal cortex (OFC) and hippocampus, displayed a numerically negative association between PPL and CBF (Table 1). The effects were statistically significant in the parietal, occipital, temporal and prefrontal cortex (p_FWER < 0.02, Figure 3(b) to (d)) and for the ACC and PCC (p_FWER < 0.02).

All 13 regions displayed a numerically negative association between SDI and CBF (Table 1). Effects were statistically significant for the parietal, occipital, prefrontal and temporal cortex (p_FWER < 0.0009, Figure 3(f) to (h)) and the ACC, PCC and putamen (p_FWER < 0.008). Our exploratory analysis showed similar associations between estimated receptor occupancy and CBF (Supplementary Figure S2 and Table S1).

Intake of ketanserin was not associated with a statistically significant change in global CBF (p = 0.35, Table 2), nor changes in any regional CBF (all p_FWER>0.19, Table 2). Intake of ketanserin was not associated with a statistically significant change in ICA diameter (p = 0.99), see Table 2.

Table 2.

Ketanserin and drug-x-time effects.

	Ketanserin effect				Drug-x-time
MEASURE	ß	SE	PFWER	Δ%	ß	SE	PFWER	Δ%
Global CBF	−2.05	1.05	0.35	2.30	−3.54	2.38	0.14	−5.09
Parietal	−0.84	0.92	0.95	1.32	−7.86	1.91	0.0005***	−11.72
Occipital	−2.09	1.23	0.50	3.66	−4.00	1.92	0.20	−7.03
Prefrontal	−2.30	1.02	0.19	1.31	−3.37	2.11	0.47	−4.71
Temporal	−0.88	1.10	0.98	3.79	−2.87	2.12	0.66	−4.40
OFC	−2.15	1.56	0.66	3.60	1.77	2.60	0.98	2.85
ACC	−0.56	1.48	0.99	0.77	−3.47	2.54	0.63	−4.55
PCC	−2.89	2.90	0.92	3.60	−6.04	4.36	0.62	−7.40
Insula	−1.18	1.39	0.96	1.76	−2.62	2.38	0.82	−3.66
Putamen	−1.54	1.31	0.83	2.86	−1.30	2.22	0.99	−2.38
Caudate	−1.92	1.31	0.62	4.19	0.65	2.11	1.00	1.45
Thalamus	−1.10	2.13	0.99	2.11	1.21	3.63	1.00	2.47
Hippocampus	−0.93	1.25	0.98	1.79	4.96	2.68	0.31	9.36
Amygdala	−0.56	1.11	0.99	−1.12	−2.46	3.06	0.96	−5.07
ICA diameter	−0.001	0.05	0.99	−0.02	−0.39	0.12	0.007**	−10.05

Expected change in outcome measures (cerebral blood flow and internal carotid artery diameter) as estimated by the Linear Mixed Effects model, Ketanserin effect is presented on the left side of the table, while the drug-x-time interaction effect for psilocybin versus ketanserin is shown on the right side, ß represents the estimated change between pre-drug and +80 minutes post-drug administration, Δ% denotes the estimated percent change for ketanserin (left) and the estimated relative percent change of psilocybin versus ketanserin (right), PFWER indicates family-wise error rate corrected p-values, except for Global CBF and ICA diameter, which are uncorrected, CBF measures are reported for global and regional areas, Brain regions include OFC (Orbitofrontal cortex), ACC (Anterior cingulate cortex), and PCC (Posterior cingulate cortex), SE: standard error, Significance levels: *p < 0,05, **p < 0,01, ***p < 0,001.

Contrasting ketanserin and psilocybin effects on CBF and ICA diameter

Contrasting ketanserin and psilocybin effects on global CBF at 80-minutes post-drug administration relative to baseline, we did not see a statistically significant drug-x-time interaction effect (p = 0.14; Table 2 and Figure 4).

Figure 4.

Effects of ketanserin and psilocybin on cerebral blood flow and ICA diameter. Global CBF, Parietal CBF, and ICA diameter shown before and +80-min after administration of psilocybin and ketanserin. Red and green circles and lines represent the ketanserin and psilocybin treatments, respectively. Smaller points denote individual observations, population mean values indicated by larger circles and error bars representing one standard error on the mean.

However, we did find a statistically significant drug-x-time interaction effect on parietal cortex CBF such that the psilocybin-induced decrease in CBF was significantly larger than the effect of ketanserin (ß = −7.86, 95% CI = [−12.99, −2.74] ml/100g/min, p_FWER = 0.0005; Table 2 and Figure 4). No other regions showed statistically significant drug-x-time interaction effects on CBF (Table 2). We also observed a significant drug-x-time interaction effect on the ICA diameter (ß = −0.39, 95% CI = [−0.649, −0.133] mm, p_FWER = 0.007; Table 2 and Figure 4).

Associations between ICA diameter and PPL or SDI

We found that psilocybin intake was associated with a decrease in ICA diameter (Figure 5), with a significant negative association between ICA diameter and both PPL (ß = −0.02, 95% CI = [−0.03, −0.015] mm per µg/L of psilocin, p < 0.0001) and SDI (ß = −0.03, 95% CI = [−0.038, −0.014] mm per unit SDI, p < 0.0001). This corresponds to a 10.5% decrease in ICA diameter at peak PPL and a 6.6% decrease at peak SDI.

Figure 5.

Associations between internal carotid artery (ICA) diameter and plasma psilocin level (PPL) and subjective drug intensity (SDI). (a) ICA diameter changes over time following psilocybin administration. Box plots show distribution at each time point, with individual measurements as dots. Asterisks indicate means. (b) Association between ICA diameter and PPL. Each colour represents a unique individual and (c) Association between ICA diameter and SDI. Associations with ICA diameter are adjusted for age, sex.

Discussion

Here, we evaluated and contrasted the acute effects of psilocybin and ketanserin, drugs with agonist and antagonist effects at the 5-HT2AR, respectively, on global and regional CBF in healthy individuals. Psilocybin intake was associated with up to an 11.6% reduction in global CBF; we show here for the first time that this change is significantly associated with both PPL and SDI, which we have previously shown to be tightly related to brain 5-HT2AR occupancy, ⁴⁶ associated with changes in BOLD functional network connectivity ⁴⁷ , occurrence of dynamic brain states ⁴⁹ , and certain forms of brain-entropy. ⁴⁹ Additionally, we observed a significant decrease in ICA diameter following psilocybin administration, with a 10.5% reduction at peak PPL, highlighting the physiological effect of psilocybin on cerebral vasculature. Neither CBF nor ICA diameter were significantly changed following ketanserin, indicating that 5-HT2AR antagonism does not affect CBF. ⁶¹ In the parietal cortex, psilocybin intake was associated with a significantly larger decrease in CBF compared to that of ketanserin. These findings reinforce an acute modulation of CBF by psilocybin across many brain regions and directly map these changes onto acute pharmacological and subjective effects.

We provide the first evidence that both PPL and SDI, acquired at the time of each individual scan session, are significantly negatively correlated with global and regional CBF, reinforcing a close link between the acute effects of psilocybin, associated subjective effects, and CBF. Our observation that CBF is reduced following psilocybin administration is somewhat consistent with previous studies, but certain discrepancies are noteworthy.^41–43 In a smaller sample of 15 healthy participants, Carhart-Harris and colleagues reported negative associations between perceptual effects of psilocybin and CBF in the ACC, PCC and thalamus. ⁴¹ We observed similar negative associations between PPL and SDI and CBF within the ACC and PCC, regions often the focus of psychedelic resting-state fMRI literature stemming from being elements of the default-mode network and regions with high 5-HT2AR density.^39,62–64 However, we did not observe significant associations between PPL nor SDI and thalamus CBF (see Table 1), the eponymous brain region for the thalamic gating model of psychedelic action.^65,66 Our observation of a limited effect in the thalamus is seemingly consistent with a study by Lewis and colleagues, who report decreased CBF in 58 individuals, i.e., that which they report as “global CBF” or “gCBF”, in large clusters distributed across many brain regions including seemingly only a small subsection of the left thalamus. ⁴² Lewis and colleagues reported lower CBF following psilocybin administration within many regions, including the frontal, parietal, temporal, and occipital cortices, as well as the ACC. Although these observations are consistent with ours, they also reported significant reductions in CBF in the insula, caudate, putamen and amygdala.

We observed a statistically significant negative association between CBF and SDI (but not PPL) in the putamen, with numerically similar but non-significant associations in other subcortical regions. Although Lewis and colleagues reported lower CBF in the hippocampus and thalamus, these effects appear to be confined to relatively smaller subsections, which somewhat aligns with our findings of non-significant associations in these areas. There are notable sources of heterogeneity between the studies, which may explain the discrepant regional results: 1) Carhart-Harris and colleagues administered psilocybin intravenously, whereas Lewis and colleagues and we administered psilocybin perorally, i.e., as it is currently administered in clinical trials, ³ 2) Carhart-Harris and colleagues estimated pulsed ASL, whereas Lewis and colleagues performed pcASL, as did we. Approximately half of our data were acquired with a 2D single-PLD sequence, similar to the sequence used by Lewis and colleagues, whereas the other half of our data were acquired with a 3D multi-PLD sequence. Nevertheless, despite regional and magnitude differences, our observation that global and regional CBF is reduced following psilocybin is generally consistent with the previous literature, which we show is significantly correlated with PPL and SDI.

Interestingly, we observed more brain regions showing a significant association between CBF and SDI than PPL and with higher statistical significance (see Table 1). This is qualitatively similar to what we observed when evaluating associations with resting-state functional connectivity in a subset of the current included participants (see Figure 3 from cit. ⁴⁷ ). Our estimated maximal change in CBF, i.e., model estimated decrease at peak observed PPL and SDI, was similar in most neocortical regions i.e. temporal, parietal and occipital, in the range: [−11% to −15.5%], but diverged in deeper brain structures, where the effect of PPL was in the range: [+1% to −13.9%] and SDI in the range [−5.2% to −14.5%]. SDI scores are more bimodally distributed than PPL, which may affect related statistical associations. Notably, 5-HT2AR occupancy by psilocin is related non-linearly to plasma psilocin levels. ⁴⁶ If we assume that most of the acute neural effects of psilocybin are due to 5HT2AR agonism, then mapping linear relations between plasma drug levels and neural changes, e.g., CBF, may inadequately capture the true relation between these variables. This may contribute to our observed differences in estimated associations with PPL and SDI; how best to model these associations should be evaluated further in future studies.

Although CBF was numerically decreased across brain regions following ketanserin administration, this effect was small, and not statistically significant (Table 2). Similarly, ketanserin produced negligible perceptual effects. Our lab previously showed with [11C]Cimbi-36 PET that 20 mg ketanserin produces ∼65–75% 5-HT2AR occupancy 1–3 hours after oral administration, near to its estimated maximal occupancy (i.e, 77%) ⁵¹ and the maximal occupancy reported for other 5-HT2AR antagonists/inverse agonists.^50,67,68 The observation that 5-HT2AR blockade with ketanserin at this occupancy level does not substantively affect CBF suggests that endogenous 5-HT2AR signalling at rest plays a limited role in regulating CBF. Directly comparing effects of psilocybin and ketanserin on CBF, we observed a statistically significant drug-x-time interaction effect within the parietal cortex such that the psilocybin-related decrease was greater (i.e., more negative) than the ketanserin-related change. The parietal cortex is the region wherein we observed the most pronounced negative association between CBF and both PPL and SDI, i.e., a decrease in CBF of ∼15% at peak drug and experience level. The parietal cortex contains regions involved in higher-order association networks including the default mode and frontoparietal networks, alterations in the functional connectivity of which have been linked to acute psychedelic effects. ⁴¹ Although other regions in those networks showed directionally similar drug-x-time interaction effects, they were not statistically significant, emphasizing that larger samples are needed to better resolve head-to-head drug effects. We interpret the associations with PPL and SDI as supportive of a substantial psilocybin effect on CBF.

It is not immediately straight-forward how to reconcile the generally observed decrease in CBF with previous psilocybin studies evaluating effects on cerebral metabolic rate of glucose (CMRglc) with [18F]-FDG PET. One previous study reported numerically decreased global but both increased (e.g., anterior cingulate) and decreased (e.g., thalamus) regional CMRglc following psilocybin administration in the context of a word repetition/association task. ⁶⁹ However, a second study that did not include any such task reported a global increase in CMRglc of ∼10–25%, most markedly in the frontal, anterior cingulate gyrus, and temporomedial cortex. ⁴⁵ This discrepancy is somewhat unexpected considering that CMRglc and CBF are normally tightly coupled.^70,71 However, decoupling between CBF and CMRglc has been reported previously, e.g., caffeine has been shown to decrease CBF^72–75 and increase energy metabolism,^76,77 and previous studies have speculated that psilocybin-induced reductions in CBF are not due to effects of stimulation of the neuronal 5-HT2ARs, but other receptor targets and direct effects on vascular tone.^41,42 One possible explanation for the discrepancy between the current results and the previous reports of increased CMRglc following psilocybin ingestion could be the agonist activity of psilocybin at the inhibitory serotonin 1A receptor (5-HT1AR). The 5-HT1AR is expressed in various brain regions and acts both as an auto- and hetero-receptor and has, upon stimulation, been shown to lead to reduced cerebral blood volume in animal models.^19,78,79

We also observed that ICA diameter was significantly negatively associated with both PPL and SDI, the first human in vivo evidence that psilocybin induces cerebral vasoconstriction. Our psilocybin observations are consistent with previous studies showing 5-HT2AR is expressed in the peripheral vasculature, ³⁰ where it induces vasoconstriction,^31,33,80,81 and preclinical research shows 5-HT-induced vasoconstriction in the carotid arteries of rats.^82,83 This highlights an important consideration for studies estimating CBF effects of 5-HT2AR agonist drugs: pcASL is inversely related to carotid blood flow velocity in a non-linear fashion, with the effects being most pronounced for increases in flow velocity.^84,85 A psilocin-induced vasoconstriction could reduce the amount of arterial blood being labelled, thereby underestimating CBF. Additionally, both preclinical and clinical studies have shown that psilocybin can induce mild hyperthermia,^86,87 which in animals induces carotid artery constriction. ⁸⁸ Based on our current observation and previously suggested best practice, ⁸⁹ future studies of psychedelic effects on CBF should measure vasoactive effects with, e.g., TOF angiography and phase-contrast mapping to directly measure ICA diameter and blood flow velocity, respectively. This will assist the interpretation of CBF effects and provide critical insights into the hemodynamic changes induced by these drugs.

Multimodal imaging techniques, such as simultaneous [18F]-FDG PET and pcASL, may also help resolve whether the apparent discrepancy between CMRglc and CBF reflects a decoupling of CBF from energy metabolism following psilocybin administration. Future studies using more selective 5-HT2AR agonists (such as the phenethylamine psychedelics ⁹⁰ ) and 5-HT1AR antagonists could help disentangle various sources of CBF alterations in CBF studies using polypharmacological drugs like psilocybin and ketanserin. These approaches would provide more detailed information about the relation between psilocybin, 5-HT2AR, and hemodynamic responses, clarifying the contributions of neuronal and non-neuronal factors to associated signals that are a common source of interest. ³⁸

One of the previous studies that evaluated acute psilocybin effects on CBF considered two analysis strategies, one of which involved adjusting for global CBF. ⁴² This was motivated in part by a previous study advocating for this strategy. ⁹¹ We did not apply this adjustment because our repeated measures model relates within individual changes in CBF to PPL or SDI, which we view as particularly desirable for assessing associated drug effects.

Our study is not without its limitations. Employing a double-blind design could have limited potential biases compared to our single-blind design. Notwithstanding, maintaining participant blinding in psychedelic studies is very difficult due to the intense subjective effects.^{4,11,17,92–94} Brain imaging data were collected across two scanners with distinct pcASL sequences (2D vs. 3D acquisition protocol) because MR1 became unavailable during the study and the MR1 sequence was not available on the MR2 scanner. These represent hardware and software sources of variation, which we adjust for by including scanner as a variable in our statistical models. This may potentially limit our ability to resolve drug-related effects. As we have discussed previously, ¹³ hysteresis effects, i.e., different associations during the ascent and descent phases, cannot be ruled out; we do not have sufficient data to directly estimate hysteresis effects. We corrected for motion and excluded participants with excessive motion, yet it may remain a confound as head motion is increased while on psychedelics.^41,93 Psilocybin has been found to increase body temperature, blood pressure, and heart rate with reported increases in systolic and diastolic blood pressure of 15–20 mmHg and 7–15 mmHg respectively at comparable doses. ⁸⁷ These changes fall within normal cerebral autoregulation range (60–150 mmHg MAP), ⁹⁴ and observed CBF decreases rather than increases suggest minimal blood pressure influence suggesting it does not underlie our observed CBF effect. We did not measure end-tidal CO2, which could potentially influence CBF measurements; however, regular communication with participants during scanning did not suggest that participants exhibited elevated breathing or hyper-ventilation. A previous study used a nasal cannula attached to a capnograph to collect data but did not apply these measurements to correct CBF data. ⁴¹ Incorporating this measure into future studies could improve the estimation of drug effects on CBF. It is also important to note that although we observed some grey-matter weighted CBF estimates outside the typically expected range of 40–80 ml/100g/min, ⁸⁹ most scans were within this range, and such values are not uncommon in pcASL imaging. Although alternative modeling approaches could be considered (e.g., winsorizing values), we decided to retain these scans based on data quality and consistency within participants.

Overall, we find that CBF decreases acutely following psilocybin administration and it is negatively associated with PPL and SDI. By contrast, ketanserin did not significantly affect CBF. ICA diameter significantly decreased following psilocybin, highlighting the need to resolve physiological mechanisms important for interpreting psychedelic effects on hemodynamic related signals. These findings add to our understanding of the neurobiological effects of psychedelic drugs in healthy participants. Although our results cannot be directly extrapolated to clinical populations, these insights into how psilocybin modulates brain physiology may inform future research exploring therapeutic mechanisms in disorders such as depression.

Supplementary material

Supplemental material for this article is available online.