PsilocybinPlacebo

Psilocybin-induced changes in brain network integrity and segregation correlate with plasma psilocin level and psychedelic experience

This fMRI study (n=15) investigated the effects of psilocybin (14-21mg/70kg) on the brain and found that the higher the psilocin (active metabolite of psilocybin) and subjective drug experience (SDI) correlated with lower network integrity and segregation (less top-down, more bottom-up).

Authors

  • Gitte Knudsen
  • Patrick Fisher

Published

European Neuropsychopharmacology
individual Study

Abstract

The emerging novel therapeutic psilocybin produces psychedelic effects via engagement of cerebral serotonergic targets by psilocin (active metabolite). The serotonin 2A receptor critically mediates these effects by altering distributed neural processes that manifest as increased entropy, reduced functional connectivity (FC) within discrete brain networks (i.e., reduced integrity) and increased FC between networks (i.e., reduced segregation). Reduced integrity of the default mode network (DMN) is proposed to play a particularly prominent role in psychedelic phenomenology, including perceived ego-dissolution. Here, we investigate the effects of a psychoactive oral dose of psilocybin (0.2-0.3 mg/kg) on plasma psilocin level (PPL), subjective drug intensity (SDI) and their association in fifteen healthy individuals. We further evaluate associations between these measures and resting-state FC, measured with functional magnetic resonance imaging, acquired over the course of five hours after psilocybin administration. We show that PPL and SDI correlate negatively with measures of network integrity (including DMN) and segregation, both spatially constrained and unconstrained. We also find that the executive control network and dorsal attention network desegregate, increasing connectivity with other networks and throughout the brain as a function of PPL and SDI. These findings provide direct evidence that psilocin critically shapes the time course and magnitude of changes in the cerebral functional architecture and subjective experience following psilocybin administration. Our findings provide novel insight into the neurobiological mechanisms underlying profound perceptual experiences evoked by this emerging transnosological therapeutic and implicate the expression of network integrity and segregation in the psychedelic experience and consciousness.

Unlocked with Blossom Pro

Research Summary of 'Psilocybin-induced changes in brain network integrity and segregation correlate with plasma psilocin level and psychedelic experience'

Introduction

Contemporary neuroimaging frameworks propose that behaviour and conscious experience emerge from coordinated activity among large-scale functional brain networks. These networks are identified with resting-state BOLD fMRI and are characterised by strong functional connectivity (FC) within a network (network integrity) and weak FC between networks (network segregation). Earlier work has suggested that psychedelics, including psilocybin, reduce within-network FC and increase between-network FC, producing a more desegregated and entropic functional architecture; however, neuroimaging studies to date have not directly related these network changes to plasma levels of psilocin (the active metabolite) or tracked them across the full acute oral psilocybin time course (ascent, peak, descent). Madsen and colleagues set out to map how plasma psilocin level (PPL) and moment-to-moment subjective drug intensity (SDI) relate to changes in network integrity and segregation after a clinically relevant oral psilocybin dose. The study tests the hypotheses that PPL and SDI would correlate negatively with default mode network (DMN) and overall within-network RSFC (reflecting reduced integrity) and positively with between-network RSFC (reflecting reduced segregation). To probe spatial specificity, the investigators also examined RSFC within and between individual networks and performed voxel-level, spatially unconstrained analyses using local correlation (LCOR) and global correlation (GCOR), measures of local integrity and whole-brain connectivity respectively.

Methods

Fifteen healthy adults (mean age 34.3 ± 9.8 years, six females) with no or limited prior psychedelic experience participated. Participants received an oral psilocybin dose of approximately 0.2 mg/kg (n=4) or 0.3 mg/kg (n=11), yielding a relative dose mean ± SD of 0.26 ± 0.04 mg/kg and an absolute dose mean ± SD of 19 ± 3.5 mg. Participants were blinded as to whether they would receive psilocybin or a non-psychedelic drug. Two psychologists provided preparation, support and integration. Structural T1-weighted and BOLD rs-fMRI data were acquired on a Siemens Prisma 3T scanner using a 64-channel head coil. Post-drug resting scans were collected at approximately 40, 80, 130 and 300 minutes after administration (each rs-fMRI acquisition lasted 10 minutes). Immediately after each scan, participants provided an SDI rating on a 0–10 Likert scale, then a blood sample was drawn from an intravenous catheter to measure free (unconjugated) psilocin using ultra performance liquid chromatography with tandem mass spectrometry. End-of-day questionnaires assessed the overall psychedelic experience (11D-ASC, MEQ30, EDI). Image preprocessing used SPM12, with denoising and FC estimation in CONN. Analyses modelled individual and group time courses of PPL and SDI, including an Emax model for the SDI–PPL relation. Within- and between-network FC were computed as averages of Fisher-transformed r-to-z values across pairs of regions for seven predefined resting-state networks (36 ROIs). Linear mixed-effects regression, accounting for repeated measures per subject, tested associations between FC and PPL or SDI; results are reported as correlation coefficients (R) and unstandardised regression coefficients (β) with 95% confidence intervals. Family-wise error rate (FWER) was controlled at 5% with adjusted p-values (pFWER < 0.05 considered significant). Voxel-level LCOR and GCOR maps entered separate mixed models with PPL or SDI; clusterwise inference used an estimated voxel-level threshold p < 0.001 and a cluster extent threshold of 560 voxels (AFNI tools). Finally, a predicted response profile for a 25 mg clinical psilocybin dose was estimated by spline-fitting the 0.3 mg/kg psilocin time course and scaling to an approximate Cmax of 18.7 µg/L, then projecting corresponding SDI and RSFC values using the study models.

Results

Psychedelic phenomenology and pharmacokinetics: End-of-day questionnaires confirmed that the administered doses produced a profoundly altered state of consciousness on 11D-ASC, MEQ30 and EDI. PPL and SDI exhibited similar temporal trajectories and correlated positively across the acute session; one participant showed a delayed but rapid increase in SDI consistent with slower pharmacokinetics. Considerable interindividual variability was observed in Cmax and Tmax. Network integrity and segregation (network-level analyses): Across seven networks, within-network integrity measures averaged over networks showed a negative association with PPL and SDI, but the overall within-network effect did not reach statistical significance after multiple-comparison correction (pFWER = 0.067). Examining individual networks, PPL correlated negatively with DMN integrity; a similar negative association was observed for the salience network (SAN). Between-network analyses demonstrated statistically significant reductions in segregation as a function of PPL for DMN with SAN and DMN with the executive control network (ECN, corresponding to the frontoparietal control network). Effects on integrity and segregation were more pronounced for higher-order networks (DMN, SAN, ECN, dorsal attention network (DAN)) than for lower-level unimodal networks such as the sensorimotor (SMN) and auditory (AN) networks. Voxel-level local and global correlation: LCOR analyses showed negative correlations between PPL and local connectivity in regions central to the DMN and ECN and in visual cortex—areas with high 5-HT2AR expression. GCOR analyses indicated positive associations between PPL and whole-brain connectivity in regions of the DAN and ECN (bilateral superior frontal gyri and bilateral intraparietal sulci). SDI produced similar GCOR associations and additionally implicated a cluster in the dorsomedial cingulate gyrus, supplementary motor area and thalamus. Notably, thalamic GCOR correlated positively with SDI. Secondary observations and quality control: SDI correlated with retrospective questionnaire measures (global ASC, EDI and MEQ30), supporting its use as a real-time index of psychedelic phenomenology. Head motion correlated positively with PPL but was generally of small magnitude; quality-control FC diagnostics suggested successful noise removal in included datasets though some datasets were excluded. The sample size limited power to detect subtler effects.

Discussion

Madsen and colleagues interpret their findings as evidence that psilocin critically shapes both the time course and magnitude of psilocybin-induced changes in cerebral functional architecture and subjective experience. The principal observations are that rising PPL (and concordant increases in SDI) are associated with reduced network integrity—most clearly in the DMN and SAN—and with decreased network segregation manifested as increased between-network synchronisation. Voxel-level results corroborate these effects, showing local integrity reductions in DMN/ECN-related regions and increased whole-brain connectivity in frontoparietal and midline regions of DAN and ECN; thalamic connectivity increased alongside SDI, consistent with impaired thalamic gating. The authors situate these results within prior psychedelic neuroimaging and theoretical work. The pattern of disrupted higher-order network integrity together with desegregation between heteromodal and unimodal systems accords with the proposed REBUS (relaxed beliefs under psychedelics) framework, which links 5-HT2AR agonism to a less constrained, more entropic functional organisation. Comparisons with earlier BOLD studies reveal both concordant findings (increased frontoparietal connectivity after psilocybin) and discrepancies (studies reporting decreased whole-brain connectivity), which the investigators attribute to differences in drug formulation and pharmacokinetics (i.v. versus oral), preprocessing choices such as global-signal regression, head motion and analytic methods. The authors emphasise that the tight correlation between PPL and SDI supports psilocin as an objective determinant of subjective experience and that pharmacokinetic variability across individuals likely contributes to variability in phenomenal effects and possibly clinical response. They recommend that future psilocybin studies measure PPL to account for such variability. Limitations acknowledged include a modest sample size, absence of a dedicated placebo condition (despite participant blinding), potential residual confounding from head motion, and the possibility of hysteresis (different ascent/descent relations) which could not be fully assessed with the available time points. The investigators conclude that their multimodal, time-resolved approach links pharmacology, subjective experience and functional network dynamics across ascent, peak and descent phases, implicating network integrity and segregation as neural correlates of the psychedelic state and consciousness.

View full paper sections

RESULTS

PPL and SDI time course. The time course of PPL and SDI was visualized for each individual and for a group approximation (Figand). The relation between PPL and SDI (Fig) was modelled using an E max -model (constraints: SDI min = 0, SDI max = 10): ‫ܫܦܵ‬ ൌ ௌூ ೌೣ ‫כ‬ ೞ ா ఱబ ା ೞ , where "n" represents the Hill factor. Network analysis. Within-network and between FC was computed as the average of Fishertransformed r-to-z values between regions within a given network or between a network pair for each rsfMRI acquisition. Linear mixed-effects regression analysis, which takes into account repeated observations for each subject, was used to model the association between FC and PPL and SDI, respectively (see SI for further information). Effects are reported in correlation coefficients (R) and unstandardized regression coefficients (β) with associated 95% confidence intervals (CI). Correction for multiple comparisons. Unless otherwise stated, the family-wise error rate (FWER) for all statistical tests was controlled at 5% using adjusted p-values (p FWER ), where p FWER below 0.05 was considered statistically significant (see SI for more details).

CONCLUSION

Here we evaluated how plasma psilocin and the psychedelic experience map onto RSFC measures of functional network integrity and segregation. As hypothesized, PPL correlated negatively with DMN network integrity. A similar negative correlation was observed for the SAN. The association with overall network integrity was similarly negative, albeit not statistically significant (p FWER = 0.067). Similar associations were observed for SDI, implicating functional network integrity in phenomenal experience. PPL and SDI correlated negatively with LCOR in regions central to the DMN and ECN and in visual cortex-all regions with high 5-HT2AR levels. These findings suggest that network integrity is also impaired locally and are consistent with previous studies. Our results provide novel empirical evidence that psilocin shapes the integrity of DMN and SAN and implicate these neuropharmacological dynamics in the transformation of consciousness from the normal to the psychedelic state. Consistent with our hypotheses, overall network segregation decreased as a function of PPL and SDI (Fig). This shows that functional networks increasingly synchronize as psilocin engages cerebral targets, including the 5-HT2AR. This finding of reduced network segregation is consistent with previous observations of a functionally more connected brain in the psychedelic state induced by psilocybinand lysergic acid diethylamide (LSD) (Robin L.. Interestingly, however, effects of PPL and SDI on individual network integrity and segregation were observed only for some networks (Fig). Specifically, effects of PPL and SDI on individual network integrity appeared more pronounced for higher-level networks such as DMN and SAN compared to lower-level unimodal networks such as SMN and AN. Further, desegregating effects between individual networks matched the repeating spatial motif of brain networks previously described in the literature. In this motif, the DMN is situated at one end, juxtaposed to heteromodal networks (ECN, DAN or SAN). These heteromodal networks are, in turn, situated next to unimodal networks such as auditory or visual networks, which reside at the other end. We found that as a function of PPL and SDI, segregation is reduced in a similar pattern for 1) DMN with DAN, ECN and SAN, and 2) DAN and ECN with unimodal sensory networks. Thus, although global network segregation is reduced, this effect appears more strongly driven by desegregation of adjacent functional networks. The GCOR analysis showed increased WBC as a function of PPL and SDI for frontoparietal and midline regions within the DAN and ECN. Thus, the spatially unconstrained GCOR analysis and the spatially constrained networks analyses converge, showing desegregation among DMN, ECN, DAN and sensory networks as PPL rises and psychedelic phenomenology intensifies. The DMN is believed to contribute to selfhood/egoand internal mentation (i.e., cognition dependent on constructed representations with minimal external stimuli); the DAN is important in externally oriented attention, and the ECN and SAN have been implicated as regulators of attention. Interestingly, subjective psychedelic experience includes unitive experiences and the perceived reduction in both sense of self and of the borders between self and the external world. It is possible that psilocin-induced disruption of higher level networks, including the DMN, coupled with a "flowing together" of normally segregated functional streams related to internally directed and externally directed attention contributes to the psychedelic unitive experience, which may be therapeutically important for clinical outcomes of psilocybin therapy. Two other fMRI studies used WBC analyses, which yield measures similar to the GCOR analysis. The first study showed increased WBC for frontoparietal networks, SAN and DMN, but not DAN, after i.v. psilocybin. The second study reported only widespread decreased WBC when global-signal regression (GSR) was not used but increased WBC in sensory regions and reduced in associative and subcortical regions when GSR was used. Although Tagliazucchi and colleagues in the former study did not identify DAN as exhibiting increased WBC after psilocybin, our findings agree that other frontoparietal regions increase WBC after psilocybin. Findings from the latter WBC study by Preller and colleagues are neither in line with our findings nor the findings of Tagliazucchi and colleagues. Interstudy discrepancies may be influenced by differences related to psilocin pharmacokinetics (i.v. vs oral psilocybin formulation), BOLD signal denoising procedures (e.g., use of GSR), head motion or analytical method. Although future studies are warranted to more firmly establish psilocybin effects on voxel-level GBC, an important strength of our study resides in the direct linking of WBC with psilocin-induced changes in neurotransmission and phenomenal experience (as indexed by PPL and SDI). The recently proposed "relaxed beliefs under psychedelics (REBUS) and the anarchic brain" theory hypothesizes that a core mechanism of 5-HT2AR agonist psychedelic compounds is impairment of the function of DMN and other high order networks (e.g., ECN), resulting in a less ordered (more entropic) and more desegregated functional architecture. Importantly, we show that psilocin dose-dependently reduces network integrity and segregation (Fig), and our findings are thus compatible with neurobiological mechanisms proposed in the REBUS hypothesis. Two recent studies showed reduced DMN integrity after challenge with the κ-opioid receptor agonist salvinorin A (psychoactive constituent of salvia divinorum)and 3,4-methylenedioxymethamphetamine (MDMA). DMN RSFC was also decreased after separate challenge with ketamine and midazolam. Thus, it is clear that reductions in RSFC within the DMN is not unique to 5-HT2AR agonist psychedelics. However, there is an overlap in the subjective experiences induced by MDMA, salvia divinorum, psilocybin, ketamineand LSD, and it would be interesting to learn the extent to which changes in network integrity (including the DMN) and segregation map onto aspects of phenomenal experience induced by different psychoactive compounds. Interestingly, the GCOR analysis also showed that thalamic GCOR correlated positively with SDI (Fig), compatible with previously observed substantive increases in thalamic RSFC after LSD. These findings are compatible with the thalamic gating theory of psychedelic drug action, which posits that thalamic filtering of external and internal information to the cortex is impaired by psilocin, resulting in information overload. Our observation of SDI-dependent increase in thalamic GCOR indicates that impaired thalamic gating contributes to psychedelic experiences in the context of psilocybin. We replicate our previous finding that PPL and SDI display similar temporal trajectories and correlate positively (Fig 1). Notably, one participant stood out by exhibiting a "break-through" response style, i.e., delayed but rapid increase in SDI ratings (Fig 1 , subject color-coded in pink). We speculate this may be explained at least partly by slow pharmacokinetics in this individual. Our observation that PPL and SDI correlate suggests that PPL is a an objective and quantitative determinant of overall subjective experience after psilocybin. Our results also support our previous observation that SDI correlates positively with retrospective questionnaire assessments of the psychedelic experience, including global ASC questionnaire score, EDI and MEQ30 (Fig). This indicates that SDI measures appraised integrated perceptual changes related to psychedelic phenomenology. Consistent with previous psilocybin pharmacokinetics studies employing similar peroral doses (0.2-0.3 mg/kg), we observed considerable interindividual variability in maximum PPL (C max ) and time to reach C max (t max ) (Table). Considering this pharmacological variability and the strong correlations of PPL with both phenomenal experience and FC changes, it is likely that interindividual differences in subjective effects and clinical response, which sometimes has been ascribed to psychological factors, are at least partly explained by interindividual differences in PPL (R. L.. Thus, our results indicate that future psilocybin studies would benefit from measuring PPL. Our study is not without limitations. A larger sample size would have made it possible to detect more subtle effects. However, multiple measurements were done for each subject over time; such a strategy increases statistical power. A placebo condition could have helped identify potential expectancy effects on RSFC and SDI not tied to psilocybin. Nevertheless, participants were blind to the receipt of psilocybin and dose. Further, our study evaluated associations between RSFC, PPL and SDI, which are not necessarily inflated by an upward bias due to expectancy. It is difficult to imagine how placebo could mimic the tight associations with PPL across a similar time course and magnitude. Head motion correlated positively with PPL (Fig) but was generally of small magnitude, and quality control functional connectivity (QC-FC) plots indicated that noise was successfully removed in included, but not in excluded, datasets as part of the denoising procedure. Nevertheless, head motion effects need to be considered carefully. Hysteresis effects, i.e., different associations during ascent and descent phases, can confound our analysis strategy because we incorporated data points collected throughout the entire acute phase into a linear model. We note we have few data points to firmly exclude hysteresis effects but believe the problem is likely of smaller magnitude given the tight association of PPL and SDI and also the limited hysteresis effects of LSD. In conclusion, the present study evaluated real-time measures of PPL, alterations in phenomenal experience and FC measures of network integrity and segregation, covering ascent, peak and descent phases of medium to high and clinically relevant psilocybin doses in individuals with no or limited experience with psychedelic drugs. These findings demonstrate that psilocin critically shapes the time course and magnitude of changes in the cerebral functional architecture after psilocybin and implicate the expression of network integrity and segregation as important for consciousness and psychedelic experience.

Study Details

Your Library