Effects of classic psychedelic drugs on turbulent signatures in brain dynamics

This study (n=20) investigated the changes in the brain’s functional hierarchy associated with psilocybin and LSD using a novel turbulence framework that allowed researchers to determine the local level of synchronisation. This framework produced detailed signatures of turbulence-based hierarchical change for each psychedelic drug, supporting the hypothesis that psychedelics modulate functional hierarchy.

Abstract

“Psychedelic drugs show promise as safe and effective treatments for neuropsychiatric disorders, yet their mechanisms of action are not fully understood. A fundamental hypothesis is that psychedelics work by dose-dependently changing the functional hierarchy of brain dynamics, but it is unclear whether different psychedelics act similarly. Here, we investigated the changes in the brain’s functional hierarchy associated with two different psychedelics (LSD and psilocybin). Using a novel turbulence framework, we were able to determine the vorticity, i.e., the local level of synchronisation, which allowed us to extend the standard global time-based measure of metastability to become a local-based measure of both space and time. This framework produced detailed signatures of turbulence-based hierarchical change for each psychedelic drug, revealing consistent and discriminate effects on a higher-level network, i.e., the default mode network (DMN). Overall, our findings directly support a prior hypothesis that psychedelics modulate (i.e., ‘compress’) the functional hierarchy and provide a quantification of these changes for two different psychedelics. Implications for therapeutic applications of psychedelics are discussed.”

Authors: Josephine Cruzat, Yonatan S. Perl, Anira Escrichs, Jakub Vohryzek, Christopher Timmerman, Leor Roseman, Andrea I. Luppi, Agustin Ibanez, David Nutt, Robin Carhart-Harris, Enzo Tagliazucchi, Gustavo Deco & Morten L. Kringelbach

Notes

This study uses the same participants as Carhart-Harris et al., 2016.

Summary

Using a novel turbulence framework, we investigated the changes in the brain’s functional hierarchy associated with two different psychedelics (LSD and psilocybin). The results revealed consistent and discriminate effects on a higher-level network, i.e., the default mode network.

In recent years, psychedelic drug research has made a strong comeback, promising to deliver effective and safe treatments for neuropsychiatric disorders such as treatment-resistant depression and addiction. Yet, a deeper knowledge of how psychedelics function in the human brain is needed to ensure best use of these compounds. Classic psychedelics act by agonizing 5-HT2A receptors, which causes spike-wave-decoupling and dysregulation of spontaneous population level activity. These effects may be mapped to subjective phenomena like ego dissolution and unitive experiences.

The present paper uses a novel turbulence framework inspired by turbulence theory of physics to determine the functional hierarchy of any brain state, by calculating measures of information processing in the human brain inspired by turbulence theory but applied to neuroscience.

A model-based approach was used to explain the statistical dependencies of the neuronal dyna mics observed in the empirical neuroimaging data. This allowed us to demonstrate that different psychedelics work in different ways.

We used fMRI to investigate changes in brain hierarchy induced by different psychedelic drugs in healthy volunteers. The results showed that the brain exhibits a turbulent dynamic intrinsic backbone that facilitates large-scale network communication.

We explored changes in information transmission flow across spatial and temporal scales using four distinct measures: turbulence, information transfer, information cascade flow, and information cascade.

Compared to placebo, psychedelics induce significant increases in turbulence in the brain. These increases were found across all spatial scales but were more pronounced at the longer ones. We found that psychedelics increase information transfer across all spatial scales in the brain, favouring information transmission. We performed the exact computations presented in Figure 2B, but now for the information cascade measurement. We found that LSD and psilocybin increase information cascade flow at all scales compared with placebo, and that psilocybin increases information cascade flow at higher spatial scales. We calculated the node variability of the local synchronization for each condition, and then computed the absolute difference of the node-level turbulence between the psychedelic state and the placebo at =0.12. This allowed us to describe the precise signature of turbulence-based hierarchical change for each psychedelic drug. We applied a whole-brain computational modeling approach to study the effects of LSD and psilocybin on brain activity.

We constructed a whole-brain dynamical model based on a supercritical Hopf bifurcation and used the exponential distance rule as a cost-of-wiring principle. Each model was fitted to optimally reproduce the empirical spatiotemporal dependencies of each brain state. The optimal operating point was determined by the minimum Euclidean distance between the empirical and the simulated functional connectivity. The susceptibility measure and the information encoding capacity measure the sensitivity of the whole-brain model to react to external perturbations. LSD and psilocybin decrease susceptibility and increase information encoding capacity, respectively, following the psychedelic drug administration.

Here we used two independent fMRI datasets to determine the psychedelic-specific modulation of the brain’s dynamic functional hierarchy and assess its impact on information processing. Our results revealed generally consistent increases in turbulent signatures across the psychedelics. Psychedelics have been shown to increase the entropy of spontaneous brain activity, broaden the repertoire of connectivity states, and enhance global connectivity between high-level networks and the rest of the brain. Turbulence related metrics can be used to understand information transfer in the brain.

Compared with placebo, both psychedelics promoted greater information transmission in the brain, specifically in the spatial and temporal domains.

The turbulence framework provides a principled, mechanistic way to describe information transmission across spacetime. Psychedelics appear to increase the transmission of information through long-range spatial scales, i.e., smaller lambdas, which is consistent with their purported ability to facilitate psychological insight.

We explored turbulence in the brain to identify the primary brain areas driving the turbulent dynamic core. We found that psilocybin directly impacts on high-level networks, whereas LSD have stronger effects on primary visual-sensory areas. We used a model-based approach to investigate how the whole-brain dynamics underlying each brain state impacts the brain’s capacity to encode external perturbations. We found that the brain becomes less sensitive to external perturbations under psychedelics, and enhances specificity when dealing with information processing.

Perturbations in dynamical models of whole-brain activity have been shown to dissociate different brain states and reveal the underlying detailed causal mechanisms. The present study used a similar model-based approach to investigate how the information encoding capability of brain states is different in the LSD state. Psychedelic drugs appear to broaden the brain’s dynamical repertoire and enhance global functional connectivity, suggesting that they also promote greater spread of neural activity.

Psychedelics exhibit a particular pattern of turbulent dynamics in the brain that may relate to their characteristic effects on conscious experience.

We applied a novel turbulence framework to two psychedelic datasets in human participants to characterise the functional hierarchy under the acute effects of different compounds.

The National Research Ethics Service Committee approved both studies, and Imperial College London sponsored the research. All participants provided written informed consent prior to participation.

Twenty healthy volunteers were recruited via word of mouth and underwent physical and mental health screening before participating in the LSD study. All participants provided full disclosure of their drug use history in a psychiatric interview. Participants attended two scanning sessions, one with LSD and the other with placebo, at least two weeks apart. They rated their experience on the Visual Analog Scale after each scan.

In a counter-balanced design, 15 participants were injected intravenously with either psilocybin (2 mg dissolved in 10 mL of saline, 60-s injection) or a placebo during two 12-min eyes-closed resting-state fMRI scans. The subjective effects of psilocybin were felt almost immediately after injection and sustained for the remainder of the scanning session.

Imaging was performed on a 3T GE HDx system with fast spoiled gradient echo scans and BOLD-weighted fMRI using a gradient-echo planer imaging sequence. 15 subjects were included in the analysis with 434 TRs each.

Psilocybin Imaging acquisitions were identical to the LSD experiment with the following exceptions: BOLD-weighted fMRI data were acquired at TR/TE = 3000/35ms, Field-of-View = 192mm.

Resting-state pre-processing included removal of the first three volumes, functional realignment, slice-timing correction, identification of outlier scans, normalization to Montreal Neurological Institute (MNI-152) standard space with 2mm isotropic resampling resolution, and spatial smoothing with a Gaussian kernel of 6mm FWHM.

We used the Human Connectome Project (HCP) database comprising diffusion spectrum and T2-weighted neuroimaging data from 32 healthy subjects and processed the data using a q-sampling imaging algorithm implemented in DSI studio. We obtained the structural connectomes from the Schaefer 1000 parcellation.

The level of amplitude turbulence, D, is defined as the standard deviation across time and space of the modulus of local Kuramoto order parameter, R . This measure can be used to characterize the brain vortex space over time. The spatial information transfer shows how the information travels across space at a specific scale, and how the correlation between two brain areas decreases with distance.

The information cascade flow is the stream of information between a given scale and a subsequent lower scale in consecutive time steps. It is computed as the time correlation between the Kuramoto local order parameter in two consecutive scales and times.

We use the discrete version of the node-level Kuramoto order parameter in equation 2 to compute the node variability of local synchronization over a given parcellation. The spatial scaling is defined by the Euclidian distance between the brain areas n and p in MNI space.

We built a whole-brain computational model that describes the transition from noise-induced oscillations to fully sustained oscillations by the normal form of a supercritical Hopf bifurcation. The model is characterized by two parameters: the multiplicative factor, G, and the local bifurcation parameter, aj. The model considered 1,000 cortical brain areas and used the exponential distance rule to obtain the underlying anatomical matrix. The spontaneous local dynamics was described by the normal form of a supercritical Hopf bifurcation.

The frequency of each brain area was determined from the empirical fMRI data. This normal form has a supercritical bifurcation at an = 0.

A functional connectivity measure was applied to the BOLD signal to determine the global coupling factor.

The structure functions are describing the evolution of the functional connectivity (FC) as function of the Euclidean distance between equally distant nodes, and the sensitivity of the Hopf model to perturbations was calculated by measuring the modulus of the local Kuramoto order.

The information encoding capability (I) is the standard deviation across trials of the difference between the perturbed and unperturbed Kuramoto order parameter across time.

We used a permutation-based paired t-test to assess differences between conditions in turbulence, information transfer, flow, and capability. We also used the Wilcoxon rank-sum method to assess differences between conditions in susceptibility and information encoding capacity measures.

AMERICA [ReDLat, supported by National Institutes of Health, National Institutes of Aging, Alzheimer’s Association, Rainwater Charitable foundation – Tau Consortium, and Global Brain Health Institute] G.D., HBP SGA3 Human Brain Project Specific Grant Agreement 3, SGR Research Support Group support, Neurotwin Digital twins for model-driven non-invasive electrical brain stimulation, euSNN European School of Network Neuroscience, Corticity, FLAG – ERA JTC 2017, and Center for Music in the Brain are supported by the authors.

The turbulence framework uses model-free and model-based measures to determine the functional hierarchy in brain dynamics. It uses a Stuart Landau non-linear oscillator to describe the dynamics of each brain area and is connected through the anatomical connectivity to simulate the global dynamics induced by each psychedelic.

The seven Yeo resting-state networks on the human brain are rendered, and the turbulence framework is used to measure the degree of participation of the Yeo RSNs.

Within a model-free framework, we investigated how psychedelics modulate information processing at the whole-brain level, characterized by turbulence and information transfer measurements. We found that both LSD and Psilocybin increase the level of turbulence compared to placebo, particularly at larger spatial scales.

Psychedelics increase information transfer across space at all spatial scales. Psilocybin shows a steeper slope-scale than LSD, denoting higher sensitivity to differences in the degree of transmission of information across scales.

Figure 3 shows that under the psychedelic state, information flow across scales and information cascade flow increases, and that the information flow across scales increases for both LSD and Psilocybin, compared to placebo.

We computed the node variability of the local synchronization for each psychedelic and placebo, at each spatial scale. The differences were greatest at larger spatial scales, and the most significant differences were found in the SOM and DAN networks.

We used a model-based approach to assess the effects of external perturbations on brain dynamics. We found that both psychedelics drugs significantly decreased the susceptibility level and increased the information capacity compared to placebo.