Dynamical exploration of the repertoire of brain networks at rest is modulated by psilocybin
Louis-David Lord, Paul Expert, Selen Atasoy, Leor Roseman, Kristina Rapuano, Renaud Lambiotte, David J. Nutt, Gustavo Deco, Robin L. Carhart-Harris, Morten L. Kringelbach, Joana Cabral,
NeuroImage, 2019, 199, 127–142
A B S T R A C T
Growing evidence from the dynamical analysis of functional neuroimaging data suggests that brain function can be understood as the exploration of a repertoire of metastable connectivity patterns (‘functional brain networks’), which potentially underlie different mental processes. The present study characterizes how the brain’s dynamical exploration of resting-state networks is rapidly modulated by intravenous infusion of psilocybin, a tryptamine psychedelic found in “magic mushrooms”. We employed a data-driven approach to characterize recurrent functional connectivity patterns by focusing on the leading eigenvector of BOLD phase coherence at single-TR resolution. Recurrent BOLD phase-locking patterns (PL states) were assessed and statistically compared pre- and post-infusion of psilocybin in terms of their probability of occurrence and transition profiles. Results were validated using a placebo session. Recurrent BOLD PL states revealed high spatial overlap with canonical resting-state networks. Notably, a PL state forming a frontoparietal subsystem was strongly destabilized after psilocybin injection, with a concomitant increase in the probability of occurrence of another PL state characterized by global BOLD phase coherence. These findings provide evidence of network-specific neuromodulation by psilocybin and represent one of the first attempts at bridging molecular pharmacodynamics
and whole-brain network dynamics.
Brain dynamics can be understood as the exploration of activity configurations over both space and time (Tononi and Edelman, 1998; Watanabe et al., 2014; Cabral et al., 2017b; Gu et al., 2018). This exploration may be defined in terms of trajectories within a fixed repertoire of metastable activity patterns, which potentially underlie different brain processes. Indeed, a robust repertoire of large-scale functional networks has been consistently detected across individuals and neuroimaging modalities, not only during task performance but also during rest (Damoiseaux et al., 2006; De Luca et al., 2006; Mantini et al., 2007; Seeley et al., 2007; Musso et al., 2010; Brookes et al., 2011; Yeo et al., 2011; Hipp et al., 2012), indicating that brain function involves the coordinated integration of information over a repertoire of spatially distributed networks of specialized brain areas (Varela, 1979; Bressler and Menon, 2010; Menon, 2011; Cavanna et al., 2017; Lord et al., 2017). While the mechanisms driving the spontaneous formation and dissolution of functional networks remain under debate (Cabral et al., 2017a), brain function has recently been explored in terms of transitions between recurrent states of functional connectivity (Hansen et al., 2015; Cabral et al., 2017a; Cavanna et al., 2017; Tewarie et al., 2019). Although the functional relevance of these explorative dynamics during rest remains unclear (Christoff et al., 2016), recent evidence suggests that transitions between brain states are organized in a hierarchical manner (Vidaurre et al., 2017) and relate to cognitive function (Cabral et al., 2017b).
Novel insights into the neurobiological correlates of functional network states and their relationship to behavior in both health and disease can be gained by understanding how specific psychoactive compounds modulate the relative stability of functional networks over time, as well as the transitions between them. Toward this aim, the present study investigates how psilocybin (4-phosphoryloxy-N,Ndimethyltryptamine) changes the exploration of the brain’s dynamical repertoire. Psilocybin is a prodrug of psilocin (4-OH-N,N-dimethyltrayptamine) – a psychoactive ingredient found in so-called “magic mushrooms” and classical tryptamine ‘psychedelic’. The subjective effects of psilocybin/psilocin include broadly unconstrained perception and cognition, hyper-associative cognition and, at higher doses, a breakdown in the perception of time, space and selfhood (Griffiths et al., 2006; Carhart-Harris et al., 2014). Psilocybin provides a promising experimental framework for linking molecular pharmaco-dynamics to changes in the brain’s dynamical repertoire because its potent psychoactive effects are selectively due to its agonist activity at the serotonin 2A (5-HT2A) receptor (McKenna et al., 1990; Vollenweider et al., 1998; Passie et al., 2002; Gonzalez-Maeso et al., 2007) – see Beliveau et al. (2017) for the highest resolution in vivo mapping of 5-HT2A receptor densities in the human brain. Furthermore, investigating how psilocybin modulates the exploration of functional network states over time may help better understand the functional mechanisms underlying the recently demonstrated therapeutic potential of psilocybin (and other indolealkylamines) for disorders including depression, anxiety and addiction (McKenna et al., 1990; Griffiths et al., 2006; Grob et al., 2011; Meltzer et al., 2012; Johnson et al., 2014; Carhart-Harris et al., 2016).
The first fMRI investigation of psilocybin’s effects on the human brain consisted of a task-free paradigm in which healthy subjects were intravenously injected with the compound inside the scanner (or a placebo, at least 1 week apart) to characterize changes in brain activity under the influence of the drug (Carhart-Harris et al., 2012). The first dynamical analysis of this dataset found increased variance of intra-network synchrony over time for a number of canonical resting-state networks under psilocybin (Carhart-Harris et al., 2014). Somewhat consistently, a subsequent analysis of this same fMRI dataset using sliding-window correlations within a specific network of four brain regions (comprising the bilateral hippocampus and anterior cingulate cortex) revealed both a larger repertoire of functional motifs under psilocybin as well as greater entropy of the motif sequence (Tagliazucchi et al., 2014). Another experiment using MEG revealed increased Lempel-Ziv complexity -a measure of neural signal diversity calculated at the single-channel level (Lempel and Ziv, 1976)- following psilocybin administration relative to placebo (Schartner et al., 2017). However, it is important to clarify that an increase in local complexity does not necessarily imply higher randomness at the larger scale. This seeming paradox can be explained by the findings in a recent study (Atasoy et al., 2018), where higher coherence – expressed as increased power and energy of connectome harmonic brain states under psilocybin – was accompanied by an enlarged repertoire of brain states, leading to higher spatial and temporal variability. Additionally, other works have shown that the serotonergic psychedelics psilocybin and LSD induce an alternative type of functional integration characterized by greater global integration (Petri et al., 2014; Roseman et al., 2014; Tagliazucchi et al., 2016). However, despite the consistent advancements, how psilocybin modulates the relative occupancy of specific functional networks over time and how it relates to the subjective psychedelic experience has not yet been explored.
Recently, a number of methodological approaches have been proposed to analyze BOLD connectivity dynamics at high temporal resolution (i.e., single volume/TR), focusing either on BOLD co-activation patterns (Tagliazucchi et al., 2012; Liu and Duyn, 2013; Karahanoglu and Van De Ville, 2015), or on BOLD phase coherence patterns (Glerean et al., 2012; Hellyer et al., 2015; Cabral et al., 2017b). While co-activation approaches (in their variant forms) are only sensitive to simultaneity in the data, phase coherence techniques can, by definition, capture temporally delayed relationships, which may be more sensitive to capture the ultra-slow oscillatory processes governing the formation of resting-state networks – as proposed by recent experimental and computational studies (Deco et al., 2009; Cabral et al., 2014a, 2017a; Ponce-Alvarez et al., 2015; Deco and Kringelbach, 2016; Gutierrez-Barragan et al., 2018; Roberts et al., 2019).
To identify recurrent patterns of BOLD phase coherence across subjects and quantify differences in the exploration of the repertoire of functional networks in the current dataset, we employed a recently developed data-driven approach, the Leading Eigenvector Dynamics Analysis (LEiDA), which captures instantaneous phase-locking patterns with reduced dimensionality by considering only the relative phase of BOLD signals (i.e., how all BOLD phases project into their leading eigenvector at each discrete time point) (Cabral et al., 2017b; Figueroa et al., 2019). LEiDA appears as a valuable step in the search for functional network recurrences in dynamical analysis because the reduced dimensionality (from an NxN matrix to a 1xN vector) allows for better convergence of the clustering algorithm, revealing robust BOLD phase-locking patterns that consistently reoccur across fMRI scans for different subjects. However, despite returning meaningful functional subsystems in previous studies (Cabral et al., 2017b; Figueroa et al., 2019), the recurrent patterns identified by LEiDA have not yet been qualitatively compared to canonical resting-state networks (Damoiseaux et al., 2006; Yeo et al., 2011).
In the current work, we therefore employed the LEiDA approach to identify recurrent BOLD phase-locking patterns (PL states) and quantified differences in their probability of occurrence and transition profiles before and after the infusion of psilocybin, while subjects were inside the MRI scanner. The validity of the results was subsequently verified using the placebo dataset, and the repertoire of PL patterns returned by LEiDA was compared to well-established resting-state networks described in the literature.