LSD and psilocybin flatten the brain’s energy landscape: insights from receptor-informed network control theory

Abstract

Psychedelics like lysergic acid diethylamide (LSD) offer a powerful window into the function of the human brain and mind, by temporarily altering subjective experience through their neurochemical effects. The RElaxed Beliefs Under Psychedelics (REBUS) model postulates that 5-HT2a receptor agonism allows the brain to explore its dynamic landscape more readily, as suggested by more diverse (entropic) brain activity. Formally, this effect is theorized to correspond to a reduction in the energy required to transition between different brain-states, i.e. a ″flattening of the energy landscape.″ However, this hypothesis remains thus far untested. Here, we leverage network control theory to map the brain′s energy landscape, by quantifying the energy required to transition between recurrent brain states. In accordance with the REBUS model, we show that LSD reduces the energy required for brain-state transitions, and, furthermore, that this reduction in energy correlates with more frequent state transitions and increased entropy of brain-state dynamics. Through network control analysis that incorporates the spatial distribution of 5-HT2a receptors, we demonstrate the specific role of this receptor in flattening the brain′s energy landscape. Also, in accordance with REBUS, we show that the occupancy of bottom-up states is increased by LSD. In addition to validating fundamental predictions of the REBUS model of psychedelic action, this work highlights the potential of receptor-informed network control theory to provide mechanistic insights into pharmacological modulation of brain dynamics.

Prefrontal contributions to the stability and variability of thought and conscious experience

“…The REBUS model proposes a unifying account of how serotonergic psychedelic compounds affect conscious experience [32]. Serotonergic psychedelics are a specific class of psychoactive compounds including but not limited to lysergic acid diethylamide (LSD), psilocybin, and dimethyltryptamine (DMT), that principally act upon cortical serotonin 2A receptors [54][55][56][57][58][59][60][61][62][63][64] to achieve their profound psychological effects [32,65,66] and therapeutic potential [67][68][69][70][71][72]. Moving forward, we use the term psychedelics to refer to these serotonergic psychedelics.…”

Larger lesion volume in people with multiple sclerosis is associated with increased transition energies between brain states and decreased entropy of brain activity

“…As previously described 16,17 , NCT can be used to understand how the structural (white matter) connectome constrains dynamic brain state changes. Specifically, it can be used to compute the minimum energy ( E m ) required to transition between pairs of brain states or to remain in the same state.…”

“…This approach was successfully used in healthy controls (HC) to identify dynamic brain states that occur while performing a working memory task as well as quantify the effects of psychedelics. 16,17 In addition to the identification of the brain dynamic states, the same studies used the network control theory (NCT) approach [18][19][20] to identify the minimum energy required to transition between these dynamic brain states. However, no study to date has applied brain activity clustering and NCT in a population of pwMS, let alone investigate differences in brain dynamics and energetics across disability subgroups in MS. Entropy of brain activity over time is a fundamental metric to quantify the amount of regularity/unpredictability in BOLD time series.…”

Network Analysis of Time Series: Novel Approaches to Network Neuroscience

“…While both have been informative (discussed blow), a common limitation of these measures is typically small number of "states" (often < 10), which is likely not enough to fully characterize a manifold. Combined slidingwindows functional connectivity analysis and k-means clustering can reveal differences in the state transition structure of the brain under different conditions (Li et al, 2019;Lord et al, 2019;Schumacher et al, 2019;Singleton et al, 2021), however a statespace model of 5-10 unique "states" may be too impoverished to extract detailed information about the structure of the intrinsic neural manifold. Hidden Markov models have been similarly applied to multiple neuroimaging modalities, including spiking networks in rodents (Jones et al, 2007), MEG data from humans (Baker et al, 2014), and fMRI data from humans (Eavani et al, 2013;Chen et al, 2016).…”