Multiscale dynamic mean field model to relate resting-state brain dynamics with local cortical excitatory-inhibitory neurotransmitter homeostasis in health and disease

Abstract

Previous neuro-computational studies have established the connection of spontaneous resting-state brain activity with "large-scale" neuronal ensembles using dynamic mean field approach and showed the impact of local excitatory−inhibitory (E−I) balance in sculpting dynamical patterns.Here, we argue that whole brain models that link multiple scales of physiological organization namely brain metabolism that governs synaptic concentrations of gamma-aminobutyric acid (GABA) and glutamate on one hand and neural field dynamics that operate on the macroscopic scale. The multiscale dynamic mean field (MDMF) model captures the synaptic gating dynamics over a cortical macrocolumn as a function of neurotransmitter kinetics. Multiple MDMF units were placed in brain locations guided by an anatomical parcellation and connected by tractography data from diffusion tensor imaging. The resulting whole-brain model generates the resting-state functional connectivity and also reveal that optimal configurations of glutamate and GABA captures the dynamic working point of the brain, that is the state of maximum metsatability as observed in BOLD signals. To demonstrate test-retest reliability we validate the observation that healthy resting brain dynamics is governed by optimal glutamate-GABA configurations using two different brain parcellations for model set-up. Furthermore, graph theoretical measures of segregation (modularity and clustering coefficient) and integration (global efficiency and characteristic path length) on the functional connectivity generated from healthy and pathological brain network studies could be explained by the MDMF model. In conclusion, the MDMF model could relate the various scales of observations from neurotransmitter concentrations to dynamics of synaptic gating to whole-brain resting-state network topology in health and disease.Key words: GABA, Glutamate, metastability, resting-state functional connectivity, structural connectivity, network measures, epilepsy, and schizophrenia. and structured resting-state dynamics of brain. The problem is further compounded by the fact that brain parameters, such as the proportion of different neurons and neurotransmitter types (glutamate, gamma-aminobutyric acid (GABA), etc.) with their associated synaptic properties cannot be manipulated independently in living system to delineate their role in brain function (Prinz, 2008). Due to these in vivo practical constraints, physiologically realistic whole brain network models are favourable candidates to overcome and manipulate experimentally inaccessible parameter(s) of neuronal networks in brain. Following this line of reasoning, mean field approach, consisting of coupled stochastic differential equations could provide appropriate neuro-computational framework to study missing link between micro-scale state variables with whole-brain network dynamics (Breakspear et al., 2017). Moreover, the computational brain network modeling could be used to explore the underlying mechanisms of human brain functions, such as at the resti...

Brain Connectivity Studies on Structure‐Function Relationships: A Short Survey with an Emphasis on Machine Learning

“…By investigating cortical heterogeneity across the entire brain, the authors explained how fluctuations around a dynamic equilibrium state of a core brain, represented by eight identified brain areas, could drive functional network state transitions [ 2 , 93 , 94 ]. Hence, in the work of Naskar et al [ 95 ], the resting brain is considered to represent a dynamically metastable system with frequent switches between several metastable states driven by multiplicative noise. Metastability can be quantified in such coupled oscillator systems by the standard deviation of one of the two time-dependent Kuramoto order parameters, which measures the phase coherence of a set of oscillators in the weak coupling limit [ 2 ].…”