Excitation-Inhibition Balance: A Study of Brain Network Dynamics Using Statistical Mechanics

The brain is a complex dynamical system whose functional properties are largely determined by the characteristics of its neurons and patterns of synaptic connectivity. Neural activity coordinat-ed across different scales from neuronal circuits to large-scale brain networks gives rise to complex cognitive functions. Bridging the gap between micro- and macro-scale processes, we present a novel framework based on the Maximum Entropy model from statistical physics, inferring a structurally informed network of brain function named the hybrid resting state structural connectome (rs-SC). This new hybrid network is shown to outperform one inferred from an unconstrained model in simulating brain dynamics and reconstructing functional correlations. The rs-SC is used to probe macro-level excitation-inhibition (E/I) balance in 76 cognitively intact middle-aged apolipoprotein E (APOE) ε4 carriers with non-carriers (16M/22F), demonstrating a shift in E/I balance in female carriers that may confer greater vulnerability to Alzheimer’s disease (AD) neuropathology. Applying the framework to a large neuroimaging database of 1326 subjects at various stages along the AD spectrum, graph theoretical analysis was used to calculate global and local efficiency on the FC network (measures of network integration and segregation) while a measure of excitation-inhibition (E/I) balance was computed on the rs-SC network. Our results indicate a strong associa-tion between APOE and AD (status based on the clinical dementia rating scale) for males with respect to global efficiency, but not females, whereas global E/I balance showed strong association with females, but not males. Further, at a local level, E/I balance was significant in the hippocampus, amygdala, and frontal gyrus for females with APOE- ε4 as well as the anterior parahippocampal gy-rus for females with suspected AD. Collectively, these results suggest that while traditional measures of functional connectivity are commonplace in neuroimaging, novel methods based on microscale neuronal activity reveal more granular network-level vulnerabilities that may serve as biomarkers of micro-to-macro dysfunction before the onset of cognitive impairment.