A multilayer network model of neuron-astrocyte populations in vitro reveals mGluR₅inhibition is protective following traumatic injury

Abstract

Despite recent advances in understanding neuron-astrocyte signaling, little is known about astrocytic modulation of neuronal activity at the population level, particularly in disease or following injury. We used high-speed calcium imaging of mixed cortical cultures in vitro to determine how population activity changes after disruption of signaling and mechanical injury. We constructed a multilayer network model of neuron-astrocyte connectivity, which captured unique topology and response behavior not evident from analysis of single cell type networks. mGluR5 inhibition decreased neuronal, but not astrocytic, activity and functional connectivity following traumatic injury, and also altered higher-order topological properties of multilayer networks. Comparison of spatial and functional community structure revealed that astrocyte segments of the same cell often belong to separate functional communities based on neural connectivity. Our findings demonstrate the utility of multilayer network models for characterizing the multiscale connectivity of two distinct but functionally dependent cell populations.

Neuronal Graphs: A Graph Theory Primer for Microscopic, Functional Networks of Neurons Recorded by Calcium Imaging

“…In silico controls allow scientists to compare neuronal graphs with random graphs that have similar properties, e.g., edge, degree distribution, etc., but lack any controlled organisation. Such comparisons can be used to (a) confirm that properties of neuronal graphs are statistically significant; (b) provide a baseline from which different experiments can be compared; and (c) be used to guide the formation on new computational models that lead to a better understanding of neural mechanisms of computation (e.g., Schroeder et al, 2019 ).…”

Neuronal graphs: a graph theory primer for microscopic, functional networks of neurons recorded by Calcium imaging

“…Such comparisons can be used to a) confirm that properties of neuronal graphs are statistically significant; b) provide a baseline from which different experiments can be compared and c) be used to guide the formation on new computational models that lead to a better understanding of neural mechanisms of computation, e.g. [60].…”