Structural Characterization of Oscillations in Brain Networks with Rate Dynamics

TL;DR

This paper provides structural conditions for oscillations in neural networks using a linear-threshold model, linking network structure and inputs to oscillatory behavior through stability analysis.

Contribution

It introduces a systematic framework to determine oscillations in neural networks based on network structure and external input, using a classical neural mass model.

Findings

Necessary and sufficient conditions for oscillations in various network configurations.

Connection between lack of stable equilibria and oscillatory activity.

Numerical validation of the stability-based approach.

Abstract

Among the versatile forms of dynamical patterns of activity exhibited by the brain, oscillations are one of the most salient and extensively studied, yet are still far from being well understood. In this paper, we provide various structural characterizations of the existence of oscillatory behavior in neural networks using a classical neural mass model of mesoscale brain activity called linear-threshold dynamics. Exploiting the switched-affine nature of this dynamics, we obtain various necessary and/or sufficient conditions on the network structure and its external input for the existence of oscillations in (i) two-dimensional excitatory-inhibitory networks (E-I pairs), (ii) networks with one inhibitory but arbitrary number of excitatory nodes, (iii) purely inhibitory networks with an arbitrary number of nodes, and (iv) networks of E-I pairs. Throughout our treatment, and given the arbitrary dimensionality of the considered dynamics, we rely on the lack of stable equilibria as a system-based proxy for the existence of oscillations, and provide extensive numerical results to support its tight relationship with the more standard, signal-based definition of oscillations in computational neuroscience.