On bursting solutions near chaotic regimes in a neuron model

Abstract

In this paper, we use mathematical analysis to study the transition of dynamic behavior in a system of two synaptically coupled Hindmarsh-Rose (HR) neurons, based on its flow-induced Poincaré map. Numerical simulations have shown that the individual HR neuron has chaotic behavior, but neurons become regularized when coupled. Using a geometric method for dynamical systems, we begin with an investigation of the bifurcation structure of its fast subsystem. We show that the emergence of regular patterns out of chaos is due to a topological change in its underlying bifurcations. Then we focus on the transitional phase of coupling strength, where the bursting solutions need to pass near two homoclinic bifurcation points located on a branch of saddle points, and we study the flow-induced Poincaré maps. We observe that as the gap between the homoclinic points narrows, the image of the return map moves away from chaotic regions where winding numbers vary abruptly. That, along with Lyaponov exponent calculations, reveals the fine dynamics in the pathway across chaotic bursting behavior and regular bursting of coupled HR neurons as the synaptic coupling strength of the neurons increases. The main contribution of this paper is the mathematical description of the transition of synaptically coupled neurons from chaotic trajectories to regular burst phases using dynamical system tools such as Poincaré maps.