Closed-loop control of a minimal central pattern generator network

Abstract

Many central pattern generators (CPGs) are built on a basic circuit of reciprocally inhibitory neurons, the minimal configuration that can produce distinct rhythmic patterns for controlling antagonistic muscles. In this paper, we use a closed-loop to control the rhythm of a CPG model based on this minimal network architecture. The closed-loop adapts the maximum conductance of the inhibition to achieve a regularized rhythm departing from an otherwise highly irregular spiking–bursting activity. Our results show that the closed-loop conductance control can rapidly find the best connectivity that leads to the regularization goal. We also discuss several external stimuli that lead to regularized CPG rhythms. We argue that the proposed closed-loop approach can be used to control CPG rhythms in a wide variety of applications in experimental and theoretical research.