Model of a central pattern generator using a spiking neural network, included different types of neurons

Abstract

The performance of various locomotor acts in animals is ensured by the central pattern generators (CPG). The CPG is an efficient system that is capable of independent generation of activity patterns, while the operation of the CPG can be corrected due to sensory feedback. Such properties of CPG are successfully used in robotics. In this paper, we propose a half-central CPG model using Izhikevich neurons. We selected some parameters for CPG neurons, thereby obtaining membrane properties that are close to those of real brain cell membranes that were found in natural neuronal networks. The pacemaker neuron is responsible for generating the rhythm in the network, it is able to generate spikes with a frequency range of 0.3–7.7 Hz, depending on the input signal, which affects the speed of the locomotor cycle. The model is implemented in the Arduino software environment, and in the future, it can be interfaced with a robotic installation that performs rhythmic locomotor patterns.