Centralization of Muscular Control in Blaberus discoidalis as a Variable of Terrain Roughness

Abstract

The purpose of this study was to illustrate how the neuromuscular architecture of Blaberus discoidalis cockroaches changes in response to variations in their environment. The species Blaberus discoidalis shares the single motor neuron physiology of Periplaneta americana, widely used in past research in the field, and thus the extensions of its leg muscles were easy to record via electromyography (EMG). By tracking the behavior of the insect’s legs, inferences can be made about how the animal’s body and brain compensate for perturbations to its running gait. Understanding this control architecture could lead to more robust locomotion systems in robotic design. In this experiment, the insects ran over terrain of varying roughness, quantified by the standard deviation (cm) of a gaussian distribution of pillars. While they ran, EMG electrodes recorded muscle activation in the rear and mid left legs, and a high-speed camera recorded a video in the top-down plane. Using mutual information as a metric, a K nearest neighbor algorithm derived centralization based on the number and timing of the action potentials in each stride. It was hypothesized that the level of centralization exhibited by the insects will increase as the roughness of the terrain they run over increases. The results showed a near linear increase in centralization as terrain roughness increased, supporting the hypothesis. This suggests that greater perturbations in gait causes muscle control architecture to become more centrally mediated. Future research should be conducted to confirm these results and explore how this response interacts with other stabilizing behaviors.