Locomotion emerges due to interactions between various subsystems such as sensory systems for vision. Rather than analyzing all the subsystems in detail which would be very complex, we use a metric that captures some property of the interactions between them. Our chosen metric is centralization which measures the coupling between neuromechanical modules that mediate responses to perturbations. This metric has been identified as an important parameter for assessing locomotion and has been used to describe both robotic and biological systems. However, little has been done to investigate how this metric may change with different circumstances such as terrain complexity and speed.
Muscle action potentials from the hind and mid limbs as well as the tarsi positions are recorded from freely running cockroaches over terrain of varying roughness. From this we can use an empirical, model-free, information theory-based measurement to estimate how centralized the neuromechanical architecture of the cockroach is.
The results from this study hope to help us further understand how cockroach’s navigate complex terrain as well as suggest which regimes could be more advantageous.