Error Function of Mathematical-Biological Hybrid Central Pattern Generator Optimization
Jassem N Bourahmah
Georgia State University USA
Co-Author(s): Jassem Bourahmah, Deniz Alacam, Akira Sakurai, Paul Katz and Andrey Shilnikov
Abstract:
Central pattern generators (CPGs) are biological neural circuits that produce rhythmic outputs in the absence of rhythmic input. We previously developed a detailed mathematical model of the CPG that controls the swimming behaviors of the sea slug Melibe leonina. The math-CPG model can produce network bursting with specific rhythmic patterns and realistic responses to perturbations recorded in the bio-CPG. We have created a hybrid system that uses electrophysiological recordings to train and optimize the computational model of the Melibe swim CPG. We also developed an iterative optimization algorithm to automate the calibration of the biologically plausible models of individual neurons and slow and fast chemical synapses. To create an effective error function for the optimization algorithm we combine and weigh 5 features of neural activity against the biological and math-CPG.