Computational model of a behaviour in c. elegans and a resulting framework for modularizing dynamical neuronal structures

The work presented in this dissertation grew out of a study of a physiologically based computational model of the tap withdrawal response in the nematode Caenorhabditis elegans. A computational model using all available anatomical and physiological data was unable to explain a dynamic property of the circuit: the ability of the behaviour to continue after the termination of the stimulus. To account for this behavioural observation, a novel approach was taken: a neuronal circuit was engineered from a set of modules each consisting of several physiologically realistic model cells. The mathematical dynamics of the resulting neuronal circuit produced an output that was similar to the behaviour observed in the intact worm and shows that neuronal network dynamics could account for the behaviour. In the course of this study, it became clear that little is known about the modular properties of neuronal dynamics. This dissertation presents an approach for combining non-linear neuronal circuits into larger systems using dynamical modules (dymods), and a set of tools for studying dymods, and discusses a research strategy for studying the modular properties of neuronal dynamics.