Computational Modeling of Neural Circuit-Like Coupled Belousov-Zhabotinsky Reactions

Computational Modeling of Neural Circuit-Like Coupled Belousov-Zhabotinsky Reactions A thesis presented to the Interdepartmental Program in Neuroscience. Graduate School of Arts and Science Brandeis University Waltham, Massachusetts By Alexander Mitchell We seek to adapt simplified circuit models of neural central pattern generators (CPGs) for use in a system of chemical oscillators using the Belousov-Zhabotinsky reaction (BZ). The BZ reaction lends itself well to the analysis of these complex oscillators, as it shares a number of properties with model neurons. BZ produces periodic all-or-nothing spike-like behavior, this spike behavior can be induced or prevented by perturbations to the reaction mixture, and critically reactors can be chemically coupled to one another. Microreactors of BZ mixture can be coupled with either hollow channels for the diffusion of aqueous solution (activator coupling) or with PDMS-filled channels for bromine (inhibitor coupling). Using existing computational models for single BZ reactors, we constructed and tested various simple arrangements of wells and chemical couplings that resembled different CPGs from the crustacean stomatogastric ganglion, lamprey spinal cord movement, and others. Using MatLab, we simulated the dynamics of oscillations in these arrangements under different conditions that could be used to disrupt the circuit, such as the presence chemical gradients across channels, light exposure, and the physical dimensions and arrangement of the circuits themselves. We then used these conditions to control the frequency and phase relationships of BZ oscillators. We successfully constructed models for a number of potential circuits and characterized their behavior. We hope that this work will encourage future exploration into applying neuroscience models to further the growing field of BZ research.