Mariia Iudina: Exploring Optimal Control Strategies for a State-Switching Task for the Wilson-Cowan Model of Neural Population Dynamics

BCCN Berlin / Technische Universität Berlin

Abstract

 

This work investigates the Wilson-Cowan model, which describes the dynamic interactions between excitatory and inhibitory neural populations in neural networks. The study examines the model's parameters and behavior. The application of Optimal Control Theory to the model is discussed, with a particular focus on the adjoint method. Experiments focus on the bistable regime of the Wilson-Cowan model using specific parameter pairs. The results show that limited-strength input pulses can effectively shift population activity to a desired state. The research concludes that the strength of the control signal and the initial state of the system are critical factors in controlling neural dynamics.

 

Additional Information

Master Thesis Defense

 

Organized by

Prof. Dr. Klaus Obermayer   & Prof. Dr. Henning Sprekeler   / Lisa Velenosi