| Description |
Title: "Neural network model of spatial sound source segregation"
Committee:
Dr. H. Steven Colburn (Co-Advisor, Chair)
Dr. Kamal Sen (Co-Advisor)
Dr. Gerald Kidd
Dr. Barbara Shinn-Cunningham
Abstract:
The cocktail party problem presents the challenge of how we can understand individual sound sources in complex listening environments. Though it has been studied extensively, the neural mechanisms are not understood and hearing impaired individuals experience difficulty with the cocktail party problem in noisy, everyday listening environments. This thesis project aims to create a neural network model of the cortical circuitry behind spatial sound source segregation, with the goals of using the model to provide mechanistic insights in neural processing and potentially implementing it in hearing assistive devices to help users overcome noisy, everyday listening environments. The modeling is implemented in MATLAB using basic neural modeling approaches and signal processing techniques, incorporating frequency and temporal processing in the neural and signal domain. |
