Data-Driven Modeling Approach to Studying the Relationship between Structure and Behavior of Neuronal Networks

Abstract

Novel experimental techniques in neuronal networks have made high dimensional datasets of synaptic connectivity and neuronal activity available, which open new vistas into the immense complexity of neuronal networks. The curse of dimensionality in connectivity and activity patterns, however, poses severe challenge to measuring and finding statistical relationships between their patterns. One approach to addressing this problem is applying dimensionality reduction processes to reduce the descriptions of connectivity and activity, and then building appropriate models to link such reduced descriptions. In this talk, I will present my data-driven modeling work based on such an approach. Our goal is to provide models used to study or predict the relationship between low-dimensional structure and behavior of neuronal networks. ?