UC Davis

Recurrent event data, characterized by repeated occurrences of events associated with one or more subjects, is prevalent in various fields, including neuroscience, healthcare, and social science. Recent technological advancements in neuroscience have significantly increased the availability of such data, presenting both new opportunities and challenges. This dissertation focuses on the statistical analysis of two common types of recurrent event data in neuroscience: neural firing activity and functional connections between neurons. For each type of data, we discuss the unique challenges posed by the data and propose efficient statistical methods to address the challenges. The proposed methods aim to identify groups of neurons with similar activity patterns while accommodating temporal disparities among neurons. We establish conditions for the identifiability of model parameters. We conduct extensive numerical experiments to evaluate the empirical performance of the proposed methods. Additionally, we apply the proposed methods to real-world neural data to reveal distinct roles of neurons and identify representative neural activity patterns.