UC Irvine

Computational modeling of neuronal networks enables the study of variables in isolation while approximating the biological state. In conditions such as the exposure to radiation and epilepsy, a large number of structural and network parameters are altered, making the role of any individual abnormality unclear. The goal of the presented work is to develop realistic computational models of the hippocampus for the incorporation of experimental observations and to shed light on the relative importance and deleteriousness of pathological alterations. The Introduction summarizes our motivations and the utility of realistic computational models, particularly for the hippocampus, and gives a background for the abnormalities present in the irradiated and epileptic conditions. Chapter 1 describes our work translating our previous model to be compatible with parallel computing and then expanding the model to run at full-scale, with over a million neurons. It then outlines a methodology for the generation of computational models for the dendritic trees of granule cells, the most prevalent cell type in the dentate gyrus, using previous tools that contained a microscopic focus. In Chapter 2, we report an entirely new methodology for morphology generation that instead shifts the focus to the macroscopic neuroanatomy, growing dendrites within a realistic three-dimensional structure and enabling the population-level study of morphology. Chapter 3 describes a study in which our computational modeling was used to interpret experimental observations in area CA1 of the hippocampus after exposure to proton radiation. The study reports long-term but subtle changes in the passive properties of pyramidal neurons, the principal excitatory cell type in CA1, which were found in a computational model to have a surprisingly dramatic effect on network function. Chapter 4 reviews the ever-growing observations of the non-recurrent microscopic nature of seemingly repetitive macroscopic events in epilepsy. Chapter 5 provides the immediate next steps and future directions for computational modeling in health and disease, building on the foundation and framework provided in the previous chapters to suggest several avenues to bring computational models ever closer to the experimental, biological, and clinical conditions.