The brain is a highly interconnected system in which molecular, cellular, anatomical, and functional networks interact across multiple levels of organization. Understanding how these networks are structured and how their interactions shape brain function requires methods capable of capturing relationships within complex, high-dimensional biological data. For this thesis, two computational tools were developed to facilitate network construction, analysis, and visualization.
NETSCOPE uses information-theoretic measures to reconstruct networks from molecular datasets. It was applied to gene-expression data from the mouse barrel cortex to study the regulatory networks underlying both cortical laminar differentiation, and experience-dependent plasticity. The resulting networks revealed modular patterns with key genes forming highly interconnected network hubs.
NeuroCarta uses publicly available data to reconstruct and analyze whole-brain, mesoscale anatomical networks. It's application revealed predominantly local and intrahemispheric connectivity, alongside hubs involved in the integration and distribution of information. Reconstruction of cell type-specific networks was used to investigate neuromodulation in the mouse whisker system, identifying previously unreported links and widespread inhibitory pathways.
Together, these approaches demonstrate how network analysis can link different levels of brain organization and provide a framework for understanding the brain as an integrated, multiscale system.