COMPUTATIONAL SEGMENTATION OF WHITE MATTER ANATOMY: METHODS, INSIGHTS, AND STANDARDS

Abstract

The brain is fundamentally an information processing and behavioral control system. The key to achieving this role is the ability to move information about the brain in a fast, reliable, and organized fashion. The axons of neurons stand as the primary means of achieving this in the brain. However, as brains became larger and more gyrified the routes between grey matter structures became correspondingly longer and more complicated. This has required axons to form a complex network of bundles in order to maintain connectivity between distal regions, giving rise to the tissue known as “white matter”. Although the resultant architecture has been studied for hundreds of years, much is still unknown. This has hindered efforts to associate characteristics of the white matter with human behavior, development, and disorders. Here, we seek to ameliorate this. In this thesis, we present a three component body of work designed to help shed light on the white matter. The first component clarifies several white matter tracts which may facilitate a more complex system of information processing between the human dorsal (“what”) and ventral (“where”) visual streams. The second is a comprehensive review of our contemporary understanding of gross white matter architecture, featuring considerations of mutual insights from human and non-human primate studies, as well as apparent discrepancies between accounts. This work responds to recent calls for the formation of a consensus regarding the ontology and taxonomy of white matter. The final component responds to calls for more transparent and well-documented digital white matter segmentation methods, and is an interactive, online resource. It serves as both an educational resource and a transparent documentation of methodology. Ultimately, it is hoped that this body of work will support research in the field of white matter anatomy, across a broad range of approaches and endeavors.