Mathematics · Geometry and Topology
Geometric representation of neuroanatomical data observed in mouse brain at cellular and gross levels
See on the mapAbstract
This dissertation studies two problems related to geometric representation of\nneuroanatomical data: (i) spatial representation and organization of individual neurons,\nand (ii) reconstruction of three-dimensional neuroanatomical regions from sparse two-dimensional\ndrawings. This work has been motivated by nearby development of new\ntechnology, Knife-Edge Scanning Microscopy (KESM), that images a whole mouse\nbrain at cellular level in less than a month.\nA method is introduced to represent neuronal data observed in the mammalian brain at\nthe cellular level using geometric primitives and spatial indexing. A data representation\nscheme is defined that captures the geometry of individual neurons using traditional\ngeometric primitives, points and cross-sectional areas along a trajectory. This\nrepresentation captures inferred synapses as directed links between primitives and\nspatially indexes observed neurons based on the locations of their cell bodies. This\nmethod provides a set of rules for acquisition, representation, and indexing of KESMgenerated\ndata.\nNeuroanatomical data observed at the gross level provides the underlying regional\nframework for neuronal circuits. Accumulated expert knowledge on neuroanatomical organization is usually given as a series of sparse two-dimensional contours. A data\nstructure and an algorithm are described to reconstruct separating surfaces among\nmultiple regions from these sparse cross-sectional contours. A topology graph is defined\nfor each region that describes the topological skeleton of the region’s boundary surface\nand that shows between which contours the surface patches should be generated. A\ngraph-directed triangulation algorithm is provided to reconstruct surface patches\nbetween contours. This graph-directed triangulation algorithm combined together with\na piecewise parametric curve fitting technique ensures that abutting or shared surface\npatches are precisely coincident. This method overcomes limitations in i) traditional\nsurfaces-from-contours algorithms that assume binary, not multiple, regionalization of\nspace, and in ii) few existing separating surfaces algorithms that assume conversion of\ninput into a regular volumetric grid, which is not possible with sparse inter-planar\nresolution.