Fiber Stipples for Crossing Tracts in Probabilistic Tractography

Abstract

Given diffusion weighted magnetic resonance (dMRI) data, tractography methods may reconstruct estimations of neural connections of the human brain, so called tractograms. Probabilistic tractography algorithms generate a scalar value for each point of the brain, which describes the confidence of an existing structural connection to a predefined seed region. Recently presented Fiber-Stippling is a promising tool to effectively visualize such scalar values on axis aligned cutting planes. However, Fiber-Stippling only works with principal diffusion directions and cannot handle complex tract configurations, such as overlapping or crossing tracts, which are very important to neuroscience. In this work we present an illustrative technique for probabilistic tracts in such configurations, which is based on Fiber-Stippling. Our technique supports multiple diffusion directions as given by high angular resolution diffusion images (HARDI) and hence can visualize crossing tracts, while preserving all of the advantages of Fiber-Stippling. We solve this by visually supporting the stipples, while not altering the original visualization metaphor. Our work is an important contribution to adequate visualization of neuroanatomy, as crossing tracts are a frequent phenomen inside of the human brain. Moreover, our technique may be customized to crossing line fields in general.