A Parametric Approach to Evaluating the Statistical Significance of Pathway Dependent Diffusion Measures

Mark J. Lowe¹, Ken E. Sakaie¹, Jian Lin¹, Lael Stone², Micheal D. Phillips¹

We have previously introduced pathway dependent diffusion measures as a way of determining functionally relevant disease burden in multiple sclerosis. This method determines the mean diffusion properties (i.e. FA, radial diffusivity, longitudinal diffusivity) along a pathway determined with probabilistic tractography. There is a significant drawback to the method in that the sensitivity can be quite low due to the fact that tissue diffusion parameters will vary substantially along some pathways. In this study, we utilize a parametric approach to determining local pathway diffusion properties and derive a statistical approach to determine if a given fiber path contains significant disease burden.