An Evaluation of the Robustness of Tract-Based Measures of White Matter Integrity

Mingyi Li¹, Jian Lin, Ken Sakaie, Erik Beall, Lael Stone, Robert Bermel, Micheal D. Phillips, Mark J. Lowe

Track based measures of white matter integrity are increasingly being employed in studies of populations with possible compromised white matter integrity. Identifying fiber tracks in populations with diseased white matter has obvious possible confounds. Various methods have been proposed to deal with this, including 1) the use of healthy control-based tracking atlases as well as 2) non-tensor based probabilistic tracking methods that are less susceptible to problems from regions of low anisotropy. To date, there has been no validation of either method with regard to establishing either 1) robust pathway measures in the case of (1), or 2) unbiased pathway identification in the case of (2). We present here a study based on non-rigid image registration to validate the robustness of establishing comparable pathways in a healthy control population and a population of Multiple Sclerosis (MS) patients. Our method suggests that 1) non-rigid image registration of tracks generated in healthy controls to patient diffusion data and 2) tracking in both patients and controls with a non-tensor based method can result in robust track-based measures of white matter integrity.