Meeting Banner
Abstract #1438

Probabilistic Fiber Tracking Using the Residual Bootstrap with Constrained Spherical Deconvolution MRI

Ben Jeurissen1, Alexander Leemans2, Jacques-Donald Tournier3,4, Jan Sijbers1

1Visionlab, Dept. of Physics, University of Antwerp, Antwerp, Belgium; 2CUBRIC, School of Psychology, Cardiff University, Cardiff, Wales, UK; 3Brain Research Institute, Florey Neuroscience Institutes (Austin), Melbourne, Victoria, Australia; 4Dept. of Medicine, University of Melbourne, Melbourne, Victoria, Australia

We present a new probabilistic tractography algorithm based on constrained spherical deconvolution (CSD) of diffusion weighted (DW) MRI and the residual bootstrap. Using simulations we show that the residual bootstrap is an accurate method to characterize CSD fiber trajectory uncertainty. As opposed to classic bootstrapping, the residual bootstrap only requires a single DW acquisition, making it clinically feasible. We compare our algorithm to state-of-the-art probabilistic diffusion tensor imaging (DTI) tractography, using both simulated and real DW data. Our algorithm is shown to be much less prone to fiber dispersion than probabilistic DTI tractography in regions of multiple fiber orientations.