Christopher T. Nguyen1,2, Roland G. Henry2,3, SungWon Chung2,3
1Bioengineering, University of California Berkeley, Berkeley, CA, USA; 2Center for Molecular and Functional Imaging, Department of Radiology and Biomedical Imaging, University of California San Francisco, San Francisco, CA, USA; 3Graduate Group in Bioengineering, University of California San Francisco and Berkeley, San Francisco, CA, USA
Recently, noninvasive techniques using diffusion MRI were used to construct in-vivo brain connectivity matrices and networks. While exciting, the potential variability of these connectivity matrices is not known and will depend on several factors including data acquisition and fiber tracking algorithms. Using residual bootstrap, we propose a method to measure the errors in whole brain connectivity matrices and present results based on a deterministic fiber-tracking algorithm. The resulting matrix demonstrated high variability in many of the pairs of ROIs especially in known crossing fiber regions.