Meeting Banner
Abstract #1668

Estimation of the Uncertainty of Diffusion MRI Fiber Tracking Parameters with Residual Bootstrap

Christopher Tam Nguyen1, SungWon Chung2, Jeffrey I. Berman1, Roland G. Henry1

1Radiology and Biomedical Imaging, University of California, San Francisco, San Francisco, CA, United States; 2Radiology, Brigham and Women's Hospital, Harvard University, Boston, MA, United States


Fiber tracking (FT) based on diffusion MR has important applications for structural connectivity analyses of brain diseases and pre-operative FT of the brain. The residual bootstrap (RB) analysis on voxelwise DTI parameters is not appropriate to characterize the uncertainty in the large 3D regions defined by FT. Therefore, we will illustrate the appropriate implementation of RB to obtain the uncertainty of fiber tracking parameters (FTPs) such as number of streamlines (NOS). We validated our method with a Monte Carlo simulation showing that RB accurately estimated the SE of the NOS.