Rate acceleration for estimators of integral curves from diffusion tensor imaging (DTI) data

The components of a tensor field MM are coefficients in a certain system of regression equations. To estimate the integral curve starting at a given point and driven by the principle eigenvector field corresponding to MM, recently an estimation procedure has been constructed in Carmichael and Sakhanenko (2014). We propose a acceleration to this method which dampens the asymptotic mean integrated squared error and boosts the convergence rates. Supporting theoretical results and simulations are presented.