The Elliptical Cone of Uncertainty in Diffusion Tensor Imaging and Its Normalized Measures

INTRODUCTION The diffusion tensor is always measured in the presence of background noise. Recent experimental observations by Jeong et al.[1] and Lazar et al.[2,3] have shown that the cones of uncertainty (COU) [4,5,6] of the major eigenvectors of the diffusion tensors in the brain are mostly elliptical. Analytical approaches have also been developed to characterize the elliptical cone of uncertainty [7] through nonlinear least squares estimations [8]. Here, we present a new technique to construct the COU based on the inverse of the Gnomonic projection and two normalized geometric measures associated with the COU—the normalized areal and circumferential measures.