Validation of MRI-Based Fiber-Tracking Results

Abstract

A new approach to estimating the probability of the fiber-track existence has been developed and validated on the basis of digital phantoms and diffusion-weighted magnetic resonance imaging measurements on human brain. The proposed approach includes two parameters—the local probability of the tract direction and the Shannon entropy. The local probability of the tract direction characterizes how well tracts “fit” the direction information in a voxel. The Shannon entropy characterizes the shape of diffusivity in the voxel. Calculations on digital phantoms and in vivo data showed that the combination of these parameters makes it possible to verify fiber-tracking results.