Progress in brain fiber voxel microstructure estimation algorithms

Abstract

Diffusion magnetic resonance imaging (dMRI) is the only available technique for the noninvasive and in vivo reconstruction of white matter microarchitecture. A crucial aspect of the image results is the accuracy of the estimation of fiber orientation. Recently, a number of high angular resolution imaging techniques have been used to estimate fiber orientation. In this paper, we review several white fiber reconstruction algorithms, including the widely used spherical deconvolution. We describe advances in the spherical deconvolution optimization algorithm by the development of the optimization algorithm, including L2 regularization, L1 regularization, L1, and L2 regularization. In addition, we conduct experiments to analyze typical reconstruction algorithm. Finally, future research and development of reconstruction algorithm are discussed.