Department of Electrical and Computer Engineering, University of Illinois at Urbana-Champaign, USA.
IEEE Trans Med Imaging. 2011 Nov;30(11):1933-40. doi: 10.1109/TMI.2011.2158654. Epub 2011 Jun 7.
A multishot data acquisition strategy is one way to mitigate B0 distortion and T2∗ blurring for high-resolution diffusion-weighted magnetic resonance imaging experiments. However, different object motions that take place during different shots cause phase inconsistencies in the data, leading to significant image artifacts. This work proposes a maximum likelihood estimation and k-space correction of motion-induced phase errors in 3D multishot diffusion tensor imaging. The proposed error estimation is robust, unbiased, and approaches the Cramer-Rao lower bound. For rigid body motion, the proposed correction effectively removes motion-induced phase errors regardless of the k-space trajectory used and gives comparable performance to the more computationally expensive 3D iterative nonlinear phase error correction method. The method has been extended to handle multichannel data collected using phased-array coils. Simulation and in vivo data are shown to demonstrate the performance of the method.
多连发数据采集策略是减轻高分辨率扩散加权磁共振成像实验中 B0 失真和 T2∗模糊的一种方法。然而，在不同的射束中发生的不同的物体运动导致数据中的相位不一致，从而导致显著的图像伪影。本工作提出了一种在 3D 多发扩散张量成像中进行运动诱导相位误差的最大似然估计和 k 空间校正。所提出的误差估计是鲁棒的、无偏的，并且接近克拉美-罗下限。对于刚体运动，所提出的校正方法可以有效地去除运动诱导的相位误差，而与所使用的 k 空间轨迹无关，并具有与更计算密集的 3D 迭代非线性相位误差校正方法相当的性能。该方法已扩展到处理使用相控阵线圈采集的多通道数据。模拟和体内数据表明了该方法的性能。
