A robust method for suppressing motion-induced coil sensitivity variations during prospective correction of head motion in fMRI
A robust method for suppressing motion-induced coil sensitivity variations during prospective correction of head motion in fMRI
复制标题
DOI：
10.1016/j.mri.2016.06.005
复制
发表时间：
2016-10-01
影响因子：
2.5
通讯作者：
Graham, Simon J.
中科院分区：
文献类型：
--
作者：
Faraji-Dana, Zahra;Tam, Fred;Graham, Simon J.
前瞻性运动校正是抑制fMRI期间头部运动影响的一个有前途的候选解决方案，理想情况下允许成像平面相对于运动的头部保持固定。然而，残留的信号伪影可能仍然存在，因为相对于固定的多通道接收器线圈（具有不均匀的灵敏度图）的头部运动可能会引入与弱fMRI BOLD信号相当的不必要的信号变化（类似于1.5-3.0 T时的1%-4%）。目前的工作旨在研究这些残余伪影的大小，并描述通过调整灵敏度图来反映头部和线圈之间相对位置变化的前瞻性运动校正的机制。数值模拟用于人体功能磁共振成像实验。仿真结果表明，在一个常用的12通道头部线圈内进行轴向成像时，头部在平面内旋转5度会产生接近3%的伪影信号变化。随后，6名年轻人在有和没有明显头部运动的情况下进行了成像，并使用前瞻性获取校正（PACE）方法进行了和没有前瞻性运动校正，并进行了和没有敏感性图调整。通过比较不同实验条件下的三个数据质量指标（激活体素数、激活重叠的Dice系数、基线fMRI时间序列数据的时间标准偏差），可以看出，灵敏度图调整与PACE相结合强有力地保护了感兴趣的伪影。结论是，与固定多通道线圈相关的头部运动可能会对前瞻性运动校正的fMRI产生不利影响，并且灵敏度图调整可以在3.0 T. (C) 2016 Elsevier Inc.减轻这种影响。版权所有。
Prospective motion correction is a promising candidate solution to suppress the effects of head motion during fMRI, ideally allowing the imaging plane to remain fixed with respect to the moving head. Residual signal artifacts may remain, however, because head motion in relation to a fixed multi-channel receiver coil (with non-uniform sensitivity maps) can potentially introduce unwanted signal variations comparable to the weak fMRI BOLD signal (similar to 1%-4% at 1.5-3.0 T). The present work aimed to investigate the magnitude of these residual artifacts, and characterize the regime over which prospective motion correction benefits from adjusting sensitivity maps to reflect relative positional change between the head and the coil. Numerical simulations were used to inform human fMRI experiments. The simulations indicated that for axial imaging within a commonly used 12-channel head coil, 5 degrees of head rotation in-plane produced artifact signal changes of similar to 3%. Subsequently, six young adults were imaged with and without overt head motions of approximately this extent, with and without prospective motion correction using the Prospective Acquisition CorrEction (PACE) method, and with and without sensitivity map adjustments. Sensitivity map adjustments combined with PACE strongly protected against the artifacts of interest, as indicated by comparing three metrics of data quality (number of activated voxels, Dice coefficient of activation overlap, temporal standard deviation of baseline fMRI timeseries data) across the different experimental conditions. It is concluded that head motion in relation to a fixed multi-channel coil can adversely affect fMRI with prospective motion correction, and that sensitivity map adjustment can mitigate this effect at 3.0 T. (C) 2016 Elsevier Inc. All rights reserved.