Imaging Research, Sunnybrook Health Sciences Centre, Toronto, Ontario M4N 3M5, Canada and Department of Medical Biophysics, University of Toronto, Ontario M4N 3M5, Canada.
Med Phys. 2011 Aug;38(8):4634-46. doi: 10.1118/1.3583814.
Functional magnetic resonance imaging (fMRI) is limited by sensitivity to millimetre-scale head motion. Adaptive correction is a strategy to adjust the imaging plane in response to measured head motion, thereby suppressing motion artifacts. This strategy should correct for motion in all six degrees of freedom and also holds promise for through-plane motion that creates "spin-history" artifact that cannot easily be removed by postprocessing methods. Improved quantitative understanding of the MRI signal behavior associated with spin-history artifact would be useful for implementing adaptive correction robustly.
A numerical simulation was developed to predict MRI artifact signal amplitude in a single-slice for simple motions, implemented with and without adaptive correction, and compared with experiment by imaging a phantom at 3.0 T. Functional MRI was also performed of a human volunteer to illustrate adaptive correction in the presence of spin-history artifact.
Good agreement was achieved between simulation and experimental results. Although time-averaged artifact signal amplitude was observed to correlate linearly with motion speed, artifact time-courses were nonlinearly related to motion waveforms. In addition, experimental results demonstrated effective adaptive correction of spin-history artifact when the phantom underwent complex motions. Adaptive correction during human fMRI suppressed spin-history artifacts and spurious activations associated with task-correlated motion.
Overall, this work suggests that adaptive correction, especially when implemented with minimal lag between motion measurement and scan plane update, may help to expand the populations for which fMRI can be performed robustly.
功能磁共振成像（fMRI）受到毫米级头部运动的灵敏度限制。自适应校正策略是一种根据测量的头部运动调整成像平面的策略，从而抑制运动伪影。该策略应校正所有六个自由度的运动，并且对产生“自旋历史”伪影的平面内运动也有一定的纠正作用，这种伪影不易通过后处理方法去除。改进对与自旋历史伪影相关的 MRI 信号行为的定量理解，对于稳健地实施自适应校正将非常有用。
开发了一个数值模拟，用于预测在简单运动下单个切片的 MRI 伪影信号幅度，模拟中实施了和未实施自适应校正，并与在 3.0 T 下对幻影进行成像的实验进行了比较。还对人类志愿者进行了功能磁共振成像，以说明在存在自旋历史伪影的情况下自适应校正。
模拟和实验结果之间取得了很好的一致性。虽然观察到平均伪影信号幅度与运动速度呈线性相关，但伪影时程与运动波形呈非线性相关。此外，实验结果表明，当幻影经历复杂运动时，自适应校正可以有效地校正自旋历史伪影。在人类 fMRI 期间，自适应校正抑制了与任务相关运动相关的自旋历史伪影和虚假激活。
总的来说，这项工作表明，自适应校正，特别是在运动测量和扫描平面更新之间实现最小滞后时，可能有助于扩大可以稳健地进行 fMRI 的人群。
