On measuring head motion and effects of head molds during fMRI.
On measuring head motion and effects of head molds during fMRI.
复制标题
在功能磁共振成像期间测量头部运动和头模的影响。
DOI：
10.1016/j.neuroimage.2020.117494
复制
发表时间：
2021
期刊：
影响因子：
5.7
通讯作者：
Power,JonathanD
中科院分区：
文献类型：
--
作者：
Lynch,CharlesJ;Voss,HenningU;Silver,BenjaminM;Power,JonathanD
头部运动会导致功能性磁共振成像(FMRI)中公认的伪影，这在静息状态研究中可能尤其有问题(Power等人，2012年)。许多组织试图开发方法，以识别和移除这些临时后的伪影，或减少扫描期间的运动(Ciric等人，2017；Greene等人，2018)。我们最近报道了一种在扫描过程中减少头部运动的方法，方法是使用适合头部线圈内的定制研磨聚苯乙烯泡沫塑料头部模具来物理约束头部(Power等人，2019b)。对13名7~28岁的受试者进行了4次扫描，每次扫描时间为4.7min，其中2次使用头部模型，2次不使用头部模型，每次扫描时间约19min。在受试者内部，头部模型减少了大的间歇性运动的发生，减少了由呼吸周期引起的小的、始终存在的运动的大小，并减少了扫描中与距离相关的fMRI信号协方差(运动伪影的特征)。随后，其他研究人员也使用了每名受试者总计约35分钟的6次fMRI测试的受试者内比较，报告称，通过将受试者的额头贴在头线圈上提供触觉反馈，可以可靠地减少使用视觉刺激和按钮反应的任务范式中的受试者运动(Krause等人，2019年)。这两种干预措施似乎都是预防头部运动的有希望的方法。我们饶有兴趣地阅读了Jolly及其同事的文章(Jolly等人，2020)，试图重新证明我们在《Power等人》中报道的头部模具减少运动的效果(Power等人，2019b)。在大多数方面，他们都没有做到这一点。乔利等人。推测(Power等人，2019b)中运动减少的发生是因为我们研究的人群主要由儿童和青少年组成，而他们研究的是年轻人，或者是因为我们的扫描与他们较长的扫描相比是“简短的”。后一种推测的逻辑并不明确，但似乎要么是我们的分析容易受到噪音和类型1错误的影响，要么是因为较长的扫描允许出现某些类型的运动，这些运动要么会导致霉菌，要么无法改善。我们对乔利等人的某些方面感到惊讶。论文和写作澄清了一些与测量头部运动有关的问题，并分享了与他们的推测有关的数据。
Head motion causes well-recognized artifacts in functional magnetic resonance imaging (fMRI) that can be especially problematic in resting state studies (Power et al., 2012). Numerous groups have attempted to develop methods either to identify and remove these artifacts post-hoc, or to reduce motion during scans (Ciric et al., 2017; Greene et al., 2018). We recently reported on a method to reduce head motion during scans by physically restraining the head using custom-milled Styrofoam head molds that fit inside head coils (Power et al., 2019b). Thirteen subjects ages 7–28 were scanned 4 times for 4.7 min at rest, twice with and twice without head molds, totaling~ 19 min per subject. Within-subject, having head molds on reduced the occurrence of large intermittent motions, reduced the size of small, always-present motions caused by respiratory cycles, and reduced distance dependent fMRI signal covariance (a signature of motion artifact) in the scans. Subsequently, other investigators, also using within-subject comparisons over 6 fMRI runs totaling~ 35 min per subject, reported that providing tactile feedback by taping subject's foreheads to the head coil reliably reduced subject motion in task paradigms using visual stimuli and button-pressing responses (Krause et al., 2019). Both interventions appear to be promising ways to prevent head motion.We read with interest the article by Jolly and colleagues (Jolly et al., 2020), an attempt to re-demonstrate the motion-reducing effects of head molds we reported in (Power et al., 2019b). In most respects, they failed to do so. Jolly et al. speculate that the motion reductions in (Power et al., 2019b) occurred because we studied a population largely composed of children and adolescents, whereas they studied young adults, or because our scans were “brief” in comparison to their longer scans. The logic of the latter speculation is not made explicit, but it seems to be either that our analyses were susceptible to noise and type 1 error, or that longer scans permit certain kinds of motion to emerge that molds either cause or would be unable to ameliorate. We were surprised at some aspects of the Jolly et al. paper and write to clarify some issues related to measuring head motion, and also to share data that bears on their speculations.