The Time-Varying Network Patterns in Motor Imagery Revealed by Adaptive Directed Transfer Function Analysis for fMRI
The Time-Varying Network Patterns in Motor Imagery Revealed by Adaptive Directed Transfer Function Analysis for fMRI
复制标题
DOI：
10.1109/access.2018.2875492
复制
发表时间：
2018
期刊：
影响因子：
3.9
通讯作者：
Tao Zhang;Mengchen Li;Li Zhang;B. Biswal;D. Yao;Peng Xu
中科院分区：
文献类型：
--
作者：
Tao Zhang;Mengchen Li;Li Zhang;B. Biswal;D. Yao;Peng Xu
运动想象（MI）是一种多维的高级认知能力，涉及多个大脑区域的协调贡献[即，补充运动区（SMA），运动前皮层（M1）和后顶叶皮层]及其耦合。然而，MI期间这些激活区域之间的动态相互作用仍不清楚。在这里，我们应用自适应定向传递函数（ADTF）来跟踪MI期间激活区域之间随时间变化的连接模式。我们发现，连接模式是不同的，在两个MI任务，并随时间动态变化，代表特殊的状态依赖和时间依赖的时变连接模式。我们的研究结果表明，左前额叶（aIns），对侧SMA，对侧M1，作为重要的因果关系的目标，在网络的重组在不同的阶段发挥了关键作用，这意味着在MI期间存在着一个分层的网络重组。此外，我们发现左侧和右侧MI的偏侧化发生在MI中期，并通过对侧SMA调制的有效连接来反映。此外，采用图形分析，以进一步表征这些激活区域之间的相互作用的时间演变。我们还发现，网络效率与网络的连接模式在MI协调。总的来说，这些研究结果的基础上ADTF分析扩大了我们的理解，随时间变化的网络组织在MI。
Motor imagery (MI) is a multi-dimensional high-level cognitive ability that involves coordinated contributions from multiple brain regions [i.e., supplemental motor area (SMA), premotor cortex (M1), and posterior parietal cortex] and their couplings as well. However, the dynamic interactions among these activated regions during MI are still unclear. Here, we applied the adaptive directed transfer function (ADTF) to track time-varying connectivity patterns among activated regions during MI. We found that the connectivity patterns are different in two MI tasks and dynamically changes over time, representing special state-dependent and timing-dependent time-varying connectivity patterns. Our findings indicate that left anterior insula (aIns), contralateral SMA, and contralateral M1, which served as important causal targets, play crucial roles in reorganization of the network at different stages, implying that there exists a hierarchical network reorganization during MI. In addition, we found that the lateralization of the left- and right-hand MI occurred in the MI middle stages and was reflected by the effective connectivities modulated by contralateral SMA. Moreover, a graph analysis was adopted to further characterize the temporal evolution of these interactions among activated regions. We also found that network efficiencies are coordinated with the connectivity pattern of networks during MI. Collectively, these findings based on the ADTF analysis expand our understanding of the time-varying network organization in MI.