The role of transient spectral 'bursts' in functional connectivity: A magnetoencephalography study
The role of transient spectral 'bursts' in functional connectivity: A magnetoencephalography study
复制标题
DOI：
10.1016/j.neuroimage.2020.116537
复制
发表时间：
2020-04-01
期刊：
影响因子：
5.7
通讯作者：
Brookes, Matthew J.
中科院分区：
文献类型：
--
作者：
Seedat, Zelekha A.;Quinn, Andrew J.;Brookes, Matthew J.
神经振荡主导着宏观脑活动的电生理测量，这些节律的波动为了解皮质兴奋、抑制和连接提供了一个有洞察力的窗口。然而，近年来，平稳变化的振荡的“经典”图景受到了挑战，因为有一种观点认为，许多“振荡”实际上可能是由活动中点状高幅度突发的重复形成的，其频谱组成与传统定义的频率范围(例如，α/β频段)相交。这一发现为可测量的大脑活动提供了一种新的解释，然而，无论是检测脉冲的方法手段，还是它们与其他发现(例如连接)的联系都尚未确定。在这里，我们使用一种新的方法来检测脑磁图(MEG)数据中的脉冲。我们证明了时延嵌入式隐马尔可夫模型(HMM)可以用来描述单区域突发，这与现有技术是一致的。然而，与现有技术不同的是，隐马尔可夫模型在时间序列数据中寻找特定的频谱模式。我们描述了静息状态脑磁图数据中突发持续时间、出现频率和幅度在大脑皮层的分布。在一项运动任务中，我们展示了运动相关的β减少和运动后的β反弹是如何由突发发生的变化驱动的。最后，我们证明了β带功能连接体可以用一种简单的爆发重叠的测量来得到，并且在不同区域的重合爆发对应于一个高度相干的周期。综上所述，本文提供了一种新的突发识别和连通性分析方法，这将对未来神经振荡的研究具有重要意义。
Neural oscillations dominate electrophysiological measures of macroscopic brain activity and fluctuations in these rhythms offer an insightful window on cortical excitation, inhibition, and connectivity. However, in recent years the 'classical' picture of smoothly varying oscillations has been challenged by the idea that many 'oscillations' may actually be formed from the recurrence of punctate high-amplitude bursts in activity, whose spectral composition intersects the traditionally defined frequency ranges (e.g. alpha/beta band). This finding offers a new interpretation of measurable brain activity, however neither the methodological means to detect bursts, nor their link to other findings (e.g. connectivity) have been settled. Here, we use a new approach to detect bursts in magnetoencephalography (MEG) data. We show that a time-delay embedded Hidden Markov Model (HMM) can be used to delineate single-region bursts which are in agreement with existing techniques. However, unlike existing techniques, the HMM looks for specific spectral patterns in timecourse data. We characterise the distribution of burst duration, frequency of occurrence and amplitude across the cortex in resting state MEG data. During a motor task we show how the movement related beta decrease and post movement beta rebound are driven by changes in burst occurrence. Finally, we show that the beta band functional connectome can be derived using a simple measure of burst overlap, and that coincident bursts in separate regions correspond to a period of heightened coherence. In summary, this paper offers a new methodology for burst identification and connectivity analysis which will be important for future investigations of neural oscillations.