New approaches to the analysis of coordinated neuromagnetic source activity

从历史上看，神经学家研究的是大脑中哪些区域参与了感知、认知的特定功能。然而，通常情况下，当我们说话、思考或移动时，整个大脑都在工作。因此，目前的研究集中在大脑区域是如何连接的。在脑电(EEG)中将多个电极连接在头皮上，在脑磁图(MEG)中将多个磁场传感器布置在头部周围，可以无创地测量大脑活动。我们比较两个电极或脑磁图传感器上的信号，寻找具有共同节律的振荡。我们的假设是，两个传感器上的同步振荡表明潜在的神经种群相互作用。利用脑磁图，我们可以定位神经活动，并将同步性分析应用于神经磁源活动。结果表明，比传感器数据分析所能提供的更高的精度，哪些大脑区域是功能连接的。脑磁源分析并不总是完美的，皮质源不能完全分开，特别是如果它们位于很近的附近。由于这种限制，源信号可能表现出明显的同步，这与功能连接性无关。为了克服虚假同步的影响，我们将开发一系列数据分析技术，这些技术表明真正的功能协调，并对扩散源活动的影响不敏感。通过模拟研究，我们将了解MEG源数据中同步度量的特性。我们将在知觉和认知研究中分析脑磁图数据。对于所有的研究，我们都提出了一个可靠的假设，即大脑活动与功能相关。根据源信号的生理特性，设计了分析策略。具体地说，我们将使用诸如同步的时间变化、相干性的假想部分、大脑活动的频谱分量之间的同步、相位重置的统计、刺激诱发的振荡的变化以及来自非线性动力学的测量，这些测量不受来源分析中的限制的影响。这一结果将允许在基础研究和临床诊断方面进行新的和改进的应用。

Historically, neuroscientists investigated which brain area is involved in specific functions of perception, cognition. However, generally the whole brain is engaged when we speak, think, or move. Thus current research focuses on how brain areas are connected. We can measure brain activity noninvasively with multiple electrodes attached to the scalp in electroencephalography (EEG) and many magnetic field sensors arranged around the head in magnetoencephalography (MEG). We compare the signals at two electrodes or MEG sensors and search for oscillations with common rhythms. Our hypothesis is that synchronous oscillation at two sensors indicates that the underlying neural populations interact with each other. With MEG we can localize the neural activity and apply the analysis of synchrony to neuromagnetic source activity. The results indicate with much higher precision, which brain areas are functional connected than the analysis of sensor data can provide. MEG source analysis is not always perfect, and the cortical sources cannot be separated entirely, specifically if they are located in close vicinity. Because of such limitation, the source signals may show apparent synchrony, which is not related to functional connectivity. To overcome the effects of spurious synchrony, we will develop a series of data analysis techniques, which indicate true functional coordination and are insensitive to effects of spreading source activity. From simulation studies we will learn about the properties of synchrony measures in MEG source data. We will analyze MEG data in studies of perception and cognition. For all studies we develop a solid hypothesis about the functionally relevant brain activity. Based on the physiological properties of the source signals we design our analysis strategy. Specifically we will employ measures like temporal changes in synchrony, the imaginary part of coherence, synchrony between spectral components of the brain activity, statistics of phase resets, changes in stimulus-induced oscillations, and measures from nonlinear dynamics, which are not affected by limitations in the source analysis. The results will allow new and improved applications in basic research and clinical diagnostics.