Pantazis Dimitrios, Nichols Thomas E, Baillet Sylvain, Leahy Richard M
Signal & Image Processing Institute, University of Southern California, Los Angeles, CA 90089-2564, USA.
Inf Process Med Imaging. 2003 Jul;18:512-23. doi: 10.1007/978-3-540-45087-0_43.
We describe the use of non-parametric permutation tests to detect activation in cortically-constrained maps of current density computed from MEG data. The methods are applicable to any inverse imaging method that maps event-related MEG to a coregistered cortical surface. To determine an appropriate threshold to apply to statistics computed from these maps, it is important to control for the multiple testing problem associated with testing 10's of thousands of hypotheses (one per surface element). By randomly permuting pre- and post-stimuius data from the collection of individual epochs in an event related study, we develop thresholds that control the familywise (type 1) error rate. These thresholds are based on the distribution of the maximum intensity, which implicitly accounts for spatial and temporal correlation in the cortical maps. We demonstrate the method in application to simulated data and experimental data from a somatosensory evoked response study.
我们描述了使用非参数置换检验来检测从脑磁图（MEG）数据计算出的电流密度的皮质约束图中的激活情况。这些方法适用于任何将事件相关脑磁图映射到配准皮质表面的逆成像方法。为了确定适用于从这些图计算出的统计数据的合适阈值，控制与检验数万个假设（每个表面元素一个）相关的多重检验问题很重要。通过在事件相关研究中对来自各个时期的数据进行刺激前和刺激后数据的随机置换，我们开发了控制族系（I型）错误率的阈值。这些阈值基于最大强度的分布，该分布隐含地考虑了皮质图中的空间和时间相关性。我们在体感诱发电位研究的模拟数据和实验数据中演示了该方法的应用。
