Topographic phase maps using iterative independent component analysis

In this study, we employed our recently developed iterative independent component analysis (iICA) procedure to measure single-trial EPs from auditory N100 recordings of 21 normal subjects obtained with a whole-head, 256-channel dense-array EEG (eJEEG) scanner. To study the general applicability of the method, it was also applied to five magnetoencephalographic (MEG) datasets. For each channel, single trial responses were extracted using the iICA approach followed by separating trials into two groups one with trials having the same and the other having the opposite phase as the average response. Topographic phase maps were constructed by calculating, for each channel, the percentage of trials that go in-phase (%IP) with the average. The phase maps were found to follow a trend that was similar to amplitude distribution. With EEG, the channels with high %IP were clustered mainly along the midline while the ones on the temporal region had lower %IP. With MEG, the nature of phase distribution was opposite to EEG with temporal regions having higher %IP and central ones having low %IP values. The MEG phase distribution also showed higher concentration on the right temporal region relative to the left, consistent with the idea that distribution of N 100m sources cover a larger area in the primary auditory cortex on the right.