Finally, we used independent component analysis (ICA) to further correct for ocular artifacts, such as microsaccades. ICA provides a more stringent method for correcting artifact, and it appears to effectively eliminate ocular artifacts that may have been overlooked in the aforementioned conventional threshold-based artifact rejection procedure (Keren et al. 2010; McMenamin et al. 2010; Shackman et al. 2010; Delorme et al. 2012). The ocular correction ICA was performed using the value trigger algorithm in semiautomatic mode to detect components associated with eye movements (Makeig et al. 1997). Decomposition of components was performed with the biased, extended Infomax method on whole data with all the channels included. All subsequent computations were performed on the reference-free CSD estimates that were corrected for artifacts.
