Research Article

PWC-ICA: A Method for Stationary Ordered Blind Source Separation with Application to EEG

Figure 1

Flow chart summarizing the PWC-ICA algorithm. The left column depicts the conceptual progression of the algorithm proceeding from top to bottom, whereas the right column summarizes the central mathematical operation at each step. Note that the two side-by-side bubbles connected by a dashed line at the complex ICA step illustrate the parallel interpretation between the dynamic state of the signal and the complexification of the phase space. The operators and indicate the linear transformation from the pairwise space to the dynamic phase space. Arrays and are the “base” and “velocity” components of the pairwise vectors in transformed to the dynamical phase space. The array represents the complexification of the phase space vectors, is the complex ICA demixing matrix, and is the PWC-ICA () demixing solution.