| pubmed-article:17271826 | pubmed:abstractText | A robust constrained blind source separation algorithm (CBSS) has been developed here for an effective removal of eye muscle artifacts from electroencephalograms (EEG). Presently, clinicians reject a data segment if the patient blinked or spoke. The rejected data segment may contain important information that may be masked by the artifact. In the CBSS technique we exploit a reference signals as a constraint. The constrained problem is then converted to an unconstrained problem by means of nonlinear penalty functions weighted by the penalty terms. This leads to the modification of the overall cost function, based on the natural gradient algorithm (NGA), by incorporating a reference signal. | lld:pubmed |
