Adaptive noise cancellation for removing cardiac and respiratory artifacts from EEG recordings

There are always many artifacts in electroencephalogram (EEG) recordings, which present serious problems for EEG interpretation and analysis. An adaptive approach is proposed to remove the cardiac and respiratory artifacts from the EEG. It makes use of two reference signals collected from the interference sources. The algorithm of recursive least squares (RLS) is used to simultaneously regulate the coefficients of the parallel filters. To evaluate the performance, the simulation and the spectrum analysis were carried out by using simulation data and real-life EEG data. The results show the approach is effective.