Reduction of alpha distortion in event related potentials

A frequently used brain monitoring technique in cognitive neuroscience research are Event Related Potentials (ERPs). ERPs are electrical brain responses to specific sensory, cognitive and motor events in electroencephalographic (EEG) data. However, ERPs are known to have a very low signal-to- noise-ratio and suffer from artifacts present in EEG data like ocular and muscular artifacts. Another common type of distortion is alpha activity. The aim of this study is to remove this activity with different mathematical approaches. Classical methods like filtering and independent component analysis (ICA) do not yield good results in the reduction of alpha distortion. Therefore, we propose Parallel Factor Analysis or PARAFAC as a method for reducing the alpha activity. We will show why filtering and ICA fail to solve this problem and how PARAFAC does succeed in improving the quality of the measured ERPs.