P300 spatial filtering and coherence-based channel selection

Spatial filtering is an important technique used in electroencephalography to enhance signal-to-noise ratio and to reduce the data dimensionality. In the context of Brain-Computer Interfaces, the Common Spatial Patterns method is widely used for classification of motor imagery events, however it is not very often used for classification of event related potentials such as P300. In this paper we show that Common Spatial Patterns is an effective approach to improve P300 classification rates. It is proposed a Bayesian methodology for feature combination that overcomes the limitations of the feature method used in motor imagery. Also, a method for channel selection based on interchannel coherence is proposed, reducing the number of channels and improving the classification results.