Wavelet transform for the identification of P300

Abstract

The reliability of a newly developed algorithm for the identification of the P300 component of event-related potentials based on a continuous wavelet transform was investigated. The electroencephalogram records of one participant made by using a three-stimulus paradigm (a kind of the odd-ball paradigm) were analyzed. The accuracy of identification of certain wavelet types for the detection of P300 was from 76.32 to 86.84%. Thus, relatively simple algorithms for processing and classifying the electroencephalogram record signal show acceptable results in terms of the accuracy of identification of the P300 component of event-related potentials based on randomly selected data.