Detection of P300 in a BCI speller

Abstract

This paper has presented the algorithm for P300 Speller Paradigm. The dataset used has been dataset II of the BCI competition III 2005. The novelty in the approach is that for each epoch the dataset has been split, organized and then preprocessed by average function of number of raw / column that was intensified. For each character classification is based on the Euclidean distance between a pattern P300 calculated on the training set and patterns of six rows and six columns calculated on the testing set. © 2011 Springer-Verlag.