Classification of steady state visual evoked potentials by Multi-Class T-Weight Method

In this paper, Multi-Class T-Weight Method (MCTW) is presented for classification in brain-computer interface (BCI) systems. Proposed method is an extension of the existing Improved T-Weight method for multi-class problems. The method was tested on the frequency and correlation based features obtained from electroencephalogram data of 20 Subjects in a steady state visual evoked potential (SSVEP) based offline BCI classification task. Obtained classification performances with different classifiers show that the MCTW method compete with the other well-known classifiers like linear discriminant analysis (LDA) and support vector machines (SVMs). Therefore, it can be used in classifying SSVEP based electroencephalogram data with proper features.