Abstract
Brain Computer Interface (BCIs) enables the direct communication between human and computer by analyzing the feature of brain waves. We use P300, which is one of the event-related potential (ERP), as the feature in BCI. P300 speller is a communication tool which allows a user to input letters only by thoughts. Pattern recognition methods such as StepWise Linear Discriminant Analysis (SWLDA) are used for the discrimination of P300. However, SWLDA needs large calculation time, because it uses not only the feature of P300 but also other information. Therefore, it is expected that the calculation time becomes shorter by extracting the feature of P300 and using it for discrimination. In this study, we employ Independent Component Analysis (ICA) to extract the feature of P300. We also consider the feature among some electrodes in this paper.
