Common Spatio-Temporal Patterns for the P300 Speller

A brain-computer interface (BCI) is a device that provides an alternate non-muscular communication/control channel for individuals with severe neuromuscular disabilities. The P300 event-related potential has been demonstrated to be a reliable signal for controlling a BCI. The ultimate goal is to continue io improve the classification speed and accuracy of a P300-based BCI. The method of common spatial patterns (CSP) has proven success with sensorimotor rhythm-based BCIs and, with some modifications, can also be used to accurately classify the P300. The present method, common spatio-temporal patterns (CSTP), extends CSP by incorporating time-delay embedding to extract the prominent spatio-temporal patterns corresponding to each class. The results indicate that CSTP is capable of identifying a decomposition subspace that accurately classifies the P300. In addition, this subspace can be visualized to provide useful insight regarding the discriminable spatio-temporal characteristics of the P300