Adaptation method for BCI system on subject

Most of high accuracy BCI systems based upon the evoked potential P300 requires a training session for each subject before it can be used. This paper proposes a method of the adaptation process to improve the accuracy and avoid the training session. The process starts with a universally trained (not trained for a specific subject) classifier. The goal is to develop a classifier optimized for the subject who is using the system. The adaptation strategy includes personalization of the classifier. Results showed that for the electrodes set with 8 electrodes we can obtain a classifier with high accuracy and the process is able to perform the adaptation while the system is being used.