A review of P300-Speller signal classification methods in brain-computer interface

The Brain Computer Interface (BCI) System is part of the neural technology that transmits command from the human brain to the computer. BCI is currently one of the fastest growing area of research. BCI programs are applicable in a variety of fields, including medicine, education, self-regulation, games and entertainment, manufacturing, security, and marketing. Electronic devices can be controlled using a brain signal called electroencephalography (EEG) to record the brain's electrical activity. In EEG, the P300 wave is a positive peak of an event-related potential (ERP) that occurs 300 milliseconds recorded by the EEG. A major method of BCI research is hence the specific pattern based on P300, stimuli that are less likely to occur than other stimuli and thus the subject responds to the unexpected occurrence of these stimuli with respect to a series of stimuli. Identify the standard provided quickly. In this paper, machine learning algorithms including DCPM, LDA, SWLDA, SVM and etc, that are used for feature extraction and classification in the design OF P300-spell based BI-PC interface systems (BCI), are reviewed. And after reviewing and comparing the classification methods, we finally have the final summary of our article, which shows the results of the review research, the adaptive classification method for supervised and unsupervised, better performance than the classification it is static.