A Novel Motion-Onset N200\P300 Brain-Computer Interface Paradigm*

Abstract

The event related potential (ERP) component P300 and N200 are considered to be the most valuable electrophysiological indicators to reflect cognitive function. The traditional rare-event P300-BCI paradigm usually only takes P300 component as the target feature but ignores the N200 component. In this paper, we proposed a novel motion-onset N200\P300 brain-computer interface (BCI) paradigm, which could evoke significant N200 and P300 responses simultaneously. To evaluate the practicality of the proposed novel BCI paradigm and the robustness of the evoked N200\P300 components, three different classifiers of linear discriminant analysis (LDA), stepwise linear discriminant analysis (SWLDA) and support vector machine (SVM) with different algorithm principles were used to analyze the recognition accuracy. We also compared the motion-onset N200\P300 data with an N200-free portion to evaluate the impact of N200 component on the improvement of the BCI accuracy. Experimental results showed that, by means of this N200\P300 combination feature, the BCI accuracy significantly increased and the false positive rate significantly decreased, indicating that the proposed motion-onset N200\P300 BCI paradigm has superior performance than a traditional P300-BCI paradigm.