Sequence detection analysis based on canonical correlation for steady-state visual evoked potential brain computer interfaces.

摘要: Abstract Background: Steady-state visual evoked potential (SSVEP) has been widely applied to develop brain computer interface (BCI) systems. The essence of SSVEP recognition is recognize the frequency component target stimulus focused by a subject significantly present in EEG spectrum. New method: In this paper, novel statistical approach based on sequence detection (SD) proposed for improving performance recognition. This method uses canonical correlation analysis (CCA) coefficients observe signal sequence. And then, threshold strategy utilized Results: result showed classification with longer duration time window achieved higher accuracy most subjects. average costing per trial was lower than predefined time. It implicated that our could improve speed BCI system contrast other methods. Comparison existing method(s): comparison resultful algorithms, experimental SD better those using used CCA-based and two newly least absolute shrinkage selection operator (LASSO) model as well multivariate synchronization index (MSI) method. Furthermore, information transfer rate (ITR) obtained three methods participants. Conclusions: These conclusions demonstrated promising high-speed online BCI.