Semi-Supervised Learning for Auditory Event-Related Potential-Based Brain–Computer Interface

A brain-computer interface (BCI) is a communication tool that analyzes neural activity and relays the translated commands to carry out actions. In recent years, semi-supervised learning (SSL) has attracted attention for visual event-related potential (ERP)-based BCIs motor-imagery as an effective technique can adapt variations in patterns among subjects trials. The applications of SSL techniques are expected improve performance auditory ERP-based well. However, there no conclusive evidence supporting positive effect on BCIs. If could be verified, it will helpful BCI community. this study, we assessed effects two public datasets-AMUSE PASS2D-using following machine algorithms: step-wise linear discriminant analysis, shrinkage spatial temporal least-squares support vector machine. These backbone classifiers were firstly trained by labeled data incrementally updated unlabeled every trial testing based approach. Although few datasets negatively affected, most apparently improved all cases. overall accuracy was logarithmically increased with additional data. This study supports encourages future researchers apply them