Cross-Subject Transfer Learning on High-Speed Steady-State Visual Evoked Potential-Based Brain-Computer Interface

Steady state visual evoked potential (SSVEP)-based brain computer interfaces have shown its robustness in achieving high information transfer rate. State-of-the-art training-based SSVEP decoding methods such as extended Canonical Correlation Analysis (CCA) and Task-Related Component Analysis (TRCA) are the major players that elevate the efficiency of SSVEP-based BCI through individualized calibration process. However, the calibration may be time-consuming in order to collect sufficient training templates, deteriorating the practicality in real-world context. This study aims to develop a cross-subject transferring approach to reduce the need of training data from a new user. The proposed least-squares transformation (LST) method was able to significantly reduce the training templates required for TRCA-based SSVEP detection on a 40-class SSVEP dataset from 8 subjects. This study shed light on plug-and-play high speed SSVEP-based BCI for further practical applications.