Simulation study of a P300 speller for single-lead hybrid BCI

Measurement of EEG with single-lead is desirable to reduce human stress. Some researchers use ECG to enhance a BCI system. We consider measurement of ECG and EEG with one channel. We evaluate whether separated EEG from artificial mixed signal of these signals can be used for BCI analysis. In our research, this single-lead hybrid BCI is applied to P300 speller. ECG and EEG are separated with nonlinear state-space projection. As a preprocessing of classification, single channel independent component analysis (SCICA) is used, and characters are predicted. We showed accuracy of prediction by separated EEG is 91%–100% in comparison with raw EEG. We conclude our proposed system is usable for actual single-lead hybrid BCI system.