Amplitude-Phase Information Measurement on Riemannian Manifold for Motor Imagery-Based BCI

Phase synchronization phenomena are directly connected with the underlying neural mechanisms of certain cognitive processes. However, only amplitude information is utilized in most electroencephalogram (EEG)-based brain-computer interfaces (BCIs). Few existing methods can simultaneously measure and phase required for classification. In this study, a novel common amplitude-phase measurement (CAPM) method proposed. This capable jointly measuring EEG signals on Riemannian manifold. The proposed CAPM comprises two-step approach. First, graph embedding dimensionality reduction while performing spatial-spectral filtering. excellent capturing intrinsic features contained by physiological signal. Second, to enhance robustness, classifier designed incorporate regularized linear regression computation distance. Experimental results two BCI competition datasets demonstrate yield high classification performance. promising tool analyzing characteristics exhibits great potential applications.