Tangent space spatial filters for interpretable and efficient Riemannian classification

Abstract

Our work promotes the theoretical understanding about Riemannian geometry based BCI classification and allows for more efficient classification as well as the removal of artifact sources from classifiers built on Riemannian methods.