 This paper proposes a novel intersubject transfer learning approach for improving spec-based BCI performance. It involves training spatial filters using multiple covariance maximization to extract SSV-paralleted information from the training trials. Then, the transferred spatial filters are used to construct two new transferred templates, which are then compared to the original template to calculate the contribution scores of each source subject. A four-dimensional feature vector is finally constructed for SSV-P detection. Experimental results show that the proposed method outperforms other methods in terms of accuracy and robustness. This article was authored by Yu Zhong, Shen Kuan Xie, Zhao Yongshu, and others.