 This paper proposes a novel approach for extracting frequency components from multi-channel EEG data. The proposed method combines tensor-to-vector projection, TVP, fast Fourier transform, FFT, common spatial patterns, CSP, and feature fusion to create a new feature set. The authors compared this new feature set against other existing approaches on two data sets and found that it outperformed them in terms of classification accuracy. Additionally, they found that the proposed method was able to capture both narrow band and broadband frequency components, which were not captured by previous approaches. This article was offered by UPay, Zheguoluo, Hong Youxiao, and others.