 The proposed cross-space convolutional neural network, CSCNN, is a novel approach to decoding neurological activities using motor imagery electroencephalography, MIEG. It combines multiple views of data from different sources, including time, frequency, and space domains, to create a more accurate representation of the subject's brain activity. Additionally, it incorporates customized characteristics such as CBCSP and DMS clustering to capture the subject's specific rhythm and source distribution information. The result is a more accurate and robust classification algorithm than existing methods. In testing, the proposed method achieved an accuracy of 96.05% with real MRI information and 94.79% without MRI in the private dataset. Furthermore, it outperformed the state-of-the-art algorithms by 1.98%, demonstrating its superior performance. This article was authored by Nghu, Yanlu, Siki Zhang, and others. We are article.tv, links in the description below.