In electroencephalogram (EEG) brain-computer interfaces (BCI), the performance of systems deteriorates especially when the number of channels is larger. Therefore, it is important to select the suitable channels for classification of different motor imagery tasks. In this paper, the optimal combination of channels selection method is presented. Based on the l 1 norm of common spatial pattern (CSP) features, every channel' contribution score is obtained under right hand motor imagery task and right foot motor imagery task. Then, all channels are ranked by the two kinds of scores. At last, the optimal combination of channels is selected by looking for the highest classification accuracy rate based on support vector machines (SVM) for every combination of channels using two types of channels ranking. Experimental results show that this method can not only reduce the number of channels effectively but also increase the average classification accuracy rate. This demonstrates the usefulness of our optimal combination of channels selection method in BCI systems.

No abstract