Preprocessing of EEG for imagery movement based on integration of multi-domain

In previous EEG preprocessing algorithms of brain-computer interface, there are problems such as huge amounts of processing EEG data and the ignorance of EEG´s variance from person to person. This paper has proposed a novel method of preprocessing algorithm based on integration of multi-domain, using Fisher distance to select the sampling electrodes and spatial preprocessing is added to the traditional algorithm which is just based on time-frequency domain. It is proved to be effective and practical in overcoming the above drawbacks with experiments of EEG collection. The experiment results show that the proposed method can reduce more than 96.9% of the total processing EEG data and decrease 95.8% of the BCI system´s total running time while remain almost equal classification accuracy, contributes to the online application.