AOAR: An automatic ocular artifacts removal approach for multi-channel EEG data based on NMF and EMD

Electroencephalogram (EEG) signals are inevitably interfered by artifacts during the acquisition process. These artifacts make analysis and interpretation of EEG data difficult. A major source of artifacts in EEG is the ocular activity. Therefore, it is important to remove the ocular artifacts before further processing the EEG data.In this study, we proposed an automatic ocular artifacts removal (AOAR) method for EEG signals based on non-negative matrix factorization (NMF) and empirical mode decomposition (EMD). First, the amplitude of EEG data was normalized in order to ensure the non-negativity of EEG data. Then, the normalized EEG data were decomposed into a set of components using NMF. The components containing ocular artifacts were extracted automatically through fractal dimension. Subsequently, the temporal activities of these components were adaptively decomposed into some intrinsic mode functions (IMFs) by EMD. The IMFs corresponding to ocular artifacts were removed. Finally, the denoised EEG data were reconstructed.The proposed method was tested against the other seven methods. In order to assess the effectiveness and reliability of the AOAR method in processing EEG data, experiments on ocular artifacts removal were performed using semi-simulated EEG data. Experimental results indicated that the proposed method was superior to other methods in terms of root mean squared error, signal noise rate and correlation coefficient, especially in the cases with lower signal noise rate. To further evaluate the application potentials of the proposed method in real life, the proposed method with the counterparts were applied to preprocess the real EEG data recorded from children with and without attention-deficit/hyperactivity disorder (ADHD). After artifacts rejection, the ERP feature was extracted for classification. The AOAR performed the best for distinguishing the ADHD children from others.These results indicate that the proposed AOAR method has great prospects in removing ocular artifacts from EEG.