Approaches for the Improvement of Motor-Related Patterns Classification in EEG Signals

We apply artificial neural network (ANN) for recognition and classification of electroencephalographic (EEG) patterns associated with motor imagery for untrained subjects. Classification accuracy is optimized by reducing complexity of input experimental data. For multichannel EEG recorded by the set of 31 electrodes, we select an appropriate type of ANN which reaches 80±10% accuracy for single trial classification. We analyze the time-frequency structure of EEG signals and find that motor-related features associated with left- and right-leg motor imagery, are more pronounced in the mu (8-13 Hz) and delta (1-5 Hz) brainwaves. Based on the obtained results, we propose the optimization approach by pre-processing the EEG signals with a low-pass filter with different cutoffs. We demonstrate that the filtration of high-frequency spectral components significantly enhances the classification performance: up to 90±5% using 8 electrodes only.