Feature Extraction of Brain–Computer Interface Electroencephalogram Based on Motor Imagery

Brain-computer interface (BCI) is a system that allows people to communicate or control external devices simply by using information from the brain without relying on the peripheral nervous system and muscles; BCI technology has great potential application value in motor function assistance and motor function rehabilitation and has become a new research hotspot in the fields of machine learning, biomedical engineering and computer communication. The feature extraction of motor imagery electroencephalogram (EEG) is to find the most effective characteristics of complex EEG signal that can represent the consciousness task, to differentiate the feature vectors extracted from different consciousness tasks, and to maximize the correlation between the feature vector and the consciousness task. On the basis of summarizing and analyzing previous research works, this paper proposes a new EEG feature extraction algorithm based on common spatial pattern (CSP) and adaptive auto-regressive (AAR), and demonstrates feasibility of band energy, sample entropy and order accumulation to be the characteristics of motor imagery classification, and finally compares the classification effects of linear discrimination classifier, common space classifier and Bayesian classifier. The simulation results show that the proposed method and algorithm can effectively extract the features of EEG signals during motor imagery. The research results of this paper provide a reference for the further study of feature extraction of brain-computer interface EEG.