Motor imagery EEG Recognition based on Biomimetic Pattern Recognition

This paper improves Biomimetic Pattern Recognition based on Hyper Sausage Neuron and applies it in the study of Motor Imagery EEG recognition. The paper uses the datasets from previous Brain-Computer Interface Competitions to test the accuracy and efficiency of the results, and compares them with those of SVM and BP. The results show that: with sufficient training set, the performance of Biomimetic Pattern Recognition is close to those of SVM and BP; but when the size of training set is small, SVM and BP´s classification performance degrades obviously, while Biomimetic Pattern Recognition demonstrates clear advantages.