MRP electroencephalogram signal feature extraction method based on correlation coefficient and MRP electroencephalogram signal feature extraction device thereof

Abstract

The invention discloses a MRP (movement-related potential) electroencephalogram signal feature extraction method based on the correlation coefficient and a MRP electroencephalogram signal feature extraction device thereof. The method comprises the steps that zero mean processing is performed on each MRP electroencephalogram signal sample and tag in a training set; an objective function based on the correlation coefficient is maximized according to the electroencephalogram signal samples and tags after zero mean processing so as to obtain a set of spatial filters; the output vectors after spatial filtering of the MRP electroencephalogram signals after zero mean processing are calculated according to each spatial filter, and the sum of all the elements of the output vectors is taken to act as one feature value of electroencephalogram signal feature vectors; and the feature vector of each electroencephalogram signal is extracted according to the spatial filter set learnt in the training phase when the acquired electroencephalogram signals are received in the brain-machine interface working process. The method is high in robustness and can obtain the relatively satisfactory result when the training set samples are insufficient.

Description

technical field [0001] The invention relates to the technical field of neural engineering, in particular to a correlation coefficient-based MRP EEG feature extraction method and a device thereof. The invention also relates to a brain-computer interface system. Background technique [0002] As an emerging discipline, neuroengineering focuses on studying the functions of the central and peripheral nervous systems and manipulating their behaviors with engineering techniques. Brain-Computer Interface (BCI) technology is one of the most popular research directions in neural engineering. Brain-computer interface technology is of great significance in the fields of recovery of human motor function, control of external equipment in special environments, and driverless cars for the disabled. [0003] Electroencephalogram (electroencephalogram, EEG), also known as brain wave, is the cortical spontaneous potential pattern recorded by means of metal electrodes and conductive glue in e...

AI Technical Summary

Problems solved by technology

[0006] However, in BCI research, considering the mental load of the subjects, the number of training samples is often small, and the number of training samples is even smaller after classification, making it difficult to accurately estimate its center, thus affecting the optimization of the spatial filter , resulting in low accuracy of feature extraction

Images