Method for improving brain-computer interface performance based on dynamic inverse learning network

Abstract

The invention provides a method for improving brain-computer interface performance based on a dynamic inverse learning network. The method comprises the following steps: performing preprocessing and feature extraction on an acquired P300 signal to obtain a data set; constructing a corresponding dynamic inverse learning network for the obtained data set to train, identify and classify; averaging the category probability output by the to-be-tested P300 signal by the constructed dynamic inverse learning network to obtain a classification result of the to-be-tested P300 signal; and combining the obtained identification classification result with a P300 spelling device interface to obtain a final spelling character. According to the invention, character recognition is carried out by combining aP300 spelling interface after preprocessing, feature extraction, neural network model recognition classification and integrated averaging. 100% of accuracy and 98% of accuracy are achieved on the second BCI competition data set IIb and the third BCI competition data set II respectively.

Description

technical field [0001] The invention relates to the field of EEG signal recognition control, in particular to a method for improving the performance of a brain-computer interface based on a dynamic inverse learning network. Background technique [0002] The brain-computer interface system is committed to establishing a simple and direct communication platform for patients with movement disorders. Visual stimulation is an effective way to obtain behavioral intention from brain activity. Subjects will produce P300 signals due to visual stimulation, which are easier to be detected than other EEG signals. In general, character spelling is considered a challenging study requiring high accuracy and information transfer rate. The main challenges faced by P300 signals in character spelling are low signal-to-noise ratio, high dimensionality, classifier variability, and overfitting problems leading to classification difficulties. [0003] In existing systems (S.Kundu and S.Ari,"P30...

AI Technical Summary

Problems solved by technology

The main challenges of the P300 signal in character spelling are low signal-to-noise ratio, high dimensionality, classifier variability, and overfitting problems leading to classification difficulties

[0003] In existing systems (S.Kundu and S.Ari,"P300 detection with brain-computer interface application using PCA and ensemble of weighted SVMs,"IETE Journal of Research,vol.64,no.3,pp.406-414,2018 .) The pre-processing uses the method of down-sampling to process the original signal, which can easily lead to signal distortion

Images