Brain electrical classification method for entropy value based on dynamic function connection

Abstract

The invention discloses a brain electrical classification method for an entropy value based on a dynamic function connection. The method comprises the steps of firstly preprocessing an acquired original brain electrical signal, and then filtering the preprocessed brain electrical signal; calculating a phase relation, of the brain electrical signal of each frequency band, between every two channelsat each time point by use of a phase synchronization analysis method to obtain a dynamic function connection matrix; then, calculating a time domain entropy of a phase relation value between two channels one by one, and obtaining an information entropy of each edge to measure the complexity of the time domain of each edge of a brain electrical functional network; respectively taking a dynamic functional connection entropy of each frequency band as a classification feature of the brain electrical functional network, training an adaptive boosting classifier to obtain multiple adaptive boostingclassifiers and corresponding classification correct rates; and performing combined classification on a sample in a voting manner. With the brain electrical classification method for the entropy valuebased on the dynamic function connection, the problem of low classification accuracy rate in an existing brain electrical signal classification method is solved.

Description

technical field [0001] The invention belongs to the technical field of EEG signals, in particular to an EEG classification method based on the entropy value of dynamic functional connections. Background technique [0002] As a combination of electroencephalogram (EEG) technology and complex network theory, EEG signal data classification methods have become one of the hotspots in the field of brain science. However, due to the limitation of the principle of the traditional EEG signal data classification method, the classification accuracy is low, which seriously affects its application value. [0003] Traditional EEG signal data classification methods mainly include: time-domain-based analysis methods, frequency-domain-based analysis methods, and time-frequency analysis methods. Classification features are obtained after spectrum analysis. These two methods have strict requirements for EEG signal preprocessing, and EEG is a non-stationary signal. Using these two methods will...

AI Technical Summary

Problems solved by technology

[0005] The object of the present invention is to provide a kind of EEG classification method based on the entropy value of dynamic functional connection, which solves the problem of low classification accuracy in the existing EEG signal classification method

Images