Motor imagery classification method based on convolutional neural network

Abstract

The invention discloses a convolutional neural network method based on a parallel multi-scale time convolution kernel, and mainly solves the problems of low detection accuracy and difficulty in effectively detecting imaginary movement of a user in the prior art. According to the implementation scheme, the method comprises the following steps: collecting imaginary motion electroencephalogram data,preprocessing the imaginary motion electroencephalogram data, and making a data set by using the preprocessed electroencephalogram data; constructing a convolutional neural network, training the convolutional neural network by using the training set and the verification set, testing the convolutional neural network by using the test set, and performing fine tuning on the tested convolutional neural network by using the electroencephalogram data of the testee to obtain a final convolutional neural network suitable for the testee to perform an online experiment; and obtaining an online imagination motion electroencephalogram signal of the testee in real time, and sending the online imagination motion electroencephalogram signal to the final convolutional neural network to obtain a real-timeclassification result. The method can effectively detect the imaginary movement of the user, improves the classification accuracy of the imaginary movement electroencephalogram signals, can be used for medical service, and serves as an auxiliary tool to participate in rehabilitation treatment of stroke patients.

Description

technical field [0001] The invention belongs to the field of information technology, in particular to a classification method of electroencephalogram signals, which can be used in medical services. Background technique [0002] EEG signals are generated by the bioelectrical activity of neurons in the brain, which belongs to spontaneous potential activity. EEG signals are usually divided into different rhythms according to the frequency bands: δ, θ, α, β, and γ. EEG signals of different rhythms can reflect different physiological and psychological state information of the human body, among which: [0003] The δ rhythm is mainly located in the 1-4Hz frequency band, which mainly reflects the deep sleep state of people or special brain diseases; [0004] The θ rhythm is mainly located in the 4-8Hz frequency band, which mainly reflects the state of early sleep, meditation, sleepiness or depression; [0005] The α rhythm is mainly located in the 8-12Hz frequency band. In additio...

AI Technical Summary

Problems solved by technology

Since the existing neural network methods use a single time convolution kernel in the time domain convolution, it is impossible to fully extract the characteristics of the EEG signal, which affects the final classification result.

Images