Steady-state visual evoked potential signal classification method based on convolutional neural network

Abstract

A steady-state visual evoked potential signal classification method based on a convolutional neural network comprises the following steps: firstly, presenting checkerboard stimuli flipped at differentfrequencies to a user at the same time, and acquiring an electroencephalogram signal when the user watches a specific target by using an electroencephalogram acquisition device; making original multi-channel electroencephalogram signals when the user watches different stimulation targets into a data set with labels, and dividing the data set into a training set, a verification set and a test set;inputting the training set into a designed deep convolutional neural network model for training, performing network optimal parameter selection by using a verification set, and finally inputting thetest set into the trained deep convolutional neural network model to complete the recognition of the stimulation target. Accurate identification of steady-state visual evoked potential signals can beachieved, the characteristic of adaptively extracting signal characteristics is achieved, manual preprocessing is not needed, and meanwhile individual difference can be better adapted through data learning.

Description

technical field [0001] The invention relates to the technical field of steady-state visual evoked potential brain-computer interface, in particular to a method for classifying steady-state visual evoked potential signals based on a convolutional neural network. Background technique [0002] Brain-computer interface (brain-computer interface, BCI) is a technology that does not depend on the normal output pathway of the brain, but directly realizes the communication between the brain and external devices such as computers. A means of communicating and controlling the external environment, such as manipulating a wheelchair through brain ideas. Commonly used brain-computer interface signal types include steady-state visual evoked potential (SSVEP), motor imagery, P300, etc. Among them, SSVEP has the advantages of strong stability and simple operation, and has become a widely used Brain-computer interface input signal. [0003] SSVEP is the response of the brain's visual system...

AI Technical Summary

Problems solved by technology

These traditional signal processing methods generally require a long period of visual stimulation to achieve better classification results, resulting in low recognition efficiency; and these methods use manual feature extraction to easily lead to information loss. Using the same recognition method for different users, Individual differences are not considered, so the recognition accuracy is low, which limits the engineering application of SSVEP-BCI

Images