Motor imagery electrocorticogram (EEG) signal classification method based on independent component analysis

Abstract

The invention discloses a motor imagery electrocorticogram (EEG) signal classification method based on independent component analysis, which comprises the following steps: S1, collecting the EEG signals and preprocessing the EEG signals, and randomly dividing the preprocessed EEG signals into a training set and a test set; S2, sequentially selecting the training set data to carry out independent component analysis and calculation on the single test sample data, and automatically identifying and acquiring the motion related component based on the spatial distribution pattern of the source; S3:classification and recognition of motor imagery based on zero training classifier; S4: optimizing the selection of leads by using the training set data, substituting the optimized leads into the testset, and looping the steps S2 and S3 to obtain the final classification recognition rate. The invention can reduce the spatial model matching problem caused by the difference between the collected EEGdata, and has high recognition accuracy to the motor imagery EEG signals.

Description

technical field [0001] The invention relates to the technical field of brain-computer interface, in particular to a motor imagery EEG signal classification method based on independent component analysis. Background technique [0002] Brain-computer interface (Brain Computer Interaction, BCI) is a new type of human-computer interaction, by reading and analyzing the subject's brain neurophysiological signals, to achieve direct control of the human brain on external devices. Electroencephalography (EEG) signals use scalp electrodes to record the synchronous electrophysiological activities of brain neurons. Due to its high time resolution and easy operation, it is widely used in brain-computer interface systems. [0003] Studies have shown that both the real movement of the limbs and the imaginary movement of the brain will cause changes in the blood flow and brain metabolism in specific areas of the cerebral motor cortex, which are manifested as the energy attenuation and incre...

AI Technical Summary

Problems solved by technology

[0005] The ICA algorithm mainly has the following shortcomings in the implementation process of the MIBCI system: (1) It is sensitive to the selection of leads, and the ICA spatial domain filter is designed by using different lead data, and the classification performance obtained is quite different

Using more leads will increase the possibility of introducing noise interference, while fewer leads will not provide enough information to separate the task-related neural sources; (2) Sensitive to data quality, the ICA algorithm can learn from multi-lead The original EEG signal has a fixed spatial position and an independent signal source in the time domain, but some sudden artifacts (such as: caused by sudden movement of the body or electrode shedding, etc.) cannot be obtained from the original EEG signal due to its short duration. Therefore, the quality of the data is also one of the key factors affecting the ICA algorithm; (3) It is difficult to match the model between different subjects or between data collected in different periods of the same subject

Due to individual differences, the EEG signals between different subjects are quite different. Even for the same subject, due to the changes of many factors such as the subject's mental state and environment in different periods, the spatial filter is not effective in the subject. When cross-validating between subjects (Subject-to-Subject) and between groups (Sesssion-to-Session), the classification performance obtained is low

Images