Using advanced neural networks to classify imagined movements from brain signals

Classification accuracies of 83.4%, 89.1%, and 73.3% were achieved on three different datasets.

• A novel data augmentation method and deep learning model were proposed to improve the decoding performance of motor imagery electroencephalography (MI-EEG).
• The raw EEG signals were converted into time-frequency maps using continuous wavelet transform to serve as model input.
• An advanced generative adversarial network was developed to effectively expand the training dataset, enhancing data quality.
• Validation showed that the generative network produced more realistic data, which is associated with improved classification results.
• The proposed model demonstrated superior performance compared to existing methods, indicating potential for better MI classification accuracy.

Simplified