EEGTCW: Electroencephalogram-based Chinese Words Decoding

With the evolution of artificial intelligence technology and the exploration of brain science, more and more researchers are committed to the decoding of brain signals, and the application of brain-computer interface(BCI) is constantly updated. Electroencephalogram(EEG) signals are non-invasive signals that have attracted increasing attention due to their non-transplantability and high temporal resolution. This paper focuses on the decoding of EEG signals to Chinese words, and proposes a novel framework, EEGTCW, which is based on the BiGRU-Attention-CNN classification model and data augmentation method to decode EEG signals into categories corresponding to Chinese words on a self-built dataset. The classification accuracy reaches 0.88, while exploring the gener-alization performance of the model under different batches of a single subject and generalization performance across subjects. It has certain reference value for the improvement of BCI assistant system for speech disorder people.