CogniNet: A Deep Learning Model for the Prediction of Motor-Imagery EEG Signals
摘要
This study primarily delved into the development of a deep learning model specialized for the prediction of motor-imagery EEG signals. In this endeavor, the model was able to provide a validation accuracy of 72.37% against a testing accuracy of 99.68% – with the model possessing a parameter count of 4,356,227. Given the prevalence of overfitting into this particular model, it is highly suggested that future implementations place emphasis on the methodologies available for dimension matching for the integration of pre-trained deep learning models into that of MI-EEG prediction.