AUTHOR=Li Rui , Liu Di , Li Zhijun , Liu Jinli , Zhou Jincao , Liu Weiping , Liu Bo , Fu Weiping , Alhassan Ahmad Bala 

TITLE=A novel EEG decoding method for a facial-expression-based BCI system using the combined convolutional neural network and genetic algorithm

JOURNAL=Frontiers in Neuroscience

VOLUME=Volume 16 - 2022

YEAR=2022

URL=https://www.frontiersin.org/journals/neuroscience/articles/10.3389/fnins.2022.988535

DOI=10.3389/fnins.2022.988535

ISSN=1662-453X

ABSTRACT=Multiple types of brain control systems have been applied in the field of rehabilitation. As an alternative scheme to balance the user’s fatigue and the classification accuracy of brain computer interface (BCI) system, facial expression based brain control technology was serves as a novel BCI system has been proposed. Unfortunately, the existing machine learning algorithms fail to meet the most relevant features of EEG signal that further limit the performance of the classifier. To address this problem, an improved classification method was proposed for facial expression based BCI system using a Convolutional Neural Network (CNN) combined Genetic Algorithm (GA) in this study. The CNN was applied to extract features and classify them. The GA was used in the process of hyper-parameters selection to extract the most relevant parameters for classification. To validate the superiority of proposed algorithm used in this study, the performance of various experimental results was systematically evaluated and the trained CNN model was constructed to control an intelligent car in real time. The average accuracy from all subjects was 89.21±3.79% and the highest accuracy was up to 97.71±2.07%, respectively. The superior performance of proposed algorithm was demonstrated by both offline and online experiments. All the experimental results demonstrated that the performance of our improved FE-BCI system outperforms the state-of-art methods in the term of facial expression based BCI system (FE-BCI).