Improved Concentrated Mental State Classification Through EEG Signal Augmentation and One-Dimensional Convolutional Neural Network

This work categorizes mental states into relaxed, neutral, and concentrated states based on electroencephalogram (EEG) signals. This classification of mental states is helpful in human–machine interaction, brain-computer interface systems, and cognitive modeling of the human brain. Here, a simple approach of data augmentation/up-sampling has been proposed with a deep learning model for the classification of EEG signal, because the database used in this study has less number of samples for the training of the model. This is a common problem with other databases also, the proposed approach can be easily applied to other databases also. One-Dimensional Convolutional Neural network (1-D CNN) is utilized for performing classification. Classification results show significant improvement in classification accuracy after the use of the proposed data augmentation method. After data augmentation, on average classification accuracy is increased by 58.09%, as compared to an original database.