A Novel Hybrid Deep Learning Approach for Classification of Cognitive States Using EEG Signals

Abstract

Cognitive states are crucial for understanding the complex functioning of the human brain, which determine our perceptions and interactions with the surrounding environment. These states include different mental processes, such as reasoning, problem-solving, decision-making, critical thinking, and information retention. The accurate identification and classification of these states necessitates the analysis of electroencephalogram (EEG) signals, which provide invaluable insights into the neural dynamics associated with these cognitive processes. In this paper, we discuss a novel hybrid deep learning method integrates convolutional neural network (CNN) with bidirectional long short-term memory (Bi-LSTM) for classifying the cognitive states. An EEG benchmark dataset is used to validate the model’s performance. Additionally, a comparison with other standard deep learning models is done. The results exhibit the effectiveness of the proposed model, attaining an accuracy of 85%, outperforming other models in terms of accuracy, precision, and recall, f1- score and AUC score. This approach leverages the strength of both CNN and Bi-LSTM, with CNN effectively extracting spatial features and the Bi-LSTM capturing long-term dependencies in the temporal sequence of the signals. The outcomes of this study contribute to advancing EEG signal analysis and gaining deeper insights into human cognition.