Modelling EEG Signals for Mental Confusion using CNN-Bi-LSTM with Attention

The popularity of e-learning continues to grow, yet accurately identifying a student's cognitive state during online education remains a significant challenge. One of the major issues students encounter in Massive Open Online Courses (MOOCs) is the lack of real-time feedback mechanisms. This often leads to a loss of focus or confusion during lessons, with no effective way for students to communicate their difficulties to educators. To tackle this problem, we propose a CNN-BiLSTM model with an integrated attention mechanism to analyze Electroencephalogram (EEG) data. This model is designed to capture both spatial and temporal features of brain activity, providing deeper insights into cognitive states. By leveraging this approach, we aim to surpass traditional methods such as Support Vector Machines (SVM), KNearest Neighbors (KNN), and basic deep learning models like RNNs and LSTMs. Our solution delivers real-time analysis to help educators tailor their teaching strategies, thereby enhancing student engagement and learning outcomes in MOOCs. Currently, our model achieves an accuracy of 80%, and we are optimizing it further with the goal of reaching a minimum accuracy of 95%. Unlike traditional classroom settings, where teachers can gauge student understanding through direct interaction, online education presents a unique challenge due to the lack of personal engagement. This gap can adversely affect the quality of education if not addressed.