Use of brain-computer interface in educational paradigm

Brain-computer interface (BCI) technology has been utilized in research for several decades. Using BCI, a person interacts with their environment by controlling it using their brain signals. Brain signals can also help determine the level of confusion among students. In this study, we discuss using BCI to measure students' concentration levels. Using student data obtained from Kaggle, a publicly accessible data source, we deployed a variety of deep learning and machine learning algorithms that include support vector machines, logistic regression, decision trees, random forests, multi-layer perceptron, deep neural networks, and bidirectional long short-term memory were used in our comparative analysis. We proposed 1D CNN and DNN models in this research study. Although LSTM performs better when time stamps are present in the data, our analysis of the findings revealed that deep neural networks performed better in this case, with an accuracy of 99%. Another 1D CNN model is proposed to compare its performance with the DNN model. We found that the 1DCNN model is more stable than the DNN model but has slightly lower accuracy, i.e., 95%. Additionally, deep learning algorithms typically employ the rectified linear unit (ReLU) to mitigate vanishing gradient issues, resulting in a well-performing model. In contrast, 'RELU' has reversed the findings of BiSTM in this investigation, reducing accuracy from 82 to 67%. We have also conducted numerous experiments to identify the optimal parameter values for deep neural networks (DNNs) that yield the highest accuracy, specifically 99.9%. Therefore, DNN can be utilized in a real-time setting for the development of a confusion detection system, and BCI can be incorporated into the educational paradigm to reap the benefits of technological progress.