Recent developments in Brain-Computer Interface technologies have increased the ability to personalize Learning by detecting and recognizing participants' cognitive and emotional affective states. From a global point of view, working in monitoring and identifying the emotional and cognitive states of the students will establish the basis to incorporate a higher level of academic supervision and control of student performance. In this paper, the authors used an Open EEG (Electroencephalography) Dataset to analyze the correlation between student brainwaves in the frontal lobe when watching videos categorized in levels of confusion by using Machine Learning algorithms. It was found that alpha1, beta1, and theta bands are highest correlated to confusion with p=0.012, p=0.085, and p=0.0016, respectively. A comparison between some algorithms showed that the Convolutional Neural Network (CNN) + LSTM Model presented 75% highest accuracy. Furthermore, the level of student cognitive engagement was computed, in terms of those three brainwaves, obtaining a p= 0.625. The results suggest that Machine Learning is a powerful tool for analyzing brain activity. This paper contributes to neurosciences applied to Personalized Learning.
Cognitive Load. Task Complexity. Confusion. Engagement. EEG. Machine Learning. Personalized Learning.