Studying the effect of lecture content on students’ EEG data in classroom using SVD

The recent innovation in technology led to huge advancement in Human-Computer Interface (HCI) systems and applications. Detection of brain activities is the vital element in these applications. This paper is employing Singular Value Decomposition (SVD) on EEG data acquired simultaneously from students in classroom to detect the changes of brain activities during learning process. Situational interest of subjects and the learning materials were evaluated through questionnaires. After preprocessing and segmentation of the data, SVD was applied on each segment separately. The 2-norms of the singular values were compared to the subject baseline and the overall result complied with the questionnaire result. Furthermore, feeding these features to Support Vector Machine (SVM) classifier achieved 83.3% accuracy in differentiating between high and low situationally interested students. It is therefore, suggested that SVD could be applied successfully to detect changes in students’ brain activities in classrooms.