Monitoring attentional states in education using fNIRS
E-learning has been an educational tool since the 19th century, that with the popularization of videoconferences. it was consolidated with the creation of the first university in this modality, the Open University (OU) in the United Kingdom.
Since then, interactivity in virtual educational environments has key concepts, such as the exercises used to reinforce the content learning, capturing the attention and engagement of the student.
This work aims to investigate the high and low levels of task engagement through functional signals of near infrared tissue spectroscopy (fNIRs) of the pre-frontal cortex. The sample is composed of higher education students and post-graduate students while
watching video courses from an e-learning course. To model the data, the
implementation of supervised learning algorithms for classification of correct and incorrect responses to a questionnaire were used, based on oxyhemoglobin and deoxyhemoglobin prefrontal concentrations. The implementation of decision tree did not result in a suitable classifier, while the random forest and
logistic regression with LASSO achieved satisfactory results.