Assessing a Single Student's Concentration on Learning Platforms: A Machine Learning-Enhanced EEG-Based Framework

TL;DR

The paper addresses the problem of assessing a single student’s concentration during online learning using EEG and a personalized machine learning pipeline. It develops a workflow from data acquisition with a Muse headband through feature extraction across five EEG bands, feature selection, and hyperparameter tuning, culminating in a random forest classifier. The approach achieves high test accuracies in both computer-based and VR learning contexts (97.6% and 98%, respectively) and demonstrates strong discrimination via ROC AUC of 1.00 across classes. The work highlights the potential for personalized, EEG-based learning analytics while noting limitations in participant diversity and suggesting future validation with broader cohorts and tasks for real-world deployment.

Abstract

This study introduces a specialized pipeline designed to classify the concentration state of an individual student during online learning sessions by training a custom-tailored machine learning model. Detailed protocols for acquiring and preprocessing EEG data are outlined, along with the extraction of fifty statistical features from five EEG signal bands: alpha, beta, theta, delta, and gamma. Following feature extraction, a thorough feature selection process was conducted to optimize the data inputs for a personalized analysis. The study also explores the benefits of hyperparameter fine-tuning to enhance the classification accuracy of the student's concentration state. EEG signals were captured from the student using a Muse headband (Gen 2), equipped with five electrodes (TP9, AF7, AF8, TP10, and a reference electrode NZ), during engagement with educational content on computer-based e-learning platforms. Employing a random forest model customized to the student's data, we achieved remarkable classification performance, with test accuracies of 97.6% in the computer-based learning setting and 98% in the virtual reality setting. These results underscore the effectiveness of our approach in delivering personalized insights into student concentration during online educational activities.

Figures (6)

• Figure 1: Muse S EEG Headband: the green and yellow labels correspond to the international 10-20 EEG electrode placement standard
• Figure 2: The International 10-20 EEG Electrode Placement Standard b5: the yellow labels are electrodes of Muse headband and green labeled sensor is deployed as a reference point for calibration
• Figure 3: Example of EEG Variations over Time: the x-axis indicates time (in seconds), and the y-axis represents the logarithm of power spectral density across different frequency bands (delta, theta, alpha, beta, gamma)
• Figure 4: Validation Accuracy Variation: the x-axis represents the number of selected features, with an interval of 5, and y-axis stands for validation accuracy
• Figure 5: Generalizability Representation: these ROC curves illustrate the performance of the RF model on the (a) non-VR and (b) VR test subset, achieving an AUC of 1.00 for each class