Date of Defense
19-11-2024 2:00 PM
Location
E1-1012
Document Type
Dissertation Defense
Degree Name
Doctor of Philosophy in Informatics and Computing
College
CIT
Department
Computer Science
Keywords
Brain Computer Interface, Brain Synchronization, Cognitive Abilities, Eye-tracking, Machine Learning, Neurofeedback, Remote Learning.
Abstract
Attention and cognitive engagement are crucial factors in remote learning environments, where the absence of physical presence often diminishes learning outcomes. Traditional methods for assessing these cognitive states, such as observation and self-reporting, are limited by subjectivity and inefficiency. Automated solutions, particularly those based on biometric data like EEG and eye-tracking, offer a more accurate and scalable alternative. However, developing robust systems that leverage biometric data in real-time presents significant challenges. These include handling large volumes of complex data, ensuring low-latency processing, and adapting machine learning models to diverse learning environments and individual cognitive states. Additionally, the integration of neurofeedback and brain-to-brain synchronization in online learning is still an emerging area of research, posing further challenges in understanding and improving cognitive alignment between students and instructors. To address these challenges, we developed an integrated system that combines machine learning, EEG-based neurofeedback, and brain-to-brain synchronization analysis to enhance cognitive engagement in remote learning environments. We employed several machine learning models, including CNN, logistic regression, and KNN, to classify cognitive states based on EEG and eye-tracking data. The CNN model outperformed others, achieving an area under the curve (AUC) of 0.98, demonstrating its effectiveness in interpreting subtle cognitive patterns. Additionally, we introduced a real-time neurofeedback system, which provided participants with immediate feedback based on their brain activity, resulting in significant improvements in attention as measured by increased beta and gamma wave activity across multiple sessions. Power spectral density (PSD) analysis and low-latency feedback processing further validated the system’s efficacy in enhancing attention and cognitive engagement. Furthermore, we explored brain-to-brain synchronization between students and instructors during remote learning sessions. Using EEG data and the KNN algorithm, we successfully identified synchronization patterns that reflected a shared cognitive focus between participants. This analysis provided valuable insights into how synchronization and intermittent desynchronization influence learning outcomes and cognitive alignment. The findings of this study demonstrate the potential for machine learning and neurofeedback systems to improve remote learning by fostering higher levels of engagement and cognitive resonance. In conclusion, the proposed system offers a promising solution to the challenges of assessing and enhancing cognitive engagement in online learning environments. By automating the detection of cognitive states and improving brain-to-brain synchronization through neurofeedback, this research has the potential to revolutionize how educators assess and enhance cognitive performance in digital classrooms. Future work will focus on scaling the system, exploring personalized neurofeedback, and expanding the use of multimodal data to further improve learning outcomes.
Included in
IMPROVING STUDENTS’ COGNITIVE ABILITIES IN REMOTE LEARNING ENVIRONMENT USING BRAIN COMPUTER INTERFACE AND EYE-TRACKING
E1-1012
Attention and cognitive engagement are crucial factors in remote learning environments, where the absence of physical presence often diminishes learning outcomes. Traditional methods for assessing these cognitive states, such as observation and self-reporting, are limited by subjectivity and inefficiency. Automated solutions, particularly those based on biometric data like EEG and eye-tracking, offer a more accurate and scalable alternative. However, developing robust systems that leverage biometric data in real-time presents significant challenges. These include handling large volumes of complex data, ensuring low-latency processing, and adapting machine learning models to diverse learning environments and individual cognitive states. Additionally, the integration of neurofeedback and brain-to-brain synchronization in online learning is still an emerging area of research, posing further challenges in understanding and improving cognitive alignment between students and instructors. To address these challenges, we developed an integrated system that combines machine learning, EEG-based neurofeedback, and brain-to-brain synchronization analysis to enhance cognitive engagement in remote learning environments. We employed several machine learning models, including CNN, logistic regression, and KNN, to classify cognitive states based on EEG and eye-tracking data. The CNN model outperformed others, achieving an area under the curve (AUC) of 0.98, demonstrating its effectiveness in interpreting subtle cognitive patterns. Additionally, we introduced a real-time neurofeedback system, which provided participants with immediate feedback based on their brain activity, resulting in significant improvements in attention as measured by increased beta and gamma wave activity across multiple sessions. Power spectral density (PSD) analysis and low-latency feedback processing further validated the system’s efficacy in enhancing attention and cognitive engagement. Furthermore, we explored brain-to-brain synchronization between students and instructors during remote learning sessions. Using EEG data and the KNN algorithm, we successfully identified synchronization patterns that reflected a shared cognitive focus between participants. This analysis provided valuable insights into how synchronization and intermittent desynchronization influence learning outcomes and cognitive alignment. The findings of this study demonstrate the potential for machine learning and neurofeedback systems to improve remote learning by fostering higher levels of engagement and cognitive resonance. In conclusion, the proposed system offers a promising solution to the challenges of assessing and enhancing cognitive engagement in online learning environments. By automating the detection of cognitive states and improving brain-to-brain synchronization through neurofeedback, this research has the potential to revolutionize how educators assess and enhance cognitive performance in digital classrooms. Future work will focus on scaling the system, exploring personalized neurofeedback, and expanding the use of multimodal data to further improve learning outcomes.