Predicting Cognitive Load Using Sensor Data in a Literacy Game

Overview

• Educational games are being increasingly used for self-paced learning
• Tracking student affect and cognitive load is a challenge for educators and designers
• Existing assessments in game-based learning environments (GBLEs) focus more on outcomes than processes
• This study collected physiological and pupil data to develop a model that tracks cognitive load during gameplay

Plain English Explanation

Educational games are becoming more common as a way for students to learn at their own pace. However, it can be difficult for teachers and game designers to monitor how students are feeling and how much mental effort they are exerting while playing these games. Typical assessments in game-based learning tend to focus more on the final results rather than the learning process itself, which means they may miss important aspects like the student's emotional state and cognitive load.

To address this issue, the researchers in this study collected various types of data from students as they played an online literacy game. They looked at physiological measures like sweat response and heart rate, as well as pupil data, to develop a model that could track the students' cognitive load - in other words, how much mental effort they were expending during the game.

The researchers found that their multimodal model was able to distinguish different levels of cognitive load with reasonable accuracy. This suggests that these types of data could be used to create more personalized learning experiences, where the game adapts to the student's cognitive state in real-time.

Technical Explanation

The researchers collected physiological and pupil data from students as they played an online literacy game, with the goal of developing a model to track cognitive load in game-based learning environments (GBLEs). They recorded electrodermal activity (EDA), heart rate, and pupil data during gameplay.

Using this multimodal data, the researchers trained a model to classify different levels of cognitive load. Their model achieved a Kappa score of 0.417 and an accuracy of 70%, indicating that it was able to distinguish between different cognitive load states with reasonable performance.

The results suggest that including affect-related features, such as EDA and heart rate, is important for accurately predicting cognitive load. This builds on recent findings that using multiple data channels can be beneficial for modeling latent aspects of learner processes.

The researchers propose that this type of cognitive load tracking could now be used to facilitate the creation of personalized learning experiences in game-based learning environments, where the game dynamically adapts to the student's cognitive state.

Critical Analysis

The study demonstrates the feasibility of using physiological and pupil data to track cognitive load in game-based learning, which is a valuable contribution to the field. However, the researchers acknowledge that their model's performance, while promising, is still relatively modest and could be improved with further refinement.

Additionally, the study was conducted in a controlled laboratory setting, and it's unclear how well the model would generalize to more naturalistic, real-world gameplay scenarios. Further research is needed to validate the approach in more diverse and ecologically valid learning environments.

Another potential limitation is that the study focused solely on cognitive load, without considering other important factors like learner engagement, motivation, and emotional state. A more holistic approach that incorporates these additional aspects of the learning process may provide a richer understanding of the learner experience.

Overall, the research presents an important step forward in the quest to better monitor and support self-paced learning in game-based environments. By continuing to explore multimodal approaches to tracking learner processes, researchers and designers can work towards creating more personalized and effective educational games.

Conclusion

This study demonstrates the potential of using physiological and pupil data to track cognitive load in game-based learning environments. The researchers developed a multimodal model that was able to distinguish different levels of cognitive load with reasonable accuracy, highlighting the importance of including affect-related features like electrodermal activity and heart rate.

The findings suggest that this type of cognitive load tracking could be used to facilitate the creation of more personalized learning experiences, where educational games dynamically adapt to the student's cognitive state. While further research is needed to validate the approach in more naturalistic settings, this study represents an important step towards better understanding and supporting the learning process in game-based learning environments.