Enhancing the feasibility of cognitive load recognition in remote learning using physiological measures and an adaptive feature recalibration convolutional neural network.

The precise assessment of cognitive load during a learning phase is an important pathway to improving students' learning efficiency and performance. Physiological measures make it possible to continuously monitor learners' cognitive load in remote learning during the COVID-19 outbreak. However, maintaining a good balance between performance and computational cost is still a major challenge in advancing cognitive load recognition technology to real-world applications. This paper introduced an adaptive feature recalibration (AFR) convolutional neural network to overcome this challenge by capturing the most discriminative physiological features (EEG and eye-tracking). The results revealed that the optimal average classification accuracy of the feature combination obtained by the AFR method reached 95.56% with only 60 feature dimensions. Additionally, compared with the best result of the conventional correlation-based feature selection (CFS) method, the introduced AFR algorithm achieved higher accuracy and cheaper computational cost, as well as a 2.06% improvement in accuracy and a 51.21% reduction in feature dimension, which is more in line with the requirements of low delay and real-time performance in practical BCI applications.

Emotional Activity Is Negatively Associated With Cognitive Load in Multimedia Learning: A Case Study With EEG Signals.

We aimed to investigate the relationship between emotional activity and cognitive load during multimedia learning from an emotion dynamics perspective using electroencephalography (EEG) signals. Using a between-subjects design, 42 university students were randomly assigned to two video lecture conditions (color-coded vs. grayscale). While the participants watched the assigned video, their EEG signals were recorded. After processing the EEG signals, we employed the correlation-based feature selector (CFS) method to identify emotion-related subject-independent features. We then put these features into the Isomap model to obtain a one-dimensional trajectory of emotional changes. Next, we used the zero-crossing rate (ZCR) as the quantitative characterization of emotional changes ZCR EC . Meanwhile, we extracted cognitive load-related features to analyze the degree of cognitive load (CLI). We employed a linear regression fitting method to study the relationship between ZCR EC and CLI. We conducted this study from two perspectives. One is the frequency domain method (wavelet feature), and the other is the non-linear dynamic method (entropy features). The results indicate that emotional activity is negatively associated with cognitive load. These findings have practical implications for designing video lectures for multimedia learning. Learning material should reduce learners' cognitive load to keep their emotional experience at optimal levels to enhance learning.