Related Experiment Video
Updated: Jan 15, 2026
10:41
Using Electroencephalography Measurements and High-quality Video Recording for Analyzing Visual Perception of Media Content
Published on: May 26, 2018
7.3K
Detecting mind wandering via EEG and facial video features
Shaohua Tang1,2,3, Chunbo Jiang4, Zheng Li5,6
1Department of Systems Science, Faculty of Arts and Sciences, Beijing Normal University, Zhuhai, 519087, Guangdong, China.
Behavior Research Methods
|October 10, 2025
Summary
Detecting mind wandering (MW) in online learning is improved using multimodal data from electroencephalography (EEG) and webcam video. Individual differences and diverse training data enhance classification accuracy for better educational tools.
Area of Science:
- Cognitive Science
- Educational Technology
- Neuroscience
Background:
- Mind wandering (MW) is a common challenge in online education, where attention shifts away from learning tasks.
- Detecting MW is crucial for developing adaptive learning systems.
- Multimodal data offers potential for more accurate MW detection.
Purpose of the Study:
- To develop a classification scheme for mind wandering detection using electroencephalograph (EEG) signals and facial video.
- To analyze feature contributions from EEG and video data.
- To explore correlations between self-reported confidence, mental state stability, and MW classification performance.
Main Methods:
- Collected data from 26 college students during a video-based learning task.
- Utilized a probe-based sample extraction method for EEG data.
- Employed a random forest algorithm with features from EEG and facial video.
- Evaluated model performance using within-participant and leave-one-participant-out (LOPO) cross-validation.
Main Results:
- Combined EEG and video features outperformed single modalities (AUC = 0.68 within-participant, 0.56 LOPO).
- Individual differences significantly impacted performance; including participant-specific data improved AUC by 10%.
- Higher introspective confidence correlated positively with performance; mental state stability improved cross-participant accuracy.
Conclusions:
- Multimodal approaches show significant potential for accurate mind wandering detection.
- Individual differences and diverse training data are critical for robust MW detection systems.
- Findings offer insights for improving real-world MW detection in educational settings.
