Journal article
Knowledge representation learning with EEG-based engagement and cognitive load as mediators of performance
Behaviour & information technology, Vol.44(13), pp.3180-3193
08/09/2025
DOI: 10.1080/0144929X.2024.2438776
Abstract
Educational and instructional research has provided contrasting results regarding the best representation of numerical information, with the two most common being tabular and graphical representations. This motivated us to examine the issue using a novel approach. We employed electroencephalography (EEG) with event-related synchronisation and desynchronisation (ERS/ERD) to model cognitive load, and alpha and beta band powers to model engagement. We conducted an experiment to measure the cognitive load and engagement of 48 Oregon State University students and compared their performances with respect to these two representations. Structural equation models (SEMs) were constructed to investigate the potential mediation of learning performance by engagement and cognitive load. The results indicate that learning performance was fully mediated by cognitive load and engagement. Relative to graphs, tables produced a higher cognitive load and engagement, and subsequently, greater overall learning performance. Current results showed that different representations could yield significant differences in learning performance, and that an understanding of representation-elicited affective behaviours has considerable potential for future online learning instructional design.
Details
- Title: Subtitle
- Knowledge representation learning with EEG-based engagement and cognitive load as mediators of performance
- Creators
- Yuzhi Sun - Oregon State UniversityDavid A. Nembhard - Harvey Mudd College
- Resource Type
- Journal article
- Publication Details
- Behaviour & information technology, Vol.44(13), pp.3180-3193
- DOI
- 10.1080/0144929X.2024.2438776
- ISSN
- 0144-929X
- eISSN
- 1362-3001
- Publisher
- Taylor & Francis
- Number of pages
- 14
- Grant note
- Oregon State University
- Language
- English
- Electronic publication date
- 12/11/2024
- Date published
- 08/09/2025
- Academic Unit
- Business Analytics
- Record Identifier
- 9984759115902771
Metrics
12 Record Views