Abstract Background The rapid growth of online learning platforms has improved educational access globally. However, most systems rely on static content delivery, providing identical materials to all learners regardless of individual abilities or learning pace. This uniform approach disregards variations in cognitive capacity, causing excessive mental strain for some and insufficient challenge for others, resulting in reduced engagement, poor retention, and ineffective learning outcomes. Methods The proposed system estimates cognitive load in real-time using non-intrusive interaction data: response time, accuracy, error rate, number of attempts, and help usage. A novel metric, the Cognitive Load Stability Index (CLSI), quantifies a learner’s mental effort and consistency. A rule-based adaptive engine uses CLSI thresholds to dynamically adjust content difficulty and provide structured support (hints, explanations), enabling real-time personalization without compromising privacy or scalability. Results The adaptive system demonstrated significant improvements over static systems: approximately 25–35% increase in accuracy, ~ 30% reduction in error rates, and ~ 40% increase in user engagement. The CLSI metric effectively detected cognitive load variations, enabling timely adjustments to content delivery and support. Conclusion Integrating cognitive load estimation into adaptive learning systems enhances personalization and learning efficiency. The CLSI metric provides a meaningful approach to monitor cognitive stability. The system improves engagement, reduces cognitive overload, and enhances outcomes while maintaining a non-intrusive, privacy-preserving design suitable for large-scale deployment.
Sharma et al. (Wed,) studied this question.