A multimodal investigation of learner attention and cognitive load in linear and non-linear educational content

The measurement of real-time cognitive load remains challenging in e-learning environments, while self-report measures provide limited insight into learners' moment-to-moment cognitive responses. This study examines whether combining eye-tracking and physiological monitoring can provide such an objective account across different instructional content formats. A Tobii Pro Fusion eye tracker and Polar heart rate sensor were used in a within-subjects study where 52 participants engaged with four types of educational content: text-heavy linear, text-heavy non-linear, graph-heavy linear, and graph-heavy non-linear. Based on gaze patterns, heat maps, and heart rate data, two main reading strategies emerged: F-pattern readers (n = 23, 44%) and Z-pattern readers (n = 29, 56%). Graph-heavy content was associated with longer completion times and higher physiological markers of cognitive effort. Heart rate increases of more than 30% over individual baseline were consistently associated with periods of higher cognitive strain, suggesting a possible threshold for adaptive instructional intervention. Mean heart rate rose from 92.08 bpm for text-heavy linear material to 122.75 bpm for graph-heavy non-linear content, indicating greater physiological engagement as instructional complexity increased. Although these results suggest practical directions for adaptive e-learning systems, learning analytics dashboards, and content delivery in LMS platforms like Moodle, Blackboard, and Canvas, they are a foundation rather than definitive design rules because this was an exploratory, descriptive study. Based on this, three outputs are proposed: a real-time intervention protocol framework, a personalisation model based on objective learner profiling, and evidence-based content creation guidelines.

This work is licensed under a Creative Commons Attribution 4.0 International License.