Science University Research Symposium (SURS)

EYE-TRACKING ANALYSIS OF ATTENTION PATTERNS ACROSS LEARNING MODALITIES: COMPARING VIDEO, AUDIO, AND TEXT-BASED EDUCATIONAL MATERIALS

Publication Date

Fall 11-10-2025

College

College of Sciences & Mathematics

Department

Psychological Science, Department of

SURS Faculty Advisor

Lingfei Luan

Presentation Type

Poster Presentation

Abstract

Research has suggested that although different learning formats produce similar learning outcomes, there may be significant differences in each format's ability to engage students and maintain their attention (Zhang et al., 2024). In the current learning environment, students' success has become increasingly dependent on sustained engagement while competing media constantly vie for their attention. This study investigates how AI-generated educational content across different modalities (audio, video, text) influences attentional engagement using a novel electrooculography (EOG) approach. Our innovative methodology combines two breakthrough elements: (1) AI-generated stimuli ensuring perfect content equivalence across modalities, eliminating confounding variables from human-created materials, and (2) real-time EOG vector analysis that creates continuous attention maps rather than traditional binary fixation measures. Participants experience identical educational content through randomly assigned modalities while EOG sensors track horizontal and vertical eye movements, integrating these into dynamic attention vectors relative to the stimulus presentation area. This approach captures not just whether students look at content, but the stability and consistency of their visual attention throughout the learning experience. We expect to identify distinct attentional signatures for each modality, with video showing highest initial engagement but greatest variability, text demonstrating steady but moderate engagement, and audio revealing unique patterns of visual anchoring behavior. This research pioneers the use of AI-controlled content generation and continuous EOG vector analysis in educational research, offering the first empirical framework for matching learning modalities to attentional capacity.

This document is currently not available here.

Share

COinS