Analysis of Complex Eye Movements in Virtual Reality Using the Concepts of Visual Entropy and Visual Groupings

The association between two visual scanpath measures—visual entropy and visual groupings—in eye movement data collected from a VR-based semantic network task is examined. Visual entropy quantifies gaze dispersion, while visual groupings capture structured gaze transitions. Eye movements were collected while participants created a semantic network using ten objects based on the relationship between the objects in the VR environment. Scanpath sequences were analyzed to compute visual entropy and identify visual groupings. Results showed that participants with higher visual entropy exhibited broader categorical groupings and multi-level structures, whereas those with lower entropy demonstrated localized gaze patterns and simpler networks with single connections. These findings suggest that visual entropy and visual groupings capture complementary aspects of gaze behavior, offering insights into visual interrogation strategies in VR environments. This study highlights the potential of these measures for assessing user engagement and learning performance in an advanced VR learning environment.