Eye Tracking in Virtual Reality : Detecting Age-Related Cognitive Decline Through Naturalistic Tasks

Abstract

Eye movements, recorded in virtual reality (VR) environments which simulate everyday activities, could provide objective markers of cognitive decline in aging populations. Using a head-mounted display with integrated eye-tracking, this thesis examined visual behavior in younger (n = 38, age 18-35 years) and older (n = 23, age 60-85 years) during a VR-based prospective memory task. While Older Adults showed longer fixation durations (M = 325 ms vs. M = 312 ms, p = 0.02) and smaller saccade amplitudes (M = 2.88° vs. M = 3.02°, p = 0.035), the most pronounced differences emerged in attention allocation patterns, with Older Adults spending significantly more time on task-irrelevant objects (66.6% vs. 61.4%, p < .001). Importantly, within the Older Adults who spent more time looking at irrelevant items demonstrated poorer cognitive performance (r = -0.84, p = .001). These findings suggest that attention control efficiency, rather than basic oculomotor function, may serve as a sensitive marker for detecting cognitive differences in aging populations, particularly when assessed through task-relevant behaviors in ecologically-relevant VR tasks.