Investigating the Impact of a Simulated Gaze-Contingent Central Scotoma on Eye Movements Using Augmented Reality

Central Vision Loss (CVL) due to conditions such as Age-related Macular Degeneration (AMD) can significantly impair daily activities by disrupting the ability to process fine visual details. When the foveal region is compromised, individuals rely on compensatory mechanisms such as peripheral vision strategies to navigate and interact with their surroundings. Augmented Reality Head-Mounted Displays (AR-HMDs) offer a promising tool for investigating visual adaptation by simulating central vision deficits in controlled experimental settings. This study explores the impact of a gaze-contingent central scotoma simulation on visual strategies using the Microsoft HoloLens 2. The scotoma was implemented in Unity as an opaque white region that dynamically followed the participant's gaze, effectively blocking central visual input. A cohort of normally sighted participants performed two visually demanding tasks, reading and visual search, under two conditions: with and without a simulated central scotoma. The reading task consisted of two parts: in the first, participants read aloud a series of short standard texts, while in the second, they were required to read aloud numbers embedded within phrases, forcing them to focus on numerical recognition rather than linguistic context. The visual search task required participants to locate a target letter ("T") among distractor letters ("L") as quickly as possible. Reading time, search time, and error rates were recorded for both tasks, while Eye–Tracking (ET) data, were analyzed to extract fixation and saccade parameters. Two computational approaches were employed to classify eye movements: a velocity-based threshold method and a Hidden Markov Model (HMM) leveraging velocity and acceleration as classification features. Statistical analyses revealed significant differences across most metrics, indicating a substantial impact of the simulated scotoma on visual behavior. Participants in the scotoma condition exhibited a greater number of fixations and saccades, along with an increased total fixation duration, suggesting a heightened effort to compensate for the absence of foveal vision. Despite these changes, the average duration of individual fixations and saccades remained stable across conditions, implying that while visual exploration became more fragmented, the time required to process individual fixations was unaffected. These findings align with existing literature on central vision deficits, reinforcing the hypothesis that individuals adapt to vision loss through increased ocular exploration. The higher fixation and saccade rates suggest an effort to reconstruct missing visual information, albeit at the cost of increased cognitive load. Longer reading times and higher error rates further indicate a deterioration in reading efficiency, likely due to the inability to process entire words in a single fixation. In the visual search task, prolonged reaction times highlight a reduced efficiency in target detection, reflecting the reliance on peripheral vision for scene analysis. This study underscores the effectiveness of Augmented Reality (AR)-based scotoma simulations in examining adaptive visual strategies and provides a framework for future investigations into compensatory mechanisms in visual impairment.