Yue Guzhang¹, Krishnaveni Nagarajan¹, Yesol Yu¹, Martina Poletti¹; ¹University of Rochester

Simulated scotomas are often used to study how central vision loss affects reading. However, due to technical limitations in eyetracking, most studies employ relatively large text sizes and scotomas occluding an area much larger than the foveola. Under these conditions, saccadic behavior is substantially altered. Yet it remains unclear whether much subtler disruptions to central vision also produce measurable changes in eye movements during reading, changes that could serve as early biomarkers for certain forms of retinal degeneration. Here, we used a high-resolution eyetracker paired with a gaze-contingent display that enables precise localization of the line of sight. We simulated a small scotoma centered on the preferred locus of fixation, rendered with a Gaussian transparency window (SD = 0.2°) that blended with the background. Participants (N = 8 emmetropes) read brief paragraphs (0.2° x-height). After each paragraph, they answered true/false comprehension questions and completed a word-recognition task indicating whether four target words had appeared in the passage. Participants showed an 8% drop in comprehension accuracy (P < 0.01) and a 9% drop in word-recognition accuracy (P < 0.01) in the scotoma condition compared to the no-scotoma condition. The small central scotoma also altered fixational eye movements; the duration of intersaccadic periods increased by an average of 42 ms (P < 0.01), and ocular drifts became significantly less curved and covered a larger spatial span relative to baseline. Additionally, in the scotoma condition, both saccade and microsaccade rates decreased (by 10% and 12.3%, respectively, P < 0.01), and the rate of regressive microsaccades increased.

Acknowledgements: This work was funded by NIH Grant EY029788 to MP and NIH grant EY001319 to the Center for Visual Science.