Exploring the Role of Tunnel Vision Simulation in the Design Cycle of Accessible Interfaces
Rie Kamikubo, Keita Higuchi, Ryo Yonetani, Hideki Koike, Yoichi Sato · 2018 · Proceedings of the 15th International Web for All Conference (W4A 2018) · doi:10.1145/3192714.3192822
Summary
This paper investigates whether gaze-contingent tunnel vision simulation can serve as a reliable tool within the design cycle for accessible web interfaces, enabling sighted developers to experience and evaluate prototypes under simulated visual field loss. The challenge motivating this work is that despite the emphasis on involving users with disabilities in accessible interface development, active involvement of low vision users is difficult in practice: the diverse range of visual abilities (even within a single condition like Retinitis Pigmentosa) complicates finding representative participants and conducting controlled evaluations. The simulation uses a Tobii eye tracker to estimate gaze direction on a 22-inch monitor and renders only a 5-degree visual angle of content around the point of gaze in real time, simulating the tunnel vision experienced by people with RP or glaucoma. The research follows a three-phase design cycle: Phase 1 involved user research with 5 low vision participants (3 with RP) to establish requirements for web navigation under reduced peripheral vision; Phase 2 validated the simulation by comparing saccadic eye movement patterns across sighted, simulated tunnel vision, and RP conditions (6 participants each); Phase 3 designed five navigation aid prototypes and evaluated them through usability testing (12 simulated participants) and user testing (6 RP participants). The navigation aids use visual markers to guide users to important page regions: three relative designs (RA: arrows, RL: compass line, RI: inverted colours following gaze) and two absolute designs (AL: radial lines emanating from targets, AA: vector field arrows).
Key findings
The validation study confirmed that the simulation produced gaze patterns statistically consistent with those of RP participants (chi-squared p=.218, not significantly different) and significantly different from normal sighted viewing (p=.001). Both simulated and RP participants made large saccadic eye movements to compensate for loss of spatial understanding — median 3 and 5 large saccades per target search respectively, versus 1 for sighted participants. In usability testing, the absolute prototype AL (radial lines) was favoured by both simulated and RP groups for direction, distance, ease of task, and overall preference. Both groups found relative navigation cues required more perceptual effort than static absolute ones. Importantly, design preference rankings between simulated and RP participants were largely consistent, validating the simulation’s utility for comparing design alternatives. However, contradictions emerged in specific cases: RA (relative arrows) was most preferred by simulated participants for search tasks but ranked lowest by RP participants, who reported it was too difficult to detect with their reduced acuity. Similarly, RI (inverted colours) empowered RP participants to see content better but confused simulated participants with normal acuity. In user testing of online shopping tasks, the AL navigation aid significantly improved layout comprehension for both groups (p=.01 for simulated) and reduced self-reported cognitive effort in locating page elements, though it did not improve overview comprehension or actual search speed.
Relevance
This paper makes a methodological contribution to accessible design by empirically assessing when and how simulation can supplement (not replace) user involvement in the design cycle. The finding that simulation reliably supports early-stage design decisions — comparing design alternatives and identifying elementary UI problems — while failing to capture nuances related to actual visual acuity, content visibility, and detection thresholds provides practical guidance for developers. The three design qualities identified (Designing: quickly probe early ideas; Evaluation: capture subjective preferences through controlled simulated trials; Identification of Requirements: observe human factors from simulated perspectives) map simulation to specific development phases. For accessibility practitioners, the key takeaway is that simulation is most valuable for rapid iteration in early prototyping and for identifying obvious usability issues, but user research with people who have actual visual impairments remains essential for requirements gathering, validating visual design details, and understanding lived experiences. The gaze-contingent simulation approach itself — requiring only a consumer eye tracker and display — is technically accessible to development teams.
Tags: low vision · tunnel vision · retinitis pigmentosa · simulation · eye tracking · accessible design · design methodology · usability testing · web accessibility · visual field loss