Gaze Compensation as a Technique for Improving Hand-Eye Coordination in Prosthetic Vision

Abstract

Purpose: Shifting the region-of-interest within the input image to compensate for gaze shifts ("gaze compensation") may improve hand-eye coordination in visual prostheses that incorporate an external camera. The present study investigated the effects of eye movement on hand-eye coordination under simulated prosthetic vision (SPV), and measured the coordination benefits of gaze compensation. Methods: Seven healthy-sighted subjects performed a target localization-pointing task under SPV. Three conditions were tested, modeling: retinally stabilized phosphenes (uncompensated); gaze compensation; and no phosphene movement (center-fixed). The error in pointing was quantified for each condition. Results: Gaze compensation yielded a significantly smaller pointing error than the uncompensated condition for six of seven subjects, and a similar or smaller pointing error than the center-fixed condition for all subjects (two-way ANOVA, P < 0.05). Pointing error eccentricity and gaze eccentricity were moderately correlated in the uncompensated condition (azimuth: R2 = 0.47; elevation: R2 = 0.51) but not in the gaze-compensated condition (azimuth: R2 = 0.01; elevation: R2 = 0.00). Increased variability in gaze at the time of pointing was correlated with greater reduction in pointing error in the center-fixed condition compared with the uncompensated condition (R2 = 0.64). Conclusions: Eccentric eye position impedes hand-eye coordination in SPV. While limiting eye eccentricity in uncompensated viewing can reduce errors, gaze compensation is effective in improving coordination for subjects unable to maintain fixation. Translational Relevance: The results highlight the present necessity for suppressing eye movement and support the use of gaze compensation to improve hand-eye coordination and localization performance in prosthetic vision.