Abstract
Purpose :
Although a gaze feedback system is effective for retinal prostheses using an external camera (Hozumi et al ARVO 2016), there are issues of processing time and device weight in conventional eye-tracking devices. In this study, we propose a new eye-tracking method that estimates direction of gaze using stimulus artifacts generated by a retinal prosthesis. To confirm its feasibility, we evaluated the accuracy of the eye-tracking method using animal models.
Methods :
In this study, a stimulating electrode was implanted into the scleral pocket created at the posterior pole of the eyes of anesthetized cats (n = 3). Stimulus pulses of 0.5 ms duration and 0.1–0.5 mA amplitude were applied to the eye via the stimulating electrode. A pair of recording electrodes was placed on the medial and lateral canthus of the implanted eye. Stimulus artifacts were evaluated by measuring the electrical potential difference between the recording electrodes. The recording trials were repeated 100 times. To measure gaze direction, we illuminated the eye and projected the fundus image on the screen placed 1.0 m in front of the eye. Changes in gaze direction were evaluated by measuring the shift in the fundus image on the screen. Finally, we investigated the relationship between the amount of shift in gaze direction and changes in the amplitude of stimulus artifacts.
Results :
The amplitude of the stimulus artifact was significantly correlated with gaze direction in all animals (#1, r = 0.996 and p < 0.01; #2, r = 0.970 and p < 0.01; #3, r = 0.997 and p < 0.01; Pearson correlation analysis). Standard deviations of the residual error of linear regression were 0.1, 0.5, and 0.2 degrees in #1, #2, and #3, respectively.
Conclusions :
These findings demonstrate that gaze direction can be estimated using stimulus artifacts evoked by retinal prosthesis without delay of processing time.
This is an abstract that was submitted for the 2017 ARVO Annual Meeting, held in Baltimore, MD, May 7-11, 2017.