Purchase this article with an account.
G. Dagnelie, M. Barry, L. Yang; Effects of Gaze-locking on Visual Performance Under Simulated Prosthetic Vision (SPV) Conditions. Invest. Ophthalmol. Vis. Sci. 2009;50(13):4234.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
To compare performance of normally and partially sighted subjects on visually-guided motor tasks under free-viewing and gaze-locked SPV conditions in a head-mounted display. In free view, subjects can freely move the eyes to inspect the low-pass filtered, pixelized image of the task, while in the gaze-locked condition (representative of an electrode array fixated on the retina) the image follows the subject’s eye movements.
We retrospectively analyzed timing data collected on 3 tasks: 1) Wayfinding in a virtual maze, traversing 10 rooms; 2) Maze tracing on a tablet computer; 3) Covering the (between 2 and 16) white fields in a modified checkerboard with black checkers. The same normally (4) and partially (1) sighted subjects, familiar with other free-viewing tasks in SPV participated in all 3 tasks. Tasks were learned in free view first, and repeated until performance reached a plateau. They were then repeated in gaze-locked condition until the subject either reached a plateau or improved over free-viewing performance by 30%.
For each task a free-viewing plateau was reached after 1-6 one-hour sessions, with the highest number of sessions required for the virtual wayfinding and checkers tasks, and some subjects requiring up to twice as many sessions as others. Gaze-locked task learning required 2-3 times as many sessions to reach criterion performance. For the wayfinding and maze tracing tasks, plateau performance was statistically indistinguishable in both conditions, for all subjects; in the checkers task the offset time, i.e., the time to place the first checker, also reached the same plateau in both conditions, but all subjects improved performance speed for additional checkers, in most cases by at least 30% after 8-17 gaze-locked training sessions.
It appears that gaze-locking is not an impediment to task performance in SPV and that under certain conditions it may even allow subjects to exceed previously achieved practice levels for motor task performance as they learning to master the gaze-locked viewing condition. Looking ahead, these findings suggest that performance of visual prosthesis recipients with a head-mounted camera will not be limited by loss of eye movements as a means of visual target inspection.
This PDF is available to Subscribers Only