Purchase this article with an account.
Ben McIntosh, Noelle Stiles, Mark Humayun, Armand Tanguay; Visual Prosthesis Simulation: Effects of Foveation on Visual Search. Invest. Ophthalmol. Vis. Sci. 2013;54(15):1057.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
A visual prosthesis simulator with the capacity to simulate several retinal diseases including Retinitis Pigmentosa (RP) and Age-Related Macular Degeneration (AMD), as well as to provide a visual representation of prosthetic vision, was constructed to further simulate both eye-directed and head-directed percepts using normally sighted volunteers wearing a head-mounted display with an external scene camera. Simulated head and eye pointed modes allow the effects of foveation on functional tasks with prosthetic vision to be studied in an effort to optimize the capabilities of future retinal prostheses. To quantify the difference between retinal prostheses with and without foveation ability, a visual search task was performed using the visual prosthesis simulator. The response time to find a target object was hypothesized to be a linear function of the number of items displayed (as with normal vision), and to exhibit an increased slope in the head-pointed mode as compared with the eye-pointed mode.
The visual prosthesis simulator was used to simulate blindness due to both RP and AMD with a 32 × 32 microstimulator array, presented in both eye-pointed and head-pointed camera modes. In both camera modes, the percept was stabilized in the gaze direction using an eye-tracker. A MATLAB program was used to both execute the visual search protocol including automated double-blind presentation of target and distractor objects and to record response times.
Four subjects performed the visual search task with black letters on a white screen spanning a 90° field of view while wearing the visual prosthesis simulator in both head-pointed and eye-pointed camera modes. The letter “F” was used as a target object amid letter “E” distractor objects. Response time as a function of total items on the screen was found to be approximately linear in all cases with increasing values and slopes in eye-pointed and head-pointed modes respectively.
All subjects showed a significant linear trend in response time plots. In each case, the slopes and intercepts of the linear regressions were significantly lower in the eye-pointed case than in the head-pointed case, suggesting an overall improvement in function.
This PDF is available to Subscribers Only