Abstract
Purpose :
Phosphenes anticipated to be evoked by an intracortical visual prosthesis would be spatially scattered throughout the visual field such that phosphene size would increase with eccentricity while density would decrease. Sighted individuals presented with simulated prosthetic vision were tested with tasks involving object recognition, eye-hand coordination and wayfinding to determine the ability of a future implant recipient to perceive phosphene-based images.
Methods :
Three subjects performed three tasks each. In the object recognition task they counted the 4 to 12 white fields in an irregular checkerboard; in the eye-hand coordination task they placed black checkers on these white fields. In the wayfinding task, they advanced through 5 different virtual mazes using a game controller. Simulated prosthetic vision was a pattern of 128 or 64 dots adjusted to each subject’s prior mapping estimations of the dot positions. Subjects were fitted with a head mounted display to present simulated prosthetic vision under gaze-locked conditions. Dot intensities were quantized to 4 or 8 grayscale levels (gsl). Three trials were conducted for each combination of gsl levels and dot pattern.
Results :
The average maze completion time and errors with 8 gsl decreased over time across sessions: from 17.5 to 2.6 mins, and from 2.7 to 0.8 (S1); 17.7 to 1.6, 2.0 to 0.3 (S2); and 17.8 to 2.1, 1.8 to 0.8 (S3). Learning was observed both within and across sessions. Task performance with 8 or 4 gsl was not significantly different. With practice, and averaged across subjects, time and error rate for 128 dots, 8 gsl, reached 69±12% and 99±35%, resp., of those for 64 dots. Subjects indicated using outlines rather than details for 64-dot navigation. In the checkerboard tests, average counting time with 64 dots was higher in S1 (p<0.18) and S2 (p<0.19) than with 128 dots. The percentage of correctly placed checkers decreased with increasing number of white fields on the board. Average counting time decreased across sessions: 57.2±15.0 s to 28.9±8.2 s (S1), 78.2±20.1 to 40.2±10.3 (S2), 65.2±21.6 to 34.6±12.4 (S3).
Conclusions :
These simulation test results raise the confidence that for a phosphene count of 128, or perhaps 64, in an intracortical prosthesis may allow the recipient to perform visual tasks. Learning effects and data from additional subjects are currently being examined.
This is an abstract that was submitted for the 2016 ARVO Annual Meeting, held in Seattle, Wash., May 1-5, 2016.