This proposal is part of an ongoing e?ort to restore sight to the blind with a thalamic visual prosthesis. Such devices use microelectrodes placed in the lateral geniculate nucleus of the thalamus (LGN), part of the brain that relays signals from the eye to the rest of the visual system, in order to generate visual images. Each microelectrode creates one pixel-like element of an arti?cial image that is called a phosphene. While phosphene location is predictable, it is not known how their size or color are in?uenced by electrode placement or strength of electrical stimulation for thalamic devices. Some work has been done in retina with humans, and in cortex with both humans and animal models, but we lack a rigorous study for LGN. The LGN architecture, with macroscopic segregation of magno-, parvo-, and koniocellular pathways not found elsewhere, suggests that observations made in other systems may not be applicable. We propose investigating LGN phosphenes to measure size, hue (color), and persistence, and then to unify this knowledge by using phosphenes to present simple visual objects. The work will be done in a behaving non-human primate model, where behavioral reports from simple tasks will convey the animal's perception of phosphenes and synthetic letters. Match-to-sample tasks will be used for size and hue. A memory saccade task will be used for persitence. A match-to-sample task will be used for letter identity. In the match-to-sample tasks, the animals are brie?y shown a cue on a computer screen, and then a choice of targets that vary in size, hue, or letter identity. They must look to the target that matched the cue in order to receive a reward. In the memory saccade task, the animal is shown two targets, and after a delay must look to the one that was illuminated longer. Tasks are presented in blocks of interleaved trials with occasional probes with cues presented through microstimulation rather than on the screen. Responses in probe trials provide insight into animal perceptions when compared with responses to control trials. By the end, we will know how phosphene size, hue, and persistence vary with electrode placement and with strength of microstimulation. We will also have demonstrated an LGN prosthesis conveying simple visual scenes. The knowledge will be critical to development of a device for use in blind humans.
Blindness is faced by millions of people due to diseases like glaucoma, retinitis pigmentosa and macular degeneration, or trauma from accidents or military combat. A visual prosthesis device to restore sight could be made by sending the output of a digital camera to the visual part of the brain. This research will try to understand how to control the pixels in a direct-to-brain display in order to make a visual prosthesis.
