Fine-Tuning Strategies for Phosphene Maps in Blind Individuals
Abstract
Focus Area: Simulation studies to assess stimulation parameters needed for an effective cortical artificial vision device. Restoring vision to the blind through electrical stimulation of the visual cortex, using implanted electrodes, has been a pursuit of researchers since the 1960s. While being a laudable goal, creating a successful cortical visual prosthesis is nearly as complex as the US endeavor to put a man on the moon, which coincidentally also started in the 1960s. For the Apollo project, it took major commitments for the development of rocket engines and a lunar lander, but it also took careful calculations of the amount of air, food, and water the astronauts would need to have on board to survive and return, and countless other details. We have landed humans on the moon; we have yet to clinically deploy a cortical visual prosthesis. Stimulation of the brain through electrodes creates scattered dots of light (also called phosphenes) that the user perceives in their visual field. A collection of phosphenes is supposed to form an image -- like a scoreboard in a ballpark. To develop a cortical prosthesis, it is not just the material of the electrodes, the type of stimulation, or the quality of the brain surgery that spell the difference between success and failure, but also the knowledge how to make an understandable image out of the phosphenes. For each electrode, the location of its phosphene is often unpredictable, so a critical condition for creating an image made up of phosphenes is to have a precise map that tells the computer, transforming the camera image to a pattern of electrical stimuli in the brain, which electrode to stimulate so that each phosphene appears in the desired location in the person’s visual field. In our laboratory, we have worked out several methods to create these phosphene maps, as well as ways to make the process as efficient as possible. These methods have only been tested in sighted individuals, simulating what prosthetic vision may look like, since there is not currently a prosthesis that can be implanted and tested in a blind human. However, we already know that the maps are not always accurate in sighted volunteers, and we suspect that they will not be any more accurate in real prosthesis wearers. Using a faulty map will cause the prosthesis wearer to see distorted images, in which straight lines may be curved or otherwise deformed, and more complex shapes may be difficult to recognize -- making an implanted prosthesis much less useful than it otherwise might be. In order to optimize the vision restoration provided by a cortical visual prosthesis, the objectives of the proposed studies are to use simulation studies in sighted individuals to: (1) Detect, in a given phosphene map, which phosphenes are displaced, by activating small sets of phosphenes intended to form a straight line or other simple geometric shape, and by asking the volunteer to indicate which phosphenes seem out of place, and in which direction they are displaced from where they should be. We will design a control box or tablet app that allows the volunteer to enter the distortion information, and to replace misfitted phosphenes by others that are nearby, in order to find the best match to form the desired shape. Our volunteers will be sighted, but all phosphene patterns will be presented in a video headset that mimics a visual prosthesis: Just like the blind person, they can only judge phosphene locations and geometric shapes relative to other phosphenes, not to an outside visual point of reference. (2) Test the control box, or app, in up to half a dozen sighted volunteers, and provide feedback regarding its ease of use, and suggestions how it can be improved. All volunteers will be given the same distorted maps, to allow comparison of their efficiency and accuracy. The outcome of this experiment will be a robust strategy for detection and elimination of distortions in phosphene maps. (3)
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Oct 29, 2018
- Source ID
- W81XWH1710621
Entities
People
- Gislin Dagnelie
Organizations
- Johns Hopkins University
- United States Army