Optimising phosphene vision using a realistic phosphene vision simulator

Keywords

Loading...
Thumbnail Image

Issue Date

2022-08-25

Language

en

Document type

Journal Title

Journal ISSN

Volume Title

Publisher

Title

ISSN

Volume

Issue

Startpage

Endpage

DOI

Abstract

In a world where people largely navigate by sight, blindness presents a signi cant obstacle in the lives of millions of people. Visual cortical implants are an emerging neuroprosthetics technology, which aims to provide a rudimentary level of vision to blind people. The technology does this by making use of 'phosphenes', small dots of light that are perceived without light entering the eye. It makes use of the retinotopic organization of the visual system to provide meaningful phosphene perception through localized electrical stimulation of the visual cortex via micro-electrodes inserted into the brain tissue. As implementing this neuroprosthetic technology in the visual cortex involves an invasive surgical procedure, many experiments regarding phosphene vision are performed using simulations of phosphene vision and sighted subjects. These experiments aim to solve problems such as comparing image processing strategies to translate a real world image to a phosphene perception. However, the simulation strategies used for these experiments are often highly simpli ed, which limits the validity of the results of the experiments when making statements about real phosphene vision. In this study, we rst analyse the shortcomings of previously used phosphene simulation methods. Next, we have created a more biologically realistic simulator, implementing knowledge of phosphene perception and interactions that arise from various simulation parameters. Using this simulator, it will be possible to test new algorithms for phosphene representation in a way that is closer to the expected percept generated by real visual cortical neuroprosthetics. When compared to a simpli ed method of phosphene simulation, this study nds that in high phosphene resolution conditions (32x32), the more realistic method of phosphene simulation allows for a higher performance in an object detection task. In a low resolution (8x8), the simpli ed and complex conditions perform similarly. This con rms the method of phosphene simulation could a ect the conclusion of an experiment investigating di erent phosphene resolutions for phosphene vision. Therefore, carefully choosing a method of phosphene simulation is an important step when investigating phosphene vision using simulation.

Description

Citation

Faculty

Faculteit der Sociale Wetenschappen