A computational model to design wide field-of-view optic nerve neuroprostheses

Abstract

Neuroprostheses based on retinal stimulation (RS) allows many individuals affected by retinal degeneration to partially restore visual perception but the produced phosphenes are confined into a narrow region of the visual field. Optic nerve stimulation (ONS) has the potential to produce visual perceptions spanning the whole visual field, but its exploitation is challenging since it produces very elongated phosphenes that cannot be easily organized into meaningful percepts. Here, to address this issue, we introduced a geometrical model that allows us to convert firing rate patterns in the retina and optic nerve into visual perceptions and vice versa. Then, we developed and extensively characterized a method to estimate the best perceptions that can be elicited through a given electrode configuration. This method was used to qualitatively compare ONS and RS also using a set of static and dynamic visual scenes through simulated prosthetic vision (SPV) experiments with healthy subjects. Both simulations and SPV experiments showed that it might be possible to reconstruct natural visual scenes using reasonable amounts of active sites, whose arrangement in the optic nerve section exploits purely geometrical factors. The ability of ONS to cover the whole visual field, allowed perception of much more detail in dynamic scenarios than what is possible with RS, where the narrowing of the visual field results in a limited ability to visualize the scene. Our findings suggest that ONS could represent an interesting approach for vision restoration and that our model can be used to optimize it.