Please welcome our CCN Seminar guest speaker E.J. Chichilinsky who will be presenting, “Replicating the Neural Code: Toward a High-Fidelity Artificial Retina” in our CCN Classroom on Monday, June 27th from 1:15PM-2:30PM. Please note that the lecture will be 1 hour, and the remaining 15 minutes will be left for Q&A.
Host: Eero Simoncelli
Title: Replicating the Neural Code: Toward a High-Fidelity Artificial Retina
Abstract: Electronic interfaces to the retina provide an opportunity to leverage our extensive knowledge about the neural circuitry and function of the retina to restore or even enhance vision. However, although existing implants that stimulate retinal ganglion cells for vision restoration demonstrate proof of principle, they are not designed to reproduce the spatiotemporally precise and cell-type specific patterns of neural activity that mediate natural vision, and they produce very limited artificial vision. I will argue that to reproduce the neural code of the retina will require devices that both stimulate and record at single-cell, single-spike resolution. I will then describe our work using large-scale multi-electrode stimulation and recording from the macaque and human retina ex vivo indicating that, in some cases, such resolution is possible. I will also show cases in which it fails, and propose that we can optimize artificial vision in such cases by adapting implant function to the host neural circuitry. Finally, I will briefly introduce the Stanford Artificial Retina Project which aims to develop a retinal implant that more faithfully reproduces the neural code of the retina, and will discuss the implications for scientific investigation of the visual system and for other neural interfaces of the future.