Project

Back to overview

I-See - Improving intracortical visual prostheses using complex coding and spontaneous activation states

English title I-See - Improving intracortical visual prostheses using complex coding and spontaneous activation states
Applicant Herzog Michael
Number 198552
Funding scheme ERA-NET NEURON
Research institution EPFL SV BMI LPSY SV 2807 (Bâtiment SV)
Institution of higher education EPF Lausanne - EPFL
Main discipline Neurophysiology and Brain Research
Start/End 01.01.2022 - 31.12.2024
Approved amount 365'034.00
Show all

Keywords (1)

Neuroprosthetics

Lay Summary (German)

Lead
Michael Herzog
Lay summary
Viele Blinde können nicht von retinalen Prothesen profitieren, da ihre Defizite kortikaler Natur sind. Hier können kortikale Prothesen helfen. In diesem Projekt werden wir eine Kamera und deep network benützen, um Objekte aus natürlichen visuellen Szenen zu extrahieren. Wir werden dann versuchen, das Wesentliche dieser Objekte zu extrahieren, um damit höhere visuellen Areale zu stimulieren und damit direkt die Wahrnehmung dieser Objekte auszulösen. 
Direct link to Lay Summary Last update: 17.12.2020

Lay Summary (English)

Lead
Michael Herzog
Lay summary
Many blind people cannot use retinal implants because their deficits are cortical. Hence, there is a need to develop cortical implants. We will use a camera and deep neural networks, which extract objects from the visual scene. We will then study how we can best stimulate the human brain to elicit a percept of the object extracted from the deep network. One of the major questions in this project will be what is crucial for an object and how are objects represented in higher visual areas.
Direct link to Lay Summary Last update: 17.12.2020

Responsible applicant and co-applicants

Employees

Name Institute

Project partner

Abstract

For the large group of blind patients who cannot profit from a retinal implant, intracortical visual prostheses offer great promise. However, intracortical prostheses have had limited success, mainly because they require strong stimulation currents, which generate only non-specific percepts consisting of large spots of light. Here we address these limitations by exploring a fundamentally different approach: we will target complex response properties of neural populations in areas beyond primary visual cortex to generate more specific percepts and link electrical stimulation patterns in a closed-loop setup to the extensive ongoing activity in visual cortex to greatly reduce required stimulation currents. To achieve this goal we will bring together scientists from different fields and complementary experimental methods, supported by a strong backbone in computational and theoretical neuroscience.
-