Science & Technology

Advanced computer models enable improvements in “bionic eye” technology

Researchers have developed an advanced, experimentally validated computer model that reproduces the shape and location of millions of nerve cells in the eye, as well as the physical and network characteristics associated with them. Researchers have focused on models of nerve cells that transmit visual information from the eye to the brain, and have identified ways to increase clarity and give color vision to future artificial retinal devices.

Researcher at Keck School of Medicine USC Develop signals that can bring color vision and improved transparency to prostheses for the visually impaired.

Millions of people are facing vision loss due to degenerative eye disease. Retinitis pigmentosa hereditary disease alone affects 1 in 4,000 people worldwide.

Today, there are technologies available to provide partial vision to people with the syndrome. Argus II, the world’s first artificial retina, reproduces some of the functions of the eye that are essential to vision, allowing the user to perceive movement and shape.

While the field of artificial vision is still in its infancy, for hundreds of users around the world, “Bionic Eye” enriches the way we interact with the world on a daily basis. For example, looking at the contours of an object allows you to safely navigate an unfamiliar environment.

That’s just the beginning. Researchers are calling for future improvements in technology with ambitious goals in mind.

Genre mosquitoes

Gianluca Lazzi, PhD. Credit: Gianluca Lazzi

“Our current goal is to develop a system that truly mimics the complexity of the retina,” said Gianluca Lazzi, a professor of ophthalmology and electrical engineering at the Keck School of Medicine at USC and the Department of Ophthalmology and Electrical Engineering at USC Bitervi Engineering. I will.

He and his USC colleagues have made progress in two recent studies using advanced computer models of what is happening in the retina. Their experimentally validated models reproduce the shape and location of millions of nerve cells in the eye, as well as their associated physical and network properties.

“We can now model what we couldn’t even see before,” said Lazzi, professor of engineering Fred H. Cole and director of the USC Institute for Technology and Medical Systems. “We can mimic the behavior of the nervous system, so we can really understand why it does.”

Researchers have focused on models of nerve cells that transmit visual information from the eye to the brain, and have identified ways to increase clarity and give color vision to future artificial retinal devices.

Eyes, bionics, etc.

To understand how computer models can improve the eye of bioengineering, it is helpful to know a little about how vision occurs and how artificial vision works.

When the light enters a healthy eye, the lens focuses it on the retina behind the eye. Cells called photoreceptors convert light into electrical impulses that are processed by other cells in the retina. After processing, the signal is passed to the ganglion cells. Ganglion cells transmit information from the retina to the brain through a long tail called an axon that is bundled to make up the optic nerve.

Photoreceptors and processing cells die of degenerative eye disease. Retinal ganglion cells usually continue to function longer. Argus II signals these cells directly.

“In these unfortunate situations, there is no longer a proper set of inputs to the ganglion cells,” Razzi said. “As engineers, we ask how we can provide that electrical input.”

The patient receives a small eye implant with an array of electrodes. These electrodes are activated remotely when a signal is sent from special eyeglasses with a camera. The pattern of light detected by the camera determines which retinal ganglion cells are activated by the electrodes and sends a signal to the brain that results in the perception of a black-and-white image containing 60 dots.

Computer models bring new advances

Under certain conditions, the electrodes in the implant accidentally stimulate the axons of cells adjacent to the target. For users of bionic eyes, this off-target stimulation of axons provides the perception of elongated shapes instead of dots.In a study published in IEEE Transactions for Neural Systems and Rehabilitation Engineering, Lazzi and his colleagues have deployed a computer model to address this issue.

“You want to activate this cell, but you don’t want to activate adjacent axons,” Razzi said. “Therefore, we tried to design an electrical stimulation waveform that more accurately targets cells.”

Researchers used models of two subtypes of retinal ganglion cells, at the single-cell level and in large networks. They identified a pattern of short pulses that preferentially targeted the cell body with less off-target activation of axons.

Another recent study in the journal Science report We investigated how to apply the same computer modeling system to the same two cell subtypes to encode colors.

This study is based on a previous study that showed that people using Argus II perceive a change in color with changes in the frequency of an electrical signal (the number of times the signal is repeated in a particular time period). .. Using this model, Lazzi and his colleagues have developed a strategy for adjusting the frequency of the signal to create a blue perception.

Beyond the possibility of adding color vision to bionic eyes, hue encoding can be combined with artificial intelligence in future system-based advances, so the most important elements around a person, such as the face and doorways, stand out. I will.

“There’s a long way to go, but we’re heading in the right direction,” Razzi said. “We can give intelligence to these prostheses, and knowledge is powerful.”


“Color and Cell Selectivity of Retinal Ganglion Cell Subtypes by Frequency Modulation of Electrical Stimulation”, Javad Paknahad, Kyle Loizos, Lan Yue, Mark S. Humayun, Gianluca Lazzi, March 4, 2021 Science report..
DOI: 10.1038 / s41598-021-84437-w

Javad Paknahad, Kyle Loizos, Mark Humayun, Gianluca Lazzi, Targeted Stimulation of Retinal Ganglion Cells in Supraretinal Prostheses: Multiscale Computational Studies, September 29, 2020, IEEE Transactions for Neural Systems and Rehabilitation Engineering..
DOI: 10.1109 / TNSRE.2020.3027560

About research

Both studies were conducted by the same USC research team. The first authors of both are Javad Paknahad, a graduate student in electrical engineering. The other author is Dr. Mark Humayun, co-inventor of Kyle Loizos and the Argus II retinal prosthesis.

The· Science report This study was supported by the National Science Foundation (1833288), the National Institutes of Health (R21EY028744, U01EB025830), and studies to prevent blindness.

Disclosure: Mark Humayun, MD, PhD are co-inventors of the Argus implant series and are paid royalties.

Advanced computer models enable improvements in “bionic eye” technology Advanced computer models enable improvements in “bionic eye” technology

Back to top button