February 25, 2021

I, Science

The science magazine of Imperial College

The first wave of retinal prosthetics can return some sensation to the eyes of blind people. But in future could we restore realistic vision ...

Miikke Tetho sits in a darkened room with a camera observing his every move. The table in front of him is covered in black felt and on it are two objects. “The one on the left looks kind of round,” says his voice from behind a pair of large, heavy-looking goggles.

“Don’t touch it yet,” says a voice from off screen. “What about the one on the right?”

“That one looks a bit like it’s curving, like this,” says Tetho, gesturing with his hand. “Actually I think it’s a banana.”

The sound of applause erupts. “Well, that’s what it looks like,” announces Tetho. He has just seen and picked up a piece of fruit – yet only a month ago he was blind.

Tetho, 46, from Finland, was wearing a retinal prosthesis constructed by German company Retina Implant AG. His results were some of the best the trial saw. Since then these bionic eyes have been gaining popularity: the ‘Argus II’, a similar implant manufactured by US company SecondSight, was given the stamp of approval by the US Food and Drug Administration earlier this year.

These implants work by mimicking the photoreceptor cells found in our retinas. In many cases of blindness these light-sensitive cells have stopped working, but the nerves in the retina that communicate with the brain remain intact. The implants effectively take over this role by sensing light and stimulating the nerve cells.

The implant is most often positioned on the back of the eyeball rather than inside it, which means the surgery required is less risky. A camera worn on the patient’s head then beams visual information to the implant and each pixel of the array responds by stimulating the nearby nerve cells. Patients fitted with the devices end up ‘seeing’ in white dots and flashes – essentially what sighted people experience when they rub their eyes. These flashes are grouped into shapes and that’s what Tetho could see when he identified the curve of the banana.

“It’s not really fair to call this ‘seeing’ though,” says Konstantin Nikolić, a researcher at Imperial’s Department of Electronic and Electrical Engineering who studies retinal prosthetics. “What they’ve achieved so far is fantastic, but we have to be honest and say: ‘that’s not proper vision’.”

Nikolić cites an example of a clinical trial conducted in 2012 where patients were fitted with implants and tested on their ability to point out a white square on a black background. 96% were better at locating the square when their implant was switched on, compared to when it was switched off. Yet when the same patients were asked to say which way a white line moved across the black background, only 57% showed any measurable improvement with the prosthesis. This sensitivity is enough to enable blind people to write their name and make out the kerb as they’re walking along a street. But Nikolić says he’s not satisfied with stopping there. He is part of a small group of scientists who want to build prosthetics that can restore more life-like vision. The situation is finally starting to look hopeful, but there have been many hurdles in their path.

For one thing, when light hits the eyes’ light-sensitive cells, the process that transforms that light into a signal is complicated. The eye is not like a digital camera where pixels simply record the light that hits them: at least three types of cell modify the signals before they are sent to the brain to be constructed into a picture. These cells don’t have linear connections to one another, but fan out in intricate patterns that aren’t easy to understand or predict.

In addition, the photoreceptor cells in the central part of the retina are so tightly-packed, that even the most advanced artificial retinas cannot match their resolution. One pixel on the electrical implant could cover hundreds of nerve cells. That means a single pixel would activate many neural pathways, resulting in a blurred picture.

There’s another reason why pictures aren’t as crisp as real life. Because most systems are situated behind the eyeball, the visual data has to be sent to the implant using wireless signals. But at the same time electrical power has to be beamed to the array. These two electrical fields can interfere with each other leading to confused information transfer.

As an engineer, Nikolić has been focusing on solutions to these practical problems, but lately he has become excited about a completely new approach to repairing vision.

The idea began in 2003, when Georg Nagel of the Max Planck Institute for Biophysics, Germany, discovered a new protein in a species of green alga. This protein, called channelrhodopsin-2 (ChR2), allows charged particles to enter the algal cells, but only if exposed to light. This discovery created huge excitement as it resembles the way that light triggers photoreceptors to send chemical signals to our nerve cells.

“In our eyes we have photoreceptor cells – which are light sensitive – and then we have ion channels, which are separate,” explains Nikolić. “But with the ChR2 you have a sensor and an ion channel all in one.”

In 2005, Nagel teamed up with colleagues from Stanford University and showed it was possible to splice ChR2 into nerve cells. When light was shone on the cells they would emit neural signals – and the response was accurate to within thousandths of a second. Now scientists are hoping that this technology, known as optogenetics, could bypass all the complex machinery of the eye and confer realistic vision. The idea would be to transfer the genes into retinal nerve cells with an injection and make the nerve themselves responsive to light.

A start-up company in Paris called GenSight, fronted by Nikolić’s colleague Botand Roska, is now experimenting with this very process. The difficulty is ensuring that the gene is transferred only to the cells where it’s needed – and making sure it stays there. It wouldn’t do for patients to end up with teeth or thumbs that fire signals to the brain in response to sunlight.

For blind people like Miikke Tetho this will be an exciting prospect. The current prosthesis models offer blind people the chance to make out the silhouette of the kerbside. But if the promise of optogenetics comes true, a single injection could be a permanent end to many types of blindness. This would be no world of white flashes; in the future blind people might see in colour.


IMAGE: attila acs