In some forms of blindness, including age-related macular degeneration, most of the eye is perfectly fine. The cells of the retina that convert light into electrical pulses may die off, but the cells that support them, including the nerves that process these signals and relay them to the brain, are still intact. This raises the prospect that the eye's infrastructure can be used to help restore vision. Stimulate the remaining neurons in response to light, and they'll happily take the signals and feed them into the visual centers of the brain.
There are a number of ways of going about this, but one of the more promising is some form of retinal implant. These devices can take incoming light, convert it into an electrical signal, and feed that directly into the neurons within the retina. The problem right now is that these things are bulky and complex, requiring wires, external power sources, and the like. In this week's Nature Photonics, researchers report on a novel method to get rid of some of the complexity: implant a photovoltaic device directly into the retina.
This doesn't entirely eliminate the complexity, but it makes the most important parts—the ones that reside inside the retina—significantly simpler. It also eliminates their need for an external power source. The idea is appealingly easy. Photovoltaic devices work by converting light energy into free electrons, producing a current. By injecting that current into the appropriate layer of the eye, it's possible to use it to stimulate the nerves that were normally receiving signals from the retina's light-receptive cells.
In practical terms, the authors' device involves an array of silicon photodiodes etched in a flexible silicon substrate. When struck by light, these will produce a small voltage difference. They're layered atop a set of iridium oxide electrodes that extend into the neural layers that relay signals to the brain. When the array is hit by light, the charge that's generated stimulates these neurons, and the brain receives a signal letting it know.
Because they're embedded into the retina itself, these sensors receive some of the benefits we normally associate with the eye, like our ability to turn the eye to focus on specific objects. Over time, the nerve cells also grow in around the electrodes, increasing the contact and lowering the amount of charge required to generate a signal.
But the biggest bonus is that there's no external power supply required. Once the implants are in place, they'll continue to produce charge using nothing more than the energy provided by the incoming light.
That's the good part. The downside is that ambient light isn't sufficient to generate much of a current. In fact, it's too dim "by a factor of at least 1,000," according to the authors. So, is this good for anything?
The authors suggest there is a way to get the system to work. The photodiodes can be stimulated by an infrared laser, and they were able to show that the laser could trigger nerve activity in rats fitted with one of the devices. The nerves didn't require that much energy before they fired, either: the authors estimate that the laser caused some small amount of local heating, but nowhere near to the levels that would be considered unsafe for either instantaneous or chronic exposure.
They envision that the laser would be embedded in a set of goggles that would register the outside world, and hand off the images to a portable computer. That would convert the image to an appropriate series of laser pulses, targeted at different areas of the photovoltaic array. They envision using a single laser and a series of mirrors, much like those used in Texas Instruments DLP display technology. Most other retinal implant systems require some form of external camera system, though, so this isn't a weakness compared to competing approaches.
(What I do suspect is that it undercuts the authors' argument that being able to turn the eye towards an object of interest is an advantage, as the direction of the goggles would seem to be the key.)
This also isn't going to restore high-resolution color vision any time soon. The best most current systems can achieve is somewhere around 20/1,200 vision.
Even with all the limits of the system, there are a couple of things to like about it. One is that the parts that actually go into the eye are simple and self-contained, and can just be left alone once they're implanted. The second is that the complicated stuff—the sensor, laser, and portable computer system—are all external, and can be swapped out at will. This means most conceivable upgrades to the system can be as simple as a hardware swap that may actually be easier than upgrading a typical hard drive.
Nature Photonics, 2012. DOI: 10.1038/NPHOTON.2012.104 (About DOIs).
reader comments
18