November 18, 2010 – Original source: New Scientist
(a) is the original image presented to the mice, (b) shows the image actually seen by the mice using the prosthetic retina system, and (c) is what the mice would have seen using existing retinal implant technology (Image: Sheila Nirenberg)
A prosthetic device capable of deciphering the code that allows the brain to make sense of visual stimuli has allowed blind mice to see clearly.
A human version could one day give back vision to the millions of people who have lost their sight through retinal degeneration.
In a healthy eye, light enters the retina and activates light-sensitive photoreceptors which process and send information to ganglion cells that in turn transmit the message on to the brain as a pattern of electrical activity.
People with retinal degeneration lose the use of their photoreceptor cells. As a result their retinas can no longer pick up images, so no visual information gets sent to the ganglion cells, or to the brain.
Prosthetic retinas usually work by stimulating the ganglion cells directly.
However, they don’t use the patterns, or “code” that the ganglion cells normally use when communicating with the brain, and so are only capable of restoring a person’s ability to see simple things like spots of light and edges.
Sheila Nirenberg at Weill Cornell Medical College in New York and colleagues compared the signals going into healthy mice retinas with the output from ganglion cells, and produced an algorithm to recreate this pattern.
Nirenberg’s team then used gene therapy to transfer the protein channelrhodopsin into the ganglion cells of blind mice, in order to make them responsive to light.
To mimic the behaviour of the photoreceptors, they used a pair of glasses with a tiny embedded video camera, computer processor and Nirenberg’s algorithm to encode images into rapid pulses of light which are shone into the eye using LEDs.
This drives the channelrhodopsin-treated ganglion cells to send signals to the brain that it can interpret in the normal way.
To test the device, Nirenberg measured the ganglion cell output when an image of a baby’s face was presented to blind mice.
Patterns from mice with the new prosthetic system resembled the output from sighted mice, whereas those from devices that lacked the code presented a more pixelated version of the face.
This prosthetic also stimulates tens of thousands of ganglion cells that express channelrhodopsin, compared to current prosthetics which use electrodes to stimulate only a handful of cells.
Mimic real vision
“We think this system gives enough information to reconstruct faces, newsprint, landscapes, essentially anything,” says Nirenberg.
The group is about to start tests on primates in the hope that human clinical trials can soon follow. They are collaborating with Bill Hauswirth, a geneticist from the University of Florida who has developed a safe and effective way of delivering genes into primates.
They also hope to team up with Second Sight, a retinal prosthesis company based in California.
The research was presented at this week’s Society for Neuroscience meeting in San Diego.