A MICROCHIP that might one day restore vision to people who suffer from certain diseases of the retina is one of the most extraordinary examples of how scientists are combining silicon and neural information processing.
A few weeks ago, scientists completed animal experiments that offered encouragement that this may one day be possible in the Retinal Implant Project, a collaboration between the Massachusetts Eye and Ear Infirmary, Massachusetts Institute of Technology, and Harvard Medical School.
By the turn of the century, they hope to carry out the first tests of a device, a microchip surgically implanted in the eye that works in conjunction with a miniature camera and laser mounted on a pair of glasses.
The glasses capture visual images with an electronic analogue to film known as a charge-coupled device. This converts a scene into digital form, so it can be sent by infra-red laser to the implanted microchip. The microchip is designed to then pass this information, in the form of electrical impulses, to the brain via an array of electrodes in the retina.
The device would be a partial substitute retina, a prosthesis for people who have lost the eye's light receptors - the retina 's rod and cone cells - through important causes of blindness such as retinitis pigmentosa and age-related macular degeneration.
Led by Prof John Wyatt and Dr Joseph Rizzo, the team recently finished a series of trials of a prototype implanted in the eye of a rabbit. The test devices delivered tiny electrical currents to ganglion cells - the nerve bodies on the inside surface of the retina that feed into the optic nerve - and produced measurable activity in the animals' brains.
Other work by the team has shown how the microchip can be powered by the laser, and show the biocompatibility of the implant materials. However, "We're facing a number of very significant obstacles, one after the other," admitted Dr Rizzo.
Researchers still have to determine the health of ganglion cells in the eyes of the blind, though a team at Johns Hopkins University Hospital in Baltimore has evidence that they can function in about 70 per cent of their normal numbers in some patients.
The scientists also need to develop an implantable chip that is biocompatible and relies on electrodes that do not tear or poison the delicate retina, which has the mechanical properties of wet tissue paper.
Later this year the group hopes to show that the prototype can be left inside a rabbit eye safety. Finally, they must face the biggest problem of all: getting the chip to package visual information in a form the brain can use.
