The long-term aim of the project is to devise a system whereby a video camera fitted to a pair of spectacles would deliver electrical stimulation directly to an array of micro-electrodes permanently embedded in the visual cortex of the patient's brain, so giving the blind patient an ability to perceive light.
"Our visual world consists of images rather than discrete points of light," explains Professor Richard Normann, head of the department of bio-engineering at the University of Utah. "We have to find out whether patterned stimulation via a large number of individual electrodes will give rise to a complex image rather than just a diffuse blob of light." The answer to this question may soon be provided by a 100-electrode array constructed by Professor Normann's group, and currently under test.
But would the very coarse degree of vision provided by such technology actually be useful? Other members of the University of Utah team devising a model of the likely visual field of an implanted patient have used data on the way in which a retinal image is mapped on to the visual cortex. The result is about what you would see if you made a hole in a sheet of paper with the point of a ball-point pen, held it up close and looked through it. Researchers were able to establish that this rather small visual field would be sufficient to permit reading rates of up to 170 words per minute, compared with a normal reading rate of about 250 words per minute for the average fully sighted reader.
Professor John Cronly-Dillon of Umist is optimistic but cautious about the possibilities of this technology: "While we have a fair understanding of the nervous system, and a technology which is conceptually quite straightforward, it could actually be very difficult to make the whole thing compatible with the intricate systems that the brain uses to process visual information."
The problem is that sight is not simply a matter of seeing a large number of points of light all joined together. This is because a tremendous amount of "image processing" takes place to enable us to recognise structures, shapes, relative sizes and movement. Also, by stimulating brain cells directly, we may be throwing away a lot of information that is vital for useful vision.
A parallel approach to the problems of prosthetic vision is that taken by a multi-disciplinary team headed by Dr Joseph Rizzo, a neuro-ophthalmologist at the Massachusetts Eye & Ear Infirmary (MEEI) and Professor John Wyatt, Professor of Electrical Engineering at the Massachusetts Institute of Technology (MIT). Their goal is to develop a microchip that can be attached not to the brain but to the surface of the retina.
The thinking behind this approach is that common diseases of the retina which cause blindness involve only the rods and cones, while other cells that connect the eye to the brain remain intact.
Stimulating the cells of the optic nerve directly using a chip implanted on to the retina may make it easier to harness the brain's image-processing functions and avoids the requirement for brain surgery that placement of a cortical chip would involve.
Quite apart from the issues of image-processing and biocompatibility common to all such projects, the MEEI/MIT approach involves the added difficulty of attempting to implant a device into the eye. There is the sheer delicacy of the retina (described by one MIT researcher as having the consistency of wet tissue paper). In addition, the eye is capable of swift movement and acceleration, such that an implant of any sizeable mass would be prone to being dislodged and to causing retinal damage while doing so. The MEEI/MIT team has devised an implant roughly 0.5mm in thickness which can be constructed to exactly match the curvature of the subject's retina.
The MEEI/MIT scenario proposes the use of a tiny video camera mounted on spectacle frames to record images which are then digitally encoded and transmitted into the eye by a low-power laser also mounted on the spectacle frames. This low-power laser beam (which is invisible to the naked eye) not only transmits data, but also powers the microchip, so avoiding any need for a physical linkage between the implant and the external environment.
The microchip decodes the data fed to the implant and translates it back into a pattern of electrical impulses which directly stimulate the array of electrodesoverlying the retina.
But how far has this technology actually progressed? In animal experiments, Dr Rizzo has now developed surgical techniques for safely affixing the implant to the retina, and has just completed a year-long study on the biocompatibility of implant materials in the eye of animals.
The team now expects to implant a fully functional device within an animal eye within the next 12 months, stimulate it with a laser, and determine if the resulting signals can be detected in the visual cortex.
The two technologies described here are by no means mutually exclusive, since the MEEI/MIT retinal implant is only feasible in those patients whose ganglion cells (which connect eye to brain) remain healthy and whose ocular physiology is otherwise normal, while the Utah/Umist approach could assist those in whom there has been physical damage to the eye and the optic nerve itself.
"We are certainly not home free with artificial vision," Utah's Professor Normann said recently, "but the technologies we have developed will begin to allow us to ask whether it could possibly become practical."
The MEEI/MIT team are equally cautious, but perhaps more upbeat. "Our next major step is to develop a working implant, verify its operation and perfect its design in a series of animal experiments. Timing and progress depends very much on funding and experimental uncertainties, but the ultimate goal is to reach the stage of experimental testing in human volunteers," said Dr Rizzo, who has a healthy respect for those who are sceptical about the project. "It's up to us to show that the device works - we very literally have to make the sceptics see the light."Reuse content