Seeing is quite a complicated business for Marie. A video camera attached to a plastic cap on her head broadcasts the images it records as radio waves, which are picked up by a "stimulator" that has been implanted in a small depression carved from the inside of her skull. A wire snakes from the stimulator, around the outside of her brain, to the back of Marie's right eye where it is clamped on to her optic nerve. Once the camera is running, the stimulator turns the radio signals into electrical impulses, which then flow along the cable, up the optic nerve and into the vision centre at the back of her brain.

By moving her head slowly, she can make the camera scan a big white board that is embedded with live pixels, rather like a stadium scoreboard. This way, images of letters or simple shapes, made up of the flashing pixels, can be "seen". Marie, not her real name, is one of a handful of astoundingly brave and determined guinea pigs who are attempting to translate the long-standing dream of bionically enhanced humans into reality.

After decades of sci-fi style predictions about the potential benefits to medicine from robots and microchips, there are signs that the long-trailed "miracle" treatments for certain conditions may finally be coming close to realisation. Last week, Kevin Warwick, a professor in cybernetics from the university of Reading, underwent a two-hour operation to have a chip attached to nerves in his left arm. "We should harness the best of machine intelligence for ourselves," he said. "We should build it into our own bodies." But while Warwick, who hopes his experiment will allow him eventually to communicate with his wife through thought alone, admits his goal is still some way off, other applications of "bionic" technology, with direct therapeutic benefits, may be considerably closer.

Marie is blind as a result of developing retinitis pigmentosa, which has destroyed the rods and cone cells in her retina. The elaborate set-up that allows her to make out a few shapes in the lab is known as Microsystem-based Visual Prosthesis (MiViP) and it's one of the projects described in a book published last month by Time reporter James Geary. "This is not true vision," stresses professor Claude Veraart, the researcher leading the project at the University of Louvaine in Brussels. "And it's definitely not a cure for blindness, but it is something to help people cope better with their impairment."

We are obviously a long way here from the seamlessly integrated, performance-enhancing implants of Blade Runner, yet our expectations are constantly being raised by over-optimistic predictions. In 1999 the blind singer Stevie Wonder commented that he hoped one day to be able to see again thanks to an artificial retina that was being developed. The story became big news and an ophthalmologist interviewed by ABC's Barbara Walters agreed that something like that might well be available in two to three years time - ie, around now.

Exactly how close we are to some sort of pacemaker for the eye was set out in Science magazine in February, which reported that at least a dozen groups are now trying to restore vision by stimulating the brain, the optic nerve or the retina. About 50% of all cases of blindness are caused by damage to the retina, so researchers at University Eye Hospital in Tubingen, Germany, have been experimenting with implanting a thin plate with thousands of light sensitive microphotodiodes directly into a damaged retina, which can then send information back up into the brain. (A normally functioning retina has about 130 million photoreceptors.)

Another approach is to implant a chip connected to an outside camera, possibly located in a plastic lens that replaces the natural lens of the eye, directly into the cells which feed into the optic nerve. So far the only miracles involved here have been ones of miniaturisation. Patients who have had retinal implants inserted experimentally just for a few hours have reported sensations of light patterns and a couple have been able to identify simple geometric patterns. "It proves," the researchers conclude optimistically, "the feasibility of generating perception of light patterns in blind people."

Besides allowing the blind to see, the other resonant goal of this research is the aim of making the lame walk. This is an area where another celebrity, paralysed "superman" Christopher Reeve, has been instrumental in keeping hopes of a breakthrough high for the 10,000 or so Americans who suffer spinal cord injuries every year. He's been quoted as saying that he'll be walking by 2005, but if the latest development is anything to go by an awful lot of effort will also be involved.

Last month the University of Arkansas announced that a quadraplegic who had been confined to a wheelchair, was now able to walk about 1,000ft "with only moderate fatigue". The key to his transformation was a set of implanted electrodes that delivered low-frequency electrical pulses to the man's spine which then stimulated walking movements. But before the implant he had trained for four months, suspended over a moving treadmill, which passively moved his leg. The implant he was given is known as Functional Electrical Stimulation (FES) and it has a number of applications.

It's at the heart of one of the few commercial bionic systems, known as Freehand, currently being used by Brian Holgerson, a Danish tetraplegic with limited movement in his left arm and shoulder. Developed by the Cleveland Functional Electrical Stimulation Centre in the US, Freehand allows Holgerson to pick up and release objects. Eight electrodes about the size of small coins have been implanted into the muscles of his forearm and these are linked by ultra-thin wires to an FES implant in his chest. This in turn is connected to another implant - a "position-sensing unit" - attached to his right shoulder, which he can still voluntarily control.

Now when he wants to pick something up, he makes an upward movement with his right shoulder and this triggers the position-sensing unit, which activates the stimulator, which sends a current to the muscles that control the grasping movement. A downwards movement tells his hand to let go. Although strange to use at first, he says it now feels very natural and he has regained the ability to hold a cup or a pen and lift a fork.

Rather than having to learn arbitrary muscle movements, however, one of the goals of this research is to link it in with commands from the brain. The idea is quite simple. Whenever we plan or carry out movements, large groups of neurons in the brain produce patterns of electrical impulses. Analyse those patterns to work out which ones go with what activity and you could stimulate the FES used in a system such as Freehand directly. This is already being tried out in the lab on a quadriplegic called Jim Jatich, from Akron, Ohio, who was injured in a diving accident 20 years ago. He wears something that looks like a shower cap that is dotted with electrodes which are sensitive to brain waves. Wires lead from the electrodes to a computer that converts the brain signals into electrical signals that then activate the FES implant.

After extensive bio-feedback training, which taught Jatich to regulate the beta-rhythm in his brain, he was able to move a cursor up or down on a computer screen just by thinking about it. Then the computer was linked to his implanted FES unit and by thinking about moving the cursor up, he opened his hand. By thinking "down", his hand closed. Now he can pick up a drinking glass or a fork by thought alone.

Unravelling the details of exactly what goes on when the brain is thinking of an action has only been experimented on animals to date. In January, researchers at the California Institute of Technology announced that they had trained a monkey to move a cursor on the screen just by having the desire to do it. Having identified with high-tech brain scans which brain regions were active when it was simply planning to move, they then implanted electrodes in that clump of cells, known as the posterior parietal cortex, and recorded from it while he was playing a simple touch-screen computer game. This allowed them to work out, with 90% accuracy, which pattern of neural firing appeared when he was planning to reach in a particular direction.

The implications for life-enhancing devices are obvious. But not everyone is impressed. Some disabled campaigners believe the whole field is over-hyped and that the facilities and the money would be far better spent producing more tangible benefits right now. As one commented: "A wheelchair is actually a rather good way of getting around, but what I'd like is a good way of controlling my bladder."

· Jerome Burne is editor of the monthly newsletter Medicine Today (www.medicine-today.co.uk). James Geary's book, The Body Electric: An Anatomy of the New Bionic Senses, is published by Weidenfeld and Nicolson