Science: Triumph of mind over matter

Now that science has enabled a man to control a computer by thought alone, where do we go from here?

Since the first computers were developed in the 1940s, the trend has been more powerful, cheaper and smaller machines. Computers which once took up whole rooms can now fit in a jacket pocket. Advances in microprocessors and better engineering techniques mean that faster and cheaper computers continue their rapid development. Computers any smaller, though, are harder to achieve.

This is not because scientists are unable to shrink the electronics, but because of the way users communicate with the machine. To work with a human, a computer needs a screen and some kind of mouse, keyboard or pen for putting in and manipulating information. These devices take up space. A computer which could take signals from the brain - rather than from brain, to hand, to keyboard, and to chip - has to be truly miniature. It also has to be far faster and easier to use. Mental, not physical, dexterity would be all that is required.

That sort of computer might not be fantasy. Researchers in the United States claim to have developed a device which lets severely disabled people control a computer cursor by thought alone. The researchers, led by Professor Roy Bakay, of the department of neurology at Emory University in Atlanta, have developed tiny implants that are put in patients' brains. The implants, small glass cones with a miniature electrode, are placed inside the motor cortex, the area of the brain which governs movement. In time, the patient's own nerves grow inside the cones, encouraged by chemicals the team extract from the knees.

Once the nerves have grown, they connect to the electrodes inside the cones, allowing the computer to detect brain signals via a small transmitter located just inside the patient's skull. The American team implanted the cones in two patients, one a 57-year-old stroke victim and the other, a woman suffering from a degenerative neurological disease. The woman has since died from her condition, but the man is now able to use the system to control a computer cursor to pick phrases on a screen, and communicate with the outside world.

The patient had to learn how to control the cursor by thinking about moving a part of his body. At the moment, the computer is limited to only very simple commands - the cones can only detect thoughts to move the cursor up and down, left or right. Even so, the benefits to a man who is almost totally paralysed and had no other way of communicating, cannot be underestimated.

The Emory research is one of a growing number of projects which are looking for alternative ways for humans to communicate with computers. Much of it, like Professor Bakay's project, is aimed at helping the disabled. Other, better known applications, such as handwriting or speech recognition, are available on personal computers and other electronic devices already.

"What it means is that you have another channel to communicate with computers," explains Dr Peter Robinson, a scientist who researches the way computer applications can help disabled people, at Cambridge University's computing laboratory. The problem is that the channel has a very limited "bandwidth": it only carries small amounts of information. "The challenge is to make the best use of this limited bandwidth," he says. The capacity of implants in the brain, though, is far greater than a more common solution to the problem: electrodes attached to the skin. "If you can go down to the cortex, the signals will be 100,000 to a million times stronger than signals you can pick up on someone's head," says Dr Stephen Roberts, at the department of electrical and electronic engineering at Imperial College in London.

Most research into controlling computers uses non-invasive techniques such as pads on a person's skull or on an arm or leg. Computer scientists and engineers are simply not permitted to work inside the human body. There are ethical and medical questions surrounding implants which mean it is unlikely that they would be allowed, let alone acceptable, for able- bodied people.

Instead, the Imperial researchers are working with mathematics to decode poor quality signals. "The other way is to tease information out of tiny signals," says Dr Roberts. "We work with dirty signals, but we are doing advanced computer analysis." He believes that joining his research on signals with the American work on implants could bring even greater progress. "The future lies in coupling the two together," he says. "At the moment, the kind of performance they [the American team] are getting is about the same as we get with electrodes on people's heads." The medical applications for computers in the body are extensive. As well as controlling computers, researchers have used computer signals to bridge damage to patients' spinal cords; they cannot walk, but they do regain some sensation. Researchers at Dundee University are working on computers which recognise gestures and sign language as well as speech. Other scientists have developed electronic cornea implants, and computer-assisted hearing. Even a decade ago, these advances would have sounded as far-fetched - and unpalatable - as brain implants do today.

Computing experts believe that demand from the public will eventually overcome squeamishness about integrating electronics within the human body. "It would be foolish indeed to say we would never get full interactivity through an implanted chip," explains Professor Peter Cochrane, head of BT's research labs at Martlesham in Suffolk. "We will accept this technology by default, because it is attractive."

Uses for PCs in or on the body extend from business to entertainment and defence. At Imperial, Dr Roberts points to devices which control computers through the electrical conductivity of the skin. This changes depending on whether we are tense or relaxed - the theory behind the lie detector. According to Dr Roberts, games manufacturers are interested in bio-feedback, as are the military, who could use it to allow pilots to control aircraft, or to fire weapons, under extreme "G" force conditions.

More peacefully, the computing giant IBM has developed a technology known as the personal area network (PAN), as well as a wearable computer. The PAN uses the human body to conduct signals between computers. This means that we could exchange details with each other with a simple handshake: a true electronic business card.

Smart phones could be built which recognise the user and can download their phone book into the telephone's memory, or use computer code for security when making calls to, say, a bank. In time, linking electronics to the body will change the way we think about computers. "The reason the wearable PC is attracting interest is because we want the advantages of the technology but don't want to do things differently," explains Vincent Smith, who is programmes manager at IBM's personal systems group in the UK. "We want information to be available in a natural and intuitive way."