Imagine being able to type without touching a keyboard but by simply looking at a computer screen and thinking what you want to say.

It might sound like something out of an episode of Doctor Who but if a Maltese scientist has his way, such a device should be on the market in a matter of years.

In fact, the special keyboard already exists but only in laboratories.

Christopher James, a professor in healthcare technology, wants to make it accessible to paralysed people who have no other way of communicating.

Prof. James, who grew up in Malta and graduated in electrical engineering from the University of Malta, explained that electrodes attached to the patient's head would measure brain activity - signals that are a couple of million times less powerful than a nine-volt battery - and allow him to write just through brain power.

Speaking to The Times from his home in England, Prof. James, originally from Tarxien, said the keys on the virtual keyboard flash individually. "When the key the person is looking at flashes, there is a small recognition response in the brain, which is picked up by a computer and associated with the character that was flashing," he explained.

The system is still slow, with a first-time user taking about 40 seconds to type one character. "This is relatively slow but when you consider that these people have no other means of communication it's lightning speed," he said.

However, it is a case of practice making perfect and the time can be reduced to five seconds a character, making communication a tangible possibility.

The system, he said, operated over brain-computer interface, which is used to establish communication with people who would otherwise not be able to communicate. These would include people who cannot even blink, patients suffering from the so-called locked-in syndrome, whose brain is working but cannot communicate messages to the rest of the body.

"They can think of the conversation they want to have, so by tapping into the brain activity the message can be conveyed using a computer," Prof. James, director of the Institute of Digital Healthcare at the University of Warwick, explained.

Although brain-computer interface started developing 15 years ago, the systems are still laboratory-based rather than being used by the people who really need them. He said that while some patients worked with laboratories and were given prototypes, there were still no products on the market.

Part of the reason is cost. Prof. James explained that most laboratories used 16-electrode systems, which needed complex amplifiers costing about £8,000. Prof. James, who just moved from the University of Southampton, where he started the research, is working to develop a system that works with two electrodes, pushing the cost of the amplifier down to about £1,000.

"We want to offer a guarantee of a form of reliable communication for people who cannot communicate in any other way."

However, this is not the sort of thing that can be used for hours on end because the electrodes - that are attached to the head with a special gel - will start falling off and the person can get tired.

Prof. James estimates that the system would cost between £2,000 and £3,000, a relatively small price to pay for communication.

"If everything goes well, within two years we hope to have trialled it with real locked-in patients, got their feedback, adjusted it and made it available," he said.

The problem remains that the market for such products is relatively small, making investors reluctant to put money in them. But games using the same brain-computer interface are already on the market, which will hopefully make the more noble projects affordable.

Prof. James has just finished an experiment through which two people in different locations communicated over the internet through brain activity. "One of them was imagining making a movement and at the other end the other person interpreted the information through fast flashing lights," he said.

He also developed a wheelchair controlled through brain power, simply by imagining moving left or right. But it is not as easy as it sounds. "You cannot think of anything else or be having a conversation while controlling the wheelchair," he said, adding that the wheelchair was unable to start or stop.

However, he is confident that the system will be developed further and there are laboratories working on developing these extremely intelligent wheelchairs that can be controlled by people who have no, or very little, ability to move.

"This is not telepathy," he stressed. "You can't imagine a cat and I would know you are thinking about a cat."