London, England (CNN) -- Researchers are working on a breakthrough in artificial limb technology -- a prosthetic hand that can actually feel.
The SmartHand project is funded by the European Union and is a collaboration between researchers from across the continent. It has produced a prototype motorized prosthetic hand that researchers say gives unprecedented sensory feedback.
Fredrik Sebelius, of Lund University, in Sweden, is one of those working on the project. He told CNN that the SmartHand is able to exploit the fact that many amputees experience what he terms a "phantom hand."
"If you push the skin on an amputee's forearm, they feel like you are pushing on their phantom fingers," Sebelius told CNN.
When an amputee imagines moving a "phantom hand," signals are sent down nerve fibres in the remaining part of the amputated arm to activate muscles that would have moved the fingers.
Myolelectric signals from those muscles are recorded by electrodes applied to the forearm and then transmitted to motors in the artificial hand.
It's a technique that has been used in prosthetic limbs for decades, but Sebelius says the SmartHand gives much more control than other systems.
It also allows sensory information to be detected and transmitted from several sensors in each prosthetic finger, meaning users can actually "feel" objects they hold in the SmartHand.
"The big difference between our system and others is the sensory feedback", Sebelius told CNN.
"Sensors in the prosthesis pick up tactile information, which is relayed to actuators on the arm that pass on the sensory feedback, and this hasn't been done before,"
Sebelius gives the example of a pressure sensor on the artificial index finger sending a signal to forearm. By targeting the area of the forearm that activates the part of the brain associated with the index finger, the signal from the finger is "felt" by the brain.
He says the prosthesis could be commercially available within two years, but that the current technology is only suitable for amputations below the elbow. Upper arm amputees don't have enough muscles associated with hand movement to control the SmartHand.
Martin Twiste, senior lecturer of prosthetics and orthotics at the University of Salford, in England, told CNN that he did not know of any commercially available prosthetic hands that gave this kind of sensory feedback.
But he said the challenge with relaying sensory information from a prosthetic hand is sending the signals to the right place.
"Any sensory information from the prosthetic hand has to be fed back to the residuum (remainder of the amputated arm) and then to the brain," he told CNN. "The difficulty is where do you feed it back to?"
"If you have several electrodes on the residuum it's very difficult to place the electrodes accurately enough for the amputee to distinguish, say, the index finger from the middle finger."
One potential solution for upper arm amputees being explored by U.S. firm Deka Research and Development is to control an artificial arm using foot pedals.
Another method uses "Targeted Muscle Reinnervation," a technique developed by Dr Todd Kuiken at the Rehabilitation Institute of Chicago. This involves transferring the remaining nerves from an amputated limb to other muscles -- for example the pectoral muscle in the chest.
That means that when someone thinks about moving their amputated hand, they activate the muscle in their chest, and the myolelectric signals from that muscle can be used to control a prosthetic hand.
Researchers from the Johns Hopkins University Applied Physics Laboratory have developed a prototype prosthetic limb that uses this technique as part of a U.S. Defense Advanced Research Projects Agency-sponsored project.
But another solution is to directly attach electrodes to nerve bundles in the remaining part of the amputated arm, recording signals from the nerves, rather than from muscles.
Some of the SmartHand researchers have been working on this technology and Sebelius says developing this kind of "neural interface" is the long-term goal of the project.
Although neural interfaces have been trialled in animals, Sebelius says there are a number of problems that have to be overcome before the technology can be made commercially available for humans.
"The neural interface has to be implanted in the body, which brings problems of biocompatibility," Sebelius told CNN.
"A common problem is for the interface to be rejected by the body, then you get a lot of tissue forming around the interface and it doesn't function correctly."