Engineering the Sense of Touch


by Jim Schnabel

December 22, 2009

The idea of being able to feel things with an artificial hand might seem like science fiction. But sensory devices that register pressure, texture and temperature can already be engineered into a prosthetic limb, and in laboratories in the United States and Europe, connecting such devices to amputees’ sensory nerves is becoming a reality. Much of the research in the area now is aimed at perfecting these connections through the development of durable machine-to-nerve interfaces.

“In the past, people have used cuff electrodes, or electrodes that wrap around a nerve,” says Paul Cederna, a surgeon and nerve-interface developer at the University of Michigan. “People also have tried putting probes into nerves. Unfortunately both of those create scarring around the nerve, so that they stop working over months and years.”

Cederna announced at a meeting of the American College of Surgeons in Chicago on Oct. 14 that his team’s new sensory nerve interface had worked for the duration of a nine-month study in rats, and throughout the study had enabled the animals to feel a tickling stimulus in their artificial paws. “It looks really good so far,” Cederna says.

Cup-shaped and permeable, Cederna’s interface uses an electrically conductive polymer, instead of metal, to transmit signals, and is lined with biological “scaffold” material derived from human muscle tissue. The amputee’s own muscle cells are placed inside, where they grow into the scaffold and soon attract nearby blood vessels. A nerve fiber with the desired function  is also placed in the scaffold, and in principle forms a working connection to the muscle cells, similar to the “neuromuscular junction” it would form in an ordinary limb.

Cederna says that this basic junction technology can be used to take electrical sensory signals coming from the prosthesis and use them to stimulate the appropriate sensory nerve. It also can transmit the impulses of a motor nerve to the appropriate mechanical actuator within a prosthesis.

“In the operating room we can identify, for example, a sensory nerve that used to go to the thumb,” he says, “and put one of these little cup [junctions] on the end of that nerve.”

Cederna’s group is applying for FDA approval to test the technology in human amputees. “We’re really excited about this,” he says, “and I think ultimately this is going to make a difference.”

Less direct but more natural?

“Contacting directly to the peripheral nerve is a very exciting idea,” says Todd Kuiken, a physician and biomedical engineer at the Feinberg School of Medicine at Northwestern University who also directs the Neural Engineering Center for Artificial Limbs at the Rehabilitation Institute of Chicago. “But it’s very, very hard.”

Instead of hunting for individual nerve fibers, Kuiken prefers a “targeted reinnervation” strategy, in which bundled nerve ends at the site of amputation are reimplanted elsewhere in the body, for example in the upper arm or the chest. In their new home, these nerve bundles sprout motor and sensory nerve fibers that form natural nerve-ends in skin and muscle. Pressure upon such a “reinnervated” skin patch thus creates the sensation, for the amputee, of pressure upon the missing limb. Kuiken wants to link pressure-signals from transducers on, say, an artificial hand, to a device—called a “tactor”—that creates pressure over the appropriate places on a reinnervated skin patch.

“We hope to get to the point someday,” Kuiken says, “of maybe having sensors in each finger and little things that push on different parts of your reinnervated skin so you feel different contact with different parts of your [artificial] hand.”

One advantage of this method, he adds, is that the reinnervated skin automatically develops an array of sensory abilities similar to those it had in the lost limb. “Skin has special transducers for picking up pressure, initial contact, vibration and so on,” Kuiken says.

Kuiken and his colleagues described how the technique can reproduce tactile sensations with reasonable accuracy, in a recent paper in the journal Brain. But the Northwestern researchers also are trying to use the technique to take motor signals from reinnervated muscles and transmit them to a prosthesis. Although the technique is already under study in a small number of human amputees, “we’ve got lots and lots of things that we’re trying to fix,” he says. “There’s no lack of challenges.”

Other approaches

Both Kuiken and Cederna have been funded by the Department of Defense, whose Advanced Research Projects Agency has spent $100 million over the past several years on a “Revolutionizing Prosthetics” program that includes more than a dozen U.S. laboratories.

A large European/Israeli research consortium also has made progress on a sensate “SmartHand,” and recently reported a basic, proof-of-principle demonstration in a human subject. Their sensory strategy is similar to that of Kuiken’s group but avoids reinnervation surgery.

“We do not use any nerve-machine interface by introducing electrodes into components of the nervous system,” explains team member Goran Lundborg, a hand surgeon and neuroscientist at Sweden’s Lund University. Instead Lundborg and his colleagues make use of the reorganization in the sensory cortex of the brain that seems to occur automatically when a hand is amputated. “The cortical hand area is invaded by an expanding forearm area,” Lundborg says. “One result is that there is a mapping of the missing hand in the [lower] forearm. This map can vary much between patients, but usually it is present.”

Thus, as in Kuiken’s approach, sensors in the Smart Hand transmit to this “hand map” via small servomotors that press against the skin. “Stimulation of the thumb part of the hand map is perceived as true feeling in the missing thumb, and the same goes for all the fingers,” Lundborg says.

Researchers generally agree that the development of this artificial sensory technology is still in its early stages. Ideally it would restore to an amputee tactile and even pain sensitivity in different parts of a prosthesis, as well as “proprioception”—the sense of where a limb is and how it is oriented. But restoring proprioception is an especially big hurdle, Kuiken says, because scientists don’t yet have a good understanding of how the nervous system produces this sense.

In any case, Cederna says, the aspect of prosthetic technology most needed to overcome such hurdles—and the aspect that traditionally has lagged furthest behind—is the nerve interface, so it’s good that some progress is being made. “The nerve interface piece,” he says, “is clearly that piece which is going to make a difference in giving somebody a prosthesis that will allow them to play the piano, or use a keyboard, for example.”