ST. LOUIS — Research on prosthetic hands has come a long way, but most of it has focused on improving the way the body controls the device.

Now, it may also be possible for prosthetic hands to send signals back to the body and “tell” it information about what the bionic hand is touching, according to a new study.

Recently, researchers at the Defense Advanced Research Projects Agency, the research arm of the U.S. military, implanted an array of small electrodes into the region of the brain that controls movement in a woman who is paralyzed.

The electrodes communicated electrical activity from the brain’s motor cortex, via wires, to a prosthetic arm that the woman was able to move through a wide range of motions.

Then the research team asked, “Can we run the experiment in reverse? Can we do for sensation what we did for the motor system?” said Justin Sanchez, program manager of the DARPA biological technologies office, in a presentation he gave on Thursday at the Wait, What? A Future Technology Forum, which DARPA hosted in St. Louis.

To answer this question, the researchers worked with a 28-year-old man who is paralyzed. They implanted an electrode array in both his motor cortex and sensory cortex, the brain region that recognizes tactile sensations such as texture and pressure.

Wires from the motor cortex array controlled the hand, as they did for the female volunteer, and sensors in the hand also conveyed information, via another set of wires, back to the array in the sensory cortex.

The researchers showed that this feedback system allowed the hand to communicate directly with the brain. In a video included in Sanchez’s presentation, a researcher blindfolded the man and then gently pressed on different fingertips in the prosthetic hand.

The volunteer was able to identify which fingertip was being touched with “nearly 100 percent accuracy” even without seeing it, according to a DARPA press release about the research.

People who have prosthetic hands today rely on being able to see what the hand is doing to control it, said Sliman Bensmaia, an associate professor of neuroscience at the University of Chicago.

But people will never be able to use these hands with dexterity until they can feel what they are doing without looking at them, he said. Bensmaia did preliminary research for Sanchez’s team on how to make the electrode array work in the sensory cortex.

“On the short term, you want to know whether you are touching an object, and how much pressure you are exerting on it, those basic things that you need to hold things,” Bensmaia said. But as the technology progresses, touch sensors may also be able to convey temperature and texture, he added.

Although the current demonstration is the first of a prosthetic hand directly communicating with the brain, other researchers have demonstrated that they can send messages from sensors in the prosthetic hand to electrodes implanted in nerves in the arm that then communicate with the brain.

“(However), in situations where people have spinal cord injury, so they are quadriplegic … you probably couldn’t give them sensation back through the nerves,” because they have lost the use of the nerves in their arm, said Dr. Paul S. Cederna, professor of plastic surgery and biomedical engineering at the University of Michigan.

Devices on the market now rely on either body power, in which a healthy part of the body controls the prosthetic through cables and harnesses, or myoelectric devices, in which electrical signals from muscles attached to the prosthetic control it.

Researchers are also working on developing highly sensitive prosthetic arms that can recreate nearly every motion of a real arm, and bionic hands that can be controlled through an iPhone.

The big benefit of Sanchez’s approach is being able to use prostheses for people with spinal cord injuries, Cederna said. The 28-year-old man in the current demonstration has been paralyzed for more than a decade because of a spinal cord injury.

Although Cederna was not involved in Sanchez’s research, he conducts DARPA-funded research on how to improve control of prosthetic devices through peripheral nerves, such as those in the arm.

The idea of implanting an electrode array into the brain to either control or receive signals from a prosthetic limb is big step forward, but it is not ready for prime time yet. “The biggest challenge, once you put that electrode into the brain, you develop scarring around the electrode, and that makes it increasingly difficult to pick up the signals it needs to pick up,” Cederna said.

Researchers are working hard to develop electrode arrays that work for longer periods of time, Bensmaia said. Currently electrode arrays in the motor cortex only work for a few years, although arrays in the sensory cortex appear to be more stable, he added.