Monkeys Use Brains to Move Computer Cursor: Study

Thu Jun 6, 5:35 PM ET

By Merritt McKinney

NEW YORK (Reuters Health) - US scientists have developed a system that allows monkeys to control a cursor on a 3-D computer screen without lifting a finger.

It may sound like a couch potato's dream, but the technology, which is still experimental, has the potential to allow people with paralyzed arms to control a robotic arm with their brains, according to one of the researchers.

In experiments with monkeys, Dr. Andrew B. Schwartz of The Neurosciences Institute in San Diego, California, and a team at Arizona State University in Tempe placed electrodes into a brain region called the cerebral cortex. The electrodes recorded the electrical activity of a tiny sample, 50 to 80 cells, of the neurons involved in movement, while monkeys used their hands to move a sphere on a 3-D screen.

During the experiments, the monkeys could not see their arms moving, but they could see the results of their movements on the computer screen. Once these signals were intercepted by the electrodes, they were sent to a computer that matched them to specific arm movements.

Using this catalog of movement-related signals, the monkeys were able to move the cursor without using their hands, according to a report in the June 7th issue of the journal Science.

An interesting observation, Schwartz told Reuters Health, was that monkeys learned the task by changing the signals that neurons use to control movement. As the monkey learned to improve, so did the computer program.

The next step, Schwartz explained, is to replace the cursor with a robot arm that a monkey can use to grab food when its arms are restrained. Eventually, this approach might work for people with paralyzed arms, he suggested. Schwartz added that the research may lead to the day when electrical stimulation of a patient's muscles could allow the patient to move his or her own arms instead of robotic ones.

The study shows that only a few dozen neurons are needed to control accurate hand movements, according to Drs. Peter Konig and Paul F.M.J. Verschure of the University/ETH Zurich, Switzerland. "Their results seem to suggest that effective encoding of motor actions can be accomplished with a very small number of neurons," they note in a related editorial.

In the report, Schwartz and colleagues point out that several human case studies have suggested that cells in the cortex can regain the ability to perform certain tasks even after long stretches of complete inactivity. The results of the present study indicate that the activity of neurons can be reorganized in a matter of minutes, they note.

With the aid of a computer program that can adapt to these changes, it is possible "to achieve brain-controlled virtual movements with nearly the same accuracy, robustness, and speed as normal arm movements," the researchers conclude.

SOURCE: Science 2002;296:1829-1832.