Monkey think, cursor do

By Eric Smalley, Technology Research News

Computers fall far short of being able to read our minds. This could change. Brown University researchers have shown that, with the right filters, computers can interpret the electrical signals brain cells send to move limbs.

The researchers implanted electrodes in the brains of three rhesus monkeys and recorded neural activity as each monkey used a hand control to move the cursor on a computer screen. The implant, a centimeter-wide silicon chip covered with tiny spikes, recorded the signals from a small number of motor neurons in the monkeys' brains.

The researchers then built a mathematical algorithm that converted these neural signals into a control signal that moved the cursor. The algorithm translated the brain signals into computer signals in real-time, which allowed the monkeys to pursue a moving spot on the computer screen with the cursor just by moving their arms.

The algorithm averages the signals from 7 to 30 motor neurons to estimate where each monkey intends to move its hand. "It's as if each neuron gets a series of votes on where it thinks the hand is," said Mijail Serruya, a graduate student at Brown University. "Some of the votes relate to how the neuron feels right now, some relate to how it felt up to one second ago. The model... uses this to guess new hand positions from the neural activity alone," he said.

The researchers' system was able to produce a control signal after recording only a few minutes of the monkeys' manual control of the cursor, according to Mijail Serruya, a graduate student at Brown University.

"The scientific principle of decoding [motor neuron] activity rapidly, online, in a useful manner is now proven," said Serruya. the method could eventually help people who are paralyzed control electronic devices, he said. "This paves the way for possible development of a medical device that could help paralyzed patients."

One monkey eventually learned to control the cursor without visibly moving its arm. The researchers could not determine whether the monkey was using subtle muscle movements to produce the neural signals, however, and so do not yet know whether thought alone can be used to produce the control signal.

The neural control was as efficient as hand control at the task of pursuing the spot on the screen, said Serruya.

Paralyzed humans have already used brain implants to control computer screen cursors. But in those experiments, the subjects took months to learn how to use the system, said Serruya. "Any neuroprosthetic system requires both the machine and the person to learn," he said. "We believe that by having our machine -- the mathematical algorithm -- do a lot of learning, it makes it much easier and faster for the subject to learn their part."

In a similar experiment last year, researchers at Duke University, the Massachusetts Institute of Technology and the State University of New York Health Science Center used a monkey's motor neuron signals to control a robotic arm. In that case, the robot arm simply mimicked the actions of the monkey's arm, and the monkey did not consciously control the robot arm.

The Brown University monkeys controlled the cursors consciously in order to win rewards.

The researchers are considering applying their technique to other output devices, said Serruya. It's too soon to estimate when or if the technique could be applied to humans, he said.

Serruya's research colleagues were Matthew R. Fellows and John P. Donoghue of Brown University, and Nicholas G. Hatsopoulos and Liam Paninski who are now at the University of Chicago. They published the research in the March 14, 2002 issue of the journal Nature. The research was funded by the National Institute of Neurological Diseases and Stroke, the Defense Advanced Research Projects Agency (DARPA) and the Burroughs Welcome Foundation.

Timeline:   Unknown
Funding:   Government, Corporate
TRN Categories:   Biotechnology; Human-Computer Interaction
Story Type:   News
Related Elements:  Technical paper, "Instant Neural Control of a Movement Signal," Nature, March 14, 2002


March 20/27, 2002

Page One

Monkey think, cursor do

DNA solves big problem

Carving beams shrink circuits

Lasers snatch free-floating DNA

Etching could boost disk capacity


Research News Roundup
Research Watch blog

View from the High Ground Q&A
How It Works

RSS Feeds:
News  | Blog  | Books 

Ad links:
Buy an ad link


Ad links: Clear History

Buy an ad link

Home     Archive     Resources    Feeds     Offline Publications     Glossary
TRN Finder     Research Dir.    Events Dir.      Researchers     Bookshelf
   Contribute      Under Development     T-shirts etc.     Classifieds
Forum    Comments    Feedback     About TRN

© Copyright Technology Research News, LLC 2000-2006. All rights reserved.