Robotic arms controlled by human brain.

This is so fucking cool.

A wheelchair-mounted robotic arm controlled by thought alone has been created by scientists at the University of South Florida.

The device could give people with amytrophic lateral sclerosis (ALS) or full body paralysis the ability to perform simple day to day functions that would otherwise be impossible.

"We aren't reading people's thoughts," said Redwan Alqasemi, a scientist at the University of South Florida who, along with Rajiv Dubey and Emanuel Donchin of USF, helped develop the software and hardware. "This is the first time a person with severe disabilities like ALS can perform daily activities for themselves."

Over time, patients with ALS slowly lose control over their muscle movement, losing the ability to move their arms, legs and eventually all muscles except those around the eye. Patients with ALS have fully functional brains, but have no way to express their thoughts.

EEG scans offer one way for patients with ALS to communicate with the outside world. By fitting patients with a head cap equipped with electrodes and filled with an electrically conductive gel, scientists can monitor particular kinds of electrical impulses coursing through the brain.

In this case, the scientists monitor a particular brain wave called P300, so-called because it lasts about one-third of a second. Reading P300 waves is basically like reading a person's thoughts, but only in the most coarse kind of way.

For the wheelchair-mounted robotic arm, the person in the wheelchair looks at directional arrows flashing across a small screen. When the arrow points in the direction that they want to go, their brain lights up on the EEG, and the wheelchair or robotic arm moves accordingly.

This doesn't happen at the speed of thought, however. Turning the wheelchair or moving the robotic arm takes about seven seconds as the arrows cycle across the screen. The wheel chair or arm continues in that direction until it receives a new command.

The wheelchair or arm could move faster, but it might not move as accurately, said Alaqsemi. The next step for the USF scientists is to refine the model's hardware and software, to increase speed and reliability while cutting down on weight.

"Every pound you take off the robotic arm is another pound of payload that can be lifted," said Alqasemi.

Right now the robotic arm can lift about four pounds, about the weight of a gallon of milk. In the next version Alqasemi hopes to double the payload.

Lifting a door handle or moving a gallon of milk may seem like simple tasks, but according to Jonathan Wolpaw, who builds brain computer interfaces at the Wadsworth Center in New York, using thought-controlled devices is harder than simply just thinking.

"Our normal muscle movements require practiced skill and control," said Walpaw. "Controlling brain activity is also a skill that requires practice."

Reading P300 brain waves is a good system, argues Walpaw, because it doesn't take a lot of practice to train the brain. With only one WMRA built so far and no current plans to commercialize the design, not many people will get the chance for their brain to learn the new skill. But when commercial models appear in several years, even slow brain computer interfaces could make the impossible, possible.

"It would allow patients with severe disabilities the ability to control their own environment and have some form of independent mobility," said William Heetderks, Director of the National Institute of Biomedical Imaging and Bioengineering. "It would be very valuable to these individuals."
- Source