Bretl takes EEG research to new heights

03/22/2010

It’s a crisp, early spring day, ideal for CSL researcher Tim Bretl and his neuroengineering students to fly remote-controlled planes. Except they’re not using remotes.

Students use EEG technology to control a flight simulator.
Students use EEG technology to control a flight simulator.

They’re using their minds to control the planes.

Bretl, an aerospace engineering professor, is researching the use of electroencephalography (EEG) connected to brain-machine interfaces (BMIs) to signal movement in aircraft. In other words, the researchers are using “mind-reading” technology to control planes in the sky. CSL faculty member Todd Coleman, an electrical and computer engineer, is working closely with Bretl on this endeavor.

EEG is a noninvasive measurement of brain activity recorded by an electrode cap worn on a person’s scalp. A BMI is a direct communication pathway between a person and a computer. In this case, the EEG is connected to a computer, and the voltage between electrodes is analyzed by Bretl’s team. If an individual imagines moving his left arm, brain activity will stimulate specific electrodes, which the computer reads.

In the recent past, researchers have studied the use of EEG and BMIs to help people with disabilities communicate simple instructions, such as spelling text or moving a wheelchair. For instance, if a disabled individual imagines himself moving left, the computer will recognize this and move his wheelchair to the left.

However, Bretl wanted to take this to the next level.

“I wanted to push the state of the art,” Bretl said. “I wanted to connect EEG to a vehicle with complex dynamics. Not necessarily because I’m interested in a brain-controlled aircraft, but because it’s way beyond the current state of the art. It’s a good test to see how far we could go.”

Because controlling aircraft is far more advanced than present EEG/BMI applications, Bretl said he has encountered challenges.

“What makes this problem hard is that EEG is such a noisy reflection of what the pilot wants. I'm collaborating with Todd (Coleman) and using tools from control and feedback information theory to say precisely how the aircraft should fly in order to best match the pilot's intent,” Bretl said. “We have a nice framework and we know how to apply it, but we still have a lot of questions.”

Further, Bretl explained that he wants to convey to the aircraft not just what direction he wants it to move at a specific moment, but the entire path of its trip.

“When a subject thinks ‘go right’, he’s not saying ‘turn to the right, right now’, but ‘somewhere along the line, I want to deviate to the right’,” Bretl said. “How to actually communicate this clearly is something we’re still working on.”

Bretl said that while this technology may never be used for flight in the real world – “Is this a practical way to fly a single aircraft? Probably not!” – he’s using the complex dynamics of aircraft as a foundation for advanced applications in other areas, such as helping the disabled gain mobility.

“We want them to drive cars, draw pictures, do the things that you and I take for granted,” he said. “These things are challenging, in the same way that the aircraft is a bigger challenge.”

Abdullah Akce, a graduate student in Computer Science and a member of Bretl’s team, said experiments so far have been fairly successful. However, its success depends largely on the subject.

“Some subjects are able to make the aircraft follow a predetermined route with close accuracy,” Akce said. “But other subjects have a more difficult time tracing the route.”

If his research continues to be successful, Bretl predicts that BMIs based on EEG may, in some cases, prove better than traditional interfaces like a keyboard or a joystick.

“We’re in the process now of doing systematic trials with human subjects. It’s a long road, and we’re in the middle of it,” he said. “Things are very much revving up now. I’m happy with our progress.”