With a headset of electrodes secured to his skull, Marvin Hinkley concentrated on the word "push," and the base of an electric wheelchair a few feet away slowly moved forward.

With a headset of electrodes secured to his skull, Marvin Hinkley concentrated on the word "push," and the base of an electric wheelchair a few feet away slowly moved forward.

Hinkley, a Southern Oregon University computer science student, demonstrated Wednesday how an electroencephalogram headset reading thought patterns might be used to move a wheelchair, turn on a light and complete other household tasks.

"The thoughts are very hard to generate at will," said Hinkley, who couldn't get the wheelchair to stop after it started moving.

Hinkley, 30, and fellow computer science student Rachel Albright, 22, gave a demonstration on the brain-computer interface during SOU's annual Southern Oregon Arts & Research event, which began Monday and ends today on the SOU campus.

The two seniors worked on the system as part of their capstone for computer science with an emphasis in programming.

The technology to read brainwaves and send them to a computer and then to a device to perform an action is nothing new, Hinkley said, but perfecting it and marketing it to the public is.

Most studies on the assistive technology have been done only for academic purposes, never from a commercial standpoint, Hinkley said.

If he can work out the kinks of the system, Hinkley would like to patent it and make it available to people with disabilities.

Reading a person's thoughts would be useful if someone had a disability that didn't allow for much physical movement, he said.

"If they're in a situation where they can't use a blow tube or a joystick — this is for them," he said.

Hinkley and Albright trained the headset-computer pair to identify some of Hinkley's brain impulses — the thought patterns he has when focusing his thoughts on actions such as pushing, pulling and going left or right.

The training takes place using a computer program with a floating cube which the user attempts to move forward and back, left and right. As the user continues with the program, the computer gets better at recognizing the intended thoughts.

"You have to generate a very discreet thought," said Hinkley. "And it becomes more accurate the more you train it."

Once the device is in use, electrodes attached to Hinkley's head read his brainwaves and send the message of a certain thought wirelessly to a computer interface that Albright helped design. The computer sends the command to the item, such as a wheelchair or light, and the item reacts.

The technology is strong, Hinkley said, but a lot of development and testing still needs to take place before it could be commercially marketable.

During Wednesday's presentation, Hinkley was able only to move the wheelchair forward.

"We have a good proof of concept, but there's still quite a bit of work to do," he said. "It's definitely a work in progress. I'm not sure if it's ready to be outside of a controlled environment."

Hinkley said before the product could be sold, he would need to be confident that an error wouldn't occur that could potentially put a disabled user at risk.

Albright said that the headset also can be used only on people with no hair or very short hair, because the electrodes need a good connection to the scalp to work properly.

"I would have had to shave my head," she said.

Hinkley bought a developer version of the EEG headset, and has other ideas for assistive technologies he could create, including using thoughts to generate text.

Teresa Ristow is a freelance writer living in Ashland. Email her at teresa.ristow@gmail.com.