"Imagine being totally trapped in your body and not being able to even communicate with anybody. You'd go nuts," says computer scientist Jessica Bayliss of the University of Rochester. That can happen to tens of thousands of people suffering from neuromuscular diseases. So Bayliss is developing a technology to help them stay connected.
Her system taps directly into the mind via electrodes that generate EEGs, graphs of brain electrical activity. When someone sees an object that elicits a strong reaction— a friend waving or a lion on the loose, for example— a spike of recognition appears in the EEG one third of a second later. Bayliss's computer program can then pick out these recognition signals. In a virtual reality simulation of driving, her setup identified when a person saw a red light 85 percent of the time. That kind of quick, reliable response could enable an invalid to control an electric wheelchair. In another test, she displayed a virtual-reality apartment in which each appliance was marked with a blinking light. The lights switched on at different times. If her system picked up an EEG signal a third of a second after the TV blinked, the computer knew that the person wanted to turn on the TV.
"The nice thing about this stimulus-reaction thing is it's a normal response, nearly everybody has it, and it's something that's not easily damaged," Bayliss says. Unlike current systems to aid the handicapped, therefore, hers should require little training to use. Soon she hopes to restore to some what most of us take for granted: the ability to control everyday objects without even thinking about it.
