When it comes to virtual environments, most systems require
participants to navigate by using a joystick or some other hand-held
device. This results in contradictory stimuli to the brain.
The person may have the illusion of walking through a computer-generated world,
but in the real world, they are moving a stick or pressing a button,
reducing the sense of being immersed in the virtual environment.
The system — under development by researchers at the Graz
University of Technology in Austria, the University College London,
and Guger Technologies, also in Graz — eliminates that confusion.
It consists of an electrode cap, which is worn by the user, an
electroencephalogram amplifier, and a computer loaded with specialized
software.
In experiments conducted over a five-month period, Leeb's team tested
the system on three healthy people between the ages of 23 and 30.
They were first asked to calibrate the system with their own brainwaves by performing a number of feedback tests.
The tests involved moving
a bar between cross hairs on a computer screen, navigating through a
virtual space while wearing a head-mounted display, and, finally,
sitting in an immersive room, called a cave, and imagining their
character walking to the end of a street in a virtual-reality city.
In each run, the computer prompted the participant with a visual cue,
such as an arrow pointing down, and an auditory stimulus, such as
double beep, to think about moving their hands or feet.
If the
participant imagined the wrong movement, their virtual character did
not move or some cases moved backward. A system gave the user a score
based on the accumulated forward distance imagined.
All test participants successfully navigated within the three
different environments and two of the participants achieved 100
percent success.
Leeb attributes the high rate to the feedback
of virtual reality imagery, which has not been incorporated by other
research groups.
Although the cues increase performance, Leeb would like to eventually
move away from using them altogether. But that presents a
technological hurdle, as it forces the machine to recognize the
correct brain signals from among all the other signals firing.
"You have always to distinguish the true signal from noise," said
assistant professor Andrea Kuebler, an expert in brain-machine
interfaces as the University of Tübingen in Germany. "To classify the
important stuff, that is the challenge."
