Brainwave interface goes 2D
Technology Research News
would be useful to be able to move a cursor around a computer screen just
by thinking about where you wanted it to go.
Researchers have been working on the task for more than a decade;
some recent progress involved monkeys moving a cursor to specific spots
on a computer screen. The price of being able to do this was fairly high,
however -- electrodes implanted in the brain.
Researchers from the New York State Department Of Health have shown
that it is possible to use brainwaves picked up by electrodes attached to
the outside of the scalp to move a cursor around a computer screen. Previous
efforts at monitoring brain waves through the scalp only allowed movement
along one axis, like up and down.
The Noninvasive Brain-Computer Interface allows people to learn
to move a cursor around a two-dimensional computer screen by controlling
the electrical noise that the brain makes as it functions, said Jonathan
Wolpaw, chief of the Laboratory of Nervous System Disorders at the Wadsworth
Center of the New York State Department of Health. The keys to the system
are recent improvements in signal processing and an algorithm that adjusts
to the way an individual controls brainwave oscillations.
The method promises to enable people who are severely disabled to
use a computer cursor. It could allow people who are completely locked in
their bodies -- without control even of breath or eye movements -- to communicate,
said Wolpaw. It may eventually be useful for people who are less severely
disabled, and, long-term, may be appropriate in specific situations for
anyone who needs to operate a computer hands-free, Wolpaw said.
In a four-person study of the interface, the researchers showed
that it is possible for people to learn to move a cursor to one of 16 areas
of a computer screen by changing the idling rhythms of the brain.
The method uses electrodes to record the rhythms, or oscillations
of the electric field, that occur near the sensory motor parts of the brain
-- the same electroencephalograph (EEG) signals that have been recorded
for 70 years for various clinical purposes. The oscillations are not part
of the brain's operations as far as we understand, said Wolpaw. "It's like
the noise a car engine makes when a car is operating."
Different areas of the brain produce a variety of rhythms that reflect
brain activity. "The question is how much control can people develop over
them," said Wolpaw. "In our case we're giving the brain the opportunity
to learn how to use the brainwaves we record for communication control,"
The program adapts in real-time to best leverage the way people
are changing their brainwave rhythms, said Wolpaw. This is key to being
able to use non-invasive sensors. "The big advance... is the on-line algorithm
has a lot more built-in adaptations," he said. "It's continually adjusting
to optimize the translation of whatever control the person has of cursor
At the same time, the adjustments encourage the person to improve
control. "This is really the interaction of two adaptive controllers," said
Wolpaw. "That's what goes on over a series of training sessions -- the system
and the user adapt to each other."
The researchers use brainwaves from the surface of the scalp above
the left and right sensory motor cortices, which are located directly under
the portion of the skull that rises above the ears. These are the areas
of the cortex that are most directly involved in movement and sensation.
The rhythms generated by the neurons in these areas "are basically... idling
rhythms -- what those areas of the cortex tend to [generate] when they are
awake and alert, but they're not doing anything," said Wolpaw. "What we
have shown is that people can use various kinds of mental imagery to make
[these] rhythms larger or smaller."
A person learning to use the control might start out thinking about
walking or moving a hand, said Wolpaw. The right sensory motor cortex controls
the left side of the body and vice versa. "If you think about moving your
right hand, generally sensory motor rhythms will tend to get smaller over
the left side of your brain." Relaxing increases the amplitude of the rhythm.
The rhythm amplitude gets converted into movement of the cursor
on a screen, said Wolpaw. If the rhythm gets larger the cursor goes up the
screen and if it gets smaller the cursor goes down the screen, and he said.
"People start out with various kinds of motor imagery until they find what's
best for moving the cursor in one direction verses the other."
A person acquires the skill of controlling brainwaves over a series
of 24-minute training sessions. Once the person masters moving the cursor
up verses down, he moves on to two-dimensional control.
Although the cursor is controlled using the signals from two electrodes,
the researchers' subjects wear a cap of sixty-four electrodes during the
training sessions. This is so the researchers can find ways to get the most
out of the signals. "We look... to see where the control is on the head
and what frequencies they're controlling and we can adjust the program off-line
to focus on those rhythms," he said.
Although people picture moving a hand or leg when starting to use
the system, as they gain more control it becomes more difficult for them
to explain exactly what they are doing to move the cursor, said Wolpaw.
"Generally after people get better, and particularly when they go on to
two dimensions, they can't really tell you what they're doing," he said.
"They really can't explain [in the same] way you really can't explain how
you raise your arm -- it becomes more like a normal motor skill."
The researchers' subjects had from 10 to 30 hours of two-dimensional
training. Two of the users are more skilled than the other two, but all
four are able to control the cursor with about a third of the speed and
accuracy of a joystick, he said. "They can hit as many as 16 different target
locations... around a screen."
Initially, the skill requires considerable concentration. As people
get better they can use the system to answer questions. "They can hear a
question, they can think of the answer, and they can use [brainwaves] to
move a cursor to the answer," said Wolpaw. "So it becomes more like a normal
skill as you get better at it -- it requires less undivided attention."
In the short-term, the method could find practical use for people
who are severely paralyzed, and medium-term for people who are perhaps not
as disabled, said Wolpaw. "For the foreseeable future it is going to be
for people with disabilities," he said.
Long-term, the method could find use under very specialized circumstances
for people who are not disabled, "but I don't see that as an a major use...
for quite awhile," said Wolpaw.
Although the brainwave control could be used as an additional communication
channel, it remains to be seen whether it could be added without impinging
on the function of the normal channels people used to communicate with and
control computers, said Wolpaw.
The researchers' next step is to use the system for a project based
in Germany that is designed to train people who are disabled to use the
system. "We're developing [and] implementing clinical applications," said
Wolpaw. "We're setting up to show that it can actually improve the quality
of life." Applications include word processing and environmental control.
The researchers are also working on a selection, or grasp function.
This will make it possible to move the cursor over an object without selecting
it, and will also enable the method to be used with robotic arms, said Wolpaw.
They are also working on three-dimensional cursor control.
Wolpaw's research colleague was Dennis J. McFarland. The work appeared
in the December 6, 2004 issue of the Proceedings of the National Academy
of Sciences. The research was funded by the National Institutes of Health
(NIH) and the James S. McDonnell Foundation.
Funding: Government, Private
TRN Categories: Applied Technology; Biotechnology; Human-Computer
Story Type: News
Related Elements: Technical paper, "Control of a Two-dimensional
Movement Signal by a Noninvasive Brain-Computer Interface in Humans," Proceedings
of the National Academy Of Sciences, December 6, 2004
February 9/16, 2005
Atomic scopes eye
Brainwave interface goes
marks to digital text
View from the High Ground:
Xerox's Hervé Gallaire
on plastic speed circuits
promises tight spots
nanocrystal transistor shines
Research News Roundup
Research Watch blog
View from the High Ground Q&A
How It Works
News | Blog
Buy an ad link