A non-invasive device that allows severely paralysed people to
interact with a computer via their brain signals has been improved
to make it a viable commercial product by Cambridge Consultants
(CCL).
Known as a brain-computer interface (BCI), the device uses
electroencephalography (EEG) to detect microvolt brain signals,
then applies an adaptive algorithm that focuses on the EEG features
the person is best able to control.
Those signals can be mapped onto functions for tasks such as
manoeuvring a cursor around a PC screen, or for spelling out words
for a speech synthesiser.
The system, developed by researchers at the Wadsworth Center, a
New York State health unit, originally used a large, $13,000
64-channel amp and the user had to wear a bulky cap to apply 64
electrodes to the skull.
CCL’s input involved reducing the size, complexity and cost to
make it more suitable for home or hospital use.
CCL also modified the on-screen menu system through which brain
signals are acquired, and added a synthetic voice – a requirement
if potential users were to claim financial assistance from US
medical insurance organisations.
“The sensor cap was designed to be worn for an hour or two at a
time,” said Mark Manasas, manager for surgical and interventional
devices in CCL’s Boston office. “We looked at it from the point of
view of a user wearing it for most of the day, and reduced the
number of sensors from 64 to eight.”
Accuracy was maintained through improvements in the spatial
mapping and signal processing of the EEG.
“In the future we’d love to get to active electrodes on the cap,
which amplify the signal right at the scalp, but for weight reasons
and medical regulations the first step is just to reduce the number
of sensors and keep them passive.”
The cost of the enhanced system has been reduced to $5,000, the
amount users can claim from Medicare for a speech-assist
device.
Stimulating spikes in brain activity
Systems for brain-computer interaction use two different
approaches, the so-called p300 response, and sensory motor rhythms
(SMR). The p300 technique elicits a microvolt spike in brain
activity, which has a particular waveform, by using a ‘surprise’
visual, audio or tactile stimulus.
In this case the system displays a grid of symbols on the PC
screen and the user concentrates on the letters of the word they
want to spell out. “The key is, there is a delay of about 300ms
from the time the response happens on the screen to the time of the
brain spike,” says CCL’s Mark Manasas.
“As the screen flashes a matrix of items, a user focuses on one
of the cells in the matrix. If you want to spell the word ‘BOY’ you
might have a matrix of letters on the screen and as the B flashes
the brain spikes.”
The computer would then look back at what flashed on the screen
300ms before the spike and associate it with the letter B.
www.cambridgeconsultants.com