Paralysed people can now control artificial limbs by thought alone.
Ian Sample reports

Thursday March 31, 2005
The Guardian

There's a hand lying on the blanket on Matt Nagle's desk and he's
staring at it intently, thinking "Close, close," as the scientists
gathered around him look on. To their delight, the hand twitches and
its outstretched fingers close around the open palm, clenching to a
fist.
In that moment, Nagle made history. Paralysed from the neck down after
a vicious knife attack four years ago, he is the first person to have
controlled an artificial limb using a device chronically implanted
into his brain.

The experiment took place a few months ago as part of a broader trial
into what are known in the business as brain-computer interfaces.
Although it is early days, aficionados of the technology see a world
where brain implants return ability to those with disability, allowing
them to control all manner of devices by thought alone. There are huge
hurdles ahead. No one knows how much information we can usefully
decipher from the electrical fizz of the brain's 100bn neurons. More
importantly, scientists are still in the dark as to what effect, if
any, long term implants will have on the human brain, or how its
circuitry will cope with the new tasks demanded of it.

Nagle got involved in the latest trial after hearing about John
Donoghue, a professor of neuroscience at Brown University on Rhode
Island, whose company Cyberkinetics has developed an implant called
BrainGate. Under Donoghue's instruction, Nagle was given a general
anaesthetic before a disc the size of a poker chip was cut from his
skull. After making an incision in the brain's protective membrane, a
tiny array of 96 hair-thin electrodes, each protruding about a
millimetre, was pressed onto the surface of his brain, just above a
region of the sensory motor cortex that is home to the neuronal
circuitry governing arm and hand movement. With the electrodes in
position, the bony disc was replaced, leaving room for a tiny wire to
connect the electrodes to a metal plate the size of a 10p piece that
sits on Nagle's head like a button.

To read brain signals from Nagle's motor cortex, Donoghue's
researchers attach an amplifier to the metallic button on his head and
run a cable to a computer. When he's hooked up, the tiny voltages of
the sparking neurons beneath the electrodes produce a series of
brainwaves that dance on the computer screen.

Since having the electrodes implanted in June last year, Nagle has
been test-driving the technology, seeing what he, and it, are capable
of. "We're evaluating his ability to do a whole range of things. We've
hooked him to a computer that lets him turn a TV on and off, change
channel and turn the volume up and down," says Donoghue.

The success of the technology relies on being able to decipher
accurately the electrical activity within Nagle's brain and turn it
into useful actions. The trials started tentatively. Nagle had been
unable to move any of his limbs for nearly four years. The scientists
had no idea how this would have affected the brain signals that
normally control movement. Would they have fizzled out through lack of
use, much as muscles waste away in the wheelchair-bound? "No one knew
if it would work in someone with these injuries, but simply by asking
him to imagine moving we got useful signals and it was amazing. I was
overwhelmed by how beautifully the cells were still working," says
Donoghue.

Getting the signals is one thing; deciphering them is another. But
Donoghue's team found that some simple rules held - if the brain
wanted to move the hand to the right, certain cells would fire a rapid
series of impulses. If the brain was willing the hand to move left,
the cells fired a different number of times. Other information, such
as where the hand should end up, what trajectory it should take, and
how quickly it should move, is also embedded in the electrical
signals.

Part of the difficulty in reading brain signals is that while even a
simple movement such as raising a hand requires electrical signals
from many regions of the brain, the implanted electrodes pick up just
a tiny fraction of those that fire. "We're recording only a dozen or
so, when a million might be active," says Donoghue, who likens the
process to dropping a microphone into a crowded room and trying to get
the gist of all the conversations going on.

The limitations of taking signals from just a few active neurons have
become apparent in the trial. Many of the tasks Nagle is set involve
moving a cursor around a screen by thinking which way it should move.
But the cursor jiggles, making it difficult to select icons on the
screen with any precision. "We could smooth it out using software, but
at the moment, we want to see if Matthew can learn to control the
wobble," says Donoghue, who is recruiting four other patients to
complete the trial. "If he can do that, he can use computer software
to answer emails, and if he can do that, he could be employed."

Ultimately, Donoghue says there should be no need to connect cables to
peoples' heads to read their minds. Miniaturisation should bring
smaller devices that can be powered through unbroken skin and transmit
signals wirelessly from the brain to a processor worn on a belt that
triggers the intended device.

If all goes according to plan, Donoghue's trial, designed to explore
how well a variety of people can control different devices by the
power of thought, will be completed in about 18 months. He's not the
only one keen to find out just how useful such devices could be. At
Duke University in North Carolina, Miguel Nicolelis is in the final
stages of getting permission to fit 16 quadriplegic patients - half in
the US, half in Brazil - with brain implants for a period of 30 days.
Initially the trial will look at whether the patients' brains still
produce useful motor signals. "Then, we want to see if these patients
can control a robotic arm that can reach and grab objects, and how
well their brains get used to it," says Nicolelis.

In previous studies, his team showed that when monkeys had their
brains hooked up to robotic arms, they assimilated the arm,
effectively making it their own. "Their brains actually incorporated
the robotic arm by dedicating neuronal space to it. We want to see if
the same thing happens in humans," he adds.

For all the promise brain implants hold, there are some that believe
they are not the best bet for many patients. Implants suffer from a
number of drawbacks, the first being that they demand invasive
surgery, with attendant risks. Second, implanted electrodes cause at
least some inflammation of the brain tissues they push into. As well
as obvious medical concerns, if the inflammation is significant, it
can smother any signals the electrodes might pick up.

"Every one you put in gives some inflammation, but it's minor. We're
still working on making electrodes more biocompatible, but we've got
monkeys who have so far survived for nearly five years with implants
and they are fine," says Nicolelis. "The thing is, to do what we want
to do, to get that level of control, you have to get into the brain."

Nicolelis says his goal is to use brain implants to allow the disabled
to walk again. He has already started designing a wearable robotic
"exoskeleton" that could help power paralysed legs - think Wallace and
Gromit's The Wrong Trousers, only with better control. Nicolelis is
also developing something called "shared control" in which a robotic
limb is triggered by a basic command from the brain, but refines and
carries out the movement itself, using pre-programmed intelligence.
"The hurdles ahead, after finding even better electrodes, are
developing prosthetics that are more amenable to brain control," he
says.

Many of the labs looking at brain implants started out doing basic
research into understanding how small numbers of neurons worked. The
research required the development of thin wire electrodes that could
cosy up to individual neurons, a legacy that led to fully implantable
devices. But for many applications, simpler signals, that can be
picked up without undergoing major surgery, may suffice.

At the Wadsworth Centre, the laboratory arm of the New York State
health department, John Wolpaw and his team recently proved that a hat
not unlike a swimming cap peppered with electrodes could pick up clear
enough signals to allow the wearer to move a cursor around a computer
screen. "There was an unsupported assumption that to get that kind of
control, you needed to implant, but our work showed that's not the
case. These systems can do better than a lot of people give them
credit for," says Wolpaw.

Instead of tapping into the brain's natural signals for moving limbs,
Wolpaw's system picks up changes in general brain activity that the
patient must learn to control. "We look at rhythms on the EEG that are
normally just idling, but we've shown that by using mental imagery,
people can learn to make the signals stronger or weaker and we can
translate that into cursor movement," says Wolpaw.

Wolpaw's patients are trained over 10 sessions, during which about 80%
learn to control their brainwaves well enough to move a cursor around
a screen. In time, most can do other things, such as think of answers
to questions to select on screen, without it interrupting their
control. The risks of the technique are undoubtedly fewer than for
full brain implants, though questions remain about the effects of
forcing the brain to change its activity, in a way the electrodes can
pick up. "It's probably just like learning anything else. There's been
no indication that any of this does anything harmful, and it's hard to
see how it could, but we can't say for sure," says Wolpaw.

While Wolpaw has achieved control many thought impossible without
implanting electrodes directly into the brain, he feels a third
technique, called electrocorticography, or Ecog, might have the
brightest future. Ecog involves a smaller operation to place a small
sheet of electrodes on the surface of the brain. "With this, you get
strong signals, you can pick them up from smaller areas but you're not
sticking something into the brain," he says. Preliminary trials show
patients can learn to use Ecog devices much faster than electrodes
placed on their scalps.

More than likely is that all three techniques will co-develop, each
finding its own niche. Full implants may only be worthwhile for the
severely disabled, who need to control complex machinery, such as
prosthetic limbs, with their thoughts. For many though, regaining even
the most minor level of independence would help. "One fellow said to
me, 'I just want to be able to scratch my nose'," says Donoghue. "It's
easy to forget the kinds of extraordinary things people can't
accomplish. If you can do something that lets them reach out to the
world even a little, it can make a huge difference."