A MONKEY is learning to manipulate an animated arm on a computer screen in
Richard Andersen鈥檚 lab in Pasadena. But it is not pressing keys or moving a
mouse. This monkey is being taught to control the arm by brain power alone.
The experiment forms part of a two-year programme at the California Institute
of Technology to develop an electronic interface with the brain to control arm
movement鈥攁 鈥渘eural prosthesis鈥�. Andersen predicts that in future such
prostheses will allow paralysed patients to operate a mechanical arm, say, or
even surf the Internet.
Other research groups are working on the same problem by tapping into the
brain鈥檚 motor cortex, where instructions for actions are generated. But Andersen
believes that the parietal lobe may prove very fruitful.
Advertisement
Three years ago, his team located an area, the parietal reach region (PRR),
where motor planning is carried out in monkeys. Neurons here 鈥渃arry early
cognitive plans鈥攅arly decisions to do something鈥�, he says.
PRR neurons don鈥檛 concern themselves with muscle forces. They leave such
details to the motor cortex. They fire when a plan is formulated and fall silent
as soon as the action begins, making them ideal candidates for a neural reach
prosthesis. Because PRR neurons don鈥檛 bother with fine detail, fewer signals
need to be recorded to figure out what the monkey intends.
Nevertheless, recording those signals is technically very difficult. But
preliminary results make Andersen optimistic. The next step will be to convert
them into commands for a mechanical arm. Here Andersen and his colleague Joel
Burdick have found another obstacle.
Activity in the parietal planning regions is encoded according to an
eye-centred view of the world. For movement, that input must be translated into
a different reference frame. You may look at an apple and know where it is in
relation to your eyes, but to take a bite you must know where your mouth is. For
that conversion, the parietal region takes in extra input, such as where the
eyes are looking. That activity must be measured to design prosthetics.
It is a tough problem. But Andersen believes it can be solved, and that an
aid for paralysis will one day come within our grasp.
