Working in Movement

Humans move to do the things they want to get done. If you want to scratch your nose, write a quadratic equation or throw a baseball, you have to move your arm to do so. And we know that humans form neural images, or schema, of the arm moving before it actually moves. A previous post lets you experience this first hand.

Such schema are an integral part of learning any motor skill, even when that skill involves an external device. Think of learning to ride a bicycle, for example. When the bike becomes an accurate part of the schema, the relatively complex act of riding a bicycle can be left to the subconscious. Learning has occurred.

But what if you couldn't move your arm? Would there still be a schematic representation of that movement in the brain? According to Duke University neurosurgeon and professor of neurobiology Dennis Turner, M.D, the answer is yes. "It's known that patients who don't have use of their arm still show in MRI studies that the control centers in the brain are working normally. When they are asked to imagine moving their arm, the control centers become active," says Turner.

If actual movement isn't possible could external devices take their place? Could these devices be controlled by the brain's intact schema? Earlier this year, Duke researchers demonstrated that monkeys could learn to control a robotic arm used to play simple computer games. Evidently, the little guys brains were up to the task of learning to use the arm with their thoughts. (An earlier post, Monkeying with Robots has more on this.)

But humans who had lost their movement ability would certainly have more complex requirements than a robotic arm capable of making simple movements. The neural signals needed to begin the control process would need to be more complex, and come from deeper inside their brains. The brain and nervous system would need to be capable of producing a complex steam of signals for there to be any hope of developing such devices, let alone people being able to learn how to use them. Is this possible?

Another experiment at Duke suggests very strongly that it is. Turner and colleague Miguel Nicolelis, M.D. collected data from electrodes implanted into the brains of patients undergoing surgery for Parkinson's disease or other tremor related disorders. During each of 11 such operations, the researchers asked the patients to play a video game controlled with a hand held device. (Apparently, such surgery is done without general anesthetic.)

Signals from the electrodes were rich enough in information to accurately predict the hand movements made in the game. This accuracy in prediction is critical in being able to control external devices accurately, especially ones controlled by thought.

This is very promising. The remarkable learning capacity of the human nervous system can be tapped into and augmented, even when the rest of the system is compromised severely. But it is only a first step, providing an initial proof of principle that human application of brain-machine interfaces is possible. But there is still lots of work to be done. The next step planned involves implanting experimental electrode arrays in quadriplegics and leaving them there for a while. This will involve implanting the arrays in specific brain regions and then asking them to perform specific tasks, exploring which tasks are optimally controlled by that region.