"Helping Hands," Feature Article, December 2002

helping hands

A robot's touch may have the power to restore lost motor skills.

This article was prepared by staff writers in collaboration with outside contributors.

In the field of medical rehabilitation, robots can act as external aids to assist people in such tasks as lifting a fork or spoon. But what if robots could be used to retrain the human nervous system? Many brain-injured patients might one day relearn to perform coordinated tasks on their own.

Working toward this goal in the Sensory Motor Performance Program at the Rehabilitation Institute of Chicago is James Patton, a neuroscientist who combines his medical training with a background in mechanical engineering. Patton and a team that includes therapists, doctors, and others at the institute are working with robots in the hope of training people to regain control over their movements.

The research is funded by the National Institutes of Health, specifically the National Institute of Child Health and Human Development.

When commands are issued by the nervous system to generate a movement, they are typically dispatched to several muscles that drive the limb along the intended path.

It is known that a nervous system possesses the natural capacity to adapt to environmental forces. For instance, when a force field systematically disturbs arm motion, people learn to anticipate and compensate for the force. When the disturbing force field is unexpectedly removed, healthy subjects make erroneous movements in the opposite direction of the perturbing force field. This behavior is called an after-effect.

Patton is trying to reverse-engineer this phenomenon to learn the forces that will result in desirable after-effects. He believes that if he can custom design a force field to produce a desired trajectory resulting in an after-effect of adaptation, he would be showing the potential for neural rehabilitation. Thus, he would be able to use these methods to teach motor skills to some brain-injured patients.

Using Simulink from The MathWorks in Natick, Mass., Patton starts with a dynamic simulation of an arm and
the nervous system that controls it. The simulation predicts how people will respond to a set of forces.

A MathWorks simulation predicts how a person will react to the force of a robotic arm. Adaptation could be a key to neural recovery.

Patton has used adaptive training techniques in healthy people to determine the robot's ability to modify the subjects' hand movements to a specified trajectory, such as a curved path of some predetermined shape. And even though the desired after-effect is eventually canceled out in healthy people, who recognize it as a mistake, Patton believes that this won't happen in stroke patients, because for them the after-effect is a healthier movement.

In one preliminary study, Patton asked nine hemiparetic stroke patients to participate in his adaptive training experiment. The nine patients were trained by making directed movements in the presence of a force field specifically designed so that when it was unexpectedly removed, a desired trajectory would result. The system measured the forces and motions of each patient's movements.

As Patton explained it, patients would be asked to reach for an object, but because of their impairment would err in the attempt. By using the interference of the robot arm, he was essentially "enhancing the error," so the after-effect would put the patient on target.

Patton said the method differs from that of a therapist, who would attempt to facilitate movement by showing the patient how to move correctly.

All but one subject showed a significant shift toward the desired trajectory, even though none of them was given a hint of what it was. Essentially, the method tricked the nervous system into generating a new motor command.

Although at this time there is no evidence proving that patients can preserve their after-effects, some patients do appear to preserve limited features for the duration of the experiment. Patton believes that prolonging the training over many days could lead to the desired outcome.