Of the 540,000 Americans who will survive a stroke this year, some 70 percent will wind up with a paralyzed arm or hand. Physical therapists can train a stroke victim to use his “good” side more effectively, but healing the damaged side has proved difficult. Even if the patient is treated in the first three months, when 90 percent of recovery typically occurs, success is usually limited. Improvement after that time can be close to zero. That’s where the robots come in.
Working with a team of mechanical engineers at MIT, Bruce Volpe, a professor of neurology and neuroscience at Weill Medical College of Cornell University, has developed the first robot-assisted stroke-therapy device for upper limbs, and the results have been remarkable.
Clinical trials Volpe conducted at the Burke Medical Research Institute in White Plains show that recovery doubled for patients who received both human- and robot-assisted therapy, as opposed to just human therapy, within the first month after a stroke. And a study completed in December showed that patients who were months, even years, beyond the three-month barrier recovered 25 to 30 percent use of their arms in addition to the improvement they had made prior to that point.The robots work like this: A patient rests his arm in a plastic trough while his hand is strapped by Velcro to the “hand” of the robot, a metal cone controlled by motorized titanium rods. The patient then tracks a Pac-Man-like figure on the video screen. “If you can’t move, the robot will move you,” says Volpe. “If you begin to move, it will correct and guide you.” Three robots each “teach” an isolated motion: moving the arm from side to side, lifting the arm up and down, and rotating the wrist.
The key differences between robots and humans are intensity and accuracy. In 45 minutes, a robot can move a person’s arm 1,500 times in the same path at the same speed, as opposed to the dozens of vaguely accurate times a human therapist can make the same movement. “As the movement becomes more accurate, your brain remembers the trajectory,” says Volpe. “We think the brain is updating its performance records.”
Patients’ performance records are stored in the robots’ computer memories, and Volpe thinks the data could someday help doctors better understand how the brain “learns” from the movement of the arms.
“Remember,” says Volpe, “not too many lifetimes ago, when a person had a stroke you said, ‘Well, keep them comfortable.’ ” Not anymore. “Now what we’re saying is, we don’t know what the residual learning capacity of the learning machine called the brain is. We don’t know how far a person can come back.”
Volpe likes to tell the story of an elderly patient he treated four years after a stroke. The man later came to him crying. Volpe was concerned at first, but then realized the man was crying tears of joy. “He said, ‘I’m a meat-and-potatoes guy.
Last night my wife bought me a two-pound Angus for the first time in five years, and I was able to cut it myself.’ ”