It’s become a truism that things are progressing today at a rate our grandparents, even our parents, could never have imagined. That being so, the next steps in tele-rehabilitation can’t be very far in the future. One of them, in fact, may be a robot developed by scientists at the Massachusetts Institute of Technology.

Prof. Neville Hogan, director of MIT’s Newman Laboratory for Biomechanics and Human Rehabilitation, his MIT colleague, Hermano I. Krebs, a research scientist in mechanical engineering, and Drs. Mindy L. Aisen, Fletcher H. McDowell and Bruce T. Volpe of the Burke Rehabilitation Center in White Plains, NY, recently completed a 20-patient study of robot-assisted stroke rehabilitation therapy. A second, larger study is currently underway.

According to Prof. Hogan, the initial study involved patients who had had strokes affecting only one side of the body who were enrolled in the study after their initial period of recovery – about two weeks after their stroke. All of the patients received conventional therapy, and the experimental group also received robot-assisted therapy. The patients’ therapists did not know which patients were receiving the additional therapy.

"The patients in the experimental group were set up with a brace that helped them hold onto a handle on the end of the robot," said Prof. Hogan. "The elbow was also supported, so the arm didn’t have to hold its own weight."

During the therapy, continued Prof. Hogan, "patients saw a computer screen with an image and a cursor. We asked them to do very simple things with the mouse to move the cursor on the screen – for example, to move a red dot on top of a yellow dot. To do that, they had to move their hand. That’s where the robot came in. At first, patients are unable to make reaching movements with their affected arm, so the robot makes the movement for them. They experience the movement and see the movement and they get an audio reward (a beep). This is repeated many times over a period of six to eight weeks."

"The important thing is that the patient gets the movement experience and sees it happening." said Prof. Hogan. "Information is going back up the spinal cord to the brain."

"The majority of patients begin to show some recovery over time. As they recover, they begin to generate movement. If they are able to make the movement, the robot doesn’t help them move; it just ensures that they acquire the target. It keeps the movement inside a ‘virtual slot’ – if the patient is moving correctly, the robot does nothing; if the movement is going wrong, the robot will resist the wrong movement. As the patient becomes more able to do the task, the robot does less and less, and, eventually, does nothing. The continuum is from active assistance to passive guidance to independent movement.

"The model we are using here is called ‘active assisted therapy.’ The robot controls physical interaction with a human patient. The robot is able – in a somewhat crude way – to mimic what a human therapist would do, using visual and auditory feedback."

According to Prof. Hogan, the outcome of the first study showed that recovery was more than two times better with the robot-assisted therapy than with standard therapy alone. This was important for two reasons, said Prof. Hogan.

"The first reason is that, although therapy that involves the passive movement of limbs is standard practice, there continued to be debate as to whether that really makes any difference in recovery," said Prof. Hogan. "The main purpose of our first study was to test the hypothesis that passive movement does improve recovery. We showed that the answer is yes, and we showed it with an objective answer from a machine without a vested interest in the answer."

"Second, our data indicate that we can see significant differences in the way that recovery proceeds depending on the location of the stroke. This fits very well with what we know about the normal functioning of the brain. For instance, when the high cortical regions are affection, the patient’s aim is off, but the vigor of his movement is unimpaired. With deep brain lesions, however, the aim is ok, but vigor is seriously impaired."

"This second study should be done by the end of summer or early fall," said Prof. Hogan. "So far, everything confirms the results of the first study. We are continuing our studies because we want to understand fully what’s going on between the patient and the robot – for safety concerns, primarily. When the robot exerts force, we have to make sure that it’s gentle enough – we don’t want to hurt the patient. This robot has been designed to be unusually compliant – if you push it, it gets out of the way."

Could such a device be used as part of a home-based telerehabilitation program? Prof. Hogan is optimistic. Although the technology is not currently configured for distance healthcare applications, there is no reason it can’t be adapted.

First of all, size isn’t a problem, he said. "It’s quite small even now. We set out to design a machine that is small and fairly benign-looking. It weighs 80 pounds, so, technically, it’s portable. We anticipate that the commercial design will be even lighter and cheaper."

Bandwidth isn’t a limiting factor, either. "The communication requirements for this kind of tele-therapy are extremely modest," Prof. Hogan said. "It has tiny bandwidth requirements. And the additional bandwidth requirement for the robot is equally tiny. ISDN lines are more than we need. The typical Internet modem has more bandwidth than we need."

So, what are the limiting factors?

One of them is that, at the present time, the robot is only a prototype, and its functions are limited to upper arm movements, said Prof. Hogan. "We have units available for wrist and fingers, but we haven’t used them in clinical trials yet, and we’re working on designs for other parts of the body, too. These are probably six months to a year away before we could begin clinical trials. We’re looking into FDA approval for the robot, but we can’t predict how long that might take."

In the final analysis, said Prof. Hogan, "It’s the ability to have the robot work with the patient independently that will allow us to send the robot home with the patient. One of the problems with that, of course, is ensuring patient compliance. For that reason, the solution we envision is more like telementoring – that is, having the clinician interact with the robot and the patient. The patient would work with the robot at home, and the robot would be linked to the clinician’s office or clinic. The robot would keep records of what the patient has done and send them to the clinician. The patient could check in with the clinician and show what he has been doing, and the clinician, in turn, could demonstrate new exercises and program the robot with new instructions remotely. This could be real-time or store-and-forward."

The benefits to this kind of approach are obvious, said Prof. Hogan. "If you look at the cost of providing therapy in a community care center, then having a machine that can work with patients by itself would be a great benefit. The main thing we need to do is show that it can be done, it can be done safely, and it works for recovery of patients."

Contact information:
Dr. Mark Malagodi, Artsco – artsco@telerama.lm.com
Pat Aydelott, Rehab Dimensions - 724-733-1333
Dr. Neville Hogan, MIT - neville@MIT.EDU