TWO days before his 40th birthday, Michael Marin of New York suffered a stroke that left him paralysed on one side of his body. After three years of physiotherapy, he had regained control of his left leg, but not the use of his arm.
In an attempt to recover some arm movement, last August Marin began three months of therapy with two trial robotic rehabilitation devices. He is now able to do push-ups. “They’re not perfect gym push-ups, but I’m getting there,” he says.
Each year 700,000 people in the US and 130,000 in the UK suffer a stroke, making it the biggest cause of severe disability in both countries. Standard physiotherapy can work well for some people, but it is expensive so patients often spend only a short time with a therapist. The standard of therapy can also vary hugely between practitioners, producing variable results. Improving stroke rehabilitation could clearly benefit millions.
Advertisement
Now it looks like robotic devices that help patients practise moving paralysed limbs could do just that, and are showing great promise in trials. “Robots offer consistency,” says Steven Cramer, a neurologist at the University of California, Irvine, who is working on a robot to restore hand and wrist use. Unlike people, robots are very good at repeating precisely the same movement over and over. “They can also record accurately what a patient does, connect with computers in a way humans cannot, and communicate accurately from a distance, opening the door for tele-rehab,” Cramer says.
Robots designed to restore movement to the shoulder and elbow, leg, wrist, hand or ankle are already being tested with people. “In the next five to 10 years, therapeutic robots will be present in all clinics and at home,” says Hermano Igo Krebs, a robotic rehabilitation specialist at the Massachusetts Institute of Technology. Such robots should also help people with other diseases that affect movement, such as Parkinson’s and multiple sclerosis, he says.
Cramer’s team, for example, has developed a prototype device called Howard, or Hand-Wrist Assisting Robotic Device, to help people regain the ability to grasp and release objects. Three straps secure the pneumatically operated robot around the hand, while the patient’s arm is fixed in a padded splint mounted on a movable platform. By moving the pads next to the patient’s fingers and hand, the robot can flex or extend all four fingers together, or the thumb or wrist alone. Sensors measure the movement of each pad, which allows software to track the patient’s movements as they go through a series of exercises involving grasping and releasing objects, or playing computer games that require them to squeeze virtual lemons to fill a cup, for example.
In its “active assist” mode, the software monitors to what extent the patient is able to carry out an action unassisted, and tells the robot to provide the appropriate level of pressure to help them to complete the move if they cannot do it on their own. Requiring a patient to start the movement and then actively helping them to complete it is crucial, Cramer says. “In completing the task, you’re teaching the sensory cortex what that feels like. The sensory cortex talks to the motor cortex. So by doing this, you’re instructing the motor cortex in how the movement should go.”
Cramer recently carried out a study in which 13 people who had suffered a stroke at least three months before, and so were past the time when they were likely to experience spontaneous recovery, received 15 2-hour sessions with Howard over three weeks. By the end of the trial, their ability to undertake real-world tasks, such as gripping a glass or picking up a phone, had improved by almost 10 per cent. A different test of their overall manual dexterity, which involved moving blocks from one side of a box to another, showed a 20 per cent improvement. “These were highly significant gains,” says Cramer.
Marin, meanwhile, owes his improvement to two devices: a wrist rehabilitation robot and a robot called MIT-Manus, which exercises the shoulder and elbow, both developed at MIT by Krebs and his colleagues. Marin was part of an ongoing trial of the robots. The wrist robot comprises two side-mounted motors attached to a support frame, which allows the patient to flex or extend the wrist, and move the hand left or right. Patients wear a virtual reality headset and perform a variety of simple exercises. As with Howard, if the patient can’t complete a task, the robot steps in. Marin says the robotic assistance is barely noticeable. “In the beginning it was hard for me to move, but the robot was so smooth you didn’t really notice.”
Patients taking part in the three-month trial at Burke Rehabilitation Hospital in White Plains, New York, which will ultimately involve 200 people, receive an hour of therapy three times a week. Krebs will present preliminary results, based on 36 volunteers, at a conference on rehabilitation robotics in Noordwijk, the Netherlands, in June. These show improvements of about 10 per cent in arm and wrist movements, something that for patients with severe, chronic impairment is “remarkable and very promising”, says Krebs.
Meanwhile, the US Department of Veterans Affairs is running a clinical trial of MIT’s wrist robot, alongside arm and hand robots, at four hospitals. It is also testing an “anklebot” on stroke patients at a hospital in Baltimore, and on people with multiple sclerosis at a hospital in West Haven in Connecticut. The anklebot, also developed at MIT, fits around the leg like a brace to provide assistance and support when walking. It is designed to help prevent falls, as well as improving people’s mobility and gait after a stroke. “Based on the outcome, the VA might adopt robotics in all its rehabilitation facilities,” says Krebs. The West Haven trial should also reveal whether other patient groups can show the same levels of benefit from robot therapy as stroke patients.
Krebs hopes that patients like Marin will soon have access to entire “gyms” of therapeutic robots, each for a different part of the body. Such facilities could be particularly useful for practising coordinated and physically demanding movements such as climbing stairs, says Etienne Burdet, who is developing a range of robotic rehabilitation devices at Imperial College London, including a robot designed to help patients recover the fine movements needed for handwriting. This is because although therapists can help people exercise individual muscles, physically supporting a patient while they carry out a coordinated task is much more difficult. Robots, in contrast, can easily provide both support and assistance, Burdet says.
Even better, studies have shown that patients strongly prefer robotic therapy to home-based exercises, so patients will be more likely to complete the rehab programmes, says David Reinkensmeyer, a biomedical engineer also at UC Irvine. He has developed a number of rehabilitation robots, including a robotic arm therapist called ARM Guide. However, he cautions that such active-assist studies usually compare the results of robot therapy with standard physiotherapy regimes, rather than with a matched amount of non-assisted therapy.
“We don’t know if you would get the same benefit if you turned the motors off and practised the same amount,” Reinkensmeyer says, “although we do know that it would be more frustrating, and clinical compliance to such a regime might ultimately be less.”
Human physiotherapists need not fear they will be pushed out of the picture entirely by robotic colleagues, however. With robots taking care of the movement training, therapists will have more freedom to focus on educating patients to live with their disability, and on managing pain.
If Marin’s experience is anything to go by, patients will take well to such a regime. “The robot has made such a difference to my life. I have seen a physical difference that has translated into my emotional well-being,” he says. “I really hope these robots get used all over the world.”
“I have seen a physical difference that has translated into my emotional well-being”
