Novel Rehabilitation Methods for Stroke, Cerebral Palsy, and Spinal Cord Injury

Traditional physical therapy for neurologic conditions can be boring for the patient and tiring for the physical therapist, making it difficult to put in the required number of training hours. That is why researchers are developing a new generation of physical therapy tools that use video games, robotics, and virtual reality. Few of these products are available today, but the hope is that someday patients with cerebral palsy (CP) and those recovering from a stroke or spinal cord injury (SCI) will have access to tools that will make them eager to exercise, resulting in improved function.

Researchers at Rutgers University in Piscataway, New Jersey, have designed systems for upper extremity rehabilitation after a stroke that link video games to a sensor on a glove, sleeve, or other item of clothing. Researchers at the New Jersey Institute of Technology (NJIT) in Newark also are designing virtual reality games for children with CP; one system involves the use of a robotic arm; another uses a camera aimed at the user's hands. Finally, researchers at the University of Michigan, Ann Arbor, are experimenting with various methods for rehabilitation for SCI, including treadmill training with a robotic exoskeleton.

VIDEO GAMES IN STROKE REHAB

In New Jersey, researchers from the Department of Electrical and Computer Engineering at Rutgers University and the University of Medicine and Dentistry of New Jersey (UMDNJ) in Newark are doing pioneering research on the use of video games for upper extremity rehabilitation after a stroke.

One rehabilitation system involves use of a video game with a sensing glove that detects movements of the patient's fingers or wrist. This combination creates a virtual reality environment where the user is able to control actions in a lifelike way by controlling a virtual hand.

"The idea is that we want to make rehabilitation fun, as opposed to what is happening today--which is boring," said Grigore Burdea, PhD, professor of electrical and computer engineering at Rutgers. "If you entertain the patients, and they want to exercise in the form of a game, then you are halfway there. This way, they do not mind exercising for a longer duration."

Burdea said that his team is working on low-cost games using the wired-glove system that can be played at home, which would save the expense of frequent visits to an outpatient clinic. Such systems have the potential to become networked to a health care facility. "In the future, you can envision a system in which the patient does some exercises at home and the doctor or therapist monitors them at a distance," he said.

An early version of the upper extremity system used a personal computer (PC) and a CyberGlove (Immersion Corp.) with exercises programmed in WorldToolKit. Trials in stroke patients with chronic symptoms performed at the UMDNJ produced encouraging results, and the patients enjoyed using the system.¹

"The patients were exercising for an hour, and they wanted to do more," Burdea said. However, at a cost of more than $17,000, the system was for experimental use only, being too expensive for easy clinical adoption.

In response, the team developed a system costing less than $600 that uses a modified Xbox gaming system (Microsoft) and a low-cost P5 glove (Essential Reality). The P5 glove does not have the capability to provide force feedback or to measure the individual joints of each finger, but it does measure flexion in the fingers. In addition, an infrared light-emitting diode mounted on the back of the hand measures the position of the wrist in relation to a base station tower.

At the 5th International Workshop of Virtual Rehabilitation, which took place this past August in New York City, the team described 2 exercises that they created for the P5 glove system. In the finger range-of-motion game, the patient begins by resting the hand flat and viewing 4 vertical bars of pixels that obscure a pleasant image, such as a hummingbird feeding on an exotic flower. Each bar corresponds to 1 finger. To reveal the image, the patient bends each finger and "scrapes off" the masking pixels.

In the finger velocity training game, the patient also begins by resting the hand flat. The screen depicts a corresponding hand representation with a butterfly that flies around its palm. The user is instructed to clench the hand into a fist as quickly as possible; the butterfly will "fly away" if the patient flexes the fingers with enough velocity to exceed a therapist-selected threshold.

The P5 glove has less accuracy and resolution than specialized equipment, such as the CyberGlove, but Burdea explained that it could provide home training that would supplement--not replace--training at a rehabilitation center.

"Therapists always become concerned about whether this is going to replace them," said Burdea. "The answer is no. . . . [the system] is picking up a lot of the menial jobs and empowering the therapist to see more patients and deal with them on a more intellectual level." Another advantage of these kinds of systems, said Burdea, is the ability to automatically get real-time data instead of having to write notes and type them into a computer.

Burdea said that the system will need further fine-tuning before it can be evaluated in clinical trials. The Xbox 3 and P5 glove are no longer produced, so newer systems would use components such as the Sony PlayStation. The team is currently focusing its attention on a system for rehabilitation of the shoulder after a stroke.

CEREBRAL PALSY REHAB

Not far away, researchers at the Rehabilitation Engineering Research Center at NJIT are working on several rehabilitation systems for children with CP, most notably, a robotic arm and an inexpensive video game system.

The robotic arm system pairs a commercially available force-controlled robotic arm with a 3-dimensional computer game. According to Richard Foulds, PhD, director of the center and associate chair for research in the NJIT Department of Biomedical Engineering, the goal is to get children with CP to initiate purposeful arm movements under robotic guidance, stimulating neurons to take over those movements.

The team is using the Haptic Master (Moog FCS Robotics, the Netherlands), a robotic arm that contains a gripper or trough where the hand and arm rest. The device has the ability to move up and down, slide in and out, and rotate, leading the arm through a normal range of motion.

During therapy, the user views a computer screen while wearing stereo glasses, creating a 3-dimensional effect. A red dot on the screen--which appears to be floating--corresponds with the location of the hand. The user is then given a task, such as pushing a button. As the user attempts to reach forward, the Haptic Master provides the arm with the force it needs to move. If the person moves off of the correct trajectory, the robotic arm will push the arm back on course. Users can also play a game of checkers or air hockey, with the checker piece or air hockey mallet linked to their arm movement.

Foulds emphasizes that the goal of the therapy is not to strengthen the muscles, but to strengthen neural connections. The keys to such rehabilitation are intention and repetition. "It's the intention that we believe will drive the growth of neural cells," he says. He predicted that with numerous repetitions, increasing numbers of neurons would take over the role of moving the limb.

The team has been designing the system with input from patients with CP who have hemiplegia; Foulds says that this represents about 10% to 20% of the CP population. The advantage of this approach is that the high-functioning arm can be used to define the normal range of motion, making it easier for the device to direct the less functional arm.

About half a dozen patients have used the device in preliminary studies. Foulds reported that the team expects to receive full institutional review board (IRB) approval this year to begin a 40-person randomized clinical trial in early 2007. "We're not seeing whether someone can play air hockey better," Foulds remarked. "The idea is whether someone can start making more functional use of their impaired limb."

The team also has designed an inexpensive rehabilitation system that combines simple computer games with a camera. The camera is aimed at the child's hands and tracks their movement in 2 dimensions: side-to-side and up-and-down.

He explained that the system is being designed to detect differences between the limbs and to adjust the behavior of the game accordingly. So, if the child is hemiplegic, the game will encourage the child to use the less functional arm by making it easier to score points with that arm. "We're writing the software ourselves, so we can create games that are therapeutically valuable," said Foulds.

He described one game that involves reaching out to swat flying saucers. Two shapes on the computer screen represent the user's 2 hands. The user can make the shapes move with small hand movements. A gentle swat will cause the saucer to move, and a hard swat will cause it to explode. The more explosions, the higher the score.

"We can teach the flying saucers to know which is the child's good arm and which is the child's bad arm," Foulds said. The game can be adjusted so that the flying saucers will gravitate toward the less functional arm and require less force to explode when that arm is used. "The idea is to encourage the use of the limb that is not encouraged in everything else the child does during the day," he said.

This game has been in development since early 2006, and the team is hoping to gain IRB approval to begin a 40-person trial in January 2007. After the team has produced some data, they plan to make the games available for download at no charge.

"We're also going to tell people how to design their own games and hopefully share them with other people," he said. "So many things for people with disabilities cost a fortune. This is one that ought to cost them nothing. You buy a $29 Web camera and put it onto your PC and just run the software," he said.

Foulds said that even if the game turns out to have no rehabilitative value, it will be valuable simply for allowing children with CP to play an enjoyable game, either alone or with someone else. The camera would sense the difference in hand speeds between the players and adjust the game speed accordingly. "I mean, how many kids with cerebral palsy can take part in something where they're competing against one of their siblings?" he said.

Another team at Rutgers has developed a stroke rehabilitation device called the hand-arm rehabilitation interface (HARI). The interface consists of a chair-mounted support that tracks and positions the arm, a sleeve that senses the force of muscle activity, and computer software that directs the user through specific protocols to promote range of motion and grip control. HARI provides feedback to the user and monitors daily progress.

"While it's important to work on restoring ambulation after a stroke, it's also important to focus on neurorehabilitation of the upper extremities," said project leader William Craelius, PhD, associate professor in the Department of Biomedical Engineering at Rutgers University and the Bioengineering Division of the UMDNJ. "Now, for the first time, we have an excellent tool to help therapists guide patients through this arduous task."

To play a video game, such as a simulation of racing a car, the user places an arm in the sleeve and then in the arm support. The sleeve measures muscle activity and gives biofeedback on wrist and finger motions. "Even if their effort doesn't result in actual joint movement, they can still see their residual muscle activity." Patients are scored on the basis of range of motion, consistency, stamina, and jerkiness. After patients have mastered the sensor sleeve, they can move on to using an attachment that measures the force of their grip. Craelius said that patients who exercise for an hour a day, several days a week, usually see their scores improve--which is often far more motivating than verbal encouragement from their physical therapist.

The team presented results on the gripper version of the device this past September at the American Congress of Rehabilitation Medicine/American Society of Neurorehabilitation Joint Annual Conference in Boston. The study included 20 patients with hemiparalysis due to stroke and was conducuted at the JFK Johnson Rehabilitation Institute (JRI) in Edison, New Jersey. The patients underwent therapy for 1 hour 2 or 3 times a week; 6 weeks of therapy was shown to be highly effective in improving upper limb function.² The system is now being used to treat patients at JRI.

Craelius said that the HARI will be commercially available in early 2007 from Nian-Crae, Inc in Somerset, New Jersey. The cost is expected to be below $25,000 for clinic use; patients will be able to lease the device. The team also has done some pilot studies using a bracelet version of the device in children with CP at Children's Specialized Hospital in Mountainside, New Jersey.

LEARNING TO WALK AFTER SCI

At the Human Neuromechanics Laboratory at the University of Michigan, researchers are studying the best ways to help persons with SCI learn how to walk again. Daniel Ferris, PhD, the laboratory's director, explained that his team is looking at 3 basic approaches: body weight-supported treadmill training with manual assistance, treadmill training with a robotic exoskeleton, and recumbent stepping. Although all 3 of these approaches are in use, including a limited number of robotic exoskeletons, it is unclear how effective they are. Ferris's group is looking at the effectiveness of each method by examining their physiologic effects on patients.

Traditional weight-supported treadmill training requires a harness to support the patient over a treadmill and a therapist to manually move the patient's hips and legs. The procedure is extremely tiring for therapists, and the patient's legs can be too spastic to move properly--which is why the use of robotic exoskeleton to move the legs is so appealing. However, there are two schools of thought on providing manual or robotic assistance. The first is that firm assistance is required to establish a normal walking pattern; the other is that minimal assistance should be provided so that patients do not become reliant on the assistance and can better learn to walk on their own.

"One of the focuses of our studies here is to determine exactly how much of a difference manual or robotic assistance plays," said Ferris. He added that initial studies indicate that patients continue to activate their muscles during both manual and robotic assistance and do not become passive participants as some have feared.^3,4 "The patients do indeed activate their legs and take part in a stepping process, even when you assist them," he said.

Recumbent stepping therapy is an appealing option for home use because it does not require expensive equipment or physical therapists. "It doesn't allow patients to practice walking exactly, but it allows patients to practice the basic motion of stepping." Ferris pointed out that patients with good use of their arms can be completely self-sufficient when using the device, "because they can grab ahold of the handles to assist their own legs in doing the motion."

Ferris said that his team has found, through computational analysis of the motor pattern during walking and recumbent stepping, that the spinal cord and brain appear to use the same neural networks to control both activities.⁵ "That's important, because it says that if you activate the neurons during recumbent stepping, it's probably the same pathways that you're using for walking. If they were completely different pathways, it would be like trying to learn how to play golf by practicing basketball."

He said that his group has been performing ongoing studies on the effect of upper limb movement on lower limb activation⁶ and will soon begin studying the effect of an hour-a-day, 5-day-a-week recumbent stepping program on walking patterns in patients with chronic SCI.

REFERENCES1. Merians AS, Jack D, Boian R, et al. Virtual reality-augmented rehabilitation for patients following stroke. Phys Ther. 2002;82:898-915.
2. Wininger MT, De Laurentis KJ, Morris TT, et al. A guided exercise tool with quantitative measures of rehabilitation progress for upper-limb motor control following stroke. Paper presented at: American Congress of Rehabilitative Medicine/American Society of Neurorehabilitation Joint Educational Conference; September 27-30, 2006; Boston.
3. Domingo A, Sawicki GS, Ferris DP. Kinematics and muscle activity of individuals with incomplete spinal cord injury during treadmill stepping with and without manual assistance. J Neuroengineering Rehabil. In review.
4. Sawicki GS, Domingo A, Ferris DP. The effects of powered ankle foot orthoses on muscle activation and joint kinematics during walking by individuals with incomplete spinal cord injury. J Neuroengineering Rehabil. 2006;3:3.
5. Stoloff RH, Zehr EP, Ferris DP. Recumbent stepping has similar but simpler neural control compared to walking. Ex Brain Res. In press.
6. Ferris DP, Huang HJ, Kao PC. Moving the arms to activate the legs. Exerc Sport Sci Rev. 2006;34:113-120.