2000 Conference Proceedings

Go to previous article 
Go to next article 
Return to 2000 Table of Contents

An Internet-Based, Force-Feedback Rehabilitation System for Arm Movement after Brain Injury

David Reinkensmeyer, Chris Painter, Stanley Yang, Erin Abbey, Brian Kaino
Department of Mechanical and Aerospace Engineering
Center for Biomedical Engineering
4200 Engineering Gateway
University of California, Irvine, 92697-3975
Email: dreinken@uci.edu

We describe a computer-based system for providing arm movement therapy to people with brain injury. The system uses a force feedback joystick to apply therapeutic patterns of force to the arm as the user plays video games.It operates over the Internet using Java, and can track user movement recovery over time and report it to a remote location.The system can also be customized to provide a personalized program of therapeutic exercise.Initial use of the system by three individuals with stroke is reported.


In the U.S. each year, over 400,000 people lose strength, coordination, and sensation in the arm and hand due to stroke or traumatic brain injury. With practice, however, they often improve their ability to reach, grasp, and manipulate objects. Significant improvements are possible even several years after brain injury [1]. Physical and occupational therapy are an essential part of recovery, and typically include hands-on manipulations, such as manually assisting in desired arm movements.These manipulations are designed to increase sensory input and stimulate regeneration in the nervous system. Sensory motor skills improve with increasing levels of therapy [2].

Unfortunately, because of economic pressures on the U.S. health care system, individuals with a brain injury are going home sooner and getting less hands-on therapy. Compounding the problem, there is a current lack of technology available for practicing self-therapy at home. To address these needs, several research groups, including our own, are developing robotic devices that physically interact with people to stimulate the senses of touch and movement, and to retrain coordinated movement [3,4]. While preliminary clinical trials are promising for both improved evaluation [5-7] and therapy [8,9], key practical problems with these devices are their cost and size, which may ultimately limit their accessibility and practicality, especially for home use.

The advent of dynamic force feedback technology for the PC, coupled with the networking power of the Internet, could solve these problems. Commercial force feedback joysticks designed primarily for gaming applications can not only sense a person’s movement, but can also apply forces during movement. Like existing robotic therapy devices, such devices could be used to stimulate the sense of touch and movement, and could apply therapeutic patterns of forces to the hand and arm as the user attempts to move. Unlike larger robotic devices, however, force feedback joysticks could become truly accessible personal movement trainers because they are already in mass production and can be purchased at low cost. By networking them to rehabilitation centers through the Internet, such devices could provide a means for an individual with a brain injury to access a personalized program of therapeutic exercises, customized by a rehabilitation expert. Also, networking could provide a means for the rehabilitation expert to track the user's sensory motor performance while the user stays at home. Force feedback movement therapy could thus become a viable aspect of telerehabilitation services [10].

This paper describes the design and initial testing of an Internet-based, force-feedback physical rehabilitation system for training and assessing arm movement after brain injury. The mechanical and software design of the system are described first. Then, initial use of the system by three individuals with stroke is reported. In conclusion, future development of the system is described.

Mechanical Design

The Microsoft Sidewinder Force Feedback Pro Joystick was used as the force feedback platform for this project. The joystick required two mechanical modifications in order to make it usable by individuals with a brain injury:

Handle Re-Design: Many individuals with brain injury lose dexterity of the fingers, making grasping difficult. Although the Microsoft handle is ergonomically designed, individuals with brain injury have difficulty in securely grasping it. To address this problem, a new handle was designed that clips on to the preexisting handle. The initial step in creating the new handle design was to survey seven individuals with a brain injury to determine the criteria for a basic shape that would comfortably hold the hand. Many prefabricated splints, made out of plastics that can be molded when heated, are available for brain-injured people with spastic hands. Four of these splints (conical, U-shaped, ball, anti-spasticity) were tested by seven individuals with brain injury to determine the best splint with regard to comfort and accessibility. The volunteers who were surveyed agreed that the splint with a small conical radius grip, and the splint with a U-shaped grip were the most comfortable and the easiest to hold. It was determined that by slightly modifying the U-shaped splint it could be made to mate securely with the joystick handle without using fasteners, and it was thus possible to simply clip it on to the joystick. A Velcro strap was added over the back of the hand to help keep the hand engaged to the splint. The resulting design (see figure 1 below) holds the hand comfortably and securely, and allows users with one impaired hand to place their impaired hand in the handle using their unimpaired hand. A similar clip-on handle was designed for the left hand, making the joystick usable by individuals with impairment on either side.

Arm Support System: The Microsoft Sidewinder Force Feedback Pro Joystick can exert a peak force of approximately 1 Kg, which is not enough to move the arm against gravity. A commercially available articulating armrest was thus integrated into the system (Ergorest, Oscar Dellert AB, Finland). The armrest, which was designed to support the arm during computer keyboard usage, consists of a forearm tray connected to a low-friction, three-joint mechanical linkage. It allows the forearm to be positioned and oriented arbitrarily in the horizontal plane and is height adjustable.

To hold the joystick and armrest, a wood base with a cutout in the shape of the joystick base was made, and a protruding wood beam was attached to the base for clamping the arm support at the appropriate height. To use the system, the wood base is clamped to a table, the joystick is placed inside the cutout to keep it from slipping, and the armrest is adjusted to support the user's arm (see Figure 2 below).

The resulting system allows comfortable movement of the hand in a workspace of approximately six by six centimeters in the horizontal plane.

Software Design

Design Philosophy: Two approaches to the software design were considered. The first was to develop a stand-alone program that could be installed on a PC, to control the joystick and to transfer information through a temporary network connection. The second approach was to incorporate force feedback into a web page using a Java applet, and to use the web page to transfer desired information. The second approach was chosen for two reasons. First, since Java applets are downloaded onto the host computer each time a web page is accessed, new versions of the therapy program could in effect be automatically installed each time a user accessed the therapy web page. This was attractive not only for facilitating rapid modifications during the early development of the system, but also eventually for providing a means for personalization of therapy plans.In effect, Java therapy applets designed at a remote center for each user could be downloaded "on the fly". The second advantage was the platform independence of Java. Using any Java-enabled web browser, users on a variety of computing platforms, with a variety of input devices, could access the therapy program, running it on their own processor. It should be noted, however, that device and platform independent standards for force feedback have to our knowledge not yet been implemented, and thus the force feedback system described here works currently only with the Microsoft joystick on a PC using Internet Explorer.

In order to control the joystick through a Java applet, we used FEELtheWEB software developed by Immersion Corporation. FEELtheWEB is an ActiveX control for Internet Explorer that can apply forces to the joystick and receive function calls through HTML.

Prototype Therapy Game: A prototype therapy game with force feedback was developed to test the system. The user begins by logging into the system home page using a personal username and password. The user is then taken to a therapy page that downloads a Java applet containing the therapy game onto his or her computer. In the therapy game, the user is asked to move a circular cursor into a square target with the joystick as quickly and accurately as possible. After a delay, if the user has not yet moved into the target, the motors on the joystick activate and assist the hand in moving into the target. This continues for a predetermined number of attempts with the target changing positions to span a 360-degree range of movement directions. For each attempt, the game tracks and records a variety of movement parameters, including the initial and final joystick position, the time when the joystick forces turned on, the total time to complete the trial, and the distance (in pixels) the cursor moved before and after the forces were activated. Upon completion of the game, the data is posted to the server computer.

Preliminary Testing

Three people with hemiparesis resulting from stroke tested the joystick system at a local outpatient rehabilitation center.The Institutional Review Board of UCI gave approval for the experiments, and the subjects provided informed consent. Subjects 1, 2, and 3 were aged 53, 32, and 52, and were 5, 2, and 4 years post-stroke respectively. They had different levels of arm impairment, classifiable respectively as mild, moderate, and severe according to a clinical scale. The subjects played a modified version of the therapy game, moving the circle cursor into the target squares in a series of ninety-six trials, with joystick forces assisting for half the trials and deactivated for the other half.The delay between presentation of a new target and activation of force feedback was set to 0.5 seconds in the assisted trials.

It was found that the forces helped the subjects move more quickly into the target, with the greatest improvement shown for the most impaired subject (see Figure 3 below). These results demonstrate that the joystick system is strong and accurate enough to assist in moving impaired arms.

The same three subjects are currently using the system several times a week to practice arm movement.By analyzing data posted to our server, changes in the subjects’ movement ability with time can be tracked (see Figure 4 below).

The subjects have used the system to practice hundreds of movements over three weeks, although it is too early to draw conclusions about the efficacy of the therapy.

Discussion and Future Directions

The force feedback therapy system described in this paper has several advantages: It is ergonomic for individuals with a brain injury. It does not require ability to grasp, and an individual with a one-sided impairment can attach his or her hand to it without assistance. It is strong and accurate enough to assist in moving impaired arms to desired locations, provided the arm is supported in an armrest. It is inexpensive. The joystick and arm support currently retail for around $100. Robotic systems that could achieve similar functionality currently cost tens of thousands of dollars. It can track user movement recovery over time, storing performance data on a remote server. Its software is rapidly modifiable and customizable. Its software is structured to be platform independent (currently for non force-feedback pointing devices, and in the future, for force feedback devices). Software Development Plans: Software is currently being developed to display online measures of the user's movement performance as a function of movement practice. This software will allow users to self-monitor movement recovery. For example, a person who has just suffered a stroke, and has been released from the hospital, will be able to logon daily from home, practice moving, and determine if his or her arm movement ability has changed. In addition, this capability will ultimately allow remotely located rehabilitation professionals to evaluate users' progress and to make therapy recommendations. Force feedback is also being incorporated into other games besides the simple prototype game described in this paper. A wide range of freeware Java games are available on the Internet, such as card games, sports simulations, and arcade-style games. A selection of these games is being modified to include force feedback that physically assists in playing the game.Other force-enabled games targeted at training specific deficits in sensory motor performance common after brain injury are also being designed.

Future Directions: Several key questions for future research are:

To what extent does practicing therapy with the system improve arm movement ability? The system described here provides a powerful new tool for identifying the relationship between movement therapy and recovery. It is hoped that the system can be distributed to a large number of users, and that this relationship can be characterized for different types of brain injuries at different times post injury. What are the optimal movement therapy techniques? Our initial efforts have focused on using force feedback to assist in movement. However, it is unclear whether such assistance, as opposed to movement practice without assistance or, alternately, with resistance, is the optimal therapy technique. The joystick system provides a means to test what therapy techniques work best for which user groups.

Can the system be applied to rehabilitation of other user populations? With appropriately designed software, people with cerebral palsy, spinal cord injury, Parkinson's disease, orthopedic injuries, and cognitive impairments may also benefit from the Internet-based, force feedback therapy.


The authors gratefully acknowledge the support of the Microsoft Foundation, through a New Discoveries Grant, and the assistance of Richard Grauman and the staff and clients of the Rehabilitation Institute of Southern California.


Taub E, Wolf SL. Constraint induced movement techniques to facilitate upper extremity use in stroke patients. Topics in Stroke Rehabil 1997;3(4):38-61.

Sunderland A, Tinson DJ, Bradley EL, Fletcher D, Langton Hewer R, Wade DT. Enhanced physical therapy improves recovery of arm function after stroke. A randomized controlled trial. J. Neurology, Neurosurgery, and Psychiatry 1992;55:530-535.

Reinkensmeyer DJ, Dewald JPA, Rymer WZ. Robotic devices for physical rehabilitation of stroke patients: Fundamental requirements, target therapeutic techniques, and preliminary designs. Technology and Disability 1996;5:205-215.

Reinkensmeyer DJ, Hogan N, Krebs HI, Lehman SL, Lum PS. Rehabilitators, robots, and guides: New tools for neurological rehabilitation. Biomechanics and Neural Control of Movement, J. Winters and P. Crago, eds., Springer-Verlag. 1999, in press

Reinkensmeyer DJ, Dewald JPA, Rymer WZ. Guidance based quantification of arm impairment following brain injury: A pilot study. IEEE Transactions on Rehabilitation Engineering, 1999;7(1): 1-11.

Lum PS, Burgar CG, Kenney D, and Van der Loos, HFM. Quantification of force abnormalities during passive andactive-assisted upper-limb reaching movements in post-stroke hemiparesis. IEEE Transactions on Biomedical Engineering, 1999; 46(6):652-662

Krebs HI, Aisen ML, Volpe BT, Hogan N. Quantization of continuous arm movements in humans with brain injury. Proc. Nat. Acad. Sci USA, 1999: 96, 4645-4649.

Reinkensmeyer DJ, Schmit BD, Rymer WZ. Assessment of active and passive restraint during guided reaching after chronic brain injury. Annals of Biomed Engineering, 1999; 27(6):in press.

Aisen ML, Krebs HI, Hogan N, McDowell F, Volpe BT. The effect of robot-assisted therapy and rehabilitative training on motor recovery following stroke. Archives of Neurology 1997;54:443-445.

Lum PS, Van der Loos HFM, Shor P, Burgar CG. A robotic system for upper-limb exercises to promote recovery of motor function following stroke. Proceedings Sixth Int. Conf. on Rehab. Robotics, 1999; 235-239. Rosen MJ. Telerehabilitation. NeuroRehabilitation,1999;12:11-26.

Go to previous article 
Go to next article 
Return to 2000 Table of Contents 
Return to Table of Proceedings

Reprinted with author(s) permission. Author(s) retain copyright.