HI Krebs1 N Hogan2 WK Durfee3 and HM Herr4

1 Department of Mechanical Engineering, Massachusetts Institute of Technology and Adjunct Assistant Research Professor of Neuroscience, The Winifred Masterson Burke Medical Research Institute, Weill Medical College of Cornell University, New York, USA;2Departments of Mechanical Engineering, and Brain and Cognitive Sciences, Massachusetts Institute of Technology, Cambridge, USA; 3Department of Mechanical Engineering, University of Minnesota, Minnesota, USA and 4 The Media Lab and The Harvard/MIT Division of Health Science and Technology, Cambridge, USA

12.1 Overview

One overarching goal drives our research and development activities: to revolutionize rehabilitation medicine with robotics, mechatronics, and information technologies that can assist movement, enhance treatment and quantify outcomes. In this chapter, we present three fronts of this revolution: rehabilitation robotics, orthotics, and prosthetics.

The first and newest approach, rehabilitation robotics, has grown significantly in the last 10 years (cf. special issue of the Journal of Rehabilitation Research and Development, 37(6) of Nov/Dec 2000; International Conference on Rehabilitation Robotics -ICORR 2001, 2003 and 2005). Previously, robotics were incorporated into assistive devices to help the physically challenged accommodate their impairment. Rehabilitation robotics, by contrast, fashions a new class of interactive and user-friendly robots that enhance the clinicians' goal of facilitating recovery by not only evaluating but also by delivering measured therapy to patients. Krebs and Hogan review pioneering clinical results in the field, discuss the growing pains of forging a novel technology, and outline the potential for a brilliant future.

Of the other two activities, we will limit our discussion to mechatronic systems. Orthotics and pros-thetics may be considered as a category of assistive robotics. While the previous high water mark for mechatronic assistive technology occurred during the Vietnam War decades of 1960s and 1970s, recent advancements in materials, computers, and neuro-connectivity (neuro-prostheses) have reinvigorated research in this field. In fact, the lack of equivalent advances in realm of energy storage represents the only major hurdle preventing the realization of practical versions of Hollywood's fancies such as Star Trek's Commander Data or the Terminator. Durfee reviews pioneering developments in orthotics, Krebs and Hogan review upper-limb prostheses, and Herr reviews lower extremity prostheses. We will also discuss some emerging developments that could render some science fictions into reality.

12.2 Rehabilitation robotics

Rehabilitation robotics encompasses an emerging class of interactive, user-friendly, clinical devices designed to evaluate patients and, also deliver therapy. Robots and computers are being harnessed to support and enhance clinicians' productivity, thereby facilitating a disabled individual's functional recovery. This development represents a shift from earlier uses of robotics as an assistive technology for the disabled. The new focus on mechanisms of recovery and evidence-based treatment together with developments facilitating safe human-machine interaction has paved the way for the surge in academic research, which started in early 1990s.

We can group devices into two main categories for the upper and lower extremity. For upper, extremity, Erlandson et al. (1990) described a patented robotic "smart exercise partner" in which the recovering stroke patient executes general spatial motions specified by the robot. Positive results using that system in a clinical setting were reported (Erlandson, 1995). However, patients had to be capable of moving independently or using the contralateral limb to move and guide the impaired limb (self-ranging). Independent movement is also essential to use of Rosen's 3-D controllable brake device (Maxwell, 1990), Rahman's functional upper limb orthosis (Rahman et al., 2000), and Burdea's pneumatically actuated glove (Merians et al., 2002). Other upper extremity robotic tools differ insofar as they do not require patients to be capable of independent movement; controlled forces can be exerted to move the patient or to measure aspects of motor status such as spasticity, rigidity or muscle tone. Lum et al. (1993, 1995) described the design and application of robotic assistive devices focused on bi-manual tasks to promote motor recovery. More recently, Lum et al. (Burgar et al., 2000; Reinkensmeyer et al., 2000a; Lum et al., 2002) used a commercial PUMA robot augmented by improved sensors to implement the Mirror Image Movement Enabler (MIME) system, in which the robot moves the impaired limb to mirror movements of the contralateral limb. Harwin et al. (2001) is using another commercial robot (Fokker) to move the impaired arm in the recently initiated European Union sponsored project. While attempts to adapt or re-configure industrial robots for use in rehabilitation robotics appears to be a reasonable approach it suffers from a critical drawback: some 20 years of experience with industrial robots shows that low impedance comparable to the human arm cannot practically be achieved with these machines (intrinsically high impedance machines). In contrast to these approaches other groups have developed robotic technology configured for safe, stable and compliant operation in close physical contact with humans. For example, the MIT-MANUS robot developed in the Newman Laboratory for Biomechanics and Human Rehabilitation was specifically designed for clinical neurologic applications and ensures a gentle compliant behavior (Hogan et al., 1995). Other low-impedance rehabilitation devices are Reinkensmeyer's ARM Guide (2000) and Furusho's EMUL (2003). Operationally, these robots can "get out of the way" as needed. They can therefore be programmed to allow the recovering stroke survivor to express movement, in whole or in part, even when the attempts are weak or uncoordinated. Whether this feature is crucial for effective therapy remains unproven but its importance for obtaining uncorrupted measurements of a patient's sensorimotor function has been established unequivocally (Krebs et al., 1998, 1999a; Reinkensmeyer et al., 2002; Rohrer et al., 2002).

Evolving lower extremity devices are inspired mainly by gym machines and orthoses rather than byre-configured industrial robots. The best examples are Hesse's Elliptical Gait Trainer, Yaskawa's therapeutic exercise machine (TEM), and the Lokomat (Sakaki et al., 1999; Colombo et al., 2000; Hesse and Uhlenbrock, 2000; see Volume II, Chapters 3 and 19). Not unlike their upper-extremity industrial robot counterparts, however, these designs suffer from a high impedance drawback. Most are presently being re-designed to modulate their impedance and to afford an interactive experience similar to the low-impedance lower extremities devices under development at MIT in the Newman Laboratory and at University of California Irvine.

Accompanying the vigorous development of rehabilitation robotic devices is an equivalent growth in clinical evaluations of device performance. Results include new insights into the recovery mechanisms for a variety of conditions, and into the rehabilitation techniques that best engage those mechanisms. Areas of research focus include not only stroke, but also motor deficits associated with diverse neurologic, orthopedic, arthritic conditions. A number of studies demonstrated the exciting opportunities and benefits of integrating robotic technology into patients' daily rehabilitation program (e.g., Aisen et al., 1997; Krebs et al., 1998, 2000; Volpe et al., 1999, 2000, 2001; Burgar et al., 2000; Colombo et al., 2000; Hesse and Uhlenbrock, 2000; Reinkensmeyer et al., 2000, 2002; Lum et al., 2002; Fasoli et al., 2003,2004; Ferraro et al., 2003).

Figure 12.1. Rehabilitation robot modules during clinical trials at the Burke Rehabilitation Hospital (White Plains, NY). (a) and (b) show the shoulder and elbow robot (MIT-MANUS) and the wrist robot, and (c) the anti-gravity spatial module.

This chapter presents only our own results to delineate the potential of the technology and future directions for rehabilitation robotics. To date, we have deployed three distinct robot modules in collaborating clinical institutions1 for shoulder and elbow, wrist, and spatial movements (Fig. 12.1).

Volpe (2001) reported results of robotic training with 96 consecutive inpatients admitted to Burke Rehabilitation Hospital (White Plains, NY) who met inclusion criteria and consented to participate. Inclusion criteria were diagnosis of a single unilateral stroke within 4 weeks of admission to the study; the ability to understand and follow simple directions;

1 Hospitals presently operating one or more MIT-MANUS class robots include Burke (NY), Spaulding (MA), Helen Hayes (NY), Rhode Island (RI) Rehabilitation Hospitals, and the Baltimore (MD) and Cleveland (OH) Veterans Administration Medical Centers.

and upper limb weakness in the hemiparetic arm (i.e., a strength grade of 3/5 or less in muscle groups of the proximal arm) as assessed with the standardized Medical Research Council battery. Patients were randomly assigned to either an experimental or control group. The sensorimotor training for the experimental group consisted of a set of "video games" in which patients were required to move the robot end-effector according to the game's goals. If the patient could not perform the task, the robot assisted and guided the patient's hand. The sensorimotor training group received an additional 4-5 h per week of robot-aided therapy while the control group received an hour of weekly robot exposure.

Although patient groups were comparable on all initial clinical evaluation measures, the robot-trained group demonstrated significantly greater motor

Table 12.1. Mean interval change in impairment and disability measure for inpatients (significance P < 0.05). For all evaluations higher scores indicate better performance. MP was only evaluated for shoulder and elbow movements.

Between group comparisons: final Robot-trained Control evaluation minus initial evaluation (n = 55) (n = 41) P-value

Impairment measures (±SEM)

Between group comparisons: final Robot-trained Control evaluation minus initial evaluation (n = 55) (n = 41) P-value

Impairment measures (±SEM)

Fugl Meyer shoulder/elbow (FM-se) max/42

Was this article helpful?

0 0
Losing Weight Without Starving

Losing Weight Without Starving

Tired of Trying To Loose Weight And It Never Works or You Have To Starve Yourself Well Here's A Weight Loss Plan That takes Care of Your Weight Problem And You Can Still Eat. In This Book, You’ll Learn How To Lose Weight And Not Feel Hungry! In An Easy Step-By-Step Process That Enables You To Feel Good About Loosing Weight As Well As Feeling Good Because Your Stomach Is Still Full.

Get My Free Ebook

Post a comment