Available online at www.sciencedirect.com
Medical Engineering & Physics 30 (2008) 1387–1398
A perspective on intelligent devices and environments in medical rehabilitation夽 Rory A. Cooper a,b,∗ , Brad E. Dicianno a,b,c , Bambi Brewer a,b , Edmund LoPresti b,d , Dan Ding a,b , Richard Simpson a,b,c , Garrett Grindle a,b , Hongwu Wang a,b a
Human Engineering Research Laboratories, Department of Veterans Affairs, Rehabilitation Research and Development Service, VA Pittsburgh Healthcare System, United States b Departments of Rehabilitation Science & Technology, Bioengineering, and Physical Medicine & Rehabilitation, University of Pittsburgh, United States c Center for Assistive Technology, University of Pittsburgh Medical Center, United States d AT Sciences, Inc., United States Received 6 May 2008; received in revised form 24 September 2008; accepted 25 September 2008
Abstract Globally, the number of people older than 65 years is anticipated to double between 1997 and 2025, while at the same time the number of people with disabilities is growing at a similar rate, which makes technical advances and social policies critical to attain, prolong, and preserve quality of life. Recent advancements in technology, including computation, robotics, machine learning, communication, and miniaturization of sensors have been used primarily in manufacturing, military, space exploration, and entertainment. However, few efforts have been made to utilize these technologies to enhance the quality of life of people with disabilities. This article offers a perspective of future development in seven emerging areas: translation of research into clinical practice, pervasive assistive technology, cognitive assistive technologies, rehabilitation monitoring and coaching technologies, robotic assisted therapy, and personal mobility and manipulation technology. Published by Elsevier Ltd on behalf of IPEM Keywords: Rehabilitation; Intelligent systems; Machine learning; Physical impairment; Wheelchairs; Cognitive impairment
1. Introduction There is a large and growing segment of our world population—people with reduced functional capabilities due to aging or disability. The number and percentages of people in need of advanced assistive technology are increasing every year. About 60 million Americans have a disability that affects one or more of their major life activities [1]. Perceptive, cognitive, and musculoskeletal diseases that impair
夽 Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the National Science Foundation. ∗ Corresponding author at: Human Engineering Research Laboratories (151-R1), 7180 Highland Drive, VA Pittsburgh Healthcare System, Pittsburgh, PA 15206, United States. Tel.: +1 412 365 4850; fax: +1 412 365 4858. E-mail address:
[email protected] (R.A. Cooper).
1350-4533/$ – see front matter. Published by Elsevier Ltd on behalf of IPEM doi:10.1016/j.medengphy.2008.09.003
motor skills dramatically increase with age. A number of subpopulations are of particular interest. In 2030, over 20% of the U.S. population will be over 65 years of age, with one in two working adults serving as informal caregivers [1]. Globally, the number of people older than 65 years is anticipated to double between 1997 and 2025. There is little debate that the 76 million American children born between 1945 and 1964 represent a cohort that is significant on account of its size [1]. Boomers account for about 39% of Americans over the age of 18 and 29% of the total population [1]. Adults with disabilities comprise approximately 21,455,000 of the 169,765,000 of working-age individuals in the US. However, only 30% of adults with disabilities are employed [2]. In Japan, the percentage of people of the age of 65 is also on the rise and it is project that by 2030 that approximately 30% of the population will be over 65 [3]. In Europe it is projected that by 2060 that 30% of the population will be over 65 [4]. As individuals, families, communities, and a planet, we are facing
1388
R.A. Cooper et al. / Medical Engineering & Physics 30 (2008) 1387–1398
new technical and social challenges to attain, prolong, and preserve quality of life. Recent advancements of technologies, including computation, robotics, machine learning, communication, and miniaturization of sensors bring us closer to futuristic visions of compassionate intelligent devices and technologyembedded environments. While many intelligent systems have been developed, most of them are for manufacturing, military, space exploration, and entertainment. Their use for improving health-related quality of life has been treated as a specialized and minor area. Assistive technology, for example, has fallen in the cracks between medical and intelligent-system technologies. The missing element is a basic understanding of how to relate human functions (physiological, physical, and cognitive) to the design of intelligent devices and systems that aid and interact with people. The purpose of this manuscript is to highlight some of the emerging research topics in medical rehabilitation that should be possible because of advances in technology. It is not intended as an exhaustive review of assistive technology, but rather to provide a perspective into some of the requirements, challenges, and possibilities of future assistive technology.
2. Translation of research into clinical practice Rehabilitation engineering research is generally conducted within the scope of health professions, basic science, and engineering programs [5]. Rehabilitation engineers (RE) may define their occupational roles as primarily involving clinical care and service delivery, design and development, or research, or may be involved in a combination of these activities [6]. The field of rehabilitation engineering integrates clinical care and research, allowing each to influence the future direction of the other. Obtaining reimbursement for advanced medical equipment and technology can be challenging, especially if there is insufficient scientific evidence to justify the utility, efficacy, durability, or impact on function that a device has. Research can fill the gaps in scientific knowledge that aids consumers and clinicians in selection of appropriate devices and treatments and lends credence to why they may be a necessary part of medical care. For instance, insurance policies on reimbursement for power seat functions for power wheelchairs often have little or no basis in the scientific literature [7]. To this end, research that adds to the scientific knowledge in this domain aids not only clinical decision-making but also justification of those choices. For example, the effects of power features such as tilt-in-space and recline on seating interface pressures have been a major topic of interest [8–11]. Alternatively, rehabilitation engineering research fuels the development of evidence-based practice. The aforementioned studies on power seat functions have contributed significantly to the development of a clinical position paper from Rehabilitation Engineering and Assistive Technology Society of North America (RESNA) [12]. Likewise, well-
designed biomechanics studies on manual wheelchair users have elucidated the biomechanical forces, transfer techniques, and propulsion patterns that are significant sources of upper limb pain, injury, and disability; these studies have helped to formulate clinical practice guidelines that are now standards of medical care in spinal cord injury [13]. Technology transfer is the process of developing practical applications for the results of scientific research. Funding for research and development may be the largest and most significant obstacle. Often, partnerships between academic institutions and industry partners are then necessary to bring technology to the market. Scientific information must be published and disseminated. Finally, consumers and clinicians must have access to this information and be able to incorporate it into their lives and clinical practice [5].
3. Pervasive assistive technology A dilemma faced by clinicians is reconciling the fact that observations of patients only occur during infrequent, face-to-face meetings in a clinic or laboratory, while ideal observations are assessments that reflect the patient’s capabilities in the real world, where distractions are present and multi-task performance is often required. As such, there is a need for ecologically valid tests that provide information about a person’s ability to function in a real-life environment [14]. One way to obtain ecologically valid measures is through the use of ubiquitous computing. Sensors integrated into the patient’s environment as part of a “smart home” can allow healthcare professionals to obtain a much clearer view of the patient’s condition than is available from short periods of monitoring in a clinical setting [15]. “Pervasive healthcare technology” offers the potential for continuous measurement, processing and communication of physiological and physical parameters from patients to service providers, family and other support people [16–20]. Technological advances in application-specific integrated circuits (ASICs), battery capacity and wireless technology have resulted in reduced size of medical sensors and systems, the ability to communicate wirelessly, and the ability to operate on batteries for prolonged periods of time [15,21]. Devices can therefore be integrated into the patient’s environment, or even the patient’s clothing, allowing healthcare providers to obtain a much clearer view of the patient’s condition than is available from short periods of monitoring in the hospital or doctor’s office [15] The appropriate design and integration of different kinds of sensors, as well as the appropriate medical algorithms to process the data could offer new possibilities for preventing health risks [15,18,22], managing chronic diseases [16,18,19,23–27], and providing support to elders and people with disabilities living independently [16,19,22,26,28,29]. Pervasive healthcare technology may allow “smart homes” to identify changes in health status more quickly
R.A. Cooper et al. / Medical Engineering & Physics 30 (2008) 1387–1398
[30] by detecting symptomatic performance failures or deteriorating behavioral patterns. This is particularly important because these patterns often occur only under specific conditions that are not easily replicated in the clinic (e.g., exiting and returning to the kitchen too many times, in confusion, when preparing a meal, because of an idiosyncratic configuration of items in the refrigerator; failing to take medications correctly, but only when the television is on or the telephone rings). These types of subtle failures may be noted most during the early stages of a disease process, or actually remain undetected because the resident is still able to recover from or mask the failure. In these situations, family members or other observers must typically serve as informants. An application of pervasive healthcare technology that is already being explored is the study of falls and fall prevention [31–34]. Pervasive healthcare technology may be used to detect when a fall occurs, along with the causes of the fall. Even more importantly, pervasive healthcare technology can detect falls that occur without consequences, which may be forgotten and go unreported, even though it might have clinical relevance. Pervasive healthcare technology may also be used to document spontaneous compensatory strategies. This is particularly important for individuals who are able to function effectively in the home because it is familiar (e.g., avoiding falls as long as the furniture is never moved), but who may experience difficulty during clinical evaluation. In such cases, information about the conditions under which an individual avoids error, may allow caregivers and clinicians to postpone more disruptive interventions, such as placement in a supervised setting or increasing attendant care. Pervasive healthcare technology may also provide information to caregivers about changes that can be made in the home to promote increased use of compensatory strategies that might not have otherwise been considered (e.g., improving accurate medication compliance by disabling the television until a certain time; changing the furniture pattern in a room where repeated performance errors are noted, to a layout where such errors never occur). An example of a pervasive healthcare technology intended for use in the home is the Independent LifeStyle Assistant (ILSA) [35,36], which passively monitors the behavior of inhabitants and can alert caregivers in the event of an emergency (e.g., a fall). ILSA uses machine learning techniques to identify mappings between sensor readings and activities [35,36]. This allows ILSA to raise alerts when unlikely sensor readings are observed. ILSA is an agent-based system, with separate agents for monitoring medication use, monitoring mobility, providing scheduled reminders, and learning patterns of behavior. Each agent relies on sensors and actuators, which may be shared with other agents, and is responsible for situation assessment within its domain area. Each agent also maintains a library of domain-specific plans for (1) recognizing user intent and (2) choosing system responses. The development and evaluation of ILSA demonstrated the potential utility of monitoring technology, but also illus-
1389
trated the importance of false alarms. In a system composed of (fallible) sensors operating without human intervention, false alarms are inevitable and represent a significant hurdle to consumer adoption. In the case of ILSA, excessive “no motion” alerts were a source of significant aggravation for older adults and their caregivers [37]. Any system that is deployed in the real world will need to strike a balance between sensitivity and false alarms when identifying events of interest. While a significant amount of effort is being devoted to the development of instrumentation, much less attention has been paid to developing techniques to analyze and incorporate this data into clinical practice [38–43]. Simply put, our ability to collect data greatly exceeds our ability to utilize data to make clinical decisions [44]. The greatest obstacle to incorporating quantitative sensor data into clinical practice is the amount of data that sensors can produce, which can quickly become overwhelming. What is needed; is not just tools for analyzing sensor data but tools which allow clinicians to integrate large amounts of quantitative sensor data with the limited qualitative data obtained during clinical visits. Of course, as this technology develops, so does the need for integrating ways to protect privacy.
4. Cognitive assistive technologies More than 21 million persons living in the United States have a cognitive disability and this number is expected to increase rapidly as the nation’s population ages [45]. A much greater number of people worldwide have cognitive impairment, and the numbers are growing rapidly as the percentage of the world population ages. Cognitive impairment is a substantial limitation in one’s capacity for mental tasks, including conceptualizing, planning, sequencing thoughts and actions, remembering, interpreting subtle social cues, and manipulating numbers and symbols [46,47]. The diagnostic and statistical manual of mental disorders (DSM-IV) defines a person with cognitive disability as one who is “significantly limited in at least two of the following areas: self-care, communication, home living, social/interpersonal skills, self-direction, use of community resources, functional academic skills, work, leisure, health and safety” [48]. Cognitive disabilities may be developmental (intellectual disability, autism, learning disability) or acquired later in life (traumatic brain injury, stroke, and dementia). People with cognitive impairments often have difficulty with executive functions and prospective memory, including such tasks as organizing a schedule, initiating activities, and remembering to perform the appropriate task at the correct time. External cueing systems can assist people with cognitive disabilities by reminding them to perform a task at the appropriate time or by providing guidance through a task [49–51]. Such systems can range from low-tech, paper-andpencil solutions, to mainstream voice recorders and PDA’s,
1390
R.A. Cooper et al. / Medical Engineering & Physics 30 (2008) 1387–1398
to specialized software designed for people with cognitive impairments [50–52]. A device may be able to provide more appropriate or efficient reminders if it is able to detect a person’s context; for example his or her location, the state of a task he or she is performing, or the presence of other people [53,54]. Such a system might monitor a scheduled task to determine whether a person is having difficulties or needs reminders, or could continuously monitor a person’s actions in order to deduce what activity the person is attempting, as well as what help would be appropriate. Researchers have developed systems which can adapt cueing strategies according to a user’s preferences and past performance [54,55], resulting in improved performance of activities of daily living [54]. To recognize information about activities which people are performing, researchers have utilized video data [56], accelerometers [57], and radio-frequency identification (rfID) tags [58]. A special case of context-aware task guidance is way finding assistance; that is, using sensor information about a person’s location to assist the person in navigating through his or her environment [59]. Indoor environments can be instrumented with sensors such as radio-frequency beacons while the global positioning system (GPS) can be used in some outdoor environments. Using knowledge of the person’s location, activities, and intent, a system could provide cues appropriate to that location or provide prompts to help the person navigate to another location. In addition to providing prospective memory support (e.g., regarding upcoming tasks to be performed), technology can provide assistance with retrospective memory (e.g., regarding past events). Such a system could help orient someone with dementia to recent events, reduce the incidence of recurring questions directed at caregivers, or provide the basis for conversations with family members, peers, or professional caregivers [60]. Such a system could utilize continual monitoring of the person’s activities combined with some combination of automatic data mining to identify significant events and a user interface to assist the user and a caregiver to select and label events [60]. Reminiscence aids could also serve to foster conversations about events in the more distant past, about which a person with dementia may have more intact memories once the memories are triggered [61]. Assistive technology can also support decision-making, leading to greater self-determination. Researchers are developing technologies to support decision-making in vocational and personal health domains [62]. Adequate and high quality caregiver support is a critical issue given the scarcity and high turnover in professional caregivers for people with cognitive disabilities, and the physical and emotional demands often experienced by family members who care for people with cognitive disabilities. It is therefore necessary to understand when and how to engage the caregiver, and the information needed to effectively resolve problems [63]. Cognitive telerehabilitation technologies can support caregivers in setting up tasks and schedules, monitoring clients’ progress, or interacting with the client in an emergency, even if the caregiver and
client are geographically separated [64–67]. Such systems can increase the independence of the person with a disability while providing both reassurance and reduced responsibility for the caregiver. Usability for the caregiver is key, to ensure that the technology does not represent an additional burden instead of a benefit [65,67]. Cognitive disabilities often result in an inability of the brain to properly process and integrate sensory information. This can make it difficult to read printed text or to produce text with a pen or keyboard. By allowing information to be presented in different ways, computers can provide people with the flexibility to utilize their strengths and accommodate information processing deficits. Computers offer options such as changing text size and contrast or changing the color of the text and/or the background [68]. These interventions can make text easier to read for some people with learning disabilities. To further make information accessible, computers can provide speech synthesis as an alternative or adjunct to printed text [69]. Computers also offer alternatives for text production, such as speech recognition. One study that highlights the potential of this technology, demonstrated that students with learning disabilities achieved higher standardized scores on written essays when using speech recognition than when using standard text production (p < 0.05), and their essays were longer and had a higher proportion of words with seven or more letters [69]. Cognitive impairments can also lead to social and behavioral difficulties. Some individuals experience poor impulse control or a tendency toward compulsive behavior patterns. Examples can range from simple hand-wringing or rocking to self-injuring behaviors, and from poor recognition of others’ personal space to hostile behavior. Other people with cognitive disabilities are easily overwhelmed by environmental stimuli, and therefore may have difficulties with concentration and social engagement. Difficulties in processing visual information about faces or auditory information about a person’s tone of voice can also impair a person’s ability to recognize social cues. Other people simply have difficulty learning social norms for behavior. Some work has considered the application of assistive technology to neurobehavioral changes. For example, an automated cueing system (such as those used for task reminders) can also provide cues related to a behavioral modification program—reinforcing desirable behaviors and providing cues not to engage in undesirable behaviors. Findings suggest that a client’s behavior could be moderated using alternate cues that are based on the requirements of different social settings [70]. A more advanced behavioral prompting system might therefore use information about a person’s social or environmental context to provide more appropriate behavioral cues [70]. Other devices have attempted to use vibration or deep pressure to provide sensory stimulation, and thereby help people with autism or other impairments relax and concentrate [71]. Researchers are also investigating the use of virtual reality (VR) to help train people with cognitive impairments to perform real-life tasks, including
R.A. Cooper et al. / Medical Engineering & Physics 30 (2008) 1387–1398
social skills training [72,73]. Virtual environments allow for controllable input stimuli, gradual modifications to the environment to support generalization, and safety compared to real-world situations [72]. Some evidence has shown benefits of VR applications for people with autism and attention deficit disorder for skills related to environmental safety and social skills [72,73]. While there remain significant technological and social barriers to matching people with cognitive disabilities with appropriate technologies, there are also many promising trends. Computer technology provides increasing flexibility for people of all abilities, such as a large variety of input devices, customizable menus, or customized web pages, thus offering features with special importance for those with cognitive disabilities. The increased pace of modern society has placed and emphasis on developing memory and organization enhancing technologies, such as PDAs, web calendars, or GPS navigation devices, for all people, which will be of great benefit to people with cognitive disabilities. Meanwhile, more attention is being focused on the design of technologies specific to people with cognitive disabilities. These technologies vary widely in terms of the difficulties they address and the environments where they will be used. However, they share a need to involve people with cognitive disabilities and their caregivers in the design process, in order to better understand the abilities and needs of these populations.
5. Rehabilitation monitoring and coaching technologies Despite attempts at using simple reminders (e.g., timers, pressure monitors, PDA reminders/surveys), user’s guides (e.g., handouts, note cards), and consumer booklets developed to promote clinical practice guidelines, users do not seem to follow clinician instructions [74]. Novel approaches are emerging that use machine learning and artificial intelligence for real-time coaching of the person with disabilities for long-term monitoring of the person’s use of the equipment and to provide hard information for clinicians to use to augment their education of people with disabilities. This research is patient-centered and is likely to allow clinicians and people with disabilities to take more proactive roles in ways technology can enhance their lives; for example by avoiding detrimental health affects of prolonged seated postures. Important strides have been made in the development of a wheelchair activity monitor, which is a cornerstone of a ‘research module’ for the National Institute on Disability and Rehabilitation Research (NIDRR) Spinal Cord Injury Model Systems [75]. This device was used on both power and manual wheelchairs [76,77]. In a more recent study, improvements were made to a caster data logger (CDL) for electrical powered wheelchair (EPW) and were used to collect EPW usage data. Wearable sensor and notification platforms have been developed for context aware computing research [78,79]. One
1391
example of this technology, the eWatch, fits into a wristwatch making it instantly viewable and socially acceptable. The eWatch is capable of sensing temperature, light, two axes of acceleration and audio input at user controllable sampling rate up to 100 kHz. The user who wears the eWatch can be notified of an event or desired activity using a 128 × 64 pixel display, an LED, vibrating motor and tone generating buzzer [79]. The eWatch platform has been used to identify previously visited locations. Data were recorded from the light and audio sensors at several different locations including the user’s apartment, office, lab, street locations, interior of a bus, restaurant and supermarket. Using the Welch’s power spectrum, principle component analysis and it’s the nearest neighbor method with a fivefold cross-validation, the classification with the light sensor alone gave an overall result of 84.9% correctly classified samples, and the overall recognition accuracy with the audio sensor was 87.4%. Both sensors combined gave the best results of 91.4% [79]. Prior experiments also demonstrated the versatility of using eWatch located in various body positions for detecting a wide range of activities such as standing, sitting, walking, running, ascending stairs, and descending stairs [78]. Low fidelity sensors (such as accelerometers) and machine learning algorithms can be used to recognize propulsion patterns for manual wheelchair users. This provides insight into how the users should perform these actions, as well as guidance to perform them correctly, namely, it can aid in the training of manual wheelchair users to avoid the more strenuous propulsion techniques, which can lead to repetitive strain injuries. Four classic propulsion patterns have been identified by user studies, which are semicircular, single loop over, double loop over and arcing. Data collected using all four propulsion patterns on a variety of surface types were analyzed using two common machine learning algorithms: knearest neighbor (kNN) and support vector machines (SVM) with a radial basis function (RBF) kernel were compared. Accuracies of over 90% were achievable even with a simple classifier such as k-nearest neighbor (kNN). Results show that the classification accuracy for different propulsion patterns (semicircular, single loop over, double loop over and arcing) on different surfaces (asphalt, tile, medium pile and low pile carpet) using SVM is different.
6. Human device interfaces, assessment, and training Clinical problems drive focused research studies. For example, the majority of control interfaces used for such devices as power wheelchairs, alternative and augmentative communication devices, environmental control, adaptive automobile driving, and computer access are suboptimal for those with severe upper limb impairments or movement disorders [74]. Engineers are often faced with challenging cases in which customization and fitting of a control interface may still not meet all of the users’ needs. Recognizing the need for improved devices has driven researchers to develop
1392
R.A. Cooper et al. / Medical Engineering & Physics 30 (2008) 1387–1398
promising and sophisticated control interfaces such as isometric joysticks [74,75] that can be highly customized for individual user needs. We have shown how, through careful software algorithm selection, it is possible to achieve performance with isometric technology that is equivalent to that achieved with a traditional control interfaces for individuals with upper limb impairments [74]. And for those with movement disorders such as in cerebral palsy, isometric technology even outperforms traditional controls [76]. Virtual reality and simulated computer-based environments are being used to assess and train computer access tasks, ambulation with prosthetic limbs, and simulated power wheelchair driving using a variety of control interfaces. One of the main reasons electronic devices and mechatronic systems are inaccessible is that they treat all users the same way, and usually do not know much about each user. This is particularly unfortunate given the amount of accessibility software that comes pre-installed on most computers or is available online including pointer configuration utilities, screen magnifiers, and onscreen keyboards. Not only do most users not know what technology they need to make an electronic technology accessible, insurance limitations and high cost can prevent them from having an assistive technology clinician assess their ability. As a result, it is not uncommon for people to end up using either no accessibility tools, or tools that are not suited for their abilities. Machine learning techniques may be built based on learned statistical models of pointing (direction signaling) performance. Steps in automatically detecting characteristics of persons with a wide range of abilities based on observing user input events have been made. Learned statistical models of mouse use have been built to understand more about a user’s performance to distinguish between novice and expert menu use with 91% accuracy, without using a task model [77]. Pre-existing datasets from Keates, Koester and Trewin may be used to build learned statistical models on pointing data collected in a laboratory setting from individuals with varying ability to use computer pointing devices [78]. These datasets have been used to distinguish between pointing behaviors from individuals with pointing problems versus individuals without with 92.7% accuracy. Further studies have been able to distinguish between pointing data from young adults and adults versus older adults versus individuals with Parkinson’s disease with 91.6% accuracy. These results suggest that it may be feasible to use such models to automatically identify computer users who would benefit from accessibility tools [78]. People with severe motor disabilities, such as very highlevel spinal cord injury, amyotrophic lateral sclerosis, and brainstem stroke, have few means to communicate with the outside world. Brain computer interface (BCI) technology is intended to establish direct communication between the brain and external devices to provide effective communication and to eventually restore volitional movement [79,80]. The current approach is to extract control signals from recording
neural activity by inserting microelectrodes into the cortex. This approach is invasive, and there are barriers to the long-term stability of the recorded signals [81]. Electroencephalography (EEG) is non-invasive, but it has very limited information content due to the low-pass filter properties of the skull. Recently, electrocorticography (EcoG) has received attention as a potential modality for BCI. It has high spatial and temporal resolution, and several studies have shown that subjects can achieve effective control of 1D and 2D cursors in a very short period of time [82]. Human trials of BCI technology have primarily relied on microelectrodes that penetrate the cortex with limited control of computer cursors achieved [83,84]. Micro-ECoG is a minimally invasive approach, and may produce reliable highresolution recording of neural activity. A high-performance BCI system could bring substantial clinical benefit for patients with motor disabilities without significant additional risk.
7. Robotic assisted therapy One application of rehabilitation engineering that has recently received a lot of attention is the use of technology, specifically robotics, to augment traditional physical and occupational therapy. Two of the most well-known systems are the InMotion2 (formerly known as the MIT-MANUS) from Interactive Motion Technologies, Inc. and the Lokomat from Hocoma, but numerous other systems have been developed. As these systems have been reviewed elsewhere [80–83], we will not attempt to catalog them all here. Rather, we will rather briefly review the potential utility of therapeutic robots and the clinical results that have been obtained thus far. Work in this area has focused primarily on individuals with stroke, though similar techniques could be applied to other populations. We define a therapeutic robot as a system that senses the motions of the users, uses that information to make decisions, and provides visual and haptic feedback to the user. Robots have been developed to address both upper and lower extremity rehabilitation. Therapeutic robots for the upper extremity usually take the form of a robotic arm that is grasped by the user [84–86] or a glove or exoskeleton worn on the upper limb [87–90]. As an example, the Pneu-WREX is an exoskeleton worn on the arm; this device can allow the user to move independently while providing assistance when needed [88,91]. Therapeutic robots for the lower extremity are usually exoskeletons for gait training on a treadmill [92,93] or programmable footplates [94,95]. An example is the Haptic Walker, in which an individual with hemiparesis is supported while footplates assist him or her to complete a repetitive gait cycle with adjustable stride length and velocity [96]. Therapeutic robots have the potential to increase the amount of therapy received by an individual, the individual’s enjoyment of that therapy, and the quality of the assistance provided during therapy. Much recent research, particularly
R.A. Cooper et al. / Medical Engineering & Physics 30 (2008) 1387–1398
the results of constraint-induced movement therapy [96,97], indicate that intense, repetitive practice leads to greater functional gains. Yet the costs of therapist time limit the amount of therapy that is available to each individual. Robots can be used to increase the amount of therapy provided to each individual, thus improving the outcome of therapy. Robots can also increase the options available for telerehabilitation [98,99], which could increase the accessibility and convenience of increased amounts of therapy. The repetitive nature of therapy can decrease patient interest and motivation. The use of robots enables repetitive exercise to be transformed into an entertaining game. Many therapeutic robots provide game-like visual feedback in order to increase patient interest. For example, the GameCycle (Three Rivers holding Company LLC, Mesa, AZ) interacts with a Nintendo GameCube, making bilateral upper extremity exercise part of a racing game [100]. The Nitendo Wii has also been proposed for use in rehabilitation (USA today 7/24/2007 “Wii speeds up the rahab process”), but evidence for the Wii’s effectiveness as a rehabilitation tool not yet available. However, motor learning literature indicates that movements toward a functional object differ from those with no object as a target [101,102]. For this reason, some researchers have focused on robotic systems that can be used while interacting with real-world objects [102,103]; others have created virtual reality games based upon activities of daily living [104], since training in a virtual environment has been shown to transfer to a similar real-world task [105–107]. Optimal game interfaces for therapy robots will have to maximize both entertainment value and transfer to real-world tasks in order to obtain the best outcome of rehabilitation. In traditional physical and occupational therapy, a therapist often assists an individual with a movement that he or she cannot complete alone. Robots can precisely measure the characteristics of an individual’s movement, and exert force feedback via actuators (motors) in the system. Thus, they are ideally suited to provide individualized assistance or resistance to a user. Robotic assistance is particularly important for lower extremity rehabilitation since partial weight supported treadmill training often requires great effort from multiple therapists [103–105]. Much recent research has focused on what type of force feedback is most beneficial in a rehabilitation context. Stein et al. [108] found no difference between assistive and resistive training, raising the possibility that the effects of massed practice may be more important that the type of feedback provided by the robot. On the other hand, Takahashi et al. [87] present evidence that robot-assisted grasping led to greater rehabilitation gains than unassisted grasping. Assistance provided by the robot must be designed to prevent individuals from remaining passive while the robot performs the target movement, because active participation leads to greater motor improvement [109]. Emken et al. [110] has shown that individuals act to minimize a combination of error and effort. This means that individuals will tend to act so as to max-
1393
imize of the robot and minimize their own effort, and the robot controller must be adjusted to prevent this [91]. Additional work in this area is needed to determine the amount and type of robotic assistance that enables the best rehabilitation outcome. In addition, researchers are also investigating less conventional types of assistance during rehabilitation, such as the reverse force field training explored by Patton et al. [111]. Clinical trials have shown similar gains between robotassisted and conventional therapy. Extensive trials of the MIT-MANUS, a system for robot-assisted reaching in the horizontal plane, have shown that individuals that received robotic therapy in addition to conventional therapy showed improved motor function relative to individuals who received only conventional therapy [84,112–115]. Rehabilitation with the mirror image motion enabler (MIME), a system focusing on gross movement of the upper extremity, was compared to neurodevelopmental therapy. For individuals with subacute and chronic stroke, robotic therapy was associated with greater motor gains in the short term, but the two groups were equivalent at a 6-month follow-up [116,117]. Similarly, Takahashi et al. [87] showed greater improvements for robotassisted grasping relative to unassisted grasping, but these differences were not apparent at a 1-month follow-up. Kahn et al. [118] found that the results robot-assisted reaching were similar to those obtained with free, unassisted reaching for individuals with chronic stroke. In terms of the lower extremity, Werner et al. [95] showed that therapy with the Haptic Walker led to greater improvements in gait over treadmill training with partial body weight support; again, the differences disappeared by a 6-month follow-up. Mayr et al. [93] showed gait training with the Lokomat led to greater improvements in gait than conventional (Bobath) therapy, but follow-up data was not reported. In general, the long-term effects of robotic therapy appear to be similar to those of conventional therapy. This may be because the effects of similar amounts of any movement practice outweigh the differences between methods. While we cannot claim that robotic therapy is superior to conventional therapy, this is not important to the goals of therapeutic robotics. The goal of this technology is not to replace physical and occupational therapists. Rather, it is to augment the work of these therapists so that each individual can receive enough therapy to reach his or her optimal rehabilitation outcome. With additional research, robots can also assist therapists by providing the precise amount of assistance needed to allow an individual to complete a task while still encouraging the user to actively contribute as much as possible. Finally, the use of therapeutic robotic in combination with video games or virtual environments can also increase the entertainment value of rehabilitation, encouraging less motivated individuals to participate in the intensive practice required for a positive therapy outcome. By combining the expertise of therapists with the potential of robotic technology, each individual can be helped to maximize his or her recovery from disease or injury.
1394
R.A. Cooper et al. / Medical Engineering & Physics 30 (2008) 1387–1398
8. Personal mobility and manipulation technology For people who require assistance with both mobility and manipulation, technology provides few practical solutions. A survey of practicing clinicians reported that between 10% and 40% of their clients who desired powered mobility (power wheelchairs or scooters) could not be fitted with them because sensory impairments, poor motor function or cognitive deficits made it impossible for the clients to safely drive using existing controls [119–121] This lack of technological solutions leave many people with disabilities dependent on human caregivers to perform daily tasks. The need for human assistance is one of the top items in all barriers surveys of people with severe disabilities, older adults with physical frailty and their caregivers [122]. An ideal technology would effectively combine manipulation and mobility assistance with perception and decision-making wherever a person goes. In order to make interaction easier, a wide range of natural and intuitive interfaces that reduce the time to complete tasks and produce fluid human-like motions are needed [123]. Several user needs that have been identified include: (1) providing ubiquitous manipulation assistance for those who need it; (2) making mobility and manipulation assistance easier to use; (3) extending the range of environments of mobility devices, and particularly handling small barriers such as a set of steps to enter a friend’s house; and (4) providing increased independence by reducing the reliance on constant attendant care. Simple manipulators have been appended to wheelchairs, but their interfaces – typically on/off or keyboard control of single joints – are rudimentary by contemporary robotics standards. Even with coordinated joint control, teleoperation of a manipulator is hard for almost anyone; teleoperating a manipulator while simultaneously controlling a wheelchair is daunting. This makes many everyday tasks, such as opening a door, very difficult to accomplish using an assistive device for manipulation. Our approach is to integrate the base and two robotic arms into a unified and coordinated robotic system that is controlled by a computer network to give unprecedented capabilities. The urrent Personal Mobility and Manipulation Appliance (PerMMA) prototype combines two MANUS robotic arms [124] modified to include additional degrees of freedom through a custom powered track system with a Permobil C500 that has been converted to a mobile robot [125] (Fig. 1). The powered track allows storage of the robot arms behind the wheelchair for a narrower profile, and also side-by-side operation for bi-manual tasks. The C500 has had its controller and algorithms replaced by a custom-built system that uses a network of computers for advanced sensing, processing, coordination, and control. Currently the network includes computers for machine vision, path planning, safety monitoring, and advanced control algorithms. In order to increase to ensure that the system meets user’s needs, we are employing an approach where users and clinicians are involved throughout the entire design and
Fig. 1. PerMMA controller include single board computer, interface with other sensors, encoders and computers.
development process. Neither the experts (clinicians or engineers) nor the users possess the complete or most critical knowledge. A systematic mapping is required of the opinions and viewpoints of the different interested stakeholders, their experiences and requirements that can contribute to the developing of a successful system. We are applying a systematic approach to user involvement characterized by: (1) incorporation of universal design principals; (2) active involvement of all relevant stakeholders and a clear understanding of the circumstances of the target user and task requirements; (3) multi-disciplinary research, design, and development processes involving users, engineers, designers, marketing, clinicians, service delivery, social and health professionals; and (4) an interactive development process.
9. Summary As the number of people with reduced functional capabilities due to aging and disabilities increases, advances in assistive technologies are needed. Traditional robotics and intelligent systems research (e.g., factory automation and entertainment) has not fulfilled the increasing and multiple requirements of the people with disabilities’ needs. Researchers need to leverage recent advancements of technologies to bring us closer to futuristic visions of compassionate intelligent devices and technology-embedded environments. This manuscript covers a few of the emerging lines of research in of engineering in medical rehabilitation and offers a perspective into the characteristics and requirements of future assistive technology.
Conflicts of interest None.
R.A. Cooper et al. / Medical Engineering & Physics 30 (2008) 1387–1398
Acknowledgements This material is based upon work supported by the National Science Foundation under Cooperative Agreement EEC-0540865. This material is also based upon work supported by the Office of Research and Development, Rehabilitation Research & Development Service, Department of Veterans Affairs, Grant# B3142C.
References [1] Center for Health Workforce Studies. The Impact of the aging population on the health workforce in the United States. School of Public Health, University at Albany; 2006. [2] Cooper RA. Quality-of-life technology [a human-centered and holistic design]. Engineering in Medicine and Biology Magazine, IEEE 2008;27(2):10–1. [3] Diagram MF. White paper of the elderly. Tokyo: Japan National Council of Social Welfare; 2001. [4] Eurostat. Population projections 2008–2060. News release 119. 26 August 2008. [5] Brandt EN, Pope AMP. Enabling America: assessing the role of rehabilitation science and engineering. National Academy Press; 1997. [6] Cooper RA, Ohnabe H, Hobson DA, editors. An introduction to rehabilitation engineering. New York: CRC Press; 2006. p. 19–45. [7] Dicianno BE, Tovey E. Power mobility device provision: understanding medicare guidelines and advocating for clients. Archives of Physical Medicine and Rehabilitation 2007;88(6):807–16. [8] Aissaoui R, Lacoste M, Dansereau J. Analysis of sliding and pressure distribution during a repositioning of persons in a simulator chair. IEEE Transaction on Neural System Rehabilitation Engineer 2001;9(2):215–24. [9] Pellow TR. A comparison of interface pressure readings to wheelchair cushions and positioning: a pilot study. Cananda J Occup Ther 1999;66(3):140–9. [10] Stinson MD, Porter-Armstrong A, Eakin P. Seat-interface pressure: a pilot study of the relationship to gender, body mass index, and seating position. Archives of Physical Medicine and Rehabilitation 2003;84(3):405–9. [11] Vaisbuch N, Meyer S, Weiss PL. Effect of seated posture on interface pressure in children who are able-bodied and who have myelomeningocele. Disability and Rehabilitation 2000;22(17): 749–55. [12] Dicianno BE, Margaria E, Arva J, Lieberman JM, Schmeler MR, Souza A, et al. RESNA position on the application of tilt, recline, and elevating leg rests for wheelchairs. Approved by RESNA Board of Directors; 2008. [13] Boninger M, Stripling T. Preserving upper-limb function in spinal cord injury. Archives of Physical Medicine and Rehabilitation 2007;88(6):817–1817. [14] Moseley AM, Lanzarone S, Bosman JM, van Loo MA, deBie RA, Hassett L. Ecological validity of walking speed assessment after traumatic brain injury: a pilot study. Journal of Head Trauma Rehabilitation 2004;19(4):341–8. [15] Raskovic D, Martin T, Jovanov E. Medical monitoring applications for wearable computing. The Computer Journal 2004;47(4): 495–504. [16] Korhonen I, Bardram JE. Guest editorial: introduction to the special section on pervasive healthcare. IEEE Transactions on Information Technology in Biomedicine 2004;8(3 (September)):229–34. [17] Dishman E. Inventing wellness systems for aging in place. IEEE Computer Magazine 2004;May:34–41.
1395
[18] Lymberis A, Olsson S. Intelligent biomedical clothing for personal health and disease management: state of the art and future vision. Telemedicine Journal and e-Health 2003;9(4):379–86. [19] Rialle V, Duchene F, Noury N, Bajolle L, Demongeot J. Healthsmart” home: information technology for patients at home. Telemedicine Journal and e-Health 2002;8(4):395–409. [20] Cuddihy P. Home assurance system. Pervasive Computing 2004; April–June:48. [21] Culler D, Estrin D, Srivastava M. Guest editors’ introduction: overview of sensor networks. Computer 2004;37(8):41–9. [22] Suzuki R, Ogawa M, Otake S, Izutsu T, Tobimatsu Y, Izumi S-I, et al. Analysis of activities of daily living in elderly people living alone: single-subject feasibility study. Telemedicine Journal and e-Health 2004;10(2):260–76. [23] Woolliscroft J, Koelling TM. Guest editorial: redesigning the management of chronic illness. Telemedicine Journal and e-Health 2004;10(2):118–21. [24] Finkelstein SM, Speedie SM, Demiris G, Veen M, Lundgren JM, Potthoff S. Telehomecare: quality, perception, satisfaction. Telemedicine Journal and e-Health 2004;10(2):122–8. [25] Lankford DA. Wireless CPAP patient monitoring: accuracy study. Telemedicine Journal and e-Health 2004;10(2):2004. [26] Kobb R, Hoffman N, Lodge R, Kline S. Enhancing elder chronic care through technology and care coordination: report from a pilot. Telemedicine Journal and e-Health 2003;9(2):189–95. [27] Suzuki R, Ogawa M, Tobimatsu Y, Iwaya T. Time-course action analysis of daily life investigations in the welfare techno house in Mizusawa. Telemedicine Journal and e-Health 2001;7(3): 249–59. [28] Chumbler NR, Mann WC, Wu S, Schmid A, Kobb R. The association of home-telehealth use and care coordination with improvement of functional and cognitive functioning in frail elderly men. Telemedicine Journal and e-Health 2004;10(2):129–37. [29] Sato T, Harada T, Mori T. Environment-type robot system “Robotic Room” featured by behavior media, behavior contents, and behavior adaptation. IEEE/ASME Transactions on Mechatronics 2004;9(3 (September)):529–34. [30] Hayes TL, Pavel M, Schallau PK, Adami AM, editors. Unobtrusive monitoring of health status in an aging population. UbiHealth 2003: the 2nd international workshop on ubiquitous computing for pervasive healthcare applications. 2003. [31] Allum JH, Carpenter MG. A speedy solution for balance and gait analysis: angular velocity measured at the centre of body mass. Current Opinions in Neurology 2005;18(1):15–21. [32] Nelson A, Powell-Cope G, Gavin-Dreschnack D, Quigley P, Bulat T, Baptiste AS, et al. Technology to promote safe mobility in the elderly. Nursing Clinics of North America 2004;39(3):649–71. [33] Wu G, Xue S. Portable preimpact fall detector with inertial sensors. IEEE Transactions on Neural Systems and Rehabilitation Engineering 2008;16(2):178–83. [34] Sixsmith A, Johnson N. A smart sensor to detect falls in the elderly. IEEE Pervasive Computing 2004;3(2):42–7. [35] Kiff LM, Haigh KZ, Sun X. Mobility monitoring with the Independent LifeStyle Assistant (I.L.S.A.). In: International conference on aging, disability and independence (ICADI). 2003. [36] Guralnik V, Haigh KZ. Learning models of human behavior with sequential patterns. In: AAAI workshop on automation as caregiver. AAAI Press; 2002. [37] Karen H, Kiff LM, Ho G. The independent LifeStyle Assistant: lessons learned. Assistive Technology 2006;18(1):87–106. [38] Chan M, Hariton C, Ringeard P, Campo E, editors. Smart house automation system for the elderly and the disabled. In: IEEE international conference on systems, man and cybernetics. Vancouver: BC Canada IEEE Press; 1995. [39] Ciampolini A, editor. Distributed medical diagnosis with abductive logic agents. In: Workshop on agents applied in health care, at the European conference on artificial intelligence (ECAI). 2002.
1396
R.A. Cooper et al. / Medical Engineering & Physics 30 (2008) 1387–1398
[40] Larsson JE, Hayes-Roth B. Guardian: an intelligent autonomous agent for medical monitoring and diagnosis. IEEE Intelligent Systems and their Applications 1998;13(1):58–64. [41] Factor M. A parallel software architecture for building intelligent medical monitors. International Journal of Clinical Monitoring and Computing 1990. [42] Mabry SL, Etters T, Schneringer T, Edwards N, Hug C, editors. Addressing the complexity of patient monitoring with multiagents and modular logic. In: International conference on autonomous agents. Melbourne, Australia: ACM Press; 2003. [43] Mabry SL, Schneringer T, Etters T, Edwards N, editors. Intelligent agents for patient monitoring and diagnostics. In: ACM symposium on applied computing. Melbourne, Florida: ACM Press; 2003. [44] Winters J, Herman W. Report of the workshop on home care technologies for the 21st century. Washington, DC: Catholic University of America; 1999. [45] Braddock D, Rizzolo MC, Thompson M, Bell R. Emerging technologies and cognitive disability. Journal of Special Education Technology 2004;19(4):49–56. [46] Braddock D. Emerging research frontiers and partnerships in developmental disabiliites. In: Proceedings of the Association of University Centers on disabilities. Bethesda, MD; 2001. [47] Braddock D. Emerging research frontiers and partnerships in developmental disabilities. In: Association of University Centers on disabilities. Bethesda, MD; 2001. [48] American Psychiatric Association Task Force on DSM-IV, Diagnostic and statistical manual of mental disorders: DSM-IV-TR. 4th text revision, xxxvii. Washington, DC: American Psychiatric Association; 2000. p. 943. [49] Kime S, Lamb D, Wilson B. Use of a comprehensive program of external cueing to enhance procedural memory in a patient with dense amnesia. Brain Injury 1995;10:17–25. [50] Kirsch NL, Levine SP, Fallon-Kreuger M, Jaros L. The microcomputer as an ‘orthotic’ device for patients with cognitive deficits. Journal of Head Trauma Rehabilitation 1987;2(4):77–86. [51] Cole E. Cognitive prosthetics: an overview to a method of treatment. NeuroRehabilitation 1999;12:39–51. [52] LoPresti EF, Mihailidis A, Kirsch N. Technology for cognitive rehabilitation and compensation: state of the art. Neuropsychological Rehabilitation 2004;14(1–2):5–39. [53] Friedman, M. A wearable computer that gives context-sensitive verbal guidance to people with memory or attention impairments. In: Proceedings of the Rehabilitation Society of North America (RESNA). 1993. p. 199–201. [54] Mihailidis A, Fernie G, Barbenel JC. The use of artificial intelligence in the design of an intelligent cognitive orthotic for people with dementia. Assistive Technology 2001;13:23–39. [55] Pollack ME, Brown L, Colbry D, McCarthy CE, Orosz C, Peintner B, et al. Autominder: an intelligent cognitive orthotic system for people with memory impairment. Robotics and Autonomous Systems 2003;44:273–82. [56] Ben-Arie J, Wang Z, Pandit P, Rajaram S. Human activity recognition using multidimensional indexing. IEEE Transaction on Pattern Anala Mach Intell 2002;24(8):1091–104. [57] Bao L, Intille SS. Activity recognition from user-annotated acceleration data. Pervasive 2004:1–17. [58] Phillipose M, Fishkin KP, Perkowitz M, Patterson DJ, Fox D, Kautz H, et al. Inferring activities from interactions with objects. IEEE Pervasive Computing 2004;3(4):50–7. [59] A. Kutiyanawala, V. Kulyukin, E. LoPresti. A rollatormounted way finding system for the elderly: a smart world perspective. In: Proceedings of ACM SIGACCESS; 2006. p. 245–6. [60] Lee ML, Dey AK. Capture & access lifelong assistive technology for people with episodic memory impairment. Working notes. In: University of Rochester assisted cognition workshop. 2007. p. 49–57.
[61] Alm N, Astell A, Ellis M, Dye R, Gowans G, Campbell J. A cognitive prosthesis and communication support for people with dementia. Neuropsychological Rehabilitation 2004;14(1/2):117–34. [62] Stock SE, Davies DK, Secor RR, Wehmeyer ML. Self-directed career preference selection for individuals with intellectual disabilities: using computer technology to enhance self-determination. Journal of Vocational Rehabilitation 2003;19:95–103. [63] Carmien S, Dawe M, Fischer G, Gorman A, Kintsch A, Sullivan JF. Socio-technical environments supporting people with cognitive disabilities using public transportation. ACM Transactions on Computer-Human Interaction 2005;12(2):233–62. [64] Jonsson B, Svensk A. Isaac a personal digital assistant for the differently abled. In: Proceedings of the 2nd TIDE congress. 1995. p. 356–61. [65] Cole E, Ziegmann M, Wu Y, Yonker V, Gustafson C, Cirwithen S. Use of “therapist-friendly” tools in cognitive assistive technology and telerehabilitation. In: Proceedings of the RESNA 2000. p. 31–3. [66] Jinks A, Kellinger F, Garber HJ. Handheld PC/web-based task reminder for brain-injured adults. In: Proceedings of the RESNA 2004. [67] LoPresti EF, Simpson RC, Kirsch N, Schreckenghost D, Hayashi S. Distributed cognitive aid with scheduling and interactive task guidance. Journal of Rehabilitation Research and Development 2008;45(4):505–22. [68] Gregor P, Newell AF. An empirical investigation of ways in which some of the problems encountered by some dyslexics may be alleviated using computer techniques. In: Proceedings of the assets. 2000. p. 85–91. [69] Raskind MH, Higgins EL. Effects of speech synthesis on the proofreading efficiency of postsecondary students with learning disabilities. Learning Disability Quarterly 1995;18(Spring):141–58. [70] Kirsch NL, Shenton M, Spirl E, Simpson R, LoPresti E, Schreckenghost D. An assistive-technology intervention for verbose speech after traumatic brain injury: a single case study. Journal of Head Trauma Rehabilitation 2004;19(5):366–77. [71] Grandin TT. Calming effects of deep touch pressure in patients with autistic disorder, college students and animals. Journal of Child and Adolescent Psychopharmacology 1992;2(1):63–70. [72] Strickland D. Virtual reality for the treatment of autism. In: Riva G, editor. Virtual reality in neuro-psych-physiology. IOS Press; 1997. p. 81–94. [73] Rizzo AA, Buckwalter JG, Bowerly T, Humphrey LA, Neumann U, Van Rooyen A, et al. The virtual classroom: a virtual reality environment for the assessment and rehabilitation of attention deficits. Revista Espanola de Neuropsicologia 2001;3(3):11–37. [74] Brian Haynes R, McDonald HP, Garg AX. Helping patients follow prescribed treatment: clinical applications. The Journal of American Medical Association 2002;288(22):2880–3. [75] Spaeth DM, Arva J, Cooper RA. Application of a commercial datalogger for rehabilitation research. In: Proceedings of the 23rd annual rehabilitation engineering & assistive technology conference. 2000. p. 381–3. [76] Cooper RA, Thorman T. Driving characteristics of EPW users: how far, fast, and often do people drive? Archives of Physical Medicine and Rehabilitation 2002;83:250–5. [77] Tolerico ML, Ding D, Cooper RA, Spaeth DM, Fitzgerald SG, Cooper R, et al. Assessing the mobility characteristics and activity levels of manual wheelchair users. Journal of Rehabilitation Research and Development 2007;44(4): 561–72. [78] Maurer U, Smailagic A, Siewiorek D, Deisher M. Activity recognition and monitoring using multiple sensors on different body positions. In: IEEE international workshop on wearable and implantable body sensor networks. 2006. p. 113–6. [79] Maurer U, Rowe A, Smailagic A, Siewiorek D. eWatch: a wearable sensor and notification platform. IEEE International Workshop on Wearable and Implantable Body Sensor Networks. 2006.
R.A. Cooper et al. / Medical Engineering & Physics 30 (2008) 1387–1398 [80] Brewer BR, McDowell SK, Worthen-Chaudhari LC. Poststroke upper extremity rehabilitation: a review of robotic systems and clinical results. Topics in Stroke Rehabilitation 2007;14(November–December):22–44. [81] Hesse S, Schmidt H, Werner C, Bardeleben A. Upper and lower extremity robotic devices for rehabilitation and for studying motor control. Current Opinion in Neurology 2003;16(December):705–10. [82] Reinkensmeyer DJ, Emken JL, Cramer SC. Robotics, motor learning, and neurologic recovery. Annual Review of Biomedical Engineering 2004;6:497–525. [83] Fasoli SE, Krebs HI, Hogan N. Robotic technology and stroke rehabilitation: translating research into practice. Topics in Stroke Rehabilitation 2004;11(Fall):11–9. [84] Krebs HI, Hogan N, Aisen ML, Volpe BT. Robot-aided neurorehabilitation. IEEE Transactions on Rehabilitation Engineering 1998;6(March):75–87 (a publication of the IEEE Engineering in Medicine and Biology Society). [85] Burgar CG, Lum PS, Shor PC, Machiel Van der Loos HF. Development of robots for rehabilitation therapy: the Palo Alto VA/Stanford experience. Journal of Rehabilitation Research and Development 2000;37(November–December):663–73. [86] Amirabdollahian F, Loureiro R, Gradwell E, Collin C, Harwin W, Johnson G. Multivariate analysis of the Fugl-Meyer outcome measures assessing the effectiveness of GENTLE/S robot-mediated stroke therapy. Journal of Neuroengineering and Rehabilitation 2007;4:4. [87] Takahashi CD, Der-Yeghiaian L, Le V, Motiwala RR, Cramer SC. Robot-based hand motor therapy after stroke. Brain 2008;131(February):425–37. [88] Wolbrecht E, Leavitt J, Reinkensmeyer DJ, Bobrow J. Control of a pneumatic orthosis for upper extremity stroke rehabilitation. In: IEEE engineering in medicine and biology conference. 2006. p. 2687–93. [89] Nef T, Riener R. ARMin: design of a novel arm rehabilitation robot. In: International conference on rehabilitation robotics. 2005. [90] Stein J, Narendran K, McBean J, Krebs K, Hughes R. Electromyography-controlled exoskeletal upper-limb-powered orthosis for exercise training after stroke. American Journal of Physical Medicine & Rehabilitation/Association of Academic Physiatrists 2007;86(April):255–61. [91] Reinkensmeyer DJ, Wolbrecht E, Bobrow J. A computational model of human–robot load sharing during robot-assisted arm movement training after stroke. In: Proceedings of the annual international conference of the IEEE Engineering in Medicine and Biology Society, vol. 1. IEEE Engineering in Medicine and Biology Society. 2007. p. 4019–23. [92] Hocoma. Available at: http://www.hocoma.ch/web/en/header/ index.html. [93] Mayr A, Kofler M, Quirbach E, Matzak H, Frohlich K, Saltuari L. Prospective, blinded, randomized crossover study of gait rehabilitation in stroke patients using the Lokomat gait orthosis. Neurorehabilitation and Neural Repair 2007 Jul-Aug;21:307–14. [94] Hesse S, Schmidt H, Werner C. Machines to support motor rehabilitation after stroke: 10 years of experience in Berlin. Journal of Rehabilitation Research and Development 2006 Sep-Oct;43:671–8. [95] Werner C, Von Frankenberg S, Treig T, Konrad M, Hesse S. Treadmill training with partial body weight support and an electromechanical gait trainer for restoration of gait in subacute stroke patients: a randomized crossover study. Stroke 2002 Dec;33:2895–901. [96] Taub E, Uswatte G, Pidikiti R. Constraint-induced movement therapy: a new family of techniques with broad application to physical rehabilitation—a clinical review. Journal of Rehabilitation Research and Development 1999 Jul;36:237–51. [97] Wolf SL, Winstein CJ, Miller JP, Thompson PA, Taub E, Uswatte G, et al. Retention of upper limb function in stroke survivors who have received constraint-induced movement therapy: the EXCITE randomised trial. Lancet Neurology 2008;7(January):33–40. [98] Reinkensmeyer DJ, Pang CT, Nessler JA, Painter CC. Webbased telerehabilitation for the upper extremity after stroke. IEEE
[99]
[100]
[101]
[102]
[103]
[104]
[105]
[106]
[107]
[108]
[109] [110]
[111]
[112]
[113]
[114]
[115]
[116]
1397
Transactions on Neural Systems and Rehabilitation Engineering 2002;10(January):102–8 (a publication of the IEEE Engineering in Medicine and Biology Society). Carignan CR, Krebs HI. Telerehabilitation robotics: bright lights, big future? Journal of Rehabilitation Research and Development 2006;43(September–October):695–710. Widman LM, McDonald CM, Abresch RT. Effectiveness of an upper extremity exercise device integrated with computer gaming for aerobic training in adolescents with spinal cord dysfunction. The Journal of Spinal Cord Medicine 2006;29:363–70. Trombly CA, Wu C. Effect of rehabilitation tasks on organization of movement after stroke. American Journal of Occupational Therapy 1999;53:333–44. Wisneski KJ, Johnson MJ. Quantifying kinematics of purposeful movements to real, imagined, or absent functional objects: implications for modelling trajectories for robot-assisted ADL tasks. Journal of Neuroengineering and Rehabilitation 2007;4:7. Takahashi CD, Der-Yeghiaian L, Le VH, Cramer SC. A robotic device for hand motor therapy after stroke. In: International conference on rehabilitation robotics. 2005. Sanchez RJ, Liu J, Rao S, Shah P, Smith R, Rahman T, et al. Automating arm movement training following severe stroke: functional exercises with quantitative feedback in a gravity-reduced environment. IEEE Transactions on Neural Systems and Rehabilitation Engineering 2006;14 (september):378–89 (a publication of the IEEE Engineering in Medicine and Biology Society). Rose FD, Attree EA, Brooks BM, Parslow DM, Penn PR, Ambihaipahan. Training in virtual environments: transfer to real world tasks and equivalence to real task training. Ergonomics 2000;43: 494–511. Holden M, Todorov E, Callahan J, Bizzi E. Virtual environment training improves motor performance in two patients with stroke: case report. Neurology Report 1999;23:57–67. Todorov E, Shadmehr R, Bizzi E. Augmented feedback presented in a virtual environment accelerates learning of a difficult motor task. Journal of Motor Behavior 1997;29(June):147–58. Stein J, Krebs HI, Frontera WR, Fasoli SE, Hughes R, Hogan N. Comparison of two techniques of robot-aided upper limb exercise training after stroke. American Journal of Physical Medicine & Rehabilitation/Association of Academic Physiatrists 2004;83(September):720–8. Lotze M, Braun C, Birbaumer N, Anders S, Cohen LG. Motor learning elicited by voluntary drive. Brain 2003;126(April):866–72. Emken JL, Benitez R, Sideris A, Bobrow JE, Reinkensmeyer DJ. Motor adaptation as a greedy optimization of error and effort. Journal of Neurophysiology 2007;97(June):3997–4006. Patton JL, Kovic M, Mussa-Ivaldi FA. Custom-designed haptic training for restoring reaching ability to individuals with poststroke hemiparesis. Journal of Rehabilitation Research and Development 2006;43(September– October):643–56. Krebs HI, Volpe BT, Aisen ML, Hogan N. Increasing productivity and quality of care: robot-aided neuro-rehabilitation. Journal of Rehabilitation Research and Development 2000;37(November–December):639–52. Volpe BT, Krebs HI, Hogan N. Is robot-aided sensorimotor training in stroke rehabilitation a realistic option? Current Opinion in Neurology 2001;14(December):745–52. Volpe BT, Krebs HI, Hogan N, Edelstein OL, Diels C, Aisen M. A novel approach to stroke rehabilitation: robot-aided sensorimotor stimulation. Neurology 2000;54(May):1938–44. Volpe BT, Krebs HI, Hogan N, Edelsteinn L, Diels CM, Aisen ML. Robot training enhanced motor outcome in patients with stroke maintained over 3 years. Neurology 1999;53(November): 1874–6. Lum PS, Burgar CG, Shor PC, Majmundar M, Van der Loos M. Robot-assisted movement training compared with conventional therapy techniques for the rehabilitation of upper-limb motor func-
1398
[117]
[118]
[119] [120]
R.A. Cooper et al. / Medical Engineering & Physics 30 (2008) 1387–1398 tion after stroke. Archives of Physical Medicine and Rehabilitation 2002;83(July):952–9. Lum PS, Burgar CG, Van der Loos M, Shor PC, Majmundar M, Yap R. MIME robotic device for upper-limb neurorehabilitation in subacute stroke subjects: a follow-up study. Journal of Rehabilitation Research and Development 2006;43(September–October):631–42. Kahn LE, Lum PS, Rymer WZ, Reinkensmeyer DJ. Robot-assisted movement training for the stroke-impaired arm: does it matter what the robot does? Journal of Rehabilitation Research and Development 2006;43(September–October):619–30. Simpson RC. Smart wheelchairs: a literature review. Journal of Rehabilitation Research and Development 2005;42(4):423–36. Simpson RC, LoPresti E, Cooper RA. How many people would benefit from a smart wheelchair? Journal of Rehabilitation Research and Development 2008;45(1):53–72.
[121] Fehr L, Langbein WE, Skaar SB. Adequacy of power wheelchair control interfaces for persons with severe disabilities: a clinical survey. Journal of Rehabilitation Research and Development 2000;37(3):353–60. [122] Cooper RA, Dan Ding, Grindle GG, Hongwu Wang. Personal mobility and manipulation using robotics, artificial intelligence and advanced control. In: 29th annual international conference of the IEEE. Engineering in Medicine and Biology Society; 2007. p. 4368–71. [123] Goodrich MA, Schultz AC. Human-robot interaction: a survey. Foundations and Trends in Human-Computer Interaction 2007;1(3 (January)):203–75. [124] MANUS robotic arms. Available at: http://www.exactdynamics.nl/. [125] Permobil USA. Available at: http://www.permobilusa.com/browse. php?id=37.
