Technology-supported training of arm-hand skills in stroke

Impaired arm-hand performance is a serious consequence of stroke that is associated with reduced self-efficacy and poor quality of life. Task-oriented arm training is a therapy approach that is known to improve skilled arm-hand performance, even in chronic stages after stroke. At the start of this project, little knowledge had been consolidated regarding taskoriented arm training characteristics, especially in the field of technology-supported rehabilitation. The feasibility and effects of technology-supported client-centred task-oriented training on skilled arm-hand performance had not been investigated but to a very limited degree. Reviewing literature on rehabilitation and motor learning in stroke led to the identification of therapy oriented criteria for rehabilitation technology aiming to influence skilled arm-hand performance (chapter 2). Most rehabilitation systems reported in literature to date are robotic systems that are aimed at providing an engaging exercise environment and feedback on motor performance. Both, feedback and engaging exercises are important for motivating patients to perform a high number of exercise repetitions and prolonged training, which are important factors for motor learning. The review also found that current rehabilitation technology is focussed mainly on providing treatment at a function level, thereby improving joint range of motion, muscle strength and parameters such as movement speed and smoothness of movement during analytical movements. However, related research has found no effects of robot-supported training at the activity level. The review concluded that a challenge exists for upper extremity rehabilitation technology in stroke patients to also provide more patienttailored task-oriented arm-hand training in natural environments to support the learning of skilled arm-hand performance. Besides mapping the strengths of different technological solutions, the use of outcome measures and training protocols needs to become more standardized across similar interventions, in order to help determine which training solutions are most suitable for specific patient categories. Chapter 4 contributes towards such a standardization of outcome measurement. A concept is introduced which may guide the clinician/researcher to choose outcome measures for evaluating specific and generalized training effects. As an initial operationalization of this concept, 28 test batteries that have been used in 16 task-oriented training interventions were rated as to whether measurement components were measured by the test. Future research is suggested that elaborates the concept with information on the relative weighing of components in each test, with more test batteries (which may lead to additional components) and by adding more test properties into the concept (e.g. psychometric properties of the tests, possible floor- or ceiling effects). Task-oriented training is one of the training approaches that has been shown to be beneficial for skilled arm-hand performance after stroke. Important mechanisms for motor learning that are identified are patient motivation for such training, and the learning of efficient goaloriented movement strategies and task-specific problem solving. In this thesis we operationalize task-oriented training in terms of 15 components (chapter 3). A systematic review that included 16 randomized controlled trials using task-oriented training in stroke patients, evaluated the effects of these training components on skilled arm-hand performance. The number of training components used in an intervention aimed at improving arm-hand performance after stroke was not associated with the post-treatment effect size. Distributed practice and feedback were associated with the largest post-intervention effect sizes. Random practice and use of clear functional training goals were associated with the largest follow-up effect sizes. It may be that training components that optimize the storage of learned motor performance in the long-term memory are associated with larger treatment effects. Unfortunately, feedback, random practice and distributed practice were reported in very few of the included randomized controlled trials (in only 6,3 and 1 out of the 17 studies respectively). Client-centred training, i.e. training on exercises that support goals that are selected by the patients themselves, improves patient motivation for training. Motivation in turn has proven to positively influence motor learning in stroke patients, as attention during training is heightened and storage of information in the long-term memory improves. Chapter 5 reports on an interview of 40 stroke patients, investigating into training preferences. A list of 46 skills, ranked according to descending training preference scores, was provided that can be used for implementation of exercises in rehabilitation technology, in order for technologysupported training to be client-centred. Chapter 6 introduces T-TOAT, a technology supported task-oriented arm training method that was developed together with colleagues at Adelante (Hoensbroek, NL). T-TOAT enables the implementation of exercises that support task-oriented training in rehabilitation technology. The training method is applicable for different technological systems, e.g. robot and sensor systems, or in combination with functional electrical stimulation, etc. To enable the use of TTOAT for training with the Haptic Master Robot (MOOG-FCS, NL), special software named Haptic TOAT was developed in Adelante together with colleagues at the Centre of Technology in Care of Zuyd University (chapter 6). The software enables the recording of the patient’s movement trajectories, given task constraints and patient possibilities, using the Haptic Master as a recording device. A purpose-made gimbal was attached to the endeffector, leaving the hand free for the use and manipulating objects. The recorded movement can be replayed in a passive mode or in an active mode (active, active-assisted or activeresisted). Haptic feedback is provided when the patient deviates from the recorded movement trajectory, as the patient receives the sensation of bouncing into a wall, as well as feeling a spring that pulls him/her back to the recorded path. The diameter of the tunnel around the recorded trajectory (distance to the wall), and the spring force can be adjusted for each patient. An ongoing clinical trial in which chronic stroke patients train with Haptic-TOAT examines whether Haptic Master provides additional value compared to supporting the same exercises by video-instruction only. Together with Philips Research Europe (Eindhoven,Aachen), the T-TOAT method has been implemented in a sensor based prototype, called Philips Stroke Rehabilitation Exerciser. This system included movement tracking sensors and an exercise board interacting with real life objects. A very strong feature of the system is that feedback is provided to patients (real-time and after exercise performance), based on a comparison of the patient’s exercise performance to individual targets set by the therapist. Chapter 7 reports on a clinical trial investigating arm-hand treatment outcome and patient motivation for technology-supported task-oriented training in chronic stroke patients. It was found that 8 weeks of T-TOAT training improved arm-hand performance in chronic stroke patients significantly on Fugl-Meyer, Action Research Arm Test, and Motor Activity Log. An improvement was found in health-related quality of life. Training effects lasted at least 6 months post-training. Participants reported feeling intrinsically motivated and competent to use the system. The results of this study showed that T-TOAT is feasible. Despite the small number of stroke patients tested (n=9), significant and clinically relevant improvements in skilled arm-hand performance were found. In conclusion, this thesis has made several contributions. It motivated the need for clientcentred task-oriented training, which it has operationalized in terms of 15 components. Four of these 15 components were identified as most beneficial for the patient. A prioritized inventory of arm-hand training preferences of stroke patients was compiled by means of an interview study of 40 subacute and chronic stroke patients. T-TOAT, a method for technology-supported, client-centred, task-oriented training, was conceived and implemented in two target technologies (Haptic Master and Philips Stroke Rehabilitation Exerciser). Its feasibility was demonstrated in a clinical trial showing substantial and durable benefits for the stroke patients. Finally, the thesis contributes towards the standardization of outcome measures which is necessary for charting progress and guiding future developments of technology-supported stroke rehabilitation. Methodological considerations were discussed and several suggestions for future research were presented. The variety of treatment approaches and the various ways of support and challenge that are offered by existing rehabilitation technologies hold a large potential for offering a variety of extra training opportunities to stroke patients that may improve their arm-hand performance. Such solutions will be of increasing importance, to alleviate therapists and reduce economic pressure on the health care system, as the stroke incidence is increasing rapidly over the coming decades