Executive Summary
In the modern world, stroke is considered as one of the most common causes of disability in adults, affecting over twenty million people every year. Being caused by abnormalities in blood flow, the condition has a variety of complications that are detrimental to the quality of life. Hemiparesis, one of its most frequent consequences that affect up to eighty percent of stroke survivors, presents an incomplete paralysis of only one side of the body.
This condition manifests itself in numerous symptoms, ranging from muscle weakness to the presence of abnormal synergy causing problems with motor control and the successful completion of everyday tasks. Patients with hemiparesis have issues with both upper and lower extremities, but upper limb rehabilitation can require more effort since fine motor skills are needed almost for any task.
Considering the amount of time to be paid to an individual during rehabilitation training sessions and the shortage of physiotherapists, post-stroke treatment can be quite costly. Understanding this problem, researchers and robot programmers all over the world are tasked with the development of new options facilitating the work of rehabilitation specialists. Since the middle of the 1990s, numerous types of robotic rehabilitation devices based on the use of exoskeletons and end effectors have been presented.
Modern robots for upper limb training differ in terms of the degrees of freedom, the type of feedback, and the available modes of training. Numerous studies included in the annotated bibliography report or cite some positive outcomes of robot-assisted interventions used to facilitate arm rehabilitation of stroke survivors with hemiparesis. The use of robotic devices in rehabilitation training is also beneficial to task distribution and performance tracking since it reduces the workload of physiotherapists and measures patient performance accurately. Based on the evidence from the studies, robot-assisted interventions in arm rehabilitation of hemiparesis post-stroke patients cause improvements in sensorimotor function measured with the help of the Fugl-Meyer scale and increase limb mobility.
The latter, however, does not always lead to complete recovery and the ability to handle objects of any size and shape. Considering the subjective outcomes of robot-assisted training, the patient-reported effectiveness of such interventions varies depending on the type of robot, being high for some used devices such as HandMentor and indefinite for pilot projects and custom devices.
In the end, although the use of rehabilitation robots is generally associated with positive changes in motor planning scores and limb mobility, there is a lack of evidence proving its advantages over non-robotic training sessions with therapists. Instead, it has been demonstrated that a combination of robotic and non-robotic methods of rehabilitation is more effective in reducing arm impairment in sub-acute and acute stages of post-stroke recovery compared to only non-robotic interventions.
Overall, given the heterogeneity of the reviewed studies in terms of quality and sample size, it cannot be recommended to use robot-assisted interventions in any patient with hemiparesis. More studies comparing different types of robots in terms of effectiveness should be done to provide more detailed recommendations concerning their use. At this stage, healthcare providers are advised to avoid using robotic interventions in patients with arm fractures and uncontrolled spasticity and decide on the intensity and length of training sessions based on the stage of recovery.
Introduction
Stroke is widely recognized as one of the most dangerous medical conditions due to mortality rates and a large number of complications that prevent stroke survivors from living a full-fledged life. In the United States, it causes about seven million cases of disability annually (Cherry et al., 2017).
Nowadays, hemiparesis or the incomplete paralysis of one side of the body is considered to be among the most common effects of the condition in question, affecting from one to two-thirds of stroke survivors (Caimmi et al., 2016). Considering the negative effects of arm weakness on the quality of life, modern researchers are concerned with the development of new and effective rehabilitation options adjusted to post-stroke patients’ unique health needs.
Discussing the motor rehabilitation of the upper limbs, it is pivotal to pay attention to the avenues of research that are recognized as promising. For instance, the emergence of robotic technology has caught the interest of specialists in physical rehabilitation and paved the way for the development of rehabilitation robotics. Nowadays, robotic rehabilitation systems are used in different countries to treat both chronic and subacute stroke patients and to encourage motor function improvement (Caimmi et al., 2016).
In general, the implementation of repetitive interventions based on the use of rehabilitation robots is associated with increases in both upper and lower limb mobility, but the underlying mechanisms of such effects are associated with many research gaps (Dierick et al., 2017; Roy, Forrester, Macko, & Krebs, 2013).
In comparison to traditional therapies, robot-aided interventions sometimes demonstrate better effects in post-stroke patients with motor impairments, but there are also studies showing no significant differences between their effects (Caimmi et al., 2016; Nathan, Johnson, & McGuire, 2009). This review is aimed at studying the potential of different robot-aided interventions for upper extremity rehabilitation and identifying any research gaps to improve practical recommendations concerning post-stroke limb training.
Rehabilitation Robots, Robot-Assisted Interventions, and Associated Benefits
Considering the devastating effects of stroke on individuals’ motor performance, the successful post-stroke physical rehabilitation is a process that requires the concerted efforts of healthcare professionals, and the use of high-quality rehabilitation equipment with known effectiveness. Due to the development of computer systems that can control robots and enable them to perceive and process sensory information, many types of robotic devices have appeared during the recent decades (Hagiu, 2016; Hughes et al., 2011).
As for the principles of classification of tools for robotic rehabilitation, some researchers such as Hagiu (2016) emphasize the importance of the affected extremities and distinguish between the devices aimed at lower or upper limb rehabilitation. Apart from helping to increase stroke survivors’ self-service abilities, medical robots can be used in other conditions, including motor impairments after accidents (Hagiu, 2016). However, post-stroke patients present the largest category of people using medical robots for rehabilitation (Hughes et al., 2011; Hagiu, 2016).
One of the first rehabilitation robots, Mit-Manus, was presented in the middle of the 1990s, and it was aimed at facilitating horizontal movements in post-stroke patients with upper extremity paresis (Hagiu, 2016). After these early successes, more devices for upper limb rehabilitation were invented, including those focusing on vertical movements and involving multiple groups of muscles (Hagiu, 2016). In some instances, the rehabilitation of post-stroke patients can be facilitated with the help of industrial robots. For example, the study by Caimmi et al. (2016) describes the applications of industrial robots supplied with end effectors in post-stroke rehabilitation.
Grasp-assistive devices are among the trends in robot-assisted rehabilitation of the upper limbs. As an example, according to Nathan et al. (2009), devices for functional electric stimulation are being replaced by robots that allow facilitating the restoration of fine manipulation tasks such as snatching objects that are tiny or ball-shaped. With that in mind, new robots’ ability to properly measure patients’ finger movements presents an important avenue of research in the field.
Ten years ago, as is clear from the study by Nathan et al. (2009), the development of cost-effective solutions uniting the benefits of grasp-assistive devices and functional electrical stimulation was one of the most promising ideas peculiar to robot-aided rehabilitation after a stroke. In particular, the integration of similar devices with commonly used robots such as ADLER attracted different researchers’ attention.
The robot-assisted rehabilitation of the extremities is facilitated with the help of different approaches to exercise control. As for this review, arm rehabilitation, post-stroke patients with upper limb issues can use robotic systems programmed in five different ways depending on the desired effect of the exercise. To begin with, modern robots for upper extremity training allow creating a passive motion helping to prevent the stiffness of muscles that is common during the first phases of recovery (Poli, Morone, Rosati, & Masiero, 2013).
In such exercises, the device based on robotic technologies helps to move the patient’s affected arm and facilitates the mobilization of the individual’s joints (Poli et al., 2013). The next programming option such as the active exercising mode with no assistance from robotic devices can be used in the end stages of post-stroke recovery when patients are capable of performing exercises on their own and no longer need active help (Poli et al., 2013).
The process of post-stroke recovery usually involves numerous unsuccessful attempts to execute exercises. Therefore, modern rehabilitation robots have the mode that involves active assistance; it is used when patients try to make arm movements and still need help to prevent improper positions and traumas (Poli et al., 2013). As for the next option that is available to post-stroke patients with paretic arms, the so-called resistive mode allows executing exercises against the robot’s force, which helps to overcome physical weakness and strengthen the affected muscles in a non-traumatic way (Poli et al., 2013).
Finally, the exercising modes in upper extremity rehabilitation robots allow implementing both unimanual and bimanual exercises (Poli et al., 2013). The necessity of the latter is explained concerning the number of everyday tasks that require the symmetry of arm movements, but the ideas concerning the increased effectiveness of bimanual exercises do not find extensive support.
Robotic rehabilitation devices available today greatly vary when it comes to the number of functions, exercising modes, and the opportunities for patient-robot communication. In their review, Poli et al. (2013) single out eight types of robots that are different in terms of the degrees of freedom (the degree of flexibility), the focus of rehabilitation exercises, mechanical peculiarities, the presence of additional devices such as visual displays, and the release of feedback.
Being among the functions of new-generation rehabilitation robots, the provision of the so-called extrinsic or objective feedback allows reporting the results of exercises, thus increasing the levels of patient involvement and improving motor learning (Poli et al., 2013).
Feedback is often listed among the key factors that explain the growing popularity of robot-assisted interventions and their effectiveness for patients’ ability to care about themselves daily. According to Liu, Li, and Lamontagne (2018), the presence of feedback has a significant influence on the acquisition and restoration of patients’ motor skills since it informs the users of rehabilitation robots about movement quality in terms of easiness and freedom or the degree to which their attempts to perform exercises are successful. With that in mind, the opportunities to improve the quality of the released feedback interest modern developers of rehabilitation robots.
Continuing on the role of feedback in rehabilitation robotics, it is important to review two popular approaches to its use in exercises aimed at the restoration of limb function. Using the first one known as the error reduction approach to limb rehabilitation, developers emphasize the role of the optimal movement patterns defined based on the motor performance of patients’ non-affected arms or, in some cases, therapists’ observations (Liu et al., 2018).
The paradigm implies that frequent demonstration of the preferred line of movement helps the users of robotic devices to imitate it and reduce the incidence of deviations from it (Liu et al., 2018). Taking it into account, in the paradigm being discussed, extrinsic feedback is used to set limitations and inform users if they make mistakes when performing exercises. In contrast, the approach that emphasizes the role of resistance and error augmentation in rehabilitation is based on the idea that all mistakes related to the trajectory of movement can be beneficial in terms of the speed of learning and motor adaptation (Liu et al., 2018).
Researchers supporting this paradigm believe that instead of preventing any errors, it is possible to use artificially created barriers to mobilize the musculoskeletal system and encourage users to find ways to help them to move their limbs properly and adapt to physical difficulties.
Stroke-Related Hemiparesis and Rehabilitation Needs
As has been mentioned earlier, stroke belongs to the most common causes of physical disability in adults since it involves blood flow abnormalities resulting in the death of healthy brain cells. Being one of the most well-known physical effects of stroke, hemiparesis is a cause of movement deficits in the side of the body contralateral to the damage center in the brain (Poli et al., 2013). In contrast to other conditions such as hemiplegia, hemiparesis does not involve a complete paralysis of the upper or lower extremities and therefore, patients with paretic limbs have more treatment options helping them to improve movement impairments (Poli et al., 2013).
In addition to being a common consequence of stroke, hemiparesis can be congenital or result from traumatic brain injuries or even brain tumors (Poli et al., 2013). In the cases of hemiparesis, the most effective approaches to treatment should be chosen based on the cause of the condition and patients’ specific needs. Some authors whose works are included in the literature review touch upon the unique rehabilitation needs of stroke survivors with paretic limbs.
The stroke-related weakness of one side of the body manifests itself in a range of symptoms. As for the key signs of the condition in question, they are presented by muscular weakness peculiar to specific muscles, abnormalities in muscle tone responsible for the inability to maintain posture, and several similar consequences (Poli et al., 2013).
Due to problems with lower limb motor units that are common after strokes, patients with hemiparesis experience significant issues when trying to maintain equilibrium and perform some basic daily tasks that require balance. In terms of the most visible manifestations of the condition, stroke survivors with hemiparetic extremities demonstrate such symptoms as the lack of arm or leg mobility and the presence of pathological synergistic movements (Poli et al., 2013).
Typically, the latter tend to occur during the early stages of post-stroke rehabilitation. They involve a series of involuntary movements after the attempts to make a specific movement involving extremities of the affected side of the body (Poli et al., 2013). The abnormal patterns of synergistic movement cause numerous inconveniences since they prevent stroke survivors from completing the majority of everyday tasks such as dressing, brushing the teeth, or holding cutlery.
Being associated with persistent movement impairments, hemiparesis significantly affects stroke survivors’ quality of life. Despite that, hemiparesis patients have certain chances to recover from this condition. The situation with motor and functional recovery is much worse for individuals with more severe complications after strokes. For instance, according to the systematic literature review, when it comes to the prognosis for initial paralysis, less than fifteen percent of patients with this issue completely recover the limbs’ motor function (Hughes et al., 2011).
In hemiparesis, the rehabilitation of the upper extremities presents a pivotal task due to the likeliness of limitations related to self-care and independence. For example, it is known that the success of daily routine tasks involving different parts of the upper limbs greatly depends on the quality of fine motor skills (Liu et al., 2018).
Despite the significance of rehabilitation needs related to the function of the human arms, a few studies conducted in the 1980s suggest that the percent of people with stroke-related conditions who manage to achieve the functional level of upper limb activity ranges from 35 to 70 (Liu et al., 2018). Therefore, as is clear from the existing literature, the estimates concerning stroke-related hemiparesis recovery rates may vary.
The rehabilitation needs of patients with hemiparesis resulting from a stroke primarily refer to the prevention of further declines in self-care functions. Recognizing the pivotal role of self-care in successful disease recovery, modern healthcare specialists facilitate these patients’ reintegration into work and social life by providing focused care and rehabilitation exercises of high intensity (Poli et al., 2013).
Taking into consideration the high costs of individual training sessions with physiotherapists and limited budgetary resources, the development of new treatment options for hemiparesis patients is widely supported (Poli et al., 2013). Thus, the most recent advances in rehabilitation robotics are related to both effectiveness and financial considerations.
The traditionally used rehabilitation interventions for the treatment of post-stroke hemiparesis include exercises helping to increase the range of motion that are classified into three categories according to the type of movement. ROM exercises for upper limb hemiparesis are active in the movement is initiated by patients, passive if they do not control arm movement, or active-passive when patients’ and therapists’ concerted efforts take place (Liu et al., 2018; Poli et al., 2013).
Continuing on the key rehabilitation tasks, it needs to be noted that hemiparesis patients’ needs related to the restoration of coordination and spatial recognition skills require the use of stretching exercises that can be helpful in individuals with non-severe mobility issues. Considering the nature of the problem in question, approaches to the rehabilitation of stroke survivors with hemiparetic limbs should involve exercises chosen based on patients’ specific situation, the severity of hemiparesis, general health, and other factors such as the presence of traumas or long-term musculoskeletal conditions.
In the end, hemiparesis belongs to the number of conditions that significantly affect patients’ ability to care for themselves and live independently due to many symptoms related to movement coordination, synergy, the accuracy of movements, and the range of motion. Despite these symptoms’ impact on the ability to perform daily tasks, hemiparesis does not involve a full paralysis of upper and lower limbs, which increases these patients’ chances of recovery. Considering the symptoms of the condition, the associated rehabilitation needs include increasing the range of motion and motor control and eliminating abnormalities related to muscle synergy.
Robot-Assisted Interventions in Paretic Arm Rehabilitation and Their Effectiveness
As is clear from modern researchers’ findings, the use of robotic rehabilitation systems is associated with numerous advantages referring to improvements in both patients’ locomotor activity and care coordination. For instance, according to Zollo et al. (2011), medical robots contribute to the quality of training that improves cognitive and physical functions, thus helping to optimize the structure of rehabilitation strategies.
Moreover, the use of robotic devices improves the quality of rehabilitation interventions due to its effects on task distribution. When machines become responsible for the tasks that involve physical exertion or unpleasant physical manipulations, operators are enabled to focus on the exercises and performance (Zollo et al., 2011). About the latter, rehabilitation devices provide new opportunities for tracking individuals’ progress due to the accuracy of performance measurements (Zollo et al., 2011).
Various exoskeleton-based rehabilitation devices are extremely good at controlling patients’ arm joints and bones. Nevertheless, due to the peculiarities of their construction, they are not easy to attach and their use may require substantial adaptation efforts (Bertomeu-Motos et al., 2018).
More than that, in some cases, the use of exoskeletons for rehabilitation exercises does not guarantee success due to the threat of traumas stemming from the risk of misalignment between the device and the limbs (Bertomeu-Motos et al., 2018). When it comes to the use of end-effector robots in arm rehabilitation, it is associated with higher levels of user-friendliness because of a small number of the upper limb parts involved in exercises. Due to these robots’ construction, they can be adapted to different conditions and therefore, be helpful not only in the cases of arm function impairments after a stroke but also in the treatment of traumatic injuries (Bertomeu-Motos et al., 2018).
At the same time, however, the joint configuration of these robots allows them to provide data related only to the trajectory of the robot’s end-effector device and the interaction of the device and the limb (Bertomeu-Motos et al., 2018). In this situation, the opportunities to improve rehabilitation interventions based on the arm joints data are limited.
In the context of rehabilitation robotics, it is pivotal to discuss modern authors’ opinions concerning the role of robot-assisted training sessions concerning conventional rehabilitation techniques. According to the review article by Poli et al. (2013), modern rehabilitation robots are generally considered as a successful supplement to conventional therapies using non-robotic approaches to the recovery of limb functions. Apart from being regarded as a method that makes rehabilitation therapy more intensive and facilitates its frequent use, robot-aided rehabilitation presents a means of stimulating movement in a well-organized and controlled way (Poli et al., 2013).
Robotic devices are believed to demonstrate a high potential when it comes to rehabilitation – in many instances, they can be used under the care of a therapist to improve the results of non-robotic approaches by increasing training intensity (Poli et al., 2013).
According to the conclusions made by Poli et al. (2013) in their review article, robotic devices for limb rehabilitation add to the effectiveness of other post-stroke treatment methods since they allow increasing the length of separate training sessions at no cost in quality. In addition to that, robots are regarded as extremely helpful when rehabilitation interventions are expected to save therapists’ time and provide treatment flexibility by using several functional modes.
Along with similar questions, the effectiveness of different types of feedback in the promotion of motor learning was one of the key topics that interested researchers in the field of rehabilitation robotics in the 2000s (Poli et al., 2013; Hagiu, 2016). Thus, the great role of visual, kinematic, and verbal feedback in motor learning abilities of healthy test subjects was demonstrated by Van Vliet and Wulf in 2006 and involved some implications to post-stroke robot-assisted rehabilitation (Poli et al., 2013).
Three years later, the study conducted by a group of Italian researchers including Casadio, Giannon, Morasso, and Sanguineti proved the effectiveness of both visual and proprioceptive feedback in improving the upper and lower limb motor control in post-stroke hemiparesis patients (Hagiu, 2016).
In addition to visual feedback released with the help of special displays, modern devices widely use haptic communication to provide users with assistance during training sessions. For instance, the assistive glove to be integrated into different robotic systems such as ADLER designed by Nathan et al. (2009) utilizes a peculiar fingertip construction to make use of haptic feedback. Therefore, the importance of extrinsic feedback belongs to the key themes linked to the outcomes of robot-assisted interventions, and its quality can be potentially related to rehabilitation results in post-stroke populations.
In their qualitative work, Cherry et al. (2017) study a sample of post-stroke veterans with hemiparesis using robotic devices for home-based upper or lower limb training and generalize on the patient-perceived pros and cons of robotic rehabilitation. HandMentor, a robotic device providing active assistance in rehabilitation, was used for upper limb training two hours a day for three months. To promote progress, the participants were required to raise the bar regularly and proceed from the easiest exercises to the most difficult ones.
The thematic analysis of interviews conducted by Cherry et al. (2017) demonstrates the following themes associated with the use of HandMentor: increases in the mobility of impaired arms observed by the test subjects or their caregivers/healthcare providers, the improved patient-perceived unity of the mind and the body, flexibility in treatment organization, and reduced anxiety levels. Despite the presence of barriers related to the proper use of robotic devices, the positive impact of rehabilitation robots on the mobility of the impaired extremities was among the most represented themes, indicating the success of active assistance exercises in arm rehabilitation.
Similar to Cherry et al. (2017), Hughes et al. (2011) analyze the perceived effectiveness of robot-assisted interventions used to treat stroke survivors with severe hemiparesis. In the experiment conducted by Hughes et al. (2011), five patients with chronic hemiparesis received eighteen hours of robot-assisted training with the help of a device using a robotic arm and the repetitive method of system control (ILC) mediated by electric stimulation.
The intervention involved a series of robot-assisted tracking tasks performed using almost thirty different movement trajectories (Hughes et al., 2011). The differences between the pre- and post-intervention isometric force and Fugl-Meyer motor scores were significant (the p-value of 0.02), as distinct from the results of the action research arm test (Hughes et al., 2011). Concerning the study’s qualitative component, the thematic analysis of structured/half-structured 30-minute interviews helped retrieve the following themes: the usability of the robotic system, the participants’ divided opinions concerning the short-term effectiveness of the system, and all patients’ willingness to recommend the intervention to other stroke survivors (Hughes et al., 2011).
Despite the absence of changes in ARAT pre- and post-intervention scores that would be clinically relevant, some participants observed positive changes related to the degree of impairment and limb function (Hughes et al., 2011). Thus, the patient-reported effectiveness of the intervention varied from person to person.
Some statements found in the existing literature prove that the success of robot-assisted interventions for upper limb rehabilitation is inextricably connected to the stage of post-stroke recovery and parts of the arm involved in exercises. An example of such conclusions is present in the article by Hagiu (2016) containing references to the systematic review of randomized controlled trials conducted by Norouzi-Gheidari, Archambault, and Fung in 2012. According to the review of twelve RCTs published between 1997 and 2009, robot-assisted interventions aimed at paretic arm rehabilitation do not excel conventional therapies in terms of effectiveness (Hagiu, 2016).
However, in the acute and subacute stages of post-stroke recovery, the combination of conventional and robot-assisted interventions is more effective for increasing the mobility of paretic elbows and shoulders compared to conventional therapies (Hagiu, 2016). Therefore, robot-assisted rehabilitation interventions are not more effective for arm rehabilitation in hemiparesis than non-robotic training sessions with therapists.
To define the degree to which robots are effective for upper limb rehabilitation, attention should also be paid to the characteristics of patients with hemiparesis. The study by Colombo, Sterpi, Mazzone, Delconte, and Pisano (2013) reviews the hypothesis developed by Kwakkel et al. in 2006. According to it, stroke patients are different in terms of their capacity to recover, and progress in rehabilitation may depend on the intensity and length of training sessions, exercise environment, and similar factors instead of being predetermined only by the reduction of neurological impairments (Colombo et al., 2013).
Based on this suggestion, the outcomes of rehabilitation therapy may need to be analyzed concerning patients’ initial conditions and other factors. In their literature review section, Colombo et al. (2013) also mention the study conducted in 2008 by a group of researchers led by Volpe. According to its results, intensive movement therapies delivered by robots (InMotion2) and therapists are effective for arm motor performance in hemiparesis patients, but robot-assisted training sessions do not cause more significant improvements compared to conventional training (Colombo et al., 2013). Thus, despite robots’ overall effectiveness for upper extremity rehabilitation, their advantages over non-robotic training sessions with therapists remain a doubtful idea.
Among the chosen articles, some works are focusing on the effectiveness of end-effector rehabilitation robots in the restoration of arm function. As an example, the study by Zollo et al. (2011) is devoted to the outcomes of robot-assisted interventions based on the use of InMotion 2 and InMotion3 to improve chronic stroke patients’ ability to control the affected arm. After six weeks of robot-aided therapy for wrists and both elbows and shoulders involving three types of games, statistically, significant reductions in the levels of impairment were found (Zollo et al., 2011).
Discussing the results, the researchers report an increase of Fugl-Meyer scores by almost 10% and a 7.4% increase in motor planning scores (Zollo et al., 2011). These results, the researchers believe, align with the findings reported by Brewer and her colleagues in 2007, according to which robot-aided therapies that are goal-directed are extremely effective for arm rehabilitation in post-stroke hemiparesis in terms of functional improvement and the involved muscles’ strength (Zollo et al., 2011). These findings support the hypothesis concerning the effectiveness of robot-assisted interventions for arm rehabilitation in hemiparesis.
To contribute to the field, Colombo et al. (2013) have studied a sample of forty stroke patients (both subacute and chronic) with unilateral brain lesions resulting in limb impairments. All patients had at least three weeks of upper extremity training (5 hours every week) using MEMOS – elbow-shoulder manipulators with two degrees of freedom and a robotic workstation based on haptic technology (Colombo et al., 2013).
Training sessions consisted of the phases of unassisted movement and robot-assisted exercises. Based on the analysis of patients’ Fugl-Meyer arm motor scores and muscle plasticity before and after the experiment, both sub-acute and chronic patients managed to improve their impairments (F = 44.25, p < 0.001) (Colombo et al., 2013). However, about the FM scale, sub-acute test subjects’ improvement (12.05±9.61) was more obvious than that of chronic stroke patients (3.00±2.61) (Colombo et al., 2013).
The results justify conclusions that in patients in the sub-acute phase of recovery, longer rehabilitation programs are needed to achieve maximum positive results (Colombo et al., 2013). Also, they demonstrate that robot-assisted interventions positively impact the motor recovery of the impaired arm in post-stroke patients.
The use of robots for rehabilitation is generally associated with positive outcomes both for patients and operators. In general, speaking about the benefits of robot-assisted interventions for stroke patients, modern researchers claim that this approach to rehabilitation allows increasing training intensity, focusing on the role of patients in exercises, and using repetition to achieve positive results (Zollo et al., 2011). These conclusions are supported by Nathan et al. (2009) who regard motivation and repetition as the essentials of successful rehabilitation after a stroke. Therefore, many researchers support the opinion that robot-assisted rehabilitation is effective in setting priorities and attracting more attention to patients, their limb functioning, and exercise performance.
The development of rehabilitation options aimed at improving the functions of different parts of the upper extremities presents an important task in rehabilitation robotics. Nathan et al. (2009) believe that in the majority of cases, robotic interventions for upper limb recovery emphasize the importance of reaching tasks, whereas tasks that involve fingers remain underestimated. The attempts to broaden the range of exercises available in robot-assisted rehabilitation resulted in the creation of new multi-purpose robots such as ADLER (Nathan et al., 2009). In comparison to robots that facilitate only arm movement, ADLER uses assistive devices to improve patients’ ability to grasp objects, thus focusing on fine movements (Nathan et al., 2009).
Because of the chosen research question, it is pivotal to characterize the state of knowledge concerning the types of devices specifically aimed at upper limb functional improvement. Robotic devices facilitating the rehabilitation of the upper extremities can vary depending on the way that they are attached to the impaired limb. According to Bertomeu-Motos et al. (2018), the motor function of the upper extremities can be improved with the help of exoskeleton-based devices that align robot axes with the axes of extremity segments controlling specific joints.
The second well-known type of device used in arm rehabilitation is presented by end-effector robots that focus on the distal segments of the limbs (Bertomeu-Motos et al., 2018). Judging from the differences related to their operation, the two classes of rehabilitation robots are dissimilar in terms of benefits for patients who have difficulties with arm movement.
Discussion
The effectiveness of robot-assisted interventions in the rehabilitation of the upper extremities in post-stroke subjects with hemiparesis presents an important question, and modern researchers try to approach it in different ways. For instance, in addition to measuring the objective changes in arm movement characteristics, there have been some attempts to study these interventions’ outcomes in terms of subjective experiences.
As is demonstrated in the previous section, the studies focusing on patient-perceived effectiveness of rehabilitation robots are not widely represented in the chosen literature. Some of the reviewed studies demonstrate that patients’ perceptions of the effectiveness of robot-assisted interventions vary, but many of the research subjects are still willing to recommend robotic rehabilitation to friends and acquaintances with stroke-related arm impairments.
Apart from the importance of patient-reported benefits of robotic rehabilitation, many studies discuss the outcomes of robot-assisted arm training concerning conventional approaches. Although many researchers observe arm motor improvements in their test subjects, which point at the effectiveness of robots for rehabilitation, some statements regarding robotic devices and their ability to replace other methods of post-stroke rehabilitation may seem too bold.
For instance, although robotic rehabilitation has benefits related to performance analysis and the costs of treatment, there is no solid evidence that it can replace physiotherapy. However, it is accepted that a combination of the two approaches can be more effective for upper limb rehabilitation in post-stroke populations than conventional therapies.
Conclusion/Recommendations
In the end, rehabilitation robots are widely used to facilitate recovery from different conditions such as stroke or car accident-related traumas. Concerning their effectiveness for arm rehabilitation in hemiparesis patients, it is demonstrated in the reviewed literature concerning some types of robots such as HandMentor, InMotion2, InMotion3, and the MEMOS robotic system. However, despite their benefits for upper limb functioning and mobility, robotic-assisted interventions alone are not enough for arm rehabilitation. Based on the literature review and the existing research gaps, the following recommendations can be provided:
- More studies are required to comparatively analyze different types of robots for arm rehabilitation with attention to practical outcomes;
- In acute and sub-acute stages of stroke recovery, the use of both robot-assisted interventions and conventional therapies maximizes positive outcomes;
- Compared to chronic stroke patients, individuals in the sub-acute phase need longer training sessions to improve arm use;
- Robot-assisted interventions should not be used in stroke patients who have arm fractures and severe upper limb spasticity.
References
Bertomeu-Motos, A., Blanco, A., Badesa, F. J., Barios, J. A., Zollo, L., & Garcia-Aracil, N. (2018). Human arm joints reconstruction algorithm in rehabilitation therapies assisted by end-effector robotic devices. Journal of Neuroengineering and Rehabilitation, 15(1), 1-11.
Caimmi, M., Visani, E., Digiacomo, F., Scano, A., Chiavenna, A., Gramigna, C.,… Panzica, F. (2016). Predicting functional recovery in chronic stroke rehabilitation using event-related desynchronization-synchronization during robot-assisted movement. BioMed Research International, 2016, 1-11.
Cherry, C. O. B., Chumbler, N. R., Richards, K., Huff, A., Wu, D., Tilghman, L. M., & Butler, A. (2017). Expanding stroke telerehabilitation services to rural veterans: A qualitative study on patient experiences using the robotic stroke therapy delivery and monitoring system program. Disability and Rehabilitation: Assistive Technology, 12(1), 21-27.
Colombo, R., Sterpi, I., Mazzone, A., Delconte, C., & Pisano, F. (2013). Robot-aided neurorehabilitation in sub-acute and chronic stroke: Does spontaneous recovery have a limited impact on outcome? NeuroRehabilitation, 33(4), 621-629.
Dierick, F., Dehas, M., Isambert, J. L., Injeyan, S., Bouché, A. F., Bleyenheuft, Y., & Portnoy, S. (2017). Hemorrhagic versus ischemic stroke: Who can best benefit from blended conventional physiotherapy with robotic-assisted gait therapy? PloS One, 12(6), e0178636.
Hagiu, B. A. (2016). The physiotherapist will be replaced by robot? Sport & Society/Sport si Societate, 16, 53-57.
Hughes, A. M., Burridge, J., Freeman, C. T., Donnovan-Hall, M., Chappell, P. H., Lewin, P. L.,… Dibb, B. (2011). Stroke participants’ perceptions of robotic and electrical stimulation therapy: A new approach. Disability and Rehabilitation: Assistive Technology, 6(2), 130-138.
Liu, L. Y., Li, Y., & Lamontagne, A. (2018). The effects of error-augmentation versus error-reduction paradigms in robotic therapy to enhance upper extremity performance and recovery post-stroke: A systematic review. Journal of Neuroengineering and Rehabilitation, 15, 1-25.
Nathan, D. E., Johnson, M. J., & McGuire, J. R. (2009). Design and validation of low-cost assistive glove for hand assessment and therapy during activity of daily living-focused robotic stroke therapy. Journal of Rehabilitation Research & Development, 46(5), 587-602.
Poli, P., Morone, G., Rosati, G., & Masiero, S. (2013). Robotic technologies and rehabilitation: New tools for stroke patients’ therapy. BioMed Research International, 2013, 1-8.
Roy, A., Forrester, L. W., Macko, R. F., & Krebs, H. I. (2013). Changes in passive ankle stiffness and its effects on gait function in people with chronic stroke. Journal of Rehabilitation Research & Development, 50(4), 555-572.
Zollo, L., Rossini, L., Bravi, M., Magrone, G., Sterzi, S., & Guglielmelli, E. (2011). Quantitative evaluation of upper-limb motor control in robot-aided rehabilitation. Medical & Biological Engineering & Computing, 49(10), 1131-1144.