The future role of robots in neuro-rehabilitation

KEYWORDS:

• Evidence
• neuro-rehabilitation
• physiotherapy
• robot-mediated therapy
• stroke

1. Past and present

The World Health Organization estimates that about 15% of the world’s population has some form of disability [1]. Rehabilitation has a key role in decreasing the level of disability. Application of advanced technologies in rehabilitation is a promising opportunity to attain this goal.

The development of rehabilitation robots started in the late 1980s. The following decade was a pioneering phase. After the year 2000, the first representatives of commercially available robots appeared. These devices can assist in practicing upper or lower limb movements and motor relearning, and in developing proprioception, cognitive functions, and attention. There is equipment that patients can use to practice the same movements as with the robots, but it does not provide mechanical assistance; so patients have to rely on their own strength. The emphasis is on high repetition, interactive and personalised therapy. The aim is to attain a higher level of function in a shorter time frame. The philosophy of the application of robots in rehabilitation is not to replace the therapist, but to widen treatment options [2].

The two main goals of therapy with rehabilitation robots are to develop upper limb function and to support gait re-education.

Rehabilitation robots are used mainly following central nervous system damage, primarily after stroke. Multiple clinical trials and meta-analyses have been conducted regarding these robots. Concerning the efficacy of electromechanical arm training, Mehrholz et al. [3] analysed 45 randomized controlled trials with 1619 participants. They found that this kind of therapy promotes improvement in arm function and muscle strength, as well as execution of activities of daily living. Nevertheless, the methodologies of the studies were quite different, and 24 different devices were used.

Robot-mediated training on a treadmill is a widely used method for gait re-education. Mehrholz et al. [4] selected 36 trials with 1472 subjects in a Cochrane review and found that post-stroke patients who received such training in addition to traditional physiotherapy were more likely to achieve independent walking, than subjects who received traditional therapy alone.

These results suggest that robot-mediated therapy gives certain advantages for patients, at least in motor relearning; but several unanswered questions remain: Which kind of exercises are most beneficial, in which phase, how often, and for how long in order to yield optimal results?

To compare routinely used types of upper limb therapy, Pollock et al. evaluated 40 reviews covering 18 different interventions (including robotics) in a Cochrane review [5] and stated that the available evidence is insufficient to determine the most effective therapy for improving upper limb function. The authors urged researchers to conduct large randomized controlled trials to find the best practice.

Although thousands of patients took part in research trials, it is difficult to compare the results. As yet, no unified methodology or generally used outcome measures for the trials exist; this makes evaluation uncertain [6]. However, some initiatives have been proposed for harmonizing methodologies. The European Network for Robotics in Neurorehabilitation (in the scope of the European Cooperation in Science and Technology – EU COST Action) has published recommendations on protocols for assessment in clinical practice and research in rehabilitation robotics [7]. There is another reason that can make the evaluation more difficult: there are non-robotic devices which provide similar exercises but without robotic assistance. In several studies such devices are used and mentioned as ‘robots’.

Studies conducted on the basis of harmonized methodologies could result in a higher level of evidence, which could help create clinical guidelines for application of rehabilitation robotics. According to the present American guidelines for rehabilitation for adults post-stroke, the application of robotic devices for improving mobility ‘may be considered’; and in improving upper limb function, it is ‘reasonable to consider’ [8]. The British guidelines accredited by the National Institute of Health and Care Excellence (NICE) recommend the consideration of electromechanically assisted gait training for those who cannot walk independently after stroke, but also recommend offering robot-assisted arm training to patients only in the context of a research study [9]. The reason for these quite cautious guidelines could be that only weak evidence can be established based on the diverse, hardly comparable trials. It seems that low treatment dose and duration can negatively affect the results. Nevertheless, robot-mediated therapy is considered also because it has very few side effects [10].

Most of the devices make it possible to exercise in a virtual environment [11]. Very few robots provide practice using real objects in a real environment; one of them is the Reharob system [12]. Application of virtual reality is easier to solve technically than practicing in a real environment, but does it have the same efficacy? The rehabilitation goal is not simply to improve upper limb function in any field, but to decrease the dependency level of the patient in activities of daily living. Reaching higher scores in video games is surely useful, but it does not automatically mean that the patient will be able to dress, eat, and bathe more independently. Moreover, video games exploit an important but narrow part of the range of motion of the upper extremities.

The structure of the brain is not static. It is dynamic, and plasticity is its fundamental property. The therapist aiding a patient in the process of relearning a function must take plastic changes into account and conduct therapy accordingly [13]. For this reason, we should know which kinds of robot-mediated exercises are the most useful for stimulating plasticity [14].

To answer the questions at hand, there are several ongoing clinical trials. The Robot Assisted Training for the Upper Limb after Stroke (RATULS) study [15] probably involves the most participants, more than 700 patients in three randomized groups. The aim of this huge trial is not only to compare the efficacies of the usual, the enhanced traditional, and the robot-mediated therapies, but also to conduct economic analysis. The RATULS trial results are expected to be available at the end of 2019.

Artificial intelligence (AI) of robots can give further options. It is not enough if robots are simply therapist extendors and save human efforts when executing dull and exhausting exercises. Artificial intelligence could be an additional advantage of robots: continuously evaluating the patients’ status and choosing the next exercise accordingly.

The human therapist is able to deal with the whole patient at the same time: both physically and mentally. There is a continuous multilevel interaction between the therapist and the subject. Current rehabilitation robots can move only certain segments of the limbs, not the whole body; they have restricted sensory input, and they are able to make only quite simple decisions. It would be reasonable to involve at least the whole upper limb from the shoulder to the fingers, because in real life people use these parts together, in coordination with each other. We can suppose that moving the whole upper extremity is necessary for restoring correct inter-joint coordination. Nevertheless, in earlier stages, where more serious palsy is present, the involvement of one or two joints can be enough. As the patient recovers, the involvement of more segments is required. Currently, two robots are necessary in order to conduct exercises with both the proximal and distal part of the limb simultaneously. Technical development should make it possible to do this with one device.

There are advanced technological solutions in gait re-education for each stage of rehabilitation: robots are useful even in the earliest stage, to help mobilize the patient into a vertical position.

In the early stage, robots may be involved in treadmill training; later, treadmill training can continue without robots. Later still, patients may walk in their real environment with wearable exoskeletons, which represent a relatively new type of rehabilitation robotics. Nevertheless, the first clinical trials and reviews are promising; they can provide the ability to walk even for certain non-ambulatory patients with spinal cord injury [16]. These robots can serve not only as therapeutic, but also as assistive device.

Opposite to the pharmaceutical trials, in case of robot mediated rehabilitation larger studies are usually executed after commercialization. It would be reasonable to prove the efficacy of robotic devices before putting on the market. This is not only a commercial, but also an ethical requirement.

1. Future challenges

In conclusion, there are several issues to be solved in the field of rehabilitation robots and many questions to be answered. The following points should be considered in the future.

Clinically:

• Harmonized methodologies for clinical trials, including assessments that make the trials comparable and suitable for meta-analysis are needed. Research guidelines for robotic trials could be very useful. Clear definitions and names for robotic and non-robotic high tech devices should be necessary not to confuse them.

• The above studies would provide guidance for answering questions regarding the procedures for patient selection and treatment. Trials should aim to find which patients may benefit more from the robotic rehabilitation [17].

• Once we have evidence, and not only experience, clinical guidelines should be developed.

Technically:

• Maximize the full range of motion.

• Gain more sophisticated sensory input from the patient.

• Use AI to allow decisions to be made regarding the therapeutic programme.

• Develop exercises that use real objects, not just virtual reality, to practice functions that are activities of daily living.

Other aspects:

• Develop wider options for home use of robots.

• Develop wider options for application of robots not only as training devices but also as assistive devices in daily life.

• Improve the cost:benefit ratio.

• Development of AI will make necessary to pay much attention to ethical rules [18].

The use of robots and other interactive devices is a promising opportunity to widen treatment options and improve outcomes of patients with disabilities that are mainly due to central nervous system impairments. Clinical trials suggest that current robotic tools can provide certain advantages for patients. It also seems possible that the usefulness of robots can be improved by further technical development, together with their targeted setting in the rehabilitation programme.

Declaration of interest

The authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

Reviewer disclosures

Peer reviewers on this manuscript have no relevant financial or other relationships to disclose.

Additional information

Funding

This paper was not funded.