Introduction

Various motor disturbances occur in more than 80% of survivors after stroke and traumatic brain injury (TBI). These violations is one of the main causes of disability. Analysis of statistical data shows that incidence of cerebrovascular diseases has been steadily increasing in recent years not only in elderly people but also among young employable people [1]. Most victims with TBI are also employable people. These are men younger 40 years and the most active part of the Russian population [2]. These data increase the relevance of the most effective rehabilitation for complete and rapid recovery of patients. Severe motor disorders is the main cause of disability after focal brain lesion. These are paresis, low or high muscle tone, followed by limb deformation, pain, impaired active or passive function of limb in everyday life, etc. New principles of rehabilitation with active inclusion of innovative methods have been actively introduced in the last decade. Moreover, searching for new rehabilitation methods is continued. Motor rehabilitation is started at the intensive care units immediately after hospitalization. Correction of life-threatening disorders is associated with postural treatment, passive kinesiotherapy, early verticalization, etc. However, there is still no sufficient evidence base for the use of mechanized, robotic and virtual technologies in acute period disease. As soon as clinical stabilization is achieved, various kinesiotherapeutic programs, new methods based on biological feedback (BFB), transcranial magnetic stimulation, botulinum therapy, kinesiotherapy, metabolic therapy [3], etc. are applied.

Robotic and computerized techniques have increasingly been included in rehabilitation programs of patients with neurological profile in recent years. Various studies confirmed that these approaches could successfully supplement conventional methods although modern methods of neurorehabilitation cannot completely replace traditional techniques. Medicine is a special area for another perspective method — virtual reality (VR). This approach has been constantly improved recently.

VP (Latin virtus — possible, potential; realis — existing, real) is an artificial world created by technical means and transmitted to a person through the senses (vision, hearing, smell, touch) [4]. The first VR device was proposed in 1967 and looked like a massive helmet that was hung from the ceiling. However, realism of transmitted image was low and required significant refinement. Improvement of this system is ongoing today [5].

There are seven mechanisms for moving a user into a virtual environment: with visualization (direct, indirect, augmented reality, avatar, tracking, combined) and without visualization [6]. However, it is still unclear what imaging mechanism is preferable in a particular clinical case. There is probably a difference in effectiveness of various visualization mechanisms during motor recovery of upper and lower extremities. So, some studies [7, 8] shown that “avatar” is preferable for restoration of muscle strength of the upper extremities. It is known that most exercises performed by hands require visual information about limb position in relation to the environment. For example, the correct capture of an object is not possible without this option. “Avatar” visualization mechanism results adequate representation of the human hands movements in a virtual environment.

VR mechanism is quite complex. Imagination of movement is a relatively new approach to rehabilitation. Presumably, imaginary movement activates motor cortex of the contralateral hemisphere as intensively as voluntary movement. In addition, the process of imaginary movement included cerebellum, subcortical nuclei, parietal lobe and circular gyrus in addition to motor cortex [9, 10]. EEG confirms these data. Voluntary and imaginary movements are associated with desynchronization of μ-rhythm (10—12 Hz), i.e. one can observe activation of patient’s brain during VR-assisted rehabilitation [11, 12]. Motor rehabilitation may be complicated by concomitant disturbances of sensitivity and coordination [6, 13].

In recent years, the number of studies on the use of VR in medical rehabilitation programs has been steadily increasing. The leading positions are occupied by Korea and the USA. The most popular VR systems for medical rehabilitation of patients after stroke are EyeToy and Sony PlayStation, Nintendo Wii, Microsoft Kinect and GestureTek IREX [14]. Both specialized medical programs and video games are used in these systems. Registration of movements can be carried out by joysticks, touch gloves, suits, exoskeletons, etc.

The most famous are EyeToy video games (Sony PlayStation II console) with imitation of boxing, basketball, bowling, etc. [15]. D. Rand et al. [16], G. Yavuzer G. et al. [17] found that patients with mild-to-moderate motor impairment achieve the best results using EyeToy games. In Russia, the first research devoted to motor rehabilitation of patients after stroke followed by paresis of the arm was carried out in 2010. Virtual environment (EyeToy game, Sony PlayStation II console) was applied in this study. The technique improves muscle strength in proximal and distal parts of the upper limb, increases speed and accuracy of movements [18]. Saposnik G. et al. [19] studied rehabilitation of patients after stroke using Nintendo Wii video games. Each game is aimed at training a specific muscle group. For example, imitation of bowling involves proximal parts of the arm, while imitation of tennis – distal muscles. E. Pietrzak et al. (2014) analyzed 13 studies devoted to the use of video games in rehabilitation of patients after stroke and noted the effectiveness of Nintendo Wii [5, 20]. Chirivella P. et al. (2014) analyzed application of NeuroAtHome rehabilitation software platform with a Microsoft Kinect sensor for recovery of the lost functions in patients after stroke. In this system, visualization mechanism is an avatar, virtual environment is a street, a sports hall, etc. Exercises improve function of the upper and lower extremities, tilts and rotation of the head and trunk. The authors noted that the NeuroAtHome VR system with a Microsoft Kinect sensor can increase motor activity and improve rehabilitation potential [21, 22]. T. Vanbellingen et al. [23] analyzed other video games in rehabilitation after stroke and found improvement of hand strength and dexterity too. A significant positive aspect was preserved motivation to continue further rehabilitation in 3 months after its initiation.

Patients need to adapt to the environment, restore basic skills including everyday ones after a stroke. Training of fine motor skills is the most valuable. Therefore, the most popular VR-assisted rehabilitation techniques are based on flexion and extension of the hand and fingers [18, 24]. L. Piron et al. [25] formed a skill to put virtual envelopes into the mailbox in patients after a stroke followed by hemiparesis. These exercises improved fine motor skills and increased hand movement speed. Jack D. and Boian R. [26] first used VR system to restore upper limb function. They used a sensor glove (CyberGlove) connected to a personal computer with inbuilt sensors. These sensors recorded joint movements in all planes. Many researchers developed software for training only flexion of the hand and capture. However, Trombly C.A. et al. [27] noted that extension of the fingers clenched into a fist is extremely difficult for patients after stroke followed by spastic hemiparesis. This pattern and movement disorders are due to increased tone of flexor muscles and weakness of extensor muscles. Therefore, function of extensor muscles should be obligatory considered in development of rehabilitation programs. In addition, touching the subject is extremely essential besides movements of the hand in a virtual world per se, i.e. proprioceptive effect is important. Combination of VR with assistive technologies was proposed in one of the research centers in Chicago [24]. X. Luo et al. [28] successfully combined pneumatic orthosis with cybernetic glove for rehabilitation of patients with spastic hemiparesis. A helmet was put on the patient. Inbuilt display showed the image of the virtual environment. The objectives were to take and move the can with a drink, return the can to its original position and unclench the fingers. A pneumatic orthosis helped to unclench the fingers. S. Adamovich et al. [29]modernized sensory glove so that the patient has a sensation of a real object. It is still unclear whether the computer can provide a sufficient flow of proprioceptive information. M. Agostini et al. [5, 30]used a real object to obtain tactile feedback and proved that this method contributes to a better recovery of hand motor function. Contactless VR Rehabilitation Gaming System was studied for rehabilitation of patients after stroke in one of the centers of Canada. The authors found improved muscle strength in distal parts of the arm (hand) after training [31]. M. Iosa et al. [5, 32] used Leap Motion system with an infrared sensor for rehabilitation of patients after stroke followed by paresis of the upper limb. The authors noted increase of muscle strength in distal parts of the hand and improvement in fine motor skills. Current gloves had inbuilt joysticks, power sensors, etc. recording finger movements, so that a patient is able to control movements of the hand and fingers.

M. Cameirão et al. [31] were ones of the first who used mechanical trainer with integrated VR (Armeo Spring) to restore hand function. This complex provides anti-gravitational support for the hand, tactile feedback and bimanual training with limb weight unloading. This is achieved due to sensors inbuilt into exoskeletal structure and transmitting movements to the virtual hand. Multiple-component feedback is valuable to use mechanisms of retraining of motor control. The authors observed a long-lasting (over 3 months) significant improvement of hand muscle strength after a course of rehabilitation course. However, spastic disorders were the same.

S.V. Kotov et al. (2014), A.A. Frolov et al. (2016), Kovyazina M.S. et al. (2019) studied the complex "brain-computer interface + exoskeleton" in patients after stroke followed by upper limb paresis. The authors noted positive changes in neurological status including significant increase of the volume of movements and muscle strength in paretic limb, slight decrease of spastic disturbances and improved coordination. Accordingly, these changes increased degree of independence in everyday life. In a recent study, the authors propose a comprehensive approach to recovery of impaired motor function. This approach includes a preliminary psychological setup (priming) [33—35]. A. Vourvopoulos A. et al. [36] recommend the use of EEG-based “brain-computer interface” for accurate interpretation of brain signals. This approach would allow observing functional improvement after rehabilitation courses and affecting brain plasticity through visualization of cerebral electrical activity. However, F. Carrera Arias [37] reported that application of EEG-based “brain-computer interface” for video game can cause more significant frustration and negative mood in a patient compared with traditional game with a keyboard and mouse.

There are few studies devoted to application of VR in patients with focal brain lesion in acute and subacute periods of injury. C. Yin et al. [38] showed the outcomes of VR-assisted rehabilitation program in acute period of stroke for recovery of upper limb function. There were no significant differences between new technologies and traditional treatment. However, study demonstrates VR as an alternative method in acute period of stroke . W. Kim et al. [39] analyzed patients in subacute period of stroke. The effectiveness of VR system for recovery of upper limb function was similar to the efficacy of other methods. Nevertheless, VR system induced more hand movements compared with conventional therapy. The authors suggested this method as an adjunctive therapy in early rehabilitation after stroke. Currently, Q. Huang [40] et al. are studying the mechanism of hand motor function restoration and effectiveness of immersive VR-based rehabilitation in subacute period of stroke.

Focal brain lesion results impairment of walking and balance. Undoubtedly, walking function depends on coordinated activity of several analyzers. Information from the visual analyzer (optical flow) is perceived not only by the occipital lobe, but also parietal and temporal cortex. These data should coincide with proprioceptive and vestibular signals. If the optical flow does not correspond to proprioceptive data, regulation of walking is disrupted and the gait parameters are involuntarily changed. Accordingly, favorable results of walking function recovery may be achieved by using of indirect visualization mechanism, i.e. visual movement of the environment in accordance with movement of the body [41, 42]. A. Lamontagne et al. [43] revealed movement speed augmentation in adults after stroke followed by hemiparesis by almost 50% as soon as optical flow was changed. P.O. Koptelova et al. [44] reported similar outcomes in young and old people. The obtained data confirmed, summarized and supplemented the results of previous researches performed on treadmills. Changed visual information led to correction of the walking speed (length and frequency of the step). The technique may be used for motor regime enlargement [45—47].

H. Yano et al. [48] used a step-simulator synchronized with virtual space movement (Gait Master2) and noted a significant improvement of speed and symmetry of walking in patients with hemiparesis after previous stroke. S. Peurala et al. [49] demonstrated increased walking speed in patients with previous stroke after 3 weeks of training. The authors used an orthosis with electromechanical transmission (Lokomat) in conjunction with traditional rehabilitation methods [49]. M. Pohl et al. (2007), A. Mayr et al. (2007), L. Forrester et al. (2011) observed significant increase of walking speed if robotic methods were included into the rehabilitation program in early period of stroke [50—52]. V.F. Pyatin and A.V. Zakharov [53] reported increase of motor activity in acute period of stroke after VR-assisted training with imaginary visualization of legs. Maximum result was achieved for the first 3—5 courses and positive outcomes significantly correlated with duration of exercises. It is known that the main load falls on a healthy leg in patients after stroke with lower limb paresis. D. Kumar et al. [54] performed VR-based balance training using two Wii weight devices which measured the pressure of each leg. This system encouraged participants to use the lower limb with impaired motor function in tasks to shift the weight. Currently, Wii Fit Balance platform is one of the popular non-immersive VR-based systems for balance training in patients after stroke. Systematic review of C. Garcia-Munoz C. et al. [55] showed that Wii Fit Balance system is an affordable rehabilitation method and characterized by similar effectiveness compared with traditional methods of balance training in patients after stroke. D. Cano Porras et al. (2019) and P. Dominguez-Tellez et al. (2019) showed a significant effect of VR on gait and balance restoration in patients with previous stroke compared to traditional rehabilitation methods [56, 57]. A. Kim [58] reported that new motor skills acquired during VR-assisted exercises are easily transferred to real locomotion. ViRTAS trial devoted to VR-assisted gait training in subacute period of stroke is currently being performed [59].

To date, there is no single algorithm for evaluating functions before and after VR-based training. F. Felipe et al. (2019) in a systematic review (1836 articles) showed that the Berg balance scale, Fugl—Meyer assessment scale and Stroke Impact Scale are the most commonly used tools for assessing balance, neurological functions and quality of life, respectively, in patients with previous stroke undergoing VR-based rehabilitation [60]. These scales may be recommended for application in clinical practice.

Studies devoted to VR-based rehabilitation of patients with focal brain lesion become better every year and meet the requirements. Two years ago, the Cochrane systematic review showed that training with VR and interactive video games after stroke did not have significant advantages over traditional rehabilitation methods. These methods were recommended to increase the total training time (over 15 hours of exposure) [61]. Recent systematic review and meta-analysis of H. Lee [62] showed moderate effect of VR-based programs. This effect was equal for recovery of arm and leg function. VR-based programs were the most effective to achieve specific results – augmentation of muscle strength, volume of movements, improvement of balance, gait and daily activity [62].

P. Fishbein et al. [63] proved that multitasking improves rehabilitation outcomes. In the study, patients walked on a treadmill and simultaneously performed one or two tasks by the hands using the VR system. Better recovery was noted in those who performed two tasks. Thus, the authors propose not only to introduce VR into clinical practice, but also combine exercises for simultaneous performing of several tasks [63].

Virtual programs close to real life conditions of patient and improving adaptation to the environment are of particular interest. P. Weiss et al. (2004), D. Rand et al. (2007) reported the use of VIVID IREX platform for rehabilitation of patient after stroke followed by hemiparesis. Patient moved between rows with products in a virtual store, selected certain products, examined them, made a decision and formed a basket. It should be noted that training improved motor function and vertical stability [16, 64, 65]. P. Grewe P. et al. [66] updated virtual store, tested and demonstrated its effectiveness. In a new program (OctaVis 360° VR equipment), all goods looked like ordinary products of various brands, which can be found in German supermarkets. S. Subramanian et al. [67] showed improved accuracy and volume of hand movements in patients with hemiparesis after previous stroke who trained pressing a button with a finger in a virtual elevator.

A. Aramaki et al. (2019), S. Hajesmaeel Gohari et al. (2019) proved that inclusion of VR-based techniques into rehabilitation programs increases motivation for training [68, 69]. Therefore, the use of virtual environments in home rehabilitation of patients with long-term motor deficiency and the need for continuous training looks extremely interesting and perspective.

S. Nijenhuis et al. (2015), M. Sivan et al.(2014), F. Wittmann et al. (2016) analyzed patients' interest to restore impaired functions in outpatient conditions. Long-term medical rehabilitation program included robotic methods — The Activities of Daily Living Exercise Robot (ADLER), Computer Assisted Arm Rehabilitation (CAAR), ArmeoSenso, respectively. The authors found increased motivation for training, better restoration of muscle strength in the upper limb especially in its proximal parts [70—72]. S. Lee et al. (2018) proved that combination of VR and functional electrical stimulation improves upper limb function in patients with previous stroke [73]. M. Fu et al. (2019) developed and successfully conducted the research of a new method of home rehabilitation for patients with moderate or severe paresis of the upper limb. This method included a combination of functional contralateral electrical stimulation and video games [74]. N. Faric et al. [75] analyzed 498 reviews of 29 the most popular VR products, summarized data on the needs, expectations and preferences of players. The authors revealed that the game should be realistic, intuitive and ensure development of certain skill. Errors and low quality of the graphics led to disappointment. The main goal of VR-based rehabilitation is to engage in physical activity. Therefore, it is necessary to avoid negative factors and take into account patient's preferences in choosing a virtual game.

An important aspect affecting restoration of impaired functions through activation of neuroplasticity processes is the number of repetitions of the exercise. G. Kwakkel et al. [76] believe that overall duration of virtual therapy in rehabilitation program should be near 16 hours to achieve favorable clinical result. S. Nijenhuis et al. [70] applied VR-based training lasting 30 minutes per day for 6 days a week (up to 18 hours for six weeks). It was revealed that duration of training over 100 min per week was followed by augmentation of the muscle strength and sleight of the hand while variety of games increased motivation for further training. T. Platz et al. [77]recommend prolonged virtual therapy with higher intensity.

In accordance with the principles of evidence-based medicine, clinical results of VR-based rehabilitation should be supported by neuroimaging data. There are few researches devoted to analysis of the effect of VR using functional magnetic resonance imaging (fMRI) and EEG. S. Cramer et al. [78]analyzed activation of certain brain areas during VR-based rehabilitation measures in patients with previous stroke. The greatest activation was recorded near the damaged cerebral cortex.

M. Toyokura M. et al. [79] found that more difficult exercises were accompanied by activation of larger area of sensorimotor cortex. X. Bao, Y. Mao. et al. [80] explained motor restoration by functional reorganization of the cortex. A. Turolla A. et al. [81] suggested a combination of mechanized and virtual techniques in order to accelerate the processes of cortical reorganization and faster recovery of movements. Lee S.H., Kim Y.M. et al. [82]instrumentally confirmed the need for bilateral training. For a long time, the question was unclear whether it is necessary to see virtual limbs for achievement of better outcomes.

C. Modrono et al. [83] reported similar neuronal activity in both the presence and absence of virtual limbs. These findings give more freedom for development of new VR systems.

Interaction of drug therapy with VR-based methods is also of great interest besides introduction of robotic techniques into virtual reality per se. G. Samuel et al. [84] showed safety and better effectiveness of levodopa combined with virtual therapy for improvement of upper limb recovery. Combination of levodopa with virtual therapy significantly improved functional outcomes. Fugl—Meyer score was higher by 16.5 points in this group compared with the control group (levodopa intake alone). The ARAT scoring system gave similar results. Further large-scale studies with longer follow-up period are required to obtain more accurate data.

M. Islam and I. Brunner [85] analyzed the costs of including VR-based techniques into rehabilitation program for patients in subacute period of stroke to improve upper limb function (VIRTUES). Undoubtedly, VR-assisted rehabilitation is more expensive compared to traditional rehabilitation measures. However, preliminary analysis showed that these costs are balanced by reduced time of contact with a doctor and improved patient's motivation for further training [84].

However, general approaches, algorithms and protocols of neurorehabilitation are still absent. Exercises and game models are based on individual knowledge. Assessment of the effectiveness of motor rehabilitation is an extremely important aspect for development of rehabilitation programs and prediction of adequate rehabilitation. Introduction of robotic highly intelligent technologies into medical and diagnostic schemes seems to be an extremely interesting area of rehabilitation especially in patients with severe neurological deficiency.