Innovative Technologies in the Neurorehabilitation of Traumatic Brain Injury: A Systematic Review

Motor and cognitive rehabilitation in individuals with traumatic brain injury (TBI) is a growing field of clinical and research interest. In fact, novel rehabilitative approaches allow a very early verticalization and gait training through robotic devices and other innovative tools boosting neuroplasticity, thanks to the high-intensity, repetitive and task-oriented training. In the same way, cognitive rehabilitation is also evolving towards advanced interventions using virtual reality (VR), computer-based approaches, telerehabilitation and neuromodulation devices. This review aimed to systematically investigate the existing evidence concerning the role of innovative technologies in the motor and cognitive neurorehabilitation of TBI patients. We searched and reviewed the studies published in the Cochrane Library, PEDro, PubMed and Scopus between January 2012 and September 2022. After an accurate screening, only 29 papers were included in this review. This systematic review has demonstrated the beneficial role of innovative technologies when applied to cognitive rehabilitation in patients with TBI, while evidence of their effect on motor rehabilitation in this patient population is poor and still controversial.


Introduction
Traumatic Brain Injury (TBI) is non-progressive damage to the brain followed by a violent and rapid external force applied to the skull. TBI affects around 64-74 million persons each year worldwide and causes a variety of physical, motor, speech, and cognitive deficits that can have a long-term detrimental impact [1,2]. In fact, beyond motor impairment, attention, memory, affectivity, behaviour and executive dysfunctions may occur after the brain damage involving the frontal and temporal lobes, especially in the basal areas [3]. Diffuse axonal injury is often the cause of the worst functional outcomes [2]. In this context, both motor and cognitive neurorehabilitation for TBI patients is fundamental to its beneficial and effective role in improving patient outcomes and quality of life [4].
Physiotherapy treatments are focused on recovering and/or improving balance and gait ability, activating the locomotor centres of the central nervous system (CNS) and, at the same time, strengthening the postural control needed for deambulation [5]. However, conventional rehabilitation techniques have some limits that may undermine the outcomes, including the absence of a standardised training environment, adaptable supports to more functionally train the patients as well as the ability to increase therapy intensity and dose with a reduced physical burden for the therapists [6]. Indeed, for example, physiotherapists have difficulties in ensuring spatial and temporal symmetry between the steps for severely affected patients, making the repeatability of the exercises imprecise [7]. New approaches for verticalization and gait training have been employed to overcome these problems, which are based on high-intensity training with a high number of task-oriented repetitions. all combined with the expression "traumatic brain injury". Moreover, we have also analysed the references of the selected articles in order to obtain a complete search.

PICO Evaluation
We defined our combination of search terms using a PICO (population, intervention, comparison, outcome) model. The population was limited to moderate to severe TBI patients; intervention included all the innovative rehabilitative approaches, including robotic devices, virtual reality and computer-based training, telerehabilitation, neuromodulation; the comparison was evaluated considering the standard cognitive and motor rehabilitation techniques; and outcome included any motor and cognitive improvements shown by the patients, efficacy of treatment.

Inclusion and Exclusion Criteria
The inclusion criteria were (i) patients affected by moderate to severe TBI; (ii) randomised clinical trials (RCT), pilot studies and systematic reviews published between January 2012 and September 2022; (iii) English language; and (iv) published in a peerreviewed journal. Exclusion criteria were (i) case reports, case control and retrospective studies, and narrative reviews; (ii) studies with other kinds of electromedical devices; (iii) studies involving mild TBI, children and adolescents; (iv) other pathologies (i.e., vascular accidents, ischaemic and/or hemorrhagic).

Literature Selection
Two reviewers (MB and RDL) screened for relevance to the motor and cognitive effects of innovative rehabilitative approaches. Then, abstracts of the remaining articles were read, and those not meeting the eligibility criteria were excluded. The full text of all potential articles was evaluated in depth. In case of uncertainty, or when the abstract was not available, the entire article was read. If a disagreement was present, a third reviewer solved the concern (RSC). A PRISMA flow diagram was used for study selection (Figure 1). Since this was intended only as a comprehensive review, we did not perform a meta-analysis of the data. Of the 290 screened studies, 41 were selected, and eventually 26 were included in this review as they met the inclusion criteria ( Figure 1).

Study Quality Assessment
In addition, three authors rated the studies included in this review, using the Physiotherapy Evidence Database (PEDro) scoring system [23] to assess the methodological quality of studies. PEDro assesses 11 areas of study quality that are answered with a "yes" (score = 1) or "no" (score = 0). The first item is a measure of external validity and is not used in calculating the final score (i.e., sum of items 2-11).

Study Quality Assessment
In addition, three authors rated the studies included in this review, using the Physiotherapy Evidence Database (PEDro) scoring system [23] to assess the methodological quality of studies. PEDro assesses 11 areas of study quality that are answered with a "yes" (score = 1) or "no" (score = 0). The first item is a measure of external validity and is not used in calculating the final score (i.e., sum of items 2-11).

Results
Twenty-six articles were eventually analysed as they met our inclusion criteria. Given the different devices we dealt with, the results have been divided into six sections: Robotic and Virtual Systems for Motor Rehabilitation (n = 8); Humanoid Robots (n = 1); Virtual Reality Systems for Cognitive rehabilitation (n = 7); Computer-based rehabilitative approach (n = 2); Tele-rehabilitation (n = 3); Neuromodulation and combined approaches (n = 5).

Results
Twenty-six articles were eventually analysed as they met our inclusion criteria. Given the different devices we dealt with, the results have been divided into six sections: Robotic and Virtual Systems for Motor Rehabilitation (n = 8); Humanoid Robots (n = 1); Virtual Reality Systems for Cognitive rehabilitation (n = 7); Computer-based rehabilitative approach (n = 2); Tele-rehabilitation (n = 3); Neuromodulation and combined approaches (n = 5).
The PEDro analysis revealed that among the selected studies, four studies presented a score of 6, one study with a score of 7, two studies with a score of 5, while the others ranged between 4-2.
Most of the included articles, independently of the device used, presented a higher prevalence of cognitive than motor outcomes (Table 1). The authors presented that the administration of Wii balance board is more useful to informally assess or treat static balance alteration rather than for dynamic balance.  Telehealthcare system demonstrated to be not inferior to usual in-person care, to manage DoC due to TBI. In this study, the authors concluded that the use of tDCS can be an emerging tool for the treatment of the neuropsychological sequelae due to TBI, also in the subacute stage of recovery.

Robotic and Virtual Systems for Motor Rehabilitation
Among the 26 articles included in this review, we found only 2 articles [12,24] on the use of robotic verticalization with the Erigo device in sTBI patients. Specifically, De Luca et al. [12] administered autobiographical music stimulation plus robotic verticalization to the experimental group with a better improvement in the level of awareness and consciousness. Taveggia et al. [24] found that the use of the Erigo device may reduce orthostatic hypotension in vs. and MCS patients, thanks to its gradual verticalization and stepping, with constant and non-invasive monitoring of vitals parameters.
Three studies [25][26][27] investigated the use of RAGT by means of the Lokomat device, a robotic orthosis that guides the hip and the knee through a treadmill on which the patient's body weight may be supported/reduced. In detail, one study [25] found a 21%, 45% and 38% increase in patients' self-selected walking velocity when comparing the G-EO system, Lokomat and partial-body weight-supported treadmill training, respectively. In addition, the Lokomat and G-EO groups presented a more symmetrical gait pattern due to the advancement of the paretic limb. Williams et al. [26] suggested that RAGT may be a safe and feasible intervention that can be used as part of a physical therapy plan of care in adults with DoC due to TBI. Another study [27] compared the effects of the Lokomat with manually assisted treadmill training (MATT) in 16 TBI patients with limitations in ambulation. Beyond a significant improvement in step-length symmetry, the Lokomat required fewer staff and less manual effort with decreased staffing costs.
Only three studies [28][29][30] included the use of VR in the achievement of motor outcomes. One randomised controlled trial [28] used balance-based physical therapy by means of a Nintendo Wii in addition to their standard physical therapy regimen for improving static balance. The advanced training was personalised to the patient's specific needs and included different surfaces, bases of support, visual inputs, as well as dual-task and dynamic interactive activities. Postural treatment using VR (i.e., virtual games with Kinect Motion sensor) was investigated by a pilot study [29] which trained and improved limb coordination, posture and gait after 15 training sessions. Another study [30], comparing VR training using the Xbox Kinect games with a home-based exercise program, found that both groups improved in balance without any statistical differences.

Humanoid Robots
We found only one study [31] on the use of Humanoid robots, known as social robots, in the treatment of cognitive functions and social skills in TBI. The authors [31] showed the positive effects of the human robot Pepper on the improvement of cognitive and emotive processes, communication, and social skills compared to traditional cognitive treatment.

Virtual Reality Systems for Cognitive Rehabilitation
Five clinical studies [32,33,36,37] and two systematic reviews [34,35] on the cognitive effects of VR systems in TBI Neurorehabilitation were included. In particular, VR was used in two pilot studies [32,33] for the training of attention processes: a visuo-haptic virtual environment and a virtual touch modality by means of the Virtual Reality Rehabilitation System (VRRS) were used to improve cognitive function. The two systematic reviews investigated VR gaming for cognitive treatment of moderate TBI, demonstrating that VR tools may improve memory and executive function in patients with TBI, while for attention training, weak evidence still exists [34,35]. A pilot study [36] carried out in TBI military patients tested the effect of NeuroDRIVE on driving abilities, cognitive performance and neurobehavioral symptoms, and demonstrated statistically-significant improvements in working memory and visual search/selective attention. Another pilot study [37] considered a virtual supermarket environment, namely the VMall, to perform occupational therapy compared to non-VR occupational therapy cognitive retraining. The authors suggested that cognitive treatment that focuses on mediating strategies can improve executive functions and also IADL independently of the use of VR environments. Finally, the BTs-Nirvana non-immersive VR system was used in a clinical trial with one hundred TBI patients [38], showing a great improvement in specific cognitive domains, such as cognitive flexibility, attentional shifting, visual search, and executive and visuospatial functions that are necessary for planning and managing daily life.

Computer-Based Rehabilitative Approach
We found only two papers [39,40] about the use of computer-based interventions to improve cognitive functioning in subjects following TBI. Lebowitz and colleagues [39] evaluated the feasibility and utility of computer-based mental exercises for visual processing and memory, i.e., the Cortex with InSight (Posit Science Corporation, San Francisco, California), for individuals with a history of TBI at home. This advanced training optimised the main cognitive functions, including attention, memory and information processing. Zickefoose et al. [40] examined the potential effect of two computerised brain game software programs in participants with severe TBI with a mean time post-injury of 4 years. Two participants received either four weeks of computer-based Attention Process Training (APT-3) first or Lumosity brain games first and were observed for 30 min per day, five days per week. Then, they switched to the other intervention for another four weeks. All participants significantly improved their level of difficulty on intervention tasks, and there was a trend toward generalisation to daily tasks.

Tele-Rehabilitation
We found one systematic review [41] and two studies [42,43] on the use of information and communication technologies, including telephone, messaging, e-mail, and video conferencing systems, in the rehabilitation of TBI. Ownsworth et al. [41] evaluated the efficacy of telerehabilitation through telephone-based (10 studies) and Internet-based (3 studies) interventions for adults with TBI, showing that the structured telephone interventions were effective in improving mood symptoms, functional status, emotional well-being, QoL and sleep quality. De Luca et al. [42], in their feasibility/usability study in severe TBI patients, found that the Virtual Reality Rehabilitation System (VRRS) had a beneficial effect on the patient's motivation during the training, encouraging the return home and the continuity of care in the territory. Raso et al. [43] confirmed the potential use of telehealth for the management of DoC due to TBI through an advanced video conferencing telehealth system at home.

Neuromodulation and Combined Approaches
In one pilot study, the authors [44] suggested a potential role of neuromodulation, when applied to the left dorsolateral prefrontal cortex, in improving attention during cognitive training after TBI. The systematic review [45] involved five articles, two about repetitive transcranial magnetic stimulation (rTMS) and three on transcranial direct current stimulation (tDCS), suggesting that neural plasticity changes induced by Non-invasive brain stimulation (NIBS) may contribute to greater improvements when combined with rehabilitation. Another study [46] used NIBS to improve divided attention and to boost neural plasticity with a normalisation of the abnormal hyperactivations of the temporal, middle frontal and postcentral gyrus. These results were confirmed and implemented by Ulam et al. [47], suggesting that 10 anodal tDCS sessions may positively modulate cortical excitability in patients with TBI. On the contrary, Leśniak et al. provided insufficient evidence to support the efficacy of repeated a-tDCS for enhancing the rehabilitation of memory and attention in patients after severe TBI [48].

Discussion
To the best of our knowledge, this is the first systematic review that investigates the use of robotics and other technologies in moderate to severe TBI rehabilitation. Indeed, most existing reviews are focused only/mainly on VR approaches and cognitive telerehabilitation [33,34,40] and do not offer an overview of the most innovative approaches for both motor and cognitive rehabilitation in this patient population.
In Table 2, we have summarised the main innovative neurorehabilitation tools that may positively affect outcomes in TBI, providing information on their structure, function and use in the different stages/degrees of the disease.
Recently, the interest in the development of rehabilitative robotic systems has grown strongly. In fact, traditional approaches can improve functional recovery but require a considerable commitment of human resources, and it can be very challenging in patients with severe disabilities. Therefore, robotic devices represent valid support, as they allow repetitive, intensive and more lasting therapy, also guaranteeing a more objective assessment of the patients' damage and outcomes. In particular, robotic motor systems allow a very early verticalization, as well as walking training, once the patients' vital signs are stable. In fact, some devices, like the Lokomat, are capable of assisting a passive gait, so they can also be used in sTBI [49]. Afterwards, the gradual recovery of muscle strength will allow the therapist to reduce the BWS with the robotic device [50]. In fact, these robotic motor devices guarantee better postural and trunk control, decreasing muscle stiffness and acting on the improvement of cardio-circulatory and respiratory functions [51]. Noteworthy, available Clinical Practice Guidelines to improve locomotor function following brain injury [52] indicate that clinicians should offer walking training at moderate to high intensities or virtual reality-based training to ambulatory individuals greater than six months following the brain injury acute-onset to improve walking speed or distance. Indeed, Esquenazi et al. [25] showed that an intensive RAGT with G-EO system or Lokomat in chronic TBI patients can improve velocity, endurance and symmetry and lengths of steps. Otherwise, weak evidence suggests that strength training, circuit (i.e., combined) training or cycling training at moderate to high intensities, and virtual reality-based balance training may improve walking speed and distance in these patient groups. However, the optimal therapeutic intervention to achieve full gait recovery remains unknown, but it is evident that any rehabilitative effort to guide motor recovery should encounter fundamentals of neuroplasticity and be related to clinical prognostic indicators (i.e., disease severity, vital signs, intracranial pressure, cerebral perfusion pressure) [53,54]. Table 2. Description of the main innovative neurorehabilitation tool used in TBI patients.

Robotic Device Short Description Picture
Erigo [12,24] Erigo is a robotic tilt To the best of our knowledge, this is the first systematic review that investigates the use of robotics and other technologies in moderate to severe TBI rehabilitation. Indeed, most existing reviews are focused only/mainly on VR approaches and cognitive telerehabilitation [33,34,40] and do not offer an overview of the most innovative approaches for both motor and cognitive rehabilitation in this patient population.
In Table 2, we have summarised the main innovative neurorehabilitation tools that may positively affect outcomes in TBI, providing information on their structure, function and use in the different stages/degrees of the disease.
Recently, the interest in the development of rehabilitative robotic systems has grown strongly. In fact, traditional approaches can improve functional recovery but require a considerable commitment of human resources, and it can be very challenging in patients with severe disabilities. Therefore, robotic devices represent valid support, as they allow repetitive, intensive and more lasting therapy, also guaranteeing a more objective assessment of the patients' damage and outcomes. In particular, robotic motor systems allow a very early verticalization, as well as walking training, once the patients' vital signs are stable. In fact, some devices, like the Lokomat, are capable of assisting a passive gait, so they can also be used in sTBI [49]. Afterwards, the gradual recovery of muscle strength will allow the therapist to reduce the BWS with the robotic device [50]. In fact, these robotic motor devices guarantee better postural and trunk control, decreasing muscle stiffness and acting on the improvement of cardio-circulatory and respiratory functions [51]. Noteworthy, available Clinical Practice Guidelines to improve locomotor function following brain injury [52] indicate that clinicians should offer walking training at moderate to high intensities or virtual reality-based training to ambulatory individuals greater than six months following the brain injury acute-onset to improve walking speed or distance. Indeed, Esquenazi et al. [25] showed that an intensive RAGT with G-EO system or Lokomat in chronic TBI patients can improve velocity, endurance and symmetry and lengths of steps. Otherwise, weak evidence suggests that strength training, circuit (i.e., combined) training or cycling training at moderate to high intensities, and virtual reality-based balance training may improve walking speed and distance in these patient groups. However, the optimal therapeutic intervention to achieve full gait recovery remains unknown, but it is evident that any rehabilitative effort to guide motor recovery should encounter fundamentals of neuroplasticity and be related to clinical prognostic indicators (i.e., disease severity, vital signs, intracranial pressure, cerebral perfusion pressure) [53,54]. Table 2. Description of the main innovative neurorehabilitation tool used in TBI patients.

Robotic Device Short Description Picture
Erigo [12,24] Erigo is a robotic tilt table used in the early stage of recovery after an acute TBI. It allows an early and gradual robotic verticalization combined with a cyclic leg movement in order to stimulate CNS through critical afferent stimuli. The tilt table inclination can be regulated by the therapists, who can control the gradualness of verticalization (i.e., from 45° to 90°) as well as the stepping speed, according to the patients' needs. In addition, the Erigo device also can improve cardiocirculatory stability through muscle activation, pump function and venous return.

Robotic Device Short Description Picture
Lokomat [25][26][27] The Lokomat is a robotic gait-assisted device widely used in the neurorehabilitation of walking. It consists of an external gait orthosis integrated with a computer-controlled linear actuators at each hip and knee joint, in addition to an advanced system of body weight support system and a treadmill. The therapists can control the level of gait assisted-support, the force and the pattern of gait movement. Lokomat [25][26][27] The Lokomat is a robotic gait-assisted device widely used in the neurorehabilitation of walking. It consists of an external gait orthosis integrated with a computer-controlled linear actuators at each hip and knee joint, in addition to an advanced system of body weight support system and a treadmill. The therapists can control the level of gait assisted-support, the force and the pattern of gait movement.

GEO-O [25]
GEO SYSTEM is a robotic end effector system that simulates the repetitive training of locomotion in everyday life, such as walking on the flat, ascending and descending stairs. The patient's feet are secured to the platforms moving in all directions (i.e., upwards, downwards, forwards and backwards) with the assistance of six engines. There is also an Evolution version that also includes the use of an immersive scenario to fully involve the person in rehabilitation therapy in dual-task activities.

Human-Robot Short Description Picture
Robot-PEPPER [31] Pepper is one of the most widely known social humanoid robots, which is used to recognize faces and basic human emotions. This humanoid robot was conceptualised for human interaction, thanks to the conversation and his colourful touchscreen.

GEO-O [25]
GEO SYSTEM is a robotic end effector system that simulates the repetitive training of locomotion in everyday life, such as walking on the flat, ascending and descending stairs. The patient's feet are secured to the platforms moving in all directions (i.e., upwards, downwards, forwards and backwards) with the assistance of six engines. There is also an Evolution version that also includes the use of an immersive scenario to fully involve the person in rehabilitation therapy in dual-task activities. Lokomat [25][26][27] The Lokomat is a robotic gait-assisted device widely used in the neurorehabilitation of walking. It consists of an external gait orthosis integrated with a computer-controlled linear actuators at each hip and knee joint, in addition to an advanced system of body weight support system and a treadmill. The therapists can control the level of gait assisted-support, the force and the pattern of gait movement.

GEO-O [25]
GEO SYSTEM is a robotic end effector system that simulates the repetitive training of locomotion in everyday life, such as walking on the flat, ascending and descending stairs. The patient's feet are secured to the platforms moving in all directions (i.e., upwards, downwards, forwards and backwards) with the assistance of six engines. There is also an Evolution version that also includes the use of an immersive scenario to fully involve the person in rehabilitation therapy in dual-task activities.

Human-Robot Short Description Picture
Robot-PEPPER [31] Pepper is one of the most widely known social humanoid robots, which is used to recognize faces and basic human emotions. This humanoid robot was conceptualised for human interaction, thanks to the conversation and his colourful touchscreen.

Human-Robot Short Description Picture
Robot-PEPPER [31] Pepper is one of the most widely known social humanoid robots, which is used to recognize faces and basic human emotions. This humanoid robot was conceptualised for human interaction, thanks to the conversation and his colourful touchscreen. Lokomat [25][26][27] The Lokomat is a robotic gait-assisted device widely used in the neurorehabilitation of walking. It consists of an external gait orthosis integrated with a computer-controlled linear actuators at each hip and knee joint, in addition to an advanced system of body weight support system and a treadmill. The therapists can control the level of gait assisted-support, the force and the pattern of gait movement.

GEO-O [25]
GEO SYSTEM is a robotic end effector system that simulates the repetitive training of locomotion in everyday life, such as walking on the flat, ascending and descending stairs. The patient's feet are secured to the platforms moving in all directions (i.e., upwards, downwards, forwards and backwards) with the assistance of six engines. There is also an Evolution version that also includes the use of an immersive scenario to fully involve the person in rehabilitation therapy in dual-task activities.

Human-Robot Short Description Picture
Robot-PEPPER [31] Pepper is one of the most widely known social humanoid robots, which is used to recognize faces and basic human emotions. This humanoid robot was conceptualised for human interaction, thanks to the conversation and his colourful touchscreen.

Human-Robot Short Description Picture
Robot-NAO [55] Nao is another humanoid robot characterised by its small size, and it is able to interact with adult, adolescent and paediatric patients. It is equipped with sensors that allow it to walk, dance, speak, and recognize faces and objects. It can be used to provide social activities and to recognize emotions, giving sensorial feedback to patients. Robot-NAO [55] Nao is another humanoid robot characterised by its small size, and it is able to interact with adult, adolescent and paediatric patients. It is equipped with sensors that allow it to walk, dance, speak, and recognize faces and objects. It can be used to provide social activities and to recognize emotions, giving sensorial feedback to patients.

Virtual Reality System Short Description Picture
BTS Nirvana [38] BTS NIRVANA is an innovative therapeutic system that assists the rehabilitation process of patients affected by neurological diseases, thanks to its multi-sensorial stimulation. The patients can move or manipulate specific objects in different ways (i.e., balls, flowers, and butterflies) or create specific combinations (i.e., colour-number) with a dynamic involvement in the virtual environment. During the interaction between the patient and the screen, the system produces audio and video feedback (using the sprite activity). The difficulty of exercises increases the base of the number of distractors and reduces the time available for the execution.

VRRS Virtual Reality
Rehabilitation System [33] Il Khymeia VRRS-Virtual Reality Rehabilitation System-is the most widely used for VR training and teletraining in clinical practice. The VRRS, in fact, is conceptualised with a "central HUB" that can be connected via USB, a series of specialised peripherals fully synchronised and integrated with the system. The VRRS is equipped with exercise modules for cognitive, language, postural, and motor rehabilitation. These virtual exercises can be selected and included in the rehab program by the therapist, who can shape the difficulty in relation to the time of execution and the type of activity.

Virtual Reality System
Short Description Picture

BTS Nirvana [38]
BTS NIRVANA is an innovative therapeutic system that assists the rehabilitation process of patients affected by neurological diseases, thanks to its multi-sensorial stimulation. The patients can move or manipulate specific objects in different ways (i.e., balls, flowers, and butterflies) or create specific combinations (i.e., colour-number) with a dynamic involvement in the virtual environment. During the interaction between the patient and the screen, the system produces audio and video feedback (using the sprite activity). The difficulty of exercises increases the base of the number of distractors and reduces the time available for the execution. Robot-NAO [55] Nao is another humanoid robot characterised by its small size, and it is able to interact with adult, adolescent and paediatric patients. It is equipped with sensors that allow it to walk, dance, speak, and recognize faces and objects. It can be used to provide social activities and to recognize emotions, giving sensorial feedback to patients.

Virtual Reality System Short Description Picture
BTS Nirvana [38] BTS NIRVANA is an innovative therapeutic system that assists the rehabilitation process of patients affected by neurological diseases, thanks to its multi-sensorial stimulation. The patients can move or manipulate specific objects in different ways (i.e., balls, flowers, and butterflies) or create specific combinations (i.e., colour-number) with a dynamic involvement in the virtual environment. During the interaction between the patient and the screen, the system produces audio and video feedback (using the sprite activity). The difficulty of exercises increases the base of the number of distractors and reduces the time available for the execution.

VRRS Virtual Reality
Rehabilitation System [33] Il Khymeia VRRS-Virtual Reality Rehabilitation System-is the most widely used for VR training and teletraining in clinical practice. The VRRS, in fact, is conceptualised with a "central HUB" that can be connected via USB, a series of specialised peripherals fully synchronised and integrated with the system. The VRRS is equipped with exercise modules for cognitive, language, postural, and motor rehabilitation. These virtual exercises can be selected and included in the rehab program by the therapist, who can shape the difficulty in relation to the time of execution and the type of activity.

VRRS Virtual Reality
Rehabilitation System [33] Il Khymeia VRRS-Virtual Reality Rehabilitation System-is the most widely used for VR training and teletraining in clinical practice. The VRRS, in fact, is conceptualised with a "central HUB" that can be connected via USB, a series of specialised peripherals fully synchronised and integrated with the system. The VRRS is equipped with exercise modules for cognitive, language, postural, and motor rehabilitation. These virtual exercises can be selected and included in the rehab program by the therapist, who can shape the difficulty in relation to the time of execution and the type of activity. Robot-NAO [55] Nao is another humanoid robot characterised by its small size, and it is able to interact with adult, adolescent and paediatric patients. It is equipped with sensors that allow it to walk, dance, speak, and recognize faces and objects. It can be used to provide social activities and to recognize emotions, giving sensorial feedback to patients.

Virtual Reality System Short Description Picture
BTS Nirvana [38] BTS NIRVANA is an innovative therapeutic system that assists the rehabilitation process of patients affected by neurological diseases, thanks to its multi-sensorial stimulation. The patients can move or manipulate specific objects in different ways (i.e., balls, flowers, and butterflies) or create specific combinations (i.e., colour-number) with a dynamic involvement in the virtual environment. During the interaction between the patient and the screen, the system produces audio and video feedback (using the sprite activity). The difficulty of exercises increases the base of the number of distractors and reduces the time available for the execution.
VRRS Virtual Reality Rehabilitation System [33] Il Khymeia VRRS-Virtual Reality Rehabilitation System-is the most widely used for VR training and teletraining in clinical practice. The VRRS, in fact, is conceptualised with a "central HUB" that can be connected via USB, a series of specialised peripherals fully synchronised and integrated with the system. The VRRS is equipped with exercise modules for cognitive, language, postural, and motor rehabilitation. These virtual exercises can be selected and included in the rehab program by the therapist, who can shape the difficulty in relation to the time of execution and the type of activity. Neuro-DRIVE System [36] The Neuro-DRIVE system uses a virtual reality driving simulator. It consists of a curved screen in addition to a driving console similar to a typical automobile. Each driving console presents turn signals, gas and brake pedals, a steering wheel, a digital dashboard, and a seat belt. The patient is seated in front of the screen, holding the steering wheel and pushing the pedals while interacting with the virtual environment stimuli.
People with TBI are affected by cognitive-behavioural problems, and many devices have been developed and used in research and clinical practice to deal with these problems. Concerning human-like robots, also called Socially Assistive Robots (SARs), such as PEPPER and Nao, primarily assist people in social interactions (e.g., speaking, driving, remembering, observing, and entertaining). In TBI patients, SARs [31] enhance mood, social relationships among patients, and emotional expression of brain injury survivors. Assad-Uz-Zaman [55] implemented the software-library exercise programs of Nao robots to be administered in individuals with upper limb impairment and also in the TBI population.
VR systems, indeed, are considered one of the most innovative technologies to potentiate both motor and cognitive rehabilitation. In fact, it is evident that by using this tool, several very intense types of exercise can be adapted to achieve specific outcomes for TBI patients. In fact, VR can be used to provide the patient with repetitive, task-oriented training due to the multisensorial feedback, thus potentiating the use-dependent plasticity of the sensory-motor cortex [56][57][58]. Patients interacting with a virtual environment could be trained in a more ecological way and can perform all those activities that are dangerous to be performed in real life. VR, by providing individuals with "knowledge of results" and "knowledge of performance" may favour reinforcement learning and then functional recovery. Nonetheless, our systematic analysis showed poor findings supporting VR for balance training using Nintendo Wii and Xbox Kinect sensors [29,30], and the VR training was no more effective than the standard treatment. Some specific immersive VR devices, such as the Computer-Assisted Rehabilitation Environment (CAREN), may lead to better results. The tool is able to provide the patient with a sense of presence, acting through both bottom-up (i.e., the neurosensorial feedback given by the treadmill and the six degrees of freedom balance) and top-down (VR acoustic and visual) modalities. Then, it has been successfully used to improve motor and cognitive abilities, vestibular dysfunction, as well as mood and anxiety disorders, especially in military medicine [59][60][61].
Computer-based treatment (C-BT) has been used for cognitive training in patients with moderate to severe TBI, and some authors have demonstrated that enjoyment and active involvement may lead to better results [40]. A critical review of the literature suggested that C-BT seems promising as an approach to improve cognition, with regard to working memory, in TBI patients [62]. Notably, it has been hypothesised that attention improvements with this tool are related to a combination of factors, such as young age, years since the trauma, closed head injury, higher pre-intervention scores on the Test of Everyday Attention, and this should be taken into account when dealing with C-BT.
Another growing issue concerns the use of telemedicine and ehealth in the rehabilitation field [63,64], although, in our review, we found only a few articles dealing with TR in the TBI population. Despite the poor available data, it is suggested that TR could be of help by overcoming the displacement of therapists or patients and reduction of patient hospitalisation times and costs for both patients and healthcare providers [65].
People with TBI are affected by cognitive-behavioural problems, and many devices have been developed and used in research and clinical practice to deal with these problems. Concerning human-like robots, also called Socially Assistive Robots (SARs), such as PEPPER and Nao, primarily assist people in social interactions (e.g., speaking, driving, remembering, observing, and entertaining). In TBI patients, SARs [31] enhance mood, social relationships among patients, and emotional expression of brain injury survivors. Assad-Uz-Zaman [55] implemented the software-library exercise programs of Nao robots to be administered in individuals with upper limb impairment and also in the TBI population.
VR systems, indeed, are considered one of the most innovative technologies to potentiate both motor and cognitive rehabilitation. In fact, it is evident that by using this tool, several very intense types of exercise can be adapted to achieve specific outcomes for TBI patients. In fact, VR can be used to provide the patient with repetitive, task-oriented training due to the multisensorial feedback, thus potentiating the use-dependent plasticity of the sensory-motor cortex [56][57][58]. Patients interacting with a virtual environment could be trained in a more ecological way and can perform all those activities that are dangerous to be performed in real life. VR, by providing individuals with "knowledge of results" and "knowledge of performance" may favour reinforcement learning and then functional recovery. Nonetheless, our systematic analysis showed poor findings supporting VR for balance training using Nintendo Wii and Xbox Kinect sensors [29,30], and the VR training was no more effective than the standard treatment. Some specific immersive VR devices, such as the Computer-Assisted Rehabilitation Environment (CAREN), may lead to better results. The tool is able to provide the patient with a sense of presence, acting through both bottom-up (i.e., the neurosensorial feedback given by the treadmill and the six degrees of freedom balance) and top-down (VR acoustic and visual) modalities. Then, it has been successfully used to improve motor and cognitive abilities, vestibular dysfunction, as well as mood and anxiety disorders, especially in military medicine [59][60][61].
Computer-based treatment (C-BT) has been used for cognitive training in patients with moderate to severe TBI, and some authors have demonstrated that enjoyment and active involvement may lead to better results [40]. A critical review of the literature suggested that C-BT seems promising as an approach to improve cognition, with regard to working memory, in TBI patients [62]. Notably, it has been hypothesised that attention improvements with this tool are related to a combination of factors, such as young age, years since the trauma, closed head injury, higher pre-intervention scores on the Test of Everyday Attention, and this should be taken into account when dealing with C-BT.
Another growing issue concerns the use of telemedicine and ehealth in the rehabilitation field [63,64], although, in our review, we found only a few articles dealing with TR in the TBI population. Despite the poor available data, it is suggested that TR could be of help by overcoming the displacement of therapists or patients and reduction of patient hospitalisation times and costs for both patients and healthcare providers [65]. In this way, telehealth systems allow cost-effectiveness avoiding feelings of abandonment and stress from both patients and caregivers [66].
Finally, NIBS is demonstrating a positive effect in promoting neuroplasticity and functional recovery following a brain injury [45], thus further potentiating cognitive rehabilitation effects [44]. Better and long-lasting results are achieved when neuromodulation is applied in combination with other methodologies, including computer-assisted training [46]. Moreover, among the potential sites of stimulation, the dorsolateral prefrontal cortex (DLPFC) seems to be preferential for its association with attention and working memory function [46,48], although it is under debate which kind of current should be used (cathodal vs. anodal) and the time/frequency of the stimulation protocol. However, it seems that the combined application of anodal and cathodal current on the right DLPFC produces improvement in cognitive performance [67].
To summarise, since various innovative neurorehabilitation approaches exist, clinicians, as well as researchers, do not always agree on which technology is better to train TBI patients. The approach's choice should take into account the site and the extent of brain damage, the age and the severity of the clinical status (i.e., a disorder of consciousness, severe vs. moderate motor and cognitive disability) [68]. In addition, patients affected by TBI (or at least their caregivers) should be familiar with technical issues to not yield the training more challenging [42], especially when they have to use virtual reality, computer-based approaches and telerehabilitation. In detail: (1) Patients with a diagnosis of DoC or sTBI need to achieve an early and gradual verticalization in order to avoid deterioration of the autonomic nervous system and bedridden complications [12,24]. For these reasons, the Erigo device could be a useful tool in clinical practice to meet patient necessities. In addition, sTBI could benefit from RAGT through Lokomat, which can assist a passive gait-increasing BWS, always monitoring vital parameters to guarantee a safe and feasible rehabilitative intervention [26]. (2) Patients affected by moderate TBI may gain balance and coordination thanks to VR exercises, which are also known to promote the enjoyment and active involvement of the patients [28][29][30], even if the evidence found is not sufficient to support its systematic use in TBI clinical practice. Instead, the GE-O system seems to be a valid tool for gait training in moderate TBI patients to improve endurance and walking speed [25]. (3) On the other hand, the C-BT, as well as VR systems, are widely used in severe and moderate TBI patients to train cognitive functions. VRRS allows a specific subdomain training and, thanks to its big screen, provides a wider view of space and facilitates the execution of cognitive tasks [33,38], whereas moderate TBI could take advantage of C-BT, which requires more control in the upper limb [39,40]. (4) Another useful innovative approach is TR; despite the poor evidence in the TBI population, it can be useful in ensuring continuity of care between discharge from the hospital and return to home [42], avoiding travelling costs and geographical barriers for patients living far from metropolitan areas [65,66]. Specifically, the VRRS for TR is a feasible tool for moderate TBI patients that allows therapists to plan a specific training session (i.e., motor, with regard to upper limb, and cognitive) remotely supervised by the Tele-CockPit workstation [42]. (5) Last, NIBS in the moderate to severe TBI population is a promising approach, especially used to stimulate cognitive functions [67]. However, it is still an emerging approach and there is not enough data to confirm its use in clinical practice. Possible future directions could investigate the efficacy of NIBS combined with other technologies such as TR, C-BT and VR.
This systematic review has some limitations to acknowledge, including a lack of bias analysis, the wide variability of available technologies and an absent meta-analysis of the data. The selected studies also presented limitations, including small sample sizes, lack of control group and lack of long-term follow-up evaluations. We have instead comprehensively considered all the technologies and their effects used in the motor and cognitive neurorehabilitation of moderate to severe TBI patients.

Conclusions
This review provides clinically useful evidence on the safety and potential efficacy of technologies in TBI in an attempt to bridge the gap between research and clinical practice in this frail patient population. Our results revealed that the use of innovative technologies in TBI neurorehabilitation provided improvements in motor and, above all, cognitive outcomes. Although evidence on the effectiveness of advanced motor rehabilitation is still poor, it seems to be promising, at least in some sTBI. For these reasons, further and larger studies are needed to identify whether and to what extent patients with sTBI may benefit from robotic devices, VR and neuromodulation.