Immersive VR for upper-extremity rehabilitation in patients with neurological disorders: a scoping review

This scoping review was performed in accordance with the Arksey and O’Malley scoping review methodology [19], which involves five mandatory steps, namely: (1) defining the research question; (2) identifying relevant studies; (3) study selection; (4) data extraction; and (5) analyzing the data, summarizing and reporting the results. An optional sixth step involves discussions among the raters. The available literature was summarized using the Preferred Reporting Items for Systematic Reviews and Meta-Analyses for Scoping Reviews (PRISMA-ScR) guidelines [20].

We based the study questions on the Sample, Phenomenon of Interest, Design, Evaluation, Research type (SPIDER) approach, commonly used for qualitative, quantitative and mixed methods reviews [21]:

Based on the objective of this scoping review, we identified the following research question: “What is known about the use of immersive VR for upper-extremity rehabilitation in patients with neurological diseases?”.

Based on the identified research question, we included the articles that met the following eligibility criteria:

We excluded the articles that met the following exclusion criteria:

To answer the research question, an additional inclusion criterion was applied to the full text of the eligible papers:

The clinical outcome metrics were considered valid if they were among the standard measures for clinical assessment of upper-limb function. Examples include the Fugl-Meyer Assessment-Upper Extremity (FMA-UE) [22], Modified Ashworth Scale (MAS) [23], Test of Upper Limb Apraxia (TULIA) [24], Box and Blocks Test (BBT) [25], and Action Research Arm Test (ARAT) [26]. The kinematic outcome metrics were considered valid if they provided an objective assessment of upper-limb function. Examples include the Range of Motion (ROM) and kinetic measures such as force and torque.

To identify potentially relevant studies to be included in this scoping review, a systematic search—from inception to end of December 2023—was conducted on the following bibliographic databases: Scopus, PubMed, and IEEE Xplore.

Based on the identified research question, the systematic search was conducted using specific key terms, in Medical Subject Headings (MeSH) where applicable, to search through titles, abstracts, and keywords. The full search strategy is described in detail in Additional file 1: Table S1. The selected key terms were Head-Mounted Display, Virtual Reality, Upper limb, and Rehabilitation.

Data extracted from the articles obtained through the search process described above were downloaded and organized using Microsoft Excel. Duplicates were removed. All titles and abstracts were first independently screened for relevance by MC and EL using the inclusion and exclusion criteria described above. No instances of conflicting opinions came up during the screening phase that necessitated the involvement of another author for resolution. Relevant full-text papers were then further screened for final inclusion by MC and EL using the full-text inclusion criterion described above. From each article, one reviewer (M.C.) extracted the data pertaining to the information needed to answer the research question. The extracted data concerned the general characteristics of the study (i.e., participants’ neurological condition, time since injury, sample size, age of the population, participants’ previous experience with VR, aims of the work, study type, and study design), experimental protocol (i.e., type of task, session duration, participants pre-training, experiment duration, recorded physiological signals, and outcome measures), virtual environment (i.e., HMD device, control modality, custom-built or commercial virtual environment, game engine, and explicit use of the audio), and assessment of cybersickness (i.e., questionnaires and other types of assessment).

One reviewer (MC) conducted a thematic analysis to identify the factors acting as barriers and facilitators in the adoption of immersive VR for rehabilitation according to the authors of each study. This analysis relied on the primary results and discussions chapters of each article, aiming to extract and retrieve the key points emphasized by the respective authors. Initially, key points representing the obstacles, challenges, and limitations were identified and subsequently condensed into one or two representative words. Subsequently, a statistic for each keyword was built by counting their frequency of occurrence across all the papers. Lastly, to facilitate discussion and categorization, we organized these terms into three clusters: technology, training and usability.

The search yielded 309 articles from Scopus (n = 181), PubMed (n = 117), and IEEE Xplore (n = 11). After removing 96 duplicates, a total of 213 articles were screened by title and abstract and 195 were subsequently excluded, showing an almost perfect level of agreement between two authors M.C. and E.L. (Cohen’s Kappa coefficient = 0.914 [27]). The full text of 18 articles was evaluated for eligibility and 5 articles were eventually excluded because they did not meet the inclusion criteria for full text described in the Methods section (Cohen’s Kappa = 1.0 between MC and EL). Thus, the final synthesis included a total of 13 records (Fig. 1).

A comprehensive summary of key information extracted from the studies, including pathology, time since injury, sample size, population age, participants’ prior experience with VR, research aims, and study type and design is provided in Table 1. The analysis of these results reveals a notable trend: the use of immersive VR for upper-limb rehabilitation has emerged relatively recently, with the oldest paper included in this review published in 2016.

Among the studies in acute and subacute stroke patients, rehabilitation tasks are quite balanced (Fig. 2). Specifically, mirror therapy was used in one study [30], and motor tasks in two studies [31, 37]. Conversely, the studies in chronic patients mainly used motor tasks (n = 3) [32‐34] and less, mirror therapy (n = 1) [35] and motor imagery (n = 1) [36].

In terms of homogeneity regarding clinical and demographic characteristics within subgroups, the three studies within the acute plus subacute stroke group focusing on motor tasks exhibit similar age ranges (Table 1): Heinrich et al. [30] reported participants’ ages ranging from 51 to 75 with a mean of 62 years, the single subject in the case study by Park et al. [31] was 56 years old, and Huang et al. 2018 [37] reported participants’ ages ranging from 44 to 79 years with a mean age of 69 years. The time since injury reported was similar for two studies: Heinrich et al. [30] (mean: 2.04 months; std: 1.87), and Huang et al. [37] (mean: 2.19 months; std: 1.13), while it cannot be compared as it was not reported by Park et al. [31]. The three studies on chronic stroke patients in the motor task subgroup show comparable clinical profiles in terms of time since injury: Erhardsson et al. [32] (mean: 2.51 years; std: 1.96), Marin Pardo et al. [33] (mean: 3.17 years; std: 1.03), and Lee et al. 2020 [34] (range: 3 years; std: 5.16). Furthermore, the age profiles are consistent between Erhardsson et al. [32] (range: 48–74 years; mean: 60.6) and Marin Pardo et al. [33] (range: 42–66 years; mean: 56.3). Only Lee et al. [34] have a population with a wider age range, starting from a lower minimum (range: 19–70 years; mean: 40.2).

As regards, instead, the comparison between studies in acute plus sub-acute and chronic stroke patients, there are some differences including the duration of a single session and the total number of sessions conducted (Table 2). For studies with participants in the sub-acute post-stroke stage [30, 31, 37], who are more sensitive to the intensity and duration of exercises because of their fragile condition, the average session duration was 13 ± 6 (mean ± std) minutes, and the average number of sessions was 14 ± 6. In contrast, in the studies involving chronic stroke patients [32‐36], longer experimental sessions were performed, lasting 40 ± 17 min on average. In this case, the total number of sessions was lower, 11 ± 4 sessions on average. These findings suggest that the duration and repetition of the experimental protocol may vary based on the stroke phase, resulting in personalized approaches tailored to individual patient’s conditions and needs. However, it is important to note that the lack of studies with a direct comparison between interventions for both sub-acute and chronic stroke patients, makes it challenging to conclusively determine whether the resulting difference is primarily influenced by participants stroke condition or by the implementation of diverse tasks across studies.

The types of rehabilitation tasks are almost evenly distributed among the studies in patients with phantom limb pain and complex regional pain syndrome. Specifically, two studies used mirror therapy [39, 40], and the other three employed motor tasks [41‐43] (Fig. 2). Furthermore, there was significant variability in the time since injury, 9 ± 8 years on average across studies, with a minimum of 5 months and a maximum of 20 years. The average number of sessions performed was 3 ± 2, with each session lasting 25 ± 18 min. The only study in patients with multiple sclerosis recruited participants 15.38 ± 9.95 years post-injury, who performed ten 20-min sessions of VR-based motor tasks.

Regarding the homogeneity within subgroups, the two studies within the group of phantom limb pain and complex regional pain syndrome on motor tasks recruited quite different populations (Table 1). The single subject in the study by Chau et al. [41] was recruited 5 months after injury, whereas the participants of the study by Osumi et al. [42] was recruited after an average of 20.13 years (std: 10.48). The age of the two populations, instead, is comparable: Chau et al. [41] (49 years) and Osumi et al. [42] (range 43–64 years; mean: 52.1). On the other hand, the two studies on mirror therapy reflect similar characteristics in both the clinical and demographic domains: Won et al. [39] (mean time since injury: 5 years; age range: 19–60 years, std: 44 years) and Osumi et al. [40] (mean time since injury: 11.63 years, std: 10.14; age range: 23–71 years, std: 48.1 years).

We analyzed the geographic distribution of the studies categorized by the type of neurological condition (Fig. 3). We established the geographical location of each study by considering the clinical sites where the participants were recruited and where the studies were conducted.

In the case of studies focused on stroke, we observed that half of the studies were evenly distributed between the United States of America (n = 2) [33, 35] and South Korea (n = 2) [31, 34]. The remaining studies were conducted in various European countries, including Germany [30], Sweden [32], and Portugal [36], with one study conducted in Australia [37]. When examining studies related to upper-limb pain syndromes, we observed an equal distribution between the United States of America (n = 2) [39, 41] and Japan (n = 2) [40, 42]. Lastly, the single study on patients with multiple sclerosis was conducted in Switzerland [43].

We reported the HMD, control modality, and VR environment selected by the authors of the studies included in this scoping review (Table 3; Fig. 4). As regards the choice of the HMD, the Oculus (Facebook Technologies, USA) was the preferred option and was used in 9 out of 13 studies [30, 33, 35‐37, 39, 40, 42, 43]. The HTC Vive (HTC Corporation, Taiwan) was used in the 4 remaining studies [31, 32, 34, 41]. As for the control modality, the most popular option (n = 6) [32, 34, 35, 39, 41, 43] included hand controllers for tracking the position of the hands in space. However, holding a hand controller can be challenging or even impossible for individuals affected by hand motor impairment. To solve this issue, the authors of studies [32, 35, 41] employed bands or straps to secure the controller to the participants' arms. Alternatively, three studies implemented hand or finger tracking using the Leap Motion (Ultraleap, USA) infrared camera [30, 40, 42], and one study [43] reported the usage of the markerless tracking included in the Oculus Quest 2 system. Markerless tracking methods offer a significant advantage by eliminating the need for the subjects to hold or wear controllers, allowing them to move without constraints. At the same time, markerless systems may exhibit reduced robustness when dealing with freely moving upper arms in a large space [30]. The other control solutions adopted were electromyographic (EMG) signals of arm muscles [33], EEG signals during motor imagery [36], and a robotic device that tracked the finger’s position and force while assisting the participant’s movements [37].

In most of the cases (9 out of 13 studies) [30, 31, 33, 36, 37, 39, 40, 42, 43], the virtual environment was custom-built, and when reported, the Unity 3D game engine software (Unity Technologies, USA) was utilized. In the other cases, commercially available games were adopted [32, 34, 35, 41]. Regarding the use of audio during the interventions, only three studies explicitly reported it [36, 41, 42]. In all these cases, the audio was used for giving feedback and instructions to the participants.

The studies included in this scoping review reported motor or neurological recovery after fully immersive VR-based rehabilitation, as described in Table 4. In Fig. 5 we report the pre- and post-rehabilitation values of the primary outcome metric of each study; this metric was chosen either as the primary metric suggested by the authors, if available, or as the most used metric among the other studies included in this scoping review, to facilitate inter-study comparison. Almost all the studies, 12 out of 13 [30‐37, 40‐43], reported an improvement in at least one of the clinical metrics following the treatment for some or all the participants. Significant improvements were observed in 5 out of 8 studies focusing on stroke patients [30, 33, 34, 36, 37]. Additionally, 3 out of 4 studies on pain relief reported significant improvements in at least one of the outcome measures [40‐42]. Finally, the single study on multiple sclerosis reported a significant improvement in one of the metrics evaluated [43]. This demonstrates that immersive VR is a valid approach to perform rehabilitation of various types of neurological disorders. Among the benefits reported, stroke survivors exhibited both physical improvements, such as increased strength, expanded range of motion, and enhanced coordination [31, 33, 36, 37], and cognitive improvements, such as better attention, memory, and executive functions [33, 36]. Finally, participants with stroke also reported functional benefits, such as increased independence in activities of daily living [31], improved mobility, and enhanced quality of life [33, 36]. Studies involving participants with pain-related syndromes generally reported significant pain alleviation.

Immersive VR is commonly associated with a phenomenon known as cybersickness [51]. Cybersickness refers to unpleasant and negative effects that individuals may experience when using immersive VR. It is characterized by symptoms like motion sickness, including nausea, dizziness, disorientation, headache, sweating, and general discomfort [52]. The manifestations of cybersickness differ from person to person, with some people experiencing mild discomfort or disorientation and others encountering more severe symptoms that could impede their ability to use VR or force them to interrupt the VR experience. These symptoms can manifest during or after VR use and may persist for different periods of time [53].

Only 4 of the analysed 13 studies [30, 33, 35, 39] included an assessment of cybersickness (Table 5). Three studies [30, 33, 39] assessed cybersickness using the Simulator Sickness Questionnaire (SSQ) [54], whereas one study [35] employed the Motion Sickness Susceptibility Questionnaire Short Form (MSSQ-Short) [55]. None of the four articles reported significant cybersickness symptoms in the participants. Among the remaining studies, only three of them discussed the cybersickness issue [31, 32, 34]. However, none of them reported adverse effects or problems related to cybersickness in the participants.

The most recurrent barriers in the use of immersive VR for upper limb rehabilitation included motion tracking and limited adoption.

As previously mentioned, the use of controllers to track the affected limb can be problematic for some participants, even though a potential solution is to secure the controller to the arm using straps. On the other hand, the use of infrared and depth-sensing cameras, such as the Leap Motion, may introduce errors and limitations like high latency or insufficient accuracy [30]. Another solution, not exploited by any of the studies included in this review, is the usage of the Vive tracker when using the Vive HMD (HTC Corporation, Taiwan). The Vive tracker is a compact device that can effortlessly be attached to the participants' arms without impeding their movements. This device provides an accurate real-time tracking system [56] that can address the tracking errors and related issues. Additionally, both HTC and Oculus offer built-in hand-gesture recognition capabilities. However, none of the studies included in this scoping review exploited this functionality. Regarding the choice of the technology, most of the studies (Fig. 4) used the Oculus device [30, 33, 35‐37, 39, 40, 42, 43], but the authors did not specify the rationale behind their choice. We speculate that one reason could be the price difference between the Oculus and HTC devices: the Oculus devices tend to be slightly cheaper than HTC devices (100$ difference approximately). Another contributing factor could be the ease of use and calibration of the Oculus devices: unlike HTC devices, Oculus systems do not require the placement of base stations in the room to track the headset and controllers, thus simplifying the setup procedures.

It has been reported [34] that the adoption of immersive VR may face obstacles due to generational differences and caregivers' limited technical experience. To address these challenges, it is essential to focus on training and education initiatives. Additionally, improving user-friendliness and developing more plug-and-play systems can enhance overall ease of use. Nevertheless, given the swift global surge in technology adoption, we anticipate that substantial advancements in these aspects are not only achievable but also imminent.

Regarding the facilitators, the immersive and interactive nature of VR environments, along with the element of fun and challenge, can motivate participants to exert more effort during rehabilitation exercises [30, 31, 33‐35]. In addition, VR is suitable for telerehabilitation where patients perform the exercises from the convenience of their own home, guided, if needed, by voice instructions from the therapist. Telerehabilitation can reduce or even eliminate the need for patients to be physically present in the clinical facilities, reducing the inconvenience and costs associated with transfers. Furthermore, VR facilitates the rehabilitation of patients who find it difficult to undergo conventional therapy due to sever motor deficits [33]. Also, the availability of objective performance measures provides clinicians with a comprehensive picture of the patients’ progress and status [31, 37]. Finally, the flexibility of VR allows for the development of personalized interventions and their adaptation over time, ultimately increasing their effectiveness [30, 32, 33]. All these elements highlight the VR potential of allowing a patient-centric approach that can enhance the quality of care and improve functional outcomes. More effective rehabilitation can reduce the frequency of interventions and, combined with the affordability of VR technology, reduce overall healthcare costs [37].

Immersive VR has shown promising potential for upper-limb pain relief therapy [40, 41]. A notable advantage of immersive VR is that it can integrate a variety of rehabilitation exercises that may not be implemented in traditional non-virtual environments, such as the creation of a more “sophisticated” immersive form of mirror-therapy [5, 57]. The virtual setting can overcome the constraints of the physical world, allowing innovative and tailored approaches to improve the effectiveness and engagement of the rehabilitation process. It is worth noting that most of the studies in our review created custom immersive VR environments. A custom-built environment offers the possibility to be tailored to the intended use-case in terms of therapists’ requirements and patients’ needs [30, 33]. However, the development of a 3D VR environment compatible with HMDs requires technical skills and expertise. To promote the wider adoption of immersive VR as a rehabilitation tool, there is a need for more commercial games specifically designed for this purpose.

Furthermore, most of the studies included in the analysis did not employ structured questionnaires for assessing the presence of cybersickness symptoms. This may contribute to underreporting [34] and, consequently, a deficiency in comprehensive evaluation. By collecting information on participants' prior experience with VR, implementing quantitative measures of cybersickness, and establishing standardized protocols tailored to specific pathologies, researchers and therapists could better assess and address the potential adverse effects of immersive VR-based rehabilitation.

Finally, as depicted in Fig. 5, we found a lack of standardization in reporting the primary clinical outcomes, even within the same study group (same neurological disease and rehabilitation task). Indeed, despite using the same metric (mostly FMA-UE for stroke studies and SF-MPQ for phantom limb pain studies), studies reported different ranges of values due to differences in the measurement methodology of the metric. This discrepancy makes it challenging to compare the results across different studies and approaches. A more standardized approach and guidelines for the choice of metrics for specific neurological disorders are needed in this field.

Most of the problems reported by the authors are related to the research methodology, and in particular to study design [30, 32, 34‐36, 40, 42] and limited sample size [30, 33‐37, 39, 41, 43]. In particular, the authors emphasize that the number of participants plays a crucial role in determining the true effectiveness of immersive VR for upper-limb rehabilitation. Similarly, study design limitations, such as a limited number of interventions and the absence of randomized or controlled designs, pose challenges when evaluating the advantages of immersive VR for neurorehabilitation. These limitations extend to the possibly to compare these benefits with traditional rehabilitation therapies. In future studies, it is advisable to address these issues by incorporating a larger participant pool within randomized controlled trials.

Envisaging the future of upper-limb rehabilitation through immersive VR interventions reveals several critical considerations highlighted in this scoping review.

Firstly, the establishment of standards and protocols is essential to ensure consistency among studies and facilitate quantitative comparisons of diverse interventions. Customizing VR interventions based on individual patient needs, including motor skills, disability levels, and personal preferences, is crucial for optimizing intervention effectiveness. Additionally, it is imperative to correlate quantifiable metrics extracted from gamified tasks with clinical outcomes to establish benchmarks for patients’ capacity during upper limb VR-rehab tasks. This step is vital for comparisons across different groups and tracking rehabilitation progress over time.

Secondly, understanding participant acceptability and adherence to VR interventions is paramount. The integration of qualitative and quantitative feedback obtained during gamified tasks should be supported by standardized protocols and questionnaires related to usability, sense of presence, and cybersickness. This aspect is crucial for the development of VR-based medical devices accepted in clinical settings, enabling manufacturers to create devices specifically tailored for clinical purposes.

Thirdly, expanding the scope of VR interventions to address various conditions requires assessing suitability across diverse age groups, disability levels, and clinical scenarios. Evaluating the long-term effects of VR interventions on upper limb functions is essential for a comprehensive assessment. Additionally, exploring new access technologies, such as brain-computer interfaces, holds promise in enhancing accessibility to VR-based rehabilitation, allowing individuals with motor disabilities to access more effective therapies by sending commands directly through the brain.

Considering the barriers and facilitators identified in this review, prioritizing the avoidance of potential adverse effects should be fundamental in VR rehabilitation protocols. A comprehensive collection of patient information regarding adverse effects from immersive VR usage is vital for targeted interventions and preventing negative impacts on patients.

Integrating VR into clinical care necessitates strategic planning, including examining effective strategies for incorporating VR rehabilitation into standard clinical practice. This involves healthcare provider training, ensuring patient accessibility, and active involvement of patients, caregivers, healthcare providers, technology developers, and researchers in a comprehensive and collaborative manner, utilizing participatory and user-centred design in the development of gamified tasks.

Collectively incorporated into future studies, these elements hold significant potential to profoundly shape the future landscape of rehabilitation therapies for upper-limb neurological disorders.