A Feasibility Study on the Application of Virtual Reality Technology for the Rehabilitation of Upper Limbs After Stroke

According to the World Health Organization, 15 million people worldwide suffer a stroke each year. Studies have shown that 86% of stroke patients will suffer from upper limb dysfunction after the onset of the disease [1] and that 30% [2] to 66% [3, 4] still suffer from upper limb dysfunction 6 months after the onset of the disease. Upper limb dysfunction can seriously affect a patient’s motor function and daily life. Therefore, upper limb rehabilitation is particularly important.

Rehabilitation therapy is characterized by high intensity, repeatability, functionality and task specificity [5, 6]. Traditional rehabilitation methods not only lack interestingness but also fail to give patients successful feedback and experience [7]. Previous research studies have shown that the participation and enthusiasm of stroke patients are key factors affecting the rehabilitation outcome, and there is a significant positive correlation between patients’ active continuous participation and the rehabilitation effect [8]. Therefore, it is of great significance to identify an active and effective rehabilitative method for improving the upper limb function of stroke patients. Researchers have proposed the application of VR technology for the rehabilitation of upper limbs after stroke as one such method [9,10,11].

VR is a computer simulation technology that can create virtual worlds and experiences within them, and it has three basic aspects: immersion, interaction and imagination [12]. Users interact with the virtual environment created by computers in a variety of ways to create an immersive feeling [13, 14]. With the application of VR technology, patients can transmit their actions to the computer through input devices (such as data gloves, motion capture devices, and computer mice) and receive sensory feedback, such as visual, auditory or touch feedback, from the output device [6]. Furthermore, the visual and proprioceptive feedback received by the users can also encourage target behaviours of the users, maintain their motivation and enthusiasm, and help them achieve a happy, successful emotional experience, thereby encouraging them to practice continuously until they learn the behaviours. VR provides technical support for three key elements of rehabilitation: repetition, performance feedback and motivation maintenance. This training method not only considerably reduces the amount of human and material resources needed for training but also increases the interestingness of the treatment to stimulate the enthusiasm of the patients, transform passive treatment into active treatment, and improve treatment efficiency [15].

Although increasingly more research studies on VR applications in rehabilitation have been conducted in recent years, most studies have focused on the effect of rehabilitation after stroke, and there are few studies on how to design VR systems for upper limb rehabilitation. Thus, the aim of this study was to review and analyse the existing literature to explore the clinical feasibility of VR applications in stroke upper limb rehabilitation and determine which VR systems are suitable for the rehabilitation of upper limbs after stroke.

The effect of VR on the rehabilitation of upper limbs in stroke has been confirmed by many clinical studies. Saposnik et al. believe that the combination of VR and video games is a novel and potentially useful technology that can be combined with traditional rehabilitation treatments to improve the function of upper limbs after stroke [11]. Because this kind of training is more interesting, patients show stronger rehabilitation motivation in the VR environment than in the traditional rehabilitation environment [16]. In addition, Laver et al. conducted a systematic evaluation of the clinical trials of VR application in stroke rehabilitation in recent years, and the results showed that there was a trend suggesting that higher dose (more than 15 h of total intervention) was preferable as were customised virtual reality programs [17]. Hatem also demonstrated through a systematic literature review that the combination of VR with other rehabilitation therapies is superior to rehabilitation therapy alone [18].

The Application of VR Combined with Other New Technologies.

The combination with other new technologies has yielded complementary effects, indicating the range of potential applications is large. The three main combinations involving VR are as follows:

• Combination of VR and rehabilitation robot technology. At present, rehabilitation robot technology has been widely used in the field of clinical rehabilitation. In these systems, mechanical feedback, visual feedback and tactile feedback devices are mostly used as input and output devices so that patients can carry out rehabilitation training assisted by rehabilitation robots in a virtual environment [19]. The assistance of the machine makes it possible for patients with severe motor dysfunction to participate in rehabilitation training; in addition, it can also correct poor postures and movements of patients and improve the rehabilitation effect [20].

• Combination of VR and telerehabilitation technology. VR telerehabilitation systems can facilitate convenient and economical exchanges of information and rehabilitation training platforms for patients and rehabilitation therapists. Patients can use VR equipment to participate in rehabilitation training in a virtual environment, as required. For rehabilitation therapists, after they receive the relevant parameters and videos the of patients’ movements, they can select and tailor the rehabilitation training mode according to patients’ individual conditions. Therefore, the combination of VR technology and telerehabilitation technology is helpful to solve the problem that scientific rehabilitation training cannot be carried out because of the geographical remoteness and limited rehabilitation conditions [21, 22].

• Combination of VR and brain-computer interface technology (BCI). The BCI technology can be used to detect the electroencephalogram (EEG) signals generated by the brain when planning volitional movements, identify the signals corresponding to patients’ movement intentions by extracting and classifying the signal features, and convert these signals into control commands to control external devices, thereby assisting patients in generating the corresponding actions [23]. The brain signals generated by the patients during motion planning is processed and realized through the VR system. The combination of these two techniques can help patients with minimal residual motor function complete rehabilitation and promote the remodelling of neural function.

VR can provide patients with more active, colourful and inspiring situational feedback. Therefore, VR and other rehabilitation technologies can be combined to enhance the efficacy of this rehabilitation technology.

In addition, VR technology may be used to develop a relatively objective evaluation standard for efficacy. There are more than 40 evaluation methods for the rehabilitation of upper limbs after stroke, and the most commonly used methods include the Fugl-Meyer (FMA), Wolf Motor Function Test (WMFT), and EQ-5D [24]. However, these scales all depend on the subjective evaluation of rehabilitation physicians, so it is difficult to evaluate the curative effect of the treatment comprehensively and objectively. In contrast, the parameters collected by VR equipment are relatively objective. In the future, the parameters collected during of exercise can be calculated as evaluation indexes of the rehabilitation effect.

When VR is used in the rehabilitation of upper limbs after stroke, appropriate hardware and software devices should be selected so that the patients can interact with the system.

3.1 Hardware Selection

Hardware devices, including display devices and input devices, are tools used by the user to interact with the virtual environment (software) and are means of communication between the user and virtual reality system.

Display Device.

High immersion is a feature of VR. At present, VR display devices can be categorized as either immersive or non-immersive devices. The immersive devices completely isolate the user’s vision from reality. What the user sees is a computer-generated virtual environment with a wide field of vision, which makes the user feel as if he or she is in a brand-new environment. Non-immersive devices display the virtual environment on a computer or TV screen [10]. There are different opinions on the immersive and non-immersive devices. Some studies have pointed out that there are three problems in using immersive devices [25]: (1) high requirements for 3D images in virtual environments increase the difficulty and cost of system development; (2) a certain amount of space on the ground may be required for training; and (3) VR technology can cause slight adverse reactions, such as transient dizziness. However, other studies tend to use immersive devices because compared with non-immersive devices, immersive devices are more attractive to patients and more beneficial to patient recovery.

With the rapid development of 3D imaging technology, VR software development is not difficult. In terms of the hardware, those of immersive VR devices need to be upgraded frequently, and immersive portable devices have become the mainstream VR devices. Therefore, if portable immersive devices are used for rehabilitation, they will not take up additional space. Dizziness can be resolved by avoiding jitter in the virtual screen and limiting the use time. All the problems existing in the application of immersive devices in the rehabilitation of upper limbs of stroke that are mentioned above can be solved. Therefore, immersive devices are a better choice.

Input Device.

The VR system interface is a tool for users to interact with the virtual environment and a communication channel between users and virtual reality. It is related to the gameplay of games. Low usability of a VR system interface design will affect the consistency of operation. Currently, VR devices that are on the market have different input methods due to the different brands and different system designs used. For example, the interaction mode of Google Cardboard is gaze and click; HTC Vive uses two six-degree of freedom controllers; Oculus uses the Xbox One controller but will eventually use the 6DOXF dual controller. All of these systems enable users to use a more advanced immersive interaction mode. In addition, there are other types of input devices, such as hand tracking devices.

As shown in Table 1, when VR technology is applied to stroke rehabilitation, its common input devices include computer mice [26], motion capture devices [10], data gloves [27], and rehabilitation robots [28]. When a mouse is used as the input device, the hand is fixed due to the posture needed to hold the mouse, which is not conducive to the rehabilitation of the patient, and the applications are relatively limited. Motion capture devices are suitable for rehabilitation with a large range of motion, as data on finger movements and other fine movements cannot be monitored. The positions of hand tracking devices can be flexibly adjusted according to the needs of the user, and they have better flexibility and portability in tracking the movements during rehabilitation. When VR is combined with a rehabilitation machine, the data collected by the rehabilitation machine is more accurate. For example, it can measure joint mobility, the average joint movement speeds, and other kinematic parameters during specific training tasks (motion accuracy, motion trajectory smoothness, motion trajectory length, movement coordination, etc.), making it easier for people to monitor the rehabilitation effect.

3.2 Software Design

The interaction logic of different software is very different. The type of VR software selected determines the interaction mode between the patients and VR rehabilitation system.

VG or VE.

There are two main types of VR that have been applied in the field of stroke upper limb rehabilitation. One is a virtual reality game (VG), in which there are settings for barriers, difficulty, rewards and leaderboards. The other is a simple virtual reality environment (VE), in which people and objects operated by the users simply interact with the objects in the virtual environment. As we mentioned earlier, patients’ own motivation plays an important role in determining the outcome of treatment [8], so it is necessary to compare the two types of applications and determine which one is more likely to motivate patients for rehabilitation.

According to the flow theory proposed by Hungarian psychologist Csikszentmihalyi [29], when a person is completely immersed in an activity, he or she will ignore the existence of other things. This state is called flow in which there is great joy, and people are often willing to pay a great price to experience this state. When people are fully engaged in an activity and are filtering out all irrelevant perceptions, we say that they enter a state of immersion. And, it is in this state that they are at their happiest moment. VE presents tasks in a repetitive manner, which makes it difficult for patients to stay interested in prolonged exercise. Compared with VE, VG is more interactive and interesting. In the process of completing the game level, patients constantly challenge themselves, improve their skills and immerse themselves in the game.

When experiencing VG, stroke patients have visual feedback information, including scores, feedback on user actions (whether they are correct or not), tips on how activities should be carried out and incentives, as well as feedback sounds such as key triggers and alarms, which are designed to attract users’ attention and encourage them to complete tasks easily. The game has difficulty level or barrier settings, which can arouse the patient’s internal drive for excellence and enhance the patient’s motivation to complete rehabilitation.

VG Design for Rehabilitation.

The only criterion used to judge whether a game is good or not is whether the players like the game. In game design, the enjoyment of players is the most important issue [30]. Based on Cziksentmilalyi’s flow theory, Sweetser proposed the game flow model, which contains eight core elements [30]:

• Focus. The game should require concentration, and players should be able to focus on the game.

• Challenge. The game should be sufficiently challenging and meet the player’s skill level.

• Player skills. The game must support the development and mastery of the player’s skills.

• Control. Players should have a sense of control over their actions in the game.

• Clear goals. The game should provide clear goals for players at the appropriate time.

• Feedback. Players must receive the appropriate feedback at the appropriate time.

• Immersion. Players should experience in-depth and relaxed efforts in the game.

• Social interaction. The game should support and create opportunities for social interaction.

This model is universal for game design, but designing games for rehabilitation patients has its particularity; that is, the design of rehabilitation games should not only satisfy the users’ needs for playing games but also achieve the goal of rehabilitation.

Although there is no uniform standard for the design of stroke rehabilitation games, it can be concluded from the relevant research that the following attributes are desirable in stroke rehabilitation games:

• Direct game cues. Although games have immediate and continuous feedback, the majority of stroke users are elderly people, and it is difficult for them to understand indirect action cues [9]. For example, it is difficult for them to understand how to perform vertical motions (raising and lowering arms) to control the motion of a game character in a horizontal position. Therefore, the cues in the game should be intuitive to most stroke rehabilitation patients.

• Difficulty settings for rehabilitation. The difficulty of the game should be set according to the rehabilitation course of the target stroke patient. Some researchers think that the difficulty of the game should be determined by the rehabilitation physician, and the difficulty of the game should be set according to a set of basic difficulty values provided by the rehabilitation doctor [9]. However, after communicating with rehabilitation doctors, we learned that they do not know all the rehabilitation methods, many of which are summarized by long-term medical practice. Therefore, when a game is used in rehabilitation, if the combination of the rehabilitation method and game is not familiar to the rehabilitation doctor, many tests need to be performed to set the difficulty level and duration according to the rehabilitation doctor’s advice.

• Fun but low-intensity games. Through more than 300,000 questionnaires, QuanticFoundry Company found that the main motivation of players over the age of 36 is fantasy and completion and that completion is the most stable game motivation, which does not weaken with age and is ranked as one of the top three motivations for both male and female players. Research has pointed out that older video game players prefer simple games, and games designed for elderly people should also take into account their decreased sensitivity and longer response time (e.g., using large font, allowing delayed response) [31].

All of these are useful for VR rehabilitation game design. In addition, Steam Store is considered the database of the game world, and player evaluation is the best screening procedure. Although the VGs on Steam are not designed for stroke patients, the VG reviews can help us understand the problems of VG. There are many VR games on Steam that receive many poor reviews. By analysing a large number of player reviews on Steam, we determined that the problems are related to the following three aspects:

• VR technology, such as picture quality and action control. The immersive experience of VR games is based on the advancement of science and technology. There are some problems in the technology used in games that received poor reviews. In addition to dizziness, delay and other problems mentioned above, the problems also include rough pictures, the system freezing and loopholes. An unreasonable motion control designs are another major problem of VG. If the game control design is unreasonable, players will often collide with objects in the game or even walk through objects directly. Some specific situations can cause discomfort for the users, such as being in high or narrow spaces.

• Specific designs, such as plot and level designs. Regarding VR systems, the game concept and design itself are the core aspects of the game, and they are what lead the players to have fun. Some games with overstep stories, boring repetitive operations and poor designs commonly receive poor comments from Steam players.

• Cooperation between the VR hardware and game operation. The operations of some games are not coordinated with the configuration of VR hardware devices, leading to players’ difficulty in operation.

Therefore, in the design of stroke upper limb rehabilitation games, attention should be paid to avoiding the above problems.

Besides, stroke patients have support needs for immersive training, because a large portion of them are elderly and they are not familiar with modern new technologies. The training program should account for the patients’ education level, life and working experience of the elderly. Thus, they could be provided with suitable learning support need and, therefore, are more likely to obtain amusement by choosing different immersion programs from the VR game.

In conclusion, the application of VR in the field of stroke rehabilitation can make the long-term rehabilitation process interesting rather than boring and can stimulate the enthusiasm of patients to improve the efficiency of treatment. Therefore, in-depth discussions on the content designs and applications of VR rehabilitation games and methods of increasing the interestingness of rehabilitation may be important directions of future research. In addition, the VR and other technologies can also be integrated to achieve complementary benefits. For example, one can try to combine VR with rehabilitation robot and brain computer interface technologies. There are still many unresolved issues to be studied and discussed regarding the integration n of VR with other technologies.

Furthermore, regarding VG design, no studies on the needs of stroke patients as a user group have been conducted. This study can only summarize the characteristics of VR and existing problems in the context of rehabilitation of upper limbs after stroke. Therefore, it is suggested that in the future, researchers who develop VR rehabilitation games investigate the needs of target users, such as the preferred game types, screen styles, and game music. After the user needs are determined, the background, scenes, elements, tasks, levels, rules, and interface for the game and the user control method can be designed according to the user’s specific needs and the gameplay enjoyment model. Therefore, it is expected that a VG design that meets the needs of the target users can be designed.

In this feasibility study, the findings showed the clinical feasibility of the application of VR for upper limb rehabilitation after stroke, as well as the existing problems that need to be solved. These problems are important targets of designing a VR system suitable for stroke upper limb rehabilitation.