This is an open-access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in JMIR Serious Games, is properly cited. The complete bibliographic information, a link to the original publication on https://games.jmir.org, as well as this copyright and license information must be included.
The increasing number of children with cerebral palsy (CP) has a serious impact on individuals, families, and society. As a new technology, virtual reality (VR) has been used in the rehabilitation of children with CP.
This study aimed to systematically evaluate the effect of VR training on balance, gross motor function, and daily living ability in children with CP.
PubMed, Embase, The Cochrane Library, Web of Science, and China National Knowledge Infrastructure databases were searched by computer, with the search period being from the establishment of each database to December 25, 2021, to collect randomized controlled trials (RCTs) on the effects of VR training on balance, gross motor function, and daily living ability in children with CP. The Cochrane risk of bias assessment tool was used to conduct quality assessment on the included literature, and RevMan software (version 5.3) was used to analyze data.
A total of 16 articles were included, involving 513 children with CP. VR training can improve the balance function (Pediatric Balance Scale: mean difference 2.06, 95% CI 1.15-2.97;
VR training can significantly improve the balance function and gross motor function of children with CP, but the effect on the daily living ability of children with CP remains controversial.
Cerebral palsy (CP) is a nonprogressive, persistent syndrome occurring in the brain of the fetus or infant [
Virtual reality (VR) refers to a virtual environment that is generated by a computer and can be interacted with. VR can mobilize the visual, auditory, tactile, and kinesthetic organs of children with CP, so that they can actively participate in the rehabilitation exercise. In this way, the central nerve conduction and peripheral motor control of children can be coordinated and unified, which is conducive to the rehabilitation of children [
It has become a focus of scholars in China and abroad to improve the motor ability, abnormal posture, and quality of life of children with CP. Previous meta-analyses have shown that VR can significantly improve children’s hand function, balance function, gross motor function, and walking function [
This study followed the requirements of the International Meta-analysis Writing Guidelines (The PRISMA [Preferred Reporting Items for Systematic Reviews and Meta-Analyses] statement for studies that evaluate health care interventions: explanation and elaboration) for the selection and use of methods.
In this study, 2 researchers independently retrieved 5 databases including PubMed, Embase, The Cochrane Library, Web of Science, and China National Knowledge Infrastructure (CNKI) to find randomized controlled trials (RCTs) of VR training effects on balance, gross motor function, and daily living ability in children with CP. Moreover, the sources of the CNKI database were limited to Chinese Social Sciences Citation Index and Chinese Science Citation Database, and the retrieval dates of all databases were from the establishment of each database to December 25, 2021. Supplementation measures were the tracking of relevant systematic reviews and references of the included literature. The retrieval adopted the method of combining subject words with free words—the Boolean operators “AND” and “OR” were used to combine and connect—and was determined after repeated prechecks. English search terms included “cerebral palsy,” “CP,” “children of cerebral palsy,” “cerebral palsy children,” “virtual reality,” “VR,” “virtual environment,” “video game,” “clinical trial,” “randomized controlled trial,” etc. An example of the literature retrieval strategy for the PubMed database is given (
The inclusion criteria were as follows:
Population for research objects: clinically diagnosed children with spastic CP; their race and gender are not limited, and they are aged <16 years
Interventions for experimental group: VR training, VR training combined with conventional rehabilitation training, or VR training added on the basis of control group training
Comparison for control group interventions: daily physical activities, balance training, conventional rehabilitation training, or comprehensive rehabilitation training, etc
Outcome for outcome indicators: balance function was evaluated by Berg Balance Scale (BBS) and Pediatric Balance Scale (PBS); gross motor function was evaluated using the Gross Motor Function Measure Scale (GMFM), including GMFM-66, GMFM-E and GMFM-88; and the ability of daily living was assessed by Pediatric Evaluation of Disability Inventory (PEDI) and The Functional Independence Measure for Children (WeeFIM)
Study design for study type: RCTs
The exclusion criteria were as follows: non-RCTs; republished papers or papers with poor-quality evaluation; literature not in Chinese and English; full text not available; outcome indicators did not meet the requirements of this study or data could not be extracted; and intervention group content did not meet the requirements.
In this study, 2 researchers independently retrieved literature from 5 databases and downloaded the retrieved literature into EndNote X9 software (Clarivate) in batches. After all the documents from the 5 databases were retrieved, duplicate documents were first removed in Endnote X9 software. Second, preliminary screening was carried out by reading the title and abstract. Third, the papers were screened according to the inclusion and exclusion criteria of this study. Finally, full-text readings were conducted to identify studies for final inclusion.
In all, 2 researchers extracted the data of the basic information and outcome indicators of the included papers and contacted the original author by email for unclear or missing data in the study. When the information extracted by 2 researchers was inconsistent, a third researcher would participate in a discussion to reach a consensus. The extracted information included basic information (author, year, country, sample size, type of CP, and patient age); experimental characteristics (intervention content, single intervention duration, frequency, and cycle); and outcome indicators. The researchers set up a table by reading the articles that met the inclusion criteria in detail and recording the relevant information.
This study used The Cochrane Collaboration’s Tool for Assessing Risk of Bias to analyze the literature using 7 methods: random sequence generation, allocation concealment, blinding of subjects and researchers, blinding of raters, incomplete outcome data, selective reporting, and other biases. These analysis results were categorized into 3 types of quality evaluation: low risk, unclear, and high risk. This process was carried out by 2 researchers independently. In case of disagreement, a third researcher would join in to discuss and make a decision. The quality of the literature was divided into 3 levels: grade A (low risk, meeting 4 or more items); grade B (low risk, meeting 2 or 3 items); and grade C (low risk, meeting 1 or no items, bias likely to occur) [
Data analysis was performed using RevMan software (version 5.3; Cochrane) and following the PRISMSA guidelines. The Q statistic test (
In all, 2 researchers searched 5 databases, including PubMed (n=82), Embase (n=191), The Cochrane Library (n=147), Web of Science (n=359), and CNKI (n=11). A total of 793 papers were retrieved, including 11 Chinese papers, 759 English papers, and 3 papers [
Literature screening process.
A total of 16 articles representing16 RCTs were included—more specifically, 5 Chinese articles and 11 English articles. The publication period is from 2013 to 2021, and the countries of publication are China, Turkey, India, and South Korea. A total of 513 children with CP were included, and the CP types included spastic hemiplegia and spastic diplegia. The intervention of the experimental group included virtual time-limited training and the combination of VR training and regular rehabilitation training, etc; and the intervention of the control group included daily physical activities and regular rehabilitation training. The intervention duration of VR training was from 15-60 minutes, the frequency was from 2-6 times a week, and the cycle was from 3-12 weeks. (
A total of 16 RCTs were included in this study, and their risk of bias was shown in
A total of 126 cases of children with CP in 6 RCTs participated in the VR training and had relevant scoring conducted by using PBS evaluation, as shown in
A total of 123 cases of children with CP in 3 RCTs participated in the VR training and had relevant scoring conducted by using BBS evaluation, as shown in
Forest plot of the effect of virtual reality on the Pediatric Balance Scale scores in children with cerebral palsy. IV: inverse variance.
Forest plot of the effect of virtual reality on the Berg Balance Scale scores in children with cerebral palsy. IV: inverse variance.
A total of 236 cases of children with CP in 7 RCTs participated in VR training for the impact on their gross motor function, as shown in
A subgroup analysis in terms of CP type, training frequency, and period is shown in
Forest plot of the effects of virtual reality on the gross motor function in children with cerebral palsy. IV: inverse variance.
Subgroup analysis of the effects of virtual reality training on gross motor function in children with cerebral palsy.
Group | Study, n | Model | SMDa (95% CI) | |||
|
||||||
|
Hemiplegia | 2 | 0 | Fixed effects model | 0.72 (0.30 to 1.14) | <.001 |
|
Diplegia | 2 | 0 | Fixed effects model | 0.28 (–0.24 to 0.81) | .29 |
|
Other | 3 | 52 | Random effects model | 0.64 (–0.00 to 1.29) | .05 |
|
||||||
|
≤4 days/week | 4 | 0 | Fixed effects model | 0.29 (–0.11 to 0.68) | .15 |
|
>4 days/week | 3 | 0 | Fixed effects model | 0.86 (0.51 to 1.22) | <.001 |
|
||||||
|
<6 weeks | 2 | 0 | Fixed effects model | 0.72 (0.30 to 1.14) | <.001 |
|
≥6 weeks | 5 | 31 | Fixed effects model | 0.53 (0.19 to 0.87) | .002 |
aSMD: standardized mean difference.
A total of 274 children with CP from 7 RCTs participated in the research on the VR training effect on the daily living ability of CP children, as shown in
To explore sources of heterogeneity, a sensitivity analysis was performed by successive elimination of studies (see
After excluding Atasavun Uysal et al [
A subgroup analysis of VR system type, training frequency, and training period is presented in
Forest plot of the effects of virtual reality on the daily living ability in children with cerebral palsy. IV: inverse variance.
Combined effect of daily living ability after excluding individual studies.
Study eliminated | SMDa (95% CI) | Model | |||
Acar et al [ |
0.48 (0.07 to 0.89) | .02 | 58 | .03 | Random effects model |
Atasavun Uysal et al [ |
0.55 (0.30 to 0.81) | .001 | 36 | .16 | Fixed effects model |
Jha et al [ |
0.35 (–0.13 to 0.84) | .15 | 70 | .006 | Random effects model |
Sahin et al [ |
0.30 (–0.19 to (0.79) | .23 | 68 | .009 | Random effects model |
Tarakci et al [ |
0.34 (–0.13 to 0.82) | .16 | 70 | .006 | Random effects model |
Xu et al [ |
0.29 (–0.18 to 0.77) | .23 | 67 | .01 | Random effects model |
Zhao et al [ |
0.27 (–0.17 to 0.70) | .23 | 62 | .02 | Random effects model |
aSMD: standardized mean difference.
Forest plot of the effects of virtual reality on the daily living ability in children with cerebral palsy (excluding Atasavun Uysal et al [
Subgroup analysis of the effects of virtual reality training on daily living ability in children with cerebral palsy.
Group | Study, n | Model | SMDa (95% CI) | ||||||||
|
|||||||||||
|
Wii | 2 | 54 | Random effects model | 0.09 (–0.66 to 0.85) | .81 | |||||
|
Kinect | 3 | 0 | Fixed effects model | 0.69 (0.35 to 1.03) | <.001 | |||||
|
Other | 1 | N/Ab | N/A | 0.75 (0.17 to 1.33) | .01 | |||||
|
|||||||||||
|
≤4 days/week | 4 | 36 | Fixed effects model | 0.39 (0.07 to 0.70) | .02 | |||||
|
>4 days/week | 2 | 0 | Fixed effects model | 0.85 (0.42 to 1.28) | <.001 | |||||
|
|||||||||||
|
≤6 weeks | 4 | 60 | Random effects model | 0.48 (–0.03, to 1.00) | .06 | |||||
|
>6 weeks | 2 | 0 | Fixed effects model | 0.61 (0.19 to 1.03) | .005 |
aSMD: standardized mean difference.
bN/A: not applicable.
Since the BBS indicators used in the study were limited, only the PBS, gross motor function, and daily living ability indicators were tested by Begg funnel plot. The test result of PBS was
VR technology is a new practical technology; it integrates computer software and hardware, artificial intelligence, sensing, simulation, and other scientific technologies to give users an immersive sense and can provide users with the ability to interact with virtual objects. At present, VR technology has been widely used in the field of medicine to promote the rapid recovery of patients [
The balance function depends on the central nervous system in many aspects. In addition to the central nervous system injury, children with CP also have skeletal deformity, triceps spasm in the calf, increased muscle tension on one side, and clubfoot, which seriously affect the balance ability of children with CP [
VR training, which provides children with CP with a similar environment to the real world, gives a visual, auditory, and kinesthetic stimuli and interaction; activates the brain specific movement area; increases the blood flow of motor cortex; and prompts the cortical neural and cortex tissue improvement and restructuring. Therefore, the movement function of children can be compensated and activated, and motor function is improved [
In this study, WeeFIM and PEDI were used to evaluate the daily living ability of children with CP. The results showed that there was significant heterogeneity between studies and that VR training did not significantly improve the daily living ability of children with CP compared to the control group. To reduce the heterogeneity between studies, the method of successive elimination of studies was used for sensitivity analysis, and the results showed that the heterogeneity came from Atasavun Uysal et al [
Our study found that VR can improve balance function and gross motor function in children with CP, which is consistent with previous studies. After removing the literature with high heterogeneity [
This study only included Chinese- and English-language literature; the indicators included were limited to GMFM, BBS, and PBS; and the number of studies included in several groups was relatively small in the subgroup analysis, so the conclusions obtained have certain limitations. The content of VR training is diverse; the content of regular rehabilitation is not consistent; and the duration, frequency, and cycle of training are also different, which may also be the reason for the heterogeneity of studies. Part of the literature included in this study did not clarify whether allocation concealment was used, the researchers or subjects were blinded, or the evaluators were blinded, so there may be bias in the research results.
Meta-analysis results showed that compared to the control group, VR training significantly improves the balance function and gross motor function of children with CP, but the impact on the social function of children with CP is still controversial. Therefore, more RCTs with high quantity and quality are suggested to be performed in the future in an effort to further confirm the treatment effects of VR training on balance function, gross motor function, and the daily living ability among children with CP, as well as to offer more solid evidence for clinical trials. Therefore, we suggest carrying out more high-quality RCTs with large samples in the future to further confirm the efficacy of VR training on balance ability, gross motor function, and the daily living ability of children with CP and provide more reliable evidence for clinical practice.
Literature retrieval strategy for the PubMed database.
Basic features of the included papers.
Risk of bias of the included papers.
Begg funnel plots.
Berg Balance Scale
China National Knowledge Infrastructure
cerebral palsy
Gross Motor Function Measure Scale
mean difference
Pediatric Balance Scale
Pediatric Evaluation of Disability Inventory
Preferred Reporting Items for Systematic Reviews and Meta-Analyses
randomized controlled trial
standardized mean difference
virtual reality
The Functional Independence Measure for Children
This work was supported by the Key Laboratory of Human Sports Ability Development and Guarantee (11DZ2261100).
None declared.