Effect of active video games on cognitive functions in healthy children and adolescents. Systematic review of randomized controlled studies

Aims: To examine the effect of active video game interventions (exergames) on cognitive functions in healthy children and adolescents. Methods: A systematic review was conducted following the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) statement, using PubMed, Scopus, and Web of Science databases. The search was restricted to randomized controlled studies evaluating the effect of exergames in a young healthy population on some cognitive variable. Results: Five studies met the inclusion criteria. All studies identified an improvement in some cognitive variable after exergame interventions. Only one study implemented an acute protocol of active video games finding a positive effect on cognitive flexibility. Concerning the chronic exergame interventions (range 4–24 weeks), favorable effects on global executive functions and inhibitory control were identified. Conclusions: The results of the included studies suggest a favorable acute effect and positive chronic effects of active video game interventions on cognitive variables in healthy children and adolescents. However, these effects appear to be inconclusive given the low number of studies, and the overall methodological quality and risks of bias. Thus, it is necessary to support these findings with future research.


Introduction
Exergames, or active video games, are video games where the person interacts with the console by performing physical activity. Examples of gaming systems that use these types of video games are Xbox Kinect, Nintendo Switch, and PlayStation Move, where the games are usually based on sports or dance activities 1,2 . These consoles interpret the player's body movements, for example, through cameras or hand controls, to interact with the video game. Thus, exergames increase energy expenditure in preadolescents, ranging from light to vigorous intensities, depending on the video game 3 .
In children and adolescents, there are known favorable effects of the use of active video games on health. For example, it has been found that they generate a better physiological and psychological response compared to sedentary activities of the same type (such as sedentary video games or watching television) 4 . Additionally, these physiological and psychological responses are similar to those observed in cycling or walking laboratory exercise tests, or physical activity in the field 4 . In addition, it has been identified that active video games increase energy expenditure compared to inactive video games 5 . At the same time, children and adolescents, after interventions through exergames, show improvements in upper limb strength 6 , physical activity level 7 , maximal oxygen consumption 8 , body composition 9 , and body mass index 8 . Thus, this type of video game seems to be a promising tool to promote health in the pediatric population.
In recent years, there has been a growing interest in studying the effect of exergames on cognitive function. For example, in girls and boys with attention deficit hyperactivity disorder, three weekly 30-min active video game sessions per week for eight weeks improved executive function 10 . Similarly, in children and adolescents with autism, a favorable effect of exergames on cognitive variables was identified 11 . On the other hand, a study in children and adolescent cancer survivors did not find an effect after eight weeks of active video games, compared to cognitive training 12 . Of note, a recent meta-analysis that included participants with neurological disabilities of all ages identified a positive effect of exergames on attention, executive function, and perception 13 . However, these findings are limited by the high heterogeneity of the results for attention and executive function, and the diversity of participant characteristics and study designs included. In the case of the healthy pediatric population, discrepancies have been found in the effect of exergames on cognitive variables, e.g., inhibitory control, reporting both favorable 14 and null effects 15 . Furthermore, to our knowledge, there are no systematic reviews that address the effect of exergames on cognitive variables in healthy children and adolescents.
Thus, although the effect of exergames on cognition has been studied in a young population, reviews have generally focused on participants with neurological disabilities or pathologies 13,16 . Thus, the aim of this systematic review is to examine the effect of exergame interventions on cognitive functions in apparently healthy children and adolescents.

Methods
This systematic review was conducted according to the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines 17 .

Search strategy
PubMed, Scopus, and Web of Science databases were searched using the following combination of keywords: ("active video games" OR exergame OR exergaming OR "video game" OR "computer game" OR "virtual reality" OR "Nintendo Wii" OR "dance dance revolution" OR "computer interaction" OR kinetic) AND (children OR teenagers OR adolescent OR young) AND (cognition OR "cognitive functions" OR "cognitive development" OR "executive functions"). The search was conducted with no date limit until August 2021.

Study selection criteria
Eligible studies had to meet the following Participants, Intervention, Comparator, Outcomes, and Study Design (PICOS) criteria: i) Participants: Healthy children or adolescents, without chronic pathologies or cognitive impairment; ii) Intervention: Acute or chronic exposure to exergames; iii) Comparator: Studies that in addition to the intervention group contain groups that do not perform exergames; iv) Outcomes: Present pre-and post-intervention measurements of some cognitive component, indicating the effect or result; and v) Study Design: Randomized controlled trials. Studies published in English and Spanish were considered.

Search, filtering, and selection of studies
The search in each database, using the combination of keywords previously described . The study selection process began with an initial filter in which the articles were analyzed according to title and abstract. Subsequently, the studies that passed the first filter were analyzed by full text. The reason for exclusion of the studies reviewed by full text was made explicit. Gray literature was excluded during the filtering and article selection process. References of studies that met the inclusion criteria, as well as systematic reviews related to exergames and cognition, were analyzed for possible eligible studies.

Data extraction and synthesis
Three reviewers (A.V.O., J.V.A., D.T.T.), independent and blinded, extracted the following information from the included studies: author, year of publication, study design, participant characteristics (country of origin, age, number of participants, and sex), exergame intervention characteristics (exergame type, session duration, session frequency, intervention duration, intensity), comparator group characteristics (indications or control group exposure), and outcomes (cognitive function assessed, instrument used, intervention results). Disagreements were resolved by a fourth reviewer (J.S.M.). The information collected was synthesized in a descriptive table, sorting the studies according to the duration of the intervention.

Assessment of methodological quality and risk of bias
Three reviewers (A.V.O., J.V.A., D.T.T.), independent and blinded, assessed the methodological quality of the included studies using the ten-point PEDro scale 18 . Disagreements were resolved by a fourth reviewer (J.S.M.). Methodological quality was categorized as high when the study scored six or higher, and as low for articles scoring five or lower 19 .
The risk of bias assessment was performed by three reviewers (A.V.O., J.V.A., D.T.T.), independent and blinded, using the Cochrane Bias Assessment Tool for Randomized Trials (RoB 2) 20 . Disagreements in the assessment were resolved by a fourth reviewer (J.S.M.). This tool assesses the risk of bias in five domains: i) bias from the randomization process; ii) bias to deviations from intended interventions; iii) bias from missing outcome data; iv) bias in outcome measurement; and v) bias from selection of the reported outcome. In addition, an overall risk of bias is derived based on the results of these five dimensions. This tool categorizes the risk of bias into three levels: i) low risk of bias; ii) some risk of bias concerns; and iii) high risk of bias.

Effect size
The effect size for each cognitive variable in each study was calculated using Cohen's d (difference between two means divided by a pooled standard deviation for two independent variables), considering the pre and post data of the intervention and control groups. WebPlotDigitizer software was used to extract and estimate data from figures when the article did not present numerical values 21 . When the article presented results as median, minimum, and maximum, the mean and standard deviation were estimated 22 . In the case of not having data from the control group, the intragroup experimental effect size was calculated. Positive values of the effect size express benefits on the cognitive variable in favor of the experimental group. The cutoffs 0.1, 0.3, 0.5, 0.7, and 0.9 were used for small, moderate, large, very large, and extremely large effect sizes, respectively 23 .

Search, filtering, and selection of studies
The process of searching, filtering and selection of studies is summarized in the flow diagram illustrated in Figure 1. By applying the search strategy in the declared databases, 3,748 articles were identified, and after eliminating duplicates, 2,898 records were analyzed according to title and abstract. After applying this initial filter, 20 studies were analyzed according to full text. The reason for exclusion of the studies analyzed by full text is described in Figure 1. Finally, five studies met the inclusion criteria and were included in the present systematic review 14,15,24,25,26 , comprising nine reports of the included studies considering the experimental groups with identified exergames.

Characteristics of the included studies
The characteristics of the selected studies are summarized in Table 1. Regarding the age of the participants, three of the five included studies were in the age range between 10-19 years 15,25,26 , while two studies were between 7-9 years of age 14,24 . Regarding the sex of the participants, only one study included exclusively males 15 , while the remaining studies included both males and females 14,24,25,26 . Regarding the characteristics of the interventions, only one study evaluated the acute effect of active video games 15 , while the rest involved interventions of four weeks 24,25 , ten-weeks 26 , and 24 weeks 14 of duration. The time range of the exergame sessions was from 10 minutes 24 to 60 minutes 25 , while the frequency of weekly sessions of the chronic interventions ranged from one 14,26 to five days 24,25 . The most commonly used platform for exergaming was Xbox Kinect 14,15,24 followed by Nintendo Wii 25,26 . Regarding the cognitive variable evaluated, most of the studies (3/5) used the Delis-Kaplan Executive Function System (D-KEFS) to measure executive functions, while the rest evaluated the effect of active video games on inhibitory control using the Go/No-Go test or Simon task.

Effect of exergames on cognition in a young healthy population
First, on the results of the studies that evaluated the effect of active video games on executive functions, favorable findings were found. On the one hand, two studies found a favorable effect of exergames on global executive function 25,26 . From these studies, we identified a large effect in a group exposed to exergames (dE post-pre = 0.58) 25 and a very large effect compared to the control group (dE1 vs C = 0.80) 26 . In more detail, only the competitively applied exergame, as opposed to the cooperative one, increased executive function performance 26 . On the other hand, another study identified that, acutely, active video games applied with higher cognitive engagement, as opposed to active video games with lower cognitive engagement and the control group, increased the cognitive flexibility component of executive functions (dE1 vs C = 0.41, moderate effect), but not the fluency and inhibition components 15 .
Secondly, the two studies that evaluated the chronic effect of active video games on inhibitory control found a positive effect. On the one hand, the study by Layne et al. identified that four weeks of 10 min of exergames, five days a week, improves performance in the Go/No-Go test, decreasing reaction time (dE vs C = 1.55, extremely large effect) and enhancing inhibitory response control (dE vs C = 1.21, extremely large effect) 24 . On the other hand, the study by Šlosar et al. found that one and two weekly sessions of exergames in addition to tennis training for six months decreased reaction time in congruent (small and trivial effects, respectively) and incongruent tasks of inhibitory control (small and large effects, respectively), although with no difference in response accuracy 14 .

Assessment of methodological quality and risk of bias
The result of the evaluation of the methodological quality of the included studies is described in Table 2. The average PEDro scale score was 5.0  1.0 points. Three of the five included studies had a low methodological quality 15,24,25 . The remaining studies were of high quality, although at the lower limit 14,26 . The criterion which all the studies met was randomization, while the items that none met were concealed allocation, and blinding of therapists and assessors.   Figure 2 summarizes the results of the risk of bias assessment of the included studies. In relation to overall bias, two studies presented high risk of bias 14,25 , two studies showed some concerns 24,26 , and only one study achieved a low overall risk of bias, as well as in all domains 15 . In particular, all studies achieved a low risk in the domains of missing outcome data and outcome measurement 14,15,24,25,26 .
In relation to the bias of the randomization process, four of the five studies showed some concerns 14,24,25,26 , mainly for not making the randomization mechanism explicit. Only one study had a high risk of bias in the deviation from the intended interventions 25 and two studies a high risk of bias in the selection of the reported outcome 14,25 .

Discussion
The aim of this review was to investigate the literature on the effect of exergames on cognitive functions in healthy children and adolescents. Based on the five studies that met the inclusion criteria, we identified that there is evidence, albeit with limitations, of a favorable acute effect after an exergame intervention on cognitive flexibility, and a chronic effect of exergames on inhibitory control and global executive functions.
The effects of exergames on cognition in healthy children and adolescents appear to be similar to those reported in systematic reviews that included children with neurological disabilities 13 and healthy adolescents 27 . However, the comparison of the findings of these studies with our results is limited, because in one review only two of the 13 included studies were in a young population 13 , while in the other review only one of the 17 included studies was in adolescents 27 . We did not identify other systematic reviews including children and adolescents that evaluate the effect of exergames on any of the cognitive variables we identified (e.g., executive function). Therefore, it is difficult to compare our findings with previous reviews, considering the limited number of studies on the subject.
The effect of exergames on cognition would be justified based on the aerobic exercise and the cognitive demand of this intervention. The acute and chronic effects of aerobic exercise on executive functions and attention in preadolescents is well known 28 . Moreover, sports or curricular exercise classes improve general executive functions and inhibitory control 29 . Several mechanisms have been proposed to support cognitive improvements after exercise, such as increased cerebral vascularization and blood flow, increased neurotrophic factors, gray matter, and functional connectivity 30 . In particular, postexercise improvements in executive functions, such as inhibitory control and cognitive flexibility, are related to an increase in volume and activation of the prefrontal area 31,32,33,34 . In addition, a synergistic effect between exercise and cognitive training on cognitive functions or dual tasks has been reported 35,36 . Studies in older adults have found that interventions with active video games improve the performance of executive functions and global cognition, as well as a greater efficiency of prefrontal area activation 37,38 . However, to our knowledge there are no studies that evaluate the effect of exergames on brain structure and function in children and adolescents, so future studies should elucidate this unknown.
Some characteristics of exergame interventions appear to impact cognition in a young healthy population. For example, the included studies comparing different exergame protocols found that the more cognitively engaged 15 , more physically active 25 and a competitive modality 26 showed greater cognitive improvements in healthy children and adolescents. Furthermore, the study by Šlosar et al. identified that virtual tennis exergame interventions, in addition to traditional tennis training, improved inhibitory control compared to tennis-only sessions 14 . This suggests that adding exergame sessions could enhance the effect of traditional sports or exercise interventions. Thus, considering the synergistic effect between exercise and cognitively demanding activities, exergames seem to be an alternative to stimulate cognition in healthy children and adolescents. However, there are characteristics of exergame interventions, such as exercise intensity, where dosage is not clear, because none of the included studies objectively dosed or monitored exercise intensity. Some meta-analyses have found exercise intensity to be a moderator of cognitive effects 39,40 . Therefore, future studies should objectify and analyze the role of exercise intensity on cognition in exergame interventions. Finally, the limited number of studies published on this topic should be considered, so more evidence is needed to elucidate the dose-response of exergames, and which characteristics of the intervention are the most appropriate to enhance the effect on cognition.
Considering the findings of exergaming interventions on cognition in healthy children and adolescents, this intervention could have a transfer to the school context and academic performance. For example, one study identified that in adolescents with developmental coordination disorders, eight weeks of gaming using a Wii Fit console (30-minute sessions, three times per week) after school or at recess improved executive functions compared to the control group 41 . Similarly, four weeks of active video games on weekends increased executive functions in preschoolers compared to traditional physical activity 42 . Regarding academics, it has been identified that the application of exergames during the school year improves academic performance in mathematics 43 . Finally, the high degree of acceptance and enjoyment reported when using exergames in children and adolescents should be considered 44,45 , as well as the favorable effect on the level of physical activity 44 and physical capacity 46 , taking into account that this population has low levels of physical activity 47,48 , and that they perceived a greater decrease during the COVID-19 pandemic 49 . Thus, although the evidence is scarce, it seems that exergames could be considered as a complementary activity to physical education classes, as well as at home, to improve cognition, academic performance, and physical activity levels.
The included studies presented, in general, a low methodological quality and a risk of bias with considerations or high, limiting the conclusions from the results. First, the nature of the design of the included studies should be considered, making it impossible to achieve a higher methodological quality score. For example, compliance with blinding of both participants and therapists is difficult in this these types of field studies under natural conditions, given the knowledge of the intervention applied. However, to improve methodological quality, future studies should strive to meet criteria such as the presentation of participants' results, statistical comparison, and variability statistics, among others. On the other hand, because this review included exclusively randomized trials, this criterion was met by all studies on the PEDro scale. However, when assessing the bias of the randomization process, it was identified that most of the studies did not make the procedure explicit, which raises some concerns. All the above, added to the other domains of methodological quality and risk of bias, should be considered when interpreting the results of the studies and the conclusions generated in the present review, as it would also help future studies on this subject to improve the quality of the evidence.

Limitations and Strengths
On the one hand, the present review has as a strength the application of the recommendations of the PRISMA statement for reporting the systematic review. In addition, it is noteworthy that the processes of filtering, selection, data extraction, assessment of methodological quality and risk of bias were performed by three reviewers, independent and blinded, and disagreements resolved by a fourth reviewer. On the other hand, this systematic review is not without limitations. One of them is the number of databases included, which was limited by the availability of access. It is known that the number of databases can modify the conclusions of reviews and meta-analyses 50 . Therefore, the conclusions of this review should be interpreted considering this limitation. Another limitation was the impossibility of performing a meta-analysis due to the limited number of cognitive variables in common among the included studies.
In addition to the above, the variability of tools used to assess different cognitive variables should be considered, which could limit the comparison of findings between studies and conclusions. Furthermore, the limited number of included studies limits the interpretation of the characteristics of the exergame intervention (intensity, duration, among others) on cognition in this population. Thus, future studies comparing the effect of different exergame protocols (duration, intensity, modality, etc.) on cognition in healthy children and adolescents are needed to elucidate the dose-response of this intervention. To complement the above, future studies should evaluate the effect of exergames on learning, school performance, and behavior, to consider their application in the educational context. Finally, considering the methodological quality and risk of bias identified in the included studies, the results of this review should be interpreted with caution. Therefore, future studies on this topic should consider increasing the methodological quality and decreasing the risk of bias, based on the domains evaluated in this review, to obtain more reliable conclusions on the effect of active video games on cognition in healthy children and adolescents.

Conclusions
This systematic review identified a limited number of studies evaluating the effect of exergames on cognition in healthy children and adolescents. Based on these studies, there is evidence that exergames generate favorable acute and chronic effects on executive functions, although in general the studies present a low methodological quality and high risk of bias or with considerations. We recommend increasing the methodological quality of future research evaluating the effect of exergames on cognitive variables in a young population to corroborate the findings.