The non-fitness-related benefits of exergames for young individuals diagnosed with autism spectrum disorder: A systematic review

Background: Autism Spectrum Disorder (ASD) is a common neurodevelopmental condition characterized by social-communication deficits and is often accompanied by an array of autism-specific traits, including cognitive and motor differences. Research has begun to investigate the utility of exergames (virtual exercise games) for combating some of the functionally impairing facets of ASD. However, no review to date has investigated the positive effects of exergaming for young autistic individuals, exclusive of their effects on fitness-related/health-related outcomes. Methods: We systematically reviewed the literature for articles published between 1990 and 2020, which aimed to examine the effectiveness of exergames for targeting some of the facets observed in autistic children, exclusive of physical fitness-related outcomes. Results: Out of 193 relevant articles, 10 met the inclusion criteria. Some studies showed initial beneficial effects of exergaming for executive function and motor skills, however more work is needed to examine the beneficial effects of exergaming for social development in young autistic individuals. Conclusion: The present review suggests the potential use of exergaming for enhancing some functions in ASD and future research is needed to expand the use of exergames for supporting social-communication skills. Furthermore, additional studies using rigorous research designs are needed to draw stronger conclusions about the effects of exergames for the young autistic population.


Autism
Autism Spectrum Disorder (ASD) 2 is a long-term neurodevelopmental condition accompanied by a markedly focused repertoire of populations, with online gaming increasing from childhood to adolescence in autistic individuals (Engelhardt, Mazurek, & Hilgard, 2017;Mazurek & Wenstrup, 2013;Mazurek, Shattuck, Wagner, & Cooper, 2012).It is thought that videogames are particularly attractive to autistic individuals as they often offer low social demand, are consistent and predictable, and allow the user/player to determine the pace of the activity, taking active control (Durkin, 2010;Engelhardt et al., 2017).An innovative solution to increase PA in sedentary populations and draw up on the interests of videogaming platforms in many, can be found in the application of 'exergaming'.The term 'exergame' has many definitions associated with it.For example, Bogost (2007) defines exergames as "the combination of exercise and video games" (pg.294), whereas as Oh and Yang (2010) redefined exergames as "an experiential activity in which playing exergames or any videogames requires physical exertion or movements that are more than sedentary activities and also include strength, balance, and flexibility activities".More recently, Kappen, Mirza-Babaei, and Nacke (2019) defined exergames as "the combination of gaming technologies and exercise routines to motivate physical activity among individuals or groups" (pg.142).For the purpose of this review, we will be drawing upon Kappen et al.'s (2019) definition, where exergames are the combination of gaming technologies and exercise routines encouraging physical activity in autistic children.Exergames; combining videogaming technology with exercise, could offer the perfect platform to increase PA in young autistic individuals.Furthermore, the application of exergames may harness the potential therapeutic benefits of PA for reducing some of the features associated with ASD that impact Fig. 1.PRISMA diagram for the flow of information through the different phases of the systematic review (Moher et al., 2009).

Table 2
Characteristics of studies that met all the inclusion criteria, including study design, sample size, exergame platform used for the intervention and duration of intervention, outcome measures and findings.quality of life, whilst also being engaging and enjoyable for many autistic individuals.Due to the vast array of available exergames previous research has indicated that exergames can be effective interventions for supporting an array of developmental skills, including social, cognitive, and motor development, and also rehabilitation in many different populations and clinical samples.For example, exergames have been observed to have some positive effects for self-concept, situational interest and motivation, enjoyment, psychological and social well-being, symptomatology and different learning experiences for TD children and adolescents (see Joronen, Aikasalo, & Suvitie, 2017 for review).In addition, children and adults with cerebral palsy have shown positive responses to participating in Wii Sports, which involves performing the actions associated with tennis, baseball, bowling, golf and boxing, and Wii Fitness (Deutsch, Borbely, Filler, Huhn, & Guarrera-Bowlby, 2008;Jelsma, Pronk, Ferguson, & Jelsma-Smit, 2013).Similarly, individuals with down syndrome have also shown significant improvements in physical fitness and functional outcomes (Perrot, Maillot, Le Foulon, & Rebillat, 2020) and motor proficiencies after participating in Wii-based exergames (Silva et al., 2017).Furthermore, exergaming was an effective intervention for temporarily enhancing balance and reducing fatigue in individuals diagnosed with Parkinson's disease after 12 weeks of Wii Fit games (Ribas, Alves da Silva, Corrêa, Teive, & Valderramas, 2017).Kappen et al. (2019) provides a good overview of the diversity and use of exergames for older adults.Therefore, due to the rapid development of computer technology, promising results in various clinical populations, and young autistic individual's interest in electronic media (i.e., videogames), more research is needed to better understand and establish the benefits of exergames for ASD.
Many studies have focused on the impact of exergames for increasing PA levels in autistic individuals, in an attempt to decrease sedentary behaviors and improve fitness-related outcomes such as heart rate and calories burned (Fang, Aiken, Fang, & Pan, 2018;Finkelstein, Barnes, Wartell, & Suma, 2013;Finkelstein, Nickel, Barnes, & Suma, 2010;Hilton et al., 2015;Jozkowski, Lichtenwalner, & Cermak, 2016).Whilst these studies have shown promising results for fitness-related outcomes, they do not account for how exergames could influence any of the well-recognized facets associated with ASD.Therefore, some studies have since focused their efforts on examining the effect of exergames for some of the specific domains that are commonly affected in ASD, such as motor skills, cognitive function or social capabilities.However, no review to date has acknowledged and highlighted the potential beneficial effects of exergames for alleviating the impairing symptomology associated with the ASD, exclusive of PA levels/fitness-related outcomes.
Therefore, the purpose of this systematic review is two-fold.First, we aim to pool all the existing research relating to the nonfitness/non-physical-health-related effects of exergames; such as cognitive, social, or motor functions, for young individuals diagnosed with autism.Secondly, we aim to collate the available evidence that suggests exergames are a viable vehicle for therapeutic intervention targeting some of the limiting differences observed in ASD.In addition, we hope to provide a direction for future research to develop tailored exergame-based interventions for young autistic individuals that will support the development of communication skills, a core facet of ASD, in a person-centered manner; promoting quality of life and well-being.

Research question
To provide a summary of the current evidence relevant to the non-fitness-related impacts of exergames for young autistic individuals and to investigate whether exergames can elicit positive effects for autistic individuals, we utilized the PICO strategy to devise a central research question.PICO is a widely recognized model often used for structuring research questions in connection with evidence syntheses, where: P: population, patient or problem; I: intervention, C: control or comparative intervention, and O: outcome (Eriksen & Frandsen, 2018).Using the PICO framework, the following research question was generated; "To what extent can exergames have non-fitness-specific beneficial effects for young individuals diagnosed with Autism Spectrum Disorder?".We appreciate that some studies may use varying study designs; including, within-subject study designs, comparing the effects of an exergame by pre-and post-intervention tests and an initial review of the literature suggested that there were limited studies available for review that included a comparison group.Therefore, we elected to not specify the inclusion of a control group or comparative intervention in our research question; considering the limitations this begets in the discussion.

Inclusion and exclusion criteria
Studies were included in the review if they met the inclusion criteria.For example, studies were included if they were in English, published between 1990 and 2020, included participants aged between 5 and 18 years old, where most participants had a diagnosis of ASD, used an exergame as an intervention, and measured the effect of the intervention with reported outcome measures that focused on skills such as social, cognitive or motor proficiencies.Studies were excluded if the papers were not in English, participants were older than 18 years old, and/or if the outcome measures focused only on PA levels or fitness-related outcomes; for example, heart rate, P.O.Morris et al. calories burned or moderate-to-vigorous PA levels, as this was not the main focus of the systematic review.Additionally, conference papers and review articles were excluded (Table 1).

Selection of studies and extraction of data
Following the PRISMA guidelines for systematic reviews, an orderly procedure was carried out to select the most relevant articles to include in the review from the literature search (Moher, Liberati, Tetzlaff, Altman, & Group, 2009).During the first phase of the procedure, we combined all extracted articles from the literature search into a single folder, removing any duplicates.Adhering to the inclusion and exclusion criteria, we selected articles to exclude based on their titles and then articles to exclude based on their abstracts.Using the 'Dual Independent Review Approach' of the search results, the first author and an independent reviewer (TM) both screened the titles and abstracts to ensure reproducibility in the inclusion and exclusion criteria (Stoll et al., 2019).The remaining articles were all read fully and brief notes were made on each of the study's characteristics, including title and author, study design, participants used, intervention utilized, and outcome measures, whilst still assessing them against the inclusion and exclusion checklists (Table 1).Additionally, the reference lists of relevant articles were read, seeking studies that were appropriate for inclusion but had been missed by the search parameters (n = 3).Papers that met all the inclusion criteria and were relevant to the central research question were all read fully again in order to finalize articles to include for evidence synthesis (Fig. 1).
Once the final studies to include in the review had been selected, data was extracted to retrieve particular information regarding each study, including; title, authors and date, study sample and size, exergame platform used, outcome measures utilized, findings, and conclusions.

Assessment of risk of bias for included studies
As reported in the Cochrane handbook, bias "is a systematic error, or deviation from the truth, in results or inferences" (Boutron et al., 2019), where different biases can lead to an over or underestimation of the true effect of an intervention.It is important to assess bias in all studies included in the review to ensure accurate conclusions can confidently be made from the synthesis of the available literature.However, it is often impossible to know the extent that biases have affected the results of a particular study.Therefore, it is more appropriate to reflect on the 'risk' of bias of a particular study.We employed the Cochrane risk assessment guidelines to determine whether the risk of bias was low, high, unclear or not applicable (if non-randomized study) for selection bias (random sequence generation and allocation concealment), performance bias (blinding of participants and personnel), detection bias (blinding of outcome assessment), attrition bias (incomplete outcome data), reporting bias (selective reporting), and any other potential bias.

Study selection
The initial literature search identified 193 studies that met some of the key terms relating to the research question (Scopus, n = 57; Web of Science, n = 29; and EBSCOhost, n = 107).An additional three studies were identified by examining the reference list of relevant papers.After removing duplicates, screening papers based on titles, and then screening based on abstracts, 15 potential papers remained.To this end, the level of agreement between the two reviewers, who both screened the articles by titles and abstracts, was calculated using Cohen's Kappa (McHugh, 2012).The weighted kappa score for agreement between the reviewers was 0.797, 95% Cl: 0.626-0.969,indicating substantial agreement (Landis & Koch, 1977).

Characteristics of participants
As this review focuses on the effects of exergaming for young autistic individuals, we chose not to include studies that had participants over the age of 18 years old.All participants were aged between 5 and 18 years old, with one exception.Anderson-Hanley et al.'s (2011) pilot study II included individuals aged between 8 and 21 years old.We have included this pilot study in the table of characteristics, however will not be drawing conclusions from this part of their study due to the participant's ages.The average age of participants included in the review, excluding Lau et al.'s (2020) study as they did not report mean age, was 9.56 years old.
Most participants had a diagnosis of ASD, according to the Diagnostic and Statistical Manual of Mental Disorders IV or V (DSM-IV; DSM-V; American Psychiatric Association, 2013).The Autism Diagnostic Observation Schedule-Second Edition (ADOS-2; Lord et al., 2012) and the Gilliam Autism Rating Scale-Second Edition (GARS-2; Gilliam, 2005) were also used to assess autism severity and P.O.Morris et al. confirm a diagnosis.Some studies, such as Lau et al. (2020) and Flynn and Colon (2016) included participants that had various difficulties or intellectual disabilities, but where the majority of participants had an autism diagnosis.Additionally, typically developing (TD) participants were also included in Edwards et al. (2017) for between-group comparisons.

Characteristics of interventions
Interventions varied in duration.Some studies concluded within a day, whilst others lasted 12 weeks.Flynn and Colon (2016) and Anderson-Hanley et al. (2011) both used a single, 20-minute intervention session to assess the effect of their chosen dance-based exergame on participant's EF, repetitive behaviors, and cognition.In comparison, Hilton et al. (2014) andRafiei Milajerdi et al. (2020) conducted their interventions across 8 weeks, in which participants were required to complete several sessions of their chosen exergame; Makoto Arena and Kinect, respectively, each week.Chung et al. (2015) and Lau et al. (2020) both utilized the Kinect platform, requiring participants to complete up to 3 exergame sessions a week, for 12 weeks.However, Chung et al. (2015) directed autistic children to play the Kinect-based exergame with their TD sibling in the home environment, whilst Lau et al. (2020) asked autistic children to be 'paired-up' by their teachers and complete the intervention in a school environment.Across all studies, the average intervention duration was 6 weeks and the most popular platform utilized was Kinect, followed equally by Makoto Arena and Wii.The most popular type of active video games played were dance-based games, such as Wii's Just Dance or Kinect's Dance Dance Revolution.However, sports-based games such as Sports Season 1 and 2 or Kinect Tennis were also played frequently.

Characteristics of outcome measures used
The studies included in this review examined a variety of potentially beneficial outcomes of exergames for young autistic individuals including social, cognitive, and motor benefits.They did not, however, focus on or investigate the fitness-related benefits of exergames, as this was not the main aim of the review.Overall, 7 of the studies investigated the effect of exergames on cognition, most frequently EF, 7 of the studies examined motor function, and only 2 of the studies explored social behaviors.It is therefore apparent that some studies investigated the effects of exergames on both cognition and motor skills (Anderson-Hanley et al., 2011;Caro et al., 2017;Collins et al., 2015;Hilton et al., 2014).
To assess EF, tasks such as the Stroop Task (Stroop, 1935), Wisconsin Card Sorting Task (WCST; Grant & Berg, 1948), and the Digit Span Backwards Task were used (see Ramsay & Reynolds, 1995 for an overview).They were often administered pre-and post-intervention.To examine the effect of exergames on motor skills and motor proficiencies most studies enlisted the Bruininks-Oseretsky Test of Motor Proficiency-Second Edition (BOT-2; Bruininks & Bruininks, 2005), however one study used the Movement Assessment Battery for Children-Second Edition (MABC-2; Henderson et al., 2007) and another used the Test of Gross Motor Development-Third Edition (TGMD-3; Ulrich, 2016).When exploring the effect of exergames on social behaviors, Pope et al. (2019) and Chung et al. (2015) both videotaped their participants during intervention sessions, subsequently observing the tapes to score behaviors according to a coding manual and to assess changes in social behaviors.

Executive function
Overall, some studies found within this review reported that exergaming produced certain benefits for young autistic individuals.Collins et al. (2015), Flynn and Colon (2016), Hilton et al. (2014), andRafiei Milajerdi et al. (2020) all noted improvements in EF following their exergame-based interventions.Both Collins et al. (2015) and Hilton et al. (2014) utilized the Behavior Rating Inventory of Executive Function (BRIEF) to measure changes in EF and both observed improvements in all areas of the inventory.Collins et al. (2015) reported a significant improvement in global executive composite (P < .05)and metacognition index (P < .01)and Hilton et al. (2014), reported a significant improvement in the specific areas of working memory (P < .05)and global executive composite (P < .05).Rafiei Milajerdi et al. (2020) used the WCST as a measure of EF.They showed that on average the exergaming group were able to significantly improve their correct responses on the WCST, between pre-and post-test from 0.50 ± 0.71-1.52 ± 2.00, which was greater than both the control and SPARK (Sports, Play and Active Recreation for Kids) groups included within the study (P < .05);thus, suggesting an improvement in set-shifting and mental flexibility due to exergaming.

Motor Skills
Caro et al. ( 2017), Collins et al. (2015), and Hilton et al. (2014) all reported an improvement in motor skills.Caro et al. ( 2017) demonstrated that 'FroggyBobby' was able to increase the percentage of time participants spent performing aimed limb movements from 56% in the second session to over 96% in the twelfth session, suggesting a significant improvement in motor coordination (P < .05).Collins et al. (2015) compared the means of pre-and post-test scores for the BOT-2 during their study, with results suggesting an improvement in all areas of the BOT-2 and a significant improvement in the motor area of strength and agility (p < .01)with a large effect size (r = .54).Similarly, Hilton et al. (2014) also noted a significant improvement in the motor area of strength and agility in the BOT-2 scores following their 'Makoto Arena' intervention (P < .05),with a small to medium effect size (d =.46).Edwards et al. (2017) suggested that during their study exergames were unable to produce an improvement in actual motor skills but instead were able to increase the participant's perceived object-motor skills by approximately 2 units, from an average score of 27.36 ± 3.85-29.45± 3.91, on the Pictorial Scale of Perceived Movement Skill Competence (P < .05),which in future may have led to an improvement in actual motor skills.
In comparison, Lau et al. (2020) reported that exergames were unable to improve motor proficiencies, body composition or PA levels over and above that of their control group during their study, as no significant improvements were observed between the P.O.Morris et al. intervention and control group.Despite children's BOT-2 scores increasing from 53.85 ± 15.25 at pre-test to 57.36 ± 15.00 at post-test in the intervention group and only increasing by two units from pre-test to post-test for the control group, after adjustment for the intervention group relative to the control group, differences for motor proficiency were found to be non-significant between groups.

Social skills
Chung et al. (2015) noted no substantial improvement in social behaviors, including positive joint affect, reciprocal communication, or aggression within their participants, following their active videogame intervention in comparison to the sedentary videogame condition.Pope et al. (2019) reported that exergames, mostly 'Just Dance 4', were able to promote an increase in PA levels, on-task behaviors, and communication.However, their comparison condition, Adaptive Physical Education (AdPE), was able to achieve this on a greater scale.For example, percentage of time spent communicating during the AdPE condition was 34.0% ± 11.8 compared to 16.1% ± 10.1 in the exergaming condition.

Enjoyment
In addition to positive outcomes for the development of specific skills, many of the studies included in this review reported that children enjoyed participating in their exergaming interventions.Both Flynn and Colon (2016) and Collins et al. (2015) measured participant's responses to participating in their exergaming interventions via questionnaires.Collins et al. reported that 88% of their participants experienced some degree of enjoyment from playing the 'Makoto Arena' exergame and Flynn and Colon (2016) reported that on average all participants enjoyed their exergaming experience.Similarly, Chung et al. (2015) received parent feedback Fig. 2. Risk of bias summary for each study that met all the inclusion criteria, including high (red), low (green), unclear risk (yellow) of bias and not applicable (black).
following their exergaming intervention inferring that the 'Kinect' games were a positive and enjoyable experience, which was engaging and social for their children.Likewise, Caro et al. (2017) suggested that 99% of the emotions displayed by participants, whilst playing the exergame, were positive emotions.

Risk of bias
Risk of bias, including low, high, and unclear risk, were reported for all studies according to selection, performance, detection, attrition and reporting bias (Fig. 2).However, the risk level was not applicable for all categories in each study.For example, studies that did not include two different groups (i.e intervention vs. control), did not need to randomly assign participants to different groups.Therefore, selection bias, with regards to random sequence generation, was not applicable for these studies.
Most studies were unable to blind participants due to the nature of the interventions; however, it was believed this would not affect the outcomes of the studies.Additionally, most studies did not blind personnel during the outcome assessments, which may have led to detection bias.Authors that accounted for this either by having multiple coders and assessors and reporting interrater reliability for their outcome measures were regarded as having low detection bias.However, studies that did not reference how they attempted to limit detection bias or failed to note how detection bias may have impacted their results were regarded as having an unclear or high risk of bias.In addition, the authors of the Caro et al. (2017) study were involved in the development of the exergame used in their deployment study and therefore were noted as potentially having a high risk of bias for 'other biases'.

Discussion
In answer to the research question "To what extent can exergames have non-fitness-specific beneficial effects for young individuals diagnosed with Autism Spectrum Disorder?" the results suggest that interventions based on exergames can produce some beneficial effects for young autistic individuals.These included improving EF, increasing motor proficiencies and, to some extent, promoting communication owing to the diversity of exergames and skills needed to play various exergames.Additionally, results from the studies included within this review suggest that participants enjoyed the exergames, which is a key component to helping autistic individuals engage with interventions (Finke Erinn et al., 2018).Flynn and Colon (2016), Collins et al. (2015), and Chung et al. (2015) reported that children enjoyed participating in their exergaming interventions, either via self-reported questionnaires or feedback from parents.Importantly, no study reported any adverse consequences for children participating in their exergaming-based intervention.

Implications for practice and future work
Online media such as videogames and exergames are believed to be intrinsically motivating for many autistic individuals and, falling in line with the available literature, the results from this review suggest that exergames are engaging, motivating, and enjoyable for young autistic individuals (Caro, Martínez-García, & Kurniawan, 2020;Finkelstein et al., 2010;Lyons, 2015).Furthermore, the availability of many different exergames that promote varying beneficial effects and involve different skills allow children to choose what games they want to play and have autonomy over their choices.It is thought motivation is essential for learning and is a key component in engaging with and committing to learn a new activity.Further, it is argued that the most optimal context for development is one that is selected based on the motivation and interests of the individual (Deci & Ryan, 2000;Koegel, 2000;Gee, 2009).Thus, illuminating why exergames may elicit positive changes in executive function and motor skills, and highlighting the strong potential exergames have for developing further skills, such as self-regulation, cooperative play, or communication skills, in young autistic individuals.
Recognizing that autistic individuals contend with various barriers throughout their day-to-day lives, challenges presented within videogames may be more manageable, enjoyable, and rewarding than the tasks presented in their everyday lives (Durkin, Boyle, Hunter, & Conti-Ramsden, 2013).Previous studies have indicated that individuals on the autistic spectrum play videogames for a variety of reasons, including emotional regulation, immersion, social interaction, achievement, creativity, and mental stimulation (Finke Erinn, Hickerson Benjamin, & Kremkow Jennifer, 2018;Mazurek, Engelhardt, & Clark, 2015).Given that exergames are observed to elicit some beneficial effects, are intrinsically motivating, and enjoyable for autistic individuals, it seems plausible that exergames can offer a viable platform for developing an array of skills in ASD in a therapeutic capacity, which is well-suited to the individual's interests.Exergames could be implemented into practice, with future work confirming the optimal duration and frequency of exergaming interventions and the ideal games to play for targeting specific domains of ASD.Further, exergames could readily and regularly be integrated into or in addition to schools' physical education programs to improve EF and motor skills in autistic children, following more rigorous studies confirming their benefits.
Despite poor social and communication skills being a core diagnostic trait of ASD and having a great impact on quality of life (QoL; Chiang & Wineman, 2014;Kuhlthau et al., 2010), only two studies were found by the systematic search to investigate the effect of exergaming on social behaviors in young autistic individuals.Both Chung et al. (2015) and Pope et al. (2019) observed communication and social behaviors during their exergaming conditions.Chung et al. (2015) reported no significant changes to positive joint affect, reciprocal communication, or aggression during their intervention condition; however, this study only included 3 participants who had an autism diagnosis in dyads with their TD siblings.Pope et al. (2019) noted an improvement in on-task behavior and communication in autistic participants (n = 5), however their AdPE condition produced a greater improvement in both outcome measures.As these were both observational studies with limited sample sizes, it may be beneficial to complete a pre-and post-intervention study design with a larger sample of autistic individuals; testing social skills using standardized measures before and after the intervention in order P.O.Morris et al. for stronger conclusions to be drawn from the findings.
It is noted that many autistic participants had a high affinity for dance-based exergames.Dance is a type of PA that emphasizes the importance of movement and fitness in a variety of ways but also allows one to express their feelings and emotion, communicating through the use of rhythm and body positions (Calvo, D'Mello, Gratch, & Kappas, 2015;Camurri, Lagerlöf, & Volpe, 2003).The use of dance has been extended to therapeutic practices, where dance movement therapy (DMT) is employed for a variety of clinical populations, including those with depression and anxiety (Adam, Ramli, & Shahar, 2016;Karkou, Aithal, Zubala, & Meekums, 2019), Parkinson's disease (de Natale et al., 2017;dos Santos Delabary, Komeroski, Monteiro, Costa, & Haas, 2018), and autism (Scharoun, Reinders, Bryden, & Fletcher, 2014;Takahashi, Matsushima, & Kato, 2019).Importantly DMT has been observed to improve communication skills in young autistic individuals, with dance and movement therapists reporting promising results through the successful use of mirroring, synchronous movement interaction, and rhythm (Cozolino, 2014;Devereaux, 2012;Martin, 2014;Tortora, 2005;Field et al., 2001;Koch et al., 2015).The techniques of mirroring and rhythm have consistently been observed to improve communication skills and social development in autistic children (Morris, Hope, Foulsham, & Mills, 2021).Both of these elements are found within dance-based exergames such as 'Just Dance' (Wii), which requires the player(s) to rhythmically mirror/copy the avatar on screen.Due to this cross over between exergames and successful elements of DMT, it seems plausible that dance-based exergames may still be able to improve social skills, despite the previous findings from Pope et al. (2019) and Chung et al. (2015).More recently, 'Just Dance' as compared to another almost identical movement-based intervention has been observed to produce significant positive psychological and psychosocial effects in a large sample of TD primary school children (n = 417; Quintas, Bustamante, Pradas, & Castellar, 2020).Therefore, future work should attempt to evaluate the beneficial effects of dance-based exergames on the social skills of young autistic individuals utilizing randomized controlled trials.Researchers could recruit an appropriate number of participants to allow for pre-and post-test comparisons of outcome measures between an intervention group (dance-based exergame) and a control group, where validated outcome measures relate to social skills; such as the Social Responsiveness Scale (Constantino, 2013), Social Skills Improvement Scale (Gresham & Stephen, 2007), the Emotion Regulation and Social Skills Questionnaire (ERSSQ; Beaumont & Sofronoff, 2008) and/or coded observations in a given context (Heimann, Laberg, & Nordoen, 2006;Nadel et al., 2000;White, Keonig, & Scahill, 2007).
The findings from this review suggest an exergame intervention based on a timeframe of 6-weeks or greater, with participants completing the exergame for 2-to-3 times per week may be an optimal starting point for future studies.If the results from such a study demonstrated beneficial effects, for example an improvement in the ERSSQ, which was specifically designed for use in the autistic population, exergames could be more confidently integrated into practice for the development of social skills in the autistic population.

Limitations
Some of the studies included in this review presented encouraging results for improving or maintaining specific domains affected within ASD.However, they are not without their limitations; restricting the conclusions drawn regarding the effectiveness of exergames for autistic individuals.Out of ten studies, only two included a control group; the first was a randomized controlled trial (Rafiei Milajerdi et al., 2020) and the second was a two-arm, single-blinded trial (Lau et al., 2020).One other study included an active comparison task for children to also complete (Pope et al., 2019).Although non-randomized controlled trials are useful for feasibility studies and still allow for the critical examination and analysis of the available evidence, it is somewhat difficult to draw conclusions that the intervention generates the observed outcome in studies that do not include control or comparison groups (Gerstein, McMurray, & Holman, 2019).Additionally, the absence of a control group prevents comparison of the experimental group to the changes that might be seen due to normal development, on-going therapies, or familiarity with the assessments (Collins et al., 2015).For example, whilst Caro et al. (2017) suggested that attention and enjoyability was maintained throughout their exergaming intervention, there is no indication or comparison available of what attention and enjoyability levels were prior to participating in the game or during a different intervention.To overcome this limitation and to confidently state exergames can elicit positive effects in young autistic individuals for a variety of traits, future studies need to include a control or active comparison group.
Small sample sizes were another reoccurring limitation across the studies reviewed.Recruiting participants with ASD can present potential barriers.For example, when data is collected in schools there often needs to be specific student-to-teacher ratios, and so, if limited teachers are available this can also limit the number of children available to participate in the study.Additionally, families with autistic children already face an array of challenges and require additional support, therefore participating in scientific research on top of educational, social, and therapy-based commitments may prove too much, and again limit the number of young autistic individuals available to participate.The overall sample size of autistic participants ranged from n = 3 to n = 125.Over half the studies included in this review had less than 17 participants overall (Anderson-Hanley et al., 2011;Caro et al., 2017;Chung et al., 2015;Collins et al., 2015;Hilton et al., 2014;Pope et al., 2019).However, most of these studies utilized a within-subject design, which is more practical than randomly assigning participants to an intervention type.This study design also has greater statistical power than a between-subject design when working with a small sample size (Bellemare, Bissonnette, & Kröger, 2014;Hilton et al., 2014;Thompson & Campbell, 2004).Only one study reportedly carried out a power analysis to determine the sample size needed for their analyses.Rafiei Milajerdi et al. (2020) concluded that with three groups at an 80% power with an a = 0.05, 60 participants would be sufficient to detect medium group, time, and group x time effects.To more robustly assess the effectiveness of exergames as a therapeutic intervention for individuals who have a diagnosis of ASD, it would be beneficial for future studies to conduct experiments with larger sample sizes, using power analyses to determine sufficient sample sizes, which will allow for the inclusion of control groups and between-group comparisons.
No study included in this review chose to conduct a follow-up study.Despite some interventions lasting up to 12-weeks and P.O.Morris et al. producing beneficial effects, it was not clear whether these beneficial effects continued after the intervention ceased.Therefore, it is somewhat difficult to suggest long-term and applied utility of exergaming for ASD.Previous research has indicated that participation in PA interventions is able to produce long-lasting beneficial effects for autistic individuals (Gabriels, Pan, Guérin, Dechant, & Mesibov, 2018;Movahedi, Bahrami, Marandi, & Abedi, 2013).For example, Nicholson, Kehle, Bray, and Heest (2011) demonstrated that antecedent PA was able to improve academic engagement four weeks post-intervention and Bahrami, Movahedi, Marandi, and Abedi (2012) described behavioral gains that were maintained long-term (1 month), following a 14-week martial arts-based training program in autistic children.As a result, future work may wish to explore the maintenance of improvements following exergaming by performing a follow-up assessment.

Conclusion
This systematic review collates the available evidence that suggests exergames may be a viable vehicle to elicit some beneficial effects in young autistic individuals, exclusive of PA levels/fitness-related outcomes.Research suggests that exergames may reduce repetitive behaviors, improve EF, and increase motor skills to some extent.Future work should strive to determine the effects of exergames on social skills in young autistic individuals.Furthermore, prospective work is needed to overcome many of the limitations highlighted by this review; for example, establishing the optimal duration and frequency of exergames or whether positive results are maintained post-intervention.Caution is warranted to not overstate the benefits of exergames for the autistic population without the addition of robust studies to the literature, such as adequately powered randomized control trials to confirm the potential benefits of exergame that have been highlighted in this review.Such work may allow for exergames to be integrated into clinical practice or school routines, where exergames are utilized to support specific domains of ASD that negatively impact an individual's quality of life, in an enjoyable and person-centered manner, which is specifically tailored to the individual's interest and motivation.
P.O.Morris et al.

Table 2
(continued ) (continued on next page) P.O.Morris et al.

Table 2
(continued ) (continued on next page) P.O.Morris et al.

Table 2
(continued ) (continued on next page) P.O.Morris et al.