Effects of different physical activity interventions on children with attention-deficit/hyperactivity disorder: A network meta-analysis of randomized controlled trials

Background Previous studies have shown that physical activity interventions positively affect core symptoms and executive functioning in children with attention-deficit/hyperactivity disorder (ADHD). However, comparisons between different physical activity interventions still need to be made. This study is the first to analyze the effects of 10 different types of physical activity on children with ADHD through a network meta-analysis. Methods PubMed, Embase, Web of Science, and Cochrane Library databases were searched for randomized controlled trials on the effects of physical activity interventions on children with ADHD. The search time frame was from database creation to October 2022. Two investigators independently performed literature screening, extraction, and quality assessment. Network meta-analysis was performed with Stata 15.1. Results A total of 31 studies were included, and the results indicated that perceptual-motor training was the most effective in terms of motor ability and working memory (SUCRA = 82.7 and 73.3%, respectively). For attention problems and cognitive flexibility, aquatic exercise was the most effective (SUCRA = 80.9 and 86.6%, respectively). For social problems, horsemanship was the most effective (SUCRA = 79.4%). For inhibition switching, cognitive-motor training was the most effective (SUCRA = 83.5%). Conclusion Our study revealed that aquatic exercise and perceptual-motor training had a superior overall performance. However, the effects of various physical activity interventions on different indicators in children with ADHD can vary depending on the individual and the intervention’s validity. To ensure an appropriate physical activity intervention is selected, it is important to assess the severity of symptoms exhibited by children with ADHD beforehand.


Introduction
Attention-deficit/hyperactivity disorder (ADHD) is a neurodevelopmental disorder that affects approximately 7.2% of children worldwide (Thomas et al., 2015;Thapar et al., 2017). Its primary characteristics are inattention, impulsivity, and hyperactivity (American Psychiatric Association [APA], 2013), and can be divided into three distinct subtypes: inattentive, hyperactiveimpulsive, and combined inattentive and hyperactive-impulsive (American Psychiatric Association [APA], 2013). It has been shown that children with ADHD often develop diverse problems, including sleep disturbances, distractibility, motor deficits, decreased social skills, and decreased academic performance (Kim et al., 2011;Konicarova et al., 2014;Schneider et al., 2016;Yu et al., 2019). These issues are also persistent, frequently remaining when patients reach puberty and adulthood (Edition, 2013). Therefore, it is highly detrimental to the development of pediatric patients and hurts their physical and mental health, academic growth, and socialization process.
Given this, the treatment of pediatric patients with ADHD is of utmost importance. The most commonly used treatment modality is medication, such as methylphenidate (MPH) (Barkley and Poillion, 1994;Welsch et al., 2021), but it may cause side effects such as headache, stomach pain, and decreased appetite (De Sousa and Kalra, 2012). Meanwhile, in the past two decades, non-pharmacological interventions for ADHD have been rapidly developed and used (Cortese et al., 2022), such as physical activity interventions, neurofeedback interventions, and cognitive interventions (Jensen and Kenny, 2004;Sánchez-López et al., 2015;Sani et al., 2022), due to concerns about the side effects and long-term effects of pharmacological treatments (Coghill, 2019). Physical activity interventions, in particular, have gained traction due to their lower cost, ease of implementation, capacity to improve physical fitness, and additional benefits (Cornelius et al., 2017).
Previous research has uncovered a strong link between physical activity and various functions in individuals with ADHD. Barnard-Brak et al. (2011) utilized data from the Early Childhood Longitudinal Study, Kindergarten cohort (ECLS-K) to demonstrate that structured physical activity was associated with a decrease in ADHD symptoms over time. This may be due to the stimulation of the catecholamine system, which is known to be impaired in individuals with ADHD ( Barnard-Brak et al., 2011).
A recent study conducted by Fard et al. (2022) investigated the impact of physical activity on the physical and mental health of children and adolescents with ADHD, with self-esteem as a moderating factor. The results indicated that physical activity and health levels are integral components of well-being for this population and that self-esteem could be a potential mediator for the connection between physical activity and health outcomes (Fard et al., 2022). Some previous meta-analyzes have also shown evidence of better efficacy of physical activity in patients with ADHD. Cerrillo-Urbina et al. (2015) explored the impact of physical activity on core symptoms of attention, impulsivity, anxiety, and executive functioning in patients with ADHD. The results showed that physical activity was more effective than non-physical activity, particularly aerobic exercise (Cerrillo-Urbina et al., 2015). Zang (2019) assessed the effects of physical activity interventions compared to non-physical activity interventions in children with ADHD. The findings indicated that physical activity interventions had a significant positive effect on anxiety and depression, aggressive behavior, thinking, and social problems in children with ADHD (Zang, 2019). Lambez et al. (2020) performed a meta-analysis evaluating the effects of non-pharmacological treatments for ADHD on cognitive functioning. The interventions studied included neurofeedback, cognitive behavioral therapy, cognitive training, and physical exercise. Physical exercise was found to have the greatest mean effect size, particularly for inhibition (Lambez et al., 2020). Seiffer et al. (2022) studied the efficacy of moderate to vigorous exercise (MVPA) on children with attention deficit hyperactivity disorder (ADHD), focusing on the intensity component of physical activity. The study indicated that MVPA was the most effective treatment for ADHD and that it might be used as an alternative (Seiffer et al., 2022). Collectively, these findings suggest that physical exercise may be an effective treatment option for ADHD patients.
However, previous meta-analyzes have largely compared physical and non-physical activity, without examined the potential distinctions between different types of physical activity interventions. The types of physical activity are diverse and include many types of aquatic exercise, ball games, mind-body exercise, and high-intensity interval training. Therefore, what specific types of physical activity provide the most significant benefit to pediatric patients with ADHD? Through a network meta-analysis of randomized controlled trial studies of physical activity in pediatric patients with ADHD, this study provides valuable information for selecting the best physical activity for treating pediatric patients with ADHD.

Protocol and registration
The meta-analysis was conducted using the Cochrane Handbook for Systematic Reviews of Interventions, and the findings were reported according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyzes (PRISMA) statement (Higgins et al., 2019;Page et al., 2021). This network meta-analysis was prospectively registered in PROSPERO (CRD 42022363255).

Data sources and search strategy
We conducted a comprehensive search of four databases (PubMed, Web of Science, Embase, and the Cochrane Library) to identify relevant studies. Search strings included physical activity interventions, age ranges, and outcomes related to patients with ADHD. The search was performed up to October 2022. Supplementary Appendix A shows the detailed search strings for this search. Li et al. 10.3389/fnins.2023.1139263 FIGURE 1 Preferred Reporting Items for Systematic Reviews and Meta-Analyzes flow diagram of the study process. from databases and reference lists. The titles and abstracts of the research were first used to determine their relevance. Then, relevant full-text studies were retrieved and evaluated for inclusion. Any disagreements were resolved through discussion and consensus.

Inclusion and exclusion criteria
This systematic review employed specified inclusion criteria. Each study met the following criteria: (1) Only randomized controlled trials were included, and observational and cross-sectional studies were excluded. (2) The range of age participants in the sample must be 18 years or less. (3) The physical activity intervention had to contain a sports or physical activity component. Studies without physical activity intervention were excluded. (4) The study must report data on indicators of motor skills, attention problems, social problems, cognitive flexibility, inhibitory switching, and working memory in children with ADHD before and after the intervention. Studies that do not report on these indicators must be excluded.
(6) We only analyzed papers written in English and excluded papers written in other languages.

Data extraction
The data were extracted to a standardized Excel spreadsheet. Two authors collected the required data separately from the included studies. Disagreements encountered during the process were resolved through discussion with the group. The following data were extracted from the final study: author, year, country, subject characteristics, intervention characteristics, and ADHDrelated outcome indicators.

Quality assessment
The risk of bias was assessed using the Cochrane System Risk of Bias Assessment tool via Review Manager 5.4 software, which evaluates the studies' quality on seven indicators: 1. Random sequence generation; 2. Allocation concealment; 3. Blinding of    participants and personnel; 4. Blinding of outcome assessment; 5. Incomplete outcome data; 6. Selective reporting; and 7. Other bias.

Statistical analysis
We computed the standardized mean difference (SMD) and 95% CIs for continuous outcomes. The P-value of the chi-square test and the I2 index from the heterogeneity test were used to express the level of statistical heterogeneity. Different effect models were selected according to the level of heterogeneity of the test data. When the level of heterogeneity was low, a fixed-effects model (P ≥ 0.1 and I2 ≤ 50%) was used to analyze the data. Otherwise, a random-effects model (P < 0.1 or I2 values >50%) was used (Higgins et al., 2003).
According to the PRISMA NMA recommendations, we aggregated and analyzed NMA data using Markov chain Monte Carlo simulation chains in a Bayesian-based framework and Stata software (version 15.1) (Moher et al., 2015;Vats et al., 2019). We will employ the nodal method to quantify and demonstrate the congruence between indirect and direct comparisons, as obtained by Stata software instructions. If the p-value is greater than 0.05, the agreement test is passed.
Network meta-analysis was performed by employing a Bayesian model. The data were preprocessed using network group commands, and a mesh evidence map was drawn. The dots in the mesh evidence plot represent one intervention type, and the larger the area of its dots represents, the more significant the number of patients included in the study for the intervention. The line connecting the two dots is a direct comparison of the two interventions, and the thickness of the line represents the number of included studies. The larger the number of included studies, the thicker the line (Chaimani et al., 2013). The effects of the different movement methods were ranked. The effects of the different exercise modalities were ranked, the surface under the cumulative ranking curve (SUCRA) was obtained, and the probability ranking was plotted in a table. SUCRA is expressed as a percentage. The larger the percentage, the more effective the intervention. Additionally, to check for publication bias and minor sample study effects, we generated funnel plots for outcome indicators with study numbers >10 and used symmetry criteria to check (Khera et al., 2016). Stata15.1 was used to perform all statistical analyzes.

Trial selection
A total of 3,052 citations are yielded in the initial search of electronic databases, and an additional seven documents were manually searched. After removing duplicate studies (n = 1,129), 1,930 relevant papers remained. Subsequently, through screening, 1,809 papers were removed, and 121 papers suitable for full-text review remained, of which 90 were further eliminated. Finally, 31 studies were adopted for quantitative synthesis (Figure 1).

Risk of bias
Eighteen studies (58.1%) had a low risk of bias with respect to random sequence generation. Twenty-one studies (67.7%) had a low risk of bias with respect to allocation concealment. Sixteen studies (51.6%) had a low risk of bias with respect to the blinding of participants and personnel. Twenty-four studies (77.4%) had a low risk of bias with respect to the blinding of outcome assessments. Twenty-nine studies (93.5%) had a low risk of bias with respect to incomplete outcome data. Thirty studies (96.8%) had a low risk of bias with respect to selective reporting. Other biases are not known. Details of the evaluation of bias results for the included literature are shown in Figures 2A, B.

Network meta-analysis
The complete NMA figure will be presented in Supplementary Appendixes B1-6.

Motor ability
Seven studies (Chang et al., 2014;Yazd et al., 2015;Pan et al., 2016;Meßler et al., 2018;Oh et al., 2018;Da Silva et al., 2020;Sani et al., 2022) reported on the motor ability of children with ADHD, and a total of nine interventions are involved. As shown in Table 2 Table 3, there is no statistical significance for each intervention in the network meta-analysis results. In SUCRA, aquatic exercise ranked first in terms of the probability of the effect of different interventions on the reduction of attention problems (SUCRA: 80.9%, as shown in Supplementary Appendix C2).

Social problems
Five studies (Pan et al., 2016;Meßler et al., 2018;Oh et al., 2018;Ahn et al., 2021;Sabzi et al., 2021) reported on the social problems of children with ADHD, and a total of six interventions are involved. As shown in Table 4, the statistically significant results of the network meta-analysis were as follows: Aquatic exercise [MD = −3.70, 95% CI = (−5.03, −2.37)] was more effective than that with no intervention. In the SUCRA, aquatic exercise ranked first in terms of the probability of the effect of different interventions on the reduction of social problems (SUCRA: 79.4%, as shown in Supplementary Appendix C3).

Working memory
Ten studies (Kang et al., 2011;Hoza et al., 2015;Ziereis and Jansen, 2015;Bustamante et al., 2016;Janssen et al., 2016;Benzing et al., 2018;Benzing and Schmidt, 2019;Hattabi et al., 2019;Kadri et al., 2019;Milligan et al., 2019) reported on the working memory of children with ADHD, and a total of nine interventions are involved. As shown in Table 7, the statistically significant results of the network meta-analysis were as follows: cognitive-motor training [MD = 9.45, 95% CI = (2.39, 16.51)] was more effective than traditional aerobic exercise. In SUCRA, perceptual-motor training ranked first in terms of the probability of effectiveness of different interventions on working memory (SUCRA: 73.3%, as shown in Supplementary Appendix C6).

Publication bias
As is vividly shown in Supplementary Appendix D, funnel plots were employed to detect publication bias, while no significant publication bias was revealed by the visual inspection of funnel plots for all indicators.

Discussion
In this study, motor ability, attention problems, social problems, cognitive flexibility, inhibition switching, and working memory are adopted as outcome indicators to compare the effects of different interventions on each outcome indicator. As shown in Table 8, it has been shown in our current study that perceptualmotor training, traditional aerobic exercise, as well as aquatic exercise were the top three interventions for benign development in motor ability. When it comes to attention problems, aquatic exercise, pharmacotherapy, and cognitive-motor training were the top three interventions to reduce attention problems. As for the indicator of the social problem, horsemanship, pharmacotherapy, and aquatic exercise were the top three interventions in reducing social problems. In terms of cognitive flexibility, aquatic exercise, mind-body exercise, and cognitive intervention were the top three interventions to increase cognitive flexibility. For inhibition switching, cognitive-motor training, perceptual-motor training, and combination exercise were the top three interventions to reduce inhibition switching time. Finally, in terms of working memory indicators, perceptual-motor training, pharmacotherapy, and horsemanship were the top three interventions for enhancing working memory. It has been shown in our findings that there is no single intervention most effective across all outcome indicators, and different interventions may be more effective for different outcomes.
Perceptual-motor training is the best physical activity intervention for children with ADHD regarding motor ability and working memory. This type of training combines physical activities such as coordination, balance, and strength with perceptual tasks (Hattabi et al., 2019). Previous research has demonstrated a strong correlation between motor behavior and underlying perceptual processes (Chu and Reynolds, 2007). In particular, when physical activity is designed to improve attention, it will contribute to developing executive functions (Piek et al., 2004;Hung et al., 2013). By combining training activities with perceptual tasks, there is potential for an overall improvement in motor ability and working memory in children with ADHD (Mandich et al., 2001).
The aquatic exercise was the intervention with the highest frequency (4 sessions) in the top three rankings for all outcome indicators and the best physical activity intervention in terms of both attention problems and cognitive flexibility. Aquatic exercise is a form of physical activity in which the training process is completed in an aquatic environment. Due to the fluid nature of water, physical activity in an aquatic environment requires participants to constantly pay attention to the environment's fluctuations (Vivas et al., 2011). At the same time, the buoyancy effect of water provides an auxiliary force, resistance, or support, which makes physical activity in the water environment safer, and children's activity can be more active (Broach and Dattilo, 1996).

Frontiers in Neuroscience
For example, swimming in water sports is a highly coordinated and lateralized sport requiring control of the upper and lower limbs in an aquatic environment (Colgate and Lynch, 2004). This feature may allow for further activation of brain regions in the prefrontal cortex and amygdala, thus contributing to improved attentional problems and cognitive flexibility (Faw, 2003). Horsemanship is the best physical activity intervention in terms of indicators of social problems. Horsemanship is a physical activity modality through learning activities with horses as a vehicle (Kern et al., 2011). It has been shown that because equestrian learning requires participants to establish trust and frequent interaction with the horse, it contributes to developing participants' social competence (Hauge et al., 2014) and self-efficacy (Bizub et al., 2003). With this mutual relationship with the horse, children experience the horse's feelings, which are then internalized in their behavior, enabling further development of empathy. This change will likely transfer to human interactions (Granados and Agis, 2011). At the same time, the horse's rhythmic activity also improves the participants' physiological responses to stress and impulsivity (Tyler, 1994;Jang et al., 2015).
Cognitive-motor training is the best physical activity intervention for inhibiting conversion indicators. Cognitivemotor training is an intervention that integrates cognitive and motor tasks to promote an individual's physical and mental health (Amini et al., 2022). It has been shown that performing two or more cognitive-motor tasks simultaneously, such as computation in postural training and movement under computer games, will contribute more to improvements in cognitive domains compared to single-task training (van der Niet et al., 2016;Luder et al., 2018;Schmidt et al., 2020) while reducing reaction time (Wollesen et al., 2020). Cognitive-motor training requires participants to use both skill and cognitive effort to cope with unpredictable stimuli from the external environment (Chuang et al., 2015). Therefore, some researchers have suggested that this may improve participants' executive functioning, including improvements in inhibitory switching (Kunstler et al., 2018;Gao et al., 2019).
In conclusion, physical activity interventions have varying levels of effect on different indicators related to the symptoms of children with ADHD. This impact is dependent on the components, characteristics, and settings of the intervention. Nevertheless, physical activity interventions have been found to have numerous advantages across multiple indicators.

Strengths and limitations
One advantage of our current study is that we are the first network meta-analysis of the effects of physical activity on symptoms related to children with ADHD, which provides some scientific reference for selecting appropriate physical activity therapy for children with ADHD. The second advantage is that this study explored the effects of different physical activities on different symptom indicators in children with ADHD, which can provide some scientific reference for targeted treatment. The third advantage is that the current study only included studies from randomized controlled trials and excluded observational and crosssectional studies, which helped to enhance the reliability of the findings. However, our reticulated meta-analysis also has some limitations that may affect the interpretation of the results. First, the relatively small number of available studies and the limited number and sample size of studies included in the analysis makes it difficult to give a particularly robust conclusion. Second, the outcome indicators that could be included are still limited. In the future, more outcome indicators of symptoms related to children with ADHD should be included based on an adequate number of studies. Finally, findings should be interpreted with caution because of the small number of studies and the limited evidence for direct comparisons of some interventions. Relevant studies should be further expanded to provide evidence with higher confidence.

Conclusion
Our current study showed that the overall performance of aquatic exercise and perceptual-motor training was better. However, different physical activity interventions have different validity and individual differences regarding their effects on different indicators in children with ADHD. Therefore, to ensure that the most suitable physical activity intervention is chosen, it is essential to accurately assess each child's specific ADHD symptoms before implementation.

Author Contributions
CL and DL conceived and designed the study. DL and DW collected the data. DL, DW, and WC analyzed and interpreted the data. DL drafted the manuscript. JY and WZ revised the manuscript. All authors have read and agreed to the published version of the manuscript, and contributed to the study conception and design.