The impact of long-term exercise on motor skills in children with ADHD: a three-level meta-analysis

Abstract

Objective

This study aims to systematically evaluate the intervention effects of long-term exercise on motor skills in children with ADHD (Attention Deficit and Hyperactivity Disorder) using a three-level meta-analysis.

Methods

A computer-based search was conducted in four databases: Embase, The Cochrane Library, PubMed, and Web of Science, for randomized controlled trials related to the effects of long-term exercise on motor skills in children with ADHD, with search dates ranging from database inception to December 3, 2024. The risk of bias in the included studies was assessed using the Rob 2.0 tool. Effect sizes were synthesized, and influence analysis, moderator analysis, and publication bias assessment were conducted using the metafor package in R software. Evidence quality was evaluated using GRADEpro.

Results

A total of 9 studies were included. The intervention effect of long-term exercise on motor skills in children with ADHD was statistically significant (g = 0.72, 95% CI = (0.31, 1.14), P = 0.001), with a moderate level of evidence. Diagnosis method, exercise type, duration, frequency, and motor skill type were not found to be moderating factors (P > 0.05 for all).

Conclusion

Long-term exercise can enhance motor skills in children with ADHD; however, this finding should be interpreted with caution. Further validation through large-scale randomized controlled trial is required in future research.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12887-025-05857-5.

Introduction

Attention Deficit and Hyperactivity Disorder (ADHD) is a common neurodevelopmental disorder during childhood, with a global prevalence ranging from 5.9–7.1% [1, 2]. ADHD is primarily characterized by attention deficits, hyperactivity, and impulsivity, with approximately 30–50% of affected children exhibiting motor skill impairments [3].Compared to typically developing peers, children with ADHD experience a delay of approximately two years in developing motor skills [4, 5]. Children with ADHD often experience difficulties in both gross and fine motor skills. Gross motor skills involve moving large muscle groups or the whole body. Delayed development of gross motor skills can result in clumsiness and lack of coordination in daily activities and may lead to reluctance to participate in activities due to low self-esteem, which in turn affects their performance in both academic and physical activities [6]. This can negatively affect their social skills, peer relationships, and overall physical and mental health [7, 8]. Fine motor skills refer to the movement of small muscles or muscle groups, such as those in the wrist and fingers. Children with ADHD commonly experience fine motor deficits, particularly in tasks such as handwriting and tying shoelaces [9]. These difficulties negatively impact their daily life and academic performance [10]. The symptoms of ADHD often persist into adulthood, leading to widespread and negative impacts on academic performance, social life, and other areas, making it one of the major public health concerns and drawing global attention [3].

Recent studies have focused on the motor skills of children with ADHD, suggesting that improvements in motor skills may also contribute to the amelioration of symptoms in other areas [11]. Although medication can improve symptoms in children with ADHD, its side effects are considerable, prompting an increasing number of researchers to explore non-pharmacological alternatives. Exercise, due to its low cost and ease of implementation, has been used as an adjunctive or supplementary treatment for children with ADHD [12]. In addition, long-term exercise may yield more beneficial effects, as it has the potential to improve both brain and physical behaviors, with these effects being more enduring [13]. Current research indicates that exercise can improve motor skills in children with ADHD. Tascioglu et al. conducted studies on children with ADHD using Power Exercises and Traditional Strength Exercises, and the results showed that both types of exercise improved motor skills and physical function in these children [14]. Similarly, Ziereis et al. also demonstrated that exercise can enhance motor skills in children with ADHD [12].

Currently, several studies have used traditional meta-analysis to explore the effects of exercise on motor skills in children with ADHD, but the results are inconsistent or insufficient. Sun et al.‘s meta-analysis suggests that exercise can improve motor skills in children with ADHD compared to medication, but it did not further investigate the role of moderating factors [15]. Vysniauskas et al.‘s meta-analysis also indicates a positive effect of exercise on motor skills in children with ADHD [16]. However, Zhang et al. found that exercise improves gross motor skills in children with ADHD but has no effect on fine motor skills [17]. Differences in exercise protocols may lead to varying outcomes, such as variations in exercise type, duration, frequency, and other parameters [18]. Furthermore, previous studies primarily employed traditional meta-analysis techniques, which did not incorporate all effect sizes, thereby potentially weakening the statistical power of the analysis [19].

Based on this, the present study employs a three-level meta-analysis that incorporates all available effect sizes related to motor skills in children with ADHD, providing a comprehensive assessment of the impact of long-term exercise interventions on motor skills in this population. Additionally, this meta-analysis explores the influence of various moderating factors (e.g., exercise type, frequency, duration, diagnostic criteria, and motor skill categories). The aim is to offer evidence-based recommendations for clinical practice, particularly highlighting the potential of long-term exercise to facilitate significant and sustained improvements in motor skills for children with ADHD.

Methods

To ensure the systematicity and reproducibility of the meta-analysis, this study followed the PRISMA 2020 guidelines for literature search, screening, coding, quality assessment, publication bias, and evidence grading, and the results were reported accordingly [20]. The research protocol has been registered in the PROSPERO international systematic review registration platform (https://www.crd.york.ac.uk/PROSPERO/) (CRD42024621980).

Eligibility criteria and selection

Inclusion criteria

Population

Children with ADHD included in this study meet the diagnostic criteria of the International Statistical Classification of Diseases and Related Health Problems (ICD), the Diagnostic and Statistical Manual of Mental Disorders (DSM), or have been diagnosed by a clinical professional. Based on a review of previous studies, this study defines childhood as the age range of 6–14 years [21, 22].

Intervention

In the experimental group, at least one subgroup must have engaged in some form of physical activity, including aerobic exercise, stretching exercises, or exergaming. Long-term exercise is defined as any “bodily movement produced by skeletal muscles with the expenditure of energy” sustained for a minimum duration of three weeks, as previous research suggests that three weeks of exercise can improve motor skills in children with ADHD [23, 24].

Control

The control group does not engage in any form of exercise and only receives routine treatment, standard pharmacological interventions, or remains in a waitlist control condition.

Outcome

In this study, we define motor skills as the abilities related to voluntary movements that involve coordination and control of the body and its parts [25]. Motor skills can be further categorized into fine motor skills (e.g., hand-eye coordination, manual dexterity) and gross motor skills (e.g., running, balance, strength, agility, and bilateral coordination). Speed and agility, balance, bilateral coordination, and strength are considered part of motor skills, while endurance and flexibility are categorized under physical fitness [26]. The study employs objective motor skill testing methods to assess the motor skills before and after exercise, such as the Bruininks-Oseretsky Test of Motor Proficiency (This tool measures a range of motor skills and uses a composite structure organized around the muscle groups and limbs involved in movements for individuals aged 4–21 years. It includes four motor area composites: fine manual control, manual coordination, body coordination, and strength and agility [21]) and the Bruininks-Oseretsky Test of Motor Proficiency (This tool is suitable for assessing the motor skills of children aged 5–14 years, including running speed and agility, balance, bilateral coordination, strength, fine and gross motor skills, upper limb coordination, response speed, visual motor control, and upper limb speed and dexterity. Subscales are typically summed to generate a score for fine and gross motor skills [27]).

Study design

The study design is a randomized controlled trial.

Exclusion criteria

Studies were excluded if the experimental group did not involve exercise or if the intervention combined exercise with treatments provided to the control group. Additionally, the following types of literature were excluded: case studies, review articles, conference papers, and abstracts. Non-English publications, duplicate studies, and studies for which data could not be obtained after contacting the original authors via email were also excluded.

Search strategy

The researchers conducted a computer-based search in PubMed, Embase, Cochrane Library, and Web of Science (WOS) for relevant literature on long-term exercise interventions for motor skills in children with ADHD, with the search period extending from the inception of each database to December 3, 2024. The search strategy combined both subject headings and free text terms, and Boolean operations using “OR” and “AND” were applied to link the terms. The keywords included: exercise, “physical activity,” swimming, badminton, “attention deficit disorders with hyperactivity,” ADHD, “attention deficit disorder,” “motor development,” “gross motor,” “fine motor,” “motor skills,” child, and children. The detailed search strategy is provided in Appendix 1.

Literature screening and data extraction

Literature screening

Two researchers independently screened the literature according to the inclusion and exclusion criteria. First, the retrieved literature was imported into Endnote X9 to eliminate duplicate literature and read the titles and abstracts of literature for preliminary screening. Secondly, the full-text reading of the screened literature was conducted for re-screening. Then the final included literature would be determined. If two researchers have disagreements, a third researcher would join in the discussion to make the final decision.

Data extraction and coding strategy

Two authors extracted data separately by using a pre-developed extraction form in Microsoft Excel. The agreement of extraction between the two was 98.10%. If the data was missing or unclear, the original author would be contacted through email; when the information extracted by two researchers was inconsistent, a third researcher would join in the discussion and make a decision together. The information extracted by the researchers included: author(s), publication year, country, sample size, age, intervention type, intervention duration, session duration, intervention frequency, diagnostic method, and outcome measures.

We categorized motor skills into gross motor skills, fine motor skills, and mixed motor skills. Additionally, exercise modalities were classified into traditional exercise methods (e.g., swimming, judo, table tennis) and novel exercise methods (e.g., exergaming, virtual reality-based games) [28]; the intervention durations were divided into 3–6 weeks and 8–12 weeks, as existing research suggests that exercise interventions lasting 3–6 weeks can lead to initial improvements in ADHD symptoms, while interventions exceeding 8 weeks result in more substantial improvements [29, 30]; intervention frequencies into 1–2 times per week and 3 times per week [29]; and diagnostic methods into DSM, ICD, and clinical diagnosis.

Methodological quality assessment and evidence grading evaluation

This study assessed the risk of bias using the Cochrane Risk of Bias Assessment Tool for Randomized Controlled Trials (RoB 2.0). RoB 2.0 evaluates bias across five domains: “Randomisation process,” “Deviations from the intended interventions,” “Missing outcome data,” “Measurement of the outcome,” and “Selection of the reported result.” The risk of bias is classified into three levels: “Low risk of bias,” “Some concerns,” and “High risk of bias.” If all domains are rated as “Low risk of bias,” the overall risk of bias is also considered “Low risk of bias.” If at least one domain is rated as “Some concerns” but no domain is rated as “High risk of bias,” the overall risk of bias is categorized as “Some concerns.” However, if any domain is rated as “High risk of bias,” the overall risk of bias is classified as “High risk of bias” [31].

The quality of outcome evidence was assessed using GRADEpro software. “High” indicates strong confidence that the predicted value is close to the true value; “Moderate” indicates some confidence, but there is a possibility of discrepancy; “Low” suggests limited confidence, with a greater chance of the predicted value differing from the true value; and “Very low” suggests minimal confidence, with a large likelihood of significant difference.

Two researchers independently screened the literature according to the inclusion and exclusion criteria. If two researchers have disagreements, a third researcher would join in the discussion to make the final decision.

Statistical analysis

This study conducted a three-level meta-analysis using the random-effects model with the metafor package in R version 4.3.0 while modifying the syntax based on previous tutorials [19, 32]. This model considers the following three categories of effect size variability: sampling variance (level 1); within-study variance (level 2); and between-study variance (level 3). Pooled effect sizes using Hedges’ g (g) and 95% confidence intervals (CI) of exercise on motor skills. When the effect size was divided according to Hedges’ g, P < 0.05 was statistically significant. Hedges’ g < 0.2 refers to a small effect, 0.2–0.49 is a small-to-moderate effect, 0.5–0.79 is a medium effect, and ≥ 0.8 is a large effect.95% prediction intervals (PI) were calculated to determine the expected range in which an effect size in future identical studies will fall [33].

The overall heterogeneity was assessed using the Q test, and the within-study variance (level 2) and between-study variance (level 3) were further examined through one-sided log-likelihood ratio tests to determine the distribution of heterogeneity. Funnel plots and multilevel Egger’s regression tests were used to examine the possibility of publication bias [34], If publication bias is detected, a trim-and-fill method will be used for further testing [35]. To test whether influential cases distort the result of the meta-analysis, we conducted the influence analysis [32].

The data extracted in this study consisted of the means and standard deviations for the baseline and post-test scores of both the experimental and control groups. Change values were derived using a conversion formula, and effect size merging and analysis were subsequently performed on the change values.

The meta-analysis followed these steps: (1) Synthesizing effect sizes to examine the intervention effects of long-term exercise on motor skills in children with ADHD; (2) Identifying outliers; (3) Removing outliers, conducting effect size synthesis, and exploring the influence of moderating factors.

Results

Literature search results

A computer-based search was conducted in PubMed (n = 9), Cochrane (n = 21), Web of Science (n = 63), and Embase (n = 22), with an additional 3 articles identified through other sources, yielding a total of 118 articles. First, duplicate articles (n = 27) were removed using EndNote X9, leaving 91 articles for title and abstract screening, resulting in the exclusion of 63 irrelevant articles. Next, the remaining 28 articles were downloaded for reading, among which 7 conference abstracts were excluded. Full-text reading was then conducted for the remaining 21 articles for re-screening. Finally, 5 articles did not include outcome measures for motor skills, 1 article was a conference proceeding, and 6 articles had inaccessible data. Ultimately, 9 articles were included in the analysis [21, 22, 24, 36–41], as shown in Fig. 1.

Fig. 1.

Literature Screening Process

Study characteristics

A total of 9 studies were included, originating from China, Switzerland, Brazil, the Netherlands, Iran, and Germany, with publication years ranging from 2015 to 2022. The studies involved 312 children with ADHD, diagnosed according to DSM, ICD, or clinical diagnosis. The age range of the participants was 6–14 years, with a higher proportion of boys than girls. One study assessed motor skills using graphomotor function tests; three studies utilized the Movement Assessment Battery for Children-Second Edition (MABC-2); three studies employed the Bruininks-Oseretsky Test of Motor Proficiency-Second Edition (BOT-2); one study used the German Motor Test; and one study evaluated motor skills using corporal coordination tests, as shown in Table 1.

Table 1.

Basic information of the included studies

Included literature	Country	Number/Gender	Age (Years)	Diagnostic tools	Outcomes
Chang et al., 2022	China	E1:16	E1:8.31 ± 1.3	DSM-IV	GFT
		E2:16	E2:8.38 ± 1.2
		C:16/	C:8.38 ± 1.31
		Girls:9	Range: None
		Boys:39
Pan et al., 2016	China	E:16	E:8.93 ± 1.49	DSM-IV	BOT-2
		C:16/	C:8.87 ± 1.56
		Girls:0	Range: 6–12
		Boys:32
Pan et al., 2017	China	E:12	E:9.63 ± 2.48	DSM-IV-TR	BOT-2
		C:12/	C:9.38 ± 2.69
		Girls:0	Range: 7–14
		Boys:24
Benzing et al., 2018	Switzerland	E:28	E:10.46 ± 1.3	ICD-10	German motor test
		C:23/	C:10.39 ± 1.44
		Girls:9	Range: 8–12
		Boys:42
Silva et al., 2020	Brazil	E:10	E:12 ± 1	DSM-IV	Corporal coordination tests
		C:10/	C:12 ± 2/
		Girls:14	Range:11–14
		Boys:6
Weerdmeester et al., 2016	Netherlands	E:37	E:9.84 ± 1.71	Doctor’s diagnosis	MABC-2
		C:36/	C:9.69 ± 1.79
		Girls:15	Range:6–13
		Boys:58
Yazd et al., 2015	Iran	E1:12	E1:7.7 ± 1.3	DSM-IV-TR	BOT-2
		E2:12	E2:8.7 ± 2
		C:12/	C:8 ± 1.9
		Girls:6	Range:6–13
		Boys:30
Meßler et al., 2016	Germany	E:14	11 ± 1	ICD-10	MABC-2
		C:14/	Range:6–12
		Girls:0
		Boys:28
Ludyga et al., 2022	Switzerland	E:29	E:10 ± 1.2	DSM-5	MABC-2
		C:28	C:10.8 ± 1.2

Note: E 1: Experimental Group (1) E2: Experimental Group (2) C: Control Group. DSM-IV: the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition. BOT-2: the Bruininks-Oseretsky Test of Motor Proficiency, Second Edition. DSM-IV-TR: Diagnostic and Statistical Manual of Mental Disorders (4th ed., text rev). ICD-10: the International Statistical Classification of Diseases and Related Health Problems. MABC-2: Movement Assessment Battery for Children. GFT: Graphomotor function tests

Among the nine included studies, the exercise interventions varied. Two studies implemented table tennis, with one study incorporating two forms: traditional table tennis and table tennis in a virtual environment. One study employed SDHRP, a multi-component, multi-modal exercise program. One study utilized high-intensity interval training (HIIT), one focused on swimming, one used judo, and two studies applied exergaming or video game-based interventions. Regarding the duration of the intervention, four studies lasted for 12 weeks, two for 8 weeks, two for 3 weeks, and one for 6 weeks. In terms of exercise frequency, one study prescribed sessions once a week, four studies three times a week, and four studies twice a week. The duration of each exercise session was not reported in two studies; among the remaining studies, session lengths were reported as 60 min (two studies), 70 min (one study), 90 min (one study), 30 min (one study), and 15 min (one study), as shown in Table 2.

Table 2.

Characteristics of interventions and main outcomes of included studies

Included literature	Intervention forms	Feature of intervention (Duration/Frequency/Session time/)	Results (post-test mean, standard deviation, and P-value)
Chang et al., 2022	E1: Actual table tennis E2: Simulated (Wii) table tennis C: Conventional therapy	12week/3times a week/60minutes	SPC: E1:56.5(17.48), E2:55.25(24.67), C:44.56(21.86),P<0.01. RT: E1:113.83(49.69), E2:164.63(56.5), C:208.63(66.67),P<0.01. TA: E1:8.44(5.21), E2:8.09(3.63), C:13.88(5.22),P<0.01. MSV: E1:27.57(10.97), E2:27.74(7.90), C:19.95(4.96),P<0.05.
Pan et al., 2016	E: Table tennis C: Conventional therapy	12week/2times a week/70minutes	FMC: E:54.38(14.56), C:50.19(12.94),P>0.05. MC: E:60.44(8.85), C:59.88(8.99),P<0.01. BC: E:57(7.96), C:69.50(6.86),P>0.05. SA: E:69.5(6.86), C:61.25(7.75),P<0.01. TMC: E:64.13(9.50),C:55.88(10.74),P<0.01.
Pan et al., 2017	E: SDHRP (multicomponent exercise) C: Conventional therapy	12week/1time a week/90minutes	TMC: E:57.25(10.86), C:44.50(5.84),P<0.05. FMC: E:45.67(6.07), C:36.17(4.57),P<0.05. MC: E:60.25(10.69), C:50.58(7.72),P<0.05. BC: E:52(8.01), C:43.5(4.19),P<0.05. SA: E:62.75(10.78), C:53.92(9.53),P<0.01. FMP: E:13.17(2.62),C:8.92(2.68),P>0.05. FMI: E:13.33(2.90),C:9.58(3.09),P>0.05. MD: E:19.58(4.60),C:16.08(4.89),P<0.05. ULC: E:18.50(5.42),C:14.92(3.70),P>0.05. BC: E:18.58(3.96),C:15.58(2.02),P<0.05. B: E:13.17(3.30),C:10.17(2.59),P>0.05. RSA: E:22.58(3.70),C:20.08(3.99),P>0.05. S: E:17.83(4.99),C:13.75(4.41),P>0.05.
Benzing et al., 2018	E: Exergaming intervention C: Conventional therapy	8week/3times a week/30minutes	BB: E:103.53(11.10), C:101.35(10.47), P = 0.16. JS: E:110.48(9.68), C:105.81(11.40), P = 0.04. SU: E:93.91(10.73), C:94.14(19.37), P = 0.23. PU: E:108.47(10.29), C:105.95(10.79), P = 0.04. LJ: E:99.53(9.40), C:99.68(9.77), P = 0.55. SR: E:104.55(12.24), C:101.40(8.75), P = 0.22. TS: E:103.77(8.51), C:101.27(6.39), P = 0.01
Silva et al., 2020	E: Swimming C: Conventional therapy	8week/2times a week/45minutes	Lower limbs: E:76(5), C:59(9), P = 0.05. Laterality: E:108(8), C:85(11), P = 0.04. B: E:108(11), C:102(14), P = 0.78. Velocity: E:93(13), C:96(9), P = 0.68. Flexibility: E:24(4), C:21(7), P = 0.049. Abdominal resistance: E:38(3), C:28(4), P = 0.037.
Weerdmeester et al., 2016	E: Videogame C: Conventional therapy	3week/2times a week/15minutes	FMS: E:4.82(2.30),C:4.57(2.08),P<0.01. GMS: E:9.05(3.54),C:8.14(3.08),P>0.05.
Yazd et al., 2015	E1: Motor-perceptual training combine drug therapy E2: Motor-perceptual training C: Conventional therapy	6week/3times a week/Unclear	FMS: E1:38.10(11.01), E2:42.00(2.33), C: 28.33(7.96),P<0.01. GMS: E1:49.50(9.95), E2:52.10(12.63),C:32.08(9.67),P<0.01.
Meßler et al., 2016	E: High-intensity interval training C: Standard multimodal therapy	3week/3times a week/Unclear	MD: E:9.4(2.5), C:8.6(3.4), P = 0.79. BS: E:12.8(3.0), C:9.1(2.5), P = 0.03. SB: E:10.4(2.6), C:9.2(3.5), P = 0.35. TS: E:11.1(2.5), C:8.7(3.5), P = 0.58.
Ludyga et al., 2022	E: Judo training C: Wait-list control	12week/2times a week/60minutes	E: 38.1(26.9), C: 45.6(28.4), P = 0.33.

Note: E 1: Experimental Group (1) E2: Experimental Group (2) C: Control Group. SPC: Short Paragraph Copy. RT: Response Time. TA: Time to Automation. MSV: Mean Stroke Velocity. FMC: Fine manual control. MC: Manual coordination. BC: Body coordination. SA: Strength and agility. TMC: Total motor composite. FMP: Fine motor precision. FMI: Fine motor integration. MD: Manual dexterity. ULC: Upper-limb coordination. BCD: Bilateral coordination. B: Balance. RSA: Running speed and agility. S: Strength. BB: Balancing backwards. JS: Jumping sideways. SU: Sit-ups. PU: Push-ups. LJ: Long-jump. SR: Stand and reach. FMS: Fine motor skills. GMS: Gross motor skills. BS: Ball skills. SB: Static/dynamic balance. TS: Total score. MD: Manual dexterity

Risk of bias and evidence quality assessment

A total of 9 studies described the “Randomisation process.” 6 studies were assessed as having a low risk of bias in the domain of “Deviations from the intended interventions,” while 7 studies exhibited a low risk of bias in “Missing outcome data.” Additionally, 3 studies were identified as having a low risk of bias in “Measurement of the outcome,” and 6 studies demonstrated a low risk of bias in “Selection of the reported result.” Overall, 1 studies were categorized as low risk, 3 studies as high risk, and 5 studies were classified as having “Some concerns.” See Fig. 2.

Fig. 2.

The risk of bias of the included studies

We used GRADEpro to assess the quality of evidence regarding the intervention effect of long-term exercise on motor skills in children with ADHD. Since all the studies included in this research were randomized controlled trials, the initial evidence level was considered high. However, due to the presence of publication bias, the evidence level was downgraded by one level, resulting in a final evidence level of moderate. See Table 3.

Table 3.

Level of evidence for outcome indicators

Outcome	Evaluation of Evidence Quality Level					Relative Effect Size	Level
	1	2	3	4	5
Motor skills	N	N	N	N	D*	0.31, 1.14	Moderate

Note: 1 for Study limitation. 2 for Inconsistency. 3 for Indirectness. 4 for Imprecision. 5 for Publication bias. D for Downgrade 1 Level. N for Not Downgraded. * This study has publication bias

Results of meta-analyses

To investigate the intervention effects of exercise on motor skills in children with ADHD, we included all relevant effect sizes and conducted a three-level meta-analysis to assess the effect. The three-level random-effects model indicated that exercise improves motor skills in children with ADHD, with a pooled effect size of g = 0.79, 95% CI = (0.33, 1.25), P = 0.001, and a 95% PI = (-0.54, 2.12). The two-level random-effects model also showed that exercise improves motor skills in children with ADHD, with a pooled effect size of g = 0.85, 95% CI = (0.65, 1.04), P < 0.001, and a 95% PI = (0.17, 1.63). See Appendix 2.

Influence analysis

To examine whether outliers influenced the results of the meta-analysis, we conducted an influence analysis. The results revealed two outliers [22], as shown in Appendix 3. To explore the impact of these outliers, we performed a meta-analysis excluding the outliers. The three-level random-effects model indicated that exercise improves motor skills in children with ADHD, with a pooled effect size of g = 0.72, 95% CI = (0.31, 1.14), P = 0.001, and a 95% PI = (-0.44, 1.88). The two-level random-effects model also showed that exercise improves motor skills in children with ADHD, with a pooled effect size of g = 0.76, 95% CI = (0.58, 0.94), P < 0.001, and a 95% PI = (0.58, 0.94), as shown in Fig. 3.

Fig. 3.

Intervention Effects of Long-Term Exercise on Motor Skills in Children with ADHD (Excluding Outliers). Note: As shown in Appendix 2

Heterogeneity analysis

The log-likelihood ratio test revealed that the within-study variance (level 2) was not significant, LRT = 0.00, P = 0.05, indicating that the three-level model was not superior to the two-level model. However, the between-study variance (level 3) was significant, LRT = 12.85, P < 0.001, suggesting that the three-level model was superior to the two-level model. In the total variance, the sampling variance (level 1) accounted for 56.43%, the within-study variance (level 2) accounted for 1.59%, and the between-study variance (level 3) accounted for 43.57%.

Moderator analyses

Due to the lack of information on exercise intensity and duration in some studies, we conducted moderator effect analyses on exercise type, intervention duration, exercise frequency, diagnostic methods, and types of motor skills. See Table 4.

Table 4.

Moderating effect of exercise on improving intervention effect in children with ADHD

	k	g	95%CI	F	DF	P	Level2 Vaniance	Level3 Vaniance
Exercise type				0.17	1	0.68	1.64	45.49
Traditional forms of exercise	22	0.78	0.27, 1.28			0.004
New forms of exercise	26	0.64	0.08, 1.21			0.027
Exercise duration				0.16	1	0.70	1.24	48.16
3–6 weeks	10	0.85	0.06, 1.64			0.037
8–12 weeks	38	0.66	0.12, 1.20			0.018
Exercise frequency				0.37	1	0.55	1.20	48.02
1–2 times per week	25	0.59	-0.02, 1.20			0.056
3 times per week	23	0.86	0.21, 1.52			0.011
diagnostic methods				1.40	2	0.26	1.76	38.67
DSM-IV	35	0.96	0.48, 1.43			<0.001
ICD-10	11	0.31	-0.48, 1.09			0.43
Doctor	2	0.22	-0.98, 1.42			0.71
Motor Skills				0.48	2	0.617	2.33	40.69
Fine skills	21	0.83	0.35, 1.31			0.001
Gross skills	23	0.71	0.24, 1.17			0.003
Mix	4	0.47	-0.27, 1.21			0.20

We categorized exercise type into “traditional exercise methods” and “novel exercise methods,” and found that exercise type did not have a moderating effect (F = 0.17, P = 0.68). The “traditional exercise methods” showed a statistically significant effect (g = 0.78, 95% CI = (0.27, 1.28), P = 0.004), and the “novel exercise methods” also demonstrated a statistically significant effect (g = 0.64, 95% CI = (0.08, 1.21), P = 0.027).

We categorized the exercise duration into “3–6 weeks” and “8–12 weeks,” and found that exercise duration did not have a moderating effect (F = 0.16, P = 0.70). The “3–6 weeks” group showed a statistically significant effect (g = 0.85, 95% CI = (0.06, 1.64), P = 0.037), and the “8–12 weeks” group also demonstrated a statistically significant effect (g = 0.66, 95% CI = (0.12, 1.20), P = 0.018).

We categorized exercise frequency into “1–2 times per week” and “3 times per week,” and found that exercise frequency did not exhibit a moderating effect (F = 0.37, P = 0.55). The “1–2 times per week” group did not show a statistically significant effect (g = 0.59, 95% CI = (-0.02, 1.20), P = 0.056), whereas the “3 times per week” group demonstrated a statistically significant effect (g = 0.86, 95% CI = (0.21, 1.52), P = 0.011).

We categorized the diagnostic methods into “DSM-IV,” “ICD-10,” and “clinical diagnosis,” and found that the diagnostic method did not exhibit a moderating effect (F = 1.40, P = 0.26). The “DSM-IV” group showed a statistically significant effect (g = 0.96, 95% CI = (0.48, 1.43), P < 0.001), while the “ICD-10” group did not show a statistically significant effect (g = 0.31, 95% CI = (-0.48, 1.09), P = 0.43), nor did the “clinical diagnosis” group (g = 0.22, 95% CI = (-0.98, 1.42), P = 0.71).

We categorized motor skills into “Fine skills,” “Gross skills,” and “Mix,” and found that the type of motor skill did not exhibit a moderating effect (F = 0.48, P = 0.617). The “Fine skills” group showed a statistically significant effect (g = 0.83, 95% CI = (0.35, 1.31), P = 0.001), as did the “Gross skills” group (g = 0.71, 95% CI = (0.24, 1.17), P = 0.003). However, the “Mix” group did not show a statistically significant effect (g = 0.47, 95% CI = (-0.27, 1.21), P = 0.20).

Publication bias

Funnel plots and Egger’s regression test were employed to assess publication bias. The results revealed asymmetry in the funnel plot, and Egger’s regression test also indicated the presence of publication bias regarding the effect of exercise on basic motor skills in children with ADHD (t = 5.49, P < 0.001). We applied the trim-and-fill method to impute values on the left side of the funnel plot. The results showed that the effect size did not change significantly (g = 0.50, 95% CI = (0.31, 0.70), P < 0.001). See Fig. 4.

Fig. 4.

Publication bias of the intervention effect of long-term exercise on motor skills in children with ADHD

Discussion

Our study indicates that exercise can improve motor skills in children with ADHD. Similar results have been reported in previous studies. A meta-analysis by Sun et al. showed that exercise can improve motor skills in children with ADHD [15], and a meta-analysis by Vysniauske et al. also demonstrated that exercise has a positive effect on motor skills in children with ADHD [16]. This may be because long-term, regular exercise can enhance brain structure and function [42]. Additionally, most of the studies included in this research involved racket sports (e.g., table tennis), cycling, and swimming. Racket sports, which engage small muscle groups in the fingers and wrists, contribute to the development of fine motor skills in children with ADHD. On the other hand, activities such as swimming and cycling, which involve larger muscle groups, help lay the foundation for fine motor development by improving body balance, coordination, and muscle strength, all of which contribute to the development of fine motor skills in the hands. This finding aligns with previous research, which suggests that young children initially use large muscle groups in their arms to perform writing tasks, and as practice and experience increase, smaller muscle groups in the hands and fingers take over the control of writing tools, thereby facilitating the development of fine motor skills [43]. Consequently, the development of gross motor skills in children occurs earlier, providing a foundation for fine motor development [43]. However, physical activity has a more beneficial effect on the motor skills of children with ADHD compared to typically developing children. Pan et al. demonstrated that after physical exercise, children with ADHD showed greater improvements in “fine manual control,” “manual coordination,” “body coordination,” “strength and agility,” and most other motor functions compared to typically developing children [37]. This may be due to the fact that children with ADHD generally have poorer motor skills compared to their peers. Tascioglu et al. compared children with ADHD to typically developing children before the exercise intervention, and the results showed that children with ADHD were weaker in most motor skills. However, after the physical activity intervention, the motor skills of the children with ADHD improved to the level of their typically developing peers [14].

Our findings indicate that the type of exercise, duration of the exercise intervention, and exercise frequency are not significant moderators. In terms of exercise type, we classified it into traditional and novel forms for the first time. Both traditional and novel exercise types were found to improve the motor skills of children with ADHD, with moderate effect sizes for both. Whether traditional or novel exercise forms, both require participation in physical activities, which engage the body’s muscles and thereby enhance motor skills. Regarding the duration of exercise, we observed for the first time that 3–6 weeks yielded large effect sizes, while 8–12 weeks resulted in moderate effects. Li et al. suggested that increasing the duration does not necessarily improve motor skills in young children [44]. Additionally, long-term interventions may face challenges such as reduced adherence and motivation [45], or the “ceiling effect” where continued intervention with the same type of exercise may not yield additional benefits [46]. In terms of exercise frequency, we found for the first time that “1–2 times/week” had no effect, while “3 times/week” resulted in a large effect size. Similar conclusions have been drawn in other studies; for example, Engel et al.’s systematic review suggests that teacher-led exercise interventions need to be conducted three times a week or more to effectively improve basic motor skills in children with intellectual disabilities [47]. However, other studies argue that the total intervention duration does not have a significant relationship with intervention effects [18]. Therefore, further research is needed to explore the effects of different intervention elements on the improvement of motor skills in children with ADHD.

We found that neither the ADHD diagnostic method nor the classification of motor skills served as significant moderating factors influencing the effects of long-term exercise interventions on motor skills in children with ADHD. Although the moderator analysis did not yield statistically significant differences, it was conducted to explore any potential heterogeneity arising from the use of varying diagnostic criteria. In terms of motor skill classification, we found that long-term exercise improved both gross and fine motor skills in children with ADHD, with fine skills showing a large effect size (g = 0.83) and gross skills showing a moderate effect size (g = 0.71). This finding is inconsistent with previous research. Zhang et al. found that exercise could improve gross motor skills but not fine motor skills in children with autism/ADHD [17]. There may be three possible reasons for the discrepancy: first, our study focused solely on ADHD children, while Zhang et al. included both autism and ADHD children, without accounting for the heterogeneity arising from different populations; second, we categorized novel exercises, such as virtual reality, as part of the intervention. Many virtual reality games now incorporate physical activity elements [24], which not only improve mental health but also promote physical well-being [48].

Study strengths and limitations

The strengths of our study are as follows: (1) We only included randomized controlled trials, which enhances the reliability of the results. (2) We were the first to use a three-level meta-analysis approach to combine all effect sizes related to motor skills, which also increases the reliability of our findings. (3) We were the first to explore moderating factors that may influence motor skills in children with ADHD, providing a theoretical basis for future research.

Of course, there are certain limitations in our study. (1) We only included English-language studies and did not screen or include studies in other languages, which may have contributed to the publication bias observed in our research. (2) The outcome indicators included in the study were assessed using different standards, with varying levels of reliability and sensitivity, which may have compromised the credibility of the results. (3) Due to missing information in some of the included studies, we were unable to conduct a full analysis of moderating factors such as ADHD subtypes, exercise duration and intensity. Additionally, we could not perform a moderating factor analysis for the subtype of ADHD in the children. Future research should adhere to stricter study design protocols to obtain more rigorous results.

Prospects for future research

This study only included randomized controlled trials, but most of the studies did not implement allocation concealment, blinding of the researchers, or blinding of outcome assessors. Future research should conduct larger, more rigorously designed studies to explore the effects of exercise interventions on motor skills in children with ADHD. Efforts should also be made to standardize diagnostic methods in future studies to avoid the impact of diagnostic variations. Due to the inclusion of only nine studies in this research, and the lack of clarity in some of the studies’ information, a more detailed subgroup analysis could not be conducted. Additionally, the presence of publication bias may have influenced the findings. Therefore, future research should ensure a clear study design and comprehensive reporting of key information, incorporate studies in multiple languages or with different research designs, and further investigate the effects of exercise components (e.g., exercise type, duration, and intensity) as well as their dose-response relationship. Further validation of the intervention effects of long-term exercise on motor skills in children with ADHD is warranted. Finally, this study only included English language literature and journal articles, excluding studies in other languages as well as conference papers. Future research should incorporate literature in other languages and of different types to mitigate publication bias.

Conclusions

Long-term exercise can improve motor skills in children with ADHD, with a moderate level of evidence. However, this finding should be interpreted with caution. Further research is needed to validate the effects of long-term exercise on motor skills in children with ADHD.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Abbreviations

ADHD

Attention Deficit and Hyperactivity Disorder

B

Balance

BB

Balancing backwards

BC

Body coordination

BCD

Bilateral coordination

BOT-2

Bruininks-Oseretsky Test of Motor Proficiency-Second Edition

BS

Ball skills

C

Control Group

CI

Confidence Intervals

DSM

Diagnostic and Statistical Manual of Mental Disorders

DSM-IV

Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition

DSM-IV-TR

TR-Diagnostic and Statistical Manual of Mental Disorders (4th ed., text rev)

E1

Experimental Group 1

E2

Experimental Group 2

FMC

Fine manual control

FMI

Fine motor integration

FMP

Fine motor precision

FMS

Fine motor skills

GFT

Graphomotor function tests

GMS

Gross motor skills

HIIT

High-Intensity Interval Training

ICD

International Statistical Classification of Diseases and Related Health Problems

LJ

Long-jump

MABC-2

Movement Assessment Battery for Children-Second Edition

MC

Manual coordination

MD

Manual dexterity

MSV

Mean Stroke Velocity

PI

Prediction intervals

PU

Push-ups

RSA

Running speed and agility

RT

Response Time

S

Strength

SA

Strength and agility

SB

Static/dynamic balance

SPC

Short Paragraph Copy

SR

Stand and reach

SU

Sit-ups

TA

Time to Automation

TMC

Total motor composite

TS

Total score

ULC

Upper-limb coordination

WOS

Web of Science

Author contributions

C.G. W. wrote the manuscript, extracted data, and analyzed data. C.G. W, J.L. Z, and Y.Y. extracted data, analyzed data, and produced pictures. Y. Y. constructed the framework of the study and revised the manuscript.

Funding

This work was supported by The National Social Science Fund of China (Educational Science): A study on the development model and co-parenting path of young children’s physical ability (No. BLA220239).

Data availability

The datasets used and/or analyzed during the current study are available from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

This article is a systematic review and meta-analysis and does not involve any new studies with human participants performed by the authors. All included studies were conducted in accordance with the Declaration of Helsinki.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.