The Effects of Physical Exercise on Functional Outcomes in the Treatment of ADHD: A Meta-Analysis

Abstract

Objective: An increasing number of studies suggest possible beneficial effects of exercise in alleviating ADHD functional outcomes. The current study provides a quantitative meta-analysis of the available studies investigating this relationship. Method: Studies reporting on the effects of physical exercise on motor skills and executive functions in children with ADHD were identified through Cochrane, PsycInfo, PubMed, Web of Science databases. Ten publications were selected. Random-effects model was used to calculate effect sizes. Results: There was a significant effect of exercise on ADHD functional outcomes (g = 0.627). Longer exercise intervention duration was consistently associated with larger effect sizes. Effect sizes were not related to exercise intensity, mean age of participants, or gender distribution. Conclusion: Results suggest that exercise has a modest positive impact on ADHD functional outcomes, such as executive functions and motor skills, with longer interventions yielding better results.

Keywords

ADHD exercise executive function motor performance children

Introduction

With a prevalence rate of 3% to 5%, ADHD is one of the most frequently diagnosed neurodevelopmental disorders in school-aged children (Faraone, Pucci, & Coghill, 2009; Gordon Millichap, 2011). Its essential features are developmentally inappropriate levels of inattention and/or hyperactivity–impulsivity (Gordon Millichap, 2011). Children with ADHD often show co-occurring disruptive behaviors, and are at a higher risk to develop anxiety, depression, and learning disorders (Barkley, 2006; Weiss & Hechtman, 1993). Frequently co-occurring neurocognitive sequels of ADHD include impaired executive functioning (EF), slower and more variable response speed, increased aversion to delay, and motor skills deficits (Boonstra, Oosterlaan, Sergeant, & Buitelaar, 2005; Piek, Pitcher & Hay, 1999; Willcutt, 2010).

Currently, central nervous system stimulant medication and behavioral interventions are two treatment options that are empirically supported and universally recognized as effective for amelioration of ADHD symptoms. However, up to 30% of children do not respond favorably to medication or are unable to tolerate frequent occurring side effects, including insomnia, appetite suppression, growth retardation, and headaches (Connor, 2006; Wigal, Emmerson, Gehricke, & Galassetti, 2013). Moreover, medication does not yield a long-term benefit, as its effectiveness is limited to the period of active drug administration (Chronis, Jones, & Raggi, 2006; Pelham & Fabiano, 2008).

Behavioral interventions such as behavioral parent training and behavioral classroom management provide an alternative evidence-based approach (Pelham & Fabiano, 2008; The Multimodal Treatment Study of Children with Attention-Deficit/Hyperactivity Disorder (MTA) Cooperative Group, 1999). Although this treatment modality offers improvement in specific areas that are affected less by pharmacological interventions, such as parenting behavior and family functioning, the effects are hard to maintain after termination of treatment (Chronis et al., 2006). Furthermore, behavioral interventions are relatively difficult to implement, costly both in terms of financial expenditures and time, and may be less effective than stimulants (MTA Cooperative Group, 1999). Behavioral interventions may also pose a heavy burden on the adults in charge of implementation, that is, teachers and/or parents, as behavioral interventions require long-term commitment to keep up high levels of fidelity and intensity (Chronis et al., 2001).

Having in mind the pervasive and chronic nature of ADHD and limitations of traditional treatments, there is a rising need for an approach that could be used as an alternative or an adjunct to existing treatments. Physical exercise seems to stand out as a promising intervention. Findings from behavioral, neurocognitive, neuroimaging, and physiological studies suggest that physical exercise might not only temporarily improve the symptoms of ADHD but also touch on the underlying physiological mechanisms and potentially change the developmental trajectory of the brain (Berwid & Halperin, 2012). Most of the studies examining the relationship between physical exercise and functional outcomes in ADHD population have focused on EF and motor skills. The current study provides a meta-analysis of findings available so far.Physical exercise is defined as body movement induced by contraction of skeletal muscle resulting in increased energy expenditure with the intention to improve or maintain physical fitness (Howley, 2001). Exercise interventions are relatively easy to implement and follow, because such interventions can be offered in a highly protocolled fashion, have essentially no side effects, and are much more affordable financially in comparison with medication or behavioral interventions. Maintenance of exercise after the intervention period requires commitment of the parents, but the parental involvement is not as complex as required for behavioral parent training. First of all, physical exercise interventions have to be implemented only during a limited time period during the day, whereas behavioral parent training requires parental vigilance to monitor target behaviors and take action whenever the child exhibits target behaviors. Second, it does not require continuous and costly supervision from a behavioral therapist as the procedure is easily mastered after initial demonstration. Finally, physical exercise programs could be scheduled directly after school at sports clubs or even at schools. This way it would require almost no commitment of the parents.

Several distinct lines of research provide evidence for the potential utility of physical exercise in developing new strategies for treating ADHD. One line of evidence comes from neuropharmacological studies showing that stimulant medications act as dopaminergic and noradrenergic agonists (Wigal et al., 2013). Given the effectiveness of medication in treating the core symptoms of ADHD, it has been argued that catecholamine dysfunction resulting in unbalanced and dysregulated levels of dopamine and norepinephrine might be central underlying mechanism of the disorder (Lenz, 2012; Solanto, 2002; Wigal et al., 2013). Behaviorally, these deficits translate into disturbed EF (Arnsten & Casey, 2011). As physical exercise produces an increase in dopamine and norepinephrine and affects their regulation, it has been suggested that it may improve EF in ADHD (Gapin, Labban, & Etnier, 2011; Lenz, 2012; Wigal et al., 2013). Another line of evidence comes from experimental studies on healthy adults and typically developing children showing that physical exercise improves EF (Hillman et al., 2009; Sibley & Etnier, 2003; Verburgh, Königs, Scherder, & Oosterlaan, 2014).

It is generally agreed that moderate-to-vigorous exercise intensity appears to be the most effective regarding EF and motor functioning (Best, 2010). However, optimal duration of exercise intervention is still being discussed. The effects of acute exercise on EF have been well documented (Sibley & Etnier, 2003; Verburgh et al., 2014). In their meta-analysis on the immediate effects of physical exercise, Chang, Labban et al. (2012) found that benefits of exercise on EF are larger when physical activity is maintained for at least 20 min. However, the effects of long-term / chronic physical exercise are still unknown, mainly due to the lack of available studies (Verburgh et al., 2014). Furthermore, although research on type of exercise still is scarce, it has been suggested that exercise interventions that require complex and controlled movement and cognition have greater impact on functional outcomes in comparison with repetitive aerobic or treadmill exercises (Best, 2010).

Research consistently shows that majority of children with ADHD have deficits both in EF and in gross and fine motor skills (Keiser, Schoemaker, Albaret, & Geuze, 2015; Pennington & Ozonoff, 1996; Shallice et al., 2002; Willcutt, Doyle, Nigg, Faraone, & Pennington, 2005). This finding might suggest that children with ADHD would possibly benefit from physical exercise even more than healthy children given that they have more room to improve. Motor and executive control systems follow similar developmental trajectories (Diamond, 2000) and are related to parallel frontal–subcortical circuits (Mahone & Wodka, 2008). For instance, inhibitory control is crucial for generating motor responses leading to completion of goal-directed behavior (Vasserman, Bender, & MacAllister, 2013).

EFs are defined as the top-down cognitive processes that enable organization, planning, and control of goal-directed behavior (Banich, 2009; Schreiber, Possin, Girard, & Rey-Casserly, 2013). Three core EFs are usually identified: inhibition, working memory, and cognitive flexibility (Lehto, Juujarvi, Kooistra, & Pulkkinen, 2003). A meta-analysis of 83 studies has shown that children with ADHD exhibit significant deficits in response inhibition, working memory, planning, and vigilance (Willcutt et al., 2005). Motor skills define the ability to co-ordinate movements to achieve intended goals. Children with ADHD show difficulties in sustaining simple motor acts and controlling their movements; they require more time for motor tasks in comparison with their peers (Keiser et al., 2015). Moreover, it has been reported that approximately 50% of children with ADHD also have comorbid Developmental Coordination Disorder (DCD), which indicates motor–perceptual impairment (Gillberg & Kadesjö, 2009).

Although there are already a few meta-analyses concerning the relationship between physical exercise and depression, anxiety, and mental health outcomes in general (e.g., Ahn & Fedewa, 2011; Bartley, Hay, & Bloch, 2013; Cooney et al., 2013; Stanton & Reaburn, 2014), there is no meta-analysis that would focus on the effects of physical exercise in ADHD. Here we provide a comprehensive meta-analysis of existing studies on the effects of physical exercise on the alleviation of the ADHD functional outcomes—EF and motor skills. We also aim to identify shortcomings of already used study designs and establish potential moderators.

Method

This meta-analysis was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines (Moher, Liberati, Tetzlaff, Altma, & The PRISMA Group, 2009).

The Search Process

The electronic literature search was conducted simultaneously and independently by two authors, using combinations of the following search terms: (ADHD OR “attention deficit” OR hyperactiv*) AND (sport* OR exercis* OR “physical activity”). The databases examined were PubMed, PsycINFO, Web of Science, and Cochrane Central Register of Controlled Trials. The search was a priori restricted to studies conducted in humans with age span ranging from 0 to 18 years and published in English up to January 1, 2015. Subsequently, references from retrieved studies were examined to locate any additional relevant studies. For detailed information on the study search and selection process, see the PRISMA flow chart in Figure 1.

Figure 1.

PRISMA flow chart depicting study search and selection process.

Inclusion Criteria

After duplicate removal, the titles, abstracts, and keywords of identified studies were screened for relevance. The decision on final inclusion was based on full text articles. Studies were included if (a) the study tested the effects of a physical exercise intervention on ADHD symptoms as compared with the effects of a no-exercise condition (either control group, pretest measure, or both); (b) the study design was a randomized controlled trial, quasi-experimental study or clinical trial; (c) participants had a diagnosis of ADHD of any subtype or met criteria for clinical levels of ADHD symptoms on validated rating scales; (d) the study included children and adolescents up to the age of 18; (e) the results on ADHD-related functional outcomes or core symptoms were reported; and (f) an exercise intervention was implemented.

Coding of Studies

Studies that met the inclusion criteria were coded for participant and study design characteristics by two independent authors (R.V. and L.V.). The following data were extracted from each study when possible: participant characteristics including mean age, gender, and stimulant medication use; total number of participants; diagnostic assessment procedures; study design (crossover, parallel or single-group pretest–posttest); randomization (yes/no); intervention characteristics (frequency: total number of exercise sessions; intensity: moderate, vigorous, unknown; and duration: total time in minutes); and geographic region of study. The results pertaining to ADHD functional outcomes were further categorized into two subgroups: EFs and motor skills. Other possible measures of ADHD-related outcomes, such as ADHD core symptoms (e.g., Kang, Choi, Kang, & Han, 2011), behavior (e.g., McKune, Pautz, & Lombard, 2003), social skills (e.g., Smith et al., 2013), memory (e.g., Craft, 1983), and academic performance (e.g., Pontifex, Saliba, Raine, Picchietti, & Hillman, 2013), were originally considered but later renounced due to a lack of studies (i.e., less than two studies per outcome) necessary for conducting a quantitative meta-analysis.

Effect Size

All effect sizes were based on standardized differences in performance of individuals in the exercise condition versus a no-exercise comparison, that is, control group or pretest measure. In some studies, healthy children were considered as a control group (e.g., Craft, 1983; Pontifex et al., 2013; Tantillo, Kesick, Hynd, & Dishman, 2002). However, to increase homogeneity of study designs included in the current meta-analysis, for those studies only the group of children with ADHD diagnosis was taken into account and such studies were treated as having a single-group pretest–posttest design.

Effects from individual studies were categorized into one of the two outcome domains for ADHD functional outcomes listed above. Some studies used more than one measure for representing the same construct or provided separate results for test subscales. In such cases, multiple effect sizes were averaged so that each study would contribute only one effect size per outcome domain. This was done to facilitate statistical independence of the data (Lipsey & Wilson, 2001).

Effect sizes were computed with regard to study design characteristics and available data. When possible, calculation was based on exercise and control group means and standard deviations. However, when means and standard deviations were not reported, effect sizes were calculated using F, t, p, and r values (Rosenthal, 1994). In line with recommendations by Morris (2008), Ashford, Davids, and Bennett (2009), Borenstein, Hedges, Higgins, and Rothstein (2009), and Ray and Shadish (1996), effect sizes were computed differently for specific study designs. In exceptional cases, when pre–post correlation was not reported, it was set to ρ = 0.5. This value represents moderate correlation that is consistent with approximate typical test–retest reliability for psychometric tests (Nigg, Lewis, Edinger, & Falk, 2012).

Statistical Analyses

Data were entered into Comprehensive Meta-Analysis, Version 2.2.064 (CMA, Biostat Inc, Englewood, New Jersey) for calculation of effect sizes and further analysis of results. Statistical significance was set at p < .05. The direction of the effect was calculated so that positive effect sizes would indicate beneficial effect of exercise on ADHD functional outcomes.

Prior to calculating an overall effect size estimate, each effect size was weighted by the inverse of its variance (Ashford et al., 2009). An overall cumulative meta-analysis was conducted for ADHD functional outcomes (i.e., motor skills and EF outcomes combined). Statistical significance of the pooled effect sizes was assessed using a z test.

Between-trial heterogeneity was assessed by calculating I² statistic, which expresses the ratio of observed variance between outcomes to the total observed variance in effect sizes (Huedo-Medina, Sanchez-Meca, Marin-Martinez, & Botella, 2006). I² values were interpreted as low (25%), moderate (50%), and high (75%; Higgins, Thompson, Deeks, & Altman, 2003). Following recommendations by Field and Gillett (2010), random-effects model was used, a priori, given the expected heterogeneity of sample characteristics, implementation of exercise interventions, and assessments of ADHD functional outcomes.

Potential moderators that could account for variability across studies were examined. Subgroup analyses were conducted for categorical variables, such as specific outcomes (EF, motor skills), study design (single-group pretest–posttest, crossover, and parallel groups), and exercise intensity (moderate, moderate-to-vigorous and vigorous). Meta-regression analyses were conducted for the potential continuous moderator variables: exercise duration, percentage of girls included in the sample, and mean age of participants.

Publication bias was examined based on funnel plot asymmetry inspection and quantified by the Egger’s linear regression method (Egger, Davey Smith, Schneider, & Minder, 1997). Next, in the case of publication bias, the trim-and-fill procedure was performed (Duval & Tweedie, 2000). The purpose of this method is to provide an estimate of the effect size after taking the potential bias into account (Borenstein, 2005).

Results

Description of Studies

A total of 10 studies and 54 effect sizes (total N = 300) were included in the meta-analysis. The number of participants in studies ranged from 14 to 84. The mean age of participants across all studies was 9.3 (SD = 1.9). Girls comprised 20% of the total study population. Table 1 provides design and participant characteristics of the included studies.

Table 1.

Summary of Design and Participant Characteristics of Studies Included in Meta-Analyses (Ordered by Year of Publication).

Study	Country	Study design	Randomized	n	Female (%)	Participants’ M age	Medication use (%)
Craft (1983)	The United States	Single group	No	31	0	8.6	0
Tantillo, Kesick, Hynd, and Dishman (2002)	The United States	Crossover	Yes	18	44	10.0	100
Medina et al. (2010)	Brazil	Single group	No	25	0	9.5	64
Ahmed and Mohammed (2011)	Saudi Arabia	Parallel group	Yes	84	36	13.9	n/r*
Chang, Liu et al. (2012)	China	Parallel group	Yes	40	8	10.4	50
Verret, Guay, Berthiaume, Gardiner, and Béliveau (2012)	Canada	Parallel group	No	21	10	9.1	67
Kosari, Hemayat-Talab, Arab-Ameri, and Keyhani (2013)	Iran	Parallel group	No	20	0	8.8	n/r*
Pontifex, Saliba, Raine, Picchietti, and Hillman (2013)	The United States	Crossover	Yes	20	30	9.5	0
Smith et al. (2013)	The United States	Single group	No	14	57	6.7	0
Chang et al. (2014)	China	Parallel group	No	27	15	8.4	48

Note. n/r = not reported.

Stimulant medication users comprised 41% of the total population across studies. However, medication use was not included as a moderator in meta-regression analysis due to great variability in strategies used to account for possible medication effects in individual studies (i.e., stopping vs. continuing the medication intake; different time intervals between discontinuing medication and experimental procedures, etc.).

The majority of studies provided exercise interventions of moderate intensity (n = 4), two studies implemented moderate-to-vigorous exercise approach (n = 2), one study used a vigorous exercise condition, and another one implemented both types of interventions (moderate and vigorous) for boys and girls, respectively (Tantillo et al., 2002). These effect sizes were assumed to be independent and considered as four separate studies (as in Sonuga-Barke et al., 2013). Two studies (Craft, 1983; Kosari, Hemayat-Talab, Arab-Ameri, & Keyhani, 2013) did not report on exercise intensity.

Participants received one to 36 sessions of exercise (M = 10, SD = 13.6), and total exercise intervention duration ranged from 10 to 1,540 min (M = 492, SD = 641.8). EF outcomes were reported in seven studies, motor skills outcomes in six studies. In case a study involved outcome measures for both EF and motor skills, the effect sizes were averaged within the study prior to conducting an overall meta-analysis on ADHD functional outcomes. Characteristics of exercise interventions and outcome measures are presented in Table 2.

Table 2.

Summary of Exercise Intervention and Outcome Characteristics of Studies Included in Meta-Analyses (Ordered by Year of Publication).

Study	Exercise			Number of effect sizes per outcome
Study	Intensity	Total duration (minutes)	Total number of sessions	Executive function	Motor skills
Craft (1983)	N/R	10	1	1
Tantillo, Kesick, Hynd, and Dishman (2002)	Moderate	26.23^a	1		1
	Vigorous	26.23^a	1		1
	Moderate	19.55^b	1		1
	Vigorous	19.55^b	1		1
Medina et al. (2010)	Vigorous	30	1	11
Ahmed and Mohammed (2011)	Moderate	1,540	30		1
Chang, Liu et al. (2012)	Moderate	30	1	7
Verret, Guay, Berthiaume, Gardiner, and Béliveau (2012)	Moderate-to-vigorous	1,350	30	5	3
Kosari, Hemayat-Talab, Arab-Ameri, and Keyhani (2013)	N/R	810	18		4
Pontifex, Saliba, Raine, Picchietti, and Hillman (2013)	Moderate	20	1	3
Smith et al. (2013)	Moderate-to-vigorous	1,080	36	10	2
Chang et al. (2014)	Moderate	1,440	16	1	2

Mean duration of exercise for boys.

Mean duration of exercise for girls.

Overall Effects

A significant and medium-sized effect of physical exercise on ADHD functional outcomes was found (g = 0.627, 95% confidence interval [CI] = [0.273, 0.982], p = .001; see Figure 2 for a forest plot), indicating that physical exercise has a positive effect on ADHD functional outcomes. The test of heterogeneity indicated considerable and statistically significant heterogeneity across the effect sizes included in the meta-analysis (I² = 78%, p < .001).

Figure 2.

Forest plot for random-effects meta-analysis of the functional ADHD outcomes (executive function and motor outcomes) after exercise intervention.

Publication Bias

Visual inspection of the funnel plot revealed evidence for bias toward publication of positive study results. Egger’s test confirmed this (p = .004). A trim-and-fill estimation suggested that the addition of three studies would be sufficient to result in a non-significant aggregated Hedge’s g of 0.280 (95% CI = [−0.142, 0.701]).

Subgroup Analyses

Stratification by outcome (EFs vs. motor skills) revealed significant point estimates for studies reporting EF outcomes (g = 0.535, 95% CI = [0.022, 1.048], p = .041) and for those examining motor skills outcomes (g = 0.818, 95% CI = [0.357, 1.279], p = .001). The difference between these two estimates was not significant (p = .42). High and significant heterogeneity was found both among studies reporting EF–related outcomes (I² = 84%, p < .001) and motor skills (I² = 75%, p < .001).

Studies using single-group pretest–posttest and crossover designs did not yield significantly larger effect sizes than parallel group studies. In fact, the aggregated effect sizes from studies with parallel group designs were the largest (g = 1.345, 95% CI = [0.465, 2.225], p = .003) when compared with aggregated effect sizes from studies with crossover designs (g = 0.402, 95% CI = [−0.111, 0.914], p = .125) and single-group pretest–posttest designs (g = 0.148, 95% CI = [−0.086, 0.382], p = .216). Similarly, exercise intensity did not prove to be a significant moderator either. Results are summarized in Table 3.

Table 3.

Summary of Results Grouped by Moderators: Type of Outcome Measure, Study Design, and Exercise Intensity.

Selection	K	Effect size		Test result		Heterogeneity statistics
Selection	K	g	SE	z	p	Q	df (Q)	I ²	p
Outcome
Executive functions	7	0.535	0.262	2.043	.041	38.005	6	84%	.000
Motor skills	9	0.818	0.235	3.479	.001	31.544	8	75%	.000
Study design
Crossover	5	0.402	0.261	1.536	.125	13.536	4	70%	.009
Parallel groups	5	1.345	0.449	2.996	.003	25.629	4	84%	.000
Single group	3	0.148	0.119	1.238	.216	1.645	2	0%	.439
Exercise intensity^a
Moderate	6	0.558	0.272	2.049	.040	24.852	5	80%	.000
Moderate-to-vigorous	2	0.490	0.241	2.032	.042	0.323	1	0%	.570
Vigorous	3	0.683	0.431	1.585	.113	10.276	2	81%	.006

One study (Kosari, Hemayat-Talab, Arab-Ameri, & Keyhani, 2013) did not report exercise intensity and was not included in this subgroup analysis.

Meta-Regression

A series of meta-regression analyses investigated moderating effects of exercise duration, gender distribution, and mean age of participants. It was found that longer exercise intervention duration was associated with larger effect sizes (Tau-squared = 0.234, Q_b = 6.391, p = .01). The gender distribution and mean age were not significantly associated with outcomes (p = .11 and p = .14 respectively).

Discussion

The main goal of current meta-analysis was to determine the magnitude of impact that physical exercise has on EFs and motor skills in children with ADHD. The obtained results based on 10 studies and a total of 300 participants indicate that moderate-to-vigorous physical activity moderately and significantly improves functional outcomes for children with ADHD when compared with control conditions.

Subgroup analysis revealed that the use of EFs versus motor skills as outcomes did not have a significant influence on the pooled effect size. Nevertheless, physical exercise had a somewhat larger effect on motor skills (g = 0.818) in comparison with EFs (g = 0.535). This was expected given that motor skills measured in the included studies, namely, speed, lack of motor persistence, strength, object control, and so on, are directly targeted and trained during exercise interventions. It is important to note that the effect of exercise on motor skills is large and comparable with the effect sizes generated by stimulant medication, which range from 0.83 for sustained-release formulas to 0.90 for immediate-release formulas (Faraone, Biederman, & Mick, 2006). Etnier et al. (1997) reported an even larger effect size (d = 1.47) for the effects of exercise on motor skills in healthy individuals of all ages following an exercise intervention. This finding is extremely important given the fact that children with ADHD manifest atypical motor development which impedes their ability to perform well on various daily and school tasks (Mahone & Wodka, 2008).

The effect of exercise on EFs was moderate and significant, yet the interpretation of this outcome is not straightforward. The problem is that many tasks used for measuring EFs involve a combination of EF skills (e.g., the Test of Everyday Attention requires a range of skills such as inhibition and mental shifting). Some of the studies included in the current meta-analysis reported composite scores of EF, whereas others provided separate scores for different aspects of EF. Due to a small amount of studies available, it was decided to pool all results together and use composite EF scores. However, there is evidence that not all EFs are equally sensitive to the effect of exercise (Gapin & Etnier, 2010). Therefore, in the future, it would be important to use measures that tap into one single aspect of EF to better understand which functions are affected most by physical activity.

Despite initial concerns regarding quality of included study designs, crossover and single-group pretest–posttest designs did not produce significantly different effect size estimates from methodologically more rigorous between-group studies. This finding is important as it justifies the decision to pool results from different study designs to avoid loss of data. If the moderator analyses were more reliable (e.g., if there were more studies included in the meta-analysis), this particular result could strengthen the conclusion that exercise is an effective intervention given that effects from methodologically sound studies are as large as those from methodologically weaker studies.

Furthermore, the use of moderate, moderate-to-vigorous, or vigorous exercise intensity did not significantly affect the derived effect sizes. However, an important limitation in this subgroup analysis was that some of the included studies did not consistently check and report objective physiological measures (e.g., heart rate) of participants to ensure that exercise intensity was appropriately maintained.

Research in mature populations suggests that catecholamine response to exercise in women is smaller than in men (Davis, Galassetti, Wasserman, & Tate, 2000) and it is well established that girls and boys follow different developmental trajectories. Besides, the manifestation of ADHD symptoms among the genders is different (Mahone & Wodka, 2008), although meta-regression analysis showed that studies with larger proportion of girls did not yield lower effect sizes. Also, although cognitive functioning in childhood seems to be more sensitive to exercise effects than in adolescence (Best, 2010), mean age of study was not found to be a significant moderator either.

The only variable that was established to be consistently associated with larger effect sizes was longer exercise intervention duration. It is interesting to compare this finding with results from Verburgh et al. (2014) who reported a small and insignificant effect size (d = 0.160) of chronic exercise on EFs in healthy preadolescent children, whereas the effect of acute exercise in the same population was moderate and significant (d = 0.570).

Although the findings of this meta-analysis are promising, the limitations of these results are equally important. There are several general methodological concerns. First of all, multiple subgroup and meta-regression analyses increase the risk of false negative and false positive findings due to the small number of studies they are based on (Thompson & Higgins, 2002). Therefore, the moderator analyses in this study should be regarded as explorative. Moreover, the scarce amount of studies and study design characteristics such as small study samples, lack of control groups and blinding procedures, insufficient methodological rigor, and inconsistent reporting, all limit the reliability and generalizability of the findings of this meta-analysis. Just as the literature investigating the effects of exercise on cognitive functioning in healthy populations, the studies included in this meta-analysis can also be described as being limited by a lack of consistency in methodology and by failure to use theory-driven designs (Chang, Labban et al., 2012). All in all, research in the field is still in its infancy, and there are many methodological issues to tackle.

Moreover, evidence for publication bias was detected. According to McLeod and Weisz (2004), population effects might be overestimated due to publication bias if unpublished studies that usually yield smaller effects are not included in the analysis. The current meta-analysis only included published studies. Also, we had to exclude some published studies with insignificant results due to a lack of available data and no response from the authors on the request to provide the data; thus, the detected publication bias seems a plausible and true phenomenon. When corrected for publication bias by performing a trim-and-fill procedure (Duval & Tweedie, 2000), the magnitude of the effect size dropped to 0.280. This suggests that the observed aggregated effect size (g = 0.627) is an overestimation of the true effect. However, it is problematic to distinguish between the effects of study heterogeneity and publication bias with sparse data (Peters et al., 2010).

Implications for Future Research

Given that the interaction between exercise and ADHD functional outcomes is complex and supposedly influenced by a great variety of factors, future research should aim at untangling the conditions under which exercise yields the best outcomes. Based on the results and important information that was missing in some studies included in the current meta-analysis, we would like to provide some guidelines for future studies.

First of all, research should focus on carefully conducted experimental studies aimed at identifying specific exercise parameters that generate the best results including exercise type (acute vs. chronic), intensity (moderate vs. vigorous), duration, and frequency. It is especially important to study the effect that exercise duration has on specific outcomes as it appears to be an important moderator of the effect. Another problem is the lack of reporting of objective exercise intensity measures. Moreover, a greater variety of exercise modalities (i.e., aerobic exercise, yoga, etc.) should be explored as most of the studies included in this meta-analysis used only individual treadmill exercise. For example, yoga has been shown to improve cognitive performance in typically developing children (Manjunath & Telles, 2001) and therefore might be a promising type of exercise in children with ADHD.

Second, as already mentioned here and in narrative reviews (e.g., Best, 2010; Wigal et al., 2013), to confirm hypotheses on underlying physiological mechanisms that mediate the effect of exercise, future studies should actually measure the physiological changes (e.g., catecholamine levels) that accompany exercise among children with ADHD in addition to evaluating functional outcomes, such as motor skills or EFs.

Finally and most importantly, future research should focus more on the behavioral outcomes and use standardized ADHD diagnostic tests to see how exercise affects ADHD symptoms. This step is crucial in determining the usefulness of exercise interventions in the context of empirically validated treatments such as stimulant medication and behavior modification.

Conclusion

This meta-analysis established that exercise has a moderate and significant dose-response effect on motor skills and EFs in children with ADHD. Motor skills seem to be especially sensitive to the impact of exercise.

This is an important finding given that children with ADHD have deficits in EF and motor skills. However, methodological limitations, such as the lack of methodologically sound studies investigating exercise effects in children with ADHD, make these results too weak to justify any clinical recommendations at this point. It also impedes any attempts to identify and analyze the predictors of better response to exercise.

We hope that the current meta-analysis serves its purpose in attracting more scientific attention to the topic and provides important directions for future research by establishing a moderate positive effect of exercise, highlighting shortcomings of available studies and methodological inconsistencies between these studies. If the effects of exercise for ADHD are better substantiated in the future, we might be looking at a powerful complementary or alternative treatment.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Author Biographies

Ruta Vysniauske is a graduate of MSc clinical psychology at Leiden University (the Netherlands) and clinical child psychologist at Vilnius University Children’s Hospital, Affiliate of Vilnius University Hospital Santariskiu Klinikos (Lithuania).

Lot Verburgh is a post-doctoral researcher at the Clinical Neuropsychology section of Vrije Universiteit Amsterdam (the Netherlands). Her research concentrates on the relationship between sports, physical activity, and neurocognitive functioning in children.

Jaap Oosterlaan is a professor of clinical neuropsychology at the Vrije Universiteit Amsterdam (the Netherlands) and the director of Follow Me, a follow-up program at the Emma’s Children’s Hospital Academic Medical Center and Vrije Universiteit Medical Center Amsterdam. His research focuses on disruptive behavior disorders and medical conditions affecting the central nervous system in children.

Marc L. Molendijk is a post-doctoral researcher and assistant professor at the Clinical Psychology unit of Leiden University (the Netherlands) and the Leiden Institute of Brain and Cognition (Leiden University Medical Center, the Netherlands). His main research interest is on the relation between biology and behavior.

References

Ahmed

G. M.

Mohamed

(2011). Effect of regular aerobic exercises on behavioral, cognitive and psychological response in patients with attention deficit-hyperactivity disorder. Life Science Journal, 8(2), 366-371.

Ahn

Fedewa

A. L.

(2011). A meta-analysis of the relationship between children’s physical activity and mental health. Journal of Pediatric Psychology, 36, 385-397.

Arnsten

A. F. T.

Casey

B. J.

(2011). Prefrontal cortical organization and function: Implications for externalizing disorders. Biological Psychiatry, 69, 1131-1132.

Ashford

Davids

Bennett

(2009). Difference-based meta-analytic procedures for between-participant and/or within-participant designs: A tutorial review for sports and exercise scientists. Journal of Sports Sciences, 27, 237-255.

Banich

M. T.

(2009). Executive function: The search for an integrated account. Current Directions in Psychological Science, 18, 89-94.

Barkley

R. A.

(2006). History. In Barkley

R. A.

(Ed.), Attention-deficit hyperactivity disorder: A handbook for diagnosis and treatment (3rd ed., pp. 3-76). New York, NY: Guilford Press.

Bartley

C. A.

Hay

Bloch

M. H.

(2013). Meta-analysis: Aerobic exercise for the treatment of anxiety disorders. Progress in Neuro-Psychopharmacology & Biological Psychiatry, 45, 34-39.

Berwid

O. G.

Halperin

J. M.

(2012). Emerging support for a role of exercise in attention-deficit/hyperactivity disorder intervention planning. Current Psychiatry Reports, 14, 543-551.

Best

J. R.

(2010). Effects of physical activity on children’s executive function: Contributions of experimental research on aerobic exercise. Developmental Review, 30, 331-351.

10.

Boonstra

A. M.

Oosterlaan

Sergeant

J. A.

Buitelaar

J. K.

(2005). Executive functioning in adult ADHD: A meta-analytic review. Psychological Medicine, 35, 1097-1108.

11.

Borenstein

(2005). Software for publication bias. In Rothenstein

H. R.

Sutton

A. J.

Borenstein

(Eds.), Publication bias in meta-analysis: Prevention, assessment and adjustments (pp. 193-220). Chichester, UK: John Wiley & Sons, Ltd.

12.

Borenstein

Hedges

L. V.

Higgins

J. P. T.

Rothstein

H. R.

(2009). Introduction to meta-analysis. Chichester, UK: John Wiley & Sons, Ltd.

13.

Chang

Y. K.

Labban

J. D.

Gapin

J. I.

Etnier

J. L.

(2012). The effects of acute exercise on cognitive performance: A meta-analysis. Brain Research, 1453, 87-101.

14.

Chang

Y. K.

Liu

H. H.

Lee

Y. H.

(2012). Effect of acute exercise on executive function in children with attention deficit hyperactivity disorder. Archives of Clinical Neuropsychology, 27, 225-237.

15.

Chang

Y. K.

Hung

C. L.

Huang

C. J.

Hatfield

B. D.

Hung

T. M.

(2014). Effects of an aquatic exercise program on inhibitory control in children with ADHD: a preliminary study. Archives of Clinical Neuropsychology, 29, 217-223.

16.

Chronis

A. M.

Fabiano

G. A.

Gnagy

E. M.

Wymbs

Burrows-MacLean

Pelham

W. E.

(2001). Comprehensive, sustained behavioural and pharmacological treatment for ADHD: A case study. Cognitive and Behavioural Practice, 8, 346-359.

17.

Chronis

A. M.

Jones

H. A.

Raggi

V. L.

(2006). Evidence-based psychosocial treatments for children and adolescents with attention-deficit/hyperactivity disorder. Clinical Psychology Review, 26, 486-502.

18.

Connor

D. F.

(2006). Stimulants. In Barkley

R. A.

(Ed.), Attention-deficit hyperactivity disorder: A handbook for diagnosis and treatment (3rd ed., pp. 608-648). New York, NY: Guilford Press.

19.

Cooney

G. M.

Dwan

Greig

C. A.

Lawlor

D. A.

Rimer

Waugh

F. R.

. . . Mead

G. E.

(2013). Exercise for depression. Cochrane Database of Systematic Reviews, 9, CD004366.

20.

Craft

D. H.

(1983). Effect of prior exercise on cognitive performance tasks by hyperactive and normal young boys. Perceptual & Motor Skills, 56, 979-982.

21.

Davis

S. N.

Galassetti

Wasserman

D. H.

Tate

(2000). Effects of gender on neuroendocrine and metabolic counterregulatory responses to exercise in normal man. Journal of Clinical Endocrinology & Metabolism, 85, 224-230.

22.

Diamond

(2000). Close interrelation of motor development and cognitive development and of the cerebellum and prefrontal cortex. Child Development, 71, 44-56.

23.

Duval

Tweedie

(2000). Trim and fill: A simple funnel-plot-based method of testing and adjusting for publication bias in meta-analysis. Biometrics, 56, 455-463.

24.

Egger

Davey Smith

Schneider

Minder

(1997). Bias in meta-analysis detected by a simple, graphical test. British Medical Journal, 315, 629-634.

25.

Etnier

J. L.

Salazar

Landers

D. M.

Petruzzello

S. J.

Han

Nowell

(1997). The influence of physical fitness an exercise upon cognitive functioning: A meta-analysis. Journal of Sport & Exercise Psychology, 19, 249-277.

26.

Faraone

S. V.

Biederman

Mick

(2006). The age-dependent decline of attention deficit hyperactivity disorder: A meta-analysis of follow-up studies. Psychological Medicine, 36, 159-165.

27.

Faraone

S. V.

Pucci

Coghill

(2009). Pharmacotherapy for attention-deficit-hyperactivity disorder. European Psychiatric Review, 2(1), 42-52.

28.

Field

A. P.

Gillett

(2010). Expert tutorial: How to do a meta-analysis. British Journal of Mathematical and Statistical Psychology, 63, 665-694.

29.

Gapin

J. I.

Etnier

J. L.

(2010). The relationship between physical activity and executive function performance in children with attention-deficit hyperactivity disorder. Journal of Sport & Exercise Psychology, 32, 753-763.

30.

Gapin

J. I.

Labban

J. D.

Etnier

J. L.

(2011). The effects of physical activity on attention deficit hyperactivity disorder symptoms: The evidence. Preventive Medicine, 52, S70-S74.

31.

Gillberg

Kadesjö

(2009). ADHD with developmental coordination disorder. In Brown

T. E.

(Ed.), ADHD comorbidities: Handbook for ADHD complications in children and adults (1st ed., pp. 306-314). Washington, DC: American Psychiatric Press.

32.

Gordon Millichap

(2011). Attention deficit hyperactivity disorder handbook: A physician’s guide to ADHD (2nd ed.). New York, NY: Springer Science & Business Media.

33.

Higgins

J. P.

Thompson

S. G.

Deeks

J. J.

Altman

D. G.

(2003). Measuring inconsistency in meta-analyses. British Medical Journal, 327, 557-560.

34.

Hillman

C. H.

Pontifex

M. B.

Raine

L. B.

Castelli

D. M.

Hall

E. E.

Kramer

A. F.

(2009). The effect of acute treadmill walking on cognitive control and academic achievement in preadolescent children. Neuroscience, 159, 1044-1054.

35.

Howley

E. T.

(2001). Type of activity: Resistance, aerobic and leisure versus occupational physical activity. Medicine & Science in Sports & Exercise, 33, S364-S369.

36.

Huedo-Medina

T. B.

Sanchez-Meca

Marin-Martinez

Botella

(2006). Assessing heterogeneity in meta-analysis: Q statistic or I2 index? Psychological Methods, 11, 193-206.

37.

Kang

K. D.

Choi

J. W.

Kang

S. G.

Han

D. H.

(2011). Sports therapy for attention, cognitions and sociality. Sports Medicine, 32, 953-959.

38.

Keiser

M. L.

Schoemaker

M. M.

Albaret

J. M.

Geuze

R. H.

(2015). What is the evidence of impaired motor skills and motor control among children with attention deficit hyperactivity disorder (ADHD)? Systematic review of the literature. Research in Developmental Disabilities, 36, 338-357.

39.

Kosari

Hemayat-Talab

Arab-Ameri

Keyhani

(2013). The effect of physical exercise on the development of gross motor skills in children with attention deficit/hyperactivity disorder. Zahedan Journal of Research in Medical Sciences, 15(2), 74-78.

40.

Lehto

J. E.

Juujarvi

Kooistra

Pulkkinen

(2003). Dimensions of executive functioning: Evidence from children. British Journal of Developmental Psychology, 21, 59-80.

41.

Lenz

T. L.

(2012). A pharmacological/physiological comparison between ADHD medications and exercise. American Journal of Lifestyle Medicine, 6, 306-308.

42.

Lipsey

M. W.

Wilson

(2001). Practical meta-analysis (Applied Social Research Methods). Thousand Oaks, CA: SAGE.

43.

Mahone

E. M.

Wodka

E. L.

(2008). The neurobiological profile of girls with ADHD. Developmental Disabilities Research Reviews, 14, 276-284.

44.

Manjunath

N. K.

Telles

(2001). Improved performance in the tower of London test following yoga. Indian Journal of Physiology and Pharmacology, 45, 351-354.

45.

McKune

A. J.

Pautz

Lombard

(2003). Behavioural response to exercise in children with attention-deficit/hyperactivity disorder. Sports Medicine, 15(3), 17-21.

46.

McLeod

B. D.

Weisz

J. R.

(2004). Using dissertations to examine potential bias in child and adolescent clinical trials. Journal of Consulting and Clinical Psychology, 72, 235-251.

47.

Medina

J. A.

Netto

T. L. B.

Muszkat

Medina

A. C.

Botter

Orbetelli

. . . Miranda

M. C.

(2010). Exercise impact on sustained attention of ADHD children, methylphenidate effects. Attention Deficit and Hyperactivity Disorders, 2, 49-58.

48.

Moher

Liberati

Tetzlaff

Altma

D. G.

The PRISMA Group. (2009). Preferred reporting items for systematic reviews and meta-analyses: The PRISMA Statement. Annals of Internal Medicine, 151, 264-269.

49.

Morris

S. B.

(2008). Estimating effect sizes from pretest-posttest-control group designs. Organizational Research Methods, 11, 364-386.

50.

MTA Cooperative Group. (1999). 14-month randomized clinical trial of treatment strategies for attention deficit hyperactivity disorder. Archives of General Psychiatry, 56, 1073-1086.

51.

Nigg

J. T.

Lewis

Edinger

Falk

(2012). Meta-analysis of attention-deficit/hyperactivity disorder or attention-deficit/hyperactivity disorder symptoms, restriction diet, and synthetic food color additives. Journal of the American Academy of Child & Adolescent Psychiatry, 51, 86-97.

52.

Pelham

W. E.

Jr. Fabiano

G. A.

(2008). Evidence-based psychosocial treatments for attention-deficit/hyperactivity disorder. Journal of Clinical Child & Adolescent Psychology, 37, 184-214.

53.

Pennington

B. F.

Ozonoff

(1996). Executive functions and developmental psychopathology. Journal of Child Psychology and Psychiatry, 37, 51-87.

54.

Peters

J. L.

Sutton

A. J.

Jones

D. R.

Abrams

K. R.

Rushton

Moreno

S. G.

(2010). Assessing publication bias in meta-analyses in the presence of between-study heterogeneity. Journal of the Royal Statistical Society, 173, 575-591.

55.

Piek

J. P.

Pitcher

T. M.

Hay

D. A.

(1999). Motor coordination and kinaesthesis in males with attention-deficit-hyperactivity disorder. Developmental Medicine & Child Neurology, 41, 159-165.

56.

Pontifex

M. B.

Saliba

B. J.

Raine

L. B.

Picchietti

D. L.

Hillman

C. H.

(2013). Exercise improves behavioural, neurocognitive, and scholastic performance in children with attention-deficit/hyperactivity disorder. The Journal of Pediatrics, 162, 543-551.

57.

Ray

J. W.

Shadish

W. R.

(1996). How interchangeable are different estimators of effect size? Journal of Consulting and Clinical Psychology, 64, 1316-1325.

58.

Rosenthal

(1994). Parametric measures of effect size. In Cooper

Hedges

L. V.

(Eds.), The handbook of research synthesis (pp. 231-244). New York, NY: Russell Sage Foundation.

59.

Schreiber

J. E.

Possin

K. L.

Girard

J. M.

Rey-Casserly

(2013). Executive function in children with attention deficit/hyperactivity disorder: The NIH EXAMINER battery. Journal of the International Neuropsychological Society, 20, 41-51.

60.

Shallice

Marzocchi

G. M.

Coser

Del Savio

Meuter

R. F.

Rumiati

R. I.

(2002). Executive function profile of children with attention deficit hyperactivity disorder. Developmental Neuropsychology, 21, 43-71.

61.

Sibley

B. A.

Etnier

J. L.

(2003). The relationship between physical activity and cognition in children: A meta-analysis. Pediatric Exercise Science, 15, 243-256.

62.

Smith

A. L.

Hoza

Linnea

McQuade

J. D.

Tomb

Vaughn

A. J.

. . . Hook

(2013). Pilot physical activity intervention reduces severity of ADHD symptoms in young children. Journal of Attention Disorders, 17, 70-82.

63.

Solanto

M. V.

(2002). Dopamine dysfunction in AD/HD: Integrating clinical and basic neuroscience research. Behavioural Brain Research, 130, 65-71.

64.

Sonuga-Barke

E. J. S.

Brandeis

Cortese

Daley

Ferrin

Holtmann

. . . European ADHD Guidelines Group. (2013). Nonpharmacological interventions for ADHD: Systematic review and meta-analyses of randomized controlled trials of dietary and psychological treatments. The American Journal of Psychiatry, 170, 275-289.

65.

Stanton

Reaburn

(2014). Exercise and the treatment of depression: A review of the exercise program variables. Journal of Science and Medicine in Sport, 17, 177-182.

66.

Tantillo

Kesick

C. M.

Hynd

G. W.

Dishman

R. K.

(2002). The effects of exercise on children with attention-deficit/hyperactivity disorder. Medicine & Science in Sports & Exercise, 34, 203-212.

67.

Thompson

S. G.

Higgins

J. P. T.

(2002). How should meta-regression analyses be undertaken and interpreted? Statistics in Medicine, 21, 1559-1573.

68.

Vasserman

Bender

H. A.

MacAllister

W. S.

(2013). Motor skills development in children with inattentive versus combined subtypes of ADHD. Applied Neuropsychology: Child, 3, 145-151.

69.

Verburgh

Königs

Scherder

E. J.

Oosterlaan

(2014). Physical exercise and executive functions in preadolescent children, adolescents and young adults: A meta-analysis. British Journal of Sports Medicine, 48, 973-979.

70.

Verret

Guay

M. C.

Berthiaume

Gardiner

Béliveau

(2012). A physical activity program improves behaviour and cognitive functions in children with ADHD: An exploratory study. Journal of Attention Disorders, 16, 71-80.

71.

Weiss

Hechtman

(1993). Hyperactive children grown up (2nd ed.). New York, NY: Guilford Press.

72.

Wigal

S. B.

Emmerson

Gehricke

J. G.

Galassetti

(2013). Exercise: Applications to childhood ADHD. Journal of Attention Disorders, 17, 279-290.

73.

Willcutt

E. G.

(2010). Attention-deficit/hyperactivity disorder. In Yeates

K. O.

Ris

M. D.

Taylor

H. G.

Pennington

B. F.

(Eds.), Pediatric neuropsychology: Research, theory, and practice (2nd ed., pp. 393-418). New York, NY: Guilford Press.

74.

Willcutt

E. G.

Doyle

A. E.

Nigg

J. T.

Faraone

S. V.

Pennington

B. F.

(2005). Validity of the executive function theory of attention-deficit/hyperactivity disorder: A meta-analytic review. Biological Psychiatry, 57, 1336-1346.