Sweat it out? The effects of physical exercise on cognition and behavior in children and adults with ADHD: a systematic literature review

As attention–deficit/hyperactivity disorder (ADHD) is one of the most frequently diagnosed developmental disorders in childhood, effective yet safe treatment options are highly important. Recent research introduced physical exercise as a potential treatment option, particularly for children with ADHD. The aim of this review was to systematically analyze potential acute and chronic effects of cardio and non-cardio exercise on a broad range of functions in children with ADHD and to explore this in adults as well. Literature on physical exercise in patients with ADHD was systematically reviewed based on categorizations for exercise type (cardio versus non-cardio), effect type (acute versus chronic), and outcome measure (cognitive, behavioral/socio-emotional, and physical/(neuro)physiological). Furthermore, the methodological quality of the reviewed papers was addressed. Cardio exercise seems acutely beneficial regarding various executive functions (e.g., impulsivity), response time and several physical measures. Beneficial chronic effects of cardio exercise were found on various functions as well, including executive functions, attention and behavior. The acute and chronic effects of non-cardio exercise remain more questionable but seem predominantly positive too. Research provides evidence that physical exercise represents a promising alternative or additional treatment option for patients with ADHD. Acute and chronic beneficial effects of especially cardio exercise were reported with regard to several cognitive, behavioral, and socio-emotional functions. Although physical exercise may therefore represent an effective treatment option that could be combined with other treatment approaches of ADHD, more well-controlled studies on this topic, in both children and adults, are needed.


Introduction
Attention-deficit/hyperactivity disorder (ADHD) is one of the most common neurodevelopmental disorders diagnosed in childhood, with symptoms continuing into adolescence and adulthood [American Psychiatric Association (APA) 2013]. The vast consequences of ADHD on individuals' lives stress the importance of proper and scientifically based treatment options. In the past decades, the prescription of stimulant medication has substantially increased and has become a matter of social debate . In this literature review, we systematically reviewed the potential benefits of an alternative or additional treatment option for ADHD, namely, physical exercise.

Symptoms and cognition of patients with ADHD
The Diagnostic and Statistical Manual of Mental Disorders (DSM-5) describes inattention, impulsivity, impaired inhibition, and hyperactivity as the most common ADHD symptoms [American Psychiatric Association (APA) 2013], giving rise to problems in everyday life, social, academic, and occupational domains (Gapin et al. 2011;Hoza et al. 2005;Loe and Feldman 2007;Uekerman et al. 2010). Furthermore, more specific deficits in higher cognitive functions (executive functions) have often been demonstrated in ADHD (Barkley 2004). These functions facilitate self-monitored and self-regulated goal-directed behavior (Gioia et al. 2001). For example, deficits have frequently been reported in working memory, response inhibition, cognitive flexibility, risk taking, and planning (Barkley, 1997;Groen et al. 2013;Pliszka et al. 2006;Semrud-Clikeman et al. 2008;Toplak et al. 2009). Impairments of cognitive flexibility, in particular, may play an important role in behavior regulation and may underlie the often seen perseverative, non-adaptive, and dysregulated behavior in children with ADHD (Lezak 2004).

Advantages and disadvantages of pharmacological treatment of ADHD
For several decades, stimulant medication (mainly methylphenidate, MPH) has constituted the most frequently applied treatment of ADHD (Huang and Tsai 2011;Pliszka et al. 2000;Tucha et al. 2006). On a neurobiological level, MPH increases dopamine and norepinephrine levels in the prefrontal cortex (PFC) (Oades et al. 2005;Pliszka 2005) and generally enhances central nervous system (CNS) activity (Semrud-Clikeman et al. 2008). On a behavioral level, an increase in alertness and a decrease in antisocial behaviors (e.g., aggression) and impulsivity have been reported following stimulant drug treatment (Rhodes et al. 2006;Semrud-Clikeman et al. 2008;Wilson et al. 2006). Stimulants have also been shown to positively affect various aspects of cognition, including domains of attention (e.g., mental alertness), executive functioning, memory, and visuospatial functioning as well as cognition related to self-regulation (Huang and Tsai 2011;Rhodes et al. 2006;Semrud-Clikeman et al. 2008;Tucha et al. 2006;Wilson et al. 2006). Furthermore, MPH was found to improve quality of life and academic achievement (Huang and Tsai 2011).
Despite the existence of well-elaborated evidence-based guidelines (NICE Guidelines on ADHD 2008) and the fact that efficacy and safety of stimulants have been proved in a considerable number of studies and meta-analyses (e.g., Maia et al. 2014;Rubia et al. 2014;Thapar and Cooper 2016), uncertainties still exist concerning stimulant drug treatment for ADHD. These include sub-optimal treatment in clinical practice (Hodgkins et al. 2013), uncertainty about the physical safety and potential long-term and side effects of stimulant medication (Biederman et al. 1991;Buitelaar and Medori 2010;Dela Pena and Cheong 2013;Owens et al. 2000;Sonuga-Barke et al. 2009;Stein et al. 2003;Taylor et al. 2004; Van der Heijden et al. 2006;Wilens et al. 2008) and a relatively high non-response rate of 10-25 % (Banaschewski et al. 2006;Taylor et al. 2004). Besides pharmacological approaches to ADHD treatment, there are non-pharmacological treatment options, such as cognitive trainings, neurofeedback, and behavioral therapy, diets, and combined interventions. Studies on these alternative treatments are relatively scarce, but do show improvements in ADHD symptoms with small to medium effect sizes (see for a meta-analysis: Sonuga-Barke et al. 2013). However, many of these studies are flawed with methodological weaknesses, and the beneficial effects often disappear when controlling for blinded assessment.

General effects of physical exercise in relation to ADHD
Recently, physical exercise has been proposed as an alternative or additional treatment for ADHD. Particularly for children, physical exercise has been raised as a safe and effective ADHD symptom management method, having beneficial effects on both cognition and behavior (Gapin et al. 2011). Physical exercise can be subdivided into several categories. First, cardio exercise raises the heart rate, stimulates perspiration, and makes one out of breath. This category comprises activities such as running, cycling, dancing and treadmill exercise. Cardio exercise has been associated with several ameliorations of physical and cognitive functions (e.g., executive functions) in healthy children (Best 2010;Hill et al. 2011). The second category is non-cardio exercise which is tranquil, e.g., yoga and tai chi, and has also been linked to both physical and cognitive gains in healthy children (Field 2012). Another categorization of exercise is based on the duration of effects, which can be classified as either acute (i.e., effect measurable shortly after the physical activity) or chronic (i.e., effects extending after a considerable period of rest).
Well-known general positive effects of exercise include the enhancement of overall physical fitness, growth and density of bone minerals, and the reduction of obesity and inflammation (Field 2012). In addition, positive effects of exercise on psychological and cognitive functions have been described (Hill et al. 2011;Hillman et al. 2008). Sibley and Etnier (2003) observed acute as well as chronic effects of various cardio and non-cardio exercises on perceptual skills, intelligence, academic achievement, developmental level and performance on verbal and mathematic tests in children and adolescents (4-18 years). Furthermore, improvements of executive functions of children have been demonstrated following cardio exercise (Best 2010). These benefits in executive functions were strongest in exercise types that demand cognitive engagement, entailing cooperation awareness, anticipation, task-demands and team-sport strategic thinking, compared with exercise types that do not require such cognitive involvement (Best 2010).
It is assumed that the mechanisms by which physical exercise provokes beneficial psychological and/or cognitive effects are an increase in vagal activity, anti-pain, and anti-depression neurotransmitters (e.g., serotonin) as well as a decrease in stress hormones (Field 2012). On the one hand, acute exercise (i.e., cognitive examination shortly upon completing an exercise bout) is thought to enhance cognitive functioning by immediate neurochemical responses. For example, neurological effects of exercise as described by Lojovich (2010) and Knaepen et al. (2010) include increases in brain-derived neurotrophic factor, levels of synaptic proteins, glutamate receptors, and the availability of insulin-like growth factor, which altogether seem to contribute to cell proliferation and neural plasticity . The cognitive enhancements are, furthermore, related to increased arousal and blood flow in the PFC (Lambourne and Tomporowski 2010). On the other hand, chronic exercise (i.e., cognitive examination after several weeks of regular physical exercise) could indirectly promote more permanently improved cognition and learning through morphological brain changes and improved cardio-respiratory functioning (Best 2010). This cognitive enhancement and improvements in the PFC are possibly cumulative and, therefore, lasting over time (Pesce 2009). Long-term cognitive benefits of regular cardio exercise are, furthermore, suggested to promote the accelerated cerebellar development in children (Bishop 2007) and the processes of healthier cerebrovascular aging, such as positive influences on blood flow, cell maintenance, and circulating catecholamines (Bolduc et al. 2013).
Reviewing the effects of physical exercise in patients with ADHD Seven prior literature reviews addressing the effects of physical exercise in children with ADHD have led to promising conclusions. Gapin et al. (2011) concluded in their review of six studies that physical exercise has both acute and chronic positive effects on behavioral and cognitive measures in children with ADHD. The authors suggested physical exercise as a potential supplement to medication. Archer and Kostrzewa (2012) reviewed three studies on this topic and reported a reduction of ADHD symptoms following moderate intensity cardio exercise, including impulse control, inattentiveness, stress, negative affect (e.g., depression), anxiety, and bad conduct. These reductions were related to an increased level of brainderived neurotrophic factor, which is typically reduced in patients with ADHD. Two biopsychologically oriented reviews describing the relation between intense cardio exercise treatment and enhancement of cognitive and behavioral functioning of children with ADHD suggested positive influences on the structure, function, and growth of the brain as underlying mechanisms (Berwid and Halperin 2012;Halperin and Healey 2011). A review by Wigal and others (2013) on the effect of physical exercise on the physiology of childhood ADHD stressed that physical activity and stimulant medication both affect the dopaminergic and noradrenergic systems. One recently published review described ameliorated executive functions and social functioning and decreased levels of ADHD symptoms following the short-term aerobic exercise (Cerillo-Urbina et al. 2015). A final recently published review paper stated that especially mixed exercise programs appear beneficial for ADHD symptomatology and fine motor skills (Neudecker et al. 2015).
Despite the high informative value of the existing reviews, they present with several shortcomings. First, not all currently available studies on cognitive and behavioral gains of physical exercise in children with ADHD were considered so far as the field is rapidly expanding. Second, not all of the former reviews systematically categorized the specific types of exercise and outcome measures, preventing clear conclusions with regard to effectiveness and implementation (concerning for instance exercise type and frequency and expected advantages). Third, the value of the available reviews is limited, since information concerning the cognitive and behavioral outcome measures has not been considered sufficiently. Recently, the literature on this topic is growing, making frequent, critical and updated overviews necessary. The current literature review, therefore, aimed to provide a systematic, up-to-date overview on the effects of cardio and non-cardio exercise types on cognitive, behavioral/socio-emotional, and physical/ (neuro)physiological outcome measures in children with ADHD, thereby also addressing the duration of these effects (acute or chronic). To better interpret and describe the existing literature, a quality screening of the included studies was performed. Furthermore, we described similar literature in adults with ADHD in an exploratory way, providing a more lifespan perspective on the effects of physical exercise in patients with ADHD.

Literature search
A literature search of scientific published literature in the databases of PubMed, PsycINFO and Web of Knowledge revealed 25 research articles published before April 2016 that described exercise in relation to cognition and behavior in children with ADHD and 4 articles related to this topic in adults with ADHD. The keywords ''ADHD, '' ''children,'' ''adults,'' ''physical,'' and ''exercise'' were combined with words related to physical exercise or outcome measures, such as ''activity, '' ''sports,'' ''yoga,'' ''neuropsychology,'' ''cognition,'' ''executive,'' ''functioning,'' and ''treatment''. Only articles, including an ADHD group, describing some form of exercise and that were written in English, were included. In addition, relevant articles cited in the included papers were added. Through a systematic categorization of these articles, the outcomes were organized within a descriptive table (see Table 1). To provide additional grounds for comparison, effects sizes were calculated when possible (if not already presented in the original paper) by using Cohen's d and its interpretation classification index (Cohen 1988), describing small effects (0.2 B d \ 0.5), medium effects (0.5 B d \ 0.8), and large effects (d C 0.8). Effect sizes or ranges are shown in Table 1. However, there was unfortunately a profound variation in the characteristics of the study samples of the reviewed articles (e.g., medication use and control conditions), the type of functions assessed, the (un)availability of outcome statistics, the type of (neuropsychological) tests applied, and the comparisons performed (e.g., within or between subjects or mixed designs). We, therefore, deemed it unjustified to reliably compare effect sizes across studies.

Analysis
To provide a systematic overview of the existing literature, categorizations were made for three variables; exercise type (cardio versus non-cardio), duration of effects (acute versus chronic), and outcome measures (cognitive, behavioral/socio-emotional, and physical/(neuro) physiological).
Acute effects of exercise were defined as the effects of physical exercise immediately after the workout, with a maximum of 24 h; hence, the outcome measures stemmed from the same full day as the exercise intervention. Chronic effects of exercise were defined as outcomes lasting longer than 24 h after the exercise intervention, with assessments after one to 10 weeks, depending on the follow-up period of included studies. This classification into acute and chronic effects was made, because physical aftereffects of exercise were thought to last for the first full day but to diminish after a resting period during the night. Persisting effects after nocturnal rest and recovery are considered to be long-lasting.
Cardio exercise included all types of workout that lead to an increased heart rate and oxygen use and that are performed for a relatively long duration, such as (treadmill) running, (ergo meter) cycling, swimming, and jumping. Any exercise type that is performed at a lower energy level and does not intensely increase the heart rate was classified as non-cardio exercise, including yoga, walking, and playground activity.
Outcome measures of the reviewed papers were classified into one of three categories, namely, ''cognitive outcome measures'' [including intelligence scores and (neuropsychological) tests for attention, planning, inhibition and memory], ''behavioral and socio-emotional outcomes'' (comprising parent and/or teacher questionnaires on the behavioral functioning of children, e.g., ADHD symptoms), and ''physical and (neuro)physiological outcomes'' (e.g., sheer physical/physiological effects).
Finally, we screened the included papers for their methodological quality to weigh the descriptions of the studies and the conclusions of this review. Two independent raters classified the following four important quality determinants of treatment studies as adequate (A), inadequate (IA), not applicable (NA) or not reported (NR). First, ADHD-diagnosis was assessed by standardized measures (e.g., DSM, ICD, ARS, Connor's rating scale) to diagnose or operationally define behaviors and symptoms of participants. Second, sample size: for detecting a medium effect size (f = 0.25) in the most commonly employed design of the studies included in this review (a within-between group interaction in a repeated measures ANOVA with two groups (e.g., ADHD versus control) and one within-subjects variable (e.g., pre-measurement versus post-measurement), 17 participants are needed per group (with a power of 0.80 and an alpha of 0.05). When more withinsubjects variables are added, a fewer participants are necessary (e.g., adding a low versus high intensity exercise condition reduces the needed participants to 12 per group) but when a control group is omitted, more participants are necessary to demonstrate a pre-post effect (34 participants are needed). Third, control condition/group: either a between group or a within subject comparison was made comparing exercise to some other condition without exercise. Last, control for medication use: either all participants were on medication, naïve for medication or off medication during the treatment/control condition, or it was checked whether medication influenced the results (e.g., by comparing subgroups).

Cardio exercise: acute effects
After cardio exercise (e.g., treadmill running, cycling), several positive effects on (higher) cognitive functioning of children with ADHD were found. Response inhibition, cognitive control, attention allocation (based on event-related brain potentials; Pontifex et al. 2013), cognitive flexibility, processing speed, and vigilance were found to S6 A. E. Den Heijer et al. Sweat it out? The effects of physical exercise on cognition and behavior in children and… S7 ADHD in the exercise group showed improved accuracy (Go/No-Go Task) with the no-gostimulus (d = 0.9) as well as coordination of motor skills, whereas accuracy with the no-gostimulus remained unchanged over time in the control group (d = -0.04). Physical exercise was therefore linked to increased levels of restraint/ behavioral inhibition in children with ADHD Unchanged/reversed effect: go-stimuli scores were respectively longer and less accurate than in the control group, with no effect of time P/NP: Post vs. pre: The physical exercise group and not the control group achieved increased levels of hand-eye-coordination in throwing and bilateral hand-eye-coordination and dexterity (as measured by target throwing scores and bead moving scores of the Basic Motor Ability Test-Revised) S12 A. E. Den Heijer et al. Both ADHD groups showed worse total motor proficiency scores than TD children; but the ADHD training group performed better than the ADHD non-training group at total motor proficiency, fine manual control, manual coordination, body coordination, strength and agility, manual dexterity and bilateral coordination (Bruininks-Oseretsky Test of Motor Proficiency, second edition) Non-significant: fine motor precision and integration, upper-limb coordination, balance, running speed and agility and strength P/NP: Post vs. pre, within-group: The ADHD training group showed improvements in all motor and fitness measures. The ADHD and TD non-training groups improved in some but not all motor skills or fitness measures In conclusion: children with ADHD thus exhibit low motor proficiency, cardiovascular fitness and flexibility at baseline, and these functions benefit from physical activity wave activity -theta-and decreased fast wave activity -beta-in EEG studies which is linked to ADHD symptoms) displayed reduced theta power waves after neurofeedback and medication interventions but not after physical activity on an EEG-analysis. The medication and neurofeedback groups showed reduced theta power post-treatment during a resting condition. Solely the medication group displayed this reduction in an effortful task condition S14 A. E. Den Heijer et al. Sweat it out? The effects of physical exercise on cognition and behavior in children and… S15 Sweat it out? The effects of physical exercise on cognition and behavior in children and… S17 be improved in children with ADHD when assessed immediately after physical exercise (Chang et al. 2012;Hartanto et al. 2015;Medina et al. 2010;Pontifex et al. 2013;Smith et al. 2013). Academic performance was also found to be improved in both children with ADHD and healthy children following physical exercise, including enhanced levels of reading comprehension and arithmetics (Pontifex et al. 2013). However, not all assessed aspects of academic and cognitive functioning were sensitive to the effects of physical exercise, with unchanged performance levels of verbal short-term and working memory, sub-indexes of executive function performances (e.g., perseverative responses), spelling, reading, and color naming (Chang et al. 2012;Medina et al. 2010;Pontifex et al. 2013;Smith et al. 2013;Ziereis and Jansen 2015). With respect to behavioral and school-day functioning, parents and teachers reported children with ADHD to display overall improvements after physical activity (e.g., diminished interruption and unintentional aggression). No effects, however, were found concerning intentional aggression, language use and following directions (Smith et al. 2013). The authors did not report whether the raters were blind to the treatment of the child.
Finally, regarding physical/(neuro)physiological acute effects of cardio exercise, improved motor impersistence (defined as the inability to sustain motor acts, such as keeping the mouth open, protruding the eyes, centrally fixing the eyes or shutting the eyelids) was reported in boys with ADHD. This effect was, however, not found in girls, lacking a clear explanation (Tantillo et al. 2002). Pontifex et al. (2013) described physical exercise-induced improvements in neurocognitive function and response production as measured by event-related potentials (ERPs), cautiously linking this to behavioral and academic gains. Wigal et al. (2013) demonstrated abnormal physiological after effects of physical exercise in children with ADHD (including more blunted (nor)epinephrine responses and no dopamine increase) which are in contrast to the effects found in healthy controls. Another study stressed that children with ADHD off MPH showed different, attenuated physiological responses to exercise compared to children with ADHD on MPH, with the latter having a higher submaximal heart rate during exercise (Mahon et al. 2008). However, other cardiorespiratory measures and perceived exertion were found to be unaffected by the use of MPH (presumably due to the small sample size (n = 14); Mahon et al. 2008).

Cardio exercise: chronic effects
Cardio exercise (e.g., running and jumping) has also been linked to longer lasting effects on cognition in children with ADHD, resulting in improved attention (including auditory sustained attention and selective attention/information processing), executive functioning (including set shifting, (accuracy of) response inhibition and planning), verbal working memory, and cognitive speed Choi et al. 2014;Gapin and Etnier 2010;Kang et al. 2011;Smith et al. 2013;Verret et al. 2012;Ziereis and Jansen 2015). However, a lack of robustness of chronic effects on cognition after cardio exercise is shown by some studies not reporting affected functions [e.g., (working) memory], and by studies not confirming significant beneficial long-term effects in the areas of inhibition, processing speed, planning, memory span, and continuous motor timing (Gapin and Etnier 2010;Smith et al. 2013). The lack of agreement in findings might partly be explained by the small sample sizes examined in the latter two studies (n = 14 and n = 18, respectively, which is less than the requested n = 34 for a one-factorial within subject design) and their consequential low statistical power.
A number of studies provided ample evidence for chronic effects of cardio exercise on parent or teacher ratings and observations of a broad range of behavioral and socio-emotional outcomes. Overall functioning, general ADHD symptomatology, attention, the child's self-esteem, anxiety, depression, somatic complaints, academic and classroom behavior, and social behavior were reported to be improved after cardio exercise, extending beyond the day on which the physical exercise took place (Ahmed and Mohamed 2011;Choi et al. 2014;Kang et al. 2011;Lufi and Parish-Plass 2011;Smith et al. 2013;Verret et al. 2012). Social behavior in this regard refers to general social skills, social competency, cooperativeness, perceived social, thought and attention problems, interrupting behavior and (unintentional) aggression. No consistent evidence was found in these studies, however, for longterm improvements of hyperactivity, assertiveness, taskorientation, self-control, emotional behavior, and oppositional/defiant social behavior (e.g., language use, intentional aggression, not following adult directions, and teasing). The fact that effects were not significant is presumably (partly) explained by design limitations (e.g., lack of a control group or small sample size; Kang et al. 2011;Smith et al. 2013). In line with this lack of robustness, McKune et al. (2003) presented comparable improvements in both the physical exercise and the control group with regard to attention, behavior, and emotional skills. These effects were thought to be (partly) explained by the extra attention children received in both conditions. Last, motor performance scores, including locomotion, hand-eye coordination, and general motor skills (e.g., fine motor skills, strength, and agility) were found to benefit from physical exercise McKune et al. 2003;Pan et al. 2014;Verret et al. 2012;Ziereis and Jansen 2015). Although Pan and colleagues (2014) described overall improvement in motor skills and coordination, performance of some motor skills (e.g., balance and running speed) improved comparably in children with ADHD who performed physical exercise and a control group without exercise, which points to test-retest effects (Pan et al. 2014). Regarding neurophysiological improvements, Choi et al. (2014) found in a functional magnetic resonance imaging study that brain activity enhanced within the right frontal and temporal cortices during a set-shifting task, but that brain activity within several other cortices remained unchanged. An EEG/ERP study showed inconsistent findings, with an increased No-Go N2 amplitude but no increased No-Go P3 amplitude (which are linked to improved executive function) in a physical activity group after 10 weeks of exercise, and no reduction in EEG theta power (which is often increased in children in ADHD, and are associated with more ADHD symptoms; Janssen et al. 2016a, b).

Non-cardio exercise: acute effects
The number of studies on acute effects of non-cardio exercise is scarce, and their conclusions are all limited by inadequate study designs that were justly addressed by the authors of these studies (e.g., small sample sizes, absence of a control group/condition). One study reported improved anxiety and conduct as well as less hyperactivity, inappropriate emotions and daydreaming in children with ADHD following tai chi sessions, but asocial behavior remained unchanged (Hernandez-Reif et al. 2001). Another study showed that although the attention of children with ADHD significantly improved after a walk in the park, it remained unclear whether this effect could be attributed to the physical activity or the natural environment (Taylor and Kuo 2009). A case study stated that playground activity appeared to diminish hyperactive behavior, but this improvement might (partly) be due to the reinforcing and appraising character of the situation rather than the physical activity itself (Azrin et al. 2006).

Non-cardio exercise: chronic effects
Similar to the acute effects of non-cardio exercise, information about the chronic effects of non-cardio exercise is scarce. The enhanced behavioral/socio-emotional measures as described by Hernandez-Reif et al. (2001) remained on a long-term basis as well. Maddigan et al. (2003) described positive effects of yoga on several general cognitive and neuropsychological functions (i.e., the ability to do homework and to cope with stressful situations, balancing, flexibility, and attention). This study, however, claimed that solid conclusions cannot be drawn from the results, because of the small sample size included (N = 10) as well as methodological limitations related to the study execution (e.g., unstable attendance to experimental sessions). Another study also reported improvements of behavior and emotions following yoga (e.g., in terms of ADHD symptomatology and social behavior; Jensen and Kenny 2004). The underlying behavior observation ratings were, however, ambiguous, with more positive ratings of teachers than parents, possibly due to the fact that teachers were observing the children, while they were under the influence of stimulant medication. Furthermore, as shown by behavioral improvements in the social intervention control group, positive effects were not limited to the yoga group. Finally, the sample size included in this study was too small for sufficient power (N = 19, Jensen and Kenny 2004).
Preliminary findings on the effect of physical exercise on adults with ADHD Although the literature on the effect of physical activity on (cognitive) functioning in adults remains particularly scarce to date, a number of studies presented promising results. Two studies described a positive relation between engaging in physical exercise and behavioral/socio-emotional outcomes, including enhanced levels of self-reported motivation and vigor, and lower levels of impulsivity, worrying, intrusive and worrisome thoughts, confusion, fatigue and depression (Abramovitch et al. 2013;Fritz and O'Connor 2016). The findings were not robust for all measures, as Fritz and O'Connor (2016) failed to find improvements of hyperactivity or cognition despite examining a sample of adequate size (N = 36).
With regard to non-cardio physical activity, Fuermaier et al. (2014a, b) reported promising effects of Whole Body Vibration (WBV) on cognitive functioning in adults with ADHD. WBV is a passive exercise method in which people are physically activated by exposure to environmental vibration by sitting or standing on a vibrating plate. In a group study, acute effects of WBV treatment on attention performance were found, with larger beneficial effects in adults with ADHD (medium effect size) than healthy adults (small effect size; Fuermaier et al. 2014a). It seems noteworthy to point out that these effects could be observed after a WBV treatment of only a few minutes. A case study about a young male adult with ADHD who underwent WBV treatment during 2 weeks, described long-lasting acute effects (assessed after nocturnal rest, within 24 h following the two-week intervention) on cognition (e.g., inhibition, flexibility, self-reported levels of attention). However, since most of the cognitive measures returned to their baselines levels after 2 weeks of no WBV, chronic effects were not found. The acute effect of WBV was substantially larger than the test-retest effect in a healthy control group receiving no treatment (N = 6). Reaction times, distractibility, and working memory were unaffected by WBV treatment. Table 2 displays the classification of four important methodological quality determinants of the reviewed studies. As shown in Tables 1 and 2, most studies adequately described the diagnosis of the included patients with ADHD (23 out of 29 studies). Almost half of the studies were, however, underpowered because of too small sample sizes (13 out of 27 studies; 2 studies were case reports). Although the majority of the studies adequately included and described a control condition/group (18 out of 29 studies), a substantial share of the papers did not include an adequate control condition/group without exercise. Most of the papers reported stimulant medication use of the included patients, and sufficiently described whether patients took their regular medication, were medication naïve, whether medication use was abstained during the experiment, or whether medication influences were controlled for in the analysis (22 out of 27 studies). Overall, it can be concluded that especially small sample sizes and the absence of adequate control conditions/groups were often limiting the reviewed studies.

Discussion
In this paper, the literature on the potential acute as well as chronic effects of both cardio and non-cardio physical exercise on cognitive, behavioral/socio-emotional, and physical/(neuro)physiological outcome measures in patients with ADHD was reviewed.
First, cardio exercise in children with ADHD appeared to be related to acute positive changes in a wide variety of these outcome measures with effect sizes (when reported) up to 1.26 (Cohen's d). However, findings were selective, with some outcome measures showing an improvement following exercise and with other measures showing no improvement. Second, studies on chronic effects following cardio exercise implied that children with ADHD also benefit on a longer term (i.e., after a substantial period of rest) from cardio physical activity as indicated by significant improvements with effect sizes (when reported) up to 0.96 (Cohen's d).
Selective effects were observed here as well by some measures not showing improvements. It should be considered that some null findings may have been caused by inadequate research designs (see Results section). Several reported differences that did not reach significance were in the expected direction. Although most studies on cardio exercise generally found improvements on cognitive functions that are often impaired in ADHD, such as speed of information processing, attention, inhibition, and flexibility (Ahmed and Mohamed 2011;Chang et al. 2012;Chang et al. 2014;Choi et al. 2014;Gapin and Etnier 2010;Hartanto et al. 2015;Kang et al. 2011;McKune et al. 2003;Medina et al. 2010;Pontifex et al. 2013;Smith et al. 2013;Verret et al. 2012;Ziereis and Jansen 2015), some studies pointed to a lack of robustness by not (consistently) revealing comparable improvements across studies Gapin and Etnier 2010;Medina et al. 2010;Pontifex et al. 2014;Smith et al. 2013;Ziereis and Jansen 2015). For example, working memory and memory functions were found to be improved on the long term by exercise in one study (Ziereis and Jansen 2015) but not in others (Gapin and Etnier 2010;Medina et al. 2010;Smith et al. 2013). This inconsistency in findings may be related to differing methodological study qualities, which will be further discussed under 'limitations.' Overall, it can be stated that the reviewed studies describe acute as well as chronic beneficial effects of cardio exercise on a wide variety of cognitive and behavioral functions in children with ADHD, but that cardio exercise does not necessarily result in functional improvement in all areas.
With regard to non-cardio exercise and acute effects, studies suggested cognitive, behavioral and motor improvements in children with ADHD, with effect sizes (when reported) up to 0.77. Drawing firm conclusions seems, however, premature, because of the third variables potentially explaining or moderating these effects (e.g., environmental factors, reinforcement and motivation) and because of the limited number of studies performed. Moreover, most studies were characterized by methodological weaknesses (e.g., small and heterogeneous groups, lacking control group/condition and logistical problems with study execution in home settings). Although some chronic beneficial effects were demonstrated for non- Sweat it out? The effects of physical exercise on cognition and behavior in children and… S21 cardio exercise on general cognitive and behavioral functions, valid conclusions cannot be drawn either because of weak study designs or the absence of theoretical support provided for the findings. In the light of the limitations discussed, clear conclusions of the acute and chronic effects of non-cardio exercise on children with ADHD remain elusive to date. Based on our current knowledge, cardio exercise appears to be a more promising treatment method for children with ADHD than non-cardio exercise with regard to both acute and chronic cognitive and behavioral effects, but more well-designed studies are needed.
The literature on the effects of physical exercise in adults with ADHD also remains relatively scarce. In the available studies, beneficial effects (of medium size) of both active (leisure time sports activities with a strong aerobic component) and passive physical activity (WBV) were described to be related to improved cognitive and behavioral functions including attention, inhibition, motivation and impulsivity (Abramovitch et al. 2013;Fritz and O'Connor 2016;Fuermaier et al. 2014a, b). Not all assessed functions improved and the duration of the effects of passive physical remains to be elucidated, but the fact that positive findings were shown in these studies is promising when considering physical exercise interventions for adults with ADHD.
In view of the findings of this literature review, the question arises as to why physical exercise could improve cognitive and/or behavioral functioning in ADHD. As mentioned in the introduction, the underlying working mechanism of chronic exercise likely entails enhanced neural growth and development which may have long-term implications (Best 2010;Bishop 2007;Bolduc et al. 2013;Halperin et al. 2012;Pesce 2009). Especially when physical exercise is performed at young age, long-term positive responses of cell-proliferation might take place (as demonstrated in animal research; Halperin et al. 2012;Kim et al. 2004). Regular physical exercise may, therefore, be especially beneficial for children with ADHD, because cardio exercise increases specific catecholamines and proteins/enzymes that are typically reduced in ADHD (e.g., dopamine, tyrosine hydroxylase and brain-derived neurotrophic factor; Chang et al. 2012;Hattori et al. 1994;Kim et al. 2011). In addition, cognitive functions that have often been found impaired in patients with ADHD, such as executive functions, improved after acute exercise in the majority of the reviewed studies. According to Chang and colleagues (2012), this can be explained by an exercise-induced improved allocation of attention resources and by positive influences on the dorsolateral PFC. Interestingly, memory improvements after acute cardio exercise in healthy adults appear to be related to better-facilitated molecular mechanisms of memory encoding and consolidation (Roig et al. 2013).
It is assumed that physical exercise entails similar neurobiological effects as stimulants (e.g., increased availability of monoaminergic catecholamines in the brain, Fritz and O'Connor 2016;Wigal et al. 2013) and that these effects result in improved functioning in overlapping areas of cognition. A number of studies even showed (cognitive) gains of physical exercise on top of medication treatment (i.e., when children were on stimulant medication during the exercise bouts and tests; Choi et al. 2014;Gapin and Etnier 2010;Jensen and Kenny 2004;Kang et al. 2011;Maddigan et al. 2003;Mahon et al. 2008;McKune et al. 2003;Tantillo et al. 2002;Verret et al. 2012). Regular physical exercise could, therefore, be an effective (additional) treatment option for children with ADHD (in this review, exercise bouts of 30 min were most common with a range of 1-90 min). This would especially be beneficial in the following cases: (1) when deficits are not normalized by pharmacological treatment (e.g., see Tucha et al. 2006), (2) when pharmacological treatment is not the first choice treatment, (3) when effects of pharmacological treatment on a child's problem behaviors are inconsistent (Swanson et al. 2011) or (4) when children suffer from milder disturbances which might benefit from physical exercise, thus possibly eliminating the need for pharmacological treatment. The advantages of physical exercise are broad: exercise can be combined with all treatment approaches currently applied in ADHD (including pharmacological treatment), it is cheap, non-invasive, and easy to implement, it has additional health benefits (e.g., potential prevention of chronic diseases and obesity), and can improve psychological well-being (e.g., Warburton et al. 2006). However, exercise programs should at all times be individually adapted and potential risks as well as physical and health issues have to be taken into consideration. Based on the demonstrated effectiveness and the practical implementation in everyday life, daily exercise bouts of 30 min appear reasonable.

Limitations
Despite the existing evidence for effectiveness, the present findings ought to be interpreted with caution as the majority of described studies suffered from one or several methodological shortcomings. The methodological quality of the included papers was screened by classifying four important quality determinants for each paper (see Table 2). Although the majority of the studies adequately assessed the ADHD symptoms by means of standardized measures and controlled for the use of medication, almost half of the studies were underpowered, and one-third of the studies did not include an adequate control group/condition without exercise. We weighed these limitations in this review, and are therefore reluctant to draw conclusions about the beneficial effects of non-cardio exercise. However, a substantial amount of studies on the effects of aerobic exercise that did adopt rigorous methodological designs point to beneficial short-and long-term effects for patients with ADHD. This finding stresses the importance of developing exercise programs for children and adults with ADHD and of rigorously evaluating their effects by means of randomized controlled trials. By including adequate control groups/conditions, potential confounding variables can be ruled out, such as received attention, environmental effects as well as possible testing/learning effects of repeated administration of the same tests. Furthermore, none of the studies examining behavioral/socioemotional outcomes reported whether raters were blind to the treatment condition of the participants. To overcome bias or a placebo effect, it is important for future studies that either raters are blind to the treatment conditions or that more objective assessments are used. Another factor influencing the results of the studies may be an unequal gender distribution in favor of boys. This, however, is not necessarily a weakness, since ADHD prevalence rates are higher in boys relative to girls (Gapin and Etnier 2010), but articles reporting gender differences in physical exercise effectiveness point to the need of equal numbers of boys and girls with ADHD in studies (e.g., Tantillo et al. 2002). Last, exercise bouts were generally not standardized but rather varying with regard to type, intensity, and duration. Although this is on the one hand a strength, since it shows effectiveness at varying and personalized intensities, it is on the other hand a weakness as it is not well controlled, making clear conclusions difficult.
In addition to these study-specific problems, there were some limitations on the review level as well. Reliably comparing studies and drawing general conclusions was complex, since studies differed considerably in many aspects (i.e., study designs, type of cognitive tests used, measurement outcomes, exercise types and duration and sample sizes), making direct comparison difficult. Furthermore, studies examined samples of children and adolescents with an age ranging from 5 to 18 years. Since symptomatology of ADHD and cognitive functions change with age and individual development (e.g., Schmidt and Petermann 2009), it would have been desirable to analyze whether specific age groups benefit more from certain forms of exercise (cardio versus non-cardio). However, this was not possible because of the small number of studies. This difficulty is even complicated by the fact that children of different age groups were pooled in the same samples (e.g., Medina et al. 2010: age range 7-15 years) which might even have reduced the impact of exercise on functioning. A final limitation on the review level is that, despite the growing research focus on alternative ADHD treatments, only a small number of studies are available.

Future perspectives
Given the promising current body of evidence regarding the cognitive, behavioral/socio-emotional, and physical/ (neuro)physiological effects of physical exercise in the treatment of ADHD symptoms, further research is necessary to substantialize these findings. It is of importance to clearly distinguish between, on the one hand, the exercise type (cardio versus non-cardio) and, on the other hand, the duration of the effects (acute versus chronic). This distinction could help establish the type of exercise that is most effective for treating ADHD symptoms in clinical practice. Even though especially cardio exercise appears promising, both cardio and non-cardio exercise should be further examined in well-controlled studies to allow more definite conclusions, e.g., by means of randomized-controlled trials. Future studies should further explore the potential treatment effects and implementation possibilities of physical exercise in patients with ADHD and would be advised to take the following issues into account. First, studies should consider multiple assessment moments by applying designs which allow the measurement of both acute and longitudinal effects. Second, studies should include adequate control groups/conditions without exercise to capture the true effect of exercise, and to control for confounding variables, such as the attention received or test-retest effects. In addition, more general aspects with regard to the methodological quality should be considered, including larger sample sizes, the use of a wide variety of cognitive, behavioral/socio-emotional and physical/ (neuro)physiological outcome measures, the use of raters who are blind for the treatment condition, and controlled intensity and duration of the bouts of exercise. Control groups of healthy children or adults should be included to assess whether exercise induced improvements are (1) normalizing functioning of patients with ADHD and (2) whether improvements are comparable to the improvements described in healthy people. Finally, future studies should address the interaction of physical exercise and stimulant medication in a controlled design, examining possible complementary and/or differing effects.
Due to the promising findings and other advantages of physical exercise implementation (e.g., inexpensive, easy to apply, additional health benefits), future studies should focus on questions like ''Are there responders and nonresponders?,'' ''Do certain dysfunctions improve more by exercise than others?,'' ''What physical activity (e.g., running, yoga, etc.) has the best outcome for certain behaviors (e.g., cognition, social behavior)?,'' ''What is a realistic exercise plan (e.g., with regard to intensity, frequency, duration, etc.)?,'' and ''Does regular physical exercise have an additional effect to other treatments (e.g., stimulant drug treatment) on symptoms of ADHD?.'' A reasonable method for daily implementation should be addressed by considering potential practical difficulties of physical exercise performance on, for instance, school and work days. In this light, passive exercise (e.g., whole body vibration) could also be considered as a practical and wellimplementable form of exercise. Future studies should thus further explore both active and passive forms of exercise and their possibilities for implementation in the treatment of children and adults with ADHD.
Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://crea tivecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.