The Influences of Environmental Enrichment, Cognitive Enhancement, and Physical Exercise on Brain Development: Can we Alter the Developmental Trajectory of ADHD?

Abstract

Attention-deficit/Hyperactivity Disorder (ADHD) is characterized by a pervasive pattern of developmentally inappropriate inattentive, impulsive and hyperactive behaviors that typically begin during the preschool years and often persist into adulthood. The most effective and widely used treatments for ADHD are medication and behavior modification. These empirically-supported interventions are generally successful in reducing ADHD symptoms, but treatment effects are rarely maintained beyond the active intervention. Because ADHD is now generally thought of as a chronic disorder that is often present well into adolescence and early adulthood, the need for continued treatment throughout the lifetime is both costly and problematic for a number of logistical reasons. Therefore, it would be highly beneficial if treatments would have lasting effects that remain after the intervention is terminated. This review examines the burgeoning literature on the underlying neural determinants of ADHD along research demonstrating powerful influences of environmental factors on brain development and functioning. Based upon these largely distinct scientific literatures, we propose an approach that employs directed play and physical exercise to promote brain growth which, in turn, could lead to the development of potentially more enduring treatments for the disorder.

INTRODUCTION

Attention-deficit/Hyperactivity Disorder (ADHD) is a chronic, highly prevalent neurodevelopmental disorder which affects as many as 9% of school-age children (Pastor & Reuben, 2008). ADHD typically emerges during the preschool years, persists through adolescence and into adulthood for many afflicted individuals, and causes significant functional disability throughout the lifespan (American Psychiatric Association Task Force on DSM-IV, 2000). Several effective empirically-validated psychopharmacologic and behavioral interventions for ADHD are currently available. These interventions have been shown to ameliorate the core symptoms of ADHD and to improve academic performance across a wide array of school-related measures. However, such gains are rarely maintained after the termination of treatment (Chronis, Pelham, Jr., Gnagy, Roberts, & Aronoff, 2003; Chronis et al., 2004; Jensen et al., 2007), and relatively few individuals with ADHD receive effective treatment throughout the full course of their disorder (Corkum, Rimer, & Schachar, 1999; Jensen et al., 2007; MTA Cooperative Group, 1999; MTA Cooperative Group, 2004; Perwien, Hall, Swensen, & Swindle, 2004; Sanchez, Crismon, Barner, Bettinger, & Wilson, 2005; Weiss, Gadow, & Wasdell, 2006; Molina et al., 2009). The need for continued treatment throughout the lifetime is both costly and problematic for a number of logistical reasons. Therefore, it would be highly beneficial if a treatment could be employed that would have lasting effects that remain after the active intervention is terminated.

Recent research has begun to elucidate the underlying neural (Casey et al., 1997; Castellanos, 2001; Makris et al., 2007; Schulz et al., 2004; Schulz, Newcorn, Fan, Tang, & Halperin, 2005a; Seidman, Valera, & Makris, 2005; Shaw et al., 2006; Shaw et al., 2007b; Shaw et al., 2007a) and neurocognitive (Sergeant, Oosterlaan, & van der Meere, 1999; van Mourik, Oosterlaan, & Sergeant, 2005; Willcutt, Doyle, Nigg, Faraone, & Pennington, 2005) determinants of ADHD. Additionally, a distinct scientific literature, largely consisting of research with animals, indicates that an array of neurodevelopmental processes facilitate efficient neurotransmission and are highly responsive to environmental influences. The bringing together of these scientific literatures may provide valuable insights for the development of more enduring treatments for ADHD.

OVERVIEW OF EVIDENCE-BASED TREATMENTS FOR ADHD

Pharmacological interventions, most prominently stimulant medication, and behavioral interventions in the forms of parent training and contingency management in the classroom, are considered the best-supported treatments for ADHD. These interventions result in significant benefits for children with ADHD in multiple domains of functioning (Conners, 2002; Greenhill, Halperin, & Abikoff, 1999; Pelham, Jr., Wheeler, & Chronis, 1998; Pelham, Jr. & Fabiano, 2008; Spencer et al., 1996). For example, studies of behavioral parent training (BPT) for ADHD have demonstrated improvements in ADHD symptoms (Anastopoulos, Shelton, DuPaul, & Guevremont, 1993; Sonuga-Barke, Daley, Thompson, Laver-Bradbury, & Weeks, 2001; Cunningham, Bremner, & Boyle, 1995), as well as co-occurring oppositional problems and impairment in children (Erhardt & Baker, 1990; Pisterman et al., 1989; Pisterman et al., 1992). BPT also improves parental functioning (e.g., decreased stress, enhanced competence) (Anastopoulos et al., 1993; Pisterman et al., 1992; Sonuga-Barke et al., 2001). Moreover, behavior contingency management in the classroom yields improvements in teacher reports of children’s functioning, observed behavior of children with ADHD in the classroom setting, as well as better academic productivity (Abramowitz, O'Leary, & Rosen, 1987; Fabiano et al., 2007; Pelham, Jr. et al., 1998; Hoffman & DuPaul, 2000). While data indicate that behavioral treatments are not, on the group level, as effective as carefully-monitored stimulant medication, the large Multimodal Treatment Study of ADHD (MTA Study) found that their pure behavioral treatment group improved to a similar degree as their community treated control group, the majority of whom received medication treatment from their provider outside of the study (MTA Cooperative Group, 1999).

Stimulant medication has been shown to improve functioning in children with ADHD, with studies demonstrating improvements in core symptoms of ADHD, compliance, aggression, and academic productivity (Conners, 2002; Greenhill et al., 1999; Spencer et al., 1996). Effect-size calculations from both behavioral interventions and stimulant medication studies demonstrate that these interventions result in substantial improvements across domains of functioning (Conners, 2002; Pelham, Jr. et al., 2008). Collectively, compelling evidence indicates that behavioral interventions and stimulant medication improve the functioning of children with ADHD and, in some cases, have additional benefits for their families. However, noted below, are several limitations to these treatments which necessitate further investigation into alternative interventions for children with ADHD.

Key Limitations of Current Evidence-Based Interventions for ADHD

Although clearly efficacious, there are limitations to current evidence-based interventions. Stimulant medication is an easy intervention to implement, but many parents prefer not to use medication as a treatment for their child with ADHD (Pisecco, Huzinec, & Curtis, 2001; Power, Hess, & Bennett, 1995). Moreover, a substantial number of children experience notable side effects with stimulant medication (MTA Cooperative Group, 2004; Swanson et al., 2006; Swanson et al., 2007a; Wigal et al., 2006) that may prohibit continued use. In addition, recent concerns have been raised by the American Heart Association (American Heart Association, 2008) suggesting problematic interactive effects of stimulant medication with underlying cardiac conditions that may further limit the acceptability of stimulant medication for children with ADHD, or at least make it less palatable to parents. Behavioral interventions, in contrast, are more palatable but much more difficult to implement, generally quite costly, and may be less effective than stimulant medications (MTA Cooperative Group, 1999).

In addition, there are several limitations that both stimulant medication and behavioral interventions share in common. First, although both treatments are efficacious, the behavior of a significant number of children with ADHD is not normalized by the use of these interventions (Hoza et al., 2005; Swanson et al., 2001). Swanson and colleagues (Swanson et al., 2001) found that 32–64% of children continued to exhibit clinically significant levels of ADHD despite intensive stimulant medication and behavioral treatment regimens. Moreover, Hoza et al. (Hoza et al., 2005) found that despite intensive treatment over 14 months, children continued to have difficulties in peer relationships. Therefore, although children with ADHD may do significantly better with stimulant medication and/or behavioral interventions relative to baseline, they still appear deviant relative to their peers in key areas of functioning.

Furthermore, treatment effects rarely persist past the point of active dosing/ implementation (Chronis et al., 2003; Chronis et al., 2004). This suggests that implementation of behavioral interventions and stimulant medication temporarily suppresses behavioral difficulties and that these difficulties resurface when treatment is no longer active. As such, there are no apparent changes in the underlying deficits that produce the behavioral manifestations of ADHD.

In addition, although ADHD is considered a chronic condition that impacts functioning across multiple settings, which would thus require long-term treatment that impacts each affected setting, long-term adherence to both stimulant medication and behavioral interventions is often poor. For instance, there is considerable noncompliance with stimulant medication (MTA Cooperative Group, 2004; Corkum et al., 1999; Jensen et al., 2007; Perwien et al., 2004; Sanchez et al., 2005; Weiss et al., 2006). More than half of children prescribed stimulant medication stop receiving treatment within a school year (Sanchez et al., 2005), and most maintain good adherence for fewer than two months (Perwien et al., 2004). It has been estimated that fewer than 10% of children with ADHD persist with long-term medication treatment (Weiss et al., 2006). For behavioral interventions to be implemented over the long-term, willingness among key adults (teachers and parents) to implement highly intensive interventions over long periods of time, with high levels of fidelity, is necessary but extremely challenging (Chronis et al., 2001; Witt, 1986).

Lastly, the collective evidence suggests that stimulant medication use has few, if any, long-term benefits for children with ADHD (Charles & Schain, 1981; Loe & Feldman, 2007; Molina et al., 2007; Paternite, Loney, Salisbury, & Whaley, 1999; Satterfield, Hoppe, & Schell, 1982; Swanson et al., 2007b; Weiss & Hechtman, 1993; Molina et al., 2009). This finding also applies to behavioral interventions (Molina et al., 2007; Pelham, Jr. et al., 2008). Thus, although acute benefits of stimulant medication and behavioral interventions are well-documented in the literature, the lack of normalization of functioning for many children following treatment, lack of generalization of treatment effects, difficulties in long-term adherence, and lack of clear improvement in long-term functioning following the use of these interventions are discouraging.

Given these findings, an alternative approach to ameliorating the difficulties of children with ADHD is indicted. The extensive literature indicating neurobiological deficits in ADHD, and the strong body of evidence regarding neural rehabilitation, suggests that research examining the impact of pharmacologic as well as environmental manipulations on neural growth, and lasting cognitive function, could provide a potentially fruitful avenue for novel treatment approaches for this highly impairing and oftentimes life-long disorder.

NEURAL SUBSTRATES OF ADHD

Neuroimaging Evidence

Neuroimaging studies suggest an important role for frontostriatal circuits along with a wide array of other cortical and subcortical brain regions in the pathophysiology of ADHD. Numerous structural MRI studies have reported smaller regions in the prefrontal cortex (PFC) of youth with ADHD (Castellanos et al., 1996; Castellanos et al., 2002a; Durston et al., 2004; Filipek et al., 1997; Hynd, Semrud-Clikeman, Lorys, Novey, & Eliopulos, 1990; Kates et al., 2002; Mostofsky, Cooper, Kates, Denckla, & Kaufmann, 2002; Sowell et al., 2003). Similarly, studies have found reduced caudate nucleus size (Filipek et al., 1997; Hynd et al., 1993), reduced volume in the globus pallidus (Castellanos et al., 1996); and reduced callosal area in ADHD (Baumgardner et al., 1996; Giedd et al., 1994; Hynd et al., 1991; Hill et al., 2003; Hynd et al., 1991; Lyoo et al., 1996; Semrud-Clikeman et al., 1994).

Structural anomalies have also been reported in several brain regions outside of the frontal lobes and striatum. Studies have noted reduced volume in several parts of the cerebellum in ADHD (Castellanos et al., 2002a; Castellanos et al., 1996; Durston et al., 2004; Posner & Petersen, 1990) as well as posterior cortical anomalies (Durston et al., 2004; Filipek et al., 1997; Sowell et al., 2003). Additionally, several (Berquin et al., 1998; Castellanos et al., 1996; Castellanos, 2001; Castellanos et al., 2002a; Hill et al., 2003), but not all (Durston et al., 2004; Filipek et al., 1997; Hynd et al., 1990; Mostofsky et al., 2002), studies report reduced whole brain size in children with ADHD. A key review of the structural MRI literature (Seidman et al., 2005) found support for the notion that ADHD is associated with frontostriatal abnormalities. However, an increasing number of studies, demonstrating widespread neural abnormalities affecting other cortical regions and the cerebellum, were highlighted.

Recently, studies have begun to use MRI to examine differences in cortical thickness between individuals with and without ADHD. In a series of developmentally-sensitive longitudinal studies Shaw et al. (Shaw et al., 2006; Shaw et al., 2007a; Shaw et al., 2007b) reported that children with ADHD followed a similar sequential pattern of cortical development, yet were delayed by as much as 2 – 3 years, depending upon the specific cortical region (Shaw et al., 2007a). These findings were interpreted as suggesting developmental delays rather than permanent anomalies in ADHD. Additional research from this group (Shaw et al., 2006) found links between cortical thickness and clinical outcome such that ADHD children with worse outcome had “fixed” thinning of the left medial PFC and those with better outcomes had right parietal normalization, which was suggested to represent compensatory cortical change. Similarly, McAlonan et al (McAlonan et al., 2009) reported an association between the magnitude of age-related growth in the anterior cingulate, striatum and medial temporal cortex, and improvements in neurocognitive measures of response inhibition in school-age children with ADHD. Further, better clinical outcome was linked to the presence of the DRD4-7-repeat allele (Shaw et al., 2007b), which was associated with thinner right orbitofrontal/inferior prefrontal and posterior parietal cortices. This thinning was most apparent in childhood and largely resolved in adolescence. Yet, these executive function networks which include the dorsolateral PFC, anterior cingulate cortex and inferior parietal lobe remain thinner in adults with ADHD (Makris et al., 2007). A recent review focusing on trajectories of brain development (Shaw, Gogtay, & Rapoport, 2010) suggests that remission in ADHD with age may be linked to a normalization of initial delays or deficits in brain networks, whereas ADHD persistence into adolescence may be associated with a lack of normalization and perhaps a more deviant trajectory of brain growth.

Consistent with the structural MRI data, functional MRI (fMRI) studies provide compelling data indicating functional brain differences between individuals with ADHD and controls. These functional studies have linked neural anomalies to deficiencies in a wide array of neurocognitive processes including inhibitory control (Casey et al., 1997; Durston et al., 2003; Schulz et al., 2004; Schulz et al., 2005a; Tamm, Menon, Ringel, & Reiss, 2004; Vaidya et al., 1998), conflict resolution (Bush et al., 1999; Tamm et al., 2004), motor control (Rubia et al., 1999), timing (Durston et al., 2007), attention (Cao et al., 2008; Sonuga-Barke & Castellanos, 2007; Uddin et al., 2008a) and working memory (Valera, Faraone, Biederman, Poldrack, & Seidman, 2005). Although findings have not always been consistent, many studies have reported greater and more diffuse PFC and basal ganglia responses to increased cognitive demands in participants with ADHD. These findings are reminiscent of the immature brain function associated with susceptibility to interference in young children (Casey, Tottenham, & Fossella, 2002).

Recently, there has been an emerging shift in focus among neuroimaging studies from the examination of specific brain regions to an examination of networks or how specific regions are connected or interact (for review, see (Konrad & Eickhoff, 2010). This research, which is in its infancy and still has many methodological issues to tackle, has focused largely, on two proposed neural systems. The Default Mode Network (DFN), comprised of the precuneus, posterior cingulate cortex, the medial prefrontal cortex and portions of the parietal cortex (Schilbach, Eickhoff, Rotarska-Jagiela, Fink, & Vogeley, 2008), has been shown to be most active during rest, and to deactivate during active task engagement (Raichle et al., 2001). In contrast, a second network, which includes the dorsolateral prefrontal cortex, the intraparietal sulcus, and the supplementary motor area is more quiescent during rest and becomes activated during periods of task engagement. This latter system seems more linked to increased alertness and response preparation than to any specific cognitive process (Konrad et al., 2010). These two neural systems have been described as being ‘anti-correlated’ such that when the task positive fronto-parietal system turns-on in response to task demands, the DFN becomes less active, and when at rest the DFN is most active, and activity in the fronto-parietal system diminishes (Fox et al., 2005; Sonuga-Barke et al., 2007).

When comparing ADHD participants to controls, some data suggest, reduced functional activity and connectivity within the DFN during rest (Castellanos et al., 2008; Uddin et al., 2008b), but others report an overactive DFN during quiescence (Tian et al., 2006; Tian et al., 2008). Sonuga-Barke & Castellanos (Sonuga-Barke et al., 2007) have hypothesized that, among those with ADHD, the DFN does not adequately deactivate during task engagement, which in turn interferes with the neural circuits that underlie the ability to efficiently engage in active tasks. According to this hypothesis, the interference results in frequent errors and lapses in attention, as evidenced by the frequently reported increased reaction time variability associated with ADHD (Castellanos & Tannock, 2002b; Kuntsi, Oosterlaan, & Stevenson, 2001; Russell et al., 2006). Notably, psychostimulant treatment has been found to improve DFN suppression during active task engagement in youth with ADHD (Peterson et al., 2009), as well as to decrease reaction time variability (Spencer et al., 2009).

Finally, recent research has also examined connectivity during active task engagement. Two studies (Rubia et al., 2010; Vloet et al., 2010) have reported decreased functional connectivity during active task performance in children with ADHD, which may in part be normalized by treatment with stimulant medication (Rubia et al., 2009). However, in adults with ADHD a more complicated pattern of regional increases and decreases in functional connectivity has been reported (Wolf et al., 2009).

Thus, when taken together, neuroimaging studies provide compelling evidence for neurodevelopmental differences between individuals with and without ADHD that span a wide array of brain regions and neural circuits.

Neuropsychological Findings

Consistent with the neuroimaging data, neuropsychological data provide clear evidence that children with ADHD have impairments in a wide array of neurocognitive domains. Much of the literature has focused on executive functions (EFs) which are mediated by circuits involving the PFC. Studies most consistently find deficits in inhibitory control (Barkley, 1997; Casey et al., 1997; Durston et al., 2003), regulation of attention (Pennington, Groisser, & Welsh, 1993; Swanson et al., 1991; Halperin et al., 1993; Johnson et al., 2008), working memory (Castellanos et al., 2002b; Martinussen, Hayden, Hogg-Johnson, & Tannock, 2005), planning (Pennington et al., 1993; Barkley, 1997), and shifting sets (Seidman et al., 1995; Reader, Harris, Schuerholz, & Denckla, 1994; Hall, Halperin, Schwartz, & Newcorn, 1997). Willcutt et al’s. (Willcutt et al., 2005) meta-analysis of 83 studies (total N = 6,703), which focused on 13 EF tasks, clearly indicated that groups of children with ADHD perform more poorly than controls on many EF measures, with effect sizes generally in the medium range (.46 - .69). However, because of the relatively modest effect sizes, it was concluded that EF weaknesses are “neither necessary nor sufficient to cause all cases of ADHD.” Nonetheless, EF deficits likely contribute to the functional impairments that are present in many children with ADHD.

Despite the common conceptualization of ADHD as a disorder of EFs (Barkley, 1997), children with ADHD also differ from controls on a broad range of non-EF functions such as motor coordination (Blondis, 1999; Carte, Nigg, & Hinshaw, 1996; Kadesjo, Kadesjo, Hagglof, & Gillberg, 2001; Sheppard, Bradshaw, Georgiou, Bradshaw, & Lee, 2000; Steger et al., 2001), perception (Garcia-Sanchez, Estevez-Gonzalez, Suarez-Romero, & Junque, 1997; Mangeot et al., 2001), language (Beitchman, Tuckett, & Batth, 1987; Beitchman et al., 1996; Carte et al., 1996; Humphries, Koltun, Malone, & Roberts, 1994; Purvis & Tannock, 1997; Tirosh & Cohen, 1998), visuomotor integration (Raggio, 1999), and learning and memory (Felton, Wood, Brown, Campbell, & Harter, 1987). Meta-analyses (Frazier, Demaree, & Youngstrom, 2004; van Mourik et al., 2005; Willcutt et al., 2005) indicate an array of executive and non-executive function deficits in children with ADHD. Furthermore, increased variability in reaction time (RT) is among the most consistently reported deficits in children with ADHD (Castellanos et al., 2002b; Russell et al., 2006; Spencer et al., 2009). Additional presumably “non-executive” parameters have been shown to differentiate ADHD from non-ADHD children, including signal detectability (d’) and response bias (lnβ) variables from continuous performance tests (CPTs; (Losier, McGrath, & Klein, 1996). Kuntsi and colleagues (Kuntsi et al., 2001) reported that RT variability discriminated ADHD children from controls better than measures of inhibition and working memory, and Epstein and colleagues (Epstein et al., 2003) reported variability in d’, lnβ, and RT to be strongly associated across multiple ADHD symptom domains. A recent study (Rommelse et al., 2007) reported that after controlling for “lower order” cognitive processes, there was little evidence for primary EF deficits in children with ADHD. As such, some have hypothesized that the EF deficits frequently observed in children with ADHD may be due to deficiencies in largely subcortical, regulatory systems, rather than cortical EF circuitry per se (Douglas, 1999; Halperin & Schulz, 2006; Rommelse et al., 2007; Sergeant et al., 1999; Sonuga-Barke et al., 2007).

In addition to more broadly assessing neurocognitive deficits in individuals with ADHD, several studies have focused more specifically on the three distributed neural networks hypothesized to underlie attention (Posner et al., 1990). One meta-analysis of 14 studies found little or no evidence for any visuospatial attention deficits in ADHD, including those functions typically attributed to the anterior or executive attention system (Huang-Pollock & Nigg, 2003). However, more recently, Johnson et al. (Johnson et al., 2008) reported evidence for impaired alerting and conflict resolution in youth with ADHD as measured using the Attention Network Test (ANT; Fan, McCandliss, Sommer, Raz, & Posner, 2002). As posited by Posner and colleagues, the alerting network is driven largely by noradrenergic input from the locus coeruleus to the frontal and parietal lobes, whereas conflict resolution is more closely linked to a dopaminergic system involving the basal ganglia, anterior cingulate and the dorsolateral prefrontal cortex (Posner et al., 1990; Posner, Sheese, Odludas, & Tang, 2006).

From this review of the neuropsychological and neuroimaging studies, we can see that the findings do not support simple models that posit that ADHD is due to dysfunction of a few isolated brain regions, including the PFC. Rather, individuals with ADHD differ from controls on measures derived from multiple brain regions and across a wide array of neurocognitive domains. Further, studies examining cortical thickness provide compelling evidence that careful examination of developmental trajectories may be key to uncovering the neural substrates of ADHD, and that clinical outcomes might be linked to cortical development and compensatory mechanisms that develop throughout childhood.

NEURODEVELOPMENTAL PERSPECTIVES ON ADHD

Throughout the past decade investigators have increasingly embraced the notion that ADHD is best viewed within the context of a developmental trajectory rather than that of a static medical condition (Halperin et al., 2006; Sagvolden, Johansen, Aase, & Russell, 2005; Sonuga-Barke & Halperin, 2009; Taylor, 1999), and notably, that this developmental trajectory may be different for boys and girls (Mahone & Wodka, 2008). It is through this developmental perspective that one is most likely to identify the mediators and moderators of the diverse outcomes characteristic of youth with ADHD. Compelling data support an array of hypotheses that posit executive (Pennington & Ozonoff, 1996), motivational (Sonuga-Barke, Taylor, Sembi, & Smith, 1992; Sonuga-Barke, Williams, Hall, & Saxton, 1996), inhibitory control (Barkley, 1997), cognitive-energetic (Sergeant et al., 1999) and reward-related (Sagvolden et al., 2005) deficits as being central to the etiology of ADHD. More recently, it has been proposed that there may be multiple causal pathways that contribute to the emergence of ADHD in early childhood (Nigg, Willcutt, Doyle, & Sonuga-Barke, 2005; Sonuga-Barke, 2002). However, little consideration has been afforded to pathways out of ADHD or the mechanisms by which the severity of the disorder diminishes over time, which could serve as the conceptual basis for the development of novel, perhaps enduring, interventions.

In an attempt to explain the commonly observed diminution in ADHD symptoms across the lifespan, Halperin and colleagues (Halperin et al., 2006; Halperin, Marks, & Schulz, 2008) have posited distinct neurocognitive mechanisms for the etiology of and relative recovery from ADHD. Specifically, they hypothesized that ADHD is caused by non-cortical neural dysfunction that is present early in ontogeny and remains relatively static throughout life. In this regard, the disorder is never really “cured” or “out-grown.” Nonetheless, they posited that the reduction of symptoms oftentimes seen over development are at least partially accounted for by the degree to which prefrontally-mediated EFs and other higher cortical functions, which emerge throughout childhood and adolescence, can compensate for these more primary and enduring subcortical deficits.

As such, the PFC and other cortical regions may be involved in the reduction in severity of symptoms and impairment from ADHD rather than the etiology of the disorder. The efficiency of later cortical developmental processes may determine the extent to which an individual can compensate for or have a diminution of ADHD symptoms. This hypothesis is consistent with data suggesting that trajectories of cortical development throughout middle childhood might be closely linked to ADHD outcomes (Shaw et al., 2006; Shaw et al., 2007b), and that lower reaction time variability is associated with increased activation of the prefrontal circuit (prefrontal cortex and caudate) in children with ADHD (Suskauer et al., 2008). This latter finding was interpreted as prefrontal compensation for a more primary motor deficiency. In addition, fMRI findings indicate that PFC activation in response to inhibition in adolescents with childhood ADHD corresponds to the persistence of symptoms, such that those who are less symptomatic appear more like never-ADHD controls (Schulz et al., 2005b; Schulz et al., 2005a). Finally, longitudinal data (Halperin, Trampush, Miller, Marks, & Newcorn, 2008) examining neurocognitive functioning in adults who had ADHD in childhood, indicate that only those in whom ADHD has persisted differ from controls on measures of EFs such as working memory, sustained attention and inhibitory control.

This hypothesized developmentally-related compensation for earlier deficits may also be related to changes in functional connectivity in the brain that occurs over development. For example, Dosenbach et al. (Dosenbach et al., 2007) described two distinct “top-down” functional control networks in adults; a frontoparietal network involved in adaptive online task control and a cinguloopercular network involved in more stable set control. When compared across groups of children, adolescents and adults, a clear pattern of developmentally-related differences emerged such that short-range connections associated with these networks decreased whereas long-range connections increased across development (Fair et al., 2007). These findings suggest greater segregation between networks as well as improved integration within networks with increased age. It is possible that these developmental processes that support the maturation of this dual-network control system play a critical role in the hypothesized compensation for earlier deficits in youth with ADHD. Notably, this pattern of developmental decreases in local connectivity and increases in long-range connectivity does not appear limited to these specific neural systems. Rather, it appears to be a more general developmental principle that operates throughout the brain (Fair et al., 2009; Supekar, Musen, & Menon, 2009).

Finally, it has been hypothesized that the persistence of ADHD into adulthood, or the relative diminution of symptoms across the lifespan, is genetically mediated (Faraone, 2004). Supporting this hypothesis, Kuntsi et al. (Kuntsi, Rijsdijk, Ronald, Asherson, & Plomin, 2005) found in a longitudinal study of 4,000 twin pairs that symptom stability in ADHD from early-to-mid childhood was primarily due to shared genetic influences. Further, at the molecular level, children with ADHD who had at least one copy of the DRD4 7-repeat allele had a distinct trajectory of cortical development characterized by normalized cortical thinning, were less likely to maintain a diagnosis of ADHD at follow-up, had higher IQs, and exhibited better global functioning than children with ADHD without a DRD4 7-repeat allele; despite no differences in ADHD symptoms at baseline. This is consistent with findings of Swanson et al. (Swanson et al., 2000) who reported that the presence of the DRD4 7-repeat allele in children with ADHD was associated with fewer neuropsychological deficits. Additionally, the presence/absence of the 7-repeat allele may also underlie a key gene x environment interaction in relation to ADHD, its trajectory, and potential treatment response. As compared to children without the DRD4 7-repeat allele, the behavior of those with a 7-repeat allele was found to be more sensitive to the quality of parenting received (Sheese, Voelker, Rothbart, & Posner, 2007) and to have a better treatment response to a psychosocial parenting intervention (Bakermans-Kranenburg, van Ijzendoorn, Pijlman, Mesman, & Juffer, 2008). These data suggest that there may be genetic differences in the degree to which both the environment and delivered treatments influence developmental trajectories in children.

Thus, from a developmental perspective, interventions that promote neural growth, functional connectivity, and lasting cognitive facilitation could potentially provide a fruitful avenue for novel treatment approaches for this highly impairing oftentimes life-long disorder. Further, it is possible, that some children may be more or less genetically susceptible to respond to such interventions. As described below, accumulating evidence clearly indicates that early brain development is highly susceptible to environmental influences.

NEURAL DEVELOPMENT

Human central nervous system development proceeds in a systematic manner that begins before conception and continues at least into early adulthood. Nevertheless, brain development is non-linear, and progresses in a localized and region-specific manner that coincides with functional maturation (Huttenlocher & Dabholkar, 1997; Keshavan, Anderson, & Pettegrew, 1994). Of particular relevance are the relatively late and protracted development of the basal ganglia and cerebral cortex, including the PFC, and the motor and higher cortical functions that they mediate (Asato, Terwilliger, Woo, & Luna, 2010; Barkovich, 2005; Barnea-Goraly et al., 2005; Benes, 1989; Chelune, Ferguson, Koon, & Disckey, 1986; Giedd et al., 1996; Giedd et al., 1999; Gogtay et al., 2004; Goldman, 1971; Goldman & Galkin, 1978; Huttenlocher & de Court, 1987; Huttenlocher, 1990; Huttenlocher et al., 1997; Hynd et al., 1993; Keshavan et al., 1994; McKay, Halperin, Schwartz, & Sharma, 1994; Mrzljak, Uylings, van Eden, & Judas, 1990; Passler, Isaac, & Hynd, 1985; Welsh, Pennington, & Groisser, 1991; Yakovlev & Lecours, 1967).

The human brain develops primarily in utero and is approximately 80% of adult size by the age of 2 years (Giedd et al., 1999; Kretschmann, Kammradt, Krauthausen, Sauer, & Wingert, 1986). The process of myelination also begins in utero and proceeds rapidly up to age 2 years (Brody, Kinney, Kloman, & Gilles, 1987; Kinney, Brody, Kloman, & Gilles, 1988), but continues well into adolescence and early adulthood (Asato et al., 2010). The first two years of life are also a period of rapid synapse formation that occurs at varying times and rates in different brain regions, reaching maximum density at age 3 months in the auditory cortex and at age 15 months in the PFC, and resulting in an overproduction of synapses (Huttenlocher et al., 1987; Huttenlocher et al., 1997). Synaptogenesis is followed by a plateau phase extending over several years, during which neurons begin to form complex dendritic trees (Mrzljak et al., 1990). These two processes seem to account for the increase in cortical gray matter in childhood found in MRI studies (Giedd et al., 1996; Giedd et al., 1999; Gogtay et al., 2004; Shaw et al., 2006; Shaw et al., 2007a).

Subsequent brain development, beginning at about 5 years of age, is marked as much by cortical organization and refinement as by neuronal growth. Cortical gray matter continues to thicken during the school-age years with about half of the cortical regions attaining peak thickness by the median age of 7.5 years (Shaw et al., 2007a), and cortical thickness peaking at around 10.2 to 12.8 years in the parietal cortex and around 11.0 to 12.1 years in the frontal lobe (Giedd et al., 1996; Giedd et al., 1999; Gogtay et al., 2004). Throughout this period the experience-dependent pruning of inefficient synapses in the cortex is also taking place (Giedd et al., 1996; Giedd et al., 1999; Gogtay et al., 2004; Huttenlocher et al., 1987; Huttenlocher et al., 1997). Synaptic pruning progresses in a region-specific manner and eventually reduces synaptic density to 60% of maximum (Huttenlocher et al., 1987; Huttenlocher et al., 1997), although it is mostly after puberty that the developmental process of cortical thinning occurs. These processes all facilitate efficient neural transmission and are necessary for the functional maturation of the brain. As in humans, this plateau phase of cortical thickness lasts through puberty in monkeys and is thought to be indicative of the need for consistent and high synaptic density during the formative years when learning and experiences are most intense (Bourgeois, Goldman-Rakic, & Rakic, 1994). While much of neural development is likely to be genetically programmed, considerable data indicate that variability in development across the lifespan is at least in part related to one’s experiences.

Environmental Influences on Neurodevelopment

The notion that life experiences and one’s environment can influence both cognitive and neural development was first clearly articulated in 1949 by Hebb (Hebb, 1949), who postulated that synapses are strengthened when presynaptic fibers repeatedly participate in activating the postsynaptic neuron. Subsequently, Hubel and Wiesel’s (Hubel & Wiesel, 1970; Wiesel & Hubel, 1965) classic experiments demonstrated environmental influences on neural plasticity and neurodevelopment in the visual system in cats. In the 1960s Rosensweig and colleagues (Rosenzweig, Krech, Bennett, & Diamond, 1962; Rosenzweig, 1966; Rosenzweig & Bennett, 1969; Rosenzweig, Love, & Bennet, 1968) systematically began to examine the influence of environmental manipulations on brain weight, cortical thickness and the structure of dendrites, as well as on cognitive functioning. In the 1970s, Greenough et al. (Greenough, Volkmar, & Juraska, 1973b; Greenough & Volkmar, 1973a; Greenough, West, & DeVoogd, 1978) expanded this work to include the effects of environmental enhancements on a range of parameters of brain function including dendritic branching, spine density, synaptogenesis, angiogenesis, and gliogenesis.

Since these early seminal studies, an abundant scientific literature, primarily, but not exclusively, in non-human species, has examined the impact of environmental influences on molecular and cellular aspects of neural development as well as on an array of behavioral and cognitive functions. The precise environmental manipulations required to best facilitate specific aspects of neural development, the critical periods in development when they are optimally applied, and the duration necessary for such interventions have remained somewhat elusive. However, there is no longer doubt that brain development is highly responsive to increased levels of physical activity/exercise as well as environmental enrichment. While a full review of this extensive literature is beyond the scope of this paper, a brief overview is warranted. A plethora of studies in rodents have shown that environmental enrichment increases neuronal size, dendritic branching and spine number, synaptic density and overall neurotransmission in the neocortex (Diamond, Krech, & Rosenzweig, 1964; Diamond, 1967; Globus, Rosenzweig, Bennett, & Diamond, 1973; Green & Greenough, 1986; Greenough et al., 1973b; Greenough et al., 1973a; Greenough et al., 1978; Leggio et al., 2005; Rosenzweig & Bennett, 1996). Environmental enrichments also enhance neurogenesis (Kempermann, Kuhn, & Gage, 1998; Nilsson, Perfilieva, Johansson, Orwar, & Eriksson, 1999), long-term potentiation (LTP; Duffy, Craddock, Abel, & Nguyen, 2001), neurotrophin levels (Ickes et al., 2000; Pham, Soderstrom, Winblad, & Mohammed, 1999; Pham, Winblad, Granholm, & Mohammed, 2002), dendritic spine growth and branching (Faherty, Kerley, & Smeyne, 2003; Green, Greenough, & Schlumpf, 1983; Rampon et al., 2000), synaptophysin levels (Frick & Fernandez, 2003; Nithianantharajah, Levis, & Murphy, 2004), and nerve growth factor mRNA and CREB gene expression (Torasdotter, Metsis, Henriksson, Winblad, & Mohammed, 1996; Torasdotter, Metsis, Henriksson, Winblad, & Mohammed, 1998; Williams et al., 2001) in the dentate gyrus of the hippocampus. Consistent with the effects of environmental enrichment on neurodevelopment, numerous studies have shown that it significantly improves performance on an array of spatial and nonspatial memory tasks in rats (Leggio et al., 2005; Nilsson et al., 1999) and mice (Kempermann, Kuhn, & Gage, 1997; Williams et al., 2001).

Evidence for the impact of environmental enrichment in humans is far more limited and centers largely on the concept of ‘cognitive reserve’ as related to the emergence of Alzheimer’s disease. Epidemiologic data provide a compelling link between increased participation in intellectual and social activities in daily life and a slower cognitive decline in the elderly (Scarmeas & Stern, 2004). Furthermore, cognitive exercises show promise as effective interventions for slowing the trajectory of cognitive and functional decline that is associated with dementia (Gates & Valenzuela, 2010).

In addition to environmental enhancements, physical exercise, most commonly in the form of wheel running in rodents, has also been shown to enhance neural growth and development, as well as associated behavioral and cognitive functions. Physical exercise increases levels of synaptic proteins (Vaynman, Ying, Yin, & Gomez-Pinilla, 2006), glutamate receptors (Farmer et al., 2004) and the availability of brain-derived neurotrophic factor (BDNF) (Berchtold, Chinn, Chou, Kesslak, & Cotman, 2005) and insulin-like growth factor-1 (Trejo, Carro, & Torres-Aleman, 2001), all of which can enhance neural plasticity. Specifically, treadmill exercises in rats increase cell proliferation via insulin-like growth factor-1 (Trejo et al., 2001) and cell proliferation and survival in the dentate gyrus (van Praag, Kempermann, & Gage, 1999b). These neural changes in response to physical exercise are accompanied by behavioral changes such that physical exercise enhances spatial learning (Fordyce & Farrar, 1991; Fordyce & Wehner, 1993) and passive avoidance memory (Samorajski et al., 1985).

Physical exercise has also been reported to increase BDNF levels (Ferris, Williams, & Shen, 2007; Gold et al., 2003; Rasmussen et al., 2009; Seifert et al., 2010; Strohle et al., 2010; Tang, Chu, Hui, Helmeste, & Law, 2008; Zoladz et al., 2008), enhance cognitive performance (Baker et al., 2010; Ferris et al., 2007; Laurin, Verreault, Lindsay, MacPherson, & Rockwood, 2001), and to promote brain health (Colcombe et al., 2003) in human adults. One study that examined BDNF levels in children with ADHD reported that as compared to controls, those with ADHD had increased levels of BDNF, and that BDNF levels correlated with neuropsychological performance as indicated by number of errors on a continuous performance test (Shim et al., 2008). However, BDNF change in response to exercise was not assessed.

Exercise and environmental enhancements also have been shown to have a positive effect on outcome in several animal models of neurodevelopmental disorders. Wheel running from an early age delays the onset of motor deficits in knock-out mouse models of Huntington’s disease (van Dellen, Cordery, Spires, Blakemore, & Hannan, 2008), and complex motor training in rats enhances synaptogenesis in the motor cortex and motor coordination following lesions of the sensorimotor cortex (Jones, Chu, Grande, & Gregory, 1999). Perhaps more relevant to the current thesis, chronic exercise blunted the developmental rise of blood pressure while increasing glutamic acid decarboxylase mRNA in the caudal hypothalamus in spontaneously hypertensive rats, which are commonly used as an animal model for ADHD (Little, Kramer, Beatty, & Waldrop, 2001). Finally, rat pups born from alcohol-intoxicated mothers have been shown to have diminished c-Fos activity in the hippocampus along with an array of neurobehavioral deficits. Postnatal treadmill exercise not only enhances c-Fos activity in the hippocampus (Sim et al., 2008), but exercise and various forms of environmental enrichment reduce the severity of behavioral deficits associated with maternal alcohol intoxication (Hannigan, O'Leary-Moore, & Berman, 2007). Thus, at least in animals, interventions involving environmental enrichment and physical exercise can yield lasting positive effects on behavioral anomalies due to both genetic and environmental etiologies.

Most rodent studies focusing on environmental enrichments employ a combination of social (e.g., multiple animals in a cage), cognitive (e.g., toys, tunnels) and motor (e.g., running wheels) stimuli. As such, it is difficult to determine the extent to which any one of these contributes to the positive outcome. Data suggest that both exercise and environmental enhancement contribute to positive neurodevelopmental outcomes, although not necessarily in an identical manner. Some evidence suggests that exercise affects the brain similarly to environmental enrichment, with changes including enhanced hippocampal LTP (Kim et al., 2004) and neurogenesis (van Praag, Christie, Sejnowski, & Gage, 1999a; van Praag et al., 1999b), as well as increased hippocampal and neocortical neurotrophin mRNA expression (Neeper, Gomez-Pinilla, Choi, & Cotman, 1995; Neeper, Gomez-Pinilla, Choi, & Cotman, 1996). Other studies suggest that individual elements may have different effects on neural plasticity and development. Manipulations involving problem solving and coordination increase synapse formation in the cerebellar cortex (Kleim, Vij, Ballard, & Greenough, 1997; Kleim et al., 1998), whereas repetitive physical exercise increases cerebellar blood vessel density (Black, Isaacs, Anderson, Alcantara, & Greenough, 1990). Exercise has also been found to improve spatial memory and preserve cognitive function in aging rats (Anderson et al., 2000; Fordyce et al., 1991) and mice (Anderson et al., 2000; Fordyce et al., 1993; van Praag et al., 1999a). Nevertheless, some data (Faherty et al., 2003) suggest that environmental enhancements are more effective for facilitating neural changes than exercise alone, while others (Lambert, Fernandez, & Frick, 2005) suggest that exercise, but not cognitive stimulation improves spatial memory.

Thus, preclinical data indicate that environmental enhancement and exercise have positive morphological, chemical and cognitive effects on the brain across the lifespan. Much of the human literature, with moderate success, has focused on the latter end of the developmental spectrum, with the aim of addressing cognitive declines characteristic of the elderly. These data suggest that cognitive and social stimulation, as well as physical exercise, may delay the onset of dementia and other neurodegenerative diseases (Scarmeas & Stern, 2004; Ravaglia et al., 2008; Kramer & Erickson, 2007; Arkin, 2007; Rolland et al., 2007). Furthermore, in adults, cognitive training increases cortical brain activity (Olesen, Westerberg, & Klingberg, 2004) and alters dopamine D1 receptor binding in both the prefrontal and parietal cortices (McNab et al., 2009).

Although early childhood is the period of greatest neural plasticity, and the most prominent cell proliferation in response to exercise has been shown to occur at younger ages in rats (Kim et al., 2004), far less research has been conducted with children. Two studies have examined exercise regimes in children with ADHD; one reporting a blunted catecholamine response (Wigal et al., 2003) and the other showing changes in eye blink responses and reductions in motor impersistence that was specific to boys (Tantillo, Kesick, Hynd, & Dishman, 2002). However, neither study examined clinical response nor brain-related changes following exercise. With regard to cognitive enhancement, Rueda et al. (Rueda, Rothbart, McCandliss, Saccomanno, & Posner, 2005) showed that attention training in preschool children yielded changes in executive functioning and parallel alterations in event related responses (ERPs) reflecting patterns typical of older individuals. Thus, literature in children is sparse, but the extensive body of animal research suggests that environmental enrichment and early brain rehabilitation could play a key role in the treatment, and possibly even the amelioration, of ADHD symptomatology through improving the core neural deficits associated with the disorder.

COGNITIVE TRAINING PARADIGMS IN CHILDREN WITH ADHD

In recent years there has been a growing interest in neurocognitive intervention approaches to treating childhood ADHD with the ultimate aim of developing more lasting treatments for the disorder (Toplak, Connors, Shuster, Knezevic, & Parks, 2008). The majority of these approaches have targeted either attention or working memory (Karatekin, 2006; Kerns, Eso, & Thomson, 1999; Klingberg, Forssberg, & Westerberg, 2002; Klingberg et al., 2005; O'Connell, Bellgrove, Dockree, & Robertson, 2006; Rapport et al., 1996; Shalev, Tsal, & Mevorach, 2007). The basic premise of these approaches is that deficits in the targeted cognitive domain are causally related to ADHD symptoms and that remediation of these deficits will lead to lasting improvements. Collectively these studies have shown that cognitive interventions for children with ADHD improve working memory, inhibition, attention, and nonverbal reasoning ability, and may reduce behavioral symptoms of ADHD as reported by parents and/or teachers (Kerns et al., 1999; Klingberg et al., 2005; Shalev et al., 2007).

Several studies have evaluated the effects of cues to improve attention in children with ADHD (Karatekin, 2006; O'Connell et al., 2006; Rapport et al., 1996). Karatekin (Karatekin, 2006) showed that adolescents with ADHD had deficits in accuracy and reaction time during antisaccade tasks relative to typically developing adolescents, but improved with the implementation of visual and auditory cognitive scaffolds. O'Connell et al. (O’Connell et al., 2006) reported that although alerting cues improved short-term sustained attention in children with ADHD, they did not improve sustained attention throughout the experimental task. Finally, Rapport et al. (Rapport et al., 1996) compared methylphenidate to an attentional training intervention in two girls with ADHD and found that both improved performance on a continuous performance task and on the Matching Unfamiliar Figures Test. However, these cognitive effects were not maintained during withdrawal of these interventions, and neither was the short-term impact of methylphenidate and the attention training system on observed behavior.

Other studies have focused on improving attention in children with ADHD through more intensive, targeted attention-focused training programs. Kerns et al. (Kerns et al., 1999) evaluated the program Pay Attention!, which uses a set of visual and auditory activities designed to train different levels of attention, including sustained, selective, alternating, and divided attention. Results indicated improvements on several neurocognitive and academic achievement measures compared to controls. However, only minimal changes in ADHD behavioral symptoms were reported by teachers, and parents reported no changes. More recently, a computerized progressive attentional training (CPAT) program was evaluated in 20 children with ADHD (Shalev et al., 2007) and showed improved performance on nonstandardized academic tests (i.e., reading comprehension, speed of copying) as well as parent-reported inattentive symptoms at home.

In addition to attention-based interventions, the Cogmed Working Memory Training Program (Cogmed) has demonstrated efficacy for enhancing working memory and reducing behavioral symptoms of ADHD (Klingberg et al., 2002; Klingberg et al., 2005). In two studies involving children with ADHD, Cogmed was compared to an identical computer program that implemented low difficulty working memory tasks unadjusted to meet the working memory levels of the child. In a multisite clinical trial (Klingberg et al., 2005), beneficial effects of Cogmed were reported in school-aged children with ADHD. Significant intervention effects were found on measures of visual (Span board) and verbal (Digit Span from the WISC-III) working memory, nonverbal complex reasoning (Raven’s Progressive Matrices), and response inhibition (Stroop task), compared to the low-level working memory computer task. Notably, results from this study also found maintenance of treatment gains on several outcomes at 3-month follow-up. However, treatment effects were found on parent but not teacher ratings of children’s ADHD behavior and oppositionality.

The studies discussed above have generally focused on the ecological validity of the intervention, as related to behavioral changes, but few have assessed the important question of neural plasticity or brain changes as a function of cognitive training. To our knowledge only one such study has been conducted with children. Rueda et al. 2005 showed that a brief 5-day attention training intervention with preschool children resulted in changes in executive functioning as well as in changes in event related responses (ERPs) that seemed to parallel normal developmental brain responses. Specifically, trained 4-year-old children generated a pattern of ERPs that were similar to typically-developing 6-year-olds, and trained 6-year-olds’ ERP patterns became more adult-like. Similarly, in adults, cognitive training, specifically in the form of the Cogmed working memory system, has been shown to increase brain activity in parietal and frontal regions linked to working memory (Olesen et al. 2004) and, importantly, to alter dopamine D1 receptor binding in both the prefrontal and parietal cortices (McNab et al.2009), suggesting a potential mechanism of action.

Thus, some progress has been made in developing non-pharmacologic interventions to ameliorate underlying cognitive dysfunctions in children with ADHD, and importantly, preliminary data suggest that such interventions have the potential to modify brain function. Notably, however, these approaches focus on specific cognitive domains (i.e., attention and working memory), primarily involve intensive computer-based interventions, and have reported some difficulties with compliance and/or retention (e.g., only 18/27 children completed Klingberg et al’s (2005) treatment study). The focus on a specific cognitive domain is problematic because ADHD is a heterogeneous disorder, with most children experiencing impairments in multiple cognitive domains and only a portion have deficits in any single domain (Nigg et al., 2005). Therefore, the narrow focus may limit effectiveness across a wide range of children with the disorder.

Additionally, while the use of computers can provide well-controlled and highly sophisticated interventions with built-in scaffolding of difficulty, generalizability to the real world might be a concern. Most children with ADHD experience considerable difficulties in the social realm; and it is unclear whether improvements on computer tasks will generalize to the outside world in which they must function.

Finally, many of the current training paradigms require intensive effort over extended periods of time on a daily basis. For example, Cogmed requires 40 minutes a day, 5 days per week for 5 consecutive weeks. Reported difficulties in retention rates for such interventions indicate that these approaches may be unpalatable to children and families (Klingberg et al., 2005).

Cognitive Enhancement in the Child’s Social Milieu

Although not directly targeting children with ADHD, Diamond et al. (Diamond, Barnett, Thomas, & Munro, 2007) evaluated “Tools of the Mind” (Tools), a comprehensive evidence-based preschool curriculum targeting EF development which is delivered within a social context as opposed to relying on computer administration. This intervention is relevant to ADHD given the noted difficulties that many children with ADHD have with EFs (Doyle, 2006; Willcutt et al., 2005) and the notion that improved EF over development should reduce the severity of ADHD symptoms and impairments (Halperin and Schulz, 2006). Tools is based on Luria’s (Luria, 1966) and Vygotsky’s (Vygotsky, 1978) work on higher mental functions and operates on the premise that a comprehensive system of activities undertaken within the context of a social environment promotes EF development. In Tools, techniques for supporting, training, and challenging EFs are interwoven within classroom activities. For example, while children are learning language skills or math, they also receive training in executive skill development. Teachers are taught how to implement and support the use of these activities throughout the preschool day. In a sample of preschool children from low-income urban preschools, Tools was compared to the school district’s existing literacy program, which focused on similar academic content to that of Tools but did not address EF development (Diamond et al., 2007). Results of this randomized trial demonstrated that preschool children in Tools-based classrooms significantly improved performance on experimental measures of inhibitory control, working memory, and cognitive flexibility. Behavioral changes were not assessed in this study.

Although the Tools program is a preventive intervention that has been used successfully in a high-risk population rather than in children with ADHD per se, it has design aspects that are appealing to cognitive-based interventions for children with ADHD. First, the intervention is focused on the early childhood period - a time when executive and other higher cortical functions are emerging developmentally. This may be preferable to a remediation approach of teaching skills to children well past the point when these skills have typically developed (Tremblay, 2006). Second, key adult figures within the lives of these children (i.e., teachers) implement this intervention within the setting that these children live and learn (i.e., school), intensifying the dose of intervention that a child receives. Moreover, the intervention is seamlessly integrated within the school context and relies on activities that children find to be enjoyable and that teachers can readily support - making the intervention more likely to be sustained.

FUTURE DIRECTIONS FOR NEUROCOGNITIVELY-BASED TREATMENTS OF ADHD

As described above, an important step forward would be to consider more ecologically valid and easily generalizable delivery mechanisms for these types of interventions. Diamond et al.’s (Diamond et al., 2007) work sets the scene for this as their intervention was integrated into the everyday school program. While this is an ideal platform for intervention, it is unlikely to be possible to have all school programs adapted; and even if it were, it would be very difficult for teachers to individualize the program to children’s specific needs. Therefore the field would be well-served by trying to design ecologically valid home-based interventions to complement what is available to date.

One way to do this would be to administer the intervention in the form of “play” (ranging from quiet play to that involving physical exercise) within the family. Most studies examining the role of play in development have focused on its role in the development of social skills and interpersonal development. As children move from early “parallel play” into social interactive play, they learn to “read” others’ intentions, take turns, regulate their emotions and behavior, and engage in the give-and-take of interpersonal relations. It is within this social context that many of the higher-order executive functions, that may be uniquely human, are most likely to develop and flourish (Diamond et al., 2007; Luria, 1966; Vygotsky, 1978). Although research is limited, juvenile play has been posited to play a role in facilitating neural development. Similar to the effects physical exercise on brain functioning in humans (Gold et al. 2003; Ferris, Williams & Shen, 2007; Rasmussen et al. 2009; Seifert et al. 2010; Strohle et al. 2010; Tang et al. 2008; Zoladz et al. 2008), Gordon, Burke, Akil, Watson, & Panksepp (2003) reported that play increases BDNF, a key modulator of neuronal development and plasticity, in the amygdala and dorsolateral PFC of juvenile rats. Similarly, juvenile play in rats stimulates c-Fos gene expression in a number of brain areas (Gordon, Kollack-Walker, Akil, & Panksepp, 2002). Finally, “chronic play therapy” reduced right frontal lesion-induced hyperactivity in juvenile rats, and enhanced access to rough-and-tumble play in normal animals improved performance on indices of behavioral inhibition (Panksepp, Burgdorf, Turner, & Gordon, 2003). Based on these and similar findings, Panksepp and colleagues (Panksepp, 2007; Panksepp et al., 2003) have speculated that the impulsive and hyperactive behaviors characteristic of ADHD may reflect overactive playful urges in some children and that enhanced access to play may have therapeutic benefits. However, data supporting or refuting this speculation are sparse.

Thus the creative use of directed play, which incorporates cognitive challenges and physical exercise, may have the potential to serve as a vehicle for treatment of children with ADHD. In particular, the intrinsic rewarding qualities of play (i.e., it’s fun) makes it an ideal delivery system for treatment of children with ADHD, who are known to be highly responsive to continuously rewarding contingencies; yet highly resistant to more effortful and less rewarding (i.e., less fun) tasks. In addition, physically and cognitively demanding play within a social context can be used to mold and develop social skills and, when structured in the appropriate manner, can potentially enhance cognitive and behavioral development, and neural growth. Importantly, we do not propose that this approach to intervention would necessarily target or remediate the core neural pathology that causes ADHD. Rather, it would facilitate development and growth of a wide array of cortical regions and their associated functions, which in turn, would allow for the implementation of compensatory mechanisms that have the potential to improve functioning in individuals with the disorder.

There are numerous common children’s games, activities and exercises that would be easy for families to play together and involve the use of wide ranging brain areas, targeting the diffuse neurocognitive deficits that have been associated with ADHD. For example, the game “S]imon-says” involves inhibitory control, “my grandmother went to the market” involves working memory, and “hopscotch” involves physical exercise and requires motor control. We hypothesize that an intervention that focuses on encouraging parents to play these games in a structured and incremental manner with their children, and encouraging children to play these games with siblings and friends, on a daily basis, if complied with, would have an impact on both neural development and behavioral regulation. As compared to other more effortful and less palatable interventions, compliance may be less of a problem with such an intervention. Also, unlike other neurocognitive training programs that continue for a fixed period of time, at which point training ends, in this case, the goal would be for the wide ranging play activities to continue long after the termination of the active intervention. While in the short-term it is unlikely that such an intervention would generate effect sizes comparable to medications, the goal is to impact the long-term trajectory of the disorder. As such, extended follow-up periods will be required to make comparisons to well-validated treatments (i.e., stimulant medication; behavior modification).

In addition, beyond directed play and exercises with parents, it is possible that engaging children in sports, nature (Kuo & Taylor, 2004), and other group-based cognitively and physically challenging activities may provide an avenue for neural and cognitive growth that would serve to facilitate the diminution of ADHD severity across development. Such activities could be provided in schools (Diamond et al., 2007), after school programs, or in summer camps, although it would be important that the activities continue over an extended period of time.

Although many questions remain regarding the clinical utility these cognitive and exercise-related interventions for children with ADHD, there are several clear directions for future work in this area that, ultimately, will allow us to determine the usefulness of this treatment modality within the context of well-established interventions for children with ADHD.