This document discusses using a systems neuroscience approach to better understand attention-deficit/hyperactivity disorder (ADHD). It summarizes that:
1) ADHD was traditionally thought to reflect dysfunction in the prefrontal-striatal circuit, but recent evidence suggests other brain networks are also involved.
2) Resting-state functional MRI is a useful tool for mapping intrinsic brain connectivity networks and their interactions, which can provide insight into the pathophysiology of ADHD.
3) Defining large-scale neural systems through approaches like resting-state fMRI can help formulate a more comprehensive brain model of ADHD beyond just the prefrontal-striatal circuit.
This document discusses using a systems neuroscience approach to better understand attention-deficit/hyperactivity disorder (ADHD). It summarizes that:
1) ADHD was traditionally thought to reflect dysfunction in the prefrontal-striatal circuit, but recent evidence suggests other brain networks are also involved.
2) Resting-state functional MRI is a useful tool for mapping intrinsic brain connectivity networks and their interactions, which can provide insight into the pathophysiology of ADHD.
3) Defining large-scale neural systems through approaches like resting-state fMRI can help formulate a more comprehensive brain model of ADHD beyond just the prefrontal-striatal circuit.
This document discusses using a systems neuroscience approach to better understand attention-deficit/hyperactivity disorder (ADHD). It summarizes that:
1) ADHD was traditionally thought to reflect dysfunction in the prefrontal-striatal circuit, but recent evidence suggests other brain networks are also involved.
2) Resting-state functional MRI is a useful tool for mapping intrinsic brain connectivity networks and their interactions, which can provide insight into the pathophysiology of ADHD.
3) Defining large-scale neural systems through approaches like resting-state fMRI can help formulate a more comprehensive brain model of ADHD beyond just the prefrontal-striatal circuit.
This document discusses using a systems neuroscience approach to better understand attention-deficit/hyperactivity disorder (ADHD). It summarizes that:
1) ADHD was traditionally thought to reflect dysfunction in the prefrontal-striatal circuit, but recent evidence suggests other brain networks are also involved.
2) Resting-state functional MRI is a useful tool for mapping intrinsic brain connectivity networks and their interactions, which can provide insight into the pathophysiology of ADHD.
3) Defining large-scale neural systems through approaches like resting-state fMRI can help formulate a more comprehensive brain model of ADHD beyond just the prefrontal-striatal circuit.
Copyright:
Attribution Non-Commercial (BY-NC)
Available Formats
Download as PDF, TXT or read online from Scribd
Download as pdf or txt
You are on page 1of 10
Special Issue: Cognition in Neuropsychiatric Disorders
Large-scale brain systems in ADHD:
beyond the prefrontalstriatal model F. Xavier Castellanos 1,2 and Erika Proal 1,3 1 Phyllis Green and Randolph Cowen Institute for Pediatric Neuroscience, Child Study Center, NYU Langone School of Medicine, New York, NY, USA 2 Nathan Kline Institute for Psychiatric Research, Orangeburg, NY, USA 3 NeuroIngenia, Mexico, D.F., Mexico Attention-decit/hyperactivity disorder (ADHD) has long been thought to reect dysfunction of prefrontalstriatal circuitry, with involvement of other circuits largely ig- nored. Recent advances in systems neuroscience-based approaches to brain dysfunction have facilitated the de- velopment of models of ADHD pathophysiology that en- compass a number of different large-scale resting-state networks. Here we review progress in delineating large- scale neural systems and illustrate their relevance to ADHD. Werelatefrontoparietal, dorsal attentional, motor, visual and default networks to the ADHD functional and structural literature. Insights emerging from mapping intrinsicbrainconnectivitynetworks provideapotentially mechanistic framework for an understanding of aspects of ADHD such as neuropsychological and behavioral in- consistency, andthepossible roleof primaryvisual cortex in attentional dysfunction in the disorder. A systems neuroscience approach to ADHD Attention-decit/hyperactivity disorder (ADHD), the most common neurodevelopmental disorder occurring in child- hood, is characterized by developmentally excessive levels of inattention, impulsivity and hyperactivity [1]. The worldwide prevalence of ADHD has been estimated at 5.3% [2] although a national survey in the USA found parent-reported ADHD in 9.5% of school-age children [3]. ADHD was once thought to be limited to childhood, but its continuation into adolescence and adulthood is no longer in doubt [4]. However, despite its substantial eco- nomic impact and life-long psychosocial and psychiatric burden, ADHD remains among the most controversial of psychiatric diagnoses. Primarily on the basis of lesion studies in animals and humans, the imaging community initially embraced a prefrontalstriatal model of ADHD that was expanded to include cerebellar involvement [5]. Prefrontal striatal cir- cuits underpin executive function, and dysfunction in such processes has long been considered an important neuro- psychological correlate of ADHD [6]. This model has been largely supported by an ever-increasing number of struc- tural and functional imaging studies [7,8], but divergent evidence such as the involvement of occipital or temporal cortex [9] has tended to be ignored because of the initially reasonable assumption that unexpected results probably represent false positives. However, accumulating evidence suggests that the prefrontalstriatal model of ADHD should be extended to include other circuits and their interrelationships from the perspective of systems neuro- science [10,11]. We suggest that formulation of a more inclusive brain model of ADHD is facilitated by the new paradigm of resting-state functional magnetic resonance imaging (R-fMRI), which is increasingly revealing the intrinsic functional architecture of the brain [12]. Finally, we speculate that modulation of neural networks through imaging-guided transcranial direct current electrical stim- ulation (tDCS) may provide novel therapeutic opportu- nities for disorders such as ADHD. Resting-state functional magnetic resonance imaging Resting-state functional imaging, that is, imaging without a specic task (Box 1), is not new. It dates from the earliest electroencephalography (EEG) and positron emission to- mography studies [13]. What has only recently been appre- ciated is that large-scale neural systems exhibit synchronous intrinsic uctuations at rates 10100 times slower than the usual EEGfrequencies [14]. These uctua- tions persist during tasks, rest, wakefulness [14], sleep and even anesthesia [15] and their correlations reect the underlying connectivity of the functional units of the brain. In other words, task-based imaging is no longer the only means of identifying neural networks because intrinsic Review Glossary Anticorrelations: negative correlations in which one value increases as the other decreases. These are observed in R-fMRI data, even in the absence of regression with the global signal as a nuisance covariate, but doing so enhances their detection. The neurophysiological significance of cerebral anticorrelations remains unknown but they are generally observed between competing neural systems such as the default network and the frontoparietal control network. Electrocorticography (ECoG): electroencephalography with the electrodes applied directly to the exposed surface of the brain to record electrical activity from the cerebral cortex. ECoG may be performed either in the operating room during surgery or outside of surgery. Because a craniotomy is required to implant the electrode grid, ECoG is an invasive procedure. ECoG is the gold standard for defining epileptogenic zones in clinical practice. ECoG data have served to validate R-fMRI findings as relevant to neuronal processes, and not simply ascribable to hemodynamic or physiological epiphenomena. Systems neuroscience: a subdiscipline of neuroscience and systems biology that studies the function of neural circuits and systems. It is an umbrella term, encompassing a number of areas of study concerned with how nerve cells behave when connected together to form neural networks. Corresponding author: Castellanos, F.X. ([email protected]). 1364-6613/$ see front matter 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.tics.2011.11.007 Trends in Cognitive Sciences, January 2012, Vol. 16, No. 1 17 relationships are continuously encoded in the spontaneous activity of the brain, and these can be most easily appre- ciated during rest [16,17]. The patterns formed by these relationships can be displayed as statistical maps that have the same appearance as task-evoked activation maps, but they do not represent the invariants associated with specic cognitions or behaviors [18]. Instead, they reect correlations resulting from stochastic (i.e. probabilistic) neural activity transmitted differentially across synaptic connections that vary in strength according to the life experience of the individual [19]. The patterns of synchrony of these high-amplitude, albeit ultra-slow, uctuations are extraordinarily robust across distinct populations and differences in scanner eld strength or scanning parameters [20], and are stable in testretest designs [21,22]. They have been validated in cross-species studies [15,23] and in humans with electrocorticography [17], and are exquisitely sensitive to age factors during development [24] and to psychopa- thology [25]. Along with recent results from resting-state functional connectivity approaches to ADHD [26], these converging lines of evidence support our overarching hy- pothesis that ADHD results from dysregulated or aber- rant interactions within and among large-scale neural systems. Dening neural systems in the human brain Neuronal connectivity can be dened at the microscale, in terms of single neurons, at the macroscale, at the level of brain regions and their pathways, and at an intermedi- ate level of minicolumns and their connection patterns [27]. Given currently available imaging methods and in- formatics capacity, the macroscale level is the most feasi- ble for achieving a rst draft of the human brain connectome [27], which is currently under way (http:// www.humanconnectomeproject.org/). Macroscale imaging based on the diffusion of water (diffusion tensor imaging) has begun to reveal the micro- structure of major white-matter tracts but is not yet capa- ble of providing a comprehensive survey of brain networks. Classical lesion studies, which were the basis for identify- ing the systems underlying language, motor control and perception, have been updated with modern imaging meth- ods and analytical techniques and continue to inform our understanding of neural systems [28]. Task-based functional imaging has also revealed many of the necessary elements of brain circuitry, but each individual contrast provides only a narrow-angle focus. When aggregated in meta-analyses, the results of thou- sands of such contrasts cumulatively delineate large-scale brain networks [16]. However, this objective is most ef- ciently achieved through R-fMRI, which captures the full repertoire of functional networks utilized by the brain in action [16]. Accordingly, R-fMRI methods have been used to identify the default network [29], the dorsal and ventral attentional networks [30], and motor, visual and executive control systems [31] across laboratories [20] and clinical populations [25]. The remarkable replicability of neural networks in healthy young adults was recently demon- strated quantitatively in 1000 participants [12]. The data were subdivided into a discovery set of 500 and a replica- tion set of 500. Nearly all (97.4%) cortical vertices were assigned to the same seven cortical networks in the dis- covery and replication data sets. The parcellation of the human cerebral cortex based on all 1000 subjects is shown in Figure 1 and is freely available (http://surfer.nmr.mgh. harvard.edu/fswiki/CorticalParcellation_Yeo2011). At this level of resolution, the seven major networks can be heuris- tically identied as sensorimotor and primary visual cortex, limbic, dorsal attention, ventral attention, frontoparietal control and default networks. Although these are not the only relevant subdivisions of the cortex, they serve as refer- ence networks that canbe fruitfully examined inADHDand other clinical conditions. As the eld advances, we antici- pate that these networks will be fractionated and designat- ed in accordance with their functional ontologies, as illustrated later when we discuss the default network [32]. Candidate neural systems in ADHD Recent conceptualizations of ADHD have taken seriously the distributed nature of neuronal processing [10,11,33,34]. Most of the candidate networks have focused on prefrontalstriatalcerebellar circuits, although other posterior regions are also being proposed [10]. Until now, the evidence proposed in support of a particular hypothe- sized circuit has consisted mostly of between-group differ- ences in task-based fMRI activations [34] or anatomic volumetric differences [8]. Such results provide indirect evidence of validity, but individually they only illuminate subsets of circuit components. Their generalizability is also usually limited to the specic construct of interest and the population sampled, and by idiosyncratic methodological factors. Fortunately, the neural substrates of functional circuits that are identied piece-wise through task-based fMRI studies are continuously represented in the brain in Box 1. Mapping of intrinsic functional connectivity Functional magnetic resonance imaging (fMRI) without an explicit task, also known as resting-state fMRI (R-fMRI), has facilitated delineation of the intrinsic functional architecture of the brain based on the detection of patterns of coherence in low-frequency (<0.1 Hz) spontaneous fluctuations in blood oxygenation level dependent (BOLD) signals [13,92]. Acquisition of R-fMRI data can be performed on any scanner capable of registering BOLD signals, as long as participants remain still and supine for upwards of 5 min, preferably with eyes open to minimize heterogeneity of arousal levels. The exploitation of R-fMRI has increased dramatically in recent years, driven by the wealth of information provided and the ease of data collection across all clinical populations and throughout the entire lifespan. Although an ever-increasing number of techniques are being developed to harness the voluminous information present in even brief records of R-fMRI data, the most commonly applied involve regression analyses of explicitly selected regions of interest (seeds), on the one hand, and data-driven independent component analyses (ICA) on the other [93]. These methods converge in detecting multiple large-scale neural systems, including the most frequently examined, the default network, that represent universal features of the adult brain architecture [16]. R-fMRI is reliable and reproducible, and the signals yielded are so robust as to allow aggregation of raw data across multiple scanners and populations, even without prior coordination of data acquisition protocols [20]. This has advanced the goal of open sharing of primary data and analytical methods such as the 1000 Functional Connectomes Project (http://fcon_1000.projects.nitrc.org/), the ADHD-200 Sample (http://fcon_1000.projects.nitrc.org/indi/adhd200/) and the Human Connectome Project (http://www.humanconnectomeproject.org/). Review Trends in Cognitive Sciences January 2012, Vol. 16, No. 1 18 the form of intrinsic connectivity networks that are most easily measured during rest [16]. The recent compilation of reference networks for healthy young adults [12] raises questions of whether these circuits will provide a brain- based perspective for the process of characterizing brain behavior relationships across the lifespan and in clinical populations. Here, we briey review the recent ADHD neuroimaging literature within the context of these refer- ence resting-state functional networks [12]. Frontoparietal network The frontoparietal control circuit (Figure 1) includes the lateral frontal pole, anterior cingulate cortex (ACC), dor- solateral prefrontal cortex (dlPFC), anterior PFC (aPFC), lateral cerebellum, anterior insula, caudate and inferior parietal lobe [35]. This network is also known as the executive control circuit [36] because it underpins goal- directed executive processes and provides the exibility to congure information processing in response to changing task demands [37]. Executive control systems guide deci- sion making by integrating external information with internal representations. In ADHD, investigations of the most-studied executive control decits have focused on motor inhibition. Multiple studies have found hypoactivation in frontostriatal and frontoparietal circuits during inhibitory tasks in children with ADHD [7,34]. Besides parietal areas, nearly all the remaining regions implicated in the prefrontalstriatal cerebellar model of ADHD [8,38] are components of the frontoparietal circuit: ACC, aPFC, dlPFC, frontal pole, cerebellumand caudate. For example, it has been reported that the dorsal ACC is hypoactivated in ADHD during go/ no go, response inhibition and attentional tasks [11,34,39 41]. Similarly, dlPFC and ventrolateral PFC are hypoacti- vated in various tasks ranging from working memory to time discrimination [34,4143]. Involvement of the fronto- parietal network has also been conrmed by resting-state studies in ADHD [4447].
Left hemisphere Right hemisphere 7-network parcellation (N=1000) Visual Key: Somatomotor Dorsal attention Ventral attention Limbic Frontoparietal Default TRENDS in Cognitive Sciences Figure 1. Coarse (7-network) parcellation of the human cerebral cortex obtained through clustering of R-fMRI data of 1000 subjects. At this resolution, association cortex is distinguished from primary sensorimotor cortex. The association networks converged on and extended networks previously described in the resting-state literature, including the dorsal attention, ventral attention, frontoparietal control, and default networks. Adapted, with permission, from [12]. Review Trends in Cognitive Sciences January 2012, Vol. 16, No. 1 19 The frontoparietal network has been situated both spa- tially and conceptually as an intermediate systembetween two other major networks in the brain, the default network and the dorsal attentional network [35], which we address next. Dorsal and ventral attentional networks Figure 1 shows the reference dorsal and ventral attention- al networks [30], which form key components of the atten- tional regulatory systems of the brain [48]. The ventral attentional network, closely related to circuits referred to as the salience network [36] or the cinguloopercular net- work [49], is involved in monitoring for salient (behavior- ally relevant) stimuli and in interrupting ongoing activity when appropriate. The ventral attentional network is anchored by the temporoparietal junction, the supramar- ginal gyrus, frontal operculum and anterior insula [48]. The network most likely to be affected by the ventral is the dorsal attentional network, which mediates goal-di- rected, top-down executive control processes, particularly in reorienting attention during visual attentional function- ing. Its key nodes are the intraparietal sulcus (BA 40) and the frontal eye elds (BA 6), which are the main regions involved in attention shifting and in the control of spatial attention [48]. The literature does not support clear involvement of the ventral attentional network in ADHD, but it is also not yet possible to discard its potential participation. By contrast, abnormalities in precentral and parietal regions associat- ed with the dorsal attentional network clearly emerge in ADHD [7,34,50]. For example, during the performance of executive and response inhibition tasks, bilateral parietal regions (BA7, BA40) were among the main areas in which controls demonstrated signicantly greater probability of activation relative to ADHD subjects, along with motor regions (BA6) [7]. More recent studies have shown greater activation of the parietal cortex of ADHD patients during response inhibition [51,52]. In addition, abnormal patterns of parietal activity have been reported during working memory [5355] and attentional tasks [50,5661]. Visual network The visual cortex and the lateral temporal MT+ region are related to the superior parietal lobule and intraparietal sulcus, which are part of the dorsal attentional network. MT+ is also coupled to frontal regions such as precentral cortex and the frontal eye elds. MT+ is strongly function- ally correlated with primary visual areas such as V1 and V3 [12]. The occipital cortex has not previously been considered to be relevant to ADHD, even though neuroimaging studies in ADHD have found repeated differences in medial occip- ital cortex (BA18, BA19) [7,51,56,62,63]. Occipital cortex interacts with the dorsal attentional network to maintain attention [64] and suppress attention to irrelevant stimuli [65]. Failure to ignore extraneous stimuli is one of the core symptoms of ADHD. A recent structural neuroimaging study in medication-nave adults with ADHD found signif- icant bilateral reduction of gray-matter volume only in early visual cortex [66]. In a 33-year follow-up of childhood ADHD, persistence of the diagnosis was associated with decreased cortical thickness in medial occipital cortex among other regions (Figure 2) [9]. In functional studies, children with ADHD show deactivation of parietal and occipital regions during spatial tasks [62,63] whereas adults with ADHD show occipital hyperactivation on inhi- bition, working memory and attentional tasks [51,56,67]. A resting-state study in children with ADHD found
Ventral t - s t a t i s t i c s 4.5 2.7 Right Ventral Dorsal Left Non-ADHD (n=57) > ADHD persistents (n=17) FDR 0.05 Dorsal TRENDS in Cognitive Sciences Figure 2. Cortical thickness analysis reveals occipital involvement in ADHD. In a 33-year longitudinal follow-up study, adults with ADHD persisting from childhood showed significantly decreased cortical thickness in multiple regions, including medial occipital cortex (arrow) relative to non-ADHD controls. Reproduced, with permission, from [9]. Review Trends in Cognitive Sciences January 2012, Vol. 16, No. 1 20 decreased small-world network nodal efciency in multiple brain regions including visual cortex [47]. These ndings suggest that visual function and its regulation by atten- tional processes should be further investigated in ADHD. Motor network The rst brain network identied by characterizing intrin- sic functional connectivity was the motor system [14]. As recently reviewed, R-fMRI analyses detect synchrony in spontaneous low-frequency uctuations between primary motor cortex, primary sensory cortex, secondary sensory cortex, supplementary motor area (SMA), ventral premo- tor cortex, putamen, thalamus and cerebellum [68]. Remarkably, despite the incontrovertible salience of motoric hyperactivity in children with ADHD, there have been few neuroimaging studies of the motor system in ADHD[34,6971]. When performing simple motor tapping, children with ADHD exhibited decreased activation in primary motor cortex relative to controls [69]. Intra-subject variability, which is generally increased in ADHD[72], was positively related to pre-SMA activation in children with ADHD, whereas in healthy controls variability was in- versely related to pre-SMA activation [70]. In a study of adults with ADHD during paced and unpaced tapping, hypoactivations in ADHD were found both in timing-relat- ed circuits and in motor and premotor cortex [71]. In a non- imaging study that directly probed the motor system, intracortical inhibition was measured with short-interval paired-pulse transcranial magnetic stimulation [73]. Chil- dren with ADHD showed markedly reduced cortical inhi- bition, which was correlated with deciencies in motor performance [73]. This recent literature suggests that
(b) (a) (d) PHC Rsp pIPL vMPFC TempP TPJ HF+ dMPFC pIPL Rsp PHC HF+ vMPFC TempP LTC dMPFC TPJ Key: (c) L PCC 0.000 - 0.249 0.250 - 0.449 0.500 - 0.749 R HF+ LTC TempP TPJ Rsp aMPFC dMPFC pIPL PHC vMPFC LTC PCC vMPFC aMPFC dMPFC Rsp TempP HF+ PHC PCC pIPL TPJ LTC LTC Medial temporal lobe subsystem Dorsal medial prefrontal cortex subsystem y = -20 z = -16 z = 2 x = -10 x = -50 aMPFC TRENDS in Cognitive Sciences Figure 3. Fractionation of the default network. Default network core hubs are shown in yellow, the dorsomedial prefrontal cortex subsystem is shown in blue, and regions comprising the medial temporal lobe subsystem are in green. (a) The 11 seeds defined a priori using functional connectivity approaches. (b) The 11 seeds projected onto an inflated brain. (c) Correlation strengths among regions within the default network are shown using network centrality measures. The size of the circle represents the centrality of a given node. The anterior medial prefrontal cortex (aMPFC) and posterior cingulate cortex (PCC) are the core hubs of the network and both are significantly connected to every other node. Negative correlations are shown with a dotted line. (d) The two clusters resulting from centrality analyses. dMPFC, dorsomedial prefrontal cortex; TPJ, temporoparietal junction; LTC, lateral temporal cortex; TempP, temporal pole; vMPFC, ventromedial prefrontal cortex; pIPL, posterior inferior parietal lobe; Rsp, retrosplenial cortex; PHC, parahippocampal cortex; HF+, hippocampal formation. Reproduced, with permission, from [32]. Review Trends in Cognitive Sciences January 2012, Vol. 16, No. 1 21
x = 0 y = -16 z = 18 9.5 -9.5 z - s c o r e z - s c o r e (a) (b) (c) 4 3 2 1 0 -1 -2 -3 -4 Time (in TRs) (d) r = .67 r = .40 0.4 0.35 0.3 0.25 0.2 0.15 0.1 0.05 0 0.4 0.35 0.3 0.25 0.2 0.15 0.1 0.05 0 -1 -0.98 -0.96 -0.94 -0.92 Default mode/task-positive correlation I n c o n g r u e n t
R T
C V C o n g r u e n t
R T
C V Default mode/task-positive correlation -0.9 -0.89 -0.86 -0.84 -0.82 -1 -0.98 -0.96 -0.94 -0.92 -0.9 -0.88 -0.86 -0.84 -0.82 -0.8 TRENDS in Cognitive Sciences Figure 4. Anticorrelations between neural networks. (a) Mid-sagittal, coronal and axial views of anticorrelated networks extracted through region-of-interest-based functional connectivity analyses. The task-positive network shown in yelloworange includes the frontoparietal network; the default network is shown in purple. (b) Mid- sagittal, coronal and axial views of anticorrelated networks extracted through independent component analyses showing substantial overlap of the two methods. The frontoparietal network is shown in yelloworange and the default network in purple. (c) Time series of default and frontoparietal networks for one participant with Pearson r=0.97 during performance of a slow event-related Eriksen flanker task. (d) The strength of this relationship was inversely related to intra-subject variability of response time across participants. Reproduced, with permission, from [74]. Review Trends in Cognitive Sciences January 2012, Vol. 16, No. 1 22 continued examination of the motor system in ADHD would be productive. The default network The most-studied intrinsic functional connectivity system is known as the default network of the brain (Figure 1) because its high-amplitude uctuations, which are consis- tently diminished during cognitive tasks and increased during rest, were described as representing the physiolog- ical baseline of the brain [13,29]. As shown in Figure 3, the default network contains two hubs, the anterior medial PFC (aMPFC) and posterior cingulate cortex (PCC), and two subcomponent systems, the dorsomedial prefrontal cortex (dMPFC) subsystem and the medial temporal lobe (MTL) subsystem [32]. In a tour de force model of how to combine task-based and resting-state data, Andrews- Hanna et al. established that the dMPFC subsystem is activated when subjects perform self-referential cognitive processes anchored in the present; the MTL subsystem is preferentially activated by cognitions regarding projection of ones self into the future [32]. Beyond its roles intypically developing individuals, the default network is implicated across the full range of psychiatric disorders [36]. Intriguingly, default network uctuations are 1808 out of phase with uctuations in networks that are activated during externally oriented tasks, presumably reecting competition between opposing processes for processing resources [74]. Stronger negative correlation between de- fault and frontoparietal control networks and greater co- herence within networks is related to better behavioral performance, as shown in Figure 4 [74]. This is consistent with the nding that diminished suppression of default network activity is associated with attentional lapses [75] and with the suggestion that inter-individual differences in performance are related to the efciency of interactions among brain regions [76]. In 2007, Sonuga-Barke and Castellanos suggested that ADHDcould be considered a default network disorder [77]. They reasoned that the default network in ADHDmight be refractory to regulation by other neural systems, and thus would produce intrusions into or disruptions of ongoing cognition and behavior, which would manifest as periodic lapses in on-task performance, a hallmark of ADHD [72]. Decreased default network coherence has been found in ADHD[78] and decreased default network suppression has been related to increased intra-individual variability in a small sample of children with ADHD [79]; ongoing studies will test the hypothesis that intercorrelation between the default network and cognitive control networks [74] under- pins ADHD attentional lapses. In the meantime, an interesting result was obtained by comparing healthy young subjects scanned after rested wakefulness and after 24 h of full sleep deprivation [80]. Sleep deprivation produced an increase in intra-subject variability and degraded attentional performance. These were paralleled by decreases in default network functional connectivity and weaker anticorrelation between the de- fault network and anti-correlated regions [74,80]. Deter- mination of whether similar effects are found in participants with ADHD is likely to be informative. Treatment with methylphenidate normalized default network suppression in ventromedial PFC and PCC in 16 youths with ADHD, each scanned twice [81]. An inter- action among methylphenidate, motivational level (high and low incentives) and diagnosis was found in default network suppression during a go/no-go task [82]. Control children deactivated the default network under both high and low incentive conditions, similarly to children with ADHD who were scanned when on methylphenidate. By contrast, children with ADHD scanned while off medica- tion only deactivated the default network during the high Box 2. Transcranial direct current stimulation Transcranial direct current stimulation (tDCS) is a noninvasive neurostimulation technique that uses small electrodes as pathways for delivering low-amplitude electric current to cerebral regions of interest. This technique is considered a promising tool for clinical populations because of its safe application in humans [94]. Transcranial DCS is performed through a battery-powered device that emits a constant current through two electrodes. Each device has a positively charged electrode (anode) and a negatively charged electrode (cathode). The applied current is diminished substantially in crossing the scalp, but sufficient electricity flows into the brain to produce neuronal effects. Application of anodal direct current increases and of cathodal direct current decreases the underlying cortical excitability. The extent of neuronal effects depends on stimulation duration, electrode size and current density [95,96]. Recent pre- and post-tDCS R-fMRI sessions led to the suggestion that neurostimulation may have therapeutic relevance for ADHD. Real but not sham tCDS applied to left and right dlPFC produced decreases in default network synchrony and increases in antic- orrelated network coherence [87]. Stimulation over primary motor cortex modulated functional connectivity of corticostriatal and thalamocortical circuits [88]. Real versus sham anodal tDCS of dlPFC significantly enhanced default and frontoparietal network synchrony, which may underlie reports of improvements in cognitive performance [89]. Box 3. Questions for future research Is increased intra-individual response time variability in ADHD ascribable to abnormalities within a single system, such as the default network, or to the interrelationships among default, executive control and limbic-motivational networks? Can imbalances between the dorsal attentional network and/or frontoparietal network, on the one hand, and the default network, on the other, be redressed through intracranial direct current electrical stimulation? Are these networks useful units for examining the effects of pharmacological and behavioral treat- ments? Are they relevant to lack of response to treatments? How do the seven large intrinsic connectivity networks map onto striatal [97], thalamic [98] and cerebellar [99] connectivity circuits? What is the appropriate resolution level to examine cortico striatothalamocortical and corticothalamocerebellar circuitry in ADHD? Are the seven large networks overly inclusive? Do visual network abnormalities in ADHD relate to its linkage to the dorsal attentional network? Are they primary or compensa- tory? How are they related to inattention symptoms? Can neurobiological subtypes of ADHD be established on the basis of neural network profiles? Can such neural network profiles be used to track treatment response? How will ADHD-related differences in neural network profiles change across development in cross-sectional and longitudinal studies? How do the default network subcomponent functions (self-related processing in the present vs the future) relate to ADHD symptoms? Is the medial temporal subcomponent linked to future prospection associated with faulty decision making in ADHD and related disorders? Review Trends in Cognitive Sciences January 2012, Vol. 16, No. 1 23 incentive condition [82]. The authors concluded that nor- malization of default network suppression by either meth- ylphenidate or increased incentives points to dysregulation of the default network rather than to its fundamental impairment. Conversely, abnormalities residing in the default network are suggested by the emergence of signi- cantly greater gray-matter volume in precuneus and PCC in a structural meta-analysis of ADHD[8]. Taken together, these ndings suggest that the interplay of default, cogni- tive control and limbic networks is likely to be a key factor in suboptimal neural functioning in ADHD. A counter-argument to the above analysis could be that we have simply proposed that the entire brain is involved in ADHD. Although such an argument is not without merit, in that global volumetric reductions have been consistently related to the disorder [5], we believe it is far more likely that interactions among the candidate functional networks we have identied will form distin- guishable neurobiological patterns that can provide the basis for meaningful subtyping of this heterogeneous con- dition. Concluding remarks Functional connectivity reveals replicable brain networks that are likely to be relevant to our understanding of brain behavior relationships in disorders such as ADHD. Char- acterization of the spatial extent of such networks [18] or their intra- or inter-network coherence for individuals has become feasible [74,8385]. What is nowneeded is to relate such brain network proles [12] to neuropsychological and clinical measures [86]. The networks we have mentioned are unlikely to be exhaustive or equally relevant to all individuals with ADHD, but they provide a straightfor- ward framework for converging attempts to parse the pertinent dimensions of symptoms and constructs, in keeping with the US National Institute of Mental Health Research Domains Criteria project (http://www.nimh.nih. gov/research-funding/rdoc/nimh-research-domain-criteria- rdoc.shtml). Brain networks are situated in the conceptual sweet spot between genes and behaviors, and represent the most tractable opportunities to formulate hypotheses link- ing these multiple levels (Box 3). However, in the absence of manipulation, neuroimaging methods remain correlational and unable to inform on causal mechanisms. Nevertheless, imaging pre- and post-treatment can reveal biomarkers linked to causal pathways. Besides pharmacological and behavioral treat- ments for ADHD, novel approaches such as tDCS should be considered (Box 2). Non-invasive tDCS can produce tran- sient increases or decreases in cortical excitability which target specic regions and circuits and their interactions [8790]. Despite substantial evidence that tDCS modu- lates neural processes, its clinical benets have not been demonstrated convincingly, even for chronic pain [91]. We suggest that future tDCS studies could use R-fMRI to select candidate patients and circuits, and that imaging be used to document the appropriate placement of stimu- lating electrodes. Evidence of short-term improvement in symptoms and corresponding changes in the circuits tar- geted could then be used to justify more prolonged treat- ment regimens, with the goal of determining whether transcranial electrical stimulation holds therapeutic prom- ise in ADHD(see also Box 3 for a list of questions for future research). References 1 American Psychiatric Association (2000) Diagnostic and Statistical Manual of Mental Disorders, (4th edn), American Psychiatric Association text revision 2 Polanczyk, G. et al. (2007) The worldwide prevalence of ADHD: a systematic review and metaregression analysis. Am. J. Psychiatry 164, 942948 3 Centers for Disease Control (2010) Increasing prevalence of parent- reported attention-decit/hyperactivity disorder among children United States, 2003 and 2007. MMWR Morb. Mortal. Wkly. Rep. 59, 14391443 4 Simon, V. et al. (2009) Prevalence and correlates of adult attention- decit hyperactivity disorder: meta-analysis. Br. J. Psychiatry 194, 204211 5 Krain, A.L. and Castellanos, F.X. (2006) Brain development and ADHD. Clin. Psychol. Rev. 26, 433444 6 Castellanos, F.X. et al. (2006) Characterizing cognition in ADHD: beyond executive dysfunction. Trends Cogn. Sci. 10, 117123 7 Dickstein, S.G. et al. (2006) The neural correlates of attention decit hyperactivity disorder: an ALE meta-analysis. J. Child Psychol. Psychiatry 47, 10511062 8 Nakao, T. et al. (2011) Gray matter volume abnormalities in ADHD: voxel-based meta-analysis exploring the effects of age and stimulant medication. Am. J. Psychiatry 168, 11541163 9 Proal, E. et al. (2011) Brain gray matter decits at 33-year follow-up in adults with attention-decit/hyperactivity disorder established in childhood. Arch. Gen. Psychiatry 68, 11221134 10 Makris, N. et al. (2009) Towards conceptualizing a neural systems- based anatomy of attention-decit/hyperactivity disorder. Dev. Neurosci. 31, 3649 11 Bush, G. (2010) Attention-decit/hyperactivity disorder and attention networks. Neuropsychopharmacology 35, 278300 12 Yeo, B.T. et al. (2011) The organization of the human cerebral cortex estimated by functional connectivity. J. Neurophysiol. 106, 11251165 13 Raichle, M.E. and Snyder, A.Z. (2007) Adefault mode of brain function: a brief history of an evolving idea. Neuroimage 37, 10831090 14 Fox, M.D. and Raichle, M.E. (2007) Spontaneous uctuations in brain activity observed with functional magnetic resonance imaging. Nat. Rev. Neurosci. 8, 700711 15 Vincent, J.L. et al. (2007) Intrinsic functional architecture in the anaesthetized monkey brain. Nature 447, 8386 16 Smith, S.M. et al. (2009) Correspondence of the brains functional architecture during activation and rest. Proc. Natl. Acad. Sci. U.S.A. 106, 1304013045 17 Keller, C.J. et al. (2011) Intrinsic functional architecture predicts electrically evoked responses in the human brain. Proc. Natl. Acad. Sci. U.S.A. 108, 1030810313 18 Mennes, M. et al. (2010) Inter-individual differences in resting state functional connectivity predict task-induced BOLD activity. Neuroimage 50, 16901701 19 Deco, G. et al. (2011) Emerging concepts for the dynamical organization of resting-state activity in the brain. Nat. Rev. Neurosci. 12, 4356 20 Biswal, B.B. et al. (2010) Toward discovery science of human brain function. Proc. Natl. Acad. Sci. U.S.A. 107, 47344739 21 Shehzad, Z. et al. (2009) The resting brain: unconstrained yet reliable. Cereb. Cortex 19, 22092229 22 Van Dijk, K.R. et al. (2010) Intrinsic functional connectivity as a tool for human connectomics: theory, properties, and optimization. J. Neurophysiol. 103, 297321 23 Margulies, D.S. et al. (2009) Precuneus shares intrinsic functional architecture in humans and monkeys. Proc. Natl. Acad. Sci. U.S.A. 106, 2006920074 24 Power, J.D. et al. (2010) The development of human functional brain networks. Neuron 67, 735748 25 Zhang, D. and Raichle, M.E. (2010) Disease and the brains dark energy. Nat. Rev. Neurol. 6, 1528 26 Konrad, K. and Eickhoff, S.B. (2010) Is the ADHD brain wired differently? A review on structural and functional connectivity in attention decit hyperactivity disorder. Hum. Brain Mapp. 31, 904916 Review Trends in Cognitive Sciences January 2012, Vol. 16, No. 1 24 27 Sporns, O. et al. (2005) The human connectome: a structural description of the human brain. PLoS Comput. Biol. 1, e42 28 Glascher, J. et al. (2010) Distributed neural system for general intelligence revealed by lesion mapping. Proc. Natl. Acad. Sci. U.S.A. 107, 47054709 29 Buckner, R.L. et al. (2008) The brains default network: anatomy, function, and relevance to disease. Ann. N. Y. Acad. Sci. 1124, 138 30 Fox, M.D. et al. (2006) Spontaneous neuronal activity distinguishes human dorsal and ventral attention systems. Proc. Natl. Acad. Sci. U.S.A. 103, 1004610051 31 Damoiseaux, J.S. et al. (2006) Consistent resting-state networks across healthy subjects. Proc. Natl. Acad. Sci. U.S.A. 103, 1384813853 32 Andrews-Hanna, J.R. et al. (2010) Functional-anatomic fractionation of the brains default network. Neuron 65, 550562 33 Durston, S. et al. (2011) Differentiating frontostriatal and fronto cerebellar circuits in attention-decit/hyperactivity disorder. Biol. Psychiatry 69, 11781184 34 Rubia, K. (2011) Cool inferior frontostriatal dysfunction in attention- decit/hyperactivity disorder versus hot ventromedial orbitofrontal limbic dysfunction in conduct disorder: a review. Biol. Psychiatry 69, e69e87 35 Vincent, J.L. et al. (2008) Evidence for a frontoparietal control system revealed by intrinsic functional connectivity. J. Neurophysiol. 100, 33283342 36 Menon, V. (2011) Large-scale brain networks and psychopathology: a unifying triple network model. Trends Cogn. Sci. 15, 483506 37 Liston, C. et al. (2006) Anterior cingulate and posterior parietal cortices are sensitive to dissociable forms of conict in a task-switching paradigm. Neuron 50, 643653 38 Valera, E.M. et al. (2007) Meta-analysis of structural imaging ndings in attention-decit/hyperactivity disorder. Biol. Psychiatry 61, 1361 1369 39 Durston, S. et al. (2007) Neural and behavioral correlates of expectancy violations in attention-decit hyperactivity disorder. J. Child Psychol. Psychiatry 48, 881889 40 Konrad, K. et al. (2006) Dysfunctional attentional networks in children with attention decit/hyperactivity disorder: evidence from an event- related functional magnetic resonance imaging study. Biol. Psychiatry 59, 643651 41 Smith, A.B. et al. (2008) Reduced activation in right lateral prefrontal cortex and anterior cingulate gyrus in medication-naive adolescents with attention decit hyperactivity disorder during time discrimination. J. Child Psychol. Psychiatry 49, 977985 42 Schneider, M. et al. (2006) Anatomical and functional brain imaging in adult attention-decit/hyperactivity disorder (ADHD) a neurological view. Eur. Arch. Psychiatry Clin. Neurosci. 256 (Suppl. 1), i32i41 43 Vaidya, C.J. and Stollstorff, M. (2008) Cognitive neuroscience of attention decit hyperactivity disorder: current status and working hypotheses. Dev. Disabil. Res. Rev. 14, 261267 44 Cao, Q. et al. (2006) Abnormal neural activity in children with attention decit hyperactivity disorder: a resting-state functional magnetic resonance imaging study. NeuroReport 17, 10331036 45 Zang, Y.F. et al. (2007) Altered baseline brain activity in children with ADHD revealed by resting-state functional MRI. Brain Dev. 29, 8391 46 Cao, X. et al. (2009) Abnormal resting-state functional connectivity patterns of the putamen in medication-naive children with attention decit hyperactivity disorder. Brain Res. 1303, 195206 47 Wang, L. et al. (2009) Altered small-world brain functional networks in children with attention-decit/hyperactivity disorder. Hum. Brain Mapp. 30, 638649 48 Corbetta, M. et al. (2008) The reorienting system of the human brain: from environment to theory of mind. Neuron 58, 306324 49 Dosenbach, N.U. et al. (2008) Adual-networks architecture of top-down control. Trends Cogn. Sci. 12, 99105 50 Stevens, M.C. et al. (2007) An FMRI auditory oddball study of combined-subtype attention decit hyperactivity disorder. Am. J. Psychiatry 164, 17371749 51 Dillo, W. et al. (2010) Neuronal correlates of ADHD in adults with evidence for compensation strategies a functional MRI study with a Go/No-Go paradigm. Ger. Med. Sci. 8, Doc09. DOI: 10.3205/000098 52 Karch, S. et al. (2010) Neural correlates (ERP/fMRI) of voluntary selection in adult ADHD patients. Eur. Arch. Psychiatry Clin. Neurosci. 260, 427440 53 Wolf, R.C. et al. (2009) Regional brain activation changes and abnormal functional connectivity of the ventrolateral prefrontal cortex during working memory processing in adults with attention-decit/ hyperactivity disorder. Hum. Brain Mapp. 30, 22522266 54 Burgess, G.C. et al. (2010) Attentional control activation relates to working memory in attention-decit/hyperactivity disorder. Biol. Psychiatry 67, 632640 55 Bayerl, M. et al. (2010) Disturbed brain activation during a working memory task in drug-naive adult patients with ADHD. NeuroReport 21, 442446 56 Schneider, M.F. et al. (2010) Impairment of fronto-striatal and parietal cerebral networks correlates with attention decit hyperactivity disorder (ADHD) psychopathology in adults a functional magnetic resonance imaging (fMRI) study. Psychiatry Res. 183, 7584 57 Rubia, K. et al. (2007) Temporal lobe dysfunction in medication-naive boys with attention-decit/hyperactivity disorder during attention allocation and its relation to response variability. Biol. Psychiatry 62, 9991006 58 Tamm, L. et al. (2006) Parietal attentional system aberrations during target detection in adolescents with attention decit hyperactivity disorder: event-related fMRI evidence. Am. J. Psychiatry 163, 1033 1043 59 Rubia, K. et al. (2009) Methylphenidate normalises activation and functional connectivity decits in attention and motivation networks in medication-naive children with ADHD during a rewarded continuous performance task. Neuropharmacology 57, 640652 60 Cubillo, A. et al. (2010) Reduced activation and inter-regional functional connectivity of fronto-striatal networks in adults with childhood attention-decit hyperactivity disorder (ADHD) and persisting symptoms during tasks of motor inhibition and cognitive switching. J. Psychiatr. Res. 44, 629639 61 Banich, M.T. et al. (2009) The neural basis of sustained and transient attentional control in young adults with ADHD. Neuropsychologia 47, 30953104 62 Silk, T. et al. (2008) Dysfunction in the fronto-parietal network in attention decit hyperactivity disorder (ADHD): an fMRI study. Brain Imag. Behav. 2, 123131 63 Vance, A. et al. (2007) Right parietal dysfunction in children with attention decit hyperactivity disorder, combined type: a functional MRI study. Mol. Psychiatry 12, 826832 64 Shulman, G.L. et al. (2009) Interaction of stimulus-driven reorienting and expectation in ventral and dorsal frontoparietal and basal ganglia cortical networks. J. Neurosci. 29, 43924407 65 Capotosto, P. et al. (2009) Frontoparietal cortex controls spatial attention through modulation of anticipatory alpha rhythms. J. Neurosci. 29, 58635872 66 Ahrendts, J. et al. (2011) Visual cortex abnormalities in adults with ADHD: a structural MRI study. World J. Biol. Psychiatry 12, 260270 67 Hale, T.S. et al. (2007) Atypical brain activation during simple and complex levels of processing in adult ADHD: an fMRI study. J. Atten. Disord. 11, 125140 68 Deco, G. and Corbetta, M. (2011) The dynamical balance of the brain at rest. Neuroscientist 17, 107123 69 Mostofsky, S.H. et al. (2006) Atypical motor and sensory cortex activation in attention-decit/hyperactivity disorder: a functional magnetic resonance imaging study of simple sequential nger tapping. Biol. Psychiatry 59, 4856 70 Suskauer, S.J. et al. (2008) fMRI of intrasubject variability in ADHD: anomalous premotor activity with prefrontal compensation. J. Am. Acad. Child Adolesc. Psychiatry 47, 11411150 71 Valera, E.M. et al. (2010) Neural substrates of impaired sensorimotor timing in adult attention-decit/hyperactivity disorder. Biol. Psychiatry 68, 359367 72 Castellanos, F.X. et al. (2005) Varieties of attention-decit/ hyperactivity disorder-related intra-individual variability. Biol. Psychiatry 57, 14161423 73 Gilbert, D.L. et al. (2011) Motor cortex inhibition: a marker of ADHD behavior and motor development in children. Neurology 76, 615 621 74 Kelly, A.M.C. et al. (2008) Competition between functional brain networks mediates behavioral variability. Neuroimage 39, 527537 75 Weissman, D.H. et al. (2006) The neural bases of momentary lapses in attention. Nat. Neurosci. 9, 971978 Review Trends in Cognitive Sciences January 2012, Vol. 16, No. 1 25 76 Rypma, B. et al. (2006) Neural correlates of cognitive efciency. Neuroimage 33, 969979 77 Sonuga-Barke, E.J. and Castellanos, F.X. (2007) Spontaneous attentional uctuations in impaired states and pathological conditions: a neurobiological hypothesis. Neurosci. Biobehav. Rev. 31, 977986 78 Castellanos, F.X. et al. (2008) Cingulateprecuneus interactions: a new locus of dysfunction in adult attention-decit/hyperactivity disorder. Biol. Psychiatry 63, 332337 79 Fassbender, C. et al. (2009) A lack of default network suppression is linked to increased distractibility in ADHD. Brain Res. 1273, 114128 80 De Havas, J.A. et al. (2011) Sleep deprivation reduces default mode network connectivity and anti-correlation during rest and task performance. Neuroimage DOI: 10.1016/j.neuroimage.2011.08.026 81 Peterson, B.S. et al. (2009) An FMRI study of the effects of psychostimulants on default-mode processing during Stroop task performance in youths with ADHD. Am. J. Psychiatry 166, 12861294 82 Liddle, E.B. et al. (2011) Task-related default mode network modulation and inhibitory control in ADHD: effects of motivation and methylphenidate. J. Child Psychol. Psychiatry 52, 761771 83 Cox, C.L. et al. (2010) Your resting brain CAREs about your risky behavior. PLoS ONE 5, e12296 84 Di Martino, A. et al. (2009) Autistic traits in neurotypical adults are related to cinguloinsular functional connectivity. Am. J. Psychiatry 166, 891899 85 Koyama, M.S. et al. (2011) Resting-state functional connectivity indexes reading competence in children and adults. J. Neurosci. 31, 86178624 86 Chabernaud, C. et al. (2011) Dimensional brainbehavior relationships in children with attention-decit/hyperactivity disorder. Biol. Psychiatry DOI: 10.1016/j.biopsych.2011.08.013 87 Pena-Gomez, C. et al. (2011) Modulation of large-scale brain networks by transcranial direct current stimulation evidenced by resting-state functional MRI. Brain Stimul. DOI: 10.1016/j.brs.2011.08.006 88 Polania, R. et al. (2011) Modulating corticostriatal and thalamo cortical functional connectivity with transcranial direct current stimulation. Hum. Brain Mapp. DOI: 10.1002/hbm.21380 89 Keeser, D. et al. (2011) Prefrontal transcranial direct current stimulation changes connectivity of resting-state networks during fMRI. J. Neurosci. 31, 1528415293 90 Alon, G. et al. (2011) Non-invasive electrical stimulation of the brain (ESB) modies the resting-state network connectivity of the primary motor cortex: a proof of concept fMRI study. Brain Res. 1403, 3744 91 OConnell, N.E. et al. (2011) Non-invasive brain stimulation techniques for chronic pain. A report of a Cochrane systematic review and meta- analysis. Eur. J. Phys. Rehabil. Med. 47, 309326 92 Lowe, M.J. (2010) A historical perspective on the evolution of resting- state functional connectivity with MRI. MAGMA 23, 279288 93 Margulies, D.S. et al. (2010) Resting developments: a review of fMRI post-processing methodologies for spontaneous brain activity. MAGMA 23, 289307 94 Utz, K.S. et al. (2010) Electried minds: transcranial direct current stimulation (tDCS) and galvanic vestibular stimulation (GVS) as methods of non-invasive brain stimulation in neuropsychology a review of current data and future implications. Neuropsychologia 48, 27892810 95 DaSilva, A.F. et al. (2011) Electrode positioning and montage in transcranial direct current stimulation. J. Vis. Exp. DOI: 10.3791/2744 96 Zaghi, S. et al. (2010) Noninvasive brain stimulation with low-intensity electrical currents: putative mechanisms of action for direct and alternating current stimulation. Neuroscientist 16, 285307 97 Di Martino, A. et al. (2008) Functional connectivity of human striatum: a resting state fMRI study. Cereb. Cortex 18, 27352747 98 Zhang, D. et al. (2008) Intrinsic functional relations between human cerebral cortex and thalamus. J. Neurophysiol. 100, 17401748 99 Buckner, R.L. et al. (2011) The organization of the human cerebellum estimated by intrinsic functional connectivity. J. Neurophysiol. 106, 23222345 Review Trends in Cognitive Sciences January 2012, Vol. 16, No. 1 26