Participants with attention-deficit/hyperactivity disorder displayed greater resting-state functional connectivity between the anterior cingulate cortex (ACC) and superior parietal and cerebellar areas than did control subjects. A, Left ACC (x = 18, y = −50, z = 64). B, Right ACC (x = 34, y = −76, z = −30). Values next to the images indicate the sections in axial view.
In control subjects, the ventral attention network showed more significant resting-state functional connectivity bilaterally than did participants with attention-deficit/hyperactivity disorder. A, Left temporoparietal junction and left middle temporal gyrus (x = −62, y = −4, z = −8). B, Left ventral frontal cortex (VFC) and left inferior operculum (x = −30, y = 6, z = 30). C, Right VFC and left precentral gyrus (x = −50, y = 2, z = 28). Values next to the images indicate the sections in axial view.
eTable. Demographic and clinical data of ADHD participants and healthy controls
eFigure 1. Heightened right lateral functional connectivity for ADHD patients compared to healthy controls within the cognitive control network between the right DLPFC and posterior cingulate cortex
eFigure 2. Association between Wender scores and connectivity between the left TPJ and the left middle temporal gyrus
McCarthy H, Skokauskas N, Mulligan A, Donohoe G, Mullins D, Kelly J, Johnson K, Fagan A, Gill M, Meaney J, Frodl T. Attention Network Hypoconnectivity With Default and Affective Network Hyperconnectivity in Adults Diagnosed With Attention-Deficit/Hyperactivity Disorder in Childhood. JAMA Psychiatry. 2013;70(12):1329-1337. doi:10.1001/jamapsychiatry.2013.2174
Copyright 2013 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.
The neurobiological underpinnings of attention-deficit/hyperactivity disorder (ADHD) and particularly those associated with the persistence of ADHD into adulthood are not yet well understood. The correlation patterns in spontaneous neural fluctuations at rest are known as resting-state functional connectivity (RSFC) and could characterize ADHD-specific connectivity changes.
To determine the specific location of possible ADHD-related differences in RSFC between adults diagnosed as having ADHD in childhood and control subjects.
Using resting-state functional magnetic resonance imaging, we calculated and compared functional connectivity from attention, affective, default, and cognitive control networks involved in the psychopathology of ADHD between the ADHD and control groups.
University psychiatric service and magnetic resonance imaging research center.
Sixteen drug-free adults (5 women and 11 men; mean age, 24.5 years) diagnosed with combined-type ADHD in childhood and 16 healthy controls matched for age (mean age, 24.4 years), sex, handedness, and educational level recruited from the community.
Functional magnetic resonance imaging.
Main Outcomes and Measures
Connectivity data from ventral and dorsal attention, affective, default, and cognitive control networks and ADHD symptoms derived from ADHD-specific rating instruments.
Adults with ADHD showed significantly decreased RSFC within the attention networks and increased RSFC within the affective and default mode and the right lateralized cognitive control networks compared with healthy controls (P < .01, familywise error for whole-brain cluster correction). Lower RSFC in the ventral and dorsal attention network was significantly correlated with higher levels of ADHD symptoms (P < .001).
Conclusions and Relevance
These RSFC findings might underpin a biological basis for adult ADHD and are functionally related to persistent inattention, disturbance in cognitive control, and emotional dysregulation in adults with ADHD. These findings need to be understood in the context of all aspects of brain function in ADHD.
Attention deficit-hyperactivity/disorder (ADHD) is characterized by the onset of developmentally inappropriate levels of impaired inattention, hyperactivity, and impulsivity before the age of 7 years.1 Attention deficit-hyperactivity/disorder is one of the most common mental disorders, with a worldwide prevalence estimate of 5.29% and a heritability estimate of about 76%.2 Attention deficit-hyperactivity/disorder is no longer considered a disorder exclusive to childhood; Faraone et al3 estimate that 15% of adults with a previous diagnosis of childhood ADHD meet the full criteria for ADHD, 65% show partial remission, and only 20% show complete remission in adulthood.
Accurately diagnosing ADHD in adults is difficult, because the diagnostic criteria for ADHD are founded on research conducted among individuals aged 4 to 17 years.4 The clinical presentation of ADHD changes with increasing age3; the hyperactive and impulsive symptoms of childhood ADHD become less pronounced, whereas inattentive symptoms persist.5 Emotional dysregulation and disorganization characteristically accompany persistent inattentive symptoms in adult ADHD.6 Attention-deficit/hyperactivity disorder in adulthood is socially and occupationally debilitating and is significantly associated with comorbid depression, anxiety, and substance abuse.4
Neuroimaging may provide some insights into the neurobiological underpinnings of the persistence and remission of ADHD symptoms during the disease course from childhood to adulthood. Most ADHD-related functional magnetic resonance imaging (fMRI) differences have been found in task-based studies.7 Researchers have proposed that ADHD-related deficiencies in functional circuits, identified through task-based fMRI studies, may sculpt the topography of resting-state networks.8,9 Thus, a growing interest in the use of resting-state fMRI, which does not require the use of a task, has become a popular means of complementing the results of task-based fMRI studies while providing a new perspective on the psychopathology of ADHD.10
Resting-state fMRI allows for the examination of large-scale neural systems that exhibit spontaneous synchronous fluctuations during goal-directed behavior and non–goal-directed behaviors such as wakeful rest, sleep, and anesthesia.9 These low-frequency (<0.1 Hz) spontaneous fluctuations in blood oxygen level–dependent (BOLD) signal correlate with interactions between adjacent and nonadjacent brain areas that form spatially distributed networks of brain function.11 Functional connectivity is the observed correlation in spontaneous neural activity between brain areas at rest.8
To improve our understanding of resting-state functional connectivity (RSFC) in adults with a childhood diagnosis of ADHD, we extracted the coordinates for regions of interest involved in the psychopathological features of ADHD12 from 5 neural networks.13,14 These coordinates were used as correlates with all other time series in the brain to determine temporally coherent networks of RSFC.14
The first network included was the affective network. This network contains integrated regions of the affective subdivision of the anterior cingulate cortex (ACC), amygdala, nucleus accumbens, hypothalamus, anterior insula, hippocampus, and orbitofrontal cortex, with reciprocal connections to autonomic, visceromotor, and endocrine systems.14 This network is involved in emotional regulation and monitoring of the salience of motivational stimuli.14 Affective disorders are highly comorbid with adult ADHD,4 and previous resting-state studies examining affective disorders and children with ADHD have found increased connectivity compared with controls within affective network regions, which were attributed to emotional control deficits.14- 16 These findings suggest increased sensitivity to affective stimuli in ADHD.17 Heightened functional connectivity for participants with ADHD compared with control subjects within this network was anticipated.
The second network, the ventral attention network, uses the temporoparietal junction (TPJ) and ventral frontal cortex (VFC) to reorient attention to salient behaviorally relevant stimuli.13 A recent meta-analysis of task-based fMRI studies found that an ADHD group had consistently hypoactivated ventral attention network regions compared with a control group in tasks requiring manipulation of attention in response to target stimuli.12 Because the neural deficits apparent during task-based studies in ADHD may indicate an RSFC deficit,8 hypoconnectivity throughout this network was expected.
The third network, the bilateral dorsal attention network, uses regions such as the intraparietal sulcus (IPS) and the frontal eye field (FEF) to enable the control of spatial attention through the selection of sensory stimuli based on internal goals or expectations and links them to appropriate motor responses.13 A meta-analysis of fMRI task studies12 showed that for participants with ADHD, intraparietal regions of the dorsal attention network were routinely hypoactivated. In addition, reduced functional connectivity of the dorsal attention network has been found in resting-state fMRI studies of child and adult samples with ADHD.18,19 Thus, we assumed that the ADHD group would elicit hypoconnectivity compared with the control group within this network.
The fourth network, the cognitive control network, uses regions such as the bilateral dorsolateral prefrontal cortex (DLPFC), presupplementary motor area, inferior frontal junction, anterior insular cortex, dorsal premotor cortex, and posterior parietal cortices.20 This network is involved in inhibition and goal-directed decision making, which is impaired in ADHD.21 A recent meta-analysis22 of inhibition and attentional tasks found DLPFC activity to be reduced in participants with ADHD compared with controls. At rest, the ADHD group previously displayed decreased functional connectivity involving this network compared with controls.23 These results suggest that use of this network is diminished in the participants with ADHD compared with controls.23
The fifth network is the default mode network, which contains the precuneus/posterior cingulate cortex, medial prefrontal cortex, and dorsal ACC and acts as a form of functional connectivity baseline thought to reflect intrinsic brain activity.8 Typically, the spontaneous fluctuations emitted by the default mode network are routinely heightened at rest and are then reduced during goal-directed tasks that activate frontal attentional networks.9 Resting-state studies21,24,25 investigating the default mode network in ADHD found heterogeneous results in the form of reduced functional connectivity between the medial prefrontal cortex and the posterior cingulate/precuneus cortices within the default mode network in participants with ADHD compared with controls. Also, connectivity was increased between the default mode network and sensory brain regions in participants with ADHD compared with controls.15 Therefore, we aimed to confirm reduced functional connectivity within default mode regions and increased connectivity between the default mode network and the sensory regions.
The aim of this study was to explore RSFC in ADHD and to determine the localization and specificity of ADHD-related connectivity differences between adults diagnosed with ADHD in childhood and controls by examining these 5 predefined neural networks. We sought to confirm our hypothesis that participants with ADHD would show increased connectivity in the affective network and less connectivity within the attention and cognitive control networks compared with the control group. Moreover, we aimed to confirm increased connectivity between the default mode network and sensory brain regions and reduced connectivity within the default mode network itself in ADHD.
Sixteen adults with combined-type ADHD who underwent careful clinical assessment as children (mean [SD] age at diagnosis, 8.9 [2.1] years) when taking part in genetic and neuropsychological studies26,27 were compared with 16 healthy controls matched for age, sex, handedness, and educational level. Full-scale IQ also did not differ between groups significantly (Supplement [eTable]). Handedness was determined using the Edinburgh Inventory.28 Educational and occupational attainments were based on the Hollingshead Four Factor Index of Social Status.29 Full-scale IQ was measured using the Wechsler Adult Intelligence Scale, Fourth Edition30; the subscales of Verbal Comprehension, Perceptual Reasoning, Working Memory, and Processing Speed were used to compute full-scale composite IQ scores.
Two participants with ADHD were medication naive. Ten participants with ADHD were drug-free for a mean of 11.6 (SD, 4.2) years but had a history of methylphenidate hydrochloride treatment (mean [SD] months of medication, 20.6 [32.8]). Four participants were still being treated with methylphenidate and were required to undergo a washout period 48 hours before investigation. Exclusion criteria consisted of previous head injury with loss of consciousness, comorbid psychiatric disorder or disease, a history of hydrocortisone use, and current alcohol or substance abuse and/or dependency. All participants were interviewed and underwent diagnosis and screening for any potential comorbidity according to exclusion criteria using the Structured Clinical Interview for DSM-IV Axis I Disorders31 by a psychiatrist (D.M., J.K., or T.F.).
Self-administered and observer-rated scales were completed by all participants for diagnostic and correlational purposes. The rating scales used were the Structured Clinical Interview for DSM-IV Axis II Disorders Personality Questionnaire32; the Wender Utah Rating Scale (WURS)33 with a control cutoff score of 46; the Conners Adult ADHD Observer and Self-report rating scales in which T scores of greater than 65 across all 4 scales are indicative of clinically elevated symptoms34; the Hamilton Rating Scale for Depression35; and the Beck Depression Inventory.36 Ethical approval for the study was granted by the ethics committees of The Adelaide and Meath Hospital, Dublin, and St James’ Hospital. Each participant was given detailed oral and written information and had to sign an informed consent form before participation. Participants received €30 ($40).
Functional and structural images from each participant were obtained with an MRI scanner (Philips Achieva; Philips Medical Systems Netherland BV) operating at 3 T. The functional images were collected in single runs using gradient echoplanar imaging (echo time, 28 milliseconds; repetition time, 2000 milliseconds; field of view, 131 mm; flip angle, 90°) sensitive to BOLD contrast (T2* weighting). A total of 37 contiguous sections with 3.2-mm thickness were acquired parallel to the anterior-posterior commissure plane (3 mm approximately isotropic resolution), providing complete brain coverage. The fMRI run included 220 volumes acquired continuously lasting 7.2 minutes in total. Structural data (for definitive atlas transformation) included a high-resolution sagittal, 3-dimensional, T1-weighted, turbo gradient-echo sequence scan (echo time, 3.9 milliseconds; repetition time, 8.5 milliseconds; inversion time, 1060 milliseconds; flip angle, 8°; 256 × 240 acquisition matrix, 1 × 1 × 1-mm voxels).
We used SPM software (SPM8; Wellcome Trust Centre for Neuroimaging [http://www.fil.ion.ucl.ac.uk/spm/software/spm8/]) to preprocess fMRI data in the following steps: first, the compensation of systematic, section-dependent time shifts; second, the elimination of systematic odd- and even-section intensity differences due to interleaved acquisition; and third, rigid body correction for interframe head motion within and across runs. Data were excluded if motion parameters exceeded 3 mm in any direction or 3.0° of any angular motion during the scan. Next, we coregistered the structural T1 image to the functional scans. Spatial normalization to standard 3 × 3 × 3-mm Montreal Neurological Institute space was then applied to the functional images and to the structural image to allow for intersubject analysis. Functional resting-state data were then spatially smoothed (smoothing full width at half maximum, 8 mm).
Using resting-state software (CONN; National Institutes of Health Blueprint for Neuroscience Research [http://www.nitrc.org/projects/conn/]), the data were temporally band-pass filtered (filter range, 0.009-0.08). Several sources of spurious variance along with their temporal derivatives then were removed from the data by linear regression, such as signal from regions centered in the white matter, cerebrospinal fluid, and movement. CONN implements the component-based noise-correction method strategy for physiological and other noise source reduction.37 This method has the advantage of not requiring external monitoring of physiological fluctuations. Compared with methods that rely on global signal regression, the component-based noise-reduction method allows for interpretation of anticorrelations because there is no regression of the global signal within this method. This approach may enhance the sensitivity and specificity of positive correlations and produce comparable negative correlations.38 This regression procedure removes fluctuations unlikely to be involved in specific regional correlations.
To compute functional connectivity maps corresponding to a selected seed region of interest, we correlated the regional time course against all other voxels within the brain. Based on data from 2 previous resting-state studies,13,14 we explored connectivity within the affective network, ventral attention system, dorsal attention system, cognitive control network, and default mode network. Correlation maps were produced by extracting the BOLD time course from a seed region and then by computing the correlation coefficient between that time course and the time course from all other brain voxels. The following seed regions of interest with a 5-mm radius were extracted for these 5 specific neural networks: the ACC (x = ±10, y = −35, z = 2) in the affective network, the TPJ (x = ±53, y = −48, z = 20) and VFC (x = ±37, y = −18, z = 1) in the ventral attention system, the IPS (x = ±27, y = −58, z = 49) and FEF (x = ±24, y = −13, z = 51) in the dorsal attention system, the DLPFC (x = ±36, y = 27, z = 29) in the cognitive control network, and the precuneus (x = ±7, y = −60, z = 21) in the default mode network. For the ventral (TPJ and VFC) and dorsal (IPS and FEF) attention network, we extracted 2 connectivity maps. The principal technique used was computation of whole-brain, voxelwise intrinsic functional connectivity maps.
We analyzed resting-state fMRI data in SPM8 using 2-sample t tests to determine significant differences in functional connectivity between ADHD participants and controls using a familywise error whole-brain–corrected threshold of P < .01. The threshold was reduced to P < .01 from P < .05 because we analyzed connectivity within the 5 different networks. Mean connectivity data were extracted from areas that showed significant differences between ADHD participants and controls. We used commercially available software (SPSS, version 16; SPSS Inc) to calculate Spearman rank correlation coefficients between ADHD symptoms from the Conners Adult ADHD Self-report rating scale and the WURS using a Bonferroni-corrected P threshold of .004 based on 12 extracted areas. Because the Beck Depression Inventory and Hamilton Rating Scale for Depression scores were different between ADHD participants and controls (although the values and variance were still below the threshold for depression), these were added as covariates in additional analysis of covariance in SPM8 to observe the difference between the ADHD and control groups.
Compared with the control group, participants with ADHD displayed significantly greater functional connectivity bilaterally throughout the affective network (Figure 1) and the dorsal attention, in the right hemisphere of the cognitive control network (Supplement [eFigure 1]), and in the right hemisphere of the default mode network (Table 1).
For ADHD participants compared with controls, significantly less functional connectivity was found between the following regions in the ventral attention network: the left TPJ and the left temporal gyrus, the left VFC and the left frontal inferior operculum, and the right VFC and the left precentral gyrus (Table 2 and Figure 2). In the dorsal attention network, participants with ADHD showed significantly reduced functional connectivity between the right IPS and the right fusiform gyrus compared with the control group (Table 2). The cognitive control network showed significantly reduced functional connectivity between the left DLPFC and the right inferior occipital lobe for the ADHD group compared with the control group (Table 2). All group comparison results held true when the Beck Depression Inventory and Hamilton Rating Scale for Depression scores were included as covariates.
To correct for multiple correlations, P ≤ .004 was used as the significance threshold. Higher WURS scores were significantly negatively correlated with reduced functional connectivity in the ventral attention network between the left TPJ and left middle temporal gyrus (r = −0.81; P < .001) (Supplement [eFigure 2]). A reduction in functional connectivity between the left TPJ and left middle temporal gyrus was also significantly negatively correlated with higher DSM-IV inattentive symptoms (r = −0.72; P = .001), higher DSM-IV hyperactive/impulsive symptoms (r = −0.54; P = .02), and higher total DSM-IV symptom scores (r = −0.59; P = .01).
Correlations that did not reach significance after considering multiple testing but were P < .05 uncorrected included higher levels of ADHD symptoms during childhood. These symptoms were measured with the WURS and negatively correlated with a reduction in functional connectivity in the dorsal attention network between the left FEF and the inferior occipital lobe (r = −0.53; P = .03) and between the right FEF and the left middle occipital gyrus (r = −0.66; P = .005). Reduced functional connectivity in the ventral attention network between the left TPJ and left middle temporal gyrus also correlated significantly with higher Hamilton Rating Scale for Depression scores in controls (r = −0.52; P = .04).
No significant correlations were found when considering multiple correlations for the ADHD group. With the less conservative threshold of P < .05, uncorrected findings can be reported that need further replication: higher WURS scores significantly negatively correlated with reduced functional connectivity in the ventral attention network between the right TPJ and right fusiform gyrus (r = −0.51; P = .046). Reduced functional connectivity in the dorsal attention network between the left FEF and the inferior occipital lobe was negatively correlated with higher total DSM-IV-TR symptom scores (r = −0.58; P = .02). Reduced functional connectivity between the right IPS and right fusiform gyrus was significantly negatively correlated with higher DSM-IV-TR inattentive symptoms (r = −0.56; P = .02) and with higher total DSM-IV-TR symptom scores (r = −0.52; P = .04) in the ADHD group.
To our knowledge, this study represents the first time the affective, ventral, and dorsal attention networks have been explored in tandem with the cognitive control and default mode networks to determine the localization and specificity of ADHD-related RSFC differences in adults diagnosed with ADHD in childhood. A main finding within our study was the pronounced lack of significant RSFC within the ventral attention network for ADHD participants compared with controls. The ventral attention network directs attention toward salient stimuli, and an activation increase within the dorsal attention network occurs with the ventral attention network when reorientation of attention is required.13 Disrupted information exchange between the ventral attention network and the dorsal attention network has been linked to adult ADHD symptoms, such as intensified distractibility and perseverance.39 Because hypofunction of the ventral attention network has been found during task-based fMRI studies,12 our RSFC finding bolsters the dysregulation apparent within this network for participants with ADHD.40 The link between ADHD and hypofunction of the ventral attention network is supported by our correlational analysis finding. We found that even for controls, and less significantly for ADHD participants, higher WURS and overall DSM-IV symptom scores were significantly associated with reduced functional connectivity between the left TPJ of the ventral attention network and the left middle temporal gyrus. The latter region mediates attentional processing of auditory information, which is necessary for rehearsing verbally mediated task strategies in ADHD.41 Participants with ADHD have been found to display slower reaction times and hypoactivation of this region compared with controls during aural and visual attention tasks.41,42 This finding suggests that elevated ADHD symptoms in controls and participants with ADHD could indicate a reduced capacity for control of auditory and visuospatial attention.
Compared with the control group, we found significantly increased RSFC for ADHD participants between the bilateral FEF and occipital areas within the dorsal attention network. Heightened occipital region RSFC in ADHD has been linked to poor inhibition of sensory perception.43 First, inhibitory control may require a greater effort for ADHD participants compared with controls owing to the lack of ventral attention network RSFC for which they may compensate.44 Second, this finding may result from the dysfunctional interplay between the IPS and bilateral FEF and their afferent connections outside the dorsal attention network.45 Moreover, we found reduced RSFC—although not significantly reduced after correction for multiple comparisons—for ADHD participants compared with controls between the right IPS and the right fusiform gyrus, which is involved in semantic retrieval.40 Hypoactivity of the fusiform gyrus in ADHD has been linked to poor inhibition of emotional memory.40
We found more RSFC for the ADHD participants than for the control group throughout the affective network. Participants with ADHD displayed significantly more RSFC between the right ACC of the affective network and the right cerebellum than did controls. The cerebellum is bidirectionally linked with regions modulating motor and affective output, including the parietal cortices and prefrontal regions.46 Similar changes in the cerebellum have also been found in adults with major depressive disorder and children and adolescents with ADHD.15,16,47 We found significantly higher RSFC between the left ACC and the bilateral superior parietal lobules in the ADHD compared with the control groups (P < .001 and P = .02, respectively), which to date has been found solely in children with ADHD.18,23
However, our affective network results are dissimilar to those of reduced affective network RSFC commonly found in adults with ADHD.19,48 Different exposure to methylphenidate across studies, which has been linked to the normalization of ACC dysregulation in ADHD,49 may account for these differences. The heterogeneity of these results warrant further research in longitudinal and treatment studies to clarify the effect of heightened affective network RSFC on emotional control in ADHD.
Participants with ADHD exhibited significantly less connectivity between the left DLPFC and the inferior occipital lobe than did the controls but displayed more RSFC in the cognitive control network between the right DLPFC and the posterior cingulate cortex. The DLPFC has been linked to adaptive control,20 and the posterior cingulate is associated with introspection.48 In more than 50% of cases, ADHD in adulthood is associated with a comorbid affective disorder.4 This finding is relevant to our cognitive control network finding because greater right DLPFC RSFC has been linked to heightened negative affect in those with and without major depressive disorder.50
The RSFC within the default mode network was not significantly changed with regard to a P < .01 whole-brain correction. However, we found a trend for heightened RSFC within the default mode network between the right precuneus and the right precentral gyrus in ADHD participants compared with controls. The overactivation of the precuneus during task engagement in ADHD has been linked to instances of error and attention lapses.48 The right precentral gyrus of the posterior frontal lobe is involved in the execution of sensorimotor demands, hyperfunction of which has been linked to impaired motor inhibition in ADHD.12 Our finding supports that of Tian et al,51 who found heightened RSFC between the precuneus and sensory motor regions in ADHD participants; this result was linked to the processing of somatosensory information at rest more than in controls. In addition, our finding may also reflect compensation within the ADHD group for the dysregulated attention network connectivity that typically coordinates sensorimotor attention.44
Our findings are contrary to previous findings of reduced functional connectivity among the superior precuneus, medial prefrontal cortex, and posterior cingulate cortices in ADHD participants.21,24,25 One study seeded their default mode network RSFC study of adults with ADHD within the posterior cingulate cortex25 and another examined RSFC in this network using the frontal control region as a seed region.48 Our findings may differ from these studies because we examined default mode network connectivity according to published methods13,14 and, unlike Uddin et al,25 we did not examine RSFC within the default mode network exclusively.
Our findings must be considered in light of limiting factors. First, the size and heterogeneous treatment history of the ADHD group limited a comparison between medication-naive vs treated subgroups. Within the ADHD group, 2 participants (13%) were treatment naïve, 10 (63%) were drug free but had a previous history of methylphenidate drug treatment spanning a mean of 20.6 months, and 4 (25%) withheld their stimulant medication for 48 hours. A history of treatment with stimulant medication may be relevant because improved suppression of the default mode network during task engagement52 and cognitive-enhancement effects in adults with ADHD53 have been linked to a history of stimulant medication therapy. At present, few longitudinal studies have explored the effect of stimulant medication from childhood to adulthood. Thus, more prospective, longitudinal, long-term studies are needed to investigate the effect of long-term stimulant treatment on brain structure and function.
Second, those still being treated within our investigation may have been susceptible to rebound effects caused by abrupt discontinuation of stimulant medication therapy that has been linked to heightened motor and ACC activity for ADHD participants.54 To account for a rebound effect, we applied a washout period of 48 hours for participants using medication. The examination of a medication-naive sample is considered optimal and could alleviate concerns about the effect of medication on RSFC in adult ADHD.
The inclusion of ADHD participants who underwent careful diagnosis in childhood and again in adulthood by the Structured Clinical Interview for DSM-IV Axis I Disorders, all of whom still fulfilled diagnostic criteria for combined-type ADHD, allowed us to examine the developmental progress of ADHD from childhood into adulthood.19 Potential recall bias associated with retrospective diagnoses of ADHD in childhood was also avoided by studying this specific group.3 Our results are valuable because they were corrected for multiple comparisons and were strong enough to survive conservative Bonferroni testing, which reduced type I errors.
Our findings shed light on ADHD-related RSFC and its role in symptoms presented by adults in whom childhood ADHD does not remit. Affective network hyperconnectivity might be related to emotional problems seen in adults with ADHD. In addition, ventral attention network hypoconnectivity may underlie attentional impairment in ADHD and seems to be closely related in particular to persistent inattention. Finally, the tentative compensation with hyperconnectivity of regions within the dorsal attention, cognitive control, and default mode networks requires future examination. These ADHD-related functional connectivity findings need to be understood in the context of a complex picture with social, environmental, physiological, biological, and genetic factors influencing brain function in ADHD.
Submitted for Publication: October 30, 2012; final revision received March 20, 2013; accepted March 21, 2013.
Corresponding Author: Thomas Frodl, MD, MA, Department of Psychiatry and Institute of Neuroscience, Trinity College Dublin, Lloyd Building, Dublin 2, Ireland (firstname.lastname@example.org).
Published Online: October 16, 2013. doi:10.1001/jamapsychiatry.2013.2174.
Author Contributions:Study concept and design: McCarthy, Skokauskas, Mullins, Johnson, Meaney, Frodl.
Acquisition of data: McCarthy, Mulligan, Donohoe, Mullins, Kelly, Fagan, Gill, Frodl.
Analysis and interpretation of data: McCarthy, Johnson, Frodl.
Drafting of the manuscript: McCarthy, Skokauskas, Kelly, Johnson, Meaney, Frodl.
Critical revision of the manuscript for important intellectual content: McCarthy, Skokauskas, Mulligan, Donohoe, Mullins, Johnson, Fagan, Gill, Meaney, Frodl.
Statistical analysis: McCarthy, Johnson, Frodl.
Obtained funding: Skokauskas, Mulligan, Johnson, Gill, Frodl.
Administrative, technical, and material support: Skokauskas, Mullins, Kelly, Johnson, Fagan, Meaney, Frodl.
Study supervision: Meaney, Frodl.
Conflict of Interest Disclosures: None reported.
Funding/Support: The study was supported by grant H01421 from the Health Research Board, Ireland (Drs Skokauskas, Johnson, Gill, and Frodl), and Stokes Professorship grant G20330 from the Science Foundation (Dr Frodl).
Role of the Sponsors: The Health Research Board Ireland and the Science Foundation had no role in the design and conduct of the study; in the collection, analysis, and interpretation of the data; or in the preparation, review, or approval of the manuscript.