Aberrant Cross-Brain Network Interaction in Children With Attention-Deficit/Hyperactivity Disorder and Its Relation to Attention Deficits: A Multisite and Cross-Site Replication Study

Annual Research Review: Translational machine learning for child and adolescent psychiatry

Children and adolescents could benefit from the use of predictive tools that facilitate personalized diagnoses, prognoses, and treatment selection. Such tools have not yet been deployed using traditional statistical methods, potentially due to the limitations of the paradigm and the need to leverage large amounts of digital data. This review will suggest that a machine learning approach could address these challenges and is designed to introduce new readers to the background, methods, and results in the field. A rationale is first introduced followed by an outline of fundamental elements of machine learning approaches. To provide an overview of the use of the techniques in child and adolescent literature, a scoping review of broad trends is then presented. Selected studies are also highlighted in order to draw attention to research areas that are closest to translation and studies that exhibit a high degree of experimental innovation. Limitations to the research, and machine learning approaches generally, are outlined in the penultimate section highlighting issues related to sample sizes, validation, clinical utility, and ethical challenges. Finally, future directions are discussed that could enhance the possibility of clinical implementation and address specific questions relevant to the child and adolescent psychiatry. The review gives a broad overview of the machine learning paradigm in order to highlight the benefits of a shift in perspective towards practically oriented statistical solutions that aim to improve clinical care of children and adolescents.

Differential structure-function network coupling in the inattentive and combined types of attention deficit hyperactivity disorder

The heterogeneous presentation of inattentive and hyperactive-impulsive core symptoms in attention deficit hyperactivity disorder (ADHD) warrants further investigation into brain network connectivity as a basis for subtype divisions in this prevalent disorder. With diffusion and resting-state functional magnetic resonance imaging data from the Healthy Brain Network database, we analyzed both structural and functional network efficiency and structure-functional network (SC-FC) coupling at the default mode (DMN), executive control (ECN), and salience (SAN) intrinsic networks in 201 children diagnosed with the inattentive subtype (ADHD-I), the combined subtype (ADHD-C), and typically developing children (TDC) to characterize ADHD symptoms relative to TDC and to test differences between ADHD subtypes. Relative to TDC, children with ADHD had lower structural connectivity and network efficiency in the DMN, without significant group differences in functional networks. Children with ADHD-C had higher SC-FC coupling, a finding consistent with diminished cognitive flexibility, for all subnetworks compared to TDC. The ADHD-C group also demonstrated increased SC-FC coupling in the DMN compared to the ADHD-I group. The correlation between SC-FC coupling and hyperactivity scores was negative in the ADHD-I, but not in the ADHD-C group. The current study suggests that ADHD-C and ADHD-I may differ with respect to their underlying neuronal connectivity and that the added dimensionality of hyperactivity may not explain this distinction.

Empirical Bayes estimation of pairwise maximum entropy model for nonlinear brain state dynamics

The pairwise maximum entropy model (pMEM) has recently gained widespread attention to exploring the nonlinear characteristics of brain state dynamics observed in resting-state functional magnetic resonance imaging (rsfMRI). Despite its unique advantageous features, the practical application of pMEM for individuals is limited as it requires a much larger sample than conventional rsfMRI scans. Thus, this study proposes an empirical Bayes estimation of individual pMEM using the variational expectation-maximization algorithm (VEM-MEM). The performance of the VEM-MEM is evaluated for several simulation setups with various sample sizes and network sizes. Unlike conventional maximum likelihood estimation procedures, the VEM-MEM can reliably estimate the individual model parameters, even with small samples, by effectively incorporating the group information as the prior. As a test case, the individual rsfMRI of children with attention deficit hyperactivity disorder (ADHD) is analyzed compared to that of typically developed children using the default mode network, executive control network, and salient network, obtained from the Healthy Brain Network database. We found that the nonlinear dynamic properties uniquely established on the pMEM differ for each group. Furthermore, pMEM parameters are more sensitive to group differences and are better associated with the behavior scores of ADHD compared to the Pearson correlation-based functional connectivity. The simulation and experimental results suggest that the proposed method can reliably estimate the individual pMEM and characterize the dynamic properties of individuals by utilizing empirical information of the group brain state dynamics.

Exploring characteristic features of attention-deficit/hyperactivity disorder: findings from multi-modal MRI and candidate genetic data

The current study examined whether machine learning features best distinguishing attention-deficit/hyperactivity disorder (ADHD) from typically developing children (TDC) can explain clinical phenotypes using multi-modal neuroimaging and genetic data. Cortical morphology, diffusivity scalars, resting-state functional connectivity and polygenic risk score (PS) from norepinephrine, dopamine and glutamate genes were extracted from 47 ADHD and 47 matched TDC. Using random forests, classification accuracy was measured for each uni- and multi-modal model. The optimal model was used to explain symptom severity or task performance and its robustness was validated in the independent dataset including 18 ADHD and 18 TDC. The model consisting of cortical thickness and volume features achieved the best accuracy of 85.1%. Morphological changes across insula, sensory/motor, and inferior frontal cortex were also found as key predictors. Those explained 18.0% of ADHD rating scale, while dynamic regional homogeneity within default network explained 6.4% of the omission errors in continuous performance test. Ensemble of PS to optimal model showed minor effect on accuracy. Validation analysis achieved accuracy of 69.4%. Current findings suggest that structural deformities relevant to salience detection, sensory processing, and response inhibition may be robust classifiers and symptom predictors of ADHD. Altered local functional connectivity across default network predicted attentional lapse. However, further investigation is needed to clarify roles of genetic predisposition.

Altered core networks of brain connectivity and personality traits in internet gaming disorder

BACKGROUND AND AIMS

Although the Internet has provided convenience and efficiency in many areas of everyday life, problems stemming from Internet use have also been identified, such as Internet gaming disorder (IGD). Internet addiction, which includes IGD, can be viewed as a behavioral addiction or impulse control disorder. This study investigated the altered functional and effective connectivity of the core brain networks in individuals with IGD compared to healthy controls (HCs).

METHODS

Forty-five adults with IGD and 45 HCs were included in this study. To examine the brain networks related to personality traits that influence problematic online gaming, the left and right central executive network (CEN) and the salience network (SN) were included in the analysis. Also, to examine changes in major brain network topographies, we analyzed the default mode network (DMN).

RESULTS

IGD participants showed lower functional connectivity between the dorsal lateral prefrontal cortex (DLPFC) and other regions in the CEN than HC participants during resting state. Also, IGD participants revealed reduced functional connectivity between the dorsal anterior cingulate cortex and other regions in the SN and lower functional connectivity in the medial prefrontal cortex of the anterior DMN. Notably, in IGD individuals but not HC individuals, there was a positive correlation between IGD severity and effective connectivity and a positive correlation between reward sensitivity and effective connectivity within the ventral striatum of the SN.

CONCLUSIONS

Problematic online gaming was associated with neurofunctional alterations, impairing the capacity of core brain networks.

Altered coupling of default-mode, executive-control and salience networks in Internet gaming disorder

Background:

Recently, a triple-network model suggested the abnormal interactions between the executive-control network (ECN), default-mode network (DMN) and salience network (SN) are important characteristics of addiction, in which the SN plays a critical role in allocating attentional resources toward the ECN and DMN. Although increasing studies have reported dysfunctions in these brain networks in Internet gaming disorder (IGD), interactions between these networks, particularly in the context of the triple-network model, have not been investigated in IGD. Thus, we aimed to assess alterations in the inter-network interactions of these large-scale networks in IGD, and to associate the alterations with IGD-related behaviors.

Methods:

DMN, ECN and SN were identified using group-level independent component analysis (gICA) in 39 individuals with IGD and 34 age and gender matched healthy controls (HCs). Then alterations in the SN-ECN and SN-DMN connectivity, as well as in the modulation of ECN versus DMN by SN, using a resource allocation index (RAI) developed and validated previously in nicotine addiction, were assessed. Further, associations between these altered network coupling and clinical assessments were also examined.

Results:

Compared with HCs, IGD had significantly increased SN-DMN connectivity and decreased RAI in right hemisphere (rRAI), and the rRAI in IGD was negatively associated with their scores of craving.

Conclusions:

These findings suggest that the deficient modulation of ECN versus DMN by SN might provide a mechanistic framework to better understand the neural basis of IGD and might provide novel evidence for the triple-network model in IGD.

Structural Covariance Networks in Children with Autism or ADHD

Background

While autism and attention-deficit/hyperactivity disorder (ADHD) are considered distinct conditions from a diagnostic perspective, clinically they share some phenotypic features and have high comorbidity. Regardless, most studies have focused on only one condition, with considerable heterogeneity in their results. Taking a dual-condition approach might help elucidate shared and distinct neural characteristics.

Method

Graph theory was used to analyse topological properties of structural covariance networks across both conditions and relative to a neurotypical (NT; n = 87) group using data from the ABIDE (autism; n = 62) and ADHD-200 datasets (ADHD; n = 69). Regional cortical thickness was used to construct the structural covariance networks. This was analysed in a theoretical framework examining potential differences in long and short-range connectivity, with a specific focus on relation between central graph measures and cortical thickness.

Results

We found convergence between autism and ADHD, where both conditions show an overall decrease in CT covariance with increased Euclidean distance between centroids compared with a NT population. The 2 conditions also show divergence. Namely, there is less modular overlap between the 2 conditions than there is between each condition and the NT group. The ADHD group also showed reduced cortical thickness and lower degree in hub regions than the autism group. Lastly, the ADHD group also showed reduced wiring costs compared with the autism groups.

Conclusions

Our results indicate a need for taking an integrated approach when considering highly comorbid conditions such as autism and ADHD. Furthermore, autism and ADHD both showed alterations in the relation between inter-regional covariance and centroid distance, where both groups show a steeper decline in covariance as a function of distance. The 2 groups also diverge on modular organization, cortical thickness of hub regions and wiring cost of the covariance network. Thus, on some network features the groups are distinct, yet on others there is convergence.

Salience network coupling is linked to both tobacco smoking and symptoms of attention deficit hyperactivity disorder (ADHD)

INTRODUCTION

Attention deficit hyperactivity disorder (ADHD) symptoms, even those below diagnostic threshold, enhance the likelihood of nicotine dependence, suggesting a neurobiological link between disorders. Of particular interest is the salience network (SN), which mediates attention to salient internal/external stimuli to guide behavior and is anchored by the dorsal anterior cingulate cortex (dACC) and bilateral anterior insula (AI). Disrupted interactions between the SN and the default mode (DMN) and central executive networks (CEN) have been noted in both ADHD and nicotine dependence. Further, enhanced intra-SN coupling between the dACC-AI influences aspects of nicotine dependence such as reactivity to smoking cues.

METHODS

To identify links between SN functional connectivity and ADHD symptoms in nicotine dependence, we compared 21 nicotine dependent individuals with 17 non-smokers on ADHD symptoms as measured by the ADHD self-report scale (ASRS) and resting state intra and inter-SN functional connectivity.

RESULTS

Relative to healthy controls, nicotine dependent individuals had significantly higher ASRS scores and greater dACC-AI coupling. No group differences were noted on inter-SN network coupling. A significant association was found between ASRS and dACC-AI coupling both in the entire cohort and specifically when evaluating nicotine dependent individuals alone.

CONCLUSIONS

The greater ASRS scores in nicotine dependent individuals is in line with existent literature and the stronger dACC-AI coupling in smokers further supports the role of this network in nicotine dependence. The significant association between dACC-AI coupling and ASRS suggests that intra-SN coupling strength may impact neurocognitive functioning associated with both ADHD symptoms and nicotine dependence.

Predicting clinical symptoms of attention deficit hyperactivity disorder based on temporal patterns between and within intrinsic connectivity networks

Attention deficit hyperactivity disorder (ADHD) is a common brain disorder with high prevalence in school-age children. Previously developed machine learning-based methods have discriminated patients with ADHD from normal controls by providing label information of the disease for individuals. Inattention and impulsivity are the two most significant clinical symptoms of ADHD. However, predicting clinical symptoms (i.e., inattention and impulsivity) is a challenging task based on neuroimaging data. The goal of this study is twofold: to build predictive models for clinical symptoms of ADHD based on resting-state fMRI and to mine brain networks for predictive patterns of inattention and impulsivity. To achieve this goal, a cohort of 74 boys with ADHD and a cohort of 69 age-matched normal controls were recruited from the ADHD-200 Consortium. Both structural and resting-state fMRI images were obtained for each participant. Temporal patterns between and within intrinsic connectivity networks (ICNs) were applied as raw features in the predictive models. Specifically, sample entropy was taken asan intra-ICN feature, and phase synchronization (PS) was used asan inter-ICN feature. The predictive models were based on the least absolute shrinkage and selectionator operator (LASSO) algorithm. The performance of the predictive model for inattention is r=0.79 (p<10^-8), and the performance of the predictive model for impulsivity is r=0.48 (p<10^-8). The ICN-related predictive patterns may provide valuable information for investigating the brain network mechanisms of ADHD. In summary, the predictive models for clinical symptoms could be beneficial for personalizing ADHD medications.

Novel subgroups of attention-deficit/hyperactivity disorder identified by topological data analysis and their functional network modular organizations

Attention-deficit/hyperactivity disorder (ADHD) is a clinically heterogeneous condition and identification of clinically meaningful subgroups would open up a new window for personalized medicine. Thus, we aimed to identify new clinical phenotypes in children and adolescents with ADHD and to investigate whether neuroimaging findings validate the identified phenotypes. Neuroimaging and clinical data from 67 children with ADHD and 62 typically developing controls (TDCs) from the ADHD-200 database were selected. Clinical measures of ADHD symptoms and intelligence quotient (IQ) were used as input features into a topological data analysis (TDA) to identify ADHD subgroups within our sample. As external validators, graph theoretical measures obtained from the functional connectome were compared to address the biological meaningfulness of the identified subtypes. The TDA identified two unique subgroups of ADHD, labelled as mild symptom ADHD (mADHD) and severe symptom ADHD (sADHD). The output topology shape was repeatedly observed in the independent validation dataset. The graph theoretical analysis showed a decrease in the degree centrality and PageRank in the bilateral posterior cingulate cortex in the sADHD group compared with the TDC group. The mADHD group showed similar patterns of intra- and inter-module connectivity to the sADHD group. Relative to the TDC group, the inter-module connectivity between the default mode network and executive control network were significantly increased in the sADHD group but not in the mADHD group. Taken together, our results show that the data-driven TDA is potentially useful in identifying objective and biologically relevant disease phenotypes in children and adolescents with ADHD.

Catechol-O-methyltransferase polymorphism is associated with the cortico-cerebellar functional connectivity of executive function in children with attention-deficit/hyperactivity disorder

The cerebellum, although traditionally considered a motor structure, has been increasingly recognized to play a role in regulating executive function, the dysfunction of which is a factor in attention-deficit/hyperactivity disorder (ADHD). Additionally, catechol-O-methyltransferase (COMT) polymorphism has been reported to be associated with executive function. We examined whether the cortico-cerebellar executive function network is altered in children with ADHD and whether COMT polymorphism is associated with the altered network. Thirty-one children with ADHD and thirty age- and IQ-matched typically developing (TD) controls underwent resting-state functional MRI, and functional connectivity of executive function-related Crus I/II in the cerebellum was analysed. COMT Val158Met genotype data were also obtained from children with ADHD. Relative to TD controls, children with ADHD showed significantly lower functional connectivity of the right Crus I/II with the left dorsolateral prefrontal cortex. Additionally, the functional connectivity of children with ADHD was modulated by COMT polymorphism, with Met-carriers exhibiting significantly lower functional connectivity than the Val/Val genotype. These results suggest the existence of variations, such as ethnic differences, in COMT genetic effects on the cortico-cerebellar executive function network. These variations contribute to heterogeneity in ADHD. Further neuroimaging genetics study might lead to the development of fundamental therapies that target ADHD pathophysiology.

Inconsistency in Abnormal Brain Activity across Cohorts of ADHD-200 in Children with Attention Deficit Hyperactivity Disorder

Many papers have shown results from the multi-site dataset of resting-state fMRI (rs-fMRI) in attention deficit hyperactivity disorder (ADHD), a data-sharing project named ADHD-200. However, few studies have illustrated that to what extent the pooled findings were consistent across cohorts. The present study analyzed three voxel-wise whole-brain metrics, i.e., amplitude of low-frequency fluctuation (ALFF), regional homogeneity (ReHo), and degree centrality (DC) based on the pooled dataset as well as individual cohort of ADHD-200. In addition to the conventional frequency band of 0.01-0.08 Hz, sub-frequency bands of 0-0.01, 0.01-0.027, 0.027-0.073, 0.073-0.198, and 0.198-0.25 Hz, were assessed. While the pooled dataset showed abnormal activity in some brain regions, e.g., the bilateral sensorimotor cortices, bilateral cerebellum, and the bilateral lingual gyrus, these results were highly inconsistent across cohorts, even across the three cohorts from the same research center. The standardized effect size was rather small. These findings suggested a high heterogeneity of spontaneous brain activity in ADHD. Future studies based on multi-site large-sample dataset should be performed on pooled data and single cohort data, respectively and the effect size must be shown.

Altered functional connectivity in default mode network in Internet gaming disorder: Influence of childhood ADHD

OBJECTIVE

Internet gaming disorder (IGD) is a type of behavioral addiction characterized by abnormal executive control, leading to loss of control over excessive gaming. Attention deficit and hyperactivity disorder (ADHD) is one of the most common comorbid disorders in IGD, involving delayed development of the executive control system, which could predispose individuals to gaming addiction. We investigated the influence of childhood ADHD on neural network features of IGD.

METHODS

Resting-state functional magnetic resonance imaging analysis was performed on 44 young, male IGD subjects with and without childhood ADHD and 19 age-matched, healthy male controls. Posterior cingulate cortex (PCC)-seeded connectivity was evaluated to assess abnormalities in default mode network (DMN) connectivity, which is associated with deficits in executive control.

RESULTS

IGD subjects without childhood ADHD showed expanded functional connectivity (FC) between DMN-related regions (PCC, medial prefrontal cortex, thalamus) compared with controls. These subjects also exhibited expanded FC between the PCC and brain regions implicated in salience processing (anterior insula, orbitofrontal cortex) compared with IGD subjects with childhood ADHD. IGD subjects with childhood ADHD showed expanded FC between the PCC and cerebellum (crus II), a region involved in executive control. The strength of connectivity between the PCC and cerebellum (crus II) was positively correlated with self-reporting scales reflecting impulsiveness.

CONCLUSION

Individuals with IGD showed altered PCC-based FC, the characteristics of which might be dependent upon history of childhood ADHD. Our findings suggest that altered neural networks for executive control in ADHD would be a predisposition for developing IGD.

Abnormal Resting-State Functional Connectivity of Insular Subregions and Disrupted Correlation with Working Memory in Adults with Attention Deficit/Hyperactivity Disorder

OBJECTIVES

Executive function (EF) deficits are major impairments in adults with attention deficit/hyperactivity disorder (ADHD). Previous studies have shown that the insula is involved in cognitive and EFs. However, the insula is highly heterogeneous in function, and few studies have focused on functional networks which related to specific insular subregions in adults with ADHD. We explored the functional networks of the insular subregions [anterior insula (AI), mid-insula (MI), and posterior insula (PI)]. Furthermore, their correlations with self-ratings of ecological EFs, including inhibition, shifting, and working memory were investigated.

METHODS

Resting-state functional magnetic resonance imaging data in 28 adults with ADHD and 30 matched healthy controls (HCs) were analyzed. The seed-based resting-state functional connectivity (RSFC) of the insular subregions was evaluated. We also investigated their associations with the Behavior Rating Inventory of Executive Function-Adult Version (BRIEF-A) inhibition, working memory, and shifting factor scores.

RESULTS

Compared with HCs, adults with ADHD showed altered RSFC of the AI, with the precuneus, precentral gyrus, and inferior temporal gyrus extended to the middle temporal gyrus, lingual gyrus, and superior occipital gyrus, respectively. There were no significant differences in RSFC of the MI and PI between the two groups. Within the HC group, working memory scores were associated with the RSFC of AI with precuneus and temporal gyrus. However, there was no correlation between these variables in the ADHD group.

CONCLUSION

The study evaluated RSFC patterns of the insular subregions in adults with ADHD for the first time. Altered RSFC of the AI which is a crucial region of salience network (SN) and part of regions in default mode network (DMN), were detected in adults with ADHD in both results with and without global signal regression (GSR), suggesting that disrupted SN-DMN functional connectivity may be involved in EF impairments in adults with ADHD, especially with respect to working memory. Deficits of the AI which is involved in salient stimuli allocation, might be associated with the pathophysiology of ADHD. The inconsistent results of MI and PI between analyses with and without GSR need further exploration.

Neural Correlates of Symptom Improvement Following Stimulant Treatment in Adults with Attention-Deficit/Hyperactivity Disorder

OBJECTIVE

The purposes of this study were to examine the impact of 3 weeks of amphetamine administration on intrinsic connectome-wide connectivity patterns in adults with attention-deficit/hyperactivity disorder (ADHD) and explore the association between stimulant-induced symptom improvement and functional connectivity alteration.

METHODS

Participants included 19 adults (age 20-55 years) diagnosed with ADHD using the Diagnostic and Statistical Manual of Mental Disorders, 4th ed., Text Revision (DSM-IV-TR) criteria per the Adult Clinician Diagnostic Scale taking part in amphetamine trials. For each patient, two 6-minute resting-state functional magnetic resonance imaging (R-fMRI) scans were acquired at baseline and after treatment. A fully data-driven multivariate analytic approach (i.e., multivariate distance matrix regression [MDMR]) was applied to R-fMRI data to characterize the distributed pharmacological effects in the entire functional connectome. Clinical efficacy was assessed using ADHD rating scale with adult prompts and the Adult Self-Report Scale v1.1 Symptom Checklist. We linked stimulant-induced functional connectivity changes to symptom amelioration using Spearman's correlation.

RESULTS

Three weeks of administration of a stimulant significantly reduced ADHD symptoms. MDMR-based analyses on R-fMRI data highlighted the left dorsolateral prefrontal cortex (DLPFC, a key cognitive control region) and the medial prefrontal cortex (MPFC, the anterior core of default network) whose distributed patterns of functional connectivity across the entire brain were altered by psychostimulants. Follow-up intrinsic functional connectivity revealed that stimulants specifically decreased the positive functional connectivity between DLPFC-insula, DLPFC-anterior cingulate cortex, and MPFC-insula. Importantly, these functional connectivity changes are associated with symptom improvement.

CONCLUSION

These results suggested that ADHD is associated with increased functional integration or decreased functional segregation between core regions of cognitive control, default, and salience networks. The apparent normalization of intrinsic functional interaction in these circuits (i.e., increased functional segregation) may underlie the clinical benefits produced by 3 weeks of amphetamine treatment.

Intrinsic Functional Connectivity in Attention-Deficit/Hyperactivity Disorder: A Science in Development

Functional magnetic resonance imaging (fMRI) without an explicit task, i.e., resting state fMRI, of individuals with Attention-Deficit/Hyperactivity Disorder (ADHD) is growing rapidly. Early studies were unaware of the vulnerability of this method to even minor degrees of head motion, a major concern in the field. Recent efforts are implementing various strategies to address this source of artifact along with a growing set of analytical tools. Availability of the ADHD-200 Consortium dataset, a large-scale multi-site repository, is facilitating increasingly sophisticated approaches. In parallel, investigators are beginning to explicitly test the replicability of published findings. In this narrative review, we sketch out broad, overarching hypotheses being entertained while noting methodological uncertainties. Current hypotheses implicate the interplay of default, cognitive control (frontoparietal) and attention (dorsal, ventral, salience) networks in ADHD; functional connectivities of reward-related and amygdala-related circuits are also supported as substrates for dimensional aspects of ADHD. Before these can be further specified and definitively tested, we assert the field must take on the challenge of mapping the "topography" of the analytical space, i.e., determining the sensitivities of results to variations in acquisition, analysis, demographic and phenotypic parameters. Doing so with openly available datasets will provide the needed foundation for delineating typical and atypical developmental trajectories of brain structure and function in neurodevelopmental disorders including ADHD when applied to large-scale multi-site prospective longitudinal studies such as the forthcoming Adolescent Brain Cognitive Development study.