Towards the identification of imaging biomarkers in attention-deficit/hyperactivity disorder

Abstract

Attention-deficit/hyperactivity disorder (ADHD) is one of common neurodevelopmental disorder that develops early in the childhood. In the recent years, a growing body of neuroimaging evidence proposed that widespread alteration or maturational delay of intrinsic functional network could play a central role in pathophysiology of ADHD. However, large heterogeneity of previous literature have confounded the discovery of cardinal features within intrinsic funcitonal organization and their contribution on symptom phenotype. In the current study, I investigated 1) discriminative features from multimodal imaging and candidate genetic data, and their contribution to ADHD symptomatology, and 2) potential attributes within intrinsic functional network relevant to pharmacological and behavioral interventions using resting-state functional MRI.

First, I identified mediating effect of GRIN2B and DRD4 polymorphism on local functional connecitivty and its dynamic repertoire which underpinned the gene-brain-behavior relationship hypothesis in ADHD pathophysiology. In addition, experimental results showed that combination of genetic and multimodal neuroimaging data could achieve excellent classification accuracy, and explain the considerable variance of ADHD symptoms. Key contributors of the best classification model were microstructural changes of white matter tracts that connects frontal cortex to far-reaching brain regions, widespread morphological changes in fronto-temporo-parietal areas, and the functional polymorphisms of dopamine receptor related genes. Model with functional connectivity features showed less predictive power for classification, but outperformed the best model in estimation of behavioral performance. Next, multivariate approach using random Forest revealed clinical improvement after methylphenidate medication could be explained by combination of inter- and intra-regional changes among resting-state functional networks, especially reduced local spontaneous acitvity of bilateral superior parietal cortices as well as segregation of large-scale functional networks. Furthermore, neural correlate of the equine-assisted activities and therapy was found in the right precuneus, which was associated with symptom improvement.

To summarize, I tried to sketch out implication of intrinsic functional architecture to ADHD from the integrative and neurobiological perspective. Regarding to experimental results, ADHD pathophyiology could be associated with multiple systemic impairment in default mode network, salient network, sensory/motor network, and part of task-positive network. Current findings would provide important clues for searching novel diagnostic and prognostic markers for ADHD.