Multimodal Neuroimaging-based Prediction of Adult Outcomes in Childhood-onset ADHD using Ensemble Learning Techniques

Abstract

Attention deficit/hyperactivity disorder (ADHD) is a highly prevalent and heterogeneous neurodevelopmental disorder, currently relaying on subjective symptom observations for diagnosis.Machine learning classifiers have been utilized to assist the development of neuroimaging-based biomarkers for objective diagnosis of ADHD. However, the existing basic model-based studies in ADHD reported suboptimal classification performances and inconclusive results, mainly due to the limited flexibility for each type of basic classifiers to appropriately handle multi-dimensional source features with various properties. In this study, we proposed to apply ensemble learning techniques (ELTs) in multimodal neuroimaging data collected from 72 young adults, including 36 probands (18 remitters and 18 persisters of childhood ADHD) and 36 group-matched controls. All the currently available optimization strategies for ELTs (i.e., voting, bagging, boosting and stacking techniques) were tested in a pool of semi-final classification results generated by seven basic classifiers. The highdimensional neuroimaging features for classification included regional cortical gray matter (GM) thickness and surface area, GM volume of subcortical structures, volume and fractional anisotropy of major white matter fiber tracts, pair-wise regional connectivity and global/nodal topological properties of the functional brain network for cue-evoked attention process. As a result, the bagging-based ELT with the base model of support vector machine achieved the best results, with the most significant improvement of the area under the receiver of operating characteristic curve (0.89 for ADHD vs. controls, and 0.9 for ADHD persisters vs. remitters). We found that features of nodal efficiency in right inferior frontal gyrus, right middle frontal (MFG)-inferior parietal (IPL) functional connectivity, and right amygdala volume significantly contributed to accurate discrimination between ADHD probands and controls; higher nodal efficiency of right MFG greatly contributed to inattentive and hyperactive/impulsive symptom remission, while higher right MFG-IPL functional connectivity strongly linked to symptom persistence in adults with childhood ADHD. Our study also suggested that considering their solidly improved robustness than the commonly implemented basic classifiers, ELTs may have the potential to identify more reliable neurobiological markers for severe brain disorders.structures estimated from structural MRI data, volume and FA of major WM fiber tracts derived from DTI data, the pair-wise regional connectivity and global/nodal topological properties (i.e., global-, local-, and nodal-efficiency, etc.) of the cue-evoked attention processing network computed from task-based fMRI data. Based on findings from existing studies from our group and other teams, we hypothesized that structural and functional alterations in frontal, parietal, and subcortical areas and their interactions would significantly contribute to accurate discrimination of ADHD probands (adults diag...

Application and research progress of machine learning in the diagnosis and treatment of neurodevelopmental disorders in children

“…Based on boosting and bagging, a stacking technique using different models for integration has emerged ( 48 ), however, literature related to NDDs is few; therefore, the application value warrants further investigations.…”