Identification of internet gaming disorder individuals based on ventral tegmental area resting-state functional connectivity

Xinwen Wen,Yawen Sun,Yuzheng Hu,Dahua Yu,Yan Zhou,Kai Yuan

DOI: https://doi.org/10.1007/s11682-020-00391-7

2020-10-10

Brain Imaging and Behavior

Abstract:Objective neuroimaging markers are imminently in need for more accurate clinical diagnosis of Internet gaming disorder (IGD). Recent neuroimaging evidence suggested that IGD is associated with abnormalities in the mesolimbic dopamine (DA) system. As the key nodes of the DA pathways, ventral tegmental area (VTA) and substantia nigra (SN) and their connected brain regions may serve as potential markers to identify IGD. Therefore, we aimed to develop optimal classifiers to identify IGD individuals by using VTA and bilateral SN resting-state functional connectivity (RSFC) patterns. A dataset including 146 adolescents (66 IGDs and 80 healthy controls (HCs)) was used to build classification models and another independent dataset including 28 subjects (14 IGDs and 14 HCs) was employed to validate the generalization ability of the models. Multi-voxel pattern analysis (MVPA) with linear support vector machine (SVM) was used to select the features. Our results demonstrated that the VTA RSFC circuits successfully identified IGD individuals (mean accuracy: 86.1%, mean sensitivity: 84.5%, mean specificity: 86.6%, the mean area under the receiver operating characteristic curve: 0.91). Furthermore, the independent generalization ability of the VTA RSFC classifier model was also satisfied (accuracy = 78.5%, sensitivity = 71.4%, specificity = 85.8%). The VTA connectivity circuits that were selected as distinguishing features were mainly included bilateral thalamus, right hippocampus, right pallidum, right temporal pole superior gyrus and bilateral temporal superior gyrus. These findings demonstrated that the potential of the resting-state neuroimaging features of VTA RSFC as objective biomarkers for the IGD clinical diagnosis in the future.

neuroimaging