多模态脑网络偏侧性在注意缺陷多动症研究中的应用

Attention deficit hyperactivity disorder (ADHD) is one of the most common psychological diseases in children, and has serious impacts on patients. Its clinical symptoms will persist into adulthood and induce other diseases, so the study of ADHD is of great significance. Many neurophysiological and anatomical studies have shown that abnormal of hemispheric specialization is one of the pathogenesis of ADHD. However, the study on specialization of ADHD brain network is still immature, and the effect of age on specialization is not concerned. Therefore, based on the whole-brain network and intrinsic functional network, this project aims to explore the specialization of ADHD patients in different age groups. Firstly, this project will study the definition methods of multimodal connections and network threshold in order to improve the construction methods of whole-brain networks. Then, module partitioning algorithms will be used to define repeatable intrinsic functional networks. With the help of graph theory and information theory, the hemispheric asymmetry of network structure and differences between intra- and inter-hemispheric connections will be analyzed from the perspective of whole-brain and intrinsic functional network in ADHD. By using the feature selection algorithm and clinical features, stable and sensitive network specialization indicators will be selected. The purpose of this study is to find objective imaging indicators and provide a new theoretical basis and auxiliary diagnosis model for the clinical diagnosis of ADHD.

注意缺陷多动症(ADHD)是一种儿童最常见的精神疾病，其临床症状会持续到成年且诱发其它疾病，因此对ADHD研究意义重大。许多神经生理学和解剖学研究表明大脑偏侧性异常是ADHD的发病机制之一，但目前ADHD脑网络偏侧性研究尚不成熟，且未关注年龄对偏侧性的影响。因此本项目针对不同年龄段的ADHD患者，拟利用全脑网络和固有功能网络，探讨由于网络连接异常引起的偏侧性。首先基于精细脑图谱，研究多模态网络连接定义与网络阈值定义方法，构建合理的多模态全脑网络；然后利用基于模块划分算法，构建可重复的固有功能网络；并借助图论和信息论方法对ADHD患者，从全脑和局部网络角度探讨左右半球拓扑结构的不对称性，以及半球内和半球间的连接差异性；最后借助特征选择算法并结合临床特征，筛选稳定且对ADHD敏感的网络偏侧性指标，挖掘不同年龄段ADHD患者的客观影像学指标，为ADHD的临床诊断提供新的理论依据和辅助诊断模型。

偏侧性是指大脑左右半球在结构和功能上存在的不对称性，其源于胚胎时期且随着年龄、环境和经验的积累而发生变化,是人脑发育的一个基本特征。基于神经影像技术和复杂网络理论的大脑网络模型，有助于从多角度、多层次评估由于脑区间的交互引起的大脑偏侧性。本项目首先基于多模态神经影像技术，研究了不同模态下大脑网络的构建方法，并重点分析了精神疾病在全脑网络拓扑结构存在的异常模式。然后，项目进一步探究了大脑网络的偏侧性机制，基于构建的大脑半球网络模型，项目从模块化、rich club、可控性等不同网络拓扑角度探究左右半球网络的不对称性，同时研究了半球内与半球间的连接不对称性。基于上述提到的拓扑特性，项目从多角度、多层次分析了精神疾病与性别等因素对大脑偏侧程度的影响。最后，针对于ADHD疾病，项目从多模态全脑网络、大脑网络偏侧性等多角度挖掘了ADHD疾病的客观影像学指标，为ADHD的临床诊断提供新的理论依据。综上所述，本项目的研究方向是计算机科学、脑科学、神经科学和临床医学等多学科的交叉领域，重点围绕着大脑偏侧性与脑疾病的诊断等重要脑科学问题开展的一系列研究。