Brain Network Mechanism And Diagnostic Model Of Attention Deficit Hyperactivity Disorder Based On Magnetic Resonance Imaging

Attention Deficit Hyperactivity Disorder(ADHD),is a mental disorder that occurs frequently in adolescents.The main clinical symptoms are: difficulty in concentration,excessive activity,and poor impulse control.Many people with ADHDstill have these symptoms in adulthood,which will negatively affect their academic,physical and mental health,family life and social ability in adulthood.However,the etiology and pathogenesis of ADHD are still unknown.At present,the clinical diagnosis of ADHD mainly relies on the fifth edition of the "Diagnostic and Statistical Manual of Mental Disorders" in the United States.This scale diagnosis will bring certain problems,that is,it is not objective enough.This will affect the accurate diagnosis and treatment of ADHD.As a result,more and more researchers are investigating the brain mechanism and automatic diagnosis of ADHD.Similarly,in order to explore the diagnostic factors of objectivity,this study will carry out two aspects of research: explore the brain network mechanism and diagnostic model of ADHD based on magnetic resonance imaging.The main research work is as follows:The first part is the study of the brain network mechanism of ADHD based on magnetic resonance imaging.The purpose of the study was to explore imaging markers for ADHD.Voxelbased morphometry(VBM)and phase synchronization analysis methods are mainly used to extract image markers of ADHD.The results of the study found that there are differences in brain regions between ADHD patients and normal controls,which has promoted further development of ADHD brain mechanism research.The second part is the research of ADHD diagnosis model based on dynamic functional connection.The purpose of the study was to explore an automatic diagnostic model for ADHD.Because researchers have discovered that revealing the complex and changing characteristics of the brain network and its mechanism is not enough to consider only static functional connections,because static functional connections cannot explain the time-varying and dynamic information interactions of the brain,so dynamic functional connections came into being.Many clinical studies have shown that dynamic functional connection analysis can provide better evidence for pathological investigation and auxiliary diagnosis of clinical diseases.So this research uses a dynamic functional connection method,combined with support vector machines and convolutional neural network classifiers to carry out research.The general process is to generate dynamic functional connection data through the method of dynamic functional connection estimation through resting functional magnetic resonance data,and then perform feature extraction on the obtained data,use the obtained features to perform clustering,and finally,classify each category..The classification method used is a more suitable classification method for machine learning in the neural field: support vector machine;however,the traditional artificial feature extraction method is more simple in features,because the artificially extracted features can be interpreted,which are deep,abstract,Unexplainable features cannot be obtained through manual intervention,so in order to effectively extract the deep spatial information of ADHD data,a convolutional neural network algorithm is introduced in this study.This part of the study can effectively achieve the classification of ADHD patients and normal subjects,and the classification accuracy rate is more than 90%.This research mainly focuses on two aspects of ADHD brain mechanism research and diagnostic model.In the research of ADHD brain mechanism,the voxel-based morphological analysis method and phase synchronization analysis method can effectively find the image markers of ADHD.Provide new ideas and directions for the next stage to study the etiology of ADHD in specific brain regions;in the study of automatic diagnostic models,the ADHD diagnosis model of dynamic functional connection and convolutional neural network can prove that the dynamic functional connection is in the classification study of ADHD Effectiveness.