A Research Of Brain Community Reconfiguration In ADHD Patients Based On A Time-varying Multilayer Network

Attention-deficit/hyperactivity disorder(ADHD)is a common neurodevelopmental disease.It often occurs in childhood and most patients’ symptoms will persist into adulthood.Adult patients with ADHD have difficulty concentrating during study and work,and have problems in controlling emotions and actions,which brings a heavy burden to the patients themselves,their families and society.Therefore,researchers from various fields have been working hard to explore the pathogenesis of ADHD and auxiliary diagnosis methods.Functional magnetic resonance imaging(f MRI)technology has promoted the vigorous development of brain science research.The brain network construction method based on graph theory provides convenience for exploring the abnormal brain structure and function of patients with mental diseases.Current research has proved that the brain network topology of ADHD patients is abnormal.However,these studies often treat the brain as a static monolithic structure,ignoring the dynamic characteristics of the brain.In recent years,research on mental diseases using dynamic network has become a hot spot.While conducting dynamic network analysis,more and more researchers are beginning to think that different network layers do not exist independently,but are related to each other.Specifically in time networks,the association between network layers is time-dependent,which results in a time-varying multi-layer network.Community is a widespread structure in complex networks,and it is no exception in brain networks.While the brain network is changing,the members of the community and the connections between members will be continuously reconfigured to achieve the purpose of maintaining normal physiological activities.Current studies have confirmed that the abnormal brain community structure of ADHD patients is related to cognitive impairment,but the dynamic changes in the brain community of patients are still unclear.Based on this current situation,the purpose of this research is to use computer technology to build a new network model to mine the dynamic change patterns of the brain community structure of ADHD patients.The main research contents and results are as follows:(1)Construct a time-varying multilayer network model,conduct community detection and evaluate the recognition resultsAiming at the time dependence between network layers in the dynamic analysis process,this study constructed a time-varying multilayer network based on the slice processing of the f MRI signal,and took the time factor into account in the model through inter-layer coupling.In order to solve the problem of community matching between different network layers,a multilayer community detection algorithm is used to identify the community of the constructed network.When evaluating the scale and quality of the results of community division,it is found that more community structures are formed in the brains of ADHD patients,and there are fewer members in the community,and the stability of the community is also lower.This part of the results proves that the brain community of ADHD patients is more prone to change.(2)Exploring the abnormal brain community reconfiguration patterns of ADHD patients from the frequency perspective.Aiming at the problem that brain connections in ADHD patients are more likely to change,this research studied the abnormal community changes of patients from the perspective of reconstruction frequency.Calculated flexibility,cohesion,disjointedness,promiscuity and temporal core-peripheral structure.The results of the experiment found that patients’ flexibility and cohesion were significantly increased,indicating that their brain network organization is more unstable,and tends to switch belonging communities in groups.In addition,this study also found that part of the brain area of ADHD patients has changed in the role of maintaining normal physiological activities in the resting state.(3)Exploring the abnormal brain community reconfiguration patterns of ADHD patients from the perspective of probability.Targeted on the problem of regional alliances in different empirical functional networks in the process of community reconfiguration,this research studied the abnormal community changes of patients from the perspective of probability.By calculating the module allegiance matrix,recruitment coefficient and integration coefficient,it is found that the brain regions of ADHD patients do not have special alliance preferences,but they tend to be in the same community as the nodes in their own experience function network;the recruitment coefficient and integration coefficient of patients are significantly lower than those controls,it shows that the communication ability within or between some functional networks of ADHD patients is impaired,and the subcortical network of the patients is disintegrated,and the thalamus is in an isolated state.In conclusion,this research is the first attempt to use a multilayer network model and a multilayer community detection algorithm to explore the community changes of ADHD.The experiment also proposes community reconfiguration mode mining methods from two different perspectives.The experimental results support the dynamic network research perspective of ADHD,provide a new basis for explaining the physiological defects caused by such neurological diseases,and also provide a new opportunity for the auxiliary diagnosis of ADHD.