Network Entropy Analysis Of Resting State Functional Connection In Schizophrenia

Schizophrenia(SC)is a kind of neurodegenerative disease,patients with cognitive,memory,emotional,motor perception and other aspects have varying degrees of impairment.The abnormal brain signal in SC patients may lead to the disorder of pathophysiology.At present,the diagnosis of SC still mainly depends on the behavior score of patients.Due to the lack of understanding of the etiology and the relatively single diagnosis method,it is of great significance to study a biomarker that is helpful for diagnosis and treatment.The development of neuroimaging provides a better means for the study of mental diseases.Functional magnetic resonance imaging(f MRI)can more accurately observe the structural and functional abnormalities of the brain.Functional connectivity has become an important tool to study the changes of brain function in neurodegenerative diseases.On this basis,the brain network theory also provides a new perspective for the study of mental illness.Graph theory based network topological attribute index has been widely studied in the field of brain network,but the research on dynamic characteristics of brain network is relatively less.Complexity is used to measure the dynamic characteristics of the system.As a typical complexity index,entropy method has the advantages of small amount of analysis data,strong anti-interference ability and simple algorithm,which is widely used in many fields.At present,the research of entropy method in brain network is mainly on the exploration and innovation of algorithm,so it is of great significance to apply entropy method to brain network to study SC.In this study,functional entropy analysis was performed based on functional brain network of resting state f MRI data of SC patients(1)It is found that the functional entropy is significantly correlated with the functional connectivity strength and network topology attributesBased on two public datasets of UCLA-LA5 c and SLIM and the power brain template,the correlation between functional entropy and network topology attributes and functional connectivity strength is analyzed.The results show that functional entropy and functional connectivity strength show significant positive correlation at both global and module levels,while functional entropy and functional connectivity strength show significant positive correlation with clustering coefficient,shortest path length,local efficiency and modularity,and significant negative correlation with global efficiency and participation coefficient.In addition,the node function entropy and node efficiency showed a significant positive correlation in all brain regions.Through the relationship among functional entropy,functional connectivity strength and network topology attribute,it is clear that functional entropy,as an index of complexity,can effectively describe functional brain network.(2)The abnormal pattern of functional entropy in the functional brain network of SC patients was foundIn this study,functional entropy was used to analyze the functional brain network of SC patients and controls.The results showed that the global functional entropy of SC patients was significantly lower than that of the control group.Significant differences were found in the node functional entropy of paracentral lobule,posterior central gyrus,occipital lobe,cerebellum,frontal lobe,paracentral gyrus and thalamus.Significant differences were found in the module functional entropy of default mode network and visual network.In the correlation analysis with clinical scales,the functional entropy of suboccipital gyrus was significantly negatively correlated with brief psychiatric scale and positive symptom rating scale,the functional entropy of paracentral lobule was negatively correlated with positive symptom rating scale score,and the modular functional entropy of visual network was significantly negatively correlated with brief psychiatric scale and positive symptom rating scale respectively.(3)Gauss functional entropy was proposed to analyze the abnormal patterns of functional brain network in SC patientsGaussian functional entropy was proposed and the functional brain networks of patients with SC and control group were analyzed.The results showed that the global Gaussian functional entropy,the node Gaussian functional entropy in paracentral lobule,frontal lobe of posterior central gyrus,temporal lobe,central sulcus tectum,superior marginal gyrus,precuneus lobe,fusiform gyrus,angular gyrus,lingual gyrus,occipital lobe,cuneiform lobe,talus fissure,cerebellum,cingulate gyrus and thalamus,as well as the sensory and motor network,the global Gaussian functional entropy in SC patients and control group were significantly higher than those in control group The module Gaussian function entropy of auditory network is significantly different from that of visual network.The node Gaussian functional entropy of the paracentral lobule,middle temporal gyrus,medial and paracingulate gyrus and the module Gaussian functional entropy of the visual network were significantly correlated with the clinical scale.(4)The classification model based on functional entropy and Gaussian functional entropy is establishedIn this study,based on Gaussian functional entropy and functional entropy,four classifiers are used to construct classification models for SC.the results show that Gaussian functional entropy has higher classification accuracy than functional entropy,and has better discrimination for SC,and the classification effect of support vector machines(SVM)classifier is the best.In conclusion,as a new measure of functional brain network complexity,Gaussian functional entropy is helpful to the diagnosis of SC in clinic.To sum up,this study studies the relationship between functional entropy and functional connectivity strength and network topological attributes to clarify the effectiveness of functional entropy in describing functional brain networks.And the functional entropy analysis of the functional brain network of SC patients is carried out,and the method is improved.Gaussian functional entropy is proposed as a new index of brain network complexity,which not only enriches and develops the analysis method of brain imaging,but also has great significance for the diagnosis and treatment of neurodegenerative diseases.