Analysis Of Dynamic Functional Connection Of Human Brain Network

The structure of the human brain is far more complex than other parts of the human body.Research on the relationship between brain network ch aracteristics and brain diseases has always been a research hotspot.With the deepening of research on brain function networks,people have found that the functional connection state of the human brain under the time domain sequence will show unique charac teristics.People call this characteristic dynami c functional connection.The dynamic functional connection characteristics of the human brain are in Specific image characteristics are reflected in f MRI images.Therefore,extracting dynamic functional connections based on f MRI images and performing clust ering and classification identification is an effectiv e way to realize the diagnosis of typical brain diseases.This thesis is aimed at patients with depression and epilepsy to analyze the dynamic functional connection of brain networks under abnormal physiological conditions.The specific research content is as follows:1)Analysis of brain dynamic intera ction abnormalities in patients with depression and epilepsy based on K-means clustering.First,preprocess the acquired f MRI data from the spatial and temporal dimensions according to the principles and characteristics of magnetic resonance imaging;then use the sliding window to extract the BOLD signal from the region of interest,and build a dynamic functional connection matrix based on the Pearson corre lation coefficient.K-means performs cluster analysis on the dynamic functional connection status of n ormal and abnormal brain networks.The results showed: normal subjects divided into two categories brain network status,depression and epilepsy patients with brain disease state network is divided into five categories;while the domain status changes depr ession,epilepsy patients and healthy subjects connected with significant network functions difference.Depression and abnormal changes in the dynamic int eractive network with epilepsy showed that there funct ion caused by the disease of depression and epilepsy brain connections change.2)Identification of patients with depression and epilepsy based on LLE dimensionality reduction and SVM classification.Fi rstly,the LLE correlation analysis method is used to reduce the dimensionality of the extracted dynam ic function connection matrix to complete the dimensionality reduction of the dynamic function connection matrix in the entire data space;Zhen using poly nomial as the kernel function SVM classifier to classi fy the dynamic functional connection features after dimensionality reduction,to solve the non-linear classification problem of abnormal physiological state and normal human brain.The experimental results show that the method can effectively use the featu re matrix to describe the dynamic characteristics of different regions of interest,and provide support for the research on the dynamic characteristics of the human brain network and brain diseases.