Research On Biological Markers Of Alzheimer's Disease Based On Brain Functional Network

With the increasing aging of the population and social pressure,the number of patients with neurodegenerative diseases such as Alzheimer,Parkinson and Brain Injury has increased sharply.As a severe mental disability,Alzheimer's disease is one of the few diseases that cannot be prevented,cured or even slowed down.Modeling human brain network based on fMRI image and using graph theory analysis methods to study the changes of topological properties in human brain network caused by diseases are hot topics in current research.Based on the background and significance of the research,the thesis completed the construction of voxel-level brain function network and the extraction of biological markers of Alzheimer's disease.Firstly,the brain-level functional network constructed by the traditional method will mask the difference of details between the voxels in the brain region,so we construct a voxel-level brain function network.In this thesis,based on the BOLD signal value of the preprocessed fMRI data,the voxel point is selected as the node of the network,and the correlation coefficient between the time series of the node is used as the edge of the network.Then,the frequent subgraph mining algorithm is used to extract the sub-patterns that frequently appear in the network constructed based on the two types of training data(AD and NC),and the biomarkers of Alzheimer's disease were extracted after further deletion of the mined subpatterns based on the graph similarity calculation method.Finally,using the classification algorithm,the extracted markers were verified on the test data set.After analyzing the topological structure and the extracted biological markers of the brain function network constructed by different sizes,it was found that for the markers of NC subjects,there is a high correlation between the posterior cingulate gyrus,hippocampus,lateral temporal cortex,medial prefrontal cortex and other brain regions.These brain regions and voxels are also the main components of markers.For the markers of AD subjects,the connectivity between these brain regions and the voxels in the brain region was significantly reduced.The results of the thesis study are basically consistent with the actual clinical findings and the existing research results.However,voxel-level brain function networks have more granular findings for changes in connections within the brain.This also shows that the use of voxel-level brain function network has a certain significance for the extraction of biological markers of Alzheimer's disease.