Static And Dynamic ICA Study Based On Schizophrenia FMRI Data

Schizophrenia is a serious and heterogeneous mental disease with extensive structural and functional abnormalities of the brain.Finding the key neurophysiological defects in schizophrenia has been a breakthrough in the cure of schizophrenia.Functional magnetic resonance imaging(fMRI),which perfectly combines brain activity with high-resolution MRI technology,has become an effective method to study brain dysfunction in patients with various mental diseases.Previous studies have all implicitly assumed that functional connectivity is constant during resting-state fMRI scans.However,a growing body of research suggests that even in a resting state,functional connections are not static.The utility of potential biomarkers for dynamic functional connectivity has been proposed in Parkinson's disease,schizophrenia and epileptic.The aim of this thesis is to use these methods to find abnormal static and dynamic functional connections between brain functional networks in patients with schizophrenia.This thesis selected the object of study for the 21 neuroleptic-naive schizophrenia(NNS),60 neuroleptic-treated schizophrenia(NTS)and58 healthy controls(HC).Firstly,we extracted the resting state functional network of the brain based on independent component analysis(ICA).Then,the time course of each component was divided into consecutive windows through the sliding time window approach,and the K-means clustering method was applied on window-correlated matrices to identify different states.Static and dynamic functional connectional differences are compared among groups.Finally,the relationship between the static functional connectivity,stability of dynamic functional networks and clinical symptoms was also studied.The 51 effective independent components obtained by ICA were grouped into seven networks.When compared in pairs,there were significant differences in the static functional connectivity between the other networks and the DMN of NTS and HC,NNS and HC.Two typical functional connectivity states of NTS,NNS and HC are identified by K-means clustering(state I is the connectivity within the network while state ? is the connectivity between the networks).It has been discovered that there are significant differences between the two states of NTS and NNS,HC,and it shows that the fractional windows of two states of NTS was significantly different from NNS and HC.Eventually,a positive correlation was uncovered between the number of state transitions and the daily medication dosage in the NTS group.We found three interesting results,which are as follows.Firstly,abnormal connectivity in the DMN is a key neurophysiological defect of schizophrenia.Secondly,taking antipsychotics may lead to weakenedconnectivity within brain networks and increased connectivity between networks.Finally,the brain's dynamic network becomes more instability after taking antipsychotics,and this instability is positively correlated with the daily medication dosage.These views lead us to believe that our findings can provide clues to new drug formulations.