Research On Dynamic Evolution Modeling Of Brain Network And Steady State Identification Method

Human brain is a changing system essentially.Resting state f MRI brain network reconstruction technology provides powerful support for the research of human brain.At present,most studies have focused on the study of the static characteristics of the human brain,and the study of dynamic characteristics is still in its infancy.Since there are various factors in the data acquisition process of the resting state functional magnetic resonance imaging,the signals collected will not directly enter the stable stage.At present,most of the studies in the resting state f MRI data based on the experience of the unified value always removed before a few data,but rarely consider the experimental conditions,the specific subjects influence factors of state and individual differences,this approach may lead to unreliable results.This thesis,starting from the overall system and brain area network spatiotemporal properties,with dynamic characteristics of brain area network as the basis,this thesis presents a quantitative judgment of resting state functional magnetic resonance imaging(f MRI)evaluation method in BOLD data when the signal into a stable state of the.This method firstly constructs the whole brain area sampling state observation matrix at a single time point,and then through unsupervised clustering to obtain the set of States,on the basis of the dynamic evolution of brain network automaton theory established time model,which can be based on the BOLD signal when the judge entered into a stable state model,provides a quantitative analysis method for judging the critical point of steady state BOLD signals.In this thesis,the experimental verification and analysis are carried out using the resting state f MRI data of different subjects.The experiment is divided into two parts,to verify the experimental results of dynamic time automaton brain network evolution model based on the time model can describe the human network state transition rule and the evolution processof different test data is universal,and can identify the abnormal evolution of subjects.BOLD signal stable survival analysis results show the critical point state based on the steady state evaluation method based on timed automaton model can be calculated quantitatively at different time points into the steady state probability,and by comparing to find the stable critical point,above work as research on the dynamic characteristics of deep brain network provides the basis.