Nonlinear Mechanism Research Of Brain Network Based On Resting-state Image Of Functional Magnetic Resonance Imaging

Human brain is a complex functional network,discussi ng about the nonlinear dynamical mechanism of the brain system has become a hot spot in the field of brain science in recent decades.However,most of these studies mainly focus on the nonlinearity of time series of local brain regions,while intuitive nonlinear studies of the entire brain system are still lacking.This paper aims to conduct a comprehensive nonlinear analysis on the brain network based on surrogate data method and graph theory,and observe its topological performance under that mechanism.The main research works of this paper are summarized as follows:(1)Three surrogate data methods: random shuffle,amplitude adjusted Fourier transform(AAFT),and static transformation autoregressive process(STAP)were used to construct linear surrogate dat a sets of the time series of90 brain regions of interest(ROI)in healthy young people and conduct nonlinear testing,the observed time series were derived from the resting-state image of functional magnetic resonance imaging(f MRI).After a single-sample bilateral t-test,the original time series of each brain region and their surrogates are found to have significant statistical differences(test statistic S>1.96),thereby verifying the local nonlinear dynamics of 90 ROIs.This method is of simple principle and provides a good tool for interpreting the dynamical character of local brain regions.(2)Based on the nonlinear analysis of local brain region in(1),combined with graph theory,we continued to conduct nonlinear testing on the functional brain network of healthy young people.The f MRI time series of each brain region was firstly random shuffled to construct linear surrogate data sets,and then the mutual information correlation coefficient between each pair of time series were calculated as the fu nctional connection to construct the original brain network and its surrogates.After a single-sample bilateral t-test,we find that there are significant small-world topological differences between original brain network and its surrogates(test statistic S> 1.96),thus proving the global nonlinearity of the functional brain network.This method can be applied to the nonlinear examination and analysis of the dynamical mechanisms of the brain in different states,and provides a new idea for the clinical dia gnosis of brain functional diseases.(3)Based on the results of local and global nonlinear testing of the brain above,graph theory was used to deeper analyze the functional brain networks of healthy young people.We find that there are a set of core node s with significant rich club efficiency in the brain network.These nodes are mainly distributed in the frontal,occipital,temporal,and insular of the brain,where information and energy are highly concentrated.This special node feature prompts the brain network to exhibit a significant small-world feature.Therefore,it is further explained that the brain is not a simple linear system that distributes tasks equally to each brain area in the process of operation,but a complex nonlinear system.In addition,this method can effectively detect the core nodes that have a key role in the coupled dynamical brain system,and is able to play a role in locating lesions in clinical auxiliary diagnosis.