Research On Complex Brain Effective Network Using Dynamic Causal Modeling

The human brain is a highly complex functional system.By constructing a network with neural units as nodes and neural connections as lines,the brain,which is regarded as sparse distribution of complex network systems formed by plenty of interaction neurons,can explore the working mechanism and the function realization process.However,this analysis lacks authenticity and validity because the simple brain network composed of nodes and connections does not reflect the actual exchange of directional information in the brain.Therefore,in this article,the Dynamic Causal Modeling(DCM)is used to construct the brain effective network,and the information exchange relationship of the relevant brain regions under the specific experimental stimulation is investigated by using the brain' s information automatic processing index?mismatch negativity(MMN)as the research object.Compared with functional magnetic resonance imaging(fMRI),the high temporal resolution of Electroencephalogram(EEG)can promptly explore the temporal changes in the activation of brain regions,and more conducive to exploring the potential dynamic changes in the brain.In addition,EEG has the advantage of noninvasive,portability and high temporal resolution,which makes it to be one of the most widely used methods in the study of brain cognitive function.In this article,the DCM algorithm based on EEG is deeply researched.The neuron cluster model is used as the foundation of neuron dynamics and the state equation is used to reflect the evolution of a neuron activity map to another neuron activity.The output equation reflects the relationship between the observed signal and mapping of neuronal activity.Then,by using Bayesian theory to estimate the parameters and model selection,the model with the highest fitting degree data is obtained,and the information transmission process in each brain region is comparatively analyzed.In this article,the effective brain network based on EEG data is constructed by using the MMN as the research object,the direct information exchange activity of the brain under non-attention auditory stimulation is revealed,and the intuitive and the effectiveness of the brain effective network constructed by DCM are analyzed with the related research of MMN.Firstly,the experiment is designed according to the auditory oddball experimental paradigm to extract MMN,the EEG data is preprocessed,the extracted MMN waveform is analyzed the spatial distribution of potential of changes by using EEG topography.Secondly,a head model and a source model are constructed to obtain a forward model,and then the EEG data is inversed by using the sparse solution under the framework of the Bayesian theory.The region of interest is selected according to the result of source location on the experimental stimuli.Finally,the study of EEG-based DCM algorithm and construct the brain effect network with DCM.Finally,using DCM to construct a brain effective network based on the EEG data obtained from experiments,the directional information interaction between the brain and the relevant brain regions stimulated by specific experiments is analyzed.We also find that DCM can directly and effectively reflect the effective connection between brain related brain regions.Overall,DCMs established by using EEG data,can reveal the brain connection relationship at the level of neurons.In practice,the construction of DCM network for the brain automatic processing index——MMN,can describe the effective connectivity of the brain under the specific stimulus,which is conducive to some brain disease diagnosis and adjuvant therapy.