| The human brain is a delicate and harmonious superior system which can process information in the currently known universe, and it can receive, transfer, process, syncretize information to complete the high-level function of brain by the joint coordination of multiple brain regions. The two most fundamental organizational principles of the human brain functions are functional separation and functional integration are. Functional separation is the basis of the hypothesis of functional orientation, however, functional integration is that multiple brain areas coordinate to complete the human behavioral function, and it is the theoreyical basis of the brain connectivity. It is usually in terms of functional connectivity and effective connectivity to investigate human brain. Effective connectivity of the human brain regions had been a focused research which measures whether there was connectivity correlation between brain regions as well as the direction and strength of the connectivity. Many methods were used to investigate the effective connectivity, such as structural equation model (SEM), dynamic causal modeling (DCM), granger causality analysis (GCA). GCA model is simple and with no need for model as a priori assumptions, it is widely used to explore the effective connectivity between different brain regions.Effective connectivity of studies usually rely on functional magnetic resonance imaging (fMRI) technology. fMRI has become an important technical means for brain connectivity research because of its non-invasive, high spatial and temporal resolution and good repeatedly. According to the collection status, fMRI is divided into Task-state fMRI and Resting-state fMRI. Due to the seriously patients unable to complete the experimental task, Task-state fMRI in the study of a number of diseases is limited, However, the subjects in the Resting-state fMRI scanning process does not need to do any task, just keep quiet, Resting-state fMRI is widely used in bipolar disorder, schizophrenia, depression and other neurological diseases.Granger causality analysis is one of the important approaches to study the effective connectivity in the Resting-state fMRI. How to select the modeling order in granger causality analysis of fMRI is one of the key problems need to solve. In the use of granger causality analysis in the Resting-state fMRI, researchers usually use the order one, however, the first order model can correctly reflect the causality between brain areas has yet to be further research. Main research content of this paper is the granger causality analysis model order selection method. In this paper, the main research content is the model order selection methods in granger causality analysis. Real effective connectivity form between the human brain regions is not yet clear, so we use simulation data to study the impact of the model order in granger causality analysis results and scientific granger causality model order selection method is given. On this basis, granger causality analysis was applied to resting state fMRI data of bipolar disorder in order to investigate effective connectivity.Bipolar disorder is one of the most debilitating illnesses, it is reported that the amygdala and prefrontal has a key role in the pathogenesis of bipolar disorder. In this paper,15patients with bipolar disorder and15healthy control subjects participated to complete the functional magnetic resonance imaging scans in the resting state, granger causality analysis was applied to explore the effective connectivity between amygdala and intermedial prefrontal cortex (mPFC) of the two groups. The results indicated that compared with healthy controls, there is no connectivity between the right amygdala and the intermedial prefrontal cortex, while there are still effective connectivity between the left amygdala and the intermedial prefrontal cortex, but its connectivity form is abnormal in bipolar disorder. This study maybe of great significance to the diagnosis and treatment of bipolar disorder in the future. |
PDF Full Text Request