Research Methods And Comparisons Of Two Causal Network

The human brain is known as one of the most complex systems which consists of about 1011 neurons, in which 1015 neural synaptic connections are included. Researchers believe that mental illness may have an impact on the structure and function of the brain, so it is particularly important to bave a study of brain networks.In recent years, with the deepening of the brain network research, many brain imaging technologies have been developed, such as functional magnetic resonance imaging (fMRI). With the rapid development of the brain imaging technology, as well as the progress of experimental technology, people can get high precise medical image data, which makes the complex brain research possible. In this paper, the analysis of the network is based on functional magnetic resonance imaging data.Effective connectivity is an important content in the study of brain networks. Granger causality analysis is one of the effective connectivity and widely applied in neuroscience, economics, signal processing, engineering and other fields. In recent years, it has been gradually applied in effective connectivity and data analysis of brain functional magnetic resonance imaging.Pair-wise Granger causality analysis cannot distinguish the effects among network nodes of the brain are direct or indirect. Therefore, the connection between the multiple nodes may appear "false connection". However conditional Granger causality analysis and Silencing method can compensate for this defect. The two methods can determine the connection between two brain regions is direct or indirect through the third brain region. Thus, the indirect connection will be removed, and the direct connection will be keeped. All of these making the network connection more clear.In this paper, there are two brain network connectivity methods were utilized:conditional Granger causality analysis and Silencing method. The two methods were first applied to build networks on the toy model, then the two methods were applied to the analysis of the task state fMRI data, in which brain effective connectivity of 43 schizophrenia patients and 33 normal controls were analysed. During the analyse, a significant difference between the normal group and the patient group was obvious. What’s more, we found that attention network,and frontal lobe are the mainly concentrated areas of ROIs and differential effective connectivity, respectively. In addition, for the network built by conditional Granger method, we obtained the overall network connection diagram of the normal group and the patient group, the results show that the frontal lobe becomes the network connection center. For Silencing method, we also obtained the frequency distribution histogram of Silencing value of the normal group and the patient group, and found that many brain regions were damaged, such as middle frontal gyrus.By comparing the established networks of two methods, we found that the results of conditional Granger causality analysis is better than that of Silencing method. So conditional Granger causality analysis can be regarded as a good way to study the brain network connection.