Analysis Of Connectivity Features In Brain Functional Networks Based On Data And Model

The research of brain function network is a hot research topic in recent years. Most researches are based on the analysis of the separation of data and model. The diagnosis of medical data is more helpful, when the disease characteristics are obvious. But the diagnosis effect is not obvious, when the disease characteristics are not obvious. The research of the model method is biased to the complexity of the method, and it is divorced from analysis of the characteristics of the medical data. This paper adopts the method that combines the data model to analyze the connectivity of brain functional network. This method not only pays attention to the characteristics of medical data, but also explores the inherent law of the disease in the model.The research carrier of this paper is mainly based on the resting state fMRI data of normal people and patients with brain diseases. The main content of this paper is that activity degree model, complex network model,search algorithm model and correlation analysis model are applied to analyze the degree, clustering coefficient, network diameter, small world characteristics and connectivity features of network node. The main contents are the combination of data model, and they are divided into three parts:First, analysis of brain functional network characteristics based on node activity, a model for judging brain active degree is proposed based on node activity, which is used to study the characteristics of nodes and network structures of brain function networks. Firstly, functional Magnetic Resonance Imaging (fMR1) data were employed to construct the brain functional network of patients with brain disease, and node degree, clustering coefficient and average distance were calculated. Then these indexes were compared with those of normal subjects and the differences were compared. Secondly, the model of activity degree is proposed on the basis of the above network theory,and we apply it to the connectivity analysis of the node and network of the brain functional network. Finally, the distributions of active degree in various brain regions as well as their connection states in brain structure were analyzed, through the activity of brain functional network nodes. The Low Frequency Amplitude (ALFF) values of normal subjects and stroke patients were measured respectively in the experiment, and the active degree of corresponding nodes was compared and judged. Experimental results further verify the feasibility of using node activity to analyze the characteristics of brain functional networks.Second, functional connectivity analysis of brain default mode networks using Hamiltonian path model, we study the functional connectivity of the brain default mode network by using the Hamiltonian path model. Firstly, the brain DMNs in resting state are constructed with the employment of fMRI data. Then, the Dijkstra algorithm is used to calculate the shortest path length of the node which represents each brain region, and the Hamiltonian path of the default network is solved through the improved adaptive ant colony algorithm. Secondly, the improved adaptive ant colony algorithm is used to solve the Hamiltonian path distance of the default mode network. Finally,complex network analysis methods are introduced to discuss the node and network properties of brain functional connectivity in both normal subjects and stroke patients. The experimental result demonstrated that there are some significant differences in the properties of the DMNs between stroke patients and normal subjects, especially the length of Hamiltonian path. It also verifies the effectiveness on studying the functional connectivity of the brain DMNs by applying the proposed method of Hamiltonian path.Third, analysis of individual differences in brain functional network based on correlation coefficient model, correlation analysis model is applied to study the similarities and differences between brain functional networks.Firstly, adjacency matrixes of brain function networks are constructed using fMRI data in resting state. Then, the adjacency matrix is transformed into two value matrix, and the node degree, the average distance of the network and the clustering coefficient are calculated. Finally, correlation analysis models are introduced to discuss the node degree distribution, network diameter and small world property of brain functional networks in autism children. The experimental result demonstrated that individual network had a negative correlation with the average network in the brain function network of autisms,and the correlation between them was very weak. But the brain functional networks between individuals and individuals were positively correlated, and the correlation of them was stronger. The brain function network of autism has the same network diameter under the threshold of our experiment.In this paper, we combine the fMRI data with the activity model, the ant colony algorithm model and the correlation model to analyze the connectivity of the brain function network. It was found that the connectivity between fMRI data of patients with cerebral apoplexy and autism was different. At the same time, the characteristics of the internal connection of the disease are similar, which provides a scientific basis for clinical diagnosis.