The Pattern Research Of Brain Functional Networks And Its Application With FMRI

Brain is known as the most complicated system of information processing in nature. It can accomplish a series of complex functions by continuously organizing and re-organizing its functional connectivity. In this dissertation, the intrinsic organizational mechanisms of the brain function are investigated in the point of view of function integration, and the pathophysiological alterations of brain functional networks are explored using functional magnetic resonance imaging (fMRI) datasets. Specific data processing approaches are explored to analyze the brain functional networks in depth, and valuable results are provided for references in the study of schizophrenic pathology, and also as criterions of schizophrenic discrimination. The main contents and contributions of the dissertation are as follows:The brain functional networks for gum-chewing are studied. The information processing and transmission streams in the brain functional network during chewing are investigated using activation area detection approaches and complex network theory. The functional network for gum-chewing, which is constructed by extracting those chewing-related activated voxels, is proved to be a scale-free small-world network. Then, by computing the centrality indices and community structure of this network, it is found that:1) The neocortical hubs of the network are distributed in the sensory and motor cortices,2) The kernel nodes with largest betweenness and occupying the shortest pathway for information transmission, are located in the thalamus and lentiform nucleus; 3) The functional communities that incorporated in the chewing task, involve not only those brain regions related to motor and rhythm, i.e., the central gyrus, the lentiform nucleus and the cerebellum, but also the cognition-related regions such as the insula and the frontal cortex. These results suggest that a cortico-thalamic-cerebellar neural circuit could be responsible for the information processing and transmitting during chewing.The characteristics of the brain functional networks during resting are analyzed in this dissertation. The resting brain functional network is constructed based on the complex network theory to explore its structure and topology. The result shows that the resting brain functional network is a scale-free small-world network. By computing the centrality indices of the functional network, we find that the precuneus and cingulate gyrus are connected widely with many other nodes in the network, and acted as the hub areas with shortest paths connecting to other nodes. Furthermore, by studying its community structure with a probabilistic mixture model, the whole network is found to consist of 10 sub-networks, including the visual system, auditory system, motor system, default-mode network and the brain regions associated with the executive and working memory functions. These findings suggest that the resting brain functional network is composed of sub-networks which are both independent and overlapping each other. The precuneus and cingulate gyrus, which are the kernel nodes of the network, play the important roles in dispatching and transferring information.Brain functions of the schizophrenics with positive and negative symptoms are also examined in the senses of function asymmetry properties and network topological structure, by means of network analysis and other methods. Firstly, altered functional connectivity asymmetries of both symptoms are examined from perspective of functional integration. Functional connectivity maps, which involve the dorsolateral prefrontal cortex, superior temporal gyrus, middle temporal gyrus and hippocampus as regions of interest (ROIs), are created. A novel voxel-based two-level asymmetry analysis method is then proposed to investigate the differences among the positive group, negative group and control group in asymmetry patterns of the functional connectivity within groups and between groups. The results show that the positive group has abnormal leftward asymmetry whereas the negative group has abnormal rightward asymmetry. In addition, the strength of functional asymmetry in frontal lobe, temporal lobe, cingulate gyrus and right precuneus are found to be significant correlated with the symptom ratings of PANSS (Positive and Negative Syndrome Scale). After that, the differences of topological properties of brain functional networks between the two schizophrenic groups are studied based on the complex network theory in the manner of integration. All three groups show small-world properties but these properties are reduced in the schizophrenics, nevertheless the reduction in the negative group is a bit higher than that in the positive group. These results demonstrate that the brain functional networks are altered in schizophrenics, and the positive group and negative group show different network topology. These findings suggest that the negative and positive subtypes of schizophrenics may display a different basis of neurobiology.The relationship between functional connectivity and structural connectivity are analyzed by combining fMRI and diffusion tensor imaging (DTI). A new approach is proposed to locate the ROIs in the functional connectivity analysis from the results of DTI. DTI data are first taken to compute the fractional anisotropy (FA) map of each individual. A quasi-optimal voxel morphological method is then introduced to construct the white matter template to normalize those FA maps. After that, abnormal brain areas in the FA maps are extracted, and the correspondent gray matter areas are taken in the construction of the functional connectivity maps. These procedures are applied on experimental datasets from a group of early-onset schizophrenics during resting. These results demonstrate that, in early-onset schizophrenics, there are both structural and functional abnormalities, and support the view that the white matter lesions may disrupt the neural circuits between the frontal regions and other regions, and affect the functional connectivity in the frontal cortex.