Research On Task-state FMRI Data And Rest-state FMRI Data Using Independent Component Analysis

Functional Magnetic Resonance Imaging(fMRI) is a functional brain imaging technique which was developed in the early nineties. Compared with other non-invasive brain function positioning technology,the advantages of fMRI are non-radioactive, non-invasive, repeatable,high temporal and spatial resolution.Therefore fMRI has become the most effective technique for detecting brain activity. In recent years, fMRI has been the rapid development and wide application in brain science and neuroscience,which from the study of human brain motor, sensory, visual and auditory functions to the study of human cognition, emotion, thought, intelligence, and learning. In addition, functional magnetic resonance imaging is also used in clinical medicine. There are model-driven research approach and data-driven research approach for functional magnetic resonance imaging data.Independent component analysis is a common data-driven approach.This paper research on task-state fMRI data and rest-state fMRI data using independent component analysis. Firstly independent component analysis is applied to feature binding of color and shape. Draw a conclusion the role of various brain regions in feature binding. Provide a theoretical basis for the establishment of the cognitive model of visual feature binding. Secondly independent component analysis is applied to the diagnosis and identification of depression. Analysis of the abnormal brain regions between depression and normal. Provide supplementary diagnostic methods for depression. The main work are as follows: 1. For research purposes, analysis and comparison of the common independent component algorithm. Choose the most suitable algorithm according to the experimental fMRI data.2. Research on fMRI data in task-state using independent component analysis method. Study feature binding of color and shape using independent component analysis. Firstly collect task-state fMRI data.Extract independent component after preproccess. Select independent component corresponding to the experimental task using multiple regression analysis and correlation analysis. Calculate independent component using cognitive subtraction and draw a conclusion the role of various brain regions in feature binding. Further analysis activation of brain regions and cognitive processes in the feature binding task. Establish theoretical basis for revealing the human visual perception process. Provide theoretical basis for the establishment of the cognitive model of visual feature binding.3. Research on fMRI data in rest-state using independent component analysis. Research on depression fMRI data and normal fMRI data using independent component analysis. Firstly collect depression fMRI data and normal fMRI data in rest-state. The data is divided into training group and testing groups. Training group is divided into depression group and the normal group. Extract team independent component of training group. Use the default network which was proposed by Greicius. Select the depression group reference component and normal group reference component. Analyze each reference component in testing group to achieve the classification of normal and depression. Detect significant differences brain areas in brain function network connection between depression and normal. Provides a method for the diagnosis of depression.4. Compare independent component analysis in task-state and resting-state.Compare them from experimental range, experimental design, Independent component extraction and selection, purpose and results of the independent component analysis. Effectively extract activated brain areas in experiment task using correlation analysis and regression analysis. Extract lesions of the brain area from neurological patients or psychiatric patients by analysis default network.Provides a basis for selection independent component analysis of fMRI data under different states.