Application Of 2-D FDR In Neuroimaging Data

The brain has been the most complex and elaborate system known to human so far. It is made up of brain regions which are linked anatomically and independent and influence one another functionally. In recent years, the emergence and rapid development of the brain imaging technique, especially the magnetic resonance imaging, MRI has provided the imaging tool to get all kinds of data of the brain, and lots of research is devoted to using neuroimaging data to look for the biomarkers of a variety of brain diseases, so as to discuss their pathogenesis.Looking for the reliable biomarkers of brain disease is a challenging job. At present, research on the resting-state networks mainly involves study of a single disease. Looking for the index of significant difference between the patients with a single mental disease, such as depression and schizophrenia and normal people and using that as potential biomarkers.In fact, a large number of experiments have shown that a lot of mental diseases share a pathological basis. Finding out the common biological markers of mental diseases makes more sense than seeking the biological marker of a single disease, but looking for the common biological markers of various mental diseases is a more difficult task. Concerning this issue, Gan came up with a new effective method—2-D fdr(local false discovery rate)in 2015. The method uses the joint distribution of the data and has a higher utility for the information among the data set of differentdiseases is fully retained.Therefore, the paper collects the Case 15 and Case 41 of fMRI data,which shows the patients with bipolar disorders in the period of mania and depression respectively and Case 49 of data of the normal control group and uses the method of 2-D fdr to study the connectivity differences of the common functions of the patients with bipolar disorders under different morbidity. We think the abnormal functional connection that coexists under different morbidity is more likely to be the biological marker of the patients with bipolar disorders, which provide some guidance for clinical practice.Our experiment results show that under the different morbidity, the patients with bipolar disorders has thirteen connections significantly different from those of the normal. These connections are mainly distributed in the brain regions of forehead brain regions, thalamus,fusiform gyrus, cingulated and lenticular nucleus. As to manic phase,there are also connections among the frontal lobes, and abnormal connections among front lobe, back rectus and cingulated gyrus, while depression phase also exists in the abnormal connection between frontal lobe and lingual gyrus, and hippocampus and fusiform gyrus.