Brain Functional Pattern Recognition Based On Structural Priori And Its Application In Migraine Patients

The relationship between structure and function of substance has always been a hot topic in natural science research.The structure is the basis of function,and function can act on the structure.The human brain has its special structure and complex function.The abnormal performance of the human brain can be attributed to the abnormal structure of the brain and the abnormal function of the brain.By studying the relationship between structure and function,we can find better features to express the brain network,so as to provide better detection indicators for clinical practice and help patients recover early.Migraine is a neurogenic chronic pain disease.At present,it is believed that migraines are caused by the abnormal cerebral cortex.Its pathological mechanism is not clear,and there are still many difficulties in clinical treatment.Therefore,it is an urgent task to find specific biomarkers for migraine patients.Magnetic resonance imaging is a non-invasive imaging method,which is very suitable for the study of the brain and is also widely used in clinical practice.This thesis uses the structural and functional magnetic resonance imaging and analysis method based on voxel and cortex,using the stepwise function connection,cortex thickness analysis in migraine patients studied the functional and structural abnormalities,and on this basis developed a fusion method of building the structure and function of the new brain network biology characteristics of migraine to pattern recognition.The main content of this thesis is divided into the following three aspects:First,in the voxel level space,we used functional magnetic resonance imaging,under the hypothesis of pain matrix,the method of using meta-analysis,find the interested area associated with pain management,calculate the migraine patients and normal person and all other parts of the brain in areas of interest gradually function connection,and obtained the optimal number of two groups of people.Through the replacement test,we found that there were differences in the transmission of pain information between migraine patients and healthy people in three regions,and the optimal pathway was shortened in all migraine patients,respectively in the left and right insula and left posterior central gyrus.Then,we analyzed the correlation between the clinical praxeology scores and the difference in the number of optimal steps between the two groups and found that the estrogen level was significantly correlated with the number of optimal steps.We conclude that migraine patients do have abnormalities in the function of communication compared with normal people,and this abnormality may be related to the decrease of estrogen levels.Second,in the cortex level space,we use the structure of magnetic resonance images,the five kinds of the morphological index,on the basis of the former research,we chose the gray matter of the cerebral cortex thickness index to measure migraine sufferers structural differences in the brain,through statistical tests and multiple correction,we can see that migraine patients after the prefrontal cortex and cingulate cortex,orbital frontal lobe,lateral prefrontal cortex thickness thinning of leaf area,the thin area is in have selfknowledge,self-policing default network,higher order cognitive control network,and somatosensory network,and in the pain matrix are the important nodes of pain processing.I believe that the thinning of gray matter tissue,which contains the most abundant neurons in the brain compared with other tissues,is an important reason for abnormal pain processing in migraine patients.We infer that the brain function of migraine patients may be limited due to the lack of neuronal activity.Third,both structural and functional abnormalities have been demonstrated in migraine patients,and we would like to further identify biomarkers with migraine characteristics that can provide an effective means for clinical diagnosis.We proposed a new brain network construction method based on the combination of structural and functional features.We normalized the cortical thickness difference of the brain regions with structural differences as the weight value of the functional expression,and used structural differences to limit the idea of functional expression,and constructed a structural priori functional network.After determining the feature space,we selected linear regression,partial least squares regression,support vector regression,and gradient lifting regression tree to fit the target variable with physiological characteristics(estrogen level).After evaluating the fitting effectiveness of the four models,we identified the characteristics of biomarkers with migraine characteristics based on feature importance judgments.