A Study Of Abnormal Whiter Matter Of Patient With Primary Dysmenorrhea Based On MVPA

Primary dysmenorrhea(PDM),defined as painful menstrual cramps of uterine, is a widely presented gynecological disorder for women at reproductive age. Epidemiological studies reported that 40% to 84.8% of female adolescents have experienced PDM and 15% have had severe pain. The significant influence of PDM on female had caused tremendous socioeconomic loss. In order to find an effective way of treating the PDM, it is needed to further investigate the neurophysiological mechanism of this disorder. In recent years, neuroimaging techniques have been widely applied to investigate the abnormal brain of patients with chronic pain..Currently, a lot of studies related to the chronic pain have reported the findings of alterations in functional activity, gray matter(GM) and white matter(WM) in the brain of patients. Some research found the changes in functional activity and GM in the brain of patients with PDM, however, no alterations of WM were reported in the PDM patients. The PDM, as a kind of chronic pain, should also lead to the changes of WM. Meanwhile, the diffusion tensor imaging(DTI) technique can assess the orientation and connectivity of WM in vivo. Therefore, the present paper primarily utilizes the DTI to research the changes in WM structure of brain in patients with PDM.Firstly, we implement skeleton-based analytical way to investigate the differences in WM in patients with PDM. The result found that the patients have increased fractional anisotropy(FA), decreased mean diffusivity(MD) and radial diffusivity(RD) in many WM regions, compared to the healthy control subjects. However, there are no higher FA, lower MD and RD in healthy controls’ brain. The significant differences regions between patients and healthy controls involve the corpus callosum, fornix, internal capsule, external capsule, corona radiata, superior longitudinal fasciculus, cingulate gyrus, thalamic, etc. In addition, we found in some WM structures the augmented FA accompanying with the reduced MD and RD in patients with PDM. These regions included the corpus callosum, anterior corona radiata, posterior corona radiata, superior corona radiata, right superior longitudinal fasciculus.Secondly, we implemented the support vector machine to screen the significantdifferences of regions and measures so that we could obtain those vital regions and measures for analyzing the mechanism of neurophysiology of PDM. We extracted the mean value of measures in significant difference regions on the each measures, then put these mean values as features into the linear support vector machine to train the model, simultaneously calculated the weight of each feature. After training, we tested the performance of the trained model with another set of subjects. When testing was done, we got rid of the feature with the lowest weight, then iterated the procedure of training and testing model until all the features were eliminated. Thus we could gain the WM regions resulting the best testing performance, including corona radiata, corpus callosum, cingulate gyrus, superior longitudinal fasciculus, thalamic, related to the pain sensation. In conclusion, all these studies demonstrated that the patients with PDM had abnormal alterations in WM structures, these altered areas mainly involved the pathway of perceiving and controlling pain.