Magnetic Resonance Diffusion Tensor Imaging Studies In Medial Temporal Lobe Epilepsy

Epilepsy is a common neurological disorder, and its electrophysiology characteristic isabnormally high excitability and synchronization of the neural activity. It may lead to themicroenvironment changes in the brain locally, thus cause the molecular motion changes of waterdiffusion. Using diffusion tensor imaging could extract, descript and study the diffusion anisotropy ofwater molecules; it provides a safe and effective method for pathophysiological mechanism ofepilepsy.This study focuses on medial temporal lobe epilepsy, whose pathology characteristic is based onthe hippocampal sclerosis. The conventional imaging examination is difficult to observe theirsymptoms related brain structure abnormalities; this influenced the clinical diagnosis and treatment ofthe mTLE patients seriously. Due to the loss of the hippocampus neurons, the intercellular space ofbroadening and then changes the water molecules diffusion movement, leading to the degree ofdiffusion anisotropy and the average diffusion coefficient. So we can get mTLE patients structureform of the damaged areas related parameters by DTI data processing. This study aims to multi-modalanalysis for mTLE patients by3D T1imaging and DTI, combing with gray and white matter toobserve the abnormal brain structure of mTLE patients, Paper main research content is as follows:1. This study research the image denosing processing for DTI image by the partial differentialequation (PDE). We established the selective smoothing diffusion, directed diffusion, fourth orderpartial differential and the total variational diffusion based on P-M model for the original DTI imagesand color-FA images. At the same time, we used Tukey double weight function for correction andperformance testing. This experiment found that PDE for DTI denosing can be continuous domainmodeling, making the image model does not depend on image edge and has been simplified, theoperation is fast, accurate and stable.2. This study adopted a optimization which based on voxel-based morphometry (VBM) dataprocessing method for gray matter of mTLE patients. This optimized algorithm for brain analysisdetect the gray matter of the differences locally, and detect gray matter volume and density variationmore accurately, thus to evaluate corresponding abnormal neural anatomical morphology objectivelyand comprehensively. We found gray matter volume reduced in bilateral temporal lobe, frontal lobeand thalamic of mTLE patients.3. This study deal with the white matter fiber tracking based on the tract-based spatial statistics,TBSS) for mTLE patients, and does the spatial correlation analysis between the abnormal whitematter integrity and gray matter volume. We found that abnormal white matter fiber almost in the brain stem fibers, project fibers, associate fibers, contact fibers and the limbic system fibers; Inaddition, the gray matter and white matter abnormalities in mTLE patients have high correlationcoefficient and strong spatial distribution uniformity, especially in the limbic system.