A Systematic Characterization of Structural Brain Changes in Epilepsy Using Magnetic Resonance Imaging

Abstract

Grey matter and white matter changes in epilepsy detected by Voxel-based morphometry (VBM) and Tensor-based morphometry (TBM) show notable similarities and differences to date. Therefore, a direct comparison of results using VBM and TBM is necessary to understand how different methods are sensitive in detecting structural brain changes in epilepsy. 45 patients with epilepsy (27.62 ±12.8 years old) and 46 healthy controls (34.25 ±16.0 years old) were scanned using a 3 Tesla MRI scanner. T1 weighted brain images were acquired, pre-processed, and further analyzed using Computational Anatomy Toolbox (CAT12). For VBM, voxel-wise grey matter volumes, white matter volumes and cerebrospinal fluid volumes were computed and smoothed (Full width at half maximum, FWHM= 8 mm) for each participant. For TBM, voxel-wise Jacobian determinant images were obtained and smoothed (FWHM= 8 mm). The group differences between patients and healthy controls were obtained using two-sample T-tests. Conjunction analyses were performed to explore the sensitivity of different methods in detecting grey matter and white matter. Our results showed that although both methods detected widespread structural changes, there are similarities and differences in grey matter and white matter findings detected by different methods. Only 35.07 % of grey matter (cerebellum, right temporal gyrus, left cuneus and left superior frontal) were commonly detected by VBM and TBM while only 26.08 % of white matter (anterior lobe of cerebellum, left occipital, frontal, and right temporal regions) were commonly detected by VBM and TBM. Therefore, we conclude that VBM and TBM are differently sensitive in detecting structural brain changes in epilepsy.