cortical_myelin_segmentation

The Matlab routines allow to segment heavily and sparsely myelinated cortical layers from high-resolution (0.8 mm isotropic and higher) MP2rage images.. In short, the intensities of the MP2rage sequence outputs images are re-scaled and a UNI/T1map ratio image (RATIO) and INV1/INV2 (INVC) image calculated. The cortical rim and adjacent CSF and white matter are isolated and k-means clustering used to identify 6 intensity clusters corresponding to a white matter and csf cluster, two partial volume clusters and a sparsely and highly myelinated rim clusters. Please see Word document for detailed step-by-step instructions. Contact of author Susanne Mueller M.D. Center for Imaging of Neurodegenerative Diseases VAMC San Francisco 4150 Clement Street San Francisco, CA, 94121 USA USA e-mail: [e-mail] You need to have Matlab with the Image Processing, Statistics, Curve Fitting, and Optimization toolboxes and spm12 (https://www.fil.ion.ucl.ac.uk/spm/) Make sure you have the cortex_2_layer directory in your Matlab search path. The Matlab routines are not for commercial use. Disclaimer: Use of the Matlab routines at your own risk.

本Matlab脚本可从高分辨率（各向同性0.8mm及以上）的MP2rage图像中，对髓鞘富集与稀疏分布的大脑皮层分层实现分割。简言之，首先对MP2rage序列输出图像的灰度强度进行重缩放，并计算UNI/T1映射图比值图像（RATIO）与INV1/INV2（INVC）图像。随后分离出皮层边缘及其邻近的脑脊液（Cerebrospinal Fluid, CSF）与白质区域，并采用k-means聚类算法识别6个灰度簇：分别对应白质簇、脑脊液簇、两个部分容积簇，以及髓鞘稀疏与高度富集分布的皮层边缘簇。详细的分步操作指南请参阅配套Word文档。作者联系方式：苏珊·穆勒医学博士（Susanne Mueller M.D.），美国旧金山退伍军人医疗中心（VAMC）神经退行性疾病影像中心，地址：美国加利福尼亚州旧金山市克莱门特街4150号，邮编94121；电子邮箱：[e-mail]。使用前需确保您的Matlab环境已安装图像处理、统计、曲线拟合与优化工具箱，以及spm12软件（下载地址：https://www.fil.ion.ucl.ac.uk/spm/）。请确保cortex_2_layer目录已添加至Matlab的搜索路径中。本Matlab脚本仅可用于非商业用途。免责声明：使用本Matlab脚本风险自负。