Figure 1 | Scientific Reports

Figure 1

From: Tract-based white matter hyperintensity patterns in patients with systemic lupus erythematosus using an unsupervised machine learning approach

Figure 1

Preprocessing workflow. Workflow of the fully automated approach. 3D-FLAIR images are reoriented and co-registered to the T1-weighted (T1w) before WMH segmentation. White matter hyperintensities (WMH) segmentation is performed with the Lesion Segmentation Toolbox-Lesion Growth Algorithm (LST-LGA) using T1w and FLAIR images. The volume and number of WMH are extracted from the WMH maps in T1-space. The WMH probability maps are transformed to Montreal Neurological Institute (MNI) space by applying the transformation from the T1w images. Those maps are masked by the Johns Hopkins University (JHU) white matter (WM) probability atlas to obtain the tract specific WMH volumes. To quantitatively assign WMH to specific WM tracts the probability values of superimposed voxels on the lesion map and the WM tract are multiplied and the resulting product is summed over the entire tract. FLAIR, Fluid-attenuated inversion recovery; FSL, FMRIB Software Library; FLIRT, Linear Image Registration Tool; BET, Brain Extraction Tool; ANTs, Advanced Normalization Tools.

Back to article page