White Matter Hyperintensity (WMH) Segmentation Evaluation Tools

## OverviewThis repository contains MATLAB scripts for evaluating agreement between manual and automated white matter hyperintensity (WMH) segmentations derived from brain MRI data.The tools were developed to support quantitative comparison of WMH segmentation approaches used in the associated manuscript:Al-Bachari SA, et al, "Comparison of Automated White Matter Lesion Segmentation Approaches for Use in Large, Multi-Site Data Analyses in Parkinson's Disease", Frontiers in Neuroscience, 2026, under reviewThe repository includes scripts for:Inter-rater Dice Similarity Coefficient assessment between two manual WMH segmentation masks.Evaluation of automated WMH segmentation outputs against a manual reference segmentation.Voxel-wise and lesion-wise performance analysis of multiple automated WMH segmentation approaches.Lobar WMH volume, Dice Similarity Coefficient and log absolute volume difference analysis using a FreeSurfer white matter lobar parcellation registered to FLAIR space.The automated methods supported by the evaluation scripts include:LPA, using FLAIR only.LPA, using bias-corrected FLAIR.BIANCA Leave-One-Out, using bias-corrected FLAIR.BIANCA Test, using bias-corrected FLAIR.FreeSurfer, using T1-weighted MRI, registered to FLAIR space.FreeSurfer, using bias-corrected T1-weighted MRI, registered to FLAIR space.UNet-PGS, using bias-corrected T1-weighted MRI, registered to FLAIR space.---