An automated pipeline for extracting quantitative histological metrics for voxelwise MRI-histology comparisons

We propose an automated pipeline to extract a quantitative measure for histological staining density (stain area fraction) from immunohistochemical (IHC) stains for voxelwise MRI-histological comparisons. This pipeline addresses IHC artefacts and is generalisable to multiple stains. We applied the pipeline to post-mortem human brain data to generate stain area fractions maps for myelin, neurofilaments, and microglia. We use this pipeline to help perform whole-slide voxelwise comparisons between MRI- and IHC-derived parameters. Our results question some previously reported MRI-histology correlations and emphasize the need to consider multiple microstructural sources when explaining MRI signals.