hippunfold 0.5.1

Snakemake BIDS app for hippocampal unfolding & subfield segmentation

This tool aims to automatically model the topological folding structure of the human hippocampus. It is currently set up to use sub-millimetric T2w MRI data, but may be adapted for other data types. This can then be used to apply the hippocampal unfolding methods presented in DeKraker et al., 2019, and ex-vivo subfield boundaries can be topologically applied from DeKraker et al., 2020.

The overall workflow can be summarized in the following steps:

0. Resampling to a 0.3mm isotropic, coronal oblique, cropped hippocampal block

1. Automatic segmentation of hippocampal tissues and surrounding structures via deep convolutional neural network U-net Li _et al_., 2017 _OR_ Manual segmentation of hippocampal tissues and surrounding structures using this protocol

2. Post-processing via fluid label-label registration to a high resolution, topoligically correct averaged template

3. Imposing of coordinates across the anterior-posterior, proximal-distal, and laminar dimensions of hippocampal grey matter via solving the Laplace equation

4. Extraction of a grey matter mid-surface and morpholigical features (thickness, curvature, gyrification index, and, if available, quantitative MRI values sampled along the mid-surface for reduced partial-voluming)

5. Quality assurance via inspection of Laplace gradients, grey matter mid-surface, and flatmapped features

6. Application of subfield boundaries according to predifined topological coordinates