Unveiling the Hippocampus: Automated Morphometry and Subfield Segmentation with HippUnfold

HippUnfold: Automated hippocampal opening for morphometry, subfield segmentation and more.

My son has just published a new article.

Cloud computing is used to perform automatic hippocampus imaging. Each scan takes about 20 minutes.

The archicortical structure of the hippocampus is also different in terms of its folding patterns between individuals. HippUnfold is a robust and automated BIDS-App that can be used to define and index individual-specific hippocampal folds in MRI. It’s similar to the popular tools for neocortical reconstructed. Topology is the basis of homology, and such tailoring is crucial for inter-individual alignement. This topological framework allows for qualitatively new analyses on morphological or laminar structures in the hippocampus and its subfields. This is crucial for improving current neuroimaging analysis at both a micro- and mesoscale. HippUnfold combines state-of the-art deep-learning with previously developed topological constrains to create uniquely folded surfaces that fit a subject’s hippocampal configuration. It can be used with sub-millimetric MRIs and may even extend to microscopic images. This paper describes the unique power of HippUnfold for feature extraction and highlights its unique value when compared to other hippocampal analysis methods.

Brain Imaging Data Standards, deep learning, hippocampal subfields, hippocampus (human), image segmentation and magnetic resonance imaging.

(c) 2022, DeKraker et al.


Leave a Reply

Your email address will not be published. Required fields are marked *