SliceMap: an algorithm for automated brain region annotation.

Abstract

SUMMARY: Many neurodegenerative disorders, such as Alzheimer's Disease, pertain to or spread from specific sites of the brain. Hence, accurate disease staging or therapy assessment in transgenic model mice demands automated analysis of selected brain regions. To address this need, we have developed an algorithm, termed SliceMap, that enables contextual quantification by mapping anatomical information onto microtome-cut brain slices. For every newly acquired high-resolution image of a brain slice, the algorithm performs a coarse congealing-based registration to a library of pre-annotated reference slices. A subset of optimally matching reference slices is then used for refined, elastic registration. Morphotextural metrics are used to measure registration performance and to automatically detect poorly cut slices. We have implemented our method as a plugin for FIJI image analysis freeware, and we have used it to regionally quantify tau pathology in brain slices from a tauopathy (P301S) mouse model. By enabling region-based quantification, our method contributes to a more accurate assessment of neurodegenerative disease development. AVAILABILITY AND IMPLEMENTATION: The method is available as a plugin for FIJI from https://github.com/mbarbie1/SliceMap/, along with an example dataset and user instructions. CONTACT: winnok.devos@uantwerpen.be. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.