Share Email Print

Proceedings Paper

Automatic registration of imaging mass spectrometry data to the Allen Brain Atlas transcriptome
Author(s): Walid M. Abdelmoula; Ricardo J. Carreira; Reinald Shyti; Benjamin Balluff; Else Tolner; Arn M. J. M. van den Maagdenberg; B. P. F. Lelieveldt; Liam McDonnell; Jouke Dijkstra
Format Member Price Non-Member Price
PDF $14.40 $18.00

Paper Abstract

Imaging Mass Spectrometry (IMS) is an emerging molecular imaging technology that provides spatially resolved information on biomolecular structures; each image pixel effectively represents a molecular mass spectrum. By combining the histological images and IMS-images, neuroanatomical structures can be distinguished based on their biomolecular features as opposed to morphological features. The combination of IMS data with spatially resolved gene expression maps of the mouse brain, as provided by the Allen Mouse Brain atlas, would enable comparative studies of spatial metabolic and gene expression patterns in life-sciences research and biomarker discovery. As such, it would be highly desirable to spatially register IMS slices to the Allen Brain Atlas (ABA). In this paper, we propose a multi-step automatic registration pipeline to register ABA histology to IMS- images. Key novelty of the method is the selection of the best reference section from the ABA, based on pre-processed histology sections. First, we extracted a hippocampus-specific geometrical feature from the given experimental histological section to initially localize it among the ABA sections. Then, feature-based linear registration is applied to the initially localized section and its two neighbors in the ABA to select the most similar reference section. A non-rigid registration yields a one-to-one mapping of the experimental IMS slice to the ABA. The pipeline was applied on 6 coronal sections from two mouse brains, showing high anatomical correspondence, demonstrating the feasibility of complementing biomolecule distributions from individual mice with the genome-wide ABA transcriptome.

Paper Details

Date Published: 21 March 2014
PDF: 7 pages
Proc. SPIE 9034, Medical Imaging 2014: Image Processing, 90343M (21 March 2014); doi: 10.1117/12.2043653
Show Author Affiliations
Walid M. Abdelmoula, Leiden Univ. Medical Ctr. (Netherlands)
Ricardo J. Carreira, Leiden Univ. Medical Ctr. (Netherlands)
Reinald Shyti, Leiden Univ. Medical Ctr. (Netherlands)
Benjamin Balluff, Leiden Univ. Medical Ctr. (Netherlands)
Else Tolner, Leiden Univ. Medical Ctr. (Netherlands)
Arn M. J. M. van den Maagdenberg, Leiden Univ. Medical Ctr. (Netherlands)
B. P. F. Lelieveldt, Leiden Univ. Medical Ctr. (Netherlands)
Delft Univ. of Technology (Netherlands)
Liam McDonnell, Leiden Univ. Medical Ctr. (Netherlands)
Jouke Dijkstra, Leiden Univ. Medical Ctr. (Netherlands)

Published in SPIE Proceedings Vol. 9034:
Medical Imaging 2014: Image Processing
Sebastien Ourselin; Martin A. Styner, Editor(s)

© SPIE. Terms of Use
Back to Top