Interlocking fMRI and Allen Brain Atlas: paving the way for new investigations of structural-functional relationships in (transgenic) mice
-
1
I. f. Phamacology, Germany
I. Purpose:
Functional brain imaging by MRI (fMRI) is non-invasive and allows for repetitive, non-invasive measurements of CNS functions. Recent developments allowed investigating brain structure specific function in transgenic animals. This allows detailed investigations of the impact genetic modifications on brain functions. Moreover various databases are established by 3D mapping of gen specific expressions. Here we establish a workflow to fuse all such datasets in a common reference (atlas) system supporting mutual data probing.
II. Methods:
The major image analysis tool used in our working group, MagnAn, is based on IDL (® ExelisVIS), a high-level development environment for data analysis and visualization. MagnAn offers also routines for segmentation, file conversion and registration of fMRI-images. Registration with MagnAn is done by calling the appropriate Advanced Normalization Tools (ANTs), which allows registering images at sub-voxel level. FMRI datasets consists of functional data as well as anatomical data at identical positions as the functional data.
Registration Workflow
Because the anatomical data have higher spatial resolution they provide more information content for establishing the registration process. Therefore, registration transformation was derived from anatomical data and in a second step applied to the functional data. First the MRI data were converted from lab own format to nifty format including setting the correct orientation matrix and Right-Superior-Anterior (RSA)-orientation. As a fixed image, the ABA mouse template with the voxel size of 0.1 mm x 0.1 mm x 0.1 mm is used.
1) Registration of anatomical fMRI mouse images
As a first step, the brain segmentation of anatomical fMRI mouse images with dimensions: 256x256x26 (high resolution RARE images) was performed within MagnAn.
Next, a bias field correction is performed, as implemented in the ANTs N3 routine.
For registration of the anatomical images (as moving images) to the ABA mouse template (as a fixed image), the following steps are made:
(1) Center of mass alignment (which is an important step for optimizing the further steps);
(2) Generating a composite transform by calling “antsRegistration.exe”, which contains:
- a derived initial moving translation (type = affine transform);
- a rigid transform (type = Euler 3D transform);
- an affine transform (type = affine transform);
- a warping (type = displacement field transform).
Multiple metrics were tested and the metric chosen for computing the transformations was “mutual information (- mi)”, because of the multimodal characteristic of the used data, as well as for speed considerations. The exact parameters established for the registration were found using an empirical process.
The obtained transformation matrices and the displacement vector field are stored in a dedicated folder “ANTS_matrix” for being accessible for further application for registration both of anatomical and of functional images.
(3) Application of the transformations (rigid, affine and warping) by calling the ANTs routine “antsApplyTransforms.exe”.
The obtained warped output images have the dimensions of the ABA mouse template [114x80x132] and are cback-onverted afterwards to our lab own format for further analysis with the MagnAn routines.
2) Registration of functional fMRI mouse images
By using the General Linear Model (GLM) approach offered by BrainVoyager, functional maps (z-score-images) are obtained, which have to be registered to the ABA mouse template for identifying activated brain regions. This serves as a first step of further data analysis for stimulus driven as well as resting-state data leading. Based on this brain structure identification further analyses like graph theoretical analysis are routinely performed.
First, the functional images are converted to “.nifti”-format and then up scaled to the resolution of [256x256x26], to match the size of the anatomical images.
Next the transformation matrices and the displacement vector field obtained in the first step of registration of the corresponding anatomical images to the ABA mouse template are used for application to the functional images (as moving images) by calling the appropriate ANTs routine.
The obtained warped output images have the dimensions of the ABA mouse template [114x80x132] and are back-converted afterwards from “.nifti”- to “.float”-array-format for further processing.
III. Results:
We obtained a powerful pipeline tool allowing to fuse datasets from non-invasive functional imaging methods (BOLD fMRI) resp. anatomical imaging methods (MRI) to 3D gen expression patterns (ABA) and high resolution connectivity data (ABA). The software tools are smoothly integrated in the normal fMRI analysis workflow. Establishing the registration transformations takes around 5 min per dataset on average desktop PC's; applying the transformations to the functional data only takes a few seconds.
IV. Conclusion:
The presented tool allows for obtaining anatomical and functional MRI datasets registered to the ABA atlas. Because the ABA serves as the most detailed database for genetic as well as connectivity data, our methods allows for future detailed investigations of the relationships between structural and functional brain regions and there genetic expression pattern.
Acknowledgements
BMBF NeuroImpa TP4: 01EC1403C and NeuroRad TP D: 02NUK034D
Keywords:
fMRI,
BOLD,
Allen Brain Atlas,
Transgenic mice,
image registration,
Ants
Conference:
Neuroinformatics 2016, Reading, United Kingdom, 3 Sep - 4 Sep, 2016.
Presentation Type:
Poster
Topic:
Neuroimaging
Citation:
Ivan
CI,
Konerth
LC and
Hess
A
(2016). Interlocking fMRI and Allen Brain Atlas: paving the way for new investigations of structural-functional relationships in (transgenic) mice.
Front. Neuroinform.
Conference Abstract:
Neuroinformatics 2016.
doi: 10.3389/conf.fninf.2016.20.00064
Copyright:
The abstracts in this collection have not been subject to any Frontiers peer review or checks, and are not endorsed by Frontiers.
They are made available through the Frontiers publishing platform as a service to conference organizers and presenters.
The copyright in the individual abstracts is owned by the author of each abstract or his/her employer unless otherwise stated.
Each abstract, as well as the collection of abstracts, are published under a Creative Commons CC-BY 4.0 (attribution) licence (https://creativecommons.org/licenses/by/4.0/) and may thus be reproduced, translated, adapted and be the subject of derivative works provided the authors and Frontiers are attributed.
For Frontiers’ terms and conditions please see https://www.frontiersin.org/legal/terms-and-conditions.
Received:
30 Apr 2016;
Published Online:
18 Jul 2016.
*
Correspondence:
Prof. Andreas Hess, I. f. Phamacology, Erlangen, Germany, andreas.hess@fau.de