Patch flattening for high resolution topographical surveys of MRI & histology without surface meshes

Poster No:

1286 

Submission Type:

Abstract Submission 

Authors:

Omer Faruk Gulban1,2, Konrad Wagstyl3, Rainer Goebel4,2, Renzo Huber4

Institutions:

1Maastricht University, Mastricht, Netherlands, 2Brain Innovation, Maastricht, Netherlands, 3UCL, London, United Kingdom, 4Maastricht University, Maastricht, Netherlands

First Author:

Omer Faruk Gulban  
Maastricht University|Brain Innovation
Mastricht, Netherlands|Maastricht, Netherlands

Co-Author(s):

Konrad Wagstyl  
UCL
London, United Kingdom
Rainer Goebel  
Maastricht University|Brain Innovation
Maastricht, Netherlands|Maastricht, Netherlands
Renzo Huber  
Maastricht University
Maastricht, Netherlands

Introduction:

Partial coverage (not whole brain) images possess analysis challenges in the daily-lives of high-resolution imaging researchers (e.g. layer-(f)MRI, histology). One of these challenges is the problem of flattening patches of the cortex to study the cortical topography (e.g. activations, laminar changes). Flattening of the cortex is addressed for the whole-brain imaging cases since decades [1, 2, 3]. However, flattening of the cortex is not a straightforward process when combined with partial coverage images focused at small regions of interest (e.g. high resolution ultra-high field MRI). Here we demonstrate a method (inspired by the local Carthesian grid approach in [4]) that works in volume space without any requirement to use triangular meshes to represent the cortical surfaces at any stage of the algorithm. Out method is specifically developed for -but not limited to- flattening patches of the cortex. We implement our patch flattening method as a C++ program and make this tool available free and open source within LayNii v1.7.0 [5].

Methods:

Our patch flattening method only requires a segmented volume file. We start from a 4 class segmentation. Note that we are not providing a solution for tissue segmentation but expecting users to segment their data in their own preferred way. However we offer a program "LN2_RIMIFY" to convert any segmented nifti file into the format preferred by LayNii.

1. On the segmented volume file (referred as the 'rim' file in LayNii), we run our layering program "LN2_LAYERS" to measure (equi-volume) cortical depths of each voxel. We also create a middle gray matter volume. Note that, we are not using any "surfaces" (a.k.a. triangular meshes) that are commonly used in conventional MRI software. In addition, LN2_LAYERS can estimate and export cortical curvature measurements.

2. We select a single voxel that will be the center or our flattened cortex patch. For example, this voxel can be selected as the closest middle gray matter voxel to the centroid of a given activity blob. We use an interactive volume visualization and manipulation software to select this voxel (e. g. Freesurfer [1, 2], BrainVoyager [3], FSLeyes [6], ITK-SNAP [7]).

3. Using the segmented volume image together with the middle gray matter image where a center of the patch voxel is selected, we generate and parametrize a geodesic disc around the selected voxel. Then, we find 4 points that are equally distant to each other on the perimeter of the disc (using LN2_MULTILATERATE). This step parametrizes (injects a 2D coordinate system) on the subset of the middle gray matter voxels. Then the coordinates (called UV coordinates) are propagated to the rest of the cortical thickness to cover the whole cortical depth.

4. We use the 'UV coordinates' together with the equi-volume depth measurement to flatten a 3D image (e.g. T1w, or activation map) onto a new arbitrarily sized 3D rectangular lattice. This 3D flat projection with depth image is exported as a nifti file to be explored through aforementioned volume visualization softwares.

Results:

Figure 1 and 2 demonstrates cortical patch flattening processing steps (as briefly outlined above in the Methods section) together with the flat projections of curvature together with MRI and histology measurements.
Supporting Image: figure1.png
   ·Figure 1. Steps to flatten T2* in vivo MR images at 350 micron iso. resolution [8]. Region of interest is Heschl’s Gyrus. Cortical depth profiles of the flat images show penetrating vessels.
Supporting Image: figure2.png
   ·Figure 2. Steps to flatten BigBrain image at 100 micron iso. res. [9, 10]. Region of interest is Calcarine Sulcus. Cortical depth profiles show cortical layers and penetrating vessels.
 

Conclusions:

Our new programs "LN2_MULTILATERATE" and "LN2_PATCH_FLATTEN" are able to flatten patches of cortex without requiring triangular surface meshes used in the conventional cortical flattening tools [1, 2, 3]. This is beneficial as it allows users to segment smaller volumes of cortex while offering the ability to conduct topographical surveys of the cortical landscape at very high resolutions. In addition, our patch flattening method benefits studies where every imaging minute is valuable by making it possible to drop the whole brain imaging requirement.

Modeling and Analysis Methods:

Methods Development 1

Neuroanatomy, Physiology, Metabolism and Neurotransmission:

Cortical Anatomy and Brain Mapping 2

Novel Imaging Acquisition Methods:

Anatomical MRI

Keywords:

ANGIOGRAPHY
Cortex
Cortical Columns
Cortical Layers
Data analysis
HIGH FIELD MR
MR ANGIOGRAPHY
MRI
STRUCTURAL MRI
Other - 7T

1|2Indicates the priority used for review

My abstract is being submitted as a Software Demonstration.

Yes

Please indicate below if your study was a "resting state" or "task-activation” study.

Other

Healthy subjects only or patients (note that patient studies may also involve healthy subjects):

Healthy subjects

Was any human subjects research approved by the relevant Institutional Review Board or ethics panel? NOTE: Any human subjects studies without IRB approval will be automatically rejected.

Yes

Was any animal research approved by the relevant IACUC or other animal research panel? NOTE: Any animal studies without IACUC approval will be automatically rejected.

Not applicable

Please indicate which methods were used in your research:

Structural MRI
Postmortem anatomy

For human MRI, what field strength scanner do you use?

7T

Which processing packages did you use for your study?

Other, Please list  -   LayNii

Provide references using author date format

[1] Fischl, B., Sereno, M. I., Tootell, R. B. H., & Dale, A. M. (1999). High-resolution intersubject averaging and a coordinate system for the cortical surface. Human Brain Mapping, 8(4), 272–284. https://doi.org/10.1002/(SICI)1097-0193(1999)8:4<272::AID-HBM10>3.0.CO;2-4

[2] Fischl, B., Sereno, M. I., & Dale, A. M. (1999). Cortical surface-based analysis: II. Inflation, flattening, and a surface-based coordinate system. NeuroImage, 9(2), 195–207. https://doi.org/10.1006/nimg.1998.0396

[3] Goebel, R. (2012). BrainVoyager--past, present, future. NeuroImage, 62(2), 748–756. https://doi.org/10.1016/j.neuroimage.2012.01.083

[4] Kemper, V. G., De Martino, F., Emmerling, T. C., Yacoub, E., & Goebel, R. (2018). High resolution data analysis strategies for mesoscale human functional MRI at 7 and 9.4T. NeuroImage, 164, 48–58. https://doi.org/10.1016/j.neuroimage.2017.03.058

[5] Huber, L., ..., Gulban, O. F. (2020). LayNii: A software suite for layer-fMRI. BioRxiv. https://doi.org/10.1101/2020.06.12.148080

[6] McCarthy, Paul. (2020). FSLeyes (Version 0.34.0). Zenodo. http://doi.org/10.5281/zenodo.3937147

[7] Yushkevich, P. A., Piven, J., Hazlett, H. C., Smith, R. G., Ho, S., Gee, J. C., & Gerig, G. (2006). User-guided 3D active contour segmentation of anatomical structures: significantly improved efficiency and reliability. NeuroImage, 31(3), 1116–1128. https://doi.org/10.1016/j.neuroimage.2006.01.015

[8] Gulban, O. F. (2020). Chapter 6: In vivo T2* imaging of human auditory cortex at 350 μm isotropic resolution. Imaging the human auditory system at ultrahigh magnetic fields. ProefschriftMaken. https://doi.org/10.26481/dis.20201006og

[9] Amunts, K., Lepage, C., Borgeat, L., Mohlberg, H., Dickscheid, T., … Evans, A. C. (2013). BigBrain: An Ultrahigh-Resolution 3D Human Brain Model. Science, 340(6139), 1472–1475. https://doi.org/10.1126/science.1235381

[10] Wagstyl, K., Larocque, S., Cucurull, G., Lepage, C., Cohen, J. P. … Evans, A. C. (2020). BigBrain 3D atlas of cortical layers: Cortical and laminar thickness gradients diverge in sensory and motor cortices. PLOS Biology, 18(4), e3000678. https://doi.org/10.1371/journal.pbio.3000678