Print  Close

Quantifying human infra- and supra-granular layer properties using high-resolution ex vivo MRI

Poster No:

2120 

Submission Type:

Abstract Submission 

Authors:

Oula Puonti1,2, Xiangrui Zeng2,3, Areej Sayeed2,3, Rogeny Herisse2,3, Jocelyn Mora2,3, Kathryn Evancic2,3, Divya Varadarajan2,3, Yael Balbastre2,3, Irene Costantini4,5,6, Marina Scardigli4,5, Josephine Ramazzotti4,5, Danila DiMeo4,5, Giacomo Mazzamuto4,5,7, Luca Pesce4,5, Niamh Brady4,5, Franco Cheli4,5, Francesco Saverio Pavone4,5,7, Patrick Hof8, Robert Frost2,3, Jean Augustinack2,3, André van der Kouwe2,3, Juan Eugenio Iglesias2,3, Bruce Fischl2,3

Institutions:

1Danish Research Centre for Magnetic Resonance, Hvidovre, Region H, 2Department of Radiology, Athinoula A. Martinos Center for Biomedical Imaging, MGH, Charlestown, MA, 3Harvard Medical School, Dept. of Radiology, Boston, MA, 4National Research Council - National Institute of Optics (CNR-INO), Sesto Fiorentino, Italy, 5European Laboratory for Non-Linear Spectroscopy (LENS), Sesto Fiorentino, Italy, 6Department of Biology, University of Florence, Florence, Italy, 7Department of Physics and Astronomy, University of Florence, Florence, Italy, 8Nash Family Department of Neuroscience, Friedman Brain Institute, Icahn School of Medicine, New York, NY

First Author:

Oula Puonti  
Danish Research Centre for Magnetic Resonance|Department of Radiology, Athinoula A. Martinos Center for Biomedical Imaging, MGH
Hvidovre, Region H|Charlestown, MA

Co-Author(s):

Xiangrui Zeng  
Department of Radiology, Athinoula A. Martinos Center for Biomedical Imaging, MGH|Harvard Medical School, Dept. of Radiology
Charlestown, MA|Boston, MA
Areej Sayeed  
Department of Radiology, Athinoula A. Martinos Center for Biomedical Imaging, MGH|Harvard Medical School, Dept. of Radiology
Charlestown, MA|Boston, MA
Rogeny Herisse  
Department of Radiology, Athinoula A. Martinos Center for Biomedical Imaging, MGH|Harvard Medical School, Dept. of Radiology
Charlestown, MA|Boston, MA
Jocelyn Mora  
Department of Radiology, Athinoula A. Martinos Center for Biomedical Imaging, MGH|Harvard Medical School, Dept. of Radiology
Charlestown, MA|Boston, MA
Kathryn Evancic  
Department of Radiology, Athinoula A. Martinos Center for Biomedical Imaging, MGH|Harvard Medical School, Dept. of Radiology
Charlestown, MA|Boston, MA
Divya Varadarajan  
Department of Radiology, Athinoula A. Martinos Center for Biomedical Imaging, MGH|Harvard Medical School, Dept. of Radiology
Charlestown, MA|Boston, MA
Yael Balbastre  
Department of Radiology, Athinoula A. Martinos Center for Biomedical Imaging, MGH|Harvard Medical School, Dept. of Radiology
Charlestown, MA|Boston, MA
Irene Costantini  
National Research Council - National Institute of Optics (CNR-INO)|European Laboratory for Non-Linear Spectroscopy (LENS)|Department of Biology, University of Florence
Sesto Fiorentino, Italy|Sesto Fiorentino, Italy|Florence, Italy
Marina Scardigli  
National Research Council - National Institute of Optics (CNR-INO)|European Laboratory for Non-Linear Spectroscopy (LENS)
Sesto Fiorentino, Italy|Sesto Fiorentino, Italy
Josephine Ramazzotti  
National Research Council - National Institute of Optics (CNR-INO)|European Laboratory for Non-Linear Spectroscopy (LENS)
Sesto Fiorentino, Italy|Sesto Fiorentino, Italy
Danila DiMeo  
National Research Council - National Institute of Optics (CNR-INO)|European Laboratory for Non-Linear Spectroscopy (LENS)
Sesto Fiorentino, Italy|Sesto Fiorentino, Italy
Giacomo Mazzamuto  
National Research Council - National Institute of Optics (CNR-INO)|European Laboratory for Non-Linear Spectroscopy (LENS)|Department of Physics and Astronomy, University of Florence
Sesto Fiorentino, Italy|Sesto Fiorentino, Italy|Florence, Italy
Luca Pesce  
National Research Council - National Institute of Optics (CNR-INO)|European Laboratory for Non-Linear Spectroscopy (LENS)
Sesto Fiorentino, Italy|Sesto Fiorentino, Italy
Niamh Brady  
National Research Council - National Institute of Optics (CNR-INO)|European Laboratory for Non-Linear Spectroscopy (LENS)
Sesto Fiorentino, Italy|Sesto Fiorentino, Italy
Franco Cheli  
National Research Council - National Institute of Optics (CNR-INO)|European Laboratory for Non-Linear Spectroscopy (LENS)
Sesto Fiorentino, Italy|Sesto Fiorentino, Italy
Francesco Saverio Pavone  
National Research Council - National Institute of Optics (CNR-INO)|European Laboratory for Non-Linear Spectroscopy (LENS)|Department of Physics and Astronomy, University of Florence
Sesto Fiorentino, Italy|Sesto Fiorentino, Italy|Florence, Italy
Patrick Hof  
Nash Family Department of Neuroscience, Friedman Brain Institute, Icahn School of Medicine
New York, NY
Robert Frost  
Department of Radiology, Athinoula A. Martinos Center for Biomedical Imaging, MGH|Harvard Medical School, Dept. of Radiology
Charlestown, MA|Boston, MA
Jean Augustinack  
Department of Radiology, Athinoula A. Martinos Center for Biomedical Imaging, MGH|Harvard Medical School, Dept. of Radiology
Charlestown, MA|Boston, MA
André van der Kouwe  
Department of Radiology, Athinoula A. Martinos Center for Biomedical Imaging, MGH|Harvard Medical School, Dept. of Radiology
Charlestown, MA|Boston, MA
Juan Eugenio Iglesias  
Department of Radiology, Athinoula A. Martinos Center for Biomedical Imaging, MGH|Harvard Medical School, Dept. of Radiology
Charlestown, MA|Boston, MA
Bruce Fischl  
Department of Radiology, Athinoula A. Martinos Center for Biomedical Imaging, MGH|Harvard Medical School, Dept. of Radiology
Charlestown, MA|Boston, MA

Introduction:

Analyzing the structure of the human cortex is challenging due to its highly folded nature and microscopic cyto- and myeloarchitectural detail. Histology allows for visualizing the structure of the cortex at a single cell resolution. However, tissue cutting and mounting makes it infeasible to align serial sections to the level of accuracy that is required for precise modeling of laminar and cytoarchitectonic boundaries of the cortical layers in 3D [1,2]. Conventional in vivo MRI provides 3D whole-brain scans that can be used to construct surface models of the cortical sheet but lacks the resolution to extend this to the cortical laminae. In contrast, the spatial resolution of ex vivo MRI scans have been pushed down to 100 micrometers (um) [3], where the mesoscopic structure of the cortex becomes visible. Here we use an existing dataset of ex vivo MRI scans acquired at 120um resolution [4], and construct surface models of the border between the infra- and supra-granular layers.

Methods:

The MRI scans are first processed using a state-of-the-art, cascaded, multi-resolution U-Net architecture [5] to obtain probabilistic segmentations (values between 0 and 1) of the infra- and supra-granular layers. The white matter (WM), granular, and pial surfaces are placed using a modified version of the FreeSurfer recon-all pipeline, which was tailored to handle the additional surface (granular) and the high-resolution data. The modified pipeline: (i) generates a full volumetric segmentation at 1mm isotropic resolution using a contrast-agnostic neural network [6-8]; (ii) upsamples the WM mask to 120um resolution; (iii) refines the 120um WM mask with the confidence maps for increased accuracy; (iv) generates a pseudo T1-weighted image from the WM mask and confidence masks, which includes the infra- and supra-granular layers by linearly scaling the confidence values; and (v) processes this synthetic image with the recon-all surface placement pipeline [9,10] to produce the final surfaces and a cortical parcellation.

Results:

Fig 1A shows the MRI scan, surfaces, volumetric segmentation, and parcellation for a sample case. To study if we can detect architectonic borders of the primary visual area (V1), we computed intensity gradients at 20 different cortical depths between the WM and pial [2]. The gradient magnitudes were averaged over the depth excluding the first and last two depths to avoid partial volume effects. Fig 1B shows the gradient magnitude on the inflated surface of three different subjects, and Fig 1C shows the border of the V1 label, mapped from the fsaverage template, overlaid on the gradient magnitude. The label border matches the areas with increased gradient magnitude, especially on the lateral part of the V1 – indicating that the borders of V1 can be directly detected using the high-resolution surface models and MRI data. Fig 2A&B show the average thickness of the infra- and supra-granular layers (top) and its variance (bottom) over four subjects. Fig 2C shows the correlation of the curvature magnitude and the infra-granular layer thickness: the gyri (light gray) have a positive correlation whereas the sulci (dark gray) have a negative correlation, indicating that the infra-granular layers expand at the gyral crowns and compress at the sulcal fundi – which is consistent with histological studies [11] and computational cortical layer models [12].
Supporting Image: figure1.png
Supporting Image: figure2.png
 

Conclusions:

We have presented a preliminary quantitative analysis of the infra- and supra-granular layer properties using high-resolution ex vivo MRI data. The dataset will be expanded to include surfaces and cortical parcellations for 17 subjects, which allows for assessing the individual variability of the infra- and supragranular layers and the cortical region borders. The models can be used to inform in vivo MRI analysis paving a way for more accurate modeling of the cortex in neuroscience studies. The scans, surfaces and parcellations will be made freely available for download from the DANDI data archive.

Modeling and Analysis Methods:

Image Registration and Computational Anatomy
Methods Development
Segmentation and Parcellation 2

Neuroanatomy, Physiology, Metabolism and Neurotransmission:

Cortical Anatomy and Brain Mapping 1
Cortical Cyto- and Myeloarchitecture

Keywords:

Cortical Layers
Data analysis
HIGH FIELD MR
Modeling
STRUCTURAL MRI

1|2Indicates the priority used for review

Provide references using author date format

[1] Wagstyl K., et al. (2020) "BigBrain 3D atlas of cortical layers: Cortical and laminar thickness gradients diverge in sensory and motor cortices", PLoS Biol 18(4): e3000678

[2] Fischl B. and Sereno M.I. (2018), "Microstructural parcellation of the human brain", Neuroimage, Vol. 182, pp. 219-231

[3] Edlow, B.L., et al. (2019) "7 Tesla MRI of the ex vivo human brain at 100 micron resolution", Scientific Data, 6(244)

[4] Costantini, et al., (2023), "A cellular resolution atlas of Broca’s area", Science Advances, 9(41):eadg3844

[5] Zeng X., et al. (2023), "Segmentation of supragranular and infragranular layers in ultra-high resolution 7T ex vivo MRI of the human cerebral cortex", Submitted to arXiv.

[6] Gopinath, K., et al. (2023), "Cortical Analysis of Heterogeneous Clinical Brain MRI Scans for Large-Scale Neuroimaging Studies", In: Greenspan, H., et al. Medical Image Computing and Computer Assisted Intervention – MICCAI 2023. MICCAI 2023. Lecture Notes in Computer Science, vol 14227.

[7] Billot B., et al. (2023), "SynthSeg: Segmentation of brain MRI scans of any contrast and resolution without retraining", Medical Image Analysis, Vol 86:102789

[8] Iglesias J.E., et al. (2023), "SynthSR: A public AI tool to turn heterogeneous clinical brain scans into high-resolution T1-weighted images for 3D morphometry", Science Advances, 9:eadd3607

[9] Dale, A.M., et al. (1999), "Cortical Surface-Based Analysis I: Segmentation and Surface Reconstruction", NeuroImage, Vol. 9, pp. 179-194

[10] Fischl, B.R., et al. (1999), "Cortical Surface-Based Analysis II: Inflation, Flattening, and Surface-Based Coordinate System", NeuroImage, Vol. 9, pp. 195-207

[11] Bok, S. T. (1929). Der Einfluss der in den Furchen und Windungen auftretenden Krümmungen der Grosshirnrinde auf die Rindenarchitektur. Zeitschrift für die gesamte Neurologie und Psychiatrie, 121, 682-750.

[12] Waehnert M.D., et al., (2014), "Anatomically motivated modeling of cortical laminae", NeuroImage, Vol. 93