Benjamin Ades-Aron¹, Hong-Hsi Lee¹, Heidi Schambra², Dmitry S. Novikov¹, Els Fieremans¹, and Timothy Shepherd^1
1Radiology, NYU School of Medicine, New York, NY, United States, ²Neurology, NYU Langone, New York, NY, United States

Diffusion MRI should be sensitive to early pathology or functional re-organization changes for small internal brainstem structures associated with ischemia, multiple sclerosis or neurodegeneration. Application of diffusion MRI to brainstem studies is challenged by limited spatial resolution, image distortion from skull base artifacts and bias introduced if diffusion contrast is also used for structure segmentation. We describe and evaluate a novel combination of Fast Gray Matter Acquisition T1 Inversion Recovery (FGATIR) denoising with deep learning, multi-modal nonlinear image co-registration and super-resolution techniques to improve the accuracy of small internal brainstem structure segmentation on advanced diffusion MRI data.

Maria Guadalupe Garcia Gomar¹, Kavita Singh¹, and Marta Bianciardi^1
1Radiology, Athinoula A. Martinos Center for Biomedical Imaging, Massachusetts General Hospital and Harvard Medical School, Boston, MA, United States, Charlestown, MA, United States

Brainstem nuclei such as the medullary reticular formation, raphe nuclei, pontine nuclei, locus coeruleus and subcoeruleus, among others, play a crucial role in arousal and sleep. Despite their involvement in critical functions, their study in living humans is challenging due to limited resolution and contrast of conventional MRI. First, we precisely delineated nine arousal/sleep brainstem nuclei in 7 Tesla MRI of healthy humans and generated their validated in-vivo stereotaxic probabilistic atlas. Second, we evaluated their structural connectivity using 7 Tesla High-Angular-Resolution-Diffusion-Imaging and probabilistic tractography. This atlas and connectome might help evaluating disorders of consciousness, sleep disorders and neurological disorders.

Bradley Karat¹, Jordan DeKraker¹, Uzair Hussain², and Ali Khan^1,2
1Department of Neuroscience, Schulich School of Medicine & Dentistry, Western University, London, ON, Canada, ²Centre for Functional and Metabolic Mapping, Robarts Research Institute, Western University, London, ON, Canada

Aberrations of hippocampal subfields and microstructure are well studied, however, current in-vivo measures of microstructure are non-specific towards intra-hippocampal pathways and neurites within subfields. This is partly due to its complex and folded structure. Although highly folded, the hippocampus maintains an intrinsic coordinate system with well defined axes along the anterior-posterior, proximal-distal, and inner-outer dimensions. Importantly, hippocampal microstructure tends to align with these axes. Specific macrostructural measures such as gradient vector fields can be derived from these axes. Leveraging these measures, we were able to find systematic correlations between macrostructure and  local microstructure in the hippocampus. 

Ivan Maximov^1,2, Dennis van der Meer², Ann-Marie de Lange², Tobias Kaufmann², and Lars T. Westlye^2
1Department of Health and Functioning, Western Norway University of Applied Sciences, Bergen, Norway, ²NORMENT, University of Oslo, Oslo, Norway

While diffusion MRI (dMRI) has been established as a powerful tool for studying human brain white matter, its potential for characterising structural grey matter properties waits for to be fully realised. dMRI is capable of providing unique complementary information about cortical microstructure, function and connectivity. Here, we describe the spatial distribution and individual differences in cortical grey matter microstructure across 17,646 UK Biobank (UKB) participants using three diffusion MRI approaches, namely, diffusion tensor imaging, kurtosis tensor imaging and spherical mean technique, and demonstrate its predictive value for brain age prediction using machine learning.

Iain P Bruce¹, Yixin Ma¹, and Allen W Song^1
1Brain Imaging and Analysis Center, Duke University, Durham, NC, United States

The cerebral cortex is comprised of short column structures that span across the cortical layers between white matter and the pial surface. Subtle abnormalities in this microstructure, which may not be visible on anatomical scans, can have a large effect on brain function and development. In an effort to delineate such abnormalities, this study presents a technique to accurately and effectively observe the cortical depth dependence of diffusion anisotropy and diffusivity in-vivo using sub-millimeter isotropic resolution diffusion tensor imaging.

Tonima Sumya Ali¹ and Fernando Calamante^1,2
1School of Biomedical Engineering, The University of Sydney, Sydney, Australia, ²Sydney Imaging, The University of Sydney, Sydney, Australia

We employ structural and diffusion MRI metrics to evaluate their effectiveness for cortical segregation based on myelination, neurite density and microstructure in cortex. Three cortical measures (T1w/T2w, FA and total apparent fibre density, AFD_total, demonstrated significant inter-correlations. Two track-weighted parameters (FOD and FA, measured in white matter adjacent to cortex), also showed significant correlations to cortical FA and AFD_total. The spatial cortical patterns corresponding to these parameters suggest they provide complementary information on cortical features, and may help improve our understanding of cortical organisation in vivo and eventually lead to a further multi-parametric model for cortical parcellation.

Robert Jones¹, Qiyuan Tian¹, Chiara Maffei¹, Jean Augustinack¹, Aapo Nummenmaa¹, Susie Huang¹, and Anastasia Yendiki^1
1Athinoula A. Martinos Center for Biomedical Imaging, Massachusetts General Hospital and Harvard Medical School, Charlestown, MA, United States

We propose an approach to extracting a generalized anisotropy profile from the ensemble average propagator, which can be obtained either from diffusion spectrum imaging (DSI) data or from a generalized DSI analysis of diffusion MRI data acquired on q-shells. We validate this approach on data from ex vivo human tissue, acquired at 9.4T with 0.25 mm spatial resolution. We show that the proposed anisotropy profile can be used to distinguish between various white- and gray-matter structures, revealing fundamental differences in their microstructure.

Emmanuelle Weber¹, Erpeng Dai¹, Christoph Leuze¹, Nikola T. Markov², Nicholas Tran¹, Mariko Bennett¹, Samuel Baker¹, Kerriann Casey¹, Kalanit Grill-Spector¹, Keren Haroush¹, and Jennifer A. McNab^1
1Stanford, Stanford, CA, United States, ²Buck Institute, Novato, CA, United States

High-resolution (0.15 mm isotropic) diffusion MRI of newborn and an adult marmoset brains were acquired. For the adult marmoset brain the diffusion MRI data of anterior and posterior parts of the brain were separately acquired and subsequently co-registered. Across the whole brain, fractional anisotropy (FA) was lower and mean diffusivity (MD) was higher (Fig. 4) in the newborn compared to the adult. The biggest differences were observed in white matter. Cingulate cortex region A25 showed the smallest difference, which aligns with previous studies showing its early stage development

Sebastian Papazoglou¹, Mohammad Ashtarayeh¹, Martina F. Callaghan², Mark D. Does^3,4,5,6, and Siawoosh Mohammadi^1,7
1Department of Systems Neurosciences, University Medical Center Hamburg-Eppendorf, Hamburg, Germany, ²Wellcome Centre for Human Neuroimaging, UCL Queen Square Institute of Neurology, University College London, London, United Kingdom, ³Department of Biomedical Engineering, Vanderbilt University, Nashville, TN, United States, ⁴Institute of Imaging Science, Vanderbilt University Medical Center, Nashville, TN, United States, ⁵Department of Radiology and Radiological Sciences, Vanderbilt University Medical Center, Nashville, TN, United States, ⁶Department of Electrical Engineering, Vanderbilt University, Nashville, TN, United States, ⁷Department of Neurophysics, Max Planck Institute for Human Cognitive and Brain Sciences, Leipzig, Germany

It is unknown whether the accuracy of DWI-model based estimates of the axonal water fraction could be enhanced by additional calibrations. In this study we compared three DWI-models (WMTI, NODDI, NODDIDTI) using histology data as gold standard in ex vivo mouse models with a broad dynamic range of axonal metrics. We found that all DWI-models improved with additional calibration. Models with fixed diffusivities benefited efficiently from an additional scaling, while the WMTI model was stronger affected by an additional offset. Furthermore, without any additional calibration, the axonal compartment biomarker from NODDIDTI could explain the data better than WMTI or NODDI.

Junye Yao¹, Zihan Zhou¹, Benjamin C. Tendler², Karla L. Miller², Lei Zhang³, Keqing Zhu³, Aimin Bao³, Hongjian He¹, and Jianhui Zhong^1,4
1Center for Brain Imaging Science and Technology, College of Biomedical Engineering and Instrumental, Zhejiang University, Hangzhou, China, ²Wellcome Centre for Integrative Neuroimaging, FMRIB, Nuffield Department of Clinical Neurosciences, University of Oxford, London, United Kingdom, ³National Human Brain Bank for Health and Disease, School of Brain Science and Brain Medicine, Zhejiang University, Hangzhou, China, ⁴Department of Imaging Sciences, University of Rochester, Rochester, NY, United States

Diffusion weighted images of formalin-fixed human brain hemispheres were used to study b-value dependence of diffusion parameters with the diffusion tensor and diffusion kurtosis model. Consistent decreases of FA with increasing b-values were observed in all hemispheres. Meanwhile, the mean kurtosis was found to increase with the discrepancy of FA between high and low b-value datasets. These results indicate that the b-value dependence of FA could be explained by non-gaussian diffusion effects consistent with tissue microstructure.

Alex Colin Valcourt Caron¹, Amir Shmuel², Ilana R. Leppert², and Maxime Descoteaux^1
1Université de Sherbrooke, Sherbrooke, QC, Canada, ²Montreal Neurological Institute, McGill University, Montreal, QC, Canada

The validation of advanced diffusion MRI methods requires acquisition at high-resolution and high SNR, which is difficult to obtain in humans. The use of anesthetized Non-Human Primates (NHPs) is key to unlocking valid microstructural maps. However, data acquisition and analysis tools must be adapted and configured for successful application to NHPs imaging. Here we propose a processing pipeline allowing for the analysis of diffusion data using DTI, CSD, and DIAMOND reconstruction, tailored for the analysis of data from the macaque brain. We present results obtained by applying the pipeline to a dataset acquired with a minimal CUSP 90 directions scheme.

Luke Joel Edwards¹, Evgeniya Kirilina^1,2, Carsten Jäger^1,3, Kirsten Garus⁴, Markus Cremer⁴, Katrin Amunts^4,5, and Nikolaus Weiskopf^1,6
1Department of Neurophysics, Max Planck Institute for Human Cognitive and Brain Sciences, Leipzig, Germany, ²Center for Cognitive Neuroscience Berlin, Freie Universität Berlin, Berlin, Germany, ³Paul Flechsig Institute of Brain Research, Leipzig University, Leipzig, Germany, ⁴Institute of Neuroscience and Medicine, Research Centre Jülich, Jülich, Germany, ⁵C. and O. Vogt Institute for Brain Research, Heinrich Heine University Düsseldorf, Düsseldorf, Germany, ⁶Felix Bloch Institute for Solid State Physics, Leipzig University, Leipzig, Germany

Higher order diffusion modelling (diffusional kurtosis and multi-tissue CSD) of a whole post mortem human brain with excellent tissue quality was enabled by ultra-high b values from a whole-body 3T scanner with ultra-strong gradients. This was complemented by a novel gradient nonlinearity correction scheme to interpolate signals onto diffusion shells before estimation of fibre distributions. The brain is part of the BigBrain initiative, and will undergo an extensive atlasing procedure. This dataset will thus extend the studies possible on the future BigBrain atlas, allowing investigation into diffusion microstructure imaging and tractography.

Michael Paquette¹, Cornelius Eichner¹, Guillermo Gallardo¹, and Alfred Anwander^1
1Neuropsychology, Max Planck Institute for Human Cognitive and Brain Sciences, Leipzig, Germany

The quantification of axon diameter from electron microscopy (EM) images requires a choice of heuristics to approximate the dimensions. This choice impacts the resulting size distributions and should be adapted to the task. In diffusion MRI, complex shapes are encoded based on their directional mean squared displacement. We compute diffusion-specific ground truth diameters using Monte-Carlo simulations and evaluate typical choices of EM image analysis heuristics against them. Further, we propose a new and simple heuristic based on the mean square displacement inside an ellipse to compute better axon diameter quantifications from electron microscopy.

Milena Capiglioni^1,2,3, Analía Zwick¹, Pablo Jimenez¹, and Gonzalo Álvarez^1,2
1Laboratorio de Espectroscopía e Imágenes por RMN, Departamento de Física Médica, Centro Atómico Bariloche - CNEA - CONICET, Bariloche, Argentina, ²Instituto Balseiro, CNEA, Universidad Nacional de Cuyo, Bariloche, Argentina, ³Support Center for Advanced Neuroimaging (SCAN), Institute for Diagnostic and Interventional Neuroradiology, University Hospital of Bern, University of Bern, Bern, Switzerland

Extracting quantitative microstructure information of living-tissue by non-invasive imaging is an outstanding challenge for understanding disease mechanisms. We introduce a method to make selectively images of microstructure-sizes by probing molecule diffusion with MRI. It relies on designing dynamical control of nuclear spins to sense magnetization “decay-shifts” rather than the commonly used spin-echo decay-rates. The feasibility and performance of the method are illustrated with proof-of-principle experiments and simulations on typical size-distributions of white-matter in the mouse brain. These results position spin-echo decay-shifts as a promising MRI tool to perform non-invasive histology without assuming a microstructure distribution model.

Mohammad Ashtarayeh¹, Laurin Mordhorst¹, Maria Morozova^2,3, Tobias Streubel^1,2, Jan Malte Oeschger¹, Joao Periquito⁴, Andreas Pohlmann⁴, Henriette Rusch³, Carsten Jäger², Thoralf Niendorf⁴, Nikolaus Weiskopf^2,5, Markus Morawski^2,3, and Siawoosh Mohammadi^1,2
1Department of Systems Neurosciences, University Medical Center Hamburg-Eppendorf, Hamburg, Germany, ²Department of Neurophysics, Max Planck Institute for Human Cognitive and Brain Sciences, Leipzig, Germany, ³Paul Flechsig Institute of Brain Research, University of Leipzig, Leipzig, Germany, ⁴Berlin Ultrahigh Field Facility (B.U.F.F.), Max-Delbrueck-Center for Molecular Medicine in the Helmholtz Association, Berlin, Germany, ⁵Felix Bloch Institute for Solid State Physics, Faculty of Physics and Earth Sciences, Leipzig University, Leipzig, Germany

We used a new method for validation of MRI-based axon radius index (ARI) mapping using large-scale light microscopy (lsLM) that provides a good representation of the fraction of large axons - the main contributors to the MRI-based ARI. The proposed method captures 100-1000 times more axons than current standard small field of view microscopy. We showed that the one-to-one correspondence between MRI-based ARI and lsLM-based effective axon radius is superior to the current standard method.

Omar Narvaez¹, Maxime Yon², Alejandra Sierra¹, and Daniel Topgaard^3
1A.I. Virtanen Institute for Molecular Sciences, University of Eastern Finland, Kuopio, Finland, ²CEMHTI, French National Centre for Scientific Research, Paris, France, ³Department of Chemistry, Lund University, Lund, Sweden

Nonparametric distributions of cell sizes or diffusion tensors have recently been applied to analyze clinically relevant data acquired with advanced diffusion encoding schemes building on oscillating gradients, targeting the frequency-dependence and cell size, or more general q-vector trajectories focusing on the tensorial aspects. We introduce nonparametric D(ω)-distributions as a joint analysis framework taking both frequency-dependence and tensorial properties into account, and demonstrate the approach with ex vivo rat brain data acquired with gradient waveforms exploring the relevant dimensions of the tensor-valued encoding spectrum b(ω).

Ricardo Coronado-Leija¹, Santiago Coelho¹, Omar Narvaez², Jorge Larriva-Sahd³, Alonso Ramirez-Manzanares⁴, Luis Concha³, Dmitry S. Novikov¹, and Els Fieremans^1
1Radiology, New York University School of Medicine, New York, NY, United States, ²University of Eastern Finland, Kuopio, Finland, ³Instituto de Neurobiologia, Universidad Nacional Autonoma de Mexico, Queretaro, Mexico, ⁴Centro de Investigacion en Matematicas, Guanajuato, Mexico

In this work, we evaluated the parameters of the Standard Model (SM), augmented with a free water compartment, on ex vivo diffusion MRI, to detect longitudinal changes caused by axonal degeneration. Parameters of SM were estimated from the rotational invariants of the diffusion signal using a supervised machine learning approach. Axonal degeneration was induced by nerve injury using a rat retinal ischemia model. Comparisons with 2D histology derived metrics revealed that SM parameters are sensitive to changes caused by axonal degeneration, particularly axon loss. SM parameters also reveal presence of microglia, and increased orientation dispersion.

Naila Rahman^1,2, Kathy Xu², Nico Arezza^1,2, Kevin Borsos^1,2, Matthew Budde³, Arthur Brown^2,4, and Corey Baron^1,2
1Medical Biophysics, Western University, London, ON, Canada, ²Robarts Research Institute, London, ON, Canada, ³Neurosurgery, Medical College of Wisconsin, Milwaukee, WI, United States, ⁴Anatomy and Cell Biology, Western University, London, ON, Canada

Microstructural diffusion MRI (dMRI) improves the specificity required to detect microstructure changes related to pathophysiology. Oscillating gradient spin-echo (OGSE) dMRI is sensitive to structural disorder and microscopic anisotropy (µA) dMRI is sensitive to water diffusion anisotropy independent of neuron fiber orientation. In this work, OGSE and µA protocols were implemented to enable in vivo longitudinal scanning at 9.4T. Preliminary data in a rodent model of traumatic brain injury (TBI) revealed changes in mean diffusivity dependence on OGSE frequency post-TBI and changes in spherical tensor kurtosis (sensitive to cell size heterogeneity), compared to healthy mice.

Sidsel Winther^1,2, Mariam Andersson^1,2, Henrik Lundell², and Tim Dyrby^1,2
1DTU Compute, Technical University of Denmark, Kongens Lyngby, Denmark, ²Danish Research Center for Magnetic Resonance, Functional and Diagnostic Imaging and Research, Copenhagen University Hospital Hvidovre, Hvidovre, Denmark

Magnetic susceptibility of myelin induces morphology- and orientation-dependent perturbations of the B_0-field. At the microstructural scale of brain white matter, the main contribution to these perturbations comes from myelin. Here, we show that a consequence of this is a fiber-orientation-dependent bias of the DWI-signal across white matter regions in an ex vivo rat brain. This implies that biophysical modeling must regard susceptibility at the microstructural scale in order not to become biased as well. Higher field strength will increase the bias and vice versa. Hence, especially pre-clinical scans are affected.

Noam Omer¹, Ella Wilczynski¹, Neta Stern¹, Tamar Blumenfeld-Katzir¹, Meirav Galun², and Noam Ben-Eliezer^1,3,4,5,6
1Department of Biomedical Engineering, Tel-Aviv University, Tel-Aviv, Israel, ²Department of Computer Science and Applied Mathematics, Weitzman institute of science, Rehovot, Israel, ³Department of Orthopedics, Shamir Medical Center, Zerifin, Israel, ⁴Sackler Faculty of Medicine, Tel-Aviv University, Tel-Aviv, Israel, ⁵Sagol School of Neuroscience, Tel-Aviv University, Tel-Aviv, Israel, ⁶Center for Advanced Imaging Innovation and Research (CAI2R), New-York University Langone Medical Center, New York, NY, United States

Multicomponent T₂ analysis (mcT₂) can provide a clinically-useful myelin biomarker in vivo. Its clinical applicability, however, is hindered by lack of gold standard technique that can overcome the ambiguity of fitting several T₂ components to a single experimental signal. In this study we aimed to validate the utility of a novel data-driven mcT₂ mapping algorithm for quantifying myelin content. To that end, we applied the mcT₂ algorithm to 14 mice divided into two groups of mice: cuprizone-induced demyelination model, and controls. Results show excellent agreement between the mcT₂ based myelin biomarker and ground truth quantification of myelin from immunohistochemical staining.