Nick Scholand^1,2,3, Christina Graf¹, and Martin Uecker^1,2,3
1Institute of Medical Engineering, Graz University of Technology, Graz, Austria, ²Department of Interventional and Diagnostic Radiology, University Medical Center Göttingen, Göttingen, Germany, ³German Centre for Cardiovascular Research, Göttingen, Germany

The Bloch equations describe the effects of relaxation and external fields on the magnetization. Its coefficients are time-dependent, requiring computationally expensive techniques for its solution in the generic case. Currently, most techniques used in MRI are based on temporal discretization, but this requires very small time steps to achieve sufficiently small approximation errors.
In this work, we investigate a simulation technique based on a precise precomputation of state-transition matrices and compare its accuracy and efficiency with techniques based on ODE solvers and with matrix-based techniques using temporal discretization.

Ke Wang^1,2, Anastasios Angelopoulos¹, Alfredo De Goyeneche¹, Amit Kohli¹, Efrat Shimron¹, Stella Yu^1,2, Jitendra Malik¹, and Michael Lustig^1
1Electrical Engineering and Computer Sciences, University of California, Berkeley, Berkeley, CA, United States, ²International Computer Science Institute, University of California, Berkeley, Berkeley, CA, United States

Deep-learning (DL)-based MRI reconstructions have shown great potential to reduce scan time while maintaining diagnostic image quality. However, their adoption has been plagued with fears that the models will hallucinate or eliminate important anatomical features. To address this issue, we develop a framework to identify when and where a reconstruction model is producing potentially misleading results. Specifically, our framework produces confidence intervals at each pixel of a reconstruction image such that 95% of these intervals contain the true pixel value with high probability. In-vivo 2D knee and brain reconstruction results demonstrate the effectiveness of our proposed uncertainty estimation framework.

Jonathan Endres¹, Hoai Nam Dang¹, Felix Glang², Alexander Loktyushin², Simon Weinmüller¹, and Moritz Zaiss^1,2
1Department of Neuroradiology, Universitätsklinik Erlangen, Erlangen, Germany, ²Magnetic Resonance Center, Max-Planck-Institute for Biological Cybernetics, Tübingen, Germany

We propose an alternate method for simulating arbitrary MRI sequences, forming complete Bloch simulations. It is closely related to Extended Phase Graphs but imposes no restrictions on the alignment of timing or gradients, extends it to support T₂' relaxation and 3D gradient encoding, is fully differentiable, provides additional insight into the composition of the measured signal while outperforming isochromat-based Bloch simulations in execution speed by using an analytical description of the signal instead of relying on a Monte-Carlo simulation.

Santiago Coelho¹, Els Fieremans¹, and Dmitry S. Novikov^1
1Center for Advanced Imaging Innovation and Research (CAI2R), Department of Radiology, New York University, School of Medicine, New York, NY, United States

We studied the symmetries of the diffusion and covariance tensors, both accessible through a general b-tensor acquisition. We showed that there is an unexplored contrast in the non-symmetric part of the covariance tensor and use it to compute the size-shape covariance of compartmental diffusion tensors. We measure all rotational invariants of the cumulant tensors in a normal volunteer, explore the novel size-shape contrast, and provide the relations between these invariants and typical diffusion contrasts.

William M Wells¹, Stephan E Maier¹, and Carl-Fredrik Westin^1
1Brigham and Women's Hospital, Harvard Medical School, Boston, MA, United States

There is great interest to quantify the spectrum of diffusivities that underlie the observed diffusion signal decay; separate tissue compartments can be identified by their spectral peaks. This spectrum is defined by the inverse Laplace transform, but unfortunately this transform is very sensitive to noise omnipresent in diffusion MRI. We present a Bayesian method of inverse Laplace transform that uses Gibbs sampling to provide spectra along with an estimate of the noise related uncertainty in the spectra; this uncertainty information is valuable in interpreting the results. This is applied to multi-b high SNR endorectal coil diffusion data of the prostate.

Elizabeth Powell¹, Yolanda Ohene^2,3, Marco Battiston⁴, Laura M Parkes^2,3, and Geoff JM Parker^5,6
1Centre for Medical Image Computing, Department of Computer Science, University College London, London, United Kingdom, ²Division of Neuroscience and Experimental Psychology, University of Manchester, Manchester, United Kingdom, ³Geoffrey Jefferson Brain Research Centre, University of Manchester, Manchester, United Kingdom, ⁴NMR Research Unit, Queen Square Multiple Sclerosis Centre, Department of Neuroinflammation, UCL Queen Square Institute of Neurology, University College London, London, United Kingdom, ⁵Centre for Medical Image Computing, Department of Medical Physics and Biomedical Engineering, University College London, London, United Kingdom, ⁶Bioxydyn Limited, Manchester, United Kingdom

We propose compartmental modelling of blood-brain barrier (BBB) water exchange measurements using diffusion-filtered exchange imaging (FEXI), with the aim of providing greater biophysical insight into BBB function than is possible using the apparent exchange rate (AXR) approach. As relaxation time differences between blood and extravascular tissue have not yet been accounted for in FEXI-based BBB permeability measurements, we use simulations to quantify potential biases in exchange rates from both the AXR and compartmental approaches. Finally, we evaluate the repeatability of the AXR and compartmental models in a cohort of healthy subjects.

Emily Louise Baadsvik¹, Markus Weiger¹, Romain Froidevaux¹, Wolfgang Faigle², Benjamin Victor Ineichen³, and Klaas Paul Pruessmann^1
1Institute for Biomedical Engineering, ETH Zurich and University of Zurich, Zurich, Switzerland, ²Neuroimmunology and MS Research Section, Neurology Clinic, University of Zurich, University Hospital Zurich, Zurich, Switzerland, ³Department of Neuroradiology, Clinical Neuroscience Center, University Hospital Zurich, University of Zurich, Zurich, Switzerland

Various techniques for myelin mapping based on signal from the lipid-protein bilayer have been proposed, and a common way of validating these techniques is using ex-vivo tissue samples. However, it is still unclear to what extent experimental factors such as tissue storage conditions and processing affect MR signal. In this work, we investigate how long-term deep-frozen storage impacts tissue signal, and evaluate whether signal component mapping is feasible in non-D₂O-exchanged samples. We also determine whether animal tissue can act as a substitute for human tissue and investigate signal differences between white and grey matter.