Jelle Veraart¹, Daan Christiaens², Erpeng Dai³, Luke J. Edwards⁴, Vladimir Golkov⁵, Siawoosh Mohammadi⁶, Kurt G. Schilling⁷, Mohammad Hadi Aarabi⁸, Benjamin Ades-Aron^1,9, Nagesh Adluru¹⁰, Sahar Ahmad¹¹, Santiago Aja-Fernandez¹², Andrew L. Alexander¹⁰, Mariam Andersson¹³, Elixabete Ansorena¹⁴, Fakhereh Movahedian Attar⁴, Arnaud Attye¹⁵, Dogu Baran Aydogan^16,17, Steven H. Baete¹, Gianpaolo Antonio Basile¹⁸, Giulia Bertò¹⁹, Ahmad Beyh²⁰, Alberto Cacciola¹⁸, Maxime Chamberland²¹, Fernando Calamante²², Jenny Chen¹, Jian Chen²³, Shin Tai Chong²⁴, Santiago Coelho¹, Luis Concha²⁵, Ricardo Coronado-Leija¹, Michiel Cottaar²⁶, Daniel Cremers⁵, Szabolcs David^14,27, Alberto De Luca¹⁴, Flavio Dell'Acqua²⁰, Thijs Dhollander²³, Jason Druzgal²⁸, Tim B. Dyrby^13,29, Hernández-Torres Enedino¹³, Oscar Esteban³⁰, Els Fieremans¹, Leon Fonville³¹, Björn Fricke³², Martijn Froeling³³, Ian Galea³⁴, Gabriel Girard^35,36, Francesco Grussu³⁷, Chin-Chin Heather Hsu^38,39, Yung-Chin Hsu⁴⁰, Khoi Huynh¹¹, Matilde Inglese^41,42, Heide Johansen-berg²⁶, Derek K Jones⁴³, Kouhei Kamiya⁴⁴, Claire Kelly^45,46, Ahmad Raza Khan⁴⁷, Ali Khan⁴⁸, Yi-Chia Kung²⁴, Alberto Lazari²⁶, Alexander Leemans¹⁴, Laura Mancini^49,50, Ivan I. Maximov⁵¹, Harri Merisaari⁵², Malwina Molendowska⁴³, Benjamin T. Newman²⁸, Michael D. Noseworthy⁵³, Dmitry S. Novikov¹, Raquel Perez-Lopez³⁷, Franco Pestili¹⁹, Tomasz Pieciak¹², Marco Pizzolato²⁹, Álvaro Planchuelo-Gómez¹², Paul Polak⁵⁴, Erika P. Raven¹, Ricardo Rios-Carrillo²⁵, Viljami Sairanen⁵⁵, Simona Schiavi⁴¹, Pohchoo Seow⁵⁶, Dmitri Shastin⁴³, Yao-Chia Shih^56,57, Lucas Soustelle⁵⁸, Yue Sun⁵⁹, Karsten Tabelow⁶⁰, Chantal MW Tax^14,43, Guillaume Theaud⁶¹, Sjoerd B. Vos⁶², Ryckie G. Wade⁶³, Li Wang⁵⁹, Limei Wang⁵⁹, Thomas Welton^64,65, Lars T. Westlye⁶⁶, Stefan Winzeck^67,68, Joseph Yuan-Mou Yang^69,70, Pew-Thian Yap¹¹, Yukai Zou³⁴, Jennifer A. McNab³, Bennett A. Landman⁷, Nikolaus Weiskopf^4,71, and Maxime Descoteaux^61
1Center for Biomedical Imaging, Dept. Radiology, NYU School of Medicine, New York City, NY, United States, ²Department of Electrical Engineering (ESAT-PSI), KU Leuven, Leuven, Belgium, ³Department of Radiology, Stanford University, Stanford, CA, United States, ⁴Department of Neurophysics, Max Planck Institute for Human Cognitive and Brain Sciences, Leipzig, Germany, ⁵Computer Vision Group, Technical University of Munich, Munich, Germany, ⁶Department of Systems Neuroscience, University Medical Center Hamburg-Eppendorf, Hamburg, Germany, ⁷Vanderbilt University Medical Center, Nashville, TN, United States, ⁸Department of Neuroscience and Padova Neuroscience Center, University of Padova, Padua, Italy, ⁹Electrical and Computer Engineering, New York University, Brooklyn, NY, United States, ¹⁰Waisman Center and Department of Radiology, UW-Madison, Madison, WI, United States, ¹¹Department of Radiology and Biomedical Research Imaging Center (BRIC), University of North Carolina at Chapel Hill, Chapel Hill, NC, United States, ¹²LPI, ETSI Telecomunicacion, Universidad de Valladolid, Valladolid, Spain, ¹³Danish Research Centre for Magnetic Resonance, Copenhagen University Hospital Amager and Hvidovre, Copenhagen, Denmark, ¹⁴Image Sciences Institute, UMC Utrecht, Utrecht, Netherlands, ¹⁵Sydney Imaging, University of Sydney, Sydney, Australia, ¹⁶A.I. Virtanen Institute for Molecular Sciences, University of Eastern Finland, Kuopio, Finland, ¹⁷Department of Biomedical Engineering, Aalto University, Espoo, Finland, ¹⁸Brain Mapping Lab, Department of Biomedical, Dental Sciences and Morphological and Functional Images, University of Messina, Messina, Italy, ¹⁹Department of Psychology, The University of Texas at Austin, Austin, TX, United States, ²⁰NatBrainLab, Department of Forensic and Neurodevelopmental Sciences, King's College London, London, United Kingdom, ²¹Donders Institute for Brain, Cognition, and Behavior, Radboud University, Nijmegen, Netherlands, ²²Sydney Imaging and School of Biomedical Engineering, University of Sydney, Sydney, Australia, ²³Developmental Imaging, Murdoch Children's Research Institute, Melbourne, Australia, ²⁴Institute of Neuroscience, National Yang Ming Chiao Tung University, Tapei, Taiwan, ²⁵Institute of Neurobiology, Universidad Nacional Autónoma de México, Querétaro, Mexico, ²⁶Nuffield Department of Clinical Neurosciences, University of Oxford, Wellcome Centre for Integrative Neuroimaging, FMRIB, Oxford, United Kingdom, ²⁷Department of Radiation Oncology, UMC Utrecht, Utrecht, Netherlands, ²⁸Department of Radiology and Medical Imaging, University of Virginia, Charlottesville, VA, United States, ²⁹Department of Applied Mathematics and Computer Science, Technical University of Denmark, Kongens Lyngby, Denmark, ³⁰Department of Radiology, Lausanne University Hospital and University of Lausanne, Lausanne, Switzerland, ³¹Department of Brain Sciences, Division of Psychiatry, Imperial College London, London, United Kingdom, ³²Department of Systems Neuroscienc, University Medical Center Hamburg-Eppendorf, Hamburg, Germany, ³³Department of Radiology, University medical center Utrecht, Utrecht, Netherlands, ³⁴Department of Clinical Neurosciences, University of Southampton, Southampton, United Kingdom, ³⁵CIBM Center for Biomedical Imaging, Lausanne, Switzerland, ³⁶Radiology Department, Centre Hospitalier Universitaire Vaudois and University of Lausanne, Lausanne, Switzerland, ³⁷Radiomics Group, Vall d’Hebron Institute of Oncology, Vall d’Hebron Barcelona Hospital Campus, Barcelona, Spain, ³⁸Institute of Neuroscience, National Yang Ming Chiao Tung University, Taipei, Taiwan, ³⁹Center for Geriatrics and Gerontology, Taipei Veterans General Hospital, Tapei, Taiwan, ⁴⁰AcroViz Inc., Taipei, Taiwan, ⁴¹Department of Neurosciences, Rehabilitation, Ophthalmology, Genetics, Maternal and Child Health (DiNOGMI), University of Genoa, Genoa, Italy, ⁴²IRCCS Ospedale San Martino, Genoa, Italy, ⁴³Cardiff University Brain Research Imaging Centre (CUBRIC), Cardiff University, Cardiff, United Kingdom, ⁴⁴Department of Radiology, Toho University, Tokyo, Japan, ⁴⁵Turner Institute for Brain and Mental Health, School of Psychological Sciences, Monash University, Melbourne, Australia, ⁴⁶Victorian Infant Brain Studies, Murdoch Children's Research Institute, Melbourne, Australia, ⁴⁷Department of Advanced Spectroscopy and Imaging, Centre of Biomedical Research, Lucknow, India, ⁴⁸Department of Medical Biophysics, Western University, London, ON, Canada, ⁴⁹Lysholm Department of Neuroradiology, National Hospital for Neurology and Neurosurgery UCLH NHS FT, London, United Kingdom, ⁵⁰Brain Repair and Rehabilitation, University College London, London, United Kingdom, ⁵¹Western Norway University of Applied Sciences, Bergen, Norway, ⁵²Department of Clinical Medicine, University of Turku, Turku, Finland, ⁵³Electrical and Computer Engineering, McMaster University, Hamilton, ON, Canada, ⁵⁴School of Biomedical Engineering, McMaster University, Hamilton, ON, Canada, ⁵⁵BABA Center, Pediatric Research Center, Department of Clinical Neurophysiology, Children’s Hospital, University of Helsinki and Helsinki University Hospital, Helsinki, Finland, ⁵⁶Diagnostic Radiology, Singapore General Hospital, Singapore, Singapore, ⁵⁷Graduate Institute of Medicine, Yuan Ze University, Taoyuan City, Taiwan, ⁵⁸Aix Marseille Univ, CNRS, CRMBM, Marseille, France, ⁵⁹Developing Brain Computing Lab, Department of Radiology and Biomedical Research Imaging Center, University of North Carolina at Chapel Hill, Chapel Hill, NC, United States, ⁶⁰Weierstrass Institute for Applied Analysis and Stochastics, Berlin, Germany, ⁶¹Sherbrooke Connectivity Imaging Laboratory (SCIL), Université de Sherbrooke, Sherbrooke, QC, Canada, ⁶²Centre for Medical Image Computing, University College London, London, United Kingdom, ⁶³Leeds Institute for Medical Research, University of Leeds, Leeds, United Kingdom, ⁶⁴National Neuroscience Institute, Singapore, Singapore, ⁶⁵Duke-NUS Medical School, Singapore, Singapore, ⁶⁶Department of Psychology, University of Oslo, Oslo, Norway, ⁶⁷Department of Computing, Imperial College London, London, United Kingdom, ⁶⁸Department of Medicine, University Division of Anaesthesia, University of Cambridge, Cambridge, United Kingdom, ⁶⁹Department of Neurosurgery, The Royal Children's Hospital, Melbourne, Australia, ⁷⁰Neuroscience Research, Murdoch Children's Research Institute, Melbourne, Australia, ⁷¹Leipzig University, Felix Bloch Institute for Solid State Physics, Faculty of Physics and Earth Sciences, Leipzig, Germany

The preprocessing of dMRI data sets is a critical step in the experimental workflow that, in general, improves the data reliability. We  provide a comprehensive survey of the preprocessing workflows for dMRI data within our research community, assessing their variability and quantifying  its impact on the dMRI metric reproducibility and inter-site comparisons.  We observed that the lack of a standardized preprocessing pipeline across researchers is a significant source of variability that might lower the reproducibility of studies. These preliminary results highlight the need for harmonization of the image preprocessing workflow for dMRI data to promote more robust and reproducible analyses.

Chitresh Bhushan¹, Radhika Madhavan², Amod Jog³, Nastaren Abad¹, Brice Fernandez⁴, Luca Marinelli¹, H. Doug Morris⁵, Maureen Hood^5,6, J Kevin DeMarco⁶, Robert Y Shih^5,6, Gail Kohls⁶, Kimbra Kenney^5,6, Vincent B Ho^5,6, and Thomas K Foo^1,5
1GE Research, Niskayuna, NY, United States, ²GE Healthcare, Niskayuna, NY, United States, ³GE Research, Bangalore, India, ⁴GE Healthcare, Versailles, France, ⁵Uniformed Services University of the Health Sciences, Bethesda, MD, United States, ⁶Walter Reed National Military Medical Center, Bethesda, MD, United States

Mild traumatic injury (mTBI) patients exhibit acute symptoms including headaches and memory problems that sometimes persist for months, though CT/MRI scans appear normal. The purpose of this study was to identify functional biomarkers of mTBI during the 3-months following injury using seed-based functional connectivity derived from multi-echo resting state functional MRI.  In this work, we use multi-band multi-echo resting state fMRI to estimate seed-based functional connectivity. We propose a novel quantitative analysis methodology of suitable for voxel-wise single-subject assessment of disrupted functional connectivity by using an outlier-analysis approach.

William Kim¹, Guocheng Jiang^1,2, Nicholas Luciw^1,2, and Bradley MacIntosh^1,2
1Hurvitz Brain Sciences Program, Sunnybrook Research Institute, Toronto, ON, Canada, ²Department of Medical Biophysics, University of Toronto, Toronto, ON, Canada

Anatomical T1-weighted imaging allows us to examine relationships of regional morphology across the brain through structural covariance. Here, we investigated structural covariance stability using decreasing amounts of T1-weighted imaging data from the UK Biobank. Starting from 1,753 individuals, we found that it is possible to drastically reduce the sample and still maintain adequate stability (78% agreement with ~87 individuals). We note, however, that stability was regionally variable; lateral and cortical regions were least affected by sample size while medial and subcortical regions were most affected. These findings may inform sample size considerations for MRI-based structural covariance in large population studies.

Xiaowei Zhuang^1,2, Zhengshi Yang¹, Tim Curran³, Rajesh Nandy⁴, Mark Lowe⁵, and Dietmar Cordes^1,3
1Lou Ruvo Center for Brain Health, Cleveland Clinic, Las Vegas, NV, United States, ²Interdisciplinary neuroscience PhD program, University of Nevada, Las Vegas, Las Vegas, NV, United States, ³University of Colorado Boulder, Boulder, CO, United States, ⁴University of North Texas Health Science Center at Fort Worth, Fort Worth, TX, United States, ⁵Cleveland Clinic, Cleveland, OH, United States

To improve the accuracy of subject-level activation detections in noisy fMRI data, models to optimize voxel-wise smoothness levels for both isotropic Gaussian filter and spatially adaptive steerable filters are proposed. The smoothing step with currently optimized FWHM is incorporated into the optimization algorithm and solved efficiently using a sequential quadratic programming solver. Results from both simulated data and real episodic memory data indicate that a higher detection sensitivity for a fixed specificity can be achieved with the proposed method as compared to the widely used univariate general linear models with various levels of smoothness.

Michelle Karker^1,2, Douglas Noll^1,2,3, Benjamin M. Hampstead^4,5, and Scott Peltier^1,2
1Biomedical Engineering, University of Michigan, Ann Arbor, MI, United States, ²Functional MRI Laboratory, University of Michigan, Ann Arbor, MI, United States, ³Radiology, University of Michigan, Ann Arbor, MI, United States, ⁴Mental Health Service, VA Ann Arbor Healthcare System, Ann Arbor, MI, United States, ⁵Research Program on Cognition and Neuromodulation Based Interventions, Psychiatry, University of Michigan, Ann Arbor, MI, United States

Partial least squares regression with feature selection (PLS-BETA) was applied to task-based and resting-state connectivity data. Leveraging a MCI-relevant object-location or face-name task illuminates relationships with measures of total cognition (RBANS_total) and memory (RBANS_delayed). This provides support for the use of clinically-relevant tasks in cases where a “driven” connectivity network may elucidate pathological changes.

Ping Wang¹, Nicholas J. Sisco¹, Aimee Borazanci², and Richard D. Dortch^1
1Translational Neuroscience, Barrow Neurological Institute, Phoenix, AZ, United States, ²Neurology, Barrow Neurological Institute, Phoenix, AZ, United States

Quantitative magnetization transfer (QMT) imaging using the selective inversion recovery (SIR) approach has been successful in estimating fundamental tissue parameters, such as the PSR (macromolecular-to-free proton pool size ratio) and the R_1f (relaxation rate of the free pool). Although SIR allows one to perform the QMT experiment using conventional inversion recovery sequences, it is hampered by long scan times. Previously we have shown that Compressed SENSE (CS-SENSE) accelerates whole-brain SIR-QMT imaging, allowing whole-brain scanning within clinically practical scan times. In this study, we systematically investigate the effect of this acceleration on test-retest repeatability for PSR and R_1f.

Dipal Patel¹, Alexandru Badalan¹, Zaki Ahmed^1,2, and Ives R. Levesque^1,3
1Medical Physics Unit, McGill University, Montreal, QC, Canada, ²Mayo Clinic, Rochester, MN, United States, ³Research Institute of the McGill University Health Centre, Montreal, QC, Canada

Roman Fleysher¹, Lazar Fleysher², Namhee Kim³, Michael L Lipton¹, and Craig A Branch^1
1Gruss Magnetic Resonance Research Center, Department of Radiology, Albert Einstein Colledge of Medicine, Bronx, NY, United States, ²Biomedical Engineering and Imaging Institute, Department of Radiology, Mount Sinai Medical Center, New York, NY, United States, ³Department of Neurological Sciences, Rush Medical College, Chicago, IL, United States

Image interpolation is inextricable in many image analysis steps including registration of low resolution images to a standard high resolution template. Interpolated images are often smoothed and their voxel intensities are not statistically independent which severely complicates subsequent statistical analysis at the group level. Difficult to account correlations lead to higher than expected false-positive rates. We propose a new, noise-fill, image interpolation method which avoids both spatial blurring and loss of statistical independence of the voxel intensities. We show that noise-fill interpolation improves sensitivity of lesion detection in simulated patient-specific voxel-wise cluster analyses compared to other typically used interpolation methods.

Fatemeh Zabihollahy¹, Steven S. Raman¹, Pornphan Wibulpolprasert², Robert Reiter³, Holden Wu¹, and Kyung Hyun Sung^1
1Radiology, University of California, Los Angeles, Los Angeles, CA, United States, ²2Department of Diagnostic and Therapeutic Radiology, Ramathibodi Hospital, Bangkok, Thailand, ³Urology, University of California, Los Angeles, Los Angeles, CA, United States

^{Multiparametric magnetic resonance imaging (mpMRI) has a significant impact on prostate cancer (PCa) diagnosis. However, it is important to realize its accuracy for PCa detection. In this study, we compare mpMRI with whole-mount histopathology (WMHP) images as a reference to discover the limitation of mpMRI for PCa lesion localization. The results are presented as a spatial probability map, corresponding to the prostate sector map used in Prostate Imaging Reporting and Data System version 2.1 (PI-RADSv2.1), to highlight the regions on the prostate glands that require further attention from clinicians for a more accurate diagnosis of PCa and its treatment planning.}

Brett Levac¹ and Jonathan I Tamir^1,2,3
1Electrical and Computer Engineering, University of Texas, Austin, TX, United States, ²Oden Institute for Computational Engineering and Sciences, University of Texas, Austin, TX, United States, ³Department of Diagnostic Medicine, University of Texas, Austin, TX, United States

PROPELLER based acquisitions have the unique ability to give low resolution images for each echo train acquired and are often used for motion correction. However, motion correction with PROPELLER can often be hindered due to the low resolution nature of each shot. We propose a technique which leverages recent advancements in super resolution neural networks to enhance low resolution PROPELLER shots for better inter-shot motion estimation. 

Brian Nghiem^1,2, Zhe Wu¹, Melissa Haskell³, Lars Kasper¹, and Kamil Uludag^1,2
1BRAIN-To Lab, University Health Network, Toronto, ON, Canada, ²Medical Biophysics, University of Toronto, Toronto, ON, Canada, ³Electrical Engineering and Computer Science, University of Michigan, Ann Arbour, MI, United States

We investigated the performance of CNN-assisted joint estimation in two cases of severe motion corruption in a 2D slice of T2w FSE MRI. We showed that the inclusion of the CNN can help speed up convergence of the joint estimation algorithm, corroborating previous findings. We also showed one case in which joint estimation failed to converge to the correct image and motion parameters, with and without the CNN. A more exhaustive study is required to confirm whether deep learning can help joint estimation salvage otherwise unsalvageable corrupted data.

Nutandev Bikkamane Jayadev¹, Pierre-Gilles Henry¹, and Dinesh Deelchand^1
1University of Minnesota, MINNEAPOLIS, MN, United States

MR spectroscopy is prone to motion artifacts due to long scan times. Existing motion correction techniques require additional hardware or are inherently slow. We demonstrate a fast and accurate prospective motion correction method for MRS using spiral navigators.  A multi-slice to volume registration approach accelerates the motion parameter determination and achieves artifact free spectra at intended volume of interest. 

Sampada Bhave¹, Jennifer Wagner¹, and Hassan Haji-valizadeh^1
1Canon Medical Research USA, Mayfield Village, OH, United States

A modified COCOA framework was introduced to suppress motion artifacts. The data convolution operation in COCOA was extended to include k-space neighbors in the slice dimension along with those in the slice. Regularization in the data convolution operation was used as an alternative to the data combination in COCOA. The proposed technique provided high motion artifact suppression as compared to the standard COCOA synthesis.

Bramsh Qamar Chandio^1,2, Tamoghna Chattopadhyay³, Conor Owens-Walton³, Julio E. Villalon Reina³, Leila Nabulsi³, Sophia I. Thomopoulos³, Javier Guaje⁴, Eleftherios Garyfallidis⁴, and Paul M. Thompson^3
1Department of Intelligent SystemsEngineering, School of Informatics, Computing, and Engineering, Indiana University Bloomington, Bloomington, IN, United States, ²Imaging Genetics Center, Mark and Mary Stevens Neuroimaging and Informatics Institute, Keck School of Medicine, University of Southern California, Marina Del Rey, CA, United States, ³Imaging Genetics Center, Mark and Mary Stevens Neuroimaging and Informatics Institute, Keck School of Medicine, University of Southern California, Marina del Rey, CA, United States, ⁴Department of Intelligent Systems Engineering, School of Informatics, Computing, and Engineering, Indiana University Bloomington, Bloomington, IN, United States

We propose a dimensionality reduction method, based on the bundle-based minimum distance metric and the UMAP technique, to disentangle and visualize clusters in whole-brain tractography. In multishell diffusion MRI data from 141 elderly subjects, 30 tracts were extracted per subject using auto-calibrated RecoBundles in DIPY. A (141x30)x(141x30) bundle distance matrix was calculated and fed into UMAP. Embedding space maps showed that the same bundles were consistently mapped across subjects, making it easier to identify outliers and define clusters for population-based statistical analysis.

Georgy Guryev¹, Eugene Milshteyn^2,3, Ilias I Giannakopoulos^4,5, Elfar Adalsteinsson^1,6,7, Lawrence L. Wald^2,3,7, and Jacob K White^1
1Electrical Engineering and Computer Science, Massachusetts Institute of Technology, Cambridge, MA, United States, ²Department of Radiology, Massachusetts General Hospital and Harvard Medical School, Athinoula A. Martinos Center for Biomedical Imaging, Charlestown, MA, United States, ³Harvard Medical School, Boston, MA, United States, ⁴Center for Advanced Imaging Innovation and Research (CAI2R), Department of Radiology, New York University Grossman School of Medicine, New York, NY, United States, ⁵The Bernard and Irene Schwartz Center for Biomedical Imaging (CBI), Department of Radiology, New York University Grossman School of Medicine, New York, NY, United States, ⁶Institute for Medical Engineering and Science, Massachusetts Institute of Technology, Cambridge, MA, United States, ⁷Harvard-MIT Division of Health Sciences Technology, Cambridge, MA, United States

It was recently demonstrated that the combination of fast low-resolution tissue mapping and fast voxel-based field simulation can be used to perform a patient-specific MR safety check in minutes[4,5].However,field simulation required several of those minutes,making it too slow to perform the dozens of simulations that would be needed for patient-specific optimization.In this abstract we describe a compressed-perturbation-matrix technique that nearly eliminates the computational cost of including complex coils (or coils+shields) in voxel-based field simulation of tissue,thereby reducing simulation time from minutes to seconds.The approach is demonstrated on a wide variety of head+coil and head+coil+shield configurations,using the latest implementation of MARIE2.0.