Mark Sundman¹, Hing-Chiu Chang¹, Laurent Petit², Shayan Guhaniyogi¹, Christopher Petty¹, Allen Song¹, and Nan-kuei Chen^1
1Brain Imaging and Analysis Center, Duke University Medical Center, Durham, North Carolina, United States, ²Groupe d'Imagerie Neurofonctionnelle (GIN) - UMR5296, CNRS, CEA, Université de Bordeaux, Bordeaux, France

This research introduces a novel mothed for high-resolution diffusion tensor imaging (DTI) that is capable of achieving submillimeter isotropic resolution (0.85 x 0.85 x 0.85 mm3) in vivo for the human brain using a normal 3T clinical scanner. This high-resolution DTI enables both researchers and physicians to more accurately identify white matter structures and fiber tracts in vivo. To demonstrate its utility and potential, we identified structures at this submillimeter isotropic resolution that are indiscernible at conventional resolutions.

Thomas Witzel¹, Aapo Nummenmaa¹, Qiuyun Fan¹, Susie Yi Huang¹, and Lawrence Leroy Wald^1,2
1Department of Radiology, Harvard Medical School, A.A.Martinos Center for Biomedical Imaging, Massachusetts General Hospital, Charlestown, MA, United States,²Harvard-MIT Division of Health Sciences and Technology, Cambridge, MA, United States

Fractional anisotropy (FA) is sensitive to intravoxel orientation dispersion, which may lead to reduced or very low FA values in regions of strong fiber crossings. Recent new measures of microscopic anisotropy have been proposed that are not influenced by the macroscopic orientation dispersion in the voxel, including the rotationally invariant eccentricity metric as well as the qMAS derived uFA. We use the rotationally invariant eccentricity approach to generate a metric which is insensitive to macroscopic orientation dispersion and combine the data with a conventional q-ball acquisition to demonstrate this metric's independence from the underlying crossing structure.

Lars Müller¹, Andreas Wetscherek¹, and Frederik Bernd Laun^1
1Medical Physics in Radiology, German Cancer Research Center (DKFZ), Heidelberg, Germany

Double wave vector diffusion MRI gives the opportunity to measure microscopic anisotropy, what, however, requires to reliably acquire diffusion weighted images with many diffusion encoding directions. This in turn might lead to problems in determining trustworthy values because of eddy current induced image distortions. We present a sequence design that compensates for eddy currents, by using additional refocussing pulses and bipolar gradients. We show the effectivity of our sequence by determining the coefficient of variation of the signal intensity over 36 diffusion direction combinations. Additionally we present evidence that eddy current compensation is needed to determine reliable anisotropy values in-vivo.

Benoit Scherrer¹, Ali Gholipour¹, Onur Afacan¹, Sanjay P. Prabhu¹, and Simon K. Warfield^1
1Harvard Medical School, Boston Children's Hospital, Boston, Massachusetts, United States

We propose a novel non-Cartesian multi-shot DWI technique that achieves high resolution dense k-space sampling by acquisition of multiple anisotropically oversampled acquisitions (so-called "shots") and by reconstruction in the image space. Our approach leverages the coupling between DW images by simultaneously performing model-based HR reconstruction and estimation of a diffusion compartment imaging (DCI) tissue model that describes the local tissue microstructure at each voxel. Our combined approach also enables reconstruction from shots with different subsets of gradients, providing robustness to patient motion and potential for acceleration.

Steven Baete^1,2 and Fernando Emilio Boada^1,2
1Center for Advanced Imaging Innovation and Research (CAI2R), NYU School of Medicine, New York, NY, United States, ²Center for Biomedical Imaging, Dept. of Radiology, NYU School of Medicine, New York, NY, United States

The angular resolution of Diffusion Spectrum Imaging (DSI) is crucial in accurately determining the complex distributions of intravoxel fiber orientations in the human brain. The recently proposed Radial q-space sampling for DSI (RDSI) has been shown to have improved angular resolution over conventional Cartesian sampling. Further improvements in Orientation Distribution Function (ODF) sampling and angular resolution of DSI can be obtained by optimizing q-space sampling density in RDSI. This can be done using variable density in q-space to minimize the ODF variance, which does not require extra time when a RDSI multi-echo stimulated echo sequence is used.

William Perrault¹, Tanguy Duval¹, and Julien Cohen-Adad^1,2
1Polytechnique de Montreal, Montreal, Quebec, Canada, ²Functional Neuroimaging Unit, CRIUGM, University of Montreal, Montreal, Quebec, Canada

Two diffusion sequences designed to measure axonal restricted compartment were compared: Pulsed Gradient Spin Echo (PGSE) and Non-Uniform Oscillating Gradient Spin Echo (NOGSE). The validity of NOGSE analytical equations was assessed using Camino. A bootstrap procedure tested the robustness and accuracy of both sequences at different gradient strengths: 40, 80 and 300 mT/m. Results suggest that (i) the precision of axon diameter estimation increases with gradient strength for both techniques and that (ii) accuracy was higher for the NOGSE technique. The current trends pushing for stronger gradients on clinical scanners, NOGSE technique seems a good candidate for quantifying axon microstructure.

Victor B. Xie^1,2 and Ed X. Wu^1,2
1The University of Hong Kong, Laboratory of Biomedical Imaging and Signal Processing, Hong Kong SAR, China, ²The University of Hong Kong, Department of Electrical and Electronic Engineering, Hong Kong SAR, China

Diffusion weighted (DW) images acquired with EPI sequence are prone to Nyquist ghost and geometric distortion. In this work, we applied a new phase labeling scheme we recently proposed to simultaneously correct both artifacts in DWI images. The artifacts in resultant phantom images are largely corrected, demonstrating that the proposed method can effectively remove both Nyquist ghost and geometric distortion in the DW-EPI images through a phase lebeling acquisition and post-processing.

Hadrien Dyvorne¹, Rafael O'Halloran¹, and Priti Balchandani^1
1Radiology, Icahn School of Medicine at Mount Sinai, New York, NY, United States

We present a novel Diffusion-Weighted Matched-phase Spin Echo acquisition (DW-MASE) that uses semi-adiabatic radiofrequency pulses to provide increased robustness to B1 field inhomogeneity for ultrahigh field brain diffusion-weighted imaging. We show that DW-MASE leads to increased signal in regions suffering from inhomogeneous B1 when compared to conventional single-refocused diffusion. The impact of improved data quality on diffusion metrics is shown in a high resolution white matter tractography example acquired in a healthy volunteer.

Erpeng Dai¹, Zhe Zhang¹, Xiaodong Ma¹, Bida Zhang², Bin Xie², Chun Yuan^1,3, and Hua Guo^1
1Center for Biomedical Imaging Research, Department of Biomedical Engineering, School of Medicine, Tsinghua University, Beijing, China, ²Healthcare Department, Philips Research China, Shanghai, China, ³Department of Radiology, University of Washington, Seattle, WA, United States

Blipped CAIPI has been proposed for single-shot EPI multiband acquisition. However, single-shot EPI is susceptible to geometric distortions, which limit its application for high resolution diffusion weighted (DW) imaging. Multi-shot EPI may alleviate this with interleaved acquisition. Thus a generalized blipped CAIPI sequence and reconstruction framework was proposed for multi-shot DW EPI. The propsed method has achieved less distortions than blipped CAIPI single-shot EPI amd comparable FA maps with single band excitation. Thus this method will be valuable for fast high resolution DWI or DTI in the future.

Po Wei Cheng^1,2, Yung Hao Chuang^1,2, Yun An Huang^2,3, Edzer L. Wu², Tzi Dar Chiueh³, and Jyh Horng Chen^1,2
1Graduate Institute of Biomedical Engineering and Bioinformatics,National Taiwan University, Taipei, Taiwan, Taiwan, ²Interdisciplinary MRI/MRS Lab, Department of Electrical Engineering, National Taiwan University, Taipei, Taiwan, ³Department of Electrical Engineering, National Taiwan University, Taipei, Taiwan

Diffusion tensor imaging or diffusion weighted imaging was used to study the tissue orientation and the diffusivity properties. However, the long scan time for applying different diffusion gradient in diffusion weighted imaginglimit the clinical application in spine, especially high resolutionimaging and 3D modeling. In this study, we applied the accelerated technique single carrier Wideband MRI with W=5acceleration on the 3D rat spine diffusion study. The results demonstrated the capacity to acquire 3D DTI imaging on rat spine by using Wideband MRI technique, it also reduced the scan time from 22.5 hours to 4.5 hours. We are looking into higher acceleration rate and higher imaging resolution of rat spinal diffusion study, and implementing on the clinical study in the near future.

Steven Baete^1,2, Jose Raya², Florian Knoll^1,2, Gene Young Cho^2,3, Prodromos Parasoglou^1,2, Ryan Brown^1,2, Tobias Block^1,2, Ricardo Otazo^1,2, Jenny Bencardino⁴, and Eric Sigmund^1,2
1Center for Advanced Imaging Innovation and Research (CAI2R), NYU School of Medicine, New York, NY, United States, ²Center for Biomedical Imaging, Dept. of Radiology, NYU School of Medicine, New York, NY, United States, ³Sackler Institute of Graduate Biomedical Sciences, NYU School of Medicine, New York, New York, United States, ⁴Radiology, NYU School of Medicine, New York, NY, United States

The feasibility of dynamic diffusion tensor imaging is demonstrated both in a rotating gel phantom and in vivo using a novel Multiple Echo Diffusion Tensor Imaging (MEDITI) method. In MEDITI both diffusion and image encoding are compressed. Specifically, each of multiple echoes generated by five RF-pulses are encoded with different diffusion weighting using a highly efficient STAR k-space trajectory for each echo. Diffusion weighted images are reconstructed using a novel multidimensional compressed sensing approach. The resulting dynamic DTI can be used to study transient DTI changes, such as in skeletal muscle following exercise, where traditional DTI methods lack temporal resolution.

Ganesh Adluru¹, Bradley D. Bolster Jr², Robert Frost³, Lorie Richards⁴, and Edward V.R. DiBella^1
1Radiology, University of Utah, Salt Lake City, Utah, United States, ²Siemens Healthcare, Salt Lake City, Utah, United States, ³FMRIB Centre, Nuffield Department of Clinical Neurosciences, University of Oxford, Oxford, United Kingdom, ⁴Occupational Therapy, University of Utah, Salt Lake City, Utah, United States

Readout-Segmented EPI (RS-EPI) acquisition is a promising approach for high quality diffusion imaging. With its short echo spacing times compared to the standard single shot EPI sequence, RS-EPI has less blurring and distortions and allows high spatial resolution acquisitions. However with long diffusion preparation time for each segment, scan time increase is almost proportional to the number of segments making the RS-EPI technique less practical especially for diffusion acquisitions with a large number of diffusion directions. Here we present a framework to speed up RS-EPI by combining simultaneous multi-slice acquisitions with constrained reconstructions for k-space undersampling. We use a patch-based low rank reconstruction to remove undersampling artifacts.

Xia Zhao¹, Michael Langham¹, Cheng Li¹, and Hee Kwon Song^1
1Laboratory for Structural NMR Imaging, University of Pennsylvania, Philadelphia, PA, United States

In this work, the performance of diffusion-weighted DESS sequence with golden angle radial acquisition is compared with the standard DW-EPI sequence for imaging the brain. Results show that radial DW-DESS sequence can provide high resolution diffusion images without the problem of susceptibility-related artifacts and N/2 ghosting. It is also demonstrated that self-gating and phase correction techniques can be utilized to compensate for motion-induced phase inconsistencies among views and improve image quality.

Jian Cheng^1,2, Dinggang Shen³, Pew-Thian Yap³, and Peter Basser^1
1Section on Tissue Biophysics and Biomimetics (STBB), PPITS, NICHD, NIH, Bethesda, MD, United States, ²The Intramural Research Program (IRP), NIBIB, Bethesda, MD, United States, ³Department of Radiology and BRIC, The University of North Carolina at Chapel Hill, NC, United States

In this abstract, we propose two novel algorithms based on the spherical code concept, i.e., Iterative Maximum Overlap Construction (IMOC) to generate a sampling scheme from discretized sphere, and a constrained non-linear optimization (CNLO) method to update a given initial scheme on a continuous sphere. Compared with the state-of-the-art methods based on spherical code and electrostatic energy minimization, the sampling schemes by the proposed methods yield larger angular separations in shells and better rotational invariance.

Joseph L Holtrop^1,2 and Bradley P Sutton^1,2
1Bioengineering, Univerity of Illinois Champaign-Urbana, Urbana, Illinois, United States, ²Beckman Institute, University of Illinois at Urbana-Champaign, Urbana, Illinois, United States

Much work has gone into making diffusion weighted imaging faster and more SNR efficient. Methods to achieve this have included using multi-shot for reduced distortions, 3D excitations for improved SNR efficiency, and simultaneous multi-slice for higher acceleration factors. In this work we combine all of this techniques and introduce a pulse sequence with 3D navigator that enables efficient acquisition of data capable of combining all these techniques. The addition of a blipped spiral-in navigator combined with a spiral readout, allows for a time efficient way to collect diffusion data and produces high quality diffusion weighted images.

Ping-Huei Tsai^1,2, Hua-Shan Liu^2,3, Hsiao-Wen Chung⁴, Chia-Feng Lu², Fei-Ting Hsu², Li-Chun Hsieh², and Cheng-Yu Chen^1,2
1Department of Radiology, School of Medicine, College of Medicine, Taipei Medical University, Taipei, Taiwan, ²Imaging Research Center and Department of Medical Imaging, Taipei Medical University Hospital, Taipei Medical University, Taipei, Taiwan, ³Graduate Institute of Clinical Medicine, Taipei Medical University, Taipei, Taiwan, ⁴Graduate Institute of Biomedical Electronics and Bioinformatics, National Taiwan University, Taipei, Taiwan

q-space imaging (QSI) is an advanced diffusion-weighted imaging technique to explore the water displacement profile. The purpose of this study is to assess the effects of maximal b value and sampling interval on water displacement profile in simulation and validate it in clinical application of the brain. Our preliminary finding demonstrated that the proposed sampling pattern provides an alternative to obtain QSI images and reliable MD and ZDP measurements with a shorter acquisition time, which may be helpful in the implementation of in vivo brain q-space imaging in clinical application.

Jacques-Donald Tournier¹, Emer Hughes¹, Nora Tusor¹, A. David Edwards¹, and Joseph V Hajnal^1
1Centre for the Developing Brain, Kings College London, London, London, United Kingdom

High angular resolution diffusion imaging (HARDI) sequences are now starting to be used for neonatal investigations. However, the neonatal brain is very different from the adult brain, and sequence parameters should be optimised specifically for use in this cohort. A recent study used a data-driven approach to optimise HARDI parameters for use in the adult brain, based on the angular frequency content of the DW signal. In this study, we use this approach to investigate the optimal imaging parameters for use in neonates.

Eric Aliotta^1,2 and Daniel B. Ennis^1,2
1Radiological Sciences, UCLA, Los Angeles, CA, United States, ²Biomedical Physics IDP, UCLA, Los Angeles, CA, United States

T2 and ADC maps were jointly reconstructed from signals with varying TEs and b-values. Joint T2 and ADC maps were compared to independent reference maps in the brain in five healthy volunteers as well as in a polyethylene glycol (PEG) and gadolinium T2-ADC phantom. Comparisons showed no significant difference between the T2 or ADC maps in the healthy subjects or the PEG phantom. Joint acquisition and estimation of T2 and ADC maps is thus feasible in the brain and shows no loss of accuracy compared to reference T2 and ADC mapping sequences that require twice the scan time.

Jacques-Donald Tournier^1,2, Emer Hughes^1,3, Nora Tusor^1,3, Stamatios N. Sotiropoulos⁴, Saad Jbabdi⁴, Jesper Andersson⁴, Daniel Rueckert⁵, A. David Edwards^1,3, and Joseph V Hajnal^1,2
1Centre for the Developing Brain, Kings College London, London, London, United Kingdom, ²Department of Biomedical Engineering, Kings College London, London, London, United Kingdom, ³Department of Perinatal Imaging & Health, Kings College London, London, London, United Kingdom, ⁴FMRIB Centre, University of Oxford, Oxford, United Kingdom, ⁵Department of Computing, Imperial College London, London, United Kingdom

A number of recently proposed methods make use of data acquired using multi-shell HARDI, characterised by the number of b-values used, their actual values, and the number of DW directions acquired per b-value shell. To date, these schemes have been optimised with respect to a particular reconstruction algorithm, with no guarantee of suitability for other methods. In this study, we present a data-driven approach to optimise these protocols, and apply it to design a multi-shell scheme suitable for use in neonatal imaging.

Pavan Poojar¹, Bikkemane Jayadev Nutandev¹, Arush Honnedevasthana Arun¹, Antharikshanagar Bellappa Sachin Anchan¹, Ramesh Venkatesan², and Sairam Geethanath^1
1Dayananda Sagar Institutions, Bangalore, karnataka, India, ²Wipro-GE Healthcare, karnataka, India

Diffusion weighted magnetic resonance imaging data with high isotropic resolution can be obtained at 1.5T. However, extensive brain coverage takes longer acquisition times. Most significant part of the brain covers the center of the k-space forming an arbitrary shape. Proposed method demonstrates optimal gradient waveform design for such arbitrary k-space trajectory using active contour technique for six data sets. Result shows Apparent Diffusion Coefficient (ADC) maps computed for different acceleration factors. Current and future work involves optimizing gradient waveforms for different anatomies and integrating this work to the scanner.

Yu-Chien Wu^1
1Radiology and Imaging Sciences, Indiana University School of Medicine, Indianapolis, IN, United States

A new sequence, Rotating Short-Axis (RSA) echo-planar imaging (EPI), may address some issues associated with single-shot spin-echo EPI sequences for diffusion MRI. RSA has a higher signal-to-noise ratio due to the shorter echo time, minimal geometric distortion from shorter echo spacing, and a faster imaging speed due to partial k-space sampling. RSA works better in high spatial resolution diffusion MRI. Here we performed computer simulations to quantitatively evaluate blade sampling and composite reconstruction in RSA methods with respect to effective image resolution, artifacts, and directional estimates of fiber orientations. The results were compared with pseudo gold standard, single-shot spin-echo EPI.

Suguru Yokosawa¹, Hisaaki Ochi¹, and Yoshitaka Bito^2
1Central Research Laboratory, Hitachi, Ltd., Kokubunji-shi, Tokyo, Japan, ²Hitachi Medical Corporation, Kashiwa, Chiba, Japan

In this work, we proposed that retrospective motion correction of diffusion-weighted images by using the optimum order for measuring diffusion directions is proposed. As for this method, retrospective motion correction is performed by using intermediate mean-diffusivity (MD) maps generated from respective groups of six consecutive DW images. The proposed method can reduce calculation error caused by subject motion in diffusion MRI without the need for additional scan time.

Zhuo Shi¹, Xinming Zhao¹, Han Ouyang¹, and Lizhi Xie^2
1Department Of Imaging Diagnosis,Cancer Hospital, Chinese Academy of Medical Sciences & Peking Union, Beijing, Beijing, China, ²GE Healthcare China, Beijing, China

The focus of this study, is to perform quantitative analysis including signal-to-noise ratio (SNR), apparent diffusion coefficient (ADC), qualitative artifacts severity, subjective image quality of the breath-hold (BH), respiratory-triggered (RT) and free-breathing (FB) DWI between the 1.5T and 3T techniques. To determine the best DWI acquisitions for upper abdomen, and demonstrate it as a reference for the clinical work in the future.

Andreas Wetscherek¹ and Frederik Bernd Laun^1
1Medical Physics in Radiology, German Cancer Research Center (DKFZ), Heidelberg, Germany

Shota Ishida¹, Tosiaki Miyati¹, Naoki Ohno¹, Tomohiro Chigusa², Hikari Usui³, Masaaki Hattori⁴, Yuki Hiramatsu⁴, Satoshi Kobayashi⁵, and Toshifumi Gabata^5
1Division of Health sciences, Graduate School of Medical Sciences, Kanazawa University, Kanazawa, Ishikawa, Japan, ²Okazaki City Hospital, Okazaki, Aichi, Japan,³Yokohama City University Hospital, Yokohama, Kanagawa, Japan, ⁴School of Health sciences, College of Medical, Pharmaceutical and Health sciences, Kanazawa University, Kanazawa, Ishikawa, Japan, ⁵Department of Radiology, Kanazawa University School of Medicine, Kanazawa, Ishikawa, Japan

To clarify relations and mechanisms of perfusion and water fluctuation, we assessed the influence of blood flow on ADC change in cardiac cycle (delta-ADC) using original cranial MRI-phantom. With use of low b-values, both maximum ADC and minimum ADC during the pulsation period increased with the simulated total cerebral blood flow. However, with high b-values, minimum ADC did not increase with the simulated total cerebral blood flow. Therefore, water fluctuation was predominant with high b-values. Delta-ADC in combination with high b-values makes it possible to minimize the direct effect of blood flow.

Tan Guo¹, Dandan Zheng², Min Chen¹, and Juan Chen^1
1Department of Radiology, Beijing Hospital, Beijing, Beijing, China, ²GE Healthcare, China, Beijing, China

Meniscal tears are an indication for arthroscopic knee surgery. The tears in the vascular portion of the meniscus is called ¡°red zone¡± and tears in the avascular portion is defined ¡°white zone¡±. Although the demarcation of the zones can be defined in anatomy, it is hard to gain a boundary through conventional MR sequence. Intravoxel incoherent motion theory provides information about microcirculation of blood in addition to the pure molecular diffusion. In this study, rFOV DWI technology and IVIM model were used to estimate the vasculature of the red and white zone of meniscus by using the IVIM perfusion parameters.

Kaining Shi¹, He Wang², Guang Cao³, Ying Qi⁴, and Xiaoming Wang^4
1Imaging Systems Clinical Science, Philips Healthcare (China), Beijing, China, ²Philips Research (China), Shanghai, China, ³Imaging Systems Clinical Science, Philips Healthcare (China), Hongkong, China, ⁴Radiology Department, Shengjing Hospital of China Medical University, Shenyang, Liaoning, China

The nonlinear bi-exponential curve-fitting in the Intravoxel Incoherent Motion (IVIM) model is sensitive to noise and time-consuming. In this work, fussy clustering technique is used to improve the reliability of curve-fitting and reduce the total calculation time.16 b-values DWI data of 2 PRES patients and 2 volunteers was processed by the fussy clustering analysis method. The new algorithm achieved brain segmentation successfully and generated similar parameters as the pixel-by-pixel approach, with 1.3-3.3% time cost and 11.4~79.0% curve-fit residual.

Alexander D. Cohen¹, Kimberly R. Pechman¹, Mona Al-Gizawiy¹, and Kathleen M. Schmainda^1,2
1Radiology, Medical College of Wisconsin, Milwaukee, WI, United States, ²Biophysics, Medical College of Wisconsin, Milwaukee, WI, United States

The DWI signal deviates from monoexponential behavior at high b-values in tissue. In this study, two techniques were used to fit a biexponential model to this signal: a non-linear least squares approach and a Bayesian approach. Biexponential DWI parameters were compared between fitting techniques and between tumor and normal tissue in a rat U87 brain tumor model. Bayesian modeling proved superior for differentiating tumor from GM. This technique also resulted in qualitatively better-looking maps with enhanced tumor to gray matter contrast compared to the traditionally used least-squares approach. There were also statistically significant differences between modeling techniques for several parameters.

Alessandra Caporale^1,2, Marco Palombo^2,3, and Silvia Capuani^2,4
1Physics Department, University 'Sapienza', Rome, ITALY, Italy, ²Physics Department, CNR-IPCF Roma Sapienza University of Rome, Rome, ITALY, Italy,³CEA/DSV/12BM/MIRCen, Fontenay-aux-Roses, FRANCE, France, ⁴Center for Life NanoScience@LaSapienza,Istituto Italiano di Tecnologia, Rome, ITALY, Italy

Anomalous diffusion (AD) stretched-exponential -imaging model, was applied to investigate a fixed mouse spinal cord by using gradient strength varying PGSTE sequence at 9.4T. To highlight the new additional information provided by AD approach, we compared AD images with results obtained by using conventional DTI, relaxometry and histological data. Mean  (M), -anisotropy (A), par and ortho are able to detect microstructural information of white matter (WM) in spinal cord more specific and complementary to those provided by DTI. Specifically, the correlation between A and myelin fraction is in agreement with the anisotropic rearrangement of myelin along WM fibers.

Yang Fan¹, Bing Wu², and Jia-Hong Gao^1
1Center for MRI Research, Peking University, Beijing, Beijing, China, ²GE Healthcare, Beijing, Beijing, China

A statistically stationary anomalous diffusion model has been proposed in this study. It offers a theoretical foundation to stretched exponential diffusion model. Its validity was proved through numerical simulation as well as in-vivo datasets. This model may aid further understanding of the anomalous diffusion process in neural tissues.

Allen Q. Ye¹, Rodolfo Gatto¹, Luis Colon-Perez², Thomas Mareci², Gerardo Morfini¹, and Richard Magin^1
1University of Illinois at Chicago, Chicago, IL, United States, ²University of Florida, Gainesville, FL, United States

By using continuous time random walk (CTRW) assumptions for diffusion, we have increased sensitivity to micro-architectural heterogeneity and tortuosity through fractional order parameters α and β. We explore the utility of α to distinguish brain tissue alterations in R6/2 mice, a well-characterized Huntington’s disease (HD) model. To this end, we imaged 21 day and 60 day post-natal mice and found changes in α in the corpus callosum. Additional microscopy and histology gives us clues as to the reason behind the change in α. This work is a step toward using fractional order modeling to stage progression of HD.

Wen-Chau Wu^1,2
1National Taiwan University, Taipei, Taiwan, ²Graduate Institute of Biomedical Electronics and Bioinformatics, National Taiwan University, Taipei, Taiwan

Diffusion kurtosis (DK) imaging is an extended diffusion MR imaging technique that permits evaluation of diffusion heterogeneity, a common characteristic of biological tissue that causes non-Gaussian diffusion. Jensen and coworkers recently proposed a fast DK imaging method based on 3-b-value acquisition and closed-form expressions. Although time efficient, the method may be prone to extreme variability of a single data point (as compared with to relatively time-consuming fitting over multiple data points). In this study, we investigated the variability of the indexes derived with fast DK imaging in the context of signal-to-noise ratio. The sensitivity of the indexes was compared with the corresponding indexes derived from the stretched exponential model imaging, another extended diffusion imaging technique.

Keith Hulsey¹, Matthew Lewis¹, Yin Xi¹, Qing Yuan¹, and Robert Lenkinski^1
1Radiology, The University of Texas Southwestern Medical Center, Dallas, TX, United States

This study investigates the accuracy and precision of diffusion parameters derived from fitting MR diffusion weighted data. A five parameter diffusion model was fit to simulated data using a non-linear least squares fit and through maximum likelihood estimators to compare the results obtained using two fitting methods. The repeatability of diffusion estimates for data collected from a phantom which incorporates diffusion and flow was investigated and compared to the precision of diffusion values calculated for the simulations.

Kouhei Kamiya¹, Yuichi Suzuki², Shota Tanaka³, Akitake Mukasa³, Masaaki Hori⁴, Harushi Mori¹, Akira Kunimatsu¹, Nobuhito Saito³, Shigeki Aoki⁴, and Kuni Ohtomo^1
1Department of Radiology, The University of Tokyo, Bunkyo, Tokyo, Japan, ²Department of Radiological Technology, The University of Tokyo Hospital, Bunkyo, Tokyo, Japan, ³Department of Neurosurgery, The University of Tokyo, Bunkyo, Tokyo, Japan, ⁴Department of Radiology, Juntendo University School of Medicine, Bunkyo, Tokyo, Japan

Prior studies have applied DTI to distinguish glioblastoma infiltration from vasogenic edema. This study investigated the utility of neurite orientation dispersion and density imaging (NODDI) within the peritumoral region of glioblastoma, by comparing with vasogenic edema in meningioma and metastasis. Voxel-by-voxel scattered plot demonstrated that peritumoral region of glioblastoma can be discriminated from vasogenic edema by combination of the diffusion metrics. Though ADC and FA were still more important discriminating factors, addition of NODDI parameters seemed to improve the specificity for tumor infiltration. Our results suggest that NODDI parameters can provide additional information to DTI, and may hopefully differentiate the highly-infilatrated area from less- or non-infiltrated area in the future.

Birgitte Fuglsang Kjølby¹, Steen Jakobsen², Jonas Brorson Jensen², Lea Hougaard Pedersen³, Louise M Rydtoft¹, Sune N Jespersen^1,4, and Brian Hansen^1
1CFIN, Aarhus University Hospital, Aarhus, Denmark, ²Nuclear Medicine and PET Center, Aarhus University Hospital, Aarhus, Denmark, ³Research Lab. for Biochemical Patology, Aarhus University Hospital, Aarhus, Denmark, ⁴Dept. of Physics and Astronomy, Aarhus University, Aarhus, Denmark

Magnetic Resonance (MR) Diffusion Kurtosis Imaging (DKI) is sensitive to tissue microstructure but routine clinical utilization has been held back by time consuming acquisition and postprocessing. Recently, a method for fast estimation of mean diffusivity, fractional anisotropy, mean kurtosis and fractional kurtosis anisotropy was proposed. Here we investigate the technique's ability to differentiate normal and diseased tissue using fixed kidneys from a GMO model of kidney fibrosis. We conclude that mean kurtosis and fractional kurtosis anisotropy are candidates of detecting fibrotic kidney tissue while mean diffusivity and fractional anisotropy are insensitive to differences between wild type and GMO kidneys.

Yasuhiko Tachibana^1,2, Takayuki Obata¹, Hiroki Tsuchiya¹, Tokuhiko Omatsu¹, Riwa Kishimoto¹, Koji Kamagata³, Masaaki Hori³, Shigeki Aoki³, and Tomio Inoue^2
1Research Center of Charged Particle Therapy, National Institute of Radiological Sciences, Chiba, Chiba, Japan, ²Department of Radiology, Yokohama City University, Yokohama, Kanagawa, Japan, ³Department of Radiology, Juntendo University, Tokyo, Japan

To compensate for the problem of diffusion kurtosis imaging (DKI) (i.e., requirement of large data, large noise generated by post-processing), a robust method based on diffusion-tensor imaging was designed to estimate diffusional kurtosis parallel and perpendicular to neuronal fibers. Map quality was improved by this method even with encoding directions reduced to six. The map was also accurate because its root-mean-square-error compared to the standard map (conventional calculation with 64 encoding directions) was smaller than by the conventional method, with differences being significant except one case.

Chris J Conklin^1,2, Devon M Middleton^2,3, Jürgen Finsterbusch⁴, Mahdi Alizadeh^2,3, Scott H Faro^2,3, Pallav Shah², Laura Krisa^5,6, Rebecca Sinko⁶, Joan Z Delalic¹, MJ Mulcahey⁶, and Feroze B Mohamed^2,3
1Electrical Engineering, Temple University, Philadelphia, PA, United States, ²Radiology, Temple University, Philadelphia, PA, United States, ³Bioengineering, Temple University, Philadelphia, PA, United States, ⁴Systems Neuroscience, University Medical Center Hamburg-Eppendorf, Hamburg, Germany, ⁵Physical Therapy, Thomas Jefferson University, Philadelphia, PA, United States, ⁶Occupational Therapy, Thomas Jefferson University, Philadelphia, PA, United States

This work investigates Diffusion Kurtosis Imaging (DKI) of the pediatric spinal cord using an inner Field of View (iFOV) sequence. DKI offers the potential to probe the underlying tissue microstructure by looking at the non-Gaussianity of water displacement. Both normal and Spinal Cord Injured (SCI) patients between the ages of 6 and 16 were acquired and analyzed. FA values were comparable to that shown in current literature and statistically significant differences were found in kurtosis metrics between groups. The results are encouraging and warrant additional study with a larger population.

Weiwei Wang¹, Rui Hu¹, Ziheng Zhang², Qingwei Song¹, Ailian Liu¹, and Yanwei Miao^1
1Radiology Department, the First Affiliated Hospital of Dalian Medical University, Dalian, Liaoning, China, ²GE Healthcare China, Beijing, China

DTI can investigate the WM microstructural abnormalities in AD, while not for GM due to the microstructural abnormalities. Diffusional kurtosis imaging (DKI) enables the precise quantification of the diffusional kurtosis, a measure of non-Gaussian diffusion, in GM. From the study, we investigated the diffusion complexity of gray nucleus in AD using DKI and found that the DKI parameters between AD and healthy controls showed significant difference. The changes of DKI parameters demonstrate this feature and show correlations with the mental state of AD. It is useful for evaluating the structure change of gray nucleus in AD patients.

Arash Latifoltojar¹, Margaret Hall-Craggs², Alan Bainbridge², Stuart Taylor³, Nikos Dikaios³, Kwee Yong³, Neil Rabin², and Shonit Punwani^3
1University College London, London, London, United Kingdom, ²University College London Hospital, London, United Kingdom, ³University College London, London, United Kingdom

Whole body diffusion weighted imaging (WBDWI) has been used for initial assessment and monitoring treatment response of multiple myeloma patients. In the standard diffusion model, the displacement of freely mobile water molecules is considered to have Gaussian distribution. However, water diffusion in tissue is more complex and several diffusion models (non-Gaussian diffusion models) have been proposed to account for this behavior and provide a more comprehensive analysis of DWI quantitative derived parameters. In this study, we investigated the application of different diffusion models for assessment of WBDWI quantitative parameters in patients with multiple myeloma.

Jordan Kovar¹, Rao Gullapalli², and Jiachen Zhuo^2
1Physics & Mathematical Sciences, Worcester Polytechnic Institute, Worcester, Massachusetts, United States, ²Diagnostic Radiology and Nuclear Medicine, University of Maryland School of Medicine, Baltimore, Maryland, United States

Biophysical modeling of Diffusion Kurtosis Imaging (DKI) data can provide direct white matter characterization for intra- and extra-axonal water diffusion properties. Human and animal studies have indicated that these parameters are sensitive to specific axonal structure changes, such as demyelination and axonal loss. However how exactly the underlying tissue property affects these WM parameters are little known. In this study is we used Monte Carlo simulation to investigate the effect of a few biophysical property of the axonal structure (e.g. axonal packing density, axon radius and axonal permeability) to the DKI WM parameters.

Edward S. Hui¹, G. Russell Glenn², Joseph A. Helpern³, and Jens H. Jensen^4
1Diagnostic Radiology, The University of Hong Kong, Pokfulam, Hong Kong, ²Neurosciences & Center for Biomedical Imaging, Medical University of South Carolina, Charleston, SC, United States, ³Radiology, Neurosciences & Center for Biomedical Imaging, Medical University of South Carolina, Charleston, SC, United States,⁴Radiology & Center for Biomedical Imaging, Medical University of South Carolina, Charleston, SC, United States

A tissue modeling framework, referred to as kurtosis analysis of neural diffusion organization (KANDO), is proposed to help provide a biophysical interpretation of the diffusion metrics estimated with diffusional kurtosis imaging (DKI). The framework supports a variety of multiple Gaussian compartment models. The model parameters are determined by minimizing the square of the Frobenius norm of the difference between the model and measured kurtosis tensors. KANDO differs from most other tissue modeling methods for diffusion MRI in that it only utilizes the information contained in the diffusion and kurtosis tensors, which can both be estimated with DKI.

Edward S. Hui¹ and Jens H. Jensen^2,3
1Department of Diagnostic Radiology, The University of Hong Kong, Pokfulam, Hong Kong, China, ²Department of Radiology and Radiological Science, Medical University of South Carolina, Charleston, South Carolina, United States, ³Center for Biomedical Imaging, Medical University of South Carolina, Charleston, South Carolina, United States

We have recently extended conventional single-pulsed-field-gradient (s-PFG) DKI to double-PFG (d-PFG) diffusion MRI sequence, known as double-pulsed DKI (DP-DKI). Owing to the fact that DKI isolates the second order contributions to the d-PFG signal, the 6D diffusional kurtosis encodes unique information beyond that available from s-PFG sequences. This study demonstrates the feasibility of in vivo human DP-DKI at 3 T.

Rachael LeeAnn Deardorff¹, Emilie T McKinnon¹, Tara Eckenrode Sokolowski¹, Jens H Jensen¹, Masaaki Hori², Varan Govind³, and Joseph A Helpern^1
1Department of Radiology & Radiological Science, Medical University of South Carolina, Charleston, South Carolina, United States, ²Department of Radiology, Juntendo University School of Medicine, Tokyo, Japan, ³Miller School of Medicine, University of Miami, Miami, Florida, United States

Kurtosis Imaging Network (KIN) will create an open source database for normal healthy controls to establish a standard range of kurtosis values within each population. This database will also allow for quantitative comparisons between sites, vendors, and various protocol parameters. Finally, KIN will also help develop a strong collaborative research network for troubleshooting and creating future studies. We have initially analyzed data from 5 selected sites and although protocol parameters and vendors differed between groups, the diffusion metrics in the body of the corpus callosum for healthy controls were not significantly different.

Xingju Nie¹, Dorothea Rosenberger², Aurelie Ledreux³, Ann-Charlotte Granholm³, Heather Boger³, and Maria Falangola^1,3
1Radiology and Center for Biomedical Imaging, Medical University of South Carolina, Charleston, South Carolina, United States, ²Anesthesiology, University of Utah, Utah, United States, ³Neuroscience, Medical University of South Carolina, Charleston, South Carolina, United States

Benzodiazepines (BZD) are widely prescribed among older adults, often for anxiety, depression and insomnia. Midazolam (MDZ) is the most commonly used BZD premedication for sedation and in the intensive care unit. However, the mechanisms of a possible MDZ neuroprotection or neurotoxicity effects on brain microenvironment are not fully understood. This study investigates if MDZ administration in rodents causes changes in the cerebral microenvironment as defined by diffusional kurtosis imaging (DKI). We detected DK metrics increase reflecting changes in the cerebral microenvironment of the cortex, striatum, thalamus and hippocampus of rats exposed to MDZ, which may be related to mitochondrial abnormalities.

Miyuki Takasu¹, Koichi Oshio², Yuji Akiyama¹, Ryuji Akita¹, Kazushi Yokomachi¹, Yoko Kaichi¹, Shuji Date¹, and Kazuo Awai^1
1Diagnostic Radiology, Hiroshima University Hospital, Hiroshima, Hiroshima, Japan, ²Department of Diagnostic Radiology, Keio University, Tokyo, Japan

Non-Gaussian diffusion methods permitting the analysis of the diffusion-weighted signal over a larger range of b-values have gained an increasing importance in tissue characterization. Our purpose was to investigate the applicability and the performance of gamma distribution model in differentiating vertebral lesions in human subjects. For each of normal spinal bone marrow and lesions, ƒÆ, ƒÈ, the area fraction of D < 1.0mm2/s (frac<1), the area fraction of D > 3.0 mm2/s (frac>3), PG (D) and K was measured. ƒÈ, frac <1, PG (D), and K proved to be useful for differentiation of malignant lesions from benign lesions.

Zaiyi Liu¹, Xin Chen², Zelan Ma¹, and Zhongping Zhang^3
1Radiology, Guangdong General Hospital, Guangdong Academy of Medical Sciences, Guangzhou, Guangdong, China, ²Radiology, Guangzhou First People's Hospital, Guangzhou Medical College, Guangzhou, Guangdong, China, ³GE Healthcare China, Beijing, Beijing, China

Non-Gaussian diffusion models DKI and stretched exponential was used to evaluate the micro-structural change in the ultra-early phase of Sorafenib administration.

Pedro A. Gómez^1,2, Tim Sprenger^1,2, Marion I. Menzel², and Jonathan I. Sperl^2
1Technical University Munich, Munich, Germany, ²GE Global Research, Munich, Germany

Estimating the diffusional kurtosis tensor requires fitting a model with 22 free parameters to noisy diffusion signals, and is subject to low accuracy. We propose a variation of the model that makes use of the main directions of diffusion, only requiring the fitting of 10 parameters. Monte Carlo simulations and experiments on volunteer datasets indicate that the reduced version of the model has less bias than the full model, particularly in white matter areas with high fractional anisotropy.

Carson Ingo¹, Yu Fen Chen², Todd B. Parrish², Andrew G. Webb¹, and Itamar Ronen^1
1C.J. Gorter Center for High Field MRI, Department of Radiology, Leiden University Medical Center, Leiden, Netherlands, ²Department of Radiology, Northwestern University, Chicago, IL, United States

Tim Schakel¹, Bjorn Stemkens¹, Hans Hoogduin², and Marielle Philippens^1
1Radiotherapy, UMC Utrecht, Utrecht, Netherlands, ²Radiology, UMC Utrecht, Utrecht, Netherlands

In this work, we integrated a multi-acquisition Dixon method into a DW-SPLICE sequence to facilitate water-fat separation. The results of the water-fat separation were used to achieve good fat suppression in an undistorted DW image in difficult areas, such as the head and neck region.

Saad Jbabdi¹, Sean Foxley¹, and Karla L Miller^1
1FMRIB Centre, University of Oxford, Oxford, Oxfordshire, United Kingdom

Diffusion imaging with SSFP offers great temporal CNR, and is particularly suited for post-mortem whole brain imaging. We have previously shown that using multiple flip angles helps improve CNR homogeneity across the brain. Here, we use a combination of analytic and Monte Carlo simulations, coupled with real post-mortem human brain data, to demonstrate how we can model multiple flip angle DW-SSFP with a simple Gamma distribution of ADCs.

Tokunori Kimura¹, Naotaka Sakashita¹, and Yutaka Machii^2
1Clinical Application Research and Development Dept., Toshiba Medical Systems corp., Otawara, Tochigi, Japan, ²MRI development dept., Toshiba Medical Systems corp., Otawara, Tochigi, Japan

The purpose of this study was to propose a short-TE computed diffusion imaging (cDWI) technique for enhancing CNR for short T2 and low ADC tissues to backgrounds. Here we assessed for simulation, phantom and volunteer brain. The CNRs for the fiber portions-to-background were enhanced and T2 shine-thorough effects in water or CSF were reduced in short-TE cDWI images. In conclusion, the short-TE cDWI technique is useful for suppressing tissues of high ADC with longer T2 and enhancing tissues of low ADC with short T2; and thus regarded as clinically useful for enhancing neuronal fibers (neurography) or short T2 tumors.

Julian Emmerich¹, Lars Müller¹, Andreas Wetscherek¹, and Frederik Bernd Laun^1
1Medical Physics in Radiology, German Cancer Research Center (DKFZ), Heidelberg, Germany

Due to the strong gradients used for diffusion weighting, vibrations still play an important role in measuring the apparent diffusion coefficient (ADC). The effect of vibrations on ADC measurements was investigated in an apparent exchange rate measurement with two diffusion weightings. A wooden frame was built that enables us to place the phantoms (suspensions of yeast in water) inside the scanner without direct physical contact. The standard deviation over six gradient directions is approximately reduced by a factor one half, while the ADC, calculated in a ROI, remains nearly unchanged. ADC maps are more homogenous using the wooden frame.

Rafael O'Halloran¹, Chen Yang¹, and Junqian Xu^1
1Radiology, Icahn School of Medicine at Mt Sinai, New York, NY, United States

In Simultaneous multi-slice (SMS) middle slices are inevitably acquired at the beginning of each diffusion-encoded volume, making SMS more sensitive to artifacts due to transient eddy currents than conventional diffusion-weighted MRI (dMRI). In conventional dMRI slices affected by transient eddy currents are often at the edge of the slice group and can be safely ignored. In SMS, however, the majority of the simultaneously excited slices are in the middle of the slice group and transient eddy currents cause the slices in the initial a few excitations to have different geometric distortions than the slices in the later excitations, when eddy currents have achieved steady state. In this work the artifact is explained and demonstrated in scans of a phantom and human brain. A simple post-processing correction is proposed and demonstrated to be effective in correcting the artifact. The importance of realizing and addressing this artifact is heightened by the number of studies using SMS dMRI.

Katharina Paul¹, Andreas Graessl¹, Jan Rieger^1,2, Darius Lysiak^1,2, Till Huelnhagen¹, Lukas Winter¹, Robin Heidemann³, Tobias Lindner⁴, Stefan Hadlich⁵, Paul-Christian Krueger⁵, Soenke Langner⁵, Oliver Stachs^4,6, and Thoralf Niendorf^1,7
1Max-Delbrueck Centre for Molecular Medicine, Berlin Ultrahigh Field Facility (B.U.F.F.), Berlin, Berlin, Germany, ²MRI.TOOLS GmbH, Berlin, Germany, ³Siemens Healthcare Sector, Erlangen, Germany, ⁴University Medicine Rostock, Pre-clinical Imaging Research Group, Rostock, Germany, ⁵University of Greifswald, Institute for Diagnotic Radiology and Neuroradiology, Greifswald, Germany, ⁶University Medicine Rostock, Department of Ophthalmology, Rostock, Germany, ⁷Experimental and Clinical Research Center, a joint cooperation between the Charité Medical Faculty and the Max-Delbrueck-Center, Berlin, Germany

Diffusion-weighted imaging (DWI) of the orbit is an emerging MRI application to provide guidance during diagnostic assessment and treatment of ophthalmological diseases. The standard approach for DWI is EPI. However, EPI is prone to severe geometric distortions, especially at high and ultrahigh magnetic fields. Realizing these constraints and the potential of ocular DWI, this work uses diffusion sensitized multi-shot RARE for ophthalmic MRI. A comparison in terms of geometric distortions of ms-RARE with two EPI variants is performed for brain and ocular imaging at 3.0 T and at 7.0 T.

Longchuan Li^1,2, Jaekeun Park², Yuguang Meng³, Todd Preuss⁴, Xiaodong Zhang³, and Xiaoping Hu^2
1Department of Pediatrics, Marcus Autism Center, Children's Healthcare of Atlanta, Emory University, Atlanta, GA, United States, ²Biomedical Imaging Technology Center, School of Medicine, Emory University, Atlanta, GA, United States, ³Yerkes Imaging Center, Yerkes National Primate Research Center, Emory University, GA, United States, ⁴Division of Neuropharmacology and Neurologic Diseases, Emory University, GA, United States

2D or 3D standard spin-echo diffusion sequences have been commonly used for collecting diffusion MRI data on postmortem monkey brains. Because of its low sampling efficiency, many parameters including diffusion directions, spatial resolution and diffusion weighting need to be compromised, limiting its application in brain connectivity studies. Here, we developed a pipeline that can acquire dMRI data on ex-vivo monkey brains with high angular resolution (N=128), high spatial resolution (0.5mm isotropic), and high diffusion weightings (b=2000-4000) within 40 hours. Such framework may bridge the rich connectivity information in monkey tracer studies and that in humans via diffusion tractography.

Shayan Guhaniyogi¹, Mei-Lan Chu¹, Hing-Chiu Chang¹, Allen Song¹, and Nan-Kuei Chen^1
1Brain Imaging and Analysis Center, Duke University, Durham, NC, United States

Despite the advantages of multishot EPI over singleshot acquisitions in diffusion tensor imaging, its primary drawback is increased sensitivity to patient motion. While phase errors and pixel misregistrations among shots are commonly addressed in multishot motion correction schemes, the altered diffusion-encoding of each shot due to motion is often neglected. We therefore present a reconstruction technique which corrects all three motion-induced errors in multishot diffusion EPI. The technique is shown to improve both image quality and tensor calculations, and is expected to be valuable for clinical and neuroscience applications requiring accurate high resolution diffusion tensor information.

Wenchuan Wu¹, Peter Koopmans¹, Robert Frost¹, and Karla L Miller^1
1FMRIB Centre, Nuffield Department of Clinical Neurosciences, University of Oxford, Oxford, Oxfordshire, United Kingdom

Several correction methods (e.g. direct combination, PEN) have been proposed to reduce slab boundary artifacts in 3D multi-slab diffusion imaging. However, these methods are only capable of removing aliasing artifacts and require long scan time. In this work, a method compatible with optimal TRs is proposed to reduce slab boundary artifacts by jointly estimating slab profile and image. Initial results show the superior performance of our proposed method over the other two methods.

Michael Herbst^1,2, Benjamin Zahneisen¹, Benjamin Knowles², Maxim Zaitsev², and Thomas Ernst^1
1University of Hawaii, Honolulu, Hawaii, United States, ²University Medical Center Freiburg, Freiburg, Germany

In this work multiplexed sensitivity encoding is combined with the continuous prospective motion correction of a segmented diffusion weighted acquisition to achieve high resolution DTI. The results show that prospective motion correction during DWI substantially improved image quality in the case of subject motion. In fact, our approach resulted in a considerable image quality improvement even without intentional subject motion. In conclusion, continuous prospective motion correction allows for high resolution multi-shot DWI both in the presence of substantial and also microscopic head motion.

Sébastien Bär¹, Matthias Weigel², Jürgen Hennig¹, Dominik Von Elverfeldt¹, and Jochen Leupold^1
1Department of Radiology, Medical Physics, University Medical Center, Freiburg, Freiburg, Germany, ²Radiological Physics, University of Basel Hospital, Basel, Switzerland

In application fields of bSSFP, like MR-microscopy or molecular imaging, strong gradients mostly available at ultra-high field small animal scanners, are necessary in order to achieve sub-milimeter image resolutions. Due to relatively high b-values of these gradients, strong diffusion effects are induced and particularly read out gradients lead to a substantial drop of the steady state signal. In this work, we demonstrate the optimization of the flip angle in dependency of T₁, T₂ and b-value of the readout gradient in a standard bSSFP sequence.

Hing-Chiu Chang¹, Mei-Lan Chu¹, Mark Sundman¹, and Nan-Kuei Chen^1
1Brain Imaging and Analysis Center, Duke University Medical Center, Durham, North Carolina, United States

High-resolution and high-quality DWI can be achieved by using different multi-shot EPI acquisition techniques. Recently, the multiplexed sensitivity-encoding encoded reconstruction method (MUSE) was developed to remove aliasing artifacts in interleaved DW-EPI to shot-to-shot phase variation without need requiring external of navigator echoes. MoreoverThe MUSE algorithm integrated implemented with projection onto convex sets based algorithm framework (POCSMUSE) further enables the flexibility of high-resolution DWI obtained with arbitrary k-space trajectoriesy for multi-shot based acquisition with phase variation correction. In this study, first, a noveln interleaved block-segmented (iblocks) DW-EPI with 1) inherent and self-navigated phase variation correction and 2) reduced geometric distortion is purposeddeveloped;. Secondsecond, the new developed POCSMUSE algorithm is used to reconstruct the iblocks DWI image data with phase variation correctioninherent phase correction.

Ivana Drobnjak¹, John Lyon¹, Andrada Ianus¹, Daniel C Alexander¹, and Tim B Dyrby^2
1Centre for Medical Image Computing, Department of Computer Science, University College London, London, London, United Kingdom, ²Copenhagen University Hospital Hvidovre, Danish Research Centre for Magnetic Resonance, Hvidovre, Denmark

Axon diameter provides information about the performance of white matter pathways, and imaging it could provide an important insight into brain operation. Whilst majority of current diffusion imaging methods use standard PGSE sequence, various authors suggest that OGSE offers benefits over PGSE for imaging small pores. Here we investigate this by comparing PGSE and trapezoidal OGSE on a monkey corpus callosum. We find that optimised OGSE outperforms the optimised PGSE protocol by increasing sensitivity to smaller axon diameter. Optimized OGSE waveforms have low frequency, a novel finding, since traditionally high frequency has been considered to increase sensitivity to small sizes.

Tzu-Cheng Chao^1,2, Jr-Yuan George Chiou³, Stephan E. Maier³, and Bruno Madore^3
1Department of Computer Science and Information Engineering, National Cheng-Kung University, Tainan, Taiwan, ²Institute of Medical Informatics, National Cheng-Kung University, Tainan, Taiwan, ³Department of Radiology, Brigham and Women's Hospital, Harvard Medical School, Boston, M.A., United States

High angular resolution diffusion imaging is a useful tool in exploring cerebral microstructure, but long scan times and geometric distortions have impeded its translation to clinical practice. In the present work, a strategy integrating advantages from accelerated multishot diffusion imaging, multiplexed sensitivity encoding and compressed sensing is proposed to provide reductions in scan time and in geometric distortions. Four-fold improvements in distortion level and as much as three-fold reductions in scan time were achieved for HARDI acquisitions.

Denis Kokorin¹, Jürgen Hennig¹, and Maxim Zaitsev^1
1Department of Radiology, Medical Physics, University Medical Center Freiburg, Freiburg, Germany

Parallel spatially-selective excitation allows for a reduction of the FOV along the PE direction. This concept was developed and examined for multi-slice inner volume imaging of the spinal cord in this work. The method was tested in humans for DWI applications on a 3T MRI system with an 8-channel TxArray extension combined with an 8-channel transmit array. An advantage of using parallel excitation includes shorter durations of multidimensional pulses. Nonetheless, technological challenges must be overcome before the method presented can be used in clinical practice. This study describes initial experience obtained in the spinal cord with parallel transmission and discusses the advantages and disadvantages of employing 2D parallel excitation.

Wenchuan Wu¹, Peter Koopmans¹, and Karla L Miller^1
1FMRIB Centre, Nuffield Department of Clinical Neurosciences, University of Oxford, Oxford, Oxfordshire, United Kingdom

3D multi-slab diffusion imaging suffers from slab boundary artifacts. If the effects were purely multiplicative, the calculation of diffusion measures should not be affected after normalization by the b=0 image. However, several studies have shown slab boundary artifacts in diffusion tensor metrics. In this work, we investigated the effects of saturation, partial volume and B0 inhomogeneity on slab boundary artifacts and diffusion tensor measures (ADC and FA). We found that the slab boundary artifacts are modulated by many factors, which may not be effectively normalized by b=0 image and result in errors in diffusion measures.

Andreia C. Silva¹, Eleni Demetriou¹, Magdalena Sokolska¹, Mohamed Tachrount¹, Niall Colgan², Bernard Siow², Mark F. Lythgoe², Xavier Golay¹, and Hui Zhang^3
1Department of Brain Repair and Rehabilitation, UCL Institute of Neurology, London, London, United Kingdom, ²Centre for Advanced Biomedical Imaging, University College London, London, United Kingdom, ³Department of Computer Science and Centre for Medical Image Computing, University College London, London, London, United Kingdom

Neurite Orientation Dispersion and Density Imaging (NODDI) is a recent model-based diffusion MRI technique that has gained rapid uptake in the clinical setting owing to the availability of an economical protocol. This work aims to empirically determine a similar protocol for the preclinical setting that provides the best trade-off between acquisition time and accuracy of parameter estimation. By assessing a broad range of the subset protocols extracted from a rich dataset with many b-values and repetitions, we find that NODDI parameters can be accurately estimated with a protocol that is less than an hour.

Iain P Bruce¹, Hing-Chiu Chang¹, Nan-Kuei Chen¹, and Allen W Song^1
1Brain Imaging and Analysis Center, Duke University, Durham, NC, United States

Despite many significant advances in the field of high-resolution diffusion MRI, most current diffusion imaging techniques are limited by low SNR and extensive data acquisition times. To address these limitations, this study presents a technique for combining 3-dimensional Fourier encoding with multi-band imaging to acquire a volume of diffusion images with improved SNR and shortened acquisition time. With the use of simultaneous slice excitation, we were able to achieve sufficiently high temporal resolutions while retaining the high SNR of 3D MRI. It is anticipated that an imaging protocol of this kind will enable a wide adoption of high-resolution diffusion MRI.

Miguel Molina-Romero^1,2, Jonathan I. Sperl², Tim Sprenger^1,2, Pedro A. Gómez^1,2, Xin Liu^1,2, Ek T. Tan³, Christopher J. Hardy³, Luca Marinelli³, Bjoern Menze¹, Derek K. Jones⁴, and Marion I. Menzel^2
1Technical University Munich, Garching, BY, Germany, ²GE Global Research, Garching, BY, Germany, ³GE Global Research, Niskayuna, NY, United States, ⁴Cardiff University Brain Research Imaging Centre (CUBRIC), Cardiff University, Cardiff, Wales, United Kingdom

Varying the diffusion mixing time () in a Stejskal-Tanner experiment allows one to obtain information about the tissue microstructure. These experiments require either high-gradient-field scanners, long scanning times, or prior knowledge of the fiber orientation. On the other hand, sampling the full q-space allows one to work with no model constraints in the propagator space and potentially might reveal further tissue information. However, a full DSI acquisition for a given set of more than one  is clinically not feasible in terms of measuring time. Therefore, we need a technique that allows combining DSI acquisition and different in clinical time. In this abstract, we present a compressed sensing algorithm and a study of five different denoising associated techniques that reduce the measuring time up to a factor of R=4.

Hans Knutsson^1,2, Magnus Herbertsson³, and Carl-Fredrik Westin^1,4
1Biomedical Engineering, Linköpings Universitet, Linköping, ÖG, Sweden, ²CMIV, Linkoping University, Linköping, ÖG, Sweden, ³Mathematics, Linköpings Universitet, Linköping, ÖG, Sweden, ⁴Radiology, Brigham and Women's, Harvard Medical School, Boston, MA, United States

A thorough understanding the process of restricted diffusion in the presence of a magnetic field gradient is required for a correct interpretation of the measurements attained by any given diffusion MR scan. A number of results exist indicating to that the commonly used concept of q-space holds an oversimplification of the process. We present a novel local frequency analysis of the process showing that this is indeed the case and discuss some important consequences. We show that the basis functions that correspond to present clinical diffusion sequences are in fact very far from the Fourier basis predicted by the short pulse approximation and, to complicate things further, it is clear that the basis created will be dependent on the geometry of the individual compartments present in one voxel.

Matteo Figini¹, Alessandro Scotti¹, Stefania Marcuzzo², Silvia Bonanno², Pia Bernasconi², Victoria Moreno Manzano³, José Manuel Garcia Verdugo⁴, Renato Mantegazza², Ileana Zucca⁵, and Maria Grazia Bruzzone^6
1Scientific Direction, Fondazione IRCCS Istituto Neurologico "Carlo Besta", Milan, Milan, Italy, ²Neurology IV - Neuromuscular Diseases and Neuroimmunology Unit, Fondazione IRCCS Istituto Neurologico "Carlo Besta", Milan, Italy, ³Neuronal and Tissue Regeneration laboratory, Centro de Investigación Príncipe Felipe, Valencia, Spain, ⁴Unidad de Neurobiología comparada, Universidad de Valencia, Valencia, Spain, ⁵Scientific Direction, Fondazione IRCCS Istituto Neurologico "Carlo Besta", Milan, Italy, ⁶Neuroradiology Unit, Fondazione IRCCS Istituto Neurologico "Carlo Besta", Milan, Italy

In this work three diffusion MRI protocol for the study of the mouse spinal cord were compared. They are different both in the number of directions of the diffusion gradients and in the diffusion weightings (b-values). The comparison was based on the variability of the estimated diffusion parameters in both GM and WM and on the detection of microstructural alterations, highlighted by electron microscopy, in a murine model of amyotrophic lateral sclerosis. The overall best-performing protocol was the one with an intermediate number of diffusion directions (12) and a relatively high b-value (b=1200 s/mm²).

TOMOYA NAKAMURA¹, Isao Muro², Nao Kajihara², Shuhei Shibukawa², and Tetsuo Ogino^3
1Tokai University Hospital, Isehara, Kanagawa, Japan, ²Tokai University Hospital, Kanagawa, Japan, ³Philips Healthcare Asia Pacific, Tokyo, Japan

We achieved well visualization of the cardiac diffusion-weighted imaging with 2nd moment nulling diffusion gradient. The purpose of this study was to assess the IVIM of heart using 2nd moment nulling pulse and high b-value. The ADC and f using 2nd moment nulling pulse were significantly lower than using 1st moment nulling pulse. Thus, 2nd moment nulling pulse is insensitive to phase dispersion caused by pulsatile motion. In conclusion, we achieved improvement of heart IVIM.

Jens Sjölund^1,2, Markus Nilsson³, Daniel Topgaard³, Carl-Fredrik Westin^1,4, and Hans Knutsson^1,5
1Linköping University, Linköping, Sweden, ²Elekta Instrument AB, Stockholm, Sweden, ³Lund University, Sweden, ⁴Brigham and Women’s Hospital and Harvard Medical School, MA, United States, ⁵Center for Medical Image Science and Visualization (CMIV), Linköping, Sweden

We propose a new framework for numerical optimization of diffusion encoding gradient waveforms. The formulation allows explicit control of hardware requirements, including maximum gradient amplitude, slew rate, heating and positioning of RF pulses. The power of the approach is demonstrated by a comparison with previous work on optimization of isotropic diffusion sequences, showing possible gains in diffusion weighting or in heat dissipation.

T. Duval¹, T. Witzel², B. Keil², L. L. Wald², V. Smith², E. Klawiter², and J. Cohen-Adad^1,3
1Institute of Biomedical Engineering, Polytechnique Montréal, Montréal, Québec, Canada, ²A.A. Martinos Center for Biomedical Imaging, Massachusetts General Hospital, Harvard Medical School, Charlestown, Massachusetts, United States, ³Functional Neuroimaging Unit, CRIUGM, Université de Montréal, Montréal, Québec, Canada

Claudio Stamile¹, Gabriel Kocevar¹, Françoise Durand-Dubief^1,2, François Cotton^1,3, Carole Frindel¹, Salem Hannoun¹, and Dominique Sappey-Marinier^1,4
1CREATIS (CNRS UMR5220 & INSERM U1044), Université Lyon 1, INSA-Lyon, Villeurbanne, France, ²Service de Neurologie A, Hôpital Neurologique, Hospices Civils de Lyon, Bron, France, ³Service de Radiologie, Centre Hospitalier Lyon-Sud, Hospices Civils de Lyon, Pierre-Benite, France, ⁴CERMEP - Imagerie du Vivant, Université de Lyon, Bron, France

In this work, we present a fully automated SVM method for subject classification in three groups: healthy control subjects, relapsing-remitting and primary progressive Multiple Sclerosis (MS) patients based on the diffusion information obtained from several WM fiber bundles. The classification performance results suggest that each WM fiber bundle contributes differently to the classification of the MS clinical form. Moreover, we show that the classification result depends on the diffusion metrics used to study the fiber bundle. This result could be useful to identify a specific diffusion metric that better characterizes the WM fiber bundle.

Claudio Stamile¹, Gabriel Kocevar¹, François Cotton^1,2, Françoise Durand-Dubief^1,3, Salem Hannoun¹, Carole Frindel¹, David Rousseau¹, and Dominique Sappey-Marinier^1,4
1CREATIS (CNRS UMR5220 & INSERM U1044), Université Lyon 1, INSA-Lyon, Villeurbanne, France, ²Service de Radiologie, Centre Hospitalier Lyon-Sud, Hospices Civils de Lyon, Pierre-Benite, France, ³Service de Neurologie A, Hôpital Neurologique, Hospices Civils de Lyon, Bron, France, ⁴CERMEP - Imagerie du Vivant, Université de Lyon, Bron, France

In this work we present a new sensitive automated method to detect small longitudinal variations in diffusion parameters along the fiber bundle. We applied this method on two WM fiber bundles, namely the left and right Corticospinal Tracts and Inferior Fronto-Occipital Fasciculi of two MS patients. This method improved the to detect longitudinal variations. It also allows the distinction between “pathological” and “normal appearing” fibers, both coexisting in a bundle. This new approach offered the potential to study the relationship between lesions and distant regions of white matter that are connected by a subset of “pathological” fibers.

Yu-Jen Chen¹, Yun-Chin Hsu¹, Yu-Chun Lo¹, Shur-Fen Susan Gau², and Wen-Yih Isaac Tseng^1,3
1Center for Optoelectronic Medicine, National Taiwan University College of Medicine, Taipei, Taipei, Taiwan, ²Department of Psychiatry, National Taiwan University Hospital, Taipei, Taipei, Taiwan, ³Molecular Imaging Center, National Taiwan University, Taipei, Taipei, Taiwan

In this study, we examined the performance of predicting adult females with ADHD based on the patterns of altered tract integrity over the whole brain. The whole-brain tract information was compared with predefined differences between ADHD and healthy participants to calculate an index as the similarity with ADHD. Our results showed that the prediction performance was high (AUC > 0.8) when we compared the steps with high ES (ES > 0.65) and more continuous neighbors along tracts (CS > 4). To conclude, the information of the whole-brain tracts estimated by tract-based automatic analysis method is potentially useful for predicting adult female with ADHD.

Michael Dayan¹, Elizabeth Monohan², Sneha Pandya¹, Amy Kuceyeski¹, Thanh Nguyen¹, Susan Gauthier², and Ashish Raj^1
1Radiology, Weill Cornell Medical College, New York, NY, United States, ²Neurology, Weill Cornell Medical College, New York, NY, United States

Tract profiles offer spatial information lacking in typical analysis averaging diffusion MRI metric over reconstructed tracts. Limited research has been performed to jointly analyze multimetric tract profiles and a novel “profilometry” framework is proposed to do so. An example of application to 141 multiple sclerosis (MS) patients and 15 healthy controls (HC), with both group and single-case comparison, based on myelin water fraction and radial diffusivity profilometry is presented. Significant differences between MS and HC around the lesion locations were found. Single-case analysis revealed that profilometry visualization could be useful to detect deviation from normative data. Profilometry is suggested to provide a more specific characterization of tract microstructure. An open-source package implementing profilometry will be made available.

J. Mitra¹, S. Ghose¹, K-K. Shen¹, K. Pannek², P. Bourgeat¹, J. Fripp¹, O. Salvado¹, J. L. Mathias³, D. J. Taylor⁴, and S. Rose^1
1Australian e-Health & Research Centre, CSIRO Digital Productivity Flagship, Herston, QLD, Australia, ²Imperial College London, London, United Kingdom, ³School of Psychology, University of Adelaide, Adelaide, SA, Australia, ⁴Dept. of Radiology, The Royal Adelaide Hospital, Adelaide, SA, Australia

Patients with mild TBI sustain diffuse axonal injury (DAI) which is microscopic in nature and difficult to detect using conventional MRI. Diffusion MRI, along with probabilistic tractography, is ideally suited to detect DAI within specific white matter (WM) pathways. These approaches, based on measures of structural connectivity, can be used to identify damaged neural pathways in group-wise analyses of TBI and healthy control cohorts. We present a new method to identify significantly different and discriminative structural connections between the TBI and healthy control groups by integrating a statistical GLM-based (generalized linear model) network clustering and random forest classifier.

G. Russell Glenn¹, Joseph A. Helpern², Ali Tabesh³, and Jens H. Jensen^3
1Neurosciences & Center for Biomedical Imaging, Medical University of South Carolina, Charleston, SC, United States, ²Radiology, Neurosciences, & Center for Biomedical Imaging, Medical Univesity of South Carolina, Charleston, SC, United States, ³Radiology & Center for Biomedical Imaging, Medical Univesity of South Carolina, Charleston, SC, United States

Fractional anisotropy (FA) is a commonly used parameter from diffusion MRI data to quantify features of tissue microstructure. However, FA may take on small values despite significant diffusional anisotropy in regions with multiple, non-uniform fiber bundle orientations. This study will compare FA to measures of anisotropy that utilize higher order diffusion information from the kurtosis tensor calculated from diffusional kurtosis imaging (DKI), including kurtosis fractional anisotropy (KFA) and generalized fractional anisotropy (GFA). Both KFA and GFA are shown provide distinct and complementary information when compared to FA and may be particularly useful in regions where WM fiber bundles intersect.

Szabolcs David¹, Chantal M. W. Tax¹, Max A. Viergever¹, Anneriet Heemskerk¹, and Alexander Leemans^1
1Image Sciences Institute, University Medical Center Utrecht, Utrecht, Utrecht, Netherlands

Different implementations of processing steps such as correction for subject motion and artifacts (e.g., eddy current induced distortions and susceptibility related deformations), spatial normalization, and estimation approach used during diffusion MRI analysis may provide different results, but whether such potential differences are significant for a typical tract analysis remains unknown. In this work, we investigated the effect of tensor estimation approach and interpolation strategy on diffusion measures for a tractography based study.

Marco Palombo^1,2, Matthias Vandesquille^1,2, and Julien Valette^1,2
1CEA/DSV/I2BM/MIRCen, Fontenay-aux-Roses, France, France, ²CEA-CNRS URA 2210, Fontenay-aux-Roses, France, France

Here we propose a hybrid scheme of Parallel Tempering (PT) and Levenberg-Marquardt (LM) approaches, named PT/LM, for optimization problems of non-linear models. The aim is to design an efficient and stable pipeline for non-Gaussian diffusion metrics estimation from noisy diffusion-weighted MRI data. Diffusional kurtosis (K) and stretched exponential () models were investigated. Our numerical and experimental results, performed on ex-vivo healthy mouse brain at 11.7T, demonstrate that the proposed novel hybrid scheme improves the efficiency and stability of conventional LM based pipelines, providing less grainy non-Gaussian K- and -maps, with higher contrast-to-noise ratio.

Marion I Menzel¹, Tim Sprenger^1,2, Ek T Tan³, Valdimir Golkov^1,2, Christopher J Hardy³, Luca Marinelli³, and Jonathan I Sperl^1
1Diagnostics, Imaging and Biomedical Technologies Europe, GE Global Research, Munich, Germany, ²Technical University Munich, Munich, Germany, ³GE Global Research, Niskayuna, NY, United States

Common diffusion MRI data processing is based on the magnitude, neglecting any phase in the underlying DWI. The observed net phase has a variety of contributing sources (B0 inhomogeneity, eddy currents, motion, etc.) which are difficult to disentangle. Separating these phase contributions however is advantageous, as the phase contains information on coherent motion (i.e. brain pulsatility); and as processing of DWI to obtain parametric quantities like DTI and Kurtosis benefits from taking real valued data, as magnitude processing introduces Rician bias. This work examines robustness of phase sensitive reconstruction applied to DSI data in phantoms and in vivo human brain.

Hossein Ragheb¹, Neil A. Thacker¹, Jean-Marie Guyader², Stefan Klein², and Alan Jackson^3
1Centre for Imaging Sciences, Faculty of Medical and Human Sciences, University of Manchester, Manchester, United Kingdom, ²Biomedical Imaging Group Rotterdam, Departments of Medical Informatics and Radiology, Erasmus MC, Rotterdam, Netherlands, ³The Wolfson Molecular Imaging Centre, Faculty of Medical and Human Sciences, University of Manchester, Manchester, United Kingdom

We develop an effective motion correction method leading to determination of accurate ADC, suitable for use in patient management and drug trials. We demonstrate improvements in ADC estimates which also maintain the biological structures essential for analysis of tissue heterogeneity. our local-rigid alignment (LRA) is compared with previously published non-rigid alignment (NRA). Performance of these methods is evaluated using metrics computed from regional ADC histograms. While NRA has the advantages of being applicable on the whole volume and is fully automatic, LRA is much faster and provides advantages with regard to data smoothness by avoiding interpolation and sub-sampling.

Marco Palombo^1,2, Dianwen Zhang³, Chen Zhu⁴, Julien Valette¹, Alessandro Gozzi⁵, Angelo Bifone⁵, Andrea Messina⁶, Gianluca Lamanna⁷, and Silvia Capuani^6,8
1CEA/DSV/I2BM/MIRCen, Fontenay-aux-Roses, France, France, ²IPCF-UOS Roma, Phys. Dpt., Sapienza University, Rome, Rome, Italy, ³ITG, Beckman Institute, UIUC, Urbana, Illinois, United States, ⁴College of Economics & Management, CAU, Beijing, China, ⁵IIT, Center for Neuroscience and Cognitive Systems @ UniTn, Rovereto, Italy, ⁶Physics Dpt., Sapienza University, Rome, Italy, ⁷INFN, Pisa Section, Pisa, Italy, ⁸IPCF-UOS Roma, Phys. Dept., Sapienza University, Rome, Italy

The application of graphics processing units (GPUs) for diffusion-weighted NMR (DW-NMR) images reconstruction by using non-Gaussian diffusion models is presented. The image processing based on non-Gaussian models (such as Kurtosis and stretched exponential) currently are time consuming for any application in real-time diagnostics. Non-Gaussian diffusion imaging processing was implemented on the massively parallel architecture of GPUs, by employing a scalable parallel LM algorithm (GPU-LMFit) optimized for the Nvidia CUDA platform. Our results demonstrate that it is possible to reduce the time for massive image processing from some hours to some seconds, finally enabling automated parametric non-Gaussian DW-NMR analyses in real-time.

Daniel Olson¹, Volkan Arpinar², and L Tugan Muftuler^2
1Biophysics, Medical College of Wisconsin, Milwaukee, Wisconsin, United States, ²Neurosurgery, Medical College of Wisconsin, Wisconsin, United States

Diffusion Kurtosis Imaging (DKI) is becoming increasingly popular in diffusion weighted imaging due to its higher sensitivity to tissue microstructure compared to conventional DTI while remaining within a clinically acceptable scan time. However, the kurtosis tensor model is not as robust to noise resulting in implausible convergence of the fitting algorithm that may be mistaken as pathology. Several approaches have been proposed including outlier removal, directional weighting and regularization, and a sparsity constraint. We quantify the accuracy of each method in simulations and demonstrate performance differences with in vivo human brain data.

Lorenza Brusini¹, Mauro Zucchelli¹, Alessandro Daducci², Cristina Granziera^3,4, and Gloria Menegaz^1
1Computer Science, University of Verona, Verona, Verona, Italy, ²EPFL, Lausanne, Switzerland, ³Siemens Healthcare IM BM PI & Department of Radiology, CHUV, Lausanne, Switzerland, ⁴Department of Clinical Neurosciences, CHUV, Lausanne, Switzerland

Micro-structural indexes based on a novel reconstruction method for diffusion MRI data (SHORE) have recently been proposed and require proper validation. In this work, we derive and analyze SHORE descriptors on a group of human Diffusion Spectrum Imaging (DSI) data. The study aimed at determining their descriptive power, reliability and relations with established indexes of connectivity microstructure. Results suggest that such new biomarkers are sensitive to tissue microstructure properties and allow discriminating not only between white (WM) and gray (GM) matter but also regions with different WM topologies, besides providing an estimate of the ensemble average axons’ diameter.

Martijn Froeling¹, Peter R Luijten¹, and Alexander Leemans^2
1Radiology, UMC Utrecht, Utrecht, Netherlands, ²Image Sciences Institute, UMC Utrecht, Utrecht, Netherlands

In this study, we have investigated the use of Intra voxel incoherent motion (IVIM) modeling in combination with moment nulled diffusion gradients to correct DW imaging data for perfusion and partial volume effects and have tested this method using diffusion tensor imaging (DTI) and constrained spherical deconvolution (CSD) analysis. We have shown that using Asymmetric bipolar diffusion gradients in combination with IVIM correction this bias can be minimized. Furthermore, using 2nd order moment nulled DW sequences prevents signal attenuation due to intra voxel de-phasing originating from velocity and acceleration gradients within a voxel resulting in more reliable IVIM fits.

Hamza Farooq¹, Junqian Xu², Essa Yacoub³, Tryphon Georgiou¹, and Christophe Lenglet^3
1Electrical and Computer Engineering, University of Minnesota, Minneapolis, Minnesota, United States, ²Department of Radiology, Icahn School of Medicine, The Mount Sinai Hospital, New York, United States, ³Center for Magnetic Resonance Research, Department of Radiology, University of Minnesota, Minneapolis, Minnesota, United States

We suggest an improved fitting algorithm for Brain white matter microstructure models to extract parameters like axon radius, fiber orientation and volume fractions of different compartments. Suggested is fast and accurate as compared to traditional techniques. Further, it does not simplify or linearize model to extract the information.

Hajime Tamura¹, Shunji Mugikura², Yoshiaki Komori³, Kazuomi Yamanaka⁴, and Hideki Ota^2
1Graduate School of Medicine, Tohoku University, Sendai, Japan, ²Diagnostic Radiology, Tohoku University Hospital, Sendai, Japan, ³Siemens Japan K.K., Tokyo, Japan, ⁴Radiology, Tohoku University Hospital, Sendai, Japan

IVIM imaging of the brain requires both high signal to noise ratio and exclusion of partial-volume effects of free water. This may be achieved by addition of inversion-recovery DWI for b = 0 and 1000 [s mm^-2] without the penalty of a long scan time.

Samuel Deslauriers-Gauthier¹, Pina Marziliano², Michaël Paquette¹, and Maxime Descoteaux^1
1SCIL, Computer science department, Université de Sherbrooke, Montréal, Québec, Canada, ²School of Electrical and Electronic Engineering, Nanyang Technological University, Singapore

We present a new local reconstruction algorithm for diffusion MRI based on a new sampling theorem for non-bandlimited functions on the sphere. The performance of this algorithm is illustrated using simulated low b-value DTI-like data.

Sung-Chieh Liu¹, Yu-Jen Chen¹, Yun-Chin Hsu¹, Tzung-Jeng Hwang², Hai-Gwo Hwu², and Wen-Yih Isaac Tseng^1,3
1Center for Optoelectronic Medicine, National Taiwan University College of Medicine, Taipei, Taipei, Taiwan, ²Department of Psychiatry, National Taiwan University Hospital, Taipei, Taipei, Taiwan, ³Molecular Imaging Center, National Taiwan University, Taipei, Taipei, Taiwan

Schizophrenia (SZ) has been widely considered as a disorder of connectivity between components of large-scale brain networks. Diffusion tensor imaging (DTI) studies have revealed altered white matter structural integrity in some brain regions in SZ, such as cingulum bundles (CB), uncinated fasciculus (UF), corpus callosum (CC), fornix, etc. In this study, we used diffusion spectrum imaging (DSI) tractography and anatomy to find physical connections between pairs of brain regions. Our aim was to find significant differences in structural connectivity strengths between pairs of brain regions, including direct and indirect structural connection, between cortical and subcortical regions between SZ and controls.

Chia-Chu Chou^1,2, Prashant Raghavan¹, Dheeraj Gandhi¹, Rao P Gullapalli¹, and Jiachen Zhuo^1
1Diagnostic Radiology and Nuclear Medicine, University of Maryland School of Medicine, Baltimore, Maryland, United States, ²Electrical and Computer Engineering, University of Maryland, College Park, Maryland, United States

Thalamotomy targeting VIM has been proved to be an efficient treatment against tremor. However, the localization process for VIM is challenging. Atlas-based coordinate methods commonly adopted does not account for patient specific anatomy variation. In this study, we showed preliminary evidence that we can use probabilistic tractography to identify cortical region that has unique connectivity with VIM in the thalamus. Our tracked VIM region from cortical connectivity showed good correspondence (67% volume overlap) with the altas-based VIM mask. We expect this new method can further facilitate the process of future thalatomy surgical planning.

Oleg P Posnansky¹, Myung-Ho In¹, and Oliver Speck^1
1Institute of Experimental Physics, Department of Biomedical Magnetic Resonance, Otto-von-Guericke University, Magdeburg, Germany

In this study, an analysis of probabilistic fiber tracking was performed using extended point spread function mapping to demonstrate that accurate distortion correction and combination of echo-planar images with opposite phase-encoding polarity is important to optimally preserve spatial information. Reverse gradient polarity methods suggested previously were also applied for comparison. The results demonstrate that the probabilistic connectivity density can be improved by these corrections. Therefore, optimally preserving spatial information from the image pair can refine the structural connectivity data.

Lee Bremner Reid^1,2, Kerstin Pannek³, Roslyn Boyd², and Stephen Rose^1
1e-Health Research Centre, CSIRO, Brisbane, Queensland, Australia, ²Queensland Cerebral Palsy and Rehabilitation Research Centre, University of Queensland, Queensland, Australia, ³Department of Computing, Imperial College London, London, United Kingdom

A novel mesh-based fMRI protocol was developed to seed and constrain diffusion tractography. This method was used to investigate the integrity of corticospinal tracts in 13 subjects with unilateral cerebral palsy for whom cortical parcellation with Freesurfer was difficult or impossible. Corticospinal tracts were seeded from fMRI hand-tapping tasks. A correlation was suggested between motor ability and the integrity the tracts defined by this method for the impaired hand, but not unimpaired hand. This method allows diffusion analysis of subjects with pathologies that normally would prevent automated analysis, in a manner which avoids blurring of fMRI activation across sulci.

G. Russell Glenn¹, Joseph A. Helpern², Ali Tabesh³, and Jens H. Jensen^3
1Neurosciences & Center for Biomedical Imaging, Medical University of South Carolina, Charleston, SC, United States, ²Radiology, Neurosciences, & Center for Biomedical Imaging, Medical Univesity of South Carolina, Charleston, SC, United States, ³Radiology & Center for Biomedical Imaging, Medical Univesity of South Carolina, Charleston, SC, United States

This study optimizes the performance of the diffusional kurtosis imaging (DKI) approximation of the diffusion orientation distribution function (dODF) for white matter fiber tractography and demonstrates its reproducibility for DKI scans with different image acquisition protocols. It is found that a radial weighting power of α=4 enhances the performance of the DKI dODF while avoiding negative effects of too strong of radial weighting. In addition, fiber tracking results from DKI scans ranging in acquisition time from 51 minutes to 5.5 minutes demonstrated good qualitative consistency, suggesting DKI based WM FT can be performed on DKI datasets consistent with clinical workflows.

Hamied A Haroon¹, Claude J Bajada², Hojjatollah Azadbakht¹, and Sha Zhao^1
1Centre for Imaging Sciences, The University of Manchester, Manchester, England, United Kingdom, ²School of Psychological Sciences, The University of Manchester, Manchester, England, United Kingdom

If CSF is not correctly masked out in high angular-resolution diffusion images (HARDI) then tractography can generate spurious and random anatomical connections outside of the brain’s parenchyma, via voxels containing CSF. We present a method based on double inversion recovery (DIR) to segment and mask out CSF efficiently and improve cortical tractography. DIR data are acquired at the same resolution and geometry as HARDI data, and therefore makes DIR an ideal technique to mask out CSF in HARDI data as they are in the same space and the voxels have the same partial volume effects.

Elisa Scaccianoce^1,2, Maria Marcella Laganà¹, Francesca Baglio¹, Giuseppe Baselli², and Flavio Dell'Acqua^3
1Don Carlo Gnocchi Foundation ONLUS, IRCCS S. Maria Nascente, Milano, Italy, ²Department of Electronics, Information and Bioengineering, Politecnico di Milano, Milano, Italy, ³NATBRAINLAB, Department of Neuroimaging, Institute of Psychiatry, Psychology and Neuroscience, King’s College, London, United Kingdom

Yuchuan Hu¹, LinFeng Yan¹, Lang Wu², DanDan Zheng³, TianYong Xu³, Wen Wang⁴, and GuangBin Cui^1
1Department of Radiology, Tangdu Hospital, Fourth Military Medical University, Xi¡¯an, Shaanxi, China, ²Center for Clinical and Translational Science, Mayo Clinic, Minnesota, United States, ³MR Research China, GE Healthcare China, Beijing, China, ⁴Fourth Military Medical University, Shaanxi, China

The method of IVIM imaging uses a biexponential model to extract the perfusion-related information from a diffusion sequence. The low b-values distribution and number of excitation (NEX) might influence the accuracy of pseudodiffusion parameter derived from IVIM in brain.The study has suggested that the low b-value distribution affected the f value derived from IVIM sequence.

Ileana O. Jelescu^1,2, Magdalena Zurek¹, Kerryanne Winters^1,2, Jelle Veraart^1,2, Anjali Rajaratnam^1,2, Timothy M. Shepherd^1,2, Dmitry S. Novikov^1,2, Sungheon G. Kim^1,2, and Els Fieremans^1,2
1Center for Biomedical Imaging, Dept. of Radiology, NYU Langone Medical Center, New York, New York, United States, ²Center for Advanced Imaging Innovation and Research, Dept. of Radiology, NYU Langone Medical Center, New York, New York, United States

In this 18-week longitudinal study, we quantified in vivo changes in both conventional MRI (MTR, T₂, radial diffusivity) metrics and White Matter Tract Integrity (WMTI) parameters (derived from diffusion kurtosis imaging) during cuprizone-induced WM degeneration and subsequent recovery, in the splenium of the mouse corpus callosum. All relevant MRI metrics were affected by the cuprizone treatment and, with the exception of MTR, partially recovered after treatment ended. WMTI seems to disentangle between effects of acute and prolonged exposure to cuprizone, via the different rates of changes in axonal water fraction and extra-axonal radial diffusivity. Histological validation is underway.

Sean Foxley¹, Saad Jbabdi¹, Stuart Clare¹, Moises Fernandez¹, Connor Scott², Olaf Ansorge², and Karla Miller^1
1FMRIB Centre, University of Oxford, Oxford, OXON, United Kingdom, ²Nuffield Department of Clinical Neurosciences, University of Oxford, Oxford, OXON, United Kingdom

Use of diffusion weighted steady-state free precession (DW-SSFP) has been demonstrated to be well suited for post-mortem human brain MRI. In this work we present principle diffusion direction (PDD) estimates and deterministic tractography results of high-resolution post-mortem human brain DW-SSFP data collected at 7T. Data were acquired with 0.5mm micron isotropic resolution over 90 directions. PDD and tractography maps demonstrate very thin structures that are invisible to our 1mm acquisition protocol. The increased resolvability of smaller structures suggests that considering the time commitment required to perform this acquisition, these data hold particular value.

Kristi Clark¹, Farshid Sepehrband^2,3, Alexander Talishinsky⁴, Samuel Barnes⁵, Russell Jacobs⁵, Shagun Mehta⁴, Celia Williams⁴, and Carol Miller^4
1Institute for Neuroimaging and Informatics, University of Southern California, Los Angeles, CA, United States, ²Centre for Advanced Imaging, The University of Queensland, Brisbane, Australia, ³Queensland Brain Institute, The University of Queensland, Brisbane, Australia, ⁴Department of Pathology, University of Southern California, Los Angeles, CA, United States, ⁵Beckman Institute, California Institute of Technology, Pasadena, CA, United States

Advanced microstructural models of diffusion imaging (dMRI) have the potential to quantify microstructural changes and map connections. However, these models have been difficult to validate against a biological gold standard due to the limitation of histological techniques, which require physical cutting of the tissue to view changes two-dimensionally. An exciting new technique called CLARITY has shown that by dissolving the lipids in a volume of brain tissue, it is now possible to study neuroanatomy, especially connectivity, in three dimensions. In this project, adult human hippocampal tissue processed with CLARITY is used to validate microstructural dMRI models.

Poonam Rana¹, Sushanta Kumar Mishra¹, Mamta Aryabhushan Gupta¹, Richa Trivedi¹, B S Hemanth Kumar¹, and Subash Khushu^1
1NMR Research Centre, Institute of Nuclear Medicine and Allied Sciences, DRDO, Delhi, Delhi, India

Ionizing radiation exposure induced systemic inflammatory response might play an important role in influencing CNS function. The present study has been carried out to identify the changes occurring in brain with special reference to hippocampus within 24 hours of radiation exposure using MR techniques. The results showed changes in mean diffusivity and radial diffusivity at 3 and 24 hrs post irradiation compared to controls. The restricted diffusion as observed in DTI along with increased pro-inflammatory cytokine gene expression by real time-PCR and GFAP expression by immunohistochemistry are suggestive of systemic inflammatory response induced neuroinflammatory changes in brain.

Wieke Haakma^1,2, Bas Jongbloed³, Martijn Froeling¹, Clemens Bos¹, Stephan H. Goedee³, Michael Pedersen⁴, Ludo van der Pol³, Alexander Leemans⁵, and Jeroen Hendrikse^1
1Department of Radiology, University Medical Center Utrecht, Utrecht, Utrecht, Netherlands, ²Department of Forensic Medicine & Comparative Medicine Lab, Aarhus University, Aarhus, Central Denmark, Denmark, ³Department of Neurology, University Medical Center Utrecht, Utrecht, Utrecht, Netherlands,⁴Department of Clinical Medicine - Comparative Medicine Lab, Aarhus University, Aarhus, Central Denmark, Denmark, ⁵Image Sciences Institute, University Medical Center Utrecht, Utrecht, Utrecht, Netherlands

Multifocal motor neuropathy (MMN) is a rare immune-mediated disorder that affects 1-2 person per 100.000. Due to the progressive pure motor weakness, MMN is often confused with amyotrophic lateral sclerosis (ALS). Diffusion tensor imaging (DTI) can be helpful in evaluating changes of diffusion values, which may aid in this discrimination. We studied DTI of the forearm nerves in patients with MMN, ALS and healthy controls. MMN patients revealed decreased fractional anisotropy and axial diffusivity compared to healthy controls. These preliminary results indicate a potential role of DTI in discriminating between MMN and ALS.

Yu-Chun Lin¹, Chun-Chieh Wang², Gigin Lin¹, and Jiun-Jie Wang^3
1Department of Diagnostic Radiology, Chang Gung Memorial Hospital, Taoyuan, Taiwan, Taiwan, ²Department of Radiation Oncology, Chang Gung Memorial Hospital, Linkou, Taiwan, ³Department of Medical Imaging and Radiological Sciences, Chang Gung University, Yaoyuan, Taiwan

We propose a method to improve diffusion MRI and facilitate the matching between MR imaging and tissue. A cryostat embedding medium was used to cover the examined target during MRI studies. The examined tissue was sectioned in parallel with the imaging plane. Phantom experiments demonstrated that the embedding improved the magnetic field inhomogeneity. Animal experiments revealed significantly reduced distortions in diffusion-weighted images in both the axial and coronal planes. The in vivo MR images were easily matched with histological specimens in a slice-to-slice fashion. This method is easy and straightforward to be used in preclinical studies.

Mukesh Kumar¹, Shilpi Modi¹, Pawan Kumar¹, and Subash Khushu^1
1NMR Research Centre, Institute of Nuclear Medicine and Allied Sciences (INMAS), New Delhi, Delhi, India

The aim of present study was to investigate micro-structural disruption of white matter tracts in alcohol dependant subjects by using Tract-Based Spatial Statistics (TBSS) of DTI measures. DTI was performed on twenty-nine control subjects and twenty-eight alcoholic patients. Our result showed decreased fractional anisotropy in the anterior thalamic radiation, inferior fronto-occipital fasciculus, superior and inferior longitudinal fasciculus, forceps minor,corticospinal, cingulum, uncinate fasciculus in alcohol dependant subjects as compared to controls. These findings suggest that the microstructural changes in these white matter tracts may contribute to underlying dysfunction in cognitive and behaviour as observed in them.

Madhura Baxi^1,2, Jennifer Robinson^1,3, Paul Waggoner⁴, Ronald Beyers¹, Edward Morrison⁵, Nouha Salibi^1,6, Thomas S. Denney Jr.^1,3, Vitaly Vodyanoy⁵, and Gopikrishna Deshpande^1,3
1AU MRI Research Center, Dept. of Electrical & Computer Engineering, Auburn University, Auburn, Al, United States, ²Psychiatry Neuroimaging Laboratory, Department of Psychiatry, Brigham and Women’s Hospital, Harvard Medical School, Boston, MA, United States, ³Dept. of Psychology, Auburn University, Auburn, Al, United States, ⁴Canine Detection Research Institute, Auburn University, Auburn, Al, United States, ⁵Dept. of Anatomy, Physiology & Pharmacology, Auburn University, Auburn, Al, United States, ⁶MR R&D, Siemens Healthcare, Malvern, PA, United States

DTI-based atlas has been created for a canine model which could be used to investigate various white matter diseases. DTI tractography based structural connectivity between Anterior Cingulate Cortex (ACC) and Posterior Cingulate Cortex (PCC) regions of the default mode network was computed in dogs and was compared with that in humans to investigate evolution of cognitive functions in humans and provide structural basis for the dissociation of anterior and posterior parts of Default Mode Network (DMN) found in a recently conducted resting state fMRI study.

Malek I Makki¹ and Jeremy J Laukka^2
1MRI Research, University Children Hospital of Zurich, Zurich, Switzerland, ²Neuroscience and Neurology, University of Toledo, Toledo, OH, United States

DTI was performed on twelve patients with Pelizaeus-Merzbacher disease. These had different PLP1 mutation categories: null, moderate, and severe. Patients with moderate mutation exhibited the lowest radial diffusion and ADC and the highest FA in the splenium. This suggested hypomyelination and axonopathy. We also observed significant differences in radial diffusion and anisotropy between moderate and null mutations in the splenium showing that these patients have mild reduction in myelin with generally preserved axons

Joong Kim¹ and David L Brody^1
1Washington University School of Medicine, St. Louis, MO, United States

Mono-exponential single tensor based diffusion tensor imaging, multi-tensor based generalized q-ball imaging (GQI), and non-mono exponential based diffusion kurtosis imaging (DKI) were performed in ex vivo human frontal cortex at 11.7T with high spatial resolution. Test-retest based analysis showed that GQI and DKI have superior reproducibility than conventional FA suggesting better reliability for assessments of tissue pathology. Furthermore, unlike anisotropy measurements, kurtosis analysis did not show low signal in crossing fiber regions. Thus, in assessing complex white matter for pathology, kurtosis measurements may avoid some of the false positives to which anisotropy-based measurements are susceptible.

Sheye Aliu¹, David Newitt¹, Wen Li¹, Jessica Gibbs¹, Lisa Cimino², Eunhee Kim², Savannah Partridge³, Patrick Bolan⁴, Thomas Chenevert⁵, Mark Rosen⁶, and Nola Hylton^1
1Radiology & Biomedical Imaging, University of California at San Francisco, San Francisco, CA, United States, ²ECOG-ACRIN Cancer Research Group, PA, United States, ³Radiology, University of Washington School of Medicine, Seattle Cancer Care Alliance, WA, United States, ⁴Center for Magnetic Resonance Research, University of Minnesota, MN, United States, ⁵Radiology, University of Michigan Health System, MI, United States, ⁶Radiology, University of Pennsylvania, PA, United States

A quality assessment and ranking system for quantitative breast DWI was developed and piloted in the ACRIN 6698 trial. Initial testing suggests that such a system is useful for quality control in large multicenter trials. With improvements, we believe that this system is extensible for other applications.

Zhongping Zhang¹, Bing Wu¹, Jin Wang², and Zhenyu Zhou^1
1GE Healthcare China, Beijing, Beijing, China, ²Radiology, The Third Affiliated Hospital of Sun Yat-sen University, Guangzhou, Guangdong, China

IVIM has been widely used in seperating diffusion and perfusion process of human liver. However, the perfusion-related IVIM parameters (D* and f) suffers a large calculation bias due to the existence of Rician noise in diffusion-weighted images. In this study, diffusion-weighted images were denoised with a Rician nonlocal means filter prior to IVIM quantification. The Rician noise level was greatly reduced and the SNR was improved. The Chi2 value significantly decreased after the Rician denoising, which suggests improvement of IVIM calculation. The accuracy of IVIM metrics may be improved by denoising diffusion weighted images with a Rician noise filter.

Raphael Shih Zhu Yiin¹, Giuliano Scattoli¹, Dow-Mu Koh¹, David J Collins², Martin O Leach², and Matthew D Blackledge^2
1Department of Radiology, The Royal Marsden Hospital, Sutton, Surrey, United Kingdom, ²CR-UK and EPSRC Cancer Imaging Centre, Sutton, Surrey, United Kingdom

Whole body diffusion weight imaging (WB-DWI) is being used to evaluate nodal disease in lymphoma and metastatic nodal disease. However, there is an overlap in the MR diffusion properties of malignant and non-malignant lymph nodes. An improved understanding of the DWI characteristics of normal lymph nodes would inform disease assessment. Our study established the whole body distribution of normal lymph nodes and showed the mean ADC value for the nodes to be similar to other published studies. We also defined the full ADC histogram characteristics of normal lymph nodes, which may inform future disease assessment when compared with diseased states.

Jiancheng Zhuang¹ and Bosco S. Tjan^1
1University of Southern California, Los Angeles, California, United States

We describe an approach of using a readout-segmented echo-planar imaging sequence combined with parallel imaging and a two-dimensional navigator-based reacquisition to investigate the diffusive water transport in in-vivo human eyes. The results show the difference of water diffusive movement across different eye structures. It demonstrates feasibility of in-vivo detection of water transport in human eyes at a high spatial resolution (0.38 mm by 0.38 mm in-plane resolution) on a typical clinic scanner by using this readout-segmented echo-planar imaging sequence.

Nicola Martini¹, Simona Celi^1,2, Daniele Della Latta¹, Daniele De Marchi¹, Giuseppe Valvano^1,3, Angelo Monteleone¹, Vincenzo Positano⁴, Maria Filomena Santarelli^4,5, Sergio Berti¹, Marco Solinas¹, Luigi Landini^1,3, and Dante Chiappino^1
1Fondazione G.Monasterio CNR-Regione Toscana, Massa, MS, Italy, ²Scuola Superiore Sant'Anna, Pisa, PI, Italy, ³Department of Information Engineering, University of Pisa, Pisa, PI, Italy, ⁴Fondazione G.Monasterio CNR-Regione Toscana, Pisa, PI, Italy, ⁵Institute of Clinical Physiology, CNR, Pisa, PI, Italy

Diffusion tensor imaging (DTI) provides noninvasive information on tissue microstructure. However, the spatial resolution of DTI scans is typically limited by low signal-to-noise ratio and artifact issues that prevent its potential application to small structures, such as vessel walls. In this study we explore the use of a multi-shot EPI sequence for high resolution DTI scans of ex-vivo samples of the human aorta. Different scan protocols are compared to test the influence of scan parameters (resolution, b-value, number of diffusion gradient directions, number of signal averages) on DTI quantitative parameters.

Wieke Haakma^1,2, Michael Pedersen³, Martijn Froeling², Lars Uhrenholt⁴, Jeroen Hendrikse², Alexander Leemans⁵, and Lene Warner Thorup Boel^4
1Department of Forensic Medicine & Comparative Medicine Lab, Aarhus University, Aarhus, Central Denmark, Denmark, ²Department of Radiology, University Medical Center Utrecht, Utrecht, Utrecht, Netherlands, ³Department of Comparative Medicine Lab - Clinical Institute, Aarhus University, Central Denmark, Denmark, ⁴Department of Forensic Medicine, Aarhus University, Aarhus, Central Denmark, Denmark, ⁵Image Sciences Institute, University Medical Center Utrecht, Utrecht, Netherlands

Diffusion tensor imaging (DTI) allows evaluation of microstructural properties of tissue and is therefore an emerging imaging technique to investigate post-mortem tissue. In this work we examine the architecture and the difference in diffusion values of the lumbosacral plexus in post-mortem subjects using DTI. Reconstructions of the lumbosacral plexus show the architecture of the spine and plexus, lower mean diffusion values compared to in vivo results and nerve injuries in one trauma case. We expect that this technique can provide a valuable contribution to the understanding of pathogenesis and disease progression in peripheral neurological disorders in the future.

Qiuju Li¹, Qiyong Guo¹, Zhoushe Zhao², Jiahui Li¹, Bing Yu¹, and Yu Shi^1
1Radiology, shengjing hospital, Shenyang, Liaoning, China, ²General Electronic Company Healthcare (China), General Electronic Company Healthcare (China), Beijing, Beijing, China

A multi-b DWI was to investigate the value of aquaporins function in diagnosis of early liver fibrosis in patients. At low-b values, the resulted ADCs reflected the hepatic blood perfusion information while the functional changes of the AQPs on cell membranes at high b values. The "standard" ADC value obtained by conventional method and low-b, mid-b and high-b ADC value mesured by tri-exponential model all can distinguish normal from cirrhosis of the liver tissue. The "standard" ADC value can be hardly used to grade early liver fibrosis. However, both low-b values and high-b ADC values have the potential to detect F1 from F0.

Olivier Reynaud^1,2, Kerryanne V Winters^1,2, Dmitry S Novikov^1,2, and Sungheon Gene Kim^1,2
1Center for Advanced Imaging Innovation and Research (CAI2R), Department of Radiology, New York University School of Medicine, New York, NY, United States,²Bernard and Irene Schwartz Center for Biomedical Imaging, Department of Radiology, New York University School of Medicine, New York, NY, United States

In this study we show that the surface to volume ratio (S/V) is a natural candidate to study in vivo the restrictions in glioblastoma in the very short diffusion time regime. OGSE-DWI data acquired on preclinical scanners (G<760 mT/m, frequency in range [60-225] Hz) demonstrates excellent agreement with Mitra regime (linear fit of ADC vs square-root of diffusion time: R2=0.95) inside the tumor (N=13), suggesting that S/V can be derived in a robust manner. S/V correlates partially with ADC calculated with PGSE (Spearman coefficient -0.43). Both metrics sense different scales and provide complementary information regarding tumor growth, treatment or microstructure.

Nuno Pedrosa de Barros^1,2, Urspeter Knecht², Roland Wiest², and Johannes Slotboom^2
1University of Bern, Bern, Bern, Switzerland, ²Institute for Diagnostic and Interventional Neuroradiology, Bern, Bern, Switzerland

MR-spectroscopy and MR-diffusion have proven to provide important information for initial brain tumor diagnostics as well as for tumor progression evaluation. For advanced longitudinal analysis of brain tumor patients, tools enabling combined advanced image and spectral analysis including co-registration are of utmost importance. However, image co-registration may have negative impact on the information extracted from ADC-maps. This study analysed the impact of co-registration on the histogram analysis of ADC maps of brain tumor patients examined by MRS.

Qing Yuan¹, Daniel N Costa^1,2, Julien Sénégas³, Yin Xi¹, Andrea J Wiethoff^2,4, Robert E Lenkinski^1,2, and Ivan Pedrosa^1,2
1Radiology, UT Southwestern Medical Center, Dallas, TX, United States, ²Advanced Imaging Research Center, UT Southwestern Medical Center, Dallas, TX, United States, ³Philips Research Laboratories, Hamburg, Germany, ⁴Philips Research North America, Briarcliff Manor, New York, United States

The purpose of this study was to compare a simplified IVIM (sIVIM) model with commonly used monoexponential and biexponential models in diffusion characterization of prostate cancer and noncancerous prostate tissues. The sIVIM approach models the perfusion effect with a Delta function at b=0 s/mm2, which allows assessment of true tissue diffusion with fewer b-values therefore reducing scan time. Our results demonstrated that the sIVIM model provided ADC estimates in prostate cancer and noncancerous prostate tissues in central gland and peripheral zone equivalent to the biexponential model, and it showed better correlation with tumor aggressiveness than the monoexponential model.

Tao GONG¹, bin wang², guangbin WANG³, and shuhui YUAN^4
1Shandong Medical Imaging Research Institute, Shandong University, Shandong, Jinan, China, ²binzhou medical university, Shandong, yantai, China, ³Shandong Medical Imaging Research Institute, Shandong University, jinan, China, ⁴binzhou medical university, yantai, China

Objectives To evaluate the ability of DTI and DTT to different the PCa from BPH. Methods 46 patients£¨ 30 BPH,16 CG-PCa£© underwent routine MRI and DTI. ADC value and FA value of CG-PCa and BPH acquired from DTI . For DTT maps, two observers record the score using a four-point scale . Results ADC values between the two groups were statistically significant difference. While FA values between the two groups have no statistically difference. The DTT map score between the two groups for two viewers were all statistically significant difference. Conclusions DTI and DTT have potential for differentiating CG-cancer from BPH.

Jeong Hee Yoon¹, Jeong Min Lee¹, Mun Young Paek², Sangwoo Lee³, and Joon Koo Han^1
1Radiology, Seoul National University Hospital, Seoul, Seoul, Korea, ²Siemens Healthcare Korea, Seoul, Korea, ³Samsung electronics, Seoul, Seoul, Korea

the purpose of this study is to determine whether T2 correction on diffusion weighted imaging (DWI) using intravoxel incoherent motion (IVIM) model can provide different perfusion fraction (f) on DWI without T2 correction in Sorafenib treated VX2 liver tumors in rabbits

Zhuangzhen He¹, Yunbin Chen¹, Youping Xiao¹, Minfeng Li¹, Weibo Chen², and He Wang^3
1Fujian Province Cancer Hospital, Fuzhou, Fujian, China, ²Philips Healthcare, Shanghai, China, ³Philips Research China, Shanghai, China

Carsten Funck¹, Frederik Bernd Laun¹, and Andreas Wetscherek^1
1Medical Physics In Radiology, German Cancer Research Center (DKFZ), Heidelberg, Germany

Aim of this work was to measure the diffusion coefficient of human blood samples dependent on diffusion time T and diffusion gradient profile. A phantom is presented which allows one to maintain the sample temperature and to prevent the blood cells from sedimentation. Measured ADCs decrease with increasing T for the flow-compensated gradient profile and stabilize on a lower level in the bipolar case. The effects can be explained considering blood as a multi-compartment system. Differences between individual blood samples are, however, large compared to these variations. IVIM applications might benefit from individually measured diffusion coefficients of blood.

Shuixing Zhang¹, Wenbo Chen¹, Long Liang¹, Kannie W.Y. Chan², Yuguo Li², Bin Zhang¹, Guanshu Liu², and Changhong Liang^1
1Radiology, Guangdong Academy of Medical Sciences/Guangdong General Hospital, Guangzhou, Guangdong, China, ²Russell H. Morgan Department of Radiology and Radiological Sciences, Division of MR Research, The Johns Hopkins University School of Medicine, Baltimore, MD, United States

Contrast-induced nephropathy(CIN) is a common iatrogenic kidney disease that can cause long-term morbidity and mortality in elderly patients and patients with pre-existing kidney insufficiency or diabetes. The aim of our study is to examine the feasibility of using Intravoxel Incoherent Motion (IVIM) MRI to simultaneously measure the pathological changes in kidney diffusion and perfusion in the course of (CIN). Our results showed that the kidney perfusion and diffusion as measured by IVIM are well-correlated with those measured using conventional methods, indicating IVIM MRI can be used as an effective tool for the diagnosis and staging of CIN.

Gregory Lemberskiy^1,2, Andrew Rosenkrantz¹, Henry Rusinek¹, Els Fieremans¹, and Dmitry S Novikov^1
1Bernard and Irene Schwartz Center for Biomedical Imaging, Department of Radiology, New York University School of Medicine, New York, NY, United States,²Sackler Institute of Graduate Biomedical Sciences, New York University School of Medicine, New York, NY, United States

We report the first known in vivo observation of time dependence in benign prostate and use effective medium theory to quantify the associated length scale of tissue heterogeneity.

Robbert Harms¹, Silvia de Santis^1,2, Matteo Bastiani¹, Rainer Goebel¹, and Alard Roebroeck^1
1Maastricht University, Maastricht, Limburg, Netherlands, ²CUBRIC Cardiff University, Cardiff, United Kingdom

Large multi-shell dMRI data is increasingly used to estimate models of brain white matter that include microstructural information such as axonal density, diameters and/or dispersion. As model sophistication increases, computation times are becoming a bottleneck. We present a GPU accelerated modular and extensible toolbox for diffusion microstructure modeling. The toolbox is open source, python scriptable and compatible to other packages. Its flexibility makes it highly useful for the model developer in prototyping and testing models and its speed renders it essential in large group or population studies. We apply the toolbox to commonly used models and show orders-of-magnitude speedups.

Sheryl L Herrera¹, Morgan E Mercredi¹, Trevor J Vincent^2,3, Richard Buist⁴, and Melanie Martin^2,5
1Physics & Astronomy, University of Mantioba, Winnipeg, Manitoba, Canada, ²Physics, University of Winnipeg, Winnipeg, Manitoba, Canada, ³Physics, University of Toronto, Toronto, Ontario, Canada, ⁴Radiology, University of Mantioba, Winnipeg, Manitoba, Canada, ⁵Physics & Astronomy, Radiology, University of Mantioba, Winnipeg, Manitoba, Canada

There is an increasing drive to use diffusion spectroscopy to infer the sizes of structures in samples yet pulsed gradient spin echo limits the obtainable sizes. This work provides experimental evidence for using oscillating gradient spin echo sequences to infer the sizes of small structures. Structures of radius 0.6 ± 0.5 µm were inferred using this method which is 2-10 times smaller than structures measured with AxCaliber and more accurate than previous measurements. This work lays the foundation for inferring the size of submicron structures, such as axon diameters in samples using MRI.

Rutger Fick¹, Mauro Zucchelli², Gabriel Girard^1,3, Gloria Menegaz², Maxime Descoteaux³, and Rachid Deriche^1
1Team Athena - INRIA, Sophia Antipolis, Alpes Maritimes, France, ²University of Verona, Verona, Italy, ³Sherbrooke Connectivity Imaging Lab (SCIL), Computer Science Department, Quebec, Canada

In clinical applications for diffusion MRI the maximum number of samples is often limited by practical constraints. Given a clinically feasible 64 direction acquisition, we compare the single-shell approach, which samples at a single b-value, with a multi-shell approach, which spreads the same number of samples over multiple b-values. We show that using the continuous analytical bases 3D-SHORE and MAP-MRI on multi-shell data we can have a better estimation of both the white matter directionality using the ODF, and microstructural features such as the axon radius, without increasing the number of samples.

Gaetan Duchene¹, Frank Peeters¹, and Thierry Duprez^1
1Medical Imaging, Université Catholique de Louvain, Brussels, Brussels, Belgium

Pore size distribution (PSD) estimation has been proposed using double pulsed field gradient sequences with zero mixing time and infinite diffusion time under the short gradient pulse approximation. Such an idealization is not valid for clinical scanners. In this work, we simulated data from MRI scanners (with finite gradient duration). We investigated the resulting systematic errors on the estimated PSDs as well as random errors due to noise, and discuss the choice of experimental parameters. We found that accurate estimations are possible if the exact gradient waveform is taken into account in the estimation procedure.

Marco Palombo^1,2, Chloé Najac^1,2, and Julien Valette^1,2
1CEA/DSV/I2BM/MIRCen, Fontenay-aux-Roses, France, France, ²CEA-CNRS URA 2210, Fontenay-aux-Roses, France

Here we investigate the ability of double polarity stimulated-echo (STE) and magic-ratio stimulated-echo (MAGSTE) sequences in counteracting susceptibility-induced-background-gradients (Gint).In order to clarify if they compensate Gint effects on DW-NMR signal, numerical simulations were performed.Our results demonstrate that: 1) double polarity STE and MAGSTE fully compensate only for Gint constant and equal during preparation and read interval; 2) only MAGSTE fully compensates for Gint varying in time in a piecewise constant manner; 3) neither double polarity STE nor MAGSTE can compensate for random and uncorrelated Gint variations in space and time, which is the most common condition in heterogeneous living tissues.

Mohammed Salman Shazeeb¹ and Bashar Issa^1
1Department of Physics, UAE University, Al-Ain, Abu Dhabi, United Arab Emirates

Brain vasculature is conventionally represented as 2D or 3D straight cylinders when simulating BOLD contrast effects in fMRI. In reality, the vasculature is more complicated with branching and coiling especially in tumors. In this study, we introduce a simple cylinder fork model to reflect the curvature of vessels and performed simulations to study the effect of the cylinder fork bifurcation angle on the transverse relaxation times at different vessel diameters. This model clearly showed a dependence of the bifurcation angle on the relaxation times which could potentially be used as a tool to differentiate between normal and tumor vessels.

Mohammed Salman Shazeeb¹ and Bashar Issa^1
1Department of Physics, UAE University, Al-Ain, Abu Dhabi, United Arab Emirates

Straight cylinders are conventionally used when simulating BOLD contrast effects in fMRI. But the vasculature is more complicated with branching and coiling especially in tumors which can affect relaxation times. Diffusion can also introduce changes in the relaxation mechanisms within tumors. In this study, we use a simple cylinder fork model to reflect the curvature of vessels and performed simulations at different diffusion rates to study the effect of the bifurcation angle and diffusion on transverse relaxation times. This model clearly indicated a dependence of relaxation times on the bifurcation angle which varied between low and high diffusion constants.

Hua Li¹, Xiaoyu Jiang¹, Jingping Xie¹, J. Oliver Mclntyre¹, John C. Gore¹, and Junzhong Xu^1
1Institute of Imaging Science, Vanderbilt University, Nashville, TN, United States

The apparent diffusion coefficients (ADCs) within tumors were usually reported to increase after treatment, presumably because of the decreased cellularity. However, the cell membrane permeability was reported to increase following apoptotic cell death, which could also increase ADC. Due to short diffusion times, the oscillating gradient spin-echo (OGSE) methods are assumed to be relatively independent of changes of permeability. In the current study, we developed an experimental protocol to selectively alter cell membrane permeability, and then investigated its influence on diffusion measurements over a broad range of effective diffusion times (0.4 - 3000 ms).

Simona Schiavi¹, Houssem Haddar¹, and Jing-Rebecca Li^1
1DeFI, INRIA, École Polytechnique, Palaiseau, France

DMRI has been established as a useful tool to obtain voxel-level information on tissue microstructure. An important quantity measured in dMRI is the apparent diffusion coefficient (ADC), and it has been well established by in-vivo brain imaging experiments that the ADC depends significantly on the diffusion time. To aid in understanding and interpreting the experimentally measured ADC, we derive a new mathematical model of the voxel-level ADC that is diffusion time-dependent and accurate over a large range of diffusion times, using a linearization of the Bloch-Torrey equation around low b-values. The goal is to use this time-dependent ADC model in the future to identify parameters of tissue microstructure from the experimentally measured ADC at different diffusion times.

Andrada Ianus¹, Ivana Drobnjak¹, and Daniel C Alexander^1
1Centre for Medical Image Computing, Department of Computer Science, UCL, London, United Kingdom

Current methods of calculating microscopic anisotropy lack specificity in the case of substrates with a distribution of pore sizes. Here we present for the first time a model-based approach for estimating explicit microstructural features, such as pore size and elongation that are invariant to size distribution. Single pulsed field gradient (PFG) measurements are not sufficient to recover this information, while double and triple PFG measurements yield accurate estimates for all the substrates considered in the simulation.

Matthew Budde¹ and Nathan Skinner^2
1Neurosurgery, Medical College of Wisconsin, Milwaukee, Wisconsin, United States, ²Biophysics, Medical College of Wisconsin, Milwaukee, Wisconsin, United States

Diffusion tensor imaging has demonstrated a unique ability to detect nervous system injury with fractional anisotropy (FA), but it is limited by assumptions that can complicate and confound its sensitivity and interpretation. Novel diffusion weighted sequences, including the double pulsed field gradient (dPFG) and q-vector Magic Angle Spinning (qMAS) have reported to have high sensitivity to the microscopic effects of injury without the complicating effects of fiber configuration, or macroscopic anisotropy. Herein, we simulated these sequences in a model of acute axonal injury characterized by beaded axons to reveal the sensitivity to injury and ability to resolve microscopic injury.

Michael Paquette¹, Chantal M.W. Tax², Alexander Leemans², and Maxime Descoteaux^1
1Sherbrooke Connectivity Imaging Lab (SCIL), Université de Sherbrooke, Sherbrooke, Quebec, Canada, ²University Medical Center Utrecht, Image Sciences Institute, Utrecht, Netherlands

Diffusion MRI is commonly used to perform tractography, in which the local reconstruction algorithms prioritize the angular accuracy of the diffusion directions, describing the underlying white matter bundles. This is why single-shell HARDI is generally the acquisition scheme of choice. One downside of single-shell acquisitions is their inability to be applied in high-order models that need some information about the radial diffusion such as MAP, kurtosis or multi-compartment models which provide interesting microstructural diffusion features. But can we get satisfactory orientation reconstructions from multiple b-values diffusion data without increasing the total number of measurements?

Kwan-Jin Jung^1
1Radiology, University of Louisville, Louisville, KY, United States

The evaluation of sequence and its parameters for the diffusion tensor imaging is difficult because of the physiology noise and the lack of a proper diffusion phantom. Available phantoms are structured to provide the proton signal outside the tube, which is opposite to the neuronal fiber. The magnetic susceptibility of the tubes in the phantom induces a directional variation of the MR signal. Osage orange is a fruit with columns running from the center core to the surface. Each column was found to provide anisotropic diffusion. Therefore, it was used as a diffusion phantom to evaluate the slice-accelerated diffusion sequence.

Maximilian Pietsch¹ and J-Donald Tournier^1
1Centre for the Developing Brain, King's College London, London, London, United Kingdom

Diffusion tensor imaging (DTI) of white matter is routinely used to investigate myelin loss in demyelinating diseases. In literature, radial diffusivity is often considered to correlate with demyelination, and fractional anisotropy (FA) with axonal integrity. We performed Monte Carlo DTI simulations of axons modelled as randomly distributed parallel cylinders in two different set-ups for a range of packing densities: Demyelination with constant axon packing density resembling acute lesions and demyelination with constant extracellular volume fractions modelling chronic lesions. We show that radial diffusivity increases in the first scenario and decreases in the second with FA being affected in both cases.

Mauro Zucchelli¹, Maxime Descoteaux², and Gloria Menegaz^1
1Computer Science, University Of Verona, Verona, Verona, Italy, ²Sherbrooke Connectivity Imaging Lab (SCIL), Université de Sherbrooke, Sherbrooke, Quebec, Canada

The Human Connectome Project (HCP) high resolution data represents the state of the art for diffusion imaging in human brain. 3D-SHORE is an analytical basis for diffusion MRI which is able to reconstruct the Ensemble Average Propagator (EAP) from multi-shell data for the characterization of the diffusion process in each voxel. In this work we calculated the analytical expressions for the recently introduced EAP-derived microstructural indices including RTOP, RTAP and RTPP for 3D-SHORE model. The use of the HCP data in combination with these biomarkers provides an estimation of microstructural changes between the different brain tissues.

Katharina Göbel¹, Johannes Gerlach², Robert Kamberger³, Jochen Leupold¹, Dominik von Elverfeldt¹, Carola Haas², Jan G. Korvink³, Jürgen Hennig¹, and Pierre LeVan^1
1Medical Physics, Dept. of Radiology, University Medical Center Freiburg, Freiburg, Germany, ²Experimental Epilepsy Research, University Medical Center Freiburg, Freiburg, Germany, ³Dept. of Microsystems Engineering (IMTEK), Technical Faculty, University of Freiburg, Freiburg, Germany

Organotypic hippocampal slice cultures are a well established neuronal culture system that combines the advantages of cell culturing with a neuronal network tightly reflecting the in vivo state. Our aim is to investigate these changes during epileptogenesis using high spatial resolution MR microscopy, whose non invasiveness allows continuous longitudinal monitoring. GRE and EPI-DTI sequence protocols were adapted in order to resolve hippocampal cytoarchitecture and connectivity patterns of epileptic and healthy tissue and a correspondence was established with immunohistochemical changes. This provides deeper insights into the dynamic processes of epileptogenesis at the cellular level in in vitro preparations.

Nathan P Skinner^1,2 and Matthew D Budde^1,3
1Biophysics Graduate Program, Medical College of Wisconsin, Milwaukee, WI, United States, ²Medical Scientist Training Program, Medical College of Wisconsin, Milwaukee, WI, United States, ³Department of Neurosurgery, Medical College of Wisconsin, Milwaukee, WI, United States

Recently proposed double pulsed field gradient and q-vector magic angle spinning techniques aim to circumvent limitations of diffusion tensor imaging in resolving tissue microstructure. We applied these sequences to fixed spinal cord and compared derived microstructure measurements to diffusion tensor parameters obtained from conventional pulsed gradient spin echo images. Fractional eccentricity and microscopic fractional anisotropy parameters derived with these techniques exhibit less difference between gray and white matter anisotropy along with reduced interference from fiber crossing in the cord compared to standard fractional anisotropy. Thus, improved microstructure determination combined with greater efficiency of these sequences warrants consideration for future applications.

Brian Hansen¹, Torben E. Lund¹, Ryan Sangill¹, and Sune N. Jespersen^1,2
1Center for Functionally Integrative Neuroscience, Aarhus University, Aarhus, Denmark, ²Department of Physics and Astronomy, Aarhus University, Aarhus, Denmark

Neuritic density is tightly coupled to normal brain function and its modulation is a central structural part of experience-dependent remodeling of the brain. This plasticity has many facets that do not readily yield themselves to scrutiny in vivo. However, when combined with modeling the diffusion weighted MR signal can be used to investigate brain cytoarchitecture. One such method is NODDI. Here we present and evaluate an alternative strategy aimed at mapping neurite density in vivo. We demonstrate agreement between ground truth values and fast estimate of neurite density in fixed rat brain and discuss preliminary results in human brain.

Nabraj Sapkota^1,2, John Rose³, Scott Miller⁴, Beth Bowman⁴, Lubdha Shah⁴, Erica Bisson⁵, Sook Yoon^1,6, and Eun-Kee Jeong^1,7
1Utah Center for Advanced Imaging Research, University of Utah, SLC, UT, United States, ²Department of Physics, University of Utah, SLC, UT, United States,³Department of Neurology, University of Utah, SLC, UT, United States, ⁴Department of Radiology, University of Utah, SLC, UT, United States, ⁵Department of Neurosurgery, University of Utah, SLC, UT, United States, ⁶Department of Multimedia Engineering, Mokpo National Engineering, Jeonnam, Korea, ⁷Department of Radiology, Korea University, Seoul, Korea

By using diffusion sensitized imaging (DSI), the intra-cellular fraction (ICF) has been reported as 20-30% and is considered as a non-expected fraction. Since extra-cellular fraction has been reported as 20-30% in the brain, it has been expected 70 – 80% of ICF which is not true in white matter where myelin occupied large fraction. Using the Monte Carlo simulation (MCS) and DSI measurement, we estimate intra-axonal fraction in cervical spinal cord (CSC) white matter. The DSI measurement estimated 41% of the contribution from intra-axonal space in pig CSC, which fairly agrees with light microscopy result.

Carlos J Perez-Torres¹, Tsen-Hsuan Lin¹, Chia-Wen Chiang¹, Peng Sun¹, Yong Wang^1,2, Anne H Cross^2,3, and Sheng-Kwei Song^1,2
1Radiology, Washington University, Saint Louis, MO, United States, ²Hope Center for Neurological Disorders, Washington University, Saint Louis, MO, United States,³Neurology, Washington University, Saint Louis, MO, United States

Atrophy assessed by structural MRI is a common image marker of axon/neuron loss in CNS diseases. However, atrophy can be underestimated by confounding increase in tissue volume resulting from elevated cellularity and edema resulting from inflammation. With diffusion basis spectrum imaging, we can quantify an “axon volume” which is a combination of the percent of signal associated with fibers and optic nerve volume, removing non-fiber partial volume effect. This “axon volume” detected axon loss in the presence of confounding inflammation at the onset of optic neuritis in the experimental autoimmune encephalomyelitis (EAE) mouse model.

Pew-Thian Yap¹, Yong Zhang², and Dinggang Shen^1
1Department of Radiology, University of North Carolina, Chapel Hill, North Carolina, United States, ²Department of Psychiatry & Behavioral Sciences, Stanford University, California, United States