Hui Zhang¹, Penny L Hubbard², Geoff J M Parker², and Daniel C Alexander^1
1University College London, London, United Kingdom, ²Manchester Academic Health Sciences Centre, Manchester, United Kingdom

Axon diameter mapping using diffusion MRI provides more specific biomarkers than DTI indices. Earlier works assume a model of strictly parallel axons. However, such approximation is inadequate for most white matter regions in which axons fan or bend, resulting in significant orientation dispersion. Such dispersion, if unaccounted for, leads to overestimation of axon diameters. We ameliorates this problem by proposing a model that captures orientation dispersion explicitly. We demonstrate that recovery of axon diameters is possible even in the presence of orientation dispersion, supporting accurate axon diameter mapping in a much wider set of white matter than previously possible.

Evren Ozarslan^1,2, Noam Shemesh³, Cheng Guan Koay^1,4, Yoram Cohen³, and Peter Joel Basser^1
1STBB / PPITS / NICHD, National Institutes of Health, Bethesda, MD, United States, ²Center for Neuroscience and Regenerative Medicine, USUHS, Bethesda, MD, United States, ³School of Chemistry, Tel Aviv University, Tel Aviv, Israel, ⁴Department of Medical Physics, University of Wisconsin, Madison, WI, United States

Previous methods to determine the axon diameter distribution (ADD) from MR data assume a known statistical distribution of compartment sizes. To overcome this limitation, theoretical relationships between the MR signal intensity and the moments of a general distribution of cylindrical compartments are established. A one-dimensional simple harmonic oscillator based reconstruction and estimation (1D-SHORE) framework was used as a numerical tool to estimate these moments. Results on simulated and real MR data obtained from controlled water-filled microcapillaries demonstrate the power of this approach to create contrast based not only on the mean compartment size but also its variance.

Daniel Barazany¹, Derek Jones², and Yaniv Assaf^1
1Neurobiology, Tel aviv university, Tel Aviv, Israel, ²CUBRIC, School of Psychology, Cardiff University, Wales, UK

So far, the ability to map the axon diameter distribution (ADD) along a given pathway was only possible with AxCaliber if the fiber orientation was known and so this precludes any tract-specific assessment. We extended the AxCaliber framework to 3D, and were able to achieve the ADD in any voxel of the brain and any fiber orientation. In the work, we analyzed the two main fiber fascicles of the porcine spinal cord which are known to have different ADDs. AxCaliber was able to distinguish between them and calculate their ADDs.

Leigh A. Johnston^1,2, David Wright², Rick H.H.M. Philipsen³, Scott C. Kolbe², James A. Bourne⁴, Iven M.Y. Mareels¹, and Gary F. Egan^2
1Electrical and Electronic Engineering and NICTA VRL, University of Melbourne, Parkville, VIC, Australia, ²Howard Florey Institute, Florey Neuroscience Institutes, Parkville, VIC, Australia, ³Technical University of Eindhoven, Netherlands, ⁴Australian Regenerative Medicine Institute, Monash University, Australia

The non-exponential signal decay observed in q-space diffusion acqusitions is derived for restricted diffusion in cylinders, using a probabilistic approach based on extreme value theory and expectation over continuum distributions for geometry-dependent apparent diffusion coefficients. Simulation and experimental results demonstrate the accuracy of the resultant non-exponential signal decay and the ability to infer axon densities without need for pulse duration or diffusion time approximations.

Bibek Dhital¹, Christian Labadie^1,2, Harald E Möller¹, and Robert Turner^1
1Max Planck Institute for Human Cognitive and Brain Sciences, Leipzig, Germany, ²Laboratoire de Spectrométrie Ionique et Moléculaire, Université Claude Bernard Lyon 1, France

We made MR measurements of diffusion up to very high b-factor in excised human corpus callosum, over a wide temperature range. Above a freezing phase transition at -20 °C, data showed a robust bi-exponential dependence on b-factor. Below this temperature only half of the slow component remained, suggesting two water pools within this component. An Arrhenius plot revealed significantly different activation energies for the fast and slow components. The lower of these corresponds well to breaking a hydrogen bond in a locally ordered region. This suggests that unfrozen water consists of hydration layers close to the membranes.

Henry H. Ong¹, and Felix W. Wehrli^1
1Laboratory for Structural NMR Imaging, Department of Radiology, University of Pennsylvania, Philadelphia, PA, United States

Knowledge of axon morphology would provide important insight into neural function, anatomy and pathology. Q-space imaging (QSI) offers potential for indirect assessment of WM architecture and already has been used to measure intra-cellular volume fraction and mean axon diameter. Here, we examine the feasibility of QSI to measure axon diameter distribution (ADD) from white matter tracts in healthy mouse spinal cords using the displacement probability density function. The results show that QSI-derived ADDs semi-quantitatively agree with histologic ADDs and consistently illustrate the relative differences in ADD between WM tracts.

Dmitry S Novikov¹, and Valerij G Kiselev^2
1Radiology, NYU School of Medicine, New York, NY, United States, ²Diagnostic Radiology, University Hospital Freiburg, Freiburg, Germany

Diffusion coefficient is known to reveal the surface-to-volume ratio of restrictions at short diffusion times. Sufficiently short diffusion times are practically achievable with the oscillating gradient technique or with the CPMG refocusing train in the presence of a static gradient. Interpetation of such measurements relies on representing the short-time diffusivity limit in the frequency domain. For both of these oscillating techniques, we derive exact expressions for the high-frequency behavior of the diffusion coefficient, applicable to probe the surface-to-volume ratio of restrictions. We also describe how to calculate the effect of restrictions for arbitrary gradient waveforms.

Manisha Aggarwal¹, Susumu Mori¹, and Jiangyang Zhang^1
1Russell H. Morgan Department of Radiology and Radiological Science, Johns Hopkins University School of Medicine, Baltimore, MD, United States

Diffusion measurements with conventional pulsed gradient diffusion MRI experiments reflect the combined effects of restriction barriers to water diffusion at multiple spatial scales, but cannot distinguish between different spatial scales. With oscillating diffusion gradients, it is possible to probe diffusion at separate spatial scales by varying the modulation frequency of the oscillating gradients. In this study, three dimensional diffusion tensor imaging (DTI) of perfusion-fixed mouse brains using oscillating diffusion-sensitizing gradients is presented. The resulting diffusion tensor spectrum D(f) revealed, for the first time, unique frequency-dependent tissue contrasts in the mouse cerebellum and hippocampus.

Noam Shemesh¹, Daniel Barzany², Ofer Sadan³, Yuval Zur⁴, Daniel Offen⁵, Yaniv Assaf², and Yoram Cohen^1
1School of Chemistry, Tel Aviv University, Tel Aviv, Israel, ²Department of Neurobiology, Tel Aviv University, Israel, ³Department of Neurology, Tel-Aviv Medical Center, Tel Aviv University, Israel, ⁴GE Healthcare, Israel, ⁵Laboratory of Neurosciences, Felsenstein Medical Research Center, Department of Neurology, Rabin Medical center and Tel Aviv University, Israel

Double-Pulsed-Field-Gradient (d-PFG) MR is emerging as a useful methodology for depicting underlying microstructural information in scenarios where conventional single-PFG (s-PFG) are very limited, such as when anisotropic compartments are randomly oriented. Here, we used d-PFG MRI on phantoms, pig spinal cord and rat brain. The angular variations in the E(ø) data could be easily observed for all specimens even in the raw data; furthermore, the presence of modulated E(ø) plots in grey matter tissues revealed that water is diffusing in randomly oriented anisotropic compartments having different eccentricities, which appeared as different patterns within the cortex of the rat brain.

Hui Zhang¹, Daniel Barazany², Yaniv Assaf², Henrik M Lundell³, Daniel C Alexander¹, and Tim B Dyrby^3
1University College London, London, United Kingdom, ²Tel Aviv University, Tel Aviv, Israel, ³Copenhagen University Hospital Hvidovre, Hvidovre, Denmark

Axon diameter and density provide information about the function and performance of white matter pathways. Direct measurement of such microstructure features offers more specific biomarkers than DTI indices. Many techniques to measure axon diameter statistics using diffusion MRI have been proposed in the literature, ranging from model-based approaches to Q-space imaging, but little is known of their relative performance and consistency. This work compares several representative model-based approaches quantitatively to gain insight into how the choices of tissue model and imaging protocol impact the estimation of microstructural features.