Acquisition for Microstructural Imaging
Click on to view the abstract pdf. Click on to view the recorded presentation.
Wednesday 9 May 2012
Plenary Hall  10:00 - 12:00 Moderators: Mark D. Does, Valerij G. Kiselev

10:00 0348.   Introduction
Mark D. Does
10:12 0349.   
Feasibility of low q-space diffusion MRI at 1.5T
Henry H. Ong1, Yusuf Bhagat1, Jeremy Magland1, and Felix W. Wehrli1
1Laboratory for Structural NMR Imaging, Departement of Radiology, University of Pennsylvania School of Medicine, Philadelphia, PA, United States

A major limitation of q-space imaging is the need for strong gradient amplitudes for sufficient displacement resolution to accurately assess axonal architecture in white matter. Low q-value diffusion MRI extracts architectural information by fitting the q-space echo decay at low q-values (q-1 >> mean axon diameter) thus obviating the need for strong gradient amplitudes, which greatly facilitates clinical implementation. Here, we investigate the feasibility of this method to assess axon architecture in excised fixed porcine spinal cords at 1.5T. The results demonstrate the potential of low q-value diffusion MRI to estimate regional mean axon diameter and intracellular volume fraction.

10:24 0350.   
Mapping the axon diameter index in the corpus callosum is clinically feasible
Torben Schneider1, Claudia A M Wheeler-Kingshott1, and Daniel C Alexander2
1NMR Research Unit, Department of Neuroinflammation, UCL Institute of Neurology, University College London, London, United Kingdom, 2Centre for Medical Image Computing, Department of Computer Science, University College London, London, United Kingdom

Diffusion imaging derived in-vivo estimates of axonal microstructure have great potential as clinical biomarkers in various neurological diseases, but current approaches suffer from long scan times and infeasible hardware requirements. In this study we demonstrate a novel imaging protocol that achieves high-resolution in-vivo axon diameter estimates in the corpus callosum in only 35 minutes of scan time. We show agreement of our estimates with reported histological assessment of the corpus callosum and show good intra- and inter-subject reproducibility in 5 healthy subjects.

10:36 0351.   Dependence of Temporal Diffusion Spectroscopy on Axon Size in White Matter Tracts of Rat Spinal Cord
Junzhong Xu1, Kevin D Harkins1, R Adam Horch1, Mark D Does1, and John C Gore1
1Institute of Imaging Science, Vanderbilt University, Nashville, TN, United States

Conventional diffusion measurements use pulsed gradient spin echo methods with relatively long diffusion times, and require high b values to increase sensitivity to smaller axons. In the current work, an oscillating gradient spin echo (OGSE) method was used to acquire temporal diffusion spectra with relatively short diffusion times and low b values. The results show that the temporal diffusion spectrum is sensitive to the mean axon size of WM tracts of fixed rat spinal cord. A new parameter R, i.e. dispersion rate of ADC vs gradient frequency, is suggested as a sensitive indicator of mean axon size.

10:48 0352.   Assessing exchange between multiple compartments using multi-directional double wave diffusion sequences
Casper Sønderby1, Henrik Lundell1, and Tim B. Dyrby1
1Danish Research Centre of Magnetic Resonance, Copehagen University Hospital Hvidovre, Hvidovre, Denmark

The apparent exchange rate (AXR) has recently been proposed as a model free measure of water diffusional exchange processes. The dependency of AXR upon measurement directions is investigated. It is proposed that different compartmental exchange rates are measured when altering the measurement direction and hence AXR is not a rotational invariant measure. A novel double wave diffusion sequence is proposed and used for ex-vivo experiments in gray matter of a fixated monkey brain and the rotational dependency of AXR is demonstrated.

11:00 0353.   Apparent Diffusion Coefficient Pattern Under Different Diffusion Times in OGSE
Yanwei Wang1, Blake Walters1, and Jae K. Kim1
1Thunder Bay Regional Research Institute, Thunder Bay, On, Canada

This work has used Monte Carlo simulation to study the apparent diffusion coefficient (ADC) behavior at different diffusion time under different microstructure models and multiple gradient powers in OGSE. Study shows a strong dependence of ADC on the diffusion time and consistently convergent behavior which is microstructure specific. This converged free diffusion coefficient at sufficiently fast diffusion time (high frequency) would be used to distinguish the microstructure. An experiment using PGSE and customized OGSE in a clinical GE3.0T at 5 Gauss per cm is reported. Consistent behavior with simulation result is achieved which highlights the potential ability of OGSE.

11:12 0354.   In Vivo Investigation of Restricted Diffusion in Human Brain Using Oscillating Diffusion Gradients
Anh Tu Van1, Samantha J Holdsworth1, and Roland Bammer1
1Center for Quantitative Neuro Imaging, Radiology, Stanford University, Palo Alto, CA, United States

Cosine-modulated diffusion gradients were used in combination with tetrahedral diffusion encoding to achieve the minimum diffusion time of 4 ms at an acceptable b-value and echo time on a standard 3 T human MRI system. At these short diffusion times, on three healthy volunteers, the measured ADCs at the genu of the corpus callosum start to show the time dependent characteristic of restricted diffusion.

11:24 0355.   DTI using modulated gradients at short effective diffusion times
Henrik Lundell1, Casper Kaae Sønderby1, Lise Vejby Søgaard1, and Tim Bjørn Dyrby1
1Copenhagen University Hospital Hvidovre, Danish Research Centre for Magnetic Resonance, Hvidovre, Copenhagen, Denmark

Assessment of short term self diffusion processes using oscillating gradient techniques is a topic of recent interest with promising applications in microstructure imaging. In this study we implement a trapezoidal modulated gradient sequence for diffusion tensor imaging and apply it on a perfusion fixed monkey brain sample. We find increased diffusivities in white matter at short effective diffusion times, but overall maintained anisotropy. However, new features emerge at short diffusion times that could be related to microstructures not visible with conventional techniques.

11:36 0356.   
Measurements of diffusion at short and long time diffusion times at 17T in the rat brain in vivo and postmortem using OGSE and PGSE methods and a biexponential model
Nadya Pyatigorskaya1, Olivier Reynaud1, Luisa Ciobanu1, and Denis Le Bihan1
1NeuroSpin, Gif sur Yvette, Ile de France, France

OGSE and PGSE methods were combined to obtain diffusion-weighted in-vivo and post-mortem images of 30 rats' brain gray matter at ultrahigh magnetic field (17.2T) at various diffusion times. Diffusion parameters were estimated using a biexponential model. ADC decreases with increasing diffusion time and upon death due to increase of slow-diffusing pool fraction, which results from the interaction of water molecules with cell membranes. The slow and fast diffusion coefficients remain constant. OGSE approach combined with the biexponential model, which allows controlling the time of membranes-water molecules interaction, may be a powerful tool for elucidating the mechanisms of diffusion in tissues.

11:48 0357.   Optimised oscillating gradient diffusion MRI for the estimation of axon radius in an ex-vivo rat brain
Bernard Siow1,2, Andrada Ianus1, Ivana Drobnjak1, Mark F Lythgoe2,3, and Daniel C Alexander1
1Centre for Medical Image Computing, UCL, London, United Kingdom, 2Centre for Advanced Biomedical Imaging, UCL, London, United Kingdom,3Institute of Child Health, UCL, London, United Kingdom

We map axon radius index in the corpus callosum of an ex-vivo rat brain using orientationally invariant optimised oscillating gradient spin-echo (OGSE) diffusion MRI protocols. Reliable estimates of axon diameter estimates require high gradient amplitudes and short diffusion times. OGSE sequences can have have shorter diffusion times compared to Pulsed Gradient Spin Echo (PGSE) sequences, thus allowing shorter length scales to be probed. We optimised OGSE protocols and estimated tissue microstructure parameters by adapting the ActiveAx technique. We found that the estimated parameters are consistent over all OGSE protocols investigated. Furthermore, our results suggest that the square wave OGSE protocol has greater precision compared to the other OGSE protocols investigated.