Fingerprinting Orientation Diffusion Functions in Diffusion MRI detects smaller crossing angles
Presented During:
Tuesday, June 27, 2017: 10:30 AM - 10:43 AM
Vancouver Convention Centre
Room:
Ballroom AB
Submission No:
1741
Submission Type:
Abstract Submission
On Display:
Monday, June 26 & Tuesday, June 27
Authors:
Steven Baete1,2, Ying-Chia Lin1,2, Martijn Cloos1,2, Fernando Boada1,2
Institutions:
1Center for Advanced Imaging Innovation and Research (CAI2R), NYU School Of Medicine, New York, United States, 2Center for Biomedical Imaging, Dept of Radiology, NYU School Of Medicine, New York, United States
First Author:
Steven Baete
-
Lecture Information
|
Contact Me
Center for Advanced Imaging Innovation and Research (CAI2R), NYU School Of Medicine|Center for Biomedical Imaging, Dept of Radiology, NYU School Of Medicine
New York, United States|New York, United States
Center for Advanced Imaging Innovation and Research (CAI2R), NYU School Of Medicine|Center for Biomedical Imaging, Dept of Radiology, NYU School Of Medicine
New York, United States|New York, United States
Introduction:
Higher Angular Resolution Diffusion Imaging (HARDI) methods, such as Diffusion Spectrum Imaging (DSI[1]) and multishell Q-ball imaging[2] are robust tools for studying in vivo white matter architecture. These methods capture the complex intravoxel crossings [1] in Orientation Distribution Functions (ODFs). To use these ODFs in tractography algorithms the fiber directions in each voxel must be identified. Limited angular resolution and intrinsic ODF peak width [3] make it however difficult to correctly estimate fiber directions when the relative angle between the bundles is small [4,5]. Most methods fail to detect crossing angles less than 40° [4,5]. Even after deconvolving the ODFs with a Fiber Response Function, it remains difficult to reliably detect crossing angles smaller than 30°[6].
Here we propose a new approach inspired by key concepts first introduced in MR Fingerprinting [7,8]. Instead of a dictionary with spin evolutions at different T1 and T2 relaxation times, we generate a library of ODF-fingerprints and identify the fiber directions of ODFs by assessing the similarity between the measured data and the elements in our library (Fig 1a). We demonstrate this method on both simulated and in vivo measured ODFs.
Here we propose a new approach inspired by key concepts first introduced in MR Fingerprinting [7,8]. Instead of a dictionary with spin evolutions at different T1 and T2 relaxation times, we generate a library of ODF-fingerprints and identify the fiber directions of ODFs by assessing the similarity between the measured data and the elements in our library (Fig 1a). We demonstrate this method on both simulated and in vivo measured ODFs.
Methods:
A library of ODF-fingerprints is generated by simulating diffusion weighted signals for a number of possible fiber combinations (up to 2 fibers; main fiber along the Z-axis, angles of subsequent fibers sampled on a 642-point tessellation of the unit sphere; fiber FA ranging from 0.4 to 0.8 in steps of 0.1; assumption of cylindrical fibers (λ2=λ3) with a simple diffusion tensor model; fiber bundle volume relative to the voxel size ranging from 0 to (1-water component) in steps of 0.1; 10% water component; ADC=1.5mm2/s) on a Radial Diffusion Spectrum Imaging grid (236 q-space samples on four shells, b=1000,2000,3000,4000s/mm2 (simulations), and b=200,1500,2750,4000s/mm2 (in vivo)[9]). The ODF-fingerprints are then subsequently calculated from the simulated diffusion weighted signals. For each measured ODF-trace, we find the matching ODF-fingerprint by searching for the ODF-fingerprint with the largest dot-product[7,8]. The proposed method is compared to peak identification with DSIStudio, MRtrix3 (sh2peaks) and FSL (qboot).
In an in-house simulation, ODF-samples are simulated as above, but with random fiber directions. Rician noise is added where necessary. In vivo DSI acquisitions are acquired on a 3T clinical scanner (Prisma, Siemens, Erlangen; 20ch head coil; bmax=4000, TR/TE=2600/114ms, 220mm FoV, 2.2x2.2x2.2mm resolution, 2xGRAPPA, SMS4; healthy male volunteer, 29 y/o). The images are denoised [10] and corrected for susceptibility, eddy currents and subject motion using eddy from the FSL Library. RDSI reconstructions were performed offline (Matlab, Mathworks) and displayed with DSIStudio.
In an in-house simulation, ODF-samples are simulated as above, but with random fiber directions. Rician noise is added where necessary. In vivo DSI acquisitions are acquired on a 3T clinical scanner (Prisma, Siemens, Erlangen; 20ch head coil; bmax=4000, TR/TE=2600/114ms, 220mm FoV, 2.2x2.2x2.2mm resolution, 2xGRAPPA, SMS4; healthy male volunteer, 29 y/o). The images are denoised [10] and corrected for susceptibility, eddy currents and subject motion using eddy from the FSL Library. RDSI reconstructions were performed offline (Matlab, Mathworks) and displayed with DSIStudio.
·Figure 1
Results:
The simulations show that the fingerprinting based approach to finding fiber directions from ODFs is able to more accurately identify smaller crossing angles (Fig 1b-e). At the same time, it also improves the precision of the fiber direction. Other methods seemingly have better precision because they often are unable to resolve a second fiber bundle if it crosses at a small angle. Although there is no gold standard for the in vivo dataset, it is evident that the fingerprinting based method is able to identify smaller crossing angles (Fig 2a-d) which aids the tractography algorithms (Fig 2e-h).
·Figure 2
Conclusions:
The applications of key concepts from fingerprinting to the ODF based fiber direction identification task improves the detection of fiber pairs crossing at small angles while simultaneously maintaining fiber direction precision. This will allow fiber tracking algorithms to more accurately display neuronal tracts and calculate brain connectivity. Future work will focus on more detailed diffusion models and phantom based validation.
Acknowledgements:This project is supported in part by PHS Grants R01CA111996, R01NS082436 and R01MH00380.
Acknowledgements:This project is supported in part by PHS Grants R01CA111996, R01NS082436 and R01MH00380.
Imaging Methods:
Diffusion MRI
Modeling and Analysis Methods:
Diffusion MRI Modeling and Analysis 1
Neuroanatomy:
White Matter Anatomy, Fiber Pathways and Connectivity 2
Keywords:
MRI
Tractography
White Matter
WHITE MATTER IMAGING - DTI, HARDI, DSI, ETC
1|2Indicates the priority used for review
Would you accept an oral presentation if your abstract is selected for an oral session?
I would be willing to discuss my abstract with members of the press should my abstract be marked newsworthy:
Please indicate below if your study was a "resting state" or "task-activation” study.
By submitting your proposal, you grant permission for the Organization for Human Brain Mapping (OHBM) to distribute the presentation in any format, including video, audio print and electronic text through OHBM OnDemand, social media channels or other electronic media and on the OHBM website.
Healthy subjects only or patients (note that patient studies may also involve healthy subjects):
Internal Review Board (IRB) or Animal Use and Care Committee (AUCC) Approval. Please indicate approval below. Please note: Failure to have IRB or AUCC approval, if applicable will lead to automatic rejection of abstract.
Please indicate which methods were used in your research:
For human MRI, what field strength scanner do you use?
Which processing packages did you use for your study?
Provide references in author date format
[2] Tuch, D.S. (2004), ‘Q-ball imaging’, Magn. Reson. Med., vol. 52, pp. 1358-72.
[3] Barnett, A. (2009), ‘Theory of Q-ball Imaging Redux: Implications for Fiber Tracking’, Magn. Reson. Med., vol 62, pp 910-923.
[4] Kuo, L.-W. (2008), ‘Optimization of diffusion spectrum imaging and q-ball imaging on clinical MRI system’, NeuroImage, vol 41, pp 7-18.
[5] Jeurissen, B. (2013), ‘Investigating the Prevalence of Complex Fiber Configurations in White Matter Tissue with Diffusion Magnetic Resonance Imaging’, Human Brain Mapping, vol. 34, pp. 2747-66.
[6] Tournier, J.-D. (2008), ‘Resolving crossing fiberes using constrained spherical deconvolution: Validation using diffusion-weighted imaging phantom data’, NeuroImage, vol 42., pp. 617-25.
[7] Ma, D. (2013), ‘Magnetic Resonance Fingerprinting’, Nature, vol 495, pp. 187.
[8] Cloos, M. (2016), ‘Multiparametric imaging with heterogeneous radiofrequency fields’, Nature communications, pp. 12445.
[9] Baete, S. (2016), ‘Radial q-Space Sampling for DSI’, Magn. Reson. Med. vol 76, pp. 769-780.
[10] Veraart J. (2016), ‘Denoising of diffusion MRI data using Random Matrix Theory’, NeuroImage, vol 142, pp. 394–406.