Diffusion MR Phantom for the evaluation of structural connectivity pipelines
Over the past decade, a series of methods have been proposed in high angular resolution diffusion imaging for the local reconstruction of fiber configurations in diffusion MRI, and for fiber tractography, paving the way for the study of connectomics. The tractogram, which is a result of this data processing, will be sensitive to the selection of a method at each step of this pipeline. Therefore, the validation of these methods is an important field of study and a great challenge, since there is no reliable ground truth for in vivo images, and there exist to date few realistic phantoms of diffusion MRI.
We designed Phantomαs, a tool to efficiently create a set of white matter fiber bundle geometries, from which diffusion MR images are simulated. Each fiber bundle connects two regions of a surface, which simulates the white matter/gray matter interface. The simulation of MR signal attenuation accounts for the partial volume effect (crossing fibers, interface WM-GM, interface WM-CSF), and uses a multi-compartment model accounting for intra- and extra-axonal water diffusion. This phantom served as a tool for the organization of the 2nd HARDI Reconstruction Challenge, organised at the International Symposium on Biomedical Imaging (2013, San Francisco, CA, USA).
A. Daducci, E. Caruyer, M. Descoteaux, J.-C. Houde and J.-P. Thiran. HARDI Reconstruction Challenge 2013.
Design of multi-shell sampling schemes with uniform coverage
In the context of multi-shell acquisition in Q-space imaging, we have proposed an electrostatic energy repulsion, generalised to multi-shell, for angular uniform sampling. This video illustrates the energy minimization on 3 shells, and the set of directions we get after reprojecting all the sampling points to the unit sphere. Further details can be found in the following reference. The incremental acquisition scheme can be generated and downloaded here.
Emmanuel Caruyer, Christophe Lenglet, Guillermo Sapiro, Rachid Deriche. Design of multishell sampling schemes with uniform coverage in diffusion MRI. Magnetic Resonance in Medicine, Wiley, 2013. <http://dx.doi.org/10.1002/mrm.24736>
Online estimation of the ODF in Constant Solid Angle
This video illustrates the incremental acquisition (top left), the corresponding diffusion weighted images, and the reconstruction of an orientation distribution function (ODF) field. The reconstructed ODF is calculated in constant solid angle, using a Kalman filter. This is a synthetic field, that mimics two fibers, crossing at the center of the slice. The local model for diffusion weighted images synthesis uses the mixture of Gaussian distribution. Further details can be found in the following reference. The incremental acquisition scheme can be generated and downloaded here.
Emmanuel Caruyer, Iman Aganj, Christophe Lenglet, Guillermo Sapiro, Rachid Deriche. Motion Detection in Diffusion MRI via Online ODF Estimation. International Journal of Biomedical Imaging, 2013, 2013 (Article ID 84936), pp. 1-8. <http://dx.doi.org/10.1155/2013/849363>