Efficient experimental designs for isotropic generalized diffusion tensor MRI (IGDTI)

Abstract

Purpose: We propose a new generalized diffusion tensor imaging (GDTI) experimental design and analysis framework for efficiently measuring orientationally averaged diffusionweighted images (DWIs), which remove bulk signal modulations attributed to diffusion anisotropy and quantify isotropic higher-order diffusion tensors (HOT). We illustrate how this framework accelerates the clinical measurement of rotationinvariant tissue microstructural parameters derived from HOT, such as the HOT-Trace and the mean t-kurtosis. Theory and Methods: For a large range of b-values, we compare orientationally averaged DWIs measured with high angular resolution diffusion imaging to those obtained with the proposed isotropic GDTI (IGDTI) experimental design. We compare rotation-invariant microstructural parameters measured with IGDTI to those derived from HOTs measured explicitly with GDTI.

Measuring non-parametric distributions of intravoxel mean diffusivities using a clinical MRI scanner

“…The mADC-weighted signal computed from multiple DWIs acquired with LTE using the standard diffusion sequence (Stejskal and Tanner, 1965) with gradient orientations uniformly sampling the unit sphere can be modulated by anisotropic restrictions and their orientational dispersion, while the IDE signal provides isotropic encoding in all microscopic water pools, regardless of shape and orientation (Topgaard, 2017;Westin et al, 2016). While mADC-weighted DWIs can be obtained with larger b-values, better SNR, and a well-defined diffusion time (Avram et al, 2018b), IDE DWIs can be acquired faster and provide a more specific and quantitative assessment of intravoxel water diffusivities.…”

A robust deconvolution method to disentangle multiple water pools in diffusion MRI

“…Furthermore, in the proposed method, we used the geometric mean, which does not allow us to capture the intrinsic anisotropy of tissues, FW and pseudo-diffusion. Nonetheless, this is generally well accepted for FW, 6 has been shown to be adequate on brain data at multiple diffusion values 65 and is not expected to be detrimental on pseudo-diffusion estimates. 66,67 In addition, as shown in this work, if part of the data is acquired with higher angular resolution, anisotropic models can still be applied after estimation of the isotropic components.…”

Retaining information from multidimensional correlation MRI using a spectral regions of interest generator

“…11 6 2 ms) with 21 logarithmically sampled τ 1 values ranging from 14.3 to 800 ms by using an IR-DWI-EPI sequence; a 1D T 2 −weighted data set ( = b 0) with 20 logarithmically sampled τ 2 values ranging from 11.6 to 120 ms by using a DWI-EPI sequence. For diffusion encoding, we used the isotropic generalized diffusion tensor MRI (IGDTI) acquisition protocol to achieve an efficient orientationally averaged DW signal 33 with the following parameters: 21 linearly sampled b-values ranging from 400 to 16000 s/mm 2 in 3 directions, 16 linearly sampled b-values ranging from 4000 to 16000 s/mm 2 in 4 directions, and 11 linearly sampled b-values ranging from 8000 to 16000 s/mm 2 in 6 directions, using the efficient gradient sampling schemes in Table 2 of 33 . Additional DW parameters were δ = 4 ms and ∆ = 15 ms.…”

“…It should be noted that to avoid computational instability and infeasible acquisition time, the diffusion was not characterized using a tensor distribution 32 . Instead, we investigated the orientationally averaged diffusivity, D , encoded by the isotropic generalized diffusion tensor MRI (IGDTI) acquisition protocol 33 . This type of diffusion encoding increases the contrast given by local anisotropy, and is not intended to measure the isotropic diffusion in the system.…”