Detection of Microscopic Diffusion Anisotropy in Human Brain Cortical Gray Matter in Vivo with Double Diffusion Encoding
Marco Lawrenz1 and Juergen Finsterbusch1

1Systems Neuroscience, University Medical Center Hamburg-Eppendorf, Hamburg, Germany


Double diffusion encoding experiments with two weighting periods applied successively in the same acquisition offer access to microscopic tissue properties. Rotationally invariant measures of the so-called microscopic diffusion anisotropy as a marker for cell or compartment shape have reliably been determined in brain white matter. In this study, it is demonstrated that microscopic diffusion anisotropy can also be detected in cortical gray matter in vivo and measures of it can be determined extending first evidences presented recently. However, an inversion recovery pulse is required to null white matter signals and avoid partial volume effects.


Double diffusion encoding (DDE) or double wave vector experiments (DWV) [1,2] using two diffusion weighting periods applied successively in a single acquisition (Fig. 1) become more and more popular, in particular since they offer access to microscopic tissue properties [1-4]. For a long mixing time $$$\tau_m$$$ between the weightings, the microscopic diffusion anisotropy causes a signal difference between the parallel/anti-parallel and orthogonal weighting combinations [1]. Unlike DTI, this anisotropy effect has been detected in human brain matter in vivo even in a region-of-interest (ROI) which macroscopically appears isotropic [5]. Thus, it is very promising to use DDE to investigate microscopic diffusion anisotropy in cortical gray matter (GM) structures. On a standard whole-body MR system the acquisition is challenging since the typical WM signal modulation can be expected to be much larger. Nevertheless, in this study, the detection of microscopic diffusion anisotropy in cortical GM in vivo is presented extending first evidences [6] by a more systematic investigation of the WM partial volume effect and the determination of the rotationally invariant MA measure.


DDE experiments were performed on a 3T whole-body MR system (TIM Trio, Siemens Healthcare) on healthy volunteers after their informed consent was obtained. A spin-echo echo-planar imaging sequence (Fig. 1) using 4.0 x 4.0 x 3.0 mm3 resolution (TE/TR = 155 ms/6.5 s) was applied. Only a single transverse slice was measured to minimize table vibration effects; a slice positioning in the centrum semi-ovale avoided regions with field inhomogeneities. A dodecane (C12H26) phantom that is not expected to exhibit anisotropic diffusion was used as a reference. The two diffusion-weighting periods were applied with a b value of 500 s mm-2 each, a diffusion time $$$\Delta$$$ of 25 ms, a mixing time $$$\tau_m$$$ of 45 ms, and a gradient pulse duration $$$\delta$$$ of 22 ms. An inversion-recovery pulse was applied prior to the excitation (Fig. 1) with different inversion times TI. All 64 combinations of 8 directions sampling a circle in steps of 45° were applied for the three coordinate planes. With one image without diffusion weighting, the total acquisition time for one plane was 7 min 29 s. To estimate the MA index, a rotationally invariant measure of microscopic diffusion anisotropy [5], 96 directional combination of 18 directions were involved (11 min 10 s). Averaged MR images of all parallel/anti-parallel and all orthogonal combinations, their difference and MA maps were calculated. Furthermore, the signal variation with the angle $$$\theta$$$ was analyzed in different ROIs (WM, GM, outside brain).

Results and Discussion

The best suppression of WM structures is observed for a TI = 630 ms (Fig. 2a and b). The individual MR signal curves (Fig. 2e) for the GM ROI chosen (Fig. 2d), exhibit the pattern typical for microscopic diffusion anisotropy, minima for 90°and 270° and maxima for 0° and 180°. However, the relative modulation depends on the inversion time which indicates that partial volume effects of WM contribute to the modulation. Thus, nulling of WM signals (as obtained at TI = 630 ms) is essential to obtain the GM anisotropy effects. At TI = 630 ms, the difference map (Fig. 2c, 630 ms) reveals a slight difference in voxels containing GM while the difference in WM regions close-by is significantly lower, if any. The corresponding mean signal curves (Fig. 3) reveal the typical anisotropy feature with 3-5 % signal deviation in all three principle planes (red curves) whereas for a WM ROI (blue curves) only a minor difference between parallel/anti-parallel and orthogonal combinations and a small deviation (below 1%) of the signal curve from a flat line is observed. Thus, it seems very unlikely that partial volume effects with WM – which seem present at other TIs – cause the observed signal difference in GM at TI = 630 ms.

The mean signal averaged over the respective ROIs (Fig. 4) support the findings of a distinct anisotropy present in GM regions, whereas WM, reference phantom as well as outer regions show an absence of microscopic diffusion anisotropy. A minor systematic but non-crucial artifact seems to be present in the dodecane phantom in the yz-plane. Furthermore, with a mean value of 0.2 +/- 0.1 in all cortical GM voxels, also the rotational invariant (microscopic diffusion anisotropy) MA measure shows a lower value than human WM as expected (see also Fig. 5).


Microscopic diffusion anisotropy was detected in cortical gray matter with minimized confounds of white matter partial volume effects and MA maps could be extracted which could help to characterize tissue microstructure in healthy and pathologic brain structures.


No acknowledgement found.


[1] Mitra PP, Multiple wave-vector extensions of the NMR pulsed-field-gradient spin-echo diffusion, Phys. Rev. B 51, 15074 (1995)

[2] Shemesh N, Conventions and nomenclature for double diffusion encoding (DDE) NMR and MRI, Magn. Reson. Med. (EPub) DOI: 10.1002/mrm.2590, (2015)

[3] Cheng Y, Multiple scattering by NMR, J. Am. Chem. Soc. 121, 7935 (1999)

[4] Özarslan E, Compartment shape anisotropy CSA revealed by double pulsed field gradient MR, J. Magn. Reson. 199, 56 (2009)

[5] Lawrenz M, Double-wave-vector diffusion-weighted imaging reveals microscopic diffusion anisotropy in the living human brain, Magn. Reson. Med. 69, 1072 (2014)

[6] Lawrenz M, Evidence for the detection of microscopic diffusion anisotropy in human brain gray matter in vivo, Proc. ISMRM 2014, p. 2638


Fig. 1: Basic pulse sequence for the DDE experiment used in the present study. An initial adiabatic inversion recovery (IR) pulse is applied to null the MR signal of the white matter.

Fig. 2: MR images (parallel/orthogonal/difference, a-c) with different TI, respectively, demonstrate the WM nulling for TI = 630 ms and significant contributions of WM otherwise. For the cortical GM ROI depicted (d), the angular MR signal curves (e) were obtained. Regions with extreme signal dropouts or increases were ignored (arrows).

Fig. 3: MR signal variation (left) for the angular signal curves in different regions-of-interest (right): cortical GM (a, red), WM (b, blue), and in the reference phantom (c, black) in the three principal planes.,

Fig. 4: Mean signal difference (a: absolute and b: relative) between parallel and orthogonal DDE measurements for cortical GM, WM, regions outside the brain/phantom and in the reference phantom in the three principal planes.

Fig. 5: MA map derived from three in vivo DDE measurements.

Proc. Intl. Soc. Mag. Reson. Med. 24 (2016)