Assessment of Diffusion Time Dependence of Diffusion Kurtosis in Rat Spinal Cord
Nathaniel D Kelm1,2, Kevin D Harkins2, and Mark D Does1,2,3,4

1Biomedical Engineering, Vanderbilt University, Nashville, TN, United States, 2Vanderbilt University Institute of Imaging Science, Vanderbilt University, Nashville, TN, United States, 3Radiology and Radiological Sciences, Vanderbilt University School of Medicine, Nashville, TN, United States, 4Electrical Engineering, Vanderbilt University, Nashville, TN, United States


Diffusion kurtosis imaging (DKI) is an extension of DTI with the ability to provide additional information about tissue microstructure. To evaluate the diffusion time dependence of diffusion kurtosis, kurtosis is measured perpendicular to white matter tracts in rat spinal cord for diffusion times ranging from 12 to 100 ms. In this study, kurtosis increased as a function of diffusion time in white matter, yet decreased in gray matter. Assessing the change in diffusion kurtosis across diffusion time could potentially inform upon the underlying white matter microstructure.


Diffusion kurtosis imaging (DKI) is a higher-order extension of diffusion tensor imaging (DTI) shown to provide additional information about tissue microstructure1. Although the dependence of the water diffusion MRI signal on diffusion time has previously been assessed for a wide spectrum of diffusion times in a variety of tissue types, current knowledge is limited regarding the effects that changes in diffusion time have on diffusion kurtosis. Previous studies have examined the diffusion time dependence of kurtosis through simulations2, applied to ischemic stroke3, in rat brain4-6, and in muscle7, but experimental data of kurtosis perpendicular to white matter (WM) tracts covering a wide spectrum of diffusion times have not previously been shown. In this study, apparent diffusion kurtosis is measured perpendicular to white matter tracts in ex vivo rat spinal cord for diffusion times ranging from 12-100 ms.


Adult female Sprague-Dawley rats (N=3) were perfusion fixed with 4% paraformaldehyde (PFA) in PBS and the spinal cords were excised, post-fixed, and washed in PBS before imaging. Single-slice 2D imaging of the cervical spinal cord was performed on a 15.2T 11-cm bore Bruker scanner with slice thickness = 2 mm, FOV = 4.8 x 4.8 mm2, and matrix size = 48 x 48 for an in-plane resolution of 100 µm. Diffusion-weighted imaging (DWI) data were acquired using a diffusion-weighted stimulated echo sequence with TR/TE = 1800/16 ms. Diffusion weighting was achieved with δ = 5 ms, Δ = 12, 25, 40, 60, and 100 ms, 5 b-values ranging from 0 to 16000 s/mm2, 1 direction (perpendicular to the spinal cord), and 20 signal averages with gradient polarity reversal for a scan time of ≈ 12h. After images were zero-padded 2x, perpendicular diffusivity (D) and kurtosis (K) were estimated for each diffusion time using a non-linear least squares approach. ROI analysis was performed for six different white matter tracts: dorsal corticospinal (dCST), funiculus gracilis (FG), funiculus cuneatus (FC), rubrospinal (RST), vestibulospinal (VST), and reticulospinal (ReST), as well as gray matter (GM).


Fig. 1 shows maps of K for diffusion times of 12 and 40 ms. Visually, there is an apparent increase in K of WM and a slight decrease in K of GM as Δ increases from 12 to 40 ms. Fig. 2 shows a plot of ROI means of K for Δ = 12­ to 100 ms. Each of the 6 white matter tracts demonstrated an increase in K (20-42%) as Δ increased from 12 to 100 ms, whereas K decreased in GM (23%). Additionally, in general, white matter tracts with smaller axons and higher axon density (e.g. dCST and FG) had higher K than those with larger axons and lower axon density (e.g. FC and VST).


This study showed an increase in K in WM and a decrease in K in GM of rat spinal cord between diffusion times of 12 and 100 ms. Previous simulation work demonstrated an initial rise in kurtosis at short diffusion times and a decrease at longer diffusion times, with the time where maximum kurtosis occurs dependent on the tissue microstructure2. Experimentally, this was shown in rat brain gray matter, where kurtosis peaked at 10 ms and decreased at longer diffusion times6, consistent with the decrease demonstrated in GM in this study. However, for the diffusion times utilized in this study, K in WM only increased, indicating that the period of time where kurtosis rises is much longer than that shown in other tissues. This is potentially due to the significant contrast in water diffusion in the extra-axonal space versus the highly restricted intra-axonal space, which prevents the diffusion displacement distribution from approaching Gaussian for a relatively long period of time. Because of this, assessment of diffusion kurtosis changes as a function of diffusion time could provide useful information concerning WM microstructure. Additionally, significant changes in K for these common diffusion times demonstrate the importance of considering diffusion time when comparing values of kurtosis across different studies.


The diffusion time dependence of diffusion kurtosis was examined in ex vivo rat spinal cord, demonstrating an increase in WM K and a decrease in GM K. Future work will focus on extending the range of diffusion times to observe whether K reaches a maximum at longer diffusion times and experiments with other tissue types such as rat brain white matter or rat peripheral nerve.


NIH R01 EB001744, NIH S10 RR029523


1. Jensen JH, et al. MRM. 2005;53(6):1432-1440. 2. Fieremans E, et al. NMR Biomed. 2010;23(7):711-724. 3. Lätt J, et al. in Proc. ISMRM. 2009, p. 40. 4. Wu EX and Cheung MM. NMR Biomed. 2010;23(7):836-848. 5. Portnoy S, et al. MRM. 2012;69(4):1131-1145. 6. Pyatigorskaya N, et al. MRM. 2013;72(2):492-500. 7. Marschar AM, et al. JMRI. 2014;41(6):1581-1590.


Figure 1. Parameter maps of K in rat spinal cord for Δ = 12 and 40 ms.

Figure 2. Plot of K versus Δ for Δ = 12 to 100 ms. Each dot represents an ROI mean with error bars spanning the mean ± SEM.

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