Diffusional kurtosis imaging (DKI)
DKI in QSIRecon
DKI reconstruction is supported in QSIRecon using the DIPY package: see init_dipy_dki_recon_wf().
This is accessible in a reconstruction specification by using a node with action: DKI_reconstruction and software: Dipy.
Also see qsirecon.interfaces.dipy.KurtosisReconstruction (primary DKI node),
qsirecon.interfaces.dipy.KurtosisReconstructionMicrostructure (when wmti is set to true),
and qsirecon.interfaces.dipy.KurtosisReconstructionMSDKI (when msdki is set to true).
DKI Foundational Papers
DKI Concept Introduction: DKI was first introduced by Jensen et al. (2005) as an extension of DTI to measure diffusion non-Gaussianity in tissues. Their seminal work defined diffusional kurtosis as a quantitative marker, showing that normal white matter has substantially higher kurtosis than gray matter. Building on this, Lu et al. (2006) provided the first full mathematical derivation of the kurtosis tensor, introduced rotational invariants like mean kurtosis (MK), reported reproducible MK values, and showed that kurtosis anisotropy can reveal complex fiber geometries. Jensen and Helpern’s 2010 review (Jensen and Helpern, 2010) consolidated the DKI model, formalizing MK, as well as axial and radial kurtosis (AK, RK) as rotationally invariant descriptors, discussed practical acquisition requirements and highlighted DKI’s sensitivity to tissue heterogeneity.
DKI Methodological Improvements: Tabesh et al. (2011) introduced constrained least-squares estimation to ensure physically valid DKI fits, defined Kurtosis Fractional Anisotropy (KFA), and provided closed-form formulas for MK and RK. Fieremans et al. (2011) extended DKI toward microstructural modeling of white matter, deriving Axonal Water Fraction (AWF) and extra-axonal tortuosity from DKI data and aligning DKI-derived parameters with known tissue features. Henriques et al. (2021) introduced Mean Signal DKI (MSDKI), which estimates DKI parameters more robustly.
DKI Studies Across The Lifespan
Normal aging: Falangola et al. (2008) showed age-related changes in DKI metrics across the healthy lifespan; MD increased and FA decreased in the oldest group, while MK exhibited distinct trends across the lifespan.
Early development: Paydar et al. (2014) demonstrated that FA and MK rise with age in WM, but MK continues to increase after FA plateaus; MK also revealed GM maturation undetectable by FA.
Adult lifespan and aging white matter: Coutu et al. (2014) found that MK and kurtosis anisotropy declines with age and that MK shows a clearer linear association with advancing age than FA/MD, indicating progressive loss of microstructural complexity in WM.
DKI Methodological Warnings and Caveats
Acquisition and fitting constraints: DKI’s higher-order model (4th-order tensor) requires multiple high-b shells and many directions, leading to longer scans and lower SNR (Steven et al., 2014). However most modern multi-shell dMRI scans are compatible with DKI.
Interpretation pitfalls (lack of specificity): DKI metrics are not tissue-specific; MK aggregates different sources (density, dispersion, heterogeneity). Alves et al. (2022) caution that the “main caveat of DKI is that different kurtosis sources are all conflated”.
Noise, artifacts, and reproducibility: Unconstrained DKI can yield non-physical or variable estimates; regularized estimation with plausible bounds improves reproducibility (Henriques et al., 2021). Further, if data is not adequately denoised and de-Gibbs’ed there will be holes in the maps. MKDKI can improve this.
References
Jens H. Jensen, Joseph A. Helpern, Arun Ramani, Hanzhang Lu, and Krzysztof Kaczynski. Diffusional kurtosis imaging: the quantification of non-gaussian water diffusion by means of magnetic resonance imaging. Magnetic Resonance in Medicine, 53(6):1432–1440, 2005. doi:10.1002/mrm.20508.
H. Lu, J. H. Jensen, A. Ramani, and J. A. Helpern. Three-dimensional characterization of non-gaussian water diffusion in humans using diffusion kurtosis imaging. NMR in Biomedicine, 19(2):236–247, 2006. doi:10.1002/nbm.1020.
Jens H. Jensen and Joseph A. Helpern. Mri quantification of non-gaussian water diffusion by kurtosis analysis. NMR in Biomedicine, 23(7):698–710, 2010. doi:10.1002/nbm.1518.
Ali Tabesh, Jens H. Jensen, Babak A. Ardekani, and Joseph A. Helpern. Estimation of tensors and tensor-derived measures in diffusional kurtosis imaging. Magnetic Resonance in Medicine, 65(3):823–836, 2011. doi:10.1002/mrm.22655.
Els Fieremans, Jens H. Jensen, and Joseph A. Helpern. White matter characterization with diffusional kurtosis imaging. NeuroImage, 58(1):177–188, 2011. doi:10.1016/j.neuroimage.2011.06.006.
Rafael N. Henriques, Marta M. Correia, Maurizio Marrale, Esther Huber, John Kruper, Steve Koudoro, Jason D. Yeatman, Eleftherios Garyfallidis, and Ariel Rokem. Diffusional kurtosis imaging in the diffusion imaging in python project. Frontiers in Human Neuroscience, 15:675433, 2021. doi:10.3389/fnhum.2021.675433.
M. F. Falangola, J. H. Jensen, J. S. Babb, C. Hu, F. X. Castellanos, A. Di Martino, S. H. Ferris, and J. A. Helpern. Age-related non-gaussian diffusion patterns in the prefrontal brain. Journal of Magnetic Resonance Imaging, 28(6):1345–1350, 2008. doi:10.1002/jmri.21604.
A. Paydar, E. Fieremans, J. I. Nwankwo, M. Lazar, H. D. Sheth, V. Adisetiyo, J. A. Helpern, J. H. Jensen, and S. S. Milla. Diffusional kurtosis imaging of the developing brain. AJNR American Journal of Neuroradiology, 35(4):808–814, 2014. doi:10.3174/ajnr.A3764.
Jean-Philippe Coutu, J. Jean Chen, H. Diana Rosas, and David H. Salat. Non-gaussian water diffusion in aging white matter. Neurobiology of Aging, 35(6):1412–1421, 2014. doi:10.1016/j.neurobiolaging.2013.12.001.
Andrew J. Steven, Jiachen Zhuo, and Elias R. Melhem. Diffusion kurtosis imaging: an emerging technique for evaluating the microstructural environment of the brain. American Journal of Roentgenology, 202(1):W26–W33, 2014. doi:10.2214/AJR.13.11365.
Rita Alves, Rafael Neto Henriques, Leevi Kerkelä, Cristina Chavarrías, Sune N. Jespersen, and Noam Shemesh. Correlation tensor mri deciphers underlying kurtosis sources in stroke. NeuroImage, 247:118833, 2022. doi:10.1016/j.neuroimage.2021.118833.
Rafael N. Henriques, Sune N. Jespersen, Derek K. Jones, and Jelle Veraart. Toward more robust and reproducible diffusion kurtosis imaging. Magnetic Resonance in Medicine, 86(3):1600–1613, 2021. doi:10.1002/mrm.28730.