Darryl McClymont1, Irvin Teh1, Hannah Whittington1, Vicente Grau2, and Jurgen Schneider1
1Division of Cardiovascular Medicine, University of Oxford, Oxford, United Kingdom, 2Department of Engineering Science, University of Oxford, Oxford, United Kingdom
Diffusion kurtosis imaging provides higher-order
information about diffusion. However, the quadratic term in the diffusion
kurtosis model produces undesirable behaviour at high b-values as a result of
the negative tails of the diffusivity distribution. A truncated normal distribution
has been proposed to address this in one dimension. This work extends this
concept to a multivariate truncated normal distribution, and extends the range
of b-values over which kurtosis can be estimated. The proposed model is fit to diffusion
data from rat hearts, and yields kurtosis values that are consistent with the DKI
model.
