Influence of Maximal B-Value, Fit Polynomial and Number of Diffusion Directions on the Measured Kurtosis: A Phantom Study

Tristan Anselm Kuder¹, Bram Stieltjes, Amir Moussavi¹, Frederik Bernd Laun¹

The diffusion kurtosis is obtained by fitting a polynomial to the logarithmic diffusion weighted signal acquired at different b-values. The aim of this study was to evaluate factors influencing the measured kurtosis. A strong dependence on the maximum b-value and the fitting polynomial was found. The cubic fit does not eliminate the dependence on the maximum b-value and causes a larger uncertainty in the measured kurtosis values compared to the quadratic fit. Thus, the quadratic fit is preferable. Fitting the Diffusion Kurtosis Tensor with 15 directions is unstable and at least 30 gradient directions should be used.