Optimized workflow for diffusion kurtosis imaging of newborns

Diffusional kurtosis imaging (DKI) is a recently proposed extension of the conventional DTI model. It has been shown to offer more sensitive characterization of neural tissues than DTI. So far, DKI has only been applied to adult human and small animal studies, but not yet to human newborns. In this work, we present an optimized workflow for the acquisition and processing of DKI images of newborns. First, optimal set of diffusion weighting gradients for DKI studies of newborn subjects is proposed. Optimized gradients allow to estimate DKI parameters with the highest precision. Next, preprocessing and segmentation of the DKI data is considered, including motion correction, eddy currents suppression, intensity modulation and gradients reorientation. Finally, statistics of estimated diffusion and kurtosis parameters for different neonatal brain tissues are calculated.