The use of novel gradient directions with DTI to synthesize data with complicated diffusion behavior

This study demonstrates a new technique for synthesizing diffusion tensor imaging (DTI) data sets that exhibit complex diffusion characteristics by performing operations on acquired DTI data of simple structures with anisotropic diffusive properties. The motivation behind this technique is to characterize the behavior of noise in complicated data using a phantom. Compared to simulations, an advantage to this approach is that the acquired data contain noise characteristic of the scanner and protocol. Using this technique, a simple capillary phantom is employed to infer the quality of data for more clinically realistic tissue structures (e.g., crossing fiber tracts). A water-filled phantom containing capillary arrays was constructed to demonstrate this technique, which uses a DTI protocol with typical clinical parameters. Eigenvalues and fractional anisotropy were calculated for the initial prolate data. Data were adjusted to synthesize different apparent diffusion coefficient (ADC) spatial distributions, which were compared to theoretical and analytical models. RMS differences and volumetric overlap between expected and measured ADC distributions were quantified for all synthesized distributions. Differences between synthesized and actual distributions were discussed.