DSI Studio provides a tool to simulate diffusion MR images. The simulation function is located at "Diffusion MRI simulation" button in the main window. The interface asks for a b-table (a text file), mean diffusivity, SNR, FA, and crossing angles. The simulation generates .src file, which can be loaded into Matlab for further processing or used in diffusion MR reconstruction.
This simulation tool was used in two journal publications . The generated images can be downloaded.
The simulation model is a multiple tensor model :
where S0 is the b0 signal; f0, f1 ,and f2 are the volume fractions of the background diffusion, the first fiber, and the second fiber, respectively; D0, D1 and D2 are the corresponding diffusion tensors.
Rician noise is added to simulate the image noise pattern.
More detailed description can be found in the references .
The simulation iterates through each FA, crossing angle, and trial iterations, generate one slice for each combination. Each slice image has a dimension of 60-by-60 pixels, and the fibers phantom has a dimension of 64-by-64 pixels, located at the center of the images. The background region simulates the free water diffusion.
To generate the simulation images. DSI Studio needs a text file that records the b-values and gradient vectors. The formate of the text file is as follows.
where the first column is the b-value, and the rest of the columns stand for the gradient unit vector.
Exemplary tables can be found in the attachment at the bottom of this page.
The simulated images and analysis results of the paper are open to public inspection and use. You may use them for any purpose. For details about the simulation parameters, please refer to the PLOS One paper.
F-C. Yeh, V. J. Wedeen, & W-Y. I. Tseng. “Generalized Q-Sampling Imaging.” IEEE Transaction on Medical Imaging, 29:1626-1635, 2010.(pdf)
F-C. Yeh and W-Y. I. Tseng, “Sparse Solution of Fiber Orientation Distribution Function by Diffusion Decomposition”, PLOS One, 2013, 8(10): e75747.(pdf)