Limitations and requirements of diffusion tensor fiber tracking: An assessment using simulations

Diffusion tensor fiber tracking potentially can give information about in vivo brain connectivity. However, this technique is difficult to validate due to the lack of a gold standard. Fiber tracking reliability will depend on the quality of the data and on the robustness of the algorithms used. Information about the effects of various anatomical and image acquisition parameters on fiber tracking reliability may be used in the design of imaging sequences and of tracking algorithms. In this study, tracking was performed on two different simulated models to study the effects on tracking quality of SNR, anisotropy, curvature, fiber cross-section, background anisotropy, step size, and interpolation. Tracking was also performed on volunteer data to assess the relevance of the simulations to real data. Our results show that, in general, tracking with high SNR and high anisotropy using interpolation and a low step size gives the most reliable results. Partial volume effects are shown to have a detrimental effect when the background is anisotropic and when tracking narrow fibers. The results derived from real data show similar trends and thus support the findings of the simulations. These simulations may therefore help to determine which structures can be tracked for a given image quality. (C) 2002 Wiley-Liss, Inc.