An error analysis of probabilistic fibre tracking methods: average curves optimization

Fibre tractography using diffusion tensor imaging is a promising method for estimating the pathways of white matter tracts in the human brain. The success of fibre tracking methods ultimately depends upon the accuracy of the fibre tracking algorithms and the quality of the data. Uncertainty and its representation have an important role to play in fibre tractography methods to infer useful information from real world noisy diffusion weighted data. Probabilistic fibre tracking approaches have received considerable interest recently for resolving orientational uncertainties. In this study, an average curves approach was used to investigate the impact of SNR and tensor field geometry on the accuracy of three different types of probabilistic tracking algorithms. The accuracy was assessed using simulated data and a range of tract geometries. The average curves representations were employed to represent the optimal fibre path of probabilistic tracking curves. The results are compared with streamline tracking on both simulated and in vivo data.