A novel white matter fibre tracking algorithm using probabilistic tractography and average curves

Abstract

This paper presents a novel white matter fibre tractography approach using average curves of probabilistic fibre tracking measures. We compute "representative" curves from the original probabilistic curve-set using two different averaging methods. These typical curves overcome a number of the limitations of deterministic and probabilistic approaches. They produce strong connections to every anatomically distinct fibre tract from a seed point and also convey important information about the underlying probability distribution. A new clustering algorithm is employed to separate fibres into branches before applying averaging methods. The performance of the technique is verified on a wide range of seed points using a phantom dataset and an in vivo dataset. © 2010 Springer-Verlag.