Several methods for the reconstruction of white matter tracts from DT-MRI have been proposed in the literature. We present a novel approach for white matter tractography based on sequential importance sampling and resampling (SBR). By modelling all possible fibre paths originating from a starting point as a distribution, a probability can be assigned for each path based on its fitness to the measured DT data and its shape. We use SISR to sequentially build up a set of weighted samples representing this distribution. Connectivity can then be derived from these weighted samples. Method A fibre path x of length N originating from a point x0 can be defined as x = {xo, _.., XN}, where xi is represented in a state space. Let n be the target distribution for all paths originating in XO. Assume that the probability for each path, up to a normalizing constant, can be written as