Q-ball imaging has the ability to discriminate multiple intravoxel fiber populations within regions of complex white matter architecture. This information can be used for fiber tracking; however, diffusion MR is susceptible to noise and multiple other sources of uncertainty affecting the measured orientation of fiber bundles. The proposed residual bootstrap method utilizes a spherical harmonic representation for high angular resolution diffusion imaging (HARDI) data in order to estimate the uncertainty in multimodal q-ball reconstructions. The accuracy of the q-ball residual bootstrap technique was examined through simulation. The residual bootstrap method was then used in combination with q-ball imaging to construct a probabilistic streamline fiber tracking algorithm. The residual bootstrap q-ball fiber tracking algorithm is capable of following the corticospinal tract and corpus callosum through regions of crossing white matter tracts in the centrum semiovale. This fiber tracking algorithm is an improvement upon prior diffusion tensor methods and the q-ball data can be acquired in a clinically feasible time frame.