In contrast to the more common Diffusion Tensor Imaging (DTI), High Angular Resolution Diffusion Imaging (HARDI) allows superior delineation of angular microstructures of brain white matter, and makes possible multiple-fiber modeling of each voxel for better characterization of brain connectivity. However, in the context of image registration, the question of how much information is needed for satisfactory alignment remains unanswered. Low order representation of the diffusivity information is generally more robust than the higher order representation, but the latter gives more information for correct fiber tract alignment. However, higher order representation, when naïvely utilized, might not necessarily be conducive to improving registration accuracy since similar structures with significant orientation differences prior to proper alignment might be mistakenly taken as non-matching structures. We propose in this paper a hierarchical spherical harmonics based registration algorithm which utilizes the wealth of information provided by HARDI in a more principled means. The image volumes are first registered using robust, relatively direction invariant features derived from the diffusion-attenuation profile, and their alignment is then refined using spherical harmonic (SH) representation of gradually increasing order. This progression of SH representation from non-directional, single-directional to multi-directional representation provides a systematic means of extracting directional information from the HARDI data. Experimental results show a significant increase in registration accuracy over a state-of-the-art DTI registration algorithm.