Vocal learning through imitation underlies both human speech acquisition and song acquisition in oscine birds; both processes depend on auditory information. In songbirds, a specialized forebrain pathway is responsible for producing the learned temporal and acoustic features of vocalizations, and auditory input reaches every level of this pathway. Nucleus robustus archistriatalis (RA) is the source of the final common output from this pathway; RA is topographically organized into subregions that control the syringeal, respiratory and other effectors involved in vocal production. The acoustic features of learned vocalizations are primarily produced by specific patterns and combinations of syringeal muscle activity, while the overall temporal structure is primarily under respiratory control. In RA, and other vocal control structures, the individual bird's own learned song (BOS) is the most effective stimulus for eliciting auditory responses. Some neurons are 'combination-selective' in that they respond maximally to stimuli consisting of sequences of syllables from song. The recording sites that respond selectively to BOS tend to be located in more ventral parts of RA, the subregion that projects to motor neurons controlling syringeal muscles. These observations do not distinguish between motor feedback and perceptual hypotheses about the function of auditory responses in vocal motor pathways but are consistent with the idea that such responses may reflect a specific pattern of interaction between sensory and motor events that reflects vocal learning.