Learned Birdsong and the Neurobiology of Human Language

The trait of vocal learning is rare, and it is a critical behavioral substrate for song in song learning birds and spoken language in humans. A common feature of species that have this trait (songbirds, parrots, hummingbirds, and humans) is that they have forebrain to brainstem systems for vocal control, whereas those that produce only innate sounds have only the brainstem vocal system. We have found that the vocal learning systems of song-learning birds are embedded within a motor system also present in vocal non-learning birds but that is involved limb and body movements. The song-learning and adjacent motor systems share many features in common, including motor-driven gene expression and connectivity into two sub-networks – an anterior pathway that is necessary for song learning and a posterior pathway that is necessary for song production. Comparative analyses with mammals indicate parallels with motor learning pathways and the spoken language brain systems in humans, respectively. To explain these findings, I propose a motor theory for the origin of vocal learning, where ancient brain systems used to control movement and motor learning gave rise to brain systems to learn and produce song and spoken language. I propose that the pre-existing system is a fundamental design of the vertebrate brain, consisting of the two motor sub-pathways (anterior and posterior), which during embryonic development form parallels systems to control different muscle groups that are innervated by sensory systems for feedback control of different motor behaviors. When vocal learning evolves, this pre-existing motor system is potentially duplicated and then connected to muscles of the vocal organ (syrinx in birds or larynx in humans) to control a specialized form of learned movement control - song and speech, with specialized regulation of genes involved in neural connectivity and neural protection. In this manner, the evolution of brain pathways for vocal learning may have evolved independently of a common ancestor, but dependent on a pre-existing motor learning pathway scaffold that then diverged.