Neural circuits for motor exploration and learning

Most human motor behaviors, such as speech or a piano concerto, are not innately programmed but are learned through a gradual process of trial and error. Learning requires exploration and the evaluation of subsequent performance. How are these processes implemented in the brain, and how do they go awry in disease? Songbirds provide a powerful model system to address these questions. Before they develop mature songs, young songbirds ‘babble’—producing highly variable vocalizations that underlie a process of trial-and-error. To investigate the neural mechanisms underlying exploration during learning, I recorded and manipulated neural activity in the basal ganglia, thalamus, and motor cortex-like nuclei in singing juvenile birds. Though the thalamus is traditionally considered a relay between the basal ganglia and cortex, I found that the thalamus, and not its inputs from the BG, was required for vocal variability during babbling. Meanwhile, the BG were required for song learning over time. Currently, my lab is pursuing three specific aims to study precisely how the BG support song learning. First, we are combining neural recordings with acoustic biofeedback to understand how neurons encode how ‘good’ (or ‘bad’) the song sounds. Second, we are developing optogenetic techniques to manipulate the activity of specific neuron subtypes in freely moving, singing birds. Finally, we are developing novel technologies to massively expand the number of neurons we can record simultaneously in singing birds. Basal ganglia circuits in songbirds and humans are very similar, and our overarching goal is to discover basic functions in a tractable model system that may ultimately provide insights into BG diseases such as Parkinson’s, Huntington’s and dystonia.