Multi-dimensional tuning in motor cortical neurons during active behavior

A region within songbird cortex, AId (dorsal intermediate arcopallium), is functionally analogous to motor cortex in mammals and has been implicated in vocal learning during development. AId thus serves as a powerful model for investigating motor cortical contributions to developmental skill learning. We made extracellular recordings in AId of freely behaving juvenile zebra finches and evaluated neural activity during diverse motor behaviors throughout entire recording sessions, including song production as well as hopping, pecking, preening, fluff-ups, beak interactions with cage objects, scratching, and stretching. A large population of single neurons showed significant modulation of activity during singing relative to quiescence. In addition, AId neurons demonstrated heterogeneous response patterns that were evoked during multiple movements, with single neurons often demonstrating excitation during one movement type and suppression during another. Lesions of AId do not disrupt vocal motor output or impair generic movements, suggesting that the responses observed during active behavior do not reflect direct motor drive. Consistent with this idea, we found that some AId neurons showed differential activity during pecking movements depending on the context in which pecks occurred, suggesting that AId circuitry encodes diverse inputs beyond generic motor parameters. Moreover, we found evidence of neurons that did not respond during discrete movements but were nonetheless modulated during active behavioral states compared to quiescence. Taken together, our results support the idea that AId neurons are involved in sensorimotor integration of external sensory inputs and/or internal feedback cues to help modulate goal-directed behaviors. SIGNIFICANCE STATEMENTMotor cortex across taxa receives highly integrated, multi-modal information and has been implicated in both execution and acquisition of complex motor skills, yet studies of motor cortex typically employ restricted behavioral paradigms that target select movement parameters, preventing wider assessment of the diverse sensorimotor factors that can affect motor cortical activity. Recording in AId of freely behaving juvenile songbirds that are actively engaged in sensorimotor learning offers unique advantages for elucidating the functional role of motor cortical neurons. The results demonstrate that a diverse array of factors modulate motor cortical activity and lay important groundwork for future investigations of how multi-modal information is integrated in motor cortical regions to contribute to learning and execution of complex motor skills.