Abstract: Neurons coordinate patterns of muscle activity to produce an astonishing variety of behaviors. However, the biological and computational bases of sensorimotor control remain mysterious, in part due to a lack of experimental hardware and computational frameworks for examining motor signals. To address these challenges, my group combines physiological, computational, and engineering approaches to understand motor control across species and behaviors. My talk will provide an overview of three projects. First, physiological and computational studies of vocal production in songbirds reveal that neurons in the motor system employ millisecond-resolution spike timing codes to regulate vocal behavior, highlighting the need to examine spiking codes across cortical, basal ganglia, and spinal motor circuits. Second, to examine spiking codes across motor circuits, we have developed electrode arrays for examining spiking patterns in muscle tissue during natural behaviors. These “Myomatrix” arrays reveal the spatiotemporal structure of muscle activity at motor-unit resolution across effectors (forelimb, trunk, orofacial, respiratory, and vocal muscles) and species – including humans – during unconstrained behavior. Third, in-progress studies examining muscle spike trains in mice (locomotion) and monkeys (reaching movements) reveal how complex patterns of motor unit activity shape skilled forelimb control.https://scholarblogs.emory.edu/soberlab/