The production of birdsong involves complex respiratory motor gestures shaped by precise coordination of neural and muscular systems. In this talk, I’ll present a framework for understanding birdsong as a sequence of simpler motor instructions, each generated by a minimal excitable system. Using air sac pressure recordings from singing canaries, we model individual syllables as sequences of transient responses of a two-dimensional Wilson–Cowan-type system. By fitting these transients to observed pressure patterns using a differential evolution algorithm, we obtain reconstructions of the motor patterns underlying song with high fidelity. We then apply unsupervised dimensionality reduction and clustering to the extracted transients, identifying a compact set of shared motor primitives across birds with different vocal learning histories. This suggests that birdsong is built from a reusable repertoire of dynamical modules, shedding light on the structure of learned motor behavior and its neural underpinnings. I’ll discuss how this approach opens new avenues for studying motor control and learning in a broad range of behaviors. [1] Agustín Carpio Andrada and Gabriel B. Mindlin. “Decomposition of respiratory motor patterns during birdsong production in terms of excitable transients.” Chaos, Solitons & Fractals, 2025