Complete neuroanatomy of dragonfly wings suggests direct sensing of aeroelastic deformations

Can mechanosensors in animal wings allow reconstruction of the wing aeroelastic states? Little is known about 15 how flying animals utilize wing mechanosensation to monitor the dynamic state of their highly deformable wings. 16 Odonata, dragonflies and damselflies, are a basal lineage of flying insects with excellent flight performance, and 17 their wing mechanics have been studied extensively. Here, we present a comprehensive map of the wing sensory 18 system for two Odonata species, including both the external sensor morphologies and internal neuroanatomy. 19 We identified eight morphological classes of sensors; most were mechanosensors innervated by a single neuron. 20 Their innervation patterns and morphologies minimize axon length and allow morphological latency 21 compensation. We further mapped the major veins of another 13 Odonata species across 10 families and 22 identified consistent sensor distribution patterns, with sensor count scaling with wing length. Finally, we 23 constructed a high-fidelity finite element model of a dragonfly wing for structural analysis. Our dynamic loading 24 simulations revealed features of the strain fields that wing sensor arrays could detect to encode different wing 25 deformation states. Taken together, this work marks the first step toward an integrated understanding of fly-by-26 feel control in animal flight.