Meeting Abstract
Animals rely on information pooled across multiple sensory modalities to control locomotion. In effect, sensory feedback in behavior comprises numerous parallel feedback loops, collectively shaping the behavioral response. We show how these parallel pathways provide robustness via redundancy. These attributes, while beneficial to behavior, present experimental and analytical confounds in identifying the contributions of individual sensory pathways. Addressing this challenge, we present a novel experimental paradigm leveraging sensory conflict for the study of multisensory behaviors in animals. In this approach, animals are presented concurrent and uncorrelated stimuli across different modalities. In contrast to prevailing approaches, in which sensory pathways are isolated by means of inhibition or ablation, sensory conflict maintains the sensory machinery intact; a control theoretic analysis disentangles the contributions of sensory pathways from the ensemble. To demonstrate this paradigm, we examine the joint visual and mechanosensory control of flight in the hawkmoth. In the laboratory, freely flying moths feed from two-part, robotically actuated flowers. This allows independent control of the visual and mechanosensory (via proboscis) cues by motion of the flower facade and nectar spur respectively. As the flower moves side-to-side, moths follow to maintain their relative position. Empirical models (fit to input—output tracking data from an assay of sensory conflict conditions) suggest that each pathway alone would be sufficient to mediate behavior. As such, the pathways insure against damage or inhibition of the other. Extrapolating these models, we predict the behavioral response to a set of sensory isolation experiments (technically infeasible), and demonstrate that these experiments do not reliably reflect the contributions of parallel pathways.
