Go with the flow : visually mediated flight control in bumblebees

Abstract: Despite their small brains and tiny eyes, flying insects are capable of detecting and avoiding collisions with moving obstacles, and with remarkable precision they navigate through environments of different complexity. For this thesis, I have investigated how bumblebees use the pattern of apparent image motion that is generated in their eyes as they move through the world (known as optic flow), in order to control flight. I analysed the speed and position of bumblebee (Bombus terrestris) flight trajectories as they negotiated arenas of different dimensions and visual complexity. I also investigated the impact of optic flow on bumblebee learning flights, a special kind of flight designed to memorise the location of the nest or a newly discovered food source. The general aim of my research has been to understand how flying insects use vision to actively control their flight. The viewing angle at which optic flow is measured has important consequences for flight in densely cluttered environments, where timely control of position and speed are necessary for effective collision avoidance. I therefore investigated when, and how, bumblebees respond to sudden changes in the magnitude of optic flow. My results reveal that the visual region over which bumblebees measure optic flow is determined by the location in the frontal visual field where they experience the maximum magnitude of translational optic flow. This strategy ensures that bumblebees regulate their position and speed according to the nearest obstacles, allowing them to maximise flight efficiency and to minimise the risk of collision. My results further demonstrate that, when flying in narrow spaces, bumblebees use optic flow information from nearby surfaces in the lateral visual field to control flight, while in more open spaces they rely primarily on optic flow cues from the ventral field of view. This result strengthens the finding that bumblebees measure optic flow for flight control flexibly in their visual field, depending on where the maximum magnitude of translational optic flow occurs. It also adds another dimension to it by suggesting that bumblebees respond to optic flow cues in the ventral visual field if the magnitude is higher there than in the lateral visual field. Thus, the ability to flexibly use the surrounding optic flow field is of great importance when it comes to the control of cruising flight. For this thesis I also investigated the impact of ventral and panoramic optic flow on the control of learning flights in bumblebees. The results show that the presence of ventral optic flow is important for enabling bumblebees to perform well-controlled learning flights. Whether panoramic optic flow cues are present or not does not strongly affect the overall structure of the learning flight, although these cues might still be involved in fine-scale flight control. Finally, I found that, when the availability of ventral optic flow is limited to certain heights, bumblebees appear to adjust their flight parameters to maintain the perception of ventral optic flow cues. In summary, the results compiled in this thesis contribute to a better understanding of how insects use visual information to control their flight. Among other findings, my results emphasize the importance of a being able to flexibly measure optic flow in different parts of the visual field, something that enhances bees’ ability to avoid collisions.