Visual place learning by walking bumblebees in virtual reality

Central-place foragers, such as bumblebees, must navigate complex and ever-changing environments to locate food sources and return to their nests. Visual cues in the vicinity of the target play a crucial role in this process, particularly in distinguishing between otherwise visually similar but nutritionally distinct targets. In this study, we used a virtual reality (VR) trackball system and a free-walk paradigm to investigate how bumblebees utilize visual information for place-learning. We specifically examined how bees resolve conflicts between previously learned and currently available visual cues surrounding a target, how the perceived valence of a target changes through association with rewards or punishments, and how they distinguish between visually identical targets based on differences in surrounding visual features. We found that bumblebees prioritize spatial proximity, probably using motion parallax information over static image matching strategies. Furthermore, bees demonstrated flexibility in associating visual features with rewards or punishments by adapting to reversed valence pairings. However, potential intrinsic preferences for specific colours and patterns influenced learning success. These findings challenge traditional snapshot navigation models and suggest that walking bumblebees employ flexible strategies for solving spatial learning tasks in virtual and real environments.