Bumblebees display stimulus-specific persistence behaviour after being trained on delayed reinforcement

In uncertain environments, animals often face the challenge of deciding whether to stay with their current foraging option or leave to pursue the next opportunity. The voluntary decision to persist at a location or with one option is a critical cognitive ability in animal temporal decision-making. Little is known about whether foraging insects form temporal expectations of reward and how these expectations affect their learning and rapid, short-term foraging decisions. Here, we trained bumblebees on a simple colour discrimination task whereby they entered different opaque tunnels surrounded by coloured discs (artificial flowers) and received reinforcement (appetitive sugar water or aversive quinine solution depending on flower colour). One group received reinforcement immediately and the other after a variable delay (0–3 s). We then recorded how long bees were willing to wait/persist when reinforcement was delayed indefinitely. Bumblebees trained with delays voluntarily stayed in tunnels longer than bees trained without delays. Delay-trained bees also waited/persisted longer after choosing the reward-associated flower compared to the punishment-associated flower, suggesting stimulus-specific temporal associations. Strikingly, while training with delayed reinforcement did not affect colour discrimination, it appeared to facilitate the generalisation of temporal associations to ambiguous stimuli in bumblebees. Our findings suggest that bumblebees can be trained to form temporal expectations, and that these expectations can be incorporated into their decision-making processes, highlighting bumblebees’ cognitive flexibility in temporal information usage.