Abstract
In this study, I developed a novel free-moving paradigm to investigate cross-modal associative learning in bumble bees. As opposed to commonly used associative odour learning experiments that restrain the bees and/or use the proboscis extension reflex (PER), this paradigm allows readout for odour identity that is independent of odour valence. I present a series of experiments to refine the apparatus and the experimental protocols to support cross-modal associative learning for bees. The initial classical condition experiment using colour-sugar reward associations did not demonstrate associative learning. Only the bees with prior exposure to colour stimuli learnt. Bees also learnt to associate a location (left or right side of the arena) during location-reward learning, and they succeeded during a final cross-modal associative learning paradigm to go to a certain location after smelling a certain odour. Although there was no significant outcome during the test trials in all experiments, successful learning was visible during the training when the sugar reward was presented – but hidden behind a wall from the bee when it made the decision. This makes it unlikely that bees made the correct choice because they could see or smell the sugar reward. Although the experiments did not demonstrate a significant associative learning outcome, the study established an important foundation for future research on insect associative learning. The findings from this study contribute to the development of assessment tools for animal cognitive research.