Statistical Olfactory Learning in Honey Bees

Statistical learning is a key mechanism for detecting regularities in sensory inputs ^1,2 . Among its functions is the ability to extract regularities from sequences (of sounds, objects, odors, etc.) ³ , enabling species to predict future events and guide behavior. This capacity has been demonstrated in vertebrates ² , including human infants ⁴ , non-human primates ⁵ , and birds ⁶ . However, the minimum computational architecture required for statistical learning remains unclear. To address this issue, we studied statistical learning in the honey bee ( Apis mellifera ), an invertebrate model for learning studies ⁷ . We show that bees learn and recall the temporal structure of sequences of odorants, suggesting that statistical learning is a fundamental component of a conserved cognitive toolkit present even in invertebrates.