Data for "Automated video tracking and flight analysis show how bumblebees solve a pattern discrimination task using active vision"

Active vision, the sensory-motor process through which animals dynamically adjust visual input to sample and prioritise relevant information via photoreceptors, eyes, head, and body movements, is well-documented across species. In small-brained animals like insects, where parallel processing may be limited, active vision for sequential acquisition of visual features might be even more important. We investigated how bumblebees use active vision to distinguish between two visual patterns: a multiplication sign and its 45° rotated variant, a plus sign. By allowing bees to freely inspect patterns, we analysed their flight paths, inspection times, velocities and regions of focus through high-speed videography. We observed that bees tended to inspect only a small region of each pattern, with a preference for lower and left-side sections, before accurately accepting target or rejecting distractor patterns. The specific pattern areas scanned differed between the plus and multiplication signs, yet flight behaviour remained consistent and specific to each pattern, regardless of whether the pattern was rewarding or punishing. Transfer tests showed that bees could generalise their pattern recognition to partial cues, maintaining scanning strategies and selective attention to learned regions. These findings highlight active vision as a crucial aspect of bumblebees' visual processing, where selective scanning behaviours during flight enhance discrimination accuracy and enable efficient environmental analysis and visual encoding.