Stimuli that fit: a biology-aligned approach to numerical cognition research

Understanding the mechanisms underlying animal cognition requires experimental designs and analyses that respect the biological and perceptual constraints of the species being studied. We address an ongoing debate regarding the role of visual spatial frequency in numerical cognition studies, using honeybees (Apis mellifera) as a model system. Prior analyses that excluded the low spatial frequencies known to be fundamental to how these insects use their spatial vision may have introduced spurious correlations between visual features and numerical cues. By recalculating stimulus properties while incorporating the biologically relevant spectrum of spatial frequencies perceivable by a bee, our results suggest that numerical cognition in honeybees is not driven by spatial frequency but instead reflects true numerosity processing. This work highlights the necessity of aligning experimental design and analysis with the sensory and perceptual abilities of the studied species to ensure biologically meaningful conclusions. To support this goal, we provide a general framework for designing experiments that respect the psychophysical constraints of animal models in numerical cognition studies.