Pattern recognition in honeybees: multidimensional scaling reveals a city-block metric

Area of Science:

• Animal behavior
• Cognitive neuroscience
• Sensory perception

Background:

• Understanding insect visual perception is crucial for deciphering cognitive processes in non-human animals.
• Previous research suggests various modes of pattern recognition across species.

Purpose of the Study:

• To investigate how bees (Apis mellifera) discriminate between complex ring patterns based on number and size.
• To determine the mathematical model that best describes the bees' perceptual metric for these patterns.

Main Methods:

• Training bees to differentiate ring patterns varying in ring number and size.
• Conducting transfer tests to assess the independence of size and pattern discrimination.
• Applying multidimensional scaling with Minkowski metrics to model the perceptual data.

Main Results:

• Bee size discrimination was largely independent of pattern type, and vice versa.
• The city-block metric, not the Euclidean metric, best described the bees' perception of ring-pattern dissimilarity.
• Bees appear to derive overall pattern dissimilarity by additively combining component differences.

Conclusions:

• Bees utilize an analytic processing mode for complex visual patterns, summing individual feature differences.
• This contrasts with holistic processing, where the overall pattern is perceived as a unified whole.
• The findings offer insights into the evolution and mechanisms of visual perception across different species.