Individual behavioral type captured by a Bayesian model comparison of cap making by sponge crabs

Area of Science:

• Behavioral Ecology
• Statistical Modeling
• Animal Behavior

Background:

• Animal personality is shaped by environmental interactions, necessitating methods to quantify individual behavioral differences.
• Understanding individual behavioral types is crucial for studying animal personality.
• The sponge crab (Lauridromia dehaani) provides a model for studying repeated behaviors like cap construction.

Purpose of the Study:

• To propose and validate a statistical method for quantifying individual behavioral types in animal behavior studies.
• To test the hypothesis that individual sponge crabs exhibit unique behavioral types in their cap-making behavior.
• To assess the appropriateness of assuming 'individual' behavior in repeatedly measured data.

Main Methods:

• Conducted behavioral experiments with sponge crabs using artificial sponges of varying sizes.
• Employed a novel Bayesian model-based comparison method, including hierarchical models, to analyze behavioral data.
• Utilized the widely applicable information criterion (WAIC) for model comparison from a predictive standpoint.

Main Results:

• Sponge crabs smaller than approximately 9 cm were observed to construct caps.
• Body size significantly influenced sponge choice, cap trimming, and cavity size, but not latency.
• The study successfully captured 'behavioral type' as a probabilistic structure within the behavioral data.

Conclusions:

• The developed statistical approach effectively quantifies individual behavioral types, confirming their existence in sponge crabs.
• Body size is a key factor influencing cap-making behavior in this species.
• The statistical methodology is broadly applicable to diverse data types (behavioral, physiological, morphological) for detecting underlying group structures.