Multi-scale inference of interaction rules in animal groups using Bayesian model selection.

Area of Science:

• Collective animal behavior
• Biophysics
• Complex systems modeling

Background:

• Inferring animal interaction rules often uses large-scale system behavior analysis.
• Fine-scale animal movements offer a more detailed approach to validating interaction rules.

Purpose of the Study:

• To compare various models of collective motion using Bayesian methodology.
• To analyze the collective motion of glass prawns (Paratya australiensis) and their phase transition behavior.

Main Methods:

• Bayesian methodology for model comparison.
• Fitting mean-field, spatial Markovian, and non-Markovian models to prawn collective motion data.
• Analyzing fine-scale interactions and global dynamics.

Main Results:

• Glass prawns exhibit a density-dependent phase transition to collective motion.
• Mean-field models capture global dynamics but not fine-scale interactions.
• Markovian models capture fine-scale interactions but not global dynamics.
• Non-Markovian models accurately reproduce both fine-scale and global dynamics.

Conclusions:

• Prawn movement is influenced by recent past interactions, not just current neighbors.
• Realistic self-propelled particle models for collective motion should incorporate memory and non-Markovian effects.