Behavioral sequences across multiple animal species in the wild share common structural features

Area of Science:

• Animal behavior
• Behavioral ecology
• Ethology

Background:

• Animal behavior is often analyzed as sequences of discrete activity bouts.
• Previous studies, mainly in invertebrates, indicated complex temporal structures like multiple timescales and long-range memory.
• Generalizability of these findings to diverse taxa and natural environments remained uncertain.

Purpose of the Study:

• To investigate the statistical structure of behavioral sequences in wild social mammals.
• To determine if patterns observed in invertebrates apply to other species in natural settings.
• To identify common principles governing behavioral sequence dynamics across species.

Main Methods:

• Utilized accelerometer data to infer behavioral states in meerkats, white-nosed coatis, and spotted hyenas.
• Analyzed behavioral sequences to identify statistical properties like hazard functions and bout duration distributions.
• Applied power-law analysis to assess the predictability of future behavioral states.

Main Results:

• Behavioral sequences exhibited consistent structuring across all species, individuals, and behavioral states.
• Bout durations showed heavy-tailed distributions and decreasing hazard functions.
• Predictability of future states decreased with time, following a truncated power-law with similar exponents across species.

Conclusions:

• Consistent statistical principles govern behavioral sequences in wild social mammals, suggesting widespread phenomena.
• Potential explanations include environmental nonstationarity, self-reinforcement, and behavioral hierarchy.
• These findings point to fundamental properties of behavioral dynamics driving convergent patterns in nature.