The relevance of a right scale for sampling when studying high-resolution behavioral dynamics

Area of Science:

• Ethology
• Behavioral Ecology
• Quantitative Biology

Background:

• Small vertebrates, often used in behavioral studies, possess high metabolic rates and potentially enhanced temporal perception.
• The selection of appropriate temporal scales for evaluating behavioral dynamics remains under-examined.

Purpose of the Study:

• To investigate the temporal organization of behaviors at fine-grain resolutions (sampling interval ≤1s).
• To gain insights into behavioral dynamics and redefine how behavioral events are defined.
• To explore the impact of sampling resolution on the analysis of behavioral event durations.

Main Methods:

• Statistical analysis of high-resolution behavioral datasets from Japanese quail (Coturnix japonica).
• Examination of 17 defined behaviors, focusing on event durations and probability distributions.
• Comparison of data across different sampling resolutions to assess distortion effects.

Main Results:

• The majority of studied behaviors in Japanese quail predominantly last less than 300 milliseconds, with some events shorter than 70 milliseconds.
• Behaviors involving spatial displacement show distorted probability distributions and overestimated durations with insufficient sampling resolution (e.g., 1s).
• Behaviors without spatial displacement, like vigilance, retain non-Gaussian, power-law distributions indicative of long-term memory, irrespective of sampling resolution.

Conclusions:

• Accurate quantification of behavioral dynamics requires considering species-specific temporal scales and data distributions.
• Sampling resolution significantly impacts the analysis of behaviors with spatial displacement.
• A proposed hierarchical model offers a framework for defining diverse behavioral types and distributions.