Long time scales, individual differences, and scale invariance in animal behavior

Area of Science:

• Animal behavior analysis
• Complex systems science
• Biophysics

Background:

• Analyzing animal behavior in natural contexts generates vast datasets with correlations across multiple timescales.
• Challenges include limited independent samples in single-animal data and distinguishing individual differences from temporal correlations.

Approach:

• Developed a novel analysis scheme to address data challenges in animal behavior studies.
• Applied this scheme to spontaneous walking behavior data from flies.
• Investigated correlations over time scales ranging from seconds to one hour.

Key Points:

• Evidence found for scale-invariant correlations in fly behavior over nearly three decades of time.
• Three distinct correlation measures consistently support a single underlying scaling field.
• The scaling field dimension was determined to be Δ = 0.180 ± 0.005.

Conclusions:

• The proposed analysis scheme effectively handles complexities in behavioral data, separating individual variation from temporal correlations.
• Scale-invariant correlations are a significant feature of spontaneous animal behavior, observable across a wide range of timescales.
• This research provides a new framework for understanding the fundamental principles governing animal movement and behavior.