Topological data analysis captures complex behavioral dynamics during naturalistic social interaction between...

Area of Science:

• Ethology
• Complex Systems Science
• Computational Biology

Background:

• Understanding naturalistic social behavior requires analyzing complex, moment-to-moment dynamics.
• Traditional time-aggregated analysis methods are insufficient for capturing these fine-grained interactions.
• Topological Data Analysis (TDA) offers novel tools for characterizing complex time-series dynamics.

Purpose of the Study:

• To adapt and apply Temporal Mapper, a TDA tool, for analyzing animal postural dynamics during naturalistic social interaction.
• To characterize the complex temporal dynamics of ferret social behavior using network analysis.

Main Methods:

• Utilized an existing dataset of domestic ferret (Mustela putorius furo) social interactions.
• Employed posture estimation to track animal movements during naturalistic interactions.
• Applied Temporal Mapper to represent postural dynamics as transition networks.

Main Results:

• Posture states were found to be smaller and more widespread during active social interaction compared to non-social activities.
• The sequence length of postural states before behavioral transitions was more consistent during social interaction.
• Social activity is characterized by a broad range of unstable postural states in consistent sequences.

Conclusions:

• Temporal Mapper effectively captures rich animal postural dynamics in naturalistic social settings.
• Social interaction in ferrets involves dynamic, consistently sequenced postural shifts.
• This TDA approach offers a powerful method for analyzing complex behavioral dynamics across species and paradigms.