Social information in equine movement gestalts

Area of Science:

• Ethology and Animal Behavior
• Evolutionary Biology
• Machine Learning in Biology

Background:

• Signal evolution theories propose that behaviors can gain communicative value detached from original functions.
• This transition may be constrained by signaling costs, such as predation risk.
• Understanding social information transfer in group-living animals is crucial.

Purpose of the Study:

• To investigate if kinematic movement patterns in horses convey rich social information.
• To explore the potential of machine learning in analyzing animal movement for behavioral insights.
• To examine how movement gestalts relate to signal evolution theories.

Main Methods:

• Collected inertial movement sensing data from free-ranging horses (Equus caballus) at multiple locations.
• Applied a machine learning algorithm to extract kinematic gestalt profiles from movement data.
• Analyzed movement patterns to identify individual characteristics and social information.

Main Results:

• Successfully extracted multi-layered information from horse movement patterns.
• Discriminated between individual horses based on identity, breed, sex, and personality traits from movement data.
• Demonstrated that movement gestalts provide a rich source of passively available social information.

Conclusions:

• Domestic horses, and likely other group-living animals, can access substantial social information through movement gestalts.
• Movement patterns serve as a reliable, passively transmitted source of information about conspecifics.
• Findings contribute to understanding signal evolution, particularly concerning the communicative value of movement.