Nonlinear analysis of irregular animal vocalizations

Area of Science:

• Bioacoustics
• Animal Communication
• Nonlinear Dynamics

Background:

• Animal vocalizations exhibit a wide spectrum of complexity, from periodic sounds to atonal noise.
• Nonlinear dynamics, including limit cycles and chaos, are increasingly observed in animal sounds, suggesting functional roles.
• Understanding these dynamics offers insights into neural control versus biomechanical properties of vocal production.

Purpose of the Study:

• To quantify the degree of nonlinearity in animal vocalizations using the deterministic versus stochastic (DVS) prediction technique.
• To investigate the relationship between vocal complexity, harmonic content, and nonlinear dynamics in macaque screams, piglet screams, and dog barks.

Main Methods:

• Application of the deterministic versus stochastic (DVS) prediction technique to analyze macaque screams, piglet screams, and dog barks.
• Quantification of nonlinearity through the low-dimensional nonlinearity measure (LNM), derived from DVS prediction.
• Correlation analysis between harmonic-to-noise ratio (HNR) and LNM in dog barks.

Main Results:

• DVS prediction effectively quantified nonlinearity in macaque screams, piglet screams, and dog barks.
• Vocalizations with strong harmonic components showed higher deterministic nonlinearity compared to stochastic prediction.
• A positive correlation was found between harmonic-to-noise ratio (HNR) and the low-dimensional nonlinearity measure (LNM) in dog barks.

Conclusions:

• Nonlinear analysis, particularly DVS prediction, is valuable for studying animal vocalizations with significant harmonic content or low-dimensional chaos.
• The low-dimensional nonlinearity measure (LNM) can indicate the presence of low-dimensional attractors in animal sounds.
• Findings suggest that nonlinear dynamics play a role in the acoustic structure and potentially the communication function of animal vocalizations.