Echoes of self: Understanding acoustic structure and informational content in common marmoset ( ) phee sequences

Kristin Meshinska,Judith M Burkart,Matthew BV Bell,Kaja Wierucka

DOI: https://doi.org/10.1101/2024.04.14.589400

2024-05-16

Abstract:Communication in social animals relies heavily on acoustic cues, yet many species possess limited vocal repertoires. To compensate, animals often produce vocalizations in sequences, potentially enhancing the diversity of transmitted information. However, the significance of repeated elements within these sequences remains poorly understood. Here, we investigated the spectro-temporal structure of elements within common marmoset ( ) phees, a long-distance contact call, often produced in sequences. Employing machine learning techniques (random forests) and linear mixed effects models, we explored how elements varied based on their position within sequences and assessed their ability to encode identity and sex information. Additionally, we examined similarities between elements occupying the same position in different sequences. Our results reveal structural differences both within and between sequences, with variations observed in phees at different positions within the same sequence, yet similarities found between first elements of distinct sequences. Notably, all phees encoded caller identity and sex information, with varying accuracy across positions within sequences, indicating a higher encoding of sex information at the beginning of the sequence and a greater emphasis on identity in subsequent elements. These findings suggest that repeated sequences may be functionally diverse structures, enriching the complexity of animal communication systems.

Animal Behavior and Cognition