A group of Mongolian gerbils. Image credit: H. Zell (CC BY-SA 3.0)
Every time you speak, the sounds coming out of your mouth may carry more meaning that you may have intended; they may reveal, for example, which country, city or even neighborhood you may be coming from. Indeed, the vocal patterns that humans use to communicate differ from one population to the next, creating an array of languages, dialects and accents.
Such diversity has also been identified in various social species across the animal kingdom. Naked mole rats, for instance, which live underground in complex societies, exhibit different ‘dialects’ depending on their group of origin. Yet studying the vocal patterns of animals has remained difficult, especially for species inhabiting burrows or other environments difficult to access.
Aiming to bypass these limitations, Peterson et al. adopted a ‘naturalistic’ approach that allowed them to capture the vocal calls of three families of Mongolian gerbils living undisturbed in enclosures that mimic features of their natural environment. These animals spend their lives underground in tight-knit families, with multiple groups often being in close proximity. Researchers have speculated that individuals may rely on vocal cues to identify whether they are part of the same colony, as they are often too far from each other to rely on sight or smell.
Over half a million vocalizations obtained continuously through the course of 20 days were analyzed using an artificial intelligence technique known as unsupervised machine learning. The analyses helped add new types of calls to the gerbil vocal repertoire, but also highlighted its complexity. In particular, they revealed that the animals could combine individual vocal elements into complex sequences. More importantly, this approach showed that gerbil families have vocal dialects that are stable across weeks, with each group displaying a preference for certain call types (i.e. words) and certain sequential patterns (i.e. phrases).
These findings demonstrate the benefits of the approach developed by Peterson et al. for the study of animal vocalizations. Going forward, they also suggest that the Mongolian gerbil could be used as an animal model to study the neural basis of vocal communication.
