At the beginning, there was sound, and yet the world was silent until millions of years later, when animals evolved the ability to hear. This allowed animals to locate sound sources and hunt in total darkness. In time, sounds also enabled animals to communicate socially to cooperate as well as compete with each other.

Despite such humble beginnings, vocal communication is now widespread among many taxa and human language is considered to be one of defining characteristics of our species. Yet our understanding of how the brain enables vocal communication, especially in the mammal, remains quire rudimentary.

In this paper (Okobi, Banerjee et al, 2019), we attempted to lay down a roadmap to investigate neural mechanisms that enable vocal interactions. Our journey began with the search for a new model system, and the cloud forests of Costa Rica provided the perfect backdrop.

Cloud forests of Costa Rica are home to a species of highly vocal rodents that our collaborator Steven M. Phelps, PhD, at the University of Texas at Austin has been studying for more than a decade. These rodents, known as Alston’s singing mice (Scotinomys teguina), produce a highly stereotypic sequence of notes strung together in a “song”. Male mice are thought to use these songs for aggression as well as mate selection. As the name suggests, these rodents sing spontaneous “solo” songs as well as “duets” with another male. The latter is known as counter-singing and resembles the back-and-forth structure (turn-taking) of human conversation.

By quantitative behavioral analysis, we discovered that conversation between males of S. teguina follows strict rules. Specifically, a vocal partner often interrupts songs of newly arrived males. In response, the interrupted male goes silent, listens as the other mouse completes his song, and within a fraction of a second initiates a new song in response. These mice are polite enough to take turns to sing!

Vocal interactions in a small rodent that are not too shy to sing in a lab allowed us to investigate the neural mechanisms that generate vocal communication in the mammalian brain. By electrical microstimulation of anterolateral motor cortex, we could cause reliable flexion of vocal musculature. Electrical stimulation of this part of motor cortex, which we refer to as the orofacial motor cortex (OMC), even disrupted ongoing singing behavior. Moreover, mild focal cooling of OMC slowed down the progression of the motor sequence, resulting in considerably longer songs. But what about conversation? Silencing OMC activity with drugs, we found that the animals could still sing their songs but could no longer participate in vocal interactions.

All of these perturbation experiments taken together suggest a hierarchical organization in which song production is mediated by subcortical structures while the processes capable of coordinating the rapid sensorimotor coupling required during conversation are controlled cortically. To the best of our knowledge, this is the first direct demonstration of cortical control of vocal interactions in a mammal.

There is growing emphasis on using naturalistic behaviors in neuroscience. The “singing mouse” model system is a rare combination of natural behavior and experimental tractability in terms of underlying neural circuits analysis. Understanding the neural basis of vocal behaviors in a mammal may not only help us better understand human communication, but it also offers an alternative approach for investigating the neural mechanisms of perception, cognition, and action in an ethologically-relevant natural behavior.

—Arkarup Banerjee, PhD

Read the paper “Motor cortical control of vocal interaction in neotropical singing mice” in Science, published March 1, 2019.