Turn-taking in mouse vocalization

Carl Zimmer, These Mice sing to One Another – Politely, NYTimes, Feb 28, 2019.

It was once believed that monkeys and apes were not able to exert voluntary cortical control over their vocalizations; that was the case when I published my book on music, Beethoven's Anvil, in 2001. We've since learned that's not true. Now, it seems, (some) mice also have cortical control over vocalization.

Alston’s singing mice sometimes belt out a song when they’re alone, but they’re especially vocal when other mice are around. Males sing as a way to fight over territory with other males, and both males and females sing to one another during courtship.

Working with Steven M. Phelps, a biologist at the University of Texas at Austin, Dr. Long set up a home for the mice in his lab to study their brains.

“They’re kind of divas,” he said. “They need exercise equipment in their cages and specialized diets. But they thrive here.”

One day, Andrew M. Matheson, one of Dr. Long’s graduate students, noticed something odd about two male mice in neighboring cages. Instead of singing over each other, they sounded like they were having a conversation.

Dr. Long and his colleagues eventually discovered that Mr. Matheson’s hunch was correct. The singing mice never overlapped: Each mouse would wait for the other to stop, and then start up within a fraction of a second. [...]

So the researchers began probing the brains of the mice, searching for the neurons that led them to be “polite” raconteurs.

In one experiment, the researchers cooled down patches of mouse brain by a few degrees, slowing the neurons. One patch in the mouse cortex is essential for controlling their singing, the scientists found. If this section is cooled, the mouse sings extended songs, adding on extra notes.

The researchers also injected nerve-blocking drugs into this brain patch and then played a recording of another male. Drugged males often failed to sing back. And when they did, they were slow to begin, taking seconds to start their own song.

Dr. Long thinks this region of the mouse cortex is crucial to the mice’s special communication. “We think of it as a conductor,” he said. “It allows the animals to sing in this turn-taking way.”

How long is that fraction of a second?

It's this turn-taking that's interesting. Human turn-taking is precisely timed and implies that the interlocutors are synchronized to a common pulse. Is that the case for these Alston’s singing mice? 

I once emailed the late Walter Freeman about human turn-taking. I was interested in how quickly one turn followed upon another, on the order of 10s of milliseconds. I was curious about the conduction speed in the nervous system, which is much slower than electrical signals in wires. Was it slow enough that instant turn-taking was possible only if one's vocalization was initiated before the interlocutor had stopped talking? That is, you can't take the cessation of speech as a cue for initiating your reply. The reply has to be initiated a fraction earlier. Freeman replied that, yes, that's how it would have to be.

Human brains and vocal apparatus are larger than those of mice (think, e.g., of breath control) and so conduction times are longer. But perhaps those mice are still subject to the same limitation.