Skip to main content
Only a few animals, such as songbirds, whales, and dolphins, are known to be vocal learners, modifying the sequence or pitch of their sounds based on what they hear from other members of their species. New evidence suggests mice can be added to that list.
Investigator, The Rockefeller University
Only a few animals, such as songbirds, whales, and dolphins, are known to be vocal learners, modifying the sequence or pitch of their sounds based on what they hear from other members of their species. New evidence suggests mice can be added to that list.


Unlike human speech, most of the squeaks, calls, and songs produced in the animal kingdom do not have to be learned—the vast majority of animals rely on innate vocalizations for their acoustic communications. Only a few, such as songbirds, whales, and dolphins, are known to be vocal learners, modifying the sequence or pitch of their sounds based on what they hear from other members of their species. New evidence suggests another animal can be added to that list: mice.

In addition to their familiar squeaking, male mice whistle songs too high for the human ear to hear. These ultrasonic vocalizations—long considered innate—are thought to be used to attract females. A new study from Howard Hughes Medical Institute investigator Erich Jarvis finds that mice can change their songs based on those of other mice—and they have the brain connections to back up their vocal learning. The study is published in the journal PLoS ONE on October 10, 2012.

Jarvis’s lab at the Duke University Medical Center uses songbirds to study how the brain generates, perceives, and learns vocalizations. About seven years ago, he decided it would be good to compare some of these processes to those of mice, since he knew—or thought he did—that they didn’t learn their vocalizations.

The brains of both humans and songbirds have a strong, direct connection between the cortex in the forebrain, where the brain manages fine motor control, and the part of the brainstem that tells the vocal organ (the larynx in humans and the syrinx in birds) how to move. Neurons in the cortex reach their axons all the way to the brainstem. Jarvis says that when scientists have examined the brain of animals that are not vocal learners—including non-human primates and birds without vocal learning—this direct connection has been absent.

In keeping with these observations, scientists had thought that in mice, the connection between these parts of the brain would be indirect, making a couple of stops on the way. But when Jarvis looked for evidence of this in the scientific literature, he found that those brain circuits had not been mapped in mice. He and Gustavo Arriaga, a graduate student in his lab at the time and the first author of the new paper, decided this would be a good place to start.

The two scientists did not find what they expected. In their first experiments, they encouraged male mice to sing by dampening their bedding with female urine, then looked at their brains to see what areas were active. “Lo and behold, we found a highly localized cortical region,” Jarvis says—a spot in the cortex that was active in singing mice. “You don’t need fine motor control from this region of the forebrain for innate behaviors,” he says. “So that was the first surprise. Then it took off from there.”

When they used a tract-tracing protein to trace the path of neurons from the larynx to the brain, they found that there was a direct connection between the cortex and the part of the brainstem that controls the larynx. That connection was not as strong as the connection in songbirds, however.

Next, Jarvis and his colleagues put male mice with different songs together to see if they would change their tunes. The researchers had noticed that males from two mouse strains had different songs—the males in one strain sang consistently higher pitches than the other. They made 12 pairs, taking a male from each strain and putting the two together in a cage with a female, since males usually won’t sing without females around. Then they recorded and analyzed the acoustics of their ultrasonic songs. Over the next eight weeks, the males that sang the higher songs shifted the pitch of their songs downward. After eight weeks, all but one pair was singing songs that were much closer in pitch. The other pair’s songs had matched almost perfectly at six weeks, then split again.

Jarvis concludes that scientists had some misconceptions about vocal learning. They had thought the world could be divided into animals that can learn and change their vocalizations, like humans and parrots, and animals that can’t, like chimpanzees. Brain anatomy was thought to agree with that dichotomy. But the mouse seems to be intermediate, with some learning and a weak connection between the cortex and the part of the brainstem that moves the larynx. There may be a continuum from species that do a lot of vocal learning and those who make only innate sounds, Jarvis says.

“When it comes to something that people consider unique to humans, we tend to have this automatic thought that nobody else has it until proven otherwise,” Jarvis says. People used to think that humans were the only animals to use tools, for example, but now tool use has been found in many other animals, from crows to chimpanzees. “What it’s really telling me overall is, we’re not really in a black and white world; there are shades of gray in between and differences and diversity and variety in the biology,” Jarvis says.