University of Southern California

USC Neuroscience

Mapping the Brain for Vocal Learning
By Susan Andrews

Loud and lively singers, some liken the sound of finches to toy trumpets — ka-ching, beep-beep, oi, a-ha, or da-de-da!

Both male and female zebra finches sound off in chattering trills and calls. However, it is only the male of the species that sings. The father passes along his song to his sons primarily for mating purposes — the song is used to court discerning females. Unlike the nightingale male, who boasts a rich repertoire of 300 love songs, the zebra finch only has one song of between five and nine syllables.

As in humans, the vocal learning of songbirds is based on first memorizing vocal sounds from an adult male tutor (i.e., the father, in zebra finches). Then the feedback of self-produced vocalizations is used to fine-tune vocalizations until the output matches the neural memory of those sounds — essentially what begins as babble turns into understandable language.

According to Sarah Bottjer, professor of Biological Sciences and Psychology, songbirds are a great model system for understanding auditory learning because they can be investigated in ways that cannot be achieved in humans.

Mammals and birds have comparable specialized forebrain regions for vocalizing — songs for birds and speech for humans. The neural circuits that underlie vocal learning in songbirds fall into one of two functional categories: a basic motor pathway for vocal production, and a pathway through the basal ganglia in juveniles for motor performance and cognition. The basal ganglia circuit is located in the part of the brain that becomes dysfunctional in humans causing Parkinson’s, Huntington’s and other diseases.

“For vocal learning in songbirds that circuit has a special function that is very important in the initial phases of vocal learning, but we have not been able to ascribe specific functions to different parts of that neural circuit until now,” Bottjer said. “As is true with both songbirds and humans, there is a critical juncture when we need to hear certain types of auditory experiences early in development if we are ever going to have a completely normal sequence of acquisition of vocal communication.”

What Bottjer and other neuroscientists studying vocal learning in songbirds have found is if the basal ganglia neural circuit is damaged during early vocal learning, acquisition of vocal communication is completely disrupted. But if disruption to the circuit occurs later on as adults, it has no obvious effect. Once songbirds have learned and acquired normal vocal communication, it is no longer affected by damage to that pathway — they are by adulthood “hardwired.”

What Bottjer has discovered in her latest study is that the basal ganglia pathway turns out to have two parallel circuits, rather than a single neural circuit as previously thought.

One pathway in the basal ganglia necessary for vocal learning in birds involves a group of neurons in the basal ganglia called the LMAN core. The focus of Bottjer’s new work, in the journal Nature Neuroscience, is a newly identified pathway that involves neurons that surround the LMAN core — a region called the LMAN shell. Bottjer explained that the LMAN shell was serendipitously detected years ago by Frank Johnson, a postdoctoral scholar who worked in her lab and is now a faculty member at Florida State University.

“Both songbirds and humans have to hear and evaluate the match between auditory feedback of their own vocal utterances and the target or template of the sound,” Bottjer said. “Over time and practice, we and they get better. When the babble turns into words or songs, the brain is saying ‘good match!’ We think it is likely that this neural circuit is helping with this process.”

In her study, funded by the NIH, the new LMAN shell pathway was surgically lesioned; however the damage did not induce immediate disruption to vocal behavior. Instead there was a profound effect on vocal learning. Specifically, said Bottjer, the young songbirds did not produce a good vocal match to the song of their tutors. The songbirds also never developed a stable temporal sequence. For example, instead of singing A, B, C, D, they sang several different sequences (such as C, D, B, A; D,B,C,A and A,C,D,B).

The LMAN shell pathway might also be important for integrating different sensory modalities. Socialization in both birds and humans is very important for vocal learning. Birds, like humans, benefit from social interactions and thus do not learn well from tape recordings of vocal sounds. If they don’t see and interact with their tutors or even see plastic models of their tutors, young birds don’t learn to vocalize very well. This may also extend to motivational cues from the tutor.

“The applied value of research in this area beyond Parkinson’s and Huntington’s diseases is pivotal in learning how normal vocal learning is carried out and how speech is acquired,” Bottjer explained. “This and future research will assist in treating stuttering and other disorders of vocal communication. Stuttering affects one percent of the population and diseases such as autism, Parkinson’s, Alzheimer’s, and Fragile X syndrome all have major effects on the ability to communicate vocally.”

Bottjer would like to investigate the activity of individual neurons as the birds are producing vocal behavior, hearing what they produced, and observing and recording their learning and practicing behaviors.

“We will also look for cross-talk between pathways — integration and information sharing that is occurring between the parallel circuits in the basal ganglia as we expect there to be,” she said

Complex learned behavior and dissecting the underlying neural circuitry and the function of each pathway is a formidable job.

“Not just my lab but all of the labs in the aggregate working on this have made amazing progress in beginning to really understand the specific functions of all of these different neural circuits,” Bottjer said.