Singing In the Brain

Jarvis was trained in molecular biology in Rivka Rudner’s lab at Hunter and began his neuroscience career at Rockefeller University in Fernando Nottebohm’s group, using songbird communication as a model system to dissect the molecular biology of speech and vocal learning in the brain. “Rockefeller was a place where I had a lot of scientific freedom. The philosophy there is if there is a high probability of an experiment working, then you’re not doing the right experiment, and if it has a high probability of failure, then it could make a big impact in science. I am looking forward to that scientific environment, which is hard to find. And I am looking forward to being closer to my family. What I am not looking forward to is the car noise, the pollution, and the cold weather,” says the native New Yorker.

“I’m now considering studying the neurobiology of dance in parrots and humans. If I can do that, I will bring all of my passions together.”

Jarvis is also looking forward to opportunities to perform with one of the city’s many dance troupes. At the High School of the Performing Arts in NYC, Jarvis majored in ballet because “if you learn something complex, it will make everything easier. I guess I still think that way,” says Jarvis. He had scholarships to study at the Joffrey Ballet and the Alvin Ailey schools and still continues to dance and perform. In college, he chose a career in science over one in dance; in graduate school, Jarvis began pioneering work to understand the mechanisms of how songbirds learn to sing, and has since spearheaded evolutionary and sequencing studies of bird species while continuing to research the genetics and neuronal circuitry of vocal communication.

Here, he discusses how Saturday Night Fever kick-started his dance career, how his family inspired his scientific pursuits, and why he chose to study the brain.

JARVIS JUMPS IN

Citizen of the world. Jarvis was born in Harlem and grew up in the Bronx and Queens. “Most of my family were singers and musicians and we were expected to become musicians and singers as well. Right before I entered high school, the movie Saturday Night Fever came out. I began to imitate John Travolta and was winning dance contests in Connecticut where I lived with my mother, Valeria McCall. I thought maybe I had dance talent, so I auditioned for the High School of the Performing Arts and got in. I realize now how unusual that school was. We were being trained in arts and academics, but on top of that, on how to be a person—to think creatively and to be good citizens and neighbors rather than to compete with one another. That training has influenced who I am now as a scientist. I think that I take a more creative, collaborative, and inclusive approach than some of my colleagues do.”

Grasp on life. In high school, Jarvis enjoyed science, and biology in particular. “I was thinking, ‘What do I want to do for the rest of my life that would make this world a better place?’ because that is what my mother taught me, to do something that has a good impact on society. Of all the things I enjoyed, I felt that as a scientist I had a higher probability of impacting society than as a dancer,” he says.  Jarvis was also influenced by his father, James Jarvis, who had wanted to become a scientist in the 1960s. “He had a tragic life. He finished high school at 15, but dropped out of college and got into drugs. He was abusive to my mother and to us, his kids, which is why my mother left. I never saw him as a father but as a friend. He was homeless when I was in high school, living in caves in upstate New York to figure out how humans invented civilization. He lived with us for a bit when I lived with my grandfather and he helped me with calculus homework. He had this romantic vision of becoming a scientist that I inherited, but also mental health problems. I learned later that he was partly schizophrenic and paranoid. He was eventually shot and killed outside one of his caves by a group of teens who were shooting homeless people as gang initiations. All of this was influencing my view of life and my transition from high school to college.”

An auspicious start. In 1983, Jarvis entered Hunter College. He fell in love with lab work after joining Rudner’s lab, which worked on bacterial protein synthesis genes. Jarvis double majored in biology and mathematics because he couldn’t decide whether he wanted “to study how the brain worked or the origins of the universe,” he says. In the four years he conducted research in Rudner’s lab, Jarvis was an author on seven papers, including three on which he was first author. In one, he mapped the chromosomal organization of the ribosomal RNA genes in Bacillus subtilis. Jarvis also convinced Rudner to allow him to give a talk on bacterial genome rearrangements at a genetics conference. “She had said no at first, but I insisted and she told me that if I got the next set of results, I could give the talk,” says Jarvis. “For me, a person coming out of a community of color with disadvantages, what was great at Hunter was that the faculty would take you under their wing if you had the ambition and the will, which I don’t see at many of the top research schools.”  

JARVIS'S JOURNEY

A window into learning. For his PhD, Jarvis wanted to apply his molecular biology training to study complex traits such as learning and language. He chose to remain in New York City and attend Rockefeller University. In 1988, he joined Nottebohm’s lab, which was studying songbirds because, like humans, these birds had the ability to imitate new sounds. To figure out the molecular mechanisms of vocal learning, Jarvis first used classical conditioning, teaching adult canaries to associate song with a mild shock to their feet. He found that expression of the ZENK gene in the forebrain increased in the birds taught the association, suggesting that the gene may be involved in song-linked memory. But Jarvis says he couldn’t find any gene changes in the brain region that most interested him, the song-learning nuclei of the forebrain. After finding too many confounding variables in classical conditioning experiments, Jarvis discovered that all he had to do was let the birds sing, and he found singing-driven gene expression in the song-learning nuclei. Since then, he has used the birds’ natural behaviors rather than an artificial setup.  

Against the odds. Jarvis’s first few years at Rockefeller were rough for him. He did not think that he fit into the culture, both because of the lack of diversity, and because of the sink-or-swim environment. He also had a wife and a young daughter and son to take care of. “But I learned how to hustle, got over the isolation, and then everything picked up,” he says. At the end of graduate school, Jarvis ignored career advice to choose a new lab for his postdoctoral training. “This was the best lab in the country, and the world, for what I was doing, and the experiments were finally working, so why would I switch?” he says. Jarvis stayed in Nottebohm’s lab for another three years as a postdoc, focusing on the pathways of vocal learning in the brains of songbirds. “This ability to imitate sound is different than regular learning, it’s a form of specialized learning not found in many animals,” he says. Jarvis found that brain gene expression varies depending on whether it is linked to the production or perception of song in birds, and also that, depending on the type of song a male songbird sings—either a courtship song to a female or one sung alone for practice—the resulting brain activation patterns differed dramatically.

Song to speech. In 1998, Jarvis moved to Duke University to set up his own lab to continue to study song learning as a window into language. “As I was writing grant proposals, I realized the knowledge of how to translate findings in bird brains to human brains was limited. I thought that, at some point, as we make discoveries in songbirds, there would be someone else who would try to test these findings in humans; but I didn’t see that happening, so I thought I’d try it theoretically myself,” he says. Jarvis began to do meta-analyses to make sense of how songbird research applies to human learning, culling from work going back as far as 150 years. The project culminated in a 2004 paper in which Jarvis hypothesized that vocal learning in both birds and humans evolved from an ancestral and basic neural network within the vertebrate brain. “Researchers were proposing that songbirds are the best model for human speech and showed parallel behaviors, but at the neurobiology level hardly anyone was making that jump between humans and songbirds,” he says. In that same time period, Jarvis co-led a consortium of scientists with Tony Reiner of the University of Tennessee Health Science Center that resulted in a major revision of the avian brain nomenclature, which has led to an appreciation of the homologies between the avian and the vertebrate brains. “The result helped  justify avian brain experiments for biomedical research  because they could now be more easily related to mammal studies.”

Jarvis and his students continued to study the evolution of vocal learning, which resulted in the motor theory for the origin of vocal learning, published in 2008. Based on comparisons of neural activity related to vocal learning and movement control in bird brains, Jarvis and his colleagues posited that the neural circuit for song learning in birds, and likely in humans, emerged by duplication of an ancient motor pathway in the brain. “I began to realize that the vocal-learning circuits, including those for human speech that researchers considered cognitively advanced, are really basic motor-learning pathways. It was the first time that I started thinking that this dichotomy in the linguistics and neurobiology communities of a separation between speech and spoken language is false,” says Jarvis. “I didn’t see any evidence of this separation from the neurobiology studies. No one has found a language part of the brain that is separate from speech production or processing parts.”

Bird tree. Since his days at Rockefeller, Jarvis had been thinking about understanding the genes common to songbirds and humans—as well as to other vocal-learning species such as parrots, hummingbirds, bats, whales, and dolphins—but there were few tools to study those genes. By 2010, only two bird genomes had been sequenced, the zebra finch, a songbird, and the chicken, a nonvocal learner, which was not enough to screen for genes associated with vocal learning. It was also unclear if vocal learning really did evolve multiple times among birds. So Jarvis helped lead several consortia, including one at BGI in China that sequenced 45 bird species and resulted in eight articles published in a 2014 special issue of Science, along with 20 publications in other journals. “We created the first genome-scale family tree of any vertebrate class, and we confirmed that vocal learning did evolve multiple times,” says Jarvis.  He and his collaborators are now working on sequencing additional avian genomes.

Comparing the bird genomic data with human brain gene-expression data provided by the Allen Institute for Brain Science in Seattle, “we saw lots of convergent gene expression changes within the vocal-learning brain circuits of birds and in the speech brain regions of humans that were not present in nonhuman primates or in pigeons and doves and mice,” he says. “I rarely give up, but before this genomes project, I was thinking, ‘I’m not sure I will ever get to finding out whether there are these convergent genes in my lifetime,’ because we didn’t have the resources. These papers were like a closing of a major chapter for me. Now we can start to manipulate these genes in mice or other species. I don’t know if we can do these manipulations and achieve engineering brain circuits for vocal learning behavior, but I am more optimistic now than I was before.”

JARVIS'S JUGGLING

Stayin’ alive. “At some point, I thought I would give up on dance, but I’m still doing it! I’ve never stopped dancing, although I did stop performing for a bit. I gave three performances this past year with the James Cobo salsa dance team. Twenty years ago, I thought by the time I was 50 I would be too old for dancing. After I got tenure, I started performing again. I had this fear that I wouldn’t accomplish everything I want to before I die, and I see my dancing connected with staying healthy and alive a lot longer to accomplish my scientific goals.”

Moving to the beat. In 2008 and 2009, researchers demonstrated that dancing in synch to a rhythm is not just intrinsic to humans, but that other species capable of vocal learning can also spontaneously learn to dance. (See “John Iversen Explores Our Perception of Musical Rhythm.”) “In our analysis across species, we found that the genes common to song pathways in different bird species also affect motor circuits involved in coordinating movements in parrots, who are the better vocal learners and dancers,” says Jarvis. “This is the start of my going full circle. I’m now considering studying the neurobiology of dance in parrots and humans. If I can do that, I will bring all of my passions together.” 

Greatest Hits

• Showed that brain activation patterns in songbirds differ depending on the social function of the communication
• Discovered relationships between brain pathways for vocal learning in song-learning birds and in humans
• Proposed a theory that vocal-learning circuits in birds and humans likely emerged by duplication of an ancient motor pathway in the brain and that there are no language-specific brain regions separate from speech brain regions
• Co-led an international team that made major revisions in the century-old terminology describing avian brains and contributed to a new understanding of vertebrate brain evolution
• Co-led an international team that sequenced the genomes of 45 bird species and used these plus 3 older sequences to create the first genome-scale phylogenetic tree of a vertebrate group, leading to the discovery that the same specialized form of genes that allow songbirds to learn to sing are also specialized in the human brain regions used to learn speech