From contemporary syntax to human language’s deep origins

Like father, not like son

The song of songbirds is a learned, complex behavior and subject to strong selective forces. However, it is difficult to tease apart the influence of the genetic background and the environment on the expression of individual variation in song. Scientists from the Max Planck Institute for Ornithology in Seewiesen in collaboration with international researchers now compared song and brain structure of parents and offspring in zebra finches that have been raised either with their genetic or foster parents. They also varied the amount of food during breeding. Remarkably, both song and the underlying brain structure had a low heritability and were strongly influenced by environmental factors.

A central topic in behavioral biology is the question, which aspects of a behavior are learned or expressed due to genetic predisposition. Today it is known that our personality and behavior are far less determined by the genetic background. Especially during development environmental factors can shape brain and behavior via so-called epigenetic effects. Thereby hormones play an important role. A shift in hormone concentrations in early life can have long lasting effects for an organism, whereas the same change in adults often may show only short-term changes. However, whether the influence of the environment has either strong or weak effects can largely depend on the individual genetic predisposition. However, it is relatively hard to discriminate the effects of the environment from that of the genes.

An attempt to tease apart these effects has been conducted by researchers from the Max Planck Institute for Ornithology in collaboration with an international team of scientists in zebra finch breeding pairs. Using partial cross-fostering the researchers exchanged half of the eggs within a nest making them to “cuckoo’s eggs”. Therefore half of the hatchlings were raised by their genetic parents, whereas the other half was raised by their foster parents. In addition they modified the availability of food by mixing the seeds with husks so that the parents had to spend more time searching for food. When the male offspring were adult at 100 days the researchers recorded their songs and analyzed the underlying neuroanatomy. This partial cross-fostering design enabled the researchers to tease apart the involvement of genotype, the rearing environment and nutritional effects to variation in song behavior and brain structure.

The results showed that heritability values were low for most song characteristics, except the number of song syllables and maximum frequency. On the other hand the common rearing environment including the song of the foster father mainly predicted the proportion of unique syllables in the songs of the sons, moreover this relationship was dependent on food availability. Even more striking results were found when the researchers investigated the brain anatomy. Heritability of brain variables was very low and strongly controlled by the environment. For example, an emergence of a clear relationship between brain mass and genotype is prevented by varying environmental quality.

This result was quite surprising as previous studies have found a high heritability of the song control system in the songbird brain, however these studies did not account for variation of the rearing environment. ”Being highly flexible in response to environmental conditions can maintain genetic variation and influences song learning and brain development” says Stefan Leitner, senior author of the study.

How Birds and Babies Learn to Talk

Few things are harder to study than human language. The brains of living humans can only be studied indirectly, and language, unlike vision, has no analogue in the animal world. Vision scientists can study sight in monkeys using techniques like single-neuron recording. But monkeys don’t talk.

However, in an article published in Nature, a group of researchers, including myself, detail a discovery in birdsong that may help lead to a revised understanding of an important aspect of human language development. Almost five years ago, I sent a piece of fan mail to Ofer Tchernichovski, who had just published an article showing that, in just three or four generations, songbirds raised in isolation often developed songs typical of their species. He invited me to visit his lab, a cramped space stuffed with several hundred birds residing in souped-up climate-controlled refrigerators. Dina Lipkind, at the time Tchernichovski’s post-doctoral student, explained a method she had developed for teaching zebra finches two songs. (Ordinarily, a zebra finch learns only one song in its lifetime.) She had discovered that by switching the song of a tutor bird at precisely the right moment, a juvenile bird could learn a second, new song after it had mastered the first one.

Thinking about bilingualism and some puzzles I had encountered in my own lab, I suggested that Lipkind’s method could be useful in casting light on the question of how a creature—any creature—learns to put linguistic elements together. We mapped out an experiment that day: birds would learn one “grammar” in which every phrase followed the form of ABCABC, and then we would switch things up, giving them a new target, ACBACB (the As, Bs, and Cs were certain stereotyped chirps and peeps).

The results were thrilling: most of the birds could accomplish the task. But it was clearly difficult—it took several weeks for them to learn the new grammar—and it was challenging in a particular way. While the birds showed no sign of needing to relearn individual sounds, the connections between individual syllables, known as “transitions,” proved incredibly difficult. The birds proceeded slowly and systematically, incrementally working out each transition (e.g., from C to B, and B to A). They could not freely move syllables around, and did not engage in trial and error, either. Instead, they undertook a systematic struggle to learn particular connections between specific, individual syllables. The moment they mastered the third transition of the sequence, they were able to produce the entire grammar. Never, to my knowledge, had the process of learning any sort of grammar been so precisely articulated.

We wrote up the results, but Nature declined to publish them. Then Dina and Ofer speculated that our findings might be more convincing if they were true for not only zebra finches (hardly the Einsteins of the bird world) but for other species as well. Ofer contacted a Japanese researcher, Kazuo Okanoya, who he thought might be able to gather data for Bengalese finches, which have a more complex grammar than zebra finches. Amazingly, the Bengalese finches followed almost exactly the same learning pattern as the zebra finches.

Then we decided to test our ideas about the incrementality of vocal learning in human infants, enlisting the help of a graduate student I had been working with at N.Y.U., Doug Bemis. Bemis and Lipkind analyzed an old, publicly available set of human-babbling data, drawn from the CHILDES database, in a new way. The literature said that in the later part of the first year of life, babies undergo a change from “reduplicated” babbling—repeating a syllable, like bababa—to “variegated” babbling—often switching between syllables, like babadaga. Our birdsong results led us to wonder whether such a change might be more piecemeal than is commonly presumed, and our examination of the data proved that, in fact, the change did not happen all at once. It was gradual, with new transitions worked out one by one; human babies were stymied in the same ways that the birds were. Nobody had ever really explained why babbling took so many months; our birdsong data has finally yielded a first clue.

Today, almost five years after Lipkind and Tchernichovski began developing the methods that are at the paper’s core, the work is finally being published by Nature.

What we don’t yet know is whether the similarity between birds and babies stems from a fundamental similarity between species at the biological level. When two species do something in similar ways, it can be a matter of “homology,” a genuine lineage at the genetic level, or “analogy,” which is independent reinvention. It will likely be years before we know for sure, but there is reason to believe that our results are not purely an accident of independent invention. Some of the important genes in human vocal learning (including FOXP2, the gene thus far most decisively tied to human language) are also involved in avian vocal learning, as a new book, “Birdsong, Speech, and Language,” discusses at length.

Language will never be as easy to dissect as birdsong, but knowledge about one can inform knowledge about the other. Our brains didn’t evolve to be easily understood, but the fact that humans share so many genes with so many other species gives scientists a fighting chance.

Birdsong study pecks at theory that music is uniquely human

A bird listening to birdsong may experience some of the same emotions as a human listening to music, suggests a new study on white-throated sparrows, published in Frontiers of Evolutionary Neuroscience.

“We found that the same neural reward system is activated in female birds in the breeding state that are listening to male birdsong, and in people listening to music that they like,” says Sarah Earp, who led the research as an undergraduate at Emory University.

For male birds listening to another male’s song, it was a different story: They had an amygdala response that looks similar to that of people when they hear discordant, unpleasant music.

The study, co-authored by Emory neuroscientist Donna Maney, is the first to compare neural responses of listeners in the long-standing debate over whether birdsong is music.

“Scientists since the time of Darwin have wondered whether birdsong and music may serve similar purposes, or have the same evolutionary precursors,” Earp notes. “But most attempts to compare the two have focused on the qualities of the sound themselves, such as melody and rhythm.”

Earp reviewed studies that mapped human neural responses to music through brain imaging.

She also analyzed data from the Maney lab on white-throated sparrows. The lab maps brain responses in the birds by measuring Egr-1, part of a major biochemical pathway activated in cells that are responding to a stimulus.

The study used Egr-1 as a marker to map and quantify neural responses in the mesolimbic reward system in male and female white-throated sparrows listening to a male bird’s song. Some of the listening birds had been treated with hormones, to push them into the breeding state, while the control group had low levels of estradiol and testosterone.

During the non-breeding season, both sexes of sparrows use song to establish and maintain dominance in relationships. During the breeding season, however, a male singing to a female is almost certainly courting her, while a male singing to another male is challenging an interloper.

For the females in the breeding state every region of the mesolimbic reward pathway that has been reported to respond to music in humans, and that has a clear avian counterpart, responded to the male birdsong. Females in the non-breeding state, however, did not show a heightened response.

And the testosterone-treated males listening to another male sing showed an amygdala response, which may correlate to the amygdala response typical of humans listening to the kind of music used in the scary scenes of horror movies.

“The neural response to birdsong appears to depend on social context, which can be the case with humans as well,” Earp says. “Both birdsong and music elicit responses not only in brain regions associated directly with reward, but also in interconnected regions that are thought to regulate emotion. That suggests that they both may activate evolutionarily ancient mechanisms that are necessary for reproduction and survival.”

A major limitation of the study, Earp adds, is that many of the regions that respond to music in humans are cortical, and they do not have clear counterparts in birds. “Perhaps techniques will someday be developed to image neural responses in baleen whales, whose songs are both musical and learned, and whose brain anatomy is more easily compared with humans,” she says.

Baby songbirds learn to sing by imitation, just as human babies do. So researchers at Harvard and Utrecht University, in the Netherlands, have been studying the brains of zebra finches—red-beaked, white-breasted songbirds—for clues to how young birds and human infants learn vocalization on a neuronal level.

While a baby bird mimicking the chirps of his “tutor” may seem far removed from human learning, the researchers at the two universities found that the songs of the birds and human language are both processed in similar areas on the left sides of the two very different brains. The discovery was published last month in the Proceedings of the National Academy of Sciences.