National Geographic : 2015 Jan
66 national geographic • january 2015 travel more quickly. Idle circuits die through the severing of connections, known as synaptic pruning. Between the ages of one and five, and then again in early adolescence, the brain goes through cycles of growth and streamlining, with experience playing a key role in engraving the circuits that will endure. How nature and nurture combine to shape the brain is nowhere more evident than in the development of language ability. How much of that comes hardwired, and how do babies acquire the rest? To learn how researchers are answering that question, I visit Judit Gervain, a cognitive neuroscientist at Paris Descartes Uni- versity who has spent the past decade probing the linguistic acumen of children, ranging in age from days to a few years. We meet on the steps of Robert-Debré Hospital in Paris, where Gervain is readying an experiment on newborns. I follow her into a room down the hall from the maternity ward. The morning’s first sub- ject is wheeled in on a cart, swaddled in a pink polka-dot blanket, with dad in tow. A research assistant slips a skullcap studded with button- like sensors onto the infant’s head. The plan is to image the baby’s brain while playing a variety of audio sequences, like nu-ja-ga. But before any observations can begin, the baby emits a series of high-pitched cries, making it known he isn’t going to submit. The assistant hurriedly removes the cap, and the dad cradles the baby. After they leave, Gervain, who had just be- come a mother a few months earlier, tells me that such failures are not uncommon. Anoth- er newborn—also accompanied by dad—is wheeled in. Gervain’s assistant follows the same protocol, and this time the observing goes off without a hitch. The baby sleeps through it. Gervain and her colleagues have used a similar setup to test how good newborns are at discrimi- nating between different sound patterns. Using near-infrared spectroscopy, the researchers im- aged the brains of babies while they heard audio sequences. In some, the sounds were repeated in an ABB structure, such as mu-ba-ba; in others, an ABC structure, such as mu-ba-ge. The researchers found that brain regions responsible for speech and audio processing responded more strongly to the ABB sequences. In a later study they found that the newborn brain was also able to distin- guish between audio sequences with an AAB pat- tern and those with an ABB pattern. Not only could babies discern repetition, they also were sensitive to where it occurred in the sequence. Gervain is excited by these findings because the order of sounds is the bedrock upon which words and grammar are built. “Positional infor- mation is key to language,” she says. “If some- thing is at the beginning or at the end makes a big difference: ‘John killed the bear’ is very different from ‘The bear killed John.’ ” That the baby brain responds from day one to the sequence in which sounds are arranged suggests that the algorithms for language learn- ing are part of the neural fabric infants are born with. “For a long time we had this linear view. First, babies are learning sounds, then they are understanding words, then many words togeth- er,” Gervain says. “But from recent results, we know that almost everything starts to develop from the get-go. Babies are starting to learn grammatical rules from the beginning.” Researchers led by Angela Friederici, a neu- ropsychologist at the Max Planck Institute for Human Cognitive and Brain Sciences in Leipzig, Germany, have found evidence of such compre- hension in an experiment with four-month-old German babies exposed to an unfamiliar lan- guage. The children first heard a series of Italian sentences representing two types of construction: “The brother can sing” and “The sister is singing.” After three minutes they listened to another set of Italian sentences, some of which were grammati- cally incorrect, along the lines of “The brother is sing” and “The sister can singing.” During this phase the researchers measured the infants’ brain activity using tiny electrodes placed on the scalp. In the first round of testing the babies showed a similar brain response to both correct and incor- rect sentences. A few rounds of training later, the infants exhibited very different activation patterns when they heard erroneous constructions. In just 15 minutes the babies appeared to have In Patricia Kuhl’s lab at the University of Washington, researchers study brain activity in babies less than a year old using a magnetoencephalography device, which measures the magnetic field around a baby’s scalp, to reveal the pattern of neurons firing.