By Cindy Moffett
“Welcome Leo” appears on a hand-lettered sign adorning the sky-blue wall of the waiting area. A brightly colored cartoon mural featuring flowers, butterflies, and small woodland creatures provides a soothing backdrop.
When 14-month-old Leo and his mother arrive, they are greeted by three cheerful psychology students. Leo’s mother sits on a sofa and sets him on the floor between her feet. One of the students delivers some toys for Leo to enjoy while he gets acquainted with the new environment.
Young Leo may not know it yet, but he is about to assist UT’s Infant Language and Perceptual Learning Lab with some very serious science.
Jessica Hay is an associate professor of psychology and director of the lab. She recently received a five-year $1.3 million grant from the National Institutes of Health to study language acquisition in children from six months to 36 months old.
“I’m looking at how kids are able to track patterns or statistics in the language they hear and then use that information to learn something new,” Hay explained.
The project has three distinct parts—how infants find and remember words heard in a continuous stream of speech, how specifically infants represent words, and how background noise affects learning. About 1,200 infants will take part in nine different studies.
Through this research, Hay hopes to better understand how infants learn language and to apply that knowledge when challenges such as hearing loss occur at an early age and delay language learning.
Pattern Recognition
Graduate student Sara Parvanezadeh explains Leo’s experiment to his mother. “We are looking at the ability to isolate words in a stream of speech. Your job is basically to be a fancy chair for your child. It’s very important not to direct where he is looking. I’ll give you headphones to listen to masking music,” she says.
Leo’s task is to recognize words in an unfamiliar language. In the English he has heard from birth, certain syllables are more likely to be linked. For instance, ma-ma is expected, but ma-dad is not. This is called transitional probability.
“Infants are really good at tracking regularities in language,” Hay noted. “They can track the likelihood that two sounds will go together.”
During the exercise, Leo will listen to a recording spoken in Persian. Four key words will be dispersed throughout the passage. In two of the words—patoo and limu—the first syllable predicts the second syllable 100 percent of the time. Leo should recognize these two words with high transitional probability more easily than the other two—biza and shena—where the two syllables occur together much less often.
“We can measure how long kids are listening to these different types of words,” Hay said. “If there is a difference in the listening times, we know they’ve learned something from the language.”
Stress patterns can present an additional challenge in word recognition. In English, stress generally falls on the first syllable as in MA-ma, BA-by, and KIT-ty. In Persian, stress is more likely to fall on the second syllable.
“Can kids track the regularities when there’s stress on the second syllable? Perhaps they need both transitional probability and familiar stress patterns,” Hay said.
Sounds of Science
Leo and his mother are soon ushered into a listening booth, where they are seated in the center of three blank computer screens. Mom dons a set of headphones and holds Leo in her lap as he begins looking and listening.
The study begins with a colorfully animated pinwheel appearing on one of the screens, drawing Leo’s attention. For about three minutes, he hears 12 Persian sentences repeated three times.
The spinning pinwheel then switches to a different screen and the testing phase begins. Leo shifts his focus and a key word repeats clearly and slowly—pahTOO, pahTOO, pahTOO.
“Kids will look in the direction of a sound they’re interested in listening to, and when they get bored they’ll look away,” Hay explained.
Video of Leo’s actions is recorded and used to measure the amount of time he listens to the sounds. When Leo looks away, the pinwheel moves to another screen and a different key word is played. The process continues for five to 10 minutes as Leo hears each of the four test words.
“This is an example of one test condition,” Hay added. “One age group, one condition. In the next condition, we’ll change the stress pattern. Each study will have three or four conditions, and could take two years to collect the data.”
After his test, Leo returns to the play area and is given a T-shirt emblazoned with “Little Scientist,” which he thoughtfully chews in the way of all babies. This little scientist can now be proud of his contribution to the complex field of infant language and perceptual learning.
For more information about participating in this research, visit the lab’s website.
To learn more about child development research at UT, visit the Child Development Research Group’s website.