Children from the poorer strata of society begin life not only with material disadvantages but cognitive ones. Decades of research have confirmed this, including a famous 1995 finding by psychologists Betty Hart and Todd Risley: By age four children reared in poverty have heard 30 million fewer words, on average, than their peers from wealthier families. That gap has been linked to shakier language skills at the start of school, which, in turn, predicts weaker academic performance.

But the sheer quantity of words a toddler hears is not the most significant influence on language acquisition. Growing evidence has led researchers to conclude quality matters more than quantity, and the most valuable quality seems to be back-and-forth communication—what researchers variously call conversational turns, duets or contingent talk.

A paper published last week in Psychological Science brings a new dimension of support to this idea, offering the first evidence these exchanges play a vital role in the development of Broca’s area, the brain region most closely associated with producing speech. Further, the amount of conversational turns a child experiences daily outweighs socioeconomic status in predicting both activity in Broca’s area and the child’s language skills.

The study, from the lab of neuroscientist John Gabrieli of Massachusetts Institute of Technology, involved 36 children, ages four to six, from a range of socioeconomic backgrounds. It had three components: First, researchers used standardized tests to evaluate each child’s verbal ability and derive a composite score. Second, the brain of each child was scanned using functional magnetic resonance imaging (fMRI) while the child listened to very short (15-second) stories. Lastly, adult–child communication at home was evaluated for two days using a state-of-the art recording and analysis system called LENA (Language Environment Analysis) to measure adult speech, the child’s utterances and their “conversational turns”—paired exchanges separated by no more than five seconds.

The researchers confirmed the classic 1995 finding that, overall, kids from wealthier families hear more words. And although their sample was small, they even confirmed the 30-million-word gap between the poorest and richest children. But what correlated most closely with a child’s verbal score was not the number of words he or she heard but the number of conversational turns. And these exchanges were the only aspect of language measured by LENA that correlated with the intensity of activity seen in Broca’s area during the fMRI story session. “We found that by far the biggest driver for brain development was not the number of words spoken but the conversations,” Gabrieli says. And although on average parents with greater income and education have more of these verbal exchanges with their young children, “there’s pretty good diversity,” he notes. In other words, some low-income parents engaged in a lot of conversation with their child and some wealthier parents conversed relatively little.

The researchers calculated that a child’s verbal ability score increased by one point for every additional 11 conversational exchanges per hour.

How exactly exposure to these exchanges alters Broca’s area is a question Gabrieli’s team is exploring in subsequent research. “We know that greater activation in Broca’s area was associated with better verbal abilities overall, so it seems like greater activation is good,” he says. One possibility is back-and-forth communication promotes more connections between brain cells in that region.

The study is a “very, very important” addition to a growing body of work, says developmental psychologist Kathryn Hirsh-Pasek, director of the Infant Language Laboratory at Temple University, who was not involved in the work. “We have known for quite a while that conversational turns—or what in my work we call conversational duets—are very important for building a foundation for language and maybe for learning generally. What hadn’t been done is to link it where we knew it had to be linked—to changes in the brain.”

Verbal exchanges have two components that children must master: temporal contingency and semantic contingency—essentially, understanding the timing of human conversation and how to respond meaningfully. Research, including Hirsh-Pasek’s, has shown children cannot learn this from watching television, although they can learn it via video-chat technology such as Apple’s FaceTime.

Contingent language begins in infancy—well before words emerge—when parents begin “cooing” and “gooing” at their babies, who respond in kind. Socioeconomic differences in this behavior arise during the first year of life, according to a 2017 study of 141 11-month-olds by Michelle McGillion of The University of Sheffield in England and colleagues.

Research in this area has big implications for parents and caregivers. The search is on for interventions that will increase adult–child conversation and boost early language skills, especially for families living in poverty. McGillion’s study, for instance, showed language learning took off for babies in low-income settings when caregivers were given instructions to spend 15 minutes a day engaging their infant by commenting on whatever the baby looked at. Unfortunately, the improvements did not persist at age two with this low-intensity intervention.

Encouraging conversation seems particularly necessary in an era when both children and adults are spending more time with devices and less in face-to-face communication. “The exchanges are not only about words but about feelings, about paying attention to someone else,” Gabrieli observes. Hearing language from television or Alexa, he says, “does very little compared to these exchanges.”

Hirsh-Pasek shares this concern about technology. One 2017 study she co-authored found that when a cell phone call interrupts an interaction in which a parent is teaching a child a new word, the learning is lost.

While we are fiddling with our digital devices, “evolution is screaming at us,” she says. “It’s saying, ‘Hey, in case you didn’t notice, there’s another human in the room—pay attention.’ If we learn better how to follow the eyes of our child and comment on what they are looking at, we will have strong language learners. And language is the single-best predictor of school readiness—in math, social skills and reading skills. It is the foundation for learning.”