What is found in a good conversation? It is certainly correct to say words—the more engagingly put, the better. But conversation also includes “eyes, smiles, the silences between the words,” as the Swedish author Annika Thor wrote. It is when those elements hum along together that we feel most deeply engaged with, and most connected to, our conversational partner, as if we are in sync with them.

Like good conversationalists, neuroscientists at Dartmouth College have taken that idea and carried it to new places. As part of a series of studies on how two minds meet in real life, they reported surprising findings on the interplay of eye contact and the synchronization of neural activity between two people during conversation. In a paper published on September 14 in Proceedings of the Natural Academy of Sciences USA, the researchers suggest that being in tune with a conversational partner is good but that going in and out of alignment with them might be better.

Making eye contact has long been conceived as acting like a cohesive glue, connecting an individual to the person with whom they are talking. Its absence can signal social dysfunction. Similarly, the growing study of neural synchrony has focused on the positive aspects of alignment in brain activity between individuals.

In the new study, by using pupil dilation as a measure of synchrony during unstructured conversation, psychologist Thalia Wheatley and graduate student Sophie Wohltjen found that the moment of making eye contact marks a peak in shared attention—and not the beginning of a sustained period of locked gazes. Synchrony, in fact, drops sharply after looking into the eyes of your interlocutor and only begins to recover when you and that person look away from each other. “Eye contact is not eliciting synchrony; it’s disrupting it,” says Wheatley, senior author of the paper.

Why would this be? Conversation requires some level of synchrony, but Wheatley and lead study author Wohltjen speculate that breaking eye contact ultimately propels the conversation forward. “Perhaps what this is doing is allowing us to break synchrony and move back into our own heads so that we can bring forth new and individual contributions to keep the conversation going,” Wohltjen says.

“It’s a fantastic study,” says psychiatrist and social neuroscientist Leonhard Schilbach of the Max Planck Institute of Psychiatry in Munich, who studies social interaction but was not involved in the research. He applauds the design of the experiment to replicate natural encounters and the focus on free-form conversation. The results suggest, he says, that “interpersonal synchrony is an important aspect of social interactions but may not always be desirable.”

Others in the field are drawn to the researchers’ creative way of thinking about conversation, which is described as “a platform where minds meet” in the paper. “Such a conceptualization may inspire other researchers to think about conversation differently and study it more deeply,” says Juliana Schroeder, a social psychologist at the Haas School of Business at the University of California, Berkeley, who also was not involved in the research.

The new work builds on an earlier study by Wheatley and psychologist Olivia Kang, now at Harvard University, who showed that pupillary synchrony serves as a measure of shared attention. Our pupils get larger and smaller as a reflexive response to changes in light but also, to a lesser degree, when we are physiologically aroused. Kang and Wheatley tracked eye movements in speakers as they recounted positive or negative memories about their life. Then the researchers tracked the eye movements of people listening to the same stories at a later point in time. They found that the pupil dilation of the listeners synchronized to that of the speakers when there were emotional peaks in the stories. “We knew this was a marker of people being on the same page as each other,” Wheatley says.

For the current paper, Wohltjen wanted to extend those earlier findings by studying face-to-face conversation in order to see how eye contact might influence shared attention in real time. She put 186 psychology students at Dartmouth, all relative strangers, into conversational pairs and asked them to talk for 10 minutes about anything they wanted while she tracked their eye movements. Participants also watched videos of their conversations and rated their remembered level of engagement minute by minute.

“We expected that eye contact worked like a cattle prod to get two people back onto the same wavelength,” Wohltjen says. If that were so, the onset of eye contact should have led to a subsequent increase in pupillary synchrony. Instead the researchers found the opposite: a peak in synchrony at the onset followed by a decrease. But they also found that participants reported being more engaged when they were making eye contact. “We thought, ‘Perhaps this making and breaking of eye contact must do something to help the conversation,’” Wohltjen says.

Previous studies of eye contact have generally been passive, as in Wheatley and Kang’s earlier work. The real-world design of Wohltjen’s experiment served as a reminder that most people naturally look at and away from each other many times during a conversation. Holding someone’s gaze for too long—or not at all—can seem awkward. As the researchers thought further about what eye contact might be doing for us, they turned to the literature on creativity. There they recognized the constraints of too much synchrony. “If people are trying to innovate in some way, you don’t want people in lockstep with each other,” Wheatley says. “You want people to [say], ‘What if we did this? What if we did that?’ You need people to be providing their independent insights and building that way.”

The idea that eye gaze can be used to modulate synchrony is intriguing to other researchers. “The elegant experimental approach [in this paper] might be helpful to quantitatively investigate psychiatric conditions, which can be described as ‘disorders of social interaction,’” says Schilbach, who has studied gaze and other elements of social interaction in autism.

The findings also help explain the frustrations of Zoom and other video conferencing platforms, on which real eye contact is nearly impossible to make—or break—because of the positioning of cameras and windows on screens. (The paper’s publication prompted a lively discussion of just that phenomenon on Twitter.)

Wheatley can imagine follow-up studies that examine a variety of conversational contexts. How does the dynamic dance between making and breaking synchrony play out when a parent is instructing a child, for instance? Presumably, in that situation, a parent would be hoping for the child’s full attention and therefore complete synchrony. On the other hand, perhaps the study helps explain why long car rides, in which people do not look at each other the whole time, are often conducive to deep conversation.

“There might be an optimal sweet spot in this coupling, decoupling thing—where people are really listening to each other, but they’re also fueling the conversation with new ideas,” Wheatley says. “Those conversations might be the most fun.”

A version of this article with the title “Losing Contact” was adapted for inclusion in the December 2021 issue of Scientific American.