Harvard Medicine

More... Share to Twitter Share to Facebook
Saving Face

Infants may be more skilled than adults at facial recognition.

BABY ON BOARD: The young participants in Charles Nelson's studies give us a glimpse into how facial recognition works.<br/><br/>Photo by Margaret Lampert The six-month-old sits cradled on his mother’s lap in the dark, watching images flash across the screen in front of him. But the colorful characters of Sesame Street and Yo Gabba Gabba aren’t on view. Instead, the child, whom we’ll call Tommy, is watching a series of human faces; an eye tracker is measuring his reactions as part of an ongoing study about facial recognition. It’s the first of several visits Tommy will make to the Laboratory of Cognitive Neuroscience at Children’s Hospital Boston. While the significance of his participation might be lost on the youngster—despite a mild cold, he sucks calmly on his pacifier throughout the hourlong session—his responses will aid scientists’ understanding of how we develop the ability to recognize and process faces.

“Our faces contain a wealth of information, from our identity, age, gender, and race to our emotions and intentions,” says Charles Nelson, the lab’s director and an HMS professor of pediatrics and neuroscience. “For babies, who are preverbal, much depends on their ability to recognize and read faces.”

In fact, infants may be more skilled than adults at facial recognition, as Nelson and his colleagues discovered nearly a decade ago when they tested babies’ ability to discriminate among different human and simian faces. Facial decoding skills develop during the first six months of life, when babies can easily distinguish among human and monkey mugs, a talent that ebbs with time. By nine months, children can still differentiate among human faces but can no longer recognize different monkey faces. As Nelson explains, our facial processing skills depend in part on our experience: The more we’re exposed to other people, the better able we are to distinguish between individuals. Unless we regularly hang out with monkeys, we’re apt to think they all look alike.

To prove this point, Nelson gave a book of macaque faces to the parents of 13 six-month-olds and asked them to spend a few minutes every day showing their children the faces. Another 13 babies received no macaque training. After three months, the babies who had regularly viewed monkey faces could differentiate between individual monkeys; the control babies could not. The results, says Nelson, “show that our experiences play a huge role in what faces we can and can’t recognize.” Ongoing studies by Nelson and others analyzing our ability to differentiate faces of other races, ethnicities, and genders suggest similar effects.

More than just cocktail-party fodder, Nelson’s findings offer stunning insights into how our visual system develops. Back at the lab, Tommy wears a cap netted with electrodes, which measure how his brain reacts to visual stimuli. As faces bearing exaggerated emotions—joy, sadness, fear—flit across the screen, a computer generates a graph based on his responses. If Tommy’s facial processing skills are similar to those of other babies his age, he’ll focus the longest on fearful expressions, even though he doesn’t yet understand what they mean. Research suggests that babies are typically first attracted to fearful faces and later to expressions of sadness, happiness, and anger. This decoding helps us form relationships and understand each other.

In contrast, babies at high risk for autism—typically those who have an older sibling with the developmental disorder—appear to lack face-reading expertise. “Children and adults with autism tend to avoid looking directly at other people’s faces, particularly the eyes,” explains Nelson. “If they focus on anything, it’s the mouth or the edges of the face.” They also don’t seem to recognize faces well: When exposed to images of their mother’s face or that of a stranger, most children are drawn to Mom’s visage. Those at risk for autism, however, respond better to strangers’ faces, if they respond at all.

Findings like these may enhance the diagnosis and treatment of autism symptoms. “Our research has the potential to help us identify kids with autism as early as six months,” says Nelson; the disorder often isn’t diagnosed until the child is at least two years old. “If we can detect it earlier, too.”


Add new comment