Peekaboo is a classic game for babies for a reason. Infants are fascinated by unexpected actions and images, yet when it comes to human characteristics they tend to focus on the familiar. By tracking babies’ gazes, pupil dilation and brain activity, researchers have uncovered several patterns in what piques their interest.

The examples below highlight key factors that get an infant’s attention and the ways these curiosities contribute to our understanding of their cognitive development.

Vanishing acts
The age-old “now you see it, now you don’t” trick may not even fool an infant. In a classic 1985 experiment researchers showed four- to five-month-old infants scenarios in which a box is obscured by a curtain. In one version the box becomes visible again once the curtain is pulled back but in the other it seems to have vanished while behind the curtain. The babies spent more time looking at the illogical scenario, suggesting they understand that objects continue to exist even when out of view.

Three-month-old infants prefer faces from their own ethnic group, researchers found in a 2005 study of 64 Caucasian babies. Tracking the tots’ gazes as they looked at images of faces revealed that they spent more time looking at faces of their own ethnicities than those of other ethnicities. Newborns, however, demonstrated no such preference, implying that exposure to people from various races during early development may shape the way in which adults perceive ethnic differences.

Infants raised primarily by women look longer at images of female faces than male faces, according to a 2002 study of 102 babies. The inverse holds true as well, suggesting that a baby’s preferences for human faces is influenced by the gender of his or her primary caregiver.

Babies are biased toward people who speak their native languages. In a 2013 study, 16 infants watched a video in which a native English speaker and a native French speaker spoke in their respective languages to the infants then silently held up a toy. When the 10-month-olds, who were all raised among English speakers, got the chance to play with the toys in real life, they showed a preference for the toys held by the English speaker. Language, the researchers propose, may be a kind of filter that guides social learning.

Math Errors
Even babies as young as six to nine months old can recognize errors in simple math, according to a study conducted on 26 babies in 2006. To convey that 1 + 1 = 2, for example, the experimenters showed the babies a video of one puppet sitting on a table then placed an opaque box in front of the puppet. With the original puppet still obscured, they next placed another puppet behind the box. In the correct scenario, moving the box away reveals two puppets whereas in an incorrect scenario moving it might show only one. They found that the regions of the infants’ brains that showed activity when looking at incorrect math are the same regions in adults that respond to errors. This similarity hints that the development of executive control, which will one day allow the babies to respond to such errors, is already under way at a very young age.

Nonsensical Behavior
Infants perk up when observing irregular or nonsensical feeding habits, researchers found in 2009 by measuring pupil dilation. In one scenario, for example, the six- and 12-month-old babies observed a woman feeding her friend normally (fork to mouth), in another she feeds the table while her friend waits open-mouthed. Because the infants became more engaged, as indicated by increased pupil dilation, when observing the irrational feeding behavior, the researchers contend that even at a very young age anticipation of goals contributes to their understanding of everyday actions.