Babies and eye tracking

By the time a baby is 4 months old, she has begun to amass an impressive array of skills. She might be able to roll over, as Nora is demonstrating in this picture. She will almost certainly be able to follow an object with her eyes as it moves across her field of vision. But research about the ability of babies to follow objects has had conflicting results. In some cases, babies this age seem to understand that when one object passes out of view behind another, it's still "there," just hidden from view.

Consider a simple example: a single ball moves back and forth behind a rectangular box, like this:

As an adult, you probably imagine the ball passing behind the rectangle and reappearing on the other side. So you wouldn't be surprised to see this result when the rectangle is removed:

Two-month old infants, however, are surprised by this. After seeing the first animation, they stare at the second animation for longer than this one:

Researchers have concluded from this that two-month-old infants don't understand the concept of one object moving behind another. They simply see the first and third animations as similar, and the second as different, and therefore more interesting. By the time they are four months old, babies respond the same way adults do, and are more surprised by the "disappearing" ball than the smoothly animated ball.

But if a wider rectangle is substituted (a 17-centimeter rectangle instead of a 10-centimeter one), so that the ball takes longer to pass behind it, then suddenly four-month-olds start to respond the same way as younger infants: they are more surprised by the second animation than the third. Do infants lose their ability to understand how objects move depending on the objects' size?

A team led by J. Gavin Bremner suspected that the key wasn't so much the size of the objects, but the duration of time that an object was occluded by another. They developed a set of five experiments to test this notion.

There are several different ways you can manipulate the amount of time an object is occluded. You can use a larger occluder, you can change the size of the ball (a larger ball would disappear for a shorter amount of time), or you can move the ball faster or slower.

The team started by using a larger ball: a 17.4 centimeter ball with a 17.7 centimeter occluder. This is the same size occluder that had baffled four-month-olds before, but now they responded in a manner consistent with adults: After watching this larger ball pass behind a rectangle, they were more surprised by the "disappearing" ball than by the smoothly animated ball.

They found similar effects when they moved a small ball quickly while it was behind the occluder. As long as the time behind the occluder was short, four-month-olds behaved as if they "knew" the ball was passing behind the occluder, acting surprised when the ball later appeared to be disappearing.

In a separate experiment, when the ball moved slowly behind the same-width occluder, babies were equally surprised by both the disappearing and smoothly animated balls.

In their final experiment, the team used a narrow occluder and a slow-moving ball, so that the time of occlusion was the same as with the fast ball and wide occluder. Once again, babies were surprised by disappearing balls.

Bremner et al. conclude that four-month-olds' ability to see motion as continuous depends both on the time and physical length of occlusion. If an occluder is too long, or if the duration is too long, then babies appear to see the balls on either side of the occluder as two independent objects. Or perhaps even this is too strong a claim: perhaps what infants perceive are continuous motions -- the babies see the connection between motions if they are close enough, either spatially or temporally, but otherwise, the motions are deemed unrelated.

Bremner, J.G., Johnson, S.P., Slater, A., Mason, U., Foster, K., Cheshire, A., & Spring, J. (2005). Conditions for young infants' perception of object trajectories. Child Development, 76(5), 1029-1043.