When is an object not an object? : Insights from infants

Johnson, S. P. and Bremner, J. G. and Slater, A. and Mason, U. and Foster, K. (2001) When is an object not an object? : Insights from infants. Journal of Vision, 1 (3): 459. ISSN 1534-7362

Abstract

The visual system can track several objects on multiple trajectories, even when temporarily occluded. The ontogenetic origins of these processes have been proposed to lie in perception of objects as spatiotemporally bounded and continuous ("object continuity") from the start in infancy. Using habituation and eye tracking methods, however, we have obtained data indicating that young infants are poor at perceiving objects across occlusion. After habituation to a ball that moved repeatedly behind an occluding box and out again (occlusion time 733 ms), 4-month-olds looked longer at a "complete" trajectory, in which the ball moved back and forth with no occluder, relative to a "broken" trajectory, in which it disappeared and then reappeared. This preference reversed when the ball was occluded for a shorter time (67 ms). Six-month-olds, however, responded to object continuity for the longer occlusion duration, and 2-month-olds performed poorly even at the shortest duration, suggesting that object continuity emerges gradually between birth and 6 months. Using an eye tracker, we found that 4-month-olds rarely anticipate the ball's re-emergence, although performance in this condition was better than chance, suggesting some level of understanding object continuity. Six-month-olds anticipate on significantly more trials. When 4-month-olds first viewed the moving ball with no occluder, however, anticipations were much more reliable, suggesting that one mechanism of development of object continuity is simply experience seeing objects move. Object continuity, therefore, does not seem to be an innate visual function. Instead, experience viewing objects in motion may be crucial for its development.