Characterizing the information content of a newly hatched chick's first visual object representation

Abstract

How does object recognition emerge in the newborn brain? To address this question, I examined the information content of the first visual object representation built by newly hatched chicks (Gallus gallus). In their first week of life, chicks were raised in controlled-rearing chambers that contained a single virtual object rotating around a single axis. In their second week of life, I tested whether subjects had encoded information about the identity and viewpoint of the virtual object. The results showed that chicks built object representations that contained both object identity information and view-specific information. However, there was a trade-off between these two types of information: subjects who were more sensitive to identity information were less sensitive to view-specific information, and vice versa. This pattern of results is predicted by iterative, hierarchically organized visual processing machinery, the machinery that supports object recognition in adult primates. More generally, this study shows that invariant object recognition is a core cognitive ability that can be operational at the onset of visual object experience. This study measured the information content of the first visual object representation built by newly hatched chicks. The results showed that chicks build object representations that contain both object identity information and view-specific information, akin to the object representations built by adult primates. This study indicates that invariant object recognition is a core cognitive ability that can be operational at the onset of visual object experience.