Abstract
Humans are perceptually sensitive to the configuration of features in faces and objects and can accurately distinguish a real face or object from a scrambled one. In contrast, for textures, humans are relatively insensitive to variation in feature configuration. However, the cortical representation of objects in human ventral visual cortex is non-selective for feature configuration, suggesting objects are encoded as a texture-like feature representation. It is widely believed that faces are a special category, so it is possible face-selective domains of visual cortex are uniquely selective for face feature configuration. Further, it remains unknown whether this configural selectivity is dependent on exposure to faces during early visual development. Here, we recorded multi-unit electrophysiological activity from inferior temporal (IT) cortex of 6 macaques—4 reared typically and 2 reared without experience of faces during infancy—as they viewed images in rapid serial visual presentation. We presented natural images of faces and objects, as well as synthesized images, generated via deep texture synthesis, which contained similar features to their corresponding natural images in a scrambled configuration. We found that in typically-reared monkeys, IT cortex was more selective for face feature configuration than object feature configuration. The population response to a natural face was more distinct from scrambled faces than the response to a natural object was from scrambled objects. However, in the face-deprived monkeys, IT activity was not significantly more selective for face feature configuration than for objects. Thus, in the absence of early childhood exposure, faces, like objects, were encoded in a texture-like representation. These results provide evidence that although the neural representation of faces in IT is uniquely selective for feature configuration compared to other objects, this configural selectivity for face features is neither innate, nor a fixed property of face stimuli, but rather is dependent on early experience with faces.