Abstract
Models of human face perception propose parallel pathways. One pathway is responsible for processing changeable aspects of faces such as expression, and the other pathway is responsible for relatively invariant aspects such as identity. However, it remains unclear what image properties are encoded in these different visual pathways. Here, we ask how image properties from the internal and external features of the face contribute to the neural representation of facial identity and expression. In the first experiment, we measured patterns of brain activity using fMRI while participants viewed images posing different facial expressions (fear, anger, disgust, sadness, happiness). Next, we asked whether these neural patterns of response could be explained by variation in the image properties of the face when posing different facial expressions. We found that neural patterns of response to facial expressions in face-selective regions could be predicted by image properties from the internal, but not the external features of the face. In the second experiment, we measured patterns of brain activity using fMRI while participants viewed images of faces with different identities. We then asked whether these neural patterns of response could be explained by variation in the image properties across faces with different identities. In contrast to facial expression, we found that neural patterns of response to facial identity in face-selective regions could be predicted by the image properties from the external, but not the internal features of the face. These results reveal a direct link between image properties from different parts of the face and the neural representation of facial expression and identity in the human brain.
Meeting abstract presented at VSS 2015