Abstract
In human fMRI studies, response amplitude of the superior intra-parietal sulcus (IPS) has been shown to track the encoding and maintenance of task relevant visual information, such as color, shape, and orientation. Consistent with this finding, we recently showed that, depending on the task demand, shape, location, or the conjunction of the two could be decoded from fMRI response patterns in this brain region. Here we investigated whether representations formed in superior IPS are limited to visual sensory information, or whether they can be extended to abstract visual information, such as viewpoint-invariant face identities. To address this question and to reduce the contribution of low-level visual features, we used face images of Leonardo Dicaprio and Matt Damon, two well-known actors matched in overall appearance. To further encourage the formation of viewpoint-invariant face identity representations, we varied viewpoint, hairstyle and facial expression of the face images and constructed two image sets for each actor. Observers viewed each image set multiple times and detected the occasional presence of an oddball face drawn from one of eight other male actors (oddball face responses were removed from subsequent analysis). In superior IPS, we found significantly higher correlations of fMRI response patterns between two sets of faces belonging to the same than different actors. This effect was not caused by the encoding of the actors’ names, as the effect went away when we repeated the task but only showed the actors’ names. This effect was also absent in the fusiform face area, the lateral occipital and the inferior IPS regions. Task relevant visual information can therefore be robustly represented in superior IPS even for abstract visual information such as face identity. This capability likely places the superior IPS as a key neural mechanism mediating the moment-to-moment visual information processing in the human brain.
Meeting abstract presented at VSS 2013