Abstract
We used fMRI to ask whether stimulus driven face selectivity exists in human prefrontal cortex, as found in the macaque (OScalaidhe et al., 1999). In Experiment 1, faces, bodies, tools, and scenes were presented in a blocked design, with either a free viewing task (N=8) or a 1-back task (N=8). Unbiased ROI analyses showed right posterior inferior frontal gyrus (pIFG) selectivity for faces in the 1-back task, and for faces and bodies with passive viewing. The right fusiform face area (rFFA) showed a comparable (although more selective) pattern. Experiment 2 (N=14) tested faces, bodies, and objects, and their component parts, in an event-related passive viewing paradigm. The global rFFA response was stronger than in pIFG, but both areas showed similarly strong body and face selectivity (relative to objects) for wholes, with little (rFFA) or no (pIFG) selectivity for parts. Experiment 3 (N=6) focused on the contribution of the eyes. Whole faces, faces with the eyes masked, pairs of eyes alone, and flowers were passively viewed in a blocked design. The two areas responded differently: while the rFFA responded maximally to whole faces, the pIFG responded most to the eyes alone, and did not distinguish between faces with eyes masked and flowers. We conclude that there is significant face-selectivity in human pIFG. Experiment 3 indicates that this selectivity is driven by different aspects of the face compared to rFFA, likely including the social, emotional, or attentional characteristics of human eyes.