Abstract
The fusiform face area (FFA) and occipital face area (OFA) are part of a network of face-selective neural populations in the ventral visual pathway. The FFA is typically associated with "holistic" face processing whereas the OFA is typically associated with the visual processing of face features, although the two presumably work together to achieve whole-face representations. We used an fMRI half-field repetition suppression technique to study the neural representation of hemifield-split chimeric faces. Observers viewed chimeric faces centrally, and the faces either repeated in full, changed in full, changed to the right of midline (RVF change) or changed to the left of midline (LVF change). We observed considerable evidence of repetition suppression when whole faces were repeated, both in bilateral face-selective cortex (FFA; OFA; and face-selective posterior fusiform, pFus), and also in some regions of retinotopic visual cortex. Whole-face repetition suppression was strongest in bilateral OFA and pFus. Of particular interest was the degree to which the OFA, pFus and FFA voxels in each hemisphere showed a contralateral bias for half-face changes. We found that OFA and pFus voxels showed clear contralateral bias for half-face changes, but nevertheless showed substantial "release" from repetition suppression for half-face changes in the ipsilateral visual field. In contrast, the FFA showed equivalent release from repetition suppression for half-face changes (RVF and LVF, no contralateral bias) and whole-faces changes. Whole-brain analysis showed greater overall contralateral bias in the right hemisphere as compared to the left, including portions of face-selective cortex. Our findings are consistent with greater receptive fields sizes in the FFA as compared to those in more posterior face-selective areas. Interestingly, our half-field method revealed results that were not observed in previous fMRI studies of contralateral bias for face stimuli in which entire faces were viewed in either the RVF or the LVF.
Meeting abstract presented at VSS 2017