Abstract
Prior research on the position sensitivity of category selective regions has shown a coupling between face selectivity and foveal representations and place selectivity and sensitivity to the periphery (Levy 2001), as well as decreased tuning to position as one ascends the ventral stream hierarchy (Schwartzlose 2008). However, most studies have examined position sensitivity by comparing discrete locations in the visual field. We extend this prior work by using continuous retinotopic mapping and measuring population receptive fields (pRFs) in category selective regions. 12 subjects participated in an fMRI experiment that contained both a category localizer as well as retinotopic mapping using flickering checkerboard stimuli. By fitting a pRF model to each voxel, we determined the retinal position and the spatial extent of the visual field which best matched the observed BOLD response within each voxel. Then, in each individual brain we defined category-selective ROIs as well as visual field maps. Category-selective ROIs were subdivided by their intersection with visual field maps. Consistent with Aracaro et al 2009, place-selective voxels in the collateral sulcus overlap at least partially with the anterior ventral visual field maps (VO1-2; PHC1-2). pRF fits to these place-selective voxels show large receptive fields that have a strong bias towards the upper visual field and the periphery. By contrast, face-selective ROIs on the ventral surface (IOG-, pFus-, mFus-faces) did not overlap with known visual field maps, but contain pRFs that are foveally centered and their combined coverage of the visual field coverage is generally contained within the central 10째. Interestingly, pRFs in right hemisphere face selective ROIs show greater coverage of the ipsilateral visual field than those on the left, which may be related to the often-reported right hemisphere dominance for face processing in humans.
Meeting abstract presented at VSS 2014