Population receptive field mapping reveals different spatial computations across body- and face-selective regions

Abstract

Human high-level visual cortex contains clusters of neurons that preferentially respond to different categories of visual stimuli such as faces, bodies, places, and words. While past work has shown that face, word, and place regions seem to prefer stimuli presented in different portions of the visual field, no work has been done examining the spatial computations of body-selective regions in visual cortex. In this thesis, we perform population receptive field (pRF) mapping which is a powerful methodology for modeling the portion of visual space to which a given neuron (or a voxel in the case of fMRI) responds, known as the receptive field. Combining pRF mapping with a visual category localizer using fMRI, we demonstrate for the first time that body-selective regions contain uniquely peripheral and large receptive fields compared to face-selective regions, and that pRF properties even vary significantly within different body-selective regions. These findings hold implications for theories on visual cortex development, as body-selective regions contain pRFs that are much more peripheral than predicted by their cortical location.