Abstract
Scene perception relies on a set of cortical regions, such as the parahippocampal place area (PPA), retrosplenial complex (RSC), and occipital place area (OPA), exhibiting dissociable functional selectivity to various scene properties. Debates are ongoing about what specific types of visual information are represented in these regions to mediate their different functions. This fMRI study examined the neural bases of functional dissociations of scene-selective regions by selectively biasing visual inputs from the magnocellular (M) and parvocellular (P) cells and comparing patterns of their cortical projections. We manipulated 96 scene images to create stimuli that biased M- (low-contrast, achromatic images defined by luminance) or P- (defined by iso-luminant, red-green contrast) responses, adjusted for each participant’s thresholds. Twenty-four participants performed indoor/outdoor categorization of M- or P-stimuli presented for 1 second. Participants were significantly faster at categorizing P-biased scenes than M-biased, interestingly contrasted to the M-advantages reported in fearful/neutral face categorization (Cushing et al., 2019). For fMRI data analyses, each participant’s PPA, RSC, and OPA were functionally defined using a separate localizer scan. The PPA was significantly more responsive to P-stimuli in general, whereas the RSC showed greater responses to M-stimuli with a slight preference for outdoor images. The OPA activations did not show systematic M/P bias. We next tested whether the PPA and RSC preferences for P- and M-stimuli, respectively, were specific to the task of scene processing. In separate fMRI runs, participants viewed rapid flashes of an achromatic, low-spatial-frequency grating (M-biased) or slow alterations of a red-green, high-spatial-frequency grating (P-biased). Although the M-bias became weaker in the RSC, the P-bias of the PPA remained robust for the scene-irrelevant gratings. Our findings together demonstrate differential processing biases of the scene-selective regions for visual attributes conveyed from the retina to the cortex, facilitating the efficient perception of complementary scene properties.