Yu Zhao¹ (), Matthew W. Shinkle¹, Mark D. Lescroart¹; ¹University of Nevada, Reno

Recent neuroimaging work has used sensitive scans and analyses to describe several small, previously overlooked regions that respond to scenes, faces, and words. This suggests that there may be elements of the representation of these or other categories that known regions do not capture. Inspired by this work, we investigate whether there may be additional body-selective regions by analyzing a large-scale 7T fMRI dataset, the Natural Scenes Dataset. We first analyzed the high-field category localizer data in NSD. Contrasts of bodies and limbs versus objects revealed small and as-yet undescribed patches of selectivity in consistent anatomical locations across subjects: one located posterior to the parahippocampal place area (PPA) and another superior to the retrosplenial complex (RSC) along the parieto-occipital sulcus. We provisionally term these areas the Ventromedial Body Area (VMBA) and the Medial Body Area (MBA), respectively. Next, to validate selectivity and characterize tuning in these regions in naturalistic contexts, we analyzed responses to the NSD main experiment. We developed a voxelwise encoding model with features capturing body locations and counts in each stimulus image. This model accurately predicted responses in both VMBA and MBA and in established body-selective areas EBA and FBA. Variance partitioning against models of low-level visual features (Gabor filters) and scene category labels showed that body-related features explained unique variance in VMBA and MBA responses. Finally, analyses of encoding model weights and highly stimulating NSD images suggested that, while EBA and FBA are selective for centrally-presented bodies, VMBA and MBA are selective for multiple bodies in the peripheral visual field. Together, our results reveal two novel cortical regions which play a complementary role to established body-selective areas. These new regions may provide a basis for understanding bodies or actions in their spatial contexts and/or for directing attention to bodies in the visual periphery.

Acknowledgements: Supported by NSF CAREER # 2340895 to M.D.L.