Chelsea Kim¹, Reebal Rafeh², Taylor Schmitz², Marieke Mur^1,3; ¹Department of Psychology, Faculty of Social Science, Western University, ²Department of Physiology and Pharmacology, Schulich School of Medicine & Dentistry, Western University, ³Department of Computer Science, Faculty of Science, Western University

Everyday vision involves multiple visual objects competing for attention, yet most neuroscience studies still rely on static, isolated stimuli. Here, we emulated naturalistic vision using simultaneously presented dynamic videos of faces and scenes. We used a recently developed frequency-tagged fMRI method (ft-fMRI; Ngo, Rafeh et al. 2024) to tag visual cortical BOLD responses synchronized to either face or place stimulation frequencies. We examined if object-object competition produces mutual suppression of face- and place-responsive regions in the visual cortex. Three adults freely viewed face videos (multiple identities, neutral/positive expressions) and place videos (e.g., kitchen, stairs, bedroom) in a 3T MRI scanner. Luminance contrast was modulated at 0.1 Hz for face videos and 0.125 Hz for place videos, allowing us to temporally disentangle responses in the frequency domain. Videos appeared centrally (face-center, place-center) or in a face-center-place-surround configuration; place videos were also shown in isolation in the surround. Each participant viewed each condition for eight minutes. Our ft-fMRI paradigm evoked oscillatory BOLD responses at the tagged frequencies. Centrally presented stimuli activated the expected face- and place-selective regions in lateral and ventral temporal cortex. Face videos also drove a lateral temporal social visual pathway (Pitcher & Ungerleider, 2021). When faces and places were presented simultaneously, face-evoked and place-evoked responses were reduced compared to isolated presentation, consistent with mutually suppressive competitive interactions between category-responsive pathways. We demonstrate that ft-fMRI provides access to fine-grained cortical maps of competing object representations under naturalistic viewing. This opens the door to investigating cortical computations subserving object-object competition and attention across the visual hierarchy and in lateral temporal pathways relevant to social cognition. Because the method leverages free viewing and frequency tagging, it holds strong potential for studying multi-object processing in infants and other challenging populations, providing a powerful new tool for naturalistic visual neuroscience.

Acknowledgements: NSERC