Corentin Jacques¹ (), Jacques Jonas^1,2, Sophie Colnat-Coulbois^1,3, Bruno Rossion^1,2; ¹Université de Lorraine, CNRS, F-54000 Nancy, France, ²Université de Lorraine, CHRU-Nancy, Service de Neurologie, F-54000 Nancy, France, ³Université de Lorraine, CHRU-Nancy, Service de Neurochirurgie, F-54000 Nancy, France

Visual object recognition is a fundamental human brain function supported by a bilateral network of brain regions extending ventrally from the occipital pole to the anterior temporal cortex. Within this network, category-selective brain regions exhibit differential activity to their preferred category (e.g., faces) relative to other categories, as measured with various neuroimaging methods. While category-selectivity is generally defined as an increased neural activity for the preferred category, activity decreases to a specific category have also sometimes been described. Here, we investigate selective neural activity increase and decrease to face relative non-face stimuli with intracerebral recordings across the whole ventral occipitotemporal cortex (VOTC) of a large sample of participants (N=140). Objectively tagged face-selective responses in the high-frequency broadband range (HFB: 30-160 Hz) were found in 10% of recording sites distributed over the whole VOTC, with regional peaks of activity around and along the fusiform gyrus, leading to four key observations. First, face-selective HFB neural activity is characterized both by response increase (signal+ : 57% of sites) and decrease (signal- : 43% of sites) relative to other non-face categories, with a clear lateral-medial anatomical distinction along the VOTC for signal+ and signal- face-selective sites. Second, signal+ activity manifests clearer markers of face-selectivity: higher amplitude and face-selectivity, right hemispheric dominance,... Third, while increases and decreases exhibited mostly overlapping time-courses, slightly shorter response durations were found for response decrease. Four, signal- sites reflect either a lower response to faces than nonface objects (~70% of signal- sites) or an active suppression in response to face images (~30% of sites). Overall, our observations suggest a potential role for category-selective decreases in neural activity in visual (face) recognition.

Acknowledgements: Funding: ERC AdG HUMANFACE 101055175