Xieyi Liu³, Yipeng Li³, Wei Jin³, Pinglei Bao^1,2,3; ¹School of Psychological and Cognitive Sciences, Peking Univ., Beijing, China, ²IDG/McGovern Institute for Brain Research, Peking Univ., Beijing, China, ³Peking-Tsinghua Center for Life Sciences, Peking Univ., Beijing, China

Object images consist of both shape and texture, two fundamental components of visual recognition. Previous behavioral studies have identified shape as the primary feature for object classification in humans, with evidence indicating that the inferotemporal (IT) cortex is organized primarily around object shape, underscoring its crucial role in classification. However, the mechanisms by which the IT cortex extracts and utilizes texture information for object classification remain poorly understood. To address this gap, we conducted functional magnetic resonance imaging (fMRI) studies on macaques to identify regions within the IT cortex that are selective for animate shape and texture. For texture, we contrasted animate textures (e.g., fur) with inanimate textures (e.g., wood), while for shape, we contrasted silhouette images of animate versus inanimate objects. Both contrasts revealed several bilateral regions extending from the posterior to the anterior IT cortex. Interestingly, the shape- and texture-selective regions in the middle IT cortex were anatomically adjacent but exhibited minimal overlap, whereas in the anterior IT cortex, these two regions showed significant overlap. Electrophysiological recordings using Neuropixels further confirmed the presence of two distinct neuronal populations in the middle IT cortex of macaques, each with a strong preference for either animate texture or shape. This clear separation was not observed in the anterior IT cortex, where neurons showed a combined preference for both shape and texture. These findings suggest that object classification in the middle IT cortex involves parallel processing pathways for shape and texture information, with distinct neuronal clusters dedicated to each feature type. In the anterior IT cortex, however, these features are integrated, resulting in a more comprehensive representation. Our results reveal a novel hierarchical structure within the IT cortex: object information is decomposed in the early IT, but integrated at a later stage.