Neural and computational dissociations between objects, scenes, and near-scale reachspaces
暂无分享,去创建一个
Space-related processing engages a network of brain regions separate from those engaged in object-related processing. This dissociation has largely been explored using images depicting a navigable scale of space compared to singleton objects. However, this scheme does not account easily for near-scale reachable spaces, which are not navigable but typically contain more than one object. To examine how these views are processed in the brain, human participants underwent functional neuroimaging in which brain responses to near-scale “reachspaces” were compared with responses to scenes and objects. We found evidence for three regions that prefer reachspaces to both scenes and objects: one in ventral visual cortex, one in occipito-parietal cortex, and one in superior parietal cortex. Furthermore, we found that both objectand scene-preferring ROIs were substantially driven by reachspaces, although to an intermediate degree. Finally, we provide computational evidence using deep convolutional neural networks that these three scales of space have separable visual features, potentially accounting for some of the differences in neural representation. Taken together, these results show that perceptual processing of reachspaces may require specialized neural circuits, and may also draw on both objectand scene-based processes.
[1] Marcel A. J. van Gerven,et al. Deep Neural Networks Reveal a Gradient in the Complexity of Neural Representations across the Ventral Stream , 2014, The Journal of Neuroscience.
[2] N. Kanwisher,et al. The lateral occipital complex and its role in object recognition , 2001, Vision Research.
[3] Nancy Kanwisher,et al. A cortical representation of the local visual environment , 1998, Nature.