Similarity judgments and cortical visual responses reflect different properties of object and scene categories in naturalistic images

Similarity judgments and cortical visual responses reflect different properties of object and scene categories in naturalistic images
复制标题

DOI:
10.1016/j.neuroimage.2019.04.079
复制
发表时间:
2019-08-15
期刊:
影响因子:
5.7
通讯作者:
Baker, Chris I.
Baker, Chris I.
中科院分区:
医学1区
文献类型:
--
作者:
King, Marcie L.;Groen, Iris I. A.;Baker, Chris I.

文献摘要

被引文献

相似文献

据报道,许多因素构成了高级人类视觉皮层中复杂图像的表征,包括类别(例如面部,物体,场景),动画和现实世界的大小,但这种组织在多大程度上反映了对现实世界刺激的行为判断尚不清楚。在这里,我们比较了来自明确的行为相似性判断和超高场(7 T)的功能磁共振成像的人类视觉皮层的多个样本的一组不同的自然图像从48个对象和场景类别的表示。虽然有一个显着的相关性之间的相似性判断和功能磁共振成像反应,有两个代表性的空间之间的显着差异。行为判断主要揭示了人为(包括人类)和自然(包括动物)图像之间的粗略划分,概念相关的类别(例如运输,动物)的清晰分组,而这些概念分组在功能磁共振成像表征中基本上是不存在的。相反,fMRI反应似乎主要反映了人类和非人类面孔/身体与所有其他类别的分离。此外,将行为和fMRI表征空间与来自深度神经网络(DNN)层的空间进行比较,显示出与最顶层行为和中间层fMRI的强烈对应关系。这些结果表明,在高层次的视觉皮层和行为相似性判断的本地化反应之间的复杂关系-每个域反映了图像的不同属性,和高层次的视觉皮层的反应可能对应于中间阶段的处理之间的基本视觉功能和概念类别,占主导地位的行为反应。
Numerous factors have been reported to underlie the representation of complex images in high-level human visual cortex, including categories (e.g. faces, objects, scenes), animacy, and real-world size, but the extent to which this organization reflects behavioral judgments of real-world stimuli is unclear. Here, we compared representations derived from explicit behavioral similarity judgments and ultra-high field (7T) fMRI of human visual cortex for multiple exemplars of a diverse set of naturalistic images from 48 object and scene categories. While there was a significant correlation between similarity judgments and fMRI responses, there were striking differences between the two representational spaces. Behavioral judgements primarily revealed a coarse division between man-made (including humans) and natural (including animals) images, with clear groupings of conceptually-related categories (e.g. transportation, animals), while these conceptual groupings were largely absent in the fMRI representations. Instead, fMRI responses primarily seemed to reflect a separation of both human and non-human faces/bodies from all other categories. Further, comparison of the behavioral and fMRI representational spaces with those derived from the layers of a deep neural network (DNN) showed a strong correspondence with behavior in the top-most layer and with fMRI in the mid-level layers. These results suggest a complex relationship between localized responses in high-level visual cortex and behavioral similarity judgments - each domain reflects different properties of the images, and responses in high-level visual cortex may correspond to intermediate stages of processing between basic visual features and the conceptual categories that dominate the behavioral response.