Abstract
Distinct concepts, such as animals and tools, often share vast and systematic differences in visual features. To what extent the neural representations in the category-selective regions in the occipitotemporal cortex contain visual vs. conceptual information? While it has been suggested that animacy, shape, and spatial frequency are three main factors to explain the lateral-to-medial activation patterns in the occipitotemporal cortex, these factors are often confounded (e.g., animals tend to have curvilinear shapes whereas tools are often elongated). To investigate the effects of these three factors on the representations in the animal- and tool-selective regions, we used low- or high-passed images of round or elongated animals and tools that shared comparable image statistics, as measured by GIST descriptors. Participants (n=20) performed a one-back task in the MRI scanner. Although the animal and tool images were indistinguishable based on image statistics, univariate analyses revealed robust category selectivity for such images in the animal-selective regions (bilateral lateral occipital complex and lateral fusiform) and tool-selective regions (left posterior middle temporal gyrus and medial fusiform), which were defined in a localizer with animal and tool images of naturally varied image statistics. Representational similarity analyses further revealed that the activation patterns in the animal-selective regions were most correlated with a model that represents only animal information, whereas the activation patterns in the tool-selective regions were most correlated with a model that represents only tool information. Moreover, shape or spatial frequency, compared with category information, appeared to contribute less significantly to the neural representations in these regions. Taken together, the distinction between animal and object representations in the occipitotemporal cortex is unlikely merely due to differences in low-level visual or shape properties. Instead, the involvement of higher-level conceptual influences may shape category selectivity and play a critical role for interpreting visual input during categorization.