Integration of local features into visual shapes in the human visual cortex

Abstract

The perception of visual shapes entails that local features are integrated into global visual forms. The human lateral occipital complex (LOC) has been implicated in shape processing. We tested whether the LOC is involved in the integration of local features into visual shapes using event-related fMRI at a 1.5T scanner. The stimuli consisted of a. random patterns; that is, displays of randomly oriented and positioned gabor elements and b. target shapes; that is, displays with a closed contour of collinear gabor elements embedded into a background of randomly oriented gabors. We observed stronger fMRI responses in the LOC for target shapes than for random patterns, suggesting that the LOC represents visual shapes and not simple image features. We further manipulated the detectability of the target shapes by varying the alignment of the gabor elements. Misalignment of the local gabors resulted in decreased target detection performance and fMRI responses in the LOC. In contrast, we observed improved detection performance and increased responses in the LOC when the segmentation of the target shapes from their background was facilitated by additional visual cues, such as motion or disparity. Our findings suggest that neural populations in the LOC are involved in the integration of local image features and the representation of visual shapes.

Footnotes

 Supported by Max Planck Society