Abstract
Shape-selective regions have been identified in multiple human visual system regions, but it remains unclear what aspects of shape information drive these diverse regions. “Shape” can mean scale-invariant image features, but can also denote unitization of a spatial region. This study used simple rectangles to identify cortical regions that responded to integrated spatial dimensions, figural goodness, or both, and the extent to which these responses depended on selective attention. 15 subjects participated in 4.5 hours of fMRI each while performing 1-back matching tasks. Within a given trial block, rectangles varied by aspect ratio, surface area, or major axis length. Trial blocks also differed according to the portrayal of the rectangles: either a closed line drawing of the rectangle, fragmented vertex lines corresponding to the rectangle's corners, or rotated vertices that reduced the rectangle's figural goodness. The visual matching task formed part of an intermodal attention experiment with four main conditions: visual task only, visual task with ignored auditory stimuli, auditory task only, and auditory task with ignored visual stimuli. We contrasted the two bimodal conditions (visual task with ignored sounds, and vice-versa), which contained identical stimuli, to examine the effects of selective attention on visual responses. Imaging results showed that anterior ventral visual cortex was divided into lateral and medial bands that responded selectively to variations in aspect ratio and single-dimension changes, respectively. The anterior inferior temporal sulcus, extending to perirhinal cortex, was selective for aspect ratio vs. area changes; anterior parahippocampal cortex was activated more by major axis changes than by either aspect ratio or area changes. The intraparietal sulcus showed attention-dependent activation following a lateral (well-formed figures) to medial (poorly-formed) gradient. The fact that rectangle line drawings produced extensive and robust extrastriate activation emphasizes the explanatory power of perceptual integration for understanding visual cortex functional organization.
Supported by grant DCD5814 to DLW and the VA Research Service.