Abstract
Using the same set of 3D objects defined either by texture or shading we investigated whether the 3D shape is extracted from shading and texture in the same or different cortical areas. The 3D visual stimuli depicted randomly shaped roughly spherical objects. Functional images of the whole brain of twelve human subjects were obtained with a 3T MR scanner (Philips).
In the shading experiment the surfaces of the visual stimuli were based on Lambertian shading. One set of 3D shapes and five sets of matched control stimuli that appeared perceptually as a 2D luminance patterns were created. Random effect analysis was applied to compare 3D to all 2D conditions and masked inclusively by 3D versus each of the 2D conditions from the group (fixed effect). We found bilateral activation in the posterior inferior temporal gyrus (post-ITG) and right middle fusiform gyrus (mid-FG).
In the second experiment the object surfaces were filled with random dot patterns. Two types of 3D surfaces and four 2D controls were used. In analyzing the results of this experiment, we performed the same analysis as in the shading test. Within this activation pattern, only 6 regions reached the strict significance criterion: LOS and post ITG, symmetrically in the two hemispheres, and two dorsal parietal regions: DIPSM and DIPSA activated only in the right hemisphere.
There was a more extensive activation when subjects viewed 3D objects defined by texture in comparison with viewing 3D shapes based on shading. The activation patterns overlap however in post ITG, completely in the left hemisphere and partially in the right hemisphere. These results underscore the importance of the posterior part of the LOC complex for the extraction of 3D shape information, in agreement with earlier studies (Orban et al 99, Amedi et al 2001 and James et al 2002).