fMRI correlates of visual cue combination

Recent models of cue combination (e.g. Landy et al., 1995, Vis Res, 35, 389) suggest that the perception of an object's 3D shape is mediated by the combination of different depth cues (e.g. disparity and texture). We used fMRI to investigate the neural substrates of this process. Methods: Vertical cylinders defined by binocular disparity and texture cues were rear-projected inside a 3T scanner. We studied cylinders in which both disparity and texture specified the same curvature (consistent cues), and cylinders in which the two cues specified different curvatures (inconsistent cues). Prior to conducting fMRI experiments, subjects adjusted the curvature of consistent cue cylinders so that their perceived curvature was similar to that of the inconsistent ones. Performance in this task was used to define stimuli with consistent cues that were perceived to have similar curvature to the inconsistent cue stimuli. We employed a sequential-presentation fMRI adaptation technique, in which stimulus changes result in increased fMRI responses compared with repeated presentation of the same stimuli. Results: We observed increased responses in the Lateral Occipital Complex (LOC) and hMT(V5)+ when the subjects discriminated curvature changes. In contrast, adaptation in these regions was observed when subjects perceived similar curvature. Conclusions: These findings suggest that visual areas LOC and hMT(V5)+ are involved in the processing of perceived 3D shape based on the combination of disparity and textures cues.