The question of how a unified visual percept emerges from a complex array of neural inputs is central to models of perceptual grouping. We assume that the bistable percepts that emerge from figures known as “ambiguous” because they elicit two mutually exclusive and alternating percepts engage these grouping mechanisms. High resolution (1.5 × 1.5 mm) whole brain functional magnetic resonance images, fMRI, were acquired during the alternating percepts of a “Schroeder Staircase”. “Default” (original) and “alternative” percepts were indicated by corresponding button responses using both hands. Eye positions monitored during the scans failed to confirm a correspondence between eye position and perspective, and thus ruled out the influence of eye movements in these studies. Random effects group analyses (SPM) of the BOLD signal during the presence of both perspectives compared to a resting baseline revealed the engagement of a widely distributed neural system including visual cortex, right inferior lobule, right lingual gyrus, anterior cingulate, and thalamus. Further analyses of individual voxels on the high-resolution acquisition grid for each individual subject revealed signals within the right inferior parietal lobule that were uniquely associated with default and alternative percepts. Together, these findings are consistent with a multilevel global model of distributed processes that drive these perspectives, and further suggest a role for “highly influential” local neural ensembles within single voxels located in the right inferior parietal lobule for perceptual grouping of visual input.