Publisher's Synopsis
For the successful recognition of objective, `real' motion based on visual cues it is necessary to take self-induced motion signals into account, such as those induced by eye-movements. During a series of fMRI studies we measured responses of visual and parietal regions to motion cues derived from (a) retinal motion, (b) eyemovements (visual pursuit) and (c) objective, (real) motion. We show that the recently described cingulate sulcus visual area (CSv) is not, as implied before, primarily driven by 3D self-motion cues but favoured 2D translational coherent motion over 3D expanding flow fields. Further, we found that V3A is capable of integrating retinal motion with eye-movements, thus allowing V3A to respond to object motion independent of retinal motion. This allowed us to define a new functional localizer for area V3A. Finally, we showed that activity in the foveal representation of the early visual cortex is driven by a combination of retinal input and by error signals as hypothesized by of Rao and Ballard (1999) for predictive coding. Taken together, this work provides evidence that regions V3A and CSv are key regions concerning visual self-motion processing and that early visual regions might be modulated by feedback from higher motion processing regions.
