An important goal of my research is to investigate the "constructive nature of visual perception", i.e. how the brain creates a three-dimensional world of objects from two-dimensional retinal raster images. It is astonishing that the brain is doing this transformation in such a reliable way that we normally do not realize that the visual input is not simply mapped onto the brain but is actively processed by numerous brain areas. We get a glimpse of these constructive processes when experiencing visual illusions like apparent motion or illusory contours. I have studied these and related phenomena using functional Magnetic Resonance Imaging (fMRI.) to unravel the underlying neural mechanisms.

Since the pioneering work of Ungerleider and Mishkin, much experimental evidence has supported the existence of two main visual pathways in the primate cortex; a ventral stream devoted to the fine-grained analysis of the visual scene, including the perception of form and color, and a dorsal stream which processes spatial characteristics of the visual scene and analyzes motion. I will report about fMRI studies investigating interactions of areas within and between the ventral and dorsal pathways. Some features of the dorsal pathway could be unraveled in using several motion paradigms including apparent motion, shape-from-motion and motion imagery. These experiments lead to interesting insights into the relationship between bottom-up and top-down processing in showing an activity gradient along the areas of the dorsal processing stream. Interactions between the dorsal and ventral stream were investigated using a structure-from-motion paradigm. The results of this study clearly demonstrate constructive brain processes in showing that a small number of appropriately moving dots is sufficient to differentially activate highly specialized areas in the ventral processing stream. I will further report about studies with blindsight patients which were conducted to learn about the relation between conscious perception and brain activation. After presenting stimuli in the "blind" visual field, ipsilesional extrastriate areas in the lateral occipital cortex (hMT+/V5 and LO) and the fusiform gyrus (V4/V8) were highly activated although no activity was detectable in early visual areas. These results imply that activation in ipsilesional ventral visual cortical areas is insufficient to generate conscious visual perception in hemianopic patients.