A fundamental question in systems neuroscience is: What are the organization principles of human visual cortex? Visual cortex originates in primary visual cortex and extends through a hierarchy of early, intermediate, and high-level visual regions separated across two processing streams (dorsal and ventral). While much is known about the organization of early visual cortex, much less is known about the organization of high-level visual regions in the ventral stream, which are thought to be involved in visual recognition. Current theories suggest functional distinctions between early and high-level regions in the ventral processing stream: early and intermediate visual regions contain a systematic representation of the visual field across a series of multiple maps (Wandell and Winawer, 2011), whereas higher-level regions are thought to be specialized for processing specific types of stimuli such as objects, faces, body parts, words, and places (Kanwisher, 2010). Several alternative theories suggest other principles for the organization of the ventral stream, including expertise (Tarr and Gauthier, 2000), eccentricity biases (Malach et al., 2002), or distributed representations (Haxby et al., 2001; Kriegeskorte et al., 2008). Nevertheless, two notions are common to these theories. First, it is widely accepted that different rules underlie the functional organization of high-level and early visual cortex. Second, the profile of activations in high-level visual cortex is thought to be more variable across individuals compared to early visual cortex. Contrary to the prevailing view, we propose common organization principles throughout early and high-level visual cortex, where functional regions have consistent anatomical locations and preserved spatial relationships to neighboring regions as well as retinotopic maps. Employing these principles enables the first framework for consistent parcellation of high-level visual regions, which can also be applied to other sensory and nonsensory cortical systems.
