Modularity and intersection of "what", "where" and "how" processing of visual stimuli: a new method of FMRI localization.

Research on the modularity of perceptual and cognitive processes has often pointed to a ventral-dorsal distinction in cortical pathways that depend upon the nature of the stimuli and the task. However, it is not clear whether the dorsal, occipital-parietal stream specializes in locating visual objects (i.e., a "where" stream), or taking action toward objects (i.e., a "how" stream), although there is some consensus for a ventral, occipital-temporal "what" stream that specializes in the identification of visual objects. It is also not clear to what extent word and picture processing are modular along these streams, as functional imaging maps to date have not addressed the modularity question directly. Here we present two types of functional imaging maps that directly show modularity and intersection of processing function for word and picture stimuli in tasks that require decisions about "what is", "where is", or "how do you interact with" a stimulus (N=6 participants). Our results reveal a middle dorsal "how" stream with some modular regions of activation that are distinct from activation during "where" processing, and that words and pictures involve several modular regions of activation along these streams.