Joint encoding of binocular disparity and direction of motion

Abstract

There are competing views regarding the existence of joint encoding of binocular disparity and direction of motion within the human visual system. It is known that cells within the sensory cortices can have preferential tuning to more than one dimension of a stimulus. For example, most cells in the human middle temporal area (hMT+), which is considered a critical part of the cortical motion processing pathway, are sensitive to both binocular disparity and direction of motion. An important question is how these stimulus dimensions are encoded in such cells and in what way this affects the manner in which information can be extracted from them. Here, data from both electrophysiology and computational studies is reviewed and compared in order to evaluate the current arguments for and against joint encoding of these stimulus dimensions. Physiological data shows that this type of joint encoding does exist in cats and monkeys, but there is not current evidence proving its existence in human visual cortex. The computational models that are discussed attempt to either prove or disprove the necessity of joint encoding of these dimensions, but are found to lack the generalizability needed to provide a definitive answer. To gain new insights into the possible role of joint encoding of binocular disparity and direction of motion, a novel neuroimaging experiment is proposed to investigate the existence of joint encoding in human visual cortex.