Neural correlates of 3D visual orientation

点击翻译按钮获取中文摘要

DESCRIPTION (provided by applicant): We can see things in three dimensions (3D) because the visual system reconstructs the 3D configuration of objects from their two-dimensional images projected onto the retina. Previous studies have described loss of 3D binocular vision and constructional apraxia after parietal lesions, although the neurophysiology of this effect remains poorly understood. Many studies have characterized the neural basis of binocular disparity processing, although few have dealt with the representation of 3D surface orientation. The proposed studies use a multi-faceted approach and promise to provide important insights into the role of the parietal cortex in 3D vision and visual orientation constancy. Specifically, i the proposed studies we examine neural selectivity for 3D surface orientation by presenting planar stimuli with different 3D orientations while neuronal responses are recorded from the caudal intraparietal area (CIP), anterior intraparietal area (AIP) and area V3a of macaque monkeys. We employ a multi-faceted approach, combining neural recordings, behavior, population decoding & chemical inactivation. Neurons in these areas are tuned to tilt manipulations in surface orientation, as defined by either binocular disparity or linear perspective and texture gradients. Here we test whether tuning of V3a/CIP/AIP neurons represents all combinations of tilt and slant, as well as whether tilt and slant tuning curves are independent of the defining cue, stimulus position in the frontoparallel plane and in depth. Furthermore, we also explore the role of this area in visual orientation constancy, i.e., the abiliy to maintain the percept of the scene as a whole stably oriented as head/body orientation changes in the world. We hypothesize that, through a combination of partial tuning curve shifts and gain changes, population activity in CIP and AIP achieves visual orientation constancy. This would allow a representation of the 3D orientation of objects in world coordinates, a property that is necessary to implement many of our interactions with the environment. In addition, we will use Fisher information analyses to compute how population thresholds vary as a function of tilt and slant reference orientation. We hypothesize that the properties of CIP/AIP neurons can predict behavioral thresholds and their dependence on reference orientation. Finally, we will directly test for a causal role of each of these areas in slant discrimination by first recording ad then manipulating neural activity while macaques perform a fine slant discrimination task. Neural firing rates will be analyzed using signal detection theory and neuronal sensitivity will be compared to behavioral sensitivity. Using the same dataset, we will also compute 'choice probabilities' to establish whether trial-to-trial variability in neuronal responses is correlated ith variability in perceptual choices. We will then probe for causal links between neurons in these areas and 3D surface orientation perception by employing reversible inactivation. Together, these studies constitute a state-of-the-art multi-faceted approach and will provide a vital test of the hypothesis that this circuit contributes to 3D vision.