Abstract
Depth is a frequently overlooked aspect in vision research, despite the fact that recognizing and perceiving depth cues are essential when it comes to appropriately interacting with our surroundings. Behavioral and physiological studies have provided a solid framework for understanding depth perception, but we have yet to establish the precise neural organization of depth representation in human visual cortex. Here, we map depth representations using a phase-encoded stimulus that travels along the z-axis (depth), analogous to the standard 2D retinotopic mapping paradigm (wedges and rings; Engel et. al., 1994; Sereno et. al., 1995). Our stimulus was a large 2D patch filled with black and white moving dots. The stimulus moved forwards or backwards (in alternating runs) through 13 discrete depth planes, completing a full cycle every 28 seconds. Red/green anaglyph glasses were used to achieve depth perception while in the scanner. Using a standard phase-encoded cross-correlation approach, we found voxels selective for different depth planes in several intermediate and later visual areas. Preliminary results suggest that depth representations may be organized into a large-scale map-like representation across V3d, V3A, V3B, and V7. Our findings provide critical insights regarding the neural correlates of depth representations, and we detail for the first time depth maps in human visual cortex using fMRI.
Meeting abstract presented at VSS 2015