Abstract
Natural scenes contain multiple objects that compete for neural representation in visual cortex by mutually suppressing their activity. These competitive interactions can be biased both by top-down and bottom-up factors. Here, we probe the idea that the local neural circuits mediating competitive interactions provide an interface for attentional mechanisms to operate on. If so, the magnitude of attention effects will directly correlate with the degree of neural competition. To investigate the interaction of attention and competition, 3 different types of stimulus arrays were created that varied in their degree of perceptual organization (noPO, weakPO, strongPO), inducing varying degrees of competition as shown previously. Each stimulus array was presented while subjects either attended to the array or attended away, performing a task at fixation. The stimulus consisted of four illusory contour inducers, which were displayed in the upper right visual quadrant. In the strongPO condition, inducers were rotated inward forming an illusory figure (i.e. a square). In the weakPO condition, the inducers were aligned to form a percept of a loosely defined figure. In the noPO condition, inducers were rotated outward, forming no figure. To measure the degree of competition, the stimuli were presented either sequentially (SEQ) or simultaneously (SIM). Sensory suppression indexes (SSI) were calculated for each visual area (V1-V4), for each level of PO and each attentional condition: (SEQ-SIM)/(SEQ+SIM), larger indexes represented a greater degree of competition. In the unattended conditions, SSIs were largest for the noPO condition and smallest for the strongPO condition. When the stimulus array was attended, attentional enhancement was largest for the SIM noPO condition, resulting in a larger reduction of the noPO SSIs compared to the strongPO SSIs throughout extrastriate cortex. This suggests that the amount of attentional enhancement observed is dependent upon the amount of competition left unresolved after bottom-up processes occur.