Abstract
Top-down factors such as attention and expectation have been shown to modulate visual discrimination by improving the quality of sensory processing. However, few studies have investigated these factors within the same experiment to determine if they operate via similar or independent mechanisms (Kok et al., 2012; Jiang et al., 2013). To address this question, subjects viewed a display containing flickering and randomly oriented lines, half of which were blue and half of which were red. On each trial, some percentage of the lines were rendered at the same orientation, either clockwise or counterclockwise relative to the vertical meridian. We manipulated attention (behavioral relevance) by cueing subjects to monitor either the red or blue lines (focused attention) or to monitor both (divided attention). Expectation of the target-defining orientation was also manipulated by varying the ratio between target orientations presented in a given block of trials. Finally, we examined interactions between the strength of sensory evidence and attention/expectation by manipulating the number of iso-oriented lines in the target display across two levels (low and high coherence). At high coherence levels, there is no main effect of attention or expectation on accuracy, but both factors lead to faster response times (RTs). In contrast, at low coherence levels, there is a robust effect of expectation on accuracy such that expected orientations are more likely to be correctly discriminated. However, this expectation effect is primarily observed in the divided-attention condition, suggesting that focused attention dampens the effects of expectation. In addition, at low coherence levels, there is a main effect of attention on RTs but little effect of expectation. Taken together, these results suggest that the effects of attention and expectation on behavioral performance during orientation discrimination are dissociable depending on the amount of relevant sensory information available in the display.
Meeting abstract presented at VSS 2015