Abstract
Goal: Covert spatial attention modulates visual performance via endogenous (voluntary, sustained) and exogenous (involuntary, transient) mechanisms. Most studies have investigated one type of attention or the other, and little is known about how the two interact. Here we explored how exogenous cues affect the magnitude of endogenous attentional benefits.
Methods: On each trial, spatial attention was manipulated, and two Gabor patches were simultaneously presented for 80ms (6.4° eccentricity to the left and right of fixation). A response cue appeared after their offset, indicating which was the target. Observers reported its orientation, clockwise or counterclockwise of vertical. Two endogenous and three exogenous cue conditions were crossed for a total of 6 conditions. The endogenous cues were valid or distributed. In valid trials, a white line appeared 400ms before the Gabors, pointing towards the location of the upcoming target (100% validity). In distributed trials, two white lines indicated that the target could appear at either location. The exogenous cues were valid, invalid or absent. In valid trials, a small white circle appeared for 60 ms, 100 ms prior to stimuli onset, just above the target location; in exogenous invalid trials, the circle appeared above the other stimulus; in exogenous absent trials, no circle appeared.
Results: Both exogenous and endogenous attention improved performance. For each endogenous cue condition, accuracy in exogenous valid trials was significantly greater than in exogenous invalid trials. For each exogenous cue condition, accuracy in endogenous valid trials was significantly greater than in endogenous distributed trials. There was no interaction: the magnitudes of the endogenous benefits in each exogenous condition were statistically indistinguishable from each other.
Conclusions: Exogenous attention improved performance but did not modulate the magnitude of endogenous attentional benefits. This suggests that these two types of attention draw on processing resources that are at least partially independent.
Meeting abstract presented at VSS 2013