Abstract
Many tasks are facilitated when the position of the target is known, even if the observer is fixating elsewhere. This facilitation is often ascribed to covert attention, which has been reported to increase contrast sensitivity, among other things. I sought to confirm these reports. On each trial, the observer saw two 30-ms flashes of a bright Gaussian (s = 0.37?) on a gray background and reported which of the two seemed brighter. The intensity of the brighter flash was randomly selected from a 5-dB range. Both flashes occurred in the same position, 3.6? away from a 1?-square fixation region. Trials in which fixation left this region were not analysed. Trials were also blocked by condition. In each cued condition, the position of the flashes never varied. It was indicated, 150 ms prior to each flash, by a 2? line segment with one end at fixation. In uncued conditions, the position of the flashes was randomly selected from one of four alternatives. These alternatives were indicated by the line segments used in the cued conditions. The four segments formed an X. Conditions were also either detection, in which there was no dim flash at all, or discrimination, in which the intensity of the dim flash was 5.2 times that required for 81% accuracy in the uncued detection condition. To achieve the same accuracy in the discrimination conditions, the brighter flash had to be 30% brighter than the dim flash, regardless whether the position were cued or not. That is, there was no effect of cueing on contrast discrimination. The cued detection threshold was 1 dB lower than the uncued detection threshold. This is consistent with signal-detection theory, which posits that four times as many independent channels are monitored in the uncued conditions. Neither separate paramaters for cued and uncued signal gains nor cued and uncued intrinsic uncertainty (nor both) could be warranted by c2 analyses. Thus there is no indication that covert attention can affect contrast sensitivity.