Abstract
Stimuli which are masked by metacontrast can nevertheless affect motoric responses. In a typical experiment, two squares are presented sequentially (both rotated by either 0 or 45 degree). The first square (the “prime”) is masked by the second square (“the target”), such that perceptual discrimination of the orientation of the prime is very low. However, if subjects are asked to respond quickly to the orientation of the target by pointing left or right, the prime still affects the trajectory: If the orientation of the prime is inconsistent with the orientation of the target, pointing first goes in the wrong direction. This result might be interpreted as a dissociation between perception (we cannot discriminate the prime's orientation) and action (the prime's orientation nevertheless affects pointing). However, it is very difficult to get discrimination performance really to zero, such that this dissociation might not be very convincing. Here, I tested whether the effect on action breaks down after taking very serious measures to suppress discrimination performance. For this, I used stimuli which are known to produce a good metacontrast suppression, reduced the contrast of the prime to very low values, and also presented the stimuli at unpredictable positions circular around the fixation point (metacontrast is known to be stronger if the stimuli are not fixated). Results show that with this procedure discrimination performance is almost zero, but the cost of this “perfect” suppression in perception is that the effect on action also breaks down. These results are consistent with the notion that the effects on action under normal conditions (with imperfect suppression of discrimination performance) are generated by similar sources as the residual discrimination performance.