Metacontrast masking uses a special type of masker and is distinguished from other types of masking in the pattern of results it generates. The target is often a small dot and the mask is an annulus that surrounds the dot. Each is presented very briefly (10-100 milliseconds). In a metacontrast study, the onset of the target and the mask (Stimulus Onset Asynchrony, SOA) is varied systematically, and the quality of the target's percept is measured.
The effects are quite interesting. If the target and the mask are presented simultaneously (SOA=0) the target is easily seen. If the mask follows the target by several hundred milliseconds (e.g., SOA=300 ms) the target is seen to appear and disappear before the mask arrives. For intermediate SOAs, though, the target is perceptually absent.
The effect of SOA is counterintuitive. It suggests that as the mask becomes more temporally distinct from the target it becomes a stronger masker. This seems strange because if you imagine processing of visual stimuli to proceed in the order of their arrival, any inhibitory effects of the mask should weaken as it is temporally separated from the target. Clearly something else is going on, and there are several theories of metacontrast masking.
The properties of metacontrast masking have motivated theories of visual processing at the neural level and have also been used as a basis for several theories of consciousness. For example, if the mask renders the target unperceived, then were you conscious of the target before the mask appeared? What does conscious awareness of the target mean?
This lab allows you to participate in a version of a metacontrast masking experiment.
You task is to identify (or guess) the location of the missing dot. You do this through the keyboard. Press "u" if the missing dot was on the upper left; "i" if the missing dot was on the upper right; "j" for the lower left; and "k" for the lower right. You may press multiple keys, and the computer will keep the last key pressed as your guess.
Press the space bar to start the next trial. There are 10 different SOAs and 15 trials for each SOA condition. At the end of the experiment a graph will appear that plots the percentage of correct detections as a function of SOA. You should find that detection is near 100% for SOA=0 and for long SOAs. Detection should be worse (chance is 25%) for middle SOAs.
Copyright Purdue University 1998