Abstract
Hogben and DiLollo (VR, 1974) displayed two sets of dots in succession. If superimposed, the two sets formed a 5×5 matrix with one missing location. The task was to determine that location, and has been interpreted as a test of the duration of visual persistence. Dixon & DiLollo (Cognitive Psychology, 1994) developed a temporal correlation model that accounted for a large amount of data with varying stimulus durations. The model has two main components: (1) computation of the neural response to each set of dots (which is both delayed and prolonged as compared to the actual stimulus), and (2) determination of the overlap/non-overlap ratio of the two neural response functions. Greater relative overlap predicts a higher probability of temporal integration, which in turn predicts a higher probability of detection of the empty location.
The present study was designed to test the model in a novel way. If the non-overlapping trailing portion of the temporal response to the second interval can be eliminated or reduced, the relative overlap of the two intervals will increase. A mask stimulus, consisting of random static noise, was presented at two different delay durations following the second set of dots: 0 ms and 60 ms. Detection of the missing dot position was significantly better with the shorter delay. Presumably, the mask being dissimilar to the set of dots erased the neural response to the second set of dots, thereby increasing the relative overlap of the first two sets of dots.