Abstract
To study the decay of visual Short-Term Memory for contrast, we adapt a paradigm original used by Lu, Williamson and Kaufman (Science, 258, 1668–1669, 1992) for the study of auditory STM for loudness. We explicitly create an intermediate-term memory of the average contrast of all the stimuli in the session (context memory). The STM of the contrast of a particular stimulus decays to the context memory.
In Experiment 1, a 0.27 sec exposure of a narrow-bandwidth Gabor patch of contrast 0.3 is followed after a variable delay by a second Gabor patch whose contrast is randomly chosen from a set which has a mean contrast of 0.34 in HIGH sessions and a mean of 0.26 in LOW sessions. Observers judge which stimulus, 1 or 2, has the higher contrast. Experiment 2 uses a roving standard. Experiment 3 uses a fixed standard and broad-bandwidth stimuli.
Results. Context stimuli have no effect on simultaneous judgments but influence judgments more-and-more as delay increases. STM for the contrast of fixed-contrast narrow-bandwidth Gabors is poor. With a roving standard or with fixed-standard broad-bandwidth Gabors, exponential decay of STM to context memory occurs with a time-constant of several seconds that depends on the observer. The Standard Deviation of judgments does not vary across delays. The constancy of SD across delays suggests that the same memory comparison register is used in all conditions; the contents consist of decreasing proportions of STM and increasing proportions of context memory as delay increases.
Air Force Office of Scientific Research, Life Sciences, Grant FA9550-04-1-0225