Abstract
The capacity of the visual working memory depends on the precision of the representations. Previously we have shown very limited working memory capacity for detecting changes in the shape (phase structure) of the radial frequency (RF) patterns. We tested whether the memory capacity is limited also when the shape is defined, not by the phase structure, but by the amplitude of RF patterns. Two different tasks were used for estimating the capacity: delayed discrimination and recall. The delayed discrimination thresholds were measured with 2-interval forced choice setup using Quest procedure. In the first interval, 1–6 stimulus items were presented with random amplitudes (.1–.5), RFs (3, 4 or 5), and phases (0–360 deg). In the second interval, after the 1.5 s blank period, the amplitude of one of the items was either increased or decreased. The observers' task was to say whether the changed item had higher amplitude in the first or in the second interval. In the recall experiment, the second interval consisted of a cue box and a probe item, and the observers' task was to adjust the amplitude of the probe to match the amplitude of the item in memory. Gaussian functions were fitted to the distribution of the adjustment errors. For one item, the amplitude thresholds were low (0.01–0.05) and the adjustments precise (standard deviations 0.01–0.03). As the number of items increased from 1 to 6, there was a linear, 6–14 -fold increase of the thresholds (0.14–0.29) but only 3-fold increase of the standard deviations (0.03–0.10). There were individual differences, but no sudden or complete breakdown in performance was observed for any subject. The results confirm the tradeoff between memory capacity and precision, and show that the amount of the tradeoff depends on the individual observers and the task by which the memory capacity is measured.