Abstract
Are capacity limitations in the storage of individual features due to (a) a set of discrete, fixed-resolution “slots,” or (b) a limited resource that can be allocated flexibly to provide more or less accurate representations depending on the number of items represented? To address this question, we conducted a series of color change-detection experiments. In the first experiment, we factorially varied set size (1, 2, or 3 items) and the magnitude of the color changes. For large change magnitudes, both hypotheses predict that performance should be at ceiling across these small set sizes. For small change magnitudes, the resource hypothesis predicts that performance should decline as set size increases because the amount of resources per item decreases, whereas the slot hypothesis predicts that performance should be equally degraded at all set sizes because resolution cannot vary. Consistent with the slot hypothesis, we found the decreasing the change magnitude led to equivalent impairments at all set sizes. In a second experiment, subjects were cued to remember either one or both of two items on a given trial, and small change magnitudes were always used. Consistent with Experiment 1, subjects were no more accurate when cued to remember one item than when cued to remember both items. A third experiment assessed whether requiring greater resolution for one object would lead to decreased resolution for other concurrently represented objects. Subjects were given an individual colored item and an additional three colored items to remember on each trial. The change magnitude for the individual item varied across blocks, which was intended to change the resolution requirements for this item. We found that increasing the resolution requirements for the individual item did not lead to decreased performance for the set of three items, which is consistent with the proposal that memory resolution is fixed and cannot be increased by the addition of resources.
This research was made possible by grants R01 MH63001 and R01 MH65034 from the National Institute of Mental Health.