Abstract
We can measure how many items working memory can maintain, but how well are these items represented within working memory? One way to measure representational fidelity is through a color identification task in which subjects have to recall and select a color on a color wheel. In such tasks, storing more items results in a reduction in the fidelity of retrieved representations. One potential explanation for the loss of fidelity is that subjects are encoding an ensemble representation of the individual stimuli. To test this hypothesis, colored items were presented simultaneously or sequentially, which we had expected to alter the ensemble statistics. Our hypothesis was that stimuli presented sequentially would have less interference than stimuli presented together. In a series of experiments, three colored squares appeared equidistantly from a focal point, and after a delay, subjects had to select a probed color-memory on a color wheel. To minimize blind guessing, subjects were able to opt out of responding by selecting a box in the lower right corner. In experiment 1, subjects saw 3 colors either simultaneously for 500ms or sequentially presented for 500ms each. In experiment 2, the total duration of the 3 sequential stimuli (166ms) was matched to that of simultaneous condition (500ms). To compare retention durations across different simultaneous conditions, stimuli appeared at either 500, 1000, or 1500 ms from the onset of the trial. Surprisingly, subjects’ deviations from the probed color were approximately equal in both the simultaneous and sequential conditions for experiments 1 and 2. These results indicate that the ensemble statistics did not differ between the sequential and simultaneous conditions. The implications of these results are that stimuli presented sequentially are not maintained as separate representations and exhibit interference from ensemble statistics.
Meeting abstract presented at VSS 2013