Attention restores discrete items to visual short-term memory

Abstract

When a memory is forgotten, is it lost forever? Our study shows that selective attention can restore forgotten items to visual short-term memory (VSTM). In our two experiments, all stimuli presented in a memory array were designed to be equally task relevant during encoding. During the retention interval, however, participants were sometimes given a cue predicting which of the memory items would be probed at the end of the delay. This shift in task relevance improved recall for that item. We found that this type of cuing improved recall for items that otherwise would have been irretrievable, providing critical evidence that attention can restore forgotten information to VSTM. Psychophysical modeling of memory performance has confirmed that restoration of information in VSTM increases the probability that the cued item is available for recall but does not improve the representational quality of the memory. We further suggest that attention can restore discrete items to VSTM.

Fig. 1.

Study design (a, b) and main results (c, d) for Experiment 1. Participants were instructed to look at a monitor and remember the display, which consisted of an array of items such as those depicted at the left in (a). This display was presented for 200 ms. On half the trials, after an 800-ms delay, a square appeared, either in the center of the monitor (neutral retro-cue) or in the location where a memory probe would subsequently appear (valid retro-cue). The cue was presented for 200 ms and followed by an 800-ms delay. Finally, a single memory probe was shown, rotated either 20° or 45° clockwise or counterclockwise relative to the probed item. This probe appeared for 200 ms. On the remaining trials, there was no cue; instead, the memory probe was presented 800 ms after the offset of the memory array (i.e., at the same time that the cue appeared in the cued trials; short no-cue trials) or 1,800 ms after the offset of the memory array (i.e., at the same time that the probe appeared in the cued trials; long no-cue trials). The bar graph (c) shows the proportion of correct responses as a function of cuing condition. The line graphs (d) show the proportion of correct responses as a function of the probe’s angle of rotation (20° or 45°; collapsed across set size and across clockwise and counterclockwise rotations) and as a function of the set size of the memory array (4 items or 8 items; collapsed across angle of rotation). Asterisks indicate significant differences (p < .05). Error bars represent ± 1 SEM.

Fig. 2.

Behavioral results and parameter estimates for Experiment 2. In (a), the proportion of probes judged to have been rotated clockwise relative to the memory item is shown as a function of the probe’s angle of rotation for each of the four cuing conditions. The bar graphs show (b) estimates of the probability parameter (λ; lower numbers indicate higher probability of recall) and (c) estimates of the slope (β; higher numbers indicate greater precision) in each of the four cuing conditions. Asterisks indicate significant differences between conditions (p < .05). Error bars represent ±1 SEM. In (d), the proportion of “clockwise” responses in each condition is shown as a function of angle of rotation for 1 participant who completed four sessions of the task; the slope and asymptote parameters are also given for this subject. Each data point represents 96 trials (solid lines represent model fits).

Fig. 3.

Illustrative depiction of our functional model showing persistent activation of neural representations corresponding to four memory items, as in our experiments (see also Edin et al., 2009). A P in a black square indicates the presentation of a memory probe; a C in a red square indicates the presentation of a cue. The top two panels represent a trial in which a memory item is forgotten late during the maintenance interval of a long no-cue trial (left) and a trial in which a memory item is forgotten early during the maintenance interval of a short no-cue trial (right). If a memory item is cued before it is forgotten (lower left panel), the delay activity may persist until the memory probe is presented, perhaps also allowing other task-irrelevant items to be forgotten and thereby reducing interitem competition. However, even if the item is already forgotten at the time of the cue, our results suggest that its representation in visual short-term memory can be restored by retro-cuing prior to the presentation of the memory probe (lower right panel).