Attentive and pre-attentive processes in change detection and identification

Abstract

In studies of change blindness, observers often have the phenomenological impression that the blindness is overcome all at once, so that change detection, localization and identification apparently occur together. Three experiments are described that explore dissociations between these processes using a discrete trial procedure in which 2 visual frames are presented sequentially with no intervening inter-frame-interval. The results reveal that change detection and localization are essentially perfect under these conditions regardless of the number of elements in the display, which is consistent with the idea that change detection and localization are mediated by pre-attentive parallel processes.In contrast, identification accuracy for an item before it changes is generally poor, and is heavily dependent on the number of items displayed. Identification accuracy after a change is substantially better, but depends on the new item's duration. This suggests that the change captures attention, which substantially enhances the likelihood of correctly identifying the new item. However, the results also reveal a limited capacity to identify unattended items. Specifically, we provide evidence that strongly suggests that, at least under these conditions, observers were able to identify two items without focused attention. Our results further suggest that spatial pre-cues that attract attention to an item before the change occurs simply ensure that the cued item is one of the two whose identity is encoded.

Figure 1. The sequence and stimuli used in

Experiment 1. The dimensions of the stimuli and their eccentricities are illustrated in the upper left panel. The time line indicates the sequence of events within a trial, beginning with the appearance of the fixation point and ending with the report identity display. The location of the changed item is indicated by the arrow in Frame 2 (yellow diamond to yellow triangle).

Figure 2. Frame 1 identification accuracy in

Experiment 1. The percent correct identifications, averaged across subjects, are illustrated for color changes, shape changes or combined color and shape changes. Error bars indicate standard errors of the mean (+/−1 SEM).

Figure 3. Frame 2 identification accuracy in

Experiment 1. The percent correct identifications, averaged across subjects, are illustrated for each of the 3 Frame 2 durations (50, 100 and 500 ms) and each type of change (color, shape or color and shape). Error bars indicate +/−1 SEM.

Figure 4. Identification accuracy in

Experiment 2. Accuracy rates (averaged across observers) are plotted as a function of stimulus set size (3, 6, 12 and 18 items) and exposure duration (100 or 500 ms). Error bars indicate 1 SEM.

Figure 5. The percent correct color or shape identifications on error trials in

Experiment 2. These data represent the percentage of correct shape or color reports for those trials in which the deleted item was misidentified, and compares the obtained results to those expected by random guessing. Error bars indicated +/−1 SEM.

Figure 6. The percentage of correct identifications for each cuing condition in

Experiment 3. Averaged accuracy rates for the critical item in Frame 1 (squares) and Frame 2 (circles) are illustrated for 12 item (open symbol) and 3 item (filled symbol) displays as function of the cuing conditions (valid, neutral, invalid near and invalid far). Error bars indicate +/−1 SEM. See text for details.

Figure 7. Frame 1 Identification accuracy in

Experiment 3 on Invalid trials. Accuracy is a function of distance between the cued location and the location of the change. The gradient of attention is the same for the 3 and 12 item displays.

Figure 8. Obtained Frame 1 accuracy rates compared to predictions from the 2-items encoded hypothesis.

Predictions are compared with the data from Experiment 1 (large squares) and those in Experiment 3 (small squares) for all cuing conditions and display sizes.