Abstract
In most attention capture tasks, prioritization of an object that has a unique, but non-predictive stimulus feature is thought to reflect a stimulus-driven shift of attention because subjects should treat the feature as irrelevant. If they cannot do so, however, they might intentionally shift attention to the critical feature on some trials and away from it on others. To explore this possibility, we introduced temporal predictability such that the critical feature was present at the target location every n-th trial where n was the number of items in the display. If subjects detect and use this regularity, they should find the target more efficiently with temporal predictability than without. Not surprisingly, subjects use the temporal pattern when informed of its presence in advance. When uninformed, however, no subjects explicitly noticed the pattern, and in most conditions, performance with the pattern was no different than without it; subjects did not spontaneously use the pattern to aid their search. As an additional measure of pattern detection, we included a single “surprise” trial at the end of a block of trials that violated the temporal pattern, giving the search target the critical feature earlier than expected. If subjects detected and used the pattern to guide search, they should be slower on this surprise trial, as they would have focused attention elsewhere in the display. Yet, subjects who were uninformed about the pattern were significantly *faster* on average on this surprise trial. This finding suggests that subjects implicitly detected the temporal pattern and that it influenced the speed with which they processed the target when the appearance of the critical feature violated the pattern.
Research supported by NIH grant R01 MH63773-02