Abstract
When we search for objects, we often know what we are looking for. However, sometimes we need to search when all we know is what we don’t need (e.g., looking for something in the buffet that you haven’t eaten yet). Prior behavioral research has shown that these negative searches are less efficient than search for targets with known features. In addition, when search tasks allow for either strategy, searchers elect to attend stimuli with known features. Here, we used event-related potentials to measure whether searchers choose to form positive templates when the option is available, even if instructed otherwise. Participants completed blocks of positive searches and negative searches where, before each search, a bilateral pair of colors was shown. The cued color in each pair always informed participants of the target color (positive search) or non-target color (negative search). The other color in the pair was random. In this context, we measured similar contralateral negativities to the two cues, showing that both cue types were stored in working memory as templates. As expected, negative searches were slower and less accurate than positive searches. A separate group of participants completed the same task with a critical change: every color-cue pair contained the same two colors that were used in subsequent search arrays. This meant that the non-cued color always matched the to-be-ignored color in positive search blocks, and the non-cued color always matched the to-be-attended color in negative search blocks. Contralateral negativities flipped sign in negative search blocks, demonstrating that searchers instead formed positive templates using the non-cued color. Behavioral data corroborated this conclusion. This provides new evidence that positive searches are preferred, and that the CDA reflects searchers’ decisions to use a cue.
Acknowledgement: NIE, NIMH, NIH