Purchase this article with an account.
Joseph Schmidt, Annmarie MacNamara, Gregory Hajcak, Gregory Zelinsky; ERP correlates of the target representation used to guide search. Journal of Vision 2011;11(11):1345. doi: https://doi.org/10.1167/11.11.1345.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Effective search guidance requires the creation, maintenance, and matching of a target representation in memory to objects in the environment for the purpose of generating a guidance signal. We investigated the representation used to guide search by simultaneously recording EEG activity and eye movements while observers searched for pictorially-cued or categorically-cued targets. Observers simultaneously previewed a picture and a basic-level object name for 400ms, followed by a ∼2.75 s delay and finally a search array depicting four real-world objects. Under blocked instructions either the pictorial or categorical cue (not both) was designated as the search target. The task was to fixate the target and simultaneously press a button; targets were always present. Trials were segregated into strong-guidance (target was the first object fixated) and weak-guidance (target was not the first object fixated) groups, and the contralateral delay activity (CDA, an electrophysiological marker for WM load which is modulated by task-relevant features and/or locations), was computed in response to cue onset for each group and cue condition. We evaluated a time window of 300–900 ms after cue onset and found that the magnitude and direction of the CDA at lateral-parietal sites interacted with the type of cue. Pictorial cues showed common CDA, consistent with target related visual details being held in WM. However, categorical cues showed inverted CDA (contralateral more positive than ipsilateral), possibly suggesting a target representation less reliant on visual information. Moreover, lateral-parietal sites were not found to distinguish between strong and weak search guidance. Only after exploring CDA at frontal sites in the same time window did we find activity correlated with later guidance to the search target. In accordance with the biased-competition model, we interpret this as evidence for the frontal modulation of the representation used to guide search via the selection or biasing of task-relevant target features.
This PDF is available to Subscribers Only