Abstract
Visual working memory (VWM) is a limited resource, which may be distributed discretely or continuously, as predicted by opposing theoretical models. When presented with distracting information, it is most efficient to ignore or minimally process it. However, in many situations some items are more relevant than others and the target/distractor distinction is less clear. In a discrete resource allocation model, distractor items should be ignored completely to process target items in the limited slots of memory. In contrast, in the continuous model, distractors may be processed minimally with preference given to target items. One event-related potential (ERP) associated with VWM is sustained posterior contralateral negativity (SPCN), which is typically shown to scale with increasing load. In the present study, we used the SPCN to determine the extent to which low-priority items are processed. Participants were presented with four lateralized coloured objects while recording ERPs in three cue conditions: one-cue with 100% validity (no priority to non-target items), one-cue with 50% validity (low-priority to non-target items), and four-cues with 100% validity (all items given priority). In the 50% valid condition any of the uncued items could be probed; thus, these items should be allocated a portion of VWM resources in order to report the colour correctly. Results demonstrate that in the low-priority condition the SPCN amplitude was between the one-cue and four-cue conditions; thus, these items are not processed as targets (as in the four-cue condition) or ignored (as in the one-cue condition), but are meaningfully processed according to their priority. Further, frontal markers indicate that this condition required more executive control to preferentially maintain target items while meaningfully holding the low-priority items in memory. These results suggest that ERP markers of VWM maintenance reflect minimal but meaningful processing of low-priority items, as predicted by a continuous resource model.
Meeting abstract presented at VSS 2016