On the fate of non-cued mental representations in visuo-spatial working memory: Evidence by a retro-cuing paradigm

By means of a visual retro-cuing paradigm, we investigated the influence of working memory representations that are not required for an ongoing task. A retro-cue reliably indicated the items on the left or right side of a memory array that remained relevant for a subsequent recognition task. On ERP level, the attentional selection of cued memory contents was reflected by a contralateral negativity that appeared at temporal sites in an interval following retro-cue presentation (i.e., 200-400ms). Additionally, the retro-cue caused a suppression of posterior induced alpha power contralateral to the relevant memory items reflecting the release of non-cued information from active maintenance in visuo-spatial working memory. Following the retro-cue, a central probe item was presented and participants had to state whether this stimulus was previously shown on the relevant side of the memory array. The probe was either a cued memory item (cued probe), non-cued memory item (non-cued probe) or was not shown in the previous memory array (new probe). While the same response was required, the non-cued probe condition featured slower response times referred to the new probe condition. In line with these results, a comparison of the probe-locked ERPs pointed toward additional memory retrieval and conflict control mechanisms when processing non-cued probes compared to new probes. Thus although retro-cues were used to drop non-cued information from working memory, it still interfered with the processing of new sensory information based on residual spatial representations in a passive and fragile short-term memory store.