Selective removal of visual working memory items at test

Working memory (WM) tasks often require comparing remembered items to test displays, but little is known about how people selectively remove irrelevant information at test. Across three experiments, we used contralateral delay activity (CDA) to track WM load and examine selective removal. In Experiment 1, CDA amplitudes increased with set size even when only one item was probed, suggesting minimal removal based on spatial location. Experiment 2 ruled out spatial grouping by presenting items sequentially in the same location, yet more items were retained for larger set size. In Experiment 3, however, when items belonged to distinct mnemonic categories, CDA amplitudes at test were reduced, consistent with selective removal based on category relevance. Additionally, P3 old-new effects showed that decision speed and strength were influenced by the number of items maintained. Together, these results suggest that people selectively remove WM contents based on categorical relevance, not spatial cues, enabling more efficient memory-based decisions.