Ying Zhou¹ (), Daryl Fougnie¹; ¹Department of Psychology, New York University Abu Dhabi

We can flexibly control the contents of working memory. Retro-cue paradigms reveal that explicitly cued information in working memory can be selectively prioritized (Souza & Oberauer, 2016); ‘Choose the best’ paradigms indicate that participants have metacognitive knowledge of memory quality and can use this knowledge to select the best remembered information in working memory (Fougnie et al., 2012). While both selections can enhance behavioral performance of the selected item, it remains unclear whether they rely on the same neural mechanism. We used fMRI to measure the neural activity during metacognition-based and cue-driven selections (8 participants). Specifically, on each trial, participants memorized the orientation of three items at different locations on the screen, and a cue during the memory delay would prompt a specific type of selection: either a spatial retro-cue indicating the location of the to-be-tested item or a self-report cue indicating that they would report the best-remembered item. The visual input across the two selection conditions is comparable, and the only difference is whether participants selected an item based on a retro-cue or on their metacognitive knowledge of memory quality. We found different brain regions are recruited by these two types of selection. Cue-driven selection, which required top-down spatial attention, preferentially engaged the precentral sulcus, whereas metacognition-based selection, which required an internally generated readout of memory quality, selectively recruited the dorsolateral prefrontal cortex. These findings indicate that there are separate neural mechanisms for internally- (metacognition-based) versus externally-driven (cue-driven) working memory selection.

Acknowledgements: This work was supported by the NYUAD Center for Brain and Health, funded by Tamkeen under NYUAD Research Institute grant CG012