Distinct neural correlates for focusing on similar memory contents originating from current or previous experience

Abstract

Flexible, goal-directed behavior depends on the ability to select and prioritize information from memory representations freshly encoded from the sensory stream as well as retrieved from previous experience. The spatial gating signatures of internal attention in working memory (WM) are increasingly well characterized, but it remains unclear whether the same neurophysiological mechanisms are recruited for orienting attention to items from long-term memory (LTM). We recorded EEG and eye movements while participants focused on WM versus LTM representations in a unified precision-report task. Retrocues improved performance for both WM and LTM items, but with larger behavioral gains for WM items. Neural and oculomotor markers of spatial orienting, including contralateral posterior alpha suppression and gaze biases toward remembered locations, were robust when focusing on WM items; but were reliably weakened or absent when focusing on LTM items. Multivariate pattern analyses provided complementary evidence for the recruitment of dissociable neural mechanisms when focusing on WM vs. LTM items, which unfolded with similar time courses. Together, the results establish the existence of dissociable neural mechanisms for internal attention, which can be deployed flexibly depending on the relevant memory trace to guide performance. The findings raise interesting and tractable questions about whether differences in representational formats or representational domains drive the distinct internal attention mechanisms.