Long‐Term Visual Gist Abstraction Independent of Post‐Encoding Sleep

• Published in: Journal of sleep research p. e70106
• Main Authors: Lutz, Nicolas D., Himbert, Johanna, Palmieri, Jessica, Kurz, Eva‐Maria, Raposo, Isabel, Yang, Xuefeng, Born, Jan, Rauss, Karsten
• Format: Journal Article
• Language: English
• Published: England 09.06.2025
• Subjects: sleep; gist; memory; consolidation; abstraction
• ISSN: 0962-1105; 1365-2869; 1365-2869
• DOI: 10.1111/jsr.70106

Current theories of memory processing postulate a slow transformation from episodic to abstract, gist‐like memories. We previously demonstrated that sleep shortly after learning improves gist abstraction in healthy volunteers across a one‐year retention interval using a visual version of the Deese‐Roediger‐McDermott (DRM) paradigm. Here, we investigate the temporal evolution of this effect by testing recognition performance on a similar DRM task immediately after encoding, as well as 1 week and 1 year later. Moreover, we address the role of feature overlap during encoding, using stimulus sets that are either closely related to or more distant from their common prototype. Behavioural data were obtained from N = 16 healthy volunteers in a within‐subjects design, where different sets of shapes were learned in separate experimental sessions, followed by consolidation during day‐time wakefulness or nocturnal sleep, respectively. Our results indicate high levels of (false) recognition of non‐encoded prototypes for all measurement points, including after 1 year. However, in contrast to our previous findings, gist memory was not affected by whether participants slept or stayed awake during the first 12 h after encoding. Comparisons across experiments indicate that the divergent results are due to changes in task demands rendering item and gist memory traces less distinct in the present study. Our results confirm the behavioural persistence of visual gist abstraction across extended intervals. At the same time, they highlight that sleep effects on this process are highly dependent on task demands.