Sleep and testing both strengthen and distort story recollection

Abstract

Over time, memories lose episodic detail and become distorted, a process with serious ramifications for topics such as eyewitness identification. What are the processes which contribute to such transformation over time? We investigated the roles of post learning sleep and retrieval practice in memory accuracy, transformation, and distortion, using a naturalistic story recollection task. Undergraduate students listened to a recording of the “War of the Ghosts”, a Native American folktale, and were assigned to either a retrieval practice or listen only study condition, and either a sleep or wake delay group. Sleep and retrieval practice independently resulted in more story elements being recalled accurately, and fewer importations of non-story elements, than the wake, no retrieval practice group. However, sleep and retrieval practice also led to more inferences of non-presented, but story related information. These findings suggest that both sleep and retrieval practice contribute equally to narrative memory stabilization and distortion.

Recall and recognition of discourse memory across sleep and wake

“…Following a 12-hr delay filled with either overnight sleep or daytime wakefulness, participants recalled the story in a free recall procedure, which was scored by assigning each recalled proposition into one of seven categories: accuracy, gist, omission, inference, normalisation, incorrect placement, and importation. Denis et al (2021b) found that the Sleep group outperformed the Wake group in terms of accuracy (i.e., more veridical propositions being recalled). Potentially, this is related to War of the Ghosts having weakly associated discourse elements, which, in turn, increased the likelihood of sleep exerting an effect.…”

“…Findings from a recent study support this view. Denis et al (2021b) exposed American undergraduates to a Native American folklore, entitled War of the Ghosts. First used by Bartlett (1932), the story was written in a non-Western style that is unfamiliar to most, if not all, people in Western countries.…”

“…In other words, less distortion seemed to have taken place in those who had a sleep opportunity, providing support for the proposal that a more nuanced scoring approach might be needed to fully capture the effect of sleep on discourse memory. Despite this, it is unclear if the findings from Denis et al (2021b) generalise to schema-consistent stories, as the War of the Ghosts story represents the extreme end of the schema-consistency spectrum, which is rare in daily life. In order to further our understanding of the interplay between sleep and discourse memory, it is important to increase ecological validity by using stories that are more comprehensible and representative of natural language (e.g., stories with discourse elements that have stronger associative links).…”

“…In order to further our understanding of the interplay between sleep and discourse memory, it is important to increase ecological validity by using stories that are more comprehensible and representative of natural language (e.g., stories with discourse elements that have stronger associative links). In addition, the scoring system adopted by Denis et al (2021b) is subjective in nature; for instance, the distinction between inference and importation is debatable-what is considered an inference by one person 6 may be considered an importation by another. This potentially makes it harder for future studies to replicate and highlights a clear need for more objective scoring metrics.…”

“…The four pre-registered categories to which a response in the fill-in-the-blank task was assigned We note here that while this kind of graded (vs. all-or-nothing) approach allows us to capture the nature of any errors, the process of response categorisation is subjective in nature-a limitation in the existing literature that was highlighted in the introduction (Denis et al, 2021b). As a remedy, we conducted an exploratory analysis, where we investigated the utility of a more objective scoring approach, namely Latent Semantic Analysis (Landauer & Dumais, 1997).…”