Incorporation of presleep stimuli into dream contents: evidence for a consolidation effect on declarative knowledge during REM sleep?

Presleep stimuli are often incorporated into dream contents. A study of 12 participants tested whether such incorporation improves memory consolidation. Participants memorized nonsense sentences before each sleep onset during a night, reported dreams and recalled sentences after each REM awakening, and gave unexpected morning reports. Dream reports were parsed into content units. Night reports more often contained valid incorporations than pseudo-incorporations. Content units common to night and morning reports—indicating better consolidation—were significantly more frequent for incorporated contents than for other contents. However, recall of words from incorporated sentences was not significantly better. The findings partially support a generation effect operating during sleep, benefiting retention of actively processed information.

Study at a glance

Abstract

Presleep stimuli to be retained for further recall is often incorporated into dream contents. To establish whether processing for insertion into dream contents may improve consolidation, we compared the retention rate at delayed recall of contents resulting from incorporation of presleep sentence-stimuli with those of other contents of the same dream experiences. We hypothesized that association with a cognitive task of recall facilitates access to recently acquired items of declarative knowledge such as presleep stimuli, and triggers the deep elaboration of their semantic features, which involves rehearsal. Twelve subjects were given a task of delayed recall for three nonsense sentences delivered once a time before each of the sleep (re-)onsets over an experimental night. After each awakening in rapid eye movement sleep, subjects were asked to report dream experience and recall the sentence to be retained. In the morning, after spontaneous awakening, subjects were unexpectedly requested to again report their dream experiences and to recall the stimuli. Two pairs of judges independently identified possible incorporations of the stimuli, and parsed dream reports into propositional content units. The proportion of night reports with at least one incorporation of the stimulus delivered (i.e. valid incorporations) was higher than that of reports with contents similar to a stimulus(-i) not yet delivered (forward pseudo-incorporations) or delivered prior to an earlier sleep period (backward pseudo-incorporations). The proportion of content units common to night and morning reports (considered to be better consolidated) was significantly higher for incorporated contents than for other contents, including pseudo-incorporated contents. Instead, the retention at morning recall of words of sentence-stimuli corresponding to incorporated contents was not significantly higher than that of other words. The better retention of incorporated contents provides a partial confirmation (that is, limited to the output of the processing) that a generation effect, which benefits retention of actively processed information, is operative during sleep as well as in waking.