A complementary contribution of REM sleep after targeted memory reactivation during NREM sleep for memory consolidation

Presentation of learning-related cues during NREM sleep has been shown to improve memory consolidation. Past studies suggest that REM sleep may contribute to the beneficial effect of reactivating memories during NREM sleep, but the relationship between REM sleep and induced reactivations in NREM remains unclear. We investigated whether a naturally ensuing episode of REM sleep is necessary for prior NREM targeted memory reactivation (TMR) to exert a beneficial effect on memory consolidation. Nineteen participants learned the association between prior or non-prior known objects and their names (pseudowords) in a within-subject multiple session experiment, in which TMR was subsequently performed either before (Pre-REM) or after (Post-REM) the final REM sleep episode of the night. While word-picture association recall measures did not differ between TMR conditions, we found better name recognition based on confidence ratings for words reactivated during Pre-REM TMR in contrast with associations cued during Post-REM TMR. In addition, we found distinct associations between cue-evoked sigma activity, subsequent REM theta power and TMR memory benefits which were contingent upon the level of relatedness with prior knowledge for the learned material. Although TMR may be less effective during the second half of the night, our findings suggest an interplay between NREM and REM sleep oscillatory activity for memory reactivation and consolidation processes.