Sleep Benefits Memory for Semantic Category Structure While Preserving Exemplar-Specific Information

Abstract

Semantic memory encompasses knowledge about both the properties that typify concepts (e.g. robins, like all birds, have wings) as well as the properties that individuate conceptually related items (e.g. robins, in particular, have red breasts). We investigate the impact of sleep on new semantic learning using a property inference task in which both kinds of information are initially acquired equally well. Participants learned about three categories of novel objects possessing some properties that were shared among category exemplars and others that were unique to an exemplar, with exposure frequency varying across categories. In Experiment 1, memory for shared properties improved and memory for unique properties was preserved across a night of sleep, while memory for both feature types declined over a day awake. In Experiment 2, memory for shared properties improved across a nap, but only for the lower-frequency category, suggesting a prioritization of weakly learned information early in a sleep period. The increase was significantly correlated with amount of REM, but was also observed in participants who did not enter REM, suggesting involvement of both REM and NREM sleep. The results provide the first evidence that sleep improves memory for the shared structure of object categories, while simultaneously preserving object-unique information.

Conflict of interest statement

The authors declare that they have no competing interests.

Figures

Figure 1

Stimuli and task. (a) Examples of satellite stimuli presented from the three classes Alpha, Beta, and Gamma, each labeled with unique code names. Satellites were built randomly for each participant, using the same category structure. Shared and unique features of one satellite are highlighted. (b) Training phase 1: one-by-one introduction to each satellite. (c) Training phase 2: participants attempt to fill in one missing feature of a satellite, receiving feedback. The test trials are the same but have two features missing. (d) Overview of protocol for Experiment 1. (e) Overview of protocol for Experiment 2. Time spans in (d) and (e) are not to scale.

Figure 2

Experiment 1 results. Change in proportion correct from first to second session for unique features, shared features, and novel item features. For each feature type, results are shown for low frequency (LF), medium frequency (MF), and high frequency (HF) category, as well as the average (Avg) across categories. *p < 0.05, **p < 0.01, ***p < 0.001, uncorrected. Asterisks above horizontal lines show significant differences between conditions; asterisks without bars indicate where conditions differ from zero. Error bars denote ±1 SEM.

Figure 3

Experiment 2 results. Same structure as in Fig. 2. *p < 0.05, **p < 0.01, uncorrected.

Figure 4

Relationship between NREM and REM sleep and behavioral change. (a) Improvement in low frequency shared features for the NREM group, the REM group, and for REM groups on either side of a median split by number of minutes in REM. (b–d) Relationship between minutes of nap spent in REM and performance improvement for LF, MF, and HF shared features. *p < 0.05, **p < 0.01.