Human hippocampal replay during rest prioritizes weakly learned information and predicts memory performance

Abstract

The hippocampus replays experiences during quiet rest periods, and this replay benefits subsequent memory. A critical open question is how memories are prioritized for this replay. We used functional magnetic resonance imaging (fMRI) pattern analysis to track item-level replay in the hippocampus during an awake rest period after participants studied 15 objects and completed a memory test. Objects that were remembered less well were replayed more during the subsequent rest period, suggesting a prioritization process in which weaker memories-memories most vulnerable to forgetting-are selected for replay. In a second session 12 hours later, more replay of an object during a rest period predicted better subsequent memory for that object. Replay predicted memory improvement across sessions only for participants who slept during that interval. Our results provide evidence that replay in the human hippocampus prioritizes weakly learned information, predicts subsequent memory performance, and relates to memory improvement across a delay with sleep.

Conflict of interest statement

The authors declare no competing interests.

Figures

Fig. 1

Stimuli and design. a Examples of stimuli presented from the three classes Alpha, Beta, and Gamma, labeled with unique code names. Satellites were built randomly for each participant, using the same category structure. b Sessions 1 and 2 procedures for all participants, with delay between sessions either overnight or across day depending on group

Fig. 2

Behavioral results. Change in proportion correct from the first to second session for Sleep and Wake group, for different feature types and the average across feature types; §p < 0.1, *p < 0.05, ***p < 0.001, one-tailed t-test. Asterisks above horizontal lines show significant differences between conditions; asterisks without bars indicate where conditions differ from zero. Gray circles indicate performance of individual subjects. Error bars denote ± 1 SEM; n = 12 in each group

Fig. 3

Replay analysis. a Each satellite’s template consists of a pattern of beta weights across all voxels in the hippocampus. The template is correlated with the preprocessed pattern of activity across all voxels in the hippocampus at each TR of the rest period. b This results in a matrix of correlation values between all templates and all rest period TRs, which is then thresholded to reflect “potential replays”. c The potential replay activity for a given template is summed across TRs and then correlated with memory performance. The inset scatterplot shows data for one representative subject’s Session 1 correlation

Fig. 4

Within-session replay–behavior relationships. a Correlation between memory in Session 1 and replay in Session 1, and between memory in Session 2 and replay in Session 2, in left (L Hipp) and right (R Hipp) hippocampus. b Same data, broken down by Sleep and Wake groups. c Same data, broken down by memory for unique and shared features. d Same data, broken down by memory for verbal and visual features. §p < 0.1, *p < 0.05, **p < 0.01, ***p < 0.001, t-test. Error bars denote ± 1 SEM; n = 24 for all bars except b, where n = 12 for each group

Fig. 5

Across-session replay–behavior relationships. Correlation between replay in each session and improvement in memory from the first to second session, as well as the average for each group across sessions; §p < 0.1, *p < 0.05, **p < 0.01, t-test, one-tailed for comparing Sleep to zero and Sleep to Wake. Error bars denote ± 1 SEM; n = 12 in each group