Two influential studies (Ellenbogen et al., 2006, 2009) reported that a night of sleep, unlike an equivalent day of wakefulness, protects newly learned verbal paired associates from retroactive interference, and became key evidence that sleep actively stabilizes memories rather than merely shielding them from waking interference. This evidence is now contested: a well-powered direct replication by Bailes et al. (2020) found a near-null Sleep × Interference interaction, Pöhlchen et al. (2020) failed to replicate in both nap and full-night experiments, yet Abel et al. (2023) successfully replicated the original 2006 result online while flagging design confounds (ceiling performance from high learning criteria; a confound between interference presence and retention-interval length). No dedicated meta-analysis or published multi-lab registered replication exists as of August 2026. The open problem is to determine, with adequately powered, confound-controlled evidence: (i) whether post-learning sleep reduces retroactive interference for declarative material and with what population effect size; (ii) which design and population moderators (stimulus modality, learning criteria, retention-interval structure, circadian timing, age, encoding strength) explain the discrepant findings; and (iii) which sleep-stage physiology (slow wave sleep, sleep spindles, hippocampal ripples) mediates any protection. A decisive answer adjudicates between passive-shielding and active-stabilization theories of sleep-dependent memory consolidation.

Background

Declarative memory — memory for facts and episodes, dependent on the hippocampus — is retained better when the retention interval contains sleep. A central theoretical question is why. The passive account holds that sleep merely shields new memories from the interference that information processing during wakefulness imposes; active consolidation accounts hold that sleep-specific processes reorganize and stabilize memory traces themselves. Retroactive interference — impairment of previously learned material by later learning of overlapping material — is the classic phenomenon that distinguishes these accounts: if a memory is less vulnerable to interference that arrives after sleep has ended, sleep must have changed the memory rather than merely paused further input. The standard paradigm, adapted from Barnes and Underwood's (1959) A-B/A-C transfer design, has participants learn word pairs (A-B), then either sleep or stay awake for roughly 12 hours, then learn interfering pairs sharing the same cues (A-C), and finally recall the original A-B pairs. Ellenbogen et al. (2006, 2009) reported that interference learning was far less damaging after sleep than after wake — a Sleep × Interference interaction, meaning the effect of interference on recall depended on whether sleep or wake preceded it — and these studies are described as the most cited and most influential evidence for sleep's proposed protective role. The surrounding literature was already mixed: Sheth et al. (2012) found no sleep-dependent protection in the experiment that included an adequate no-additional-learning control, while Sonni and Spencer (2015) reported protection in older adults. The contest crystallized when Bailes, Caldwell, Wamsley and Tucker (2020) ran a direct, well-powered replication of Ellenbogen et al. (2009) — a sample more than twice as large, powered at 0.90 to detect the original effect — and observed a near-zero Sleep × Interference interaction (partial eta squared, a standard ANOVA effect-size measure, of 0.003), with the original effect (partial eta squared of 0.0985) falling outside the 90% confidence interval. They concluded that a small protective effect cannot be excluded, that the original effect may have been an overestimate or a false positive, and that resolving the question requires larger or meta-analytic studies, polysomnographic measurement (sleep recording with EEG and related signals, used to identify sleep stages), and characterization of task, modality, and procedural moderators. Independent failures to replicate followed: Pöhlchen, Pawlizki, Gais and Schönauer (2020) found no protection in either a nap or a full-night experiment and concluded the stabilizing effect had been overestimated. Reviews by Cordi and Rasch (2021), Zhang, Whitehurst and Mednick (2022), and Németh et al. (2023) place these failures within a broader pattern of small, task-dependent, weakly replicable sleep-memory effects and call for higher-powered, pre-registered, multi-lab and meta-analytic approaches. Yet Abel, Nickl, Reßle, Unger and Bäuml (2023) then replicated Ellenbogen et al. (2006) successfully in an online study, while showing that the original design contains features — ceiling-level recall produced by 100% learning criteria, and a confound in which interference conditions also have longer retention intervals before the final test — that could manufacture the interaction, and concluded that it is still unclear whether sleep protects memories from retroactive interference. Mechanism-adjacent evidence is similarly preliminary: Wang et al. (2023) reported positive napping effects on resistance to interference, and Jones et al. (2023) reported that closed-loop transcranial alternating current stimulation (electrical brain stimulation timed to ongoing slow oscillations) delivered during slow wave sleep reduced retroactive interference in low-encoding participants, but with small samples and no independent replication. This is the state of the field in which the problem stands.

Problem Statement

The overarching objective is to establish whether human sleep protects newly formed declarative memories against retroactive interference, and thereby to adjudicate between passive-shielding and active-stabilization accounts of sleep-dependent memory consolidation. The problem comprises three linked questions that jointly serve this single objective. (1) Existence and size: in healthy adults, in the Barnes–Underwood-style paired-associates paradigm (A-B learning, a sleep or wake retention interval, A-C interference learning, A-B recall), is the Sleep × Interference interaction non-null — does a retention interval filled with sleep reduce the forgetting caused by subsequent interference learning, relative to an equal interval of wakefulness? The existence question is settled by a population effect-size estimate precise enough to establish or reject a non-null Sleep × Interference interaction and to estimate its magnitude; bounds that only exclude effects as large as the originally reported partial eta squared of about 0.10 do not resolve it. (2) Moderators and boundary conditions: which design and population factors account for the discrepant literature — stimulus modality (verbal versus visual paired associates), learning criteria and the resulting ceiling performance, the confound between interference presence and retention-interval length, circadian timing (time of day relative to the body's daily rhythm) of encoding and test, nap versus full-night sleep, participant age, and encoding strength? (3) Mechanism: which sleep-stage-specific physiological processes — slow wave sleep (the deepest sleep stage, dominated by large slow cortical brain waves), sleep spindles (brief 12–15 Hz thalamocortical oscillations), or hippocampal ripples (fast oscillations accompanying hippocampal sharp waves) — carry any protective effect? A resolving answer addresses all three components: for (1), the population effect-size estimate specified above; for (2), an identification of which design and population factors account for the discrepant findings; and for (3), an identification of which sleep-stage-specific processes carry any protective effect. These components form one problem rather than three: a protective effect that is real but tiny, occurs only under specific procedural conditions, or has no identifiable physiological carrier supports a different theoretical account than a robust, stage-specific, generalizable protection, so no component can be dropped without changing the scientific claim being tested.

References

  1. Barnes JM, Underwood BJ (1959). Fate of first-list associations in transfer theory. Journal of Experimental Psychology 58:97-105. PMID: 13796886. https://pubmed.ncbi.nlm.nih.gov/13796886/
  2. Ellenbogen JM, Hulbert JC, Stickgold R, Dinges DF, Thompson-Schill SL (2006). Interfering with theories of sleep and memory: sleep, declarative memory, and associative interference. Current Biology 16(13):1290-1294. doi:10.1016/j.cub.2006.05.024. https://doi.org/10.1016/j.cub.2006.05.024
  3. Ellenbogen JM, Hulbert JC, Jiang Y, Stickgold R (2009). The sleeping brain's influence on verbal memory: boosting resistance to interference. PLoS ONE 4(1):e4117. doi:10.1371/journal.pone.0004117. https://doi.org/10.1371/journal.pone.0004117
  4. Sheth BR, Varghese R, Truong T (2012). Sleep shelters verbal memory from different kinds of interference. Sleep 35:985-996. doi:10.5665/sleep.1966. https://doi.org/10.5665/sleep.1966
  5. Sonni A, Spencer RMC (2015). Sleep protects memories from interference in older adults. Neurobiology of Aging 36(7):2272-2281. doi:10.1016/j.neurobiolaging.2015.03.010. https://doi.org/10.1016/j.neurobiolaging.2015.03.010
  6. Bailes C, Caldwell M, Wamsley EJ, Tucker MA (2020). Does sleep protect memories against interference? A failure to replicate. PLOS ONE 15(2):e0220419. doi:10.1371/journal.pone.0220419. https://doi.org/10.1371/journal.pone.0220419
  7. Pöhlchen D, Pawlizki A, Gais S, Schönauer M (2020). Evidence against a large effect of sleep in protecting verbal memories from interference. Journal of Sleep Research 30(2):e13042. doi:10.1111/jsr.13042. https://doi.org/10.1111/jsr.13042
  8. Cordi MJ, Rasch B (2021). How robust are sleep-mediated memory benefits? Current Opinion in Neurobiology 67:1-7. doi:10.1016/j.conb.2020.06.002. https://doi.org/10.1016/j.conb.2020.06.002
  9. Zhang J, Whitehurst LN, Mednick SC (2022). The role of sleep for episodic memory consolidation: Stabilizing or rescuing? Neurobiology of Learning and Memory 191:107621. doi:10.1016/j.nlm.2022.107621. https://doi.org/10.1016/j.nlm.2022.107621
  10. Abel M, Nickl AT, Reßle A, Unger C, Bäuml K-HT (2023). The role of sleep for memory consolidation: does sleep protect memories from retroactive interference? Psychonomic Bulletin & Review 30(6):2296-2304. doi:10.3758/s13423-023-02264-8. https://doi.org/10.3758/s13423-023-02264-8
  11. Wang Y, Meng Y, Zhang R, Du J, Huang Y, Zheng S, Xu S, Xiao L, Xu J, Wang H, Su T, Tang Y (2023). Positive effects of napping on memory consolidation and resistance against interference. Sleep and Breathing 27(6):2499-2507. doi:10.1007/s11325-023-02822-3. https://doi.org/10.1007/s11325-023-02822-3
  12. Jones AP, Bryant NB, Robert BM, Mullins TS, Trumbo MC, Ketz N, Howard MD, Pilly P, Clark VP (2023). Closed-loop tACS delivered during slow-wave sleep reduces retroactive interference on a paired-associates learning task. Brain Sciences 13(3):468. doi:10.3390/brainsci13030468. https://doi.org/10.3390/brainsci13030468
  13. Németh D, Gerbier E, Born J, Rickard T, Diekelmann S, Fogel S, Genzel L, Prehn-Kristensen A, Payne J, Dresler M, Simor P, Mazza S, Hoedlmoser K, Ruby P, Spencer RMC, Albouy G, Vékony T, Schabus M, Janacsek K (2023). Optimizing the methodology of human sleep and memory research. Nature Reviews Psychology 3(2):123-137. doi:10.1038/s44159-023-00262-0. https://doi.org/10.1038/s44159-023-00262-0
  14. Schmidig FJ, Geva-Sagiv M, Falach R, Yakim S, Gat Y, Sharon O, Fried I, Nir Y (2024). A visual paired associate learning (vPAL) paradigm to study memory consolidation during sleep. Journal of Sleep Research 33(5). doi:10.1111/jsr.14151. https://doi.org/10.1111/jsr.14151
  15. Gasparello A, Baldassarri A, Degasperi G, Cellini N (2024). The impact of sleep on factual memory retention over 24 hr. Journal of Sleep Research 33(6). doi:10.1111/jsr.14237. https://doi.org/10.1111/jsr.14237