KurisutinaCitation: John D. Rudoy, Joel L. Voss, Carmen E. Westerberg, and Ken A. Paller. Strengthening Individual Memories by Reactivating Them During Sleep. Science 326(5956), 1079; published 20 November 2009. Peer-reviewed Brevia. DOI, public author-lab article and SOM.
Reading scope: Read the complete main article, figure/caption, all supplementary methods, and Tables S1–S3. Visually checked the main page and supplementary tables page. The combined eight-page PDF contains the article, six SOM pages, and a publisher abstract-page capture. Download succeeded over HTTPS with default certificate verification; no TLS verification bypass was used. No raw-data, code, or independent plot reanalysis was performed. Provenance records sources, versions, SHA256 hashes, and scope. No open redistribution license was identified.
Evidence, concise paraphrase: Twelve adults learned 50 object–sound–location associations before a nap. Twenty-five sounds were selected by pairing items on baseline spatial accuracy and randomly cueing one item per pair. During non-REM sleep, those cues preceded better subsequent location recall: mean error 1.07 versus 1.23 cm for cued versus uncued objects (paired t11=2.6, p<0.05). Forgetting was also smaller, with that advantage in ten participants. Five additional recruits could not sleep. A separate twelve-person wake control showed errors of 1.15 versus 1.32 cm (t11=1.4, p=0.18). Thus the raw mean advantages were approximately 0.16 and 0.17 cm; no sleep-by-cue interaction or equivalence test was reported. Sleep staging and recognition judgments supported cue presentation without reported awareness. Cue-evoked scalp potentials covaried with subsequent forgetting after an outcome-based median split. This supplies a randomized within-person behavioral cueing result, not decoding of replay content, identification of its hippocampal mechanism, or demonstration of cortical information transfer.
Method audit:
- Assignment and statistical unit: Locations were randomized for each person. Cue assignment was randomized within pairs matched on pre-nap error, preserving a useful within-person control. Reported behavioral tests use participant-level paired comparisons with eleven degrees of freedom. The items are nested within people and matched pairs; no crossed item model or item-level population inference was reported.
- State comparison: The wake group was a separate cohort performing an attention-demanding digit task, with instructions to focus on it despite the sounds. Random assignment to sleep versus wake was not described. A significant sleep result and a nonsignificant wake result do not establish a difference between states. The similar point estimates do not establish equivalence either.
- Sleep and physiology: Cue delivery occurred during slow-wave sleep in ten participants and stage 2 in two, judged offline by a rater blind to cue timing. EEG, eye, and chin electrodes were used. Stable alpha power and absent reported cue awareness support sleep delivery, but do not rule out every transient arousal or form of awareness. The repeated baseline sound also differs from the individually presented memory cues in repetition history.
- Outcome: The target is location error and change from baseline after a daytime nap. The physiological association groups trials using the later outcome; it is not a prospectively validated decoder of remembered locations or an independently manipulated mediator.
Original scientific reasoning: The sound identifies an object without specifying its randomly assigned location. An improvement in the corresponding spatial response therefore motivates a content-specific memory account more strongly than a cue that already supplies the correct location would. However, better later access does not uniquely identify strengthened storage, a particular neural replay event, or the destination of any transferred biological information.
The three replay studies now answer different questions. Girardeau links disruption of ripple-associated activity to impaired spatial performance. Bendor and Wilson link a learned cue to a change in decoded spatial reactivation. Rudoy links selected cues during sleep to later behavioral retention. Their convergence motivates a mechanism to test; it does not jointly demonstrate the entire causal sequence within one preparation or establish that the same mechanism accounts for every result.
For extraction, a reminder can change the very response being sampled. Improving historical accuracy and preserving a person's pre-cue memory are different objectives. An acquisition protocol should therefore date its observations and retain which reminders were presented, rather than treating later recall as independent confirmation of an unchanged trace.
Proposed discrimination: Use matched randomized cue assignments and a direct state-by-cue comparison with adequate precision, while measuring attention, arousal, and retention intervals. Add a separately validated content readout if testing replay mediation; then perturb that content independently of cue presentation. Evaluate multiple details and retrieval routes. For an artificial recipient, compare selective rehearsal policies on the same acquired episodes and distinguish improved factual retention from matching the person's measured change. These are proposals, not procedures demonstrated here.