Altering wake-sleep patterns can bring consciousness and the unconscious into encounter, and this encounter is brief and rarely remembered.

For supplementary question 3: first, a leave-out validation stage–experience analysis is conducted on the DREAM database; then approximately 120 healthy adults are recruited in a randomized crossover design, completing adaptation night and four conditions: full sleep, REM-directed awakening, NREM-directed awakening, and non-stage-directed awakening with matched number and duration; washout between conditions is at least one week; the total cycle is about 6–8 weeks, and the project lasts 18–24 months. Polysomnography is used throughout; REM and NREM minutes and proportions, microarousals, and total sleep time are reported separately. Experience is recorded immediately after awakening, and again in the morning and 24 hours later. Autobiographical memory is evaluated using pre-experiment sealed diaries and autobiographical interview scoring, distinguishing correct details, incorrect details, and confidence. Randomization identifies the total intervention effect, but the mediating causal role of the REM/NREM ratio depends on additional assumptions and cannot be established by proportional correlation alone.

Sleep-Stage-Targeted Awakenings and Autobiographical Memory Retrieval: A Randomized Crossover Study

Limited support does not overturn current mainstream theory; if it is strictly demonstrated that stage changes specifically facilitate verifiable autobiographical detail retrieval, the model assumption that 'this effect is explained only by total sleep, number of awakenings, and retrieval context' would be revised.

Delphine Oudiette，Paris Brain Institute，DreamTeam; Isabelle Arnulf，Sorbonne Université，Paris Brain Institute; Ken A. Paller，Northwestern University; Francesca Siclari，Netherlands Institute for Neuroscience; Naotsugu Tsuchiya，Monash University

'Supported' applies only to the operational phenomenological level: wake-sleep transitions and some sleep states can yield cognitive experiences different from normal wakefulness, with incomplete recall; it does not cover an encounter between two independent mental systems, and the original text gives no predictions for sleep dose, stage, or memory gains.

'autobiographical memory' is usually translated as 'autobiographical memory,' not 'self-created memory'; [DreamTeam](https://institutducerveau.org/equipes-recherche-linstitut-cerveau/dreamteam-sommeil-reves-et-cognition) is studying wake-sleep mixed states, but artificial sleep fragmentation may also reduce memory performance.