1 Title and one-sentence summary

Training Intensity and Targeted Lucidity Reactivation: Dream Awareness and Sleep Costs

Summary: A randomized study of 320 adults will test how presleep training dose and REM cue pairing affect objectively verified lucid dreaming and sleep costs, informing decisions about training with the lowest necessary burden.

Applicant: Future Mind Institute, a US nonprofit positioned as a neutral consciousness-and-life-sciences funder. Priority A; domain: dream consciousness and self-representation. Funding requested: $900,000 over 24 months.

2 Scientific question and background

The question is whether additional training justifies its sleep costs. Demonstrating that lucid dreams can be induced is not the objective. Aspy (2020) studied 355 participants and supported practice combinations including mnemonic induction, without isolating the effect of one bedtime suggestion. Carr et al. (2023) established the feasibility of combining presleep cognitive training with REM cues in morning laboratory naps. That work did not establish the overnight relationship between training dose and sleep disruption. Its 2020 online publication and 2023 volume publication are the same study; the two supplied entries therefore constitute one source.

Baird, Tononi, and LaBerge (2022) showed why frontal 40-Hz differences potentially attributable to eye-movement artifacts cannot independently verify lucidity. Available signaling and stimulation methods make a dose trial feasible in 2026, while transfer from morning naps to overnight sleep remains an empirical question.

3 Origin of the hypotheses

The Seth Material supplies hypotheses only. Claim waking-consciousness-7, early-sessions-6 §277, proposes that bedtime suggestion may increase experiences involving awareness during dreams. Claim waking-consciousness-8, the same book §264, proposes that excessive effort can interfere with falling asleep. We translate these propositions into measurable lucid-dream outcomes and sleep-onset latency. The material is not evidence, and objective projection is not tested.

Provenance follows the applicant-supplied dataset, described as CC BY 4.0, and research website. Version and original passages will be checked before registration. The project sits within the Institute’s public “Seth Material Research Program,” with scientific conclusions independent of that program’s source material.

4 Main hypotheses and falsifiable predictions

Paired cues are expected to increase verified success. Twenty-minute training may impose additional sleep costs without sufficient additional benefit. Separate null hypotheses specify zero marginal effects of dose and pairing on verified success and sleep-onset latency.

Planning alternatives are success probabilities of 15% versus 35% and a 10-minute latency increase. These are design assumptions, not established effects. Decision thresholds are a 10-percentage-point benefit and a 10-minute sleep-onset cost. If higher-dose costs reach the threshold while the upper confidence bound for benefit falls below its threshold, escalation will no longer be recommended.

Two doses test a contrast; they cannot establish an inverted-U curve. The challenged proposition is the protocol assumption that increasing dose within the specified range adds benefit without adding sleep cost. It is not an accepted general theory of sleep.

5 Research design

The main trial randomizes 240 participants to a 2×2 factorial design: 5 versus 20 minutes of training, crossed with paired versus unpaired sounds, with 60 participants per cell. An auxiliary cohort of 80 is randomized equally to a single-suggestion condition and a neutral active control. The suggestion condition delivers the target sentence once each evening, followed by neutral activity; both auxiliary conditions last five minutes and receive no REM cues. This comparison explores the contribution of one sentence but cannot establish equivalence with complete training. The active control estimates change under repeated measurement and diary participation, rather than an untouched natural history.

Participants will be adults aged 18–60 who are not expert practitioners. Central concealed allocation will stratify by site, baseline dream recall, and previous lucid dreaming. Each participant completes one baseline week and three training weeks, with PSG during baseline and training weeks one and three: 960 monitored nights overall.

Practice finishes before a fixed lights-out time in every group; the practice period itself is separately counted as time burden. The factorial groups receive standardized training content differing in duration. During wakefulness, sounds are either paired with awareness practice or temporally separated from it, with acoustic exposure matched. REM playback follows common rules for intensity and maximum cue count. Actual exposure and microarousals are recorded. No additional scheduled awakenings are introduced as an induction technique.

Participants cannot be blinded to duration, but superiority predictions are withheld and cue assignment is concealed where feasible. PSG scorers, eye-signal adjudicators, and analysts working with coded conditions remain blinded. Delivery technicians do not adjudicate endpoints. Every group receives identical instruction in the predetermined eye signal.

Allowing 15% attrition leaves approximately 204 factorial participants, or 102 per factor level. Conservative planning for four primary tests uses two-sided α=0.0125. A normal approximation gives approximately 80% power for 15% versus 35% success and 86% for a 10-minute latency difference with an adjusted standard deviation of 20 minutes. Interaction, auxiliary, and equivalence analyses are not assured adequate power. Simulation will validate these assumptions before preregistration. Outcomes, exclusions, missingness procedures, cue rules, multiplicity, and stopping criteria will then be frozen.

6 Data and analysis plan

The co-primary outcomes are the proportion of participants with at least one verified lucid dream across the two training monitoring nights, and mean PSG sleep-onset latency across those nights, adjusted for baseline. Verification requires a predetermined left–right eye-movement sequence during stable REM, a consistent post-awakening report of awareness, and independent agreement by two adjudicators. Absent dream recall and unsuccessful induction remain in the denominator; unavailable recordings are identified separately.

Intention-to-treat logistic models will produce standardized risk differences. Baseline-adjusted models will estimate latency differences in minutes. Holm correction will cover the four primary tests. Night-level mixed models and the dose-by-pairing interaction are secondary analyses. Additional outcomes include next-day psychomotor vigilance, dream recall, reported lucidity, total sleep time, sleep efficiency, microarousals, adherence, and total time burden.

Dream recall may mediate an intervention effect and will not be inserted post hoc into the primary model. Because REM playback follows sleep onset, its association with that night’s initial latency cannot be interpreted as an immediate cue effect.

All induction failures and between-group confidence intervals will be reported. Multiple imputation and sensitivity analyses for nonrandom missingness will address unavailable outcomes. Preregistration, code, a data dictionary, and deidentified data will be released; raw dream narratives and potentially identifiable physiological records will require controlled access. Benefit and cost will remain separate outcomes, avoiding conclusions selected through a retrospectively constructed composite score.

7 Milestones and duration

Months 1–4 cover ethics approval, independent adversarial protocol review, equipment calibration, and registration. Months 5–17 cover recruitment and 960 monitoring nights across two sites. Months 18–20 cover blinded quality control and database lock. Months 21–24 cover analysis, data release, and manuscripts. Pilot work will optimize procedures without selecting the protocol according to apparent efficacy.

8 Budget magnitude

The requested $900,000 comprises:

• Personnel: $260,000.
• PSG and facilities for 960 nights: $288,000.
• Participant compensation: $96,000.
• Equipment and maintenance: $55,000.
• Statistics and data management: $55,000.
• Independent adjudication and safety oversight: $25,000.
• Open science and dissemination: $16,000.
• Indirect costs and contingency: $105,000.

The estimate assumes existing sleep-laboratory infrastructure and an incremental PSG cost of $300 per night. Institutional quotations will accompany the application.

9 PI profile and candidate teams or institutions

The PI should combine sleep-physiology expertise, randomized-trial experience, and a record of open science. The team needs a sleep physician, statistician, technicians, and an independent skeptical collaborator.

Proposed invitations include Michelle Carr and the Université de Montréal Dream Engineering Lab for training and stimulation methods; Karen R. Konkoly for signal verification; Ken A. Paller and Northwestern collaborators for reactivation design; and Benjamin Baird for neurophysiology and artifact review. These are candidates, not confirmed collaborators. Carr’s institutional background is documented by Université de Montréal.

10 Risks and abandonment criteria

Principal threats include recall-related ascertainment bias, cue-induced microarousals, expectancy, and first-night effects. High-frequency EEG findings will remain exploratory because ocular and muscle contamination can mimic relevant activity.

The higher-dose protocol will be abandoned if the multiplicity-adjusted lower confidence bound for its latency cost is at least 10 minutes and the upper bound for its success advantage is below 10 percentage points. The paired-cue protocol will not be expanded if its benefit upper bound falls below 10 percentage points. Wide intervals mean uncertainty; nonsignificance alone does not establish ineffectiveness.

If more than 20% of monitoring nights remain uninterpretable after calibration, the verification pathway will be paused. Persistent sleep deterioration triggers independent safety review. Successful lucid-dream induction cannot establish objective out-of-body perception.

11 Ethics and compliance

Both sites will obtain ethics approval and informed consent before recruitment. Screening will address untreated sleep disorders and clinical risks that sleep disruption could aggravate. Participants can pause or withdraw; payment does not depend on induction success. Safe return transport will be arranged for participants experiencing significant sleepiness.

Consent will state that this is research rather than treatment and requires no acceptance of the source material’s worldview. Cross-border transfers will follow applicable privacy requirements. Access restrictions, conflicts of interest, and funder roles will be disclosed. Investigators retain independent publication rights, including for negative findings.

12 Fit with existing funders

Bial Foundation offers a relevant psychophysiology orientation, although its publicly described grant scale favors a methodological component rather than this entire budget. Fetzer Institute may be approached regarding experience and well-being, without making metaphysical conclusions a condition of support. These are proposed fit assessments, not expressions of funder interest.

Templeton World Charity Foundation’s ARC initiative and COGITATE provide models for preregistration, open science, and adversarial collaboration. BICS, UVA DOPS, and IONS are potential dialogue partners, not interchangeable or committed funding sources. Future Mind Institute proposes coordinating joint support, with proponents and skeptics jointly signing the decision rules before data collection. Alignment does not imply a currently open funding call or an invitation to apply.