{
  "study_id": "NCT04312269",
  "agent": "framework",
  "run": 1,
  "p_primary_hypothesis_supported": 0.15,
  "effect_direction": "favours intervention",
  "effect_size_90pct_interval": {"low": -0.35, "high": 0.8, "metric": "Cohen's d for between-group difference in WMFT change (baseline to 6 weeks), positive = more improvement with TMR"},
  "moderator_prediction": "All-phase TMR at least as large as SWS-only; the framework prices cues in deep sleep as near-weightless drip, so SWS confinement should not help and cues landing near sleep thresholds may add a little.",
  "claim_ids_used": ["PM-0095", "PM-0096", "PM-0129", "PM-0130", "PM-0132", "PM-0333", "PM-0364", "PM-0337", "PM-0851", "PM-0104", "PM-0061"],
  "rationale": "Framework: reactivation re-books what attentive daytime training stored (PM-0095, PM-0096); cues paired with learning carry the entry (PM-0333, PM-0364), so a small positive TMR effect is expected. But cues delivered in sleep resemble the drip, nearly weightless and working only by volume (PM-0129, PM-0130); the installer's quadrant is the thresholds, not deep sleep (PM-0132). Gains need daytime use to be promoted (PM-0337, PM-0851), and both arms get identical MyoCI training. General knowledge: ~70 across four arms (~17/arm) gives power only for d above ~1; WMFT timed scores are noisy in moderate-severe chronic stroke. Small positive d, likely non-significant."
}