Fig. 5

Sensitivity snapshot of CAP A‑phase burden contrasts. (a) and (b) summarise group–control differences in the A‑phase index (API) derived from automated CAP‑A detection in 60‑s windows restricted to NREM sleep (C3, C4; Methods). API quantifies A‑phase burden as time‑in‑A1/A2/A3 per NREM window and is not a CAP index; B‑phases and A–B cyclicity were not modelled. In (A), stage‑specific effects in N2 for subtypes A1 and A2 are depicted. Points denote the mean difference from controls (Δ vs. Control; API units) for each disorder; error bars indicate 95% confidence intervals. Marker shape encodes channel (circle = C3; triangle = C4). Colour encodes diagnostic group: iRBD (orange), NT1 (blue), NREM parasomnia (magenta), Fibromyalgia (teal). The horizontal grey line marks zero (no difference). Negative values indicate lower A‑phase burden than controls; positive values indicate higher burden. Across cohorts, the largest separations occur in N2 A2 (iRBD, NT1 < Control) and N2 A1 (iRBD, Fibromyalgia < Control); NREM parasomnia shows small or mixed effects. In (B) (illustrative, no inference), channel‑specific highlights illustrating our reporting convention for hemispheric analyses are shown. The panel shows contrasts selected a priori from the main analysis: NREM parasomnia in light sleep (N1) with higher A1/A2 relative to controls, most apparent in C3, and iRBD in deep sleep (N3) with higher A3 in C3. Each estimate is expressed as Δ vs. Control (API); shapes and colours as in (a). These examples demonstrate how channel‑wise reporting is handled when hemisphere‑specific effects are plausible. Scope and caveats. Estimates are descriptive summaries of stage‑wise group means versus controls; inferential statistics for all stage/subtype/channel contrasts are reported in Supplementary Tables. Because API reflects only A‑phase burden, results should not be interpreted as CAP index or as evidence about CAP cyclicity that includes A–B sequencing.