# Causal EEG nested LODO results

**Scope:** joint decision-layer selection on frozen focal-synthetic w=0.5 evidence. This does not retrain upstream models.

**Objective:** `J = 0.5 * H + 0.5 * (Acc / 100)`; selection uses inner LODO only.

| Outer held | H | Acc (%) | L | SDT (s) | Miss rate | p_stay | scale | tau (s) | threshold |
|---|---:|---:|---:|---:|---:|---:|---:|---:|---:|
| AVGC | 0.8200 | 75.75 | 0.894 | 7.35 | 0.196 | 0.9800 | 0.800 | 0.00 | 0.65 |
| DTU | 0.9225 | 90.38 | 0.942 | 1.95 | 0.136 | 0.9970 | 0.500 | 0.00 | 0.65 |
| KUL | 0.8828 | 83.08 | 0.942 | 3.08 | 0.162 | 0.9800 | 0.800 | 0.00 | 0.65 |
| UHD | 0.8621 | 80.92 | 0.922 | 1.48 | 0.175 | 0.9800 | 0.800 | 0.00 | 0.65 |
| USTC | 0.9297 | 87.14 | 0.996 | 0.17 | 0.061 | 0.9800 | 0.650 | 0.00 | 0.65 |

## Mean outer-fold metrics

- H: **0.8834**
- Accuracy: **83.45%**
- L: 0.939
- SDT: 2.81 s
- Miss rate: 0.146

## External real-switch evaluation (AASD / Carta2026)

- Selection: deployment config selected by inner LODO over all five corpus datasets only; AASD and Carta2026 labels used exclusively for the final, post-selection score.
- Frozen external config: `{"emission_scale": 0.65, "p_stay": 0.985, "tau_s": 0.0, "threshold": 0.65}`

| External test | H | Acc (%) | L | SDT (s) | Switches | Undetected | Miss rate |
|---|---:|---:|---:|---:|---:|---:|---:|
| AASD | 0.6224 | 51.18 | 0.794 | 13.00 | 4600 | 1589 | 0.345 |
| Carta2026 | 0.5556 | 49.51 | 0.633 | 26.94 | 2880 | 727 | 0.252 |

## Leakage and scope notes

- Each outer-held dataset is excluded from hyperparameter selection.
- HMM emission statistics are estimated from outer-training datasets.
- Candidate parameters are selected by inner LODO only.
- Confirm that upstream evidence caches were generated without leakage.
- The current cache is already fused at w=0.5; focal/synthetic fusion weight learning requires separate upstream streams.
- These LODO scores use the project's synthetic-switch evaluation protocol; they are not a substitute for real-switch external evaluation.
