# How EEG decoding methods separate across seven protocols

Interactive explorer

There is no single composite score here. Inspect the level-removed 2D method profile map, free protocol-by-protocol error-bar scatter, 7×7 protocol agreement matrix, and pick-two paired comparator directly from published intervals.

2D Method Map Protocol Scatter Agreement Matrix Pick-Two Compare

View 1 · Level-removed 2D profile embedding

## 2D Method Profile Map

Each of the seven accuracy protocols is standardized across all complete methods, then each method’s own mean across the seven standardized scores is subtracted. That removes overall level completely so position reflects behavioral profile shape alone: nearby points share similar relative strengths across paradigms.

Step 1 of 5

Solid dot: distinct cross-protocol profile Hollow dot: flatter profile (position less informative) Outer arc: share of 3 nearest 7D neighbours kept in 2D

Hover or arrow-key to any node to inspect its 7-protocol profile. Click two nodes to compare them side by side.

Compact neural nets

### EEGNet

Scratch · 20 epochs

[Open method page →](https://bci.report/methods/eegnet/)

7-protocol accuracy & relative profile

MI / rest 71.0% +0.39σ

BETA 8ch 55.8% +0.61σ

BETA 4ch 44.1% +0.32σ

Arithmetic 67.6% −0.12σ

P300 ERP 53.3% −0.24σ

Semantic ERP 61.4% +1.35σ

Sleep 56.5% −2.31σ

3 nearest neighbours in true 7D profile space

- Complete methods: 27
- 2D variance retained: 69.5%
- Distance correlation r: 0.92
- 3-NN preserved: 53.1%

**Flatter profiles (drawn hollow):** The following methods have cross-protocol profile variation smaller than 1.5× their interval resampling SD, so their exact position on the map is less informative: ERP-FM Base, LaBraM at Scale Molecule-L, LaBraM at Scale Quark-L, REVE Large.

**Partial-coverage methods (excluded from the 7-protocol map, 9 total):** The map performs no imputation. Methods evaluated on a subset of protocols remain available in the Protocol Scatter and Pick-Two Comparator below: BrainOmni Base (5/7), CSP+LDA (5/7), FBCCA (2/7), MDM (2/7), Standard CCA (2/7), Temporal ridge (2/7), TS+LR (2/7), xDAWN+LDA (2/7), XDAWNCov+TS+LR (2/7).

Full profile coordinates & 7-protocol scores table (27 complete + 9 partial methods)

| Method | Family | Axis 1 | Axis 2 | 3-NN kept | Profile clarity | MI / rest | BETA 8ch | BETA 4ch | Arithmetic | P300 ERP | Semantic ERP | Sleep |
| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- |
| [BrainOmni Base](https://bci.report/methods/brainomni/) | Foundation models | — | — | — | 5/7 | 58.8% — 57.0–60.8% | 21.7% — 18.6–25.0% | 13.0% — 10.6–15.6% | 61.6% — 58.4–64.9% | — | — | 67.7% — 65.8–69.5% |
| [CBraMod](https://bci.report/methods/cbramod/) | Foundation models | −0.15 | +0.05 | 1/3 | Clear | 62.9% — 59.6–66.5% | 33.7% — 29.8–37.6% | 27.6% — 24.1–31.3% | 62.3% — 58.1–66.5% | 55.0% — 53.3–56.7% | 55.8% — 54.2–57.4% | 71.3% — 68.7–73.7% |
| [CodeBrain](https://bci.report/methods/codebrain/) | Foundation models | −0.96 | −0.78 | 1/3 | Clear | 58.8% — 54.5–62.6% | 11.9% — 10.5–13.3% | 11.2% — 9.8–12.7% | 63.0% — 59.5–66.5% | 50.8% — 49.3–52.2% | 54.7% — 53.0–56.6% | 70.7% — 68.8–72.7% |
| [eeg-fm-masking MAE (r=9 cm, L=2)](https://bci.report/methods/eeg-fm-masking/) | Foundation models | +1.35 | +1.70 | 2/3 | Clear | 68.1% — 64.3–72.3% | 54.8% — 49.4–60.0% | 47.1% — 42.1–52.1% | 63.6% — 59.1–68.1% | 49.5% — 48.1–51.0% | 54.8% — 52.7–56.9% | 67.9% — 65.4–70.4% |
| [EEGMamba](https://bci.report/methods/eegmamba/) | Foundation models | +1.02 | +0.88 | 2/3 | Clear | 67.3% — 63.6–70.6% | 40.2% — 35.7–44.7% | 32.7% — 28.9–36.7% | 60.3% — 57.0–63.4% | 52.5% — 51.4–53.5% | 54.4% — 52.5–56.3% | 64.1% — 61.6–66.4% |
| [EEGPT](https://bci.report/methods/eegpt/) | Foundation models | −1.54 | −0.45 | 1/3 | Clear | 58.3% — 54.9–61.8% | 20.3% — 18.2–22.6% | 17.3% — 15.5–19.2% | 65.7% — 60.9–70.3% | 54.0% — 51.5–56.6% | 54.6% — 53.3–55.9% | 75.5% — 73.9–77.1% |
| [ERP-FM Base](https://bci.report/methods/erp-fm/) | Foundation models | −1.02 | +0.37 | 1/3 | Less informative | 54.8% — 51.8–57.8% | 18.4% — 16.3–20.5% | 20.9% — 18.4–23.5% | 61.1% — 57.3–65.1% | 51.2% — 49.3–53.2% | 53.3% — 51.7–54.8% | 71.0% — 69.5–72.6% |
| [LaBraM](https://bci.report/methods/labram/) | Foundation models | −1.49 | −0.25 | 1/3 | Clear | 53.5% — 51.4–55.4% | 10.8% — 9.5–12.2% | 12.9% — 11.2–14.6% | 64.6% — 60.5–68.6% | 49.4% — 47.7–51.2% | 53.2% — 51.3–54.9% | 71.1% — 69.0–73.2% |
| [LaBraM at Scale Atom-L](https://bci.report/methods/labram/) | Foundation models | −1.39 | +0.62 | 2/3 | Clear | 56.9% — 53.3–61.0% | 25.8% — 22.4–29.4% | 28.0% — 24.3–31.8% | 67.5% — 63.7–71.3% | 53.2% — 51.5–54.9% | 52.8% — 51.3–54.3% | 72.7% — 70.6–74.4% |
| [LaBraM at Scale Molecule-L](https://bci.report/methods/labram/) | Foundation models | −0.52 | +0.32 | 1/3 | Less informative | 59.6% — 57.2–61.4% | 26.6% — 22.9–30.5% | 28.9% — 25.1–33.0% | 63.9% — 59.9–68.0% | 51.4% — 49.2–53.5% | 54.7% — 52.9–56.5% | 71.9% — 70.2–73.4% |
| [LaBraM at Scale Quark-B](https://bci.report/methods/labram/) | Foundation models | −0.31 | −0.02 | 1/3 | Clear | 61.9% — 59.1–64.4% | 27.4% — 23.6–31.5% | 27.8% — 24.0–31.9% | 66.0% — 62.0–70.0% | 49.6% — 48.5–50.6% | 55.4% — 53.1–57.8% | 72.1% — 70.3–74.0% |
| [LaBraM at Scale Quark-G](https://bci.report/methods/labram/) | Foundation models | −0.21 | +1.24 | 1/3 | Clear | 58.7% — 57.3–60.1% | 25.0% — 21.3–28.8% | 26.8% — 22.8–31.1% | 59.4% — 56.9–61.7% | 50.3% — 48.5–52.1% | 51.8% — 50.1–53.4% | 67.3% — 65.6–68.9% |
| [LaBraM at Scale Quark-H](https://bci.report/methods/labram/) | Foundation models | −1.06 | +0.81 | 2/3 | Clear | 60.2% — 57.4–62.9% | 23.7% — 20.2–27.3% | 26.2% — 22.3–30.3% | 64.5% — 60.6–68.3% | 52.6% — 51.1–54.1% | 52.2% — 50.3–53.9% | 72.9% — 71.5–74.2% |
| [LaBraM at Scale Quark-L](https://bci.report/methods/labram/) | Foundation models | −0.37 | +0.25 | 2/3 | Less informative | 59.6% — 56.3–62.9% | 27.1% — 23.3–31.0% | 28.0% — 24.0–32.2% | 63.8% — 59.8–67.8% | 51.0% — 49.4–52.6% | 54.8% — 53.0–56.6% | 70.3% — 68.0–72.5% |
| [LUNA Base](https://bci.report/methods/luna/) | Foundation models | −0.69 | −1.14 | 2/3 | Clear | 52.6% — 50.2–54.9% | 8.6% — 7.7–9.5% | 9.0% — 8.1–10.1% | 60.7% — 56.9–64.4% | 48.7% — 46.3–51.1% | 56.2% — 54.2–58.2% | 67.5% — 65.3–69.5% |
| [LUNA Large](https://bci.report/methods/luna/) | Foundation models | −0.86 | −0.97 | 3/3 | Clear | 49.9% — 46.6–53.5% | 10.0% — 8.9–11.1% | 11.2% — 9.8–12.6% | 57.4% — 54.3–60.6% | 48.8% — 46.9–50.8% | 57.1% — 55.3–58.9% | 72.4% — 70.4–74.0% |
| [REVE Base](https://bci.report/methods/reve/) | Foundation models | −0.54 | +0.43 | 1/3 | Clear | 66.2% — 61.8–70.6% | 37.5% — 33.3–41.9% | 25.5% — 22.4–28.8% | 68.8% — 64.1–73.2% | 50.5% — 48.5–52.3% | 53.7% — 51.8–55.4% | 74.6% — 72.9–76.4% |
| [REVE Large](https://bci.report/methods/reve/) | Foundation models | +0.55 | −0.05 | 1/3 | Less informative | 70.2% — 64.0–76.1% | 52.7% — 47.7–57.8% | 46.4% — 41.7–51.2% | 69.0% — 63.3–74.7% | 54.6% — 52.7–56.5% | 59.7% — 57.5–61.8% | 76.6% — 75.2–78.0% |
| [REVE Small](https://bci.report/methods/reve/) | Foundation models | +0.39 | +0.12 | 2/3 | Clear | 61.5% — 58.3–64.8% | 39.3% — 35.2–43.6% | 35.9% — 32.0–40.0% | 65.2% — 60.5–70.0% | 54.3% — 52.9–55.9% | 56.7% — 54.5–59.0% | 64.5% — 62.5–66.6% |
| [SingLEM](https://bci.report/methods/singlem/) | Foundation models | +4.15 | −1.14 | 2/3 | Clear | 72.0% — 66.6–76.6% | 26.6% — 23.7–29.5% | 20.5% — 17.8–23.3% | 51.5% — 49.2–53.8% | 51.2% — 49.2–53.1% | 58.5% — 56.1–61.1% | 35.3% — 33.7–37.0% |
| [ST-EEGFormer Base](https://bci.report/methods/st-eegformer/) | Foundation models | −1.58 | −0.44 | 2/3 | Clear | 58.1% — 53.7–63.3% | 17.1% — 15.5–18.8% | 17.5% — 15.8–19.3% | 64.5% — 59.0–70.0% | 53.6% — 52.3–54.8% | 54.6% — 52.9–56.4% | 77.3% — 75.6–78.8% |
| [ST-EEGFormer Large](https://bci.report/methods/st-eegformer/) | Foundation models | −0.81 | −0.91 | 2/3 | Clear | 62.4% — 57.4–67.8% | 22.6% — 20.1–25.3% | 22.4% — 19.9–25.0% | 63.4% — 57.8–69.0% | 52.7% — 51.3–54.2% | 57.8% — 56.0–59.6% | 79.6% — 78.2–81.0% |
| [ZUNA 1.1](https://bci.report/methods/zuna/) | Foundation models | −0.17 | −0.17 | 1/3 | Clear | 50.7% — 49.2–52.1% | 11.9% — 10.8–13.0% | 8.4% — 7.5–9.4% | 53.3% — 49.5–57.1% | 51.6% — 50.3–53.0% | 53.9% — 52.2–55.6% | 60.1% — 57.8–62.3% |
| [Deep4Net](https://bci.report/methods/deep4net/) | Compact neural nets | +3.20 | +0.35 | 2/3 | Clear | 68.6% — 63.3–74.0% | 52.8% — 47.3–58.2% | 38.7% — 33.7–43.8% | 61.3% — 57.5–65.2% | 50.8% — 50.0–51.6% | 57.4% — 55.4–59.5% | 45.2% — 42.7–47.7% |
| [EEGNet](https://bci.report/methods/eegnet/) | Compact neural nets | +2.50 | −0.78 | 1/3 | Clear | 71.0% — 65.0–76.3% | 55.8% — 50.1–61.3% | 44.1% — 38.6–49.6% | 67.6% — 62.8–72.5% | 53.3% — 51.7–54.9% | 61.4% — 58.9–64.0% | 56.5% — 54.9–58.3% |
| [ShallowFBCSPNet](https://bci.report/methods/shallowfbcspnet/) | Compact neural nets | +1.08 | −1.45 | 1/3 | Clear | 68.8% — 63.4–73.9% | 23.0% — 20.3–25.9% | 16.2% — 14.1–18.6% | 67.9% — 63.4–72.4% | 49.7% — 49.0–50.4% | 56.7% — 54.7–58.8% | 55.0% — 53.1–56.8% |
| [CSP+LDA](https://bci.report/methods/csp-lda/) | Classical methods | — | — | — | 5/7 | 47.1% — 43.8–50.2% | — | — | 68.3% — 62.0–74.3% | 50.0% — 49.1–50.8% | 52.7% — 51.4–54.0% | 50.1% — 46.9–53.1% |
| [FBCCA](https://bci.report/methods/fbcca/) | Classical methods | — | — | — | 2/7 | — | 79.1% — 75.1–82.8% | 72.3% — 67.5–76.9% | — | — | — | — |
| [Log band-power ridge](https://bci.report/methods/#log-band-power-ridge) | Classical methods | −0.86 | −1.73 | 2/3 | Clear | 57.3% — 54.4–60.3% | 5.3% — 4.8–5.8% | 5.3% — 4.8–5.8% | 65.0% — 60.2–69.7% | 50.4% — 49.3–51.6% | 56.1% — 54.2–58.0% | 66.0% — 63.0–68.9% |
| [MDM](https://bci.report/methods/riemannian/) | Classical methods | — | — | — | 2/7 | 53.8% — 50.8–57.9% | — | — | 71.5% — 65.5–77.8% | — | — | — |
| Spectral ridge | Classical methods | +0.12 | +3.34 | 3/3 | Clear | 54.9% — 53.2–56.7% | 50.7% — 45.6–55.8% | 48.2% — 43.1–53.3% | 56.8% — 54.6–58.9% | 51.0% — 49.6–52.4% | 51.0% — 49.6–52.5% | 72.2% — 70.7–73.8% |
| [Standard CCA](https://bci.report/methods/cca/) | Classical methods | — | — | — | 2/7 | — | 63.1% — 57.2–69.0% | 57.6% — 51.8–63.5% | — | — | — | — |
| Temporal ridge | Classical methods | — | — | — | 2/7 | — | — | — | — | 53.8% — 51.9–55.6% | 56.3% — 54.4–58.4% | — |
| [TS+LR](https://bci.report/methods/riemannian/) | Classical methods | — | — | — | 2/7 | 58.9% — 54.9–63.9% | — | — | 65.5% — 59.6–71.8% | — | — | — |
| [xDAWN+LDA](https://bci.report/methods/xdawn/) | Classical methods | — | — | — | 2/7 | — | — | — | — | 60.4% — 57.7–63.2% | 56.8% — 55.1–58.6% | — |
| [XDAWNCov+TS+LR](https://bci.report/methods/xdawn/) | Classical methods | — | — | — | 2/7 | — | — | — | — | 61.9% — 58.8–65.0% | 60.3% — 58.2–62.6% | — |

Closest & most distant pairs in true 7D profile space (10 each)

| Group | Method A | Method B | 7D distance | 2D distance |
| --- | --- | --- | --- | --- |
| Closest profiles | LaBraM at Scale Molecule-L | LaBraM at Scale Quark-L | 0.24 | 0.17 |
| Closest profiles | EEGPT | ST-EEGFormer Base | 0.42 | 0.04 |
| Closest profiles | ERP-FM Base | LaBraM at Scale Molecule-L | 0.68 | 0.50 |
| Closest profiles | ERP-FM Base | LaBraM at Scale Quark-L | 0.90 | 0.66 |
| Closest profiles | LaBraM at Scale Atom-L | LaBraM at Scale Quark-H | 0.90 | 0.38 |
| Closest profiles | CodeBrain | ST-EEGFormer Large | 0.92 | 0.20 |
| Closest profiles | REVE Large | REVE Small | 1.00 | 0.24 |
| Closest profiles | LaBraM at Scale Quark-B | LaBraM at Scale Quark-L | 1.03 | 0.28 |
| Closest profiles | LUNA Base | LUNA Large | 1.08 | 0.24 |
| Closest profiles | CodeBrain | Log band-power ridge | 1.08 | 0.96 |
| Most distinct profiles | SingLEM | Spectral ridge | 6.24 | 6.02 |
| Most distinct profiles | LaBraM at Scale Atom-L | SingLEM | 6.17 | 5.81 |
| Most distinct profiles | LaBraM | SingLEM | 6.15 | 5.71 |
| Most distinct profiles | EEGPT | SingLEM | 5.87 | 5.74 |
| Most distinct profiles | SingLEM | ST-EEGFormer Base | 5.87 | 5.78 |
| Most distinct profiles | LaBraM at Scale Quark-H | SingLEM | 5.75 | 5.56 |
| Most distinct profiles | REVE Base | SingLEM | 5.70 | 4.95 |
| Most distinct profiles | LUNA Large | SingLEM | 5.59 | 5.01 |
| Most distinct profiles | ShallowFBCSPNet | Spectral ridge | 5.53 | 4.88 |
| Most distinct profiles | ERP-FM Base | SingLEM | 5.52 | 5.39 |

View 2 · Free protocol × protocol comparison

## Protocol × Protocol 2D Scatter

Choose any two accuracy protocols for the horizontal and vertical axes. Every method scored on both protocols appears with 2D crosshair 95% intervals and slate dashed chance reference lines.

Spearman ρ = −0.15 · shared methods n = 29

Hover or click any point for exact intervals. Clicking any cell in the Agreement Matrix below also loads that protocol pair here.

Compact neural nets

### EEGNet

- Motor imagery & rest: 71.0%
- Sleep staging: 56.5%

Current scatter table: MI / rest × Sleep (29 methods)

| Method | Family | Motor imagery & rest | Sleep staging |
| --- | --- | --- | --- |
| BrainOmni Base | Foundation models | 58.8% (57.0–60.8%) | 67.7% (65.8–69.5%) |
| CBraMod | Foundation models | 62.9% (59.6–66.5%) | 71.3% (68.7–73.7%) |
| CodeBrain | Foundation models | 58.8% (54.5–62.6%) | 70.7% (68.8–72.7%) |
| eeg-fm-masking MAE (r=9 cm, L=2) | Foundation models | 68.1% (64.3–72.3%) | 67.9% (65.4–70.4%) |
| EEGMamba | Foundation models | 67.3% (63.6–70.6%) | 64.1% (61.6–66.4%) |
| EEGPT | Foundation models | 58.3% (54.9–61.8%) | 75.5% (73.9–77.1%) |
| ERP-FM Base | Foundation models | 54.8% (51.8–57.8%) | 71.0% (69.5–72.6%) |
| LaBraM | Foundation models | 53.5% (51.4–55.4%) | 71.1% (69.0–73.2%) |
| LaBraM at Scale Atom-L | Foundation models | 56.9% (53.3–61.0%) | 72.7% (70.6–74.4%) |
| LaBraM at Scale Molecule-L | Foundation models | 59.6% (57.2–61.4%) | 71.9% (70.2–73.4%) |
| LaBraM at Scale Quark-B | Foundation models | 61.9% (59.1–64.4%) | 72.1% (70.3–74.0%) |
| LaBraM at Scale Quark-G | Foundation models | 58.7% (57.3–60.1%) | 67.3% (65.6–68.9%) |
| LaBraM at Scale Quark-H | Foundation models | 60.2% (57.4–62.9%) | 72.9% (71.5–74.2%) |
| LaBraM at Scale Quark-L | Foundation models | 59.6% (56.3–62.9%) | 70.3% (68.0–72.5%) |
| LUNA Base | Foundation models | 52.6% (50.2–54.9%) | 67.5% (65.3–69.5%) |
| LUNA Large | Foundation models | 49.9% (46.6–53.5%) | 72.4% (70.4–74.0%) |
| REVE Base | Foundation models | 66.2% (61.8–70.6%) | 74.6% (72.9–76.4%) |
| REVE Large | Foundation models | 70.2% (64.0–76.1%) | 76.6% (75.2–78.0%) |
| REVE Small | Foundation models | 61.5% (58.3–64.8%) | 64.5% (62.5–66.6%) |
| SingLEM | Foundation models | 72.0% (66.6–76.6%) | 35.3% (33.7–37.0%) |
| ST-EEGFormer Base | Foundation models | 58.1% (53.7–63.3%) | 77.3% (75.6–78.8%) |
| ST-EEGFormer Large | Foundation models | 62.4% (57.4–67.8%) | 79.6% (78.2–81.0%) |
| ZUNA 1.1 | Foundation models | 50.7% (49.2–52.1%) | 60.1% (57.8–62.3%) |
| Deep4Net | Compact neural nets | 68.6% (63.3–74.0%) | 45.2% (42.7–47.7%) |
| EEGNet | Compact neural nets | 71.0% (65.0–76.3%) | 56.5% (54.9–58.3%) |
| ShallowFBCSPNet | Compact neural nets | 68.8% (63.4–73.9%) | 55.0% (53.1–56.8%) |
| CSP+LDA | Classical methods | 47.1% (43.8–50.2%) | 50.1% (46.9–53.1%) |
| Log band-power ridge | Classical methods | 57.3% (54.4–60.3%) | 66.0% (63.0–68.9%) |
| Spectral ridge | Classical methods | 54.9% (53.2–56.7%) | 72.2% (70.7–73.8%) |

View 3 · Why one overall score distorts the evidence

## Do protocols reward the same methods?

Each cell reports the Spearman correlation ρ and disagreement (1 − ρ) across all methods shared by two accuracy protocols. Off-diagonal correlation has median +0.21, ranging from −0.19 (Semantic ERP vs. Sleep) to +0.95 (BETA 8ch vs. BETA 4ch). Select any off-diagonal cell to inspect that pair in the scatter above.

Each cell shows Spearman correlation ρ, disagreement (1 − ρ), and shared method count n.

| Protocol | MI / rest | BETA 8ch | BETA 4ch | Arithmetic | P300 ERP | Semantic ERP | Sleep |
| --- | --- | --- | --- | --- | --- | --- | --- |
| MI / rest | **1.00** — n=31 | **+0.73** — 1−ρ 0.27 · n=28 | **+0.55** — 1−ρ 0.45 · n=28 | **+0.14** — 1−ρ 0.86 · n=31 | **+0.30** — 1−ρ 0.70 · n=28 | **+0.56** — 1−ρ 0.44 · n=28 | **−0.15** — 1−ρ 1.15 · n=29 |
| BETA 8ch | **+0.73** — 1−ρ 0.27 · n=28 | **1.00** — n=30 | **+0.95** — 1−ρ 0.05 · n=30 | **+0.21** — 1−ρ 0.79 · n=28 | **+0.31** — 1−ρ 0.69 · n=27 | **+0.17** — 1−ρ 0.83 · n=27 | **−0.11** — 1−ρ 1.11 · n=28 |
| BETA 4ch | **+0.55** — 1−ρ 0.45 · n=28 | **+0.95** — 1−ρ 0.05 · n=30 | **1.00** — n=30 | **+0.18** — 1−ρ 0.82 · n=28 | **+0.33** — 1−ρ 0.67 · n=27 | **+0.04** — 1−ρ 0.96 · n=27 | **+0.04** — 1−ρ 0.96 · n=28 |
| Arithmetic | **+0.14** — 1−ρ 0.86 · n=31 | **+0.21** — 1−ρ 0.79 · n=28 | **+0.18** — 1−ρ 0.82 · n=28 | **1.00** — n=31 | **+0.16** — 1−ρ 0.84 · n=28 | **+0.07** — 1−ρ 0.93 · n=28 | **+0.24** — 1−ρ 0.76 · n=29 |
| P300 ERP | **+0.30** — 1−ρ 0.70 · n=28 | **+0.31** — 1−ρ 0.69 · n=27 | **+0.33** — 1−ρ 0.67 · n=27 | **+0.16** — 1−ρ 0.84 · n=28 | **1.00** — n=31 | **+0.26** — 1−ρ 0.74 · n=31 | **+0.29** — 1−ρ 0.71 · n=28 |
| Semantic ERP | **+0.56** — 1−ρ 0.44 · n=28 | **+0.17** — 1−ρ 0.83 · n=27 | **+0.04** — 1−ρ 0.96 · n=27 | **+0.07** — 1−ρ 0.93 · n=28 | **+0.26** — 1−ρ 0.74 · n=31 | **1.00** — n=31 | **−0.19** — 1−ρ 1.19 · n=28 |
| Sleep | **−0.15** — 1−ρ 1.15 · n=29 | **−0.11** — 1−ρ 1.11 · n=28 | **+0.04** — 1−ρ 0.96 · n=28 | **+0.24** — 1−ρ 0.76 · n=29 | **+0.29** — 1−ρ 0.71 · n=28 | **−0.19** — 1−ρ 1.19 · n=28 | **1.00** — n=29 |

Pick-two protocol comparison

### Compare any two methods on one protocol

Inspect point estimates and 95% intervals side by side on a shared axis. Paired differences appear only where a paired interval is published. Dropdowns are grouped by family alphabetically — never ordered by score.

[Download protocol CSV](https://bci.report/data/mi-rest-results.csv) [Open protocol page](https://bci.report/protocols/mi-rest/)

Published paired contrast **ShallowFBCSPNet − EEGNet: −2.2 pp** [−8.0, +3.4 pp]

95% interval, unadjusted · n = 10 participants · Paired interval includes zero

against the matrix's one-seed EEGNet cell; training variance not included

Default view (Motor imagery & rest): ShallowFBCSPNet balanced accuracy 68.8% (95% interval 63.4–73.9%); EEGNet 71.0% (95% interval 65.0–76.3%); chance 50.0%.

Published paired contrasts table (91 pairs)

| Protocol | Pair (left minus right) | Mean difference | Governing interval | Participants | Adjustment & notes |
| --- | --- | --- | --- | --- | --- |
| SSVEP · 8 channels | **FBCCA − Standard CCA** | +16.0 pp | [+11.6, +20.5 pp] | 70 | 98.75% interval, Bonferroni-adjusted for four contrasts |
| SSVEP · 8 channels | **FBCCA − EEGNet** | +23.3 pp | [+17.6, +29.6 pp] | 70 | 98.75% interval, Bonferroni-adjusted for four contrasts — against the matrix's one-seed EEGNet cell; training variance not included |
| SSVEP · 4 channels | **FBCCA − Standard CCA** | +14.7 pp | [+10.7, +19.1 pp] | 70 | 98.75% interval, Bonferroni-adjusted for four contrasts |
| SSVEP · 4 channels | **FBCCA − EEGNet** | +28.2 pp | [+22.8, +33.9 pp] | 70 | 98.75% interval, Bonferroni-adjusted for four contrasts — against the matrix's one-seed EEGNet cell; training variance not included |
| Motor imagery & rest | **TS+LR − Spectral ridge** | +4.0 pp | [−1.1, +11.5 pp] | 10 | 98.75% interval, Bonferroni-adjusted for four contrasts |
| Motor imagery & rest | **TS+LR − EEGNet** | −12.1 pp | [−18.1, −5.4 pp] | 10 | nominal 95% interval — against the matrix's one-seed EEGNet cell; training variance not included |
| Motor imagery & rest | **MDM − EEGNet** | −17.2 pp | [−23.7, −10.6 pp] | 10 | nominal 95% interval — against the matrix's one-seed EEGNet cell; training variance not included |
| Motor imagery & rest | **MDM − Spectral ridge** | −1.1 pp | [−4.7, +3.8 pp] | 10 | nominal 95% interval |
| Motor imagery & rest | **TS+LR − CSP+LDA** | +11.8 pp | [+7.9, +16.8 pp] | 10 | nominal 95% interval |
| Motor imagery & rest | **MDM − CSP+LDA** | +6.8 pp | [+3.3, +10.8 pp] | 10 | nominal 95% interval |
| Motor imagery & rest | **TS+LR − LaBraM** | +5.4 pp | [+0.3, +11.3 pp] | 10 | nominal 95% interval |
| Motor imagery & rest | **TS+LR − CBraMod** | −4.0 pp | [−8.4, +0.9 pp] | 10 | nominal 95% interval |
| Motor imagery & rest | **TS+LR − MDM** | +5.1 pp | [+2.3, +7.8 pp] | 10 | nominal 95% interval |
| Arithmetic & rest | **TS+LR − Spectral ridge** | +8.8 pp | [+1.6, +16.1 pp] | 36 | 98.75% interval, Bonferroni-adjusted for four contrasts |
| Arithmetic & rest | **TS+LR − EEGNet** | −2.1 pp | [−8.0, +3.8 pp] | 36 | nominal 95% interval — against the matrix's one-seed EEGNet cell; training variance not included |
| Arithmetic & rest | **MDM − EEGNet** | +3.8 pp | [−1.3, +9.3 pp] | 36 | nominal 95% interval — against the matrix's one-seed EEGNet cell; training variance not included |
| Arithmetic & rest | **MDM − Spectral ridge** | +14.7 pp | [+8.9, +20.5 pp] | 36 | nominal 95% interval |
| Arithmetic & rest | **TS+LR − LaBraM** | +0.9 pp | [−4.6, +6.6 pp] | 36 | nominal 95% interval |
| Arithmetic & rest | **TS+LR − CBraMod** | +3.2 pp | [−1.5, +8.0 pp] | 36 | nominal 95% interval |
| Arithmetic & rest | **TS+LR − MDM** | −6.0 pp | [−11.6, −0.5 pp] | 36 | nominal 95% interval |
| P300 target ERP | **XDAWNCov+TS+LR − Spectral ridge** | +11.0 pp | [+6.8, +15.1 pp] | 21 | 98.75% interval, Bonferroni-adjusted for four contrasts |
| P300 target ERP | **XDAWNCov+TS+LR − EEGNet** | +8.6 pp | [+5.2, +12.0 pp] | 21 | nominal 95% interval — against the matrix's one-seed EEGNet cell; training variance not included |
| P300 target ERP | **xDAWN+LDA − EEGNet** | +7.1 pp | [+4.3, +10.0 pp] | 21 | nominal 95% interval — against the matrix's one-seed EEGNet cell; training variance not included |
| P300 target ERP | **xDAWN+LDA − Spectral ridge** | +9.5 pp | [+6.5, +12.4 pp] | 21 | nominal 95% interval |
| P300 target ERP | **XDAWNCov+TS+LR − Temporal ridge** | +8.1 pp | [+5.3, +11.1 pp] | 21 | nominal 95% interval |
| P300 target ERP | **xDAWN+LDA − Temporal ridge** | +6.6 pp | [+3.7, +9.8 pp] | 21 | nominal 95% interval |
| P300 target ERP | **XDAWNCov+TS+LR − LaBraM** | +12.5 pp | [+8.7, +16.3 pp] | 21 | nominal 95% interval |
| P300 target ERP | **XDAWNCov+TS+LR − CBraMod** | +6.9 pp | [+4.4, +9.6 pp] | 21 | nominal 95% interval |
| P300 target ERP | **XDAWNCov+TS+LR − xDAWN+LDA** | +1.5 pp | [−0.9, +3.7 pp] | 21 | nominal 95% interval |
| Semantic target ERP | **XDAWNCov+TS+LR − Spectral ridge** | +9.3 pp | [+6.1, +12.8 pp] | 30 | 98.75% interval, Bonferroni-adjusted for four contrasts |
| Semantic target ERP | **XDAWNCov+TS+LR − EEGNet** | −1.2 pp | [−3.0, +0.6 pp] | 30 | nominal 95% interval — against the matrix's one-seed EEGNet cell; training variance not included |
| Semantic target ERP | **xDAWN+LDA − EEGNet** | −4.6 pp | [−6.5, −2.7 pp] | 30 | nominal 95% interval — against the matrix's one-seed EEGNet cell; training variance not included |
| Semantic target ERP | **xDAWN+LDA − Spectral ridge** | +5.8 pp | [+4.0, +7.8 pp] | 30 | nominal 95% interval |
| Semantic target ERP | **XDAWNCov+TS+LR − Temporal ridge** | +3.9 pp | [+2.0, +5.9 pp] | 30 | nominal 95% interval |
| Semantic target ERP | **xDAWN+LDA − Temporal ridge** | +0.5 pp | [−1.2, +2.1 pp] | 30 | nominal 95% interval |
| Semantic target ERP | **XDAWNCov+TS+LR − LaBraM** | +7.1 pp | [+4.8, +9.4 pp] | 30 | nominal 95% interval |
| Semantic target ERP | **XDAWNCov+TS+LR − CBraMod** | +4.5 pp | [+2.5, +6.4 pp] | 30 | nominal 95% interval |
| Semantic target ERP | **XDAWNCov+TS+LR − xDAWN+LDA** | +3.5 pp | [+1.8, +5.3 pp] | 30 | nominal 95% interval |
| Arithmetic & rest | **CSP+LDA − EEGNet** | +0.7 pp | [−4.3, +5.8 pp] | 36 | 95% interval, unadjusted — against the matrix's one-seed EEGNet cell; training variance not included |
| Arithmetic & rest | **Deep4Net − EEGNet** | −6.3 pp | [−10.4, −2.3 pp] | 36 | 95% interval, unadjusted — against the matrix's one-seed EEGNet cell; training variance not included |
| Arithmetic & rest | **ShallowFBCSPNet − EEGNet** | +0.3 pp | [−2.6, +3.3 pp] | 36 | 95% interval, unadjusted — against the matrix's one-seed EEGNet cell; training variance not included |
| Arithmetic & rest | **ShallowFBCSPNet − Deep4Net** | +6.6 pp | [+2.6, +10.6 pp] | 36 | 95% interval |
| SSVEP · 4 channels | **Deep4Net − EEGNet** | −5.5 pp | [−6.8, −4.1 pp] | 70 | 95% interval, unadjusted — against the matrix's one-seed EEGNet cell; training variance not included |
| SSVEP · 4 channels | **ShallowFBCSPNet − EEGNet** | −27.9 pp | [−31.9, −23.9 pp] | 70 | 95% interval, unadjusted — against the matrix's one-seed EEGNet cell; training variance not included |
| SSVEP · 4 channels | **ShallowFBCSPNet − Deep4Net** | −22.4 pp | [−26.1, −18.9 pp] | 70 | 95% interval |
| SSVEP · 8 channels | **Deep4Net − EEGNet** | −3.0 pp | [−3.9, −2.1 pp] | 70 | 95% interval, unadjusted — against the matrix's one-seed EEGNet cell; training variance not included |
| SSVEP · 8 channels | **ShallowFBCSPNet − EEGNet** | −32.7 pp | [−37.0, −28.5 pp] | 70 | 95% interval, unadjusted — against the matrix's one-seed EEGNet cell; training variance not included |
| SSVEP · 8 channels | **ShallowFBCSPNet − Deep4Net** | −29.7 pp | [−33.7, −25.8 pp] | 70 | 95% interval |
| Motor imagery & rest | **Deep4Net − EEGNet** | −2.4 pp | [−6.8, +2.3 pp] | 10 | 95% interval, unadjusted — against the matrix's one-seed EEGNet cell; training variance not included |
| Motor imagery & rest | **ShallowFBCSPNet − EEGNet** | −2.2 pp | [−8.0, +3.4 pp] | 10 | 95% interval, unadjusted — against the matrix's one-seed EEGNet cell; training variance not included |
| Motor imagery & rest | **ShallowFBCSPNet − Deep4Net** | +0.3 pp | [−6.8, +7.7 pp] | 10 | 95% interval |
| P300 target ERP | **CSP+LDA − EEGNet** | −3.3 pp | [−5.4, −1.6 pp] | 21 | 95% interval, unadjusted — against the matrix's one-seed EEGNet cell; training variance not included |
| P300 target ERP | **Deep4Net − EEGNet** | −2.5 pp | [−4.2, −0.9 pp] | 21 | 95% interval, unadjusted — against the matrix's one-seed EEGNet cell; training variance not included |
| P300 target ERP | **ShallowFBCSPNet − EEGNet** | −3.6 pp | [−5.4, −1.9 pp] | 21 | 95% interval, unadjusted — against the matrix's one-seed EEGNet cell; training variance not included |
| P300 target ERP | **ShallowFBCSPNet − Deep4Net** | −1.1 pp | [−2.1, −0.2 pp] | 21 | 95% interval |
| Semantic target ERP | **CSP+LDA − EEGNet** | −8.8 pp | [−11.1, −6.4 pp] | 30 | 95% interval, unadjusted — against the matrix's one-seed EEGNet cell; training variance not included |
| Semantic target ERP | **Deep4Net − EEGNet** | −4.0 pp | [−5.8, −2.1 pp] | 30 | 95% interval, unadjusted — against the matrix's one-seed EEGNet cell; training variance not included |
| Semantic target ERP | **ShallowFBCSPNet − EEGNet** | −4.7 pp | [−7.2, −2.2 pp] | 30 | 95% interval, unadjusted — against the matrix's one-seed EEGNet cell; training variance not included |
| Semantic target ERP | **ShallowFBCSPNet − Deep4Net** | −0.7 pp | [−3.0, +1.5 pp] | 30 | 95% interval |
| Sleep staging | **CSP+LDA − EEGNet** | −6.4 pp | [−9.3, −3.3 pp] | 20 | 95% interval, unadjusted — against the matrix's one-seed EEGNet cell; training variance not included |
| Sleep staging | **Deep4Net − EEGNet** | −11.3 pp | [−13.6, −9.0 pp] | 20 | 95% interval, unadjusted — against the matrix's one-seed EEGNet cell; training variance not included |
| Sleep staging | **ShallowFBCSPNet − EEGNet** | −1.6 pp | [−3.4, +0.4 pp] | 20 | 95% interval, unadjusted — against the matrix's one-seed EEGNet cell; training variance not included |
| Sleep staging | **ShallowFBCSPNet − Deep4Net** | +9.7 pp | [+7.0, +12.7 pp] | 20 | 95% interval |
| Motor imagery & rest | **Log band-power ridge − Spectral ridge** | +2.4 pp | [−1.2, +5.8 pp] | 10 | 95% paired participant bootstrap interval, unadjusted |
| SSVEP · 8 channels | **Log band-power ridge − Spectral ridge** | −45.4 pp | [−50.4, −40.6 pp] | 70 | 95% paired participant bootstrap interval, unadjusted |
| SSVEP · 4 channels | **Log band-power ridge − Spectral ridge** | −42.9 pp | [−47.7, −38.1 pp] | 70 | 95% paired participant bootstrap interval, unadjusted |
| Arithmetic & rest | **Log band-power ridge − Spectral ridge** | +8.2 pp | [+4.0, +12.6 pp] | 36 | 95% paired participant bootstrap interval, unadjusted |
| P300 target ERP | **Log band-power ridge − Spectral ridge** | −0.5 pp | [−2.4, +1.4 pp] | 21 | 95% paired participant bootstrap interval, unadjusted |
| Semantic target ERP | **Log band-power ridge − Spectral ridge** | +5.1 pp | [+3.1, +7.1 pp] | 30 | 95% paired participant bootstrap interval, unadjusted |
| Sleep staging | **Log band-power ridge − Spectral ridge** | −6.2 pp | [−8.5, −3.9 pp] | 20 | 95% paired participant bootstrap interval, unadjusted |
| Motor imagery & rest | **LaBraM at Scale Quark-L − LaBraM at Scale Quark-B** | −2.3 pp | [−4.8, −0.3 pp] | 10 | 95% paired participant bootstrap interval, unadjusted (21 ladder contrasts) |
| SSVEP · 8 channels | **LaBraM at Scale Quark-L − LaBraM at Scale Quark-B** | −0.4 pp | [−1.3, +0.6 pp] | 70 | 95% paired participant bootstrap interval, unadjusted (21 ladder contrasts) |
| SSVEP · 4 channels | **LaBraM at Scale Quark-L − LaBraM at Scale Quark-B** | +0.1 pp | [−0.9, +1.2 pp] | 70 | 95% paired participant bootstrap interval, unadjusted (21 ladder contrasts) |
| Arithmetic & rest | **LaBraM at Scale Quark-L − LaBraM at Scale Quark-B** | −2.1 pp | [−5.0, +0.8 pp] | 36 | 95% paired participant bootstrap interval, unadjusted (21 ladder contrasts) |
| P300 target ERP | **LaBraM at Scale Quark-L − LaBraM at Scale Quark-B** | +1.4 pp | [−0.3, +3.3 pp] | 21 | 95% paired participant bootstrap interval, unadjusted (21 ladder contrasts) |
| Semantic target ERP | **LaBraM at Scale Quark-L − LaBraM at Scale Quark-B** | −0.6 pp | [−2.8, +1.7 pp] | 30 | 95% paired participant bootstrap interval, unadjusted (21 ladder contrasts) |
| Sleep staging | **LaBraM at Scale Quark-L − LaBraM at Scale Quark-B** | −1.8 pp | [−3.3, −0.3 pp] | 20 | 95% paired participant bootstrap interval, unadjusted (21 ladder contrasts) |
| Motor imagery & rest | **LaBraM at Scale Quark-H − LaBraM at Scale Quark-L** | +0.6 pp | [−3.1, +3.5 pp] | 10 | 95% paired participant bootstrap interval, unadjusted (21 ladder contrasts) |
| SSVEP · 8 channels | **LaBraM at Scale Quark-H − LaBraM at Scale Quark-L** | −3.4 pp | [−4.4, −2.4 pp] | 70 | 95% paired participant bootstrap interval, unadjusted (21 ladder contrasts) |
| SSVEP · 4 channels | **LaBraM at Scale Quark-H − LaBraM at Scale Quark-L** | −1.7 pp | [−2.7, −0.8 pp] | 70 | 95% paired participant bootstrap interval, unadjusted (21 ladder contrasts) |
| Arithmetic & rest | **LaBraM at Scale Quark-H − LaBraM at Scale Quark-L** | +0.6 pp | [−1.7, +2.8 pp] | 36 | 95% paired participant bootstrap interval, unadjusted (21 ladder contrasts) |
| P300 target ERP | **LaBraM at Scale Quark-H − LaBraM at Scale Quark-L** | +1.6 pp | [−0.4, +3.5 pp] | 21 | 95% paired participant bootstrap interval, unadjusted (21 ladder contrasts) |
| Semantic target ERP | **LaBraM at Scale Quark-H − LaBraM at Scale Quark-L** | −2.6 pp | [−5.1, −0.1 pp] | 30 | 95% paired participant bootstrap interval, unadjusted (21 ladder contrasts) |
| Sleep staging | **LaBraM at Scale Quark-H − LaBraM at Scale Quark-L** | +2.5 pp | [+1.0, +4.2 pp] | 20 | 95% paired participant bootstrap interval, unadjusted (21 ladder contrasts) |
| Motor imagery & rest | **LaBraM at Scale Quark-H − LaBraM at Scale Quark-B** | −1.8 pp | [−5.2, +1.0 pp] | 10 | 95% paired participant bootstrap interval, unadjusted (21 ladder contrasts) |
| SSVEP · 8 channels | **LaBraM at Scale Quark-H − LaBraM at Scale Quark-B** | −3.8 pp | [−4.8, −2.8 pp] | 70 | 95% paired participant bootstrap interval, unadjusted (21 ladder contrasts) |
| SSVEP · 4 channels | **LaBraM at Scale Quark-H − LaBraM at Scale Quark-B** | −1.6 pp | [−2.5, −0.7 pp] | 70 | 95% paired participant bootstrap interval, unadjusted (21 ladder contrasts) |
| Arithmetic & rest | **LaBraM at Scale Quark-H − LaBraM at Scale Quark-B** | −1.5 pp | [−4.4, +1.3 pp] | 36 | 95% paired participant bootstrap interval, unadjusted (21 ladder contrasts) |
| P300 target ERP | **LaBraM at Scale Quark-H − LaBraM at Scale Quark-B** | +3.0 pp | [+1.3, +4.8 pp] | 21 | 95% paired participant bootstrap interval, unadjusted (21 ladder contrasts) |
| Semantic target ERP | **LaBraM at Scale Quark-H − LaBraM at Scale Quark-B** | −3.3 pp | [−5.5, −0.9 pp] | 30 | 95% paired participant bootstrap interval, unadjusted (21 ladder contrasts) |
| Sleep staging | **LaBraM at Scale Quark-H − LaBraM at Scale Quark-B** | +0.7 pp | [−0.9, +2.5 pp] | 20 | 95% paired participant bootstrap interval, unadjusted (21 ladder contrasts) |

---
Markdown copy of https://bci.report/explore/, generated from the published page. Figures are aggregate results; terms of use: https://bci.report/data-use/
