OSCR

A confidence-gated source selection strategy for cross-session transfer in brain-computer interfaces.

Code ↔ Paper

9 matches between paragraphs of the paper and lines of its authors' code, computed by the harvester (lexical-v1). Click a colored paragraph or line to see its counterpart.

The 9 matches
  1. [1] § Materials and methods › Classification ↔ src/crossda/pipelines/pipeline_utils.py, lines 237 to 287 · score 0.68 · linear discriminant, shrinkage, gamma, RBF, kernel, LDA · Report a mismatch
  2. [2] § Materials and methods › Source-selection and configuration-selection pipelines › MMP: minimum-distance multi-source pipeline ↔ src/crossda/pipelines/mmp_pipeline.py, lines 87 to 133 · score 0.65 · nearest proxy task, proxy target, proxy source, MMP, training, distance · Report a mismatch
  3. [3] § Materials and methods › Source-selection and configuration-selection pipelines › BDP: bridge-domain pipeline ↔ src/crossda/config.py, lines 40 to 151 · score 0.63 · cross validation, Far Bridge, fitted, proxy, score, domain · Report a mismatch
  4. [4] § Materials and methods › Feature extraction ↔ src/crossda/pipelines/pipeline_utils.py, lines 172 to 212 · score 0.61 · Tangent space, covariance matrices · Report a mismatch
  5. [5] § Materials and methods › Evaluation protocol ↔ src/crossda/config.py, lines 40 to 151 · score 0.57 · cross validation, Bridge Far, unweighted, folds, DWP, MMP · Report a mismatch
  6. [6] § Materials and methods › Datasets ↔ src/crossda/core/workers.py, lines 111 to 164 · score 0.55 · EEG channels, epochs, binary, classes, filtered, 35 Hz · Report a mismatch
  7. [7] § Materials and methods › Classification ↔ src/crossda/core/method_bank.py, lines 18 to 149 · score 0.54 · RBF SVM, linear, LDA, radial, sensitivity, baseline · Report a mismatch
  8. [8] § Materials and methods › Source-selection and configuration-selection pipelines › MMP: minimum-distance multi-source pipeline ↔ src/crossda/pipelines/mmp_pipeline.py, lines 1 to 35 · score 0.53 · weighted vote, MMP, distance, proxy, pipeline · Report a mismatch
  9. [9] § Materials and methods › Source-selection and configuration-selection pipelines › DWP: distance-weighted pooling ↔ src/crossda/pipelines/dwp_pipeline.py, lines 1 to 34 · score 0.52 · inverse distance weights, DWP, training, MAP, configuration · Report a mismatch

Paper

Loaded from Europe PMC by your browser; OSCR does not store it: doi.org · Europe PMC

The paper is loaded when this pane is shown.

The authors' code

Python · 287 lines · 11 KB · MIT · 2 matches

  1. """
  2. pipeline_utils.py — Shared utilities for MAP, MMP, and BDP pipelines.
  3. """
  5. from __future__ import annotations
  7. from typing import Any, Dict, List, Optional, Tuple
  9. import numpy as np
  10. from numpy.typing import NDArray
  12. from da4bci.metrics.distance import (
  13. compute_mmd,
  14. compute_wasserstein,
  15. compute_energy,
  16. compute_mahalanobis,
  17. )
  18. from da4bci.geometry.spd import compute_geodesic
  19. from da4bci import domain_adaptation
  21. from mne.decoding import CSP
  22. from pyriemann.estimation import Covariances
  23. from pyriemann.tangentspace import TangentSpace
  24. from sklearn.pipeline import make_pipeline
  25. from sklearn.discriminant_analysis import LinearDiscriminantAnalysis
  26. from sklearn.svm import SVC
  27. from sklearn.linear_model import LogisticRegression
  28. from pyriemann.classification import MDM
  31. ###############################################################################
  32. # Distance
  33. ###############################################################################
  35. def session_distance(
  36. Xs: NDArray, Xt: NDArray, dist_type: str, dist_param: Optional[dict] = None,
  37. ) -> float:
  38. dist_param = dist_param or {}
  39. dispatch = {
  40. "mmd": lambda: compute_mmd(Xs, Xt, sigma=dist_param.get("sigma", 1.0)),
  41. "wasserstein": lambda: compute_wasserstein(Xs, Xt),
  42. "energy": lambda: compute_energy(Xs, Xt),
  43. "geodesic": lambda: compute_geodesic(Xs, Xt),
  44. "mahalanobis": lambda: compute_mahalanobis(Xs, Xt),
  45. }
  46. if dist_type not in dispatch:
  47. raise ValueError(f"Unknown distance type: {dist_type!r}")
  48. return float(dispatch[dist_type]())
  51. def distance_ci(
  52. X: NDArray, Y: NDArray, dist_type: str, dist_param: Optional[dict] = None,
  53. B: int = 200, alpha: float = 0.05, est_method: str = "direct",
  54. rng: Optional[np.random.Generator] = None,
  55. ) -> Dict[str, float]:
  56. """Bootstrap confidence interval for session_distance."""
  57. X = np.asarray(X, dtype=float)
  58. Y = np.asarray(Y, dtype=float)
  59. n, m = X.shape[0], Y.shape[0]
  60. if n < 1 or m < 1:
  61. return dict(est=np.inf, lwr=np.inf, upr=np.inf)
  63. if rng is None:
  64. rng = np.random.default_rng()
  66. d_hat = np.nan
  67. if est_method == "direct":
  68. try:
  69. d_hat = session_distance(X, Y, dist_type, dist_param)
  70. except Exception:
  71. d_hat = np.inf
  73. vals = np.full(B, np.nan)
  74. for b in range(B):
  75. try:
  76. vals[b] = session_distance(
  77. X[rng.integers(n, size=n)], Y[rng.integers(m, size=m)],
  78. dist_type, dist_param,
  79. )
  80. except Exception:
  81. pass
  83. finite = vals[np.isfinite(vals)]
  84. if len(finite) < 2:
  85. est = d_hat if np.isfinite(d_hat) else np.inf
  86. return dict(est=est, lwr=np.inf, upr=np.inf)
  88. lwr = float(np.quantile(finite, alpha / 2))
  89. upr = float(np.quantile(finite, 1 - alpha / 2))
  90. est = float(np.mean(finite)) if est_method == "boot_mean" else d_hat
  91. # The direct point estimate can fail (d_hat=inf) even when the bootstrap
  92. # resamples are finite; fall back to the bootstrap mean so the reported
  93. # estimate is never inf while its own CI is finite.
  94. if not np.isfinite(est):
  95. est = float(np.mean(finite))
  96. return dict(est=est, lwr=lwr, upr=upr)
  99. ###############################################################################
  100. # Feature scaling
  101. ###############################################################################
  103. def scale_to_target(
  104. X_list: List[NDArray], X_target: NDArray,
  105. ) -> Tuple[List[NDArray], NDArray]:
  106. """Standardise sources + target to the target's column-wise mean and sd."""
  107. X_target = np.asarray(X_target, dtype=float)
  108. center = X_target.mean(axis=0)
  109. sds = X_target.std(axis=0, ddof=1)
  110. sds[~np.isfinite(sds) | (sds <= 0)] = 1.0
  111. scaled_sources = [(np.asarray(X, dtype=float) - center) / sds for X in X_list]
  112. scaled_target = (X_target - center) / sds
  113. return scaled_sources, scaled_target
  116. def compute_distance_ci_table(
  117. X_sources: List[NDArray], X_target: NDArray, dist_type: str,
  118. dist_param: Optional[dict] = None, *, scale: bool = True,
  119. B: int = 200, alpha: float = 0.05, est_method: str = "boot_mean",
  120. rng_for_index=None,
  121. ) -> List[Dict[str, Any]]:
  122. """Per-source distance-to-target CI table (optionally on target-scaled features).
  124. Returns one ``{"i": i, "est", "lwr", "upr"}`` row per source. ``rng_for_index``
  125. is a callable ``i -> Generator | None`` supplying the bootstrap Generator for
  126. source i; the distance-based pipelines differ only in this per-source seeding
  127. rule, so each passes its own. Default (None) yields nondeterministic CIs.
  128. """
  129. if scale:
  130. x_pre, t_pre = scale_to_target(X_sources, X_target)
  131. else:
  132. x_pre, t_pre = X_sources, X_target
  133. if rng_for_index is None:
  134. def rng_for_index(_i):
  135. return None
  136. return [
  137. {"i": i, **distance_ci(
  138. xs, t_pre, dist_type, dist_param,
  139. B=B, alpha=alpha, est_method=est_method, rng=rng_for_index(i),
  140. )}
  141. for i, xs in enumerate(x_pre)
  142. ]
  145. ###############################################################################
  146. # Domain adaptation
  147. ###############################################################################
  149. def apply_da(da_method, da_control, X_src, X_tgt):
  150. """Apply domain adaptation to a (source, target) feature pair.
  152. Returns the adapted ``(X_src, X_tgt)``. ``da_method == "none"`` is a
  153. pass-through. Missing da4bci result keys fall back to the raw inputs, so a
  154. DA that returns unexpected keys degrades to no-DA instead of raising. This
  155. is the single source/target contract shared by every pipeline; callers that
  156. need to treat a raised DA error specially still wrap the call in try/except.
  157. """
  158. if da_method == "none":
  159. return X_src, X_tgt
  160. da_result = domain_adaptation(
  161. source_data=X_src, target_data=X_tgt,
  162. method=da_method, control=da_control,
  163. )
  164. return (da_result.get("weighted_source_data", X_src),
  165. da_result.get("target_data", X_tgt))
  168. ###############################################################################
  169. # Feature extraction
  170. ###############################################################################
  172. def extract_features_train(
  173. X: NDArray, y: NDArray, feat_name: str, feat_params: Optional[dict] = None,
  174. ) -> Tuple[NDArray, Any]:
  175. """
  176. Fit a feature extractor on training data.
  178. Parameters
  179. ----------
  180. X : 3D array (n_trials, n_channels, n_times)
  181. y : labels
  182. feat_name : "logvar" | "CSP" | "TS"
  184. Returns (features, fitted_object)
  185. """
  186. feat_params = feat_params or {}
  188. if feat_name == "logvar":
  189. return _logvar(X), {"type": "logvar"}
  191. if feat_name == "CSP":
  192. n_components = feat_params.get("n_components", feat_params.get("ncomps", 8))
  193. csp = CSP(n_components=n_components, log=True)
  194. return csp.fit_transform(X, y), {"type": "CSP", "csp": csp}
  196. if feat_name == "TS":
  197. # Tangent-space parametrization of covariance matrices.
  198. # metric="riemann" uses the geometric (Frechet) mean as reference —
  199. # better theoretical grounding but expensive (iterative logm/expm).
  200. # metric="euclid" uses the arithmetic mean — nearly identical accuracy
  201. # but significantly faster on high-dimensional data (e.g. 3x on 60ch).
  202. # See: Dadi et al. (2019) "Benchmarking functional connectome-based
  203. # predictive models for resting-state fMRI", NeuroImage 192, Fig.A10
  204. # & Appendix A. Default changed to "euclid" based on our own benchmark
  205. # (TS_test/): 3x speedup with comparable or better accuracy when
  206. # combined with domain adaptation (SA).
  207. cov_est = Covariances(estimator="lwf")
  208. ts = TangentSpace(metric=feat_params.get("ts_metric", "euclid"))
  209. covs = cov_est.fit_transform(X)
  210. return ts.fit_transform(covs, y), {"type": "TS", "cov_est": cov_est, "ts": ts}
  212. raise ValueError(f"Unknown feature: {feat_name!r}. Supported: logvar, CSP, TS")
  215. def extract_features_test(
  216. X: NDArray, fitted_obj: Any, feat_name: str,
  217. ) -> NDArray:
  218. """Apply a fitted feature extractor to new data."""
  219. if feat_name == "logvar":
  220. return _logvar(X)
  221. if feat_name == "CSP":
  222. return fitted_obj["csp"].transform(X)
  223. if feat_name == "TS":
  224. return fitted_obj["ts"].transform(fitted_obj["cov_est"].transform(X))
  225. raise ValueError(f"Unknown feature: {feat_name!r}")
  228. def _logvar(X: NDArray) -> NDArray:
  229. """Log-variance features: (n_trials, n_channels)."""
  230. return np.log(np.var(X, axis=2) + 1e-10)
  233. ###############################################################################
  234. # Classifier dispatch
  235. ###############################################################################
  237. def get_classifier(clf_name: str, clf_params: Optional[dict] = None) -> Any:
  238. """Create an sklearn-compatible classifier instance."""
  239. p = clf_params or {}
  241. if clf_name == "lda":
  242. return LinearDiscriminantAnalysis(shrinkage=p.get("shrinkage", "auto"), solver="lsqr")
  244. if clf_name == "svm_linear":
  245. return SVC(kernel="linear", C=p.get("cost", p.get("C", 0.1)))
  247. if clf_name == "svm_radial":
  248. gamma = p.get("gamma", 0.1)
  249. return SVC(kernel="rbf", C=p.get("cost", p.get("C", 0.1)),
  250. gamma="scale" if gamma == -1 else gamma)
  252. if clf_name == "mdm":
  253. return MDM()
  255. if clf_name == "el":
  256. from sklearn.linear_model import SGDClassifier
  257. return SGDClassifier(
  258. loss="log_loss", penalty="elasticnet",
  259. l1_ratio=p.get("alpha", 0.5), max_iter=100000, random_state=42,
  260. )
  262. if clf_name == "lr":
  263. return LogisticRegression(penalty="l2", max_iter=20000, solver="lbfgs")
  265. if clf_name == "lgbm":
  266. from lightgbm import LGBMClassifier
  267. return LGBMClassifier(
  268. num_leaves=p.get("num_leaves", 31), learning_rate=p.get("learning_rate", 0.05),
  269. max_depth=p.get("max_depth", -1), n_estimators=p.get("nrounds", 200),
  270. reg_alpha=p.get("lambda_l1", 0.1), reg_lambda=p.get("lambda_l2", 0.1),
  271. subsample=p.get("bagging_fraction", 0.8), colsample_bytree=p.get("feature_fraction", 0.8),
  272. verbose=-1, random_state=p.get("seed", 123),
  273. )
  275. if clf_name == "elm":
  276. from sklearn.kernel_approximation import RBFSampler
  277. from sklearn.linear_model import RidgeClassifier
  278. return make_pipeline(RBFSampler(n_components=p.get("nhid", 500), random_state=1), RidgeClassifier())
  280. if clf_name == "catboost":
  281. from catboost import CatBoostClassifier
  282. return CatBoostClassifier(
  283. depth=p.get("depth", 6), learning_rate=p.get("lr", 0.05),
  284. iterations=p.get("iters", 500), early_stopping_rounds=p.get("es_round", 30), verbose=0,
  285. )
  287. raise ValueError(f"Unknown classifier: {clf_name!r}")

pipeline_utils.py at commit 573c896, under MIT · at the source

Overview

Authors: Yiming Shen1, David Degras1
  1. Department of Mathematics, University of Massachusetts Boston, Boston, MA, United States
Institutions: University of Massachusetts Boston (United States)
Journal: Frontiers in human neuroscience, volume 20, article 1895016
Dates: received 29 May 2026; accepted 11 August 2026; published online 28 August 2026
Type: Research article · Language: English
License: CC BY
Identifiers: DOI 10.3389/fnhum.2026.1895016 · PMID 42729429 · PMCID PMC13562021 · OpenAlex W7204496482
Open access: gold, a free copy (OpenAlex)
Status: code verified (how the code is found and verified)
Categories: EEG (modality), methods / tools (subfield)
Methods: Spectral & time-frequency, Connectivity, Statistics, Machine learning, Preprocessing
Keywords: brain-computer interface, confidence interval, cross-session transfer, longitudinal EEG, motor imagery EEG, multiple source domain adaptation, uncertainty-aware source selection
Topic: EEG and Brain-Computer Interfaces (Cognitive Neuroscience, Neuroscience), according to OpenAlex
Citations: not cited yet (Europe PMC); 50 references in the paper

Abstract

Cross-session variability remains a major obstacle to the reliable operation of motor imagery (MI)-based brain-computer interfaces (BCI), particularly when systems are reused across multiple days. When multiple prior sessions from the same subject are available, two key questions arise before domain transfer: which source sessions to select and how to effectively utilize them. We address these questions by developing confidence-gated, selective-transfer pipelines: a Minimum-Distance Multi-Source Pipeline (MMP) that only uses source sessions close to the target, and a Bridge Domain Pipeline (BDP) that exploits both near and far sources to improve robustness. We evaluated these novel pipelines against uniform-pooling (MAP) and distance-weighted-pooling (DWP) baseline methods on two public motor imagery EEG datasets under matched experimental configurations of feature extraction, classification, and domain adaptation algorithms. The benchmark results support an endpoint-specific interpretation rather than a single accuracy ranking. Specifically, MAP achieved the highest maximum-configuration accuracy, DWP demonstrated the highest average accuracy across configurations, and on the primary dataset MAP, DWP, and BDP exhibited no statistically significant differences as the top-performing pipelines for data-driven configuration selection. In contrast, MMPmta performed similarly under fixed configurations but proved less effective during data-driven configuration selection. Overall, the best-performing proposed pipeline (BDP) ranks among the highest-performing approaches with respect to accuracy while requiring substantially reduced execution time and fewer source sessions than full pooling–demonstrating that BDP transforms the CI-gated retention idea into a more reliable and computationally efficient framework for automated configuration selection. These results indicate that in cross-session MI decoding, the primary challenge in selective transfer extends beyond session selection alone to encompass how retained sessions are integrated downstream, offering significant implications for longitudinal rehabilitation and assistive BCI use.

Reproduced under the paper's license (CC BY), from the paper cited above.

Repositories

Its files are read in the Code ↔ Paper reader above, with 9 matches between paragraphs and lines of code.

Yiming-S/MSDA-Bench

License: MIT
State: the link answers, verified on 27 September 2026
Evidence: files inventoried
Commit: 9cc9097b83141299c08ad16e92bd1e2983c57590, 22 July 2026
Languages: Python (15)
Size: 137 files, 15 scripts
Software Heritage: not archived
Found in: “Data availability statement”
Holds: README, license file, CITATION.cff, environment (requirements.txt, runtime.txt, .devcontainer/devcontainer.json)
Not found: tests, continuous integration, documentation
Tools: NumPy (13 files), pandas (13 files), Plotly (12 files)
Availability: 1 check, the latest on 27 September 2026: the link answers
  • 27 September 2026: the link answers
17 files
At the source: github.com/Yiming-S/MSDA-Bench

Yiming-S/CrossPython

License: MIT
State: the link answers, verified on 27 September 2026
Evidence: files inventoried
Commit: 573c8962133d00d4eecae1aa6a07ab39c99fa22d, 22 July 2026
Languages: Python (28)
Size: 41 files, 28 scripts
Software Heritage: not archived
Found in: “Data availability statement”
Holds: README, license file, CITATION.cff, environment (pyproject.toml), tests, continuous integration, documentation
Tools: NumPy (16 files), scikit-learn (6 files), MNE-Python (3 files), pandas (3 files), LightGBM (1 file), MOABB (1 file), pyRiemann (1 file)
Availability: 1 check, the latest on 27 September 2026: the link answers
  • 27 September 2026: the link answers
30 files
At the source: github.com/Yiming-S/CrossPython

The paper's code and data availability statement is in the Data section.

Tracing map

Proposed by the machine: these links were found in the paper and verified at the source, without human review. The map will receive a Zenodo DOI once one of the paper's authors has validated it with their ORCID.

What the map holds:

  • 2 repositories of the authors' code, each at its verified commit, with its license and how the link was found in the paper;
  • 43 scripts, each with its path and the digest of its content;
  • 9 matches between paragraphs of the paper and lines of the code (method lexical-v1);
  • neither the text of the paper nor the code itself.

Its JSON (tracing-map.json) is deposited on Zenodo with its DOI once the map is validated.

Data

No dataset and no data link were found in the paper.

Data availability statement

Publicly available datasets were analyzed in this study. BNCI2014_004 is accessible through the MOABB library (Jayaram and Barachant, 2018), and Stieger2021 is available as described by (Stieger et al. 2021). The software associated with this study is publicly available on GitHub, including the MSDA-Bench benchmark dashboard (https://github.com/Yiming-S/MSDA-Bench), the CrossPython cross-session pipeline framework (https://github.com/Yiming-S/CrossPython), the DA4BCI R package (https://github.com/Yiming-S/DA4BCI), and the DA4BCI-Python package, distributed as da4bci (https://github.com/Yiming-S/DA4BCI-Python).

Reproduced under the paper's license (CC BY), from the paper cited above.

Versions

The history of this record: each version stored by the harvester or made by a correction of its authors or of the maintainers of its code, and what changed in its facts. The texts of the paper (its abstract, its availability statements) are not part of it; versions that changed only those are not listed.

Version 1, 27 September 2026: the first record

Recorded: type, language, journal, volume, pages, dates, 2 authors, 7 keywords, 38 references.

Cite

This paper

Shen, Y., & Degras, D. (2026). A confidence-gated source selection strategy for cross-session transfer in brain-computer interfaces. Frontiers in human neuroscience, 20, 1895016. https://doi.org/10.3389/fnhum.2026.1895016

BibTeX

@article{shen2026confidence,
author = {Shen, Yiming and Degras, David},
title = {{A confidence-gated source selection strategy for cross-session transfer in brain-computer interfaces}},
journal = {Frontiers in human neuroscience},
year = {2026},
month = aug,
volume = {20},
pages = {1895016},
publisher = {Frontiers Media SA},
issn = {1662-5161},
doi = {10.3389/fnhum.2026.1895016},
url = {https://doi.org/10.3389/fnhum.2026.1895016},
pmid = {42729429},
pmcid = {PMC13562021}
}

RIS

TY - JOUR
AU - Shen, Yiming
AU - Degras, David
TI - A confidence-gated source selection strategy for cross-session transfer in brain-computer interfaces
T2 - Frontiers in human neuroscience
J2 - Front Hum Neurosci
PY - 2026
DA - 2026/08/28
VL - 20
SP - 1895016
SN - 1662-5161
PB - Frontiers Media SA
DO - 10.3389/fnhum.2026.1895016
UR - https://doi.org/10.3389/fnhum.2026.1895016
LA - en
ER -

CSL-JSON

{
"id": "10.3389/fnhum.2026.1895016",
"type": "article-journal",
"title": "A confidence-gated source selection strategy for cross-session transfer in brain-computer interfaces",
"container-title": "Frontiers in human neuroscience",
"author": [
{
"family": "Shen",
"given": "Yiming"
},
{
"family": "Degras",
"given": "David"
}
],
"container-title-short": "Front Hum Neurosci",
"volume": "20",
"page": "1895016",
"DOI": "10.3389/fnhum.2026.1895016",
"PMID": "42729429",
"PMCID": "PMC13562021",
"ISSN": "1662-5161",
"publisher": "Frontiers Media SA",
"URL": "https://doi.org/10.3389/fnhum.2026.1895016",
"language": "en",
"issued": {
"date-parts": [
[
2026,
8,
28
]
]
}
}

The tracing map gets a citation of its own once an author has validated it and it has a DOI.

Similar papers

The papers with a page that share the most with this one: the tools found in their code, their categories, datasets, cited references and authors, the rarest counting most.

[1] doi:10.3389/fnhum.2026.1824613 [code]
Rotation-based metric on the Riemannian manifold of SPD matrices with applications to source data selection for brain-computer interface transfer learning.
Journal: Frontiers in human neuroscience
In common: MOABB, pyRiemann, scikit-learn, 1 other tool, 7 references
[2] doi:10.3390/brainsci13020221 [code]
Subject Separation Network for Reducing Calibration Time of MI-Based BCI.
Journal: Brain sciences
In common: MOABB, pyRiemann, MNE-Python, 4 other tools, methods / tools, 4 references
[3] doi:10.1162/imag_a_00391 [code]
Surfing beta burst waveforms to improve motor imagery-based BCI.
Journal: Imaging neuroscience (Cambridge, Mass.)
In common: MOABB, pyRiemann, MNE-Python, 3 other tools, EEG, 4 references
[4] doi:10.1038/s41597-023-02787-8 [code]
An EEG motor imagery dataset for brain computer interface in acute stroke patients.
Journal: Scientific data
In common: pyRiemann, MNE-Python, scikit-learn, 2 other tools, methods / tools, EEG, 5 references
[5] doi:10.1038/s41597-024-03501-y [code]
An EEG &amp; eye-tracking dataset of ALS patients &amp; healthy people during eye-tracking-based spelling system usage.
Journal: Scientific data
In common: MOABB, pyRiemann, MNE-Python, 3 other tools, methods / tools, EEG, 2 references
[6] doi:10.1162/imag_a_00040 [code]
Harmonizing and aligning M/EEG datasets with covariance-based techniques to enhance predictive regression modeling.
Journal: Imaging neuroscience (Cambridge, Mass.)
In common: pyRiemann, MNE-Python, scikit-learn, 2 other tools, methods / tools, EEG, 4 references
[7] doi:10.1038/s41597-025-04618-4 [code]
Lower limb motor imagery EEG dataset based on the multi-paradigm and longitudinal-training of stroke patients.
Journal: Scientific data
In common: MOABB, pyRiemann, MNE-Python, 4 other tools, EEG, 1 reference
[8] doi:10.1162/imag_a_00511 [code]
Deep Riemannian Networks for end-to-end EEG decoding.
Journal: Imaging neuroscience (Cambridge, Mass.)
In common: pyRiemann, scikit-learn, pandas, 1 other tool, EEG, 5 references
[9] doi:10.3389/fnhum.2024.1421922 [code]
Harnessing Few-Shot Learning for EEG signal classification: a survey of state-of-the-art techniques and future directions.
Journal: Frontiers in human neuroscience
In common: MNE-Python, scikit-learn, pandas, 1 other tool, EEG, 5 references
[10] doi:10.1038/s41597-025-04861-9 [code]
Dataset combining EEG, eye-tracking, and high-speed video for ocular activity analysis across BCI paradigms.
Journal: Scientific data
In common: MOABB, pyRiemann, MNE-Python, 4 other tools, methods / tools, EEG

Contribute

The authors of this paper can claim it, correct its record and validate its tracing map, and the maintainers of its code (its owner, or a public member of its organization) correct what it says of their repository; anyone signed in can ask for its removal. Every request goes to OSCR's own machine, which answers it; your account page follows them.

Sign in with ORCID to claim this paper as one of its authors, correct its record or validate its tracing map: when the paper's metadata lists your ORCID iD, you are recognized at once. Maintainers of its code: sign in with GitHub, then claim the repository on your account page.

Add the badge to the code's README

OSCR: the badge of this paper

If you are an author or a maintainer of this paper's code, you can add this badge. It says what OSCR knows of the paper (its code, its validated tracing map) and links the code to this page. Paste one of these snippets under the README's title. Keep the two marks around it, so that the badge can be replaced in place and is never added twice. OSCR changes nothing in your repository.

Markdown
<!-- oscr-badge:start -->
[![OSCR](https://openscicode.org/badge/doi/10.3389/fnhum.2026.1895016.svg)](https://openscicode.org/badge/doi/10.3389/fnhum.2026.1895016)
<!-- oscr-badge:end -->
reStructuredText
.. oscr-badge:start

.. image:: https://openscicode.org/badge/doi/10.3389/fnhum.2026.1895016.svg
   :target: https://openscicode.org/badge/doi/10.3389/fnhum.2026.1895016
   :alt: OSCR

.. oscr-badge:end
HTML
<!-- oscr-badge:start -->
<a href="https://openscicode.org/badge/doi/10.3389/fnhum.2026.1895016"><img src="https://openscicode.org/badge/doi/10.3389/fnhum.2026.1895016.svg" alt="OSCR"></a>
<!-- oscr-badge:end -->

On GitHub, open the README, choose Edit (the pencil), paste the snippet under the title, then "Commit changes".

If you choose to, you can have the badge kept up to date for you, with oscr badge --write from the command line or with the GitHub Action in your repository's workflows. Every other form and option (a card, AsciiDoc, LaTeX, shields.io) is on the badge page.

Request its removal

To ask OSCR to remove this record, the copies of its authors' scripts or its tracing map, use the removal request page: signed in, you say who you are, what to remove and why, then review and confirm the request. Published rules decide every request (how).

Discussion, reproductions, activity

Discussion: code errors and mismatches between the code and the paper about this paper and its code, from readers and its authors, beside the paper and its tracing map.

Report a mismatch between the code and the paper · Report a code error (signed in)

Reproductions: reports from readers who ran the authors' code and did not get the paper's result, what they ran, with which environment, commit and data.

Report a reproduction failure (signed in)

Versions of its code: the releases of its code tied to a version of this paper (the preprint, the accepted manuscript, the version of record…), each with the tracing map versioned with it.

Activity: what happens around this paper: new versions of its record, its map's validation, discussions and reproductions. It opens with sign-in.