Pith. sign in

REVIEW 4 major objections 7 minor 119 references

Revisiting Euclidean Alignment for Transfer Learning in EEG-Based Brain-Computer Interfaces

T0 review · 4 major / 7 minor · reviewed 2026-08-07 · deepseek-v4-flash

Pith's one-line read Euclidean alignment should be inserted between temporal and spatial filtering in EEG transfer learning pipelines.

desk verdict Useful tutorial from EA's inventor, but the central placement rule rests on a single five-subject dataset with no significance tests; treat it as guidance, not established fact. read the letter →

arxiv 2502.09203 v2 pith:P4QQ2LPY submitted 2025-02-13 cs.HC cs.LG

classification cs.HCcs.LG
keywords EEGbrain-computerinterfacetransferlearningEuclideanalignmentlabelmotorimagerydatapreprocessing
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

Euclidean alignment (EA) whitens each subject's EEG trials so that every domain's mean covariance matrix becomes the identity, making data from different subjects more similar for transfer learning. This paper argues that to get the full benefit, EA must be inserted at a specific point in the decoding pipeline: between temporal filtering and spatial filtering for classical pipelines, and immediately after temporal filtering for deep learning models. The paper confirms this placement on a five-subject motor imagery benchmark, where the recommended pipeline reaches 81.6% plug-and-play accuracy with zero labeled target data, and it surveys results from 13 BCI paradigms that each show accuracy gains from EA. If the placement rule holds, BCI researchers can adopt a standard, free preprocessing step that reduces calibration effort and improves cross-subject decoding.

What carries the argument

The central object is Euclidean alignment: for a domain with $N$ trials $X_n \in \mathbb{R}^{c\times t}$, compute the mean covariance matrix $\bar{R} = \frac{1}{N}\sum_{n=1}^{N} X_n X_n^\top$ and transform each trial by $\tilde{X}_n = \bar{R}^{-1/2}X_n$, so the domain's mean covariance becomes the identity matrix $I$. The alignment is unsupervised, uses two closed-form formulas, and works in the Euclidean space, so any subsequent Euclidean-space classifier can be applied. The paper's placement analysis uses this whitening identity as the mechanism that reduces inter-subject distribution shift, and the experiments test where in the pipeline the transformation matrix should be computed: after temporal filtering (to suppress outliers) and before spatial filtering (so the filter is trained on aligned data).

What would settle it

A direct counter-test would be to run the same three configurations (no EA, EA between temporal and spatial filtering, EA after spatial filtering) on a different motor imagery dataset such as BCI Competition IV Dataset 2a with its nine subjects and four classes; if placing EA after spatial filtering matches or beats the recommended middle placement on average across subjects, the rule would be refuted. A second check would compare deep pipelines (TF-EA-EEGNet vs. EA-EEGNet with band-pass filtering inside the network) on an ERP or SSVEP dataset to see whether the placement advantage also holds beyond motor imagery.

Watch

Extended reading notes

Core claim

On the paper's own terms, the central discovery is that Euclidean alignment should be placed between the temporal filtering and spatial filtering blocks in the traditional BCI transfer learning pipeline, and between the temporal filtering block and the deep learning module in deep pipelines. The paper's new experiments on BCI Competition IV Dataset 1 show that EA placed there almost always outperforms both no-EA and EA placed after spatial filtering, because temporal filtering improves the quality of the covariance matrices that EA whitens, and aligning before spatial filtering lets filters like RCSP be learned on distributionally similar data. The paper also compiles results from 13 BCI paradigms to argue that EA is a broadly applicable, efficient, unsupervised preprocessing step, and recommends it as a standard component for cross-subject models.

Load-bearing premise

The load-bearing premise is that one motor imagery dataset (BCI Competition IV Dataset 1, five subjects with matched left/right classes) plus a heterogeneous table of literature results across 13 paradigms is enough to establish both the placement rule and its broad generality.

Editorial extensions

If this is right

  • Motor imagery BCIs can operate with no per-subject calibration: the best traditional pipeline reaches 81.6% accuracy with zero labeled target trials, and accuracy rises only slowly as labeled target data are added.
  • For deep decoding, explicit temporal filtering before EA is necessary: EEGNet-only and EA-without-filtering stay near chance, while TF-EA-EEGNet averages 66.3% accuracy across subjects with no calibration.
  • The pipeline TF-CAR-EA-RCSP-wAR becomes a strong default baseline for cross-subject motor imagery classification, outperforming the same pipeline with no EA or with EA after spatial filtering.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The 13-paradigm evidence in Table III pools results from different datasets, classifiers, and papers; a controlled multi-dataset benchmark under identical preprocessing would make the placement rule directly comparable across paradigms.
  • The placement rationale implies that any preprocessing that improves the covariance estimate (for example, artifact removal or channel selection) should be ordered before EA, while any step that assumes aligned data should come after it.
  • If the reported convergence speedup from aligned training holds for larger models, EA could reduce the compute and data needed to pretrain large EEG foundation models, which the paper identifies as an open direction.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 7 minor

Summary. The paper revisits Euclidean alignment (EA) for transfer learning in EEG-based BCIs. It restates the EA algorithm, proves in Eq. (3) that EA whitens the mean covariance matrix of each domain, and presents new experiments on BCI Competition IV Dataset 1 (five-subject, two-class motor imagery) to argue that EA should be placed between temporal filtering and spatial filtering in traditional pipelines and after temporal filtering before deep learning modules. The paper also surveys 13 BCI paradigms in which EA has been reported to improve accuracy, compares EA with related alignment methods such as Riemannian alignment and label alignment, and discusses extensions including adversarial robustness and privacy preservation. The central prescriptive claim is that the recommended placement constitutes the 'correct usage' of EA.

Significance. If the placement rule holds, it offers a simple and actionable guideline for BCI practitioners and may explain inconsistent results in prior EA applications. The paper's literature compilation in Table III and the method comparison in Table V are useful resources, and the derivation of the core EA property in Eq. (3) is clear and correct. The paper also explicitly credits related work and quotes independent evaluations, which strengthens its tutorial value. However, the experimental support for the placement rule is narrow and lacks statistical treatment; at this stage the contribution is best characterized as a tutorial with a preliminary recommendation rather than a fully validated design rule.

major comments (4)
  1. [Sec. II-B, Fig. 4] The central placement recommendation is supported only by averages over 30 random repeats on one two-class motor-imagery dataset with five subjects, with no error bars, no per-subject variance, and no significance tests. The text uses 'always' (Sec. II-B, item 1) and 'almost always' (item 2) without statistical support. Please add per-subject results with confidence intervals or paired significance tests (e.g., Wilcoxon signed-rank across the five subjects), or substantially soften the claims to 'in our experiments on this dataset.'
  2. [Sec. II-C, Table II] The deep-learning comparison reports only three repeats per subject and no variance. The EA-EEGNet average (48.00%) is actually lower than the EEGNet-only average (49.33%), so the conclusion that 'TF-EA-EEGNet performed much better, suggesting the necessity to apply temporal filtering explicitly before EA' requires a significance check; with five subjects these differences may be within noise.
  3. [Sec. II-E, Table III] Table III documents that EA improves classification accuracy within each individual study, but those studies use different datasets, classifiers, and preprocessing pipelines, and they do not compare alternative EA placements. The abstract and conclusions claim that 'numerous experiments from 13 different BCI paradigms demonstrated its effectiveness and efficiency,' but those are literature results, not new experiments, and they do not validate the placement rule. Please separate the effectiveness claim from the placement claim and avoid implying that Table III supports the placement recommendation.
  4. [Sec. II-B, rationale paragraphs] The stated reasons for the recommended placement—that temporal filtering improves covariance-matrix quality and that EA before RCSP helps spatial filters generalize across subjects—are plausible but post hoc and are not tested by any ablation. A direct test (e.g., comparing covariance-matrix quality or RCSP filter consistency with and without EA at each position) would strengthen the causal claim; otherwise, the rationale should be explicitly labeled as a hypothesis.
minor comments (7)
  1. [Abstract and captions] There are typos: 'br ain-computer' in the abstract and 'Euclidian' in the Figure 3(a) caption; please correct them.
  2. [References] Reference [107] spells 'General Data Protection Regulation' as 'General Data Protection Regularization', and reference [118] has 'Jouranl' for 'Journal'; please fix these.
  3. [Sec. II-D] The claim that Figure 6 shows a 'clear diagonal pattern' is based on visual inspection; consider quantifying diagonal dominance with a numerical metric.
  4. [Sec. II-E] The champion and runner-up BEETL approaches are both cited as [72], which is the competition summary paper; please cite the actual team papers or clarify the attribution.
  5. [Sec. II-A] The phrase 'semi positive-definite' should be 'positive semidefinite'.
  6. [Table V] The 'Online or Offline' row is difficult to read because the repeated 'Both' entries are not separated from the following column; consider reformatting the table for clarity.
  7. [Sec. IV-D] The statement 'EA is essential in adversarial training' is too strong; the cited ABA T algorithm uses EA, but 'essential' is not established by the experiments presented.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the paper's central claims are supported by direct experiments and external benchmarks, not reduced to their inputs by construction.

full rationale

The paper's derivation chain is not circular. The key mathematical property in Eq. (3), that the mean covariance matrix after Euclidean alignment is the identity, follows algebraically from the definition of EA in Eqs. (1)-(2); it is an identity true by construction, not an empirical prediction fitted from data, so it does not constitute circularity. The central prescriptive claim, that EA should be placed between temporal filtering and spatial filtering, is introduced as a proposal from prior same-author work [14] but is then tested in new experiments on BCI Competition IV Dataset 1 in Sections II-B and II-C, comparing six traditional pipelines and three deep-learning configurations. These are genuine leave-one-subject-out evaluations with held-out target data; their weakness is limited evidence (one dataset, five subjects, no significance tests), which is a robustness concern, not a circularity concern. Table III surveys 13 paradigms and largely supports EA's effectiveness, not the placement rule; although several rows cite the author's own papers ([13], [15], [42], [11]), the table also includes many independent groups, and the effectiveness claim is independently corroborated by external works such as [30], [41], and [55]. Self-citations appear in the discussion of adversarial training and privacy-preserving decoding (e.g., [104], [105], [119]), but these are forward-looking suggestions and survey statements, not load-bearing derivation steps that reduce the paper's conclusions to its own premises. No specific equation or fitted parameter can be exhibited that makes a claimed prediction equivalent to its input, so the appropriate score is 0.

Assumptions & free parameters 0 free parameters · 3 assumptions · 0 invented entities

The EA whitening property in Eq. (3) is a parameter-free identity: once Rbar is defined as the mean covariance, transforming by Rbar^{-1/2} makes the mean covariance equal to I. No free parameters are fitted to obtain that identity. The paper's empirical recommendations rest on domain assumptions about EEG shift and on a single-dataset experiment; no new entities are introduced.

assumptions (3)
  • standard math The per-domain mean covariance matrix Rbar is invertible, so the inverse square root Rbar^{-1/2} exists.
    Needed for Eq. (2); in EEG practice with enough trials and c channels this is usually true, but singular cases are not discussed.
  • domain assumption Whitening each domain so its mean covariance becomes the identity reduces cross-subject distribution shift and improves transfer learning.
    Core premise of EA, Section II.A; the paper treats this as generally true across 13 paradigms rather than proving it per paradigm.
  • domain assumption Temporal filtering improves covariance estimation, so EA should be applied after temporal filtering and before spatial filtering.
    Explanatory claim in Section II.B observation 2; only tested on one motor imagery dataset.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Revisiting Euclidean Alignment for Transfer Learning in EEG-Based Brain-Computer Interfaces." pith.science (2026). https://pith.science/paper/P4QQ2LPY

@misc{pith2026250209203,
  author       = {Pith},
  title        = {Pith review of: Revisiting Euclidean Alignment for Transfer Learning in EEG-Based Brain-Computer Interfaces},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/P4QQ2LPY}},
  note         = {Machine review of arXiv:2502.09203}
}
read the original abstract

Due to large intra-subject and inter-subject variabilities of electroencephalogram (EEG) signals, EEG-based brain-computer interfaces (BCIs) usually need subject-specific calibration to tailor the decoding algorithm for each new subject, which is time-consuming and user-unfriendly, hindering their real-world applications. Transfer learning (TL) has been extensively used to expedite the calibration, by making use of EEG data from other subjects/sessions. An important consideration in TL for EEG-based BCIs is to reduce the data distribution discrepancies among different subjects/sessions, to avoid negative transfer. Euclidean alignment (EA) was proposed in 2020 to address this challenge. Numerous experiments from 13 different BCI paradigms demonstrated its effectiveness and efficiency. This paper revisits EA, explaining its procedure and correct usage, introducing its applications and extensions, and pointing out potential new research directions. It should be very helpful to BCI researchers, especially those who are working on EEG signal decoding.

Figures

Figures reproduced from arXiv: 2502.09203 by the authors.

Figure 1
Figure 1. Illustration of Euclidean alignment and label align [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 3
Figure 3. Different configurations of the transfer learning pi [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figure 5
Figure 5. The optimal location to perform Euclidean alignment [PITH_FULL_IMAGE:figures/full_fig_p005_5.png] view at source ↗
Figures from the paper (3 more)
Figure 6
Figure 6. Figure 6: Visualization of the EA transformation matrix [PITH_FULL_IMAGE:figures/full_fig_p005_6.png]
Figure 7
Figure 7. Figure 7: The motor imagery classification performance of six d [PITH_FULL_IMAGE:figures/full_fig_p008_7.png]
Figure 9
Figure 9. Figure 9: The aligned and augmented adversarial ensemble (A3E [PITH_FULL_IMAGE:figures/full_fig_p011_9.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

119 extracted references · 77 canonical work pages

  1. [14]

    Transfer learning for moto r imagery based brain-computer interfaces: A tutorial,

    D. Wu, X. Jiang, and R. Peng, “Transfer learning for moto r imagery based brain-computer interfaces: A tutorial,” Neural Networks , vol. 153, pp. 235–253, 2022

  2. [1]

    Facilitating calibration in high-speed BCI spellers via l everaging cross-device shared latent responses,

    M. Nakanishi, Y .-T. Wang, C.-S. Wei, K.-J. Chiang, and T. -P . Jung, “Facilitating calibration in high-speed BCI spellers via l everaging cross-device shared latent responses,” IEEE Trans. on Biomedical Engineering, vol. 67, no. 4, pp. 1105–1113, 2020

  3. [2]

    Self-paced (asynchronous) BCI cont rol of a wheelchair in virtual environments: a case study with a tetr aplegic,

    R. Leeb, D. Friedman, G. R. M¨ uller-Putz, R. Scherer, M. S later, and G. Pfurtscheller, “Self-paced (asynchronous) BCI cont rol of a wheelchair in virtual environments: a case study with a tetr aplegic,” Computational Intelligence and Neuroscience , vol. 2007, no. 1, p. 079642, 2007

  4. [3]

    BCI for stroke rehabili tation: motor and beyond,

    R. Mane, T. Chouhan, and C. Guan, “BCI for stroke rehabili tation: motor and beyond,” Journal of Neural Engineering , vol. 17, no. 4, p. 041001, 2020

  5. [4]

    Affective brain-comp uter interfaces (aBCIs): A tutorial,

    D. Wu, B.-L. Lu, B. Hu, and Z. Zeng, “Affective brain-comp uter interfaces (aBCIs): A tutorial,” Proc. of the IEEE , vol. 11, no. 10, pp. 1314–1332, 2023

  6. [5]

    Neuroprosthesis for decoding speech in a paralyzed person with anarthria,

    D. A. Moses, S. L. Metzger, J. R. Liu, G. K. Anumanchipalli , J. G. Makin, P . F. Sun, J. Chartier, M. E. Dougherty, P . M. Liu, G. M. Abrams et al. , “Neuroprosthesis for decoding speech in a paralyzed person with anarthria,” New England Journal of Medicine , vol. 385, no. 3, pp. 217–227, 2021

  7. [6]

    Walking naturally after spinal cord injury using a brain-spine interface,

    H. Lorach, A. Galvez, V . Spagnolo, F. Martel, S. Karakas, N. Intering, M. V at, O. Faivre, C. Harte, S. Komi et al. , “Walking naturally after spinal cord injury using a brain-spine interface,” Nature, vol. 618, no. 7963, pp. 126–133, 2023

  8. [7]

    China brain p roject: from bench to bedside,

    X. Liu, T. Gao, T. Lu, G. Schumann, and L. Lu, “China brain p roject: from bench to bedside,” Science Bulletin, vol. 68, no. 5, pp. 444–447, 2023

Show all 119 references
  1. [8]

    Seven technologies to watch in 2024,

    M. Eisenstein, “Seven technologies to watch in 2024,” Nature, vol. 625, no. 7996, pp. 844–848, 2024

  2. [9]

    Science in 2025: the events to watch for in the coming year,

    M. Naddaf, “Science in 2025: the events to watch for in the coming year,” Nature, vol. 637, no. 8044, pp. 9–11, 2025

  3. [10]

    Technology predictions 2025,

    D. Milojicic, “Technology predictions 2025,” Computer, vol. 58, no. 1, pp. 88–90, 2025

  4. [11]

    Transfer learning for EEG-ba sed brain- computer interfaces: A review of progress made since 2016,

    D. Wu, Y . Xu, and B.-L. Lu, “Transfer learning for EEG-ba sed brain- computer interfaces: A review of progress made since 2016,” IEEE Trans. on Cognitive and Developmental Systems , vol. 14, no. 1, pp. 4–19, 2022

  5. [12]

    A survey on negati ve transfer,

    W. Zhang, L. Deng, L. Zhang, and D. Wu, “A survey on negati ve transfer,” IEEE/CAA Journal of Automatica Sinica , vol. 10, no. 2, pp. 305–329, 2023

  6. [13]

    Transfer learning for brain-computer i nterfaces: A Euclidean space data alignment approach,

    H. He and D. Wu, “Transfer learning for brain-computer i nterfaces: A Euclidean space data alignment approach,” IEEE Trans. on Biomedical Engineering, vol. 67, no. 2, pp. 399–410, 2020

  7. [15]

    Different set domain adaptation for bra in-computer interfaces: A label alignment approach,

    H. He and D. Wu, “Different set domain adaptation for bra in-computer interfaces: A label alignment approach,” IEEE Trans. on Neural Sys- tems and Rehabilitation Engineering , vol. 28, no. 5, pp. 1091–1108, 2020

  8. [16]

    Trans- fer learning: a Riemannian geometry framework with applica tions to brain-computer interfaces,

    P . Zanini, M. Congedo, C. Jutten, S. Said, and Y . Berthou mieu, “Trans- fer learning: a Riemannian geometry framework with applica tions to brain-computer interfaces,” IEEE Trans. on Biomedical Engineering , vol. 65, no. 5, pp. 1107–1116, 2018

  9. [17]

    Channel and trial selection for reducin g covariate shift in EEG-based brain-computer interfaces,

    H. He and D. Wu, “Channel and trial selection for reducin g covariate shift in EEG-based brain-computer interfaces,” in Proc. IEEE Int’l Conf. on Systems, Man and Cybernetics , Bari, Italy, Oct. 2019

  10. [18]

    Towards adaptive class ification using Riemannian geometry approaches in brain-computer interfa ces,

    S. Kumar, F. Yger, and F. Lotte, “Towards adaptive class ification using Riemannian geometry approaches in brain-computer interfa ces,” in Int’l Winter Conference on Brain-Computer Interface , Gangwon, South Korea, Feb. 2019, pp. 1–6

  11. [19]

    Cross-data set variability problem in EEG decoding with deep learning,

    L. Xu, M. Xu, Y . Ke, X. An, S. Liu, and D. Ming, “Cross-data set variability problem in EEG decoding with deep learning,” Frontiers in Human Neuroscience, vol. 14, p. 103, 2020

  12. [20]

    Enha ncing transfer performance across datasets for brain-computer i nterfaces using a combination of alignment strategies and adaptive ba tch nor- malization,

    L. Xu, M. Xu, Z. Ma, K. Wang, T.-P . Jung, and D. Ming, “Enha ncing transfer performance across datasets for brain-computer i nterfaces using a combination of alignment strategies and adaptive ba tch nor- malization,” Journal of Neural Engineering , vol. 18, no. 4, p. 0460e5, 2021

  13. [21]

    Sub-band target alignment common spatial pattern in brain-computer interface,

    X. Zhang, Q. She, Y . Chen, W. Kong, and C. Mei, “Sub-band target alignment common spatial pattern in brain-computer interface,” Computer Methods and Programs in Biomedicine , vol. 207, p. 106150, 2021

  14. [22]

    Manifold trial selection to reduce negative transfer in mo tor imagery- based brain–computer interface,

    Z. Liang, Z. Zheng, W. Chen, J. Wang, J. Zhang, J. Chen, an d Z. Chen, “Manifold trial selection to reduce negative transfer in mo tor imagery- based brain–computer interface,” in IEEE/RSJ Int’l Conf. on Intelligent Robots and Systems , Prague, Czech Republic, Sep. 2021, pp. 4...

  15. [23]

    Transfer learning ba sed on hybrid Riemannian and Euclidean space data alignment and su bject selection in brain-computer interfaces,

    Y . Li, Q. Wei, Y . Chen, and X. Zhou, “Transfer learning ba sed on hybrid Riemannian and Euclidean space data alignment and su bject selection in brain-computer interfaces,” IEEE Access, vol. 9, pp. 6201– 6212, 2021

  16. [24]

    Cross-subject trials reweighting for enhancing motor ima gery-based brain-computer interface,

    Z. Liang, Z. Zheng, W. Chen, J. Wang, J. Zhang, J. Chen, an d H. Shi, “Cross-subject trials reweighting for enhancing motor ima gery-based brain-computer interface,” in IEEE Conf. on Industrial Electronics and Applications, 2022, pp. 943–948

  17. [25]

    CovMix: Covariance mixi ng regulariza- tion for motor imagery decoding,

    G. Zoumpourlis and I. Patras, “CovMix: Covariance mixi ng regulariza- tion for motor imagery decoding,” in 2022 10th International Winter Conference on Brain-Computer Interface , Gangwon, South Korea, 2022, pp. 1–7

  18. [26]

    Privacy-preserving domain adaptation for motor imagery-based brain-computer interf aces,

    K. Xia, L. Deng, W. Duch, and D. Wu, “Privacy-preserving domain adaptation for motor imagery-based brain-computer interf aces,” IEEE Trans. on Biomedical Engineering , vol. 69, no. 11, pp. 3365–3376, 2022

  19. [27]

    Riemannian geometric and ensemble learning for decoding c ross- session motor imagery electroencephalography signals,

    L. Pan, K. Wang, L. Xu, X. Sun, W. Yi, M. Xu, and D. Ming, “Riemannian geometric and ensemble learning for decoding c ross- session motor imagery electroencephalography signals,” Journal of Neural Engineering , vol. 20, no. 6, p. 066011, 2023

  20. [28]

    Manifold embedded instance selection to suppress negative transfer in motor imagery-based brain–computer interface,

    Z. Liang, Z. Zheng, W. Chen, Z. Pei, J. Wang, and J. Chen, “ Manifold embedded instance selection to suppress negative transfer in motor imagery-based brain–computer interface,” Biomedical Signal Process- ing and Control , vol. 88, p. 105556, 2024

  21. [29]

    Calibration-free o nline test- time adaptation for Electroencephalography motor imagery decoding,

    M. Wimpff, M. D¨ obler, and B. Y ang, “Calibration-free o nline test- time adaptation for Electroencephalography motor imagery decoding,” in Int’l Winter Conf. on Brain-Computer Interface , Gangwon, Korea, Feb. 2024, pp. 1–6

  22. [30]

    A systematic evaluation of euclidean alignment with deep l earning for EEG decoding,

    B. Junqueira, B. Aristimunha, S. Chevallier, and R. Y . d e Camargo, “A systematic evaluation of euclidean alignment with deep l earning for EEG decoding,” Journal of Neural Engineering , vol. 21, no. 3, p. 036038, 2024

  23. [31]

    Tra nsfer learning with data alignment and optimal transport for EEG b ased motor imagery classification,

    C. Chu, L. Zhu, A. Huang, P . Xu, N. Ying, and J. Zhang, “Tra nsfer learning with data alignment and optimal transport for EEG b ased motor imagery classification,” Journal of Neural Engineering , vol. 21, no. 1, p. 016015, 2024

  24. [32]

    Joint Spatio-Temp oral Filtering of Motion Imagery EEG Signals for Data Alignment in Transfer Learning,

    A. Jiang, S. Hou, Y . Tang, and Y . Zhu, “Joint Spatio-Temp oral Filtering of Motion Imagery EEG Signals for Data Alignment in Transfer Learning,” in Proc. IEEE Int’l Conf. on Acoustics, Speech and Signal Processing, Seoul, South Korea, Apr. 2024, pp. 2235–2239. 13

  25. [33]

    Riemanni an procrustes analysis: Transfer learning for brain–computer interface s,

    P . L. C. Rodrigues, C. Jutten, and M. Congedo, “Riemanni an procrustes analysis: Transfer learning for brain–computer interface s,” IEEE Trans. on Biomedical Engineering , vol. 66, no. 8, pp. 2390–2401, 2019

  26. [34]

    A novel heterogen eous transfer learning method based on data stitching for the seq uential coding brain computer interface,

    Q. Zhan, L. Wang, L. Ren, and X. Huang, “A novel heterogen eous transfer learning method based on data stitching for the seq uential coding brain computer interface,” Computers in Biology and Medicine , vol. 151, p. 106220, 2022

  27. [35]

    Cr oss- subject mindfulness meditation EEG decoding,

    A.-I. Karaiskou, C. V aron, K. Alaerts, and M. De V os, “Cr oss- subject mindfulness meditation EEG decoding,” in IEEE Int’l Conf. on Metrology for eXtended Reality, Artificial Intelligence and Neural Engineering, Milan, Italy, October 2023, pp. 781–786

  28. [36]

    M ean- based geodesic distance alignment transfer for decoding na tural hand movement from MRCPs,

    M. Xue, B. Xu, L. Li, J. Ping, M. Miao, H. Li, and A. Song, “M ean- based geodesic distance alignment transfer for decoding na tural hand movement from MRCPs,” Measurement, vol. 247, p. 116836, 2025

  29. [37]

    The perils and pitfalls of blo ck design for EEG classification experiments,

    R. Li, J. S. Johansen, H. Ahmed, T. V . Ilyevsky, R. B. Wilb ur, H. M. Bharadwaj, and J. M. Siskind, “The perils and pitfalls of blo ck design for EEG classification experiments,” IEEE Trans. on Pattern Analysis and Machine Intelligence , vol. 43, no. 1, pp. 316–333, 2021

  30. [38]

    Regularized common spatial pattern with aggregation for EEG classification in small-sample setting,

    H. Lu, H.-L. Eng, C. Guan, K. N. Plataniotis, and A. N. V en etsanopou- los, “Regularized common spatial pattern with aggregation for EEG classification in small-sample setting,” IEEE Trans. on Biomedical Engineering, vol. 57, no. 12, pp. 2936–2946, 2010

  31. [39]

    Online and offline domain adaptation for reducin g BCI calibration effort,

    D. Wu, “Online and offline domain adaptation for reducin g BCI calibration effort,” IEEE Trans. on Human-Machine Systems , vol. 47, no. 4, pp. 550–563, 2017

  32. [40]

    EEGNet: a compact convolutional neural net work for EEG-based brain-computer interfaces,

    V . J. Lawhern, A. J. Solon, N. R. Waytowich, S. M. Gordon, C. P . Hung, and B. J. Lance, “EEGNet: a compact convolutional neural net work for EEG-based brain-computer interfaces,” Journal of Neural Engineering , vol. 15, no. 5, p. 056013, 2018

  33. [41]

    Thinker invariance: enablin g deep neural networks for BCI across more people,

    D. Kostas and F. Rudzicz, “Thinker invariance: enablin g deep neural networks for BCI across more people,” Journal of Neural Engineering , vol. 17, no. 5, p. 056008, 2020

  34. [42]

    Manifold embedded knowledge transf er for brain-computer interfaces,

    W. Zhang and D. Wu, “Manifold embedded knowledge transf er for brain-computer interfaces,” IEEE Trans. on Neural Systems and Reha- bilitation Engineering , vol. 28, no. 5, pp. 1117–1127, 2020

  35. [43]

    Team cogitat at neurips 2021 : Benchmarks for EEG transfer learning competition,

    S. Bakas, S. Ludwig, K. Barmpas, M. Bahri, Y . Panagakis, N. Laskaris, D. A. Adamos, and S. Zafeiriou, “Team cogitat at neurips 2021 : Benchmarks for EEG transfer learning competition,” arXiv preprint arXiv:2202.03267, 2022

  36. [44]

    An improved vers ion of local activities estimation to enhance motor imagery class ification,

    M. M. Togha, M. R. Salehi, and E. Abiri, “An improved vers ion of local activities estimation to enhance motor imagery class ification,” Biomedical Signal Processing and Control , vol. 66, p. 102485, 2021

  37. [45]

    Inter-sub ject transfer learning using Euclidean alignment and transfer component analysis for motor imagery-based BCI,

    O. Demsy, D. Achanccaray, and M. Hayashibe, “Inter-sub ject transfer learning using Euclidean alignment and transfer component analysis for motor imagery-based BCI,” in IEEE Int’l Conference on Systems, Man, and Cybernetics , Melbourne, Australia, Oct. 2021, pp. 3176–3176

  38. [46]

    Transfer learning based on optimal transport for motor imagery brain-computer interfaces,

    V . Peterson, N. Nieto, D. Wyser, O. Lambercy, R. Gassert , D. H. Milone, and R. D. Spies, “Transfer learning based on optimal transport for motor imagery brain-computer interfaces,” IEEE Trans. on Biomed- ical Engineering , vol. 69, no. 2, pp. 807–817, 2021

  39. [47]

    Motor imagery EEG decoding using manifold embedded transfer learning,

    Y . Cai, Q. She, J. Ji, Y . Ma, J. Zhang, and Y . Zhang, “Motor imagery EEG decoding using manifold embedded transfer learning,” Journal of Neuroscience Methods, vol. 370, p. 109489, 2022

  40. [48]

    EEG signal decodin g of motor imagination based on euclidean space-weighted logistic re gression transfer learning,

    L. Chen, A. Gong, P . Ding, and et al., “EEG signal decodin g of motor imagination based on euclidean space-weighted logistic re gression transfer learning,” Journal of Nanjing University , vol. 58, no. 2, pp. 264–274, 2022

  41. [49]

    Improving NeuCube spiking neural network for EEG-based pa ttern recognition using transfer learning,

    X. Wu, Y . Feng, S. Lou, H. Zheng, B. Hu, Z. Hong, and J. Tan, “Improving NeuCube spiking neural network for EEG-based pa ttern recognition using transfer learning,” Neurocomputing, vol. 529, pp. 222–235, 2023

  42. [50]

    Intra-and inter-subject common spa tial pattern for reducing calibration effort in MI-based BCI,

    Q. Wei and X. Ding, “Intra-and inter-subject common spa tial pattern for reducing calibration effort in MI-based BCI,” IEEE Trans. on Neural Systems and Rehabilitation Engineering , vol. 31, pp. 904–916, 2023

  43. [51]

    Dual con trastive training and transferability aware adaptation for multi-s ource privacy- preserving motor imagery classification,

    J. Zhu, G. Xu, Q. Hu, B. Wang, T. Zhou, and J. Qin, “Dual con trastive training and transferability aware adaptation for multi-s ource privacy- preserving motor imagery classification,” IEEE Trans. on Instrumenta- tion and Measurement , 2023

  44. [52]

    Double stag e transfer learning for brain–computer interfaces,

    Y . Gao, M. Li, Y . Peng, F. Fang, and Y . Zhang, “Double stag e transfer learning for brain–computer interfaces,” IEEE Trans. on Neural Systems and Rehabilitation Engineering , vol. 31, pp. 1128–1136, 2023

  45. [53]

    EEG multi-domai n feature transfer based on sparse regularized Tucker decomposition ,

    Y . Gao, C. Zhang, J. Huang, and M. Meng, “EEG multi-domai n feature transfer based on sparse regularized Tucker decomposition ,” Cognitive Neurodynamics, vol. 18, no. 1, pp. 185–197, 2024

  46. [54]

    Time-frequency trans form based cross-subject EEG data augmentation,

    Z. Wang, S. Li, X. Chen, and D. Wu, “Time-frequency trans form based cross-subject EEG data augmentation,” Knowledge Based Systems , vol. 311, p. 113074, 2025

  47. [55]

    Latent alignment in deep learning models for EEG decoding,

    S. Bakas, S. Ludwig, D. A. Adamos, N. Laskaris, Y . Panaga kis, and S. Zafeiriou, “Latent alignment in deep learning models for EEG decoding,” Journal of Neural Engineering , vol. 22, p. 016047, 2025

  48. [56]

    Euclidean alignment for transfer learning in multi- band common spatial pattern,

    M. M. Amorim, L. Prata, J. S. Maur´ ıcio, A. Borges, H. Ber nardino, and G. de Souza, “Euclidean alignment for transfer learning in multi- band common spatial pattern,” in Brazilian Conf. on Intelligent Systems, Bel´ em do Par´ a, Brazil, Dec. 2024, pp. 430–443

  49. [57]

    SPDIM: Source-fre e unsuper- vised conditional and label shifts adaptation in EEG,

    S. Li, M. Kawanabe, and R. J. Kobler, “SPDIM: Source-fre e unsuper- vised conditional and label shifts adaptation in EEG,” arXiv preprint arXiv:2411.07249, 2024

  50. [58]

    TFTL: A task-free transfer learning strategy for EEG- based cross-subject & cross-dataset motor imagery BCI,

    Y . Wang, J. Wang, W. Wang, J. Su, C. Bunterngchit, and Z.- G. Hou, “TFTL: A task-free transfer learning strategy for EEG- based cross-subject & cross-dataset motor imagery BCI,” IEEE Trans. on Biomedical Engineering , vol. 72, pp. 810–821, 2024

  51. [59]

    Lightweight sour ce- free domain adaptation based on adaptive Euclidean alignme nt for brain-computer interfaces,

    H. Wang, H. Han, J. Q. Gan, and H. Wang, “Lightweight sour ce- free domain adaptation based on adaptive Euclidean alignme nt for brain-computer interfaces,” IEEE Journal of Biomedical and Health Informatics, vol. 29, pp. 909–922, 2024

  52. [60]

    Towa rds cross-brain computer interface: A prototype-supervised a dversarial transfer learning approach with multiple sources,

    J. Zhang, J. Fang, S. Liu, D. Liu, H. Wu, and J. Long, “Towa rds cross-brain computer interface: A prototype-supervised a dversarial transfer learning approach with multiple sources,” IEEE Trans. on Instrumentation and Measurement , vol. 73, p. 2528113, 2024

  53. [61]

    A relat ion feature comparison network for cross-domain recognition o f motion intention,

    J. Xu, D. Li, P . Zhou, Y . Zhang, Z. Wang, and D. Ma, “A relat ion feature comparison network for cross-domain recognition o f motion intention,” IEEE Trans. on Instrumentation and Measurement , vol. 73, p. 4008513, 2024

  54. [62]

    Suppres sion of negative transfer in motor imagery brain–computer interfa ce based on mutual information and Pearson correlation coefficient,

    F. Zhu, J. Cai, H. Zheng, Z. Liang, and Y . Zhang, “Suppres sion of negative transfer in motor imagery brain–computer interfa ce based on mutual information and Pearson correlation coefficient, ” Review of Scientific Instruments , vol. 95, no. 7, 2024

  55. [63]

    A novel multi-morphological representation approach for mu lti-source EEG signals,

    Y . Gao, Y . Liu, M. Meng, F. Fang, M. Houston, and Y . Zhang, “A novel multi-morphological representation approach for mu lti-source EEG signals,” Neurocomputing, vol. 617, p. 129010, 2025

  56. [64]

    A cross-dataset adaptive domain selection transfer learn ing framework for motor imagery-based brain-computer interfaces,

    J. Jin, G. Bai, R. Xu, K. Qin, H. Sun, X. Wang, and A. Cichoc ki, “A cross-dataset adaptive domain selection transfer learn ing framework for motor imagery-based brain-computer interfaces,” Journal of Neural Engineering, vol. 21, no. 3, p. 036057, 2024

  57. [65]

    Se lf- supervised contrastive learning for EEG-based cross-subj ect motor imagery recognition,

    W. Li, H. Li, X. Sun, H. Kang, S. An, G. Wang, and Z. Gao, “Se lf- supervised contrastive learning for EEG-based cross-subj ect motor imagery recognition,” Journal of Neural Engineering , vol. 21, no. 2, p. 026038, 2024

  58. [66]

    Fine- tuning strategies for continual online EEG motor imagery de cod- ing: Insights from a large-scale longitudinal study,

    M. Wimpff, B. Aristimunha, S. Chevallier, and B. Y ang, “ Fine- tuning strategies for continual online EEG motor imagery de cod- ing: Insights from a large-scale longitudinal study,” arXiv preprint arXiv:2502.06828, 2025

  59. [67]

    Symmetric convolutional and adversarial neural network enables improved mental stress classificati on from EEG,

    R. Fu, Y .-F. Chen, Y . Huang, S. Chen, F. Duan, J. Li, J. Wu, D. Jiang, J. Gao, J. Gu et al., “Symmetric convolutional and adversarial neural network enables improved mental stress classificati on from EEG,” IEEE Trans. on Neural Systems and Rehabilitation Engineeri ng, vol. 3...

  60. [68]

    Emotion rec ognition via multiscale feature fusion network and attention mechan ism,

    Y . Jiang, S. Xie, X. Xie, Y . Cui, and H. Tang, “Emotion rec ognition via multiscale feature fusion network and attention mechan ism,” IEEE Sensors Journal, vol. 23, no. 10, pp. 10 790–10 800, 2023

  61. [69]

    A radial basis deformable residual con- volutional neural model embedded with local multi-modal fe ature knowledge and its application in cross-subject classificat ion,

    J. Li, Y . Zhou, T. Liu, T.-P . Jung, X. Wan, D. Duan, D. Li, H . Y u, H. Song, X. Dong et al. , “A radial basis deformable residual con- volutional neural model embedded with local multi-modal fe ature knowledge and its application in cross-subject classificat ion,” Expert Sys...

  62. [70]

    Canine EEG helps human: Cross- species and cross-modality epileptic seizure detection vi a multi-space alignment,

    Z. Wang, S. Li, and D. Wu, “Canine EEG helps human: Cross- species and cross-modality epileptic seizure detection vi a multi-space alignment,” National Science Review , vol. 12, p. nwaf086, 2025

  63. [71]

    Reducing interdataset covariate shift in sleep EEG of traumatic brain injured humans and mice using transfer Eucl idean alignment,

    M. Vishwanath, S. Cao, N. Dutt, A. M. Rahmani, M. M. Lim, and H. Cao, “Reducing interdataset covariate shift in sleep EEG of traumatic brain injured humans and mice using transfer Eucl idean alignment,” Biomedical Signal Processing and Control , vol. 108, p. 107948, 2025

  64. [72]

    2021 BEETL competition: Advancing transfer learning for s ubject independence and heterogenous EEG data sets,

    X. Wei, A. A. Faisal, M. Grosse-Wentrup, A. Gramfort, S. Chevallier, V . Jayaram, C. Jeunet, S. Bakas, S. Ludwig, K. Barmpas et al. , “2021 BEETL competition: Advancing transfer learning for s ubject independence and heterogenous EEG data sets,” in NeurIPS 2021 Competitions an...

  65. [73]

    An adaptive cross-cl ass transfer learning framework with two-level alignment,

    D.-Q. Xu, Y .-J. Sun, and M.-A. Li, “An adaptive cross-cl ass transfer learning framework with two-level alignment,” Biomedical Signal Processing and Control , vol. 86, p. 105155, 2023

  66. [74]

    A cross-attention-based class alig nment network for cross-subject EEG classification in a heterogeneous spa ce,

    S. Ma and D. Zhang, “A cross-attention-based class alig nment network for cross-subject EEG classification in a heterogeneous spa ce,” Sensors, vol. 24, no. 21, p. 7080, 2024. 14

  67. [75]

    Label alignment improves EEG-based machine learning-bas ed classi- fication of traumatic brain injury,

    M. Vishwanath, N. Dutt, A. M. Rahmani, M. M. Lim, and H. Ca o, “Label alignment improves EEG-based machine learning-bas ed classi- fication of traumatic brain injury,” in Proc. 44th Annual Int’l Conf. of the IEEE Engineering in Medicine & Biology Society (EMBC) , Glasgow, UK, J...

  68. [76]

    Joint EEG feature transfer and semisupervised cross-subj ect emotion recognition,

    Y . Peng, H. Liu, W. Kong, F. Nie, B.-L. Lu, and A. Cichocki , “Joint EEG feature transfer and semisupervised cross-subj ect emotion recognition,” IEEE Trans. on Industrial Informatics , vol. 19, no. 7, pp. 8104–8115, 2023

  69. [77]

    Semi-supervised pairwise transfer learning bas ed on multi-source domain adaptation: A case study on EEG-based e motion recognition,

    C. Ren, J. Chen, R. Li, W. Zheng, Y . Chen, Y . Y ang, X. Zhang , and B. Hu, “Semi-supervised pairwise transfer learning bas ed on multi-source domain adaptation: A case study on EEG-based e motion recognition,” Knowledge-Based Systems , vol. 305, p. 112669, 2024

  70. [78]

    Evoluti onary ensemble learning for EEG-based cross-subject emotion rec ognition,

    H. Zhang, T. Zuo, Z. Chen, X. Wang, and P . Z. Sun, “Evoluti onary ensemble learning for EEG-based cross-subject emotion rec ognition,” IEEE Journal of Biomedical and Health Informatics , vol. 28, no. 7, pp. 3872–3881, 2024

  71. [79]

    Using semi-supervised domain adaptation to enhance EEG-b ased cross-task mental workload classification performance,

    T. Wang, Y . Ke, Y . Huang, F. He, W. Zhong, S. Liu, and D. Min g, “Using semi-supervised domain adaptation to enhance EEG-b ased cross-task mental workload classification performance,” IEEE Journal of Biomedical and Health Informatics , vol. 28, no. 12, pp. 7032–7039, 2024

  72. [80]

    Parallel transpor t on the cone manifold of SPD matrices for domain adaptation,

    O. Y air, M. Ben-Chen, and R. Talmon, “Parallel transpor t on the cone manifold of SPD matrices for domain adaptation,” IEEE Trans. on Signal Processing, vol. 67, no. 7, pp. 1797–1811, 2019

  73. [81]

    Custom domain adaptation: A new method for cross-subject, EEG-based cognitive load re cogni- tion,

    M. Jim´ enez-Guarneros and P . G´ omez-Gil, “Custom domain adaptation: A new method for cross-subject, EEG-based cognitive load re cogni- tion,” IEEE Signal Processing Letters , vol. 27, pp. 750–754, 2020

  74. [82]

    Impr oving transfer performance of deep learning with adaptive batch n ormaliza- tion for brain-computer interfaces,

    L. Xu, Z. Ma, J. Meng, M. Xu, T.-P . Jung, and D. Ming, “Impr oving transfer performance of deep learning with adaptive batch n ormaliza- tion for brain-computer interfaces,” in Proc. 43rd Annual Int’l Conf. of the IEEE Engineering in Medicine & Biology Society (EMBC) , Virtua...

  75. [83]

    A comprehensive survey on transfer learning,

    F. Zhuang, Z. Qi, K. Duan, D. Xi, Y . Zhu, H. Zhu, H. Xiong, a nd Q. He, “A comprehensive survey on transfer learning,” Proc. of the IEEE, vol. 109, no. 1, pp. 43–76, 2021

  76. [84]

    Domain ada ptation via transfer component analysis,

    S. J. Pan, I. W. Tsang, J. T. Kwok, and Q. Y ang, “Domain ada ptation via transfer component analysis,” IEEE Trans. on Neural Networks , vol. 22, no. 2, pp. 199–210, 2011

  77. [85]

    Return of frustratingly easy domain adaptation,

    B. Sun, J. Feng, and K. Saenko, “Return of frustratingly easy domain adaptation,” in Proc. 30th AAAI Conf. on Artificial Intelligence , vol. 6, no. 7, Phoenix, AZ, Feb. 2016, pp. 2058–2065

  78. [86]

    Crowdsourcing seizure detection: Algorithm developmen t and validation on human implanted device recordings,

    S. N. Baldassano, B. H. Brinkmann, H. Ung, T. Blevins, E. C. Conrad, K. Leyde, M. J. Cook, A. N. Khambhati, J. B. Wagenaar, G. A. Wor rell et al. , “Crowdsourcing seizure detection: Algorithm developmen t and validation on human implanted device recordings,” Brain, vol. 140, n...

  79. [87]

    EPILEPSIAE–A Euro pean epilepsy database,

    M. Ihle, H. Feldwisch-Drentrup, C. A. Teixeira, A. Wito n, B. Schel- ter, J. Timmer, and A. Schulze-Bonhage, “EPILEPSIAE–A Euro pean epilepsy database,” Computer Methods and Programs in Biomedicine , vol. 106, no. 3, pp. 127–138, 2012

  80. [88]

    Large brain mode l for learning generic representations with tremendous EEG data in BCI,

    W.-B. Jiang, L.-M. Zhao, and B.-L. Lu, “Large brain mode l for learning generic representations with tremendous EEG data in BCI,” i n Proc. Int’l Conf. on Learning Representations , Vienna, Austria, May 2024

  81. [89]

    Integrating large language model, EEG, and eye- tracking for word-level neural state classification in read ing comprehen- sion,

    Y . Zhang, Q. Li, S. Nahata, T. Jamal, S.-K. Cheng, G. Cauw enberghs, and T.-P . Jung, “Integrating large language model, EEG, and eye- tracking for word-level neural state classification in read ing comprehen- sion,” IEEE Trans. on Neural Systems and Rehabilitation Engineeri n...

  82. [90]

    EEG-based driver drowsiness es timation using feature weighted episodic training,

    Y . Cui, Y . Xu, and D. Wu, “EEG-based driver drowsiness es timation using feature weighted episodic training,” IEEE Trans. on Neural Systems and Rehabilitation Engineering , vol. 27, no. 11, pp. 2263– 2273, 2019

  83. [91]

    EEG-based user reaction time estimation using Riem annian geometry features,

    D. Wu, V . J. Lawhern, B. J. Lance, S. Gordon, T.-P . Jung, a nd C.- T. Lin, “EEG-based user reaction time estimation using Riem annian geometry features,” IEEE Trans. on Neural Systems and Rehabilitation Engineering, vol. 25, no. 11, pp. 2157–2168, 2017

  84. [92]

    Driver drowsiness estimation from EEG signals using onlin e weighted adaptation regularization for regression (OwARR),

    D. Wu, V . J. Lawhern, S. Gordon, B. J. Lance, and C.-T. Lin , “Driver drowsiness estimation from EEG signals using onlin e weighted adaptation regularization for regression (OwARR),” IEEE Trans. on Fuzzy Systems , vol. 25, no. 6, pp. 1522–1535, 2017

  85. [93]

    Spatial fil- tering for EEG-based regression problems in brain-compute r interface (BCI),

    D. Wu, J.-T. King, C.-H. Chuang, C.-T. Lin, and T.-P . Jun g, “Spatial fil- tering for EEG-based regression problems in brain-compute r interface (BCI),” IEEE Trans. on Fuzzy Systems , vol. 26, no. 2, pp. 771–781, 2018

  86. [94]

    On the vulnerability of CNN classifie rs in EEG-based BCIs,

    X. Zhang and D. Wu, “On the vulnerability of CNN classifie rs in EEG-based BCIs,” IEEE Trans. on Neural Systems and Rehabilitation Engineering, vol. 27, no. 5, pp. 814–825, 2019

  87. [95]

    White-box tar get attack for EEG-based BCI regression problems,

    L. Meng, C.-T. Lin, T.-P . Jung, and D. Wu, “White-box tar get attack for EEG-based BCI regression problems,” in Proc. Int’l Conf. on Neural Information Processing, Sydney, Australia, Dec. 2019

  88. [96]

    Adversarial attacks and defenses in physiological c omputing: A systematic review,

    D. Wu, J. Xu, W. Fang, Y . Zhang, L. Y ang, H. Luo, X. Xu, and X. Y u, “Adversarial attacks and defenses in physiological c omputing: A systematic review,” National Science Open , vol. 2, no. 1, p. 20220023, 2023

  89. [97]

    SSVEP-based brain-compute r inter- faces are vulnerable to square wave attacks,

    R. Bian, L. Meng, and D. Wu, “SSVEP-based brain-compute r inter- faces are vulnerable to square wave attacks,” Science China Information Sciences, vol. 65, no. 4, p. 140406, 2022

  90. [98]

    Active poisoning: ef ficient backdoor attacks on transfer learning-based brain-comput er interfaces,

    X. Jiang, L. Meng, S. Li, and D. Wu, “Active poisoning: ef ficient backdoor attacks on transfer learning-based brain-comput er interfaces,” Science China Information Sciences , vol. 66, p. 182402, 2023

  91. [99]

    EEG-based brain-computer inter faces are vulnerable to backdoor attacks,

    L. Meng, J. Huang, Z. Zeng, X. Jiang, S. Y u, T.-P . Jung, C. -T. Lin, R. Chavarriaga, and D. Wu, “EEG-based brain-computer inter faces are vulnerable to backdoor attacks,” IEEE Trans. on Neural Systems and Rehabilitation Engineering , vol. 31, pp. 2224–2234, 2023. [Online]. A...

  92. [100]

    Adv ersarial filtering based evasion and backdoor attacks to EEG-based br ain- computer interfaces,

    L. Meng, X. Jiang, X. Chen, W. Liu, H. Luo, and D. Wu, “Adv ersarial filtering based evasion and backdoor attacks to EEG-based br ain- computer interfaces,” Information Fusion , vol. 107, p. 102316, 2024

  93. [101]

    Tiny noise, big mistakes: Adversarial per turbations induce errors in brain-computer interface spellers,

    X. Zhang, D. Wu, L. Ding, H. Luo, C.-T. Lin, T.-P . Jung, a nd R. Chavarriaga, “Tiny noise, big mistakes: Adversarial per turbations induce errors in brain-computer interface spellers,” National Science Review, vol. 8, no. 4, p. nwaa233, 2021

  94. [102]

    Binnendijk, T

    A. Binnendijk, T. Marler, and E. M. Bartels, Brain-Computer Inter- faces: U.S. Military Applications and Implications, An Ini tial Assess- ment. Santa Monica, CA: RAND Corporation, 2020

  95. [103]

    Adversarial robustness b enchmark for EEG-based brain-computer interfaces,

    L. Meng, X. Jiang, and D. Wu, “Adversarial robustness b enchmark for EEG-based brain-computer interfaces,” Future Generation Computer Systems, vol. 143, pp. 231–247, 2023

  96. [104]

    Alignment-based adversar ial training (ABA T) for improving the robustness and accuracy of EEG-bas ed BCIs,

    X. Chen, Z. Wang, and D. Wu, “Alignment-based adversar ial training (ABA T) for improving the robustness and accuracy of EEG-bas ed BCIs,” IEEE Trans. on Neural Systems and Rehabilitation Engineeri ng, vol. 32, pp. 1703–1714, 2024

  97. [105]

    Data alignment based advers arial defense benchmark for EEG-based BCIs,

    X. Chen, T. Jia, and D. Wu, “Data alignment based advers arial defense benchmark for EEG-based BCIs,” Neural Networks , 2025, in press

  98. [106]

    Privacy-preserving brain-com puter interfaces: A systematic review,

    K. Xia, W. Duch, Y . Sun, K. Xu, W. Fang, H. Luo, Y . Zhang, D . Sang, X. Xu, F.-Y . Wang, and D. Wu, “Privacy-preserving brain-com puter interfaces: A systematic review,” IEEE Trans. on Computational Social Systems, vol. 10, no. 5, pp. 2312–2324, 2023

  99. [107]

    General Data Protection Regulation, “Regulation (EU ) 2016/679 of the European parliament and of the council of 27 April 2016 on the protection of natural persons with regard to the processing of personal data and on the free movement of such data, and repealing dire ctive 95/...

  100. [108]

    Protecting privacy of us ers in brain-computer interface applications,

    A. Agarwal, R. Dowsley, N. D. McKinney, D. Wu, C.-T. Lin , M. De Cock, and A. C. A. Nascimento, “Protecting privacy of us ers in brain-computer interface applications,” IEEE Trans. on Neural Systems and Rehabilitation Engineering , vol. 27, no. 8, pp. 1546–1555, 2019

  101. [109]

    A privacy-pre serving generative adversarial network method for securing EEG bra in signals,

    E. Debie, N. Moustafa, and M. T. Whitty, “A privacy-pre serving generative adversarial network method for securing EEG bra in signals,” in Proc. Int’l Joint Conf. on Neural Networks , Glasgow, UK, Jul. 2020, pp. 1–8

  102. [110]

    EpilepsyGAN: Synthetic epileptic brain a ctivities with privacy preservation,

    D. Pascual, A. Amirshahi, A. Aminifar, D. Atienza, P . R yvlin, and R. Wattenhofer, “EpilepsyGAN: Synthetic epileptic brain a ctivities with privacy preservation,” IEEE Trans. on Biomedical Engineering , vol. 68, no. 8, pp. 2435–2446, 2021

  103. [111]

    Privacy preserving classification of EEG data using machine learning and homomorphic encryption,

    A. B. Popescu, L. A. Taca, C. I. Nita, A. Vizitiu, R. Deme ter, C. Suciu, and L. M. Itu, “Privacy preserving classification of EEG data using machine learning and homomorphic encryption,” Applied Sciences , vol. 11, no. 16, p. 7360, 2021

  104. [112]

    Multi-source decentrali zed transfer for privacy-preserving BCIs,

    W. Zhang, Z. Wang, and D. Wu, “Multi-source decentrali zed transfer for privacy-preserving BCIs,” IEEE Trans. on Neural Systems and Rehabilitation Engineering , vol. 30, pp. 2710–2720, 2022

  105. [113]

    Source-free domain adapta tion (SFDA) for privacy-preserving seizure subtype classification,

    C. Zhao, R. Peng, and D. Wu, “Source-free domain adapta tion (SFDA) for privacy-preserving seizure subtype classification,” IEEE Trans. on Neural Systems and Rehabilitation Engineering , vol. 31, pp. 2315– 2325, 2023

  106. [114]

    Lightweight source free transfer f or privacy- preserving unsupervised motor imagery classification,

    W. Zhang and D. Wu, “Lightweight source free transfer f or privacy- preserving unsupervised motor imagery classification,” IEEE Trans. on Cognitive and Developmental Systems , vol. 15, no. 2, pp. 938–949, 2023. 15

  107. [115]

    Federated moto r imagery classification for privacy-preserving brain-computer int erfaces,

    T. Jia, L. Meng, S. Li, J. Liu, and D. Wu, “Federated moto r imagery classification for privacy-preserving brain-computer int erfaces,” IEEE Trans. on Neural Systems and Rehabilitation Engineering , vol. 32, pp. 3442–3451, 2024

  108. [116]

    Mac hine unlearning for seizure prediction,

    C. Shao, C. Li, R. Song, X. Liu, R. Qian, and X. Chen, “Mac hine unlearning for seizure prediction,” IEEE Trans. on Cognitive and Developmental Systems , vol. 16, no. 6, pp. 1969–1981, 2024

  109. [117]

    Protecting multip le types of privacy simultaneously in EEG-based brain-computer inter faces,

    L. Meng, X. Jiang, T. Jia, and D. Wu, “Protecting multip le types of privacy simultaneously in EEG-based brain-computer inter faces,” in Proc. IEEE Int’l Conf. on Systems, Man and Cybernetics , Kuching, Malaysia, Oct. 2024

  110. [118]

    User-wise pert urbations for user identity protection in EEG-based BCIs,

    X. Chen, S. Li, Y . Tu, Z. Wang, and D. Wu, “User-wise pert urbations for user identity protection in EEG-based BCIs,” Jouranl of Neural Engineering, vol. 22, no. 1, p. 016040, 2025

  111. [119]

    A3E: Aligned and augmented a dversarial ensemble for accurate, robust and privacy-preserving EEG d ecoding,

    X. Chen, T. Jia, and D. Wu, “A3E: Aligned and augmented a dversarial ensemble for accurate, robust and privacy-preserving EEG d ecoding,” https://doi.org/10.48550/arXiv.2412.11390, 2024

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.