Transfer Learning for Brain–Computer Interfaces: A Euclidean Space Data Alignment Approach
Open accessThe DOI matches its Crossref record; 50 of the 52 references it cites were found in a registry or an index; 9 of the 30 sources its claims cite were checked for support.
Abstract
OBJECTIVE: This paper targets a major challenge in developing practical electroencephalogram (EEG)-based brain-computer interfaces (BCIs): how to cope with individual differences so that better learning performance can be obtained for a new subject, with minimum or even no subject-specific data? METHODS: We propose a novel approach to align EEG trials from different subjects in the Euclidean space to make them more similar, and hence improve the learning performance for a new subject. Our approach has three desirable properties: first, it aligns the EEG trials directly in the Euclidean space, and any signal processing, feature extraction, and machine learning algorithms can then be applied to the aligned trials; second, its computational cost is very low; and third, it is unsupervised and does not need any label information from the new subject. RESULTS: Both offline and simulated online experiments on motor imagery classification and event-related potential classification verified that our proposed approach outperformed a state-of-the-art Riemannian space data alignment approach, and several approaches without data alignment. CONCLUSION: The proposed Euclidean space EEG data alignment approach can greatly facilitate transfer learning in BCIs. SIGNIFICANCE: Our proposed approach is effective, efficient, and easy to implement. It could be an essential pre-processing step for EEG-based BCIs.
Overview
CoChat Summary
TL;DR
The paper proposes Euclidean-space alignment (EA), an unsupervised transfer-learning preprocessing step that aligns EEG trials from different subjects in Euclidean space to reduce individual differences for BCI classification. Across offline and simulated online experiments on motor imagery and ERP datasets, EA generally outperformed Riemannian-space alignment and no-alignment baselines while being much faster.
Key findings
- In offline unsupervised MI classification, EA-CSP-LDA achieved average accuracies of 79.79% on MI Dataset 1 and 73.53% on MI Dataset 2a, versus RA-MDRM at 68.07% and 70.91%, CSP-LDA at 59.71% and 67.75%, and MDRM at 54.36% and 57.18%.
- In offline ERP classification, EA-xDAWN-SVM achieved an average balanced classification accuracy (BCA) of 68.80%, compared with EA-SVM 67.85%, SVM 64.64%, and xDAWN-SVM 64.60%; EA-SVM significantly outperformed SVM (p=0.0386) and EA-xDAWN-SVM significantly outperformed xDAWN-SVM (p=0.0449).
- In simulated online MI classification, EA-CSP-LDA outperformed RA-MDRM on 12 out of 16 subjects; in simulated online ERP classification, EA-xDAWN-SVM outperformed RA-MDRM on all 11 subjects.
- EA-CSP-LDA was 3.6–19.5 times faster than RA-MDRM; for example, on MI Dataset 1 it took 0.3864 s versus 7.5326 s, and on MI Dataset 2a it took 0.2405 s versus 0.8766 s.
Methods
- The proposed EA aligns time-domain EEG trials in Euclidean space by using the arithmetic mean of all covariance matrices from a subject as the reference matrix, then transforming trials so that the mean covariance matrix of each subject becomes the identity matrix; it requires no labels from the new subject.
- Experiments used two MI datasets from BCI Competition IV (Dataset 1: 7 subjects, 59-channel EEG, 100 trials per class in calibration; Dataset 2a: 9 subjects, 22-channel EEG plus 3-channel EOG, 72 trials per class for left- and right-hand MI) and one ERP RSVP dataset from PhysioNet (11 subjects, 8 channels, 5 Hz Session a, target-to-non-target ratio around 1:9).
- Evaluation included offline unsupervised leave-one-subject-out cross-validation and simulated online supervised classification, with CSP-LDA, MDRM, SVM, and xDAWN-SVM pipelines; paired-sample t-tests and area under the curve (AUC) were used for statistical comparisons.
Limitations
- The paper states that EA only compensates for covariate shift and ignores prior probability shift and concept shift; it may even increase concept shift, and bad trials or outliers can corrupt the reference matrix. It also states that the simulated online experiments are not identical to real online experiments.
- The text makes evident that EA degraded accuracy for some subjects (e.g., Subjects 4 and 9 in MI Dataset 2a, and some ERP subjects), and that RA-MDRM could not be applied to offline unsupervised ERP classification because it requires labeled non-target and target trials from the new subject.
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
4 Valid · 8 Not valid · 15 Not checked · 89 Own results and interpretations
Every source the paper cites for this claim was checked and supports it.
In the paper
“So Barachant and Congedo [2] proposed a novel approach to augment the ERP trials so that their covariance matrices can also encode the temporal information.”
What the sources say
- Supportedcites [2] [2] A. Barachant and M. Congedo, 'A plug & play P300 BCI using information geometry,' arXiv: 1409.0107 , 2014. - source
Section B (pp. 3–4) describes a prototyped ERP response obtained by averaging single-trial responses, then concatenating it with each trial to form a super trial. The resulting covariance matrices include a cross-covariance block that carries information about the ERP waveform’s temporal structure. (LLM, conf=0.94)
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
Every source the paper cites for this claim was checked and supports it.
In the paper
“However, the main discriminative information of ERP trials is carried temporally rather than spatially.”
What the sources say
- Supportedcites [2] [2] A. Barachant and M. Congedo, 'A plug & play P300 BCI using information geometry,' arXiv: 1409.0107 , 2014. - source
Section B (p. 3) says ERP discriminative information lies in the potential’s waveform rather than the spatial distribution of its variance. It also says ignoring temporal structure loses information needed for classification, supporting the paraphrase. (LLM, conf=0.93)
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Our proposed EA-CSP-LDA was 3.6-19.5 times faster than RA-MDRM, and also it had much smaller standard deviation.”
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“On average RI-MDRM outperformed RR-MDRM, and EI-CSP-LDA outperformed ER-CSP-LDA, on both datasets, suggesting that estimating the reference matrix from all imagery trials would be better than using all resting trials.”
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Our proposed EA-CSP-LDA was 3.6-19.5 times faster than RA-MDRM, and also it had much smaller standard deviation.”
plainstates an association- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
47 more claims
Create a free account to see all 116 claims and to verify the 4 references still open.
- It may state two separate claims as one. 47
- Probably not a substantial claim. 30
- The quote only partly states this claim. 1
Citations
Cites
LIBSVM: A library for support vector machines
Chang, C.-C., & Lin, C.-J. (2011) · ACM Transactions on Intelligent Systems and Technology · doi:10.1145/1961189.1961199 · Cited by 40,918 · cited 1×
“LibSVM [6] with a linear kernel was used for classification.”
Method(2008) · Journal of Machine Learning Research · Cited by 33,051
Delorme, A., & Makeig, S. (2004) · Journal of Neuroscience Methods · doi:10.1016/j.jneumeth.2003.10.009 · Cited by 24,371
Pan, S. J., & Yang, Q. (2010) · IEEE Transactions on Knowledge and Data Engineering · doi:10.1109/tkde.2009.191 · Cited by 23,387 · cited 2×
“This idea has been widely used in TL [24], [31], [40].”
BackgroundMethodBook Review: Pattern Recognition and Machine Learning
Nasrabadi, N. M. (2007) · Journal of Electronic Imaging · doi:10.1117/1.2819119 · Cited by 22,066
A Simple Sequentially Rejective Multiple Test Procedure
(1979) · Scandinavian Journal of Statistics · doi:10.2307/4615733 · Cited by 21,526 · cited 1×
“Hence, we performed Holm-Bonferroni corrections [12] on the three p-values of each dataset.”
MethodGoldberger, A. L., Amaral, L. A. N., Glass, L., Hausdorff, J. M., Ivanov, P. C., Mark, R. G., Mietus, J. E., Moody, G. B., Peng, C.-K., & Stanley, H. E. (2000) · Circulation · doi:10.1161/01.cir.101.23.e215 · Cited by 14,809
Pattern Recognition and Machine Learning
(2006) · doi:10.1007/978-0-387-45528-0 · Cited by 9,824 · cited 1×
“Usually the EEG signals are first band-pass filtered and spatially filtered to increase the signal-to-noise ratio, and then discriminative features are extracted, which are next fed into machine learning algorithms such as Linear Discriminant Analysis (LDA) and Support Vector Machine (SVM) [3] for classification.”
MethodBrain–computer interfaces for communication and control
Wolpaw, J. R., Birbaumer, N., McFarland, D. J., Pfurtscheller, G., & Vaughan, T. M. (2002) · Clinical Neurophysiology · doi:10.1016/s1388-2457(02)00057-3 · Cited by 7,927
On the Kolmogorov-Smirnov Test for Normality with Mean and Variance Unknown
Lilliefors, H. W. (1967) · Journal of the American Statistical Association · doi:10.2307/2283970 · Cited by 4,530
Cited by
Items 101–200 of 520
Wang, Y., Cui, W., Yu, T., et al. (2023) · IEEE Transactions on Neural Systems and Rehabilitation Engineering · doi:10.1109/tnsre.2023.3321414 · Cited by 25
“Although these feature fusion methods obtained a comprehensive feature, they only considered the general fusion of low-level features in Euclidean space [21].”
MethodMotor Imagery Classification for Asynchronous EEG-Based Brain–Computer Interfaces
Wu, H., Li, S., & Wu, D. (2024) · IEEE Transactions on Neural Systems and Rehabilitation Engineering · doi:10.1109/tnsre.2024.3356916 · Cited by 34
“For asynchronous BCIs, data alignment [25], source-free domain adaptation [26], and domain generalization [27] approaches may be used to further improve performance and protect user privacy.”
MethodSelf-citationTang, X., Li, X., Li, W., Hao, B., Xie, Y., & Dang, X. (2021) · IEEE Transactions on Cognitive and Developmental Systems · doi:10.1109/tcds.2021.3082648 · Cited by 26
Wu, M., Hu, S., Wei, B., & Lv, Z. (2022) · Journal of Neuroscience Methods · doi:10.1016/j.jneumeth.2022.109642 · Cited by 29
Nagarajan, A., Robinson, N., Ang, K. K., Chua, K. S. G., Chew, E., & Guan, C. (2023) · Journal of Neural Engineering · doi:10.1088/1741-2552/ad152f · Cited by 32
“Finally, another transfer learning method for MI-BCI based on the Riemannian framework was introduced by He and Wu [ 50].”
MethodQuantifying Spatial Activation Patterns of Motor Units in Finger Extensor Muscles
Jiang, X., Ren, H., Xu, K., Ye, X., Dai, C., Clancy, E. A., Zhang, Y., & Chen, W. (2020) · IEEE Journal of Biomedical and Health Informatics · doi:10.1109/jbhi.2020.3002329 · Cited by 23
“So far, a variety of modalities have been employed to develop HMI techniques, such as sEMG [39]–[41], EEG [30], [31],”
MethodPriming cross-session motor imagery classification with a universal deep domain adaptation framework
Zhang, X., Miao, Z., Menon, C., Zheng, Y., Zhao, M., & Ming, D. (2023) · Neurocomputing · doi:10.1016/j.neucom.2023.126659 · Cited by 31
“In this paper, a 200-order Blackman window bandpass FIR filter of [4-38] Hz was used to filter the EEG data.”
BackgroundMethodInfluentialWei, F., Xu, X., Li, X., & Wu, X. (2024) · IEEE Transactions on Industrial Informatics · doi:10.1109/tii.2024.3450010 · Cited by 24
Time–frequency transform based EEG data augmentation for brain–computer interfaces
Wang, Z., Li, S., Chen, X., & Wu, D. (2025) · Knowledge-Based Systems · doi:10.1016/j.knosys.2025.113074 · Cited by 29
Self-citationMotor Imagery EEG Classification Based on Transfer Learning and Multi-Scale Convolution Network
Chang, Z., Zhang, C., & Li, C. (2022) · Micromachines · doi:10.3390/mi13060927 · Cited by 24
“To solve the cross-subject problem of transfer learning, in [24,25], the data in Riemannian space and European space, respectively, were aligned and transformed, which resulted in similar data distribution of different subjects, significantly improving the problem of transfer learning.”
An Overview of Brain Fingerprint Identification Based on Various Neuroimaging Technologies
Zhang, S., Yang, W., Mou, H., Pei, Z., Li, F., & Wu, X. (2023) · IEEE Transactions on Cognitive and Developmental Systems · doi:10.1109/tcds.2023.3314155 · Cited by 29
Dual selections based knowledge transfer learning for cross-subject motor imagery EEG classification
Luo, T.-J. (2023) · Frontiers in Neuroscience · doi:10.3389/fnins.2023.1274320 · Cited by 25
Domain Generalization in Biosignal Classification
Dissanayake, T., Fernando, T., Denman, S., Ghaemmaghami, H., Sridharan, S., & Fookes, C. (2020) · IEEE Transactions on Biomedical Engineering · doi:10.1109/tbme.2020.3045720 · Cited by 25
“Examining the literature, we find that researchers in the medical field have been able to develop successful techniques for domain adaptation [2]–[4]; however there has not been any significant effort in applying domain generalization in the medical field, and the problem remains open and requires further attention.”
BackgroundChen, X., Wang, Z., & Wu, D. (2024) · IEEE Transactions on Neural Systems and Rehabilitation Engineering · doi:10.1109/tnsre.2024.3391936 · Cited by 31
“Multiple EEG data aliment approaches have been proposed, e.g., Riemannian alignment [27], EA [28] and label alignment [35].”
BackgroundMethodSelf-citationLi, Y., Wei, Q., Chen, Y., & Zhou, X. (2021) · IEEE Access · doi:10.1109/access.2020.3048683 · Cited by 24
“The reference matrix of a subject is strictly equivalent to an identity matrix [24].”
BackgroundMethodInfluentialSpatial filtering based on Riemannian distance to improve the generalization of ErrP classification
Cruz, A., Pires, G., & Nunes, U. (2021) · Neurocomputing · doi:10.1016/j.neucom.2021.10.078 · Cited by 22
Liang, Z., Zheng, Z., Chen, W., Pei, Z., Wang, J., & Chen, J. (2023) · Biomedical Signal Processing and Control · doi:10.1016/j.bspc.2023.105556 · Cited by 21
Giles, J., Ang, K. K., Mihaylova, L. S., & Arvaneh, M. (2019) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp.2019.8683331 · Cited by 26
A Cascade xDAWN EEGNet Structure for Unified Visual-Evoked Related Potential Detection
Wang, H., Wang, Z., Sun, Y., Yuan, Z., Xu, T., & Li, J. (2024) · IEEE Transactions on Neural Systems and Rehabilitation Engineering · doi:10.1109/tnsre.2024.3415474 · Cited by 20
“Moreover, Wu et al. achieved remarkable progress by aligning ERP data with Euclidean alignment and enriching features with xDAWN and tangent space mapping which secured the top spot in the RSVP detection competition at the World Robot Contest 2021 [32], [33].”
BackgroundChen, L., Yu, Z., & Yang, J. (2022) · Frontiers in Neurorobotics · doi:10.3389/fnbot.2022.958052 · Cited by 21
“He and Wu (2019) propose a Euclidean Space Data Alignment Approach to align the time-domain EEG trials in the Euclidean Space and alleviate the domain shift between different sessions and subjects successfully.”
BackgroundIntra- and Inter-Subject Common Spatial Pattern for Reducing Calibration Effort in MI-Based BCI
Wei, Q., & Ding, X. (2023) · IEEE Transactions on Neural Systems and Rehabilitation Engineering · doi:10.1109/tnsre.2023.3236372 · Cited by 23
“He and Wu [23] presented a Euclidean alignment (EA) method, in which EEG trials from different subjects are aligned in Euclidean space to make them more similar and any Euclidean space classifier can be used after EA.”
BackgroundMethodCross-Subject EEG Channel Selection Method for Lower Limb Brain-Computer Interface
Wei, M., Huang, M., & Ni, J. (2023) · International Journal of Network Dynamics and Intelligence · doi:10.53941/ijndi.2023.100008 · Cited by 21
MSFNet: A Multi-Scale Space-Time Frequency Fusion Network for Motor Imagery EEG Classification
Wang, C., Wu, Y., Wang, C., Ren, Y., Shen, J., Pang, T., Chan, C. S., Ren, W., & Yu, Y. (2024) · IEEE Access · doi:10.1109/access.2024.3351204 · Cited by 23
“In 2020, He and Wu proposed EA-CSP-LDA for MI-EEG classification and achieved an average accuracy of 73.53% on MI Dataset 2a [24].”
MethodInfluentialClass-specific regularized joint distribution alignment for unsupervised domain adaptation
Luo, T.-J. (2024) · Engineering Applications of Artificial Intelligence · doi:10.1016/j.engappai.2024.107877 · Cited by 22
Chen, Z., Yang, R., Huang, M., Wang, Z., & Liu, X. (2024) · IEEE Transactions on Emerging Topics in Computational Intelligence · doi:10.1109/tetci.2024.3359097 · Cited by 32
Wang, Z., Li, S., & Wu, D. (2025) · National Science Review · doi:10.1093/nsr/nwaf086 · Cited by 22
Self-citationZhao, Y., Dai, G., Borghini, G., Zhang, J., Li, X., Zhang, Z., Aricò, P., Flumeri, G. D., Babiloni, F., & Zeng, H. (2021) · Frontiers in Human Neuroscience · doi:10.3389/fnhum.2021.706270 · Cited by 21
“He and Wu (2019) proposed Euclidean space EEG data alignment method to minimize the distance between the mean covariance matrices in different domains by transforming and aligning the EEG data in the Euclidean space.”
MethodGiles, J., Ang, K. K., Phua, K. S., & Arvaneh, M. (2022) · Frontiers in Neuroergonomics · doi:10.3389/fnrgo.2022.837307 · Cited by 20
“…learning algorithms applied alignment of the EEG distributions between the source and target sessions (Sugiyama et al., 2008; Dalhoumi et al., 2015; Raza et al., 2016; He and Wu, 2018) or weighted the source sessions according to their similarities with the target session (Azab et al., 2019b).”
MethodTangent space alignment: Transfer learning for Brain-Computer Interface
Bleuzé, A., Mattout, J., & Congedo, M. (2022) · Frontiers in Human Neuroscience · doi:10.3389/fnhum.2022.1049985 · Cited by 35
Time–space–frequency feature Fusion for 3-channel motor imagery classification
Miao, Z., & Zhao, M. (2023) · Biomedical Signal Processing and Control · doi:10.1016/j.bspc.2023.105867 · Cited by 23
Luo, J., Cui, W., Xu, S., Wang, L., Chen, H., & Li, Y. (2024) · IEEE Transactions on Neural Systems and Rehabilitation Engineering · doi:10.1109/tnsre.2024.3359191 · Cited by 22
“For the PhysioNet dataset, the proposed CST-TVA-DRTL method significantly outperforms the xDAWN-SVM [18] approach, highlighting the weaker feature transfer capabilities of machine learning methods.”
MethodLuo, T.-J. (2024) · Expert Systems with Applications · doi:10.1016/j.eswa.2024.124673 · Cited by 21
Yang, K., Li, R., Xu, J., Zhu, L., Kong, W., & Zhang, J. (2024) · IEEE Journal of Biomedical and Health Informatics · doi:10.1109/jbhi.2024.3395910 · Cited by 18
“The RG features of this frequency range have been used wildly in classification [41], [42].”
MethodQiu, Y., Liu, H., & Zhao, M. (2025) · Applied Sciences · doi:10.3390/app15010392 · Cited by 30
Meta-Learning for Fast and Privacy-Preserving Source Knowledge Transfer of EEG-Based BCIs
Li, S., Wu, H., Ding, L., & Wu, D. (2022) · IEEE Computational Intelligence Magazine · doi:10.1109/mci.2022.3199622 · Cited by 21
“, Riemannian alignment [35], Euclidean alignment [36], label alignment [37], centroid alignment [38], adversarial training [39], [40], etc.”
BackgroundSelf-citationFacilitating applications of SSVEP-BCI by effective Cross-Subject knowledge transfer
Li, H., Xu, G., Du, C., Li, Z., Han, C., Tian, P., Li, B., & Zhang, S. (2024) · Expert Systems with Applications · doi:10.1016/j.eswa.2024.123492 · Cited by 20
An improved version of local activities estimation to enhance motor imagery classification
Togha, M. M., Salehi, M. R., & Abiri, E. (2021) · Biomedical Signal Processing and Control · doi:10.1016/j.bspc.2021.102485 · Cited by 18
Transfer Learning for EEG-Based Brain-Computer Interfaces: A Review of Progress Made Since 2016Preprint
Wu, D., Xu, Y., & Lu, B. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2004.06286 · Cited by 17
“For example, EA and LA data alignment schemes have been combined with TL classifiers [50], [60], and CSDF-ReliefF and CSDF-mRMR feature selection approaches have also been integrated with TL classifiers [114].”
Self-citationTang, J., Li, X., Yang, Y. W., & Zhang, W. (2021) · Electronics Letters · doi:10.1049/ell2.12275 · Cited by 16
Xiong, W. H., & Wei, Q. (2022) · PLoS ONE · doi:10.1371/journal.pone.0263641 · Cited by 16
“Euclidean alignment (EA) [36] is adopted in the study because it can be used as either a supervised or a unsupervised method.”
BackgroundMethodZhang, H., Zhao, Z., Liu, C., Duan, M., Lü, Z., & Wang, H. (2024) · Journal of Applied Biomedicine · doi:10.1016/j.bbe.2024.11.003 · Cited by 18
Adaptive deep feature representation learning for cross-subject EEG decoding
Liang, S., Li, L., Zu, W., Feng, W., & Hang, W. (2024) · BMC Bioinformatics · doi:10.1186/s12859-024-06024-w · Cited by 22
“These methods have demonstrated the capability to learn the domain-invariant features without leveraging the label of target EEG data [28, 29].”
Revisiting Euclidean alignment for transfer learning in EEG-based brain–computer interfaces
Wu, D. (2025) · Journal of Neural Engineering · doi:10.1088/1741-2552/addd49 · Cited by 26
“Most materials in this subsection are adapted from [13]. i.e., ¯ R is the arithmetic mean of all N covariance matrices from that domain.”
InfluentialSelf-citationWang, X., Dai, X., Liu, Y., Chen, X., Hu, Q., Hu, R., & Li, M. (2023) · Frontiers in Human Neuroscience · doi:10.3389/fnhum.2023.1175399 · Cited by 18
“For example, He and Wu (2019) proposed a weighted logistic regression transfer learning algorithm based on Euclidean Space (EA), which aligns the EEG data of the source domain and target domain in the data preprocessing stage to reduce the differences between signals.”
MethodOnline Privacy-Preserving EEG Classification by Source-Free Transfer Learning
Wu, H., Ma, Z., Guo, Z., Wu, Y., Zhang, J., Zhou, G., & Long, J. (2024) · IEEE Transactions on Neural Systems and Rehabilitation Engineering · doi:10.1109/tnsre.2024.3445115 · Cited by 19
“We analyze the influence of Euclidean Alignment (EA) [63] on OSFTL and present the average results in Table VII.”
MethodKilani, S., Aghili, S. N., & Hulea, M. G. (2023) · Applied Sciences · doi:10.3390/app13106283 · Cited by 17
Ahmadi, H., & Mesin, L. (2024) · Electronics · doi:10.3390/electronics13061033 · Cited by 14
Rácz, F. S., Kumar, S., Kaposzta, Z., Alawieh, H., Liu, D. H., Liu, R., Czoch, A., Mukli, P., & Millán, J. D. R. (2024) · Frontiers in Neuroscience · doi:10.3389/fnins.2024.1271831 · Cited by 16
“Specifically, in each pipeline a global reference matrix R was obtained as the Karcher mean of all matrices C i regardless of class or session (He and Wu, 2019) that was then used to re-center C i in the entire dataset according to Subsequently, training and testing (including CSP estimation) were…”
MethodA systematic evaluation of Euclidean alignment with deep learning for EEG decoding
Junqueira, B. R. R. S. P., Aristimunha, B., Chevallier, S., & Camargo, R. Y. D. (2024) · Journal of Neural Engineering · doi:10.1088/1741-2552/ad4f18 · Cited by 29
Cui, J., Jin, X., Hu, H., Zhu, L., Ozawa, K., Pan, G., & Kong, W. (2021) · IEEE Transactions on Cognitive and Developmental Systems · doi:10.1109/tcds.2021.3137530 · Cited by 17
“In a recent EEG-based application, the Euclidean space data alignment method [31] maps each subject’s EEG signals to a new space where the differences in the “mean covariance matrix” are minimized, thereby implicitly reducing the differences between the source and target domains.”
MethodAbdel-Ghaffar, E. A., Wu, Y., & Daoudi, M. (2022) · IEEE Access · doi:10.1109/access.2022.3147461 · Cited by 14
“He et al. [25] aligned the EEG trials in the Euclidean space which makes them more similar and enhances the learning performance for new participants, feature extraction and classification processes are performed on the aligned data.”
BackgroundCross-subject aesthetic preference recognition of Chinese dance posture using EEG
Li, J., Wu, S.-R., Zhang, X., Luo, T.-J., Li, R., Zhao, Y., Liu, B., & Peng, H. (2022) · Cognitive Neurodynamics · doi:10.1007/s11571-022-09821-2 · Cited by 18
Multi-Source geometric metric transfer learning for EEG classification
Zhang, X., She, Q., Tan, T., Gao, Y., Ma, Y., & Zhang, J. (2022) · Biomedical Signal Processing and Control · doi:10.1016/j.bspc.2022.104435 · Cited by 15
Cao, L., Chen, S., Jia, J., Fan, C., Wang, H., & Xu, Z. (2021) · Frontiers in Neuroscience · doi:10.3389/fnins.2020.629572 · Cited by 17
“Therefore, our classification result was lower than those of the above state-of-art BCI systems (Azab et al., 2019, 2020; He and Wu, 2019).”
ResultDBConformer: Dual-Branch Convolutional Transformer for EEG Decoding
Wang, Z., Wang, H., Jia, T., He, X., Li, S., & Wu, D. (2025) · IEEE Journal of Biomedical and Health Informatics · doi:10.1109/jbhi.2025.3622725 · Cited by 39
“We apply Euclidean alignment (EA) [49] to align the original trials from each subject in the Euclidean space.”
Self-citationDynamic Weighted Filter Bank Domain Adaptation for Motor Imagery Brain–Computer Interfaces
Zhang, Y., Qiu, S., Wei, W., Ma, X., & He, H. (2022) · IEEE Transactions on Cognitive and Developmental Systems · doi:10.1109/tcds.2022.3209801 · Cited by 14
“In 2019, He and Wu [40] proposed an unsupervised Euclidean space data alignment”
BackgroundSelf-citationTensor decomposition-based channel selection for motor imagery-based brain-computer interfaces
Huang, Z., & Wei, Q. (2023) · Cognitive Neurodynamics · doi:10.1007/s11571-023-09940-4 · Cited by 15
“Among them, SVM (Vapnik 2000) is one of the mostly used classifiers for MI-based BCIs (Jin et al. 2019, 2020; Lal et al. 2004; Arvaneh et al. 2011; He and Wu 2020; Azab et al. 2019).”
Interacting Federated and Transfer Learning-Aided CSI Prediction for Intelligent Cellular Networks
Sun, J., Zhang, Y., Gui, G., Zhao, H., Gacanin, H., & Sari, H. (2023) · IEEE Transactions on Vehicular Technology · doi:10.1109/tvt.2023.3290660 · Cited by 17
“Other frameworks [47], [48], [49] based on transfer learning were proposed in domains such as remote sensing, fault diagnosis, and brain-computer interfaces.”
BackgroundUnsupervised domain adaptation for cross-patient seizure classification
Wang, Z., Zhang, W., Li, S., Chen, X., & Wu, D. (2023) · Journal of Neural Engineering · doi:10.1088/1741-2552/ad0859 · Cited by 20
“It can be performed on the EEG covariance matrices in the Riemannian space [18], or on the raw EEG signals in the Euclidean space [19].”
Self-citationCross-Subject Motor Imagery Decoding by Transfer Learning of Tactile ERD
Zhong, Y., Yao, L., Pan, G., & Wang, Y. (2024) · IEEE Transactions on Neural Systems and Rehabilitation Engineering · doi:10.1109/tnsre.2024.3358491 · Cited by 18
Active poisoning: efficient backdoor attacks on transfer learning-based brain-computer interfaces
Jiang, X., Meng, L., Li, S., & Wu, D. (2023) · Science China Information Sciences · doi:10.1007/s11432-022-3548-2 · Cited by 17
“The following four datasets were used in this study, as in [18,20], including two datasets (P300 and ERN) on event-related potentials and two MI datasets in BCI Competition IV 3) .”
MethodSelf-citationRobust Collaborative Clustering of Subjects and Radiomic Features for Cancer Prognosis
Liu, H., Li, H., Habes, M., Li, Y., Boimel, P. J., Janopaul-Naylor, J. R., Xiao, Y., Ben‐Josef, E., & Yong Fan, A. (2020) · IEEE Transactions on Biomedical Engineering · doi:10.1109/tbme.2020.2969839 · Cited by 13
Khademi, S., Neghabi, M., Farahi, M., Shirzadi, M., & Marateb, H. R. (2022) · Elsevier eBooks · doi:10.1016/b978-0-323-91197-9.00004-7 · Cited by 16
Luo, T.-J. (2022) · IEEE International Conference on Bioinformatics and Biomedicine (BIBM) · doi:10.1109/bibm55620.2022.9995282 · Cited by 18
“During the classification of cross-subject MI-EEG, although the CMCA method is capable of aligning sample distribution across subjects, the recently EA-CSP-LDA [7]/ SB-TA-CSP [9] method only exerts the CSP feature.”
MethodInfluentialChu, C., Zhu, L., Huang, A., Xu, P., Ying, N., & Zhang, J. (2024) · Journal of Neural Engineering · doi:10.1088/1741-2552/ad1f7a · Cited by 12
“In contrast, He and Wu [7] proposed the Euclidean Alignment (EA) approach, which overcomes many of the drawbacks of RA.”
Multi-source transfer learning via optimal transport feature ranking for EEG classification
Li, J., She, Q., Fang, F. Z., Chen, Y., & Zhang, Y. (2024) · Neurocomputing · doi:10.1016/j.neucom.2024.127944 · Cited by 13
Jin, J., Bai, G., Xu, R., Qin, K., Sun, H., Wang, X., & Cichocki, A. S. (2024) · Journal of Neural Engineering · doi:10.1088/1741-2552/ad593b · Cited by 16
“To address this problem, He and Wu [23] designed a new Euclidean alignment (EA) approach, which transformed and aligned the EEG trials in the Euclidean space for signal processing, feature extraction, and classification.”
BackgroundMethodInfluentialA Riemannian Convolutional Neural Network for EEG-based motor imagery decoding
Li, C., & Gu, Z. (2025) · Neurocomputing · doi:10.1016/j.neucom.2025.130206 · Cited by 12
Riemannian geometry-based transfer learning for reducing training time in c-VEP BCIs
Ying, J., Wei, Q., & Zhou, X. (2022) · Scientific Reports · doi:10.1038/s41598-022-14026-y · Cited by 18
“Several DA approaches exist in literature such as Riemannian alignment (RA) 47 , Euclidean alignment (EA) 48 and online pre-alignment (OPS) 49 , depending upon how the reference matrix is defined.”
BackgroundReal-time markerless video tracking of body parts in mice using deep neural networksPreprint
Forys, B. J., Xiao, D., Gupta, P. K., Boyd, J. D., & Murphy, T. H. (2018) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.1101/482349 · Cited by 12
Foundation models for EEG decoding: current progress and prospective research
Yuxuan, Y., Wang, H., Li, C., Yiheng, P., & Luo, J. (2025) · Journal of Neural Engineering · doi:10.1088/1741-2552/ae17e9 · Cited by 23
“Finally, EA normalization [113] was applied to each session to standardize the data, enhancing stability and processing performance.”
Multi-source online transfer algorithm based on source domain selection for EEG classification
Wu, Z., She, Q., Hou, Z., Li, Z., Tian, K., & Ma, Y. (2022) · Mathematical Biosciences & Engineering · doi:10.3934/mbe.2023211 · Cited by 10
CovMix: Covariance Mixing Regularization for Motor Imagery Decoding
Zoumpourlis, G., & Patras, I. (2022) · International Winter Conference on Brain-Computer Interface (BCI) · doi:10.1109/bci53720.2022.9734883 · Cited by 15
“2: Riemannian Alignment: Typically, alignment [10], [24] on EEG signals is performed separately within each session, applying the same session-specific transformation to all trials.”
MethodInfluentialArı, E., & Taçgın, E. (2023) · Brain Sciences · doi:10.3390/brainsci13020240 · Cited by 13
“As machine learning and deep learning methods, support vector machine [25], random forest [27], linear discriminant analysis [28], autoencoders [19], artificial neural network, deep belief network, recurrent neural network and convolutional neural network models [29] have been developed and used.”
MethodA probabilistic approach for calibration time reduction in hybrid EEG–fTCD brain–computer interfaces
Khalaf, A. F., & Akçakaya, M. (2020) · BioMedical Engineering OnLine · doi:10.1186/s12938-020-00765-4 · Cited by 12
An Adaptive CSP and Clustering Classification for Online Motor Imagery EEG
Jiang, Q., Zhang, Y., Ge, G., et al. (2020) · IEEE Access · doi:10.1109/access.2020.3016700 · Cited by 11
A Hybrid Transfer Learning Approach for Motor Imagery Classification in Brain-Computer Interface
Wang, X., Yang, R., Huang, M., Yang, Z., & Wan, Z. (2021) · IEEE Global Conference on Life Sciences and Technologies (LifeTech) · doi:10.1109/lifetech52111.2021.9391933 · Cited by 11
“For data alignment, this paper adopts Euclidean Space Data Alignment method [17] to align the source and target domains in the Euclidean space.”
MethodDemsy, O., Achanccaray, D., & Hayashibe, M. (2021) · Conference proceedings/Conference proceedings - IEEE International Conference on Systems, Man, and Cybernetics · doi:10.1109/smc52423.2021.9658731 · Cited by 12
“In fact, when dealing with TL case with multiple sources, several studies pool together all sources data into one big source [8], [11].”
BackgroundMethodInfluentialA lightweight spiking neural network for EEG-based motor imagery classification
Zhang, H., Wang, H., An, J., Zheng, S., & Wu, D. (2025) · Neural Networks · doi:10.1016/j.neunet.2025.107741 · Cited by 16
Self-citationZhu, J. L., Xu, G., Hu, Q., Wang, B., Zhou, T., & Qin, J. (2023) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2023.3341121 · Cited by 13
A causal perspective on brainwave modeling for brain–computer interfaces
Barmpas, K., Panagakis, Y., Zoumpourlis, G., Adamos, D. A., Laskaris, N., & Zafeiriou, S. P. (2024) · Journal of Neural Engineering · doi:10.1088/1741-2552/ad3eb5 · Cited by 17
“By utilizing both statistical and trainable methods (inspired by convariance-based alignment methods used in Riemannian geometry [72]).”
BackgroundMethodXu, G., Wang, Z., Hu, H., Zhao, X., Li, R., Zhou, T., & Xu, T. (2024) · IEEE Journal of Biomedical and Health Informatics · doi:10.1109/jbhi.2024.3402324 · Cited by 12
“He and Wu [24] then propose an Euclidean space data alignment and thus not only Riemannian methods can be taken advantage of to realize feature extraction and classification.”
MethodRiemannian transfer learning based on log-Euclidean metric for EEG classification
Zhuo, F., Zhang, X., Tang, F., Yu, Y., & Liu, L. (2024) · Frontiers in Neuroscience · doi:10.3389/fnins.2024.1381572 · Cited by 15
“In the Riemannian framework of sample alignment-based methods, data distribution changes manifest as geometric transformations of covariance matrices, which are referred to as “covariance shift" by He and Wu (2018).”
MethodUnsupervised Domain Adaptation via Spatial Pattern Alignment for VEP-Based Identity Recognition
Zhao, H., Wang, Y., & Gao, X. (2024) · IEEE Internet of Things Journal · doi:10.1109/jiot.2024.3431233 · Cited by 11
“He and Wu [28] employed Euclidean-space alignment (EA) for motor imagery classification, surpassing Riemannian space-based methods and attaining commendable results.”
BackgroundCalibration-free online test-time adaptation for electroencephalography motor imagery decoding
Wimpff, M., Döbler, M., & Yang, B. (2024) · International Winter Conference on Brain-Computer Interface (BCI) · doi:10.1109/bci60775.2024.10480468 · Cited by 13
“The most common DG approaches for EEG are data augmentation methods [6], [9] and alignment strategies [5], [10], [11].”
BackgroundMethodWen, Y., He, W., & Zhang, Y. (2022) · Journal of Neural Engineering · doi:10.1088/1741-2552/ac93b4 · Cited by 9
“The brain-computer interface (BCI), which has been widely used in neuroscience, medical engineering, and clinical rehabilitation, offers new research avenues and methods for human parsing of brain thought models and consciousness formation [1, 2].”
Huang, Z., Zheng, W., Wu, Y., & Wang, Y. (2020) · Journal of Neuroscience Methods · doi:10.1016/j.jneumeth.2020.108855 · Cited by 10
Wang, Z., Zhang, H., Ji, Z., Yang, Y., & Wang, H. (2023) · Brain Science Advances · doi:10.26599/bsa.2023.9050013 · Cited by 9
“Wu, who was awarded in the ERP final competition of the WRC 2021, proposed a novel method for improving the classification performance of RSVP across subjects [20].”
MethodUnsupervised heterogeneous domain adaptation for EEG classification
Wu, H., Xie, Q., Yu, Z., Zhang, J., Liu, S., & Long, J. (2024) · Journal of Neural Engineering · doi:10.1088/1741-2552/ad5fbd · Cited by 10
“He and Wu [32] developed a data alignment method for aligning EEG trials of different subjects in an Euclidean space, resulting in a faster alignment process compared to the technique used in the Riemannian space.”
Shi, Y., Wang, Y., & Meng, H. (2024) · Applied Intelligence · doi:10.1007/s10489-024-05662-0 · Cited by 13
“Examples of such methods include EA [13], WTLT [3], and JDAOT-Mix [7].”
An, S., Kang, M., Kim, S., Chikontwe, P., Shen, L., & Park, S. H. (2024) · Lecture notes in computer science · doi:10.1007/978-3-031-72120-5_63 · Cited by 14
Chen, N., Zhang, Y., Wu, J., et al. (2021) · Computational Intelligence · doi:10.1111/coin.12451 · Cited by 9
“In 2019, scientists from the UCSF published a study demonstrating a BCI capable of helping neurological condition patients with speech disability .7,8 Their BCI utilized high-density electrocorticography 9 to record neural activity from a patient’s brain.”
BackgroundLou, H., Ye, Z., Yao, L., & Zhang, Y. (2023) · IEEE Transactions on Neural Systems and Rehabilitation Engineering · doi:10.1109/tnsre.2023.3252610 · Cited by 11
“In [11], the mean covariance matrix of each user is transformed into an identity matrix for the alignment of their EEG trials.”
MethodMI brain-computer interfaces: A concise overview
Mandal, S. K., & Naskar, M. N. B. J. (2023) · Biomedical Signal Processing and Control · doi:10.1016/j.bspc.2023.105293 · Cited by 10
Variance characteristic preserving common spatial pattern for motor imagery BCI
Liang, W., Jin, J., Xu, R., Wang, X., & Cichocki, A. S. (2023) · Frontiers in Human Neuroscience · doi:10.3389/fnhum.2023.1243750 · Cited by 11
“For another, some transfer learning algorithms have been successfully applied to CSP algorithms (He and Wu, 2020; Wu et al., 2022).”
MethodWei, M., Yang, R., Huang, M., et al. (2023) · IEEE Transactions on Cognitive and Developmental Systems · doi:10.1109/tcds.2023.3338460 · Cited by 9
Multiclass motor imagery classification with Riemannian geometry and temporal-spectral selection
Li, Z., Tan, X., Li, X., & Yin, L. (2024) · Medical & Biological Engineering & Computing · doi:10.1007/s11517-024-03103-1 · Cited by 10
“Instead, the Riemannian method [33–36] allows the direct manipulation of the spatial EEG signal covariance matrix using a dedicated Riemannian geometry, which allows for a better differentiation between signal and noise components, resulting in improved signal-to-noise ratio.”
Classification of motor imagery using multisource joint transfer learning
Wang, F., Ping, J., Xu, Z., & Bi, J. (2021) · Review of Scientific Instruments · doi:10.1063/5.0054912 · Cited by 10
Ensemble Voting-Based Multichannel EEG Classification in a Subject-Independent P300 Speller
Mussabayeva, A., Jamwal, P. K., & Akhtar, M. T. (2021) · Applied Sciences · doi:10.3390/app112311252 · Cited by 9
Three-stage transfer learning for motor imagery EEG recognition
Li, J., She, Q., Meng, M., Du, S., & Zhang, Y. (2024) · Medical & Biological Engineering & Computing · doi:10.1007/s11517-024-03036-9 · Cited by 8
“To address this, different approaches have been adopted to align the target and source domain, such as the Euclidean space data alignment (EA) [25], the Riemann alignment (RA) [26], and the balanced distribution adaptive algorithm (BDA) [27].”
Metadata
- Authors
- He, H., & Wu, D.
- Affiliations
- Huazhong University of Science and Technology (CN)
- ORCID
- Venue
- IEEE Transactions on Biomedical Engineering
- Published
- 2020-02
- Volume
- 67
- Issue
- 2
- Pages
- 399-410
- ISSN
- 0018-9294 (ISSN-L)
- 1558-2531
- OpenAlex
- W2963727766
- Cited by
- 569
- References
- 52
- Citations by year
- 2026: 135 · 2025: 117 · 2024: 101 · 2023: 69 · 2022: 70 · 2021: 39 · 2020: 26
- FWCI
- 19.17
- Type
- journal article
- Language
- en
- Open access
- green
- arxiv.org
- Licence
- Locations
- IEEE Transactions on Biomedical Engineering (publishedVersion)
- PubMed (publishedVersion)
- arXiv (Cornell University) (submittedVersion)
- arXiv (Cornell University)
- Topics
- Life Sciences › Neuroscience › Cognitive Neuroscience › EEG and Brain-Computer Interfaces
- Physical Sciences › Computer Science › Artificial Intelligence › Emotion and Mood Recognition
- Physical Sciences › Computer Science › Human-Computer Interaction › Gaze Tracking and Assistive Technology
- Keywords
- brain-computer interface, EEG, transfer learning, data alignment, interindividual variability, motor imagery classification, unsupervised alignment, EEG preprocessing
- MeSH terms
- Machine Learning, Algorithms, Computer Simulation, Electroencephalography, Evoked Potentials, Humans, Imagination, Signal Processing, Computer-Assisted, Databases, Factual, Brain-Computer Interfaces
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
The claim verdicts this report counts are experimental: their accuracy is not yet measured.
A language model drew the claims from the paper on 2026-10-10, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).
A second model labelled each claim on 2026-10-10 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).
Cited sources are checked with a language model’s help (gpt-6-sol).
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.