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 301–400 of 520
Li, M., Pun, S. H., & Chen, F. (2024) · Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC) · doi:10.1109/embc53108.2024.10781999 · Cited by 2
“Other data alignment algorithms should be tested in the future, such as Euclidean-space alignment approach [20], centroid alignment approach [21], etc.”
A Novel Self-Supervised Contrastive Learning Framework for Masked EEG Motor Imagery Modeling
Zhang, K., Zhou, Q., & Hu, H. (2025) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp49660.2025.10888531 · Cited by 5
Cross-subject motor imagery classification via transfer Riemannian space-based sparse coding
Yang, Y., & Luo, T.-J. (2025) · International Journal of Intelligent Computing and Cybernetics · doi:10.1108/ijicc-05-2025-0275 · Cited by 2
Xiong, W., Ma, L., & Li, H. (2025) · Scientific Reports · doi:10.1038/s41598-025-24466-x · Cited by 2
“The data preprocessing pipeline consists of three key steps: channel-wise normalization, Euclidean space alignment of the covariance matrices 30 , and trial-level global normalization.”
Cross-domain correlation analysis to improve SSVEP signals recognition in brain-computer interfaces
Hu, K., Wang, Y., Tu, K., Guo, H., & Yan, J. (2025) · Biomedical Physics & Engineering Express · doi:10.1088/2057-1976/ae2772 · Cited by 2
Robust Covert Spatial Attention Decoding from Low-Channel Dry EEG by Hybrid AI Model
Kim, D., & Lee, J. (2025) · AI · doi:10.3390/ai7010009 · Cited by 2
Cyclic deep representation-based domain adaptation for cross-subject motor imagery classification
He, M., Cao, X., & Luo, T.-J. (2026) · Biomedical Signal Processing and Control · doi:10.1016/j.bspc.2026.109762 · Cited by 3
Gao, Y., Ma, Y., Liu, Y., Yin, G., & Qin, Y. (2026) · Cognitive Neurodynamics · doi:10.1007/s11571-026-10427-1 · Cited by 2
“EA-CSP (He and Wu 2019): Solves constraint satisfaction problems using evolutionary algorithms, simulating natural selection to find solutions that meet specific constraints.”
Motor Imagery Recognition Method Based on Multisource Transfer Learning and Multiclassifier Fusion
Gao, C., & Sun, J. (2022) · Wireless Communications and Mobile Computing · doi:10.1155/2022/3893866 · Cited by 2
Ouahidi, Y. E., Drumetz, L., Lioi, G., Farrugia, N., Pasdeloup, B., & Gripon, V. (2023) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp49357.2023.10097201 · Cited by 5
“Before training our models we align our data with Euclidean alignment [27], and resample the 4 datasets by oversampling the smaller datasets.”
MethodMI-EEG Recognition Based on Euclidean Alignment and Style Transfer Mapping
Niu, X., Zhang, X., & Yang, L. (2022) · International Conference on Computational Modeling, Simulation and Data Analysis (CMSDA) · doi:10.1109/cmsda58069.2022.00039 · Cited by 2
“EA is an unsupervised data alignment method, which takes the average covariance matrix of a single user as the reference matrix and maps the data based on affine invariance so that the mapped sample covariance matrix is similar to the identity matrix [7] .”
MethodSpatial Feature Regularization and Label Decoupling Based Cross-Subject Motor Imagery EEG Decoding
Zhou, Y., Luo, T.-J., Zhang, X., & Han, T. (2023) · Lecture notes in computer science · doi:10.1007/978-981-99-8558-6_34 · Cited by 2
Beyond Within-Subject Performance: A Multi-Dataset Study of Fine-Tuning in the EEG Domain
Sartzetaki, C., Antoniadis, P., Antonopoulos, N., Gkinis, I., Krasoulis, A., Perdikis, S., & Pitsikalis, V. (2023) · Conference proceedings/Conference proceedings - IEEE International Conference on Systems, Man, and Cybernetics · doi:10.1109/smc53992.2023.10394335 · Cited by 7
“We compare the proposed fine-tuning (FT) adaptation method with other BCI adaptation methods, namely the signal-level Euclidean alignment (EA) [22], model-level knowledge distillation (KD) [23], and prediction-level thresholding [29].”
MethodTransfer Learning for P300 Speller: An Alignment Approach for Cross-Subject P300 Data
Liu, M., Ahmad, F., & Beyette, F. R. (2024) · IEEE SoutheastCon · doi:10.1109/southeastcon52093.2024.10500162 · Cited by 3
“In 2019, [11] proposed a Euclidean alignment (EA) approach to align covariance matrices.”
BackgroundA Relation Feature Comparison Network for Cross-Domain Recognition of Motion Intention
Xu, J., Li, D., Zhou, P., Zhang, Y., Wang, Z., & Ma, D. (2024) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2024.3420350 · Cited by 5
“In the actual experiment, we select source domains based on the better subjects of the datasets (five groups, A1, A3, and A7–A9) [14].”
BackgroundMethodInfluentialFasterEA-FML for EEG: Federated Meta-learning with Faster Euclidean Space Data Alignment
Yao, M., Chen, W., Xu, T., Zhang, C., Liu, J., Chen, D., & Wang, Z. (2024) · Lecture notes in computer science · doi:10.1007/978-981-97-5591-2_41 · Cited by 2
Prototypical Contrastive Domain Adaptation Network for Nonstationary EEG Classification
Li, D., Xu, J., Zhang, Y., Ma, D., & Wang, J.-H. (2024) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2024.3476618 · Cited by 5
“Therefore, to reduce the data distribution differences in different domains, we utilize the EEG data alignment (EA) method [14] to align the covariance matrices of the source and target domains in the data preparation stage after balancing the computational speed.”
MethodMultifunctional neural probe for synchronized monitoring multiple brain signals
Cao, J., Wang, L., Tu, K., Xu, Q., Xu, M., Chen, H., Wang, X., Yang, B., & Liu, J. (2025) · Sensors and Actuators A Physical · doi:10.1016/j.sna.2025.116572 · Cited by 2
AEP: An adaptive ensemble P300-BCI classifier based on user-feedback and knowledge-transfer
Huang, Z., Chen, Q., Chen, X., Zheng, W., Lin, Z., & Luo, T.-J. (2025) · Applied Intelligence · doi:10.1007/s10489-025-06612-0 · Cited by 3
“He et al. developed a BCI transfer learning approach that aligns EEG data from different subjects in the Euclidean space to make them more similar, and hence improved the performance of BCI classifiers for new subjects [26].”
TFANet: a temporal fusion attention neural network for motor imagery decoding
Zhang, C., Liu, Y., & Wu, X. (2025) · Frontiers in Neuroscience · doi:10.3389/fnins.2025.1635588 · Cited by 5
EEG-Based Driver Drowsiness Estimation Using Self-Paced Learning with Label Diversity
Xu, Y., & Wu, D. (2019) · IEEE Symposium Series on Computational Intelligence (SSCI) · doi:10.1109/ssci44817.2019.9002753 · Cited by 2
“Transfer learning [14], [15], which utilizes data from auxiliary subjects or sessions to facilitate the learning for a new subject, has been widely used in BCIs.”
BackgroundSelf-citationCross-subject MEG Transfer Learning by Riemannian Manifold and Feature Subspace Alignment
Liu, S., Yu, T., Huang, Z., & Ye, H. (2020) · International Symposium on Autonomous Systems (ISAS) · doi:10.1109/isas49493.2020.9378879 · Cited by 2
“A data alignment procedure has been proposed to notably decrease the discrepancy between source and target [9], which outperforms the Riemannian alignment approach to motor imagery paradigm.”
MethodZhang, S., Lin, J., You, Y., Yu, N., & Han, J. (2022) · Measurement and Control · doi:10.1177/00202940221105092 · Cited by 2
“Another possible reason may be that some outliers were in these subjects.(37) The CorrCA-RFG could not deal with these outliers.”
BackgroundReducing calibration time using novel hybrid transfer-learning for P300-based BCI applications
Kilani, S., Aghili, S. N., Ghassemlooy, Z. F., Hulea, M. G., & Wu, Q. (2022) · International Symposium on Communication Systems, Networks and Digital Signal Processing (CSNDSP) · doi:10.1109/csndsp54353.2022.9908000 · Cited by 2
Heterogeneous domain adaptation for intracortical signal classification using domain consensus
Dong, Y., Hu, D., Wang, S., & He, J. (2022) · Biomedical Signal Processing and Control · doi:10.1016/j.bspc.2022.104540 · Cited by 2
Comparison of Generic and Subject-Specific Training for Features Classification in P300 Speller
Mussabayeva, A., Jamwal, P. K., & Akhtar, M. T. (2020) · Cited by 2
EEG-Based Driver Drowsiness Estimation Using Feature Weighted Episodic TrainingPreprint
Cuui, Y., Xu, Y., & Wu, D. (2019) · arXiv (Cornell University) · doi:10.48550/arxiv.1909.11456 · Cited by 2
“He and Wu [32] proposed a similar EEG data alignment approach in the Euclidean space, which is more efficient than the Riemannian space data alignment approach, and can be used as a pre-processing step before any Euclidean space classifier.”
MethodSelf-citationBenzoquinoline Chemical Space: A Helpful Approach in Antibacterial and Anticancer Drug Design
Lungu, C. N., Mangalagiu, V., Mangalagiu, I. I., & MEHEDINTI, M. C. (2023) · Molecules · doi:10.3390/molecules28031069 · Cited by 1
“All biological variables, reflecting either a discrete or a continuous presence and interaction of biological entities, can be represented and measured in Euclidian space [9].”
BackgroundSingh, P., Dewangan, N. K., Potdar, R. M., et al. (2024) · Polish Journal of Medical Physics And Engineering · doi:10.2478/pjmpe-2024-0016 · Cited by 1
Chen, H., Wang, D., Xu, M., Chen, J., Zhang, Y., & Chen, Y. (2024) · Neurocomputing · doi:10.1016/j.neucom.2024.129239 · Cited by 1
Many-to-Many Transfer Learning on Motor Imagery BCI
Altindış, F., Yılmaz, B., & Congedo, M. (2024) · International Conference on Decision Aid Sciences and Applications (DASA) · doi:10.1109/dasa63652.2024.10836450 · Cited by 2
Cai, Y., & Liu, X. (2024) · Conference proceedings/Conference proceedings - IEEE International Conference on Systems, Man, and Cybernetics · doi:10.1109/smc54092.2024.10831755 · Cited by 1
Machine Learning in Recognition of Native and Artificially Generated EEGs
Rusakov, K. D., Turovsky, Y. A., & Киселев, Е. А. (2024) · Journal of Communications Technology and Electronics · doi:10.1134/s1064226924700359 · Cited by 1
“The normalization was performed using the standard normalization method, in which each value in the EEG data was transformed in such a way that the mean value was zero and the standard deviation was unity [20].”
Personalized Continual EEG Decoding: Retaining and Transferring Knowledge
Li, D., Shin, H.-B., & Yin, K. (2025) · International Winter Conference on Brain-Computer Interface (BCI) · doi:10.1109/bci65088.2025.10931666 · Cited by 3
“Euclidean alignment (EA) [19] is an alignment technique designed to reduce the inter-subject variability of EEG signals by calculating a common reference covariance matrix, thereby improving the robustness of downstream analysis and classification tasks.”
West, J., Singh, C., Cha, J. S., & Li, D. (2025) · IISE Transactions on Healthcare Systems Engineering · doi:10.1080/24725579.2025.2492555 · Cited by 1
“Additionally, the CSP methods Azab et al. (2019); Cheng et al. (2017) and euclidean alignment from He and Wu (2020) rely on calculating the covariance matrix which can be computationally expensive, especially with high-dimensional data, and might not be suitable for real time transfer learning,…”
InfluentialXie, C., Wang, L., Yang, J., & Guo, J. (2025) · Journal of Neuroscience Methods · doi:10.1016/j.jneumeth.2025.110483 · Cited by 2
Fan, C., Yang, F., Zhang, J., et al. (2025) · IEEE Transactions on Consumer Electronics · doi:10.1109/tce.2025.3573773 · Cited by 1
“Following this, we applied Euclidean alignment [34].”
Mannan, M. M. N., Lloyd, D. G., & Pizzolato, C. (2025) · International Conference on Rehabilitation Robotics · doi:10.1109/icorr66766.2025.11063010 · Cited by 1
Multi-source Dictionary Transfer Learning for few-shot motor imagery EEG classification
Li, X., She, Q., Cai, Y., Fang, F. Z., Meng, M., & Zhang, Y. (2025) · Signal Image and Video Processing · doi:10.1007/s11760-025-04532-7 · Cited by 1
“He et al. [23] extended RA to Euclidean alignment (EA), which realize the distribution alignment and EEG classification in Euclidean space.”
Ouyang, Y., Cheng, W., Wang, L., Zhu, X., & Zeng, H. (2025) · IEEE Journal of Biomedical and Health Informatics · doi:10.1109/jbhi.2025.3619419 · Cited by 1
“To minimize data discrepancies among different clients, Jia et al. [27] proposed FedBS, which incorporated Euclidean alignment (EA) [32] to reduce inter-subject variability in EEG data.”
Zhang, H., Xie, J., Liu, K., et al. (2025) · Computers in Biology and Medicine · doi:10.1016/j.compbiomed.2025.111284 · Cited by 1
Takens-Based Kernel Transfer Entropy Connectivity Network for Motor Imagery Classification
Gomez-Rivera, A., Álvarez-Meza, A. M., Peña, D. A. C., & Orozco, A. A. (2025) · Sensors · doi:10.3390/s25227067 · Cited by 3
“In fact, MI-based BCIs have proven effective in neuroprosthetic systems, thereby enhancing the quality of life for patients suffering from motor impairments, such as those resulting from spinal cord injuries, amyotrophic lateral sclerosis, or stroke [9].”
Amoozgar, R., & Hosseinzadeh, H. (2025) · doi:10.1109/ipria68579.2025.11263626 · Cited by 1
“In [ 11 ], a method for aligning EEG trials from different individuals in Euclidean space is proposed.”
Xiong, P., Ouyang, J., Wu, Z., Zhang, Y., Li, X., Chen, G., Wang, Y., & Fan, M. (2025) · Proceedings of the ACM on Interactive Mobile Wearable and Ubiquitous Technologies · doi:10.1145/3770642 · Cited by 2
“Normalization is conducted in this module, which can cope with individual differences in a real-time EEG system [93].”
Wimpff, M., Aristimunha, B., Chevallier, S., & Yang, B. (2025) · Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC) · doi:10.1109/embc58623.2025.11253543 · Cited by 9
“Specifically, we use Euclidean alignment (EA) [38], [39] in an online fashion [34] to mitigate the domain shift of the input data between sessions.”
MIRepNet: A pipeline and pre-trained model for EEG-based motor imagery classification
Liu, D., Chen, Z., Luo, J., Lian, S., Chen, Y., Hou, S., Zhu, X., & Wu, D. (2026) · Knowledge-Based Systems · doi:10.1016/j.knosys.2026.115966 · Cited by 20
“To reduce marginal distribution shifts, we adopt Euclidean alignment (EA) [18], [42] via whitening transformations.”
Self-citationKDFNet: Knowledge-data fusion network for motor imagery based brain-computer interfaces
Jiang, X., Meng, L., Chen, X., Li, W., & Wu, D. (2025) · Information Sciences · doi:10.1016/j.ins.2025.123001 · Cited by 2
Self-citationEfficient Adaptive Learning via Partial-Update Variable Step-Size LMS for Real-Time ERP Denoising
Boudiaf, M. A., Benkherrat, M., Djelel, S., Messadeg, D., & Absi, R. (2025) · Applied Sciences · doi:10.3390/app152312702 · Cited by 2
Eyvazpour, R., Farrokhi, B., & Erfanian, A. (2025) · iScience · doi:10.1016/j.isci.2025.114521 · Cited by 4
DPMMN: A dual performer-multi-modal network for emotion recognition
Gornale, S. S., Shivakumara, P., Unki, A., & Vadera, S. (2025) · Signal Processing Image Communication · doi:10.1016/j.image.2025.117464 · Cited by 1
Wu, X., Gao, W., Lu, J., & Gao, Y. (2025) · Information · doi:10.3390/info17010015 · Cited by 1
Röhr, H., Atad, D. A., Trautwein, F., Mediano, P. A. M., Dor‐Ziderman, Y., Schweitzer, Y., Berkovich‐Ohana, A., Schmidt, S., & van Vugt, M. K. (2025) · Human Brain Mapping · doi:10.1002/hbm.70440 · Cited by 2
“…such as feature-space alignment and transfer learning approaches have been suggested to address this issue by reducing inter-subject variability (He and Wu 2020; Zanini et al. 2018) and might be promising methods to improving the across-participant generalization of meditation-state classifiers…”
Wang, H., Xu, S., Guo, R., Han, J., & Huang, M.-C. (2026) · Diagnostics · doi:10.3390/diagnostics16020293 · Cited by 1
Gao, W., Liu, D., Wang, Q., Zhao, Y., & Sun, J. (2026) · IEEE Journal of Biomedical and Health Informatics · doi:10.1109/jbhi.2026.3650972 · Cited by 1
“For example, He et al. [25] align task-state EEG signals across subjects by computing the mean covariance matrix from resting-state EEG signals.”
Cripe, C. (2026) · Artificial intelligence · doi:10.5772/intechopen.1014199 · Cited by 1
Selective Forgetting in Machine Learning and Beyond: A Survey
Sha, A. S., Nunes, B. P., & Haller, A. (2026) · ACM Computing Surveys · doi:10.1145/3796542 · Cited by 18
“In the medical and healthcare sector, active forgetting approaches improve diagnostic accuracy by disregarding non-relevant features from different demographics or imaging devices, thus enhancing model transferability [94, 171, 184, 286].”
Spiking neural network for intra-cortical brain signal decoding
Yang, S., Fu, H., Zhang, H., Zhang, P., Li, W., & Wu, D. (2026) · Knowledge-Based Systems · doi:10.1016/j.knosys.2026.115508 · Cited by 5
Self-citationGao, R., Han, R., Dong, H., Suganthan, P. N., & Yu, Y. (2026) · Applied Soft Computing · doi:10.1016/j.asoc.2026.114796 · Cited by 1
FSTAM-net: A dual attention network for motor imagery decoding and visualization analysis
Fu, S., Shi, B., Wei, Y., Wu, D., Yin, S., Liu, G., Zhang, J., Zhu, A., Wang, J., & Zhou, G. (2026) · Biomedical Signal Processing and Control · doi:10.1016/j.bspc.2026.109916 · Cited by 1
Self-citationA Lightweight Dual-Attention Neural Network for Robust and Efficient EEG Motor Imagery Decoding
Wang, G., Song, X., Jiang, L., Zhang, Y., Yao, D., Lu, J., Xu, P., & Li, F. (2026) · International Journal of Neural Systems · doi:10.1142/s0129065726500267 · Cited by 2
Zhou, X., Wang, L., Zhang, L., Yao, W., & Li, G. (2026) · Journal of Neuroscience Methods · doi:10.1016/j.jneumeth.2026.110768 · Cited by 1
Al-Majed, R., Hussain, M., & Al-Ahmadi, S. A. (2026) · Engineering Science and Technology an International Journal · doi:10.1016/j.jestch.2026.102373 · Cited by 1
Abinaya, G., & Dinakaran, K. (2026) · MethodsX · doi:10.1016/j.mex.2026.103913 · Cited by 1
“Existing solutions largely address cross-subject or cross-session transfer within a single dataset through domain adaptation techniques [8] such as Euclidean alignment [9,10], covariate shift correction [11], and adversarial training [12].”
Luo, T.-J., & Zhang, X. (2026) · IRBM · doi:10.1016/j.irbm.2026.100952 · Cited by 1
An EEG Transfer Learning Algorithm Based on Mutual Information and Transfer Component Analysis
Hu, C., Cai, J., Liang, Z., Wang, K., Zhang, Y., & Chen, W. (2022) · IEEE Conference on Industrial Electronics and Applications (ICIEA) · doi:10.1109/iciea54703.2022.10006050 · Cited by 1
“EA-CSP-LDA [17]: After alignment by EA, it is then classified by CSP algorithm and finally by LAD.”
MethodHou, S., Jiang, A., Tang, Y., Li, M., & Zhu, Y. (2023) · International Congress on Image and Signal Processing (CISP) · doi:10.1109/cisp-bmei60920.2023.10373302 · Cited by 1
“Additionally, He and Wu [24] extended RA to Euclidean alignment (EA) in Euclidean space, enabling the use of any Euclidean space classifier.”
BackgroundA Debiased Domain Adaptation Framework with Minimum Class Confusion for Motor Imagery Decoding
Fan, C., Chen, Z., Song, X., Li, X., Xue, J., Li, P., & Lv, Z. (2024) · International Joint Conference on Neural Networks (IJCNN) · doi:10.1109/ijcnn60899.2024.10649993 · Cited by 1
“2) Euclidean alignment: Euclidean alignment [22], similar to Z-score, is a more effective method that can be used to reduce the discrepancy between two domains.”
MethodA class alignment network based on self-attention for cross-subject EEG classification
Ma, S., Zhang, D., Wang, J., & Xie, J. (2024) · Biomedical Physics & Engineering Express · doi:10.1088/2057-1976/ad90e8 · Cited by 2
Bi-Stream Adaptation Network for Motor Imagery Decoding
Wang, Z., Li, A., Wang, Z., et al. (2024) · doi:10.1109/embc53108.2024.10782480 · Cited by 1
“And He et al. [9] presented a low-cost unsupervised method aligning EEG trials in the Euclidean space.”
Khabti, J., Al-Ahmadi, S. A., & Soudani, A. (2024) · Sensors · doi:10.3390/s24248127 · Cited by 2
“Several papers, such as [6,38–41], utilize Riemannian geometry and alignment techniques to address the differences between subjects in transfer learning for EEG classification.”
InfluentialA Capsule Decision Neural Network Based on Transfer Learning for EEG Signal Classification
Zhang, W., Tang, X., Dang, X., & Wang, M. (2025) · Biomimetics · doi:10.3390/biomimetics10040225 · Cited by 2
“In order to accurately evaluate the method proposed in this paper, we use the same EEG signal preprocessing method as the literature [28].”
InfluentialBrain-Computer Interface (BCI) in Healthcare
Shaikh, A., Aparna, V. B., Munene, R., et al. (2025) · IGI Global eBooks · doi:10.4018/979-8-3693-7427-6.ch019 · Cited by 1
Senthil Vadivelan, D., & Sethuramalingam, P. (2025) · Computers in Biology and Medicine · doi:10.1016/j.compbiomed.2025.110675 · Cited by 1
Tiny Noise Can Make an EEG-Based Brain-Computer Interface Speller Output Anything.
Zhang, X., Wu, D., Ding, L., Luo, H., Lin, C., & Jung, T. (2020) · Cited by 1
Self-citationShen, J., Li, X., Zeng, H., & Song, A. (2019) · IEEE Symposium Series on Computational Intelligence (SSCI) · doi:10.1109/ssci44817.2019.9002946 · Cited by 1
“Recent years have witnessed an increasing interest in using the EEG covariance matrices as the input for classification in some BCI tasks [1,7,8], which has been proven to achieve reliable performance.”
BackgroundXu, D., & Li, M. (2022) · PubMed · doi:10.7507/1001-5515.202108060 · Cited by 3
Inter-session Transfer Learning in MI Based BCI for Controlling a Lower-Limb Exoskeleton
Ferrero, L., Quiles, V., Ortíz, M., Juan, J. V., Iáñez, E., & Azorín, J. M. (2022) · Lecture notes in computer science · doi:10.1007/978-3-031-06527-9_24 · Cited by 3
Different Set Domain Adaptation for Brain-Computer Interfaces: A Label Alignment ApproachPreprint
He, H., & Wu, D. (2019) · arXiv (Cornell University) · doi:10.48550/arxiv.1912.01166 · Cited by 1
“Although it has been shown to significantly improve the cross-subject classification performance in [27], it only aligns the data in the feature space, and may not work well when the source subjects and the target subject have different label spaces.”
BackgroundMethodSelf-citationWei, X., Faisal, A. A., Grosse‐Wentrup, M., Gramfort, A., Chevallier, S., Jayaram, V., Jeunet, C., Bakas, S., Ludwig, S., Barmpas, K., et al. (2022) · arXiv (Cornell University) · doi:10.48550/arxiv.2202.12950 · Cited by 42
“Euclidean Alignment is used to close the gap between subjects in source domains and target domain He and Wu (2019).”
MethodInfluentialGao, Y., cai, Z., Meng, M., et al. (2023) · SSRN Electronic Journal · doi:10.2139/ssrn.4656119 · Cited by 1
An open dataset of multidimensional signals based on different speech patterns in pragmatic Mandarin
Zhao, R., Bai, Y., Zhang, S., Zhu, J., Liu, H., & Ni, G. (2025) · Scientific Data · doi:10.1038/s41597-025-06213-z · Cited by 6
“However, these technologies have several limitations, including restricted freedom of movement, lengthy training periods, the necessity for users to concentrate on specific tasks, and communication rates that are significantly slower than natural speech 1–3 .”
Patch LearningPreprint
Wu, D., & Mendel, J. M. (2019) · arXiv (Cornell University)
Self-citationManifold Embedded Knowledge Transfer for Brain-Computer InterfacesPreprint
Zhang, W., & Wu, D. (2019) · arXiv (Cornell University) · doi:10.48550/arxiv.1910.05878
“Euclidean Alignment (EA) Although RA-MDM has demonstrated promising performance, it still has some limitations [23]: 1) it processes the covariance matrices in the Riemannian space, whereas there are very few Riemannian space classifiers; 2) it computes the reference matrix from the non-target stimuli in ERP-based BCI…”
BackgroundMethodInfluentialSelf-citationTan, Y. (2021) · Tuwhera (Auckland University of Technology)
Técnicas de aprendizado de máquina para interfaces cérebro-computador baseadas em EEG
de Almeida, W. F. (2021) · Digital Library of Theses and Dissertations (Universidade de São Paulo) · doi:10.11606/d.100.2021.tde-20062021-211101
A Fast-Incremental SVM Algorithm for Online Motor Imagery-Based Brain-Computer Interface
Jiang, Q., Zhang, Y., He, X., et al. (2021) · IEEE International Conference on Real-time Computing and Robotics (RCAR) · doi:10.1109/rcar52367.2021.9517630
Weighted Conditional Distribution Adaptation for Motor Imagery Classification
Zhang, R., Gu, F., Zou, Z., Yu, T., & Li, Y. (2021) · Lecture notes in computer science · doi:10.1007/978-3-030-87355-4_42
Cross-Sessions and Cross-Paradigms Analysis for the Problem of Brain-Computer Interface Inefficiency
Chen, W., Huang, G., Hu, Z., Zhang, L., Li, L., Liang, Z., & Zhang, Z. (2021) · International Conference on Mechatronics and Machine Vision in Practice (M2VIP) · doi:10.1109/m2vip49856.2021.9664989
“Due to the significant class imbalance in P300 experiment, we used the balanced classification accuracy as the performance measure for both paradigms[17].”
Method2020 Index IEEE Transactions on Biomedical Engineering Vol. 67
(2020) · IEEE Transactions on Biomedical Engineering · doi:10.1109/tbme.2020.3048339
Multi-Source Geometric Metric Transfer Learning for EEG ClassificationPreprint
Zhang, X., She, Q., Tan, T., et al. (2022) · SSRN Electronic Journal · doi:10.2139/ssrn.4155138
Towards Real-World Neuromonitoring and Applications in Cognitive Engineering
Wei, C.-S., & Jung, T. (2021) · Handbook of Neuroengineering · doi:10.1007/978-981-15-2848-4_79-1
Cross-subject Trials Reweighting for Enhancing Motor Imagery-based Brain-Computer Interface
Liang, Z., Zheng, Z., Chen, W., Wang, J., Zhang, J., Chen, J., & Shi, H. (2022) · IEEE Conference on Industrial Electronics and Applications (ICIEA) · doi:10.1109/iciea54703.2022.10005954 · Cited by 1
“For more details about the EA algorithm, see [18].”
MethodTowards Real-World Neuromonitoring and Applications in Cognitive Engineering
Wei, C.-S., & Jung, T. (2023) · Handbook of Neuroengineering · doi:10.1007/978-981-16-5540-1_79 · Cited by 1
An Algorithm for Motor Imagery Classification Based on Transfer Learning and Feature Fusion
Shuaibin, W., Wu, J., Zhang, D., Suo, D., & Yan, T. (2022) · ICME International Conference on Complex Medical Engineering (CME) · doi:10.1109/cme55444.2022.10063324
“Essentially, EA serves as a compensation of covariate shift [21] which means the distribution changes of the inputs.”
BackgroundMethodInfluentialWang, J., & Chen, Y. (2022) · Machine learning · doi:10.1007/978-981-19-7584-4_1
Kieu, D. H., Cuong, T. V., Nguyen, H. M., Vu, Q. N., Nguyen, V. D., & Le, H. T. (2023) · SSRN Electronic Journal · doi:10.2139/ssrn.4460037
Pousson, J. E., Shen, Y.-W., Lin, Y., Voicikas, A., Pipinis, E., Bernhofs, V., Burmistrova, L., & Griškova-Bulanova, I. (2023) · IEEE Transactions on Neural Systems and Rehabilitation Engineering · doi:10.1109/tnsre.2023.3327740 · Cited by 2
“…multiple-day analytical scenario facilitated a better understanding of the role of the frontal cortex ( e.g. , DLPFC (BA9), ACC (BA 24, 32, and 33) [8], [39], [40], [41], [42]) in intervening in emotional processes and expressing spectral signatures that are relatively resistant to ecological…”
BackgroundOptimal Transport Feature Ranking Based Multi-Source Transfer Learning for Eeg ClassificationPreprint
Li, J., She, Q., Fang, F. Z., et al. (2023) · SSRN Electronic Journal · doi:10.2139/ssrn.4657125
An novel SDA-CNN few shot domain adaptation framework for silent speech recognition
Ramkumar, N., & Renuka, D. K. (2024) · Journal of Intelligent & Fuzzy Systems · doi:10.3233/jifs-237890
“Prior to CSP, EEG signals from both domains underwent Eigenvalue Decomposition in the Euclidean space [24].”
RetractedDeep machine learning for biosignal analysis
Dissanayakage, P. T. T. (2024) · Queensland University of Technology · doi:10.5204/thesis.eprints.247177
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.