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 401–500 of 520
Jiang, A., Hou, S., Tang, Y., & Zhu, Y. (2024) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp48485.2024.10445838 · Cited by 1
“In [23], RA was extended to Euclidean alignment (EA) in Euclidean space, enabling the use of any Euclidean space classifier.”
MethodA Multi-Morphological Representation Approach for Analysis of Multi-Source Eeg SignalsPreprint
Gao, Y., Meng, M., She, Q., et al. (2024) · SSRN Electronic Journal · doi:10.2139/ssrn.4778142
Assessing the Rigor of Machine Learning in Physiological Signal Processing Applications
Ishmakhametov, N., Naser, M. Y. M., & Bhattacharya, S. (2024) · IEEE SoutheastCon · doi:10.1109/southeastcon52093.2024.10500274 · Cited by 2
Fang, H., Wang, M., Wang, Y., & Yang, Y. (2024) · doi:10.36227/techrxiv.171560997.77552223/v1 · Cited by 1
Weighted Common Spatial Pattern Based Adaptation Regularization for Multi-Source Eeg Time SeriesPreprint
Han, R., Li, Z., Zhang, Y., et al. (2024) · SSRN Electronic Journal · doi:10.2139/ssrn.4858187
A Novel Multi-Morphological Representation Approach for Multi-Source EEG SignalsPreprint
Gao, Y., Meng, M., Fang, F. Z., et al. (2024) · SSRN Electronic Journal · doi:10.2139/ssrn.4868539
Transferrable Subject-Independent Feature Representation for Discriminating EEG-Based Brain Signals
Lou, H., Ye, Z., & Yao, L. (2024) · Guidance Navigation and Control · doi:10.1142/s273748072441005x
Wang, T., Ke, Y., He, F., et al. (2024) · doi:10.1109/embc53108.2024.10782484
“Additionally, He et al. presented an approach for aligning EEG trials in the Euclidean space to address individual differences in EEG-based motor imagery classification [14].”
InfluentialZhou, J., Lian, C., Lan, X., Shi, X., Wang, L., Yan, N., & Guo, Y. (2024) · Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC) · doi:10.1109/embc53108.2024.10782219
Luo, T.-J., Li, J., Li, R., Zhang, X., Wu, S.-R., & Peng, H. F. (2022) · Journal of Integrative Neuroscience · doi:10.31083/j.jin2312218
“Specifically, the covariance matrix centroid alignment (CMCA) method [47] was applied to align filtered cross-subject MI-EEG signals.”
InfluentialEuclidean Alignment for Transfer Learning in Multi-band Common Spatial Pattern
Amorim, M. M., Prata, L., Maurício, J. S. S., et al. (2025) · Lecture notes in computer science · doi:10.1007/978-3-031-79029-4_30
Neural Signal-Based Speech Recognition: Leveraging EEG for Speech Detection
Bhutani, P., Madan, P., Mishra, U., Vishnoi, A., & Arya, P. K. (2024) · International Conference on Computational Intelligence and Network Systems (CINS) · doi:10.1109/cins63881.2024.10862994 · Cited by 1
Mannan, M. M. N., Lloyd, D. G., & Pizzolato, C. (2025) · Biosystems & biorobotics · doi:10.1007/978-3-031-77588-8_121
Graph Attention Network for Motor Imagery Classification
Kieu, H.-D. (2024) · RIVF International Conference on Computing and Communication Technologies (RIVF) · doi:10.1109/rivf64335.2024.11009062
“Several models have been proposed in the past to address the task of MI classification, such as Euclidean Alignment [4], Pre-alignment Strategy [15], Common spatial pattern [16], [17], Linear Discriminant Analysis (LDA).”
Lightweight and Efficient Motor Imagery Decoding for High-ITR Multi-task Brain-Computer Interface
Ma, X., Wang, S., Xu, X., Li, P., Hu, P., Liu, T., & Zhong, S. (2024) · International Conference on Communication and Information Processing (ICCIP) · doi:10.1145/3708657.3708692
“This phenomenon could be due to overfitting to noisy or unrepresentative data for these subjects, or variability in task engagement or signal quality [32].”
Adaptive Convex Spherical Structure Detection for Semi-supervised Feature Extraction
Zhang, X., Zhou, J. N., & Gao, C. (2025) · Communications in computer and information science · doi:10.1007/978-981-96-6951-6_15
Li, Y., & Zhang, J. (2025) · PLoS ONE · doi:10.1371/journal.pone.0327121
Jung, E., & An, J. (2025) · SSRN Electronic Journal · doi:10.2139/ssrn.5343882
Exploring Cortical Responses to Blood Flow Restriction through Deep Learning
Köllner, J., Wimpff, M., Gizzi, L., Vujaklija, I., & Yang, B. (2025) · International Conference on Rehabilitation Robotics · doi:10.1109/icorr66766.2025.11063023
“Two distinct approaches were employed for this: standardization, as has been done with similar data for DL-models [16], and Euclidean-space alignment [17].”
A domain adaptation method based on domain selection and dual-space feature extractor
Xiao, N., & Li, M. (2025) · Journal of Physics Conference Series · doi:10.1088/1742-6596/3079/1/012067
From signals to decisions in noninvasive neural technologies
Cabestaing, F., & Chevallier, S. (2025) · Elsevier eBooks · doi:10.1016/b978-0-443-24824-5.00020-x · Cited by 1
Brain-Robot Interface-Based Sound Source Navigation of an Assistive Robot in Industry 4.0Preprint
Saravanan, M., & Chowdhury, A. R. (2025) · Research Square · doi:10.21203/rs.3.rs-6928093/v1
A Robust Brain-Computer Interface for Reliable Cognitive State Classification and Device Control
Derkach, M., Matiuk, D., Skarga-Bandurova, I., et al. (2024) · International Conference on Dependable Systems, Services and Technologies (DESSERT) · doi:10.1109/dessert65323.2024.11122175
“Recent studies have shown that the application of data normalization significantly improves classification accuracy in BCI systems by reducing variability across subjects and sessions [3].”
InfluentialShi, K., Huang, R., Li, Z., Lyu, J., Zhao, Y., Song, G., Cheng, H., & Zhang, J. (2025) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra55743.2025.11128281
“Other research efforts have focused on using data alignment methods [10], [11] and lightweight network structures [12]–[16] to extract shared features across the frequency, time, and spatial domains of MI-EEG data from different subjects.”
Han, D., Ren, Y., Wang, Y., & Deng, C. L. (2025) · Magnetic Medicine · doi:10.1016/j.magmed.2025.100038 · Cited by 1
Synthetica: The First AI-Native Cohort and the Democratic Risks of Synthetic InfrastructuresPreprint
Milelli, R. (2025) · SSRN Electronic Journal · doi:10.2139/ssrn.5450674
BGTransform: a neurophysiologically informed EEG data augmentation framework
Yue, J., Xiao, X., Zhang, H., Xu, M., & Ming, D. (2025) · Journal of Neural Engineering · doi:10.1088/1741-2552/ae0c3a
“Moreover, by preserving neurophysiological fidelity, BGTransform-produced data may improve model robustness and generalization across subjects and sessions [44].”
EEG-based motor execution classification of upper and lower extremities using machine learning
Korkmaz, I., & Tepe, C. (2025) · Computer Methods in Biomechanics & Biomedical Engineering · doi:10.1080/10255842.2025.2566260
Hybrid curriculum and transfer learning method for cross-subject motor imagery recognition
Gao, C., Li, X., Cheng, S., Liu, W., & Guo, S. (2025) · Expert Systems with Applications · doi:10.1016/j.eswa.2025.130221
Khoshkavandi, P., Shamsollahi, M. B., & Nasrabadi, A. M. (2025) · doi:10.1109/icee67339.2025.11213879
“Brain signals associated with motor activity are inherently noisy and often tricky to decode, complicating achieving real-time, reliable control over assistive devices [2].”
NeuroCommTrainer: Toward an Adaptive and Wearable Multimodal Brain–Computer Interface
Scherer, J., Finke, A., Everding, V., Lindenbaum, L., Kayser, C., & Kissler, J. M. (2025) · Brain Connectivity · doi:10.1177/21580014251393151
SDDA: Spatial Distillation Based Distribution Alignment for Cross-Headset EEG Classification
Liu, D., Li, S., Wang, Z., Li, W., & Wu, D. (2025) · IEEE Transactions on Biomedical Engineering · doi:10.1109/tbme.2025.3631604 · Cited by 9
“To address this, we introduce DA, which further reduces the distribution shifts via: 1) Input-space data normalization using session-wise Euclidean alignment (EA) [24].”
Self-citationPFC-TTA: Parameter-Free Classifier for Test Time Adaptation in Emotion Recognition
Shao, H., Dong, L., Li, L., & Zhang, L. (2025) · International Joint Conference on Neural Networks (IJCNN) · doi:10.1109/ijcnn64981.2025.11227210
“During the transfer to the target domain, unlabeled electroencephalography (EEG) trials are processed and aligned in the same way, followed by analysis utilizing the ensemble of encoders [13].”
Toward Memory-Efficient Continual Adaptation for MI-EEG Decoding in BCIs
Li, D. X., Shin, H.-B., & Lee, S. (2025) · IEEE Transactions on Systems Man and Cybernetics Systems · doi:10.1109/tsmc.2025.3630757
“EA [14] is a data preprocessing technique designed to mitigate domain shifts before the data enters the neural network.”
DS-tCNN:A Subject-Specific Feature Decoupling Method for SSVEP-Based Brain-Computer Interfaces
Ma, J., Li, P., Li, X., et al. (2025) · Journal of Physics Conference Series · doi:10.1088/1742-6596/3147/1/012017
Wei, R., Hua, C., Chen, J., et al. (2025) · IEEE Transactions on Neural Systems and Rehabilitation Engineering · doi:10.1109/tnsre.2025.3639091
“Another promising strategy involves transfer learning [12] or domain adaptation techniques [13], which reduces the distributional di ff erences between subjects by fine-tuning models with a small amount of subject-specific data.”
Sub-Group Partition Strategy for RSVP-based Collaborative Brain-Computer Interfaces
Si, Y., Wang, Z., Zhao, X., Xu, T., Zhou, T., & Hu, H. (2025) · Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC) · doi:10.1109/embc58623.2025.11252828
“Additionally, Euclidean alignment (EA) has been proposed to reduce ERP signal variability across sessions [8].”
A Study of Brain-Computer Interface Recognition Performance Crossing Action Observation Paradigms
Hu, G., Zeng, F., Tang, H., et al. (2025) · doi:10.1109/embc58623.2025.11254040
“Cross-domain algorithms aim to develop a generalized pattern recognition model for EEG-based BCI systems, optimized for diverse subjects, sessions, devices, and tasks[5].”
Fu, R., Liu, Y., Lu, B., et al. (2025) · Biomedical Signal Processing and Control · doi:10.1016/j.bspc.2025.109286
EEG-based meditation decoding: tackling subject variability with spatial and temporal alignment
Karaiskou, A.-Ι., Varon, C., Musluoglu, C. A., Alaerts, K., & Vos, M. D. (2025) · Journal of Neural Engineering · doi:10.1088/1741-2552/ae2b0f · Cited by 1
“One common strategy is to use Euclidean geometry, where the Euclidean mean covariance matrix of each subject captures non-task-related information, facilitating the creation of domain-invariant representations [12, 16, 17].”
InfluentialXu, M., Gao, X., Li, F., Zhang, B., & Chen, Y. (2025) · Journal of Neural Engineering · doi:10.1088/1741-2552/ae2d9a
“The specifics of each subject-adaptation method are detailed as follows: EA [37]: A Euclidean-space-based alignment approach that projects EEG data into a shared Euclidean space and employs deep learning for subsequent feature extraction.”
Zhong, Y., Qi, W., Wen, H., Yao, L., & Wang, Y. (2026) · IEEE Transactions on Cognitive and Developmental Systems · doi:10.1109/tcds.2026.3651312
“2) Within-Task EA: To achieve the alignment of the data distribution between different subjects and tasks, we employed a domain adaptation technique called EA, as described in the work by He et al. [72].”
CTSSP: A temporal–spectral-spatial joint optimization algorithm for motor imagery EEG decoding
Pan, L., Wang, K., Yi, W., Zhang, Y., Xu, M., & Dong, M. I. N. G. (2026) · Journal of Neural Engineering · doi:10.1088/1741-2552/ae34ea · Cited by 1
“To ensure consistency in EEG signal distributions, we applied Euclidean alignment (EA) [35] and whitening to both the training and testing data.”
TVGNet: A novel hybrid EEG motor imagery decoding network
Yao, Z., Xing, X., & Gao, M. (2026) · Biomedical Signal Processing and Control · doi:10.1016/j.bspc.2026.109601
Brain–Computer Interfaces (BCIs) in STEM Learning
Cano, S. (2026) · Artificial intelligence · doi:10.5772/intechopen.1013817
M, D. K., Parashiva, P. K., & Vinod, A. P. (2025) · Conference proceedings/Conference proceedings - IEEE International Conference on Systems, Man, and Cybernetics · doi:10.1109/smc58881.2025.11343716
“Euclidean Alignment (EA) is an unsupervised method that maps EEG trials into a common feature space, aligning their covariance matrix with the identity matrix [10].”
SPD-DFNet: A SPD Manifold Dual Flow Unsupervised Domain Adaptation Network for EEG Decoding
Cheng, J., Chen, J., Ran, R., et al. (2025) · IEEE International Conference on Bioinformatics and Biomedicine (BIBM) · doi:10.1109/bibm66473.2025.11356666
“The first type reduces domain differences using suitable metrics to improve model generalization, with examples like EA1 [8].”
Bidirectional Time-Frequency Pyramid Network for Enhanced Robust EEG Classification
Hong, J., Li, S., Jian, M., et al. (2025) · IEEE International Conference on Bioinformatics and Biomedicine (BIBM) · doi:10.1109/bibm66473.2025.11357122
“Furthermore, to address cross-subject variability in EEG signals, we apply Euclidean Alignment (EA) [21] as a pre-processing step.”
Adaptive Split-MMD Training for Small-Sample Cross-Dataset P300 EEG Classification
Chen, W., & Arnaud Delorme, A. (2025) · IEEE International Conference on Bioinformatics and Biomedicine (BIBM) · doi:10.1109/bibm66473.2025.11356119
“These domain shifts also make small-sample adaptation hard to stabilize [1].”
Wang, F., Chen, Y., Wang, P., Gong, A., Xu, J., & Fu, Y. (2026) · Scientific Data · doi:10.1038/s41597-026-06708-3 · Cited by 2
“Nevertheless, high noise levels, considerable variability across sessions and individuals, and limited spatial resolution make efficient decoding of handwriting-related EEG signals a major challenge 20,21 .”
Fang, H., Wang, M., Zhang, Z., Wang, Y., & Yang, Y. (2026) · IEEE Transactions on Cognitive and Developmental Systems · doi:10.1109/tcds.2026.3661567 · Cited by 1
“However, current emotion recognition methods generally focus on real-valued EEG co-variance and coherence matrices [18], [19], [21], which has limited ability to capture phase information in EEG functional networks.”
InfluentialFu, F. (2026) · Research Square · doi:10.21203/rs.3.rs-8802883/v1
Ding, Z., Zhu, H., Pan, J., Lazreq, Y., & Xu, X. (2026) · SSRN Electronic Journal · doi:10.2139/ssrn.6235622
A simple deep transfer learning model with feature alignment block for motor imagery decoding
Liu, H., Li, M., Yang, Y., & Li, Z. (2026) · Computer Methods in Biomechanics & Biomedical Engineering · doi:10.1080/10255842.2026.2627492
Peng, H., Min, X., Luo, T.-J., Peng, H., Lu, X., Li, J., & Li, R. (2026) · Entertainment Computing · doi:10.1016/j.entcom.2026.101104
FusionGen: Feature Fusion-Based Few-Shot EEG Data Generation
Chen, Y., Liu, D., Yang, X., Xu, X., Chen, B., & Wu, D. (2025) · IEEE International Conference on Computer Vision Workshops · doi:10.1109/iccvw69036.2025.00042
“To mitigate these effects and improve consistency, we adopt Euclidean alignment (EA) [7, 29], a simple yet effective whitening-Table Next, each trial is whitened by Since the second-order statistics of the aligned trials become identity, effectively reducing covariance shifts.”
Self-citationMAGCANet: A multiscale adaptive graph-convolutional attention network for MI-EEG decoding
Zhu, X., Yin, G., Shi, D., Wang, L., Yan, J., Feng, D., Wei, Z., Wang, Y., Wang, B., Tan, S., et al. (2026) · Biomedical Physics & Engineering Express · doi:10.1088/2057-1976/ae4c94
Abinaya, G., & Dinakaran, K. (2026) · SSRN Electronic Journal · doi:10.2139/ssrn.6350471
An fNIRS-Based Multi-Sensor Fusion Approach for Passenger Comfort Assessment
Zhao, Y., Yu, Y., Shi, K., Deng, R., & Cheng, H. (2025) · IEEE International Conference on Intelligent Transportation Systems (ITSC) · doi:10.1109/itsc60802.2025.11423300
An fNIRS-Based Multi-Sensor Fusion Approach for Passenger Comfort Assessment
Zhao, Y., Yu, Y., Shi, K., Deng, R., & Cheng, H. (2025) · IEEE International Conference on Intelligent Transportation Systems (ITSC) · doi:10.1109/itsc60802.2025.11423047
Liu, X., Lin, P., Guo, A., Liang, T., Li, J., & Li, Z. (2026) · Annals of Biomedical Engineering · doi:10.1007/s10439-026-04079-9 · Cited by 1
“Therefore, a series of preprocessing measures are needed to remove noise, artifacts and improve signal quality before EEG signal analysis is performed [16].”
Adaptive Human Digital Twin (aHDT) in Metaverse: A Model for Enhancing Physical Safety
Diak, V., & Diak, A. (2026) · IntechOpen eBooks · doi:10.5772/intechopen.1013725
EDAPT: towards calibration-free BCIs with continual online adaptation
Haxel, L., Kapoor, J., Ziemann, U., & Macke, J. H. (2026) · Journal of Neural Engineering · doi:10.1088/1741-2552/ae5689 · Cited by 6
“To minimize the distribution shift between the source population and the target subject, we standardize the second-order statistics of the input data [24–26].”
Feature alignment and enhancement network with guided tuning for non-stationary EEG classification
Li, D., Wang, J., Xu, J., & Zhang, Y. (2026) · Journal of Neural Engineering · doi:10.1088/1741-2552/ae512e
“In cross-subject classification, RA-MDRM [33] and EA-CSP-LDA [34] perform covariance comparisons on source and target domain data, resulting in accuracy rates that are 4.5% and 1.53% lower, respectively, than the method proposed in this paper.”
InfluentialArtificial Intelligence-Based Brain-Computer Interfaces: A Key Direction in BCI Development
Li, Z., Yan, L., Cao, Y.-Y., Wang, F., Chen, Y., Li, T., Zhao, L., Xu, J.-P., & Fu, Y. (2025) · Journal of Brain-Computer Interface · doi:10.58195/bci.2025.1.010
Meng, W., Wu, J., Hou, F., et al. (2026) · SSRN Electronic Journal · doi:10.2139/ssrn.6510367
Enhancing real-time cross-subject EEG decoding via pseudo-label tuning and adaptation
Lin, Z., Wu, Y., Dai, C., & Guo, Y. (2026) · Applied Soft Computing · doi:10.1016/j.asoc.2026.115150 · Cited by 1
Luo, T.-J. (2026) · Cognitive Neurodynamics · doi:10.1007/s11571-026-10450-2
Angkan, P., Jalali, A., Hungler, P. C., & Etemad, A. (2026) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp55912.2026.11462985
Liu, J., Zhang, J., Hu, C., et al. (2026) · Systems · doi:10.3390/systems14050476
Deep Neural Networks Based on Wavelet Drive for Motor Imagery EEG Decoding
Zhou, P., Cao, D., & Zhao, Y. (2025) · China Automation Congress (CAC) · doi:10.1109/cac67268.2025.11487551
An Investigation on Motor Imagery EEG Classification Using Random Forest Algorithm
Helen, R., Thenmozhi, T., & Vijay, M. (2026) · Lecture notes in networks and systems · doi:10.1007/978-981-95-7241-0_7
Cheng, J., Ran, R., & Fang, B. (2026) · Neural Networks · doi:10.1016/j.neunet.2026.109076
Combating Temporal Misalignment: A Phase-Guided Method for Motor Imagery EEG Classification
Liang, W., Zhao, F., Li, X., & Jin, J. (2026) · International Conference on Intelligent Perception and Autonomous Control (IPAC) · doi:10.1109/ipac69021.2026.11542582
CroSS:Cross-subject Visual Brain Decoding via Soft Alignment and Subject-aware Batch Normalization
Kim, J.-M., Choi, W., Choi, Y., Kim, S., & Kam, T. (2026) · ACM Symposium on Applied Computing · doi:10.1145/3748522.3779936
Cross-Subject Cognitive Load Recognition via Meta-Learning
Liang, S., Zhou, Y., Zhou, S., et al. (2026) · doi:10.1109/wccct69960.2026.11549701
“These methods aim to improve adaptation to EEG data from new subjects through transfer learning [8] or domain adaptation strategies [9].”
Li, J., Li, Y., Shi, W., Kang, R., & Wang, H. (2026) · Annals of Biomedical Engineering · doi:10.1007/s10439-026-04220-8 · Cited by 1
Tang, M., Li, P., Peng, Z., et al. (2026) · Biomedical Signal Processing and Control · doi:10.1016/j.bspc.2026.110786
Self-citationFCB–FLADA: A domain-adaptive deep framework for cross-subject EEG-based VR locomotion classification
Daşdemir, Y. (2026) · Biomedical Signal Processing and Control · doi:10.1016/j.bspc.2026.110869
E2EG2G: Latent node generation and Graph-to-Grid conversion for unified EEG decoding
Jin, M., Yu, L., Wei, X., & Li, K. (2026) · Information Fusion · doi:10.1016/j.inffus.2026.104580
Reducing Calibration Time in EEG Motor Imagery BCI: A Two-Stage Approach for Clinical Deployment
Grecu, C.-S., Botezatu, N.-A., & Lupu, R. G. (2026) · World Congress on Medical Physics and Biomedical Engineering, September 7 - 12, 2009, Munich, Germany · doi:10.1007/978-3-032-24045-3_15
Айтим, А., & Bagybek, Z. (2026) · IEEE International Conference on Smart Information Systems and Technologies (SIST) · doi:10.1109/sist61674.2026.11596503
“Alignment-based strategies can improve cross-subject performance by reducing distributional differences prior to classification [9].”
PGAP: Purity-Guided Active Prompting for EEG Decoding With LLMs [Research Frontier]
Luo, J., Wang, Z., Liu, D., & Wu, D. (2026) · IEEE Computational Intelligence Magazine · doi:10.1109/mci.2026.3665159
“Euclidean (EA) [49] was applied to each subject ’ s data.”
Self-citationConsumer Trait Prediction from EEG Using a Novel SpectralTraitNet for Neuromarketing Application
Matavalam, C. P., Sagila Gangadharan, K., & Vinod, A. P. (2026) · IEEE International Conference on Human-Machine Systems (ICHMS) · doi:10.1109/ichms69701.2026.11602183
“To mitigate the domain shift in cross-subject classification, we then applied Euclidean Alignment (EA) [10], an unsupervised domain adaptation method that maps trial covariance matrices into a common reference space.”
Shehu, O., Teliti, D., Karlık, B., & Kevrić, J. (2026) · The Open Neuroimaging Journal · doi:10.2174/0118744400492718260720045955
Liu, T. (2026) · SSRN Electronic Journal · doi:10.2139/ssrn.7197508
Xu, H., Sun, Y., Zhou, Z., Han, C., Huang, J., & Wang, J. (2026) · IEEE Sensors Journal · doi:10.1109/jsen.2026.3715901 · Cited by 2
“Euclidean alignment (EA) reduces intersession drift by normalizing covariance structure toward the identity matrix [34].”
Deep Cholesky space adaptation for cross-session motor imagery decoding
Wang, X., Qi, W., Yang, W., Hu, B., Shi, J., & Wang, W. (2026) · Knowledge-Based Systems · doi:10.1016/j.knosys.2026.116794
A Lightweight Subject-adaptive EEG Network for Driver Drowsiness Detection
Hou, Q. (2026) · Scientific Research Bulletin · doi:10.71052/srb2024/dgii1415
Park, Y. K., Shin, D.-S., & Kim, H.-K. (2026) · Applied Sciences · doi:10.3390/app16168042
Golovan, K. S. (2026) · SSRN Electronic Journal · doi:10.2139/ssrn.7304321
Meng, W., Hou, F., Wu, J., et al. (2026) · Biomedical Signal Processing and Control · doi:10.1016/j.bspc.2026.111272
Li, R., Bian, S., Deng, Y., Wei, X., Yang, S., Liu, Y., & Chen, J. (2026) · Frontiers in Neurorobotics · doi:10.3389/fnbot.2026.1858496
“It effectively expands the dimensionality of brain control commands and enhances system robustness, thereby providing a theoretical basis for group control in complex tasks (He and Wu, 2020; Nagender et al., 2023).”
Motor imagery classification in heterogeneous settings for brain-computer interface applications
Sharma, A., Roy, P. P., Mishra, D., & Mypati, O. (2026) · Biomedical Signal Processing and Control · doi:10.1016/j.bspc.2026.111326
Subject Identity Confounds qEEG Emotion Recognition on DEAP and DREAMER
Pandilova, E., Stojmenski, A., Chorbev, I., Petrov, M., Kitanovski, I., & Trajanov, D. (2026) · Sensors · doi:10.3390/s26175327 · Cited by 2
“Euclidean Alignment [33], which recenters each participant’s trial covariances to a common reference without requiring target labels, is a simple and widely used device that we adopt as a baseline.”
Li, J. (2026) · SSRN Electronic Journal · doi:10.2139/ssrn.7353471
Sinc-DFformer: An Intrinsically Interpretable Transformer Framework for Motor Imagery EEG Decoding
Chen, W., & Miao, M. (2026) · International Conference on Electronics and Communication, Network and Computer Technology (ECNCT) · doi:10.1109/ecnct70535.2026.11661522
“However, MI-EEG signals are characterized by low signal-to-noise ratios, strong non-stationarity, and substantial inter-subject variability, making robust feature extraction a challenging problem [1], [5], [6].”
Neural network-based dynamic functional connectivity estimation for motor imagery decoding from EEG
Zeng, X., Yu, Z., Gu, Z., Zhang, J., Li, Y., & Yu, T. (2026) · Neurocomputing · doi:10.1016/j.neucom.2026.134967
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.