Transfer Learning: A Riemannian Geometry framework with applications to Brain-Computer Interfaces
Open accessThe DOI matches its Crossref record; 37 of the 38 references it cites were found in a registry or an index.
Abstract
Objective This paper tackles the problem of transfer learning in the context of electroencephalogram (EEG)-based brain-computer interface (BCI) classification. In particular, the problems of cross-session and cross-subject classification are considered. These problems concern the ability to use data from previous sessions or from a database of past users to calibrate and initialize the classifier, allowing a calibration-less BCI mode of operation. Methods Data are represented using spatial covariance matrices of the EEG signals, exploiting the recent successful techniques based on the Riemannian geometry of the manifold of symmetric positive definite (SPD) matrices. Cross-session and cross-subject classification can be difficult, due to the many changes intervening between sessions and between subjects, including physiological, environmental, as well as instrumental changes. Here, we propose to affine transform the covariance matrices of every session/subject in order to center them with respect to a reference covariance matrix, making data from different sessions/subjects comparable. Then, classification is performed both using a standard minimum distance to mean classifier, and through a probabilistic classifier recently developed in the literature, based on a density function (mixture of Riemannian Gaussian distributions) defined on the SPD manifold. Results The improvements in terms of classification performances achieved by introducing the affine transformation are docum…
Overview
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
In the citing paper
“Experiments have shown that RA-MDRM outperformed MDRM in MI and ERP tasks [42], when compared in a TL setting.”
From Transfer Learning for Brain–Computer Interfaces: A Euclidean Space Data Alignment Approach (2020) · cites this paper as [42]
In the citing paper
“But as mentioned above, it still needs a small amount of labeled subject-specific calibration trials for ERP classification.”
From Transfer Learning for Brain–Computer Interfaces: A Euclidean Space Data Alignment Approach (2020) · cites this paper as [42]
In the citing paper
“Zanini et al. [42] proposed a novel TL approach in the Riemannian space, referred to in this paper as RA-MDRM, to improve the performance of the MDRM classifier by utilizing auxiliary data from other sessions and/or subjects when there are only a few labeled trials from a new subject.”
From Transfer Learning for Brain–Computer Interfaces: A Euclidean Space Data Alignment Approach (2020) · cites this paper as [42]
In the citing paper
“Although RA-MDRM has demonstrated promising performance in several BCI applications [42], it still has some limitations:”
From Transfer Learning for Brain–Computer Interfaces: A Euclidean Space Data Alignment Approach (2020) · cites this paper as [42]
In the citing paper
“[42] used the nontarget stimuli as the resting state in ERP, which requires that some labeled trials from the new subject must be known.”
From Transfer Learning for Brain–Computer Interfaces: A Euclidean Space Data Alignment Approach (2020) · cites this paper as [42]
2 more claims about this paper
Create a free account to see every claim other papers make about this one.
Citations
Cites
(2008) · Journal of Machine Learning Research · Cited by 33,051
Pan, S. J., & Yang, Q. (2010) · IEEE Transactions on Knowledge and Data Engineering · doi:10.1109/tkde.2009.191 · Cited by 23,387 · cited 1×
“The second improvement relates to the problem of transfer learning [30].”
BackgroundOptimizing Spatial filters for Robust EEG Single-Trial Analysis
Blankertz, B., Tomioka, R., Lemm, S., Kawanabe, M., & Muller, K.-R. (2008) · IEEE Signal Processing Magazine · doi:10.1109/msp.2008.4408441 · Cited by 2,080 · cited 2×
“First, EEG signals of a training set are transformed through frequency and/or spatial filters in order to extract discriminant features [8], [16].”
BackgroundMethodA Riemannian Framework for Tensor Computing
Pennec, X., Fillard, P., & Ayache, N. (2006) · International Journal of Computer Vision · doi:10.1007/s11263-005-3222-z · Cited by 1,550 · cited 1×
“Some of them try to find the right value through numerical procedure like deterministic line search [17], [26], simple or stochastic gradient descent [10], [31].”
BackgroundSingle-trial analysis and classification of ERP components — A tutorial
Blankertz, B., Lemm, S., Treder, M., Haufe, S., & Müller, K.-R. (2011) · NeuroImage · doi:10.1016/j.neuroimage.2010.06.048 · Cited by 1,138 · cited 4×
“This is done by supervised techniques like, for instance, Linear Discriminant Analysis (LDA) [9].”
MethodInfluentialRegularizing Common Spatial Patterns to Improve BCI Designs: Unified Theory and New Algorithms
Lotte, F., & Guan, C. (2011) · IEEE Transactions on Biomedical Engineering · doi:10.1109/tbme.2010.2082539 · Cited by 1,045 · cited 1×
“Some examples of transfer learning applied to BCI problem can be found in [15], [21], [27] and [36].”
BackgroundBhatia, R. (2009) · Princeton University Press eBooks · doi:10.1515/9781400827787 · Cited by 1,027
Multiclass Brain–Computer Interface Classification by Riemannian Geometry
Barachant, A., Bonnet, S., Congedo, M., & Jutten, C. (2012) · IEEE Transactions on Biomedical Engineering · doi:10.1109/tbme.2011.2172210 · Cited by 878 · cited 7×
“In [2] the authors used a basic classifier, named Minimum Distance to Mean (MDM), which takes into account distances on the manifold between the observations and some reference points of the classes, known as centers of mass, means, or barycenters.”
MethodInfluentialSelf-citationA Differential Geometric Approach to the Geometric Mean of Symmetric Positive-Definite Matrices
Moakher, M. (2005) · SIAM Journal on Matrix Analysis and Applications · doi:10.1137/s0895479803436937 · Cited by 702 · cited 1×
“Since P (n) is a Riemannian manifold of non-positive curvature, existence and unicity of the Riemannian mean can be proved [1], [28].”
BackgroundBCI Competition 2008 { Graz data set B
Leeb, R., & Brunner, C. (2008) · Cited by 615
Research works printed but not linked: 1
- said, 2015, Riemannian Gaussian distributions on the space of covariance matrices, arXiv 1507 01760v1
Cited by
Items 101–200 of 411
Luo, T.-J. (2024) · Expert Systems with Applications · doi:10.1016/j.eswa.2024.124673 · Cited by 21
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 can be broadly classified into three groups based on the type of transferred knowledge: instance [14, 15], feature [16–19], and classifier [20–24] transfer.”
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
Feng, Y., Li, M., Lou, S., Zheng, H., Gao, Y., & Tan, J. (2021) · Journal of Computing and Information Science in Engineering · doi:10.1115/1.4049895 · Cited by 19
Ultra Efficient Transfer Learning with Meta Update for Cross Subject EEG ClassificationPreprint
Duan, T., Chauhan, M., Shaikh, M. A., Chu, J., & Srihari, S. N. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2003.06113 · Cited by 20
Transfer Learning in Motor Imagery Brain Computer Interface: A Review
Li, M., & Xu, D. (2022) · Journal of Shanghai Jiaotong University (Science) · doi:10.1007/s12204-022-2488-4 · Cited by 21
“Zanini et al. [75] proposed a method to perform an affine transformation on the covariance matrix of each session or subject for EEG-based BCI classification, which can center the data relative to the reference covariance matrix.”
Similarity Distance Learning on SPD Manifold for Writer Independent Offline Signature Verification
Zois, E. N., Tsourounis, D., & Kalivas, D. S. (2023) · IEEE Transactions on Information Forensics and Security · doi:10.1109/tifs.2023.3333681 · Cited by 20
Facilitating 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
Liu, L., Li, J., Ouyang, R., Zhou, D., Fan, C., Liang, W. Z., Li, F., Lv, Z., & Wu, X. (2024) · Journal of Neuroscience Methods · doi:10.1016/j.jneumeth.2024.110132 · Cited by 21
Workshops of the eighth international brain–computer interface meeting: BCIs: the next frontier
Huggins, J. E., Krusienski, D. J., Vansteensel, M. J., Valeriani, D., Thelen, A., Stavisky, S. D., Norton, J. J. S., Nijholt, A., Müller-Putz, G., Kosmyna, N., et al. (2022) · Brain-Computer Interfaces · doi:10.1080/2326263x.2021.2009654 · Cited by 21
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.”
BackgroundFormulating Divergence Framework for Multiclass Motor Imagery EEG Brain Computer Interface
Kumar, S., Reddy, T. K., Arora, V., & Behera, L. (2020) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp40776.2020.9053168 · Cited by 17
“Zanini et al. [18] proposed a framework to optimize the classification performance based on an Affine transformation strategy on riemannian manifolds.”
MethodMellot, A., Collas, A., Rodrigues, P. L. C., Engemann, D., & Gramfort, A. (2023) · Imaging Neuroscience · doi:10.1162/imag_a_00040 · Cited by 16
Xiong, W. H., & Wei, Q. (2022) · PLoS ONE · doi:10.1371/journal.pone.0263641 · Cited by 16
“Several DA approaches exist in literature such as RA [35], EA [36] and PS [37], depending upon how the reference matrix is defined.”
BackgroundA 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
“[16] used the recentering to the resting state [52] instead of the geometric mean of trials to compare to EA.”
InfluentialSpatial covariance improves BCI performance for late ERPs components with high temporal variability
Aydarkhanov, R., Ušćumlić, M., Chavarriaga, R., Gheorghe, L., & Millán, J. D. R. (2020) · Journal of Neural Engineering · doi:10.1088/1741-2552/ab95eb · Cited by 16
“For example, the two distributions of covariance matrices can be matched by alignment of geometric means (see equation (2)) [9] or by translation, stretching and rotation of the whole distribution [27].”
BackgroundA Novel Transfer Support Matrix Machine for Motor Imagery-Based Brain Computer Interface
Chen, Y., Hang, W., Liang, S., Liu, X., Li, G., Wang, Q., Qin, J., & Choi, K. (2020) · Frontiers in Neuroscience · doi:10.3389/fnins.2020.606949 · Cited by 16
“The source EEG data are obtained through either instance selection or importance sampling cross-validation (Li et al., 2010; Hossain et al., 2016, 2018; Zanini et al., 2018).”
MethodLuo, T.-J. (2022) · IEEE International Conference on Bioinformatics and Biomedicine (BIBM) · doi:10.1109/bibm55620.2022.9995282 · Cited by 18
“To fairly compare and evaluate the proposed FWR-CSP-JPDA method, six state-of-the-arts methods, MDM [4], CSP-JDA [14], CSP-TJM [15], MEKT [8], SB-TA-CSP [9], METL [10], are selected for experiments.”
BackgroundMethodInfluentialReputation-Based Federated Learning Defense to Mitigate Threats in EEG Signal Classification
Zhang, Z., Li, P., Hammadi, A. Y. A., Guo, F., Damiani, E., & Yeun, C. Y. (2024) · International Conference on Computer and Automation Engineering · doi:10.1109/iccae59995.2024.10569874 · Cited by 24
“The rapid advancement of brain-computer interface (BCI) technology has led to an increasing interest in collecting and examining electroencephalogram (EEG) data [1].”
BackgroundA Lightweight Riemannian Covariance Matrix Convolutional Network for PolSAR Image Classification
Shi, J., Wang, W., Jin, H., Nie, M., & Ji, S. (2024) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2024.3428835 · Cited by 18
“Some covariance matrix-based deep learning methods [40], [41], [42], [43] in Riemannian space have been initially developed and applied to simple computer vision tasks.”
MethodChangoluisa, V., Varona, P., & Rodríguez, F. B. (2020) · IEEE Access · doi:10.1109/access.2020.3000187 · Cited by 18
“However, when looking for the optimal precision of a BCI, one may want to consider methods of electrode selection adapted to the subject and robust between sessions [21]–[24].”
Ultra Efficient Transfer Learning with Meta Update for Continuous EEG Classification Across Subjects
Duan, T., Chauhan, M., Shaikh, M. A., Chu, J., & Srihari, S. N. (2021) · doi:10.21428/594757db.6bc1ca44 · Cited by 22
“RA-MDRM [15] utilized covariance matrix from different subjects and forms a calibration less system suitable for low resource scenarios.”
MethodThe Analysis of Event-Related Potentials
Congedo, M. (2018) · Biological and medical physics, biomedical engineering · doi:10.1007/978-981-13-0908-3_4 · Cited by 17
“This approach as well as other classifiers based on Riemannian geometry have been shown to possess good accuracy, generalization and robustness properties (Congedo et al., 2017; Mayaud et al., 2016; Yger et al., 2016; Zanini et al, 2017)”
BackgroundSelf-citationPan, L., Wang, K., Huang, Y., Sun, X., Meng, J., Yi, W., Xu, M., Jung, T., & Ming, D. (2025) · Neural Networks · doi:10.1016/j.neunet.2025.107511 · Cited by 20
Kilani, S., Aghili, S. N., & Hulea, M. G. (2023) · Applied Sciences · doi:10.3390/app13106283 · Cited by 17
Cross-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
Multiview Contrastive Learning for Unsupervised Domain Adaptation in Brain–Computer Interfaces
Asgarian, S., Wang, Z., Wan, F., Wong, C. M., Liu, F., Mohsenzadeh, Y., Wang, B., & Ling, C. X. (2024) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2024.3366285 · Cited by 16
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
EEG‐Based Emotion Recognition Datasets for Virtual Environments: A Survey
Hamzah, H. A., & Abdalla, K. K. (2024) · Applied Computational Intelligence and Soft Computing · doi:10.1155/2024/6091523 · Cited by 20
Latent alignment in deep learning models for EEG decoding
Bakas, S., Ludwig, S., Adamos, D. A., Laskaris, N., Panagakis, Y., & Zafeiriou, S. P. (2025) · Journal of Neural Engineering · doi:10.1088/1741-2552/adb336 · Cited by 25
“Among the state-of-the-art methods for subject adaptation are Riemannian techniques, which perform standardization of subject-wise distributions on the semi-positive definite (SPD) manifold of EEG covariance matrices [5, 17–19].”
Abdel-Ghaffar, E. A., Wu, Y., & Daoudi, M. (2022) · IEEE Access · doi:10.1109/access.2022.3147461 · Cited by 14
“This property of Riemannian distance is called the congruence invariance, which means that, the distance between any two SPD matrices is invariant with respect to any linear invertible transform in the data space [24], [26].”
BackgroundMethodEffects of Cognitive Distraction on Upper Limb Movement Decoding From EEG Signals
Fei, W., Bi, L., Wang, J., Xia, S., Fan, X., & Guan, C. (2022) · IEEE Transactions on Biomedical Engineering · doi:10.1109/tbme.2022.3187085 · Cited by 15
“The good robustness of the Riemannian Manifold-based methods to cognitive distraction might be owed to the affine invariance of Riemannian Manifold-based features [42].”
MethodMulti-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
Mind wandering state detection during video-based learning via EEG
Tang, S., Liang, Y., & Li, Z. (2023) · Frontiers in Human Neuroscience · doi:10.3389/fnhum.2023.1182319 · Cited by 18
“Riemannian feature extraction The Riemannian geometry-based approach has achieved stateof-the-art results on various EEG-based BCI, for example, for BCI based on motor imagery (Zanini et al., 2018; Chu et al., 2020), P300 (Li et al.”
BackgroundRevisiting Riemannian geometry-based EEG decoding through approximate joint diagonalization
Kalaganis, F. P., Laskaris, N., Oikonomou, V. P., Nikopolopoulos, S., & Kompatsiaris, I. Y. (2022) · Journal of Neural Engineering · doi:10.1088/1741-2552/aca4fc · Cited by 14
“…covariance matrices (each stemming from different subjects and/or recording) so that each set of matrices has the identity matrix as its geodesic mean [28], serves as a scaling factor in order to normalize the energy of the source signals (each signal normalized by a different scaling factor).”
BackgroundKulwa, F., Li, Y., Samuel, O. W., Zhang, H., Oyemakinde, T. T., Asogbon, M. G., McEwan, A. A., & Li, G. (2024) · Biomedical Signal Processing and Control · doi:10.1016/j.bspc.2024.106446 · Cited by 14
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
Classification Methods for EEG Patterns of Imaginary Movements
Kapralov, N., Nagornova, Z. V., & Shemyakina, N. V. (2021) · Informatics and Automation · doi:10.15622/ia.2021.20.1.4 · Cited by 15
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
“[26], the trial-wise covariance matrices of each session are re-centered (i.e. aligned) with respect to a reference covariance matrix that corresponds to the resting state.”
BackgroundMethodRiemannian 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
Transfer learning and clustering analysis of epileptic EEG signals on Riemannian manifold
Xu, H., He, H. S., Xue, W., Dai, Z., & Hao, Y. (2023) · Applied Soft Computing · doi:10.1016/j.asoc.2023.110656 · Cited by 13
Unsupervised 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].”
A Transfer Learning Framework for RSVP-based Brain Computer Interface
Wei, W., Qiu, S., Ma, X., Li, D., Zhang, C., & He, H. (2020) · Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC) · doi:10.1109/embc44109.2020.9175581 · Cited by 13
“In the study of EEG transfer learning, based on the MDRM classifier, Zanini [11] promoted a Riemannian geometry transfer framework that works well on motor imagery (MI).”
MethodDomain Adaptation with Optimal Transport on the Manifold of SPD matricesPreprint
Yair, O., Dietrich, F., Talmon, R., & Kevrekidis, I. G. (2019) · arXiv (Cornell University) · doi:10.48550/arxiv.1906.00616 · Cited by 30
“In Zanini et al. [2018], the augmented covariance matrices were computed as follows.”
MethodInfluentialAghili, S. N., & Erfanian, A. (2023) · IEEE Access · doi:10.1109/access.2023.3313260 · Cited by 12
“signals [46], [47], [48], [49], [50], [51], [52], [53].”
Riemannian Channel Selection for BCI With Between-Session Non-Stationarity Reduction Capabilities
Sadatnejad, K., & Lotte, F. (2022) · IEEE Transactions on Neural Systems and Rehabilitation Engineering · doi:10.1109/tnsre.2022.3167262 · Cited by 16
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
“However, variations in data distribution due to individual differences between subjects can significantly degrade the decoding performance of MI-BCI [22].”
BackgroundMethodA 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
Demsy, 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 the preprocessing step, Zanini et al. [7] proposed Riemannian alignment, which aims to center the covariance of resting state EEG signals of all domains to a reference covariance matrix.”
MethodDoes Meta-Learning Improve EEG Motor Imagery Classification?
Wu, X., & Chan, R. H. M. (2022) · Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC) · doi:10.1109/embc48229.2022.9871035 · Cited by 16
“We have compared its prediction accuracy with other methods like RA-MDRM, CRAM, and EEGNet.”
BackgroundInfluentialChu, 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
“The resting state in MI tasks refers to moments when subjects are not performing any tasks, making it suitable as a reference state for EEG data [6, 8].”
Calibration-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].”
BackgroundMethodMishra, P. K., Jagadish, B., Kiran, M. P. R. S., Rajalakshmi, P., & Reddy, D. S. (2018) · IEEE International Conference on e-Health Networking, Applications and Services (Healthcom) · doi:10.1109/healthcom.2018.8531086 · Cited by 12
“In [6], authors proposed an affine transformation based MDRM classifier which helps in cross subject training.”
BackgroundMethodResultInfluentialDomain Adaptation Using Riemannian Geometry of Spd Matrices
Maman, G., Yair, O., Eytan, D., & Talmon, R. (2019) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp.2019.8682989 · Cited by 14
“This problem calls for domain adaptation, where a given model or representation, which is well performing in a particular domain, is adapted to a different, yet related domain [7–9].”
BackgroundTHE RIEMANNIAN MINIMUM DISTANCE TO MEANS FIELD CLASSIFIER
Congedo, M., Rodrigues, P. L. C., & Jutten, C. (2019) · HAL · doi:10.3217/978-3-85125-682-6-02 · Cited by 14
“In adaptive mode instead, the centers of mass are initialized by a database of previous users for a naive user and/or a database of the same user for a non-naive user thanks to transfer learning strategies [12, 13], then the centers of mass are adapted to the user while the BCI is operated [14].”
BackgroundSelf-citationBrain–Computer Interfaces for Spinal Cord Injury Rehabilitation
Merante, A., Zhang, Y., Kumar, S., & Nam, C. S. (2020) · Cognitive science and technology · doi:10.1007/978-3-030-34784-0_16 · Cited by 13
Yu, H., Mu, Q., Liu, C., Wang, S., & Sun, J. (2025) · Neural Regeneration Research · doi:10.4103/nrr.nrr-d-25-00217 · Cited by 20
Harneshing the Potential of EEG in Neuromarketing with Deep Learning and Riemannian Geometry
Georgiadis, K., Kalaganis, F. P., Oikonomou, V. P., Nikolopoulos, S., Laskaris, N., & Kompatsiaris, I. Y. (2023) · Lecture notes in computer science · doi:10.1007/978-3-031-43075-6_3 · Cited by 11
Semi-supervised multi-source transfer learning for cross-subject EEG motor imagery classification
Zhang, F., Wu, H., & Guo, Y. (2024) · Medical & Biological Engineering & Computing · doi:10.1007/s11517-024-03032-z · Cited by 13
“For example, Zanini et al. introduced Riemann alignment as a data preprocessing strategy in [70].”
Retain and Adapt: Online Sequential EEG Classification With Subject Shift
Duan, T., Wang, Z., Shen, L., Doretto, G., Adjeroh, D., Li, F., & Tao, C. (2024) · IEEE Transactions on Artificial Intelligence · doi:10.1109/tai.2024.3385390 · Cited by 15
Xu, 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
A 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
Rammy, S. A., Abbas, W., Mahmood, S. S., Riaz, H., Rehman, H. U., Abideen, R. Z. U., Aqeel, M., & Zhang, W. (2020) · Journal of Applied Biomedicine · doi:10.1016/j.bbe.2020.12.004 · Cited by 13
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
“…the covariance matrices of EEG signals are all symmetric positive definite (SPD) matrices and lie in the Riemannian manifold [11], Zanini et al. [12] proposed a Riemannian Alignment (RA) approach which uses the reference matrix to align the covariance matrices of EEG trials in Riemannian…”
MethodTransfer Learning for the Riemannian Tangent Space: Applications to Brain-Computer Interfaces
Bleuzé, A., Mattout, J., & Congedo, M. (2021) · International Conference on Engineering and Emerging Technologies (ICEET) · doi:10.1109/iceet53442.2021.9659607 · Cited by 15
“Then, all trials are recentered (parallel transported to the identity (6)), which, as discussed in the introduction, is a common step to all transfer learning procedures.”
InfluentialSelf-citationMental Workload Classification Method Based on EEG Cross-Session Subspace Alignment
Qu, H., Zhang, M., & Pang, L. (2022) · Mathematics · doi:10.3390/math10111875 · Cited by 11
“Non-stationary EEG leads to a decrease in generalization capability and classification accuracy if the classifier only used the data from one previous session as its training set [27].”
BackgroundCusinato, R., Gross, S., Bainier, M., Janz, P., Schoenenberger, P., & Redondo, R. L. (2024) · Journal of Neuroscience Methods · doi:10.1016/j.jneumeth.2024.110155 · Cited by 12
“…strategies to minimize such limitations (Malafeev et al., 2018; Miladinovi ´ c et al., 2019): Cross-subject learning, for example, builds a classifier based on data from past subjects (Zanini et al., 2018), and population subgrouping can be deployed in large datasets (Nasiri and Clifford, 2021).”
Riemannian 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
Multiclass Classification Framework of Motor Imagery EEG by Riemannian Geometry Networks
Shi, Y., Jiang, A., Zhong, J., Li, M., & Zhu, Y. (2024) · IEEE Journal of Biomedical and Health Informatics · doi:10.1109/jbhi.2024.3496757 · Cited by 12
Boosting automated sleep staging performance in big datasets using population subgrouping
Nasiri, S., & Clifford, G. D. (2021) · SLEEP · doi:10.1093/sleep/zsab027 · Cited by 14
Xu, L., Ma, Z., Meng, J., Xu, M., Jung, T., & Ming, D. (2021) · Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC) · doi:10.1109/embc46164.2021.9629529 · Cited by 15
“A possible solution to alleviate the impact of cross-dataset variability is to apply the whitening procedure [15] for each subject.”
BackgroundActive Brain-Computer Interfacing for Healthy Users
Shishkin, S. L. (2022) · Frontiers in Neuroscience · doi:10.3389/fnins.2022.859887 · Cited by 11
“Further development of transfer learning (e.g., Zanini et al., 2017; Fahimi et al., 2019; Dehghani et al., 2021) and more recent meta-learning (Li et al., 2021; Bhosale et al., 2022; Wei et al., 2022) approaches make possible applying a classifier trained on large multisubject datasets to the data…”
BackgroundMulti-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
MI 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
Huang, Z., Zheng, W., Wu, Y., & Wang, Y. (2020) · Journal of Neuroscience Methods · doi:10.1016/j.jneumeth.2020.108855 · Cited by 10
Eder, M., Xu, J., & Grosse‐Wentrup, M. (2024) · Journal of Neural Engineering · doi:10.1088/1741-2552/ad6793 · Cited by 18
“In particular, extensions of Riemannian pipelines for transfer learning have been developed but have not been incorporated [38].”
Liu, D. H., Hsieh, J.-C., Alawieh, H., Kumar, S., Iwane, F., Pyatnitskiy, I., Ahmad, Z. J., Wang, H., & Millán, J. D. R. (2024) · Journal of Neural Engineering · doi:10.1088/1741-2552/ad9edf · Cited by 10
Sarikaya, M. A., & İnce, G. (2025) · PeerJ Computer Science · doi:10.7717/peerj-cs.2649 · Cited by 11
Manifold-based multi-branch transfer learning for MI-EEG decoding
Li, H., An, J., Juan, R., Benko, T. P., Perc, M., Dang, W., & Gao, Z. (2025) · Chaos Solitons & Fractals · doi:10.1016/j.chaos.2025.116326 · Cited by 10
Zhu, J. L., Xu, G., Lin, Z., Long, J., Zhou, T., Sheng, B., & Yang, X. (2025) · IEEE Transactions on Automation Science and Engineering · doi:10.1109/tase.2025.3604283 · Cited by 12
“1) Classical classifiers for MI-based BCIs, i.e. , CSP [39], EA [40], MDRM [41], RA [42], EEGNet [43] and DeepConvNet [38].”
Cross-Subject & Cross-Dataset Subject Transfer in Motor Imagery BCI systems
Zaremba, T., & Atyabi, A. (2022) · International Joint Conference on Neural Networks (IJCNN) · doi:10.1109/ijcnn55064.2022.9891904 · Cited by 15
“The study reported 71% classification accuracy for subject transfer using BCI Comp. IV dataset 4a. Zanini et al., [38] proposed a mechanism for subject election in a subject transfer study.”
Yan, W., He, B., & Zhao, J. H. (2023) · Frontiers in Neuroscience · doi:10.3389/fnins.2023.1161511 · Cited by 9
“, 2008; Li and Guan, 2008), and unsupervised (Blumberg et al., 2007; Vidaurre et al., 2011; Kindermans et al., 2014; Zanini et al., 2017).”
BackgroundSpectrum-Weighted Tensor Discriminant Analysis for Motor Imagery-Based BCI
Huang, S., Chen, Y., Wang, T., & Ma, T. (2020) · IEEE Access · doi:10.1109/access.2020.2995302 · Cited by 10
“, [6], [36]) have reached state-of-the-art performances on multiple BCI problems; deep learning methods (e.”
BackgroundGuidelines to use Transfer Learning for Motor Imagery Detection: an experimental study
Bougrain, L., Rimbert, S., Rodrigues, P. L. C., Canron, G., & Lotte, F. (2021) · International IEEE/EMBS Conference on Neural Engineering · doi:10.1109/ner49283.2021.9441254 · Cited by 10
“, Riemannian transfer learning algorithms [7], [8].”
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
Said, S., Heuveline, S., & Mostajeran, C. (2022) · IEEE Transactions on Information Theory · doi:10.1109/tit.2022.3199479 · Cited by 9
“They have found successful applications in fields such as brain-computer interface analysis and artificial intelligence [4], [5].”
BackgroundSelf-citationReducing the Dimensionality of SPD Matrices with Neural Networks in BCI
Peng, Z., Li, H., Zhao, D., & Pan, C. (2023) · Mathematics · doi:10.3390/math11071570 · Cited by 12
An efficient domain adaptation framework for cross-subject EEG classification
Luo, T.-J. (2023) · IEEE International Conference on Bioinformatics and Biomedicine (BIBM) · doi:10.1109/bibm58861.2023.10385973 · Cited by 8
“Researchers treated the samples as multivariate time-series, and perform alignment on the covariance matrices, since the covariance represents the samples distribution, such as Riemannian alignment (RA) [6] and Euclidean alignment (EA) [7] method.”
MethodThree-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].”
Wang, H., Han, H., Gan, J. Q., & Wang, H. (2024) · IEEE Journal of Biomedical and Health Informatics · doi:10.1109/jbhi.2024.3463737 · Cited by 11
“She et al. [14] extend the RA method, which minimizes the marginal probability distribution on the Riemannian manifold and subsequently transfers the manifold features with conditional probability distribution adaptation.”
RGB-D Cameras and Brain–Computer Interfaces for Human Activity Recognition: An Overview
Iadarola, G., Mengarelli, A., Iarlori, S., Monteriù, A., & Spinsante, S. (2025) · Sensors · doi:10.3390/s25206286 · Cited by 8
“For achieving this objective, some promising solutions are currently available, allowing also to address the cross-session issue [165].”
Multi-metric domain adaptation for unsupervised transfer learning
Yang, H.-W., He, H., Li, T., Bai, Y., & Zhang, W.-Z. (2020) · IET Image Processing · doi:10.1049/iet-ipr.2019.1434 · Cited by 8
“Many metric learning-based domain adaptation studies are currently emerging, such as Riemann distance-based transfer learning [11, 12] and other metric-based methods [13, 14].”
BackgroundRuan, Y., Du, M., & Ni, T. (2022) · Frontiers in Psychology · doi:10.3389/fpsyg.2022.899983 · Cited by 11
“For example, Zanini et al. (2018) proposed a transfer learning model using Riemannian geometry to handle across-session and across-subject classification tasks in brain-computer interface (BCI).”
MethodSegment alignment based cross-subject motor imagery classification under fading data
Wan, Z., Yang, R., Huang, M., Alsaadi, F. E., Sheikh, M. M., & Wang, Z. (2022) · Computers in Biology and Medicine · doi:10.1016/j.compbiomed.2022.106267 · Cited by 8
Guo, S., Wang, Y., Liu, Y., Zhang, X., & Tang, B. (2025) · Biomedical Signal Processing and Control · doi:10.1016/j.bspc.2025.107756 · Cited by 7
Wu, H., & Wu, D. (2022) · Brain Science Advances · doi:10.26599/bsa.2022.9050001 · Cited by 7
“[19] proposed Riemannian alignment (RA), which is a transfer learning approach in the Riemannian space.”
BackgroundRaoof, I., & Gupta, M. K. (2023) · Multimedia Tools and Applications · doi:10.1007/s11042-023-15900-1 · Cited by 8
“To deal with the cross-subject/ session MI EEG data, authors in [175] proposed a symmetric heterogenous TL-based Riemannian geometry.”
Signal alignment for cross-datasets in P300 brain-computer interfaces
Song, M., Gwon, D., Jun, S. C., & Ahn, M. (2024) · Journal of Neural Engineering · doi:10.1088/1741-2552/ad430d · Cited by 7
“Also, various types of means can be used, such as the Euclidean mean for Euclidean alignment [42], the Riemannian mean for Riemannian alignment [43], and the log-Euclidean mean for log-Euclidean alignment [41].”
MethodMeenakshinathan, J., Gupta, V., Reddy, T. K., Behera, L., & Sandhan, T. (2024) · Medical & Biological Engineering & Computing · doi:10.1007/s11517-024-03137-5 · Cited by 7
“One such Riemannian alignment approach to center the EEG covariance matrices towards a common reference was proposed by Zanini et al. [9].”
A novel multi-morphological representation approach for multi-source EEG signals
Gao, Y., Liu, Y., Meng, M., Fang, F. Z., Houston, M. J., & Zhang, Y. (2024) · Neurocomputing · doi:10.1016/j.neucom.2024.129010 · Cited by 10
Metadata
- Authors
- Zanini, P., Congedo, M., Jutten, C., Said, S., & Berthoumieu, Y.
- Affiliations
- Centre National de la Recherche Scientifique (FR)
- Venue
- IEEE Transactions on Biomedical Engineering
- Published
- 2018-05
- Volume
- 65
- Issue
- 5
- Pages
- 1107-1116
- ISSN
- 0018-9294 (ISSN-L)
- 1558-2531
- OpenAlex
- W2746829572
- Cited by
- 412
- References
- 38
- Citations by year
- 2026: 48 · 2025: 53 · 2024: 78 · 2023: 53 · 2022: 62 · 2021: 37 · 2020: 49 · 2019: 25 · 2018: 7
- FWCI
- 12.72
- Type
- journal article
- Language
- en
- Open access
- green
- hal.science
- Locations
- IEEE Transactions on Biomedical Engineering (publishedVersion)
- PubMed (publishedVersion)
- HAL (submittedVersion)
- HAL (submittedVersion, other-oa)
- Funders
- ERC (2012-ERC-AdG-320684)
- 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
- transfer learning, brain-computer interface, EEG classification, cross-subject classification, Riemannian geometry, symmetric positive definite matrices, covariance matrix, affine transformation, minimum distance classifier
- MeSH terms
- Machine Learning, Electroencephalography, Humans, Models, Theoretical, 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.
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.