Signal Processing Approaches to Minimize or Suppress Calibration Time in Oscillatory Activity-Based Brain–Computer Interfaces
Open accessThe DOI matches its Crossref record; 86 of the 88 references it cites were found in a registry or an index.
Abstract
One of the major limitations of brain-computer interfaces (BCI) is their long calibration time, which limits their use in practice, both by patients and healthy users alike. Such long calibration times are due to the large between-user variability and thus to the need to collect numerous training electroencephalography (EEG) trials for the machine learning algorithms used in BCI design. In this paper, we first survey existing approaches to reduce or suppress calibration time, these approaches being notably based on regularization, user-to-user transfer, semi-supervised learning and a priori physiological information. We then propose new tools to reduce BCI calibration time. In particular, we propose to generate artificial EEG trials from the few EEG trials initially available, in order to augment the training set size. These artificial EEG trials are obtained by relevant combinations and distortions of the original trials available. We propose three different methods to do so. We also propose a new, fast and simple approach to perform user-to-user transfer for BCI. Finally, we study and compare offline different approaches, both old and new ones, on the data of 50 users from three different BCI data sets. This enables us to identify guidelines about how to reduce or suppress calibration time for BCI.
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
“In addition, the BCI system requires a long calibration phase before each time it is used because, to achieve a good performance, every subject’s BCI system needs to be trained by their own EEG signals and cannot use other’s EEG signals [11].”
From Transfer Learning Algorithm of P300-EEG Signal Based on XDAWN Spatial Filter and Riemannian Geometry Classifier (2020) · cites this paper as [11]
In the citing paper
“In addition, the BCI system requires a long calibration phase before each time it is used because, to achieve a good performance, every subject’s BCI system needs to be trained by their own EEG signals and cannot use other’s EEG signals [11].”
From Transfer Learning Algorithm of P300-EEG Signal Based on XDAWN Spatial Filter and Riemannian Geometry Classifier (2020) · cites this paper as [11]
In the citing paper
“Classification must generalize across subjects, since per-subject calibration delays setup and limits scalability [4]”
From ERP-XTTN: interpretable prototype-guided cross-attention for cross-subject ERP classification (2026) · cites this paper as [4]
Citations
Cites
Caruana, R. (1997) · Machine Learning · doi:10.1023/a:1007379606734 · Cited by 6,636 · cited 1×
“…guide the optimization algorithms towards good solutions, similar to that of the other users, thus enabling a better learning from little data. b) Multi-task learning: Multi-task learning consists of learning simultaneously multiple tasks that share some constraints or prior information [37].”
BackgroundSemi-Supervised Learning Literature Survey
Zhu, X. (2005) · Minds at UW (University of Wisconsin) · Cited by 4,505
A well-conditioned estimator for large-dimensional covariance matrices
Ledoit, O., & Wolf, M. (2004) · Journal of Multivariate Analysis · doi:10.1016/s0047-259x(03)00096-4 · Cited by 3,046 · cited 3×
“More precisely, it is known that when covariance matrices are estimated from too little training data, their largest and smallest eigen values will be, respectively, over and under estimated [32].”
BackgroundMethodInfluentialA review of classification algorithms for EEG-based brain–computer interfaces
Lotte, F., Congedo, M., Lécuyer, A., Lamarche, F., & Arnaldi, B. (2007) · Journal of Neural Engineering · doi:10.1088/1741-2560/4/2/r01 · Cited by 2,665 · cited 2×
“The LDA classifier uses a linear hyperplane to separate feature vectors from two classes [10].”
MethodSelf-citationOptimizing 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 3×
“Since the variance of a band-pass filtered signal corresponds to the band-power of the signal in that band, CSP optimizes spatial filters that lead to optimally discriminant band-power features [29].”
BackgroundMethodInfluentialMotor imagery and direct brain-computer communication
Pfurtscheller, G., & Neuper, C. (2001) · Proceedings of the IEEE · doi:10.1109/5.939829 · Cited by 1,990 · cited 1×
“As an example, during left hand motor imagery (i.e., the imagination of a left hand movement), the EEG band power in the (cid:2) (8–12 Hz) and (16–24 Hz) frequency bands will decrease in the right sensorimotor cortex [30].”
BackgroundA tutorial on Principal Components Analysis
Smith, L. I. (2002) · Otago University Research Archive (University of Otago) · Cited by 1,636
Filter Bank Common Spatial Pattern Algorithm on BCI Competition IV Datasets 2a and 2b
Ang, K. K., Chin, Z. Y., Wang, C., Guan, C., & Zhang, H. (2012) · Frontiers in Neuroscience · doi:10.3389/fnins.2012.00039 · Cited by 1,264 · cited 1×
“…and classifier optimization from limited training data), several algorithms being available to optimize such bands with CSP and LDA; see, e.g., [75]–[81]; 2) outlier/ artifact removal in small sample size settings, as when little data is available, a single outlier (artefacts being…”
MethodSingle-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 1×
“Since this is a simple, computationally efficient and parameter free method, we advocate that it should become a standard tool for BCI design, including for oscillatory activity-based BCI (it is already becoming a rather standard tool for ERP classification [74]).”
MethodAn EEG-based brain-computer interface for cursor control
Wolpaw, J. R., McFarland, D. J., Neat, G. W., & Forneris, C. A. (1991) · Electroencephalography and Clinical Neurophysiology · doi:10.1016/0013-4694(91)90040-b · Cited by 1,114 · cited 2×
“, [2] and [3].”
Research works printed but not linked: 3
- tu, 0, Dynamical ensemble learning with model-friendly classifiers for domain adaptation, Proc IEEE ICPR, 1181
- lu, 0, Boosting-based subject-independent brain computer interface, Proc ICPR
- meng, 2014, Improved semi-supervised adaptation for small training data set in brain-computer interface, IEEE J Biomed Health Inform, 18
Cited by
Items 101–200 of 328
Wei, W., Qiu, S., Zhang, Y., Mao, J., & He, H. (2022) · Journal of Neural Engineering · doi:10.1088/1741-2552/ac5eb7 · Cited by 24
“Currently, for paradigms such as MI and P300, zerocalibrationmethods can be classified as ensemble and pooled methods [31].”
Mind the Traps! Design Guidelines for Rigorous BCI Experiments
Jeunet, C., Debener, S., Lotte, F., Mattout, J., Scherer, R., & Zich, C. (2018) · CRC Press eBooks · doi:10.1201/9781351231954-32 · Cited by 33
“As far as LDA is concerned, when little data is available, regularized LDA (notably shrinkage LDA) should be used preferably to LDA, as they were shown to lead to higher accuracy (Blankertz et al, 2011; Lotte 2015b).”
BackgroundMethodSelf-citationA Brute-Force CNN Model Selection for Accurate Classification of Sensorimotor Rhythms in BCIs
Abibullaev, B., Dolzhikova, I., & Zollanvari, A. (2020) · IEEE Access · doi:10.1109/access.2020.2997681 · Cited by 25
EEG data augmentation using Wasserstein GAN
Bouallegue, G., & Djemal, R. (2020) · International Conference on Sciences and Techniques of Automatic Control and Computer Engineering · doi:10.1109/sta50679.2020.9329330 · Cited by 37
“In addition, Lotte [4] used combinations of distortions of original EEG segments to generate artificial ones.”
MethodMirzabagherian, H., Menhaj, M. B., Suratgar, A. A., Talebi, N., Sardari, M., & Sajedin, A. (2023) · Computers in Biology and Medicine · doi:10.1016/j.compbiomed.2023.107159 · Cited by 26
Spatial filtering based on Riemannian distance to improve the generalization of ErrP classification
Cruz, A., Pires, G., & Nunes, U. (2021) · Neurocomputing · doi:10.1016/j.neucom.2021.10.078 · Cited by 22
Challenges of neural interfaces for stroke motor rehabilitation
Vidaurre, C., Irastorza-Landa, N., Sarasola-Sanz, A., Insausti-Delgado, A., Ray, A. M., Bibián, C., Helmhold, F., Mahmoud, W. J., Ortego-Isasa, I., López‐Larraz, E., et al. (2023) · Frontiers in Human Neuroscience · doi:10.3389/fnhum.2023.1070404 · Cited by 32
“The working hypothesis is that Brain-computer (BCIs) or Brain-machine Interfaces (BMIs) might strengthen the ipsilesional sensorimotor loop and foster neuroplasticity that facilitates motor recovery by triggering Heabbian plasticity (Birbaumer and Cohen, 2007; Jackson and Zimmermann, 2012; López-Larraz et al., 2018c).”
BackgroundInfluentialHigashi, H., Rutkowski, T. M., Tanaka, T., & Tanaka, Y. (2016) · IEEE Journal of Selected Topics in Signal Processing · doi:10.1109/jstsp.2016.2599297 · Cited by 23
“To alleviate such problems, unsupervised dimensionality reduction techniques, such as principal component analysis (PCA) [11], regularizations [12]–[16], calibrations with data from different users or sessions [10], [17], [18], and the generation of artificial signals [13] are often used.”
MethodReducing the Calibration Time in Somatosensory BCI by Using Tactile ERD
Yao, L., Jiang, N., Mrachacz‐Kersting, N., Zhu, X., Farina, D., & Wang, Y. (2022) · IEEE Transactions on Neural Systems and Rehabilitation Engineering · doi:10.1109/tnsre.2022.3184402 · Cited by 23
Real World BCI: Cross-Domain Learning and Practical Applications
Gordon, S. M., Jaswa, M., Solon, A. J., & Lawhern, V. J. (2017) · Proceedings of the 2017 ACM Workshop on An Application-oriented Approach to BCI out of the laboratory · doi:10.1145/3038439.3038444 · Cited by 20
Guo, M., Han, X., Liu, H., Zhu, J., Zhang, J., Bai, Y., & Ni, G. (2025) · Annals of the New York Academy of Sciences · doi:10.1111/nyas.15288 · Cited by 27
“Following the method described by Lotte, 39 the method involves representing N From these segments, new samples can be generated by concatenating segments randomlyselectedfromdifferenttrainingtrials.”
Reducing Response Time in Motor Imagery Using A Headband and Deep Learning
García-Moreno, F. M., Bermúdez-Edo, M., Garrido, J. L. G., & Rodríguez‐Fórtiz, M. J. (2020) · Sensors · doi:10.3390/s20236730 · Cited by 24
“As mentioned above, traditional EEG devices are high-cost, invasive (electrodes implanted in the brain), intrusive and require user-training and long calibration time [19,21].”
BackgroundEEG-based emergency braking intention detection during simulated driving
Liang, X., Yu, Y., Liu, Y., Liu, K., Liu, Y., & Zhou, Z. (2023) · BioMedical Engineering OnLine · doi:10.1186/s12938-023-01129-4 · Cited by 24
“Related studies have shown that the shrinkage LDA classifier outperforms the standard LDA classifier in EEG classification tasks [27, 34, 35].”
MethodTransfer Learning for SSVEP-based BCI Using Riemannian Similarities Between Users
Kalunga, E. K., Chevallier, S., & Barthélemy, Q. (2018) · European Signal Processing Conference (EUSIPCO) · doi:10.23919/eusipco.2018.8553441 · Cited by 27
“Nonetheless, this calibration is a source of fatigue, frustration and could induce loss of performance for the real task to come [6], [7].”
BackgroundChoo, S., Park, H., Jung, J., Flores, K., & Nam, C. S. (2024) · Neural Networks · doi:10.1016/j.neunet.2024.106665 · Cited by 25
Giles, J., Ang, K. K., Phua, K. S., & Arvaneh, M. (2022) · Frontiers in Neuroergonomics · doi:10.3389/fnrgo.2022.837307 · Cited by 23
“CSP-based BCI commonly require 15–30 min of calibration, collecting 40, 60 or even 80 trials per class to train the BCI (Lotte, 2015; Proesmans, 2016).”
Subdomain Adversarial Network for Motor Imagery EEG Classification Using Graph Data
Li, X., Tang, X., Qiu, S., Deng, X., Wang, H., & Tian, Y. (2023) · IEEE Transactions on Emerging Topics in Computational Intelligence · doi:10.1109/tetci.2023.3301385 · Cited by 23
“All of the above methods have achieved impressive performances, but require access to sufficiently labeled data to train subject-specific BCI models [19], [20], [21].”
BackgroundA Neuroergonomics Approach to Measure Pilot’s Cognitive Incapacitation in the Real World with EEG
Dehais, F., Somon, B., Mullen, T. R., & Callan, D. E. (2020) · Advances in intelligent systems and computing · doi:10.1007/978-3-030-51041-1_16 · Cited by 19
Robust Motor Imagery Classification Using Sparse Representations and Grouping Structures
Oikonomou, V. P., Nikolopoulos, S., & Kompatsiaris, I. Y. (2020) · IEEE Access · doi:10.1109/access.2020.2997116 · Cited by 20
“More information on regularized CSP could be found in [31].”
BackgroundAdaptive 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
“Reducing or eliminating the calibration time is of great importance [7], particularly for disabled subjects with limited motor functions.”
Abibullaev, B., Kunanbayev, K., & Zollanvari, A. (2022) · IEEE Transactions on Human-Machine Systems · doi:10.1109/thms.2022.3189576 · Cited by 19
“Nevertheless, machine learning techniques that could achieve this goal are categorized as methods for calibration time reduction or methods for calibration time suppression (or zero-calibration methods) [71].”
MethodLiang, S., Xuan, C., Hang, W., Lei, B., Wang, J., Qin, J., Choi, K., & Zhang, Y. (2023) · IEEE Transactions on Neural Systems and Rehabilitation Engineering · doi:10.1109/tnsre.2023.3300961 · Cited by 23
“To examine the efficacy of the proposed ISCS EEG data augmentation method, we compared it with the following mainstream methods: 1) noise addition [35]; 2) sliding window [36]; 3) segmentation and recombination in time-frequency domain [33], and 4) Mixup [34].”
MethodInfluentialMeinel, A., Castaño-Candamil, S., Blankertz, B., Lotte, F., & Tangermann, M. (2018) · Neuroinformatics · doi:10.1007/s12021-018-9396-7 · Cited by 20
Self-citationZhou, B., Wu, X., Ruan, J., Zhao, L. V., & Zhang, L. (2019) · Biomedical Signal Processing and Control · doi:10.1016/j.bspc.2019.01.017 · Cited by 19
Cai, Z., Wang, L., Guo, M., Xu, G., Guo, L., & Li, Y. (2022) · International Journal of Neural Systems · doi:10.1142/s0129065722500058 · Cited by 21
Xiong, W. H., & Wei, Q. (2022) · PLoS ONE · doi:10.1371/journal.pone.0263641 · Cited by 16
“Lotte proposed three artificial data generation techniques [22, 23], i.”
BackgroundEEG-TCNTransformer: A Temporal Convolutional Transformer for Motor Imagery Brain–Computer Interfaces
Nguyen, A. H. P., Oyefisayo, O., Pfeffer, M. A., & Ling, S. H. (2024) · Signals · doi:10.3390/signals5030034 · Cited by 26
Architectural Review of Co-Adaptive Brain Computer Interface
Singh, A. S. G., Lal, S. P., & Guesgen, H. W. (2017) · Asia-Pacific World Congress on Computer Science and Engineering (APWC on CSE) · doi:10.1109/apwconcse.2017.00044 · Cited by 18
“Lotte [29] in his paper survey all existing approaches to suppress calibration time and also proposed some new methods to reduce calibration time.”
BackgroundSong, Y., Yang, L., Jia, X., & Xie, L. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2102.04456 · Cited by 32
“A trial was segmented into eight segments, as suggested in [54].”
Time Majority Voting, a PC-Based EEG Classifier for Non-expert Users
Dou, G., Zhou, Z., & Qu, X. (2022) · Lecture notes in computer science · doi:10.1007/978-3-031-17618-0_29 · Cited by 21
A 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
“One well-known technique is the Riemannian Alignments (RA) [ 3, 26, 35, 52], which aims to reduce the distance between the subjects’ covariance matrices in the Riemannian manifold [49].”
Neurophysiologically Meaningful Motor Imagery EEG Simulation With Applications to Data Augmentation
Galván, C. M., Spies, R. D., Milone, D. H., & Peterson, V. (2024) · IEEE Transactions on Neural Systems and Rehabilitation Engineering · doi:10.1109/tnsre.2024.3417311 · Cited by 16
“We briefly describe them below: • Segmentation and recombination (SR): training trials (2 s) were splitted into 8 segments and new trials were created by concatenating random analogous segments from trials of the same class, as in [51].”
MethodSocial anxiety prediction based on ERP features: A deep learning approach
Tian, X., Zhu, L., Zhang, M., Wang, S., Lu, Y., Xu, X., Jia, W., Zheng, Y., & Song, S. (2024) · Journal of Affective Disorders · doi:10.1016/j.jad.2024.09.006 · Cited by 19
Spatial-Temporal Mamba Network for EEG-Based Motor Imagery Classification
Yang, X., & Jia, Z. (2024) · Lecture notes in computer science · doi:10.1007/978-981-96-0821-8_28 · Cited by 33
Miao, F., Li, Y., & Zhao, X. (2022) · IEEE Transactions on Neural Systems and Rehabilitation Engineering · doi:10.1109/tnsre.2022.3213848 · Cited by 16
“…class data, e.g., Sun et al. [24] proposed augmenting the minority class by geometric transformations such as flips and translations, F. Lotte [25] proposed amplification of the minority class samples using segmentation and recombination of the original signal, J. Fan, et al. [26] generated…”
BackgroundVedernikov, A., Kumar, P., Chen, H., Seppänen, T., & Li, X. (2024) · IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) · doi:10.1109/cvprw63382.2024.00475 · Cited by 24
“In contrast, the Segmentation and Recombination (S&R) technique [22], applied in the temporal domain, effectively augments signal data by preserving its unique patterns (Fig.”
MethodMotor task-to-task transfer learning for motor imagery brain-computer interfaces
Gwon, D., & Ahn, M. (2024) · NeuroImage · doi:10.1016/j.neuroimage.2024.120906 · Cited by 20
“1C) (Dai et al., 2020; Lotte, 2015; Yang et al., 2021).”
Yadav, H., & Maini, S. (2025) · Computers in Biology and Medicine · doi:10.1016/j.compbiomed.2025.110937 · Cited by 21
Fan, Y., Mao, H., & Li, Q. (2023) · IEEE Journal of Biomedical and Health Informatics · doi:10.1109/jbhi.2023.3248139 · Cited by 15
“In addition, it shortens the response time of BCI applications, the preparation time of the acquisition device, and the calibration time of subjects [32].”
BackgroundData Augmentation for P300-based Brain-Computer Interfaces Using Generative Adversarial Networks
Kunanbayev, K., Abibullaev, B., & Zollanvari, A. (2021) · International Winter Conference on Brain-Computer Interface (BCI) · doi:10.1109/bci51272.2021.9385317 · Cited by 15
“…number of conventional data augmentation techniques have been already used: signal segmentation and recombination, and artificial trail generation [8], [9], rotational data augmentation [10], upsampling minority class, change in morphology, dataset-to-dataset transfer [9]; label-based data…”
MethodKo, W., Jeon, E., Yoon, J. S., & Suk, H. (2022) · Scientific Reports · doi:10.1038/s41598-022-08490-9 · Cited by 20
“In essence, this work is inspired by Lotte’s work 21 , which demonstrated the use and effects of artificial EEG samples in constructing a BCI system.”
BackgroundMethodInfluentialFunctional Mapping of the Brain for Brain–Computer Interfacing: A Review
SINGH, S. P., Mishra, S., Gupta, S., Padmanabhan, P., Lu, J. X., Colin, T. K. A., Tsai, Y. T., Kejia, T., Sankarapillai, P., Mohan, A., et al. (2023) · Electronics · doi:10.3390/electronics12030604 · Cited by 17
Dual-TSST: A Dual-Branch Temporal-Spectral-Spatial Transformer Model for EEG Decoding
Li, H., Zhang, H., & Chen, Y. (2025) · IEEE Journal of Biomedical and Health Informatics · doi:10.1109/jbhi.2025.3577611 · Cited by 27
“Following the guidelines set forth in references [16], [31] Here, to capture local details in the temporal dimension as much as possible, the time convolution kernel size is set to be relatively small.”
MethodBrainWave-Scattering Net: a lightweight network for EEG-based motor imagery recognition
Barmpas, K., Panagakis, Y., Adamos, D. A., Laskaris, N., & Zafeiriou, S. P. (2023) · Journal of Neural Engineering · doi:10.1088/1741-2552/acf78a · Cited by 18
“However, the scarcity of labelled EEG data due to the lengthy acquisition process (e.g. [26–28]) prohibits the training of deep BCI personalized networks leading often to overfitting.”
BackgroundCen, H., Zhao, M., Cui, K., Tian, F., Zhao, Q., & Hu, B. (2024) · Biomedical Signal Processing and Control · doi:10.1016/j.bspc.2024.106957 · Cited by 14
Jiménez-Guarneros, M., Fuentes-Pineda, G., & Grande-Barreto, J. (2024) · IEEE Transactions on Computational Social Systems · doi:10.1109/tcss.2024.3519300 · Cited by 22
“Then, UDA is used to train a classifier model that learns a shared feature space between the source and target domains, reducing differences in their data distributions [20].”
A self-supervised data augmentation strategy for EEG-based emotion recognition
Qiao, Y., & Zhao, Q. (2025) · Information Fusion · doi:10.1016/j.inffus.2025.103279 · Cited by 13
Data augmentation in Riemannian space for Brain-Computer Interfaces
Kalunga, E. K., Chevallier, S., & Barthélemy, Q. (2015) · HAL · Cited by 18
Design and Implementation of an Asynchronous BCI System With Alpha Rhythm and SSVEP
Zhang, L., Wu, X., Guo, X., Liu, J., & Zhou, B. (2019) · IEEE Access · doi:10.1109/access.2019.2946301 · Cited by 14
“Typical BCI systems are generally composed of units for signal acquisition, feature extraction, classification and command output [46], [47].”
BackgroundSomon, B., Giebeler, Y., Darmet, L., & Dehais, F. (2022) · Frontiers in Neuroergonomics · doi:10.3389/fnrgo.2021.802486 · Cited by 15
“As advocated by Lotte (2015), shrinkage (or regularization) is used for the estimation of covariance matrices.”
MethodAlbuquerque, I., Monteiro, J., Rosanne, O., & Falk, T. H. (2022) · Frontiers in Artificial Intelligence · doi:10.3389/frai.2022.992732 · Cited by 15
“This is achieved by collecting a (usually small) number of labeled examples from this particular user and retraining or pruning the model to finetune it to the new user (Lotte, 2015).”
BackgroundLi, Z., Chen, B., Zhu, N., Li, W., Liu, T., Guo, L., Han, J., Zhang, T., & Yan, Z. (2025) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2025.3527489 · Cited by 12
“Following [26], we divide the trials of the same category into T s segments, and then concatenate segments from different and randomly selected training trials of the same category while maintaining the original time order.”
Kim, D.-K., Kim, Y.-T., Jung, H. R., Kim, H., & Kim, D. (2021) · International Winter Conference on Brain-Computer Interface (BCI) · doi:10.1109/bci51272.2021.9385340 · Cited by 11
“[15, 16].”
EEG-Based Emotion Recognition for Modulating Social-Aware Robot Navigation
Chang, Y., & Sun, L. (2021) · Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC) · doi:10.1109/embc46164.2021.9630721 · Cited by 13
“According to this updated review, the shrinkage linear discriminant analysis (LDA) classifier [33] has been shown to be effective with little training data.”
BackgroundTanaka, H., Watanabe, H., Maki, H., Sakti, S., & Nakamura, S. (2019) · Frontiers in Computational Neuroscience · doi:10.3389/fncom.2019.00015 · Cited by 13
“We assume that this system can also be used for assessing people who exhibit the anomalies of semantic context sensitivity (e.g., autism spectrum, dementia, Olichney et al., 2008; Pijnacker et al., 2010; O’Connor, 2012; Tanaka et al., 2012, 2015, 2017a,b, 2018a; Ujiro et al., 2018).”
BackgroundAlbuquerque, I., Monteiro, J., Rosanne, O., Tiwari, A., Gagnon, J.-F., & Falk, T. H. (2019) · Conference proceedings/Conference proceedings - IEEE International Conference on Systems, Man, and Cybernetics · doi:10.1109/smc.2019.8914469 · Cited by 13
“This is achieved by collecting a (usually small) number of labelled examples from this particular user and retraining or pruning the model to fine-tune it to the new user [9].”
BackgroundElectromyography for Teleoperated Tasks in Weightlessness
Hagengruber, A., Leipscher, U., Eskofier, B. M., & Vogel, J. (2021) · IEEE Transactions on Human-Machine Systems · doi:10.1109/thms.2020.3047975 · Cited by 11
“Other interfaces like electroencephalography (EEG) still require long and intense calibration routines, which limits their use in practice [37], especially with regard to the time constraints of a parabolic flight.”
BackgroundJacaruso, L. C. (2021) · Informatics in Medicine Unlocked · doi:10.1016/j.imu.2021.100729 · Cited by 11
“Lotte [42] proposes a way to generate new data by using Fourier transforms and principal component analysis in the context of brain-machine interfaces.”
MethodTransfer learning with large-scale data in brain-computer interfaces
Wei, C.-S., Lin, Y., Wang, Y.-T., Lin, C., & Jung, T. (2016) · Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC) · doi:10.1109/embc.2016.7591768 · Cited by 11
“drawback of other current TL approaches [2]-[4] and the”
BackgroundAs above, so below? Towards understanding inverse models in BCI
Lindgren, J. T. (2017) · Journal of Neural Engineering · doi:10.1088/1741-2552/aa86d0 · Cited by 10
“A cost of using the statistical, data-driven methods is that a significant amount of training data might be needed to learn a representation and a classifier that are able to generalize well to new, unseen data [59].”
Batch Mode Query by Committee for Motor Imagery-Based BCI
Hossain, I., Khosravi, A., Hettiarachchi, I. T., & Nahavandi, S. (2018) · IEEE Transactions on Neural Systems and Rehabilitation Engineering · doi:10.1109/tnsre.2018.2883594 · Cited by 12
“Therefore, attempt to reduce the training data to attain robust performance has an utmost significance in BCI system development process [10].”
BackgroundOn the use of generative deep neural networks to synthesize artificial multichannel EEG signals
Özdenizci, O., & Deniz Erdogmus, B. (2021) · International IEEE/EMBS Conference on Neural Engineering · doi:10.1109/ner49283.2021.9441381 · Cited by 11
A CNN model with feature integration for MI EEG subject classification in BMIPreprint
Roy, A. M. (2022) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.1101/2022.01.05.475058 · Cited by 12
“In this procedure, each training EEG trial for the particular class has been subdivided into multiple segments and then new trials are generated by combining segments collected from different and randomly selected trials from the same training class data [55].”
Wronkiewicz, M., Larson, E., & Lee, A. K. C. (2016) · Journal of Neural Engineering · doi:10.1088/1741-2560/13/5/056017 · Cited by 10
“Applied to BCI research, however, this ‘ transfer learning ’ technique is focused on eliminating lengthy calibration session needed in most BCIs ( for review, see Lotte 2015 ) .”
MethodWu, D., Lawhern, V. J., Gordon, S. M., Lance, B. J., & Lin, C. (2016) · Conference proceedings/Conference proceedings - IEEE International Conference on Systems, Man, and Cybernetics · doi:10.1109/smc.2016.7844330 · Cited by 12
“To facilitate their real-world applications, it is very important to minimize or even completely eliminate this calibration [10], [25], [28].”
BackgroundShang, G., Li, L., Zhang, L., Liu, X., Li, D., Qin, G., & Li, H. (2023) · Applied Sciences · doi:10.3390/app132011158 · Cited by 10
Zou, Y., Zhao, X., Chu, Y., Zhao, Y., Xu, W., & Han, J. (2018) · Medical & Biological Engineering & Computing · doi:10.1007/s11517-018-1917-x · Cited by 11
Data augmentation for cross-subject EEG features using Siamese neural network
Fu, R., Wang, Y., & Jia, C. (2022) · Biomedical Signal Processing and Control · doi:10.1016/j.bspc.2022.103614 · Cited by 13
Analysis of Minimal Channel Electroencephalography for Wearable Brain–Computer Interface
Suwannarat, A., Pan-ngum, S., & Israsena, P. (2024) · Electronics · doi:10.3390/electronics13030565 · Cited by 12
Liang, Y., Zhang, C., An, S., Wang, Z., Shi, K., Peng, T., Ma, Y., Xie, X., He, J., & Zheng, K. (2024) · Journal of Neural Engineering · doi:10.1088/1741-2552/ad4743 · Cited by 12
“Following [49], we use the short-time Fourier transform (STFT) to convert each channel of the training sample into a time–frequency representation (we used 250 ms-long hamming windows with 50% overlap) while maintaining the original time window order.”
Jiménez-Guarneros, M., & Fuentes-Pineda, G. (2025) · Knowledge-Based Systems · doi:10.1016/j.knosys.2025.113238 · Cited by 13
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
Krol, L. R., Klaproth, O. W., Vernaleken, C., Rußwinkel, N., & Zander, T. O. (2021) · Elsevier eBooks · doi:10.1016/b978-0-12-821413-8.00009-9 · Cited by 13
Zou, Y., Zhao, X., Chu, Y., Xu, W., Han, J., & Li, W. (2022) · Biomedical Signal Processing and Control · doi:10.1016/j.bspc.2022.103576 · Cited by 10
Reducing 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 13
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].”
Deng, X., Huo, H., Ai, L., Xu, D., & Li, C. (2025) · Sensors · doi:10.3390/s25092922 · Cited by 12
“One approach utilized to generate new data samples is the Segment and Reassemble (S&R) technique [45].”
Adaptive Fuzzy-Convolution and TSK-Guided Attention for Interpretable EEG MI Decoding
Sun, Y., Yao, J., Chen, T., Xia, K., Jiang, Y., & Qian, P. (2025) · IEEE Transactions on Fuzzy Systems · doi:10.1109/tfuzz.2025.3643920 · Cited by 10
“To augment the limited training samples and mitigate overfitting, we adopted the Segmentation and Reconstruction (S&R) data augmentation strategy [29].”
Ground Truth Dataset for EEG-Based Emotion Recognition With Visual Indication
Yang, G., Jiao, R. L., Jiang, H., & Zhang, T. (2020) · IEEE Access · doi:10.1109/access.2020.3030680 · Cited by 9
“Some researchers have generated EEG data by applying a geometric transformation to the original data [34], [35].”
MethodGuidelines 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
“This has enabled to reduce or even suppress the need for calibration for the target domain [4], [3].”
BackgroundSelf-citationTrambaiolli, L. R., & Falk, T. H. (2018) · Elsevier eBooks · doi:10.1016/b978-0-12-812892-3.00010-8 · Cited by 12
“…doors for adaptive systems capable of dynamically re-estimating the statistical properties of the BCI algorithm, for example, changing the classification weights of channels and measures, and achieving optimal solutions to daily signal changes (Amiri et al., 2013; Müller-Putz et al., 2011, 2015).”
BackgroundEEG-based neuroergonomics for 3d user interfaces: opportunities and challenges
Frey, J., Hachet, M., & Lotte, F. (2017) · Le travail humain · doi:10.3917/th.801.0073 · Cited by 13
“For instance, in Frey, Appriou, Lotte, & Hachet (2015); Frey, Appriou, Lotte, & Hachet (2016) we explored how EEG can be used to monitor visual comfort during the display of stereoscopic images.”
InfluentialSelf-citationXiong, Y., Li, J., Zhuang, Y., Zhao, X., Xu, Y., & Jie, L. (2025) · Biomedical Signal Processing and Control · doi:10.1016/j.bspc.2025.108421 · Cited by 9
D-FaST: Cognitive Signal Decoding With Disentangled Frequency–Spatial–Temporal Attention
Chen, W., Wang, C., Xu, K., Yuan, Y., Bai, Y., & Zhang, D. (2024) · IEEE Transactions on Cognitive and Developmental Systems · doi:10.1109/tcds.2024.3370261 · Cited by 8
“For each record, the temporal segment following the cue occurrence was extracted, in line with the guidelines presented in references [7], [8], [14], [62].”
MethodRao, Z., Zhang, R., He, S., Zhou, Y., Lu, Z., Li, K., & Li, Y. (2024) · IEEE Sensors Journal · doi:10.1109/jsen.2024.3510059 · Cited by 12
“Consequently, the features learned by the nonupdated model may not match those of the target patient, leading to imbalanced classification [22].”
InfluentialXue, Y., Tang, J., He, F., Xu, M., & Qi, H. (2019) · IEEE Access · doi:10.1109/access.2019.2940593 · Cited by 7
Are users' traits informative enough to predict/explain their mental-imagery based BCI performances?
Benaroch, C., Lotte, F., & Jeunet, C. (2019) · HAL · doi:10.3217/978-3-85125-682-6-39 · Cited by 10
“EEG signals were spatially filtered using 3 sets of Common Spatial Pattern (CSP) filters [21] and classifier using a shrinkage Linear Discriminant Analysis (sLDA) classifier [22].”
MethodSelf-citationEEG Generation of Virtual Channels Using an Improved Wasserstein Generative Adversarial Networks
Li, L.-L., Cao, G., Liang, H., Chen, J., & Zhang, Y.-P. (2022) · Lecture notes in computer science · doi:10.1007/978-3-031-13841-6_36 · Cited by 9
Decoding neural activity to assess individual latent state in ecologically valid contexts
Gordon, S. M., McDaniel, J. R., King, K., Lawhern, V. J., & Touryan, J. (2023) · Journal of Neural Engineering · doi:10.1088/1741-2552/acee20 · Cited by 8
“This was done as a form of data augmentation, which has been shown to improve the generalizability and robustness of machine learning models across a variety of domains (Shorten & Khoshgoftaar, 2019), including BCI decoding (Lotte, 2015; Zhang, et al., 2019).”
MethodTrambaiolli, L. R., Tiwari, A., & Falk, T. H. (2021) · Frontiers in Neuroergonomics · doi:10.3389/fnrgo.2021.678981 · Cited by 9
“, 2021b), artificially generated training samples (Lotte, 2015), or self-recalibrating classifiers (Bishop et al.”
BackgroundAugmented Reality Integrated Brain Computer Interface for Smart Home Control
Sahal, M., Dryden, E., Halac, M., Feldman, S. J., Heiman-Patterson, T. D., & Ayaz, H. (2021) · Lecture notes in networks and systems · doi:10.1007/978-3-030-80285-1_11 · Cited by 7
Zhou, Z., Dou, G., & Qu, X. (2022) · Communications in computer and information science · doi:10.1007/978-3-031-19679-9_16 · Cited by 11
Transformer-Based Multiscale 3-D Convolutional Network for Motor Imagery Classification
Su, J., An, S., Wang, G., Sun, X., Hao, Y., Li, H., & Gao, Z. (2025) · IEEE Sensors Journal · doi:10.1109/jsen.2025.3528009 · Cited by 8
“In [26], one of the data augmentation methods proposed in [29], namely the segmentation and reconstruction (S&R), was employed to evaluate its effectiveness in enhancing the classification performance of MI EEG signal.”
Murtaza, H., Ahmed, M., Murtaza, G., Zafar, S. N., & Nazir, T. (2025) · Biomedical Signal Processing and Control · doi:10.1016/j.bspc.2025.107552 · Cited by 9
Yuan, X. M., Shi, K., Jiang, N., & He, J. (2025) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2025.3593511 · Cited by 6
“To mitigate this, we adopted a segmentation and reconstruction (S&R) strategy in the time domain, as suggested by previous studies [22], [23], to augment the training data.”
Alexopoulou, A., Pergantis, P., Koutsojannis, C. M., Triantafillou, V., & Drigas, A. S. (2025) · Sensors · doi:10.3390/s25051342 · Cited by 29
“Technically speaking, BCI-VR systems are primarily classified based on the BCI paradigms they employ: motor imagery (MI) potential; event-related potential (ERP), such as P300; and steady-state visual evoked potentials (SSVEPs) [5,7,9,11].”
SRI-EEG: State-Based Recurrent Imputation for EEG Artifact Correction
Liu, Y., Höllerer, T., & Sra, M. (2022) · Frontiers in Computational Neuroscience · doi:10.3389/fncom.2022.803384 · Cited by 7
“Depending on the features of interest, EEG based BCIs can be categorized into two classes (Lotte, 2015): (1) event-related potential (ERP) based BCI, and (2) oscillatory activity based BCI.”
MethodMulti-Armed Bandits in Brain-Computer Interfaces
Heskebeck, F., Bergeling, C., & Bernhardsson, B. (2022) · Frontiers in Human Neuroscience · doi:10.3389/fnhum.2022.931085 · Cited by 6
“Optimal Calibration Data Another solution to the problem with calibration time (Lotte, 2015) is to collect calibration data cleverly.”
BackgroundMethodInfluentialEissa, T., Huber, M., Obermayer-Pietsch, B. M., Linkohr, B., Peters, A., Fleischmann, F., & Žigman, M. (2024) · PNAS Nexus · doi:10.1093/pnasnexus/pgae449 · Cited by 8
Cross-subject decoding of eye movement goals from local field potentials
Angjelichinoski, M., Choi, J. S., Banerjee, T., Pesaran, B., & Tarokh, V. (2020) · Journal of Neural Engineering · doi:10.1088/1741-2552/ab6df3 · Cited by 7
Metadata
- Authors
- Lotte, F.
- Affiliations
- Centre Inria de l'université de Bordeaux (FR)
- Corresponding author
- Lotte, Fabien
- ORCID
- Lotte, Fabien
- Venue
- Proceedings of the IEEE
- Published
- 2015-06
- Volume
- 103
- Issue
- 6
- Pages
- 871-890
- ISSN
- 0018-9219 (ISSN-L)
- 1558-2256
- OpenAlex
- W1610928686
- Cited by
- 327
- References
- 88
- Citations by year
- 2026: 29 · 2025: 45 · 2024: 36 · 2023: 28 · 2022: 38 · 2021: 30 · 2020: 36 · 2019: 29 · 2018: 25 · 2017: 12 · 2016: 16 · 2015: 3
- FWCI
- 11.10
- Type
- journal article
- Language
- en
- Open access
- green
- inria.hal.science
- Locations
- Proceedings of the IEEE (publishedVersion)
- HAL (submittedVersion, other-oa)
- Topics
- Life Sciences › Neuroscience › Cognitive Neuroscience › EEG and Brain-Computer Interfaces
- Physical Sciences › Computer Science › Human-Computer Interaction › Gaze Tracking and Assistive Technology
- Physical Sciences › Computer Science › Artificial Intelligence › Emotion and Mood Recognition
- Keywords
- brain-computer interface, calibration time reduction, EEG, neural oscillations, semi-supervised learning, regularization, data augmentation, training data size
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.