MIRepNet: A Pipeline and Foundation Model for EEG-Based Motor Imagery Classification

Dingkun Liu; Dongrui Wu; Jingwei Luo; Shijie Lian; Zhu Chen

REVIEW 5 major objections 7 minor 2 cited by

MIRepNet: A Pipeline and Foundation Model for EEG-Based Motor Imagery Classification

T0 review · 5 major / 7 minor · reviewed 2026-08-06 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read A motor-imagery-specific EEG foundation model, fed through a fixed motor-cortex channel template and pretrained with masked reconstruction plus classification, can decode a new user from fewer than 30 trials per class and beat both…

desk verdict Potentially useful few-shot MI decoding results, but the main comparison is confounded by unequal preprocessing and the 'first' claim needs cleaning up before this is citable. read the letter →

arxiv 2507.20254 v1 pith:GMDDM476 submitted 2025-07-27 cs.CV

Dingkun Liu , Zhu Chen , Jingwei Luo , Shijie Lian , Dongrui Wu This is my paper · ORCID

classification cs.CV

keywords motorimageryEEGfoundationmodelbrain-computerinterfacechanneltemplatemaskedtokenreconstructionfew-shotcalibrationEuclideanalignmentcross-datasettransfer

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper argues that EEG foundation models should be built for one brain-computer-interface paradigm at a time, rather than trained on many paradigms at once. It introduces MIRepNet, a motor-imagery (MI) foundation model whose preprocessing maps any headset's electrodes onto a fixed template over motor-cortex regions, and whose pretraining combines masked token reconstruction with supervised MI classification. The central claim is that after pretraining, a new subject's imagined movements can be decoded accurately from fewer than 30 trials per class and about 10 fine-tuning epochs, across five public datasets and headsets with 3 to 22 channels. If true, it would mean that short calibration sessions suffice to deploy MI-driven devices such as stroke-rehabilitation exoskeletons, and that paradigm-specific foundation models outperform both paradigm-agnostic models and specialist models trained from scratch.

What carries the argument

The central object is the neurophysiologically informed channel template: a fixed list of electrode sites over the frontal-central, central, centro-parietal, and temporal scalp regions, reconstructed from any headset by inverse-distance weighting. It carries cross-headset transfer by encoding the MI-specific prior that discriminative sensorimotor rhythms live over C3, C4, and nearby sites. The second mechanism is the hybrid pretraining objective, L = L_rec + L_cls, which jointly minimizes masked temporal-spatial token reconstruction error and cross-entropy MI classification loss. The template standardizes arbitrary electrode layouts, and the hybrid objective supplies the few-shot adaptation that the paper demonstrates.

What would settle it

Take a high-density recording, hold out the exact template channels, interpolate them from the remaining channels with inverse-distance weighting, and compare the interpolated to the recorded signals; if the correlation is low or the reconstructed channels are spatially smeared, the cross-headset transfer mechanism is not doing what the paper claims. A direct experiment on BNCI2014004 comparing MIRepNet with and without template interpolation would also separate preprocessing gains from learned-representation gains.

Watch

Extended reading notes

Core claim

The central discovery is that aligning heterogeneous MI recordings to a neurophysiologically informed channel template, then pretraining with a hybrid objective, yields representations that transfer to new subjects with very little data. MIRepNet's pipeline first bandpass-filters 8-30 Hz, resamples to 250 Hz, maps each headset's channels onto a fixed FC/C/CP/T template by inverse-distance weighting, and whitens each subject's trials by Euclidean alignment. A Transformer encoder then processes temporal-spatial tokens; 50% are masked and reconstructed while the same encoder drives a supervised MI classifier. On BNCI2014001, BNCI2015001, BNCI2014004, AlexMI, and BNCI2014001-4, covering 47 subjects, MIRepNet fine-tuned on 30% of one session reports average accuracies of 81.77, 81.67, 82.36, 69.79, and 64.14 percent, above both specialist models trained on the same 30% and generalist models fine-tuned on 80% of a session.

Load-bearing premise

The pipeline assumes that a fixed set of scalp positions over the motor cortex can be reconstructed by distance-weighted interpolation from whatever electrodes a headset actually has, even a three-channel headset, and the text never lists those positions or checks the reconstruction error.

Editorial extensions

If this is right

If MIRepNet is right, a stroke-rehabilitation BCI can be calibrated from a short session of under 30 trials per class instead of requiring large amounts of the new user's data.
The same pretrained weights transfer across headsets with different electrode layouts, including a 3-channel cap, as long as the channel-template interpolation covers the motor region.
The hybrid objective is the reason for the gain: ablations show that removing pretraining or removing masked reconstruction lowers accuracy on all five downstream tasks.
Mask ratio is not a sensitive hyperparameter; 10% to 90% masking all land within roughly two points of the best average, so the model does not need per-dataset tuning.
Paradigm-specific pretraining is preferable whenever the deployment paradigm is known in advance, which the paper argues is the typical real-world setting.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

A direct test the paper does not run: replace the learned Transformer with a shallow classifier on the same template-plus-Euclidean-alignment features; if accuracy holds, the gains come from preprocessing, not representation learning.
The subject-screening step keeps only subjects whose within-subject classifier passes a threshold, which may bias pretraining toward easy-to-decode users and could overstate few-shot gains for the harder users assistive BCIs most need.
The same logic suggests building separate foundation models for SSVEP, ERP, and sleep staging, each with its own cortical priors, instead of one mixed-paradigm model.
Because the template sites are never enumerated, auditing whether 3-channel recordings are actually reconstructed from nearby electrodes is necessary before clinical deployment; if they are not, low-density headsets may need a different alignment.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Short version: the paper does something genuinely worth looking at—pretraining an MI-specific EEG model on seven datasets and fine-tuning on under 30 trials per class of a new subject, with reported gains of 3–8 points over strong baselines across five held-out datasets. The ablation shows pretraining and the self-supervised loss each help, and the mask-ratio analysis suggests the method is not hyper-sensitive. If the results hold, this is directly useful for low-calibration BCI applications. So credit where earned: the evaluation is broad (47 downstream subjects), internally consistent, and the hybrid pretraining combination is not something I have seen reported before.

The soft spot is the comparison. MIRepNet gets the full pipeline: 8–30 Hz filtering, resampling, inverse-distance interpolation to a channel template, and Euclidean alignment. The specialist and generalist baselines are described as trained or fine-tuned on the target session, but nothing says they receive the same preprocessing. For BNCI2014004, which has three bipolar channels, interpolating to a dense FC/C/CP/T template could plausibly account for a large chunk of the margin all by itself. There is no ablation that feeds identical interpolated inputs to the baselines or removes the template from MIRepNet. That is a load-bearing omission, and the reader's stress-test lands.

Other issues are real but less severe. The channel template electrode set is never enumerated, so the interpolation cannot be checked. The subject-screening threshold is unspecified, which matters because selecting only good subjects inflates apparent pretraining quality. The text also has an internal inconsistency on whether the generalist 80%-finetuned baseline used only the target subject's session or other subjects' data—that needs to be clarified. The 'first MI foundation model' claim is weakened by the authors' own CLEAN-MI paper and by the table rows labeled 'MI-FM' and Section V.D calling the model 'RepMI'; either the naming history should be explained or the claim softened. No significance tests back 'significantly outperforming', and the three-seed variances (0.10–0.42) look smaller than one would expect, though that alone is not damning.

Bottom line: this deserves a serious referee, but not acceptance as-is. I would send it out, ask for code and pretrained weights, an enumerated template, a controlled preprocessing comparison or an ablation isolating template/EA, and a cleaned-up narrative about what is actually new relative to CLEAN-MI/RepMI. If those come back, this could be a useful paper. If the authors cannot or will not release the code and clarify the baseline protocol, the central claim stays unverified.

Referee Report

5 major / 7 minor

Summary. The paper introduces MIRepNet, described as the first EEG foundation model tailored specifically to the motor imagery (MI) paradigm. The method consists of a preprocessing pipeline (8-30 Hz band-pass filtering, resampling to 250 Hz, subject screening, inverse-distance interpolation to a fixed FC/C/CP/T channel template, and Euclidean alignment) and a hybrid pretraining scheme combining masked token reconstruction with supervised MI classification. The authors pretrain on seven MI datasets and evaluate on five downstream datasets (47 subjects), comparing with nine specialist models and five generalist EEG foundation models. They report consistent improvements, e.g., 81.77% vs. 73.67% on BNCI2014001, with only 30% of a single subject's trials used for fine-tuning.

Significance. If the reported results withstand scrutiny, a paradigm-specific MI foundation model requiring fewer than 30 trials per class for a new subject would be a practically valuable contribution to BCI calibration. The paper has notable strengths: the evaluation covers five downstream datasets from MOABB, uses standard public baselines, reports standard deviations across three runs, and includes an ablation of pretraining and self-supervision. The authors also state that code will be released. However, the central claim is currently conditional because MIRepNet is inseparable from its preprocessing pipeline, and the baselines are not demonstrably given the same inputs. The significance therefore depends on the outcome of a controlled comparison that isolates the learned representation from the preprocessing.

major comments (5)

[§IV.C, Tables II–VII] The SOTA claim is confounded by an asymmetric preprocessing comparison. MIRepNet receives band-pass filtering, inverse-distance interpolation to the channel template, and Euclidean alignment before being fed to the model (Eqs. 1–6), while the specialist baselines in Tables II–VII are described only as '30% Trained' and the generalist baselines as '30%/80% Finetuned.' Section IV.B states only that all signals were resampled to 250 Hz; it does not state that baselines received the same template interpolation or EA. On BNCI2014004, which has only C3, Cz, and C4, MIRepNet reconstructs the full template while specialist baselines see the raw three-channel montage. The reported 3–8 point margins can therefore be explained by the preprocessing itself. The authors should rerun all baselines on identical template+EA inputs and add MIRepNet variants without template interpolation and without EA.
[§III.B.3, Eqs. (2)–(4)] The channel template is never enumerated. The text defines 'a fixed set of C electrodes' over FC/C/CP/T regions but does not give the electrode names, their coordinates φ(t_i), or the value of C. This makes it impossible to check the inverse-distance interpolation, especially for low-density recordings such as the three-channel BNCI2014004, where most template sites lie far from any recorded electrode. The authors should specify the full template and, if possible, report reconstruction error or a sensitivity analysis for sparse montages.
[Table VIII] The ablation study does not isolate the preprocessing components. It compares only 'w/o Pre-training' and 'w/o Self-supervised' variants against the full MIRepNet. Since the headline claim is that the learned foundation-model representation outperforms specialists, the absence of ablations that remove the channel template or EA means the contribution of those preprocessing steps to the reported gains is unknown. Add ablations with raw montage inputs and without EA to substantiate the attribution.
[Table I and §V.B.8] The dataset description is internally inconsistent. Section V.B.8 states that BNCI2014001 includes two sessions of 288 trials per subject, which would be 5,184 trials for 9 subjects, but Table I reports 1,296 total trials. The relationship between BNCI2014001 and BNCI2014001-4, and the exact 30% calibration split, are not defined. In addition, the abstract's 'fewer than 30 trials per class' claim appears not to hold for BNCI2015001, where 30% of 200 trials per subject with two classes is exactly 30 trials per class. Clarify the trial counts and the calibration protocol.
[§I and Ref. [13]] The manuscript claims 'the first EEG foundation model tailored for the MI paradigm' but does not discuss the authors' own CLEAN-MI pipeline (ref. [13]), which appears to overlap with the preprocessing ideas in Section III.B (subject selection, channel template). The paper should state explicitly which components are inherited from CLEAN-MI, which are new to MIRepNet, and how the novelty claim is consistent with that prior work.

minor comments (7)

[Eq. (4)] Equation (4) has an index inconsistency: the left-hand side uses X'[b,i,t], the sum runs over c, but the weight is written as W_ij instead of W_i,c (or W_i,k). Please correct the notation.
[Section III.A and III.B.3] The symbol C is used both for the number of classes in the problem definition (Section III.A) and for the number of template electrodes (Section III.B.3). Use distinct symbols to avoid ambiguity.
[Section V.D] The model is referred to as 'RepMI' in several places (e.g., 'A key benefit of RepMI' and 'this fast convergence demonstrates RepMI's practicality') although the paper's model is named MIRepNet. Unify the name throughout.
[Table I and Section V.B.1] BNCI2014002 is listed with 14 subjects in Table I but Section V.B.1 says 13 participants. Please verify the correct count.
[Fig. 5 and Tables IX–XIII] Figure 5's caption says the mask-ratio sensitivity is shown on 'three downstream MI datasets,' but Tables IX–XIII report results on five downstream datasets. Update the caption.
[Section IV.B] The 'significantly outperforming' claim in the abstract is not supported by any statistical test. Given that standard deviations are reported, paired significance tests or confidence intervals across subjects should be added for the main comparisons.
[Section V.B.3] 'datatset' is a typo for 'dataset'.

Circularity Check

0 steps flagged · score 0.0 of 10

No circular derivation: MIRepNet's SOTA claim rests on held-out downstream trials, not on its own preprocessing or a self-citation chain; the main caveat is comparison fairness, not circularity.

full rationale

MIRepNet's principal claim is an empirical benchmark result. Pretraining uses seven MI datasets, while downstream evaluation uses five disjoint datasets (47 subjects), and reported accuracies are computed on trials held out after fine-tuning on 30% of a subject's session. Nothing in the paper defines the target accuracy in terms of the channel-template interpolation, Euclidean alignment, or pretraining losses; these are inputs that could plausibly help or hurt, and the ablations in Table VIII show that pretraining and the self-supervised term contribute measurable margins. The main legitimate concern is comparison fairness: specialist and generalist baselines may not receive the same template interpolation and EA as MIRepNet, so the 3-8 point margins could reflect preprocessing rather than the learned representation. That is a validity or confound issue, not a circular reduction, because no equation equates the reported metric to a fitted input and no prediction is constructed from the quantity it claims to predict. The only self-citation of note is reference [13] (CLEAN-MI by the same group), used in the introduction to support paradigm-specific neurophysiological differences; this is a standard claim independently elaborated in Section V and is not load-bearing for the empirical result. The 'first' priority claim and the 'MI-FM' table label are presentation and historical issues, not circular derivations. Accordingly, no quoted reduction exists under the requested standard, and the paper should receive a non-circular finding.

Assumptions & free parameters 6 free parameters · 5 assumptions · 1 invented entities

The central claim rests on three groups of choices. First, hand-chosen data-construction parameters: the 8-30 Hz band, 250 Hz resampling, the unlisted channel template electrodes, and the unstated subject screening threshold. Second, imported domain assumptions: ERD/ERS neuroscience, the transfer value of masked reconstruction, and the properties of Euclidean alignment (whose key citations are self-citations by the senior author). Third, an invented methodological construct, the channel template, whose electrode set is not specified, so its contribution cannot be independently evaluated. The empirical accuracy numbers are not circular by construction, but the paper's incremental contribution over the authors' own CLEAN-MI pipeline is not delineated.

free parameters (6)

subject_screening_accuracy_threshold = Not reported in text
Section III.B.2 excludes pretraining subjects whose within-subject classifier accuracy falls below a 'predefined performance threshold'. The threshold and the classifier are unspecified, and this choice directly shapes the pretraining data composition.
channel_template_electrode_set = Not enumerated (FC, C, CP, T regions)
Section III.B.3 defines a fixed set of C template electrodes over FC, C, CP, and T regions, but the actual electrodes are never listed. This set is the interpolation target for every dataset, so the model's input space is undefined in the text.
temporal_kernel_size_and_stride (kt, st) = Not reported in text
Eq. (7) defines the temporal convolution embedding that produces tokens, but kernel size and stride are unspecified, so the token sequence length W, and therefore H', cannot be derived from the paper.
spatial_convolution_encoder_parameters = Not reported in text
Eq. (8) compresses spatial information with convolution parameters that are never given, including output channels and the resulting W'.
bandpass_range_and_target_sample_rate = 8-30 Hz, 250 Hz
Section III.B.1 chooses the 8-30 Hz band and 250 Hz resampling based on MI neurophysiology. This is standard for MI decoding but still a hand-chosen modeling decision.
masking_ratio_alpha = 0.5
Section IV.B sets the mask ratio to 50%. Section IV.E shows accuracy is robust across 10-90% masking, so this parameter is low concern.

assumptions (5)

domain assumption ERD/ERS over the sensorimotor cortex is the dominant MI discriminative signal, justifying the FC/C/CP/T template and the 8-30 Hz band.
Sections III.B.1, III.B.3, and V.A rely on standard, well-cited neuroscience (Neuper, Pfurtscheller, and related literature). The paper does not validate this assumption on its own data.
domain assumption Masked token reconstruction loss Lrec produces representations that transfer to downstream MI decoding.
Section III.D.1 assumes the MAE/LaBraM/EEGPT transfer story carries over to MI. No analysis in the paper demonstrates why reconstruction helps MI specifically, beyond the ablation in Table VIII.
domain assumption Euclidean alignment whitening (Eqs. 5-6) removes cross-subject second-order shifts without destroying class-discriminative structure.
Section III.B.4 relies on the known properties of EA from He and Wu [18] and Wu [42], both self-citations. The assumption is standard but is imported rather than demonstrated here.
domain assumption The 30% finetune and 80% baseline protocols measure the same quantity: per-subject held-out decoding accuracy.
Section IV.C first states that adapting a generalist model to subject i requires combining data from other subjects, then says generalist baselines were fine-tuned with 80% of the target session. The two statements are not reconciled, so protocol parity is ambiguous.
domain assumption MOABB dataset labels and trial structures are correct as provided.
Section IV.A takes all pretraining and downstream data from MOABB [31] and does not audit labels, event timing, or trial validity beyond its own subject screening.

invented entities (1)

MIRepNet channel template (fixed FC/C/CP/T electrode set)
purpose: Unify arbitrary electrode configurations into a common spatial layout for cross-dataset pretraining and finetuning.
Introduced in Section III.B.3 without enumerating the electrodes. The only evidence for its validity is the benchmark tables in this paper; no external falsifiable handle (e.g., a neurophysiological prediction about which electrodes matter) is provided.

how reviews work

0 comments

Cite this review

Pith. "Pith review of MIRepNet: A Pipeline and Foundation Model for EEG-Based Motor Imagery Classification." pith.science (2026). https://pith.science/paper/GMDDM476

@misc{pith2026250720254,
  author       = {Pith},
  title        = {Pith review of: MIRepNet: A Pipeline and Foundation Model for EEG-Based Motor Imagery Classification},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/GMDDM476}},
  note         = {Machine review of arXiv:2507.20254}
}

read the original abstract

Brain-computer interfaces (BCIs) enable direct communication between the brain and external devices. Recent EEG foundation models aim to learn generalized representations across diverse BCI paradigms. However, these approaches overlook fundamental paradigm-specific neurophysiological distinctions, limiting their generalization ability. Importantly, in practical BCI deployments, the specific paradigm such as motor imagery (MI) for stroke rehabilitation or assistive robotics, is generally determined prior to data acquisition. This paper proposes MIRepNet, the first EEG foundation model tailored for the MI paradigm. MIRepNet comprises a high-quality EEG preprocessing pipeline incorporating a neurophysiologically-informed channel template, adaptable to EEG headsets with arbitrary electrode configurations. Furthermore, we introduce a hybrid pretraining strategy that combines self-supervised masked token reconstruction and supervised MI classification, facilitating rapid adaptation and accurate decoding on novel downstream MI tasks with fewer than 30 trials per class. Extensive evaluations across five public MI datasets demonstrated that MIRepNet consistently achieved state-of-the-art performance, significantly outperforming both specialized and generalized EEG models. Our code will be available on GitHub\footnote{https://github.com/staraink/MIRepNet}.

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 2 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

EEG-FM-Compass: Progress, Benchmarking, and Future Directions for EEG Foundation Models
cs.LG 2026-01 conditional novelty 6.0 of 10

A unified benchmark of 12 EEG foundation models across 13 datasets finds specialists remain competitive and larger pre-trained models do not consistently improve downstream decoding.
The 1st International Workshop on Disentangled Representation Learning for Controllable Generation (DRL4Real): Methods and Results
cs.LG 2025-08 unverdicted

A workshop summary describing nine papers on disentangled representation learning applied to controllable generation, editing, video, 3D, EEG, and compression.

Reference graph

Works this paper leans on

42 extracted references · 35 canonical work pages · cited by 2 Pith papers

[13]

CLEAN-MI: A scalable and efficient pipeline for constructing high-quality neurodata in motor imagery paradigm,

D. Liu, Z. Chen, and D. Wu, “CLEAN-MI: A scalable and efficient pipeline for constructing high-quality neurodata in motor imagery paradigm,” arXiv preprint arXiv:2506.11830 , 2025

arXiv 2025
[1]

Brain computer interfaces, a review,

L. F. Nicolas Alonso and J. Gomez Gil, “Brain computer interfaces, a review,” Sensors, vol. 12, no. 2, pp. 1211–1279, 2012

work page 2012
[2]

Characteri- zation of EEG—a comparative study,

N. Kannathal, U. R. Acharya, C. M. Lim, and P. Sadasivan, “Characteri- zation of EEG—a comparative study,”Computer Methods and Programs in Biomedicine, vol. 80, no. 1, pp. 17–23, 2005

work page 2005
[3]

Usability and cost-effectiveness in brain-computer interaction: is it user throughput or technology related?

A. V ourvopoulos and S. B. I. Badia, “Usability and cost-effectiveness in brain-computer interaction: is it user throughput or technology related?” in Proc. of the 7th Augmented Human Int’l Conf. 2016 , 2016, pp. 1–8

work page 2016
[4]

Switching EEG headsets made easy: Reducing offline calibration effort using active wighted adaptation regularization,

D. Wu, V . J. Lawhern, W. D. Hairston, and B. J. Lance, “Switching EEG headsets made easy: Reducing offline calibration effort using active wighted adaptation regularization,” IEEE Trans. on Neural Systems and Rehabilitation Engineering, vol. 24, no. 11, pp. 1125–1137, 2016

work page 2016
[5]

Spatial distillation based distribution alignment (SDDA) for cross-headset EEG classification,

D. Liu, S. Li, Z. Wang, W. Li, and D. Wu, “Spatial distillation based distribution alignment (SDDA) for cross-headset EEG classification,” arXiv preprint arXiv:2503.05349 , 2025

arXiv 2025
[6]

EEGformer: A transformer–based brain activity classification method using EEG signal,

Z. Wan, M. Li, S. Liu, J. Huang, H. Tan, and W. Duan, “EEGformer: A transformer–based brain activity classification method using EEG signal,” Frontiers in Neuroscience, vol. 17, p. 1148855, 2023

work page 2023
[7]

BIOT: Biosignal transformer for cross-data learning in the wild,

C. Yang, M. Westover, and J. Sun, “BIOT: Biosignal transformer for cross-data learning in the wild,” Advances in Neural Information Processing Systems, vol. 36, pp. 78 240–78 260, Dec. 2023

work page 2023

Show all 42 references

[8]

BENDR: Using trans- formers and a contrastive self-supervised learning task to learn from massive amounts of EEG data,

D. Kostas, S. Aroca-Ouellette, and F. Rudzicz, “BENDR: Using trans- formers and a contrastive self-supervised learning task to learn from massive amounts of EEG data,” Frontiers in Human Neuroscience , vol. 15, p. 653659, 2021

2021
[9]

CBraMod: A criss-cross brain foundation model for EEG decoding,

J. Wang, S. Zhao, Z. Luo, Y . Zhou, H. Jiang, S. Li, T. Li, and G. Pan, “CBraMod: A criss-cross brain foundation model for EEG decoding,” arXiv preprint arXiv:2412.07236 , 2024

2024
[10]

Large brain model for learning generic representations with tremendous EEG data in BCI,

W.-B. Jiang, L.-M. Zhao, and B.-L. Lu, “Large brain model for learning generic representations with tremendous EEG data in BCI,” in Proc. Int’l Conf. on Learning Representations , Vienna, Austria, May. 2024

2024
[11]

EEGPT: Pretrained transformer for universal and reliable representation of EEG signals,

G. Wang, W. Liu, Y . He, C. Xu, L. Ma, and H. Li, “EEGPT: Pretrained transformer for universal and reliable representation of EEG signals,” Vancouver, Canada, Dec. 2025, pp. 39 249–39 280

2025
[12]

Neurolm: A universal multi- task foundation model for bridging the gap between language and EEG signals,

W. Jiang, Y . Wang, B.-l. Lu, and D. Li, “Neurolm: A universal multi- task foundation model for bridging the gap between language and EEG signals,” in The Thirteenth Int’l Conf. on Learning Representations , Singapore, APR. 2025

2025
[14]

A survey of wearable lower extremity neurorehabilitation exoskeleton: Sensing, gait dynamics, and human–robot collaboration,

J. Li, X. Gu, S. Qiu, X. Zhou, A. Cangelosi, C. K. Loo, and X. Liu, “A survey of wearable lower extremity neurorehabilitation exoskeleton: Sensing, gait dynamics, and human–robot collaboration,” IEEE Trans. on Systems, Man, and Cybernetics: Systems , vol. 54, no. 6, pp. 3675– ...

2024
[15]

Paradigm shifts in the neuropsychology of epilepsy,

B. Hermann, D. W. Loring, and S. Wilson, “Paradigm shifts in the neuropsychology of epilepsy,” Journal of the Int’l Neuropsychological Society, vol. 23, no. 9-10, pp. 791–805, 2017

2017
[16]

Motor imagery and its practical application,

O. Mokienko, L. Chernikova, A. Frolov, and P. Bobrov, “Motor imagery and its practical application,” Neuroscience and Behavioral Physiology , vol. 44, no. 5, pp. 483–489, 2014

2014
[17]

Motor imagery and EEG-based control of spelling devices and neuroprosthe- ses,

C. Neuper, G. R. M ¨uller-Putz, R. Scherer, and G. Pfurtscheller, “Motor imagery and EEG-based control of spelling devices and neuroprosthe- ses,” Progress in Brain Research , vol. 159, pp. 393–409, 2006

2006
[18]

Transfer learning for brain-computer interfaces: A euclidean space data alignment approach,

H. He and D. Wu, “Transfer learning for brain-computer interfaces: A euclidean space data alignment approach,” IEEE Trans. on Biomedical Engineering, vol. 67, no. 2, pp. 399–410, 2019

2019
[19]

Optimizing spatial filters for robust EEG single-trial analysis,

B. Blankertz, R. Tomioka, S. Lemm, M. Kawanabe, and K.-R. Muller, “Optimizing spatial filters for robust EEG single-trial analysis,” IEEE Signal processing magazine , vol. 25, no. 1, pp. 41–56, 2007

2007
[20]

Deep learning with convolutional neural networks for EEG decoding and visualization,

R. T. Schirrmeister, J. T. Springenberg, L. D. J. Fiederer, M. Glasstetter, K. Eggensperger, M. Tangermann, F. Hutter, W. Burgard, and T. Ball, “Deep learning with convolutional neural networks for EEG decoding and visualization,” Human Brain Mapping , vol. 38, no. 11, pp. 539...

2017
[21]

EEGNet: a compact convolutional neural network for EEG–based brain–computer interfaces,

V . J. Lawhern, A. J. Solon, N. R. Waytowich, S. M. Gordon, C. P. Hung, and B. J. Lance, “EEGNet: a compact convolutional neural network for EEG–based brain–computer interfaces,” Journal of Neural Engineering , vol. 15, no. 5, p. 056013, 2018

2018
[22]

FBCNet: A multi–view convolutional neural network for brain-computer interface,

R. Mane, E. Chew, K. Chua, K. K. Ang, N. Robinson, A. P. Vinod, S.-W. Lee, and C. Guan, “FBCNet: A multi–view convolutional neural network for brain-computer interface,” arXiv preprint arXiv:2104.01233, 2021

2021 arXiv
[23]

IFNet: An interactive frequency con- volutional neural network for enhancing motor imagery decoding from EEG,

J. Wang, L. Yao, and Y . Wang, “IFNet: An interactive frequency con- volutional neural network for enhancing motor imagery decoding from EEG,” IEEE Trans. on Neural Systems and Rehabilitation Engineering , vol. 31, pp. 1900–1911, 2023

1900
[24]

ADFCNN: attention-based dual-scale fusion convolutional neural network for motor imagery brain–computer interface,

W. Tao, Z. Wang, C. M. Wong, Z. Jia, C. Li, X. Chen, C. P. Chen, and F. Wan, “ADFCNN: attention-based dual-scale fusion convolutional neural network for motor imagery brain–computer interface,” IEEE Trans. on Neural Systems and Rehabilitation Engineering , vol. 32, pp. 154–165, 2023

2023
[25]

EEG conformer: Convolutional transformer for EEG decoding and visualization,

Y . Song, Q. Zheng, B. Liu, and X. Gao, “EEG conformer: Convolutional transformer for EEG decoding and visualization,”IEEE Trans. on Neural Systems and Rehabilitation Engineering , vol. 31, pp. 710–719, 2022

2022
[26]

EDPNet: An efficient dual prototype network for motor imagery EEG decoding,

C. Han, C. Liu, C. Cai, J. Wang, and D. Qian, “EDPNet: An efficient dual prototype network for motor imagery EEG decoding,” arXiv e- prints, pp. arXiv–2407, 2024

2024
[27]

Random forests in non–invasive sensorimotor rhythm brain-computer interfaces: a practical and convenient non-linear classifier,

D. Steyrl, R. Scherer, J. Faller, and G. R. M ¨uller-Putz, “Random forests in non–invasive sensorimotor rhythm brain-computer interfaces: a practical and convenient non-linear classifier,” Biomedical Engineer- ing/Biomedizinische Technik, vol. 61, no. 1, pp. 77–86, 2016

2016
[28]

Physiobank, physiotoolkit, and physionet: components of a new research resource for complex physiologic signals,

A. L. Goldberger, L. A. Amaral, L. Glass, J. M. Hausdorff, P. C. Ivanov, R. G. Mark, J. E. Mietus, G. B. Moody, C.-K. Peng, and H. E. Stanley, “Physiobank, physiotoolkit, and physionet: components of a new research resource for complex physiologic signals,” Circulation, vol. 1...

2000
[29]

A large EEG database with users’ profile information for motor imagery brain- computer interface research,

P. Dreyer, A. Roc, L. Pillette, S. Rimbert, and F. Lotte, “A large EEG database with users’ profile information for motor imagery brain- computer interface research,” Scientific Data , vol. 10, no. 1, p. 580, 2023

2023
[30]

Evaluation of EEG oscillatory patterns and cognitive process during simple and compound limb motor imagery,

W. Yi, S. Qiu, K. Wang, H. Qi, L. Zhang, P. Zhou, F. He, and D. Ming, “Evaluation of EEG oscillatory patterns and cognitive process during simple and compound limb motor imagery,” PloS One , vol. 9, no. 12, p. e114853, 2014

2014
[31]

MOABB: trustworthy algorithm bench- marking for BCIs,

V . Jayaram and A. Barachant, “MOABB: trustworthy algorithm bench- marking for BCIs,” Journal of Neural Engineering , vol. 15, no. 6, p. 066011, 2018

2018
[32]

A fully automated trial selection method for optimization of motor imagery based brain– computer interface,

B. Zhou, X. Wu, Z. Lv, L. Zhang, and X. Guo, “A fully automated trial selection method for optimization of motor imagery based brain– computer interface,” PloS One, vol. 11, no. 9, p. e0162657, 2016

2016
[33]

EEG dataset and openbmi toolbox for three BCI paradigms: An investigation into BCI illiteracy,

M.-H. Lee, O.-Y . Kwon, Y .-J. Kim, H.-K. Kim, Y .-E. Lee, J. Williamson, S. Fazli, and S.-W. Lee, “EEG dataset and openbmi toolbox for three BCI paradigms: An investigation into BCI illiteracy,” GigaScience, vol. 8, no. 5, p. giz002, 2019

2019
[34]

EEG datasets for motor imagery brain–computer interface,

H. Cho, M. Ahn, S. Ahn, M. Kwon, and S. C. Jun, “EEG datasets for motor imagery brain–computer interface,” GigaScience, vol. 6, no. 7, p. gix034, 2017

2017
[35]

Review of the BCI competition IV,

M. Tangermann, K.-R. M ¨uller, A. Aertsen, N. Birbaumer, C. Braun, C. Brunner, R. Leeb, C. Mehring, K. J. Miller, G. R. M ¨uller-Putz et al., “Review of the BCI competition IV,” Frontiers in neuroscience, vol. 6, p. 55, 2012

2012
[36]

Autocalibration and recurrent adaptation: Towards a plug and play online ERD-BCI,

J. Faller, C. Vidaurre, T. Solis-Escalante, C. Neuper, and R. Scherer, “Autocalibration and recurrent adaptation: Towards a plug and play online ERD-BCI,” IEEE Trans. on Neural Systems and Rehabilitation Engineering, vol. 20, no. 3, pp. 313–319, 2012

2012
[37]

Brain–computer communication: motivation, aim, and impact of exploring a virtual apartment,

R. Leeb, F. Lee, C. Keinrath, R. Scherer, H. Bischof, and G. Pfurtscheller, “Brain–computer communication: motivation, aim, and impact of exploring a virtual apartment,” IEEE Trans. on Neural Systems and Rehabilitation Engineering , vol. 15, no. 4, pp. 473–482, 2007

2007
[38]

Commande robuste d’un effecteur par une interface cerveau–machine EEG asynchrone,

B. Alexandre, “Commande robuste d’un effecteur par une interface cerveau–machine EEG asynchrone,” Universit´e de Grenoble , 2006

2006
[39]

Pre-stimulus sensorimotor rhythms influence brain–computer interface classification performance,

C. L. Maeder, C. Sannelli, S. Haufe, and B. Blankertz, “Pre-stimulus sensorimotor rhythms influence brain–computer interface classification performance,” IEEE Trans. on Neural Systems and Rehabilitation Engi- neering, vol. 20, no. 5, pp. 653–662, 2012

2012
[40]

Transfer learning for motor imagery based brain-computer interfaces: A tutorial,

D. Wu, X. Jiang, and R. Peng, “Transfer learning for motor imagery based brain-computer interfaces: A tutorial,” Neural Networks, vol. 153, pp. 235–253, 2022

2022
[41]

Affective brain-computer interfaces (aBCIs): A tutorial,

D. Wu, B.-L. Lu, B. Hu, and Z. Zeng, “Affective brain-computer interfaces (aBCIs): A tutorial,” Proc. of the IEEE , vol. 11, no. 10, pp. 1314–1332, 2023

2023
[42]

Revisiting Euclidean alignment for transfer learning in EEG-based brain-computer interfaces,

D. Wu, “Revisiting Euclidean alignment for transfer learning in EEG-based brain-computer interfaces,” Journal of Neural Engineering , vol. 22, p. 031005, 2025

2025

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.

[1] [13]

CLEAN-MI: A scalable and efficient pipeline for constructing high-quality neurodata in motor imagery paradigm,

D. Liu, Z. Chen, and D. Wu, “CLEAN-MI: A scalable and efficient pipeline for constructing high-quality neurodata in motor imagery paradigm,” arXiv preprint arXiv:2506.11830 , 2025

arXiv 2025

[2] [1]

Brain computer interfaces, a review,

L. F. Nicolas Alonso and J. Gomez Gil, “Brain computer interfaces, a review,” Sensors, vol. 12, no. 2, pp. 1211–1279, 2012

work page 2012

[3] [2]

Characteri- zation of EEG—a comparative study,

N. Kannathal, U. R. Acharya, C. M. Lim, and P. Sadasivan, “Characteri- zation of EEG—a comparative study,”Computer Methods and Programs in Biomedicine, vol. 80, no. 1, pp. 17–23, 2005

work page 2005

[4] [3]

Usability and cost-effectiveness in brain-computer interaction: is it user throughput or technology related?

A. V ourvopoulos and S. B. I. Badia, “Usability and cost-effectiveness in brain-computer interaction: is it user throughput or technology related?” in Proc. of the 7th Augmented Human Int’l Conf. 2016 , 2016, pp. 1–8

work page 2016

[5] [4]

Switching EEG headsets made easy: Reducing offline calibration effort using active wighted adaptation regularization,

D. Wu, V . J. Lawhern, W. D. Hairston, and B. J. Lance, “Switching EEG headsets made easy: Reducing offline calibration effort using active wighted adaptation regularization,” IEEE Trans. on Neural Systems and Rehabilitation Engineering, vol. 24, no. 11, pp. 1125–1137, 2016

work page 2016

[6] [5]

Spatial distillation based distribution alignment (SDDA) for cross-headset EEG classification,

D. Liu, S. Li, Z. Wang, W. Li, and D. Wu, “Spatial distillation based distribution alignment (SDDA) for cross-headset EEG classification,” arXiv preprint arXiv:2503.05349 , 2025

arXiv 2025

[7] [6]

EEGformer: A transformer–based brain activity classification method using EEG signal,

Z. Wan, M. Li, S. Liu, J. Huang, H. Tan, and W. Duan, “EEGformer: A transformer–based brain activity classification method using EEG signal,” Frontiers in Neuroscience, vol. 17, p. 1148855, 2023

work page 2023

[8] [7]

BIOT: Biosignal transformer for cross-data learning in the wild,

C. Yang, M. Westover, and J. Sun, “BIOT: Biosignal transformer for cross-data learning in the wild,” Advances in Neural Information Processing Systems, vol. 36, pp. 78 240–78 260, Dec. 2023

work page 2023