Artificial Intelligence · arXiv:2611.17031 v1 · submitted 12 Nov 2026 · 51 references
Your Vision Foundation Model Is Secretly an EEG Foundation Model
The paper says
“Electroencephalography (EEG) foundation models are pretrained on thousands of hours of recordings, yet their frozen features transfer poorly to downstream tasks, and this transfer does not consistently improve with larger models or more pretraining data.”
more of the abstract
We show that a vision foundation model that has never seen EEG is a stronger frozen EEG encoder than these specialized models, once the signals are rendered as images that a vision transformer can read. Our method, SeeEEG, keeps the vision backbone frozen and learns only how EEG is displayed to it. Each channel occupies a dedicated row of patches and is drawn either as a waveform at a fixed gain or as a scalogram whose frequency range spans the patch height, so that task-relevant patterns fall within individual tokens. The montage, filtering, and contrast parameters of this display are initialized from a standard clinical montage and optimized with the task loss through the frozen backbone, which amounts to at most 1,113 trainable parameters per view. We compare SeeEEG with seven EEG foundation models on eight clinical, sleep, affective, and brain-computer interface datasets, with every encoder frozen and paired with identical readout heads. With a frozen DINOv3 ViT-B, SeeEEG outperforms every EEG foundation model in balanced accuracy on all eight datasets and reaches a mean rank of 1.3 in Cohen kappa or AUROC, compared with 2.9 for the strongest baseline, REVE. The learned displays transfer without further optimization to the video encoder V-JEPA 2.1, which outperforms every EEG foundation model on both metrics on six of the eight datasets. Replacing DINOv3 with a randomly initialized ViT-B lowers the score by 15.2 points on average, so the ability to read rendered EEG comes from visual pretraining.
from the abstract
Pith says
Machine review · deepseek-v4-flash · 13 Nov 2026 · what was read was not recorded · not checked by a person
It can be wrong. Challenge this review or reply to it.
Holds up, with conditions. Frozen ViT beats all EEG foundation models on 8 datasets—the evidence is strong and the method is clean. The cross-fitting concern is real but the paper acknowledges it and the main result is robust enough that I'd send this to review. This line is wrong
1 major objection · 6 minor
How often objections from reviews like this one are right: not yet measured
Not checked: figures, code, what the 51 references say.
The paper says, at §3.3, Eq. (7)
The review did not quote this passage.
Pith says: Objection 1 of 1 · major
The cross-fitting procedure prevents the display from exploiting subject-identity shortcuts on between-subject tasks, but for the three within-subject datasets (SEED, BCIC2020-3, PhysioMI) the folds split trials rather than subjects, so each head h−k still trains on other trials of the same subjects whose windows appear in fold Fk.
more of objection 1
The paper acknowledges the general shortcut (b98-b99) but provides no empirical measure of whether the display exploits session-, block-, or subject-level artifacts that correlate with labels within a subject (for example, recording-session drift, persistent eye-blink patterns, or impedance changes). This is load-bearing because two of the larger margins in Table 2 (BCIC2020-3 kappa +8.1, SEED kappa +6.3 over REVE) and the PhysioMI gain arise on within-subject tasks, so the 'all eight datasets' claim rests in part on this unresolved point. Because a trial-specific artifact would not generalize to held-out trials, the danger is not random trial noise but session- or block-level confounds shared across trials of the same subject. I request an empirical probe rather than a formal guarantee: for the within-subject datasets, either (i) split the cross-fitting folds by recording session or contiguous time blocks instead of random trials and report the resulting margins, or (ii) compare the cross-fitted display against a naively jointly trained display and against a display trained on permuted labels to bound the fraction of the gain that could come from trial- or session-level confounds.
6 minor objections
The paper says, at §5.1, Table 2
Pith says: minor objection 1 of 6
Please report per-seed standard deviations or confidence intervals for the test scores: with three seeds, several margins over the best baseline are small (PhysioMI balanced accuracy +2.1, TUAB +0.5), and the reader cannot tell whether those particular comparisons are resolved beyond noise even though the direction is consistent across datasets.
The paper says, at §4, Datasets
Pith says: minor objection 2 of 6
The paper repartitions validation and test subjects for HMC relative to the protocol of Xiong et al. [44]; please justify this deviation and state explicitly that all baselines were re-evaluated on the same repartitioned split so the comparison remains apples-to-apples.
show 4 more
The paper says, at Table 3 caption
Pith says: minor objection 3 of 6
The 'Best EEG FM' column should state explicitly that it is the per-dataset maximum over the seven baseline models rather than a single fixed model, since the gain of +4.6 points is computed against that per-dataset maximum.
The paper says, at §5.3, Figure 3
Pith says: minor objection 4 of 6
The per-channel kappa improvements (median 15.8 to 25.8 overall, 16.0 to 36.0 for left- versus right-hand imagery) are reported without spread or sample counts; adding the distribution across channels would strengthen the claim that the learned display increases task-relevant information in the frozen tokens.
The paper says, at §5.5, Figure 4
Pith says: minor objection 5 of 6
The per-window asymmetry index (0.45 for lateralized versus 0.07 for generalized discharges) is reported as a single number without variability; a per-window distribution or the number of recognized test windows would support the physiological localization claim.
The paper says, at Abstract and §5.1
Pith says: minor objection 6 of 6
The mean rank of 1.3 mixes Cohen kappa (multi-class tasks) with AUROC (binary tasks); please state in the main text which datasets contribute which metric to the ranking, as the abstract currently leaves this implicit.
What would settle it
Run an ablation on a within-subject dataset (e.g., PhysioMI or SEED) where the cross-fitting procedure is replaced by a naive joint training (no cross-fitting). If the performance difference between the two is small (e.g., < 2% in balanced accuracy), then the cross-fitting is not critical and the concern is mitigated. If the difference is large, it indicates that the display is learning shortcuts that the cross-fitting suppresses, and the paper should report this and discuss the implications for generalization. Additionally, compute the correlation between the learned display's output and subject identity; if it is high, the display is exploiting subject-specific traits.
from the skeptic's reading
How the review went
4 readings by deepseek-v4-flash · 13 Nov 2026
- Reader holds up, with conditions
- Skeptic agrees
- Editor wrote the verdict above
- Referee major revision: 1 major objection, 6 minor
Not run on this paper: the simulated authors' defence.
Pith did not check this review's objections against the paper.
The review in full
The review in brief, machine-written
Guoan,
This paper does something genuinely surprising: a frozen DINOv3 ViT-B/16, with no EEG pretraining, outperforms every EEG foundation model on all eight datasets. The method is clean—learn the display parameters (montage, filter, contrast) through the frozen backbone, keep the backbone frozen, and the display stays a valid page of EEG traces. The ablation in Table 3 is convincing: token-local strips already beat the best EEG foundation model on 6/8 datasets, and the learned display pushes that to 8/8. The image controls in Table 4 show that the performance depends on visual pretraining (random init drops 15.2 points) and on the token-local layout (half-row shift drops up to 29.7 points). The evidence maps in Figure 4 are a nice touch—they show the model localizes periodic discharges to the correct hemisphere, learned from window-level labels alone.
The cross-fitting concern the stress-test raises is real but the paper already acknowledges it (b98-b99) and implements a reasonable mitigation: cross-fitting across subjects for between-subject tasks, across trials for within-subject tasks. The paper does not provide an empirical test of whether this fully prevents subject-specific shortcuts, and that is a soft spot. But I think it's a minor one: the main result is robust across 8 datasets with different task structures, and the ablation in Table 3 shows that even the clinical montage (no learning) beats EEG foundation models on 6/8 datasets. The learned display adds 4.6 points on average, so the core claim does not rest entirely on the cross-fitting procedure. A more serious concern is the lack of error bars in Table 2—the paper reports means over three seeds but no standard deviations or confidence intervals. That is a standard expectation for a paper making comparative claims, and it should be addressed.
Who is this for? Anyone working on EEG representation learning, foundation models, or cross-modal transfer. The paper is well-written, the experiments are thorough, and the central claim is supported. I would cite it. I would bring it to a reading group. I would send it to peer review—the authors should add error bars and ideally an empirical check of the cross-fitting assumption, but the paper is solid enough to deserve a serious referee.
Best, [Your name]
The strongest objection
The paper's central claim is that a frozen ViT, with a learned display, outperforms all EEG foundation models. The weakest assumption is that the cross-fitting procedure (Eq. 7) prevents the display from exploiting subject-specific or trial-specific artifacts. The paper acknowledges this shortcut (b98-b99) and attempts to mitigate it by cross-fitting across subjects for between-subject tasks and across trials for within-subject tasks. However, the effectiveness of this mitigation is not empirically validated. For within-subject benchmarks (e.g., SEED, BCIC2020-3, PhysioMI), the folds split trials, not subjects. If the display learns trial-specific artifacts (e.g., noise patterns, eye blinks, or other non-neural signals that correlate with the label within a subject), the cross-fitting across trials may not prevent this, as the same subject appears in both training and validation folds. The paper does not provide a formal guarantee or an empirical test (e.g., comparing to a naive joint training, or showing that subject-specific traits are not learned). This is load-bearing because if the display exploits such shortcuts, the claim that the ViT is a better EEG encoder would be undermined, as the performance would not generalize to new subjects or new recording sessions. The paper's own limitation statement (b98-b99) confirms the concern but does not resolve it.
The reader's weakest_assumption correctly identifies the cross-fitting procedure and the potential for subject-specific shortcuts. The reader notes that the paper acknowledges this but does not provide a formal guarantee or empirical test. My analysis agrees with this assessment. The reader's verdict is CONDITIONAL, and I find that this is appropriate. The paper's central claim is strong and well-supported by the ablation studies and image controls, but the lack of validation for the cross-fitting procedure is a notable weakness. The concrete test I propose would directly address this concern. I do not find any other load-bearing concern that would change the verdict. The paper is otherwise sound, and the method is novel and significant. The agreement_with_reader is 'agree' because the reader's weakest_assumption is the same as mine.
The referee's report
If the comparative result holds, this is a significant challenge to the premise that EEG-specific pretraining is necessary for frozen-encoder transfer: it shows that a vision backbone that never saw EEG, combined with a learned display, can match or beat seven EEG-specific foundation models under a controlled protocol. The experimental design is a genuine strength: 8 datasets, 7 baselines, all encoders frozen with identical heads, validation-based selection of display, head, and ensemble with the test set untouched, cross-fitting to prevent head-display shortcuts, transfer of the same learned displays to two other backbones, and a random-init control whose 15.2-point drop ties the gain to visual pretraining. The paper also argues fairly by noting that on TUEV, TUAB, and TUSL the EEG baselines saw the relevant TUH recordings during pretraining and still lose to SeeEEG.
Pith's reading of the paper (a machine's summary, not the authors')
The paper argues that a frozen vision transformer (ViT), pretrained on natural images, can serve as a stronger encoder for EEG signals than any existing EEG foundation model—provided the EEG is rendered into an image that respects the ViT's patch structure. The authors introduce SeeEEG, a method that learns only how to display the EEG (montage, filtering, contrast) while keeping the vision backbone completely fixed. Across eight diverse EEG datasets, SeeEEG with DINOv3 ViT-B achieves a mean rank of 1.3 in Cohen kappa or AUROC, outperforming seven specialized EEG foundation models on every dataset. The key implication is that the visual priors from large-scale image pretraining already capture much of the structure needed for EEG analysis, reducing the need for massive EEG-specific pretraining.
What carries the argument. Token-local rendering: each EEG channel occupies a dedicated row of 16×16 patches, so that every token holds one channel over 16 time samples (80 ms at 200 Hz), either as grayscale amplitude (value view) or as a scalogram with 16 logarithmically spaced frequencies from 1 to 64 Hz. This layout ensures that task-relevant patterns fall within individual tokens, leveraging the ViT's patch embedding while keeping spatial relationships across electrodes within the image. The display module (montage weights, filter gains, contrast parameters) is initialized from a clinical montage and optimized with the task loss through the frozen backbone, using at most 1,113 trainable parameters per view.
Core claim. A frozen DINOv3 ViT-B/16, when fed EEG signals rendered as token-local strips (one channel per row of patches, either as a waveform or scalogram), outperforms every EEG foundation model (BENDR, BIOT, LaBraM, EEGPT, CBraMod, CSBrain, REVE) in balanced accuracy on all eight datasets tested, reaching a mean rank of 1.3 in Cohen kappa or AUROC versus 2.9 for the strongest baseline REVE. The learned displays transfer to other frozen vision backbones (V-JEPA 2.1, supervised ViT) without re-optimization, and a randomly initialized ViT drops performance by 15.2 points on average, confirming that the ability comes from visual pretraining, not architecture.
The premise it rests on. The cross-fitting procedure (Section 3.3) assumes that subject-specific traits are not correlated with the label in a way that survives splitting by subject (or by trial for within-subject tasks), and that the display cannot learn trial-specific artifacts that generalize to held-out subjects.
If the paper is right (Pith's inferences, not the authors')
EEG foundation models may not need massive EEG-specific pretraining; a frozen vision backbone with a learned display can exceed their performance.
Future advances in vision foundation models (e.g., larger ViTs, better self-supervised methods) can directly benefit EEG analysis by simply replacing the backbone and refitting the lightweight display and head.
The same principle likely extends to other multichannel physiological signals that experts interpret visually, such as MEG and ECG.
The learned displays are interpretable: the evidence maps localize class-discriminative patterns to specific electrode pairs and time intervals, consistent with clinical knowledge.
Because the vision backbone stays frozen, the approach is computationally efficient for multi-task deployment—only the display and head (few thousand parameters) need to be stored per task.
The success of token-local strips suggests that the ViT's patch embedding and attention mechanism, tuned to natural image statistics, can be repurposed for time-frequency analysis when the signal is arranged to match its inductive biases—effectively treating the ViT as a universal feature extractor for locally structured signals.
The cross-fitting procedure (Section 3.3) is a practical but heuristic defense against shortcut learning; a more rigorous theoretical understanding of when this procedure guarantees out-of-distribution generalization would strengthen the claim.
The paper implicitly challenges the value of continued scaling of EEG-specific pretraining corpora, suggesting that the bottleneck may be in the input representation rather than the model capacity or data volume.
One could test the limits of this approach by applying SeeEEG to signals that lack a natural visual analogy (e.g., high-density EEG with many channels, or non-periodic signals) to see where the visual prior fails.
Check this review
Pith stored no record of this review's passes and took no fingerprint of it. Reviews made by code from 13 Nov 2026 on carry both.
How often reviews like this one are right
Reviews like this one: written by deepseek-v4-flash, where what was read was not recorded.
How often objections from reviews like this one are right: not yet measured
Reviews by people
This review was written by a machine. No person has reviewed this paper on Pith yet. A scientist listed on Pith can review it for a fee, and their review is published here whatever it concludes.
The paper
Abstract
Electroencephalography (EEG) foundation models are pretrained on thousands of hours of recordings, yet their frozen features transfer poorly to downstream tasks, and this transfer does not consistently improve with larger models or more pretraining data. We show that a vision foundation model that has never seen EEG is a stronger frozen EEG encoder than these specialized models, once the signals are rendered as images that a vision transformer can read. Our method, SeeEEG, keeps the vision backbone frozen and learns only how EEG is displayed to it. Each channel occupies a dedicated row of patches and is drawn either as a waveform at a fixed gain or as a scalogram whose frequency range spans the patch height, so that task-relevant patterns fall within individual tokens. The montage, filtering, and contrast parameters of this display are initialized from a standard clinical montage and optimized with the task loss through the frozen backbone, which amounts to at most 1,113 trainable parameters per view. We compare SeeEEG with seven EEG foundation models on eight clinical, sleep, affective, and brain-computer interface datasets, with every encoder frozen and paired with identical readout heads. With a frozen DINOv3 ViT-B, SeeEEG outperforms every EEG foundation model in balanced accuracy on all eight datasets and reaches a mean rank of 1.3 in Cohen kappa or AUROC, compared with 2.9 for the strongest baseline, REVE. The learned displays transfer without further optimization to the video encoder V-JEPA 2.1, which outperforms every EEG foundation model on both metrics on six of the eight datasets. Replacing DINOv3 with a randomly initialized ViT-B lowers the score by 15.2 points on average, so the ability to read rendered EEG comes from visual pretraining.
The paper's own sentences, arranged, kept beside the review and separate from it. Each label is drawn from a closed vocabulary; the number beside it is the probability the classifier gave that label. No sentence in this record was written by a model.
abstract compiled · 2026-11-13
The authors' main claim, as the classifier picked it out 1.00
“We show that a vision foundation model that has never seen EEG is a stronger frozen EEG encoder than these specialized models, once the signals are rendered as images that a vision transformer can read.”
How the main claim is stated: asserted 1.00
What the abstract itself states, as probabilities: claims first 0.08 · states limitation 0.39 · human participants 0.73 · controlled comparison 0.95 · randomized assignment 0.02 · states quantitative result 0.99 · compares with prior results 0.98 · mentions code or data availability 0.03
The abstract, sentence by sentence
This table labels the abstract only (arXiv listing abstract v1). The number is the probability the classifier gave the label.
| Sentence | Role | The authors' own work | States the main conclusion | Strength |
|---|---|---|---|---|
| Electroencephalography (EEG) foundation models are pretrained on thousands of hours of recordings, yet their frozen features transfer poorly to downstream tasks, and this transfer does not consistently improve with larger models or more pretraining data. | background 0.80 | 0.09 | 0.03 | asserted 1.00 |
| We show that a vision foundation model that has never seen EEG is a stronger frozen EEG encoder than these specialized models, once the signals are rendered as images that a vision transformer can read. | result 0.66 | 0.93 | 0.90 | asserted 1.00 |
| Our method, SeeEEG, keeps the vision backbone frozen and learns only how EEG is displayed to it. | method 0.86 | 0.81 | 0.12 | asserted 0.98 |
| Each channel occupies a dedicated row of patches and is drawn either as a waveform at a fixed gain or as a scalogram whose frequency range spans the patch height, so that task-relevant patterns fall within individual tokens. | method 1.00 | 0.84 | 0.03 | asserted 0.98 |
| The montage, filtering, and contrast parameters of this display are initialized from a standard clinical montage and optimized with the task loss through the frozen backbone, which amounts to at most 1,113 trainable parameters per view. | method 1.00 | 0.92 | 0.03 | asserted 1.00 |
| We compare SeeEEG with seven EEG foundation models on eight clinical, sleep, affective, and brain-computer interface datasets, with every encoder frozen and paired with identical readout heads. | method 0.97 | 0.94 | 0.04 | asserted 0.94 |
| With a frozen DINOv3 ViT-B, SeeEEG outperforms every EEG foundation model in balanced accuracy on all eight datasets and reaches a mean rank of 1.3 in Cohen kappa or AUROC, compared with 2.9 for the strongest baseline, REVE. | result 1.00 | 0.96 | 0.58 | asserted 1.00 |
| The learned displays transfer without further optimization to the video encoder V-JEPA 2.1, which outperforms every EEG foundation model on both metrics on six of the eight datasets. | result 1.00 | 0.92 | 0.15 | asserted 1.00 |
| Replacing DINOv3 with a randomly initialized ViT-B lowers the score by 15.2 points on average, so the ability to read rendered EEG comes from visual pretraining. | interpretation 0.63 | 0.96 | 0.08 | asserted 0.98 |
What kind of paper the classifier reports this is: method or tool 0.99 · evidence from observation or experiment 0.90
What was examined, and what was not
Sources read
- arXiv listing abstract v1
Not examined
- full text
- figures and tables
- references and citing sentences
- relations to other papers
Levels run
- P abstract
Not run
- S sections
- T body sentences
- C citations
- R cross-paper relations
How the labels were made
Compiled by classifier-1.13.0 via TypeSafe API, question bank abstract-p1-20260923, 2026-11-13: 1 request, 7,224 input tokens.
Every quoted sentence on this record is the authors'. Every label is from Pith's closed vocabulary, with the probability shown beside it. This record as JSON
Figures (4)
Figures are the authors', reproduced from the paper.
References (51)
- [1] Diego Alvarez-Estevez and Roselyne Rijsman. Haaglanden Medisch Centrum sleep staging database. PhysioNet, 2022. Version 1.1. doi:10.13026/t79q-fr32. 5 on Pithdoi:10.13026/t79q-fr32
- [2] Diego Alvarez-Estevez and Roselyne M. Rijsman. Inter- database validation of a deep learning approach for automatic sleep scoring.PLOS ONE, 16(8):e0256111, 2021. 5 on Pith
- [3] Parastoo Azizeddin, Omid Sharafi, and Maryam M. Shanechi. Multimodal LLMs can learn to read brain signals: A vision– language model for unified multi-task EEG decoding, 2026. NeurIPS 2026 Workshop on Foundation Models for the Brain and Body. 2, 3 on Pith
- [4] Randall Balestriero and Yann LeCun. Proving and improv- ing the Platonic representation hypothesis with JEPAs, 2026. arXiv:2610.39302. 2 on PithPith's reviewarXiv
- [5] Benjamin Blankertz, Ryota Tomioka, Steven Lemm, Motoaki Kawanabe, and Klaus-Robert M ¨uller. Optimizing spatial filters for robust EEG single-trial analysis.IEEE Signal Pro- cessing Magazine, 25(1):41–56, 2008. 4 on Pith
- [6] Mouxiang Chen, Lefei Shen, Zhuo Li, Xiaoyun Joy Wang, Jianling Sun, and Chenghao Liu. VisionTS: Visual masked autoencoders are free-lunch zero-shot time series forecast- ers. InProceedings of the 42nd International Conference on Machine Learning, pages 8979–9007. PMLR, 2025. 2, 3 on Pith
- [7] Yassine El Ouahidi, Jonathan Lys, Philipp Th¨olke, Nicolas Farrugia, Bastien Pasdeloup, Vincent Gripon, Karim Jerbi, and Giulia Lioi. REVE: A foundation model for EEG – adapt- ing to any setup with large-scale pretraining on 25,000 sub- jects. InAdvances in Neural Information ... on Pith
- [8] Ary L. Goldberger, Luis A. N. Amaral, Leon Glass, Jeffrey M. Hausdorff, Plamen Ch. Ivanov, Roger G. Mark, Joseph E. Mietus, George B. Moody, Chung-Kang Peng, and H. Eugene Stanley. PhysioBank, PhysioToolkit, and PhysioNet: Com- ponents of a new research resource for complex ph... on Pith
- [9] Meysam Golmohammadi, Amir Hossein Harati Nejad Tor- bati, Silvia Lopez de Diego, Iyad Obeid, and Joseph Picone. Automatic analysis of EEGs using big data and hybrid deep learning architectures.Frontiers in Human Neuroscience, 13: 76, 2019. 4 on Pith
- [10] Yuan Gong, Yu-An Chung, and James Glass. AST: Audio spectrogram transformer. InProc. Interspeech 2021, pages 571–575, 2021. 3 on Pith
- [11] Amir Harati, Meysam Golmohammadi, Silvia Lopez, Iyad Obeid, and Joseph Picone. Improved EEG event classification using differential energy. In2015 IEEE Signal Processing in Medicine and Biology Symposium (SPMB), pages 1–4, 2015. 5 on Pith
- [12] Bo Hjorth. An on-line transformation of EEG scalp potentials into orthogonal source derivations.Electroencephalography and Clinical Neurophysiology, 39(5):526–530, 1975. 4 on Pith
- [13] Minyoung Huh, Brian Cheung, Tongzhou Wang, and Phillip Isola. Position: The Platonic representation hypothesis. In Proceedings of the 41st International Conference on Machine Learning, pages 20617–20642, 2024. 2, 8 on Pith
- [14] Ji-Hoon Jeong, Jeong-Hyun Cho, Young-Eun Lee, Seo-Hyun Lee, Gi-Hwan Shin, Young-Seok Kweon, Jos´e del R. Mill´an, Klaus-Robert M¨uller, and Seong-Whan Lee. 2020 interna- tional brain–computer interface competition: A review.Fron- tiers in Human Neuroscience, 16:898300, 2022. 5 on Pith
- [15] Wei-Bang Jiang, Li-Ming Zhao, and Bao-Liang Lu. Large brain model for learning generic representations with tremen- dous EEG data in BCI. InInternational Conference on Learn- ing Representations (ICLR), pages 16405–16426, 2024. 2, 3, 5, 6 on Pith
- [16] Agnan Kessy, Alex Lewin, and Korbinian Strimmer. Optimal whitening and decorrelation.The American Statistician, 72 (4):309–314, 2018. 4 on Pith
- [17] Demetres Kostas, St´ephane Aroca-Ouellette, and Frank Rudz- icz. BENDR: Using transformers and a contrastive self- supervised learning task to learn from massive amounts of EEG data.Frontiers in Human Neuroscience, 15:653659, on Pith
- [18] Na Lee, Konstantinos Barmpas, Yannis Panagakis, Dimitrios Adamos, Nikolaos Laskaris, and Stefanos Zafeiriou. Are large brainwave foundation models capable yet? Insights from fine- tuning. InProceedings of the 42nd International Conference on Machine Learning, pages 32878–32888... on Pith
- [19] Chenqi Li, Yu Liu, Shuo Zhang, Timothy Denison, and Tingt- ing Zhu. Standing on the shoulders of giants: Rethinking EEG foundation model pretraining via multi-teacher distilla- tion. InAdvances in Neural Information Processing Systems, on Pith
- [20] Zekun Li, Shiyang Li, and Xifeng Yan. Time series as images: Vision transformer for irregularly sampled time series. In Advances in Neural Information Processing Systems, pages 49187–49204, 2023. 3 on Pith
show 31 more
- [21] Dingkun Liu, Yuheng Chen, Zhu Chen, Zhenyao Cui, Yaozhi Wen, Jiayu An, Jingwei Luo, and Dongrui Wu. EEG-FM- Compass: progress, benchmarking and future directions for EEG foundation models.National Science Review, page nwag466, 2026. 2, 3 on Pith
- [22] S. L ´opez, G. Suarez, D. Jungreis, I. Obeid, and J. Picone. Automated identification of abnormal adult EEGs. In2015 IEEE Signal Processing in Medicine and Biology Symposium (SPMB), pages 1–5, 2015. 5 on Pith
- [23] S. L´opez, A. Gross, S. Yang, M. Golmohammadi, I. Obeid, and J. Picone. An analysis of two common reference points for EEGs. In2016 IEEE Signal Processing in Medicine and Biology Symposium (SPMB), pages 1–5, 2016. 4 on Pith
- [24] Silvia L´opez de Diego. Automated interpretation of abnor- mal adult electroencephalograms. Master’s thesis, Temple University, 2017. 5 on Pith
- [25] Andreas Miltiadous, Katerina D. Tzimourta, Theodora Afran- tou, Panagiotis Ioannidis, Nikolaos Grigoriadis, Dimitrios G. Tsalikakis, Pantelis Angelidis, Markos G. Tsipouras, Euri- pidis Glavas, Nikolaos Giannakeas, and Alexandros T. Tzallas. A dataset of scalp EEG recordings o... on Pith
- [26] Lorenzo Mur-Labadia, Matthew Muckley, Amir Bar, Mido Assran, Koustuv Sinha, Mike Rabbat, Yann LeCun, and Nico- las Ballas. V-JEPA 2.1: Unlocking dense features in video self-supervised learning. InComputer Vision – ECCV 2026, pages 671–689. Springer Nature Switzerland, 2026. 6 on Pith
- [27] Iyad Obeid and Joseph Picone. The Temple University Hos- pital EEG data corpus.Frontiers in Neuroscience, 10:196, on Pith
- [28] Robert Oostenveld and Peter Praamstra. The five percent elec- trode system for high-resolution EEG and ERP measurements. Clinical Neurophysiology, 112(4):713–719, 2001. 4 on Pith
- [29] Maxime Oquab, Timoth´ee Darcet, Th´eo Moutakanni, Huy V . V o, Marc Szafraniec, Vasil Khalidov, Pierre Fernandez, Daniel Haziza, Francisco Massa, Alaaeldin El-Nouby, Mah- moud Assran, Nicolas Ballas, Wojciech Galuba, Russell Howes, Po-Yao Huang, Shang-Wen Li, Ishan Misra, Mich... on Pith
- [30] Alec Radford, Jong Wook Kim, Chris Hallacy, Aditya Ramesh, Gabriel Goh, Sandhini Agarwal, Girish Sastry, Amanda Askell, Pamela Mishkin, Jack Clark, Gretchen Krueger, and Ilya Sutskever. Learning transferable visual models from natural language supervision. InProceedings of the... on Pith
- [31] Herbert Ramoser, Johannes M ¨uller-Gerking, and Gert Pfurtscheller. Optimal spatial filtering of single trial EEG during imagined hand movement.IEEE Transactions on Rehabilitation Engineering, 8(4):441–446, 2000. 4 on Pith
- [32] Simon Roschmann, Quentin Bouniot, Vasilii Feofanov, Iev- gen Redko, and Zeynep Akata. Unlocking pretrained vi- sion transformers for time series classification. InGer- man Conference on Pattern Recognition (GCPR), 2026. arXiv:2506.08641. 2, 3 on PithPith's reviewarXiv
- [33] Gerwin Schalk. EEG motor movement/imagery dataset. Phy- sioNet, 2009. Version 1.0.0. doi:10.13026/C28G6P. 5 on Pithdoi:10.13026/c28g6p
- [34] Gerwin Schalk, Dennis J. McFarland, Thilo Hinterberger, Niels Birbaumer, and Jonathan R. Wolpaw. BCI2000: A general-purpose brain-computer interface (BCI) system. IEEE Transactions on Biomedical Engineering, 51(6):1034– 1043, 2004. 5 on Pith
- [35] Oriane Sim´eoni, Huy V . V o, Maximilian Seitzer, Federico Baldassarre, Maxime Oquab, Cijo Jose, Vasil Khalidov, Marc Szafraniec, Seungeun Yi, Micha¨el Ramamonjisoa, Francisco Massa, Daniel Haziza, Luca Wehrstedt, Jianyuan Wang, Tim- oth´ee Darcet, Th ´eo Moutakanni, Leonel Se... on Pith
- [36] Andreas Steiner, Alexander Kolesnikov, Xiaohua Zhai, Ross Wightman, Jakob Uszkoreit, and Lucas Beyer. How to train your ViT? Data, augmentation, and regularization in vision transformers.Transactions on Machine Learning Research, on Pith
- [37] Yulin Sun, Xiaopeng Si, Runnan He, Xiao Hu, Peter Smielewski, Wenlong Wang, Xiaoguang Tong, Wei Yue, Mei- jun Pang, Kuo Zhang, Xizi Song, Dong Ming, and Xiuyun Liu. An automated classifier of harmful brain activities for clinical usage based on a vision-inspired pre-trained fr... on Pith
- [38] Christopher Torrence and Gilbert P. Compo. A practical guide to wavelet analysis.Bulletin of the American Meteorological Society, 79(1):61–78, 1998. 3 on Pith
- [39] Eva von Weltin, Tameem Ahsan, Vinit Shah, Dawer Jamshed, Meysam Golmohammadi, Iyad Obeid, and Joseph Picone. Electroencephalographic slowing: A primary source of error in automatic seizure detection. In2017 IEEE Signal Process- ing in Medicine and Biology Symposium (SPMB), pages 1–5, on Pith
- [40] Guangyu Wang, Wenchao Liu, Yuhong He, Cong Xu, Lin Ma, and Haifeng Li. EEGPT: Pretrained transformer for universal and reliable representation of EEG signals. InAdvances in Neural Information Processing Systems, pages 39249–39280, on Pith
- [41] Jiquan Wang, Sha Zhao, Zhiling Luo, Yangxuan Zhou, Hait- eng Jiang, Shijian Li, Tao Li, and Gang Pan. CBraMod: A criss-cross brain foundation model for EEG decoding. InIn- ternational Conference on Learning Representations (ICLR), pages 75310–75346, 2025. 2, 3, 5, 6 on Pith
- [42] Ross Wightman. PyTorch image models. https : //github.com/huggingface/pytorch- image- models, 2019. 6 on Pith
- [43] Jiamin Wu, Zichen Ren, Junyu Wang, Pengyu Zhu, Yonghao Song, Mianxin Liu, Qihao Zheng, Lei Bai, Wanli Ouyang, and Chunfeng Song. AdaBrain-Bench: Benchmarking brain foundation models for brain-computer interface applications, on Pith
- [44] Wei Xiong, Jiangtong Li, Jie Li, Kun Zhu, and Changjun Jiang. EEG-FM-Bench: A comprehensive benchmark for the systematic evaluation and diagnostic analyses of EEG foundation models. InProceedings of the 43rd International Conference on Machine Learning, pages 140324–140359. PM... on Pith
- [45] Chaoqi Yang, M. Brandon Westover, and Jimeng Sun. BIOT: Biosignal transformer for cross-data learning in the wild. In Advances in Neural Information Processing Systems, pages 78240–78260, 2023. 2, 3, 5, 6 on Pith
- [46] Liuyin Yang, Qiang Sun, Ang Li, and Marc M. Van Hulle. Are EEG foundation models worth it? Comparative evaluation with traditional decoders in diverse BCI tasks. InInterna- tional Conference on Learning Representations (ICLR), 2026. 2, 3 on Pith
- [47] Wei-Long Zheng and Bao-Liang Lu. Investigating critical frequency bands and channels for EEG-based emotion recog- nition with deep neural networks.IEEE Transactions on Autonomous Mental Development, 7(3):162–175, 2015. 5 on Pith
- [48] Jiahao Zhou, Chenghao Xu, Wei Wang, Erkun Yang, and Cheng Deng. EEGiT: Teaching vision transformers to under- stand the EEG signal. InProceedings of the IEEE/CVF Con- ference on Computer Vision and Pattern Recognition (CVPR), pages 40438–40447, 2026. 2, 3 on Pith
- [49] Yuchen Zhou, Jiamin Wu, Zichen Ren, Zhouheng Yao, Wei- heng Lu, Kunyu Peng, Qihao Zheng, Chunfeng Song, Wanli Ouyang, and Chao Gou. CSBrain: A cross-scale spatiotempo- ral brain foundation model for EEG decoding. InAdvances in Neural Information Processing Systems, pages 87150–87195, on Pith
- [2025] arXiv:2507.09882. 2, 3 on PithPith's reviewarXiv
- [2026] To appear. arXiv:2603.04478. 2, 3, 6 on PitharXiv
Challenges and discussion
Disagree with the review, or know something it missed? Leave a comment, or challenge this review formally: a challenge becomes a permanent entry in this paper's record, and its decision is published here.
Come back by email
A short list of the week's newly reviewed papers. Nothing is sent until you confirm from your inbox, and every email has a link that stops it. No account.
No comments yet.
Sign in to comment