Pith. sign in

REVIEW 4 major objections 5 minor 144 references

Bridging Brain with Foundation Models through Self-Supervised Learning

T0 review · 4 major / 5 minor · reviewed 2026-08-06 · deepseek-v4-flash

Pith's one-line read This survey argues that self-supervised learning on unlabeled brain signals can grow into brain foundation models, but 64 reviewed studies show the field has not yet reached task-agnostic, zero-shot representations.

desk verdict A useful survey of self-supervised learning for brain signals, but its central Table 3 misclassifies the modalities of core foundation models, so the landscape claims need a full recheck before the paper can be trusted. read the letter →

arxiv 2506.16009 v1 pith:IC4KTCM2 submitted 2025-06-19 cs.LG

classification cs.LG
keywords EEGself-supervisedlearningbrainfoundationmodelscontrastivemaskedmodelingbrain-computerinterfacemultimodalEEG-to-textdecoding
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 surveys the young field that trains foundation models on brain signals, mostly EEG, using self-supervised learning. It organizes the field into self-predictive methods (masked modeling, autoregressive prediction) and contrastive methods (instance discrimination, multimodal alignment), and it reviews the resulting brain foundation models, evaluation protocols, and datasets. The central message is that SSL pretraining helps but has not yet delivered the task-agnostic, zero-shot neural representations that foundation models achieve in language and vision. A second load-bearing claim is that some EEG-to-text results may be inflated by teacher forcing and a lack of input sensitivity, so the survey calls for stricter evaluation baselines.

What carries the argument

The organizing machinery is a two-way taxonomy of self-supervised learning objectives, together with the standard evaluation ladder of KNN, linear probing, and full fine-tuning. A brain signal is treated as a matrix $X \in \mathbb{R}^{C\times T}$ of channels by time points, segmented into fixed-length patches that serve as tokens; models either reconstruct masked or future patches (self-predictive) or align positive pairs while separating negatives (contrastive). A second key mechanism in the most ambitious models is a vector-quantized neural tokenizer that turns continuous EEG patches into discrete codebook tokens, letting an LLM-style transformer treat EEG as a foreign language, as in LaBraM and NeuroLM. The survey uses this machinery to compare models, to explain why pretraining benefits transfer, and to identify where current SSL objectives miss the neurobiological structure of brain signals.

What would settle it

Take any published EEG-to-text model and re-evaluate it with teacher forcing removed and with a noise-only control input; if the noise control performs as well as the EEG-conditioned model, the survey's claim that reported results are inflated is confirmed, whereas a clear EEG-minus-noise accuracy gap would undercut that claim.

Watch

Extended reading notes

Core claim

The paper's discovery is a structured map of how self-supervised learning is being applied to build brain foundation models, together with a critical assessment of how far the field has actually come. After systematically selecting 64 studies from 2020 through early 2025, it finds that masked modeling and autoregressive prediction dominate self-predictive pretraining, while contrastive learning is used both within a single brain modality and across modalities such as EEG, fMRI, text, and images. Models like LaBraM, BrainWave, and NeuroLM show that large-scale pretraining on heterogeneous EEG corpora improves downstream classification and even enables some few-shot or instruction-tuned behavior. However, the paper concludes that current brain foundation models still depend heavily on task-specific fine-tuning and have yet to bridge the gap toward truly task-agnostic, general-purpose neural representations with zero-shot capabilities. It also reports evidence from a study by Jo et al. that many EEG-to-text models may not be learning from EEG at all, since they perform similarly when trained on random noise and their scores drop sharply when teacher forcing is removed at evaluation.

Load-bearing premise

The review's conclusions depend on the 64 selected studies reporting their methods and performance numbers accurately, yet the paper's own tables include rows marked 'No mention found' and it cites a study showing some EEG-to-text models fail to learn from EEG at all.

Editorial extensions

If this is right

  • If the survey's assessment holds, future brain foundation models should not simply copy NLP and vision SSL recipes but should design pretext tasks grounded in functional connectivity, hierarchical processing, and oscillatory dynamics.
  • EEG-to-text results reported with teacher forcing at evaluation should be re-run without it and against noise-only baselines before being trusted.
  • Because MEG and iEEG consistently outperform EEG in SSL decoding benchmarks, portable MEG technology and invasive recordings may set the upper bound for what EEG-only self-supervised models can achieve.
  • Adopting a common evaluation protocol of frozen-backbone KNN and linear probing would make reported gains across brain foundation models directly comparable.

Reading between the lines

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

  • The authors do not say so explicitly, but their evidence implies that the field's near-universal fine-tuning requirement is itself a signal that current pretext tasks are not learning the right invariances, and that a benchmark measuring zero-shot transfer on held-out tasks would expose this clearly.
  • A natural extension is to combine vector-quantized neural tokenization with montage-aware or anatomy-aware channel grouping, since the survey's own discussion suggests that random patch masking loses the spatial topology that makes brain signals structured.
  • The noise-ablation test cited for EEG-to-text could be applied just as easily to EEG-to-image decoding, where similar inflated-reporting risks may exist but have not yet been systematically checked.
  • If optically pumped magnetometer MEG becomes portable and affordable, the same SSL methods reviewed here may shift their center of gravity from EEG to MEG, because the paper documents consistent MEG advantages in decoding accuracy.
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.

Referee Report

4 major / 5 minor

Summary. This manuscript is a survey of self-supervised learning (SSL) applied to brain signals, with a focus on EEG. It proposes a taxonomy that divides SSL into self-predictive and contrastive methods, reviews brain foundation models and their evaluation, discusses multimodal integration of brain signals with text, images, and audio, catalogs EEG datasets, and concludes with a critical discussion of open challenges. The central claims are that the field can be organized along this dichotomous taxonomy, that current brain foundation models do not yet achieve task-agnostic zero-shot capability, and that some EEG-to-text results are inflated due to teacher forcing and lack of input sensitivity.

Significance. If the survey's synthesis were accurate, it would provide a useful entry point for researchers entering this rapidly growing area. The paper has clear strengths: it follows a PRISMA-based selection procedure, it covers a broad range of recent models and datasets, and it includes a well-calibrated warning about the reliability of EEG-to-text evaluation based on the Jo et al. study (Section 8.2). The proposed self-predictive vs. contrastive taxonomy is a reasonable organizing frame. However, the value of a survey depends on the fidelity of its reported landscape, and the table of brain foundation models contains modality classifications that are verifiably wrong for several prominent models. Because the cross-modality discussion in Section 8.1 builds on this landscape, the inaccuracies are load-bearing and require correction before the survey can be considered reliable.

major comments (4)
  1. [Table 3 and Section 8.1] The modality column of Table 3 misclassifies several foundation models in ways that are verifiable from the cited primary sources: BrainLM [83] is an fMRI foundation model but is listed as EEG; Brain-JEPA [37] is described in the text as 'particularly suited for fMRI data' but appears in an EEG row; AnatCL [82] is trained on anatomical MRI, and its own row reports '21,155 MRI images' yet is categorized as EEG; BrainSegFounder [85] is a 3D neuroimage segmentation model; and FM-APP [87] is an fMRI-to-sMRI model. These errors affect the set of models treated as EEG foundation models, the EEG-vs-other-modality comparisons in Section 8.1, and the overall taxonomy. Please correct the modality entries against the primary papers and re-evaluate any claims that depend on the resulting distribution.
  2. [Section 7.5 and Section 4.1] The manuscript contains structural inconsistencies that impede verification: Section 7.5 refers to 'Table 5' for EEG datasets, but the datasets table is labeled 'Table 4', and Section 4.1 appears twice, once for data extraction before Section 4 and once for data processing inside Section 4. This makes it difficult to follow the survey's own organization and suggests the manuscript has not been carefully assembled. Please reconcile the section and table numbering.
  3. [Sections 4.2, 5, and 8.2] Several citations are incomplete placeholders rather than references: Section 4.2 has '[ref]' for LLaMA and Claude, Section 5 has '[ref]' for Vision Transformers, and Section 8.2 has 'BART [], T5 [], PEGASUS []'. A survey must provide complete bibliographic entries for every cited work; these placeholders make it impossible for readers to check the claims that depend on them. Please fill in all missing references before resubmission.
  4. [Tables 1, 2, and 3] The tables repeatedly report 'No mention found' for performance metrics and omit entries in the 'Code' and 'Pretrained data size' columns. While it is acceptable to state that a primary paper does not report a value, the current phrasing is informal and ambiguous: it does not distinguish 'not reported' from 'not found in our extraction'. Please standardize the notation (e.g., 'Not reported') and ensure that every row has a consistent level of detail so that the comparative claims in the text can be evaluated.
minor comments (5)
  1. [Section 5 (Evaluation of SSL-FMs)] The text says 'Linear Evaluation (known as liner-probing evaluation)'; the correct term is 'linear probing'.
  2. [Section 5 (BrainWave)] The phrase 'BrainWave the lased version of them' appears to be a typo; it should likely read 'the latest version'.
  3. [Section 8.2] The phrase 'papers refereed by [140]–[144]' should be 'papers reviewed by' or 'papers referenced by', as 'refereed' has a different meaning.
  4. [Table 3 (CBraMod row)] The URL for CBraMod is split across lines as 'https: //github.com/wjq-learning/CBraMod'; please fix the line break so the URL is a single string.
  5. [Table 3 (BrainBERT row)] BrainBERT is described in the text as an intracranial recording model (iEEG/ECoG), but the table lists its modality as 'EEG, audio'; please reconcile this with the primary source.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the survey's claims are descriptive syntheses of external studies, with no derivation chain that reduces to its own inputs.

full rationale

This paper is a literature survey, not a derivation. Its central claims — that SSL for brain signals can be organized into self-predictive and contrastive learning, that certain brain foundation models exist with given properties, and that current SSL brain models still rely heavily on fine-tuning — are supported by cited external studies and by the paper's own summaries of those studies. I checked the load-bearing reasoning in Sections 4, 5, and 8. No equation in the paper defines an output in terms of an input it then claims to predict; no fitted parameter is renamed as a prediction; and no 'uniqueness theorem' or ansatz is imported from the authors' prior work. The author self-citations (e.g., [6]–[9], [95], [96], [98]) appear only as background references on EEG motor-imagery deep learning and do not justify any central premise of the survey. The skeptical observation that Table 3 mislabels modalities of several foundation models is a correctness or internal-consistency problem, not circularity: the survey's taxonomy is not defined in terms of those models in a way that forces the conclusions. The paper also explicitly flags its own limitation that current EEG-to-text results may be inflated by teacher forcing and input insensitivity, relying on the independent external study by Jo et al. [138], which further indicates the survey is not circularly self-confirming. Score 0 is therefore the honest finding.

Assumptions & free parameters 0 free parameters · 3 assumptions · 0 invented entities

No new parameters, models, or entities are introduced. The survey's claims rest on secondary synthesis of published work; the main assumptions are the accuracy of primary studies, the completeness of the literature search, and the chosen taxonomy.

assumptions (3)
  • domain assumption Reported results in the 64 surveyed studies are accurate and representative of the field.
    The survey aggregates metrics and method descriptions from primary papers without re-running experiments; Section 3 and Tables 1 to 3 rely on these reports.
  • domain assumption The PRISMA search over Web of Science and PubMed, plus reference tracing, covers the relevant literature.
    Section 3 defines inclusion criteria; if these databases miss major venues, the synthesis could be biased. The survey notes that some arXiv preprints were added through tracing.
  • ad hoc to paper The binary taxonomy dividing SSL into self-predictive and contrastive learning is a valid organizing frame.
    Section 4 states that the authors find this division the most intuitive; other taxonomies exist and this choice affects how methods are grouped and compared.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Bridging Brain with Foundation Models through Self-Supervised Learning." pith.science (2026). https://pith.science/paper/IC4KTCM2

@misc{pith2026250616009,
  author       = {Pith},
  title        = {Pith review of: Bridging Brain with Foundation Models through Self-Supervised Learning},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/IC4KTCM2}},
  note         = {Machine review of arXiv:2506.16009}
}
read the original abstract

Foundation models (FMs), powered by self-supervised learning (SSL), have redefined the capabilities of artificial intelligence, demonstrating exceptional performance in domains like natural language processing and computer vision. These advances present a transformative opportunity for brain signal analysis. Unlike traditional supervised learning, which is limited by the scarcity of labeled neural data, SSL offers a promising solution by enabling models to learn meaningful representations from unlabeled data. This is particularly valuable in addressing the unique challenges of brain signals, including high noise levels, inter-subject variability, and low signal-to-noise ratios. This survey systematically reviews the emerging field of bridging brain signals with foundation models through the innovative application of SSL. It explores key SSL techniques, the development of brain-specific foundation models, their adaptation to downstream tasks, and the integration of brain signals with other modalities in multimodal SSL frameworks. The review also covers commonly used evaluation metrics and benchmark datasets that support comparative analysis. Finally, it highlights key challenges and outlines future research directions. This work aims to provide researchers with a structured understanding of this rapidly evolving field and a roadmap for developing generalizable brain foundation models powered by self-supervision.

Figures

Figures reproduced from arXiv: 2506.16009 by the authors.

Figure 8
Figure 8. Overview of self-supervised learning strategies: (top) self￾predictive learning methods, which reconstruct or predict parts of the input (e.g., masked, corrupted, or future data); (bottom) contrastive learning, which aligns representations of positive pairs (e.g., augmented views or cross-modal pairs) while separating negatives. Before applying SSL strategies, it is essential to preprocess neural data to address the… view at source ↗

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

144 extracted references · 61 canonical work pages

  1. [83]

    BrainLM: A foundation model for brain activity recordings,

    J. O. Caro et al., “BrainLM: A foundation model for brain activity recordings,” in The Twelfth International Conference on Learning Representations, 2024

  2. [37]

    Brain-JEPA: Brain dynamics foundation model with gradient positioning and spatiotemporal masking,

    Z. Dong et al., “Brain-JEPA: Brain dynamics foundation model with gradient positioning and spatiotemporal masking,” Adv. Neural Inf. Process. Syst., vol. 37, pp. 86048–86073, 2024

  3. [82]

    Anatomical Foundation Models for Brain MRIs,

    C. A. Barbano, M. Brunello, B. Dufumier, and M. Grangetto, “Anatomical Foundation Models for Brain MRIs,” arXiv Prepr. arXiv2408.07079, 2024

  4. [85]

    BrainSegFounder: towards 3D foundation models for neuroimage segmentation,

    J. Cox et al., “BrainSegFounder: towards 3D foundation models for neuroimage segmentation,” Med. Image Anal. , vol. 97, p. 103301, 2024

  5. [87]

    FM -APP: Foundation Model for Any Phenotype Prediction via fMRI to sMRI Knowledge Transfer,

    Z. He et al. , “FM -APP: Foundation Model for Any Phenotype Prediction via fMRI to sMRI Knowledge Transfer,” IEEE Trans. Med. Imaging, 2024

  6. [1]

    A cookbook of self -supervised learning,

    R. Balestriero et al., “A cookbook of self -supervised learning,” arXiv Prepr. arXiv2304.12210, 2023

  7. [2]

    Extracting and composing robust features with denoising autoencoders,

    P. Vincent, H. Larochelle, Y. Bengio, and P. -A. Manzagol, “Extracting and composing robust features with denoising autoencoders,” in Proceedings of the 25th international conference on Machine learning, 2008, pp. 1096–1103

  8. [3]

    Stacked denoising autoencoders: Learning useful representations in a deep network with a local denoising criterion.,

    P. Vincent, H. Larochelle, I. Lajoie, Y. Bengio, P. -A. Manzagol, and L. Bottou, “Stacked denoising autoencoders: Learning useful representations in a deep network with a local denoising criterion.,” J. Mach. Learn. Res., vol. 11, no. 12, 2010

Show all 144 references
  1. [4]

    Efficient estimation of word representations in vector space,

    T. Mikolov, K. Chen, G. Corrado, and J. Dean, “Efficient estimation of word representations in vector space,” arXiv Prepr. arXiv1301.3781, 2013

  2. [5]

    Attention is all you need,

    A. Vaswani et al. , “Attention is all you need,” in Advances in neural information processing systems, 2017, pp. 5998–6008

  3. [6]

    Dynamic convolution with multilevel attention for EEG -based motor imagery decoding,

    H. Altaheri, G. Muhammad, and M. Alsulaiman, “Dynamic convolution with multilevel attention for EEG -based motor imagery decoding,” IEEE Internet Things J., p. 1, 2023

  4. [7]

    Attention -Inception and Long Short -Term Memory - based Electroencephalography Classification for Motor Imagery Tasks in Rehabilitation,

    S. U. Amin, H. Altaheri, G. Muhammad, M. Alsulaiman, and A. Wadood, “Attention -Inception and Long Short -Term Memory - based Electroencephalography Classification for Motor Imagery Tasks in Rehabilitation,” IEEE Trans. Ind. Informatics , vol. 18, no. 8, pp. 5412–5421, 2022

  5. [8]

    A Multi -Branch Convolutional Neural Network with Squeeze - and-Excitation Attention Blocks for EEG -Based Motor Imagery Signals Classification,

    G. A. Altuwaijri, G. Muhammad, H. Altaheri, and M. Alsulaiman, “A Multi -Branch Convolutional Neural Network with Squeeze - and-Excitation Attention Blocks for EEG -Based Motor Imagery Signals Classification,” Diagnostics, vol. 12, no. 4, pp. 1 –16, 2022

  6. [9]

    Deep learning techniques for classification of electroencephalogram (EEG) motor imagery (MI) signals: a review,

    H. Altaheri et al., “Deep learning techniques for classification of electroencephalogram (EEG) motor imagery (MI) signals: a review,” Neural Comput. Appl., vol. 35, no. 20, pp. 14681–14722, 2023

  7. [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 12th International Conference on Learning Representations, ICLR 2024, 2024

  8. [11]

    Combining brain–computer interfaces and assistive technologies: state -of-the-art and challenges,

    J. del R. Millán et al., “Combining brain–computer interfaces and assistive technologies: state -of-the-art and challenges,” Front. Neurosci., vol. 4, p. 161, 2010

  9. [12]

    L. J. Greenfield, J. D. Geyer, and P. R. Carney, Reading EEGs: A practical approach. Lippincott Williams & Wilkins, 2012

  10. [13]

    Signal quality of simultaneously recorded invasive and non-invasive EEG,

    T. Ball, M. Kern, I. Mutschler, A. Aertsen, and A. Schulze - Bonhage, “Signal quality of simultaneously recorded invasive and non-invasive EEG,” Neuroimage, vol. 46, no. 3, pp. 708 –716, 2009

  11. [14]

    E. R. Kandel, J. H. Schwartz, T. M. Jessell, S. Siegelbaum, A. J. Hudspeth, and S. Mack, Principles of neural science , vol. 4. McGraw-hill New York, 2000

  12. [15]

    CHB-MIT Scalp EEG Database

    “CHB-MIT Scalp EEG Database.” [Online]. Available: https://archive.physionet.org/physiobank/charts/chbmit.png. [Accessed: 12-Apr-2020]

  13. [16]

    Masked autoencoders are scalable vision learners,

    K. He, X. Chen, S. Xie, Y. Li, P. Dollár, and R. Girshick, “Masked autoencoders are scalable vision learners,” in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2022, pp. 16000–16009

  14. [17]

    On the opportunities and risks of foundation models,

    R. Bommasani et al., “On the opportunities and risks of foundation models,” arXiv Prepr. arXiv2108.07258, 2021

  15. [18]

    Brain -conditional multimodal synthesis: A survey and taxonomy,

    W. Mai, J. Zhang, P. Fang, and Z. Zhang, “Brain -conditional multimodal synthesis: A survey and taxonomy,” IEEE Trans. Artif. Intell., 2024

  16. [19]

    Preferred reporting items for systematic reviews and meta - analyses: the PRISMA statement,

    D. Moher, A. Liberati, J. Tetzlaff, D. G. Altman, and P. Group, “Preferred reporting items for systematic reviews and meta - analyses: the PRISMA statement,” PLoS med , vol. 6, no. 7, p. e1000097, 2009

  17. [20]

    Self-taught learning: transfer learning from unlabeled data,

    R. Raina, A. Battle, H. Lee, B. Packer, and A. Y. Ng, “Self-taught learning: transfer learning from unlabeled data,” in Proceedings of the 24th international conference on Machine learning, 2007, pp. 759–766

  18. [21]

    Auto-encoding variational bayes,

    D. P. Kingma and M. Welling, “Auto-encoding variational bayes,” arXiv Prepr. arXiv1312.6114, 2013

  19. [22]

    Glove: Global vectors for word representation,

    J. Pennington, R. Socher, and C. D. Manning, “Glove: Global vectors for word representation,” in Proceedings of the 2014 conference on empirical methods in natural language processing (EMNLP), 2014, pp. 1532–1543

  20. [23]

    Unsupervised visual representation learning by context prediction,

    C. Doersch, A. Gupta, and A. A. Efros, “Unsupervised visual representation learning by context prediction,” in Proceedings of the IEEE international conference on computer vision , 2015, pp. 1422–1430

  21. [24]

    Colorful image colorization,

    R. Zhang, P. Isola, and A. A. Efros, “Colorful image colorization,” in Computer Vision–ECCV 2016: 14th European Conference, Amsterdam, The Netherlands, October 11-14, 2016, Proceedings, Part III 14, 2016, pp. 649–666

  22. [25]

    Unsupervised learning of visual representations by solving jigsaw puzzles,

    M. Noroozi and P. Favaro, “Unsupervised learning of visual representations by solving jigsaw puzzles,” in European conference on computer vision, 2016, pp. 69–84

  23. [26]

    Context encoders: Feature learning by inpainting,

    D. Pathak, P. Krahenbuhl, J. Donahue, T. Darrell, and A. A. Efros, “Context encoders: Feature learning by inpainting,” in Proceedings of the IEEE conference on computer vision and pattern recognition, 2016, pp. 2536–2544

  24. [27]

    Unsupervised representation learning by predicting image rotations,

    S. Gidaris, P. Singh, and N. Komodakis, “Unsupervised representation learning by predicting image rotations,” arXiv Prepr. arXiv1803.07728, 2018

  25. [28]

    Momentum contrast for unsupervised visual representation learning,

    K. He, H. Fan, Y. Wu, S. Xie, and R. Girshick, “Momentum contrast for unsupervised visual representation learning,” in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2020, pp. 9729–9738

  26. [29]

    A simple framework for contrastive learning of visual representations,

    T. Chen, S. Kornblith, M. Norouzi, and G. Hinton, “A simple framework for contrastive learning of visual representations,” in International conference on machine learning , 2020, pp. 1597 – 1607

  27. [30]

    Bootstrap your own latent -a new approach to self-supervised learning,

    J.-B. Grill et al., “Bootstrap your own latent -a new approach to self-supervised learning,” Adv. Neural Inf. Process. Syst., vol. 33, pp. 21271–21284, 2020

  28. [31]

    Bert: Pre - training of deep bidirectional transformers for language understanding,

    J. Devlin, M. -W. Chang, K. Lee, and K. Toutanova, “Bert: Pre - training of deep bidirectional transformers for language understanding,” in Proceedings of the 2019 conference of the North American chapter of the association for computational linguistics: human language technol...

  29. [32]

    Improving language understanding by generative pre -training,

    A. Radford, K. Narasimhan, T. Salimans, and I. Sutskever, “Improving language understanding by generative pre -training,” 2018

  30. [33]

    Language models are unsupervised multitask learners,

    A. Radford, J. Wu, R. Child, D. Luan, D. Amodei, and I. Sutskever, “Language models are unsupervised multitask learners,” OpenAI blog, vol. 1, no. 8, p. 9, 2019

  31. [34]

    Investigating critical frequency bands and channels for EEG-based emotion recognition with deep neural networks,

    W.-L. Zheng and B.-L. Lu, “Investigating critical frequency bands and channels for EEG-based emotion recognition with deep neural networks,” IEEE Trans. Auton. Ment. Dev., vol. 7, no. 3, pp. 162– 175, 2015

  32. [35]

    Deap: A database for emotion analysis; using physiological signals,

    S. Koelstra et al., “Deap: A database for emotion analysis; using physiological signals,” IEEE Trans. Affect. Comput., vol. 3, no. 1, pp. 18–31, 2011

  33. [36]

    A modality -independent proto -organization of human multisensory areas,

    F. Setti et al. , “A modality -independent proto -organization of human multisensory areas,” Nat. Hum. Behav. , vol. 7, no. 3, pp. 397–410, 2023

  34. [38]

    Brant: Foundation Model for Intracranial Neural Signal,

    D. Z. Zhang, Z. Z. Yuan, Y. Yang, J. R. Chen, J. J. Wang, and Y. F. Li, “Brant: Foundation Model for Intracranial Neural Signal,” ADVANCES IN NEURAL INFORMATION PROCESSING SYSTEMS 36 (NEURIPS 2023) , no. 37th Conference on Neural Information Processing Systems (NeurIPS). 2023

  35. [39]

    CBraMod: A Criss-Cross Brain Foundation Model for EEG Decoding,

    J. Wang et al., “CBraMod: A Criss-Cross Brain Foundation Model for EEG Decoding,” in The Thirteenth International Conference on Learning Representations, 2025

  36. [40]

    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,” Adv. Neural Inf. Process. Syst. , vol. 37, pp. 39249 – 39280, 2024

  37. [41]

    Simmim: A simple framework for masked image modeling,

    Z. Xie et al., “Simmim: A simple framework for masked image modeling,” in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2022, pp. 9653–9663

  38. [42]

    EEGFormer: Towards Transferable and Interpretable Large-Scale EEG Foundation Model,

    Y. Chen et al. , “EEGFormer: Towards Transferable and Interpretable Large-Scale EEG Foundation Model,” in AAAI 2024 Spring Symposium on Clinical Foundation Models, 2024

  39. [43]

    GEFM: Graph - Enhanced EEG Foundation Model,

    L. Wang, T. Suzumura, and H. Kanezashi, “GEFM: Graph - Enhanced EEG Foundation Model,” arXiv.org, 2024

  40. [44]

    NeuroLM: A Universal Multi-task Foundation Model for Bridging the Gap between Language and EEG Signals,

    W. Jiang, Y. Wang, B. Lu, and D. Li, “NeuroLM: A Universal Multi-task Foundation Model for Bridging the Gap between Language and EEG Signals,” in The Thirteenth International Conference on Learning Representations, 2025

  41. [45]

    Promoting Cross -Modal Representations to Improve Multimodal Foundation Models for Physiological Signals,

    C. Fang et al. , “Promoting Cross -Modal Representations to Improve Multimodal Foundation Models for Physiological Signals,” in NeurIPS, 2024

  42. [46]

    Knowledge-guided EEG Representation Learning,

    A. Kommineni, K. Avramidis, R. Leahy, and S. Narayanan, “Knowledge-guided EEG Representation Learning,” in 46th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC), 2024, pp. 1–6

  43. [47]

    Neuro-GPT: Towards A Foundation Model For EEG,

    W. Cui et al. , “Neuro-GPT: Towards A Foundation Model For EEG,” Proceedings - International Symposium on Biomedical Imaging, no. 21st IEEE International Symposium on Biomedical Imaging (ISBI). 2024

  44. [48]

    EEGPT: Unleashing the Potential of EEG Generalist Foundation Model by Autoregressive Pre -training,

    T. Yue et al. , “EEGPT: Unleashing the Potential of EEG Generalist Foundation Model by Autoregressive Pre -training,” arXiv Prepr. arXiv2410.19779, 2024

  45. [49]

    Enhancing EEG -to-Text Decoding through Transferable Representations from Pre -trained Contrastive EEG -Text Masked Autoencoder,

    J. Wang, Z. Song, Z. Ma, X. Qiu, M. Zhang, and Z. Zhang, “Enhancing EEG -to-Text Decoding through Transferable Representations from Pre -trained Contrastive EEG -Text Masked Autoencoder,” PROCEEDINGS OF THE 62ND ANNUAL MEETING OF THE ASSOCIATION FOR COMPUTATIONAL LINGUISTICS, ...

  46. [50]

    Neuro-BERT: Rethinking Masked Autoencoding for Self -Supervised Neurological Pretraining,

    D. Wu, S. Li, J. Yang, and M. Sawan, “Neuro-BERT: Rethinking Masked Autoencoding for Self -Supervised Neurological Pretraining,” IEEE J. Biomed. Heal. Informatics, 2024

  47. [51]

    Brainbert: Self -Supervised Representation Learning for Intracranial Recordings,

    C. Wang et al. , “Brainbert: Self -Supervised Representation Learning for Intracranial Recordings,” 11th Int. Conf. Learn. Represent. ICLR 2023, 2023

  48. [52]

    Representation learning with contrastive predictive coding,

    A. van den Oord, Y. Li, and O. Vinyals, “Representation learning with contrastive predictive coding,” arXiv Prepr. arXiv1807.03748, 2018

  49. [53]

    Exploring simple siamese representation learning,

    X. Chen and K. He, “Exploring simple siamese representation learning,” in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2021, pp. 15750–15758

  50. [54]

    Emerging properties in self -supervised vision transformers,

    M. Caron et al. , “Emerging properties in self -supervised vision transformers,” in Proceedings of the IEEE/CVF international conference on computer vision, 2021, pp. 9650–9660

  51. [55]

    Barlow twins: Self -supervised learning via redundancy reduction,

    J. Zbontar, L. Jing, I. Misra, Y. LeCun, and S. Deny, “Barlow twins: Self -supervised learning via redundancy reduction,” in International conference on machine learning , 2021, pp. 12310 – 12320

  52. [56]

    VICReg: Variance - Invariance-Covariance Regularization For Self -Supervised Learning,

    A. Bardes, J. Ponce, and Y. Lecun, “VICReg: Variance - Invariance-Covariance Regularization For Self -Supervised Learning,” in ICLR 2022-International Conference on Learning Representations, 2022

  53. [57]

    MBrain: A Multi-channel Self-Supervised Learning Framework for Brain Signals,

    D. H. Cai, J. R. Chen, Y. Yang, T. Liu, Y. F. Li, and ACM, “MBrain: A Multi-channel Self-Supervised Learning Framework for Brain Signals,” PROCEEDINGS OF THE 29TH ACM SIGKDD CONFERENCE ON KNOWLEDGE DISCOVERY AND DATA MINING, KDD 2023, no. 29th ACM SIGKDD Conference on Knowledg...

  54. [58]

    BENDR: Using Transformers 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 Transformers and a Contrastive Self-Supervised Learning Task to Learn From Massive Amounts of EEG Data,” Front. Hum. Neurosci., vol. 15, 2021

  55. [59]

    wav2vec 2.0: A framework for self-supervised learning of speech representations,

    A. Baevski, Y. Zhou, A. Mohamed, and M. Auli, “wav2vec 2.0: A framework for self-supervised learning of speech representations,” Adv. Neural Inf. Process. Syst., vol. 33, pp. 12449–12460, 2020

  56. [60]

    Self-supervised contrastive learning for EEG-based cross-subject motor imagery recognition,

    W. Li et al., “Self-supervised contrastive learning for EEG-based cross-subject motor imagery recognition,” J. Neural Eng., vol. 21, no. 2, 2024

  57. [61]

    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,” Adv. Neural Inf. Process. Syst., vol. 36, pp. 78240–78260, 2023

  58. [62]

    EEG-Language Modeling for Pathology Detection,

    S. Gijsen and K. Ritter, “EEG-Language Modeling for Pathology Detection,” arXiv Prepr. arXiv2409.07480, 2024

  59. [63]

    Learning transferable visual models from natural language supervision,

    A. Radford et al. , “Learning transferable visual models from natural language supervision,” in International conference on machine learning, 2021, pp. 8748–8763

  60. [64]

    Videoclip: Contrastive pre -training for zero -shot video-text understanding,

    H. Xu et al. , “Videoclip: Contrastive pre -training for zero -shot video-text understanding,” arXiv Prepr. arXiv2109.14084, 2021

  61. [65]

    Active contrastive learning of audio -visual video representations,

    S. Ma, Z. Zeng, D. McDuff, and Y. Song, “Active contrastive learning of audio -visual video representations,” arXiv Prepr. arXiv2009.09805, 2020

  62. [66]

    Cross-modal contrastive learning for speech translation,

    R. Ye, M. Wang, and L. Li, “Cross-modal contrastive learning for speech translation,” arXiv Prepr. arXiv2205.02444, 2022

  63. [67]

    Multi -modal cross -domain self -supervised pre - training for fMRI and EEG fusion,

    X. Wei et al. , “Multi -modal cross -domain self -supervised pre - training for fMRI and EEG fusion,” Neural Networks, vol. 184, 2025

  64. [68]

    Brant -X: A Unified Physiological Signal Alignment Framework,

    D. Z. Zhang, Z. Z. Yuan, J. R. Chen, K. R. Chen, Y. Yang, and A. C. MACHINERY, “Brant -X: A Unified Physiological Signal Alignment Framework,” PROCEEDINGS OF THE 30TH ACM SIGKDD CONFERENCE ON KNOWLEDGE DISCOVERY AND DATA MINING, KDD 2024, no. 30th ACM SIGKDD Conference on Know...

  65. [69]

    Self -supervised cross-modal visual retrieval from brain activities,

    Z. S. Ye, L. A. Yao, Y. Zhang, and S. Gustin, “Self -supervised cross-modal visual retrieval from brain activities,” PATTERN Recognit., vol. 145, 2024

  66. [70]

    BELT-2: Bootstrapping EEG -to-Language representation alignment for multi -task brain decoding,

    J. Zhou, Y. Duan, F. Chang, T. Do, Y. -K. Wang, and C. -T. Lin, “BELT-2: Bootstrapping EEG -to-Language representation alignment for multi -task brain decoding,” arXiv Prepr. arXiv2409.00121, 2024

  67. [71]

    Optimising EEG decoding with refined sampling and multimodal feature integration,

    A. Akbarinia, “Optimising EEG decoding with refined sampling and multimodal feature integration,” arXiv Prepr. arXiv2409.20086, 2024

  68. [72]

    Decoding Natural Images from EEG for Object Recognition,

    Y. Song, B. Liu, X. Li, N. Shi, Y. Wang, and X. Gao, “Decoding Natural Images from EEG for Object Recognition,” in The Twelfth International Conference on Learning Representations, 2024

  69. [73]

    NeuroBind: Towards Unified Multimodal Representations for Neural Signals,

    F. Yang et al. , “NeuroBind: Towards Unified Multimodal Representations for Neural Signals,” arXiv.org, 2024

  70. [74]

    A Knowledge-Driven Cross -view Contrastive Learning for EEG Representation,

    W. Weng, Y. Gu, Q. Zhang, Y. Huang, C. Miao, and Y. Chen, “A Knowledge-Driven Cross -view Contrastive Learning for EEG Representation,” arXiv.org, 2023

  71. [75]

    Uncovering the structure of clinical EEG signals with self-supervised learning,

    H. Banville, O. Chehab, A. Hyvärinen, D. -A. Engemann, and A. Gramfort, “Uncovering the structure of clinical EEG signals with self-supervised learning,” J. Neural Eng., vol. 18, no. 4, p. 46020, 2021

  72. [76]

    Toward Generalizing Visual Brain Decoding to Unseen Subjects,

    X. Kong, K. Huang, P. Li, and L. Zhang, “Toward Generalizing Visual Brain Decoding to Unseen Subjects,” in The Thirteenth International Conference on Learning Representations, 2025

  73. [77]

    Language models are few -shot learners,

    T. Brown et al., “Language models are few -shot learners,” Adv. Neural Inf. Process. Syst., vol. 33, pp. 1877–1901, 2020

  74. [78]

    Sparks of Artificial General Intelligence: Early experiments with GPT-4

    S. Bubeck et al., “Sparks of Artificial General Intelligence: Early experiments with GPT-4.” ArXiv, 2023

  75. [79]

    Brant -2: Foundation model for brain signals,

    Z. Yuan, D. Zhang, J. Chen, G. Gu, and Y. Yang, “Brant -2: Foundation model for brain signals,” arXiv e -prints, p. arXiv - 2402, 2024

  76. [80]

    BrainWave: A Brain Signal Foundation Model for Clinical Applications,

    Z. Yuan, F. Shen, M. Li, Y. Yu, C. Tan, and Y. Yang, “BrainWave: A Brain Signal Foundation Model for Clinical Applications,” arXiv Prepr. arXiv2402.10251, 2024

  77. [81]

    {MAEEG}: Masked Auto -encoder for {EEG} Representation Learning,

    H.-Y. S. Chien, H. Goh, C. M. Sandino, and J. Y. Cheng, “{MAEEG}: Masked Auto -encoder for {EEG} Representation Learning,” in NeurIPS 2022 Workshop on Learning from Time Series for Health, 2022

  78. [84]

    BrainMAE: A Region - aware Self -supervised Learning Framework for Brain Signals,

    Y. Yang, Y. Mao, X. Liu, and X. Liu, “BrainMAE: A Region - aware Self -supervised Learning Framework for Brain Signals,” arXiv Prepr. arXiv2406.17086, 2024

  79. [86]

    CEReBrO: Compact Encoder for Representations of Brain Oscillations Using Efficient Alternating Attention,

    A. Dimofte et al. , “CEReBrO: Compact Encoder for Representations of Brain Oscillations Using Efficient Alternating Attention,” arXiv Prepr. arXiv2501.10885, 2025

  80. [88]

    An Approach to Building Foundation Models for Brain Image Analysis,

    D. Karimi, “An Approach to Building Foundation Models for Brain Image Analysis,” in International Conference on Medical Image Computing and Computer-Assisted Intervention, 2024, pp. 421–431

  81. [89]

    FoME: A Foundation Model for EEG using Adaptive Temporal -Lateral Attention Scaling,

    E. Shi et al. , “FoME: A Foundation Model for EEG using Adaptive Temporal -Lateral Attention Scaling,” arXiv Prepr. arXiv2409.12454, 2024

  82. [90]

    Large Cognition Model: Towards Pretrained EEG Foundation Model,

    C.-S. Chen, Y. -J. Chen, and A. H. -W. Tsai, “Large Cognition Model: Towards Pretrained EEG Foundation Model,” arXiv Prepr. arXiv2502.17464, 2025

  83. [91]

    MEET: A Multi-Band EEG Transformer for Brain States Decoding,

    E. Shi et al., “MEET: A Multi-Band EEG Transformer for Brain States Decoding,” IEEE Trans. Biomed. Eng. , vol. 71, no. 5, pp. 1442–1453, 2024

  84. [92]

    Meta Transfer of Self -Supervised Knowledge: Foundation Model in Action for Post-Traumatic Epilepsy Prediction,

    W. Cui, H. Akrami, G. Zhao, A. A. Joshi, and R. M. Leahy, “Meta Transfer of Self -Supervised Knowledge: Foundation Model in Action for Post-Traumatic Epilepsy Prediction,” ArXiv, p. arXiv- 2312, 2023

  85. [93]

    Towards a foundation model for brain age prediction using covariance neural networks,

    S. Sihag, G. Mateos, and A. Ribeiro, “Towards a foundation model for brain age prediction using covariance neural networks,” arXiv Prepr. arXiv2402.07684, 2024

  86. [94]

    Self - supervised contrastive pre -training for time series via time - frequency consistency,

    X. Zhang, Z. Zhao, T. Tsiligkaridis, and M. Zitnik, “Self - supervised contrastive pre -training for time series via time - frequency consistency,” Adv. Neural Inf. Process. Syst. , vol. 35, pp. 3988–4003, 2022

  87. [95]

    Physics - Informed Attention Temporal Convolutional Network for EEG - Based Motor Imagery Classification,

    H. Altaheri, G. Muhammad, and M. Alsulaiman, “Physics - Informed Attention Temporal Convolutional Network for EEG - Based Motor Imagery Classification,” IEEE Trans. Ind. Informatics, vol. 19, no. 2, pp. 2249–2258, 2023

  88. [96]

    Attention based Inception model for robust EEG motor imagery classification,

    S. U. Amin, H. Altaheri, G. Muhammad, M. Alsulaiman, and W. Abdul, “Attention based Inception model for robust EEG motor imagery classification,” in 2021 IEEE International Instrumentation and Measurement Technology Conference (I2MTC), 2021, pp. 1–6

  89. [97]

    MOABB: trustworthy algorithm benchmarking for BCIs,

    V. Jayaram and A. Barachant, “MOABB: trustworthy algorithm benchmarking for BCIs,” J. Neural Eng., vol. 15, no. 6, p. 66011, 2018

  90. [98]

    REH-MI: EEG Motor Imagery Dataset from the Same Limb for Rehabilitation Applications

    H. Altaheri, G. Muhammed, S. U. Amin, and M. Alsulaiman, “REH-MI: EEG Motor Imagery Dataset from the Same Limb for Rehabilitation Applications.” IEEE Dataport, 2025

  91. [99]

    Neurobolt: Resting-state eeg-to-fmri synthesis with multi-dimensional feature mapping,

    Y. Li et al., “Neurobolt: Resting-state eeg-to-fmri synthesis with multi-dimensional feature mapping,” Adv. Neural Inf. Process. Syst., vol. 37, pp. 23378–23405, 2024

  92. [100]

    CATD: Unified Representation Learning for EEG -To-fMRI Cross-Modal Generation,

    W. Yao, Z. Lyu, M. Mahmud, N. Zhong, B. Lei, and S. Wang, “CATD: Unified Representation Learning for EEG -To-fMRI Cross-Modal Generation,” IEEE Trans. Med. Imaging, 2025

  93. [101]

    Towards neural foundation models for vision: Aligning eeg, meg and fmri representations to perform decoding, encoding and modality conversion,

    M. Ferrante, T. Boccato, and N. Toschi, “Towards neural foundation models for vision: Aligning eeg, meg and fmri representations to perform decoding, encoding and modality conversion,” in ICLR 2024 Workshop on Representational Alignment, 2024

  94. [102]

    Deep Representation Learning for Open Vocabulary Electroencephalograph y -to-Text Decoding,

    H. Amrani, D. Micucci, and P. Napoletano, “Deep Representation Learning for Open Vocabulary Electroencephalograph y -to-Text Decoding,” IEEE J. Biomed. Heal. Informatics, pp. 1–12, 2024

  95. [103]

    LLMs Help Alleviate the Cross-Subject Variability in Brain Signal and Language Alignment,

    L. Yifei, Y. Hengwei, and L. Shuhang, “LLMs Help Alleviate the Cross-Subject Variability in Brain Signal and Language Alignment,” 2025

  96. [104]

    Decoding Brain Representations by Multimodal Learning of Neural Activity and Visual Features,

    S. Palazzo, C. Spampinato, I. Kavasidis, D. Giordano, J. Schmidt, and M. Shah, “Decoding Brain Representations by Multimodal Learning of Neural Activity and Visual Features,” IEEE Trans. Pattern Anal. Mach. Intell., vol. 43, no. 11, pp. 3833–3849, 2021

  97. [105]

    Decoding visual neural representations by multimodal learning of brain -visual-linguistic features,

    C. Du, K. Fu, J. Li, and H. He, “Decoding visual neural representations by multimodal learning of brain -visual-linguistic features,” IEEE Trans. Pattern Anal. Mach. Intell., vol. 45, no. 9, pp. 10760–10777, 2023

  98. [106]

    DreamDiffusion: High -Quality EEG-to-Image Generation with Temporal Masked Signal Modeling and CLIP Alignment,

    Y. Bai, X. Wang, Y. -P. Cao, Y. Ge, C. Yuan, and Y. Shan, “DreamDiffusion: High -Quality EEG-to-Image Generation with Temporal Masked Signal Modeling and CLIP Alignment,” in European Conference on Computer Vision, 2024, pp. 472–488

  99. [107]

    Seeing through the brain: image reconstruction of visual perception from human brain signals,

    Y.-T. Lan et al., “Seeing through the brain: image reconstruction of visual perception from human brain signals,” arXiv Prepr. arXiv2308.02510, 2023

  100. [108]

    BrainECHO: Semantic Brain Signal Decoding through Vector -Quantized Spectrogram Reconstruction for Whisper -Enhanced Text Generation,

    J. Li, Z. Song, J. Wang, M. Zhang, and Z. Zhang, “BrainECHO: Semantic Brain Signal Decoding through Vector -Quantized Spectrogram Reconstruction for Whisper -Enhanced Text Generation,” arXiv Prepr. arXiv2410.14971, 2024

  101. [109]

    VIT -LENS: Towards Omni -modal Representations,

    W. X. Lei et al. , “VIT -LENS: Towards Omni -modal Representations,” 2024 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR) , no. IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). pp. 26637–26647, 2024

  102. [110]

    Deep learning human mind for automated visual classification,

    C. Spampinato, S. Palazzo, I. Kavasidis, D. Giordano, N. Souly, and M. Shah, “Deep learning human mind for automated visual classification,” in Proceedings of the IEEE conference on computer vision and pattern recognition, 2017, pp. 6809–6817

  103. [111]

    End -to-end translation of human neural activity to speech with a dual –dual generative adversarial network,

    Y. Guo, T. Liu, X. Zhang, A. Wang, and W. Wang, “End -to-end translation of human neural activity to speech with a dual –dual generative adversarial network,” Knowledge-based Syst., vol. 277, p. 110837, 2023

  104. [112]

    Neural and physiological data from participants listening to affective music,

    I. Daly et al. , “Neural and physiological data from participants listening to affective music,” Sci. data, vol. 7, no. 1, p. 177, 2020

  105. [113]

    Extracting Temporal-Spectral- Spatial Representation of EEG Using Self -Supervised Learning for the Identification of Children with ASD,

    H. Chen, O. Gaoxiang, and X. Li, “Extracting Temporal-Spectral- Spatial Representation of EEG Using Self -Supervised Learning for the Identification of Children with ASD,” in 2023 IEEE 13th International Conference on CYBER Technology in Automation, Control, and Intelligent Sy...

  106. [114]

    SparrKULee: A speech -evoked auditory response repository of the KU Leuven, containing EEG of 85 participants,

    L. Bollens, B. Accou, M. Gillis, W. Verheijen, and T. Francart, “SparrKULee: A speech -evoked auditory response repository of the KU Leuven, containing EEG of 85 participants,” 2023

  107. [115]

    TIE -EEGNet: Temporal information enhanced EEGNet for seizure subtype classification,

    R. Peng et al. , “TIE -EEGNet: Temporal information enhanced EEGNet for seizure subtype classification,” IEEE Trans. Neural Syst. Rehabil. Eng., vol. 30, pp. 2567–2576, 2022

  108. [116]

    A large and rich EEG dataset for modeling human visual object recognition,

    A. T. Gifford, K. Dwivedi, G. Roig, and R. M. Cichy, “A large and rich EEG dataset for modeling human visual object recognition,” Neuroimage, vol. 264, p. 119754, 2022

  109. [117]

    A mind -brain-body dataset of MRI, EEG, cognition, emotion, and peripheral physiology in young and old adults,

    A. Babayan et al. , “A mind -brain-body dataset of MRI, EEG, cognition, emotion, and peripheral physiology in young and old adults,” Sci. data, vol. 6, no. 1, pp. 1–21, 2019

  110. [118]

    EEG dataset and OpenBMI toolbox for three BCI paradigms: an investigation into BCI illiteracy,

    M.-H. Lee et al., “EEG dataset and OpenBMI toolbox for three BCI paradigms: an investigation into BCI illiteracy,” Gigascience, vol. 8, no. 5, p. giz002, 2019

  111. [119]

    MPED: A multi -modal physiological emotion database for discrete emotion recognition,

    T. Song, W. Zheng, C. Lu, Y. Zong, X. Zhang, and Z. Cui, “MPED: A multi -modal physiological emotion database for discrete emotion recognition,” IEEE Access, vol. 7, pp. 12177 – 12191, 2019

  112. [120]

    You snooze, you win: the physionet/computing in cardiology challenge 2018,

    M. M. Ghassemi et al. , “You snooze, you win: the physionet/computing in cardiology challenge 2018,” in 2018 Computing in Cardiology Conference (CinC) , 2018, vol. 45, pp. 1–4

  113. [121]

    Emotionmeter: A multimodal framework for recognizing human emotions,

    W.-L. Zheng, W. Liu, Y. Lu, B. -L. Lu, and A. Cichocki, “Emotionmeter: A multimodal framework for recognizing human emotions,” IEEE Trans. Cybern., vol. 49, no. 3, pp. 1110 –1122, 2018

  114. [122]

    Automated interpretation of abnormal adult electroencephalograms,

    S. López, I. Obeid, and J. Picone, “Automated interpretation of abnormal adult electroencephalograms,” MS Thesis, Temple Univ., 2017

  115. [123]

    DREAMER: A database for emotion recognition through EEG and ECG signals from wireless low-cost off -the-shelf devices,

    S. Katsigiannis and N. Ramzan, “DREAMER: A database for emotion recognition through EEG and ECG signals from wireless low-cost off -the-shelf devices,” IEEE J. Biomed. Heal. informatics, vol. 22, no. 1, pp. 98–107, 2017

  116. [124]

    Differential entropy feature for EEG-based emotion classification,

    R.-N. Duan, J.-Y. Zhu, and B.-L. Lu, “Differential entropy feature for EEG-based emotion classification,” in 2013 6th international IEEE/EMBS conference on neural engineering (NER) , 2013, pp. 81–84

  117. [125]

    Detection of seizures in intracranial EEG: UPenn and Mayo Clinic’s seizure detection challenge,

    A. Temko, A. Sarkar, and G. Lightbody, “Detection of seizures in intracranial EEG: UPenn and Mayo Clinic’s seizure detection challenge,” in 2015 37th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC) , 2015, pp. 6582–6585

  118. [126]

    Montreal Archive of Sleep Studies: an open‐access resource for instrument benchmarking and exploratory research,

    C. O’reilly, N. Gosselin, J. Carrier, and T. Nielsen, “Montreal Archive of Sleep Studies: an open‐access resource for instrument benchmarking and exploratory research,” J. Sleep Res., vol. 23, no. 6, pp. 628–635, 2014

  119. [127]

    Review of the BCI competition IV,

    M. Tangermann et al., “Review of the BCI competition IV,” Front. Neurosci., vol. 6, p. 55, 2012

  120. [128]

    A multimodal database for affect recognition and implicit tagging,

    M. Soleymani, J. Lichtenauer, T. Pun, and M. Pantic, “A multimodal database for affect recognition and implicit tagging,” IEEE Trans. Affect. Comput., vol. 3, no. 1, pp. 42–55, 2011

  121. [129]

    Application of machine learning to epileptic seizure onset detection and treatment

    A. H. Shoeb, “Application of machine learning to epileptic seizure onset detection and treatment.” Massachusetts Institute of Technology, 2009

  122. [130]

    BCI2000: a general -purpose brain -computer interface (BCI) system,

    G. Schalk, D. J. McFarland, T. Hinterberger, N. Birbaumer, and J. R. Wolpaw, “BCI2000: a general -purpose brain -computer interface (BCI) system,” IEEE Trans. Biomed. Eng. , vol. 51, no. 6, pp. 1034–1043, 2004

  123. [131]

    Indications of nonlinear deterministic and finite- dimensional structures in time series of brain electrical activity: Dependence on recording region and brain state,

    R. G. Andrzejak, K. Lehnertz, F. Mormann, C. Rieke, P. David, and C. E. Elger, “Indications of nonlinear deterministic and finite- dimensional structures in time series of brain electrical activity: Dependence on recording region and brain state,” Phys. Rev. E , vol. 64, no. 6...

  124. [132]

    Self -Supervised Transformer Model Training for a Sleep-EEG Foundation Model,

    M. Ogg and W. G. Coon, “Self -Supervised Transformer Model Training for a Sleep-EEG Foundation Model,” bioRxiv, 2024

  125. [133]

    Decoding speech perception from non -invasive brain recordings,

    A. Défossez, C. Caucheteux, J. Rapin, O. Kabeli, and J. -R. King, “Decoding speech perception from non -invasive brain recordings,” Nat. Mach. Intell. , vol. 5, no. 10, pp. 1097 –1107, 2023

  126. [134]

    Learning from models beyond fine-tuning,

    H. Zheng et al., “Learning from models beyond fine-tuning,” Nat. Mach. Intell., vol. 7, no. 1, pp. 6–17, 2025

  127. [135]

    Brain -to-Text Decoding: A Non -invasive Approach via Typing,

    J. Lévy et al. , “Brain -to-Text Decoding: A Non -invasive Approach via Typing,” Meta-Research, 2025

  128. [136]

    Applications of OPM -MEG for translational neuroscience: a perspective,

    M. Brickwedde et al. , “Applications of OPM -MEG for translational neuroscience: a perspective,” Transl. Psychiatry, vol. 14, no. 1, p. 341, 2024

  129. [137]

    Visual decoding and reconstruction via eeg embeddings with guided diffusion,

    D. Li, C. Wei, S. Li, J. Zou, H. Qin, and Q. Liu, “Visual decoding and reconstruction via eeg embeddings with guided diffusion,” arXiv Prepr. arXiv2403.07721, 2024

  130. [138]

    Are eeg-to-text models working?,

    H. Jo, Y. Yang, J. Han, Y. Duan, H. Xiong, and W. H. Lee, “Are eeg-to-text models working?,” arXiv Prepr. arXiv2405.06459 , 2024

  131. [139]

    Open vocabulary electroencephalography-to- text decoding and zero -shot sentiment classification,

    Z. Wang and H. Ji, “Open vocabulary electroencephalography-to- text decoding and zero -shot sentiment classification,” in Proceedings of the AAAI Conference on Artificial Intelligence , 2022, vol. 36, no. 5, pp. 5350–5358

  132. [140]

    BELT: Bootstrapped EEG -to-Language Training by Natural Language Supervision,

    J. Z. Zhou, Y. Q. Duan, Y. C. Chang, Y. K. Wang, and C. T. Lin, “BELT: Bootstrapped EEG -to-Language Training by Natural Language Supervision,” IEEE Trans. NEURAL Syst. Rehabil. Eng., vol. 32, pp. 3278–3288, 2024

  133. [141]

    Aligning Semantic in Brain and Language: A Curriculum Contrastive Method for Electroencephalography-to-Text Generation,

    X. C. Feng, X. C. Feng, B. Qin, and T. Liu, “Aligning Semantic in Brain and Language: A Curriculum Contrastive Method for Electroencephalography-to-Text Generation,” IEEE Trans. NEURAL Syst. Rehabil. Eng., vol. 31, pp. 3874–3883, 2023

  134. [142]

    UniCoRN: Unified cognitive signal reconstruction bridging cognitive signals and human language,

    N. Xi, S. Zhao, H. Wang, C. Liu, B. Qin, and T. Liu, “UniCoRN: Unified cognitive signal reconstruction bridging cognitive signals and human language,” in Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), 2023, pp. ...

  135. [143]

    Semantic -aware contrastive learning for electroencephalography -to-text generation with curriculum learning,

    X. Feng, X. Feng, and B. Qin, “Semantic -aware contrastive learning for electroencephalography -to-text generation with curriculum learning,” arXiv Prepr. arXiv2301.09237, 2023

  136. [144]

    DeWave: Discrete EEG Waves Encoding for Brain Dynamics to Text Translation,

    Y. Q. Duan, J. Z. Zhou, Z. Wang, Y. K. Wang, and C. T. Lin, “DeWave: Discrete EEG Waves Encoding for Brain Dynamics to Text Translation,” ADVANCES IN NEURAL INFORMATION PROCESSING SYSTEMS 36 (NEURIPS 2023) , no. 37th Conference on Neural Information Processing Systems (NeurIPS). 2023

Pith tools

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