J Shanghai Jiaotong Univ Sci

Journal of Shanghai Jiaotong University(Science) >

2026 , Vol. 31 >Issue 3: 783 - 799

DOI: https://doi.org/10.1007/s12204-026-2922-0

Celebrating 30 Years

EEG Foundation Models for Brain-Computer Interfaces: Progress and Future Directions

Expand
  • School of Computer Science, Shanghai Jiao Tong University, Shanghai 200240, China

Received date: 2026-01-12

  Revised date: 2026-03-10

  Accepted date: 2026-03-24

  Online published: 2026-05-08

Fold

Abstract

Electroencephalography (EEG) foundation models are increasingly used as general-purpose backbones for brain-computer interfaces (BCIs) by leveraging large-scale pretraining and task-specific adaptation. This review summarizes recent progress in EEG foundation models from three perspectives: datasets and task coverage, with emphasis on how generalization goals are operationalized by split protocols and concrete evaluation procedures; model design choices, including input construction and tokenization, masked pretraining objectives, and Transformer backbones for spatiotemporal modeling across heterogeneous channel layouts; and downstream adaptation, comparing linear probing, full fine-tuning, and parameter-efficient tuning, while clarifying the conditions under which each setting is most informative. We emphasize that reported gains are often protocol-dependent, as differences in task scope, preprocessing, training budget, and baseline selection can substantially affect comparability and the extent to which conclusions generalize. Finally, we outline future directions for EEG foundation models in BCI, focusing on standardized evaluation infrastructure, EEG-tailored modeling choices, and deployment-aware adaptation under real-world constraints.

Cite this article

Dai Renjie, Dong Shenhua, Lyu Baoliang, Zheng Weilong . EEG Foundation Models for Brain-Computer Interfaces: Progress and Future Directions[J]. Journal of Shanghai Jiaotong University(Science), 2026 , 31(3) : 783 -799 . DOI: 10.1007/s12204-026-2922-0

References

[1] Willsey M S, Shah N P, Avansino D T, et al. A high-performance brain–computer interface for finger decoding and quadcopter game control in an individual with paralysis [J]. Nature Medicine, 2025, 31(1): 96-104.
[2] Willett F R, Avansino D T, Hochberg L R, et al. High-performance brain-to-text communication via handwriting [J]. Nature, 2021, 593(7858): 249-254.
[3] Wu D R, Lu B L, Hu B, et al. Affective brain–computer interfaces (aBCIs): A tutorial [J]. Proceedings of the IEEE, 2023, 111(10): 1314-1332.
[4] He Z X, Cai M H, Li L T, et al. EEG-EMG FAConformer: Frequency aware conv-transformer for the fusion of EEG and EMG [C]//2024 IEEE International Conference on Bioinformatics and Biomedicine. Lisbon: IEEE, 2025: 3258-3261.
[5] Abiri R, Borhani S, Sellers E W, et al. A comprehensive review of EEG-based brain–computer interface paradigms [J]. Journal of Neural Engineering, 2019, 16(1): 011001.
[6] Li M G, Xu D Q. Transfer learning in motor imagery brain computer interface: A review [J]. Journal of Shanghai Jiao Tong University (Science), 2024, 29(1): 37-59.
[7] Huang J Y, Yang P F, Wan B, et al. KDLPCCA-based projection for feature extraction in SSVEP-based brain-computer interfaces [J]. Journal of Shanghai Jiao Tong University (Science), 2022, 27(2): 168-175.
[8] Thuwajit P, Rangpong P, Sawangjai P, et al. EEGWaveNet: Multiscale CNN-based spatiotemporal feature extraction for EEG seizure detection [J]. IEEE Transactions on Industrial Informatics, 2022, 18(8): 5547-5557.
[9] Kong G W, Li C, Peng H, et al. EEG-based sleep stage classification via neural architecture search [J]. IEEE Transactions on Neural Systems and Rehabilitation Engineering, 2023, 31: 1075-1085.
[10] He Z X, Li H Y, Yuan G, et al. Multimodal cardiovascular risk profiling using self-supervised learning of polysomnography [J]. Sleep, 2025: zsaf371.
[11] Dai R J, Hu K, Yin H, et al. Self-supervised EEG representation learning based on temporal prediction and spatial reconstruction for emotion recognition [C]//47th Annual Meeting of the Cognitive Science Society. San Francisco: Cognitive Science Society, 2025: 2217-2224.
[12] Sadiya S, Alhanai T, Ghassemi M M. Artifact detection and correction in EEG data: A review [C]//2021 10th International IEEE/EMBS Conference on Neural Engineering. Online: IEEE, 2021: 495-498.
[13] Li D S, Ren K, Wang Y S, et al. Learning topology-agnostic EEG representations with geometry-aware modeling [C]// 37th Conference on Neural Information Processing Systems. New Orleans: NIPS, 2023: 53875-53891.
[14] Lacourse K, Yetton B, Mednick S, et al. Massive online data annotation, crowdsourcing to generate high quality sleep spindle annotations from EEG data [J]. Scientific Data, 2020, 7: 190.
[15] Duan R N, Zhu J Y, Lu B L. Differential entropy feature for EEG-based emotion classification [C]//2013 6th International IEEE/EMBS Conference on Neural Engineering. San Diego: IEEE, 2014: 81-84.
[16] Albalawi H, Song X M. A study of kernel CSP-based motor imagery brain computer interface classification [C]//2012 IEEE Signal Processing in Medicine and Biology Symposium. New York: IEEE, 2013: 1-4.
[17] Craik A, He Y T, Contreras-Vidal J L. Deep learning for electroencephalogram (EEG) classification tasks: A review [J]. Journal of Neural Engineering, 2019, 16(3): 031001.
[18] Li M G, Wei L N. Motor imagery classification based on plain convolutional neural network and linear interpolation [J]. Journal of Shanghai Jiao Tong University (Science), 2024, 29(6): 958-966.
[19] Ma J, Ren Z, Zhang T T, et al. Transformer-based contrastive learning method for automated sleep stages classification [J]. Journal of Shanghai Jiao Tong University (Science), 2025, 30(4): 720-732.
[20] Devlin J, Chang M W, Lee K, et al. BERT: Pre-training of deep bidirectional transformers for language understanding [C]// 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies. Minneapolis: ACL, 2019: 4171-4186.
[21] He K M, Fan H Q, Wu Y X, et al. Momentum contrast for unsupervised visual representation learning [C]//2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Seattle: IEEE, 2020: 9726-9735.
[22] Zhao W X, Zhou K, Li J Y, et al. A survey of large language models [PP/OL]. arXiv (2023-05-31) [2025-12-21]. https://doi.org/10.48550/arXiv.2303.18223.
[23] Li R, Wang Y T, Zheng W L, et al. A multi-view spectral-spatial-temporal masked autoencoder for decoding emotions with self-supervised learning [C]// 30th ACM International Conference on Multimedia. Lisboa: ACM, 2022: 6-14.
[24] Li Y, Chen J, Li F, et al. GMSS: Graph-based multi-task self-supervised learning for EEG emotion recognition [J]. IEEE Transactions on Affective Computing, 2023, 14(3): 2512-2525.
[25] Zhang Z, Liu Y, Zhong S H. GANSER: A self-supervised data augmentation framework for EEG-based emotion recognition [J]. IEEE Transactions on Affective Computing, 2023, 14(3): 2048-2063.
[26] Hu K Y, Dai R J, Chen W T, et al. Contrastive self-supervised EEG representation learning for emotion classification [C]//2024 46th Annual International Conference of the IEEE Engineering in Medicine and Biology Society. Orlando: IEEE, 2024: 1-4.
[27] Zhang D, Yuan Z, Yang Y, et al. Brant: Foundation model for intracranial neural signal [C]// 37th Conference on Neural Information Processing Systems. New Orleans: NIPS, 2023: 26304-26321.
[28] Jiang W B, Zhao L M, Lu B L. Large brain model for learning generic representations with tremendous EEG data in BCI [PP/OL]. arXiv (2024-05-29) [2025-12-21]. https://doi.org/10.48550/arXiv.2405.18765.
[29] Cui W H, Jeong W, Thölke P, et al. Neuro-GPT: Towards a foundation model for EEG [C]//2024 IEEE International Symposium on Biomedical Imaging. Athens: IEEE, 2024: 1-5.
[30] Shi E Z, Zhao K, Yuan Q L, et al. FoME: A foundation model for EEG using adaptive temporal-lateral attention scaling [PP/OL]. arXiv (2024-09-19) [2025-12-21]. https://doi.org/10.48550/arXiv.2409.12454.
[31] Wang G, Liu W, He Y, et al. EEGPT: Pretrained transformer for universal and reliable representation of EEG signals [C]// 38th Conference on Neural Information Processing System. Vancouver:  NIPS, 2024: 39249-39280.
[32] Yuan Z Z, Shen F Q, Li M, et al. BrainWave: A brain signal foundation model for clinical applications [PP/OL]. V7. arXiv (2025-09-01) [2025-12-21]. https://doi.org/10.48550/arXiv.2402.10251.
[33] Jiang W B, Wang Y, Lu B L, et al. NeuroLM: A universal multi-task foundation model for bridging the gap between language and EEG signals [PP/OL]. V3. arXiv (2025-05-20) [2025-12-21]. https://doi.org/10.48550/arXiv.2409.00101.
[34] Wang J, Zhao S, Luo Z, et al. CBraMod: A criss-cross brain foundation model for EEG decoding [PP/OL]. V6. arXiv (2025-11-06) [2025-12-21]. https://doi.org/10.48550/arXiv.2412.07236.
[35] Li Z Y, Zheng W L, Lu B L. Gram: A large-scale general EEG model for raw data classification and restoration tasks [C]// 2025 IEEE International Conference on Acoustics, Speech and Signal Processing. Hyderabad: IEEE, 2025: 1-5.
[36] Ouahidi Y E, Lys J, Thölke P, et al. REVE: A foundation model for EEG: Adapting to any setup with large-scale pretraining on 25,000 subjects [PP/OL]. arXiv (2025-10-24) [2025-12-21]. https://doi.org/10.48550/arXiv.2510.21585.
[37] Zhou Y, Wu J, Ren Z, et al. CSBrain: A cross-scale spatiotemporal brain foundation model for EEG decoding [PP/OL]. arXiv (2025-06-29) [2025-12-21]. https://doi.org/10.48550/arXiv.2506.23075.
[38] Li A, Wang Z K, Yang L Y, et al. CoMET: A contrastive-masked brain foundation model for universal EEG representation [PP/OL]. arXiv (2025-08-30) [2025-12-21]. https://doi.org/10.48550/arXiv.2509.00314.
[39] Chen Z G, Qin C X, You W L, et al. HEAR: An EEG foundation model with heterogeneous electrode adaptive representation [PP/OL]. arXiv (2025-10-14) [2025-12-21]. https://doi.org/10.48550/arXiv.2510.12515.
[40] Chen Z S, Zhang Y W, Lan Q Z, et al. Uni-NTFM: A unified foundation model for EEG signal representation learning [PP/OL]. arXiv (2025-09-29) [2025-12-21]. https://doi.org/10.48550/arXiv.2509.24222.
[41] Jia T W, Meng L B, Li S Y, et al. Federated motor imagery classification for privacy-preserving brain-computer interfaces [J]. IEEE Transactions on Neural Systems and Rehabilitation Engineering, 2024, 32: 3442-3451.
[42] Pfurtscheller G, Neuper C. Motor imagery and direct brain-computer communication [J]. Proceedings of the IEEE, 2001, 89(7): 1123-1134.
[43] Morgan S T, Hansen J C, Hillyard S A. Selective attention to stimulus location modulates the steady-state visual evoked potential [J]. Proceedings of the National Academy of Sciences of the United States of America, 1996, 93(10): 4770-4774.
[44] Acharya U R, Vinitha Sree S, Swapna G, et al. Automated EEG analysis of epilepsy: A review [J]. Knowledge-Based Systems, 2013, 45: 147-165.
[45] Boostani R, Karimzadeh F, Nami M. A comparative review on sleep stage classification methods in patients and healthy individuals [J]. Computer Methods and Programs in Biomedicine, 2017, 140: 77-91.
[46] Alvi A M, Siuly S, Wang H. Neurological abnormality detection from electroencephalography data: A review [J]. Artificial Intelligence Review, 2022, 55(3): 2275-2312.
[47] Gruzelier J H. EEG-neurofeedback for optimising performance. I: A review of cognitive and affective outcome in healthy participants [J]. Neuroscience & Biobehavioral Reviews, 2014, 44: 124-141.
[48] Tangermann M, Müller K R, Aertsen A, et al. Review of the BCI competition IV [J]. Frontiers in Neuroscience, 2012, 6: 55.
[49] Schalk G, McFarland D J, Hinterberger T, et al. BCI2000: A general-purpose brain-computer interface (BCI) system [J]. IEEE Transactions on Biomedical Engineering, 2004, 51(6): 1034-1043.
[50] Lee M H, Kwon O Y, Kim Y J, et al. EEG dataset and OpenBMI toolbox for three BCI paradigms: An investigation into BCI illiteracy [J]. GigaScience, 2019, 8(5): giz002.
[51] Schirrmeister R T, Springenberg J T, Fiederer L D J, et al. Deep learning with convolutional neural networks for EEG decoding and visualization [J]. Human Brain Mapping, 2017, 38(11): 5391-5420.
[52] Wang Y J, Chen X G, Gao X R, et al. A benchmark dataset for SSVEP-based brain–computer interfaces [J]. IEEE Transactions on Neural Systems and Rehabilitation Engineering, 2017, 25(10): 1746-1752.
[53] Liu B C, Huang X S, Wang Y J, et al. BETA: A large benchmark database toward SSVEP-BCI application [J]. Frontiers in Neuroscience, 2020, 14: 627.
[54] Obeid I, Picone J. The temple university hospital EEG data corpus [J]. Frontiers in Neuroscience, 2016, 10: 196.
[55] Shoeb A H. Application of machine learning to epileptic seizure onset detection and treatment [D]. Massachusetts: Massachusetts Institute of Technology, 2009.
[56] Kemp B, Zwinderman A H, Tuk B, et al. Analysis of a sleep-dependent neuronal feedback loop: The slow-wave microcontinuity of the EEG [J]. IEEE Transactions on Biomedical Engineering, 2000, 47(9): 1185-1194.
[57] Quan S F, Howard B V, Iber C, et al. The sleep heart health study: Design, rationale, and methods [J]. Sleep, 1997, 20(12): 1077-1085.
[58] Khalighi S, Sousa T, Santos J M, et al. ISRUC-Sleep: A comprehensive public dataset for sleep researchers [J]. Computer Methods and Programs in Biomedicine, 2016, 124: 180-192.
[59] Alvarez-Estevez D, Rijsman R M. Inter-database validation of a deep learning approach for automatic sleep scoring [J]. PLoS One, 2021, 16(8): e0256111.
[60] Zheng W L, Lu B L. Investigating critical frequency bands and channels for EEG-based emotion recognition with deep neural networks [J]. IEEE Transactions on Autonomous Mental Development, 2015, 7(3): 162-175.
[61] Koelstra S, Muhl C, Soleymani M, et al. DEAP: A database for emotion Analysis;Using physiological signals [J]. IEEE Transactions on Affective Computing, 2012, 3(1): 18-31.
[62] Katsigiannis S, Ramzan N. DREAMER: A database for emotion recognition through EEG and ECG signals from wireless low-cost off-the-shelf devices [J]. IEEE Journal of Biomedical and Health Informatics, 2018, 22(1): 98-107.
[63] Soleymani M, Lichtenauer J, Pun T, et al. A multimodal database for affect recognition and implicit tagging [J]. IEEE Transactions on Affective Computing, 2012, 3(1): 42-55.
[64] He K M, Chen X L, Xie S N, et al. Masked autoencoders are scalable vision learners [C]//2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition. New Orleans: IEEE, 2022: 15979-15988.
[65] Oord A, Vinyals O, Kavukcuoglu K. Neural discrete representation learning [C]// 31st Conference on Neural Information Processing Systems. Long Beach: NIPS, 2017: 6309-6318.
[66] Foumani N M, Ghane S, Nguyen N, et al. EEG-X: Device-agnostic and noise-robust foundation model for EEG [PP/OL]. arXiv (2025-11-12) [2025-12-21]. https://doi.org/10.48550/arXiv.2511.08861.
[67] Vaswani A, Shazeer N, Parmar N, et al. Attention is all you need [C]// 31st Conference on Neural Information Processing Systems. Long Beach: NIPS, 2017: 1-11.
[68] Wang S L, Deng Y S, Bao Z H, et al. NeuroTTT: Bridging pretraining-downstream task misalignment in EEG foundation models via test-time training [PP/OL]. V2. arXiv (2025-10-01) [2025-12-21]. https://doi.org/10.48550/arXiv.2509.26301.
[69] Hu E J, Shen Y, Wallis P, et al. LoRA: Low-rank adaptation of large language models [PP/OL]. V2. arXiv (2021-10-16) [2025-12-21]. https://doi.org/10.48550/arXiv.2106.09685.
[70] Jeon J, Jeong S, Shon Y, et al. Parameter-efficient transfer learning for EEG foundation models via task-relevant feature focusing [C]// 34th ACM International Conference on Information and Knowledge Management. Seoul: ACM, 2025: 1068-1077.
[71] Suzumura T, Kanezashi H, Akahori S. Graph adapter of EEG foundation models for parameter efficient fine tuning [PP/OL]. V2. arXiv (2025-02-18) [2025-12-21]. https://doi.org/10.48550/arXiv.2411.16155.
[72] Del Pup F, Zanola A, Tshimanga L F, et al. The role of data partitioning on the performance of EEG-based deep learning models in supervised cross-subject analysis: A preliminary study [J]. Computers in Biology and Medicine, 2025, 196: 110608.
[73] Arvaneh M, Guan C T, Ang K K, et al. EEG data space adaptation to reduce intersession nonstationarity in brain-computer interface [J]. Neural Computation, 2013, 25(8): 2146-2171.
[74] Fazli S, Popescu F, Danóczy M, et al. Subject-independent mental state classification in single trials [J]. Neural Networks, 2009, 22(9): 1305-1312.
[75] Lawhern V J, Solon A J, Waytowich N R, et al. EEGNet: A compact convolutional neural network for EEG-based brain–computer interfaces [J]. Journal of Neural Engineering, 2018, 15(5): 056013.
[76] Jing J, Ge W D, Hong S D, et al. Development of expert-level classification of seizures and rhythmic and periodic patterns during EEG interpretation [J]. Neurology, 2023, 100(17): e1750-e1762.
[77] Wu J M, Ren Z C, Wang J Y, et al. AdaBrain-bench: Benchmarking brain foundation models for brain-computer interface applications [PP/OL]. V2. arXiv (2025-08-05) [2025-12-21]. https://doi.org/10.48550/arXiv.2507.09882.
[78] Xiong W, Li J T, Li J, et al. EEG-FM-bench: A comprehensive benchmark for the systematic evaluation of EEG foundation models [PP/OL]. arXiv (2025-08-25) [2025-12-21]. https://doi.org/10.48550/arXiv.2508.17742.

Options
Outlines

/

〈 〉