Harnessing Spectrum Video for Subject-Level Few-Shot and Cross-Montage EEG Generalization

Wei Wang, Fang He, Yifan Li, Wanying Qu, Yawei Li, Quanying Liu, Yanwei Fu
Proceedings of the 43rd International Conference on Machine Learning, PMLR 306:126598-126613, 2026.

Abstract

Existing EEG models are limited by electrode heterogeneity and rigid "channel-first" architectures that treat sensors as independent features. We propose Brain Signal Rendering (BSR), which reinterprets EEG as a physical projection of neural activity and transforms raw signals into structured spatiotemporal tensors (termed Spectrum Videos), enabling the transfer of rich priors from video foundation models. By utilizing VideoMAE for self-supervised pre-training, BSR learns robust, layout-agnostic spatiotemporal representations that preserve neural topology. We further employ subject-level few-shot learning and introduce cross-montage fine-tuning to rigorously evaluate generalization across subjects and electrode configurations. Experiments show that VideoMAE model integrated with the BSR framework significantly outperforms state-of-the-art spectrum based methods, providing a scalable and data-efficient foundation for generalizable EEG modeling. Our code is available at https://github.com/yanweifu-sii/BSR-VideoMAE.

Cite this Paper

BibTeX
@InProceedings{pmlr-v306-wang26bt, title = {Harnessing Spectrum Video for Subject-Level Few-Shot and Cross-Montage {EEG} Generalization}, author = {Wang, Wei and He, Fang and Li, Yifan and Qu, Wanying and Li, Yawei and Liu, Quanying and Fu, Yanwei}, booktitle = {Proceedings of the 43rd International Conference on Machine Learning}, pages = {126598--126613}, year = {2026}, editor = {Zhang, Tong and Dudik, Miroslav and Jaggi, Martin and Agarwal, Alekh and Li, Sharon and Schuurmans, Dale and Zhu, Jerry and Berkenkamp, Felix and Dong, Hanze and Bietti, Alberto}, volume = {306}, series = {Proceedings of Machine Learning Research}, month = {06--11 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v306/main/assets/wang26bt/wang26bt.pdf}, url = {https://proceedings.mlr.press/v306/wang26bt.html}, abstract = {Existing EEG models are limited by electrode heterogeneity and rigid "channel-first" architectures that treat sensors as independent features. We propose Brain Signal Rendering (BSR), which reinterprets EEG as a physical projection of neural activity and transforms raw signals into structured spatiotemporal tensors (termed Spectrum Videos), enabling the transfer of rich priors from video foundation models. By utilizing VideoMAE for self-supervised pre-training, BSR learns robust, layout-agnostic spatiotemporal representations that preserve neural topology. We further employ subject-level few-shot learning and introduce cross-montage fine-tuning to rigorously evaluate generalization across subjects and electrode configurations. Experiments show that VideoMAE model integrated with the BSR framework significantly outperforms state-of-the-art spectrum based methods, providing a scalable and data-efficient foundation for generalizable EEG modeling. Our code is available at https://github.com/yanweifu-sii/BSR-VideoMAE.} }
Endnote
%0 Conference Paper %T Harnessing Spectrum Video for Subject-Level Few-Shot and Cross-Montage EEG Generalization %A Wei Wang %A Fang He %A Yifan Li %A Wanying Qu %A Yawei Li %A Quanying Liu %A Yanwei Fu %B Proceedings of the 43rd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2026 %E Tong Zhang %E Miroslav Dudik %E Martin Jaggi %E Alekh Agarwal %E Sharon Li %E Dale Schuurmans %E Jerry Zhu %E Felix Berkenkamp %E Hanze Dong %E Alberto Bietti %F pmlr-v306-wang26bt %I PMLR %P 126598--126613 %U https://proceedings.mlr.press/v306/wang26bt.html %V 306 %X Existing EEG models are limited by electrode heterogeneity and rigid "channel-first" architectures that treat sensors as independent features. We propose Brain Signal Rendering (BSR), which reinterprets EEG as a physical projection of neural activity and transforms raw signals into structured spatiotemporal tensors (termed Spectrum Videos), enabling the transfer of rich priors from video foundation models. By utilizing VideoMAE for self-supervised pre-training, BSR learns robust, layout-agnostic spatiotemporal representations that preserve neural topology. We further employ subject-level few-shot learning and introduce cross-montage fine-tuning to rigorously evaluate generalization across subjects and electrode configurations. Experiments show that VideoMAE model integrated with the BSR framework significantly outperforms state-of-the-art spectrum based methods, providing a scalable and data-efficient foundation for generalizable EEG modeling. Our code is available at https://github.com/yanweifu-sii/BSR-VideoMAE.
APA
Wang, W., He, F., Li, Y., Qu, W., Li, Y., Liu, Q. & Fu, Y.. (2026). Harnessing Spectrum Video for Subject-Level Few-Shot and Cross-Montage EEG Generalization. Proceedings of the 43rd International Conference on Machine Learning, in Proceedings of Machine Learning Research 306:126598-126613 Available from https://proceedings.mlr.press/v306/wang26bt.html.

Related Material