arXiv cs.AI· Urban \v{S}irca, Maryam Alimardani, Stefanos Zafeiriou, Konstantinos Barmpas·· 9 天前AI 评分22
超越准确率：EEG 基础模型的鲁棒性、可解释性与表达力评估
Beyond Accuracy: Robustness, Interpretability and Expressiveness of EEG Foundation Models
AI 导读
一项研究对六个 EEG 基础模型和一名监督基线在十个数据集上从鲁棒性、可解释性与表达力三个层面展开评估，发现没有单一模型能主导所有失效模式。最抗噪声的模型在通道 dropout 下反而最脆弱，而此前归因于预训练表征质量差的 head-only 性能不佳，很大程度上可由池化策略解释。研究建议未来评估应分别报告各类扰动的鲁棒性、生成归因图并检验多种池化策略。
正文 · 原文
Abstract:EEG foundation models (EEG-FMs) have been evaluated predominantly on clean, in-distribution accuracy, demonstrating modest gains over supervised baselines and weak frozen representations. This study examines whether these conclusions hold beyond clean accuracy by evaluating six EEG-FMs and a supervised baseline across ten datasets along three layers of analysis: (i) Robustness: we apply test-time perturbations including additive noise, random and region-based channel dropout and region-specific noise injection. Our analyses show that no single model dominates all failure modes. The most noise-robust model is among the most fragile under channel dropout and much of the dropout fragility disappears when channels are removed rather than zero-padded. (ii) Interpretability: using attribution methods in EEG-FMs, we show that models broadly concentrate relevance on task-appropriate brain regions consistent with known neurophysiology. (iii) Expressiveness: we demonstrate that the poor head-only performance previously attributed to low-quality pre-trained representations is largely explained by the pooling strategy and that EEG-FMs possess sufficient representational capacity when their token-level embeddings are preserved. Furthermore, with block-wise probing and attention analysis we show that late blocks are repurposed during fine-tuning, while early blocks already hold task-related information. Our results show that conclusions about EEG-FMs depend on evaluation choices and we recommend that future evaluation of EEG-FMs should report robustness per perturbation type, produce attribution maps and examine multiple pooling strategies.
| Subjects: | Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Human-Computer Interaction (cs.HC) |
| Cite as: | arXiv:2605.17562 [cs.LG] |
| (or arXiv:2605.17562v2 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2605.17562 arXiv-issued DOI via DataCite |
Submission history
From: Urban Širca [view email]
[v1]
Sun, 17 May 2026 17:47:00 UTC (37,592 KB)
[v2]
Thu, 8 Oct 2026 16:45:23 UTC (36,436 KB)
来源：arXiv cs.AI · arxiv.org