发表机构
University of Manchester; ETH Zürich; Shanghai Jiao Tong University; Anhui University; ELLIS Institute Finland; Aalto University; Institute of Automation, Chinese Academy of Sciences; University of Chinese Academy of Sciences（曼彻斯特大学; 苏黎世联邦理工学院; 上海交通大学; 安徽大学; ELLIS芬兰研究所; 阿尔托大学; 中国科学院自动化研究所; 中国科学院大学）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
提出NeurDuo-EEG，一种具有持久状态和显式记忆的EEG基础模型，通过多时间尺度记忆管理实现连续EEG建模，在五个基准中四个达到最优，并支持高效流式推理。
AI 中文摘要
脑电图（EEG）连续记录数小时，相关动态跨越从毫秒到小时的时间尺度。然而，大多数EEG基础模型独立处理固定窗口，限制了它们捕捉长时间尺度动态中编码信息的能力。状态空间架构支持持久递归处理，但长程信息仍隐式压缩在递归状态中。我们提出NeurDuo-EEG，一种具有通道分辨持久记忆的因果EEG基础模型。NeurDuo-EEG引入多时间尺度记忆管理，具有学习巩固和选择性检索，以固定大小状态实现对连续EEG的持久建模。它在17个公共数据集的3,955小时EEG上预训练，使用多通道自回归预测离散频谱码。在三个短窗口和两个长序列下游任务中，NeurDuo-EEG在五个基准中的四个上达到最佳性能，包括所有三个短窗口任务和癫痫检测，其中AUC-PR从$0.285$提高到$0.471$，优于最强的非NeurDuo基线。NeurDuo-EEG在睡眠分期上也保持竞争力，并支持高效流式推理，随着可用历史增长到一小时，每块延迟几乎恒定。值得注意的是，Small变体仅用4.7M骨干参数就实现了这一点。这些结果证明了持久、多时间尺度建模对长序列和短窗口EEG分析的价值。我们的代码可在https URL获取。
英文摘要
Electroencephalography (EEG) is recorded continuously over hours, with relevant dynamics spanning timescales from milliseconds to hours. Most EEG foundation models nevertheless process fixed windows independently, limiting their ability to capture information encoded in long-timescale dynamics. State-space architectures enable persistent recurrent processing, but long-range information remains implicitly compressed in recurrent states. We present NeurDuo-EEG, a causal EEG foundation model with channel-resolved persistent memory. NeurDuo-EEG introduces multi-timescale memory management with learned consolidation and selective retrieval, enabling persistent modelling of continuous EEG with fixed-size state. It is pre-trained on 3,955 hours of EEG from 17 public datasets using multichannel autoregressive prediction of discrete spectral codes. Across three short-window and two long-sequence downstream tasks, NeurDuo-EEG achieves the best performance on four of five benchmarks, including all three short-window tasks and seizure detection, where AUC-PR improves from $0.285$ to $0.471$ over the strongest non-NeurDuo baseline. NeurDuo-EEG also remains competitive on sleep staging and supports efficient streaming inference, with nearly constant per-chunk latency as the available history grows to one hour. Notably, the Small variant achieves this with only 4.7M backbone parameters. These results demonstrate the value of persistent, multi-timescale modelling for both long-sequence and short-window EEG analysis. Our code is available at https://github.com/YifaNNW/NeurDuo-EEG.
