医学 AI - arXivDaily 专题

2602.07628 2026-06-19 cs.AI cs.LG 版本更新 90%

SleepMaMi: A Universal Sleep Foundation Model for Integrating Macro- and Micro-structures

SleepMaMi：一种融合宏观与微观结构的通用睡眠基础模型

Keondo Park, Younghoon Na, Yourim Choi, Hyunwoo Ryu, Hyun-Woo Shin, Hyung-Sin Kim

发表机构 * Graduate School of Data Science, Seoul National University, Seoul, South Korea（首尔国立大学数据科学研究生院，韩国首尔）； Department of Biomedical Sciences, Seoul National University College of Medicine, Seoul, Republic of Korea（首尔国立大学医学院生物医学科学系，韩国首尔）； Obstructive Upper Airway Research (OUaR) Laboratory, Department of Pharmacology, Seoul National University College of Medicine, Seoul, Republic of Korea（首尔国立大学医学院药理学系阻塞性上气道研究（OUaR）实验室，韩国首尔）； Department of Otorhinolaryngology-Head and Neck Surgery, Seoul National University Hospital, Seoul, Republic of Korea（首尔国立大学医院耳鼻喉头颈外科系，韩国首尔）

专题命中健康监测：睡眠基础模型SleepMaMi融合宏观微观结构

AI总结提出SleepMaMi睡眠基础模型，通过分层双编码器设计（宏观编码器建模整夜时间依赖，微观编码器捕捉生物信号短时特征），结合人口统计引导对比学习和混合掩码自编码器训练，在超过2万条PSG记录上预训练，在下游任务中优于或匹配现有基础模型。

Comments 8 pages, Appendix 9 pages

详情

AI中文摘要

虽然向统一基础模型的转变已经彻底改变了许多深度学习领域，但睡眠医学仍然主要局限于专注于局部微观结构特征的特定任务模型。这些方法常常忽略多导睡眠图（PSG）丰富的多模态背景，并且未能捕捉整夜睡眠的全局宏观结构。为了解决这个问题，我们引入了SleepMaMi，一种睡眠基础模型，旨在掌握长达一小时的睡眠架构和细粒度信号形态。我们的框架采用分层双编码器设计：宏观编码器用于建模整夜时间依赖，微观编码器用于从生物信号中捕捉短期特征。宏观编码器通过人口统计引导对比学习进行训练，该学习将夜间睡眠模式与客观受试者元数据（如年龄、性别和BMI）对齐，以优化全局表示。微观编码器通过混合掩码自编码器（MAE）和多模态对比目标进行优化。在超过20,000条PSG记录（158K小时）的大规模语料库上预训练，SleepMaMi在多样化的下游任务套件中优于或匹配现有的最先进基础模型，展示了在临床睡眠分析中卓越的泛化能力和标签高效适应能力。

英文摘要

While the shift toward unified foundation models has revolutionized many deep learning domains, sleep medicine remains largely restricted to task-specific models that focus on localized micro-structure features. These approaches often neglect the rich, multi-modal context of Polysomnography (PSG) and fail to capture the global macro-structure of a full night's sleep. To address this, we introduce SleepMaMi , a Sleep Foundation Model engineered to master both hour-long sleep architectures and fine-grained signal morphologies. Our framework utilizes a hierarchical dual-encoder design: a Macro-Encoder to model full-night temporal dependencies and a Micro-Encoder to capture short-term characteristics from biosignals. Macro-Encoder is trained via Demographic-Guided Contrastive Learning, which aligns overnight sleep patterns with objective subject metadata, such as age, sex and BMI to refine global representations. Micro-Encoder is optimized via a hybrid Masked Autoencoder (MAE) and multi-modal contrastive objective. Pre-trained on a massive corpus of $>$20,000 PSG recordings (158K hours),SleepMaMi outperforms or matches state-of-the-art existing foundation models across a diverse suite of downstream tasks, demonstrating superior generalizability and label-efficient adaptation for clinical sleep analysis.

URL PDF HTML ☆

赞 0 踩 0