撰文丨王聪 编辑丨王多鱼 排版丨水成文 心房颤动 （ Atrial fibrillation，简称房颤） 是一种常见且严重的心率失常，在全球范围内与较高的发病率和死亡率相关。 心电图 （ECG） 被认为是诊断房颤的 金标准。然而，目前的心电图检查主要仅在出现症状时或偶尔体检时使用，因为其测量方式需要接触皮肤。这种局限性导致难以捕捉到早期的心颤发作，从而错 失了及时干预的机会。 2025 年 5 月 20 日，中国科学技术大学 陈彦 团队在 Nature Communications 期刊发表了题为： Atrial fibrillation detection via contactless radio monitoring and knowledge transfer 的研究论文。 该研究开发了一个基于 无线电技术 和 人工智能 的 非接触式房颤检测系统 ，该系统有助于在传统临床诊断路径之前检测出房颤。 该系统的优异性能源于两项关键创新： 首先 ，研究团队设计了一种 毫米波雷达 ，采用混合信号处理算法来精确捕捉毫米级的心脏机械运动。其次，通过利用 知 识迁移技术 和 心脏兴奋-收缩偶联机制 ，借助基于现有大规 ...

Core Insights - A new non-contact atrial fibrillation (AF) detection system has been developed by a team of Chinese scientists, utilizing radar sensing and artificial intelligence (AI) technology to monitor arrhythmia symptoms wirelessly, potentially allowing for earlier detection and intervention compared to traditional clinical methods [1][3]. Group 1: Technology and Methodology - The system captures sub-millimeter heart movements remotely using radio signals and employs a knowledge transfer-driven AI framework to identify AF patterns from electrocardiogram (ECG) diagnostics [3]. - The innovative approach establishes a mapping between cardiac electrical activity and mechanical motion, leveraging validated ECG signal features to assist neural networks in recognizing abnormal mechanical fluctuations specific to AF [3]. Group 2: Clinical Evaluation - The research team evaluated the non-contact AF detection system using data from 6,258 outpatient patients, including 229 AF patients, and found that the system's sensitivity and specificity in detecting AF were comparable to that of ECG [3]. - Further testing on 27 patients during regular sleep demonstrated the system's potential in detecting the presence and episodes of AF [3][4]. Group 3: Future Implications - The non-contact AF detection system shows promise for deployment in everyday life, which could facilitate large-scale early screening and proactive management of AF patients [4].