Department of Computing, University of Turku, Finland.
Department of Biomedical Engineering, University of California Irvine, Irvine, CA, United States.
Comput Biol Med. 2024 Sep;179:108679. doi: 10.1016/j.compbiomed.2024.108679. Epub 2024 Jul 20.
Sleep staging is a crucial tool for diagnosing and monitoring sleep disorders, but the standard clinical approach using polysomnography (PSG) in a sleep lab is time-consuming, expensive, uncomfortable, and limited to a single night. Advancements in sensor technology have enabled home sleep monitoring, but existing devices still lack sufficient accuracy to inform clinical decisions. To address this challenge, we propose a deep learning architecture that combines a convolutional neural network and bidirectional long short-term memory to accurately classify sleep stages. By supplementing photoplethysmography (PPG) signals with respiratory sensor inputs, we demonstrated significant improvements in prediction accuracy and Cohen's kappa (k) for 2- (92.7 %; k = 0.768), 3- (80.2 %; k = 0.714), 4- (76.8 %, k = 0.550), and 5-stage (76.7 %, k = 0.616) sleep classification using raw data. This relatively translatable approach, with a less intensive AI model and leveraging only a few, inexpensive sensors, shows promise in accurately staging sleep. This has potential for diagnosing and managing sleep disorders in a more accessible and practical manner, possibly even at home.
睡眠分期是诊断和监测睡眠障碍的重要工具，但在睡眠实验室中使用多导睡眠图 (PSG) 的标准临床方法既耗时、昂贵、不舒服，又仅限于一晚上。传感器技术的进步使得家庭睡眠监测成为可能，但现有设备的准确性仍然不足以为临床决策提供信息。为了解决这一挑战，我们提出了一种深度学习架构，该架构结合卷积神经网络和双向长短期记忆模型，可准确分类睡眠阶段。通过在光体积描记 (PPG) 信号中补充呼吸传感器输入，我们证明了在使用原始数据进行 2-（92.7%；k=0.768）、3-（80.2%；k=0.714）、4-（76.8%；k=0.550）和 5-（76.7%；k=0.616）睡眠分类时，预测准确性和 Cohen's kappa (k) 显著提高。这种相对可移植的方法使用了一个不那么密集的人工智能模型，并且只利用了少数几个便宜的传感器，有望更准确地分期睡眠。这有可能以更便捷、实用的方式诊断和管理睡眠障碍，甚至可能在家中进行。
