Department of Biobehavioral Health, The Pennsylvania State University, University Park, Pennsylvania, USA; Proactive Life, Inc, New York, New York, USA.
Department of Biobehavioral Health, The Pennsylvania State University, University Park, Pennsylvania, USA.
Sleep Health. 2023 Oct;9(5):596-610. doi: 10.1016/j.sleh.2023.07.001. Epub 2023 Aug 10.
Commonly used actigraphy algorithms are designed to operate within a known in-bed interval. However, in free-living scenarios this interval is often unknown. We trained and evaluated a sleep/wake classifier that operates on actigraphy over ∼24-hour intervals, without knowledge of in-bed timing.
Actigraphy counts from ActiWatch Spectrum devices.
Sleep staging derived from polysomnography, supplemented by observation of wakefulness outside of the staged interval. Classifications from the Oakley actigraphy algorithm were additionally used as performance reference.
Adults, sleeping in either a home or laboratory environment.
Machine learning was used to train and evaluate a sleep/wake classifier in a supervised learning paradigm. The classifier is a temporal convolutional network, a form of deep neural network.
Performance was evaluated across ∼24 hours, and additionally restricted to only in-bed intervals, both in terms of epoch-by-epoch performance, and the discrepancy of summary statistics within the intervals.
Performance of the trained model applied to the Multi-Ethnic Study of Atherosclerosis dataset.
Over ∼24 hours, the temporal convolutional network classifier produced the same or better performance as the Oakley classifier on all measures tested. When restricting analysis to the in-bed interval, the temporal convolutional network remained favorable on several metrics.
Performance decreased on the Multi-Ethnic Study of Atherosclerosis dataset, especially when restricting analysis to the in-bed interval.
A classifier using data labeled over ∼24-hour intervals allows for the continuous classification of sleep/wake without knowledge of in-bed intervals. Further development should focus on improving generalization performance.
常用的活动记录仪算法旨在已知的卧床时间间隔内运行。然而，在自由生活场景中，这个间隔通常是未知的。我们训练和评估了一种睡眠/唤醒分类器，该分类器可以在不了解卧床时间的情况下，在大约 24 小时的活动记录仪数据上运行。
活动记录仪计数来自 ActiWatch Spectrum 设备。
多导睡眠图衍生的睡眠分期，辅以分期间隔外清醒状态的观察。奥克利活动记录仪算法的分类结果也被用作性能参考。
成人，在家庭或实验室环境中睡眠。
机器学习用于在监督学习范例中训练和评估睡眠/唤醒分类器。分类器是一种时间卷积网络，是一种深度学习网络形式。
在大约 24 小时内评估性能，并且还限制在仅卧床间隔内，从逐epoch 性能以及间隔内汇总统计数据的差异两个方面进行评估。
训练后的模型在多民族动脉粥样硬化研究数据集上的性能。
在大约 24 小时内，时间卷积网络分类器在所有测试的指标上都产生了与奥克利分类器相同或更好的性能。当将分析限制在卧床间隔内时，时间卷积网络在几个指标上仍然有利。
在多民族动脉粥样硬化研究数据集上的性能下降，尤其是当将分析限制在卧床间隔内时。
使用标记超过 24 小时的数据的分类器可以在不了解卧床间隔的情况下连续分类睡眠/唤醒。进一步的发展应侧重于提高泛化性能。
