Validation of 'Somnivore', a Machine Learning Algorithm for Automated Scoring and Analysis of Polysomnography Data
Validation of 'Somnivore', a Machine Learning Algorithm for Automated Scoring and Analysis of Polysomnography Data
复制标题
DOI：
10.3389/fnins.2019.00207
复制
发表时间：
2019-03-18
影响因子：
4.3
通讯作者：
Gundlachla, Andrew L.
中科院分区：
文献类型：
--
作者：
Allocca, Giancarlo;Ma, Sherie;Gundlachla, Andrew L.
多导睡眠图数据的手动评分是劳动密集型和耗时的，并且大多数现有软件不考虑主观差异和用户可变性。因此，我们评估了一种有监督的机器学习算法Somnivore（TM），用于自动化唤醒-睡眠阶段分类。我们设计了一种算法，该算法在简短的手动评分之后从各种输入通道中提取特征，并为每个记录提供自动的唤醒-睡眠阶段分类。对于算法验证，从独立实验室获得多导睡眠图数据，包括正常、认知受损和酒精治疗的人类受试者（总n = 52）、发作性睡眠小鼠和药物治疗的大鼠（总n = 56）和鸽子（n = 5）。用于验证的训练和测试集先前由来自每个实验室的1-2名受过训练的睡眠技术人员手动评分。F-测量用于评估分类器输出和人类评分员一致性的统计分析的精度和灵敏度。该算法在所有人类数据中与手动视觉评分具有高度一致性（清醒0.91 +/- 0.01; N1 0.57 +/- 0.01; N2 0.81 +/- 0.01; N3 0.86 +/- 0.01; REM 0.87 +/- 0.01），这与所有阶段的手动评分员间一致性相当。同样，在所有啮齿动物（清醒0.95 +/- 0.01; NREM 0.94 +/- 0.01; REM 0.91 +/- 0.01）和鸽子（清醒0.96 +/- 0.006; NREM 0.97 +/- 0.01; REM 0.86 +/- 0.02）数据中观察到高度一致性。还检查了从单个信号输入、简单阶段重新分类、自动去除过渡时期和训练集大小的分类器学习的效果。总之，我们已经开发了一个多导睡眠图分析程序，用于对不同物种的数据进行自动睡眠阶段分类。Somnivore能够对实验和临床睡眠研究进行灵活、准确和高通量的分析。
Manual scoring of polysomnography data is labor-intensive and time-consuming, and most existing software does not account for subjective differences and user variability. Therefore, we evaluated a supervised machine learning algorithm, Somnivore (TM), for automated wake-sleep stage classification. We designed an algorithm that extracts features from various input channels, following a brief session of manual scoring, and provides automated wake-sleep stage classification for each recording. For algorithm validation, polysomnography data was obtained from independent laboratories, and include normal, cognitively-impaired, and alcohol-treated human subjects (total n = 52), narcoleptic mice and drug-treated rats (total n = 56), and pigeons (n = 5). Training and testing sets for validation were previously scored manually by 1-2 trained sleep technologists from each laboratory. F-measure was used to assess precision and sensitivity for statistical analysis of classifier output and human scorer agreement. The algorithm gave high concordance with manual visual scoring across all human data (wake 0.91 +/- 0.01; N1 0.57 +/- 0.01; N2 0.81 +/- 0.01; N3 0.86 +/- 0.01; REM 0.87 +/- 0.01), which was comparable to manual inter-scorer agreement on all stages. Similarly, high concordance was observed across all rodent (wake 0.95 +/- 0.01; NREM 0.94 +/- 0.01; REM 0.91 +/- 0.01) and pigeon (wake 0.96 +/- 0.006; NREM 0.97 +/- 0.01; REM 0.86 +/- 0.02) data. Effects of classifier learning from single signal inputs, simple stage reclassification, automated removal of transition epochs, and training set size were also examined. In summary, we have developed a polysomnography analysis program for automated sleep-stage classification of data from diverse species. Somnivore enables flexible, accurate, and high-throughput analysis of experimental and clinical sleep studies.