机器学习建立脑电联合指标用于监测麻醉状态下的意识水平

麻醉深度监测一直是临床麻醉的难点，麻醉深度与认知功能等术后并发症的发生具有一定的联系，自发脑电图(EEG)和听觉诱发电位(AEP)被用来监测麻醉期间患者的意识水平，但还存在一定的缺陷，通过一定的机器学习算法优化组合这些指标可能有助于取长补短，发现更好的指标。

现分享发表在PLoS One. 2020;15(8)的文章Machine learning for a combined electroencephalographic anesthesia index to detect awareness under anesthesia.译文供参考。

摘要译文

机器学习建立脑电联合指标用于监测麻醉状态下的意识水平

自发脑电图(EEG)和听觉诱发电位(AEP)已被用来监测麻醉期间的意识水平。由于两种信号反映不同的神经通路，两种信号的参数组合可以提供麻醉期间大脑状态的更广泛信息。

适当的参数选择和组合到一个单一的指数对于利用这种潜力是至关重要的。机器学习领域提供了参数选择和组合的算法。

在这项研究中，几个已经建立的机器学习方法，包括选择合适信号参数的方法和分类算法被用来构建一个指数来预测麻醉患者的反应性。

本文分析了支持向量机、人工神经网络和贝叶斯学习算法中的几种分类算法。

根据意识和无意识转换的数据，采用自动化方法建立的EEG和AEP信号参数组合模型，最大预测概率为0.935，高于0.916(EEG参数)和0.880(AEP参数)的交叉验证。

这表明机器学习技术可以成功地应用于改进的联合EEG和AEP参数来区分意识和无意识。

Machinelearning for a combined electroencephalographic anesthesia index to detectawareness under anesthesia

Spontaneouselectroencephalogram (EEG) and auditory evoked potentials (AEP) have beensuggested to monitor the level of consciousness during anesthesia. As bothsignals reflect different neuronal pathways, a combination of parameters fromboth signals may provide broader information about the brain status duringanesthesia. Appropriate parameter selection and combination to a single indexis crucial to take advantage of this potential. The field of machine learningoffers algorithms for both parameter selection and combination. In this study,several established machine learning approaches including a method for theselection of suitable signal parameters and classification algorithms areapplied to construct an index which predicts responsiveness in anesthetizedpatients. The present analysis considers several classification algorithms,among those support vector machines, artificial neural networks and Bayesianlearning algorithms. On the basis of data from the transition betweenconsciousness and unconsciousness, a combination of EEG and AEP signalparameters developed with automated methods provides a maximum predictionprobability of 0.935, which is higher than 0.916 (for EEG parameters) and 0.880(for AEP parameters) using a cross-validation approach. This suggests thatmachine learning techniques can successfully be applied to develop an improvedcombined EEG and AEP parameter to separate consciousness from unconsciousness.

本文由“小麻哥的日常”授权转载  

打赏