Data_Sheet_1_Permutation entropy is not an age-independent parameter for EEG-based anesthesia monitoring.docx

BackgroundAn optimized anesthesia monitoring using electroencephalographic (EEG) information in the elderly could help to reduce the incidence of postoperative complications. Processed EEG information that is available to the anesthesiologist is affected by the age-induced changes of the raw EEG. While most of these methods indicate a “more awake” patient with age, the permutation entropy (PeEn) has been proposed as an age-independent measure. In this article, we show that PeEn is also influenced by age, independent of parameter settings.
MethodsWe retrospectively analyzed the EEG of more than 300 patients, recorded during steady state anesthesia without stimulation, and calculated the PeEn for different embedding dimensions m that was applied to the EEG filtered to a wide variety of frequency ranges. We constructed linear models to evaluate the relationship between age and PeEn. To compare our results to published studies, we also performed a stepwise dichotomization and used non-parametric tests and effect sizes for pairwise comparisons.
ResultsWe found a significant influence of age on PeEn for all settings except for narrow band EEG activity. The analysis of the dichotomized data also revealed significant differences between old and young patients for the PeEn settings used in published studies.
ConclusionBased on our findings, we could show the influence of age on PeEn. This result was independent of parameter, sample rate, and filter settings. Hence, age should be taken into consideration when using PeEn to monitor patient EEG.

背景：针对老年患者采用脑电图（electroencephalogram, EEG）信息优化麻醉监测，有助于降低术后并发症的发生率。麻醉医师可获取的处理后脑电图信息，会受到原始脑电图因年龄增长产生的变化的影响。尽管多数此类监测方法显示患者随年龄增长呈现“更清醒”的状态，但排列熵（permutation entropy, PeEn）曾被提出作为一种与年龄无关的监测指标。本文研究证实，排列熵（PeEn）同样会受年龄影响，且该效应与参数设置无关。
方法：本研究回顾性分析了300余例患者的脑电图数据，这些数据采集于无刺激的稳态麻醉状态下；我们针对经多种频段滤波处理的脑电图，计算了不同嵌入维度m下的排列熵（PeEn）值。我们构建线性模型以评估年龄与排列熵之间的关联。为将本研究结果与已发表文献进行对比，我们还采用了逐步二分类法，并使用非参数检验与效应量开展组间两两比较。
结果：除窄频段脑电图活动外，所有参数设置下均观察到年龄对排列熵存在显著影响。对二分类数据的分析同样显示，已发表研究中采用的排列熵参数设置下，老年与青年患者间存在显著差异。
结论：基于本研究结果，我们证实了年龄对排列熵的影响效应，且该效应不受参数、采样率与滤波设置的干扰。因此，在采用排列熵监测患者脑电图时，应将年龄因素纳入考量范围。