麻醉过程中脑电信号处理方法研究

摘要

Anesthesia plays an essential role in clinical operations. Guiding anesthesia by EEG signals is one of the most promising methods at present and it has obtained good results. The analysis and process of the EEG signals in anesthesia can provide clean signal for further research. This paper used variance threshold method to remove the mutation fast and large interfering signals; and used notch filter to remove frequency interference, smoothing filter to remove baseline drift and Butterworth low-pass filter to remove high frequency noise at the same time. In addition to this, the translation invariant wavelet method to remove interference noise on the signals which was after the classical filter and retained non-stationary characteristics was used to evaluate parameter calculation. By comparing the calculated parameters from treated signal using this paper’s methods and untreated signal and standard signal, the standard deviation and correlation has been improved, particularly the major parameters BetaR, which provides better signal for integration of multi-parameter to evaluate depth of anesthesia index for the latter.%麻醉是临床手术中必不可少的关键环节，通过脑电信号来指导麻醉手术是目前最有潜力的方法之一，其已获得较好的效果。该文通过对麻醉手术中脑电信号的分析处理，为进一步的研究指导麻醉手术指标提供干净的信号。采用方差阈值法去除采集过程中突变较大的干扰信号；陷波器去除工频干扰、平滑滤波去除基线漂移和巴特沃斯低通滤波器去除高频干扰信号；小波平移不变量法对经过经典滤波器并且保留非平稳特征的信号去除干扰噪声来评估麻醉深度的参数计算。比较由该文方法处理的信号、未经处理的信号和标准信号计算得到的参数，标准差和相关度都有了提高，特别是主要参数BetaR，为后期多参数融合评价麻醉深度指标提供了较好的信号。