计算机分析脑电波信号的算法研究及实现方法

本文描述了一种EEG自动分析中排除伪差的基于知识的系统，对脑电图伪差及癫痫病棘波的识别方法进行了探讨：运用时城分析法并全面地利用16个通道上可得到的前后信息识别伪差；用波形分析法来识别癫痫病棘波。初步实验结果表明，本方案预处理方法简便，波形特征易于提取，识别算法灵活，功能齐全，能识别各种通常在EEG中出现的伪差，在适当的抽样速率下并能准确检测棘波。

The Algorithm Research and Methods of a Computerized Automated Eeg Analysis System

In the paper a know/edge-based system is described to eliminate artifacts in the automated analysis of EEG. The recognition method of EEG artifacts and Epileptogenic spikes are discussed with emphasis. The system makes use of time domain analysis to reject the artifacts, recognizes Epileptogenic spikes by waveform analysis and makes a preliminary diagnosis. The preliminary result shows the preprocessing is simple, the waveform feature is easy to be extracted, artifacts that often found in EEG can be recognized, and Epileptogenic spikes can be correctly detected at proper speed.