中文摘要：

目的为了研究不同的实验任务是否会影响EEG脑力负荷预测的准确率,方法本文采用了一种名为原功率调制（SPo C）新型的特征提取方式,结合传统的信号分类方法对EEG信号进行了分类。使用实验模拟脑力劳动的四个阶段,采用SPo C这种盲源信号分离方法提取脑电信号的特征,使用线性判别方法对在此基础上对代表不同信息加工阶段的四种任务的不同难度等级的脑电信号进行了分类。通过对分类准确性的统计,研究了脑电信号与任务难度等级的关系,更深入的研究了不同任务（即四种信息加工阶段）与脑电信号的关系,讨论了实验任务对EEG脑力负荷预测准确率的影响。结果分析了不同实验对象分别在四种实验任务重的脑电分布图,并对SPo C脑图模式进行了分析。计算了各实验的平均分类准确率。两难度实验任务分类准确率最高达0.98以上,不同任务之间的分类准确率平均值之间的差值不超过0.09。结论实验任务对分类准确率没有显著性影响,实验者对分类准确率有显著性影响。

英文摘要：

Objective In order to find whether the four stages of brain information processing will impact the classification results of the analysis of the Electroencephalograph （EEG） for mental workload; Methods a novel feature extraction approach combined with a classical classification method has been used. Four different experiments have been chosen to simulate four stages of brain information processing. A source power correlation （SPoC） Ap- proach was used for signal processing. After that, a series of statistical analysis has been done to find the relationship between the tasks and the classification accuracy rate. Results The brain maps of each subject in different tasks has been analyzed, combined with SPoC patterns. The mean accuracy of each task was calculated. The highest accuracy rate between the two different classes exceeded up to 0.98. The highest and lowest mean accuracy differed by only 0.09. Conclusion It has been found that the tasks has not significant effect on accuracy rate, but the subjects in the experiments are significantly relate to the accuracy rate.