中文摘要：

引入了符号动力学方法分析认知事件相关电位（ERP）的复杂度．以混合模型生成的随机时间序列为例，对近似熵和符号熵作了比较．应用符号熵分析了Oddball范式中不同任务条件（靶刺激和非靶刺激）下的ERP的复杂度．研究发现，额区、中央区和顶区的ERP复杂度在刺激呈现后的任务加工时间段内显著减小（非靶刺激和靶刺激分别在刺激呈现后200-300和400-500ms），而且靶刺激ERP复杂度大约在P300成分的峰值时刻达到最小值，在响应之后逐渐回升．这表明基于符号动力学的复杂度分析能够反映认知任务加工的时间过程，并且能够显著区分两种任务类型．

英文摘要：

The symbolic dynamics was introduced to analyze the complexity of cognitive event-related potentials （ ERPs）. This method was first applied to the case of the MIX （ p ） system, and then was compared with the approximate entropy which had been successfully used to analyze the biological and medical experimental data. The complexity of ERPs in the Oddball paradigm was calculated with the symbolic dynamics. It was found that the complexity of ERPs over the frontal, central, and parietal areas declined significantly during the processing of tasks, reached at the minimum corresponding to the latency of P300 component, and rose back after the response was performed. In addition, this method was sensitive to the discrimination between target stimuli and non-target stimuli. The results indicate that the complexity based on the symbolic dynamics can sensitively reflect the processing of the cognitive task, and will be a promising tool to analyze the neural activity associated with the cognitive task.