中文摘要：

目的基于大鼠前额叶皮层多通道局部场电位和锋电位的频谱相干编码，研究不同模态的神经信号如何相互协同编码工作记忆行为事件。方法应用Cerebus多通道信号采集系统，记录大鼠Y迷宫工作记忆行为事件中在体前额叶皮层上同时记录的2类不同模态的多通道神经信号：连续的多通道局部场电位（LFPs）和离散的多通道锋电位（Spikes）。利用短时傅里叶（STTF）计算多通道LFPs的功率谱密度（PSD）。发展和应用Multitaper频谱相干编码方法，选取窗口500ms，移动步长125ms计算行为事件参考点前后5s共10s时间长度的多通道LFPs-Spikes的频谱相干值。结果2只大鼠各10次实验结果在第2-4秒、0．3～15Hz的频谱相干值（0．6524±0．0197）与其他时间和频率频谱相干值（0．3214±0．0164）之间的差异具有统计学意义（P〈0．05）.结论大鼠在Y迷宫工作记忆行为事件中的多通道LFPs-Spikes协同编码并预测了该工作记忆行为事件。

英文摘要：

Objective To investigate how neural signals coherently encode working memory based on spectrum coherence of two types of multi-channel neural signals from rat prefrontal cortex. Methods Cerebus 128 Data Acquisition System was introduced to record two types of multi-channel neural singals： continuous time series （local field potentials, LFPs ） and discrete time series （spikes）, which were acquired on prefrontal cortex of rats during working memory task of Y maze. Short-Time Fouries Transform （STFT） was applied to calculate the power spectrum density （PSD） of LFPs. Experiment data of multi-channel LFPs-Spikes were obtained from time interval of 5s before and after the event. Spectrum coherence of LFP and spike from each channel was calculated in the Muhitaper window with 500ms length sliding from the beginning and 125ms overlapping. Results Spectrum coherence of 0.3-15 Hz multi-channel LFPs-Spikes at 2-4 s increased with statistical significance （P〈 0.05 ）. Conclusion Our results suggest that spectrum coherence of multi-channel LFPs-Spikes encodes and predicts the event effectively.