中文摘要：

利用自适应自回归（adaptive autoregressive,AAR）模型和卡尔曼滤波器算法,分析小鼠海马CA1区场电位ripple高频振荡的时频特性.研究发现,与传统的基于短时傅立叶变换的实时频谱分析方法相比,利用AAR模型以及卡尔曼滤波器算法的参数化方法在对ripple高频振荡信号进行实时频谱分析时,具有更高的时域和频域分辨率.因此,基于卡尔曼滤波器得到的ripple能量变化,可更为准确、实时地反映ripple高频振荡的发生与变化过程.

英文摘要：

This paper studied high frequency ripple（100~200 Hz） oscillations in hippocanpal CA1 area by applications of adaptive autoregressive（AAR） model and Kalman filter.Compared with traditional real time frequency analysis of time seies based on short term Fourier transfrom（STFT）,improved time and frequency resolutions in time-frequency representation could be achieved by parametric methold obtained by AAR model and Kalman filter algorithms.Thus,the occurance of ripple oscillations and the variation of ripple power could be addresed more accurate by Kalman filter than that of STFT.