Bondar' A T, Fedotchev A I
Institute of Cell Biophysics, Russian Academy of Sciences, Pushchino.
Zh Vyssh Nerv Deiat Im I P Pavlova. 2000 Nov-Dec;50(6):933-42.
The possible mechanisms which determine the temporal dynamics of discrete narrow-band spectral components of human EEG recorded by a single electrode in the state of rest were analyzed. The dynamics of short-segment spectra was observed by application of Fast Fourier Transform (FFT) to 5-s EEG epochs successively shifted by 0.32 s. For each subject the matrices were formed and presented in a graphic mode. Matrix rows represented the number of points in each short-segment spectrum, and the columns represented the number of short-segment spectra. The columns reflect the amplitude dynamics of a given frequency, and power transition between the columns reflects the frequency dynamics. The most common type of the amplitude dynamics consisted in short (2-8 s) periods of stable activity of the discrete spectral components replaced by symmetrical bifurcation or confluence of spectral peaks. The obtained results suggest by the presence of both additive and multiplicative mechanisms of oscillatory interactions in the EEG. More detailed analysis of the amplitude-modulated EEG processes is provided by application of some additive features of the FFT to both EEG and computer-simulated signals.
分析了在静息状态下由单个电极记录的人类脑电图离散窄带频谱成分的时间动态的可能机制。通过对连续移动0.32秒的5秒脑电图片段应用快速傅里叶变换（FFT）来观察短片段频谱的动态。为每个受试者形成矩阵并以图形模式呈现。矩阵行表示每个短片段频谱中的点数，列表示短片段频谱的数量。列反映给定频率的幅度动态，列之间的功率转换反映频率动态。幅度动态最常见的类型是离散频谱成分的短（2 - 8秒）稳定活动期，随后被频谱峰值的对称分叉或合并所取代。所得结果表明脑电图中存在振荡相互作用的加法和乘法机制。通过将FFT的一些加法特征应用于脑电图和计算机模拟信号，对幅度调制的脑电图过程进行了更详细的分析。
