Development Studies of Measuring and Processing Systgem of EEG Activity during Orking and Sleeping

脑电(EEG)分析的目的是从大脑状态和脑功能中提取重要信息，在频域上，功率谱方法通常是通过将总频段划分为几个带宽来实现的。在时间域中，Fujimoris分析作为过零方法的扩展被用来表征脑电波。在本研究中，发展了一种考虑波浪方向的波浪识别方法。波的方向意味着从波底和波峰计算脑电波的幅度。该方法提高了波的识别率。接下来，在与传统的1 kHz采样方法相反的100 Hz方法下对EEG数据进行采样。在100赫兹采样后，应用干涉函数，极大地提高了脑电数据的约简能力，对睡眠脑电波的刻画效果较好。这些研究以《脑电分析中的波形识别研究》、《脑电中纺锤波频率分析的测量方法》为题发表。为了进一步研究睡眠纺锤波的频率分析，我们应用希尔伯特变换法测量了睡眠纺锤波的瞬时频率。希尔伯特方法在有噪声干扰的数据中被认为是弱的。但在我们的主轴分析中，波在频率带宽上是有限的。这在纺锤波的频率分析中产生了良好的结果。E将这里开发的这些方法应用于EEG数据。结果表明，这些方法在脑电信号分析和非平稳数据分析中是有用的。

The purpose of the EEG( electroencephalography ) analysis is to exitract the importand information from the brain state and the brain functgion.In the frequency domain, power spectrum methods are often used by dividing the total frequency range into several bandwidths. In the time domain, Fujimoris analysis has been developed to characterize the EEG waves as an extension of the zero- crossing method. In this study, the wave recognition method has been developed with consideration of the wave direction. The wave direction implies computation of the amplitude of the EEG wave from both the bottom and the peak of waves. The method here improves the recognition rate of the waves. Next, the EEG datga have been sampled under the 100 Hz method contrary to the conventional 1 KHz sampling method. After the 100 Hz sampling, an interporation function was applied, The method developed here greatly improves the EEG data reduction and shows good results of the characterization of the sleeping EEG waves. These studies are published as paper titles " Studies of waveform recognition in the EEG analysis","Measurement method for frequency analysis of spindle waves in EEG". For the further study of the frequency analysis of the sleep spindle waves, we applied the Hilbert transform method measuring the instantaneous frequency of the spindle waves. The Hilbert method is said to be weak in the data which are disturbed by noise. But in our spindle anlysis, the wave is limitted in the frequency bandwidths. This produces good results in the frequency analysis of the spindle waves. E applied these methods developed here to the EEG data. it was clarified that these methods are usefull in the EEG analysis and monstationary data.