Eeg Signal Processing Based On Wavelet Transform

The brain electrophysiological signals, including electroencephalogram (EEG) and evoked potentials(EP), contain a lot of important information about the physiological states and functional activities of brain. How to process and analyze EEG effectively is significant for neural science and neurosis diagnosis. The wavelet transform, which produces a good local representation of the signal in both time and frequency domain, provides an important tool in signal analysis and feature extraction. The main research work of this paper focuses on the application of wavelet transform technique in EEG analysis, which could be summarized as the following three aspects.(a)EEG signal denoising method: After the discussion of the application of wavelet thresholding denoising method in EEG signal disposing, an improved wavelet threshold function is presented in this thesis. Subsequently, we use a fast and fixed point algorithm for the processing of EEG., and then put forward a reformative method based on median filter, which is called adaptive wavelet thresholding method. The numerical experiments manifests that this method could give better denoising effect.(b)Extraction of Evoked Potential: The evoked potential (EP) is an important tool for neuroscience studies and clinical diagnosis. However, extracting EPs with traditional averaging method always results in a serious loss of transient information, adding pain to the testee , as well as the reduction of reliability. So it is urgent to develop a new method which could extract the EPs fast. In this thesis, we study the application of wavelet method and Third-Order Corrections(TOC) based filtering technology in this field, combining TOC with wavelet thresholding method creatively, and present a new technique for the fast extraction of EPs, and then point out that the new thresholding function-TOC based filter method is an effective method.(c)Sleep EEG processing: The electroencephalogram(EEG) plays an importantrole in sleep research. Multichannel electroencephalogram (EEG) data collection and analysis has become an significant tool in this field. Wavelet transform is not applied widely in sleep EEG research, and the correlative production is very limited. This thesis summarize the application of wavelet transform in sleep EEG processing in recent years, such as the detection of REM , Spindles and k-complex, and then discusses the feasibility and effectiveness of this application.