| EEG(Electroencephalogram) instrumentation is an important assistant tool for the detection of brain illness in clinical, and it is an important symbol for judging the performance of a EEG instrument that whether the feature information in EEG signals can be extracted effectively. Based on the knowing and analysis for EEG signal processing and the developed actuality of electroencephalograph inside and outside of China at present, this dissertation makes use of the magnitude achievement in the area of signal processing, and computer software and hardware techniques and microelectronics technique, furthermore, integrates the virtual instrument technology developed at a very fast speed during the recent years, to construct the virtual EEG measuring and analysis instrumentation system. Which is the inevitable result that virtual instrument technique is infiltrated into biomedical field, and will bring out great impact for biomedical electronics and biomedical instruments fields of our country. EEG signal is a kind of complex non-stationary signal, the feature information in EEG signals can't be extracted sufficiently only relying on traditional signal processing methods in time domain and frequency domain. On the basis of the summarization of the excellences and limitation of all kinds of time-frequency analysis theory, this dissertation discusses thorough the application in extracting of feature information in EEG signals of Gabor transform, wavelet and wavelet packet transform, Wigner distribution and Choi-Williams distribution, Matching Pursuit(MP), Independent Component Analysis(ICA), multi-resolution time-frequency analysis based on wavelet packet decomposition and short time Fourier transform(STFT), and ANN. Because of the diversity and complexity of feature waveform in EEG signals, the all feature information in EEG signals can not be extracted efficiently by one of time-frequency analysis methods only, it is necessary for using multi time-frequency analysis methods synthetically. The time-frequency analysis methods discussed above are integrated into the virtual EEG measuring and analysis instrumentation developed by the author. The concept of EEG basic rhythms frequency bands relative intensity ratio(BRIR) is introduced, and Gabor transform is used to realize the function of detecting BRIR automatically, to help doctors judge the restrain case of EEG rhythm rightly; ICA, multi-resolution wavelet transform, adaptive-filtering processing, continuous wavelet transform, ANN and expert knowledge rules are utilized synthetically to extract spike wave, sharp wave and slow wave in epileptic EEG; appropriate time-frequency analysis methods, such as wavelet transform and Wigner distribution etc., can be chosenself-adaptively based on the difference of waveform feature in EEG signal to show the feature information in EEG signal furthest. At the same time, EEG topography detection and the function of analysis automatically for sleep stages are integrated into the instrumentation considering the demands for the development characteristic and clinical diagnosis of electroencephalograph. The function of EEG topography can be realized by utilizing Gabor transform and sphere interpolation method; and the function of sleep stages analysis can be realized based on multi-resolution time-frequency analysis method and sleep stages rules. The application of zero-phase error data filter in EEG signal processing is also discussed in this dissertation. During the course of the realization of the instrumentation, error-phase data filter is used to preprocess the sampled EEG data and extract the EEG rhythm frequency band information with zero-phase error.In the virtual EEG measuring and analysis instrumentation system, it is not only that the functions of strong feature information extraction of EEG signal are integrated, but also multi displaying controlling parameters are set to get various flexible display manners by considering the convenience that doctors look at the EEG waveforms in clinical application. Furthe...
|
PDF Full Text Request