Research On Multi-parameter Sleep Stage And Its Application

With the increasing pressure from the life and work and also population agingaround the world today, insomnia has become one of the most popular healthproblems and also a danger disease since it can seriously damage people’s physicaland mental health and reduce the quality of our life. So, it is an important medical andbiomedical engineering task to evaluate the sleep quality and look for the more secureand reliable method to improve sleep quality.Accurate sleep stage is the basis to valuate sleep quality and diagnose disease.The existing automatic sleep staging methods mostly use EEG signal, which is muchwell developed and cannot match the higher accuracy rate people demanded. So, it isan hot research to explore the change rules of other physiological signals, design themore effective multi-parameters sleep staging method and put it into application.This paper is aimed to a design multi-parameter sleep stage method whichcombines EEG and ECG signals. And then use this method to evaluate the therapeuticeffect of low frequency magnetic (LFM) for insomnia.The research contents of this paper are as follows: first, we explore the changerules of sleep EEG and ECG characteristics in the time domain, frequency domainand nonlinear. Second, against information redundancy, we come up with a newautomatic sleep staging method which reduce the dimensionality of parameters usingPrincipal Component Analysis (PCA) and then classify with Support Vector Machine(SVM). Third, we compare the multi-parameter staging method with EEG stagingmethod, result shows that multi-parameter method have higher accuracy rate. Last, wedesign experiment which asks subjects sleep with LFM. By analyzing the sleepstructure before and after treatment using the algorithm we designed and the Arsenal’ssleep scale we find that LFMS can treatment insomnia effectively.To sum up, this dissertation researches into the change rules of sleep EEG andECG and designs a new multi-parameter sleep staging method, and uses this methodto evaluate the therapeutic effect of LFMS, which provide experimental foundationfor sleep study.