The Research And Design Of Neonates Cerebral Function Monitor And Analysis Of AEEG Based On Nonlinear Dynamics

Amplitude-integrated electroencephalogram, also known as cerebral function monitor, is a kind of non-invasive detection for neonatal brain function. CFM with intuitive graphical is easy to use and analyze, and can be using for long time in the bedside. Nowadays, more and more researchs show that aEEG has high accuracy on the early diagnosis and prognosis of neonatal brain injury. In view of the CFM produced only by Brainz, Lectromed Devices Ltd, Olympic Medcial, and so on a small number of corporations, because of their possession of intellectual property rights, they sell it with high price. As a result, the development of CFM with independent intellectual property rights can bring very important economic and social benefits, which including the detection algorithm of related disease for auto-diagnosis and early warning of neonatal brain disease.In this paper, according to the framework of CFM proposed by Pamela Prior and Douglas Maynard, we divided CFM into infant EEG acquisition module and total controlling module. In the nfant EEG acquisition module, we use a special three-level amplification circuit and driver circuit, and acquire good EEG. In the total controlling module, we filtered out the noises from EEG and extracted the usefull signal using the design principle of integer coefficient filter metioned in chapter two. At the same time, we designed the amplitude-integrated algorithm according to that the peak-to-peak amplitude of the EEG bursts defines the upper border of the aEEG trace, and the peak-to-peak amplitude of the interburst EEG defines the lower border of the trace.In order to eliminate the subjective factors of the traditional methods of aEEG analysis based on its waveform and amplitude, through analyzing the aEEG deeply, we firstly applied the non-linear dynamics to aEEG automatic diagnosis, and combined the mathematical statistics. Finally compared with tradional time-domain analysis method, it not only eliminated the subjective factors effectively, but also achieved remarkable results. In details we calculated and analyzed the correlation dimension, Lyapunov exponent and approximate entropy of the aEEG for convulsions and normal neonates; then made them as a three-dimensional vertor; finally we can easily find the aEEG of convulsions and normal neonates were respectively distributed into two parts in three-dimensional space. So this method was an efficient approach to the quantitative analysis of aEEG.