基于非线性动力学的振幅整合新生儿脑电图分析

Analysis of Neonatal Amplitude Integrated Electroencephalogram Based on Nonlinear Dynamics

振幅整合脑电图(Amplitude integrated electroencephalogram,aEEG)即脑功能检测(Cerebral function monitor,CFM)是一种无创的脑功能检测技术,对新生儿脑损伤早期诊断及预后评价具有很好的准确性。如今,医生主要是依据振幅整合脑电图的波形和幅度对其进行分类,进而对新生儿的脑功能做出正确的诊断。对于波形和振幅特征不太明显的情况,只能依靠医生的经验进行判断,带有一定的主观因素,不能给出准确的诊断信息。本文利用非线性动力学方法对惊厥新生儿和正常新生儿的aEEG的关联维数、Lyapunov指数和近似熵进行计算分析,并将这三个特征量作为三维向量,结果在三维空间中能有效地区分惊厥新生儿与正常新生儿,表明关联维数、Lya-punov指数和近似熵能够反映aEEG的内在信息,可以作为一种研究aEEG的新方法。

Amplitude integrated electroencephalogram （aEEG）, also known as cerebral function monitor （CFM）, is a non-invasive detection of brain function, having good accuracy in early diagnosis and prognosis evaluation of neonatal brain damage. Today, doctors classify amplitude integrated electroencephalogram mainly based on its waveform and amplitude, then they make correct diagnosis of brain function of neonates. However, in some cases, the amplitude and waveform of aEEG are not very clearly shown, the only way is relying on doctors＇ experience to give out judgement, which has some subjective factors, so doctors can not present accurate diagnostic information. To solve this problem by the use of non-linear dynamics, we calculate and analyze the correlation dimension, Lyapunov exponent and approximate entropy of the aEEG for neonates with convulsions and for normal neonates~ then we take these three characters as a three-dimensional vector; finally, the aEEG of neonates with convulsions and the aEEG of normal neonates are distributed into two parts in three-dimensional space, thus the correlation dimension, Lyapunov exponent and approximate entropy of the aEEG can reflect the internal information of neonatal brain function. Therefore, it can be used as a new method for studying neonatal aEEG.