The Analysis Of Improved Relative Transfer Entropy And Transfer Entropy For Adaptive Template Method Based On EEG

The wealth of neurons distribute on cortex, after a violent or playful activity, these neurons can produce a series of biological signals which call electroencephalogram(EEG) signal. This signal exists in the central nervous system, and is a spontaneous electrical activity. Some years ago, the medical field had already paid attention to the analysis of biological signals. Some techniques have been applied to the clinical, including a large number of physiological and pathological information, EEG attracts widespread intention.This thesis was mainly focused on the following three aspects:Firstly, the relative transition entropy analysis of epilepsy electroencephalogram signal.Extracting the advantages of the relative entropy and the transfer entropy, This letter presents a new method called relative transition entropy which is based on forward sequence and its reverse sequence. Applying this method to study the irreversibility of normal and epileptic EEG, which shows that the value of epileptic entropy is less than the normal. All these indicate that the relative transfer entropy can be used as a parameter on irreversible degree of physical process, which leads to an important significance of distinguishing whether the patient suffers from epilepsy disease or not by using EEG.Secondly, the Multiscale relative transition entropy analysis of electroencephalogram.In this letter we analysis the white noise and pink noise, the adolescent and adults EEG as well as normal and epileptic EEG. The results indicate that the tendency of different types EEG among the different scales are distinctive as well as the multiscale entropy can distinguish between different physiological and pathological signals.Thirdly, transfer entropy analysis of EEG-based adaptive template method.We propose a new method that is based on adaptive template and apply this method to human EEG signals and investigate its statistical properties, using the transfer entropy proposed by Thomas Schreiber. The results of experiment show that both the youth and adults, as compared with the convenient base template of symbolic transfer entropy. this new method can improve the coupling of their EEG signals and be better to capture the signal dynamic information as well as the change of dynamical complexity.