Wavelet Coding For Neural Spikes And Its Application On Epilepsy Study

Due to the limitation of rate coding and temporal coding about firing of neuron ,we raise and perform wavelet coding which can analyze signals in time-frequency field ,and we try to use wavelet coding to investigate the characteristic patterns of firing of neuron in temporal lobe epilepsy cat hippocampus. Our study provide a feasible and effective coding method for studying the mechanism of epilepsy ?the hypothesis of supersynchronization firing of neuron.The main purpose of this thesis are as follows. First, to investigate the character and approach of wavelet coding with the point of view of information coding. Second, to perform wavelet coding on characteristic scales on supersynchronization firing of epilepsy neuron, in order to provide a feasible coding method for the hypothesis of supersynchronization firing.Our thesis includes two parts. The first part is to simulate firing of neuron by using of Chay model in order to prove the feasibility and effect of wavelet coding. In second part, on the basis of the feasibility of wavelet coding, we perform wavelet coding on interspike intervals of neuron in normal and temporal lobe epilepsy cat hippocampus.The main methods of this thesis is to get interspike intervals form signal at first, and then do wavelet coding on interspike intervals which have dynamic information, it includes, to select suitable mother wavelet and scales, to select characteristic component and do wavelet coding.The results are as follows. First, after wavelet transform performed, the beating and bursting signal is decomposed into different components in 5 scales. TheAbstractcharacteristic component is scale 1 and scale 2. Second, after wavelet coding performed in 5 scales, the interspike intervals of normal neuron is different form it of temporal lobe epilepsy neuron on scale 1, scale 2 and scale 3. Third, the type of firing of temporal lobe epilepsy neuron is belong to bursting.It can be concluded that we can perform wavelet coding on one or several characteristic component. After wavelet coding performed, the signals which are not very different from each other on whole may show some differences on one or several characteristic scales. It can be induced from the results of wavelet coding of bursting-firing that bursting-firing may be consisted of several beating-firing. Wavelet coding is good for studying the supersynchronization firing of temporal lobe epilepsy neurons which discharge like bursting-firing.