Entropy Coding Of Neural Firing And Its Application On Epilepsy Pathogenesy Study

We raised and performed Shannon entropy coding method due to the limitation of rate coding method, inter-spike interval (ISI) temporal coding method and some entropy coding methods of neural spike. We verified availability of our method and studied problem of select bin width by using simulation data, then studied Shannon entropy coding of normal rats and temporal epilepsy rats hippocampus neural firing,it made foundation for the study of hypothesis of super synchronization neural firing, finally, comparison between Shannon entropy coding and ISI temporal coding was made and showed the superiority of our method.The main methods of this thesis is to get inter-spike intervals form signals at first, and then do Shannon entropy coding which have dynamic information, it includes, make inter-spike interval histogram,calculate entropy of different bin width and do Shannon entropy coding on the foundation of suitable selection of bin width.Results are simulation train of neural bursting firing and beating firing has different entropy coding pattern, because of that entropy of bursting firing simulation train is apparently lower than entropy of beating firing simulation train. And entropy of temporal epilepsy rats hippocampus neural firing(6.36±0.207 bit) is apparently lower than the entropy of normal rats hippocampus neural firing(7.44±0.472 bit), P<0.05.It can be concluded that Shannon entropy coding can quantitative analysis different types of nonlinear neural firing effetely due to the simulation study and has superiority to ISI temporal coding, because ISI temporal coding can only qualitative analysis different types of coding pattern. Robustness of Shannon entropy coding depends on selection of proper bin width and our study showed that region between 0.265 ms to 0.485 ms is the best choice when doing Shannon entropy coding of normal rats and temporal epilepsy rats hippocampus neural firing. The conclusions above can support the study of super synchronization firing of temporal lobe epilepsy neurons, because we can use entropy to apart super synchronization firing neurons from the neuron pool and make foundation of neuron pool entropy coding study.