Effects of firing patterns on spike strain classification performance of liquid state machine

For the various changes of the environment inside and outside the body, the nervous system expresses the information with different firing patterns. The key point of revealing the neural coding mechanism is to analysis the neuronal firing activity. In particular, Spiking and bursting are two most common firing patterns. Relevant researches have demonstrated their remarkable performance on information processing and transmission. In this paper, we have consciously compared the effects of bursting and spiking patterns on the classification capability of spike trains based on Liquid State Machine (LSM). According to calculating the mean squared error(MSE), the simulation results show that bursting pattern has more advantages for spike trains classification tasks.