The blind pattern recognizing in neural spike train

A method of blind pattern recognition in a neural spike train is proposed. Examples of pattern recognition in a simulation experiment and a neural spike train of a rat brain are given. It is shown that the neural spike train is the carrier of neural information and that just three to seven codes (in pattern) can describe completely the information contained in more than 1000 data sequences using the proposed method. The favored pattern can be extracted in the absence of a priori knowledge and when patterns occur randomly and local pattern features are destroyed. Because the method can extract candidate pattern sets directly and easily, and it uses the techniques of parallel match, fast shuffling, and optimum search, its efficiency and detectability are far superior to those of some previously proposed methods. Examples of pattern recognition in a simulation experiment and in a neural spike train of a rat brain are given