A Comparative Study of Pattern Detection Algorithm and Dynamical System Approach Using Simulated Spike Trains

We apply two different approaches—pattern detection algorithm and dynamical system analysis—to study sets of simulated spike trains produced by chaotic attractors and Poisson processes. We show that both algorithms are able to detect a deterministic activity in the chaotic spike trains and they are tolerant to the presence of noise in input data. A method for noise filtering in input data series is proposed and its application is demonstrated for the simulated data sets.