Method for estimating neural network topology based on SPIKE-distance

Abstract

To understand information processing in the brain, it is important to clarify the neural network topology. We have already proposed the method of estimating neural network topology only from observed multiple spike sequences by quantifying distance between spike sequences. To quantify distance between spike sequences, the spike time metric was used in the conventional method. However, the spike time metric involves a parameter. Then, we have to set an optimal parameter in the spike time metric. In this paper, we used the SPIKE-distance instead of the spike time metric and applied a partialization analysis to the SPIKE-distance. The SPIKE-distance is a parameter-free measure which can quantify the distance between spike sequences. Using the SPIKE-distance, we estimate the network topology. As a result, the proposed method exhibits higher performance than the conventional method.