A new spike train distance measure

In this paper a new spike train distance measure has been proposed, which is purely dependent on statistical features of the two spike trains, the distance between which is to be measured. The accuracy can be improved by increasing the number of features. Here we have selected only two features and reported performance with high accuracy on spike trains generated by (1) Poisson process and (2)Izhikevich method . We have done performance analysis by two different methods namely, (1) area under ROC curve and (2) area under quasi-ROC curve. It has been indicated how accuracy can be further improved by increasing the number of features.