Measuring Spike Train Correlation with Non-Parametric Statistics Coefficient
Jorge Humberto Soletta, Fernando Daniel Farfán, C.J. Felice · IEEE Latin America Transactions · 2015
Measure correlation between spike trains is a fundamental step for the study of neural systems. There are many alternatives to measure correlation, but not all possess the required properties. In this paper we propose to use non-parametric coefficients of correlation, coefficients Spearman and Kendall. To analyze their properties were generated computationally trains of spikes that simulate different experimental conditions, then the proposed coefficients were calculated and compared with the Pearson coefficient. The results show that under certain experimental conditions Kendall coefficient is more appropriate to quantify correlations between spikes trains.