Sequential non-parametric change-point detection for time series models: assessing the functional dynamics of neuronal networks
Fabio Rigat
CRiSM, Department od Statistics, University of Warwick, UK

This work illustrates the application of a novel sequential non-parametric method to detect significant changes of the functional connectivity of neuronal networks using in-vivo multiple spike trains recordings obtained via multi-electrode arrays. Rather than relying on a parametric specification of the parameters' evolution, their dynamics are assessed as a change-point problem in discrete time within a hypothesis testing framework. The Kullback-Leibler divergence between the posterior distributions of different sets of data under the same model is proposed as a test statistic. Markov chain Monte Carlo posterior simulation is used in general to approximate the value of the Kullback-Leibler statistic and its critical region under the null hypothesis of no change. The changes in the network parameters detected by the Kullback-Leibler statistic explain variations of the baseline neuronal spiking rates as well as of the functional connections between neurons across different experiments.