Abstract:
|
This talk focuses on the statistical analysis of neural spike trains, which are instantaneous electrochemical waves through which nerve cells communicate. Frequently, these spikes are modelled as realizations of an inhomogeneous Poisson process. In this talk we outline the limitations of such a model and introduce the Skellam process with resetting as a powerful alternative. Skellam process with resetting is defined as the difference between two independent Poisson processes with a modification to accommodate the neuronal refractory period. The interesting problem of inter-spike interval (ISI) distribution is addressed, and the closed form ISI distribution within Skellam process with resetting is derived. Also introduced is the multivariate Skellam distribution as well as the multivariate Skellam process with resetting. We show that unlike traditional models in the literature, capturing negative correlation among spike trains (inhibitory neurons) is not a limitation for this new framework. Computationally efficient methods are illustrated in simulations, and real data analysis provide promising results for this new approach.
|
ASA Meetings Department
732 North Washington Street, Alexandria, VA 22314
(703) 684-1221 • meetings@amstat.org
Copyright © American Statistical Association.