Predicting Neural Spike Train Using Peers' Activities: Likelihood-Based Approach

• Activity Number: 629
• Type: Contributed
• Date/Time: Thursday, August 5, 2010 : 8:30 AM to 10:20 AM
• Sponsor: ENAR
• Author(s): Ruiwen Zhang*+
• Companies: The University of North Carolina at Chapel Hill
• Keywords: neuronal spike train ; point process ; likelihood function ; peer prediction

Abstract

Neurons can produce action potentials with high temporal precision. The discrete, all-or-nothing nature of a sequence of neuronal action potentials together with their stochastic structure suggests that a spike train can be viewed as a point process. Likelihood function starting from the point process has been derived and popularly used in the field of neuron spike train analysis to estimate some biologically interpretable parameters. In this paper, a general statistical framework based on the point process likelihood function is constructed to predict neuron spike trains. From the result, both the real data analysis and the simulation, we can draw two conclusions as (a) better predictions can be made by involving more peer information and (b) one can have unbiased estimation by using the likelihood-based approach which is validated in the simulation.