Loading Events

Statistics Colloquium by Melissa Swann ’20, Wednesday, February 19

Wed, February 19th, 2020
1:10 pm
- 1:50 pm

  • This event has passed.
Image of Stetson Court classroom

An Inhomogeneous Poisson Process for Modeling Neural Spike Data by Melissa Swann ’20, 1:10 – 1:50 pm, Stetson Court Classroom 105, Statistics Colloquium

Abstract:  When we want to model events that occur over time or space, we can use a Poisson process.  This model makes two major assumptions: increments are stationary (meaning that the rate of event occurrence is constant over time or space) and non-overlapping increments are independent.  In practice, however, these assumptions are not always valid.  In neuroscience, for example, researchers model the baseline behavior of neurons to better understand the properties of well-functioning brains. Their research has shown that the rate of neural spikes is a function of other covariates and is therefore not constant over time. We will look at an example in which a primate performs a two-dimensional limb-reaching activity.  Using an inhomogeneous Poisson process, which relaxes the stationarity assumption, we will discuss how to model the primate’s neural spikes as a function of speed and direction.

Event/Announcement Navigation

Related Events