"Everything should be made as simple as possible, but not simpler." - Albert Einstein

Seminar Announcement

Dynamical Multiple-Input, Single-Output Model of Neural Spike Transformation

Yan Catherine Tu, M.S. Candidate, Department of Statistics, Colorado State University

Tuesday, April 6, 2010

10:00 a.m., room 008, Statistics Bldg

ABSTRACT

In this paper, we consider the problem of modeling neural signal transformation. A dynamic Multiple-Input, Single-Output model of neural information communication is proposed. Each input neuron and the output neuron have a functional relationship which is approximated by polynomial splines. A penalized likelihood approach is implemented for simultaneous parameter estimation and variable selection. The notion of sparsity in parameter estimation has been generalized to function estimation. Two different types of functional sparsity are of particular interest: global sparsity and local sparsity. Computation of the penalized approach is rather challenging. The one-step estimator based on the group bridge approach for maximizing the penalized likelihood is proposed. The performance of the proposed method is assessed using Monte Carlo simulation studies.

Advisory Committee:

Dr. Haonan Wang, Advisor
Dr. F. Jay Breidt, Committee Member
Dr. Phil Chapman, Committee Member
Dr. Jie Luo, Electrical & Computer Engineering, Outside Committee Member