Realistic Computer Modeling of the Mammalian Olfactory Cortex

Author: Vanier, Michael Christopher

Year: 2001

Degree: Dissertation (Ph.D.)

Advisor: Bower, James M.

Committee Members: Pine, Jerome; Bower, James M.; Barr, Alan H.; Goodman, Rodney M.; Koch, Christof

Option: Computation and Neural Systems

DOI: 10.7907/QM8V-8404

Abstract

A combination of experimental and computer modeling techniques were used to investigate the dynamics and computational functions of the rat olfactory (piriform) cortex. Experimental characterization of synaptic response to afferent and associational fiber voltage shocks were performed, in the presence and absence of the neuromodulator norepinephrine. This data was used to generate computer models of synaptic transmission in piriform cortex. Models of pyramidal neurons and feedback inhibitory interneurons were constructed which accurately match intracellular experimental data in the presence and absence of norepinephrine. In order to achieve this, parameter search tools for automatically matching computer models of neurons to data were developed. Models of feedforward inhibitory interneurons were also constructed. An abstract spike generating model of the olfactory bulb was built. These components were combined to create a realistic computer model of the piriform cortex. This model can accurately replicate the response of the real system to a strong shock stimulus, as reflected in current source density plots. Two versions of the model were created to model the oscillatory response of the system to week shots. The first model replicates the surface field potential with considerable accuracy, but fails to replicate the current source density data. The second model replicates the current source density data and suggests a new organizing principle for the olfactory system based on non-overlapping neuronal groups. This hypothesis is experimentally testable.

Files