Year of Award

2018

Document Type

Dissertation

Degree Type

Doctor of Philosophy (PhD)

Degree Name

Mathematics

Department or School/College

Mathematical Sciences

Committee Chair

Professor Emily F. Stone

Committee Co-chair

None

Commitee Members

Dave Patterson, Brian Steele, Jonathan Bardsley, Nathan Insel

Keywords

short-term synaptic plasticity, mutual information, cholinergic modulation, hippocampal GABAergic synapses, epsilon machines, synaptic filtering, interneuron-pyramidal cell synapses, causal state splitting reconstruction

Publisher

University of Montana

Subject Categories

Computational Neuroscience | Dynamic Systems | Other Applied Mathematics | Other Statistics and Probability | Probability | Statistical Theory

Abstract

Neurons convey information about the complex dynamic environment in the form of signals. Computational neuroscience provides a theoretical foundation toward enhancing our understanding of nervous system. The aim of this dissertation is to present techniques to study the brain and how it processes information in particular neurons in hippocampus.

We begin with a brief review of the history of neuroscience and biological background of basic neurons. To appreciate the importance of information theory, familiarity with the information theoretic basics is required, these basics are presented in Chapter 2. In Chapter 3, we use information theory to estimate the amount of information postsynaptic responses carry about the preceding temporal activity of hippocampal interneuron synapses and estimate the amount of synaptic memory. In Chapter 4, we infer parsimonious approximation of the data through analytical expression for calcium concentration and postsynaptic response distribution when calcium decay time is significantly smaller that the interspike intervals.

In Chapter 5, we focus on the study and use of Causal State Splitting Reconstruction (CSSR) algorithm to capture the structure of the postsynaptic responses. The CSSR algorithm captures patterns in the data by building a machine in the form of visible Markov Models. One of the main advantages of CSSR with respect to Markov Models is that it builds states containing more than one histories, so the obtained machines are smaller than the equivalent Markov Model.

Share

COinS
 

© Copyright 2018 Elham Bayat Mokhtari