AUTOMATIC WAVEFORM DETECTION AND MODEL PARAMETER ESTIMATION FOR OPTOKINETIC NYSTAGMUS.
Item
-
Title
-
AUTOMATIC WAVEFORM DETECTION AND MODEL PARAMETER ESTIMATION FOR OPTOKINETIC NYSTAGMUS.
-
Identifier
-
AAI8302546
-
identifier
-
8302546
-
Creator
-
SINGH, AVTAR.
-
Contributor
-
Frederick E. Thau
-
Date
-
1982
-
Language
-
English
-
Publisher
-
City University of New York.
-
Subject
-
Engineering, Electronics and Electrical
-
Abstract
-
The purpose of this dissertation is to develop algorithms that detect and separate the slow and quick eye movements, and estimate parameters of the slow eye movement control system. The slow eye movement control system is important since it gives clues to the pathological state of the neural system where these movements are generated.;The detection algorithm compares a window of eye position samples to templates of quick and slow eye movements. Based on minimizing the probability of error, a decision function is formulated to assign the latest received sample to a slow or a rapid eye movement. Two kinds of errors can be made in the detection process. The first kind is--when all the samples in the window are from one kind of movement and it is detected as the other kind. This kind of error is due to the inherent noise on the samples. The second kind of error occurs when the data window consists of samples from both kinds of movements. This error is due to transitions between movements, their temporal relations and the window width used. An expression for the probability of error accounting for these errors is derived.;Once a quick phase or a saccade is detected, it is removed by subtracting the saccadic jump and extrapolating the slow eye position during a saccade. This generates a cumulative slow phase position whose differentiation yields the slow phase eye velocity.;The slow phase velocity obtained by processing an optokinetic nystagmus is utilized in the parameter estimation of a model for the system which generates optokinetic nystagmus (OKN) and optokinetic after-nystagmus (OKAN). The estimated parameters are used to evaluate the model.;Contributions of this research lie in the application of detection theory, nonlinear digital filtering, and parameter estimation to the oculomotor system analysis. The computer implementation of signal processing algorithms done here is potentially important for the development of automated diagnostic techniques.
-
Type
-
dissertation
-
Source
-
PQT Legacy CUNY.xlsx
-
degree
-
Ph.D.
-
Program
-
Engineering
