Abstract

Optokinetic nystagmus (OKN) is a fundamental gaze-stabilising response in which eye movements attempt to compensate for the retinal slip caused by self-motion. The OKN response consists of a slow following movement made in the direction of stimulus motion interrupted by fast eye movements that are primarily made in the opposite direction. The timing and amplitude of these slow phases and quick phases are notably variable, but this variability is poorly understood. In this study I performed principal component analysis on OKN parameters in order to investigate how the eigenvectors and eigenvalues of the underlying components contribute to the correlation between OKN parameters over time. I found three categories of principal components that could explain the variance within each cycle of OKN, and only parameters from within a single cycle contributed highly to any given component. Differences found in the correlation matrices of OKN parameters appear to reflect changes in the eigenvalues of components, while eigenvectors remain predominantly similar across participants, and trials. I have developed a linear and stochastic model of OKN based on these results and demonstrated that OKN can be described as a 1st order Markov process, with three sources of noise affecting SP velocity, QP triggering, and QP amplitude. I have used this model to make some important predictions about the optokinetic reflex: the transient response of SP velocity, the existence of signal dependent noise in the system, the target position of QPs, and the threshold at which QPs are generated. Finally, I investigate whether the significant variability within OKN may represent adaptive control of explicit and implicit parameters. iii

Keywords

optokinetic nystagmus, markov model, principal components analysis, cost model, quick phase interval duration, distribution analysis, positional error

Document Type

Thesis

Publication Date

2012

Download

Share

COinS