Monash University

Restricted Access

Reason: Access restricted by the author. A copy can be requested for private research and study by contacting your institution's library service. This copy cannot be republished

Investigating binocular rivalry in healthy individuals and bipolar disorder: Excluding confounds and optimising methods for large-scale endophenotype studies

posted on 2017-02-07, 04:34 authored by Phillip Cheuk Fung Law
Binocular rivalry (BR) occurs when two dissimilar images concurrently presented to corresponding retinal locations of each eye stochastically alternate in perception. A slow rate of BR has been proposed as a potential endophenotype for the heritable psychiatric condition, bipolar I disorder (BD), because BR rate is slower in BD than in healthy individuals and the trait is under substantial genetic influence. The evidence thus far indicates that the slow BR trait meets several criteria to be considered an endophenotype for BD (i.e., high sensitivity, heritability and reliability). However, definitively assessing the endophenotype utility of the BR trait requires large-scale datasets (N=1,000s to 10,000s), and existing in-lab BR viewing methods cannot feasibly achieve such massive recruitment targets. A main aim of this thesis is therefore to investigate a method of inducing BR that is suitable for an online BR test, so that this method may be applied in large-scale endophenotype studies of BR. A further main aim is to directly examine whether group separation between BD and healthy groups, using BR rate, can be maximised by manipulating various stimulus parameters that also determine both stimulus strength and the degree of perceptual mixing between each eye‘s presented image (mixed-percept duration; MPD). Furthermore, because the slow BR trait could conceivably reflect eye movement (EM) dysfunction in BD, the thesis also aims to examine this proposal.
   Chapter 1 comprehensively evaluates various aspects of viewing methods for BR research, with a view to determining the most suitable method for online BR testing. This published critical review proposes the use of anaglyphs — i.e., cardboard frame glasses with complementary monochrome (red and blue) filters — for online BR testing, as a feasible, low-cost strategy for achieving the very large datasets required to assess the clinical and endophenotype potential of the slow BR trait.
   Chapter 2 presents a published experiment that examines the relationship between EM profiles and BR rates in 40 healthy individuals. All subjects underwent EM tasks and separate BR tasks using a range of drifting stimuli with different speeds and sizes. Correlations were conducted to determine significant associations between each EM measure and BR rate for all stimulus conditions. Bayesian analyses were also performed. This chapter provides evidence for no association between each EM measure and BR rate for all stimuli. This experiment suggested the need for directly examining the relationship between EMs and BR rate in subjects with BD.
   Chapter 3 presents an experiment that repeats the preceding experiment, this time in 20 subjects with BD and 20 age- and sex-matched healthy controls. All subjects underwent the same EM tasks used in the preceding experiment and separate BR tasks using various drifting stimuli with different drift speeds, including stationary green gratings. In addition to the statistical analyses performed in the preceding experiment, between-group comparisons were also conducted to probe group differences in EM measures. Frequentist and Bayesian statistics were performed. The results indicated no EMs that were substantially associated with BR rate, where those EMs were also substantially different between BD and control groups. This chapter provides evidence that EM profiles do not explain the slow BR endophenotype for BD.
   Chapter 4 directly assesses the effects of stimulus strength variation on BR rate in 40 healthy individuals. Greater stimulus strength is associated with particular physical manipulations of the BR stimulus such as higher contrast, faster drift speed, and brighter luminance. However, the sensitivity function is not always monotonic. In this repeated measures within-subjects experiment, all subjects completed BR tasks using high-strength orthogonally drifting green gratings that varied in drift speed and size. MPD was also assessed with a view to minimise this parameter within a given BR observation period, to provide a more representative and accurate measure of an individual‘s true BR rate. Results indicated that stimuli drifting at 8 cycles/sec in a 1.5° aperture induced a faster BR rate in healthy individuals, without increasing MPD, but that direct examination of stimulus factors was required in BD subjects. The results also confirmed Levelt‘s fourth proposition using drift speed as the stimulus strength factor, with the vast majority of healthy individuals exhibiting this stimulus modulation effect on BR rate.
   Chapter 5 directly examined individuals‘ BR rates in 20 subjects with BD and 20 age- and sex-matched healthy controls, using different high-strength stimuli (i.e., green gratings, anaglyph gratings and anaglyph rings) that varied in drift speed, as well as lower-strength stationary green gratings. This repeated-measures within-subjects experiment also retested individuals‘ BR rates for each stimulus on a separate day. MPD was also assessed. The results indicated that high-strength 4 cycles/s green gratings and 4 cycles/s anaglyph gratings are the most suitable BR stimuli to use to further investigate the clinical and endophenotype potential of the slow BR trait. The results again showed individual variation in stimulus-strength modulation of BR rate for the vast majority of healthy controls. However, almost half of BD subjects showed the opposite effect to Levelt‘s (1965) fourth proposition with consistently greater magnitudes of stimulus-strength modulation than controls. The findings demonstrated that the anaglyph 4 cycles/s gratings are a valid and reliable method for inducing BR in an online platform for use in large-scale clinical and endophenotype studies of the slow BR trait.
   The present thesis contributes to clinical BR literature and advances the field through demonstrating that: (i) EMs do not explain the slow BR trait in BD; (ii) an anaglyph-based BR test website is a feasible, low-cost, valid and reliable strategy for achieving very large BR datasets for use in clinical and genetic studies; and (iii) high-strength gratings drifting at 4 cycles/s in a 1.5° aperture are optimal stimuli for such studies. The work in the present thesis also contributes to general psychophysical knowledge regarding stimulus-strength modulation of BR rate and individual variation therein.


Principal supervisor

Steven Miller

Additional supervisor 1

Trung Ngo

Additional supervisor 2

Caroline Gurvich

Year of Award


Department, School or Centre

Central Clinical School

Campus location



Doctor of Philosophy Psychology

Degree Type



Faculty of Medicine Nursing and Health Sciences