Variation of the Stiles-Crawford Effect with Accommodation ... · Variation of the Stiles-Crawford...

Variation of the Stiles-Crawford Effect with

Accommodation and Myopia

Nisha Singh BSc (Optom)

A thesis submitted in fulfilment of the requirements of the degree of Doctor of Philosophy

School Of Optometry Institute of Health and Biomedical Innovation

Queensland University of Technology 2009

Keywords

ii

Keywords Stiles-Crawford effect Accommodation Myopia Psychophysical technique Multifocal electroretinogram Aberrations Accommodative lag

Abstract

iii

Abstract

Background: Mechanical forces either due to accommodation or myopia may

stretch the retina and/or cause shear between the retina and choroid. This can be

investigated by making use of the Stiles-Crawford effect (SCE), which is the

phenomenon of light changing in apparent brightness as it enters through different

positions in the pupil. The SCE can be measured by psychophysical and objective

techniques, with the SCE parameters being directionality (rate of change across the

pupil), and orientation (the location of peak sensitivity in the pupil).

Aims: 1. To study the changes in foveal SCE with accommodation in emmetropes

and myopes using a subjective (psychophysical) technique. 2. To develop and

evaluate a quick objective technique of measuring the SCE using the multifocal

electroretinogram.

Methods: The SCE was measured in 6 young emmetropes and 6 young myopes for

up to 8 D accommodation stimulus with a psychophysical technique and its variants.

An objective technique using the multifocal electroretinogram was developed and

evaluated with 5 emmetropes.

Results: Using the psychophysical technique, the SCE directionality increased by

similar amounts in both emmetropes and myopes as accommodation increased, with

an increase of 15-20% with 6 D of accommodation. However, there were no

significant orientation changes. Additional measurements showed that most of the

change in the directionality was probably an artefact of optical factors such as

higher-order aberrations and accommodative lag rather a true effect of

accommodation. The multifocal technique demonstrated the presence of the SCE, but

results were noisy and too variable to detect any changes in SCE directionality or

orientation with accommodation.

Conclusion: There is little true change in the SCE with accommodation responses up

to 6 D in either emmetropes or myopes, although it is possible that substantial

changes might occur at very high accommodation levels. The objective technique

using the multifocal electroretinogram was quicker and less demanding for the

subjects than the psychophysical technique, but as implemented in this thesis, it is

not a reliable method of measuring the SCE.

Contents

iv

Table of Contents

Keywords ..................................................................................................................... ii

Abstract ...................................................................................................................... iii

Table of Contents ...................................................................................................... iv

Table of Figures ....................................................................................................... viii

Table of Tables ......................................................................................................... xii

Statement of original authorship ........................................................................... xiv

Acknowledgements ................................................................................................... xv

CHAPTER 1: INTRODUCTION ............................................................................. 1

1.1 Background ........................................................................................................ 1 1.2 Research objectives ............................................................................................ 3 1.3 Thesis outline ..................................................................................................... 4

CHAPTER 2: LITERATURE REVIEW ................................................................. 5

2.1 The Stiles-Crawford effect ................................................................................. 5 2.1.1 Mathematical description ............................................................................ 5 2.1.2 Basis of the SCE .......................................................................................... 7 2.1.3 Photoreceptor alignment ............................................................................. 8 2.1.4 Effects of luminance and wavelength ......................................................... 9 2.1.5 Peak location and directionality ................................................................ 10 2.1.6 Measurement techniques ........................................................................... 11

2.1.6.1 Psychophysics .................................................................................... 11 2.1.6.2 Electrophysiological measurement .................................................... 12 2.1.6.3 Fundus reflectometry technique including Adaptive optics ophthalmoscopy ............................................................................................. 18

2.2 Accommodation and the Stiles-Crawford effect .............................................. 20 2.2.1 Mechanism of accommodation and retinal deformation during accommodation .................................................................................................. 21 2.2.2 Characteristics of accommodation ............................................................ 22 2.2.3 Influence of accommodation on the Stiles-Crawford effect ..................... 23 2.2.4 Measurement of accommodation .............................................................. 24 2.2.5 Effect of phenylephrine on accommodation ............................................. 25

2.3 Myopia and the Stiles-Crawford effect ............................................................ 26

Contents

v

2.3.1 Myopia and accommodation ..................................................................... 27 2.3.2 Myopia induced changes to the retina ....................................................... 28

2.4 Summary .......................................................................................................... 29

CHAPTER 3: THE STILES-CRAWFORD EFFECT IN EMMETROPES AND MYOPES USING A PSYCHOPHYSICAL TECHNIQUE – METHODS ......... 31

3.1 Introduction ...................................................................................................... 31 3.2 Subjects ............................................................................................................ 32 3.3 Apparatus ......................................................................................................... 33

3.3.1 Background channel .................................................................................. 35 3.3.2 Test channel .............................................................................................. 35 3.3.3 Subject alignment ...................................................................................... 36 3.3.4 Focus setting to stimulate accommodation or to correct refractive errors 36 3.3.5 Replacement of 1mm aperture with 4 mm aperture in the background channel ............................................................................................................... 36

3.4 Procedure ......................................................................................................... 38 3.5 Mathematical fits of Stiles-Crawford effect ..................................................... 41 3.6 Calibration of the PowerRef II ......................................................................... 42 3.7 Accommodative stimulus-response functions ................................................. 43 3.8 Control measurements ...................................................................................... 44

3.8.1 Does source S2 directionality influence the Stiles- Crawford measurement? ............................................................................................................................ 44 3.8.2 Does aperture movement influence the Stiles-Crawford measurement? .. 46 3.8.3 Effect of sampling intervals on Stiles-Crawford measurement ................ 53 3.8.4 Influence of contact lens, aberrations, and accommodative lag on the Stiles-Crawford effect ........................................................................................ 54

3.9 Peak-finding technique .................................................................................... 63 3.10 Aberration measurements for emmetropes .................................................... 65

3.10.1 Subjects ................................................................................................... 66 3.10.2 Instrumentation ....................................................................................... 66 3.10.3 Calibration of the instrument .................................................................. 67 3.10.4 Procedure to measure ocular aberrations ................................................ 68 3.10.5 Data analysis ........................................................................................... 69

CHAPTER 4: THE STILES-CRAWFORD EFFECT IN EMMETROPES AND MYOPES USING A PSYCHOPHYSICAL TECHNIQUE - RESULTS ............ 70

4.1 Emmetropes ..................................................................................................... 70 4.1.1 PowerRef II calibration ............................................................................. 70 4.1.2 Accommodative stimulus-response functions: 1 mm vs. 4 mm aperture . 71 4.1.3 Accommodation responses of subjects during Stiles-Crawford measurements ..................................................................................................... 73 4.1.4 Changes in the Stiles-Crawford effect with accommodation .................... 75

4.2 Myopes ............................................................................................................. 78 4.2.1 PowerRef II calibration ............................................................................. 78 4.2.2 Accommodative stimulus-response functions .......................................... 79

Contents

vi

4.2.3 Accommodation responses of subjects during Stiles-Crawford measurements ..................................................................................................... 80 4.2.4 Changes in Stiles-Crawford effect with accommodation .......................... 81

4.3 Changes in Stiles-Crawford effect with accommodation for combined groups ................................................................................................................................ 84 4.4 Peak-finding technique ..................................................................................... 86 4.5 Aberration measurements of emmetropes ........................................................ 87 4.6 Discussion ........................................................................................................ 91

4.6.1 Changes in the SCE directionality with accommodation .......................... 91 4.6.2 Changes in the SCE peak location with accommodation .......................... 91 4.6.3 Comparison of SCE with accommodation between emmetropes and myopes ............................................................................................................... 92

4.7 Summary .......................................................................................................... 93

CHAPTER 5: THE STILES-CRAWFORD EFFECT USING THE MULTIFOCAL ELECTRORETINOGRAM ....................................................... 94

5.1 Introduction ...................................................................................................... 94 5.2 Methodology .................................................................................................... 95

5.2.1 Subjects ..................................................................................................... 95 5.2.2 Apparatus .................................................................................................. 95 5.2.3 MfERG stimulus ....................................................................................... 97 5.2.4 Experimental procedure ............................................................................ 97 5.2.5 Calibration of the Photorefractor ............................................................... 99 5.2.6 Accommodative stimulus-response function .......................................... 100

5.3 mfERG data analysis ...................................................................................... 101 5.4 Control measurements .................................................................................... 104

5.4.1 Effect of different sizes of aperture A1 on the measured SCE using mfERG .......................................................................................................................... 104 5.4.2 Retinal illuminance measurement ........................................................... 107 5.4.3 Comparison of mfERG measurements with LCD and CRT monitors .... 108

5.5 Results ............................................................................................................ 110 5.5.1 Photorefractor calibration functions ........................................................ 110 5.5.2 Accommodative stimulus-response functions ......................................... 111 5.5.3 Accommodation responses of subjects during SCE measurements ........ 112 5.5.4 SCE measurement using mfERG stimulus for ........................................ 113 0 D and 6 D accommodation stimuli ................................................................ 113

5.5.4.1 Change in the SCE directionality with accommodation .................. 113 5.5.4.2 Shift in peak sensitivity of the SCE with accommodation ............... 117

5.6 Discussion ...................................................................................................... 119 5.6.1 Changes in the SCE directionality with accommodation ........................ 119 5.6.2 Changes in the SCE peak location with accommodation ........................ 120 5.6.3 Comparison with Sutter’s study .............................................................. 120 5.6.4 Comparison between subjective and objective techniques of measuring the SCE ................................................................................................................... 121

5.7 Summary ........................................................................................................ 122

Contents

vii

CHAPTER 6: SUMMARY AND FUTURE DIRECTIONS .............................. 123

6.1 Summary ........................................................................................................ 123 6.2 Discussion and conclusion ............................................................................. 124 6.3 Future research directions .............................................................................. 125

References ............................................................................................................... 129

APPENDICES ........................................................................................................ 145

Appendix 1: SCE apparatus .................................................................................. 145

Appendix 2: Calibration of the LCD monitor ..................................................... 146 A2.1 Introduction ................................................................................................. 146 A2.2 Apparatus .................................................................................................... 149 A2.3 Warm-up characteristics .............................................................................. 150 A2.4 Voltage-luminance relationship .................................................................. 153 A2.5 Spatial uniformity of the LCD .................................................................... 159 A2.6 Temporal luminance profile ........................................................................ 160 A2.7 Spectral characteristics ................................................................................ 161 A2.8 Chromaticity constancy of the LCD primaries ........................................... 162 A2.9 Channel independence ................................................................................ 164 A2.10 Channel constancy .................................................................................... 167 A2.11 Summary ................................................................................................... 170

Appendix 3: Publications/Presentations .............................................................. 172

Contents

viii

Table of Figures

Figure 2.1. A typical Stiles-Crawford function. ........................................................... 5

Figure 2.2. Waveform representation of a full-field ERG. ........................................ 12

Figure 2.3. Sample of a VERIS 61 hexagonal stimulus array. ................................... 15

Figure 2.4. Waveform representation of a first order mfERG response. ................... 16

Figure 2.5. Schematic diagram explaining the optical Stiles-Crawford effect (OSCE). .................................................................................................................................... 19

Figure 3.1. Schematic representation of the two-channel Maxwellian view apparatus. .................................................................................................................................... 34

Figure 3.2. Movement of the test spot for SCE measurement ................................... 40

Figure 3.3. Experimental set-up for calibration of the PowerRef II .......................... 43

Figure 3.4. ND filter transmission (%) plotted against mean pixel brightness of the test image .................................................................................................................... 45

Figure 3.5. Test spot image brightness as a function of source S2 position ............... 46

Figure 3.6. Image brightness as a function of A2 aperture position for different locations of source S2 ................................................................................................. 47

Figure 3.7. Image brightness as a function of aperture A2 position for the horizontal meridian (A) and for the vertical meridian (B) .......................................................... 48

Figure 3.8. Schematic representation of the modified apparatus ............................... 60

Figure 3.9. Mean thresholds of three runs for the central location of a 1-mm diameter pupil; without filter and with 0.4 and 0.6 ND filters ................................................. 61

Figure 3.10. Schematic representation of the peak-finding technique ...................... 64

Figure 3.11. Slider positions for two observers for different trial lens powers ......... 68

Figure 4.1. Relationship between PowerRef II measured refraction and trial lens induced refraction for the 6 emmetropic subjects ...................................................... 71

Figure 4.2a. Accommodative stimulus-response curves with (A) 4 mm aperture and (B) 1 mm aperture ...................................................................................................... 72

Contents

ix

Figure 4.2b. Mean accommodative stimulus-response curves of subjects with 4 mm and 1 mm apertures …………………………………………………………………72

Figure 4.3. Accommodative response for emmetropic subjects as a function of accommodative stimulus ............................................................................................ 74

Figure 4.4. SCE results for subjects EM and AM for 0 D and 6 D accommodative stimuli along the horizontal meridian ....................................................................... 75

Figure 4.5. Changes in ρx (A), ρy (B), xc (C) and yc (D) are plotted against the accommodative response for emmetropes ................................................................ 77

Figure 4.6. Relationship between PowerRef II measured refraction and trial lens induced refraction for the 6 myopic subjects ............................................................. 78

Figure 4.7. Accommodative stimulus-response curves for myopes .......................... 79

Figure 4.8. Accommodative responses measured during the SCE task are plotted against the accommodative stimulus .......................................................................... 80

Figure 4.9. Changes in ρx (A), ρy (B), xc (C) and yc (D) are plotted against the accommodative response for myopes ....................................................................... 83

Figure 4.10. Changes in ρx(A), ρy, (B), xc (C) and yc (D) are plotted against the accommodative response for combined data of emmetropes and myopes. ............... 85

Figure 4.11. Mean of the SCE peak pupil locations from both directions for 0 D (unaccommodated) and 6 D (accommodated) accommodative stimuli ..................... 86

Figure 4.12. The accommodative responses versus accommodation stimulus for six subjects ....................................................................................................................... 88

Figure 4.13. Higher-order RMS aberrations plotted against the accommodation response for a 5 mm pupil .......................................................................................... 89

Figure 4.14. Spherical aberration coefficients 04C as a function of accommodation

response for 5 mm diameter pupil .............................................................................. 90

Figure 5.1. Schematic representation of the apparatus using multifocal electroretinogram ....................................................................................................... 96

Figure 5.2. Picture of the stimulus array used to elicit mfERG responses ................ 97

Figure 5.3. Schematic diagram of a first order mfERG response ............................ 101

Figure 5.4. MfERG responses of the right eye of a subject AM showing (a) raw data (without filtering), and (b) filtered data. .................................................................. 102

Figure 5.5. Diagram of the stimulus array showing peak hexagon as ring 1 ........... 103

Figure 5.6. Implicit time (plot A) of N1 and P1 and amplitudes (plot B) of N1 and N1P1 of the peak response waveform with two aperture sizes ................................ 105

Contents

x

Figure 5.7. Filtered ring average plots of subject AM for (A) 12mm and (B) 18 mm aperture sizes for a smaller stimulus size; (C) 18 mm aperture size for the larger stimulus size ............................................................................................................. 106

Figure 5.8. Relationship between slope of pupillary brightness profile and trial lens induced refraction for all subjects ............................................................................ 110

Figure 5.9. Accommodative stimulus-response data for all subjects ....................... 111

Figure 5.10. Mean accommodative responses for subjects as a function of accommodative stimulus during SCE recording ...................................................... 112

Figure 5.11. Filtered ring average waveforms around peak response of subject AM for 0 D and 6 D accommodation .............................................................................. 113

Figure 5.12. Normalized mean N1 and N1P1 amplitudes as a function of pupil eccentricity between 0 D and 6 D accommodation ................................................. 115

Figure 5.13. Amplitudes of ring average waveforms for the combined data of two runs plotted against ring eccentricity for 0 D and 6 D accommodation .................. 116

Figure 6.1. Schematic representation of an adaptive optics - retinal imaging system .................................................................................................................................. 127

Figure 6.2. Flow chart demonstrating (A) the retinal image processing algorithm and (B) the procedures required for determining the SCE parameters. .......................... 128

Figure A1.1. Photograph of the two-channel Maxwellian view SCE apparatus…..145

Figure A2.1. Structure of a subpixel of an LCD. ..................................................... 147

Figure A2.2. Luminance for cold and warm starts (A) and chromaticity measurements for cold starts (B and C) as a function of time ................................. 151

Figure A2.3. Change in total colour difference (C, jnd) for cold start with respect to the final stable colour achieved. ............................................................................... 153

Figure A2.4. Measured LUT for the three colour channels. .................................... 155

Figure A2.5. Inverse LUT for red channel ............................................................... 156

Figure A2.6. Input-output relationships for the R, G, B channels after gamma correction. ................................................................................................................. 158

Figure A2.7. Luminance (cd/m2) and colorimetric values for the monitor at various locations ................................................................................................................... 159

Figure A2.8. Oscilloscope traces for a pseudorandom sequence of mfERG stimulus for both CRT and LCD systems .............................................................................. 161

Figure A2.9. Spectral radiance distribution of the Red (solid), Green (dashed) and Blue (dotted) channels of the LCD display .............................................................. 162

Contents

xi

Figure A2.10. Measured chromaticities of red, green and blue primaries of the Viewsonic monitor in CIE 1931 system (A) and CIE 1964 system (B) .................. 163

Figure A2.11. Channel independence for a channel at two output levels 10% and 90% of maximal output when the other two channels are at 0% and 50 % of maximal output ...................................................................................................................... 165

Figure A2.12. Measured spectral output of individual and combined channels at the maximum operating levels, that is, 98 % of the maximum ...................................... 166

Figure A2.13. Spectral radiance measurements of red, green and blue channels as a function of wavelength (nm) at operating levels 90% and 50% .............................. 168

Figure A2.14. Chromaticity coordinates (u, v) and differences in u and v from the 90 % values are plotted against channel operating levels for each channel.................. 169

Figure A2.15. Calculated colour difference (C*uv) plotted against operating level for each channel. ............................................................................................................ 170

Contents

xii

Table of Tables

Table 3.1. Combinations of ND filters inserted in front of the auxiliary source to match its brightness to S1 ........................................................................................... 37

Table 3.2. SCE measurements of one subject from the main experiment with maximum range of aperture movement using method I. ........................................... 50

Table 3.3. SCE measurements of one subject from the main experiment with maximum range of aperture movement using method II. .......................................... 51

Table 3.4. Uncorrected and corrected SCE parameter fits using method I ................ 52

Table 3.5. Similar to Table 3.4, but using method II. ................................................ 53

Table 3.6. SCE directionality (ρx) and peak location (xc) with different sampling intervals for three subjects .......................................................................................... 54

Table 3.7. Mean directionality (ρx, ρy) and peak locations (xc, yc) of two SCE measurements for no contact lens, +5 D contact lens and -5 D contact lens conditions .................................................................................................................................... 55

Table 3.8. Directionality (ρx) and peak locations (xc) of three SCE measurements for no trial lens, +1.25 D trial lens and -1.25 D trial lens conditions .............................. 56

Table 3.9. Directionality and peak locations of two runs of SCE measured using main apparatus for in-focus, 1 D hyperopic defocus and 1 D myopic defocus conditions .................................................................................................................................... 57

Table 3.10. Higher order RMS and spherical aberration of right eye for 6 mm pupil, with and without the contact lenses ............................................................................ 58

Table 3.11. Directionality and peak locations of two runs of SCE using modified apparatus for in-focus, 1 D hyperopic defocus and 1 D myopic defocus conditions. 62

Table 3.12. A comparison of percentage changes in SCE directionality of all conditions for both 6 mm and 5 mm pupil sizes ........................................................ 63

Table 4.1. Means (first entry) and differences (second entry) of parameter fits for two SCE runs at different accommodative stimuli for emmetropes ................................. 76

Table 4.2. Means (first entry) and differences (second entry) of parameter fits for two SCE runs at different accommodative stimuli for myopes ........................................ 82

Table 4.3. Comparison of rates of changes in ρx, ρy, xc and yc with accommodation for emmetropes, myopes and combined data of emmetropes and myopes ................ 85

Contents

xiii

Table 4.4. Mean ± SD of the SCE peak pupil locations obtained from both directions for 0 D (unaccommodated) and 6 D (accommodated) accommodative stimuli ....... 87

Table 5.1. Measurements of retinal illuminance from the mfERG stimulus .......... 108

Table 5.2. N1 and P1 implicit times (ms) of rings 1-4 measured with the LCD and CRT displays. ........................................................................................................... 109

Table 5.3. Comparison of ρ mean (first entry) from N1 and N1P1 amplitudes of the ring average waveforms between 0 D and 6 D accommodation .............................. 117

Table 5.4. Estimated peak locations for 0 D and 6 D accommodation stimuli along horizontal (x) and vertical (y) meridians in the pupil. ............................................. 118

Table A.1 Chromaticity coordinates and maximum luminance for the channels following colorimetric characterization. .................................................................. 155

Statement of authorship

xiv

Statement of original authorship The work contained in this thesis has not been previously submitted to meet the

requirements of an award at this or any other higher education institution. To the best

of my knowledge and belief, the thesis contains no material previously published or

written by another person except where due reference is made.

Signature: ____ ___________________________ Date: _________15.10.2009____________________________

Acknowledgements

xv

Acknowledgements I would like to thank my supervisor Professor David Atchison for his assistance and

support he has extended to me throughout my work towards this thesis, and for his

informed guidance and advice. I also extend my sincere gratitude to my associate

supervisor Dr Sanjeev Kasthurirangan, for his assistance and encouragement. His

patience and willingness to discuss the minutiae of the different obstacles I

encountered while working on this project were invaluable.

My thanks also go to Professor Brian Brown for his guidance in carrying out the

second study using multifocal electroretinogram and also for his help in analysis of

the data. I am also grateful to Dr Andrew Zele for his assistance and advice in

performing the calibration of the LCD monitor.

Thanks must go to my fellow postgrads, labmates and staff at the School of

Optometry, IHBI in QUT. Thank you all for your encouragement and support all the

while. I also express my appreciation to all the participants for lending me their

valuable time and eyes in pursuit of scientific knowledge.

For financial support, I thank QUT for providing me the scholarship and the School

of Optometry for partly supporting my living allowance.

This thesis is dedicated to my family – My dad, whose blessings are always with me,

my mom, brother, sister, sister-in-law, brother-in-law, niece and nephews for their

unending love and support. They have always supported me in my endeavours,

always given me the strength and encouragement to follow my dreams.

Nisha Singh, October 2009.

Chapter 1 Introduction

1

Chapter 1

Introduction

1.1 Background The brightness of light changes with its entry location in the pupil of the human eye,

with the light entering through the periphery appearing less bright than light passing

through the centre. This differential light sensitivity is called the Stiles-Crawford

effect of the first kind (SCE I) after its discoverers, W. S. Stiles and B. H. Crawford

(Stiles & Crawford, 1933). Later Stiles (1937) measured a change in the perceived

hue and saturation with light entry location in the pupil, and termed this phenomenon

the Stiles-Crawford effect of the second kind (SCE II) or chromatic Stiles-Crawford

effect. The research described here relates to SCE I only, and for simplicity SCE I

will be referred to as SCE in this thesis.

Stiles and Crawford (1933) hypothesized that the SCE is of retinal origin and later

studies attributed the apparent luminosity change with pupil-entry location to the

angular sensitivity of the cone photoreceptors (Laties, Liebman & Campbell, 1968;

Wright & Nelson, 1936). It is now accepted that the cones act as optical waveguides

collecting light over a restricted angular extent and directing the incident light along

the outer segments of the photoreceptors (Pask & Stacey, 1998; Toraldo di Francia,

1949).

Investigations have been performed to determine the role of the SCE in visual

performance (Atchison, Joblin & Smith, 1998; Atchison & Scott, 2002a) and to

understand photoreceptor optics (Pask & Stacey, 1998; Snyder & Pask, 1973;

Toraldo di Francia, 1949). The SCE has been identified as having a minor role in


2

improving spatial visual performance because it reduces the effects of defocus and

aberrations to a small extent (Atchison, Joblin & Smith, 1998; Atchison & Scott,

2002b). It is possibly important in reducing the effects of intraretinal scatter on visual

performance.

There have been several studies examining the influence of the SCE on the retinal

point spread function and the modulation transfer function of the eye (Atchison,

Joblin & Smith, 1998; Atchison, Scott, Joblin & Smith, 2001; Atchison, Scott, Strang

& Artal, 2002; Zhang, Ye, Bradley & Thibos, 1999). In these studies, the SCE was

described as an alteration of pupil function, also known as an apodization, in which

the density of a transmission filter increases from its centre toward its edge. The SCE

in single dimension is typically described as a Gaussian distribution

2

max )(max

xxe

where η is brightness sensitivity at a given pupil-entry location, x is the pupil entry

location (in millimetres), xmax is the pupil-entry location with maximum brightness

sensitivity, ηmax is the maximum brightness sensitivity and ρ is the shape factor

(related to steepness of the curve and hence represents the photoreceptors

directionality). When described in ln units, this equation becomes a parabola.

2

maxmax )(lnln xx

Variations in luminance and wavelength influence the SCE. Crawford (1937) found

that foveal SCE directionality was essentially the same at zero and high luminance

background levels, but parafoveal (5°, 14° eccentricity) directionality increased

gradually from low to high luminances. Stiles (1937) noticed that SCE directionality

depended upon stimulus wavelength. The foveal SCE function had greatest

directionality for stimuli of short wavelengths, had slightly less directionality for

long wavelengths, and had much less directionality for mid-spectrum wavelengths.

Stiles (1939) attributed this to higher directional sensitivity of short and long

wavelength-sensitive cones than of the medium wavelength sensitive cones. The

SCE is also affected by retinal pathologies such as retinitis pigmentosa, retinal


3

detachment, macular edema, age related macular degeneration, and central serous

choroidopathy (Dunnewold, 1964; Enoch & Lakshminarayanan, 1991; Enoch &

Tobey, 1981; Fankhauser, Enoch & Cibis, 1961; Smith, Pokorny & Diddie, 1978;

Smith, Pokorny & Diddie, 1988). Shearing effects between the retina and the choroid

due to accommodation (Blank, Provine & Enoch, 1975; Enoch, 1975) and tractional

strains/stress in the retina due to elongation of eyeball in myopic eyes (Choi, Enoch

& Kono, 2004) have been reported to affect the SCE. High accommodation was

reported to cause a nasal shift of the peak position of the SCE in the horizontal

meridian (Blank, Provine & Enoch, 1975; Enoch, 1975). Differences in SCE between

emmetropes and myopes have been reported (Choi, Garner & Enoch, 2003a). No

study has evaluated changes in SCE directionality with accommodation, either for

emmetropes or myopes. The research described here seeks to understand the

influence of accommodation on the SCE in emmetropic and myopic subjects

(Chapters 3 and 4).

Although the SCE is interesting from a research perspective, its use as a clinical tool

to determine photoreceptor/retinal integrity has been limited because the

conventional, psychophysical method of measuring it is time consuming and needs

excellent co-operation from subjects. A few objective techniques such as

reflectometry and retinal imaging using adaptive optics have been used in recent

years. Another objective technique to measure the SCE using multifocal

electroretinogram (mfERG) proposed by Sutter (1997) has not been explored further.

Its usefulness in terms of measuring the SCE has not been established. Therefore, the

latter part of this research will evaluate accommodation-induced changes on the SCE

with a faster objective technique using the multifocal electroretinogram (Sutter,

1997) (Chapter 5).

1.2 Research objectives Mechanical stresses on the retina due to accommodation (Blank, Provine & Enoch,

1975) and myopia (Choi, Enoch & Kono, 2004) have been reported to influence the

SCE. There is no study in regard to changes in SCE directionality with


4

accommodation. No studies have investigated the interaction between

accommodation and myopia in influencing the SCE. The first objective of this

research is to measure and compare the SCE with accommodation in emmetropes

and myopes.

In recent years there has been a growing interest in developing quick and easy

methods of measuring the SCE (Burns, Wu, Delori & Elsner, 1995; Burns, Wu, He

& Elsner, 1997; Gorrand & Delori, 1995). One small scale study has been done using

the mfERG (Sutter, 1997) and the technique was found to be much quicker in

comparison to the psychophysical technique. The second objective of this research is

to develop a quick objective technique using mfERG to measure the SCE and

compare it with the psychophysical technique.

1.3 Thesis outline Chapter 2 gives an overview of the Stiles-Crawford effect, its relation with

accommodation, and different techniques for measuring it. The experimental work is

presented in Chapters 3 to 6. Chapter 3 describes the psychophysical techniques of

measuring the SCE and investigating its changes with accommodation in

emmetropes and myopes. Methodologies of control and supporting experiments

related to the psychophysical techniques and their results are described in Chapter 3.

Chapter 4 presents and discusses the results of the psychophysical investigations.

Chapter 5 describes an objective technique of measuring the SCE using the

multifocal electroretinogram, and the chapter compares and contrasts the results of

psychophysical and objective techniques. The concluding Chapter 6 has a general

summary and discusses future directions of the research.

Chapter 2 Literature Review

5

Chapter 2

Literature review

2.1 The Stiles-Crawford effect

2.1.1 Mathematical description The peak sensitivity and steepness (or directionality) of the Stiles-Crawford effect

(SCE) are usually obtained from a mathematical fit to a plot of sensitivity (or

threshold) against the pupil-entry location across the horizontal or vertical meridians

(Figure 2.1).

Pupil Entry Location (mm)

-4 -3 -2 -1 0 1 2 3 4

Th

resh

old

(ln

un

its)

5.6

5.8

6.0

6.2

6.4

6.6

6.8

7.0

7.2

7.4

7.6

7.8

NASALTEMPORAL

Figure 2.1. A typical Stiles-Crawford function along the horizontal meridian from this study.

Threshold for seeing a small spot of light is plotted against the pupil entry location. The

continuous curve is the fit to the data.


6

Various mathematical functions have been proposed to describe the SCE. The most

commonly used function is the parabolic function as originally proposed by Stiles

(1937),

2

max10max )(loglog xx (2.1a)

(or)

2maxmax )(lnln xxe (2.1b)

where η is brightness sensitivity at a given pupil-entry location, x is the pupil entry

location (in millimetres), xmax is the pupil-entry location with maximum brightness

sensitivity, ηmax is the maximum brightness sensitivity and is the steepness

parameter (represent photoreceptor directionality) of the SCE.

The coefficients e and 10 are related by

)10ln(10 e (2.2)

Equations 2.1 are sometimes expressed in the equivalent forms,

2

max )(max

xxee

(2.3a)

2

max10)(

max10xx (2.3b)

A limitation of these functions is that they do not provide accurate fits beyond 3mm

from the peak of sensitivity or if data are not symmetrical about the peak. Using the

natural logarithm of the sensitivity, a Gaussian function has been demonstrated to

provide a better fit to the foveal SCE than the parabolic functions of equation (2.1)

when points of more than ±3 mm from the peak of the SCE are included (Safir &

Hyams, 1969; Safir, Hyams & Philpot, 1970). The form of this Gaussian function is

2)(

21ln cxBAeKK (2.4)

Here c is the location of the peak, K2 is the horizontal asymptote, K1 is a constant

that depends on the optics of the apparatus used to measure SCE, and A and B are

parameters of the spread. Functions other than the Gaussian and the parabola have

also been proposed which are mostly truncated sine and cosine functions of the

exponential function.


7

Although the equation 2.4 provides better fits over the entire pupil, the parabolic

function (equation 2.1) or its equivalent Gaussian function (equation 2.3) is used

predominantly in the literature to describe SCE data within 3 mm of the peak

location. Therefore, following convention the parabolic function (equation 2.3a) was

selected for the present studies.

2.1.2 Basis of the SCE The origin of the SCE was assumed to be retinal by Stiles and Crawford (1933)

because the SCE was found to be stronger at photopic levels than at scotopic levels

of retinal illumination (Stiles & Crawford, 1933). Variations in the amplitude of

electroretinogram (ERG) with pupil entry positions are also a strong indicator of the

retinal origin of the SCE (Armington, 1967). Other evidence comes from cases of

retinal pathology, where the optics of the eye is unaffected but the orientation of the

photoreceptors is affected, such as in retinal detachment, after photocoagulation,

scleral buckles, and Fuch’s coloboma (Bailey, Lakshminarayanan & Enoch, 1994;

Fankhauser, Enoch & Cibis, 1961).

Wright and Nelson (1936) considered that photoreceptors function as light guides,

since it is known that the receptor cell membrane has a higher index of refraction

than the internal cell plasma and the extracellular space. This allows for the angular

dependence of trapping of light rays. This explains differences in SCE with

wavelength (since refractive index varies with wavelength). O’Brien (1946; 1951)

further elaborated on this theory by studying the angular dependence of the light rays

that enter the outer segment of the photoreceptors using geometrical optics. Toraldo

di Francia (1949) proposed that each receptor should be treated as a tiny antenna

(waveguide) whose directional behaviour is described by wave optics rather than by

geometrical optics. The waveguide theory of photoreceptor directionality forms the

basis of the modern theories of SCE origin. It is based on the model of a single

photoreceptor comprising inner and outer optical waveguide segments that are

connected by a tapered section called the ellipsoid (Enoch & Lakshminarayanan,


8

1991; Pask & Stacey, 1998; Snyder & Pask, 1973; Stacey & Pask, 1994). Waveguide

models are highly dependent on the refractive indices of the receptors.

2.1.3 Photoreceptor alignment Histological investigations from diverse vertebrate retinas including fish, amphibians

birds and mammals have shown that the photoreceptors are aligned essentially

somewhere near the centre of the pupil (Laties, 1969; Laties, Liebman & Campbell,

1968).

Photoreceptor alignment may be governed by a phototropic mechanism or light-

guided mechanism that actively aligns the retinal receptors toward the pupil of the

eye so as to improve capture of light (Enoch & Lakshminarayanan, 1991). Evidence

for phototropism comes from realignment of receptors towards the exit pupil, after

resolution of pathologies affecting the outer retina (Enoch & Lakshminarayanan,

1991) and from displaced and abnormal pupil studies (Applegate & Bonds, 1981;

Dunnewold, 1964; Enoch & Birch, 1981). Other evidence for phototropism is the re-

alignment of the SCE peak toward the pupil centre in a patient following removal of

a cataract that obscured all but the margin of the pupil on one side (Smallman,

MacLeod & Doyle, 2001).

As well as phototropic mechanisms, factors such as retinal tractional forces may

influence photoreceptor alignment. There is evidence that retinal trauma and certain

pathologies affecting the outer retina such as retinitis pigmentosa, retinal detachment,

macular oedema, age-related maculopathy, central serous choroidopathy and Best’s

disease can cause reduction in directionality and a shift in peak position of the SCE

of more than one mm (Bailey, Lakshminarayanan & Enoch, 1994; Dunnewold, 1964;

Enoch & Lakshminarayanan, 1991; Fankhauser, Enoch & Cibis, 1961;

Lakshminarayanan, Bailey & Enoch, 1997; Smith, Pokorny & Diddie, 1978; Smith,

Pokorny & Diddie, 1988). Saccadic eye movements of 5° amplitude can also shift

the SCE peak by 0.6 mm; this has been suggested to be due to the shearing forces

acting on the retina during rapid eye movements (Richards, 1969). Accommodation


9

has been suggested to affect the photoreceptor alignment due to a shear effect

between the retina and the pigment epithelium or choroid (Blank, Provine & Enoch,

1975; Enoch, 1975) (see section 2.2.1).

2.1.4 Effects of luminance and wavelength The SCE directionality is reduced at scotopic levels compared with photopic

luminance levels (Crawford, 1937; Stiles, 1939). This reduction is attributed to the

change of the functioning receptor type from cone to rod. Crawford (1937)

investigated the effect of luminance on the SCE for both foveal and parafoveal

locations. At the fovea, SCE directionality was unaffected by background luminance

while at 5° from the fovea directionality increases progressively from zero to high

background luminances. At both 5° and 14° parafoveal locations, the curves resulting

from high background luminance approached that of the foveal curves, while the

curves at zero luminance were flat (Crawford, 1937). Because of the difference

between the photopic and scotopic SCEs, Crawford (1937) and Stiles (1939)

suggested that rods have no directional sensitivity. Later, Van Loo and Enoch (1975)

demonstrated the presence of a scotopic SCE using a careful signal detection

procedure at a 6 degree location in the nasal or temporal visual field in three

observers. There was a gradual transition between the small scotopic SCE and the

larger photopic SCE in the mesopic range (Enoch & Lakshminarayanan, 1991).

Stiles (1937, 1939) noticed that the SCE directionality changes systematically with

stimulus colour. Varying the wavelength and keeping the same photopic luminance

for a foveal stimulus, he found that the directionality was highest for short

wavelengths, slightly less for the long wavelengths and much lower for medium

wavelengths. Stiles (1937) suggested that this was because of different directional

properties for each of the three cone types. Enoch and Stiles (1961) reached the same

conclusion using colour matching rather than an incremental threshold technique.

Stiles (1939) also demonstrated the wavelength dependence of the SCE at 5 degrees

from the fovea. This had the least effect at blue and green areas of the spectrum, but

had an increased effect for long-wavelengths.


10

2.1.5 Peak location and directionality The SCE peak is assumed to reflect the principal alignment tendencies of the

photoreceptors in the tested area (Enoch & Bedell, 1981; Enoch & Hope, 1972b;

Laties, 1969; Laties, Liebman & Campbell, 1968). The SCE ρ is assumed to reflect

the directionality (or angular tuning) of the population of photoreceptors being tested

(Crawford, 1937; Safir & Hyams, 1969; Stiles, 1937). Two extended psychophysical

studies provided normative data for foveal SCE peak location and directionality in

photopic conditions. Dunnewold (1964) determined the peak location of the SCE in

18 pairs of normal eyes and 11 single eyes of observers whose untested eye had a

disease. The mean and standard deviations for the SCE peak locations were

0.37±0.78 mm nasally and 0.29±0.80 mm superiorly from the centre of the dilated

pupil. Applegate and Lakshminarayanan (1993) generated both horizontal and

vertical data sets from 49 eyes. Their means and standard deviations for the SCE

peak locations were 0.51±0.72 mm nasal and 0.20±0.64 mm superior relative to the

first Purkinje image. The mean values and standard deviations for ρ10 were

0.047±0.013 mm-2 horizontally and 0.053±0.012 mm-2 vertically. These values were

based on equation (2.1a), and converting the ρ10 values to ρe values according to

equation 2.2 gives 0.108±0.030 mm-2 and 0.122±0.028 mm-2 along horizontal and

vertical meridians, respectively.

The photoreceptor directionality increases generally as a function of retinal

eccentricity (Bedell & Enoch, 1979; Enoch & Hope, 1973; Stiles, 1939). Stiles

(1939) found higher directionality parafoveally at 5° than foveally. Enoch and Hope

(1973) found that SCE directionality increased gradually from zero to 2° eccentricity,

but the directionality was relatively stable between 2 and 10°. Bedell and Enoch

(1979) found that the SCE directionality at 35° in the temporal retina was similar to

that at the fovea. Choi, Garner & Enoch (2003a) reported lower directionality at the

fovea than at 5º–15º in the peripheral temporal retina and at 5º-10º in the nasal retina

in different refractive error groups (emmetropes, moderate myopes, and hyperopes).

However, high myopes had smaller directionality at the fovea and 10º nasal retina

than at 5º in the nasal retina and at 5º–15º in the temporal retina. Generally, the

emmetropes had greater SCE directionality than myopes across the retina, suggesting


11

that the receptors are better aligned with each other in emmetropes (Choi, Garner &

Enoch, 2003a).

2.1.6 Measurement techniques Various techniques have been used to determine the SCE, but the main technique

since the discovery of the SCE in 1933 has been the psychophysical (subjective)

technique. Specifically, the measurement of the SCE involves relative luminous

efficiency measurements at a number of positions across the pupil. A few objective

techniques such as the electrophysiological technique, photometric matching, flicker

photometry and fundus reflectometry have been employed to measure the SCE

(Armington, 1967; Bailey & Heath, 1978; Burns, Wu, He & Elsner, 1997; Sternheim

& Riggs, 1968; Sutter, 1997).

The following sections provide an overview of a few techniques to measure the

Stiles-Crawford effect.

2.1.6.1 Psychophysics To investigate variations in luminous efficiency as a function of pupillary entry

location, Stiles & Crawford (1933) designed a Maxwellian-view system to project

two small light beams at different pupillary locations. One beam entered the pupil

centrally and the other beam was presented at discrete positions across either the

horizontal or vertical meridians. Subjects were required to match the brightness of

the two beams based on an increment threshold procedure. This psychophysical

technique, although with some variations, has been the main method of measuring

the SCE. This technique requires significant co-operation and concentration from the

subject and is time consuming (up to one hour of testing). The version of the

technique used in this work is described in detail in Chapter 3.


12

2.1.6.2 Electrophysiological measurement The conventional (full-field) electroretinogram (ERG) is a measure of the gross

potential of the whole retina to a flash of light and thus reflects the activity of all the

cells in the retina (Heckenlively & Arden, 2006; Lam, 2005). The resulting

waveform is characterised by the amplitude, latency, and duration of its individual

components. The ERG is recorded at the cornea and is generated by the radial

currents that arise either directly from retinal neurons or by extracellular movements

of potassium [K+] and sodium [Na+] ions from the activity of retinal neurons

(Heckenlively & Arden, 2006; Lam, 2005). The four major cell types contributing to

the ERG waveform are retinal pigment epithelium cells, photoreceptors, bipolar cells

and Müller cells.

The first human ERG waveform was obtained by Kahn & Lowenstein (1908). Granit

(1947) proposed an improved interpretation of the ERG waveform that is still in use.

He interpreted ERG as a combination of three processes (or potentials) that he called

P I, P II and P III, named for the sequence of disappearance under ether anaesthesia.

This classification has been widely accepted. This model of Granit was revised by

Rodieck (1972) and incorporated new features reflecting improved understanding of

the anatomy of photoreceptors and their function. The waveforms are usually a

summation of the a-wave (photoreceptors) and b-wave (ON bipolar cells and Müller

cells) (Figure 2.2) (Heckenlively & Arden, 2006; Lam, 2005). The waveform can be

characterised by the amplitude, latency and duration of its individual components.

Figure 2.2. Waveform representation of a full-field ERG. Amplitudes and implicit times of the

waveform are labelled.

a-wave amplitude

a-wave implicit time

b-wave amplitude

b-wave implicit time

Time (ms)

a-wave amplitude


b-wave amplitude


a-wave amplitude


b-wave amplitude


Time (ms)


13

In ERG waveform analysis, it is common to measure implicit times and amplitude.

Implicit time is the measured time from flash onset to the trough of the a-wave or

from flash onset to the peak of the b-wave. The implicit time of the a-wave provides

a valuable index in evaluating retinal degeneration arising from photoreceptor

dysfunction (Lam, 2005).

The amplitude of the a-wave is measured from the baseline resting potential to the

trough of the negative deflection. This amplitude gives information about

photoreceptor transduction and is dependent upon both the functional integrity of the

photoreceptors and their ability to absorb light quanta (Breton, Schueller, Lamb &

Pugh, 1994; Perlman, Gdal-On, Miller & Zonis, 1985). The amplitude of the b-wave

of the ERG is measured from the negative trough of the a-wave to the positive peak

of the b-wave. The b-wave amplitude will thus vary as a function of the amplitude of

the a-wave and is a result of the light-evoked depolarization of the ON bipolar

neurons (Stockton & Slaughter, 1989).

The conventional (full-field) ERG is a well established and widely used test of retinal

function, but responses from local areas cannot be obtained with it. The recently

developed multi-stimulus electroretinogram, or multifocal electroretinogram

(mfERG), allows simultaneous stimulation and recording of multiple retinal areas.

MfERG was introduced by Sutter & Tran (1992) to explore the field topography of

the ERG components in response to photopic local luminance modulation. The

individual hexagons of the stimulus are turned on and off according to the

“pseudorandom m-sequence” flicker stimulation technique (fast-sequence mfERG)

in which each hexagon goes through a pseudo-random sequence of black-and-white

presentations with a 0.5 probability of being either black or white. This allows fast

simultaneous recording of many local responses from the retina (Sutter & Tran,

1992).

The stimulus pattern usually consists of black and white hexagons presented on a

CRT display (Figure 2.3). Arrays of 61 or 103 hexagons covering the central 20° to

50° of the retina are commonly used. Usually the hexagons increase in size with


14

increase in eccentricity so that the local responses are of approximately equal

amplitude in control subjects (central hexagon covers about 1°-2° and the most

peripheral hexagons are about 4 to 5 times larger) (Sutter & Tran, 1992). However,

patterns of equal size hexagons can be generated and may be useful in special cases

such as patients with eccentric fixation or when using specialized flicker sequences.

The choice to vary the pattern depends on balancing the need for spatial resolution,

signal-to-noise ratio, and length of recording (Poloschek & Sutter, 2002). Higher

resolution patterns require longer recording times which might not be suitable for

clinical use (Heinemann-Vernaleken, Palmowski, Allgayer & Ruprecht, 2001).

In the VERIS (Visual-Evoked Response Imaging System) software (EDI, Redwood

City, CA), each hexagon starts at a different point in the m-sequence which allows

analysis of the responses with a fast algorithm (Sutter & Tran, 1992). The contrast

between the lightened and darkened stimulus elements should be 90 % or greater.

The background area of the monitor that is, outside the area of stimulus hexagons

should have a luminance equal to the mean luminance of the stimulus array

(Marmor, Hood, Keating, Kondo, Seeliger & Miyake, 2003). The stimulus array is

generated as an achromatic (black-and-white or colour-and-background) flicker

changing with each frame of the display, usually every 13.3 ms at a frame rate of 75

Hz, by means of a customised Macintosh video card (EDI, Redwood City, CA).

More sophisticated modes to display the stimulus pattern have been introduced such

as digital stimulus delivery methods (LCD) (Keating, Parks, Malloch & Evans,

2001), mainly to avoid artefacts produced by CRT monitors such as high temporal

frequency and luminance artefacts (García-Pérez & Peli, 2001; Zele & Vingrys,

2005).

A central fixation dot or cross is available with most stimulus programs for subjects

to maintain stable fixation. For clinical use, the subject’s viewing distance is

generally 32 cm or 37 cm. The retinal signals are recorded with corneal contact lens

electrodes or thread (Dawson-Trick-Litzkow (DTL)) electrodes. The contact lens

electrodes give larger amplitudes than the DTL electrodes but are less comfortable

(Coupland, 1991; Esakowitz, Kriss & Shawkat, 1993). Skin cup electrodes on the


15

forehead and at the lateral canthus serve as ground and reference electrodes,

respectively.

Figure 2.3. Sample of a VERIS 61 hexagonal stimulus array. The individual hexagons are

stimulated to alternate between black and white according to the pseudo-random binary m-

sequence.

The International Society of Clinical Electrophysiology and Vision (ISCEV) has

suggested guidelines for basic cone-mediated mfERG recording to make data

comparable between different studies (Marmor, Hood, Keating, Kondo, Seeliger &

Miyake, 2003). Several technical aspects with regards to stimulus delivery,

electrodes, amplifiers, filter bandwidth as well as interference and noise reduction

have been reviewed by Keating, Parks & Evans (2000).

Unlike the full–field ERG, the mfERG is not the directly recorded responses, but a

mathematical extraction of recorded signals. The mfERG allows determination of

first-, second- and higher-order responses, also known as kernels. For the SCE,

which is a function of cone photoreceptors, only first-order responses are of interest.

The first-order response is a biphasic wave with an initial negative component (N1)

followed by a positive component (P1). There may be a second negative component

(N2) after the positive component (P1) (Figure 2.4). N1 and P1 latencies are the time

measurements taken from the onset of the stimulus to the N1 trough and P1 peak,

respectively. The N1-amplitude is measured from the baseline to the N1 trough


16

whereas the P1-amplitude (N1P1) is measured from the N1 trough to the P1 peak.

This first-order response is derived from the average retinal response following a

white frame (or flash response) and then subtracting all responses following a black

frame. It is thought to represent activity from the outer to middle retinal layers,

especially bipolar cells and photoreceptors (Hood, 2000; Hood, Frishman, Saszik &

Viswanathan, 2002). By slowing the stimulus sequence (below 75 Hz), the biphasic

mfERG waveform can be compared with the photopic a- and b-waves of the full-

field ERG (Hood, Seiple, Holopigian & Greenstein, 1997). The first-order response

is assumed to be a linear response but it can also contain some non-linearities due to

inner retinal cell contributions and lateral interaction (Sutter & Tran, 1992). These

non-linearities can be better extracted from the second-order response. The second-

order response measures how the mfERG response is influenced by the adaptation to

successive flashes and is considered to be a measure of the outer plexiform layer

response (Hood, 2000).

Figure 2.4. Waveform representation of a first order mfERG response. Amplitudes and implicit

times of the waveform are labelled.

Hood, Frishman, Saszik & Viswanathan (2002) performed an in vivo study on

primates to derive the origins of mfERG waveforms by pharmacologically blocking

particular cells and circuits in the retina. They found that the descending arm of N1 is

mainly related to the onset of the off-bioplar cells with a small contribution from the

N1

P1

N2

N1

P1

N2

N1-implicit time

N1-

amp

litud

e

P1-implicit time

N1

P1-

am

plit

ude

Descending arm of N

1

Asc

endi

ng a

rm o

f P1

0 80 ms

20 nVN1

P1

N2

N1

P1

N2

N1-implicit time

N1-

amp

litud

e

P1-implicit time

N1

P1-

am

plit

ude

Descending arm of N

1

Asc

endi

ng a

rm o

f P1

N1-implicit time

N1-

amp

litud

e

P1-implicit time

N1

P1-

am

plit

ude

Descending arm of N

1

Asc

endi

ng a

rm o

f P1

0 80 ms

20 nV

0 80 ms

20 nV


17

cone photoreceptors and the ascending arm of P1 is related to on- and off-bipolar

cells as well as a small contribution from the cones (see Figure 2.4). Under photopic

conditions, it is believed that the rod system does not provide any contribution to

mfERG responses, as it is suppressed by the high frequency and high luminance

stimulus (Hood, Frishman, Saszik & Viswanathan, 2002; Keating, Parks & Evans,

2000). Since the mfERG is predominantly generated by bipolar cells, whose

contribution must be driven by photoreceptors, only damage to the cones or bipolar

cells greatly decreases mfERG amplitude (Hood, Odel, Chen & Winn, 2003).

Damage to retinal ganglion cells or amacrine cells has only slight effects on the

waveform (Fortune, Johnson & Cioffi, 2001; Hood, Greenstein, Holopigian, Bauer,

Firoz, Liebmann, Odel & Ritch, 2000), and the contribution of ganglion cells to the

mfERG response is revealed under specific conditions such as by chemically

blocking the action potential of ganglion cells (Hood, 2000; Sutter & Bearse, 1999).

The ERG is very useful in assessing a range of ocular conditions (Feigl, Brown,

Lovie-Kitchin & Swann, 2004; Feigl & Zele, 2008). The equipment can also be

modified to measure the SCE by imaging the stimulus on the pupil rather than on the

retina to allow objective measurement of the SCE. Armington (1967) was the first to

use the ERG to measure the SCE, with a grid of alternating black and white stripes

presented in Maxwellian view to elicit electrical responses from retinal receptors. He

plotted the reciprocal of the relative luminances required to produce response

amplitudes of 6.5 µv as a function of pupil entry location. The peak sensitivity was

near the centre of the pupil. The ERG and psychophysical functions of individual

subjects were similar in appearance. Sutter (1997) developed the technique further

using the mfERG and like Armington imaged the stimulus at the pupil. This gave

him 103 simultaneous pupillary locations to test the central 20 or 50 degrees

diameter of the central retina in 8 minutes of testing. This method induced less

fatigue than the psychophysical method. Beresford, Crewther & Crewther (1999)

used mfERG to determine SCE in vivo in chicken eyes.


18

2.1.6.3 Fundus reflectometry technique including Adaptive optics ophthalmoscopy Most of the light that reaches the fundus is absorbed in either the retina or the

pigment epithelium and little is reflected from the retina and back out of the pupil.

Some of the light is incident on the photoreceptor apertures and reflected back down

the axis of the photoreceptor, while other light that misses the photoreceptors is

reflected back from the extracellular space between the photoreceptors. The reflected

light that misses the photoreceptors is scattered back diffusely and uniformly fills the

pupil, while light reflected down the axis of the photoreceptors is directed towards

the centre of the pupil (van Blockland and van Norren, 1986). Thus, light is

distributed in the pupil plane in accordance with the waveguide properties of the

receptors (van de Kraats, Berendschot & van Norren, 1996). Each photoreceptor

guides light toward a common point near the centre of the pupil, resulting in a

roughly Gaussian-shaped light distribution, which is known as the optical Stiles-

Crawford effect (OSCE) (Burns, Wu, Delori & Elsner, 1995; Gorrand & Delori,

1997; Gorrand & Delori, 1995). For the reflectometrically-determined SCE, the

distribution of light in the plane of the pupil can be described as the sum of two

components 2

max )( xxAeB (2.5)

where B is the intensity of a diffuse component (independent of the pupil-entry

position) and A is the maximum intensity of the directional component (Figure 2.5).

The directionality parameter of the OSCE is about twice that obtained with

psychophysical SCE techniques because the light has double-passed through the

waveguides (Burns, Wu, Delori & Elsner, 1995; Gorrand & Delori, 1995; He,

Marcos & Burns, 1999). Chen & Makous (1989) showed that most of the light that

enters the margin of the pupil and is absorbed by the cones has previously passed

through other cones. Such recaptured light may have a larger contribution on

absorption than on reflection, which would result in a broadening of the

psychophysical SCE with respect to the optical SCE. This was demonstrated

theoretically by Vohnsen, Iglesias and Artal (2005).


19

Figure 2.5. Schematic diagram explaining the optical Stiles-Crawford effect (OSCE). [Modified

from Burns, Wu, Delori & Elsner (1995)].

The OSCE has been studied using a variety of objective reflectometers. Krauskopf

(1965) introduced the reflectometry technique to measure photoreceptor alignment

by flood illuminating the retina, and this was further refined by van Blokland & van

Norren (1986) and Gorrand & Delori (1995). Burns, Wu, He & Elsner (1997)

improved the design to obtain the SCE by illuminating the retina with well-

controlled pupil entry locations. Delint, Berendschot & van Norren (1997) performed

fundus reflectometry with a scanning laser ophthalmoscope and Zagers, van de

Kraats, Berendschot & van Norren (2002) measured the optical SCE across the

visible spectrum using an imaging spectrograph.

In vivo imaging of the human retina using adaptive optics has gained popularity in

recent times. The first demonstration of adaptive optics correction with retinal

images was given by Liang, Williams & Miller (1997). Adaptive optics corrects

higher order aberrations of the human eye and improves the retinal image to the

extent that it is possible to visualize photoreceptors. The adaptive optics system

consists of a wavefront sensor which measures the eye’s aberrations and a

deformable mirror which is used to correct the aberrations. Roorda and Williams

(1999) determined the arrangement of S, M and L cones in the living human eye

using adaptive optics. Roorda and colleagues used scanning laser ophthalmoscopy

with adaptive optics for in vivo imaging of the photoreceptors, nerve fibres and flow

Illumination light

Guided componentDiffuse component

Photoreceptor Illumination light


Photoreceptor


Photoreceptor


20

of white blood cells in retinal capillaries (Roorda, Romero-Borja, Donnelly,

Queener, Hebert & Campbell, 2002).

Roorda & Williams (2002) used an adaptive optics ophthalmoscope to measure the

angular tuning or directionality of the cones in the living human eye. They imaged

the retina at 1 degree nasal from the fovea with seven different entrance beam

locations and measured the averaged reflected intensity of each cone. This was based

on the fact that the amount of light scattered should be equivalent to the amount of

light coupled into the cone from illumination light. The ρ values obtained by their

technique were twice those of the psychophysical technique by Applegate and

Lakshminarayanan (1993) but similar to other OSCE determinations (Burns, Wu,

Delori & Elsner, 1995; Gorrand & Delori, 1995). Recently, the OSCE has been

measured using adaptive optics with optical coherence tomography. Gao, Cense,

Zhang, Jonnal & Miller (2008) measured the directionality of several layers of the

retina and found that the directionality for the photoreceptor inner/outer segments

junction was correlated to the psychophysical SCE measurement, whereas that for

the posterior tip of the outer segment of the photoreceptors was consistent with

optical SCE measurements. Reflections from the retinal pigment epithelium were

highly insensitive to beam entry position.

Thus, a similar technique based on the adaptive optics technique of Roorda and

Williams (2002) has been proposed to measure the SCE directionality and orientation

with accommodation (see section 6.3 for details).

2.2 Accommodation and the Stiles-Crawford effect

Accommodation is the ability of the eye to alter its power in order to see objects at

different distances clearly. The farthest and closest object points along the range of

clear vision are called the far and near points, respectively. The difference between

the vergences corresponding to the far and near points is called the amplitude of


21

accommodation (Donders, 1864; Duane, 1912). Many individuals do not

accommodate accurately to bring a near target into focus and may also over-

accommodate to distant targets. The under-accommodation is referred to as a lag of

accommodation and the over-accommodation is referred to as a lead of

accommodation (Rosenfield, 2009)

2.2.1 Mechanism of accommodation and retinal deformation during accommodation Although various theories for the mechanism of accommodation have been proposed,

the Helmholtz-Fincham model of accommodation is widely accepted. According to

this model, the ciliary body contracts in response to a near stimulus. Tension on the

zonular fibres is released and subsequently the crystalline lens adopts a more curved

shape and centre thickness is increased under the influence of the elastic lens

capsule. This results in an increase in the powers of the lens and of the eye (Fincham,

1937).

During accommodation, ciliary muscle contraction pulls the anterior margin of the

retina (ora serrata) and stretches the retina anteriorly by almost 0.05 mm per dioptre

of accommodation stimuli (Moses, 1987). Enoch (1973) calculated that the increase

in retinal area was 30 mm2 for ten dioptres of accommodation (2.4 % area increase)

caused by the accommodation induced forward retinal translation. Hollins (1974)

used a Maxwellian view apparatus to show that the central region of the retina also

stretches by approximately 4.5 % during high (9D) accommodation demand. From

the nasal shift in the peak of the SCE, Blank, Provine & Enoch (1975) suggested that

the anterior stretch in the retina caused a shear effect between the retinal receptors

and the underlying interdigitating microfibrils of the pigment epithelium or between

the retina and the choroid. Accommodation induced retinal stretch has also been

estimated by measuring the retinal movements at the temporal retinal margin using a

transcleral illumination technique and by determining the shift in the blind-

spot/fovea position using a modified perimetry test. An average retinal stretch of


22

0.07 mm/D anteriorly along the sclera was found (Enoch, Moses, Nygaard & Allen,

1983).

2.2.2 Characteristics of accommodation An accommodation response can be stimulated by presenting a target at different

distances from the eye. The accommodation response plotted as a function of the

accommodation stimulus yields the stimulus-response function which generally has a

sigmoid shape (Ciuffreda, 1991; Morgan, 1944). In general, an accommodation

stimulus-response curve shows a lead of accommodation at distance, a linear portion

with a slope of less than one and a saturation-point at the maximum amplitude of

accommodation (Charman, 1982; Charman, 1999; Ciuffreda, 1991).

Blur is considered to be the primary drive for the accommodation system (Collins,

1937; Kruger & Pola, 1986; Kruger & Pola, 1987). Even-error cue (blur) in itself

provides information only about the size of the accommodative error, not about its

direction. Odd-error cues like chromatic aberration and spherical aberration, and the

SCE itself have been suggested to provide directional information to guide the

response (Kruger & Pola, 1985). Collins (1937) observed that the accommodation

response shows microfluctuations during steady state viewing of a target. The

characteristics of these microfluctuations have been extensively studied: their

magnitude is small (≤ 0.5 D) and they are dominated by a high-frequency component

(approximately 1.0 – 2.3 Hz) and a low-frequency component (< 0.6 Hz). The high

frequency component is correlated to arterial pulse (Winn, Pugh, Gilmartin &

Owens, 1990), while the low-frequency component is correlated to other aspects of

the cardio-pulmonary system (Collins, Davis & Wood, 1995). The low frequency

fluctuations are suggested to be used to maintain the steady state response of

accommodation because their magnitudes increase when the depth of focus is

increased such as when viewing through small pupils and in low luminance

(Charman & Heron, 1988; Gray, Winn & Gilmartin, 1993).


23

2.2.3 Influence of accommodation on the Stiles-Crawford effect As mentioned in section 2.2.1, the stretch in the retina during accommodation can

cause physiological disruptions. Few studies have been done to investigate stretching

in the retina due to accommodation. Blank & Enoch (1973) used a monocular

bisection technique, in which subjects fixated at the central line of three short vertical

parallel lines separated horizontally. While adjusting the position of the central line,

the subjects bisected the visual space (imaginary line segment) located between the

two peripherally located lines into two subjectively equal angular or spatial

segments. While making the spatial bisection, if a subject set the midline to one side

then this indicated that he overestimated the angular or spatial subtense of the short

side of the bisected visual segment. Using the technique, Blank & Enoch (1973)

demonstrated that marked accommodation induces substantial distortions in

monocular space perception in the horizontal meridian, suggesting horizontal retinal

stretch.

In order to predict the changes in photoreceptor alignment in the foveal region with

accommodation, Blank, Provine & Enoch (1975) measured the SCE peak location

for a zero accommodation stimulus and a high accommodation (9D) stimulus in three

subjects. With increase in the accommodation stimulus, the peak of the SCE shifted

between 0.5 mm to 1.5 mm nasally, but shifted little vertically. They suggested that

the SCE peak shift was due to either a shearing effect between the receptors and

pigment epithelium or between the retina and the choroid caused by the anterior

(forward) stretching during high accommodation (Blank, Provine & Enoch, 1975;

Enoch, 1975). In a pilot study, Atchison (unpublished) measured the SCE as a

function of accommodation in a 22-year-old subject along the horizontal meridian.

The S-C function co-efficient increased with accommodation, approximately

doubling from 0.08 mm-2 at 0 D stimulus to 0.16 mm-2 at 8 D stimulus. Also,

evidence of a slight nasal shift in the peak was observed.


24

There are no other published studies evaluating the changes in the SCE peak with

accommodation and no study has investigated changes in the SCE directionality with

accommodation. Moreover, Blank, Provine & Enoch (1975) did not measure the

accommodation responses. Thus, the research described here seeks to understand the

changes in the SCE directionality and peak location with increase in accommodation

responses.

2.2.4 Measurement of accommodation Accommodation response measurement requires a device that does not interfere

significantly with the viewing conditions. Over the past 50 years many devices have

been developed to quantify the refractive status of the eye, but only a few have a

temporal resolution high enough to study the dynamic characteristics of

accommodation. Open-view autorefractors such as the Shin-Nippon SRW-5000 have

been used to measure static accommodation, but photoretinoscopy is better placed to

measure both static and continuous accommodation under natural viewing conditions

as the instrument is positioned away from the subject (Choi, Weiss, Schaeffel,

Seidemann, Howland, Wilhelm & Wilhelm, 2000; Schaeffel, Wilhelm & Zrenner,

1993; Wolffsohn, Hunt & Gilmartin, 2002).

The first commercial photorefractor using eccentric photorefraction (the TOMEY

ViVA, Fortune Optical, Padova) had limited ability to measure astigmatism and poor

accuracy perhaps because it used a single light source at one eccentricity (Thompson,

Li, Peck, Howland, Counts & Bobier, 1996). In 1997, a new infrared

photoretinoscope with six infrared LED segments arranged at 30°, 90° and 150° and

their opposite orientations 210°, 270° and 330° was proposed by Gekeler, Schaeffel,

Howland & Wattam-Bell (1997).

Another photorefractor named the PowerRefractor (MultiChannel Systems,

Tübingen, Germany) measures refractive error over a range of -8 D to +6 D at 25 Hz

with pupils >3 mm diameter (in practice > 4 mm) and allows continuous as well as

static measurements from a distance of 1 metre (Wolffsohn, Hunt & Gilmartin,


25

2002). It records pupil size in both eyes. It is well suited for measurements in infants

and has similar reliability and accuracy to other autorefractors (Choi, Weiss,

Schaeffel, Seidemann, Howland, Wilhelm & Wilhelm, 2000; Hunt, Wolffsohn &

Gilmartin, 2003). Seidemann and Schaeffel (2003) found that the lag of

accommodation measured by the PowerRefractor at 3D was in the range of published

values measured using the Canon R-1 open-field autorefractor (Abbott, Schmid &

Strang, 1998; McBrien & Millodot, 1986). The PowerRefractor has been used to

study the effect of chromatic aberration on accommodation and the effect of

longitudinal chromatic aberration on emmetropization in chickens (Seidemann &

Schaeffel, 2002). In recent years, the PowerRefractor has been used to measure the

dynamics of the accommodation responses (Kasthurirangan, Vilupuru & Glasser,

2003; Schaeffel, Howland, Weiss & Zrenner, 1993).

The latest model of the PowerRefractor is the PowerRef II of Plusoptix AG

(Nürnberg, Germany). This allows elaborate testing under three modes: the

‘GazeScan’ mode permits measurements and visualizations of the fixation angle

and/or the strabismus angle, the ‘Full Scan’ mode allows binocular full refraction and

measurement of pupil size, and the ‘Dynamic Scan’ mode allows measurements of

temporal changes in pupil size and accommodation (Jainta, Jaschinski & Hoormann,

2004).

In the present study, an important concern is to measure accommodation responses

during experiments. Therefore, considering the effectiveness of the PowerRef II, the

PowerRef II and instruments based on it have been chosen for use in this study to

measure continuous accommodation responses at regular intervals without

interfering with the subjects’ view of the target during the SCE measurements.

2.2.5 Effect of phenylephrine on accommodation

The SCE is usually measured across at least a 6 mm diameter pupil and therefore the

pupil needs to be dilated but with minimal influence on accommodation in this study.

Phenylephrine is a sympathomimetic drug that is used clinically to dilate the pupil


26

without significant cycloplegic effects (Culhane, Winn & Gilmartin, 1999; Ostrin &

Glasser, 2004). 10% phenylephrine is used for treatment of pupillary block glaucoma

and 2.5% phenylephrine is used diagnostically for fundus examination.

Phenylephrine in high doses (2 drops of 10% solution) causes a reduction in the

subjective amplitude of accommodation (Biggs, Alpern & Bennett, 1959) and slows

accommodative response times (Mordi, Tucker & Charman, 1986). The subjective

amplitude of accommodation is also reduced with 2.5% phenylephrine and this

reduction is enhanced with prior instillation of a topical anaesthetic (Mordi, Lyle &

Mousa, 1986). Eyes with heavily pigmented irides tend to show a slightly greater

reduction in amplitude than eyes with light pigment irides (Mordi, Lyle & Mousa,

1986). The effects with 2.5% phenylephrine will have little implications on this study

because subjects are given longer to achieve an accurate response and the subjects

are young enough that the levels of accommodation required are less than the

amplitude available even with phenylephrine. In summary, to dilate the pupil a

mydriatic must be used and 2.5% phenylephrine affects accommodation least of the

available options.

2.3 Myopia and the Stiles-Crawford effect Myopia, sometimes called near-sightedness or short-sightedness, is a refractive

defect of the eye in which light from a distant target is focused in front of the retina

when accommodation is relaxed (Atchison & Smith, 2000). Myopia has a major

global impact, as approximately a billion of the 6 billion people in the world are

myopes.

Industrialized countries in South–East Asia have some of the highest prevalence rates

of myopia worldwide among their urban populations (Dandona, Dandona,

Naduvilath, Srinivas, McCarty & Rao, 1999; Dandona, Dandona, Srinivas, Sahare,

Narsaiah, Munoz, Pokharel & Ellwein, 2002; Fan, Lam, Lam, Lau, Chong, Cheung,

Lai & Chew, 2004; Hosaka, 1988; Lin, Shih, Tsai, Chen, Lee, Hung & Hou, 1999;

Saw, Tong, Chua, Chia, Koh, Tan & Katz, 2005). There are variations in myopia

prevalence and severity between different races and ethnic groups (Saw, Katz,


27

Schein, Chew & Chann, 1996). Wu (2001) noticed ethnic differences in the

prevalence of myopia even after adjusting for education. There is a high prevalence

of myopia in Singapore (79.3%), with Chinese people having higher rates (82.2%)

compared with Indian people (68.7%) and Malays (65.0%) (Wu, Seet, Yap, Saw,

Lim & Chia, 2001). Within a single racial or ethnic group, the prevalence of myopia

varies with occupation (Saw, Katz, Schein, Chew & Chann, 1996).

One of the oldest arguments about myopia is whether it results from hereditary

factors or from environmental influences. Both the environment, including diet and

close work habits and hereditary factors contribute to the development of myopia

(Ciuffreda & Wallis, 1998; Edwards, 1996; Gardiner, 1958; Ong & Ciuffreda, 1995;

Rosenfield & Gilmartin, 1998; Tan, Ng, Lim, Ong, Snodgrass & Saw, 2000; Zadnik,

Satariano, Mutti, Sholtz & Adams, 1994).

Various attempts have been made to arrest myopia progression. Different optical

modes like orthokeratology (Cho, Cheung & Edwards, 2005; Khoo, Chong & Rajan,

1999) and progressive lenses (Gwiazda, Hyman, Hussein, Everett, Norton, Kurtz,

Leske, Manny, Marsh-Tootle & Scheiman, 2003), and drugs such as atropine (Shih,

Chen, Chou, Ho, Lin & Hung, 1999) and pirenzepine (Saw, Gazzard, Au Eong &

Tan, 2002) have been used in different studies, but a safe and effective treatment has

not been found yet.

2.3.1 Myopia and accommodation Several studies investigating accommodative changes in various refractive groups

have shown that myopes tend to underaccommodate to near targets more than

emmetropes (Abbott, Schmid & Strang, 1998; Gwiazda, Thorn, Bauer & Held,

1993). McBrien and Millodot (1986) found that subjects with late-onset myopia

(onset after 15 years of age) have the highest lags of accommodation followed by

those with early-onset myopia (onset before 15 years of age) and then emmetropes.

Myopes have significantly larger depth of subjective focus than emmetropes

(Rosenfield & Araham-Cohen, 1999).


28

The accommodation stimulus-response curves in myopic children (aged between 6

and 18 years) are shallower than those of emmetropic children (Gwiazda, Thorn,

Bauer & Held, 1993). Abbott, Schmid & Strang (1998) replicated Gwiazda et al.’s

study and found similar results in an adult population, with the accuracy of

accommodation response reduced significantly in progressing myopes in comparison

to emmetropes. The age of onset had no significant effect on accommodation

response and they suggested classifying myopia according to the rate of progression

rather than the age of onset. The closed-loop accommodation response after

variations in near fixation period was investigated in emmetropes, early-onset

myopes and late-onset myopes by use of a dynamic tracking infrared optometer

(Culhane & Winn, 1999). Culhane & Winn (1999) demonstrated significantly longer

accommodation response times in late-onset myopes after a sustained near vision

task than for emmetropes and early onset myopes. Also, subjects with late-onset

myopia show larger accommodation fluctuations and reduced performance under

dynamic tasks (Seidel, Gray & Heron, 2003). Rosenfield & Araham-Cohen (1999)

found greater blur threshold for myopes than emmetropes whereas Schmid, Iskander,

Li, Edwards & Lew (2002) found no significant difference in blur threshold between

myopes and emmetropes.

Tractional forces in the retina due to myopia (Choi, Enoch & Kono, 2004) and

accommodation (Blank, Provine & Enoch, 1975) have been found to affect

photoreceptor orientation. So far no study has investigated the relationship between

accommodation and myopia in regards to their influence on the SCE. This research

will determine if accommodation influences the SCE in myopes differently from that

in emmetropes.

2.3.2 Myopia induced changes to the retina Various factors contribute to the refraction of the eye, including the axial length and

the refractive powers of the cornea and lens. Axial length is the dominating factor in

myopia, since it has been consistently shown that myopic eyes are longer than

emmetropic and hypermetropic eyes (Carney, Mainstone & Henderson, 1997;


29

Grosvenor & Scott, 1994; Strang, Schmid & Carney, 1998). The vitreous chamber

elongates as myopia develops (Kinge, Midelfart, Jacobsen & Rystad, 1999). The

mechanical stress thus applied to the posterior sclera may lead to various

pathological anomalies, such as chorioretinal atrophy, myopic crescent and retinal

detachment. In extreme cases, posterior staphylomas may develop. Choi, Enoch &

Kono (2004) suggested that these mechanical forces (both static and transient)

extending across the retina with the origin near the optic nerve head due to

elongation of the eyeball may be the likely cause of transient variations in the

photoreceptor alignment. Fluctuations in photoreceptor alignment have been

demonstrated in longer eyes, whether myopic or emmetropic, by repeatedly testing

three retinal locations (the fovea, and 22° and 27° in the nasal retina) over a period of

time (Choi, Enoch & Kono, 2004).

In emmetropes, the receptors across the retina are aligned towards the centre of the

exit pupil (Enoch & Bedell, 1981; Enoch & Birch, 1981; Enoch & Tobey, 1981;

Laties, Liebman & Campbell, 1968). This is also the case for myopes for the central

and temporal retina, but receptors in the nasal retina point towards the nasal pupil

(Choi, Garner & Enoch, 2003a). Also, in post-photorefractive (myopic) keratectomy

subjects and the myopic eyes of anisometropic subjects, Choi, Garner & Enoch

(2003b) found nasal alignment of photoreceptors in the nasal retina.

2.4 Summary

It is well established that the SCE directionality and peak location can be influenced

by various factors such as luminance, wavelength and mechanical forces (retinal

pathologies, accommodation or myopia). This thesis focuses on measuring the

changes in the SCE directionality and peak location with accommodation and

myopia. Only one study by Blank, Provine & Enoch (1975) has investigated shift in

the peak of the SCE with accommodation, and this was for only one accommodation

stimulus (9D). They found significant shifts of between 0.5 mm and 1.5 mm in three

subjects, possibly due to the stretch in the retina or shear between choroid and retinal


30

layers during accommodation. No study has investigated changes in the SCE

directionality with accommodation. In myopes, the elongation of the eyeball has

been suggested to stretch the retina with mechanical force originating near the optic

nerve head and thus causing transient variations in the photoreceptors alignment and

nasal shift in the photoreceptors orientation in the nasal retina (Choi, Enoch & Kono,

2004; Choi, Garner & Enoch, 2003a; Choi, Garner & Enoch, 2003b). Therefore, the

prime objective of this thesis is to measure changes in the SCE with accommodation

in emmetropes and myopes using a psychophysical technique.

The most popular technique to measure the SCE has been the psychophysical

(subjective) technique, but this technique is time-consuming (~ 1 hour per run) and

requires excellent co-operation from the subjects. In recent years a few quick

objective techniques such as using the electrophysiological technique, photometric

matching, flicker photometry, fundus reflectometry, and retinal imaging using

adaptive optics have been proposed. Sutter (1997) performed a pilot study to

demonstrate the use of mfERG to measure the SCE. He showed that the technique

using mfERG could measure the SCE from 103 pupillary locations testing the central

20°-50° in the retina in a short duration of 8 minutes. However, the technique using

mfERG has not been compared quantitatively with the psychophysical technique.

Thus, the secondary objective of this thesis is to explore the mfERG technique to

measure the SCE and compare it with the classic psychophysical technique.

Chapter 3 SCE using a psychophysical technique - Methods

31

Chapter 3

The Stiles-Crawford effect in emmetropes and myopes using a

psychophysical technique – Methods

3.1 Introduction The objectives of this study are to measure and compare the changes in the SCE with

accommodation in emmetropes and myopes. The study described in this and the next

chapter investigated the changes in the Stiles-Crawford effect (SCE) with

accommodation in emmetropes and myopes using a psychophysical technique. To be

more specific, this study determined the changes in the photoreceptors’ angular

tuning (directionality) and alignment of the photoreceptors (peak sensitivity) with

accommodation. The study involved some control experiments investigating factors

that can influence the SCE measurements.

Only one study (Blank, Provine & Enoch, 1975) has investigated the shift in the SCE

peak sensitivity location with one accommodation stimulus (9D) using a rapid

technique. This study found a significant shift between 0.5 mm – 1.5 mm in nasal

direction in the pupil. Blank et al. suggested that the transient shear effect between

the photoreceptors and the pigment epithelium-choroid during accommodation may

displace the peak of the SCE. A pilot study by Atchison (unpublished) on one subject

demonstrated an increase in directionality of the SCE with 8 D accommodation. The

study presented here provides a more comprehensive investigation of changes in the

SCE with accommodation.


32

The tractional forces acting in the retina of myopic eyes have been found to produce

some transient changes in photoreceptors’ orientation (Choi, Enoch & Kono, 2004;

Choi, Garner & Enoch, 2003a; Choi, Garner & Enoch, 2003b). I was interested to

determine whether photoreceptors could be further affected in myopic eyes during

accommodation. To investigate the interaction between accommodation and myopia

in influencing the directionality and alignment of the photoreceptors, the SCE was

measured for different accommodation levels in myopes as well as in emmetropes.

The SCE was measured using the classical psychophysical technique for four levels

of accommodation (0, 2, 4 and 6 D) in 6 emmetropes and 6 myopes. In addition, an

experiment was performed using a similar technique to that of Blank, Provine &

Enoch (1975) to determine peak sensitivity.

The methodologies using the psychophysical technique are described in this chapter

and the results are presented in chapter 4.

3.2 Subjects Twelve subjects (6 emmetropes and 6 myopes) with normal visual acuity and

oculomotor functions were recruited for the study. The subjects ranged in age from

16-30 years old (mean ± SD: 23 ± 3 years). An additional 52 year old, 2.25 D

myopic subject was used for some of the pilot investigations. Subjects were students

and staff of the Queensland University of Technology, Brisbane. Research was

approved by the Queensland University of Technology Human Research Ethics

Committee.

Each subject underwent a short optometric examination containing a brief history,

visual acuity, refraction, and slit-lamp evaluation. Each subject had 6/6 Snellen

distance visual acuity and had no ocular disease or surgery. For the emmetropic

group, the refractive corrections of right (tested) eyes were within ± 0.50 DS with

astigmatism < 0.75 DC. For the myopic group, refractive corrections of right eyes

ranged from -2.00 DS to -6.50 DS (mean: -5.00 D ± 1.50 D) and astigmatism ranged

from -0.25 DC to -2.25 DC. Refraction was measured in both eyes using an infrared


33

autorefractor (SRW-5000; Shin-Nippon, Tokyo, Japan) (Mallen, Wolffsohn,

Gilmartin & Tsujimura, 2001). Subjective amplitude of accommodation was

measured with a hand-held Badal optometer and push-up method using a near visual

acuity chart (N8 font). The measured subjective amplitudes of accommodation were

found to be 7 D to 12 D, and this is normal for the age group (Duane, 1912; Turner,

1958). Wide anterior chamber angles (Van Herick ratio of 1:2 or wider) (Van

Herick, Shaffer & Schwartz, 1969) and no history of glaucoma or adverse reactions

to the phenylephrine eye drop was checked before dilatation.

After stimulating accommodation of 6 D the apparatus had a limited range to correct

myopia, so spherical refractions of all myopic subjects were corrected with spherical

soft contact lenses (Optic Zone Diameter (OZD): 9mm up to power -5D; Freshlook

UV, CIBAVISION). A pupil diameter of at least 7 mm was required to measure the

SCE over a 6 mm pupil because the pupil was aligned with respect to the corneal

reflex and the spot that moved across the pupil to measure the SCE was 1 mm in

diameter. The optic zone diameter of the contact lenses were less than 9 mm for

powers more negative than 5 D, and therefore three subjects (JP, YT and RSK) with

myopia more than 5 D were only partly corrected with contact lenses (JP -4.50 D,

YT -3.50 D, RSK -4.50 D). Two subjects (JP and RSK) had astigmatism > 0.75 DC,

and this was corrected by placing cylindrical trial lenses (JP -0.25 D x 180°, RSK -

0.25 D x 160°) midway between lenses L3 and L4 in the apparatus (see Figure 3.1).

Since the effective power at the eye of a lens placed at this position is 4 times its

actual power, only -0.25 D cylindrical trial lenses were used to correct 1.00 D

astigmatism. Subject RSK with astigmatism of -2.25 D x 160° preferred an effective

correction of only -1.00 D x 160° in the set-up.

3.3 Apparatus The SCE was measured with a two-channel Maxwellian view system to image two

light sources at the plane of the entrance pupil (Figure 3.1). This set up is similar to

that described by Applegate and Lakshminarayanan (1993). The sources S1 and S2

are the red channels of diffuse light-emitting diodes (Kingbright LF593MBGMBW,


34

dominant wavelength for a range of standard illuminants approximately 620 nm)

limited by 4 mm and 1 mm diameter apertures, respectively. Beam splitter BS1

combines the two channels.

Figure 3.1. Schematic representation of the two-channel Maxwellian view apparatus. S1, S2 are

light sources limited by 1mm and 4 mm apertures, respectively; L1, L2 are 100 mm focal length

lenses; L3, L4 are 200 mm focal length relay lenses; L5 is a 100 mm focal length Badal lens; BS1,

BS2 are beam splitters; M1 is a front surface mirror; M2, M3, M4 and M5 are reflecting right

angle prisms forming the optical trombone (translator), used to produce defocus or stimulate

accommodation; H is a hot mirror; VC is a video camera; Mo is a monitor. Inset A shows

apertures A1 and A2, aperture A2 is moved as necessary to keep it appearing in the middle of A1.

Inset B shows the subject’s view. S1 and S2 are conjugate with the pupil and A1 and A2 are

conjugate with the retina. (A coloured photograph of the apparatus is shown in Appendix 1).

Eye

Alignment Ring

Inset BS1

S2

L1

L2

A1

A2

BS1 BS2

L3 L4

M2 M5

M3 M4

L5

M1 Optical Trombone

MoVC

PowerRefractor

H

ND Filter

Ba

ckgroundch

annel

Test Channel

0.6° 7°

A2 A1

Inset A

Flashes at 2 HzEye

Alignment Ring

Inset BInset BS1

S2

L1

L2

A1

A2

BS1 BS2

L3 L4

M2 M5

M3 M4

L5

M1 Optical Trombone

MoVC

PowerRefractor

H

ND Filter

Ba

ckgroundch

annel

Test Channel

0.6° 7°

A2 A1

Inset A

Flashes at 2 Hz


35

3.3.1 Background channel The light source S1, limited by a 4 mm diameter aperture, providing a steady

background is imaged in the plane of the entrance pupil by an achromatic collimating

lens L1, a pair of achromatic relay lenses L3 and L4, and achromatic Badal lens L5.

The 12 mm circular field stop A1 conjugate with the retina defines the angular

subtense (7°) of the background field and contains a target to aid fixation (Figure 3.1,

inset A).

3.3.2 Test channel

The light source S2, limited by a 1 mm diameter aperture, is imaged in the plane of

the entrance pupil by lenses L2, L3, L4 and L5. The 1 mm circular field stop A2

conjugate with the pupil defines the angular subtense (0.6°) of the test field (Figure

3.1, inset A). To the subject, the test field appears superimposed centrally on the

background field (Figure 3.1, inset B). The images of S1 and S2 at the pupil plane are

the same sizes as S1 and S2.

S2 is electronically square-wave flickered at 2 Hz. The illuminance control is a 200

Hz pulse width modulation output gated with the 2 MHz HCII E clock of a

computer. This gives bursts of 250 ns pulses every 5 ms, with the illuminance being

varied by the number of these pulses over a 4-log-unit luminance range. The subject

has a control box with a rotating knob to adjust light intensity in 0.05 ln (0.0217 log)

unit steps. The position of S2 is changed with stepper motors under computer control

along horizontal and vertical meridians – this alters the entry point of light into the

eye. Any deviation of the test field from its central location in the background, due to

optical aberrations or residual defocus of the subject’s eye, can be corrected by

adjusting A2 horizontally and vertically (the effect of moving A2 on threshold is

investigated in section 3.8.2).


36

3.3.3 Subject alignment The subject’s head position is fixed with a bite bar mounted on a XYZ movement

controller. An alignment ring containing eight evenly spaced infrared light emitting

diodes is used to illuminate the eye. Beamsplitter BS2 and front surface mirror M1

provide a magnified view of the eye, along with the first Purkinje image (anterior

corneal reflection) of the illumination ring on a monitor Mo via video-camera VC

(Sony CCD-IRIS Black & White) (Figure 3.1). Alignment of a subject’s eye is

achieved by moving the subject’s head in the X, Y and Z directions to maintain the

first Purkinje image in focus and in the centre of a reticule on the monitor. The first

Purkinje image is used as a reference, rather than the geometric centre of the entrance

pupil, because its position is not affected by changes in pupil size and shape due to

accommodation and phenylephrine (used to dilate pupils in the experiments).

3.3.4 Focus setting to stimulate accommodation or to correct refractive errors Stimulus to accommodation is provided by an optical trombone that gives a range of

vergences from +14.0 D to -9.0 D. One pair of prisms (M2 and M5) is stationary

while another pair (M3 and M4) is on a movable translator (Figure 3.1). As the Badal

lens L5 has a 100 mm focal length, 1 mm movement of the optical trombone is

equivalent to 0.2 D change in stimulus. The reference position of the translator was

obtained by adjusting the optical trombone until A1 appeared optimally focused

while viewing it through a telescope, focused for infinity, placed at the subject’s eye

position.

3.3.5 Replacement of 1mm aperture with 4 mm aperture in the background channel In early pilot measurements, the reference source S1 was limited by a 1 mm diameter

aperture as in previous studies (Applegate & Lakshminarayanan, 1993). The


37

accommodative response was low or more variable than expected. This was

attributed to the large depth-of-focus provided by such a small aperture. The 1 mm

aperture was replaced with a 4 mm aperture to decrease the depth of focus. The

accommodative stimulus-response functions with both apertures are compared in

section 4.1.2.

Changing the aperture size of S1 altered the background retinal illuminance. In order

to equalize retinal illuminance for 1 mm and 4 mm apertures, a neutral density filter

was placed in front of source S1 when used with the 4 mm aperture. The required

neutral density filter was identified by a subjective brightness matching test. Three

observers compared the brightness of S1 (1 mm aperture) with an auxiliary light

source (Quartz-diode lamp with a 620 nm interference filter) placed to the side of the

apparatus. An aperture stop was placed in front of the auxiliary light source to

produce a 5 degree field at 50 cm. Combinations of ND filters were inserted in front

of the auxiliary light source to make the match (Table 3.1). The luminance of the

auxiliary light source (with the selected 3.5 ND filter) was measured with a Topcon

BM-7 luminance colorimeter to be 11.1 cd/m2. The brightness of the auxiliary light

source was matched to S1 (4 mm aperture) by placing additional ND filters in front of

S1. A 0.9 ND filter was the middle of the matching range for the three observers

(Table 3.1), and this filter was retained in subsequent experiments.

Table 3.1. Combinations of ND filters inserted in front of the auxiliary source to match its

brightness to S1 with 1 mm aperture, and combinations of ND filters inserted in front of S1 with

4 mm aperture to match its brightness to the auxiliary light source with 3.5 ND filter.

Observer ND filter match

(1mm aperture) ND filter match (4mm aperture)

DAA 3.40 to 3.60 0.80 to 1.00

SK 3.40 to 3.60 0.75 to 1.05

NS 3.35 to 3.65 0.80 to 1.00


38

3.4 Procedure At the start of each experimental session, alignment of the apparatus was checked by

passing a laser beam through the centres of all the optical elements. The position of

the test source S2 was calibrated each time with respect to the reference source S1 by

aligning their images on a reticule on the monitor Mo. This was achieved by placing

a sheet at the position of a subject’s eye and viewing the images of S1 and S2 on the

the monitor. The reticule was moved on the monitor to align it with the image of S1.

If the image of S2 was not aligned with S1, S2 was re-aligned manually. Mirror M1

was altered so that the images through it were also aligned with the images on the

sheet.

At the first visit, each subject was shown the apparatus, a diagram of the subject’s

view (see Figure 3.1) and given a brief description of the study. The subject’s right

eye was dilated with a 2.5 % phenylephrine eye drop approximately 35 - 40 minutes

before the experiment. For subjects with dark irides, 0.4 % benoxinate was also used

prior to phenylephrine to facilitate rapid penetration and longer action of the drug.

Additional drops of 2.5% phenylephrine were instilled if the pupil size was less than

7 mm. All through the experiment the pupil alignment was monitored by the video

camera and the pupil size was measured with the PowerRef II. A pupil diameter of at

least 7 mm size was required to measure the SCE over a 6 mm pupil because the test

spot image at the pupil has a diameter of 1 mm. Once pupil diameter was larger than

7 mm, as measured with the PowerRef II, a contact lens was inserted in the myopic

subjects’ test eyes.

The subject’s dental impression was moulded on a bite bar. Once the mould was set

and the subject was comfortable, the bite bar was attached in the system. The subject

was asked to bite the bite bar and a preliminary alignment was made by adjusting

XYZ positions until the subject could see the background target and the corneal

reflex was in clear focus on the monitor. The room illumination was lowered during

the experiment to avoid any distractions. The subject was asked to maintain clear

fixation at the centre of the cross hair of the background target. The subject was

given a response box containing a dial to control the luminance of the test spot and a


39

button to indicate the brightness match. At each pupil-entry location the subject’s

task was to increase the luminance of the test spot (if necessary) until it was easily

seen and then to slowly reduce the luminance of the test spot until it just disappeared

into the background (descending method of adjustment) and to push the button.

The pupil was sampled across 25 points in 0.50 mm steps with 13 points each along

horizontal and vertical meridians (1-dimensional) and duration per run was

approximately 30 minutes. For one subject (NS), the pupil was sampled over 49

points in a grid fashion (2-dimensional) in 0.75 mm intervals across a 6 mm pupil

and the duration per run was approximately one hour. At each position, 3

measurements were taken. For ease of analysis, 2-dimensional data of subject NS

were combined with 1-dimensional data of other subjects.

For the SCE measured along horizontal and vertical meridians, the test spot was

driven in a sequence from the nasal to the temporal side in the horizontal meridian

and superior to inferior direction in the vertical meridian across the pupil (Figure

3.2A). For the SCE measured at 49 points, the test spot moved across the pupil from

the nasal to the temporal side (Figure 3.2B).


40

Figure 3.2. Movement of the test spot for SCE measurement. (A) SCE measured along

horizontal and vertical meridians; arrows and numbers indicate direction and sequence of the

test spot movement (B) SCE measured over 49 point / grid. Circles are the number of the pupil

entry point tested. N, T, S and I indicate nasal, temporal, superior and inferior positions in the

pupil.

A randomized manner of pupil sampling would be preferable to decrease any bias

associated with sequential sampling, but the stepper motor could not be moved

quickly enough to achieve this.

The SCE was measured for 0, 2, 4 and 6 D accommodative stimuli. As one subject

(LS) from the emmetropic group and one subject EK from the myopic group could

accommodate for a 8D stimulus, this stimulus amplitude was used also for these two

subjects. Two runs were done using the accommodative stimuli, in the order 0, 4, 2

and 6 D. Sometimes the order could not be followed either due to a small pupil size

or subject inability to accommodate to the required stimulus level. The runs were

done over a period of a few days, with usually at least two runs performed in a day.

At each stimulus demand, accommodation was measured eight times for all subjects

except subject NS for whom accommodation was measured fourteen times: before

commencing a run, after every fourth pupil entry position, and after completing the

run. The PowerRefractor was used in ‘dynamic mode’ that allows measurement of

temporal changes in accommodation. Each time in a run, the accommodation

I

Pupil

19

18

17

16

12

11

109

8

7

25 23 22 21 20

15

14

13 6 5 4 3 2 1 49 46 40 33 25 17 10 4 1

47 41 34 26 18 11 5

48 42 35 27 19 12 6

43 36 28 20 13

29

45 39 32 24 16 9 3

44 38 31 23 15 8 2

37 30 22 14 7

21

24 NT NT

I

S

(A) (B)

S

I

Pupil

19

18

17

16

12

11

1099

88

7

2525 2323 22 21 20

15

14

13 66 55 4 33 2 11 4949 4646 4040 3333 2525 1717 10 4 11

4747 4141 3434 2626 1818 11 5

4848 4242 3535 2727 1919 1212 6

4343 3636 2828 2020 1313

2929

4545 3939 3232 2424 1616 9 3

4444 3838 3131 2323 1515 8 2

3737 3030 2222 1414 7

2121

24 NT NT

I

S

(A) (B)

S


41

measured was the mean of ten PowerRefractor measurements in a trace. The mean of

all the measurements was considered as the accommodative response for a particular

run. The coefficient of variation (COV = mean/standard deviation) was calculated to

assess the stability of accommodation during each run. The coefficient of variation

was in the range 1.9 % to 16.5 % (for all the runs of each of the twelve subjects),

which is acceptable if we consider accommodative lag of the order of 1 D.

A short practice run for 7 points along the horizontal meridian was performed by all

subjects to become accustomed to the procedure. The data from the practice runs

were assessed immediately, and if standard deviations at more than two positions

were greater than 0.25 ln (0.11 log) units the practice run was repeated until the

subject’s performance was improved.

3.5 Mathematical fits of Stiles-Crawford effect The SCE was obtained by plotting the threshold (ln units) against pupil entry

locations and by fitting the equation 2

max )(max

xxee

using the scientific program

Matlab (Mathworks Inc., Natick, MA, version R2006a). The program uses nonlinear

least-squares regression to fit SCE models through experimentally measured SCE

data. The core function that performs the regression is NLINFIT from Matlab's

Statistics Toolbox. It uses the Gauss-Newton algorithm with Levenberg-Marquardt

modifications for global convergence. A model of the form Y=k*exp(-Rho*(X-

Xc)^2) was used for fitting 1D SCE data sets and Y=k*exp(-Rho_x*(X-Xc)^2 -

Rho_y*(Y-Yc)^2) for 2D data sets.

Fits were done separately along horizontal and vertical meridians except for the

subject NS, for whom two-dimensional fits were done. The parameters obtained were

directionality (ρ) and peak sensitivity location (xmax) of the SCE. The goodness of fit

was evaluated by examination of adjusted R2.


42

Two runs were done for each accommodation stimulus. Some runs were repeated if

the adjusted R2 of the fit was < 0.80 and/ or the standard deviations of three threshold

measurements for a pupil entry location were > 0.25 ln units for more than 3 points

(pupil locations) in the data. Finally, only two measurements for each

accommodation stimulus were considered in the analyses (section 4.1.4 –

emmetropes and 4.2.4 - myopes).

3.6 Calibration of the PowerRef II During the experiment, the PowerRef II was used to monitor the refraction of the

right eye while the subject fixated at the centre of the target in the reference channel.

The PowerRef II was aligned to the eye via hot mirror H (Figure 3.1). Continuous

measurement of accommodation during the experiment in this study was limited

because the hot mirror blocks infrared light from the eight LEDs alignment ring and

obstructs the view of the subject’s pupil in the monitor. Therefore, accommodation

was measured at regular intervals during each run by temporarily inserting the hot

mirror in front of the subject’s eye quickly to minimally interfere with the subject’s

state of accommodation. The slope measured by the PowerRef II was converted to

refraction using a calibration function obtained for each subject, using a calibration

procedure similar to that described by Schaeffel, Wilhelm & Zrenner (1993).

For calibration, an additional Badal lens system was set-up adjacent to the SCE

apparatus to provide a distant target stimulus to the left eye (Figure 3.3). The system

contained a black on white target illuminated by a white LED, whose image was at

infinity when it was positioned 20 cm from the +5 D Badal lens. This set-up was

placed on a mechanical stage to move in XYZ directions. With the right eye

occluded, the subject was asked to view the target with the left eye and directed the

examiner to move the target to be in the middle of the Badal lens field. For myopes,

contact lenses were inserted in the left eye to enable the subjects to see the fixation

target clearly. The right eye was then uncovered to perform the calibration

procedure. Ophthalmic trial lenses from +6 to -4 D in 1 D steps were placed in a lens

holder in front of the right eye while the right eye looked straight into the SCE


43

apparatus. Both SCE light sources S1 and S2 were turned off and stray light shielded

from the right eye to eliminate any stimulus for accommodation to the right eye. The

induced refractive error created by the trial lens was plotted against the slope of the

pupil intensity profile. A linear equation was fitted to the data for each subject to

describe the relationship between refraction and intensity slope. The refractions

obtained from the fitted function were then converted to accommodation.

Figure 3.3. Experimental set-up for calibration of the PowerRef II.

3.7 Accommodative stimulus-response functions

Stimulus-response functions were measured separately from the main experiment.

Stimuli for accommodation were provided by moving the optical trombone from 0 to

9 D positions in 1 D steps. Three myopic subjects (JP, YT, RSK), whose refractive

errors were partly corrected by the trombone and partly by contact lenses, could only

be provided with accommodation stimuli up to 7 D. The subject was first made

SC

E apparatus

+5D lens

Target

Badal Lens

PowerRefractor

Hot Mirror

Badal lens system used during calibration

Subject

Trial Lens

White


44

comfortable and stabilized on the bite-bar. Subject’s alignment was ensured as

described in section 3.4. Stimuli were presented in a random order to avoid bias in

the responses. Subjects were instructed to maintain clarity of the target at all times.

Results are shown in sections 4.1.2 (emmetropes) and 4.2.2 (myopes).

3.8 Control measurements Four control experiments were performed to assess factors that can influence SCE

measurements. The first control experiment investigated the influence of

directionality of light output source S2 as it traverses along horizontal and vertical

meridians on its perceived brightness. The second control experiment investigated

the influence of movements of aperture A2 on the brightness of source S2. The third

control experiment investigated the influence of sampling rate across the pupil in

order to achieve a good balance between the number of pupil positions investigated

and the time to take measurements. A fourth control study investigated the possible

effects of contact lenses, aberrations and errors of accommodation.

3.8.1 Does source S2 directionality influence the Stiles- Crawford measurement? During the experiments, source S2 traverses horizontally and vertically in a plane

perpendicular to the apparatus optic axis. To investigate whether these translations

affect SCE measurements because of source directionality, the luminance of the test

spot image was measured at different positions of source S2 along horizontal and

vertical meridians while keeping the aperture A2 centred. Images of the test spot

were captured with a camera (PixeLINK) placed at the position of subject’s pupil.

First, relative illuminance measures were obtained where test spot images were

captured through neutral density (ND) filters placed midway between the relay lenses

L3 and L4 for the central location of the test spot. The ND filters ranged in optical

density from 0.1 to 0.5 in 0.1 steps. The mean pixel brightness of each image was


45

calculated using a Matlab program. Transmission was calculated for each ND filter

image using the formula

Transmission = 1/10ND -------------------------------- (3.1) The transmission values of ND filters were plotted against mean pixel brightness of

the images (Figure 3.4).

Mean Pixel Brightness

0 5 10 15 20 25 30

ND

Filt

er

Tra

ns

mis

sio

n (

%)

20

40

60

80

100

120

Pixel brightness of test imageFit

% Transmission = 3.22 * pixel brightness + 26.67r ² = 0.94

Figure 3.4. ND filter transmission (%) plotted against mean pixel brightness of the test image.

Circles represent the mean luminance of the test spot for each ND filter. The solid line is the best

linear fit.

Based on the fit obtained with the ND filters, the change in test spot image brightness

for various source positions was calculated. Images of the test spot image for various

positions of source S2 were captured. Source S2 was moved in 2 mm steps from its

central position out to ± 4 mm horizontally (nasal to temporal in the pupil) while

keeping its vertical position fixed. S2 was then moved vertically in 2 mm steps from

its central position out to ± 4 mm vertically (inferior to superior in the pupil) while

keeping its horizontal position fixed. The variation in image brightness was plotted

as a function of source S2 position (Figure 3.5).


46

Horizontal position of source S2 (mm)

4 6 8 10 12 14 16 18

Ima

ge

Bri

gh

tne

ss (

%)

96

98

100

102

104

106

108

110

Brightness of test spot image

Vertical position of source S2 (mm)

4 6 8 10 12 14 16 18

Ima

ge

Bri

gh

tne

ss (

%)

96

98

100

102

104

106

108

110

Figure 3.5. Test spot image brightness as a function of source S2 position (A) horizontal

movement; (B) vertical movement.

Figure 3.5 shows approximately 11 % and 4 % variation in brightness of the test spot

image in horizontal and vertical directions, respectively, for about 8 mm movement.

This variation is small compared with the typical brightness change of 66 %

measured ± 3 mm from the peak location of the SCE (for ρ = 0.12 mm-2). Thus,

source directionality can be ignored in measurements.

3.8.2 Does aperture movement influence the Stiles-Crawford measurement? As mentioned in section 3.3.2, residual defocus and ocular aberrations associated

with movement of S2 cause offsets in the perceived position of A2 away from the

centre of A1. An x-y adjustment of A2 is made to restore alignment (see Figure 3.1).

I examined if the position of A2 might affect the apparent brightness of source S2 at

the pupil. This is really an additional test of the influence of the output directionality

of S2. The source S2 was placed at a central position in the SCE apparatus. Images of

this source for various positions of aperture A2 were captured. Aperture A2 was

moved in 1 mm steps from its central position out to ± 3 mm horizontally (nasal to

temporal in the pupil) while keeping its vertical position fixed. A2 was then moved

vertically in 1 mm steps from its central position out to ± 3 mm vertically (inferior to


47

superior in the pupil) while keeping its horizontal position fixed. This process was

repeated four more times with source S2 positioned 3 mm to the left, 3 mm to the

right, 3mm above and 3 mm below the central position. The percentage changes in

image brightness were calculated from the linear regression fit given in Figure 3.4.

The variation in image brightness was plotted as a function of aperture A2 position

for each position of source S2 (Figure 3.6).

10 11 12 13 14 15 16 17 18 19

Imag

e Il

lum

ina

nce

(%

)

70

80

90

100

110

120

130

14010 11 12 13 14 15 16 17 18 19

S2: 14.5H, 10V S2: 11.5H, 10V

10 11 12 13 14 15 16 17 18 19

10 11 12 13 14 15 16 17 18 19

Ima

ge

Illu

min

ance

(%

)

70

80

90

100

110

120

130

140

S2: 8.5H, 10V

Position of aperture A2 (mm)

10 11 12 13 14 15 16 17 18 19

70

80

90

100

110

120

130

140

70

80

90

100

110

120

130

140

S2: 11.5H, 13V

Position of aperture A2 (mm)

10 11 12 13 14 15 16 17 18 1970

80

90

100

110

120

130

140

70

80

90

100

110

120

130

140

Horizontal

Vertical

S2: 11.5H, 7V

Figure 3.6. Image brightness as a function of A2 aperture position for different locations of

source S2. Bottom axes refer to horizontal movement of aperture and top axes refer to vertical

movement of aperture. Filled squares represent image brightness when the aperture was moved

vertically. Open squares represent image brightness when the aperture was moved horizontally.

The dashed line along 100 is the expected brightness for the central position with ‘0’ ND filter.

The vertical centre position of the aperture was 15 mm when the aperture was moved

horizontally and the horizontal centre position of the aperture was 14 mm when the aperture

was moved vertically.


48

From Figure 3.6, it is evident that the image brightness changes on moving the

aperture over a 6 mm range. During the main experiment, aperture position was

documented every time it was moved.

For ease of analysis, horizontal movements of the aperture were considered for SCE

measured along horizontal pupil meridian and vertical movements of the aperture

were considered for SCE measured along the vertical pupil meridian. Linear

regression fits were obtained for each source position from Figure 3.6 (Figure 3.7).

To estimate the effect of changes in image brightness with aperture A2 movement on

the measured SCE, the thresholds obtained during the main SCE experiment were

recalculated for all pupil locations to adjust for the difference in spot brightness with

aperture location. This is all described in the rest of this section.

Horizontal position of aperture A2 (mm)

10 11 12 13 14 15 16 17 18 19

Imag

e il

lum

inan

ce (

%)

70

80

90

100

110

120

130

140

Nasal 3mm

Centre

Temporal 3mm

7.24*x - 0.12r ² = 0.99

7.25*x - 1.82r ² = 0.95

5.62*x + 14.42r ² = 0.92

Nasal:

Temporal:

Centre:

Vertical position of aperture A2 (mm)

10 11 12 13 14 15 16 17 18 19

Imag

e ill

um

inan

ce (

%)

70

80

90

100

110

120

130

140

Inferior 3mm

Centre

Superior 3mm

5.68*x + 27.35r ² = 0.99

5.35*x + 31.09r ² = 0.99

4.40*x + 41.64r ² = 0.98

Inferior:

Superior:

Centre:A B

Figure 3.7. Image brightness as a function of aperture A2 position for the horizontal meridian

(A) and for the vertical meridian (B). In plot A, open circles, filled circles and inverted filled

triangles represent the aperture positions for centre, nasal and temporal locations of the test

spot in the pupil, respectively. In plot B, open circles, filled circles and inverted filled triangles

represent the aperture positions for centre, inferior and superior locations of the test spot in the

pupil. The linear regression equations are shown for data obtained for each test spot position

respectively.

The linear regression equations for each position of the source along a meridian were

compared statistically using F-tests to find whether one equation alone could be used

for different meridians.


49

Along the horizontal meridian, the linear regressions for central and nasal source

locations were not significantly different (F2,10 = 0.59, p = 0.57), linear regressions

for central and temporal locations were significantly different (F2,10 = 8.6, p = 0.007),

and linear regressions for nasal and temporal locations were significantly different

(F2,10 = 18.8, p = 0.0004). Therefore, one equation could not be used to re-calculate

the subjective brightness for each position along horizontal meridian. The equation

for nasal location was used for both nasal and central locations while the equation for

temporal location was used for the temporal location. Along the vertical meridian,

linear regressions for central and inferior source locations were not significantly

different (F2,10 = 0.88, p = 0.44), linear regressions for central and superior locations

were significantly different (F2,10 = 8.6, p = 0.007), and linear regressions for inferior

and superior locations were significantly different (F2, 10 = 23.9, p = 0.0002). The

equation for inferior location was used for both inferior and centre locations while

the equation for superior location was used for the superior location.

Two SCE measurements of the subject from the main experiment identified with the

maximum range of aperture position along both the horizontal (13.3 mm to 15.2 mm)

and the vertical meridians (12.2 mm to 14.8 mm) from two different runs were

considered to evaluate the maximum effect of aperture location on measured SCE.

Two methods of correcting thresholds were used. For method I, corrected thresholds

were determined from linear regressions described in Figure 3.7. For example, for a

source location 3 mm nasal in the pupil, a horizontal aperture location of 15.2 mm

was used. This aperture location is substituted in the corresponding equation in

Figure 3.7 (in this case y = 7.24*15.2 - 0.12) and is subtracted from 100% to obtain a

brightness threshold adjustment of +9.9 %. This brightness threshold adjustment was

multiplied by the threshold and then added to the threshold to obtain the corrected

threshold. The corrected thresholds of all pupil locations for horizontal and vertical

pupil meridians are given in Table 3.2.


50

Table 3.2. SCE measurements of one subject from the main experiment with maximum range of

aperture movement along (A) the horizontal meridian and (B) along the vertical meridian,

showing thresholds and corrected thresholds obtained using method I.

A – SCE along horizontal meridian Aperture location (mm)

Pupil location (mm) Threshold (ln) Threshold Corrected threshold (ln) Corrected threshold

15.2 3 6.42 612 6.51 67315.2 2.5 6.03 416 6.12 45715.2 2 5.72 304 5.81 33514.3 1.5 5.88 356 5.91 36814.3 1 5.48 239 5.51 24714.3 0.5 5.62 275 5.65 28514.3 0 5.33 207 5.36 21414.1 -0.5 5.63 279 5.56 26114.1 -1 5.29 199 5.23 18614.1 -1.5 5.55 258 5.49 24214.1 -2 6.17 477 6.10 44713.3 -2.5 6.58 719 6.46 64113.3 -3 6.91 1002 6.79 893

B – SCE along vertical meridian Aperture location (mm)

Pupil location (mm) Threshold (ln) Threshold Corrected threshold (ln) Corrected threshold

14.8 -3 6.70 812 6.81 90514.2 -2.5 6.63 756 6.71 817

14 -2 6.23 505 6.29 540

13.8 -1.5 6.04 421 6.10 445

13.8 -1 5.76 318 5.82 336

13.8 -0.5 5.86 350 5.91 370

13.8 0 5.74 310 5.79 328

13.3 0.5 5.82 336 5.82 337

13.3 1 6.03 416 6.03 417

13.3 1.5 6.24 512 6.24 513

13.3 2 6.59 728 6.59 730

13.1 2.5 7.11 1221 7.10 121212.2 3 7.35 1558 7.30 1485

For method II, a reference aperture location was taken as the aperture location when

aligning the SCE apparatus with a laser. The aperture position during the experiment

for each pupil-entry location of the source was subtracted from the reference aperture

location. This difference was substituted as the ‘x’ value in the appropriate linear

regression equation in Figure 3.7. The reference aperture position was 14.9 mm in

the horizontal meridian and 14.0 mm in the vertical meridian. For example, for a


51

source location 3 mm nasal in the pupil, a horizontal aperture location was 15.2 mm.

The reference horizontal aperture location 14.9 mm was subtracted from this aperture

location. This difference is multiplied by the slope of the corresponding linear

regression determined in Figure 3.7 (in this case y = 7.24*(14.9-15.2)) and divided

by 100 to obtain a brightness threshold adjustment of +2.2%. Similar to method I,

this brightness threshold adjustment was multiplied by the threshold and then added

to the threshold to obtain the corrected threshold. The corrected thresholds of all

pupil locations for horizontal and vertical pupil meridians are given in Table 3.3.

Table 3.3. SCE measurements of one subject from the main experiment with maximum range of

aperture movement along (A) the horizontal meridian and (B) along the vertical meridian,

showing thresholds and corrected thresholds obtained using method II.

A – SCE measurement along horizontal meridian Aperture Location (mm)

Pupil Location (mm) Threshold (ln) Threshold Corrected threshold (ln) Corrected threshold

15.2 3 6.42 612 6.40 59915.2 2.5 6.03 416 6.01 40715.2 2 5.72 304 5.70 29814.3 1.5 5.88 356 5.92 37214.3 1 5.48 239 5.52 24914.3 0.5 5.62 275 5.66 28714.3 0 5.33 207 5.37 21614.1 -0.5 5.63 279 5.67 29114.1 -1 5.29 199 5.34 20714.1 -1.5 5.55 258 5.60 27014.1 -2 6.17 477 6.21 49913.3 -2.5 6.58 719 6.66 78313.3 -3 6.91 1002 7.00 1092


52

B – SCE measurement along vertical meridian Aperture Location (mm)

Pupil Location (mm) Threshold (ln) Threshold Corrected threshold (ln) Corrected threshold

14.8 -3 6.70 811 6.65 77414.2 -2.5 6.63 756 6.62 74814.0 -2 6.23 505 6.23 50513.8 -1.5 6.04 421 6.05 42613.8 -1 5.76 318 5.77 32213.8 -0.5 5.86 350 5.87 35413.8 0 5.74 310 5.75 31313.3 0.5 5.82 336 5.85 34713.3 1 6.03 416 6.06 42913.3 1.5 6.24 512 6.27 52813.3 2 6.59 728 6.62 75113.1 2.5 7.11 1221 7.15 126912.2 3 7.35 1558 7.43 1681

The corrected thresholds (ln) obtained using method I (Table 3.2) and method II

(Table 3.3) for all pupil locations were re-fitted using the Matlab program to obtain

corrected SCE directionalities and peak locations (Table 3.4 and Table 3.5).

Table 3.4. Uncorrected and corrected SCE parameter fits using method I. x and y are the

directionality of the SCE along horizontal and vertical meridians respectively; xc and yc are the

peak locations of the SCE along horizontal and vertical meridians respectively.

ρx (mm-2) xc (mm) ρy (mm-2) yc (mm)

Uncorrected SCE

0.140 +0.24 0.149 -0.33

Corrected SCE (method I)

0.139 +0.10 0.148 -0.26

Corrected SCE – uncorrected SCE

-0.001 -0.14 -0.001 +0.07


53

Table 3.5. Similar to Table 3.4, but using method II.


Uncorrected SCE

0.140 +0.24 0.149 -0.33

Corrected SCE (method II)

0.138 +0.30 0.148 -0.37

Corrected SCE – uncorrected SCE

-0.002 +0.06 -0.001 -0.04

Comparing Table 3.4 and Table 3.5 reveals small reductions in directionality along

both horizontal and vertical meridians with both methods (greatest difference being

0.002 mm-2). The peak sensitivity was shifted in opposite directions in the pupil after

analysing the data with both the methods (method I: temporal (-0.14 mm) and

superior (+0.07 mm); method II: nasal (+0.06 mm) and inferior (-0.04)) but the

overall shift was very small (< 0.15 mm). The changes shown are for the case where

the aperture movement was the maximum, and it is likely that SCE changes for most

if not all other subject/accommodation combinations would be smaller.

To summarise this and the previous section 3.8.1, test source S2 directionality and

movement of aperture A2 have only minor influences on the SCE. I believe that I can

safely ignore them.

3.8.3 Effect of sampling intervals on Stiles-Crawford measurement In order to minimize durations of subject runs without loss of accuracy in performing

the psychophysical task, the number of points needed to be sampled across the pupil

was investigated. The SCE fits were compared for 0.25 mm, 0.50 mm, 0.75 mm and

1.0 mm sampling intervals data for three subjects. The pupil was dilated using the


54

same protocol as described in section 3.4. The SCE measurements were performed

by sampling across a 6 mm pupil at 0.25 mm intervals along horizontal meridian (25

points) for 0 D accommodation stimulus. The data for 0.50 mm (13 points), 0.75 mm

(9 points) and 1.0 mm (7 points) sampling intervals were extracted from the data of

0.25 mm sampling. All the data were fitted with the SCE equation (2.3a) using the

Matlab program. The individual directionality and peak location across different

sampling intervals are shown in Table 3.6.

Table 3.6. SCE directionality (ρx) and peak location (xc) with different sampling intervals for

three subjects.

0.25 0.50 0.75 1.00 0.25 0.50 0.75 1.00

subject 1 0.099 0.099 0.096 0.095 + 0.29 + 0.27 + 0.27 + 0.28

subject 2 0.095 0.102 0.098 0.105 + 0.79 + 0.70 + 0.58 + 0.56

subject 3 0.090 0.096 0.100 0.100 + 0.95 + 0.78 + 0.75 + 0.62ρx (mm-2) xc (mm)

sampling (mm)

Measurement with a 0.25 mm sampling interval required approximately 30 minutes

along one meridian which was twice the time required for 0.50 mm sampling

interval. A one-factor within subjects ANOVA showed no significant differences

between the four sampling intervals (F3,6=1.08; p=0.43). Although SCE directionality

and peak location did not differ between different sampling intervals, the 1.0 mm

sampling interval reduced greatly the number of points measured across the pupil.

Therefore, a sampling interval of 0.50 mm was used in the study except for two-

dimensional measurements for subject NS for which the grid sampling interval was

0.75 mm (49 points).

3.8.4 Influence of contact lens, aberrations, and accommodative lag on the Stiles-Crawford effect To determine if use of contact lenses in the main study can influence the SCE in

myopes, the SCE was measured with a +5 D contact lens, a -5 D contact lens and

without a contact lens for 2 D myopic subject DAA. The contact lenses used were


55

spherical soft contact lenses with an optic zone diameter of 9 mm (Freshlook UV,

CIBAVISION). Pupil dilation was achieved with one drop of 1% cyclopentolate. A

drop of 2.5 % phenylephrine was also used if the required pupil size of 7 mm was not

achieved. For each correction condition, two runs were taken, using 49 points in a

two–dimensional grid across a 6 mm pupil with 0.75 mm sampling intervals. Three

measurements were taken at each pupil–entry location. Contact lens induced defocus

was compensated with movement of the optical trombone prior to runs.

Table 3.7 shows that SCE directionality along both horizontal (ρx) and vertical (ρy)

meridians increased considerably with the +5 D contact lens (about 55%), but only

slightly with the -5 D contact lens (about 12%), relative to the no contact lens

condition. The SCE peak shifted slightly temporally with both contact lenses relative

to the no contact lens condition.

Table 3.7. Mean directionality (ρx, ρy) and peak locations (xc, yc) of two SCE measurements

along horizontal and vertical meridians for a 6 mm pupil for no contact lens, +5 D contact lens

and -5 D contact lens conditions. The second entry in each cell is the difference between two

runs.


No CL 0.145, 0.001 -0.05, 0.03 0.106, 0.007 -0.57, 0.03

-5D CL 0.157, 0.008 -0.35, 0.08 0.123, 0.020 -0.70, 0.21

+5D CL 0.216, 0.014 -0.26, 0.11 0.171, 0.041 -0.40, 0.23

To check whether the contact lenses truly influenced the results, another experiment

was conducted with lenses from a trial lens set. Three runs of SCE measurements

were performed with each of -1.25 D and +1.25 D trial lenses and without a trial

lens. The trial lenses were placed midway between lenses L3 and L4 in the apparatus

(Figure 3.1). Since the effective power of a lens placed at this position is 4 times its

actual power, a 1.25 D lens induces 5 D defocus. Trial lens induced defocus was

compensated with movement of the optical trombone prior to runs. Each run

comprised 13 points sampled in 0.50 mm interval along the horizontal meridian in


56

the pupil. Three measurements were taken at each pupil-entry location across a 6 mm

diameter pupil.

Table 3.8. Directionality (ρx) and peak locations (xc) of three SCE measurements along

horizontal meridian for a 6 mm pupil for no trial lens, +1.25 D trial lens and -1.25 D trial lens

conditions. SD represents the standard deviation of three runs.

ρx (mm-2) xc (mm)

Run 1 Run 2 Run 3 Mean±SD Run 1 Run 2 Run 3 Mean±SD

No trial lens 0.119 0.140 0.126 0.128±0.011 -0.08 -0.02 -0.07 -0.06±0.03

Positive trial lens

0.133 0.144 0.159 0.145±0.013 -0.08 -0.12 -0.07 -0.09±0.03

Negative trial lens

0.141 0.156 0.154 0.150±0.008 -0.28 -0.13 -0.12 -0.18±0.09

Table 3.8 shows that SCE directionality increased with both trial lenses relative to

the no trial lens condition, but this was relatively small (about 15% and less than

0.022 mm-2).

Comparing the contact lens and trial lens experiments, what stands out is that the +5

D contact lens produced a large increase in SCE directionality, while the -5 D contact

lens and the trial lenses produced much smaller effects. The results indicate that

while contact lenses can influence the SCE measurements, this influence is likely to

be small for the negative power lenses used in the main study. As we are interested in

comparing changes in SCE with accommodation and whether this is different in

myopes than in emmetropes, the effect of negative contact lenses is not likely to be

critical.

Higher order aberrations may also contribute to artefactual increases in SCE

directionality. With increasing accommodation, spherical aberration shows a

systematic change in the negative direction, usually becoming negative at 2–3 D of

accommodation (Cheng, Barnett, Vilupuru, Marsack, Kasthurirangan, Applegate &


57

Roorda, 2004; Ninomiya, Fujikado, Kuroda, Maeda, Tano, Oshika, Hirohara &

Mihashi, 2002; Plainis, Ginis & Pallikaris, 2005). The changes in spherical

aberration are much greater than the changes in other higher order aberration co-

efficients (Cheng, Barnett, Vilupuru, Marsack, Kasthurirangan, Applegate & Roorda,

2004). Small accommodative lags for higher accommodation levels can add to the

influence of negative spherical aberration and contribute to increased thresholds at

the pupil periphery and increasing directionality of the SCE.

To investigate further whether aberrations can influence the SCE, another experiment

was performed to measure the SCE in-focus and for defocus of 1 D hyperopic

(negative) defocus and 1 D myopic (positive) defocus for subject DAA. Defocus was

induced by moving the optical trombone (Figure 3.1). Two runs each of 13 points in

0.50 mm sampling intervals along the horizontal meridian for a 6 mm diameter pupil

were performed. Three measurements were taken at each pupil-entry location.

Table 3.9 shows that SCE directionality increased for myopic defocus and decreased

for hypermetropic defocus relative to the in-focus condition by 24 % and 14 %,

respectively.

Table 3.9. Directionality and peak locations of two runs of SCE measured for 6 mm pupil along

horizontal meridian using main apparatus for in-focus, 1 D hyperopic defocus and 1 D myopic

defocus conditions. Also, mean and difference between 2 runs for ρx and xc are shown in third

and sixth columns, respectively.

ρx (mm-2) xc (mm)

Run 1 Run 2 Mean, difference


In-focus 0.126 0.137 0.132, 0.011 -0.28 -0.35 -0.32, 0.06

Hypermetropic defocus

0.119 0.113 0.116, 0.164 -0.29 +0.10 -0.10, 0.39

Myopic defocus

0.164 0.164 0.164, 0.000 -0.21 -0.30 -0.26, 0.09


58

In addition to these measurements, aberrations of the subject’s eye were measured

using a Hartmann-Shack wavefront sensor without and with the contact lenses

(details of the procedure are given in section 3.10). Table 3.10 shows the higher

order RMS and spherical aberration coefficient C(4,0)

(American National Standards Institute, 2004) measured with and without the contact

lenses for 6 mm diameter pupil. The higher order RMS and spherical aberrations

were considerably higher with the +5 D contact lens than with the -5 D contact lens

or without a contact lens for a 6 mm pupil size. This shows clearly that the positive

contact lens increases the higher order aberrations, and particularly spherical

aberration, in the eye.

Table 3.10. Higher order RMS and spherical aberration of right eye for 6 mm pupil, with and

without the contact lenses. SD represents the standard deviation of three measurements.

No CL±SD +5D CL±SD -5D CL±SD

HO RMS (μm) 0.28±0.01 0.85±0.07 0.23±0.03

C(4, 0) (μm) +0.19±0.01 +0.59±0.01 +0.11±0.02

The results of the SCE experiments, together with the aberration measurements,

indicate that aberrations of the eye may influence the SCE, particularly when acting

in the same direction as any defocus, as occurred for DAA when positive spherical

aberration combined with positive defocus and may occur for an accommodating eye

when negative spherical aberration combines with a lag of accommodation.

Because of this potential problem in interpreting changes in the SCE with

accommodation, another way of investigating the SCE psychophysically was

considered in which the SCE apparatus was modified to minimize the influence of

aberrations. In the modified apparatus, rather than the source with variable pupil

entry image position and variable luminance being seen as a small spot on a much

larger background provided by the fixed source, the source with fixed position and

variable luminance was seen as a small spot on a much larger background provided

by the source with variable pupil position. Blurring of a large background should


59

have less influence on thresholds than the blurring of a small test spot. The

modification was based on the apparatus of Enoch and Hope (1972a) (compare

Figure 3.8 with Figure 3.1).

Apertures A1 and A2 were swapped (in the Figure A1 has become A2 and vice versa)

and light source S1 was flashed rather than light source S2. S2, L2 and A2 now form

the background channel, while S1, L1 and A1 form the test channel. In the main

experiment, source S1 providing the background channel was limited by a 4 mm

aperture to provide a reasonable stimulus for accommodation, but in this case a 4 mm

aperture for S2 would have provided a blurred target due to aberrations when viewed

through the pupil periphery. Thus, both S1 and S2 were limited by 1 mm apertures.

The background field aperture (A2) was translated as necessary during SCE

measurements so that A1 appeared in its centre. This modified apparatus cannot be

used for the main SCE experiment because accommodation stimulation is poor

through a 1mm pupil aperture (see section 4.1.2), but it can be used to give an

indication as to the effects of aberrations on SCE measurements.


60

Figure 3.8. Schematic representation of the modified apparatus. Inset A shows apertures A1 and

A2. Inset B shows subject’s view. Other details are as for Figure 3.1.

The luminance of the background target was determined using an indirect method

similar to that described in section 3.3.5. Subject DAA performed the task of

brightness matching because he performed the experiment with the modified

apparatus. He selected ND filter combinations (± 0.1 ND filter) that produced the

closest brightness match of an auxiliary source to the background field. Using this

method, the background field luminance was determined to be 16 cd/m2, which is

about 40% greater than that of the main apparatus.

The technique depends on the Weber relationship which means that the effective

luminance Le of the background remains high enough, as its pupil entry position

varies, to ensure that the ratio L/Le remains constant. Here Le is the SCE-weighted

luminance of the background and L is the increment threshold of the test spot. A

small experiment was performed to test this. For a hypothetical observer with SCE

directionality ρ of 0.12 mm-2, using equation 2.3a the effective luminance of the

Alignment Ring

Eye

Inset BS1

S2

L1

L2

A1

A2

BS1 BS2

L3 L4

M2 M5

L5

M1 Optical Trombone

MoVC

ND Filter

Background channel

Te

st Channel

0.6° 7°

A1 A2

Inset A

Flashes at 2 Hz

M3 M4

Alignment Ring

Eye

Inset BInset BS1

S2

L1

L2

A1

A2

BS1 BS2

L3 L4

M2 M5

L5

M1 Optical Trombone

MoVC

ND Filter

Background channel

Te

st Channel

0.6° 7°

A1 A2

Inset A

Flashes at 2 Hz

M3 M4


61

background will be reduced at the edge of a 6 mm diameter pupil, relative to that at

the centre, by 1.08 ln units (0.47 log units). A series of three runs, each with three

measurements of thresholds was taken for the background entering through the

centre of the pupil, without and with 0.4 ND and 0.6 ND filters placed between the

relay lenses (L3 and L4) (Figure 3.8). If the Weber law holds for the different filter

levels, because the filters affect both the background and the spot, there will be no

change in the threshold luminance of the test spot from that occurring for the no filter

condition. Results are shown in Figure 3.9. This shows no effect with the 0.4 ND

filter, but a small increase in threshold with the 0.6 ND filter of 0.12 ln unit (0.05 log

unit). Translated to SCE measurements, this suggests that a small, but probably

negligible increase in the directionality could occur. It was thus valid to proceed with

the experiment with the modified apparatus, although it is preferable to use higher

luminance light sources.

ND Filter

0.0 0.2 0.4 0.6 0.8

Th

esh

old

(ln

un

its)

4.75

4.80

4.85

4.90

4.95

5.00

5.05

5.10

5.15

Threshold

Figure 3.9. Mean thresholds of three runs for the central location of a 1-mm diameter pupil;

without filter and with 0.4 and 0.6 ND filters. Error bars represent standard deviations.

Two runs along the horizontal meridian with 0.50 mm sampling interval were

performed for in-focus, 1 D hypermetropic defocus and 1 D myopic defocus

conditions for subject DAA.


62

Table 3.11 shows the SCE fits with the modified apparatus for the 3 defocus

conditions. This Table should be compared with Table 3.9 which shows results with

the main apparatus. The main difference is that the increase in SCE directionality (24

%) with positive defocus found with the main apparatus does not occur with the

modified apparatus. In fact, there is a 15% decrease.

Table 3.11. Directionality and peak locations of two runs of SCE analysed for 6 mm pupil along

horizontal meridian (ρx) using modified apparatus for in-focus, 1 D hyperopic defocus and 1 D

myopic defocus conditions. Also, mean and difference between two runs for ρx and xc are shown

in third and sixth columns respectively.

ρx (mm-2) xc (mm)



In-focus 0.115 0.099 0.107, 0.016 +0.08 +0.01 +0.05, 0.07

Hypermetropic defocus

0.093 0.083 0.088, 0.010 -0.02 -0.12 -0.07, 0.10

Myopic defocus

0.093 0.092 0.093, 0.001 -0.25 -0.06 -0.16, 0.19

The comparison provides strong evidence that aberrations and defocus can influence

the SCE results for the main experiment. Because of this, analysis of the experiments

in this section was also done at 5 mm pupil diameter as this may be affected less by

aberrations and defocus than the 6 mm pupil. The SCE functions were refitted over 5

mm pupils (deleting 12 points for 2-D data and deleting 2 points for 1-D data). Table

3.12 shows that the percentage changes in SCE directionality for all the conditions

were similar between 6 mm and 5 mm pupil, with a maximum difference of 9%. This

comparison suggests that SCE results for both 6 mm and 5 mm pupil sizes are

similarly affected by accommodative lag and aberrations. Therefore, fits to a 6 mm

pupil can be safely used for analysis of the main experiment (Chapter 4) as it also

involves more points across the pupil generating a better fit in comparison to a 5 mm

pupil. However, these optical defects must be taken into account in interpretation of

results.


63

Table 3.12. A comparison of percentage changes in SCE directionality of all conditions for both

6 mm and 5 mm pupil sizes relative to their respective no contact lens, no trial lens and in-focus

conditions. Positive and negative signs indicate increase and decrease in percentage changes,

respectively.

Condition 6 mm 5 mm

Contact lens +5 D CL +55 % +46 %

-5 D CL +12 % +8 %

Trial lens +1.25 D TL +13 % +8 %

-1.25 D TL +17 % +18 %

Defocus – main apparatus

1 D myopic +24 % +27 %

1 D hypermetropic

-12 % -15 %

Defocus – modified apparatus

1 D myopic -13 % -7 %

1 D hypermetropic

-18 % -13%

3.9 Peak-finding technique Blank, Provine & Enoch (1975) measured the shift in the peak of the SCE with

accommodation via a fast procedure described as the “peak-finding” technique. They

found a nasal shift of approximately 1.5 mm across three emmetropic subjects for 9

D accommodation stimulus. For another measure of changes in the SCE with

accommodation, I performed another study using the peak-finding technique, which

is described below.

The apparatus for the psychophysical study (see section 3.3) was modified. Two

round apertures of 0.30 mm diameter separated by 2 mm replaced the 1 mm aperture

of source S2 in the test channel (Figure 3.10, Inset A). The apertures were covered by

linearly polarized filters oriented at 90° to each other. Aperture A2 of the test channel

was replaced by two oval apertures each measuring 4 mm x 3 mm and separated by 1

mm (Figure 3.10, Inset B). The apertures formed the target. The apertures were


64

covered by linearly polarized filters oriented at right angles to each other. The target

and the light source apertures could be orientated horizontally or vertically, but in

order to investigate the peak along the horizontal meridian, I displaced the apertures

of the light source horizontally and the apertures of the target vertically. The

background target at A1 served as an accommodation and fixation target.

In this arrangement, one light source aperture was viewed through one part of the

target and the other light source aperture was viewed through the other half of the

target. The two halves of the target appeared equally bright when the two light source

apertures were equally displaced about the position corresponding to the SCE peak in

the pupil.

Figure 3.10. Schematic representation of the peak-finding technique. Inset A shows the

apertures of source S2. Inset B shows apertures A1 and A2. Inset C shows the subject’s view.

Alignment Ring

S1

S2

L1

L2

A1

A2

BS1 BS2

L3 L4

M2 M5

L5

Eye

Monitor

Video Camera

M1 Optical Trombone

(Mo)

(VC)

Inset A

S2

Inset C

Ba

ckgroundch

annel

Test Channel

7°

A2 A1

Inset B

M3 M4

Alignment Ring

S1

S2

L1

L2

A1

A2

BS1 BS2

L3 L4

M2 M5

L5

Eye

Monitor

Video Camera

M1 Optical Trombone

(Mo)

(VC)

Inset A

S2

Inset C

Ba

ckgroundch

annel

Test Channel

7°

A2 A1

Inset B

7°

A2 A1

Inset B

M3 M4


65

Five of the emmetropic subjects who participated in the psychophysical study were

recruited for this study. To avoid any influence of contact lens aberrations on the

results, myopes were not included. Subjects’ right eyes were dilated with a 2.5 %

phenylephrine eye drop approximately 35-40 minutes before the experiment.

Subjects were asked to bite on the bite–bar and eye alignment was maintained as

described previously in section 3.4. The stimulus for accommodation was provided

by moving the optical trombone. Two experimenters were involved, one translating

the light source while the other experimenter maintained the eye alignment and

recorded the results. Subjects fixated at the centre of the two illuminated ovals of the

viewing target (Figure 3.10, Inset C). The first experimenter translated the light

source S2 slowly from one side to the other until the subject could perceive a

difference in brightness of the two apertures of the target. The procedure was

repeated in the other direction. The subject’s task was to indicate when the two

apertures of the target appeared equally bright. The experimenter approached this

position in both directions, in order to counteract any subject bias in judgement.

For two subjects EM and NS, 6 sets of measurements were made for each of 0 D and

6 D accommodation stimuli (one set comprising single approaches from both

directions). For the other three subjects (AGK, AM and ST), 12 sets of

measurements for each of 0 D and 6 D were done in a sequence that began with 6

sets for 0 D stimuli followed by 6 sets for 6 D stimuli, followed by a 5 min break and

then by 6 sets for 6 D stimuli and 6 sets for 0 D stimuli. The average peak locations

obtained from the two directions were averaged. Results of this experiment are

described in section 4.4.

3.10 Aberration measurements for emmetropes In section 3.8.4, it was found that the aberrations in the eye can potentially influence

the SCE measurement for high accommodative stimuli. Therefore, it was important

to appreciate ocular aberrations and to know the changes in aberrations with

accommodation. This experiment was conducted after completing other experiments


66

reported in this chapter and only on the emmetropic subjects who were readily

available.

3.10.1 Subjects Right eyes of six emmetropic subjects who participated in the main experiment were

tested.

3.10.2 Instrumentation The monochromatic aberration function of the eye was measured using the Complete

Ophthalmic Analysis System-HD (COAS-HD, Wavefront Sciences, Inc.,

Albuquerque, NM, USA), which uses the Shack-Hartmann principle (Thibos, 2000).

COAS-HD provides a real time display of the pupil image, which is used to measure

pupil size to the nearest 0.1 mm. COAS-HD uses an 840 nm infrared super-

luminescent diode as the radiation source and a lenslet array of 33 × 44 (a total of

1452 lenslets). The diameter of each lenslet is 144 μm. The lenslet array samples the

exiting wavefront every 210 µm in pupil plane and allows approximately 600 sample

points within a 6 mm diameter pupil. The software can record Shack-Hartmann

images and pupil size with an exposure time of about 130 ms for each frame capture.

The fixation target is a circular grid with a red illuminating spot in the centre of the

grid. The data extracted from COAS-HD consist of a set of Zernike coefficients in

the format recommended by the Optical Society of America (OSA) (Thibos,

Applegate, Schwiegerling, Webb & VSIA Standards Taskforce Members, 2002) .

The COAS-HD has a slider control panel that enables manual correction of the

refraction.


67

3.10.3 Calibration of the instrument In the usual use of the COAS-HD instrument, it performs a “fogging” of the fixation

target during every measurement to encourage relaxation of accommodation of the

eye under examination. A calibration procedure was performed to set the COAS

slider control to correspond to different accommodation stimuli. The calibration

procedure involved two observers and a telescope placed at the position of the eye,

38 mm from the front of the COAS-HD. Each observer adjusted the eye piece to

focus the telescope at infinity. Ophthalmic trial lenses (-6 D to +8 D in 1 D intervals)

were placed in front of the objective lens of the telescope while the observers looked

through the telescope at the fixation target. The precision of the slider position

settings was ± 0.1 D. The observer moved the slider until the fixation target was in

focus for each trial lens.

For most trial lenses the slider position settings of the two subjects were within 0.1D.

The settings were averaged. A quadratic regression of mean slider position on trial

lens power was performed (Figure 3.11). This equation was used to determine the

slider setting required for a given accommodation stimulus (which had the opposite

sign to the trial lens power).


68

Trial Lens Power (D)

-8 -6 -4 -2 0 2 4 6 8

Sli

der

Po

siti

on

(D

)

-12

-10

-8

-6

-4

-2

0

2

4

6

Slider Position (Observer 1)

Slider Position (Observer 2)

y = -1.25 + 1.02*x - 0.03*x^2r ² = 1.00

Figure 3.11. Slider positions for two observers for different trial lens powers. A quadratic fit

was made to the means of the observers’ settings.

3.10.4 Procedure to measure ocular aberrations Aberrations were measured for 0, 2, 4 and 6 D accommodation stimuli. One subject

(LS) performed the SCE task for 8 D accommodation and therefore for this subject

aberrations were measured for 8 D accommodation also. The required stimulus was

provided by moving the slider according to the fit shown in Figure 3.11. The

aberrations were measured over a 6 mm pupil up to 6th order Zernike polynomials.

Two measurements were made for each accommodation stimulus. Each subject’s

right eye was dilated with one drop of 2.5% phenylephrine. A pupil size between 6

mm and 6.5 mm was required as at higher-order aberrations and larger pupils, some

of the radiation passing through the pupil can be vignetted by an aperture stop that

reduces the influence of the corneal reflex. This vignetting can be recognized by the

maximum analysable pupil size being smaller than the actual pupil size. The room

illumination was controlled to get the pupil size within the required range. The

subject positioned his/her head on the chin rest and fixated on the centre of the


69

fixation target. The operator manually aligned the subject's pupil centre with the

optical axis of the device with the aid of six dots (that lie on a circle concentric with

the pupil) displayed on a video monitor. This ensured that the subject's line-of-sight

was coaxial with the instrument's optical axis.

3.10.5 Data analysis In order to confirm that the subjects responded well to the accommodative stimuli

provided during the experiment, the accommodative response was evaluated by

calculating the defocus using the second–order ( 02C ), fourth–order ( 0

4C ) and sixth–

order ( 06C ) Zernike coefficients according to: 20

604

02 /)72451234( RCCC ,

where coefficients are in µm and R is the pupil radius in mm (Atchison, 2004).

Although aberration measurement was attempted for a 6 mm pupil, the pupil

diameter for the 6 D accommodation stimuli was less than 6 mm for some subjects,

and therefore, the Zernike expansion coefficients were obtained for a 5 mm pupil

diameter. The coefficients were "scaled" to a 5 mm pupil diameter using the standard

employed by COAS software, which re-calculated Zernike expansion coefficients

(up to 6th order) after discarding the parts of the Shack-Hartmann image outside the 5

mm zone.

The subjects’ aberrations for a particular stimulus were represented by the average

Zernike coefficients of two measurements. Results are presented in section 4.5.

Chapter 4 SCE using a psychophysical technique - Results

70

Chapter 4

The Stiles-Crawford effect in emmetropes and myopes using a

psychophysical technique - Results

4.1 Emmetropes

4.1.1 PowerRef II calibration Individual PowerRef II calibration functions for each subject were used to obtain

accommodative responses from the measured slopes of pupil intensity gradient with

the PowerRef II. Individual calibration functions were linear over the range of

refractive errors from +6 to -4 D (r2 ranged from 0.969 to 0.995). Figure 4.1 shows

combined data and a combined calibration.


71

PowerRef II measured refraction (D)

-8 -6 -4 -2 0 2 4 6 8

Ind

uc

ed r

efr

acti

on

(D

)

-8

-6

-4

-2

0

2

4

6

8

AMEMNSLSAKST

Induced refraction = 0.80*PowerRef II measured refraction - 0.90 r ² = 0.97

Figure 4.1. Relationship between PowerRef II measured refraction and trial lens induced

refraction for the 6 emmetropic subjects. Individual calibration functions were obtained for

each subject by fitting linear regression equation to the data of each subject (not shown). A

linear equation is fitted to the cumulative data (line shown) to demonstrate overall linear

relation between induced refraction and PowerRef II measured refraction.

4.1.2 Accommodative stimulus-response functions: 1 mm vs. 4 mm aperture

To obtain stimulus-response functions with the source S2 having 1 mm and 4 mm

limiting apertures in the reference channel, the stimuli for accommodation were

provided by moving the optical trombone from 0 D to 9 D positions in 1 D steps as

explained in section 3.7. The accommodative responses are plotted against

accommodative stimuli in both conditions for individual subjects in Figure 4.2a and

for the group in Figure 4.2b.


72

Accommodation Stimulus (D)

-1 0 1 2 3 4 5 6 7 8 9 10

Ac

com

mo

dat

ion

Res

po

nse

(D

)

-1

0

1

2

3

4

5

6

7

8

9

10

AM

EM

NS

LS

AK

ST

4 mm aperture

-1 0 1 2 3 4 5 6 7 8 9 10

1 mm aperture

A B

Figure 4.2a. Accommodative stimulus-response curves with (A) 4 mm aperture and (B) 1 mm

aperture. The results for different subjects are represented as different symbols. The dashed

line represents the 1:1 line. The variability within a run across all subjects ranged from 0.03 D

to 0.32 D, but for clarity variability is not shown.


-1 0 1 2 3 4 5 6 7 8 9 10

Ac

com

mo

dat

ion

Res

po

nse

(D

)

-1

0

1

2

3

4

5

6

7

8

9

10

4 mm aperture 1 mm aperture

Figure 4.2b. Mean accommodative stimulus-response of subjects with 4 mm (filled circles) and 1

mm apertures (open circles). The error bars are the standard deviations of the means. The

dotted line represents the 1:1 line.


73

It is evident from the Figure 4.2 that the 4mm aperture provided more accurate and

less variable accommodative responses than the 1mm aperture. This justifies using

the larger aperture size in the main SCE experiment. For both sizes, lag of

accommodation increased with stimulus amplitude beyond 4 D in five out of six

subjects. Subject AM (filled circles) showed the lowest responses for 8 D and 9 D

stimuli with the 4 mm aperture and also reported poor clarity of the target during

measurement.

4.1.3 Accommodation responses of subjects during Stiles-Crawford measurements The PowerRef II measurements were taken at every fourth pupil entry position

during SCE measurements. The PowerRef II measurements were converted to

refractions using the linear equations fitted to the data of induced refractive error by

the trial lens and the slope of the pupil intensity profile for each subject (see section

3.6 for details). The mean of these measurements was considered as the

accommodative response for a particular run. The mean accommodative responses

of two runs for each subject are plotted against accommodative stimuli in Figure 4.3.


74

Accommodative Stimulus (D)

-1 0 1 2 3 4 5 6 7 8 9

Acc

om

mo

da

tive

Res

po

nse

(D

)

-1

0

1

2

3

4

5

6

7

8

9AM EMNSLSAKST

Figure 4.3. Accommodative response for emmetropic subjects as a function of accommodative

stimulus. Each subject is represented by a different symbol. The variability between the two

runs across all subjects ranged from 0.15 D to 1.27 D, but for clarity variability between the two

runs is not shown.

Subjects showed lead of accommodation for 0 D stimuli and lag with the increase in

stimulus (except LS). In general, subjects found the focusing task difficult for 6 D

and 0 D, yet were able to perform well by taking a few short breaks during the

experiment. At high accommodation stimulus another issue was a strong pupil

constriction and so an additional dose of 2.5 % phenylephrine was instilled 50

minutes after the first instillation when the effect of the drug started to decrease

(Mordi, Lyle & Mousa, 1986). LS and NS could clear the target at 8 D stimulus but it

was difficult to achieve pupil diameters >7mm for NS at this stimulus. Therefore

SCE could not be measured for NS at 8 D stimulus. As the stimulus-response

functions were not very different when done separately (Figure 4.2A) or during SCE

measurements (Figure 4.3), I made the tentative conclusion that SCE measurements

did not interfere with accommodative responses.


75

4.1.4 Changes in the Stiles-Crawford effect with accommodation Individual SCE directionality changes along the horizontal meridian with

accommodation for subjects EM and AM are shown in Figure 4.4. Plots A and B

show raw thresholds and plots C and D show fitted thresholds as a function of pupil

entry location. Plots C and D shows that SCEs are steeper for 6 D accommodation

stimulus than for the 0 D accommodation stimulus in both these subjects.

Pupil Entry Location (mm)-4 -3 -2 -1 0 1 2 3 4

Th

resh

old

(ln

un

its

)

5.6

5.8

6.0

6.2

6.4

6.6

6.8

7.0

7.2

7.4

7.6

7.8

0D 6D


-4 -3 -2 -1 0 1 2 3 4

Th

resh

old

(ln

un

its)

5.6

5.8

6.0

6.2

6.4

6.6

6.8

7.0

7.2

7.4

7.6

7.8

EM AM


-4 -3 -2 -1 0 1 2 3 4

Th

res

ho

ld (

ln u

nit

s)

5.6

5.8

6.0

6.2

6.4

6.6

6.8

7.0

7.2

7.4

7.6

7.8

A B

D


-4 -3 -2 -1 0 1 2 3 4

Th

resh

old

(ln

un

its

)

5.6

5.8

6.0

6.2

6.4

6.6

6.8

7.0

7.2

7.4

7.6

7.8

0D

6D

C

Figure 4.4. SCE results for subjects EM and AM for 0 D and 6 D accommodative stimuli along

the horizontal meridian. (A) and (B) are raw thresholds plotted against pupil entry locations,

and (C) and (D) are fits plotted against pupil entry locations. The 6 D accommodation fits have

been shifted vertically by 0.06 ln units in both subjects so that the peaks for 0 D and 6 D curves

coincide.

Table 4.1 summarizes SCE and accommodation results for all subjects. The

parameters are based on fitting across a 6 mm pupil diameter.


76

Table 4.1. Means (first entry) and differences (second entry) of parameter fits for two SCE runs

at different accommodative stimuli for emmetropes. AS is accommodation stimulus and AR is

accommodation response, respectively.

Subject AS (D) AR (D) ρx (mm-2) ρy ( mm-2) xc (mm) yc ( mm)

EM 0 -0.1, 0.1 0.102, 0.014 0.095, 0.001 +0.48, 0.01 +0.10, 0.172 1.9, 0.2 0.122, 0.019 0.109, 0.000 +0.43, 0.25 -0.05, 0.034 3.7, 0.3 0.126, 0.005 0.121, 0.004 +0.33, 0.14 +0.19, 0.216 4.9, 0.3 0.139, 0.001 0.126, 0.005 +0.43, 0.07 -0.02, 0.04

AM 0 -0.1, 0.0 0.091, 0.003 0.109, 0.014 +0.38, 0.14 -0.43, 0.272 1.6, 0.1 0.104, 0.003 0.151, 0.047 +0.32, 0.00 -0.37, 0.054 3.6, 0.6 0.131, 0.019 0.134, 0.015 +0.14, 0.10 -0.24, 0.056 4.9, 1.8 0.143, 0.008 0.162, 0.025 +0.20, 0.08 -0.22, 0.20

AK 0 0.3, 0.0 0.083, 0.003 0.093, 0.024 +1.02, 0.27 +0.03, 0.062 1.9, 0.1 0.078, 0.025 0.094, 0.028 +1.37, 0.07 -0.15, 0.174 3.5, 0.2 0.102, 0.003 0.123, 0.010 +0.66, 0.38 -0.51, 0.096 5.2, 0.1 0.11, 0.018 0.119, 0.054 +1.16, 0.56 -0.38, 0.10

LS 0 0.2, 0.6 0.124, 0.002 0.125, 0.003 -0.31, 0.05 +0.24, 0.062 1.9, 0.3 0.117, 0.010 0.126, 0.018 -0.47, 0.16 +0.34, 0.144 4.2, 0.5 0.114, 0.025 0.134, 0.057 -0.60, 0.03 +0.32, 0.126 7.2, 0.9 0.125, 0.009 0.167, 0.078 -0.30, 0.01 +0.23, 0.108 8.1, 0.4 0.147, 0.019 0.162, 0.035 -0.24, 0.03 +0.25, 0.09

ST 0 1.0, 0.4 0.110, 0.034 0.118, 0.001 +0.17, 0.09 +0.39, 0.082 1.1, 0.2 0.114, 0.010 0.144, 0.025 +0.16, 0.02 +0.43, 0.264 3.8, 0.3 0.149, 0.007 0.124, 0.010 +0.07, 0.03 +0.42, 0.046 4.9, 0.0 0.138, 0.019 0.127, 0.027 +0.23, 0.22 +0.69, 0.06

NS 0 0.4, 0.3 0.139, 0.002 0.126, 0.025 +0.01, 0.29 -0.05, 0.092 1.8, 0.1 0.152 , 0.010 0.152, 0.010 +0.19, 0.15 -0.18, 0.014 3.7, 0.3 0.160, 0.007 0.160, 0.007 +0.19, 0.12 +0.07, 0.096 5.8, 0.5 0.168, 0.007 0.168, 0.007 +0.14, 0.04 +0.12, 0.01

Changes in ρx, ρy, xc and yc for all subjects are plotted against accommodative

response (Figure 4.5). For subject LS, data for 8 D accommodation stimuli were also

included in the analysis. The change in a parameter for a particular subject,

accommodation stimulus and run was obtained by subtracting the parameter value, at

the least accommodative response for that subject, from the parameter. Results from

both runs are shown.


77

Accommodation Response (D)

-1 0 1 2 3 4 5 6 7 8 9

Ch

ang

e in

x

(mm

-2)

-0.08

-0.06

-0.04

-0.02

0.00

0.02

0.04

0.06

0.08

0.10

0.12

0.14y = 0.003*x + 0.011r ² = 0.083; p = 0.04

Accommodation Response (D)-1 0 1 2 3 4 5 6 7 8 9

Ch

ang

e in

y

(mm

-2)

-0.08

-0.06

-0.04

-0.02

0.00

0.02

0.04

0.06

0.08

0.10

0.12

0.14EM

AM

AK

LS

ST

NS

y = 0.005*x - 0.001r ² = 0.261; p = 0.0001


-1 0 1 2 3 4 5 6 7 8 9

Ch

ang

e in

xc

(mm

)

-1.5

-1.0

-0.5

0.0

0.5

1.0

1.5y = 0.004*x - 0.006r ² = 0.001; p = 0.82


-1 0 1 2 3 4 5 6 7 8 9

Ch

ang

e in

yc

(mm

)

-1.5

-1.0

-0.5

0.0

0.5

1.0

1.5y = -0.001*x - 0.004r ² = 0.0004; p = 0.89

A B

C D

N = 6

Figure 4.5. Changes in ρx (A), ρy (B), xc (C) and yc (D) are plotted against the accommodative

response for emmetropes. Pupil size is 6 mm. In all the graphs, each subject is represented by a

different symbol. The solid lines in the plots are the linear fits to the data.

The regression analysis showed that the directionality along both horizontal and

vertical meridians increased significantly with accommodation (ρx slope = +0.003

mm-2/D, r2 = 0.08, p = 0.04; ρy slope = +0.005 mm-2/D, r2 = 0.26, p = 0.0001), but

that the peak locations did not show any significant changes with accommodation.

Because each subject had a number of observations, these were not independent. A

linear mixed model was applied to take this into account (Statistical Packages for the

Social Sciences version 16), but there were negligible changes to the probabilities.


78

4.2 Myopes

4.2.1 PowerRef II calibration

A combined PowerRef II calibration functions for myopic subjects is shown in

Figure 4.6. As for emmetropes, to obtain accommodative responses from all subjects,

individual PowerRef II calibration functions for each subject were used. Individual

calibration functions for myopes were also reasonably linear over the range of trial

lens induced refractive errors from +6 to -4 D (r2 ranged from 0.88 to 0.98). The

calibration functions for myopes were more scattered and thus variable when

compared to calibration functions of emmetropes.

PowerRef II Measured Refraction (D)

-8 -6 -4 -2 0 2 4 6 8

Ind

uce

d R

efra

ctio

n (

D)

-8

-6

-4

-2

0

2

4

6

8

EK JP PM YT WF SK

Induced refraction = 0.83*PowerRef II measured refraction + 0.26 r ² = 0.84

Figure 4.6. Relationship between PowerRef II measured refraction and trial lens induced

refraction for the 6 myopic subjects. Individual calibration functions were obtained for each

subject by fitting linear regression equation to the data of each subject (not shown). A linear

equation is fitted to the cumulative data (line shown) to demonstrate overall linear relation

between induced refraction and PowerRef II measured refraction.


79

4.2.2 Accommodative stimulus-response functions Accommodative responses are plotted against accommodative stimuli for individual

subjects in Figure 4.7.


-1 0 1 2 3 4 5 6 7 8 9 10

Acc

om

mo

dat

ion

Res

po

nse

(D

)

-1

0

1

2

3

4

5

6

7

8

9

10EK PM JP WF YT SK

Figure 4.7. Accommodative stimulus-response curves for myopes. The results for different

subjects are represented as different symbols. The dashed line represents the 1:1 line. The

variability within a run across all subjects ranged from 0.15 D to 1.27 D, but for clarity

variability is not shown.

For three subjects (JP, YT, SK), the residual refractive error that was corrected with

the trombone was subtracted from the PowerRef II measured refractions.

Three subjects (EK, PM, SK) demonstrated increased lag of accommodation beyond

1 D of accommodation stimulus. Subject PM showed higher lags than the other

subjects. The two of the subjects (YT and JP) had stimulus-response curves close to


80

1:1 line. Subject WF exhibited a lead of accommodation for lower accommodative

stimuli.

4.2.3 Accommodation responses of subjects during Stiles-Crawford measurements

The mean of the accommodative responses during the two runs for each subject are

plotted against accommodative stimuli in Figure 4.8. In general, subjects found the

focusing task difficult for 6 D and 0 D, yet were able to perform reasonably well by

taking a few short breaks during the experiment. Two of the subjects (EK and WF)

were able to clear the target at 8D stimuli. Subject EK performed the SCE task for

8D accommodative stimulus and subject WF was not available for the 8D SCE task.


-1 0 1 2 3 4 5 6 7 8 9

Ac

com

mo

dat

ion

Res

po

nse

(D

)

-1

0

1

2

3

4

5

6

7

8

9EKPMJPWFYTSK

Figure 4.8. Accommodative responses measured during the SCE task are plotted against the

accommodative stimulus. For clarity, variability between the two runs is not shown. The

variability between the two runs across all subjects ranged from 0.15 D to 1.27 D, but for clarity

variability between the two runs is not shown.


81

The accommodative responses were variable among subjects and two subjects (EK

and PM) could not accommodate well for 4 D and 6 D accommodative stimuli.

Subject PM showed higher lag of accommodation than other subjects, but reported

that the target was clear and did not report any difficulty with the experiment. Two

subjects (EK and PM) who showed increase lag of accommodation in Figure 4.7

showed increased lag here too. Contrary to Figure 4.7, subject SK showed relatively

linear 1:1 stimulus-response function during the SCE task. The other three subjects

(JP, WF and YT) showed linear 1:1 stimulus-response function in both Figure 4.7

and Figure 4.8.

4.2.4 Changes in Stiles-Crawford effect with accommodation Table 4.2 summarizes the SCE and accommodation results for all myopic subjects by

showing the means and differences between two runs. The parameters are based on

fitting across a 6 mm pupil diameter.

Changes in ρx, ρy, xc and yc for all subjects are plotted against accommodative

response in Figure 4.9, together with linear fits. For subject EK, data for 8 D

accommodation stimuli were also included in the analysis. As for the emmetropes

(section 4.1.4) the change in a parameter for a particular subject, accommodation

stimulus and run was obtained by subtracting the parameter value, at the least

accommodative response for that subject, from the parameter. Results from both runs

are shown.


82

Table 4.2. Means (first entry) and differences (second entry) of parameter fits for two SCE runs

at different accommodative stimuli for myopes. AS and AR are accommodation stimulus and

accommodation response, respectively.

Subject AS (D) AR (D) ρx (mm-2) ρy ( mm-2) xc (mm) yc ( mm)EK 0 0.4, 0.2 0.114, 0.014 0.087, 0.017 -0.34, 0.09 -1.12, 0.50

2 1.8, 0.0 0.097, 0.003 0.095, 0.018 -0.13, 0.17 -1.13, 0.254 2.8, 0.1 0.112, 0.010 0.101, 0.006 -0.05, 0.16 -1.16, 0.356 4.3, 0.0 0.139, 0.004 0.097, 0.003 +0.12, 0.11 -1.50, 0.108 4.9, 0.8 0.165, 0.008 0.158, 0.006 +0.14, 0.16 -0.85, 0.23

PM 0 0.1, 0.0 0.161, 0.000 0.120, 0.004 +0.19, 0.29 -0.47, 0.202 0.9, 0.5 0.167, 0.007 0.158, 0.018 +0.11, 0.18 -0.38, 0.224 2.1, 0.7 0.176, 0.008 0.188, 0.037 -0.05, 0.11 -0.31, 0.096 1.6, 0.7 0.242, 0.037 0.235, 0.011 -0.01, 0.04 -0.15, 0.30

JP 0 1.5, 0.1 0.106, 0.038 0.126, 0.001 +0.26, 0.36 -0.05, 0.132 2.7, 0.7 0.128, 0.006 0.133, 0.016 +0.36, 0.09 -0.13, 0.004 4.9, 0.2 0.135, 0.009 0.146, 0.033 +0.16, 0.17 +0.03, 0.166 6.7, 0.1 0.160, 0.012 0.174, 0.015 +0.29, 0.25 +0.05, 0.33

WF 0 1.3, 06 0.122, 0.003 0.113, 0.016 -0.06, 0.16 +0.53, 0.242 2.6, 1.5 0.118, 0.006 0.123, 0.027 -0.09, 0.03 +0.98, 0.254 4.4, 0.2 0.126, 0.001 0.115, 0.004 +0.06, 0.08 +1.04, 0.136 6.1, 0.2 0.130, 0.030 0.128, 0.013 +0.16, 0.08 +0.84, 0.11

YT 0 1.6, 0.8 0.126, 0.008 0.132, 0.005 -0.03, 0.05 -0.85, 0.122 2.8, 0.3 0.145, 0.009 0.099, 0.008 +0.03, 0.08 -0.86, 0.344 5.3, 0.3 0.121, 0.014 0.131, 0.011 +0.11, 0.24 -0.44, 0.036 5.7, 0.2 0.117, 0.028 0.129, 0.006 +0.01, 0.30 -0.27, 0.20

SK 0 1.5, 0.5 0.113, 0.001 0.084, 0.033 +0.32, 0.02 +0.36, 0.012 2.1, 0.2 0.137, 0.002 0.115, 0.023 +0.32, 0.09 +0.49, 0.294 4.8, 0.2 0.123, 0.029 0.119, 0.000 +0.50, 0.32 +0.56, 0.236 5.8, 0.3 0.140, 0.039 0.123, 0.009 +0.40, 0.30 +0.65, 0.10


83


-1 0 1 2 3 4 5 6 7

Ch

an

ge

in

x

(m

m-2

)

-0.08

-0.06

-0.04

-0.02

0.00

0.02

0.04

0.06

0.08

0.10

0.12

0.14EK

PM

JP

WF

YT

SK

y = 0.004*x + 0.011r ² = 0.067; p = 0.10


-1 0 1 2 3 4 5 6 7

Ch

ang

e i

n

y (

mm

-2)

-0.08

-0.06

-0.04

-0.02

0.00

0.02

0.04

0.06

0.08

0.10

0.12

0.14y = 0.004*x + 0.003r ² = 0.035; p = 0.19


-1 0 1 2 3 4 5 6 7

Ch

an

ge

in

xc

(mm

)

-1.5

-1.0

-0.5

0.0

0.5

1.0

1.5y = -0.054*x + 0.123r ² = 0.114; p = 0.02

Accommodation Response (D)-1 0 1 2 3 4 5 6 7

Ch

ang

e i

n y

c (

mm

)

-1.5

-1.0

-0.5

0.0

0.5

1.0

1.5 y = -0.039*x + 0.008 r ² = 0.043; p = 0.11

N = 6A B

C D

Figure 4.9. Changes in ρx (A), ρy (B), xc (C) and yc (D) are plotted against the accommodative

response for myopes. Pupil size is 6 mm. In all the graphs, each subject is represented by a

different symbol. The solid lines in the plots are the linear fits to the data.

The regression analysis showed that the directionality increased along both

horizontal and vertical meridians, but that neither of the slopes was significantly

different from zero. Only peak pupil position along the horizontal meridian (xc)

showed a significant change; this was in the temporal direction in the pupil with

increasing accommodation (slope = -0.054 mm/D, r2 = 0.11, p = 0.02). As for

emmetropes, linear mixed model was applied to take into account the lack of

independence of observations, and again there were negligible changes to the

probabilities.


84

4.3 Changes in Stiles-Crawford effect with accommodation for combined groups

Although SCE directionality changed significantly with accommodation for

emmetropes but not for myopes, the rates of change for both ρx and ρy with

accommodation were similar in both groups (Table 4.3). Therefore, the data from

both emmetropic and myopic groups for both pupil sizes were combined to

determine the overall change in SCE with accommodation (Table 4.3, last column,

Figure 4.10).

With regression analysis, both ρx and ρy increased significantly with increasing

accommodation (ρx slope = +0.003 mm-2/D, r2 = 0.07, p = 0.01, ρy slope =

+0.005mm-2/D, r2 = 0.11, p = 0.001). Peak location did not change significantly with

accommodation in either horizontal or vertical meridians (Figure 4.10). When a

linear mixed model was applied to take into account the lack of independence of

observations, there were negligible changes to the probabilities.


85

Accommodative Response (D)

-1 0 1 2 3 4 5 6 7 8 9

Ch

ang

e in

x

(mm

-2)

-0.08

-0.06

-0.04

-0.02

0.00

0.02

0.04

0.06

0.08

0.10

0.12

0.14

MyopesEmmetropes

y = 0.003*x + 0.011r ² = 0.067; p = 0.01

Accommodative Respose (D)-1 0 1 2 3 4 5 6 7 8 9

Ch

ang

e in

y

(m

m-2

)

-0.08

-0.06

-0.04

-0.02

0.00

0.02

0.04

0.06

0.08

0.10

0.12

0.14y = 0.005*x + 0.001r ² = 0.111; p = 0.001

Accommodative Response (D)-1 0 1 2 3 4 5 6 7 8 9

Ch

ang

e in

xc

(mm

)

-1.5

-1.0

-0.5

0.0

0.5

1.0

1.5y = -0.018*x + 0.048r ² = 0.021; p = 0.15

Accommodative Response (D)

-1 0 1 2 3 4 5 6 7 8 9

Ch

ang

e in

yc

(mm

)

-1.5

-1.0

-0.5

0.0

0.5

1.0

1.5y = -0.015*x - 0.010r ² = 0.014; p = 0.24

N = 12A B

C D

Figure 4.10. Changes in ρx(A), ρy, (B), xc (C) and yc (D) are plotted against the accommodative

response in panels A, B, C and D, respectively. Pupil size is 6 mm. Filled circles represent

emmetropes and open circles represent myopes. The solid lines are the linear fits to the

combined data of emmetropes and myopes.

Table 4.3. Comparison of rates of changes in ρx, ρy, xc and yc with accommodation for

emmetropes, myopes and combined data of emmetropes and myopes for a 6 mm pupil.

Significant values are indicated with asterisks.

Slopes Emmetropes Myopes Combined

ρx (mm-2/D) +0.003* +0.004 +0.003*

ρy (mm-2/D) +0.005* +0.004 +0.005*

xc (mm/D) +0.006 -0.054* -0.018

yc (mm/D) -0.003 -0.039 -0.015


86

4.4 Peak-finding technique This section presents the results for the SCE function peaks of emmetropes obtained

with the peak-finding technique of Blank, Provine & Enoch (1975), described in

section 3.9, and compares them with the results obtained with the main SCE

technique.

Subjects

AK AM ST EM NS

Pu

pil

pea

k (m

m)

-1.0

-0.8

-0.6

-0.4

-0.2

0.0

0.2

0.4

0.6

0.8

1.0

UnaccommodatedAccommodated

Nasal

Temporal

Figure 4.11. Mean of the SCE peak pupil locations from both directions for 0 D

(unaccommodated) and 6 D (accommodated) accommodative stimuli. Dashed line indicates the

reference position in the pupil. Error bars represent standard deviations of the peak settings.


87

Table 4.4. Mean ± SD of the SCE peak pupil locations obtained from both directions for 0 D

(unaccommodated) and 6 D (accommodated) accommodative stimuli and the resulting peak

shift ± SD. Statistics using t-tests are shown in the fifth and sixth columns

Subject Unaccommodated

(mm) Accommodated

(mm) Peak Shift

(mm) t p-values

AK -0.46 ± 0.25 -0.09 ± 0.24 +0.37 ± 0.24 3.73 0.001

AM +0.38 ± 0.28 +0.14 ± 0.17 -0.24 ± 0.23 -3.31 0.003

ST +0.23 ± 0.07 +0.27 ± 0.13 +0.04 ± 0.10 0.88 0.390

EM -0.33 ± 0.20 -0.15 ± 0.22 +0.18 ± 0.21 1.67 0.121

NS +0.16 ± 0.17 0.00 ± 0.14 -0.16 ± 0.16 -1.75 0.112

The peak location settings for five subjects are shown for 0 D and 6 D

accommodation stimuli in Figure 4.11 and Table 4.4. Two subjects showed shifts in

the nasal direction in the pupil with accommodation, while three showed shifts in the

temporal direction. All the shifts were < 0.40 mm, and only two subjects showed

significant shifts, AK in the nasal direction and AM in the temporal direction. These

significant shifts were in the same direction to those in the main experiment,

although the absolute values were larger (compare Table 4.4 with Table 4.1). The

peak shifts of the SCE across the group were small and not systematic. This supports

the main study which found no significant changes in peak of the SCE with

accommodation for emmetropes (section 4.1.4).

4.5 Aberration measurements of emmetropes This section illustrates the results of aberration measurements of emmetropes

obtained using the COAS-HD instrument (section 3.10).


88

Figure 4.12 shows the accommodative responses of emmetropic subjects for different

accommodative stimuli. The results show good accuracy for most subjects for the

range of accommodation stimuli.


-2 0 2 4 6 8 10

Ac

com

mo

dat

ion

Res

po

nse

(D

)

-2

0

2

4

6

8

10AMNSLSEMAKST

Figure 4.12. The accommodative responses versus accommodation stimulus for six subjects. The

dashed line represents an ideal 1:1 relationship.

Figure 4.13 shows the higher order RMS aberrations (third to sixth order) as a

function of accommodation response. Although large individual variations occurred,

the higher order RMS increased significantly with accommodation (slope = +0.020

μm/D, r2 = 0.30, p = 0.004). The higher order RMS increased significantly with

accommodation for three subjects (AM, LS and AGK) (r2 ranged from 0.94 to 0.98,

p < 0.05).


89


-2 0 2 4 6 8 10

Hig

he

r o

rde

r R

MS

(m

)

0.0

0.1

0.2

0.3

0.4

0.5

0.6AM NS LSEMAKST

y = 0.020*x + 0.179r ² = 0.30; p = 0.004

Figure 4.13. Higher-order RMS aberrations plotted against the accommodation response for a 5

mm pupil. Each data point represents the average of two measurements. The line is the linear

regression fitted to all data.


90


-2 0 2 4 6 8 10

Sp

he

rica

l ab

err

ati

on

co

effi

cien

t (

m)

-0.5

-0.4

-0.3

-0.2

-0.1

0.0

0.1

AMNS LSEMAKST

y = -0.036*x + 0.020r ² = 0.66; p < 0.0001

Figure 4.14. Spherical aberration coefficients 04C as a function of accommodation response for 5

mm diameter pupil. Each data point represents the average of two measurements. The line is

the linear regression fitted to the cumulative data.

Spherical aberration coefficient 0

4C decreased with accommodation and became

negative for all the subjects (Figure 4.14). The rate of change of 04C was linearly

related to the accommodative response for the subject group (slope = - 0.036 µm/D,

r2 = 0.66, p < 0.0001). There were no significant changes for individual third-order

aberration coefficients as a function of accommodation.

To summarize, this section has investigated changes in higher-order aberrations with

accommodation in a group of six emmetropic subjects. Similar to other studies

(Cheng, Barnett, Vilupuru, Marsack, Kasthurirangan, Applegate & Roorda, 2004;

Ninomiya, Fujikado, Kuroda, Maeda, Tano, Oshika, Hirohara & Mihashi, 2002;

Plainis, Ginis & Pallikaris, 2005), increase in higher order RMS aberrations and

negative changes in spherical aberration as accommodation increased have been

demonstrated.


91

4.6 Discussion

4.6.1 Changes in the SCE directionality with accommodation The changes in the foveal SCE were investigated with increase in accommodation

stimuli using psychophysical techniques. A significant increase in SCE directionality

was found in both horizontal and vertical meridians with increase in accommodation

for fits to 6 mm pupil in emmetropes (Figure 4.5). For myopes, the SCE

directionality did not change significantly with accommodation along either

horizontal or vertical meridian for fits to a 6 mm pupil (Figure 4.9).

When the SCE data of emmetropes and myopes are combined, significant changes in

directionality were found in both horizontal and vertical meridians with

accommodation (ρx slope = +0.003 mm-2/D and ρy slope = +0.005 mm-2/D),

corresponding to a modest increase in directionality of 15 - 25 % at 6 D

accommodation (Figure 4.10).

In general, subjects accommodated close to the desired amount but two myopic

subjects demonstrated considerable lags for higher accommodative stimuli (Figure

4.8). Additional measurements showed that aberrations and accommodative lag

affect the SCE (section 3.8.4). Spherical aberration increased with accommodation in

emmetropic subjects (section 4.5) in the same direction as accommodative lag and

was likely to have influenced the SCE, although the magnitude of this effect is not

known. This indicates that some of the increase in SCE directionality with

accommodation is due to blur increasing the threshold of peripheral pupil points.

4.6.2 Changes in the SCE peak location with accommodation Although xc changed significantly in myopes with 6 D accommodation (6 mm pupil),

overall there was no significant shift in peak location (Figure 4.10). The results of


92

this study conflict with those of Blank, Provine & Enoch (1975) who found a

substantial nasal shift of up to 1.5 mm with 9 D accommodation stimulus in three

subjects.

Due to the difference in measuring techniques between the main study and Blank et

al’s study, I conducted a peak-finding study similar to that of Blank et al. (section

3.9). Although two of five subjects showed significant shifts of the horizontal

component of the peak with a 6D accommodation stimulus, these were less than 0.4

mm and there was no significant pattern across the group (section 4.4).

There were two major differences between the main study and the study of Blank et

al.. Firstly, they used high-power soft contact lens to help stimulate accommodation

whereas I used an optical trombone. A pilot study in section 3.8.4 showed that

contact lenses can change the SCE (although this would seem to be more marked for

positive than negative lenses), so it is possible that the high power lenses might have

caused artefactual changes in the SCE, particularly if the lenses were not well

centred. Secondly, Blank et al.’s 9 D stimulus would have been more effective than

a 6 D stimulus in causing peak shifts if horizontal retinal stretching really occurs in

accommodation and is greater in the temporal than in the nasal retina.

In view of results from the main experiment and the peak–finding technique, I

conclude that the SCE peak sensitivity location changes little and not systematically

with accommodation for up to 6 D accommodation stimulus in young subjects.

4.6.3 Comparison of SCE with accommodation between emmetropes and myopes

The SCE became steeper in both emmetropes and myopes with accommodation with

significant increase of SCE directionality in emmetropes (Figure 4.5). The rates of

change of directionality in the two groups were similar along both horizontal and

vertical meridians for a 6 mm pupil (Table 4.3). The peak location did not show any

systematic shift with accommodation in either emmetropes or myopes.


93

4.7 Summary 1. There were increases of up to 15-25% in foveal SCE directionality with

accommodation up to 6D in horizontal and vertical meridians. However, additional

experiments described in section 3.8.4 indicated that at least some of the increase

was an artefact of higher order aberrations and accommodative lag rather being a true

effect of accommodation.

2. Contrary to the study of Blank, Provine & Enoch (1975), both the main study and

peak-finding technique found small and non-systematic shifts in SCE peak

sensitivity with 6 D accommodation. The actual shift in peak positions was small

with the biggest shift being approximately 0.35 mm.

3. Both emmetropes and myopes tend to have a similar change in directionality per

dioptre of accommodation along both horizontal and vertical meridians. The shift in

foveal SCE peak location with accommodation was also comparable between both

the groups.

4. The psychophysical technique for measuring the SCE was very time consuming

and demanding for subjects, and this could have affected results if subjects tired,

particularly for the higher accommodative stimuli. The following study, using the

objective technique, is directed to overcome this limitation.

Chapter 5 SCE using multifocal electroretinogram

94

Chapter 5

The Stiles-Crawford effect using the multifocal electroretinogram

5.1 Introduction As stated earlier, the psychophysical technique of measuring the Stiles-Crawford

effect (SCE) was lengthy and very demanding for the subjects. Recently, objective

techniques such as fundus reflectometry, adaptive optics ophthalmoscopy, and

electrophysiology have been used to measure SCE faster than the psychophysical

technique (Burns, Wu, Delori & Elsner, 1995; Roorda & Williams, 2002; Sutter,

1997).

Sutter (1997) introduced an objective technique of measuring the SCE using the

multifocal electroretinogram (mfERG). The local responses of the retina (SCE) were

obtained in a matter of minutes with this technique (Sutter, 1991; Sutter & Tran,

1992). The method of measuring the SCE using the mfERG is based on the principle

of the mfERG, but instead of imaging the multifocal stimulus on the retina, the

stimulus is imaged in the pupil plane. Sutter suggested that it is quicker and less

demanding than the psychophysical technique but no quantitative comparison has

been done between the two techniques.

In order to avoid subjective factors and experiment duration from obscuring subtle

accommodation related changes in SCE, an electrophysiology based technique,

similar to that described by Sutter (1997) was attempted. The main objectives of this

study were a) to investigate the changes in the SCE with accommodation and b)


95

compare subjective (psychophysical) and objective measurements of the SCE. The

SCE was measured at two accommodative states – 0 D and 6 D in six emmetropic

subjects. Myopes were not included in this study due to the higher variability and

higher lags in accommodative responses seen in this group in Chapter 4.

5.2 Methodology

5.2.1 Subjects Six emmetropic subjects who participated in the psychophysical experiment (Chapter

4) were recruited for this study. The research was approved by the Queensland

University of Technology Human Research Ethics committee.

5.2.2 Apparatus

The Stiles-Crawford effect was measured with a multifocal ERG (VERISTM Science

5.2X, Electrodiagnostic Imaging Inc., Redwood City, CA, USA), viewed through a

Maxwellian-viewing system (Figure 5.1). The multifocal stimulus was presented on

a LCD display M2. The multifocal stimulus was imaged at the pupil rather than at the

retina by the mirror, relay lenses L1 and L2 and the Badal lens L3. The retinal

illuminance was approximately 8770 trolands (tds) (the method for determining

retinal illuminance is given in section 5.4.2). An 18 mm diameter aperture A1 with a

cross-hair target served as a fixation target and limited the field of view to 13.8°.

Accommodation was stimulated by moving A1 axially. The stimulus array was

minified 4.3 times on to the pupil plane. A firewire camera (PixeLINK) with a 55

mm lens (Edmund Scientific 52-271) viewed a subject’s eye through a hot mirror BS

(transmits visible, reflects infrared). The camera was connected to a laptop computer

M1 through the FireWire IEEE-1394 Repeater port (Toshiba Corporation). The pupil

image displayed on the computer monitor M1 was used for ensuring pupil alignment.

The camera also acted as a photorefractor to measure refraction and calculate

accommodation. This custom-photorefractor was equipped with twenty infrared light


96

emitting diodes (LEDs) (wavelength 890 nm) covering the lower half of the camera

lens.

A custom-photorefractor rather than the PowerRef II (used in the psychophysical

experiment) was used primarily because of space considerations and to monitor pupil

centration. The PowerRef II has a bulky computer mounted camera and needs a 1m

distance whereas the custom-photorefractor is a compact system that can be placed

closer to the apparatus. Pupil centration with respect to the apparatus can be carefully

monitored with the higher resolution camera of the custom-photorefractor but this

could not be done with the PowerRef II.

Figure 5.1. Schematic representation of the apparatus using multifocal electroretinogram. The

multifocal display on monitor M2 is reflected from a mirror towards the eye through a pair of

relay lenses L1 and L2 and Badal lens L3. Lenses L1, L2 and L3 have the same focal length (75

mm). An aperture A1 (18mm subtending 13.8 degrees at the eye) with a cross target is used in

the light path to avoid stray light and present a fixation target. A1 can be moved axially to

stimulate accommodation. During alignment of the apparatus an aperture with frosted glass

(A2) was used in the place of the pupil of the eye. During the experiment, pupil alignment and

refraction was monitored with a photorefractor focussed on the eye through a hot mirror

beamsplitter BS, which reflects infrared and transmits visible light.

L1L3L2

Movement to stimulate accommodation

Multifocal display – LCD M2

Eye

A1

Photorefractor and alignment camera

BS

Computer monitor M1

Mirror

Frosted glass with an aperture (temporarily placed at the position of eye during alignment of the apparatus)

A2

Reticule used for aligning the pupil

L1L3L2

Movement to stimulate accommodation

Multifocal display – LCD M2

Eye

A1

Photorefractor and alignment camera

BS

Computer monitor M1

Mirror

Frosted glass with an aperture (temporarily placed at the position of eye during alignment of the apparatus)

A2

Reticule used for aligning the pupil


97

5.2.3 MfERG stimulus

The visual stimulus array was displayed on the LCD monitor M2 (Viewsonic,

Viewsonic Corporation, Walnut, CA) driven at a frame rate of 75 Hz and consisted

of 61 equal-sized black and white hexagons (Figure 5.2). Several aspects of the

monitor were assessed such as voltage-luminance relationship, warm-up

characteristics, spatial variance, spectral characteristics (See Appendix 2 for details).

The size of the array on the monitor was 156 mm horizontally and 148 mm

vertically. Each hexagon was temporally modulated between black and white

according to a pseudorandom binary m-sequence (213 – 1 steps in length). The

luminance of the monitor after 60 min of warm-up was 280 cd/m2 for the white

hexagons and 2 cd/m2 for the black hexagons (measured with an optiCALTM

photometer, Cambridge Research Systems) (Appendix 2).

Figure 5.2. Picture of the stimulus array used to elicit mfERG responses. The stimulus array

consisted of 61 equal-sized hexagons.

5.2.4 Experimental procedure A frosted glass with an aperture A2 was placed at the position of the pupil to check

the alignment of the apparatus (Figure 5.1). The mfERG stimulus was imaged on the

frosted glass and A2 was opened to the size of the stimulus array image and viewed


98

on the monitor M1. A transparency with a reticule on it matching the size of the

aperture was placed on the monitor. At the start of each experiment, the alignment of

the eye was ensured by placing the transparency on the monitor to coincide with the

view of the aperture; during the experiment, the eye was located in this position. The

mfERG stimulus size as measured on the frosted glass was 8.5 mm horizontally and

7.3 mm vertically. The mfERG display monitor was warmed-up for at least 50

minutes before beginning an experiment.

In the experimental session, 1 drop of 0.4 % benoxinate was instilled in the subject’s

right eye, followed by 1 drop of 2.5% phenylephrine. An additional drop of 2.5%

phenylephrine was instilled after one hour if the pupil size was less than 9 mm. Pupil

diameter of 8–10 mm is desired for improved mfERG signal-to-noise ratio

(Gonzalez, Parks, Dolan & Keating, 2004; Keating, Parks & Evans, 2000). The

subject’s dental impression was moulded on a bite bar. Once the mould was set and

the subject was comfortable, the bite bar was mounted on a XYZ movement

controller and attached to the apparatus. The subject was asked to bite lightly on the

bite-bar so that the influence of the temporalis muscle (which moves while biting)

was minimal. After subject alignment, the room illumination was reduced to help

maintain a large pupil. Subjects were asked to maintain clear fixation at the centre of

the cross hair target and not to blink for the duration of each segment of mfERG

recording (13.37 seconds). Pupil size was monitored by centring the pupil’s image on

the monitor reticule with respect to the corneal reflex. A pupil diameter of at least 9

mm was required to image the whole mfERG stimulus (see section 5.3 for an

explanation of how I dealt with pupils smaller than this). Continuous monitoring of

the subject’s pupil alignment was not possible as the light from the photorefractor

infrared LEDs interfered with the mfERG signal. Therefore, alignment was checked

after every fourth segment measurement of mfERG. The same pupil image captured

to check for alignment was used to assess accommodation.

After maximum dilation of the pupil, the skin areas for electrode placement were

cleaned with single-use alcohol swabs. Gold skin electrodes were filled with

ultrasonic gel and attached to the cleaned sites with adhesive tape. The reference

electrode was positioned near the temporal orbital rim and the ground skin electrode


99

was placed on the forehead. Unlike Sutter (1997) who used Burian-Allen contact lens

electrodes, I used DTL electrodes which give poorer signal-to-noise ratio

(Esakowitz, Kriss & Shawkat, 1993) and lower amplitude responses (Coupland,

1991; Esakowitz, Kriss & Shawkat, 1993), but are more comfortable to the subjects

for extended recording periods. DTL electrodes have been successfully used in the

past to record mfERG in patients with various diseases of the retina (Feigl, Brown,

Lovie-Kitchin & Swann, 2004; Kretschmann, Bock, Gockeln & Zrenner, 2000). The

DTL-electrode (active electrode) was positioned inside the upper margin of the lower

eyelid. All three electrodes were connected to their respective inputs of the amplifier.

The impedance between the skin electrodes (reference and ground) was less than 10

K for each mfERG recording. Amplifier gain was x 100,000 and filters were set

from 10 Hz to 300 Hz. The SCE was measured for 0 D and 6 D accommodative

stimuli. One subject (LS) could accommodate for an 8 D stimulus and so an 8 D

instead of 6 D stimulus was used for this subject. Two runs were performed for each

accommodative stimulus. SCE measurement was obtained for 0 D accommodation

stimulus followed by 6 D stimulus, and repeated for 0 D and 6 D accommodative

stimuli in the same order. Three subjects (AM, LS and NS) performed all four

measurements in one session, subjects AGK and EM in two days (two measurements

each day), and subject ST in four days (one measurement each day). Each SCE

measurement was divided into 32 segments, lasting 13.37 seconds each. ERG

segments with noticeable artefacts due to blinks or small eye movements were

detected online, discarded and immediately re-recorded. The mean accommodation

level of eight measurements was calculated as the accommodative response during

one SCE measurement. Total recording time for a run including capturing of pupil

images was approximately 10-15 minutes.

5.2.5 Calibration of the Photorefractor The photorefractor calibration procedure was similar to that performed in the

psychophysical experiment (see section 3.6) except that a visible blocking, infrared

pass filter (Kodak Wratten Filter #89B, high pass at 700 nm) was placed in front of

the right eye to prevent it from seeing the left eye fixation target while allowing the

photorefractor to measure refraction through the filter. Ophthalmic trial lenses from


100

+8 D to -6 D in 1 D steps were placed in front of the right eye in a trial lens frame.

Pupil images of the right eye obtained through the trial lenses were analysed with

custom software written in Matlab (The Mathworks Inc., Natick, MA, version

R2008a). Refraction induced by the trial lenses was calculated after compensating

for a 16 mm vertex distance. The quality of the pupil images were evaluated

subjectively. Pupil brightness profiles with excessive noise and excessively bright or

dark regions were excluded from analysis. The relation between pupillary brightness

profile and induced refraction was described with a linear regression function. This

linear regression was used to convert pupillary brightness profiles measured during

SCE measurement to refraction. Accommodative response was calculated by

subtracting refraction for a particular accommodative stimulus from 0 D refraction.

Change in refraction instead of raw refraction values was considered as

accommodation in order to eliminate any bias in baseline refraction measurements by

the photorefractor (such as due to the use of infra-red light and small uncorrected

refractive errors). The photorefractor was not calibrated against a clinical

autorefractor or subjective refraction as the purpose was not to measure absolute

refraction but only relative changes in refraction with accommodation. Therefore, the

photorefractor as used here is suitable for accommodation measurements and not

necessarily for baseline refractive error measurements. Results are presented in

section 5.5.1.

5.2.6 Accommodative stimulus-response function

Accommodative stimulus-response function was measured with a similar procedure

to that described in section 3.7 for the psychophysical experiment. Electrodes were

removed from the subject’s eye and face. Stimuli for accommodation were provided

by moving aperture A1 from 0 D to 8 D positions in 1 D steps. The left eye of the

subject was occluded and the right eye used for fixation of the accommodative target.

At each accommodative stimulus, a pupil image was obtained with the

photorefractor. Results of this experiment are described in section 5.5.2.


101

5.3 mfERG data analysis

The first negative and first positive deflections in the first-order mfERG response

(Figure 5.3) are termed N1 and P1, respectively. The amplitude of N1 was measured

from baseline to the first negative peak. The amplitude of P1 was measured from the

baseline to the first positive peak. The trough-to-peak amplitude (N1P1-amplitude)

was measured from the first negative peak to the first positive wave peak (Figure

5.3). Implicit time was measured from stimulus onset to the negative peak and the

positive peak (N1 and P1). To characterize the SCE from mfERG responses, N1 and

N1P1 amplitudes (nV) and N1 and P1 implicit times (ms) across pupil location were

considered because both N1 and P1 have contributions from the photoreceptors (see

Section 2.1.6.2 for details).

Figure 5.3. Schematic diagram of a first order mfERG response. Amplitude and implicit time of

the waveform are labelled.

The mfERG responses were noisy and variable in amplitude from hexagon to

hexagon as shown in Figure 5.4a. Major sources of noise were poor electrode

contacts, poor grounding or electrical interference from other equipments. Noise

N1 - amplitude

P1 - amplitude

N1P1 - amplitude

N1 – implicit time

P1 – implicit time

0 80 ms

20 nV

N1 - amplitude

P1 - amplitude

N1P1 - amplitude

N1 – implicit time

P1 – implicit time

0 80 ms

20 nV

0 80 ms

20 nV


102

reduction was achieved with better electrode contact, grounding, use of proper digital

filters such as power line filter (50 Hz) and low-pass filter (49 Hz) to the obtained

responses. The filters minimized the noise without distorting the signal (Komáromy,

Brooks, Dawson, Källberg & Ollivier, 2002; Marmor, Hood, Keating, Kondo,

Seeliger & Miyake, 2003).

The 80 ms signal response was analysed after two iterations of artefact removal as

recommended by VERISTM Science and ISCEV (International Society for Clinical

and Electrophysiology of Vision) standards. The two runs for each stimulus and

subject were combined using VERISTM software and the combined data were filtered

using digital filters (Figure 5.4b). The filtered data were exported to MS Excel

2003TM (Microsoft Corporation).

Figure 5.4. MfERG responses of the right eye of a subject AM showing (a) raw data (without

filtering), and (b) filtered data.

Since the size of the stimulus image on the pupil was 8.5 mm horizontally and 7.3

mm vertically, a pupil size of at least 9 mm was required for the whole stimulus to be

imaged on the pupil. For 6 D accommodation stimulus, the pupil constricted to 7 mm

to 8 mm for some subjects therefore, the mfERG responses were analysed across a 7

mm pupil instead of 9 mm pupil for all runs of both accommodation levels. To obtain

(a) (b)


103

mfERG responses for a 7 mm pupil, the outermost ring of the mfERG data was

excluded (shown by the dashed line in Figure 5.5).

To obtain the Stiles–Crawford (S–C) function, i.e. ERG response across pupil

location, an analysis based on ring averaged waveform was used rather than an

analysis based on each waveform. The local responses can be low in signal-to-noise

ratio and therefore, signal averaging methods might be a better approach to enhance

the signal of interest while minimizing extraneous noise (Parks, Keating, Evans,

Williamson, Jay & Elliott, 1997). mfERG data were grouped into concentric rings,

with ring 1 representing the peak response waveform (maximum N1P1 amplitude)

and rings 2-5 corresponding to the successive annuli of stimulation in the pupil. Most

of the mfERG data did not have a complete ring of hexagons and the number of

hexagons in peripheral rings was not uniform from subject-to-subject as the peak

response position varied across subjects. For example in Figure 5.5, ring 4 and ring 5

shown by white and black hexagons respectively are not complete rings and for

analysis an average of the hexagons present in an incomplete ring was considered.

To know the position (co-ordinates) of the hexagons in the pupil, each hexagon

position was translated to its respective location in the pupil of 7 mm diameter. The

hexagon position with the waveform of maximum N1P1 amplitude was considered to

be the SCE peak location in the pupil.

Figure 5.5. Diagram of the stimulus array showing peak hexagon as ring 1. Successive rings

were rings of the hexagons around the peak hexagon shown in black and white colour rings. The

outermost ring of hexagons in white (indicated with dashed line) was excluded from the analysis.

2 3 4 51 2 3 4 51 2 3 4 51 2 3 4 51


104

The directionality of the SCE was obtained for both accommodation levels (0 D and

6 D) by fitting individual Gaussian functions to N1 and N1P1 amplitudes of ring

average waveforms. As there were only five rings plotted across the peak response,

to perform the Gaussian fit the data points (amplitudes) were duplicated and plotted

along the negative x-axis after shifting the peak response to 0. The Gaussian function

was2

0 )(* xxeay where ‘a’ is the peak sensitivity, ‘ρ’ is the directionality and x0

is the peak of the function (but not the actual peak in the pupil).

5.4 Control measurements

5.4.1 Effect of different sizes of aperture A1 on the measured SCE using mfERG

Aperture A1 controls the angular subtense of the retinal field and thus the area of the

retina stimulated by the mfERG stimulus varies with change in aperture A1 size. To

select an aperture size and stimulus size to test the smallest field size possible and at

the same time give good signals for the experiment, the SCE measurement using

mfERG recordings were performed for the right eye of two subjects AM and PG for

12 mm (9.2º) and 18 mm diameters (13.8°) of aperture A1. The procedure for testing

subjects’ right eyes was similar to that described in section 5.2.5. The mfERG

stimulus sizes used were 5.9 mm x 5.6 mm (small) and 8.5 mm x 7.3 mm (large) as

measured at the pupil plane. The stimulus size at the pupil was varied by changing

the size of the stimulus on the monitor using VERISTM software. Both the subjects

performed two runs for each aperture size with smaller stimulus and only subject AM

performed the experiment with the larger stimulus size. The two runs were combined

using VERISTM and the data were filtered.

The P1 implicit time (ms) and N1, N1P1 amplitudes (nV) of the peak response

waveform for the smaller stimulus size are shown in Figure 5.6. N1 and P1 implicit

times were longer with the 12 mm aperture compared with those for the 18 mm

aperture. The amplitudes for N1 and N1P1 were larger with the 18 mm aperture than


105

with the 12 mm aperture, but subject AM showed smaller N1 amplitude with the 18

mm aperture than with the 12 mm aperture.

N1 P1

Imp

lici

t ti

me

(mse

c)

0

10

20

30

40

50

60AM-12mm AM-18mm PG-12mm PG-18mm

N1 N1P1

Am

pli

tud

e (n

V)

0

10

20

30

40

50

60A B

Figure 5.6. Implicit time (plot A) of N1 and P1 and amplitudes (plot B) of N1 and N1P1 of the

peak response waveform with two aperture sizes, 12 mm and 18 mm for the smaller stimulus

size.

The ring average waveforms of subject AM for both the apertures and larger stimulus

size are shown in Figure 5.7. For panels A (12 mm aperture) and B (18 mm aperture)

in Figure 5.7, the peripheral rings were attenuated for the smaller stimulus size

whereas in panel C (18 mm aperture) with larger stimulus size the rings were more

defined (compare rings 4 & 5 between A, B & C). This suggests that the amount of

light entering the eye with the smaller stimulus size was not sufficient to excite the

retina and to get good signals across the pupil.


106

Figure 5.7. Filtered ring average plots of subject AM for (A) 12mm and (B) 18 mm aperture

sizes for a smaller stimulus size; (C) 18 mm aperture size for the larger stimulus size. All

waveforms are at the same scale.

To summarize, longer implicit times with 12 mm aperture (Figure 5.6) and

attenuation of peripheral rings with the stimulus size of 5.9 mm x 5.6 mm (also with

18mm aperture) (Figure 5.7) were found. Reduced retinal illuminance from the small

aperture and reduced field of view from the smaller stimulus (leading to fewer

photoreceptors being stimulated) most likely resulted in delayed responses with only

the central response being well defined. A larger stimulus size (8.5 mm x 7.3mm, on

the pupil plane) and aperture size of 18mm was efficient in eliciting good signal and

defined waveforms across the pupil and was used in this study.

A B CA B C


107

5.4.2 Retinal illuminance measurement

Retinal illuminance was measured to ensure that the mfERG responses to measure

the SCE were obtained from the cones (i.e. under photopic conditions). An indirect

method similar to that described in Chapter 3 (section 3.3.5) was used to measure

retinal illuminance. Three observers (DAA, NS and BB) compared the brightness of

the mfERG stimulus (not flickering) with an auxiliary light source (white fluorescent

light) placed to the side of the apparatus. An aperture stop was placed in front of the

auxiliary light source to produce a similar field of view to that used in the

experiment. Each subject’s right eye was dilated and the pupil was aligned as

described in section 5.2.4. Various combinations of ND filters were inserted in front

of the auxiliary light source in order to match the brightness of the stationary mfERG

stimulus with the source. The luminance of the auxiliary light source was measured

with a luminance colorimeter (Topcon BM-7, Tokyo, Japan) with a 1-degree field

and three measurements were taken for each combination of matching ND filters.

From the luminance, effective retinal illuminance was calculated for 8.5 mm pupil

size. An “effective retinal illuminance” takes account of the resulting changes in

response produced by the directional sensitivity of the retina (SCE). The effective

retinal illuminance was computed as

I (trolands) = luminance (cd/m2) x A’,

where I is the retinal illuminance and A’ is the effective area. A’ is the product of

pupil area and photometric efficiency, where photometric efficiency is calculated as

2/)1(2

re r and accounts for the SCE ( is the directionality of the SCE and

was taken as 0.12 mm-2 for the average population (Applegate & Lakshminarayanan,

1993)).

The mean retinal illuminance for the three subjects was 8770 trolands (tds) (Table

5.1), which is well beyond the rod saturation level of 400 tds (Hood & Finkelstein,

1986). Thus, the cones should be the only sensitive photoreceptors at this level.


108

Table 5.1. Measurements of retinal illuminance from the mfERG stimulus for three subjects are

shown. ND filter match is the combination of ND filters ± range.

Subject ND filter match

Mean of three luminance

measurements

(cd/m2)

Retinal illuminance for 8.5 mm pupil

(tds)Retinal illuminance

(log tds)DA 1.3 ± 0.15 410 9507 4.0NS 1.5 ± 0.15 278 6442 3.8BB 1.15 ± 0.1 447 10360 4.0

Mean 8770 3.9

5.4.3 Comparison of mfERG measurements with LCD and CRT monitors

An LCD monitor to display the mfERG stimulus was used in this study whereas

CRT monitors are used conventionally to measure the mfERG responses. Keating,

Parks, Malloch & Evans (2001) compared mfERG response from CRT vs LCD

displays and suggested that CRT systems are better for examining non-linear aspects

of the multifocal response whereas LCD systems can separate onset and offset

components. A detailed calibration of the LCD monitor (Viewsonic VX 922,

ViewSonic Corporation, Walnut, CA) performed for the purposes of this study (see

Appendix 2, section A2.4) showed that the monitor can be used for display of the

achromatic mfERG stimulus when used within its operating limits. Consistent with

other studies (Brainard, Pelli & Robson, 2002; Feigl & Zele, 2008; Keating, Parks,

Malloch & Evans, 2001), it was found that this LCD monitor reaches its maximum

approximately 5.5 ms from the start of the video signal, whereas a CRT provides a

short pulse very soon after the onset of the video frame. Therefore, the mfERG

recordings with the LCD and a CRT as stimulus delivery systems were compared to

determine if there is any delay in the multifocal responses with the LCD.


109

Two mfERG recording sessions each with the LCD (used in this study - Viewsonic,

75 Hz frame rate, 280 cd/m2) and a CRT (FIMI MD0709BRM, a 21” monochrome

display, 75 Hz frame rate, 100 cd/m2) systems were performed for one subject (AM)

without the optical system, that is, the mfERG stimulus was imaged on the retina.

The pupil was dilated using the same protocol as for 5.2.4. The mfERG stimulus size

used in the main experiment (8.5 mm x 7.3 mm) was used here. Two runs were

combined using VERISTM software and the outermost ring was excluded. The N1 and

P1 implicit times showed a mean delay of 6.4 ms and 6.7 ms with the LCD compared

to a CRT display across all the four rings. Table 5.2 shows N1 and P1 implicit times

of rings 1-4 measured with the LCD and CRT displays.

Table 5.2. N1 and P1 implicit times (ms) of rings 1-4 measured with the LCD and CRT displays.

N1 – implicit time (ms) P1 – implicit time (ms)

Ring LCD CRT Difference LCD CRT Difference

1 20.8 14.2 6.6 36.7 30.0 6.7

2 20.8 14.2 6.6 35.0 28.3 6.7

3 20.0 13.3 6.7 34.2 27.5 6.7

4 20.0 14.2 5.8 35.0 28.3 6.7

Therefore, the mfERG responses in this study are expected to be delayed by

approximately 6.5 ms with the LCD display in comparison to a CRT display.

Variations in the responses are expected because of differences in timing of the

delivery of the stimulus signal. However, it is assumed that same cells in the retina

produce the waveforms by either of the displays.


110

5.5 Results

5.5.1 Photorefractor calibration functions

Individual calibration functions for five subjects were linear with r2 > 0.80

(individual r2 ranged from 0.87 to 0.96) except for one subject ST (open squares, r2 =

0.61) (see procedure in section 5.2.5). Figure 5.8 shows combined data and the

combined calibration linear regression fit. Subject ST’s data were excluded from the

combined linear regression fit.

Slope of Pupillary Brightness Profile

-0.4 -0.3 -0.2 -0.1 0.0 0.1

Ind

uc

ed R

efr

acti

on

(D

)

-8

-6

-4

-2

0

2

4

6

8

AMLSNSAKEMST

Induced refraction = -28.34*Photorefractor slope - 5.24r ² = 0.88

Figure 5.8. Relationship between slope of pupillary brightness profile and trial lens induced

refraction for all subjects. Individual calibration functions were obtained by fitting a linear

regression equation to the data of each subject (not shown). A cumulative linear regression

(solid black line) fitted to all data except subject ST’s data has an r2 of 0.88 demonstrating

validity of the calibration function.


111

5.5.2 Accommodative stimulus-response functions Accommodative stimulus-response functions for all subjects are shown in Figure 5.9.

Changes in refraction instead of raw refraction values are plotted in order to

eliminate any bias in baseline refraction measurements by the photorefractor (such as

the use of infra-red light and small uncorrected refractive errors). With increasing

accommodative stimulus, most subjects demonstrated increasing accommodative

responses. Subject ST showed higher lag and noisier accommodative responses than

other subjects. He did not demonstrate a linear photorefractor calibration function

previously (see section 5.5.1). The noise in ST’s accommodative data could be a

combination of poor accommodation and invalid photorefractor calibration function.

The other subjects demonstrated a classical pattern of linear accommodative

responses at small stimulus amplitudes (< 4 D) followed by lag at higher stimuli (≥ 5

D).


-2 0 2 4 6 8 10

Acc

om

mo

dat

ion

Resp

on

se (

D)

-2

0

2

4

6

8

10

AMLSNSAKEMST

Figure 5.9. Accommodative stimulus-response data for all subjects. The results for different

subjects are represented as different symbols. The dashed line represents the 1:1 line. All other

subjects except for ST demonstrated increasing accommodative responses with higher stimulus

amplitudes.


112

5.5.3 Accommodation responses of subjects during SCE measurements

The mean of eight photorefractor measurements during each SCE recording was

considered as the refraction for a particular run. Accommodation was calculated as

described in section 5.5.2 i.e. by subtracting refraction measurement for an

accommodation stimulus from 0 D refraction for each subject individually.

Accommodative responses are plotted against accommodative stimuli for individual

subjects in Figure 5.10. For 6 D and 8 D (LS only) stimuli all subjects demonstrated

greater lags of accommodation than seen in Figure 5.10, possibly due to the longer

duration and mild discomfort during mfERG measurements. Subject ST

demonstrated the least accommodative response for a 6 D stimulus. Subject ST

consistently showed lowest accommodative responses in both Figures 5.9 and 5.10

and unacceptable photorefractor calibration. Due to the poor calibration function, the

refraction and accommodation data are unreliable for subject ST.


0 2 4 6 8 10

Acc

om

mo

dat

ion

Res

po

nse

(D

)

0

2

4

6

8

10

AM LS NS AK EM ST

Figure 5.10. Mean accommodative responses for subjects as a function of accommodative

stimulus during SCE recording. Each subject is represented by a different symbol. Error bars

represent differences between two runs.


113

5.5.4 SCE measurement using mfERG stimulus for

0 D and 6 D accommodation stimuli

5.5.4.1 Change in the SCE directionality with accommodation

Amplitudes of N1 and N1P1 for the ring average waveforms (mean of all the

hexagons for each ring) were calculated. In Figure 5.11 ring average waveforms

around the peak response of subject AM for 0 D and 6 D accommodative stimuli are

shown. It can be seen that N1 and N1P1 amplitudes decrease from ring 1 to ring 5.

Implicit time did not change with eccentricity or accommodation for any of the

subjects for either of the accommodation stimuli which suggests that the implicit

time is not influenced by the SCE.

Figure 5.11. Filtered ring average waveforms around peak response of subject AM for 0 D and 6

D accommodation, illustrating amplitudes (nV) and implicit times (ms) for N1 and N1P1. Ring 1

represents the peak of the SCE. The scales are different for enhanced illustration of all the

waveforms.

Amplitude (nV) Implicit time (ms)

N1

88.5 19.2

N1P1 155.4 33.3

0 D0 D6 D

N1

45 18.3

96.5 33.3

N1P1

N1

19.01 19.2

N1P1

42.4 35.0

N1

12.57 25.8

N1P1

35.1 49.2

N1

N1P1 64.3 34.2

105.14 18.3

207 33.3

55.88 19.2

131.4 33.3

32.84 20.8

47.1 35.0

0.58 22.2

5.31 33.3

24.8 21.7

64.2 35.0

Amplitude (nV) Implicit time (ms)

22.5 18.3


114

Figure 5.12 shows normalized mean amplitudes of five subjects for both N1 and

N1P1 from ring 1 towards ring 5, that is, from the centre to the periphery of the pupil

for both 0 D and 6 D accommodation stimuli. The normalization of amplitudes for

each subject was done by dividing each ring amplitude by the ring 1 amplitude. The

amplitude decreases from ring 1 outwards. As subject ST‘s refraction and

accommodation responses were not reliable (section 5.5.3), his data were excluded

from this analysis.

Statistical analysis of the data was conducted using the Statistical Packages for the

Social Sciences (SPSS Inc., version 15.0). Repeated measures analysis of variance

(ANOVA) was performed to determine if there were any differences in mfERG

responses (N1 and N1P1 amplitudes) with pupil eccentricity and accommodation.

Retinal eccentricity (concentric rings 1-5) and accommodation (0 D and 6 D) were

within-subject factors. ANOVA showed significant change in N1 and N1P1

amplitudes with ring eccentricity (p<0.0001) but not with accommodation (N1, p =

0.06; N1P1, p = 0.79). There was no significant interaction between accommodation

levels (0 D and 6 D) and mean of the rings for both N1 (p = 0.21) and N1P1 (p =

0.24) amplitudes.


115

EccentricityR1 R2 R3 R4 R5

No

rmal

ized

Am

plit

ud

e (n

V)

0.0

0.2

0.4

0.6

0.8

1.0 0 D6 DN1

EccentricityR1 R2 R3 R4 R5

No

rmal

ized

Am

plit

ud

e (n

V)

0.0

0.2

0.4

0.6

0.8

1.0 0 D6 DN1P1

Rings: F4,16 = 196.79, p < 0.0001

Accommodation: F1,4 = 0.08, p = 0.79

Rings: F4,16 = 25.89, p< 0.0001

Accommodation: F1,4 = 6.52, p = 0.06

Figure 5.12. Normalized mean N1 and N1P1 amplitudes as a function of pupil eccentricity

between 0 D and 6 D accommodation. Abbreviations R1 = ring 1, R2 = ring 2, R3 = ring 3, R4 =

ring 4 and R5 = ring 5. Note that the 8 D data for subject LS was analysed with 6 D of other

subjects and subject ST was excluded.


116

The SCE directionality was obtained by fitting Gaussian functions to N1 and N1P1

amplitudes for both accommodation levels for each subject as described in section

5.3 (see Figure 5.13 and Table 5.3).

0 1 2 3 4 5

Am

pli

tud

e (

nV

) 0

20

40

60

80

100

120

140

160

0 1 2 3 4 5

0

20

40

60

80

100

120

140

160

0 1 2 3 4 5

0

50

100

150

200

250

300

350

AMNSLSAKEM

Eccentricity

0 1 2 3 4 5

0

50

100

150

200

250

300

350

N1 N1

N1P1 N1P1

0 D 6 D

A B

C D

Figure 5.13. Amplitudes of ring average waveforms for the combined data of two runs plotted

against ring eccentricity for 0 D and 6 D accommodation. A, B shows N1-amplitude and C, D

shows N1P1-amplitude. The lines are the Gaussian function fits for individual subjects.

Eccentricity = 0 corresponds to Ring 1, Eccentricity = 1 corresponds to Ring 2, Eccentricity = 2

corresponds to Ring 3, Eccentricity = 3 corresponds to Ring 4 and Eccentricity = 4 corresponds

to Ring 5. Legend for all panels is shown in A and B.

The Gaussian function fits for both N1 and N1P1 amplitudes were generally good

with four subjects having correlation coefficients ranging from 0.71 to 0.97 for both

accommodation stimuli and subject LS having a poor fit of r = 0.50 for N1 for 6 D

accommodation stimulus. The estimated ρ values for 7 mm pupil from the Gaussian

function fits to the N1 and N1P1 amplitudes for the ring average waveforms did not


117

show significant change with accommodation for five subjects (paired t-test, N1 p =

0.24; N1P1 p = 0.39) (Table 5.3).

Table 5.3. Comparison of ρ mean (first entry) from N1 and N1P1 amplitudes of the ring average

waveforms between 0 D and 6 D accommodation. AS and AR are accommodation stimulus and

accommodation response, respectively. Second entries in AR and ρ are the differences between

two runs and the standard errors of the Gaussian functions fitted to the combined data of two

runs, respectively.

Subject AS (D) AR (D)

N1 N1P1AM 0 0 0.566, 0.138 0.112, 0.027

6 3.6, 0.3 0.097, 0.049 0.142, 0.041

NS 0 0 0.247, 0.082 0.119, 0.0296 4.5, 0.1 0.223, 0.080 0.177, 0.028

LS 0 0 0.080, 0.038 0.044, 0.0158 4.5, 0.3 0.053, 0.056 0.118, 0.050

AK 0 0 0.133, 0.050 0.194, 0.0346 4.7, 0.7 0.091, 0.021 0.077, 0.015

EM 0 0 2.003, 0.947 1.785, 0.6396 5.2, 0.1 0.149, 0.040 0.112, 0.027

ρ (mm-2

)

5.5.4.2 Shift in peak sensitivity of the SCE with accommodation

As mentioned earlier, the peak response waveform (hexagon) with maximum N1P1

amplitude was considered to correspond to the peak of the SCE. The peak locations

were very variable between the subjects. To get a better estimate of the peak of the

SCE, analysis based on the peak hexagon (ring 1) and surrounding ring of hexagons

(ring 2) were performed as


118

Amplitude

xAmplitudex ,

Amplitude

yAmplitudey

where amplitude is N1P1 amplitude of N1P1, x and y are co-ordinates of hexagon

location in the pupil in millimetres, and x and y are the estimated co-ordinates of

the peak location in the pupil in millimetres. Table 5.4 shows the estimated peak in

the pupil for five subjects for their combined data of two runs for 0 D and 6 D

accommodation stimuli.

Table 5.4. Estimated peak locations for 0 D and 6 D accommodation stimuli along horizontal (x)

and vertical (y) meridians in the pupil.

Subject Peak Location (mm)

0 D 6 D

x y x y

AM 2.35 1.42 3.33 0.91

LS 2.22 1.30 0.21 -0.02

NS 1.57 0.02 1.55 0.04

AK 3.19 1.15 2.42 2.45

EM 0.13 0.06 3.18 0.97

The SCE peaks for three subjects were more peripheral than expected (approximately

2-3 mm) from the centre of the pupil along either of the meridians for both

accommodation stimuli. Paired t-tests showed no significant difference between the

peaks for 0 D and 6 D accommodation stimulus in either horizontal (p=0.79) or

vertical meridian (p=0.87). Along the horizontal meridian in the pupil, LS and EM

demonstrated shifts of 2 mm temporally and 3 mm nasally, respectively and along

the vertical meridian subjects LS and AK demonstrated shifts of 1.3 mm in inferior

and superior direction, respectively, with accommodation. Other subjects showed

shift < 1mm with 6 D accommodation, which is comparable with the peak shifts

from the psychophysical techniques (Chapter 4). Two subjects showed large shifts

(2-3 mm) in the SCE peak with 6 D accommodation, but the peak did not shift

systematically across the group.


119

5.6 Discussion

5.6.1 Changes in the SCE directionality with accommodation The mfERG technique successfully demonstrated the presence of SCE. While the

amplitudes of mfERG responses changed with eccentricity (SCE), the measured SCE

did not change with accommodation. Increased noise and variability in the mfERG

measured SCE makes this technique less reliable than the psychophysical study for

measuring SCE or for understanding changes in SCE with accommodation.

The lack of significant changes in SCE directionality with accommodation suggests

that probably mfERG based SCE measurement is not sensitive to the changes in the

SCE or the technique is too variable to give any meaningful results. The subjects

(except subject EM) elicited lower accommodation responses for 6 D stimulus

(Figure 5.10) in comparison to the psychophysical study (Figure 4.3). Discomfort

due to the electrode in the eye and refraining from blinking could have led to reduced

accommodative responses or affected the accuracy of photorefractor measurements

due to any reduction in tear film quality. As a precaution any photorefractor image

that looked saturated in brightness or demonstrated a noisy pupil brightness profile

were excluded from the analysis. The lack of significant levels of accommodation

could have led to only small changes in SCE within the noise level of the mfERG

based technique. To better understand the source of variability, another study

investigating repeatability of the measurements over a period of four days was

conducted for a subject DAA, with two runs performed each day. The maximum

variability in directionality across four days of measurements with respect to the Day

1 measurement for N1 and N1P1 amplitudes were 64% and 36%, respectively. Thus,

the technique was highly variable in determining SCE directionality.


120

5.6.2 Changes in the SCE peak location with accommodation

The SCE peak locations for both accommodation stimuli and shift in peak from 0 D

to 6 D accommodation stimuli were variable among subjects (Table 5.4). Two

subjects showed large shifts in the SCE peak location of approximately 3 mm along

horizontal meridian of the pupil with 6 D accommodation. In the psychophysical

study the shift in SCE peak was not more than 0.4 mm for any subject. Some of the

variation in SCE peak in the mfERG study could be because the pupil was not

monitored continuously and subjects were advised not to bite firmly on the bite-bar

(section 5.2.4). As a consequence, maintaining head stability could be an issue for

some subjects but it seems most unlikely that this would shift the SCE peak by 3

mm. The repeatability of measurements assessed across four consecutive days on one

subject (DAA) had SCE peak locations at 3 mm on two days and within 2 mm from

centre of the pupil on the other two days. Thus, the SCE measured using mfERG was

too noisy to give meaningful peak locations. The technique could be improved to

some extent by having a better way of monitoring the pupil continuously, for

example replacing the infrared LEDS by a light source which has minimal

interference with the mfERG signal and maintaining the stability of subject’s head,

for example using chin and head rest as well as the bite-bar.

5.6.3 Comparison with Sutter’s study

The technique used here was similar to that used by Sutter (1997) but the studies

were different in a few aspects. The active electrode used in this study was a DTL

thread electrode whereas Sutter used a contact lens electrode that gives larger

amplitude responses (Coupland, 1991; Esakowitz, Kriss & Shawkat, 1993) but is

also less comfortable for the subjects. Sutter used larger fields of 20° and 50°,

whereas a smaller field size of 13.8° was used here to sample a retinal area closer to

the fovea without compromising response signal (section 5.4.1). A delay of

approximately 5 ms in implicit time of the mfERG responses relative to that of


121

Sutter’s study was found. The delay in implicit time was due to the LCD monitor

being used to display the mfERG stimulus in this study, whereas Sutter used a CRT

monitor (see section 5.4.3). Sutter sampled more pupil locations (103 as compared

with 61 in the current study) that could have provided a better estimate of the SCE

directionality and peak location. Comparing signal-to-noise ratio (SNR) between the

data from the current study with those of Sutter’s data, the SNR from Sutter’s data

was approximately 6.3 (for the peak waveform) while the current study produces

SNR of approximately 4.6 (of peak waveform from a combined filtered data for 0 D

accommodation stimulus of subject AM). Thus, SNR was higher for Sutter’s data

than those of our data which suggest that the contact lens electrodes are more

sensitive to measure the changes in the SCE in comparison to the DTL electrodes.

5.6.4 Comparison between subjective and objective techniques of measuring the SCE

One objective of this study was to compare subjective (psychophysical) and

objective measurements of SCE. The objective technique using mfERG was quicker

than the psychophysical technique. One run of a SCE measurement with mfERG

took around 15 minutes compared to 40 minutes in the psychophysical technique.

However, the mfERG technique needs 15 minutes of preparation time to place the

electrodes on the subject’s face. Unlike the psychophysical technique, the objective

technique does not require any task to be done by the subjects except for fixation and

accommodation. However, it does not have the sensitivity to measure the SCE in the

foveal region as is done in the psychophysical study. The signal response in the

control study described in section 5.4.1 was poor for a field less than 10°. Although

the objective technique using mfERG was quicker and less tiring than the

psychophysical technique, it produces variable and less reliable data to accurately

describe the SCE.


122

5.7 Summary 1. mfERG amplitude changed with pupil entry location but not with accommodation.

While the technique demonstrates the presence of SCE, it was not sensitive enough,

as implemented here, to demonstrate a change in the SCE directionality with

accommodation. The SCE peak was variable among the subjects and did not show

systematic change with accommodation.

2. mfERG provides a quick but not a reliable way of measuring SCE in comparison

to the psychophysical study.

Chapter 6 Summary and future directions

123

CHAPTER 6

Summary and future directions

6.1 Summary

Mechanical stresses on the retina have been shown to alter the regular arrangement

or tuning characteristics of foveal photoreceptors. Any such mechanical stress caused

by myopia and/or accommodation can potentially be evaluated by studying the

Stiles-Crawford effect (SCE). The first objective of the research presented in this

thesis was to measure and compare the changes in the foveal SCE with

accommodation in emmetropes and myopes. The second objective was to develop a

quick objective technique to measure the SCE and compare it with the classic

subjective (psychophysical) technique.

Chapter 3 describes the development of a psychophysical technique to measure the

foveal SCE. Necessary calibration and control measurements are also described in

this chapter. Chapter 4 describes a study using this technique that investigated

changes in foveal SCE directionality and peak in young emmetropes and myopes

with accommodation for 0 D to 8 D stimuli. SCE directionality increased 15-20 %

with 6 D of accommodation, but with no apparent shift in the peak of the SCE. The

increase in SCE directionality with accommodation was similar between emmetropes

and myopes. Additional experiments indicated that much of this increase in

directionality was not a true effect of accommodation.

The psychophysical technique was laborious and demanding, especially with

extended periods of high accommodative demands. Chapter 5 describes the

development and evaluation of an objective technique of measuring the SCE using


124

multifocal ERG. With this technique, SCE could be measured four times faster than

in the psychophysical study. While a representation of SCE was evident in the

mfERG responses, the responses were very noisy and too variable to give meaningful

estimates of any changes that might occur in SCE directionality with

accommodation. The SCE peak was also variable among subjects and some subjects

showed shift in the peak location of up to 3 mm which does not seem very likely.

6.2 Discussion and conclusion

Blank, Provine & Enoch (1975) showed significant shifts of SCE peaks of 0.5 mm to

1.5 mm with 9 D accommodation stimulus whereas the psychophysical studies

presented in this thesis did not find any shift with up to 6 D accommodation

response. It is possible that changes in the SCE might occur at higher

accommodation response levels. It needs to be noted that Blank et al. did not

measure accommodation responses and these are unknown. Also, Blank et al. used

soft contact lenses to stimulate accommodation which could have caused an

artefactual shift in the SCE peak (see Table 3.7).

In this thesis additional experiments performed with soft contact lenses (+5D, -5D)

and induced defocus (+1 D, -1 D) showed that the SCE directionality increased with

a +5 D contact lens and with myopic (positive) defocus in comparison to a -5 D

contact lens and hypermetropic (negative) defocus (Tables 3.7 and 3.9). Also, higher

order aberrations were higher with the + 5 D contact lens than with -5 D contact lens

or no contact lens (Table 3.10). This finding suggested that aberrations acting in the

same direction as defocus may artificially increase the SCE directionality. Similarly,

spherical aberrations were found to decrease or became negative with increase in

accommodation which can combine with lag of accommodation (hypermetropic

defocus) and can artificially increase the SCE directionality. As an example of this

artefact, induced positive defocus led to significant increase in the SCE directionality

with the set-up, but not with a slightly different set-up that was less susceptible to

optical confounds. Therefore, most of the increase found in this study could be an

artefact of aberrations and accommodative lag rather than a true effect of

accommodation.


125

Similar to Sutter’s study, I found that the mfERG technique to measure the SCE was

quick and less tiring for the subjects (Chapter 5). The technique presented in this

thesis was not sensitive and repeatable in identifying small changes in the SCE such

as with accommodation. The DTL electrode was used in the current study whereas

Sutter used a Burrian-Allen electrode which gives bigger amplitudes. The SNR was

lower for the current study than for Sutter’s study. Additionally, the mfERG stimulus

size used was larger than the dilated pupil size of 7 mm and therefore, for analysis

the peripheral ring of hexagons was removed from the mfERG stimulus which

reduced the number of locations analysed across the pupil which also led to increased

noise in the data. Therefore, the technique could be improved by having a smaller

mfERG stimulus size, by using Burrian-Allen electrodes and by testing more

locations across the pupil.

In conclusion, this thesis reported little change in the SCE directionality and no shift

in the peak of the SCE with up to 6 D accommodation, either in emmetropes or

myopes. The objective technique using the multifocal electroretinogram was a rapid

and less laborious technique than the psychophysical technique. Although there

could be improvements in the technique, as used in this thesis it does not appear to be

a reliable method to measure the SCE.

6.3 Future research directions

The present research relates to foveal SCE, moderate levels of accommodation and

comparison of one subjective and one objective technique. Another area of research

is comparison of foveal vs non-foveal SCE at high levels of accommodation in

younger myopes and emmetropes with new measurement techniques. With advances

in technology, single photoreceptors can be monitored with accommodation instead

of pooled tuning characteristic as measured with the psychophysical SCE.

Blank, Provine & Enoch (1975) observed that marked accommodation (9 D) causes

the retina to elongate asymmetrically, more so in the nasal than in the temporal side

due to the presence of the optic nerve head. Measurement of the SCE at different


126

retinal locations at high levels of accommodation may help better understanding of

accommodation influence on the retina. For example, 10-18 year old emmetropes

and myopes can be tested to elicit high accommodative response amplitudes. It

would be useful to develop fast, robust techniques to measure the SCE especially

when high levels of accommodation are involved. In addition, any new technique

should be carefully evaluated for confounding factors such as optical artefacts as

carried out in this thesis.

The techniques used in this research cannot describe changes in individual

photoreceptor orientations with accommodation. Measurements of these might

identify areas of retina susceptible to myopic traction prior to prominent clinical

signs. An objective technique for measuring SCE by determining photoreceptor

directionality and orientation is briefly described here. This is based on the adaptive

optics technique of Roorda and Williams (2002). The system will consist of several

elements (Figure 6.1). A super luminescent diode will be the light source for ocular

wavefront aberration measurement. A Hartmann–Shack sensor conjugated with the

pupil will measure wavefront slopes at different pupil points and wavefront

aberrations of the eye can be reconstructed from its measurement across the pupil. A

deformable mirror conjugate with the pupil and the sensor will correct these

aberrations. An illumination source such as the Xenon flash lamp will be triggered

when aberrations are low enough. About 1 degree retinal area can be illuminated and

light reflected back from this area will be collected by a scientific high performance

CCD camera to form images of photoreceptors.

An optical trombone will be used to vary defocus and stimulate accommodation to a

fixation target independent of the deformable mirror. The fixation target can be

moved transversely to measure different retinal locations. The 2 mm aperture PA

conjugate with the pupil will be moved laterally to change illumination entry

positions at the pupil. The photoreceptor images obtained at different pupil positions

will have different intensity distributions for a particular retinal location, and will be

used to determine changes in the SCE with accommodation.


127

Figure 6.2 (A) outlines an image processing algorithm to enhance raw images of the

retinal photoreceptors. Figure 6.2 (B) gives a flow chart for further processing that

will be required to determine SCE directionality and orientation parameters.

Figure 6.1. Schematic representation of an adaptive optics - retinal imaging system.

Fixation target

High performance CCD

Eye

PA

Hartmann-Shack Sensor

Deformable Mirror

Super Luminescent Diode

Xe Flash Arc lamp

Optical trombone


128

Figure 6.2. Flow chart demonstrating (A) the retinal image processing algorithm and (B) the

procedures required for determining the SCE parameters.

Oversampling

Fourier domain low-pass filtering

Morphological opening

Cross correlation

Spline interpolation of cross correlation

Subsequent sub-pixel registration

Images registration

Raw retinal images with background removed

Register other pupil position images with central image

Measure average intensity of cone on central image

Measure average intensities of cones on peripheral images

Fit intensity values for each cone with Gaussian function

Correct residual optical blur and get average photoreceptor disarray

Correct finite entrance beam aperture by deconvolution

SCE parameters

Identify cone locations on central pupil image

A B

Oversampling



Cross correlation



Images registration


Oversampling



Cross correlation



Images registration








SCE parameters








SCE parameters


A B

References

129

References This list also includes references appearing in Appendix 2.

Abbott, M.L., Schmid, K.L., & Strang, N.C. (1998). Differences in the accommodation stimulus response curves of adult myopes and emmetropes. Ophthalmic and Physiological Optics, 18, 13-20.

American National Standards Institute (2004). American National Standard for Ophthalmics - Methods for reporting optical aberrations of the eye, ANSI Z80.28-2004.

Applegate, R.A., & Bonds, A.B. (1981). Induced movement of receptor alignment toward a new pupillary aperture. Investigative Ophthalmology and Visual Science, 21, 869-873.

Applegate, R.A., & Lakshminarayanan, V. (1993). Parametric representation of Stiles-Crawford functions: normal variation of peak location and directionality. Journal of the Optical Society of America A. Optics and Image Science, 10, 1611-1623.

Armington, J.C. (1967). Letter to the editor: Pupil entry and the human electroretinogram. Journal of the Optical Society of America, 57, 838-839.

Atchison, D.A. (2004). Recent advances in representation of monochromatic aberrations of human eyes. Clinical and Experimental Optometry, 87, 138-148.

Atchison, D.A., Joblin, A., & Smith, G. (1998). Influence of Stiles-Crawford effect apodization on spatial visual performance. Journal of the Optical Society of America A. Optics and Image Science, 15, 2545-2551.

Atchison, D.A., & Scott, D.H. (2002a). Contrast sensitivity and the Stiles-Crawford effect. Vision Research, 42, 1559-1569.

Atchison, D.A., & Scott, D.H. (2002b). The Stiles-Crawford effect and subjective measurement of aberrations. Vision Research, 42, 1089-1102.

Atchison, D.A., Scott, D.H., Joblin, A., & Smith, G. (2001). Influence of Stiles-Crawford effect apodization on spatial visual performance with decentered pupils. Journal of the Optical Society of America A. Optics and Image Science, 18, 1201-1211.

References

130

Atchison, D.A., Scott, D.H., Strang, N.C., & Artal, P. (2002). Influence of Stiles-Crawford apodization on visual acuity. Journal of the Optical Society of America A. Optics and Image Science, 19, 1073-1083.

Atchison, D.A., & Smith, G. (2000). Optics of the human eye. (p. 269). Oxford: Butterworth-Heinemann.

Bailey, J.E., & Heath, G.G. (1978). Flicker effects on receptor directional sensitivity. American Journal of Optometry and Physiological Optics, 55, 807-812.

Bailey, J.E., Lakshminarayanan, V., & Enoch, J.M. (1994). Photoreceptor orientation in iris coloboma. Optometry and Vision Science, 71, 120-124.

Bastani, B., Cressman, B., & Funt, B. (2005). Calibrated color mapping between LCD and CRT displays: A case study. Color Research and Application, 30, 438-447.

Bedell, H.E., & Enoch, J.M. (1979). A study of the Stiles-Crawford (S-C) function at 35 degrees in the temporal field and the stability of the foveal S-C function peak over time. Journal of the Optical Society Of America, 69, 435-442.

Beresford, J.A., Crewther, S.G., & Crewther, D.P. (1999). A technique for in vivo measurement of photoreceptor orientation in the chicken retina. Australian and New Zealand Journal of Ophthalmology, 27, 241-243.

Berns, R.S. (1996). Methods for characterizing CRT displays. Displays, 16, 173-182.

Berns, R.S., Motta, R.J., & Gorzynski, M.E. (1993). CRT colorimetry. Part I: theory and practice. Color Research and Application, 18, 299-314.

Biggs, R.D., Alpern, M., & Bennett, D.R. (1959). The effect of sympathomimetic drugs upon the amplitude of accommodation. American Journal of Ophthalmology, 48, 169-172.

Blank, K., & Enoch, J.M. (1973). Monocular spatial distortions induced by marked accommodation. Science, 182, 393-395.

Blank, K., Provine, R.R., & Enoch, J.M. (1975). Shift in the peak of the photopic Stiles-Crawford function with marked accommodation. Vision Research, 15, 499-507.

Brainard, D.H., Pelli, D.G., & Robson, T. (2002). Display characterization. In: J. Hornak (Ed.) The encyclopedia of imaging science and technology (pp. 172-188). New York: Wiley.

References

131

Breton, M.E., Schueller, A.W., Lamb, T.D., & Pugh, E.N., Jr. (1994). Analysis of ERG a-wave amplification and kinetics in terms of the G-protein cascade of phototransduction. Investigative Ophthalmology and Visual Science, 35, 295-309.

Burns, S.A., Wu, S., Delori, F., & Elsner, A.E. (1995). Direct measurement of human-cone-photoreceptor alignment. Journal of the Optical Society of America A. Optics and Image Science, 12, 2329-2338.

Burns, S.A., Wu, S., He, J.C., & Elsner, A.E. (1997). Variations in photoreceptor directionality across the central retina. Journal of the Optical Society of America A. Optics and Image Science, 14, 2033-2040.

Carney, L.G., Mainstone, J.C., & Henderson, B.A. (1997). Corneal topography and myopia. A cross-sectional study. Investigative Ophthalmology and Visual Science, 38, 311-320.

Charman, W.N. (1982). The accommodative resting point and refractive error. Ophthalmic Optician, 22, 469-473.

Charman, W.N. (1999). Near vision, lags of accommodation and myopia. Ophthalmic and Physiological Optics, 19, 126-133.

Charman, W.N., & Heron, G. (1988). Fluctuations in accommodation: a review. Ophthalmic and Physiological Optics, 8, 153-164.

Chen, B., & Makous, W. (1989). Light capture by human cones. Journal of Physiology, 414, 89–109.

Cheng, H., Barnett, J.K., Vilupuru, A.S., Marsack, J.D., Kasthurirangan, S., Applegate, R.A., & Roorda, A. (2004). A population study on changes in wave aberrations with accommodation. Journal of Vision, 4, 272-280.

Cho, P., Cheung, S.W., & Edwards, M. (2005). The longitudinal orthokeratology research in children (LORIC) in Hong Kong: a pilot study on refractive changes and myopic control. Current Eye Research, 30, 71-80.

Choi, M., Weiss, S., Schaeffel, F., Seidemann, A., Howland, H.C., Wilhelm, B., & Wilhelm, H. (2000). Laboratory, clinical, and kindergarten test of a new eccentric infrared photorefractor (PowerRefractor). Optometry and Vision Science, 77, 537-548.

Choi, S.S., Enoch, J.M., & Kono, M. (2004). Evidence for transient forces/strains at the optic nerve head in myopia: repeated measurements of the Stiles-Crawford Effect of the first kind (SCE-I) over time. Ophthalmic and Physiological Optics, 24, 194-206.

References

132

Choi, S.S., Garner, L.F., & Enoch, J.M. (2003a). The relationship between the Stiles-Crawford effect of the first kind (SCE-I) and myopia. Ophthalmic and Physiological Optics, 23, 465-472.

Choi, S.S., Garner, L.F., & Enoch, J.M. (2003b). Stiles-Crawford effect of the first kind (SCE-I) in post-photorefractive keratectomy and anisometropic subjects. Ophthalmic and Physiological Optics, 23, 473-476.

Ciuffreda, K.J. (1991). Accommodation to gratings and more naturalistic stimuli. Optometry and Vision Science, 68, 243-260.

Ciuffreda, K.J., & Wallis, D.M. (1998). Myopes show increased susceptibility to nearwork aftereffects. Investigative Ophthalmology and Visual Science, 39, 1797-1803.

Collins, G. (1937). The electronic refractometer. British Journal of Physiological Optics, 1, 30-42.

Collins, M.J., Davis, B., & Wood, J.M. (1995). Microfluctuations of steady-state accommodation and the cardiopulmonary system. Vision Research, 35, 2491-2502.

Coupland, S. (1991). Electrodes for clinical electrophysiology testing. In: J.R. Heckenlively, & G.B. Arden (Eds.), Principles and Practice of Clinical Electrophysiology of Vision (pp. 177-182). St. Louis: Mosby.

Crawford, B.H. (1937). The luminous efficiency of light entering the eye pupil at different points and its relation to brightness threshold measurements. Proceedings of the Royal Society of London. Series B: Biological Sciences, 124, 81-96.

Culhane, H.M., & Winn, B. (1999). Dynamic accommodation and myopia. Investigative Ophthalmology and Visual Science, 40, 1968-1974.

Culhane, H.M., Winn, B., & Gilmartin, B. (1999). Human dynamic closed-loop accommodation augmented by sympathetic inhibition. Investigative Ophthalmology and Visual Science, 40, 1137-1143.

Dandona, R., Dandona, L., Naduvilath, T.J., Srinivas, M., McCarty, C.A., & Rao, G.N. (1999). Refractive errors in an urban population in southern India: the Andhra Pradesh eye disease study. Investigative Ophthalmology and Visual Science, 40, 2810-2818.

Dandona, R., Dandona, L., Srinivas, M., Sahare, P., Narsaiah, S., Munoz, S.R., Pokharel, G.P., & Ellwein, L.B. (2002). Refractive error in children in a rural population in India. Investigative Ophthalmology and Visual Science, 43, 615-622.

References

133

Day, E.A., Taplin, L., & Berns, R.S. (2004). Colorimetric characterization of a computer-controlled liquid crystal display. Color Research and Application, 29, 365-373.

Delint, P.J., Berendschot, T.T.J.M., & van Norren, D. (1997). Local photoreceptor alignment measured with a scanning laser ophthalmoscope. Vision Research, 37, 243-248.

Donders, F.C. (1864). On the Anomalies of Accommodation and Refraction of the Eye. (London: The New Sydenham Society.

Duane, A. (1912). Normal values of the accommodation at all ages. Journal of the American Medical Association, 59, 1010-1013.

Dunnewold, C.J.W. (1964). On the Campbell and Stiles-Crawford effect and their clinical importance. Ph. D. dissertation (Rijksuniversiteit te Utrecht, Utrecht, The Netherlands), 1-84.

Edwards, M.H. (1996). Do variations in normal nutrition play a role in the development of myopia? Optometry and Vision Science, 73, 638-643.

Enoch, J.M. (1973). Effect of substantial accommodation on total retinal area. Journal Of The Optical Society Of America, 63, 899.

Enoch, J.M. (1975). Marked accommodation, retinal stretch, monocular space perception and retinal receptor orientation. American Journal of Optometry and Physiological Optics, 52, 376-392.

Enoch, J.M., & Bedell, H.E. (1981). The Stiles Crawford Effects. In: J.M. Enoch, & F.L.J. Tobey (Eds.), Vertebrate Photoreceptor Optics (pp. 83-126). Heidelberg: Springer-Verlag.

Enoch, J.M., & Birch, D.G. (1981). Inferred positive phototropic activity in human photoreceptors. Philosophical Transactions of the Royal Society of London. Series B, Biological Sciences, 291, 323-351.

Enoch, J.M., & Hope, G.M. (1972a). An analysis of retinal receptor orientation. III. Results of initial psychophysical tests. Investigative Ophthalmology, 11, 765 - 782.

Enoch, J.M., & Hope, G.M. (1972b). An analysis of retinal receptor orientation. IV. Center of the entrance pupil and the center of convergence of orientation and directional sensitivity. Investigative Ophthalmology, 11, 1017-1021.

Enoch, J.M., & Hope, G.M. (1973). Directional sensitivity of the foveal and parafoveal retina. Investigative Ophthalmology, 12, 497-503.

References

134

Enoch, J.M., & Lakshminarayanan, V. (1991). Retinal fibre optics. In: W.N. Charman (Ed.) Visual Optics and Instrumentation, Vol. 1 of Vision and visual dysfunction (pp. 280-309). New York: MacMillan Press.

Enoch, J.M., Moses, R.A., Nygaard, R.W., & Allen, D. (1983). Perimetric techniques used to assess retinal strain during accommodaiton. Fifth International Visual Field Symposium, 413-420.

Enoch, J.M., & Stiles, W.S. (1961). The colour change of monochromatic light with retinal angle of incidence. Acta Ophthalmologica, 8, 329-358.

Enoch, J.M., & Tobey, F.L. (1981). Vertebrate Photoreceptor Optics. (p. 469). Heidelberg: Springer-Verlag.

Esakowitz, L., Kriss, A., & Shawkat, F. (1993). A comparison of flash electroretinograms recorded from Burian Allen, JET, C-glide, gold foil, DTL and skin electrodes. Eye, 7, 169-171.

Fairchild, M.D., & Wyble, D.R. (1998). Colorimetric characterization of the Apple studio display (Flat Panel LCD). Munsell Color Science Laboratory Technical Report, Rochester Institute of Technology, Rochester, NY, 1-22.

Fan, D.S., Lam, D.S., Lam, R.F., Lau, J.T., Chong, K.S., Cheung, E.Y., Lai, R.Y., & Chew, S.J.H.Y. (2004). Prevalence, incidence, and progression of myopia of school children in Hong Kong. Investigative Ophthalmology and Visual Science, 45, 1071-1075.

Fankhauser, F., Enoch, J., & Cibis, P. (1961). Receptor orientation in retinal pathology. A first study. American Journal Of Ophthalmology, 52, 767-783.

Feigl, B., Brown, B., Lovie-Kitchin, J., & Swann, P. (2004). Cone- and rod-mediated multifocal electroretinogram in early age-related maculopathy. Eye, 19, 431-441.

Feigl, B., & Zele, A.J. (2008). A method for investigating the temporal dynamics of local neuroretinal responses. Journal of Neuroscience Methods, 167, 207-212.

Fincham, E.F. (1937). The mechanism of accommodation. British Journal of Ophthalmology, 8 (Supplement), 5-80.

Fortune, B., Johnson, C.A., & Cioffi, G.A. (2001). The topographic relationship between multifocal electroretinographic and behavioral perimetric measures of function in glaucoma. Optometry and Vision Science 78, 206–214.

Gao, W., Cense, B., Zhang, Y., Jonnal, R.S., & Miller, D.T. (2008). Measuring retinal contributions to the optical Stiles-Crawford effect with optical coherence tomography. Optics Express, 16, 6486-6501.

References

135

García-Pérez, M.A., & Peli, E. (2001). Luminance artifacts of cathode-ray tube displays for vision research. Spatial Vision, 14, 201–215.

Gardiner, P.A. (1958). Dietary treatment of myopia in children. Lancet i, 1152-1155.

Gekeler, F., Schaeffel, F., Howland, H.C., & Wattam-Bell, J. (1997). Measurement of astigmatism by automated infrared photoretinoscopy. Optometry and Vision Science, 74, 472-482.

Gibson, J.E., & Fairchild, M.D. (2000). Colorimetric characterization of three computer displays (LCD and CRT). Munsell Color Science Laboratory Technical Report, Rochester Institute of Technology, Rochester, NY, 1-40.

Glasser, J. (1997). Display systems: design and applications. (Chichester: John Wiley & Sons.

Gonzalez, P., Parks, S., Dolan, F., & Keating, D. (2004). The effects of pupil size on the multifocal electroretinogram. Documenta Ophthalmologica, 109, 67-72.

Gorrand, J.-M., & Delori, F.C. (1997). A model for assessment of cone directionality. Journal of Modern Optics, 44, 473-491.

Gorrand, J.M., & Delori, F. (1995). A reflectometric technique for assessing photoreceptor alignment. Vision Research, 35, 999-1010.

Granit, R. (1947). Sensory mechanisms of the retina. With an appendix on electroretinography. (p. 412). London: Oxford University Press.

Gray, L.S., Winn, B., & Gilmartin, B. (1993). Microfluctuations of accommodation below 0.6 Hz are likely to contribute to the maintenance of sustained accommodation. Investigative Ophthalmology and Visual Science, 34 (Supplement), 1307.

Grosvenor, T., & Scott, R. (1994). Role of the axial length/corneal radius ratio in determining the refractive state of the eye. Optometry and Vision Science, 71, 573-579.

Gwiazda, J., Hyman, L., Hussein, M., Everett, D., Norton, T.T., Kurtz, D., Leske, M.C., Manny, R., Marsh-Tootle, W., & Scheiman, M. (2003). A randomized clinical trial of progressive addition lenses versus single vision lenses on the progression of myopia in children. Investigative Ophthalmology and Visual Science, 44, 1492-1500.

Gwiazda, J., Thorn, F., Bauer, J., & Held, R. (1993). Myopic children show insufficient accommodative response to blur. Investigative Ophthalmology and Visual Science, 34, 690-694.

References

136

He, J.C., Marcos, S., & Burns, S.A. (1999). Comparison of cone directionality determined by psychophysical and reflectometric techniques. Journal of the Optical Society of America A. Optics and Image Science, 16, 2363-2369.

Heckenlively, J.R., & Arden, G.B. (2006). Principles and practice of clinical electrophysiology of vision (2nd edition). (p. 977). MA: MIT Press.

Heinemann-Vernaleken, B., Palmowski, A.M., Allgayer, R., & Ruprecht, K.W. (2001). Comparison of different high resolution multifocal electroretinogram recordings in patients with age-related maculopathy. Graefe's Archive for Clinical and Experimental Ophthalmology, 239, 556-561.

Hollins, M. (1974). Does the central human retina stretch during accommodation? Nature, 251, 729-730.

Hood, D., Seiple, W., Holopigian, K., & Greenstein, V. (1997). A comparison of the components of the multifocal and full-field ERGs. Visual Neuroscience, 14, 533-544.

Hood, D.C. (2000). Assessing retinal function with the multifocal technique. Progress in Retinal and Eye Research, 19, 607-646.

Hood, D.C., & Finkelstein, M.A. (1986). Sensitivity to light. In: K.R. Boff, L. Kaufman, & J.P. Thomas (Eds.), Handbook of perception and human performance: Sensory processes and perception, Volume 1 (pp. 5-1, 5-66). New York: John Wiley & Sons.

Hood, D.C., Frishman, L.J., Saszik, S., & Viswanathan, S. (2002). Retinal origins of the primate multifocal ERG: implications for the human response. Investigative Ophthalmology and Visual Science, 43, 1673-1685.

Hood, D.C., Greenstein, V.C., Holopigian, K., Bauer, R., Firoz, B., Liebmann, J.M., Odel, J.G., & Ritch, R. (2000). An attempt to detect glaucomatous damage to the inner retina with the multifocal ERG. Investigative Ophthalmology and Visual Science, 41, 1570-1579.

Hood, D.C., Odel, J.G., Chen, C.S., & Winn, B.J. (2003). The multifocal electroretinogram. Journal of Neuroophthalmology 23, 225–235.

Hosaka, A. (1988). Population studies: Myopia experience in Japan. Acta Ophthalmologica, 185 (Supplement), 37-40.

Hunt, O.A., Wolffsohn, J.S., & Gilmartin, B. (2003). Evaluation of the measurement of refractive error by the PowerRefractor: a remote, continuous and binocular measurement system of oculomotor function. The British Journal of Ophthalmology, 87, 1504-1508.

References

137

Jainta, S., Jaschinski, W., & Hoormann, J. (2004). Measurement of refractive error and accommodation with the photorefractor PowerRef II. Ophthalmic and Physiological Optics, 24, 520-527.

Kahn, R., & Lowenstein, A. (1908). Das Electroretinogramm. Graefes Arch Clin Exp Ophthalmol, 114, 304-325.

Kasthurirangan, S., Vilupuru, A.S., & Glasser, A. (2003). Amplitude dependent accommodative dynamics in humans. Vision Research, 43, 2945-2956.

Keating, D., Parks, S., & Evans, A. (2000). Technical aspects of multifocal ERG recording. Documenta Ophthalmologica, 100, 77-98.

Keating, D., Parks, S., Malloch, C., & Evans, A. (2001). A comparison of CRT and Digital stimulus delivery methods in the multifocal ERG. Documenta Ophthalmologica, 102 95-114.

Khoo, C.Y., Chong, J., & Rajan, U. (1999). A 3-year study on the effect of RGP contact lenses on myopic children. Singapore Medical Journal, 40, 230-237.

Kinge, B., Midelfart, A., Jacobsen, G., & Rystad, J. (1999). Biometric changes in the eyes of Norwegian university students - a three-year longitudinal study. Acta Ophthalmologica Scandinavica, 77, 648-652.

Komáromy, A.M., Brooks, D.E., Dawson, W.W., Källberg, M.E., & Ollivier, F.J. (2002). Technical issues in electrodiagnostic recording. Veterinary Ophthalmology, 5, 85-91.

Krauskopf, J. (1965). Some experiments with a photoelectric ophthalmoscope. Excerpta Medica International Congress Series, 63.

Kretschmann, U., Bock, M., Gockeln, R., & Zrenner, E. (2000). Clinical applications of multifocal electroretinography. Documenta Ophthalmologica, 100, 99-113.

Kruger, P.B., & Pola, J. (1985). Changing target size is a stimulus for accommodation. Journal of the Optical Society of America A. Optics and Image Science, 2, 1832-1835.

Kruger, P.B., & Pola, J. (1986). Stimuli for accommodation: blur, chromatic aberration and size. Vision Research, 26, 957-971.

Kruger, P.B., & Pola, J. (1987). Dioptric and non-dioptric stimuli for accommodation: target size alone and with blur and chromatic aberration. Vision Research, 27, 556-567.

References

138

Lakshminarayanan, V., Bailey, J.E., & Enoch, J.M. (1997). Photoreceptor orientation and alignment in nasal fundus ectasia. Optometry and Vision Science, 74, 1011-1018.

Lam, B.L. (2005). Electrophysiology of vision: clinical testing and applications. (p. 528). Florida: Taylor & Francis Group.

Laties, A. (1969). Histological techniques for the study of photoreceptor orientation. Tissue and Cell, 63-81.

Laties, A., Liebman, P., & Campbell, C.E.M. (1968). Photoreceptor orientation in the primate eye. Nature (London), 218, 172-173.

Liang, J., Williams, D.R., & Miller, D.T. (1997). Supernormal vision and high-resolution retinal imaging through adaptive optics. Journal of the Optical Society of America A. Optics and Image Science, 14, 2884-2892.

Lin, L.L., Shih, Y.F., Tsai, C.B., Chen, C.J., Lee, L.A., Hung, P.T., & Hou, P.K. (1999). Epidemiological study of ocular refraction among school children in Taiwan in 1995. Optometry and Vision Science, 76, 275-281.

Mallen, E.A., Wolffsohn, J.S., Gilmartin, B., & Tsujimura, S. (2001). Clinical evaluation of the Shin-Nippon SRW-5000 autorefractor in adults. Ophthalmic and Physiological Optics, 21, 101-107.

Marmor, M.F., Hood, D.C., Keating, D., Kondo, M., Seeliger, M.W., & Miyake, Y. (2003). Guidelines for basic multifocal electroretinography (mfERG). Documenta Ophthalmologica, 106, 105-115.

McBrien, N.A., & Millodot, M. (1986). The effect of refractive error on the accommodative response gradient. Ophthalmic and Physiological Optics, 6, 145-149.

Metha, A.B., Vingrys, A.J., & Badcock, D.R. (1993). Calibration of a colour monitor for visual psychophysics. Behavior Research Methods, Instruments and Computers, 25, 371-383.

Mordi, J., Tucker, J., & Charman, W.N. (1986). Effects of 0.1% cyclopentolate or 10% phenylephrine on pupil diameter and accommodation. Ophthalmic and Physiological Optics, 6, 221-227.

Mordi, J.A., Lyle, W.M., & Mousa, G.Y. (1986). Effect of phenylephrine on accommodation. American Journal of Optometry and Physiological Optics, 63, 294 - 297.

Morgan, M.W. (1944). Accommodation and its relationship to convergence. American Journal of Optometry and Archives of American Academy of Optometry, 21, 183-195.

References

139

Moses, R.A. (1987). Accommodation. In: Adler’s Physiology of the Eye (pp. 291-310). St. Louis: C.V. Mosby.

Ninomiya, S., Fujikado, T., Kuroda, T., Maeda, N., Tano, Y., Oshika, T., Hirohara, Y., & Mihashi, T. (2002). Changes of ocular aberration with accommodation. American Journal of Ophthalmology, 134, 924-926.

O'Brien, B. (1946). A theory of the Stiles and Crawford effect. Journal of the Optical Society of America, 36, 506-509.

O'Brien, B. (1951). Vision and resolution in the central retina. Journal of the Optical Society of America, 41, 882-894.

Ong, E., & Ciuffreda, K.J. (1995). Nearwork-induced transient myopia. Documenta Ophthalmologica, 91, 57-85.

Ostrin, L.A., & Glasser, A. (2004). The effects of phenylephrine on pupil diameter and accommodation in rhesus monkeys. Investigative Ophthalmology and Visual Science, 45, 215-221.

Parks, S., Keating, D., Evans, A.L., Williamson, T.H., Jay, J.L., & Elliott, A.T. (1997). Comparison of repeatability of the multifocal electroretinogram and Humphrey perimeter. Documenta Ophthalmologica 92, 281-289.

Pask, C., & Stacey, A. (1998). Optical properties of retinal photoreceptors and the Campbell effect. Vision Research, 38, 953-961.

Perlman, I., Gdal-On, M., Miller, B., & Zonis, S. (1985). Retinal function of the diabetic retina after argon laser photocoagulation assessed electroretinographically. British Journal of Ophthalmology, 69, 240-246.

Plainis, S., Ginis, H.S., & Pallikaris, A. (2005). The effect of ocular aberrations on steady-state errors of accommodative response. Journal of Vision, 5, 466-477.

Poloschek, C.M., & Sutter, E.E. (2002). The fine structure of multifocal ERG topographies. Journal of Vision, 2, 577-587.

Richards, W. (1969). Saccadic suppression. Journal of the Optical Society of America, 59, 617-623.

Rodieck, R.W. (1972). Components of the electroretinogram - a reappraisal. Vision Research, 12, 773-780.

Roorda, A., Romero-Borja, F., Donnelly, W.J., Queener, H., Hebert, T.J., & Campbell, M.C.W. (2002). Adaptive optics scanning laser ophthalmoscopy. Optics Express, 10, 405-412.

References

140

Roorda, A., & Williams, D.R. (1999). The arrangement of the three cone classes in the living human eye. Nature, 397, 520-522.

Roorda, A., & Williams, D.R. (2002). Optical fiber properties of individual human cones. Journal of Vision, 2, 404-412.

Rosenfield, M. (2009). Chapter 15: Clinical assessment of accommodation. In: M. Rosenfield, & N. Logan (Eds.), Optometry: Science, Techniques and Clinical Management (2nd edition) (pp. 229-240). Oxford: Butterworth-Heinemann Elsevier.

Rosenfield, M., & Araham-Cohen, J.A. (1999). Blur sensitivity in myopes. Optometry and Vision Science, 76,

Rosenfield, M., & Gilmartin, B. (1998). Myopia and Nearwork. (p. 220). Oxford: Butterworth-Heinmann.

Safir, A., & Hyams, L. (1969). Distribution of cone orientations as an explanation of the Stiles-Crawford effect. Journal of the Optical Society of America, 59, 757-765.

Safir, A., Hyams, L., & Philpot, J. (1970). Movement of the Stiles-Crawford effect. Investigative Ophthalmology and Visual Science, 9, 820-825.

Saw, S.M., Gazzard, G., Au Eong, K.G., & Tan, D.T. (2002). Myopia: attempts to arrest progression. British Journal of Ophthalmology, 86, 1306-1311.

Saw, S.M., Katz, J., Schein, O.D., Chew, S.J., & Chann, T.K. (1996). Epidemiology of myopia. Epidemiologic Review, 18, 175-187.

Saw, S.M., Tong, L., Chua, W.H., Chia, K.S., Koh, D., Tan, D.T., & Katz, J. (2005). Incidence and progression of myopia in Singaporean school children. Investigative Ophthalmology and Visual Science, 46, 51-57.

Schaeffel, F., Howland, H., Weiss, S., & Zrenner, E. (1993). Measurement of the dynamics of accommodation by automated real time photorefraction. Investigative Ophthalmology and Visual Science, 34, 1306.

Schaeffel, F., Wilhelm, H., & Zrenner, E. (1993). Inter-individual variability in the dynamics of natural accommodation in humans: relation to age and refractive errors. Journal of Physiology, 461, 301-320.

Schmid, K.L., Iskander, D.R., Li, R.W.H., Edwards, M.H., & Lew, J.K.F. (2002). Blur detection thresholds in childhood myopia: single and dual target presentation. Vision Research, 42, 239-247.

Seidel, D., Gray, L.S., & Heron, G. (2003). Retinotopic accommodation responses in myopia. Investigative Ophthalmology and Visual Science, 44, 1035-1041.

References

141

Seidemann, A., & Schaeffel, F. (2002). Effects of longitudinal chromatic aberration on accommodation and emmetropization. Vision Research, 42, 2409.

Seidemann, A., & Schaeffel, F. (2003). An evaluation of the lag of accommodation using photorefraction. Vision Research, 43, 419-430.

Sharma, G. (2002). LCD Versus CRTs - color-calibration and gamut considerations. Proceedings of the IEEE, 90, 605-622.

Shih, Y.F., Chen, C.H., Chou, A.C., Ho, T.C., Lin, L.L., & Hung, P.T. (1999). Effects of different concentrations of atropine on controlling myopia in myopic children. Journal of Ocular Pharmacology and Therapeutics, 15, 85-90.

Smallman, H.S., MacLeod, D.I., & Doyle, P. (2001). Vision. Realignment of cones after cataract removal. Nature, 412, 604-605.

Smith, V.C., Pokorny, J., & Diddie, K.R. (1978). Color matching and Stiles-Crawford effect in central serous choroidopathy. Modern Problems in Ophthalmology, 19, 284-295.

Smith, V.C., Pokorny, J., & Diddie, K.R. (1988). Color matching and the Stiles-Crawford effect in observers with early age-related macular changes. Journal of the Optical Society of America A: Optics, Image Science and Vision, 5, 2113-2121.

Snyder, A.W., & Pask, C. (1973). The Stiles-Crawford effect - explanation and consequences. Vision Research, 13, 1115-1137.

Sperling, G. (1971). The description and luminous calibration of cathode ray oscilloscope visual displays. Behavioral Research Methods & Instrumentation, 3, 148-151.

Stacey, A., & Pask, C. (1994). Spatial-frequency response of a photoreceptor and its wavelength dependence. I. Coherent sources. Journal of the Optical Society of America A. Optics and Image Science, 11, 1193-1198.

Sternheim, C.E., & Riggs, L.A. (1968). Utilization of the Stiles-Crawford effect in the investigation of the origin of electrical responses of the human eye. Vision Research, 8, 25-33.

Stiles, W.S. (1937). The luminous efficiency of monochromatic rays entering the eye pupil at different points and a new colour effect. Proceedings of the Royal Society of London. Series B: Biological Sciences, 123, 90-118.

Stiles, W.S. (1939). The directional sensitivity of the retina, and the spectral sensitivities of the rods and cones. Proceedings of the Royal Society of London. Series B: Biological Sciences, 127, 64-105.

References

142

Stiles, W.S., & Crawford, B.H. (1933). The luminous efficiency of rays entering the eye pupil at different points. Proceedings of the Royal Society of London. Series B: Biological Sciences, 112, 428-450.

Stockton, R.A., & Slaughter, M.M. (1989). B-wave of the electroretinogram. A reflection of ON bipolar cell activity. The Journal of General Physiology, 93, 101-122.

Strang, N.C., Schmid, K.L., & Carney, L.G. (1998). Hyperopia is predominantly axial in nature. Current Eye Research, 17, 380-383.

Sutter, E.E. (1991). The fast m-transform: a fast computation of cross-correlations with binary m-sequences. Journal of Computing, 20, 686-694.

Sutter, E.E. (1997). Rapid derivation of the Stiles-Crawford function using electrophysiological responses. In: Vision Science and its Applications. OSA Technical Digest Series, Volume 1 (pp. 258-261). Washington, D.C.: Optical Society of America.

Sutter, E.E., & Bearse, M.A. (1999). The optic nerve head component of the human ERG. Vision Research, 39, 419-436.

Sutter, E.E., & Tran, D. (1992). The field topography of ERG components in man-I. The photopic luminance response. Vision Research, 32, 433-446.

Tan, G.J.S., Ng, Y.P., Lim, Y.C., Ong, P.Y., Snodgrass, A., & Saw, S.M. (2000). Cross-sectional study of near work and myopia in kindergarten children in Singapore. Annals Academy of Medicine Singapore, 29, 740-744.

Tannas, L.E. (1985). Flat panel displays and CRTS. (New York: Van Nostrand Reinhold.

Thibos, L.N. (2000). Principles of Hartmann-Shack Aberrometry. Journal of Refractive Surgery, 16, S563-S565.

Thibos, L.N., Applegate, R.A., Schwiegerling, J.T., Webb, R.H., & VSIA Standards Taskforce Members (2002). Standards for reporting the optical aberrations of eyes. Journal of Refractive Surgery, 18, S652-S660.

Thompson, A.M., Li, T., Peck, L.B., Howland, H.C., Counts, R., & Bobier, W.R. (1996). Accuracy and precision of the Tomey ViVA infrared photorefractor. Optometry and Vision Science, 73, 644-652.

Toraldo di Francia, G. (1949). Retina cones as dielectric antennas. Journal of the Optical Society of America, 39, 324.

References

143

Turner, M.J. (1958). Observations on the normal subjective amplitude of accommodation. British Journal of Physiological Optics, 15, 70-100.

van Blokland, G.J., & van Norren, D. (1986). Intensity and polarization of light scattered at small angles from the human fovea. Vision Research, 26, 485-494.

van de Kraats, J., Berendschot, T.T., & van Norren, D. (1996). The pathways of light measured in fundus reflectometry. Vision Research, 36, 2229-2247.

Van Herick, W., Shaffer, R.N., & Schwartz, A. (1969). Estimation of width of angle of anterior chamber. Incidence and sigificance of the narrow angle. Am J Ophthalmol, 68, 626-629.

Van Loo, J.A., & Enoch, J.M. (1975). The scotopic Stiles-Crawford effect. Vision Research, 15, 1005-1009.

Vohnsen, B., Iglesias, I., & Artal, P. (2005). Guided light and diffraction model of human-eye photoreceptors. Journal of the Optical Society of America A. Optics and Image Science, 22, 2318-2328.

Winn, B., Pugh, J.R., Gilmartin, B., & Owens, H. (1990). Arterial pulse modulates steady-state ocular accommodation. Current Eye Research, 9, 971-975.

Wolffsohn, J.S., Hunt, O.A., & Gilmartin, B. (2002). Continuous measurement of accommodation in human factor applications. Ophthalmic and Physiological Optics, 22, 380-384.

Wright, W.D., & Nelson, J.H. (1936). The relation between the apparent intensity of a beam of light and the angle at which it strikes the retina. Proceedings of the Physical Society London, 48, 401-405.

Wu, H., Seet, B., Yap, E.P., Saw, S., Lim, T., & Chia, K. (2001). Does education explain ethnic differences in myopia prevalence? A population-based study of young adult males in Singapore. Optometry and Vision Science, 78, 234-239.

Wyszecki, G., & Stiles, W.S. (1982). Color science: concepts and methods, quantitative data and formulae (2nd edition). (New York: John Wiley & Sons, Inc.

Zadnik, K., Satariano, W.A., Mutti, D.O., Sholtz, R.I., & Adams, A.J. (1994). The effect of parental history of myopia on children's eye size. Journal of the American Medical Association, 271, 1323-1327.

Zagers, N.P., van de Kraats, J., Berendschot, T.T., & van Norren, D. (2002). Simultaneous measurement of foveal spectral reflectance and cone-photoreceptor directionality. Applied Optics, 41, 4686-4696.

References

144

Zele, A.J., & Vingrys, A.J. (2005). Cathode-ray-tube monitor artefacts in neurophysiology. Journal of Neuroscience Methods, 141, 1-7.

Zhang, X., Ye, M., Bradley, A., & Thibos, L. (1999). Apodization by the Stiles-Crawford effect moderates the visual impact of retinal image defocus. Journal of the Optical Society of America A. Optics and Image Science, 16, 812-820.

Appendices

145

Appendices

Appendix 1: SCE apparatus

Figure A1.1. Photograph of the two-channel Maxwellian view SCE apparatus (Aerial view).

Appendices

146

Appendix 2: Calibration of the LCD monitor

A2.1 Introduction Liquid-crystal displays (LCD) are increasingly being used as computer monitors

because of their compact size and low power consumption. They offer both high

luminance (eg. 250 cd/m2) and contrast ratios (eg. 200:1), which is comparable to or

higher than Cathode Ray Tube (CRT) displays. CRT monitors are based on a raster

scan in which an electron beam scans the screen in a raster pattern (left to right, top

to bottom) (Sperling, 1971). The electron beams needed to activate the phosphors on

the screen are supplied by red, green and blue guns (cathode ray tubes). There are

three phosphors for each phosphor triad, and the colour of each pixel depends on

which phosphor the electron strikes. The LCD display is manufactured by deposition

of pixel electronics on a glass substrate. Each pixel consists of three sub-pixels,

which have red, green or blue filters in front of them, just as each pixel on a CRT has

RGB phosphors. The subpixels are made up of a group of liquid crystal molecules.

These molecules are suspended between transparent electrodes and are sandwiched

between two polarizing filters (Figure A2.1). An electric current passed through the

liquid causes the crystals to align so that light cannot pass through them. Each

crystal, therefore, is like a shutter, either transmitting or blocking the light. The

backlight is a cold cathode (Sharma, 2002; Tannas, 1985).

Appendices

147

Figure A2.1. Structure of a subpixel of an LCD.

To date, almost all the work with mfERGs has used CRT monitors. The objectives of

the following calibrations and analyses were to determine the performance of an

LCD monitor with respect to colorimetric characterization, spatial and temporal

properties and whether the performance is sufficient to consider this display for the

achromatic mfERG stimulus presentation used in this experiment.

CRT monitors produce high temporal frequency and luminance artefacts but are

powerful stimulus generators when used within their operating limits (García-Pérez

& Peli, 2001; Zele & Vingrys, 2005). Generally, the human eye has little sensitivity

to perceive high temporal frequency signals (> 40 Hz) but these artefacts can become

significant with mfERG stimuli that are presented at high temporal frequencies.

Keating, Parks, Malloch and Evans (2001) observed artefacts on the oscillatory

potential due to the refresh rate of the CRT that were not seen on an LCD display.

Few studies have compared CRT and LCD displays with regard to colour calibration

(Gibson & Fairchild, 2000; Sharma, 2002). A significant difference between CRT

and LCD monitors is the difference in their temporal properties. Brainard, Pelli and

Robson (2002) used a sequence with a single white frame followed by ten black

Coloured filter

polarizing filter

Liquid crystal molecules

Polarizing filter

Glass plates

backlight

Coloured filter

polarizing filter

Liquid crystal molecules

Polarizing filter

Glass plates

backlight

Appendices

148

frames and observed a delay in time-to-peak of 5.6 ms for an LCD monitor when

compared to a CRT, suggesting that CRTs are better than LCDs for studies involving

temporal processing than LCDs.

Research has been undertaken into the colorimetric characterization of LCDs.

Fairchild and Wyble (1998) evaluated the fundamental characteristics of an Apple

LCD Studio Display such as its spectral characteristics, chromaticity coordinates of

the primaries, additivity of the primaries and the luminance-voltage relationship.

Day, Taplin and Berns (2004) suggested that the accurate colorimetric

characterization of an LCD display would require measurements at multiple screen

positions and at different viewing angles.

Keating, Parks, Malloch and Evans (2001) compared the mfERG responses from a

CRT and an LCD. They used a fast response photodiode to measure the pulse widths

of an m-sequence that was adjusted such that the on-state consisted of four

consecutive white frames followed by four black frames and the off-state consisted

of eight black frames. They found that the LCD produces a 13.3 ms pulse which is

equivalent to a 75 Hz frame rate. A CRT can produce 2 ms pulse widths (Keating,

Parks, Malloch & Evans, 2001; Brainard, Pelli & Robson, 2002; Zele & Vingrys,

2005). Because an LCD system has a longer pulse width, it can provide information

on the retina’s ability to recover from longer periods of bleaching pulses and allow

proper separation of onset and offset components. For a CRT system, the first order

mfERG response is defined as the sum of transitions to a white stimulus minus the

sum of transitions to a black stimulus from the mean adaptation level. For an LCD

system, an offset response is generated for the white to black state, and no response

is generated for a white to white state, the first order response is the difference

between onset and offset responses. Therefore, the first order response amplitudes

from mfERG responses are larger for the CRT system than with the LCD system

(Keating, Parks, Malloch & Evans, 2001). The LCD system was suggested to be an

alternative to the CRT stimulus delivery system (Keating, Parks, Malloch & Evans,

2001). However, due to inherent limitations in the control circuitry of LCD systems,

few LCDs might be suitable for multifocal studies.

Appendices

149

The following sections evaluate several aspects of LCD monitor performance in

terms of voltage-luminance relationship, warm-up characteristics, spatial variance,

temporal channel profile, spectral characteristics, chromaticity constancy of

primaries, channel independence, and channel constancy.

A2.2 Apparatus

LCD monitor: The LCD monitor used in the experiment was a 19” (viewable

diagonal area) flat-panel RGB monitor (Viewsonic VX 922, ViewSonic Corporation,

Walnut, CA). This monitor has an antiglare glass surface. It is a thin film transistor

(TFT) active-matrix liquid crystal display with 1280 x 1024 pixel resolution with a

refresh rate of 75 Hz (13.3 ms). The monitor was set to its default internal D65 white

point setting. The monitor was connected to a Cambridge Research Systems visual

stimulus generator (VSG 2/5 and ViSaGe) to perform colorimetric and luminance

measurements using photometers. For the mfERG experiments, the monitor was

controlled by a 24-bit video card (VERISTM Inc).

Photometry: The photometers were commercially available silicon cells (optiCALTM

and colorCALTM, Cambridge Research Systems). Both the optiCAL and colorCAL

were driven by vsgDesktop display calibration software. The optiCAL photometer

(VSG 2/5) uses an RS-232 interface and can be connected to an oscilloscope using a

BNC cable. This photometer can measure luminances up to 2400 cd/m2. The

colorCAL colorimeter communicates via USB interface (1.5 Mbps) and is

compatible with latest version of VSG (ViSaGe) whose analog I/O interface can be

connected to the Viewsonic LCD monitor. The luminance range of colorCAL is 0.2 –

200 cd/m2. Due to the compatibility of the ViSaGe with the Viewsonic monitor, I

used colorCAL colorimeter to measure the voltage-luminance relationship (section

5.2.4.4) and the optiCAL photometer for other measurements such as warm-up time,

spatial variance of luminance and temporal luminance profile (sections 5.2.4.3,

5.2.4.5 and 5.2.4.6). Both the silicon cells were calibrated and warmed-up for at least

15 minutes before all measurements.

Appendices

150

Spectroradiometer: Spectral radiance measurements were made using an EPP2000

StellarNet Fiber Optic Spectrometer (StellarNet Inc) at a fixed distance of 10 mm,

perpendicular to the centre of the display surface. The spectroradiometer was set to

its default CIE standard illluminant of D65 (daylight) and provided full spectral

radiance (in W/sr/m2) at a 0.50 nm sampling resolution in the range 270–901.5 nm. A

darklight scan was done before starting reference measurements using the

spectroradiometer to eliminate background electronic signalling.

Oscilloscope: A Tektronix TDS 210 60MHz 1GS/s digital real-time oscilloscope was

used to record the temporal luminance profiles via the photometer connected to the

y-channel of the oscilloscope.

All data were collected in a dark room with minimal stray light as is typically used in

psychophysical and mfERG experiments.

A2.3 Warm-up characteristics

Some studies have reported that LCDs should be allowed to warm-up for at least

30–40 minutes (Brainard, Pelli & Robson, 2002; Fairchild & Wyble, 1998). I

determined the warm-up characteristics for the LCD monitor used in this study in a

similar way to that proposed by Metha, Vingrys and Badcock (1993) for a CRT

monitor.

Warm-up characteristics were measured for the monitor running from a “cold start”

(off period > 24 hours) and from a “warm start” (restarted after 1 hour off following

a 2 hour on period). The test image was the stationary achromatic mfERG stimulus

array used in the main study and displayed in the centre of the LCD. Luminance

measurements were made every 5 minutes for 1 hour using the optiCAL photometer

placed at the central white hexagon of the mfERG stimulus using the suction cup of

the device.

Appendices

151

Figure A2.2. Luminance for cold and warm starts (A) and chromaticity measurements for cold

starts (B and C) as a function of time. The filled circles represent data from a cold start and the

open circles represent data from a warm start.

0.188

0.189

0.190

0.191

0.192

0 10 20 30 40 50 60

Time (min)

u

0.321

0.322

0.323

0 10 20 30 40 50 60

Time (min)

v

B

C

280

290

300

310

320

0 10 20 30 40 50 60

Time (min)

Lu

min

ance

(cd

/m2 )

Cold start

Warm start

A

Appendices

152

Figure A2,2A shows that a stable luminance was achieved within 50 minutes after a

cold start and within 40 minutes after a warm start. The greatest variability in

luminance was found in the first 40 minutes of both the warm and the cold starts,

during which luminance decreased by approximately 11 %. It should be noted that

this is opposite to a CRT for which luminance increases over time (Metha, Vingrys

& Badcock, 1993). The chromaticity was measured in 1931 CIE system and then

converted to 1964 CIE (U*V*W*) system according to equation A2.1.

)3122/(6

)3122/(4

yxyv

yxxu (A2.1)

Figures A2.2 B and C show that the chromaticity changed by u = -0.002 and v =

+0.001. The warm-up duration should ensure that the human eye is not able to

perceive differences in the chromaticity luminance output during the experimental

procedure. Therefore, the warm-up time specified in terms of total colour difference

(C*uv) that takes into account luminance and chromaticity variations needs to be

considered (Wyszecki & Stiles, 1982). The total colour difference is scaled in JNDs

(just noticeable differences) or psychometric units and calculated according to

equation A2.2.

2/1222 ])()()[( finalfinalfinal WWVVUUC

where

YWvvWVuuWU );(13);(13 00 (luminance) (A2.2)

Ufinal and Vfinal are calculated from the final colour achieved (ufinal, vfinal); u0, v0 =

values of the variables u,v for the white hexagons of the mfERG stimulus.

Figure A2.3 shows the total colour difference calculated for each measured colour

with respect to the final colour achieved for a cold start. Adopting the criterion that

the screen should not vary by more than 1 JND over 1 hour of use, a warm-up time

of 50 minutes is required. I recommend that this LCD monitor should be warmed-up

for at least 50 minutes after the cold start and 40 minutes after the warm start so that

Appendices

153

the luminance and chromaticity outputs are stable. The monitor was warmed-up for

at least 60 min before any experiment was commenced.

0

5

10

15

20

25

30

35

40

0 10 20 30 40 50 60

Time (min)

del

ta C

*uv

(jn

d)

Figure A2.3. Change in total colour difference (C, jnd) for cold start with respect to the final

stable colour achieved.

A2.4 Voltage-luminance relationship

The voltage-luminance relationship describes the relationship between the voltage

signal used to drive the LCD display and the luminance output. To characterize any

computer-controlled device, this relationship needs to be modelled. The relationship

is nonlinear. Such a relationship for CRT displays is referred to as “gamma” or

“gain-offset-gamma” (GOG) (Berns, 1996; Berns, Motta & Gorzynski, 1993).

In the GOG model, the relationship between voltage and luminance is described by

the following equation,

iVV 0 (A3.3)

where V0 is the physical output or the luminance measure, Vi is the input voltage, α is

the dark light of the screen (the channel activation at zero voltage), β is a constant

and gamma () is the exponent of Vi. It has been suggested that the dark light term

“α” should be included in gamma correction, even though the ambient room

illumination which contributes to α is zero (Metha, Vingrys & Badcock, 1993).

Appendices

154

For LCD monitors, the relationship between voltage and luminance has been named

electro-optical transfer function (EOTF) (Fairchild & Wyble, 1998; Gibson &

Fairchild, 2000). Glasser (1997) reported that the transfer function of LCD and CRTs

are different such that CRT gamma or GOG models cannot be used to accurately

characterize LCD displays.

Fairchild & Wyble (1998) studied the colorimetric characterization of an Apple

Studio LCD display and observed that the gamma function or GOG model describes

the EOTF of the LCD monitor well. The model however, shows small systematic

errors (< 2%) in the fits, particularly for the red and green channels. The errors were

large at high voltages. However, the characterization performance was excellent with

one-dimensional (1D) look-up tables (LUTs). Bastani, Cressman & Funt (2005)

showed that the 3D LUT model was more accurate than the 1D LUT model, but was

too complex to be used. When Gibson & Fairchild (2000) evaluated colorimetric

characteristics of one CRT (Sony GDM-F500) and two LCD displays (SGI 1600SW

and IBM prototype), they found that both Sony and SGI displays could be

characterised using a GOG model but the IBM LCD display showed high errors with

the GOG model.

I evaluated voltage-luminance relationship of the Viewsonic LCD monitor using the

gamma model, primarily for easy determination of the inverse LUTs. The LUTs

were generated using the ViSaGe provided by Cambridge Research Systems. To

obtain a LUT, the colorCAL colorimeter was attached with the suction cup to the

centre of the test image (400 x 400 pixel square; x, y = 0.338, 0.361) displayed on

the centre of the screen. These analog outputs were digitized by an analog-to-digital

converter (Cambridge Research Systems) and extracted in readable format using

Matlab (Mathworks Inc., Natick, MA, version R2006a). The obtained LUTs were

composed of 16385 readings and there was one LUT for each red, green and blue

channel. Figure A2.4 shows the relationship between input voltage and output

luminance for each of the three channels. It should be noted that all three curves are

truncated towards the end of the curve (~ 0.98 Volts) due to the limitation of ViSaGe

graphics card in which input voltage saturates beyond voltage of 0.98 volts. The

maximum luminance of the monitor was 280 cd/m2 measured with the Optical

Appendices

155

photometer at 98 % of the maximum voltage. The entire calibration was performed

in 45 minutes.

Figure A2.4. Measured LUT for the three colour channels.

The chromaticity and luminance outputs of the Viewsonic LCD monitor are given in

Table A2.1.

Table A2.1 Chromaticity coordinates and maximum luminance for the channels following

colorimetric characterization.

R, G, B x y Measured maximum luminance (cd/m2) using

ViSaGe 255, 0, 0 (red) 0.622 0.313

44

0, 255, 0 (green) 0.259

0.615

146

0, 0, 255 (blue) 0.148

0.068

15

Appendices

156

To linearize the GOG function, an inverse function is obtained and a compensating

LUT is generated. The inverse function was obtained using an iterative procedure

and a compensating LUT was generated.

The inverse function has the form,

/10 /' VV (A2.4)

where V’ is the required output voltage to linearize the function, and V0, α, β and γ

are as defined in equation A2.3. Figure A2.5 shows the inverse function curves for

the red channel as an example.

Figure A2.5. Inverse LUT for red channel. The lower curve (solid) is the modelled gamma

function and the top curve (dashed) is the inverse of the model. The numbers indicate the input

and the output values and the arrows indicate the direction to achieve linear outputs. Schematic

is based on the Figure 11 of Metha, Badcock & Vingrys (1993).

In Figure A2.5, for a linear system a relative input voltage of 0.6 (1) should return a

relative output of 0.6. But due to the nonlinearity of the voltage-luminance

relationship, the relative output was 0.22 (12). Therefore, in order to correct the

nonlinearity an inverse function (upper dashed curve) needs to be determined that is

Appendices

157

sent to the LUT. When the desired value of 0.6 (1) enters the LUT, a value of 0.82

(3) was returned. The value 0.82 (4) as the relative input voltage would return the

desired relative output of 0.6 (5).

When the inverse function is applied to the LUTs, the relationship between input and

output voltages becomes linear (Figure A2.6). We can see in Figure A2.6 that the

obtained relationship curves do not intercept the y-axis (output) at zero but intercept

at 5% of the maximum voltage of all the channels, which suggests that there was

dark light of LCD present during the measurements. Also, the curves were truncated

beyond 98 % of the operating range probably due to limitation of the ViSaGe

graphics card. Therefore, the operating range for the Viewsonic LCD monitor should

be limited between 5 % - 98 % of the operating range.

Appendices

158

Figure A2.6. Input-output relationships for the R, G, B channels after gamma correction.

Appendices

159

A2.5 Spatial uniformity of the LCD For most CRT monitors, the luminance is highest at the centre and falls off toward

the edges. Metha, Vingrys & Badcock (1993) suggested that the image should be

small and restricted to the centre of the screen for the CRT monitor. LCD

performance is much better with regard to spatial homogeneity than that of CRTs

(Sharma, 2002). The test image used was the stationary mfERG stimulus used in this

study and displayed at the centre of the monitor (Figure A2.7). I measured the white

hexagons luminance and chromaticity coordinates (x, y) at different positions for the

LCD monitor following a warm-up time of 50 minutes. The CIE 1931 x,y

chromaticity coordinates were transformed to CIE 1964 u,v chromaticity coordinates

using equation A2.1. A 4% difference in luminance was found from centre to the

edge of the test image. The chromaticity coordinates showed a maximum change of

0.3% for u-chromaticity and no change for v-chromaticity with respect to change in

location.

Figure A2.7. Luminance (cd/m2) and colorimetric values for the monitor at various locations.

Monitor and mfERG stimulus subtended 43° x 34° and 28° x 23°, respectively at a distance of 50

cm. W is the luminance (cd/m2) and u, v are the chromaticity coordinates (CIE 1964).

mfERG stimulus (28° x 23°)

W = 272u = 0.189v = 0.213

W = 278u = 0.189v = 0.213

W = 274u = 0.190v = 0.213

W = 282u = 0.189v = 0.213

W = 270u = 0.189v = 0.213

W = 249

Monitor (43° x 34°)

mfERG stimulus (28° x 23°)

W = 272u = 0.189v = 0.213

W = 278u = 0.189v = 0.213

W = 274u = 0.190v = 0.213

W = 282u = 0.189v = 0.213

W = 270u = 0.189v = 0.213

W = 249

Monitor (43° x 34°)

Appendices

160

A2.6 Temporal luminance profile

To understand the output timing of the LCD display and compare it with a CRT

display, the temporal luminance profile was recorded. The channel activation and

decay were measured using optiCAL photometer (1° arc) for a single 3° centre

hexagon modulated according to the mfERG pseudorandom sequence. It should be

noted that signals recorded over an extended region of 1 causes temporal smearing

of the luminance profile (Zele & Vingrys, 2005).

The luminance profiles (waveforms) were measured by the oscilloscope and imaged

with a digital camera (Canon Ixus 85 IS). Comparison was made between the

waveforms obtained from the LCD (ViewSonic, RGB, 75 Hz frame rate, 280 cd/m2)

and a CRT monitor (FIMI MD0709BRM, a 21” monochrome display, 75 Hz frame

rate, 100 cd/m2). Both monitors were calibrated according to standard procedures. To

analyse the oscilloscope trace, GraphClick 3.0 software was used. The normalized

temporal profile of both CRT and LCD monitors are shown in Figure A2.8.

In Figure A2.8, the temporal luminance profile of the CRT has a short pulse of 5 ms

duration with half height of approximately 1.5 ms. The rise time to maximum was

0.5 ms and decays to noise within 5 ms. This luminance profile of the CRT acquired

with the 1 photosensor was similar to that obtained by Zele & Vingrys (2005). For

the LCD monitor, the luminance profile is not actually a pulse but a frame-length

block and decays over a longer time-frame than for a CRT. The rise time to

maximum for the LCD monitor was 6 ms and decays to noise within 14 ms (Figure

A2.8).

Appendices

161

Time (msec)

0 2 4 6 8 10 12 14

No

rma

lize

d A

mp

litu

de

0.0

0.2

0.4

0.6

0.8

1.0

1.2CRTLCD

Figure A2.8. Oscilloscope traces for a pseudorandom sequence of mfERG stimulus for both

CRT and LCD systems. The horizontal scale is the time in milliseconds (ms).

Consistent with the results reported by Keating, Parks, Malloch & Brainard (2001)

and Brainard, Pelli & Robson (2002), the time-to-peak from the start of the video

signal for the LCD luminance profile was delayed by approximately 5.5 ms in

comparison to the CRT.

A2.7 Spectral characteristics For the purpose of colour characterization, the display was allowed to warm-up for

60 minutes and a test image (400 x 400 pixel square) was displayed in the central

region of the LCD. Spectral radiance measurements were recorded in 0.50 nm

sampling intervals and analysed from 400-800 nm. The spectral output was measured

at the maximum voltage for each of the red, green and blue channels (for example,

the blue-channel characterization test image consisted of R = G = 0, B = 255). The

spectral characteristics of the Red, Green and Blue channels of the display are

illustrated in Figures A2.9.

Appendices

162

Wavelength (nm)

400 450 500 550 600 650 700

Rad

ian

ce (

W/S

t/m

2 /nm

)

0.0001

0.001

0.01

0.1

RedGreenBlue

Figure A2.9. Spectral radiance distribution of the Red (solid), Green (dashed) and Blue (dotted)

channels of the LCD display.

The peak spectral radiance of the red channel is 610 nm. The green channel has a

peak at 547 nm with two smaller peaks at 487 nm and 582.5 nm that are overlapping

with the blue and red channels, respectively. The blue channel has two peaks at 435.5

nm and 487.5 nm. This spectral distribution for the Viewsonic monitor is similar to

the spectral distribution reported by Fairchild & Wyble (1998) for the Apple Studio

display (Apple Computer, Inc.).

A2.8 Chromaticity constancy of the LCD primaries

Figure A2.10 shows the chromaticities of the red, green and blue channels operating

at 10 %, 50 % and 90 % of their maximum outputs for the LCD display in relation to

the spectrum locus on the CIE 1931 and CIE 1964 chromaticity diagram.

Appendices

163

The triangle formed by the three channel chromaticity-coordinates represents the

achievable gamut for the display in chromaticity space (Figure A2.10 A and B).

Sharma (2002) showed that LCDs have larger gamuts than those of CRTs,

particularly in the dark colour regions around black, which he suggested to be due to

the higher luminance of LCD displays than CRT displays (Figure A2.10). The

chromaticity gamut of the Viewsonic monitor extends up to the spectral locus of

chromaticity space in the red colour region. The larger gamut indicates that the

monitor can display highly saturated colours. Note that the chromaticity gamuts for

all three operating levels (10%, 50% and 100%) were similar in size which indicates

that these three different operating levels would produce the same range of colours.

In the same chromaticity diagram, the chromaticities for the white-point are also

plotted. The white-points are important for this study because the mfERG stimulus is

black-and-white hexagons. The white-points at different operating levels are very

close to each other, that is, the white-point chromaticity is independent of operating

levels of the channels.

Figure A2.10. Measured chromaticities of red, green and blue primaries of the Viewsonic

monitor in CIE 1931 system (A) and CIE 1964 system (B).

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9

x

y

50%

10%

100%

Sharma(2002)

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9

x

y

50%

10%

90%

0

0.1

0.2

0.3

0.4

0.5

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7

u

v

50%

10%

90%

A B0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9

x

y

50%

10%

100%

Sharma(2002)

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9

x

y

50%

10%

90%

0

0.1

0.2

0.3

0.4

0.5

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7

u

v

50%

10%

90%

A B

Appendices

164

A2.9 Channel independence

Each channel should function independently of other channels such that, if the output

of one channel changed, the output of other channels should not change. This is

called channel independence. To test this requirement, the spectral radiance of the

display was measured using the spectroradiometer with one channel set at 10% and

90% of maximum output while other channels were kept constant at 50% and/or 0%

output (Figure A2.11). The inset to each panel of Figure A2.11 shows the log

difference between the two curves in the corresponding panel. In Figure A2.11,

Panels A and B demonstrate the change in the red channel at 90% and 10% when the

green and blue channels were decreased from 50% to 0%, Panels C and D

demonstrate change in the green channel when the red and blue channels decreased

from 50% to 0% and Panels E and F show changes in the blue channel when the red

and green channels decreased from 50% to 0%. Comparing the inset between Panels

A and B shows that the spectral output for red channel changed less than 1 log units

at both 10% and 90 % operating levels. Similar findings can be observed for the

green and blue channels at 50% and 10% operating levels when red, blue channels

and red, green channels were varied from 0% to 50%.

Appendices

165

Wavelength (nm)

400 450 500 550 600 650 700

Rad

ian

ce (

W/S

t/m

2 /n

m)

0.0001

0.001

0.01

0.1 G, B = 0% and R = 90%

G, B = 50% and R = 90%

400 500 600 700-1.0

-0.5

0.0

0.5

1.0

Wavelength (nm)

400 450 500 550 600 650 700

Rad

ian

ce (

W/S

t/m

2 /n

m)

0.0001

0.001

0.01

0.1 G, B = 0% and R = 10%

G, B = 50% and R = 10%

400 500 600 700-1.0

-0.5

0.0

0.5

1.0

Wavelength (nm)

400 450 500 550 600 650 700

Rad

ian

ce (

W/S

t/m

2 /n

m)

0.0001

0.001

0.01

0.1 R, B = 0% and G = 90%

R, B = 50% and G = 90%

400 500 600 700-1.0

-0.5

0.0

0.5

1.0

Wavelength (nm)

400 450 500 550 600 650 700

Rad

ian

ce (

W/S

t/m

2 /n

m)

0.0001

0.001

0.01

0.1 R, B = 0% and G = 10%

R, B = 50% and G = 10%

400 500 600 700-1.0

-0.5

0.0

0.5

1.0

Wavelength (nm)

400 450 500 550 600 650 700

Rad

ian

ce

(W/S

t/m

2/n

m)

0.0001

0.001

0.01

0.1 R, G = 0% and B = 90%

R, G = 50% and B = 90%

400 500 600 700-1.0

-0.5

0.0

0.5

1.0

Wavelength (nm)

400 450 500 550 600 650 700

Rad

ian

ce (

W/S

t/m

2/n

m)

0.0001

0.001

0.01

0.1 R, G = 0% and B = 10%

R, G = 50% and B = 10%

400 500 600 700-1.0

-0.5

0.0

0.5

1.0

A B

C D

E F

Figure A2.11. Channel independence for a channel at two output levels 10% and 90% of

maximal output when the other two channels are at 0% and 50 % of maximal output. The y-

axis scales are in logarithmic scale. The insets to each panel show the log difference between the

curves in the corresponding panel. The axes for insets are same as the main plots, except the y-

axes of the insets are in log linear scale.

Channel independence can also be checked by measuring the output of each channel

separately and then checking if the measured output of a combination of channels

(white background; R+G+B=255) is equal to the sum of their individual

contributions obtained from measurement.

Appendices

166

Wavelength (nm)

400 450 500 550 600 650 700

Rad

ian

ce (

W/S

t/m

2 /nm

)

0.0001

0.001

0.01

0.1 WhiteSum of RGB guns

400 500 600 700-0.4

-0.2

0.0

0.2

0.4

Figure A2.12. Measured spectral output of individual and combined channels at the maximum

operating levels, that is, 98 % of the maximum. The y-axis is in logarithmic scale. The inset

shows the log difference between the curves. The axes for the inset is same as the main plot,

except the y-axis of the inset is in log linear scale.

Figure A2.12 shows the spectral radiance output as a function of wavelength for the

white test image (R+G+B = 255) and for all three channels operating at 100% of

their maximal output. The difference curve shown (log R+G+B – log R, G, B) in the

inset illustrates a maximum difference of 0.22 log unit between the measured and

calculated spectral output. Combined findings from Figure A2.11 and Figure A2.12

suggest that the channels function independently of each other.

Appendices

167

A2.10 Channel constancy

When the spectral power radiance of a channel is independent of its operating level,

channel constancy is achieved (Metha, Vingrys & Badcock, 1993). It is important for

the monitors to have spectral power distributions that are independent of intensity

level. To check for channel constancy, the spectral output of a channel was measured

at 90% and 50% operating levels and the radiance was plotted against the

wavelength. If channel constancy is achieved, then the curves would have similar

shape at different operating levels. Figures A2.13 A, B, and C show the spectrum of

red, green and blue channels of the LCD monitor respectively, at two operating

levels (90 % and 50 %). The curve for a channel at 90% operating level has similar

shape to that of 50% operating level. Note that for the LCD display the curves not

only reduce vertically but also shrink horizontally from 90 % to 50 % operating

level. The insets to each panel show the difference between 90 % and 50 % operating

levels. The difference was calculated after reducing the curves at 90 % operating

level to 50 % operating level. A log difference of up to 1 log units exist when the

monitor operates from 90 % to 50 % of the maximum output.

Appendices

168

400 450 500 550 600 650 700

Rad

ian

ce (

W/S

t/m

2 /n

m)

0.0001

0.001

0.01

0.1

Red 90%

Red 50%

400 450 500 550 600 650 700

Rad

ian

ce

(W/S

t/m

2/n

m)

0.0001

0.001

0.01

0.1

Green 90%

Green 50%

Wavelength (nm)

400 450 500 550 600 650 700

Ra

dia

nce

(W

/St/

m2 /

nm

)

0.0001

0.001

0.01

0.1

Blue 90%

Blue 50%

400 500 600 700-1.0

-0.5

0.0

0.5

1.0

400 500 600 700-1.0

-0.5

0.0

0.5

1.0

400 500 600 700-1.0

-0.5

0.0

0.5

1.0

A

B

C

Figure A2.13. Spectral radiance measurements of red, green and blue channels as a function of

wavelength (nm) at operating levels 90% and 50% are shown in logarithmic scale in panels A,

B, and C, respectively. The insets to each panel show the log difference between the curves in

the corresponding panel. The axes for insets are same as the main plots, except the y-axes of the

insets are in log linear scale.

Appendices

169

To substantiate the violation of channel constancy from the ideal state, the

chromaticity coordinates were analysed for 10%, 50%, and 90% operating levels. If

channel constancy was achieved, the chromaticity coordinates should not change as

output varies. Figure A2.14 shows the chromaticity coordinates (CIE 1964, uv) for

each channel (transformed from 1931 CIE system using equation A2.1) and the

relative chromatic difference for each operating level from the 90 % output when

only one channel is operating. We also analysed chromatic difference (C*uv) in

isolation by assuming that the chromaticity differences were not accompanied by any

corresponding changes in luminance (using equation A2.2). The chromatic difference

was calculated relative to the colour at 90 % operating level. This is scaled in JNDs

(Figure A2.15).

Figure A2.14. Chromaticity coordinates (u, v) and differences in u and v from the 90 % values

are plotted against channel operating levels for each channel.

Gun operating level (%)

CIE

196

4 u

,v


Dif

fere

nce

in

u o

r v

(CIE

196

4)

0

0.1

0.2

0.3

0.4

0.5

10 30 50 70 90

Green

0

0.1

0.2

0.3

0.4

0.5

10 30 50 70 90

0

0.1

0.2

0.3

0.4

0.5

10 30 50 70 90

u

vRed

Blue

-0.4

-0.3

-0.2

-0.1

0

0.1

0.2

0.3

0.4

10 30 50 70 90

u

v

-0.4

-0.3

-0.2

-0.1

0

0.1

0.2

0.3

0.4

10 30 50 70 90

-0.4

-0.3

-0.2

-0.1

0

0.1

0.2

0.3

0.4

10 30 50 70 90

Red

Green

Blue


CIE

196

4 u

,v


Dif

fere

nce

in

u o

r v

(CIE

196

4)

0

0.1

0.2

0.3

0.4

0.5

10 30 50 70 90

Green

0

0.1

0.2

0.3

0.4

0.5

10 30 50 70 90

0

0.1

0.2

0.3

0.4

0.5

10 30 50 70 90

u

vRed

Blue

-0.4

-0.3

-0.2

-0.1

0

0.1

0.2

0.3

0.4

10 30 50 70 90

u

v

-0.4

-0.3

-0.2

-0.1

0

0.1

0.2

0.3

0.4

10 30 50 70 90

-0.4

-0.3

-0.2

-0.1

0

0.1

0.2

0.3

0.4

10 30 50 70 90

Red

Green

Blue

Appendices

170

0

0.1

0.2

0.3

0.4

0.5

10 30 50 70 90


C*u

v (j

nd

)Red

Green

Blue

Figure A2.15. Calculated colour difference (C*uv) plotted against operating level for each

channel.

Figure A2.14 shows that the u-coordinates were more variable than the v-

coordinates. The green and blue colours appear to be saturated when operated at low

level (10%), with the blue colour to be the most saturated. Exceptionally, the red

channel showed a deviation in u-coordinate at 50% output from other operating

levels. Thus, the channels do not achieve constancy. From Figure A2.15, it is

confirmed that the deviations from channel constancy may be significant at operating

levels of 10 % or below, especially for the green and blue channels.

Channel constancy of a monitor is not under the control of the user and the low

operating level should be avoided for the colour representation in these LCD

monitors.

A2.11 Summary

The Viewsonic VX922 monitor used in this study should be warmed up for at least

50 minutes before starting experiments (Figure A2.2). The spatial variation in

chromaticity and luminance was approximately 4% from the centre to the edge of the

stimulus pattern (Figure A2.7). The centre of the display should be used to avoid any

large variation in the luminance across the screen. The temporal channel profile of

this LCD monitor showed a lag of 5 ms to reach maximum luminance from the start

Appendices

171

of video signal compared to a CRT, which should be accounted for in the multifocal

ERG (section 5.2.4.5).

The luminance range for the display should be limited to between 10% - 98% of the

maximum output (Figure A2.6). Similar to Sharma (2002), the chromaticity gamut

for the Viewsonic monitor is also found to be larger than that of a CRT and the

different operating levels would produce the same range of colours. The white-point

chromaticity lies at the same location in the chromaticity space for different

operating levels (Figure A2.10). Also, the channel constancy was not achieved fully,

suggesting that the lower end of the monitor’s dynamic range (< 10 %) should be

avoided for colour representation (Figure A2.15).

For this Viewsonic monitor, the channels operate independently of each other and the

sum of spectra for three channels is comparable to the spectra for white stimuli

(Figure A2.12). Although, there is a lack of channel constancy below 10% operating

level, it should not have any impact on the achromatic mfERG stimulus if the

operating range is limited to 10% - 98% (Figure A2.15).

Within the constraints specified here, the LCD monitor (Viewsonic VX 922,

ViewSonic Corporation, Walnut, CA) can be reliably used to display the achromatic

mfERG stimulus.

Appendices

172

Appendix 3: Publications/Presentations

Following are the publications and presentations which have arisen from or are

related to the work presented in this thesis:

Journal article: Singh N, Atchison DA, Kasthurirangan S, Guo H (in press). Influences of accommodation and myopia on the foveal Stiles-Crawford effect. Journal of Modern Optics accepted 3 January 2009 (DOI: 10.1080/09500340902721915).

Presentations: Singh N, Moss F, Dinh D, Barron A, Tsigournis A, Atchison DA, Lambert AJ. Effect of luminance and colour on the foveal Stiles-Crawford effect http://aaopt.org/Submission/PPP/ViewPPP.asp American Academy of Optometry Annual General Meeting, December 2005. Singh N, Atchison DA, Kasthurirangan S. Accommodation induced changes in the foveal Stiles-Crawford effect of the first kind. 3rd European Meeting in Physiological Optics, London, 8 September 2006. Atchison DA, Singh N, Kasthurirangan S, Guo H. Effect of accommodation on the Stiles-Crawford effect http://www.opticsinfobase.org/abstract.cfm?URI=FiO-2008-SThE3 Frontiers in Optics, Annual Meeting, Optical Society of America, Rochester, New York, USA, 23 October, 2008.

Variation of the Stiles-Crawford Effect with Accommodation ... · Variation of the Stiles-Crawford...

Documents

Transcript of Variation of the Stiles-Crawford Effect with Accommodation ... · Variation of the Stiles-Crawford...