Corneal topography, near work and eyelid forces Tobias F. Buehren Diplom Ingenieur (FH) Augenoptik A thesis in partial fulfilment of the requirements for the degree of Doctor of Philosophy Centre for Eye Research School of Optometry Queensland University of Technology Brisbane, Australia 2003
216
Embed
Corneal topography, near work and eyelid forceseprints.qut.edu.au/15882/1/Tobias_Buehren_Thesis.pdf · aberrations of the human eye has grown rapidly. Aberrations of the whole eye
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Corneal topography, near work
and eyelid forces
Tobias F. Buehren
Diplom Ingenieur (FH) Augenoptik
A thesis in partial fulfilment of the requirements
for the degree of Doctor of Philosophy
Centre for Eye Research
School of Optometry
Queensland University of Technology
Brisbane, Australia
2003
KEYWORDS
Videokeratoscope, Cornea, Corneal topography, reading, lid forces, aberrations,
refractive error, myopia
ABSTRACT
The cornea is the most powerful refractive component of the eye and as such, subtle
changes in corneal shape can cause substantial changes in the optical characteristics of
the eye. Monocular diplopia has previously been linked to corneal distortion
following near work in various studies but has not been investigated in detail. The
work reported in this thesis has investigated the optical effects of corneal distortions
caused by eyelid forces and demonstrated that several corneal higher and lower order
Zernike wavefront aberrations can change following reading.
Measuring subtle changes in corneal topography requires the highest possible
instrument accuracy, while software analysis tools should be able to detect and
highlight those subtle changes with high reliability. The effect of ocular
microfluctuations on the qualitative and quantitative analysis of corneal topography
was investigated. A technique was developed to measure tilt, displacement, and
cyclotorsion in multiple videokeratographs from the same cornea. This information
was used to reposition each videokeratograph according to the average position of a
sample of multiple measurements. The corneal topography of ten subjects was
measured 20 times each, using videokeratoscopy. The RMSE calculated from
difference between single videokeratographs and the average videokeratograph
decreased by an average of 24.6 % for the ten subjects’ data. The method can improve
the precision performance of videokeratoscopy in multiple measurements of corneal
topography.
I
A study was undertaken, to investigate whether there are significant changes in
corneal topography during accommodation in normal corneas and corneas that are
pathologically thinner due to keratoconus. This was done to eliminate the possibility
that changes in corneal aberrations associated with near work could be at least partly
due to corneal changes caused by the effects of accommodation. A videokeratoscope
was modified to present an accommodation stimulus that was coaxial with the
instrument’s measurement axis. Six subjects with normal corneas and four subjects
with keratoconus were studied. In the initial analysis it was found that a number of the
subjects showed significant changes in corneal topography as accommodation
changed. However further analysis showed a significant group mean excyclotorsion of
the topography maps for both accommodation stimuli compared with the 0 D
stimulus. When the excyclotorsion was accounted for, no clear evidence of
statistically significant changes in corneal topography as a result of accommodation
were found. A small ocular excyclotorsion typically accompanies accommodation and
this changes the relative orientation of the topography of the cornea.
To investigate the effects of eyelid pressure on corneal shape and corneal aberrations
during reading, twenty young subjects with normal ocular health were recruited.
Cornea1 topography of one eye was measured with a videokeratoscope prior to
reading and then again after a 60 minute reading task. Twelve of the twenty corneas
showed significant changes in central topography immediately following reading. The
location of the changes corresponded closely to the position and angle of the subject’s
eyelids during reading. Within the central region of the cornea there were significant
changes in corneal wavefront Zernike coefficients, the root-mean-square error, overall
refractive power and astigmatism. The changes observed in corneal topography
II
appear to be directly related to the force exerted by the eyelids during reading. These
findings may have important implications for the definition of refractive status and
may also aid in the understanding of the relationship between reading and the
development of refractive errors.
To study whether corneal distortions after reading significantly differ between
refractive error groups, corneal aberrations were measured before and after a period of
reading, for a group of ten young progressing myopes and a group of ten young stable
emmetropes. The major difference between the two groups was the location and
magnitude of the corneal distortions, which had a significantly larger effect on central
corneal optics in the myopic group compared to the emmetropic group. A
significantly smaller palpebral aperture for the myopic group in the reading gaze
position was the cause of this difference.
The experiments described in this thesis have shown that numerous corneal
characteristics can change due to eyelid forces during near work. The eye was shown
to undergo a small cyclotorsion during higher levels of accommodation. There was a
shift in direction of against the rule astigmatism of the cornea following reading and a
change was found for primary vertical coma and trefoil. The changes in corneal shape
following reading appear to be different in myope versus emmetropic refractive error
groups. These findings are important for our understanding of the stability of the
refractive error of the eye and could have important implications for refractive error
development.
III
CONTENTS
Abstract I
Contents IV
Statement of Authorship VIII
Acknowledgements X
Chapter 1:
1.1
1.2
1.3
1.4
1.5
Introduction
Videokeratoscopy
1.1.1 Placido-based Instruments
1.1.2 Technical requirements
1.1.3 Corneal topography reconstruction
1.1.4 Topographic displays
1.1.5 Limitations of videokeratoscopes
1.1.6 Accuracy and repeatability
Corneal shape
Corneal mechanics
1.3.1 Hydration effects
1.3.2 Tear instability
1.3.3 Mechanical properties
1.3.4 Stability of corneal shape
1.3.5 Diurnal variations
1.3.6 Corneal epithelia1 cell movement
The effect of lid pressure on corneal topography
1.4.1 Monocular diplopia associated
with near work
Eyelids and blinks
1.5.1 Blinking
1 .5.2 Eyelidpressure
Page
1
3
3
4
7
8
8
11
16
19
19
20
21
23
24
25
27
29
34
36
37
IV
1.6
1.7
1.8
Wavefront aberrations of the human eye
1.6.1 Aberroscopy
1.6.2 Measurements of monochromatic
wavefront aberrations
1.6.3 Ocular wavefront aberrations and
accommodation
1.6.4 Cornea1 and total wavefront aberrations
of the eye
1.6.5 Wavefront aberrations and myopia
Myopia and refractive error development
1.7.1 Myopia prevalence
1.7.2 Myopia etiology
1.7.3 Emmetropisation
1.7.4 Near work related myopia studies
in humans
Attempts to prevent myopia progression 1.7.5
Rationale
Chapter 2: Ocular Microfluctuations and
Video kera toscopy
2.1 Introduction
2.2 Methods
2.2.1 Regression plane - to remove tilts
2.2.2 Sphere apex - to remove x,y,z, shifts
2.2.3 Best-fit sphero-cylinder - to remove
cyclo-deviations
2.2.4 Limitations
2.2.5 Protocol
2.3 Results
2.4 Discussion
40
40
41
44
44
46
47
47
49
50
52
54
56
59
59
62
63
64
66
67
69
70
75
V
Chapter 3:
3.1
3.2
3.3
3.4
Chapter 4:
4.1
4.2
4.3
4.4
4.5
Chapter 5:
5.1
5.2
5.4
5.5
Chapter 6:
6.1
6.2
6.3
6.4
6.5
Corneal Topography and
Accommodation
Introduction
Methods
Results
Discussion
Corneal Aberrations and Reading
Introduction
Methods
Analysis
Results
Discussion
Corneal aberrations following reading
in progressing myopes
Introduction
Methods
Results
Discussion
Conclusions
Ocular micro fluctuations
Accommodation and topography
Corneal aberrations, reading and myopia
Future directions
Summary
77
77
79
85
93
97
97
98
100
104
114
120
120
122
128
146
153
153
154
155
160
165
VI
References: 166
Appendix 1: Publications from thesis A
VII
STATEMENT OF ORIGINAL AUTHORSHIP
“The work contained in this thesis has not been previously submitted for a degree or
diploma at any other higher education institution. To the best of my knowledge and
belief, the thesis contains no material previously published or written by another
Figure 3-5: Example of refractive power map output for subject 8, for a 0 D accommodation stimulus (left) and a 7 D accommodation stimulus (right).
Differences in corneal astigmatic axis are shown along with the displacement of blood vessels (indicated by arrows; white for blood vessels and grey for
astigmatic axis) relative to the videokeratoscope coordinate system. The magnitudes of both are similar and approximately 4°.
0 D 7 D
Chapter 3
3.4 DISCUSSION
We did not find clear evidence of statistically significant changes in corneal
topography as a result of accommodation in the subjects examined in this study. This
was the case for both the subjects with normal corneas and those with keratoconic
corneas. These results are consistent with previous studies on central regions of the
cornea (Fairmaid 1959; Lopping and Weale 1965; Mandell and St Helen 1968).
We did however find a significant cyclotorsion of corneal topography of many of the
eyes when changing focus from far to near distances and this may explain the
apparently contradictory findings in the literature in this field.
During fixation the amount of torsion is fixed for each eye position (Listing’s law)
(Fetter and Haslwanter 1999). In our study we induced convergence in the presence
of a deliberate head turn and therefore our results may not be directly comparable to
the findings of Fairmaid, or Lopping and Weale on corneal changes associated with
natural convergence (Fairmaid 1959; Lopping and Weale 1965). Fairmaid reported an
increase of curvature in the horizontal meridian and a decrease of curvature in the
vertical meridian while Lopping and Weale found flattening in the horizontal
meridian associated with convergence.
As in our experiment, Pierscionek et al. induced accommodation in the measured eye,
while the other eye was covered (Pierscionek et al. 2001). This produces
asymmetrical convergence (i.e. a different head position relative to normal
convergence). Pierscionek et al. found differences in central corneal curvature in at
least one principal meridian in most of the subjects tested (Pierscionek et al. 2001).
However the results of the study, which also measured accommodation stimuli of up
93
to 9 D, may have been affected by the cyclotorsional variations, which have been
reported by other authors (Allen and Carter 1967; Bannon 1971; Enright 1980) and
which also have been found in our study. Allen and Carter have shown that even with
an experimental procedure that produces asymmetrical convergence, excyclotorsion
can arise in the non-verging eye (Allen and Carter 1967).
The technique, which was used to confirm that cyclotorsional changes had caused the
apparent changes in the topography difference maps was initially developed in order
to remove ocular micromovements during videokeratoscopy (Chapter 2). By applying
the method across all accommodation conditions not only cyclotorsion was corrected
but also lateral shifts and tilts between corneal topography measurements were
minimized. It was previously shown that cyclotorsional fluctuation is a substantial
contributor to micromovements during steady fixation (Chapter 2). The order of
magnitude (2.4" f 1.1) is similar to what was found for the different accommodation
conditions in this study.
In some of my subjects, these ocular microfluctuations may have masked
cyclotorsional changes related to the different accommodation conditions. However,
within the group it still was possible to show statistically significant cyclotorsional
differences between the 0 D demand condition and the 4 D or 9 D demand conditions.
Furthermore, the comparison between the location of limbal blood vessels and the
changes in corneal astigmatic axis also supported the role of ocular cyclotorsion in
causing apparent changes in corneal shape with accommodation.
94
Chapter 3
Along with others (Allen and Carter 1967; Bannon 1971) it was found that
excyclotorsion occurred with accommodation for most of the subjects. When it was
not accounted for cyclotorsional effects, the keratoconic corneas showed the greatest
changes with accommodation due to higher levels of corneal asymmetry. In the case
of a cyclotorsion occurring between far and near viewing conditions, this would
clearly lead to larger differences occurring in keratoconic corneas rather than more
rotationally symmetric “normal” corneas. The two examples shown in
Figure 3-3 and Figure 3-4 confirm this effect. Both eyes show a significant amount of
cyclotorsion between the different accommodation conditions (Table 3-1). The
excyclotorsion in case of the keratoconic cornea of subject 8 (showing 4 D of corneal
astigmatism) caused significant height changes in the difference map. However, the
excyclotorsion in case of the normal cornea of subject 1 (showing only 0.5 D of
corneal astigmatism) did not lead to such significant changes in the difference map.
The rotation of corneal topography during accommodation has significant
implications for the accuracy of results obtained from optometers or wavefront
sensors that measure the optical characteristics of eyes during near versus far viewing
conditions and pre- versus post operative refractive surgery. As a result of
accommodation, small changes in ocular astigmatism may occur and also asymmetric
higher order aberrations such as coma will be affected. For example, ocular
cyclotorsion while fixating a near target during refractive surgery procedures may
contribute to residual astigmatism outcomes. Furthermore, analyses such as
videokeratoscope difference maps and wavefront analyses of pre- versus post-
operative refractive surgery will be affected if the change in the eye’s refractive status
leads to a change in accommodative status during the image capture procedure.
95
In summary it appears unlikely that changes occur in central corneal shape during
accommodation up to a level of 9 D. Therefore, the changes in corneal topography
that have been reported to occur in some individuals following reading can be
ascribed to lid force effects as suggested by most of these authors (Mandell 1966;
Knoll 1975; Bowman et al. 1978; Carney et al. 1981; Goss and Criswell 1992;
Kommerell 1993; Ford et al. 1997; Campbell 1998; Golnik and Eggenberger 200 1)
rather than due to forces caused by accommodation. However, we found a significant
excyclotorsion of the eye globe during accommodation, which in turn caused a
rotation of corneal topography relative to the instrument’s measurement axis. While
the amount of this cyclotorsion is small, it should be taken into account when
considering the optical characteristics of the eye for different levels of
accommodation.
96
Chapter 4
97
CHAPTER 4
Corneal Aberrations and Reading
4.1 INTRODUCTION
The anterior surface of the eye is its most powerful refractive component and as such,
subtle changes in corneal shape can cause substantial changes in its optical
characteristics. Monocular diplopia has been linked to corneal distortion following
near work in various studies dating back over 35 years (Mandell 1966; Knoll 1975;
Bowman et al. 1978; Carney et al. 1981; Goss and Criswell 1992; Kommerell 1993;
Ford et al. 1997; Campbell 1998; Golnik and Eggenberger 2001). The corneal
distortions that have been observed in these studies have been explained by sustained
(Mandell 1966; Knoll 1975; Bowman et al. 1978; Carney et al. 1981) or abnormal lid
pressure (Kommerell 1993), lid position (Goss and Criswell 1992; Ford et al. 1997),
and tear film interactions with the corneal surface during sustained close work (Ford
et al. 1997; Golnik and Eggenberger 2001). During routine topography measurements,
lid pressure effects on corneal topography have been observed in the central and
peripheral corneal shape (Lieberman and Grierson 2000; Buehren et al. 2001).
There has been conjecture that lid pressure may cause corneal astigmatism (Wilson et
al. 1982; Grey and Yap 1986). Vihlen and Wilson found no significant association
between the tension of the eyelids and the amount of corneal toricity (Vihlen and
Wilson 1983). Grey and Yap measured ocular astigmatism in subjects with three
deliberately narrowed lid positions and found a statistically significant increase of
Chapter 4 _
98
ocular with-the-rule astigmatism when the lid aperture was narrowed (Grey and Yap
1986). There is also evidence that astigmatism can be induced by chalazion (Nisted
and Hofstetter 1974; Cosar et al. 2001), lid-loading procedures used in the treatment
of lagophthalmos (Kartush et al. 1990; Brown et al. 1999; Goldhahn et al. 1999), and
after ptosis surgery (Holck et al. 1998).
While studies on monocular diplopia after near work provide useful information about
the subjectively perceived optical effects of corneal distortion, there hasn’t been a
detailed objective analysis regarding the optical characteristics of corneal distortions
following near work. In this study we have investigated the effect of one hour of
reading on corneal topography. The position of eyelids during reading, relative to the
location and size of the pupil, was measured and compared with corneal topography
changes. We have analysed the optical consequences of the changes that were
observed using traditional sphero-cylinder and corneal higher order aberrations.
4.2 METHODS
Twenty subjects (age range 20 to 37 years, mean 27 years) with healthy eyes were
recruited for the experiment, and one eye was randomly chosen for analysis. Informed
consent was obtained for all subjects. The twenty eyes had a range of refractive errors
(mean –1.6 D, range +0.25 to –6.00 D); five were emmetropic (i.e. ≤ 0.5 D in the
worst meridian), five were primarily astigmatic (i.e. ≤ 0.5 D of spherical component
and > 0.5 D of astigmatic component), five had simple myopia (i.e. ≤ 0.5 D of
astigmatism and > 0.5 D of myopic component), and five had myopic astigmatism
(i.e. > 0.5 D of myopic and astigmatic component).
Chapter 4
99
The experiment was always conducted early in the morning, approximately 2-3 hours
after the subjects woke. All subjects were given the instruction not to perform any
sustained reading (e.g. newspaper) prior to the experiment. For each subject, six
baseline videokeratographs were taken prior to reading and six videokeratographs
were again taken immediately after the 60 minutes reading task.
The Keratron videokeratoscope (EyeQuip Division, Alliance Medical Marketing,
Jacksonville, FL) was used for all corneal topography measurements. The Keratron
has been shown to have high accuracy and precision performance for inanimate test
objects (Tripoli et al. 1995; Tang et al. 2000). Prior to the study, the instrument
calibration was checked according to the manufacturer’s instructions. The
videokeratoscope is based on the placido-disk principle and enables the capture of six
consecutive videokeratographs without the requirement for immediate data
processing.
The subjects were seated and asked to read from a novel for 60 minutes. During this
time the subjects were allowed to adopt whatever head posture was comfortable.
Photopic lighting conditions were used during the reading task. The subjects’ pupil
size was measured in the photopic room condition with the subjects focused at a
distance of 33 cm.
Prior to the reading experiment, digital photography was used to document the
external ocular features using a high-resolution digital camera (Kodak DC260).
Photography was conducted with: (1) the eyes in reading gaze posture, and (2) the
subject positioned in a headrest with eyes in primary gaze. A ruler with millimetre
Chapter 4 _
100
increments was placed in the peripheral field of the captured images to allow
calibration of subsequent measurements.
To determine the approximate position of the eyelids during reading in relation to the
corneal topography, we identified iris features in the digital photographs of the eye
taken during both reading and in the headrest. We then assumed that the relative
position of the eye during videokeratoscopy was the same as that photographed when
the subject was in the headrest. This allowed us to superimpose the approximate
position of the eyelids during reading onto the corneal topography measurements
taken after reading (Figure 4-1). In this way, we could investigate the potential
association between lid position during reading and changes in topography.
After 30 minutes of reading, the subjects’ blink frequency was measured over a time
period of 3 minutes and mean blink rate per minute was later calculated. The subjects
were not informed that blink frequency was being monitored, since this may cause a
change in blink characteristics (Zaman and Doughty 1997; Cho et al. 2000).
4.3 ANALYSIS
Corneal instantaneous power, height data, and refractive power were exported from
the videokeratoscope for analysis. The instantaneous power maps were chosen to
compare corneal topography with the subject’s natural lid position during reading,
because the instantaneous power maps are most sensitive to local power changes
caused by slight variations in slope.
Chapter 4
101
To study the potential effect of reading on topography, height difference maps were
calculated. For each set of six baseline measurements and six post-reading
measurements, the effect of ocular microfluctuations were minimized using the
method described in Chapter 2. The methodology repositions a given
videokeratograph map to best approximate an “average” videokeratograph based on a
set of multiple measurements of the same cornea. This procedure involves
interpolating (bilinear) the topography data to a common grid format (256 meridians
and point spacing along the meridian of ~0.15 mm) and subsequent calculation of an
average height map for each set of maps (i.e. before and after reading).
From the six refractive power maps for each condition (i.e. pre- and post-reading) we
calculated the average, standard deviation, and the number of valid measurements at
each grid location within the map. Difference maps of pre- versus post-reading
topography were calculated along with the t-test maps showing significant areas of
change (Buehren et al. 2001). The root mean square error (RMSE) between corneal
refractive power and best-fit sphero-cylinder before and after reading was calculated
and t-tests were applied to measure the significance of changes in refractive power
between the averages of the two conditions.
This was performed for each individual’s photopic pupil size and also for fixed pupil
sizes of 2.5, 3, 4, 5, and 6 mm. For the 4 mm fixed pupil size, power matrices (Harris
2000) were used to average individual best-fit sphero-cylinders for each condition and
again to calculate the corneal changes in sphere, cylinder and astigmatic axis pre
versus post-reading. A multivariate test (Hotelling’s T2) representing a generalisation
Chapter 4 _
102
of the t statistic was used to test the significance of overall change in corneal sphero-
cylinder.
The anterior surface of the cornea was modelled as a single surface optical system in
order to derive the corneal wavefront error using a method similar to that described by
Guirao and Artal (Guirao and Artal 2000). Optical path distance (OPD) for each point
on the surface was calculated using 3-D ray tracing and the wavefront was fitted using
a set Zernike terms of up to the fourth order polynomial expansion according to the
Optical Society of America convention (OSA convention) (Thibos 2000) for pupil
sizes of 2.5 mm, 4 mm, and 6 mm (image plane at circle of least confusion, wavefront
error scaled by λ = 555 nm, refractive index n = 1.376, and midline symmetry taken
into account (Smolek et al. 2002). All wavefront coefficients were normalized to a
unit circle to enable quantitative comparison between different pupil sizes. The
wavefront was centred on the line of sight. To achieve this we calculated the average
pupil offset derived from the pupil detection system provided by the Keratron
videokeratoscope for each subject and used this offset as the new principle axis
reference point. A full 3-D ray-trace technique was applied to calculate the wavefront
error along the line of sight. T-tests were used to identify statistically significant
changes in corneal wavefront Zernike coefficients after reading.
Chapter 4
103
Figure 4-1: Method for overlaying eyelid features onto corneal topography maps. Iris features (indicated by arrows) are used in each of the images to record relative
lid position. Subject 13 (top left), lid position in primary gaze, lid position during reading (bottom left), baseline corneal topography overlaid with lid position in
primary gaze (top right), and post-reading corneal topography overlaid with lid position during reading (bottom right).
Chapter 4 _
104
4.4 RESULTS
In twelve of the twenty subjects we studied, the instantaneous power maps after
reading showed distinct band-like distortions in the superior region of the maps,
which correlated closely with the subject’s natural lid position during reading. In
Figure 4-2, two examples of videokeratograph comparisons before and after reading
for subject 1 and 15 are shown, with the overlay of the subjects’ lid position during
reading. These topography changes were often encroaching within the boundary of
the subjects’ upper pupil margin. Topography changes were also often evident in the
inferior cornea associated with the position of the lower lid margin during reading
(see subject 1, Figure 4-2). However these inferior distorted regions generally did not
encroach within the pupil zone.
Analysis of the corneal wavefronts revealed that seven wavefront coefficients were
significantly changed after one hour of reading (Figure 4-3). The terms that changed
significantly for all pupil sizes were 22Z primary astigmatism (p < 0.05 at 2.5 mm and
p < 0.01 at 4 mm and 6 mm), 13−Z primary vertical coma (p < 0.01 at 2.5 mm and
p < 0.001 at 4 mm and 6 mm), and 33−Z trefoil 30° (p < 0.05 at 6 mm and p < 0.01 at
2.5 mm and 4 mm). The change of the primary astigmatism terms was in the direction
of against-the-rule (i.e. with-the-rule astigmatism decreased or against-the-rule
astigmatism increased).
Chapter 4
105
Figure 4-2: Examples of the effect of lid position during reading upon corneal topography. Subject 1 (left) and subject 15 (right), corneal topography pre-reading
(top) versus pos-reading (bottom) is shown overlaid with the subjects’ lid position in primary gaze and during reading respectively.
Chapter 4 _
106
Other changes in Zernike terms were limited to certain pupil sizes (i.e. corneal
regions). For the 2.5 mm pupil size, the term 13Z primary horizontal coma (p < 0.05)
changed significantly. The defocus term 02Z changed for the 4 mm and 6 mm pupil
sizes (p < 0.01 at 4 mm and 6 mm). The vertical prism term 11−Z changed for the 6
mm pupil size (p < 0.05) and the secondary astigmatism component 24Z changed for
the 4 mm pupil size (p < 0.05).
There was a significant association between changes occurring in the wavefront for
the 13−Z vertical coma and the 3
3−Z trefoil 30° terms after reading. Most subjects (15
of 20) showed positive vertical coma and negative trefoil 30°, or negative vertical
coma and positive trefoil 30° in the baseline (pre-reading) measurements and this
trend was increased after reading (18 of 20). Both of these combinations of coma and
trefoil terms represent a wave-like shape (Figure 4-4 bottom) but they are opposite in
direction. After reading, there was a trend for the vertical coma term to shift in the
negative direction, whereas the trefoil 30° term generally shifted in the positive
direction. The changes in the vertical coma and trefoil 30° coefficients after reading
are shown in Figure 4-4 (top). These wave-like shape changes in the wavefront are
consistent with the changes in instantaneous power maps associated with the effect of
the upper lid margin.
Chapter 4
107
* * ** * * * * *
* * * * *
-0.15
-0.12
-0.09
-0.06
-0.03
0
0.03
0.06
0.09
Zernike Coefficients (OSA convention)
Cha
nge
in W
avef
ront
Err
or C
oeff
icie
nt V
alue
(
)
2.5mm P upil4.0mm P upil6.0mm P upil
1- 1 Z
1 1 Z
-22Z
02Z
22Z
-33Z
-13Z
1 3 Z
3 3 Z
-44Z
-24Z
0 4 Z
24Z
44Z
* = t-test significant at (p < 0.05)
Figure 4-3: The group mean change (±SE) of normalized Zernike wavefront coefficients after reading (post minus pre) for three pupil sizes (2.5 mm, 4 mm, 6 mm) is
shown. The Zernike polynomials are vertical prism 1
1−Z , horizontal prism
11Z , primary astigmatism along 45°
22−Z , defocus
02Z , primary astigmatism
22Z ,
trefoil along 30° 3
3−Z , primary vertical coma
13−Z , primary horizontal coma
13Z , trefoil
33Z , tetrafoil along 22.5°
44−Z , secondary astigmatism along 45°
24−Z ,
spherical aberration 04Z , secondary astigmatism
24Z , and tetrafoil
44Z .
Chapter 4 _
108
Correlation between Change in vertical Coma and Trefoil along 30° after reading
y = -0.8739x - 0.0056R = 0.8121
-0.06
-0.04
-0.02
0
0.02
0.04
0.06
0.08
0.1
-0.1 -0.08 -0.06 -0.04 -0.02 0 0.02 0.04
Change in primary vertical ComaWavefront Coefficient Value (λ)
Cha
ng
e in
Tre
foil
alo
ng
30°
W
ave
fro
nt
Co
eff
icie
nt
Val
ue
()
Negative Coma Positive Trefoil Negative Coma + Positive Trefoil
Figure 4-4: Correlation between change in primary vertical coma and trefoil along 30° (4 mm pupil)
following reading. Bottom panel shows the combination of negative vertical coma and positive trefoil
30° coefficients producing a “wave-like” distortion.
Chapter 4
109
An example of the optical changes in corneal refractive power following reading is
presented for subject 8 in Figure 4-5. In the difference map (pre-versus post-reading
refractive power), the superior semi-meridian shows values of up to -1.34 D change.
These changes are highly statistically significant as shown by the p-values of the t-test
map (Figure 4-5). Within a 4-mm pupil, thirteen of the twenty subjects showed
statistically significant (p < 0.001) areas of change in refractive power (Table 4-1, last
column). These significant regions of change were mostly located in the upper, and/or
the lower pupil areas. Some subjects showed small randomly distributed points of
statistically significant change, however these areas were considered to be non-
systematic and probably related to local tear instabilities rather than true changes in
corneal topography and were not classified as representing statistically significant
change (in Table 4-1, last column).
The group mean RMSE deviation from the best-fit sphero-cylinder was slightly larger
for the post-reading corneas (Table 4-1). This difference was statistically significant
when calculated for the individuals’ photopic pupil size (pre 0.23 D versus post 0.28 D,
p = 0.013) as well as for fixed pupils of 5 mm and 6 mm (pre 0.31 D versus post 0.35
D, p = 0.036 and pre 0.38 D versus post 0.42 D, p = 0.022 respectively). The increased
RMSE was not statistically significant for the 2.5, 3 and 4 mm pupils (p = 0.66, p =
0.28, and p = 0.09 respectively).
In Table 4-1 individual subject data for various refractive power changes following
reading are summarized. This includes the total refractive error, corneal best-fit sphero-
cylinder power pre and post-reading and the change in corneal sphero-cylinder for the
twenty subjects.
Chapter 4 _
110
Across the group, lid fissure width decreased from a mean of 9.4 mm (SD ± 0.9) in
primary gaze to 6.8 mm (SD ± 1.0) in the reading position. Average pupil size of the
subject group during reading in the photopic condition was 3.3 mm (SD ± 0.7). Blink-
frequency showed large variability between individuals, with an average value of 8.3
blinks/minute (range 2 to 26 blinks per minute). There was no significant correlation
between blink rate and corneal RMSE differences (R2 = 0.10).
The changes in corneal aberrations that we measured following reading were clearly
associated with forces applied by the eyelid margin to the surface of the eye. However
the role of eye movements during reading was unknown. To investigate this we
recruited one subject (subject 9) who showed obvious corneal topography changes
following the reading trial. The subject was retested on a separate morning, but this
time the subject had to stare (fixed gaze) at a single word on a page for 60 minutes in
normal reading gaze. The changes in refractive power following the fixed gaze trial
along with the results from the initial reading trial of the same subject are shown in
Figure 4-6. The corneal changes after 60 minutes of staring are less pronounced, with
values approximately half those found after 60 minutes of reading. The locations of
changes shown after both trials indicate that a similar lid position was adopted during
the two experiments. Although it is not clear whether other factors may also have
influenced these two trials, it may be speculated that eye movements during reading
contribute to the forces applied to the cornea.
Chapter 4
111
Corneal Refractive Power Analysis (Subject 8)
Figure 4-5: Average baseline refractive power map (top left) and average post-reading refractive
power map (top right) for a 6-mm pupil zone of subject 8. Bottom left: Refractive power difference
map of post-reading minus baseline refractive power for a 4 mm pupil zone. Bottom right:
Significance map based on t-tests at all points within the map. The regions where p values < 0.001 are
in black, 0.001< p < 0.05 are in grey and p values > 0.05 are in white.
60 Min Read
RMSE = 0.30 D RMSE = 0.42 D
Baseline
Difference in 4 mm Pupil P-values
Chapter 4 _
112
Table 4-1: The changes in corneal refractive power (sph, cyl, and axis) are presented as refractive error, not correction. The RMSE is the difference between the
corneal refractive power and the best-fit corneal sphero-cylinder. The topographic refractive power changes illustrate those subjects where statistically significant
areas of change occurred in refractive power within the central 4 mm of the cornea following reading (see Figure 5 example). Because of potential type 1 errors
associated with repeated statistical testing, we have chosen to highlight only corneal changes where p < 0.001 (i.e. * = p < 0.05; ** = p< 0.01; *** = p < 0.001).
After two hours of reading (Figure 5-3 top) the mean corneal map of the myopic group
shows a distinct band like distortion at less than 2 mm from the pupil centre. The mean
corneal map of the emmetropes also shows more asymmetry due to distortions in the
upper cornea at about 2.5 mm from the pupil centre. The changes are highlighted in the
difference maps showing a similar but larger and more complex distortion pattern in the
myopic group. The distortions in the myopes are more closely located to the pupil centre
and show higher statistical significance (bottom right) than in case of the emmetropic
group (bottom left).
The change of corneal wavefront RMS error after two hours of reading is shown in
Figure 5-4. What becomes evident from this analysis is that a large proportion of the
increased total wavefront RMS error change in the myopic group comes from the lower
order defocus and primary astigmatism components. In comparison, the emmetropic
group does not show such a distinct difference between total RMS and higher order
RMS changes. For a 5 mm pupil, the amount of the total higher order RMS changes in
both groups are very similar. When the RMS error is analysed in terms of the Zernike
higher order components, a more complex distortion pattern within the myopic group is
indicated by larger 4th order RMS changes and significantly larger 5th order RMS
changes (p = 0.008) in the myopes, while the 3rd order components are larger in the
emmetropes. For the smaller pupil regions of 4 mm and 3 mm, the magnitude of change
in the Zernike 4th, 5th, and 6th order terms became less significant in myopes and
emmetropes. As a consequence, within a 3 mm pupil zone the myopes show increased
higher order RMS changes. In the 3 mm pupil these changes are primarily due to the 3rd
order components in the myopes resulting in a statistically significant increase of the
total wavefront RMS error (p = 0.029) in this pupil size.
Chapter 5 _
133
Analysis of the corneal wavefronts revealed significantly changed wavefront
coefficients for both groups after one and two hours of reading. For a 5 mm pupil the
corneal wavefront coefficients before and after 2 hours of reading for the emmetropic
group and the myopic group are shown in Figure 5-5 and Figure 5-6 respectively. As in
Chapter 4, for the myopic group the changes in primary vertical coma and trefoil along
30º were similar in magnitude and as such that primary vertical coma shifted in negative
direction while the trefoil component shifted in positive direction. Statistical significant
corneal wavefront coefficients for emmetropes and myopes for all pupil sizes are
summarized in Table 5-2.
Chapter 5
134
-0.08
-0.04
0
0.04
0.08
0.12
0.16
0.2
0.24
0.28
0.32
0.36
0.4
0.44
Cha
nge
in W
avef
ront
RM
S V
alue
(mic
rons
)Myopic GroupEmmetropic Group
Figure 5-4: Group mean change (±SD) in wavefront root mean square error (RMS) after 2 hours of reading for myopes and emmetropes for three different
pupil sizes (3, 4, and 5 mm) is shown. Corneal wavefront RMS is presented as total RMS, higher order RMS, 3rd order, 4th order, 5th order, and 6th order RMS
components along with the relevant significance values for the t-test comparison between the change in value for the myopic versus emmetropic groups.
5 mm Pupil 4 mm Pupil 3 mm Pupil
Total RMS
Higher Order
3rd Order
4th Order
5th Order
6th Order
p-Value 0.079
0.726
0.212
0.076
0.008
0.404
p-Value 0.114
0.229
0.633
0.321
0.221
0.488
p-Value 0.029
0.052
0.079
0.083
0.269
0.173
Total RMS
Higher Order
3rd Order
4th Order
5th Order
6th Order
Total RMS
Higher Order
3rd Order
4th Order
5th Order
6th Order
Corneal Wavefront RMS change after 2 hours of Reading (OSA convention)
*
*
Chapter 5 _
135
Figure 5-5: Group mean wavefront coefficients (±SD) of emmetropes before and after 2 hours of reading for a 5 mm pupil sizes is shown. Zernike terms
according to OSA convention of up to the 6th order polynomial expansion are presented. Piston and the prism coefficients are excluded.
-0.5
-0.45
-0.4
-0.35
-0.3
-0.25
-0.2
-0.15
-0.1
-0.05
0
0.05
0.1
0.15
Zernike Coefficients (OSA convention)
Cha
nge
in W
avef
ront
Err
or C
oeff
icie
nt V
alue
(mic
rons
)BaselinePost 2-h read
* = t-test significant at (p < 0.05)
13−Z 3
3−Z 4
4−Z 1
3Z
33Z
24Z
04Z
24−Z 4
4Z
55−Z 3
5−Z 1
5−Z 1
5Z 35Z
55Z
66−Z 4
6−Z 2
6−Z 0
6Z
26Z
46Z
66Z
22Z
22−Z 0
2Z
* * *
Emmetrope Group: Baseline and post 2-hours reading Zernike wavefront coefficients (5 mm pupil)
Chapter 5
136
Figure 5-6: Group mean wavefront coefficients (±SD) of myopes before and after 2 hours of reading for a 5 mm pupil sizes is shown. Zernike terms
according to OSA convention of up to the 6th order polynomial expansion are presented. Piston and the prism coefficients are excluded.
-0.5
-0.45
-0.4
-0.35
-0.3
-0.25
-0.2
-0.15
-0.1
-0.05
0
0.05
0.1
0.15
Zernike Coefficients (OSA convention)
Cha
nge
in W
avef
ront
Err
or C
oeff
icie
nt V
alue
(mic
rons
)BaselinePost 2-h read
* = t-test significant at (p < 0.05)
13−Z 3
3−Z 4
4−Z 1
3Z
33Z
24Z
04Z
24−Z 4
4Z
55−Z 3
5−Z 1
5−Z 1
5Z 35Z
55Z
66−Z 4
6−Z 2
6−Z 0
6Z
26Z
46Z
66Z
22Z
22−Z 0
2Z
Myopic Group: Baseline and post 2-hours reading Zernike wavefront coefficients (5 mm pupil)
* * * *
* *
Chapter 5 _
137
Summary of significant corneal wavefront changes after 2 hours of reading for various pupil sizes:
Table 5-2: The change of various normalised corneal Zernike wavefront coefficients in microns (OSA convention) along with the relevant p-values for the
emmetropic and myopic group respectively is shown.
Emmetropic Group Myopic Group
Zernike Pupil Pupil Pupil Pupil Pupil Pupil
Microns 3 mm p-value 4 mm p-value 5 mm p-value 3 mm p-value 4 mm p-value 5 mm p-value
The difference in corneal wavefront error coefficient changes after two hours of reading
between the myopic and emmetropic groups revealed a significantly larger change of
the 22Z primary astigmatism term in direction of against-the-rule in the myopic group
(p < 0.05 at 5 mm) (Figure 5-7). The 02Z defocus term shifted toward corneal flattening
(hyperopia) in myopes (see Figure 5-6) while emmetropes tended to show slightly
steeper corneas following reading (see Figure 5-5). This led to a significant 02Z defocus
term difference (p < 0.05 at 3, 4, and 5 mm) between emmetropes and myopes. There is
little statistically significant change within the 3rd order terms, however several 4th and
higher order terms changed significantly different, confirming the more complex
distortion pattern of the instantaneous power maps in the myopes.
The effect on corneal higher order RMS of time spent reading revealed a linear trend
towards more significant changes for longer reading periods and larger pupils (4 mm
and 5 mm pupil) within the emmetropic group (Figure 5-8). After one hour of reading
the higher order wavefront RMS values show larger increases in the myopic group,
reaching statistically significant differences compared with the emmetropic group for
the three pupil sizes analysed. After two hours of reading the mean corneal higher order
wavefront RMS for myopes decreases slightly relative to the 1-hour reading trial while
emmetropes continued to increase. However, for the 4 mm and 5 mm pupils the
difference still was significantly larger in myopes. The mean corneal higher order
wavefront RMS in microns for a 3, 4, and 5 mm pupil at the baseline condition was not
significantly higher in the myopic group compared with the emmetropic group
(emmetropes: 0.07 µm, 0.11 µm, and 0.19 µm ; myopes: 0.08 µm, 0.16 µm, and 0.26
µm for a 3, 4, and 5 mm pupil respectively).
Chapter 5 _
139
-0.15
-0.125
-0.1
-0.075
-0.05
-0.025
0
0.025
0.05
0.075
0.1
Zernike Coefficients (OSA convention)
Cha
nge
in W
avef
ront
Err
or C
oeff
icie
nt V
alue
(mic
rons
) 3mm Pupil4mm Pupil5mm Pupil
Figure 5-7: Group mean difference in wavefront coefficient change (±SD) between myopes and emmetropes after 2 hours of reading for three different pupil
sizes (3, 4, and 5 mm) is shown. Zernike terms according to OSA convention of up to the 6th order polynomial expansion are presented. Piston and the prism
coefficients are excluded.
* = t-test significant at (p < 0.05)
13−Z 3
3−Z 4
4−Z 1
3Z
33Z
24Z
04Z
24−Z 4
4Z
55−Z 3
5−Z 1
5−Z 1
5Z 35Z
55Z
66−Z 4
6−Z 2
6−Z 0
6Z
26Z
46Z
66Z
22Z
22−Z 0
2Z
* * * * *
* *
* *
* * *
Corneal wavefront error change after 2 hours of reading (Difference for Myopic minus Emmetropic Group)
Chapter 5
140
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
0.45
Hig
her
orde
r wav
efro
nt R
MS
(mic
rons
)
Emmetrope
Myope
Figure 5-8: Higher order corneal wavefront RMS (±SD) before and after 1-hour and 2 hours of reading for myopic group and emmetropic group. The increase for various pupil sizes is shown.
pre-read post 1-h post 2-h
pre-read post 1-h post 2-h
pre-read post 1-h post 2-h
3 mm pupil 4 mm pupil 5 mm pupil
*
*
* *
*
* = t-test significant at (p < 0.05)
Chapter 5 _
141
The group mean characteristic shape difference between emmetropes and myopes of the
total corneal wavefront in a 5 mm pupil zone after 2 hours of reading, show increased
differences in the upper region of the wavefront (Figure 5-9 top left). The higher order
Zernike wavefront (top right) as well as 4th and 5th order components (centre right and
bottom left) show the characteristic band like distortion pattern following reading with
the 4th order component (centre right) making up for most (RMS = 0.0595 µm) of the
differences in higher order wavefront error change between the myopes and the
emmetropes.
In Table 5-3 the group mean corneal sphero-cylinder and corneal best sphere lens
(sphero-cyl and BSL measured at pre-reading circle of least confusion) before and after
1 hour and 2 hours of reading are shown along with the differences, and the statistical
significance of the changes (Hoteling T2). For the myopes, within the 3 and 4 mm pupils
the group mean sphere and cylinder after 2 hours of reading show the largest changes
with sph +0.22 (±0.35) and cyl –0.21 (±0.32) for a 3 mm pupil, while for a 4 mm pupil a
change of sph +0.22 (±0.27) and cyl –0.23 (±0.23) is shown. However only within the 3
mm pupil, the sphero-cylinder change at 1 hour after reading in the myopic versus the
emmetropic group is significantly different (Hoteling T2, p = 0.049). The changes in
corneal wavefront RMS are shown in Table 5-2 along with the t-test results for each
group. For both myopic and emmetropic groups there is a trend towards more
significant changes for larger pupils and longer reading periods.
Chapter 5
142
Wavefront Error Difference (Myopic minus Emmetropic Group)
All Coefficients Higher Order
3rd order 4th order
5th order 6th order
Figure 5-9: Group mean difference in wavefront root mean square error (RMS) between myopes
and emmetropes (myopes minus emmetropes) is shown. Corneal wavefront RMS difference is
presented as total RMS (top left) along with the sphero cylinder change, higher order RMS (top
right), 3rd order (centre left), 4th order (centre right), 5th order (bottom left), and 6th order RMS
components (bottom left).
Chapter 5 _
143
Emmetropic and myopic group data of sphero-cylinder changes
Emmetropes (n = 10) Myopes (n = 10)
Pupil 3mm Pupil
3mm
Baseline Sph Cyl A BSL HO RMS Baseline Sph Cyl A BSL HO
RMS Mean
SD 0.32
(±0.19) -0.56
(±0.39) 27 0.040 (±0.07)
0.067 (±0.013)
Mean SD
0.59 (±0.38)
-1.03 (±0.75) 3 0.070
(±0.047) 0.081
(±0.03)
Post1-h read Sph Cyl A BSL HO RMS Post1-h read Sph Cyl A BSL HO
RMS Mean
SD 0.31
(±0.18) -0.63
(±0.39) 29 0.000 (±0.08)
0.071 (±0.012)
Mean SD
0.61 (±0.38)
-0.92 (±0.70) 6 0.148
(±0.184) 0.104
(±0.04) Difference
SD 0.00
(±0.07) -0.09
(±0.07) 46 -0.040 0.004 Difference SD
0.15 (±0.24)
-0.14 (±0.18) 78 +0.079 0.023
T-Test T-Test T-Test T-Test P=0.095 p=0.229
P=0.165 p=0.043
Difference in Sphero-Cylinder change between Emmetropic and Myopic Groups Hoteling T2: p = 0.049
Post2-h read Sph Cyl A BSL HO RMS Post2-h read Sph Cyl A BSL HO
RMS Mean
SD 0.25
(±0.24) -0.57
(±0.38) 28 -0.033 (±0.129)
0.069 (±0.012)
Mean SD
0.62 (±0.34)
-0.86 (±0.71) 7 0.187
(±0.232) 0.099
(±0.05) Difference
SD -0.06
(±0.15) -0.04
(±0.05) 61 -0.07 0.005 Difference SD
0.22 (±0.35)
-0.21 (±0.32) 76 +0.117 0.018
T-Test T-Test T-Test T-Test P=0.126 p=0.365
P=0.115 p=0.356
Difference in Sphero-Cylinder change between Emmetropic and Myopic Groups Hoteling T2: p = 0.165 Pupil 4mm Pupil
4mm
Baseline Sph Cyl A BSL HO RMS Baseline Sph Cyl A BSL HO
RMS Mean
SD 0.36
(±0.19) -0.59
(±0.42) 26 0.068 (±0.051)
0.107 (±0.028)
Mean SD
0.58 (±0.40)
-1.01 (±0.76) 1 0.076
(±0.026) 0.158
(±0.08)
Post1-h read Sph Cyl A BSL HO RMS Post1-h read Sph Cyl A BSL HO
RMS Mean
SD 0.34
(±0.16) -0.63
(±0.38) 27 0.029 (±0.06)
0.120 (±0.028)
Mean SD
0.58 (±0.38)
-0.85 (±0.78) 4 0.15
(±0.14) 0.191
(±0.094) Difference
SD -0.01
(±0.07) -0.06
(±0.06) 46 -0.039 0.012 Difference SD
0.16 (±0.22)
-0.18 (±0.20) 78 +0.076 0.034
T-Test T-Test T-Test T-Test P=0.07 P=0.10
P=0.120 p=0.101
Difference in Sphero-Cylinder change between Emmetropic and Myopic Groups Hoteling T2: p = 0.110
Post2-h read Sph Cyl A BSL HO RMS Post2-h read Sph Cyl A BSL HO
RMS Mean
SD 0.28
(±0.24) -0.62
(±0.41) 27 -0.029 (±0.143)
0.128 (±0.028)
Mean SD
0.59 (±0.37)
-0.81 (±0.78) 5 0.1833
(±0.168) 0.177
(±0.065) Difference
SD -0.08
(±0.17) -0.04
(±0.05) 54 -0.097 0.021 Difference SD
0.22 (±0.27)
-0.23 (±0.23) 77 +0.108 0.019
T-Test T-Test T-Test T-Test P=0.086 p=0.067
p=0.074 p=0.46
Difference in Sphero-Cylinder change between Emmetropic and Myopic Groups Hoteling T2: p = 0.088 Pupil 5mm Pupil
5mm
Baseline Sph Cyl A BSL HO RMS Baseline Sph Cyl A BSL HO
RMS Mean
SD 0.42
(±0.21) -0.61
(±0.43) 25 0.112 (±0.048)
0.186 (±0.05)
Mean SD
0.62 (±0.41)
-1.02 (±0.76) 1 0.112
(±0.029) 0.260
(±0.11)
Post1-h read Sph Cyl A BSL HO RMS Post1-h read Sph Cyl A BSL HO
RMS Mean
SD 0.39
(±0.20) -0.63
(±0.41) 26 0.07 (±0.06)
0.202 (±0.04)
Mean SD
0.60 (±0.39)
-0.91 (±0.78) 2 0.15
(±0.08) 0.314
(±0.12) Difference
SD -0.03
(±0.08) -0.03
(±0.06) 44 -0.04 0.017 Difference SD
0.09 (±0.13)
-0.18 (±0.10) 79 +0.035 0.054
T-Test T-Test T-Test T-Test p=0.076 p=0.112
p=0.246 p=0.005
Difference in Sphero-Cylinder change between Emmetropic and Myopic Groups Hoteling T2: p = 0.167
Post2-h read Sph Cyl A BSL HO RMS Post2-h read Sph Cyl A BSL HO
RMS Mean
SD 0.33
(±0.26) -0.61
(±0.40) 27 0.02 (±0.14)
0.219 (±0.06)
Mean SD
0.615 (±0.39)
-0.86 (±0.74) 3 0.19
(±0.11) 0.30
(±0.10) Difference
SD -0.06
(±0.17) -0.05
(±0.06) 71 -0.09 0.06 Difference SD
0.16 (±0.16)
-0.18 (±0.10) 80 +0.073 0.041
T-Test T-Test T-Test T-Test p=0.106 p=0.048
p=0.073 p=0.049
Difference in Sphero-Cylinder change between Emmetropic and Myopic Groups Hoteling T2: p = 0.054
Table 5-3: The best sphere lens (BSL) change is calculated from the difference between the
baseline corneal wavefront circle of least confusion and post reading circles of least confusion at
various pupil sizes. HO RMS = Higher order wavefront root mean square error.
Chapter 5
144
Analysis of digital image processing measurement revealed a significantly smaller
palpebral aperture for the myopic group in the reading gaze position (emmetropic group
= 8.15 mm (±1.91); myopic group = 6.74 mm (±1.21); p = 0.042) but not in primary
gaze position (emmetropic group = 10.72 mm (±1.44); myopic group = 10.21 mm
(±1.61); p = 0.47). Also the distance from the upper lid margin to the pupil centre in the
myopic group was significantly smaller in reading gaze position (emmetropic group =
3.38 mm (±1.50); myopic group = 2.14 mm (±1.0); p = 0.028) but not in primary gaze
position (emmetropic group = 4.52 mm (±1.04); myopic group = 4.08 mm (±0.73); p =
0.29). No statistical significant difference was found for the distance from the lower lid
margin to the pupil centre either in primary gaze position (emmetropic group = 6.20 mm
(±0.82); myopic group = 6.13 mm (±1.00); p = 0.87) or in reading gaze position
(emmetropic group = 4.77 mm (±1.07); myopic group = 4.6 mm (±0.59); p = 0.63). A
typical example of reading gaze ocular biometrics for an emmetropic subject and a
myopic subject (Figure 5-10) is shown.
Average horizontal limbus diameter was 12.36 mm (±0.48) for the emmetropic group
and for the myopic group 11.97 mm (±0.61) (p = 0.13). The average reading distance
for emmetropes was 42.80 cm ranging from 30 to 52 cm, while myopes showed a mean
reading distance of 36.60 cm ranging from 23 to 48 cm (p = 0.087). Near work activity
scores showed 82 (±26) for the emmetropes and 97 (±31) for the myopes (p = 0.24)
respectively.
Chapter 5 _
145
Figure 5-10: Ocular biometrics data in reading gaze position using automatic image
measurement techniques for an emmetropic subject (left) and a myopic subject (right)
respectively. The information extracted from digital photographs is PUPIL d = pupil diameter,
LIMBUS d = limbus diameter, APERTURE UP = distance from limbus centre to upper lid
margin, and APERTURE DOWN = distance from limbus centre to lower lid margin.
PUPIL d = 4.13 mm LIMBUS d = 12.00 mm Aperture UP = 3.38 mm
Aperture DOWN = 4.89 mm
PUPIL d = 3.79 mm LIMBUS d = 10.79 mm Aperture UP = 2.26 mm
Aperture DOWN = 4.69 mm
Chapter 5
146
5.4 DISCUSSION
The major difference between the two groups of emmetropes and myopes found in this
study was the location and magnitude of corneal distortions following reading. This was
most noticeable in the instantaneous power difference maps and supported by the digital
image analysis of lid position during reading. For the myopic group, the location and
magnitude of distortions caused larger central corneal optical changes as compared with
the emmetropes, which was most clearly highlighted by the total RMS error changes.
Almost all subjects showed some changes in the corneal instantaneous power
topography after two hours of reading. The exception was one emmetropic subject in
whom there was virtually no change in topography apparent after two hours of reading.
This emmetropic subject also showed the largest palpebral aperture in the reading gaze
position (12.3 mm) and this provides a likely explanation for the lack of central (>5mm)
corneal topography changes.
Corneal wavefront RMS analysis revealed that distortions closer to the pupil centre not
only have a larger impact on higher order Zernike term components but also
significantly change the lower order astigmatism and defocus components. Analysis of
corneal wavefront error confirmed several of the previously reported results (Chapter 4),
which were found to be associated with the band like distortions related to lid forces.
For the myopic group an astigmatic shift in the direction of against-the-rule, as well as
significant changes of the primary vertical coma and trefoil along 30° terms were found.
While the link between lid force effects on the cornea and the changes of coma and
trefoil components is interesting, it may not always be a precise description for these
Chapter 5 _
147
corneal distortions. As distortions become more complex, a combination of secondary
astigmatism and tetrafoil components in addition to some 5th and 6th order terms may
provide a more accurate description. For larger pupils in particular, several 5th and 6th
order Zernike coefficients changed significantly in both groups indicating that Zernike
terms up to the 4th order may not always be enough to adequately describe the changes
in the corneal surface associated with lid forces.
Due to the 1 hour and 2 hour time periods the participants spent on reading, it was
possible to gain an impression of the effect of time on distortions, which showed a
group mean increase in magnitude of distortions and optical effects with increasing
time. Some subjects did not show increased higher order aberration changes after 2
hours, but rather showed a slight decrease relative to the 1-hour trial. While there was a
group mean increase in lower order terms, the change in higher order components did
not always lead to an increased higher order wavefront RMS. In one myopic subject the
corneal changes after reading even lead to a decrease in wavefront RMS error compared
with the subject’s baseline RMS. The differences between the 1-hour and 2 hour trial
might also be affected by a change in reading strategy (i.e. a change in body or head
posture or a change in book position) after a certain period of reading. To further
investigate this effect, continuous or periodic digital photos of lid and eye position
during the entire reading period would need to be taken. Reading a book for 2 hours
without any breaks is probably uncommon. In our experiment there was a short break of
just a view minutes between the 1-hour and 2 hour reading trials to collect corneal data
after the first hour of reading.
Chapter 5
148
Averaging of individuals’ corneal instantaneous power maps (Figure 5-1 to Figure 5-3)
may have resulted in distortions being masked due to differences in location on the
corneal surface. However the aim of this analysis was to illustrate the group mean trend
for emmetropes and myopes. A more accurate analysis is provided by the corneal
wavefront Zernike coefficients and RMS data.
The most commonly reported subjective visual observation that has been reported
following reading is monocular vertical doubling (Fincham 1963; Mandell 1966; Knoll
1975; Bowman et al. 1978; Carney et al. 1981; Kommerell 1993; Ford et al. 1997;
Campbell 1998; Golnik and Eggenberger 2001). Some studies have estimated the time
course of remission of distortions after reading based on subjective perception of
remission of the vertical doubling (Knoll 1975; Golnik and Eggenberger 2001). When
objectively measuring the time course of remission of the lid induced changes on the
corneal surface, there appears to be a non-linear decrease in the distortion with a large
proportion of the effect disappearing within a short time period after cessation of
reading, followed by a continuous slowing down of the process that may take more than
120 minutes after 60 minutes of reading (Chapter 4). It has also been shown that lid
related distortions in the periphery of the cornea change within seconds during the post-
blink interval (Buehren et al. 2001).
The optical effects during reading with the eyelids in reading gaze posture are likely to
be larger, taking into account short term changes that have been reported within the
post-blink interval (Buehren et al. 2001). In order to allow individual statistical analysis
to be drawn from our data and to improve accuracy, we have collected six corneal
topography measurements for each subject and each of the pre- and post reading
Chapter 5 _
149
conditions. After reading, several subjects’ data showed a small trend towards
decreasing distortions even within the short time frame it took to capture the six images
(approximately 30 sec). Furthermore, the effect of the tear meniscus at the lid margin in
cases where the upper lid position is near the upper pupil edge could also contribute to
changes in optical characteristics during reading. In order to capture the full magnitude
of the optical changes of eyes during reading it would be necessary to measure the eye’s
corneal and total wavefront aberrations after a period of reading with the eye in the
reading gaze position.
The exact mechanism of how the changes on the corneal surface occur is not clear. The
pattern of remission could indicate cell displacement similar to the mechanism that has
been proposed in orthokeratology (Swarbrick et al. 1998). However this assumption
needs to be confirmed by pachometry of the epithelium.
The likelihood of corneal distortions effecting overall visual performance during the day
will be significantly influenced by the time spent on near work activities. The sample of
myopes in this study reported that they spend about 20 % more time on near work
activities than the sample of emmetropes, although the difference was not statistically
significant. The myopes also used smaller reading distances than emmetropes, which is
likely to have an influence on eye movements during reading.
The issue of labelling subject groups as emmetropic and progressing myopes in this
study is not straightforward. If changes in topography associated with lid forces during
reading lead to myopia progression then the amount of time spent reading is an
important factor influencing myopia progression. It is conceivable that a subject
Chapter 5
150
classified as emmetropic might have significant corneal changes associated with lid
forces following reading. However if this subject performs little habitual reading, then
there would be no stimulus for myopia progression. On the other hand a subject with
lesser lid force, who reads often, could become a progressing myope. Therefore the
possible influence of lid forces during reading on myopia progression cannot be
considered in isolation from the amount of time spend reading.
Prolonged near work, which is a well established environmental risk factor associated
with myopia development (Tan et al. 2000; Hepsen et al. 2001; Saw et al. 2002) has
drawn researchers attention in the field of ocular wavefront aberrations. Changes in
spherical aberration with accommodation have been reported to decrease from positive
spherical aberration and changed to negative values with increasing levels of
accommodation (Koomen et al. 1949; Ivanoff 1956; Jenkins 1963; Atchison et al. 1995;
Ninomiya et al. 2002). Third order (coma and coma-like) aberrations are dominant for
most people (Howland and Howland 1976; Howland and Howland 1977; Walsh et al.
1984; Walsh and Charman 1985) and also have found to change with accommodation
(Atchison et al. 1995; He et al. 2000; Ninomiya et al. 2002).
One of the main interests in the field of myopia research is the mechanism of
accommodation and its possible relationship to myopia development. Using auto-
refractometers, a large number of studies have investigated whether accommodative
stimulus/response behaviour differs with refractive error (McBrien and Millodot 1986;
Rosenfield and Gilmartin 1988; Gwiazda et al. 1993; Gwiazda et al. 1995; Abbott et al.
1998). A common hypothesis is that accommodative lag results in a blurred retinal
image (hyperopic defocus), which in turn triggers eye growth. This is consistent with
Chapter 5 _
151
the results of animal studies (Schaeffel et al. 1988; Wildsoet 1997) that have
demonstrated that in the presence of artificially imposed hyperopic retinal defocus (with
negative lenses), the axial length of the eye increases, presumably to compensate for the
induced retinal defocus, and myopia develops. Reduced accommodation stimulus
responses in myopes have been reported in many studies (i.e. lags of accommodation
tend to be higher in myopes) (McBrien and Millodot 1986; Rosenfield and Gilmartin
1988; Gwiazda et al. 1993; Gwiazda et al. 1995). Conflicting reports have found that an
increased lag of accommodation accompanies rather than precedes the development of
myopia (Rosenfield et al. 2002).
Considering the various factors that can affect visual performance during the reading
process, much remains unknown. The complex optical characteristics shown by corneal
distortions and their interaction with internal ocular aberrations have not been
investigated. Studies that use wavefront sensors to investigate far and near viewing
conditions have the potential to provide more detailed information about the eye’s
optical characteristics during and after reading. While there is ongoing research
concerning the accommodation system, little is known about how corneal distortions
might affect accommodation. Furthermore the question must be raised of how much of
the apparent differences found in accommodation behaviour in myopes might have their
origin in the changes in corneal aberrations associated with reading in downgaze.
The hypothesis that corneal distortions during reading may contribute to juvenile-onset
myopia is consistent with the hypothesis of retinal image-mediated ocular growth.
Congenital ptosis has been associated with higher incidence of refractive error. In this
study it is shown that the optical effects of prolonged narrowing of the palpebral
Chapter 5
152
aperture is likely to compromise retinal image quality during reading and thereby could
have an effect on retinal image-mediated ocular growth. The results of this study have
also shown that the distance from pupil centre significantly influences optical changes
due to corneal distortion. In the group of myopes in this study, corneal distortions
following reading tended to be larger and occur closer to the central corneal area