Escolar Documentos
Profissional Documentos
Cultura Documentos
A N D
E X P E R I M E N T A L
OPTOMETRY
INVITED REVIEW
The field of aberrations of the human eye is moving rapidly, being driven by the desire
to monitor and optimise vision following refractive surgery. In this paper, I discuss the
different ways of representing aberrations of the human eye, the terminology used, how
wave aberrations are used to determine refractions, the influence of pupil size on aberrations, how to compare right and left eye aberrations, how aberrations can be manipulated into different forms, how to make corrections for changes in wavelength, the appropriate ocular axis, and corneal and lenticular components of the aberrations.
Prior to the past decade, the field of aberrations of the human eye was almost solely
the preserve of visual optical scientists,
such as myself. The reader may ask why
ophthalmic practitioners should now be
concerned about aberrations when they
have managed for years without them.
There is a number of reasons. In normal
eyes, the conventional sphero-cylindrical
errors dominate but when these are corrected, higher-order aberrations limit our
visual capabilities. In abnormal eyes with
conditions such as keratoconus, higherorder aberrations can be more important
than the conventional refractive errors.
Some treatments for refractive errors, such
as corneal refractive surgery, intra-ocular
lens implants and orthokeratology, can
unintentionally introduce considerable
amounts of higher-order aberrations.
For convenience, aberrations of optical
systems are usually divided into two
groups, the monochromatic and the chromatic aberrations. As the names suggest,
the first group refers to aberrations that
are present when only one wavelength of
light is considered, while the second group
refers to the additional aberrations that
occur when the chromatic dispersion effects of optical media are considered. This
review is about the first group, although
the reader should keep in mind that the
monochromatic aberrations at one wavelength are different from those at another
wavelength. This is particularly important
when dealing with instruments that measure refraction and/or aberrations at infrared wavelengths, and correction is required for the visible spectrum.
Aberrations are often considered to be
optical defects additional to refractive errors but with the use of Zernike aberration terminology and with instruments
that measure both, the distinction between
Clinical and Experimental Optometry 87.3 May 2004
138
aberrations and refractive errors is becoming less useful. It is becoming more common to refer to the refractive errors as
lower-order aberrations and to other effects, such as spherical aberration and
coma as higher-order aberrations. I would
like to emphasise to readers that, for normal eyes at least, the refractive errors are
the most important optical defects and
higher-order aberrations are of consequence only once these refractive errors
are corrected.
This paper addresses the representation
of monochromatic aberrations. Another
paper will consider methods used to measure aberrations, the magnitudes of aberrations in the general population and how
these are affected by various factors (for
example, age, refractive error, accommodation and refractive surgery) and how
aberrations affect spatial visual performance.
Figure 1. Wave, transverse and longitudinal aberrations of an optical system in (A) image space and (B) object space. (B) includes the OSA
co-ordinate system in the pupil. For clarity, the aberrations have been exaggerated and the object plane in (B) has been placed close to the eye
rather than at infinity.
1. Wave aberration
We can image a spherical wavefront moving outwards from an originating point
object. The aberrations of an optical system, such as the eye, prevent the wavefront
from remaining spherical as it passes
through the system. This aberrated
wavefront can be compared with an ideal
spherical wavefront, whose centre of curvature on the image side of the system is
at the ideal image position. For the ideal
wavefront the image point is located in a
position so that it produces an image that
is a perfect copy of any object (allowing
for magnification differences and diffraction effects). The aberrated wavefront may
be compared with the ideal wavefront at
any point in the optical system but a convenient position for doing this is at the exit
pupil of the system. This is the image of
the limiting aperture (stop) as seen from
the image side of the system. At any point
on the exit pupil, the departure of the
aberrated wavefront from the ideal
wavefront, multiplied by the refractive index in the vitreous, gives the wave aberration associated with that point. If the
wavefront is ahead of the ideal wavefront,
the wave aberration is positive. Wave aberrations are small quantities and are usually expressed in micrometres or wavelengths. At a wavelength of 500 nm, one
micrometre (m) is equivalent to two
wavelengths.
2. Transverse aberration
Rays describe the path of a particular point
of the wavefront through the optical system. The transverse aberration of a single
ray is its departure from its ideal position
at the image surface. Usually it has both
horizontal (') and vertical components
('). Transverse aberrations are often
expressed as distances in millimetres or as
angles in minutes of arc.
3. Longitudinal aberration
This is the departure of the intersection,
where it occurs, of a ray with a reference
axis away from its ideal intersection. It may
have different values in the horizontal and
vertical planes. It may be measured as a
length (typically millimetres) or as a power
(dioptres).
In the case of the eye, we do not have
Clinical and Experimental Optometry 87.3 May 2004
139
access to the image side of the optical system and so it is sometimes convenient to
start with the object at the image position
(on the retina) and then to measure the
aberrations of the eye on the object side
of the system (Figure 1b). Instead of being referred to the exit pupil, they are referred to the entrance pupil, which is the
image of the stop (the iris) as seen from
the object side of the optical system. Again,
the wave aberration is positive where the
wavefront is ahead of the ideal wavefront.
As the ideal condition for the eye is emmetropia, the ideal wavefront leaving the
eye will be a plane, although for reasons
of space I have shown the ideal wavefront
as converging to a plane near the eye in
Figure 1b.
The wave aberrations of an optical system such as the eye can be represented by
a wave aberration polynomial function
W(X, Y) = W1 + W2X + W3Y + W4X 2 + W5XY
+ W6Y 2 + W7X 3 + W8X 2Y + W9XY 2 + W10Y 3
+ W11X 4 + W12X 3Y + W13X 2Y 2 + W14XY 3
+ W15Y 4 + higher order terms
[1]
where X and Y are horizontal and vertical
co-ordinates in the entrance pupil. This
polynomial function is sometimes referred
to as a Taylor series. The terms are ar-
W(,) =
cnmZ nm(,)
[3]
n = 0 m = -n
n - |m| = even
m
n
m
n
|m|
n
|m|
n
{NN RR
()cos(m), for m 0
()sin(|m|), for m < 0
[4]
s=0
Nnm = 2(n + 1)
(-1)s(n-s)!
n-2s
s![0.5(n + |m|)-s]![0.5(n - |m| - s]!
[5]
[6a]
140
[6b]
for m = 0 and
W(,) = c j Z j
[7]
j=0
[8]
Number Radial
j
order n
Frequency
m
Polar representation
Monomial representation
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
0
-1
1
-2
0
2
-3
-1
1
3
-4
-2
0
2
4
-5
-3
-1
1
3
5
-6
-4
-2
0
2
4
6
1
2sin()
2cos()
62sin(2)
3(22-1)
62cos(2)
83sin(3)
8(33-2)sin()
8(33-2)cos()
83cos(3)
104sin(4)
10(44-32)sin(2)
5(64-62+1)
10(44-32)cos(2)
104cos(4)
125sin(5)
12(55-43)sin(3)
12(105-123+3)sin()
12(105-123+3)cos()
12(55-43)cos(3)
125cos(5)
146sin(6)
14(66-54)sin(4)
14(156-204+62)sin(2)
7(206-304+122-1)
14 (156-204+62)cos(2)
14(66-54)cos(4)
146cos(6)
1
2y
2x
26xy
3(2x2+2y2-1)
6(x2-y2)
8(3x2y-y3)
8(3x2y+3y3-2y)
8(3x3+3xy2-2x)
8(x3-3xy2)
10(4x3y-4xy3)
10(8x3y+8xy3-6xy)
5(6x4+12x2y2+6y4-6x2-6y2+1)
10(4x4+4x2y2-3x2-4x2y2-4y4+3y2)
10(x4-6x2y2+y4)
12(5x4y-10x2y3+y5)
12(15x4y+10x2y3-12x2y-5y5+4y3)
12(10x4y+20x2y3+10y5-12x2y-12y3+3y)
12 10x5+20x3y2+10xy4-12x3-12xy2+3x)
12(5x5-4x3-10x3y2-15xy4+12xy2)
12(x5-10x3y2+5xy4)
14(6x5y-20x3y3+6xy5)
14(24x5y-20x3y-24xy5+20xy3)
14(30x5y+60x3y3+30xy5-40x3y-40xy3+12xy)
7(20x6+60x4y2+60x2y4+20y6-30x4-60x2y2-30y4+12x2+12y2-1)
14(15x6+15x4y2-20x4+6x2-15x2y4-15y6+20y4-6y2)
14(6x6-5x4-30x4y2-30x2y4+30x2y2+6y6-5y4)
14(x6-15x4y2+15x2y4-y6)
0
1
1
2
2
2
3
3
3
3
4
4
4
4
4
5
5
5
5
5
5
6
6
6
6
6
6
6
is, these are second-order polynomial functions. These are important for determining refraction. The central term is
rotationally symmetric and is called
defocus. Its equivalent as a longitudinal
aberration is mean spherical equivalent
(also called mean sphere). On either side
of it are the astigmatisms. The polynomial
function for which m = -2 is called oblique
astigmatism and has a maximum value of
+2 along the 45 degree meridian and a
minimum value of -2 along the 135 degree
meridian. The other polynomial function,
for which m = +2 is the same but rotated
by -45 degrees to have a maximum value
along the horizontal meridian of +2 and a
minimum value along the vertical merid-
141
Z nmZ n'm' dd = 0
=0 =0
[9]
[10]
142
' or =
-l d W(X,Y )
dX
' or =
-l d W(X,Y )
dY
[11]
(Cnm )2
n >1, all m
[12]
Zernike
Name
polynomial
Coefficient
4 mm pupil ( m)
Coefficient
6 mm pupil ( m)
Z 00
Z 1-1
Z 11
Z 2-2
Z 20
Z 22
0.427
0.707
0.859
-0.025
0.309
-0.276
1.733
1.214
1.404
-0.034
1.356
-0.456
0.118
0.025
0.015
0.039
0.007
-0.009
0.228
0.098
0.046
0.141
0.006
-0.017
0.045
0.357
0.022
0.004
-0.006
-0.024
-0.002
-0.007
0.007
-0.001
0.010
0.009
0.001
-0.002
-0.001
-0.013
0.011
0.439
0.143
-0.20/-0.68 x 3
+0.21/-0.89 x 179
+0.28/-0.92 x 178
0.053
0.009
-0.007
-0.025
-0.015
0.008
-0.007
-0.027
0.015
-0.019
0.003
0.069
-0.011
0.014
-0.010
1.507
0.472
-0.79/-0.50 x 2
+0.38/-0.72 x 179
-0.05/-0.84 x 178
Z 3-3
Z 3-1
Z 31
Z 33
Z 4-4
Z -24
Z 40
Z 42
Z 44
Z 5-5
Z 5-3
Z 5-1
Z 51
Z 53
Z 55
Z 6-6
Z 6-4
Z 6-2
Z 60
Z 62
Z 6-4
Z 6-6
Piston
Vertical tilt
Horizontal tilt
Oblique astigmatism
Defocus
With-/against-the-rule
astigmatism
Oblique trefoil
Vertical coma
Horizontal coma
Horizontal trefoil
Oblique quatrefoil
Oblique secondary
astigmatism
Spherical aberration
With/against the rule
secondary astigmatism
Quatrefoil
Total RMS
Higher order RMS
Refraction 2nd order
Refraction, 2nd and 4th orders
Refraction, 2nd, 4th and 6th orders
Table 2. Zernike aberration coefficients of the eye, the wave aberration maps of which are
shown in Figure 3, and derived quantities
RMS (from third-order and above coefficients) as shown in the fifth-last and fourthlast rows, respectively, of the Table.
Figure 3 shows a three-dimensional map
of the patients aberrations, in which the
height represents the magnitude of aberration as a function of position in the
pupil. A projection (contour map) of the
aberrations is shown as part of this figure.
In Figure 3a, the tilt has been removed
(or set to zero) and the piston manipuClinical and Experimental Optometry 87.3 May 2004
143
5c 04 =
11
13
15
22
24
26
28
37
39
41
43
45
144
and at least theoretically, adding additional terms will not change the lower
terms. The Zernike fit to a wavefront aberration should contain a sufficient
number of terms to model the wavefront
aberration as closely as possible. At some
point, the error of the fit associated with a
coefficient becomes as significant as the
coefficient itself, so that the contribution
of that Zernike polynomial to the model
is no better than adding random noise. It
is important to have an appropriate test
for determining the optimal number of
terms to fit15,16 but these are not yet available in commercial instruments. Using
inappropriate terms can contribute to
giving poor estimates of refraction (see
below). Another thing that must be considered is that taking multiple measurements of the eye may give considerable
variation relative to the mean of some coefficients, which may not be significantly
different from zero.17 As a rough guide,
from my experience for eyes that seem to
have normal optics, there is probably no
need to go above the sixth-order for pupils up to at least six-millimetre diameter.
There may be occasions where higherorder aberrations are needed, for example, in radial keratotomy with eight cuts
at regular intervals, there may be significant higher-order coefficients such as c -88
and c 88.18
USING ZERNIKE ABERRATIONS TO
DETERMINE REFRACTION
INFLUENCE OF PUPIL SIZE
One possible criterion for refraction is to
provide a correcting lens that minimises
the RMS of the wave aberration. This is
achieved by neutralising the second-order
Zernike aberrations. This will be confusing to the practitioner because in the
presence of considerable higher-order
terms, particularly in the fourth-order,
refraction will be dependent on the
pupil size. As an example, comparing
refractions on this basis for pupils of twoand six-millimetre diameter can give
discrepancies of up to one dioptre in
mean spherical equivalent,6 but subjective
refractions show little change across a
wide range of pupil sizes.19-22
[17]
S = M - C/2
[18]
= [tan-1(J45/J180)]/2
[19]
must be applied:
If J180 < 0
= + 90
[19a]
[19b]
[20]
[21]
[22]
145
[23]
SE = -43 (c nm)2/R 2
[24]
[25]
[26]
and
tan-1
nm =
-m
n
m
n
( cc )
[27]
|m|
cnmZnm (,,nm)
[28]
all n and m
146
c 20 = -M.R 2/(43)
[30]
where R is the pupil semi-diameter. Probably, equation (29) gives slight overestimates because it does not take into
account the increased penetration of
infrared radiation compared with that of
visible radiation. There are also subtle
effects depending on the length of the eye
and its refractive error.38 One manufacturer gives a complex equation that allows
for increased penetration into the eye of
0.125 mm and a changing eye length depending on refractive error.39 It gives an
approximate mean spherical correction
from 840 nm to 550 nm of -0.7 D, with
very small changes for wave aberration
coefficients other than c 02.
OFF-AXIS ABERRATIONS
REFERENCE AXIS
H = S + C cos2() V = S + C sin2()
[31]
[32]
147
[36]
surface is spherical and if Q > 0 the surface steepens away from the surface and
is referred to as being oblate. A number
of authors use p to represent the
asphericity, where
15.
16.
p=1+Q
[37]
ACKNOWLEDGEMENT
17.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
148
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
Authors address:
David Atchison
School of Optometry
Queensland University of Technology
Victoria Park Road
Kelvin Grove QLD 4059
AUSTRALIA