18.11.2013 Views

Nonoptical Determinants of Aniseikonia - Investigative ...

Nonoptical Determinants of Aniseikonia - Investigative ...

Nonoptical Determinants of Aniseikonia - Investigative ...

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

<strong>Nonoptical</strong> <strong>Determinants</strong> <strong>of</strong> <strong>Aniseikonia</strong><br />

Arthur Bradley, Jeff Rabin,* and R. D. Freeman<br />

Interocular differences in apparent size (aniseikonia) are typically associated with interocular differences<br />

in refractive error (anisometropia). <strong>Aniseikonia</strong> is generally thought to reflect disparities in<br />

retinal image size that <strong>of</strong>ten accompany anisometropia. This assumption was examined with seven<br />

highly anisometropic subjects who were tested under conditions in which no substantial retinal image<br />

size differences were present. Using a dichoptic size matching task, consistent and large (mean = 22%)<br />

aniseikonias were found. Myopic anisometropes exhibit perceptual minification, while hyperopes<br />

demonstrate perceptual magnification when using their more ametropic eye. Both ultrasonic and<br />

fundus examinations <strong>of</strong> these subjects indicate that differential retinal growth or stretching is responsible<br />

for these findings. Invest Ophthalmol Vis Sci 24:507-512, 1983<br />

Substantial interocular differences in apparent size<br />

(aniseikonia) 1 are typically associated with interocular<br />

differences in refractive error (anisometropia), 2<br />

although small amounts can occur in isometropic<br />

subjects. 3 It has been suggested that both optical and<br />

anatomical or physiologic differences between the<br />

eyes may be responsible for aniseikonia. 1 However,<br />

the critical factor is generally assumed to be the difference<br />

in retinal image size produced by the anisometropia<br />

or its correction. 24 " 7<br />

Most anisometropias <strong>of</strong> greater than 2 diopters result<br />

from interocular differences in axial length rather<br />

than optical power. 8 Therefore, regardless <strong>of</strong> the magnitude<br />

<strong>of</strong> the anisometropia, refractive errors can be<br />

corrected and retinal image sizes equated in the two<br />

eyes by a simple spectacle correction placed at the<br />

anterior focal plane <strong>of</strong> the eyes. This relationship is<br />

described by Knapp's Law, 9 which <strong>of</strong>ten serves as a<br />

rule <strong>of</strong> thumb for correcting anisometropia and eliminating<br />

aniseikonia. 10 However, several recent studies<br />

report substantial amounts <strong>of</strong> aniseikonia present in<br />

anisometropes corrected with spectacle lenses."" 13<br />

These apparent failures <strong>of</strong> Knapp's Law could result<br />

from one or more <strong>of</strong> several reasons. The spectacle<br />

lens positions may have differed substantially from<br />

that <strong>of</strong> the anterior focal plane <strong>of</strong> the eye, or the<br />

anisometropias may have originated from optical and<br />

not axial length differences. In addition, it is possible<br />

that large interocular anatomical or physiological dif-<br />

From the School <strong>of</strong> Optometry, University <strong>of</strong> California,<br />

Berkeley, California.<br />

* Present address: DDEAMC, Fort Gordon, Georgia.<br />

Supported by Grant EY01175 and Research Career Development<br />

Award EY00092 from the US National Eye Institute.<br />

Submitted for publication March 22, 1982.<br />

Reprint requests: R. D. Freeman, School <strong>of</strong> Optometry, University<br />

<strong>of</strong> California, Berkeley, CA 94720.<br />

ferences were responsible. However, because complete<br />

information about lens position, corneal powers,<br />

and ocular dimensions was not provided it is impossible<br />

to discriminate between these alternatives.<br />

In the present study we have attempted to examine<br />

the origins <strong>of</strong> aniseikonia in highly anisometropic<br />

subjects and tried to evaluate the relative importance<br />

<strong>of</strong> optical and neural factors. Under conditions <strong>of</strong><br />

approximate equality <strong>of</strong> retinal image size in the two<br />

eyes we find large amounts <strong>of</strong> aniseikonia, which is<br />

most likely the result <strong>of</strong> differential growth <strong>of</strong> the two<br />

eyes.<br />

Subjects<br />

Materials and Methods<br />

Seven highly anisometropic subjects (five myopes<br />

and two hyperopes) were chosen for this study. All<br />

were examined carefully in the university eye clinic,<br />

and the resulting refractive information is presented<br />

in Table 1. The anisometropias in our sample varied<br />

from 5 to 20 diopters, and they were therefore at the<br />

very edge <strong>of</strong> the normal distribution <strong>of</strong> interocular<br />

refractive error differences. 6 However, very little interocular<br />

difference in corneal power was found<br />

(mean = 0.5 diopters). Also, the average corneal<br />

power <strong>of</strong> our sample was 43 diopters, which is at the<br />

center <strong>of</strong> the normal distribution. 14 Therefore, to a<br />

first approximation, it was reasonable to assume that<br />

we had a sample <strong>of</strong> axial anisometropes who had<br />

approximately equal and typical optical powers in<br />

their eyes. Consequently, using the calculations <strong>of</strong><br />

Gullstrand 15 for the typical eye, we estimated the anterior<br />

focal plane to lie 15 mm anterior to the cornea.<br />

Large deviations from this value are unlikely in our<br />

sample because <strong>of</strong> the homogeneity <strong>of</strong> the corneal<br />

powers (standard deviation = 1.4 diopters).<br />

0146-0404/83/0400/507/$ 1.10 © Association for Research in Vision and Ophthalmology<br />

507


508 INVESTIGATIVE OPHTHALMOLOGY & VISUAL SCIENCE / April 1983 Vol. 24<br />

Table 1. Clinical data from anisometropic subjects<br />

<strong>Aniseikonia</strong> ('%)<br />

Corneal power (D)<br />

Subject<br />

Refractive error<br />

Visual acuity<br />

SL<br />

CL<br />

H<br />

V<br />

JR<br />

AN<br />

-3.25<br />

-10-3 X 180<br />

-19-2.5 X 10<br />

piano<br />

+ 1<br />

-8 -2 X 180<br />

-0.5<br />

+5 -.5 X 50<br />

-7.5 -1X2<br />

-3 -2.25 X 10<br />

+2<br />

+7.5 -1.25 X 90<br />

-9.5 -2.5 X 180<br />

-1.5 -3X2<br />

20/20<br />

20/200<br />

20/180<br />

20/15<br />

-10.2<br />

-34.1<br />

-1.3<br />

-6.5<br />

42.25<br />

42.25<br />

41.75<br />

42.67<br />

43.25<br />

43.75<br />

44.00<br />

43.12<br />

42.00<br />

41.5<br />

44.25<br />

44.25<br />

44.50<br />

44.50<br />

44.20<br />

41.70<br />

CP<br />

20/20<br />

20/400<br />

-27.6<br />

41.50<br />

42.50<br />

44.00<br />

44.00<br />

DH<br />

BP<br />

20/15<br />

20/400<br />

20/15<br />

20/400<br />

+24.5<br />

-11.3<br />

46.50<br />

46.50<br />

CD<br />

20/20<br />

20/40<br />

+26.2<br />

R 42.60<br />

L 42.20<br />

42.25<br />

41.75<br />

DG<br />

20/400<br />

20/15<br />

-20.5<br />

44.00<br />

44.25<br />

R = right eye, L = left eye, H = horizontal, V = vertical, SL = spectacle lens, CL = contact lens.<br />

Procedure<br />

A Moller-Wedell phase differences haploscope was<br />

used to measure aniseikonia. This instrument provides<br />

a dichoptic view by phase-coupling projection<br />

and exposure <strong>of</strong> alternate images to each eye. The left<br />

eye viewed a 3° X 20 min luminous bar, while the<br />

right eye viewed a similar bar <strong>of</strong> adjustable size. Both<br />

bars were viewed in a large empty field, which helped<br />

minimize the effects <strong>of</strong> size constancy by providing<br />

very little distance information. Direct comparison<br />

eikonometry 1 was used to evaluate the aniseikonia.<br />

All subjects were instructed to vary the size <strong>of</strong> the<br />

adjustable bar until both appeared equal. The experimenter<br />

randomly adjusted the length <strong>of</strong> the variable<br />

bar between each setting, and the mean and standard<br />

deviation <strong>of</strong> a minimum <strong>of</strong> eight settings were obtained.<br />

<strong>Aniseikonia</strong> (%) was calculated in the following<br />

way: <strong>Aniseikonia</strong> (%) = ([E - A]/E) X 100%.<br />

Where E and A represent the actual lengths <strong>of</strong> the<br />

bars seen as equally long by the least and most ametropic<br />

eyes respectively. Positive values indicate perceptual<br />

magnification and negative values minification<br />

for the more ametropic eye.<br />

All subjects were tested with their refractive errors<br />

neutralized with trial lenses placed at a vertex distance<br />

<strong>of</strong> approximately 15 mm from the cornea. This distance<br />

was carefully obtained by adjusting variable<br />

head and chin rests until the subject's closed eyelid<br />

just touched a 15-mm protrusion from a lens blank<br />

mounted in the haploscope lens carrier. The trial<br />

lenses were chosen to have a small and constant shape<br />

factor <strong>of</strong> less than 1%, and therefore they magnified<br />

primarily on the basis <strong>of</strong> power. Two subjects were<br />

also tested with their contact lens corrections instead<br />

<strong>of</strong> with trial lenses.<br />

Additional examinations were conducted with several<br />

subjects to obtain fundus photographs and also<br />

A- and B-scan ultrasonographs in order to assess relative<br />

sizes and differential growths <strong>of</strong> the two eyes.<br />

A Zeiss fundus camera was used to photograph the<br />

horizontal meridian from the nasal to temporal ora<br />

serrata. Both axial lengths and overall shapes <strong>of</strong> the<br />

eyes were measured with A- and B-scans, respectively.<br />

A Sonometrics System was used, and the B-scans<br />

were obtained with a sector scan technique.<br />

Results<br />

Knapp's Law states that, in axial anisometropia,<br />

retinal image size is equal in the two eyes if the correcting<br />

lenses are placed at the anterior focal plane<br />

<strong>of</strong> the eyes. Consequently, if aniseikonia is dependent<br />

primarily on retinal image size differences, it should<br />

be negligible in our sample when corrected at a vertex<br />

distance <strong>of</strong> 15 mm. This prediction can be examined<br />

directly in Figure 1, which shows aniseikonia in percent<br />

as a function <strong>of</strong> anisometropia in diopters. Contrary<br />

to the prediction, all <strong>of</strong> our sample exhibit large<br />

amounts <strong>of</strong> aniseikonia when corrected at or around<br />

the anterior focal plane (filled circles). In fact, there<br />

is a systematic relationship between anisometropia<br />

and aniseikonia. All <strong>of</strong> the myopic anisometropes<br />

observed the 3° bar to be smaller with their more<br />

ametropic eye (perceptual minification), while the<br />

converse was true for the hyperopes.<br />

There are several alternative explanations for the<br />

results. First, it is possible that large errors were made


VINO)<br />

No. 4 NONOPTICAL DETERMINANTS OF ANISEIKONIA / Bradley er ol.' : 509<br />

in placing the trial lenses. Although conceivable this<br />

is unlikely, because errors <strong>of</strong> more than 15 mm would<br />

be required to account for these data. 1 The consistancy<br />

<strong>of</strong> the measurements make such large errors<br />

unlikely. However, we have tested the possibility <strong>of</strong><br />

such a positioning error directly by repeating the<br />

measurements with two <strong>of</strong> our sample (AN and JR)<br />

while they wore their contact lens corrections (Fig.<br />

1, open circles). Large position errors could not occur<br />

with a contact lens correction. Therefore, assuming<br />

the refracting surface <strong>of</strong> the contact lens to lie about<br />

3 mm anterior to the entrance pupil <strong>of</strong> the eye, 15 the<br />

formula for spectacle magnification (M = 1/1 - zF,<br />

where M = % magnification, z = distance <strong>of</strong> lens<br />

from entrance pupil, and F = power <strong>of</strong> correcting lens<br />

in diopters) allows the trial lens position to be estimated<br />

based on the measured aniseikonia differences<br />

under these two conditions. For example, subject AN<br />

exhibited a 28% increase in minification with her<br />

myopic eye when the trial lens was used. This places<br />

the trial lens at approximately 16 mm anterior to the<br />

entrance pupil <strong>of</strong> the eye. The corresponding prediction<br />

for subject JR was approximately 14 mm, indicating<br />

spectacle lens positions slightly closer to the<br />

eye than expected. Therefore, it is very unlikely that<br />

the observed aniseikonias were produced by significant<br />

lens positioning errors. In fact, to produce such<br />

large minification for the myopic eyes <strong>of</strong> these subjects<br />

(34% for AN and 10% for JR) the lenses would<br />

need to be much farther, rather than a little nearer<br />

to the eye than the anterior focal plane.<br />

The results shown in Figure 1 are only surprising<br />

if the ametropias are axial. However, these data are<br />

entirely consistant with those expected from refractive<br />

anisometropia, 1 where there should be very small<br />

differences in retinal image size with contact lens correction<br />

but much larger ones with spectacle corrections.<br />

Because most anisometropias greater than 2<br />

diopters are axially produced, 8 it is improbable that<br />

all seven <strong>of</strong> our sample have refractive anisometropias.<br />

The similarity in the keratometric measurements<br />

for each eye (see Table 1) also makes this unlikely.<br />

However, it is possible, for example, that differences<br />

in anterior chamber depth, or lens thickness<br />

are responsible. Therefore, in an attempt to evaluate<br />

the relative roles <strong>of</strong> both optical and axial length differences,<br />

we examined two <strong>of</strong> our sample with A-scan<br />

ultrasonography (Fig. 2). The data from each eye <strong>of</strong><br />

subjects AN and JR are given in Table 2. The anterior<br />

chamber depths and lens thicknesses are essentially<br />

identical in both eyes <strong>of</strong> each subject. However, the<br />

differences in axial length are very close to those expected<br />

if the anisometropias were axial. Subject AN<br />

has 20 diopters <strong>of</strong> myopic anisometropia and an axial<br />

length difference <strong>of</strong> 7.7 mm, which could account for<br />

2<br />

ISEIH<br />

z<br />

<<br />

CP<br />

o<br />

E<br />

min<br />

30 ~<br />

10<br />

0<br />

10<br />

20<br />

30<br />

40 "~<br />

-20<br />

ANISOMETROPIA<br />

Fig. 1. The amount <strong>of</strong> aniseikonia in percent is plotted as a<br />

function <strong>of</strong> anisometropia in diopters for each subject. All subjects<br />

were tested with trial lens corrections (filled circles), and two were<br />

also tested with their contact lens corrections (open circles). Vertical<br />

bars indicate ±1SE.<br />

over 90% <strong>of</strong> her anisometropia. The axial length difference<br />

for subject JR can account for at least 95%<br />

<strong>of</strong> his anisometropia. In conclusion, then, it is very<br />

likely that the anisometropias in our sample result<br />

primarily from axial length differences between the<br />

two eyes.<br />

The accurate lens positioning and confirmed axial<br />

nature <strong>of</strong> the anisometropia emphasize that the aniseikonias<br />

reported by our subjects did, in fact, occur<br />

with approximately congruous retinal images. Therefore,<br />

some very significant nonoptical interocular differences<br />

must be responsible. The large amount <strong>of</strong><br />

neural processing that intervenes between the retinal<br />

image and the appreciation <strong>of</strong> size (the "ocular image"<br />

1 ) provides many possible locations for the source<br />

<strong>of</strong> the aniseikonia. It is certainly possible that central<br />

differences between the inputs from the two eyes<br />

could exist in visual cortex, but the eye itself is a more<br />

likely candidate because differences in size already<br />

exist here.<br />

In order to evaluate what ocular variables were responsible<br />

for the aniseikonias, we made a detailed<br />

examination <strong>of</strong> the globe and retinae <strong>of</strong> some <strong>of</strong> our<br />

sample. First, to assess how differences in axial length<br />

affected the shape <strong>of</strong> the eye, we obtained B-scan ultrasonographs<br />

from subjects AN and JR. Complete<br />

scans <strong>of</strong> both eyes are shown in Figure 3A for one<br />

subject (AN). To facilitate interocular comparison the<br />

right and left halves <strong>of</strong> the sonographs <strong>of</strong> the right<br />

and left eyes respectively have been placed side by


INVE5TIGATIVE OPHTHALMOLOGY b VISUAL SCIENCE / April 1983 Vol. 24<br />

SUBJECT<br />

AN<br />

LE piano RE -20.25<br />

AL = 23.56mm<br />

mmm<br />

1 SUBJECT JR<br />

AL = 31.22 mm<br />

LE -12.50 RE -3.25<br />

AL = 29.13mm<br />

AL = 25.51 mm<br />

Fig. 2. A-scan ultrasonographs were obtained to evaluate the<br />

role <strong>of</strong> axial length in producing the anisometropias in our sample.<br />

Scans are shown from subjects AN (A) and JR (B). Peaks, from<br />

left to right, are produced by the following sources: First, on the<br />

far left, is the emitter artifact. The second peak indicates the position<br />

<strong>of</strong> the cornea. Then comes the peaks from the front and<br />

back surface <strong>of</strong> the lens. Finally, a large peak can be seen from the<br />

retina, choroid, sclera, and orbital tissue. AL = axial length. The<br />

quantitative details <strong>of</strong> these scans are tabulated in Table 2.<br />

side (Fig. 3B). Several aspects <strong>of</strong> globe shape can be<br />

evaluated by simple inspection <strong>of</strong> these figures.First,<br />

from the symmetry <strong>of</strong> the anterior segments it is evident<br />

that the ocular length differences are restricted<br />

to the posterior section <strong>of</strong> the eye. Also, the more<br />

myopic eye does not exhibit much shape distortion.<br />

These pictures indicate that the area <strong>of</strong> the globe covered<br />

by the retina has grown considerably larger in<br />

the more myopic eye. Therefore, when retinal image<br />

sizes have been equated in the two eyes, the image<br />

in the larger more myopic eye covers a smaller proportion<br />

<strong>of</strong> the retina. It is these subjects who report<br />

minification in their more myopic eye when the retinal<br />

images are matched in size. Therefore, it seems<br />

likely that the proportion <strong>of</strong> the retina covered by an<br />

image, and not its absolute size may determine its<br />

apparent size.<br />

Although these pictures show conclusive evidence<br />

<strong>of</strong> differential eye growth in anisometropia, it is not<br />

clear where the extra (in the case <strong>of</strong> myopes) growth<br />

occurs. Are these differences in growth spread out<br />

uniformly over the globe posterior to the ora serata,<br />

or are they concentrated at the posterior pole <strong>of</strong> the<br />

eye? The interpretation <strong>of</strong> our size matching data<br />

depends on the answer to this question, because we<br />

sampled only a very small and select region <strong>of</strong> the<br />

retina (1.5° on either side <strong>of</strong> the fovea). Therefore,<br />

to examine differential growth, or stretching across<br />

the retina we took a series <strong>of</strong> fundus shots across the<br />

horizontal meridian <strong>of</strong> the eyes from several subjects.<br />

Sample data are shown in Figure 4. A fundus comparison<br />

<strong>of</strong> an emmetropic (left) eye and highly myopic<br />

(right) eye <strong>of</strong> subject AN can be made from Figure<br />

4A. The fundus <strong>of</strong> the left emmetropic eye is<br />

ophthalmoscopically normal. The uniform coloring<br />

<strong>of</strong> the retina is typical. However, the fundus <strong>of</strong> the<br />

right myopic eye exhibits many features commonly<br />

found in highly myopic eyes that have undergone<br />

excessive growth or stretching. First, the separation<br />

<strong>of</strong> the retina, pigment epithelium, and choroid from<br />

around the optic disc is typical. Second, the pale and<br />

nonuniform fundus is indicative <strong>of</strong> stretching and<br />

thinning. Third, the thin retinal vessels, and the visible<br />

choroidal vessels also indicate stretching. It is<br />

clear from these two pictures that the area <strong>of</strong> retina<br />

that manifested perceptual minification has undergone<br />

excessive stretching. Therefore, in our experiment,<br />

where retinal image size was held constant in<br />

the two eyes <strong>of</strong> this subject (AN), the image in the<br />

myopic right eye fell on a retina containing fewer<br />

Table 2. A-scan ultrasonography data: ocular<br />

dimensions in millimeters for two myopic<br />

anisometropes whose ultrasonographs<br />

are shown in figures 2 and 3.<br />

Ocular dimention<br />

Anterior cornea/<br />

anterior lens<br />

Lens thickness<br />

Vitreous chamber<br />

depth<br />

Axial length<br />

LE<br />

3.7<br />

3.4<br />

16.4<br />

23.6<br />

LE = left eye, RE = right eye.<br />

AN<br />

RE<br />

3.6<br />

3.7<br />

23.8<br />

31.2<br />

Subject<br />

LE<br />

4.0<br />

4.1<br />

20.9<br />

29.1<br />

JR<br />

RE<br />

3.8<br />

3.8<br />

18.0<br />

25.5


No. 4 NONOPTICAL DETERMINANTS OF ANI5EIKONIA / Bradley er al. 511<br />

SUBJECT<br />

AN<br />

LE RE SUBJECT<br />

M<br />

J R<br />

Fig. 3. The shape <strong>of</strong> the globe was evaluated using B-scan ultrasonographs. A, Full scans <strong>of</strong> both the right and left eyes <strong>of</strong> subject AN.<br />

Interocular comparison is facilitated in B, where the abutting left and right halves <strong>of</strong> the left and right eyes, respectively, are shown for<br />

subjects AN and JR.<br />

photoreceptors per unit area than the emmetropic left<br />

eye. This deduction is not unreasonable considering<br />

that there is no evidence <strong>of</strong> additonal retinal .growth<br />

to compensate for the extra size <strong>of</strong> the eye. Consequently,<br />

it seems likely that apparent size is determined<br />

by the number <strong>of</strong> retinal elements stimulated<br />

and not the absolute size <strong>of</strong> the retinal image.<br />

This conclusion is supported further by the evidence<br />

given in Figure 4B, showing the fundae <strong>of</strong> two<br />

myopic eyes. The first is from the left eye <strong>of</strong> subject<br />

JR who has slightly more anisometropia and myopia<br />

than subject DG, whose right eye fundus is shown in<br />

the lower picture. Although JR has more myopia, the<br />

fundus <strong>of</strong> DG shows considerably more signs <strong>of</strong><br />

stretching. The significance <strong>of</strong> this difference can be<br />

seen by comparing the amount <strong>of</strong> perceptual minification<br />

exhibited by the myopic eyes <strong>of</strong> these two<br />

subjects (Fig. 1). Subject DG exhibits twice as much<br />

minification (20%) than JR (10%). Therefore, even<br />

for subjects with similar amounts <strong>of</strong> anisometropia<br />

the degree <strong>of</strong> aniseikonia correlates well with the<br />

amount <strong>of</strong> retinal stretching. This emphasizes both<br />

the importance <strong>of</strong> nonoptical factors in determining<br />

aniseikonia, and also the weakness <strong>of</strong> any optically<br />

based rule <strong>of</strong> thumb for correcting it.<br />

Discussion<br />

Although the possibility <strong>of</strong> nonoptical factors influencing<br />

aniseikonia has been appreciated for some<br />

time, 1 such influences are generally assumed to be<br />

secondary, and most emphasis in the clinical literature<br />

is placed on the role <strong>of</strong> retinal image size. 4 However,<br />

our results show convincingly that very substantial<br />

aniseikonias can be observed with congruous<br />

retinal images. The findings also emphasize that previous<br />

reports <strong>of</strong> large amounts <strong>of</strong> aniseikonia in anisometropes<br />

corrected with spectacle lenses 1213 are<br />

probably not due to retinal image size disparities.<br />

A)<br />

RE<br />

B)<br />

LE<br />

RE<br />

Fig. 4, Composite fundus photographs are shown here from three<br />

subjects. A, Both the left (top and right (bottom) eyes <strong>of</strong> subject<br />

AN. B, The more myopic left eye <strong>of</strong> subject JR (top) and more<br />

myopic right eye <strong>of</strong> DG (bottom).


512 INVESTIGATIVE OPHTHALMOLOGY & VISUAL SCIENCE / April 1983 Vol. 24<br />

At the outset we posed the question: are optical or<br />

nonoptical factors most important for producing aniseikonia?<br />

It turns out that these two types <strong>of</strong> interocular<br />

difference are always opposite in sign. Myopic<br />

anisometropes have optical magnification and neural<br />

minification. Conversely, hyperopic anisometropes<br />

have optical minification, but nonoptical magnification.<br />

Therefore, the relative importance <strong>of</strong> each factor<br />

can be gauged by measuring the resultant aniseikonia<br />

when both are present, that is, for the uncorrected<br />

eye. Unfortunately this experiment is very<br />

difficult to do because <strong>of</strong> the poor image quality.<br />

However, we have already measured the neural magnification<br />

differences, and it is simple to predict the<br />

optical differences. The uncorrected retinal image size<br />

differences (%) in axial anisometropia are approximately<br />

1.7 X RX, where RX is the spectacle correction<br />

necessary at the anterior focal plane. Only one<br />

<strong>of</strong> our sample (JR) had a neural difference that is<br />

smaller than the predicted retinal image size difference.<br />

All other subjects exhibited larger neural than<br />

optical differences. Therefore, we conclude that, in<br />

general, there are likely to be larger neural than optical<br />

magnification differences between the two eyes<br />

in high anisometropia.<br />

Although these findings are somewhat contrary to<br />

those expected, the explanation seems straight-forward.<br />

As the eye grows, it increases in length and<br />

circumference. Consequently, the retinal image size<br />

increases, but so does the area <strong>of</strong> retina receiving the<br />

image. To a first approximation these opposite effects<br />

cancel. However, our data indicate that the retinal<br />

stretching around the fovea may increase by a greater<br />

proportion than the axial length <strong>of</strong> the eye. Therefore,<br />

although, for example, the retinal image in an uncorrected<br />

myopic eye is very large, it may actually<br />

stimulate a smaller number <strong>of</strong> retinal photoreceptors.<br />

The clinical implications <strong>of</strong> these data are clear.<br />

First, contrary to popular belief, Knapp's Law, although<br />

an adequate approximation for equating retinal<br />

image size, cannot be used as a rule <strong>of</strong> thumb<br />

for eliminating aniseikonia in anisometropes. Second,<br />

it may seem from our data (Fig. 1) that, a contact<br />

lens correction would serve as a useful correction for<br />

these patients. However, before accepting this conclusion<br />

it is worth noting a major limitation <strong>of</strong> this<br />

study. Our results only represent a very small region<br />

<strong>of</strong> the retina, and it is clear from the fundus photographs<br />

in Figure 4 that the retina is likely to be<br />

stretched in a very nonuniform way. Consequently,<br />

what may be the perfect correction for one part <strong>of</strong><br />

the retina may introduce large aniseikonias in another<br />

part <strong>of</strong> the visual field. Indeed, it may be impossible<br />

to correct for both the anisometropia and the aniseikonia<br />

for any large region <strong>of</strong> the retina. Eliminating<br />

aniseikonia in the central retina while ignoring possible<br />

interocular differences in perceived size in the<br />

periphery may be the best compromise.<br />

Key words: Anisometropia, aniseikonia, magnification,<br />

myopia, hyperopia, neural development<br />

Acknowledgments<br />

The authors thank Drs. R. D. Stone and D. Sheets for<br />

their help in producing, respectively, the ultrasonographs<br />

and fundus photographs.<br />

References<br />

1. Ogle KN: Researches in Binocular Vision. Philadelphia, WB<br />

Saunders, 1950.<br />

2. Lancaster WB: <strong>Aniseikonia</strong>. Arch Ophthalmol 20:907, 1938.<br />

3. Carleton EH and Madigan LF: Relationships between aniseikonia<br />

and ametropia; from a statistical study <strong>of</strong> clinical cases.<br />

Arch Ophthalmol 18:237, 1937.<br />

4. Von Bahr G: An analysis <strong>of</strong> the change in perceptual size <strong>of</strong><br />

the retinal image at correction <strong>of</strong> ametropia. Doc Ophthalmol<br />

20:530, 1966.<br />

5. Mills PV: <strong>Aniseikonia</strong> in corrected anisometropia. Br Orthopt<br />

J 36:36, 1979.<br />

6. Rayner AW: <strong>Aniseikonia</strong> and magnification in ophthalmic<br />

lenses. Problems and solutions. Am J Optom 43:617, 1966.<br />

7. Straatsma BR, Heckenlively JR, Foos RY, and Shahinian JK:<br />

Myelinated retinal nerve fibers associated with ipsilateral myopia,<br />

amblyopia, and strabismus. Am J Ophthalmol 88:506,<br />

1979.<br />

8. Sorsby A, Leary GA, and Richards MJ: The optical components<br />

<strong>of</strong> anisometropia. Vision Res 2:43, 1962.<br />

9. Knapp H: The influence <strong>of</strong> spectacles on the optical constants<br />

and visual acuteness <strong>of</strong> the eye. Arch Ophthalmol Otol 1:377,<br />

1869.<br />

10. Katz M: The human eye as an optical system. In Clinical<br />

Ophthalmology, Duane, TD, editor. Hagerstown, Harper and<br />

Row, 1981, Vol. 1, Chapter 33.<br />

11. Arner RS: Eikonometer measurements in anisometropes with<br />

spectacles and contact lenses. J Am Optom Assoc 40:712,<br />

1969.<br />

12. Rose L and Levinson A: Anisometropia and aniseikonia. Am<br />

J Optom 49:480, 1972.<br />

13. Awaya S and von Noorden GK: <strong>Aniseikonia</strong> measurement by<br />

phase difference haploscope in myopic anisometropia and unilateral<br />

aphakia (with special reference to Knapp's law and comparison<br />

between correction with spectacle lenses and contact<br />

lenses). J Jpn Contact Lens Soc 13:131, 1971.<br />

14. Sorsby A, Benjamin B, Davey JB, Sheridan M, and Tanner<br />

JM: Emmetropia and its abberations. Med Res Council Spec<br />

Rep Ser No. 293. London, HMSO, 1957.<br />

15. Gullstrand A: Appendix A. In Helmholtz's Treatise on Physiological<br />

Optics. Translated from the 3d German edition,<br />

Southall, J. P. G, editor. Rochester, The Optical Society <strong>of</strong><br />

America, 1924.

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!