SEARCH

SEARCH BY CITATION

Keywords:

  • amblyopia;
  • children;
  • contrast sensitivity;
  • vision;
  • visual acuity

Abstract

  1. Top of page
  2. Abstract
  3. METHODS
  4. RESULTS
  5. DISCUSSION
  6. SUMMARY
  7. ACKNOWLEDGEMENT
  8. REFERENCES

Background:  Dichoptic visual stimulation may be achieved using shutter goggles and mirror systems. These methods vary in their feasibility for use in children. This study aims to investigate the feasibility of use of a simple trial frame-based system to evaluate interactions in children.

Methods:  Low contrast acuity, contrast sensitivity and alignment sensitivity were measured in the non-dominant eye of 10 normally-sighted children, 14 anisometropic children without amblyopia and 14 anisometropic amblyopic children (aged 5–11 years) using goggles and a trial frame apparatus (TFA). The dominant eye was either fully or partially occluded. The difference in visual functions in the non-dominant eye between the full and partial occlusion conditions was termed the ‘interaction index’. Agreement between the TFA and goggles in terms of visual functions and interactions was assessed in anisometropic children with and without amblyopia using the Bland-Altman method and t-test. Training sessions allowed subjects to become accustomed to the systems and tasks. The duration of training, the number of breaks requested by subjects and their willingness to attend further experiments were recorded in 10 subjects from each group and were compared between groups and between systems.

Results:  Both Bland-Altman and t-test methods indicated acceptable agreement between the TFA and goggles in visual function and interaction measures (p > 0.05), except for contrast sensitivity measured in anisometropic children without amblyopia (p = 0.042). For all subject groups, contrast sensitivity training was significantly longer using goggles than using the TFA (p ≤ 0.001). Significantly more breaks were requested in acuity and contrast sensitivity testing, when goggles were used than when the TFA was used (p < 0.045). Anisometropic children without amblyopia showed a significantly greater willingness to attend more experiments using the TFA than using goggles (p = 0.025).

Conclusion:  The TFA may be a useful tool in studies of interactions in amblyopes, particularly in studies of children's vision.

Dichoptic visual masking occurs when a target is presented to one eye and a mask is presented to the fellow eye. The stimulus and masking configurations may take various forms, including ‘masking light by light’, ‘masking a pattern by light’ and ‘masking a pattern by a pattern’.1 The mask may be presented either coincident with or adjacent to the stimulus.2–10 In the normal visual system, the target and the mask are processed binocularly and the mask can affect the perception of the target, resulting in an increment or a decrement in visual function thresholds of the tested eye.2,3,8–10 These masking effects are underpinned by interactions in the visual system that take place between signals from the two eyes, known as interocular interactions. In dichoptic masking studies, the impact of the masking stimulus on perception of the test stimulus is assessed. Comparison between visual functions measured under dichoptic and monocular viewing conditions allows the nature and the strength of interocular interactions to be investigated.

Mirror or prism devices and shutter goggles have been used in dichoptic masking studies;3,11–13 however, equipment of this kind may not be feasible for use in some populations.14 For this reason, a simple and portable viewing system, a ‘trial frame apparatus’ (TFA) was designed and validated in a previous study.14 The TFA consisted of a trial frame adjusted to the individual's interpupillary distance, a pinhole aperture (1.0 mm in diameter) placed in front of the tested eye and a partial occluder placed in front of the non-tested eye. The occluder was made using a +5.00 DS trial lens, on which four strips of 5.0 mm wide opaque black masking tape were placed to form a ‘star-shaped’ occlusion (Figure 1). Visual function and interocular interaction measurements made using this method were found to agree with those made using shutter goggles in children and adults with normal vision.14

image

Figure 1. The trial frame apparatus (TFA), 1.0 mm pinhole, opaque occluder and the star-shaped occluder used with the TFA. The star-shaped occluder was made using a +5.00 DS trial lens with four strips of 5.0 mm wide black masking tape. The opaque occlusion at the centre of the star-shaped occluder was 13 mm in diameter.

Download figure to PowerPoint

Amblyopia is often known as ‘lazy eye’. It refers to a disorder of the visual system that is characterised by unilateral or bilateral loss of visual acuity without pathology.15 Anisometropia, which has a prevalence of two to six per cent in school children,16,17 is one of the major causes of amblyopia. Previous studies indicate that in humans, about one-third of cases of amblyopia are caused by anisometropia, one-third by strabismus and one-third by a combination of both.18,19 Deprivation due to conditions such as congenital cataracts can also lead to amblyopia.20

Numerous studies have demonstrated that interocular interactions differ between amblyopic and normal observers. For example, binocular summation is reduced or absent in amblyopes.21,22 Holopigian, Blake and Greenwald23 found that stereoacuity and binocular summation are absent at high spatial frequencies only, while Hood and Morrison24 found a higher level of binocular contrast summation in amblyopes with binocular single vision than in normal observers, for both low and high spatial frequencies. Baker and colleagues25 found that binocular contrast summation is normal in amblyopes when contrast sensitivity is normalised across the amblyopic and the fellow eyes. Studies by Sengpiel and Blakemore26, and Smith and colleagues27 have demonstrated inhibitory interaction in animals with amblyopia and Vedamurthy and colleagues2 found significant inhibitory interaction in anisometropic amblyopic adults.2 Most of these findings suggest that abnormal interactions occur in anisometropic amblyopia. Studying interactions in anisometropic amblyopes may provide information on visual processing in these subjects and the mechanisms underlying this disorder.

Comparison between interactions in anisometropes with and without amblyopia may be informative because both groups have significant differences in refractive error between eyes but only one of these groups has developed amblyopia. The aim of this study is to investigate the feasibility of use of the TFA as a means of studying interocular interactions in children with and without anisometropic amblyopia. Specifically, to investigate:

  • 1
    Agreement in visual function and interocular interaction measures between the TFA and shutter goggles in anisometropic children with and without amblyopia and
  • 2
    The feasibility of use of the TFA as a viewing system to evaluate interactions in children.

METHODS

  1. Top of page
  2. Abstract
  3. METHODS
  4. RESULTS
  5. DISCUSSION
  6. SUMMARY
  7. ACKNOWLEDGEMENT
  8. REFERENCES

Subjects

Subjects were 5–11 years old, including 10 normally sighted children (mean age: 9.2 ± 1.5 years; seven male, three female), 14 anisometropic children without amblyopia (mean age: 8.8 ± 0.8 years; six male, eight female) and 14 anisometropic amblyopic children (mean age: 8.8 ± 1.0 years; eight male, six female). These three subject groups are referred to as ‘controls’, ‘anisometropes’ and ‘amblyopes’, respectively. All subjects were recruited through a school-based screening program in Huadu, China. Table 1 shows the clinical characteristics of the anisometropes and amblyopes. Ethical approval for the present study was obtained from the Human Research Ethics Committee of the University of New South Wales in Australia and the ethics committee of the Zhongshan Ophthalmic Center in China. Written, informed consent was obtained before recruitment from parents of all subjects. Assent was obtained from each subject. This study adhered to the tenets of the Declaration of Helsinki.

Table 1. Clinical characteristics of anisometropic children with [a] and without [b] amblyopia. VA, visual acuity measured in the screening program (using a high contrast E chart); DE, dominant eye; NDE, non-dominant eye; IDVA, interocular difference in high contrast acuity. No monocular suppression or diplopia was reported in any of these subjects (Worth four dot test).
(a) Anisometropic amblyopic children
SubjectSexAge (years)PrescriptionVA (logMAR)IDVA (logMAR)Stereoacuity (arcsec)
Right eyeLeft eyeDENDE
 1M6.4plano+3.00-0.20.10.350
 2M6.5+0.25+6.50/-2.00 × 180-0.10.50.6400
 3M7.4+6.25+7.000.10.30.2400
 4F8.30/-1.75 × 95-1.0000.10.130
 5F8.4+2.25/-0.50 × 170+0.50/-1.00 × 10-0.20.20.470
 6M8.4-4.00/-2.25 × 50-2.50/-0.50 × 17000.150.1550
 7F8.5+0.75/-0.50 × 130+3.7500.50.5100
 8F8.6+4.25/-0.50 × 155+0.25/-0.50 × 15-0.10.30.470
 9F9.1+4.50/-1.00 × 160plano-0.10.20.350
10M9.2+0.50+7.00/-2.00 × 45-0.20.20.4400
11F9.3-4.50/-1.50 × 10-1.25-0.20.40.650
12M9.5-3.25/-0.75 × 170plano00.150.1530
13M9.8-6.00/-1.75 × 40-4.25/-1.00 × 13000.10.140
14M10.1+2.50/-2.50 × 180+4.50/-2.50 × 18001.01.0400
(b) Anisometropic children without amblyopia
SubjectSexAge (years)PrescriptionVA (logMAR)Stereoacuity (arcsec)
Right eyeLeft eyeDENDE
 1M7.1-1.75-1.75/-1.00 × 850040
 2F7.20/-3.25 × 1750/-2.25 × 50070
 3F8.3-2.50-2.75/-1.25 × 165-0.1-0.140
 4M8.3-2.25-1.00-0.2-0.230
 5F8.8-4.75/-0.50 × 180-3.25/-0.50 × 1700030
 6F8.9-0.25-1.500030
 7M9.0-6.50-4.75/-0.50 × 300030
 8F9.3-5.00/-1.50 × 35-6.25/-2.00 × 1500040
 9F9.3+0.50+1.25/-0.75 × 155-0.1-0.130
10F9.3+1.00/-1.00 × 165+2.00/-2.25 × 175-0.2-0.2140
11F9.6-12.50-10.500030
12F9.7-4.50-4.50/-1.00 × 1600040
13F9.8+3.00/-0.75 × 180-1.00/-0.50 × 17500140
14M9.8-3.75-2.25/-0.75 × 1600030

A series of vision screening tests was conducted as part of the screening program, including a clinical acuity test for distance (a high-contrast E chart at 5 m), cover test for distance and near (with refractive correction if needed; correction was prescribed by one of the authors [XJL] based on both cycloplegic [1% cyclopentolate] and non-cycloplegic refractions28,29 and was applied throughout this study), binocularity (Worth four dot test at 6 m), stereopsis (Randot stereoacuity test at 40 cm), sighting dominance test30,31 and fundus examination, including assessment of central fixation (direct ophthalmoscopy). No ocular deviation or eccentric fixation was found in any of the subjects.

Subjects were included only if they had no history of ocular trauma and/or ocular pathology, no systemic disease (by self-report), no strabismus (based on cover test), no previous or current treatment of anisometropia and/or amblyopia (refractive correction or occlusion), no eccentric fixation and they met the following criteria:

CONTROLS
  • 1
    Uncorrected vision of 0.0 logMAR or better in each eye, with an interocular difference of less than 0.1 logMAR
  • 2
    spherical and cylindrical refractive error of 0.50 D or less for distance
  • 3
    stereopsis of equal to or better than 40 seconds of arc (arcsec).
ANISOMETROPES AND AMBLYOPES
  • 1
    Visual acuity for anisometropes: 0.1 logMAR or better in each eye, with an interocular difference of less than 0.1 logMAR
  • 2
    visual acuity for amblyopes: 0.1 logMAR or worse in one eye and 0.1 logMAR or better in the other eye, with an interocular difference of 0.1 logMAR or more
  • 3
    interocular difference in spherical refractive error of 0.75 DS or more for hyperopic children, 1.25 DS or more for myopic children and cylindrical refractive error of 1.00 DC or more for children with astigmatism (refractive errors were defined using negative cylinder).32

Apparatus

Visual stimuli were generated using a VSG 2/5 graphics card (Cambridge Research Systems, Rochester, UK) externally connected to an HP 8530P laptop computer and were displayed on a gamma-corrected 21-inch Trinitron Sony GDM-F520 cathode ray tube monitor. The refresh rate was 120 Hz. The mean room illuminance was 4.78 ± 2.76 Lux (Konica Minolta T-10 illuminance meter). Three visual functions were measured in each subject, namely, low contrast acuity (-0.20 Weber contrast), contrast sensitivity and alignment sensitivity. Two viewing systems were used, namely, ferro-electric shutter goggles (FE-1, Cambridge Research Systems) and the TFA. Using each of these viewing systems, each visual function was measured under both full and partial occlusion conditions (see below). For anisometropes and amblyopes, refractive error was always corrected using trial lenses.

SHUTTER GOGGLES

The shutter goggles were synchronised with the monitor so that alternate frames were presented to each eye (for example, odd-numbered frames to right eye, even-numbered frames to left eye). Thus, each eye viewed the stimuli at a refresh rate of 60 Hz. The background luminance of the monitor was fixed at 170 cd/m2 and this level was reduced to about 21 cd/m2 at each eye by the goggles. The goggles were worn using an elasticated strap and were held in place by an assistant to reduce their weight and to minimise discomfort. In the full occlusion (monocular) condition, an opaque eye-patch was used to cover the non-tested eye, with a trial frame holding the subject's refractive correction over it (if needed) and with the shutter goggles outermost. In the partial occlusion condition, subjects wore a trial frame with refractive correction (if needed) and the goggles, with no eye-patch.

TRIAL FRAME APPARATUS

The design and validation of the TFA (Figure 1) are described elsewhere.14 In both full and partial occlusion, a pinhole aperture (1.0 mm in diameter) was placed in front of the tested eye. The non-tested eye was covered using the star-shaped occluder in the partial occlusion condition and using an opaque cover for full occlusion. The star-shaped occluder consisted of one opaque black masking tape at horizontal, one at vertical and the other two at 45° and 135°. Depending on the vertex distance, the opaque occlusion at the centre of the star-shaped occluder blocked out between 22° and 24° of the central visual field of the non-tested eye but light transmission from the periphery was allowed.14 The pinhole transmits approximately 10° of central field to the tested eye and the field more peripheral to this was vignetted. Thus, in partial occlusion, the view consists of the target presented to the tested eye coincident with the central occlusion of the non-tested eye. While the experiments were conducted in a darkened room, even with a pupil diameter of 8.0 mm, the field of view through the pinhole remained below 20° and always below the peripheral field angle of the star-shaped occluder.

Stimuli and experimental tasks

A training session was conducted for each subject before the main experiments, to allow the subject to become accustomed to the tasks with each viewing system. The stimuli and tasks described here apply to the main experiments. Any differences for the training sessions are indicated in Procedures.

On each trial, a central target was presented to the tested eye only for a period of 140 ms (400 ms in training sessions), to minimise the impact of eye movements.33 With full occlusion, the non-tested eye was fully occluded, while with partial occlusion, the non-tested eye was masked with either a square low-luminance patch (viewed via shutter goggles and referred to as ‘square partial occlusion’; Figure 2) or a star-shaped occluder (viewed via the TFA; Figure 1). A fusion lock was constantly presented to both eyes. It was visible to the tested eye only in the full occlusion condition when the shutter goggles were used and in both viewing conditions when the TFA was used. Suppression markers (see below) were used with the shutter goggles (not with the TFA) in partial occlusion only to check for suppression of either eye.

image

Figure 2. Examples of stimuli presented to the tested eye [i] and the non-tested eye [ii] with partial occlusion for the acuity [A], contrast sensitivity (central target at 4 cpd is presented here, virtual target was at 6 cpd) [B] and alignment sensitivity (central target at 2 cpd is presented here, virtual target was at 6 cpd) [C] measurements when shutter goggles were used as a viewing system

Download figure to PowerPoint

CENTRAL TARGET

Acuity was assessed with a single letter E presented at one of four possible orientations (right, left, up or down) at a viewing distance of 4 m. The letter was constructed in a 5 × 5 grid, in which each stroke and gap was one-fifth of the dimension of the square grid.

The contrast sensitivity test consisted of a circular Gabor patch (vertical, at six cycles per degree [cpd]) subtending 3.5° at a viewing distance of 2 m, with the standard deviation of the Gaussian envelope 0.65°.

Alignment sensitivity was assessed with three Gabor patches at 65 per cent Michelson contrast (vertical, at six cycles per degree, presented at a viewing distance of 1 m) with the upper and lower patches in vertical alignment and the central patch displaced to the left or right relative to this alignment on each trial.

SQUARE PARTIAL OCCLUSION

The square partial occlusion was at -0.78 Weber contrast, subtending 2.3°, 3.5° and 9.5° of the visual field for acuity, contrast sensitivity and alignment sensitivity tests, respectively. This form of partial occlusion applied only to the shutter goggles.

FUSION LOCK

The fusion lock, intended to maintain in the dichoptic, partial occlusion conditions, was a ring target at -0.78 Weber contrast with a width of 0.1°, subtending 4° for acuity, 8° for contrast sensitivity and 15° for alignment sensitivity.

SUPPRESSION MARKERS

Suppression markers were four lines, two presented to each eye (subtending 0.3°, 2.5° and 1° at locations 6°, 10° and 19.5° in the periphery for acuity, contrast sensitivity and alignment sensitivity measurements, respectively). These markers were presented in the goggle square partial occlusion condition only.

The experimental tasks were identical in the training and the main experiments. These tasks were:

  • 1
    in the acuity test, to indicate by pointing the orientation of the E target
  • 2
    contrast sensitivity was measured using a temporal two-alternative forced choice method, in which subjects verbally reported whether the Gabor patch was presented in interval ‘one’ or ‘two’ after each trial and
  • 3
    in the alignment sensitivity test, the task was to point to the offset direction (left or right) of the central Gabor patch.

All responses were input to the program by the examiner (author XJL). No feedback was given.

Procedures

In the training session, the dominant eye of each subject was tested in all three visual functions with partial occlusion only, using both viewing systems (all tests were conducted in a pseudo-random order). In these sessions, a two-down one-up (2/1) single staircase method was used for acuity and contrast sensitivity testing, while a 1/1 double staircase method was used for alignment sensitivity testing. Start levels were based on previous findings to ensure subjects could easily detect the stimuli, and the step sizes were sufficiently small to ensure the stimuli were always supra-threshold during training.2,34,35 The square partial occlusion used with shutter goggles in this training session subtended 3°, 5° and 7° in the acuity, contrast sensitivity and alignment sensitivity tests, respectively. Subjects who could give 10 consecutively correct answers were regarded as having passed the training for each visual function and viewing system. Duration of training sessions for each visual function test were recorded using a stopwatch in 10 subjects of each group, for the TFA and the goggles, respectively.

The main experiments were conducted within three days of the training. The three visual functions were measured in the non-dominant eye only in each subject. Four testing conditions were used:

  • 1
    shutter goggles with full occlusion
  • 2
    shutter goggles with partial occlusion
  • 3
    trial frame with full occlusion and
  • 4
    trial frame with partial occlusion.

Different viewing systems were tested on separate days and all tests using each viewing system were conducted in a pseudo-random order. Initially, individual start level was determined using a 1/1 single staircase method in the acuity and contrast sensitivity tests and a 1/1 double staircase method in the alignment sensitivity test. Following this, thresholds were measured using a 2/1 double staircase method in the acuity and contrast sensitivity tests and a 1/1 double staircase method in the alignment sensitivity test. The step sizes were 0.08 logMAR, 3.5 dB and 1.5 arcmin in the acuity, contrast sensitivity and alignment sensitivity tests, respectively.

In both training and main experiments, an ‘unforced-choice method’36 (allowing ‘don't know’ answers) was used in acuity and contrast sensitivity tests due to long experimental duration (up to eight minutes), to enhance subjects' co-operation and to increase the likelihood of test completion.36,37 Although subjects were not forced to give a response to each trial, they tended to give as many answers as they could, with an average of only one ‘don't know’ answer given by each subject during each test. An ‘incorrect’ response was recorded for the ‘don't know’ answers. A ‘forced-choice method’38 was used in the alignment sensitivity test to reduce the variability of the subjects' judgment of alignment.

A game-like atmosphere39 was employed. At the beginning of the experiments, the child was asked to add his/her name to a list of ‘challengers’ pasted on a wall and was told that he/she was going to play three games (three visual function tests) that day. Each game included two stages (two viewing conditions) and he/she was allowed to paste a little red flag after his/her name when he/she passed a stage of a game. At the end of each day, the child who had most flags would be the ‘champion of red flags’. The examiner (XJL) frequently praised and encouraged the subjects throughout the experiments and words were carefully used in experimental task instructions to ensure they were easily understood by the subjects.

Subjects were tested in groups of two during the main experiments and thus they were tested in turns and a break was given after each visual function test under one viewing condition. Subjects were also allowed to take breaks by request at any time during a test. The number of breaks requested by subjects (excluding those given after each visual function test) was recorded in 10 subjects of each group, for the TFA and shutter goggles respectively.

After all tests, 10 subjects of each group were given a slip of paper with one question and two response options. They were asked to mark one or both of the option boxes. The question was:

‘Do you want to play some more games with us in the next semester? Please check the box to tell us which game you would like to play and cross the one you don't like. Check both of them if you like both of them and cross both of them if you don't like either of them.’

Response options were: ‘(A) Glasses with spider web or (B) Diving mask’, representing (A) TFA and (B) shutter goggles, respectively.

Data analysis

THRESHOLDS AND INTERACTION INDEX

In the single and double staircase methods, trials up to the first reversal on each track were excluded. In the acuity and contrast sensitivity tests, the mean of mid-points of peaks and valleys of the remaining reversals were taken to represent threshold. In the alignment sensitivity test, the standard deviation (SD) of the mid-points was taken to represent the variability of a subject's judgment of alignment and the reciprocal of this was recorded as alignment sensitivity.

The difference in non-dominant eye visual function between the full and partial occlusion is termed the ‘interaction index’. As a threshold measure, the acuity ‘interaction index’ was calculated by the following normalising function:

  • image

As sensitivity measures, contrast and alignment sensitivity interaction indices were calculated as follows:

  • image

For each visual function measurement, a positive index value indicates poorer non-dominant eye visual function (stronger inhibitory impact) with partial relative to full occlusion of the dominant eye. The interaction index of each visual function evaluated using the TFA was compared between anisometropes and amblyopes using the repeated measures analysis of variance (ANOVA).

Agreement between shutter goggles and the trial frame was examined for anisometropes and amblyopes, in terms of:

  • 1
    absolute measures of each visual function with partial occlusion and
  • 2
    interaction index for each of these three visual functions. Two methods were used to test for agreement: the paired-samples t-test and the Bland-Altman method.40 The Bland-Altman method includes the following four steps:
  • 1
    The means and the differences between measures using two viewing systems were calculated for each subject,
  • 2
    the differences were plotted as a function of the means,
  • 3
    the mean and standard deviation of the differences were calculated and
  • 4
    the lower and upper ‘limits of agreement’ (LOA) were calculated:
  • image

where

  • image
  • image

n = number of subjects

The trial frame and shutter goggles were considered to show very good agreement if:

  • 1
    ninety per cent of the differences (n = 14) between the goggles and the trial frame were within the two limits of agreement and
  • 2
    the standard deviation of the differences was less than the SD of the results (visual function or interaction index) obtained using the goggles.

They were considered to show lower but acceptable agreement if the SD of the differences was more than the SD of the results obtained using the goggles but equal to or less than 1.96 × SD of the results obtained using the goggles. They were considered to show poor agreement if the SD of the differences was more than 1.96 × SD of the results obtained using the goggles.

FEASIBILITY OF THE SHUTTER GOGGLES AND TRIAL FRAME

The duration of each subject training session, the number of breaks requested by subjects in the main experiments and the willingness to attend further experiments were recorded in 10 subjects from each group to assess the feasibility of the two viewing systems. Based on the children's answers to the question of willingness to attend further experiments, a tick to a viewing system was marked as ‘1’, while a cross to a viewing system was marked as ‘0’. For each test of visual function, these scores, duration of training and the number of breaks requested by subjects were each compared among the three groups of subjects for data obtained using the trial frame and shutter goggles, respectively, using the one-way ANOVA or the Kruskal-Wallis analysis of variance and were each compared between the TFA and shutter goggles using the repeated measures ANOVA or the Wilcoxon signed-rank test.

RESULTS

  1. Top of page
  2. Abstract
  3. METHODS
  4. RESULTS
  5. DISCUSSION
  6. SUMMARY
  7. ACKNOWLEDGEMENT
  8. REFERENCES

Visual functions and interocular interactions in anisometropes and amblyopes

For anisometropes and amblyopes, measures of visual function with partial occlusion using the trial frame and shutter goggles are presented in Figure 3. The ‘interaction indices’ of each visual function evaluated using the TFA and shutter goggles are presented in Figure 4. High inter-individual variation was found in all three visual functions and in interaction index of these visual functions in both anisometropes and amblyopes. The limits of agreement and the number of subjects showing differences beyond these limits and the results of agreement assessment between these two viewing systems (Bland-Altman and t-test methods) are shown in Table 2. The agreement between these two viewing systems in control subjects is discussed elsewhere.14

image

Figure 3. Non-dominant eye acuity (logMAR) [A], contrast sensitivity (dB) [B] and alignment sensitivity (arcmin-1) [C] measured using the trial frame and shutter goggles with partial occlusion on the dominant eye of 14 anisometropic children without amblyopia (anisometropes; solid circles) and 14 anisometropic amblyopic children (amblyopes; open squares)

Download figure to PowerPoint

image

Figure 4. Acuity [A], contrast sensitivity [B] and alignment sensitivity [C] interaction indices evaluated using the trial frame and shutter goggles in the non-dominant eye of 14 anisometropic children without amblyopia (anisometropes; solid circles) and 14 anisometropic amblyopic children (amblyopes; open squares). Index towards the positive y-axis indicates poorer non-dominant eye visual function with partial relative to full occlusion of the dominant eye.

Download figure to PowerPoint

Table 2. Agreement between the two viewing systems. 1. The limits of agreement (LOA) for absolute measures and ‘interaction indices’ measured using the two viewing systems (the trial frame and shutter goggles) in anisometropic children with and without amblyopia. 2. The number of subjects in whom differences between the two viewing systems (the trial frame versus shutter goggles) exceeded the two LOA (outliers) and 3. The agreement between the two viewing systems (Bland-Altman and t-test methods). For the t-test, p > 0.05 indicates no significant difference between measurements using shutter goggles and the trial frame apparatus. Differences between these two viewing systems were found only in contrast sensitivity measurements for anisometropic children without amblyopia.
 Agreement test
LOAOutlierAgreement
Bland–Altmant-test
Anisometropic amblyopic childrenAcuityFunction-0.07 ± 0.501Acceptablep = 0.338
Index-0.02 ± 0.241Acceptablep = 0.644
Contrast sensitivityFunction-0.54 ± 11.810Acceptablep = 0.745
Index0 ± 0.110Goodp = 0.874
Alignment sensitivityFunction-0.09 ± 0.550Goodp = 0.241
Index0.10 ± 0.640Goodp = 0.254
Anisometropic children without amblyopiaAcuityFunction0.04 ± 0.460Acceptablep = 0.540
Index0.05 ± 0.231Goodp = 0.120
Contrast sensitivityFunction4.59 ± 14.900Acceptablep = 0.042
Index-0.02 ± 0.090Goodp = 0.133
Alignment sensitivityFunction0 ± 0.580Acceptablep = 0.954
Index0.05 ± 0.501Acceptablep = 0.495

For both anisometropes and amblyopes, the Bland-Altman method indicated acceptable or very good agreement between shutter goggles and the TFA in all three visual function measures and in their interaction index evaluations. The t-test indicated that in each visual function, interaction index was not significantly different when evaluated using shutter goggles or the trial frame (p > 0.05). Absolute measures of acuity and alignment sensitivity also did not differ significantly between the two viewing systems (p > 0.05). Contrast sensitivity was not significantly different between viewing systems in the amblyopes (p > 0.05) but it was significantly lower (poorer) when measured using the trial frame than using goggles in the anisometropes (p = 0.042).

Acuity interaction index evaluated using the trial frame was significantly higher in amblyopes than in anisometropes (F1,13= 5.000, p = 0.044) but there was no significant difference between these two groups in contrast sensitivity (F1,13= 0.025, p > 0.05) or alignment sensitivity (F1,13= 2.483, p > 0.05) interaction index. The interaction index evaluated using shutter goggles is discussed elsewhere.41

Feasibility of use of the shutter goggles and the trial frame

The training duration of each visual function for each group is presented in Figure 5. No significant difference was found among subject groups in any measurement of visual function, using either the TFA (acuity: χ2[2]= 2.754; contrast sensitivity: χ2[2]= 5.292; alignment sensitivity: F2,27= 0.263; p > 0.05) or shutter goggles (acuity: χ2[2]= 4.055; contrast sensitivity: F2,27= 3.021; alignment sensitivity: F2,27= 0.477; p > 0.05). For all three groups of subjects, the durations of training sessions for acuity (Z = -0.930, p > 0.05) and alignment sensitivity (Z = -0.154, p > 0.05) tests were not significantly different when using the trial frame and the goggles as viewing systems. The training sessions for contrast sensitivity testing were significantly longer using goggles than using the trial frame in all groups of subjects (controls: F1,9= 21.423, p = 0.001; anisometropes: Z = -2.703, p = 0.007; amblyopes: F1,9= 30.146, p < 0.001).

image

Figure 5. Duration of training sessions (minutes) using shutter goggles and the trial frame as viewing systems in acuity, contrast sensitivity and alignment sensitivity testing in normally sighted children (controls) and anisometropic children with (amblyopes) and without (anisometropes) amblyopia. Error bars represent 95 per cent confidence intervals.

Download figure to PowerPoint

The number of breaks requested by each group is shown in Table 3. The number was not significantly different among subject groups, using either the trial frame (acuity: χ2[2]= 1.036; contrast sensitivity: χ2[2]= 0.558; alignment sensitivity: χ2[2]= 0.558; p > 0.05) or shutter goggles (acuity: χ2[2]= 0.180; contrast sensitivity: χ2[2]= 0.201; alignment sensitivity: χ2[2]= 0.330; p > 0.05). The number of breaks requested during alignment sensitivity testing did not differ significantly between the trial frame and goggle viewing systems (Z = -1.265, p > 0.05). Significantly more breaks were requested by controls (Z = -2.070, p = 0.038) and amblyopes (Z = -2.111, p = 0.035) during the acuity test and by all groups of subjects during the contrast sensitivity test (controls: Z = -2.070, p = 0.038; anisometropes: Z = -2.165, p = 0.030; amblyopes: Z = -2.041, p = 0.041) when goggles were used, compared to when the TFA was used.

Table 3. Total numbers of breaks requested by normally sighted children (controls) and anisometropic children with (amblyopes) and without (anisometropes) amblyopia during acuity, contrast sensitivity and alignment sensitivity testing using the trial frame and shutter goggles as viewing systems
 ControlsAnisometropesAmblyopes
AcuityTrial frame011
Shutter goggles878
Contrast sensitivityTrial frame211
Shutter goggles91010
Alignment sensitivityTrial frame112
Shutter goggles323

The total scores indicated by each group of subjects for willingness to attend further sessions using the trial frame and shutter goggles are shown in Table 4. The scores did not differ significantly among subject groups (TFA: χ2[2]= 2.000; goggles: χ2[2]= 1.050; p > 0.05). They were significantly higher for the trial frame than for goggles in anisometropes (Z = -2.236, p = 0.025), indicating that these subjects had significantly greater willingness to attend more experiments using the TFA than using goggles. The scores from controls (Z = -1.633, p > 0.05) and amblyopes (Z = -1.732, p > 0.05) did not differ significantly between the two viewing systems.

Table 4. Total scores from normally sighted children (controls) and anisometropic children with (amblyopes) and without (anisometropes) amblyopia indicating willingness to attend further experiments using the trial frame and shutter goggles
Subject groupTrial frameShutter goggles
Controls95
Anisometropes105
Amblyopes107

DISCUSSION

  1. Top of page
  2. Abstract
  3. METHODS
  4. RESULTS
  5. DISCUSSION
  6. SUMMARY
  7. ACKNOWLEDGEMENT
  8. REFERENCES

Shutter goggles and the trial frame generally show acceptable to good agreement in visual function measures and in interaction index evaluations in anisometropes and amblyopes, with one exception. Contrast sensitivity in anisometropes showed poor agreement between the two viewing systems based on the paired t-test (but acceptable agreement was shown based on the Bland-Altman method). These findings agree with previous work in controls,14 and suggest that the trial frame may be applicable in the evaluation of interocular interaction in children and in investigation of these interactions in amblyopes, in populations where more complex devices cannot be used. It is important to note that the TFA does not have the flexibility of shutter goggles and mirror or prism devices and it may be used to study limited forms of dichoptic masking.

The poor agreement between shutter goggles and the trial frame in contrast sensitivity measurement in anisometropes is consistent with findings in controls.14 As the size of the central occlusion presented to the non-tested eye was 22° to 24° using the trial frame and 3.5° using goggles, it is possible that the size of the central occlusion and the amount of light transmitted from the periphery had a strong impact on contrast sensitivity but a less significant impact on acuity or alignment sensitivity. Perhaps better agreement in contrast sensitivity measurement may be found between these two systems if similar sizes of central occlusion were used.

As shown in Figure 3, it is interesting to note that the acuity of the non-dominant eye measured with partial occlusion did not differ significantly between amblyopes and anisometropes (one-way ANOVA, p > 0.05), which is consistent with previous findings using shutter goggles as a viewing system.41 This may be due to the use of the pinhole in front of the tested eye that resulted in the reduction in retinal illuminance.14,41 The present study found a significantly higher acuity interaction index in amblyopes than in anisometropes (evaluated using the trial frame). This suggests that the extent to which acuity of the non-dominant eye is influenced by stimulation of the dominant eye is greater in anisometropic children with amblyopia than in those without. This finding agrees with previous work using shutter goggles;41 however, the sample size in the present study is small and the statistical power in this sample is only 38 per cent (PS Power and Sample Size Program42). Further experiments with a larger sample may improve understanding of the relationship between interactions and anisometropic amblyopia.

Two amblyopic subjects had negative acuity interaction index (Figure 4), indicating better acuity with partial rather than full occlusion in these subjects. This suggests that the relationship between interocular interaction and anisometropic amblyopia may vary in amblyopic subjects. The reason for this inconsistency is unclear due to the lack of information about the history of the development of amblyopia and about the deficits at subcortical and cortical levels in amblyopic subjects. Similar studies on subjects with a clear history of development of amblyopia and accompanied by objective measures of amblyopia-related deficits would provide a better understanding of this relationship.

The pool of subjects tested here was recruited from a larger group of 106 children.43 While parents of all children gave consent for them to be tested using the TFA, consent for testing with the shutter goggles was received from parents of only 42 of these children. Some of the parents expressed concern that the use of unfamiliar equipment could be harmful to their children's vision. In addition, not all of the 42 children were able to complete the experimental processes using the goggles. Three children in the anisometropes group and one child in the amblyopes group did not complete testing with goggles due to discomfort when wearing the goggles and their parents were not able to persuade them to continue with the tests. Thus, only 38 of the 106 children underwent all tests with both viewing systems, due to difficulties encountered in using the shutter goggles.

In addition to discomfort and parental concerns about the shutter goggles, another complication with this viewing system was that the lenses were occasionally affected by condensation. This complication is related to the fact that the present study was conducted in a warm environment, in high humidity and might not arise in cooler, less humid conditions. This factor may in part explain the differences between the two viewing systems in terms of duration of training, number of breaks and willingness to participate in further test sessions.

SUMMARY

  1. Top of page
  2. Abstract
  3. METHODS
  4. RESULTS
  5. DISCUSSION
  6. SUMMARY
  7. ACKNOWLEDGEMENT
  8. REFERENCES

Our findings indicate acceptable agreement between shutter goggles and the trial frame in acuity, contrast sensitivity and alignment sensitivity testing and in evaluation of a form of interocular interaction for these visual functions, in anisometropic children with and without amblyopia. The trial frame is better accepted by anisometropic children without amblyopia than the shutter goggles. The TFA may be a useful system in studies of interocular interaction in amblyopes, particularly in children.

ACKNOWLEDGEMENT

  1. Top of page
  2. Abstract
  3. METHODS
  4. RESULTS
  5. DISCUSSION
  6. SUMMARY
  7. ACKNOWLEDGEMENT
  8. REFERENCES

The authors thank Jinping Zheng for assistance with data collection.

REFERENCES

  1. Top of page
  2. Abstract
  3. METHODS
  4. RESULTS
  5. DISCUSSION
  6. SUMMARY
  7. ACKNOWLEDGEMENT
  8. REFERENCES