Spatial-Frequency–Dependent Changes in Cortical Activation before and after Patching in Amblyopic Children

Abstract

purpose. To examine the cortical response under transient stimulus conditions in amblyopic children before and after eye patching. To determine whether improvement in acuity is associated with spatial-frequency–dependent changes in specific peaks of the cortical response.

methods. Visual evoked potentials (VEPs) to check reversal (163–18 arc min) and onset of sine wave gratings (0.5–4 cyc/deg) were measured in 24 amblyopic children (<7 years of age) before eye patching. VEPs were repeated in nine subjects with 20/40 or better acuity after patching. Age, severity of amblyopia, and VEP amplitudes of positive peak (P)100, P1, and negative peak (N)2 were analyzed by multivariate statistics.

results. Before patching, the amblyopic eye showed decreasing amplitude with increasing spatial frequencies (P < 0.05) when compared with the nonamblyopic eye. Reduced amplitudes occurred at frequencies well below acuity. Latencies were mildly prolonged. After patching, amplitudes increased in the amblyopic eye across all spatial frequencies (ANCOVA; P < 0.0001 for each peak). However, a spatial-frequency–dependent increase in amplitude was significant only for a late negative peak (N2). The patched eye showed no significant changes.

conclusions. Recovery of acuity after eye patching is associated with an overall increase in cortical activation across a wide range of spatial frequencies below the acuity threshold. A spatial-frequency–dependent increase in a late negative peak suggests that the cortical generator of this peak demonstrates plasticity of acuity recovery.

Functional and anatomic abnormalities in striate cortex underlie the visual deficits of amblyopia. ^{1 2 3 4 5 6 7} Because the visual evoked potential (VEP) is attributed to neural activity in the visual cortex, the VEP is a practical way to measure cortical deficits that occur in anisometropic or strabismic amblyopia in humans. ^{8 9 10 11 12 13 14 15 16 17 18} A predominant finding in studies has been that the amblyopic eye has reduced amplitudes in response to contrast-reversing checks and pattern onset/offset gratings. Latencies are normal or mildly delayed. However, in these studies, the effects of spatial frequency on the amblyopic eye were not systematically examined. The reduced response in the amblyopic eye may be correlated with decreased contrast sensitivity or contrast gain, or deficits in high spatial-frequency mechanisms. To minimize the complex relationship between stimulus contrast and the VEP amplitude, Levi et al. ^{19 20 21} demonstrated spatial-frequency–dependent reductions in amplitude in adult amblyopes under steady state conditions (8 Hz) to high-contrast (44%) gratings. However, the steady state VEP cannot isolate potentially important components that underlie amblyopia, because the response is compressed into a sinusoidal waveform. 

After the amblyopic eye is patched, it can show large increases in VEP amplitude and shortening of latency. ^{11 15 16 17 22 23} Patching has the reverse effect on the nonamblyopic eye, showing amplitude reduction and variable latency increases that can persist for more than 3 months. ^{11 22 23 24} Despite extensive studies, there are limited data on the spatial-frequency–dependent cortical responses in children with isolated amblyopia during the critical period before and after patching. 

In the present study, we assessed the overall cortical response in anisometropic and strabismic amblyopia in children less than 7 years of age, which is well within the critical period. We also used the transient VEP to elucidate spatial-frequency–dependent changes in the cortical response that accompany recovery of acuity after occlusion therapy. 

Methods

Twenty-four children between the ages of 2 and 7 years (mean and median = 4.7; SD = 1.3) underwent full ophthalmic examination including acuity (Allen or Snellen optotypes), cycloplegic refraction, assessment of eye alignment, fundus examination, and stereo acuity, if they were cooperative with the Titmus test. The definition of amblyopia followed modified guidelines of the Amblyopia Treatment Study. ²⁵ Amblyopia was defined as an interocular acuity difference of 3 or more log minimum angle of resolution (logMAR) related to strabismus and/or anisometropia, with an otherwise normal eye examination. Anisometropic amblyopia was defined as amblyopia in the presence of an interocular refractive difference of 0.50 D or greater in spherical equivalent or a 1.50-D or greater difference in astigmatism in any meridian (meridional amblyopia). All anisometropic amblyopes had microtropia, defined as an abnormality of binocularity in which there is a heterotropia less than 10 prism D in conjunction with reduced stereoacuity. 

Strabismus was defined as any heterotropia at distance and/or near fixation of 10 prism D or more. Accommodative esotropia was defined as any strabismus in which horizontal alignment was established to within 10 prism D with spectacles (may include bifocals). Nonaccommodative esotropia was defined as heterotropia of 10 prism D or more with full hyperopic correction and bifocals. Eleven healthy children (2–8 years of age; mean, 5.4) with normal findings in an ophthalmic examination and 20/20 acuity served as control subjects. 

Acuity was measured at a 20-foot distance (B-VAT II Acuity Tester; Mentor, Norwell, MA), with appropriate optical correction determined by cycloplegic refraction. Optotypes were whole-line Allen or Snellen targets, or isolated optotypes if the subject was uncooperative with whole-line testing. Viewing distance was reduced for subjects unable to discriminate the 20/200 optotype and then acuity was scaled appropriately. All acuities were transformed into logMAR (e.g., 0.0 = 20/20, and 1.0 = 20/200). 

All amblyopes were treated with full-time monocular occlusion during waking hours for 1 week per year of age. All testing was performed with the subject wearing full optical correction. None of the subjects had the optical correction changed or had surgery during treatment. Full-time occlusion was repeated up to three times until the best acuity was achieved. Thereafter, part-time patching (4–6 h/d for 4–7 d/wk) was used. Twenty-one subjects returned for their first scheduled visit (mean, 70 days). Three subjects failed to return for follow-up. Visual acuity outcome in the amblyopic eye was the best acuity recorded during an average 2.6-year follow-up. The patch was removed from the occluded eye for at least 3 hours before repeat VEP testing, to avoid short-term occlusion effects. ²⁶  

Stimuli were generated by a calibrated 12-bit video system (Venus; NeuroScientific Corp., Farmingdale, NY; no longer produced) and presented on a monitor that subtended 20° × 20° at 60 cm distance. Subjects wore optical correction and the untested eye was patched. All stimuli were of saturating contrast, ²⁷ to offset the reduced contrast sensitivity of the amblyopic visual system while having little effect on the control eye. ^{19 20 21} Stimuli were contrast-reversing checkerboards (80% contrast, 1.4 Hz, 52 cd/m² mean luminance) of 163, 84, 42, and 18 arc min (spatial frequency was the fundamental frequency along the 45° diagonal). Additional stimuli were pattern-onset of sine wave gratings (99% contrast, 43 cd/m² mean luminance) of 0.5, 1, 2, and 4 cyc/deg presented for 200 ms followed by a blank screen for 800 ms at the same mean luminance. We chose these spatial frequencies because they generate a detectable VEP signal across all severity levels of amblyopia, and they overlap the spatial frequency tuning of cortical cells in monkey models of amblyopia. ⁴ Gratings were oriented horizontally and presented on large fields to minimize fixation and latent nystagmus artifact. ²⁸ Assuming total axial length accounts for most ametropia, ²⁹ retinal magnification alters the stimulus spatial frequency by a factor of 0.87 to 1.44 between the most hyperopic and myopic eyes. 

Preparation of patients for VEPs followed International Society for Clinical Electrophysiology of Vision (ISCEV) guidelines, ³⁰ using Oz as active, Cz as reference, and Pz as ground. Details of VEP recording, maturation, and scoring of waveforms are described elsewhere. ^{31 32 33 34} Briefly, check-reversal amplitude was the difference between the first major positive peak near 100 ms (P100) and the preceding negative peak. Pattern-onset P1 amplitude was the difference between the first major positive peak near 120 ms to the preceding negative peak (or baseline if absent). Pattern-onset negative peak (N)2 amplitude was the difference between the positive peak (P)1 and the major negative peak near 200 ms. Amplitudes were normalized to the largest response from the nonamblyopic eye, which reduce intersubject amplitude variation. ²²  

Statistical software (Systat, Inc., Richmond, CA) was used for all analyses. Analysis of covariance (ANCOVA) compared effects of discrete variables (e.g., amblyopic versus nonamblyopic eye, each spatial frequency), and the interaction of these variables on normalized amplitude after removing the variance associated with age and severity of amblyopia. Outcomes from the ANCOVA were (1) the main effect of eye tested, which indicated an overall change in normalized amplitudes between the amblyopic and nonamblyopic eyes after collapsing the data across spatial frequency and (2) the ANCOVA interaction, which indicated a change in response across spatial frequencies between the amblyopic and the nonamblyopic eyes. The Bonferroni test was used for multiple post hoc comparisons. 

This research adhered to the tenets of the Declaration of Helsinki. Guardians in this study provided informed consent after explanation of the nature and possible consequences of the study. 

Results

Table 1 summarizes the clinical findings. Of the 24 subjects enrolled, 11 had anisometropic amblyopia and 11 had anisometropia combined with accommodative (n = 4), nonaccommodative (n = 4), or combined (n = 3) esotropia. Two subjects had strabismic amblyopia, one with accommodative esotropia and one with non-accommodative esotropia. One (subject 11) had a bilateral medial rectus recession. Stereoacuity was 140 arc sec or worse in subjects who could perform stereo testing. At study entry, all subjects had a VEP to 163′ checks and 0.5 cyc/deg pattern-onset grating; however, because of problems with cooperation, a minimum of 20 subjects provided data for each stimulus spatial frequency. One (subject 4) was eliminated from the multiple stepwise regression analysis because limited data were collected. Analyses pooled all amblyopia types because only two subjects had strabismic amblyopia. 

Figure 1 shows intersubject variation in prepatching VEP waveforms in two representative subjects of similar age and severity of anisometropic amblyopia. The variations in waveforms from the nonamblyopic eye were typical of those in control subjects (95% confidence interval for interocular amplitude ratio, ±15%). As spatial frequency increased, P100 and P1 showed inconsistent changes in amplitude, whereas N2 consistently increased in amplitude and became more tuned. In subject 7 (Fig. 1 , top) the amblyopic right eye showed reduced amplitudes with increasing spatial frequency in response to checks and pattern onset. Latency delays were present but were longer for pattern onset. In comparison, subject 3 (Fig. 1 , bottom) showed little interocular difference in amplitudes in response to checks but significant reductions in P1 and N2 amplitudes at 4 cyc/deg. Latency delays were not present. 

Figure 2 shows normalized amplitudes versus spatial frequency in all subjects (right column). The nonamblyopic eye data suggest some spatial tuning of the N2 in this limited spatial-frequency range. In comparison, the amblyopic eye showed a relative decrease in amplitude with increasing spatial frequency at all peaks. The results in a subset of subjects with 20/150 or better acuity (Fig. 2 , left; n = 12/24) showed that amplitude reductions can occur at spatial frequencies below 4 cyc/deg, which was within their acuity limit. 

Statistical comparison of group data with respect to eye tested, normalized amplitudes, and spatial frequency was performed by ANCOVA. This analysis removes the variance associated with severity of amblyopia and age ³⁵ before comparison. The ANCOVA main effect, which indicates an overall change in normalized amplitudes between the amblyopic and nonamblyopic eyes after data are collapsed across spatial frequency, showed a significant reduction in amplitude of P100, P1, and N2 (F ratio > 15 for each peak; P < 0.0001). The ANCOVA interaction, which tests for a change in response across spatial frequency between eyes, was significant for the P100 (F = 4.0; P = 0.009), P1 (F = 2.6; P = 0.05), and N2 (F = 7.0; P < 0.0001). Severity of amblyopia, but not age, contributed significantly to the variance of normalized amplitudes for all peaks (each F-ratio > 10.0; P < 0.005). 

Latencies in the nonamblyopic eye were within the 95% confidence intervals of control subjects. The difference in interocular latencies between the amblyopic and the nonamblyopic eyes ranged from 1.7 to 18.0 ms. Compared with control subjects, amblyopes, had a significantly longer interocular latency delay of 3.4 ms at 163 arc min, 6.2 ms at 0.5 cyc/deg, and 7.6 ms at 4 cyc/deg (positive peaks; all P < 0.02). Although these interocular latencies are statistically significant, they are within 1 or 2 intervals of the VEP sampling rate (3.5 ms) and therefore have limited validity. 

Figure 3 shows the distribution of acuities before and after patching in 21 patients who returned for follow-up. Before patching, acuities ranged from 20/40 to 20/2000 (log MAR 0.3–2.0, respectively). After patching, 16 (76%) of 21 subjects improved to 20/40 or better acuity in the amblyopic eye. None of the subjects had reverse amblyopia. Of the five subjects with refractory amblyopia, two had unilateral high myopia (>9 D) and the remaining three were older than 6 years. 

Repeat VEPs were recorded in 9 (56%) of the 16 subjects with good acuity outcome (mean logMAR acuity = 0.1). The data set is incomplete because only nine patients consented to retesting. Figure 4 shows pre- and postpatching VEP waveforms from two amblyopic children. The top portion of the figure shows VEPs from an anisometrope with moderate amblyopia (subject 3). After patching, the amblyopic eye improved from 20/100 to 20/30. VEPs in the amblyopic eye showed increases in amplitude at higher spatial frequencies and sharper peak timing. The bottom of Figure 4 shows VEPs in an accommodative esotrope with dense amblyopia (subject 12). After patching, acuity in the amblyopic eye improved from 20/2000 to 20/20. VEP amplitudes increased across all spatial frequencies, more so at higher spatial frequencies. In both subjects, the nonamblyopic eye had mild amplitude variations across a range of spatial frequencies (data not shown). 

Figure 5 shows the average normalized amplitudes in nine subjects before and after patching. In the amblyopic eye, amplitudes increased across all spatial frequencies but were more pronounced at high spatial frequencies in response to check reversal and N2. The ANCOVA test showed all peak amplitudes significantly increased in the amblyopic eye after patching (F = 13.5, 19.9, and 25.8; P < 0.0001; for P100, P1, and N2, respectively). The ANCOVA interaction, which indicated a differential change in response across spatial frequency after patching, was significant for the pattern-onset N2 only (F = 3.6; P < 0.02). Although the ANCOVA interaction for the check stimuli was not significant, a post hoc t-test showed a significant increase in amplitude to the 18-arc min checks (P < 0.05). Of the covariates used (age, severity of amblyopia, and patching duration), only patching duration was a significant source of variance (P100 and P1, F = 4.8 and 5.1, respectively; P < 0.05). Although the nonamblyopic eye showed mild spatial-frequency–dependent changes in amplitude after patching, these were not significant (post hoc test; P > 0.05 for all comparisons). 

After patching average latency in the amblyopic eye decreased 4 to 22 ms depending on spatial frequency, whereas the patched nonamblyopic eye had an increase in average latency by 3 to 15 ms. However, none of these latency changes was significant by ANCOVA, (age, interocular acuity ratio, and duration of patching therapy were covariates). 

Discussion

This prospective study measured the spatial-frequency–dependent VEP cortical response to multiple stimuli before and after patching of amblyopic eyes during the critical period. Before patching, the amblyopic eye showed amplitude reduction that was spatial frequency dependent. In a subset of subjects with 20/40 or better acuity after patching, VEP amplitudes from the amblyopic eye increased across all spatial frequencies, which is consistent with overall strengthening of cortical connections. Furthermore, significant spatial-frequency–dependent increases in cortical activation were found only for a late negative peak (N2). This finding suggests that the cortical generators of N2 reflect the acuity deficit in amblyopia and the acuity recovery after patching. 

There is an overall reduction in VEP amplitude from the amblyopic eye across a range of spatial frequencies. Functional MRI in adult amblyopes also shows reduced activation of striate cortex by the amblyopic eye. ^{5 6 7} The reduction in VEP amplitude could arise from a shift in ocular dominance similar to that observed in animal studies of monocular deprivation. ^{36 37 38 39} In these studies, there is shrinkage of ocular dominance columns serving the deprived eye. However, in anisometropic amblyopia, ocular dominance columns appear normal, suggesting the anatomic changes that underlie reduced activation are more subtle. ^{40 41}  

Cortical activation showed differential losses across the range of spatial frequencies tested, suggestive of a shift in peak response. The disproportionate loss at higher spatial frequencies could arise from an increase in the size of receptive fields representing the fovea. Cortical receptive field sizes can dynamically change depending on the relative synaptic strength of vertical (geniculocortical) and horizontal (intercortical) connections. The long-range horizontal connections provide subthreshold excitation beyond the “classic” receptive field. When vertical inputs are reduced, the relative contribution of the horizontal inputs is strengthened, resulting in a receptive field size increase up to fourfold. ⁴² Therefore, the weaker vertical inputs from the amblyopic eye may unmask the horizontal connections, which would expand receptive field sizes in the central visual field. Expanded receptive fields would pool information over a larger area but compromise spatial resolution. Similar abnormalities of spatial integration may account for the increased spatial summation, crowding phenomenon, and spatial distortion that are also observed in amblyopia. 

Amblyopic eye amplitudes significantly increased at all peaks after patching (independent of age, severity of amblyopia, and patching duration). However, even when acuity recovered to 20/20 residual VEP abnormalities were still present, suggesting there are additional visual cortical deficits. Because the early VEP peaks probably arise from excitatory postsynaptic potentials in striate and extrastriate cortex, ^{43 44 45 46 47} the rapid amplitude improvement suggests strengthening of synaptic connections between visual inputs from the amblyopic eye and these cortical areas. Animal studies of reverse lid-suture deprivation (analogous to patching therapy) during the critical period support this notion. That is, cortical ocular dominance histograms shift back toward the previously deprived eye with partial regrowth of geniculocortical arbors in layer IV if deprivation and reverse lid suture occur during the critical period. ^{38 39 48} Therefore, the VEP can be a useful tool for studying cortical plasticity in humans during the critical period of visual development, which can extend to 10 years of age. ^{35 49 50} Although these comparisons to the monocular deprivation model are appealing, anatomic, and functional changes in anisometropic and strabismic amblyopic monkeys or humans are more subtle. ^{40 41}  

The selective loss of N2 at higher spatial frequencies before patching suggests that the cortical generators of this peak can underlie the acuity deficit in amblyopia. The increase of this peak after patching further suggests the cortical generators of N2 are important in cortical plasticity and acuity recovery. Although few studies describe the late negative VEP peak in amblyopia ^{9 17 31} a late negative peak response to higher spatial frequencies appears at 5 months of age and becomes progressively tuned during the critical period. ^{31 32} During the critical period, there is an age-related decrease in N2 amplitude in control subjects (4 cyc/deg; r ² = 0.51; P = 0.0006; personal observation, 2003). N2 diminishes and the waveform is dominated by a prominent early negative peak. ^{32 51} Therefore, the rapid onset of postpatching increases in N2 are unlikely to be related to maturation. The cortical generator of N2 is not known but source localization studies of late negative peaks implicate extrastriate cortex and visual association cortex. ^{46 52}  

Reverse amblyopia was not observed in this study despite a full-time patching regimen. On average, the VEP did not show significant reductions in amplitude or changes in latency in the nonamblyopic eye after patching. However, individual subjects showed amplitude reductions in the nonamblyopic eye after patching despite 20/20 acuity, suggesting functional changes in cortical activation in the patched eye that are not reflected by acuity. 

Amblyopes are known to have reduced accommodation that may improve along with visual acuity after patching across a wide range of spatial frequencies. ⁵³ Optical defocus due to accommodative inaccuracy could decrease the VEP amplitude. However, the influence of accommodative inaccuracy was unlikely, owing to the use of high-contrast stimuli and the 60-cm viewing distance. The data of Ciuffreda and Rumpf ⁵³ show that accommodation plateaus at the level of tonic accommodation (1–1.5 D), which is similar to the fixation distance in this study. 

 Supported by the William O. Rogers Trust fund awarded to Children’s Hospital and Regional Medical Center (Seattle, WA) and by an unrestricted grant from Research to Prevent Blindness to the University of Washington.

 Submitted for publication March 29, 2004, revised June 1, 2004; accepted June 21, 2004.

 Disclosure: A.H. Weiss, None; J.P. Kelly, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Table 1.

 

Summary of Patient Data

Subject	Age/Sex (y)	Diagnosis	Eye	Refraction	Initial Acuity	Follow-up Acuity*	Age at Last Follow-up
1	6.3/F	A, MER	R	+1.50, +1.75 at 90	20/20	20/20	8.7
			L	+0.25, +6.50 at 90	20/40C	20/25C
2	5.3/M	A, ET	R	+7.50 +0.50 × 75	20/100	20/60	7.8
			L	+3.00, sphere	20/20	20/20
3	5.3/F	A	R	+3.00, +0.25 × 90	20/20	20/20	6.7
			L	+6.00, sphere	20/100	20/30
4	3.8/M	A	R	−10.50 + 0.50 × 30	20/200	20/500	8.2
			L	−0.50 + 1.00 × 180	20/25	20/25
5	4.5/F	A, S	R	+1.00, sphere	20/20	20/20	4.6
			L	+1.75, +0.75, × 90	20/400	20/40
6	5.1/M	A, ET	R	+3.50, sphere, add 2.50	20/25	20/20	6.0
			L	+4.50, sphere, add 2.50	20/400	20/25
7	6.5/M	A	R	+6.50 +0.50 × 90	20/200	20/40	9.1
			L	+1.5, sphere	20/30	20/20
8	6.4/F	A	R	−0.25, +0.75 × 90	20/25	20/25	6.7
			L	−9.50, +0.50 × 65	20/333	20/100
9	3.9/M	A	R	+2.00, sphere	20/20	20/20	4.7
			L	+5.50, sphere	20/100	20/30
10	6.8/M	A	R	+6.50, sphere	20/60	NA	NA
			L	+3.50, sphere	20/25	NA
11	4.9/F	A, S, †	R	+2.00, +0.50 × 90	20/30	NA	NA
			L	+4.00, +0.50 × 105	20/666	NA
12	3.6/F	A, ET	R	+4.00, sphere	20/30	20/20	6.4
			L	+8.25, sphere	20/2000	20/20
13	6.9/M	A, ET	R	+8.00 +0.25 × 30	20/200	20/100	9.4
			L	+7.00 +0.50 × 135	20/25	20/20
14	4.8/F	A, MER	R	+1.00, +2.50 × 90	20/50	20/20	7.5
			L	+1.50, sphere	20/20	20/20
15	4.1/M	A	R	+3.00 +0.25 × 90	20/50	20/25	4.2
			L	+2.50 + 0.25 × 90	20/20	20/20
16	2.6/F	A, S	R	+5.00 + 1.00 × 95	20/2000	20/30	4.2
			L	+4.00 +0.50 × 90	20/30	20/25
17	3.5/M	ET	R	+7.50 +0.75 × 105	20/50	20/30	6.3
			L	+7.50 +0.25 × 75	20/30	20/20
18	3.8/M	S	R	+3.50 +0.50 × 90	20/30	20/20	5.1
			L	+3.50 +0.50 × 90	20/667	20/25
19	5.3/F	A	R	+2.75 +0.75 × 90	20/20	20/20	7.1
			L	+6.50 +1.50 × 105	20/50	20/30
20	6.1/F	A	R	+3.50, +1.75 × 90	20/80	20/70	6.2
			L	+0.75, +0.25 × 90	20/25	20/25
21	4.5/M	A, S, ET	R	+6.00, +0.75 × 75	20/25	20/20	7.9
			L	+6.50, +0.75 × 105	20/40	20/20
22	2/M	A, S	R	+3.00, sphere	20/30	20/20	4.4
			L	+6.00, +0.25 × 75	20/100	20/25
23	3.7/F	A, S, ET	R	+1.00, +0.75, × 070	20/40	20/20	6.3
			L	+2.25, +0.50, × 110	20/400	20/20
24	4.2/F	A, S, ET	R	+5.50, +0.50 × 90	20/70	NA	NA
			L	+4.00, +0.50 × 90	20/30	NA

 A, anisometropia; S, non-accommodative esotropia; ET, accommodative esotropia; MER, meridonal amblyopia. All anisometropes were orthotropic or had a microtropia.

*  Best acuity across follow-up.

†  Bilateral recession medial recti.

Figure 1.

 

VEPs from two anisometropic amblyopes of similar age. Stimuli were contrast-reversing checkerboards (left) and pattern onset of a sine wave grating (right). Dotted and solid traces: amblyopic and nonamblyopic eyes, respectively. Peaks scored in this study are noted on sample waveforms. Acuities in the right (RE) and left (LE) eyes are shown at left for each subject. Check sizes are in arc minutes (′), and gratings are in spatial frequencies (cycles per degree; c/d).

Figure 2.

 

Normalized amplitudes plotted versus spatial frequency for all peaks. Each VEP peak is sorted into three rows. Open and filled symbols: amblyopic and nonamblyopic eyes, respectively. Left: subset of subjects with 20/150 or better acuity (n = 12); right: average of all subjects (n = 24). *Significant differences between amblyopic and nonamblyopic eyes by paired t-test (P < 0.05). Spatial frequency for checkerboards is the dominant frequency along the 45° diagonal.

Figure 3.

 

Changes in logMAR acuity before and after patching therapy in 21 amblyopes. Average duration of patching was 108 days.

Figure 4.

 

VEPs in the amblyopic eye of a child with anisometropia (subject 3, top) and the amblyopic eye of a child with accommodative esotropia (subject 12, bottom). Thin and thick traces: before and after patching, respectively. Acuity before and after patching is noted to the left. Stimuli were contrast-reversing checkerboards (left) and pattern-onset sine wave gratings (right).

Figure 5.

 

Normalized VEP amplitudes in nine subjects with good visual outcome after patching. Data for each peak are shown across rows. Amplitudes are shown before and after patching. Spatial frequency for checkerboards is the dominant frequency along the 45° diagonal.