Cortical Reorganization After Optical Alignment in Strabismic Patients Outside of Critical Period

Abstract

Purpose: To measure visual crowding, an essential bottleneck on object recognition and reliable psychophysical index of cortex organization, in older children and adults with horizontal concomitant strabismus before and after strabismus surgery.

Methods: Using real-time eye tracking to ensure gaze-contingent display, we examined the peripheral visual crowding effects in older children and adults with horizontal concomitant strabismus but without amblyopia before and after strabismus surgery. Patients were asked to discriminate the orientation of the central tumbling E target letter with flankers arranged along the radial or tangential axis in the nasal or temporal hemifield at different eccentricities (5° or 10°). The critical spacing value, which is the minimum space between the target and the flankers required for correct discrimination, was obtained for comparisons before and after strabismus surgery.

Results: Twelve individuals with exotropia (6 males, 21.75 ± 7.29 years, mean ± SD) and 15 individuals with esotropia (6 males, 24.13 ± 5.96 years) participated in this study. We found that strabismic individuals showed significantly larger critical spacing with nasotemporal asymmetry along the radial axis that related to the strabismus pattern, with exotropes exhibiting stronger temporal field crowding and esotropes exhibiting stronger nasal field crowding before surgical alignment. After surgery, the critical spacing was reduced and rebalanced between the nasal and temporal hemifields. Furthermore, the postoperative recovery of stereopsis was associated with the extent of nasotemporal balance of critical spacing.

Conclusions: We find that optical realignment (i.e., strabismus surgery) can normalize the enlarged visual crowding effects, a reliable psychophysical index of cortical organization, in the peripheral visual field of older children and adults with strabismus and rebalance the nasotemporal asymmetry of crowding, promoting the recovery of postoperative stereopsis. Our results indicated a potential of experience-dependent cortical organization after axial alignment even for individuals who are out of the critical period of visual development, illuminating the capacity and limitations of optics on sensory plasticity and emphasizing the importance of ocular correction for clinical practice.

How we “see” the world is one of the fundamental challenges in brain science, understanding of which requires insights into both optical and cortical elements involved in visual perception. Setting fundamental constraints over retinal image quality, the eye's optics drives normal maturation of visual cortical processing.^1,2 For binocular creatures, one of the characteristic optics is that we are long-adapted to a two-eye visual mode in which two eyes coordinate in an aligned fashion to form a unifying three-dimensional representation of the external world.³ If the visual axes of two eyes are misaligned, retinal correspondence is disrupted, and strabismus may develop.⁴ Without timely and appropriate treatment, strabismus may lead to functional changes and abnormal functional connectivity of the visual cortex,^5–7 resulting in a variety of visual deficits (e.g., defective stereopsis, increased fixation instability, interocular suppression, abnormal retinal correspondence, and even amblyopia).^8–10 Following the doctrine that cortical plasticity will diminish after the critical period, the common clinical practice for strabismus is thus to correct the visual axis in early childhood.¹¹ By measuring visual crowding, a reliable psychophysical index of cortical organization, here we show that surgical alignment of the visual axis in older children and adults with strabismus but without amblyopia also leads to cortical reorganization, adding new insights on the capacity and limits of the eye's optics on sensory plasticity. 

Table.

 

Clinical Characteristics of Strabismic Individuals

 

Visual crowding characterizes humans’ inefficiency in identifying objects in clutter but cannot simply be explained by reduced visual acuity or contrast sensitivity at a given retinal location.^12–15 The magnitude of the crowding effects, usually indexed by critical spacing (i.e., the minimum spacing between a target and flankers required for reliable target identification), expands with increasing retinal eccentricity.^16–19 It has been reported that visual crowding reflects the neuroanatomic limits of cortical processing of object recognition,^14,20 with the extent of critical spacing negatively correlated with the cortical magnification factor.^21–24 Psychophysical measurement of visual crowding has thus been applied to characterize cortical plasticity following long-term adaptation to retinal lesion or central vision deprivation such as age-related macular degeneration,¹⁷ glaucoma,^25,26 and amblyopia.²⁷ On the other hand, as the common way to restore ocular alignment, empirical observations have shown that even in adults, strabismus surgery may also lead to the improvement or recovery of binocular functions.^28–31 One possible conjecture is that restoring coordination between the signals from the two eyes probably results in cortical reorganization of intracortical neuronal connections.³² 

In the present study, we used real-time eye tracking to ensure gaze-contingent display of the target and flankers at four different retinal locations and measured the critical spacing value in 12 exotropes and 15 esotropes before and after surgical operation to assess whether axial alignment could induce cortical reorganization in older children and adults with strabismus. We found that before surgery, strabismic individuals manifested significantly larger critical spacing with field- and axis-specific patterns along only the radial axis, with exotropes exhibiting stronger temporal hemifield crowding and esotropes exhibiting stronger nasal hemifield crowding, which is consistent with our previous study,³³ even though the individuals of these two studies were not from the same group. Interestingly, the critical spacing was reduced and rebalanced between the nasal and the temporal hemifields 3 months after surgical correction of ocular misalignment, which correlated with postoperative stereopsis recovery. Overall, our results revealed that the visual cortex could be reorganized in response to correction of visual alignment even in humans outside of the typical critical period. 

Methods

Patients

Individuals with horizontal concomitant strabismus but without amblyopia were recruited through the Zhongshan Ophthalmic Center, Guangzhou, China. Twelve individuals with exotropia (6 men, 21.75 ± 7.29 years, mean ± SD) and 15 individuals with esotropia (6 men, 24.13 ± 5.96 years) participated in this study. This study followed the principles of the Declaration of Helsinki and was approved by the Institutional Review Board of Zhongshan Ophthalmic Center, Sun Yat-sen University. After explaining the purpose, procedures, risks, and benefits of this study, informed consents were obtained from patients or their parents/legal guardians. 

All individuals underwent comprehensive eye examinations before and 3 months after strabismus surgery, including a review of medical history, cycloplegic refraction and determination of the best-corrected distance visual acuity, examination of eye movements, alternate prism cover tests at near and distance fixations, slit-lamp examination, funduscopic examination, and stereoacuity test (Vision Assessment Corporation Co., Elk Grove Village, IL, USA). 

Horizontal concomitant strabismus is defined as a horizontal deviation of the visual axis and an identical angle of deviation of the squinting eye relative to the other eye, regardless of the direction of the gaze.^34–36 All individuals reported alternating strabismus with shifted fixation between their eyes and ocular deviation at both near and distant. According to the Preferred Practice Patterns of the American Academy of Ophthalmology,³⁷ individuals who had clinically significant refractive error were instructed to wear spectacles for at least 3 months before the experiment, and only those who had normal or corrected-to-normal visual acuity (0.00 logMAR or better) after refractive correction were included in this study. All individuals had a history of strabismus since early childhood (3.33 ± 3.56 years old, mean ± SD) and had never received strabismus surgery before. All individuals had no history of vertical, paralytic, or restrictive strabismus; accommodative esotropia; acute concomitant esotropia; nystagmus; or a history of ocular surgery or prism correction. Individuals with any neurologic disorders or systemic diseases were also excluded. 

All individuals underwent strabismus surgery (rectus shortening and/or recession) to recover axial alignment after the preoperative measurement, and a postoperative measurement was performed 3 months after the surgery. Successful postoperative axial alignment is considered less than 10 prism diopters.³⁴ Clinical characteristics of each individual are summarized in Table. 

Apparatus

Experimental stimuli were programmed with MATLAB R2018a (MathWorks, Inc., Natick, MA, USA) using the Psychtoolbox extension (Version 3.0.14).³⁸ All stimuli were presented on a gamma-corrected liquid crystal display screen with a resolution of 1920 × 1080, a refresh rate of 144 Hz, and a uniform gray background of 54 cd/m². Patients were seated at an eye-to-screen distance of 57 cm with a chin-and-forehead rest to limit head movements. 

Stimuli

The stimuli consisted of a single black tumbling E in the visual acuity test and a trigram of black tumbling Es with the middle one as the target in the crowding test. Patients were asked to report the orientation (left, right, up, or down) of the target tumbling E by pressing the appropriate key on the keyboard. The trigrams were arranged along the radial or tangential axis at an eccentricity of 5° or 10° in the nasal or temporal hemifield, resulting in eight testing conditions appearing randomly (see Fig. 1A). 

Figure 1.

 

Schematic description of the visual crowding test. (A) Stimuli used in the visual crowding test. The stimuli consisted of a target tumbling E flanked by two tumbling Es arranged on both sides of the target along the radial or tangential axis. Critical spacing value was measured in the nasal or temporal hemifield at the eccentricities of 5° or 10°, respectively, leading to a total of eight kinds of testing stimuli (2 axes × 2 visual fields × 2 eccentricities). (B) Time course of the crowding test. After a standard five-point calibration and validation, the trial began with a central red fixation point presenting throughout the test. The stimuli appeared randomly for 250 ms, followed by an interstimulus interval for patients to respond without time limitation. After the response, there was a brief pause of 1 second and then the next trial started. A three-down, one-up staircase was adopted to track the spacing between the center of the target and the center of the flankers. (C) Eye recording–based gaze-contingent display. A video-based eye tracker was used to continuously monitor patients’ eye positions, with a gaze-contingent paradigm to ensure the display of the target and flankers in the specific visual field locations. Once the eye tracker detected patients’ involuntary eye movement toward the target letter, the program would move the target and flank letters at an identical direction and distance to offset the involuntary eye movement.

Eye Movement Recording

A high-resolution infrared-emitting video-based eye tracker sampled at 1000 Hz (EyeLink 1000; SR Research, Ottawa, Canada) was used, with a maximum spatial resolution of 0.02°, to continuously monitor patients’ eye positions. We performed and repeated a standard five-point calibration and validation sequence at the beginning of every test until the validation error was smaller than 1° on average.³⁹ The average gaze error was 0.5°, ranging from 0.1° to 1°. An eye recording–based gaze-contingent display was used to ensure the display of the target and flankers in the specific visual field locations (see Fig. 1C). Once the eye tracker detected the involuntary eye movement of the patient to the target letter, the program would move the target and flank letters at the identical direction and distance to offset the involuntary eye movement. During the whole test, a chin-and-forehead rest was used to minimize head movements and trial-to-trial variability in the estimate of gaze position. A good fixation was defined as a stable fixation for at least 250 ms and positions not exceeding 0.5° in any direction.³⁹ Trials that failed to meet the requirements of good fixation were repeated until the requirements were met. 

Design

The test was monocular with the untested eye covered by a black patch. To isolate the effect of crowding from that of simply missing the target, we conducted a visual acuity test at first to obtain the minimum letter size visible to the patients and ensured that they could see the target in isolation clearly. In the following crowding test, the initial letter size was set to 1.5 times the minimum letter size estimated from the visual acuity test, and the initial spacing between the center of the target and the center of the flankers was 0.75 times the eccentricity.⁴⁰ Critical spacing of crowding across the radial or tangential axis in the nasal or temporal visual fields was measured at two eccentricities (5° and 10°), leading to a total of eight crowding measurements (2 axes × 2 visual fields × 2 eccentricities). All individuals completed tests with their nondominant eye (determined with the Mile's test)^41,42 in both the preoperation and the postoperation measurements. 

Procedure

After completing a standard five-point calibration and validation⁴³ to provide the horizontal and vertical position measurement of the test eye, patients were instructed to fixate on the central red point. Stimuli were briefly presented for 250 ms, and eight different testing conditions were randomly interleaved. Patients were asked to identify the orientation of the target tumbling E by pressing the corresponding key on a computer keyboard. After each response, there was a short pause of 1 second, and then the next trial began (see Fig. 1B). The distance between the center of the target and the center of the flankers was tracked using a three-down, one-up staircase with a 10% step size,^44,45 in which three consecutive correct discriminations of the orientation of the target tumbling E led to a decrease of the center-to-center distance between the target and the flankers by 10% (i.e., Distance_t+1 = Distance_t × 90%), while an incorrect discrimination led to a 10% increase in the distance (i.e., Distance_t+1 = Distance_t × 110%). After six reversals, the staircase procedure ended and the average spacing of the last four reversals was regarded as the critical spacing value (i.e., the space between the center of the target and the center of the flankers required for 79.4% correct identification). The same methods of staircase and reversal were used to calculate the visual acuity threshold in the visual acuity test. 

Before the test, the patients were explained the task and given an opportunity to practice for 10 minutes. Patients who did not complete the test well or who did not maintain good fixation stability during the test were excluded. We recruited a total of 15 exotropes and 15 esotropes, and 3 exotropes were excluded due to poor fixation. As a result, 12 exotropes and 15 esotropes successfully completed both the preoperative and postoperative measurements and were included in the final analysis. 

Statistical Analysis

To reduce the influence of eccentricity,⁴⁶ we first normalized the critical spacing by dividing these values by the measured eccentricity to obtain the normalized critical spacing (NCS). A mixed repeated-measures ANOVA with axis (radial or tangential), eccentricity (5° or 10°), visual field (nasal or temporal), and operation (preoperative or postoperative) as the within-subject factors and group (exotropia or esotropia) as the between-subject factor was carried out. Since we found that the NCS of each group was comparable in both eccentricities (5° and 10°; see Supplementary Table S1), we combined the NCS measured at the same visual field across two eccentricities for the following analysis. 

The combined NCS was analyzed by repeated-measures ANOVA, with within-subject factors of axis (radial or tangential), visual field (nasal or temporal), and operation (preoperative or postoperative) for exotropia and esotropia groups separately. The pairwise post hoc comparisons (with Bonferroni correction) were performed to compare the NCS before and after the operation in each group. 

Horizontal concomitant strabismus turns its visual axis to the nasal (e.g., esotropia) or temporal (e.g., exotropia) field of vision. In order to better describe the crowding effects across different visual fields, we calculated the ratio of the NCS between the nasal and the temporal hemifields (or critical spacing ratio, CSR) as (NCS_nas – NCS_temp) / (NCS_nas + NCS_temp). CSR values of <0, 0, and >0 indicate greater crowding effects in the temporal hemifield, symmetrical crowding effects across visual fields, and greater crowding effects in the nasal hemifield, respectively.⁴⁷ The CSR was analyzed by repeated-measures ANOVA with within-subject factors of axis (radial or tangential) and operation (preoperative or postoperative) and the pairwise post hoc comparisons (with Bonferroni correction) to examine whether the CSR differed significantly before and after strabismus surgery in each group. In addition, a single-sample t-test was conducted to examine whether the CSR was significantly different from zero. 

To figure out whether strabismus surgery could improve the patient's fixation stability and thus reduce the visual crowding effects, we analyzed the eye positions using MATLAB (MathWorks) and quantified the fixation stability by calculating the bivariate contour ellipse area (BCEA) using the following equation^48–50:  

\begin{eqnarray*}{{\bf BCEA}} = \ {\boldsymbol{\pi }}{{{\bf X}}^2}{{\boldsymbol{\sigma }}_{\boldsymbol{x}}}{{\boldsymbol{\sigma }}_{\boldsymbol{y}}}\sqrt {1 - {{\boldsymbol{p}}^2}} \end{eqnarray*}

where X² is a chi-square variable with two degrees of freedom; σ_x and σ_y are the standard deviation of eye position in the horizontal and vertical meridians, respectively; and p is the correlation coefficient of the moment of the Pearson product of the horizontal and vertical eye positions. The area of 68% BCEA was used in this study as a quantitative measure of fixation instability. It refers to the area in squared degrees (deg²) of the ellipse around the stimulus containing most of the fixation points. Therefore, the smaller the BCEA value is, the more stable the gaze is. A log₁₀ transformation was used to normalize the resulting BCEA values (logBCEA). We also analyzed the distribution of fixation points in the nasal and temporal visual fields during the test to explore the influence of oculomotor bias on the detection of crowding effects. 

Data are reported as mean ± SD unless otherwise noted. All statistical analyses were performed using IBM SPSS Statistics Version 25 (IBM, Armonk, NY, USA) and JASP 0.14.1.0 (https://jasp-stats.org), with P < 0.05 as the criterion for statistical significance. 

Results

With comprehensive eye examinations, 12 individuals with exotropia (6 males, 21.75 ± 7.29 years, mean ± SD) and 15 individuals with esotropia (6 males, 24.13 ± 5.96 years) were recruited and psychophysically tested visual crowding before and 3 months after strabismus surgery. Using a real-time eye tracker and gaze-contingent display paradigm, the crowding test was performed to measure the critical spacing (i.e., the center-to-center spacing between the center target and flankers that corresponds to 79.4% correct identification) across radial and tangential axes in the nasal and temporal visual fields at two eccentricities (5° and 10°), leading to a total of eight crowding measurements (2 axes × 2 visual fields × 2 eccentricities). 

To reduce the influence of eccentricity,⁴⁶ we first normalized the critical spacing by dividing the values by the eccentricity at which they were measured to get the NCS. A mixed repeated-measures ANOVA revealed that the NCS varied significantly with surgical operation (F_{1, 25} = 24.51; P < 0.001) and axis (F_{1, 25} = 253.07; P < 0.001) but not hemifield (F_{1, 25} = 0.92; P = 0.346), eccentricities (F_{1, 25} = 1.64; P = 0.213), and groups (F_{1, 25} = 0.14; P = 0.716; see Supplementary Table S1 for details). Since the interaction of eccentricity and group was also nonsignificant (F_{1, 25} = 2.12; P = 0.158), our results indicated that the NCS was comparable across the two eccentricities in esotropia and exotropia groups. Mean NCS across eccentricities (i.e., average of the data of 5° and 10° eccentricities in the same hemifield locations) was thus used in the following analysis. 

Peripheral Visual Crowding Effects in Older Children and Adults With Exotropia Before and After Surgical Alignment

Crowding zones mapped from the NCS across nasal and temporal hemifields for 12 exotropes (EXO) before and after surgery are plotted in Figure 2. Before surgery, the average NCS of exotropes was 0.43 ± 0.09 at the radial axis and 0.21 ± 0.08 at the tangential axis in the nasal hemifield, as well as 0.58 ± 0.18 and 0.32 ± 0.26, respectively, in the temporal hemifield. After surgical alignment, the NCS decreased to 0.35 ± 0.11 at the radial axis and 0.17 ± 0.07 at the tangential axis in the nasal hemifield, as well as 0.38 ± 0.13 at the radial axis and 0.17 ± 0.06 at the tangential axis in the temporal hemifield. When a repeated-measures ANOVA was performed, the effects of surgical operation (F_{1, 11} = 13.94; P = 0.003), hemifield (F_{1, 11} = 6.77; P = 0.025), and axis (F_{1, 11} = 214.58; P < 0.001; see Supplementary Table S2 and Supplementary Table S4 for details) were significant. A pairwise post hoc comparison (with Bonferroni correction) showed a significant difference in the NCS before and after operation in the temporal hemifield (radial: t = 4.80, P < 0.001; tangential: t = 3.58, P = 0.038) but not in the nasal hemifield (radial: t = 1.91, P = 1.00; tangential: t = 0.99, P = 1.00). Moreover, the NCS along the radial axis was significantly larger in the temporal hemifield, as opposed to the nasal hemifield before the surgery (nasal versus temporal: t = 3.90, P = 0.015). However, no significant difference in the NCS between the nasal and temporal hemifields along the radial axis was found after surgery (t = 0.79, P = 1.00; Fig. 4A). 

Figure 2.

 

Crowding zones mapped from the normalized critical spacing across the nasal and temporal hemifields for 12 exotropes before and after surgery. For each hemifield, the size of the crowding zone was shown for radial and tangential axes. The solid red line represents the crowding zone in the nasal hemifield before surgery. The solid blue line represents the crowding zone in the temporal hemifield before surgery. The dotted red line represents the crowding zone in the nasal hemifield after surgery. The dotted blue line represents the crowding zone in the temporal hemifield after surgery (see color legend). (A) The crowding zone of each exotropia individual. (B) The average crowding zone of the exotropes groups (n = 12). The light shaded regions indicate the SEM across individuals.

In short, similar to our previous study, in which nasotemporal symmetry crowding effects of normal individuals and nasotemporal asymmetry crowding effects of strabismic patients were reported,³³ we found that exotropes manifested stronger peripheral visual crowding effects in the temporal hemifield (i.e., the deviating direction of fixation), and surgical correction of ocular position significantly reduced and rebalanced the crowding effects. 

Peripheral Visual Crowding Effects in Older Children and Adults With Esotropia Before and After Surgical Alignment

Crowding zones mapped from the NCS across the nasal and temporal hemifields for 15 esotropes (ESO) before and after surgery are plotted in Figure 3. Before surgery, the average NCS of esotropes was 0.54 ± 0.17 at the radial axis and 0.21 ± 0.12 at the tangential axis in the nasal hemifield, as well as 0.43 ± 0.17 and 0.19 ± 0.06, respectively, in the temporal hemifield. After surgical alignment, the NCS reduced to 0.41 ± 0.11 at the radial axis and 0.18 ± 0.05 at the tangential axis in the nasal hemifield, as well as 0.39 ± 0.08 at the radial axis and 0.16 ± 0.04 at the tangential axis in the temporal hemifield. A repeated-measures ANOVA revealed that the effects of surgical operation (F_{1, 14} = 8.96; P = 0.010), hemifield (F_{1, 14} = 8.77; P = 0.010), and axis (F_{1, 14} = 114.29; P < 0.001; see Supplementary Table S3 and Supplementary Table S5 for details) were significant. A pairwise post hoc comparison (with Bonferroni correction) showed the NCS after surgery was significantly decreased compared with the corresponding NCS before surgery in the nasal hemifield along the radial axis (t = 5.04, P < 0.001) but not along the tangential axis (t = 0.96, P = 1.00) or in the temporal hemifield (radial axis: t = 1.28, P = 1.00; tangential axis: t = 0.95, P = 1.00). Besides, consistent with our previous report,³³ the NCS in the nasal hemifield was significantly larger than that in the temporal hemifield along only the radial axis before surgery (nasal versus temporal: t = 5.55, P < 0.001), but no difference of the NCS was found between the nasal and temporal hemifields after surgery (radial axis: t = 0.72, P = 1.00; tangential axis: t = 0.91, P = 1.00; Fig. 4B). 

Figure 3.

 

Crowding zones mapped from the normalized critical spacing across the nasal and temporal hemifields for 15 esotropes before and after surgery. For each hemifield, the size of crowding zone was shown for radial and tangential axes. The solid red line represents the crowding zone in the nasal hemifield before surgery. The solid blue line represents the crowding zone in the temporal hemifield before surgery. The dotted red line represents the crowding zone in the nasal hemifield after surgery. The dotted blue line represents the crowding zone in the temporal hemifield after surgery (see color legend). (A) The crowding zone of each esotropia individual. (B) The average crowding zone of the esotropes groups (n = 15). The light shaded regions indicate the SEM across individuals.

Figure 4.

 

(A) The normalized critical spacing across different hemifields and axes for exotropes before and after surgery. (B) The normalized critical spacing across different hemifields and axes for esotropes before and after surgery. Higher values on the y-axes signify stronger crowding effects. Statistically significant difference: *P < 0.05, ***P < 0.001.

In general, we found that esotropes manifested stronger peripheral visual crowding effects in the nasal hemifield (i.e., the deviating direction of fixation), which could also be significantly reduced and became relatively nasotemporally balanced after surgical correction. 

Nasotemporal Rebalance in Visual Crowding Effects After Axial Alignment

Our results found peripheral visual crowding exhibited nasotemporal asymmetry in strabismic individuals but diminished and rebalanced after ocular alignment. To better characterize the surprising change of hemifield-specific crowding effects in older children and adults with strabismus, we calculated the ratio of the NCS between the nasal and the temporal hemifields (or CSR; see Methods). On average, the CSR was −0.14 ± 0.02 (mean ± SEM) at the radial axis and −0.12 ± 0.06 at the tangential axis for exotropes, as well as 0.13 ± 0.03 and 0.03 ± 0.04, respectively, for esotropes before surgery. After axial alignment, the corresponding value was −0.03 ± 0.04 and −0.01 ± 0.04 for exotropes and 0.01 ± 0.03 and 0.05 ± 0.04 for esotropes. In the exotropia group, the effect of surgical operation (F_{1, 8} = 6.35; P = 0.029) was significant, but the effect of the axis was not significant (F_{1, 8} = 0.68; P = 0.427) when a repeated-measures ANOVA was performed. A pairwise post hoc comparison (with Bonferroni correction) showed that the absolute value of the CSR for exotropes after surgery was significantly smaller than that before surgery along the radial axis (by 0.12, t = −2.31, P = 0.042). In the esotropia group, a repeated-measures ANOVA revealed that the effect of the interaction between surgical operation and axis was significant (F_{1, 13} = 13.17; P = 0.003), with a pairwise post hoc comparison (with Bonferroni correction) showing that the absolute value of the CSR after surgery was significantly smaller than that before surgery along the radial axis (by 0.12, t = −2.92, P = 0.011; Fig. 5). Moreover, a single-sample t-test showed that the CSR at the radial axis was significantly different from zero for both groups before surgery (EXO: t = −6.39, P < 0.001; ESO: t = 4.22, P < 0.001) but not after surgery (EXO: t = −0.88, P = 0.398; ESO: t = 0.37, P = 0.719). However, no significant difference in the CSR from zero was found at the tangential axis for both groups regardless of surgery (all P > 0.05). 

Figure 5.

 

Boxplots of the CSR of exotropia and esotropia groups before and after surgery. The solid line within each box represents the median. The box represents the interquartile range (IQR) of the data (75th to the 25th percentile, e.g., Q3–Q1), and the data points with red crosses represent outliers. Outliers were defined as values less than Q1 – 1.5 × IQR or greater than Q3 + 1.5 × IQR. The gray dashed line indicates a CSR of 0. A CSR of <0, 0, and >0 indicates greater crowding effects in the temporal hemifield, symmetrical crowding effects across visual fields, and greater crowding effects in the nasal hemifield, respectively. To avoid interference with the results, outliers were excluded from the statistical analysis. A pairwise post hoc comparison (with Bonferroni correction) was performed to test if the CSR differed significantly before and after strabismus surgery, and black asterisks above indicate a statistically significant difference: *P < 0.05. A single-sample t-test was conducted to examine whether the CSR was significantly different from zero, and green asterisks below indicate significant differences from zero: ***P < 0.001.

Correlations Between the CSR and Recovery of Stereopsis After Axial Alignment

Due to long-term ocular misalignment, most of our individuals had a loss of stereopsis. About 3 months after surgical correction, we found that 15 strabismic individuals (out of 27) partly recovered their near stereopsis. Since the change of the CSR was mainly exhibited in the radial axis, we compared the preoperative CSR at the radial axis between the individuals who manifested postoperative stereopsis recovery (n = 15) and individuals who did not (n = 12). The results showed that the absolute value of preoperative CSR of individuals with postoperative stereopsis recovery was significantly smaller than that of individuals without postoperative stereopsis recovery (t = 4.22, P < 0.001; Fig. 6A), indicating that severer preoperative nasoemporal asymmetry of visual crowding effects may predict poorer postoperative stereopsis recovery. We further analyzed the correlation between the stereopsis of individuals with postoperative stereopsis recovery and their corresponding postoperative CSR at the radial axis and found that there was a significant negative correlation between the reciprocal of near stereopsis and the absolute value of postoperative CSR (r = −0.55, P = 0.034; Fig. 6B). In other words, the more balanced the crowding effects between the nasal and temporal hemifields after surgery (i.e., the absolute value of the CSR approaching 0), the better the postoperative stereopsis of the individual. 

Figure 6.

 

(A) Comparisons of the absolute value of the preoperative CSR between individuals with postoperative stereopsis recovery (n = 15) and individuals without postoperative stereopsis recovery (n = 12). Error bars are ±1 SEM. Statistically significant difference: ***P < 0.001. (B) The relationship between the reciprocal of near stereopsis and the absolute value of the postoperative CSR. Higher values on the x-axis indicate less nasotemporal balance of crowding effects after strabismus surgery, and higher values on the y-axis indicate better stereopsis. *Statistically significant linear correlation (P < 0.05).

Fixation Stability and Nasotemporal Bias of Fixation Eye Movements Before and After Surgical Alignment

Some studies have reported increased fixation instability in patients with strabismus.^9,49,51,52 To figure out whether strabismus surgery can improve the patient's fixation stability and thus reduce the visual crowding effects, we calculated and compared the logBCEA (deg²) of individuals before and after surgical operation. There was no significant difference in the logBCEA before and after surgery for both exotropia (preoperative: 0.41 ± 0.13 vs. postoperative: 0.46 ± 0.20; t = 0.73, P = 0.47, two-tailed paired samples t-test) and esotropia groups (preoperative: 0.48 ± 0.19 vs. postoperative: 0.46 ± 0.15; t = 0.27, P = 0.79; Fig. 7B). We further analyzed the distribution of fixation points in the nasal and temporal visual fields during the test to eliminate the influence of oculomotor bias on the crowding effects detection. The results showed that the proportion of fixation points in the nasal and temporal hemifields was roughly similar and symmetrical whether before or after the operation (all P > 0.05, two-tailed paired samples t-test; Fig. 7C). 

Figure 7.

 

(A) Representation of the fixation on the visual field of a representative individual with the plot of the BCEA covering 68% of the fixation points. (B) Comparisons of the logBCEA of exotropia and esotropia groups before and after surgery. Higher values on the y-axis signify worse fixation stability. Error bars are ±1 SEM. (C) Comparisons of the proportion of fixation points between the nasal and temporal hemifields, as well as before and after surgery in different groups. Higher values on the y-axis signify greater fixation tendency. Error bars are ±1 SEM.

Discussion

In the current study, secured by a real-time eye tracker to fulfill gaze-contingent display, we found that optical realignment (i.e., strabismus surgery) could normalize the enlarged visual crowding effects, a reliable psychophysical index of cortical organization, in the peripheral visual field of older children and adults with strabismus and rebalance the nasotemporal asymmetry of crowding along the radial axis, promoting the recovery of postoperative stereopsis. Our results indicated a potential of experience-dependent cortical organization after axial alignment even for older children and adults who are out of the critical period of visual development. 

Previous neurophysiologic and brain imaging studies have suggested that crowding is a cortical phenomenon and may originate in the early visual cortex (i.e., V1), with its perceptual consequence influenced by the downstream visual areas (i.e., V2, V3, V4, or higher-level visual areas).^{14,22,24,53,54} In this study, we observed a hemifield- and radial- specific reduction of crowding effects across the visual fields after strabismus surgery. This kind of anisotropic change in the crowding effects is consistent with the horizontal deviation of the eye, and we postulate that this may be due to plasticity at the cortical level.^32,55 The transition to adulthood entails a reduction in the plastic potential of the visual cortex.^56,57 However, growing experimental studies in animal models and humans have shown that physiologic plasticity can be enabled in the adult cortex as well.^58,59 By selecting a group of older children and adults with long-term strabismus history and relatively large angles of misalignment, our findings suggest that the visual system of adulthood is still capable of reorganizing its sensory coding to adapt to improved visual input. Besides, previous studies have revealed that perceptual learning can reduce crowding effects in the normal periphery, which also imply an experience-dependent plasticity on the cortical substrates of crowding.^16,22,27 On these bases, we think strabismus surgery is more than a cosmetic surgery but also can improve visual functions even in adulthood.^60,61 We do not attribute these improvements to patients’ learning effect of the test task, because first, no feedback was given to patients during the test, and thus it was not so easy to have learning effects. Second, there was an interval of about 3 months between preoperative and postoperative measurement. Third, in our previous study,³³ we have retested the normal control group at the same interval and found that the results were similar and far lower than the improvement of surgery (see Supplementary Fig. S1). 

Although some studies have suggested an increased fixation instability in patients with strabismus as well as nasotemporal bias for motion and perception,^9,62–64 our results of fixation eye movements showed relatively stable fixation stability and nasaotemporal symmetrical fixation before and after strabismus surgery. We also excluded individuals with clinical nystagmus who are proved to be more fixation unstable and less likely to have recovery of stereopsis.⁴⁹ Thus, we also do not attribute the features and improvements of crowding effects to patients’ changes in fixation eye movements. 

What is interesting is that we found a correlation between the restored stereopsis and the extent of nasotemporal balance of crowding effects, indicating adult brain plasticity and signifying the importance of ocular correction. We speculate that the rebalanced binocular input after axial alignment may be conducive to the restoration and improvement of binocular visual functions. It is uncovered that better binocular outcomes are obtained when interocular imbalance is corrected after surgical correction,^65,66 and microstructural changes of visuospatial network tracts may be the reason.³² Whereas stereopsis is one of the key indicators for evaluating visual functions and plays an important role in daily work and life,³ our results suggest that assessment of visual crowding across visual fields has important implications for clinical practice to provide a more comprehensive evaluation of visual functions before and after surgical correction in strabismus, as well as for our understanding of peripheral vision and object recognition. 

There were some limitations of the current study. First, our study included a relatively small sample of patients and a follow-up for only 3 months. Additional investigations with larger samples and followed up by a longitudinal study are needed to help further elucidate the surgical outcomes of crowding effects and other clinical improvements of strabismus. In addition, although our results revealed the functional change of peripheral visual crowding effects of strabismic individuals, the substrate of cortical processing in object recognition following long-term adaptation to ocular deviation as well as axis realignment remains unclear, which merits further examination in future neuroimaging and electrophysiology researches. 

In conclusion, we measured visual crowding, an essential bottleneck on object recognition and a reliable psychophysical index of cortex organization, in older children and adults with strabismus before and after strabismus surgery. We found that the peripheral visual crowding was reduced and became rebalanced between the nasal and temporal hemifields after surgery, associating with a recovery of postoperative stereopsis. Our results indicated strong experience-dependent cortical reorganization after axial alignment and signify the importance of ocular correction for clinical practice. 

Acknowledgments

Supported by the Natural Science Foundation of China Grant NSFC 82271115 (MY), Guangdong Basic and Applied Basic Research Foundation (2023A1515011829) (ZL), Natural Science Foundation of China Grant NSFC 32100864 (F-FY), and Natural Science Foundation of China Grant NSFC 32071056 (C-BH). 

Disclosure: Y. Huang, None; Z. Liu, None; M. Wang, None; L. Gao, None; Y. Wu, None; J. Hu, None; Z. Zhang, None; F.-F. Yan, None; D. Deng, None; C.-B. Huang, None; M. Yu, None