Introduction

Amblyopia is a neurodevelopmental disease characterized by decreased best-corrected visual acuity (BCVA) in one or both eyes, which is due to a discordant visual experience (strabismus, anisometropia, high refractive error, or deprivation) during the critical period of visual development1. It is the most primary cause of monocular visual acuity (VA) impairment in children, affecting approximately 3–5% of the population2. In addition to VA, it also impairs binocular sensory perception, contrast sensitivity, positional certainty, and depth perception2. These deficits may be correlated with abnormal changes of visual pathway in amblyopia3.

Children with amblyopia have defects not only in visual function such as BCVA and stereoacuity, but also in behavioral performances on tasks involving visual selective attention4, cognitive processing5and simple perceptual processing6. Deficits in the response time for complex visual decisions have also been reported in adults with amblyopia, suggesting impaired higher-order perceptual performance7. However, the visual selective attention and numerical processing ability in children with anisometropic amblyopia have not been fully studied.

Numerical processing refers to the individual processing of digital stimuli, forming corresponding numerical psychological representations, and using them to perform related cognitive processing, such as quantity comparison and mathematical operations8. According to the eye-mind and immediacy assumptions, the ___location of an eye fixation and its duration can separately reflect the spatial locus and time course of numerical processing8. Therefore, the experimental data for tracking participants’ eye-fixation behavior, such as reaction time (RT), contribute to quantitatively evaluate their numerical processing ability. In addition, numerical processing has multiple effects, including the numerical size effect and numerical distance effect. Numerical size effect is characterized by the easier comparisons with smaller numbers for a constant numerical distance (e.g., 1–5 vs. 5–9). Numerical distance effect refers to decreased response time and error rates with increased numerical distance when comparing the magnitudes of the two numbers (e.g., 1–5 vs. 4–5)9. It required combined participation of numerical processing and visual selective attention.

Numerical processing involves visual selective attention ability. The orienting, division, and adjustment of visual selective attention can immediately affect numerical processing. Visual selective attention refers to the selective processing of only one stimulus while neglecting other stimuli in the presence of two or more stimuli, developing continuously in children10. Moreover, studies on functional brain images have revealed that attentional mechanism works at multiple stages in the visual system so visual selective attention is determined by different visual processing abilities at each stage, such as the lateral geniculate nucleus, V4, the frontal cortex and the parietal cortex11. Previous research revealed the impact of amblyopia on binocular visual attention and found that children with amblyopia may have deficits in the attentional pursuit of multiple moving objects when filtering out interfering substances12. While a study on visual selective attention of patients with amblyopia has shown that neither VA nor binocular function measures are significantly related to visual selective attention13. However, the differences in visual selective attention between children with anisometropic amblyopia and those who have recovered from anisometropic amblyopia have not yet been reported.

The current gold standard of successfully recovered amblyopia is full recovery of BCVA in the amblyopic eye, with the criterion of at least 20/25. The most common treatment for amblyopia is patching, while it is not the most ideal way to restore binocularity and cortical function in amblyopic patients14. And this limitation may result in remaining deficiency of a number of visual functions, visual selective attention and numerical processing in patients whose visual acuity have reached to a normal level15. Besides visual acuity, the improvement of behavioral performances also needs further exploration in recovered amblyopic patients.

Hence, the effect of anisometropic amblyopia on visual selective attention and numerical processing in children requires further study, and whether the impairment recovers with improved BCVA in children with anisometropic amblyopia remains unclear. Therefore, we conducted a study with three tasks separately, including number locating task, number comparing task and specific number comparing task, in children aged 6–8 and 9–11 to discover their visual selective attention and numerical processing ability. Differences in the performances on these tasks and the numerical distance effect were compared among children with anisometropic amblyopia, those who had recovered from anisometropic amblyopia, and children with normal vision. The results would provide more insight into the impairment and recovery of behavioral performances in children with anisometropic amblyopia.

Methods

Participants

Forty-five Chinese children aged 6–8 and 45 Chinese children aged 9–11 were recruited for this study. Children of both age groups were divided into amblyopia, recovered amblyopic, and control groups, with 15 subjects in each group. The amblyopia and recovered amblyopic groups were recruited from Beijing Tongren Hospital. The amblyopia group included anisometropic amblyopia children with interocular difference of spherical equivalent ≥ 1.50 diopter, BCVA of > 0.1 logMAR (20/25) in the amblyopic eye and an inter-eye acuity difference of two or more lines. The inclusion criteria for the recovered amblyopic group were history of anisometropic amblyopia and normal VA after treatment (at least 0.1 logMAR, 20/25 of BCVA in each eye) with an inter-eye difference of one line or less. The control group comprised age-matched healthy children with normal VA recruited from the Xuanwu District Youth Science and Technology Museum in Beijing, China. All participants were right-handed, without mental, neurological, or intellectual disorders, and excluded from strabismus and ocular structural disorders.

The experiment was conducted in accordance with the requirements of the Beijing Tongren Hospital Ethics Committee. Informed consent, consistent with the tenets of the Declaration of Helsinki, was obtained from all participants or their guardians before the experiment.

Visual function assessment

All participants underwent refraction and VA tests. BCVA was scored in logMAR units. Slit-lamp and ocular fundus examinations were performed to exclude organic lesions. The cover tests were taken to exclude strabismus. The recovery time was calculated from when the BCVA of the children with anisometropic amblyopia was enhanced to normal to the testing date when they participated in the experiment.

Tasks and procedure

The experiment consisted of three visual attention tasks with stimuli of eight numbers: 1, 2, 3, 4, 6, 7, 8, and 9 (Times New Roman font, 96 font size). Using the E-prime software (Psychology Software Tools, Pittsburgh, PA), all stimuli were presented at the center of a 17-inch standard computer screen measuring 1024 × 786 pixels with the illumination of the screen was set to 80.

The experiment was conducted in a single-personal soundproof laboratory with room temperature at 23 °C and constant room illumination. Before the experiment, each participant was given a practice block. The formal test was started only when the correct rate exceeded 90%. At the beginning of the experiment, the center of the screen presented a black “+” gaze point for 400 ms, followed by two boxes on the left and right sides of the screen with stimulation of “∗” appearing randomly in each box for 500 ms (Fig. 1).

In tasks 1 and 2, after the “∗” was eliminated, a black number appeared randomly on the right or left side and remained in view until the participant pressed a key (but not for more than 3,000 ms). In task 1, the participants were required to identify the ___location of number on the screen and press the button on the same side of the number as quickly and correctly as possible while avoiding errors. In task 2, participants needed to compare the magnitude of number and pressed the left button when the number was smaller than 5 and pressed the right button when the number was larger than 5. A blank screen was presented for 500 ms after each trial, with 96 trials per task (Fig. 1A).

For the third task, two numbers (red and green) appeared randomly in the left and right boxes for 3,000 ms after the “∗” was eliminated. Participants were required to compare the magnitude of specific colored number and press the left button when the red number was smaller than 5 and the right button when it was larger than 5. The next trial started after a blank screen for 500 ms for 109 trials (Fig. 1B). The RT was measured from the stimuli appearance until the button was pressed.

Fig. 1
figure 1

Flow chart of three visual attention tasks. A black “+” gaze point presented for 400 ms, then stimulation of “∗” appeared randomly in the right or left box for 500 ms. In task 1 and task 2 (A), a number (1, 2, 3, 4, 6, 7, 8, and 9) appeared randomly in the right or left box; in task 3 (B), two numbers (red and green) appeared randomly in the two boxes. The stimuli remained for 3,000 ms or until the participant pressed a button, then a blank screen was presented for 500 ms.

Full size image

Statistical processing

The data followed a normal distribution and were homoscedastic. Analysis of variance (ANOVA) was used to compare group differences in RT. The numerical processing ability was reflected by difference in RT of task 2 and 1, with smaller difference representing better ability. The visual selective attention was reflected by difference in RT of task 3 and 2, with smaller difference representing better attention. According to the distance from each number to 5, the stimuli in task 2 were divided into two types: far numbers (1, 2, 8, and 9) and close numbers (3, 4, 6, and 7). RT for close and far numbers were compared by t-tests to assess numerical distance effect. The distance effect size is operationalized as the difference in RT for close and far numbers, and compared using ANOVA and t-tests. The educational background was calculated by the grades of participants. All statistical analyses were performed using SPSS version 22.0 statistical software. Differences were considered statistically significant at P < 0.05.

Results

Demographic

The total sample included 45 children aged 6–8 years and 45 children aged 9–11. The 6–8 years group included 15 children with anisometropic amblyopia (7.2 ± 0.7 years, 4 females), 15 who had recovered from anisometropic amblyopia (7.2 ± 0.7 years, 6 females), and 15 with normal vision (7.5 ± 0.5 years, 3 females). The 9–11 years group included 15 children with anisometropic amblyopia (9.7 ± 0.7 years, 5 females), 15 who had recovered from anisometropic amblyopia (9.7 ± 0.9 years, 6 females), and 15 with normal vision (9.9 ± 0.7 years, 4 females). The VA of recovered amblyopic groups had reached to normal level at the age of 6.5 ± 1.3 years in children aged 6–8 and 8.6 ± 1.7 years in children aged 9–11 (t = 3.76, P < 0.001). Children in recovered amblyopic groups had recovered for 7.4 ± 11.4 months in children aged 6–8 and for 17.3 ± 16.8 months in children aged 9–11 (t = −1.89, P = 0.07). All children were carried in full-term pregnancies and had no known neurological, intellectual, or ocular structural disorders. In both age group, there were no significant differences in age, sex and educational background among amblyopia, recovered amblyopic, and control groups (Table 1). All the participants completed the three tasks with an accuracy of over 90%.

Table 1 Clinical characteristics of subjects.
Full size table

Reaction time of visual attention tasks

In both age groups, one-way ANOVA indicated that RTs of all three tasks showed statistically significant differences between the three groups (6–8 years, task 1: F[2, 42] = 9.09, P = 0.001, task 2: F[2, 42] = 6.17, P = 0.004, task 3: F[2, 42] = 8.49, P = 0.001; 9–11 years, task 1: F[2, 42] = 18.60, P < 0.001, task 2: F[2, 42] = 9.40, P < 0.001, task 3: F[2, 42] = 16.66, P < 0.001). The visual selective attention also differed between the three groups (6–8 years, F[2, 42] = 3.58, P = 0.04; 9–11 years, F[2, 42] = 6.06, P = 0.01). In contrast, the numerical processing ability had no significant differences between the three groups (6–8 years, F[2, 42] = 1.32, P = 0.28; 9–11 years, F[2, 42] = 0.47, P = 0.63) (Fig. 2).

The post hoc tests revealed that in both age groups, amblyopia groups demonstrated significantly longer RTs of all three tasks than control groups (6–8 years, task 1 P < 0.001, task 2 P = 0.01, task 3 P = 0.001; 9–11 years, task 1 P < 0.001, task 2 P < 0.001, task 3 P < 0.001). The recovered amblyopic groups also showed longer RTs of all three tasks than control groups (6–8 years, task 1 P = 0.002, task 2 P = 0.002, task 3 P = 0.001; 9–11 years, task 1 P < 0.001, task 2 P = 0.001, task 3 P < 0.001). Compared with amblyopia groups, the recovered amblyopic groups showed no significant differences in RTs of all three tasks (6–8 years, task 1 P = 0.58, task 2 P = 0.44, task 3 P = 0.78; 9–11 years, task 1 P = 0.27, task 2 P = 0.93, task 3 P = 0.44). In addition, in children aged 6–8 (Fig. 2A), only amblyopia group showed significantly worse visual selective attention than control group (P = 0.01); however, there were no differences in visual selective attention between recovered amblyopic group and control group (P = 0.29). In children aged 9–11 (Fig. 2B), both amblyopia (P = 0.002) and recovered amblyopic groups (P = 0.02) showed significantly worse visual selective attention than control group (Fig. 2).

Fig. 2
figure 2

Reaction time of three tasks and difference in reaction time between tasks in amblyopia, recovered amblyopic and control groups (mean ± standard deviation). The difference in reaction time of task 2 and 1 indicated numerical processing ability of subjects and the difference in reaction time of task 3 and 2 indicated visual selective attention. *P < 0.05.

Full size image

Numerical distance effect

The stimuli in task 2 was divided into far numbers (1, 2, 8, and 9) and close numbers (3, 4, 6, and 7) according to their distance from 5. In both age groups, RTs for far numbers were significantly shorter than that for close numbers in all three groups (6–8 years, control: t = −6.74, P < 0.001, amblyopia: t = −4.92, P < 0.001, recovery: t = −4.87, P < 0.001; 9–11 years, control: t = −5.35, P < 0.001, amblyopia: t = −4.25, P < 0.001, recovery: t = −5.50, P < 0.001), proving the existence of numerical distance effect in numerical processing in all three groups (Fig. 3).

Fig. 3
figure 3

Reaction time for far numbers (1, 2, 8 and 9) and close numbers (3, 4, 6 and 7) in amblyopia, recovered amblyopic and control groups (mean ± standard deviation). The reaction time for far numbers was shorter than that for close numbers, indicating that numerical distance effect existed. *P < 0.05.

Full size image

The distance effect size is operationalized as the difference in RTs for close and far numbers. Multivariate ANOVA comparisons of age×disorder situation revealed no significant main effect or interaction effect on the distance effect size. Considering the significant difference of cognitive ability between different ages, we further explored the effect of age on the distance effect size. The independent-sample t-tests indicated that children aged 6–8 had greater distance effect size than those aged 9–11 in control group (t = 2.15, P = 0.040); in contrast, the differences were not significant between different ages in the amblyopia (t = 1.11, P = 0.28) and recovered amblyopic groups (t = −0.38, P = 0.71) (Fig. 4).

Fig. 4
figure 4

The distance effect size at different ages in amblyopia, recovered amblyopic and control groups (mean ± standard deviation). The distance effect size was demonstrated by the difference in reaction time for far numbers and close numbers. Children aged 6–8 had a significantly larger distance effect size than those aged 9–11 in control group. *P < 0.05.

Full size image

Discussion

This study revealed that children with anisometropic amblyopia and even those recovered showed significantly longer RTs of all three visual attention tasks than control group. The amblyopia group presented worse visual selective attention than control group. The recovered amblyopic group showed worse visual selective attention than control group only in children aged 9–11 but not in children aged 6–8. The numerical distance effect was validated in all three groups, but only control group exhibited different distance effect size between ages.

In the two age groups, children with anisometropic amblyopia had longer RTs than controls in the three visual attention tasks, which indicated damaged visual attention in children with anisometropic amblyopia. Deficits in the attentional tracking of single and multiple objects in both the amblyopic and fellow eyes have also been reported in children with strabismic and anisometropic amblyopia16. Moreover, children with anisometropic amblyopia also presented worse visual selective attention than controls in both age groups. Another study also found worse visual selective attention in children with amblyopia using different visual attention tasks13. Our findings confirmed the impaired visual selective attention in children with anisometropic amblyopia. While only anisometropic amblyopia was investigated in this study, attentional deficits in strabismic and other types of amblyopia are expected in further research.

Compared to previous studies that only explored the behavioral performance in children with amblyopia, we further investigated visual selective attention performance in children who had recovered from amblyopia. We revealed that children whose BCVA recovered to normal level also had longer RTs of all three tasks than controls in our study, indicating that visual attention deficits remained in children who had recovered from anisometropic amblyopia. Moreover, we found that the recovered amblyopic group showed no significant differences in visual selective attention with control group in children aged 6–8. This result suggests that the impaired visual selective attention in children with anisometropic amblyopia may improve after recovery at younger age. Then, we found recovered amblyopic group showed worse visual selective attention than control group in children aged 9–11, which indicated that older children retained deficit visual selective attention ability after recovery. This demonstrates that the recovery of visual selective attention in children with anisometropic amblyopia is better at younger age than at older age. Younger children with better neurodevelopmental plasticity had faster recovery of their visual selective attention. But older children’s visual selective attention remained defective in the early stages of VA recovery.

The traditional treatment of amblyopia mainly focuses on the VA by forcing the amblyopic eye to work harder to enhance its vision. In addition to VA, binocular visual outcomes17and fine motor skill performance18may also improve after treatments for amblyopia19. However, the influence of traditional amblyopia treatment on visual selective attention has not been reported. Therefore, the treatment for impaired visual selective attention needs to be further studied. This finding provides a reference for the further treatment of anisometropic amblyopia, focusing on not only VA but also performances of visual selective attention.

Furthermore, our study validated the numerical distance effect, which exists in children with normal vision, in children with anisometropic amblyopia and those who had recovered from anisometropic amblyopia. Moreover, our research showed that younger children had a larger distance effect size than older children in the control group. This result is consistent with a previous study illustrating that the decreasing distance effect size with age may be attributed to increasing subjective distance between numerosities on a “mental number line”, where the magnitudes of numbers gradually increase from left to right on this line existing in mind20. However, in children with anisometropic amblyopia and those who had recovered from anisometropic amblyopia, the distance effect size in the older group was not significantly different from that in the younger group. It may due to altered “mental number line” in amblyopia and recovered amblyopic groups, the distance between numerosities on their “mental number line”, especially its developmental change with age, may differ from that in the control group. The altered “mental number line” may explain the absence of age effect on numerical distance effect in amblyopia and recovered amblyopic groups. The corresponding neural network of the “mental number line” involves bilateral parietal-frontal areas, the left supplementary motor area, and the right cerebellum21. And these brain areas were found to have functional changes in patients with amblyopia in previous studies22. Our findings on the altered numerical distance effect in children with anisometropic amblyopia and those who had recovered from anisometropic amblyopia suggest that impaired visual networks in children with anisometropic amblyopia affect not only visual selective attention but also other behavioral performances such as numerical distance effect, and the impact also remain after VA recovery.

In conclusion, this study demonstrates that children with anisometropic amblyopia exhibit not only defective visual selective attention but also altered age-related numerical distance effect. Surprisingly, the rehabilitation of these behavioral performances was not simultaneous with recovery of VA in anisometropic amblyopia, suggesting that further research should focus on the improvement of behavioral performances in children who had recovered from anisometropic amblyopia with longer observation times and the investigation of relevant treatment.