Correlation Between Quantitative Contrast Sensitivity Function and Spherical Equivalent Under Uncorrected and Corrected Conditions in Children with Ametropia

Abstract

Background

This study investigated the correlation between quantitative contrast sensitivity function (qCSF) and spherical equivalent (SE) under uncorrected and corrected conditions in children with ametropia.

Methods

All participants completed comprehensive ophthalmic evaluations, including manifest refraction and qCSF testing performed under uncorrected and corrected visual acuity conditions. The qCSF parameters included the area under the log contrast sensitivity function (AULCSF), contrast sensitivity function acuity, and contrast sensitivity at spatial frequencies of 1.0–18.0 cycles per degree (cpd). The analysis was stratified by age and SE.

Results

A total of 114 eyes from 57 children (29 boys and 28 girls; mean age: 8.56 ± 2.09 years) were included. The mean SE was −0.60 ± 1.04 D (range: −3.125 to +3 D). In children with myopia, the qCSF results were significantly better after refraction correction than those under uncorrected visual acuity and increased linearly with the degree of myopia, whereas in children with hyperopia, the qCSF parameters before and after refraction correction did not differ significantly. Generalized linear modeling identified SE was the strongest independent predictor of uncorrected qCSF performance, whereas age showed minimal influence. The differences in qCSF parameters before and after refractive correction were correlated with the SE.

Conclusions

In children without myopia, uncorrected qCSF closely matched the corrected results supporting its use for screening, whereas in children with myopia, the difference between uncorrected and corrected qCSF increased linearly with the degree of myopia. These findings demonstrate the importance of early optical correction and integrating contrast sensitivity assessment into pediatric eye-health evaluations.

Introduction

Contrast sensitivity function (CSF), evaluating the sensitivity thresholds at different spatial frequencies, provides an accurate measure of functional vision and a method for evaluation of conditions with subjective visual impairment.^1,2 The quantitative CSF (qCSF) test has broadly been used for detecting relevant functional vision changes in many diseases including age-related macular degeneration,^3–8 amblyopia,^9,10 glaucoma,¹¹ and multiple sclerosis.¹² Recent advancements in quantitative CSF (qCSF) testing enable rapid, precise measurements, making it suitable for use in pediatric eye examinations.^13–15

Although qCSF testing has been widely applied as a follow-up tool in ocular diseases such as glaucoma, age-related macular degeneration, and amblyopia, its utility as a vision screening tool in children and in myopia prevention remains underexplored.¹⁶ qCSF performance with refraction correction holds limited value when used as a screening tool as it is equivalent to wearing glasses for a vision test, and therefore does not reflect the true visual state of the child. Conversely, the results of uncorrected qCSF at different spatial frequencies can be used to assess functional visual acuity without refractive correction in children during daily activities, offering valuable information for some children who did not receive myopia correction in the early stage.

Refraction sphere and spherical equivalent (SE) have previously been identified as the major contributing factors of qCSF without refractive correction in children.¹⁷ However, unlike traditional visual acuity tests, the interpretation of uncorrected qCSF values in individual patients remains challenging because they depend on population-based comparisons rather than on absolute reference ranges.¹⁸ This gap hampers the use of uncorrected qCSF as a screening metric in pediatric clinical practice.

Unlike objective optical quality metrics such as the modulation transfer function (MTF), which primarily evaluate the optical performance of the eye, qCSF assesses functional visual performance by reflecting the combined effects of optical quality, retinal function, and neural visual processing. Therefore, qCSF provides clinically meaningful information regarding visual function under real-world viewing conditions. Because the present study aimed to evaluate the functional consequences of refractive error in children rather than optical quality alone, qCSF was selected as the primary outcome measure.

To address this, our study aimed to characterize the relationship between qCSF parameters and SE according to the degree of refractive error under uncorrected and corrected conditions. This approach has the potential to provide a robust framework for interpreting uncorrected qCSF outcomes in children, thus expanding its role as an efficient, sensitive screening tool for detecting early functional visual deficits in pediatric ametropia.

Materials and Methods

Patients

This case series included children aged 4 to 14 years who attended the Eye and ENT Hospital of Fudan University in 2023. Pediatric patients were excluded if they had (1) a history of orthokeratology lens wear, low-dose atropine use, or other related medications, or (2) a history of ocular diseases, surgeries, or trauma; systemic illnesses; or severe psychological or psychiatric disorders.

Ethical approval for this study was obtained from the Ethics Committee of the Eye and ENT Hospital of Fudan University (No. 2020107, date: 1 July, 2021), and the research was conducted in accordance with the principles of the Declaration of Helsinki. Written informed consent was obtained from the parents or legal guardian of all participants.

Eye Examinations

Axial length and corneal curvature were measured using the IOL Master 700 (Carl Zeiss Meditec, Germany). Cycloplegia was induced by administering a total of five applications of tropicamide phenylephrine eye drops (Mydrin-P ophthalmic solution; Santen, Osaka, Japan) at 5-minute intervals. Thirty minutes after the final instillation, cycloplegia was confirmed by the absence of accommodation and pupillary light reflex before refraction measurements were obtained. The RT-5100 phoropter (Nidek Technologies, Japan) was used to determine manifest refraction and evaluate corrected distance visual acuity under complete cycloplegia.

Quantitative Contrast Sensitivity Function Test

A visual stimulus consisting of Sloan-font numbers (0–9) was presented on an NEC P403 display (Gension & Waltai Digital Video System Co., Ltd., China). The monitor had a resolution of 1920 × 1080 pixels, a display area of 116.84 × 77.89 cm, a peak brightness of 700 cd/m², a standard brightness of 550 cd/m², and a contrast ratio of 4000:1.

The qCSF test used a Bayesian adaptive algorithm to present 25 stimuli across varying spatial frequencies, with each stimulus comprising three numerals shown at different contrast levels. Testing was performed under full cycloplegia, both without and with refractive correction. Participants observed the screen from a 3-meter distance in a mesopic lighting condition. Each eye was evaluated individually while the other eye was occluded. Prior to formal testing, all participants completed a brief practice session to ensure understanding of the testing procedure. Participants were instructed to report any digits they could recognize, and responses, categorized as correct, incorrect, or none, were immediately recorded by the examiner.

The system automatically generated the CSF curve. Output parameters included the area under the log CSF curve (AULCSF), CSF acuity, and contrast sensitivity values (in logarithmic units) across six spatial frequencies: 1.0, 1.5, 3.0, 6.0, 12.0, and 18.0 cycles per degree (cpd). The AULCSF reflects the overall spatial visual capacity, whereas CSF acuity and frequency-specific sensitivity values indicate high-resolution visual performance under general and targeted visual tasks, respectively.

Statistical Analysis

Continuous variables were expressed as means, standard deviations, and ranges, whereas categorical variables were reported as frequencies. The Kolmogorov–Smirnov test was used to assess data normality, and Levene’s test was used to evaluate homogeneity of variance. Stratified analyses were performed with participants stratified according to age (4–8 years and 9–14 years) to assess the effect of earlier myopia onset,¹⁹ spherical equivalent (high SE [< −1 D], low SE [−1.0 to 0 D], and hyperopia [>0 D]) because −1.00 D is an indication for prescribing spectacles. For normally distributed continuous variables, paired t-tests and one-way ANOVA were used for comparisons between two groups and multiple groups, respectively. The Bonferroni correction was applied when performing multiple post-hoc comparisons. The Wilcoxon rank-sum test was used for comparisons between two groups of non-normally distributed continuous variables. Pearson’s correlation coefficient was used to assess the correlation between continuous variables. A generalized linear model (GLM) was used to assess the influence of multiple factors, accounting for covariates such as age, sex, ocular biometric measurements, and inter-eye inclusion effects.

All statistical analyses were performed using SPSS version 25.0 (IBM Corp, Armonk, NY, USA). Statistical significance was set at P<0.05.

Results

Patient Characteristics

The study included 57 participants (29 boys and 28 girls) aged 4–14 years (mean: 8.56 ± 2.09 years). The participants had a mean axial length of 23.72 ± 0.79 mm, mean SE of −0.6 ± 1.04 D, and mean corrected distance visual acuity of 0.01 ± 0.03 LogMAR. The participant characteristics are summarized in Table 1, and the age and SE distributions are shown in Figure 1A.

Table 1.

Patient Demographics and qCSF (Quantitative Contrast Sensitivity Function) Results

Characteristic	Mean±SD	Range
Age (years)	8.56±2.09	(4, 14)
Gender (male/female)	29/28
Axial length (mm)	23.72±0.79	(21.67, 25.24)
RS(D)	−0.36±1.05	(−3.00, 3.50)
RC(D)	−0.48±0.57	(−3.75, 0)
SE(D)	−0.6±1.04	(−3.125, 3)
CDVA (LogMAR)	0.01±0.03	(0, 0.2)
K-flat (D)	42.86±1.35	(40.25, 46.23)
K-steep (D)	43.94±1.99	(41.00, 47.74)

Abbreviations: RS, Refraction Sphere; RC, Refraction Cylinder; SE, Spherical equivalent; CDVA, Corrected Distance Visual Acuity.

Figure 1.

Subject characteristics of age, refraction, and qCSF. (A) Distribution of age and spherical equivalents. (B) Mean contrast sensitivity (log units) at different spatial frequencies (left panel). Mean area under the line of the contrast sensitivity function (AULCSF) and contrast sensitivity function (CSF) acuity (right panel).

Comparison of Uncorrected and Corrected qCSF Parameters

Compared with uncorrected visual acuity, the qCSF parameters improved significantly with refraction correction (Figure 1B). The AULCSF increased from 0.63 ± 0.28 to 0.91 ± 0.17 (mean Δ: 0.29 ± 0.31, p < 0.001), and CSF acuity improved (mean Δ: 7.67 ± 8.00, p < 0.001). All contrast sensitivity (CS) values tested from 1.0 to 18.0 cpd showed statistically significant enhancements (p < 0.001; Table 2).

Table 2.

QCSF (Quantitative Contrast Sensitivity Function) Results in All Participants with Uncorrected Acuity and Refraction Correction, Respectively

Characteristic	Uncorrected	Corrected	Δ	P (paired-t)	ICC	p
AULCSF	0.63±0.28	0.91±0.17	0.29±0.31	<0.001	0.119	0.102
CSF Acuity	12.54±6.71	20.21±5.64	7.67±8.00	<0.001	0.167	0.037
CS (1.0 cpd)	1.04±0.17	1.11±0.12	0.07±0.19	<0.001	0.194	0.019
CS (1.5 cpd)	1.03±0.18	1.14±0.11	0.11±0.18	<0.001	0.182	0.026
CS (3.0 cpd)	0.88±0.29	1.11±0.14	0.23±0.32	<0.001	0.044	0.320
CS (6.0 cpd)	0.52±0.39	0.90±0.20	0.38±0.42	<0.001	0.083	0.190
CS (12.0 cpd)	0.18±0.27	0.45±0.24	0.27±0.32	<0.001	0.198	0.017
CS (18.0 cpd)	0.05±0.13	0.16±0.18	0.10±0.20	<0.001	0.193	0.019

Note: Values with statistical significance are shown in bold.

Abbreviations: AULCSF, Area Under Log CSF; CS, Contrast Sensitivity; Uncorrected, qCSF test with uncorrected acuity; Corrected, qCSF test with refraction correction.

Group Analysis

Table 3 and Figure 2 show the stratified comparisons of qCSF parameters according to age and SE. In both age groups, children with refraction correction showed significantly higher values than with uncorrected visual acuity, except for the change in CS at 1.0 cpd in the 9–14 years group which was borderline significant (p=0.053).

Table 3.

Comparison of qCSF (Quantitative Contrast Sensitivity Function) Readings in Different Groups Analyzed by Generalized Linear Model

Groups			N	AULCSF	CSF Acuity	CS (1.0 cpd)	CS (1.5 cpd)	CS (3.0 cpd)	CS (6.0 cpd)	CS (12.0 cpd)	CS (18.0 cpd)
Uncorrected	Age (years)	4~8	58	0.68±0.29	13.15±7.10	1.05±0.15	1.06±0.17	0.94±0.31	0.58±0.41	0.22±0.31	0.08±0.17
		9~14	56	0.58±0.25	11.91±6.29	1.04±0.19	1.00±0.18	0.82±0.26	0.45±0.36	0.14±0.21	0.03±0.08
	SE (D)	>0	27	0.83±0.26	15.61±6.99	1.05±0.14	1.11±0.11	1.07±0.21	0.79±0.38	0.34±0.34	0.11±0.20
		−1~0	41	0.70±0.23	13.99±6.10	1.10±0.14	1.10±0.14	0.96±0.20	0.60±0.34	0.20±0.25	0.06±0.12
		<-1	46	0.44±0.20*	9.45±5.87*	0.99±0.20	0.92±0.18*	0.69±0.30*	0.28±0.30*	0.06±0.15*	0.02±0.07
Corrected	Age (years)	4~8	58	0.90±0.17	19.54±5.34	1.11±0.11	1.14±0.11	1.10±0.15	0.89±0.20	0.42±0.24	0.14±0.16
		9~14	56	0.93±0.17	20.91±5.90	1.11±0.13	1.14±0.10	1.11±0.13	0.90±0.20	0.48±0.24	0.18±0.19
	SE (D)	>0	27	0.88±0.15	17.88±4.99#	1.08±0.11	1.12±0.10	1.09±0.10	0.87±0.18	0.38±0.25	0.09±0.17
		−1~0	41	0.89±0.18	19.93±5.51	1.10±0.13	1.13±0.11	1.09±0.16	0.87±0.22	0.43±0.23	0.13±0.16
		<-1	46	0.95±0.17	21.83±5.71	1.14±0.11	1.16±0.10	1.14±0.14	0.94±0.20	0.51±0.24	0.22±0.18

Notes: Values with statistical significance are shown in bold. There are statistical significances (p<0.05) between Uncorrected vs Corrected in different age and SE groups (Except for Age group [9~14] in CS at 1.0cpd, p=0.053; and SE group [>0] in all qCSF parameters, p>0.05, which are shown in italic). *, vs the other two SE groups, p <0.05. #, vs SE group (<-1), p <0.05.

Abbreviations: N, Number of eyes; D, Diopter; SE, Spherical Equivalent; AULCSF, Area Under Log CSF; Uncorrected, qCSF test with uncorrected acuity; Corrected, qCSF test with refraction correction.

Figure 2.

Group analyses of qCSF parameters according to age and SE. (A) Differences in qCSF parameters among age groups. (B) Differences in qCSF parameters in groups with spherical equivalents. *, vs The other two SE groups, p <0.05. #, vs SE group (<-1), p <0.05.

In the analysis stratified by SE, the group of SE < −1 D had the lowest mean AULCSF (0.44 ± 0.20), CSF acuity (9.45 ± 5.87) and CS at 1.5 to 12.0 cpd under uncorrected visual acuity, and showed the most substantial improvement following refractive correction, with the mean change in CSF acuity (21.83 ± 5.71) significantly exceeding that of the hyperopic group (SE > 0 D, 17.88 ± 4.99; p < 0.05). In contrast, children in the hyperopic group (SE > 0 D) showed only modest changes in qCSF parameters after refractive correction, none of which were statistically significant.

Correlation Analysis

The associations between qCSF parameters and other factors, including age and SE, are shown in Table 4. GLM analyses demonstrated that SE was the major factor contributing to the change in qCSF parameters under uncorrected visual acuity, whereas age had limited impact. The Pearson correlation analyses between SE and qCSF parameters are presented in Figure 3. AULCSF, CSF acuity, and all spatial frequencies were significantly correlated with SE under uncorrected visual acuity, indicating that higher degree of myopia was associated with poorer visual function. However, these relationships were not statistically significant after refractive correction.

Table 4.

Association of qCSF (Quantitative Contrast Sensitivity Function) Parameters with Other Factors Analyzed by Generalized Linear Model

Factors	qCSF	Uncorrected		Corrected		Δ
		B	p	B	p	B	p
Age	AULCSF	0.011	0.378	−0.002	0.776	−0.013	0.319
	CSF Acuity	0.315	0.314	0.017	0.949	−0.298	0.417
	1.0cpd	0.013	0.128	−0.004	0.460	−0.017	0.056
	1.5cpd	0.012	0.100	−0.003	0.545	−0.015	0.046
	3.0cpd	0.009	0.464	−0.002	0.814	−0.011	0.425
	6.0cpd	0.012	0.509	−0.004	0.707	−0.015	0.410
	12.0cpd	0.009	0.473	−0.001	0.963	−0.009	0.525
	18.0cpd	−0.002	0.764	0.000	0.982	0.002	0.819
SE	AULCSF	0.0131	<0.001	−0.013	0.416	−0.144	<0.001
	CSF Acuity	2.164	<0.001	−0.680	0.197	−2.844	<0.001
	1.0cpd	0.047	0.003	−0.018	0.118	−0.065	<0.001
	1.5cpd	0.091	<0.001	−0.015	0.129	−0.106	<0.001
	3.0cpd	0.119	<0.001	−0.014	0.289	−0.133	<0.001
	6.0cpd	0.159	<0.001	−0.015	0.442	−0.174	<0.001
	12.0cpd	0.090	<0.001	−0.014	0.546	−0.103	<0.001
	18.0cpd	0.028	0.023	−0.024	0.143	−0.053	0.004

Note: Values with statistical significance are shown in bold.

Abbreviations: AULCSF, Area Under Log CSF; CS, Contrast Sensitivity; SE, Spherical equivalent; Uncorrected, qCSF test with uncorrected acuity; Corrected, qCSF test with refraction correction.

Figure 3.

Correlations between SE and qCSF parameters. Pearson linear regressions between SE and (A) AULCSF, (B) CSF acuity, and CS (log units) at (C) 1.0 c/d, (D) 1.5 c/d, (E) 3.0 c/d, (F) 6.0 c/d, (G) 12.0 c/ d, and (H) 18.0 c/d. The solid lines illustrate the line of linear regressions, and the dotted line illustrates the confidential intervals. Top left (with refraction correction) and bottom right (with uncorrected acuity): equation of linear regression and P values.

The correlations between SE and the differences in qCSF parameters before and after refraction correction are shown in Figure 4. The changes in parameters were significantly positively correlated with SE (p < 0.01), indicating that children with higher degrees of myopia experienced greater visual improvements after refractive correction. Specifically, the strongest effects were observed between 3.0 and 12.0 cpd.

Figure 4.

Correlations between SE and the difference in qCSF parameters (uncorrected – corrected). Pearson linear regressions between SE and (A) AULCSF, (B) CSF acuity, and CS (log units) at (C) 1.0 c/d, (D) 1.5 c/d, (E) 3.0 c/d, (F) 6.0 c/d, (G) 12.0 c/ d, and (H) 18.0 c/d. The solid lines illustrate the line of linear regressions, and the dotted line illustrates the confidential intervals. Top left: equation of linear regression and P values.

Discussion

This study provides preliminary evidence regarding the potential clinical applicability of qCSF testing in pediatric refractive assessment. In children with SE > 0 D, the results of the uncorrected and corrected qCSF did not differ significantly. Conversely, in children with SE < −1 D, the magnitude of improvement in qCSF after correction increased linearly with refractive severity, establishing a quantitative link between SE and functional vision impairment. These results support a more individualized approach to functional vision assessment in children.

This study demonstrates that uncorrected myopia disproportionately impairs contrast sensitivity in children with ametropia. Notably, the greatest deficits in uncorrected vision were found at mid-to-high spatial frequencies (3.0–12.0 cpd), which are essential for critical visual tasks such as reading and face recognition and are often overlooked by conventional acuity tests.^20,21 These frequencies are particularly susceptible to optical defocus in myopic eyes, and their restoration after correction demonstrates the clinical value of refractive intervention. These findings align with previous research demonstrating that uncorrected refractive errors substantially impair functional vision, and that appropriate optical correction enhances contrast sensitivity and visual performance, especially in children with amblyopia.²²

The stratified analysis revealed that the impact of refractive correction on qCSF varied according to age group and refractive condition. Although younger children (4–8 years) showed consistent improvements across all parameters after correction, older children (9–14 years) demonstrated slightly lower gains in low spatial frequency contrast sensitivity (1.0 cpd), suggesting a potential developmental influence. This aligns with previous studies that suggest that the improvement in contrast sensitivity with age is limited. Dekker et al²³ reported an average improvement of approximately 0.3 log units in CS from age 4 to 18 years, with 90% of this change completed by the age of 12 years.

Regarding refractive status, children with SE < −1D exhibited the lowest qCSF parameters under uncorrected visual acuity, particularly in AULCSF and CSF acuity. However, after refractive correction, their CSF acuity significantly improved, even exceeding that of the hyperopic group. In contrast, children with hyperopia showed no statistically significant improvement in qCSF parameters after refractive correction. These findings indicate that optical correction effectively restores contrast sensitivity in myopic eyes and that qCSF parameters are sensitive to subtle variations in visual function across refractive groups, confirming that refractive error is not only a cause of reduced visual acuity but also of impaired functional vision, especially in tasks requiring fine contrast discrimination. Therefore, beyond traditional visual acuity assessments, contrast sensitivity testing such as qCSF can serve as a valuable tool to guide clinical decisions regarding spectacle prescription. Early correction not only improves visual clarity but may also help reduce visual strain and enhance the child’s ability to engage fully in academic and daily activities.

The correlation analysis revealed a linear correlation between uncorrected qCSF parameters and SE. GLM analysis revealed that SE strongly predicted qCSF with uncorrected visual acuity. Notably, this linear association was present both before and after refractive correction, with greater improvements in qCSF parameters observed in children with higher myopia. For example, the difference in CSF acuity before and after refraction correction increased by 2.737 units per dioptre of myopia, demonstrating the quantitative dependence of functional vision deficits on refractive status.

These findings have important clinical implications. The linearity between uncorrected qCSF and SE suggests that uncorrected qCSF measurements alone could serve as a rapid screening tool for estimating visual impairment severity in children with SE < −1D. For example, in resource-limited settings in which cycloplegic refraction is unavailable, a quick qCSF test under uncorrected conditions can provide some information of the refractive state and its functional impact.

This study established that SE and the difference between uncorrected and corrected qCSF values are linearly correlated, thus providing a practical framework for clinical application. By referencing the corrected qCSF values in different refractive states, clinicians can estimate the likely range of uncorrected qCSF values corresponding to the SE. This enables the direct comparison of a patient’s measured uncorrected qCSF with a normative range, facilitating rapid and individualized assessment of children’s functional vision. Such an approach eliminates the need for refraction correction during screening and enhances the efficiency and scalability of qCSF-based vision screening in pediatric populations.

This study has some limitations. First, this was a cross-sectional study with a relatively modest sample size and no a priori sample size calculation. Although statistically significant associations were observed, larger multi-center studies are required to confirm the generalizability of our findings. Second, cycloplegic refraction was performed using a tropicamide-phenylephrine regimen rather than atropine, which may have introduced a small degree of measurement bias. Third, variability in understanding and test performance among younger children may have influenced the measurements. Future studies incorporating repeated testing or objective reliability assessments in pediatric populations would further strengthen the robustness of the findings. Fourth, qCSF reflects the integrated effects of optical quality, retinal function, and neural visual processing, but it does not directly quantify optical quality, as MTF does. Future studies combining qCSF with objective optical quality metrics may provide a more comprehensive understanding of the mechanisms underlying visual function changes associated with refractive errors. Second, the range of refractive errors in the study population was relatively limited. Further studies are needed to assess the characteristics of the qCSF in children with high myopia.

Conclusions

This study reveals a linear relationship between SE and CSF outcomes in children with ametropia. The lack of significant difference between uncorrected and corrected qCSF in those without myopia supports the use of uncorrected qCSF for rapid screening. In contrast, a greater degree of myopia was associated with larger improvements after refraction correction, highlighting the importance of early optical intervention. These results support individualized vision assessment and indicate a need for wider adoption of qCSF in pediatric screening.

Funding Statement

This work was supported by Shanghai Rising-Star Program (23QA1401000); Healthy Young Talents Project of Shanghai Municipal Health Commission (2022YQ015); Shanghai Oriental Talents-Technology Platform Program (QNKJ2024055); Shanghai Hospital Development Center Foundation (SHDC12025144).

Author Contributions

All authors made a significant contribution to the work reported, whether that is in the conception, study design, execution, acquisition of data, analysis and interpretation, or in all these areas; took part in drafting, revising or critically reviewing the article; gave final approval of the version to be published; have agreed on the journal to which the article has been submitted; and agree to be accountable for all aspects of the work.

Disclosure

The authors declare no conflicts of interest.